The Journal of Toxicological Sciences (J

The Journal of Toxicological Sciences (J. Toxicol. Sci.) 155 Vol.41, No.1, 155-164, 2016

Original Article Detection of metabolic activation leading to drug-induced phospholipidosis in rat hepatocyte spheroids

Masashi Takagi1, Seigo Sanoh1, Masataka Santoh1, Yoko Ejiri2, Yaichiro Kotake1 and Shigeru Ohta1

1Graduate School of Biomedical and Health Sciences, Hiroshima University, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8553, Japan 2Molding Component Business Department, New business Development Division, Kuraray Co., Ltd., 1-1-3 Otemachi, Chiyoda-ku, Tokyo 100-8115, Japan

(Received September 24, 2015; Accepted December 5, 2015)

ABSTRACT — Drug-induced phospholipidosis (PLD) is one of the adverse reactions to treatment with cationic amphiphilic drugs. Recently, simple and reliable evaluation methods for PLD have been reported. However, the predictive power of these methods for in vivo PLD induction is insufficient in some cases. To accurately predict PLD, we focused on drug metabolism and used three-dimensional cultures of hepatocytes known as spheroids. Here we used the fluorescent phospholipid dye N-(7-nitrobenz-2-oxa-1,3- diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (NBD-PE) to detect PLD induction. After 48 hr exposure to 20 μM amiodarone and amitriptyline, PLD inducers, NBD-PE fluorescence in the spheroids was significantly higher than that in the control. In contrast, 1 mM acetaminophen, as a negative control, did not increase fluorescence. Furthermore, the combination of NBD-PE fluorescence and LysoTracker Red fluorescence and the accumulation of intrinsic phospholipids reflected PLD induction in spheroids. To evaluate metabolic activation, we assessed PLD induction by loratadine. NBD-PE fluorescence intensity was significantly increased by 50 μM loratadine treatment. However, the fluorescence was markedly decreased by co-treatment with 500 μM 1-aminobenzotriazole, a broad cytochrome P450 inhibitor. The formation of desloratadine, a metabolite of loratadine, was observed in spheroids after treatment with loratadine alone. These results showed that metabolic activation is the key factor in PLD induction by treatment with loratadine. We demonstrated that rat primary hepatocyte spheroid culture is a useful model for evaluating drug-induced PLD induction mediated by metabolic activation of the drug using the fluorescence probe technique.

Key words: Phospholipidosis, Hepatocyte, Spheroid, Loratadine, Metabolite, Cytochrome P450

INTRODUCTION lipophilicity with basicity conferred by amine groups. When such drugs diffuse into lysosomes, the basic groups Drug-induced phospholipidosis (PLD) is an adverse are protonated because of acidic environments, leading drug reaction that induces the accumulation of phosphol- to their accumulation in lysosomes (Kazmi et al., 2013). ipids in several organs. The presence of concentric mye- This event is considered as a critical step in phospholip- lin-like structures, called lamellar bodies, is the morpho- id accumulation. Furthermore, the inhibition of lysosomal logical hallmark of PLD (Anderson and Borlak, 2006). phospholipases by CADs, have been reported (Hostetler Cationic amphiphilic drugs (CADs), such as amiodar- and Matsuzawa, 1981; Shayman and Abe, 2013). In addi- one, which are clinically used as an antiarrhythmic, can tion, Sawada et al. (2005) reported that CADs change induce PLD (Sirajudeen et al., 2002). Amitriptyline, one gene expression by the inhibition of lysosomal phosphol- of the tricyclic antidepressants classiﬁed as CADs, is also ipase activity, lysosomal enzyme transport, and induc- a potential PLD inducer (Lüllman-Rauch, 1974). There tion of phospholipid and cholesterol biosynthesis. Song et are several reports describing the pathogenic mecha- al. (2011) reported that CADs increase phospholipid bio- nism of PLD. The chemical structures of CADs have high synthesis via peroxisome proliferator-activated receptor γ Correspondence: Seigo Sanoh (E-mail: [email protected])

Vol. 41 No. 1 156

M. Takagi et al. signaling. Recently, Ferslew and Brouwer (2014) reported that PLD markedly reduced the activity of both the organic anion transport involved in polypeptide-mediated uptake and bile salt export pump. Therefore, it is important to consider the relationship between the accumulation of phospholipids and toxicity in humans. It is necessary to predict PLD in pharmaceutical drugs to avoid associated toxicity during drug discovery and development. PLD can be directly evaluated by observ- Fig. 1. Metabolic pathway of loratadine to desloratadine. ing the appearance of lamellar bodies in histopathologi- cal analysis using electron microscopy after the admin- istration of drugs in cultured cells (Drenkhahn et al., bolic activities of drug-metabolizing enzymes are report- 1976). However, because this method takes time for eval- ed to decrease during conventional monolayer cultivation uation, development of alternative approaches via high- (Gómez-Lechón et al., 2010; Meng, 2010). Therefore, throughput screening is required. Recently, in vitro meth- three-dimensional (3-D) culture models have been devel- ods using the fluorescent imaging of phospholipids have oped to predict toxicity resulting from drug metabolism been reported (Nioi et al., 2007; Park et al., 2011; van de (Meng, 2010; Nakamura et al., 2011). 3-D culture meth- Water et al., 2011). NBD-phospholipid, which contains a ods form cell aggregations called spheroids, which are fluorescent moiety in its phospholipid structure, is often expected to maintain in vivo liver functions and cellular used for the detection of PLD. The fluorescence indicates environments. Previously, we reported the utility of sphe- phospholipid accumulation. Detection by PLD biomark- roid culture for detecting CYP-dependent metabolic acti- ers is also available (Hamaguchi et al., 2014a; Liu et al., vation of acetaminophen using a fluorescent probe (Sanoh 2014). These authors demonstrated that bis (monoacylg- et al., 2014). Hepatocyte spheroids were formed and cul- lycerol) phosphate and phosphatidylinositol composition tured using a micro-space cell culture plate (ElplasiaTM are reliable indicators of PLD induction. Furthermore, Kuraray Co., Ltd., Tokyo, Japan), and mRNA expres- in silico approaches are also useful for predicting PLD sion levels of some drug-metabolizing enzymes, such as potential. Ploemen et al. (2004) suggested the predictive CYPs, were found to persist during culture (Sanoh et al., power of various drugs for PLD using formulations and 2014). criteria such as “(ClogP)2 + (pKa)2 > 90” and “ClogP > 1, In the present study, our aim was to determine wheth- pKa > 8” that refer to lipophilicity parameters and acid er a drug metabolite is responsible for PLD and to assess dissociation constants. In silico approaches are sensitive, the utility of PLD evaluation based on drug metabolism convenient, and reliable. However, they have some limi- in hepatocyte spheroids. We report the development of a tations in prediction. simple PLD detection method using a fluorescence probe Loratadine, a histamine antagonist, is also an in vivo in rat hepatocyte spheroid culture and the contribution of PLD inducer (Hanumegowada et al., 2010). When used the metabolic activation of loratadine to PLD. in an in silico method of Ploemen (2004), loratadine is classified as a non-inducer of PLD (Hanumegowada et MATERIALS AND METHODS al., 2010). Additionally, an in vitro method using a cell line also indicated that loratadine is not a PLD inducer Materials (Kasahara et al., 2006). In this assay, the uptake of flu- Acetaminophen and 1-aminobenzotriazole (ABT) were orescent phospholipids into CHO-K1 cells was assessed purchased from Sigma-Aldrich (St. Louis, MO, USA). and loratadine did not induce fluorescence accumulation. Amitriptyline hydrochloride, loratadine, and deslorata- Loratadine is reported to be metabolized to deslorat- dine were purchased from Tokyo Chemical Industry Co., adine by cytochrome P450 (CYP) (Yumibe et al., 1996) Ltd. (Tokyo, Japan). Amiodarone was purchased from (Fig. 1). To accurately predict PLD, it is necessary to con- LKT Laboratories, Inc. (St. Paul, MN, USA). NBD- struct a cellular assay to evaluate not only parent drugs PE was purchased from Setareh Biotech, LLC (Eugene, but also their metabolites formed by drug-metabolizing OR, USA). LysoTracker Red® DND-99 was purchased enzymes. There are possibilities that metabolic activation from Life Technologies (Eugene, OR, USA). by drug-metabolizing enzymes leads to PLD induction. Hepatocytes are often used to assess drug-induced tox- Hepatocyte isolation and cell culture icity mediated by drug metabolism. However, the meta- Seven-week-old Crl:CD (SD) rats were obtained from

Vol. 41 No. 1 157

Metabolic activation in drug-induced phospholipidosis

Charles River Laboratories, Kanagawa, Japan. A two-step spheroids were washed with phenol red-free medium collagenase perfusion technique was used to isolate hepa- and observed with a FV 1000-D confocal laser scanning tocytes (Berry and Friend, 1969). The viability of suspend- biological microscope (Olympus, Tokyo, Japan). NBD-PE ed hepatocytes was determined by trypan blue assay, and fluorescence intensity values in the middle of spheroids ≥ 90% of cell viability of rat hepatocytes for spheroid for- were acquired from nine representative compartments per mation was confirmed. The study was approved by the well and normalized to the spheroid area. NBD-PE fluo- animal ethics committee of Hiroshima University. rescence was detected using a 488-nm excitation filter and Obtained primary hepatocytes were seeded at 2.5 × 105 a 520-nm emission filter. To assess metabolic activation cells/well in 24-well micro-space cell culture plates by loratadine and desloratadine, spheroids were exposed (ElplasiaTM, wide; 200 μm, depth; 50 μm) coated with to these compounds with or without 500 μM 1-aminoben- 0.01% poly-l-lysine. They were cultured in Dulbecco’s zotriazole (ABT), a broad inhibitor of CYPs. modified Eagle’s Nutrient Mixture F-12 Ham medium containing 10% fetal bovine serum, 2 mM L-glutamine, LysoTracker Red staining 100 units/mL penicillin, 100 μg/mL streptomycin, 10 mM After treatment with compounds and NBD-PE, nicotinamide, 50 μM β-mercaptoethanol, 100 nM dexam- LysoTrackerRed was also applied to spheroids. The incu- ethasone, 520 μM L-ascorbic acid, 1 μg/mL insulin, and bation concentration of LysoTracker Red was 1 μM. After 5 mM HEPES. The medium was replaced once a day. 20 min incubation with LysoTracker Red, spheroids were washed with phenol red-free medium and images were Treatment with test compounds and NBD-PE obtained with the FV 1000-D. Red fluorescence was The test compound solutions were applied to spheroid detected by a 559-nm excitation filter and a 572-nm emis- culture on day 6 after seeding of rat hepatocytes. Dime- sion filter. thyl sulfoxide was used as a vehicle with 10 and 20 μM of amiodarone and amitriptyline and 1 mM of acetami- Measurement of intrinsic phospholipid nophen. The final concentration of dimethyl sulfoxide accumulation was 0.1%. Loratadine and desloratadine were dissolved After 48 hr treatment with test compounds (20 μM in ethanol to a final ethanol concentration of 0.2%. The amiodarone, 20 μM amitriptyline, and 1 mM acetami- chemical structures of compounds used in this study are nophen), spheroids were washed with PBS and then described in Fig. 2. After 24 hr of incubation, test com- trypsinized. Eight wells per sample were collected and pounds were applied once again in 0.1% of NBD-PE replaced with 1 mL of PBS. The cell suspension was solution. NBD-PE stock solution was prepared by dis- extracted by addition of 3.75 mL of chloroform/meth- solving 2.5 mg NBD-PE in 1 mL of 50% acetonitrile anol (1/2, v/v) and vortexed for 2 min. Then, 1.25 mL and then sonicating. After an additional 24 hr incubation, chloroform was added, followed by vortexing for 30 sec.

Fig. 2. Chemical structures of (A) amiodarone, (B) amitriptyline, (C) acetaminophen, (D) loratadine, and (E) desloratadine.

Vol. 41 No. 1 158

M. Takagi et al.

After that, 1.25 mL of 1.5 M NaCl was added to wash compounds were applied for 48 hr from day 6 to day 8 the mixture, followed by vortexing for 30 sec. The chlo- in spheroids, and fluorescence intensity inside spheroids roform layer was extracted for analysis. The chloroform was measured at day 8 after the addition of NBD-PE for extracts were evaporated and residues were redissolved in 24 hr to evaluate phospholipid accumulation. To veri- 1 mL chloroform. This solution was mixed with 1 mL of fy our method, we used 10 and 20 μM amiodarone and a solution of 0.1 M FeCl36H2O and 0.4 M NH4SCN in amitriptyline as positive controls for PLD induction. A a 1:1 (v/v) ratio. The mixture was vortexed for 15 sec. higher concentration of acetaminophen (1 mM) was used The chloroform layer was collected and the absorbance as a negative control. Green emission fluorescence was at 485 nm was measured as phospholipid concentra- observed inside each spheroid by confocal laser scanning tion with a Beckman DU-640 spectrophotometer (GMI, biological microscopy (Fig. 3A). As expected, the fluores- Ramsey, MN, USA). The phospholipid was measured by cence intensity in the spheroids after exposure to 20 μM colorimetric assay following Stewart (1980). amiodarone and amitriptyline was significantly higher than that in the control. In contrast, 1 mM acetaminophen Quantification of the formation of desloratadine did not increase fluorescence (Fig. 3B). in spheroids Spheroids on day 6 were treated with 50 μM lorata- Co-localization of lysosome and NBD-PE probe dine. After 24 hr, the medium was replaced with fresh in spheroids medium containing the test compound and incubated for We observed the localization of lysosome and phos- an additional 24 hr for a total treatment period of 48 hr. pholipid accumulation using LysoTracker Red and NBD- The culture medium and hepatocytes after 24 and 48 hr PE probe, respectively. LysoTracker Red accumulates in exposure were collected by trypsinization. To 50 μL of the lysosome and displays red emission fluorescence. The cell suspension, 100 μL of acetonitrile containing the fluorescence intensities of both the LysoTracker Red and internal standard carbamazepine was added with mix- NBD-PE probe in the spheroids after exposure to 20 μM ing followed by centrifugation at 14,000 × g for 5 min. amiodarone and amitriptyline were higher than those in Supernatants were diluted with 0.1% formic acid and ana- the control and 1 mM acetaminophen. Acetaminophen did lyzed by liquid chromatography-tandem mass spectrom- not increase either fluorescence intensity. Co-localization etry (LC/MS/MS). Of the solution, 10 μL was injected of LysoTracker Red and NBD-PE was partly observed in onto an Inertsil ODS-3 column (5 μm, 50 × 2.1 mm; GL spheroids treated with amiodarone and amitriptyline. The Sciences Inc., Tokyo, Japan) connected to an HP-1100 LysoTracker Red fluorescence was increased by treatment series HPLC system. The mobile phases were mixtures of with amiodarone and amitriptyline (Fig. 4). 0.1% formic acid (solvent A) and acetonitrile (solvent B) flowing at 200 μL/min. The composition (A:B) was 90:10 Evaluating intrinsic phospholipid accumulation (initial), 40:60 (5-7 min), and return to 90:10 (7.1-11 min) by colorimetric assay for re-equilibration. The NBD-PE assay shows the cell response to external An API 2000 system (Applied Biosystems, Foster, CA, phospholipids. In addition to using the NBD-PE assay, USA) was used to perform MS/MS measurements. The we measured absorbance at 485 nm, which indicates the mass number of the molecular ion was m/z = 311.0 and total amount of intrinsic phospholipid contents in sphe- that of the product ion was m/z = 258.9. roid lysates after 48 hr exposure to 20 μM amiodarone or amitriptyline and 1 mM acetaminophen, using a color- Statistical analysis imetric assay. Absorbances in spheroids after exposure to All experiments were performed with three independ- 20 μM amiodarone or amitriptyline tended to be higher ent replications and representative experiments are shown. than those in the control although these were not signif- Data are expressed as mean + S.D. Statistical analysis was icant. In contrast, 1 mM acetaminophen did not increase performed using Dunnett’s t test. the absorbance (Fig. 5).

RESULTS Metabolic activation of loratadine for PLD In the intracellular ATP assay indicating cell viability, PLD assessment using the NBD-PE probe in 5, 10, and 50 μM loratadine showed no effects on ATP spheroids contents in spheroids, although 100 μM loratadine sig- The formation of hepatocyte spheroids was conﬁrmed nificantly reduced ATP contents in spheroids (data not at day 5 after seeding of isolated rat hepatocytes. Test shown). We accordingly evaluated phospholipid accu-

Vol. 41 No. 1 159

Metabolic activation in drug-induced phospholipidosis

(A)

(B)

Fig. 3. Phospholipidosis evaluation by the fluorescence phospholipid dye, NBD-PE. (A) Confocal laser scanning micrographs of spheroids in nine compartments treated with test compounds for 48 hr from day 6 to day 8 and dyed with NBD-PE. (B) Quantification of NBD-PE fluorescence. Results are expressed as mean ± S.D. (n = 3 independent experiments, *p < 0.05 vs. control, **p < 0.01 vs. control). The fluorescence is the average emission of nine compartments.

mulation in spheroids using three concentrations of desloratadine alone and co-treatment with ABT were loratadine (5, 10, and 50 μM) using the NBD-PE assay. observed (Fig. 6B). NBD-PE fluorescence intensity was significantly (approx- imately 2-fold) increased by 50 μM loratadine treatment, Formation of desloratadine in spheroids the highest concentration of loratadine that did not influ- We measured the formation of desloratadine in the ence cell viability. However, the green emission of NBD- culture medium and cell lysate after 24 and an addition- PE was clearly reduced by co-treatment with 500 μM al 24 hr for a total period of 48 hr treatment with 50 μM ABT (Fig. 6A). loratadine by LC/MS/MS analysis. Desloratadine forma- We also evaluated desloratadine, a metabolite of lorat- tion was spontaneously observed in spheroid culture fol- adine, alone. We confirmed the accumulation of NBD- lowing 24 and 48 hr treatments, and the concentration of PE fluorescence on treatment with 10 μM desloratadine. desloratadine was 7 μM after 48 hr exposure, which is the The intensity of fluorescence accumulation after exposure evaluation time for the NBD-PE assay. The concentration to 50 μM desloratadine increased to 4-fold of that of the of desloratadine in spheroids was suppressed to 1 μM by control. In contrast, no differences in spheroids between co-treatment with 500 μM ABT (Fig. 7).

Vol. 41 No. 1 160

M. Takagi et al.

Fig. 4. Fluorescence imaging of NBD-PE and LysoTracker Red-treated test compounds for 48 hr from day 6 to day 8 in spheroids.

Park et al., 2012; van de Water et al., 2011). NBD-PE, which is phosphatidylethanolamine with a fluorescent label, is available as a cell imaging probe to evaluate phospholipid accumulation by measurement of intracellular fluorescence. In the present study, we used micro- space cell culture plates to produce hepatocyte spheroids. Regularly sized spheroids can be formed because the bottom surface of each well has regularly spaced square compartments (200 × 200 × 50-μm depth). Furthermore, the plate contents can be easily viewed by confocal laser scanning microscopy because the bottom surface is thin. Sanoh et al. (2014) demonstrated that the plate is useful for live-cell imaging techniques using calcein-AM and Fig. 5. Intrinsic phospholipid accumulation in spheroids after monochlorobimane, which are imaging probes for cell treatment with test compounds for 48 hr from day 6 viability and intracellular glutathione, respectively. In the to day 8. Results are expressed as mean ± S.D. (n = 3 independent experiments). present study, we observed NBD-PE fluorescence for PLD assessment in spheroids. First, amiodarone and amitriptyline as positive test compounds and acetaminophen as a negative control were used to validate our method. Ami- DISCUSSION odarone and amitriptyline induced PLD in rats in vivo (Hirode et al., 2008). In contrast, acetaminophen did not It is necessary to identify the mechanism of drug-in- induce PLD (Hanumegowada et al., 2010; van de Water duced PLD and develop screening methods to avert this et al., 2011). In Fig. 3, NBD-PE fluorescence accumula- condition in drug discovery and development. Recent tions in spheroids after exposure to 20 μM amiodarone reports have described the development of screening and amitriptyline were observed, whereas a high concen- methods using cell imaging techniques (Nioi et al., 2007; tration of acetaminophen did not induce NBD-PE fluo-

Vol. 41 No. 1 161

Metabolic activation in drug-induced phospholipidosis (A) (B)

Fig. 6. Phospholipidosis evaluation by NBD-PE for (A) loratadine treatment and (B) desloratadine treatment. The test compounds were incubated for 48 hr. Results are expressed as mean ± S.D. (n = 3 independent experiments, **p < 0.01 vs. control). The ﬂuorescence is the average emission in nine compartments.

after exposure to amiodarone and amitriptyline in spheroids (Fig. 4). Funk and Krise (2012) showed lysosome expansion by induction of lysosomal cholesterol contents after exposure to CADs in human fibroblasts. The high accumulation of LysoTracker Red fluorescence alone in spheroids following treatment with amiodarone and amitriptyline may indicate lysosome expansion, given that this fluorescence was not observed in spheroids of the control or after exposure to acetaminophen. Ferslew and Brouwes (2014) suggested that the accumulation of LysoTracker Red fluorescence is an indicator predicting PLD risk in the sandwich culture of rat hepatocytes. From our findings, intracellular imaging in spheroids is also useful for predicting PLD, given that co-localization of the NBD-PE probe in lysosomes and accumulation of Fig. 7. Comparison of desloratadine formation after 24 and an LysoTracker Red probe could be observed. additional 24 hr for a total period of 48 hr treatment NBD-PE fluorescence accumulation indicates cell with 50 μM loratadine with or without ABT. Deslorat- response to external phospholipids. To confirm accumula- adine was analyzed in a mixture of spheroids and me- tion of intrinsic phospholipids induced by CADs, we per- dium. Results are expressed as mean ± S.D. (n = 3 independent experiments). formed a colorimetric assay (Stewart, 1980) (Fig. 5). The result showed that the intrinsic phospholipids had ten- dency to accumulate following the treatment with 20 μM rescence. These results qualitatively agree with those of amiodarone or amitriptyline. Several phospholipids, such other in vitro cellular imaging assays using NBD-PE flu- as phosphatidylcholine and phosphatidylethanolamine, orescence (Miyamoto et al., 2009; van de Water et al., accumulate in RAW264 cells (Hamaguchi et al., 2014a). 2011), which suggested that NBD-PE is also useful for Measurement of phospholipids by colorimetric assay sup- intracellular imaging in spheroids. ported the NBD-PE fluorescence accumulation in sphe- PLD induces the accumulation of phospholipids in lys- roids. osomes (Anderson and Borlak, 2006; Hamaguchi et al., It is also important to consider drug metabolism, which 2014b). We confirmed that fluorescence of NBD-PE was affects PLD. The observation that expression levels of partly co-localized in lysosomes using LysoTracker Red some drug-metabolizing enzymes persisted during sphe-

Vol. 41 No. 1 162

M. Takagi et al.

Table 1. Physicochemical properties and model prediction of loratadine, desloratadine. In silico prediction ClogP pKa PLD in vivo (Ploemen et al., 2004) Loratadine 5.05a 4.81a 49a Negative Positivea Desloratadine 3.45b 10.27b 117b Positive - a: Hanumegowda et al., 2010 b: Calculated using Advanced Chemistry Development (ACD/Labs) Software V11.02 (© 1994-2015 ACD/Labs)

roid culture on the micro-space cell culture plate (Sanoh based assay. Our findings support the in silico results by et al., 2014) suggests that spheroids are useful for assay- Goracci et al. (2013), which suggested that desloratadine ing drug metabolism and associated toxicity. is involved in inducing PLD. In the present study, we used loratadine as a test com- Yumibe et al. (1996) suggested that desloratadine is pound to evaluate the contribution of metabolic activation formed by O-dealkylation and followed by decarbox- to PLD. Loratadine induced PLD in rats (Hanumegowda ylation. The formation of the metabolite with amine et al., 2010). However, loratadine has been shown to be group though dealkylation such as the reaction metab- negative for PLD induction by an in silico approach using olizing loratadine to desloratadine, is observed in oth- lipophilicity parameters and acid dissociation constants as er drug metabolisms. These metabolisms including reported by Ploemen (Hanumegowda et al., 2010). Lorat- direct N-dealkylation by drug-metabolizing enzymes adine is metabolized to desloratadine by CYP3A4 and can increase the basicity of compounds, and clogP may by CYP2D6 in humans (Yumibe et al., 1996). Because decrease. Lipophilicity and basicity of metabolites may the pKa of desloratadine is higher than that of lorata- also be key factors in PLD induction. For example, dine, desloratadine has a potential to induce PLD. When Goracci et al. (2013) suggested the possibility of meta- the parameters of desloratadine are also applied in an in bolic activation by the conversion from ketoconazole to silico approach, desloratadine is determined to positive N-deacetyl ketoconazole in addition to that by lorata- (Table 1). dine. This report concluded that ketoconazole metabolism Loratadine induced NBD-PE fluorescence accumu- was responsible for PLD induced in vivo by administra- lation in spheroids, indicating PLD induction. However, tion of ketoconazole, given that ketoconazole is deter- this fluorescence accumulation was not observed in sphe- mined to negative when applied in silico methods of Plo- roids after exposure to loratadine by co-treatment with emen (Hanumegowda et al., 2010). Zhou et al. (2011) ABT, a broad CYP inhibitor. An assay of NBD-PE flu- also reported the PLD potential of parent compounds and orescence accumulation by treatment with desloratadine their metabolites. They indicated that there are some cas- alone showed a significant increase in fluorescence inten- es in which not only parent drugs but also their metab- sity induced in a dose-dependent manner by lower con- olites have PLD potential, such as in the metabolism of centrations of desloratadine than of loratadine (Fig. 6). amiodarone to desethylamiodarone and that of amitriptyl- These results indicated metabolic activation by deslorat- ine to nortriptyline, in contrast to the case of loratadine. adine of PLD through CYPs in spheroids. Furthermore, In particular, we propose that the inference of metabolic we measured the formation of desloratadine after treat- activation of an inactive parent drug to an active metab- ment with loratadine using LC/MS/MS. Approximate- olite should be made with caution, as with loratadine, ly 7 μM of desloratadine was formed in spheroid cul- because PLD risk may be falsely estimated as negative ture after 48 hr treatment with 50 μM loratadine (Fig. 7). unless metabolism is considered. Given that the fluorescence intensity in the spheroids of Sanoh et al. (2014) demonstrated that the metabolic desloratadine alone increased at 10 μM desloratadine, we activation of acetaminophen can be detected using fluores- showed that the fluorescence intensity induced by lorat- cence imaging in rat hepatocyte spheroids. In the present adine resulted in the formation of desloratadine. There study, we also found that metabolic activation associated was not a contribution of the subsequent metabolic acti- with PLD is detectable by fluorescence imaging in sphe- vation of desloratadine, given that no changes in fluores- roids. The usage of an inhibitor of drug metabolism to cence intensity were observed in spheroids after exposure predict metabolic activation may be convenient because it to desloratadine with or without ABT (Fig. 6B). We could eliminates the need to synthesize the metabolite standard. assess the metabolic activation of loratadine using a cell- Gum et al. (2001) reported that PLD induced by ABT-

Vol. 41 No. 1 163

Metabolic activation in drug-induced phospholipidosis

770, a matrix metalloproteinase inhibitor, was associated teral uptake and biliary excretion of anionic probe substrates. primarily with an amine metabolite of ABT-770 detected Toxicol. Sci., 139, 99-107. Funk, R.S. and Krise, J.P. (2012): Cationic amphiphilic drugs cause by use of a metabolite standard in rat hepatocytes. In this a marked expansion of apparent lysosomal volume: implications study (Gum et al., 2001), treatment with the parent com- for an intracellular distribution-based drug interaction. Mol. pound showed no evidence of inducing PLD in rat hepa- Pharm., 9, 1384-1395. tocytes, suggesting that sufficient metabolism did not Goracci, L., Ceccarelli, M., Bonelli, D. and Cruciani, G. (2013): occur in conventional rat primary culture. This metabol- Modeling phospholipidosis induction: reliability and warnings. J. Chem. Inf. Model., 53, 1436-1446. ic activation of ABT-770 may be detectable in spheroid Gómez-Lechón, M.J., Lahoz, A., Gombau, L., Castell, J.V. and culture. Donato, M.T. (2010): In vitro evaluation of potential hepatotox- We used hepatocytes in this assay to detect metabolic icity induced by drugs. Curr. Pharm. Des., 16, 1963-1977. activation. Because the severity and organ/tissue specifi- Gum, R.J., Hickman, D., Fagerland, J.A., Heindel, M.A., Gagne, city of PLD is thought to differ among drugs (Anderson G.D., Schmidt, J.M., Michaelides, M.R., Davidsen, S.K. and Ulrich, R.G. (2001): Analysis of two matrix metalloprotein- and Borlak, 2006), it is desirable to evaluate PLD using ase inhibitors and their metabolites for induction of phospholi- other tissues to perform accurate evaluation reflecting in pidosis in rat and human hepatocytes. Biochem. Pharmacol., 62, vivo conditions. Furthermore, it is important to consid- 1661-1673. er species differences in drug-induced PLD. Bhandari et Hamaguchi, R., Tanimoto, T. and Kuroda, Y. (2014a): Putative al. (2008) observed species differences in PLD potential biomarker for phospholipid accumulation in cultured cells treated with phospholipidosis-inducing drugs: alteration of the phos- between rat and monkey hepatocytes and suggested that phatidylinositol composition detected using high-perform- these were due to differences in drug metabolizing abili- ance liquid chromatography-tandem mass spectrometry. J. ty. Thus, it is desirable to establish methods using human Chromatogr. B Analyt. Technol. Biomed. Life Sci., 967, 110- hepatocytes. At least, spheroid culture is useful in drug 117. Hamaguchi, R., Haginaka, J., Tanimoto, T. and Kuroda, Y. (2014b): discovery and development as a cell-based assay for PLD Maintenance of luminal pH and protease activity in lysosomes/ associated with drug metabolism. late endosomes by vacuolar ATPase in chlorpromazine-treated RAW264 cells accumulating phospholipids. Cell Biol. Toxicol., ACKNOWLEDGMENTS 30, 67-77. Hanumegowda, U.M., Wenke, G., Regueiro-Ren, A., Yordanova, R., Corradi, J.P. and Adams, S.P. (2010): Phospholipidosis as a This work using a confocal laser scanning biological function of basicity, lipophilicity, and volume of distribution of microscope was performed at the Analysis Center of Life compounds. Chem. Res. Toxicol., 23, 749-755. Science, Natural Science Center for Basic Research and Hirode, M., Ono, A., Miyagishima, T., Nagao, T., Ohno, Y. and Development, Hiroshima University. This research was Urushidani, T. (2008): Gene expression profiling in rat liver treated with compounds inducing phospholipidosis. Toxicol. supported by JSPS KAKENHI Grant Number 26293029 Appl. Pharmacol., 229, 290-299. and by Kuraray Co., Ltd. Hostetler, K.Y. and Matsuzawa, Y. (1981): Studies on the mechanism of drug-induced lipidosis. Cationic amphiphilic drug Conflict of interest---- This study was supported by inhibition of lysosomal phospholipases A and C. Biochem. Kuraray Co., Ltd. as the collaborative research. Pharmacol., 30, 1121-1126. Kasahara, T., Tomita, K., Murano, H., Harada, T., Tsubakimoto, K., Ogihara, T., Ohnishi, S. and Kakinuma, C. (2006): Establish- REFERENCES ment of an in vitro high-throughput screening assay for detecting phospholipidosis-inducing potential. Toxicol. Sci., 90, 133-141. Anderson, N. and Borlak, J. (2006): Drug-induced phospholipido- Kazmi, F., Hensley, T., Pope, C., Funk, R.S., Loewen, G.J., sis. FEBS Lett., 580, 5533-5540. Buckley, D.B. and Parkinson, A. (2013): Lysosomal sequestra- Berry, M.N. and Friend, D.S. (1969): High-yield preparation of iso- tion (trapping) of lipophilic amine (cationic amphiphilic) drugs lated rat liver parenchymal cells: a biochemical and fine struc- in immortalized human hepatocytes (Fa2N-4 cells). Drug Metab. tural study. J. Cell. Biol., 43, 506-520. Dispos., 41, 897-905. Bhandari, N., Figueroa, D.J., Lawrence, J.W. and Gerhold, D.L. Liu, N., Tengstrand, E.A., Chourb, L. and Hsieh, F.Y. (2014): Di-22: (2008): Phospholipidosis assay in HepG2 cells and rat or rhe- 6-bis(monoacylglycerol)phosphate: A clinical biomarker of sus hepatocytes using phospholipid probe NBD-PE. Assay Drug drug-induced phospholipidosis for drug development and safety Dev. Technol., 6, 407-419. assessment. Toxicol. Appl. Pharmacol., 279, 467-476. Drenckhahn, D., Kleine, L. and Lüllmann-Rauch, R. (1976): Lyso- Lüllmann-Rauch, R. (1974): Lipidosis-like ultrastructural alter- somal alterations in cultured macrophages exposed to anorexi- ations in rat lymph nodes after treatment with tricyclic anti- genic and psychotropic drugs. Lab. Invest., 35, 116-123. depressants or neuroleptics. Naunyn. Schmiedebergs. Arch. Ferslew, B.C. and Brouwer, K.L. (2014): Identification of hepat- Pharmacol., 286, 165-179. ic phospholipidosis inducers in sandwich-cultured rat hepato- Meng, Q. (2010): Three-dimensional culture of hepatocytes for cytes, a physiologically relevant model, reveals altered basola- prediction of drug-induced hepatotoxicity. Expert Opin. Drug Metab. Toxicol., 6, 733-746.

Vol. 41 No. 1 164

M. Takagi et al.

Miyamoto, S., Matsumoto, A., Mori, I. and Horinouchi, A. (2009): tion mechanism and establishment of a novel in vitro screening Relationship between in vitro phospholipidosis assay using system. Toxicol. Sci., 83, 282-292. HepG2 cells and 2-week toxicity studies in rats. Toxicol. Mech. Shayman, J.A. and Abe, A. (2013): Drug induced phospholipido- Methods, 19, 477-485. sis: an acquired lysosomal storage disorder. Biochim. Biophys. Nakamura, K., Mizutani, R., Sanbe, A., Enosawa, S., Kasahara, M., Acta., 1831, 602-611. Nakagawa, A., Ejiri, Y., Murayama, N., Miyamoto, Y., Torii, Sirajudeen, K.N., Gurumoorthy, P., Devaraj, H. and Devaraj, S.N. T., Kusakawa, S., Yamauchi, J., Fukuda, M., Yamazaki, H. and (2002): Amiodarone-induced phospholipidosis: an in vivo [14C]- Tanoue, A. (2011): Evaluation of drug toxicity with hepatocytes acetate uptake study in rat. Drug Chem. Toxicol., 25, 247-254. cultured in a micro-space cell culture system. J. Biosci. Bioeng., Song, M., Kim, Y.J. and Ryu, J.C. (2011): Phospholipidosis induced 111, 78-84. by PPARγ signaling in human bronchial epithelial (BEAS-2B) Nioi, P., Perry, B.K., Wang, E.J., Gu, Y.Z. and Snyder, R.D. (2007): cells exposed to amiodarone. Toxicol. Sci., 120, 98-108. In vitro detection of drug-induced phospholipidosis using gene Stewart, J.C. (1980): Colorimetric determination of phospholipids expression and ﬂuorescent phospholipid based methodologies. with ammonium ferrothiocyanate. Anal. Biochem., 104, 10-14. Toxicol. Sci., 99, 162-173. van de Water, F.M., Havinga, J., Ravesloot, W.T., Horbach, G.J. Park, S., Choi, Y.J. and Lee, B.H. (2012): In vitro validation of drug- and Schoonen, W.G. (2011): High content screening analysis of induced phospholipidosis. J. Toxicol. Sci., 37, 261-267. phospholipidosis: validation of a 96-well assay with CHO-K1 Ploemen, J.P., Kelder, J., Hafmans, T., van de Sandt, H., van and HepG2 cells for the prediction of in vivo based phospholipi- Burgsteden, J.A., Saleminki, P.J. and van Esch, E. (2004): Use dosis. Toxicol. In Vitro, 25, 1870-1882. of physicochemical calculation of pKa and CLogP to predict Yumibe, N., Huie, K., Chen, K.J., Snow, M., Clement, R.P. and phospholipidosis-inducing potential: a case study with structur- Cayen, M.N. (1996): Identiﬁcation of human liver cytochrome ally related piperazines. Exp. Toxicol. Pathol., 55, 347-355. P450 enzymes that metabolize the nonsedating antihistamine Sanoh, S., Santoh, M., Takagi, M., Kanayama, T., Sugihara, K., loratadine. Formation of descarboethoxyloratadine by CYP3A4 Kotake, Y., Ejiri, Y., Horie, T., Kitamura, S. and Ohta, S. (2014): and CYP2D6. Biochem. Pharmacol., 51, 165-172. Fluorometric assessment of acetaminophen-induced toxicity Zhou, L., Geraci, G., Hess, S., Yang, L., Wang, J. and Argikar, U. in rat hepatocyte spheroids seeded on micro-space cell culture (2011): Predicting phospholipidosis: a fluorescence noncell plates. Toxicol. In Vitro, 28, 1176-1182. based in vitro assay for the determination of drug-phospholip- Sawada, H., Takami, K. and Asahi, S. (2005): A toxicogenomic id complex formation in early drug discovery. Anal. Chem., 83, approach to drug-induced phospholipidosis: analysis of its induc- 6980-6987.

Vol. 41 No. 1