Establishment of an In Vitro High-Throughput Screening Assay for Detec
http://www.100md.com
《毒物学科学杂志》
Pharmaceutical Research Center, Research Planning and Management
Development Research, Mochida Pharmaceutical Company Limited, Shizuoka, Japan
ABSTRACT
Excessive accumulation of phospholipids results in phospholipidosis (PL), which may interfere with cellular functions, leading to acute or chronic disease or even death. Electron-microscopic detection of cytoplasmic lamellar bodies is often used as a diagnostic criterion of PL, but a faster, more convenient procedure is required for high-throughput assay of the PL-inducing potential of candidate drugs. We have developed a 96-well microplate cell-culture method for detecting PL, using a phosphatidylcholine-conjugated dye (NBD-PC) and a fluoro-microplate reader. The fluorescence intensity due to NBD-PC was normalized to that of Hoechst33342, used as an indicator of cell number, to obtain the amount of NBD-PC taken up per living cell. To select a suitable cell type, we examined the PL-detection sensitivity of five cell lines, as well as human and rat primary hepatocyte cultures, with five cationic amphiphilic drugs (CAD) as PL inducers and a negative control compound. The cell lines CHO-K1 and CHL/IU gave the best results. The NBD-PC uptake per CHO-K1 cell showed a high correlation with the pathological score of PL for 24 compounds, including PL-positive and negative compounds. This high-throughput screening assay for PL-inducing potential (HTS-PL assay) offers high sensitivity and accuracy, and it allows simultaneous determination of cytotoxicity.
Key Words: phospholipidosis; cationic amphiphilic drug; NBD-PC; Hoechst33342; HTS-PL assay.
INTRODUCTION
Phospholipidosis (PL) is a pathological accumulation of phospholipids in tissue, characterized by the formation of lysosomal lamellar bodies, and may be induced by certain drugs with a cationic lipophilic structure (cationic amphiphilic drugs; CAD). The accumulated phospholipids interfere with cellular functions, sometimes with fatal results, as in the case of PL induced by diethylaminoethoxyhexestrol (Yamamoto et al., 1971a). Therefore, it is very important to investigate whether drugs under development have the potential to induce PL. Electron microscopic detection of cytoplasmic lamellar bodies is thought to be the most reliable method for identifying PL, but this is unsuitable for high-throughput systems, which are required for assay of large numbers of compounds generated by combinatorial chemistry. Roger et al. (1991) reported that phospholipid accumulation could be detected by means of fluorescence microscopy, following the incorporation of a fluorescent phospholipid analog, 1-acyl-2-[12-(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]dodecanoyl]-sn-glycero-3-phosphocholine (NBD-PC), into cytoplasm. Therefore, we sought to establish a high-throughput screening assay (HTS-PL assay) to assay PL-inducing potential quantitatively and rapidly in a 96-well microplate format, using NBD-PC and cultured cells. Phospholipidosis is usually induced after chronic intake of CADs, with a latency period ranging from a few days to several months, depending on species (Lullmann-Rauch et al., 1974; Seiler and Wassermann, 1975; Yamamoto et al., 1971b), strain within species (Reasor et al., 1996) and age (Kacew et al., 1982). Therefore, we examined a range of cell types (human or rat, cultured or primary-cultured) using five PL-inducing CADs and a negative control in order to find suitable cells for the assay.
MATERIALS AND METHODS
Chemicals.
Test compounds known to cause or not to cause phospholipidosis or not in animals and/or humans in vivo were purchased from the following sources: amiodarone, imipramine, promazine, chloroquine, chloramphenicol, propranolol, acetaminophen, amitriptyline hydrochloride, AY9944, chlorpromazine, clomipramine hydrochloride, clozapine, disopyramide phosphate salt, flecainide, fluoxetine hydrochloride, haloperidol, ketoconazole, loratadine, quinidine, sotalol, tamoxifen citrate, thioridazine hydrochloride, and zimelidine from Sigma Chemical (St. Louis, MO, USA); clarithromycin and erythromycin from Wako Pure Chemicals (Osaka, Japan); levofloxacin from LKT Laboratories (St. Paul, MN, USA). Chloroquine, disopyramide phosphate salt, imipramine, promazine, and propranolol were dissolved in water and other compounds were dissolved in dimethylsulfoxide (DMSO) to afford 100 mM solutions.
Cell culture.
The human hepatocellular carcinoma cell line (HepG2), diploid rat liver epithelial cell line (ARLJ301–3), Chinese hamster ovary cell line (CHO-K1), Chinese hamster lung cell line (CHL/IU), and mouse macrophage-like cell line (J744A) were purchased from Dainippon Pharmaceutical Co. Ltd. (Osaka, Japan), and all culture reagents were purchased from GIBCO BRL (Grand Island, NY, USA), unless otherwise noted. Cell lines were cultured in the following media: HepG2 and ARLJ301–3 in Williams' medium (WE) containing 10% fetal bovine serum (FBS), CHO-K1 in D-MEM/F-12 containing 10% FBS, CHL/IU in modified Eagle's medium (MEM) containing 10% calf serum (CS), J744A in Dulbecco's-MEM (D-MEM) containing 10% FBS. All media were supplemented with 100 μg/ml streptomycin and 100 U/ml penicillin. These cells were used in the logarithmic phase of growth and were seeded in wells of 96-well microplates for HTS-PL assay using NBD-PC. For the confocal laser scanning study, cell density per well was as follows: HepG2, 3 x 104 cells; ARLJ301, 2 x 104 cells; CHO-K1, 1 x 104 cells; CHL/IU, 1 x 104 cells; J744A, 4 x 104 cells. Primary hepatocyte cultures were prepared from a male Sprague-Dawley rat at 10 weeks of age by the method of Uno et al. (1992). The rat was anaesthetized with sodium pentobarbital and the liver was perfused in situ for 5 min at flow rate of 10 ml/min with Hank's balanced salt solution (Ca2+/Mg2+-free) containing 0.5 mM EGTA (Sigma) and 10 mM HEPES (pH 7.2–7.3), followed by 7 min in Hanks' balanced salt solution containing 0.05% collagenase S-1 (Nitta Gelatin Inc., Osaka, Japan), 50 μg/ml trypsin inhibitor (Sigma), 10 mg/ml bovine serum albumin (Sigma), 10 mM HEPES, and 560 μg/ml CaCl2 (pH 7.5). All solutions were kept at 40°C for use. The perfused liver was minced with surgical scissors in a Petri dish containing Hank's balanced salt solution, and filtered through a cell strainer with 100-μm nylon mesh (Becton Dickinson, San Jose, CA, USA). Rat hepatocytes were washed twice by centrifugation at 50 x g for 1 min and resuspended in culture medium for primary hepatocytes, WE with additives (insulin, hydrocortisone, transferrin, hEGF, L-ascorbic acid), selected from the HCM SingleQuots Kits (Cambrex Co., North Brunswick, NJ, USA), 100 μg/ml streptomycin, 100 U/ml penicillin, and 10% FBS. The number and viability of cells were determined using a conventional trypan blue exclusion test. Human primary plateable cryopreserved hepatocytes were purchased from In Vitro Technologies, Inc. (Baltimore, MD, USA), and were warmed quickly to 37°C in a water bath. Percoll was purchased from Sigma, and isotonic Percoll solution was prepared by mixing Percoll (90 ml) with 9% NaCl (10 ml). Dead cells were removed by centrifugation at 50 x g for 10 min in an isolation solution prepared by mixing isotonic Percoll solution (15 ml) with WE medium (35 ml). The pelleted human hepatocytes were resuspended in culture medium for primary hepatocytes, washed twice at 50 x g for 5 min, and examined with a conventional trypan blue exclusion test. Rat and human hepatocytes were plated at density of 1 x 104 rat cells or 2 x 104 human cells per well on a collagen-coated 96-well microplate (Becton, Dickinson). After an initial attachment period (about 4 h), the hepatocytes were washed once with the medium. For transmission electron microscopic examination, 10 times higher cell numbers were seeded into 6-well microplates. All cell cultures were conducted at 37°C under 95% air and 5% CO2.
Drug treatment.
Serial dilutions of compounds in the medium were prepared by using deep-well microplates (chemical preparation plate). Medium was added to columns of wells on the chemical preparation plate as follows; column A, 998 μl; columns B and D, 700 μl; columns C and E, 600 μl. A 2-μl aliquot of test compound (100 mM) dissolved in DMSO or water was added to the wells in column A, then 300 μl of medium was transferred from these wells to the wells in column B. Similarly, successive transfers resulted in final concentrations of 200, 60, 20, 6, and 2 μM in duplicate (six compounds/plate). For experiments performed to investigate the correlation between assay results and pathological changes in CHO-K1 cells, plates were similarly prepared, but with five compounds/plate in replicates of four. In chemical preparation plates, the concentrations of compounds dissolved in the medium were twice the desired final concentrations.
NBD-PC preparation.
The fluorescent phospholipid analog, 1-acyl-2-[12-(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]dodecanoyl]-sn-glycero-3- phosphocholine (NBD-PC; Avanti Polar Lipids, Inc., AL, USA) was prepared according to the method of Kremer et al. (1977). A stock solution of NBD-PC was prepared at 10 mM in chloroform, divided into 100 μl aliquots in vials, and dried. Vials were stored at –20°C and reconstituted in 100 μl of 100% ethanol for use. This solution was added to the medium to make a concentration of 80 μM; the mixture was then sonicated for 30 min and passed through a 0.2-μm disposable filter. Medium containing the fluorescent phospholipid analog is referred to as "NBD-PC medium."
HTS-PL assay using NBD-PC.
The HTS-PL assay was established by modifying Ulrich's method (1991). At 24 h after plating the cells, medium in each well of the 96-microplate was removed, 50 μl of medium containing 80 μM NBD-PC was added quickly, and then 50 μl of medium containing various test compounds at serial twofold concentration was transferred to the cell culture plate from the chemical preparation plate. At 24 h after addition of NBD-PC and test compounds, the plated cells were washed twice with phosphate-buffered saline (PBS), and 50 μl of PBS was added to each well. Fluorescence of NBD-PC taken up by cells was measured with a fluoro-microplate reader (Thermo Electron Oy, Vantaa, Finland) with an excitation wavelength of 485 nm and emission wavelength of 538 nm. After measuring the fluorescence of NBD-PC, and 50 μl of PBS containing 20 μg/ml Hoechst33342 (ICN Pharmaceutical, Inc., Costa Mesa, CA, USA) was added. Incubation was continued for 20 min at 37°C, after which the cellular fluorescence of Hoechst33342 was measured using a fluoro-microplate reader with an excitation wavelength of 355 nm and emission wavelength of 460 nm. To remove the influence of cytotoxicity, the uptake of NBD-PC per cell was calculated using the following formula.
Confocal laser scanning microscopy.
Cellular fluorescent lipid was observed using an Olympus confocal laser microscope FV1000 (Olympus, Tokyo, Japan) equipped 20x and 40x fluoro lenses. Microplates read with the fluoro-microplate reader for HTS-PL assay were also used for confocal laser scanning microscopy. Fluorescence of NBD-PC and that of Hoechst33342 were visualized as green and blue, respectively, by dual-wavelength excitation (NBD-PC, 488 nm; Hoechst33342, 405 nm) and indicated accumulation of phospholipids in cytoplasm and the location of the nucleus, respectively. In the case of rat primary hepatocytes, transmission laser micrographs were taken of the same microscopic field observed for the fluorescence of NBD-PC and Hoechst33342 to investigate cell morphology.
Transmission electron microscopy.
Cells cultured in 6-well microplates under the conditions described above were fixed with cold 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2, PB) for 2 h, then washed with PB, and postfixed with 1.0% osmium tetroxide in PB for 60 min. They were again rinsed with the same PB, dehydrated with graded ethanol through to 100%, infiltrated with Epok 812 (Oken, Tokyo, Japan), and polymerized for 3 days. Thin sections were cut with an ultramicrotome (Reichent, Vienna, Austria), and stained with saturated uranyl acetate and lead acetate. The degree of formation of lamellar bodies in lysosomes was evaluated with a JEM-100CX2 electron microscope (JEOL, Tokyo, Japan) and scored on a scale of – to ++++ (–, within normal limits; ±, very slight; +, slight; ++, moderate; +++, severe; ++++, very severe) in a blind fashion.
RESULTS
HTS-PL Assay Using Various Cultured Cells
The results of the HTS-PL assay are summarized in Figure 1. Uptake of NBD-PC by ARLJ-301 and HepG2 cells treated with amiodarone was increased approximately 50% over the corresponding controls, whereas other CADs were almost ineffective. Uptake of NBD-PC in rat primary hepatocytes treated with amiodarone was increased approximately 50%. Uptake of NBD-PC in human primary hepatocytes was unaffected by all of the CADs examined. Normalized values for rat and human primary hepatocytes were increased by treatment with promazine. In CHO-K1 and the CHL/IU, all of the CADs used at the concentration of 10 or 30 μM in this study induced accumulation of NBD-PC to about twice the control level (Fig. 1). Uptake of NBD-PC in J744A cells treated with 10 μM amiodarone was increased 200%, but other CADs had little effect (data not shown). Though the fluorescence intensity of Hoechst33342, used as an indicator of cell number, was different in various cells, chloroquine showed the strongest cytotoxicity in all of the cells, whereas chloramphenicol showed little cytotoxicity.
Confocal Laser Scanning Microscopy
Microplates used to observe the fluorescence intensity of NBD-PC and Hoechst33342 were also used for the confocal laser microscope observations, and fluorescence micrographs of cells treated with amiodarone are shown in Figure 2A. NBD-PC and Hoechst33342 were excited at 488 nm and 405 nm and showed green and blue fluorescence, respectively. Although human primary hepatocytes exhibited many fluorescent granules in both the control and amiodarone groups, the appearance frequency of fluorescent granules was obviously higher in the amiodarone group than in the control group for other cells. Uptake of NBD-PC could not be detected in the control group of CHO-K1 cells (Fig. 2B), though weak fluorescence of NBD-PC was apparent in the control groups of all other cell types (Fig. 2A). In CHO-K1 cells, the fluorescence intensity of NBD-PC was strong in the order of amiodarone, imipramine, propranolol, and control (Fig. 2B), and the fluorescence intensity of these compounds was almost correlated with results of HTS-PL assay using CHO-K1 cells (Fig.1).
Transmission laser observations of rat primary-cultured hepatocytes indicated that dead cells remained adhering to living cells. Confocal laser microscopy revealed that the dead cells bound NBD-PC nonspecifically, showing strong fluorescence. Nevertheless, the appearance frequency of fluorescent granules in living cells attached to the microplate appeared to reflect the PL-inducing action of CAD (Fig. 3).
Transmission Electron Microscopy
Figure 4 shows electron micrographs of typical cells. In human primary hepatocytes, many lamellar bodies were observed in both control and amiodarone-treated cells. However, in other cultured cells the appearance frequency of lamellar bodies was increased by CADs.
Correlation between HTS-PL Assay Results and Pathology Scores
CHO-K1 cells were treated with 24 compounds, including PL inducers and non-inducers, and the results of the HTS-PL assay (four concentrations) and the electron-microscopic pathology findings are shown in Table 1. The pathology findings were transformed into scores, and the Spearman rank correlation coefficient of the HTS-PL assay results with the pathology scores was 0.812 (Fig. 5).
DISCUSSION
This study was performed to develop a high-throughput assay for evaluation of PL-inducing potential (HTS-PL assay), by using a fluorescent probe. Target organs for PL include lung, brain, kidney, and cornea (Kodavanti and Mehendale, 1990), but little is known about the susceptibility of cell lines of various lineages. Therefore, we first examined the suitability of various types of cells (ARLJ301, HepG2, human and rat primary hepatocytes, CHO-K1, CHL/IU, and J744A), using six compounds with known PL-inducing potential (Joshi et al., 1988). ARLJ-301 and the HepG2 were considered unsuitable because they showed little response to the CADs, except for amiodarone, which induced only a 50% increase of NBD-PC uptake. The reason may be that ARLJ-301 and the HepG2 contain many lysosomes. In addition, ARLJ-301 has an unusual ultrastructure with many amorphous gaps in the cytoplasm. The macrophage-derived cell line J744A was also sensitive only to amiodarone among the CADs tested (data not shown). In human primary hepatocytes, PL was induced in both treated and control cells. Gum et al. (2001) observed drug-induced phospholipidosis in human primary hepatocytes, suggesting that it is possible that human primary hepatocytes show substantial inter-lot differences. The donor of the human hepatocytes used in our experiment was a 74-year-old diabetic. In rat primary hepatocytes, dead cells that strongly bound NBD-PC adhered to living cells, and the highly nonspecific fluorescence intensity meant that these cells were unsuitable for HTS-PL assay.
In the cases of CHO-K1 and CHL/IU, all the CADs tested induced at least 100% increase in accumulation of NBD-PC, and NBD-PC-uptake in cells treated with chloramphenicol (PL-noninducing compound) was at the same level as the control. Both cell lines were attached very strongly to the plate and were not readily lost during washing. Further, when the intensity of NBD-PC fluorescence was normalized with respect to the value of Hoechst33342 (used as indicator of amount of DNA), the degree of change versus the control increased, affording better sensitivity for detecting PL-inducing potential. According to Joshi et al. (1988), the phospholipidotic potency of various compounds was ranked as follows: amiodarone and chloroquine, "maximum effect"; imipramine and promazine "moderate effect"; propranolol, "mild effect"; chloramphenicol, "no effect." The order of the normalized values in HTS-PL assay using CHO-K1 and CHL/IU was similar to that of phospholipidotic potency reported by Joshi et al. Confocal laser microscopy revealed slight uptake of NBD-PC in the control group of CHL/IU, whereas the control CHO-K1 showed no uptake. Therefore, CHO-K1 was considered to be most suitable for use in the assay.
At concentrations of test drug that showed high cytotoxicity, the normalized value of NBD-PC uptake tended to increase sharply, and the order of PL-inducing potential ranked by normalized value became different from that at concentrations exhibiting little or no cytotoxicity. The reason for this may be that some cell fragments binding NBD-PC remained on the microplate after the cells had died, so that slight fluorescence of NDB-PC remained, whereas that of Hoechst33342 was removed with the nuclei of the dead cells, leading to an apparent increase in the normalized values. Thus, we recommend that PL-inducing potential should be evaluated at concentrations at which at least half of the cells survive in the control, as judged from the result with Hoechst33342.
We next selected 24 compounds randomly from among those used to establish representative values for PL markers (Sawada et al., 2005), and NBD-PC uptake was measured by HTS-PL assay using CHO-K1. The assay results showed a high correlation with the pathological scores for the 24 compounds. Thus, our HTS-PL assay is thought to be able to measure PL-inducing potential quantitatively. Among compounds with a pathological score of 3 or more, haloperidol had the lowest normalized value of 1.49. Accordingly, a suitable criterion for PL-inducing potential might be a normalized assay value of about 1.5 or more. However, to set the criterion for PL-inducing potential properly, it will be necessary to perform HTS-PL assays and electron microscopic observations using many more compounds.
Besides our HTS-PL assay, methods based on flow cytometric monitoring of intracellular dye after staining cellular phospholipids with nile red (Xia et al., 1997) and spectrofluorometeric measurement of fluorescence intensity in cell lysate after uptake of NBD-PC (Ulrich et al., 1991) have been reported. These methods are complex and time-consuming, whereas our method is able to detect PL more conveniently in a shorter time and is more efficient in determining these points.
Because sensitivity to the PL-inducing activity of compounds is dependent on species, strain, and age, HTS-PL assay might be not able to predict sensitivity to PL in humans reliably. However, the HTS-PL assay has high sensitivity and accuracy when only a small dose of test compound is used, and it can rapidly detect the PL induction potency of multiple compounds at the same time, with simultaneous determination of cytotoxicity. Therefore, because this assay is able to rank compounds for PL-inducing potency, it should be useful as a screening tool for PL-inducing potential of new compounds at an early stage in drug development.
ACKNOWLEDGMENTS
We thank Mr. Toru Yokoyama, Mr. Makoto Ozawa, and Mrs. Tomoko Honda for technical assistance with cell cultures. We also thank Mr. Takayuki Sekido and Mr. Shin-ichi Akimoto for their skillful assistance.
REFERENCES
Gum, R. J., Hickman, D., Fagerland, J. A., Heindel, M. A., Gagne, G. D., Schmidt, J. M., Michaelides, M. R., Davidsen, S. K., and Ulrich, R. G. (2001). Analysis of two matrix metalloproteinase inhibitors and their metabolites for induction of phospholipidosis in rat and human hepatocytes (1). Biochem. Pharmacol. 62, 1661–1673.
Joshi, U. M., Kodavanti, P. R., Coudert, B., Dwyer, T. M., and Mehendale, H. M. (1988). Types of interaction of amphiphilic drugs with phospholipid vesicles. J. Pharmacol. Exp. Ther. 246, 150–157.
Kacew, S. (1982). Alterations in newborn and adult rat lung morphology and phospholipid levels after chlorcyclizine or chlorphentermine treatment. Toxicol. Appl. Pharmacol. 65, 100–108.
Kodavanti, U. P., and Mehendale, H. M. (1990). Cationic amphiphilic drugs and phospholipid storage disorder. Pharmacol. Rev. 42, 327–354.
Kremer, J. M. H., von der Esker, W. J., Pathmamanoharan, C., and Wiersema, P. H. (1977). Vesicles of variable diameter prepared by a modified injection method. Biochemistry 16, 3932–3935.
Lullmann Rauch, R., and Reil, G. H. (1974). Chlorphentermine-induced lipidosislike ultrastructural alterations in lungs and adrenal glands of several species. Toxicol. Appl. Pharmacol. 30, 408–421.
Reasor, M. J., McCloud, C. M., Beard, T. L., Ebert, D. C., Kacew, S., Gardner, M. F., Aldern, K. A., and Hostetler, K. Y. (1996). Comparative evaluation of amiodarone-induced phospholipidosis and drug accumulation in Fischer-344 and Sprague-Dawley rats. Toxicology 106, 139–147.
Roger, G. U., Kenneth, S. K., Elena, L. S., Clay, T. C., and Leonard, C. G. (1991). An in vitro fluorescence assay for the detection of drug-induced cytoplasmic lamellar bodies. Toxicol. Meth. 1, 89–105.
Sawada, H., Takami, K., and Asahi, S. (2005). A toxicogenomic approach to drug-induced phospholipidosis: Analysis of its induction mechanism and establishment of a novel in vitro screening system. Toxicol. Sci. 83, 282–292.
Seiler, K. U., and Wassermann, O. (1975). Drug-induced phospholipidosis. II. Alterations in the phospholipid pattern of organs from mice, rats and guinea-pigs after chronic treatment with chlorphentermine. Naunyn Schmiedebergs Arch. Pharmacol. 288, 261–268.
Ulrich, R. G., Kligore, K. S., Sun, E. L., Cramer, C. T., and Ginsberg, L. C. (1991). An in vitro fluorescence assay for the detection of drug-induced cytoplasmic lamellar bodies. Toxicol. Meth. 1, 89–105.
Uno, Y., Takasawa, H., Miyagawa, M., Inoue, Y., Murata, T., Ogawa, M., and Yoshikawa, K. (1992). In vivo–in vitro replicative DNA synthesis (RDS) test using perfused rat livers as an early prediction assay for nongenotoxic hepatocarcinogens: I. Establishment of a standard protocol. Toxicol. Lett. 63, 191–199.
Xia, Z., Appelkvist, E. L., DePierre, J. W., Nassberger, L. (1997). Tricyclic antidepressant-induced lipidosis in human peripheral monocytes in vitro, as well as in a monocyte-derived cell line, as monitored by spectrofluorimetry and flow cytometry after staining with Nile red. Biochem Pharmacol. 53, 1521–1532.
Yamamoto, A., Adachi, S., Ishikawa, K., Yokomura, T., and Kitani, T. (1971a). Studies on drug-induced lipidosis. 3. Lipid composition of the liver and some other tissues in clinical cases of "Niemann-Pick-like syndrome" induced by 4,4'-diethylaminoethoxyhexestrol. J. Biochem. (Tokyo) 70, 775–784.
Yamamoto, A., Adachi, S., Kitani, T., Shinji, Y., and Seki, K. (1971b). Drug-induced lipidosis in human cases and in animal experiments. Accumulation of an acidic glycerophospholipid. J. Biochem. (Tokyo) 69, 613–615(Toshihiko Kasahara, Kazuo Tomita, Hiroyu)
Development Research, Mochida Pharmaceutical Company Limited, Shizuoka, Japan
ABSTRACT
Excessive accumulation of phospholipids results in phospholipidosis (PL), which may interfere with cellular functions, leading to acute or chronic disease or even death. Electron-microscopic detection of cytoplasmic lamellar bodies is often used as a diagnostic criterion of PL, but a faster, more convenient procedure is required for high-throughput assay of the PL-inducing potential of candidate drugs. We have developed a 96-well microplate cell-culture method for detecting PL, using a phosphatidylcholine-conjugated dye (NBD-PC) and a fluoro-microplate reader. The fluorescence intensity due to NBD-PC was normalized to that of Hoechst33342, used as an indicator of cell number, to obtain the amount of NBD-PC taken up per living cell. To select a suitable cell type, we examined the PL-detection sensitivity of five cell lines, as well as human and rat primary hepatocyte cultures, with five cationic amphiphilic drugs (CAD) as PL inducers and a negative control compound. The cell lines CHO-K1 and CHL/IU gave the best results. The NBD-PC uptake per CHO-K1 cell showed a high correlation with the pathological score of PL for 24 compounds, including PL-positive and negative compounds. This high-throughput screening assay for PL-inducing potential (HTS-PL assay) offers high sensitivity and accuracy, and it allows simultaneous determination of cytotoxicity.
Key Words: phospholipidosis; cationic amphiphilic drug; NBD-PC; Hoechst33342; HTS-PL assay.
INTRODUCTION
Phospholipidosis (PL) is a pathological accumulation of phospholipids in tissue, characterized by the formation of lysosomal lamellar bodies, and may be induced by certain drugs with a cationic lipophilic structure (cationic amphiphilic drugs; CAD). The accumulated phospholipids interfere with cellular functions, sometimes with fatal results, as in the case of PL induced by diethylaminoethoxyhexestrol (Yamamoto et al., 1971a). Therefore, it is very important to investigate whether drugs under development have the potential to induce PL. Electron microscopic detection of cytoplasmic lamellar bodies is thought to be the most reliable method for identifying PL, but this is unsuitable for high-throughput systems, which are required for assay of large numbers of compounds generated by combinatorial chemistry. Roger et al. (1991) reported that phospholipid accumulation could be detected by means of fluorescence microscopy, following the incorporation of a fluorescent phospholipid analog, 1-acyl-2-[12-(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]dodecanoyl]-sn-glycero-3-phosphocholine (NBD-PC), into cytoplasm. Therefore, we sought to establish a high-throughput screening assay (HTS-PL assay) to assay PL-inducing potential quantitatively and rapidly in a 96-well microplate format, using NBD-PC and cultured cells. Phospholipidosis is usually induced after chronic intake of CADs, with a latency period ranging from a few days to several months, depending on species (Lullmann-Rauch et al., 1974; Seiler and Wassermann, 1975; Yamamoto et al., 1971b), strain within species (Reasor et al., 1996) and age (Kacew et al., 1982). Therefore, we examined a range of cell types (human or rat, cultured or primary-cultured) using five PL-inducing CADs and a negative control in order to find suitable cells for the assay.
MATERIALS AND METHODS
Chemicals.
Test compounds known to cause or not to cause phospholipidosis or not in animals and/or humans in vivo were purchased from the following sources: amiodarone, imipramine, promazine, chloroquine, chloramphenicol, propranolol, acetaminophen, amitriptyline hydrochloride, AY9944, chlorpromazine, clomipramine hydrochloride, clozapine, disopyramide phosphate salt, flecainide, fluoxetine hydrochloride, haloperidol, ketoconazole, loratadine, quinidine, sotalol, tamoxifen citrate, thioridazine hydrochloride, and zimelidine from Sigma Chemical (St. Louis, MO, USA); clarithromycin and erythromycin from Wako Pure Chemicals (Osaka, Japan); levofloxacin from LKT Laboratories (St. Paul, MN, USA). Chloroquine, disopyramide phosphate salt, imipramine, promazine, and propranolol were dissolved in water and other compounds were dissolved in dimethylsulfoxide (DMSO) to afford 100 mM solutions.
Cell culture.
The human hepatocellular carcinoma cell line (HepG2), diploid rat liver epithelial cell line (ARLJ301–3), Chinese hamster ovary cell line (CHO-K1), Chinese hamster lung cell line (CHL/IU), and mouse macrophage-like cell line (J744A) were purchased from Dainippon Pharmaceutical Co. Ltd. (Osaka, Japan), and all culture reagents were purchased from GIBCO BRL (Grand Island, NY, USA), unless otherwise noted. Cell lines were cultured in the following media: HepG2 and ARLJ301–3 in Williams' medium (WE) containing 10% fetal bovine serum (FBS), CHO-K1 in D-MEM/F-12 containing 10% FBS, CHL/IU in modified Eagle's medium (MEM) containing 10% calf serum (CS), J744A in Dulbecco's-MEM (D-MEM) containing 10% FBS. All media were supplemented with 100 μg/ml streptomycin and 100 U/ml penicillin. These cells were used in the logarithmic phase of growth and were seeded in wells of 96-well microplates for HTS-PL assay using NBD-PC. For the confocal laser scanning study, cell density per well was as follows: HepG2, 3 x 104 cells; ARLJ301, 2 x 104 cells; CHO-K1, 1 x 104 cells; CHL/IU, 1 x 104 cells; J744A, 4 x 104 cells. Primary hepatocyte cultures were prepared from a male Sprague-Dawley rat at 10 weeks of age by the method of Uno et al. (1992). The rat was anaesthetized with sodium pentobarbital and the liver was perfused in situ for 5 min at flow rate of 10 ml/min with Hank's balanced salt solution (Ca2+/Mg2+-free) containing 0.5 mM EGTA (Sigma) and 10 mM HEPES (pH 7.2–7.3), followed by 7 min in Hanks' balanced salt solution containing 0.05% collagenase S-1 (Nitta Gelatin Inc., Osaka, Japan), 50 μg/ml trypsin inhibitor (Sigma), 10 mg/ml bovine serum albumin (Sigma), 10 mM HEPES, and 560 μg/ml CaCl2 (pH 7.5). All solutions were kept at 40°C for use. The perfused liver was minced with surgical scissors in a Petri dish containing Hank's balanced salt solution, and filtered through a cell strainer with 100-μm nylon mesh (Becton Dickinson, San Jose, CA, USA). Rat hepatocytes were washed twice by centrifugation at 50 x g for 1 min and resuspended in culture medium for primary hepatocytes, WE with additives (insulin, hydrocortisone, transferrin, hEGF, L-ascorbic acid), selected from the HCM SingleQuots Kits (Cambrex Co., North Brunswick, NJ, USA), 100 μg/ml streptomycin, 100 U/ml penicillin, and 10% FBS. The number and viability of cells were determined using a conventional trypan blue exclusion test. Human primary plateable cryopreserved hepatocytes were purchased from In Vitro Technologies, Inc. (Baltimore, MD, USA), and were warmed quickly to 37°C in a water bath. Percoll was purchased from Sigma, and isotonic Percoll solution was prepared by mixing Percoll (90 ml) with 9% NaCl (10 ml). Dead cells were removed by centrifugation at 50 x g for 10 min in an isolation solution prepared by mixing isotonic Percoll solution (15 ml) with WE medium (35 ml). The pelleted human hepatocytes were resuspended in culture medium for primary hepatocytes, washed twice at 50 x g for 5 min, and examined with a conventional trypan blue exclusion test. Rat and human hepatocytes were plated at density of 1 x 104 rat cells or 2 x 104 human cells per well on a collagen-coated 96-well microplate (Becton, Dickinson). After an initial attachment period (about 4 h), the hepatocytes were washed once with the medium. For transmission electron microscopic examination, 10 times higher cell numbers were seeded into 6-well microplates. All cell cultures were conducted at 37°C under 95% air and 5% CO2.
Drug treatment.
Serial dilutions of compounds in the medium were prepared by using deep-well microplates (chemical preparation plate). Medium was added to columns of wells on the chemical preparation plate as follows; column A, 998 μl; columns B and D, 700 μl; columns C and E, 600 μl. A 2-μl aliquot of test compound (100 mM) dissolved in DMSO or water was added to the wells in column A, then 300 μl of medium was transferred from these wells to the wells in column B. Similarly, successive transfers resulted in final concentrations of 200, 60, 20, 6, and 2 μM in duplicate (six compounds/plate). For experiments performed to investigate the correlation between assay results and pathological changes in CHO-K1 cells, plates were similarly prepared, but with five compounds/plate in replicates of four. In chemical preparation plates, the concentrations of compounds dissolved in the medium were twice the desired final concentrations.
NBD-PC preparation.
The fluorescent phospholipid analog, 1-acyl-2-[12-(7-nitro-2,1,3-benzoxadiazol-4-yl) amino]dodecanoyl]-sn-glycero-3- phosphocholine (NBD-PC; Avanti Polar Lipids, Inc., AL, USA) was prepared according to the method of Kremer et al. (1977). A stock solution of NBD-PC was prepared at 10 mM in chloroform, divided into 100 μl aliquots in vials, and dried. Vials were stored at –20°C and reconstituted in 100 μl of 100% ethanol for use. This solution was added to the medium to make a concentration of 80 μM; the mixture was then sonicated for 30 min and passed through a 0.2-μm disposable filter. Medium containing the fluorescent phospholipid analog is referred to as "NBD-PC medium."
HTS-PL assay using NBD-PC.
The HTS-PL assay was established by modifying Ulrich's method (1991). At 24 h after plating the cells, medium in each well of the 96-microplate was removed, 50 μl of medium containing 80 μM NBD-PC was added quickly, and then 50 μl of medium containing various test compounds at serial twofold concentration was transferred to the cell culture plate from the chemical preparation plate. At 24 h after addition of NBD-PC and test compounds, the plated cells were washed twice with phosphate-buffered saline (PBS), and 50 μl of PBS was added to each well. Fluorescence of NBD-PC taken up by cells was measured with a fluoro-microplate reader (Thermo Electron Oy, Vantaa, Finland) with an excitation wavelength of 485 nm and emission wavelength of 538 nm. After measuring the fluorescence of NBD-PC, and 50 μl of PBS containing 20 μg/ml Hoechst33342 (ICN Pharmaceutical, Inc., Costa Mesa, CA, USA) was added. Incubation was continued for 20 min at 37°C, after which the cellular fluorescence of Hoechst33342 was measured using a fluoro-microplate reader with an excitation wavelength of 355 nm and emission wavelength of 460 nm. To remove the influence of cytotoxicity, the uptake of NBD-PC per cell was calculated using the following formula.
Confocal laser scanning microscopy.
Cellular fluorescent lipid was observed using an Olympus confocal laser microscope FV1000 (Olympus, Tokyo, Japan) equipped 20x and 40x fluoro lenses. Microplates read with the fluoro-microplate reader for HTS-PL assay were also used for confocal laser scanning microscopy. Fluorescence of NBD-PC and that of Hoechst33342 were visualized as green and blue, respectively, by dual-wavelength excitation (NBD-PC, 488 nm; Hoechst33342, 405 nm) and indicated accumulation of phospholipids in cytoplasm and the location of the nucleus, respectively. In the case of rat primary hepatocytes, transmission laser micrographs were taken of the same microscopic field observed for the fluorescence of NBD-PC and Hoechst33342 to investigate cell morphology.
Transmission electron microscopy.
Cells cultured in 6-well microplates under the conditions described above were fixed with cold 2% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2, PB) for 2 h, then washed with PB, and postfixed with 1.0% osmium tetroxide in PB for 60 min. They were again rinsed with the same PB, dehydrated with graded ethanol through to 100%, infiltrated with Epok 812 (Oken, Tokyo, Japan), and polymerized for 3 days. Thin sections were cut with an ultramicrotome (Reichent, Vienna, Austria), and stained with saturated uranyl acetate and lead acetate. The degree of formation of lamellar bodies in lysosomes was evaluated with a JEM-100CX2 electron microscope (JEOL, Tokyo, Japan) and scored on a scale of – to ++++ (–, within normal limits; ±, very slight; +, slight; ++, moderate; +++, severe; ++++, very severe) in a blind fashion.
RESULTS
HTS-PL Assay Using Various Cultured Cells
The results of the HTS-PL assay are summarized in Figure 1. Uptake of NBD-PC by ARLJ-301 and HepG2 cells treated with amiodarone was increased approximately 50% over the corresponding controls, whereas other CADs were almost ineffective. Uptake of NBD-PC in rat primary hepatocytes treated with amiodarone was increased approximately 50%. Uptake of NBD-PC in human primary hepatocytes was unaffected by all of the CADs examined. Normalized values for rat and human primary hepatocytes were increased by treatment with promazine. In CHO-K1 and the CHL/IU, all of the CADs used at the concentration of 10 or 30 μM in this study induced accumulation of NBD-PC to about twice the control level (Fig. 1). Uptake of NBD-PC in J744A cells treated with 10 μM amiodarone was increased 200%, but other CADs had little effect (data not shown). Though the fluorescence intensity of Hoechst33342, used as an indicator of cell number, was different in various cells, chloroquine showed the strongest cytotoxicity in all of the cells, whereas chloramphenicol showed little cytotoxicity.
Confocal Laser Scanning Microscopy
Microplates used to observe the fluorescence intensity of NBD-PC and Hoechst33342 were also used for the confocal laser microscope observations, and fluorescence micrographs of cells treated with amiodarone are shown in Figure 2A. NBD-PC and Hoechst33342 were excited at 488 nm and 405 nm and showed green and blue fluorescence, respectively. Although human primary hepatocytes exhibited many fluorescent granules in both the control and amiodarone groups, the appearance frequency of fluorescent granules was obviously higher in the amiodarone group than in the control group for other cells. Uptake of NBD-PC could not be detected in the control group of CHO-K1 cells (Fig. 2B), though weak fluorescence of NBD-PC was apparent in the control groups of all other cell types (Fig. 2A). In CHO-K1 cells, the fluorescence intensity of NBD-PC was strong in the order of amiodarone, imipramine, propranolol, and control (Fig. 2B), and the fluorescence intensity of these compounds was almost correlated with results of HTS-PL assay using CHO-K1 cells (Fig.1).
Transmission laser observations of rat primary-cultured hepatocytes indicated that dead cells remained adhering to living cells. Confocal laser microscopy revealed that the dead cells bound NBD-PC nonspecifically, showing strong fluorescence. Nevertheless, the appearance frequency of fluorescent granules in living cells attached to the microplate appeared to reflect the PL-inducing action of CAD (Fig. 3).
Transmission Electron Microscopy
Figure 4 shows electron micrographs of typical cells. In human primary hepatocytes, many lamellar bodies were observed in both control and amiodarone-treated cells. However, in other cultured cells the appearance frequency of lamellar bodies was increased by CADs.
Correlation between HTS-PL Assay Results and Pathology Scores
CHO-K1 cells were treated with 24 compounds, including PL inducers and non-inducers, and the results of the HTS-PL assay (four concentrations) and the electron-microscopic pathology findings are shown in Table 1. The pathology findings were transformed into scores, and the Spearman rank correlation coefficient of the HTS-PL assay results with the pathology scores was 0.812 (Fig. 5).
DISCUSSION
This study was performed to develop a high-throughput assay for evaluation of PL-inducing potential (HTS-PL assay), by using a fluorescent probe. Target organs for PL include lung, brain, kidney, and cornea (Kodavanti and Mehendale, 1990), but little is known about the susceptibility of cell lines of various lineages. Therefore, we first examined the suitability of various types of cells (ARLJ301, HepG2, human and rat primary hepatocytes, CHO-K1, CHL/IU, and J744A), using six compounds with known PL-inducing potential (Joshi et al., 1988). ARLJ-301 and the HepG2 were considered unsuitable because they showed little response to the CADs, except for amiodarone, which induced only a 50% increase of NBD-PC uptake. The reason may be that ARLJ-301 and the HepG2 contain many lysosomes. In addition, ARLJ-301 has an unusual ultrastructure with many amorphous gaps in the cytoplasm. The macrophage-derived cell line J744A was also sensitive only to amiodarone among the CADs tested (data not shown). In human primary hepatocytes, PL was induced in both treated and control cells. Gum et al. (2001) observed drug-induced phospholipidosis in human primary hepatocytes, suggesting that it is possible that human primary hepatocytes show substantial inter-lot differences. The donor of the human hepatocytes used in our experiment was a 74-year-old diabetic. In rat primary hepatocytes, dead cells that strongly bound NBD-PC adhered to living cells, and the highly nonspecific fluorescence intensity meant that these cells were unsuitable for HTS-PL assay.
In the cases of CHO-K1 and CHL/IU, all the CADs tested induced at least 100% increase in accumulation of NBD-PC, and NBD-PC-uptake in cells treated with chloramphenicol (PL-noninducing compound) was at the same level as the control. Both cell lines were attached very strongly to the plate and were not readily lost during washing. Further, when the intensity of NBD-PC fluorescence was normalized with respect to the value of Hoechst33342 (used as indicator of amount of DNA), the degree of change versus the control increased, affording better sensitivity for detecting PL-inducing potential. According to Joshi et al. (1988), the phospholipidotic potency of various compounds was ranked as follows: amiodarone and chloroquine, "maximum effect"; imipramine and promazine "moderate effect"; propranolol, "mild effect"; chloramphenicol, "no effect." The order of the normalized values in HTS-PL assay using CHO-K1 and CHL/IU was similar to that of phospholipidotic potency reported by Joshi et al. Confocal laser microscopy revealed slight uptake of NBD-PC in the control group of CHL/IU, whereas the control CHO-K1 showed no uptake. Therefore, CHO-K1 was considered to be most suitable for use in the assay.
At concentrations of test drug that showed high cytotoxicity, the normalized value of NBD-PC uptake tended to increase sharply, and the order of PL-inducing potential ranked by normalized value became different from that at concentrations exhibiting little or no cytotoxicity. The reason for this may be that some cell fragments binding NBD-PC remained on the microplate after the cells had died, so that slight fluorescence of NDB-PC remained, whereas that of Hoechst33342 was removed with the nuclei of the dead cells, leading to an apparent increase in the normalized values. Thus, we recommend that PL-inducing potential should be evaluated at concentrations at which at least half of the cells survive in the control, as judged from the result with Hoechst33342.
We next selected 24 compounds randomly from among those used to establish representative values for PL markers (Sawada et al., 2005), and NBD-PC uptake was measured by HTS-PL assay using CHO-K1. The assay results showed a high correlation with the pathological scores for the 24 compounds. Thus, our HTS-PL assay is thought to be able to measure PL-inducing potential quantitatively. Among compounds with a pathological score of 3 or more, haloperidol had the lowest normalized value of 1.49. Accordingly, a suitable criterion for PL-inducing potential might be a normalized assay value of about 1.5 or more. However, to set the criterion for PL-inducing potential properly, it will be necessary to perform HTS-PL assays and electron microscopic observations using many more compounds.
Besides our HTS-PL assay, methods based on flow cytometric monitoring of intracellular dye after staining cellular phospholipids with nile red (Xia et al., 1997) and spectrofluorometeric measurement of fluorescence intensity in cell lysate after uptake of NBD-PC (Ulrich et al., 1991) have been reported. These methods are complex and time-consuming, whereas our method is able to detect PL more conveniently in a shorter time and is more efficient in determining these points.
Because sensitivity to the PL-inducing activity of compounds is dependent on species, strain, and age, HTS-PL assay might be not able to predict sensitivity to PL in humans reliably. However, the HTS-PL assay has high sensitivity and accuracy when only a small dose of test compound is used, and it can rapidly detect the PL induction potency of multiple compounds at the same time, with simultaneous determination of cytotoxicity. Therefore, because this assay is able to rank compounds for PL-inducing potency, it should be useful as a screening tool for PL-inducing potential of new compounds at an early stage in drug development.
ACKNOWLEDGMENTS
We thank Mr. Toru Yokoyama, Mr. Makoto Ozawa, and Mrs. Tomoko Honda for technical assistance with cell cultures. We also thank Mr. Takayuki Sekido and Mr. Shin-ichi Akimoto for their skillful assistance.
REFERENCES
Gum, R. J., Hickman, D., Fagerland, J. A., Heindel, M. A., Gagne, G. D., Schmidt, J. M., Michaelides, M. R., Davidsen, S. K., and Ulrich, R. G. (2001). Analysis of two matrix metalloproteinase inhibitors and their metabolites for induction of phospholipidosis in rat and human hepatocytes (1). Biochem. Pharmacol. 62, 1661–1673.
Joshi, U. M., Kodavanti, P. R., Coudert, B., Dwyer, T. M., and Mehendale, H. M. (1988). Types of interaction of amphiphilic drugs with phospholipid vesicles. J. Pharmacol. Exp. Ther. 246, 150–157.
Kacew, S. (1982). Alterations in newborn and adult rat lung morphology and phospholipid levels after chlorcyclizine or chlorphentermine treatment. Toxicol. Appl. Pharmacol. 65, 100–108.
Kodavanti, U. P., and Mehendale, H. M. (1990). Cationic amphiphilic drugs and phospholipid storage disorder. Pharmacol. Rev. 42, 327–354.
Kremer, J. M. H., von der Esker, W. J., Pathmamanoharan, C., and Wiersema, P. H. (1977). Vesicles of variable diameter prepared by a modified injection method. Biochemistry 16, 3932–3935.
Lullmann Rauch, R., and Reil, G. H. (1974). Chlorphentermine-induced lipidosislike ultrastructural alterations in lungs and adrenal glands of several species. Toxicol. Appl. Pharmacol. 30, 408–421.
Reasor, M. J., McCloud, C. M., Beard, T. L., Ebert, D. C., Kacew, S., Gardner, M. F., Aldern, K. A., and Hostetler, K. Y. (1996). Comparative evaluation of amiodarone-induced phospholipidosis and drug accumulation in Fischer-344 and Sprague-Dawley rats. Toxicology 106, 139–147.
Roger, G. U., Kenneth, S. K., Elena, L. S., Clay, T. C., and Leonard, C. G. (1991). An in vitro fluorescence assay for the detection of drug-induced cytoplasmic lamellar bodies. Toxicol. Meth. 1, 89–105.
Sawada, H., Takami, K., and Asahi, S. (2005). A toxicogenomic approach to drug-induced phospholipidosis: Analysis of its induction mechanism and establishment of a novel in vitro screening system. Toxicol. Sci. 83, 282–292.
Seiler, K. U., and Wassermann, O. (1975). Drug-induced phospholipidosis. II. Alterations in the phospholipid pattern of organs from mice, rats and guinea-pigs after chronic treatment with chlorphentermine. Naunyn Schmiedebergs Arch. Pharmacol. 288, 261–268.
Ulrich, R. G., Kligore, K. S., Sun, E. L., Cramer, C. T., and Ginsberg, L. C. (1991). An in vitro fluorescence assay for the detection of drug-induced cytoplasmic lamellar bodies. Toxicol. Meth. 1, 89–105.
Uno, Y., Takasawa, H., Miyagawa, M., Inoue, Y., Murata, T., Ogawa, M., and Yoshikawa, K. (1992). In vivo–in vitro replicative DNA synthesis (RDS) test using perfused rat livers as an early prediction assay for nongenotoxic hepatocarcinogens: I. Establishment of a standard protocol. Toxicol. Lett. 63, 191–199.
Xia, Z., Appelkvist, E. L., DePierre, J. W., Nassberger, L. (1997). Tricyclic antidepressant-induced lipidosis in human peripheral monocytes in vitro, as well as in a monocyte-derived cell line, as monitored by spectrofluorimetry and flow cytometry after staining with Nile red. Biochem Pharmacol. 53, 1521–1532.
Yamamoto, A., Adachi, S., Ishikawa, K., Yokomura, T., and Kitani, T. (1971a). Studies on drug-induced lipidosis. 3. Lipid composition of the liver and some other tissues in clinical cases of "Niemann-Pick-like syndrome" induced by 4,4'-diethylaminoethoxyhexestrol. J. Biochem. (Tokyo) 70, 775–784.
Yamamoto, A., Adachi, S., Kitani, T., Shinji, Y., and Seki, K. (1971b). Drug-induced lipidosis in human cases and in animal experiments. Accumulation of an acidic glycerophospholipid. J. Biochem. (Tokyo) 69, 613–615(Toshihiko Kasahara, Kazuo Tomita, Hiroyu)