Systems-ADME/Tox: Resources and Network Approaches

Journal of Pharmacological and Toxicological Methods 53 (2006) 38 – 66 www.elsevier.com/locate/jpharmtox Appraisal of state-of-the art Systems-ADME/Tox: Resources and network approaches

Sean Ekins *

GeneGo, 500 Renaissance Drive, Suite 106, St. Joseph, MI 49085, USA School of Pharmacy Department of Pharmaceutical Sciences, University of Maryland, USA

Received 23 May 2005; accepted 23 May 2005

Abstract

The increasing cost of drug development is partially due to our failure to identify undesirable compounds at an early enough stage of development. The application of higher throughput screening methods have resulted in the generation of very large datasets from cells in vitro or from in vivo experiments following the treatment with drugs or known toxins. In recent years the development of systems biology, databases and pathway software has enabled the analysis of the high-throughput data in the context of the whole cell. One of the latest technology paradigms to be applied alongside the existing in vitro and computational models for absorption, distribution, metabolism, excretion and toxicology (ADME/Tox) involves the integration of complex multidimensional datasets, termed toxicogenomics. The goal is to provide a more complete understanding of the effects a molecule might have on the entire biological system. However, due to the sheer complexity of this data it may be necessary to apply one or more different types of computational approaches that have as yet not been fully utilized in this field. The present review describes the data generated currently and introduces computational approaches as a component of ADME/Tox. These methods include network algorithms and manually curated databases of interactions that have been separately classified under systems biology methods. The integration of these disparate tools will result in systems-ADME/Tox and it is important to understand exactly what data resources and technologies are available and applicable. Examples of networks derived with important drug transporters and drug metabolizing enzymes are provided to demonstrate the network technologies. D 2005 Elsevier Inc. All rights reserved.

Keywords: Algorithms; Human; Microarray; Mouse; Networks; Rat; Software; Toxicogenomics; Toxicoproteomics

1. Introduction in the industry as they have been evaluated primarily for toxicology and metabolism assessment (Waring et al., Metabolism and safety assessment have witnessed some 2003, 2002, 2001) with some considerable focus on growth in the number of new technologies and methods hepatotoxicity (Harris, Dial, & Casciano, 2004; Hartley that have been introduced within the last decade. However, et al., 2004; Heijne et al., 2004; Huang et al., 2004; according to a recent FDA white paper there is still Liguori et al., 2005; Ulrich, Rockett, Gibson, & Pettit, considerable scope for additional new methods (FDA, 2004). For example, searching PubMed for publications in 2004). For example, recently various reports have the last 5 years with the keywords Fmicroarray and described new software and methods for metabolism toxicology_ or Ftoxicogenomics_, indicates that the accu- prediction (Balakin et al., 2004a, 2004b; Borodina et al., mulation of papers describing the latter is doubling every 2003; Borodina et al., 2004; Boyer & Zamora, 2002; year (Fig. 1) which perhaps is mirrored by the application Korolev et al., 2003). Simultaneously the use of high in the pharmaceutical industry for predictive toxicology throughput (HT) methods for genomics, proteomics and (Suter, Babiss, & Wheeldon, 2004). metabonomics have taken off in terms of their acceptance To date toxicogenomics experiments have been carried out under non-standardized conditions. Most of the studies * Corresponding author. Tel.: +1 269 930 0974; fax: +1 269 983 7654. have been conducted with rats, less often mice, using E-mail addresses: [email protected], [email protected]. multiple different microarray formats and statistical proce-

100 The perturbing effect of a molecule on the complete 80 biological system can be observed across all metabolic and signaling pathways or networks and can provide some 60 limited insight into the binding to multiple proteins or effects on gene expression simultaneously. This requires the 40

PubMed collection of high-throughput data, including global gene expression, protein content, metabolic profiles for the same 20 samples as well as individual genetic, clinical and pheno-

Number of publications in 0 typic data. However there are difficulties with such an 2000 2001 2002 2003 2004 2005 approach as there are likely to be differences between the Year Fstimulus to effect_ durations for all the gene–protein relationships (Nicholson, Holmes, Lindon, & Wilson, Fig. 1. Annual frequency of articles appearing with the words ‘‘toxicogenomics’’ (squares) or ‘‘microarray and toxicology’’ (diamonds). 2004). We can now use either the growing number of academic or commercially available pathway database and network dures (Tables 1-4). There have been relatively few cross- building tools with expression data. These enable the platform toxicogenomics studies under controlled condi- connection of interacting, differentially expressed genes as tions (Thompson et al., 2004). Despite the suggested poor networks (Barabasi & Oltvai, 2004; Hanisch, Zien, Zimmer, compatibility between the different array types, this latter & Lengauer, 2002; Ideker, Ozier, Schwikowski, & Siegal, study demonstrated a high (90%) consistency between the 2002; Ideker et al., 2001; Segal et al., 2003a; Segal, Wang, & expression of the genes that were shared between the Koller, 2003b; Spirin & Mirny, 2003; Tornow & Mewes, platforms. The development of methods to visualize such 2003) as well as allowing the reverse engineering of complex expression data has also expanded beyond the functional connections (Somogyi, Fuhrman, & Wen, 2001). widely used clustering methods (Eisen, Spellman, Brown, & The use of such network visualizations suggests an Botstein, 1998). With the outcome of microarray analysis organized modularity in complex systems (Han et al., being dependent on the widely used statistical procedures 2004) which has also been applied to interpret the applied to derive those genes that are significantly differ- connectivity of small molecules and their interaction with entially expressed (Butte, 2002), newer approaches that do proteins in the subfield of chemogenomics (Bredel & Jacoby, not necessarily require data clustering may be an advantage. 2004; Csermely, Agoston, & Pongor, 2005; Parsons et al., As rat and mouse are the most widely used in vivo toxicity 2004; Sharom, Bellows, & Tyers, 2004). The parallel models it is assumed that acute and chronic toxicity shown development of HT methods, databases, ADME/Tox model- in animals largely coincides with human toxicity. Therefore ing and systems modeling is ongoing (Ekins, Nikolsky, & differential expression patterns in animal models are also Nikolskaya, 2005e). The present review is therefore timely assumed to be predictive of the end point toxic response in as it discusses some of the data resources, limitations and human. This is not always the case due to differences technologies that are available for Systems-ADME/Tox between human and rodent physiology, genetics, metabo- (Fig. 2) along with some illustration of their applications to lism and signaling pathways. For example, the mechanism drug metabolism and drug transport which are key com- of toxicity for pyrazinamide has been reconstructed ponents of the ADME/Tox process. The ultimate aim of (Bugrim, Nikolskaya, & Nikolsky, 2004) to illustrate that this discussion is to provide awareness of an integrated the accumulation of uric acid occurs in human, but not in approach rather than a technology silo mentality, represent- mice, and this results in toxicity in the former. The relatively ing the latest proposed research model in the field (Fig. 2). poor understanding of such species differences may be reflected in the relatively large number of late stage molecules that have undergone in vivo toxicity assessment 2. Data available yet have been later withdrawn due to adverse events in humans. A recently published book provides an excellent over- In recent years the appearance of systems biology which view of toxicogenomics and the reader is referred to this to uses the relationships of all elements of biology rather than gain more insight into the applications and limitations approaching them separately has been evident and will (Hamadeh & Afshari, 2004). The growing number of likely reunite biological fields (Harrison, 2004; Hood & toxicogenomic datasets derived from in vivo studies with Galas, 2003). These systems approaches are the latest rat (Table 1) and mouse (Table 2) as well as in vitro cell incarnation of the importance of the Fparts vs wholes_ derived data (Table 3) highlights the different strains, debate (Ekins & McGowan, 2001) and interpreting ADME/ microarray types and compounds that are routinely assessed. Tox in this context may improve our understanding and Also there are numerous instances of multiple groups testing ultimate predictions (Bugrim et al., 2004; Ekins, Boulanger, the same compound at similar or different doses, e.g. well Swaan, & Hupcey, 2002a; Kitano, 2002a; Werner, 2003). known hepatotoxicants or nephrotoxicants such as clofibrate 40 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Table 1 Literature toxicogenomics data derived from rat in vivo studies Compounds Rat strain Microarray type Compound dose Microarray data Reference availability Acetaminophen; furan; Male Sprague– Phase 1 Molecular Acetaminophen Gene name, accession (Huang et al., methotrexate; Dawley VAF+ toxicology array (4500 mg/kg/day); number and fold changes 2004) methapyrilene; albino methotrexate (1 mg/ are provided in a phenytoin kg/ day); manuscript table methapyrilene (100 mg/kg/day); furan (40 mg/kg/day) or phenytoin (300 mg/ kg/day) A-277249; 3MC; Sprague–Dawley Affymetrix rat 10 or 100 mg/kg Gene name and (Waring et al., Aroclor toxicology U34 array Affymetrix ID and fold 2002) change in a manuscript table Clofibrate; paracetamol; Male Sprague– Custom chip CLO at 250 mg/kg Gene name and fold (Cunningham, benzoapyrene Dawley BP at 10 mg/kg bw change in a manuscript Liang, Fuhrman, given 3 times per table Seilhamer, & week for 2 weeks, Somogyi, 2000) APAP at 1000 mg/kg L-742694 Female 25K rat microarray L-742694 (50 mg/kg/ Tables of accession (Hartley et al., Dexamethasone Sprague–Dawley day); DEX (50 mg/ number and gene names 2004) kg/day) only in a manuscript table for liver and intestine Bromobenzene Male Wistar Custom 3000 rat gene 0.5, 2, 5 mmol/kg Gene name, accession (Heijne et al., array number and log fold 2004) change for 3 dose levels in a manuscript table. Supplemental data also available 1,25-Dihydroxyvitamin Male Sprague– Affymetrix high-density 730 ng/kg Gene name, Genbank (Kutuzova & D3 Dawley (small rat oligonucleotide accession number and Deluca, 2004) intestine) arrays (GeneChips fold change in a RG-U34A) manuscript table Fenofibrate; clofibrate; Male CD Agilent arrays Dose ranging for up Gene name and genbank (Cornwell, Souza, bezafibrate; to 14 days accession number in a & Ulrich, 2004) gemfibrozi; manuscript table ciprofibrate; beclofibrate; etofibrate Furan Male Sprague– NIEHS rat chip ¨7000 Exposed to 4 or 40 Gene ID and gene (Hamadeh et al., Dawley clones mg/kg furan for up name—binary data in a 2004) to 14 days manuscript table Dataset also available on NIEHS website Paracetamol Male F344/N Rat NIEHS tox chip 0, 50, 150, Unigene accession (Heinloth et al., (http://dir.niehs.nih.gov/ 1500 mg/kg number, gene name, fold 2004) microarray/chips.htm) change in a manuscript table Carbon tetrachloride Male Sprague– ADME Rat expression Up to 14 day Accession numbers and (Young et al., Dawley bioarray (Motorola Life treatment binary data in a 2003) Sciences) consists of manuscript table 1040 single-stranded oligonucleotide probes Carbon tetrachloride and Male Sprague– The rat CT arrays 6, 24, 72 h high and Gene names and fold (Kier et al., 2004) chloroform Dawley contain sequences from low doses changes at multiple time almost 700 rat genes points in a manuscript with known or table discovered responsiveness to toxic treatments (continued on next page) S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 41

Table 1 (continued) Compounds Rat strain Microarray type Compound dose Microarray data Reference availability Dimethylarsinic acid Female F344 Rat 10K chip (MWG 100 ppm Genbank accession (Wei, Arnold, (bladder Biotech Inc.) containing number, gene name in a Cano, & Cohen, epithelium) 10,000 genes manuscript table 2005) Dexamethasone; Female Sprague– Rat HepatoChip DEX (50 mg/kg/day); Gene name and Genbank (Meneses-Lorente troleandomycin; Dawley TAO (500 mg/kg/ accession number—data et al., 2003) miconazole; day); MIC (100 mg/ displayed as a heatmap clotrimazole; kg/day); CLOT (100 not readily extracted isoniazid; mg/kg per day); ISN from publication methylclofanapate (100 mg/kg/day); MCP (75 mg/kg/day) TCDD PeCDF PCB126 Female Harlan Affymetrix GeneChip Exposed for 13 weeks Accession number, gene (Vezina, Walker, PCB153 Sprague–Dawley Test3 arrays to toxicologically name and fold change in & Olson, 2004) equivalent doses a manuscript table Paclitaxel Male and female Genocheck 4.8K cDNA 4 mg/kg/day male; Accession numbers and (Lee et al., 2004) Sprague–Dawley 7 mg/kg/day female gene names in a manuscript table Clofibrate; gemfibrozil; Male Sprague– 4.8 K cDNA microarray Treated with each Gene name and fold (Jung et al., 2004) phenytoin Dawley VAF(+) in house compound for 24 h change in a manuscript albino and 2 weeks table Clofibrate; Sprague–Dawley Merck Drug Safety Chip 30 mg/kg/day Gene name, GenBank (Gerhold et al., dexamethasone; 1443 genes (rat, human accession number and 2001) phenobarbital; and mouse) relative fluorescence 3-methylcholanthrene levels in a manuscript table Bemitradine; clofibrate; Male CD; IGS Incyte RatGEM1.0 Low, mid and high A bar chart of 2 genes (Kramer et al., doxylamine; ¨7800 rat cDNAs doses that are affected by 2004) methapyrilene; compounds phenobarbital; tamoxifen; 2-acetylaminofluorene; 4-acetylaminofluorene; isoniazid PhIP Female Sprague– Mouse cDNA 75 mg/kg/day Gene names in a (Shan, Yu, Schut, Dawley microarray containing manuscript table & Snyderwine, 9984 cDNA clones 2004) (National Cancer Institute Ethinylestradiol Male and female Custom chip 3776 genes 0, 0.01, 0.1, and Accession number gene (Kato et al., 2004) Sprague–Dawley 1.0 ppm name and fold change at different exposure levels in a manuscript table Paraquat Male Wistar 1090 genes 7 mg/kg/day Gene name and (Satomi et al., expression ratio in a 2004) manuscript table Hexachlorobenzene Female Brown Affymetrix rat 0, 150, or 450 mg/kg Accession number, gene (Ezendam et al., Norway RGU-34A GeneChip name and fold change 2004) microarray data for multiple organs in a manuscript table N-methyl-N-nitro-V N- Rat pyloric muco- AFFYMETRIX Rat 83 mg/l AFFY ID, gene name, (Yamashita et al., nitrosoguanidine sae; male congenic Genome U34A gene symbol and fold 2004) (MNNG) rat strain that has a change in a manuscript homozygous_LIZ table transgene of BigBlue\ rat N-methyl-N-nitro-V N- Male ACI/NJcI Affymetrix GeneChip 83 mg/l from the age Accession number, gene (Abe et al., 2003) nitrosoguanidine (ACI) Rat genome U34A of 8 weeks through to name, symbol fold change arrays 40 weeks in a manuscript table; also rat vs human stomach cancer comparison Cisplatin Sprague–Dawley Different arrays, tox 0.3–5 mg/kg over a 4 to Unigene, gene ID, gene (Thompson et al., chip, incyte, phase 1, etc. 144 h name NIEHS ID and 2004) data for 5 platforms in a manuscript table 42 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Table 1 (continued) Compounds Rat strain Microarray type Compound dose Microarray data Reference availability Clofibrate Male Sprague– Atlas Rat Toxicology II High (250 mg/kg/day) or Genbank accession (Baker et al., Dawley arrays (Clontech, Palo low (25 mg/kg/day) number and data for 3 2004) Alto, CA, USA) platforms in a containing 465 genes manuscript table Di(2-ethylhexyl) Male Sprague– An in-house cDNA 20 or 2000 mg/kg Gene names, Genbank (Kijima et al., phthalate Dawley—(testes) microarray ID and fold change in a 2004) manuscript table Ciprofibrate Female Fischer An in-house cDNA 50 mg/kg body weight Gene names, symbol, (Yadetie et al., microarray accession number, mean 2003) ratio and SD Methapyrilene Male Sprague– Rat Tox Chip 1.0 10 or 100 mg/kg/day Gene Name, accession (Hamadeh et al., Dawley number and indication of 2002b) up or down regulation; original data available on NIEHS website Ecteinascidin-743 Female Wistar Custom chip cDNA 40 ug/kg Data available on (Donald et al., (ET-743) microarrays containing laboratory website—not 2002) approximately 4700 available at present hybridizable mouse expressed sequence tags derived from IMAGE clones obtained from Research Genetics (Huntsville, AL) or from the MRC Human Gene Mapping Project Clofibrate; Wyeth Male Sprague– NIEHS rat chip v1.0 Clofibrate (250 Gene names and fold (Hamadeh et al., 14,643; Gemfibrozil; Dawley VAF+ mg/kg/day; Wyeth changes in a manuscript 2002a) phenobarbital; 14,643 (250 mg/kg/ table day); Gemfibrozil (100 mg/kg/day); Phenobarbital (120 mg/kg/day) Vinclozin; procymidone Male Sprague– Clontech Atlas Rat 1.2 200 mg/kg Gene name, accession (Rosen, Wilson, Dawley—prostate Toxicology array number, average fold Schmid, & Gray, change 2005) Cisplatin Male Sprague– Rat Tox Microarrays 0.5 or 1 mg/kg/day Gene name, accession (Huang et al., Dawley VAF1 were purchased from number, fold change in 2001) albino (CRL: Phase-1 Molecular kidney CD(SD) BR Toxicology Allyl alcohol; Male Sprague– Affymetrix GeneChip Allyl alcohol (40 mg/kg Heatmap figures and a (Waring et al., miodarone; Aroclor Dawley Test 2 Array day); miodarone (100 mg/ table of Affymetrix 2001) 1254; arsenic; kg/day); Aroclor 1254 names for genes carbamazepine; (400 mg/kg/day); arsenic correlated with clinical carbon tetrachloride; (20 mg/kg/day); chemistry changes in a diethylnitrosamine; carbamazepine (250 mg/ manuscript table dimethylformamide; kg/day); carbon diquat; etoposide; tetrachloride (1000 mg/kg/ indomethacin; day); diethylnitrosamine methapyrilene; (100 mg/kg/day); methotrexate; dimethylformamide (1000 monocrotaline; mg/kg/day); diquat (17.2 3-methylcholanthrene mg/kg/day); etoposide (50 mg/kg/day); indomethacin (20 mg/kg/day); methapyrilene (250 mg/ kg/day); methotrexate (250 mg/kg/day); monocrotaline (50 mg/kg/ day); 3-methylcholanthrene (100 mg/kg/day) (continued on next page) S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 43

Table 1 (continued) Compounds Rat strain Microarray type Compound dose Microarray data Reference availability Microcystin-LR (MLR); Male Wistar Purpose-made rat DNA Various Accession number gene (Bulera et al., phenobarbital (PB); microarray (Affymetrix, name, fold changes 2001) lipopolysaccharide Santa Clara, CA) shown as a heat map (LPS); carbon containing 1,600 rat table in publication— tetrachloride (CT); DNA sequences data extraction would be thioacetamide (THA); laborious and cyproterone acetate (CPA) Acetamidofluorene; Male Sprague– Custom Rat MegaA Acetamidofluorene (200 Gene name, accession (McMillian et al., aniline; bromobenzene; Dawley cDNA chip 3434-gene mg/kg); aniline (200 mg/ number in a manuscript 2004) butyl hydroxytol; kg); bromobenzene (900 table dieldrin; disulfiram; mg/kg); butyl hydroxytol ethinyl estradiol; (1000 mg/kg); dieldrin hexachlorocyclohexane (30 and 45 mg/kg); g; 4-methylthiazole; disulfiram (2000 mg/kg); nimesulide; piperonyl ethinyl estradiol (500 butoxide; precocene I; mg/kg); pulegone; tannic acid; hexachlorocyclohexane trans-anethole gamma (40, 65, 80 mg/ kg); 4-methylthiazole (120 mg/kg); nimesulide (500 mg/kg); piperonyl butoxide (4000 mg/kg); precocene I (500 mg/kg); pulegone (400 mg/kg); tannic acid (3000 mg/ kg); trans-anethole (600 mg/kg) NIEHS website at http://dir.niehs.nih.gov/microarray/datasets/home-pub.htm. EDGE website at http://edge.oncology.wisc.edu/. and cisplatin, (Table 1). In the majority of cases the resulting by providing freely accessible microarray and other toxicity number of differentially expressed genes is a very small related data. Two of these databases being developed in the subset of the starting number on the microarray following public domain are Chemical Effects in Biological Systems clustering or other types of analysis. Upon closer examina- (CEBS) (http://www.niehs.nih.gov/nct/cebs.htm)(Mattes, tion of these publications, the majority of them either Pettit, Sansone, Bushel, & Waters, 2004; Waters et al., provide images of a heat map and/or a table listing a gene 2003) which will accommodate gene expression profiles, name accession number and expression change. Very few of proteomics and metabolomics data and allow complex the published studies (Tables 1–3) provide the original raw queries (Hood, 2003a; Mattes et al., 2004). Similar goals microarray data file at a freely accessible website, hence are being pursued in the development of the ArrayTrack restricting further analysis by scientists using other software. database at the FDA (Tong et al., 2003). The EDGE Although in some cases it is possible to cut and paste the database (http://edge.oncology.wisc.edu/edge.php), an gene expression data tables from the publication pdf files, expanding public effort at The University of Wisconsin, this is not always the case. In the worst case scenario one contains mouse gene expression profiles following treat- would have to manually retype gene lists or extract them ment with different toxic molecules (Hayes et al., 2005; from heatmaps as binary type data. As not all computational Thomas et al., 2001). These separate efforts if widely researchers will have a laboratory available to them to adopted should make published studies describing HT data generate such quantities of toxicogenomics data, the latter more readily accessible, although it might have been more points are important if we are going to continue to see efficient to evolve these into a single global database instead innovation in software development for this data. This will of fragmented repositories. require free unrestricted access to data published. Similarly Proteomics data has also been generated in a limited if we are to discern ‘‘fingerprints’’ for molecules acting with number of toxicology studies (Table 4), once again this has a similar or identical mechanism we will need databases of been produced with different strains of rats and mice, using many diverse chemical structures that have been tested in a different protein chips, 2-D gel electrophoresis (2-DIGE) similar manner. There is therefore considerable interest in and mass spectroscopy methods (e.g. MALDI-MS). The the current databases being developed by the NIEHS, FDA proteomic data is very rarely accessible to the reader for and other groups which should help to improve the situation their own computational analysis. Subsequently there have 44 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Table 2 Literature toxicogenomics data derived from mouse in vivo studies Compounds Mouse strain Microarray type Compound dose Microarray data Reference availability Benzene Male and female P53 Affymetrix and Incyte 300 ppm, 6 h per day, 5 Gene name, accession (Yoon et al., 2003) KO mice and GEM system days a week for 2 weeks number and fold change C57BL/6 data in a manuscript table Benzene Male 129/SvJ Affymetrix MG-U74Av2 100 ppm, 6 h per day, 5 Gene name, accession (Faiola, Fuller, days a week for 2 weeks number and fold change Wong, & Recio, data in a manuscript 2004) table Phenobarbital CAR and wild type NIEHS Mouse tox chip 100 mg/kg for 12 h Accession number, gene (Ueda et al., 2002) 8736 genes name and fold change for wild type and knock out mice in a manuscript table Aroclor; BNF; C57BL/6J Custom array with 1200 Aroclor (200 mg/kg); Gene names and fold (Thomas et al., ciprofloxacin; cobalt cDNAs BNF (5 mg/kg); change as a heat map 2001) chloride; TCDD; IL-6; ciprofloxacin (250 mg/ Data is also available in LPS; PCB-153; kg); cobalt chloride (60 the EDGE database phenobarbital; mg/kg); TCDD (10 ug/ phenylhyrzn; TNFa; kg); IL-6 (25 ug/kg); WY-16,463 LPS (1 mg/kg); PCB-153 (80 mg/kg); phenobarbital (100 mg/ kg/day), 3 days; phenylhyrzn (100 mg/ kg); TNFa (50 ug/kg); WY-16,463 (0.125% w/v) TCPOBOP CD-1 female Custom 9000 cDNA 1–3 h treatment, 3 mg/ Accession numbers and (Locker et al., mouse array kg body wt fold change data in a 2003) manuscript table 3H-1,2-dithiole-3-thione Male wild-type and Affymetrix murine 0.5 mmol/kg Accession number, gene (Kwak et al., (D3T) nrf2-disrupted genome U74Av2 name, fold in a 2003) GeneChip manuscript table MDMA Male albino Swiss– 15 K mouse cDNA clone 47 mg/kg, Gene name and fold (Xie et al., 2004) Webster, (neurons) set change in a manuscript table Phenytoin Female C57BL/6 and Murine genome-U74Av2 300 mg/l Gene name, accession (Trocho et al., LDLRÀ/À number and fold change 2004) in a manuscript table Genistein (1000 Ag/ ICR (testes) Custom cDNA Genistein (1000 Ag/ Accession number gene (Adachi et al., mouse/day) or microarray, containing mouse/day); name and fold change in 2004) diethylstilbestrol 1754 cDNA probes diethylstilbestrol (50 Ag/ a manuscript table—note (DES) (50 Ag/mouse/ mouse/day) very few genes day) Cocaine and Male ICR Mouse DiscoveryArrayi 40 mg kg—1 cocaine Data apparently not (Hayase, buprenorphine type I array containing once a day for 4 days; 40 available Yamamoto, 2688 brain-derived probes mg kg—1 cocaine plus Yamamoto, Muso, (Display Systems Biotech 0.25 mg kg—1 BUP for & Shiota, 2004) Inc., Copenhagen, 4 days Denmark) Methamphetamine Male C57BL/J6 Affymetrix mouse 40 mg/kg Gene accession number, (Thomas, (striatum) genechip, mg-U74A.v2. gene ID and name and Francescutti- oligonucleotides arrays, signal log ratio in a Verbeem, Liu, & 12 488 genes manuscript table Kuhn, 2004) Di(2-ethylhexyl) Male C57BL/6 Murine genome U74Av2 1.0% dietary DEHP for Genbank accession (Wong & Gill phthalate Arrays (MGU74Av2) 13 weeks numbers, gene name; log 2002) ratio data present as a bar chart in the publication; data extraction would take some effort Diethylhexylphthalate Male PPARa null and Custom made containing 1150 mg/kg/day Gene names available on (Hasmall et al., wild type 600 tox genes a bar chart—quantitative 2002) data not easily accessible S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 45

Table 2 (continued) Compounds Mouse strain Microarray type Compound dose Microarray data Reference availability Acetaminophen C57B1/6 3 129/Ola Test-2 Chips (Affymetrix) 300 mg/kg Genbank or SwissProt (Reilly et al., hybrid then individual ID and fold change data 2001) oligonucleotide in a manuscript table microarrays (Mul1K sub A and sub B; Affymetrix) that can detect the expression of 11,000 known genes and expressed sequence tags (ESTs) Cadmium chloride; Male Swiss Webster Custom Chips Various Data in table form, few (Bartosiewicz, benzo(a)pyrene (BaP); containing 148 unique genes affected for BAP Penn, & Buckpitt, and trichloroethylene genes and TCE 2001) (TCE) been only a very small number of studies that have statistically evaluated, as demonstrated in a very large combined both transcriptomic and proteomic methods with number of published examples (Dobrin, Beg, Karabasi, & a single animal strain after treatment with a drug. Hopefully Oltvai, 2004; Fiehn, 2001; Han et al., 2004; Hanisch et al., we will see this change in the future, but this will in turn 2002; Ideker et al., 2002; Jeong, Mason, Barabasi, & Oltvai, present considerable challenges as huge amounts of 2001; Jeong, Tombor, Albert, Oltvai, & Barabasi, 2000; Li proteomic data are combined with the equally large et al., 2004; Milo et al., 2002; Nikitin, Egorov, Daraselia, & transcript data files. Mazo, 2003; Pereira-Leal, Enright, & Ouzounis, 2004; Rives & Galitski, 2003; Segal et al., 2003a; Somogyi et al., 2001; Spirin & Mirny, 2003; Tornow & Mewes, 2003; 3. Network analysis and databases Vasquez, Flammini, Maritan, & Vespignani, 2003; Yeger- Lotem & Margalit, 2003; Yu, Zhu, Greenbaum, Karro, & From some of the early reviews of systems biology there Gerstein, 2004). has been discussion of its application to drug discovery (Kitano, 2002a,b) as well as the utility for ADME/Tox 3.1. Network applications (Ekins et al., 2002a, 2000b). More recently several other journals have dedicated whole issues to the field of systems For example, one group has used as an inference the biology. However one could consider this quite a broad field Bayesian network method for analysis of tissue toxicity from from network or pathway analysis to quantitative simulation microarray data as well as a mechanistic simulation for a of organelles (Vo, Greenberg, & Palsson, 2004), whole cells different pharmaceutically relevant molecule (Aksenov et and organs. It is apparent that we are now understanding al., 2005). Pathway tools and various resources have also organisms from the perspective of computationally gener- been applied to modeling the networks of nuclear hormone ated networks of protein and ligand interactions (Barabasi & receptors and their connections with other genes and small Oltvai, 2004). Network and pathway tools enable the molecules using a manually curated database, MetaDrug analysis of HT data in the context of all known interactions (Ekins, Kirillov, Rakmatulin, & Nikolskaya, 2005d)or when using a database as the source. Individual reviews MetaCore (Ekins, Bugrim, Nikolsky, & Nikolskaya, 2005). have in some cases indicated that networks will be valuable Transcriptional regulation of many transporters, CYPs and for understanding adverse events (Hood & Perlmutter, phase II enzymes are regulated by these receptors, affecting 2004), drug target identification or validation (Butcher, endogenous molecule transport, metabolism, cell growth, Berg, & Kunkel, 2004) and complex metabolic interactions proliferation and oxidative stress (Ulrich, 2003; Ulrich et al., (Nicholson et al., 2004). A general schematic has been 2004). When the signaling networks and interacting ligands generated in order to provide a description of the utilization for the transcriptional factors PPAR, FXR/RXRA, ESR1, of such pathway databases and network building algorithms AHR, HNF4A, GCR-h, MCR, CAR-beta, GCR-a, LXR-a, from the initial parsing of high throughput data to network CAR/RXR, HNF4, FXR, PXR/RXR heterodimer, PXR, comparisons and visualization (Fig. 3). High throughput AHR/ARNT heterodimer, PPARa/LXRa, VDR, PPAR-a data can be superimposed and visualized on the various are visualized a very complex picture of interactions can be protein interaction databases available. This is accomplished created (Ekins et al., 2005d). This suggests that when we by using either preset maps that capture current biological consider a molecule binding with only one nuclear receptor knowledge or by building custom interaction networks we are observing only a fraction of the likely possible using many different algorithms which can be compared and feasible interactions, based on the data gathered to date. 46 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Table 3 Literature toxicogenomics data derived from in vitro cell studies Compounds Cell type Microarray type Compound dose Microarray data Reference availability 4-Hydroxytamoxifen MCF-7 breast cancer NIEHS ToxChip 1 uM hydroxytamoxifen Data available at NIEHS (Hodges et al., estrogen microarray consisting of for a year, 10 nM website 2003) 1901 genes 17b-estradiol Trovafloxacin Human hepatocytes Affymetrix U133A array 30–800 uM 142 genes available in (Liguori et al., supplemental table—not 2005) easily extracted Estrogen MCF-7 breast cancer NIEHS ToxChip 10–10 M 17b-estradiol Data available at NIEHS (Lobenhofer et al., microarray consisting of website 2002) 1901 genes Valproic acid NMRI mice embryo Custom chip including 600 mg/kg body weight Gene symbol, gene (Kultima et al., and P19 mouse 15K mouse cDNA clone name, NIA EST log fold 2004) embryocarcinoma set change in a manuscript table Sulindac sulfide Human colorectal NIEHS human 12K chip 10 uM Genbank accession (Bottone, carcinoma SW-480 number, gene name and Martinez, Collins, and HCT-116 fold change at various Afshari, & Eling, time points in a 2003) manuscript table; data also available on NIEHS website 17Beta-estradiol; estriol; MCF-7 U95A oligonucleotide 10 nM (E2, estriol; Unigene name, gene (Terasaka et al., estrone; genistein; probe arrays (Affymetrix estrone; DES) 10 AM name and fold change 2004) diethylstilbestrol; (genistein, bisphenol A, for estrogen responsive bisphenol A; nonylphenol, and and nonresponsive in a nonylphenol; methoxychlor) manuscript table methoxychlor Ouabain; lauryl sulfate; HepG2 Clontech Atlas Human Ouabain (43 uM); lauryl Gene name, ratio, p- (Morgan et al., dimethylsulfoxide; Stress Toxicology cDNA sulfate (260 uM); value in downloadable 2002) cycloheximide; arrays (234 genes) dimethylsulfoxide (1.28 tables at journal website tolbutamide; sodium M); cycloheximide (62.5 fluoride; diethyl uM); tolbutamide (12.8 maleate; buthionine; mM); sodium fluoride (3 sulfoxamine; mM); diethyl maleate potassium bromate; (1.25 mM); buthionine; sodium selenite; sulfoxamine (30 mM); alloxan; adriamycin; potassium bromate (2.5 hydrogen peroxide mM); sodium selenite (30 uM); alloxan (130 mM); adriamycin (40 uM); hydrogen peroxide (4 mM) Aflatoxin B(1) (AFB(1)), HepG2 and primary Gene filter arrays 1.0 uM aflatoxin B1, None (Harris et al., 2-acetylaminofluorene hepatocytes containing 31,000 genes 4.0 mM acetaminophen, 2004) (2AAF), 100 uM dimethylnitrosamine dimethylnitrosamine, (DMN), 1.0 uM acetaminophen (APAP) 2-acetylaminofluorene Mitomycin C (MMC) L5178Y TK(+/-) Affymetrix mouse Low, mid and high doses Accession number, gene (Hu et al., 2004) and cisplatin (CIS), mouse lymphoma MG-U74A for MMC name, gene ID, statistical and an alkylating and MG-U745Av2 significance at each time agent, methyl (Affymetrix Inc., Santa point in a manuscript methanesulfonate Clara, CA) for all the table (MMS); other chemicals; a total indirect-acting of 9977 probe sets genotoxins included (genes or ESTs) common hydroxyurea (HU), a to these two array ribonucleotide models reductase inhibitor, taxol (TXL), a microtubule inhibitor, and etoposide (ETOP), S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 47

Table 3 (continued) Compounds Cell type Microarray type Compound dose Microarray data Reference availability Hydroxyurea L5178Y Tk_/-mouse The Twin - Chip Mouse- 10 ng/ml paclitaxel, 31.3 Gene symbol and fold (Lee et al., 2003) (a carcinogen), lymphoma 7.4K Digital Genomics ug/ml hydroxyurea, 32 change in a manuscript p-anisidine cDNA microarray ug/ml p-anisidine table (a noncarcinogen), and paclitaxel Acetaminophen; Wistar Rat DualChip rat hepato A single concentration Accession numbers and (de Longueville amiodarone; clofibrate; hepatocytes (Eppendorf, Hamburg, which varied for each gene name—fold et al., 2003) erythromycin estolate; Germany) compound changes shown as isoniazid; alpha- colored heat map—note naphtylylisothiocyanate; easily extracted from beta-naphtoflavone; publication 4-pentenoic acid; phenobarbital; tetracycline; and zileuton Bleomycin and hydrogen Mouse lymphoma Clontech Mouse 1.2K Bleomycin (2.5 and 20 Gene names and fold (Seidel, Kan, peroxide L5178Y / TK(+/ -) cDNA microarray (1185 ug/ml); hydrogen peroxide change presented as bar Stott, Schisler, & genes) (5and10ug/ml) charts in publication Gollapudi, 2003) Bupivicaine; HL-60 Agilent human cDNA 1 mM Gene name, GenBank (Unami, camptothecin microarray accession number, Shinohara, unigene and ratio in a Ichikawa, & Baba, manuscript table 2003) Bisphenol A Mouse Sertoli TTE3 IntelliGene mouse 0–400 uM Gene name, GenBank (Tabuchi & expression glass accession number and Kondo, 2003) microarrays (Version fold change at time 1.0, Takara Shuzo), points in a manuscript which were spotted with table 564 cDNA fragments of mouse known genes and approximately 301 expressed sequence tags (ESTs) Mitomycin C or Hep G2 85 human gene custom 10 um mitomycin C, Bar charts with fold (Hong, Muller, & doxorubicin array 2 um doxorubicin or changes in publication; Lai, 2003) 2% ethanol very few genes Amphotericin B Human peripheral GF211 FKnown Genes_ 5 ug/ml Accession numbers and (Cleary, Rogers, & blood mononuclear Genefilter cDNA array fold expression in a Chapman, 2001; and THP-1 (ResGen); this array manuscript table Rogers, Pearson, consists of >4000 Cleary, Sullivan, individual elements, & Chapman, each representing a 2002) known human gene Benzo(a)pyrene; diol TK6 human Human-350 microarray, 0, 0.01, 0.10 or Gene names and fold (Akerman et al., epoxide lymphoblastoid a glass slide with 350 1.0 ug/ml) change at doses in a 2004) spotted human cDNA manuscript table probes (Phase-1 Molecular Toxicology Etomoxir HepG2 Clontech Atlasi 1 mM etomoxir Gene names and fold (Merrill et al., Human Stress change in a manuscript 2002) Toxicology cDNA arrays table (234 genes) Tetrodotoxin Human glioma cell Using Affymetrix 10 and 20 mM Affymetrix ID, Genbank (Raghavendra line HTB-138 GeneChip (HG-U133A ID, gene name, gene Prasad, Qi, symbol and fold change Srinivasan, & in a manuscript table Gopalakrishnakone, 2004) Methotrexate; Human acute Affymetrix U133A chip Low and high dose and Data available as (Cheok et al., 2003) mercaptopurine lymphoblastic combination supplemental data online leukemia Prednisolone; Human acute Affymetrix U133A chip Various Data available as (Holleman et al., vincristine; lymphoblastic supplemental data online 2004) asparaginase; leukemia daunorubicin 48 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

A second study has indicated how a natural language p =2.838eÀ 31). This network also maps the Gene Ontology processing method, CCNet was used to show the genes processes for the activation of MAPK (11.8% of genes; p regulated by the nuclear hormone receptor FXR (Apic, value 9.143eÀ 07); signal transduction (33.3%; 1.600eÀ 05); Ignjatovic, Boyer, & Russell, 2005). These automated regulation of transcription, DNA-dependent (29.4%; methods enable a more complete understanding of the 2.786eÀ 04); regulation of inflammatory response (3.9%; complexity of the transcriptional factors (Ekins, Mirny, & 3.746eÀ 04), and the regulation of blood pressure (7.8%; Schueltz, 2002b; Plant, 2004; Ulrich, 2003) but ultimately 4.230eÀ 04). This example network indicates how molecules rely on the quality of the content of the underlying database of the same or different therapeutic classes could be of literature interactions. This is a key consideration that is evaluated for their effects as a graph either together, as in often overlooked. For example the gold standard database is this case, or individually. This would be useful to indicate one that is manually curated to ensure the fidelity of the potential off target effects and identify structurally dissimilar direct interaction and is preferable to one generated molecules with similar network patterns. Such networks computationally by algorithms like natural language pro- could then be compared to assess network overlap or cessing (Nikolsky, Nikolskaya, & Bugrim, 2005). The differences between molecules and their inhibition of advantage of interaction networks over clustering has been multiple proteins. This type of unique visualization of high demonstrated in one study using MetaCore (Nikolsky, throughput screening data illustrates how the target proteins Ekins, Nikolskaya, & Bugrim, 2005) by reanalysis of a may be connected as a network to infer the possible published microarray study of G0-arrested MCF-7 breast downstream effects of inhibition. cancer cells treated with estrogen and 4-hydroxytamoxifen With the burgeoning number of freely available online (Hodges et al., 2003). After producing integrated gene and commercial databases that can be used for pathway networks for each treatment, strikingly different patterns construction numbering in the hundreds, there have been were displayed although both contained early transcriptional suggestions to impose standards for model exchange, factors myc, jun and fos. Only the estrogen network featured querying and visualization (Cary, Bader, & Sander, 2005). induced genes essential for all cell cycle phases (Nikolsky et To date there has been little discussion with regards to al., 2005). Similarly microarray data for benzene toxicity standardization of ADME/Tox related databases, although (Yoon et al., 2003) has been re-analyzed by focusing on the there has been considerable discussion relating to drug genes assessed on the p53 pathway (Ekins et al., 2005e). We metabolism database generation (Erhardt, 2003). This is envisage that a database of such networks for toxic certainly an important area to address in the future. There is compounds will be used for comparing between different already a growing literature related to ADME/Tox that is molecules and used in the development of predictive partially captured in the several commercially available algorithms for Systems-ADME/Tox modeling in future. databases (Ekins et al., 2005e), but to date there have been Another approach to using such pathway approaches is to limited academic efforts to capture data for transporters visualize the results of quantitative structure activity models with the human membrane transporter database (Yan & for predicting molecules binding to enzymes, transporters, Sadee, 2000) TP-search transporter database (http://www. receptors and ion channels (Ekins, Andreyev, et al., in press, ilab.rise.waseda.ac.jp/tp-search/), drug interaction database 2005e). (http://www.druginteractioninfo.org/Databaseinfo.aspx), It is also possible to simultaneously interpret high nuclear hormone receptors (Nakata, Yukawa, Komiyama, throughput data and predictions on interaction networks, Nakano, & Kaminuma, 2002), the ADME-AP database providing a novel approach to predicting and understanding (Sun, Ji, Chen, Wang, & Chen, 2002) and PharmaGKB potential undesirable drug–drug or off target effects in the (Oliver et al., 2002), DSStox (http://www.epa.gov/nheert/ area of systems pharmacology. An example data set uses dsstox/), TOXNET (http://toxnet.nim.nih.gov/)thatare percent inhibition data for clotrimazole and ticonazole readily accessible. which were screened against many different assays at a In order to generate accessible pathways using any of single concentration in a commercially available database, the available software, a large enough set of object BioPrint (Cerep, Redmond, WA) as published recently identifiers are required to map onto the underlying data- (Fliri, Loging, Thadejo, & VOlkman, 2005). The data for base. To demonstrate this, datasets from toxicogenomics 10 assays has been arbitrarily encoded as inhibitors (>50% studies have been evaluated with both the KEGG pathway inhibition) or non-inhibitors (<50% inhibition) in a text file database and a commercially available product MetaCore which was then loaded into MetaCore. The analyzed (www.genego.com). These gene or protein lists range in network algorithm was then used which generates a large size from 21 to 1853 objects. In virtually all cases more network and fragments it into sub-networks each with a Z- identifiers are mapped to networks in MetaCore and this score and p-values for ranking according to saturation with also seems independent of the identifier type used objects from the initial gene list. The Gene-ontology (Unigene, Affymetrix, Genbank or Locuslink, Table 5). processes are also mapped to the gene list and individual These numbers will obviously change as more information networks. In this example a statistically significant network is added to the respective databases, hence, the more was generated for the different proteins (Fig. 4, objects that are mapped from a dataset. The more extensive S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 49

Table 4 Literature derived toxicoproteomics data Compounds Data source Microarray type Compound dose Microarray data Reference availability Carbon Male Wistar rats Affymetrix rat chip 8799 6–24 h exposure Genbank, SwissProt and (Fountoulakis, tetrachloride probes+proteomics binary data in a 2004) study manuscript table Paracetamol CD-1 male mice Custom mouse tox blots 150 or 500 mg/kg Gene name, IMAGE ID, (Ruepp, Tonge, with 450 genes, GenBank accession Shaw, Wallis, & RTQ-PCR+proteomics number, fold induction. Pognan, 2002) study 2-DIGE+ SwissProt identifier and MALDI-MS protein abundance change in manuscript tables Paracetamol AP-1 male mice Affymetrix murine 11K Up to 500 mg/kg Gene names and fold (Coen et al., 2004) set+proteomics study changes at multiple doses and time points in a manuscript table— proteomics data not accessible Oxazepam or Male B6C3F1 mice NIEHS Mouse Chip Oxazepam (2500 p.p.m.) Proteomics data in a (Iida et al., 2003) Wy-14,643 (8700 genes); 2-DIGE Wyeth (Wy)-14,643 manuscript table and MS (500 p.p.m.) Genbank accession number, gene name and fold change in a manuscript table; data also available to download from NIEHS website Compound A Female rats Crl:CD Proteomics, 2-DIGE and 250 mg/kg/day up to Accession number and (Meneses-Lorente (PPAR gamma (SD)IGS BR; MS 5 days protein name and et al., 2004) ligand?) average ratio in a manuscript table Bromobenzene Male Wistar rats Proteomics, 2-DIGE and 5 mmol/kg Accession number, gene (Heijne, Stierum, MS, custom 3000 cDNA name, fold change in a Slijper, van rat chip manuscript table; protein Bladeren, & van names along with bar Ommen, 2003) charts the network that can be generated (as it will consist of more oxidative stress (Ekins, Giroux, Nikolsky, Bugrim, & nodes), then a more comprehensive understanding of the Nikolskaya, 2005c). networks is possible. The data available currently in the Other important pathway/network building tools that could literature can be used to evaluate such pathway and potentially be applied to toxicogenomics data include: network generation tools. Recently we have used several Ingenuity pathways analysis (http://www.ingenuity.com), of the published studies (Tables 1–3) with MetaCore to PathArt (http://www.jubilantbiosys.com.pd.htm), Pathway visualize networks for acetaminophen, furan, carbon tetra- Assist (http://www.ariadnegenomics.com/procts/pathway. chloride, benzene and cisplatin showing genes involved in html)(Nikitin et al., 2003), and several other databases deposited at the Pathway Resource List (http://cbiio.mskcc. org/prl). All of these products have unique underlying proprietary pathway databases which are compiled manually , , , , 1970 s 1980 s 1990 s 2000 s or automatically with text mining tools or a combination of both. We are still waiting for studies that provide a comparison In vivo of different database tools or network building algorithms. In vitro Until then there may be some overlap but also some differences between the results obtained from more than one network OMICS method due to the database content and the algorithms used. The reader is recommended to evaluate for themselves several In Silico technologies and select those with the most appropriate fit to Systems Biology (?) their own specialized needs. In the next Sections the further application of some Fig. 2. The timeline for major paradigms in ADME/Tox. available network and database tools will be described with 50 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Microarray or other high throughput data

Upload list of gene/protein identifiers and fold change, significance etc (raw data or after selection of the differentially expressed Genes using other statistical methods)

Parse database and generate interaction network with different algorithms Export gene list or visualize on maps

Compare 2 or more networks, Access significance of the interaction networks Export gene list intersection, overlap etc

Determine GO processes generate interaction network with different algorithms or visualize on maps Filter Networks

Visualization

Fig. 3. Schematic for the utilization of pathway tools for assessing high throughput data. specific detailed reference to transporters and enzymes, but 3.1.1. The role of transporters these technologies can also be applied elsewhere as A diverse array of organic solutes such as nutrients, described above. neurotransmitters and drugs are transported by specialized

Fig. 4. Network for high throughput screening data for clotrimazole and ticonazole screened against 10 in vitro assays (Human Cannabinoid 1, Human cholecystokinin, CCR1, Choline transporter, Chloride channel, catechol-O-methyl transferase, COX2, CYP1A2, CYP2B6, CYP2C19), data published by Fliri and co workers (Fliri et al., 2005). Nodes highlighted with large blue circles represent assays used in the study. Red circles on these nodes represent both molecules showing >50% inhibition in the assay, blue circles represents both molecules showing <50% inhibition, chequered circles represent one molecule inhibits >50% and one inhibits <50%. Ligands (large) linked to transcriptional factors, enzymes and transporters via edges using the MetaCorei database (ww.genego.com). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 51

Table 5 Datasets acquired from publications and other sources and used to evaluate two database/pathway tools, KEGG and metaCore Dataset Number of objects/ Number of objects Number of objects mapped Reference for microarray genes /proteins in list mapped with KEGG with MetaCore on networks data gene list Tox Chip 1.0 (NIEHS)# 1853 405 1498 https://dir-apps.niehs.nih.gov/ maps/guest/clonesrch.cfm 4-Hydroxytamoxifen and 1617 434 1343 (Hodges et al., 2003) estrogen Mitochondrial proteins 722 156 388 (Gaucher et al., 2004; Taylor et al., 2003) Bromobenzene 130 41 89 (Heijne et al., 2004) (24 and 48 h) Acetaminophen# 30 19 23 (Heinloth et al., 2004) Acetaminophen 84 29 64 (Huang et al., 2004) Furan 185 64 139 (Huang et al., 2004) Tetrodotoxin* 116 31 86 (Raghavendra Prasad et al., 2004) Benzene 73 16 62 (Yoon et al., 2003) Benzene 76 5 53 (Faiola et al., 2004) Carbon tetrachloride 37 8 26 (Young et al., 2003) Estrogen 94 33 90 (Lobenhofer et al., 2002) Trovafloxacin* 142 20 82 (Liguori et al., 2005) Phenobarbital 37 13 28 (Ueda et al., 2002) L-742694 (liver)+ 45 17 19 (Hartley et al., 2004) L-742694 (intestine)+ 23 11 10 (Hartley et al., 2004) A-277249* 21 7 9 (Waring et al., 2002) All files were converted to Locuslink identifiers except as noted (#) unigene, (*) Affymetrix, (+) Genbank identifiers were used. proteins across cellular membranes. These may function as somes (Weinshilboum & Wang, 2004). A number of passive processes or active processes energized by the structurally diverse molecules bind to P-gp which is hydrolysis of ATP or coupling to the co-transport of counter expressed in many tissues and has numerous SNPs, one of ions down an electrochemical gradient such as Na+,H+ and which (C3435T) affects the expression level in the ClÀ. There are many thousands of transporters which can be duodenum and therefore can impact the absorption of classified into distinct superfamilies. One of these, the solute molecules which would be substrates for this transporter carrier class (SLC) is rapidly expanding and contains over (Sakaeda, Nakamura, & Okumura, 2002). The human 30 families and 200 members. The ATP-binding cassette proton-dependent dipeptide transporter (hPEPT1) can also (ABC) contains 7 families and over 48 members including affect the absorption of molecules in the intestine and P-glycoprotein (P-gp) and MRP subfamilies (Zhang, Knipp, recently 9 SNPs were found with only one displaying a Ekins, & Swaan, 2002a; Zhang, Phelps, Cheng, Ekins, & reduced transport capacity (Zhang et al., 2004). The sodium- Swaan, 2002b). Transporters have a key role in clinical dependent carnitine cotransporter OCTN2 can possess pharmacology with many drugs specifically targeting them. mutations and these result in primary carnitine deficiency Numerous drugs share transport pathways with nutrients which impacts fatty acid oxidation and is characterized by and transporters have a role in oral absorption, drug many clinical manifestations (Lahjouji, Mitchel, & Qureshi, bioavailability, drug resistance, excretion, and ultimately 2001). The organic cation transporter 1 (OCT1) is also pharmacokinetics and pharmacodynamics. important in the transport of numerous xenobiotics and Polymorphism of drug transporters may be a key factor endobiotics. Recently 4 SNPs were identified in the in drug interactions and lack of effectiveness. This field has Japanese population and when functionally characterized become known as pharmacogenomics and is focused on in vitro the uptake of cations was reduced significantly for understanding of inherited DNA sequence variations (poly- some of these mutations, indicating that this would likely morphisms and mutations) revealed by xenobiotics (Evans contribute to inter-individual variations in metabolism of & McLeod, 2003; Weinshilboum, 2003). Over the last few drugs which were transported via OCT1 (Sakata et al., decades many genes have been directly linked to the 2004). mechanisms of response (Evans & McLeod, 2003; Wein- shilboum, 2003) such that 20–95% of variability to drug 3.1.1.1. Clinical relevance of transporters. The pregnane response is inheritable (Evans & McLeod 2003). This X-receptor (PXR) is a transcriptional regulator of the phenotypical variability is mainly caused by single nucleo- enzyme human MDR1 (P-gp), MRPs and OATP (Wang & tide polymorphisms (SNPs) present in anywhere from 1% to LeCluyse, 2003) CYP3A (Bertilsson et al., 1998; Blumberg 50% of the population resulting in either lower protein et al., 1998; Kliewer et al., 1998) and CYP2C8/9 (Synold, activity, incorrect folding or rapid degradation via proteo- Dussault, & Forman, 2001) as well as many other genes 52 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 involved in the transport, metabolism and biosynthesis of stronger soon after exposure, before declining (Fleck et bile acids (Staudinger, Liu, Madan, Habeebu, & Klaassen, al., 2003). Some transporters may therefore be specifically 2001). However the additional receptors, CAR, FXR, LXR targeted by drugs in one tissue such as the CNS but these and other nuclear receptors take part in a complex network same transporters may also be expressed elsewhere in the of interactions to control these and other proteins. Thus, body, hence off-target effects may result in toxicity. The elucidation of the regulatory networks, which control the serotonin transporter is one such example which is expression of these transporters, is also important. To date expressed in the lungs and brain. Some substrates for this most of the research has centered on efflux transporters but channel like fenfluramine can result in primary pulmonary there has been considerable interest in uptake transporters hypertension as they accumulate in pulmonary cells (Roth- such as the organic anion transporter polypeptide (OATP, man, Ayestas, Dersch, & Baumann, 1999). Similarly P-gp is ((Kim, 2003), see also later Section on OATP). expressed at the blood brain barrier and intestine, impacting There are several specific examples of the importance of the efficacy and bioavailability of drugs. drug transporters to the clinical development of drugs. One example is the insulin sensitizer troglitazone which was 3.1.1.2. Transporter network examples: ABCA1. The withdrawn due to hepatotoxicity although the precise ABCA1 transporter mediates the first step of cholesterol mechanism appears to have been unclear until recently. transport. Mutations in this gene cause Tangier disease, The major metabolite is a sulfated species and is suspected which results in severe HDL deficiency, cholesterol of being responsible for the observed toxicity. The recent accumulation in macrophages and attendant atherosclerosis. assessment of the organic anion transporting polypeptides This transporter represents a drug target for upregulation, OATP-C and OATP8 expressed on the hepatocyte baso- modulating cholesterol metabolism and prevention of lateral membrane indicated that sulphated troglitazone has a cardiovascular disease (Oram & Lawn, 2001). In vitro, high affinity for the former and possibly lower affinity for ABCA1 can be inhibited by the sulfonylurea glybenclamide the latter (Nozawa et al., 2004). This metabolite would (Field, Burn, & Mathur, 2004; Wang, Silver, Thiele, & Tall, therefore be expected to accumulate in hepatocytes and 2001). In order to illustrate the advantages of mapping drug inhibit the bile salt export pump and Mrp2. Because transporters as networks onto functional models alongside polymorphisms have been shown for OATP-C (Tirona, other proteins, one can consider the example of ABCA1. Leake, Merino, & Kim, 2001) it is also possible that these The query of the MetaCorei database shows that this may result in the accumulation of the metabolite and in turn transporter appears on three manually curated pathway maps elicit idiosyncratic toxicity. A second relevant example of representing the Fstate of the art_ knowledge derived from the impact of transporters is the clinically significant drug– reliable, high quality literature sources. One can also use the drug interaction between cerivastatin and cyclosporine A individual maps as an interface to access the underlying which occurs via the OATP-C transporter (Shitara, Itoh, layers of information about the transporter, including the list Sato, Li, & Sugiyama, 2003). A third example are the HIV of encoding genes/splice variants with known SNPs. In protease inhibitors saquinavir, ritonavir and indinavir which addition to browsing MetaCorei, a user can also build are transported by MRP2 in vitro and other drugs such as custom networks around ABCA1 using the network- probenecid and sulfinpyrazone are able to enhance this construction tool (Fig. 5). Such a visualization utility may transport. Transport by MRP2 suggests that these com- be very helpful for identification of all putative pathways pounds will have decreased bioavailability due to increased around a particular transporter or compound of interest. The clearance and other drugs could aggravate this situation by ABCA1 network created by this tool shows that this further enhancing transport (Huisman et al., 2002). Sim- transporter at the time of writing is linked directly to ilarly the rifampicin mediated induction via PXR of MRP2 twenty-five other objects such as APOE1 and LXR. Many and P-gp in healthy subjects was found to significantly of its neighbors have their own SNPs that could be decrease the AUC and also correlated with intestinal important in determining interactions between transport of expression of these transporters. This transporter is also a drug and normal human transport of endogenous ligands inducible by cisplatin, 2-AAF, and phenobarbital (Schrenk in health or disease. et al., 2001), indicating multiple mechanisms may be involved. In other species such as rat commonly used as a 3.1.1.3. Transporter network examples: OATP. The toxicity model, orthologs of the transporters such as OATP2 OATPs are key membrane bound transporters expressed in are expressed and can be induced with ligands for PXR like many organs including intestine, liver, lung, choroid plexus, PCN (Guo, Choudhuri, & Klaassen, 2002). This is useful blood brain barrier and other organs (Tamai et al., 2000). knowledge because the advent of microarray technology This family of transporters is capable of mediating the allows one to dose a rat with a xenobiotic and assess sodium-independent transport of a diverse array of mole- thousands of genes simultaneously in a particular tissue. For cules such as steroid conjugates, organic anions and instance animals dosed with known nephrotoxins have xenobiotics by coupling uptake with efflux of bicarbonate shown some upregulation of the Na–K–Cl transporter, (Satlin, Amin, & Wolkoff, 1997), glutathione or its however the authors suggested genomic responses are conjugates (Hagenbuch & Meier, 2004). The inhibition of .Eis/Junlo hraooia n oiooia ehd 3(06 38–66 (2006) 53 Methods Toxicological and Pharmacological of Journal / Ekins S.

Fig. 5. The network of direct gene interactions around the ABCA1 transporter gene (centre) generated with MetaCorei. The interation types between nodes are shown as small colored hexagons e.g. unspecified, allosteric regulation, binding, cleavage, competition, covalent modification, dephosphorylation, phosphorylation, transcription regulation, transformation. When applicable, interactions also have a positive or negative effect and direction. Ligands (purple) linked to other proteins (blue blobs), transfactors (red), enzymes (orange). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 53 54 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 this transporter’s hepatic uptake of other compounds may be cholestatic hepatitis (Dumont et al., 1997; Meier & Stieger important for reported drug–drug interactions (Kim, 2003) 2000; Rost et al., 2003) and primary sclerosing cholangitis described earlier (Shitara et al., 2003) as well as cerivastatin (Oswald, Kullack-Ublick, Paumgarter, & Beuer, 2001). In with gemfibrozil (Shitara, Hirano, Sato, & Sugiyama, 2004). particular, OATP1B1 appears to be regulated by the liver- OATP1B1 (previous names OATP-C, LST-1, OATP2, enriched transcription factor hepatocyte nuclear factor 1a SLC21A6) represents the most studied human OATP to (HNF1a) which binds to the promoter region of this date (Meier & Stieger, 2000) and is expressed on the transporter (Jung et al., 2001). Site directed mutagenesis basolateral plasma membrane of hepatocytes. Several single of this binding site resulted in inactivation suggesting the nucleotide polymorphisms have been identified in the critical nature of the interaction with HNF1a. Bile acids OATP1B1 gene in European-Americans, African-Americans such as CDCA have been shown to transcriptionally repress (Tirona et al., 2001) and Japanese (Nozawa et al., 2002), HNF1a in vitro via inhibition of the transactivating effect of dramatically impacting the transport of ligands such as HNF4a on HNF1a (Jung & Kullak-Ublick 2003). After pravastatin (Mwinyi, Johne, Bauer, Roots, & Gerloff, 2004; screening many rat and human uptake transporters in vitro, Nishizato et al., 2003), estrone-3-sulfate (Nozawa et al., OATP1B1 was also shown to modulate the PXR response 2002; Tirona et al., 2001), Rifampin (Tirona, Leake, by controlling rifampin retention in the cell and therefore Wolkoff, & Kim, 2003) and estradiol 17h-d-glucuronide affecting the induction of CYP3A4 and other gene products (Tirona et al., 2001). such as P-gp (Tirona et al., 2003). The regulation of SLCO may be affected during Some of the literature for OATP1B1 human substrate extrahepatic cholestasis, bile duct ligation, bile salt induced data has been annotated into MetaCorei to illustrate the

Fig. 6. Networks generated with the autoexpand algorithm in MetaCorei to illustrate how ligands for the human OATP-C (OATP1B1) transporter (centre) can be interconnected with other protein, regulatory, signaling information. Information on the type of interaction between objects is hidden for clarity e.g. unspecified, allosteric regulation, binding, cleavage, competition, covalent modification, dephosphorylation, phosphorylation, transcription regulation, transformation. When applicable, interactions also have a positive or negative effect and direction. Ligands (purple) linked to other proteins (blue blobs), transfactors (red), enzymes (orange). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 55 visualization of the complex interconnections between this known inhibitors and substrates only a small number are transporter, its ligands, regulatory factors and signaling shown here for clarity. We assume that if a perturbation in molecules already in the database (Fig. 6). The network was a pathway (e.g. due to a nonfunctional enzyme) is linked generated with the autoexpand algorithm in the software, to a certain pathologic condition, a similar perturbation representing one of multiple available algorithms for caused by the interference from xenobiotic metabolism connecting genes, ligands and other objects in the database. (e.g. competitive inhibition of the same enzyme) may Clearly if more ligands and their connections are added to result in identical effects. As microarray gene expression the database, the complexity of the network will increase data is increasingly generated, the role of enzyme considerably. The OATP-C (OATP1B1) gene details can be regulation in toxicity of certain xenobiotics will become viewed upon querying the database and links are provided more apparent from either in vivo or in vitro studies. The to other public databases. This page can be used to highlight visualization of such signature gene networks involving the multiple synonyms for this gene as well as links to the transporters and enzymes, their ligands and regulatory multiple SNPs identified to date. factors will also be important for future toxicity prediction methods. We have recently generated visualizations of 3.1.1.4. Transporter microarray data. Microarrays have microarray data from MCF-7 cells treated with 4-hydrox- generally been limited in the number of transporters present ytamoxifen to show that some of the key genes involved in on them (Annereau et al., 2004) however they have been used metabolism and transport are upregulated (Ekins et al., in an attempt to correlate pharmacokinetic properties with 2005d). In addition we have made predictions with various gene expression for valacyclovir (Landowski et al., 2003)as QSAR models in MetaDrug to indicate the involvement of well as understand the expression profile in different tissues PXR, CYP3A4 and P-gp (Ekins et al., 2005e). Therefore it or cell lines upon food component or xenobiotic treatment appears likely that 4-hydroxytamoxifen could increase its (Anderle, Huang, & Sadee, 2004). This lack of transporters own metabolism as well as efflux from cells via P-gp on microarrays has prompted some groups to produce their which can be visualized on networks. Any decrease of own arrays with a heavier emphasis on transporters. These function of these enzymes or transporters in a population arrays have for example then been used to demonstrate the would likely result in changes in the metabolism and upregulation of ABC transporters and down-regulation of transport of this active metabolite, potentially impacting GST-Pi in cell lines resistant to colchicines or 9-nitro- the clinical effect. This represents one example of how camptothecin (Annereau et al., 2004). The genes that were both pathway tools, QSAR models and network building significantly up or down-regulated in this particular study algorithms can be used with different types of predicted were used to build networks with MetaCore (Fig. 7A, B) and and experimental data to allow visualization of potential the similarities between them were assessed (Fig. 7C). compound interactions or toxicity. Although there were only a small number of significantly changed genes in common (IL-8, Fos, GST-Pi, Calpactin and 3.1.3. Future network applications Ubiquitin hydrolase) it is perhaps likely that there is a much As the population ages an increasing incidence and larger common gene network that is important for drug prevalence of systemic diseases, especially chronic diseases resistance, although a much larger number of cell lines and have occurred among older adults. This has resulted in an drug treatments need to be evaluated to produce a definitive increase in medications used concomitantly by this drug resistance signature involving transporters, enzymes population which presents challenges for drug–drug and transcriptional regulators. interactions. Physiologically, elderly patients may behave differently to the younger patients for which the drugs are 3.1.2. Applications to enzymes initially developed for and tested on. Many pharmacoki- As we have already described, it may be particularly netic investigations in the elderly population reveal valuable to visualize enzymes as networks to show decreased clearance of lipophilic drugs metabolized by interactions with transcriptional regulators and ligands. the cytochrome P450 enzymes; however, few studies have For example a key enzyme is CYP3A4 which metabolizes evaluated aging-dependent or gender-related changes in 40–45% of all drugs and has relatively few SNPs specific P450 enzymes (Hunt, Westerkam, & Slave, 1992). (Ingelman-Sundberg, 2004). Using a second software suite Age-related physiological changes, such as a reduction in MetaDrugi (www.genego.com) it is possible to construct liver mass, hepatic metabolizing enzyme activity, liver a custom network around this or other drug metabolizing blood flow and alterations in plasma drug binding may enzymes (Fig. 8). In this case the gene network for account for the decreased elimination of some metabolized CYP3A4 highlights all of the major transcriptional drugs in the elderly. It is particularly difficult to separate an regulators and several more distant linked proteins and effect of aging from the variation in the rate of metabolism ligands connected on the network that may be useful for due to factors such as individual metabolic phenotype further study (Fig. 8). Substrates, inhibitors as well as (slow or fast metabolizer due to SNPs), environmental regulatory factors and other enzymes can be observed influences, concomitant disease states and drug intake connected on this network. Due to the many hundreds of (drug–drug interactions). The available data suggest that 56 .Eis/Junlo hraooia n oiooia ehd 3(06 38–66 (2006) 53 Methods Toxicological and Pharmacological of Journal / Ekins S.

Fig. 7. Networks for drug resistant cell lines generated with the Auto expand algorithm in MetaCore. A. KB-3-1 vs KB-8-5, B. ATC-DU-145 vs RCO.1, C. node overlap of A and B with initial nodes highlighted with blue or red solid circles. (Initial gene lists derived from (Annereau et al., 2004)). Information on the type of interaction between objects is hidden for clarity e.g. unspecified, allosteric regulation, binding, cleavage, competition, covalent modification, dephosphorylation, phosphorylation, transcription regulation, transformation. When applicable, interactions also have a positive or negative effect and direction. Ligands (purple) linked to other proteins (blue blobs), transfactors (red), enzymes (orange). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) .Eis/Junlo hraooia n oiooia ehd 3(06 38–66 (2006) 53 Methods Toxicological and Pharmacological of Journal / Ekins S.

Fig. 7 (continued). 57 58 .Eis/Junlo hraooia n oiooia ehd 3(06 38–66 (2006) 53 Methods Toxicological and Pharmacological of Journal / Ekins S.

Fig. 7 (continued). S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 59

Fig. 8. The network of gene interactions around the CYP3A4 enzyme (centre) from MetaCorei. Information on the type of interaction between objects is hidden for clarity e.g. unspecified, allosteric regulation, binding, cleavage, competition, covalent modification, dephosphorylation, phosphorylation, transcription regulation, transformation. When applicable, interactions also have a positive or negative effect and direction. Ligands (purple) linked to other proteins (blue blobs), transfactors (red), enzymes (orange). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) the initial doses of drugs metabolized by these enzymes of predicted and experimental data to allow visualization of should be reduced in older patients according to the clinical potential compound interactions or toxicity in elderly response. In most published studies the elderly appear at populations. We may see different gene networks high- least as responsive as the young to inducers or inhibitors of lighted as humans age and these may be modified by drug P450s (Durnas, Loi, & Cusack, 1990). More recently there treatment and coadministration. This represents an area has been some suggestion that there are age related were network analysis could be applied in the future and is reductions in function of some specific P450s such as in need of considerable further research. CYP3A4 (Patki et al., 2004) and this could occur at the The collection of microarray data in databases such as level of regulation. However earlier studies with the same CEBS, ArrayTrack and EDGE represents a future resource enzyme showed no change in clearance with age (Hunt et for computational gene network analysis. One could al., 1992). To date there has been even less examination of envisage that ultimately in each case such data is converted the transporter functions and any changes with age to one or more networks that are also displayed for the user (Kinirons & O’Mahony 2004), so the current understand- and can be used to compare treatments from in vivo and in ing of the effects of aging on metabolism and transport is vitro experiments. This would represent a different approach anything but transparent. This represents an extreme to clustering the data as currently implemented in one of challenge for the pharmaceutical industry: how to predict these efforts (Hayes et al., 2005) and may condense large whether a drug has an affinity for an enzyme or transporter amounts of experimentally derived data into a readily and whether this may be clinically important if it shows a interpreted network. decline in expression or function with aging. Any decrease of function of these enzymes or transporters in an elderly population would result in changes in the metabolism and 4. Discussion transport of metabolites, potentially impacting the clinical effect. This represents another example of how computa- Previously in this journal the progress of many tional approaches may perhaps be used with different types research groups in predicting human ADME parameters 60 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 in silico (Ekins et al., 2000b) and approaches for drug ing complex tissue-level processes as networks integrating metabolism (Ekins, Ring, Grace, McRobie-Belle, & all data types based on functional interactions. The data- Wrighton, 2000a) have been described. Both of these bases developed and used in this approach will certainly reviews commented on moving HT assays for ADME/ benefit from further annotation around the drug metaboliz- Tox much earlier in drug discovery, which had also been ing enzymes and transporters as described herein in terms indicated by other groups. The initial reviews also of transcriptional regulation and the ligands associated with highlighted the likely wealth of data that would become them, which frequently appear in the literature. It will be available and how this could be used for structure important to capture disparities in the assignment of ligands activity relationships alongside the bioactivity data in to enzymes, transporters and other proteins as well as computational models. It was noted that there was a negative data. paucity of predictive metabolism tools at that time. In In summary, although there have been numerous addition a growing number of efforts to model whole toxicogenomics studies published, there is presently a cells and organs, now a field called systems biology, relatively small number of datasets that are freely available were recognized as models that could be integrated with to perform network analysis of microarray data. The the in silico ADME approaches. In summary, since these number of studies identifying large numbers of proteins past reviews virtually all pharmaceutical companies have which are affected by molecule treatment are even scarcer attempted earlier high throughput screening for ADME/ still while there are several examples of some published Tox properties and to some extent the wider application studies that combine such data. It is hoped that the of computational approaches for physicochemical proper- numerous database initiatives for high content and toxicol- ties. Systems biology is being quite widely acknowledged ogy data that are being undertaken will improve the as the new paradigm for understanding complex bio- situation for other researchers that are not currently logical datasets derived from high throughput technolo- equipped to do such microarray studies themselves. The gies and the accumulated knowledge on human protein addition of requirements by journals to deposit such raw interactions (Hartwell, Hopfield, Leibler, & Murray, 1999; data in a freely accessible resource will aid these initiatives. Hood, 2003b). Therefore, systems biology can be defined ADME/Tox groups have seen new technologies and as the integration of genetic, proteomic, transcriptomic approaches developed over the last decade that have all and metabonomic data using computational methods been applied to identify poor compounds earlier (Fig. 2). (Nicholson & Wilson, 2003). When taken together, The latest technologies integrate network building tools information on molecular processes derived from different with high content data and databases. The current review sources represents a ‘‘universe’’ of putative biological described the limited number of networks generated for functionality of which only a small fraction of it will be ADME/Tox at present and one hopes that the impact of realized in a cell at any given time. To date systems such analyses will be commonplace in the future. Systems biology has been driven by academia and funding bodies biology is however more than just applying a network such as the NIH rather than the pharmaceutical compa- approach and hence systems-ADME/Tox will have to nies. Presently, there is a great deal of interest from evolve due to the continual pressure to develop newer scientists of all backgrounds in identifying the networks technologies. This current paradigm, combining empirical of cellular pathways and the corresponding physically data and computational methods should integrate the interacting proteins. complex data already generated, making it readily inter- The network building software for systems biology pretable and valuable for identifying the most promising described in this current review will be valuable to query compounds in the future. high throughput data and known literature interactions in order to predict potential toxicity in different species. In the future the compilation of published toxicogenomics data- Acknowledgements sets characteristic of different types of toxicity will likely be available in these software systems to act as a reference Dr. Maggie A.Z. Hupcey is gratefully acknowledged for database. It is also feasible that we will be able to generate editorial assistance. Dr. Peter W. Swaan (University of the annotated datasets which specifically address the Maryland) and Dr. Cheng Chang (Ohio State University), differences between human and rat networks implicated Dr. Steve Wright (University of Arizona), Dr. K. Sandy in toxicity. The identification of sub-network modules Pang (University of Toronto) and Dr. Craig Giroux (Wayne conserved between human and rat, distinct for toxicity State University) are kindly acknowledged for their support types or predictive for toxic end-points in human will be and discussions. My colleagues at GeneGo; Sergey possible. Such signature gene networks (Nikolsky et al., Andreyev, Andy Ryabov, Eugene Kirillov, Eugene A. 2005) can then be verified with other experimentally Rakhmatulin, Svetlana Sorokina, Andrej Bugrim, Tatiana derived data prospectively or from preexisting databases. Nikolskaya, Yuri Nikolsky, John Metz and Julie Bryant are The combined Fhigh throughput data–in silico network_ all thanked for their considerable contributions to software approaches therefore represents a novel method for access- development and data annotation. S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 61

References sulindac sulfide, in human colorectal carcinoma cells: Possible link to apoptosis. Journal of Biological Chemistry, 278, 25790–25801. Abe, M., Yamashita, S., Kuramoto, T., Hirayama, Y., Tsukamoto, T., Ohta, Boyer, S., & Zamora, I. (2002). New methods in predictive metabolism. T., et al. (2003). Global expression analysis of N-methyl-N-nitro-V N- Journal of Computer-Aided Molecular Design, 16, 403–413. nitrosoguanidine-induced rat stomach carcinomas using oligonucleotide Bredel, M., & Jacoby, E. (2004). Chemogenomics: An emerging strategy microarrays. Carcinogenesis, 24, 861–867. for rapid target and drug discovery. Nature Reviews. Genetics, 5, Adachi, T., Ono, Y., Koh, K. B., Takashima, K., Tainaka, H., Matsuno, Y., 262–275. et al. (2004). Long-term alteration of gene expression without Bugrim, A., Nikolskaya, T., & Nikolsky, Y. (2004). Early prediction of drug morphological change in testis after neonatal exposure to genistein in metabolism and toxicity: Systems biology approach and modeling. mice: Toxicogenomic analysis using cDNA microarray. Food and Drug Discovery Today, 9, 127–135. Chemical Toxicology, 42, 445–452. Bulera, S. J., Eddy, S. M., Ferguson, E., Jatkoe, T. A., Reindel, J. F., Akerman, G. S., Rosenzweig, B. A., Domon, O. E., McGarrity, L. J., Bleavins, M. R., et al. (2001). RNA expression in the early character- Blankenship, L. R., Tsai, C. A., et al. (2004). Gene expression profiles ization of hepatotoxicants in Wistar rats by high-density DNA micro- and genetic damage in benzo(a)pyrene diol epoxide-exposed TK6 cells. arrays. Hepatology, 33, 1239–1258. Mutation Research, 549, 43–64. Butcher, E. C., Berg, E. L., & Kunkel, E. J. (2004). Systems biology in drug Aksenov, S. V., Church, B., Dhiman, A., Georgieva, A., Sarangapani, R., discovery. Nature Biotechnology, 22, 1253–1259. Helmlinger, G., et al. (2005). An integrated approach for inference and Butte, A. (2002). The use and analysis of microarray data. Naturalist mechanistic modeling for advancing drug development. FEBS Letters, Review of Drug Discovery, 1, 951–960. 579, 1878–1883. Cary, M. P., Bader, G. D., & Sander, C. (2005). Pathway information for Anderle, P., Huang, Y., & Sadee, W. (2004). Intestinal membrane transport systems biology. FEBS Letters, 579, 1815–1820. of drugs and nutrients: Genomics of membrane transporters using Cheok, M. H., Yang, W., Pui, C. H., Downing, J. R., Cheng, C., Naeve, expression microarrays. European Journal of Pharmaceutical Sciences, C. W., et al. (2003). Treatment-specific changes in gene expression 21, 17–24. discriminate in vivo drug response in human leukemia cells. Nature Annereau, J. P., Szakacs, G., Tucker, C. J., Arciello, A., Cardarelli, C., Genetics, 34, 85–90. Collins, J., et al. (2004). Analysis of ATP-binding cassette transporter Cleary, J. D., Rogers, P. D., & Chapman, S. W. (2001). Differential expression in drug-selected cell lines by a microarray dedicated to transcription factor expression in human mononuclear cells in response multidrug resistance. Molecular Pharmacology, 66, 1397–1405. to amphotericin B: Identification with complementary DNA microarray Apic, G., Ignjatovic, T., Boyer, S., & Russell, R. B. (2005). Illuminating drug technology. Pharmacotherapy, 21, 1046–1054. discovery with biological pathways. FEBS Letters, 579, 1872–1877. Coen, M., Ruepp, S. U., Lindon, J. C., Nicholson, J. K., Pognan, F., Lenz, Baker, V. A., Harries, H. M., Waring, J. F., Duggan, C. M., Ni, H. A., Jolly, E. M., et al. (2004). Integrated application of transcriptomics and R. A., et al. (2004). Clofibrate-induced gene expression changes in rat metabonomics yields new insight into the toxicity due to paracetamol in liver: A cross-laboratory analysis using membrane cDNA arrays. the mouse. Journal of Pharmaceutical and Biomedical Analysis, 35, Environmental Health Perspectives, 112, 428–438. 93–105. Balakin, K. V., Ekins, S., Bugrim, A., Ivanenkov, Y. A., Korolev, D., & Cornwell, P. D., De Souza, A. T., & Ulrich, R. G. (2004). Profiling of Nikolsky, Y., et al. (2004a). Quantitative structure–metabolism relation- hepatic gene expression in rats treated with fibric acid analogs. ship modeling of the metabolic N-dealkylation rates. Drug Metabolism Mutation Research, 549, 131–145. and Disposition, 32, 1111–1120. Csermely, P., Agoston, V., & Pongor, S. (2005). The efficiency of multi- Balakin, K. V., Ekins, S., Bugrim, A., Ivanenkov, Y. A., Korolev, D., & target drugs: The network approach might help drug design. Trends in Nikolsky, Y., et al. (2004b). Kohonen maps for prediction of binding to Pharmacological Sciences, 26, 178–182. human cytochrome P450 3A4. Drug Metabolism and Disposition, 32, Cunningham, M. J., Liang, S., Fuhrman, S., Seilhamer, J. J., & 1183–1189. Somogyi, R. (2000). Gene expression microarray data analysis for Barabasi, A. -L., & Oltvai, Z. N. (2004). Network biology: Under- toxicology profiling. Annals of the New York Academy of Sciences, 919, standing the cell’s functional organization. Nature Reviews Genetics, 52–67. 5, 101–113. de Longueville, F., Atienzar, F. A., Marcq, L., Dufrane, S., Evrard, S., Bartosiewicz, M., Penn, S. G., & Buckpitt, A. (2001). Applications of gene Wouters, L., et al. (2003). Use of a low density microarray for studying arrays in environmental toxicology: Fingerprints of gene regulation gene expression patterns induced by hepatotoxicants on primary associated with cadmium chloride, benzo(a)pyrene, and trichloro- cultures of rat hepatocytes. Toxicological Sciences, 75, 378–392. ethylene. Environmental Health Perspectives, 109, 71–74. Dobrin, R., Beg, Q. K., Barabasi, A. L., & Oltvai, Z. N. (2004). Bertilsson, G., Heidrich, J., Svensson, K., Asman, M., Jendeberg, L., Aggregation of topological motifs in the Escherichia coli transcriptional Sydow-Backman, M., et al. (1998). Identification of a human nuclear regulatory network. BMC Bioinformatics, 5, 10. receptor defines a new signaling pathway for CYP3A induction. Donald, S., Verschoyle, R. D., Edwards, R., Judah, D. J., Davies, R., Proceedings of the National Academy of Sciences of the United States Riley, J., et al. (2002). Hepatobiliary damage and changes in hepatic of America, 95, 12208–12213. gene expression caused by the antitumor drug ecteinascidin-743 (ET- Blumberg, B., Sabbagh Jr., W., Juguilon, H., Bolado Jr., J., van Meter, 743) in the female rat. Cancer Research, 62, 4256–4262. C. M., Ong, E. S., et al. (1998). SXR, a novel steroid and xenobiotic- Dumont, M., Jacquemin, E., D’Hont, C., Descout, C., Cresteil, D., Haouzi, sensing nuclear receptor. Genes and Development, 12, 3195–3205. D., et al. (1997). Expression of the liver Na+-independent organic anion Borodina, Y., Rudik, A., Filimonov, D., Kharchevnikova, N., Dmitriev, A., transporting polypeptide (oatp-1) in rats with bile duct ligation. Journal Blinova, V., et al. (2004). A new statistical approach to predicting of Hepatology, 27, 1051–1056. aromatic hydroxylation sites. Comparison with model-based Durnas, C., Loi, C. M., & Cusack, B. J. (1990). Hepatic drug metabolism approaches. Journal of Chemical Information and Computer Sciences, and aging. Clinical Pharmacokinetics, 19, 359–389. 44, 1998–2009. Eisen, M. B., Spellman, P. T., Brown, P. O., & Botstein, D. (1998). Cluster Borodina, Y., Sadym, A., Filimonov, D., Blinova, V., Dmitriev, A., & analysis and display of genome-wide expression patterns. Proceedings Poroikov, V. (2003). Predicting biotransformation potential from of the National Academy of Sciences of the United States of America, molecular structure. Journal of Chemical Information and Computer 95, 14863–14868. Sciences, 43, 1636–1646. Ekins, S., Andreyev, S., Ryabov, A., Kirilov, E., Rakhmatulin, E. A., Bottone Jr., F. G., Martinez, J. M., Collins, J. B., Afshari, C. A., & Eling, Bugrim, A., et al. (in press). Computational Prediction of Human Drug T. E. (2003). Gene modulation by the cyclooxygenase inhibitor, Metabolism. Exp Opin Drug Metab Toxicol. 62 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Ekins, S., Boulanger, B., Swaan, P. W., & Hupcey, M. A. Z. (2002). and toxicology using DNA microarrays. Physiological Genomics, 5, Towards a new age of virtual ADME/TOX and multidimensional drug 161–170. discovery. Journal of Computer-Aided Molecular Design, 16, 381–401. Guo, G. L., Choudhuri, S., & Klaassen, C. D. (2002). Induction profile of Ekins, S., Bugrim, A., Nikolsky, Y., & Nikolskaya, T. (2005). Systems rat organic anion transporting polypeptide 2 (oatp2) by prototypical biology: Applications in drug discovery. In S. Gad (Ed.), Drug drug-metabolizing enzyme inducers that activate gene expression Discovery Handbook (pp. 123–183). Hoboken, NJ’ Wiley and Sons. through ligand-activated transcription factor pathways. Journal of Ekins, S., Giroux, C. N., Nikolsky, Y., Bugrim, A., & Nikolskaya, T. Pharmacology and Experimental Therapeutics, 300, 206–212. (2005). A signature gene network approach to toxicity. The Toxoco- Hagenbuch, B., & Meier, P. J. (2004). Organic anion transporting logists, 84. polypeptides of the OATP/ SLC21 family: Phylogenetic classification Ekins, S., Kirillov, E., Rakhmatulin, E., & Nikolskaya, T. (2005). A novel as OATP/ SLCO superfamily, new nomenclature and molecular/func- method for visualizing nuclear hormone receptor networks relevant to tional properties. Pflugers Archiv, 447, 653–665. drug metabolism. Drug Metabolism and Disposition, 33, 474–481. Hamadeh, H. K., & Afshari, C. A. (2004). Toxicogenomics principles and Ekins, S., & McGowan, R. J. (2001). The limits of reductionism: The applications. Hoboken’ John Wiley and Sons. shifting genomic paradigm’s impact on industry and academia. Hamadeh, H. K., Bushel, P. R., Jayadev, S., Martin, K., DiSorbo, O., Sieber, Philsophy in Science, 9, 1–23. S., et al. (2002). Gene expression analysis reveals chemical-specific Ekins, S., Mirny, L., & Schuetz, E. G. (2002). A ligand-based approach to profiles. Toxicological Sciences, 67, 219–231. understanding selectivity of nuclear hormone receptors PXR, CAR, Hamadeh, H. K., Jayadev, S., Gaillard, E. T., Huang, Q., Stoll, R., FXR, LXRa and LXRb. Pharmaceutical Research, 19, 1788–1800. Blanchard, K., et al. (2004). Integration of clinical and gene expression Ekins, S., Nikolsky, Y., & Nikolskaya, T. (2005). Techniques: Application endpoints to explore furan-mediated hepatotoxicity. Mutation Research, of systems biology to absorption, distribution, metabolism, excretion, 549, 169–183. and toxicity. Trends in Pharmacological Sciences, 26, 202–209. Hamadeh, H. K., Knight, B. L., Haugen, A. C., Sieber, S., Amin, R. P., Ekins, S., Ring, B. J., Grace, J., McRobie-Belle, D. J., & Wrighton, S. Bushel, P. R., et al. (2002). Methapyrilene toxicity: Anchorage of A. (2000). Present and future in vitro approaches for drug pathologic observations to gene expression alterations. Toxicologic metabolism. Journal of Pharmacological and Toxicological Methods, Pathology, 30, 470–482. 44, 313–324. Han, J. D., Bertin, N., Hao, T., Goldberg, D. S., Berriz, G. F., Zhang, L. V., Ekins, S., Waller, C. L., Swaan, P. W., Cruciani, G., Wrighton, S. A., & et al. (2004). Evidence for dynamically organized modularity in the Wikel, J. H. (2000). Progress in predicting human ADME parameters in yeast protein–protein interaction network. Nature, 430, 88–93. silico. Journal of Pharmacological and Toxicological Methods, 44, Hanisch, D., Zien, A., Zimmer, R., & Lengauer, T. (2002). Co-clustering of 251–272. biological networks and gene expression data. Bioinformatics, 18, Erhardt, P. W. (2003). A human drug metabolism database: Potential roles S145–S154. in the quantitative predictions of drug metabolism and metabolism- Harris, A. J., Dial, S. L., & Casciano, D. A. (2004). Comparison of basal related drug–drug interactions. Current Drug Metabolism, 4, 411–422. gene expression profiles and effects of hepatocarcinogens on gene Evans, W. E., & McLeod, H. L. (2003). Pharmacogenomics—drug expression in cultured primary human hepatocytes and HepG2 cells. disposition, drug targets, and side effects. New England Journal of Mutation Research, 549, 79–99. Medicine, 348, 538–549. Harrison, S. C. (2004). Whither structural biology. Nature Structural Ezendam, J., Staedtler, F., Pennings, J., Vandebriel, R. J., Pieters, R., Molecular Biology, 11, 12–15. Boffetta, P., et al. (2004). Toxicogenomics of subchronic hexachlor- Hartley, D. P., Dai, X., He, Y. D., Carlini, E. J., Wang, B., Huskey, S. E., obenzene exposure in Brown Norway rats. Environmental Health et al. (2004). Activators of the rat pregnane X receptor differentially Perspectives, 112, 782–791. modulate hepatic and intestinal gene expression. Molecular Pharma- Faiola, B., Fuller, E. S., Wong, V. A., & Recio, L. (2004). Gene expression cology, 65, 1159–1171. profile in bone marrow and hematopoietic stem cells in mice exposed to Hartwell, L. H., Hopfield, J. J., Leibler, S., & Murray, A. W. (1999). From inhaled benzene. Mutation Research, 549, 195–212. molecular to modular cell biology. Nature, 402, C47–C52. FDA (2004). Innovation stagnation: Challenge and opportunity on the Hasmall, S., Orphanides, G., James, N., Pennie, W., Hedley, K., Soames, critical path to new medicinal products. A., et al. (2002). Downregulation of lactoferrin by PPARalpha ligands: Fiehn, O. (2001). Combining genomics, metabolome analysis, and bio- Role in perturbation of hepatocyte proliferation and apoptosis. chemical modelling to understand metabolic networks. Comparative Toxicology Sciences, 68, 304–313. Functional Genomics, 2, 155–168. Hayase, T., Yamamoto, Y., Yamamoto, K., Muso, E., & Shiota, K. (2004). Field, F. J., Born, E., & Mathur, S. N. (2004). LXR/RXR ligand activation Microarray profile analysis of toxic cocaine-induced alterations in the enhances basolateral efflux of beta-sitosterol in CaCo-2 cells. Journal of expression of mouse brain gene sequences: A possible ’protective’ Lipid Research, 45, 905–913. effect of buprenorphine. Journal of Applied Toxicology, 24, 15–20. Fleck, C., Sutter, L., Appenroth, D., Koch, B., Meinhold, T., Pitack, M., Hayes, K. R., Vollrath, A. L., Zastrow, G. M., McMillan, B. J., Craven, M., et al. (2003). Use of gene chip technology for the characterisation of Jovanovich, S., et al. (2005). EDGE: A centralized resource for the the regulation of renal transport processes and of nephrotoxicity in comparison, analysis, and distribution of toxicogenomic information. rats. Experimental and Toxicologic Pathology, 54, 401–410. Molecular Pharmacology, 67, 1360–1368. Fliri, A. F., Loging, W. T., Thadeio, P. F., & Volkmann, R. A. (2005). Heijne, W. H., Slitt, A. L., van Bladeren, P. J., Groten, J. P., Klaassen, C. D., Biological spectra analysis: Linking biological activity profiles to Stierum, R. H., et al. (2004). Bromobenzene-induced hepatotoxicity at molecular structure. Proceedings of the National Academy of Sciences the transcriptome level. Toxicology Sciences, 79, 411–422. of the United States of America, 102, 261–266. Heijne, W. H., Stierum, R. H., Slijper, M., van Bladeren, P. J., & van Fountoulakis, M. (2004). Application of proteomics technologies in the Ommen, B. (2003). Toxicogenomics of bromobenzene hepatotoxicity: investigation of the brain. Mass Spectrometry Reviews, 23, 231–258. A combined transcriptomics and proteomics approach. Biochemical Gaucher, S. P., Taylor, S. W., Fahy, E., Zhang, B., Warnock, D. E., Ghosh, Pharmacology, 65, 857–875. S. S., et al. (2004). Expanded coverage of the human heart Heinloth, A. N., Irwin, R. D., Boorman, G. A., Nettesheim, P., Fannin, mitochondrial proteome using multidimensional liquid chromatography R. D., Sieber, S. O., et al. (2004). Gene expression profiling of rat coupled with tandem mass spectrometry. Journal of Proteome livers reveals indicators of potential adverse effects. Toxicology Research, 3, 495–505. Sciences, 80, 193–202. Gerhold, D., Lu, M., Xu, J., Austin, C., Caskey, C. T., & Rushmore, T. Hodges, L. C., Cook, J. D., Lobenhofer, E. K., Li, L., Bennett, L., Bushel, (2001). Monitoring expression of genes involved in drug metabolism P. R., et al. (2003). Tamoxifen functions as a molecular agonist inducing S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 63

cell cycle-associated genes in breast cancer cells. Molecular Cancer administered ethinylestradiol for 28 days using a microarray technique. Research, 1, 300–311. Toxicology, 200, 179–192. Holleman, A., Cheok, M. H., den Boer, M. L., Yang, W., Veerman, K.M., Kier, L. D., Neft, R., Tang, L., Suizu, R., Cook, T., Onsurez, K., et al. A. J., Kazemier, et al. (2004). Gene-expression patterns in drug-resistant (2004). Applications of microarrays with toxicologically relevant genes acute lymphoblastic leukemia cells and response to treatment. New (tox genes) for the evaluation of chemical toxicants in Sprague Dawley England Journal of Medicine, 351, 533–542. rats in vivo and human hepatocytes in vitro. Mutation Research, 549, Hong, Y., Muller, U. R., & Lai, F. (2003). Discriminating two classes of 101–113. toxicants through expression analysis of HepG2 cells with DNA arrays. Kijima, K., Toyosawa, K., Yasuba, M., Matsuoka, N., Adachi, T., Toxicology in Vitro, 17, 85–92. Komiyama, M., et al. (2004). Gene expression analysis of the rat testis Hood, E. (2003). Proteomics: Characterizing the cogs in the machinery of after treatment with di(2-ethylhexyl) phthalate using cDNA microarray life. Environmental Health Perspectives, 111, A816–A825. and real-time RT-PCR. Toxicology and Applied Pharmacology, 200, Hood, L. (2003). Systems biology: Integrating technology, biology, and 103–110. computation. Mechanism of Ageing and Development, 124, 9–16. Kim, R. B. (2003). Organic anion-transporting polypeptide (OATP) trans- Hood, L., & Galas, D. (2003). The digital code of DNA. Nature, 421, porter family and drug disposition. European Journal of Clinical 444–448. Investigation, 33,1–5. Hood, L., & Perlmutter, R. M. (2004). The impact of systems approaches on Kinirons, M. T., & O’Mahony, M. S. (2004). Drug metabolism and ageing. biological problems in drug discovery. Nature Biotechnology, 22, British Journal of Clinical Pharmacology, 57, 540–544. 1215–1217. Kitano, H. (2002a). Computational systems biology. Nature, 420, 206–210. Hu, T., Gibson, D. P., Carr, G. J., Torontali, S. M., Tiesman, J. P., Chaney, Kitano, H. (2002b). Systems biology: A brief overview. Science, 295, J. G., et al. (2004). Identification of a gene expression profile that 1662–1664. discriminates indirect-acting genotoxins from direct-acting genotoxins. Kliewer, S. A., Moore, J. T., Wade, L., Staudinger, J. L., Watson, M. Mutation Research, 549, 5–27. A., Jones, S. A., et al. (1998). An orphan nuclear receptor activated Huang, Q., Dunn II, R. T., Jayadev, S., DiSorbo, O., Pack, F. D., Farr, S. B., by pregnanes defines a novel steroid signalling pathway. Cell, 92, et al. (2001). Assessment of cisplatin-induced nephrotoxicity by 73–82. microarray technology. Toxicology Sciences, 63, 196–207. Korolev, D., Balakin, K. V., Nikolsky, Y., Kirillov, E., Ivanenkov, Y. A., Huang, Q., Jin, X., Gaillard, E. T., Knight, B. L., Pack, F. D., Stoltz, J. H., et Savchuk, N. P., et al. (2003). Modeling of human cytochrome p450- al. (2004). Gene expression profiling reveals multiple toxicity endpoints mediated drug metabolism using unsupervised machine learning induced by hepatotoxicants. Mutation Research, 549, 147–167. approach. Journal of Medicinal Chemistry, 46, 3631–3643. Huisman, M. T., Smit, J. W., Crommentuyn, K. M., Zelcer, N., Wiltshire, Kramer, J. A., Pettit, S. D., Amin, R. P., Bertram, T. A., Car, B., H. R., Beijnen, J. H., et al. (2002). Multidrug resistance protein 2 Cunningham, M., et al. (2004). Overview on the application of (MRP2) transports HIV protease inhibitors, and transport can be transcription profiling using selected nephrotoxicants for toxicology enhanced by other drugs. Aids, 16, 2295–2301. assessment. Environmental Health Perspectives, 112, 460–464. Hunt, C. M., Westerkam, W. R., & Stave, G. M. (1992). Effect of age and Kultima, K., Nystrom, A. -M., Scholz, B., Gustafson, A. -L., Dencker, L., gender on the activity of human hepatic CYP3A. Biochemical & Stigson, M. (2004). Valproic acid teratogenicity: A toxicogenomics Pharmacology, 44, 275–283. approach. Environmental Health Perspectives, 112, 1225–1235. Ideker, T., Ozier, O., Schwikowski, B., & Siegal, A. F. (2002). Discovering Kutuzova, G. D., & Deluca, H. F. (2004). Gene expression profiles in rat regulatory and signalling circuits in molecular interaction networks. intestine identify pathways for 1,25-dihydroxyvitamin D(3) stimulated Bioinformatics, 18, S233–S240. calcium absorption and clarify its immunomodulatory properties. Ideker, T., Thorsson, V., Ranish, J. A., Christmas, R., Buhler, J., Eng, J. K., Archives of Biochemistry and Biophysics, 432, 152–166. et al. (2001). Integrated genomic and proteomic analyses of a system- Kwak, M. K., Wakabayashi, N., Itoh, K., Motohashi, H., Yamamoto, M., & atically perturbed network. Science, 292, 929–934. Kensler, T. W. (2003). Modulation of gene expression by cancer Iida, M., Anna, C. H., Hartis, J., Bruno, M., Wetmore, B., Dubin, J. R., et al. chemopreventive dithiolethiones through the Keap1-Nrf2 pathway. (2003). Changes in global gene and protein expression during early Identification of novel gene clusters for cell survival. Journal of mouse liver carcinogenesis induced by non-genotoxic model carcino- Biological Chemistry, 278, 8135–8145. gens oxazepam and Wyeth-14,643. Carcinogenesis, 24, 757–770. Lahjouji, K., Mitchell, G. A., & Qureshi, I. A. (2001). Carnitine transport Ingelman-Sundberg, M. (2004). Pharmacogenetics of cytochrome P450 and by organic cation transporters and systemic carnitine deficiency. its applications in drug therapy: The past, present and future. Trends in Molecular Genetics and Metabolism, 73, 287–297. Pharmacological Sciences, 25, 193–200. Landowski, C. P., Sun, D., Foster, D. R., Menon, S. S., Barnett, J. L., Jeong, H., Mason, S. P., Barabasi, A. L., & Oltvai, Z. N. (2001). Lethality Welage, L. S., et al. (2003). Gene expression in the human intestine and and centrality in protein networks. Nature, 411, 41–42. correlation with oral valacyclovir pharmacokinetic parameters. Journal Jeong, H., Tombor, B., Albert, R., Oltvai, Z. N., & Barabasi, A. L. of Pharmacology and Experimental Therapeutics, 306, 778–786. (2000). The large-scale organization of metabolic networks. Nature, Lee, G. -J., Lee, W. -S., Jeon, K. -S., Um, C. -H., Kim, S. -J., Lee, C. -H., 407, 651–654. et al. (2004). cDNA microarray gene expression analysis and Jung, D., Hagenbuch, B., Gresh, L., Pontoglio, M., Meier, P. J., & Kullak- toxicological phenotype for anticancer drug. Journal of Veterinary Ublick, G. A. (2001). Characterization of the human OATP-C Medical Science, 66, 1339–1345. (SLC21A6) gene promoter and regulation of liver-specific OATP genes Lee, M., Kwon, J., Kim, S. N., Kim, J. E., Koh, W. S., Kim, E. J., et al. by hepatocyte nuclear factor 1 alpha. Journal of Biological Chemistry, (2003). cDNA microarray gene expression profiling of hydroxyurea, 276, 37206–37214. paclitaxel, and p-anisidine, genotoxic compounds with differing Jung, D., & Kullak-Ublick, G. A. (2003). Hepatocyte nuclear factor 1 tumorigenicity results. Environmental and Molecular Mutagenesis, alpha: A key mediator of the effect of bile acids on gene expression. 42, 91–97. Hepatology, 37, 622–631. Li, S., Armstrong, C. M., Bertin, N., Ge, H., Milstein, S., Boxem, M., et al. Jung, J. -W., Park, J. -S., Hwang, J. -W., Kang, K. -S., Y.S., L., Song, B. (2004). A map of the interactome network of the metazoan C. elegans. -S., et al. (2004). Gene expression analysis of peroxisome prolifer- Science, 303, 540–543. ators- and phenytoin-induced hepatotoxicity using cDNA microarray. Liguori, M. J., Anderson, L. M., Bukofzer, S., McKim, J., Pregenzer, J. F., Journal of Veterinary Medical Science, 66, 1329–1333. Retief, J., et al. (2005). Microarray analysis in human hepatocytes Kato, N., Shibutani, M., Takagi, H., Uneyama, C., Lee, K. -Y., Takiami, S., suggests a mechanism for hepatotoxicity induced by trovafloxacin. et al. (2004). Gene expression profile in the livers of rats orally Hepatology, 41, 177–186. 64 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Lobenhofer, E. K., Bennett, L., Cable, P. L., Li, L., Bushel, P. R., & Afshari, transport of troglitazone sulfate: Implications for understanding trogli- C. A. (2002). Regulation of DNA replication fork genes by 17beta- tazone hepatotoxicity. Drug Metabolism and Disposition, 32, 291–294. estradiol. Molecular Endocrinology, 16, 1215–1229. Oliver, D. E., Rubin, D. L., Stuart, J. M., Hewett, M., Klein, T. E., & Locker, J., Tian, J., Carver, R., Concas, D., Cossu, C., Ledda-Columbano, Altman, R. B. (2002). Ontology development for a pharmacogenetics G. M., et al. (2003). A common set of immediate-early response genes knowledge base. Package Symposion, 88–99. in liver regeneration and hyperplasia. Hepatology, 38, 314–325. Oram, J. F., & Lawn, R. M. (2001). ABCA1. The gatekeeper for eliminating Mattes, W. B., Pettit, S. D., Sansone, S. A., Bushel, P. R., & Waters, M. D. excess tissue cholesterol. Journal of Lipid Research, 42, 1173–1179. (2004). Database development in toxicogenomics: Issues and efforts. Oswald, M., Kullak-Ublick, G. A., Paumgartner, G., & Beuers, U. (2001). Environmental Health Perspectives, 112, 495–505. Expression of hepatic transporters OATP-C and MRP2 in primary McMillian, M., Nie, A. Y., Parker, J. B., Leone, A., Bryant, S., Kemmerer, sclerosing cholangitis. Liver, 21, 247–253. M., et al. (2004). A gene expression signature for oxidant stress/reactive Parsons, A. B., Brost, R. I., Ding, H., Li, Z., Zhang, C., Sheikh, B., et al. metabolites in rat liver. Biochemical Pharmacology, 68, 2249–2261. (2004). Integration of chemical-genetic and genetic interaction data Meier, P. J., & Stieger, B. (2000). Molecular mechanisms in bile formation. links bioactive compounds to cellular target pathways. Nature Bio- News in Physiological Sciences, 15, 89–93. technology, 22, 62–69. Meneses-Lorente, G., de Longueville, F., Dos Santos-Mendes, S., Bonnert, Patki, K. C., von Moltke, L. L., Harmatz, J. S., Hesse, L. M., Court, M. H., T. P., Jack, A., Evrard, S., et al. (2003). An evaluation of a low-density & Greenblatt, D. J. (2004). Effect of age on in vitro triazolam DNA microarray using cytochrome P450 inducers. Chemical Research biotransformation in male human liver microsomes. Journal of in Toxicology, 16, 1070–1077. Pharmacology and Experimental Therapeutics, 308, 874–879. Meneses-Lorente, G., Guest, P. C., Lawrence, J., Muniappa, N., Knowles, Pereira-Leal, J. B., Enright, A. J., & Ouzounis, C. A. (2004). Detection of M. R., Skynner, H. A., et al. (2004). A proteomic investigation of drug- functional modules from protein interaction networks. Proteins, 54, induced steatosis in rat liver. Chemical Research in Toxicology, 17, 49–57. 605–612. Plant, N. (2004). Interaction networks: Coordinating responses to xeno- Merrill, C. L., Ni, H., Yoon, L. W., Tirmenstein, M. A., Narayanan, P., biotic exposure. Toxicology, 202, 21–32. Benavides, G. R., et al. (2002). Etomoxir-induced oxidative stress in Raghavendra Prasad, H. S., Qi, Z., Srinivasan, K. N., & Gopalakrish- HepG2 cells detected by differential gene expression is confirmed nakone, P. (2004). Potential effects of tetrodotoxin exposure to biochemically. Toxicology Sciences, 68, 93–101. human glial cells postulated using microarray approach. Toxicon, 44, Milo, R., Shen-Orr, S., Itzkovitz, S., Kashtan, N., Chklovskii, D., & Alon, 597–608. U. (2002). Network motifs: Simple building blocks of complex Reilly, T. P., Bourdi, M., Brady, J. N., Pise-Masison, C. A., Radonovich, M. networks. Science, 298, 824–827. F., George, J. W., et al. (2001). Expression profiling of acetaminophen Morgan, K. T., Ni, H., Brown, H. R., Yoon, L., Qualls, C. W., Crosby, L. liver toxicity in mice using microarray technology. Biochemical and M., et al. (2002). Application of cDNA microarray technology to in Biophysical Research Communications, 282, 321–328. vitro toxicology and the selection of genes for a real-time RT-PCR- Rives, A. W., & Galitski, T. (2003). Modular organization of cellular based screen for oxidative stress in Hep-G2 cells. Toxicologic networks. Proceedings of the National Academy of Sciences of the Pathology, 30, 435–451. United States of America, 100, 1128–1133. Mwinyi, J., Johne, A., Bauer, S., Roots, I., & Gerloff, T. (2004). Evidence Rogers, P. D., Pearson, M. M., Cleary, J. D., Sullivan, D. C., & Chapman, for inverse effects of OATP-C (SLC21A6) 5 and 1b haplotypes on S. W. (2002). Differential expression of genes encoding immunomodu- pravastatin kinetics. Clinical Pharmacology and Therapeutics, 75, latory proteins in response to amphotericin B in human mononuclear 415–421. cells identified by cDNA microarray analysis. Journal of Antimicrobial Nakata, K., Yukawa, M., Komiyama, N., Nakano, T., & Kaminuma, T. Chemotherapy, 50, 811–817. (2002). A nuclear receptor database that maps pathways to diseases. Rosen, M. B., Wilson, V. S., Schmid, J. E., & Gray, L. E. (2005). Gene Genome Infomatics, 13, 515–516. expression analysis in the ventral prostate of rats exposed to vinclozolin Nicholson, J. K., Holmes, E., Lindon, J. C., & Wilson, I. D. (2004). The or procymidone. Reproductive Toxicology, 19, 367–379. challenges of modeling mammalian biocomplexity. Nature Biotechnol- Rost, D., Herrmann, T., Sauer, P., Schmidts, H. L., Stieger, B., Meier, P. J., ogy, 22, 1268–1274. et al. (2003). Regulation of rat organic anion transporters in bile salt- Nicholson, J. K., & Wilson, I. D. (2003). Understanding ’global, systems induced cholestatic hepatitis: Effect of ursodeoxycholate. Hepatology, biology: Metabonomics and the continuum of metabolism. Nature 38, 187–195. Reviews on Drug Discovery, 2, 668–676. Rothman, R. B., Ayestas, M. A., Dersch, C. M., & Baumann, M. H. (1999). Nikitin, A., Egorov, S., Daraselia, N., & Mazo, I. (2003). Pathway studio— Aminorex, fenfluramine, and chlorphentermine are serotonin transporter the analysis and navigation of molecular networks. Bioinformatics, 19, substrates. Implications for primary pulmonary hypertension. Circula- 2155–2157. tion, 100, 869–875. Nikolsky, Y., Ekins, S., Nikolskaya, T., & Bugrim, A. (2005). A novel Ruepp, S. U., Tonge, R. P., Shaw, J., Wallis, N., & Pognan, F. (2002). method for generation of signature networks as biomarkers from Genomics and proteomics analysis of acetaminophen toxicity in mouse complex high throughput data. Toxicology Letters, 158, 20–29. liver. Toxicology Sciences, 65, 135–150. Nikolsky, Y., Nikolskaya, T., & Bugrim, A. (2005). Biological networks Sakaeda, T., Nakamura, T., & Okumura, K. (2002). MDR1 genotype- and analysis of experimental data in drug discovery. Drug Discovery related pharmacokinetics and pharmacodynamics. Biological & Phar- Today, 10, 653–662. maceutical Bulletin, 25, 1391–1400. Nishizato, Y., Ieiri, I., Suzuki, H., Kimura, M., Kawabata, K., Hirota, T., Sakata, T., Anzai, N., Shin, H. J., Noshiro, R., Hirata, T., Yokoyama, H., et al. (2003). Polymorphisms of OATP-C (SLC21A6) and OAT3 et al. (2004). Novel single nucleotide polymorphisms of organic cation (SLC22A8) genes: Consequences for pravastatin pharmacokinetics. transporter 1 (SLC22A1) affecting transport functions. Biochemical Clinical Pharmacology and Therapeutics, 73, 554–565. and Biophysical Research Communications, 313, 789–793. Nozawa, T., Nakajima, M., Tamai, I., Noda, K., Nezu, J., Sai, Y., et al. Satlin, L. M., Amin, V., & Wolkoff, A. W. (1997). Organic anion (2002). Genetic polymorphisms of human organic anion transporters transporting polypeptide mediates organic anion/HCO3-exchange. OATP-C (SLC21A6) and OATP-B (SLC21A9): Allele frequencies in Journal of Biological Chemistry, 272, 26340–26345. the Japanese population and functional analysis. Journal of Pharmacol- Satomi, Y., Tsuchiya, W., Mihara, K., Ota, M., Kasahara, Y., & Akahori, F. ogy and Experimental Therapeutics, 302, 804–813. (2004). Gene expression analysis of the lung following paraquat Nozawa, T., Sugiura, S., Nakajima, M., Goto, A., Yokoi, T., Nezu, J., et al. administration in rats using DNA microarray. Journal of Toxicological (2004). Involvement of organic anion transporting polypeptides in the Sciences, 29, 91–100. S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66 65

Schrenk, D., Baus, P. R., Ermel, N., Klein, C., Vorderstemann, B., & Thomas, D. M., Francescutti-Verbeem, D. M., Liu, X., & Kuhn, D. M. Kauffmann, H. M. (2001). Up-regulation of transporters of the MRP (2004). Identification of differentially regulated transcripts in mouse family by drugs and toxins. Toxicology Letters, 120, 51–57. striatum following methamphetamine treatment—an oligonucleotide Segal, E., Shapira, M., Regev, A., Pe’er, D., Botstein, D., Koller, D., et al. microarray approach. Journal of Neurochemistry, 88, 380–393. (2003). Module networks: Identifying regulatory modules and their Thomas, R. S., Rank, D. R., Penn, S. G., Zastrow, G. M., Hayes, K. R., condition-specific regulators from gene expression data. Nature Pande, K., et al. (2001). Identification of toxicologically predictive Genetics, 34, 166–176. gene sets using cDNA microarrays. Molecular Pharmacology, 60, Segal, E., Wang, H., & Koller, D. (2003). Discovering molecular pathways 1189–1194. from protein interaction and gene expression. Bioinformatics, 19, Thompson, K. L., Afshari, C. A., Amin, R. P., Bertram, T. A., Car, B., i264–i272. Cunningham, M., et al. (2004). Identification of platform-independent Seidel, S. D., Kan, H. L., Stott, W. T., Schisler, M. R., & Gollapudi, B. B. gene expression markers of cisplatin nephrotoxicity. Environmental (2003). Identification of transcriptome profiles for the DNA-damaging Health Perspectives, 112, 488–494. agents bleomycin and hydrogen peroxide in L5178Y mouse lymphoma Tirona, R. G., Leake, B. F., Merino, G., & Kim, R. B. (2001). cells. Environmental and Molecular Mutagenesis, 42, 19–25. Polymorphisms in OATP-C: Identification of multiple allelic Shan, L., Yu, Y., Schut, H. A. J., & Snyderwine, E. G. (2004). Susceptibility variants associated with altered transport activity among European- of rats to mammary gland carcinogenesis by the food-derived and African-Americans. Journal of Biological Chemistry, 276, carcinogen 2-Amino-1-Methyl-6-Phenylimidazo[4,5-b]Pyridine (PhIP) 35669–35675. varies with age and is associated with the induction of differential gene Tirona, R. G., Leake, B. F., Wolkoff, A. W., & Kim, R. B. (2003). expression. American Journal of Pathology, 165, 191–202. Human organic anion transporting polypeptide-C (SLC21A6) is a Sharom, J. R., Bellows, D. S., & Tyers, M. (2004). From large networks to major determinant of the rifampin-mediated pregnane X receptor small molecules. Current Opinion in Chemical Biology, 8, 81–90. activation. Journal of Pharmacology and Experimental Therapeutics, Shitara, Y., Hirano, M., Sato, H., & Sugiyama, Y. (2004). Gemfibrozil and 304, 223–228. its glucuronide inhibit the OATP2(OATP1B1: SLC21A6)-mediated Tong, W., Cao, X., Harris, S., Sun, H., Fang, H., Fuscoe, J., et al. (2003). hepatic uptake and CYP2C8-mediated metabolism of cerivastatin— ArrayTrack—supporting toxicogenomic research at the U.S. Food and Analysis of the mechanism of the clinically relevant drug–drug Drug Administration National Center for Toxicological Research. interaction between cerivastatin and gemfibrozil. Journal of Pharma- Environmental Health Perspectives, 111, 1819–1826. cology and Experimental Therapeutics. Tornow, S., & Mewes, H. W. (2003). Functional modules by relating Shitara, Y., Itoh, T., Sato, H., Li, A. P., & Sugiyama, Y. (2003). Inhibition of protein interaction networks and gene expression. Nucleic Acids transporter-mediated hepatic uptake as a mechanism for drug–drug Research, 31, 6283–6289. interaction between cerivastatin and cyclosporin A. Journal of Trocho, C., Escola-Gil, J. C., Ribas, V., Benitez, S., Martin-Campos, J. M., Pharmacology and Experimental Therapeutics, 304, 610–616. Rotllan, N., et al. (2004). Phenytoin treatment reduces atherosclerosis in Somogyi, R., Fuhrman, S., & Wen, X. (2001). Genetic network inference in mice through mechanisms independent of plasma HDL-cholesterol computational models and applications to large-scale gene expression concentration. Atherosclerosis, 174, 275–285. data. In M. Bower, & H. Bolouri (Eds.), Computational modeling of Ueda, A., Hamadeh, H. K., Webb, H. K., Yamamoto, Y., Sueyoshi, T., genetic and biochemical networks (pp. 119–157). Cambridge, MA’ The Afshari, C. A., et al. (2002). Diverse roles of the nuclear orphan MIT Press. receptor CAR in regulating hepatic genes in response to phenobarbital. Spirin, V., & Mirny, L. A. (2003). Protein complexes and functional Molecular Pharmacology, 61,1–6. modules in molecular networks. Proceedings of the National Academy Ulrich, R. G. (2003). The toxicogenomics of nuclear receptor agonists. of Sciences of the United States of America, 100, 12123–12128. Current Opinion in Chemical Biology, 7, 505–510. Staudinger, J., Liu, Y., Madan, A., Habeebu, S., & Klaassen, C. D. Ulrich, R. G., Rockett, J. C., Gibson, G. G., & Pettit, S. D. (2004). (2001). Coordinate regulation of xenobiotic and bile acid homeo- Overview of an interlaboratory collaboration on evaluating the effects stasis by pregnane X receptor. Drug Metabolism and Disposition, of model hepatotoxicants on hepatic gene expression. Environmental 29, 1467–1472. Health Perspectives, 112, 423–427. Sun, L. Z., Ji, Z. L., Chen, X., Wang, J. F., & Chen, Y. Z. (2002). ADME- Unami, A., Shinohara, Y., Ichikawa, T., & Baba, Y. (2003). Biochemical AP: A database of ADME associated proteins. Bioinformatics, 18, and microarray analyses of bupivacaine-induced apoptosis. Journal of 1699–1700. Toxicological Sciences, 28, 77–94. Suter, L., Babiss, L. E., & Wheeldon, E. B. (2004). Toxicogenomics in Vasquez, A., Flammini, A., Maritan, A., & Vespignani, A. (2003). Global predictive toxicology in drug development. Chemistry and Biology, 11, protein function prediction from protein–protein interaction networks. 161–171. Nature Biotechnology, 21, 697–700. Synold, T. W., Dussault, I., & Forman, B. M. (2001). The orphan nuclear Vezina, C. M., Walker, N. J., & Olson, J. R. (2004). Subchronic Exposure to receptor SXR coordinately regulates drug metabolism and efflux. TCDD, PeCDF, PCB126, and PCB153: Effect on hepatic gene Nature Medicine, 7, 584–590. expression. Environmental Health Perspectives, 112, 1636–1644. Tabuchi, Y., & Kondo, T. (2003). cDNA microarray analysis reveals chop- Vo, T. D., Greenberg, H. J., & Palsson, B. O. (2004). Reconstruction and 10 plays a key role in Sertoli cell injury induced by bisphenol A. functional characterization of the human mitochondrial metabolic Biochemical and Biophysical Research Communications, 305, 54–61. network based on proteomic and biochemical data. Journal of Bio- Tamai, I., Nezu, J., Uchino, H., Sai, Y., Oku, A., Shimane, M., et al. (2000). logical Chemistry, 279, 39532–39540. Molecular identification and characterization of novel members of the Wang, H., & LeCluyse, E. L. (2003). Role of orphan nuclear receptors in human organic anion transporter (OATP) family. Biochemical and the regulation of drug-metabolising enzymes. Clinical Pharmacoki- Biophysical Research Communications, 273, 251–260. netics, 42, 1331–1357. Taylor, S. W., Fahy, E., Zhang, B., Glenn, G. M., Warnock, D. E., Wiley, S., Wang, N., Silver, D. L., Thiele, C., & Tall, A. R. (2001). ATP- et al. (2003). Characterization of the human heart mitochondrial binding cassette transporter A1 (ABCA1) functions as a choles- proteome. Nature Biotechnology, 21, 281–286. terol efflux regulatory protein. Journal of Biological Chemistry, 276, Terasaka, S., Aita, Y., Inoue, A., Hayashi, S., Nishigaki, M., Aoyagi, K., 23742–23747. et al. (2004). Using a customized DNA microarray for expression Waring, J. F., Cavet, G., Jolly, R. A., McDowell, J., Dai, H., Ciurlionis, R., profiling of the estrogen-responsive genes to evaluate estrogen activity et al. (2003). Development of a DNA Microarray for toxicology based among natural estrogens and industrial chemicals. Environmental on hepatoxin-regulated sequences. Environmental Health Perspectives, Health Perspectives, 112, 773–781. 111, 863–870. 66 S. Ekins / Journal of Pharmacological and Toxicological Methods 53 (2006) 38–66

Waring, J. F., Gum, R., Morfitt, D., Jolly, R. A., Ciurlionis, R., Heindel, M., Yamashita, S., Nomoto, T., Abe, M., Tatematsu, M., Sugimura, T., & et al. (2002). Identifying toxic mechanisms using DNA microarrays: Ushijima, T. (2004). Persistence of gene expression changes in stomach Evidence that an experimental inhibitor of cell adhesion molecule mucosae induced by short-term N-methyl-N-nitro-V N-nitrosoguanidine expression signals through the aryl hydrocarbon nuclear receptor. treatment and their presence in stomach cancers. Mutation Research, Toxicology, 181–182, 537–550. 549, 185–193. Waring, J. F., Jolly, R. A., Ciurlionis, R., Lum, P. Y., Praestgaard, J. T., Yan, Q., & Sadee, W. (2000). Human membrane transporter database: A Morfitt, D. C., et al. (2001). Clustering of hepatotoxins based on web-accessible relational database for drug transport studies and mechanism of toxicity using gene expression profiles. Toxicology and pharmacogenomics. AAPS Pharmsci, 2, E20. Applied Pharmacology, 175, 28–42. Yeger-Lotem, E., & Margalit, H. (2003). Detection of regulatory circuits by Waters, M., Boorman, G., Bushel, P., Cunningham, M., Irwin, R., Merrick, integrating the cellular networks of protein–protein interactions and A., et al. (2003). Systems toxicology and the Chemical Effects in transcription regulation. Nucleic Acids Research, 31, 6053–6061. Biological Systems (CEBS) knowledge base. EHP Toxicogenomics, Yoon, B. I., Li, G. -X., Kitada, K., Kawasaki, Y., Igarashi, K., & Kodama, 111, 15–28. Y., et al. (2003). Mechanisms of benzene-induced hematotoxicity and Wei, M., Arnold, L., Cano, M., & Cohen, S. M. (2005). Effects of co- leukemogenicity: cDNA microarray analyses using mouse bone marrow administration of antioxidants and arsenicals on the rat urinary bladder tissue. Environmental Health Perspectives, 111, 1411–1420. epithelium. Toxicology Sciences, 83, 237–245. Young, M. B., DiSilvestro, M. R., Sendera, T. J., Freund, J., Kriete, A., & Weinshilboum, R. (2003). Inheritance and drug response. New England Magnuson, S. R. (2003). Analysis of gene expression in carbon Journal of Medicine, 348, 529–537. tetrachloride-treated rat livers using a novel bioarray technology. Weinshilboum, R., & Wang, L. (2004). Pharmacogenetics: Inherited Pharmacogenomics Journal, 3, 41–52. variation in amino acid sequence and altered protein quantity. Clinical Yu, H., Zhu, X., Greenbaum, D., Karro, J., & Gerstein, M. (2004). TopNet: Pharmacology and Therapeutics, 75, 253–258. A tool for comparing biological sub-networks, correlating protein Werner, E. (2003). In silico multicellular systems biology and minimal properties with topological statistics. Nucleic Acids Research, 32, genomes. Drug Discovery Today, 8, 1121–1127. 328–337. Wong, J. S., & Gill, S. S. (2002). Gene expression changes induced in Zhang, E. Y., Fu, D. J., Pak, Y. A., Stewart, T., Mukhopadhyay, N., mouse liver by di(2-ethylhexyl) phthalate. Toxicology and Applied Wrighton, S. A., et al. (2004). Genetic polymorphisms in human proton- Pharmacology, 185, 180–196. dependent dipeptide transporter PEPT1: Implications for the functional Xie, T., Tong, L., McCann, U. D., Yuan, J., Becker, K. G., Mechan, A. O., role of Pro586. Journal of Pharmacology and Experimental Therapeu- et al. (2004). Identification and characterization of metallothionein-1 tics, 310, 437–445. and -2 gene expression in the context of ()3,4-methylenedioxymetham- Zhang, E. Y., Knipp, G. T., Ekins, S., & Swaan, P. W. (2002). phetamine-induced toxicity to brain dopaminergic neurons. Journal of Structural biology and function of solute transporters: Implications Neuroscience, 24, 7043–7050. for identifying and designing substrates. Drug Metabolism Reviews, Yadetie, F., Laegreid, A., Bakke, I., Kusnierczyk, W., Komorowski, J., 34, 709–750. Waldum, H. L., et al. (2003). Liver gene expression in rats in response Zhang, E. Y., Phelps, M. A., Cheng, C., Ekins, S., & Swaan, P. W. (2002). to the peroxisome proliferator-activated receptor-alpha agonist ciprofi- Modeling of active transport systems. Advanced Drug Delivery brate. Physiological Genomics, 15, 9–19. Reviews, 54, 329–354.