Assessment of candidate biomarkers of drug-induced hepatobiliary injury in preclinical toxicity studies

Toxicology Letters 196 (2010) 1–11

Contents lists available at ScienceDirect

Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet

Assessment of candidate biomarkers of drug-induced hepatobiliary injury in preclinical toxicity studies M. Adler a , D. Hoffmann a , H. Ellinger-Ziegelbauer b , P. Hewitt c , K. Matheis d , L. Mulrane e , W.M. Gallagher e , J.J. Callanan e,f , L. Suter g , M.M. Fountoulakis g , W. Dekant a , A. Mally a,∗ a

University of Würzburg, Würzburg, Germany Bayer Schering Pharma AG, Wuppertal, Germany Merck KGaA, Darmstadt, Germany d Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach, Germany e UCD School of Biomolecular and Biomedical Science, UCD Conway Institute, University College Dublin, Dublin, Ireland f School of Agriculture, Food Science & Veterinary Medicine, UCD Conway Institute, University College Dublin, Dublin, Ireland g Hoffmann-La Roche AG, Basel, Switzerland b c

a r t i c l e

i n f o

Article history: Received 2 February 2010 Received in revised form 18 March 2010 Accepted 24 March 2010 Available online 1 April 2010 Keywords: Preclinical safety assessment Liver Hepatotoxicity Biomarker NGAL Thiostatin

a b s t r a c t This study was designed to assess the value of a set of potential markers for improved detection of liver injury in preclinical toxicity studies. Male Wistar rats were treated with drug candidates (BAY16, EMD335823, BI-3) that previously failed during development, in part due to hepatotoxicity, at two dose levels for 1, 3 and 14 days. Concentrations of lipocalin-2/NGAL and clusterin, which are frequently overexpressed and released from damaged tissues, and thiostatin, recently identified within PredTox as being elevated in urine in response to liver injury, were determined in rat urine and serum by ELISA. This was supplemented by confirmatory qRT-PCR and immunohistochemical analyses in the target organ. Serum paraoxonase-1 activity (PON1), which has been suggested as a marker of hepatotoxicity, was determined using a fluorometric assay. Clusterin and PON1 were not consistently altered in response to liver injury. In contrast, thiostatin and NGAL were increased in serum and urine of treated animals in a time- and dose-dependent manner. These changes correlated well with mRNA expression in the target organ and generally reflected the onset and degree of drug-induced liver injury. Receiver–operating characteristics (ROC) analyses supported serum thiostatin, but not NGAL, as a better indicator of drug-induced hepatobiliary injury than conventional clinical chemistry parameters, i.e. ALP, ALT and AST. Although thiostatin, an acute phase protein expressed in a range of tissues, may not be specific for liver injury, our results indicate that thiostatin may serve as a sensitive, minimally-invasive diagnostic marker of inflammation and tissue damage in preclinical safety assessment. © 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Drug-induced liver toxicity is a major problem during drug development and the main cause for drug withdrawals from the market (Ostapowicz and Lee, 2000). For more than 60 years, assessment of drug hepatotoxicity and differentiation between various types of hepatic injury (e.g. hepatocellular vs.

Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; DILI, drug-induced liver injury; GGT, ␥glutamyltransferase; NGAL, neutrophil gelatinase-associated lipocalin; PON1, paraoxonase; 2D-PAGE, two-dimensional polyacrylamide gel electrophoresis; TMAs, tissue microarrays. ∗ Corresponding author at: Department of Toxicology, University of Würzburg, Versbacher Str. 9, 97078 Würzburg, Germany. Tel.: +49 931 20148894; fax: +49 931 20148865. E-mail address: [email protected] (A. Mally). 0378-4274/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2010.03.018

cholestatic) has been based primarily on determination of hepatic enzymes in blood, most notably alanine aminotransferase (ALT)/aspartate aminotransferase (AST) and alkaline phosphate (ALP)/␥-glutamyltranspeptidase (GGT) as indicators of hepatocellular and cholestatic liver injury, respectively. However, elevation of blood “hepatic” enzymes may be due to pathological conditions other than liver damage, such as muscle injury (Nathwani et al., 2005). Moreover, abnormal enzyme levels reveal only severe tissue injury and may not represent sensitive indicators of subtle toxic change. Hence, the need for identification and validation of additional novel biomarkers of hepatotoxicity with the ability to detect early signs of liver injury to improve safety assessment of drug candidates in preclinical studies has long been recognized. However, despite application of omics profiling strategies, including transcriptomics, proteomics and metabolomics and world-wide efforts in the field of biomarker discovery, only a few, novel candidate markers of hepatotoxicity have emerged. These

2

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

include paraoxonase (PON1), an antioxidant enzyme implicated in detoxification of organophosphates and prevention of low-density lipoprotein (LDL) oxidation, which has been shown to be modulated in serum and plasma in response to hepatic damage (Ferre et al., 2006), and clusterin, which was recently identified as part of a gene expression signature predictive of acute liver injury (Zidek et al., 2007). Importantly, clusterin is a glycoprotein synthesized and released by a variety of cells and tissues in response to injury, suggesting that determination of clusterin in body fluids may represent a sensitive non-invasive means to detect tissue injury, including hepatotoxicity. With the overall aim of assessing the value of combining results from omics technologies together with results from conventional toxicology methods for more informed decision making in preclinical safety evaluation, the Innomed PredTox consortium (www.innomed-predtox.com) extensively characterized and integrated omics responses to 14 drug candidates which previously failed during non-clinical development, in part due to hepatotoxic and/or nephrotoxic effects to obtain mechanistic insight and identify potential biomarkers (Gallagher et al., 2009). Similar to clusterin, transcriptomics results from these studies indicated that neutrophil gelatinase-associated lipocalin (NGAL), a 25 kDa secreted glycoprotein, which has been suggested as a urinary marker of renal injury (Mishra et al., 2006; Mitsnefes et al., 2007), was consistently upregulated in livers of rats following treatment with compounds associated with hepatotoxicity (unpublished data). In addition, proteomics based on two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) demonstrated an increase in urinary thiostatin protein concentrations in dosed animals compared to controls after treatment with two hepatotoxic drug candidates randomly selected for these analyses (unpublished data). The present study was conducted as part of the PredTox project to assess the performance of PON1, clusterin, NGAL, thiostatin as potential non-invasive markers of drug-induced liver toxicity, as compared to routine clinical chemistry parameters. Specifically, we were interested to determine the ability of these biomarker candidates to detect or even predict hepatotoxicity in the form of hepatocellular necrosis and/or bile duct damage induced by three drug candidates. Thus, serum PON1 activity and release of clusterin, NGAL and thiostatin into serum or urine was analyzed following 1, 3, and 14 days treatment of male Wistar rats with low and a high doses of the glucagon receptor antagonist BAY16 ((+)-(1R)-1-[4-(4-fluorophenyl)-2,6-diisopropyl-5-propylpyridin-3-yl]ethanol), the aldose reductase inhibitor EMD335823 (1-(2-trifluoromethoxyphenyl)-2-nitroethanone) and BI-3 (3-pyrrolidineacetic acid, 5-[[[4 -[imino[(methoxycarbonyl) amino]methyl][1,1 -biphenyl]-4-yl]oxy]methyl]-2-oxo-, methylester(3S-trans)) These analyses were supplemented by gene/protein expression and localizations studies in the target organ to link biomarker concentrations in serum/urine to their expression/localization in the liver to provide anchoring to the target organ and site of injury. 2. Materials and methods 2.1. Reagents Primary antibodies used were goat anti-rat lipocalin-2/NGAL (R&D Systems GmbH, Wiesbaden-Nordenstadt, Germany), mouse anti-NGAL (ABS 039-08, ABS039-32B; AntibodyShop A/S, Gentofte, Denmark), goat anti-clusterin ␣ (Santa Cruz Biotechnology, Heidelberg, Germany), chicken anti-thiostatin (Immunology Consultants Laboratory, Inc., Dunn Labortechnik GmbH, Asbach, Germany), rabbit anti-GAPDH (Sigma–Aldrich, Taufkirchen, Germany). Secondary antibodies were obtained from Santa Cruz Biotechnology, and VECTASTAIN Elite ABC Kit and DAB substrate kit were purchased from Vector Laboratories (Burlingame, USA). Recombinant rat lipocalin-2/NGAL protein was purchased from R&D Systems GmbH (Wiesbaden-Nordenstadt, Germany). Unless otherwise indicated, all other chemicals were from Roth (Karlsruhe, Germany).

Table 1 Treatment schedule of animals. Animal identification

Treatment

Total no. of doses

01–05 06–10 11–15 16–20 21–25 26–30 31–35 36–40 41–45

Vehicle Low dose High dose Vehicle Low dose High dose Vehicle Low dose High dose

1 1 1 3 3 3 14 14 14

Note: Urine was collected from animals nos. 31–45 after 1, 3 and 12 days of dosing

2.2. Animal studies All animal studies were conducted in accordance to European or national animal welfare regulations using a harmonized study protocol as previously described (Mulrane et al., 2008). SPF-bred male Wistar rats (8–10 weeks old, weighing 170–200 g), which are most commonly used in preclinical safety studies by European Pharmaceutical Industries, were obtained from the standard supplier of each participating company, i.e. Harlan Winkelmann, Borchen, Germany (BAY16 and EMD334823: Hsd Cpb:WU) or Charles River Laboratories Italia S.r.l., Calco, Italy (BI-3: Crl: Wistar (Han)). Only males were used as previous studies by the participating companies had not indicated any sex differences in the susceptibility to the drugs employed. Rats were distributed into 3 dose groups (n = 5 per group and time-point) and dosed with BAY16 ((+)-(1R)-1-[4-(4-fluorophenyl)-2,6-diisopropyl-5-propyl-pyridin-3yl]ethanol) (0, 20 and 100 mg/kg/day), EMD335823 (1-(2-trifluoromethoxyphenyl)2-nitroethanone) (0, 15 and 350 mg/kg/day) or BI-3 [3-pyrrolidineacetic acid, 5[[[4 -[imino[(methoxycarbonyl) amino]methyl] [1,1 -biphenyl]-4-yl]oxy]methyl]2-oxo-, methylester(3S-trans)] (0, 100 and 1000 mg/kg/day) by oral gavage for 1, 3 or 14 days, followed by necropsy after an overnight fasting period (Table 1). After 1, 3 and 12 days of dosing, 16 h urine samples were collected from animals treated for a total of 14 days, aliquoted and stored at −80 ◦ C until use. Blood samples were collected from all animals at the time of necropsy. Blood or serum, and urine samples were used to measure routine hematology and clinical chemistry parameters according to standard operating procedures at the participating companies, as described in detail elsewhere (Hoffmann et al., 2010). At necropsy, organs (liver, kidney) were removed, aliquoted, fixed in formalin or flash frozen in liquid nitrogen and stored at −80 ◦ C. Sections (3–4 mm) were fixed in 10% phosphate buffered formalin and subsequently embedded in paraffin blocks, sectioned, stained with hematoxylin and eosin and examined by light microscopy by the respective study pathologist. Histopathological findings were confirmed by an independent veterinary pathologist (JJC). Based on the peer review, histopathological findings involving hepatocyte cell death/necrosis, bile duct mitosis/hyperplasia and bile duct inflammation were summarized into a single histopathology score, whereby 0 ↔ lesion not observed; + ↔ minimal, ++ ↔ mild, +++ ↔ moderate and ++++ ↔ high severity of lesion. 2.3. Measurement of clusterin, NGAL, and thiostatin protein and PON1 activity in serum and urine PON1 enzyme activity in serum was determined by a fluorometric assay (Invitrogen, Molecular Probes, Karlsruhe, Germany) based on enzymatic hydrolysis of a fluorogenic organophosphate analog by PON1 (excitation/emission maxima 360/450 nm). Fluorescence intensity was measured using a Berthold Mithras LB940 fluorescence plate reader (Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany). Urinary and serum clusterin was assayed by the Rat Clusterin ELISA Kit (Biovendor Laboratory Medicine Inc., Heidelberg, Germany). Quantification of urinary and serum thiostatin was performed using a commercially available anti-rat sandwich ELISA according to the manufacturer’s instructions (ICL Inc., Dunn Labortechnik GmbH, Asbach, Germany). The absorbance was measured by a microplate reader (SpectraMax 190, Molecular Devices) at 450 nm. Results from ELISA assays were calculated using standard curves fitted with a four-parameter logistic regression and expressed as ␮g/ml serum or relative to urinary creatinine (ng/mg creatinine). Quantification of rat NGAL in urine and serum was performed by an in-house sandwich ELISA using two monoclonal antibodies (BioPorto Diagnostics, Gentofte, Denmark) as previously described (Hoffmann et al., 2010; Sieber et al., 2009). 2.4. Reverse transcription and real-time PCR Total RNA was isolated from frozen livers and kidneys using the RNeasy Mini Kit (Qiagen, Hilden, Germany) including DNA digestion according to the manufacturer’s instructions. Complementary DNA synthesis was synthesized from 1 ␮g total RNA using the VersoTM cDNA Kit (Thermo Fisher Scientific, Hamburg, Germany). Real-time PCR was carried out on a Roche LightCycler480

+++ (5/5) + (1/5), ++ (4/5) + (5/5) + (1/5), ++++ (3/5) – –

++ (1/5) + (5/5) – ++++ (2/5) ++++ (3/5) + (2/5), +++ (2/5) – –

– + (1/5)

– – + (2/5), ++ (1/5), +++ (2/5) – ++++ (5/5) – –

Hepatocyte cell death Bile duct inflammation Bile duct epithelial cell mitosis/hyperplasia Summary score

+ (1/5) – –

++++ (5/5)

+ (2/5) +++ (4/5) ++ (1/5), +++ (4/5) + (1/5), ++ (4/5) + (1/5)

++ (1/5), +++ (4/5) + (5/5) + (5/5) – – + (3/5) + (3/5)

–

+ (5/5) – + (5/5) – – – – – – – + (1/5), ++ (3/5) ++ (2/5), +++ (3/5) +++ (5/5) + (3/5) – – + (3/5) – –

+++ (5/5)

++ (2/5), +++ (3/5) ++ (1/5), +++ (4/5)

+ (1/5), ++ (2/5), +++ (2/5) + (4/5), +++ (1/5) + (5/5) – – – + (2/5), ++ (2/5) + (4/5) + (2/5)

1000 mg/kg

+ (2/5), ++ (3/5) + (1/5), ++ (2/5), +++ (2/5) ++ (3/5) ++ (1/5) + (4/5), +++ (1/5) + (1/5)

100 mg/kg 0 mg/kg

– + (5/5) – – – –

350 mg/kg 15 mg/kg

– – – – – –

0 mg/kg 100 mg/kg

– + (2/5), ++ (2/5) + (4/5) + (4/5) – –

20 mg/kg

+ (2/5) – –

BI-3

14

Major histopathological findings in liver were restricted to rats treated with a high doses of BAY16, BI-3 or EMD335823, although slight to moderate effects were also seen in individual animals after treatment with a low dose of BI-3 (Table 2). Initial changes seen after administration of BAY16 consisted of acute necrosis of the bile duct epithelium, bile duct inflammation/hyperplasia with increasing severity with time, followed by focal/multifocal hepatocyte necrosis and moderate periportal fibrosis by day 14. Although single cell death was occasionally noted

Hepatocyte cell death Bile duct inflammation Bile duct epithelial cell mitosis/hyperplasia Summary score

3.1. General toxicity, histopathology and clinical chemistry

3

3. Results

0 mg/kg

Data are presented as individual animals or mean ± standard deviation (SD). Statistical analyses were conducted by ANOVA followed by Dunnett’s post hoc test. Values significantly different from control are indicated as *p < 0.05, **p < 0.01, and ***p < 0.001. Receiver–operating characteristics (ROC) curves were plotted by entering data obtained from animals with a histopathology score >1 vs. control animals and animals without histopathological evidence of liver injury using the Graph Pad Prism 5 software package (GraphPad Software Inc., La Jolla, CA).

Hepatocyte cell death Bile duct inflammation Bile duct epithelial cell mitosis/hyperplasia Summary score

2.7. Statistical analyses

1

Livers were homogenized in lysis buffer containing RIPA-buffer (50 mM Tris–HCl pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% NP-40, 0.1% SDS), 1 mM NaF, 1 mM Na3 VO4 and protease inhibitor cocktail (Sigma–Aldrich) and centrifuged at 8000 × g for 5 min at 4 ◦ C. The protein concentration of the lysate was determined using the DC Assay method (Biorad Laboratories GmbH, München, Germany). Aliquots of 10 ␮g protein in Laemmli buffer (Sigma–Aldrich) were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred onto nitrocellulose membranes at 100 V for 75 min at 4 ◦ C using transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol). Membranes were blocked in 5% non-fat dry milk in TBST buffer (50 mM Tris–HCl, 150 mM NaCl, 0.1% Tween-20, pH 7.4) for 1 h at RT following incubation with chicken anti-thiostatin diluted 1:5000 in blocking buffer overnight at 4 ◦ C or with rabbit antiGAPDH diluted 1:20,000 for 1 h at RT. After several washing with TBST, horseradish peroxidase (HRP) conjugated antibodies were incubated at a dilution of 1:5000 in blocking buffer for 1 h at RT and detection of proteins was performed using ECL detection system (GE Healthcare, München, Germany). Densitometry analysis was performed using the Bio-Rad Gel Doc 2000 system.

EMD335823

2.6. Western blot analysis of thiostatin

BAY16

Liver and kidney sections (5 ␮m) were prepared from tissue microarray blocks (Mulrane et al., 2008) and mounted onto glass slides. Tissues were deparaffinized, rehydrated and washed in PBS. Heat-induced antigen retrieval was achieved by 4 min autoclaving in 10 mM citrate buffer, pH 6.0. For detection of NGAL, tissues were then treated with 0.1% trypsin for 2 min at 37 ◦ C. After washing in PBS, sections were blocked with 5% donkey serum for 1 h. Endogenous peroxidase was subsequently blocked by incubation with 3% H2 O2 in PBS for 10 min, followed by two additional blocking steps with 0.001% avidin/PBS and 0.001% biotin/PBS for 15 min, respectively. Between all steps, sections were washed several times with PBS. Slides were incubated with 1.3 ␮g/ml goat anti-NGAL or 2 ␮g/ml goat anticlusterin ␣ diluted in 5% donkey serum in PBS overnight at 4 ◦ C in a humidified chamber. After three washes, tissues were incubated with a biotinylated secondary antibody (donkey anti-goat; Santa Cruz Biotechnology) diluted 1:200 in PBS for 1 h at RT and subsequently washed with PBS. Immunohistochemical staining was performed using VECTASTAIN Elite ABC Kit (Vector Laboratories, Burlingame, USA). Tissues were counterstained with hematoxylin, dehydrated and mounted in Eukitt mounting medium (Sigma–Aldrich, Taufkirchen, Germany).

Histopathological change

2.5. Immunohistochemical staining of NGAL and clusterin

3

Days of treatment

(Roche, Mannheim, Germany) using the following cycling conditions: 95 ◦ C for 15 min, 45 cycles of 95 ◦ C for 15 s, 60 ◦ C for 30 s, and 72 ◦ C for 30 s. The reaction mixture (20 ␮l) consisted of 2× mastermix with SYBR Green I (Thermo Fisher Scientific, Hamburg, Germany), 2 ␮l cDNA and 105 nM of each primer. QRT-PCR primers designed for NGAL and clusterin were as previously described (Rached et al., 2008). Other primers used were as follows: T-kininogen: forward 5 -TTCACTACTTCCCAAGAAATGCT-3 and reverse 5 -AGCAACCACCTGTGATGTTG-3 (Hillmeister et al., 2008), PON1: forward 5 -GGGAAATACGTTGGATATGTC-3 and reverse 5 -ATATCGTTGATGCTAGGCAG-3 , 18S rRNA: forward 5 GTAACCCGTTGAACCCCATT-3 and reverse 5 -CCATCCAATCGGTAGTAGCG-3 . All samples were measured in duplicate and normalized against 18S rRNA. The relative quantification of gene expression was determined by the Ct method, and data are presented as mean fold mRNA expression change relative to controls (n = 5).

Table 2 Summary of histopathological changes in livers of male Wistar rats treated with BAY16, EMD335823 or BI-3 for 1, 3 and 14 days. Abbrevations: – ↔ lesion not observed; + ↔ minimal, ++ ↔ mild, +++ ↔ moderate and ++++ ↔ high severity of lesion.

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

4

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

Table 3 Summary of serum clinical chemistry data obtained from male Wistar rats treated with BAY16, EMD335823 and BI-3 for up to 14 days. Data are presented as mean ± standard deviation (n = 5). Days of treatment

BAY16

BI-3a

EMD335823

0 mg/kg

20 mg/kg

100 mg/kg

0 mg/kg

15 mg/kg

350 mg/kg

0.24 ± 0.01 0.25 ± 0.01 0.28 ± 0.02

0.24 ± 0.02 0.23 ± 0.02 0.28 ± 0.02

0 mg/kg

100 mg/kg

1000 mg/kg

0.30 ± 0.01 0.51 ± 0.03 0.30 ± 0.02*** 0.52 ± 0.04 0.38 ± 0.14 0.52 ± 0.07

0.46 ± 0.03 0.45 ± 0.04* 0.53 ± 0.05

0.47 ± 0.03 0.48 ± 0.04 0.65 ± 0.32

49 ± 6 50 ± 7 45 ± 6

53 ± 17 48 ± 7 60 ± 6

46 ± 9 67 ± 19 110 ± 72

Creatinine [mg/dl]

1 3 14

0.58 ± 0.03 0.57 ± 0.03 0.64 ± 0.02

0.57 ± 0.02 0.54 ± 0.02 0.60 ± 0.02

0.57 ± 0.03 0.63 ± 0.03 0.63 ± 0.04

Blood urea nitogen (BUN) [mg/dl]

1 3 14

39 ± 7 40 ± 5 34 ± 19

36 ± 4 38 ± 2 39 ± 3

32 ± 5 39 ± 5 47 ± 5

72 ± 7 83 ± 16 90 ± 20

73 ± 5 90 ± 14 107 ± 11

87 ± 12* 71 ± 9 106 ± 34

1.6 ± 1.1 3.0 ± 3.2 1.0 ± 0.0

2.0 ± 0.0 0.8 ± 1.5 2.6 ± 2.1

3.0 ± 2.7 10.6 ± 6.0* 39.0 ± 24.5*

0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

0.0 ± 0.0 0.0 ± 0.0 37 ± 61

0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0

1.7 ± 1.4** 1.6 ± 2.4 1.6 ± 1.3**

60 ± 2 61 ± 1 60 ± 1

58 ± 1 60 ± 2 57 ± 2

60 ± 4 54 ± 2*** 57 ± 5

59 ± 1 59 ± 1 58 ± 2

58 ± 2 58 ± 3 55 ± 2

61 ± 2 56 ± 3 45 ± 6***

␥-Glutamyltransferase 1 3 (GGT) [U/l] 14

***

Total protein [g/l]

1 3 14

62 ± 2 60 ± 2 65 ± 1

60 ± 2 60 ± 2 66 ± 2

59 ± 2* 64 ± 2** 71 ± 1***

Aspartate aminotransferase (AST) [U/l]

1 3 14

84 ± 6 73 ± 21 63 ± 11

80 ± 20 90 ± 17 60 ± 3

141 ± 92 400 ±196** 90 ± 13**

119 ± 32 124 ± 13 106 ± 20

110 ± 11 115 ± 21 125 ± 13

120 ± 15 134 ± 35 293 ± 282

193 ± 37 190 ± 25 157 ± 19

167 ± 38 172 ± 22 151 ± 27

298 ± 195 215 ± 43 144 ± 40

Alkaline phosphatase (ALP) [U/l]

1 3 14

272 ± 41 246 ± 21 244 ± 26

281 ± 46 243 ± 39 231 ± 8

267 ± 62 433 ± 57*** 293 ± 39*

226 ± 59 158 ± 33 164 ± 41

221 ± 36 153 ± 30 155 ± 49

202 ± 36 146 ± 35 216 ± 90

182 ± 25 197 ± 47 114 ± 11

149 ± 32 177 ± 36 116 ± 13

513 ± 289* 425 ± 77*** 262 ± 90**

Alanine aminotransferase (ALT) [U/l]

1 3 14

51 ± 12 50 ± 11 65 ± 7

52 ± 5 50 ± 11 64 ± 3

77 ± 46 346 ± 108*** 101 ± 26**

70 ± 9 53 ± 10 51 ± 10

71 ± 10 56 ± 5 43 ± 7

64 ± 8 68 ± 15 220 ± 253

31 ± 5 24 ± 4 24 ± 4

28 ± 14 22 ± 2 28 ± 10

169 ± 185 43 ± 14** 24 ± 4

Bilirubin [mg/dl]

1 3 14

0.07 ± 0.02 0.10 ± 0.01 0.13 ± 0.01

0.05 ± 0.01 0.10 ± 0.01 0.11 ± 0.01

0.50 ± 0.79 3.53 ± 1.36*** 0.36 ± 0.05***

0.06 ± 0.01 0.10 ± 0.03 0.08 ± 0.02

0.06 ± 0.01 0.10 ± 0.02 0.08 ± 0.01

0.07 ± 0.02 0.08 ± 0.02 0.89 ± 1.10

Triglyce rides [mg/dl]

1 3 14

40 ± 14 45 ± 24 73 ± 17

49 ± 27 37 ± 13 76 ± 23

58 ± 27 157 ± 28*** 44 ± 9

164 ± 74 40 ± 8 48 ± 10

125 ± 43 48 ± 18 43 ± 14

Glucose [mg/dl]

1 3 14

112 ± 6 83 ± 30 95 ± 24

118 ± 22 95 ± 16 80 ± 5

a * ** ***

Not determined Not determined Not determined

54 ± 25* 46 ± 13 87 ± 83 123 ± 13 91 ± 11 82 ± 2

Not determined Not determined Not determined 64 ± 33 65.6 ± 25 87.4 ± 22

64 ± 22 79 ± 27 86 ± 6

73.8 ± 11 36 ± 15 51 ± 18

147 ± 11 144 ± 20 148 ± 14

132 ± 23 143 ± 8 163 ± 4

152 ± 20 122 ± 6* 128 ± 47

Hoffmann et al. (2010). Statistical analysis was performed by ANOVA and Dunnett’s test p < 0.05. Statistical analysis was performed by ANOVA and Dunnett’s test p < 0.01. Statistical analysis was performed by ANOVA and Dunnett’s test p < 0.001.

in some of the low dose animals, similar observations were made in control animals and thus these changes were considered as background and not compound-related. Liver enzyme activities of AST, ALT, ALP, GGT as well as bilirubin levels were significantly increased as early as day 3 (Table 3). In rats treated with BAY16 at 100 mg/kg bw (animal nos. 41–45), reduced body weight relative to controls (12% by the end of treatment) accompanied by reduced food intake was seen. During the course of the study, weight loss and clinical signs of toxicity, i.e. ventral recumbency and scabs on the nose or genital region, were observed in individual animals (nos. 42, 44, 45) administered EMD335823 at 350 mg/kg bw. Histopathological lesions in livers of rats dosed with EMD335823 were seen in the same three animals after 14 days of dosing and were characterized by massive multifocal liver cell necrosis and moderate to high severity of bile duct damage (Table 2). In good agreement with the severity of hepatic lesions which were most prominent in rats nos. 42, 45, alterations in conventional clinical chemistry parameters (ALT, AST, GGT, ALP, bilirubin) indicative of hepatotoxicity were restricted to these two animals (Table 3). In addition, an increase in serum creatinine was observed in the high dose group at all timepoints but no signs of kidney injury were seen by histopathology. Hepatotoxic effects in the form of hepatocyte necrosis, bile duct inflammation and hyperplasia were observed in individual animals at all time-points in response to BI-3 (Table 2). Slight to moderate pericholangitis associated with bile duct hyperplasia occurred in

almost all animals and distribution and severity of hepatocellular vacuolation increased with time. Measurement of liver enzymes in serum showed elevated level of AST, ALT and ALP with maximal effects occurring in rats with histopathological changes (Table 3). Besides liver injury, proximal tubular damage was also observed in Table 4 Enzymatic activity of PON1 in serum of rats treated with 3 hepatotoxic drug candidates (BAY16, EMD335823, BI-3) for 1, 3 and 14 days (C = control, LD = low dose, HD = high dose). PON1 activities are presented as mean ± standard deviation (n = 5). Compound

Days of treatment

Serum PON1 Activity [Unit/ml] C

LD

HD

BAY16

1 3 14

17.8 ± 6.6 13.5 ± 6.9 19.5 ± 4.2

18.6 ± 6.9 20.1 ± 7.5 19.1 ± 6.5

13.3 ± 4.1 24.7 ± 9.6 12.6 ± 2.0

EMD335823

1 3 14

31.2 ± 4.7 36.4 ± 6.4 27.0 ± 1.7

31.3 ± 4.2 29.1 ± 2.2 27.7 ± 4.4

29.7 ± 3.2 8.1 ± 0.6a , *** 18.7 ± 6.3*

BI-3

1 3 14

30.3 ± 4.5 45.3 ± 19.9 43.4 ± 25.4

40.6 ± 14.1 25.4 ± 4.6 19.9 ± 5.2

62.2 ± 23.9* 27.3 ± 18.9 10.9 ± 3.8**

a * ** ***

n = 2. p < 0.05 (ANOVA + Dunnett’s post hoc test). p < 0.01 (ANOVA + Dunnett’s post hoc test). p < 0.001 (ANOVA + Dunnett’s post hoc test).

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

kidneys of high dose rats (nos. 41, 43) (Hoffmann et al., 2010). A marked decrease in body weight (20% compared to controls) associated with reduced food intake was recorded in high dose animals (1000 mg/kg bw) at the end of the 14-day dosing period. 3.2. Serum PON1 activity and hepatic expression Previous studies suggested that serum PON1 activity may be reduced in response to liver injury and may, therefore, serve as an indicator of hepatotoxicity (Amacher et al., 2005; Feingold et al., 1998; Ferre et al., 2006). In line with these studies, PON1 activity was significantly reduced in rats after treatment with a high dose of EMD335823 for 3 and 14 days (Table 4). Similarly, a dosedependent decrease in PON1 activity was found in serum of rats treated with BI-3 for 14 days, whereas PON1 activity appeared to increase after single administration of a high dose of BI-3. How-

5

ever, no significant differences in serum PON1 activity between controls and treated animals were observed in response to BAY16, despite the marked histopathological alterations observed in high dose animals after 3 and 14 days of dosing. Reduced PON1 activity in response to EMD335823 and BI-3 was not associated with downregulation of PON1 mRNA expression in the target organ (data not shown). 3.3. Performance of clusterin as a liver biomarker Consistent with the absence of histopathological alterations, no significant changes in serum clusterin concentrations were evident in response to low dose treatment with BAY16. In contrast, a marked increase in serum clusterin was observed in all high dose animals after 3 and 14 days (Fig. 1a), which correlated with clusterin mRNA and protein expression in the target organ (Fig. 1a

Fig. 1. Analysis of clusterin in serum, urine and liver of rats following treatment with BAY16 (a), EMD335823 (b) and BI-3 (c) for up to 14 days. Data are presented as individual animals, color coded according to histopathology scores for liver damage. Mean values of 5 individual animals per dose group are indicated by a black line. Statistical analysis was performed by ANOVA and Dunnett’s post hoc test (*p < 0.05, **p < 0.01, ***p < 0.001). Note: for urinary markers after 1 and 3 days of treatment, no corresponding histopathology readouts are available as histopathology at these time-points was performed from groups of rats treated in parallel. (d) Immunohistochemical analysis of clusterin on liver tissue microarrays following treatment with BAY16, EMD335823 and BI-3, demonstrating increased protein expression at target sites of toxicity. Increased clusterin expression was predominantly found within the bile duct epithelium (BAY16, EMD335823) or in hepatocytes (BI-3).

6

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

and d). Furthermore, immunolocalization of clusterin in liver sections demonstrated expression of clusterin in biliary epithelial cells in areas associated with inflammatory and necrotic processes (Fig. 1d). Clusterin, a secretory heterodimeric disulphide-linked glycoprotein, may be cleaved into its alpha and beta chain, which may be able to pass the glomerulus (Shannan et al., 2006) and is also known to be released into urine in response to kidney injury (Hidaka et al., 2002; Sieber et al., 2009). Therefore, we were interested to determine if enhanced expression of clusterin at other sites, in this case the liver, is also associated with increased urinary concentrations of clusterin in addition to the rise in serum. Although analysis of clusterin in urine of rats treated with BAY16 revealed increased urinary levels in 2/5 high dose animals, alterations in the target organ or in serum were not well correlated by urinary concentrations of clusterin.

In contrast to BAY16, no clear effects on serum clusterin were observed after administration of EMD335823, despite an increase in hepatic or urinary clusterin in individual high dose animals (Fig. 1b). Immunohistochemical analysis revealed positive staining of clusterin in bile duct cells of treated rats 42 and 45. Elevated concentrations of clusterin were observed in urine (Hoffmann et al., 2010) and serum of individual animals in response to 1000 mg/kg bw BI-3 (Fig. 1c). Although, increased immunoreactivity of clusterin was found in liver sections of affected rats treated with BI-3 (Fig. 1d), no alterations in clusterin mRNA were observed in rat liver in contrast to the kidney (Hoffmann et al., 2010), suggesting that the kidney as a further target of BI-3 toxicity may also be a source of clusterin and that elevated serum or urinary clusterin may be associated with kidney injury rather than liver toxicity.

Fig. 2. Analysis of NGAL in serum, urine and liver of rats treated with BAY16 (a), EMD335823 (b) and BI-3 (c). Imunohistochemical staining of NGAL in liver sections of affected high dose animals after treatment with BAY16 and EMD335823 (d). Immunolocalization of NGAL in kidney sections of rats treated with BAY16 and EMD335823 (e), demonstrating distinct vesicular localization of NGAL in proximal tubule cells in the absence of kidney injury and significant up-regulation of NGAL mRNA in kidneys (Table 4), suggesting glomerular filtration and tubular reabsorption of NGAL released into serum at the site of injury. Data are presented as individual animals, color coded according to histopathology scores for liver damage. Mean values of 5 individual animals per dose group are indicated by a black line. Statistical analysis was performed by ANOVA and Dunnett’s post hoc test (*p < 0.05, **p < 0.01, ***p < 0.001). Note: for urinary markers after 1 and 3 days of treatment, no corresponding histopathology readouts are available as histopathology at these time-points was performed from groups of rats treated in parallel.

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

3.4. Evaluation of NGAL as a candidate biomarker of drug-induced liver injury Repeated administration of 100 mg/kg bw BAY16 resulted in a significant increase in serum NGAL protein, evident even 1 day after administration of a single dose (Fig. 2a). The release of NGAL into serum was accompanied by a pronounced increase in urinary NGAL protein after 3 and 12 days (>16-fold and 37-fold, respectively). In addition, serum concentrations of NGAL were found to correlate well with hepatic NGAL gene expression and with the severity of liver injury. Similar results were observed after exposure to EMD335823, whereby NGAL was dramatically increased in urine, serum and liver, specifically in those high dose animals (nos. 42 and 45), in which pronounced histopathological changes were seen (Fig. 2c). In both studies, immunohistochemical analysis in liver tissue microarrays revealed marked induction of NGAL in both hepatocytes and biliary epithelial cells in regions associated with prominent histopathological alterations (Fig. 2d). Following treatment with BI-3, concentrations of NGAL were found to be increased in urine (Hoffmann et al., 2010) and serum of individual animals at all time-points (Fig. 2c). Despite changes at the mRNA level, no clear alterations in protein expression were seen in liver after treatment with BI-3 (data not shown). However, it is important to note that treatment with BI-3 also induced nephrotoxic effects in the form of necrosis, inflammation and dilation of proximal tubules, and previous qRT-PCR analysis revealed up-regulation of NGAL in kidney (Hoffmann et al., 2010) in addi-

7

tion to the liver. This suggests that the increase of NGAL in urine and serum may originate from tissue injury in both organs. This is in contrast to BAY16 and EMD335823, where marked NGAL gene expression changes occurred exclusively in the liver as the target organ of toxicity (note: a fourfold change in kidney mRNA of animal no. 42 (EMD335823) vs. a 650-fold change in liver mRNA). Interestingly, however, strong positive staining for NGAL was evident in proximal tubule cells in kidney sections of animals showing marked liver injury after treatment with BAY16 and EMD335823 in the absence of histopathological changes in the kidney (Fig. 2e). Although staining intensity was much more pronounced, the pattern of NGAL immunoreactivity was similar to controls, with NGAL localizing to vesicles at the apical side of proximal tubules in the outer cortex, pointing towards a physiological role of the kidney in handling NGAL. Based on these findings, it appears that the marked increase of NGAL protein in kidney and urine of rats treated with BAY16 and EMD335823 is unrelated to kidney injury but may have occurred as a result of glomerular filtration and tubular reabsorption of NGAL produced at other sites of toxicity, such as the liver. 3.5. Association between liver injury and thiostatin in urine and serum Consistent with changes in urinary thiostatin previously detected by 2D-PAGE (unpublished data), ELISA analysis confirmed the presence of thiostatin in urine of rats treated with a high dose of EMD335823 or BI-3 for 14 days (Fig. 3b and c). Follow-

Fig. 3. Analysis of thiostatin (T-kininogen, alpha-1-MAP) in serum, urine and liver of Wistar rats after exposure to BAY16 (a), EMD335823 (b) and BI-3 (c). Data are presented as individual animals, color coded according to histopathology scores for liver damage. Mean values of 5 individual animals per dose group are indicated by a black line. Statistical analysis was performed by ANOVA and Dunnett’s post hoc test (*p < 0.05, **p < 0.01, ***p < 0.001). Note: for urinary markers after 1 and 3 days of treatment, no corresponding histopathology readouts are available as histopathology at these time-points was performed from groups of rats treated in parallel. Western blot analysis of thiostatin protein expression in livers of Wistar rats treated with 100 mg/kg bw BAY16, 350 mg/kg bw EMD335823 and 1000 mg/kg bw BI-3 for 14 days (n = 5), confirming increased thiostatin expression in all affected high dose animals as compared to untreated rats (d).

8

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

Table 5 Treatment-related changes in gene expression of candidate markers in kidneys of rats after administration of low and high dose of BAY16, EMD335823, BI-3. Data are presented as mean deregulation ± standard derivation compared to controls. Gene title

Days of treatment

Fold deregulation BAY16 [mg/kg bw]

EMD335823 [mg/kg bw]

BI-3 [mg/kg bw]

0

20

100

0

15

0

100

1000

NGAL

1 3 14

1.0 ± 0.2 1.0 ± 0.5 1.0 ± 0.4

1.0 ± 0.4 1.5 ± 0.6 0.6 ± 0.2

1.1 ± 0.3 2.0 ± 0.7 1.0 ± 0.4

1.0 ± 0.4 1.0 ± 0.3 1.0 ± 0.7

0.7 ± 0.2 0.7 ± 0.4 1.4 ± 0.3

0.7 ± 0.2 1.5 ± 0.4 1.9 ± 1.6a

1.0 ± 0.6 1.0 ± 0.4 1.0 ± 0.3

0.8 ± 0.7 1.0 ± 0.4 1.7 ± 0.3

0.7 ± 0.4d 1.5 ± 1.5d 81.6 ± 151.4c , d

T-kininogen

1 3 14

1.0 ± 0.4 1.0 ± 0.7 1.0 ± 0.9

1.4 ± 0.8 1.6 ± 1.1 0.8 ± 0.6

7.4 ± 6.3 5.9 ± 3.8 1.1 ± 0.8

1.0 ± 0.5 1.0 ± 0.7 1.0 ± 0.5

1.0 ± 1.1 1.0 ± 1.1 2.5 ± 1.1

0.7 ± 0.9 2.1 ± 2.4 14.2 ± 20.6b

1.0 ± 0.4 1.0 ± 1.1 1.0 ± 1.0

2.7 ± 2.4 0.8 ± 0.3 1.0 ± 0.5

5.3 ± 5.2 13.7 ± 15.0 30.5 ± 46.6

Clusterin

1 3 14

1.0 ± 0.4 1.0 ± 0.5 1.0 ± 0.5

0.7 ± 0.3 1.3 ± 0.5 0.9 ± 0.5

1.3 ± 0.3 1.0 ± 0.3 1.2 ± 0.5

1.0 ± 0.5 1.0 ± ± 0.3 1.0 ± 0.4

0.7 ± 0.3 0.8 ± 0.5 1.7 ± 0.3

0.5 ± 0.1 1.7 ± 0.5 1.4 ± 0.8

1.0 ± 0.4 1.0 ± 0.2 1.0 ± 0.4

1.1 ± 0.5 1.0 ± 0.3 1.3 ± 0.3

1.0 ± 0.4d 2.8 ± 4.4d 39.2 ± 43.1d

a b c d

350

Slightly increased mRNA expression in animals nos. 42 and 45 only. Increased gene expression restricted to animals nos. 42 and 45. Up-regulation of mRNA restricted to animals nos. 41 and 43. From Hoffmann et al. (2010).

ing treatment with EMD335823, these changes were restricted to animals showing histopathological evidence for liver damage, and correlated well with thiostatin (T-kininogen) mRNA expression in the target organ as well as with thiostatin serum levels (Fig. 3b). Similar effects were observed in rats dosed with BI-3 at 1000 mg/kg bw, although thiostatin in serum appeared to reflect liver injury and hepatic thiostatin (T-kininogen) mRNA expression better than measurement of urinary thiostatin, with statistically significant changes in serum occurring within 1 day of administration of a single dose (Fig. 3c). Similarly, a rise in serum thiostatin was also found in response to treatment with BAY16 and was accompanied by significant mRNA up-regulation in liver and enhanced urinary excretion (Fig. 3a). In good agreement with up-regulation of thiostatin (T-kininogen) mRNA in rat liver, western blot analysis

revealed increased levels of thiostatin in livers of affected high dose animals compared to untreated controls following treatment with BAY16, EMD335823 and BI-3 for 14 days (Fig. 3d). However, commercially available antibodies against thiostatin were not suitable for immunohistochemical analysis, and we were therefore not able to discriminate between hepatocellular and hepatobiliary localization of thiostatin. While our results demonstrate that over-expression of thiostatin and secretion into serum and urine may represent an immediate response to liver injury, it is important to note that slight changes in gene expression were also observed in kidneys of affected animals although the basal level of transcription and the transcriptional response was much less pronounced as compared to the liver (Table 5).

Fig. 4. Receiver–operating characteristics (ROC) curves for (a) gene expression of candidate biomarkers and (b) serum biomarkers compared to traditional clinical chemistry parameters. The area under the ROC curve, which serves as measure for the overall ability of each biomarker to discriminate rats without signs of hepatotoxicity from those with hepatocyte and bile duct damage, is given in the legend.

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

3.6. Assessment of serum biomarker accuracy Receiver–operating characteristics (ROC) analyses were used to determine the ability of the biomarker candidates to discriminate animals without evidence of liver toxicity from those with histopathological change. Due to the large background variability observed between studies conducted at different sites, all values were normalized to corresponding controls prior to ROC analysis. These analyses revealed thiostatin (T-kininogen) as the most sensitive marker gene (AUC > 0.9) (Fig. 4a). Moreover, serum thiostatin, but not NGAL, clusterin or PON1, was found to be a better indicator of drug-induced hepatobiliary injury than the conventional parameters ALP, ALT, and AST (Fig. 4b). 4. Discussion This study was designed to evaluate the performance of a set of non-invasive candidate markers for improved detection of drug-induced liver injury as compared to traditional hepatotoxicity clinical chemistry parameters. Thus, the potential of literature and omics derived markers, i.e. PON1, thiostatin, NGAL and clusterin, as novel preclinical markers of hepatobiliary toxicity was assessed following treatment of male Wistar rats with drug candidates that were dropped from further development, in part due to hepatotoxicity. Our results suggest that a rise in serum thiostatin, which was accompanied by a corresponding increase in the rate of transcription of thiostatin (T-kininogen) in the target organ and correlated well with the progressive histopathological changes, may provide a more sensitive – albeit not necessarily specific – indicator of drug-induced liver damage than current clinical chemistry parameters, i.e. serum transaminases and alkaline phosphatase. In contrast, NGAL, PON1 and clusterin were not consistently altered and thus only provide limited additive information to the traditional liver enzymes in detecting drug-induced hepatotoxicity. Thiostatin (T-kininogen) is a rat-specific 56 kDa glycoprotein and major acute phase protein which is synthesized primarily in liver and released into plasma during acute inflammatory conditions, e.g. in response to endotoxinemia (Cole et al., 1985; Urban et al., 1979). While it has not been previously associated with drug-induced liver toxicity, proteomics profiling conducted as part of the Innomed PredTox project recently identified thiostatin as being elevated in urine of rats suffering from hepatobiliary injury induced by treatment with EMD335823 and BI-3. To further assess the sensitivity and time- and dose-dependency of the response, we used an enzyme-linked immunosorbent assay to quantify thiostatin in urine and serum along with the determination of its mRNA expression in the target organ. ELISA analysis confirmed increased urinary excretion of thiostatin in response to 14-day treatment with EMD335823 and BI-3, and revealed elevated thiostatin levels at earlier time-points than previously examined. In addition, hepatobiliary toxicity caused by a further drug candidate not previously investigated, BAY16, was also found to cause a marked rise in urinary thiostatin. However, modulation of urinary thiostatin occurred late, compared to the corresponding transcriptional changes and histopathological alterations. Therefore, we speculated that thiostatin in serum may be a more suitable marker of hepatotoxicity than thiostatin excretion into urine. Indeed, statistically significant changes in serum thiostatin were evident as early as day 1 after treatment with BAY16 and BI-3 and, while in good agreement with liver pathology, preceded alterations in the activity of serum transaminases. In support of these observations, ROC analyses confirmed serum thiostatin as a better diagnostic marker of drug-induced hepatobiliary injury than conventional clinical chemistry parameters. To date, little is known regarding the function of thiostatin. It has been speculated that as a potent

9

inhibitor of cysteine proteases, thiostatin may afford protection to healthy cells against proteolytic enzymes released during inflammation and tissue damage (Fung and Schreiber, 1987; Moreau et al., 1988). However, as the liver is known to be the primary source of plasma proteins, including thiostatin (Mann and Lingrel, 1991), it is important to note that increased hepatic expression and alterations in circulating thiostatin levels are not necessarily indicative of liver injury but may also occur in response to tissue injury at other sites, e.g. the kidney (Bandara et al., 2003). Given the sensitivity of the response, however, our results suggest that incorporation of thiostatin into the current battery of tests may enhance our ability to non-invasively diagnose drug-induced tissue damage and inflammation during preclinical safety assessment. Unfortunately, translation into the clinic may be hampered by the fact that thiostatin (T-kininogen) appears to be formed exclusively in the rat and no human equivalent has been reported to date. Similar to thiostatin, NGAL is an acute phase protein expressed by neutrophils and a variety of epithelial cells during inflammation and tissue injury (Bu et al., 2006; Cowland et al., 2006; Flo et al., 2004; Jayaraman et al., 2005). Previous studies demonstrated increased hepatic NGAL mRNA expression in mice chronically exposed to ethanol (Bykov et al., 2007), under conditions of oxidative stress (Roudkenar et al., 2007) and in acute endotoxemia (Sunil et al., 2007). Moreover, gene expression profiling in liver of rats administered model hepatotoxicants, i.e. acetaminophen, bromobenzene, carbon tetrachloride, dimethylnitrosamine, and thioacetamide, identified NGAL gene expression as significantly altered more than 2.0-fold by at least four of the five chemicals tested (Minami et al., 2005). Consistent with these findings, upregulation of NGAL was observed in livers of rats suffering from drug-induced hepatobiliary injury in the present study. Importantly, increased rates of transcription in the target organ were also associated with a rise of NGAL in serum and urine, suggesting that measurement of NGAL in body fluids may provide a diagnostic tool to assess drug toxicity. Urinary excretion of NGAL has previously been regarded as an early biomarker of acute kidney injury (Bennett et al., 2008; Han et al., 2009; Malyszko et al., 2008; Mishra et al., 2006, 2004; Mori and Nakao, 2007; Wagener et al., 2008). However, this and other reports demonstrating increased urinary NGAL accompanied by enhanced immunoreactivity against NGAL in the kidney tubule epithelium in the absence of histopathological signs of nephrotoxicity, suggest that a rise in urinary excretion of NGAL may not be specific for kidney injury but may also occur as a result of increased transcription in response to systemic inflammation or tissue damage at other sites, and subsequent renal handling of NGAL (Hoffmann et al., 2010; Smyth et al., 2009). Indeed, it is known that the low molecular weight lipocalins may be freely filtered in the glomeruli, bound to the scavenger-receptor megalin, which is expressed specifically at the brush borders of proximal tubules, taken up by endocytosis and stored in lysosomes (Hvidberg et al., 2005). Reduced activity of the antioxidant enzyme PON1, most likely as a result of decreased PON1 expression and secretion by the liver, has been suggested as an early biochemical change indicative of hepatotoxicity based on an inverse correlation between hepatic and/or serum PON1 activity and liver function in human and experimental liver disease models, including chronic hepatitis and liver cirrhosis (Amacher et al., 2005; Feingold et al., 1998; Ferre et al., 2001, 2002). Similarly, proteomics analysis of rat liver protein extracts identified PON1 as significantly down-regulated following treatment with acetaminophen and compound A (Amacher et al., 2005; Meneses-Lorente et al., 2004). Although results from our studies demonstrated reduced serum PON1 activity in the absence of changes at the mRNA level in response to liver injury induced by BI-3 and EMD335823, the lack of effects on PON1 activity in rats treated with BAY16, in which marked histopathological alter-

10

M. Adler et al. / Toxicology Letters 196 (2010) 1–11

ations were seen, suggests that serum PON1 may not represent a reliable marker of drug-induced hepatotoxicity. Moreover, while PON1 as a down-regulated marker has been recommended as a valuable addition to a liver biomarker panel (Ozer et al., 2008), the unexpected finding that PON1 activity (but not gene expression) appeared to increase after single administration of a high dose of BI-3 is difficult to reconcile with our expectations of a reliable biomarker. Similarly, a recent study reported increased rather than decreased hepatic and serum PON1 activity in parallel to a rise in serum ALT activity in rats given a single dose of cyclophosphamide (Abraham and Sugumar, 2008). In the absence of further data supporting the value of PON1 as a marker of drug-induced liver injury, these findings suggest that modulation of PON1 activity may at least in part depend on the compound rather than being a regular feature of drug-induced liver damage, although it seems indisputable that a considerable impairment of liver function may result in a decline in PON1 activity secondary to decreased PON1 expression. Clusterin, which is also under evaluation as a urinary biomarker of proximal tubule injury (Hidaka et al., 2002; Hoffmann et al., 2010; Rached et al., 2008; Sieber et al., 2009), was recently found to be significantly upregulated in livers of rats treated with a range of hepatotoxins, including carbon tetrachloride, alphanaphthyl-isothiocyanate, acetaminophen and chloroform (Zidek et al., 2007). Moreover, clusterin was among the top-scoring gene set responsible for the correct discrimination between hepatotoxic and nonhepatotoxic model compounds (Zidek et al., 2007). While these data suggested that clusterin expression and subsequent release into body fluids may serve as a promising marker of drug-induced liver injury, hepatic expression as well as concentrations of clusterin in serum did not correlate well with the histopathological changes seen in response to BI-3 and EMD335823 in the present study. The poor diagnostic value of clusterin as an indicator of druginduced liver injury is also reflected by the small area under the ROC curve. While the significant increase in urinary excretion of clusterin observed in rats treated with BI-3 is likely related to the drug’s nephrotoxic effects, it is interesting to note that elevated urinary levels were also seen in individual animals dosed with EMD335823 and BAY16, which were not associated with kidney injury. Thus, the possibility that increased urinary clusterin levels may not be as specific for kidney injury as generally assumed should be taken into consideration when applying clusterin as a novel biomarker of nephrotoxicity. In summary, comprehensive analysis of candidate serum biomarkers of liver injury and corresponding transcriptional changes in the target organ provided important information regarding the ability of a number of candidate liver markers to accurately detect drug-induced hepatotoxicity. While our data suggest that thiostatin may be a promising marker of inflammation and tissue damage in toxicity studies in rats, further evaluation of its sensitivity and specificity is needed. Conflicts of interest None. Acknowledgements This work was supported in part by funding by the 6th Research Framework Program of the European Union (Innomed PredTox, LSHB-CT-2005-518170). The contribution of all members of the PredTox Consortium (www.innomed-predtox.com) is gratefully acknowledged. The authors would also like to thank Ursula Tatsch, University of Würzburg, for excellent technical assistance. PredTox Consortium: Nycomed, Bayer Schering Pharma, Boehringer-Ingelheim, Johnson & Johnson, Lilly, Merck-Serono,

Novartis, Novo-Nordisk, Schering-Plough, Roche, Sanofi-Aventis, Servier, University of Würzburg, University College Dublin, University Hacettepe, Bio-Rad, Genedata. References Abraham, P., Sugumar, E., 2008. Increased glutathione levels and activity of PON1 (phenyl acetate esterase) in the liver of rats after a single dose of cyclophosphamide: a defense mechanism? Exp. Toxicol. Pathol. 59, 301–306. Amacher, D.E., Adler, R., Herath, A., Townsend, R.R., 2005. Use of proteomic methods to identify serum biomarkers associated with rat liver toxicity or hypertrophy. Clin. Chem. 51, 1796–1803. Bandara, L.R., Kelly, M.D., Lock, E.A., Kennedy, S., 2003. A correlation between a proteomic evaluation and conventional measurements in the assessment of renal proximal tubular toxicity. Toxicol. Sci. 73, 195–206. Bennett, M., Dent, C.L., Ma, Q., Dastrala, S., Grenier, F., Workman, R., Syed, H., Ali, S., Barasch, J., Devarajan, P., 2008. Urine NGAL predicts severity of acute kidney injury after cardiac surgery: a prospective study. Clin. J. Am. Soc. Nephrol. 3, 665–673. Bu, D.X., Hemdahl, A.L., Gabrielsen, A., Fuxe, J., Zhu, C., Eriksson, P., Yan, Z.Q., 2006. Induction of neutrophil gelatinase-associated lipocalin in vascular injury via activation of nuclear factor-kappaB. Am. J. Pathol. 169, 2245–2253. Bykov, I., Junnikkala, S., Pekna, M., Lindros, K.O., Meri, S., 2007. Effect of chronic ethanol consumption on the expression of complement components and acutephase proteins in liver. Clin. Immunol. 124, 213–220. Cole, T., Inglis, A., Nagashima, M., Schreiber, G., 1985. Major acute-phase alpha(1)protein in the rat: structure, molecular cloning, and regulation of mRNA levels. Biochem. Biophys. Res. Commun. 126, 719–724. Cowland, J.B., Muta, T., Borregaard, N., 2006. IL-1beta-specific up-regulation of neutrophil gelatinase-associated lipocalin is controlled by IkappaB-zeta. J. Immunol. 176, 5559–5566. Feingold, K.R., Memon, R.A., Moser, A.H., Grunfeld, C., 1998. Paraoxonase activity in the serum and hepatic mRNA levels decrease during the acute phase response. Atherosclerosis 139, 307–315. Ferre, N., Camps, J., Cabre, M., Paul, A., Joven, J., 2001. Hepatic paraoxonase activity alterations and free radical production in rats with experimental cirrhosis. Metabolism 50, 997–1000. Ferre, N., Camps, J., Prats, E., Vilella, E., Paul, A., Figuera, L., Joven, J., 2002. Serum paraoxonase activity: a new additional test for the improved evaluation of chronic liver damage. Clin. Chem. 48, 261–268. Ferre, N., Marsillach, J., Camps, J., Mackness, B., Mackness, M., Riu, F., Coll, B., Tous, M., Joven, J., 2006. Paraoxonase-1 is associated with oxidative stress, fibrosis and FAS expression in chronic liver diseases. J. Hepatol. 45, 51–59. Flo, T.H., Smith, K.D., Sato, S., Rodriguez, D.J., Holmes, M.A., Strong, R.K., Akira, S., Aderem, A., 2004. Lipocalin 2 mediates an innate immune response to bacterial infection by sequestrating iron. Nature 432, 917–921. Fung, W.P., Schreiber, G., 1987. Structure and expression of the genes for major acute phase alpha 1-protein (thiostatin) and kininogen in the rat. J. Biol. Chem. 262, 9298–9308. Gallagher, W.M., Tweats, D., Koenig, J., 2009. Omic profiling for drug safety assessment: current trends and public–private partnerships. Drug Discov. Today 14, 337–342. Han, W.K., Wagener, G., Zhu, Y., Wang, S., Lee, H.T., 2009. Urinary biomarkers in the early detection of acute kidney injury after cardiac surgery. Clin. J. Am. Soc. Nephrol. 4, 873–882. Hidaka, S., Kranzlin, B., Gretz, N., Witzgall, R., 2002. Urinary clusterin levels in the rat correlate with the severity of tubular damage and may help to differentiate between glomerular and tubular injuries. Cell Tissue Res. 310, 289–296. Hillmeister, P., Lehmann, K.E., Bondke, A., Witt, H., Duelsner, A., Gruber, C., Busch, H.J., Jankowski, J., Ruiz-Noppinger, P., Hossmann, K.A., Buschmann, I.R., 2008. Induction of cerebral arteriogenesis leads to early-phase expression of protease inhibitors in growing collaterals of the brain. J. Cereb. Blood Flow Metab. 28, 1811–1823. Hoffmann, D., Adler, M., Vaidya, V., Rached, E., Mulrane, L., Gallagher, W.M., Callanan, J.J., Gautier, J.C., Matheis, K., Staedtler, F., Dieterle, F., Brandenburg, A., Sposny, A., Hewitt, P., Ellinger-Ziegelbauer, H., Bonventre, J.V., Dekant, W. and Mally, A., (2010). Performance of novel kidney biomarkers in preclinical toxicity studies. Toxicol Sci, Jan 29 2010 [Epub ahead of print] kfq029 [pii] doi:10.1093/toxsci/kfq029. Hvidberg, V., Jacobsen, C., Strong, R.K., Cowland, J.B., Moestrup, S.K., Borregaard, N., 2005. The endocytic receptor megalin binds the iron transporting neutrophilgelatinase-associated lipocalin with high affinity and mediates its cellular uptake. FEBS Lett. 579, 773–777. Jayaraman, A., Roberts, K.A., Yoon, J., Yarmush, D.M., Duan, X., Lee, K., Yarmush, M.L., 2005. Identification of neutrophil gelatinase-associated lipocalin (NGAL) as a discriminatory marker of the hepatocyte-secreted protein response to IL-1beta: a proteomic analysis. Biotechnol. Bioeng. 91, 502–515. Malyszko, J., Bachorzewska-Gajewska, H., Sitniewska, E., Malyszko, J.S., Poniatowski, B., Dobrzycki, S., 2008. Serum neutrophil gelatinase-associated lipocalin as a marker of renal function in non-diabetic patients with stage 2–4 chronic kidney disease. Ren. Fail. 30, 625–628. Mann, E.A., Lingrel, J.B., 1991. Developmental and tissue-specific expression of rat T-kininogen. Biochem. Biophys. Res. Commun. 174, 417–423. Meneses-Lorente, G., Guest, P.C., Lawrence, J., Muniappa, N., Knowles, M.R., Skynner, H.A., Salim, K., Cristea, I., Mortishire-Smith, R., Gaskell, S.J., Watt, A., 2004. A pro-

M. Adler et al. / Toxicology Letters 196 (2010) 1–11 teomic investigation of drug-induced steatosis in rat liver. Chem. Res. Toxicol. 17, 605–612. Minami, K., Saito, T., Narahara, M., Tomita, H., Kato, H., Sugiyama, H., Katoh, M., Nakajima, M., Yokoi, T., 2005. Relationship between hepatic gene expression profiles and hepatotoxicity in five typical hepatotoxicant-administered rats. Toxicol. Sci. 87, 296–305. Mishra, J., Ma, Q., Kelly, C., Mitsnefes, M., Mori, K., Barasch, J., Devarajan, P., 2006. Kidney NGAL is a novel early marker of acute injury following transplantation. Pediatr. Nephrol. 21, 856–863. Mishra, J., Mori, K., Ma, Q., Kelly, C., Barasch, J., Devarajan, P., 2004. Neutrophil gelatinase-associated lipocalin: a novel early urinary biomarker for cisplatin nephrotoxicity. Am. J. Nephrol. 24, 307–315. Mitsnefes, M.M., Kathman, T.S., Mishra, J., Kartal, J., Khoury, P.R., Nickolas, T.L., Barasch, J., Devarajan, P., 2007. Serum neutrophil gelatinase-associated lipocalin as a marker of renal function in children with chronic kidney disease. Pediatr. Nephrol. 22, 101–108. Moreau, T., Esnard, F., Gutman, N., Degand, P., Gauthier, F., 1988. Cysteineproteinase-inhibiting function of T kininogen and of its proteolytic fragments. Eur. J. Biochem. 173, 185–190. Mori, K., Nakao, K., 2007. Neutrophil gelatinase-associated lipocalin as the real-time indicator of active kidney damage. Kidney Int. 71, 967–970. Mulrane, L., Rexhepaj, E., Smart, V., Callanan, J.J., Orhan, D., Eldem, T., Mally, A., Schroeder, S., Meyer, K., Wendt, M., O’Shea, D., Gallagher, W.M., 2008. Creation of a digital slide and tissue microarray resource from a multi-institutional predictive toxicology study in the rat: an initial report from the PredTox group. Exp. Toxicol. Pathol. 60, 235–245. Nathwani, R.A., Pais, S., Reynolds, T.B., Kaplowitz, N., 2005. Serum alanine aminotransferase in skeletal muscle diseases. Hepatology 41, 380–382. Ostapowicz, G., Lee, W.M., 2000. Acute hepatic failure: a Western perspective. J. Gastroenterol. Hepatol. 15, 480–488.

11

Ozer, J., Ratner, M., Shaw, M., Bailey, W., Schomaker, S., 2008. The current state of serum biomarkers of hepatotoxicity. Toxicology 245, 194–205. Rached, E., Hoffmann, D., Blumbach, K., Weber, K., Dekant, W., Mally, A., 2008. Evaluation of putative biomarkers of nephrotoxicity after exposure to ochratoxin a in vivo and in vitro. Toxicol. Sci. 103, 371–381. Roudkenar, M.H., Kuwahara, Y., Baba, T., Roushandeh, A.M., Ebishima, S., Abe, S., Ohkubo, Y., Fukumoto, M., 2007. Oxidative stress induced lipocalin 2 gene expression: addressing its expression under the harmful conditions. J. Radiat. Res. (Tokyo) 48, 39–44. Shannan, B., Seifert, M., Leskov, K., Willis, J., Boothman, D., Tilgen, W., Reichrath, J., 2006. Challenge and promise: roles for clusterin in pathogenesis, progression and therapy of cancer. Cell Death Differ. 13, 12–19. Sieber, M., Hoffmann, D., Adler, M., Vaidya, V.S., Clement, M., Bonventre, J.V., Zidek, N., Rached, E., Amberg, A., Callanan, J.J., Dekant, W., Mally, A., 2009. Comparative analysis of novel noninvasive renal biomarkers and metabonomic changes in a rat model of gentamicin nephrotoxicity. Toxicol. Sci. 109, 336–349. Smyth, R., Lane, C.S., Ashiq, R., Turton, J.A., Clarke, C.J., Dare, T.O., York, M.J., Griffiths, W., Munday, M.R., 2009. Proteomic investigation of urinary markers of carbon-tetrachloride-induced hepatic fibrosis in the Hanover Wistar rat. Cell. Biol. Toxicol. 25, 499–512. Sunil, V.R., Patel, K.J., Nilsen-Hamilton, M., Heck, D.E., Laskin, J.D., Laskin, D.L., 2007. Acute endotoxemia is associated with upregulation of lipocalin 24p3/Lcn2 in lung and liver. Exp. Mol. Pathol. 83, 177–187. Urban, J., Chan, D., Schreiber, G., 1979. A rat serum glycoprotein whose synthesis rate increases greatly during inflammation. J. Biol. Chem. 254, 10565–10568. Wagener, G., Gubitosa, G., Wang, S., Borregaard, N., Kim, M., Lee, H.T., 2008. Urinary neutrophil gelatinase-associated lipocalin and acute kidney injury after cardiac surgery. Am. J. Kidney Dis. 52, 425–433. Zidek, N., Hellmann, J., Kramer, P.J., Hewitt, P.G., 2007. Acute hepatotoxicity: a predictive model based on focused illumina microarrays. Toxicol. Sci. 99, 289–302.