TOXICOLOGICAL SCIENCES 90(2), 419–431 (2006) doi:10.1093/toxsci/kfj088 Advance Access publication October 29, 2005

Gene Expression Profiling of Nephrotoxicity from the Sevoflurane Degradation Product Fluoromethyl-2,2-difluoro-1- (trifluoromethyl)vinyl Ether (‘‘Compound A’’) in Rats

Evan D. Kharasch,*,1 Jesara L. Schroeder,† Theo Bammler,‡ Richard Beyer,‡ and Sengkeo Srinouanprachanh‡

*Department of Anesthesiology, Washington University, St. Louis, Missouri 63110–1093; †Department of Anesthesiology, University of Washington, Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 Seattle, Washington 98195; and ‡Department of Environmental Health, University of Washington, Seattle, Washington 98195

Received October 19, 2005; accepted December 25, 2005

The haloalkene fluoromethyl-2,2-difluoro-1-(trifluoromethyl)- The major degradation product of the volatile anesthetic sevo- vinyl ether (FDVE, referred to as ‘‘compound A’’ in the drug flurane, the haloalkene fluoromethyl-2,2-difluoro-1-(trifluoromethyl)- label) results from the degradation of the volatile anesthetic vinyl ether (FDVE or ‘‘compound A’’), is nephrotoxic in rats. sevoflurane in anesthesia machines by carbon dioxide absor- FDVE undergoes complex metabolism and bioactivation, which bents which contain strong base (Frink et al., 1992; Hanaki mediates the nephrotoxicity. Nevertheless, the molecular and cellu- lar mechanisms of FDVE toxification are unknown. This inves- et al., 1987). FDVE causes proximal tubular necrosis in rats tigation evaluated the expression profile of kidneys in rats when administered by inhalation or intraperitoneal injection administered a nephrotoxic dose of FDVE. Male Fischer 344 rats (Keller et al., 1995; Kharasch et al., 1997, 1998; Morio et al., (five per group) received 0.25 mmol/kg intraperitoneal FDVE or 1992). Necrosis is accompanied by increased serum creatinine, corn oil (controls) and were sacrificed after 24 or 72 h. Urine diuresis, and urinary excretion of , glucose, and renal output and kidney histological changes were quantified. Kidney tubular enzymes. In rats, FDVE undergoes a complex route of RNA was extracted for microarray analysis using Affymetrix metabolism and bioactivation, which mediates the nephrotox- GeneChipÒ Rat Expression Array 230A arrays. Quantitative real- icity (Altuntas et al., 2004; Anders, 2005; Iyer and Anders, time PCR confirmed the modulation of several . FDVE 1996, 1997; Iyer et al., 1998; Jin et al., 1995, 1996; Sheffels caused significant diuresis and necrosis at 24 h, with normal urine et al., 2004; Spracklin and Kharasch, 1996; Tong and Anders, output and evidence of tubular regeneration at 72 h. There were 2002; Uttamsingh et al., 1998). FDVE reacts with glutathione 517 informative genes that were differentially expressed >1.5-fold ( p < 0.05) versus control at 24 h, of which 283 and 234 were to form glutathione conjugates, which undergo cleavage to upregulated and downregulated, respectively. Major classes of the corresponding cysteine S-conjugates, which may undergo upregulated genes included those involved in apoptosis, oxidative N-acetylation to form nontoxic mercapturates, which are ex- stress, and inflammatory response (mostly at 24 h), and re- creted in urine. Reactive intermediates may be formed via the generation and repair; downregulated genes were generally metabolism of FDVE cysteine S-conjugates and by the sulfox- associated with transporters and intermediary metabolism. idation FDVE of cysteine S- or mercapturic acid conjugates. Among the quantitatively most upregulated genes were kidney Although the metabolism of FDVE is well-characterized, the injury molecule, osteopontin, clusterin, tissue inhibitor of metal- mechanism of nephrotoxicity is poorly understood. Like many loproteinase 1, and TNF receptor 12, which have been associated other nephrotoxic haloalkenes, renal cysteine conjugate b- with other forms of nephrotoxicity, and angiopoietin-like protein lyase-catalyzed metabolism of FDVE cysteine S-conjugates, 4, glycoprotein nmb, ubiquitin hydrolase, and HSP70. Microarray and sulfoxidation of mercapturates, are thought to mediate results were confirmed by quantitative real-time PCR. FDVE causes rapid and brisk changes in , providing nephrotoxicity in rats in vivo (Altuntas et al., 2003, 2004; potential insights into the mechanism of FDVE toxification, and Anders, 2005; Iyer et al., 1997; Kharasch et al., 1997, 1998; potential biomarkers for FDVE nephrotoxicity which are more Sheffels et al., 2004). FDVE cysteine S-conjugates metabolism sensitive than conventional measures of renal function. by renal b-lyase forms reactive intermediates (thiolate and Key Words: sevoflurane; compound A; nephrotoxicity; haloal- thioacyl fluoride) (Tong and Anders, 2002); however, it is not kene; microarray; kidney injury molecule. known if, or how, they contribute to FDVE renal toxicity in vivo. Certain FDVE S-conjugates sulfoxides are also highly reactive (Sheffels et al., 2004) and toxic in vitro (Altuntas et al., 2003), but their potential role in toxicity in vivo is similarly 1 To whom correspondence should be addressed at Clinical Research Division, Department of Anesthesiology, Washington University, 660 S. Euclid unknown. Other haloalkenes S-conjugates also undergo b- Ave., Campus Box 8054, St. Louis, MO 63110–1093. Fax: (314) 362-8571. lyase-mediated metabolism to reactive intermediates (thioacyl E-mail: [email protected]. halides, thioketenes) (Anders, 2004), and form adducts with

Ó The Author 2005. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For Permissions, please email: [email protected] 420 KHARASCH ET AL. renal and other macromolecules which are thought to Male Fischer 344 rats (12 weeks, 225–275 g) were purchased from Harlan (San participate in toxification (Anders, 2004). Mitochondria are Diego, CA). Rats were housed in individual metabolic cages, acclimated for at a major target of several haloalkene S-conjugates in isolated least 48 h prior to any experiments, provided food and water ad libitum,and maintained on a 12-h light-dark cycle (7 A.M.to7P.M.). The FDVE dose used in renal cells (Lash et al., 2000); however, it is unknown if they these experiments (0.25 mmol/kg) was based on the threshold dose for are a target of FDVE cysteine S- or mercapturic acid nephrotoxicity observed in previous experiments (Kharasch et al., 1997). Twenty conjugates. In addition to protein alkylation and mitochondrial rats were randomized to one of four treatment groups (n ¼ 5 per group): no dysfunction, other postulated mechanisms of haloalkene cyto- treatment, corn oil (24 h), FDVE (24 h), or FDVE (72 h). No-treatment controls toxicity include oxidative stress, calcium ion dysregulation, were euthanized 48 h after transfer to the metabolic cages. Corn oil controls received 2 ml/kg corn oil and were euthanized 24 h later. FDVE rats received 0.25 apoptosis, and altered expression of genes regulating cell mmol/kg in corn oil by intraperitoneal injection and were euthanized 24 or 72 h growth and differentiation (Lash et al., 2000, 2001, 2003). It later. Urine was collected in 24-h intervals. Rats were anesthetized with is unknown whether any of these pathways are operant in pentobarbital and sacrificed by cardiac exsanguination. Kidneys were immediately FDVE toxicity. excised, cut in a midtransverse plane through cortex and medullary pyramid, and Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 In humans, FDVE undergoes qualitatively similar biotrans- a section fixed in 10% neutral buffered formalin. Sections were stained with hematoxylin and eosin and periodic acid-Schiff, and histologic analysis was formation to glutathione and cysteine S-conjugates (Altuntas performed by a veterinary pathologist blinded to animal treatments. Histopath- and Kharasch, 2001, 2002; Altuntas et al., 2004; Iyer and ologic changes were recorded in regard to location, character, and severity. The Anders, 1996; Iyer et al., 1998; Kharasch and Jubert, 1999). semiquantitative severity score consisted of a range from 0 to 4 (normal, minimal, Nevertheless, exposure of surgical patients to FDVE during slight, moderate and marked, respectively), which reflected the degree and sevoflurane anesthesia has been found to have no clinically distribution of the tubular necrosis. significant effects (Bito and Ikeda, 1996; Conzen et al., 2002; Microarrays. The remainder of the kidney was saved in RNAlater and Fukuda et al., 2004; Higuchi et al., 1998; Kharasch et al., 2001, stored at 20° C for future processing. RNA was extracted from 10 to 25 mg of 2003; Obata et al., 2000). These clinical investigations have kidney tissue using the RNeasy Mini kit (Qiagen Inc, Valencia, CA) according to the manufacturer’s directions except as follows: Tissue disruption was evaluated the conventional markers of serum creatinine and performed in an ice-cold Teflon-glass homogenizer with 600 ll of RNeasy lysis urea nitrogen, and also proteinuria, glucosuria, and enzymuria. buffer. This was followed by further homogenization through a 20-g needle Nevertheless, there is concern about the sensitivity of the with a 1-ml syringe. Subsequent procedures followed the manufacturer’s conventional clinical measures of renal function and consider- directions for RNA extraction from animal tissues. RNA labeling, gene chip able effort to identify more sensitive markers of renal toxicity hybridizations, and gene chip scans were conducted at the University of Washington Center for Expression Arrays. Each total RNA sample was in humans (Hewitt et al., 2004). evaluated for quality on an Agilent 2100 bioanalyzer, then converted into Toxicogenomics holds promise for elucidating cellular biotin-labeled cRNA using the Affymetrix eukaryotic target labeling protocol mechanisms of xenobiotic renal toxification, identifying po- (www.affymetrix.com). Briefly, 5 lg of total RNA was reverse transcribed to tential site-specific biomarkers of toxicity, which could afford double-stranded cDNA via a round of transcription with Superscript II and then greater sensitivity, earlier detection, and/or quantitative esti- a round with T4 DNA polymerase. The resulting cDNA was converted to biotinylated cRNA in the presence of T7 DNA polymerase and biotin-labeled mates of tissue injury, and for forecasting potential candidate nucleotides. This resulting cRNA was fragmented and used for hybridization to protein biomarkers (Bailey and Ulrich, 2004; Hayes and an Affymetrix Rat 230A GeneChip according to standard Affymetrix protocols. Bradfield, 2005; Thukral et al., 2005). Gene expression profiles The 230A chip contains 30,248 transcripts for 28,757 genes. have been generated for ischemic renal injury and several renal Quantitation of specific gene expression. Quantitation of mRNA levels toxins known to act via diverse pathways, with the emergence for specific genes was accomplished by real-time quantitative PCR with of some expression profiles for generalized renal nephrotoxic- a fluorogenic 5# nuclease-dependant (TaqManÒ-based) quantitative gene ity and for toxin-specific injury (Amin et al., 2004; Basile expression assay using an ABI (Applied Biosystems Inc., Foster City, CA) et al., 2005; Davis et al., 2004; Devarajan et al., 2003; iScience 7900HT Fast Real-Time PCR System, as described previously (Diaz et al., 2001; Kevil et al., 2004; Lin et al., 2002). PCR primers and probes Goodsaid, 2004; Huang et al., 2001; Kramer et al., 2004; Lu¨he (Table 1) were selected using Primer Express 1.5Ò software supplied by ABI. et al., 2003; Thompson et al., 2004; Thukral et al., 2005). The The sequences of each gene were taken from Affymetrix’s NetAffxÒ Analysis purpose of this investigation was to evaluate the gene Center (www.Affymetrix.com). A reference standard was identified by eval- expression profile of kidneys in rats exposed to doses of FDVE uating similar sample types for high mRNA expression of GAPDH. After an that are known to cause nephrotoxicity. appropriate reference standard was established, this sample was serially diluted to derive a linear regression formula that was used to calculate and quantitate specific gene expression. The levels of mRNA expression of the GAPDH gene were used to normalize these data. The PCR mixture (20 ll final volume) for MATERIALS AND METHODS this assay consisted of the appropriate sense and antisense primers (0.35 lM each), 100 nM TaqMan probe, and 13 TaqManÒ Fast Universal PCR Master Mix (Applied Biosystems, Inc.). Amplification and detection of fluorescence Animal treatments. Fluoromethyl-2,2-difluoro-1-(trifluoromethyl)vinyl ether was measured using the ABI 7900 system with the following PCR reaction (FDVE, 99.92% purity) was provided by Abbott Laboratories (Abbott Park, IL). profile: 1 cycle of 95°C for 15 s, 40 cycles of 95°C for 1 s, and 60°C for 20 s. RNAlater and the RNeasy mini kit were purchased from Qiagen (Valencia, CA). GeneChipÒ Rat Expression Array 230A arrays were purchased from Affymetrix Data analysis. Image processing and expression analysis were performed (Santa Clara, CA). All experiments were approved by the University of using Affymetrix GeneChip Operating Software (GCOS). Each GeneChip Washington Animal Care and Use Committee and conducted in accordance with image underwent GCOS absolute expression analysis. The quality of the American Association for Accreditation of Laboratory Animal Care guidelines. hybridizations and overall chip performance were determined by visual HALOALKENE NEPHROTOXICITY MICROARRAY ANALYSIS 421

TABLE 1 Primers and Probes for RT-PCR

Gene Oligonucleotide sequences and conjugations Tm (°C) Product size

GAPDH Glyceraldehyde-3-phosphate dehydrogenase 125bp Sense primer 5#-TCC TGC ACC ACC AAC TGC TT-3# 59.4 Antisense primer 5#-GAG GGG CCA TCC ACA GTC TT-3# 61.4 Probe 5#-6FAM-CAC TCA TGA CCA CAG TCC ATG CCA TCA C-TAMRA-3# 65.7 CFTR2 ATP-binding cassette, subfamily C 108bp Sense primer 5#-ACg gAT AgC CTC ATT CAg ACg AC-3# 58.4 Antisense primer 5#-GAC CAT TAT CTT GTC ACT GTC CAT GA-3# 56.6 # # Probe 5 -6FAM-TCT CCC AgT gCA Cgg TCA TCA CCA TC-TAMRA-3 64.8 Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 KIM Kidney injury molecule 1 146bp Sense primer 5#-GAG TTC ATT AGA GCC ATT TCC ACT CC-3# 57.4 Antisense primer 5#-GAA AGC CTG TGT CCT GCT CTC TCT-3# 60.5 Probe 5#-6FAM-AAC TCA CCC ACT GAG CTC TGA ATT AGG TGC AG-TAMRA-3# 64.8 RGN 134bp Sense primer 5#-CTC TAG GCC TTC TAT TAA AAA CAA TCT GTA G-3# 55.3 Antisense primer 5#-CCT TAT GTA CAG CAA GTA TGG AAT GC-3# 55.9 Probe 5#-6FAM-CAA GCC CCA GAG CAA TTG ATC AGG ATG AT-TAMRA-3# 62.2 TNF rsf21A Tumor necrosis factor receptor superfamily, member 12a 120bp Sense primer 5#-CAA GCT AGG TCA CAA AGC GAC TC-3# 57.9 Antisense primer 5#-GAA CCC AGT GTT GTG TCT GCC-3# 58.6 Probe 5#-6FAM-TTG CTC CCC ACA AGT CCT GGA GGC-TAMRA-3# 66.0 HNMT Histamine N-methyltransferase 108bp Sense primer 5#-TGA AAT CTA CAA GTT TCC AAG ACA ACT TAG-3# 55.5 Antisense primer 5#-TGT TTC CAG TTG GGC AGT GTT TA-3# 57.1 Probe 5#-6FAM-CTG AGC CAC TTC CAG TTG TGA CCT TAA AAG CT-TAMRA-3# 63.0 SPP1 Secreted phosphoprotein 1 (Osteopontin) 137bp Sense primer 5#-GCC ACA GAT GAG GAC CTC ACC-3# 59.8 Antisense primer 5#-TGA CTT GAC TCA TGG CTG GTC TT-3# 58.3 Probe 5#-6FAM-AGC GTC TGA GCG TGC CCT CTG ATC AG-TAMRA-3# 66.6 HSP70 1B Heat shock 70kD protein 1B 132bp Sense primer 5#-CAC CAT CGA GGA GGT GGA TTA GA-3# 57.9 Antisense primer 5#-CTG TAA CTT TAA ACT GAA CTC CGG AGA GA-3# 57.8 Probe 5#-6FAM-TGC ACC AGC AGC CAT CAA GAG TCT GTC T-TAMRA-3# 64.8 CLU Clusterin 124bp Sense primer 5#-GAG AGG CTG ACC CAG CAG TAC AA-3# 60.7 Antisense primer 5#-TGA GGT TAG CCA GCT GGG ACA-3# 60.6 Probe 5#-6FAM-CTC CAG TCC AAG ATG CTC AAC ACC TCA TCC-TAMRA-3# 64.0 SCF21m1 Solute carrier family 21, member 1 138bp Sense primer 5#-CCA CAC TAC ACT TTA AAG CTT CCT TCA TT-3# 57.3 Antisense primer 5#-TCA ATG AAA TTA AAT ACT GTT TCA GGA ATA GG-3# 54.4 Probe 5#-6FAM-CAA TCA AGG GAA ATA TGT GTT TCC CAC ACA TCT TT-TAMRA-3# 60.6 CRFG G protein-binding protein CRFG 108bp Sense primer 5#-AGA AAA CGA GAA GAG TCT GTT CCT CC-3# 58.1 Antisense primer 5#-TCA CCA TCT TGA CAT CCC GAA G-3# 56.8 Probe 5#-6FAM-TCA CGT GGC GTT TTA GAG CAA CTC CGA-TAMRA-3# 63.4 HSP70i Heat shock protein 70kD inducible 107bp Sense primer 5#-TCC TAT GCC TTC AAC ATG AAG AGC-3# 61.0 Antisense primer 5#-GAG ATG ACC TCC TGG CAC TTG TC-3# 64.2 Probe 5#-6FAM-TCT TCT TGT CAG CCT CGC-MGB-3# 70.9

inspection of both the raw scanned data and the extracted quality control and data normalization were carried out with: GeneTrafficÒ (Iobion Infor- metrics, including performance of spike-in controls and endogenous house- matics LLC, La Jolla, CA) and Bioconductor software developed by collab- keeping genes (Actin and GAPDH). The resultant cell intensity files (CEL) orators based at the Biostatistics Unit of the Dana Farber Cancer Institute at were analyzed by the University of Washington Center for Ecogenetics and Harvard Medical School/Harvard School of Public Health [http://bioconductor. Environmental Health: Bioinformatics and Biostatistics Core. Microarray data org] (Gentleman et al., 2004). These software packages were used to hier- are available online. (Add link to supplementary data here) Statistical analysis archically cluster the data and evaluate the data statistically and qualitatively. 422 KHARASCH ET AL.

Each data set consisted of five replicates, with each replicate corresponding to 24 versus 72 h with 72-h values as baseline (24 vs. 72-h data a single animal. Data normalization was performed in GenetrafficÒ using set). Each data set consisted of an expression array from each a modified version of the Robust Multi-Array method (Irizarry et al., 2003), of five animals. A current consensus uses a filter criterion of called the GC-RMA method (Wu et al., 2004), which uses the GC content of the mismatch probes for a better background adjustment of the perfect genes that are up- or downregulated two-fold or greater. match probes. However, due to the number of replicates and the agreement Statistical analysis was performed with GenetrafficÒ to evaluate increases between them, we felt confident in using a 1.5-fold change in or decreases in gene expression using an unpaired two class t-test with the expression as the threshold to filter these three data sets. Using Benjamini-Hochberg correction for multiple comparisons (Benjamini and this criterion, the 24-h data set had 795 genes that were Hochberg, 1995). Results for comparison pairs are expressed as a fold change. Statistical significance was assigned at a minimum 1.5-fold change and p< differentially expressed versus controls by at least 1.5-fold, 0.05. No significant differences in gene expression were found between the with 414 genes upregulated and 381 downregulated. The 72-h untreated and corn oil–treated rats; thus the corn oil–treated rats were used as data set had a total of 490 genes, with 274 up- and 216 the controls for all subsequent analyses. Data sets compared were FDVE (24 h) downregulated. The 24-h versus 72-h data set had a total of 381 Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 versus corn oil, FDVE (72 h) versus corn oil, and FDVE (24 h) versus FDVE genes that were differentially expressed, with 179 upregulated (72 h). RT-PCR results are expressed as the mean ± SD. Data were analyzed by and 202 downregulated. The resulting lists of genes for each analysis of variance, and significance assigned at p<0.05. data set were further evaluated for statistical significance using an unpaired two-class t-test with an adjusted p value < 0.05 and the Benjamini-Hochberg method to correct for multiple RESULTS hypothesis testing and prevent false positives. This provided a final list of 544 informative genes with a fold change in Effects of FDVE on animal weight, urine output, and renal expression of 1.5 or greater and p<0.05 in one or more of our tubular necrosis are provided in Table 2. Significant necrosis data sets. was observed in rats given FDVE, and evidence of tubular An unsupervised hierarchial cluster analysis was performed regeneration was observed at 72 h. Significant diuresis was also on this list of 544 genes (Fig. 1). Samples clustered according observed, with return to control values at 72 h. One animal in to treatment and time, except for one FDVE-treated rat at 72 h the 72-h group was less affected (see below) that the others in which clustered with the corn oil controls. Clustering reflected this group (24, 48, and 72-h urine output was 10, 10, and 8 ml, the renal injury. The FDVE-treated animal which clustered compared with means of 29, 16, and 13 ml in the other four with the controls was the one which showed substantially less animals; renal necrosis was scant, compared with a median of 3 necrosis and urine output (see above). Overall, 72-h results in the other four animals). In the remaining animals, renal showed greater interindividual variability than the 24-h values, effects were similar to those seen previously at this intraper- which may reflect inherent variability in the capacity to recover itoneal dose (Kharasch et al., 1997; Sheffels et al., 2004) and at from FDVE. approximately 350–400 ppm-h inhaled FDVE (Keller et al., The overall list of 544 genes with 1.5-fold change and p< 1995; Kharasch et al., 1998). Previous studies have also shown 0.05 was further analyzed for group comparisons. The 24-h that FDVE renal effects were greatest after 24 h (diuresis and FDVE versus control data set had 517 genes (of which 234 osmolality) or 48 h (proteinuria), with recovery apparent at 72 h were upregulated and 283 were downregulated), the 72-h (Kharasch et al., 1997, 1998; Sheffels et al., 2004). FDVE versus control data set had no genes, and the 24-h Data were evaluated first for the qualitative amount of FDVE versus 72-h FDVE set had 127 genes (of which 72 were differential gene expression (fold change). The following upregulated and 50 were downregulated at 72 h compared with comparisons were performed: 24 h versus corn oil baseline 24 h, Tables 3 and 4). These are not independent lists of genes (24-h data set), 72 h versus corn oil baseline (72-h data set) and (Fig. 2). Of the 127 genes that vary significantly between the

TABLE 2 FDVE Nephrotoxicity

Weight (g) Urine output (ml) Renal necrosis

Group Predose 24 h 72 h 24 h 48 h 72 h 24 h 72 h

Control (no treatment) 253 ± 4 240 ± 9 10 ± 6 1 (0.5–1) Control (corn oil) 252 ± 17 251 ± 15 8 ± 1 1 (0–1) FDVE 24 h 246 ± 17 235 ± 14 23 ± 7a 3 (2–4)a FDVE 72 h 248 ± 26 237 ± 19 234 ± 16 25 ± 10a 15 ± 4a 12 ± 3 3 (0.5–4)a

Note. Urine output is expressed as mean ± SD, necrosis as median (range), n ¼ 5 per group. aSignificantly different from both controls ( p<0.05). HALOALKENE NEPHROTOXICITY MICROARRAY ANALYSIS 423

24- and 72-h time points, 97 were also in the 517-gene group which was significantly different at 24 h versus control, suggesting that these 97 genes were returning to control values at 72 h. This leaves 27 unique genes that vary at 72 h but not 24 h. The lack of statistical significance between the 72-h FDVE versus corn oil data set appeared at variance with the degree of gene modulation and was thought to be due to the 72-h FDVE-treated animal that clustered with the controls, and the resulting group variance leading to lack of statistical significance. Therefore, a reanalysis of the 72-h FDVE versus corn oil and the 24-versus 72-h data was performed with omission of the one outlier, described above, in the 72-h FDVE Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 group. After omitting this sample, the 72 h FDVE data set had 175 genes that were differentially expressed 1.5-fold versus controls, with 91 upregulated and 84 downregulated, and the 24- versus 72-h data set had 104 differentially expressed genes (56 upregulated and 48 downregulated). Of the 517 genes that varied significantly at 24 h, 75 are significantly different at 72 h versus 24 h, indicating their return to control, while 175 remained significantly different versus control. Further analysis of the 24 h data set was performed. Differentially regulated genes were grouped into major func- tional categories. To limit the size of the candidate gene set, inclusion in the table was based on a 3-fold upregulation (Table 3) or downregulation (Table 4) at 24 h. For comparison, fold-changes at 72 h are also included. Predominant classes of upregulated genes included those involved in apoptosis, oxidative stress, inflammatory response, and regeneration and repair; downregulated genes were generally associated with transporters and enzymes of intermediary metabolism. Genes associated with phase 1 xenobiotic metabolism (cytochrome P450) were downregulated, while those associated with phase 2 metabolism were upregulated. Among the quantitatively most upregulated genes were kidney injury molecule, angiopoietin- like protein 4, osteopontin, glycoprotein nmb, ubiquitin hydrolase, clusterin, and TNF receptor. Several other genes, yet unidentified, were also significantly upregulated. Additional classification of groups of differentially ex- pressed genes based on the was performed using the GenMAPP program (Doniger et al., 2003). Since the rat genome is less annotated than some others, only 236 of the 517 genes were recognized by GenMAPP. The top 20 Gene Onotology categories for up- and downregulated genes at 24 h, ranked according to z score, are listed in Table 5. Microarray results were confirmed for a subset of genes using quantitative real-time PCR (Table 6). A selected set of

FIG. 1. Hierarchical cluster diagram of renal gene expression profiles following administration of FDVE or corn oil (control). The statistical criteria was a 1.5-fold or greater change in expression and p<0.05 at 24 h, which yielded 544 genes. The colored bar at the top indicates animal treatment: corn oil control (orange), 24 h post FDVE (green), 72 h post FDVE (blue). Red and green indicate increased and decreased gene expression relative to control, respectively. 424 KHARASCH ET AL.

TABLE 3 Upregulation of Rat Kidney Gene Expression by FDVE

Fold change in expression

Functional category Gene Name Symbol 24 h 72 h 72 hra

Regeneration/Repair Kidney injury molecule KIM1/TIM1/HAVCR1 254 98 148c Angiopoietin-like protein 4 ANGPTL4 22 3.0 4.3 Osteopontin (Secreted phosphoprotein 1) OPN/SPP1 10 9.4 12.0c Glycoprotein (transmembrane) nmb GPNMB 10 5.3 8.2c Clusterin (Apolipoprotein J) CLU/apoJ/SGP2 5.8 5.9 8.2c Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 Tissue inhibitor of metalloproteinase 1 TIMP1 5.2 3.0 3.6b Galectin 3 LGALS3/MAC2 5.0 3.1 3.8c Annexin A7 ANXA7 3.6 1.5b 1.7b Fibrinogen, B beta polypeptide FGB 3.9 3.7 5.2c Fibrinogen, alpha polypeptide FGA 3.6 2.7 3.0 Minichromosome maintenance 6 MCMD6/MCM6 3.3 2.7 3.6 Inflammatory response Osteopontin (secreted phosphoprotein 1) SPP1/OPN 10 9.4 12.0c Chemokine (C-C motif) ligand 2 CCL2/MCP1 5.4 4.2 4.8c Interleukin 1 receptor-like 1 IL1RL1/T1/ST2 5.3 1.4b 1.5b Kininogen 1 KNG1 5.2 2.0 2.6c CD44 antigen CD44 3.6 5.6 7.6c Chemokine (C-X-C motif) ligand 10 CXCL10 3.1 3.5 3.8 Apoptosis Ubiquitin carboxy-terminal hydrolase L1 UCHL1 11 1.4b 1.7b Tumor necrosis factor receptor superfamily, member 12a TNFRSF12/DR3 6.4 3.9 5.0c Clusterin CLU/apoJ/SGP2 4.7 1.9 2.0 Xenobiotic metabolism UDP-glucuronosyltransferase UGT 35 1.0b 1.0b ATP-binding cassette, subfamily C, (CFTR/MRP) member 2 ABCC2/MRP2 9.8 2.1b 2.6b NAD(P)H dehydrogenase, quinone 1 NQO1 6.2 1.0b 1.0b Aldo-keto reductase family 1, member B8 AKR/FR1 6.1 1.6b 1.9b Glutathione S-transferase, mu 1 GSTM1 5.2 1.6b 1.9b Similar to arsenite inducible RNA associated protein AIRAP 4.5 1.1b 1.1b Glutathione S-transferase, Yc2 subunit YC2 4.0 1.6b 1.8b Oxidative stress Glutathione peroxidase 2 GPX2 6.9 2.0b 2.5b,c Heme oxygenase 1 HO1 6.3 1.4b 1.7b NAD(P)H dehydrogenase, quinone 1 NQO1 6.2 1.0b 1.0b Heat shock 70kD protein 1b (HSP70-2, HSP72) HSPA1B 5.7 1.4 1.0b Heat shock 70kD protein 1a (HSP70-1, HSP72) HSPA1A 5.1 1.1b 1.1b Activating transcription factor 3 ATF3 4.1 2.2b 2.4 Heat shock 27kD protein 1 (HSP27) HSPB1 3.3 1.2b 1.2b Immune response Putative interferon stimulated gene 12(b) ISG12(b) 3.6 1.4b 1.5b Transmembrane 4 superfamily, member 3 TM4SF3 3.3 3.3 3.8c Cholesterol metabolism HMG-Coenzyme A synthase 2 HMGCS2 4.5 1.9 2.6 Squalene epoxidase SQLE 3.2 1.1b 1.1b Cytoskeleton/Cell morphology Annexin 2 ANXA2 3.9 2.8 3.5c S-100 calcium binding protein A10 (calpactin) S100A10 3.1 2.0 2.1c S-100 calcium binding protein A6 (calcyclin) S100A6 3.0 3.0 3.6 Renal regulation Solute carrier family 34 (sodium phosphate), member 2 SLC34A2 5.4 4.2 4.4c Guanylate cyclase activator 2A GUXA2A, GCAP2 3.2 1.5 1.6 Miscellaneous Poliovirus receptor PVR 5.2 3.6 4.1 Unknown proteins 16 1.0b 1.0b Affymetrix probe 1368578_at 15 2.3b 2.5b HALOALKENE NEPHROTOXICITY MICROARRAY ANALYSIS 425

TABLE 3—Continued

Fold change in expression

Functional category Gene Name Symbol 24 h 72 h 72 hra

cDNA clone MGC:109086 12 3.0b 3.6b Affymetrix probe 1372510_at 4.6 1.4b 1.4 Affymetrix probe 137497_at 3.1 1.4 1.8 Affymetrix probe 1370988_at

Note. Inclusion in the table was based on 3-fold upregulation at 24 h compared to control ( p<0.05, Benjamini-Hochberg correction). Italicized genes are listed twice. Included for comparison are 72-h values. aReanalyzed without outlier, see text. Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 b72-h FDVE significantly different versus 24-h FDVE ( p<0.05, Benjamini-Hochberg correction). c72-h FDVE significantly different versus control ( p<0.05, Benjamini-Hochberg correction). up- and downregulated genes, of varying fold change, was and at lower concentrations (Chen et al., 2001; Lash et al., 2001, evaluated. RT-PCR confirmed the upregulation of kidney injury 2003). Additional studies, evaluating earlier time points, are molecule, HSP70, clusterin, osteopontin, and TNF receptor 12 needed to resolve the role of apoptosis in FDVE toxicity. Genes at 24 h after FDVE, which were identified by microarray. subtending inflammation, regeneration, and repair were already Consistent with the microarray results, kidney injury molecule upregulated at 24 h and were generally persistently elevated was the most quantitatively upregulated. RT-PCR also con- (Kim-1, osteopontin, GPNMB, clusterin, galectin, fibrinogen, firmed the downregulation of organic anion transporter 1 calpectin, calcyclin), if not even more greatly expressed (CD44) (solute carrier family 21, member 1), histamine N-methyl- at 72 h, while others had returned to near control values. This is transferase, chronic renal failure gene, and regucalcin at 24 h, consistent with the histologic evidence of regeneration which which continued to be downregulated at 72 h. was observed at 72 h. It is also consistent with the early occurrence of renal tissue repair observed after exposure to the S-conjugates of other nephrotoxic haloalkenes (Vaidya et al., DISCUSSION 2003a,b). Tissue repair processes are a major determinant of the progression of renal injury after S-conjugate toxicity and The results of this investigation demonstrate that a dose of influence whether acute injury is followed by restoration of FDVE which results in structural and functional evidence of renal tubular structural and functional integrity or by irreversible nephrotoxicity also causes a marked alteration in renal gene massive necrosis, renal failure, and death (Vaidya et al., expression profiles in Fischer 344 rats. This is the first 2003a,b). Hence, expression of regeneration and repair genes investigation profiling renal gene expression following a neph- signals not only renal injury, but also renal repair. Upregulation rotoxic fluoroalkene. Only one other investigation has reported of inflammatory genes may be predictive of downstream events; the renal expression profile in rats following a nephrotoxic however, additional studies evaluating later time points would haloalkene, specifically the chloroalkene hexachlorobutadiene be needed to for verification. In contrast to upregulation, (Thukral et al., 2005). downregulation of gene expression after FDVE generally It is likely that the expression pattern of genes upregulated persisted through the 72-h observation period. at 24 and 72 h after FDVE administration reflects both pathways The renal expression profile following FDVE administration of toxification (apoptosis, oxidative stress, and/or necrosis) and bears strong similarity to that following exposure to several response to injury (inflammation, regeneration, and repair). other nephrotoxins, which has been observed both in rodents Upregulation of genes subtending apoptosis and especially and in nonhuman primates. Recent investigations have dem- oxidative stress (UCHL1, HSPs, heme oxygenase, glutathione onstrated a suite of renal genes which are modulated in vivo in peroxidase, NADPH dehydrogenase) was generally greater at response to injury caused by a diverse group of model 24 h compared with 72 h, consistent with greater impairment of nephrotoxins, including cisplatin (Amin et al., 2004; Huang renal function at this time point and a potential association with et al., 2001), gentamycin (Amin et al., 2004; Davis et al., mechanisms of FDVE toxification. The relatively small number 2004), puromycin (Amin et al., 2004; Thukral et al., 2005), of apoptotic genes upregulated at 24 h may be interpreted as ochratoxin (Lu¨he et al., 2003), and mercuric chloride, a relative lack of involvement of apoptosis in FDVE toxicity, or 2-bromoethylamine hydrobromide, hexachlorobutadiene, mito- an indication that these are early response genes whose mycin, and amphotericin (Thukral et al., 2005). Genes highly upregulation is diminished by 24 h. Previous investigations (more than five-fold) and significantly upregulated rather with the cysteine S-conjugates of the nephrotoxin trichloroeth- consistently include glutathione transferase pi, Kim-1, osteo- ylene showed that apoptosis occurred much earlier than necrosis pontin, tissue inhibitor of metalloproteinase 1 (TIMP1), 426 KHARASCH ET AL.

TABLE 4 FDVE 24hr vs corn oil FDVE 72hr vs corn oil Downregulation of Rat Kidney Gene Expression by FDVE

Fold change in 0 expression 420 0

Function category Name Symbol 24 h 72 h 72 ha 0 Xenobiotic metabolism 0 Cytochrome P4502C CYP2C 19 9.7 16.7 97 Transporters Cationic amino acid SLC7a12/ASC2 13 13 25.5c

transporter (predicted) Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 28 Organic anion transporter SLC21A13/OATP5 7.1 3.7 5.2c protein 5 Organic anion transporter SLC21A1/OATP1 5.6 3.4 5.0c protein 1 c Organic anion transporter 19 SLC22A19/OAT5 5.0 2.8 3.6 FDVE 24hr vs FDVE 72hr Organic cation transporter 13 SLC22A13/OCTL1/ 3.8 2.8 3.6c (predicted) OCTL3/ORCTL3 FIG. 2. Venn diagram showing number of genes differentially expressed 1.5-fold or greater and p<0.05; p values were adjusted for multiple hypothesis Intermediary metabolism testing by the method of Benjamini-Hochberg. Histamine N-methyltransferase HNMT 7.3 2.7 3.7c Thioether S-methyltransferase — 3.8 1.5 1.6 (predicted) clusterin, vimentin, and fibrinogen alpha, and those also c Ornithine aminotransferase OAT 3.6 1.9 2.3 upregulated but less so (typically two- to five-fold) include Ornithine decarboxylase 1 ODC 3.2 2.5 3.2c Cysteine dioxygenase 1 CDO1 3.1 2.5 2.5c heme oxygenase, UDP-glucuronosyltransferase, glutathione Calcium Homeostasis synthetase, insulin-like growth factor binding protein 1 Regucalcin RGN/SMP30 4.6 1.9 2.4c (IGFBP1), and annexinA7. Genes significantly downregulated Cell signaling more than five-fold generally include regucalcin and several Lysophospholipase — 6.2 2.8 3.5c organic ion transporters, including organic anion transporter Guanylate cyclase 1, GUCY1B2 3.5 2.8 3.6c protein 1(SLC21A1/OATP1), organic anion transporter family soluble, beta 2 member 1A2 (SLC21A7/OATP3), and solute carrier family 22 Immune response (organic cation transporter) member 2 (SLC22A2/OCT2), and Group specific component/ GC/DBP 3.4 2.4 3.1c Vitamin D binding protein those also downregulated but less so (typically two- to five- Cell cycle fold) include solute carrier family 22 member 6 (organic anion Maternal embryonic leucine MELK/PEG3/ 3.3 1.7 2.2c transporter 1, SLC22A6/OAT1) and HSP27. Like those zipper kinase (predicted) MDK 38 modulated after FDVE, these also tend toward genes involved Miscellaneous in tissue regeneration and inflammation, which frequently Chronic renal failure gene CRFG 5.1 2.9 3.8 occur as a consequence of necrosis, and which have been b b Iodothyronine deiodinase DIO1 4.1 1.6 1.7 associated with tissue repair after various forms of acute renal Pregnancy induced growth OKL38 3.9 2.4 3.0c inhibitor failure (Huang et al., 2001; Thukral et al., 2005). One recent Alpha-synuclein SNCA 3.5 1.8 23 investigation of rat gene expression evaluating dose-response Unknown genes 6.2 3.3 4.3c effects of several model nephrotoxins showed some clustering Affymetrix probe 1372868_at 5.5 1.2b 1.3b of expression profiles according to the degree of injury, which Affymetrix probe 1377018_at 3.2 2.1 2.9 was independent of the specific nephrotoxin (Thukral et al., Affymetrix probe 1374963_s_at 3.1 1.1b 1.2b Transcribed : Affy probe 3.1 2.5 3.3c 2005). Milder forms of injury causing degeneration/regenera- 1372911_at tion resulted in upregulation of glutathione S-transferase P1, Transcribed locus: Affy probe Kim-1, osteopontin, and TIMP1 and downregulation of 1379580_at OATP1, while more severe toxicity causing necrosis resulted in upregulation of these and several other genes and down- Note. Inclusion in the table was based on 3 fold downregulation at 24 h regulation of several more transporters (Thukral et al., 2005). compared to control ( p<0.05, Benjamini-Hochberg correction). Italicized genes are listed twice. Included for comparison are 72 h values. While most of the organic ion transporters were downregulated aReanalyzed without outlier, see text. after FDVE, solute carrier family 34 (sodium phosphate), b72-h FDVE significantly different versus 24-h FDVE ( p<0.05, member 2 (SLC34A2) was upregulated, a finding similar to Benjamini-Hochberg correction). that after several other nephrotoxins (Thukral et al., 2005). c p< 72-h FDVE significantly different versus control ( 0.05, Benjamini- Interesting exceptions to the pattern of renal injury response Hochberg correction). genes frequently upregulated after acute renal injury were HALOALKENE NEPHROTOXICITY MICROARRAY ANALYSIS 427

TABLE 5 TABLE 6 Top 20 Modulated Gene Ontology Categories 24 h after FDVE RT-PCR Analysis of Gene Expression

GO identification GO Name Z score Fold change in expression

GO categories for upregulated genes corn oil FDVE FDVE Top five nonredundant controls 24 h 72 h biological processes 16125 sterol metabolism 7.83 kidney injury molecule KIM 1.0 ± 1.3 56.5 ± 15.6a 25.0 ± 19.5a,b 6066 alcohol metabolism 6.52 heat shock 70kD HSP701B 1.0 ± 0.4 5.5 ± 2.7 4.6 ± 7.2 9067 aspartate family amino acid 5.47 protein 1B (HSP70-2) biosynthesis clusterin CLU 1.0 ± 0.6 5.0 ± 1.4a 4.6 ± 2.8a 42267 natural killer cell mediated 5.47 heat shock 70kD protein HSP70i 1.0 ± 0.3 4.0 ± 0.7a 1.4 ± 1.4b Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 cytolysis 1A inducible (HSP70-1) 16096 polyisoprenoid metabolism 5.47 osteopontin (secreted SPP1 1.0 ± 0.3 3.4 ± 0.3a 4.1 ± 1.9a Top five nonredundant phosphoprotein 1) a a molecular functions TNF receptor superfamily, TNFrsf12A 1.0 ± 0.5 3.0 ± 0.5 2.6 ± 1.3 16744 transferase activity (transferring 7.74 member 12a a b aldehyde or ketonic groups) ATP-binding cassette sub- ABCC2 1.0 ± 0.3 1.8 ± 0.7 0.9 ± 0.1 4743 pyruvate kinase activity 7.74 family C (CFTR/MRP) a a 4421 hydroxymethylglutaryl-CoA 7.74 solute carrier SCF2 1.0 ± 0.6 0.1 ± 0.1 0.2 ± 0.2 synthase activity family 21, member 1 a a 46912 transferase activity 6.22 histamine HNMT 1.0 ± 0.3 0.1 ± 0.1 0.4 ± 0.6 (transferring acyl groups) N-methyltransferase a a,b 4090 carbonyl reductase (NADPH) activity 5.47 chronic renal failure gene CRFG 1.0 ± 0.8 0.1 ± 0.0 0.3 ± 0.2 a a GO categories for downregulated genes regucalcin RGN 1.0 ± 0.5 0.2 ± 0.1 0.4 ± 0.2 Top five nonredundant biological processes Note. Results are the ratio of expression levels compared with the mean 6519 amino acid and derivative metabolism 9.89 control value (mean ± SD, n ¼ 5). a 9308 amine metabolism 9.70 Significantly different versus control ( p<0.05). b 6082 organic acid metabolism 7.13 Significantly different versus 24 h ( p<0.05). 6790 sulfur metabolism 7.12 50 urea cycle 7.00 Top five nonredundant organic ion transporters), others at least 10-fold (osteopontin, molecular functions ANGPTL4, GPNMB, UCHL1, organic ion transporters), and 16624 oxidoreductase activity (acting on 8.32 the aldehyde donor) Kim-1 was upregulated approximately 100-fold. These results 8533 astacin activity 8.32 suggest that one or more of these modulated genes may serve as 4238 meprin A activity 8.32 an early and sensitive biomarker for FDVE toxicity. Charac- 9374 biotin binding 6.69 teristic or predictive candidate biomarkers for nephrotoxicity 16885 ligase activity (forming 6.69 identified most consistently in other investigations include the carbon–carbon bonds) genes for Kim-1, osteopontin, and clusterin, and others Note. Top 20 modulated Gene Ontology categories 24 h after FDVE were identified by some as potentially useful include IGFBP-1, identified using GenMAPP and MAPPfinder based on Gene Ontology (GO) alpha-fibrinogen, GSTalpha, lipocalin, TNF receptor 12a, and annotation. GO items are ranked by z score. A positive z score indicates there several transporters (SLC21a2, SLC15, SLC34a2) (Amin et al., are more genes in this GO category than would be expected by chance. A z score of 1.96 (were the model perfect) would correspond to a p value of 0.05. 2004; Davis et al., 2004; Thukral et al., 2005). Kim-1 was the gene most consistently and quantitatively upregulated in response to FDVE, as well as by several other vimentin and insulin-like growth factor binding protein 1 nephrotoxins. FDVE stimulated Kim-1 gene expression (IGFbp1), which were only increased by FDVE 1.4- and 1.9- 254-fold over control at 24 h, and at 72 h this gene was still fold, respectively, at 24 h (not shown). upregulated more than 100-fold. In a nonhuman primate model Expression profiling was a much more sensitive indicator of of antibiotic nephrotoxicity, quantitative analysis showed that FDVE renal effects than more conventional measures of renal Kim-1 had the greatest increase in gene expression among function. Whereas similar (0.25–0.3 mmol/kg) FDVE doses various biomarkers (Davis et al., 2004). Kim-1 is a type 1 caused negligible or small (two-fold) changes in serum creat- membrane glycoprotein with a large ectodomain containing inine and urea nitrogen, and urine protein excretion (Kharasch immunoglobulin and glycosylated mucin subdomains and et al., 1997; Sheffels et al., 2004), expression of several genes a short cytoplasmic tail. Kim-1 constitutive expression is low, was modulated at least 5-fold (clusterin, TIMP1, TNFRSF12, but upregulated in association with dedifferentiated and CFTR/MRP, heme oxygenase, histamine N-methyltransferase, regenerating tubular epithelial cells (Han et al., 2002). 428 KHARASCH ET AL.

Increased Kim-1 protein expression in proximal tubule epithe- N-methyltransferase by metoprine resulted in brisk diuresis lial cells of kidneys from rats treated with the haloalkene and decreased urine osmolality, by an unknown mechanism conjugate S-(1,1,2,2-tetrafluoroethyl)-l-cysteine (TFEC) was (Lecklin et al., 1999). It is unknown whether metoprine shown by immunoblotting, immunofluorescence, and immu- resulted in renal cellular toxicity, or whether histamine-N- nohistochemistry, and the Kim-1 soluble protein ectodomain methyltransferase plays a role in the diuresis and decreased and fragments were demonstrated in urine by immunoblotting urine osmolality following FDVE. Another interesting obser- (Ichimura et al., 2004). The Kim-1 ectodomain is cleaved by vation is the identification of meprin activity, and that of its matrix metalloproteinases and shed into cell culture media and parent family astacin, as downregulated Gene Ontology func- into urine following experimental nephrotoxicity and ischemia/ tional categories. Microarray analysis showed that meprin was reperfusion in rats and acute tubular necrosis in humans (Bailly downregulated 2.8-fold (not shown). Meprins are metallopro- et al., 2002; Han et al., 2002; Ichimura et al., 1998, 2004). The teases highly expressed at the brush border membrane of renal specific role of Kim-1 is unknown, but its expression in proximal tubular cells and cleave a wide variety of substrates, Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 dedifferentiated cells in response to several renal ischemic including cytokines such as osteopontin (Bond and Beynon, and toxic insults suggests that it is important in repair processes 1995; Bond et al., 2005). Meprins can be detrimental when involving dedifferentiation, migration, proliferation, and res- there is renal tissue damage and are downregulated in renal toration of cellular architecture and function (Ichimura et al., injury and in situations leading to cell death through apoptosis 2004). Early expression of Kim-1, and its detectability at or necrosis. Upregulation of TIMP1, chemokines, and osteo- degrees of injury at which conventional renal markers do not pontin observed directly on the microarrays is consistent with change, has resulted in the conclusion that Kim-1 may be the Gene Ontology identification of meprin downregulation. a sensitive biomarker for mild–moderate renal injury and The role of meprins and metalloproteases in the pathogenesis repair. The results in the present investigation are consistent of or response to FDVE nephrotoxicity merits evaluation. with that hypothesis. It remains unknown whether Kim-1 Downregulation of regucalcin by FDVE is similar to that re- protein is expressed in kidneys or excreted in urine following ported recently in response to another nephrotoxin (Thukral FDVE nephrotoxicity. et al., 2005). The physiologic role of regucalcin, found in Osteopontin expression is upregulated following several renal cortex but not medulla, is incompletely elucidated, but forms of toxic renal injury, not limited to the proximal tubule this protein may play a role in regulating intracellular calcium (Amin et al., 2004; Basile et al., 2005; Davis et al., 2004). It is homeostasis, suppress DNA synthesis, participate in protein a chemoattractant for macrophages and also inhibits apoptosis. degradation, and suppress apoptosis and cell death, and sup- Osteopontin is constitutively secreted in urine and has been pression of expression has been suggested to cause renal tubular shown to be secreted in increased amounts in at least one model cell dysfunction (Nakagawa and Yamaguchi, 2005; Yamaguchi, of nephrotoxicity (Khan et al., 2002). Clusterin gene expres- 2005). The toxicologic implication of regucalcin downregulation sion is also upregulated in response to numerous forms of renal by FDVE requires further investigation. Downregulation of toxicity, remodeling, and diseases (Amin et al., 2004; Davis chronic renal failure gene (nucleolar GTP-binding protein 1) et al., 2004). The role of clusterin in renal injury and was confirmed by RT-PCR; however, the mechanism and im- regeneration is unknown, but clusterin is detectable in urine plications of this modulation are unknown. before changes in serum creatinine after renal injury and has In summary, FDVE nephrotoxicity was associated with rapid been suggested as a potential biomarker. Upregulation of and brisk changes in gene expression. Upregulation of genes osteopontin and clusterin gene expression following FDVE involved in apoptosis, oxidative stress, and inflammatory nephrotoxicity is consistent with a role as a potential biomarker. response may provide potential insights into the mechanism Upregulation of sterol metabolism genes was observed of FDVE toxification. A suite of upregulated genes, including directly in the microarray analysis and was reflected also in kidney injury molecule, osteopontin, clusterin, tissue inhibitor the Gene Ontology categories for biological processes (sterol of metalloproteinase 1, and TNF receptor 12, have also been metabolism) and molecular functions (hydroxymethylglutaryl- associated with other forms of nephrotoxicity. Gene expression CoA synthase activity). Renal cortical cholesterol accumula- profiling may provide potential biomarkers for FDVE nephro- tion is a prominent feature after acute renal toxicity, postulated toxicity which are more sensitive than conventional measures as an adaptive response, and due in part to increased renal of renal function. Gene expression profiling may provide tubular cell cholesterol synthesis (Zager et al., 2002). potential biomarkers for evaluating the clinical effects of Less well understood are the mechanisms and implications FDVE exposure in humans. of gene downregulation following FDVE and other nephrotoxic injuries. Downregulation of several organic ion transporters by FDVE is consistent with previous effects of several other SUPPLEMENTARY DATA nephrotoxins (Amin et al., 2004; Davis et al., 2004; Thukral et al., 2005). Downregulation of histamine-N-methyltransferase Supplementary data are available online at http://toxsci. expression is an interesting observation. Inhibition of histamine- oxfordjournals.org/. HALOALKENE NEPHROTOXICITY MICROARRAY ANALYSIS 429

ACKNOWLEDGMENTS Chen, Y., Cai, J., Anders, M. W., Stevens, J. L., and Jones, D. P. (2001). Role of mitochondrial dysfunction in S-(1,2-dichlorovinyl)-L-cysteine-induced apo- The authors appreciate the excellent assistance of Dr. Mette Peters, UW ptosis. Toxicol. Appl. Pharmacol. 170, 172–180. Center for Expression Arrays. Supported by NIH grants R01DK53765, Conzen, P. F., Kharasch, E. D., Czerner, S. F., Artru, A. A., Reichle, F. M., 5U24DK058813 and NIEHS P30ES07033 to the Center for Ecogenetics and Michalowski, P., Rooke, G. A., Weiss, B. M., and Ebert, T. J. (2002). Low- Environmental Health. flow sevoflurane compared with low-flow isoflurane anesthesia in patients Conflict of Interest Statement: No author has any conflict of interest. Evan with stable renal insufficiency. Anesthesiology 97, 578–584. Kharasch formerly served as an occasional ad hoc consultant to Abbott Davis, J. W., 2nd, Goodsaid, F. M., Bral, C. M., Obert, L. A., Mandakas, G., Laboratories, which markets sevoflurane. Garner, C. E., 2nd, Collins, N. D., Smith, R. J., and Rosenblum, I. Y. (2004). Quantitative gene expression analysis in a nonhuman primate model of antibiotic-induced nephrotoxicity. Toxicol. Appl. Pharmacol. 200, 16–26. REFERENCES Devarajan, P., Mishra, J., Supavekin, S., Patterson, L. T., and Steven Potter, S.

(2003). Gene expression in early ischemic renal injury: Clues towards Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 Altuntas, T. G., and Kharasch, E. D. (2001). Glutathione S-conjugation of the pathogenesis, biomarker discovery, and novel therapeutics. Mol. Genet. sevoflurane degradation product, fluoromethyl-2,2-difluoro-1-(trifluoromethyl)- Metab. 80, 365–376. vinyl ether (Compound A) in human liver, kidney, and blood in vitro. Toxicol. Appl. Pharmacol. 177, 85–93. Diaz, D., Krejsa, C. M., White, C. C., Keener, C. L., Farin, F. M., and Kavanagh, T. J. (2001). Tissue specific changes in the expression of Altuntas, T. G., and Kharasch, E. D. (2002). Biotransformation of L-cysteine glutamate-cysteine ligase mRNAs in mice exposed to methylmercury. S-conjugates and N-acetyl-L-cysteine S-conjugates of the sevoflurane de- Toxicol. Lett. 122, 119–129. gradation product fluoromethyl-2,2-difluoro-1-(trifluoromethyl)vinyl ether (Compound A) in human kidney in vitro: Interindividual variability in N- Doniger, S. W., Salomonis, N., Dahlquist, K. D., Vranizan, K., Lawlor, S. C., acetylation, N-deacetylation, and b-lyase-catalyzed metabolism. Drug and Conklin, B. R. (2003). MAPPFinder: Using Gene Ontology and Metab. Dispos. 30, 148–154. GenMAPP to create a global gene-expression profile from microarray data. Genome Biol. 4, R7. Altuntas, T. G., Park, B. S., and Kharasch, E. D. (2004). Sulfoxidation of cysteine and mercapturic acid conjugates of the sevoflurane degradation Frink, E. J., Jr., Malan, T. P., Morgan, S. E., Brown, E. A., Malcomson, M., product fluoromethyl-2,2-difluoro-1-(trifluoromethyl)vinyl ether (compund A). Gandolfi, A. J., and Brown, B. R., Jr. (1992). Quantification of the Chem. Res. Toxicol. 17, 435–445. degradation products of sevoflurane in two CO2 absorbents during low-flow anesthesia in surgical patients. Anesthesiology 77, 1064–1069. Altuntas, T. G., Zager, R. A., and Kharasch, E. D. (2003). Cytotoxicity of S- Fukuda, H., Kawamoto, M., Yuge, O., and Fujii, K. (2004). A comparison of the conjugates of the sevoflurane degradation product fluoromethyl-2,2-difluoro- effects of prolonged (>10 hour) low-flow sevoflurane, high-flow sevoflurane, 1-(trifluoromethyl)vinyl ether (compound A) in a human proximal tubular and low-flow isoflurane anaesthesia on hepatorenal function in orthopaedic cell line. Toxicol. Appl. Pharmacol. 193, 55–65. patients. Anaesth. Intensive Care. 32, 210–218. Amin, R.P., Vickers, A. E., Sistare,F., Thompson, K. L., Roman, R.J., Lawton, M., Gentleman, R. C., Carey, V. J., Bates, D. M., Bolstad, B., Dettling, M., Dudoit, Kramer, J., Hamadeh, H. K., Collins, J., Grissom, S., et al. (2004). S., Ellis, B., Gautier, L., Ge, Y., Gentry, J., Hornik, K., et al. (2004). Identification of putative gene based markers of renal toxicity. Environ. Bioconductor: Open software development for computational biology and Health Perspect. 112, 465–479. bioinformatics. Genome Biol. 5, R80. Anders, M. W. (2004). Glutathione-dependent bioactivation of haloalkanes and Goodsaid, F. M. (2004). Identification and measurement of genomic bio- haloalkenes. Drug Metab. Rev. 36, 583–594. markers of nephrotoxicity. J. Pharmacol. Toxicol. Methods 49, 183–186. Anders, M. W. (2005). Formation and toxicity of anesthetic degradation Han, W. K., Bailly, V., Abichandani, R., Thadhani, R., and Bonventre, J. V. Annu. Rev. Pharmacol. Toxicol. 45, products. 147–176. (2002). Kidney injury molecule-1 (KIM-1): A novel biomarker for human Bailey, W. J., and Ulrich, R. (2004). Molecular profiling approaches for renal proximal tubule injury. Kidney Int. 62, 237–244. identifying novel biomarkers. Expert Opin. Drug Saf. 3, 137–151. Hanaki, C., Fujii, K., Morio, M., and Tashima, T. (1987). Decomposition of Bailly, V., Zhang, Z., Meier, W., Cate, R., Sanicola, M., and Bonventre, J. V. sevoflurane by soda lime. Hiroshima J. Med. Sci. 36, 61–67. (2002). Shedding of kidney injury molecule-1, a putative adhesion protein Hayes, K. R., and Bradfield, C. A. (2005). Advances in toxicogenomics. Chem. involved in renal regeneration. J. Biol. Chem. 277, 39739–39748. Res. Toxicol. 18, 403–414. Basile, D. P., Fredrich, K., Alausa, M., Vio, C. P., Liang, M., Rieder, M. R., Hewitt, S. M., Dear, J., and Star, R. A. (2004). Discovery of protein biomarkers Greene, A. S., and Cowley, A. W., Jr. (2005). Identification of per- for renal diseases. J. Am. Soc. Nephrol. 15, 1677–1689. sistently altered gene expression in the kidney after functional recovery Higuchi, H., Sumita, S., Wada, H., Ura, T., Ikemoto, T., Nakai, T., Kanno, M., from ischemic acute renal failure. Am. J. Physiol. Renal. Physiol. 288, and Satoh, T. (1998). Effects of sevoflurane and isoflurane on renal F953– F963. function and on possible markers of nephrotoxicity. Anesthesiology 89, Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: A 307–322. practical and powerful approach to multiple testing. J. Royal Stat. Soc. Ser. B Huang, Q., Dunn, R. T., 2nd, Jayadev, S., DiSorbo, O., Pack, F. D., Farr, S. B., 57, 289–300. Stoll, R. E., and Blanchard, K. T. (2001). Assessment of cisplatin-induced Bito, H., and Ikeda, K. (1996). Renal and hepatic function in surgical patients nephrotoxicity by microarray technology. Toxicol. Sci. 63, 196–207. Anesth. Analg. 82, after low-flow sevoflurane or isoflurane anesthesia. Ichimura, T., Bonventre, J. V., Bailly, V., Wei, H., Hession, C. A., Cate, R. L., 173–176. and Sanicola, M. (1998). Kidney injury molecule-1 (KIM-1), a putative Bond, J. S., and Beynon, R. J. (1995). The astacin family of metalloendo- epithelial cell adhesion molecule containing a novel immunoglobulin peptidases. Protein Sci. 4, 1247–1261. domain, is up-regulated in renal cells after injury. J. Biol. Chem. 273, Bond, J. S., Matters, G. L., Banerjee, S., and Dusheck, R. E. (2005). Meprin 4135–4142. metalloprotease expression and regulation in kidney, intestine, urinary tract Ichimura, T., Hung, C. C., Yang, S. A., Stevens, J. L., and Bonventre, J. V. infections and cancer. FEBS Lett. 579, 3317–3322. (2004). Kidney injury molecule-1: A tissue and urinary biomarker for 430 KHARASCH ET AL.

nephrotoxicant-induced renal injury. Am. J. Physiol. Renal Physiol. 286, Kharasch, E. D., Thorning, D. T., Garton, K., Hankins, D. C., and Kilty, C. G. F552–F563. (1997). Role of renal cysteine conjugate b-lyase in the mechanism of Irizarry, R. A., Hobbs, B., Collin, F., Beazer-Barclay, Y. D., Antonellis, K. J., compound A nephrotoxicity in rats. Anesthesiology 86, 160–171. Scherf, U., and Speed, T. P. (2003). Exploration, normalization, and Kramer, J. A., Pettit, S. D., Amin, R. P., Bertram, T. A., Car, B., Cunningham, M., summaries of high density oligonucleotide array probe level data. Bio- Curtiss, S. W., Davis, J. W., Kind, C., Lawton, M., et al. (2004). Overview on statistics 4, 249–264. the application of transcription profiling using selected nephrotoxicants for toxicology assessment. Environ. Health Perspect. 112, 460–464. Iyer, R. A., and Anders, M. W. (1996). Cysteine conjugate b-lyase-dependent biotransformation of the cysteine S-conjugates of the sevoflurane degrada- Lash, L. H., Hueni, S. E., and Putt, D. A. (2001). Apoptosis, necrosis, and tion product compound A in human, nonhuman primate, and rat kidney cell proliferation induced by S-(1,2-dichlorovinyl)-L-cysteine in primary cytosol and mitochondria. Anesthesiology 85, 1454–1461. cultures of human proximal tubular cells. Toxicol. Appl. Pharmacol. 177, 1–16. Iyer, R. A., and Anders, M. W. (1997). Cysteine conjugate b-lyase-dependent biotransformation of the cysteine S-conjugates of the sevoflurane degrada- Lash, L. H., Parker, J. C., and Scott, C. S. (2000). Modes of action of

tion product 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene (compound trichloroethylene for kidney tumorigenesis. Environ. Health. Perspect. Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021 A). Chem. Res. Toxicol. 10, 811–819. 108(Suppl. 2), 225–240. Iyer, R. A., Baggs, R. B., and Anders, M. W. (1997). Nephrotoxicity of the Lash, L. H., Putt, D. A., Hueni, S. E., Krause, R. J., and Elfarra, A. A. (2003). glutathione and cysteine S-conjugates of the sevoflurane degradation product Roles of necrosis, apoptosis, and mitochondrial dysfunction in S-(1,2- 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene (compound A) in male dichlorovinyl)-L-cysteine sulfoxide-induced cytotoxicity in primary cultures Fischer 344 rats. J. Pharmacol. Exp. Ther. 283, 1544–1551. of human renal proximal tubular cells. J. Pharmacol. Exp. Ther. 305, 1163–1172. Iyer, R. A., Frink, E. J., Jr., Ebert, T. J., and Anders, M. W. (1998). Lecklin, A., Eriksson, L., Leppaluoto, J., Tarhanen, J., and Tuomisto, L. (1999). Cysteine conjugate b-lyase-dependent metabolism of compound A (2- Metoprine-induced thirst and diuresis in Wistar rats. Acta Physiol. Scand. [fluoromethoxy]-1,1,3,3,3-pentafluoro-1-propene) in human subjects anes- 165, 325–333. thetized with sevoflurane and in rats given compound A. Anesthesiology 88, 611–618. Lin, Y. S., Dowling, A. L., Quigley, S. D., Farin, F. M., Zhang, J., Lamba, J., Schuetz, E. G., and Thummel, K. E. (2002). Co-regulation of CYP3A4 and Jin, L., Baillie, T. A., Davis, M. R., and Kharasch, E. D. (1995). Nephrotoxicity CYP3A5 and contribution to hepatic and intestinal midazolam metabolism. of sevoflurane compound A [fluoromethyl-2,2-difluoro-1-(trifluoromethyl)- Mol. Pharmacol. 62, 162–172. vinyl ether) in rats: Evidence for glutathione and cysteine conjugate formation and the role of renal cysteine conjugate b-lyase. Biochem. Lu¨he, A., Hildebrand, H., Bach, U., Dingermann, T., and Ahr, H. J. (2003). A Biophys. Res Commun. 210, 498–506. new approach to studying ochratoxin A (OTA)-induced nephrotoxicity: Expression profiling in vivo and in vitro employing cDNA microarrays. Jin, L., Davis, M. R., Kharasch, E. D., Doss, G. A., and Baillie, T. A. (1996). Toxicol. Sci. 73, 315–328. Identification in rat bile of glutathione conjugates of fluoromethyl 2,2- difluoro-1-(trifluoromethyl)vinyl ether, a nephrotoxic degradate of the Morio, M., Fujii, K., Satoh, N., Imai, M., Kawakami, U., Mizuno, T., Kawai, Y., anesthetic agent sevoflurane. Chem. Res. Toxicol. 9, 555–561. Ogasawara, Y., Tamura, T., Negishi, A., et al. (1992). Reaction of sevoflurane and its degradation products with soda lime. Toxicity of the byproducts. Keller, K. A., Callan, C., Prokocimer, P., Delgado-Herrera, M. S., Friedman, Anesthesiology 77, 1155–1164. M. B., Hoffman, G. M., Wooding, W. L., Cusick, P. K., and Krasula, R. W. Nakagawa, T., and Yamaguchi, M. (2005). Overexpression of regucalcin (1995). Inhalation toxicology study of a haloalkene degradant of sevoflurane, suppresses apoptotic cell death in cloned normal rat kidney proximal tubular Compound A (PIFE), in Sprague-Dawley rats. Anesthesiology 83, epithelial NRK52E cells: Change in apoptosis-related gene expression. J. 1220–1232. Cell. Biochem. 96, 1274–1285. Kevil, C. G., Pruitt, H., Kavanagh, T. J., Wilkerson, J., Farin, F., Moellering, D., Obata, R., Bito, H., Ohmura, M., Moriwaki, G., Ikeuchi, Y., Katoh, T., and Darley-Usmar, V. M., Bullard, D. C., and Patel, R. P. (2004). Regulation of Sato, S. (2000). The effects of prolonged low-flow sevoflurane anesthesia on endothelial glutathione by ICAM-1: Implications for inflammation. FASEB renal and hepatic function. Anesth. Analg. 91, 1262–1268. J. 18, 1321–1323. Sheffels, P., Schroeder, J. L., Altuntas, T. G., Liggitt, H. D., and Kharasch, E. D. Khan, S. R., Johnson, J. M., Peck, A. B., Cornelius, J. G., and Glenton, P. A. (2004). Role of cytochrome P4503A in cysteine S-conjugates sulfoxidation (2002). Expression of osteopontin in rat kidneys: Induction during and the nephrotoxicity of the sevoflurane degradation product fluoromethyl- ethylene glycol induced calcium oxalate nephrolithiasis. J. Urol. 168, 2,2-difluoro-1-(trifluoromethyl)vinyl ether (‘‘compound A’’) in rats. Chem. 1173–1181. Res. Toxicol. 17, 1177–1189. Kharasch, E. D., Conzen, P. F., Michalowski, P., Weiss, B. M., Rooke, G. A., Spracklin, D., and Kharasch, E. D. (1996). Evidence for the metabolism of Artru, A. A., Ebert, T. J., Czerner, S. F., and Reichle, F. M. (2003). Safety fluoromethyl-2,2-difluoro-1-(trifluoromethyl)vinyl ether (Compound A), by of low-flow sevoflurane anesthesia in patients. Anesthesiology 99, cysteine conjugate b-lyase in rats. Chem. Res. Toxicol. 9, 696–702. 752–754. Thompson, K. L., Afshari, C. A., Amin, R. P., Bertram, T. A., Car, B., Kharasch, E. D., Frink, E. J., Jr., Artru, A., Michalowski, P., Rooke, G. A., and Cunningham, M., Kind, C., Kramer, J. A., Lawton, M., Mirsky, M., et al. Nogami, W. (2001). Long-duration low-flow sevoflurane and isoflurane (2004). Identification of platform-independent gene expression markers of effects on postoperative renal and hepatic function. Anesth. Analg. 93, cisplatin nephrotoxicity. Environ. Health Perspect. 112, 488–494. 1511–1520. Thukral, S. K., Nordone, P. J., Hu, R., Sullivan, L., Galambos, E., Fitzpatrick, Kharasch, E. D., Hoffman, G. M., Thorning, D., Hankins, D. C., and Kilty, C. G. V. D., Healy, L., Bass, M. B., Cosenza, M. E., and Afshari, C. A. (2005). (1998). Role of the renal cysteine conjugate b-lyase pathway in inhaled Prediction of nephrotoxicant action and identification of candidate toxicity- compound A nephrotoxicity in rats. Anesthesiology 88, 1624–1633. related biomarkers. Toxicol. Pathol. 33, 343–355. Kharasch, E. D., and Jubert, C. (1999). Compound A uptake and metabolism to Tong, Z., and Anders, M. W. (2002). Reactive intermediate formation from the mercapturic acids and 3,3,3-trifluoro-2-fluoromethoxypropanoic acid during 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene (compound A)-derived low-flow sevoflurane anesthesia. Biomarkers for exposure, risk assessment, cysteine S-conjugate S-[2-(fluoromethoxy)-1,1,3,3,3-pentafluoropropyl]-L- and interspecies comparison. Anesthesiology 91, 1267–1278. cysteine in pyridoxal model systems. Chem. Res. Toxicol. 15, 623–628. HALOALKENE NEPHROTOXICITY MICROARRAY ANALYSIS 431

Uttamsingh, V., Iyer, R. A., Baggs, R. B., and Anders, M. W. (1998). Fate and Wu, Z., Irizarry, R., Gentlemen, R., Martinez-Murillo, F., and Spencer, F. toxicity of 2-(fluoromethoxy)-1,1,3,3,3-pentafluoro-1-propene (compound A)- (2004). A model-based background adjustment for oligonucleotide expres- derivedmercapturates inmale, Fischer 344 rats. Anesthesiology89, 1174–1183. sion arrays. J. Am. Statist. Assoc. 99, 909–917. Vaidya, V. S., Shankar, K., Lock, E. A., Bucci, T. J., and Mehendale, H. M. Yamaguchi, M. (2005). Role of regucalcin in maintaining cell homeostasis and (2003a). Renal injury and repair following S-1, 2 dichlorovinyl-L-cysteine function. Int. J. Mol. Med. 15, 371–389. administration to mice. Toxicol. Appl. Pharmacol. 188, 110–121. Zager, R. A., Shah, V. O., Shah, H. V., Zager, P. G., Johnson, A. C., Vaidya, V. S., Shankar, K., Lock, E. A., Bucci, T. J., and Mehendale, H. M. and Hanson, S. (2002). The mevalonate pathway during acute tubular (2003b).Roleoftissue repair insurvival fromS-(1,2-dichlorovinyl)-L-cysteine- injury: Selected determinants and consequences. Am. J. Pathol. 161, induced acute renal tubular necrosis in the mouse. Toxicol. Sci. 74, 215–227. 681–692. Downloaded from https://academic.oup.com/toxsci/article/90/2/419/1658441 by guest on 29 September 2021