Abstract
Mdr1a-, Bcrp-, and Mrp2-knockout rats are a more practical species for absorption, distribution, metabolism, and excretion (ADME) studies than murine models and previously demonstrated expected alterations in the pharmacokinetics of various probe substrates. At present, gene expression and pathology changes were systematically studied in the small intestine, liver, kidney, and brain tissue from male SAGE Mdr1a, Bcrp, and Mrp2 knockout rats versus wild-type Sprague-Dawley controls. Gene expression data supported the relevant knockout genotype. As expected, Mrp2 knockout rats were hyperbilirubinemic and exhibited upregulation of hepatic Mrp3. Overall, few alterations were observed within 112 ADME-relevant genes. The two potentially most consequential changes were upregulation of intestinal carboxylesterase in Mdr1a knockouts and catechol-O-methyltransferase in all tissues of Bcrp knockout rats. Previously reported upregulation of hepatic Mdr1b P-glycoprotein in proprietary Wistar Mdr1a knockout rats was not observed in the SAGE counterpart investigated herein. Relative liver and kidney weights were 22–53% higher in all three knockouts, with microscopic increases in hepatocyte size in Mdr1a and Mrp2 knockout rats and glomerular size in Bcrp and Mrp2 knockouts. Increased relative weight of clearing organs is quantitatively consistent with reported increases in the clearance of drugs that are not substrates of the knocked-out transporter. Overall, SAGE knockout rats demonstrated modest compensatory changes, which do not preclude their general application to study transporter-mediated pharmacokinetics. However, until future studies elucidate the magnitude of functional change, caution is warranted in rare instances of extensive metabolism by catechol-O-methyltransferase in Bcrp knockouts and intestinal carboxylesterase in Mdr1a knockout rats, specifically for molecules with free catechol groups and esters subject to gut-wall hydrolysis.
Introduction
Transporter gene knockout rats generated with zinc finger nucleases garnered considerable interest for mechanistic pharmacokinetic studies (Bundgaard et al., 2012; Chu et al., 2012; Huang et al., 2012; Zamek-Gliszczynski et al., 2012a, 2013). Although detailed absorption, distribution, metabolism, and excretion (ADME) experiments can be conducted in knockout mice (Tian et al., 2007; Zamek-Gliszczynski et al., 2011; Higgins et al., 2012), rats are practically advantageous and more relevant as the most often used nonclinical species in drug development. SAGE Mdr1a, Bcrp, and Mrp2 knockout rats (SAGE Labs, Inc., St. Louis, MO) were recently commercialized in the Sprague-Dawley strain, which is commonly used in toxicology, pharmacokinetic, distribution, and excretion studies. These knockout rats can be used for preclinical investigation of the pharmacokinetic role of the three ABC (ATP-binding cassette) efflux pumps with established relevance to drug disposition (Giacomini et al., 2010; Zamek-Gliszczynski et al., 2012b). Although a proprietary colony of Mdr1a knockout rats has also been reported (Chu et al., 2012), differences in housing or dietary conditions and genetic drift between small inbred populations can result in phenotypic differences (Giacomini et al., 2010). It is important for applications such as drug development for independent groups to have access to knockout rats with well-controlled and standardized husbandry to avoid these complications; in this regard, a commercial source is desirable.
Previous studies of absorption, distribution, metabolism, and excretion of various transporter probe substrates across SAGE Mdr1a, Bcrp, and Mrp2 knockout rats demonstrated the expected phenotypic changes with no surprising pharmacokinetic alterations (Zamek-Gliszczynski et al., 2012a). Independent studies established that SAGE Mdr1a knockout rats exhibited increased brain distribution of seven P-glycoprotein (P-gp) substrates quantitatively consistent with established murine models (Bundgaard et al., 2012), and SAGE Bcrp knockout rats showed expected increases in oral absorption and decreased biliary excretion of relevant probes (Huang et al., 2012). Proprietary Mdr1a knockout rats generated in the Wistar strain also exhibited increased oral absorption and central nervous system distribution of P-gp substrate drugs (Chu et al., 2012). SAGE transporter knockout rats have since been applied in drug development to investigate a worst-case drug interaction scenario with canalicular transport of metabolites for a phase II clinical development compound cleared by extensive hepatic metabolism with irreversible biliary excretion of metabolites (Zamek-Gliszczynski et al., 2013).
To date, potential gene expression and pathology changes in SAGE knockout rats have not been systematically evaluated, and these unknowns have created reservations about their broad application in the preclinical study of transporter-mediated pharmacokinetics. To address this knowledge gap, global gene expression and pathology were studied in the small intestine (duodenum, jejunum, ileum), liver, kidney (cortex, medulla), and brain tissue. Our results demonstrated overall minor compensatory changes in SAGE Mdr1a, Bcrp, and Mrp2 knockout rats that do not preclude their general application to the study of transporter-mediated pharmacokinetics. These data further support the use of these knockout rats as an alternative to murine models whenever rats are practically advantageous or more relevant.
Materials and Methods
Animals.
Male Sprague-Dawley Mdr1a, Bcrp, and Mrp2 knockout rats were purchased from SAGE Labs, Inc., and male wild type Sprague-Dawley control rats were purchased from Charles River (Portage, MI). Animals were acclimated for 4–20 days. At the time of necropsy, rats were 8–11 weeks old (290–400 g body weight). The Institutional Animal Care and Use Committee at Covance (Greenfield, IN) approved all animal procedures.
Tissue Collection.
Fasted body, liver, kidney, and brain weights were measured at necropsy. The following tissue sections were collected for histopathology and gene expression analysis: brain (frontal lobe of cerebral cortex), duodenum (midpoint), ileum (midpoint), jejunum (midpoint), kidney (cortex-anterior pole), kidney (medulla), and liver (edge of left lateral lobe). Approximately 100-mg sections of the collected tissues (≤2 mm thick) were placed into 1 ml of RNALater (Ambion, Inc., Austin, TX) immediately after collection and were stored until RNA isolation at 4°C. For microscopic histopathology examination, tissues were fixed in 10% buffered neutral formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin.
Gene Expression Analysis.
RNA from approximately 20 mg of tissue was isolated using Qiagen RNeasy columns (Qiagen, Hilden, Germany), and RNA quality was confirmed with an Agilent Bioanalyzer (Agilent Technologies, Santa Clara, CA). RNA was subsequently labeled and hybridized to RG230v2 microarrays (Affymetrix, Inc., Santa Clara, CA) according to the manufacturer’s protocol. Signal intensities were generated within Expression Console (Affymetrix) using the MAS5 algorithm with default settings and global scaling targeted to a signal of 1500. Affymetrix probe sets were annotated to the rat genome using the Bioconductor package (rat2302.db, March 2012). Transcriptional changes were initially filtered using a false discovery rate of ≤0.2 with Benjamini-Hochberg correction for multiple comparisons (Benjamini and Hochberg, 1995), a minimum signal of 250 for either the vehicle or treated samples, and an absolute fold-change of ≥1.5.
Pathology.
At necropsy, all organs were assessed for gross lesions. Microscopic examination was performed on hematoxylin and eosin-stained sections of all collected tissues. Blood was collected at necropsy for determination of the following hematology and clinical chemistry parameters: blood cell morphology, total and differential leukocyte cell counts, erythrocyte count, hematocrit, hemoglobin concentration, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, mean corpuscular volume, platelet count, and reticulocyte count; alanine aminotransferase, alkaline phosphatase, aspartate transaminase, γ-glutamyltransferase, and creatine kinase activities; total protein, albumin, globulin, blood urea nitrogen, creatinine, glucose, calcium, inorganic phosphorus, potassium, sodium, total bilirubin, cholesterol, and triglyceride concentrations; and albumin/globulin ratio.
Results and Discussion
Gene expression data supported the relevant knockout genotype (Supplemental Fig. 1). Mdr1a and Bcrp expression was reduced 8- to 23-fold and 3- to 10-fold, respectively, in all tissues from the relevant knockout rats. Mrp2 expression was 5.1-fold lower in the liver, decreased to the limit of detection in the intestine and kidney, and altogether not detected in the brain, where Mrp2 is not expressed (Agarwal et al., 2012). As expected, Mrp2 knockout rats exhibited hyperbilirubinemia (38-fold increase in total bilirubin) and upregulation of hepatic Mrp3 (2.4-fold, P < 0.05).
Significant changes in the expression of 112 ADME-relevant genes mined from the overall content of Affymetrix microarrays in SAGE knockout rats are summarized in Table 1 (complete data are reported in Supplemental Table 1). The two most consequential expression changes were upregulation of catechol-O-methyltransferase in all tissues of Bcrp knockout rats and intestinal carboxylesterase in Mdr1a knockouts. Catechol-O-methyltransferase induction is important to consider in the study of compounds containing a free catechol group, as well as in pharmacology studies with catecholamines (e.g., dopamine, epinephrine). Intestinal carboxylesterase induction is an important consideration for esters hydrolyzed in the gut wall. The pharmacokinetic relevance of catechol-O-methyltransferase and carboxylesterase induction is thus limited to catechols and esters, which account for a small fraction of drug space and can be readily identified based on chemical structure. Until future studies elucidate the magnitude of functional change that can result from alterations in gene expression, caution is warranted in rare instances of extensive metabolism by catechol-O-methyltransferase in Bcrp knockouts and intestinal carboxylesterase in Mdr1a knockout rats.
Other ADME gene expression changes reported in Table 1 are less relevant to preclinical studies of drug disposition. Downregulation of renal cytochrome P450 (P450) enzymes in all SAGE rats was the most prevalent gene expression change in this study and is also consistent with observations in proprietary Wistar Mdr1a knockout rats (Chu et al., 2012). However, the predominant drug metabolism pathway in the kidney is glucuronidation, and constitutive renal P450 expression and activity are negligible (Lash et al., 2008). Therefore, the noted decreases in the expression of renal P450s are not important to drug disposition. The 3- to 4-fold induction of hepatic Cyp2c in Mdr1a and Bcrp knockout rats could be of functional relevance; however, this change has not translated into increased systemic clearance of the prototypical substrate, paclitaxel, in Bcrp and Mdr1a knockout rats (Zamek-Gliszczynski et al., 2012a). Suppression of UDP-glucoronosyltransferase (Ugt) 2b15 and 2b17 in the jejunum was observed in all SAGE knockout rats. However, it is unlikely that SAGE knockouts have markedly altered intestinal glucuronidation activity because these two Ugts were repressed in only one specific segment of the intestine, and the other intestinal Ugt1a, Ugt2a, and Ugt2b enzymes were unperturbed. The observed 2- to 3-fold decrease in hepatic Cyp2b2 expression is consistent with previous reports in proprietary Wistar Mdr1a knockout rats and SAGE Bcrp knockouts (Chu et al., 2012; Huang et al., 2012); however, Cyp2b accounts for <1% of hepatic P450 content (Paine et al., 2006), so its suppression is unlikely to be consequential. Other noted changes immaterial to drug pharmacokinetics included the following: 1) Bcrp knockout suppression of Cyp4v3, an enzyme not known to metabolize drugs; 2) modest downregulation of renal Oatp1a1 and Fmo1, which are not involved in renal drug clearance; and 3) suppression of jejunal Cyp2d2, which is not known to be an important intestinal drug-metabolizing enzyme.
Hepatic Mdr1b P-gp upregulation has previously been reported in proprietary male Wistar Mdr1a knockout rats (Chu et al., 2012), but similar Mdr1b upregulation was not observed in SAGE rats. Mdr1a accounts for most functional P-gp expression in the intestine and brain, whereas both Mdr1a and 1b contribute to hepatic and renal P-gp expression (Cui et al., 2009). Therefore, upregulation of Mdr1b raises the concern that the model may be inappropriate for the study of P-gp-mediated biliary and urinary excretion. In SAGE Mdr1a knockout rats, hepatic and renal Mdr1b expression was not upregulated (actually decreased 2-fold), while biliary and urinary excretion of the P-gp substrate paclitaxel were significantly decreased (Zamek-Gliszczynski et al., 2012a). Thus, concerns that the impact of P-gp on in vivo pharmacokinetics will be masked by Mdr1b upregulation in SAGE Mdr1a knockout rats are not supported either at a gene expression or functional level.
Pathology observations in SAGE rats are summarized in Table 2. Relative liver and kidney weights were 22 to 53% higher in all three knockouts, with microscopic increases in hepatocyte size in Mdr1a and Mrp2 knockout rats and glomerular size in Bcrp and Mrp2 knockouts. Despite uniformly increased liver weights, only Mdr1a and Bcrp knockout rats exhibited increased alkaline phosphatase activities, suggesting that the increase may have arisen at the intestinal level. However, no microscopic differences were identified in the examined intestinal sections. The only finding in the brain was slight dilation of the lateral ventricles in Bcrp knockout animals.
Elevated relative liver and kidney weights in SAGE rats were quantitatively consistent with reported modest increases in the clearance of drugs that are not substrates of the knocked-out transporter. Clearance of the non-Bcrp substrate digoxin was increased by 33% in Bcrp knockout rats (Huang et al., 2012). Sulfasalazine, whose clearance is not mediated by P-gp, was cleared at a 30% enhanced rate in Mdr1a knockout rats, whereas clearance of the Mrp probe, carboxydichlorofluorescein, was 1.4-fold higher in both Bcrp and Mdr1a knockout rats (Zamek-Gliszczynski et al., 2012a). Clearance of the non-Mrp2 substrates loperamide, paclitaxel, and sulfasalazine was increased by 30–60% in Mrp2 knockout rats (Zamek-Gliszczynski et al., 2012a). Together, these data show that clearance of drugs that are not substrates of the knocked-out transporter can be 30–60% higher in SAGE knockout rats, which correlate with the 22–53% increase in liver and kidney relative weights. However, this trend of modestly decreased exposure of nonsubstrate drugs resulting from increased clearing capacity is irrelevant to the study of efflux transporter-mediated pharmacokinetics, whose goal is to identify increases in exposure (Bundgaard et al., 2012; Chu et al., 2012; Huang et al., 2012; Zamek-Gliszczynski et al., 2012a). Finally, prediction of the drug-interaction potential is based on the increase in exposure in the relevant knockout versus wild-type controls (Zamek-Gliszczynski et al., 2009).
Global analysis of gene expression changes in each model studied herein did not reveal robust changes at the functional (GO Ontology) or pathway (Ingenuity Pathway Analysis; Kyoto Encyclopedia of Genes and Genomes) levels. The low overall number of gene expression changes within 31,042 probe sets was consistent with the absence of major biologic or compensatory changes when analyzed using a false discovery rate ≤0.2. However, a few interesting significant gene-expression changes were noted that may be pertinent to broader transporter biology. In Mdr1a knockout rats, the RT1 class II major histocompatibility complex that is involved in antigen presentation was downregulated an order of magnitude in all tissues and >100-fold in the small intestine. Since P-gp effluxes endogenous inflammation modulators, such as steroids, prostaglandins, and cytokines, it can be postulated that the immunomodulatory function of P-gp may operate in part through the expression of this gene. Notably, P-gp knockout mice are susceptible to developing severe spontaneous intestinal inflammation (Panwala et al., 1998) but exhibit reduced dendritic cell function and autoimmune neuroinflammation (Kooij et al., 2009). Utrophin was induced in the three gut sections of all SAGE rats. Utrophin is a membrane-positioning protein involved in proper localization of the cholesterol transporter ABCA1 (Albrecht et al., 2008). Upregulation of utrophin in the small intestine of all three knockout rats suggests that it may be more broadly involved in membrane localization of ABC efflux pumps. Collagen type X α1 was specifically induced in all Mrp2 knockout rat tissues. Collagen is known to modulate Mrp expression in hepatocyte cultures (Luttringer et al., 2002), and the present findings suggest that the regulation of Mrp expression by collagen may operate as two-way crosstalk. Other robust expression changes with unclear links to transporter biology included upregulation of prolactin receptor, renal Snap91, hepatic Kcnn2, Plekhh1 in Mdr1a and Mrp2 knockouts, as well as downregulation of hepatic Inmt and Fam1031a in Mdr1a knockout rats. The complete gene expression results are available in the National Center for Biotechnology Information Gene Expression Omnibus (study GSE44962).
Overall, SAGE knockout rats demonstrated modest compensatory changes in gene expression and overall pathology, which do not preclude their general application to the study of transporter-mediated pharmacokinetics. Slightly higher clearance associated with 22–53% increases in relative liver and kidney weights are likely to be observed for nonsubstrates of the knocked-out transporter, but this is not an impediment to in vivo identification of transporters whose impairment elicits increased drug exposure. Until future studies elucidate the presence and magnitude of functional changes associated with these observations, caution is warranted in rare instances of extensive metabolism by catechol-O-methyltransferase in Bcrp knockouts and intestinal carboxylesterase in Mdr1a knockout rats, specifically for molecules with free catechol groups and esters subject to gut-wall hydrolysis.
Acknowledgments
The authors thank Kathy S. Piroozi and Tricia Wolff for overseeing proper tissue collection and analysis, and Drs. James L. Stevens and Armando R. Irizarry for insightful suggestions.
Authorship Contributions
Participated in research design: Zamek-Gliszczynski, Baker, Ryan.
Conducted experiments: Paulman, Goldstein.
Contributed new reagents or analytic tools: Ryan.
Performed data analysis: Goldstein, Paulman, Ryan, Zamek-Gliszczynski.
Wrote or contributed to the writing of the manuscript: Zamek-Gliszczynski, Goldstein, Paulman, Baker, Ryan.
Footnotes
- Received February 3, 2013.
- Accepted April 2, 2013.
This article has supplemental material available at dmd.aspetjournals.org.
Abbreviations
- ABC
- ATP-binding cassette
- ADME
- absorption, distribution, metabolism, and excretion
- Bcrp
- breast cancer resistance protein
- Mdr
- multidrug-resistance gene
- Mrp
- multidrug-resistance associated protein
- P450
- cytochrome P450
- P-gp
- P-glycoprotein
- Ugt
- UDP-glucoronosyltransferase
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics