Upstream eventActivation, Constitutive androstane receptor
Altered expression of hepatic CAR-dependent genes
Key Event Relationship Overview
AOPs Referencing Relationship
|AOP Name||Adjacency||Weight of Evidence||Quantitative Understanding|
|Constitutive androstane receptor activation leading to hepatocellular adenomas and carcinomas in the mouse and the rat||adjacent||High||Moderate|
Life Stage Applicability
|All life stages||High|
Key Event Relationship Description
Activation of CAR by an endogenous or foreign substance leads to translocation of the CAR-RXR heterodimer into the nucleus, and this dimer binds to DNA recognition elements in the regulatory region of CAR-responsive genes. CAR activation thus alters gene expression and upregulates xenobiotic-metabolizing enzymes such as CYP2B, CYP2C, CYP3A, sulfotransferases, UDP-glucuronyltransferases and glutathione transferases, as well as xenobiotic transporters such as Mrp2 and Mrp4 (Omiecinski et al., 2011a). In addition, CAR alters genes involved in lipid homeostasis, glucose utilization and energy metabolism. In rats and mice, the expression of additional genes involved in cell proliferation and apoptosis control are altered; Gadd45beta and Cdc20 are examples of genes that function in this way and are upregulated in mice within hours of treatment with a CAR activator (Peffer et al., 2018a; Peffer et al., 2018b; Tojima et al., 2012).
Evidence Supporting this KER
For the MIE of CAR activation, direct measurement of this step after exposure to a stressor in vivo is difficult to attain, because of the complex sequence of intermediate steps that follow. Nuclear translocation of a fluorescently tagged CAR molecule following activation by a stressor such as phenobarbital has been observed in vitro in hepatocytes, and occurs fairly rapidly (i.e. within a few hours) (Chen et al., 2010). More commonly, indirect measurements of downstream effects, and absence of those effects in CAR-null mice or rats, are employed to demonstrate that a particular molecule or other stressor can activate CAR (Huang et al., 2005; Peffer et al., 2018b; Peffer et al., 2007; Yamamoto et al., 2004).
Following the activation of CAR by stressors such as phenobarbital, TCPOBOP or metofluthrin in rats or mice, the in vivo and in vitro data are strong that demonstrate a characteristic pattern of changes in CAR-responsive genes. The data for these example molecules are described further within this KER.
It is highly plausible that activation of CAR would produce changes in expression of specific genes, since transcription of genes specifically associated with CAR response elements is how this nuclear receptor achieves its biological effects (Tien and Negishi, 2006). In mice, rat and human hepatocytes, CAR activation produces changes in genes related to CYP2B and CYP3A enzymes, certain sulfotransferase and glucuronyltransferase enzymes, hepatic transport proteins and genes related to lipid and glucose metabolism (Maglich et al., 2003; Omiecinski et al., 2011b; Omiecinski et al., 2011a; Tien and Negishi, 2006). In mice and rats, CAR activation has been shown to also achieve alterations in genes that give an overall pathway change associated with increased progression through the cell cycle such as Gadd45b and Cdc20 (Deguchi et al., 2009; Geter et al., 2014; Ross et al., 2010; Tojima et al., 2012).
Strong support for this linkage comes from in vivo and in vitro studies with model CAR activators, and via the absence of the same effects including specific gene expression changes in CAR knockout mice lacking the CAR receptor. In a study with TCPOBOP (a direct activator of mouse CAR), these CAR-specific changes included Gadd45b (↑ 14-fold), plus Cdc20 (↑ 37-fold) and additional cytokines, and these genes were unaffected in CAR knockout mice (Tojima et al., 2012). Gene pathways (by Ingenuity Pathway Analysis, IPA) that were altered by phenobarbital in a dose-responsive manner in CD-1 mice included “Cell cycle of chromosomal replication” and “Cell cycle: G2/M DNA damage checkpoint regulation”; these were only significantly altered above a dose level of 15 mg/kg/day (Geter et al., 2014). In addition, CAR knockout mice showed absent or greatly diminished induction of mRNA (compared to wild-type mice) for Cyp2b10 after treatment with TCPOBOP or phenobarbital (Ross et al., 2010; Tojima et al., 2012).
With humans, a selective activator of human CAR (CITCO; 6-(4-Chlorophenyl)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl)oxime) was discovered that has shown little or no ability to activate the corresponding CAR of mice and rats in CAR3 reporter assays, whereas it produced a strong positive response in reporter assays for human CAR (Maglich et al., 2003; Omiecinski et al., 2011b). Upon treatment of human hepatocytes with CITCO or a model PXR activator (rifampicin), CITCO produced a change in expression of a suite of eight marker changes that was distinct from that of the PXR activator. The degree of fold-changes varied greatly across three different human donors, but consistent CAR-related changes were seen in CYP2B6, CYP3A4, CYP2A6, ALAS, ALDH1A4, and GSTA2, with induction in 1 or 2 of the donors for the genes MDR1 and SULT1A1 (Maglich et al., 2003).
Oshida et al. (2015a) have published a “CAR signature” that is based upon changes in 83 genes consistently expressed in mice with three different CAR activators (phenobarbital, TCPOBOP and CITCO). Comparison to this CAR biomarker signature for new molecules is possible via a running Fisher’s p-value in Correlation Engine (Illumina) or via examination IPA software (Ingenuity Pathway Analysis) of the 10 most altered pathways. An approach like this is useful to establish a CAR gene expression profile has occurred, since it relies on an overall pathway rather than a small number of selected genes. For example, the triazole fungicide cyproconazole gave a highly significant response compared to the CAR signature (-log [p-values] of 8.0, 12.4 and 19.0 for 100, 200 and 400 ppm doses), and the gene signature did not match similar published biomarker signatures for AhR or PPARα (Oshida et al., 2015a; Oshida et al., 2015b; Oshida et al., 2015c; Peffer et al., 2018b).
In rats, metofluthrin has been shown to be a strong CAR activator in the liver. The key event of increased Cyp2b1/2 and/or Cyp3a1 mRNA expression levels was demonstrated at dose levels of 900 ppm and above, and they were unaffected at the NOAEL of 200 ppm in the diet. In isolated rat hepatocyte cultures treated with 50 µm metofluthrin, co-administration of siRNA for CAR (a gene silencing technique) caused a knockdown of Car mRNA expression and resulting suppression of the response to metofluthrin in terms of Cyp2b1 mRNA and Car mRNA levels (Deguchi et al., 2009), thus providing evidence for the MIE (CAR activation) and the downstream key event (altered gene expression secondary to CAR activation).
In other mammalian species such as dogs, hamsters, and non-human primates, the presence of a functional CAR receptor with a known species-specific amino acid sequence has been demonstrated (Omiecinski et al., 2011b; Reschly and Krasowski, 2006), but very little experimentation with the gene expression effects of CAR activators in these species has been described thus far in the literature.
Uncertainties and Inconsistencies
In general, CAR activators show very consistent, large fold-increases for the characteristic expression of Cyp2b isoforms across in vivo studies in multiple species and with many different molecules. While certain genes related to a pro-proliferative effect appear to be CAR-mediated and reproducibly impacted in multiple studies (Currie et al., 2014; Deguchi et al., 2009; Geter et al., 2014; Peffer et al., 2007; Tojima et al., 2012), there are examples where changes in a specific gene was not observed. For example, Ross et al. (2010) tested 80 mg/kg/day (ip dosing) phenobarbital for 4 days in WT C57BL/6J mice, and they observed a 15.8-fold increase in Cdc20, but did not see an increased expression of Gadd45b. Mapping of a specific genes’ changes following activation of CAR by a particular CAR activator may be affected by the species, strain, dose level and time point examined, as well as the other non-CAR effects of that molecule. Examining for a significant pathway change is likely to be a more reliable measure of this Key Event Relationship (Oshida et al., 2015a), but this is also somewhat dependent on the experimental design, the species and duration of treatment, and the pathway analysis tools.
Quantitative Understanding of the Linkage
A quantitative understanding supporting the linkage between CAR activation (the MIE) and measurable changes in the appropriate genes’ expression levels (KE1) is somewhat difficult to assess directly, because the methods to measure the actual event of CAR activation are largely confined to sub-molecular manipulations that are not conducive to in vivo dose-response assessment, or require in vitro techniques such as reporter assays, or nuclear translocation determinations (Chen et al., 2010; Maglich et al., 2003; Omiecinski et al., 2011b; Stanley et al., 2006). Instead, assessments of the relationship between MIE and KE1 with CAR activators have been obtained via indirect demonstration that CAR activation has occurred, by examining responses in both wild-type and CAR-null mice or rats, or by use of siRNA. As discussed above under "Empirical Evidence", data of this type for phenobarbital or TCPOBOP treatment in wild-type mice and CAR-null mice has demonstrated that changes in expression of CAR-responsive genes at a particular dose level in vivo requires the presence (and activation) of a functional CAR (Huang et al., 2005; Peffer et al., 2007; Tojima et al., 2012). Also, following metofluthrin treatment of rats, altered gene expression was observed (900 ppm and higher), and treatment of rat hepatocytes in vitro along with siRNA for CAR greatly attenuated the response of CAR-activated genes (Deguchi et al., 2009). Thus, there is a limited level of understanding of this quantitative relationship between MIE and KE1.
Because the extent or degree of activation of CAR is not readily measurable with in vivo studies, the nature of the relationship between CAR activation and altered mRNA expression (e.g. linear, exponential, other) cannot be stated. In addition, the direction of change in certain CAR-responsive genes is variable, in that CAR activation causes an increase in expression of some genes and a decrease in expression of other genes. For example, in mice treated with TCPOBOP by ip injection, Tojima et al. (2012) showed large increases in expression of Cyp2b10 (151x), Cdc20 (37x) and Gadd45b (14x), plus decreases in expression of lipid-related genes Acsl5 (0.5x), Slc21a1 (0.2x) and Hmgcs2 (0.3x). The different fold-change values for specific genes also indicates that the magnitude of the gene expression differences can be very dependent on the properties of the individual gene, and whether it is normally active or quiescent prior to treatment with a CAR activator.
The onset of CAR activation (upstream KE) and the time scale of measurable differences in gene expression (downstream KE) are influenced by variables such as pharmacokinetics and route of administration of an administered stressor. However, it is established that exposure of the livers of mice to a CAR activator (TCPOBOP) by intraperitoneal injection, thus bypassing the intestinal transit and absorption processes, was able to achieve large changes in gene expression in the liver after 12 hours (Tojima et al., 2012). The cascade of normal biological events following CAR activation by a ligand (i.e. binding, nuclear translocation, altered gene expression, and post-transcriptional modifications of mRNA) likely takes several hours to be completed, after which a measurable difference vs. control in mRNA levels is detectable by RT-PCR or microarrays. After the initial perturbation by a stressor of CAR, the time course of changes in gene expression is indefinite. That is, some gene expression changes can continue to be observed for as long as the CAR activator is present. For example, in mice treated with phenobarbital for up to 32 weeks following initiation with a single dose of diethylnitrosamine, differences in gene expression of CAR-responsive genes such as Gadd45b and Cdc2 continued to be observed throughout the entire treatment period in both non-neoplastic liver tissue and in liver tumors of treated vs. untreated animals (Phillips et al., 2009a).
Known modulating factors
Activation of PXR (NR1I2), a related nuclear receptor to CAR (NR1I3), is a possible confounding factor that may be operative for certain substances. There is much cross-talk between CAR and PXR, and similar responsive genes, and a particular agent could produce a mixed set of gene expression response by activating both PXR and CAR (Tojima et al., 2012; Stanley et al., 2006).
Known Feedforward/Feedback loops influencing this KER
In terms of this specific KER1 of activation of CAR (MIE) leading directly to altered expression of CAR-responsive genes (KE1), there are no known feedback loops that would affect this overall process. As an example of this lack of feedback alteration, when mice were treated long-term with phenobarbital, they continued to show significant changes in expression of CAR-responsive genes up through 32 weeks of treatment (Phillips et al., 2009).
Domain of Applicability
CAR receptors are present in the livers of virtually all mammalian species; however, there are important differences in protein sequence and thus ligand binding properties (Omiecinski et al., 2011b; Reschly and Krasowski, 2006). In reporter assays for mouse, rat, dog and human CAR, clear qualitative as well as quantitative differences in the ability of suspected CAR activators to activate CAR from the different species were demonstrated (Omiecinski et al., 2011b). In terms of the specific KER of CAR activation directly leading to altered gene expression specific to CAR activation, in vitro hepatocyte experiments indicate that human hepatocytes have only partial overlap with mice and rats in terms of the genes that are affected. In particular, genes that are related to CYP induction (e.g. Cyp2b isoforms) show increases in expression across mouse, rat and human if the CAR molecule for that species is activated, but the pro-proliferative gene pathways have been shown to be activated only in mice and rats (Elcombe et al., 2014; Hasmall and Roberts, 1999; Hirose et al., 2009; Lake, 2009). For example, the total number of altered genes in livers of chimeric mice that reflected human hepatocytes (293) compared to livers of similarly treated CD-1 mice (846) was much lower, and only 10 differentially expressed genes (primarily CYP genes) were common to both species’ liver samples following treatment with phenobarbital at dose levels of 1000 – 2500 ppm in the diet (Yamada et al., 2014).
AOPWIKI will not let me paste in refs. ERROR