Requirement for metabolic activation of acetylaminofluorene to induce multidrug gene expression.

Previously we have demonstrated that several xenobiotics can induce multidrug (mdr) gene expression in cultures of primary isolated hepatocytes. One of the best of these xenobiotic inducers in rat hepatocytes is 2-acetylaminofluorene (2-AAF), which induces mdr expression by an enhancement of mdr gene transcription. In all species studied to date, AAF is extensively and variously metabolized. In this study we have sought to determine if AAF per se or a metabolite is responsible for mediating the increase in mdr gene transcription and expression. This study demonstrates that AAF per se is not active, but that the effect of AAF we have observed on mdr gene transcription and expression in the rat is due to the formation of a reactive metabolite(s). Our data indicate that this reactive metabolite is probably N-acetoxy-2-aminofluorene or the sulfate ester of N-hydroxy-AAF. The requirement for the formation of one of these metabolites may explain the differences in species response to AAF, in terms of mdr gene expression, that we have observed. We hypothesize that the mechanism by which mdr gene transcription is increased in response to AAF involves a covalent interaction between a reactive metabolite and an mdr gene regulatory protein. Our current work is concerned with the exploration of this hypothesis.


Introduction
The multidrug (mdr) resistance phenotype occurs commonly in tumors prior to and after exposure to chemotherapeutic agents and can arise by several different mechanisms (1-3). One mechanism by which an mdr phenotype is developed is via overexpression of a membrane transport protein, P-glycoprotein (P-gp) (4)(5)(6)(7). This protein operates as a plasma membrane pump to transfer xenobiotic molecules across the plasma membrane and out of the cell against a concentration gradient (8)(9)(10)(11). Energy for this process is derived from the hydrolysis of ATP (9,10,12). P-gps are encoded by a family of genes consisting of two members in humans, primates, rabbits, and fish; generally three in rodents; four in the dog; and five in the pig (13). The genes generally divide into two classes: mdrl and mdr2. The mdrl genes encode a P-gp that confers drug resistance while the mdr2 genes code for a P-gp that does not.
Humans and primates have a mdrl and mdr2 gene. Rodents also have only one mdr2 gene but have two genes in the mdrl class. These genes are variously named in the literature, but we refer to them as mdrla and mdrlb in accordance with the nomenclature proposed by Hsu et al. (14). The P-gps are part of a much larger protein family known as the ABC proteins or ATPbinding cassette proteins (15)(16)(17). Members of this family include the yeast STE6 factor transporter to which P-gp is closely related and the cystic fibrosis transmembrane conductor (18,19).
The expression levels of P-gp are variable in tumors both at the time of diagnosis and after treatment with chemotherapeutic agents (20). Normal expression of P-gp is highest in the intestine, kidney (proximal convoluted tubule), on the bile canalicular surface of the hepatocyte, the blood brain barrier, in the adrenal gland, and in the uterus of a pregnant animal (mouse) (21)(22)(23)(24)(25). Tumors derived from these tissues normally have the highest pretreatment levels of P-gp (20,26,27). In contrast, tumors of the hemopoietic system appear most likely to develop expression following exposure to chemotherapeutic agents (20,(27)(28)(29). The normal distribution of Pgp expression suggests that its physiological role may be to expel naturally occurring toxins from the cell (2). This hypothesis is consistent with the substrate specificity of P-gp that is highest against naturally occurring chemotherapeutic agents (30).
Expression of members of the mdr gene family is induced in vivo following partial hepatectomy (31)(32)(33) and during chemical carcinogenesis (33). Hepatocytes derived from livers undergoing regeneration following partial hepatectomy demonstrate an increased resistance to hepatotoxins (34). Hepatocytes isolated from rats previously exposed to acetylaminofluorene (AAF) display an increased resistance to several cytotoxic drugs including actinomycin D (35). Expression of the mdr gene family can be induced in rat liver following acute exposure to several xenobiotics (36,37).

Results and Discussion
We are evaluating the hypothesis that P-gp is one of a family of proteins whose function is to pump xenobiotics and their metabolites from cells. We have called this process a phase III system, as it would form the third major response (phase I and II being metabolic transformations) that the cell can make to a xenobiotic insult ( Figure  1). By this hypothesis it would be reasonable for the regulatory regions of the P-gp genes, and other related excretory pump genes, to have common regulatory elements so their expression can be coordinately regulated by common protein in response to Environmental Health Perspectives toxic stress. Thus, we might postulate that the cell has a family of receptors that have affinities for various classes of substrates and, upon binding a particular xenobiotic, are able to induce expression of those metabolic and excretory genes whose proteins are best able to alter and excrete the xenobiotic ( Figure 1). The best clarified such receptor that may form part of this superfamily is the arylhvdrocarbon (Ah) receptor (38). Other possible candidates are the recently identified peroxisome proliferator receptor (39) and proteins that bind to the antioxidant and electrophilic responsive elements (40,41). The existence of different receptors for the different classes of xenobiotics that the cell may encounter is in concordance with the various substrate specificities of the cytochrome P450 isozymes. It would not be in the economic interests of the cell to activate the whole family of cvtochrome P450 enzymes to metabolize a xenobiotic that was a substrate for only one.
Previously we and others have demonstrated that several xenobiotics are able to induce the mdr gene expression (33,36,37,42). One of the mnost potent mdr inducers in the rat is 2-acetylaminofluorene (2-AAF). In rat liver and isolated hepatocytes, AAF is able to induce transcription of the mndrl genes and thus elevate their expression measured at both the mRNA and protein levels (37). In monkeys AAF is also able to induce mdr expression though the effect is less and more variable than that seen in the rat (Gant et al., unpublished data).
AAF is extensively and variably metabolized in all species (43,44). For this reason, we are conducting a study to determine if AAF per se or a metabolite is responsible for inducing mdr gene family transcription and expression. Differential excretion of AAF also occurs between species (44), so we also have been trying to determine if differences in the route of excretion of AAF metabolites can explain the differential ability of AAF to induce mdr gene expression between species (unpublished data). All of the work described here has been carried out in the rat primary cultured hepatocyte model as previously described (37). We first tested the effect of both the ring hydroxylated and N-hydroxy metabolites of AAF, to induce mdr expression (Figures 2A and 2B). As we had anticipated, none of the ring hydroxylated forms had any effect on mdr expression in this system. In contrast, both the N-hydroxy and its derivative the N-acetoxy-2-AAF (2-AAAF) were able to induce mdr expression and were 10-to 20-fold more potent as mdr inducers than AAF. This suggested that metabolism through hydroxylation to either the 2-AAAF or further metabolites was necessary for AAF to elicit increased expression of the mdr gene. Additionally fluorene has no ability to induce mdr gene expression, but 2-aminofluorene which can be acetylated to 2-AAF or activated per se to electrophiles, is equipotent with 2-AAF as an inducer of mdr gene expression (data not shown).
To further establish the role N-hydroxy metabolites in the induction of mdr gene expression by AAF, we attempted to block the induction of mdr gene expression by AAF using U-naphthoflavone (o-NF). oc-NF is an inhibitor of the cvtochrome P4501A isoforms that are mainly responsible for the catalysis of AAF N-hydroxylation. At concentrations of 1 to 5 pM, ux-NF was able to block the ability of AAF to induce mdr gene expression in isolated rat hepatocyte cultures ( Figure 3). Interestingly, uX-NF was also able to reducc the basal expression of the mdr gene in these cells, suggesting that there may be an endogenous inducer present that requires metabolism also through the cytochrome P4501A enzymes (Figure 3, lanes 3 and 4). To evaluate the role of deacetylases in the formation of the reactive metabolites, wc inhibited cellular deacetylases using paraoxon at 1 and 5 pM and tested the ability of AAF and 2-AAAF to induce mndr gene family expression under these circumstances. In microsomes, paraoxon completely inhibits deacetylase in the rat at 0.1 to 1.0 pM (46). Paraoxon was able to inhibit the ability of 2-AAAF to induce mdr gene expression but not that of AAF (Figure 4). This demonistrated that the formation of an ultimate electrophilic metabolite was required for the induction of mndr gene expression. IThc lack of effect of paraoxon on the ability of AAF to induce mdr gene expression can be explained by consideration of the other possible routes by which AAF can be metabolized to electrophilic metabolites. Two other routes of Environmental Health Perspectives .I.iIi.  2-AAAF (2 pM) + paraoxon (1 pM); 8) 2-AAAF (2 pM) + paraoxon (5 pM); 9) Paraoxon (10 pM). Analysis was performed as described in Figure 2A with the compounds being added to the cultures concurrently.
paraoxon. In rats, further metabolism of the 2-hydroxy-AAF by N O-acyltransferase appears to be more important as a route of carcinogenic activation than sulfation (43). Induction of mdr gene family expression by 2-AAAF is inhibited by paraoxon, which indicates that the 2-AAAF does not induce per se or decompose to the N-acetyl electrophile to form covalent adducts to proteins. Rather, it undergoes deacetylation to the N-acetoxy-2-aminofluorene and then subsequent decomposition to the electrophile.
The requirement for metabolism of AAF to the N-hydroxy-2-AAF and the potency of 2-AAAF as an inducer of the mdr gene family demonstrates that generation of reactive electrophiles and subsequent protein covalent binding may be the route by which AAF is able to induce mdr expression in the rat hepatocyte. 2-AAAF appears to require a deacetylation to N-acetoxy-2-aminofluorene to induce mdr expression. After the electrophile is formed, gene activation could occur either through a direct interaction with DNA, on a gene regulatory element, or via a protein interaction. Our working hypothesis is that induction of gene expression occurs through a specific protein/electrophile interaction. We know that elevation mdr gene expression occurs through a transcriptional mechanism in response to AAF and other xenobiotics (37) and therefore we postulate that this protein forming an adduct with an AAF metabolite is then, either alone or a heterodimer with another protein, able to activate mdr gene transcription. Our current work is aimed at identifying these proteins and the mechanism by which they are able to induce mdr and other gene family expressions.