First, we will discuss the possibility that the AHR can be considered as a target for immune modulation and treatment of diseases including autoimmunity and transplant rejection, and paradoxically, also potentially for cancer therapy depending on the ligand employed. Based on efforts at characterizing novel ligands of the AHR in relation to their interaction with the acquired immune system, we envision that ligands can either be “regulatory” or “effector”, depending on the inflammatory milieu and dosing strategies of the ligands. In the future this may form the basis for an entirely new class of drugs targeting the AHR for immunomodulation. A second novel concept in this manuscript is the ability of SU5416 to activate the AHRb and AHRd polymorphisms with similar efficacy. These two isoforms are present in different strains of mice, and have been well characterized for many ligands, particularly TCDD. For the majority of ligands studied, the AHRd isoform displays less than one-tenth the response of AHRb after binding. It has been proposed that a true endogenous ligand of the AHR would activate the two polymorphisms similarly, given the importance of the AHR in normal physiologic development, and that mice with either genotype do not display the abnormal phenotypes seen in AHR2/2 and hypomorphic mice [12]. While we initially utilized the AHRd polymorphism to narrow our search for potent ligands of the AHR, we inadvertently found that SU5416 activates these two isoforms with similar potency. This not only confirms the importance of this property of the drug in humans, who harbor the AHRd polymorphism, but also will allow the structure of SU5416 to serve as a model in our search for clinically relevant endogenous ligands of the AHR.
Results Primary Screen for Agonists of the Human AHR
To identify novel agonists of the AHR, a library of 4,160 small molecules, “The KBA library”, was screened at 10 mM per compound, by the Small Molecular Screening Facility of The Carbone Cancer Center of the University of Wisconsin School of Medicine and Public Health. This library represents the sum of three commercially available well characterized chemical libraries with a high frequency of approved drugs and prototype signaling molecules. This includes 2,000 diverse FDA approved drugs and natural products (Microsource Discovery Systems, Inc; Gaylordsville, CT); the 1280 compound LOPAC1280 library of diverse characterized compounds (Sigma; St Louis, MO); and 880 characterized compounds (Prestwick Chemicals; Illkirch, FR). In this first stage of the screen, AHR agonism was determined by monitoring the activation of the human receptor using the human 101L-hepatoma cell line that has a stably integrated “dioxinresponsive element (DRE) driven luciferase reporter [13]. At the tested concentration of 10 mM, approximately 100 compounds induced at least a three-fold increase in luciferase activity (figure 1A).
Secondary Screen for Agonists of the Murine AHRd Low Affinity Receptor
The 100 “hit compounds” from the primary screen were subsequently screened for their capacity to activate the low-affinityFigure 1. Screen of small molecule library for AHR agonists. A. A collection of 4160 compounds was screened for the induction of the DRE-driven luciferase in the human hepatoma 101L cell line. In 384-well plates, 100 mL media containing 70% confluent 101L cells was incubated with 10 mM of each test compound (1% v/v DMSO) for 24 hours. Dotted line indicates 3-fold induction. B. Screen for agonists of the AHRd. The AHRd-15 cell line was treated with 1 mM of the 98 compounds identified from the primary screen, 2 nM TCDD or DMSO and EROD activity was determined. Dashed line indicates 5-fold induction.
murine AHRd receptor isoform using the activity of the endogenous Cyp1a1 gene as a readout. To this end, we established a hepatoma cell line that expresses the AHRd receptor isoform derived from the DBA/2J mouse [14]. An AHRd-expressing cell line was generated by stably transfecting the AHRd cDNA into the rat hepatoma AHR-deficient cell line, BP8 [15]. After stable selection with G418, a subclone (AHRd-15) was analyzed for receptor expression and function. First, a western blot using an anti-AHR antibody, revealed that the AHRd-15 cells produced an immunoreactive protein band that co-migrated with a receptor species isolated from the hepatic cytosol of DBA/2J mice (approximate size 104 kDa). This band was distinct from the AHRb isoform found in C57BL/6J cytosol, which migrated at 97 kDa (figure S1A). To confirm that the AHRd-15 clone expressed a functional low affinity AHRd isoform, we examined the receptor-mediated response to the prototype agonists, TCDD and b-naphthoflavone (BNF). Increasing concentrations of TCDD induced CYP1A1-mediated EROD activity in these cells with an EC50 in the 30 nM range [16]. In contrast, the much weaker agonist, BNF, known not to induce an AHR-mediated response in the AHRd receptor isoform expressed in the hepatocytes of DBA/ 2J mice [17], was shown to be inactive at doses as high as 10 mM in the AHRd-15 cells (figure S1B). To test the ability of the 100 AHR inducers to activate the AHRd-15 cells, they were treated with each of the compounds at the dose of 1 mM, for 36 hours in 96-well plates. Only the compound SU5416, and the positive control, TCDD, induced AHRd-mediated EROD activity greater than 5-fold (figure 1B). Therefore, SU5416 was considered for further analysis.
Induction of DRE-mediated Transcription by SU5416 is AHR and ARNT Dependent
To prove that induction of the DRE was mediated through classic AHR signal transduction, and not through a VEGF-related mechanism, we employed mutant cell lines that lack expression of the AHR or ARNT. The C35 cell line, which contains a dysfunctional AHR, was utilized [18]. It was transfected with vector containing the murine AHR gene, the lacZ gene, and the luciferase reporter gene driven by 3 upstream DREs, as described in the Methods section. Controls were mock transfected with reporter plasmids and the empty vector. Cells were treated with either 3 mM SU5416 or DMSO (control). As seen in figure 2A, cells transfected with the AHR plasmid generated significant luciferase activity when exposed to SU5416 compared to DMSO. The control cells generated minimal activity. In a similar experiment, the ARNT-deficient mouse hepatoma cell line C4 was transiently transfected with plasmids encoding human ARNT, the lacZ gene, and the same DRE-driven luciferase gene, and control samples received empty vectors for ARNT [19,20]. As shown in figure 2B, after exposure to SU5416 or DMSO, activity was only seen when ARNT was transfected.
