The Identification of Naturally Occurring Neoruscogenin as a Bioavailable,Potent,and High-Affinity Agonist of the Nuclear Receptor RORα (NR1F1)

Abstract

Plants represent a tremendous structural diversity of natural compounds that bind to many different human disease targets and are potentially useful as starting points for medicinal chemistry programs. This resource is, however, still underexploited due to technical difficulties with the identification of minute quantities of active ingredients in complex mixtures of structurally diverse compounds upon raw phytomass extraction. In this work, we describe the successful identification of a novel class of potent RAR-related orphan receptor alpha (RORα or nuclear receptor NR1F1) agonists from a library of 12,000 plant extract fractions by using an optimized, robust high-throughput cell-free screening method, as well as an innovative hit compound identification procedure through further extract deconvolution and subsequent structural elucidation of the active natural compound(s). In particular, we demonstrate that neoruscogenin, a member of the steroidal sapogenin family, is a potent and high-affinity RORα agonist, as shown by its activity in RORα reporter assays and from its effect on RORα target gene expression in vitro and in vivo. Neoruscogenin represents a universal pharmacological tool for RORα research due to its specific selectivity profile versus other nuclear receptors, its excellent microsomal stability, good bioavailability, and significant peripheral exposure in mouse.

Keywords

high-throughput screening plant extracts natural products RORα NR1F1 neoruscogenin

Introduction

New target validation has always been a challenge for the pharmaceutical industry with respect to the potential therapeutic value. Information obtained through genetic gain- and/or loss-of-function approaches is of course of great value, but it does not replace target validation through pharmacological stimulation or inhibition of a target with selective and potent chemical ligands. The identification of such ligands usually takes 18 to 24 months or even more in programs where medicinal chemistry starts from high-throughput screening (HTS)–derived hit compounds identified in chemical libraries. Since it is essential to provide high-quality proof-of-concept validations early in the drug discovery process, technologies that enable the identification of potent, selective, and bioavailable chemical probes are of crucial interest, even if the resulting compounds require further optimization for development into a pharmaceutical product.

It has been recognized that natural compounds bear unique drug-likeness properties due to their immense structural diversity and their inherent propensity to interact with biological targets. Numerous examples exist to show that plant biodiversity screening has led to the identification of novel drug candidates.^1,2 Indeed, the property space of natural products does match very well with the property space of drugs on the market,³ considering that molecules derived from natural products are at the origin of nearly half of all the small-molecule drugs that have reached the market from 2000 to 2010.⁴ Despite the fact that methods compatible with high-throughput extract fractionation and target screening were initially described several years ago,^5,6 technical difficulties associated with structural elucidation of minute quantities of specific compounds in a complex matrix of plant secondary metabolites have made natural product–based drug discovery procedures unpopular in the drug discovery industry.

Nuclear receptors (NRs) constitute a large family of ligand-activated transcription factors that work in concert with other nuclear proteins to regulate the expression of genes important for development and homeostasis. NRs bind directly to specific DNA response elements in the regulatory regions of their target genes and modulate the rate of transcription in a ligand-dependent manner. NRs have a conserved domain structure composed of a variable amino-terminal A/B domain, a central highly conserved DNA-binding domain (C domain) hinge region (domain D), and a carboxy terminal ligand-binding domain (LBD; E domain). Ligand binding induces a conformational change in the receptor. For an agonist ligand, these results in the dissociation of a corepressor(s) to be replaced by a coactivator protein, whereas for an inverse agonist, it induces a conformational change that results in a decreased affinity of the receptor for the coactivator protein(s) and in the inhibition of its constitutive activity. Consistent with their key biological roles, nuclear receptors serve as important drug targets,⁷ and NR ligands represent 10% to 15% of the present pharmacopoeia.⁸

RAR-related receptors (RORs), originally identified by virtue of their homology to retinoic acid receptors (RARs), have gained broad attention because of multiple phenotypes observed in the RORα-deficient staggerer mice, which carry a natural loss-of-function mutation in the RORα gene. The staggerer mutation and/or targeted RORα deletion both result in serious defects in cerebellum development, including dramatic loss of Purkinje cells and granule cells, leading to extensive neurodegeneration of cerebellum and severe ataxia.^9,10 A variety of other systemic effects, such as circadian rhythm regulation, effects on glucose and lipid homeostasis, immune system regulation, and xenobiotic metabolism, were described in mice deficient for RORα in addition to its role in CNS development. However, the complexity of RORα biology is beyond the scope of this article, and interested readers are referred to an excellent recent review.¹¹

Previous studies examining the crystal structure of RORα identified cholesterol and cholesterol sulfate present in the ligand-binding pocket.^12,13 Subsequently, 7-oxygenated sterols (7α-hydroxycholesterol, 7β-hydroxycholesterol, 7-ketocholesterol) and (24S)-hydroxycholesterol were identified as novel RORα ligands, with comparable nanomolar binding affinities,^14,15 and recently the structures of synthetic RORα ligands, based on either the benzenesulfonamide or thiourea scaffolds, have been published.^16–18 However, no potent and bioavailable ligands are available to date to address important questions about the complex biology of the RORα receptor in vivo. We demonstrate in this work that less conventional approaches, such as the identification of receptor ligands starting from a natural product library of plant extract fractions, are an efficient and effective method to identify potent and bioavailable pharmacological probes to be used for orphan target validation studies.

Materials and Methods

Recombinant Proteins Used in the Study

DNA sequences coding for the glutathione S-transferase (GST) fragment (AAA57101:1–238), GST-RORα fusion protein, or GST-BAP fusion protein were cloned into the pRSETA (Invitrogen, Carlsbad, CA, USA). The GST-RORα construct contains the RORα sequence that corresponds to NP_599023.1:139–523. The GST-BAP construct contains the sequence that corresponds to the bacterial alkaline phosphatase fragment NP_414917.2:26–471. The three open reading frames were subsequently subcloned into a proprietary protein expression vector, and the proteins were produced in Sf9 (High-Five) cells by GTP Technology (Labège, France).

The sequences of both the TIF2 (NCOA2) fragment (NP_006531.1:413–812) and BAP (NP_414917.2:26–471) were inserted in frame in the pRSETA (Invitrogen). Three LXXLL motifs present in the TIF2 (413–812) fragment were mutated one by one into LXXAA motifs by the classic site-directed mutagenesis method (QUICK Change II Kit; Stratagene, La Jolla, CA) to yield the His-TIF2mut-BAP sequence. Both His-TIF2-BAP and His-TIF2mut-BAP fusion proteins were produced in Escherichia coli, and the purification was achieved through the immobilized-metal affinity chromatography (IMAC) at GTP Technology.

Pull-Down Assay

The pull-down assay was performed in high-binding capacity streptavidin-coated plates (Pierce, Rockford, IL, USA), which were coated overnight with biotin-conjugated, affinity-purified anti-GST antibodies (Antibodies Online, Aachen, Germany). Final antibody dilution was empirically optimized for each lot. Plates were washed with TBS-Tween (20 mM Tris-HCl [pH 7.2], 150 mM NaCl, and Tween 0.05%), and the reaction was set up in the binding buffer (15 mM Tris-HCl [pH 8.0], 15 mM glycerol, 0.15 mM EDTA, 0.04% NP40, 22.5 mM KCl, 0.25 mg/mL fatty acid–free bovine serum albumin, 1 mM dithiothreitol, and 1 mM benzamidine), in a final volume of 200 µL. For each reaction, 0.12 µg GST-RORα protein and 0.10 µg TIF2-BAP protein were used. The antiprotease tablets from Roche (Basel, Switzerland) were added (1 tablet for each 50 mL of the buffer) to prevent protein degradation. All compounds were added as DMSO solutions. 7-Dehydrocholesterol (7DC, 10 µM) was used as a reference compound. Binding reaction was performed overnight at 4 °C. Binding plates were washed with TBS-Tween (20 mM Tris-HCl [pH 7.2], 150 mM NaCl, and Tween 0.05%). Lumi-Phos WB (Pierce; 50 µL per well) was added following the washing step, and plates were incubated for 30 min at 37 °C. The luminescent signal was read with the TECAN Ultra 384-well microplate reader (TECAN, Männedorf, Switzerland). The integration time was set to 500 ms.

To discard compounds that induce a nonspecific signal, hit compounds were subsequently tested in counterscreen assays. We have used assays that measured (1) a recruitment of a mutated TIF2-BAP protein on GST-RORα, (2) a recruitment of TIF2-BAP on a GST protein fragment that contains no ROR-LBD, and (3) the effect of the compound on a chemiluminescent signal produced by the GST-BAP protein. Active compounds (primary assay) that scored as negative in all three counterscreen assays were validated as RORα ligands.

Results

Screening Assay Validation

In an effort to identify natural RORα ligands, a cell-free high-throughput assay was developed, which was sufficiently robust and sensitive to detect RORα activity in partially purified fractions produced from raw plant extracts. Among the different assays tested, a pull-down assay that measures TIF2 protein recruitment on the immobilized RORα protein (Suppl. Fig. S1) provided the most robust and reproducible signal. By systematically testing different approaches available in a comprehensive platform, it was first established that the biotin-streptavidin chemistry, when coupled with an anti-GST antibody, gave the most stable GST-RORα fusion protein attachment to a solid surface. TIF2 (NCOA2) protein was previously reported as a natural RORα cofactor protein that was important for the control of hepatic glucose production.¹⁹ In this work, the TIF2 protein fragment, which encompasses residues 413–812 and contains three functional LXXLL motifs, was used to make a hybrid reporter protein TIF2-BAP that carried the bacterial alkaline phosphatase (BAP) activity, which produced a chemiluminescent signal upon addition of the Lumi-Phos WB substrate. As shown in Figure 1A , the chemiluminescent signal (TIF2-BAP protein recruitment) rises with increasing concentration of the ligand 7DC. 7DC was previously described in the literature as a high-affinity RORα ligand and used in this study to validate the ligand-dependent character of the assay.¹³ Despite the high-binding affinity in a cell-free assay, 7DC induced only very modest RORα activity in cells, and this effect was confounded by cytotoxicity. To ensure that the observed chemiluminescent signal is specific, the ligand-dependent TIF2-BAP recruitment was demonstrated to be completely abolished when the LXXLL motifs that interact with the LBD of the receptor were mutated on the TIF2 protein ( Fig. 1B ).

Figure 1.

Development and validation of the primary screening assay. Purified GST-RORα protein was first immobilized on high-binding capacity streptavidin-coated plates. Upon several washes, a purified reporter protein (TIF2-BAP) and RORα reference agonist ligand, 7-dehydrocholesterol (7DC), were added. Binding reaction was performed overnight at 4 °C. A chemiluminescent substrate of the bacterial alkaline phosphatase (BAP) was then added upon extensive washes, and the signal was finally developed for 30 min at 37 °C in the dark. The reporter signal (relative light units [RLU]) was normalized to the basal signal, obtained for the solvent (DMSO). As shown in A, the addition of 7DC resulted in a dose-dependent increase of the reporter signal. The depicted curve is an average of 10 independent experiments. Data points are mean values. Error bars represent standard deviation. The specificity of the assay was ascertained by performing parallel binding reactions with a mutated TIF2-BAP protein. As shown in B, the addition of 7DC did not augment the recruitment of the TIF2 mut protein, where all the three LXXLL motifs were invalidated by side-directed mutagenesis.

Since the capacity for HTS was an objective, the pull-down assay was validated in the 384-well plate format, where induction with 7DC (10 µM) was on average 2.6-fold versus DMSO, and the Z′ factor values were between 0.5 and 0.7.

Plant Extract Library Screening Results

Once validated on steroid ligands, the pull-down assay was used for screening of the commercially available BILOBAC N library (Bicoll GmbH, München, Germany), which consists of 12,000 plant extract fractions (plant Profiles). These extracts were originally produced from 100 different and rare endemic Asian plant organs (leafs, stems, roots, etc.). For each plant, a set of 96 fractions originating from a nonpolar crude extract (obtained with dichloromethane [DCM] and ethyl acetate [EtOAc] extraction) and the corresponding set of 96 fractions of a more polar crude extract (obtained with EtOH extraction) were both submitted to screening. The complexity of the fractions prepared with this technology is considerably reduced as compared with the crude plant extract, and the fractionated material was easily amenable to high-throughput compatible procedures. Due to the fractionation procedure and the sequential screening of each fraction in the series, the signals measured in the neighboring wells form a continuous, bell-shaped distribution of the activity, resulting in active fraction clusters (Suppl. Fig. S2A). This is indeed very different from a random distribution pattern, usually observed in primary screens on synthetic small-molecule compound libraries. Therefore, in this study, a primary hit is identified as a bell-shaped activity cluster obtained by signals from at least three consecutive fractions.

The initial screening was performed in simplicates at a concentration of 50 µg/mL. Fractions from three different plant fraction plates (96-well microplate format) were transferred onto one 384-well reaction plate, and the remaining empty wells were filled with either DMSO or the reference compound. The primary screen was performed as described in the Materials and Methods on 12,000 fractions, which correspond on average to between 36,000 and 120,000 different compounds.²⁰ To avoid a significant background activity, we have qualified only those clusters with an activity that was at least 2 × the standard deviation (SD) of the mean induction as calculated from all plates. Furthermore, only clusters with an activity of 40% or more, as compared with the induction obtained with the reference compound (7DC), were considered. Supplemental Figure S2B shows the distribution of screening assay reference values (basal signal level, positive control [7DC] signal, and Z′ factor) for each of the 42 plates (384-well format) screened. The mean ± SD induction over DMSO, obtained for the reference compound, was 2.69 ± 0.24, and the mean ± SD Z′ factor for the screen was 0.59 ± 0.006. The control values were satisfactory for most plates. Indeed, the Z′ factor was above 0.5 for 39 of 42 plates (Suppl. Fig. S2B).

Primary screen data were analyzed manually to exclude all the active clusters that did not satisfy primary hit criteria. By applying this method, 18 potential agonist clusters were identified (Suppl. Fig. S2C) that reached the activity of at least 60% of the positive control (7DC). In addition, 50 potential agonist clusters that reached the activity of at least 40% of the positive control were also identified (Suppl. Fig. S2C). Similarly, 54 potential inverse agonist clusters were identified, to provide a total of 122 potential hit clusters. Subsequently, the specificity of hit clusters on RORα activity was confirmed and validated in most of the qualified potential agonist hit clusters by running these fraction clusters in parallel in GST-RORα and GST-alone pull-down setups. The specific RORα activity profiles measured in the primary screen were confirmed in the secondary confirmation assay for most of the hit clusters. A selection of 52 potential agonist clusters was confirmed by this approach and subsequently tested in counterscreen assays (as described in Materials and Methods), where all of the 52 confirmed hit clusters passed successfully through this last validation step.

Fine Fractionation and Structure Elucidation

Fractions of the plant extracts (plant Profiles) represent a broad logP range. Empirical methods established that the logP-value extends with high reproducibility from 10 (A01 coordinate) to 1 (H12 coordinate) on each 96-well plate for different plant species (not shown). To ensure the structural diversity of compounds picked for further evaluation and to avoid any repeated selection of similar structures from different plants, active clusters from different logP regions were chosen for further study.

Typically, a deconvolution of an active cluster into a pure compound was achieved by preparative radial thin-layer chromatography (TLC) of the combined fractions on 1-mm silica gel plates. On average, one confirmed hit cluster led to a resubmission of between 5 and 20 partially purified compounds (purity from 70%–90%) for RORα activity evaluation. In the primary screen, the most active RORα agonist cluster, 8B6-MP-068E12-F7 (from the Agavaceae, Dracaena cambodiana), had a relative activity of approximately 60% of the reference compound. This activity was reconfirmed in a separate experiment and was specific for RORα, since no activity was observed in the assay that contained no RORα protein ( Fig. 2A ). Fine fractionation and isolation of active compound(s) of the cluster 8B6-MP-068E12-F7 were performed from the crude DCM and EtOAc extracts. The extract was obtained by refluxing dried plant material in the solvents, followed by filtration and evaporation. The crude extract was subjected to dry silica gel vacuum chromatography with a DCM-methanol gradient system (60:1 → 10:1). Compounds present in the active fractions were identified by silica gel TLC comparison with active plant Profiles (detection UV [254 nm], I₂, Ce-Mo-phosphate). Compounds were enriched in the DCM-methanol = 25:1 fraction. Subsequent radial TLC on silica gel with the DCM-methanol gradient system (40:1 → 10:1) yielded 10 fine fractions for activity reconfirmation. The fine fraction 5 showed the highest activity, as determined by GST-RORα versus GST alone in the pull-down assay, with a relative activity of 310% as compared with DMSO ( Fig. 2B ). The major compound present in fraction 5 was chosen for structure elucidation studies. A final purification was conducted by a chromatography self-packed RP-18 column with acetonitrile/water, yielding 1.4 mg of pure desired compound with >90% purity (according to ¹H NMR estimation). Structure elucidation was conducted as based on ¹H NMR, ¹³C NMR, and 2D-COSY-NMR (coupled C, H-spectrum). The chemical shifts (¹H NMR and ¹³C NMR) are consistent with those reported in the literature for (25S)-ruscogenin, which is a spirostan sapogenin compound (Suppl. Fig. S3).²¹ To the best of our knowledge, this is the first description of (25S)-ruscogenin found in D. cambodiana. Based on the chemical shift in ¹H NMR (H-26a/b) and ¹³C NMR (F-ring carbons), the absolute configuration could be confirmed as 25S (Suppl. Tables S1 and S2). As measured by pull-down assay, highly purified (25S)-ruscogenin was very active, with an EC₅₀ of 0.78 µM ( Fig. 2C ).

Figure 2.

Identification and structural elucidation of a submicromolar RORα agonist ligand in the extract of the Agavaceae, Dracaena cambodiana. (A) The graph represents the relative RORα activity detected in a series of consecutive fractions (original cluster MP-068:E12–F07) from D. cambodiana. Natural compounds were extracted first with the dichloromethane (DCM) and subsequently with ethyl acetate (EtOAc). The activities of GST-RORα wells that were measured in the primary screen (gray circles) or in the confirmation run (black squares) are similar and show a bell-shaped distribution. In contrast, no ligand-dependent induction was detected in wells that contained these same fractions and the glutathione S-transferase (GST) protein with no RORα (black triangles). (B) RORα activity of semi-pure compounds, isolated from the active agonist cluster MP-068:E12–F07. The activity measured in GST-RORα wells is depicted as a series of black squares. The activity measured in GST-only wells is depicted as a series of black triangles. (C) The chemical structure of the pure RORα agonist compound isolated from fraction 5 in B. This compound has previously been described in the literature as (25S)-ruscogenin, according to structure elucidation by ¹H NMR, ¹³C NMR, and 2D-COSY-NMR. The graph represents TIF2-BAP recruitment on GST-RORα that was induced by highly purified (25S)-ruscogenin. The value of 100% corresponds to the basal signal with the solvent (DMSO).

Identification of a Potent (25S)-Ruscogenin Analogue and Functional Properties In Vitro and In Vivo

Purification of milligram amounts of (25S)-ruscogenin suitable for extensive in vitro evaluation was very time and resource consuming. Therefore, some highly purified and commercially available (25S)-ruscogenin analogues were sourced for RORα agonist activity determination. Among the closest analogues, neoruscogenin (CAS #17676-33-4) appeared as the most potent RORα agonist compound (EC₅₀ = 0.11 µM), as established by the pull-down assay ( Fig. 3A ). Furthermore, treatment with neoruscogenin, as described in detail in the supplementary materials and methods, significantly induced Gal4RE-luc reporter activity in COS7 cells that expressed Gal4(DBD)-RORα protein ( Fig. 3B ). In contrast, reporter activity was only minimally induced in COS7 cells that expressed the Gal4-DBD construct. The capacity of neoruscogenin to induce RORα target gene expression in the natural chromatin context was then demonstrated in HepG2 cells, where the expression of G6Pase and Bmal1, both well-established RORα target genes,^14,18,22 was significantly induced upon neoruscogenin treatment ( Fig. 3C ). Detailed description of all experimental procedures is provided in the accompanying supplementary materials and methods.

Figure 3.

Structurally related neoruscogenin is a more potent RORα agonist as compared with (25S)-ruscogenin. (A) The graph represents TIF2-BAP recruitment on GST-RORα (pull-down assay) that was induced by highly purified neoruscogenin (NEO). The value of 100% corresponds to the basal signal of the assay (DMSO). The chemical structure of the neoruscogenin is depicted. (B) COS7 cells were transfected with either Gal4-DBD or Gal4-RORα, in addition to Gal4-RE-luc reporter plasmid, and treated with either DMSO or 10 µM NEO. Reporter activity (relative light units [RLU]) was measured 24 h upon stimulation. (C) HepG2 hepatoma cells were treated with NEO for 24 h, RNA was harvested, and RORα target gene (G6Pase, Bmal1) expression was studied by quantitative PCR. The expression of target genes was normalized to the expression of the 36B4 gene. Treatment was not associated with toxicity as established with the (Promega, Madison, WI, USA) assay (not shown). Statistical evaluation: t test, *p < 0.05, ***p < 0.001.

Next, the selectivity, early ADME properties, and the pharmacokinetic (PK) profile of neoruscogenin were characterized before the compound was qualified for experiments in murine models. Neoruscogenin treatment did not significantly activate either steroid or xenobiotic nuclear receptors with the exception of a modest PXR activation (i.e., 1.8-fold reporter induction in cells that expressed GAL4(DBD)-PXR vs. 1.3-fold induction in cells that did not express the Gal4(DBD)-PXR) (Suppl. Fig. S4), However, the induction of Gal4(DBD)-PXR by neoruscogenin was very modest as compared with the effect of classic PXR reference ligands, such as the rifampicine (4- to 5-fold increase).

Neoruscogenin was metabolically stable, as 88% and 73% of the parent compound remained intact upon a 1-h incubation with mouse or human primary hepatic microsome preparations, respectively. Finally, classic PK studies were performed in mice (Suppl. Table S3) and showed that upon a single-dose administration (10 mpk per os), circulating neoruscogenin was detected in the plasma at a low micromolar range (C_max = 1.1 µmol/L) with a peak concentration (T_max) at 3 h, which are good results for a primary hit compound from the HTS. Significantly higher exposure was detected in brain samples. Despite rather modest plasma exposure (AUC_{0-24 h} = 3561 ng/mL·h), neoruscogenin concentration in the liver was sufficient to induce RORα-dependent target gene expression. Indeed, quantitative PCR studies performed on liver samples from mice treated with neoruscogenin show that expression of classic RORα target genes, such as Bmal1, Cyp7b1, and G6Pase,^22–25 is significantly upregulated ( Fig. 4 ), with no change in the expression of the RORα receptor itself. This finding confirmed the gene expression results obtained in HepG2 cells ( Fig. 3C ). In addition, hepatic expression of Lpin2 and Angptl4, two genes that are significantly downregulated in RORα–deficient staggerer mice,²⁶ was induced by treatment with NEO ( Fig. 4 ).

Figure 4.

Neoruscogenin (NEO) activates hepatic RORα target gene expression. RNA was extracted and gene expression was measured by quantitative PCR in liver samples from mice that were treated with NEO (3 mg/kg/d, per os) for 7 days, as described in Materials and Methods. The expression of target genes was normalized to the expression of the 36B4 gene. Statistical evaluation: t test, *p < 0.05, **p < 0.01, ***p < 0.001.

Discussion

We describe in this article a straightforward and HTS-compatible procedure that enables the identification of naturally occurring ligands of orphan receptors in fractionated plant extracts. In principle, this approach is only limited by the technical difficulty to set up a robust cell-free assay, where the ligand binding domain of the target is exposed for interactions with preprocessed phytomaterial under nondenaturing conditions. Nuclear receptor RORα, a protein that has been targeted for more than a decade to identify potent and bioavailable agonists, was used here as an example of a challenging drug target.

In this work, roughly 12,000 readily available phytomaterial fractions (plant Profiles) were used, which should correspond to a library of up to 100,000 chemical entities.²⁰ In contrast, commercially available libraries of pure plant compounds contain only several thousand molecules. The technology used in the plant Profiles approach maximizes the chemical diversity present in the sample of 100 different Asian plant species used in this study. In contrast to crude extracts, which often interfere with biological assays to give false-positive readouts, our HTS procedure run on plant extract fractions gave reproducible and robust signals. Furthermore, positive fractions with RORα activity were further deconvoluted to obtain pure and active compounds within a period of 3 months. Classic problems associated with traditional extract testing, such as precipitation, interference with signal, and nonreproducible fractionation, are readily overcome by the process described in this article with no loss in product diversity. In conclusion, a preformatted (96-well), fractionated extract approach was chosen over a library of pure natural compounds to increase the structural diversity and avoid the burden of the purification of thousands of potentially inactive compounds.

The choice of a cell-free assay with a mechanism of action that relied on a naturally occurring protein-protein interaction (NR-LBD: NR-binding domain of a cofactor protein) was very important to obtain reproducible results and a significant ligand-induced response window (i.e., signal induction with the reference compound was ≥250% as compared with the basal signal). It was also taken into consideration during the assay development phase that there was a clear advantage of using a reporter assay that is based on chemiluminescence, to avoid interference by fluorescence that is naturally produced by plant metabolites.

The purification of (25S)-ruscogenin, which was the most active hit compound identified in this campaign, appeared difficult due to its low extract concentration and abundance of similar products. For practical reasons, the second part of this study concentrated on neoruscogenin (a close analogue of (25S)-ruscogenin), which is readily commercially available in gram quantities and more potent in terms of RORα activity. Indeed, the primary objective of this study was to rapidly obtain a pharmacological probe to be used for in vitro and in vivo validation studies.

Selectivity of compounds for a chosen target and the bioavailability in experimental rodent models are important selection parameters for high-quality pharmacological probes. Indeed, the selectivity of certain classes of natural compounds such as the flavonoids is quite low, and numerous mechanistically unrelated pharmacological activities have been described for such compounds.²⁷ On the contrary, we found that neoruscogenin did not show any significant effect on both RORβ and RORγ or on several other nuclear receptors that naturally interact with diverse steroid ligands.

The bioavailability and organ exposure to neoruscogenin in mouse were sufficient to obtain a clear effect on target-related hepatic gene expression. The effect on gene expression in the CNS was not examined, but judging from the PK data, brain tissue exposure (C_max = 150-fold EC₅₀) was sufficient to induce target gene regulation and confer pharmacologic action. In medicinal chemistry programs that start from small molecules derived from synthetic libraries, it can be challenging to obtain potent and bioavailable compounds within a couple of months. Therefore, we consider that the procedure described in this article is a real alternative solution when pharmacological in vivo proof of concept is needed to make educated decisions about an orphan target.

The results presented in this article validate two major strategic objectives. First, we have validated a robust procedure to identify potent and bioavailable pharmacological probes in complex plant extract fractions. Second, we demonstrated that carefully designed natural ligand identification procedures can rapidly deliver reliable chemical tools to characterize difficult orphan targets. Ruscogenins were identified among other biologically active compounds in extracts of Radix Ophiopogon japonicus and Ruscus aculeatus, plants that are known for their medicinal properties. Pure ruscogenin administration in a mouse model of pulmonary disease has favorable effects on hemodynamics, pulmonary vascular remodeling, and prevention of pulmonary hypertension development.²⁸ In line with these findings, ruscogenins showed anti-inflammatory,^28,29 antithrombotic,³⁰ and antiapoptotic²⁸ activities in vitro. Future research will demonstrate which of these properties directly rely on RORα activation.

Footnotes

Acknowledgements

Bicoll members dedicate this article to Prof. Dr. Joachim Treusch, on the occasion of his 70th birthday.

Supplementary material for this article is available on the Journal of Biomolecular Screening Web site at .

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was partly supported by research grants from German Federal Ministry of Education and Research (BMBF, 0315012A and 0315944A) and by the EuroTransBio/OSEO grant No. PRESAGE 37149.

References

Harvey

A. L.

Natural Products in Drug Discovery. Drug Discov. Today. 2008, 13, 894–901.

Mishra

B. B.

Tiwari

V. K.

Natural Products: An Evolving Role in Future Drug Discovery. Eur. J. Med. Chem. 2011, 46, 4769–4807.

Feher

Schmidt

J. M.

Property Distributions: Differences between Drugs, Natural Products, and Molecules from Combinatorial Chemistry. J. Chem. Inf. Comput. Sci. 2003, 43, 218–227.

Newman

D. J.

Cragg

G. M.

Natural Products as Sources of New Drugs over the 30 Years from 1981 to 2010. J. Nat. Prod. 2012, 75, 311–335.

Quinn

R. J.

High-Throughput Screening in Natural Product Drug Discovery in Australia Utilising Australia’s Biodiversity. Drug Dev. Res. 1999, 46, 250–254.

Eldridge

G. R.

Vervoort

H. C.

Lee

C. M.

. High-Throughput Method for the Production and Analysis of Large Natural Product Libraries for Drug Discovery. Anal. Chem. 2002, 74, 3963–3971.

Chen

Nuclear Receptor Drug Discovery. Curr. Opin. Chem. Biol. 2008, 12, 418–426.

Overington

J. P.

Al-Lazikani

Hopkins

A. L.

How Many Drug Targets Are There?

Nat. Rev. Drug Discov. 2006, 5, 993–996.

Hamilton

B. A.

Frankel

W. N.

Kerrebrock

A. W.

. Disruption of the Nuclear Hormone Receptor RORalpha in Staggerer Mice. Nature 1996, 379, 736–739.

10.

Sidman

R. L.

Lane

P. W.

Dickie

M. M.

Staggerer, a New Mutation in the Mouse Affecting the Cerebellum. Science 1962, 137, 610–612.

11.

Jetten

A. M.

Retinoid-Related Orphan Receptors (RORs): Critical Roles in Development, Immunity, Circadian Rhythm, and Cellular Metabolism. Nucl. Recept. Signal. 2009, 7, e003.

12.

Kallen

Schlaeppi

J. M.

Bitsch

. Crystal Structure of the Human RORalpha Ligand Binding Domain in Complex with Cholesterol Sulfate at 2.2 A. J. Biol. Chem. 2004, 279, 14033–14038.

13.

Kallen

J. A.

Schlaeppi

J. M.

Bitsch

. X-ray Structure of the hRORalpha LBD at 1.63 A: Structural and Functional Data That Cholesterol or a Cholesterol Derivative is the Natural Ligand of RORalpha. Structure 2002, 10, 1697–1707.

14.

Wang

Kumar

Crumbley

. A Second Class of Nuclear Receptors for Oxysterols: Regulation of RORalpha and RORgamma Activity by 24S-hydroxycholesterol (cerebrosterol). Biochim. Biophys. Acta. 2010, 1801, 917–923.

15.

Wang

Kumar

Solt

L. A.

. Modulation of Retinoic Acid Receptor–Related Orphan Receptor Alpha and Gamma Activity by 7-oxygenated Sterol Ligands. J. Biol. Chem. 2010, 285, 5013–5025.

16.

Kim

E. J.

Yoon

Y. S.

Hong

. RORalpha-Induced Activation of AMP-Activated Protein Kinase Results in Attenuation of Hepatic Steatosis. Hepatology 2011, 55, 1379–1388.

17.

Kumar

Kojetin

D. J.

Solt

L. A.

. Identification of SR3335 (ML-176): A Synthetic RORalpha Selective Inverse Agonist. ACS Chem. Biol. 2011, 6, 218–222.

18.

Wang

Kumar

Nuhant

. Identification of SR1078, a Synthetic Agonist for the Orphan Nuclear Receptors RORalpha and RORgamma. ACS Chem. Biol. 2010, 5, 1029–1034.

19.

Chopra

A. R.

Louet

J. F.

Saha

. Absence of the SRC-2 Coactivator Results in a Glycogenopathy Resembling Von Gierke’s Disease. Science 2008, 322, 1395–1399.

20.

Enke

C. G.

Nagels

L. J.

Undetected Components in Natural Mixtures: How Many? What Concentrations? Do They Account for Chemical Noise? What Is Needed to Detect Them?

Anal. Chem. 2011, 83, 2539–2546.

21.

González

A. G.

Freire

García-Estrada

M. G.

. New Sources of Steroid Sapogenins—XIV: 25S-ruscogenin and Sansevierigenin, Two New Spirostan Sapogenins from Sansevieria trifasciata. Tetrahedron 1972, 28, 1289–1297.

22.

Akashi

Takumi

The Orphan Nuclear Receptor RORalpha Regulates Circadian Transcription of the Mammalian Core-Clock Bmal1. Nat. Struct. Mol. Biol. 2005, 12, 441–448.

23.

Kumar

Solt

L. A.

Conkright

J. J.

. The Benzenesulfoamide T0901317 [N-(2,2,2-trifluoroethyl)-N-[4-[2,2,2-trifluoro-1-hydroxy-1-(trifluoromethyl)ethy l]phenyl]-benzenesulfonamide] Is a Novel Retinoic Acid Receptor–Related Orphan Receptor-Alpha/Gamma Inverse Agonist. Mol. Pharmacol. 2010, 77, 228–236.

24.

Nakajima

Ikeda

Kimura

. Bidirectional Role of Orphan Nuclear Receptor RORalpha in Clock Gene Transcriptions Demonstrated by a Novel Reporter Assay System. FEBS Lett. 2004, 565, 122–126.

25.

Wada

Kang

H. S.

Angers

. Identification of Oxysterol 7alpha-hydroxylase (Cyp7b1) as a Novel Retinoid-Related Orphan Receptor Alpha (RORalpha) (NR1F1) Target Gene and a Functional Cross-Talk between RORalpha and Liver X Receptor (NR1H3). Mol. Pharmacol. 2008, 73, 891–899.

26.

Kang

H. S.

Okamoto

Takeda

. Transcriptional Profiling Reveals a Role for RORalpha in Regulating Gene Expression in Obesity-Associated Inflammation and Hepatic Steatosis. Physiol. Genomics 2011, 43, 818–828.

27.

Batovska

D. I.

Todorova

I. T.

Trends in Utilization of the Pharmacological Potential of Chalcones. Curr. Clin. Pharmacol. 2010, 5, 1–29.

28.

L. Q.

Zhu

Kong

. Ruscogenin Attenuates Monocrotaline-Induced Pulmonary Hypertension in Rats. Int. Immunopharmacol. 2013, 16, 7–16.

29.

Sun

Chen

Gao

. Ruscogenin Inhibits Lipopolysaccharide-Induced Acute Lung Injury in Mice: Involvement of Tissue Factor, Inducible NO Synthase and Nuclear Factor (NF)–kappaB. Int. Immunopharmacol. 2012, 12, 88–93.

30.

Kou

Tian

Tang

. Antithrombotic Activities of Aqueous Extract from Radix Ophiopogon japonicus and Its Two Constituents. Biol. Pharm. Bull. 2006, 29, 1267–1270.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.46 MB