The Greenpharma database (GPDB)
GPDB was developed to exploit the steadily growing botanical data, natural chemical structures, and the knowledge obtained from biological tests with vegetal extracts, or isolated molecules, or from scientific literature in general. In addition, it covers phytotaxonomic issues such as the family, genus, and species of organisms, their common names and synonyms, and it also provides information on their applications in traditional medicine (such as the organs targeted in the use of these organisms), as well as indicating the ethnic groups from which the data was collected. GPDB has demonstrated its usefulness in accelerating the discovery of anti-inflammatory compounds.33
Currently, the GPDB contains 150,000 molecule and 161,000 organism entries.
In the present work, we focus on this set of natural molecules as input for Selnergy screening on our targets of interest.
In silico screening with Selnergy
Protein 3D structures were either retrieved directly from published crystal structures at the Protein Data Bank34
(PDB; see http://www.rcsb.org/pdb
) or built by homology modelling.35
The crystal structures of SHBG and aromatase were downloaded from PDB (PDB: 1D2S36
). However, in the case of 5α-reductase 1, no crystal structure was available, and its three-dimensional (3D) model was generated by homology modelling.35
The accepted assumption for this approach is that two proteins with almost identical and highly phylogenetically related (homologous) amino acid sequences will share similar 3D structures. Hence, the 5α-reductase 1 model was safely constructed based on the crystal structure of 5β-reductase 1 (PDB: 3CAS38
) thanks to a 40% sequence homology, in view of a commonly accepted threshold of 25% to 30% of the minimal homology required for a length of 80 to 100 residues (typical domain length). Homology modeling was performed with biopolymer and composer modules within the Sybyl 8.0 package (Tripos International, St Louis, MO).39
was conducted with Selnergy.32
A candidate was considered to be “accepted” for virtual screening if it fulfilled the following conditions: from a test set of molecules consisting of ligands with known target activities and others that were randomly selected, Selnergy ranks that candidate among the best scoring structures.
An unattended post-processing procedure was performed to discard docking solutions falling outside active site: a spatial fit criterion is taken into account to determine whether a ligand was “correctly” docked into the binding cavity. This goal was achieved by comparing the distances between two centroids: one defining the active site of the studied protein and the other defining the docked compound. Each measured the distance from centroid to centroid: (dC-C
), which served as the spatial fit criterion. Any docking position with a dC-C
> 4 Å was considered as being “out of the protein active site,” and the protein/ligand pair was discarded from further analysis. As Selnergy can propose several poses for the same molecule, the docked pose with the shortest dC-C
was selected. A macro was written in Sybyl programming language and implemented to automate this procedure. The Sybyl programming language macro is distributed by Tripos.39
Finally, the automated screening and docking processes ended when an expert inspected the outcome based on intuitive grounds in order to discard false-positives40
in accordance with the similarities of their binding modes, and with reference to co-crystallized ligands.
Biological assays: aromatase and 5α-reductase activity tests
The aim of the bioassay was to determine the in vitro potency of putative inhibitors – found by our in-silico-screening tool, Selnergy32
– of the aromatase enzyme using human placental microsomes as a source of the enzyme, and with titrated water acting as a tracer.41
The approved aromatase inhibitor letrozole served as a reference compound.
Furthermore, in the present study the in vitro potency of honokiol as an inhibitor of the 5α-reductase isoenzymes type 1 and type 2 was investigated. Transfected human embryonic kidney cells (HEK293), stably expressing the respective isoenzymes 1 or 2 (HEK293-5α1 = HEK 1; HEK293-5α2 = HEK 2), were used as a whole-cell test system.42
On the basis of the enzymatic conversion of the C-labeled substrate androstenedione to the 5α-reductase product dihydroandrostenedione,14
the inhibitory potency was measured by photo-stimulated luminescence and was compared to those of finasteride,43
a well-known 5α-reductase inhibitor.
Aromatase assay: microsomal aromatase enzyme preparation
The microsomal fraction was prepared from freshly delivered human term placenta. The tissue was washed in ice-cold 0.15 M KCl and freed of membranes and blood vessels. The tissue was placed in 0.25 M sucrose (1.0 mL/g tissue) and cut into small pieces with surgical scissors. The tissue was homogenized in an Ultra-Turrax T25®
Works, Inc, Wilmington, NC) using twenty 10 s bursts at 20,000 rpm with 50 s cooling periods. Portions (50 mL) of the homogenate were subjected to a Kinematica Polytron ultrasonic homogenizer (Labotal Scientific Equipment, Ltd, Abu Gosh, Israel) using five 15 s bursts at speed 5 with 15 s cooling periods. The homogenate was centrifuged at 20,000 g
for 50 min (Hitachi centrifuge; Hitachi Koki Co, Ltd, Tokyo, Japan). The clear supernatant was centrifuged at 148,000 g
for 65 min (Ultraspin; LKB Instruments, Bromma, Sweden). The pellets were washed in phosphate buffer (0.05 M; pH 7.4) using a Teflon homogenizer. The washed pellets were pooled, dissolved in about 50 mL of phosphate buffer, and washed twice by centrifugation at 100,000 g
for 60 min. The final pellet was resuspended in 45 mL of phosphate buffer, and 100 μL aliquots were made from a stirring suspension, which was snap frozen and stored at −70°C. Protein contents were determined by the method of Lowry et al,45
using a SpectraMax Plus384 (Molecular Devices, LLC, Sunnyvale, CA) and bovine serum albumin as a standard.
Aromatase assay: in vitro aromatase inhibition experiment
Incubations were performed at 37°C containing [1β-3H]-androstenedione (250 nM), excess NADPH (0.24 mM), 20 μg of human placental protein, and phosphate buffer (0.05 M; pH 7.4). The final volume of the incubation mixture was 1.0 mL. Control incubations were performed without the inhibitor, and the background was determined in an incubation device without any enzymes. The inhibitor was dissolved in MeOH and serially diluted to reach the three test concentrations of 100, 10, and 1 μM. Letrozole was tested at 10 nM.
The reaction was started by the addition of protein and stopped after 20 min by the addition of dichloromethane (10 mL). Following extraction, the tubes were centrifuged at 2,000 g for 5 min in a bench top centrifuge. Thereafter, 0.5 mL of the aqueous phase was removed and placed in a tube to which 1.0 mL of 5% charcoal suspension (Norit-activated; Norit Pharmaceuticals, AC Amersfoort, Netherlands) was added. After shaking the tubes for at least 15 min at room temperature, the tubes were centrifuged at 2,500 g for 10 min. Two 0.5 mL aliquots were removed, added to scintillation vials containing 10 mL scintillation fluid (Quickszint 212; Zinsser Analytic GmbH, Frankfurt, Germany), and were counted in a liquid scintillation counter (1209 RACKBETA, LKB Wallac; LKB Instruments). Counts from the control incubations containing no enzyme were subtracted from each of the incubation counts.
Aromatase assay: data analysis
Results were expressed as the percent of inhibition relative to untreated controls. Inhibition rates were calculated out of the mean conversion rates with (n = 2) and without inhibitor (n = 4).
5α-reductase assay: cell culture
HEK 1 and HEK 2 cells were cultivated in Dulbecco’s modified Eagle’s medium (Invitrogen; Life Technologies Corporation, Carlsbad, CA) (pH 7.4) with 10% fetal calf serum, penicillin/streptomycin (100 U/mL and 100 μg/mL, respectively), and 0.5 mg/mL of Geneticin-418-sulfate. Inhibition assays were performed with HEK 1 and HEK 2 cells seeded at a concentration of 0.25 × 106 cells per 1.9 cm2 and were incubated for 20 h for attachment, doubling, and differentiation in a humidified 5% CO2 atmosphere at 37°C.
5α-reductase assay: in vitro inhibition assays
Test compounds were dissolved in dimethyl sulfoxide (DMSO) and serially diluted in Dulbecco’s modified Eagle’s medium to reach the three final test concentrations of 100 μM, 10 μM, and 1 μM. Incubation mixtures containing 0.24 mM NADPH and 50 nM [4-14C]-androstenedione at a final volume of 500 μL were pre-incubated at 37°C for 10 min. Finasteride used as an internal control was also dissolved in DMSO and diluted to the test concentrations of 100 μM and 800 μM. Controls containing the solvent only (1% DMSO) were treated the same way.
The reaction was started by removing the culture medium and adding the pre-warmed incubation mixture to the cell layers. After 30 min (HEK 1) or 15 min (HEK 2), respectively, the reaction was stopped by removing the supernatant. For the extraction of product and nonconverted substrate, 500 μL ethyl acetate was added to each sample. After 10 min shaking, samples were centrifuged for 5 min for phase separation, and the supernatant was transferred into fresh tubes. After evaporation of the solvent, the dried residues were reconstituted in 25 μL of acetone.
The reconstitutes were spotted on a high performance thin-layer-chromatography (HPTLC) plate (20 cm × 10 cm, Silicagel 60F254 with concentrating zone). The HPTLC plates were run twice in a freshly prepared solution of dichloromethane:diethylether (8:2) as a solvent. Imaging plates were exposed to the HPTLC plates for 48 h. The imaging plates were scanned using a Phosphorimager, and the spots corresponding to [4-14C]-androstenedione (A) and [4-14C]-dihydroandrostenedione were integrated using the corresponding software.
5α-reductase assay: data analysis
Results were displayed as photostimulated luminescence units photo and were corrected for the background. Conversion rates were calculated according to the following formula:
Inhibition rates, expressed as the percent of inhibition values relative to untreated controls, were calculated out of the mean conversion rates both with (n = 2) and without inhibitor (n = 2).
Quality control: acceptability of the assays
As quality controls, positive control inhibitors were included in the assay set-up. The in vitro assays are considered as acceptable if it meets the following criteria:
- A 50% inhibition of aromatase by Letrozole at a concentration of 10 nM (tested concentration corresponding to the IC50 of Letrozole).
- A 50% inhibition of 5α-reductase type 1 by Finasteride at a concentration of 800 nM (tested concentration corresponding to the IC50 of Finasteride).
- A 50% inhibition of 5α-reductase type 2 by Finasteride at a concentration of 100 nM (tested concentration corresponding to the IC50 of Finasteride).
Clinical evaluation: patch test study
Previous to the patch test study, honokiol 1% and 5% was evaluated for its irritating potential on the chorio-allantoic membrane of hen eggs and for its mutagenicity potential on the Ames test according to OCDE471. This test evaluates the acute cutaneous tolerance of the raw material and was designed and evaluated by Dermscan (Villeurbanne Cedex, France). A basic cream formula was used including 1% and 5% of honokiol. The study was realized during 48 hours using an occlusive method on 20 volunteers with normal skin.
Clinical study: evaluation of the product
The randomized, double-blind comparative study was carried out in two parallel groups: one group tested the placebo preparation (cream without the active ingredient), and the other group tested the verum product at 1% (that is placebo + 1% honokiol). The study protocol was carried out by Dermscan. The study was conducted from October to December. The objectives were to evaluate the antiaging and redensifying effects of the tested products.
The study involved 40 Caucasian healthy men from 55 to 63 years old (mean: 60 ± 1 years). The main inclusion criteria were wrinkles, crow’s feet around the eyes, and loose skin on the face.
All volunteers were divided exactly into two groups. The verum group received a cream preparation containing 1% of honokiol whereas the placebo group received the same formulation without the active ingredient (honokiol). Facial applications (on crow’s feet around the eyes) were made twice a day for 2 months. The kinetic data were assessed at three sampling points: day 0, day 28, and day 56. Our assessment criteria were as follows:
- Skin print analysis with the skin image analyzer.
- Dermis density by Dermscan.
- Macrophotographs of crow’s feet (for 10 volunteers in each group).
For statistical analysis, Microsoft Excel 2000 version 9.0 (Microsoft, Redmond, WA) was used. The statistical method used was the Student’s paired t-test.