Discovery and SAR analysis of phenylbenzo[d][1,3]dioxole-based proprotein convertase subtilisin/kexin type 9 inhibitors

Abstract Proprotein convertase subtilisin/kexin type 9 (PCSK9) has emerged as a novel therapeutic target for the development of cholesterol-lowering drugs. In the discovery of PCSK9/LDLR (low-density lipoprotein receptor) protein-protein interaction (PPI) impairing small molecules, a total of 47 phenylbenzo[d][1,3] dioxole-based compounds were designed and synthesised. The result revealed that the 4-chlorobenzyl substitution in the amino group is important for the PPI disrupting activity. In the hepatocyte-based functional tests, active compounds such as A12, B1, B3, B4 and B14, restored the LDLR levels on the surface of hepatic HepG2 cells and increased extracellular LDL uptake in the presence of PCSK9. It is notable that molecule B14 exhibited good performance in all the evaluations. Collectively, novel structures targeting PCSK9/LDLR PPI have been developed with hypolipidemic potential. Further structural modification of derived active compounds is promising in the discovery of lead compounds with improved activity for the treatment of hyperlipidaemia-related disorders.


Introduction
Lipid disorder conditions are common risk factors for cardiovascular diseases, which may occur with various factors, such as high fat diet 1 or environmental pollutants 2 . Among the lipid disorder conditions, high blood level of low-density lipoprotein cholesterol (LDL-C) has been revealed to be a key risk factor for cardiovascular conditions such as adverse atherosclerotic cardiovascular events 3 . Low-density lipoprotein receptors (LDLRs) are responsible for the uptake of cholesterol-carrying lipoprotein particles into cells and are critical in regulating the amount of LDL-C in blood 4 . LDLRs on the surface of liver cells are of particular importance for the clearance of blood LDL-C due to their primary role in removing excess cholesterol from the body. The amount of hepatic LDLR proteins is mainly regulated by proprotein convertase subtilisin/kexin type 9 (PCSK9), a proprotein convertase family protease. It is identified that LDLR is degraded along with the complexed LDL particle in lysosomes by binding to PCSK9 5 . Thus, PCSK9 is closely related to dyslipidemia resulting from decreased LDLR levels and the consequent increased circulating LDL-C.
It was reported that PCSK9 binds to the EGF(A) domain of LDLR at the surface of liver cells 6 . Impairing the PCSK9/LDLR protein-protein interaction (PPI) has been employed to improve the hepatic LDLR population and decrease circulating LDL-C 7 . In 2015, two PCSK9 monoclonal antibodies (mAbs), alirocumab 8 and evolocumab 9 , were approved by the FDA of the United States for the treatment of hyperlipidaemia. The PCSK9 mAbs exhibited remarkable clinical benefit without obvious side effects compared with the statin therapy 10,11 . However, the wide application of PCSK9 mAbs is restricted by the expensive price and inconvenient intravenous administration.
Development of small molecule PCSK9/LDLR PPI inhibitors is promising in the discovery of oral effective and low-cost drugs for hyperlipidemic therapy 12,13 . There are two kinds of small molecule PCSK9 inhibitors developed for blocking the function of PCSK9, PCSK9/LDLR PPI impairing molecules and PCSK9 synthesis/expression inhibitors 7,14 . Interfering with PCSK9/LDLR PPI has been explored by using human mAbs, peptides, peptidomimetics and small molecules [15][16][17][18] . By analysing the structure of PCSK9/LDLR binding domain, lead compounds, such as CB_36 19 , SBC-110,424 and SBC-115,076 (WO 2016/040305 A1) were derived by in silico virtual screening approach (Figure 1). Compound SBC-115,337 was developed by structural modification of previously derived molecules. The b-strand peptidomimetic MeIm 20 and RIm13 21 could effectively block the PCSK9/LDLR PPI by disturbing the b-strand mediated interactions. SRX200 (WO 2016/029037 A1) was considered to be an allosteric inhibitor by binding to an allosteric site of PCSK9.
In the current study, small molecule PCSK9 inhibitors were developed by targeting PCSK9/LDLR PPI. Due to the flat and featureless conformation of the PCSK9/LDLR binding domain ( Figure  2), hydrophobic structures with large surface area were considered to be favourable for binding of inhibitors to PCSK9 and impairing the PPI. Because of the existence of fused rings and bi-aromatic rings in the structure of PCSK9/LDLR PPI inhibitors, 5-phenylbenzo[d] [1,3]dioxole group was selected for occupying the PCSK9/ LDLR PPI interface. In the current study, 4-(6-aminobenzo[d] [1,3]dioxol-5-yl)benzoic acid was utilised as the core structure for the design of novel molecules to impair the PCSK9/LDLR PPI. Different substitutions were introduced to the carboxyl-and amino-group for the SAR analysis. The activities of the derived compounds were evaluated in the PCSK9/LDLR PPI inhibitory assay, hepatic cell-based LDLR expression test and LDL uptake investigation.

Chemistry
The target molecules were synthesised as described in Schemes 1-3. The commercially available benzo[d] [1,3]dioxol-5-amine (1) was used as the starting material for the compound synthesis. Firstly, the amino group of 1 was protected by Boc group for bromine substitution. Then, the synthesised intermediate 3 was condensed with 4-(Methoxycarbonyl)benzeneboronic acid via Suzuki coupling to afford the key intermediate 4. De-esterification under alkaline condition allowed the introduction of various substituted benzenamines and piperazines to the carboxy group. Substituted benzoic acids were introduced by deprotection of amino group and condensation reactions. Target molecules of the A series were synthesised by 3,4,5-trimethoxyaniline substitution in the carboxyl group and introduction of different benzoic acids to the amino group (Scheme 1). In the B series compounds, different amines were firstly introduced to the carboxyl group, then 4-(chloromethyl)benzoic acid substitution was performed on the amino group (Scheme 2). The C series compounds were derived by carboxyl substitution, demethylation and alcoholysis, and the final amino condensation (Scheme 3).

PCSK9/LDLR PPI inhibitory screening
The PCSK9/LDLR time-resolved fluorescence resonance energy transfer (TR-FRET) assay was performed to evaluate the activity of synthesised compounds in inhibition of PCSK9-LDLR binding. The A series compounds were firstly synthesised for the PCSK9/LDLR PPI inhibitory test. The results revealed that the 4-chlorobenzyl substituted compound A12 has the highest inhibitory activity with an inhibitory rate of 45.1% at the concentration of 10 mM (Table 1). Therefore, a group of A12 derivatives (B series) were synthesised with 4-chlorobenzyl substituted in the amino group and various substitutions introduced to the carboxylic group. Among the B series compounds, several molecules exhibited improved inhibitory activity compared with A12, such as B1, B3, B4 and B14 (Table 2). To confirm whether the 4-chloromethyl group is essential for the PPI inhibitory activity, a new series of compounds (C series) were synthesised with the 4-chloromethyl group replaced by 4-methoxymethyl moiety. The inhibitory activity was observed, most likely resulting from the 4-methoxymethyl substitution ( Table 3). The dose-dependent IC 50 calculation revealed that molecule B14 had high PCSK9/LDLR PPI inhibitory potential compared with SBC-115337 ( Figure 3). These results  suggested that molecules with 4-chlorobenzyl substitution had the potency in impairing PCSK9/LDLR PPI, indicating that the 4chloromethyl group in the active compounds is essential for the inhibitory activity.
To acquire clues for the structural modification of the derived compounds, SAR analysis was performed based on the PCSK9/ LDLR PPI inhibitory activities. Among the A series compounds (Table 1), fluorine and trifluoromethyl substitutions in the meta and para positions of phenyl ring (in R group) are favoured for the inhibitory activities, such as A1, A2, A9, and A11. While the difluoro substitutions did not significantly enhance the activities, such as A15 and A16. Chlorine, methoxy and alkyl groups substituted in the phenyl ring did not improve the PPI inhibitory activities. Molecule A12 with chloromethyl substitution in the R group exhibited the best inhibitory activity in the A series compounds.
With the 4-chloromethyl-benzoyl group substituted in the amino group, the B series compounds were synthesised by the introduction of different substitutions to the carboxyl group. From B1 to B12, halogen and hydroxyl substitutions in the para position of phenyl ring increased the PPI inhibitory activities, such as B1, B2, B3 and B4 ( Table 2). The alkoxyl-substituted compounds, such as B5, B6, B7 and B12, did not show improved inhibitory activities. Alkyl and phenyl groups substituted in the para position of R moiety did not exhibit positive effects on enhancing the inhibitory activity. Among the piperazine containing R groups, methoxy substitution in the ortho position of phenyl ring obviously increased the PCSK9/LDLR PPI inhibitory activity.
Activity decrease caused by the replacement of 4-chloromethyl group with 4-methoxymethyl in the C series compounds revealed the importance of 4-chloromethyl substitution for the inhibitory potency of the compounds (Table 3). Among the C series molecules, methoxy, fluorine and chlorine substituted the para position of phenylpiperazine ring, such as C3, C7 and C8. The methoxy group substituted in the ortho position decreased the inhibitory activity. Molecule C1 with 3,5-dimethoxyphenylamino R group also exhibited potency in the inhibition of PCSK9/LDLR PPI.

LDLR expression assay
LDLR is degraded in lysosomes by binding to PCSK9. Thus, inhibition of PCSK9/LDLR PPI could increase the LDLR level on the surface of hepatocytes. In the current study, human hepatic HepG2 cells were selected for the investigation of the active compounds for LDLR expression. SBC-115337, developed by Shifa Biomedical Corporation, as a potent PCSK9/LDLR PPI inhibitor was utilised as the positive control. All the synthesised compounds were evaluated to be safe with low inhibitory ratios at the dose of 10 mM in the MTT assay (Tables 1-3). Then, in cell western (ICW) assay was performed to detect the LDLR level on the surface of HepG2 cells. The results showed that the presence of PCSK9 significantly decreased the LDLR level compared with the control (Figure 4). The tested compound A12, B1, B3, B4 and B14 restored the LDLR protein level in a dose dependent manner compared with the positive control SBC-115337. Remarkably, the proportion of LDLR increased from 34.33% in cells treated with PCSK9 alone to 50.55% and 66.51% in cells treated with the molecule B14 at the concentrations of 5 mM and 10 mM, respectively. These results indicated that these representative compounds are effective in improving the LDLR levels induced by inhibition of PCSK9/ LDLR PPI.

LDL uptake test
Extracellular LDL uptake assay was performed to evaluate the functional effects of PCSK9/LDLR PPI inhibition. The effects of active compounds in improving the capacity of HepG2 cells to uptake the fluorescent LDL were investigated using SBC-115337 as

Binding pattern analysis
Molecular docking was performed to predict the binding mode of the active molecule B14 in the LDLR binding domain of PCSK9. As illustrated in Figure 6(a), there was no obvious binding pocket in the PCSK9/LDLR PPI interface. Hydrophobic interactions play an important role in the binding of molecule B14 to the LDLR binding site. It was revealed that Phe379 is a key residue with Pi-Pi stacking interactions with phenyl rings of molecule B14 ( Figure  6(b)). The surrounding residues, such as Pro155, Ala239, Ile369, Thr377, Cys378 and Val380 are also significant in the hydrophobic interactions formed between molecule B14 and PCSK9. Hydrogen bond interaction generated between NH of Arg194 and CO of B14 also makes contributions to the ligand-receptor binding. These results indicated that the flat structure of molecule B14 is suitable for matching the surface of PCSK9/LDLR PPI. Further structural modification of molecule B14 could be performed by    Compared with the control group, PCSK9 treatment significantly decreases the LDL uptake of HepG2 cells ( ÃÃ p < 0.01). The synthesised compounds restore the ability of HepG2 cells to uptake LDL with a dose dependent manner comparing with SBC-115337 ( Ã p < 0.05, ÃÃ p < 0.01, the PCSK9 only control was used as a comparator). increasing hydrogen bond interactions with surrounding polar residues.

Conclusion
Hyperlipidaemia is a major risk factor for cardiovascular diseases, and PCSK9 inhibition has emerged as a novel cholesterol-lowering therapy. In order to discover small molecules that could disrupt PCSK9/LDLR PPI, a total of 47 phenylbenzo[d] [1,3]dioxole containing compounds were designed and synthesised. The derived compounds were tested in the molecular PCSK9/LDLR PPI inhibitory screening, LDLR expression on the surface of hepatocytes, and hepatic-cell-based extracellular LDL uptake. Several compounds, such as A12, B1, B3, B4 and B14 exhibited potency in impairing the PCSK9/LDLR PPI. SAR analysis revealed the importance of 4chlorobenzyl substitution in the amino group for inhibitory activity. In the LDLR expression and LDL uptake studies, the tested molecules restored both LDLR expression and fluorescent LDL uptake of HepG2 cells in the presence of PCSK9. Comparing with SBC-115337, molecule B14 showed remarkable performance in the cell based functional tests. In summary, a potent lead compound (B14) was developed in the current study for the discovery of small molecule PCSK9 inhibitors that directly disrupt PCSK9/ LDLR PPI. Further structural modification of the derived active compounds would be promising in the development of hypolipidemic small molecules with improved potency.

Materials and methods
All chemicals were obtained from commercial suppliers and can be used without further refinement. All reactions were detected by TLC using 0.25 mm silica gel plate (60GF-254). UV light and ferric chloride were used to show TLC spots. 1 H NMR and 13 C NMR spectra were recorded on a Bruker DRX spectrometer at 500 MHz, using TMS as an internal standard. High-resolution mass spectra were recorded using a Thermo Scientific Q Exactive hybrid quadrupole-orbitrap mass spectrometer from Weifang Medical University.

PCSK9-LDLR TR-FRET assay
The compounds were dissolved in 100% DMSO; then 2 mL of the dilution was added to a 20 mL of reaction to keep final concentration of DMSO less than 1% in all of reactions. The binding reaction was conducted at room temperature. The reaction mixture in assay buffer contains PCSK9, the indicated amount of the inhibitor, ligand LDLR, and the reaction dyes. The reaction mixture was incubated for 120 min before detection of the TR-FRET signal. Fluorescence signals for both the donor and acceptor dyes were measured using a Tecan Infinite M1000 plate reader. TR-FRET was recorded as the ratio of the fluorescence of the acceptor and the donor dyes (acceptor/donor). The TR-FRET data were analysed using Graphpad Prism software.

Molecular docking
Molecular docking was performed using Glide in Schrodinger Suites 2018. Crystal structure of PCSK9 (PDB Entry: 3GCX) was derived from RCSB protein data bank (www.rcsb.org). Structural modifications were performed by Protein Preparation Wizard. The LDLR chain and embedded water molecules in the protein structure were removed. The default OPLS3 force field was assigned to the refined protein. The structure of ligand B14 was sketched by maestro and prepared by LigPrep. The active site was set to be an enclosing box centred on residue Phe379 with side length of 20 Å. Extra precision was selected for the docking process, and other parameters were set as default.
Cell culture and ICW assay The HepG2 cell line was grown in a humidified incubator with 5% CO 2 at 37 C and cultured in DMEM-High glucose medium with 10% (v/v) heat-inactivation foetal bovine serum, 2 mM L-glutamine and 100 U/ml penicillin-streptomycin. The expression level of LDLR in HepG2 cells was determined by In-Cell Western (ICW) Assay as described previously 16 . Briefly, cells were plated in a 96-well plate (2 Â 10 4 cells per well) and cultured for more than 24 h, followed by washing with PBS and starving overnight in DMEM without FBS. HepG2 cells were treated with 4.0 lg/ml PCSK9 (K9) alone and the tested compounds with the presence of 4.0 lg/ml of K9, and vehicle (PBS) for 8 h. Then cells were washed with PBS one time and fixed in 4% paraformaldehyde for 20 min. Cells were washed 5 times with PBS, then were blocked with 5% BSA in PBS for 1 h. LDLR monoclonal Antibody solution (1:1000 in 5% BSA in PBS, 50.0 lL/well) was incubated overnight at 4 C. Subsequently, the cells were washed 5 times with PBS. Goat anti-mouse IgG-HRP secondary antibody solution (1:2000 in 5% BSA in PBS, 50.0 lL/ well) was added and incubated 1 h at RT. Then cells were washed 5 times with PBS and added with TMB substrate, then incubated at RT until desired colour was developed. 50.0 lL/well of 2 M H 2 SO 4 were added to stop the reaction and the absorbance at 450 nm was measured using a plate reader (EnSpire, Perkin Elmer Corporation).
Fluorescent LDL uptake assay HepG2 cells (2 Â 10 4 cells per well) were seeded in black 96-well plates and cultured for more than 24 h. The following day, cells were treated with 4.0 lg/ml K9 alone and the tested compounds with the presence of 4.0 lg/ml of K9, and vehicle (PBS) for 8 h. At the end of the treatments, the culture medium was replaced with 50.0 ll/well LDL-DyLight TM 550 working solution (Cayman Chemical Company, US). The cells were additionally incubated for 16 h at 37 C. Then the cells were washed with PBS one time, then added PBS with 100.0 ll/well. The degree of LDL uptake was measured using a plate reader (EnSpire, Perkin Elmer Corporation) at excitation wavelengths 540 nm and emission wavelengths 570 nm.

Statistical analysis
All experiments were repeated at least three times unless otherwise stated. The data were represented as mean ± SD. Statistical analysis were performed with Student's t test for two group comparisons and using one-way ANOVA with Tukey's post hoc test for multigroup comparisons. p < 0.05 or p < 0.01 were considered statistically significant.

Disclosure statement
No potential conflict of interest was reported by the author(s).