Trans-(±)-TTPG-B Attenuates Cell Cycle Progression and Inhibits Cell Proliferation on Cholangiocarcinoma Cells

This research aimed to determine the target protein and molecular mechanism of trans-(±)-kusunokinin ((±)-KU) derivatives (trans-(±)-ARC and trans-(±)-TTPG-B). Molecular docking was used to predict potential synthesized (±)-KU targets among 22 proteins. The (±)-TTPG-B bound HSP90α better than EC44, native (±)-KU and (-)-KU, and (±)-KU and (−)-ARC. In contrast, (−)-ARC bound PI3K more strongly than any other test compound. CSF1R and AKR1B1 were not supposed to be the target of (±)-TTPG-B and (±)-ARC, unlike native (±)-KU. The (±)-TTPG-B bound Tyr139 and Trp162 of HSP90α. Moreover, (−)-ARC bound PI3K via hydrogen bonds and π-π stacking at distinct amino acids, which was different from the other tested compounds. Using half of the IC50 concentration, (±)-TTPG-B, (±)-KU and (±)-ARC enhanced cell cycle arrest at the G0/G1 phase after 12 h and 24 h on KKU-M213 (CCA) cells. The (±)-TTPG-B showed a stronger inhibitory effect than (±)-ARC and (±)-KU on HSP90α, PI3K, HSP90β, c-Myc, AKT, MEK1, CyclinB1, CyclinD1, and CDK1 for 24 and 48 h after treatment with the same concentration (0.015 µM). Thus, trans-(±)-TTPG-B, a newly synthesized compound, has pharmacological potential for development as a target therapy for CCA treatment.


Introduction
Cholangiocarcinoma (CCA), has been reported to have a hepatocyte origin [1] and is an adenocarcinoma that begins in the bile ducts.Most CCA patients are diagnosed with an advanced and aggressive malignancy that cannot be treated with surgery alone [2].Targeted therapy is an interesting treatment for improving overall survival and decreasing the undesirable adverse effects of cancer treatment.This treatment uses drugs that bind to a specific protein in cancer [3].The proteins in the tyrosine kinase pathway are important to cancer growth; therefore, the proteins in this pathway are good targets for targeted therapy.Fibroblast growth factor receptor 2 (FGFR2) is one of the important targets in CCA, and it is targeted by pemigatinib.This drug received approval by the United States Food and Drug Administration for the treatment of adults with unresectable, locally advanced, or metastatic CCA in 2020 [4].Zanidatamab, a human epidermal growth factor receptor 2 (HER2)-targeted bispecific antibody, is suitable for advanced, unresectable, and metastatic HER2-expressing biliary tract cancers [3].Activation of the phosphoinositide-3-kinase (PI3K)/AKT signaling pathway is frequently found in CCA [5].It has been suggested that it is a key step in cancer chemotherapy resistance, especially during DNA-damaging agent treatments such as cisplatin and oxaliplatin [6].Binimetinib (an MEK1/2 inhibitor) is used DNA-damaging agent treatments such as cisplatin and oxaliplatin [6].Binimetinib (an MEK1/2 inhibitor) is used in combination with capecitabine.This drug demonstrates promising antitumor efficacy for CCA patients whose first-line chemotherapy treatments failed, especially in patients with RAS/RAF/MEK/ERK pathway mutations [7].
Due to the different actions of these (±)-KU derivatives, a study on the mechanism of action of (±)-TTPG-B and (±)-ARC on CCA was valuable.First, (±)-ARC and (±)-TTPG-B were predicted target proteins compared with (±)-KU using molecular docking.Then, cell cycle arrest was performed followed by the determination of target proteins and their downstream proteins using Western blotting.

Molecular Docking of (±)-TTPG-B and (±)-ARC
Both (±)-TTPG-B and (±)-ARC were recently reported to have anticancer properties [13].To predict the target protein of both (±)-KU compounds, molecular docking was employed similarly to previous studies [10,11].The binding location and interactions were visualized in order to investigate the binding characteristics.Among the 22 selected proteins were the target protein of native (±)-KU and cancer-related proteins, Table 1.The binding energy (ΔGbind) of the selected synthesized compounds ((−)-ARC, (+)-ARC, (−)-TTPG-B, and (+)-TTPG-B) was lower than that of native (±)-KU and a known inhibitor of the target protein.Due to the different actions of these (±)-KU derivatives, a study on the mechanism of action of (±)-TTPG-B and (±)-ARC on CCA was valuable.First, (±)-ARC and (±)-TTPG-B were predicted target proteins compared with (±)-KU using molecular docking.Then, cell cycle arrest was performed followed by the determination of target proteins and their downstream proteins using Western blotting.

Molecular Docking of (±)-TTPG-B and (±)-ARC
Both (±)-TTPG-B and (±)-ARC were recently reported to have anticancer properties [13].To predict the target protein of both (±)-KU compounds, molecular docking was employed similarly to previous studies [10,11].The binding location and interactions were visualized in order to investigate the binding characteristics.Among the 22 selected proteins were the target protein of native (±)-KU and cancer-related proteins, Table 1.The binding energy (∆G bind ) of the selected synthesized compounds ((−)-ARC, (+)-ARC, (−)-TTPG-B, and (+)-TTPG-B) was lower than that of native (±)-KU and a known inhibitor of the target protein.The docking results were validated by performing a redocking experiment with the known inhibitor and its protein target.In comparison to the corresponding crystal structure, the criteria were based on the position of the docked pose and its binding pocket site.In the absence of the co-crystallized inhibitor, the docked position of the known inhibitor must resemble the reported binding site.All of the docked inhibitors occupied the same binding site as the experimental structure or literature-reported site.
The possible target proteins of (+)-TTPG-B were thus HSP90α and EGFR.Considering the previously reported CSF1R and AKR1B1 as the native (±)-KU target, both synthesized (±)-TTPG-B isomers showed higher ∆G bind with CSF1R but similar ∆G bind with AKR1B1.This implied that the synthesized trans-TTPG-B could act on another target, not only CSF1R and AKR1B1.

Binding Position of (±)-TTPG-B and (±)-ARC and on HSP90α
The binding site and binding interaction(s) between the compound and HSP90α were observed and are concluded in Table 2 and Figure 2. EC44, a known inhibitor of HSP90α, interacted with Tyr139 and Leu107 via π-π stacking and hydrogen bonding, respectively (Figure 2).The native (+)-KU bound to HSP90α via hydrogen bonds at Gly108 and Gln23.
The highlighted amino acids indicate those that interacted with the inhibitor.

Discussion
(±)-KU has an anti-cancer effect and could bind with CSF1R [10] and AKR1B1 [12].In our previous study, we showed that (±)-TTPG-A (or (±)-ARC) and (±)-TTPG-B were (±)-KU derivative compounds that were modified at two binding positions, including 3,4 dimethoxybenzyl butyrolactone and 1,3-benzodioxole parts.Both derivative compounds showed cytotoxicity in breast cancer, CCA, colon, and ovarian cancer cells.(±)-TTPG-B had a greater cytotoxic and apoptotic induction effect than (±)-KU.In addition, (±)-ARC induced cell cycle arrest at the S phase, whereas (±)-TTPG-B caused cell arrest at the G0/G1 phase, which is the same as (±)-KU in KKU-M213 cells [13].In contrast, in this study, half of the IC50 values of (±)-TTPG-B, (±)-ARC, and (±)-KU caused cell cycle arrest at G0/G1 at 12 and 24 h.Therefore, we hypothesized that the specific target protein of these derivatives might be a protein associated with cell signaling and cell proliferation, which remains similar to (±)-KU.Here, we used the molecular docking technique to determine the binding affinity of these two compounds with twenty-two target proteins of the native (±)-KU.

Discussion
(±)-KU has an anti-cancer effect and could bind with CSF1R [10] and AKR1B1 [12].In our previous study, we showed that (±)-TTPG-A (or (±)-ARC) and (±)-TTPG-B were (±)-KU derivative compounds that were modified at two binding positions, including 3,4 dimethoxybenzyl butyrolactone and 1,3-benzodioxole parts.Both derivative compounds showed cytotoxicity in breast cancer, CCA, colon, and ovarian cancer cells.(±)-TTPG-B had a greater cytotoxic and apoptotic induction effect than (±)-KU.In addition, (±)-ARC induced cell cycle arrest at the S phase, whereas (±)-TTPG-B caused cell arrest at the G0/G1 phase, which is the same as (±)-KU in KKU-M213 cells [13].In contrast, in this study, half of the IC 50 values of (±)-TTPG-B, (±)-ARC, and (±)-KU caused cell cycle arrest at G0/G1 at 12 and 24 h.Therefore, we hypothesized that the specific target protein of these derivatives might be a protein associated with cell signaling and cell proliferation, which remains similar to (±)-KU.Here, we used the molecular docking technique to determine the binding affinity of these two compounds with twenty-two target proteins of the native (±)-KU.
The (±)-TTPG-B compound showed a lower affinity towards CSF1R than (+)-KU and (−)-KU.The CSF1R binding site is characterized by a narrow cleft [10].The presence of a long chain of butoxy groups (-OBu) introduces steric hindrance to the amino acid at the binding site, in contrast to the smaller methoxy groups (-OMe) found in (+)-KU and (−)-KU.The observation of this feature, resulting from the longer hydrocarbon chain, was made by comparing the poses of (±)-TTPG-B with the (+)-KU poses at the CSF1R pocket, as depicted in Figure 7A.For the (±)-ARC, while the docking suggested that the hydroxy group in both forms could do so, (−)-ARC was the only form resembling the (±)-KU, while the (+)-ARC was not able to penetrate the CSF1R binding site.Thus, we speculated that, in the case of (±)-ARC, the binding with CSF1R of (±)-ARC could be different due to their different stereo-configurations; Figure 7B.
The phenolic group present in (±)-TTPG-B and (±)-ARC may potentially contribute to a decrease in the binding affinity to the hydrophobic pocket of AKR1B1.This is because the primary interactions observed in (±)-KU were π-π stacking and the hydrophobic effect [12], whereas the phenolic group is considered a polar group.Therefore, there was a reduced affinity observed between (±)-TTPG-B and (±)-ARC with AKR1B1.
made by comparing the poses of (±)-TTPG-B with the (+)-KU poses at the CSF1R pocket, as depicted in Figure 7A.For the (±)-ARC, while the docking suggested that the hydroxy group in both forms could do so, (−)-ARC was the only form resembling the (±)-KU, while the (+)-ARC was not able to penetrate the CSF1R binding site.Thus, we speculated that, in the case of (±)-ARC, the binding with CSF1R of (±)-ARC could be different due to their different stereo-configurations; Figure 7B.The phenolic group present in (±)-TTPG-B and (±)-ARC may potentially contribute to a decrease in the binding affinity to the hydrophobic pocket of AKR1B1.This is because the primary interactions observed in (±)-KU were π-π stacking and the hydrophobic effect [12], whereas the phenolic group is considered a polar group.Therefore, there was a reduced affinity observed between (±)-TTPG-B and (±)-ARC with AKR1B1.
Next, the pose position analysis of (±)-TTPG-B and (±)-ARC.The structural information of the ATP binding pocket in the N-terminal domain of HSP90 demonstrated that Next, the pose position analysis of (±)-TTPG-B and (±)-ARC.The structural information of the ATP binding pocket in the N-terminal domain of HSP90 demonstrated that hydrophobic substituents at the 6 position could reach the side chains of Phe138, Try139, and Trp162 residues and form π-π interactions.(+)-TTPG-B bound stronger to HSP90α than EC44 (a known inhibitor), (±)-KU, and (±)-ARC.(+)-TTPG-B requires Tyr139 and Trp162 via π-π stacking on HSP90α.A new series of compounds bearing 2-thioquinazolinone scaffolds were designed and synthesized as the HSP90 inhibitors.Compound 8a showed a dual mode of interactions, basically by the same type of binding forces (arenehydrogen and hydrogen bonding), but this time with even more amino acid residues (four residues named Asn51, Tyr139, Val136, and Gly135), which could be in part responsible for its higher activity on HSP90α [16].The binding positions of (+)-TTPG-B and compound 8a shared the same key amino acid as Tyr139.Therefore, (+)-TTPG-B might have the capacity to bind to HSP90α.
The main forces between EVO and the residues were the Van der Waals forces, π-π stacking, and hydrophobic interactions.The docking results of evodiamine amino deriva-  Target proteins for CCA treatment have been reported, such as FGFR, IDH1, HER2, NTRK, PI3K, and MAPK [3,20].The drugs in a clinical study for targeted therapies with FDA approval, including pemigatinib, ivosidenib, zanidatamab, and futibatinib, which target FGFR, IDH1, HER2, and FGFR, respectively [3].The PI3K-AKT and RAS-MAPK pathways are promising pathways for targeted therapy for CCA due to this pathway being related to cell proliferation, angiogenesis, and survival.Tyrosine kinase proteins of both receptor and non-receptor types are also target proteins, including HER2, EGFR, VEGFR, PDGFR, and FGFR.These proteins are the upstream PI3K-AKT and RAS-MAPK pathways [19].HSP90 is an ATP-dependent molecular chaperon highly expressed in many cancer types [21].HSP90 controls many proteins in cancer which play an essential role in cell growth and survival; hence, the development of HSP90 inhibitors can be beneficial in cancers that have a high HSP90 expression (Figure 8) [22].Taken together, (±)-TTPG-B could be a drug target for CCA-targeted therapy.However, the anti-cancer activity and side effects of this compound require more in-depth studies in the animal model.

Molecular Docking
Twenty-two proteins associated with CCA cells and a potential target protein of native (±)-KU were selected for the investigation of potential protein targets of (±)-ARC and (±)-TTPG-B.The 3D protein structures were obtained from the Research Collaboratory for Structural Bioinformatics Protein Database Protein Data Bank (RCSB PDB).The AutoDock-Tool (ADT) version 4.1 was used to separate water molecules and all co-crystallized ligands.Every co-crystallized ligand was eliminated.All polar hydrogen atoms were included in the Protein Data Bank (PDB) file format in order to simulate hydrogen bond interactions.The RCSB PDB ascension codes of the 22 selected proteins are concluded in the Supplementary Information (Table S1).
The conformers, as the SDF format file, of native (+)-KU, native (−)-KU, (+)-ARC, (−)-ARC, (+)-TTPG-B, (−)-TTPG-B, and known ligands were obtained from the PubChem database.All structure files were then converted to the PDB file format using the Online SMILES Translator and Structure File Generator (https://cactus.nci.nih.gov/translate/,accessed on 21 January 2022).ADT was used to add all of the missing polar hydrogen atoms.Finally, both protein and ligand structures were saved in a PDB and Partial Charge (Q) and Atom Type (T) (PDBQT) file.
A molecular docking study between all compounds and the 22 selected proteins was evaluated using AutoDock4 version 4.2 [10].The grid was positioned at the center of the protein molecule with an x-y-z grid point of 126-126-126 cubic angstrom (Å 3 ).The Fifty Lamarckian Genetics Algorithm (GA), which runs with a population size of 200, was used in the study.All other parameters were placed at the default value on the AutoDock4 program.
The docking score was reported as the lowest predicted binding energy (∆G bind ) in kcal/mol.The docking outcomes were validated by performing a redocking experiment with the known inhibitor and its protein target.The criteria were based on the position of the docked pose and the site of its binding pocket as compared to the corresponding crystal structure.The docked position of the known inhibitor had to resemble the reported binding site when the co-crystallized inhibitor was absent.
The Visual Molecular Dynamics (VMD) package was used for all 3D structure visualizations.Biovia Discovery Studio (DS) was used to analyze the 2D interaction between a compound and a protein based on its PDB file structure.

Cell Cycle Analysis Assay
The assay was performed using the Muse ® kit (Merck Millipore, Darmstadt, Germany) according to the manufacturer's protocol.In brief, KKU-M213 cells were seeded in 24-well plates at a density of 1 × 10 5 cells/well.Cells were starved with low fetal bovine serum in DMEM for 24 h.Then, the medium was removed and replaced with a fresh culture medium without or with at 2.24, 0.035, and 0.005 µM of native (±)-KU, (±)-ARC, or (±)-TTPG-B and incubated for 12, 24, and 48 h.Cell pellets were resuspended in 1X PBS and the cell distribution was determined using propidium iodid (PI).The percentage of the cells in the G0/G1, S, and G2/M phases was analyzed using the MUSE ® Cell Analyzer (Merck Millipore, Darmstadt, Germany).

Western Blot Analysis
KKU-M213 cells were seeded in 6-well plates with 1 × 10 6 cells/well and treated with 0.15 µM (±)-KU, (±)-TTPG-B, or (±)-ARC for 24 and 48 h.After treatment, cells were subjected to Western blot analysis, as previously reported [8].Cells were harvested by trypsinization and lysated using a RIPA buffer (Thermo Scientific, Waltham, MA, USA).According to the manufacturer's instructions, the total protein concentration was measured using the Bradford method (Bio-Rad, Hercules, CA, USA).Eighty micrograms of each protein lysate were separated on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a nitrocellulose membrane (Millipore, Billerica, MA, USA).After that, the membranes were blocked with 5% non-fat dry milk in TBST (0.1% Tween 20, 154 mM NaCl, 48 mM Tris-base, pH 6.8) to prevent the nonspecific binding.The membranes were exposed to primary antibodies, including anti-PI3K, HSP90α, HSP90β, c-Myc, CyclinD1, MEK1, AKT, (Cell Signaling Technology, Danvers, MA, USA), CDK1, CyclinB1 (Santa Cruz Biotechnology, Dallas, TX, USA), and GAPDH antibodies (Calbiochem, Darmstadt, Germany).The protein signal was visualized using the SuperSignal TM West Dura Extended Duration substrate kit (Thermo Scientific, Waltham, MA, USA), according to the protocol supplied with the kit and visualized using a CCD camera.The band intensity was analyzed by Image J (NIH, Bethesda, MD, USA).

Statistical Analysis
Data values of 3 independent experiments were analyzed by Student's t-test on Microsoft Excel and were represented as the mean ± standard deviation (SD).A p-value of less than 0.05 was considered to indicate a statistically significant difference between groups.

Figure 4 .
Figure 4.The increasing of the cell cycle arrest by native (±)-KU and its derivatives.KKU-M213 were treated with (±)-KU, (±)-ARC, or (±)-TTPG-B at 2.24, 0.035, and 0.005 µM for (A,B) 12 h, (C 24 h, and (E,F) 48 h.Non-treated cells were exhibited as a negative control.Flow cytometry was u for the determination of cell cycle distribution after staining with PI.DNA histograms displayed G0/G1, S, and G2/M phases.The G0 phase is the resting phase during which the cell has stop dividing.From the end of the previous M phase to the beginning of DNA synthesis is the G1 ph All of the chromosomes have been replicated in the S phase.The G2 phase is a period of pro synthesis and rapid cell growth to prepare the cell for mitosis.The distributions of cells phases exhibited in terms of percentage of mean ± SD (n = 3).The asterisk (*) is represents p < 0.05 w

Figure 4 .
Figure 4.The increasing of the cell cycle arrest by native (±)-KU and its derivatives.KKU-M213 cells were treated with (±)-KU, (±)-ARC, or (±)-TTPG-B at 2.24, 0.035, and 0.005 µM for (A,B) 12 h, (C,D) 24 h, and (E,F) 48 h.Non-treated cells were exhibited as a negative control.Flow cytometry was used for the determination of cell cycle distribution after staining with PI.DNA histograms displayed the G0/G1, S, and G2/M phases.The G0 phase is the resting phase during which the cell has stopped dividing.From the end of the previous M phase to the beginning of DNA synthesis is the G1 phase.All of the chromosomes have been replicated in the S phase.The G2 phase is a period of protein synthesis and rapid cell growth to prepare the cell for mitosis.The distributions of cells phases are exhibited in terms of percentage of mean ± SD (n = 3).The asterisk (*) is represents p < 0.05 when compared with the non-treated cells (control).(±)-KU; trans-(±)-kusunokinin, (±)-ARC; (±)-arctigenin, (±)-TTPG-B; trans-(±)-TTPG-B.

Figure 8 .
Figure 8.The proposed pathway of (A) native trans-(−)-kusunokinin, (B) trans-(±)-TTPG-B, and (C) trans-(±)-arctigenin on CCA cells.The b to the inhibitory effect of the compounds.The numbers represent the binding energy in kcal/mol, and the yellow box refers to the protein docking analysis.The blue and red arrows mention the reduction of the protein levels in the KKU-M213 cells at 24 and 48 h, respectivel analysis.

Figure 8 .
Figure 8.The proposed pathway of (A) native trans-(−)-kusunokinin, (B) trans-(±)-TTPG-B, and (C) trans-(±)-arctigenin on CCA cells.The bold black arrows refer to the inhibitory effect of the compounds.The numbers represent the binding energy in kcal/mol, and the yellow box refers to the proteins from the molecular docking analysis.The blue and red arrows mention the reduction of the protein levels in the KKU-M213 cells at 24 and 48 h, respectively, using Western blot analysis.

Table 2 .
Interacting amino acids of HSP90α with the tested compounds.

Table 3 .
Interacting amino acids of PI3K with the tested compounds.