Most of the methods used here have been published previously but are repeated here for clarity (Chiu et al., 2019; Lee et al., 2013; Liu et al., 2015; Liu et al., 2010; Liu et al., 2012; Pan et al., 2013a; Pan et al., 2013b; Wu et al., 2016b). Honokiol was obtained from the Wako Chemical Company (Osaka, Japan).
Cell culture
Cell culture systems were performed as previously described (Wu et al., 2016b). The cell bank of Taipei Veterans General Hospital (Taiwan) supplied the human gastric cancer cell lines, AGS (moderately differentiated gastric adenocarcinoma) and MKN 45 or SCM-1 cells (poorly differentiated gastric adenocarcinoma). Cells were grown in RPMI medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin (complete medium) at 37°C in humidified incubator with 5% CO2. During experiments, cells were plated in six-well plates cultured with serum-free medium (starved medium) overnight, and then treated with drugs (honokiol or melatonin) with various concentrations for various time intervals. In some experiments, transfection of cancer cells was performed using a Lipofectin reagent (Invitrogen) according to the manufacturer’s instructions. The efficiency of transfection (~ 90%) was determined using an equal amount of a plasmid that encoded the green fluorescent protein under the cytomegalovirus promoter.
Tissue samples and clinical data collection
This study analyzed 40 patients with gastric cancer. The institutional review board of Taichung Veterans General Hospital approved the study protocol (Approval No CE 17096B). Data obtained from the registry of the Cancer Institute of Tissue Bank included age, sex, differentiation, tumor location, tumor size, depth of tumor invasion (T stage) and number of metastatic lymph nodes (N stage) (both based on the seventh edition UICC TNM classification for gastric cancer), extent of lymph node metastasis, and Lauren classification. Correlations between HDAC3 expression and the clinico-pathologic characteristics of patients were summarized. Fisher’s exact test was used for the histologic analysis of tumor stage, tumor grade, and distant metastasis, while one-way analysis of variance (ANOVA) was used to compare clinico-pathologic characteristics. Statistical significance was set at p < 0.05.
Ligand Binding assays
Study were conducted as previously described in 10 mM 2-(N-morpholino)ethanesulfonic acid (MES) buffer containing10 mM MnCl2, 1 mM EDTA, adjusted to pH 6.0 with KOH (Davis and Sharif, 2000). Initial tissue linearity studies were carried out using 0.1 ± 200 µg of cell containing the recombinant human HDAC3 per 0.5 ml total volume in 96-well deep well assay blocks (Matrix Technologies Corp., Hudson, NH, U.S.A.), using 300 pM [3H]-Honokiol (PerkinElmer Inc.). Protein were thawed quickly, diluted to the desired concentration in the binding buffer, and mixed to a homogeneous suspension prior to dispensation. After addition of the radioligand, the assay mixtures were incubated for 120 min on a rotary shaker (50 r.p.m.). The study was terminated by rapid vacuum filtration on Whatman GF/B glass filter mats (previously soaked in 0.5% polyetheleneimine) using cold MES buffer. Bound radioligand was then quantitated by liquid scintillation counting. (Packard BioScience).
Differential scanning calorimetry (DSC)
The methodology and rationale for DSC analysis (PerkinElmer DSC 7) are summarized as described previously and references contained within(Martin et al., 2013). Although some HDAC3 molecules may have nucleotide bound at the end of the purification, this will be released from the protein before the protein unfolding.
Immuno-histochemistry
In the human gastric cancer tissues, 5 µm thick sections were cut from 10% formalin-fixed samples and stained for specific antibodies using anti-primary antibodies. Immuno-histochemistry (IHC) analysis was performed as described previously (Liu et al., 2015). The antibodies used were listed and described in supplementary Tables 1 and 2.
HDAC3 activity
A HDAC Activity Assay Kit was used to examine HDAC activity, following the manufacturer’s protocol (Abnova Corporation; Catalog Number KA3731). The assay kit provided a positive control (a HeLa nuclear extract), a deacetylated HDAC assay standard, and a control inhibitor (trichostatin A; TSA), as well as a colorimetric HDAC substrate [R-H-K-K(Ac)-AFC] to release the AFC molecule, which could be fluorometrically detected (Ex/Em = 380/500 nm).
Luciferase reporter assay
Genetic reporters are used as indicators gene expression studies with accompanying cellular events (Pan et al., 2013a). Cells at 60% confluence were co-transfected with 0.2 µg of the promoter reporter construct CCAAT/enhancer-binding protein B (C/EBPβ), NFκB activator, and 0.05 µg of a thymidine kinase promoter driven Renilla luciferase vector (pRLTK; Promega, Mannheim, Germany). The pRL-tk-LUC vector coding for a Renilla luciferase under control of a constitutively active thymidine kinase promoter was co-transfected to correct for transfection efficiency. After incubation, the cells were lysed and processed using the Dual Luciferase Kit (Promega) as described by the manufacturer. Luciferase activity was normalized against the Renilla firefly activity for transfection efficiency and recorded by a luminometer (LKB, Rockville, MD, USA). Experiments were performed in triplicate, unless stated otherwise.
Endoplasmic reticulum stress response element (ERSE)
Endoplasmic reticulum stress measurements were performed using the Cignal ERSE Reporter Luciferase assay kit (SA Biosciences, QIAGEN, Frederick, MD, USA). The ERSE reporter assay was a mixture of an ERSE-responsive luciferase construct and a constitutively expressed Renilla luciferase construct. The diluted transfection provided ready reporter, negative control, positive control formulations, and relevant nucleic acid tests. Overnight post-transfection, the transfected cells were treated with Honokiol and ER stress activator. Activities of the signaling pathways were investigated using dual luciferase assay.
Endoplasmic reticulum stress response element (ERSE)
TEM was performed as previously described (Pan et al., 2013b). Cells were treated with or without Honokiol or shHDAC3 for 18 hours and then harvested and fixed with 4% glutaraldehyde and 2.5% paraformaldehyde dissolved in 0.1 M sodium cacodylate. The cells were then post-fixed in 1% osmium tetraoxide, dehydrated in ethanol, and embedded in araldite. Sections on grids were counter-stained with uranyl acetate and lead citrate, and examined in a JEM 1200 EX TEM (JEOL, Peabody, MA, USA) at an accelerating voltage of 80 kV.
Transfection
Cancer cells AGS, SCM1, and MKN45 were transfected with shRNA (National RNAi Core Facility Platform, Taipei, Taiwan) or over-expressed plasmid 1 µg/ml pcDNA (Genome Research Center, National Yang-Ming University) using a Lipofectin reagent (Invitrogen), in accordance with the manufacturer’s instructions.
Immuno-blotting
The preparation of whole-cell lysates of gastric cancer cells for immuno-blotting was performed as previously described (e.g. Liu et al., 2015; Pan et al., 2013a). The antibodies used in the present study were also listed in supplementary Tables 1 and 2. Detection was performed by ECL (Amersham) and by chemiluminescence using Kodak X-Omat film. The immunoblot assay was independently repeated five times.
Wound-healing cell migration assay and Matrigel invasion assay
After transfection, the cells were seeded at 3.5 × 105 cells/well in 12-well plates. At 100% confluence, the cells were scraped by a sterile tip of a 1000 µl pipette to generate an artificial “wound”. Two parallel wounds were created using a plastic pipette tip. The cells were further grown in a culture medium with 2% FBS. The images were collected at 0 and 18 h using a microscope. Migration capacity was quantified by measuring the change of “wound” width. This assay was independently repeated five times.
For the Matrigel invasion assay, 3 × 105 cells/well were seeded in the upper chamber that was coated with Matrigel (BD Bioscience, San Diego, CA, USA). After 48 h at 37°C and 5% CO2, the cells present on the lower surface of the insert were stained with Diff-Quik stain (Biochemical Sciences, Inc., Swedesboro, NJ, USA). Cells that invaded through the Matrigel-coated membrane were microscopically enumerated. Cell staining from three randomly selected fields were photographed using a CKX41 inverted microscope (Olympus Corp). The mean value was recorded. All experiments were performed in triplicate; each being repeated at least five times. And this variability of the mean values is represented by the SD.
Experimental animals
All of animal design and models are used have approval and all procedures has been conform to the guidelines from the NIH Guide for the Care and Use of Laboratory Animals. All animal studies were approved by the appropriate institutional ethical committee of the Taichung Veterans General Hospital Taiwan (Approval No La-1061488). For euthanasia, mice during continuous deep were anaesthetized with isoflurane inhalation (induction: 3%, maintenance: 1–2%) in medical air (0.4 L/min). The animals were sacrificed by cervical dislocation under isoflurane.
Xenograft tumor mouse model and positron emission tomography-computed tomography (PET/CT)
Cell culture systems were used as described previously (Chiu et al., 2019). Imaging studies were performed using positron emission tomography-computed tomography (PET/CT) also as previously described. To evaluate for peritoneal metastasis, MKN45 cells were inoculated into the peritoneal cavity of BALB/c nude mice. Peritoneal tumors in the nude mice were established by PET/CT surveillance 5–7 days after the injection. The mice were given intra-peritoneal injections of Honokiol (5 mg/kg/twice per week) for 28 days.
Changes in Honokiol on the peritoneal dissemination were evaluated by PET/CT. The mice were sacrificed under anesthesia and examined macroscopically for the presence of peritoneal metastasis. The tumors were excised, cut into blocks, fixed in 10% formalin, and embedded in paraffin blocks or snap-frozen in liquid nitrogen. All images were analyzed and interpreted by senior nuclear medicine physicians with all available clinical information. Correlative conventional imaging was used for anatomic guidance.
TOP/FOP luciferase reporter assay
To assess the transcriptional activity of β-catenin in gastric cancer cells, the TOP/FOP reporter system using the dualluciferase kit (Dual-GloTM Luciferase Assay System, Promega, Madison, WI, USA) were used. Cells were transiently transfected with 1 µg of constitutively active vector encoding thymidine kinase promoter-Renilla luciferase reporter plasmid (pRL-TK) (Promega) and β-catenin responsive firefly luciferase reporter plasmid TopFlash (Millipore, Billerica, MA, USA), or the negative control FopFlash (Millipore) using lipofectin. Cells were harvested after 24 hrs in culture and both firefly and Renilla luciferase activities were measured in duplicate/triplicate according to the manufacturer’s instructions. The firefly luciferase activity was normalized against the Renilla luciferase activity and the fold increase in TOPFlash activity compared to FOPFlash was reported.
To assess the function of Honokiol on β-catenin transcriptional activity, cells were co-transfected with TopFlash or FopFlash and small hairpin RNA (shRNA) using lipofectin. The cells were harvested after 36 hrs for luciferase measurements. Luminescence was read using the Sirius luminometer (Berthold Detection System, Pforzheim, Germany).
Statistical analyses
Values were presented as mean ± SD. Analysis of variance (ANOVA), followed by Fisher’s least significant difference test, was performed for all data. Statistical significance was set at p < 0.05.