Animal models
All animal experiments were performed under the guideline, approved by Institutional Animal Care and Use Committee (IACUC) of the Ulsan National Institute of Science and Technology (UNSIT). For liver pathology study, CCl4 was freshly dissolved in olive oil in one to three ratio and 6-week-old mice was administered with CCl4 mixture (0.5 μl/g) through oral gavage or intraperitoneal (i.p.) injection every 5 days up to 6 months. DEN-induced HCC mice model was established by single intraperitoneal injection of DEN (25 mg/kg) to infants at postnatal 2 weeks. Littermate male mice were used as control for all experiments. For BDL, mice were euthanized by isoflurane gas and their abdomen was incised to reveal common bile duct. After common bile duct was ligated, the abdominal cavity was closed by separate suture of peritoneum and skins and restoration of surgical wounds were found in 3 to 5 days. The mice model was dissected 10 days and their liver tissues were harvested for histological assessment
For enhancer-deficient Eo771 tumor allografts, Eo771 cells were cultured in HG medium for one week.
The cells were then implanted orthotopically, to the inguinal mammary fat pad of male wildtype C57BL/6J mice (0.8 × 106 cells per mouse). Tumors were harvested two weeks after implantation. For the diabetic breast cancer mouse model, STZ was freshly prepared in 0.1M citrate buffer (pH 4.5) and C57BL/6J wildtype female mice (8 weeks) were given either vehicle (citrate buffer) or STZ (50 mg/kg) via intraperitoneal injections for five consecutive days. Fasting blood glucose levels were assessed three days after the final STZ/vehicle treatment, and Eo771 cells (0.8 × 106 cells per mouse) were orthotopically injected to the inguinal mammary fat pad. Tumor growth was monitored, and tumors were harvested once tumor volumes reached approximately 800 mm3. For drug combination treatment, Eo771 cells (0.8 × 106 cells per mouse) were implanted orthotopically, to inguinal mammary fat pads.
Three days after cancer cell implantation, a single shot of either vehicle or STZ (110 mg/kg) was administered to the tumor-bearing mice. Fasting blood glucose levels were measured three days after STZ/vehicle administration, and the mice were injected with either vehicle or a lapatinib combined with DAPT for six consecutive days. Lapatinib was freshly prepared in 30% polyethylene glycol 400 (PEG400), 10% DMSO, and 10% Tween 80 and administered orally (100 mg/kg), while DAPT was freshly prepared in corn oil (90%) and administered intraperitoneally (25 mg/kg). Tumors were harvested two days after treatment.
15
For PyMT/FP635 breast cancer model, female mice (8 weeks) were given intraperitoneal injections of cisplatin, which was sonicated for 5 min before use. Mice were received cisplatin (2.5 mg/kg) in the first week of administration and then lower dose of cisplatin (1.5 mg/kg) for the remaining treatment.
Either NC13 or Vms (20 mg/kg) were mixed in a powdered chow and they were fed to mice during the administration. Breast cancer growth was monitored in a every week basis, by the in vivo fluorescence imaging system (Bruker, Germany). Cisplatin administration was terminated after 10 weeks of initial treatment, and tissue samples were harvested from the lung and tumors. For syngeneic tumor implant model, wildtype BALB/C mice (9 week) were given powdered mix of chow and the NC13 or Vms compounds (20mg/kg) until sacrifice.
One week of pretreatments of NC13 or Vms were given to mice, and 4T1 cells (0.5 × 105/100uL) were orthotopically injected to mammary fat pad. When the tumor got to around 100 mm3, intraperitoneally injection of cisplatin (2.5 mg/kg) was given to the mice 2 times in a week. Digital caliper was used to measure tumor growth, and animals were dissected after 6 week of cisplatin treatment.
Plant material and isolation of bioactive components
Korea Research Institute of Bioscience and Biotechnology (KRIBB, Cheong-ju si, Chungcheongbuk- do, Korea) provided the natural compounds used in this study, where the sample voucher specimen (KRIB-0020697) was deposited. The isolation process of Vms was documented in previous reports (18).
Briefly, three times of methanol (MeOH) extraction was performed in dried leaves and stems of the NC13 (2.0 kg) at 25℃ to acquire solid extract (198.7 g). Preparative reverse phase chromatography (GRACE C18, 10 µm, Grace Davison Discovery Sciences, CA, USA) was performed with the MeOH extract (1.0 g), which was isocratically eluted with methanol (25%). The eluted samples were then fractionated into Frs. 1 to 6, and they were further concentrated by using rotary evaporator in a pressure reduced condition. Isolation of fraction #4 was carried out using RPC-18 column (Zeoprep C18, 10 μm, 20 × 250 mm, Louisville, USA) in a middle performance chromatography and then elution was carried out with a gradient aqueous methanol solution (20 to 100%) to be sub-fractionated into F#4 a to e.
NC106 (10.7 mg) was obtained from the F#4b subfractions by using semiparative HPLC (Atlantis T3, 5 μm, 19 × 250 mm, Waters, USA, 18% acetonitrile aqueous solution). Vms (23.5 mg) and NC105 (12.4 mg) were obtained from the F#4c subfractions by using semipreparative HPLC (Synergy Polar-RP 4 μm, 21.2 × 250 mm, Phenomenex, CA, USA, 22% acetonitrile aqueous solution). To obtain NC107 (6.1 mg), separation of the F#4d was carried out using Sephadex LH-20 (Pharmacia Biotech AB, Sweden, 90% methanol aqueous solution)
16 Analysis of Human data
TCGA database was used to obtain gene expression profile of HCC cohorts, through the GDC (https://gdc-portal.nci.nih.gov) for several analysis (TCGA-LIHC). In the TCGA-LIHC projects, the upper quantile of standardized FPKM (FPKM-UQ) was extracted from gene expression profile (RNA- sequencing) and clinical information of the HCC patients were matched accordingly. R studio (https://www.rstudio.com) was used to analysis and GraphPad Prism (GraphPad Software) was used to visualize the results. Expression levels of COL6A3 were analyzed in both tumor and matched non-tumor liver tissues from HCC subjects (n = 50). The FPKM-UQ values of non-tumor liver tissues were designated as reference to normalize that of tumor tissue. R-derived clinical information and RNA-seq data were used to generate a Kaplan-Meier survival analysis and log-rank P-value for COL6A3 levels in HCC cohort. HCC patients were stratified by COL6A3 expression level with the lowest 60% quantile of FPKM-UQ as low group and the highest 40% quantile as high group. Similarly, Non-tumor cohorts were stratified with the lowest 23% as low group and highest 77% as high group.
For clinical data analysis in breast cancer cohort, survival annotations and RNA sequencing data of 1,077 breast cancer patients were downloaded from TCGA. To analyze gene expression profiles, FPKM counts were normalized through Log2 transformation. A hallmark Notch signaling gene set (M5903 from GSEA) was used to define the NOTCH score that was used to categorize TCGA cohort based on NOTCH activity. For each patient, the score was determined using the singscore method implemented in the R software singscore_1.8.0 (R version 4.0.2). The scores were then stratified based on the median value. To analyze open-chromatin accessibility in the Nrg1 enhancer region, a subset of patients with ATAC-Seq data (n=74) identified in previous reports [14] were stratified based on median NOTCH activity (n=37 per group). To analyze survival, breast cancer patients were initially separated by low- or high-NRG1 groups on the basis of NRG1 levels and then on HER2-status in TCGA. Individuals were classified into either HER2-positive (n=161) or HER2-negative (n=916) groups. Kaplan‒Meier curves were generated for both HER2-positive and HER2-negative groups using survival parameters (PFI and DFI) obtained from previously published data [15]. Statistical significance was calculated using Gehan- Breslow-Wilcoxon test implemented in GraphPad Prism 7.
Cell culture
For liver pathology experiments, primary NPCs and hepatocytes were isolated from liver tissues from wildtype C57BL/6 mouse, according to a two-step collagenase perfusion protocol. Liver tissues were harvested and cut into small fragments to be filtered with cell strainers (100 μm). This crude suspension of cells contained NPCs and hepatocytes. The hepatocyte cell fraction was separated by low-speed
17
centrifugation (25 × g for 5 min) and washed with M199/EBSS medium (Hyclone, SH30253.01, Logan, UT, USA) for three times.The remaining supernatant from the hepatocyte pellet was further centrifuged (300 × g for 5 minutes) to isolate the NPC pellets. The NPC fraction was washed with M199/EBSS medium for three times. The isolated cells were subsequently plated on a cell culture dish and incubated overnight before treatment.
The murine breast cancer cell line Met1 was isolated from MMTV-PyMT mice, while Eo771 and 4T1 cancer cell lines were acquired from the ATCC. Human breast cancer cell lines SK-BR-3, BT-474, T47- D, MDA-MB-231, ZR75-30, BT-20, and MCF7 were also acquired from ATCC. Cell culture medium was supplemented with penicillin/streptomycin (1%) and FBS (10%), and the cells were maintained in a 5% CO2 humidified incubator. To establish a cell line-based model system for studying hyperglycemia-driven breast cancer cells, a low glucose (1 g/L) adaptation period was applied. All the breast cancer cells were maintained in low glucose DMEM for at least 3 days and then transferred to high glucose (4.5 g/L) DMEM until the indicated time points.
Wound healing migration assay
In 96-well ImageLock plates (Essen BioScience, USA), MDA-MB-231 cells were plated and incubated until reaching full confluency (95 to 100%). To make a scratch in each well, Wound-Maker (Essen BioScience) was used and subsequent cell debris were washed with two times of PBS. After wash, cells were treated with serum free culture media together with compounds or cisplatin. Cellular migration was recorded every 2 h, together with cell images using IncuCyte Imaging System (Essen BioScience).
Dosage information of natural products
Multiple concentrations of NC13 or Vms were tested for in vitro treatments, and subsequently, 5 μg/mL (NC13) and 10 μM (5.244 μg/mL, Vms) were chosen for the further experiments, on the basis of CompuSyn dose analysis. For the in vivo administration of the compounds, 20 mg/kg was given to mice model based on the previous study (17). During the experimental periods, mice were given to powdered form of NC13 or Vms (20mg/kg), combined with powdered chow, which was corresponding to 97.3 mg in a day for human with 60kg weight (19). In clinical trial in USA, human subjects are dosed with NC13 drugs ranging from 80 to 160mg, and thus 20 mg/kg of administrations are accepted as physiologically appropriate.
18 Plasmid and siRNA transfection
Plasmid vectors and small interfering RNAs (siRNAs) were transiently transfected in the target cells by treating jet-OPTIMUS transfection reagent (Polyplus, #117-01) and G-Fectin (Genolution, Seoul, Korea), respectively, following the manufacturer’s protocols. Scrambled siRNAs were obtained from Genolution (Seoul, Korea). The siRNAs used for the study are listed in Table 1.
Chemical reagents
TSA and PUGNAc (Sigma, USA) were dissolved in DMSO to be stored as stock solutions (500 µM and 50 mM, respectively) and then diluted to desired concentrations for in vitro assays. For osmotic control experiments, mannitol (Sigma, USA), and D-glucose (Amresco, USA) were dissolved in PBS (1 M) before use. The gamma-secretase inhibitor, DAPT (SelleckChem, USA), and the selective RTK inhibitor, lapatinib (TCI, Tokyo, Japan), were dissolved in DMSO to be stored as 100 mM stock solutions and subsequently diluted to desired concentration for in vivo assays. STZ (Sigma, USA) was freshly prepared in 0.1M of citrate buffer (pH 4.5) just before injecting the animals. The primary antibodies used were against: H3K4me1 (Abcam, ab8895), H3K27ac (Abcam, ab4729), NRG1 (Abcam, ab27303), NOTCH1 (Abcam, ab27526), RBPJ (Abcam, ab25949), O-GlcNAc (Abcam, ab2735), P300 (Abcam, ab275378), CBP (CST, #7379), HDAC1 (CST, #34589), SETD1A (CST, #50805), GAPDH (SCBT, sc-32233), Lamin A/C (SCBT, sc-376249), and β-Actin (SCBT, sc-47778).
shRNA knockdown cell line
P300, Cbp, Setd1A, and Nrg1-targeting oligomers were annealed and cloned into pLKO.1 TRC vectors.
H293T cells were transfected with VSVG, Δ8.9, and pLKO.1 vectors, and lentivirus-containing supernatant was collected twice and then filtered with 0.45μm filters. Cancer cells were treated with lentiviral media and polybrene, and infected cells were selected by puromycin treatment. The sequences of hairpins in our study are listed in Table 2.
Immunohistochemistry
Excised tissues were fixed in a neutral buffered formalin (10%) for 2days. After sequential steps of dehydration followed by paraffin embedding, each liver tissue specimen was stained with H&E or Masson's trichrome C stain. Citrate buffer (pH6.0) was used to retrieve antigen and then tissue slides
19
were treated with hydrogen peroxide and BSA blocking buffer. The sections were incubated with primary antibodies and subsequent secondary antibodies. Images were taken by using Olympus FSX100 inverted microscope.
Immunofluorescence staining
Primary NPCs and hepatocytes were seeded on coverslips and incubated overnight. The hepatocytes were treated with recombinant ETP, 1.2 mM CCl4, and 10 µM sp600125 and their conditioned medium were collected. The hepatocyte conditioned medium was treated to NPCs. After indicated treatments, the cells were washed with two times of PBS and fixed in a PFA (4%) for 30 min at 25℃. The fixed cells were subsequently blocked by NH4Cl and permeabilized with Triton X-100 (0.1%) for 5 minutes, respectively. Primary antibodies were incubated overnight at 4℃, and secondary antibodies labeled with Alexa-Fluor 488 dye or Cy3 were incubated at 25℃ for 1 hour. DAPI was co-stained for nuclei detection. Apoptotic cells were detected by using TUNEL assay kit (Roche, #11684795910, St. Louis, MO, USA), following the manufacturer’s protocols. Slide images were taken using a confocal microscope (Olympus, Fv-1000).
Met1 and 4T1 cancer cells were plated on coverslips overnight. After HG treatment, cells were washed with two time of PBS and then fixed in a PFA (4%) in PBS for 30 min. After washing, fixed cells were blocked with NH4Cl for 5 minutes, incubated with primary antibodies against NOTCH1 and O-GlcNAc overnight at 4 °C, and then with secondary antibodies labeled with Alexa488 or 594 for an hour at 25°C.
Nucleus was stained by DAPI, and images were acquired using a confocal microscope (Olympus Fv1000).
Immunoblotting
After indicated treatment, cells were washed and lysed in NETN buffer (1 mM EDTA, 100 mM NaCl, 20 mM Tris-Cl, pH 7.5, 1% NP-40) then subjected to centrifugation. The supernatants were collected as total cell extracts. Sample quantification was performed by a BCA assay (Pierce, USA), and subsequently, 30μg of protein samples were loaded on 10% or 12% polyacrylamide gels to be resolved by SDS-PAGE and subsequently transferred to NC membranes (Amersham, UK). After blocking for 30minutes, the membranes were incubated with primary antibodies at 4℃ overnight and then secondary antibodies conjugated with infrared fluorescent dye at 25℃ for 1 hour. Odyssey CLx scanner (Li-COR Biosciences, NE, USA) was used to scan the membranes.
20 RNA isolation and qPCR
Total RNA was isolated from tissue samples or cultured cells, by using TRIzol reagent (Invitrogen, CA, USA) following the manufacturer’s protocols. MMLV-RT (Invitrogen, #28025-013) and Oligo(dT)12-18
(Invitrogen, #18418-012), were used to synthesize cDNA from 2 μg of total RNA. qPCR analysis was performed using SYBR reagent (GenetBio, Korea) in a QuantStuido5 qPCR equipment (Applied Biosystems,CA, USA).
ELISA
Serum α-fetoprotein levels in mouse were measured by using an AFP ELISA kit (R&D Systems, MN, USA), according to the manufacturer’s instructions.
ChIP and Sequential ChIP
Chromatin was sheared using a truChIP® Chromatin Shearing Kit (PN520154, Covaris, MA, USA) following the manufacturer’s instructions, but with subtle modifications. Briefly, cells were fixed with 1% formaldehyde for 8 minutes at 25℃ and then 0.125 M of glycine was treated for quenching. Cells used for histone modification (H3K4me, H3K27ac) were not fixed. The cells were then washed, resuspended in lysis buffer, and incubated on ice for 15 minutes. Nuclei were collected and chromatin shearing performed using the Covaris S220 sonicator to obtain DNA fragments 200–500 bp in size.
Sheared chromatin samples were diluted in ChIP dilution buffer (1.2 mM EDTA, 167 mM NaCl, 0.01%
SDS, 1.1% Triton-X 100, and 16.7 mM Tris-HCl, pH 8.1) and then incubated with protein A/G beads (Santa Cruz, sc-2003), rabbit serum, and salmon sperm DNA at 4 °C for 1 hour. Precleared samples were incubated at 4 °C overnight with specific antibodies and then precipitated with protein A/G bead at 4 °C for 1 h. The immunoprecipitated DNA complex was sequentially washed with a low salt concentration buffer (10 mM NaCl, 20 mM Tris-HCl pH 8.1, 1% Triton-X 100, 2 mM EDTA, and 0.1%
SDS), high salt concentration buffer (500 mM NaCl, 20 mM Tris-HCl pH 8.1, 1% Triton-X 100, 2 mM EDTA, and 0.1% SDS), LiCl buffer (10 mM Tris-HCl, pH 8.1, 1 mM EDTA, 1% deoxycholic acid, 1%
NP-40, and 0.25 M LiCl), and twice with TE buffer (1 mM EDTA and 10 mM Tris-HCl, pH 8.0). Sample elution was carried out using ChIP elution buffer (1% SDS and 100 mM NaHCO3) at 25 °C for 30 min.
For sequential ChIP assay, samples were eluted using re-ChIP elution buffer (15 mM DTT, 2% SDS, and 10 mM Tris pH 8.0) for 30 minutes at 37 °C, and subsequently diluted one twentieth in ChIP dilution buffer. Second antibody incubation, precipitation, sequential washes and elution steps were repeated.
21
Eluted samples were incubated with proteinase K, and RNase A, and then incubated at 65 °C overnight for reverse cross-linking. DNA samples were purified using a column-based kit (DN10200, Bionics, Seoul, Korea) and analyzed using qPCR. Primers used for ChIP-qPCR are listed in Table 3.
Chromatin accessibility assay
A chromatin accessibility kit was acquired from Abcam (ab185901), and the assay was carried out according to the manufacturer's instructions. Briefly, chromatin samples were isolated from LG or HG treated cells (2 x 106) and digested with nuclease mix. DNA samples were purified and analyzed by qPCR. Primer sequences used for analyzing Nrg1 enhancer and promoter are listed in Table 4.
Viability assay
Breast cancer cells were plated overnight in 96-well culture dish (2000 cells/well) and then treated with either siRNA or DAPT. The cells were then incubated in Incucyte Zoom system (Essen bioscience) for two days. Cellular proliferation was monitored by taking images every two hours. Confluency was measured using Incucyte software (Sartorius, Göttingen, Germany).
To test effects of the natural compounds in cell viability, MTT assay was carried out to assess viability of cancer cells. Cells (8 × 103 cells/well) were plated in 96-well culture dish, and the next day, cells were treated with either NC13 or Vms for 44 hours, followed by treatment of MTT reagent for another 4 hours. After aspirating media, 100 μL of DMSO was added to each well to dissolve formazan, and optical density was measured at 595 nm by using a microplate reading equipment (Tecan, Infinite- M1000).
Analysis of cytoplasmic and nuclear fractions
Nuclear and cytoplasmic extracts were prepared using a two-step hypotonic and high salt concentration buffer protocol. Harvested cells were washed with two time of cold PBS and incubated at 4℃ with ice- chilled hypotonic buffer (1 mM EDTA, 1 mM EGTA and 20 mM HEPES pH 7.9) for 20 minutes. The crude extracts were centrifuged (14,000 x g for 5 minutes, 4 °C), and cytoplasmic extracts were collected in the supernatant. The remaining nuclear pellets were re-suspended in a high salt concentration buffer (20% glycerol, 1 mM EGTA, 1 mM EDTA, 20 mM HEPES pH 7.9 and 420 mM
22
NaCl) and homogenized in an ice-cold dounce homogenizer. After 20 min on ice, the extracts were centrifuged (15,000 x g for 30 minutes, 4 °C), and nuclear extracts harvested in the supernatant.
Luciferase assay
The Nrg1 enhancer or promoter region was inserted into a pGL4.23 reporter vector (Promega, Madison, WI, USA) while Rbpj cDNA was inserted into pCMV6–Myc/Flag (Origene). Met1 cancer cells were subjected to transient transfection with control reporter or the Nrg1 enhancer reporter together with RBPJ expressing vectors. In addition, the control reporter, Nrg1 enhancer, Nrg1 promoter, or both reporter vectors were transfected into Met1 or 4T1 breast cancer cells, which were then treated with LG or HG. A Renilla luciferase construct was used to normalize transfection controls. Cell lysates were harvested, and subsequently their luciferase reporter activity was determined with the luciferase assay system (Promega), following the manufacturer’s protocol.
CRISPR/Cas9 based Nrg1 enhancer editing
CRISPR/Cas9 genome editing was used to delete parts of Nrg1 enhancer elements in Eo771 breast cancer cells using the lentivirus delivery system. Single guide RNAs (sgRNAs) targeting Rbpj, Ap1, and Gata4-binding motifs, as well as the whole Nrg1 enhancer region were designed on the basis of online software tools (https://www.benchling.com/) (listed in Table 5). sgRNA pairs were annealed and inserted into the LentiCRISPRv2 plasmid vector. To produce lentiviruses, LentiCRISPRv2 constructs were co-transfected with Vsvg and Δ8.9 enveloping/packaging vectors into H293T cells, and the conditioned medium was collected. Eo771 breast cancer cells were transduced with the conditioned medium containing lentiviral particles targeting each binding motif and the enhancer region, as well as a non-target sequence in the wildtype control. Transduced cells were selected using puromycin and single colony isolation. Standard PCR and commercial DNA sequencing were performed to validate target sequence deletions. We selected clones with the intended modifications in their target sequences:
wildtype (Non-target) control, ΔE-250, ΔE-Rbpj, ΔE-Ap1, and ΔE-Gata4.
Bisulfite sequencing
Genomic DNA (gDNA) was extracted from Met1, 4T1, and Eo771 cells grown under LG and HG conditions. Bisulfite conversions of gDNA were performed using the EpiTect bisulfite kit (Qiagen,
#59104), following the manufacturer’s instructions. Design of primer pairs for methylated sequences