Extract of Ginkgo biloba exacerbates liver metastasis in a mouse colon cancer Xenograft model

Firefly luciferase was used to illuminate colon cancer cells used in the non-invasive bioluminescent imaging. The luciferase gene was excised from the pGL3 luciferase reporter vector (www.​promega.​com) and was inserted into the pLPCX vector that contains the LTR from Moloney murine leukemia virus. The luciferase-pLPCX plasmid was then transfected into a viral packaging cell line 293 T–e for packaging retroviral particles and ecotropic delivery using SuperFect transfection reagent (Qiagen.​com). Packaging cell line 293 T was kindly provided by Dr. Michael Kilberg at the University of Florida, Gainesville. SW620 cells (www.​atcc.​org; catalog # CCL-227) were infected with the luciferase-expressing retrovirus and the resulting SW620-luc cell line was established for stably expressing luciferase enzyme for in vivo luminescent imaging. Cell line verification was performed in the laboratory according to the guidelines by ATCC (Technical Bulletin No. 8, www.​atcc.​org). Cell line authentication has been performed according to the guidelines set by UKCCCR [34] and the latest authentication tests were performed by a commercial testing laboratory (Radil lab, Columbia MO).

Extract of Ginkgo biloba (EGb) used in the study was provided by Hebei Medical College (Hebei, China) using a procedure by Baweisong and associates (Chinese patent #00129807.0). Briefly, the crude extract of Ginkgo biloba (100 g) was dissolved in 666 mL methanol and sonicated for 15 min, followed by the addition of 1200 mL Ethyl and another sonication of 30 min. The mixture was ultra-filtered and dried using vacuum pump at 75 °C for 24 h. The purified product was dissolved in boiling ethanol water (17% w/w in water), decolorized with active carbon for 5 min, and filtered. The purified product was mixed with Na₂S₂O₅ solution and 2 g of Ginkgolide B was added to the mixture and dissolved under heat. The mixture was then concentrated by filtration. EGb concentrate was adjusted to pH 7 by NaHCO₃ before autoclave sterilization. Final product was packaged into 17.5 mg per ampule vial. Characterization of the purified EGb was conducted using HPLC. HPLC-grade acetonitrile was obtained from Fisher Scientific (Fisher Scientific, USA). The reference crude extract of Ginkgo biloba leaves (Lot No. 1160–200,001) and rutin reference standard were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Ginaton was purchased from Dr. Willmar Schwabe Pharmaceuticals (Karlsruhe, Germany). All other chemicals were analytical grade. The rutin reference standard solution was accurately weighed and dissolved in methanol then diluted to appropriate concentration ranges for the establishment of calibration curves. Two g of reference of Ginkgo biloba was dissolved in 50 mL ethanol, and ultrasonic extracted for 30 min. After the addition of 50 mL ethanol, the extract was filtered through a 0.45-μm PTFE filter membrane, as EGb reference sample solution. Analysis was done using a Diamonsil C18 column (5-μm particle size, 250 × 4.6 mm i.d.). Separation was carried out with gradient elution procedure. Linear ratio of mobile phase A (acetonitrile) and B (1% H3PO4) was set as follows: 0~30 min, 15~17% A; 30~37 min, 17~16% A; 37~45 min, 16~20% A; 45~60 min, 20~25% A; 60~80 min, 25~15% A. The total run time was 80 min at a flow rate of 1 mL/min. The eluent was monitored by a diode array detector, and the detection wavelength was set at 360 nm. The injection volume of sample was 20 μL, and the column temperature was 30 °C.

Twenty-four 4-wk.-old athymic mice were purchased from Charles River Laboratories (Charles River Laboratories, Wilmington, MA) and fed a modified AIN-76 diet (product #D12010101, Research Diets, Inc., New Brunswick, NJ). Animals were housed individually in standard polycarbonate cages in a humidity- and temperature- controlled environment on a 12-h light-dark cycle. After 2-wk. adaptation, animals were intra-splenically injected with 2.5 × 10⁶ of SW620-luc colon cancer cells. Tumor development and metastasis were monitored weekly by BLI of live animals. Briefly, BLI was conducted using a custom-made imaging system for live animals (Stanford Photonics, Palo Alto, CA), which includes a photon-tight dark box, bright field illumination, warming plate and a dual micro-channel plate ICCD camera (Mega 10-Z with cathode cooling). D-luciferin was freshly dissolved in PBS as a 150 mg/mL solution before administration. Injecting volume (μL) of D-luciferin was calculated as 10× body weight (g) and D-luciferin was injected intraperitoneally into each animal. Three min after the administration of D-luciferin, each mouse was anesthetized using isoflurane/oxygen gas from a vaporizer and placed face-up on a height-adjustable tray in the imaging system for whole body imaging. General anesthesia was maintained throughout the imaging. Photon emission was collected for a 3-min period using imaging software Piper Control (Stanford Photonics, Palo Alto, CA). At 10 wks post cancer cell injection, BLI imaging was used to determine tumor development and to assign animals into Tumor (animals with tumor development) and Normal (animals without detectable tumors) for subsequent EGb treatment.

A total of 12 animals in the Tumor group and 12 in the Normal group were randomized into PBS (n = 6) and EGb (n = 6), generating 4 groups of animals with Tumor/PBS, Tumor/EGb, Normal/PBS, and Normal/EGb treatments. Normal groups of animals were included for later assessment of the potential effects of EGb in normal liver without tumor development. Animals in the EGb groups were intraperitoneally injected with EGb solution daily for 5 wks with a dose of 35 mg/kg body weight (3.5 mg/mL EGb diluted in PBS). The dose of EGb was based on a previous publication by Rojas et al. [35]. Five wks later, all animals were euthanized by CO₂ asphyxiation before being decapitated for sample collection. A part of liver, spleen and other tissues were fixed in formalin, while the remainder of the liver was snap frozen in liquid nitrogen and stored at −80 °C for further analyses. The experimental design is outlined in Fig. 1a. All experiments were conducted in accordance to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the University of Illinois Institutional Animal Care and Use Committee.

Animals were euthanized by CO₂ asphyxiation before being decapitated for sample collection. Liver tissues were fixed in 10% buffered formalin for 24 h, then trimmed and embedded in paraffin. Four-μm-thick sections were stained with hematoxylin and eosin (H&E) and tumor number and size (in diameter) were analyzed by obtaining images using a NanoZoomer Slide Scanner and NDP View software (Hamamatsu, Bridgewater, NJ). Ten 20× light microscopic fields were examined on each section for each liver. Six representative sections from different locations in each liver were analyzed. The presence of micro-metastases at necropsy was evaluated by the veterinary pathologist Dr. Stephane Lezmi.

Total RNA was isolated from mouse liver using Tri-reagent (Sigma-Aldrich, St. Louis, MO) according to the manufacturer’s instructions. RNA concentration was measured by Bio-Rad spectrophotometry (Bio-Rad laboratories Inc., Irvine, CA). cDNA was synthesized from 2 μg total RNA using High Capacity cDNA Reverse transcription Kit (Applied Biosysterms, Foster City, CA) in a DNA 2720 Thermal Cycler (Applied Biosystems, Foster City, CA). The program was as follows: 25 °C for 10 min, 37 °C for 120 min, and 85 °C for 5 s. Twenty-five ng of cDNA was analyzed using quantitative real-time PCR and relative mRNA level was detected using the SYBR Green fast master mix (Quanta BioSciences, www.​vwr.​com) with a 7300 real-time PCR system (Applied Biosystems, Foster City, CA). The reaction was as follows: 40 cycles of 95 °C for 15 min, 95 °C for 15 s and 60 °C for 60 s. After amplification, dissociation curves were acquired by stepwise increases from 55 °C to 95 °C to ensure a specific product amplified in the reaction. Standard curves with slope of −3.30 ± 0.20 and R² ≥ 0.95 were accepted. The following housekeeping genes whose expression were not affected by the treatment were used to calculate the geometric mean to normalize raw PCR values: L7a, GAPDH, Tbp, Actb, 36B4 and HPRT1. The equation for calculation of geometric mean is \( {\left(\prod \limits_{i=1}^n{a}_i\right)}^{1/n}=\sqrt[n]{a_1{a}_2\cdots {a}_n} \)where a₁-a_n are PCR results for individual housekeeping genes. Primers were designed using Vector NT I software (Invitrogen Corporation, Carlsbad, CA) and synthesized by Integrated DNA Technologies (www.​idtdna.​com). Primer sequences used in the experiment are listed in Additional file 1: Table S1.

Frozen liver (200 mg) was ground in liquid nitrogen and suspended in 500 μl of protein lysis buffer (0.125 mol/L Tris–HCl, pH 6.8, 1% SDS, 0.04% bromophenol blue, and 20% glycerol, v/v) with 1× proteinase inhibitor (Roche applied sciences, Indianapolis, IN) and phosphatase inhibitor cocktail 1 and 2 (Sigma-Aldrich, St. Lois, MO). Samples were then sonicated with 25 pulses at power setting 2 (Fisher Scientific model 100 Sonic Dismembrator, Pittsburgh, PA). Twenty μg of diluted protein were size-fractionated on a 12% Tris–HCl polyacrylamide gel and transferred to a PVDF membrane (Bio-Rad, Hercules, CA). The membrane was then incubated with a blocking solution [Tris-buffered saline/Tween (TBS/T), 20 mmol/L Tris–HCl, pH 7.6, 137 mmol/L NaCl, and 0.1% v/v Tween-20] containing optimal amount of bovine serum albumin (BSA) or nonfat dry milk (NFDM) for 1 h at room temperature. Primary antibodies were diluted in the blocking solution according to the manufacturers’ instructions and incubated with the membrane for 3 h at room temperature. Antibodies used are listed in the Additional file 1: Table S2. Primary antibody solutions were prepared as follows: phospho-p38 MAPK (1:5000 in 5% BSA), p38 MAPK (1:1000 in 10% NFDM), phospho-p44/42 and p44/42 (1:1000 in 10% NFDM), and phospho-JNK and JNK (1:1000 in 5% BSA). Subsequently, the membranes were washed with TBS/T for 5 X 5 min. HRP-conjugated secondary antibody was diluted to 1:10,000 in a blocking solution containing 1% (w/v) NFDM and incubated for 1 h at room temperature. After 5 times of 5-min wash with TBS/T, membranes were incubated with enhanced chemiluminescence reagent SuperSignal West Dura (Pierce; Rockford, IL). Signal was detected and quantified using ChemiDoc XRS imaging system (Bio-Rad).

A standard three-step streptavidin–biotin complex immunohistochemical assay was performed on paraffin sections for the detection of Ki-67 and phosphor-Histone H3 (Ser 10). Sections were deparaffinized and rehydrated in gradient concentrations of xylene and ethanol and then in distilled water. Sections were treated with 3% H₂O₂ to quench endogenous peroxidase activity. Antigen retrieval was performed by microwave pretreatment in 0.01 M citrate buffer for 25 min. After cooling, non-specific binding of the antibody to sections was blocked with 10% rabbit serum, Ki-67 antibody (BD#550609, Pharmingen, San Francisco, CA, USA) was applied in a 1:3000 dilution and incubated at 24 °C for 2 h. For pH3S10, antibody was diluted to 1:500. A subsequent incubation of 30 mins with biotinylated anti-rabbit serum was followed by a 30-min incubation using the Vectastain Elite ABC anti-rabbit kit (Vector Laboratories). The peroxidase was then developed in DAB (Diaminobenzidine, Amresco Inc. OH, USA) with H₂O₂. The sections were counterstained with hemotoxylin, dehydrated in gradient ethanol solution, cleared in xylene and mounted. One known liver tumor sample was included for positive control and for negative control when the primary antibody was replaced with 10% rabbit serum (manufacturer’s instructions for Vectastain Elite ABC anti-rabbit kit).

Early cell apoptosis was detected using the ApopTag TM peroxidase kit (Oncor, Gaithersburg, MD) per the manufacturer’s instructions. Briefly, sections were first deparaffinized and rehydrated. The cell membranes were then digested with 20 μg/mL of proteinase K. The endogenous peroxidase activity was quenched with 3.0% hydrogen peroxide and sections were incubated with the equilibration buffer containing dUTP-digoxigenin. TdT enzyme was added and the reaction was stopped by dipping the slides in the stopping buffer. Then an anti-digoxigenin antibody conjugated with peroxidase was added, followed by 0.05% of the substrate, 3′ 3′-diaminobenzidine tetrahydrochloride (DAB). One sample without the addition of reaction mixture was used as a negative control and the same sample using DNase I to induce DNA strand breaks was used as a positive control according to the manufacturer’s instructions for ApopTag®Peroxodase In Situ Apoptosis Detection Kit.

Fifty mg of frozen liver was ground by a mortar and pestle in liquid nitrogen and homogenized in PBS with 5% FBS. The homogenized samples were centrifuged twice at 1.12 × 1000 g for 2 min to collect the supernatant and the resulting cell suspension was washed twice with PBS containing 5% FBS. Cells were counted using a hemocytometer and the concentration was adjusted to ~1 × 10⁶ cells/mL. The cells were pelleted by centrifugation, resuspended in PBS, fixed, and stored at −20 °C until future analysis. Propidium iodide was used to stain DNA and cell cycle analysis was performed using a BD Biosciences LSR II flow cytometer at the Flow Cytometry Facility (The W.M. Keck Center for Comparative and Functional Genomics, University of Illinois at Urbana-Champaign).

Fisher’s exact test was used to compare numbers of implantation and liver metastases between EGb group and PBS group (Table 1). Student t-tests were used to compare EGb and PBS groups for all remaining analyses. Each data point represents the mean ± SEM. Differences were considered significant at p < 0.05. The geometric mean was used to normalize data from RT-PCR for relative mRNA expression.

In order to determine whether EGb affects tumor development in mice, athymic mice were used to establish a human colon cancer xenograft model. Human colorectal cancer cells, SW620-luc cells were injected into the spleen to induce tumor development before EGb or PBS control was injected 10 weeks later. Tumor growth in PBS- and EGb-treated mice was monitored weekly by BLI (Fig. 1b). Live monitoring of luciferase activity and location of SW620-luc tumor cells showed that bioluminescent signals increased in both PBS- and EGb-treated mice over time (Fig. 1b). Moreover, the EGb group developed tumor metastasis from spleen to an upper portion of the mouse body over time (Fig. 1b). Necropsy at the end of the study showed that the implantation rate (the percentage of mice that developed tumors) in PBS and EGb groups was 58% and 71%, respectively (Table 1). Additionally, liver metastasis rate (the percentage of mice that developed liver tumors among mice with tumor implantations) was 38% and 50%, for PBS and EGb groups respectively (Table 1). EGb significantly increased the total number of tumors in the liver compared to PBS (Fig. 1c, p  < 0.05). Furthermore, size of spleen tumors was significantly larger in EGb mice compared to PBS (Table 1). Tumor sizes from all other tissues and organs were not statistically different between EGb and PBS groups (Table 1).

The gross structure of the liver with metastasis – including necrotic cells, mitotic cells and regular hepatocytes – was histologically examined using H&E stained liver sections (Fig. 1d). EGb mice had significantly smaller necrotic areas in the liver when compared to the PBS group (Fig. 1d, p  < 0.05). Furthermore, early apoptosis analysis using the TUNEL assay confirmed that EGb significantly decreased apoptosis in the metastatic liver when compared to the PBS group (Fig. 1e, p  < 0.05). In summary, EGb treatment significantly increased metastasis to liver from the spleen, and decreased necrosis and apoptosis in tumor when compared to the PBS control group in mouse liver.

To investigate the mechanisms of increased liver metastasis by EGb, cell mitosis and proliferation within tumor metastases were examined. Number of mitotic figures (cells undergoing mitosis within a given field) was evaluated under a light microscope with 20× magnification for H&E-stained tissue sections. Overall, there were significantly more mitotic figures in tumors from the EGb group than PBS group when numbers from all tissues were combined (Fig. 2a, p  < 0.05). Specifically, EGb significantly increased the number of mitotic figures in splenic tumors (black bars, Fig. 2a, p  < 0.05) but not in tumors in either liver or other tissues (gray-shaded bars for liver and open bars for other tissues, Fig. 2a). Effects on mitosis by EGb treatment were further investigated using an antibody against phospho-histone H3 Serine 10 (H3S10p), another mitotic marker in metastatic liver tumors. H3S10p was located in proliferating tumor cells that surround the necrotic areas (Fig. 2b), and EGb significantly increased H3S10p in the metastatic liver compared to PBS treatment (Fig. 2c, p  < 0.05).

Flow Cytometry was performed to examine cell cycle progression (G1, S, and G2/M) in liver of PBS and EGb-treated mice with tumors (Fig. 2d). EGb significantly decreased the percentage of cells in G1 phase when compared to PBS control. It had no effect on the percentage of cells in G2/M phase in liver with tumors (Fig. 2e, p  < 0.05).

Ki67, a proliferation marker, was analyzed to determine the growth of tumor in the metastatic liver. Ki67-positive cells were primarily located in the proliferating areas of metastatic tumor surrounding the necrotic areas (Fig. 2f). Moreover, EGb significantly increased the percentage of Ki67-positive cells in liver tumors when compared to tumors from PBS controls (Fig. 2g, p  < 0.05). Overall, EGb significantly increased proliferation of metastatic tumor cells in mouse liver when compared to PBS treatment.

In order to determine the potential mechanism of the increased metastasis from the primary splenic site to the liver by EGb, liver tissues were collected from both Tumor and Normal groups to investigate various pathways that could affect hepatic microenvironment and impact tumor physiology. EGb-treated mice had significantly increased mRNA expression of Ndrg1 and p21, two metastasis-related genes, when compared to the PBS control, but only in animals with liver tumors (Tumor), and not in the Normal, non-tumor groups (Fig. 3a, p  < 0.05).

mRNA expression of apoptosis-related genes Timp1, Caspase-9, Bad and Apaf1 were also assessed. EGb-treated mice had significantly decreased mRNA expression of these genes when comparing to PBS controls but only in animals with liver tumors, and not in the Normal group (Fig. 3b, p  < 0.05).

EGb treatment exerted differential patterns of mRNA expression of oxidative stress-related genes. Specifically, EGb significantly decreased gene expression of Nrf2, Cat and Nox3, in the liver of mice with tumors (Tumor, Fig. 3c, p  < 0.05). On the other hand, EGb significantly increased Cox2 mRNA expression in the Tumor group when compared to the PBS control group. Interestingly, in the Normal group, EGb significantly decreased mRNA expression of Cox2 and Nox3 while Cat expression was upregulated in the liver of EGb-treated mice without tumors (Normal, Fig. 3c).

EGb significantly increased mRNA expression of genes related to mitochondrial functions including Mitofusin-2 (Mfn2) and dynamin related protein-1 (Drp1) in Normal group when compared to the PBS control group (p < 0.05), but there were no significant changes observed in the Tumor group.

Western blot was performed to investigate the activation of the MAPK/JNK signaling pathway, a major signaling pathway in tumor development in mouse liver with metastasis (Fig. 4a). MAP kinases in this pathway include extracellular signal-regulated kinases, c-Jun N-terminal or stress-activated protein kinases (JNK), ERK/big MAP kinase 1 (p44/42), and the p38 group of protein kinases (p38). EGb significantly increased hepatic protein level of phosphorylated p38 (p-p38), phosphorylated p44/42 (p-p42/44) and phosphorylated JNK (p-JNK) in the Tumor group when compared to the PBS control (Fig. 4b–p  < 0.05). In the Normal group, EGb ablated p-p38 and increased protein level of p-JNK (Fig. 4b, d). In conclusion, EGb activated MAP kinase cascades in the Tumor group, and exerted different effects in mice without tumor development (Normal group).