A hepatoprotective Lindera obtusiloba extract suppresses growth and attenuates insulin like growth factor-1 receptor signaling and NF-kappaB activity in human liver cancer cell lines

Tissue culture plates and polystyrene microtiter for ELISA as well as for fluorimetric analysis were from Nunc (Roskilde, Denmark) and Dynex (Chantilly, VA), respectively. If not stated otherwise, all reagents were purchased from Merck (Darmstadt, Germany) or Sigma-Aldrich (Deisenhofen, Germany) and were of the highest purity available. Cell culture media and solutions were purchased from Invitrogen (Karlsruhe, Germany) or Biochrom (Berlin, Germany).

Freeze-dried extracts of L.obtusiloba were obtained as described previously [31]. To obtain stock solutions, 10 mg powder was redissolved in 10 ml sterile phosphate-buffered saline (PBS) at 60°C for 30 min. Aliquots were stored at -20°C. Freshly prepared working solutions of L.obtusiloba extract were routinely standardized according to their anti-fibrotic and anti-inflammatory activity as previously described [31, 32]. Briefly, 100 μg/ml L.obtusiloba extract had to reduce proliferation of 3T3-L1 preadipoctyes by 45% and to suppress the autocrine stimulation of TGF-β expression of hepatic stellate cells by 50% before to be used in the assays with HCC cells.

The human HCC cell lines HepG2 (ATCC HV-8062), Hep3B (ATCC HV-8064), Huh-7 (JCRB 0403; Tokio, Japan) and SK-Hep1 (ATCC HTB-52) cells (Fuchs et al., 2008) were cultured in a humidified atmosphere at 37°C and 5% CO₂. Standard culture medium consisted of DMEM with 862 mg/l L-alanyl-L-glutamine, 4.5 g/l glucose, 50 μg/ml streptomycin, 50 units/ml penicillin, 50 μg/ml L-ascorbic acid, supplemented with 10% heat-inactivated fetal bovine serum (FBS). Cell layers were detached with 0.05% trypsin/0.02% EDTA solution. Cell morphology in culture was directly examined by inverse phase contrast microscopy (Zeiss, Oberkochen, Germany).

HCC cells (5 × 10³) were seeded into 96-well tissue culture plates in 100 μl standard culture medium. After 24 h, cells were cell cycle synchronized in 100 μl culture medium containing 0.2% FBS for additional 24 h. Cultures were treated with up to 200 μg/ml L.obtusiloba extract as indicated for 20 h. Proliferation was determined by adding 0.5 μCi/well [³H]-thymidine (GE Healthcare, Munich, Germany) for 4 h. Cells were fixed with 10% trichloro acetic acid and the DNA was solubilized with 200 mM NaOH, neutralized with an equal volume of 800 mM HCl and transferred to glass filter pads. Radioactive decay was monitored by liquid β-scintillation counting within 1 min (LKB Wallac Turku, Finland).

50 μl of 3 mg/ml Matrigel™ (BD Biosciences, Heidelberg, Germany) diluted in ice cold, serum free DMEM were used to coat the upper compartments of 24-well transwell inserts (BD Biosciences; pore size 8 μm) for 16 h at 37°C. 2 × 10⁵ cells diluted in 300 μl serum free medium were seeded into the upper compartments and L.obtusiloba extract was added at a final concentration of 100 μg/ml. DMEM containing 10% FBS as stimulating agent was added to the lower compartment and the plates were incubated for up to 24 h at 37°C in a humidified atmosphere with 5% CO₂. Cells that remained in the upper compartment were gently removed with a cotton swab. The inserts were then washed with PBS and invaded cells on the lower surface of the insert were fixed for 20 min with 2% glutaraldehyde in PBS and stained using 0.1% crystal violet in water. The stained cells on each insert were visualized by light microscopy and manually counted in three independent spots per insert.

Apoptosis was quantified fluorimetrically from caspase-3/7 activity. In brief, 2 × 10⁵ HCC cells in standard culture medium were seeded into 24-well tissue culture plates. Confluent cell layers were thoroughly washed with DMEM and subsequently incubated with culture medium containing 0.2% FBS for 24 h. Cells were then treated for another 24 h in the presence of 100 μg/ml L.obtusiloba extract or 100 nM staurosporine and 0.2% FBS. Apoptosis was determined using the SensoLyte™ Homogenous AFC Caspase-3/7 Assay Kit (AnaSpec, San Jose, CA) according to the manufactures instructions. Briefly, cells were lysed in 200 μl lysis buffer for 1 h at 4°C. The clear supernatant obtained after centrifugation at 2,500 × g for 30 min was stored at -80°C until measurement. Caspase 3/7-mediated conversion of the substrate N-acetyl-Asp-Glu-Val-Asp-7 amino-4 trifluoromethyl coumarin was monitored fluorometrically using a Spectramax Gemini EM microplate reader (λex: 380 nm, λem: 500 nm; Molecular Devices, Sunnyvale, CA).

HCC cells cultured in 6-well tissue culture plates with 125 ng/ml human recombinant IGF-1 (Biomol, Hamburg, Germany), 100 μg/ml L.obtusiloba extract and a combination of both for 48 h were rinsed with ice-cold PBS and lysed with a lysis-buffer containing 50 mM Tris-HCl pH 7.4, 2.25 M urea, 1.4% sodium dodecyl sulfate, 100 mM dithiothreitol, 2 mM NaVO₃, 5 mM NaF, and per 10 ml buffer one tablet of Complete Mini Protease Inhibitor cocktail (Roche, Penzberg, Germany). Aliquots of 333 μl lysate were transferred to 0.5 ml reaction tubes and frozen at 80°C. Protein content was determined using the Nano Orange Protein Assay Kit (Molecular Devices) according to the manufactures instructions. From each cell lysate, 25 μg protein per lane were separated by SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, Munich, Germany) using a tank blot apparatus (Hoefer, Holliston, MA). Membranes blocked for 1 h with 5% skim milk powder in a buffer containing 10 mM Tris, 154 mM NaCl, 0.1% Tween 20 were incubated over night at 4°C with the following specific primary antibodies with the dilution given: Akt (1:1,250), COX-2 (1:1,000), Erk1/2 (1:1,000), iNOS (1:1,500), pAkt (1:1,250), pErk1/2 (1:1,250), Stat3 (1:1,250; Cell Signaling, Beverly, MA), β-Actin (1:10,000), HIF-1α (1:2,000; Novus Biologicals, Littleton, CO, USA), IGF-1R (1:1,250), pIGF-1R (1:1,250; Imgenex, San Diego, CA) and PPARγ (1:2,000), pStat3 (1:1,250), VEGF (1:800; Santa Cruz, Santa Cruz, CA). After washing, membranes were incubated for 1 h with rabbit or mouse immunoglobulin G-specific horseradish peroxidase-labeled secondary antibodies (1:2,500; Dako, Hamburg, Germany). Bands were detected by enhanced chemiluminescence (GE Healthcare) using the Luminescent Image Analyser LAS-4000 (Fujifilm, Düsseldorf, Germany). Band intensities were quantified using Image J and normalized to the β-actin loading control.

Transfection of the cells was performed using the electroporation method and a NF-κB-luciferase reporter plasmid as described by Stroh et al. [33, 34]. Detached cells (2 × 10⁵) were resuspended in 100 μl electroporation buffer containing 90 mM phosphate buffer pH 7.2, 10 mM MgCl₂, and 50 mM glucose before 4 μg of the NF-κB-luciferase reporter plasmid pNF-κB-TA-Luc (Clontech, Mountain View, CA) were added. In an electroporation cuvette with a gap of 2 mm (Biozym, Hessisch Oldendorf, Germany), cells were subjected to single square pulses of 400 V for 400 μs (HepG2, Hep3B and Huh-7) or 600 V for 400 μs (SK-Hep1), allowed to rest for 1 min, and transferred into prewarmed standard culture medium. A total of 1 × 10⁵ transfected cells in 1 ml culture medium were seeded into a 24-well plate. Cell viability as determined by Calcein AM staining [32] was about 85% in conjunction with a cell transfection efficacy of ~75%.

Twenty hours after transfection with the NF-κB-luciferase reporter plasmid [33] cells were treated with 10 μg/ml recombinant human TNFα (Peprotech, Hamburg, Germany), 100 μg/ml L.obtusiloba extract, a combination of both and 15 nM of the NF-κB inhibitor 17-Dimethylamino-ethylamino-17-demethoxygeldanamycin (17-DMAG, InvivoGen, San Diego, CA). Cells were incubated for 24 h, washed twice with PBS, and lysed in 80 μl of reporter lysis buffer (Promega, Mannheim, Germany). Protein concentrations were determined using the Nano Orange Protein Assay Kit. Samples (20 μl) were transferred into a white 96 well plate before 60 μl of luciferase substrate were added and mixed for 5 s. Luciferase activity was measured for 0.5 s using a Mithras LB 940 luminescence reader (Berthold Technologies, Bad Wildbad, Germany). NF-κB activity was estimated as relative luminescence units (RLU) corresponding to equal protein amounts.

One way ANOVA/Tukey Tests were performed using SigmaStat for Windows (version 2.03; Systat, San Jose, CA). P < 0.05 was considered significantly different.

Effects of L.obtusiloba extract on the proliferation of human HCC cells were tested in cell-cycle synchronized cell lines. To define effective dose ranges, HCC cells in culture were treated with up to 200 μg/ml L.obtusiloba extract (Figure 1A). The range of concentration of L.obtusiloba extract and the experimental protocols were adapted from preceding studies dealing with the extract [31, 32].

L.obtusiloba extract reduced the proliferation of all four human HCC cell lines in a dose-dependent manner. The IC₅₀ values for the inhibition of the de novo DNA synthesis were approximately 100 μg/ml L.obtusiloba extract for all HCC cell lines. This concentration was used in all subsequent experiments. Induction of apoptosis due to exposure of cells with L.obtusiloba extract was determined by the enzymatic activity of pro-apoptotic caspase-3/-7 (Figure 1B). As shown for the apoptosis inducer and kinase inhibitor staurosporine used as control, all cell lines were highly susceptible to induction of apoptosis by L.obtusiloba extract as shown by 2.2- to 20-fold enhanced caspase activity. In the differentiated HCC cell lines HepG2, Hep3B and Huh-7, this effect of L.obtusiloba extract did not exceed 60% of the effect of 100 nM staurosporine. In contrast, L.obtusiloba extract provoked a caspase activity that corresponded to ~80% of apoptosis induced by staurosporine in the poorly differentiated SK-Hep1 cells (P < 0.001). Since their migratory potential mainly defines their aggressiveness, 100 mg/ml L.obtusiloba extract was applied to HCC cells in matrigel invasion assays. Again, while L.obtusiloba extract only slightly attenuated the invasion of HepG2, Huh-7 (P < 0.05) and Hep3B cells through a reconstituted basement membrane, it led to a stronger reduction of invasion in SK-Hep1 cells by 55% (P < 0.01) (Figure 1C). As for direct effects of L.obtusiloba extract on tumor cells, it diminished the invasive potential of HCC cell lines and was most effective on cells displaying a highly aggressive phenotype.

HCC represents a highly vascularized tumor entity and the tumor cells contribute to that process by production of proteins regulating angiogenesis. Thus, we next investigated whether L.obtusiloba extract impacts the expression of VEGF and HIF-1α in HCC cell lines. Linking Huh-7 to SK-Hep1 cells, stimulation with exogenous IGF-1 enhanced basal expression of VEGF by 1.4- or 3.3-fold, while in HepG2 and Hep3B no effects of IGF-1 were observed (Table 1). L.obtusiloba extract alone reduced VEGF expression in all four cell lines but strongest in Huh-7 cells. In combination with IGF-1, L.obtusiloba extract did not affect the IGF-1-induced VEGF expression in HepG2 cells, but in Hep3B, Huh-7 and SK-Hep1. The IGF-1-induced enhancement of HIF 1α expression was most prominent in differentiated HepG2 cells (3.6-fold) and intermediate in Hep3B (1.5-fold) and SK-Hep1 cells (1.3-fold). In Huh-7 cells no significant IGF-1-mediated effects on HIF 1α expression were observed. Similar to VEGF, L.obtusiloba extract distinctly reduced basal and IGF-1-induced HIF-1α expression in each of the HCC cell lines to comparable individual levels that were independent of the presence of IGF-1. These findings on VEGF and HIF-1α pointed to a strong anti-angiogenic potential of L.obtusiloba extract. Consequently, we studied the impact of L.obtusiloba extract on the expression of other proteins crucial in neo-angiogenesis.

The expression of the nuclear transcription factor PPARγ and its target genes COX-2 and iNOS are implicated in hepatocarcinogenesis and in the formation of enhanced microvessel density in HCC tissues. Effects of L.obtusiloba extract on the expression of PPARγ, COX-2 and iNOS were examined at protein level (Table 2). The expression of PPARγ in all four HCC cell lines was enhanced after stimulation with IGF-1. L.obtusiloba extract reduced both, basal and IGF-1-induced PPARγ expression with the same pattern as HIF-1α (Table 1). COX-2 was not detected in HepG2 and Huh-7 cells (Table 2). On the other hand, Hep3B and SK-Hep1 showed a high IGF-1-induced expression of COX-2 by 2.3- and 3.2-fold, respectively and with L.obtusiloba extract a reduction of both, the basal and the IGF-1-induced COX-2 expression. Hep3B and Huh 7 cells showed no expression of iNOS. In HepG2 and SK-Hep1 cells the basal expression of iNOS was enhanced by IGF-1 by 1.2- and 1.9-fold, respectively. L.obtusiloba extract reduced the basal and the IGF-1-induced iNOS expression of both cell lines by ~80%.

The IGF-1/IGF-1R axis plays an important role in angiogenesis and therefore the development of HCC. To investigate signaling pathways involved, western-blots specific for (p)IGF-1R and the activation states of its target proteins were focused on Hep3B as one out of the three less invasive HCC cells (compare Figure 1C) and the more aggressive SK-Hep1 cells. In both cell lines 100 μg/ml exogenous IGF-1 increased the phosphorylation state of the IGF-1R (Figure 2, Table 3). This IGF-1-mediated activation of the IGF-1R was strongly reduced in the presence of L.obtusiloba extract; by half in the Hep3B and to about a quarter in SK-Hep1 cells. As for the downstream signaling molecules Akt, Stat3 and Erk, L.obtusiloba extract did not alter basal phosphorylation. IGF-1 induced phosphorylation of Akt, Stat3 and Erk were tested in both cell lines. Increased pAkt levels that were at least partially abrogated by L.obtusiloba extract were found to be the most prominent effect. Treatment with L.obtusiloba extract in combination with IGF-1 markedly decreased the levels of pAkt, pStat3 and pErk in Hep3B and SK-Hep1 cells.

NF-κB is a key regulator of crucial pro-inflammatory cytokines during carcinogenesis and promotes cell survival and angiogenesis. Since L.obtusiloba extract induces apoptosis (Figure 1B) and displays anti-inflammatory activity [32], we assessed whether the extract decreases the activity of NF-κB in HCC cells (Figure 3). All four HCC cell lines transfected for transient constitutive expression of NF-κB exhibited high levels of basal NF-κB transcriptional activity of about 160 260 RLU. This activity was not significantly increased by addition of TNFα. In all cell lines, treatment of transfected cells with the specific NF-κB-inhibitor 17-DMAG reduced the activity to <10% of the basal level thus approving the function of the experimental system (data not shown). Except for HepG2 cells, L.obtusiloba extract attenuated the transcriptional activity of NF-κB to 75% (P < 0.05) of the basal level in Huh-7 and to ~65% (P < 0.001) in Hep3B cells while in the poorly differentiated SK-Hep1 cells the high basal transcriptional activity of NF-κB was reduced to 50% (P < 0.001). These results at the level of regulation clearly strengthen our conclusion that L.obtusiloba extract directly impairs the survival and the angiogenic program in HCC cells.