A purified, biologically active, soluble, recombinant human furin fragment [from Dr I. Lindberg] prepared was used for in vitro enzyme assays as previously described (10). The polyphenols were obtained from Sigma-Aldrich and a 10 mM stock solution for use in different experiments was made in ethanol, DMSO according to the instructions by the supplier. As fluorogenic peptide substrate Pyr-RTKR-AMC (Pyr, pyroglutamic acid; AMC, 7-amino-4-methylcoumarin) was used (Bachem). Uncleaved pro-diphtheria toxin was obtained from Merck and the proprotein convertase inhibitor nona-D-arginine (D9R) from Pepscan Presto, Lelystad, The Nederlands.

Cell lines including CHO (ATCC, CCL-61), RPE.40 (12) and HepG2 (ATCC, HB-8065) were grown in DMEM/F12 (1:1) (Invitrogen™) supplemented with 10% fetal calf serum (HyClone) and cultured at 37°C in a humidified 5% CO₂ atmosphere.

To test the effect of polyphenols on the processing of proprotein substrates by endogenous PCs in cells, full length cDNA encoding pro-IGF-1R, or HA-tagged pro-GPC3, which constitute established PC substrates, was transfected into CHO cells or cells of the furin-deficient cell line RPE.40. Full length human IGF-IR cDNA (hIGF-IR) [kind gift of Dr R. Kooijman, VUB, Brussels, Belgium; (13)], and full length human GPC3 cDNA [kind gift of Dr G. David; (14)] were used in the pcDNA3 expression vector. In similar co-transfection experiments, the effect of the compounds on processing of the afore-mentioned proproteins by exogenously expressed human furin, was studied in RPE.40 cells. Furthermore, to test the effect of these compounds on the maturation of PCs, expression constructs encoding wild-type human furin [the 4.1 EcoR1 FUR cDNA (4)] in pcDNA3 vector was transfected into CHO cells. Transfections were performed using Fugene™ 6 Transfection Reagent (Roche) according to the manufacturer’s protocol. About 24 h after transfection, the PC zymogens, the proprotein substrates and their corresponding processing products were studied by western blot analysis using specific antibodies.

A possible impact of polyphenols on the cleavage of the fluorogenic Pyr-RTKR-AMC peptide substrate by purified hfurin⁵⁷³in vitro was determined as follows. Experiments were performed in 96-well microtiter plates containing each a 100 μl reaction mixture. First, 2 μl of hfurin⁵⁷³ solution (10 ng/μl) was added to 83.8 μl of buffer solution [0.5% Triton X-100, 0.5 mM CaCl₂ and 100 mM HEPES (pH 7.0)]. Subsequently, one of the polyphenolic compounds (10 μl) was added at a final concentration of 10 or 20 μM and pre-incubation was performed at room temperature for 30 min. Then 4.2 μl of the fluorogenic peptide substrate Pyr-RTKR-AMC (final concentration 100 μM) was added to each well and incubation at 37°C was continued. The various solvents used were brought in HEPES buffer and have been used as corresponding vehicle controls. The FLUOstar Galaxy fluorometer (BMG Lab Technologies) was used to measure the fluorescence intensity at an emission wavelength of 460 nm at 37°C at 5-min intervals for 1 h. An excitation wavelength of 390 nm was used. The fluorometric output, which was obtained as relative fluorescence units (RFU), was plotted against time and further processed. All fluorescence data are reported as the means of the measurements of three independent experiments, with error bars showing the standard deviations. Reaction rates were measured as RFU/min. The final reaction rate was calculated by subtracting the RFU/min of the vehicle controls as background. The experiments in which only hfurin⁵⁷³ and fluorogenic substrate were mixed in the absence of inhibitor were used as the positive controls. The percentage inhibition was calculated according to the formula { [ ( RFU / min )   positive control - ( RFU / min ) inhibitor ] / ( RFU / min )   positive control } × 100 %.

Purified hfurin⁵⁷³ was preincubated with 5 μM D9R or different concentrations of polyphenols (10 μM or 20 μM) in HEPES I buffer (100 mM HEPES, 0.5% Triton X-100, 0.5 mM CaCl₂) at room temperature for 30 min. After the preincubation, 3 μg of the uncleaved diphtheria toxin (final concentration 0.3 μg/μl) was added and incubated at 37°C for 1 h. The final concentration of hfurin⁵⁷³ was 0.03 ng/μl. Processing of the uncleaved diphtheria toxin into the A- and B-fragments was evaluated by SDS polyacrylamide gel electrophoresis analysis and proteins were visualized by Coomassie Blue staining (Merck Chemicals).

The procedure for measuring the endogenous PC activity was as follows: CHO cells were grown to confluence and treated in each experiment for 24 h with different concentrations of different compounds, as indicated in the figures. Cells were washed twice with PBS buffer and then lysed in 100 μl TNT buffer [10 mM Tris-HCl (pH 7.2), 0.15 M NaCl, 0.1% Triton X-100, and protease inhibitor mix (1.0 μM E-64, 0.1 μM leupeptin, 1.0 μM pepstatin, 0.01 mM TPCK)]. The cell lysate was centrifuged in a minifuge at 13,000 rpm for 1 min. The supernatant was transferred to another tube and the protein concentration was measured using a BCA (bicinchonic acid) protein assay kit (Thermo Scientific) according to the manufacturer’s protocol. Each sample was adjusted by TNT buffer to achieve the same protein concentration. In order to measure the PC activity, 25 μl of cell lysate was added to each well of 96-well microtiter plates, then 20 μl HEPES II buffer [250 mM HEPES (pH 7.0), 0.25% Triton X-100, 12.5 mM CaCl₂, and the above-mentioned protease inhibitor mix] was added before the addition of the fluorogenic peptide substrate Pyr-RTKR-AMC (200 μM final concentration) and incubation at 37°C was continued. As controls 25 μl cell lysate, for each polyphenol concentration, added to 25 μl HEPES II buffer (pH 7.0) were used. Fluorescence was measured at 37°C at 5-min intervals for 1 h. An excitation wavelength of 390 nm and an emission wavelength of 460 nm were used. The relative PC activity, after incubation for 1 h, was plotted against the concentration of the polyphenols. Determination of PC processing activity in the medium was performed similarly.

Cells were lysed about 24 h after transfection in SDS-PAGE sample buffer (50 mM Tris-HCl, pH 6.8, 10% glycerol, 2% SDS and 4% β-mercaptoethanol) and after boiling the solution for 10 min, samples were loaded on a 10% SDS-polyacrylamide gel. After separation on the basis of molecular weight, proteins were transferred electrophoretically onto Protran^® nitrocellulose transfer membranes (Whatmann GmbH) and subjected to western blot analysis. Expression of the IGF-IR-related proteins was evaluated using a rabbit anti-hIGF-1Rα-subunit antibody (SC-712; Santa Cruz). Expression of the HA-tagged GPC3 protein was evaluated with a mouse anti-HA tag monoclonal antibody (monoclonal antibody 16B12, Covance). Expression of the pro- and mature forms of human furin was detected using the monoclonal antibody MON-148 (Enzo Life Sciences), which specifically recognizes the subtilisin-like catalytic domain of furin (15). The identity of pro-furin was confirmed with a home-made rabbit polyclonal anti-serum raised against the pro-domain of human pro-furin (designated hProfurinAb). As a control, the expression level of actin was evaluated using an anti-β-actin monoclonal antibody from Sigma-Aldrich.

To investigate the possible influence of the polyphenols on pro-furin maturation, CHO cells were transfected with an expression construct encoding full length human pro-furin. Twenty-four hours after transfection with the pro-furin encoding DNA, various concentrations of different polyphenols were added to the culture medium. After 12-h incubation, the culture medium was removed and cells were washed twice with DMEM/F12 medium and cultured overnight in serum-free medium containing the different compounds. Subsequently, culture medium was precipitated using 4 volumes of methanol and the presence of FUR gene-encoded proteins in the medium was visualised by SDS-PAGE and western blot analysis. The cells were lysed in 200 μl of TNT buffer, as described above. The lysate was used to determine the FUR gene-encoded proteins in the cell lysate by SDS-PAGE and western blot analysis using the antibodies mentioned above.

Cell viability upon treatment with different polyphenols was studied using the liver carcinoma cell line HepG2. Cells were detached by trypsinization, seeded at 5,000 cells/48-well microtiter plate and cultured for 24 h. Then they were treated with various concentrations of different polyphenols, and incubated for an additional 48 h. The effect of different polyphenols on cell growth was examined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell viability assay. Briefly, 100 μl of MTT solution (5 mg/ml; Sigma-Aldrich) was added to each well and incubated for 4 h at 37°C. The supernatant was aspirated, and the MTT formazan that was formed by metabolically viable cells was dissolved in 150 μl of DMSO, and then monitored by a Dynatech MR5000 microplate reader (Dynatech Corp., Burlington, MA) at a wavelength of 590 nm. The data are expressed as the percentage of viable cells as compared to the number of cells in the control (taken as 100% viability), as determined by the MTT reduction test.

In vitro fluorescence substrate-based screening assays are commonly used for determining enzyme activity. When the substrate encounters the active enzyme, substrate cleavage will occur and a fluorescence signal will emerge. We have tested 30 compounds. Twenty-six of these are polyphenols, 1 compound is a monophenol (salicylic acid) while 3 compounds are non-polyphenolic small molecules (caffeine, theophylline and piperine). The genuine furin inhibitor D9R was used as a control. These compounds were first tested using the in vitro fluorescence peptide substrate-based screening assay and the in vitro assay measuring the processing of uncleaved diphtheria toxin protein.

Fig. 1A shows that the majority of the polyphenols leads to a strong to moderate inhibition of the enzyme activity. Some polyphenols, with ellagic acid as the clearest example, do not provoke inhibition. The genuine PC inhibitor D9R shows a strong inhibitory effect. On the other hand, inhibition by the mono-phenolic and the non-polyphenolic control compounds is weak if at all present.

In light of the previously observed false-positive inhibitory effects of polyphenols on hfurin⁵⁷³-mediated cleavage of the fluorogenic peptide substrate in vitro(10), studies were extended to the investigation of the effect of the selected compounds on the in vitro processing of a genuine furin proprotein substrate, e.g. uncleaved diphtheria toxin. In contrast to the results obtained with the fluorogenic peptide substrate, no inhibitory effect for any of these 30 compounds was observed on the in vitro processing of the diphtheria toxin proprotein (Fig. 1B). In fact the presence of protein is sufficient to almost completely abolish the inhibitory effects of polyphenols in the observed in vitro fluorescence substrate-based screening assay (data not shown). On the contrary, the genuine PC inhibitor D9R strongly inhibits processing also in this assay.

Our results clearly shed doubt on the capacity of the polyphenols tested to inhibit PC activity in vivo, in the presence of abundant other cellular proteins, by direct interaction with the enzyme. On the other hand it cannot be excluded that polyphenols could inhibit PC activity as a result of indirect mechanisms. In order to address this question a cellular test system was used.

The small molecules tested were first incubated with CHO cells for 24 h. After the incubation, the cells were washed twice with PBS and immediately lysed in TN buffer supplemented with Triton X-100. After lysing the cells, the fluorescent AMC substrate was added and the activity of endogenous PCs was measured. The results indicate that some of the compounds show inhibitory effect in the cell-based assay in which cellular proteins are abundantly present. While a number of compounds show some dose-dependent effect on PC activity (Fig. 1C), only myricitin, octyl gallate, dodecyl gallate, ellagic acid and nordihydroguariaretic acid showed moderate to strong inhibition. Octyl gallate, dodecyl gallate and nordihydroguariaretic acid were used for further studies. It should be noted that D9R has virtually no effect on PC activity in the cellular test system which measures indirect effects on PC activity. This implies that despite the reported capacity of D9R to penetrate cells (16), this is insufficient to reduce PC activity significantly.

In order to investigate whether octyl gallate, dodecyl gallate and nordihydroguariaretic acid also inhibit endogenous PC activity in CHO cells when other than the fluorescent AMC substrate is used, two genuine PC substrates were tested. The proprotein substrates studied included pro-IGF-1R and pro-GPC3. They were expressed in CHO cell cultures after transfection.

Treatment of the transfected CHO cells with various concentrations of these three compounds revealed that the ratio between the individual proproteins and their corresponding mature processing products increased at increasing compound concentration (Fig. 2A and B). These results show an inhibitory effect of the three compounds on the processing capabilities of endogenous proprotein convertase in CHO cells with respect to the three proproteins investigated. The observation that octyl gallate, dodecyl gallate and nordihydroguariaretic acid inhibit endogenous PC-mediated processing of various proproteins raised the question as to whether all PCs are inhibited by them.

Processing of pro-GPC3 was studied in co-expression experiments using RPE.40 cells. These cells lack furin processing activity as already reported (12). In the presence of various concentrations of octyl gallate, dodecyl gallate and nordihydroguariaretic acid, the processing of pro-GPC3 was inhibited (Fig. 2C), suggesting that the endoproteolytic processing activity of endogenous proprotein convertase in RPE.40 cells is affected by these three polyphenols.

This result leaves us with the question whether furin is also inhibited by these three polyphenols. In order to answer this question RPE.40 cells were transfected with constructs for furin as well as for the pro-IGF-1R. Pro-IGF-1R appears not to be processed by the endogenous PCs in RPE.40 cells (Fig. 2D). In transfection experiments of RPE.40 cells with constructs for both pro-furin and pro-IGF-1R, octyl gallate, dodecyl gallate and nordihydroguariaretic acid appear capable of inhibiting the processing of pro-IGF-1R (Fig. 2D).

Proprotein convertase maturation, an intramolecular autocatalytic process, is required for the formation of the active enzyme. Since the inhibitory effect of octyl gallate, dodecyl gallate and nordihydroguariaretic acid on furin activity is not due to the direct inhibition of the enzyme, we next investigated whether the maturation of the enzyme would be influenced by these compounds. Therefore, the maturation of pro-furin was studied by western blot analysis in CHO cells in the presence of 10 μM of one of these three inhibitors. At a 10 μM concentration of the inhibitors, the amount of pro-furin is increasing and concomitantly, that of intracellular mature furin is decreasing (Fig. 3). The identity of pro-furin was established using rabbit polyclonal antibody hPro-furinAb which is specific for the pro-domain of human pro-furin (result not shown).

HepG2 cells form a hepatocellular carcinoma cell line in which the IGF-1R is overexpressed and in which inhibition of IGF-1R signaling with a monoclonal antibody against IGF-1R was shown to reduce cellular proliferation and to induce apoptosis (17). Overexpression of furin promoted HepG2 cell invasion in tail vein xenograft models (18).

In order to investigate whether the indirect PC inhibitory effect of octyl gallate, dodecyl gallate and nordihydroguariaretic acid could potentially be relevant for inhibition of tumor growth, the IGF-1R signaling pathway was studied in the HepG2 hepatocellular carcinoma cell line. HepG2 cells were used to study the inhibitory effect of these compounds on the processing of pro-IGF-1R. After exposure of HepG2 cells to the polyphenols a strong reduction of processing was found by nordihydroguariaretic acid and octyl gallate while there was a moderate effect with dodecylgallate (Fig. 4).

Normal cells in culture stop growing when coming in physical contact with other cells or when a certain density is reached. Malignant cells, however, form areas of densely-packed cells piling up on the original monolayer of cells. This phenomenon is called focus formation. The relative number of foci formed is a measure for the degree of cell malignancy (19). For HepG2 cells the morphological changes that indicate a reversal of the malignant phenotype have been previously described (20,21).

Cell viability/proliferation as a function of the concentration of the polyphenols was measured using the MTT assay. The relative viability/proliferation of cells grown during 48 h in culture medium containing increasing concentrations of the polyphenols is shown in Fig. 5A. Moderate concentrations (15 μM and below) did only moderately compromise cell viability/proliferation in the HepG2 cell line during the 48 h incubation period. We subsequently tested octyl gallate, dodecyl gallate and nordihydroguariaretic acid for their capacity to reduce focus formation in HepG2 cells. Cells were grown in 6-well plates for 3 weeks and medium supplemented with different concentrations of octyl gallate, dodecyl gallate or nordihydroguariaretic acid was changed every other day. The results are shown in Fig. 5B. In the absence of the inhibitors, cells continue to grow and form foci. With increasing concentrations of the inhibitors, the number and size of the foci reduced indicating a gradual reversal of the malignant phenotype of the HepG2 cells (in each experiment around 100 foci were counted in the control).

Despite the predictive value of the focus forming assay, anchorage-independent growth remains one of the hallmarks of transformation (22). We therefore also investigated the inhibition of anchorage-independent growth of HepG2 cells by octyl gallate, dodecyl gallate and nordihydroguariaretic acid. In the absence of the inhibitors, the liver carcinoma cell line readily formed multiple large colonies as illustrated in Fig. 5C. In the presence of different concentrations of the inhibitors, the colony size as well as the number of the colonies are significantly reduced (Fig. 5C), indicating that these inhibitors at the concentrations used effectively block a hallmark feature of tumor cells, i.e. anchorage-independent cell growth (in each experiment, around 50 colonies were counted in the control).