Inhibition of tumour cell growth by hyperforin, a novel ... - Nature

ã

Oncogene (2002) 21, 1242 ± 1250 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Inhibition of tumour cell growth by hyperforin, a novel anticancer drug from St. John's wort that acts by induction of apoptosis Christoph M Schempp*,1, Vladimir Kirkin2, Birgit Simon-Haarhaus1, Astrid Kersten3, Judit Kiss1, Christian C Termeer1, Bernhard Gilb4, Thomas Kaufmann5, Christoph Borner5, Jonathan P Sleeman2 and Jan C Simon1 1

Department of Dermatology, University of Freiburg, Hauptstrasse 7, D-79104 Freiburg, Germany; 2Institute of Genetics, Forschungszentrum Karlsruhe, PO Box 3640, D-76021 Karlsruhe, Germany; 3Institute of Pathology, University of Freiburg, Albertstrasse 19, D-79104 Freiburg, Germany; 4HWI Analytik, Hauptstrasse 28, 78106 Rheinzabern, Germany; 5Institute of Molecular Medicine, University of Freiburg, Breisacherstr. 66, D-79106 Freiburg, Germany

Hyperforin is a plant derived antibiotic from St. John's wort. Here we describe a novel activity of hyperforin, namely its ability to inhibit the growth of tumour cells by induction of apoptosis. Hyperforin inhibited the growth of various human and rat tumour cell lines in vivo, with IC50 values between 3 ± 15 mM. Treatment of tumour cells with hyperforin resulted in a dose-dependent generation of apoptotic oligonucleosomes, typical DNA-laddering and apoptosis-speci®c morphological changes. In MT-450 mammary carcinoma cells hyperforin increased the activity of caspase-9 and caspase-3, and hyperforin-mediated apoptosis was blocked by the broad-range caspase inhibitor zVAD.fmk. When added to MT-450 cells, hyperforin, but not paclitaxel, induced a rapid loss of the mitochondrial transmembrane potential Dcm, and subsequent morphological changes such as homogenization and vacuolization of mitochondria. Monitoring of Dcm revealed that the hyperforin-mediated mitochondrial permeability transition can not be prevented by zVAD.fmk. This indicates that mitochondrial permeabilization is a cause rather than a consequence of caspase activation. Moreover, hyperforin was capable of releasing cytochrome c from isolated mitochondria. These ®ndings suggest that hyperforin activates a mitochondria-mediated apoptosis pathway. In vivo, hyperforin inhibited the growth of autologous MT-450 breast carcinoma in immunocompetent Wistar rats to a similar extent as the cytotoxic drug paclitaxel, without any signs of acute toxicity. Owing to the combination of signi®cant antitumour activity, low toxicity in vivo and natural abundance of the compound, hyperforin holds the promise of being an interesting novel antineoplastic agent that deserves further laboratory and in vivo exploration. Oncogene (2002) 21, 1242 ± 1250. DOI: 10.1038/sj/ onc/1205190 Keywords: programmed cell death; Hypericum perforatum; mitochondrial membrane permeabilization; caspases; cytochrome c

*Correspondence: CM Schempp, E-mail: [email protected] Received 17 May 2001; revised 14 November 2001; accepted 26 November 2001

Introduction An exciting aspect of apoptosis is that it can be utilized in the treatment of cancer (Revillard et al., 1998; Nicholson, 2000). Anticancer agents with dierent modes of action have been reported to trigger apoptosis in chemosensitive cells (Hickman, 1992; Debatin, 1999). The process of apoptosis consists of dierent phases, including initiation, execution and degradation (Kroemer, 1997), and is activated by two major pathways. One of these is receptor-induced apoptosis which involves the TNF-family of death receptors, including the TNF-R1, CD95 (Fas) and TRAIL-receptors (Ashkenazi and Dixit, 1998; Martinez-Lorenzo et al., 1998). In this so-called èxtrinsic' cell death pathway, a cytoplasmic death domain forms upon ligand binding and trimerization of the receptor (Tartaglia et al., 1993), and recruits adapter proteins like TRADD and FADD. This signalling complex activates an initiator caspase (caspase-8) that, upon accumulation, in turn activates further eector caspases (Thornberry and Lazebnik, 1998; Krammer, 2000; Liu et al., 1996). The other major route leading to apoptosis is the ìntrinsic' cell death pathway which is activated by the release of proapoptotic factors from mitochondria, including cytochrome c and Apaf-1 (Thornberry and Lazebnik, 1998; Liu et al., 1996; Zamzami et al., 1995; Castedo et al., 1996; Reed, 1997; Green and Reed, 1998). In the presence of ATP, these factors recruit procaspase-9 which is converted to caspase-9 and becomes part of the so-called apoptosome complex that activates the caspase cascade (Hengartner, 2000). The release of cytochrome c from the mitochondria is regulated by members of the Bcl-2 protein family (Kroemer, 1997; Kluck et al., 1997; Wolter et al., 1997; Adams and Cory, 1998; Goping et al., 1998; Kroemer and Reed, 2000). Overexpression of Bcl-2 prevents the eux of cytochrome c from the mitochondria and the initiation of apoptosis (Kroemer, 1997; Yang et al., 1997). The release of mitochondrial factors is usually associated with a reduction in the mitochondrial transmembrane potential that can be measured by carbocyanide dyes such as DiOC6(3) (3,3'dihexyloxacarbocyanine iodide) (Zamzami et al.,

Hyperforin is a novel anticancer drug CM Schempp et al

1995; Castedo et al., 1996; Green and Reed, 1998; Kroemer and Reed, 2000). Cytotoxic drugs such as doxorubicin and paclitaxel initiate apoptosis of tumour cells via the Fas/Fas ligand system (Friesen et al., 1996; Srivastava et al., 1999). On the other hand, the plant-derived anticancer drug betulinic acid selectively inhibits neuroectodermal tumours (Pisha et al., 1995), and has been shown to induce apoptosis via the release of proapoptotic factors from mitochondria (Fulda et al., 1998). St. John's wort (Hypericum perforatum L.) is a member of the hypericaceae plant family. Recently, it has gained a reputation as an eective treatment for depression (Linde et al., 1996). In traditional European medicine, lipophilic extracts of St. John's wort have been used for the topical treatment of burns and skin injuries (Roth, 1990). One of the active ingredients of St. John's wort is the phloroglucin-derivative hyperforin (Bystrov et al., 1975; Erdelmeier, 1998). Hyperforin is an acylphloroglucinol-type compound that consists of a phloroglucinol skeleton substituted with lipophilic isoprenchains (Figure 1). It is a natural antibiotic that inhibits the growth of several gram-positive bacteria including meticillin resistant Staphylococcus aureus (Gurevich et al., 1971; Schempp et al., 1999). Recent studies suggest hyperforin is an important neurotransmitter reuptake-inhibiting constituent of St. John's wort (MuÈller et al., 1998). In a previous study we observed a dose-dependent antiproliferative eect of hyperforin in PHA-stimulated peripheral blood lymphocytes (Schempp et al., 2000). Together with the previous report that hyperforin-related substances (hyperevolutins) inhibit the growth of a colon carcinoma cell line (Decosterd et al., 1989), these ®ndings prompted us to adress the question of whether hyperforin displays antiproliferative eects on tumour cells. Hyperforin inhibited the growth of several tumour cell lines in a dose-dependent

Figure 1 Chemical structure of hyperforin

manner. Furthermore, we demonstrated in dierent readout systems that hyperforin exerts its inhibitory eects by activation of a mitochondria-mediated apoptosis pathway. Finally, in an animal model, we observed that hyperforin is able to inhibit the growth of tumour cells in vivo.

1243

Results Hyperforin inhibits tumour cell growth in vitro by induction of apoptosis The antiproliferative potential of hyperforin was ®rst evaluated in six cancer cell lines by [3H]-TdR incorporation. As shown in Figure 2a, the growth of all cell lines was inhibited by hyperforin with IC50 values (half-maximal inhibitory concentrations) of 3 ± 15 mM. The mammary carcinoma cell lines MT-450 (rat), MDA-MB-468 and MCF-7 (human) were more sensitive to hyperforin (IC5055 mM) as compared with the human squamous cell carcinoma (A-431), malignant melanoma (HT-144) and lymphoma (Jurkat) cell lines (Figure 2a). Incubation with the solvent (1% DMSO) did not aect cell proliferation (not shown). Additionally, 10 human and seven rat cell lines were screened for the cytotoxic eect of hyperforin in comparison to the antineoplastic drugs vincristine, paclitaxel and camptothecin. These drugs were tested at their usual eective in vitro concentrations, and cell proliferation was measured by BrdU incorporation. The results are summarized in Table 1. Hyperforin was eective in all cell lines except BDX2 (IC50 40 mg/ ml), whereas several of the cell lines were resistant to the other cytostatic drugs at the tested concentrations (Table 1). We next questioned whether the inhibition of tumour cells by hyperforin was associated with the induction of apoptosis. Apoptosis was determined in dierent readout systems. First, lysates of all above mentioned cell lines were evaluated for the content of apoptotic oligonucleosomes by a colorimetric ELISA. Figure 2B shows a dose-dependent increase of DNA fragmentation in all cell lines after treatment with hyperforin. Apoptosis was further con®rmed in Jurkat cells by DNA-gel electrophoresis and ultrastructural evaluation. DNA extracted from cells treated with hyperforin showed an increased generation of apoptotic DNA fragments as compared with solvent-treated control cells (Figure 3a). The DNA ladder of the solvent control results from spontaneous apoptosis in Jurkat cells growing in culture that can also be observed in Figure 2b. Electron microscopy revealed apoptosis-speci®c features of hyperforin-treated cells such as nuclear condensation, karyorhexis and cytoplasmic edema (Figure 3b). It is concluded that hyperforin is capable of inhibiting the proliferation of cancer cells by induction of apoptosis. Similar experiments also Oncogene


1244

con®rmed that hyperforin induces MT450 cells (data not shown).

apoptosis

in

a

Hyperforin-induced apoptosis of tumour cells is associated with activation of caspases To further elucidate the mechanisms of hyperforininduced apoptosis we assessed the possible involvement of caspases. MT-450 cells were incubated with 20 mM hyperforin and 10 mM paclitaxel for 12 h. The enzymatic activity of caspase-9 and caspase-3 was increased by treatment with hyperforin, whereas caspase-8 activity remained unchanged. In contrast, treatment with paclitaxel also resulted in a strong increase of caspase-8 activity (Figure 4a). Both hyperforin and paclitaxel-induced apoptosis could be blocked by the cell-permeable caspase inhibitor zVAD.fmk (Figure 4b), illustrating the functional relevance of caspase activation during drug-induced apoptosis. The absence of caspase-8 activation in the presence of caspase-9 activation during early hyperforin-induced apoptosis could be due to activation of the ìntrinsic' cell death pathway, because caspase-9 is part of the mitochondria-related apoptosome complex (Hengartner, 2000). We therefore asked whether hyperforin has a direct eect on mitochondria. Mitochondrial activation is an early event during hyperforin-mediated apoptosis

b

Figure 2 Eect of hyperforin on cell proliferation and generation of oligonucleosomes. (a) Antiproliferative potential of hyperforin against cancer cell lines. IC50 values were determined by [3H]-TdR incorporation. Cell lines: mammary carcinoma (MT450, rat; MDA-MB-468 and MCF-7, human); human squamous cell carcinoma (A-431); malignant melanoma (HT-144); lymphoma (Jurkat). Data are expressed as per cent of control (mean of three independent experiments). (b) Dose-dependent increase of DNA-fragmentation after treatment with hyperforin. Cell lysates of all above mentioned cell lines were evaluated for the content of apoptotic oligonucleosomes by a cell death detection ELISA, using biotinylated anti-histone and peroxidase-coupled anti-DNA antibodies. The amount of nucleosomes was photometrically quanti®ed at 405 nm. Data represent the means of three independent experiments Oncogene

MT-450 cells were incubated with hyperforin and paclitaxel and were stained with the carbocyanide dye DiOC6(3) to assess the mitochondrial transmembrane potential (Dcm). Preincubation of DiOC6(3) with hyperforin did not aect DiOC6(3)-induced ¯uorescence of intact mitochondria (not shown). Mitochondria of hyperforin-treated but not of paclitaxel-treated cells underwent a loss of Dcm within 30 min of treatment, as demonstrated by ¯ow cytometry and ¯uorescence microscopy (Figure 5a). Treatment with paclitaxel resulted in a loss of Dcm only at later stages of apoptosis (24 and 48 h) (data not shown). The kinetics of Dcm was assessed in MCF-7 cells using a video microscope. MCF-7 cells (27105/ml) were grown on sterile coverslips, washed and incubated for 10 min with hyperforin (20 mM) in the presence or absence of the caspase inhibitor zVAD.fmk (50 mM). Figure 5b illustrates the ¯uorescence of mitochondrial membranes after the addition of DiOC6(3). Mitochondria from untreated cells developed a stable ¯uorescence, whereas Dcm was almost completely inhibited in cells treated with hyperforin. The caspase inhibitor zVAD.fmk did not interfere with hyperforin-induced Dcm dissipation (Figure 5b). To con®rm that hyperforin induces mitochondrial changes in intact cells, we added hyperforin to MT-450 cells and performed electron microscopic analysis. MT-450 cells treated for 30 min with 20 mM hyperforin contained ultrastructurally normal mitochondria. In contrast, mitochondria in MT-450 cells incubated for 12 h with hyperforin showed a condensed appearance and vacuolization, that evolved into progressive damage of mitochondria after 24 h (Figure 5c).


1245

Table 1 IC50 values of hyperforin compared to other antineoplastic agents Compounds (mg/ml) Human cell lines A431 (squamous cell carcinoma) SB1 (melanoma) SB3 (melanoma) MV3 (melanoma) 1F6 (melanoma) HT144 (melanoma) Jurkat (T cell leukemia) MCF-7 (breast carcinoma) MDA-MB-468 (breast carcinoma) SK-OV-3 (ovary carcinoma) Rat cell lines MT450 (breast carcinoma) MAT-Lu (prostate carcinoma) AT-2.1 (prostate carcinoma) AR42J (pancreas carcinoma) ARIP (pancreas carcinoma) RG2 (glioblastoma) BDX2 (fibrosarcoma)

Hyperforin

Camptothecin

Paclitaxel

Vincristine

4.50 4.50 4.50 2.50 4.50 6.00 6.50 1.50 2.00 3.00

41.0 0.8 0.3 0.08 0.2 0.05 0.006 41.0 0.3 41.0

41.0 0.04 0.5 0.05 0.1 0.008 0.008 1.00 0.4 41.0

41.0 1.0 0.008 41.0 0.007 0.008 0.0001 0.01 0.08 41.0

1.50 9.00 1.90 1.40 2.20 2.50 40.0

0.007 41.0 41.0 41.0 0.04 41.0 41.0

0.008 41.0 0.1 41.0 0.5 41.0 41.0

0.0007 41.0 41.0 41.0 0.1 41.0 0.2

Cell proliferation was determined by BrdU incorporation. Results are shown as half maximal inhibitory concentrations (IC50 values). Drug concentrations tested were 2.5, 5, 10, 20, 40, 80 mg/ml (Hyperforin); and 0.0001, 0.001, 0.01, 0.1, 1.0 mg/ml (Camptothecin, Paclitaxel, Vincristine)

Hyperforin causes release of cytochrome c from isolated mitochondria To address the question of whether hyperforin induces cytochrome c release from mitochondria, a mitochondria-enriched fraction prepared from HEK-293 cells was incubated in the presence or absence of hyperforin (20 mM) or staurosporine (10 mM) for 0, 5, 10 and 30 min. We observed a rapid release of cytochrome c from hyperforin-treated mitochondria that was comparable to staurosporine (Figure 6). Hyperforin inhibits tumour growth in vivo To assess the potential of hyperforin to inhibit tumour growth in vivo, female Wistar Furth rats were injected subcutaneously with rat mammary carcinoma MT-450 cells. Drug treatment with equimolar (2 mm) concentrations of hyperforin or paclitaxel was initiated 15 days after tumour injection, with daily s.c. injections over a period of two weeks. As shown in Figure 6, hyperforin inhibited tumour growth to a similar extent as compared with paclitaxel. In each of the treatment groups, a complete lack of toxicity, as judged by changes in body weight or other forms of acute toxicity, was noted. Discussion Here we describe the potential of hyperforin to inhibit the growth of 17 human and rat tumour cell lines in vitro. Cytotoxic drugs act primarily by inducing apoptosis in sensitive target cells (Hickman, 1992; Debatin, 1999). We therefore evaluated apoptosis of hyperforin-treated cells in dierent readout systems. Treatment of tumour cells with hyperforin resulted in a

dose-dependent generation of apoptotic oligonucleosomes, increased DNA-laddering and apoptosis-speci®c morphological changes. Our data suggest that hyperforin acts by triggering the ìntrinsic' cell death pathway. Firstly, by using speci®c ELISAs, we found that treatment with hyperforin resulted in activation of caspase-9 and caspase-3. We tested a caspase inhibitor for its eect on cell death induced by hyperforin. Using MT-450 cells, the broad-range caspase inhibitor zVAD.fmk completely blocked apoptosis induced by hyperforin. This indicates that caspases downstream of mitochondria are important eectors of hyperforin-induced apoptosis. Secondly, when added to intact tumour cells, hyperforin but not paclitaxel speci®cally induced early mitochondrial alterations. These consisted of a loss of Dcm that could be detected as early as 30 min after treatment with hyperforin, and subsequent morphological changes such as homogenization and vacuolization of mitochondria, detected within 12 h after treatment with hyperforin. We have also shown that the caspase inhibitor zVAD.fmk did not interfere with hyperforin-induced Dcm dissipation. Thus, hyperforin can directly trigger mitochondrial permeability transition without involvement of a zVAD.fmk-inhibitable caspase. Consistently, we observed a rapid release of cytochrome c from hyperforin-treated mitochondria. These data are in favour of the hypothesis that hyperforin is able to inhibit tumour cell growth by triggering the ìntrinsic' cell death pathway. Although the early mitochondrial changes are consistent with this putative mechanism of action, we cannot rule out the possibility that hyperforin kills tumour cells by inducing some other damage to cells which then activates the apoptotic program. The precise molecular mechanisms of Oncogene


1246

a

a

b

b

Figure 3 Evaluation of apoptosis by DNA-electrophoresis and electron microscopy. (a) Apoptotic DNA fragmentation in Jurkat cells. Cells were cultured for 12 h with the solvent (1) or 20 mM hyperforin (2). DNA was extracted by NP40 lysis and separated on a 3% agarose gel. M, KiloBaseTM DNA marker. (b) Electron microscopic detection of apoptosis. Jurkat cells were incubated for 24 h in the absence or presence of 20 mM Hyperforin. Ultrathin sections (50 nm) were stained with 5% uranyl acetate and 0.2% lead citrate. Untreated cells show normal morphology (*), whereas hyperforin-treated cells show apoptosis-speci®c features such as nuclear condensation, karyorhexis (k) and cytoplasmic edema (e). Bar, 500 nm

hyperforin-induced apoptosis are currently under investigation. A highly important ®nding from the studies presented here is that in addition to inducing tumour cell apoptosis in vivo, hyperforin also inhibited the growth of autologous MT-450 breast carcinoma cells in vivo in immunocompetent Wistar rats. Tumour growth was inhibited to a similar extent as with the cytotoxic drug paclitaxel, in the absence of any signs of acute toxicity. This observation is important, because paclitaxel is a drug which is already widely used as an anticancer agent in the clinical setting (Rowinsky and Donehower, 1995). The dose of paclitaxel used in our experiment was about 10 times under the dose normally used in animal experiments. This was necessary because we wanted to use equimolar concentrations of paclitaxel and hyperforin, and we were not able to increase the concentration of the Oncogene

Figure 4 Involvement of caspases in hyperforin-induced apoptosis. (a) MT-450 cells were treated for 12 h with 20 mM hyperforin (black bars), 10 mM paclitaxel (hatched bars), or were left untreated (open bars). Cell lysates were tested for protease activity by the addition of caspase-speci®c peptides conjugated to the colour reporter p-nitroanilide. Compared with the untreated control, the enzymatic activity of caspase-9 and caspase-3 was increased by treatment with hyperforin, whereas caspase-8 activity remained unchanged. In contrast, treatment with paclitaxel also resulted in a strong increase of caspase-8 activity. (b) MT-450 cells were treated for 24 h with 20 mM hyperforin or 10 mM paclitaxel in the presence or absence of the caspase inhibitor zVAD.fmk (50 mM). Cells were then evaluated for apoptosis using a cell death detection ELISA. Data represent the means +s.d. of three experiments. **P50.01 (Students paired t-test)

lipophilic compound hyperforin in an aqeous solution. However, we are improving the solubility of hyperforin at the present and hopefully we will soon be able to present further data on the mean tolerated dose and eectivity of hyperforin in vivo. Although mammary carcinoma cells were the most sensitive to hyperforin in terms of inhibition of proliferation, hyperforin inhibited proliferation and induced apoptosis in all tumour cells tested. These


a

b

observations suggest that hyperforin might be a broadspectrum anti-tumour reagent, with activity against a wide range of dierent cancers. Signi®cant antitumour activity in the absence of toxicity in vivo is a prerequisite for the development of new anticancer drugs. Another requirement for the development of new plant-derived anticancer drugs is ample supply in terms of plant availability and yield of the chemotherapeutic agent. Acylphloroglucinol-type compounds, of which hyperforin is the principal component, represent the largest group of compounds in St. John's wort (Hypericum perforatum). Crude extracts of the drug contain up to 5% of hyperforin (Roth, 1990; Erdelmeier, 1998; Maisenbacher and Kovar, 1992). St. John's wort is found in abundance throughout Europe, North America, Asia and North Africa. Together with low toxicity in vivo and natural abundance of the compound, we suggest that hyperforin is a potential antitumour drug that thoroughly deserves further investigation.

1247

Materials and methods Hyperforin and cytotoxic drugs

c

Figure 5 Hyperforin disrupts mitochondrial membranes. (a) MT-450 cells were treated for 20 min with 20 mM hyperforin or 10 mM paclitaxel. The cells were then incubated with DiOC6(3) for 10 min and analysed on a FACScan ¯ow cytometer (upper panel) and by ¯uorescence microscopy (lower panel). The bars in the upper panel represent the percentage of DiOC6(3)-positive gated cells. Mitochondria of hyperforin-treated but not of paclitaxeltreated cells underwent a loss of Dcm within 30 min of treatment. (b) Time lapse study of Dcm. MCF-7 tumour cells attached to coverslips were incubated for 10 min with hyperforin (20 mM) in the presence or absence of the broad-range caspase inhibitor zVAD.fmk (50 mM). Subsequently, the coverslips were rinsed with PBS and were transferred into the closed bath ¯ow microchamber (378C) of the video microscope. DiOC6(3) (50 nm ®nal concentration) was injected directly into the chamber and microphotographs were taken every 5 s. The graphs illustrate the mean ¯uorescence (584 nm) of gated cells over a period of 60 min. The experiment was repeated with similar results. (c) Cultured cells were studied by electron microscopy. MT-450 cells treated with 20 mM hyperforin for 30 min show the ultrastructural details of normal mitochondria (left). In contrast, mitochondria in MT-450 cells incubated for 12 h with hyperforin begin to show a condensed appearance and vacuolization (middle), that evolves into progressive damage after 24 h (right). Bar, 150 nM

Hyperforin was extracted under protection from light and air from a supercritical CO2-extract of Hypericum perforatum (Flavex, Rehlingen, Germany). The CO2-extract was mixed with methanol (1 : 10, w : v), centrifuged and evaporated. The residue was chromatographed on a preparative silica gel plate (silica gel 60 F254, 20620 cm, 0.5 mm; mobile phase : hex-

Figure 6 Hyperforin causes release of cytochrome c from isolated mitochondria. A mitochondria-enriched fraction was prepared from HEK-293 cells as described in the Materials and methods. The isolated mitochondria were resuspended in MSH buer supplemented with proteinase inhibitors and were incubated in the presence or absence of hyperforin (20 mM) or staurosporine (10 mM). Supernatants were collected after 0, 5, 10 and 30 min and their cytochrome c content was determined using a speci®c ELISA. The experiment was repeated with similar results Oncogene


1248

a gift from Dr GP van Muijen (Dept. of Pathology, Nijmegen, The Netherlands); MAT-Lu and AT-2.1, rat prostate carcinoma (Isaacs et al., 1986); AR42J and ARIP, rat pancreas carcinoma (ATCC); RG2, rat glioblastoma (ATCC); BDX2, rat ®brosarcoma (Sleeman et al., 1996); and MT-450, rat breast adenocarcinoma (Kim, 1986; Sleeman et al., 1999). All cells, with the exception of MCF-7, were cultured in RPMI supplemented with 10% FCS, 1% L-Glutamine and 1% Penicillin/Streptomycin (all from Gibco, Eggenstein, Germany) in a humidi®ed athmosphere (5% CO2, 378C). MCF-7 cells were grown in Dulbeccos minimal essential medium supplemented with 2 mM L-glutamine, 0.1 mM nonessential amino acids, 1.0 mM sodium pyruvate, 10% FBS, 1% L-glutamine and 1% Penicillin/Streptomycin (Gibco). Proliferative assays Figure 7 Antitumour activity of hyperforin in vivo. Female Wistar Furth rats were injected subcutaneously with 56105 MT450 cells in PBS. Drug treatment was initiated 15 days after tumour cell injection when tumours were readily detectable and easily measured. The animals received 100 ml of 2 mm hyperforin (n=8), 2 mM paclitaxel (n=8), or solvent (n=8). Daily injections were administered subcutaneously at the site of the tumour cell injection for two weeks. Data are indicated as mean + standard error of mean

ane:diethylether 80 : 20) (Merck, Darmstadt, Germany), and on a silica gel column (LiChrospher# Si 10, 12 mM; mobile phase: cyclohexane : ethylacetate gradient) (Merck, Darmstadt). The resulting extracts were combined, evaporated, dried an analysed. The hyperforin content of the dried residue was 490%. The identity of hyperforin was con®rmed by HPLC/PDA, 1H-NMR-, 13C-NMR-, IR- and and UVspectral analysis. The purity of hyperforin was assessed by two dierent validated HPLC-methods. All steps of the isolation were controlled by analytical HPLC. To obtain hyperforin aliquots, the compound was dissolved in methanol and known amounts of hyperforin were evaporated in brown glass tubes. The tubes were stoppered and were stored at 7308C under a liquid nitrogen atmosphere. Repeated analytical controls revealed a maximal decrease of hyperforin of 0.5% during a storage period of 6 months. Paclitaxel was obtained from Tocris (Tocris Cookson, MO, USA). Camptothecin, staurosporine and vincristine were obtained from Sigma (Sigma, St. Louis, MO, USA). All drugs were dissolved in DMSO and further diluted with phosphatebuered saline (PBS) to obtain stock solutions. Further dilutions were made with medium to obtain ®nal concentrations as indicated. The maximal ®nal DMSO concentration was 1%. Cell lines The following cell lines were used: A-431, human epidermoid carcinoma; HT-144, human melanoma; Jurkat, human acute T cell leukemia; MCF-7, human breast adenocarcinoma, expressing estrogen receptors; MDA-MB-468; human breast adenocarcinoma, lacking estrogen receptor expression; SKOV-3, human ovary carcinoma, lacking p53 expression; HEK-293, human embryonal kidney cells (all from ATCC, Manassas, VA, USA); SB1 and SB3, human melanoma, were a gift from Dr CM Hamby (New York Medical College, Valhalla, NY, USA); MV3 and 1F6, human melanoma, were Oncogene

Adherent cells (all except MT-450 and Jurkat) were collected with Trypsin/EDTA (Gibco). All cells were washed twice with PBS, resuspended in medium and further cultured (16105 cells/ml) in 96-well microtitre plates (Costar, Cambridge, MA, USA). Following addition of hyperforin (5 ± 180 mM) the cells were incubated for 24 h. Cells were then pulsed with 1 mCi [3H]-TdR/ well for 18 h and [3H]-TdR incorporation was determined by liquid scintillation spectroscopy using a Top-Count (Canberra Packard, Frankfurt, Germany) (Schempp et al., 2000). Data are expressed as per cent of control (mean of three independent experiments). Alternatively, the incorporation of 5-bromo-2'-deoxyuridine (BrdU) into the DNA of proliferating cells was measured using a colorimetric BrdU cell proliferation ELISA (Roche Molecular Biochemicals, Mannheim, Germany). The assay was performed according to the manufacturer's instructions. Photometric determination of apoptosis Cells (16104 cells/ml) were treated with various hyperforin concentrations for 24 h as described above. Cells were then evaluated for apoptosis using a cell death detection ELISA (Cell Death Detection ELISAPLUS, Roche Molecular Biochemicals). The principle of this test is the detection of monoand oligonucleosomes in the cytoplasmic fractions of cell lysates using biotinylated anti-histone and peroxidase-coupled anti-DNA antibodies. The amount of nucleosomes is photometrically quanti®ed at 405 nm by the peroxidase activity retained in the immunocomplexes. Data are expressed as mean of three independent experiments. In a separate experiment, MT-450 cells (16104 cells/ml) were treated for 24 h with 20 mM hyperforin or 10 mM paclitaxel in the presence or absence of the caspase inhibitor zVAD.fmk (50 mM) (R&D Systems, Wiesbaden, Germany) and apoptosis was determined as described above. DNA gel electrophoresis Apoptotic DNA fragments were isolated from Jurkat cells treated for 12 h with hyperforin (20 mM) as described (Herrmann et al., 1994; Schempp et al., 2001). Brie¯y, cells were washed and pelleted by centrifugation. Cell pellets were treated for 10 s with lysis buer (1% NP40; 20 mM EDTA; 50 mM Tris-HCl, all from Sigma). The cells were then mixed with 1% SDS (Sigma), treated for 2 h with RNase (5 mg/ml) (Boehringer, Mannheim, Germany) at 568C and digested with proteinase K (2.5 mg/ml) (Sigma) for 2 h at 378C. After addition of 10 m ammonium acetate (Merck, Darmstadt, Germany) the DNA was precipitated with 100% ethanol at


7208C, resuspended in gel loading buer (10 mM Tris-HCl, 1 mM EDTA, Sigma) and separated on 3% agarose gels together with a KiloBaseTM DNA marker (Pharmacia, Freiburg, Germany). Electron microscopy Jurkat cells (16106/ml) were treated with 20 mM hyperforin for 24 h, and MT-450 cells were treated with 20 mM hyperforin for 30 min, 12 h and 24 h, respectively. After centrifugation, cell pellets were ®xed for 12 h in 3% cacodylate-buered glutaraldehyde. Ultrathin sections (50 nm) were stained with 5% uranyl acetate and 0.2% lead citrate. Examination was carried out using a Philips CM10 electron microscope (Philips, Eindhoven, The Netherlands). Caspase assay

1249

Cytochrome c release from isolated mitochondria 56107 HEK-293 cells were harvested, washed in PBS and resuspended in MSH buer (210 mM mannitol, 70 mM sucrose, 20 mM HEPES, 1 mM EDTA, pH 7.4) plus protease inhibitors. The cells were lysed by passing them ®ve times through a 21G and a 23G needle. Nuclei and cellular debris were removed by centrifugation at 5006g for 5 min. The post-nuclear supernatant was centrifuged at 51006g for 10 min, and the obtained pellet was washed twice in PBS. The resulting crude mitochondria were resuspended in MSH buer and stored at 48C until use. Mitochondria were treated with hyperforin (20 mM) or staurosporine (10 mM), and mitochondrial supernatants were obtained by high speed centrifugation (15 0006g, 5 min, 48C) after 0, 5, 10 and 30 min. The cytochrome c content of the supernatants was measured using a speci®c ELISA (R&D systems).

To determine the enzymatic activity of caspases-3, -8 and -9, semi-quantitative colorimetric ELISA's were performed. MT450 cells (16105/ml) were treated for 12 h with 20 mM Hyperforin or 10 mM Paclitaxel. Cells were lysed in ice-cold hypotonic buer containing 20 mM Tris-HCl (pH 7.2), 1 mM EDTA and protease inhibitors to prevent nonspeci®c cleavage of proteins. Homogenates were clari®ed by centrifugation for 10 min at 16 0006g. Cell lysates were then tested for protease activity by the addition of caspase-speci®c peptides conjugated to the colour reporter p-nitroanilide (R&D systems). The cleavage of the peptide by the caspase releases the chromophore, which is quantitated spectrophotometrically at 405 nm. The level of caspase enzymatic activity is directly proportional to the colour reaction.

Female Wistar Furth rats were injected subcutaneously with 56105 MT-450 cells in PBS. Drug treatment was initiated 15 days after tumour injection, a time at which tumours were readily detectable and could easily be measured. The animals received 100 ml of 2 mM drug solution (hyperforin, eight per group; paclitaxel, eight per group) or the solvent (10% DMSO, eight per group). Daily injections of the drugs/ control were administered subcutaneously at the site of the tumour cell injection for two weeks. Tumours were measured with a micrometre calliper every third day throughout the study (Sleeman et al., 1999). Studies with laboratory animals were approved by the local Ethical Review Board.

Determination of the mitochondrial membrane potential Dcm

Statistical analysis

MT-450 cells (16106/ml) were treated for 20 min with 20 mM Hyperforin or 10 mM Paclitaxel. The cells were then incubated for 10 min with DiOC6(3) (50 nm; Molecular Probes, Eugene, OR, USA) at 378C, washed in PBS and analysed on a FACScan ¯ow cytometer (Becton Dickinson, San Jose, CA, USA) and by ¯uorescence microscopy (Zeiss, Jena, Germany). The kinetics of Dcm were assessed in MCF-7 cells using timelapse video microscopy as described previously (TILL Photonics, Martinsried, Germany) (Termeer et al., 2001). MCF-7 cells (26105/ml) were grown on sterile coverslips for 12 h. The cells were washed and were incubated for 10 min with hyperforin (20 mM) in the presence or absence of the caspase inhibitor zVAD.fmk (50 mM). Subsequently, the coverslips were rinsed with PBS and were transferred into the closed bath ¯ow micro-chamber (378C) of the video microscope. DiOC6(3) (50 nM ®nal concentration) was injected directly into the chamber and microphotographs were taken every 5 s over a period of 60 min. The mean ¯uorescence (584 nm) of 30 randomly selected cells per ®lm were calculated with the TILL VisionTM v3.3 software.

The mean and standard deviation of several experiments is shown, unless otherwise indicated. The number of experiments and the con®rmatory tests used are indicated in the ®gure legends.

Tumour growth in vivo

Acknowledgments We thank Norma Howells and Jonathan Ward for animal care; Marco Averbeck and Bernhard Kremer for technical assistance with the video microscope; and Stefan Martin for critically reading the manuscript. CM Schempp was supported by the Carl und Veronika Carstens-Stiftung and the Paul und Yvonne Gillet-Stiftung (S012/01.221/97). T Kaufmann and C Borner were supported by the Swiss National Science Foundation (31-57236.99). Judit Kiss is a scienti®c exchange fellow of the Semmelweis University, Budapest, Hungary, supported by a grant from the University Medical Center Freiburg.

References Adams JM and Cory S. (1998). Science, 281, 1322 ± 1325. Ashkenazi A and Dixit VM. (1998). Science, 281, 305 ± 308. Bystrov NS, Chernov BK, Dobrynin VN and Kolosov MN. (1975). Tetrahedron Lett., 32, 2791 ± 2794. Castedo M, Hirsch T, Susin SA, Zamzami N, Marchetti P, Macho A and Kroemer G. (1996). J. Immunol., 157, 512 ± 521. Debatin KM. (1999). Adv. Exp. Med. Biol., 457, 237 ± 244.

Decosterd LA, Stoeckli-Evans H, Chapuis JC, Msonthi JD, Sordat B and Hostetmann K. (1989). Helv. Chim. Acta., 72, 464 ± 471. Erdelmeier CAJ. (1998). Pharmacopsychiatry, 31 (Suppl), 2 ± 6. Friesen C, Herr I, Krammer PH and Debatin KM. (1996). Nat. Med., 2, 574 ± 577. Oncogene


1250

Oncogene

Fulda S, Scadi C, Susin SA, Krammer PH, Kroemer G, Peters ME and Debatin KM. (1998). J. Biol. Chem., 273, 33942 ± 3398. Goping IS, Gross A, Lavoie JN, Nguyen M, Jemmerson R, Roth K, Korsmeyer SJ and Shore GC. (1998). J. Cell. Biol., 143, 207 ± 215. Green D and Reed JC. (1998). Science, 281, 1309 ± 1312. Gurevich AI, Dobrynin VN, Kolosov MN, Popravko SA, Ryabova ID, Chernov BK, Derbentseva NA, Aizenman BE and Garagulya AD. (1971). Antibiotiki, 16, 510 ± 513. Hengartner MO. (2000). Nature, 407, 770 ± 776. Herrmann M, Lorenz HM, Voll R, GruÈnke M, Woith W and Kalden JR. (1994). Nucl. Ac. Res., 22, 5506 ± 5507. Hickman JA. (1992). Cancer Met. Rev., 11, 121 ± 139. Isaacs JT, Isaacs WB, Feitz WF and Scheres J. (1986). Prostate, 9, 261 ± 281. Kim U. (1986). J. Surg. Oncol., 33, 151 ± 165. Kluck RM, Bossy-Wetzel E, Green DR and Newmeyer DD. (1997). Science, 275, 1132 ± 1136. Krammer PH. (2000). Nature, 407, 789 ± 795. Kroemer G and Reed JC. (2000). Nat. Med., 6, 513 ± 519. Kroemer G. (1997). Nat. Med., 3, 614 ± 620. Linde K, Ramirez G, Muirow CD, Pauls A, Weidenhammer W and Melchart D. (1996). Brit. Med. J., 313, 253 ± 258. Liu X, Kim CN, Yang J, Jemmerson R and Wang X. (1996). Cell, 86, 147 ± 157. Maisenbacher P and Kovar KA. (1992). Planta Med., 58, 351 ± 354. Martinez-Lorenzo MJ, Alava MA, Gamen S, Kim KJ, Chuntharapai A, Pineiro A, Naval J and Anel A. (1998). Eur. J. Immunol., 28, 2714 ± 2725. MuÈller WE, Singer A, Wonnemann M, Hafner U, Rolli M and SchaÈfer C. (1998). Pharmacopsychiatry, 31 (Suppl 1): 16 ± 21. Nicholson DW. (2000). Nature, 407, 810 ± 816. Pisha E, Chai H, Lee IS, Chagwedera TE, Farnsworth NR, Cordell GA, Beecher CWW, Fong HHS, Kinghorn AD, Brown DM, Wani MC, Wall ME, Hieken TJ, Das Gupta TK and Pezzuto JM. (1995). Nat. Med., 1, 1046 ± 1051.

Reed JC. (1997). Cell, 91, 559 ± 562. Revillard JP, Adorini L, Goldman M, Kabelitz D and Waldmann H. (1998). Immunol. Today, 19, 291 ± 293. Roth L. (1990). Hypericum ± Hypericin. Ecomed, Germany: Landsberg/Lech, pp. 97 ± 109. Rowinsky EK and Donehower RC. (1995). New Engl. J. Med., 332, 1004 ± 1014. Schempp CM, Pelz K, Wittmer A, SchoÈpf E and Simon JC. (1999). Lancet, 353, 2129. Schempp CM, Winghofer B, LuÈdtke R, Simon-Haarhaus, SchoÈpf E and Simon JC. (2000). Brit. J. Dermatol., 142, 979 ± 984. Schempp CM, Simon-Haarhaus B, Termeer CC and Simon JC. (2001). FEBS Lett, 493, 26 ± 30 Sleeman JP, Arming S, Moll J, Hekele A, Rudy W, Hofmann M, Kreil G, Ponta H and Herrlich P. (1996). Cancer Res., 56, 3134 ± 3141. Sleeman JP, Kim U, LePendu J, Howells N, Coquerelle T, Ponta H and Herrlich P. (1999). Oncogene, 18, 4485 ± 4494. Srivastava RK, Sasaki CY, Hardwick JM and Longo DL. (1999). J. Exp. Med., 190, 253 ± 263. Tartaglia LA, Ayres TM, Wong GHW and Goeddel DV. (1993). Cell, 74, 845 ± 853. Termeer CC, Johannsen H, Braun T, Renkl A, Ahrens T, Denfeld RW, Lappin MB, Weiss JM and Simon JC. (2001). J. Leuk. Biol., 70, 715 ± 722. Thornberry NA and Lazebnik Y. (1998). Science, 281, 1312 ± 1316. Wolter KG, Hsu YT, Smith CL, Nechushtan A, Xi XG and Youle RJ. (1997). J. Cell. Biol., 139, 1281 ± 1292. Yang J, Liu X, Bhalla K, Kim CN, Ibrado AM, Cai J, Peng TI, Jones DP and Wang X. (1997). Science, 275, 1129 ± 1132. Zamzami N, Marchetti P, Castedo M, Zanin C, VayssieÁre JL, Petit PX and Kroemer G. (1995). J. Exp. Med., 181, 1661 ± 1672.