Genistein is a potent plant-derived isoflavone displaying estrogenic activity at low (nanomolar) concentrations and antiproliferative and antiangiogenic properties at higher concentrations (above 10–50 μM). The antiproliferative potential of genistein has made it an interesting candidate for cancer chemotherapy at high concentrations; however, the potential for genistein toxicity in neurons at such concentrations has not been previously addressed. We show that genistein is toxic to rat primary cortical neurons at a concentration of 50 μM, whereas daidzein, a structural analog, shows no toxicity at similar concentrations. The dying cells display an apoptotic morphology that is characterized by fragmented nuclei, appearance of apoptotic bodies, DNA laddering, and caspase-dependent poly(ADP-ribose) polymerase cleavage. This cell death is partially dependent on caspase activity, independent of estrogen receptors, and does not result in a significant loss of Bcl-2 or Bcl-XL protein. Genistein exposure induces delayed and prolonged activation of p42/44 mitogen-activated protein kinase (MAPK) and p38 MAPK but not c-Jun NH2-terminal kinase. The specific p42/44 MAPK kinase inhibitor PD98059 (50 μM) partially blocks genistein-induced apoptosis, whereas the p38 MAPK inhibitor SB202190 (10 μM) has no effect. Genistein elevates intracellular calcium and both 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester (1 μM) and dantrolene (10 μM) inhibit genistein-induced apoptosis, suggesting a link between genistein-induced intracellular calcium release and apoptosis. The combination of dantrolene and PD98059 block genistein-induced apoptosis in an additive manner compared with either compound alone. These findings provide evidence for a proapoptotic function of p42/44 MAPK and raise caution about potential side effects in the nervous system with genistein use as a high-dose therapeutic agent.
The plant-derived isoflavone genistein has gained popularity as an over-the-counter alternative to estrogen replacement therapy and, at higher doses, it has been suggested to be an effective tumor suppressive agent. Due to impressive antiproliferative effects in cancerous cells in vitro, it has been suggested that doses of genistein in the high micromolar range may be useful in chemotherapy (Kelloff et al., 1996; Polkowski and Mazurek, 2000). On a cellular level, genistein appears to elicit estrogen-like proliferative effects at nanomolar doses (Martin et al., 1978) but reduces tumor cell proliferation (Dixon-Shanies and Shaikh, 1999) and blocks angiogenesis (Fotsis et al., 1993) at micromolar doses. It has been suggested that these effects occur through cell cycle inhibition (Matsukawa et al., 1993;Alhasan et al., 1999). Due to a reported lack of toxicity in untransformed fibroblasts and primary keratinocytes in vitro (Fioravanti et al., 1998; Alhasan et al., 1999), it has been postulated that genistein targets tumor-derived rapidly dividing but spares nondividing cells. Using in vitro cultures of postmitotic primary cortical neurons, we show that genistein is a potent proapoptotic agent, raising the possibility of in vivo neurotoxicity at high genistein concentrations.
At high doses, genistein has multiple intracellular effects, including inhibition of the activity of certain tyrosine kinases such as Src and the epidermal growth factor receptor (Akiyama et al., 1987), inhibition of topoisomerase II (Markovits et al., 1989), and alterations in phosphatidylinositol turnover. Additionally, genistein has been reported to affect the mitochondrial membrane permeability transition pore (Yoon et al., 2000) and ion channel activity (Paillart et al., 1997; Huang et al., 1999). The multiple intracellular effects of genistein with potential therapeutic value impart to the agent a unique pharmacological profile that merits further investigation.
Neurotoxicity can be necrotic (lytic) or apoptotic in nature. Necrotic stimuli cause a rapid loss of plasma and nuclear membrane integrity, whereas apoptosis involves the activation of signal transduction pathways, caspase activation, plasma membrane blebbing, and DNA laddering. In many systems it has been determined that the MAPK family of proteins, particularly p42/44 MAPK, p38 MAPK, and JNK play an important role in apoptosis. It has been well established that the activation of the JNK and p38 MAPK pathways leads to the phosphorylation of a variety of proapoptotic downstream effectors, whereas the p42/44 MAPK pathway is more often associated with cell survival (Xia et al., 1995; Cross et al., 2000). However, there are also reports suggesting that each pathway can act in either an anti- or proapoptotic manner. For example, the JNK pathway has been reported in certain instances to induce proliferation rather than apoptosis (Nishina et al., 1997; Smith et al., 1997), whereas the p42/44 MAPK pathway has been reported to induce apoptosis in neurons as well as other cell types (Stanciu et al., 2000; Wang et al., 2000).
Additionally, perturbations in intracellular calcium levels have been associated with cellular apoptosis. It has been established that misregulation of intracellular calcium can lead to mitochondrial dysfunction, activation of caspases, activation of calcium-sensitive proteases and kinases, and cell death (Kruman and Mattson, 1999). Given its therapeutic potential, we were interested in determining whether genistein is capable of inducing apoptosis or necrosis in neurons, and, if so, whether calcium and the MAPK family of proteins are involved in mediating this toxicity.
Materials and Methods
Primary Neuron Isolation and Culture.
Cortices were dissected from E18 Sprague-Dawley rat pups and dissociated as described previously (Singer et al., 1996). Briefly, cortices were minced and dissociated with trypsin (250 μg/ml) for 10 min at 37°C. The tissue was centrifuged in a DNase/trypsin inhibitor solution and triturated through a Pasteur pipet. Suspended cells were underlayed with bovine serum albumin, centrifuged, and resuspended in phenol red-free neurobasal medium (Invitrogen, Carlsbad, CA) supplemented with B27 (Invitrogen), l-glutamine (0.5 mM), and gentamycin (50 μg/ml). Cells were plated at a density of 3 × 105 cells/cm2 on poly-d-lysine-coated plates and were allowed to mature for 5 days before use. This protocol has been reported to produce cultures with >99% neuronal composition (Brewer et al., 1993) and the consistent near-absence of glia in the cultures was confirmed by morphological inspection. Morphological determination showed that cells with phase dark cell bodies were healthy neurons, whereas apoptotic neurons showed signs of membrane blebbing, loss of neurites, and the formation of phase-light apoptotic bodies.
Reagents and Treatments.
All drugs were made as 1000× stocks in DMSO and stored at −20°C away from light. Genistein and daidzein were obtained from ICN Biochemicals (Cleveland, OH). PD98059 and SB202190 were obtained from Calbiochem (San Diego, CA). BAPTA-AM and dantrolene were both obtained from Sigma Chemical (St. Louis, MO). ICI 182,780 was obtained from Tocris Cookson (Ballwin, MO). Inhibitors were added 15 min prior to genistein treatment except where noted otherwise.
To purify fragmented genomic DNA from primary cortical neurons, a modified version of the method of Yan et al. (2000) was used. Briefly, 3 × 106 cells were lysed in buffer (10 mM Tris, pH 7.4, 5 mM EDTA, 1% Triton) for 20 min at 4°C. The samples were then centrifuged at 4°C for 20 min at 11,000g. The supernatant was removed and incubated with 100 μg/ml RNase A for 1 h at 37°C. Proteinase K was then added (100 μg/ml) and the samples were incubated for 1 h at 56°C. The samples were ethanol precipitated and washed with 70% ethanol. The DNA was resuspended in 50 μl of nuclease-free water and one-half was run on a 1.8% agarose gel and analyzed for the presence of a laddering pattern.
The live/dead assay was obtained from Molecular Probes (Eugene, OR) and used according to the manufacturer's instructions. Briefly, cells were grown on Nalge permanox plastic tissue culture slides (Fisher Scientific, Pittsburgh, PA) and treated as indicated in the figure legends. The cells were then incubated with calcein AM (1 μM) and ethidium homodimer (2 μM) for 15 min at 37°C, washed twice with phosphate-buffered saline, and coverslips were applied. Green fluorescent cells were the product of mitochondrial cleavage of calcein AM and were observed using a Nikon Optiphot 2 microscope (Nikon, Tokyo, Japan) with the EF-D fluorescence attachment and G-1B and DM510 filters and counted as living cells. Red-orange fluorescence from the nuclei of dead cells was observed. The red fluorescence resulted from the binding of the cell-impermeable fluorescent dye ethidium homodimer to the nuclei of cells with compromised plasma membranes. The number of living cells (green cytoplasm) plus dead cells (red nuclei) was used as the total cell population.
For determination of plasma membrane integrity loss, LDH release into the extracellular medium was measured using the cyto-tox96 nonradioactive assay from Promega (Madison, WI). This assay measures the formation of a red formazan product after the conversion of lactate and NAD+ to pyruvate and NADH. The assay was used according to the manufacturer's instructions. Briefly, cells were plated on a 24-well plate. One well was lysed in a 1% Triton X-100 solution for 45 min at 37°C to control for maximum LDH release. Culture medium (50 μl) was transferred to a 96-well plate and incubated for 30 min at room temperature in the dark with the colorimetric reagent. Acetic acid (1 M) was added to stop the reaction and color development was read on a Packard SpectraCount 96-well spectrophotometer (Packard Instruments, Meredin, CT) at 490 nm.
Phase Contrast Microscopy.
Phase contrast micrographs were taken to show that the results seen in the cytotoxicity assays correlate with the number of phase-dark cells seen by morphological inspection. Micrographs were taken using a Zeiss microscope (Carl Zeiss Inc., Thornwood, NY) and Nikon Coolpix digital camera.
Cells grown on tissue culture slides were incubated in Hoescht 33342 (10 μg/ml) for 5 min at 37°C. The cells were then washed twice with phosphate-buffered saline and fixed with 4% paraformaldehyde for 10 min at 4°C. Cells were visualized within 1 week by using a Nikon Diaphot 200 inverted microscope with excitation at 360 nm and emission at 510 nm. Images were captured with a Princeton Instruments (Monmouth Junction, NJ) charge-coupled device. Nuclei were identified as normal, fragmented, or condensed. Fragmented or condensed nuclei were classified as apoptotic.
Cells were harvested into a lysis buffer consisting of 2.5 mM HEPES, pH 7.5, 10% glycerol, 5 mM EDTA, 5 mM EGTA, 100 mM NaCl, 100 mM Na pyrophosphate, 50 mM NaF, 0.1 mM NaVO4, 1% Triton X-100, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml aprotinin, 2 μg/ml pepstatin. Total protein (25 μg) was run on a 10 to 20% Tris-glycine polyacrylamide gel (Novex, San Diego, CA). The proteins were transferred to a polyvinylidene difluoride membrane (Immobilon P; Millipore, Bedford, MA). Transferred membranes were blocked for 1 h in 5% milk in Tris-buffered saline + 0.05% Tween 20. Primary antibody was added in 5% milk and the blot was incubated overnight at 4°C. The blots were washed 3 × 5 min and incubated with secondary antibody (anti-mouse horseradish peroxidase or anti-rabbit horseradish peroxidase; Santa Cruz Biotechnology Santa Cruz, CA) at 1:1000 in 5% milk. The blots were washed 3 × 5 min and exposed to the Renaissance chemiluminescent detection system (PerkinElmer Life Science Products, Boston, MA) and exposed to film. Antibodies used were PARP (1:500; Transduction Laboratories, Lexington, KY), Bcl-2 (1:200; Santa Cruz Biotechnology), Bcl-XL (1:500; Transduction Laboratories), active JNK (1:1000; New England Biolabs, Beverly, MA), JNK (1:1000; New England Biolabs), active p42/44 MAPK (1:1000; New England Biolabs), total p42/44 MAPK (1:5000, extracellular signal receptor-activated kinase-2; Santa Cruz Biotechnology), active p38 MAPK (1:1000; New England Biolabs), p38 MAPK (1:1000; New England Biolabs), and Phospho-ATF (1:1000; New England Biolabs).
This assay was performed according to the manufacturer's instructions for the P38 kinase activity kit (Cell Signaling, Beverly, MA). Briefly, Cells were treated and lysates harvested as described above. Lysates were immunoprecipitated using an agarose bead-immobilized active p38 antibody for 16 h at 4°C. The immunoprecipitate was then incubated in the presence of purified ATF and ATP in a kinase buffer (ingredients) for 30 min at room temperature. Western blotting was performed on the mixture and phosphorylated ATF was detected using a phospho-ATF antibody.
Intracellular Calcium Measurements.
Intracellular free calcium ([Ca2+]i) was quantified by fluorescence ratiometric imaging of the calcium probe fura-2 (Molecular Probes). Rat primary cortical neurons were cultured in neurobasal medium on glass coverslips (Bioptechs, Butler, PA) and loaded with 10 μM fura-2/acetoxymethyl ester (30-min incubation at 37°C). The fura-2 was dissolved in DMSO to a final [DMSO] of <0.25% in the culture medium. Cells were washed twice with fresh medium and incubated for another 30 min before experiments. The coverslip was mounted into a temperature-controlled live cell chamber system (Bioptechs) buffered in Gey's balanced salt solution (Sigma Chemical) at 37°C. Calcium imaging was carried out using a Nikon/MetaFluor software system (Universal Imaging, West Chester, PA) with a 20× objective. Drugs were administered through a pump (Instech, Plymouth Meeting, MA) connected to the closed cell chamber. The average [Ca2+]i of individual neurons was determined from the ratio of fluorescence emissions resulting from two different excitation wavelengths (340 and 380 nm). The system was calibrated using a fura-2 calcium imaging calibration kit (Molecular Probes) according to the formula [Ca2+]i −Kd[(R −Rmin)/(Rmax− R)](F380 max/F380 min). For each treatment 50 to 100 cells were chosen for statistical analysis. The results were presented as fluorescence ratio (R) of the two wavelengths.
Analysis and Statistics.
All experiments were analyzed using GraphPad Prism version 3.00 (GraphPad Software, San Diego, CA) and the one-way analysis of variance was used to determine statistical significance. Dunnett's post test was used to compare treatment groups to control if overall p < 0.05. Graphic manipulation was conducted using Photoshop 5.5 (Adobe Software, Seattle, WA).
Genistein Induces Cytotoxicity in Primary Cortical Neurons.
Genistein clearly induced a potent cytotoxicity in primary cortical neurons within 24 h of treatment (Fig.1). This toxicity appeared at 50 μM and is evidenced by a reduction in the number of green fluorescent cell bodies (living) and an increase in the number of red fluorescent nuclei (dead) (Fig. 1, B and D). Consistent with the loss of mitochondrial function, genistein treatment also increased in the release of LDH into the medium (Fig. 1C).
The structural analog daidzein (Fig. 1A) has estrogen receptor agonist properties similar to genistein (Kuiper et al., 1998) but is inactive with respect to tyrosine kinase and topoisomerase II inhibition. Daidzein did not induce cell death even at doses of 100 μM (Fig. 1, B–D), suggesting that genistein acts through tyrosine kinase and topoisomerase-mediated mechanisms and not through estrogen receptors or nonspecific intracellular binding interactions.
Genistein Induces Apoptosis in Primary Cortical Neurons.
The cell death associated with genistein toxicity is apoptotic in nature (Fig. 2). Genistein induced a DNA laddering pattern that is consistent with the intranucleosomal cleavage resulting from caspase-activated endonucleases (Fig. 2A). No fragmentation above baseline levels was seen with 100 μM daidzein, consistent with its lack of cytotoxicity. PARP cleavage was evident at 24 h in a dose-dependent manner as a result of genistein but not daidzein treatment (Fig. 2B). The accumulation of a 26-kDa cleavage product is a hallmark of specific caspase cleavage as opposed to general proteolytic degradation (Lazebnik et al., 1994). In Fig. 2C, we show that morphological changes in the cell membrane after genistein treatment are consistent with the formation of apoptotic bodies. The presence of phase-dark neurons with intact neuronal processes is nearly lost by 24 h after genistein treatment (Fig. 2C, top right). Additionally, genistein induces nuclear condensation and fragmentation (Fig. 2C, bottom right), which is consistent with an apoptotic paradigm of cell death.
Genistein-Induced Cell Death Is Independent of Estrogen Receptors and Dependent on Caspase Activity.
Figure3A shows that genistein toxicity is not affected by the estrogen receptor antagonist ICI 182,780 (1 μM), a compound that induces no toxicity alone at this concentration. Figure 3B shows that genistein toxicity is attenuated by 60% in the presence of the pan-caspase inhibitor carbobenzoxy-Asp-Glu-Val-Asp-fluoromethylketone (50 μM), a compound that causes no toxicity alone at this concentration. The caspase-3-specific inhibitor carbobenzoxy-Val-Ala-Asp-fluoromethylketone (50 μM) is unable to significantly attenuate the toxicity associated with genistein treatment.
Genistein Does Not Down-Regulate Bcl-2 or Bcl-XL over Time Course of PARP Cleavage.
A further investigation of the time course of the development of a PARP cleavage product is shown in Fig.4. Accumulation of this caspase cleavage product was first noticeable between 3 and 6 h and became robust by 24 h. In contrast, levels of immunoreactive Bcl-2 and Bcl-XL, two proteins that are often down-regulated in apoptosis, were unchanged throughout the time course of genistein-induced apoptosis.
p42/44 MAPK Is Involved in Genistein-Induced Apoptosis.
We next investigated whether members of the MAPK family of proteins are activated in these cells during the course of genistein-induced apoptosis. The dual phosphorylation of the MAPK family proteins on a threonine and tyrosine in a conserved motif has been shown to confer activity on these kinases (Anderson et al., 1990). We measured activation by immunoblotting for the dual phosphorylated form of the enzyme and by comparing the immunoreactivity of the activated state to the total immunoreactivity of the protein. We found (Fig.5) that the JNK protein does not undergo profound phosphorylation during the course of genistein-induced apoptosis. However, we do see a reduction in JNK phosphorylation at 24 h, which is concurrent with a loss in total JNK protein relative to total intracellular protein and is likely due to proteolytic cleavage. Thus, we chose not to continue investigating the role of JNK in genistein-induced apoptosis. The p38 protein, however, underwent significant activation that was maximal at 6 h after genistein treatment. In addition, p42/44 MAPK was also significantly activated, beginning at 6 h after treatment and remaining elevated through the 24-h time point (Fig. 5).
We were interested to determine whether the activation of these MAPK family proteins is necessary for genistein-induced apoptosis. To examine the role of the MAPK proteins, we used the specific inhibitors PD98059, which has been shown to block activation of mitogen-activated protein kinase kinase 1 (the direct activator of p42/44 MAPK) and SB202190, which is a specific inhibitor of p38 MAPK with little effect on other members of the MAPK family (Frantz et al., 1998). Figure6A shows that PD98059 (50 μM) blocks genistein-induced activation of p42/44 MAPK at 6 h and SB202190 (10 μM) blocks the ability of genistein to induce p38-dependent phosphorylation of the ATF substrate at 3 h. In Fig. 6B we showed that PD98059 (50 μM) was able to reduce cell death by 40%, whereas SB202190 (10 μM) had no significant effect. This result was confirmed by an examination of DNA fragmentation (Fig. 6C) and caspase-dependent PARP cleavage (Fig. 6D). Neither compound caused significant toxicity in the absence of genistein. These results suggest that although genistein is able to activate p38 MAPK, this activation is not directly involved in the apoptotic pathway. Activation of p42/44 MAPK, however, appears to be required for apoptosis and is upstream of the caspase-dependent cleavage of PARP.
Intracellular Calcium Is Involved in Genistein-Induced Apoptosis.
We determined that in response to genistein treatment (100 μM), cytosolic calcium levels are significantly elevated at 30 min and 3 h after treatment (Fig.7A). Pretreatment (30 min) with dantrolene (10 μM), an inhibitor of endoplasmic reticulum calcium release, blocks the appearance of a genistein-induced calcium increase at 30 min after treatment (Fig. 7B). We then examined the role of intracellular calcium accumulation in genistein-induced apoptosis. We used the cell membrane-permeable calcium chelator BAPTA-AM (1 μM) as well as dantrolene (10 μM) in these experiments. We found that both BAPTA-AM and dantrolene were able to significantly reduce cell death (Fig. 7C), DNA fragmentation (Fig. 7D), and caspase-dependent PARP cleavage (Fig. 7E) at concentrations that do not cause cellular toxicity alone.
Calcium and p42/44 MAPK Are Acting in Parallel to Cause Apoptosis.
To determine whether activation of the p42/44 MAPK protein and the release of intracellular calcium were occurring through dependent or parallel pathways, we pretreated cells with dantrolene (10 μM) in combination with either PD98059 (50 μM) or SB202190 (10 μM) before addition of genistein (100 μM). The application of both PD98059 and dantrolene reduced the cellular toxicity by 75% (Fig. 8A), which was additive with respect to the effects of either PD98059 or dantrolene alone on genistein toxicity (Figs. 6B and 7C). Morphologically, the cells appeared similar to vehicle-treated controls (Fig. 8B). However, the combination of SB202190 and dantrolene reduced the genistein-induced toxicity by 30% (Fig. 8A), a decrease that was not significantly greater than that induced by dantrolene alone (Fig.7C). This finding lends further support to our hypothesis that p38 MAPK activation is not involved in genistein-induced apoptosis in primary neurons. The combination of PD98059 and dantrolene reduced genistein toxicity to near baseline levels, suggesting that these inhibitors are blocking the primary intracellular apoptotic pathways stimulated by genistein.
The data presented here provide evidence that micromolar concentrations of genistein are capable of inducing apoptosis in primary cortical neurons. This finding has important ramifications for genistein's therapeutic potential as an anticarcinogenic agent and also provides insight into mechanisms of apoptosis in neuronal cells. The results demonstrate that genistein-induced apoptosis involves activation of the caspase family of proteases (not caspase-3) as well as intranucleosomal cleavage of the genomic DNA, but is not associated with down-regulation of the expression of the antiapoptotic proteins Bcl-2 and Bcl-XL. Additionally we have demonstrated that two members of the MAPK family of proteins (p38 and p42/44) are converted to the dual phosphorylated active form in response to genistein. p42/44 MAPK activation is proapoptotic in these cells, whereas p38 MAPK activation does not appear to be critical to the apoptotic pathway. Finally, we have shown that calcium release from intracellular stores is involved in genistein-induced toxicity, and blockade of both calcium release and p42/44 MAPK activation can reduce the toxicity to levels that are morphologically similar to controls. Both p42/44 MAPK and calcium release are upstream of caspase activation as well as cell death as evidenced by the ability of PD98059 and dantrolene to block both DNA fragmentation and caspase-dependent PARP cleavage. Together, these results represent an atypical pathway of neuronal cell death that lends support to the theory that the p42/44 MAPK pathway can play a proapoptotic role under certain conditions as well as providing an example of intracellular signal transduction pathway convergence in apoptosis (Fig. 9).
Bcl-2 Family Proteins.
The Bcl-2 family of proteins has been shown to play a crucial role in apoptosis, acting in either a pro- or antiapoptotic manner to regulate the release of cytochrome cfrom the mitochondria (Adams and Cory, 1998). Cytochrome crelease is an apoptotic event that leads to the loss of the mitochondrial potential, intracellular free radical damage, and the activation of caspases. This sequence of events varies depending on the apoptotic stimulus; however, Bcl-2 has been shown to protect cells from apoptosis in response to a variety of inducers (Zhong et al., 1993). In many cell lines, the induction of apoptosis reduces the protein levels of Bcl-2 or Bcl-XL, which correlates with the extent of apoptosis (Isoherranen et al., 1999; Migita et al., 1999). In particular, genistein is reported to down-regulate Bcl-2 in MDA-MB-435 cells (Li et al., 1999). Therefore, it was surprising that Bcl-2 and Bcl-XL expression were unchanged during the course of genistein-induced apoptosis. It is possible that although genistein does not alter total expression levels, it may be affecting the intracellular distribution or phosphorylation of these proteins. Future studies will address that possibility.
The phosphorylation of p42/44 MAPK in response to genistein was delayed, robust, and prolonged. This activation profile is distinct from the rapid, transient p42/44 MAPK activation induced by antiapoptotic growth factors and is consistent with the profile of activation occurring after apoptotic or necrotic stimuli in neurons as well as other cell types (Stanciu et al., 2000). PD98059, the specific inhibitor of p42/44 MAPK kinase (MAPKK1), which has little effect on other kinases, including the other members of the MAPK family (Alessi et al., 1995), was used to block activation of p42/44 MAPK. PD98059 was able to block both cytotoxicity and PARP cleavage, suggesting that MAPK activation is upstream of caspase activation and is critical for apoptosis induction. Although other groups have demonstrated p42/44 MAPK activation after apoptotic or necrotic stimuli, in many instances PD98059 is ineffective at blocking cell death (Gouni-Berthold et al., 2001), suggesting that extracellular signal receptor-activated kinase activation may play a variety of roles, depending on the cell type and toxic stimulus.
We also determined that genistein does not profoundly affect JNK phosphorylation, suggesting that this pathway is unlikely to be involved in genistein-induced apoptosis. It is possible that our detection methods were unable to reveal isoform-specific activation of JNK; however, the degree of protection conferred by PD98059 and dantrolene together suggests that any contribution of the JNK pathway to genistein-induced apoptosis is likely to be small. These results represent a growing complexity in the role for p42/44 MAPK and JNK activation in neuronal cells.
Calcium is an intracellular signal that is tightly regulated and can lead to growth, differentiation, and apoptosis, depending on the timing and extent of release. Genistein treatment leads to the elevation of intracellular calcium levels, a phenomenon that is blocked by dantrolene treatment. We found that BAPTA-AM and dantrolene were both able to block genistein-induced apoptosis and PARP cleavage. This suggests that genistein-induced release of calcium from intracellular stores is important for apoptosis induction and that it is occurring upstream of caspase activity. In addition, the combined effect on cell death of blocking intracellular calcium release by using dantrolene and blocking p42/44 MAPK activation was additive relative to the effect of each inhibitor alone. This leads us to believe that genistein is initiating multiple parallel intracellular pathways that play a critical role in apoptosis.
Proposed Mechanism of Action.
It is likely that the unusual mechanism by which genistein induces apoptosis is a result of the multiple intracellular effects of the compound (Fig.9). The lack of effect of ICI 182,780 on genistein-induced apoptosis and the lack of cytotoxicity resulting from daidzein treatment suggest that estrogen receptor-dependent mechanisms are not involved. Daidzein's lack of effect also rules out involvement of direct effects of genistein on sodium channels and γ-aminobutyric acid receptors, where daidzein has been reported to have similar activity to genistein (Paillart et al., 1997; Huang et al., 1999). This leads us to the conclusion that the tyrosine kinase and topoisomerase inhibition properties of genistein are the primary intracellular targets for apoptosis induction, although a recent report suggests that direct action on the mitochondrial permeability transition pore may be involved as well (Yoon et al., 2000). Tyrosine kinase inhibitors have been found to induce apoptosis under certain situations in cultured cells (Ohigashi et al., 2000) and may be involved in genistein-induced apoptosis of neurons.
Topoisomerase inhibition is another likely candidate for the mechanism by which genistein induces apoptosis in neurons. Etoposide and camptothecin, which inhibit topoisomerase I and II, respectively, can induce neuronal apoptosis (Nakajima et al., 1994; Morris and Geller, 1996). Additionally, reports have demonstrated the presence of active topoisomerase II in rat brain (Holden et al., 1990). The mechanism for topoisomerase inhibition-induced cell death in neurons is unknown, as is the role for topoisomerase activity in postmitotic cells. It is likely that the combination of tyrosine kinase inhibition and topoisomerase inhibition, as well as potential direct effects on mitochondria are resulting in the apoptotic pathway we have observed in primary cortical neurons.
Our findings raise questions regarding potential side effects of genistein when used therapeutically at high doses for chemotherapy. The consumption of soy has been shown to lead to plasma concentrations of genistein in the high nanomolar to low micromolar range (Watanabe et al., 1998), much lower than the doses used here. However, the plasma genistein concentrations achieved from either a high-dose oral supplement or an injected or intravenous therapy have been unstudied and it is likely that the concentration of genistein in the blood could reach levels in the high micromolar range, particularly with a nonoral route of administration. Free genistein has been found in the brain of rats consuming dietary genistein, suggesting that genistein is capable of crossing the blood-brain barrier (Chang et al., 2000). Additionally, levels of genistein in the range that we have tested (50–100 μM) seem to be required for the tumor-suppressive activities of genistein in vitro (Polkowski and Mazurek, 2000). Therefore, strategies are being investigated to boost the maximum levels of plasma genistein. Although such studies should not be discouraged, we have shown that the levels of genistein required in vitro to elicit proapoptotic effects in tumor cells are similar to doses that are toxic to cultured neurons and lead to neuronal apoptosis. Thus, the potential for neurotoxicity of genistein needs to be considered in future studies.
We thank the following people for helpful discussions and comments: Dr. Jennifer Fitzpatrick, Dr. Pamela McMillan, Dr. Amy Mize, Dr. Robert Shapiro, Dr. Nephi Stella, and Christian Wade.
This work was supported by National Institutes of Health Grant T32 GM07270 (to N.J.L.), a project in Alzheimer's Disease Research Center Grant P50AG05136, and National Institutes of Health Grant NS20311 (to D.M.D.). This work was presented at the Society for Neuroscience meeting 2000 [Soc Neurosci Abstr (2000) vol 26, abstract 780.1].
- mitogen-activated protein kinase
- c-Jun NH2-terminal kinase
- dimethyl sulfoxide
- 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid-acetoxymethyl ester
- lactate dehydrogenase
- activating transcription factor
- intracellular calcium concentration
- 1-[2-(5-carboxyoxazol-2-yl)-6-aminobenzofuran-5-oxy]-2-(2′-amino-5′-methylphenoxy)-ethane-N,N,N′,N′-tetraacetic acid
- The American Society for Pharmacology and Experimental Therapeutics