|Home | About | Journals | Submit | Contact Us | Français|
We previously showed that treatment of prostate cancer cells with soy isoflavones and radiation resulted in greater cell killing in vitro, and caused down-regulation of NF-κB and APE1/Ref-1. APE1/Ref-1 functions as a redox activator of transcription factors, including NF-κB and HIF-1α. These molecules are upregulated by radiation and implicated in radioresistance of cancer cells. We extended our studies to investigate the role of HIF-1α survival pathway and its upstream Src and STAT3 molecules in isoflavones and radiation interaction. Radiation induced phosphorylation of Src and STAT3 leading to induction of HIF-1α. Genistein, daidzein or a mixture of soy isoflavones did not activate this pathway. These data were observed both in PC-3 (AR−) and C4-2B (AR+) androgen-independent cell lines. Pre-treatment with isoflavones inhibited Src/STAT3/HIF-1α activation by radiation and nuclear translocation of HIF-1α. These findings correlated with decreased expression of APE1/Ref-1 and DNA binding activity of HIF-1α and NF-κB. In APE1/Ref-1 cDNA transfected cells, radiation caused a greater increase in HIF-1α and NF-κB activities but this effect was inhibited by pre-treatment with soy prior to radiation. Transfection experiments indicate that APE1/Ref-1 inhibition by isoflavones impairs the radiation-induced transcription activity of NF-κB and HIF-1α. This mechanism could result in inhibition of genes essential for tumor growth and angiogenesis, as demonstrated by inhibition of VEGF production and HUVECs tube formation. Our novel findings suggest that the increased responsiveness to radiation mediated by soy isoflavones could be due to pleiotropic effects of isoflavones blocking cell survival pathways induced by radiation including Src/STAT3/HIF-1α, APE1/Ref-1 and NF-κB.
In recent years, many dietary compounds, have been recognized as cancer chemopreventive agents and found to exert their anti-tumor activity by inhibiting signal transduction pathways essential for tumor growth, invasion and metastasis1. We and others have shown that dietary agents can be exploited to improve the therapeutic efficacy of radiotherapy and chemotherapy for cancer 2, 3. We clearly showed that soy isoflavones act as potent radiosensitizers in prostate cancer (PCa) 4–9. In the United States, PCa is the most commonly diagnosed cancer in men as well as the second leading cause of male cancer deaths. The American Cancer Society estimated that 218,890 new cases of PCa will be diagnosed in 2007 and 27,050 men will die of the disease10, 11. Localized PCa is sensitive to conventional radiotherapy using megavoltage photons (X-rays), yet the disease was not eradicated in a significant proportion of patients resulting in clinical recurrence and progression of PCa 12. Epidemiological studies indicate that there is an inverse association between prostate cancer risk and consumption of phytoestrogens, mainly soy isoflavones 13. The isoflavones in soybeans mainly include genistein, daidzein and glycitein. Genistein, the most active biological compound of soy, inhibited the growth of human PCa cells in vitro by affecting the cell cycle and inducing apoptosis 14. To improve the therapeutic efficacy of radiotherapy for PCa, we have previously demonstrated that, when given prior to radiation, pure genistein potentiates radiation-induced tumor cell killing of human PC-3 PCa cell line in vitro 4, 6. In vivo, using a metastatic orthotopic PC-3 xenograft tumor model in nude mice, we showed that treatment with genistein combined with prostate tumor irradiation significantly enhanced inhibition of prostate tumor growth and increased mouse survival 5, 8. However, we found that pure genistein administered alone caused increased spontaneous metastasis to regional lymph nodes5. To investigate further this intriguing finding, we have compared a mixture of soy isoflavones (genistein, daidzein and glycitein) to pure genistein and found that soy isoflavones mixture had comparable activity than pure genistein causing cancer cell apoptosis and affecting signal transduction molecules in vitro 8, 9. Furthermore, when used to treat orthotopic prostate tumors in mice, the soy isoflavones mixture did not cause increased spontaneous metastasis, in contrast to pure genistein, and potentiated radiation therapeutic effect8. In the current study, we have further investigated the effect of the soy isoflavones mixture on cell survival molecular pathways and also determined the activity of two major isoflavones genistein and daidzein.
The molecular mechanism of interaction between soy isoflavones and radiation remains to be clarified. In previous studies, we have identified NF-κB and APE1/Ref-1 as two potential molecular targets which cross-talk could be involved in increased cell killing of PC-3 cells by soy isoflavones and radiation 6, 9. Both molecules are important signaling molecules involved in cell death or survival pathways. NF-κB correlates with tumor progression and is a major transcription factor involved in the synthesis of critical cell survival proteins in response to cellular stress, including radiation 15,16, 17. Apurinic/apyrimidinic (AP) endonuclease 1/redox factor-1 (APE1/Ref-1) is a multifunctional protein involved in DNA repair that also functions as a redox activator of cellular transcription factors, including NF-κB3, 18–20. Both NF-κB and APE1/Ref-1 were expressed by PC-3 cells and were activated and upregulated by radiation but were inhibited by soy isoflavones in vitro and in vivo 6, 9. Furthermore, we showed that pre-treatment of PCa cells with soy isoflavones completely inhibited activation of NF-κB DNA binding activity and upregulation of APE1/Ref-1 induced by radiation, in vitro and in PC-3 prostate tumors in vivo 6, 9. APE1/Ref-1 down regulation by soy isoflavones could affect additional signaling pathways activated by radiation.
APE1/Ref-1 is also responsible for redox-activation of the hypoxia-inducible factor 1 (HIF-1). HIF-1, consists of α and β subunits and is a major transcription factor involved in cell response to hypoxia21. Cellular levels of HIF-1α and APE1/Ref-1 redox stabilization of the HIF-1α protein are critical for its nuclear translocation and DNA binding and transcriptional activity.22, 23. Hypoxia is a common feature of many cancers that contributes to local and systemic tumor progression and compromises radiotherapy and chemotherapy 24, 25. In the PC-3 animal tumor model, focal HIF-1α expression was observed in radiation-treated prostate tumors but was minimal after treatment with soy isoflavones and radiation 8. Based on these in situ observations, the goals of the current study were to investigate the effect of soy isoflavones on HIF-1α. We also investigated the effect of soy isoflavones on signaling events upstream of HIF-1α. The c-Src is a proto-oncogene, a non receptor protein tyrosine kinase that acts upstream of STAT3 and HIF-1α and has been associated with cancer cell growth, migration, invasion and angiogenesis 26–29. Inhibition of STAT3 altered tumor growth and angiogenesis by blocking HIF-1α expression and transcription of its downstream gene vascular endothelial growth factor (VEGF), thus establishing STAT3 as an upstream molecule of HIF-1α 30. Previous studies have demonstrated that genistein caused inactivation of Src 31. In this study, we investigated whether soy isoflavones alter the Src/STAT3/HIF-1α signaling pathway, which could contribute to the mechanism of increased response of PCa to radiation. Our data show that radiation induced phosphorylation of Src and STAT3, and caused a significant induction of HIF-1α protein whereas soy isoflavones did not activate this pathway. Interestingly, we found that pre-treatment of PCa cells with soy isoflavones inhibited radiation-induced activation of the Src/STAT3/HIF-1α signaling pathway and these results correlated with decreased expression of APE1/Ref-1 and decreased DNA binding activity of NF-κB and HIF-1α.
PC-3 human PCa cell line (AR- and non responsive to androgen) was cultured in F-12K culture medium (CM) containing 7% heat-inactivated fetal bovine serum (FBS) with supplements 4, 6. The C4-2B human PCa cell line (AR+ and non responsive to androgen) was cultured in RPMI 1640 CM containing 10% heat-inactivated FBS with supplements 32, 33.
The soy isoflavones mixture G2535 consists of 43% genistein, 21% daidzein, 2% glycitein, 2.5 % protein, 11.9% fat, 1.7% water, with the remainder being carbohydrate (manufactured by Organic Technologies and obtained from NIH) 8, 9. Soy isoflavones, pure genistein (LKT Laboratories) and pure daidzein (Biomol International) powders were dissolved in 0.1 mol/L Na2CO3 and were further diluted in CM to obtain final concentrations of 30–60 μmol/L 8, 9. Control cells were incubated with equivalent dilutions of Na2CO3 in CM. Cells were treated when ~75% confluent.
For Western blot analysis, cells were washed with phosphate buffered saline (PBS) and scraped. For protein extraction, cell lysis was done using modified RIPA buffer 34. Protein samples (20–40 μg) were loaded and separated on 7.5%, 10% or 12% SDS-PAGE gel and transferred to Trans-Blot membranes (Bio-Rad, Hercules, CA). Membranes were immunoblotted with primary antibodies (Ab) directed against HIF-1α (R&D systems, Minneapolis, MN), phospho-Src (Tyr 416), total Src, phospho-STAT3 (Tyr 705) and total STAT3 (Cell Signaling Technology, Danvers, MA). Monoclonal anti-APE1/Ref-1 Ab was obtained from Novus Biologicals (Littleton, CO). Antibodies were diluted in a range of 1:500 to 1:2500 6, 9. Membranes were incubated in IgG-HRP secondary Abs. Membranes were re-probed with anti-β-actin Ab (Santa Cruz, CA) as an internal control. Membranes were incubated in SuperSignal® West Pico Chemiluminescent Substrate (Pierce, Rockford, IL), and analysed in Bio-Rad ChemiDoc XRS imaging system. AlphaEaseFC™ imaging software (AlphaInnotech, San Leandro, CA) was used to quantify resultant bands 6, 9.
PC-3 cells were cultured onto cover slips in six-well plates and treated with 30 μM soy isoflavones for 48 hrs, then irradiated with 3 Gy. Three hours after radiation, cells were fixed with 4% formaldehyde, permeabilized with 0.5% Triton X100 and incubated overnight at 4°C with anti-HIF-1α Abs (Exalpha Biologicals). After washing with PBS, cells were exposed for 2 hrs to FITC-conjugated secondary goat anti-mouse IgG (Alexa Fluor 488, Invitrogen, Carlsbad, CA) prepared in DAPI and 0.05% Triton X 100 (8). Stained cells were analyzed using NIKON E-800 UV microscope for detection of DAPI and FITC staining.
Nuclear extracts were prepared from treated cells using Sigma CelLytic™ NuCLEAR™ Extraction Kit (Sigma-Aldrich, St. Louis, MO) 6, 9. Electrophoretic mobility shift assay (EMSA) was accomplished as previously detailed 9, 35. Briefly, 5 μg of nuclear proteins were incubated with IRDye-700 labeled NF-κB oligonucleotide or IRDye-700 labeled HIF-1α oligonucleotide (LI-COR Biosciences, Lincoln, NE) 9. The DNA-protein complex was run on 8.0% native polyacrylamide gel using TBE buffer, pH 8.4. The bands were visualized by Odyssey Infrared Imaging System using Odyssey Software Release 1.1. Equal protein loading was ensured by immunoblotting 10 μg of nuclear protein with retinoblastoma antibody.
For APE1/Ref-1 over-expression, PC-3 cells were transfected either with empty plasmid pCMV6-XL5 (pCMV6) or with the plasmid containing APEX nuclease 1 (pCMV6.APE1) (OriGene, Rockville, MD). PC-3 cells were transfected with the cDNA plasmid vectors (5 μg DNA) as previously described 9. After 6 hrs incubation, medium was replaced and cells were further incubated for 24 hrs, then, treated with 60 μM soy for 24 hrs followed by 3 Gy radiation9. Whole cell lysates or nuclear extracts were used in Western blots. HIF-1α and NF-κB EMSA were performed using nuclear extracts. For APE1/Ref-1 under-expression, PC-3 cells were transfected twice with 200pmole of validated siRNA against APE1/Ref-1 (Ambion, Austin TX). After 48 h siRNA incubation, cells were treated with 60 μM soy for 24h and nuclear extracts were used in Western blot or EMSA. GAPDH siRNA was used as a negative control.
PC-3 cells cultured for 24hrs, in serum-free RPMI 1640, were treated with 30 μM soy isoflavones in 1% FBS for 24 h, then irradiated with 3 Gy. At 6hrs after radiation, the conditioned media were collected and stored at −20°C 36. VEGF secretion was determined using VEGF ELISA kit (R&D System) as previously described 36. The conditioned media were also used to test for human umbilical vascular endothelial cells (HUVECs) tube formation in a Matrigel in vitro assay as detailed previously 36. Each well was photographed using an inverted microscope with digital camera.
Comparisons between means of various treatment groups in Western blot and EMSA assays were analyzed by two-tailed unpaired Student’s t test. A p-value less than 0.05 was considered statistically significant 9.
To further explore the molecular pathways involved in the mechanism of interaction between radiation and soy isoflavones, we have investigated the modulation of the transcription factor HIF-1α and its upstream signaling molecules Src and STAT3 by radiation in PC-3 cells. Cells were irradiated with 3 Gy photons and processed for whole cell protein extraction at 1, 3, 5 and 24 hr after radiation. Protein extracts were used in Western blots to analyze the expression of Src and STAT3 molecules and their respective phosphorylated molecules as well as to detect the expression of HIF-1α and APE1/Ref-1. Src phosphorylation (at Tyr 416) was induced by radiation and increased within 1–5 hrs of radiation with a slight decrease by 24 hrs confirming that it is an early event in response to radiation (Fig. 1A). This effect was stronger with a higher dose of 5 Gy (data not shown). Radiation caused significant phosphorylation of STAT3 (at Tyr 705) measurable at 1–24 hr after radiation (Fig. 1A). Radiation induced significant HIF-1α expression at 5hr after radiation and higher levels were observed at 24 hrs (Fig. 1A). Radiation caused upregulation of APE1/Ref-1 detectable by 1 hr after radiation and persisting up to 24 hrs after radiation (Fig. 1A), as previously demonstrated 9.
To investigate the role of HIF-1α in soy isoflavones mediated inhibition of PC-3 cells, hypoxic conditions were created in vitro by treatment of PC-3 cells with 100 μM of cobalt chloride CoCl2 for 3hrs 37. Under such conditions, we observed a significant induction of HIF-1α (Fig. 1B, lane 2). Treatment of PC-3 cells for 24 hrs with 60 μM pure genistein or 60 μM soy isoflavones did not induce HIF-1α in vitro (Fig. 1B, lanes 3, 4). Furthermore, pre-treatment of cells with soy isoflavones for 24 hrs prior to CoCl2 induced hypoxia, inhibited subsequent HIF-1α induction by CoCl2 (Fig. 1B, lanes 5, 6). The effect of soy isoflavones on HIF-1α upstream molecule STAT3 was compared to radiation effect. Based on experimental conditions determined in experiments presented in Fig. 1A, PC-3 cells were treated with 60 μM pure genistein or 60 μM soy isoflavones or with 3Gy photon radiation and incubated for 24 hrs. As shown above in Fig. 1A, radiation caused phosphorylation of STAT3 and induction of HIF-1α whereas genistein or soy isoflavones did not (Fig. 1C). Genistein or soy isoflavones reproducibly decreased APE1/Ref-1 expression (Fig. 1C) as shown previously 9.
The bioactive components of the soybeans are the three isoflavones genistein, daidzein, glycitein. Several studies have shown that genistein is the major and most bioactive compound of soy but daidzein is also a significant isoflavone and has shown anti-cancer activity whereas glycitein is a minor component2, 14, 38, 39. To investigate whether daidzein plays a role in the effect of soy isoflavones on STAT3/HIF-1α pathway, PC-3 cells were treated with 30μM of pure daidzein or pure genistein or the mixture of soy isoflavones for 72 hrs. For this experiment, we selected milder doses of isoflavones and longer incubation times to detect differences in molecular responses between each treatment. The phosphorylation of STAT 3 and HIF-1α expression were inhibited by daidzein, genistein and soy isoflavones at comparable levels (Fig. 1D). Inhibition of APE1/Ref-1 also occurred with daidzein but at a lower level than with genistein and was more pronounced with the mixture of soy isoflavones than with each isoflavone (Fig. 1D).
It should be noted that there are variations between experiments on the basic levels of HIF-1α in control cells, which are due to incubation time and the status of the cells. Nevertheless, our main findings demonstrating a significant increase of HIF-1α with radiation and no increase with soy isoflavones are very consistently reproduced (Fig. 1A-D).
To investigate whether the effects of soy isoflavones and radiation observed in the PC-3 PCa cell line, that is AR- and androgen-independent, were reproduced in another PCa cell line with a phenotype of advanced PCa, we tested the C4-2B cell line that is AR+ but androgen-independent 32, 33. C4-2B cells were treated with either with 3 Gy radiation for 1 and 3 hrs or with 30μM of pure daidzein or pure genistein or the mixture of soy isoflavones for 72 hrs, and then tested for HIF-1α and APE1/Ref-1 expressions. As observed with PC-3 cells, HIF-1α expression was significantly increased at 3 hrs after radiation (Fig. 1E). Like PC-3 cells, C4-2B showed a high constitutive APE1/Ref-1expression that was significantly increased by radiation within 3 hrs (Fig. 1E). Genistein caused a more significant inhibition of HIF-1α and APE1/Ref-1 expressions than daidzein and the mixture of soy isoflavones was more effective than either isoflavone alone (Fig. 1F).
The effect of pure genistein and soy isoflavones combined with radiation on the Src/STAT3/HIF-1α signaling pathway was tested. PC-3 cells were pre-treated with 30 μM genistein or 30 μM soy for 72 hr then treated with 3 Gy radiation and processed for Western blot analysis at 5 hr post-radiation. A lower but biologically effective dose of genistein or soy isoflavones (30 μM versus 60 μM in previous experiments) was used because of a longer pre-treatment of 72 hr prior to radiation, as determined previously 9. Treatment of PC-3 cells with CoCl2 was used as internal control for simulating hypoxic conditions and resulted in significant phosphorylation of Src and STAT3 associated with significant induction of HIF-1α expression (Fig. 2). Radiation of PC-3 cells also caused significant phosphorylation of Src and STAT3, and significant induction of HIF-1α compared to control cells, whereas these molecules were not activated by soy isoflavones (Fig. 2). Interestingly, pre-treatment with pure genistein or soy isoflavones significantly inhibited radiation-induced phosphorylation of Src and STAT3 (Fig. 2). Similar to isoflavones inhibition of CoCl2-induced HIF-1α expression, isoflavones pre-treatment of PC-3 cells also significantly inhibited radiation-induced expression of HIF-1α (Fig. 2), concomitant with a decrease in APE1/Ref-1 protein expression, as previously demonstrated 9.
Immunofluorescence analysis of HIF-1α cellular expression and localization revealed that untreated PC-3 cells showed minimal expression of HIF-1α in the nucleus and some expression in cytoplasm (Fig. 3A). Following radiation, an increase in HIF-1α cell staining was observed in the cytoplasm and was particularly intense in the nuclei of the cells (Fig. 3B) whereas cells treated with soy isoflavones showed only cytoplasmic staining and minimal nuclear staining (Fig. 3C). When cells were pre-treated with soy then irradiated, cells showed much lower levels of HIF-1α nuclear staining compared to the high intensity staining seen in cells treated with radiation only (Fig. 3D).
We tested the effect of soy isoflavones and radiation on HIF-1α DNA binding activity to assess HIF-1α function as a transcription factor. PC-3 cells were pre-treated with 30 μM genistein or 30 μM of soy isoflavones for 72 hr followed by 3 Gy photon radiation or with each modality separately. At 3hr post-radiation, cells were processed for extraction of nuclear proteins, which were used to determine HIF-1α DNA binding activity by EMSA. HIF-1α DNA binding activity was induced by radiation but not by genistein or soy isoflavones (Fig. 4A). Pre-treatment of PC-3 cells with pure genistein or soy isoflavones significantly inhibited radiation-induced HIF-1α DNA binding activity (Fig. 4A). As an internal control for simulating hypoxic conditions, treatment with CoCl2 caused an increase in HIF-1α DNA binding activity similar to that observed with radiation (Fig. 4A). To determine the role of APE1/Ref-1 in regulating HIF-1α transcription activity, PC-3 cells were transfected with APE1/Ref-1 cDNA plasmid (pCMV6.APE1), then treated with soy isoflavones for 24hrs followed by 3 Gy radiation. At 3 hrs after radiation, cells were tested for APE1/Ref-1 expression and HIF-1α DNA binding activity. Transfection of PC-3 cells with empty pCMV6 vector did not affect APE1/Ref-1 expression compared to control non-transfected PC-3 cells. The effects of soy isoflavones and/or radiation on PC-3 cells transfected with empty pCMV6 vector paralleled those observed with non-transfected PC-3 cells 9, including an increase in nuclear APE1/Ref-1 expression induced by radiation and a marked decrease caused by soy isoflavones alone and combined with radiation (Fig. 4B). In transfection experiments with pCMV6.APE1, aimed at transient expression of the APE1 gene within 24 hrs, transfected PC-3 cells showed a significant 1.5-fold increase in nuclear expression of APE1/Ref-1 compared to non-transfected control cells, and to cells transfected with empty pCMV6 vector (Fig. 4B). Treatment of PC-3/APE1 transfected PC-3 cells with soy isoflavones caused a significant 50% decrease in APE1/Ref-1 expression (Fig. 4B), as shown previously 9. In the current study, we also evaluated the effect of radiation on APE1/Ref-1 over-expressing PC-3 cells and found that radiation caused a further 30% increase in APE1/Ref-1 expression, however, that increase was markedly inhibited by pre-treatment with soy prior to radiation (Fig. 4B). HIF-1α DNA binding activity was induced by APE1/Ref-1 transfection, compared to pCMV6 transfected cells, but reduced by soy treatment (Fig. 4B). A further significant 1.6-fold increase in HIF-1α activity was observed with radiation in over-expressing PC-3 cells and correlated with APE1/Ref-1 increase in these cells. Soy isoflavones pre-treatment prior to radiation significantly inhibited activation of HIF-1α DNA binding activity by radiation concomitant with decreased expression of APE1/Ref-1.
Nuclear and cytosolic expressions of HIF-1α protein were also tested in APE1/Ref-1 over-expressing PC-3 cells and showed induction of expression both in the cytosol and nuclear fractions (Fig. 4C). Interestingly, following radiation, HIF-1α expression seems to be more prominent in the nuclear than in the cytosolic fractions (Fig. 4C). Cells treated with soy alone or combined with radiation showed more HIF-1α expression in the cytosolic fraction then in the nuclear fraction in correlation with decreased APE1/Ref-1 expression suggesting that soy isoflavones could inhibit nuclear translocation of HIF-1α protein upregulated by radiation and/or by APE1/Ref-1 over-expression (Fig. 4C). These data corroborate our microscopic localization findings of minimal HIF-1α staining in the nucleus of cells treated with soy or soy combined with radiation in contrast to the intense HIF-1α nuclear staining caused by radiation treatment alone (Fig. 3B)
To elucidate further the role of APE1/Ref-1 and its cross-talk with NF-κB in the interaction between radiation and soy isoflavones, we tested both modalities in APE1/Ref-1 transfected PC-3 cells, using identical experiments as those described above for HIF-1α activity. In control pCMV6 transfected cells, NF-κB DNA binding activity was increased by radiation but significantly decreased by soy combined with radiation (Fig. 5A), as shown with non-transfected cells 9. In pCMV6.APE1 transfected cells, a significant increase in NF-κB DNA binding activity was observed compared to pCMV6 transfected cells, which correlated with increased expression of APE1/Ref-1 (Fig. 5A). This activity remained strong following radiation but was significantly decreased by soy alone or combined with radiation concomitant with decreased APE1/Ref-1 (Fig. 5A). To confirm the interaction and molecular cross-talk between APE1/Ref-1 and NF-κB, PC-3 cells were transfected with siRNA against APE1/Ref-1 to induce transient knockdown of the APE1 gene within 48 hrs, then treated with soy for 24 hrs. APE1.siRNA transfected cells showed a significant 50% decrease of APE1/Ref-1 nuclear expression and decreased NF-κB DNA binding activity (Fig. 5B). Soy treatment of APE1-siRNA transfected cells caused a further and significant 50% reduction in APE1/Ref-1 expression that was associated with an additional 50% decrease in NF-κB activity compared to untreated APE1-siRNA cells (Fig. 5B). This decrease was more marked than that observed in control siRNA transfected cells treated with soy. Control siRNA did not affect APE1/Ref-1 nuclear expression.
We have shown that soy isoflavones inhibited the increase in DNA binding activity of HIF-1α and NF-κB that was induced by radiation (Fig. 4,,5).5). Both transcription factors were found to be involved in the transcription of VEGF gene, therefore we investigated the effect of soy isoflavones on VEGF secretion by PC-3 cells. Conditioned media from cells treated with 30μM soy for 24 hrs then with 3Gy radiation were harvested, cleared of cellular debris by centrifugation then tested in VEGF ELISA assay. Compared to untreated cells, PC-3 cells treated with soy isoflavones showed a significant 80% inhibition in the VEGF levels measured in the media (p<0.05, Fig. 6A). The effect of a low dose of 3 Gy radiation alone measured at 3 hrs after cell irradiation was comparable to that of control cells, in the in vitro conditions of this experiment (Fig. 6A). Cells pre-treated with soy and then irradiated showed also 80% inhibition of VEGF secretion, akin to soy treatment alone, probably due to the effect of soy on the cells blocking VEGF production (Fig. 6A). As an indirect measure of angiogenesis, conditioned media from the same experiment were also tested for induction of HUVECs tube formation in a matrigel in vitro assay. Tube formation was observed in PC-3 untreated control cells and in radiation-treated cells whereas tube formation was disrupted by treatment with soy or soy combined with radiation (Fig. 6B).
Two important signaling molecules APE1/Ref-1 and NF-κB, which are activated by radiation and involved in cell survival pathways, are inhibited by soy isoflavones both in vitro and in vivo, resulting in increased cell killing when cells are treated with soy and radiation 3, 9. Downregulation of APE1/Ref-1 expression by soy isoflavones could alter its function as a redox activator of NF-κB, and block activation of radiation-induced NF-κB dependent survival pathways 9. Consistent with our findings, other studies emphasize that selective targeting of APE1/Ref-1 could enhance radiotherapy 40–42. Targeting APE1/Ref-1 with soy isoflavones could also suppress the activities of other transcription factors, including HIF-1α that is dependent on APE1/Ref-1 redox activation 42. APE1/Ref-1, NF-κB and HIF-1α molecules are upregulated and or activated by radiation and they have been implicated in radioresistance of cancer cells 3, 9, 24, 25, 42–44. Recent studies have established activation of Src and STAT3 as upstream signaling events leading to induction of HIF-1α 30, 45, 46. Our results showed that radiation of PC-3 cells induced phosphorylation of Src (Tyr 416) and STAT3 (Tyr 705) and a progressive increase in HIF-1α protein expression. This activation of Src/STAT3/HIF-1α pathway by radiation correlated with an increase in APE1/Ref-1 protein expression. In contrast to radiation, soy isoflavones neither cause phosphorylation of Src and STAT3 nor induction of HIF-1α protein in PC-3 cells. In the current study, we have compared the effect of a mixture of soy isoflavones to pure genistein and pure daidzein. Several studies have shown that genistein is the major compound of soy and the most bioactive 2, 14. Daidzein is also a significant isoflavone in the soybean and has shown anti-cancer activity whereas glycitein is a minor component38, 39. We found that daidzein inhibited activation of STAT3 and decreased expression of HIF-1α and APE1/Ref-1 but to a lower extent than pure genistein and often the mixture of soy isoflavones showed greater activity than each separate isoflavone in PC-3 cells. The inhibitory effects of isoflavones on HIF-1 α and APE1/Ref-1 were also reproduced with another PCa cell line, the C4-2B cells that are AR+ and androgen-independent, a phenotype more representative of advanced refractory PCa 32, 33. Radiation induced HIF-1 α and upregulated APE1/Ref-1 in C4-2B cells, akin to its effects on the AR- PC-3 cells. These data indicate that modulation of these molecules by soy isoflavones or radiation occurs independently of the status of AR.
Interestingly, soy isoflavones inhibited radiation-induced HIF-1α confirming that soy isoflavones inhibit induction of HIF-1α under conditions of hypoxia caused by oxidative stress. Studies on cellular localization of HIF-1α, by Western blot analysis and by microscopic analysis of immunofluorescent stained cells, both suggest that soy isoflavones could inhibit nuclear translocation of HIF-1α protein that is upregulated by radiation (Fig. 3, ,4C).4C). These in vitro data correlated with our in vivo findings showing decreased HIF-1α expression in murine PC-3 prostate tumors treated with soy isoflavones and radiation 8. In agreement with our findings, recent studies have revealed increased tumor radiosensitivity by decreasing levels of HIF-1α 47, 48. Furthermore, our current studies show that soy isoflavones-mediated inhibition of HIF-1α induction by radiation could also involve inhibition of signaling events that are upstream of HIF-1α. Pre-treatment with soy isoflavones prior to radiation also significantly inhibited radiation-induced activation of phosphorylation of Src and STAT3. These data were reproducible with either pure genistein or a mixture of soy isoflavones confirming the biological activity of genistein. These findings demonstrate that soy isoflavones inhibit the Src/STAT3/HIF-1α signal transduction pathway that is activated in response to radiation. The signal transduction pathway involving Src/STAT3/HIF-1α is critical for cancer growth and inhibition of Src-STAT3 signaling induces growth arrest and apoptosis of tumor cells in vitro and in vivo 26–30. Inhibition of radiation-induced phosphorylation of Src and STAT3 by soy isoflavones was correlated with decreased levels of HIF-1α protein, corroborating previous studies showing that Src and STAT3 are upstream molecules of HIF-1α 30, 45, 46.
Associated with the Src/STAT3/HIF-1α signaling pathway, a molecular cross-talk between APE1/Ref-1 and HIF-1α was found to be essential for HIF-1α to fulfill its role as a transcription factor of genes involved in tumor progression. Nuclear translocation of HIF-1α was found to be dependent on APE1/Ref-1 redox stabilization of the HIF-1α protein and was shown to be a critical step in the hypoxic response 23, 42. In the current study, we demonstrated that radiation not only increases the expression of HIF-1α protein but also increased HIF-1α DNA binding activity, and was concomitant with increased APE1/Ref-1 expression, as observed previously with NF-κB activity 9. Soy isoflavones blocked radiation-induced HIF-1α and NF-κB DNA binding activities, probably interfering with transcription of downstream genes important for tumor growth, especially VEGF. Interestingly, both NF-κB and HIF-1α molecules are upregulated by radiation and are involved in the transcriptional regulation of VEGF 49. We found that in contrast to radiation, soy isoflavones treatment alone or in combination with radiation caused a significant inhibition of VEGF production in PC-3 cells and inhibition of angiogenesis as measured by HUVECs tube formation. These data suggest that soy isoflavones downregulate transcription of VEGF and, in turn, inhibit angiogenesis. Simultaneous downregulation of NF-κB and HIF-1α in PCa cells by inhibition of APE1/Ref-1 with soy isoflavones could decrease both cell survival and VEGF-mediated angiogenesis, and thereby enhance tumor radiosensitivity.
This conclusion is validated by our transfection studies using APE1/Ref-1 over-expressing PC-3 cells, in which radiation caused further APE1/Ref-1 increase resulting in a significant increase in HIF-1α DNA binding activity whereas both effects were inhibited by soy pre-treatment prior to radiation. Interestingly, soy seemed to inhibit the translocation of HIF-1α protein, an effect shown to be dependent on redox activation mediated by APE1/Ref-1 23, 42, and our results correlated with soy-induced decreased levels of APE1/Ref-1. Soy isoflavones also consistently inhibited radiation-induced increase in NF-κB DNA binding activity both in PC3 cells and in APE1/Ref-1 over-expressed PC-3 cells. Moreover, we now showed that APE1-siRNA transfection of PC-3 cells caused decreased NF-κB activity associated with decreased APE1/Ref-1 while in these cells, soy further decreased both NF-κB activity and APE1/Ref-1. These findings confirm a cross-talk between APE1/Ref-1, NF-κB and HIF-1α and indicate a critical role for APE1/Ref-1 in the mechanism of interaction between soy isoflavones and radiation that results in the inhibition of NF-κB and HIF-1α transcription of genes essential for tumor cell survival, tumor growth and angiogenesis. This novel approach using safe dietary agents to enhance the efficacy of radiotherapy is promising for the design of new clinical strategies for the treatment of prostate cancer.
This study was supported in part by the American Cancer Society (grant ROG-06-097-01 to G.G.H.), the Department of Defense (DMAD17-03-1-0042 to FHS) and National Institutes of Health (5R01CA108535-05 to FHS).
We thank Kristin Dominiak and Amit Patel for excellent technical assistance.
Novelty and Impact: We report original and novel findings demonstrating that soy isoflavones inhibit the activation of the Src/STAT3 signaling pathway by radiation and radiation-induced HIF-1α expression and activity contributing to increased response of cancer cells to radiation. We showed that these findings correlated with decreased expression of APE1/Ref-1 resulting in decreased DNA binding activity of HIF-1α and NF-κB, thereby inhibiting transcription of downstream genes essential for tumor growth and angiogenesis.