The present study shows that oral gavage of 6 μmol d,l-sulforaphane (SFN), a synthetic analogue of cruciferous vegetable-derived L isomer, thrice per week beginning at 6 weeks of age, significantly inhibits prostate carcinogenesis and pulmonary metastasis in TRAMP mice without causing any side effects. The incidence of the prostatic intraepithelial neoplasia and well-differentiated (WD) carcinoma were ∼23% to 28% lower (P < 0.05 compared with control by Mann-Whitney test) in the dorsolateral prostate (DLP) of SFN-treated mice compared with controls, which was not due to the suppression of T-antigen expression. The area occupied by the WD carcinoma was also ∼44% lower in the DLP of SFN-treated mice relative to that of control mice (P = 0.0011 by Mann Whitney test). Strikingly, the SFN-treated mice exhibited ∼50% and 63% decrease, respectively, in pulmonary metastasis incidence and multiplicity compared with control mice (P < 0.05 by t test). The DLP from SFN-treated mice showed decreased cellular proliferation and increased apoptosis when compared with that from control mice. Additionally, SFN administration enhanced cytotoxicity of cocultures of natural killer (NK) cells and dendritic cells (DC) against TRAMP-C1 target cells, which correlated with infiltration of T cells in the neoplastic lesions and increased levels of interleukin-12 production by the DC. In conclusion, the results of the present study indicate that SFN administration inhibits prostate cancer progression and pulmonary metastasis in TRAMP mice by reducing cell proliferation and augmenting NK cell lytic activity. [Cancer Res 2009;69(5):2117–25]
Epidemiologic studies continue to support the premise that dietary intake of cruciferous vegetables may be protective against different types of malignancies, including prostate cancer (1–3). The anticarcinogenic effect of cruciferous vegetables is attributed to organic isothiocyanates (ITC), which are generated via myrosinase-mediated hydrolysis of corresponding glucosinolates (4, 5). The ITC compound (−)-1-isothiocyanato-(4R)-(methylsulfinyl)-butane (L-SFN) and its synthetic analogue d,l-sulforaphane (SFN) have generated a great deal of research interest because of their remarkable anticancer effects (4–7). For example, L-SFN and SFN are equipotent inducers of quinone reductase in Hepa1c1c7 murine hepatoma cells (6). The L-SFN or SFN administration has been shown to offer protection against 9,10-dimethyl-1,2-benzanthracene–induced mammary cancer in rats, azoxymethane-induced colonic aberrant crypt foci in rats, and benzo(a)pyrene–induced forestomach cancer in mice (8–10). Dietary feeding of SFN and its N-acetylcysteine conjugate inhibited malignant progression of lung adenomas induced by tobacco carcinogens in A/J mice (11).
Recent studies, including those from our laboratory, have documented novel cellular responses to SFN treatment in cultured human cancer cells, including growth arrest, induction of apoptosis and autophagy, inhibition of histone deacetylase, binding to proteins, and sensitization of cells to TRAIL-induced apoptosis (12–21). The mechanism by which SFN causes growth arrest and apoptosis induction has been extensively studied in human prostate cancer cells (13, 14, 16–18, 22). For example, we have shown previously that exposure of human prostate cancer cells to SFN results in G2-M phase cell cycle arrest via checkpoint kinase 2–mediated phosphorylation of cell division cycle 25C (13). We also showed that the SFN-induced apoptosis was selective toward prostate cancer cells and linked to generation of reactive oxygen species (15, 16). Moreover, oral gavage of SFN in male nude mice significantly retarded the growth of PC-3 human prostate cancer xenografts in vivo (22).
Demonstration of in vivo efficacy of potential cancer chemopreventive agents in suitable animal models is a prerequisite for their further clinical development. The present study was designed to test whether SFN administration offers protection against prostate carcinogenesis in a transgenic mouse model of prostate cancer, TRAMP (transgenic adenocarcinoma of mouse prostate). We now show that oral gavage of SFN significantly prevents prostate cancer progression and pulmonary metastasis in TRAMP mice without causing any side effects or affecting the T-antigen expression.
Materials and Methods
Synthesis of SFN. SFN was synthesized as described by Conaway and colleagues (11). The structure of SFN was confirmed by proton nuclear magnetic resonance and mass spectrometry. The purity of the SFN (>99% pure) preparation was determined by high-performance liquid chromatography (HPLC), as described by Conaway and colleagues (11). SFN was stored at −20°C and found to be stable for at least 6 mo.
Cell viability and apoptosis assays. Monolayer cultures of TRAMP-C1 cells, a generous gift from Dr. Barbara Foster (Roswell Park Cancer Institute), were maintained in DMEM supplemented with 10% heat-inactivated fetal bovine serum, 10−8 mol/L 5α-androstan-17β-ol-3-one, and antibiotics in a humidified atmosphere of 95% air and 5% CO2. Stock solution of SFN was prepared in DMSO, and an equal volume of DMSO (final concentration, <0.1%) was added to controls. Effect of SFN treatment on TRAMP-C1 cell viability was determined by trypan blue dye exclusion assay, as described by us previously (23). The proapoptotic effect of SFN was assessed by the analysis of cytoplasmic histone-associated DNA fragmentation and subdiploid cells, as described by us previously (23, 24).
Immunoblotting. The control and SFN-treated cells were lysed, as described by us previously (25). Prostate/tumor tissues were thawed on ice, and samples weighing ∼100 mg were minced and homogenized in 5 volumes of ice-cold PBS containing protease and phosphatase inhibitors. Tissue homogenates were centrifuged at 5,000 × g for 10 min at 4°C to remove tissue debris followed by centrifugation at 14,000 × g for 30 min to obtain clear supernatants. Immunoblotting was performed as described by us previously (23–25).
SFN treatment. We selected TRAMP mice for the studies described herein because (a) a well-defined course of disease progression with similarities to human prostate carcinogenesis from histologic prostatic intraepithelial neoplasia (PIN) to well-differentiated (WD) and poorly differentiated (PD) carcinoma and distant site metastasis renders TRAMP mice suitable for chemoprevention studies (26–28), (b) the chemoprevention studies in TRAMP mice can be completed in a reasonable time frame (27, 28), and (c) the TRAMP model has been used successfully to test chemopreventive efficacy of several natural agents (29–32). Male TRAMP[C57BL/6xFVB]F1 hybrid mice were generated by breeding female TRAMP on a C57BL/6 background with nontransgenic FVB male mice. Male TRAMP mice were identified by PCR using DNA obtained from tail clippings (27). After transgene verification, male TRAMP mice were maintained in a climate-controlled environment with a 12-h light/12-h dark cycle and fed food and water ad libitum. At 6 to 7 wk of age, mice were randomized into two groups. Mice in the control group (n = 14) received 0.1 mL PBS by oral gavage thrice per week (Monday, Wednesday, and Friday), whereas the experimental mice (n = 13) received 6 μmol SFN in 0.1 mL PBS by oral gavage thrice per week for 17 to 19 wk. Body weights of the control and SFN-treated mice were recorded once each week beginning at the onset of the study. The mice were sacrificed 24 h after the last administration of the vehicle or SFN by CO2 inhalation. A portion of the prostate/tumor tissue was placed in 10% neutral buffered formalin and paraffin-embedded.
Pathologic evaluation and scoring of tumor stage. Ten randomly selected fields of H&E-stained sections of the dorsolateral prostate (DLP) from individual mice of both control and SFN-treated groups were independently scored by two researchers for incidence and percentage of area corresponding to each pathologic stage. Pathologic grading was performed according to the recommendations of Greenberg and colleagues (27, 28), as described by us previously (32). Lung and pelvic lymph nodes were evaluated for the presence of metastasis by two independent researchers.
Pharmacokinetic study. Groups of male nontransgenic mice (n = 3, 6–8 wk of age) were gavaged with 6 μmol SFN in 0.1 mL PBS. Mice were sacrificed at specified time intervals, and plasma and lung tissue were collected and kept frozen at −80°C. Plasma (0.1 mL) or a 1:3 (w/v) homogenate of lung tissue in PBS (0.2 mL) was mixed with 10 μL of 10 μg/mL alyssin (internal standard, LKT Laboratories). Lung homogenates were mixed with 1 mL of acetonitrile and centrifuged at 16,000 × g for 5 min at room temperature. The resulting supernatants were evaporated to dryness under a gentle stream of nitrogen at 37°C and reconstituted in 200 μL distilled water/formic acid (200:10, v/v). Plasma samples were mixed with 5 μL formic acid. Each reconstituted lung or plasma sample was applied to a Bond Elute (C2, 500 mg, 3 mL) solid-phase cartridge (Varian) preconditioned with 2 mL of methanol and 2 mL of 1% formic acid. The cartridges were washed with 2 mL of distilled water/methanol (95:5, v/v). The SFN and the internal standard were eluted with 2 mL of acetonitrile, and the eluates were evaporated to dryness under a gentle stream of nitrogen at 37°C. Each dried residue was reconstituted in 100 μL of acetonitrile/distilled water (15:85, v/v) and transferred to an autosampler vial, and 10 μL were injected into the HPLC system. The HPLC system consisted of an Agilent model 1100 autosampler and binary pump (Agilent Technologies) and a Phenomenex Hydro-Synergi RP (4 μm, 100 × 2 mm) reverse-phase analytic column. A gradient mobile phase consisting of solvent A (acetonitrile/formic acid, 99.9%/0.1%) and solvent B (water/formic acid, 99.9%/0.1%) was used for separation. The initial condition was 15% solvent A (2 min), which increased linearly in 8 min to 40% solvent A. This condition was held for 5 min, and then solvent A was linearly returned to 15% at 16 min. This condition was held for 8 min. The flow rate was 0.2 mL/min. The SFN and the internal standard were detected using a ThermoFinnigan MSQ Mass Spectrometer (Thermo Electron) operating in positive electrospray, single-ion mode monitoring m/z 177.7 for SFN and m/z 191.7 for alyssin. Concentrations of SFN were determined by calculating the ratio of SFN peak area to that of the internal standard and back-calculating the concentration of SFN in the unknown samples with a concomitantly analyzed standard curve in the respective matrix. Standard curves were linear over the range of 10 to 10,000 ng/mL and displayed acceptable performance (accuracy within 15% of nominal concentrations and precision <15% at all concentrations). Pharmacokinetic data were analyzed noncompartmentally with PK Solutions 2.0 (Summit Research Service5
Immunohistochemical analyses of T-antigen, proliferating cell nuclear antigen, CD31, and E-cadherin expression and quantitation of apoptotic bodies. Tissues were sectioned at 4 to 5 μm thickness, deparaffinized, and rehydrated. Immunohistochemical analysis of T-antigen, proliferating cell nuclear antigen (PCNA), CD31, and E-cadherin expression was performed, as described by us previously (32). Apoptotic bodies were visualized by TUNEL staining, as described by us previously (32). The images were analyzed using Image ProPlus 5.0 software (Media Cybernetics) for quantitation of T-antigen, PCNA, and E-cadherin expression, as well as for analysis of microvessel number (CD31 staining).
Cytotoxicity of natural killer cells and dendritic cells. After transgene verification, male TRAMP mice (ages, 6–7 wk) were randomized into two groups with six mice per group. The mice were gavaged thrice weekly (Monday, Wednesday, and Friday) with either 6 μmol SFN in 0.1 mL PBS or vehicle (control) for 13 wk. We reasoned that SFN-mediated inhibition of prostate carcinogenesis in TRAMP mice may be preceded by augmentation of natural killer (NK)/dendritic cell (DC) function, which prompted us to terminate this experiment at an earlier time point. The mice were then sacrificed, and the spleen and bone marrow were harvested from mice of each group for isolation of NK cells and DCs, respectively, as described by us previously (32). Cytotoxicity of NK and/or DC toward TRAMP-C1 target cells was determined using the Cytotox 96 nonradioactive assay, as described previously by us (32). Quantitation of T-cell and NK cell infiltration in the DLP was performed by immunohistochemical analysis of CD3+ and CD57+ cells, respectively, using the double-stain kit from Biocare Medical according to the manufacturer's protocol.
Quantitation of interleukin-12 and IFNγ release. Spleen-derived NK cells and bone marrow–derived DCs from control and SFN-treated mice were cultured for 5 d, and the culture media were then collected and analyzed for IFNγ and interleukin-12 (IL-12) release, respectively, using ELISA kits from Biosource.
SFN treatment decreased viability of TRAMP-C1 cells by causing apoptosis. As can be seen in Fig. 1A, the viability of TRAMP-C1 cells was decreased significantly after a 24-hour exposure to SFN in a dose-dependent manner. The SFN-mediated suppression of TRAMP-C1 cell viability correlated with apoptosis induction, as revealed by a dose-dependent increase in cytoplasmic histone-associated DNA fragmentation (Fig. 1B) and enrichment of subdiploid apoptotic fraction (Fig. 1C). Moreover, consistent with the results in human prostate cancer cells (16, 22), SFN treatment caused induction of multidomain proapoptotic protein Bax and down-regulation of antiapoptotic Bcl-xL protein (Fig. 1D).
Oral administration of SFN inhibited prostate cancer progression in TRAMP mice. Next, we proceeded to determine the effect of SFN administration on prostate carcinogenesis using TRAMP mice. We have shown previously that thrice weekly oral gavage of 6 μmol SFN is highly effective in suppressing growth of PC-3 xenografts in male nude mice (22). We used a similar dosing regimen in the present study to determine efficacy of SFN against prostate carcinogenesis in TRAMP mice. Representative H&E staining in sections of DLP from control and SFN-treated mice are depicted in Fig. 2A. The DLP of the vehicle-treated control TRAMP mice exhibited low-grade and high-grade PIN and WD and PD carcinomas along with areas consistent with normal prostate. The incidence of the PIN and WD carcinoma in the DLP of mice treated with SFN was lower by 28% (P = 0.006 by Mann-Whitney test) and 23% (P = 0.0104), respectively, in comparison with control mice (Fig. 2B). The area occupied by the WD in the DLP of SFN-treated mice was lower by ∼44% (P = 0.0011 by Mann-Whitney test) compared with control mice (Fig. 2C). Although the area occupied by the PIN in the DLP of SFN-treated mice was lower by ∼24% compared with control, the difference did not reach statistical significance. The body weights of the control and SFN-treated mice did not differ significantly throughout the experimental protocol (results not shown). The mice of both groups were monitored on alternate days for signs of stress throughout the experimental protocol. Except for some mice that were moribund due to large tumor burden before the termination of the experiment, control and SFN-treated mice did not exhibit signs of stress, including food and water withdrawal, impaired posture or movement, ruffled fur, areas of redness or swelling, and indigestion. Moreover, the vital organs (liver, kidney, lung, and heart) from control and SFN-treated mice were histologically normal, except for microscopic evidence of pulmonary metastasis. However, the SFN-mediated inhibition of prostate carcinogenesis was not due to suppression of T-antigen expression (Fig. 2D).
SFN administration inhibited pulmonary metastasis. Figure 3A depicts H&E staining in lung section of a representative mouse of both the control and SFN treatment group. The incidence of pulmonary metastasis was ∼80% in vehicle-treated control mice, which was reduced to ∼40% in the SFN treatment group (Fig. 3B). The incidence of metastasis to pelvic lymph nodes was comparable in control and SFN-treated mice (Fig. 3B). Multiplicity of pulmonary metastatic lesions in mice treated with SFN was lower by ∼63% (P = 0.042 by t test) compared with control mice (Fig. 3C).
Pharmacokinetic measurements. We found that SFN was absorbed rapidly and reached a peak plasma concentration (Cmax) of 3,410 ng/mL at a Tmax of ∼42 minutes (Fig. 3D). The terminal elimination half-life of SFN in plasma was ∼255 minutes. The Cmax of SFN in the lung was ∼3,686 ng/g with a Tmax of ∼15 minutes. The area under the concentration versus time curve between 0 and 8 hours (AUC0-8) of SFN in lung (5.80 μg/g·h) was 47% higher than the AUC0-8 of SFN in plasma (3.94 μg/mL·h).
SFN administration decreased cell proliferation. We determined expression of PCNA, a well-accepted marker of cell proliferation (33), by immunohistochemistry to test whether SFN-mediated suppression of prostate carcinogenesis was due to inhibition of cellular proliferation. As can be seen in Fig. 4A, PCNA expression was ∼40% lower in the DLP of mice treated with SFN compared with control mice (P = 0.028 by t test). The fraction of TUNEL-positive apoptotic bodies was ∼1.9-fold higher in the DLP of SFN-treated mice compared with the controls, but the difference was statistically insignificant due to large data scatter, especially in the SFN treatment group (Fig. 4B).
SFN exhibits antiangiogenic activity in vitro and in vivo (34). To test whether SFN administration caused suppression of neovascularization in vivo, the DLP from vehicle-treated control mice and SFN-treated mice were stained for angiogenic marker CD31. Immunohistochemical staining for CD31 in the DLP of representative mouse from both the control and SFN groups are shown in Fig. 4C (vessels are identified by arrows). The average number of vessels in the DLP did not differ significantly between control and SFN-treated mice (Fig. 4C).
E-cadherin is considered a suppressor of invasion and growth of many epithelial cancers (35). We, therefore, compared expression of E-cadherin in the DLP of control and SFN-treated mice. Expression of E-cadherin was slightly higher in the DLP of SFN-treated mice compared with that of control mice, but the difference was statistically insignificant. These results indicated that the SFN-mediated inhibition of prostate carcinogenesis correlated with decreased cellular proliferation, as well as a modest increase in the number of apoptotic bodies in the DLP.
Effect of SFN administration on levels of Bcl-2 family proteins. We compared expression of Bcl-2 family proteins using prostate/tumor supernatants to determine the mechanism of SFN-mediated apoptosis. As can be seen in Fig. 5, the levels of proapoptotic Bcl-2 family proteins Bad, Bak, Bax, and Bid were higher by ∼1.33-fold, 1.16-fold, 1.27-fold, and 1.9-fold, respectively, in the prostate of SFN-treated mice compared with controls. Additionally, SFN administration resulted in a marked decrease (∼63% decrease compared with control) in the level of antiapoptotic protein Mcl-1 compared with control (Fig. 5). The level of cleaved poly(ADP-ribose) polymerase (PARP) was ∼2.7-fold higher in the prostate of SFN-treated mice compared with the control (Fig. 5).
SFN augmented cytotoxic activity of NK/DC cocultures against TRAMP-C1 cells. The NK cells play an important role in immunosurveillance during tumorigenesis (36, 37). The NK cells mediate cytotoxicity against tumor cells by multiple mechanisms, including granule-mediated exocytosis and induction of apoptosis (36). To assess if SFN administration affected NK cell/DC function, we determined the cytotoxicity of NK cells and/or DC isolated from control and SFN-treated mice toward TRAMP-C1 as a target cell line. The results in Fig. 6A show that the NK cells isolated from the spleen of SFN-treated TRAMP mice were more cytotoxic against TRAMP-C1 cells compared with the NK cells isolated from control mice (23% versus 50% cytotoxicity for NK cells isolated from control and SFN-treated mice, respectively; P < 0.05 by t test). It is well established that interactions of NK cells with DC result in coactivation of both cell types (38). We, therefore, determined the lytic functions of NK cell and DC cocultures. The cytotoxic activity of NK cells against TRAMP-C1 cells was increased even more with coculture of NK cells with DC at 10:8 ratio compared with NK cells alone. This effect correlated with a significant increase in IL-12 production by the DC isolated from the bone marrow of SFN-treated mice compared with control mice (Fig. 6B,, left). On the other hand, the IFNγ production by the NK cells did not differ significantly between control and SFN-treated mice (Fig. 6B , right). Altogether, these results indicate that SFN administration stimulates immune response by augmenting NK cell lytic function through DC-mediated IL-12 production.
SFN administration caused T-cell infiltration in the prostate tumor. Next, we explored the possibility of whether the prostate tumor from SFN-treated mice was infiltrated by carriers of an adaptive immune response, such as T cells. We stained prostate tumors from control and SFN-treated mice using anti-CD3 (a marker of T cells) and anti-CD57 (a marker of NK cells) antibodies. The results can be visualized in Fig. 6C, where NK cells are stained in brown and the T cells are stained in pink. The prostate tumors from SFN-treated mice exhibited a modest yet statistically significant increase in T-cell, but not NK cell, infiltration compared with control (Fig. 6D). Taken together, our results indicate that SFN administration increases immunosurveillance by boosting both the NK cell lytic function and T-cell infiltration.
The present study shows that SFN not only suppresses the growth of a TRAMP tumor–derived cell line in culture but also prevents incidence and/or burden of PIN and well-differentiated carcinoma and incidence and multiplicity of lung metastasis in male TRAMP mice without causing weight loss or affecting the T-antigen expression. The possibility that higher concentration of SFN and/or a more intensive dosing regimen (e.g., daily administration) results in better chemopreventive response cannot be ruled out. Likewise, it is possible that earlier onset of SFN administration (e.g., starting at 4–5 weeks of age) leads to even greater protection against prostate carcinogenesis in TRAMP mice. Additional work is needed to systematically explore these possibilities.
The bioavailability of SFN in TRAMP mice is yet to be evaluated, but this variable has been determined in rats. The SFN was detectable in the plasma after 1 hour of oral gavage of rats with 50 μmol SFN (39). The maximal plasma concentration of SFN was found to be ∼20 μmol/L after 4 hours of oral gavage and declined with a half-life of 2.2 hours (39). Hanlon and colleagues (40) have shown that SFN is rapidly absorbed with an absolute bioavailability of ∼82% in rats (40). The maximal plasma concentration of SFN and its cysteine conjugate varied between 0.65 and 0.82 μmol/L in healthy human volunteers, who consumed a test meal of broccoli soup (150 mL) containing 100 g of broccoli (41). The SFN concentration effective against prostate carcinogenesis and pulmonary metastasis in TRAMP mice is well within the range that may not only be generated through dietary intake of cruciferous vegetables but can also be safely given to humans without any harmful side effects (5, 42).
We have shown previously that SFN treatment suppresses growth of cultured human prostate cancer cells by causing G2-M phase cell cycle arrest and apoptosis induction (13, 16, 21, 22). The present study reveals that a cell line derived from the prostate tumor of a TRAMP mouse is also highly sensitive to growth inhibition and apoptosis induction by SFN treatment. We also found that oral administration of SFN causes a statistically significant decrease in cell proliferation in the DLP of TRAMP mice, as evidenced by reduced expression of PCNA. Moreover, the average number of TUNEL-positive apoptotic bodies was markedly higher in the DLP of SFN-treated mice in comparison with controls. Thus, it is reasonable to conclude that SFN-mediated inhibition of prostate carcinogenesis in TRAMP mice in vivo is mediated by reduced cellular proliferation and increased apoptosis.
Distant site metastasis is a major cause of death in prostate cancer patients (43). The pathogenesis of metastasis is complex and involves a series of molecular events, including synthesis and secretion of angiogenic factors to promote neovascularization, motility, and local invasion of the host stroma, entry into the circulation, detachment, and extravasation (44). Steps leading to metastasis are complex and regulated by multiple molecules, including growth factors, matrix metalloproteinases, and cell adhesion molecules (44). For example, loss of expression of cell adhesion molecules, especially E-cadherin, is believed to be important for development of metastatic lesions (35). Inhibition of prostate carcinogenesis and metastasis by the plant flavonoid apigenin in TRAMP mice was shown to correlate with retained expression of E-cadherin (29). The present study reveals that the SFN administration inhibits pulmonary metastasis incidence and multiplicity in TRAMP mice. However, the inhibitory effect of SFN administration against pulmonary metastasis seems independent of changes in E-cadherin expression or inhibition of angiogenesis. Delineation of the precise mechanisms by which SFN administration inhibits pulmonary metastasis awaits further work.
The present study shows, for the first time, that treatment of TRAMP mice with SFN leads to increased production of IL-12 by DC. This effect results in augmentation of NK cell cytotoxicity toward a prostate cancer cell line and increased efficiency of DC activation of NK cell cytotoxic function after their coculture. Previous studies have documented the capacity for DC to internalize, process and present tumor cell-derived antigens, and thereby to stimulate the antitumor function of both CTLs and IFNγ-producing NK cells (45, 46). In addition to the stimulatory effects of SFN administration on in vitro functions of NK cells and DC, we observed an increase in T-cell infiltration in the prostate tumors of SFN-treated TRAMP mice in comparison with control mice. Taken together, our results suggest that SFN can enhance immune function toward prostate tumors by stimulating IL-12 release by DC, leading to an increased level of cytotoxicity of NK cells and ultimately to a greater infiltration of prostate tumors by T cells, which results in a lower tumor burden and less metastases.
In conclusion, the results of the present study indicate that oral administration of SFN significantly inhibits prostate carcinogenesis and pulmonary metastasis in TRAMP mice without causing any harmful side effects or reducing T-antigen expression.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: National Cancer Institute US PHS grants CA115498 and CA101753.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.