Lycopene inhibits IGF-I signal transduction and growth in normal prostate epithelial cells by decreasing DHT-modulated IGF-I production in co-cultured reactive stromal cells

Abstract

Prostate stromal and epithelial cell communication is important in prostate functioning and cancer development. Primary human stromal cells from normal prostate stromal cells (PRSC) maintain a smooth muscle phenotype, whereas those from prostate cancer (6S) display reactive and fibroblastic characteristics. Dihydrotestosterone (DHT) stimulates insulin-like growth factor-I (IGF-I) production by 6S but not PSRC cells. Effects of reactive versus normal stroma on normal human prostate epithelial (NPE or PREC) cells are poorly understood. We co-cultured NPE plus 6S or PRSC cells to compare influences of different stromal cells on normal epithelium. Because NPE and PREC cells lose androgen receptor (AR) expression in culture, DHT effects must be modulated by associated stromal cells. When treated with camptothecin (CM), NPE cells, alone and in stromal co-cultures, displayed a dose-dependent increase in DNA fragmentation. NPE/6S co-cultures exhibited reduced CM-induced cell death with exposure to DHT, whereas NPE/PRSC co-cultures exhibited CM-induced cell death regardless of DHT treatment. DHT blocked CM-induced, IGF-I-mediated, NPE death in co-cultured NPE/6S cells without, but not with, added anti-IGF-I and anti-IGF-R antibodies. Lycopene consumption is inversely related to human prostate cancer risk and inhibits IGF-I and androgen signaling in rat prostate cancer. In this study, lycopene, in dietary concentrations, reversed DHT effects of 6S cells on NPE cell death, decreased 6S cell IGF-I production by reducing AR and β-catenin nuclear localization and inhibited IGF-I-stimulated NPE and PREC growth, perhaps by attenuating IGF-I's effects on serine phosphorylation of Akt and GSK3β and tyrosine phosphorylation of GSK3. This study expands the understanding of the preventive mechanisms of lycopene in prostate cancer.

Introduction

Prostate stromal and epithelial cell interactions are important in tissue function and play a role in the development of prostate cancer ( 1 ). Paracrine interactions between cultured prostate epithelial cells and stromal fibroblasts have been investigated for gene expression and secretory activities using co-culture models ( 2–4 ). Interactions between stromal and epithelial cells in hormone-responsive tissues are increasingly recognized as an important area of research highlighted for many diseases, especially cancer ( 5 ). Many factors may be involved in the bidirectional signaling between the stromal and epithelial cells. These signals are seminal in regulating tissue homeostasis, also the loss of which may contribute to progression of prostate cancer ( 6 ). IGF-I is an important factor in mediating stromal–epithelial cell signaling; it induced endometrial epithelial cell proliferation only when co-cultured with stromal cells, an effect that was blocked by adding IGF-I antibody ( 7 ). We recently reported that primary human stromal cells isolated from a normal prostate (PRSC) maintain a smooth muscle phenotype, whereas those obtained from prostate cancer (6S) display more reactive and fibroblastic characteristics, and that DHT modulates IGF-I, IGFBP-2 and IGFBP-3 expression in 6S but not PRSC cells ( 8 ). To date, there is little information regarding the impact of reactive versus normal stroma on cell proliferation or death of normal prostate epithelial (NPE) cells.

NPE cells lose androgen receptor (AR) expression in culture; thus, androgen effects in prostate epithelial plus stromal cell co-cultures must be modulated by associated stromal cells, which do express AR. Androgen binding to AR transforms it to an active conformation and initiates AR translocation to the nucleus, with consequent regulation of gene expression ( 9 , 10 ). We have demonstrated that testosterone (T) and DHT can induce 6S but not PRSC stromal cells to increase IGF-I production ( 8 ). AR also interacts with β-catenin protein, a co-transcription factor in the Wnt pathway, modulating gene expression ( 11–15 ). However, the role of AR–β-catenin complex in androgen-modulated IGF-I production is unknown.

Elevated levels of IGF-I have been associated with increased risk of prostate cancer ( 16 ). IGF-I regulates the growth of cells through two complementary mechanisms. IGF-I stimulates cell proliferation and inhibits cell death, including apoptosis ( 17–20 ) through activation of phosphatidylinositol 3-kinase and subsequent activation of the downstream protein serine/threonine kinase, Akt ( 18 , 21 , 22 ).

Lycopene and other tomato-derived carotenoids exert anti-proliferative effects on prostate and other cancer cells in vitro and in xenografted mice ( 23–26 ). Lycopene also induces mitochondrial apoptosis in human LNCaP prostate cancer cells ( 27 ). Supra-dietary concentrations of lycopene significantly inhibit DNA synthesis in primary human prostate epithelial cells ( 28 ). Both epidemiologic and case–control studies have suggested that increased consumption of tomato products and greater blood concentrations of lycopene are associated with a reduced risk of prostate cancer ( 29–34 ). In one recent pilot study, patients with localized prostate cancer given antioxidant tablets including 10 mg lycopene/day exhibited reduced prostate-specific antigen (PSA) velocity during 1 year of follow-up ( 28 ). Lycopene inhibits IGF-I and androgen signaling in normal rat prostate and in rat prostate cancer models ( 35 , 36 ). Lycopene, in dietary ( 37 ) or supra-dietary concentrations ( 38 ), inhibits prostate cancer cell IGF-I receptor (IGF-R) expression induced by high concentrations of IGF-I, and inhibits Akt phosphorylation in prostate cancer cells incubated in serum-rich medium ( 37 ). Lycopene, in dietary concentrations was ineffective in decreasing prostate tumor volume and circulating PSA levels in a mouse xenograft model of human prostate cancer ( 39 ). However, supplementation with lycopene extracts was effective in decreasing abnormally elevated IGF-I concentrations in colon cancer patients ( 40 ).

The objectives of the current study were to examine (i) the role of stromal-epithelial cell communication in modulating actions of DHT on NPE cell proliferation and death, (2) the factors controlling NPE cell death in stromal-epithelial co-culture and (3) the mechanisms by which lycopene, in dietary concentrations, inhibits DHT and IGF-I effects, and the importance of stromal cell type on mediating such effects, in NPE cells.

Materials and methods

Human prostate stromal and epithelial cells

Normal prostate stromal (PRSC) cells (Cambrex, Walkersville, MD) and 6S prostate cancer-derived stromal cells were isolated as described previously ( 8 ) and grown in Dulbecco's Modified Eagle's Medium: F12 (1:1) medium (Invitrogen, Gaithersburg, MD), with penicillin (100 U/ml), streptomycin (100 μg/ml), L -glutamine (292 μg/ml) (Invitrogen) and 5% fetal bovine serum (Hyclone Laboratories, Logan, UT) at 37°C in 5% CO ₂ as described previously ( 8 ). NPE (PREC) cells (derived from a 19-year-old man; Cambrex) were maintained in PREGM medium (Cambrex) containing bovine pituitary extract (4 μl/ml), insulin (1 μl/ml), hydrocortisone (1 μl/ml), gentamicin sulfate (1 μl/ml), retinoic acid (1 μl/ml), transferrin (1 μl/ml), triiodothyronine (1 μl/ml), epinephrine (1 μl/ml) and recombinant human epidermal growth factor (1 μl/ml). NPE cells (primary cells isolated from a 50-year-old man; gift of Mark Stearns, Drexel University, PA) were grown in keratinocyte serum-free media (Invitrogen) containing penicillin (100 U/ml) and streptomycin (100 μg/ml). Similar to many other prostate primary epithelial cell preparations, both epithelial cell lots were AR negative (data not shown).

Antibodies

Antibodies against IGF-I and Insulin like growth factor-I receptor (IGF-R) were purchased from Sigma–Aldrich (St Louis, MO). Antibodies against AR, pS(9)GSK3β, GSK3 and normal IgG without sodium azide were obtained from Santa Cruz (Santa Cruz, CA). β-catenin antibody and GSK3β were purchased from BD Biosciences (San Jose, CA). The antibody staining pY(279/216)GSK3 was obtained from Upstate (Charlottesville, VA), and those staining pS(473)Akt and Akt from Cell Signaling Technology (Danvers, MA).

Cell growth/death by MTT assay

NPE (or PREC) cells were seeded into a 96-well plate with 500–1500 cells per well in their respective maintenance media overnight. Media were removed, and F12:199 (1:1) medium (Invitrogen) containing penicillin (100 U/ml), streptomycin (100 μg/ml) and 2% charcoal/dextran treated fetal bovine serum (Hyclone Laboratories) (designated as 2% serum medium) was added in the absence or presence of 0.15 or 1 ng/ml IGF-I (Sigma–Aldrich), without or with lycopene (Sigma–Aldrich) for 24 h. The (3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide (MTT) assay (Promega, Madison, WI) was performed according to the protocol of the company, as reported previously ( 8 ). The MTT assay was used with the NPE monoculture assay, as it quantifies both cell growth and, conversely, cell death. NPE cells were treated overnight with or without 10 nM DHT, 0.15 ng/ml IGF-I and/or 1 μm lycopene (2 M lycopene in tetrahydrofuran stored in −80°C). For all experiments, the stock lycopene was diluted using tetrahydrofuran and added to fresh media in varying concentrations, depending on cell culture volumes. Cells were pre-treated overnight with lycopene before adding DHT, IGF-I, etc. Media were not changed after addition of lycopene. For controls, the cells without lycopene were also treated with the same volume of tetrahydrofuran (0.1%) as was used in all the experiments containing lycopene. The next day, cells were treated with or without 1 μg/ml camptothecin (CM) for 6 h, and MTT dye was added to each of the wells. After overnight incubation, MTT stopping buffer was added to each of the wells and samples were analyzed by spectrophotometer. The concentrations of IGF-I were chosen based on previous growth response studies of IGF-I on NPE cells (data not shown).

Cell co-cultures

Stromal cells were seeded into 6- or 24-well plates in 2% serum-containing medium. Epithelial cells were grown on 30 or 12 mm Millicell cellular inserts (Millipore, Bedford MA). Prior to seeding, inserts were coated with Matrigel (1:10 dH ₂ O) (BD Biosciences, Bedford, MA). Cells were grown in monoculture for 24 h to allow for cell adhesion. Media were removed and cellular inserts were added to stromal containing wells and co-cultured in media described above for specific epithelial cell type. Cells were co-cultured for 72 h after combination and treatment with DHT, IGF-I or lycopene (Sigma–Aldrich).

Cell death analysis

Cell death was induced by the addition of CM (Sigma–Aldrich) to cell co-cultures for 24 h. The cell death enzyme-linked immunosorbent assay (Roche Applied Science, Indianapolis, IN) was performed according to manufacturer's protocol. Briefly, cells were grown as described above on 30 or 12 mm inserts. Following hormone and/or CM treatment, inserts were removed and washed three times with phosphate-buffered saline to remove any media. Wash solutions were centrifuged and resultant pellets were re-suspended in 500 μl of provided incubation buffer and added to epithelial cell containing inserts for lysis. After 30 min, lysates were centrifuged at 14 000 r.p.m. for 10 min. Supernatant containing only the cytoplasmic fraction (fragmented DNA) was removed and added to sample plates. Prior to sample addition, plates were coated with primary anti-histone antibody. After 90 min incubation with samples, plates were rinsed three times with washing buffer and incubated with secondary detection antibody for 90 min. After three more washings, substrate solution was added and the absorbance of each plate was read at 405 nm. The results were normalized to cell numbers and protein concentrations.

To explore the effects of DHT on NPE cell death, NPE cells were co-cultured with normal stromal PRSC cells or 6S prostate cancer-derived stromal cells, in the absence or presence of various concentrations of DHT (0.1, 10 and 1000 nM) for 72 h, followed by treatment with CM (0, 0.1, 0.5 and 2 μg/ml) for 24 h, after which cell death assays were performed. To block IGF-I induced by DHT, anti-IGF-I (2.5 μg/ml) and anti-IGF-R (1 μg/ml) antibodies were added into the co-culture (6S/NPE) following DHT treatment. The same amount of normal IgG was added to the control group. There was no difference with or without the addition of normal IgG in the control group in the cell death assay (data not shown). The remaining procedures were the same as those described.

Determination of IGF-I RNA concentrations using quantitative real-time polymerase chain reaction

6S stromal cells were seeded into six-well plates with 2% serum medium for 2 days. Media were removed, and fresh 2% serum-containing medium was added in the absence or presence of 100 nM DHT, without or with 1 μM lycopene, for 72 h. Quantitative real-time polymerase chain reaction measurement of IGF-I RNA was performed as described previously ( 8 ).

Cell lysis and fractionation

6S stromal cells were seeded and treated as described above. After 24 h treatment, cells were lysed using a cell lysis buffer (Cell Signaling Technology) for western blots, or using buffers from a kit (nuclear and cytoplasmic extraction reagents, Pierce, Rockford, IL) for cell fractionation, which was done according to the manufacturer's protocol.

Modulation of Akt/GSK3 phosphorylations by IGF-I/lycopene

NPE cells (PRECs) were seeded into 24-well plates with maintenance medium overnight. The next day, medium was replaced with 2% serum-containing medium without or with 1 μM lycopene. After 24 h, cells were treated with 1 ng/ml IGF-I for 0, 1, 2 and 4 h, respectively. Cells were then lysed using the lysis buffer from Cell Signaling Technology.

Western blot

Protein concentrations were assessed in the above cell lysates and cell fractionation extracts using the bicinchoninic acid protein (BCA) assay kit (Pierce). Western blotting has been described previously ( 41 ).

Statistical analysis

All results are presented as mean values ± SE of three to five separate experiments. The statistical analysis was performed using the JMP (SAS Institute, Cary, NC) statistical package based on t -test. An adjusted P value of 0.05 was considered significant. Densitometric values for protein bands of phosphorylated signaling intermediates were reported in arbitrary units above background values after normalization to total signaling intermediate protein levels.

Results

DHT, lycopene and IGF-I effects on NPE cell death in co-cultured prostate epithelial plus stromal cells

We have shown that IGF-I is induced by DHT in 6S cells ( 8 ), whereas CM, a DNA topoisomerase I inhibitor, causes breaks in DNA strands ( 42 ) and is widely used to induce cell death in vitro . DNA damage is thus an initial event of cell death. We investigated whether the IGF-I secreted by stromal cells exerts protective effects on prostate epithelial cells treated with CM. The cell death kit from Roche is able to directly measure the event. As shown in Figure 1A , NPE cells co-cultured with normal prostate primary stromal (PRSC) cells and treated with CM (2 μg/ml) exhibited a peak, ∼2-fold stimulation of cell death as measured by DNA fragmentation ( P < 0.001) compared with cells not treated with CM, independent of DHT administration. There was a non-significant trend to increasing cell death with non-maximal doses of CM. In comparison, control-treated NPE cells co-cultured with cancer-associated (6S) stromal cells exhibited >2-fold ( P < 0.001) increase ( Figure 1B ) in cell death after exposure to 2 μg/ml CM, whereas DHT at all concentrations tested significantly inhibited CM-induced cell death ( P < 0.001).

Fig. 1.

Effects of DHT, IGF-I and lycopene on CM-modulated cell death in co-cultured prostate epithelial + stromal cells. (A) Normal human prostate epithelial (NPE) cells were co-cultured with normal prostate stromal (PRSC) cells for 72 h in the absence or presence of varying concentrations of DHT, without and with addition of differing amounts of CM, and epithelial cell death was measured in a DNA fragmentation assay. All values shown represent the mean values ± SEM of three to five experiments performed in triplicate. *** P < 0.001 versus control. (B) Normal human prostate epithelial (NPE) cells were co-cultured with reactive, fibroblastic, prostate cancer-derived stromal (6S) cells for 72 h in the absence or presence of varying concentrations of DHT, without and with addition of differing amounts of CM, and epithelial cell death was measured in a DNA fragmentation assay. All values shown represent the mean values ± SEM of three to five experiments performed in triplicate. *** P < 0.001 versus control. (C) Normal human prostate epithelial (NPE) cells were co-cultured with prostate stromal (6S) cells for 72 h in the absence or presence of varying concentrations of lycopene (LP), without or with co-administration of DHT, CM or both, and epithelial cell death was measured in a DNA fragmentation assay. All values shown represent the mean values ± SEM of three to five experiments performed in triplicate. * P < 0.05 versus control, ** P < 0.01 versus any other treatments with 1 μM lycopene, *** P < 0.001 versus any other treatment without lycopene. (D) Normal human prostate epithelial (NPE) cells were co-cultured with prostate stromal (6S) cells for 72 h in the presence of normal IgG or anti-IGF-I and IGF-R, without or with co-administration of DHT, CM or both, and epithelial cell death was measured in a DNA fragmentation assay. All values shown represent the mean values ± SEM of three experiments performed in triplicate. *** P < 0.001 versus any other treatment.

To determine whether lycopene affected DHT's anti-cell death actions, NPE cells were co-cultured with 6S cells and treated with different concentrations of lycopene without or with 10 nM DHT for 72 h, followed by treatment without or with CM (2 μg/ml) for 24 h and measurement of cell death by DNA fragmentation. As shown in Figure 1C , administration of lycopene alone, or DHT + lycopene, did not induce significant cell death of the NPE cells co-cultured with 6S stromal cells. CM greatly increased NPE cell death in the absence of DHT treatment ( P < 0.001), whereas CM-induced cell death was inhibited substantially by the addition of DHT ( P < 0.001). However, if cells were pre-treated with lycopene (0.3 and 1 μM), the DHT inhibitory effects were reversed, and cells were returned to the higher cell death state ( P < 0.05 and P < 0.01), although lower than that seen with CM alone.

Because DHT increases IGF-I production in 6S cells but not PRSC cells ( 8 ) and IGF-I inhibits cell death including apoptosis ( 18–20 ), IGF-I might be a factor reducing CM-induced NPE cell death in 6S/NPE co-culture ( Figure 1B and C ). To determine this, we added antibodies to IGF-I and IGF-R to the 6S/NPE co-culture. Figure 1D reveals that the antibodies exerted similar effects as did normal IgG on NPE death in cells that were left untreated or treated with 1 μg CM or 10 nM DHT, whereas DHT failed to rescue NPE death induced by CM ( P < 0.001) after anti-IGF-I and anti-IGF-R treatment. Because IGF-R mediates biological activities of both IGF-I and IGF-II ( 43 ), anti-IGF-R antibody might also block IGF-II activity. However, DHT, even at a 100 nM concentration, does not increase IGF-II production in 6S cells (our unpublished results). Therefore, IGF-I was involved in mediating the DHT effect to prevent NPE cell death in CM-treated co-cultures.

Effects of lycopene on DHT-induced IGF-I mRNA expression in 6S stromal cells

Because IGF-I was a factor controlling NPE cell death and lycopene reversed DHT-rescued cell death ( Figure 1C and D ), we evaluated the effects of lycopene on DHT-induced IGF-I expression in 6S cells. Accordingly, 6S cells were treated with DHT (100 nM) for 72 h, which increased IGF-I mRNA about 6-fold ( P < 0.001; Figure 2 ) above that of control cells, similar to our previous results ( 8 ). Administration of 1 μM lycopene alone did not significantly affect IGF-I mRNA production, whereas lycopene inhibited the DHT-induced rise of IGF-I mRNA by 50% ( P < 0.001). We did not perform the same experiments using PRSC cells as these cells do not increase IGF-I mRNA induced by DHT ( 6 ).

Fig. 2.

Effects of DHT, lycopene and DHT + lycopene on IGF-I RNA production by 6S prostate stromal cells. Reactive, fibroblastic prostate stromal (6S) cells were cultured for 72 h in the absence or presence of DHT, lycopene or both, and IGF-I RNA was measured by quantitative real-time polymerase chain reaction. All data shown represent mean values ± SE averaged from three separate experiments, each performed in triplicate; *** versus control, P < 0.001, ** versus control, P < 0.01.

Effects of lycopene on DHT- mediated nuclear co-localization of AR and β-catenin in 6S stromal cells

DHT binding to the AR changes AR conformation and increases AR protein levels ( 9 , 10 ), probably due to increasing AR stability. AR and β-catenin form a complex and modulate gene expression in several cell lines ( 11–15 ). Because lycopene reduced DHT-induced IGF-I expression in 6S cells ( Figure 2 ), lycopene might modulate AR protein expression and association with the co-activator, β-catenin. To determine the effects, 6S cells were left untreated or treated with 100 nM DHT and/or 1 μM lycopene for 24 h. Cells were lysed for western blotting to detect levels of AR and β-catenin protein. DHT administration increased AR stability ( Figure 3A ), whereas lycopene reduced the DHT effect to about 70% ( P < 0.05, Figure 3B ). β-Catenin levels in the lysates did not differ significantly after any of the treatments ( Figure 3A ). We did not detect any change of GSK3β in the above lysates either (data not shown). DHT treatment has been shown to promote AR and β-catenin accumulation in cell nuclei ( 15 ). 6S cells were treated with control or DHT and/or lycopene as above, followed by cell fractionation, after which the effects on nuclear localization of AR and β-catenin were measured. DHT dramatically increased nuclear localization of AR and β-catenin ( Figure 3C ), whereas treatment with DHT and lycopene decreased nuclear localization of AR to ∼60% ( P < 0.01, Figure 3C and D ) and of β-catenin to ∼50% ( P < 0.01, Figure 3C and E ). Administration of lycopene alone increased nuclear β-catenin but not nuclear AR. T cell factor-4 (TCF-4) served as a loading control for standardizing protein amounts.

Fig. 3.

Effects of DHT, lycopene or DHT + lycopene on AR and β-catenin expression in 6S prostate stromal cell lysates and nuclear co-localization. (A) Lycopene reduces the effects of DHT on AR in 6S cell lysates. Untreated or treated 6S cells were lysed and the lysates were examined by direct immunoblotting using the indicated antibodies. These blots represent three separate experiments with similar results. (B) Lycopene reduces DHT-induced AR stability in 6S lysates. Columns represent mean values of AR intensities ± SE averaged from three separate experiments including the first blot on panel (A). Only the intensities of AR from cells treated with DHT or DHT + lycopene were compared. Equal loadings are confirmed by a GSK3β antibody staining (data not shown); * versus DHT treated: P < 0.05. (C) Lycopene decreases the effects of DHT on nuclear co-localization of AR and β-catenin in 6S cells. T cell factor-4 was a loading control. These blots represent three separate experiments with similar results. (D) Lycopene reduces DHT-induced nuclear AR in 6S cells. Columns represent mean values of AR intensities ± SE averaged from three separate experiments including the third blot on panel (A). Only the intensities of AR from cells treated with DHT or DHT + lycopene were compared; ** versus DHT treated: P < 0.01. (E) Lycopene reduces DHT-induced nuclear β-catenin in 6S cells. Columns represent mean values of β-catenin intensities ± SE averaged from three separate experiments including the fourth blot on panel (A). Only the intensities of β-catenin from cells treated with DHT or DHT + lycopene were compared; ** versus DHT treated: P < 0.01.

Roles of IGF-I and lycopene in mediating CM-induced effects in prostate epithelial cells

We speculated that IGF-I, even at the low concentrations (∼0.15 ng/ml) found to be secreted by 6S stromal cells after 72 h treatment with 100 nM DHT ( 8 ), would be sufficient to increase NPE cell growth or decrease death by CM. Reversing DHT effects on NPE cell death in co-culture ( Figure 1C ), lycopene not only reduced IGF-I expression in 6S cells ( Figure 2 ) but also directly exert effects on NPE cells. To assess these possibilities, we used the MTT assay to quantify both cell growth and, conversely, cell death in our monocultures. The results are shown in Figure 4A . In the absence of CM, treatment with 0.15 ng/ml IGF-I induced significant NPE cell growth to about 140% ( P < 0.001), compared with that of control-treated cells. Cell growth was similar using IGF-I at 1 versus 0.15 ng/ml (data not shown). Treatment with lycopene or DHT alone did not significantly affect NPE cell growth. Moreover, lycopene (1 μM) reduced IGF-I-induced cell growth to control levels. In comparison, CM (1 μg/ml) caused about 50% cell death ( P < 0.001), whereas 0.15 ng/ml IGF-I, but not 10 nM DHT, rescued the cells from death ( P < 0.001). In addition, 1 μM lycopene inhibited the cell death rescue ( P < 0.001) induced by IGF-I.

Fig. 4.

Effects of DHT, IGF-I, lycopene and CM on cell proliferation or death of NPE (PREC) cells. (A) Normal human prostate epithelial (NPE) cells were cultured overnight in the absence or presence of DHT, IGF-I and/or lycopene. The next day, cells were treated with or without 1 μg/ml CM for 24 h. Then, cell growth and death were measured in an MTT assay. All values shown represent the mean values ± SEM of three experiments performed in three to eight replicates; *** P < 0.001 versus control. (B) Normal human prostate epithelial (PREC) cells were cultured for 24 h in the absence or presence of IGF-I, lycopene or both, and cell growth was measured in an MTT assay. All values shown represent mean values ± SE averaged from five separate experiments, each performed with four to six replicates. *** P < 0.001 versus any other treatment. The measurements of the control cells were arbitrarily set as one in both experiments [(A) and (B)].

A separate epithelial cell lot was used to further verify that lycopene alone is able to inhibit IGF-I-stimulated NPE cell growth. PREC cells were left untreated or treated with IGF-I (1 ng/ml) and/or lycopene (1 μM) for 24 h. This concentration of IGF-I was found to be similar to concentration used above (0.15 ng/ml) in effects on NPE cell growth. IGF-I-treated PREC cell growth increased to 160% ( P < 0.001) of control values, whereas lycopene addition decreased IGF-I-induced cell growth back to levels which were not significantly different from that of the control cells ( Figure 4B ). Lycopene alone did not change cell growth significantly during the treatment interval (24 h), whereas incubation of cells with lycopene or lycopene/IGF-I for an additional 16–24 h resulted in slower growth (data not shown).

Effects of lycopene on IGF-I-mediated actions on Akt and GSK3 in normal prostate epithelial (PREC) cells

Because lycopene inhibited NPE or PREC cell growth ( Figure 4 ), it might affect IGF-I signaling. To determine this, we focused on the effects of IGF-I on Akt/GSK3 phosphorylation by pre-treating PREC without or with 1 μM lycopene for 16 h followed by treatment with 1 ng/ml (∼0.13 nM) IGF-I for 0, 1, 2 and 4 h. Higher concentrations of IGF-I were not included in order to keep the IGF-I dose consistent with cell growth studies and so it would be comparable to that of IGF-I secreted from DHT-treated 6S cells. Figure 5A illustrates representative immunoblots of time courses of Akt and GSK3 phosphorylation and of intermediate proteins in PREC cells. There were no significant differences in phosphorylation of Akt or GSK3 in cells pre-treated with or without lycopene, and with IGF-I at 0 h ( Figure 5B–D ) in a low-serum medium. IGF-I elicited an ∼30% increase of Ser (473) phosphorylation of Akt at 2 h, compared with control values at 0 h, and the increase was sustained without lycopene pre-treatment through the 4 h time point. In comparison at the 4 h time point, Akt phosphorylation in cells pre-treated with lycopene decreased to ∼70% of the control value at 0 h ( Figure 5B ). IGF-I increased Ser9 phosphorylation of GSK3β (inactive form) by about 2- to 3-fold ( P < 0.01) at 2 and 4 h, but did not increase the phosphorylation significantly in the cells pre-treated with lycopene, compared with that of the control. Lycopene lessened the IGF-I-induced serine phosphorylation of GSK3β by more than 2-fold at 4 h ( P < 0.01; Figure 5C ). Moreover, IGF-I decreased tyrosine (279/216) phosphorylation of GSK3 (active forms) to ∼35% of the control ( P < 0.05) at 1 and 2 h, whereas the phosphorylation was maintained in cells pre-treated with lycopene ( Figure 5D ).

Fig. 5.

Effects of lycopene, IGF-I or lycopene + IGF-I on Akt and GSK3 in normal prostate epithelial (PREC) cells. (A) Lycopene decreases IGF-I effects on Akt and GSK3 in PREC cells. These blots represent three separate experiments with similar results. (B) IGF-I increases but lycopene decreases phosphorylation of Akt in PREC cells. Columns represent mean values of pAkt intensities ± SE normalized by Akt staining and averaged from three separate experiments including the first and second blots on panel (A); ** versus *: P < 0.05. (C) IGF-I increases serine phosphorylation of GSK3β but lycopene lessens the effect in PREC cells. Columns represent mean values of pSGSK3β intensities ± SE normalized by GSK3β staining and averaged from three separate experiments including the third and fourth blots on panel (A); ** and *** versus control, and *** versus *: P < 0.01. (D) IGF-I decreases but lycopene maintains tyrosine phosphorylation of GSK3 in PREC cells. Columns represent mean values of pYGSK3 intensities ± SE normalized by GSK3 staining and averaged from three separate experiments including the fifth and sixth blots on panel (A); * versus any other treatments except **: P < 0.05.

Discussion

This study demonstrates that (1) the potent androgen, DHT, exerts pro-proliferative, anti-cell death effects on NPE cells (which lack an AR), when co-cultured with stromal fibroblasts (6S cells) derived from prostate cancer tissue but not when co-cultured with normal prostate stromal (PRSC) cells. It appears that CM induced more NPE cell death when co-cultured with 6S cells. It is difficult to speculate as to why this is so because so little is known about the differences between normal and cancer-related stroma cells. The unknown constitutive secreted factors from 6S cells might play some role; (2) IGF-I secreted by DHT-treated 6S cells is a factor exerting the aforementioned anti-cell death effects and (3) the carotenoid lycopene, in dietary concentrations, counteracts the effects of DHT, at least in part, by reducing DHT-stimulated IGF-I production via decreasing nuclear co-localization of AR and β-catenin in 6S cells and by attenuating activity of the IGF-I/Akt/GSK3 signaling axis in NPE cells.

We employed NPE cells (or PREC) in these experiments, rather than prostate cancer epithelial cells, so as to delineate response of the former cells to local environmental changes, particularly the stromal cell environment, and to use this in vitro co-culture model to investigate mechanisms of occurrence and prevention of prostate cancer. NPE cells obtained from the normal prostate of a 50-year-old donor were compared with PREC cells from a healthy 19-year-old man, and the two cell lines responded similarly to treatments.

Lycopene-treated NPE cells, when exposed to CM even in the presence of DHT, were as vulnerable to increased cell death as were the control cells ( Figure 1C ). By identifying IGF-I as a factor controlling NPE cell death ( Figure 1D ), and lycopene effects on IGF-I expression in 6S cells ( Figure 2 ) as well as GSK3 activities in PREC cells ( Figure 5 ), we now suggest that lycopene inhibits DHT's anti-cell death effects by downregulation of IGF-I production in stromal cells and by attenuating IGF-I signaling in epithelial cells. In 6S/NPE co-cultures, lycopene effects on CM-induced NPE cell death mimicked those of anti-IGF-I and IGF-R antibodies ( Figure 1C and D ), which sequester and block IGF-I from NPE cells extracellularly, whereas lycopene functions intracellularly ( Figures 2 , 3 and 5 ).

Increasing IGF-I production in 6S cells may require increasing translocation of AR and β-catenin to nuclei. During DHT stimulation in AR-positive cells, AR and its co-activator, β-catenin, are translocated into cell nuclei to initiate gene transcription. The two proteins are associated with each other in nuclei of 3T3-L1 cells ( 15 ). In comparison, lycopene decreased AR protein expression but blocked AR nuclear translocation only slightly ( Figure 3A–D ), whereas lycopene reduced nuclear translocation of β-catenin induced by DHT ( Figure 3C and E ). These effects of lycopene on nuclear translocation of the AR and β-catenin may have accounted for the 50% reduction in IGF-I production observed in 6S cells ( Figure 2 ). Tomato lycopene extract was reported to decrease IGF-I levels in colon cancer patients ( 40 ). Perhaps, colon cancer cells treated with lycopene also exhibit decreased IGF-I by reducing AR and β-catenin nuclear co-localization. The mechanism may also explain why lycopene reduced PSA velocity ( 28 ), given that PSA is a target gene of DHT/AR/β-catenin signaling axis using a reporter assay in LNCaP cells ( 14 ).

The influences of the IGF-I/Akt/GSK3 signaling pathway on cell survival and cell proliferation have been extensively studied ( 17–20 ), and various mechanisms by which Akt/GSK3 contributes to cell functioning have been proposed ( 44–48 ). In the current study, we used low concentrations (0.15 or 1.0 ng/ml) of IGF-I, similar to those produced by 6S stromal cells in response to DHT ( 8 ). Using this paradigm, we detected substantial IGF-I-mediated cell death rescue and cell growth ( Figure 4 ), and phosphorylation changes of Akt and especially of GSK3, at the higher IGF-I concentration. However, we observed that lycopene inhibited IGF-I-induced cell growth, maintained Akt phosphorylation (active form) similar to or lower than that of control-treated cells and maintained GSK3β Ser9 phosphorylation (inactive form) or GSK3 Tyr(279/216) phosphorylation (active forms) similar to that of control-treated PREC cells, independent of IGF-I treatment ( Figures 4 and 5 ). IGF-I at low concentrations might exert minimal effects on Akt phosphorylation. There are three forms of Akt in mammalian cells, and the antibody recognizes only one form of Akt phosphorylation. IGF-I, at the concentration used, affected phosphorylation of GSK3, one of the Akt substrates, whereas lycopene counteracted the IGF-I effects or maintained GSK3 activities in PREC cells.

Although administration of lycopene alone increased nuclear β-catenin in 6S stromal cells ( Figure 3A ), it is possible that lycopene reduces β-catenin signaling in vivo . In AR-positive cancer-derived prostate stromal (6S) cells, in the presence of DHT, lycopene decreased β-catenin nuclear localization ( Figure 3C and E ). In AR-negative epithelial (PREC) cells, lycopene maintained GSK3 active forms in the presence of IGF-I ( Figure 5A and D ), and there should be less β-catenin accumulation than that in the cells treated with IGF-I alone. One possible mechanism by which lycopene alone increased nuclear β-catenin in 6S stromal cells might be via its effects on low density lioprotein (LDL) receptor-related protein 5/6. Lycopene increases macrophage LDL receptor activity ( 49 ) suggesting that lycopene can associate with the LDL receptor. One of the Wnt co-receptors is LDL receptor-related protein 5/6 ( 50 ). Thus, lycopene may also associate with the receptor and may affect the protein associations in 6S cells. Altering protein associations in Wnt-signaling pathway may mediate β-catenin stabilization ( 39 ). Although the effects of lycopene on β-catenin signaling in vivo remain uncertain, our data ( Figures 3 and 5 ) suggest that lycopene may reduce the in vivo signaling.

In conclusion, this study provides novel information regarding two issues. First, it establishes a role for prostate stromal cells in regulating the actions of DHT and IGF-I on prostate epithelial cell proliferation and death, and highlights the significant functional differences between the effects of cancerous versus normal stroma on normal epithelial cells based on differences in DHT-modulated IGF-I production. Second, our data suggest that lycopene inhibits DHT- and IGF-I signaling by decreasing IGF-I expression and by attenuating IGF-Is’ effects on serine phosphorylation (pSs) of Akt and GSK3β and tyrosine phosphorylation (pYs) of GSK3 ( Figure 6 ). The in vivo effects of lycopene on human prostate cell growth and function, and IGF-I signaling, require further study.

Fig. 6.

Schematic model depicting lycopene's inhibition of DHT-induced IGF-I production in 6S cells and attenuation of IGF-I effects on Akt/GSK3 phosphorylation in PREC cells.

Funding

The Intramural Research Program of the National Center for Complementary and Alternative Medicine, National Institutes of Health, Bethesda, MD 20892.

Abbreviations

Abbreviations
 
• AR

  androgen receptor

 
• CM

  camptothecin

 
• NPE

  normal prostate epithelial

 
• PSA

  prostate-specific antigen

The authors thank Dr Irini Manoli and Dr Vernon Steele for their constructive critiques of the manuscript.

Conflict of Interest Statement : None declared.