The Prostate 45:304–314 (2000)
Alterations of Prostate Biomarker Expression and Testosterone Utilization in Human LNCaP Prostatic Carcinoma Cells by Garlic-Derived S-Allylmercaptocysteine John Thomas Pinto,1,4* Changhong Qiao,1 Jie Xing,1 Brian P. Suffoletto,1 Kristin B. Schubert,1 Richard S. Rivlin,1,4 Robert F. Huryk,2,3 Dean J. Bacich,2,3 and Warren D.W. Heston2,3 1
Nutrition Research Laboratory, Memorial Sloan-Kettering Cancer Center, New York, New York 2 Urology Research Laboratory, Memorial Sloan-Kettering Cancer Center, New York, New York 3 George M. O’Brien Urology Research Center, Memorial Sloan-Kettering Cancer Center, New York, New York 4 Clinical Nutrition Research Unit, Memorial Sloan-Kettering Cancer Center, New York, New York BACKGROUND. This study determined the effects of S-allylmercaptocysteine (SAMC), a phytoconstituent from garlic, on the expression of androgen-responsive biomarkers, prostate specific antigen (PSA), and prostate specific membrane antigen (PSMA), in human prostatic carcinoma cells (LNCaP). METHODS. Secretion of PSA was determined as well as the activity of PSMA measured as a function of its ability to hydrolyze poly-␥-glutamated folate and N-acetylaspartylglutamate (NAAG). Folate hydrolase capacity was also determined in SAMC-treated cells grown in charcoal stripped fetal calf serum (CS-FCS). In addition, testosterone disappearance was measured from culture media of SAMC-treated LNCaP and PC-3 cells as well as from cell free lysates. RESULTS. PSA secretions were significantly decreased compared to control values at 1 day (8.4 ± 2.6 vs. 18.9 ± 1.7, P < 0.01), 4 days (18.9 ± 5.3 vs. 73.8 ± 4.4, P < 0.001), and 6 days (35.6 ± 2.1 vs. 96.5 ± 17.9 ng/105 cells, P < 0.01; mean ± SD). By contrast, PSMA activity measured
Abbreviations used: BCA, bicinchonic acid; CS-FCS, Charcoalstripped fetal calf serum; FCS, Fetal calf serum; LNCaP, human adenocarcinoma prostatic cell line isolated from lymph node; MTXglu3, Methotrexate triglutamate; NAAG, N-acetyl-␣-aspartylglutamate; NAALADase, N-acetylated alpha-linked acidic dipeptidase; PAP, Prostatic acid phosphatase; PBS, phosphate buffer saline; PSA, Prostate specific antigen; PSMA, Prostate specific membrane antigen; SAC, S-allylcysteine; SAMC, S-allylmercaptocysteine. Contract grant sponsor: Clinical Nutrition Research Unit; Contract grant number: DK/CA 47650; Contract grant sponsor: National Institutes of Health; Contract grant number: CA 29502. *Correspondence to: Dr. John Thomas Pinto, Memorial SloanKettering Cancer Center, 1275 York Avenue Box 140, New York, NY 10021. E-mail:
[email protected] Received 27 December 1999; Accepted 16 June 2000
© 2000 Wiley-Liss, Inc.
SAMC and Expression of Prostate Biomarkers
305
as either folate hydrolase or NAAG dipeptidase (NAALADase) activity increased in cells treated with SAMC. PSMA-folate hydrolase activity in SAMC-treated cells grown in CS-FCS increased beyond that observed in cells grown in CS-FCS alone. Pre-exposure of LNCaP cells to SAMC resulted in enhanced rate of testosterone disappearance from culture media at 6 hr (P < 0.01) and at 48 hr (P < 0.001) compared to media from cells not previously exposed to SAMC. Results similar to these were also observed in androgen-independent PC-3 cells treated with SAMC. In lysates of SAMC-treated LNCaP cells, the rate of testosterone catabolism was twice that from phosphate buffered saline (PBS)-treated cells. SAMC-treated LNCaP cells grown in media supplemented with testosterone temporarily exhibited enhanced growth over a 2 day period but cell numbers declined later to levels similar to those of SAMC treatment. CONCLUSIONS. These results show that SAMC exhibits differential effects on recognized biomarkers for LNCaP cells similar to those produced by androgen deprivation and strongly suggests that this effect may be mediated, in part, by diminishing the trophic effects of testosterone, likely by converting it to metabolites less reactive toward androgen receptors. Prostate 45:304–314, 2000. © 2000 Wiley-Liss, Inc. KEY WORDS:
allylsulfides; testosterone; PSA; PSMA; folate hydrolase; N-acetyl-␣aspartylglutamate
INTRODUCTION Epidemiological studies, as well as investigations using animal and cell culture models, have shown that allium compounds derived from garlic display potent antitumor activities [1–3]. In particular, the risk of prostate cancer shown in a 328 man, case-controlled epidemiologic study of diet and prostate cancer was significantly less in individuals who regularly consumed garlic food items and garlic supplements [4]. Although the molecular basis for their anticancer effect have not been precisely identified, several allium derivatives have been shown to induce a number of enzymes associated with phase I and II metabolism of xenobiotic compounds [5–7] as well as down-regulating other enzymes that enhance risk from carcinogen exposure and/or are associated with modifying signal transduction pathways [8–15]. Previous investigations by our laboratory on human prostatic carcinoma cells have focused on the antiproliferative activity of two water-soluble allium compounds, S-allylcysteine (SAC), and S-allylmercaptocysteine (SAMC) [16]. Of these two derivatives, SAMC has been shown to produce a marked and prolonged antiproliferative effect on the human prostatic carcinoma cell line (LNCaP). LNCaP cells were selected for our in vitro studies on allylsulfide derivatives. Although LNCaP cells are poorly differentiated human prostatic epithelia [17], they are responsive to androgens, and express cell markers similar to those of normal prostatic epithelial cells, namely prostatic acid phosphatase (PAP), prostate specific antigen (PSA), and prostate specific membrane antigen (PSMA) [18].
Recent advances in early diagnosis of prostate cancer, the extent of tumor progression and metastasis, as well as the determination of treatment efficacy have all involved monitoring these biomarkers [19]. Thus, it is important to study which factors affect expression of prostate biomarkers and to explore those mechanisms by means of which allium derivatives affect prostatic cancer proliferation [20]. Prostatic secretions of PAP and PSA vary during prostate cancer progression. Unlike PAP, PSA reasonably correlates with tumor burden and serves as a prognostic indicator for metastatic disease [21]. Studies in LNCaP cell cultures have shown that testosterone exposure increases PSA secretion [22] but has little or no effect upon enhancing PAP secretion [23]. Conversely, media depleted of testosterone [23] or supplemented with testosterone antagonists such as finasteride [24,25] results in decreases in the expression of these biomarkers by LNCaP cells and consequently their secretion into culture media. By contrast to its effect on expression of PSA, testosterone deprivation or use of the nonsteroidal, antiandrogen, casodex [26], causes an increase in PSMA expression as determined by levels of mRNA and protein [27,28]. The underlying significance of increased expression of PSMA by testosterone-deprivation requires elucidation. By contrast to PSA, which is secreted into plasma, PSMA is an integral protein of the prostate cell membrane and has been identified by us to be a unique folate hydrolase, capable of hydrolyzing progressively gamma-linked glutamate moieties from poly-␥glutamated folate and pteroyl derivatives [29]. The uniqueness of this enzyme is further exemplified by its
306
Pinto et al.
Fig. 1. Structural formula for the water soluble constituent of aged garlic, S-allylmercaptocysteine (SAMC). This compound is formed by reaction of allicin with the amino acid, cysteine. Cys = −CH2−CH (NH2) COOH.
capacity to hydrolyze an ␣–linked glutamyl moiety from the neural dipeptide transmitter, N-acetylaspartylglutamate, thus expressing N-acetylated ␣–linked acidic dipeptidase (NAALADase) activity [30]. PSMA is a type II transmembraneous glycoprotein that has a molecular weight of 100,000 kD and has recently been designated as a glutamate carboxypeptidase II (EC 3.4.17.21) by the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology [31, see review 32]. Testosterone metabolism occurs through the mixed function oxidase cytochrome P450 system (1A1, 1A2, 3A3, 3A4, and 3A5) [33] and allium derivatives are known to regulate the activity of one or more of these enzymes [5,6,8,9]. The balance of testosterone metabolites generated through the combined action of constitutive and inducible P450 enzymes may represent a critical mechanism in the development of prostate cancer. The present investigation examined whether the antiproliferative effect of SAMC on LNCaP cells and, subsequently, the expression of distinct testosteroneresponsive prostate biomarkers, namely PSA and PSMA, may be mediated, in part, by diminishing the trophic effects of testosterone, and if so by possibly converting it to metabolites less reactive toward androgen receptors. MATERIALS AND METHODS Chemicals and Reagents S-allylmercaptocysteine (SAMC) was generously supplied by Wakunaga of America Co., Ltd. (Mission Viejo, CA). The structural formula of SAMC, a cysteinyl disulfide, is illustrated in Figure 1. This derivative can be formed in vivo from the interaction of the garlic constituent, allicin, and the amino acid, cysteine. Methotrexate triglutamate (4-NH2-10-CH3-PteroylGlu4 [MTXglu3]), a substrate for PSMA-folate hydrolase, was purchased from Dr. B. Schircks Laboratories (Jona, Switzerland) and analyzed by HPLC to be >98%
pure. N-acetyl-␣-aspartylglutamate (NAAG), the substrate for NAALADase, was purchased from Sigma Chemical Company (St. Louis, MO). Radiolabeled NAAG (NAAG, glutamate-3,4-3H; 49.4 Ci/mmol) was purchased from New England Nuclear™ Life Science Products, Inc. (Boston, MA) and used to monitor PSMA-NAALADase activity. AG 1-X8 (formate) resin was obtained from Bio-Rad Laboratories (Melville, NY). All other reagents were obtained from Sigma Chemical Co. Sep-Pak威 Plus C-18 Cartridges for HPLC separation of testosterone were purchased from Waters Corporation (Milford, MA). Growth Conditions of Human Prostate Adenocarcinoma Cells (LNCaP and PC-3) LNCaP cells are poorly differentiated lymph nodederived human prostatic epithelia [17]. The rationale for using these cells is that they are responsive to androgens and express cell markers similar to those of normal prostatic epithelial cells. PC-3 prostatic carcinoma cells are highly transformed and lack an androgen receptor [34]. Although they exhibit low activities of testosterone 5 ␣-reductase, they do not express either PSA or PSMA. Thus, these biomarkers could not be monitored in PC-3 cells after treatment with SAMC. This cell line, however, was used to examine the ability of androgen-receptor negative cells to metabolize testosterone after exposure to SAMC. LNCaP and PC-3 cells were maintained in RPMI1640 medium supplemented with non-essential amino acids, 5 mM glutamine, and 5% heat-inactivated fetal calf serum (FCS). Cells were routinely passaged when they were 70–80% confluent in either flasks or plates. In studies to determine effects of testosterone depletion after SAMC treatment, both LNCaP and PC-3 cell lines were used. In addition, to examine effects of testosterone depletion after SAMC treatment of cells, heat-inactivated FCS was replaced with charcoalstripped FCS. Cells were plated in T-25 tissue culture flasks containing 5 mL of medium and incubated at 37°C in a humidified atmosphere of 5% CO2. Cell numbers were determined using a Model Z F Coulter Counter (Beckman-Coulter Electronic, Inc., Brea, CA). Prostate cells were harvested from plates by either trypsinization or gentle scraping into phosphate buffered saline (136.9 mM NaCl, 2.68 mM KCl, 8.10 mM Na2HPO4, 1.47 mM KH2PO4, pH 7.34, PBS) and centrifuged at 500g to obtain a cell pellet. Sedimented cells were routinely rinsed twice with 15 mL volumes of PBS. Treatment of LNCaP and PC-3 Cultures with SAMC and Testosterone Equal numbers of cells (2 × 104) were seeded in T-25 culture flasks. Twenty-four hours later, cells were
SAMC and Expression of Prostate Biomarkers treated with 50 L of SAMC dissolved in sterile PBS (final culture concentration, 260 M). Controls were treated with a similar volume of PBS. Following addition of SAMC, cultures were typically re-incubated at intervals from 1 to 7 days. Previous studies on growth of LNCaP cells by our laboratory [16] demonstrated that SAMC in the dose range of 50–300 M is growth inhibitory without producing appreciable cytotoxicity. From the perspective of determining the effectiveness of chemopreventive agents on secretion of PSA, it is important to use a dose that is growth inhibitory and not cytotoxic. For this reason, we chose the dose of 260 M since consistent growth inhibition of LNCaP cells occurs in culture for at least 1 week [16]. At the end of each experimental period, media from SAMC-treated cells and PBS-treated controls were removed and stored frozen at −20°C for PSA determinations (LNCaP cells only). Adherent cells were carefully rinsed twice with PBS to remove dead and dying cells and then trypsinized (0.025% trypsin-EDTA) to re-suspend cells. Aliquots of each cell culture were obtained and cell numbers determined using a Model Z F Coulter Counter (Beckman-Coulter Electronic, Inc.). For PSMA-hydrolase determinations, prostate cells (LNCaP cells only) were harvested from plates by scraping at 4°C into 50 mM Tris buffer pH 7.4 and centrifuging at 500g to obtain a cell pellet. All studies were performed in triplicate and SAMC was only administered to cultures once after cells had re-established adhesion to culture flasks (day 0). To determine whether testosterone can reverse the growth inhibitory effect of SAMC, LNCaP cells were grown in regular media supplemented with 100 nM testosterone in the presence and absence of SAMC (260 M) for 6 days. This concentration of testosterone combined with that in media containing 5% regular serum is approximately five times total plasma testosterone concentration of 20 nM [35]. LNCaP cells were seeded in six-well plates and allowed 24 hr to reattach. On day 0, four treatment groups were established with five samples per group: PBS-treated, PBS plus testosterone (100 nM), SAMC-treated (260 M), and SAMC plus testosterone. Both testosterone and SAMC were added on day 0. Cells were counted on day 0, 1, 3, 4, 5, and 6. This experiment was performed twice. Radioimmunoassay for PSA PSA values were determined using radioimmunoassay kits purchased from Hybrid tech, Incorporated (San Diego, CA). Assays were performed following the manufacturer’s recommended procedure. Aliquots of media were removed after 1, 4, and 6 days in
307
culture with SAMC, cellular debris was removed by centrifugation, and the clear supernatant fraction was stored at −20°C until analysis. PSA values were reported relative to number of viable cells.
Preparation of Cell Membranes and Determination of PSMA-Hydrolase Activity Cells were sonicated in 50 mM Tris pH 7.4 buffer (2 × 10 sec pulses at 20 mWatts) in an ice-bath. Membrane fractions were obtained by centrifuging lysates at 100,000g for 30 min. The supernatant fractions were discarded and pelleted membranes were resuspended by gentle trituration and re-sedimented at 100,000g for 30 min through 10 mL of cold 50 mM Tris pH 7.4 buffer. Washed membrane fractions were dissolved in 50 mM Tris pH 7.4 buffer containing 0.1% Triton X-100 (Tris/Triton). PSMA-hydrolase activity was determined by measuring both folate hydrolase and NAALADase reactions using this membrane preparation [29,36]. Since PSMA-hydrolase acts upon both pteroyl polygammaglutamates (conjugated folates) and the excitatory neurotransmitter, N-acetylaspartylglutamate (NAAG), both reactions were monitored in these studies: In brief, folate hydrolase activity was determined using capillary electrophoresis as described earlier [29]. The standard assay mixture contained 50 M MTXGlu3, 50 mM Tris-0.5% Triton X-100 buffer (pH 7.4), and enzyme (2–5 g protein) to a final volume of 100 L. Reactions were conducted for times varying between 0 and 60 min at 37°C and were terminated in a boiling water bath. Samples were stored frozen (−20°C) until analysis by capillary separation of MTXGlun analogues using a Spectra Phoresis 1000 instrument (Thermo Separation, San Jose, CA) and absorbance monitored at 300 nm. Data were recorded with an IBM computer using CE-1000 software (Thermo Separation) and reported as mol MTXglun hydrolyzed per hr per mg protein. N-acetylated ␣–linked acidic dipeptidase (NAALADase) activity was measured using the method previously reported [36]. Reactions were initiated by the addition of tritiated NAAG into 50 mM Tris-HCl pH 7.4 buffer containing ∼2–5 g of LNCaP membrane protein (100 L total volume). Reaction mixtures were incubated for 20 min at 37°C and stopped with addition of ice-cold 0.1 M sodium phosphate buffer (pH 7.4). Aliquots (100 L) of the stopped reaction mixture were applied to minicolumns packed with AG 1-X8 anion exchange resin and (3H) glutamate was eluted with 1.0 M formic acid. Radioactivity was measured using scintillation spectrometry and data reported as nmoles glutamate hydrolyzed per min per mg protein.
308
Pinto et al.
Testosterone Supplementation and Determination The effects of SAMC on testosterone concentrations were initially examined in LNCaP cells pre-treated for a 7 day period with 500 M SAMC. We chose this higher dose and time schedule to allow LNCaP cells sufficient time in culture to adapt or modify endogenous activities of microsomal and/or phase II enzymes shown previously to respond to allium compounds [5,6]. Control cells were grown in media containing an equivolume of PBS. On day 8, both SAMC- and saline-treated cells were washed free of their respective culture media, trypsinized, and equinumber of cells (1 × 106) were re-cultured in 100 mm plates containing fresh media in the absence of SAMC or PBS vehicle. LNCaP cells were allowed to re-attach (24 hr) before cultures were supplemented with testosterone (2 g/mL). This supra-physiologic concentration of testosterone was selected early in our studies to be certain that the disappearance rate of testosterone could be accurately assayed within the limits of the methodology employed, particularly at the very low concentrations to be expected at the longer time points in SAMC-treated cells. Subsequent studies used 100 nM testosterone which represent five to six times the physiological dose and based on controlled studies did not exhibit discernible adverse effects on growth of LNCaP cells during the time course of 1 to 6 days in culture. At 3, 6, and 48 hr posttestosterone administration, aliquots of media were removed, extracted with chloroform/methanol (2:1 v/v), and analyzed for concentration of testosterone remaining in the media. Plates incubated over the same time intervals and containing media supplemented with testosterone in the absence of cells were carried along to determine recoveries and extraction efficiencies. Following centrifugation at 1,000g, the organic epiphase was removed, evaporated to dryness under nitrogen and re-dissolved in absolute methanol. Following a 1:3 dilution of methanolic fractions with the starting mobile phase (methanol-water-acetonitrile (80:18:2, v/v/v), 10 l aliquots were injected onto an Ultrasphere 5 , 4.6 × 250 mm, C18 column and eluted using a gradient mobile phase of methanol-wateracetonitrile (80:18:2 to 39:60:1) [37] at a flow rate of 1 mL/min. Peak areas were monitored using ultraviolet detection at 256 nm and areas under the concentration curves were analyzed using Perkin-Elmer software. The final concentrations of testosterone are reported as ng testosterone per 105 cells. In subsequent studies, we determined that significant and reproducible acceleration of the testosterone removal rate could be achieved after 4 days in culture with 260 M SAMC. We hypothesize that this effect may represent induction of testosterone metabolizing
and/or conjugating enzymes. Accordingly, both LNCaP and PC-3 cells were cultured for 4 days in the presence and absence of 260 M SAMC, harvested and cell free lysates prepared by brief sonication (2 × 10 sec pulses at 20 mWatts). Following centrifugation at 10,000g to remove debris, supernatant fractions were incubated at 30°C in the presence of 100 M NADPH and 1 g of testosterone. In addition, known concentrations of testosterone were added to cell lysates without incubation, extracted at time 0 and testosterone determined. After a 30 min incubation, the reaction was stopped by the addition of 2:1 mixture of chloroform/methanol and testosterone analyzed as described above. Data are reported as ng testosterone per mg protein. Protein Determination Protein concentrations of isolated membrane or supernatant fractions were determined by incubating diluted aliquots with BCA reagent (Pierce Chemical Co., Rockford, IL) at 37°C for 30 min. The spectrophotometric quantitation of protein was conducted by determining the absorbance at 562 nm against bovine serum albumin standard. Statistical Analysis Data were analyzed by using the Statgraphics version 4.0 program (Statistical Graphics Corporation, Rockville, MD). Data are expressed as mean ± SD. Dunnett’s t-test was used to compare the significance of differences between PBS-treated controls and SAMC-treated cells. RESULTS Growth Arrest of LNCaP Cells by SAMC Previous studies using the water-soluble allium derivative, SAMC, demonstrated that growth of LNCaP cells in SAMC-treated media is markedly inhibited after 1 to 3 days in culture and that cell proliferation remains arrested after replacement with fresh media for at least 4 days. The present study extends these observations and examines the effect of SAMC administration on the expression of recognized biomarkers of LNCaP cells in culture and on testosterone metabolism in both LNCaP and PC-3 cells. Since cell population densities as well as growth rates can influence secretion of prostate biomarkers, cultures were typically harvested before they reached 70 to 80% confluence representing the log phase in LNCaP cultures.
SAMC and Expression of Prostate Biomarkers
Fig. 2. Effects of SAMC on prostate specific antigen (PSA) secretion by LNCaP cells. LNCaP cells were plated at a density of 2 × 104 in T25 flasks and treated on the following day with SAMC (260 µM). On day 1, 4, and 6 post-SAMC treatment, aliquots of media were removed and analyzed for PSA using radioimmunoassay. Cells were counted and PSA levels normalized to ng PSA/105 cells. Data are expressed as mean ± SD from triplicate determinations.
Effect of SAMC on PSA Secretion The accumulation of PSA secretions (ng/105 cells; mean ± SD) into culture media was approximately linear with time for up to 6 days. In cultures treated with SAMC (260 M), PSA secretions were significantly reduced compared to corresponding controls at 1 day (P < 0.01), 4 days (P < 0.001), and 6 days (P < 0.01; Fig. 2). PSA concentrations were standardized relative to number of viable cells and the decrease in secretion was disproportionately greater compared to the reduction in LNCaP cell growth, suggesting either diminished expression of PSA per cell or diminished rates of secretion.
309
with the known effects of androgen deprivation on expression of these biomarkers, we next examined PSMA activity in LNCaP cells grown in media containing 5% FCS and 5% charcoal-stripped FCS both in the presence and absence of SAMC. Figure 3 illustrates folate hydrolase activity and the sequential hydrolysis of the gamma-linked triglutamated MTX (MTXglu3) with subsequent formation of MTXglu2, MTXglu1, and MTX after 60 min of incubation. As PSMA-folate hydrolase exhibits exopeptidase activity that successively cleaves the terminal glutamate, it is important to note that each subsequent polyglutamated MTX derivative becomes a substrate for this reaction until MTX is formed. Since MTX is the final product of the reaction, the significance of difference between PBS controls and SAMC treatments was determined using only this fraction. Thus, PSMA-folate hydrolase progressively liberates all of the possible glutamates from MTXGlu3 with accumulation of MTX. Results in Figure 3A illustrate that SAMC evokes approximately a two-fold increase in PSMA-folate hydrolase activity when compared to values in PBStreated controls. LNCaP cells grown in media containing charcoal-stripped FCS alone (Fig. 3B) show a pattern of folate hydrolase activity comparable to that observed in SAMC-treated cells grown in regular 5% FCS (Fig. 3A). Cells grown in 5% charcoal-stripped FCS and SAMC exhibit a further increase in hydrolase activity over cells grown in 5% charcoal-stripped serum alone (P < 0.05; Fig. 3B). SAMC Increases Testosterone Removal from Culture Media of LNCaP Cells Since LNCaP cells have mutated androgen receptors but still respond to testosterone, we examined the rate of testosterone removal from media by cells cul-
Effect of SAMC on PSMA-Hydrolase Activity Since PSMA exhibits reactivity towards polygammaglutamated folates as well as towards NAAG, both reactions, i.e., folate hydrolase and NAALADase, were monitored following treatment of LNCaP cells with SAMC. By contrast to PSA formation, the hydrolytic activity of PSMA measured as nmoles glutamate formed from NAAG significantly increased over time. Table I illustrates the increase in PSMA-NAALADase activity in isolated plasma membranes to approximately three-fold after SAMC treatment at 1 day and 4 days (P < 0.001) and approximately four-fold at 6 days over control values (P < 0.001). Because an increase in PSMA activity and a concomitant decrease in PSA secretion are in accordance
TABLE I. PSMA-NAALADase Activity in LNCaP Cells Exposed to SAMC for 1, 4, and 6 Days* NAALADase activity (nmol glutamate ⭈ min−1 ⭈ mg−1 protein) Treatment Day 1 4 6
PBS
SAMC
P value
38 ± 17 56 ± 13 71 ± 20
110 ± 25 181 ± 30 296 ± 36
