Clinical & Experimental Metastasis 17: 307–314, 1999. © 1999 Kluwer Academic Publishers. Printed in the Netherlands.
307
Role of neurotrophins and neurotrophin receptors in the in vitro invasion and heparanase production of human prostate cancer cells E. Timothy Walch & Dario Marchetti Section of Molecular Cell Biology, Department of Cancer Biology, Box 108, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX , USA Received 22 December 1998; accepted in revised form 28 April 1999
Key words: heparanase, invasion, neurotrophins, prostate cancer, p75 neurotrophin receptor (p75NTR)
Abstract The role of the neurotrophins (NTs) and their corresponding receptors (NTRs) TrkA, TrkB, TrkC, and p75NTR in neoplasia has received relatively little attention. However, because malignant cell migration within the prostate occurs predominantly by direct extension around prostatic nerves, the presence and possible upregulation of NTs from autocrine/paracrine sources and NTR expression within prostate epithelial tumor cells may be important in metastasis. We have been addressing their expression and interactions in human prostate cancer cell lines (LNCaP, PC-3, and DU145) and their role in prostate cancer invasion. In this study, we demonstrated that nerve growth factor (NGF), the prototypic NT, and NT-4/5 increased in vitro invasion through a reconstituted basement membrane and induced time- and dose-dependent expression of heparanase, a heparan sulfate-specific endo-β-D-glucuronidase, an important molecular determinant of tumor metastasis. The NT effects were most marked in the DU145 brain-metastatic cells and were detected at NT concentrations sufficient to fully saturate both low- and high-affinity NTRs. Additionally, we characterized the molecular expression of NT high-affinity (Trk) and low-affinity (p75NTR) receptors in these cell lines by reverse transcription-polymerase chain reaction. These lines had negligible trkA and trkC expression, although trkB was expressed in the three prostatic tumor cell lines examined. The brain-metastatic DU145 cells were also positive for p75NTR. Our data showed that the NTs and NTRs are important in metastasis and that their expression coincides with transformation to a malignant phenotype capable of invasion along the perineural space and extracapsular metastasis to distant sites. These findings set the stage for more research into this area as related to prostate cancer evolution and may improve therapy for prostate cancer metastasis.
Introduction Prostate cancer has become the most common malignant neoplasia in men in North America [1]. Because the incidence of prostate carcinoma increases rapidly after age 50 [2] and the average age of male adults is increasing, the incidence of prostate cancer will also increase during the next several decades [2]. Understanding the molecular mechanisms that regulate prostate cancer growth and metastasis may help us develop clinical stategies for the treatment and control of the disease. The growth of the prostate is regulated by humoral factors such as steroids, extracellular matrix components and autocrine/paracrine growth factors, all of which may be interdependent to varying degrees. The consensus view from a large number of studies [1–8] is that paracrine growth factors secreted by stromal cells and autocrine growth factors secreted by epithelial cells mediate the development and proliferation of normal and neoplastic human prostate tissue. One of these factors is nerve growth Correspondence to: Dr Dario Marchetti, Department of Neurosurgery, UTHouston Medical School, Suite 7.136, 6431 Fannin St., Houston, TX 77030, USA. Tel: +1-713-500-6137; Fax: +1-713-500-7787; E-mail:
[email protected]
factor (NGF) [9]. The mature form of NGF, β-NGF, is relatively abundant within the prostates of several species, including humans [9–14]. Recently NGF was confirmed to be the prototypic member of a family of molecules called the neurotrophins (NTs). NTs are a group of specific neurotrophic factors distinct from the neuropoietic cytokines (i.e., interleukin-6 and ciliary neurotrophic factor) and the acidic/basic fibroblast growth factors. Presently the family of human mammalian NTs consists of four structurally and functionally related proteins: NGF, the prototypic NT; brain-derived neurotrophic factor (BDNF); neurotrophin-3 (NT-3); and neurotrophin-4/5 (NT-4/5). The NTs can be distinguished based on their distinct patterns of spatial and temporal expression as well as their different effects on neuronal targets [15–17]. The biological effects of NTs are mediated through two unrelated classes of cell-surface membrane receptors. All NTs interact with a transmembrane protein without a direct catalytic function, p75NTR [18, 19], as well as with distinct members of a subfamily of receptor tyrosine kinases known as Trks [20]. NGF interacts with TrkA, BDNF and NT-4/5 with TrkB, and NT-3 with TrkC. In addition to their normal physiological roles, several NTs and their corresponding re-
308 ceptors have been shown to induce a variety of pleiotropic responses in malignant cells, including enhanced tumor invasiveness and chemotaxis. Specifically, we observed that NTs affect the metastasis of human malignant melanoma to the brain, implicating NTs and their receptor subtypes in this metastatic cancer through paracrine mechanisms involving brain-invasive melanoma cells and the adjacent brain and stromal tissues [21–25]. When we focused our attention on studying invasion-related properties of melanoma cells as regulated by previous incubation with NTs, we found that exposure to selected NTs resulted not only in an enhanced ability of brain-metastatic melanoma cells to penetrate a basement membrane-like matrix in vitro but, importantly, in an increased activity of the matrix-degrading heparanase. Heparanase is an endo-β-D-glucuronidase, the main molecular determinant of such NT-driven matrix degradation [21, 26-28] and an important enzyme associated with tumor metastasis as previous studies have indicated [21]. Prostate cancer metastasis occurs predominantly by direct extension around prostatic nerves. This perineural space invasion follows the course of nerve branches and so allows prostate cancer to spread outside the prostatic capsule. Consequently, the provocative hypothesis that nerves may be attracted by putative growth factors elaborated by the epithelium was formulated [10, 29]. Since NGF production is controlled by androgen [30] and exogenous β-NGF, the mature and biologically active form of NGF, has been shown to stimulate the growth of prostate epithelial cells [31, 32], NGF has been considered a potential candidate responsible for these factors-related influences and causing nerves to exhibit invasive-like properties. Because prostate cancers can be aggressive even when small, there is a need to define how they invade the basement membrane and to determine additional ways of detecting and treating them. In the study reported here, we measured the levels of p75NTR and the various Trk receptors by reverse transcription (RT)-polymerase chain reaction (PCR) in the human prostatic cancer cell lines LNCaP, PC-3, and the brain-metastatic variant DU145 and the effects of NTs on the invasion and heparanase production in these cell lines by using newly developed and specific assays.
Materials and methods Tumor cells and culture conditions The androgen-responsive human prostate adenocarcinoma cell line, LNCap, the brain-metastatic prostate cancer cell line DU145, the androgen-independent human prostate carcinoma cell line PC-3, the human neuroblastoma cell line SH-SY5Y, and the human uterine adenocarcinoma cell line RL95 were purchased from American Type Culture Collection (Manassas, VA). PC-3 cells and their metastatic variants are recognized as a model of malignant human prostate carcinoma [41]. The human brain-metastatic cell line 70W was a generous gift of Dr Robert Kerbel (Sunnybrook Health Science Center, Toronto, Ontario, Canada) and maintained as previously described [21]. Human metastatic melanoma
E. Timothy Walch and Dario Marchetti derivatives of the cell line SK-MEL were generously provided by Dr Anthony Albino (Naylor Dana Institute for Disease Prevention, Valhalla, NY) and maintained as previously described [42]. MDA PCA-2a and -2b cells were a generous gift from Dr Nora Navone (Department of Genitourinary Medical Oncology, M. D. Anderson Cancer Center, Houston, TX). The tumor cell lines were cultured in 1:1 (v/v) Dulbecco’s modified Eagle’s medium (DMEM)/F-12 containing 10% FBS. They were used between passages 5 and 9 and passaged before reaching confluence by using Ca2+ and Mg2+ -free phosphate-buffered saline (PBS) containing 2 mM ethylendiaminetetraacetic acid (EDTA). The cells were cultured in a humidified 5% CO2 in air atmosphere in 1:1 DMEM/F-12 containing 5% FBS without antibiotics. All cell lines were periodically checked for Mycoplasma contamination with a Geneprobe kit (San Diego, CA), and only Mycoplasma-free cells were used in this study. Primers and RT-PCR analysis Total RNA was obtained from asynchronous cultures of the human prostatic carcinoma, melanoma, and neuroblastoma cell lines mentioned above by using TriReagent (Molecular Research Center, Cincinnati, OH). Total RNA was used as template for first-strand cDNA synthesis with Moloney murine leukemia virus reverse transcriptase and random hexamer primers (Perkin Elmer, Foster City, CA). Oligonucleotide primers (GIBCO BRL, Bethesda, MD) for PCR analysis of TrkA, TrkB, TrkC and p75NTR were selected to cross the intron boundaries so that only cell lines expressing the receptor would produce a signal of the proper size. Positive controls for RT-PCR were SK-MEL147 and 70W human melanoma cells for p75NTR transcript expression, SH-SY5Y human neuroblastoma cells for trkA/B transcript expression and 70W human melanoma cells for trkC transcript expression. The primers, salt concentration, and annealing temperature were those specified for the GeneAmp RNA PCR kit (Perkin Elmer). The amplification profile consisted of an initial template denaturation step at 94 ◦ C for 1 min followed by 35 cycles of 94 ◦ C for 45 sec, 57 ◦ C for 1 min, and 72 ◦ C for 1 min. The amplification products were analyzed on a 1% agarose/TBE gel by using a 100-bp ladder size standard (GIBCO BRL). The oligonucleotide primers (GIBCO BRL) for PCR analysis of the trkA receptor were selected to cross the intron boundaries between exons 11 and 12 of trkA, so that only cell lines that express this receptor would produce a specific 210-bp signal. The following oligonucleotide primers were used for human trkA: forward primer, 50 -TCCTTTCTACGCTGCTCCTT-30 and reverse primer, 50 -AGGCATCACAGAAGTATTGT-30. The primer sequences were based on the published genomic DNA sequence for trkA [44]. Since there are no the published genomic sequences for p75NTR, trkB, and trkC, the primers were specifically designed to cross the intron boundaries and produce correct-sized fragments as determined from a comparison of published cDNA sequences and genomic sample controls. The primers for PCR analysis of the other NTR and their positions were as follows: human p75NTR forward primer 50 -157bp-TGCTGCTGTTGCTGCTTCTG-30
Neurotrophins and prostate cancer invasion and reverse primer 50 -GTTCCACCTCTTGAAGGCTATG30 -941bp which were expected to produce a PCR signal of 784 bp; human trkB forward primer 50 -1382bpGAGCATCTCTCGGTCTATGC-30 and human trkB reverse primer 50 -GTCAAACCCTCTGAATTCAGTGCT-30 2055bp which were expected to produce a PCR signal of 673 bp; and human trkC forward primer 50 -1945bp-CATGAGCACATTGTCAAGTTC-30 and human trkC reverse primer 50 -ACCACTAGCCAAGAATGT-30 2637bp which were expected to produce a PCR signal of 692 bp [45]. Chemoinvasion assays and NT treatment The invasive ability of tumor cells after NT treatment was assayed by using Transwell cell culture chambers (Costar Inc., Cambridge, MA) and monitored by fluorescence plate scanner analysis with the Cytofluor 2350 (Millipore, Bedford, MA) [26]. Reconstituted basement membrane (MatrigelTM; Becton Dickinson Inc., Bedford, MA) was diluted 1:20 in cold DMEM/F-12 without phenol red (Sigma Chemical Co.) and applied ('20 µg/100 µl/filter) to the upper surface of each Transwell (8 µm pore size/6.5 mm diameter) polycarbonate filter insert. The lower chamber contained a mixture of 0.5% low-gelling agarose (FMC Bio-Products, Rockland, ME), 0.5% gelatin (Sigma Chemical Co.), 10 nM N-formylMet-Leu-Phe chemotactic peptide (Sigma Chemical Co.), and 25% fresh conditioned medium from MDA-PCA cells conditioned for 48 h. Both N-formyl-Met-Leu-Phe peptide and conditioned medium were placed in the lower chamber as chemoattractants, which significantly increased invasive values, being a possible source of paracrine motility signals. The total volume in the lower chamber was 380–390 µl, depending on the degree of gel drying. The bottom of each well (24-well plates, Costar) was previously coated with 5 µg/cm2 fibronectin (Becton-Dickinson Inc., Bedford, MA). Cells (3.0 × 104 /Transwell filter) suspended in serumfree, phenol red-free. DMEM/F-12 were seeded into the upper compartment. After a 72-h incubation at 37 ◦ C with purified human NTs (2–400 ng/ml), the upper chambers were carefully removed and the bottom chambers treated with 4 µM calcein-AM for 45 min at 37 ◦ C. Alternatively, a monoclonal antibody to human NGF (NGF-MAB, 20 µg/ml) or a recombinant NGF molecule unable to bind p75NTR (RH-NGF, 100 ng/ml) were incubated with cells into the upper compartments using the same conditions. After several washes, the plates were analyzed for fluorescence with a Cytofluor 2350 fluorescence plate scanner (BB4 settings), and the formation of calcein was monitored by measuring the increase in fluorescence at 530 nm. The accuracy of the invasion assay was assessed by replicate determinations of cell number, which were obtained by measuring specific fluorescence on the bottom of the Transwell filters after the assay was complete. Trypsin/EDTA (0.5% /1 mM, 0.6 ml of solution) was added to the bottom chamber, and incubation was allowed to proceed for 10 min at 37 ◦ C. To insure that all cells were removed, 100 µl of a trypsin/EDTA solution was placed on top of each filter for 10 min. Then,
309 the cover on each Transwell was removed, and the plate was sharply struck to assure that all cells detached from the membranes. The Transwells were also observed under a microscope to see whether all the cells had been removed. A second trypsinization step was usually not required. Once the cells rounded and detached, the trypsin/EDTA was removed from the top, bottom, and sides of Transwells and the membrane was quickly rinsed. The detached cells in the lower chamber were counted by the calcein-AM method with the Cytofluor 2350 [21]. The invasive profiles were determined by running appropriate controls without NGF. The ratios of arbitrary fluorescence units (A.F.U.) to cell number in a cellular growth assay were determined. The data from this assay therefore reflect the number of the cells that completely traversed the MatrigelTM -coated barrier after 72 h of incubation in the presence of NTs. Data were collected from repeated experiments to insure that the invasion assay was both reproducible and accurate. The percentages of invasive cells that attached to the lower side of the Transwells filters did not vary significantly among the cell lines studied. There was background fluorescence and a decrease in fluorescence when the cells were plated on the bottom versus the top. These control experiments helped us to quantitate background fluorescence. Only specific fluorescence (total minus background) was analyzed. We also performed a functional analysis of several batches of MatrigelTM and used only one in the invasion assays. Aliquots were kept frozen at −20 ◦ C until used. The data were statistically analyzed using Student t test and P values determined for each experimental condition. Heparanase activity assay Heparanase activity was determined by the degradation of purified cell-surface [35 S]heparan sulfate (HS) from human adenocarcinoma RL95 cells, the rationale being that RL95 are able to grow in sulfate-depleted media and most (>95%) of glycosaminoglycans (GAG) chains synthesized by these cells are HS [26], thus providing a convenient substrate to investigate heparanase action and heparanase regulation by NT. We subsequently measured heparanase activity by high-speed gel permeation chromatography as previously reported [26, 27], with some modifications. Briefly, subconfluent cells were harvested and solubilized in 50 mM Tris-HCl, pH 7.5, 0.05% NaN3 , and 0.5% Triton-X-100, 1 mM phenylmethylsulphonylfluoride (PMSF; Sigma Chemical Co.) and 5 mM N-ethylmaleimide (NEM; Sigma Chemical Co.) as protease inhibitors for 30 min at 4 ◦ C. The cell extracts were then centrifuged at 12 000 × g for 30 min at 4 ◦ C. Cell extracts (50–70 µg protein) were then incubated with 5 µg of radiolabeled HS and 2-[N-morpholino]ethanesulfonic acid (MES) buffer (Sigma Chemical Co.), pH 5.0, for 18 h at 37 ◦ C (10–100 µl final reaction volume). To avoid decreases in HS as a result of nonspecific binding to proteins, heparin solution (100 µg/ml final concentration) was added to each tube for 15 min. The reactions were then terminated by heating the samples for 15 min at 95 ◦ C and centrifuging them. A delipidation step was also applied to the cell extracts after the heparanase assay and before high-pressure liquid chro-
310 matography (HPLC) analysis. Supernatants (50 µl) were injected into a Waters 600E HPLC system (Milford, MA) equipped with a TSK gel G3000 PWX2 column (7.8 mm × 30 cm; 6 µm particle size) from Toso Haas (Montgomeryville, PA). An elution rate of 0.50 ml/min was used, and fractions were collected. The radioactivity in the fractions was measured with a liquid scintillation counter (LKBPharmacia, Uppsala, Sweden, 75% efficiency). Heparanase activity was defined as the decrease in the area of the high Mr half of the HPLC peak, indicating degradation of highmolecular weight HS to intermediate- and low-molecular weight components. Profiles of reaction mixtures without cell extract were always run as controls, both at the beginning and the end of HPLC analysis; these samples contained an equivalent volume of MES buffer or double-distilled water. Data displayed are means of 3 independent experiments with quadruplicate assays performed in each experiment. The data were statistically analyzed using Student t test and P values determined for each experimental condition as shown. Agarose gel electrophoresis of HS subpopulations Each reaction mixture consisted of 10–20 µl of heparanase enzyme source mixed with electrophoresis buffer [final concentrations of 0.025% (w/v) bromophenol blue, 0.025% (w/v) xylene cyanol, 2.5% (w/v) Ficoll (type 400) in doubledistilled H2 O] [26, 27]. Before electrophoresis, sodium dodecyl sulfate [a final concentration of 1% (w/v)] was added to each sample for 15 min at room temperature (25 ◦ C) to dissociate 35 S-labeled digestion products from molecular weight complexes consisting of radiolabeled glycosaminoglycans chains of known size [26]. Then, electrophoresis in 1.2% (w/v) agarose gel electrophoresis was performed at 75 V for 1 h at 25 ◦ C or until the samples migrated approximately two-thirds of the entire gel length. Autoradiography was performed on the dried gel by exposure to X-AR 5 film (Kodak, Rochester, NY) for 3–7 days. In the figure, the direction of electrophoretic mobility is from top to bottom.
Results Enhancement of in vitro invasion of prostate cancer cells by NGF The effect of the prototypic NT, NGF (in its mature form, β-NGF), on the invasive capacity of the LNCap, PC-3, and DU145 cell lines was determined by seeding cells onto MatrigelTM -coated filters in Transwell units [26, 46]. The brain-metastatic line DU145 was stimulated by NGF to invade the Matrigel barrier at a higher rate than the androgenresponsive line LNCap or the androgen-insensitive line PC-3 (Figure 1). The concentration of NGF that stimulated maximum invasion was 100 ng/ml, or approximately 4 nM (Figure 1). The stimulation of DU145 was biphasic; at higher concentrations of NGF (200 to 400 ng/ml), the invasion values decreased. All three cell lines were also unable to penetrate the filters in the absence of MatrigelTM during
E. Timothy Walch and Dario Marchetti
Figure 1. Effect of NGF on the invasion of the human prostate cancer cell lines LNCap, DU145 and PC-3 cells into Matrigel-coated filters. The cells invading the lower Transwell chambers were counted by a chemoinvasion assay involving the Cytofluor 2350 system (see ‘Materials and methods’). The cell numbers are related to specific fluorescence values. The remaining cells attached to the lower side of Transwell filters were detached from the membrane with trypsin/EDTA solution and quantified with the Cytofluor system. Bars represent the standard deviation of the mean ± SD of 4 independent experiments, with each experiment performed in quadruplicate. – ∗ P value < 0.01; P value < 0.06; P value < 0.001.
a 72 h incubation with NGF. Furthermore, when a monoclonal antibody to NGF was added to the upper chambers, the cells did not invade the lower chamber (Table 1). Invasion was also abolished in presence of a recombinant human NGF molecule unable to bind p75NTR [59] (Table 1). The data displayed were statistically significant and P values were determined for each experimental condition as shown (Figure 1, Table 1). Effects of other NTs on in vitro invasion of human prostate cancer cells To assess other members of the NT family, namely BDNF, NT-3, and NT-4/5, LNCap, PC-3, and DU145 cells were incubated with HPLC-purified and biologically active preparations of those human NTs at the optimal concentration of 100 ng/ml seen for NGF. Only NGF and NT-4/5 stimulated invasion, and only in DU145 and, to some extent, PC-3 cells (Figure 2). LNCap cells were not invasive in response to NT treatment. BDNF and NT-3 did not have any dramatic effects on the invasion patterns of the 3 cell lines studied. Invasion values was negligible when specific NT antibodies were present in the upper compartment of Transwell chambers, as seen for NGF (data not shown). Characterization of NTRs in prostate cancer cells by RT-PCR Because receptors generally occur in low abundance relative to their ligands, we analyzed the expression of p75NTR, trkA, trkB, and trkC by RT-PCR amplification of transcripts. Figure 3 shows that p75NTR amplification products of approximately 783 bp were detected in SK-MEL 147 cells,
Neurotrophins and prostate cancer invasion
311
Table 1. Invasion of MatrigelTM -coated filters in human prostate cancer cells: effect of NGF, NGF-MAB and RH-NGF. Cell type
Time (h)
−NGF
+NGF
+NGF-MAB
+RH-NGF
LNCaP
0 72 0 72 0 72
1.00 ± 0.01b 1.03 ± 0.5 1.00 ± 0.01 1.24 ± 0.3 1.00 ± 0.01 1.13 ± 0.2
1.02 ± 0.2 2.24 ± 0.7∗ 1.00 ± 0.1 2.92 ± 0.7∗ 1.00 ± 0.3 31.0 ± 9.3∗
1.00 ± 0.1 1.42 ± 0.3∗ 1.00 ± 0.5 1.13 ± 0.2∗ 1.00 ± 0.5 3.30 ± 0.2∗∗
1.03 ± 0.1 1.06 ± 0.1∗∗ 1.02 ± 0.1 1.46 ± 0.4∗ 1.03 ± 0.4 3.15 ± 0.1∗
PC-3 DU145
a Cells (3.0 × 104 /chamber) from the selected lines were plated into MatrigelTM -coated invasion chambers containing only serum-free medium (100 µl) or equivalent amount of serum-free medium with either 1) purified and biologically active human NGF (100 ng/ml) or, 2) a monoclonal antibody to human NGF (NGF-MAB, 20 µg/ml) coincubated with NGF or, 3) a recombinant human NGF unable to bind p75NTR (100 ng/ml). b Fluorescence units (FU) at time 0 were considered equal to 1.00. Data displayed are means of 3 independent experiments with quadruplicate assays performed in each experiment. – ∗ P value < 0.01; ∗∗ P value < 0.06.
Figure 2. Effects of NTs on the invasion of the human prostate cancer cell lines LNCap, DU145, and PC-3 into MatrigelTM -coated filters. The data are the means ± SD of 3 independent experiments, with each experiment performed in quadruplicate. – ∗ P value < 0.01; P value < 0.06; P value < 0.001.
which are known to contain elevated amounts of p75NTR [42], and, to a lesser extent, in DU-145. No amplification products for p75NTR were detected in PC-3 or LNCap cells (Figure 3). When we analyzed the same cell lines for amplification products corresponding to trkA, trkB, and trkC, there were differences in the expression of the various Trk members. Neither trkA or trkC were detected in LNCap, DU145, and PC-3, but a specific amplification products for trkB was observed (Figure 4). NT enhancement of heparanase production Because NTs may be involved in modulating a variety of aspects of tumor invasion including increased heparanase production [21–23, 25–27], we investigated heparanase expression and its regulation by NTs in LNCap, PC-3, and DU145. We assayed heparanase by HPLC and agarose gel
Figure 3. Expression of human p75NTR measured by RT-PCR in the prostatic cancer cell lines LNCap, DU145, and PC-3. RT-PCR was performed as described in Materials and Methods. Lane 1, 100-bp marker ladder; Lane 2, blank control for RT-PCR reaction; Lane 3, human melanoma cell line SK-MEL147 (positive control); Lane 4, DU145 cell line; Lane 5, LNCap cell line; Lane 6, PC-3 cell line. The RT-PCR reaction products were found to be approximately 784 bp. The data are representative of 2 experiments. Several experiments were performed with a variety of primers, confirming expression of p75NTR in DU145 cells.
electrophoresis of HS degradation products. In the HPLC heparanase assay, heparanase activity was defined as decrease in the area of the high-Mr half of the HS peak. Such decrease and consequent shift in the elution profile, indicates the specific degradation of high-molecular weight heparan sulfate to intermediate- and low-molecular weight components. The elution profiles of [35S]HS incubated with PC-3 and DU145 cell extracts showed increased heparanase activity (Figure 5). These results were confirmed by agarose gel electrophoresis analysis with purified cell-surface [35 S]HS [26, 27] incubated with LNCap, PC-3, and DU145 cell lysates. Again, the high-molecular weight [35 S]HS was almost completely degraded when PC-3 and DU145 cell extracts (50– 70 µg protein/assay) were used in the heparanase assay, whereas when LNCap cell extract was present, only minimal
312
E. Timothy Walch and Dario Marchetti Table 2. Heparanase activity in human prostatic cancer cells in the absence and presence of NGF and NT-4/5.a Cell type
−NGF
+NGF
−NT-4/5
+NT-4/5
LNCap PC-3 DU145
1.30 ± 0.4b 1.44 ± 0.3 1.62 ± 0.2
1.67 ± 0.5∗∗ 5.96 ± 0.4∗ 6.82 ± 0.1∗∗
1.46 ± 0.1 1.31 ± 0.1 1.88 ± 0.3
1.88 ± 0.5∗ 9.79 ± 0.6∗ 13.8 ± 0.2∗∗
a The cells were exposed to human NGF or NT-4/5 (100 ng/ml) at 37 ◦ C for 72 h. The data shown represent subtraction of control values (no cellular extract in the heparanase assay) from test sample data. Heparanase activity was defined as decrease in the area of the high-Mr half of the HS peak. Such decrease indicates the specific degradation of high-molecular weight heparan sulfate to intermediate- and low-molecular weight components. b Mean cpm × 103 ± SD from 3 separate experiments. – ∗ P value < 0.01; ∗∗ P value < 0.02.
Figure 4. Expression of human trkB measured by RT-PCR in LNCap, DU145, and PC-3. Lane 1, 100-bp marker ladder; Lane 2, blank control for RT-PCR reaction; Lane 3, DU145; Lane 4, LNCap; Lane 5, PC-3. These results are representative of 2 experiments. Several experiments were performed with a variety of primers confirming expression of trkB but not trkC or trkA in human prostate cancer cell lines.
sis. Treatment with NGF or NT-4/5 resulted in heparanase production, as revealed by a shift in the [35 S]HS peak to a lower molecular weight, suggesting degradation of HS and increased heparanase activity. The increased heparanase activity was consistently most pronounced in the PC-3 and DU145 cells (Table 2). Samples prepared from the control cultures without NGF or NT-4/5 showed a smaller shift in the elution position of HS, suggesting normal heparanase levels. The data displayed were statistically significant and P values were determined for each experimental condition as shown.
Discussion
Figure 5. Heparanase activity in the human prostatic cancer cell lines LNCap, PC-3, and DU145. (Left panel) Elution profiles of cell-surface [35 S]HS ( ) incubated with cell extracts of LNCap ( ), PC-3 ( ), and DU145 ( ). (Right panel) Agarose gel electrophoresis of purified high-molecular weight cell surface [35 S]HS subpopulations after incubation with no extract (Lane 1) and extracts from PC-3 (Lane 2), LNCap (Lane 3), and DU145 (Lane 4) cells.
degradation was observed (Figure 5). That these results reflected heparanase action and not other enzymatic activities was confirmed by performing several control experiments (i.e., no specific intra-chain HS degradation by contaminating sulfatases or heparitinases) as previously reported [26, 27]. Because there was a correlation between the amounts of cell extract and heparanase activity and increased invasive ability for NGF and NT-4/5 but not BDNF or NT-3, we incubated LNCap, PC-3, and DU145 cells with purified and biologically active NGF and NT-4/5 (100 ng/ml), lysed the cells, and performed a heparanase assay and HPLC analy-
Stromal cells have been implicated in the regulation of prostatic growth via the secretion of androgen-dependent paracrine growth factors that are mitogenic for prostatic epithelial cells [39]. These stromal cell-derived mitogenic factors include hepatocyte growth factor, basic fibroblast growth factor, and, importantly, NGF [30–33]. In addition, the low-affinity p75NTR has been found in prostate cancer biopsy samples, and its expression is progressively lost during neoplastic progression of prostatic cells, strongly suggesting that NGF and/or other NTs are involved in specific stages of the metastatic cascade [35, 37]. NT-4/5 was identified in whole-organ homogenates of the human prostate, and the highest levels of NT-4/5 mRNA are found in the prostate [34]. The cellular localization and function of NT-4/5 within the prostate remain to be established. The presence of the low-affinity p75NTR may be therefore useful as a diagnostic tool, as are prostate-specific antigen serum levels, which tend to be relatively low even in well-differentiated cancers. Furthermore, proteins secreted from human prostate stromal cells induce chemotaxis and chemokinesis of prostate tumor cell lines in vitro, which can be prevented with an anti-NGF antibody [30, 38]. These data suggest a paracrine interaction between stromal cells in the prostate and epithelial cells and explain the responsiveness of prostate tumor cells to NGF. However, the exact role of NGF and NTs in prostate biology and the significance of NTR in prostatic carcinoma is far from clear and remains elusive. NGF immunoreactivity has been detected in human prostate adenocarcinoma tissues and cell lines, suggesting that NGF is a mitogen or
Neurotrophins and prostate cancer invasion survival factor [12]. NGF is the prototypical member of a family of mammalian neuronal survival and differentiation factors, known as the NTs, that also includes BDNF, NT3, and NT-4/5 [15–20]. Receptors for and sensitivity to NTs have been found in a number of tumors, including neuroblastoma [53], gliomas [54], medulloblastoma [55], insulinoma [56], prolactinoma [57], and melanoma [21–23]. Of particular interest is the relationship between NTs and invasion of malignant melanoma during metastasis to the brain [21–23]. Melanoma cell invasion in vitro is enhanced by NGF and NT-3 [26], and an inverse expression of NTs and NTR was observed in vivo at the invasion front of human melanoma brain metastases, suggesting that NTs are paracrine growth factors involved in brain metastasis [22]. Paracrine NTs result in the degradation and invasion of brain basement membrane by mechanisms involving a NT-induced expression of important basement membrane-degradative enzymes like heparanase [26–28]. Our studies have linked NGF and other NTs with invasiveness of prostate cancer cell lines. The results reported here confirm and extend earlier results on the presence and role of NGF in prostate cancer cell lines and tissue [47, 58] and demonstrate that NTs have a modulating effect on the in vitro invasive capacity of prostate cancer cells and regulation of heparanase activity, thus extending the biological relevance and importance of this HS-degrading enzyme to tumor metastasis [48, 51]. These results are consistent with the notion of a paracrine role for the significant levels of NGF and NT-4/5 in the prostate [9, 10, 34]. The DU145 cell line was derived from a prostatic carcinoma metastasis to the brain, where selected NTs are present [16, 17, 50]. Our data showed that the increased invasive behavior of the DU145 human prostate cancer cell line in vitro was the only biological effect of NGF and NT-4/5 observable in prostate cancer cells. The findings pertaining to p75NTR and trkB presence, particularly in DU145, further confirm the important role for presence of specific NTRs in invasive processes, as seen originally in melanoma cell systems [21– 24]. However, at the level of the analysis reported here, a strict correlation between trkB and invasive potential was not present. It is conceivable that NT synthesis under in vivo conditions results in a local accumulation of NT activity, leading to relevant paracrine stimulation of prostate cancer cell migration and invasion. However, an autocrine mechanism cannot be ruled out, particularly for DU145 cells, which are known to secrete active NGF [33, 58], and for other androgen-insensitive cell lines known to contain NGF and NT-4/5 transcripts [49]. Overall, our observations may support the concept that autocrine expression of NGF, NT4/5, and/or other NTs is up-regulated in prostate epithelial cells after their transformation to an androgen-refractory phenotype. Therefore, autocrine NGF and NT-4/5 may stimulate proliferation and invasion of tumor cells via the p75NTR and TrkB receptors. One explanation is that during malignant transformation of the prostate, the androgen-insensitive epithelial tumor cells at least partially escape paracrine dependence on stromal cell-derived NTs by the acquisition of autocrine expression of NTs, which may contribute, in
313 part, to the proliferation of the transformed epithelial cells. However, further work is needed to prove this hypothesis. Malignant cell migration within the prostate occurs predominantly by direct extension around prostatic nerves [29]. When the cancer is outside the prostate, autocrine as well as paracrine expression may provide a sufficient source of NTs to maintain the viability of the tumor cells during metastasis to and growth in distant sites. Questions remain to be answered about the mechanisms and biological principles involved in the invasion of perineural space, a relatively common feature of prostate cancer. That the perineural space is not a lymphatic channel makes this route of dissemination even more unusual. Hence, the expression of NTR on prostate cancer cells together with autocrine/paracrine expression of NTs may coincide with transformation to a malignant phenotype capable of invasion along the perineural space and so extracapsular metastasis to distant sites. The role of NTs and their receptors in the biology of prostate cancer metastasis, as suggested in this paper, thus deserves further investigation because it may be important in prostate cancer prognosis. Acknowledgements We acknowledge the technical expertise of Mr Gavin Doughty and the editorial help of Dr Maureen E. Goode, Department of Scientific Publications, M.D. Anderson Cancer Center. Supported by NIH grant R29-CA64178 (to D.M.) and a grant from the Prostate Cancer Research Foundation of M.D. Anderson Cancer Center (to D.M.) References 1.
2. 3.
4. 5.
6.
7.
8.
9. 10. 11.
Gloeckler Reis LA, Hankey BF, Edwards BK (eds). Cancer Statistics View, 1973–1987. Bethesda, MD: The Surveillance Program-DCPC, NIH, US Department of Health and Human Services 1990; 1–357. Carter HB, Coffey DS. The prostate: an increasing medical problem. Prostate 1990; 16: 39–48. Cunha GR. The role of androgens in epithelio-mesenchymal interactionsinvolved in prostatic morphogenesis in embryonic mice. Anat Rec 1980; 175: 87–96. Cunha GR, Chung LWK, Shannon JM, Reese BA. Stromal-epithelial interactions in sex differentiation. Biol Reprod 1980; 22: 19–42. Chung GR, Cunha GR. Stromal-epithelial interactions: II. Regulation of prostatic growth by embryonic urogenital sinus mesenchyme. Prostate 1983; 4: 503–11. Sugimura Y, Cunha GR, Donjour A et al. Whole autoradiography study of DNA synthetic activity during postnatal development and induced regeneration in the mouse prostate. Biol Reprod 1986; 34: 985–95. Sugimura Y, Norman JT, Cunha GR, Shannon JM. Regional differences in the inductive activity of the mesenchyme of the embryonic mouse urogenital sinus. Prostate 1985; 7: 253–60. Djakiew D, Delsite R, Pflug B et al. Regulation of growth by a nerve growth factor-like protein which modulates paracrine interactions between a neoplastic epithelial cell line and stromal cells of the human prostate. Cancer Res 1991; 51: 3304–10. Harper GP, Barde YA, Burnstock YA et al. Guinea pig prostate is a rich source of nerve growth factor. Nature 1979; 279: 160–3. Harper GP, Thoenen H. The purification of nerve growth factor from bovine seminal plasma. J Neurochem 1980; 34: 893–903. Graham C, Lynch JH, Djakiew D. Distribution of nerve growth factor-like protein and nerve growth factor receptor in human benign prostatic hyperplasia and prostatic adenocarcinoma. J Urol 1992; 147: 1444–7.
314 12. MacGrogan D, Saint-Andre J-P, Dicou E. Expression of nerve growth factor receptor genes in human tissues and in prostatic adenocarcinoma cell lines. J Neurochem 1992; 59: 1381–91. 13. Shikata H, Utsumi N, Hiramatsu M et al. Immunohistochemical localization of nerve growth factor and epidermal growth factor in guinea pig prostate gland. Histochemistry 1984; 80: 4111–3. 14. Murphy RA, Watson AY, Rhodes JA. Biological sources of nervegrowth factor. Appl Neurophysiol 1984; 47: 33–42. 15. Levi-Montalcini R. The nerve growth factor 35 years later. Science 1987; 237: 1154–64. 16. Ip NY, Yancopoulos G. Neurotrophic factors and their receptors. Prog Brain Res 1995; 105: 189–95. 17. Snider WD. Functions of the neurotrophins during nervous system development: what knockouts are teaching us. Cell 1994; 77: 627–38. 18. Bothwell M. Functional interactions of neurotrophins and neurotrophin receptors. Annu Rev Neurosci 1995; 18: 223–53. 19. Chao MV. Neurotrophin receptors: a window into neuronal differentiation. Neuron 1992; 9: 583–93. 20. Barbacid M. Structural and functional properties of the Trk family of neurotrophin receptors. Ann NY Acad Sci 1995; 766: 442–58. 21. Marchetti D, Menter DG, Jin L et al. Nerve growth factor effects on human and mouse melanoma cell invasion and heparanase production. Int J Cancer 1993; 55: 692–9. 22. Marchetti D, McCutcheon I, Ross HM, Nicolson GL. Inverse expression of neurotrophin receptor and neurotrophins at the invasion front of brain-metastatic human melanoma tissues. Int J Oncol 1995; 7: 87–94. 23. Nicolson GL, Menter DG, Herrmann JL et al. Tumor metastasis to brain: role of endothelial cells, neurotrophins, and paracrine growth factors. Crit Rev Oncogenesis 1994; 5: 451–71. 24. Herrmann JL, Menter DG, Hamada J et al. Mediation of NGFstimulated extracellular matrix invasion by the human melanoma low-affinity p75 neurotrophin receptor: melanoma p75 functions independently of trkA. Mol Biol Cell 1993; 4: 1205–16. 25. Marchetti D, Parikh N, Sudol M, Gallick GE. Stimulation of the protein tyrosine kinase c-Yes but not c-Src by neurotrophins in human brain-metastatic melanoma cells. Oncogene 1998; 16: 3253–60. 26. Marchetti D, McQuillan D, Spohn WC et al. Neurotrophin stimulation of human melanoma cell invasion: Selected NT enhancement of heparanase activity and heparanase degradation of specific heparan sulfate subpopulations. Cancer Res 1996; 56: 2856–63. 27. Marchetti D. Specific degradation of subendothelial matrix proteoglycans by brain-metastatic melanoma and brain endothelial cell heparanases. J Cell Physiol 1997; 172: 334–42. 28. Nicolson GL, Nakajima M, Herrmann JL et al. Malignant melanoma metastasis to brain: role of degradative enzymes and responses to paracrine growth factors. J Neurooncol 1994; 18: 139–49. 29. DeSchryver-Kecskemeti K, Balogh K, Neet KE. Nerve growth factor and the concept of neural-epithelial interactions. Arch Pathol Lab Med 1987; 111: 833–5. 30. Persson H, Ayer-Lievre C, Soder O et al. Expression of beta-nerve growth factor receptor mRNA in Sertoli cells downregulated by testosterone. Science 1990; 247: 704–8. 31. Djakiew D, Pflug BR, Delsite R et al. Chemotaxis and chemokinesis of human prostate tumor cell lines in response to human prostate stromal cell secretory proteins containing a nerve growth factor-like protein. Cancer Res 1993; 53: 1416–20. 32. Chung LWK, Li W, Gleave ME et al. Human prostate cancer model: roles of growth factors and extracellular matrices. J Cell Biochem 1992; 16H: 99–105. 33. MacGrogan D, Saint-Andre JP, Dicou E. Expression of nerve growth factor receptor genes in human tissues and in prostatic adenocarcinoma cell lines. J Neurochem 1992; 34: 1381–91. 34. Ip NY, Ibanez CF, Nye SH et al. Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution, and receptor specificity. Proc Natl Acad Sci USA 1992; 89: 3060–64. 35. Graham CW, Lynch JH, Djakiew D. Distribution of nerve growth factor-like protein and nerve growth factor receptor in human benign prostatic hyperplasia and prostatic adenocarcinoma. J Urology 1992; 147: 1444–47. 36. Pflug BR, Onoda M, Lynch JH, Djakiew D. Reduced expression of the low-affinity nerve growth factor receptor in benign and malignant
E. Timothy Walch and Dario Marchetti
37.
38. 39.
40.
41.
42.
43.
44. 45.
46.
47.
48. 49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
human prostate tissue and loss of expression in four human metastatic prostate tumor cell lines. Cancer Res 1992; 52: 5403–6. Pflug BR, Dionne C, Kaplan DR et al. Expression of a Trk high affinity nerve growth factor receptor in the human prostate. Endocrinology 1995; 136: 262–8. Angelsen A, Sandvik AK, Syrensen U et al. NGF-β, NE-cells and prostatic cancer cell lines. Scan J Urol Nephrol 1998; 32: 7–13. Cunha GR, Chung LWK, Shannon JM et al. Hormone-induced morphogenesis and growth: role of mesenchymal-epithelial interactions. Recent Progr Horm Res 1983; 51: 3304–10. Kawaja MD, Crutcher KA. Sympathetic axons invade the brains of mice overexpressing nerve growth factor. J Comp Neurol 1997; 383: 60–72. Ware JL, Maygarden SJ. Metastatic diversity in human prostatic carcinoma: implications of growth factor receptors for the metastatic phenotype. Pathol Immunopathol Res 1989; 8: 231–8. Papandreou C, Bogenrieder T, Loganzo F. Mutation and expression of the low-affinity neurotrophin receptor in human malignant melanoma. Melanoma Res 1996; 6: 373–8. Navone NM, Olive M, Ozen M et al. Establishment of two human prostate cancer cell lines derived from a single bone metastasis. Clin Cancer Res 1997; 3: 2493–500. Greco A, Villa R, Pierotti M. Genomic organization of the human NTRK1 gene. Oncogene 1996; 13: 2463–6. Ichaso N, Rodriguez RE, Martin-Zanca D, Gonzalez-Sarmiento R. Genomic characterization of the human trkC gene. Oncogene 1998; 17: 1871–5. Albini A, Iwamoto Y, Kleinmann HK et al. A rapid in vitro assay for quantitating the invasive potential of tumor cells. Cancer Res 1987; 47: 3239–45. Paul AB, Grant ES, Habib FK. The expression and localisation of βnerve growth factor (β-NGF) in benign and malignant human prostate tissue: relationship to neuroendocrine differentiation. Br J Cancer 1996; 74: 1990–6. Nakajima M, Irimura T, Nicolson GL. Heparanase and tumor metastasis. J Cell Biochem 1986; 36: 157–67. Dalal R, Djakiew D. Molecular characterization of neurotrophin expression and the corresponding tropomyosin receptor kinases (trks) in epithelial and stromal cells of the human prostate. Mol Cell Endocrinology 1997; 134: 15–22. Korsching S, Auburger G, Heumann R et al. Levels of nerve growth factor and its mRNA in the central nervous system of the rat correlate with colinergic innervation. Eur Mol Biol Org 1985; 4: 1389–93. Kosir MA, Quinn CCV, Zukowski KL et al. Human prostate carcinoma cells produce extracellular heparanase. Journ Surgical Res 1997; 67: 98–105. Chang S-M, Chung LWK. Interaction between prostatic fibroblast and epithelial cells in culture: role of androgens. Endocrinology 1989; 125: 2719–27. Nakagawara A, Azar CG, Scavarda NJ, Brodeur GM. Expression and function of TRK-B and BDNF in human neuroblastomas. Mol Cell Biol 1994; 14: 759–67. Hamel W, Westphal M, Szonyi E et al. Neurotrophin gene expression by cell lines derived from human gliomas. J Neurosci Res 1993; 34: 147–57. Keles GE, Berger MS, Schofield D, Bothwell M. Nerve growth factor receptor expression in medulloblastoma and the potential role of nerve growth factor as a differentiating agent in medulloblastoma cell lines. Neurosurgery 1993; 32: 274–80. Missale C, Boroni F, Losa M et al. Nerve growth factor suppresses the transforming phenotype of human prolactinomas. Proc Natl Acad Sci USA 1993; 90: 7961–5. Polak M, Scharfmann R, Seilheimer B et al. Nerve growth factor induces neuron-like differentiation of an insulin-secreting pancreatic beta cell line. Proc Natl Acad Sci USA 1993; 90: 5781–5. Geldof AA, De Kleijn MAT, Rao BR, Newling DWW. Nerve growth factor stimulates in vitro invasive capacity of DU145 human prostate cancer cells. J Cancer Res Clin Oncol 1997; 123: 107–12. Ibanez CF, Efbendal T, Barbany G et al. Disruption of the lowaffinity receptor-binding site in NGF allows neuronal survival and differentiation by binding to the trk gene product. Cell 1992; 69: 329–41.
