Mammary carcinoma F-7 (MCF-7), human breast T-47D cancer cells, and human prostate DU-145 cancer cells were purchased from American Type Culture Collection (Manassas, VA) and cultured as advised by the manufacturer. Anandamide and palmitoylethanolamide were synthesized as described previously (2). Arvanil[ N-(3-methoxy-4-hydroxy-benzyl)-arachidonoylamide], a hybrid agonist at CB1 and vanilloid receptors, was synthesized as described previously (17A ). 2-AG and the synthetic cannabinoid HU-210 were donated by Prof. R. Mechoulam, Hebrew University of Jerusalem (Jerusalem, Israel). SR141716A and SR144528 were gifts from Sanofi Pharmaceuticals, Inc. (Montpellier, France). (R)-Methanandamide and BML-190 were purchased from Biomol (Plymouth Meeting, PA), capsazepine was obtained from Alexis Corp. (San Diego, CA), and human PRL and recombinant β-NGF were purchased from Sigma. [³H]SR141716A (55 Ci/mmol) was purchased from Amersham Pharmacia Biotech (Aylesbury, UK).

Total RNA was prepared from MCF-7, T-47D, and DU-145 cells by the acid guanidium thiocyanate-phenol-chloroform method. Complementary DNA (cDNA) synthesis was performed in a 20-μl reaction mixture containing 75 mm KCl, 3 mm MgCl₂, 10 mm dithiothreitol, 1 mm deoxy-NTPs, 50 mm Tris-HCl (pH 8.3), 5 μg total RNA, 20 U ribonuclease inhibitor (Roche Molecular Biochemicals, Indianapolis, IN), 0.125 A₂₆₀ units hexanucleotide mixture (Roche Molecular Biochemicals) for random priming, and 200 U Moloney murine leukemia virus reverse transcriptase (Superscript, Life Technologies, Inc., Gaithersburg, MD). The cDNA reaction mixture was incubated at room temperature for 10 min and at 37 C for 90 min, and the reaction stopped by heating at 98 C for 5 min followed by lowering the temperature to 4 C. RT-PCR amplification was performed using 0.5–3 μl cDNA reaction mixture and 2 U Thermus icelandicus DNA polymerase (Red-Hot, Advanced Biotechnologies, Columbia, MD) in a 25-μl PCR reaction mixture containing 25 mm Tris-[hydroxymethyl]-methyl-propane-sulfonic acid, sodium salt buffer (pH 9.3), 50 mm KCl, 1.5 mm MgCl₂, 1 mm β-mercaptoethanol, 12% dimethylsulfoxide, 250 μm of each deoxy-NTP, and 0.5μ m each of 5′- and 3′-primers. Reactions were performed in a PE Gene Amp PCR System 9600 thermocycler (Perkin-Elmer Corp., Norwalk, CT). The amplification protocol consisted of an initial denaturation of 5 min at 95 C, followed by 25–40 cycles of 1 min at 95 C, 1 min at 55 C, and 2 min at 75 C. A final cycle of 10 min was carried out at 75 C. The primers used were: CB1 sense primer, 5′-CGC AAA GAT AGC CGC AAC GTG T-3′; CB1 antisense primer, 5′-CAG ATT GCA GTT TCT CGC AGT T-3′; CB2 sense primer, 5′-TTT CCC ACT GAT CCC CAA TG-3′; CB2 antisense primer, 5′-AGT TGA TGA GGC ACA GCA TG-3′; β₂-microglobulin sense primer, 5′-CCA GCA GAG AAT GGA AAG TC-3′; andβ ₂-microglobulin antisense primer, 5′-GAT GCT GCT TAC ATG TCT CG-3′. The expected sizes of the amplicons were 244 bp for CB1, 337 bp for CB2, and 268 bp forβ ₂-microglobulin. Theβ ₂-microglobulin housekeeping gene expression was used to evaluate any variation in the RNA content and cDNA synthesis in the different preparations. Furthermore, the PCR primers for β₂-microglobulin were selected on the basis of the sequence of the β₂-microglobulin gene (NCBI accession no. M17987) by including the intron 402-1017. These primers, in the presence of contaminant DNA, would have generated an amplicon of the expected size of 886 bp. Ten microliters of the PCR products were analyzed by electrophoresis on 2% agarose gels (MS agarose, Roche Molecular Biochemicals) in 1 × Tris-acetate-EDTA buffer at 4 V/cm for 4 h. Ethidium bromide (0.1 μg/ml) was included in both the gel and the electrophoresis buffer, and the PCR products were detected by exposure under UV light. No PCR product was detected in the absence of cDNA, primers, or Red-Hot DNA polymerase.

Binding assays in MCF-7, T-47D, and DU-145 cells were carried out using the filtration procedure reported previously (13), and membranes were prepared as therein described, except for the absence of phenylmethylsulfonylfluoride from the binding buffer. The binding of increasing concentrations (100–10000 pm) of[ ³H]SR141716A to aliquots (0.4 mg total proteins) of these membranes and the displacement of a fixed concentration (300 pm) of[ ³H]SR141716A by increasing concentrations (0.025, 0.1, 0.5, 1.0, and 5.0 μm) of anandamide or arvanil were measured in equilibrium assays. SR141716A (10μ m) was used to determine nonspecific binding. Receptor binding results were analyzed with GraphPad software (GraphPad Software, Inc., San Diego, CA). Scatchard curves for the binding of [³H]SR141716A were used to calculate the binding capacity (B_max) and K_d for this ligand using nonlinear regression, and one- and two-site analyses were compared to determine better fit values (r² = 0.88 for one-site binding). Displacement curves (calculated by means of Pharm/PCS software[ Microcomputer Specialists (MCS), Philadelphia, PA)] were used to calculate the K_i values for anandamide by inserting the corresponding IC₅₀ values from the best-fitting curves into the Cheng-Prusoff equation.

Cell proliferation assays were carried out according to the method previously described (13) in 6-well dishes containing subconfluent cells (at a density of about 50,000 cells/well). With MCF-7 and T-47D cells test substances were introduced 3 h after cell seeding and then daily at each change of medium. Cells were treated with trypsin and counted by a hemocytometer 4 days after the addition of test substances. No significant decrease in cell viability (as assessed by trypan blue) was observed with up to 100μ m anandamide. With DU-145 cells, we used a previously described procedure (18). Substances were added 3 h after cell seeding (50,000 cells/well). PRL (1 mIU/ml) or vehicle was then added after 24 h with the change of medium, in the presence of the test substances or vehicle. After 72 h, cells were treated with trypsin and counted by a hemocytometer. To study the effect of NGF on MCF-7 cell proliferation, we used the previously described procedure (19). Twenty-four hours after cell seeding (50,000 cells/well), the medium was changed to serum-free medium, and cells were starved for 24 h. Cells were then treated with serum-free medium containing β-NGF (100 ng/ml) plus test substances or vehicle and treated with trypsin, then counted after 48 h. Means were compared by means of unpaired Student’s t test, using P < 0.05 as the threshold for statistical significance.

Cells in 100-mm petri dishes were treated with test substances under the same conditions those described above for cell proliferation assays. Cells were then washed twice with 137 mmNaCl, 3 mm KCl, 12 mm Na₂HPO₄, and 2 mm KH₂PO₄ (pH 7.4) and lysed with a lysis buffer consisting of 50 mm Tris-HCl (pH 7.4), 1 mm EDTA, 150 mm NaCl, 1 mm Na₃VO₄, 1 mm NaF, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mm phenylmethylsulfonylfluoride, 1% Triton X-100 and 1μ g/ml each of aprotinin, leupeptin, and pepstatin A. Lysates were loaded onto gels containing, respectively, 10% and 7.5% polyacrylamide for blotting of PRLr and CB1 or Trk, respectively. Proteins were transferred to nitrocellulose membranes, which were then incubated first for 1 h at room temperature with the first antibody, i.e. either antihuman PRL receptor monoclonal antibody (U5, purchased from Affinity BioReagents, Inc., Golden, CO; 1:1000), antimouse Trk monoclonal antibody (B-3, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; 1:500), or CB1 polyclonal antibody (Cayman, Ann Arbor, MI; 1:800) and then with the appropriate horseradish peroxidase-labeled second antibody conjugates (1:5000; Bio-Rad Laboratories, Inc., Hercules, CA). Bands were visualized by the enhanced chemiluminescence technique (Bio-Rad Laboratories, Inc.). The anti-Trk and CB1 antibodies cross-react with human Trk and CB1.

The molecular targets for the action of each of the substances used in this study are listed in Table 1. The effects of anandamide, the synthetic cannabinoid HU-210, the metabolically stable anandamide analog (R)-methanandamide, 2-AG, and the anandamide congener, palmitoylethanolamide, on human EFM-19 breast cancer cells and of anandamide on MCF-7 and T-47D cells have been described previously (13). As shown in Table 2, the IC₅₀ values for the antiproliferative effect of anandamide on MCF-7 and T-47D cells, after only a few subculturing passages, were 1.4 and 1.9μ m, respectively. IC₅₀ values as high as 7 μm were instead found for anandamide when using MCF-7 cells that had undergone between 10–15 subculturing passages. 2-AG was less active than anandamide in T-47D, but not in MCF-7 (Table 2) or EFM-19 (13), cells. As shown in Fig. 1A, the selective CB2 agonist BML-190 (20) did not exert any antiproliferative action on HBCCs up to a 10-μm concentration, whereas arvanil, a CB1 agonist (see below), was a potent inhibitor of HBCC proliferation. The selective CB2 antagonist SR144528 (21) did not reverse the antiproliferative action of anandamide and arvanil, whereas the CB1 antagonist SR141716A did, in agreement with our previous report (13). The vanilloid receptor antagonist, capsazepine (not shown), also counteracted the effect of arvanil, in agreement with the capability of this compound to activate both CB1 and vanilloid receptors, whose presence in HBCCs was recently suggested (17A ).

In agreement with a previous study (19), starved MCF-7 cells start proliferating after treatment with 100 ng/ml human β-NGF. A 91± 13% (n = 16) increased proliferation was observed vs.that of vehicle-treated cells. This process was blocked by submicromolar doses of anandamide, 2-AG, (R)-methanandamide, the synthetic cannabinoid HU-210, and arvanil, but not by BML-190 or palmitoylethanolamide (Fig. 1B). The antiproliferative effects of anandamide and arvanil were attenuated by SR141716A, but not by SR144528 (Fig. 1C). Capsazepine also counteracted the effect of arvanil (data not shown).

MCF-7 cells express both low (p75) and high (Trk) affinity NGF receptors (19). We detected the latter receptor as a single band of approximately 140 kDa using the Western immunoblotting technique carried out by means of a monoclonal anti-Trk antibody. As shown in Fig. 2, anandamide, (R)-methanandamide, HU-210, and arvanil, but not palmitoylethanolamide and BML-190, suppress the expression of Trk in starved, β-NGF-treated MCF-7 cells. The effect of anandamide was dose related (IC₅₀ = 0.6μ m) and, like that of HU-210, was reversed by SR141716A. The effect of arvanil was partly counteracted by either SR141716A or capsazepine.

As shown in Fig. 3A, the two HBCC lines studied express the CB1 mRNA as detected from the finding of a RT-PCR-amplified cDNA fragment of the expected size obtained by using oligonucleotide primers with the same sequence as human CB1 cDNA (see Ref. 22 for a review). HBCCs also express CB2 mRNA as detected by using the same technique (23). When using the same amount of total RNA (from the same preparation), and the same nonsaturating number of amplification cycles, the CB2 signal was reproducibly much less intense than the CB1 signal, and, in T-47D cells, hardly visible. Using the same procedure, probes and amplification conditions, CB2 and CB1 transcripts of the same size were found in human lymphocytes, the former always in higher amounts than the latter (data not shown). The presence of CB1-like receptors in MCF-7 cells was confirmed by Western immunoblotting, which showed the presence of a single band with an apparent molecular mass of approximately 57 kDa (Fig. 3B). This value is compatible with the human CB1 receptor molecular mass if glycosylation of this protein (as described in Ref. 22) is taken into account. This value is very similar to the molecular mass (∼58 kDa) of CB1-immunoreactive proteins detected in various rat and human cells that express CB1 receptors (24). We found that the abundance of the CB1 transcript, compared with that of theβ ₂-microglobulin transcript, as well as the intensity of the CB1-immunoreactive band decreased when MCF-7 cells underwent 10–15 subculturing passages (Fig. 3b and data not shown).

Membrane preparations from MCF-7 and T-47D cells contain specific binding sites for [³H]SR141716A as detected by the filtration binding assays used in this study. Table 3 shows the B_max and K_d values for[ ³H]SR141716A binding sites in each HBCC line, calculated from the corresponding Scatchard plots.[ ³H]SR141716A specific binding in these two cell lines, determined in the presence of 10 μm SR141716A, depended on the number of subculturing passages and varied between 10–40% for MCF-7 cells and 15–30% for T-47D cells. Anandamide K_i values for the displacement of[ ³H]SR141716A specific binding were 1.50± 0.31 and 0.85 ± 0.25 μm (n = 3) for T-47D and MCF-7 cells, respectively. The K_i value for arvanil in MCF-7 cells was 0.25 ± 0.10 μm (n = 3).

In agreement with a previous study (18), DU-145 cells proliferated more rapidly in response to human PRL; a 25 ± 1% increase in cell proliferation (n = 19) was observed with 1 mIU/ml of the hormone after a 3-day treatment. Longer treatments of DU-145 cells with PRL or higher doses of the hormone did not lead to a further increase in cell proliferation (18) (data not shown). Anandamide, 2-AG, and HU-210 were potent inhibitors of PRL-induced, but not basal, DU-145 cell proliferation (Fig. 4A). (R)-Methanandamide and arvanil were not less potent than anandamide. The effect of anandamide was reversed by SR141716A, but not by SR144528, and was not mimicked by BML-190 and palmitoylethanolamide (Fig. 4, A and B). Both SR141716A and capsazepine, but not SR144528, counteracted the effect of arvanil (Fig. 4B and data not shown).

As shown in Fig. 4C, DU-145 cells express CB1 and, to a lesser extent, CB2 mRNA, as determined from the finding of RT-PCR-amplified cDNA fragments. DU-145 cells also contain specific binding sites for[ ³H]SR141716A, whose B_max and K_d values are shown in Table 3, and an approximately 57-kDa protein detected by Western immunoblotting (Fig. 3B). Anandamide displaced[ ³H]SR141716A from DU-145 cell membranes with a K_i of 160 ± 19 nm(n = 3).

Using the Western immunoblotting technique carried out with a specific monoclonal antibody, DU-145 cells were shown to contain measurable levels of the high molecular mass (∼100 kDa) form of the PRLr. Anandamide inhibited the expression of this receptor in a fashion sensitive to the presence of SR141716A (Fig. 4D).