The murine melanoma cell line B16-F10 was purchased from the American Type Culture Collection (Manassas, VA, USA). C57BL/6 mice (Harlan, Mexico City, Mexico) were used in this study. The experiments with mice were conducted in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the Declaration of Helsinki.

4-Hydroxycoumarin (4-Hydroxy-2H-1-benzopyran-2-one [cat. no. H23805]) and its vehicle, ethanol, were purchased from Sigma (St. Louis, MO, USA). Antibodies against paxillin (sc-5574 rabbit polyclonal), β-tubulin (D-10 mouse monoclonal), FAK (H-1 mouse monoclonal), and phosphotyrosine (PY20 mouse monoclonal) as well as siRNAs were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Control siRNA (sc-37007) is a non-targeting 20–25 nt siRNA that will not lead to the specific degradation of any known cellular mRNA, while paxillin siRNA (sc-36197) is a pool of 3 target-specific 20–25 nt siRNAs.

B16-F10 cells were routinely cultured at 37°C in a humid 5% CO₂ atmosphere, using RPMI-1640 containing 10% fetal bovine serum (FBS). For all experiments cells were seeded at a density of 3 × 10⁴ cells/cm². In experiments with 4-HC, cells were incubated overnight and then exposed for 24 h to 500 μM 4-HC (dissolved in ethanol) or 0.75% ethanol (control) in serum-free medium. The concentration and exposure time used here were previously reported 26, and are those that induce changes in actin cytoskeleton, impairing cell adhesion and motility.

Transfection with siRNA (60 pmols) was carried out with Lipofectamine 2000 (Invitrogen, Rockville, MD, USA) according to the procedure recommended by the manufacturer.

Treated cells were washed twice with ice-cold PBS and lysed in cold lysis buffer [50 mM Tris (pH 8.0), 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% SDS] containing a protease inhibitor cocktail (Roche, Indianapolis, IN, USA). Samples containing 60 μg of total protein were separated by SDS-PAGE and transblotted onto PVDF membranes. The membranes were blocked and then probed with anti-paxillin or with anti-phosphotyr¹¹⁸-paxillin antibodies. On both cases the membranes were stripped and reprobed with anti-β-tubulin antibody. After incubation with the corresponding secondary antibody, the immunoreactive bands were visualized by chemiluminiscence and a densitometric analysis was carried out using ImageJ NIH software 27.

Cellular localization of paxillin was evaluated on cells seeded on Labtek chambers (Nunc, Rochester, NY, USA). After treatment, cells were fixed with 4% formaldehyde in PBS and permeated during 4 min at room temperature with 0.1% Triton X-100 diluted in PBS. Then, paxillin was labeled using anti-paxillin antibody followed by a secondary antibody conjugated with Alexa-546 (Molecular Probes, Eugene, OR, USA). After extensive washing, the slides were mounted and analyzed with a Nikon epifluorescence microscope (Melville, NY, USA).

Total RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer's instructions and quantified spectrophotometrically at 260 nm. Synthesis of cDNA was carried out using Super-Script reverse transcriptase (Invitrogen) and oligo-dT as primer. Semiquantitative PCR was performed in 50 μl of a reaction mixture containing 1× PCR buffer, 2.5 mM MgCl₂, 0.2 mM dNTPs, 1·U AmpliTaq DNA polymerase (Invitrogen), 0.2 μM of each primer, and cDNA obtained from 175 ng of total RNA. Reactions were cycled 30 times through 30 s at 94°C, 60 s at 55°C, and 60 s at 72°C. The paxillin primers used were previously reported by Mazaki et. al. 14 and are as follows: pan-paxillin: sense 5'-aacaagcagaagtcagcagagcc-3', antisense 5'-ctagcttgttcaggtcggac-3' (amplicon length: 582 bp for α isoform and 684 bp for β isoform); β-paxillin: same sense primer than for pan-paxillin, antisense 5'-ctctccatccactctctgtt-3' (503 bp). The GAPDH primers 28 were: sense 5'-accacagtccatgccatcac-3', antisense 5'-tccaccaccctgttgctgta-3' (452 bp). PCR products were resolved by 2% agarose gel electrophoresis and stained with ethidium bromide. Negative controls lacking reverse transcriptase were run in parallel to confirm that samples were not contaminated with genomic DNA.

Real time PCR for ARM-1 was performed in a GeneAmp 5700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) as described previously [29 javascript:popRef('b12')]. Briefly, 0.5× of SYBR Green I (Molecular Probes, Eugene, OR, USA) was added to the reaction mixture described above. The primers for amplification of ARM-1 were: sense 5'-ggacagcttggccctctcat-3'; antisense 5'-gggcaaatcacaatcaccactac-3'. Reactions for GAPDH were performed with the same primers used in semiquantitative PCR. Real time-PCR reactions were cycled 35 times through 30 s at 95°C, 30 s at 60°C, 60 s at 72°C, and 5 s at temperature of fluorescence acquisition (FA). Data were analyzed with the GeneAmp 5700 SDS software version 1.3 (Applied Biosystems).

Lysates from adhered cells were obtained with cold lysis buffer containing phosphatase inhibitors [1 mM sodium pyrophosphate; 1 mM sodium orthovanadate; 50 mM sodium fluoride]. Lysates were immunoprecipitated with anti-FAK antibody bound to protein A- agarose beads. The immunoprecipitated proteins were recovered adding SDS sample buffer, separated by SDS-PAGE, transblotted onto PVDF membranes, and probed with anti-phosphotyrosine and anti-paxillin antibodies. The same membranes were reprobed with anti-FAK antibody.

Pull-down assays for Rac-1 were performed using EZ-Detect activation kit from Pierce (Rockford, IL, USA), according to the manufacturer's instructions. Briefly, 10⁶ cells were lysed in 400 μl of cold lysis buffer as described above. The samples were mixed with GST-PAK1-PBD, loaded on SwellGel immobilized glutathione disks, and incubated with constant agitation at 4°C for 2 h. Then, the disks were washed and the glutathione-bound Rac-1 (active form) was recovered adding SDS sample buffer [60 mM Tris-HCl (pH 6.8); 2% SDS; 0.05% 2-mercaptoethanol; 1% glycerol; and 0.05% bromophenol blue]. The isolated proteins as well as aliquots of cell lysates were separated by electrophoresis and Rac-1 was detected by immunoblotting using anti-Rac-1 monoclonal antibody.

Cell viability was determined using the neutral red accumulation assay 30. The cells were plated on 96-well plates and then treated as described above. After 12, 24, 36 or 48 h of exposure to 4-HC, the medium was changed and the cells were incubated for 90 min under cell culture conditions with 50 μg/ml neutral red. After this incubation the cells were fixed on the plate with an aqueous solution containing 1% formaldehyde and 1% calcium chloride and then lysed with 50% ethanol/49% water/1% acetic acid. The concentration of accumulated neutral red as a marker for cell viability was measured spectrophotometrically at 560 nm.

Cell survival was evaluated accordingly to the methodology reported by Franken et. al. 31. Briefly, cells were harvested and counted after treatment. Cell suspensions from each treatment were diluted in RPMI-1640 with 5% FBS, and immediately re-plated in 6-well plates at a density of 20 cells/cm². The plates were incubated until cells in control wells have formed sufficiently large colonies. After that, the colonies were fixed with 6% glutaraldehyde and stained with 0.5% crystal violet. The plates were photographed and their digital images were analyzed with ImageJ NIH software 27 to known the colony number.

B16-F10 cells were treated in vitro, as mentioned under "cell culture and treatments". After treatment, cells were detached with a non-enzymatic cell dissociation buffer [4 mM EDTA in Ca²⁺ and Mg²⁺-free PBS], resuspended in Hank's balanced salt solution and immediately injected into the tail vein of 8-week old, male C57BL/6 mice. Each mouse received 10⁶ cells. At 2 weeks after i.v. injection mice were euthanized, lungs were excised and the metastatic tumors counted in a blind manner under a dissecting microscopy.

Total RNA, free of DNA contamination was used for the DD-RT-PCR as described by Liang and Pardee 32. Briefly, cDNA was synthesized with 200 ng of RNA and 1 μM of primer of sequence 5'-t₁₂ca-3' to anneal. A control reaction lacking reverse transcriptase was included to confirm absence of non-specific amplification from genomic DNA. The cDNAs corresponding to 20 ng of RNA were PCR amplified in the presence of 0.4 μM of [α-³⁵S] dATP (37 GBq/pmol) using primers of sequence 5'-t₁₂ca-3' (1 μM) and 5'-gttgcgatcc-3' (0.2 μM). PCR reactions were cycled 35 times through 50 s at 95°C, 90 s at 40°C, and 60 s at 72°C. The heat denatured PCR products were electrophoresed on a urea-PAGE gel (48% urea and 6% acrylamide). The gels were dried and exposed to X-ray films at -70°C. After 10–12 h the X-ray film was developed until the DNA bands were clearly seen on the film. The DNA bands which were differentially displayed in the autoradiograph were visually selected, marked on the gel and the band was cut with a sterile razor. The DNA extracted from the gel was PCR amplified using the same set of primers used in the DD-RT-PCR analysis, cloned, and sequenced.

To evaluate the organ-specific adhesion of tumor cells in vitro, the method of Vink et. al. 33 was used with modifications in incubation times. Briefly, fresh lungs obtained from healthy C57BL/6 mice were embedded in TissueTeK O.C.T compound (Poly-Labo, Strasbourg, France) and frozen at -196°C. Fresh cryostat sections (8–10 μm thick) mounted on glass slides were first incubated with PBS/BSA 3% for 60 min in a humid chamber. After treatment, melanoma cells were detached with a non-enzymatic cell dissociation buffer, resuspended in medium, and placed on tissue sections for 30 min at 37°C with gentle agitation. The slides were washed 3 times with medium to remove nonattached cells and then fixed in paraformaldehyde 3% in PBS. Slides were stained with hematoxiline-eosine and the numbers of cells attached to the cryostat sections were counted in 10 microscopic fields.

Analysis of paxillin expression by immunoblotting showed that B16-F10 melanoma cells normally express α and β isoforms of paxillin (Figure 1A, control lane). Treatment of cells with 4-HC decreased the expression of both paxillin isoforms. Densitometric analyses of the immunoblots revealed that in 4-HC-treated cells, the β-paxillin/tubulin and α-paxillin/tubulin ratios decreased to 35 and 50% of the control values, respectively (Figure 1A). Additionally, we performed semiquantitative RT-PCRs in order to evaluate whether the decrements of paxillin elicited by 4-HC were produced through changes in mRNA levels. Using pan-paxillin primers as well as specific β-paxillin primers we found decreased mRNA expression in cells that were treated with 4-HC (Figure 1B). Cellular localization of paxillin can regulate its activation; thus, we studied the paxillin localization through immunofluorescence assays. In control cells paxillin is distributed in a punctate pattern, presumably forming part of focal adhesions. In contrast, in B16-F10 cells treated with 4-HC paxillin localization was predominantly perinuclear (Figure 1C). As previously reported 26, 4-HC impaired the formation of lamellipodia and induced shrinkage of the outer envelope, forming a round cell with radial filopodia attached to adhesion points. Paxillin can be phosphorylated on several tyrosines 34; among them, tyrosines 31 and 118 have shown importance in the generation of signals necessary for adhesion and migration 712. Thus, we studied if the reduction in paxillin expression produced by 4-HC correlated with changes in the level of phosphotyr¹¹⁸-paxillin. Even when both isoforms can be phosphorylated, we were able to detect phosphotyr¹¹⁸ only in the α isofom of paxillin (Figure 1D). For this isoform, the phosphotyr¹¹⁸-paxillin/paxillin ratio is 85% of the control value, indicating that the reduction in phospho-paxillin is mainly caused by paxillin downregulation.

Paxillin participates in the regulation of different signaling pathways. We studied the activation of FAK, a key partner in paxillin-meditated signaling. Basal tyrosine phosphorylation of FAK was found in serum starved B16-F10, as previously reported 35 (Figure 2A, control lane). In 4-HC-treated cells, the phospho-FAK/FAK ratio decreased to 46% of the control value (Figure 2A). However, levels of FAK-bound paxillin did not change significantly (Figure 2A) suggesting that even when paxillin is downregulated, the remaining paxillin supports FAK binding.

The activation of Rac-1, which promotes the formation of lamellipodia, can be triggered by tyrosine-phosphorylated paxillin; therefore, we studied the activation of Rac-1 performing pull-down assays. 4-HC decreased the amount of GTP-bound Rac-1 without changing Rac-1 expression (Figure 2B), reducing the active Rac-1/total Rac-1 ratio to 35% of the control value.

Reduction of paxillin expression as well as decreased FAK activation might affect cell proliferation, survival or apoptosis/anoikis 36. For that reason, we studied the viability of cell cultures exposed to 4-HC during different times. In presence of FBS, 4-HC did not modify cell proliferation in comparison with control cells (Figure 3A, solid lines). When cultures were treated in serum-free medium, we found that 4-HC caused a slight, non significant reduction on the number of viable cells at times longer than 24 h (Figure 3A, dashed lines). Congruously, 4-HC did not affect the ability of single cells to grow into a colonies (Figure 3B), suggesting that cell survival is unaffected by 4-HC.

In order to analyze the relevance of the effects produced by 4-HC in metastasis, we evaluated the effect of 4-HC in an experimental metastasis model. In that model, B16-F10 cells previously treated in vitro were injected into the tail vein of C57BL/6 mice. When control cells were introduced, they metastasize to the lungs and form 55.5 ± 10.1 (mean ± SE; n = 9) black spherical colonies. In contrast, 4-HC-treated cells formed only 8.8 ± 2.4 (mean ± SE; n = 9) of such colonies (Figure 4). Data were analyzed with computer-assisted t test and the difference was highly significant.

In order to study if paxillin inhibition was enough to block the formation of metastasis, we analyzed the metastatic potential of B16-F10 cells transfected with paxillin-siRNA. These cells showed reduced expression of both paxillin isoforms to levels similar of those found in 4-HC-treated cells (Figure 5A). Employing the model described above, we found that melanoma cells transfected with the control-siRNA formed 58.5 ± 7.0 (mean ± SE; n = 7) pulmonary tumors. When cells transfected with paxillin-siRNA were injected, the number of tumors was modest but significantly reduced to 35.7 ± 7.1 (mean ± SE; n = 7) (Figure 5B).

To evaluate other changes in gene expression in melanoma cells treated with 4-HC, we employed the differential display RT-PCR analysis. Several transcripts were differentially displayed between control and 4-HC-treated cells (data not shown). One downregulated band was purified, reamplified, cloned, sequenced, and compared to the nucleotide database in GenBank. The transcript was identified as Adhesion Regulating Molecule-1 (ARM-1; GeneBank accession: NM_019822).

Downregulation of ARM-1 was confirmed by real time RT-PCR. In 4-HC-treated cells, ARM-1/GAPDH ratio was diminished 1800-fold (Figure 6A). ARM-1 expression promote adhesion to endothelial cells 37; therefore, we studied the effect of 4-HC on the adhesion of melanoma cells to lung sections. Our results showed that 4-HC produced a 7-fold reduction in the number of adhered cells (Figures 6B and 6C), suggesting that ARM-1 downregulation may participate in the antimetastatic effect of 4-HC.