Background
Melanoma skin cancer is one of the most devastating types of cancer, extremely aggressive with high metastatic potential. Melanoma metastasis to distant organs is the primary cause of human cancer-related deaths. Worldwide, the incidence of cutaneous malignant melanoma is increasing faster than any other type of cancer. Cutaneous melanoma originates from pigment-producing melanocytes localized at the epidermal-dermal junction in human skin and develops through different steps [
1]. Among various hypotheses, it is proposed that these involve radial (RGP) and vertical (VGP) aberrant growth phases of preexisting nevi or at new site. Then to metastasize at distant sites, melanoma detach from a primary lesion, acquire motility and proteolytic activities to reach lymphatic and blood circulation and undergo growth to distinct organs, all this according to stepwise molecular changes involving defined genetic events [
2,
3]. However, the exact mechanisms underlying this devastating process are complex and somehow still poorly understood. From a molecular point of view, oncogenic activation of the mitogen-activated protein kinase (MAPK) pathway, due to somatic mutations in B-RAF (V600E), is frequently observed in melanoma (70%) [
4].
In mammals, the family of Nck (
non-
catalytic region of tyrosine
kinase) proteins is represented by two highly conserved members, Nck1 and Nck2, composed of three N-terminal SH3 (Src homology 3) domains followed by a unique C-terminal SH2 (Src homology 2) domain and devoid of any catalytic activity [
5,
6]. Like other SH2/SH3 domain-containing proteins, Nck1 and Nck2 behave as adaptor proteins by physically coupling activated membrane receptors to specific downstream effectors [
7]. In mice, individual
Nck knockout resulted in no phenotype, confirming redundancy of Nck proteins, while early embryonic lethality of the double
Nck knockout mice revealed their crucial role in embryonic development [
8]. However, regardless that Nck1 and Nck2 share high amino acid identity, and common cellular functions and binding partners, increasing evidence support specific roles and proteins interactions, as well as tissue expression patterns for these adaptors [
7,
9‐
15]. Previous studies have reported that overexpression of Nck1 in fibroblasts induces cellular transformation and that these cells form tumors in mice [
16,
17]. Furthermore, either Nck has been shown to cooperate with potent oncogenes (v-Abl and Ras) to transform cells, influence cell morphology and anchorage-independent growth [
6]. Although, these studies strongly suggest a role for Nck in cancer development, the mechanism by which Nck oncogenic potential is achieved still remains to be established.
Originally the Nck1 cDNA was isolated from a human melanoma cDNA expression library using a monoclonal antibody produced against the human melanoma-associated antigen [
5], which has no similarity with Nck1. This suggests that the Nck1 mRNA might be abundant in human melanoma. Most recently, the
Nck2 gene was found as being overexpressed in human metastatic melanoma compared with non-metastatic melanoma lesions [
18]. In agreement, the cancer microarray database Oncomine (
https://www.oncomine.org/) reports
Nck2 as a gene upregulated in several human cancer cell lines, including human melanoma. Therefore, the concept that deregulated expression of Nck adaptor proteins could contribute to promote melanoma development and/or progression deserves further investigation. In the present study, using human melanoma cell lines harboring the activating B-RAF (V600E) mutation, that are well defined for stage of cancer progression [
19,
20], we demonstrate that Nck2 protein and mRNA levels are increased in human metastatic melanoma cells compared with human primary melanoma cells that rarely metastasis. We show that Nck2 promotes cell proliferation, migration and invasion in human melanoma cells. In addition, using an
in vivo xenograft model, we provide evidence that increased Nck2 expression in human primary melanoma cells promotes melanoma-derived tumor growth rate. Collectively, our findings indicate that Nck2 plays a role in human melanoma progression.
Methods
Cell lines
The Wistar melanoma cell lines (WM278, WM1232, WM115, 1205Lu, WM164, WM1617 and 451Lu) were obtained from Dr Meenhard Herlyn (PA, USA). Human Epidermal Melanocytes (HEM) cell line was purchased from Cell Applications Inc. Murine colon carcinoma cell (CT26, CT36 and CT51) were obtained from Dr. Nicole Beauchemin (McGill University, Montreal, QC). Breast cancer cell lines (MCF10, MCF7, T47D, MDA-MB-231) were kindly provided by Dr. Morag Park (McGill University, Montreal, Qc).
Cell culture
Unless specified, all chemicals used in this study are from regular commercial sources. Cells were maintained at 37°C in 5% CO2-95%O2 atmosphere. HEK293, colon and breast cancer cell lines were grown in DMEM (Dulbecco's modified Eagle's medium) containing 10% FBS and supplemented with 100 Units/ml of penicillin, 100 μg/ml of streptomycin and 0.25 μg/ml of Amphotericin B. Melanoma cell lines were grown in RPMI 1640 supplemented with 2 mg/ml NaHCO3 and 0.3 mg/ml glutamine. MCF10 cells were grown in DMEM containing 5% Horse Serum (Invitrogen), 20 μg/ml of mouse epidermal growth factor (mEGF, Collaborative Biomedical Products), 10 μg/ml of insulin and 0.1 μg/ml of cholera toxin. To induce MCF10 cell differentiation, cells were grown in media in absence of mEGF but supplemented with 0.5 μg/ml of hydrocortisone for two days. HEM (human epidermal melanocyte) cells were grown in HEM media (Cell Applications Inc.) and cultured according to the manufacturer's instructions.
To analyze phospho-tyrosine proteins, cells were exposed to protein phosphotyrosine phosphatase inhibitor (pervanadate (Na3VO4) or bpVPhen, 100 μM, 15 min at 37°C) prior to be harvested and total cell lysates processed for anti-phosphotyrosine western blot as reported below. Alternatively, total cell lysates (2 mg protein) were incubated with indicated antibodies (4 μg) for 2 hours at 4°C and 40 μl of 50% slurry solution of Protein-A immobilized on Sepharose beads (Santa Cruz Biotech.) were added for an additional 2 hours of incubation at 4°C. Immunoprecipitated samples were washed 3X with lysis buffer before to be finally recovered in Laemmli buffer and processed for anti-phosphotyrosine western blot as reposted below.
Antibodies
Nck polyclonal antibodies were raised by immunizing rabbits with GST-Nck fusion proteins as antigens. Crude serum samples were Protein-A-purified (ProChem, MA) and further tested for Nck specificity as described below. A pan-Nck antibody (1793), which recognizes both Nck isoforms was raised against residues 1-251 containing the three SH3 domains of human Nck1 as previously reported [
21]. Nck1 antibody (2383) and Nck2 antibody (3313) were raised against isoform specific amino acid sequence in between the last SH3 and the SH2 domain of each Nck as reported earlier [
15]. Other antibodies used are: CrkII (C-18), Integrin β3 (H-96), phospho-tyrosine (clone PY99), HA (Y-11) and GFP (B-2) from Santa Cruz Biotech. Antibodies against Integrin β1 (anti-CD29, clone 18), E-Cadherin (clone 34) and N-Cadherin (clone 32) were purchased from BD (ON, Canada). Antibodies to detect vinculin (clone h-VIN-1) and Tubulin (TUB2.1) were from Sigma-Aldrich, USA. Secondary antibodies coupled to HRP were from Bio-Rad Inc. Rhodamine-coupled to mouse anti-IgG was bought at Jackson ImmunoResearch Inc. Phalloidin-coupled to AlexaFluor
®555 and 488 were purchased from Molecular Probes (Invitrogen, CA, USA)
Cell lysis and western blots
Cell lysates were prepared in lysis buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 10% Glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, 10 mM Sodium Pyrophosphate, 100 mM Sodium Fluoride, supplemented with the protease inhibitors Aprotinin and Leupeptin at 1 μg/ml and PMSF at 1 mM. Lysates normalized for protein content (Bradford protein assay, Bio-Rad) were prepared in Laemmli buffer, heated, subjected to SDS-PAGE on 10% acrylamide gels and transferred onto nitrocellulose membranes. For western blot analyses, membranes were blocked in TBS (Tris-buffered saline) containing 10% dry milk and 0.1% Tween-20, and then incubated overnight at 4°C with indicated primary antibodies appropriately diluted in the blocking solution. For pY western blot, blocking and primary antibody solution was TBS containing 5% bovine serum albumin (BSA, Sigma) and 0.1% Tween-20. Next morning, the membranes were washed twice with TBS for 5 minutes followed by two 5 minutes washes using TBS-T (TBS-0.1% Tween-20) and two 5 minutes washes with TBS. The membranes were then incubated with secondary antibody appropriately diluted in milk-blocking solution for 1 hour and washed as above. Finally, signal was detected using ECL Plus Western Blotting Detection System (GE Healthcare, UK) and XR film exposure.
RNA isolation and RT-PCR
Total RNA was isolated from melanoma cells using the TRIZOL (Invitrogen) according to the manufacturer's protocol. Briefly, cells from 100-mm dishes (1 × 106 cells) were suspended in 7.8 ml of TRIZOL. The aqueous and organic phases were separated after addition of chloroform. Precipitated RNA by isopropyl alcohol addition was washed in 70% ethanol and dissolved in RNase-free water. RNA concentration and purity (OD260/280) was measured using an Ultrospec 2100 Pro UV/visible Spectrophotometer (Fisher Scientific, ON). First-strand cDNA synthesis was generated by reverse transcriptase reaction in a final volume of 50 μl. For this, 2.0 μg of total RNA were mixed in a total reaction volume of 20 μl of RNAse free water containing 1 μM Oligo d(T)20 for Nck amplifications or 6 μg of Random Primers for 18S amplification. The reactions were incubated at 65°C for 5 min and quenched on ice. Then, the RT reaction was assembled by adding 10 μl of the 5X 1st strand buffer (Invitrogen), 5 μl of 0.1 M DTT (Invitrogen), 2.5 μl of RNase Inhibitor (40 U/μl) (Invitrogen), 2.5 μl of 10 mM dNTPs, 5 μl of 50 mM MgSO4 and 2.5 μl of Superscript III (200 U/μl) (Invitrogen). Samples were incubated at 37°C for 50 min and deactivated at 70°C for 15 min. PCR amplification was performed using 0.5 μl of cDNA template in a final volume of 50 μl containing 5 μl of 10X PCR Enhancer buffer (Invitrogen), 1.5 mM MgCl2, 0.2 mM dNTPs, 50 pM of specific forward and reverse primers, 10 μl of Amplification buffer (Invitrogen) and 1 U of Taq DNA polymerase (Invitrogen) and DEPC water. Primers used were: Nck1 forward 5'-GCCAGATTCTGCATCTCCTG-3', Nck1 reverse 5'-ACACTTGCCCAGTATTTAGG-3', Nck2 forward 5'-CGAGTACCCCGCCAATGG-3' and Nck2 reverse 5'-CCCGTCACTGAGGACCACC-3'. Reactions were carried out on PTC-100 Programmable Thermal Controller (MJ Research Inc.) according to the following program conditions: initial denaturation at 94°C for 1 min, followed by 1 min at 94°C, 30 seconds of annealing (47°C for Nck 1, 53°C for Nck 2 and 55°C for 18S) and 1 min at 72°C. The final elongation step was 10 minutes at 72°C and the samples were kept at 4°C until analysis. PCR products were separated on a 1% agarose gels and imaged using an NIH Image J 1.30 system. Fifteen, 20, 25 and 30 cycles of PCR amplification products were analyzed to confirm that the amplification was in the linear range for each gene. Ratios of Nck1 and Nck2 over 18S were calculated from similar assays performed in triplicate and repeated three times.
Cell transfection
Human HA-tagged Nck2 cDNA generously provided by Dr. Wei Li (University of Southern California, LA, CA) was subcloned into the retroviral vector pLXSN (Clonetech Laboratories Inc., CA, USA) and the viral particles produced using the GP2-293 cell line according to the manufacturers' instructions. Human Nck2 cDNA was also subcloned into the pEGFP-C1 plasmid (BD, NJ, USA). To establish stable clones of human WM278 primary melanoma overexpressing GFP or GFP-Nck2, cells plated in 100-mm dishes (1 × 106 cells) were transiently transfected with 10 μg of plasmid DNA (pEGFP or pEGFP-Nck2) using calcium phosphate and following selection with neomycin, clones were isolated, amplified and analyzed for GFP or GFP-Nck2 by western blot. 451Lu cells plated at 40-60% confluence were transfected with either 100 nM Nck2 or control siRNA 13379 (Ambion, Austin, TX) using Lipofectamine Plus reagent according to the manufacturer's protocol and analyzed for protein expression after 24 or 48 h.
Proliferation assays
Briefly, cells (4 × 103) were seeded in 96-well plates and 24, 48, 72 or 96 h after, cells were fixed by adding glutaraldehyde (20 min, final concentration 1%). Then, fixed cells were washed twice with deionised water and stained with Crystal Violet (20 min, 0.4% in 10% ethanol, Sigma). The excess of Crystal Violet was removed by washing the cells three times with water and finally, incorporated Crystal Violet was dissolved in 10% acetic acid and read at 570 nm using a spectrophotometer (Beckman Coulter). Wells without cells, but containing medium were used as blank value that was subtracted from all values. Data were expresses a raw OD at 570 nm or as ratio of OD at specific time point over initial OD at day 1.
Wound healing assays
Cell migration was assessed in classical wound healing assays. Confluent monolayer cells in a 6-well plate were wounded using a plastic pipette tip (P200) and rinsed with PBS before to add back culture medium. The bottoms of the wells were marked to indicate where the initial pictures of the wound area were taken. After 8 h incubation at 37°C, pictures (10X) of the same areas were recorded using an Axiovert 200 M microscope (Zeiss) equipped with a CoolSnap™ES camera (Photometric®, Roper Scientific) and closure of the wound evaluated using Metamorph® (V6.3, Molecular Devices Corp.).
Cell invasion assays using Transwells
Melanoma cells (1 × 105) resuspended in 10% serum containing medium were added to the top chamber of a Transwell (8 μm, DD Biosciences, NJ, USA) pre-coated with matrigel™ (BD Biosciences, NJ, USA) diluted in ice-cold PBS (175 μg/ml) at a total of 35 μg per well and allowed to migrate for 24 h. To evaluate the amount of cells that had invaded through each transwell, excess of media was discarded and the transwells were washed once with PBS and then placed in trypsin solution (0.025%) to release the invaded cells underside of the transwells and in the bottom chamber. Total invaded cells were estimated using Calcein AM (BD Biosciences, NJ, USA) as recommended by the manufacturer. Data were normalized according to the respective total amount of cells for each line plated at the same time in adjacent wells devoid of transwells to take into account variations in cell number between cell lines.
Spheroid formation and culture in 3D were performed according to the hanging drop method [
22]. Briefly, 2 × 10
4 cells in 20 μl of culture medium were suspended on the lid of tissue culture dishes containing 10 ml of culture medium for 48 h to form spheroids. Then spheroids were transferred in culture dishes containing culture medium and on 2% agar (Agar Select, Invitrogen, CA, USA) at the bottom. After 72 h of growth in suspension, individual spheroid has been transferred in 4-well plate containing 80% collagen type IV (PureCol
®, Advanced BioMatrix) in RPMI without FBS. Following 30 min at 37°C to allow collagen polymerization, 500 μl of RPMI containing 10% FBS was added to each well. Images were recorded initially and at 12-24 hr intervals as reported above for wound healing assays. Spheroid invasion was determined qualitatively as either positive or negative comparing sequential images.
Actin and focal adhesions
Cells were plated at 3 × 104 cells/well on glass coverslips pre-coated or not with various extracellular matrices and incubated in culture medium for 24 h. All steps were carried at room temperature and coverslips were rinsed with PBS between each step. Cells were fixed in freshly prepared 3.7% formaldehyde for 10 min, permeabilized in 0.2% Triton-X-100 for 5 min and blocked in 0.1% BSA for 30 min. For vinculin staining, cells were incubated with primary monoclonal anti-vinculin antibody (1:400) for 1 h and with a mixture of secondary tetramethylrhodamine isothiocyanate-conjugated phalloidin-conjugated goat mouse antibody (TRITC-GAM, Sigma) for 30 min. Actin staining was performed by incubating the coverslips for 30 min with Phalloidin-AlexaFluor®555. Coverslips were mounted by inverting them on glass slides using Prolong anti-fade mounting media (Molecular Probes). Coverslips were examined on a Zeiss Axiovert 200 M microscope (Zeiss) using 40X or oil immersion 63× objective lens. Fluorescent images were captured using a CoolSnap™ES camera (Photometric®, Roper Scientific) and analyzed using Metamorph® (V6.3, Molecular Devices Corp.).
Tumor growth in vivo
WM278 primary melanoma cells either parental, overexpressing GFP (C2) or GFP-Nck2 at low (N7) or high (N14) levels were grown in RPMI medium supplemented with 10% FBS to 80% confluency. Cells (5 × 106) resuspended in 500 μl at 50% Matrigel™ were injected subcutaneously in the right flank of 6-week-old CD-1 Nude mice (Charles River, Qc, Canada) (n = 5 for each group). Tumors development was followed for 20 weeks. Tumor size was measured every week with calipers to assess tumor volume ([πlength × width2]/6). Mice were housed in McGill University Animal facilities at the Genome building. Mice experiments were conducted under a McGill University-approved animal use protocol (Dr. P.M. Siegel) in accordance with guidelines established by the Canadian Council on Animal Care.
Data analysis and statistics
Densitometry analysis results are expressed as means ± S.E.M. Student's t test was used to evaluate the statistical significance of the results. A p ≤value 0.05 is assumed to be significant.
Discussion
Nck1 and Nck2 SH2-SH3 domain-containing proteins have been reported to be differently expressed in numerous mouse tissues [
8,
15]. In agreement with the ability of both Nck to collaborate with strong oncogenes to transform cells [
6],
Nck1 and
Nck2 genes were found upregulated in several human cancer cell lines, including melanoma (
https://www.oncomine.org/). However, Nck proteins expression levels in cancer tissues and possible mechanism(s) by which these adaptors contribute to cancer development have been poorly investigated to date. In this study, we provide evidence that Nck2 plays a role in promoting proliferation, migration and invasion of human melanoma cells
in vitro and growth of melanoma-derived tumors
in vivo, while its expression is upregulated in metastatic cancer cells, including colon, breast and melanoma.
Our investigation revealing that Nck2 overexpression in human primary melanoma cells induces metastatic characteristics point towards Nck2 sufficiency to promote metastasis phenotype. In this study, we did not address whether Nck2 is necessary for melanoma metastasis. However, we provided some insights suggesting that Nck2 could play such function. In fact, we found higher levels of Nck2 expression in metastatic compared to non metastatic cell lines in three different types of cancer. In addition, we demonstrate that depletion of Nck2 in metastatic melanoma reduces cell proliferation. This does not exclude that other yet identified players could be required to fully promote metastasis in melanoma overexpressing Nck2. None the less, our findings clearly demonstrate that overexpression of Nck2 in human primary melanoma correlates with upregulation of the total phospho-tyrosine proteins content, assembly of novel Nck2-dependent pY-protein complexes and downregulation of E- and N-cadherins, and β-1 and -3 integrins. E-cadherin, found at adherens junctions, is the principal effector of cell-cell adhesion [
40]. Loss of E-cadherin expression in cancer cells weakens cell-cell adhesion and is associated with cancer progression, invasion and metastasis [
41‐
43]. At the present time, there is no evidence for a direct link between E-cadherin and Nck2. Further investigation is required to elucidate the molecular events responsible for E-cadherin downregulation associated with overexpression of Nck2 in human primary melanoma cells and whether downregulation of Nck2 in metastatic human melanoma cells would restore E-cadherin expression remains to be determined. On the other hand, the degree of cancer cells cohesion in primary tumor also depends on the strength of cell-ECM contacts mediated by integrins [
44]. Alteration in integrins expression has been also implicated in cancer progression, invasion and metastasis [
45,
46]. Integrins signaling associated with regulation of the actin cytoskeleton leading to adhesive attachment involves the activation of the focal adhesion kinase (FAK) and the integrin-like kinase (ILK) (reviewed in [
47]). Interestingly, Nck2 has been shown to affect cell motility through its direct interaction with FAK [
48]. Moreover, increasing evidence support a close relationship between integrins and growth factor receptor tyrosine kinases to activate signaling pathways that promote proliferation and metastatic activity (reviewed in [
49]). Nck2 has been reported to function as a molecular link connecting integrins and growth factor receptor tyrosine kinases signaling pathways. In fact, Nck2 associates with numerous receptor tyrosine kinases [
17,
50‐
54] through its SH2 domain and using its third SH3 domain, it binds to a LIM domain in PINCH (Particularly Interesting Cys-His-rich Protein) [
55]. PINCH, a binding protein for ILK, plays an important role in mediating integrins-induced cell-ECM interaction by directing ILK to focal adhesions [
56]. It is recognized that the ILK-PINCH complex participates to signaling pathways regulating fundamental cellular processes (reviewed in [
57], including cell shape and migration [
58]. A crucial role for Nck in regulating these processes was particularly illustrated by the findings showing that fibroblasts derived from
Nck double knockout mice embryos display major defects in cell attachment, cell motility and actin remodeling [
8]. Therefore, increased expression of Nck2 in human primary melanoma cells may elicit protein interactions that re-wire signaling pathways in a fashion that alters focal adhesions and promotes cell motility by interacting with FAK and PINCH. Alternatively, increased expression of Nck2 could passively destroy proper stoichiometry of molecular complexes and in this manner, indirectly contributes to cancer progression by altering signaling pathways regulating the actin cytoskeleton network supporting cell migration.
Nck proteins are known to couple activated receptor tyrosine kinases, as well as non receptor tyrosine kinases, to effectors involved in signaling pathways regulating proliferation and actin cytoskeleton dynamics [
7,
14,
59‐
61]. Non-receptor protein kinases of the Src and Abl families are often overexpressed or aberrantly activated in a wide variety of human cancers and their roles in cancer progression, including proliferation, survival, motility, invasiveness, metastasis and angiogenesis, is significant. Of note, Nck directly binds to and promotes Abl activation and signaling [
62,
63], and associates with p60
v-src in vitro [
16]. c-Src has been recently reported to be overexpressed in human metastatic melanoma tumors [
64]. Interestingly, Src-dependent phosphorylation of Tks5 and cortactin recruits Nck to invadopodia, where it regulates actin assembly and ECM degradation [
65‐
67]. Invadopodia, exclusive invasive cancer cell membrane actin-based protrusions enriched in signaling and proteolytic activities, are used by invasive cancer cells to degrade the ECM and invade surrounding tissues [
68‐
71]. It is then possible that upregulation of tyrosine phosphorylated proteins and downregulation of cadherins and integrins in human primary melanoma cells that overexpress Nck2 may endow melanoma cells with altered adhesive properties and spatial relationships that favor uncontrolled proliferation, migration and invasion.
Nck1 and Nck2 proteins are highly identical, but despite high homology, redundant functions and common binding partners, increasing evidence suggest specific roles and protein interactions [
7,
9‐
12,
14], as well as specific tissue expression patterns for Nck proteins [
8,
15]. In this study, the effect of Nck1 overexpression on melanoma phenotype was not addressed. However, our results demonstrate that increased endogenous expression of Nck2 in human metastatic melanoma cells relative to primary melanoma cells and melanocytes results from increased
Nck2 transcription, suggesting that Nck1 and Nck2 promoters are under different regulatory controls.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MLC participated to most experiments and contributed to the analysis of the data. JD performed experiments to generate Figure
8 and has been essential to the study design and the draft the manuscript. SI participated in the cellular and biochemical assays. APC carried out wound healing assays experiments reported in Additional file
3. PMS contributed to the conception and achievement of
in vivo experiments. LL conceived the study, participated in its design and coordination, and to the preparation of the final manuscript. All authors read and approved the final manuscript.