Metastasis is associated with poor prognosis for melanoma responsible for about 90% of skin cancer-related mortality. To metastasize, melanoma cells must escape keratinocyte control, invade across the basement membrane and survive in the dermis by resisting apoptosis before they can intravasate into the circulation. α-Catulin (CTNNAL1) is a cytoplasmic molecule that integrates the crosstalk between nuclear factor-kappa B and Rho signaling pathways, binds to β-catenin and increases the level of both α-catenin and β-catenin and therefore has potential effects on inflammation, apoptosis and cytoskeletal reorganization. Here, we show that α-catulin is highly expressed in melanoma cells. Expression of α-catulin promoted melanoma progression and occurred concomitantly with the downregulation of E-cadherin and the upregulation of expression of mesenchymal genes such as N-cadherin, Snail/Slug and the matrix metalloproteinases 2 and 9. Knockdown of α-catulin promoted adhesion to and inhibited migration away from keratinocytes in an E-cadherin-dependent manner and decreased the transmigration through a keratinocyte monolayer, as well as in Transwell assays using collagens, laminin and fibronectin coating. Moreover, knockdown promoted homotypic spheroid formation and concomitantly increased E-cadherin expression along with downregulation of transcription factors implicated in its repression (Snail/Slug, Twist and ZEB). Consistent with the molecular changes, α-catulin provoked invasion of melanoma cells in a three-dimensional culture assay by the upregulation of matrix metalloproteinases 2 and 9 and the activation of ROCK/Rho. As such, α-catulin may represent a key driver of the metastatic process, implicating potential for therapeutic interference.
Melanoma is a highly metastatic cancer, with poor prognosis once it has metastasized. Its incidence has increased fivefold during the past three decades, thus making this malignancy a significant clinical problem. Despite their central role for disease progression, the genetic and molecular mechanisms associated with the development and progression of melanoma have not been fully understood so far.
Melanoma progression is a multistep process including the conversion of pigmented lesions via dysplastic nevi to an in situ melanoma, which grows laterally and is mostly confined to the epidermis. This stage is known as radial growth phase melanoma.1–4 If left untreated, the melanoma can progress to the vertical growth phase, which is associated with invasion of the dermis and frequently progression to metastatic melanoma.5, 6
In human epidermis, melanocytes adhere to surrounding basal keratinocytes mainly through expression of E-cadherin.7, 8 The extracellular domain of E-cadherin is responsible for homophilic interaction with adjoining cells, whereas the cytoplasmic domain is linked to the actin cytoskeleton via catenins, important for regulating melanocyte proliferation and differentiation.9 The initial escape of melanoma cells from their original location is mostly associated with transcriptomic and proteomic changes from an epithelial to a mesenchymal profile, leading to expression of mesenchymal markers such as vimentin and N-cadherin, to the loss of epithelial markers such as E-cadherin and to enhanced cell motility.8, 10, 11 This event liberates melanoma cells from keratinocyte-mediated growth and phenotypic control and allows invasion and migration of the transformed melanocytes to distant organs.12, 13 Loss of cell adhesion, repression of E-cadherin expression and increased cell mobility are collectively termed “epithelial to mesenchymal transition” (EMT).14, 15
The transcription factors Snail and Slug, E47, ZEB and Twist are known to be important players in the regulation of EMT as well as in cancer progression.16, 17 Snail, for example, can effect an EMT in melanoma cells by modulating repressing E-cadherin and inducing N-cadherin and proteases MMP2 and t-PA.13, 18
Various signaling pathways have been found to be important in melanoma progression, invasion and metastasis, including the nuclear factor-kappa B (NF-κB).19–22 NF-κB has been implicated in the transcriptional repression of E-cadherin by upregulating specific transcription factors such as Snail/Slug family, Twist or ZEB1/2.23, 24
In an attempt to identify novel proteins that modulate NF-κB signaling or mediate its crosstalk with other pathways, we recently identified α-catulin (CTNNAL1) as an IκB kinase IKK-α (IKK)-interacting protein augmenting NF-κB activation after stimulation with TNF-α or IL-1 in HeLa and HEK 293 cells. In addition, we demonstrate that α-catulin facilitates the crosstalk between the NF-κB and Rho signaling pathways. The biological consequences thereof are the promotion of cell migration and the protection of cells from apoptosis.25, 26 Together with the finding that α-catulin binds to β-catenin and increases the level of both α-catenin and β-catenin suggests that α-catulin has an important role in Wnt signaling.27
Here we show, for the first time, that in contrast to human melanocytes, α-catulin is highly expressed in melanoma cells. α-Catulin knockdown prominently altered the expression of E-cadherin and other genes known to be implicated in melanoma progression. Furthermore, knockdown of α-catulin promoted both binding of melanoma cells to keratinocytes and spheroid formation by enhanced E-cadherin expression. Consistent with the molecular changes, α-catulin provoked invasion of melanoma cells in a three-dimensional (3D) culture assay by the upregulation of matrix metalloproteinases 2 and 9 and the activation of Rho.
Material and Methods
Biochemicals and antibodies
Purified nonlabeled monoclonal mouse and rabbit antibodies were anti-CTNNAL1 (Abnova, Taipei, Taiwan), anti-SLUG (Abcam, UK), anti-CDH1, anti-CDH2 anti β-catenin (BD-Biosciences), anti-MMP2 (Abnova), anti-ZEB1 (Santa Cruz) and Isotype IgG (Sigma-Aldrich). Texas-Red-conjugated secondary anti-mouse and anti-rabbit antibodies were purchased from Jackson Immuno (Newmarket, UK). HRP-conjugated TUBB polyclonal antibody was purchased from Abnova. PDGF and TNFα were purchased from PeproTech (Rocky Hill, NJ). Latrunculin B was obtained from Calbiochem (Merck, Germany), and SRE-Luc-Reporter from Agilent Technologies (Santa Clara).
Cells and cell cultures
Human melanoma cells A375 were purchased from Sigma-Aldrich and cultivated in RPMI (Invitrogen). Metastatic melanoma from spleen (Mel.7), skin (Mel.17), lymph node (Mel.15) and ovary (Mel.14) were isolated and cultivated as described previously.28 IGR37 and IGR39 were kindly provided by Dr. Peter Petzelbauer (Department of Dermatology, Vienna) and cultivated as described previously.29 HeLa cells and melanocytes were obtained from ATCC, and human adult low calcium high temperature keratinocytes (HaCaT) cells were obtained from DKFZ (Germany).
Spheroids were generated by hanging drop method with 100 cells per drop (30 μl) or rotary wall vessel (Synthecon) and embedded into Matrigel-Collagen I matrix (BD Biosciences) after 4 days.
Quantitative real-time PCR
Total RNA was extracted using the RNeasy Mini Kit (Qiagen). RNA was reverse transcribed with the First-Strand cDNA Synthesis Kit (Roche Diagnostics, Germany) according to the manufacturer's instruction. Real-time PCR was performed with TaqMan Gene expression Master Mix and Assays with unlabeled primers and TaqMan probes (FAM dye and quencher labeled). Gene, Probe ID (Applied Biosystem): CTNNAL1, Hs00972094; E-cadherin, Hs01013953; N-cadherin, Hs01032817; Snail1, Hs00195591; Snail2, Hs00161904; ZEB1, Hs00232783; ZEB2, Hs00207691; MCAM, Hs00174838; Plakoglobin, Hs00158408; CTNNB1, Hs00355049; PTEN, Hs02621230; GAPDH, Hs99999905; MMP2, Hs01548727; MMP9, Hs00234579; ZO-1, Hs01551876; Occludin, Hs00170162. Reactions were run on the Light Cycler 480 (Roche). Threshold cycle (Ct) values of the target genes were converted to arbitrary expression values by extrapolation from the standard curve and finally normalized with GAPDH as internal control. Each experiment used at least three independent batches of RNA, and each batch was tested independently at least in triplicate.
For Western blotting, proteins were extracted from 105 cells from each cell line. Total protein extracts were separated by 10% SDS-PAGE and transferred to HybondC nitrocellulose membranes. Membranes were blocked with 5% nonfat milk in Tris-buffered saline (pH 7.4), and immunodetection was carried out using specific antibodies (see “Biochemicals and antibodies” section).
Rock-II Activity Immunoblot Kit was obtained from Cell Biolabs (San Diego, CA). Cell lysis was performed with lysis buffer (150 mM sodium chloride, 50 mM Tris and 1% Triton-X-100, pH 8.0). Kinase reaction was implemented according to the manufacturer's instruction, and phosphorylation of MYPT1 substrate was detected by Western blot. Kinase buffer containing 10 ng of active ROCK-II kinase was used as positive control.
Subcloning of myc-catulin and expression vectors
Myc-catulin DNA from CMV-Myc vector was subcloned into the lentiviral pLEX MCS vector (Thermo Fisher Scientific) by blunt end cloning.25 Lentiviral vector constructs pLEX-MCS, pLEX-JRed and pGIPZ shRNAmir CTNNAL1 and pGIPZ nonsilencing control were purchased from Thermo Scientific. pGL3 reporter vectors were purchased from Invitrogen. E-cadherin promoter (−319/+56) was cloned as described previously.30 Enhanced green fluorescent protein (EGFP) was cloned into pLex MCS vector by GenSkript (Piscataway). Sequences of shRNAmir CTNNAL1 RNAs were as follows: sh-catu1 (V3LHS_356693), sense strand 5′-AGCTCAAAGCAAGAAAACA-3′; antisense 5′-GTTTTCTTGCTTTGAGCT-3′; sh-catu2 (V3LHS_356695), sense strand 5′-AGCTTGTTGAGACCTGTCG-3′; antisense: 5′-CGACAGGTCTCAACAAGCT-3′.
Cell transfection and lentiviral infection
Melanocytes or melanoma cells (4 × 105) per well were transiently transfected by the lipofectamin-2000 method using 2 μg DNA and 7 μl lipofectamin (Invitrogen). After 6 hr, transfected cells were rinsed and incubated for additional 16–48 hr before they were used for the experiments.
For stable transfection, 5 × 106 HEK293T cells were used for the production of lentiviral stocks with a translentiviral packaging system (see “Subcloning of myc-catulin and expression vectors” section; Thermo Scientific). Melanoma cells (2 × 105) were infected with lentiviral stocks and selected with puromycin.
For the adhesion assay, 40,000 HaCaT cells were grown to confluent monolayers in matrigel-precoated (BD Biosciences) 96-well plates for 24 hr. Mel.7 cells (pGIPZ-eGFP-sh-catu2 or -non-silencing (n.s.)) and HaCaT monolayers were preincubated with anti-E-cadherin antibody or istotype control for 1 hr, and Mel.7 cells (pretreated with anti-E-cadherin or isotype control) were used to overlay the nontreated or anti-E-cadherin or isotype control pretreated HaCaT monolayers for 20 minutes. Cells were rinsed twice with PBS, and melanoma cells attached to the HaCaT monolayers were measured by fluorescence using a multiplate reader (EGFP).
Chemotaxis assays were performed as described previously.28 For the transepithelial migration assay, the filters were precoated with 10 μg/ml collagen IV and 10 μg/ml laminin (Sigma) on the upper site. HaCaT monolayers and melanoma cells were cultivated in medium with 1% FCS for 8 hr before melanoma cells were allowed to migrate for 8–24 hr.
Melanoma migration assay was performed by using the Oris Cell Migration Assay from Platypus Technologies (Madison, WI). Melanoma cells (50,000) were dispensed into each well of a 96-well microplate populated with the cell-seeding stoppers before melanoma cells were overlaid with 100,000 HaCaT cells and incubated for 24 hr before the stoppers were removed, and the cells were allowed to migrate for 48 hr. After migration, the detection mask was applied, and the migrated cells were imaged with the multiplate reader and the microscope (Leica CTR6500).
Flow cytometric analysis
For FACScan, cells (2–4 × 105) were fixed and permeabilized (3.5% formaldehyde, 0.2% Triton-X-100) and resuspended in PBS and 0.1% BSA (Biological Industries). Immunofluorescence was performed by exposing cells with antibodies (see “Biochemicals and antibodies” section) for 2 hr on ice. After staining, cells were analyzed using flow cytometry and CellQuest software (BD Biosciences).
Keratinocyte monolayer rearrangement assay was performed with the “electrical cell substrate impedance sensing” (ECIS) model 9600Z (Applied BioPhysics). The measurement system consists of a 96-well cell culture dish (96W10E plate) with ten active electrodes.31, 32 For the keratinocyte monolayer breakdown assays, 1 × 105 HaCaT cells were grown to confluent monolayers and treated with different melanoma (1 × 105) cells seeded on top of the monolayers. Keratinocyte rearrangement was assessed by continuous resistance measurements for 24 hr.
Melanoma cells were seeded on collagen IV-coated (10 μg/ml; Sigma) chamber slides. After 24 hr, cells were fixed with 3.5% formaldehyde in PBS for 30 min and permeabilized with 0.1% Triton X-100 in PBS for 5 min. Proteins were immunodetected with 5 μg/ml monoclonal anti-CDH1 and Texas-Red-conjugated anti-mouse secondary antibody. Chamber slides were mounted (Vector Laboratories) onto cover slips and viewed under a confocal microscope (Leica SP2).
The Student's paired t-test was used for analysis. Reported p values are three-tailed; p < 0.05 was considered statistically significant, and p < 0.01 and p < 0.001 were considered as statistically highly significant.
Human malignant melanoma cells express high levels of α-catulin compared to primary melanocytes
Human melanoma cell lines (A375, IGR37 and IGR39) and freshly isolated human melanoma cells (Mel.7, Mel.14, Mel.15 and Mel.17), obtained from patients with metastatic melanoma,28 were isolated and cultivated for two to five passages and analyzed for α-catulin levels in comparison to normal human primary epidermal melanocytes (normal human melanocytes (NHM-1) and NHM-2) using RT-PCR and Western blotting. Malignant melanoma demonstrated a more than threefold to ninefold increase of α-catulin mRNA expression (Fig. 1a), and a fivefold to tenfold increase of α-catulin protein expression compared to melanocytes in all samples (Fig. 1b). Samples used were isolated from patients who did not receive any treatment for at least 1 month.
α-Catulin knockdown significantly altered genes involved in melanoma progression and metastasis
There is ample evidence that α-catulin plays an important role in tumorigenesis.24, 25, 27, 33, 34 To determine the influence of α-catulin on the expression level of genes known to be involved in EMT, we infected Mel.7 with lentiviral sh-catu1 and sh-catu2 (n.s. as control) particles and A375 melanoma cells either with sh-catu2 or myc-α-catulin (n.s. or mock as control) particles and assessed α-catulin downregulation by real-time PCR analysis (Supporting Information Fig. S6) followed by Western blotting (Fig. 2a). As shown in Figure 2b, α-catulin knockdown resulted in upregulation of E-cadherin and downregulation of Snail1/2, ZEB1/2 and N-cadherin, whereas ectopic expression of α-catulin reversed these effects. To verify E-cadherin expression, A375 cells (stable infected with lentiviral EGFP n.s., sh-catu2, mock and myc-catulin) and Mel.7 cells (n.s. and sh-catu2) were transfected with E-cadherin reporter construct (pGL3-Ecadh) or pGL3 alone (background activity), and the transcriptional activation was measured by assessing luciferase activity. High expression of α-catulin diminished, whereas knockdown increased the expression of the E-cadherin reporter construct (Fig. 2c).
We subsequently determined the corresponding protein levels using flow cytometry, Western blotting and confocal microscopy. For flow cytometry, NHM and Mel.17 cells were infected with sh-catu2 or A375 cells with myc-α-catulin constructs as described above, fixed and stained with anti-E-cadherin or anti-Slug antibodies (Fig. 2d). These data confirmed that α-catulin downregulation promotes E-cadherin expression, whereas α-catulin upregulation diminished this effect. Highest upregulation of E-cadherin was observed when Mel.7 cells were transfected with sh-catu2 and cells selected for 3 days in puromycin-containing medium before analysis (Fig. 2d, right site). These results were confirmed when stable-infected Mel.7 cells were analyzed by Western blotting (Fig. 2e) and immunofluorescence using confocal microscopy (Fig. 2f). E-cadherin was mainly localized in the plasma membrane in knockdown melanoma cells, whereas control cells showed no E-cadherin.
To obtain a status overview and dependency on α-catulin of genes that are associated with tumor progression, we used the melanoma cells Mel.7, Mel.15 and Mel.17 (sh-catu2 and control) and analyzed different genes known to be involved in EMT by real-time PCR and Western blotting (Figs. 2g–2i). Notably, besides E-cadherin, N-cadherin and the associated transcription factors Snail1/2 and ZEB1/2, we found MCAM, Plakoglobin, Occludin and ZO1 and the inhibitors of NF-κB PTEN and RKIP prominently altered in α-catulin knockdown cells. Additionally, the matrix metalloproteinases MMP-2 and MMP-9 known to be involved in melanoma migration are strongly downregulated in α-catulin knockdown cells. Taken together, these data clearly demonstrate that upregulation of α-catulin reduces the expression level of E-cadherin and other genes that are crucial for EMT transition, whereas knockdown of α-catulin increases this effect.
α-Catulin knockdown promotes the adhesion to and inhibits the migration away from keratinocytes via E-cadherin binding
Since our previous results have shown that altered expression of genes and proteins involved in melanoma progression is dependent on α-catulin, we next sought to determine the functional correlation between α-catulin expression and certain steps of adhesion and migratory behavior that deemed to be important for metastasis. The following assays include the escape from keratinocyte–melanoma and melanoma–melanoma adhesion, invasion and migration through extracellular matrices (ECMs).
First, to determine whether melanoma cells could adhere to keratinocytes by means of restoring E-cadherin expression after α-catulin knockdown, HaCaT cells were grown to monolayers. Mel.7 cells (sh-catu2 or n.s.) and HaCaT monolayers were preincubated with anti-E-cadherin antibody or istotype control for 1 hr. Mel.7 cells were allowed to attach to the HaCaT monolayer for 20 min before cells were rinsed twice with PBS, and then the attached melanoma cells were measured (see Material and Methods section). As such, we observed that α-catulin knockdown melanoma cells could adhere much strongly to keratinocytes than control cells. In addition, anti-E-cadherin antibody attenuated this effect compared to isotype control (Fig. 3a).
Second, to study the influence of α-catulin in melanoma to keratinocyte monolayer attachment and reorganization, ECIS technology was applied. HaCaT monolayer reorganization can thus be monitored in real time by means of measuring the decrease (disruption) or increase (formation) in impedance over time. As such, cells that infiltrate and reorganize the monolayer can be detected, as they will lead to a changed resistance of the monolayer. HaCaT cells were cultured onto 96-well ECIS arrays and formed a confluent monolayer within 24 hr. The monolayers were challenged with A375 cells (myc-α-catulin and mock or sh-catu2 and n.s.) or with HaCaT cells as positive control. HaCaT monolayer reorganization was subsequently assessed by continuous resistance measurements. A375 cells with high levels of α-catulin (myc-α-catulin) increased the HaCaT monolayer disruption during melanoma–HaCaT rearrangement, whereas knockdown cells rearranged to form a tight monolayer within 8 hr resulting in increased impedance compared to control cells (Fig. 3b). The same effect was observed when monolayers were challenged with HaCaT cells compared to mock-transfected melanoma cells. These results demonstrate that α-catulin knockdown promotes the adhesion of melanoma cells to keratinocytes.
Third, to better understand the relevance of α-catulin overexpression in melanoma cells in respect to transmigration, HaCaT cells were seeded onto collagen IV/laminin-coated transwell culture chambers until confluent monolayers were formed. Mel.7 cells (sh-catu1, sh-catu2, n.s. or not infected) were subsequently placed onto the HaCaT monolayer in the upper chamber of the 24-well transwells. Transmigration of melanoma cells through HaCaT monolayers was allowed to proceed for 24 hr. As shown in Figure 3c, α-catulin knockdown decreased the number of melanoma cells migrating through the HaCaT monolayer by 80%, whereas control cells showed no effect. To verify these findings and to examine whether the melanoma cells migrate away from the keratinocytes, the Oris Cell Migration Assay was performed (see Material and Methods section). A375 (n.s. and sh-catu2, mock and myc-α-catulin), Mel.7 and Mel.17 cells (n.s. and sh-catu2) together with HaCaT cells were plated around cell-seeding stoppers for 8 hr before HaCaT cells were seeded once again on top of the melanoma cells, and cell–cell binding was allowed for 24 hr before stoppers were removed. The degree of HaCaT (RFP) and melanoma (EGFP) migration into the detection zone was assessed by measuring the fluorescence intensity of both cell types after 48 hr. α-Catulin knockdown decreased the migration of melanoma cells significantly compared to control cells, whereas upregulation of α-catulin promoted the migration only slightly (Fig. 3d). The highest migration was detected when A375 cells were stimulated with 10 ng/ml PDGF and the lowest when treated with 100 nM Latrunculin A, an inhibitor of actin polymerization. Taken together, these results indicate that α-catulin expression promotes the migration away from keratinocytes by downregulation of E-cadherin.
α-Catulin knockdown promotes homotypic spheroid formation via E-cadherin
As homotypic spheroids are thought to be formed by cell–cell adhesions via E-cadherin,35 we hypothesized that high expression of α-catulin might also mediate suppression of E-cadherin-mediated cell–cell adhesion in melanoma spheroid formation. Therefore, we generated spheroids of Mel.7 (sh-catu2 and control) and monitored their formation from Days 1 to 4. α-Catulin knockdown, which promotes E-cadherin expression (Fig. 2), resulted in enlarged but less melanoma spheroids compared to control spheroids by using two methods, the hanging drop (Figs. 4a and 4b) and rotary wall vessel (Supporting Information Fig. S7) methods. No difference in cell metabolism could be detected during spheroid formation between silenced and control melanoma cells (Supporting Information Fig. S8). This suggested a role of calcium-dependent cell–cell adhesion involving integrins or cadherins in the formation/maintenance of melanoma spheroids. Based on real-time PCR data, β1-integrin expression level did not change (data not shown), whereas N-cadherin protein levels were downregulated in α-catulin knockdown melanoma cells (Fig. 2). Therefore, we hypothesized that enhanced E-cadherin levels were mainly responsible for the enlarged spheroid formation. To investigate the correlation between α-catulin and E-cadherin during spheroid formation, we tested the mRNA levels on Days 0–4 using real-time PCR. α-Catulin mRNA level decreased while that of E-cadherin increased during spheroid formation (Fig. 4c). To verify that E-cadherin has a functional role in homotypic spheroid formation, Mel.7 cells (n.s. and sh-catu2) were allowed to generate spheroids for 4 days in the presence of anti-E-cadherin, isotype control or without antibody (hanging drop method). As shown in Figure 4d, spheroids that were treated with E-cadherin blocking antibody had only half the size (∼95 μm n.s. and ∼91 μm sh-catu2) compared to isotype control (∼126 μm n.s. and ∼158 μm sh-catu2) or spheroids that were left untreated. Although the sizes of the individual spheroids showed a large deviation, a clear change between control and knockdown as well as between spheroids treated with anti-E-cadherin and isotype control could be observed. Taken together, these results strongly support an important role of α-catulin for E-cadherin-dependent spheroid formation in anchorage-independent melanoma cells.
α-Catulin is required for melanoma invasion through ECMs
Having demonstrated that α-catulin promotes cell migration by activating Rho GTPases in HeLa cells25 and the expression of the matrix metalloproteinases MMP-2 and MMP-9 in melanoma cells (Figs. 2g and 2i), we next examined whether α-catulin was required for melanoma invasion through ECMs. Therefore, to investigate whether the activity of Rho was influenced by α-catulin in melanoma cells, we transfected Mel.7 knockdown and control cells with a SRE-Luc-Reporter gene either together with IKKβ expression vector or stimulated the transfected cells 24 hr later with TNFα for 8 hr. Downregulation of α-catulin decreased the basal level of the SRE-Luciferase reporter, and also after cotransfection with IKKβ or when cells were stimulated with TNFα (Fig. 5a). This is consistent with our previous finding that α-catulin is a crucial protein for regulating Rho activation. ROCK mediates Rho signaling and reorganizes actin cytoskeleton through phosphorylation of several substrates that contribute to the rearrangement of actin cytoskeleton. We further investigated the influence of α-catulin to ROCK activity by using the ROCK Activity Immunoblot Assay, in which recombinant MYPT1 is used as the ROCK substrate. These results are consistent with our concurrent finding that α-catulin promotes Rho/ROCK activity (Fig. 5b). As Rho family proteins regulate actin cytoskeleton rearrangement required for cell motility, we next tested the effect of α-catulin downregulation on cell migration and invasion. Therefore, spheroids of Mel.7 knockdown or control cells were generated and spheroids of the same size were embedded in matrigel/collagen I mixture for 24 hr. Consistent with the molecular changes (Fig. 2g), invasion and/or cell migration of the α-catulin knockdown Mel.7 cells were completely inhibited in this 3D assay (Fig. 5c). To distinguish whether α-catulin knockdown cells were blocked only by cell–cell binding via E-cadherin or although in cell migration and invasion through the ECMs, we coated the upper and lower sites of the upper 8-μm filter of the 24-well transwell plates with matrigel, col. I, II, IV, laminin and fibronectin and performed the assay using Mel.7 cells with stable-infected sh-catu1, sh-catu2 and n.s constructs. As such, α-catulin knockdown Mel.7 cells were inhibited to invade in all indicated ECMs, but mostly in matrigel followed by laminin and collagen IV (Fig. 5d). These results are consistent with our concurrent finding that α-catulin promotes the expression of the matrix metalloproteinases MMP-2 and MMP-9 that efficiently degrade collagen IV and laminin and thus can assist the melanoma cells to pass through the basement membrane.
These results further confirm that α-catulin is not only responsible for the downregulation of E-cadherin but also for melanoma invasion by the upregulation of MMP 2 and 9 and the activation of ROCK/Rho.
In our study, we showed for the first time that high expression of α-catulin plays a critical role in EMT, melanoma cell migration and invasion. We first demonstrated that α-catulin is highly expressed in melanoma cells, causing reduced E-cadherin and increased N-cadherin expression. Furthermore, upregulation of α-catulin promoted the expression of the transcription factors Snail/Slug and ZEB1/2, whereas downregulation of α-catulin showed the opposite effect (Fig. 2). Besides E-cadherin, other genes coding for crucial proteins of tight junctions or NF-κB inhibitors such as PTEN and RKIP were found to be regulated by α-catulin (Fig. 2g). We further demonstrated that knockdown of α-catulin decreases the expression level of MMP-2 and MMP-9 that have been described to be implicated in malignant melanoma progression by degrading the major components of the basal membrane collagen IV and laminin V.36, 37 As high levels of α-catulin modulate the expression of genes involved in the control of EMT (Fig. 2), it is likely that melanoma cells use the high α-catulin expression for their tumor progression, invasion and metastasis. These findings are in agreement with the notion that α-catulin is crucial for melanoma cells to escape the melanoma–keratinocyte binding by reducing E-cadherin expression.12, 13 Interestingly, knockdown of α-catulin in melanoma cells not only promoted binding to keratinocytes but also formed a dense monolayer with keratinocytes (Figs. 3a and 3b). In addition, knockdown of α-catulin inhibited migration of melanoma cells through keratinocyte monolayers, whereas high level of α-catulin promoted this effect (Figs. 3c and 3d). These results are in accordance with the finding that expression of E-cadherin in melanoma cells resulted in inhibition of cell motility and local invasion.12, 38
Major phenotypic changes were also observed during spheroid formation when melanoma cells were grown under nonadherent conditions. Although downregulation of α-catulin exhibited strong cell–cell adhesion and formed large melanoma spheroids, control cells instead formed more, but smaller and partially loose spheroids (Figs. 4a and 4b). In line with the enlarged homotypic adhesion in melanoma knockdown cells, it was noted that downregulation of α-catulin also markedly elevated the expression of E-cadherin. Furthermore, during spheroid formation, the expression level of α-catulin decreased gradually (Fig. 4c). Additionally, when E-cadherin was blocked during spheroid formation by using an anti-E-cadherin antibody that binds to the extracellular domain of E-cadherin, the size of the melanoma spheroids reduced almost by half (Fig. 4d). Although the function of E-cadherin has been reported during EMT and spheroid formation,8, 10 our study presents the first link between the loss of E-cadherin and α-catulin expression. It is noteworthy to mention that overexpression of E-cadherin in melanoma cells did not regulate α-catulin mRNA expression (data not shown), demonstrating that α-catulin expression is required for E-cadherin repression and not vice versa.
The importance of E-cadherin in tumor suppression is demonstrated by the fact that re-expression of E-cadherin reverses the tumor phenotype, suppressing invasion and metastasis.39, 40 By downregulation of E-cadherin, upregulation of the matrix metalloproteinases and Rho/ROCK activation, α-catulin promotes the invasion of melanoma cells in and through ECM (Fig. 5). In our study, we demonstrate that α-catulin expression modulates markers of the EMT and increased motility and invasion.
Because high level of α-catulin modulates the expression of genes involved in the control of EMT, it is likely that melanoma cells use the high α-catulin expression for their tumor progression, invasion and metastasis. Taken together, we conclude that α-catulin may represent a key driver of the metastatic process in human melanoma.
This work was supported by the Austrian Forschungs Förderungs Gesellschaft (FFG; Project COIN) to Christoph Wiesner.