1 Introduction
Globally, colorectal cancer (CRC) is the third most common cancer (10.2 % of all cancer cases) and the second most common cause of cancer-related mortality (9.2 %) [
1]. The high CRC-related mortality rate is closely related to the occurrence of metastasis, which is characterized by the dissemination of aggressive cancer cells from the primary tumor to distant organs. Cell migration and invasion are aberrantly activated in cancer cells during metastasis. To invade surrounding tissues cancer cells, to a considerable extent, undergo cytoskeleton rearrangements with the dynamic assembly and disassembly of F-actin-based structures, such as membrane ruffles, blebs, filopodia, lamellipodia and invadopodia [
2,
3].
Membrane protrusions are spatiotemporally regulated by classical Rho family small guanosine triphosphatases (GTPases), including RhoA, Rac1 and Cdc42, along with their downstream effectors [
4‐
6]. The activity of classical Rho GTPases is controlled via reversible actions between Rho-specific guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins by exchanging bound GDP nucleotides for GTP nucleotides and
vice versa. Specifically, several GEFs are vital for promoting Rho GTPase-mediated cell migration. The Dbl and DOCK families of proteins have been identified as specific types of GEFs. Among them, beta-Pix (βPix), which activates Rac1 and Cdc42, is one of the most frequently investigated GEFs, particularly in terms of cell spreading and migration [
7‐
9]. Previous studies have revealed that βPix promotes collective migration of endoderm cells and neural tube formation by activating downstream signaling including Rac1 and p21-activated kinase 2a (Pak2a) during embryonic development [
9‐
11]. Mechanistically, βPix plays an active role in cell migration by forming several complexes. To date, the most notable βPix-mediated multi-protein complex is the βPix-GIT-PAK complex, known to induce Rac1-mediated actin reorganization at the membrane edge and to recycle focal adhesion (FA) components in an integrin-dependent manner [
12,
13]. In addition, it has been found that the P-cadherin-βPix complex is required for Cdc42-induced cell polarity and mechanical forces during collective migration [
14]. Therefore, βPix may exert certain intracellular functions related to its binding partner.
Overexpression of βPix has been observed in patients with breast cancer and CRC, indicating its potential as a cancer biomarker [
15,
16]. Specifically, in CRC cells βPix has been found to enhance the transcriptional activity of β-catenin via direct binding, leading to cell proliferation regardless of GEF activity [
17]. Considering the intracellular functions of βPix, it is reasonable to speculate that βPix may play a role in CRC cell motility or invasion. As yet, however, the specific binding protein that controls the function of βPix in CRC remains unknown.
Dynamin 2 (Dyn2), a large GTPase, participates in the endocytic pathway. To date, three dynamin isoforms (Dyn1, Dyn2 and Dyn3) have been identified. Dyn2 is ubiquitously expressed, whereas Dyn1 is expressed only in neuronal cells and Dyn3 in the brain, lung and testis [
18]. It has been reported that Dyn2 is necessary for embryonic development and is involved in regulating integrin endocytosis and actin filament distribution during muscle maturation [
19‐
21]. In addition to the scission of newly formed vesicles from the plasma membrane, Dyn2 is reportedly responsible for cell migration by controlling microtubule and actin cytoskeleton dynamics. Furthermore, Dyn2 has been found to be involved in metastatic activity by regulating microtubule-dependent FA dynamics [
22], and the effect of Dyn2 on cell motility has been found to be largely associated with its ability to control F-actin rearrangement, including the assembly in lamellipodia and the promotion of membrane ruffling [
23]. Moreover, recent studies have shown that the interaction of Dyn2 with α-actinin promotes invadopodia formation, resulting in pancreatic cancer cell invasion. This finding suggests that Dyn2 can act as a scaffold protein to spatiotemporally modulate the actin cytoskeleton and its regulatory proteins for cancer metastasis [
24‐
26].
In the present study, we show that βPix interacts with the proline-rich domain (PRD) of Dyn2 via the SH3 domain of βPix, and co-localizes at the membrane edge. The βPix-Dyn2 complex, promoted by the Src kinase-induced phosphorylation of tyrosine at position 442 in βPix, is crucial for Rac1-mediated membrane ruffling and CRC cell invasion. Interestingly, blockade of βPix-Dyn2 complex formation by intracellular delivery of an anti-βPix-SH3 antibody impaired cell invasion. These results suggest that disruption of the βPix-Dyn2 complex may be a therapeutic strategy for treating CRC.
2 Materials and methods
2.1 Clinical data mining
Expression profiles of
ARHGEF7, which encodes βPix, in patients with CRC were obtained from the Oncomine (
www.oncomine.org) and Gene Expression Omnibus (GEO) (
www.ncbi.nlm.nih.gov/geo) databases. For transcriptional analysis in Oncomine, data with
p < 0.0001, fold change > 2, and gene rank < 10 % were used. For GEO analysis, accession numbers GSE20916 and GSE32474 were selected. Analysis of βPix protein expression in patients with CRC was performed using the Human Protein Atlas (
www.proteinatlas.org). Additional gene expression datasets for CRC (GSE29621 and GSE14333) were downloaded for overall and recurrence-free survival analyses. The distribution of βPix-Dyn2 expression in patients with CRC across the three lymph node stages was analyzed using The Cancer Genome Atlas (TCGA; Pan Cancer Atlas, 2018) from the cBioportal database.
2.2 Antibodies and reagents
In the present study, we produced monoclonal anti-βPix-SH3 and polyclonal anti-βPix-GBD antibodies against purified GST-SH3 and GST-GBD proteins. In addition, we procured antibodies against Dyn2 (C-18, #sc-6400; Santa Cruz Biotechnology, Dallas, TX, USA), E-cadherin (24E10, #3195; Cell Signaling Technology, Danvers, MA, USA), MT1-MMP (L-15, #sc-12367; Santa Cruz Biotechnology), GAPDH (6C5, #sc-32233; Santa Cruz Biotechnology), Cortactin (H-191; #sc-11408; Santa Cruz Biotechnology), GST (B-14, #sc-138; Santa Cruz Biotechnology), c-Myc (9E10, #sc-40; Santa Cruz Biotechnology), FLAG (M2, #F1804; Sigma-Aldrich, St. Louis, MO, USA), GFP (B-2, #sc-9996; Santa Cruz Biotechnology), Rac1 (#610651; BD Transduction Laboratory, San Jose, CA, USA), and phosphotyrosine (4G10; #05-321; Millipore, Burlington, MA, USA). In addition, epidermal growth factor (EGF, E9644; Sigma-Aldrich), PP2, a Src kinase inhibitor (#529573; Calbiochem, La Jolla, CA, USA), poly L-lysine hydrobromide (#P6282; Sigma-Aldrich), fibronectin (#F2006; Sigma-Aldrich) and Matrigel (#354234; Corning, Corning, NY, USA) were used.
2.3 Plasmids
Dyn2-GFP was provided by Mark A. McNiven (Mayo Clinic and Foundation, MN, USA). Paxillin cloned into pME18S-FL3 was purchased from the Korea Human Gene Bank (KHGB, #KU016281; Daejeon, South Korea). Expression vectors were generated by digesting pFlagCMV2 (#E7033; Sigma-Aldrich), pcDNA3.1 myc/His A (#V800-20; Invitrogen, Carlsbad, CA, USA), pEGFP-C1 (#6084-1; Clontech, Palo Alto, CA, USA) and pEGFP-N1 (#6085-1; Clontech) vectors with restriction enzymes. βPix mutants, i.e., SH3 (W43K), DH, Y442F, Y442E and Dyn2 (R834A and K44A), were generated using a QuickChange™ Site-Directed Mutagenesis Kit (#200518; Agilent Technologies, Santa Clara, CA, USA) following the manufacturer’s instructions. The following lentiviral shRNA oligonucleotides were used: human βPix shRNA #1 (5′-GCAAATGCTCGTACAGTCT-3′) and shRNA #2 (5′-CGACAGGAATGACAATCAC-3′) targeting the coding region of βPix, human βPix shRNA #3 (5′-TGCGAATGGAGACGATCAAAC-3′) targeting the 3′ untranslated region (UTR) of βPix, human Dyn2 shRNA #1 (5′-ATGTAGGGCAGGCCTTCTATA-3′) targeting the 3′UTR of Dyn2, and shRNA #2 (5′-CCCGTTGAGAAGAGGCTACAT-3′) targeting the coding region of Dyn2. These shRNA oligos were cloned into a pLKO.1 vector (#10878; Addgene, Cambridge, MA, USA). For overexpressing Flag-βPix using the lentiviral system, the pLenti-G418 vector generated from pLenti-puro (#39481; Addgene) was used. All constructs were verified using DNA sequencing.
2.4 Mammalian cell culture and transfection
The human colorectal adenocarcinoma LoVo, SW480 and DLD-1 cell lines were gifted by Eok-Soo Oh (Ewha Womans University). LoVo cells were maintained in Roswell Park Memorial Institute 1640 medium (RPMI-1640; #31800022; Gibco, Grand Island, NY, USA), SW480 cells were maintained in Dulbecco’s modified Eagle’s medium: Nutrient Mixture F-12 (DMEM/F-12; #12500062; Gibco) and DLD-1 cells were maintained in DMEM (#12100046; Gibco) supplemented with 10 % heat-inactivated fetal bovine serum (FBS; #US-FBS-500; GW Vitek, Seoul, South Korea), 100 units/ml penicillin and 100 µg/ml streptomycin (#LS202-02; WelGENE, Daegu, South Korea). In addition, HEK293T cells were maintained in DMEM supplemented with 10 % FBS (#US-FBS-500; GW Vitek), 100 units/ml penicillin and 100 ug/ml streptomycin (#LS202-02; WelGENE). The cells were incubated at 37 °C in a humidified incubator with 5 % CO2. For transient transfection, 1–3 µg of plasmids was transfected into HEK293T cells using the calcium phosphate precipitation method. SW480 cells were transfected using Lipofectamine 3000 Reagent (#L3000015; Invitrogen) according to the manufacturer’s instructions.
2.5 Generation of stable cell lines using a lentiviral system
βPix and Dyn2 knockdown SW480 cell lines were generated using a lentiviral system. shRNA constructs were packaged with helper plasmids pMD2.G and psPAX2 (#12259 and #12260; Addgene), which were co-transfected into HEK293T cells. Lentiviral particles containing shRNA constructs were harvested from HEK293T cells after 72 h and infected into SW480 cells using 8 µg/ml polybrene. For establishing stable knockdown cell lines, cell selection was performed by treatment with 1 µg/ml puromycin (#P8833; Sigma-Aldrich). Depleted expression of βPix and Dyn2 was verified using Western blotting. The absence of off-target shRNA effects was verified using quantitative reverse transcription-polymerase chain reaction (RT-qPCR; Supplementary Table S1). Overexpression of Flag-βPix in LoVo cells was also performed using a lentiviral system, and the cells were selected using 500 µg/ml OmniPur® G418 Sulfate (#5.09290; Calbiochem). Overexpression of Flag-βPix was verified using Western blotting.
2.6 RT-qPCR
For isolating total RNA, SW480 cells were lysed using RNAiso Plus reagent (#9109; TaKaRa, Tokyo, Japan) according to the manufacturer’s instructions. In brief, 1 µg RNA was used to synthesize complementary DNA using PrimeScript™ reverse transcriptase (#2680; TaKaRa). qPCR was performed using SYBR Premix Ex Taq II (#RR820; TaKaRa) and QuantStudio 3 (Applied Biosystems, Foster City, CA, USA). Gene expression levels were calculated using the 2−ΔΔCt method and normalized to the Ct value of GAPDH. The primers used for qPCR are shown in Supplementary Table S2.
2.7 GST pull-down assay
GST and GST-SH3 proteins were expressed in
Escherichia coli BL21 and purified using Glutathione Sepharose 4B (#17-0756-01; GE Healthcare, Buffalo Grove, IL, USA), as described previously [
27,
28]. HEK293T cells were washed with ice-cold phosphate-buffered saline (PBS) and lysed with pull-down buffer (50 mM Tris-Cl, pH 7.4, 100 mM NaCl, 0.5 % Triton X-100, 5 % glycerol, 5 mM MgCl
2, 1 mM DTT, 20 mM NaF, 1 mM aprotinin, 1 mM leupeptin, and 1 mM pepstatin). After centrifugation at 21,000 ×
g for 15 min at 4 °C, the supernatant was incubated with 5 µg GST or GST-SH3 protein in pull-down buffer for 1 h at 4 °C and combined with 20 µl Glutathione Sepharose 4B. After incubation, the beads were washed three times with pull-down buffer, after which Western blotting was performed. For the GST-PBD pull-down assay, GST-PBD purified protein, a p21-binding domain of PAK1, was employed. HEK293T cells transfected with the indicated vectors were lysed with pull-down buffer and incubated with 5 µg GST-PBD protein for 1 h at 4 °C, followed by incubation with 20 µl Glutathione Sepharose 4B. Active Rac1 was pulled down and visualized using Western blotting with an anti-Rac1 antibody (#610651; BD Transduction Laboratory).
2.8 Immunoprecipitation
At 24–36 h post-transfection, cells were washed with ice-cold PBS and lysed with IP buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 0.5 % NP 40, 1 mM EGTA, 1 mM EDTA, 20 mM NaF, 1 mM aprotinin, 1 mM leupeptin, and 1 mM pepstatin). Then, whole-cell lysates were centrifuged at 15,000 rpm for 15 min at 4 °C, and the concentration of the lysates was determined using the Bradford protein assay (#5000006; Bio-Rad, Hercules, CA, USA). In brief, 1 mg lysate was incubated with the appropriate antibodies for 2 h, followed by incubation with Protein A-Sepharose (#P3391, Millipore) for 1 h. The immunoprecipitates were washed with IP buffer and analyzed using Western blotting.
2.9 Western blotting
SW480 cells were lysed with sodium dodecyl sulfate (SDS) lysis buffer (2 % SDS, 1 mM Na3VO4, 50 mM NaF, 1 mM DTT, and 1 mM PMSF). Then, PVDF membrane-transferred proteins were incubated with primary antibodies for 12–16 h at 4 °C. Horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) were added for 1–2 h at 24 °C. Protein signals were detected using enhanced chemiluminescence (ECL; #1705061; Bio-Rad) using the Fusion Solo S imaging system (VILBER, Collegien, France). Band densities of the proteins were measured using Evolution Capt software (VILBER).
2.10 Immunocytochemistry
HEK293T cells were seeded on 0.1 mg/ml poly L-lysine- and 10 µg/ml fibronectin-coated coverslips and incubated in media for 15 min. SW480 cells, which were plated on Matrigel-coated coverslips, were fixed with 3.7 % paraformaldehyde for 15 min and permeabilized using 0.5 % Triton X-100 in PBS for 10 min. For blocking non-specific signals, the samples were incubated with blocking solutions: 2 % bovine serum albumin and 0.1 % Triton X-100 in PBS. The samples were stained with the indicated primary antibodies for 1 h at RT and then incubated with fluorescein-conjugated secondary antibodies for 1 h at RT after washing with 0.1 % Triton X-100 in PBS. Coverslips were mounted onto slides with Fluoromount-G (#0100-01; Southern Biotechnology Associates, Birmingham, AL, USA). The samples were evaluated using an ECLIPSE 80i fluorescence microscope (Nikon, Tokyo, Japan) and a Zeiss LSM700 confocal microscope (Zeiss, Oberkochen, Germany). Images were captured using a digital camera (DS-Qi2, Nikon) and processed using NIS-Elements image analysis software (Nikon). Pearson’s coefficients analysis to quantify protein colocalization was performed using NIS-Elements image analysis software (Nikon).
2.11 3-Dimensional (3D) sphere formation assay
SW480 cells (1 × 104) were seeded in 6-well ultralow attachment plates (SPL 3D™ Cell Floater, #39706; SPL Life Sciences, Gyeonggi-do, South Korea) and incubated with media supplemented with 10 % FBS. Sphere formation was observed after seven days, and sphere diameters were measured using NIS-Elements image analysis software (Nikon). For immunocytochemistry, 3D spheres were fixed using 3.7 % paraformaldehyde for 15 min, and spheroids were added to adhesive microscope slides (HistoBond® microscope slides, #0810001; MARIENFELD, Lauda-Königshofen, Germany), followed by immunocytochemistry.
2.12 Scratch wound healing migration and Matrigel invasion assays
For scratch wound healing assays, 2 × 105 LoVo cells were seeded on 10 µg/ml Matrigel-coated 24-well culture plates with media containing 10 % FBS. After 16 h of incubation, the cells were scratched with a sterile 1-ml pipette tip and then incubated in media with 10 % FBS for 48 h. For the in vitro invasion assay, 2 × 105 SW480 cells suspended in media without FBS were seeded on 10 µg/ml Matrigel-coated upper chamber membranes of a Transwell, with a pore size of 8 μm and a diameter of 6.5 mm (#35224; SPL Life Sciences). The inserts were supplemented with media containing 20 % FBS. LoVo cells (2 × 105) suspended in media without FBS were seeded on the upper chamber membranes, supplemented with 20 % FBS. After 24 h of incubation, the cells at the bottom of the inserts were fixed with methanol and stained with 0.1 % crystal violet. The cells on the upper side of the Transwell membranes (non-invasive cells) were removed using a cotton swab and imaged randomly under 10× magnification using an Olympus CKX53 inverted microscope (Olympus, Tokyo, Japan), and the number of invasive cells was enumerated in each image.
Subsequently, we measured the activity of secreted matrix metalloproteinases (MMPs). To this end, 2 × 105 SW480 cells were incubated for 16 h at 37 °C on 10 µg/ml Matrigel-coated 12-well culture plates with media containing 10 % FBS. After incubation, the cells were washed with serum-free media and incubated in fresh serum-free media for 24 h. Conditioned (supernatant) media were harvested and centrifuged at 200 × g for 3 min at 24 °C. Next, the supernatants were incubated with 10 µM fluorogenic MMP substrates (Mca-PLGL-Dpa-AR-NH2; #ES001; R&D Systems, Minneapolis, MN, USA) for 2 h at 37 °C. The fluorescence signals of samples were measured at 320-nm excitation and 405-nm emission using a Synergy HTX Multi-Mode Reader (BioTek, Winooski, VT, USA). The fluorescence signal subtracted from the background signal was normalized by the absorbance of crystal violet-stained cells at 550 nm, fixed using 3.7 % paraformaldehyde and incubated in 2 % SDS.
2.13 Time-lapse imaging
To observe lamellipodia formation and FA dynamics, 7 × 10
4 SW480 cells were seeded on Matrigel-coated glass-bottom plates 24 h post-transfection with plasmids. Next, the cells were serum-starved for 12 h and stimulated with 100 ng/ml EGF in a 5 % CO
2 chamber at 37 °C. Images were captured under the indicated conditions using a Nikon ECLIPSE Ti2 inverted microscope system (Plan Apo λ 60X Oil) with a digital camera (DS-Qi2). To examine the rates of assembly and disassembly in individual FAs, each FA was tracked, and its intensity was measured over time, as described previously [
29,
30]. The paxillin-GFP intensities in the time series were normalized using the intensity at the first point (each time point intensity/initial intensity; assembly rate calculations) or the last point (initial intensity/each time point intensity; disassembly rate calculations), and then converted by using the log scale. The log-transformed values were fitted to linear regression models to calculate assembly and disassembly rates. For the cell migration assay, 1 × 10
5 SW480 cells were plated on Matrigel-coated plates and monitored at 30-min intervals for 12 h. Cell motility was analyzed using NIS-Elements image analysis software (Nikon). The mean square displacement was calculated according to the equation
\(: {(r\left(t\right)-r(0\left)\right)}^{2}\) =
\({(x\left(t\right)-x(0\left)\right)}^{2}+{(y\left(t\right)-y\left(0\right))}^{2}\), where
\(r\left(t\right)\) is the position of a single cell at time
\(t\) and
\(r\left(0\right)\) is the initial position. Single-cell velocity was measured as
\(\sqrt{{(x\left(i\right)-x\left(i-1\right))}^{2}+{(y\left(i\right)-y\left(i-1\right))}^{2}} / dT\) at different time points.
2.14 Preparation of gold nanoparticle (AuNP)-SH3 antibody complex
AuNPs (diameter: 15 nm) were functionalized with IgG aptamers (#NES002-01; NES, Seoul, South Korea) and conjugated with monoclonal anti-βPix-SH3 or normal mouse IgGs, as previously described [
31]. AuNP-IgG and AuNP-SH3 conjugates were incubated with βPix-GFP- or Dyn2-GFP-expressing SW480 cells on Matrigel-coated glass. Antibody-conjugated AuNPs were visualized using fluorescein-conjugated secondary antibodies. The localization of βPix-GFP was evaluated after EGF stimulation for 20 min, followed by 4 h of serum starvation. For time-lapse video imaging, Dyn2-GFP-expressing SW480 cells treated with AuNP-SH3 were monitored for 30 min at 1-min intervals under EGF stimulation.
2.15 Statistical analyses
Statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). One-way analysis of variance (one-way ANOVA) was performed to compare more than two groups. Student’s unpaired t-test was used to compare two groups. For all one-way ANOVA, Tukey’s multiple comparison test was employed as the post-hoc test. One-way ANOVA F values are displayed in each figure legend as F(DFn, Dfd) (DFn as the df nominator and Dfd as the df denominator). Statistical significance was set at p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). Data are presented as the mean ± standard deviation (S.D.).
4 Discussion
In the present study, we show that a spatiotemporal regulation of the βPix-Dyn2 interaction, by cell-matrix adhesion-mediated Src kinase activation, is crucial for CRC cell invasion. The βPix-Dyn2 complex allowed cells to increase lamellipodia formation by promoting the GEF activity of βPix to Rac1. Consequently, this process induced MT1-MMP recruitment to the leading edge of the cell membrane periphery, followed by increased invasive activity (Fig.
7h).
The function of βPix as a GEF protein for small Rho GTPases is determined by the interaction of binding proteins owing to its relatively weak GEF activity [
39]. Accordingly, various proteins that bind to the SH3 domain of βPix have been identified as cellular function regulators of βPix. For example, E3 ligase Cbl competitively binds to the SH3 domain of βPix with Pak1 to simultaneously regulate Cdc42 and EGFR signaling [
40]. In addition, βPix recruitment to the leading edge by interacting with scaffold proteins, such as Scrib and STIL, has been found to control the spatial regulation of membrane dynamics during cell migration and cell polarization [
41‐
43]. Here, we identified Dyn2 as a novel binding protein regulating the GEF activity of βPix. The molecular mechanism through which binding to Dyn2 increases the GEF activity of βPix remains unclear, but when
βPix is overexpressed in cells with inhibited
Dyn2 expression, membrane localization of βPix does not occur and βPix-dependent lamellipodia formation is suppressed. In addition, lamellipodia formation did not occur despite
Dyn2 overexpression in
βPix-silenced cells (data not shown). Therefore, the spatial interaction of βPix and Dyn2 near the plasma membrane likely provides the driving force for increasing the GEF activity of βPix, resulting in active βPix-induced Rac1 and increased membrane dynamics.
On the basis of our observations, disrupting the βPix-Dyn2 interaction using the Dyn2 R834A mutant inhibited Rac1 activation, lamellipodia formation and cell invasion, although the enzymatic activity of Dyn2 was unaltered. However, Dyn2 K44A, a mutant lacking enzymatic activity, also reduced βPix recruitment at the membrane edge, decreasing lamellipodia formation and invasion. Interestingly, immunoprecipitation assays revealed that βPix continuously interacted with the Dyn2 K44A mutant (data not shown), suggesting that the GTPase activity of Dyn2 is also required for the GEF activity of βPix. Dyn2 has been proposed as a well-defined mechanoenzyme that drives membrane fission from the plasma membrane and clathrin-mediated endocytosis at endocytic sites of the plasma membrane [
25,
44]. These Dyn2 functions are largely mediated by GTPase activity, as the GTPase-deficient Dyn2 K44A mutant results in impaired endocytosis and membrane trafficking with abnormal cellular distribution [
45,
46]. In the case of Dyn2-mediated Rac1 modulation, Dyn2 is reportedly involved in Rac1 internalization into macropinosomes, as well as Rac1 recycling into integrin at the membrane edge, where lamellipodia formation is increased via the GTPase activity of Dyn2 during cell migration [
47,
48]. However, dominant-negative Dyn2 (Dyn2 K44A) increases total Rac1 activity by inhibiting Rac1 internalization, thus accumulating active Rac1 in abnormal membrane positions and decreasing Rac1 trafficking toward newly generated lamellipodia [
48]. Our results also revealed that cellular localization of Dyn2 K44A could be observed throughout the cytoplasm with reduced membrane localization in SW480 cells while maintaining the interaction with βPix. Therefore, we postulate that if βPix binds to Dyn2, βPix can activate Rac regardless of its cellular localization. Interestingly, the GTPase activity of Dyn2 is critical for membrane targeting of the βPix-Dyn2 complex, and βPix-mediated Rac1 activation at the membrane edge is required for lamellipodia formation and CRC cell invasion. Our results indicate that spatially regulated membrane localization of the βPix-Dyn2 complex via Dyn2 GTPase activity is required for Rac1-dependent membrane dynamics during CRC cell invasion.
Furthermore, our results revealed that
βPix knockdown decreased MT1-MMP localization in the membrane periphery and inhibited lamellipodia formation in CRC cells. Reportedly, actin rearrangement for invasive migration increases the concentration of MT1-MMP in lamellipodia [
49]. In HT1080 fibrosarcoma cells in fibril gel, the expression of MT1-MMP was found to be induced by active Rac1, which in turn was induced by GEF proteins [
50]. Furthermore, βPix has been found to promote Rac3 activity in serous ovarian cancer by forming a complex with β-arrestin1/integrin-linked kinase, thus supporting MT1-MMP-dependent ECM degradation at invadopodia formation [
51,
52]. Thus, it can be speculated that βPix-mediated Rac1 activation may be required for CRC cell invasion to promote actin rearrangement, lamellipodia formation, and MT1-MMP trafficking at the peripheral area of the leading edge. Additionally, we found that the downregulated Rac1 activity caused by a disrupted βPix-Dyn2 interaction via Dyn2 PRD mutant reduced lamellipodial localization of MT1-MMP and CRC cell invasion. This finding indicates that the βPix-Dyn2 complex is essential for MT1-MMP recruitment toward the leading edge of invasive CRC cells in a Rac1-dependent manner.
During metastatic progression, tumor cells require fine-tuned phases of metastatic cascades that can sense the surrounding tumor microenvironment, degrade the ECM, and migrate through the processed matrix [
53]. Cell-ECM adhesion-dependent signals, including integrins and growth factors, temporally coordinate the activation of downstream signaling pathway factors, such as Src family kinases, FA kinase (FAK), ERK and PI3K/AKT, for tumor cell invasion [
54,
55]. Thus, we hypothesize that the formation of the βPix-Dyn2 complex is temporally regulated under cell adhesion conditions. It has been reported that βPix is phosphorylated by Src, FAK and PAK2, and then activated and recruited at lamellipodia, thus promoting downstream signaling in several model systems [
38,
56]. In particular, phosphorylation of the tyrosine 442 residue of βPix has been shown to enhance its GEF activity of Cdc42 and promote the formation of βPix-Cdc42-Cbl complex to suppress EGFR degradation [
38,
40]. Furthermore, constitutive phosphorylation of βPix Y442 in v-Src-expressing fibroblasts interferes with EGFR homeostasis and causes cell transformation, tumorigenesis, migration and invasion in
in vivo systems [
37]. In line with the importance of βPix Y442 phosphorylation in tumor progression, we also investigated whether Src kinase-induced βPix Y442 phosphorylation modulated the invasive activity of CRC cells by temporally regulating the formation of the βPix-Dyn2 complex. How was the βPix-Dyn2 interaction temporally controlled for βPix-mediated CRC invasion? One possible scenario is that cell-ECM adhesion triggers Src kinase activation, with phosphorylated βPix then promoting the interaction of Dyn2 temporally, resulting in βPix-mediated CRC invasive migration. Interestingly, it has been reported that Dyn2 is also phosphorylated by Src and modulates its GTPase activity, endocytosis, and cancer cell invasion [
57‐
59]. Thus, another possibility is that Src-induced Dyn2 phosphorylation may support the membrane localization of βPix-Dyn2. Additionally, the GTPase activity of Dyn2 involved in clathrin-mediated endocytosis is activated by Arf6-specific GEFs, EFA6B and EFA6D [
60]. Therefore, further experiments are needed to demonstrate the impact of Dyn2 phosphorylation on βPix interaction and the role of βPix in regulating Dyn2 function, and to investigate the positive feedback mechanism of Src-βPix-Dyn2 by synergistic effects on CRC cell invasion.
For efficacious cancer treatment, numerous oncogenic proteins have been considered as molecular targets to suppress the biological functions of cancerous cells, including proliferation, differentiation and metastasis [
61]. Considering that βPix functions as a tumor-promoting protein in patients with CRC, in whom βPix amplification mainly occurs among genetic mutations, we assumed that βPix could be an effective target for cancer treatment. Indeed, the delivery of anti-βPix-SH3 antibody effectively interrupted the interaction of βPix and Dyn2 and inhibited CRC invasion. Furthermore, as antibody therapy for cancer can overcome limitations of conventional chemotherapies in terms of high toxicity, weak selectivity to tumor cells and drug resistance [
62], effective targeting of βPix via the intracellular delivery of anti-βPix-SH3 antibodies could afford a potential therapeutic strategy to suppress CRC progression, even in lung and breast cancers in which βPix is reportedly overexpressed [
15,
63]. Further investigations assessing CRC progression using
in vivo systems will highlight the clinical significance of the βPix-Dyn2 complex and simultaneously reinforce our findings regarding the function of the βPix-Dyn2 complex in CRC invasion using an
in vitro human cell line model system.
In summary, our data suggest that Dyn2 serves as a novel binding partner for βPix, assisting βPix functions in CRC progression. We also verified that the spatiotemporal regulation of βPix-Dyn2 is essential for lamellipodia formation and MT1-MMP localization at the leading edge of invading cells. Thus, Dyn2-mediated spatial and Src-mediated temporal regulation of βPix activity appears to be crucial for invasive CRC migration. Furthermore, our study implies that targeting the βPix-Dyn2 complex may be an efficient strategy for the targeted treatment of cancer.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.