Substrate stiffness and the receptor-type tyrosine-protein phosphatase alpha regulate spreading of colon cancer cells through cytoskeletal contractility

Abstract

Microenvironmental clues are critical to cell behavior. One of the key elements of migration is the generation and response to forces. Up to now, there is no definitive concept on how the generation and responses to cellular forces influence cancer-cell behavior. Here, we show that expression of receptor-type tyrosine-protein phosphatase alpha (RPTPα) in human SW480 colon cancer cells sets a threshold for the response to matrix forces by changing cellular contractility. This can be explained as an RPTPα-mediated increase in contractility with a consecutive increase in number and size of adhesion sites and stress fibers. These effects are mediated through myosin light chain kinase and largely independent of Rho/Rho-kinase (ROCK) signaling. In addition, we report that RPTPα influences spreading on low-rigidity surfaces, binding of collagen-coated beads and expression of RPTPα is required for invasion into the chorioallantoic membrane. These data suggest that force-responsive proteins such as RPTPα can influence cancer-cell behavior and identify potential targets for cancer therapy.

Introduction

About 90% of all cancers originate from epithelial tissues. The most apparent morphological change is the switch from a differentiated, epithelial morphology to a migratory and invasive phenotype. During this process, cells redistribute or downregulate cell–cell contacts and gain cellular motility. In this context, the physical properties of the extracellular matrix (ECM) have a decisive function as recently shown in a series of elegant papers (Wozniak et al., 2003; Paszek et al., 2005; Engler et al., 2006). These studies show that forces exerted on or by cells affect epithelial morphogenesis, stem-cell lineage determination and transformation through cellular contractility. However, the nature of the sensor for matrix stiffness is less well characterized (Wells, 2008). Whenever cells migrate, they develop bonds with matrix molecules for remodeling or traction force generation. Physical force is critical in the development of the cell-matrix contacts at all stages. ECM molecules bind to integrins, which initiate intracellular signaling events causing integrins to connect with the cytoskeleton. As the rearward transport of the cytoskeleton develops force on the integrin-cytoskeleton connection, the connection is reinforced in concert with the binding of focal contact (FC) proteins. This process is blocked by tyrosine phosphatase inhibitors, suggesting the involvement of tyrosine phosphatases in force-transduction pathways (Choquet et al., 1997).

The family of receptor-type tyrosine-protein phosphatase alpha (RPTPα) includes a number of proteins sharing the characteristics of an extracellular domain and two intracellular PTP-homology domains (Petrone and Sap, 2000). RPTPs have been implicated in the regulation of integrin-mediated events. Overexpression of RPTPα increases substrate adhesion (Harder et al., 1998). More intriguingly, gene inactivation of RPTPα delays spreading on fibronectin and impairs activation of Src family kinases (SFK) (Su et al., 1999) (hereafter synonymously used for c-Src, c-Yes and c-Fyn). RPTPα is also required for the force-dependent reinforcement of initial integrin-cytoskeleton linkages and force-dependent formation of focal complexes and FCs in fibroblasts (von Wichert et al., 2003) and induces assembly of the actin cytoskeleton (Zeng et al., 2003). In addition, RPTPα can interact with transmembrane proteins lacking catalytic activity such as neural cell adhesion molecules or integrins (Zeng et al., 1999; von Wichert et al., 2003). However, no soluble or cell surface anchored ligands are known. A function of RPTPα in force-dependent signaling processes is supported by studies showing the requirement of RPTPα for the rigidity response at the leading edge in fibroblasts (Jiang et al., 2006). On activation, RPTPα associates with SFK (Harder et al., 1998; Su et al., 1999) and causes SFK activation (den Hertog et al., 1993). Interestingly, it has been speculated that the ability of RPTPα to activate SFK is responsible for its capacity to transform fibroblast (Zheng et al., 1992). In addition, a supporting function for RPTPα in generation of a malignant phenotype is suspected as increased expression of RPTPα has been reported for various cancers (Tabiti et al., 1995; Wu et al., 2006). Nonetheless, the pathophysiological function of RPTPα expression during cancer progression remains elusive.

Here, we show that RPTPα is expressed in colon cancer and that expression of RPTPα regulates the formation of FC and stress fibers in colon cancer cells through cytoskeletal contractility. The RPTPα-mediated increase in contractility is mediated by SFK and myosin light chain kinase (MLCK) through a Rho/Rho-kinase (ROCK)-independent pathway. RPTPα increases the accumulation of adhesion-site proteins and improves the recognition of suboptimal mechanical stimuli. We can show that RPTPα is necessary for cell spreading on less favored, low-rigidity surfaces and for invasion in an in vivo model. Taken together, these data describe RPTPα for the first time as a regulator of cellular contractility and suggest RPTPα as a new target for the development of anti-cancer strategies.

Results

RPTPα is expressed in colon cancer samples and transiently localizes at the leading edge during cell spreading

Tissue-arrays containing 50 colorectal cancer specimens and 10 normal colon samples were subjected to immunohistochemistry for RPTPα expression. In normal tissue samples, relevant RPTPα expression was restricted to smooth muscles cells. None of the normal colonocytes expressed RPTPα in detectable quantities. In contrast, over 70% of the colon cancer samples showed expression of RPTPα (Figure 1a (arrows)). These data suggest that RPTPα is expressed in colon cancer.

Figure 1
figure1

RPTPα is expressed in human colon cancer samples and RPTPα transiently localizes to areas of new matrix contacts. (a) Left: paraffin sections (each slide containing 50 cancer samples and 10 samples of normal colon mucosa) were stained with monoclonal anti-RPTPα antibodies. White arrows indicate cells expressing RPTPα (tumor cells or myo-fibroblasts). Black arrows indicate cells not expressing RPTPα (normal colonocytes). H&E counterstain. Right: shown is the percentage of tumor samples expressing RPTPα (b) SW480 cells were transiently transfected with YFP-RPTPα (0.5 μg/ml), plated on collagen-I for 120 min and overnight and subsequently stained for paxillin and β1-integrins. Nuclei were counterstained with DAPI.

Next, subcellular localization of RPTPα in colon cancer cells was determined. Therefore, SW480 cells were transfected with an RPTPα-yellow fluorescent protein (YFP) construct. For spreading, SW480 colon cancer cells were plated on collagen-I. Cells were analyzed 120 min and 24 h after plating. Interestingly, RPTPα colocalized with paxillin and β1-integrins at the leading edge on collagen-I 120 min after plating. After 24 h, RPTPα was diffusely localized in the membrane or localized to the tip of filopodia, whereas paxillin and β1-integrins localized to FC (Figure 1b). Thus, there is evidence of a transient colocalization of RPTPα, paxillin and β1-integrins during the early phases of matrix interaction in spreading colon cancer cells.

RPTPα mediates the formation of adhesion sites and stress fibers through activation of SFK

It has been suggested that RPTPα contributes to the assembly of FC and the actin cytoskeleton in fibroblasts (von Wichert et al., 2003; Zeng et al., 2003). SW480 endogenously expressed RPTPα at moderate levels (Figure 2b). To analyze the effect of RPTPα in these cells, SW480 cells were transfected with pcDNA3-paxillin-green fluorescent protein (GFP) together with specific siRNA oligonucleotides directed against RPTPα (effects were the same for two different sequences (up to 70% reduction of endogenous protein levels)) or pcDNA3-RPTPα expression vectors (up to 250% increase over endogenous protein levels) (Figure 2b) and were plated on collagen-I for 120 min. Interestingly, there was a marked reduction in the number and size of FC in knockdown cells and no obvious filamentous actin structures could be detected (Figure 2a). In contrast, in cells overexpressing RPTPα, even more and larger FC as well as an increase in the number of stress fibers could be detected (Figure 2a). To test whether RPTPα affects the activation of downstream targets such as SFK, cells were again transfected with different siRNA oligonucleotides. After suspension, the cells were seeded on collagen-coated culture dishes and lysed at the time points indicated in the figure legends. Interestingly, there was a pronounced reduction in the SFK activity after knockdown of RPTPα, as judged by autophosphorylation. In contrast, prolonged and pronounced autophosphorylation of SFK was observed after overexpression of RPTPα (Figure 2c). To address the RPTPα-dependent function of SFK activation, cells were co-transfected with paxillin-GFP and C-terminal Src kinase (CSK). CSK counteracts RPTPα and inactivates SFK (Figure 2d, middle). In good agreement, CSK prevented the formation of adhesion sites and stress fibers (Figure 2d, left and right). These data suggest that RPTPα acts as an upstream regulator of SFK after plating on collagen and that RPTPα-dependent activation of SFK is required for the formation of FC and the actin cytoskeleton in these cells.

Figure 2
figure2

RPTPα regulates the formation of adhesion sites through SFK. (a) SW480 cells were transfected with paxillin-GFP together with pcDNA3-RPTPα or with RPTPα-specific siRNA oligonucleotides. Cells were spread on collagen-I, stained with phalloidin and the number of FCs was determined. (b) SW480 cells were transfected with pcDNA3-RPTPα or with RPTPα-specific siRNA oligonucleotides. Cells were lysed and equal amounts of protein were analyzed by western blotting using anti-RPTPα or anti-actin antibodies (c) Cells were treated as in (a) and plated on collagen-I. Cells were lysed at the time points indicated and equal amounts of protein were analyzed by western blotting using anti-Src and autophosphorylation-site-specific anti-SFK antibodies. (d) SW480 cells were transfected with paxillin-GFP and pCMV5-CSK with or without co-transfection of pcDNA3-RPTPα. Cells were spread on collagen-I and stained with phalloidin and the number of FC was determined. The effect of pCMV5-CSK on the activation state of SFK was controlled by staining cells with autophosphorylation-site-specific antibodies (d, middle). Where indicated, results are the mean (percent control)±s.d. of at least three independent experiments (n=>50 cells/experiment—if applicable) (**P<0.025; ***P<0.01).

RPTPα effects on the cytoskeleton are mediated by MLCK and independent from Rho/ROCK signaling

As the formation of stress fibers as well as the formation of FC is tightly associated with the force acting on the cytoskeleton, we were interested whether the effects of RPTPα were mediated by signaling pathways involving contractility. Actomyosin-based contractility is controlled by GTPases such as Cdc42, Rac and RhoA. RhoA activates ROCK, which in turn phosphorylates and inactivates the phosphatase that dephosphorylates myosin II regulatory light chain, resulting in increased contractility. To analyze the effect of RPTPα on Rho/ROCK-mediated contractility, cells were co-transfected with dominant-negative RhoA (N19RhoA) or were pretreated with a selective Inhibitor of ROCK, Y27632 (5 μM). Interestingly, inhibition of Rho and ROCK caused a significant decrease in the number of large adhesions after spreading and the FC resembled focal complexes. However, neither the inhibition of Rho nor direct inhibition of ROCK prevented RPTPα effects on the formation of stress fibers and FC (Figures 3a and b). Cells pretreated with an inhibitor of MLCK (Ml-7: 5 μM) also displayed smaller and less FC and stress fibers after spreading. However, when cells were co-transfected with RPTPα, there was no increase in the number of FC and stress fibers (Figure 3c). The effects of RPTPα knockdown and overexpression on the phopshorylation status of myosin light chain (MLC) and the ROCK substrate myosin phosphatase target subunit isoform 1 (MYPT1) as well as the specificity of the inhibitors were tested by western blotting (Figure 3d). Taken together, these data suggest that RPTPα effects on FC and the cytoskeleton are mediated through MLCK, but most probably independent of Rho/ROCK signaling.

Figure 3
figure3

RPTPα initiates the formation of adhesion sites through an MLCK-mediated signal-transduction pathway. (a) SW480 cells were transfected with paxillin-GFP and pUSERhoAT19N with or without pCDNA3-RPTPα. Cells were spread on collagen-I, stained with phalloidin and the number of FC was determined. (b) Cells were treated as in (a), but were pre-incubated with a specific inhibitor of ROCK (Y-27632: 5 μM) for 4 h and were spread in the presence of the inhibitor. (c) Cells were treated as in (a), but were incubated with a specific inhibitor of MLCK (Ml-7: 5 μM). (d) Left: SW480 cells were transfected with GFP and either co-transfected with pCDNA3-RPTPα or transfected with RPTPα-specific siRNA oligonucleotides or incubated with Y-27632 (5 μM) or Ml-7 (5 μM). Equal amounts of protein were analyzed by western blotting using myosin light chain (MLC) S19P or pan-MLC antibodies. Right: cells were treated as in (d). Left ROCK-kinase assays were performed by exposure of a recombinant myosin phosphatase target subunit isoform 1 (MYPT1) fragment to the ROCK immunoprecipitates. ROCK activity was quantified by western blotting for MYPT1 T696P. Where indicated, results are the mean (percent control)±s.d. of at least three independent experiments (n=>50 cells/experiment, if applicable) (*P<0.05; **P<0.025; ***P< 0.01).

RPTPα regulates cellular contractility through SFK and MLCK

It has been shown earlier that contractile force influences the formation of FC and stress fibers (Balaban et al., 2001). Therefore, we were interested whether RPTPα acts through changes in contractility. Thus, cells were plated on collagen-I-coated silicon films and the number of cells able to wrinkle the substrate was quantified (Figure 4, upper panel). Under baseline conditions, 25% of the cells were able to wrinkle the substrate. In contrast, cells transfected with siRNA oligonucleotides showed a significant reduction of cellular contractility. Vice versa, when cells were co-transfected with RPTPα and EGFP, contractility increased to 48% of all cells. Interestingly, when SFK activation was inhibited by CSK-co-transfection, basal contractility decreased. Furthermore, when CSK was co-transfected with pcDNA3-RPTPα, there was no increase in contractility (Figure 4, lower panel). These data suggest that RPTPα regulates contractility through SFK. As we have shown before that RPTPα-mediated formation of FC and stress fibers is regulated through MLCK, but independent of Rho/ROCK signaling, we were interested whether this is reflected in the contractile behavior. Indeed, when cells were pretreated with the ROCK inhibitor, contractility decreased, confirming the need of ROCK for the maintenance of a contractile phenotype. However, when cells were co-transfected with pcDNA3-RPTPα and EGFP, cells were able to increase contractility in the presence of the ROCK inhibitor. In contrast, when MLCK was inhibited, contractile force decreased and co-transfection of RPTPα did not increase contractility (Figure 4, lower panel). Taken together, these data suggest that the RPTPα effects seen on the cytoskeleton are paralleled by an increase in cellular contractility.

Figure 4
figure4

RPTPα controls cellular contractility through a RhoA/ROCK-independent and MLCK-mediated signal-transduction pathway. Upper panel: SW480 cells were either transfected with EGFP with or without co-transfection of pcDNA3-RPTPα, or with RPTPα-specific siRNA oligonucleotides together with fluorescent nonsense oligonucleotides. Alternatively, cells were transfected with EGFP together with the indicated expression vectors or siRNA oligonucleotides. In addition, cells with or without co-expression of pcDNA3-RPTPα were preincubated with specific inhibitors of ROCK (Y-27632: 5 μM) or MLCK (Ml-7: 5 μM). Cells were spread overnight on collagen-I-coated silicon sheets. Effects were quantified by counting the number of cells able to deform the substrate. Lower panel: the percentage of contractile cells under the conditions indicated. Shown is the mean ± s.d. of at least three independent experiments (**P<0.025; ***P<0.01).

RPTPα regulates recognition of physical stimuli and stimulates formation of adhesion sites through contractility

To further analyze whether RPTPα influences the way cells sense physical stimuli, we incubated collagen-I-coated beads with cells. Large beads (>3 μm) stimulate FC assembly at the bead site, independent of mechanical restraint. In contrast, cells usually do not recognize small beads (<3 μm) unless mechanically restrained (von Wichert et al., 2003). Indeed, when cells were challenged with large beads (5.9 μm diameter), they accumulated paxillin to the bead-binding site (Figure 5, upper panel). Interestingly, when RPTPα expression was silenced, the number of cells accumulating paxillin-GFP at the bead site decreased. In contrast, when RPTPα was overexpressed, an increase of cells accumulating paxillin-GFP was evident. In addition, this effect depended on the activation of SFK as judged by co-transfection of CSK. In agreement with the data on the formation of FC, accumulation depended on MLCK-mediated contractility. Treatment with Ml-7 significantly reduced the number of cells accumulating paxillin-GFP at the bead-binding site. In contrast, although inhibition of ROCK reduced bead binding under basal conditions, RPTPα could overcome the inhibition of ROCK, resulting in increased accumulation (quantitative analysis: Figure 5, lower panel). To further analyze this phenomenon, we challenged the cells with beads (1.75 μm) that are not recognized by cells under normal circumstances (Figure 6, upper panel). Interestingly, overexpression of RPTPα increased the ability of cells to accumulate paxillin around the binding site of small beads. In contrast, silencing RPTPα expression prevented accumulation of paxillin-GFP. Again, RPTPα effects were mediated by SFK/MLCK and were independent of ROCK (quantitative analysis: Figure 6, lower panel). Taken together, these data suggest that the level of RPTPα expression is an important parameter in the recognition of physical stimuli. Thus, the presence of high levels of RPTPα allows cells to respond to suboptimal levels of stimulation.

Figure 5
figure5

RPTPα regulates reinforcement of initial integrin-cytoskeleton linkages. Upper panel: SW480 cells were transfected with paxillin-GFP together with pcDNA3-RPTPα or with RPTPα-specific siRNA oligonucleotides. Alternatively, cells were transfected with paxillin-GFP and pCMV5-CSK with or without co-transfection of pcDNA3-RPTPα. In addition, cells with or without co-expression of pcDNA3-RPTPα were preincubated with specific inhibitors of ROCK (Y-27632: 5 μM) or MLCK (Ml-7: 5 μM) and were spread on collagen-I. Large beads (5.9 μm) coated with collagen were incubated with cells and the number of cells with beads causing accumulation of paxillin-GFP was determined. Pictures shown are fluorescence images (top), the corresponding DIC-images (middle) and the merge (bottom). Arrows indicate the position of the bead. Lower panel: the percentage of cells able to accumulate paxillin-GFP around beads under the conditions indicated. Shown is the mean±s.d. of at least three independent experiments (*P<0.05; **P<0.025; ***P<0.01).

Figure 6
figure6

RPTPα ameliorates the recognition of suboptimal physical stimuli. Upper panel: SW480 cells were transiently transfected with paxillin-GFP together with pcDNA3-RPTPα or with RPTPα-specific siRNA oligonucleotides. Alternatively, cells were transiently transfected with paxillin-GFP and pCMV5-CSK with or without co-transfection of pcDNA3-RPTPα. In addition, cells were pre-incubated with the indicated inhibitors and were spread on collagen-I. Small beads (1.75 μm) coated with collagen were incubated with cells and the number of cells with beads causing accumulation of paxillin-GFP was determined. Pictures shown are fluorescence images (top), the corresponding DIC-images (middle) and the merge (bottom). Arrows indicate the position of the bead. Lower panel: the percentage of cells able to accumulate paxillin-GFP around beads under the conditions indicated. Shown is the mean±s.d. of at least three independent experiments (**P<0.025; ***P<0.01).

RPTPα regulates spreading on elastic surfaces

The human body consists of different tissues with different rigidities and it has been shown that cells sense the substrate stiffness (Ingber, 2003). To analyze RPTPα-dependent effects during this process, SW480 cells were plated for 24 h on collagen-I-coated glass (very stiff) or collagen-I-coated polyacrylamide gels with different rigidities (40 N/m2/0.4% bis-acrylamide vs 4 N/m2/0.05% bis-acrylamide) (Figure 7). The spreading on glass surfaces was reduced significantly after knockdown of RPTPα. In contrast, overexpression led to a slight increase in spreading, suggesting that the spreading process is an already maximally active process. In addition, spreading on high-rigidity surfaces as well as spreading on low-rigidity surfaces greatly depended on RPTPα, as expression of RPTPα led to a significant increase of spread cells on low-rigidity surfaces (Figure 7a). Moreover, as shown before, RPTPα acted through an SFK/MLCK-mediated increase in contractility (Figure 7b). These data are in good agreement with the data obtained with the beads and suggest RPTPα as a key sensor for matrix rigidities in these cells.

Figure 7
figure7

RPTPα controls spreading of cell to surfaces of different rigidities and ameliorates the adhesion to pliable surfaces. (a) SW480 cells were transiently transfected with paxillin-GFP and were co-transfected either with pcDNA3-RPTPα or with RPTPα-specific siRNA oligonucleotides. Alternatively, cells were transiently transfected with paxillin-GFP and pCMV5-CSK with or without co-transfection of pcDNA3-RPTPα. (b) Cells with or without co-expression of pcDNA3-RPTPα were preincubated with specific inhibitors of ROCK (Y-27632: 5 μM) or MLCK (Ml-7: 5 μM). All cells (a, b) were plated on collagen-I-coated glass or polyacrylamide gels containing different concentrations of bis-acrylamide (stiff 0.4% (40 N/m2) and soft 0.05% (4 N/m2)). The number of spread cells was quantified 24 h after plating. Shown is the mean±s.d. of at least three independent experiments (*P<0.05; **P<0.025; ***P<0.01).

RPTPα regulates invasion and adhesion in vivo

To transfer these findings into an in vivo system, we used the chicken chorioallantoic membrane (CAM) as a model for invasion (Kunzi-Rapp et al., 2001). For the CAM assay, we generated cells stably transfected with siRNA vectors reaching a 90% knockdown of RPTPα mRNA, which resulted in a complete loss of detectable protein. In addition, cells stably overexpressing RPTPα were generated (Figure 8c). These cells were inoculated on the CAM and tumor formation was assessed after 4 days of incubation by histological analysis (Figure 8a). The control-SW480 colon carcinoma cells proliferated on top of the CAM and were able to penetrate into the stroma underlying the CAM (asterisk). Similarly, SW480 cells overexpressing RPTPα grew on the CAM and were also able to invade the CAM-stroma, although the invading cells seemed to be only loosely connected (Figure 8a, lower panel). Interestingly, RPTPα knockdown cells showed a decreased ability to grow on the CAM together with a drastically diminished invasion into the CAM (Figure 8, middle panel), although cell growth in vitro was comparable for all lines (data not shown). A quantitative analysis of 18 independent CAM assays is summarized in Figure 8b. Taken together, these data strongly suggest that the presence of RPTPα facilitates the invasion of cancer cells into the CAM-stroma.

Figure 8
figure8

RPTPα controls invasion in the CAM model. (a) Left: SW480 cells were stably transfected with different shRNA-plasmids targeting RPTPα (clone 1 and 2), pcDNA3-RPTPα or empty vector. A total of 1 × 106 tumor cells were implanted onto the CAM. Tumor cells with the surrounding CAM were sampled 4 days after seeding by H&E staining. Shown are representative images out of two independent eggs. Black arrows indicate cancer cells, the asterisk indicates the CAM-stroma right: schematic diagram explaining the CAM invasion assay. (b) Quantitative analysis of the CAM invasion assay. (c) SW480 cells were stably transfected with different shRNA-plasmids targeting RPTPα (two independent clones), pcDNA3-RPTPα or empty vector. Equal amounts of protein were analyzed by western blots using anti-RPTPα antibodies. Protein content was controlled by western blotting for actin. Knockdown of mRNA expression was quantified using real-time PCR. Where indicated, results are the mean (percent control)±s.d. of at least three independent experiments (**P<0.025; ***P<0.01).

Discussion

The concept that physical stimuli can modulate cellular behavior has been shown for several cell types. There is increasing evidence that cancer cells (Thamilselvan and Basson, 2004) or transformed fibroblasts (Munevar et al., 2001) respond to forces. The importance of mechanical stimuli in the initiation of focal-complex formation, the requirement of mechanical strain for maturation of a focal complex to an FC, as well as the necessity of force for the maintenance of FC underline an important function of force for ECM interaction (Balaban et al., 2001; Wang et al., 2005). Interestingly, it has been suggested that the adhesiveness of cancer cells correlates inversely with the prognosis of cancer (Benoliel et al., 2003). Recent evidence further suggests that increased force in tissues can support malignant transformation (Paszek et al., 2005). These observations suggest that force-activated signaling pathways affect cancer-cell behavior.

It has been speculated that force-induced transformation of a physical signal into a biochemical response takes place in enzymatic as well as in structural components. Although the precise nature of the force sensor remains elusive, a number of candidates have been proposed (von Wichert et al., 2003; Han et al., 2004; Sawada et al., 2006). RPTPα has been recently associated with the transmission of physical stimuli into biochemical signals. However, a detailed characterization of RPTPα-dependent signaling and its effects on cellular behavior in cancer cells is missing to date.

Here, we showed that expression of RPTPα is crucial for the interaction of epithelial cells with collagen-I surfaces in which RPTPα transiently localized to areas of new matrix contacts and colocalized with FC markers and integrins. It is important to note that the interaction of RPTPα with other transmembrane receptors seems not to be exclusive, as interaction has been shown for several proteins (Zeng et al., 1999; von Wichert et al., 2003; Bodrikov et al., 2005). RPTPα induced the formation of adhesion sites and stress fibers through sustained activation of SFK and subsequent activation of myosin through MLCK on collagen-I. These effects were largely independent from the RhoA/ROCK-mediated regulation of myosin phosphatase. Whether this is a direct effect of RPTPα/SFK on MLCK or whether additional intermediates are required needs further analysis. However, RPTPα-dependent activation of SFK was necessary for a matrix-dependent increase in cytoskeletal contractility. The data presented here are also reflected in findings by Totsukawa showing that MLCK-inhibition blocked MLC phosphorylation at the cell periphery, but not in the center. In contrast, ROCK-inhibition blocked MLC phosphorylation in the center, but not in the periphery (Totsukawa et al., 2004). Thus, the spatial regulation of MLC phosphorylation is critical and RPTPα might account for this phenomenon. The local increase in contractility enhanced the ability of cells to respond to physical stimuli that otherwise would not trigger formation of adhesion sites. It has been suggested earlier that RPTPα and integrins form a rigidity–responsive complex at the leading edge (Jiang et al., 2006). Interestingly, the rigidity response correlated with the recruitment of SFK to early adhesions (Kostic and Sheetz, 2006). These observations are in line with a threshold-dependent formation of adhesion sites according to a ligand-receptor signaling threshold model. Such a model envisages that the fraction of cells forming adhesion sites depends on the numbers of signal–receptor complexes as compared with a given threshold level. If a cell is expressing the receptor in a sufficient number and is exposed to sufficient concentrations of the cognate signal (here force), the number of signal–receptor complexes is above a certain threshold and the cell will form adhesion sites. Conversely, if a cell expresses too few receptors or if the signal level is low, the number of signal–receptor complexes decreases below the threshold and adhesion will be significantly reduced. This model suggests that cells expressing high levels of RPTPα, such as colon cancer cells, are able to overcome limitations of signal intensity (matrix force) by increased receptor levels. Although it has been shown that expression of RPTPα correlates with low tumor grade in breast cancers (Ardini et al., 2000), there are several reports that support a pro-malignant function of RPTPα. Overexpression of RPTPα is observed in advanced colon carcinoma (Tabiti et al., 1995) and expression of RPTPα correlates with gastric cancer progression including lymphovascular invasion and peritoneal dissemination (Wu et al., 2006). In addition, dedifferentiation and stroma recruitment was associated with an upregulation of RPTPα expression in squamous cell carcinoma (Berndt et al., 1999). Moreover, overexpression of RPTPα in fibroblasts results in persistent activation of pp60c-src kinase, with concomitant cell transformation and growth in soft agar (that is growth in a surrounding, which is not supporting significant cellular forces) (Zheng et al., 1992). In good agreement, RPTPα expression was tightly associated with the ability of cells to adhere to pliable surfaces and to invade into the stroma of a CAM.

Taken together, these findings strongly suggest that RPTPα sets the threshold for the force-dependent formation of adhesion sites through SFK/MLCK-mediated contractility. Consequently, RPTPα dramatically influenced the way cells dealt with suboptimal surfaces and hence allowed stromal invasion. These findings suggest RPTPα as a new target for interfering with the metastatic behavior of cancer cells.

Materials and methods

Cell culture and RNA interference

SW480 cells (ATCC) were maintained in DMEM (containing 10% (v/v) fetal calf serum) in a humidified atmosphere of 5% CO2:95% air at 37 °C and passaged every 4 days. Transient knockdown was achieved by using functionally validated Stealth (RNAi Invitrogen, Carlsbad, CA, USA) directed against PTPRA. Stable clones were generated by transfection of SW480 cells with SureSilencing-shRNA plasmids (BioMol, Hamburg, Germany) for human PTPRA, followed by selection of positive clones (for details see Supplementary methods).

CAM and tissue micro-array assay

The CAM assay was carried out as described earlier (Kunzi-Rapp et al., 2001). Briefly, 1 × 106 tumor cells in 25 μl Matrigel matrix (BD Biosciences, Franklin Lakes, NJ, USA) were implanted onto the CAM of fertilized chicken eggs on day 8 of incubation. Tumor cells with the surrounding CAM were sampled 4 days after seeding. Sections were stained with H&E for histopathology. For tissue micro-array (TMA) assays, paraffin sections (each slide containing 40 colon cancer samples, 10 metastases and 10 samples of normal colon mucosa (Imgenex, San Diego, CA, USA)) were stained after antigen retrieval with monoclonal anti-RPTPα antibodies (BD Biosciences) (1:50). Antibody binding was visualized using a biotinylated secondary antibody with avidine-conjugated peroxidase (Vector Laboratories, Burlingame, CA, USA) and DAB as a substrate (H&E counterstaining). Image acquisition was performed using an inverted microscope (Olympus IX71) (Olympus, Hamburg, Germany) connected to an RGB-camera (Qicam, Qimaging, Canada).

Western blotting

Cells were lysed in 1% NP-40 lysis buffer, protein concentrations were adjusted and Laemmli buffer was added to the lysates. Proteins were further analyzed by SDS–PAGE followed by western blotting using polyclonal phosphospecific-anti-SFK(Y416) (Cell Signaling, Beverly, MA, USA), phosphospecific-anti-myosin(S19)-, anti-myosin (both Chemicon, Temecula, CA, USA) and anti-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies or monoclonal anti-Src (Oncogene, San Diego, CA, USA) and anti-RPTPα antibodies (BD Biosciences) with immunoreactive bands being visualized by enhanced chemiluminescence detection (Pharmacia, Piscataway, NJ, USA).

ROCK-kinase assay

Immunoprecipitation was performed as described (Sahai and Marshall, 2002). Briefly, cells were lysed and lysates were precleared by incubation with protein G sepharose (Pharmacia). Precleared lysate was incubated with 5 μl of a ROCK-II antibody (Santa Cruz Biotechnology) for 30 min, followed by incubation with 50 μl of protein G sepharose beads. Beads were then washed and resuspended in kinase assay buffer. ATP (Sigma, St Louis, MO, USA) and recombinant MYPT1 substrate (Upstate, Lake Placid, NY, USA) were incubated with the bead-kinase assay buffer slurry in a reaction volume of 50 μl for 30 min at 37 °C, then stopped by SDS–PAGE buffer addition and boiled for 10 min at 95 °C. Kinase activity was detected by western blotting using phospho-MYPT1 (threonine 696) antibodies (Upstate).

Immunocytochemistry

Cells were transiently transfected with pRK5GFP-paxillin (von Wichert et al., 2003), pCMV-CSK (von Wichert et al., 2003), pSGrPTPa-516-YFP (von Wichert et al., 2003), pcDNA3-RPTPα or pUSERhoAT19N (Upstate) (all 0.5 μg/ml) using Fugene 6 (Roche, Indianapolis, IN, USA) or pretreated for 4 h with an inhibitor of ROCK (Y-27632: 5 μM) or MLCK (ML-7: 5 μM) (Chemicon). Cells were suspended for 60 min and plated for 120 min on collagen-I before fixation. For co-staining, cells were incubated with monoclonal anti-paxillin antibody (Transduction Laboratories, San Diego, CA, USA) for 1 h followed by detection with an Alexa-labeled (647 nm) secondary antibody (Molecular-Probes, Eugene, OR, USA). Image acquisition was performed using an inverted microscope (Olympus IX71) (Olympus) connected to a CCD camera (Orca, Hamamatsu, Japan). Image analysis was performed using ImageJ (v1.38).

Spreading assays on silicon sheets and polyacrylamide substrates

To visualize cellular contractility, flexible silicone substrates (dimethyl diphenyl polysiloxane, Sigma) were generated by a modification of the protocol of Harris et al. (1981). Cells were plated onto collagen-coated silicon films (collagen-I (0.1 mg/ml) (Sigma)) and contractility was assessed by formation of wrinkles. Assays were used for quantification only when 20–30% of untreated cells were able to deform the substrate. Polyacrylamide substrates of different rigidity were prepared and characterized as described earlier (Wang and Pelham, 1998). Cells were plated onto the gels for 24 h and 20 serial fields were counted ( × 20 magnification). Cells appearing phase bright were counted as non-spread, cells appearing phase dark were counted as spread. Cells that did not fit into either category were not counted. Statistical analysis was performed using ANOVA followed by a Student–Newman–Keuls post hoc test (for details see Supplementary methods).

Bead assays

A total of 1.75 and 5.9 μm beads (Polyscience, Niles, IL, USA) were coated with Collagen-I or BSA (Sigma) as described earlier (Lee et al., 1996). For bead-binding assays, paxillin-GFP-transfected cells were briefly lifted as described above and were subsequently plated on collagen-I-coated cover glass for 120 min. Then the collagen-I-coated beads were added and incubated for 60 min followed by fixation. Bead binding was assessed by fluorescence microscopy and images were analyzed using ImageJ. Beads were scored positive when there was a clearly visible accumulation of paxillin around the bead that was at least above twofold the surrounding fluorescence intensity. Statistical analysis was performed using ANOVA followed by a Student–Newman–Keuls post hoc test.

Abbreviations

CAM:

chorioallantoic membrane

CSK:

C-terminal Src kinase

ECM:

extracellular matrix

FC:

focal contact

GFP:

green fluorescent protein

MLCK:

myosin light chain kinase

MLC:

myosin light chain

MYPT1:

myosin phosphatase target subunit isoform 1

ROCK:

Rho kinase

RPTPα:

receptor-type tyrosine-protein phosphatase alpha

SFK:

Src family kinases

TMA:

Tissue micro-array

YFP:

yellow fluorescent protein

References

  1. Ardini E, Agresti R, Tagliabue E, Greco M, Aiello P, Yang LT et al. (2000). Expression of protein tyrosine phosphatase alpha (RPTPalpha) in human breast cancer correlates with low tumor grade, and inhibits tumor cell growth in vitro and in vivo. Oncogene 19: 4979–4987.

  2. Balaban NQ, Schwarz US, Riveline D, Goichberg P, Tzur G, Sabanay I et al. (2001). Force and focal adhesion assembly: a close relationship studied using elastic micropatterned substrates. Nat Cell Biol 3: 466–472.

  3. Benoliel AM, Pirro N, Marin V, Consentino B, Pierres A, Vitte J et al. (2003). Correlation between invasiveness of colorectal tumor cells and adhesive potential under flow. Anticancer Res 23: 4891–4896.

  4. Berndt A, Luo X, Bohmer FD, Kosmehl H . (1999). Expression of the transmembrane protein tyrosine phosphatase RPTPalpha in human oral squamous cell carcinoma. Histochem Cell Biol 111: 399–403.

  5. Bodrikov V, Leshchyns'ka I, Sytnyk V, Overvoorde J, den Hertog J, Schachner M . (2005). RPTPalpha is essential for NCAM-mediated p59fyn activation and neurite elongation. J Cell Biol 168: 127–139.

  6. Choquet D, Felsenfeld DP, Sheetz MP . (1997). Extracellular matrix rigidity causes strengthening of integrin-cytoskeleton linkages. Cell 88: 39–48.

  7. den Hertog J, Pals CE, Peppelenbosch MP, Tertoolen LG, de Laat SW, Kruijer W . (1993). Receptor protein tyrosine phosphatase alpha activates pp60c-src and is involved in neuronal differentiation. EMBO J 12: 3789–3798.

  8. Engler AJ, Sen S, Sweeney HL, Discher DE . (2006). Matrix elasticity directs stem cell lineage specification. Cell 126: 677–689.

  9. Han B, Bai XH, Lodyga M, Xu J, Yang BB, Keshavjee S et al. (2004). Conversion of mechanical force into biochemical signaling. J Biol Chem 279: 54793–54801.

  10. Harder KW, Moller NP, Peacock JW, Jirik FR . (1998). Protein-tyrosine phosphatase alpha regulates Src family kinases and alters cell-substratum adhesion. J Biol Chem 273: 31890–31900.

  11. Harris AK, Stopak D, Wild P . (1981). Fibroblast traction as a mechanism for collagen morphogenesis. Nature 290: 249–251.

  12. Ingber DE . (2003). Mechanobiology and diseases of mechanotransduction. Ann Med 35: 564–577.

  13. Jiang G, Huang AH, Cai Y, Tanase M, Sheetz MP . (2006). Rigidity sensing at the leading edge through alphavbeta3 integrins and RPTPalpha. Biophys J 90: 1804–1809.

  14. Kostic A, Sheetz MP . (2006). Fibronectin rigidity response through Fyn and p130Cas recruitment to the leading edge. Mol Biol Cell 17: 2684–2695.

  15. Kunzi-Rapp K, Genze F, Kufer R, Reich E, Hautmann RE, Gschwend JE . (2001). Chorioallantoic membrane assay: vascularized 3-dimensional cell culture system for human prostate cancer cells as an animal substitute model. J Urol 166: 1502–1507.

  16. Lee W, Sodek J, McCulloch CA . (1996). Role of integrins in regulation of collagen phagocytosis by human fibroblasts. J Cell Physiol 168: 695–704.

  17. Munevar S, Wang Y, Dembo M . (2001). Traction force microscopy of migrating normal and H-ras transformed 3T3 fibroblasts. Biophys J 80: 1744–1757.

  18. Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A et al. (2005). Tensional homeostasis and the malignant phenotype. Cancer Cell 8: 241–254.

  19. Petrone A, Sap J . (2000). Emerging issues in receptor protein tyrosine phosphatase function: lifting fog or simply shifting? J Cell Sci 113: 2345–2354.

  20. Sahai E, Marshall CJ . (2002). ROCK and Dia have opposing effects on adherens junctions downstream of Rho. Nat Cell Biol 4: 408–415.

  21. Sawada Y, Tamada M, Dubin-Thaler BJ, Cherniavskaya O, Sakai R, Tanaka S et al. (2006). Force sensing by mechanical extension of the Src family kinase substrate p130Cas. Cell 127: 1015–1026.

  22. Su J, Muranjan M, Sap J . (1999). Receptor protein tyrosine phosphatase alpha activates Src-family kinases and controls integrin-mediated responses in fibroblasts. Curr Biol 9: 505–511.

  23. Tabiti K, Smith DR, Goh HS, Pallen CJ . (1995). Increased mRNA expression of the receptor-like protein tyrosine phosphatase alpha in late stage colon carcinomas. Cancer Lett 93: 239–248.

  24. Thamilselvan V, Basson MD . (2004). Pressure activates colon cancer cell adhesion by inside-out focal adhesion complex and actin cytoskeletal signaling. Gastroenterology 126: 8–18.

  25. Totsukawa G, Wu Y, Sasaki Y, Hartshorne DJ, Yamakita Y, Yamashiro S et al. (2004). Distinct roles of MLCK and ROCK in the regulation of membrane protrusions and focal adhesion dynamics during cell migration of fibroblasts. J Cell Biol 164: 427–439.

  26. von Wichert G, Jiang G, Kostic A, DeVos K, Sap J, Sheetz MP . (2003). RPTP-alpha acts as a transducer of mechanical force on alphav/beta3-integrin-cytoskeleton linkages. J Cell Biol 161: 143–153.

  27. Wang Y, Botvinick EL, Zhao Y, Berns MW, Usami S, Tsien RY et al. (2005). Visualizing the mechanical activation of Src. Nature 434: 1040–1045.

  28. Wang YL, Pelham Jr RJ . (1998). Preparation of a flexible, porous polyacrylamide substrate for mechanical studies of cultured cells. Methods Enzymol 298: 489–496.

  29. Wells RG . (2008). The role of matrix stiffness in regulating cell behavior. Hepatology 47: 1394–1400.

  30. Wozniak MA, Desai R, Solski PA, Der CJ, Keely PJ . (2003). ROCK-generated contractility regulates breast epithelial cell differentiation in response to the physical properties of a three-dimensional collagen matrix. J Cell Biol 163: 583–595.

  31. Wu CW, Kao HL, Li AF, Chi CW, Lin WC . (2006). Protein tyrosine-phosphatase expression profiling in gastric cancer tissues. Cancer Lett 242: 95–103.

  32. Zeng L, D'Alessandri L, Kalousek MB, Vaughan L, Pallen CJ . (1999). Protein tyrosine phosphatase alpha (PTPalpha) and contactin form a novel neuronal receptor complex linked to the intracellular tyrosine kinase fyn. J Cell Biol 147: 707–714.

  33. Zeng L, Si X, Yu WP, Le HT, Ng KP, Teng RM et al. (2003). PTP alpha regulates integrin-stimulated FAK autophosphorylation and cytoskeletal rearrangement in cell spreading and migration. J Cell Biol 160: 137–146.

  34. Zheng XM, Wang Y, Pallen CJ . (1992). Cell transformation and activation of pp60c-src by overexpression of a protein tyrosine phosphatase. Nature 359: 336–339.

Download references

Acknowledgements

This work was supported by the DFG to GvW and FO and the Deutsche Krebshilfe to GvW and TS.

Author information

Correspondence to G von Wichert.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Krndija, D., Schmid, H., Eismann, J. et al. Substrate stiffness and the receptor-type tyrosine-protein phosphatase alpha regulate spreading of colon cancer cells through cytoskeletal contractility. Oncogene 29, 2724–2738 (2010). https://doi.org/10.1038/onc.2010.25

Download citation

Keywords

  • spreading
  • force
  • colon cancer
  • tyrosine-phosphatase
  • contractility
  • invasion

Further reading