Oncogenic Function of a Novel WD-Domain Protein, STRAP, in Human Carcinogenesis

The development and progression of malignancies is a complex multistage process that involves the contribution of a number of genes giving growth advantage to cells when transformed. The role of transforming growth factor-β (TGF-β) in carcinogenesis is complex with tumor-suppressor or prooncogenic activities depending on the cell type and the stage of the disease. We have previously reported the identification of a novel WD-domain protein, STRAP, that associates with both TGF-β receptors and that synergizes with the inhibitory Smad, Smad7, in the negative regulation of TGF-β–induced transcription. Here, we show that STRAP is ubiquitously expressed and is localized in both cytoplasm and nucleus. STRAP is up-regulated in 60% colon and in 78% lung carcinomas. Stable expression of STRAP results in activation of mitogen-activated protein kinase/extracellular signal-regulated kinase pathway and in down-regulation of the cyclin-dependent kinase inhibitor p21^Cip1, which results in retinoblastoma protein hyperphosphorylation. In addition, we have observed that Smad2/3 phosphorylation, TGF-β–mediated transcription, and growth inhibition are induced in STRAP-knockout mouse embryonic fibroblasts compared with wild-type cells. Ectopic expression of STRAP in A549 lung adenocarcinoma cell line inhibits TGF-β–induced growth inhibition and enhances anchorage-independent growth of these cells. Moreover, overexpression of STRAP increases tumorigenicity in athymic nude mice. Knockdown of endogenous STRAP by small interfering RNA increases TGF-β signaling, reduces ERK activity, increases p21^Cip1 expression, and decreases tumorigenicity. Taken together, these results suggest that up-regulation of STRAP in human cancers may provide growth advantage to tumor cells via TGF-β–dependent and TGF-β–independent mechanisms, thus demonstrating the oncogenic function of STRAP. (Cancer Res 2006; 66(12): 6156-66)

Human carcinomas arise through the accumulation of a number of genetic and epigenetic changes, functional inactivation of tumor-suppressor genes, and activation or up-regulation of cellular oncogenes. There is compelling evidence indicating that transforming growth factor-β (TGF-β) has complex roles in tumor suppression and progression that are context and stage dependent. TGF-β plays a tumor-suppressive role by its ability to maintain tissue architecture, inhibit growth, induce apoptosis, and inhibit genomic instability in nontransformed cells or tissues. Several lines of evidence suggest that carcinoma cells frequently lose the antiproliferative response to TGF-β. The role of TGF-β signaling in tumor suppression has been suggested by the presence of inactivating mutations in genes encoding TGF-β receptors and Smads. For example, in colorectal cancers, resistance to TGF-βs by functional inactivation of signaling molecules, including TGF-β type II receptor (TβRII), Smad4, and Smad2, occurs usually in the transition from late adenoma/adenocarcinoma to metastatic carcinoma (1, 2).

TGF-βs initiate signaling by binding to TβRII that results in recruitment and phosphorylation of the type I receptor (TβRI). After being activated, TβRI propagates the signal to a family of intracellular signal mediators known as Smads. Smad proteins are classified according to their structure and function in signaling by TGF-β family members. Receptor-regulated Smads, Smad2 and Smad3, are phosphorylated and activated by TβRI. Then, they form complexes with central Smad, Smad4, and translocate to the nucleus for regulating the expression of target genes. Given the involvement of TGF-β in regulation of cellular homeostasis, it is expected that there are also a number of feedback mechanisms regulating this process. A distinct class of distantly related Smads, including Smad6 and Smad7, functions as inhibitors of these signaling pathways. Smad7 expression is induced in response to TGF-β, suggesting roles in the negative feedback of these pathways.

In addition to Smads, other proteins that interact with TGF-β receptors have been identified and some of them are involved in TGF-β signaling (3). Three proteins containing WD repeats associate with and are phosphorylated by the receptors and regulate TGF-β receptor signaling. Whereas TRIP-1 interacts with TβRII (4), the β-subunit of the protein phosphatase 2A interacts with type I receptors (5). We have previously reported the identification of a novel WD40 domain–containing protein, STRAP, which interacts with both TβRI and TβRII and negatively regulates TGF-β–induced gene expression (6). WD40 domain–containing proteins, in general, seem to serve regulatory functions in various cellular processes, such as signal transduction, transcriptional regulation, RNA processing, vesicular trafficking, and cell cycle progression (7, 8). STRAP associates with Smad7, recruits it from the cytosol to the activated TβRI, stabilizes the heteromeric complex, and thus assists Smad7 in preventing Smad2 and Smad3 activation by the receptor complex (9). STRAP has been shown to be strong predictive marker of 5-fluorouracil-based adjuvant chemotherapy benefit in colorectal cancer (10). We and others have shown that Smad7 induces tumorigenicity by blocking TGF-β–induced growth inhibition and apoptosis (11). The survival rate of patients with negative Smad7 expression in human gastric carcinomas is significantly higher than that of patients with positive Smad7 expression (12). As mutations or functional inactivations of receptors and Smads are not enough to explain the broad spectrum of TGF-β unresponsiveness in human cancers, enhanced expression of the TGF-β signaling inhibitors, including STRAP, may present a novel mechanism by which a portion of human tumors become refractory to tumor-suppressor functions of TGF-β. However, nothing is known about the mechanism of function of STRAP in human cancers and the potential role of enhanced expression of STRAP in cellular proliferation and tumorigenicity. We now report that STRAP is up-regulated mostly in transformed epithelium in human colorectal and lung carcinomas. STRAP activates mitogen-activated protein (MAP) kinase (MAPK)/extracellular signal-regulated kinase (ERK) pathway and down-regulates Cdk inhibitor p21^Cip1 that leads to retinoblastoma protein (pRb) phosphorylation. TGF-β/Smad–mediated transcriptional activation and growth inhibition are induced in STRAP^−/− mouse embryonic fibroblasts (MEF) compared with STRAP^+/+ MEFs. Ectopic expression of STRAP in lung adenocarcinoma A549 and colon adenonocarcinoma FET cells inhibits TGF-β–induced growth inhibition and enhances anchorage-independent growth and tumorigenicity. Knock down of STRAP in A549 cells increases TGF-β signaling and decreases tumorigenicity. All together, our results suggest that up-regulation of STRAP in human cancers may induce cell proliferation in a TGF-β–dependent and TGF-β–independent manner; thus, STRAP may be involved in tumor progression.

Cell lines. FET cells, mink lung epithelial cells (Mv1Lu), mouse fibroblast cells (NIH-3T3), human embryonic kidney cells (293T/R), human hepatoma cells (HepG2), mouse keratinocytes (MK), mouse fibroblast cells (AKR2B), monkey kidney cells (COS-1 and COS-7), mouse mammary epithelial cells (NMuMG), rat intestinal epithelial cells (RIE), STRAP^+/+, STRAP^−/− MEFs, and human lung adenocarcinoma cells (A549) were maintained in 10% serum-containing medium supplemented with penicillin and streptomycin.

Reagents and antibodies. TGF-β1 was purchased from R&D Systems (Minneapolis, MN). The anti-Flag and anti-β-actin antibodies were purchased from Sigma Biochemicals (St. Louis, MO). Rabbit anti-Smad2 and anti-Smad3 were from Zymed Laboratories, Inc. (San Francisco, CA). Anti-Smad4, anti-p21^Cip1, anti-p27^Kip1, anti–Rb, anti-phospho-ERK, anti-ERK, anti-RhoA, and anti-poly(ADP)ribose polymerase (PARP) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phospho-Smad2 and anti-phospho-Rb antibodies were from Cell Signaling Technology (Beverly, MA).

Retroviral construct and stable STRAP clones. Mouse STRAP cDNA was cloned into BamH1 and EcoR1 sites of retroviral vector pBabe-Puro3 and the resulting construct, pBabe-Puro3-STRAP-Flag, was used for generation of stable clones in FET cells as discussed previously (11). Stable STRAP clones were maintained in the presence of 1 μg/mL puromycin and the expression of STRAP-Flag was verified by Western blotting with anti-Flag antibody.

Stable STRAP and Smad7 clones. A549 cells stably transfected with Flag-tagged STRAP-pCDNA3, pCMV-Flag-Smad7, or pCDNA3-Flag vector (Invitrogen, Carlsbad, CA) were selected with G418 (600 μg/mL) to isolate stable clones and maintained in the presence of 400 μg/mL G418. The expression of STRAP or Smad7 was verified by Western blotting with anti-Flag antibody. Stable STRAP clones in Mv1Lu cells were generated as above using 1.2 mg/mL G418.

Transcriptional response assay. Wild-type and STRAP^−/− MEFs or A549 stable clones were transiently transfected with either p3TP-Lux or (CAGA)₉ MLP-Luc with CMV-β-gal plasmids. For small interfering RNA (siRNA) transfection, A549 cells were transiently cotransfected with (CAGA)₉ MLP-Luc, siRNA, and CMV-β-gal plasmids. Transfected cells were treated with TGF-β and cell lysates were analyzed for luciferase and β-gal activities. Normalized luciferase activity was determined.

Western blot analyses. Cell lysates were prepared from MK, 293R, AKR2B, COS-1, HepG2, Mv1Lu cells, colon cancer cell lines, and human colon and lung cancer tissues. Cell lysates were also prepared from epidermal growth factor (EGF)–treated vector control clones from A549 and FET cells and from wild-type and STRAP^−/− MEFs after TGF-β treatment. For subcellular localizations of endogenous STRAP, nuclear and cytoplasmic protein extracts were prepared from 293T, FET, A549, RIE, Mv1Lu, and NMuMG cells. Lysates were analyzed by Western blotting as indicated in figure legends.

Immunohistochemical analyses. Human tissue specimens were fixed in formalin and embedded in paraffin blocks, sections were rehydrated, and target retrieval (DAKO) was done. Target was permeabilized with saponin and then incubated with anti-STRAP antibody. The DAKO Envision+ System, DAB+/Peroxidase (DAKO, Carpinteria, CA), was used to produce localized, visible staining followed by a light hematoxylin counterstain. Specificity of staining by anti-STRAP antibody was determined by preincubating the antibody with recombinant STRAP protein.

Immunofluorescence detections. Mv1Lu and NIH-3T3 cells were grown in chamber slides, fixed, permeabilized, and then used to perform immunofluorescent staining with anti-STRAP antibody using the protocol described previously (13). Stained cells were visualized by fluorescence microscope. STRAP^+/+ and STRAP^−/− MEFs were treated with TGF-β for 90 minutes and processed for immunofluorescence analyses as above using anti-Smad2 and anti-Smad4 antibodies.

In vitro kinase assay. To determine the MAPK activity, cell lysates from three stable STRAP clones and one vector control clone in Mv1Lu cells were immunoprecipitated with anti-ERK antibody. The immunocomplex was captured by protein G-Sepharose beads (Sigma Biochemicals) and then used for kinase assay using myelin basic protein (MBP) as substrate, as described previously (14). Phosphorylated MBP was identified by autoradiography.

Cell counting assay. STRAP^+/+ and STRAP^−/− MEFs were treated with 3 ng/mL TGF-β, counted, and the results were plotted. Parental A549 cells, vector control, three stable STRAP clone, and one stable Smad7 clone were treated with TGF-β and counted as above.

siRNA experiments. STRAP-specific siRNA oligonucleotides corresponding to the coding regions (515, amino acids 124-129) of human STRAP (15) were synthesized and annealed by Ambion, Inc. (Austin, TX). Lysates from siRNA-transfected A549 cells were analyzed for Western blotting with antibodies as indicated. For siRNA stable clones, double-stranded DNA oligonucleotides corresponding to the coding region (719-737) were inserted into siRNA expression vector (pSilencer 4.1-CMV neo) according to the protocol of the manufacturer (Ambion). The resultant plasmid was stably transfected into A549 cells and selected with 600 μg/mL of G418. Three stable siRNA clones (clones 7, 45, and 50) were generated that express lower levels of endogenous STRAP.

Generation of STRAP-adenovirus and infection of STRAP^−/− MEFs. To generate recombinant adenovirus vector, mouse STRAP cDNA was cloned into the shuttle vector (pAdTrack-CMV) according to pAdEasy system as described previously (16) and was transfected into adenovirus package 293T cells. Recombinant adenovirus recovered from cultured medium was used to infect STRAP^−/− MEFs, and the cell lysates were analyzed by Western blotting.

Soft agarose assay and xenograft studies. A549 stable clones (either STRAP overexpressed or knocked down) or FET stable clones were plated for soft agarose assay as discussed previously (11). For xenograft studies, 1 × 10⁶ cells from STRAP stable clones, one vector clone, and parental FET cells were injected s.c. in athymic nude mice. The animals were monitored for tumor formation every week for a total of 7 weeks and the tumors were measured as determined previously (11).

Intracellular localization of STRAP. To evaluate the function of STRAP protein and to detect the endogenous STRAP expression, we have raised rabbit antisera against recombinant full-length mouse STRAP protein. Lysates prepared from different cell lines were analyzed by Western blotting using the above anti-STRAP antibody (Fig. 1A). As a positive control, the lysate from COS-1 cells overexpressing STRAP tagged with Flag epitope was used. Exogenous STRAP was little bigger in size than the endogenous STRAP due to the epitope tag. We observed that endogenous STRAP expression varies depending on cell types. We tested the specificity of anti-STRAP antibody by first incubating the antibody with recombinant STRAP protein and then used the antigen-antibody complex for Western blot analyses. Anti-STRAP antibody recognizes endogenous STRAP, whereas incubating the antibody with recombinant STRAP protein abolished the STRAP signal as shown in Fig. 1B. The STRAP-specific band, but not the nonspecific band, was blocked by recombinant STRAP protein. Therefore, the anti-STRAP antibody is specific and STRAP is expressed in most cell types.

In an attempt to verify the subcellular localization of endogenous STRAP and to test whether there is any differential localization of STRAP in normal versus tumor cell lines, we separated cytoplasmic and nuclear fractions from either normal or tumor cells, and analyzed them by Western blotting using anti-STRAP antibody. We have observed comparable level of STRAP in both compartments in each cell line (Fig. 1C), except for RIE cells where nuclear expression of STRAP is greater. The purity of cytoplasmic and nuclear fractions was verified by Western blotting with antibodies against RhoA and PARP, respectively (Fig. 1C). These results suggest that STRAP is similarly present in the cytoplasm as well as in the nucleus. To further confirm the subcellular localization of endogenous STRAP, we did immunofluorescence analyses. Mv1Lu or NIH-3T3 cells were subjected to immunofluorescence analyses using anti-STRAP antibody that indicated the presence of STRAP in both cytoplasm and nucleus in either cell lines (Fig. 1D). We did not observe any staining with rabbit IgG (data not shown). These results suggest the presence of STRAP in both cytoplasm and nucleus.

Up-regulation of STRAP in human colorectal cancers. Enhanced expression of TGF-β signaling inhibitors in human cancers may play a role in tumor progression. As STRAP is widely expressed in different cell types and is inhibitory to TGF-β signaling, we first investigated whether the level of STRAP expression is regulated in different colon cancer cell lines that have different gene mutation status. We observed good level of expression of STRAP in all colon cancer cell lines (Fig. 2A). To investigate whether STRAP is up-regulated in human colon cancer, we examined the level of STRAP in 20 colon tumors and patient-matched adjacent normal tissues by immunoblot analyses. In normal samples, low to moderate level of expression of STRAP was observed. In contrast, relatively high levels of STRAP were present in 12 of 20 (60%) colon cancer samples (Fig. 2B). STRAP expression in tumor samples were normalized and shown in Fig. 2B. To investigate in what type of cells STRAP level is increased in colon tumors, we did immunohistochemical analyses of two tumor specimens that had been examined for the up-regulation of STRAP by immunoblot analyses (underlined). We observed the similar expression profile of STRAP in both tissues (Fig. 2C). Strong expression of STRAP was observed in transformed epithelial cells (middle and right, black arrowhead), whereas the tumor stroma showed weak expression (Fig. 2C,, middle, white arrowhead). Transformed epithelial cells within the tumor mass showed more expression of STRAP when compared with adjacent normal epithelium (Fig. 2C,, left, black arrow). Specificity of staining by anti-STRAP antibody was tested by blocking the staining of serial sections with recombinant STRAP protein (Fig. 2D). These results suggest that STRAP is up-regulated in 60% colorectal tumors, and the expression of STRAP is increased in the transformed epithelium but not in normal epithelium or in stroma.

Up-regulation of STRAP in human lung cancers. To investigate whether STRAP is induced in human lung cancers, we have examined the expression of STRAP in lung tumors and patient-matched adjacent normal lung tissues by immunoblot analyses. We have observed strong up-regulation of STRAP in 11 of 14 tumors, including squamous, adenocarcinoma, and large cell carcinoma (Fig. 3A). STRAP expression in tumor samples were normalized and shown in Fig. 3A. We investigated the expression pattern of STRAP in lung tumors, including carcinoid tumor, large cell carcinoma, squamous cell carcinoma, and adenocarcinoma (Fig. 3B). Strong expression of STRAP was observed in transformed epithelial cells (black arrowhead), whereas the tumor stroma showed weak expression (black arrow). We observed weak to moderate expression of STRAP in some pneumocytes, macrophages, and lymphocytes, which is consistent with the fact that STRAP is expressed in many cell types. These data suggest that STRAP is up-regulated in 78% of lung tumors, and the expression of STRAP is mostly increased in the transformed epithelium but not in the stroma.

STRAP activates MAP/ERK kinase/ERK pathway. MAPK/ERK has been revealed to be involved in the physiologic proliferation of mammalian cells and also to potentiate transformation. To determine the potential role of high levels of STRAP in human cancers, we first investigated whether STRAP has any effect on growth stimulatory kinase MAP/ERK kinase (MEK)/ERK pathway. We generated stable cell lines expressing STRAP-Flag or STRAP-HA using mink lung epithelial cells (Mv1Lu). This cell line was chosen because it is responsive to TGF-β and expresses low level of STRAP. Expression of STRAP was tested using anti-STRAP antibody in three stable clones, two with Flag tag (F2 and F13) and one with HA tag. We observed different levels of exogenous STRAP expression when compared with endogenous STRAP level (Fig. 4A,, top). Equal amounts of whole ERK1/2 were immunoprecipitated from total cell lysates of stable clones and the immunoprecipitates were subjected to in vitro kinase assays using MBP as substrate. We found that MAPK/ERK was activated in all three STRAP-overexpressing clones depending on the level of STRAP expression (Fig. 4A , bottom). Activation of ERK1/2 was also detected in these stable cell lines by immunoblotting with anti–phospho-ERK1/2 antibody (data not shown). However, TGF-β has no effect on STRAP-induced activation of MAPK/ERK pathway (data not shown). Thus, these data suggest that STRAP can induce MEK/ERK activity in a TGF-β–independent manner.

Down-regulation of p21^Cip1 by STRAP results in hyperphosphorylation of pRb. TGF-β exerts its growth inhibitory effects in part by inducing the expression of cell cycle suppressor proteins including CDK inhibitor p21^Cip1 (17). To determine whether STRAP has any effect on the expression of cell cycle regulatory molecules involved in G₁ arrest, equal amount of each lysate was subjected to immunoblot analyses with anti p21^Cip1 and anti p27^Kip1 antibodies. STRAP expression (F13<H4<F2) directly correlated with the decreased expression of p21^Cip1 in the order F13>H4>F2, whereas p27^Kip1 levels were unaffected (Fig. 4B). We did not observe any effect of TGF-β on STRAP-mediated down-regulation of p21^Cip1. p21^Cip1 promoter activity was reduced in STRAP clones when compared with the vector clone (data not shown). We then verified whether down-regulation of p21^Cip1 could lead to hyperphosphorylation of pRb. Cell lysates from vector and STRAP stable clones were used for Western blotting with anti-Rb and anti-phospho-Rb antibodies. We found that the down-regulation of p21^Cip1 by STRAP correlates with increased phosphorylation of pRb in the order F13<H4<F2 (Fig. 4C). These data show that STRAP-mediated down-regulation of the cyclin-dependent kinase inhibitor p21^Cip1 leads to hyperphosphorylation of pRb.

STRAP inhibits TGF-β–induced nuclear translocation of Smad2/Smad4, activation of TGF-β responsive reporter genes, and growth inhibition. As STRAP is expressed ubiquitously in most cell types, we have used STRAP-deficient MEFs to study the mechanism of STRAP function in regulating TGF-β signaling. We first confirm the expression level of STRAP in STRAP^+/+ and STRAP^−/− MEFs (18) by Western blot analysis. We found that STRAP^+/+ MEFs express a good amount of STRAP, whereas STRAP expression was not detected in STRAP^−/− MEFs as expected (Fig. 5A,, left, top). To verify whether STRAP can block endogenous or TGF-β–induced phosphorylation of Smad2, we did Western blotting using cell lysates from STRAP^+/+ and STRAP^−/− MEFs after treatment with TGF-β for different time points. Interestingly, we observed that the basal level of Smad2 phosphorylation is significantly low in STRAP^+/+ MEFs compared with STRAP^−/− MEFs (Fig. 5A,, left bottom, lanes 1 and 6). TGF-β enhances phosphorylation of Smad2 in both STRAP^−/− (lanes 2-5) and STRAP^+/+ (lanes 7-10) MEFs. However, TGF-β–induced Smad2 phosphorylation is more pronounced in STRAP^−/− MEFs when compared with wild-type MEFs, suggesting that endogenous STRAP is inhibitory to TGF-β/Smad signaling. To verify whether STRAP positively regulates MAPK/ERK pathway in MEFs, we overexpressed STRAP or GFP in STRAP^−/− MEFs by infecting adenoviruses that express either STRAP or GFP, respectively. We observed a dose-dependent expression of STRAP in adenovirus-infected STRAP^−/− MEFs, and STRAP^+/+ MEFs showed a modest level of STRAP expression, as expected (Fig. 5A,, right). Interestingly, we observed a steady increase in the phospho-ERK level that directly correlated with the amount of STRAP expression in infected STRAP^−/− MEFs. These data show a role of STRAP in the activation of MAPK/ERK pathway in MEFs. Then, we investigated the subcellular localization of Smad2 and Smad4 by immunofluorescence analyses. We found that both Smad2 and Smad4 were predominantly located in the cytoplasm in both cell lines in the absence of TGF-β. TGF-β significantly increases the nuclear translocation of both Smad2 and Smad4 in STRAP^−/− MEFs, whereas reduced amount of Smad2 and Smad4 were translocated into the nucleus in STRAP^+/+ MEFs under similar conditions (Fig. 5B). These results suggest that endogenous STRAP inhibits TGF-β–mediated nuclear translocation of both Smad2 and Smad4. To further verify whether STRAP expression in STRAP^+/+ MEFs can affect downstream transcriptional responses mediated by TGF-β, we did transient transfection analyses using TGF-β–responsive p3TP-Lux or (CAGA)₉ MLP-Luc (Fig. 5C) reporter plasmids. We observed that TGF-β strongly induces both reporter activities in STRAP^−/− MEFs in a dose-dependent manner, whereas TGF-β–induced reporter activities were strongly suppressed in STRAP^+/+ MEFs. These data suggest that endogenous STRAP in wild-type MEFs blocks TGF-β–mediated transcriptional responses. To examine the effect of STRAP on TGF-β–mediated cell growth inhibition, we did cell counting assay using STRAP^+/+ and STRAP^−/− MEFs. We observed that TGF-β significantly inhibits growth of STRAP^−/− MEFs (Fig. 5D). STRAP-expressing wild-type MEFs is not responsive to TGF-β–mediated growth inhibitory effects; rather, these cells grew continuously in the presence of TGF-β (Fig. 5D). These results suggest that STRAP can function to block TGF-β signaling downstream of its receptor complex.

Inhibition of TGF-β signaling in tumor cells by STRAP results in induced anchorage-independent growth. The up-regulation of STRAP in human tumors and blockade of TGF-β signaling by endogenous STRAP in vivo raises the possibility of involvement of STRAP in carcinogenesis. In an attempt to determine the role of STRAP up-regulation in human cancers, we stably expressed STRAP in human lung adenocarcinoma (A549) cells that have intact TGF-β signaling. Three stable STRAP clones (clones 9, 17, and 25) and one stable Smad7 clone were selected for experiments and for positive control, respectively (Fig. 6A,, left). To test whether increased expression of STRAP results in the activation of MAPK/ERK signaling, we did Western blot analyses. We observed that STRAP induces the phosphorylation of ERK in stable clones when compared with vector control clone (Fig. 6A,, right, compare lanes 1 and 3-5). As a positive control, EGF treatment induces higher levels of phosphorylation of ERK in vector clone (lanes 1 and 2). Overexpression of STRAP did not further stimulate phosphorylation of ERK in stable STRAP clones when treated with EGF (data not shown). However, ectopic expression of STRAP in these cells did not reduce the level of p21^Cip1 (data not shown). To test the role of STRAP in TGF-β signaling, clones were transiently transfected with either (CAGA)₉ MLP-Luc (Fig. 6B) or p3TP-Lux (data not shown) reporter vector and then treated with 2 or 5 ng/mL TGF-β. TGF-β strongly induces reporter activities in parental and vector control cells, whereas STRAP or Smad7-overexpressing clones inhibits TGF-β–induced reporter activation (Fig. 6B), suggesting that STRAP inhibits TGF-β–mediated transcriptional responses in this tumor cell line. To further determine the effect of stable expression of STRAP on TGF-β–mediated inhibition of cell growth, we used cell counting assay after treating the clones with TGF-β for 6 days. Stable expression of STRAP or Smad7 significantly blocked TGF-β–mediated growth inhibition when compared with that of parental or vector control cells (Fig. 6C). These data suggest that STRAP induces proliferation of A549 cells by blocking TGF-β–mediated inhibition of cell growth. To further determine whether STRAP can alter the tumorigenic properties of cells, we did an assay for anchorage-independent growth. We observed that STRAP or Smad7 stable clones produced larger colonies when compared with parental or vector control clone (Fig. 6D , left). We also noticed that all three stable STRAP clones produced 3- to 5-fold higher numbers of colonies in soft agarose when compared with parental or vector control clone (right). In addition, stable Smad7 clone produced colonies that are comparable in number and size with that produced by STRAP stable clones. Together, these data suggest that inhibition of TGF-β signaling and activation of MAPK/ERK pathway by STRAP in A549 cells may be involved in inducing tumorigenicity.

Knock down of STRAP enhances TGF-β signaling and reduces anchorage-independent growth. To confirm the physiologic role of STRAP in the regulation of MAPK/ERK activity and of p21^Cip1 expression, we did Western blot analyses using lysates from A549 cells transfected with human STRAP-specific siRNA. The level of endogenous STRAP expression was significantly reduced that resulted in a decrease in phospho-ERK levels and increase in p21^Cip1 (Fig. 7A). To test whether endogenous STRAP can block TGF-β signaling, we did TGF-β–responsive reporter assay by transient transfection of siRNA into A549 cells. We compared the reporter activation in the presence or absence of siRNA in response to TGF-β. We observed a dose-dependent induction of promoter activity when cells were transfected by increasing amount of siRNA (Fig. 7B), suggesting a strong inhibitory role of STRAP in blocking TGF-β signaling. These results are in agreement with the strong induction of TGF-β–mediated transcription in STRAP^−/− MEFs. To test the effects of endogenous STRAP on anchorage-independent growth, we generated three STRAP-specific siRNA stable clones (clones 7, 45, and 50) in A549 cells (Fig. 7C). We observed reduced levels of STRAP expression in stable clones, which correlated with reduced levels of phosphorylation of ERK in stable clones compared with parental and vector control clones (data not shown). We then verified the effect of endogenous STRAP on tumorigenicity by in vitro anchorage-independent growth assay. Interestingly, we observed significantly reduced number and size of colonies formed in soft agarose by siRNA stable clones compared with parental and vector control clone (Fig. 7D). These data further show the role of STRAP on growth and tumorigenicity.

Stable expression of STRAP in FET cells enhances tumorigenicity. To determine the potential role of high levels of STRAP expression and subsequent inhibition of tumor-suppressor function of TGF-β in cancer cells in vivo, we used FET cells, a colon adenocarcinoma-derived cell line, to stably express STRAP by using retroviral system. Exogenous STRAP expression was tested by Western blot analysis using anti-Flag M2 monoclonal antibody (Fig. 8A). We next tested whether increased expression of STRAP in colon tumor cells can lead to the activation of MAPK/ERK pathway. We observed that expression of STRAP in stable clones induced phosphorylation of ERK when compared with vector control clone (Fig. 8B, lanes 1 and 3-5). As a positive control, EGF treatment of the vector control cells induced phospho-ERK level strongly (Fig. 8B, lanes 1 and 2). However, overexpression of STRAP did not further stimulate phosphorylation of ERK in stable STRAP clones when treated with EGF (data not shown). In an attempt to verify whether STRAP can induce anchorage-independent growth of FET cells, we did in vitro soft agarose assay. We observed that all STRAP stable clones produced larger colonies in soft agarose compared with parental FET cells and vector control clone (Fig. 8C,, left). In addition, STRAP stable clones produced over 4-fold higher numbers of colonies when compared with FET cells and vector clone (right), suggesting that overexpression of STRAP enhances anchorage-independent growth of FET cells. We did in vivo tumor xenograft studies using stable FET clones. FET cells are sensitive to the growth inhibitory responses to TGF-β, do not form colonies in soft agarose, and do not readily form tumors in athymic nude mice (19, 20). As we reported previously (11), FET cells do not give rise to any tumor in 7 weeks, and the vector control clone formed only a small nodule at the site of inoculation that regressed after 5 weeks (Fig. 8D). In contrast, inoculation of stable STRAP clones resulted in palpable tumor formation after 2 weeks and they grew continuously over the period of observation (Fig. 8D). The tumors from stable STRAP cells are >11-fold after 7 weeks compared with parental and vector control cells. Therefore, these results suggest that stable expression of STRAP in FET cells induces tumorigenicity.

Cell growth is controlled by negative and positive regulatory signals, and the disruption of these signals may cause disease state. Loss of negative growth constraints may contribute to oncogenic processes, which may occur as a consequence of the loss of TGF-β–mediated tumor-suppressor functions. It has been well documented that most human cancers, including colorectal cancer, are resistant to TGF-β–mediated growth inhibition. Resistance to TGF-β growth inhibition in colorectal cancer can occur through a variety of mechanisms, including mutations or functional inactivation of TβRII, a decreased expression of either TβR1 or TβRII, and the inactivating mutations in components of the TGF-β signaling pathway, such as Smad2 or Smad4 (21–26). However, inactivating mutations of the TGF-β receptors and selected TGF-β signal transducers (Smad2/4) are not sufficient to explain the insensitivity to TGF-β–induced antitumor activity in colon cancer. Therefore, we hypothesized that abrogation of TGF-β–induced growth inhibition by the TGF-β signaling inhibitor STRAP may provide a mechanism by which human tumors become nonresponsive to TGF-β. Here, we show that the expression of STRAP is up-regulated in both human colon and lung cancers. Ectopic expression of STRAP enhances tumorigenicity of lung and colon tumor cells. STRAP inhibits TGF-β–mediated transcription and growth inhibition. In addition, STRAP induces cell growth by inhibiting the expression of CDK inhibitor p21^Cip1 that lead to phosphorylation of Rb protein, and by activating the growth proliferating MAPK activity by a TGF-β–independent mechanism. These biological functions of STRAP are further confirmed by using STRAP-null cells and by knocking down endogenous STRAP in tumor cells.

As STRAP is inhibitory to TGF-β signaling and it has growth-promoting functions, we investigated whether it is up-regulated in human cancers. Our results show that STRAP is up-regulated in 60% colorectal tumors (Fig. 2B) and 78% lung tumors, including squamous cell carcinoma, adenocarcinoma, and large cell carcinoma (Fig. 3A). STRAP is up-regulated mostly in transformed epithelial cells but not in normal epithelium or stroma of the same tumor (Figs. 2C and 3B). As we have observed that the up-regulation of STRAP takes place mostly in transformed epithelium and that STRAP is expressed in most cell types, the actual fold increase in STRAP will be attenuated when whole tumor is used for Western blot analyses. Our data is consistent with a recent report suggesting the up-regulation of human STRAP in breast cancer (27).⁵

Another study has shown that STRAP is a strong predictive marker of adjuvant chemotherapy benefit in colorectal cancer (10). Patients bearing tumors with higher amplification of STRAP had significantly better prognosis than those patients receiving adjuvant chemotherapy. However, it is not known whether amplification of STRAP is associated with overexpression of the protein product. Coupled with our previous studies (9), synergistic inhibition of TGF-β signaling by up-regulated STRAP and/or Smad7 in tumors provides a mechanism for the resistance to TGF-β antitumor effects.

In an attempt to determine the role of STRAP up-regulation in human cancers, we have observed that STRAP can activate MEK/ERK pathway that may be associated with the stimulation of cell growth. STRAP inhibits the expression of the cell cycle regulatory p21^Cip1 protein (Figs. 4B and 7A) through transcriptional repression of the promoter activity (data not shown). Down-regulation of p21^Cip1 leads to hyperphosphorylation of pRb that might contribute to cell proliferation (Fig. 4C). As STRAP is expressed ubiquitously in most cell types, it is possible that endogenous STRAP signaling is important for activation of these growth-stimulatory pathways. This is in agreement with our STRAP knockdown experiments (Fig. 7A). Interestingly, these effects of STRAP are independent of TGF-β and may be involved in tumor progression.

One of the most interesting parts of this study is to establish the inhibitory effects of endogenous STRAP on TGF-β signaling. Our previous studies involving overexpression of STRAP (6, 9) have suggested a modest inhibitory role on TGF-β transcriptional effects, as most of the cell types have good expression level of STRAP. In this study, we have used STRAP-deficient MEFs to study the mechanism of STRAP function. STRAP^−/− MEFs show a strong inhibition of TGF-β–mediated transcription and these cells are strongly growth inhibited by TGF-β when compared with wild-type MEFs. These observations are supported by strong induction of TGF-β signaling in A549 cells when endogenous STRAP is knocked down by siRNA. Interestingly, loss of inhibitory effects of STRAP on TGF-β tumor suppression function and of growth-stimulatory effects of STRAP by knocking it down results in the reduction of anchorage-independent growth of A549 cells (Fig. 7D).

Others and we have shown that Smad7 enhances tumorigenicity in pancreatic and colon cancer, respectively (11, 28). TGF-β signaling inhibitors, including Smad7, are up-regulated in several cancers. A recent study has shown that Smad7 expression is associated with poor outcome in gastric carcinomas (12). Our study suggests that STRAP is up-regulated in human cancers. It will be interesting to examine whether STRAP expression has any prognostic significance in tumor progression. Multiple lines of evidence suggest that human cancers are in general functionally resistant to TGF-β–induced tumor-suppressor function. Our studies provide a novel intracellular mechanism describing how TGF-β tumor-suppressor function is abrogated by STRAP. Regulation of cell proliferation by STRAP by both TGF-β–dependent and TGF-β–independent mechanisms and elevation of its expression in human cancers suggest its involvement in tumor progression. Thus, STRAP could be a potentially important drug target for therapeutic intervention in human cancers.

Grant support: R01 CA95195 and CA113519, a Career Development Award from Specialized Programs of Research Excellence in lung cancer (5P50CA90949), and a Clinical Innovator Award from Flight Attendant Medical Research Institute (P.K. Datta); grants R01 DK52334 and CA69457 (R.D. Beauchamp); and Frances Williams Preston Laboratories of the T.J. Martell Foundation (H.L. Moses).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank P. Soriano (Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA) and M.G. Brattain (Department of Pharmacology and Therapeutics, Roswell Park Cancer Institute, Buffalo, NY) for providing us cell lines.