Background
TGF-β and its signalling effectors regulate many aspects of tumour cell biology, such as growth arrest, and cell motility the latter of which is important for the metastatic dissemination of tumour cells from their primary location to lymph or blood vessels [
1,
2]. TGF-β's cellular activities are mediated by specific receptor complexes that are assembled upon ligand binding and comprise the TGF-β type II receptor (TβRII) and TGF-β type I receptor (TβRI/ALK5). The activated ligand-receptor complex typically activates the Smad signalling pathway. The canonical Smad signalling cascade is initiated by C-terminal phosphorylation of receptor-regulated Smad transcription factors (R-Smads) Smad2 and/or Smad3 by activated ALK5 [
3]. This allows R-Smad binding to Smad4 and translocation of the complex to the nucleus where it can recruit transcriptional coactivators or corepressors to Smad binding elements (SBEs) in the promoters of TGF-β target genes [
1,
2]. The TGF-β signalling effectors are also key players of tumour cell behaviour and are often deregulated in cancer cells [
2,
4]. For instance, human pancreatic ductal adenocarcinoma (PDAC) is characterized besides the common K-Ras mutations (representing an early event in PDAC tumourigenesis) by both TGF-βoverexpression and mutational inactivation of the tumour suppressor Smad4/DPC4, the latter being a relatively late event. Recent studies in mice have shown that blockade of TGF-β/Smad signalling and activated Ras signalling cooperate to promote PDAC progression [
5,
6]. The crucial role of the Smad pathway in PDAC formation was also highlighted in orthotopic xenotransplantation experiments with TGF-β responsive PANC-1 cells, by which we demonstrated that Smad signalling through a kinase-active version of ALK5 suppressed primary tumour growth, but enhanced metastatic progression [
7]. A recent study in breast cancer cells has revealed that TGF-β signalling was activated transiently and locally and caused a switch from cohesive movement to single cell motility and promoted haematogenous metastasis [
8].
Smad2/3 and Smad4 are direct mediators of TGF-β signalling and there is now ample evidence to suggest that Smad2 and Smad3 have distinct and non-overlapping roles in TGF-β signalling and that these differ in epithelial cells and fibroblasts [reviewed in Ref. [
9]]. However, relatively few studies on the roles of Smad2 and Smad3 in TGF-β signalling have been performed in human epithelial cells from which most cancers arise. Moreover, it remained a mystery why TGF-β can induce different functions, such as growth arrest and epithelial-to-mesenchymal transition (EMT), in the same cell lines, even though both play opposing roles in tumourigenesis [
9]. The mechanisms for the selective activation of Smad2
versus Smad3 are largely unknown but can principally occur at the level of the TβRs, nuclear import and export, protein turnover, and/or at the transcriptional level. Alternatively, Smad2
versus Smad3 responses may be selected by post-translational modifications such as differential phosphorylation at the TβR complex [
9]. It is possible that the availability of other factors such as co-repressors and co-activators determine which response is mediated by Smad3 and Smad2. Since strategies for therapeutic targeting of the TGF-β signalling pathway are being pursued, revealing the identity of factors that modulate the relative activation of Smad2 or Smad3 in the TGF-β response may provide target(s) for more effective strategies for cancer therapy.
Rac1 belongs to the Rho family of small GTPases and has been implicated in the organization of the actin cytoskeleton, the formation of lamellopodia and focal adhesions, and in endocytic vesicle trafficking and receptor endocytosis. Rac1 can also drive cell proliferation and protect cells from apoptosis through its ability to activate extracellular signal-regulated kinases (ERKs), phosphatidylinositol-3 kinase (PI3-K), and the transcription factor NFκB [reviewed in Ref. [
10]]. Activated Rac1 acts synergistically with ligand-activated epidermal growth factor-receptor (EGF-R) to stimulate pancreatic tumour cell proliferation through cyclin D1 upregulation [
11]. Rac1 has a critical role in cell migration, and in the invasive, and metastatic behavior of cancer cells [
12‐
14]. Moreover, Rac1 function is required for oncogenic K-Ras tumourigenesis and proliferation [
15]. Activation of Rac1 is accompanied by its rapid translocation from the cytosol to the cell membrane, where it exerts part of its effects as an essential subunit of the reactive oxygen species (ROS)-producing enzyme NAD(P)H oxidase [
16]. In PDAC dysregulated expression of Rac1 was observed in the tumour cell compartment [
17], along with high activity of Vav1, a guanine exchange factor (GEF), which exhibits a particularly strong guanine exchange activity for Rac1 [
11]. Also TGF-β and Rac1 signalling exert antagonistic roles in tumour cell proliferation but share common nuclear targets such as cyclin D1 and p21
WAF1[
18,
19]. Initial evidence for a role of Rac1 in TGF-β signalling came from transcriptional reporter gene assays with dominant negative (dn) and constitutively active (ca) mutants [
20] and this was followed by the demonstration that Rac1 is involved in TGF-β-induced EMT [
12]. We have shown earlier that Rac1 is rapidly activated following stimulation of PDAC cells with TGF-β1 and that dn inhibition of Rac1 activity blunted both TGF-β1-induced p38 MAPK activation and expression of the small leucine-rich proteoglycan biglycan [
21].
As mentioned above, we demonstrated in orthotopic xenotransplantation experiments that Smad signalling through a kinase-active version of ALK5 suppressed primary tumour growth and enhanced metastatic progression [
7]. However, the design of this study did not permit to test why Smad signalling exerted opposite effects on both responses and whether each response may be mediated predominantly or exclusively by only one of the two R-Smads. In this study we therefore asked whether growth inhibition and cell migration (as
in vitro correlates of tumour growth and metastasis) are controlled differentially by Smad2 and Smad3 and whether Rac1 impacts on differential activation of both R-Smads by TGF-β1. For this purpose, we utilized the well characterized PDAC cell lines PANC-1 and COLO 357 which have retained a functional TGF-β/Smad pathway [
4,
22‐
24]. Using RNA interference to specifically deplete cells of the expression of the two R-Smads, we found that TGF-β1-induced growth inhibition was dependent on Smad3 (confirming earlier observations in PANC-1 cells [
25]) while the migratory response to TGF-β1 was positively controlled by Smad2. We went on to show that Rac1 modulates TGF-β1-signalling in PDAC cells by suppressing and promoting, respectively, TGF-β1-induced activation of Smad3 and Smad2, eventually resulting in protection of PDAC cells from excessive growth inhibition by TGF-β1 and in enhanced cell migration (chemokinesis).
Discussion
In this study we initially presented evidence that TGF-β1-induced growth inhibition and cell migration in PDAC cells were differentially and selectively controlled by Smad3 and Smad2, respectively. Knockdown of Smad3 but not Smad2 relieved TGF-β1-induced growth inhibition, indicating that this response was Smad3-dependent, an observation made previously in various other cell types including PANC-1 cells [
9,
36,
25]. In contrast, knockdown of Smad2 decreased the TGF-β1-driven motility of PDAC cells revealing cell migration (or more precisely chemokinesis) to be a Smad2-specific response. This is in line with the demonstration of a crucial role of Smad2 in regulating keratinocyte migration during wound healing [
37]. We went on to describe first-time observations, namely that the effects of Smad2 depletion on TGF-β1-mediated growth inhibition and cell migration were largely mimicked by inhibition of Rac1 expression (via siRNA knockdown) or activity (via ectopic expression of a dn Rac1 mutant), or pharmacologic inhibition (via NSC23766), together suggesting a functional link between both proteins. We subsequently confirmed this assumption by showing that Rac1 inhibition abrogated TGF-β1-induced Smad2-specific C-terminal phosphorylation and transcriptional activity but increased TGF-β1-mediated p21
WAF1 expression. Another interesting and novel observation of this study was the mutual amplification of effects such that knockdown of Smad2 or inhibition of Rac1 (without direct modulation of Smad3) enhanced growth inhibition, Smad3-specific transcriptional activity, and C-terminal phosphorylation of Smad3, while knockdown of Smad3 (without direct modulation of Smad2) enhanced both Smad2-specific responses such as cellular migration (this study) and Smad2 phosphorylation by TGF-β [
25]. This suggested functional antagonism between the two R-Smads and that the ratio of Smad3 to Smad2 determines the ultimate outcome of the TGF-β response as demonstrated previously for TGF-β-induced growth inhibition in PANC-1 cells [
25].
The decreases in basal proliferation of PANC-1 and COLO 357 cells following Rac1 inhibition may be largely due to disruption of promitogenic growth factor signalling. PDAC cells, e.g. PANC-1 cells, are well known to autostimulate their proliferation in culture via secretion of EGF. Consequently, both the tyrosine kinase inhibitor tyrphostin AG1478 and the ERK inhibitor U0126 dramatically inhibited PANC-1 cell proliferation (H. U., unpublished data). The intimate relationship between the TGF-β and EGF-R pathways in growth regulation of carcinoma cells is also evident from studies showing that TGF-β1 can suppress PDAC cell proliferation by repressing EGF-R-induced ERK activation [
38] and that EGF signalling, in turn, is permissive for regulation of gene expression and growth suppression by TGF-β1 [
39]. Previous observations of TGF-β1 secretion in vitro [
33], and suppression of "basal" p-Smad2/3 levels and
BGN mRNA upon ALK5 inhibition [
21,
23,
40] clearly suggested that PANC-1 cells may also exhibit autocrine TGF-β growth inhibition. Previous studies in breast cancer cells have shown that cell cycle progression/inhibition is subject to regulation by autocrine TGF-β [
41,
42]. In order to block autocrine TGF-β signalling we used PP1, which in PDAC cells effectively blunted growth inhibition induced by exogenously added and autocrine TGF-β's (Figure
7). Importantly, in the presence of PP1 siRNA-mediated Rac1 depletion resulted in much less growth inhibition than in control transfected cells with functional TGF-β/Smad signalling. Hence, reduced DNA synthesis in cells with low Rac1 activity (and not exposed to exogenous TGF-β1) may, at least in part, be explained by increased susceptibility to autocrine growth inhibition by TGF-β's. Similar observations (an increase in growth suppression even in the absence of exogenous TGF-β) were made by Kim and coworkers [
25] upon depletion of Smad2 in PANC-1 cells and these authors showed that this response disappeared in the presence of neutralizing anti-TGF-β antibody. These results perfectly match our data on the sensitization to autocrine TGF-β responses obtained through pharmacologic inhibition of ALK5 and further support our hypothesis of Rac1-mediated control of Smad2 activation.
Interestingly, the decrease in basal and TGF-β1-induced growth upon dn Rac1 expression was accompanied by a respective increase in expression of
p21
WAF1
. In line with these results, Rac1 activity was both necessary and sufficient for suppression of p21
WAF1 in prostate cancer cells [
19].
As discussed above, the decreases in basal proliferation following Rac1 inhibition may involve both disruption of promitogenic growth factor signalling and loss of protection from autocrine TGF-β-mediated growth inhibition as a consequence of the shift from p-Smad2 to p-Smad3 signalling. Similarly, as the inhibition of Rac1 was much more effective in suppressing basal and TGF-β1-induced cell migration than was the inhibition of Smad2 expression (compare Figures
2 and
5), Rac1 is likely to control cell motility, too, in part in an autocrine TGF-β-dependent (and TGF-β-independent) fashion.
There is now ample evidence that Smad2 and Smad3 have distinct functional and non-overlapping roles in TGF-β signalling [
9] implying that intracellular factors which control the relative activation state of Smad2 versus Smad3 signalling have a central role in determining the final outcome of the TGF-β response. Here, we showed that PANC-1 cells responded to inhibition of Rac1 with a pronounced decrease in TGF-β1-mediated p-Smad2 and a slight increase in p-Smad3. In agreement with these data, dn Rac1 expression not only decreased Smad2-specific transcriptional activity (on pAR3-luc) but enhanced general Smad3-specific transcriptional activity (on pCAGA-luc). Moreover, dn Rac1 also increased
p21
WAF1
protein expression which is in line with data showing that
p21
WAF1
was transcriptionally induced by TGF-β in a Smad3-dependent manner in pancreatic, hepatic and skin cells [
30,
18,
43]. However, TGF-β-induced transcription of another reporter gene (p3TP-lux) in HepG2 cells was effectively inhibited by Rac1-N17 expression [
20] which might be explained by the fact that this plasmid is partially responsive to non-Smad (e.g. p38 MAPK) signalling. With respect to the functional antagonism observed, a likely explanation is that Smad2 and Smad3 compete with each other either i) for binding to TβRI/ALK5, ii) capture of Smad4 in the cytoplasm, or iii) recruitment of transcriptional corepressors to SBEs in the nucleus, the latter of which is normally performed by Smad2 [
1]. As a consequence, a reduction in Smad2 expression or activation would increase the ability of Smad4 to bind Smad3 on the SBEs of target gene promoters. In agreement with this possibility are experiments in PANC-1 cells, in which direct silencing of Smad2 via siRNA transfection did not only augment TGF-β1-induced Smad3 phosphorylation [
25], p21
WAF1 expression and growth inhibition (Additional file
1 Figure S1 and Ref. [
25]), but also potentiated TGF-β1-induction of Smad3-regulated genes such as
MMP2 and
BGN (see Figure
1). Indirect evidence that the endogenous ratio of Smad2 and Smad3 determines the quality of the TGF-β response was observed in Hep3B cells, in which the expression of Smad3-Smad4-dependent TGF-β target genes was further enhanced after selective knockdown of
SMAD2[
44], and in mouse keratinocytes, in which Smad2 loss led to a significant increase in Smad3-Smad4 binding to the promoter of the transcription factor Snail, Snail upregulation, and EMT [
45]. Indirect evidence that competition can be mutual comes from a study with Smad2 and Smad3-deficient fibroblasts, in which activation of the pAR3-luc reporter, though strongly suppressed in Smad2-deficient fibroblasts, was
enhanced in Smad3-null cells [
46]. Regarding the intracellular site of competition (see above) our data favour Smad recruitment or binding to ALK5 since dn Rac1 stimulated a shift from p-Smad2 to p-Smad3.
As mentioned above, Rac1 has been found to be overexpressed in PDAC [
17] along with high activity of Vav1 [
11]. Hyperactive Rac1 could therefore increase basal growth through its (TGF-β/Smad-independent) growth promoting effect and, at the same time, protect tumour cells, which have not yet accumulated inactivating mutations in the TGF-β pathway, from exaggerated growth restraints by TGF-β. More specifically, Rac1 aids cancer cells to more efficiently antagonize TGF-β1/Smad3-mediated growth inhibition via its ability to promote Smad2 activation. Interestingly, hyperactive Ras (present in both PANC-1 and COLO 357 cells) has been shown, like Rac1, to suppress ALK5-mediated Smad3 phosphorylation and growth inhibition [
47]. Oncogenic Ras-induced transformation can lead to the production of superoxide through one or more pathways involving NAD(P)H oxidase(s)/Nox1 and Rac1 [
48]. In this way Rac1 may act as a mediator of Ras-induced cell cycle progression independent of MAPK and JNK and may contribute to the unchecked proliferation of Ras-transformed cells [
48]. Notably, preliminary data from our laboratory indicate that Rac1 acts through ROS and NAD(P)H oxidase to promote Smad2 phosphorylation (H. U., unpublished observation).
The mechanism described here for Rac1 differs from the previously described ones in that it reciprocally targets Smad2 and Smad3 at the posttranscriptional level. It is widely appreciated that Rac1 acts in a prooncogenic fashion during later stages of tumour progression by promoting migration, invasion, and metastasis [
13,
14]. In addition to fundamental differences in the mechanism of Smad2 and Smad3 activation by TGF-β1, at least in PDAC cells, our study reveals that Rac1 may drive tumourigenesis in carcinoma cells with a still intact TGF-β/Smad pathway by favouring resistance to TGF-β1-mediated growth inhibition and by increasing TGF-β1-induced cell migration at the R-Smad epigenetic level.
Methods
Antibodies and reagents
TGF-β1 was purchased from R&D Systems (Wiesbaden, Germany). The antibodies and their suppliers were: Rac1, p21WAF1: BD Transduction Laboratories (Heidelberg, Germany); phospho-Smad2 (Ser465/467), phospho-Smad3 (Ser423/425)/Smad1(Ser463/465), HSP90, MYC-Tag (clone 9B11): Cell Signalling Technology (Heidelberg, Germany); Smad2: Zymed (Berlin, Germany); FAK (C-20), Smad2/3 (E-20): Santa Cruz Biotechnology (Heidelberg, Germany); β-actin, FLAG (M2): Sigma (Deisenhofen, Germany); HA (12CA5): Roche Diagnostics (Mannheim, Germany), active Rac1: NewEast Biosciences (Malvern, USA). PP1 analog, the Smad3 inhibitor SIS3, and the Rac1 inhibitor NSC23766 were purchased from Calbiochem/Merck (Darmstadt, Germany). Pharmacological inhibitors were added to cells 30 min before the addition of TGF-β1 which was used at 5 ng/ml for both PANC-1 and COLO 357 cells.
Cell lines and cell culture
Maintenance of the human PDAC cell lines PANC-1 and COLO 357 was described earlier [
23]. PANC-1 cells stably transduced with dn Rac1 retroviral vectors were cultured in the presence of 2.5 μg/ml puromycin (Sigma).
RNA isolation and RT-PCR analysis
Total RNA from PANC-1 cells was isolated with peqGOLD RNAPure (Peqlab, Erlangen, Germany) and reverse transcribed using Superscript II Reverse Transcriptase (Invitrogen). The primer sequences for BGN, β-actin, MMP-2, and TATA box binding protein (TBP) were given earlier [
7,
23]. The mRNA expression was quantified by quantitative real-time RT-PCR (qPCR) on an I-Cycler (Bio-Rad, München, Germany) with I-Cycler software (Bio-Rad). SYBR green was used for detection of amplification products. All values for BGN and MMP-2 mRNA concentrations were normalized to those for β-actin and TBP-specific transcripts in the same sample to account for small differences in cDNA input.
Construction of vectors and retroviral infection
The construction of a retroviral vector (pBABEpuro) for human dn Rac1 (T17N mutation) [
21] and of pcDNA3-based expression vectors for FLAG-tagged Smad2 and GADD45β [
40] was described previously. A cDNA insert of a MYC-tagged version of dn Rac1 was released from the pRK5-MYC vector and subcloned in pcDNA3.
Transient transfections of expression vectors and siRNAs and reporter gene assays
For transient transfections followed by immunoprecipitation (IP), PANC-1 cells were seeded at a density of 2 × 10
4 cells/cm
2 in 6-cm plates on day 1, and on day 2 were co-transfected serum-free with Lipofectamine Plus (Invitrogen) according to the manufacturer's instructions with FLAG-tagged Smad2 in combination with either empty pcDNA3 vector, HA-tagged FRNK, MYC-tagged dn Rac1 (T17N), or MYC-tagged ca Rac1 (Q61L) as indicated in the legend to Figure
7. Following removal of the transfection solution and a recovery period of 24 h in normal growth medium, cells were stimulated with TGF-β1 for 1 h. The transfected cells were then lysed in IP buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerolphosphate, 1 mM Na
3VO
4, 1 μg/ml leupeptin, 1 mM PMSF) and processed for anti-FLAG, anti-HA, and anti-MYC immunoprecipitation and immunoblotting (see below). SiRNAs specific for Rac1 (ON-TARGETplus SMARTpool, a mixture of four prevalidated siRNAs) and matched negative control (non-target control SMARTpool) were purchased from Thermo Scientific Dharmacon (Epsom, UK), while prevalidated siRNAs to Smad2 and Smad3 as well as matched control were from Qiagen (Hilden, Germany) [
33]. Rac1, Smad2/3, and negative control siRNAs were transfected twice on two consecutive days with either Lipofectamine 2000 or Lipofectamine RNAiMax (PANC-1) and HiperFect (COLO 357) (all from Invitrogen) according to the supplier's recommendations. For reporter gene assays, cells were seeded in 96-well plates and were co-transfected on the next day serum-free with either Lipofectamine Plus or Lipofectamine 2000 (Invitrogen) with various cDNAs (in pcDNA3) at an equal molar ratio together with dn Rac1 and either pAR3-luc+FAST-1, or pCAGA-luc, along with the
Renilla luciferase encoding vector pRL-TK (Promega, Mannheim, Germany). Each well received the same total amount of DNA (0.2 μg) and empty vector was added as needed. Following transfection and TGF-β1 stimulation, luciferase activities were determined with the Dual Luciferase Assay System (Promega). Pilot experiments with pCAGA-luc and increasing concentrations (5, 25, and 100 ng/well) of dn Rac1-pcDNA3 DNA indicated that the effect of dn Rac1 was dose-dependent (data not shown). In case of combined siRNA/plasmid DNA transfections PANC-1 cells underwent a first round of transfection with siRNA alone (on day 1 after seeding) and Lipofectamine RNAiMax, followed by a second round with siRNA plus plasmid DNA and Lipofectamine 2000 (on day 2). In all reporter gene assays the data were derived from 6-8 wells processed in parallel and corrected for transfection efficiency with
Renilla luciferase activity.
Immunoprecipitation and immunoblot analysis
Epitope-tagged proteins were immunoprecipitated from cellular lysates with anti-FLAG, anti-HA, or anti-MYC antibodies and Protein A Sepharose Fast Flow (Amersham Biosciences, Freiburg, Germany) or protein G Plus Sepharose (Santa Cruz Biotechnology) according to the protocol provided by the supplier, and subsequently analyzed by SDS-PAGE and immunoblotting as described in detail earlier [
23].
Proliferation and apoptosis assays
Cell counting of was performed with Cedex XS cell analysis system (Roche Diagnostics, Mannheim, Germany) according to the instruction manual. The methyl-[
3H]-thymidine incorporation assay was essentially carried out as described previously [
7,
23]. Twenty-four hours after transient transfection with expression plasmids for dn Rac1 or GADD45β, a JAM DNA fragmentation assay was performed as outlined in detail earlier [
49]. Briefly, transfected PANC-1 cells were trypsinized and reseeded at a density of 1-2 × 10
4 cells/well into 96-well flat bottom plates, allowed to adhere overnight and labelled with [
3H]-thymidine (370 KBq/μl) for 4 h. Subsequently, non-incorporated radioactivity was removed by washing the cells with PBS. Following incubation with TGF-β1 in normal growth medium for 24 h, cells were harvested by vacuum aspiration on glass fiber filters. Dried filters were counted into a liquid scintillation counter (Wallac, Switzerland). The percentage of specific DNA fragmentation, indicative of apoptosis, was calculated as: % viability = (
E/S) × 100, where
E (experimental) is cpm of retained DNA in the presence of TGF-β1 and
S (spontaneous) is cpm of retained DNA in the absence of TGF-β1.
Measurement of cell migration
Using the xCELLigence DP device from Roche Diagnostics real-time measurements of cell migration on wild type or transfected PANC-1 and COLO 357 cells were performed. 60,000-90,000 cells were seeded per well in CIM-Plates 16 (Roche Diagnostics). Prior to cell seeding the underside of the wells was coated with collagen I (400 μg/ml, Sigma, Deisenhofen, Germany) which was chosen since it represents the major matrix protein in PDAC tissue. TGF-β1 (and in some experiments pharmacologic inhibitors) were added to both lower and upper wells at the same concentration. The RTCA assay was performed as detailled by Roche Diagnostics in the instruction manual. In those experiments in which cells underwent transfection they were processed to enter the assay 24-48 hrs after the second round of transfection. In experiments involving small molecule inhibitors, cells were pretreated for 1 h before the addition of TGF-β1. Data acquisition and analysis were performed with the RTCA software (version 1.2, Roche Diagnostics) over a period of 48 h.
Statistical analysis
Statistical significance was calculated using the unpaired student's t-test. Data were considered significant at p < 0.05. Calculated levels of significance were p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
HU, SG and SS performed the experiments. FG and HL contributed to the interpretation and discussion of the results. Both HU and FF are the principal investigators and were involved in the conceptualization and discussion of the manuscript. HU wrote the manuscript. All authors read and approved the final version of the manuscript.