The interaction between epithelial cancer cells and cancer-associated fibroblasts (CAFs) has a major role in cancer progression and eventually in metastasis. In colorectal cancer (CRC), CAFs are present in high abundance, but their origin and functional interaction with epithelial tumor cells has not been elucidated. In this study we observed strong activation of the transforming growth factor-β (TGF-β)/Smad signaling pathway in CRC CAFs, accompanied by decreased signaling in epithelial tumor cells. We evaluated the TGF-β1 response and the expression of target genes including matrix metalloproteinases (MMPs) and plasminogen activator inhibitor (PAI)-1 of various epithelial CRC cell lines and primary CAFs in vitro. TGF-β1 stimulation caused high upregulation of MMPs, PAI-1 and TGF-β1 itself. Next we showed that incubation of CAFs with conditioned medium (CM) from epithelial cancer cells led to hyperactivation of the TGF-β signaling pathway, enhanced expression of target genes like PAI-1, and the expression of α-smooth muscle actin (α-SMA). We propose that the interaction of tumor cells with resident fibroblasts results in hyperactivated TGF-β1 signaling and subsequent transdifferentiation of the fibroblasts into α-SMA-positive CAFs. In turn this leads to cumulative production of TGF-β and proteinases within the tumor microenvironment, creating a cancer-promoting feedback loop.
The interaction between carcinoma cells and fibroblasts within the tumor microenvironment contributes to cancer initiation, progression and eventually metastasis in many cancer types.1, 2 Cancer-associated fibroblasts (CAFs) are a heterogeneous group of fibroblasts expressing vimentin, fibroblast-activating protein and α-smooth muscle actin (α-SMA).3, 4, 5, 6 Compared to normal mucosa the number of myofibroblasts is strongly increased in colorectal cancer (CRC),5, 7 where they influence the immune response and show increased synthesis of chemokines, cytokines, proteolytic enzymes and several extracellular matrix components.8 The origin of CAFs has not fully been clarified yet,6 but owing to primarily in vitro studies, the most commonly accepted hypothesis is that the majority of CAFs are generated by transforming growth factor-β (TGF-β) induced trans-differentiation of resident fibroblasts.1, 9 In addition to its role in CAF formation, TGF-β is involved in tumor progression/metastasis by inducing tumor angiogenesis, increasing production of extracellular matrix, proteolytic enzymes and immune suppression.10, 11, 12
TGF-β is synthesized as a latent precursor consisting of a TGF-β homodimer non-covalently linked to the latency associated protein, forming the small latent complex (SLC). The latent TGF-β binding protein (LTBP) connects the complex to the extracellular matrix as the large latent complex (LLC)13 and is important for TGF-β function.14 TGF-β activation involves removal of both TGF-β binding protein and latency associated protein and occurs through proteolytic processing (for example, plasmin15 and matrix metalloproteinases (MMPs)16) or non-proteolytically through conformational changes (for example, integrins17). The activating mechanism seems to be strongly dependent on the cell/tissue type or the experimental conditions. Active TGF-β binds to the widely expressed TGF-β type-II receptor (TβRII), which subsequently recruits and transphosphorylates the type-I TGF-β receptor activin receptor-like kinase (ALK)-5. Subsequently, intracellular signaling is initiated through phosphorylation of Smad2/Smad3 and recruitment of Smad4. Smad complexes translocate to the nucleus, where they regulate transcriptional responses.18
In this study we investigated the interaction between epithelial tumor cells and CAFs with respect to TGF-β signaling and consequent regulation of proteinase and α-SMA expression. First, α-SMA expression and TGF-β1 activity were immunohistochemically studied in human CRC specimens. Next, we assessed TGF-β1 response of both epithelial tumor cells and CAFs, the regulation of TGF-β target genes and proteinases implicated in invasion and TGF-β activation. Our data show that epithelial tumor cell derived TGF-β hyperactivates TGF-β signaling in fibroblasts, leading to the generation of CAFs, accompanied by increased TGF-β1 and MMP production, creating a cancer-progressing feedback loop.
TGF-β signaling in colorectal cancer tissues
To analyse TGF-β signaling in CRC we stained 88 CRC tissues and normal colonic mucosa for TGF-β signaling components. In normal mucosa a single layer of vimentin positive pericryptal cells co-stained for α-SMA, indicating a myofibroblast phenotype (Figure 1a, upper panel). Staining for TGF-β1 was low in normal mucosa, although epithelial cells did show nuclear phosphorylated-Smad2 (pSmad2) staining, indicating active TGF-β signaling. Nuclear pSmad2 staining showed an increasing gradient from bottom to top of the crypt (Figure 1a), consistent with the anti-proliferative and pro-apoptotic features of TGF-β in normal tissue. pSmad2 was barely detectable in normal fibroblasts or myofibroblasts.
In contrast, TGF-β staining in CRC tissues was strongly enhanced in malignant epithelial cells and tumor stroma. The majority of fibroblasts (up to 90%) was of the α-SMA-positive myofibroblast phenotype and in 89% of the patients we observed nuclear pSmad2 accumulation (Figure 1a, lower panel). Moreover, absent epithelial nuclear pSmad2 was observed in 58% of the CRCs, indicating a shift from primarily epithelial TGF-β signaling in normal tissues to mesenchymal signaling in tumors (Figure 1b, upper panel). In contrast, the lower panel in Figure 1b shows a representative picture of 42% of the tumors in which the malignant epithelial cells still displayed nuclear pSmad2 localization, in addition to the surrounding CAFs.
Next we analyzed the relationship between epithelial presence of pSmad2 and survival of CRC patients. These data revealed that patients with epithelial pSmad2 have a significant worse disease-free survival compared to patients with absent epithelial pSmad2 (Log rank 3.78, P=0.05, Figure 1c).
Characterization of TGF-β response in CAFs
To examine the TGF-β response in CAFs, we isolated CAFs from five CRC tissues. The number of α-SMA-positive CAFs was evaluated by immunocytochemistry and varied per tumor between 32% and 92% (Supplementary Figure 1A). All CAFs showed a dose-dependent TGF-β transcriptional response (up to 100-fold) as shown in Supplementary Figure 1B and a representative graph in Figure 2a. TGF-β-induced Smad2 phosphorylation was also seen in all CAFs tested (Supplementary Figure 1C). Real-time PCR analysis revealed high basal expression of collagen type 1 and plasminogen activator inhibitor (PAI)-1. Expression of MMPs in CAFs was generally high compared to CRC cells (Table 1). TGF-β1 treatment increased the expression of collagen-1, PAI-1, urokinase type plasminogen activator, various MMPs and tissue inhibitors of matrix metalloproteinases in CAFs (Figure 2b). Interestingly, the expression of TGF-β1 was induced 18-fold under TGF-β1 stimulation, indicating a strong autocrine regulatory loop. The protein expression of a selection of the upregulated genes was verified using various techniques. TGF-β1 stimulation of CAFs, which secrete low but detectable basal levels of TGF-β1 (Supplementary Figure 1D), indeed led to a strong increase of total (latent and active) TGF-β1 protein levels combined with a modest increase in TGF-β1 activity (Figure 2c). Gelatin-zymography analysis revealed strongly increased MMP-2 and MMP-9 levels in the medium of stimulated CAFs compared to non-stimulated cells (Figure 2d). To confirm the role of MMPs in invasion of these CAFs we performed a spheroid invasion assay. CAF spheroids were embedded in a collagen-1 matrix in presence of TGF-β1 or GM6001, a broad-spectrum MMP inhibitor. The data revealed that CAFs invade the collagen matrix, which is slightly induced by TGF-β1 and inhibited by GM6001, confirming the role of MMPs in CAF invasion (Figure 2e).
Characterization of TGF-β response in CRC cells
To evaluate the variation in TGF-β1 response in CRC cells we first analyzed the expression of the TGF-β type II receptor (TβRII) and Smad4 as an important downstream signaling molecule in a panel of CRC cells. HT29 and SW480 cells did not show detectable Smad4 expression, although Lovo cells had very low expression of TβRII (Figure 3a). To further analyze downstream signaling, cells were stimulated with TGF-β1 and pSmad2 was analyzed. The data revealed pSmad2 in all cells after TGF-β1 stimulation, except for Lovo cells (Figure 3b). To assess TGF-β signaling more functionally, we used the Smad3/Smad4 transcriptional reporter construct CAGA-luc in all cell lines. LS180, Lovo, SW480, SW948 and CaCo-2 cells showed no induction of luciferase activity after TGF-β1 treatment (data not shown). HT29 cells were partly responsive; only stimulation with concentrations equal to or higher than 5 ng/ml (200 pM) TGF-β1 increased transcriptional activity (Figure 3c). In contrast, HCT116 cells showed dose-dependent induction of luciferase activity up to fivefold when treated with TGF-β1 (Figure 3d). Addition of an ALK-5 kinase inhibitor (SB431542) abolished TGF-β1-induced transcriptional responses. As TGF-β can inhibit cell proliferation, we analyzed this in HT29 and HCT116 cells. These data revealed that cell proliferation was not altered after 6 h (not shown) or 72 h of TGF-β1 treatment (MTS proliferation assay, Figure 3e).
To further characterize these cells we analyzed the basal expression levels of TGF-β1, PAI-1, and invasion-related proteinases in HCT116, HT29, SW948 and CaCo-2 cells by quantitative PCR. The data showed high TGF-β1 expression in all cell lines, intermediate levels of PAI-1, and variable expression of urokinase type plasminogen activator and MMPs (Table 1). As expected, collagen-1 was not expressed. To analyze TGF-β1-mediated regulation of these genes, HT29 and HCT116 cells were treated with 5 ng/ml (200 pM) TGF-β1 and expression was analyzed by quantitative PCR. Figure 4a shows that TGF-β induces urokinase type plasminogen activator and PAI-1 expression in HT29 cells, yet several MMPs and tissue inhibitors of matrix metalloproteinases were even stronger upregulated. In contrast TGF-β1 mRNA only showed a twofold upregulation. In HCT116 cells a similar pattern was observed (Figure 4b), but the induction of MMP-2, MMP-13, PAI-1 and urokinase type plasminogen activator was much stronger in these cells. However, TGF-β1 was not changed at all in HCT116 cells after TGF-β stimulation.
Increased proteinase expression is associated with invasiveness of tumor cells through extracellular matrix degradation and by processing/release of cytokines. Figure 4c illustrates the effect of TGF-β1 in an invasion assay using HT29 spheroids embedded in a collagen matrix. TGF-β1 treatment clearly induced migration of cells from spheroids into the collagen matrix, eventually leading to distant metastasis-like cell clusters. Unfortunately, the effect on highly responding HCT116 cells could not be evaluated due to poor spheroid formation of these cells.
Latent TGF-β1 activation
Cells secrete TGF-β as a latent complex, which needs processing to become biologically active. Therefore, we evaluated whether interactions within the tumor microenvironment could enable activation of the latent TGF-β1 complex. TGF-β responsive HCT116 cells were transfected with the CAGA-luc reporter and incubated with 10 ng/ml (121 pM) recombinant small latent TGF-β1 (SLC), which did not result in increased luciferase activity, implying no activation of the latent complex (Figure 4d). HT29 and HCT116 cells secrete high levels of TGF-β1 (1 and 2 ng/ml, respectively, Figure 4e). In addition, HT29 and HCT116 cells secrete minor amounts of TGF-β3, whereas only HCT116 cells secrete TGF-β2 (Figure 4e). To analyze if these cells secrete large or small latent TGF-β1, conditioned medium (CM) was analyzed by western blot for the latency associated protein and TGF-β1. Clear bands at ∼200 kDa under non-reducing conditions indicate that both HT29 and HCT116 secrete large latent TGF-β1 (LLC, Figure 4f). To further evaluate the TGF-β1 response in these cells, HCT116 cells were stimulated with HT29 CM, leading to a minor increase in luciferase activity. In contrast, stimulation of HCT116 cells with CAF-CM, containing low levels latent and active TGF-β1 (Figure 2c), induced a fourfold increase in activity (Figure 4d), almost equal to 5 ng/ml recombinant TGF-β1. These data indicate that the interaction of tumor cells with CAF-derived factors initiates TGF-β signaling, possibly through activation of latent TGF-β1.
TGF-β signaling in CAFs and interaction with tumor cells
As we observed increased TGF-β signaling when HCT116 cells were treated with CAF-CM, we evaluated the response of CAFs after stimulation with tumor cell CM. Stimulation of CAFs with 10 ng/ml (121 pM) SLC or CaCo-2 CM did not increase luciferase activity, whereas stimulation with HT29 CM (∼1 ng/ml LLC) revealed a dose-dependent 46-fold increase in luciferase activity (Figure 5a). Moreover, treatment with HCT116 CM containing high levels (2 ng/ml) of LLC and only little active TGF-β1 (25–80 pg/ml) led to a dose-dependent 71-fold increase compared to control levels. Addition of ALK-5 kinase inhibitor completely abolished the response, confirming that the induction was TGF-β mediated (Figure 5a). Analysis of the CAF medium after incubation with HCT116 CM showed enhanced active TGF-β1 levels (Figure 5b). Furthermore, almost 95% of CAFs treated with HCT116 CM were α-SMA positive, which is comparable to addition of 5 ng/ml TGF-β1, whereas parallel incubations with SLC or HT29 CM showed no change in α-SMA expression (Figure 5c). These data indicate that the interaction of CAFs with HCT116 CM leads to hyperactivation of the TGF-β pathway, involving TGF-β activation and as a consequence α-SMA induction and myofibroblast trans-differentiation.
To evaluate the mechanism of TGF-β activation, CAFs were stimulated with HCT116 CM in the presence of inhibitors of proteases, integrins or reactive oxygen species, all implicated in latent TGF-β activation in vitro. None of the inhibitors, that is, Aprotinin (plasmin and serine protease inhibitor), GM6001 (broad range MMP inhibitor), E64 (broad range cystein protease inhibitor), specific inhibitors of MMP-2/-9, MMP-3 and MMP-13, and α2-macroglobulin (BMP-1 and MMP inhibitor), were able to inhibit the HCT116 CM-induced responses (Figure 5d). To exclude the possible necessity of mutual activating proteolytic cascades, different cocktails of inhibitors were tested. These protease inhibitor cocktails also did not inhibit HCT116 CM-induced responses, whereas the ALK-5 kinase inhibitor and neutralizing TGF-β antibody inhibited the signal by more than 90%.
Hyperactivation of TGF-β signaling in CAFs by HCT116 CM
As we observed very strong increase in luciferase activity after stimulation of CAFs with HCT116 CM, we evaluated the effects on a major TGF-β target gene, PAI-1. Figures 6a–c show a strong increase of PAI-1 expression after 6, 12 and 24 h of TGF-β1 stimulation. Interestingly, the induction of PAI-1 expression by HCT116 CM is higher and lasts longer than stimulation with TGF-β1. Especially after 24 h, no induction by TGF-β1 is detected anymore, although HCT116 CM still shows 29-fold induction of PAI-1 mRNA. We further evaluated the underlying mechanism by western blot. Analysis of downstream pSmad2 and pSmad3 revealed increased phosphorylation after stimulation with HCT116 CM compared to TGF-β1 (Figure 6d). Similarly, PAI-1 protein levels after 6 h were higher in HCT116 CM compared to TGF-β1 stimulated cells. Finally, stimulation of CAFs with HCT116 CM led to earlier and prolonged (up to 360 min) pSmad2 levels (Figure 6e) followed by higher PAI-1 protein expression. To examine whether the expression of invasion related proteinases is also enhanced upon HCT116 CM stimulation, we analyzed expression of MMP-2, MMP-3 and MMP-14. mRNA levels of MMP-2 were similar in HCT116 CM and TGF-β1 treated cells whereas MMP-3 and MMP-14 were twofold stronger induced by HCT116 CM than by TGF-β1 (Figure 6f). Furthermore, TGF-β1 mRNA levels were also strongly upregulated after HCT116 CM stimulation. Stimulation with HT29 CM showed a similar trend, although less pronounced as HCT116 CM, confirming the CAGA-luc reporter data (Figure 5a). These data indicate hyperactivated TGF-β signaling accompanied by increased expression of TGF-β regulated genes in CAFs as a result of interaction with epithelial tumor cell-derived soluble factors.
The present study shows that TGF-β signaling in CRC shifted from mainly epithelial TGF-β signaling to signaling in CAFs. In addition, we have shown that the interaction between CAFs and epithelial tumor cells leads to the hyperactivation of TGF-β signaling in CAFs, accompanied by increased expression of invasion-related proteinases and elevated TGF-β secretion.
Although most cancers originate from epithelial cells, the tumor stroma might actively participate in the progression of carcinomas, as was illustrated by the prognostic relevance of stroma quantity for the survival of CRC patients.19 Within the tumor microenvironment, especially CAFs are associated with auto- and paracrine signaling,1, 6 hereby influencing epithelial cell proliferation and enhancing the tumors’ invasive and metastatic potential.20 Interactions between cancer cells and CAFs can occur through direct cell−cell contact through for instance extracellular matrix metalloproteinases inducer, a potent inducer of the myofibroblast phenotype.21 However, the majority of interactions are mediated by soluble factors. Amongst the fibroblast-derived soluble factors affecting epithelial cancer cells are scatter factor/hepatocyte growth factor,22, 23 insulin-like growth factor, TGF-β1 and several others.24, 25 TGF-β1 is implicated in activation of fibroblasts leading to the generation of CAFs.26, 27 Our immunohistochemistry data confirmed the shift in TGF-β signaling from epithelial cells to CRC CAFs. In normal mucosa (myo-) fibroblasts showed hardly any nuclear pSmad2 staining, indicating low TGF-β signaling in the normal stroma. In CRC, however, signaling in most of the epithelial cancer cells was lost, which corresponds to previous studies showing that in later stages epithelial cells become refractory to the TGF-β growth inhibiting properties.12, 28, 29 In vitro experiments have shown that the majority of CRC epithelial tumor cell lines are not responsive to TGF-β, a phenomenon caused by mutations in the TβRII, Smad4 or other pathway components.30, 31 Our data show that even CRC cells with a functional TGF-β pathway are not growth inhibited by TGF-β1, but instead respond through increased expression of invasion related proteinases. This is further confirmed by the observation that CRC patients with present epithelial TGF-β signaling show worse disease-free survival compared with patients with absent TGF-β signaling.
With regard to CAFs we observed an enrichment of pSmad2-positive population in CRC indicating active TGF-β signaling in these cells, probably caused by increased levels of TGF-β1 in the tumor microenvironment. In vitro CAFs responded robustly to (tumor derived)-TGF-β1 by trans-differentiation into α-SMA-expressing myofibroblasts and concomitant upregulation of target genes PAI-1, collagen-1, MMPs and particularly TGF-β1. Upregulation of latent growth factors is a common phenomenon in many cancer types, but the subsequent activation is even more crucial for their effects.32, 33 For TGF-β1, we have shown previously that although latent TGF-β1 is already upregulated in pre-malignant adenomas, increased TGF-β1 activity is only observed in CRC,34 suggesting that over-activation and not over-expression is characteristic of malignant progression. In the present study we showed that CAFs were not effective in releasing TGF-β1 from the SLC. However, treatment with tumor cell-derived CM induced TGF-β signaling in CAFs and expression of target genes, including invasion-related proteinases, to a higher extent than stimulation with recombinant TGF-β1. Earlier studies already showed that coculture of dermal fibroblasts with keratinocytes increased myofibroblast numbers35 and that breast cancer cells in coculture with fibroblasts are also able to induce the myofibroblast phenotype through direct cell−cell contact.36 Our data indicate hyperactivation of TGF-β signaling in CAFs in a TGF-β1-dependent manner, with collaboration of other soluble factors, secreted by CRC cells, leading to CAF activation.
TGF-β activation occurs through proteolytic processing of the LLC by MMPs, plasmin or BMP1, followed by a conformational change generated by integrins.13, 15, 16, 37, 38, 39 Another mechanism involves TGF-β activation by reactive oxygen species.40 As we observed increased TGF-β signaling after incubation with tumor cell-derived CM containing mainly latent TGF-β1 and this effect was accompanied by increased active TGF-β1 levels, we also evaluated the potential activation mechanism of TGF-β1. No inhibition of TGF-β signaling after addition of protease inhibitors, a RGD peptide blocking integrins, or reactive oxygen species inhibitor n-acetylcysteine was observed. To exclude the possibility that cascades of proteolytic activation of proproteinases would be required cocktails of inhibitors were used, which also did not show any effect. Other factors secreted by tumor cells might also influence TGF-β effects on myofibroblast differentiation, as has been shown for interleukin-1, which can inhibit TGF-β-induced myofibroblast differentiation.35 However, incubation with an ALK5 kinase inhibitor or neutralizing pan-TGF-β antibody completely abolished the induction of target genes and reporter activity indicating a strong TGF-β dependency.
In conclusion, we have shown that the soluble, non-cell-contact-dependent interaction between epithelial tumor cells and CAFs increases proteinase expression and thereby tumor progression. We propose that tumor cells during cancer progression lose the anti-proliferative response to TGF-β and gain pro-invasive properties through increased proteinase production. Epithelial tumor cells produce high amounts of TGF-β1 LLC, which hyperactivates TGF-β signaling in fibroblasts leading to their trans-differentiation into CAFs. CAFs in turn show strongly increased production of proteinases and TGF-β1 creating a cancer progressing feedback loop (Figure 7). These data further emphasize the role of tumor-stroma interactions, in particular myofibroblasts, and validate further exploration of the tumor stroma, including TGF-β as potential therapeutic target.3, 25, 41
Materials and methods
Tissue specimens from patients undergoing surgical resection for CRC at the Department of Surgery, Leiden University Medical Centre, were collected. Tissues were formalin-fixed, dehydrated and embedded in paraffin. The patient population contained stage I−III CRC patients and is part of the cohort described before.19 Fresh tissues from the patients were used for isolation of primary CAFs. The study was performed according to the guidelines of the Medical Ethics Committee of the Leiden University Medical Centre.
Immunohistochemical staining was performed as described before.32 In short, sequential sections were deparaffinized, rehydrated and endogenous peroxidase activity was quenched. Antigen retrieval was performed by boiling in 0.01 M sodium citrate, pH 6.0, for 10 min followed by overnight incubation at room temperature with primary antibodies described in Table 2. Sections were incubated with biotinylated secondary antibodies, streptavidin−biotin complex (Dako, Glostrup, Denmark) and staining was visualized using diaminobenzidine and H2O2. Representative photomicrographs were taken with a Nikon Eclipse E800 microscope equipped with a Nikon DXM1200 digital camera (Nikon instruments, Melville, NY, USA).
Cell culture, isolation of CAFs and preparation of CM
HT29, HCT116, LS180, Lovo, SW480, SW948 and CaCo-2 cells were cultured in DMEM/F12+GlutaMAX-1 (Invitrogen, Bleiswijk, the Netherlands), 10 mM HEPES, 50 μg/ml gentamycin, 100 U/ml penicillin and 100 μg/ml streptomycin (all Invitrogen), supplemented with 10% heat-inactivated fetal calf serum (Perbio Science, Aalst, Belgium) or 20% fetal calf serum for CaCo-2 cells. Human CAFs were isolated from the non-necrotic part of the tumors. The tissue was washed with phosphate-buffered saline, cut into 5-mm pieces and incubated in tissue culture flasks. After 7–10 days, fibroblast-like cell outgrowth was observed. The fibroblast origin of these cells was confirmed by positive staining for vimentin and negative staining for pan-cytokeratin. CAFs were maintained in complete DMEM/F12 containing 10% fetal calf serum and used at passage 5–11.
CM from CRC cell lines and CAFs were prepared by seeding the cells and growing them to subconfluence. Medium was changed to serum-free DMEM/F12, containing HEPES and antibiotics as described above, and incubated for 4 days. CM used for stimulation was diluted twofold with fresh serum-free DMEM/F12. To obtain myofibroblast CM, fibroblasts were stimulated with 5 ng/ml recombinant human TGF-β1 (Peprotech, London, UK) for 24 h. After three washes with serum-free DMEM/F12, cells were incubated for 4 days with serum-free DMEM/F12. Medium was analyzed for MMP-2 and MMP-9 levels by zymography as described before42 and for TGF-β1 levels by enzyme-linked immunosorbent assay as described below.
MTS proliferation assay
HT29 and HCT116 cells were seeded in 96-well plates (2000 or 10 000 cells/well) and allowed to attach overnight. Cells were stimulated with 5 ng/ml TGF-β1 for 6 or 72 h. Subsequently the medium was changed to 100 μl complete DMEM/F12+20 μl MTS substrate (Promega, Madison, WI, USA). The metabolic activity of the cells was analyzed by absorbance change at 490 nm after 2 h.
TGF-β response assay
TGF-β response in tumor cells was determined using (CAGA)12−MLP−Luciferase promoter reporter construct.43 This construct contains 12 palindromic repeats of the Smad3/4 binding element derived from the PAI-1 promoter and was shown to be highly specific and sensitive to TGF-β. Tumor cells were seeded in 24-well plates and allowed to attach overnight. Subconfluent cells were transfected using Lipofectamin 2000 (Invitrogen) according to the manufacturer’s protocol. A β-galactosidase plasmid was co-transfected to correct for transfection efficiency. After 6 h, medium was changed to complete DMEM/F12 and the cells were incubated for 24 h and serum-starved overnight. Cells were stimulated with 0−5 ng/ml TGF-β1 or indicated conditions. After stimulation the cells were washed, lysed and luciferase activity was determined according to the manufacturer’s protocol (Promega). β-Galactosidase activity in the lysates was determined using β-gal substrate (0.2 M H2PO4, 2 mM MgCl2, 4 mM ortho-nitrophenyl-phosphate, 0.25% β-mercaptoethanol) and measuring absorbance change at 405 nm. The luciferase count was corrected for β-galactosidase activity. The relative increase in luciferase activity was calculated versus controls.
CAFs were infected using an adenoviral Ad-(CAGA)9−MLP−Luc promoter reporter construct.44 Cells (22 000/well) were infected with 1E+6 p.f.u. virus. After infection, medium was changed to complete DMEM/F12 for 24 h and the CAFs were serum-starved overnight, stimulated, lysed and the luciferase activity was determined as described above.
Western blot analysis
Western blot analysis was performed as described before.45 In short, equal protein amounts (DC protein assay, Bio-Rad Laboratories, CA, USA) were separated on 10% SDS−polyacrylamide gel electrophoresis under reducing conditions. Proteins were transferred to nitrocellulose membranes (Whattman, Dassel, Germany) and unspecific binding was blocked with 5% milk powder in tris-buffered saline containing 0.05% Tween-20 (Merck, Darmstadt, Germany). Blots were incubated overnight with mouse anti-Smad4 antibodies (Santa Cruz Biotechnologies, Santa Cruz, CA, USA), rabbit anti-TβRII (Santa Cruz), rabbit anti-pSmad2,46 rabbit anti-pSmad3 (kindly provided by Dr E Leof (Mayo Clinic, Rochester, MN ,USA), rabbit anti-PAI-1 (Santa Cruz), mouse anti-β-actin (Sigma-Aldrich, St Louis, MO, USA). Detection was performed by horseradish peroxidase-conjugated secondary antibodies (all GE Healthcare, Waukesha, WI, USA) and chemoluminescence according to the manufacturer’s protocol (Pierce, Rockford, IL, USA). Blots were stripped and reprobed with anti-β-actin antibodies as a loading control. To detect latent TGF-β complexes in CM, 10–50 μl CM was run under non-reducing conditions. Blots were probed with anti-latency associated protein (R&D Systems) or anti-TGF-β1 (Promega) antibodies and analyzed as described above.
RNA isolation and real-time PCR
Expression of TGF-β target genes and MMPs were analyzed in HT29, HCT116, SW948 and CaCo-2 cells. Cells were grown to subconfluence and stimulated with 5 ng/ml TGF-β1 or twofold diluted CM for indicated time points, harvested and RNA was isolated using RNeasy Mini Kit according to the manufacturers’ instructions (Qiagen, CA, USA). RNA concentration and purity was determined spectrometrically. Complementary DNA synthesis was performed on 1 μg RNA using random primers. cDNA samples were subjected to 40 cycles real-time PCR analysis using primers as previously described,47, 48 except for MMP-7, MMP-13 and MMP-28 primer sets that were purchased from Applied Biosystems, Bleiswijk, the Netherlands. PAI-1 quantitative PCR analysis was performed as described before.49 All values were normalized for cDNA content by GAPDH or actin expression. TGF-β does not influence GAPDH expression.50
Spheroid invasion assays
HT29 or CAF spheroids were grown as described before.51 Spheroids were collected and embedded in collagen gels, consisting of DMEM/F12 with 0–10 ng/ml (0–400 pM) TGF-β1 and 1 mg/ml collagen type 1 (Vitrogen, Nutacon, CA, USA). HT29 spheroids were analyzed at 4–14 days for invasiveness and the formation of distant metastasis-like cell clusters. CAF spheroids were embedded and treated with 5 ng/ml TGF-β1 or 10 μM GM6001 and analyzed after 1 and 2 days for invasion. Representative photomicrographs were taken using a Zeiss Axiovert microscope equipped with × 10 objective (Carl Zeiss BV, Sliedrecht, The Netherlands).
TGF-β1 activation experiments
To analyze TGF-β1 activation by tumor cells and CAFs these cells were transfected or transduced as described above. Cells were stimulated with 10 ng/ml (121 pM) rh-small latent TGF-β1 complex (R&D Systems), tumor cell CM or CAF-CM.
To examine the contribution of various proteins described to be involved in TGF-β1 activation, CAFs were stimulated with HCT116 CM in the presence of: 10 μM ALK-5 inhibitor SB431542 (Tocris, Bristol, UK), a TGF-β-neutralizing antibody (R&D Systems, 10 μg/ml), 1 μM GM6001 (broad spectrum MMP inhibitor), 1 μM specific MMP-2/9 or specific MMP-13 inhibitor, 10 μM specific MMP-3 inhibitor (all Calbiochem, San Diego, CA, USA), 1 μM Marimastat (broad spectrum MMP inhibitor, kindly provided by British Biotech Pharmaceuticals, Oxford, UK), 10 μg/ml aprotinin (serine protease inhibitor including plasmin), 20 μM E64 (cystein protease inhibitor), 100 nM α-2 macroglobulin (all Sigma-Aldrich) 100 μM RGD (H-Arg-Gly-Asp-OH) peptide (kindly provided by Dr E de Heer, LUMC, Department of Pathology) or 2.5 mM NAC (n-acetyl-cysteine, a reactive oxygen species inhibitor, Sigma-Aldrich). Combinations of inhibitors (Inhibitor cocktail-1: MMP-2/-9,-3,-13 inhibitors, aprotinin and E64; inhibitor cocktail-2: aprotinin, GM6001 and E64) were also tested. After 6 h incubation, luciferase activity was determined as described above. For CAF experiments parallel incubations were performed, medium was harvested for the TGF-β1 ELISA and cells were harvested by trypsinization. Cytospins were prepared and cells were analyzed for α-SMA content by immunocytochemistry.
De Wever O, Mareel M . Role of tissue stroma in cancer cell invasion. J Pathol 2003; 200: 429–447.
Ruiter D, Bogenrieder T, Elder D, Herlyn M . Melanoma-stroma interactions: structural and functional aspects. Lancet Oncol 2002; 3: 35–43.
Micke P, Ostman A . Tumour-stroma interaction: cancer-associated fibroblasts as novel targets in anti-cancer therapy? Lung Cancer 2004; 45 (Suppl 2): S163–S175.
Park CC, Bissell MJ, Barcellos-Hoff MH . The influence of the microenvironment on the malignant phenotype. Mol Med Today 2000; 6: 324–329.
Powell DW, Adegboyega PA, Di Mari JF, Mifflin RC . Epithelial cells and their neighbors I. Role of intestinal myofibroblasts in development, repair, and cancer. Am J Physiol Gastrointest Liver Physiol 2005; 289: G2–G7.
Kalluri R, Zeisberg M . Fibroblasts in cancer. Nat Rev Cancer 2006; 6: 392–401.
Adegboyega PA, Mifflin RC, DiMari JF, Saada JI, Powell DW . Immunohistochemical study of myofibroblasts in normal colonic mucosa, hyperplastic polyps, and adenomatous colorectal polyps. Arch Pathol Lab Med 2002; 126: 829–836.
Lieubeau B, Heymann MF, Henry F, Barbieux I, Meflah K, Gregoire M . Immunomodulatory effects of tumor-associated fibroblasts in colorectal-tumor development. Int J Cancer 1999; 81: 629–636.
Desmouliere A, Geinoz A, Gabbiani F, Gabbiani G . Transforming growth factor-β1 induces alpha-smooth muscle actin expression in granulation tissue myofibroblasts and in quiescent and growing cultured fibroblasts. J Cell Biol 1993; 122: 103–111.
Blobe GC, Schiemann WP, Lodish HF . Role of transforming growth factor-β in human disease. N Engl J Med 2000; 342: 1350–1358.
Derynck R, Akhurst RJ, Balmain A . TGF-β signaling in tumor suppression and cancer progression. Nat Genet 2001; 29: 117–129.
Akhurst RJ, Derynck R . TGF-β signaling in cancer-a double-edged sword. Trends Cell Biol 2001; 11: S44–S51.
ten Dijke P, Arthur HM . Extracellular control of TGF-β signalling in vascular development and disease. Nat Rev Mol Cell Biol 2007; 8: 857–869.
Yoshinaga K, Obata H, Jurukovski V, Mazzieri R, Chen Y, Zilberberg L et al. Perturbation of transforming growth factor (TGF)-β1 association with latent TGF-β binding protein yields inflammation and tumors. Proc Natl Acad Sci USA 2008; 105: 18758–18763.
Lyons RM, Gentry LE, Purchio AF, Moses HL . Mechanism of activation of latent recombinant transforming growth factor β1 by plasmin. J Cell Biol 1990; 110: 1361–1367.
Jenkins G . The role of proteases in transforming growth factor-β activation. Int J Biochem Cell Biol 2008; 40: 1068–1078.
Annes JP, Munger JS, Rifkin DB . Making sense of latent TGF-β activation. J Cell Sci 2003; 116: 217–224.
Meulmeester E, ten Dijke P . The dynamic roles of TGF-β in cancer. J Pathol 2011; 223: 205–218.
Mesker WE, Junggeburt JM, Szuhai K, de Heer P, Morreau H, Tanke HJ et al. The carcinoma-stromal ratio of colon carcinoma is an independent factor for survival compared to lymph node status and tumor stage. Cell Oncol 2007; 29: 387–398.
Pourreyron C, Dumortier J, Ratineau C, Nejjari M, Beatrix O, Jacquier MF et al. Age-dependent variations of human and rat colon myofibroblasts in culture: Influence on their functional interactions with colon cancer cells. Int J Cancer 2003; 104: 28–35.
Huet E, Vallee B, Szul D, Verrecchia F, Mourah S, Jester JV et al. Extracellular matrix metalloproteinase inducer/CD147 promotes myofibroblast differentiation by inducing α-smooth muscle actin expression and collagen gel contraction: implications in tissue remodeling. FASEB J 2008; 22: 1144–1154.
Kankuri E, Cholujova D, Comajova M, Vaheri A, Bizik J . Induction of hepatocyte growth factor/scatter factor by fibroblast clustering directly promotes tumor cell invasiveness. Cancer Res 2005; 65: 9914–9922.
Kankuri E, Babusikova O, Hlubinova K, Salmenpera P, Boccaccio C, Lubitz W et al. Fibroblast nemosis arrests growth and induces differentiation of human leukemia cells. Int J Cancer 2008; 122: 1243–1252.
Elenbaas B, Weinberg RA . Heterotypic signaling between epithelial tumor cells and fibroblasts in carcinoma formation. Exp Cell Res 2001; 264: 169–184.
Cirri P, Chiarugi P . Cancer-associated-fibroblasts and tumour cells: a diabolic liaison driving cancer progression. Cancer Metastasis Rev 2011; 31: 195–208.
Ronnov-Jessen L, Petersen OW, Koteliansky VE, Bissell MJ . The origin of the myofibroblasts in breast cancer. recapitulation of tumor environment in culture unravels diversity and implicates converted fibroblasts and recruited smooth muscle cells. J Clin Invest 1995; 95: 859–873.
Lewis MP, Lygoe KA, Nystrom ML, Anderson WP, Speight PM, Marshall JF et al. Tumour-derived TGF-β1 modulates myofibroblast differentiation and promotes HGF/SF-dependent invasion of squamous carcinoma cells. Br J Cancer 2004; 90: 822–832.
Leivonen SK, Kahari VM . Transforming growth factor-β signaling in cancer invasion and metastasis. Int J Cancer 2007; 121: 2119–2124.
Bierie B, Moses HL . Tumour microenvironment: TGF-β: the molecular Jekyll and Hyde of cancer. Nat Rev Cancer 2006; 6: 506–520.
Munoz NM, Upton M, Rojas A, Washington MK, Lin L, Chytil A et al. Transforming growth factor β receptor type II inactivation induces the malignant transformation of intestinal neoplasms initiated by Apc mutation. Cancer Res 2006; 66: 9837–9844.
Woodford-Richens KL, Rowan AJ, Gorman P, Halford S, Bicknell DC, Wasan HS et al. SMAD4 mutations in colorectal cancer probably occur before chromosomal instability, but after divergence of the microsatellite instability pathway. Proc Natl Acad Sci USA 2001; 98: 9719–9723.
Hawinkels LJ, Verspaget HW, van Duijn W, van der Zon JM, Zuidwijk K, Kubben FJ et al. Tissue level, activation and cellular localisation of TGF-β1 and association with survival in gastric cancer patients. Br J Cancer 2007; 97: 398–404.
Barcellos-Hoff MH, Ewan KB . Transforming growth factor-β and breast cancer: Mammary gland development. Breast Cancer Res 2000; 2: 92–99.
Hawinkels L, Verspaget HW, van der Reijden JJ, Van der Zon J, Verheijen JH, Hommes D et al. Active TGF-β1 correlates with myofibroblasts and malignancy in the colorectal adenoma-carcinoma sequence. Cancer Sci 2009; 100: 663–670.
Shephard P, Martin G, Smola-Hess S, Brunner G, Krieg T, Smola H . Myofibroblast differentiation is induced in keratinocyte-fibroblast co-cultures and is antagonistically regulated by endogenous transforming growth factor-β and interleukin-1. Am J Pathol 2004; 164: 2055–2066.
Kunz-Schughart LA, Heyder P, Schroeder J, Knuechel R . A heterologous 3-D coculture model of breast tumor cells and fibroblasts to study tumor-associated fibroblast differentiation. Exp Cell Res 2001; 266: 74–86.
Ge G, Greenspan DS . BMP1 controls TGFβ1 activation via cleavage of latent TGFβ-binding protein. J Cell Biol 2006; 175: 111–120.
Wipff PJ, Hinz B . Integrins and the activation of latent transforming growth factor β1 - An intimate relationship. Eur J Cell Biol 2008; 87: 601–615.
Nishimura SL . Integrin-mediated transforming growth factor-β activation, a potential therapeutic target in fibrogenic disorders. Am J Pathol 2009; 175: 1362–1370.
Barcellos-Hoff MH, Derynck R, Tsang ML, Weatherbee JA . Transforming growth factor-β activation in irradiated murine mammary gland. J Clin Invest 1994; 93: 892–899.
Hawinkels LJ, ten Dijke P . Exploring anti-TGF-β therapies in cancer and fibrosis. Growth Factors 2011; 29: 140–152.
Hawinkels LJ, Zuidwijk K, Verspaget HW, Jonge-Muller ES, Duijn W, Ferreira V et al. VEGF release by MMP-9 mediated heparan sulphate cleavage induces colorectal cancer angiogenesis. Eur J Cancer 2008; 44: 1904–1913.
Dennler S, Itoh S, Vivien D, ten Dijke P, Huet S, Gauthier JM . Direct binding of Smad3 and Smad4 to critical TGF β-inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. EMBO J 1998; 17: 3091–3100.
Dooley S, Hamzavi J, Breitkopf K, Wiercinska E, Said HM, Lorenzen J et al. Smad7 prevents activation of hepatic stellate cells and liver fibrosis in rats. Gastroenterology 2003; 125: 178–191.
Hawinkels LJ, Kuiper P, Wiercinska E, Verspaget HW, Liu Z, Pardali E et al. Matrix metalloproteinase-14 (MT1-MMP)-mediated endoglin shedding inhibits tumor angiogenesis. Cancer Res 2010; 70: 4141–4150.
Persson U, Izumi H, Souchelnytskyi S, Itoh S, Grimsby S, Engstrom U et al. The L45 loop in type I receptors for TGF-β family members is a critical determinant in specifying Smad isoform activation. FEBS Lett 1998; 434: 83–87.
Lindeman JH, Hanemaaijer R, Mulder A, Dijkstra PD, Szuhai K, Bromme D et al. Cathepsin K is the principal protease in giant cell tumor of bone. Am J Pathol 2004; 165: 593–600.
Sier CF, Hawinkels LJ, Zijlmans HJ, Zuidwijk K, de Jonge-Muller ES, Ferreira V et al. Endothelium specific matrilysin (MMP-7) expression in human cancers. Matrix Biol 2008; 27: 267–271.
Pardali E, van der Schaft DW, Wiercinska E, Gorter A, Hogendoorn PC, Griffioen AW et al. Critical role of endoglin in tumor cell plasticity of Ewing sarcoma and melanoma. Oncogene 2011; 30: 334–345.
Stolle K, Schnoor M, Fuellen G, Spitzer M, Engel T, Spener F et al. Cloning, cellular localization, genomic organization, and tissue-specific expression of the TGFβ1-inducible SMAP-5 gene. Gene 2005; 351: 119–130.
Hawinkels LJ, van Rossenberg SM, de Jonge-Muller ES, Molenaar TJ, Appeldoorn CC, van Berkel TJ et al. Efficient degradation-aided selection of protease inhibitors by phage display. Biochem Biophys Res Commun 2007; 364: 549–555.
We thank Dr R Hanemaaijer (TNO Quality of Life BioSciences, Leiden, The Netherlands) for helpful suggestions and reagents. Eveline de Jonge-Muller (Department of Gastroenterology-Hepatology, LUMC), Adri Mulder-Stapel (TNO) and Gabi van Pelt (Department Surgery, LUMC) are acknowledged for excellent technical support. This work was supported by the EC Tumor-host genomic project, the Centre for Biomedical Genetics, the Swedish Cancer Fonden 090773 (EW, PtD) and the Bas Mulder Award 2011 (LJACH, MP, UL2011-5051).
The authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Oncogene website
About this article
Cite this article
Hawinkels, L., Paauwe, M., Verspaget, H. et al. Interaction with colon cancer cells hyperactivates TGF-β signaling in cancer-associated fibroblasts. Oncogene 33, 97–107 (2014). https://doi.org/10.1038/onc.2012.536
- colorectal cancer
- tumor microenvironment
Crosstalk between cancer-associated fibroblasts and immune cells in the tumor microenvironment: new findings and future perspectives
Molecular Cancer (2021)
PAI-1 derived from cancer-associated fibroblasts in esophageal squamous cell carcinoma promotes the invasion of cancer cells and the migration of macrophages
Laboratory Investigation (2021)
Signal Transduction and Targeted Therapy (2021)
Neutral endopeptidase depletion decreases colon cancer cell proliferation and TGF-β1 synthesis in indirect co-cultures with normal colon fibroblasts
Clinical and Translational Oncology (2021)
Candesartan treatment enhances liposome penetration and anti-tumor effect via depletion of tumor stroma and normalization of tumor vessel
Drug Delivery and Translational Research (2021)