Squamous cell carcinoma (SCC) cells of the head and neck specifically express collagenase-3 (matrix metalloproteinase-13 (MMP-13)), the expression of which correlates with their invasion capacity. Transforming growth factor-β (TGF-β) enhances MMP-13 and collagenase-1 (MMP-1) expression and invasion of SCC cells via p38 mitogen-activated protein kinase. Here, we have examined the role of Smad signaling in regulating MMP-13 expression and in invasion of head and neck SCC cells. Treatment with TGF-β resulted in activation of Smad2 and Smad3 in SCC cells, but had no effect on their proliferation or viability. Basal activation of Smad3 and p38 was noted in SCC cells without exogenous TGF-β stimulation, and adenoviral delivery of Smad7 and dominant-negative Smad3 inhibited p38 activation in these cells. Adenoviral overexpression of Smad3 augmented the upregulatory effect of TGF-β on MMP-13 expression by SCC cells. Disruption of Smad signaling by adenoviral expression of kinase-defective TGF-β type I receptor (activin-receptor-like kinase-5), Smad7, and dominant-negative Smad3 potently suppressed the basal and TGF-β-induced expression of MMP-13 and MMP-1 in SCC cells, and inhibited their basal and TGF-β-induced invasion through Matrigel and type I collagen. Adenoviral overexpression of Smad7 in cutaneous and oral SCC cells significantly inhibited their implantation in skin of SCID mice and growth of xenografts in vivo, as compared to LacZ adenovirus-transduced control cells. Together, these results show that Smad signaling plays an important role in promoting the invasive phenotype of human head and neck SCC cells by upregulating their collagenase expression.
Tumor invasion and metastasis are critical steps in the progression of malignant tumors. Invasion of tumor cells involves degradation of the basement membrane and the stromal extracellular matrix (ECM), which results in tumor infiltration into adjacent tissue. Matrix metalloproteinases (MMPs) play a central role in tumor invasion due to their ability to degrade many ECM components and other substrates, such as growth factors, and chemokines (Folgueras et al., 2004). The human MMP gene family consists of 23 members, which can be divided into subgroups of collagenases, gelatinases, stromelysins, matrilysins, membrane-type MMPs, and other MMPs according to their substrate specificity and function (Visse and Nagase, 2003; Folgueras et al., 2004). Collagenases (MMP-1, -8, and -13) are the principal proteinases capable of cleaving native fibrillar collagens (Ala-aho and Kähäri, 2005). Collagenase-3 (MMP-13) has a notably wide substrate specificity, and its expression is limited to physiologic situations, in which rapid and effective remodeling of collagenous ECM is required, for example fetal development of bone, postnatal bone remodeling, and gingival and fetal skin wound repair (Johansson et al., 1997b; Ståhle-Bäckdahl et al., 1997; Ravanti et al., 1999, 2001). However, the wide substrate specifity of MMP-13 suggests that it is also a powerful invasion tool for cancer cells. MMP-13 is one of the few MMPs primarily expressed by tumor cells in malignant tumors, such as breast carcinomas (Freije et al., 1994), squamous cell carcinomas (SCCs) of the head and neck (Airola et al., 1997; Johansson et al., 1997a), SCCs of the vulva (Johansson et al., 1999), and primary and metastatic melanomas (Airola et al., 1999; Nikkola et al., 2001). In SCCs of the head and neck and vulva, MMP-13 is expressed primarily by cancer cells at the invading edge of the tumor and the expression correlates with poor prognosis (Johansson et al., 1997a, 1999; Cazorla et al., 1998).
Transforming growth factor-β (TGF-β) enhances MMP-13 expression and promotes invasion of head and neck SCC cells via p38 mitogen-activated protein kinase (MAPK) pathway (Johansson et al., 2000). TGF-β is a multifunctional growth factor, which exerts distinct effects on cell growth and proliferation and on ECM deposition (Moustakas et al., 2001). In general, TGF-β stimulates proliferation of fibroblastic cells, but acts as a potent growth inhibitor for cells of epithelial origin (Moustakas et al., 2001). Binding of TGF-β to its specific cell surface receptors initiates TGF-β signaling, when TGF-β receptor type II kinase phosphorylates the type I TGF-β receptor kinase (TβRI). Subsequently, TβRI, also termed as activin-receptor-like kinase-5 (ALK5), phosphorylates Smad2 and Smad3, which then associate with Smad4. Thereafter, Smads translocate into the nucleus, where they bind directly to specific DNA sequences in the regulatory regions of genes or associate with transcription coactivators or corepressors, and this way regulate the transcription of TGF-β-responsive genes, such as plasminogen-activator inhibitor (PAI)-1 (Massague and Wotton, 2000; Moustakas et al., 2001). Smad7 serves as a negative regulator of Smad signaling by interfering with the activity of ALK5 kinase, and this way specifically inhibiting the phosphorylation of Smad2 and Smad3 (Nakao et al., 1997).
In this study, we have elucidated the role of Smad signaling in regulating the invasive phenotype of human head and neck SCC cells. Our results show that TGF-β activates Smad signaling in SCC cells, but has no effect on proliferation and viability of these cells. Smad signaling regulates the basal and TGF-β-induced MMP-13 expression and invasion of these cells. Furthermore, inhibition of Smad pathway by Smad7 suppresses the implantation of human cutaneous SCC cells in skin of SCID mice and results in delayed growth of the xenografts. Together, these results provide evidence for the crucial role of Smad signaling in promoting the invasive phenotype of human SCC cells by upregulating their MMP-13 expression, and suggest Smad signaling as a novel target for anti-invasive therapy of head and neck SCCs.
Smad3 is constitutively activated in human head and neck SCC cells
Smad-signaling cascade is activated upon TGF-β stimulation in various types of cells (Moustakas et al., 2001; Siegel and Massagué, 2003). First, we characterized the activation of the Smad pathway in human SCC cells. As shown in Figure 1a, TGF-β stimulation of metastatic cutaneous SCC cells (UT-SCC-7) resulted in sustained phosphorylation of Smad2, first noted at 30 min of incubation. Interestingly, phoshorylation of Smad3 was detected already prior to TGF-β stimulation, and it was further enhanced after 30 and 60 min TGF-β treatment (Figure 1a). Basal phosphorylation of Smad3 was also detected in four other SCC cell lines (UT-SCC-2, -8, -12A, and -15), and the phosphorylation was further enhanced by TGF-β (data not shown). In some SCC cell lines also Smad2 was basally phosphorylated (not shown). In accordance with previous observations (Johansson et al., 2000), phosphorylation of p38 MAPK was noted in the absence of TGF-β (Figure 1a). No basal activation of Smad2 or Smad3 was detected in normal epidermal keratinocytes, although TGF-β treatment induced activation of Smad2 and Smad3 (Figure 1b).
Next we examined the nuclear translocation of adenovirally produced HA-tagged Smad2, Smad3, and Smad4 by immunocytochemistry. We have previously shown that these adenoviruses produce functional Smad proteins, which remain in the cytoplasm of infected human fibroblasts and are translocated to the nucleus following TGF-β stimulation (Leivonen et al., 2002). UT-SCC-7 cells were infected with the HA-tagged Smad adenoviruses, incubated with or without TGF-β for 1.5 h, and stained with HA-antibody. In untreated control cells, Smad2 was located predominantly in the cytoplasm, whereas in TGF-β-treated SCC cells, it was translocated into the nucleus (Figure 1c). In comparison, in untreated control cells infected with Smad3 or Smad4 adenovirus, HA-staining was also detected in the nucleus, and after TGF-β stimulation, Smad3 and Smad4 were located predominantly in the nucleus (Figure 1c). Specifically, approximately 30–40% of the cells expressing HA-Smad3 or HA-Smad4 showed nuclear HA-staining without TGF-β stimulation (Figure 1d).
Activation of p38 MAPK in SCC cells is Smad dependent
MAPK- and Smad-signaling pathways cooperate in mediating the cellular effects of TGF-β, and MAPKs can activate Smad3 (Brown et al., 1999; Dennler et al., 2000). As the results above show basal activation of p38 and Smad3 in SCC cells, we examined whether activation of Smad3 is dependent on p38 MAPK activity. UT-SCC-7 cells were serum starved for 18 h, and subsequently treated with SB203580 (10 μ M), a specific chemical inhibitor for p38 activity, for different periods of time. Consistent with the results above, basal activation of Smad3 was detected in UT-SCC-7 cells (Figure 2a). However, inhibition of p38 MAPK activity had no effect on the levels of phosphorylated Smad3 in these cells (Figure 2a), or in UT-SCC-2 cells (not shown). In comparison, the levels of phosphorylated CREB, a downstream substrate of p38 MAPK, were markedly reduced, initially after a 2-h SB203580 treatment, indicating that p38 activity was inhibited under these conditions (Figure 2a).
To examine, whether activation of p38 MAPK in human SCC cells is dependent on Smad signaling, we infected UT-SCC-7 cells with adenovirus RAdSmad7 to obtain overexpression of Smad7, which prevents phosphorylation of Smad2 and Smad3 by TβRI (ALK5) (Nakao et al., 1997). Control cultures were infected with empty adenovirus RAdpCA3. The cultures were then treated with TGF-β for different periods of time. Interestingly, overexpression of Smad7 prevented the basal activation of p38, Smad2, and Smad3, and also potently reduced their TGF-β-induced phosphorylation (Figure 2b). In further experiments, adenoviral expression of dominant-negative Smad3 (Smad3DN) inhibited the TGF-β-induced Smad3 and p38 activation, but had only a slight effect on TGF-β-elicited induction of Smad2 phosphorylation (Figure 2c). This implies that both basal and TGF-β-induced p38 MAPK activation in SCC cells is Smad dependent.
Head and neck SCC cells produce and activate TGF-β
Many types of malignant cells produce TGF-β, which can promote tumor growth in an autocrine or a paracrine manner (Siegel and Massagué, 2003). Therefore, we examined whether head and neck SCC cells produce TGF-β, which could lead to autocrine activation of TGF-β signaling. UT-SCC-2, -7, -8, -12A, and -15 cells and HaCaT keratinocytes were incubated in serum-free growth media for 24 h. Thereafter, conditioned media were collected and TGF-β activity was assayed using reporter cells that produce luciferase activity in response to TGF-β. As shown in Figure 3a, UT-SCC-2, -12A, and -15 cells produced relatively high levels of TGF-β, whereas UT-SCC-7, UT-SCC-8, and HaCaT cells produced relatively low amounts of TGF-β. Interestingly, the majority of TGF-β secreted by UT-SCC-2 cells was in active form (Figure 3a). Coculture experiments confirmed that UT-SCC-2 cells were able to activate TGF-β (Figure 3b). To analyse the production of different TGF-β isoforms by SCC cells, UT-SCC-2, -12A, and -15 cell-conditioned media were incubated with the reporter cells in the presence of specific neutralizing antibodies for TGF-β1, -β2, and -β3. As shown in Figure 3c, the major isoform produced by UT-SCC-2 cells was TGF-β1, whereas UT-SCC-12A and UT-SCC-15 secreted mainly TGF-β2 and TGF-β3.
Expression of MMP-13 by head and neck SCC cells is regulated by Smad signaling
We have previously demonstrated that head and neck SCC cells express MMP-13, the expression of which is enhanced by TGF-β via p38 MAPK pathway (Johansson et al., 1997a, 2000). In addition, we have observed that Smad3 mediates the effects of TGF-β on MMP-13 expression in human gingival fibroblasts (Leivonen et al., 2002). In this context, we investigated the role of Smad signaling in regulating the expression of MMP-13 in SCC cells. We utilized adenoviral gene delivery of Smad2, Smad3, Smad4, and Smad3DN, and determined the expression of MMP-13 by Northern blot hybridizations. A 24-h TGF-β treatment of cells infected with control virus RAdpCA3 resulted in 3.2-fold enhancement of MMP-13 mRNA levels, as compared to corresponding untreated cells (Figure 4a). Overexpression of Smad3 augmented the TGF-β-enhanced levels of MMP-13 mRNAs, whereas Smad2 had no marked effect on the expression of MMP-13 mRNAs (Figure 4a). Interestingly, Smad3DN suppressed the effect of TGF-β on MMP-13 mRNA levels, and reduced the basal expression of MMP-13 mRNAs by 70% (Figure 4a).
Overexpression of Smad3 alone or in combination with Smad4 also enhanced the basal expression of MMP-1 mRNA (Figure 4a). Treatment of cells with TGF-β enhanced the levels of MMP-1 mRNAs, and this effect was further enhanced by Smad3. In addition, expression of Smad3DN suppressed the effect of TGF-β on MMP-1 expression by 60%. TGF-β also enhanced the mRNA levels for PAI-1, and overexpression of Smad3 and Smad4 augmented the stimulatory effect of TGF-β, whereas Smad3DN abrogated the effect of TGF-β on PAI-1 mRNA levels (Figure 4a). We also analyzed four other head and neck SCC cell lines, and noted that Smad3DN and inhibitory Smad7 suppressed proMMP-13 production in all SCC cell lines studied, and proMMP-1 production in UT-SCC-2 and UT-SCC-12A cells (data not shown).
To examine, whether ALK5 kinase activity is required for MMP-13 and MMP-1 expression in SCC cells, UT-SCC-7 cells were infected with adenoviruses for a constitutively active mutant for ALK5 (RAdCA-ALK5) and a kinase-defective mutant for ALK5 (RAdKD-ALK5). Expression of CA-ALK5 resulted in enhanced proMMP-13 production, as compared to control virus RadLacZ-infected cells (Figure 4b). In contrast, expression of KD-ALK5 markedly suppressed both basal and TGF-β-induced production of proMMP-13 and pro-MMP-1, as potently as Smad3DN (Figure 4b). The expression of KD-ALK5 also abrogated basal phosphorylation of Smad3 in UT-SCC-7 cells (data not shown). This indicates, that activity of TβRI is required for basal production of MMP-13 and also of MMP-1 by SCC cells.
Invasion of head and neck SCC cells is dependent on Smad signaling
MMP-13 potently degrades collagens and components of the basement membranes (Knäuper et al., 1996), and expression of MMP-13 promotes invasion of malignant cells (Ala-aho et al., 2002b). SCCs of the head and neck are characterized by high invasion capacity, and inhibition of MMP-13 expression suppresses invasion of SCC cells through reconstituted basement membrane, Matrigel (Ala-aho et al., 2004). Therefore, we analyzed the effect of KD-ALK5, Smad3DN, and Smad7 on the invasion capacity of SCC cells. As shown in Figure 5a, basal invasion of UT-SCC-7 cells through type I collagen was potently (by 60–80%) inhibited by expression of KD-ALK5, Smad3DN, and Smad7. TGF-β treatment enhanced the invasion of UT-SCC-7 cells through collagen 2.5-fold, and expression of KD-ALK5, Smad3DN, and Smad7 potently inhibited this effect. Similarly, KD-ALK5, Smad3DN, and Smad7 potently (by 70–80%) inhibited basal invasion of UT-SCC-7 cells through Matrigel (Figure 5b). TGF-β also enhanced invasion of UT-SCC-7 cells through Matrigel by 1.7-fold, and adenoviral expression of KD-ALK5, Smad3DN, and Smad7 inhibited the invasion of cells under these conditions (Figure 5b).
To confirm the role of MMP-13 in TGF-β-mediated SCC cell invasion, we utilized adenoviral gene delivery of a previously characterized antisense ribozyme targeted against MMP-13 (MMP-13AsRZ) (Ala-aho et al., 2004). Interestingly, inhibition of MMP-13 expression by MMP-13AsRZ abrogated the invasion of SCC cells through type I collagen and Matrigel, and also reduced the ability of TGF-β to enhance invasion (Figure 5c and d). In addition, MMP-13 activity in the conditioned media obtained from the upper invasion chambers correlated with the invasion of SCC cells (Figure 5e and f). In contrast, the secretion of 72-kDa gelatinase (MMP-2) and 92-kDa gelatinase (MMP-9), determined by gelatinase zymography, was not affected by KD-ALK5, Smad7, or Smad3DN (not shown).
TGF-β has no effect on proliferation of SCC cells
TGF-β exerts a growth inhibitory effect on normal epithelial cells and this way serves as a tumor suppressor (Siegel and Massagué, 2003). In a cell-type-dependent manner, TGF-β upregulates the transcription of the cyclin-dependent kinase (CDK) inhibitors p21(Cip1/Waf1) and p15(Ink4b), which inhibit CDK phosphorylation of the Rb protein and thus halt cell-cycle progression in G1-phase (Datto et al., 1995; Li et al., 1995). Typically, many types of malignant cells escape the growth inhibitory effects of TGF-β (Akhurst and Derynck, 2001). In this context, we examined the ability of TGF-β to induce the expression of CDK inhibitors in head and neck SCC cells. As shown in Figure 6a, TGF-β stimulation had no effect on the expression of p15, p21, and p27 by the five SCC cell lines studied. As expected, in normal keratinocytes, p15 and p21 protein levels were upregulated after 24-h TGF-β stimulation (Figure 6b). In addition, TGF-β treatment resulted in marked increase in p21 expression in HaCaT cells after a 6- and 24-h incubation (Figure 6b).
To specifically examine the effect of TGF-β on the proliferation of head and neck SCC cells and HaCaT keratinocytes, we determined their DNA synthesis by BrdU incorporation assay at different time points after TGF-β stimulation. As shown in Figure 6c, TGF-β had no marked effect on the proliferation of UT-SCC-2 cells. In comparison, TGF-β potently (by 50%) inhibited the proliferation of HaCaT keratinocytes (Figure 6c). TGF-β had no effect on the number of viable SCC cells, as determined by MTT assay (data not shown). Furthermore, adenoviral expression of Smad3, Smad7, and Smad3DN had no effect on the proliferation or viability of SCC cells during a 96-h incubation period (data not shown). These observations show that although SCC cells are responsive to TGF-β, neither activation nor inhibition of their TGF-β signaling has an effect on their proliferation or viability.
Inhibition of Smad signaling suppresses implantation and growth of SCCs in vivo
Next we examined the role of Smad signaling in the implantation and subsequent growth of SCCs in vivo. In two independent experiments, cutaneous SCC cells (UT-SCC-7) and SCC cells from the oral cavity (UT-SCC-2) were infected with RAdSmad7 and with control virus RAdLacZ at MOI 700 for 6 h, and further incubated for 24 h to obtain overexpression of Smad7. Thereafter, the infected cells (5 × 106) in 100 μl of PBS were injected subcutaneously in the back of SCID/SCID mice. Interestingly, SCC cell implantation and subsequent growth of cutaneous xenografts in vivo in SCID mice were significantly delayed by adenoviral overexpression of Smad7 in UT-SCC-7 cells, as compared to LacZ adenovirus-transduced control cells (Figure 7a, left panel). Interestingly, the results of another experiment performed with oral SCC cell line, UT-SCC-2, showed that adenoviral expression of Smad7 entirely inhibited implantation and growth of these cells in SCID mice (Figure 7a, right panel).
The Ki67-positive proliferating cells were detected near the tumor margins in xenografts established with both RAdSmad7-infected UT-SCC-7 cells and RadLAcZ-infected control cells (Figure 7b). However, the number of Ki67-positive cells was markedly lower in RAdSmad7 xenografts than in RAdLacZ xenografts (Figure 7b). Quantitation of the relative number of Ki67-positive cells revealed that the number of proliferating cells was 80% lower in RAdSmad7-infected SCC xenografts than in RAdLacZ control SCC xenografts, and the relative number of Ki67-positive cells correlated with tumor size (Figure 7c).
MMPs are a family of proteinases that play an important role in different steps of tumor invasion and metastasis. Proteolysis of the ECM is the first step in the invasion of tumor cells, which includes crossing the basement membrane and invading the surrounding stroma consisting primarily of fibrillar collagens (Stetler-Stevenson and Yu, 2001; Ala-aho and Kähäri, 2005). MMP-13 is a collagenolytic MMP characterized by a wide substrate specificity and restricted physiological expression in vivo, and it is expressed by many types of invasive tumors, including SCCs of the head and neck (Ala-aho and Kähäri, 2005). The expression of MMP-13 correlates with the invasion capacity of malignant cells, including head and neck SCC cells, in which MMP-13 expression is enhanced by TGF-β via p38 MAPK pathway (Johansson et al., 2000; Ala-aho et al., 2004). In the present study, we have elucidated the role of Smads in the regulation of MMP-13 expression and invasion of head and neck SCC cells. Our results show that Smad signaling is activated in SCC cells by TGF-β, and specifically regulates the basal and TGF-β-induced expression of MMP-13 and MMP-1, and invasion of SCC cells. Our results also show that TGF-β has no effect on proliferation or viability of SCC cells, indicating that these cells have escaped the growth inhibitory effect of TGF-β. Furthermore, we demonstrate that inhibition of Smad signaling by Smad7 inhibits the implantation and subsequent growth of human cutaneous SCC xenografts in SCID/SCID mice.
An interesting observation in this study was that Smad3 is constitutively activated in head and neck SCC cells. This basal phosphorylation of Smad3 was not dependent on p38 MAPK activity, since inhibition of p38 activity by a chemical inhibitor, SB203580, had no effect on the basal levels of phosphorylated Smad3. Interestingly, phosphorylation of p38 MAPK was dependent on Smad signaling, which is supported by reports showing that TGF-β activates p38 through both Smad-independent pathway via TGF-β-activated kinase 1 (Yamaguchi et al., 1995; Hanafusa et al., 1999; Sano et al., 1999) and through Smad-dependent pathways (Takekawa et al., 2002; Ungefroren et al., 2003). Most SCC cell lines examined produced and activated TGF-β, which could be responsible for autocrine activation of TGF-β-receptor signaling. However, mutations in TGF-β-signaling components are frequent in various cancers, so it is possible that basal Smad3 activation is due to a mutation in for example TGF-β receptor ALK5.
Our recent observations demonstrate that in human gingival fibroblasts, Smad3 mediates the TGF-β-elicited expression of MMP-13 in cooperation with p38 MAPK (Leivonen et al., 2002). Here, adenoviral expression of Smad7 and Smad3DN resulted in suppression of basal MMP-13 expression, indicating that Smad signaling specifically upregulates the expression of MMP-13 in SCC cells. Smad-signaling pathway was also essential for the invasion capacity of SCC cells, since Smad7 and Smad3DN potently inhibited their invasion through Matrigel and type I collagen. Furthermore, inhibition of TβRI function by kinase-defective ALK5 potently suppressed the expression of MMP-13 and invasion of SCC cells, indicating that TGF-β type I receptor activity plays an important role in promoting the invasion capacity of these cells. These results are in accordance with our previous studies showing that MMP-13 production of SCC cells specifically correlates with their invasion capacity, and that inhibition of MMP-13 expression by p38 inhibition, p53, interferon-γ, or by specific ribozyme potently suppresses their invasion (Ala-aho et al., 2000, 2002a, 2004; Johansson et al. 2000).
TGF-β has been implicated in carcinogenesis in a number of studies (Derynck et al., 2001; Wakefield and Roberts, 2002). As a multifunctional growth factor, TGF-β has both tumor suppressor and tumor-promoting activities, depending on the stage of carcinogenesis and the differentiation of the tumor cell. Normal epithelial cells and transformed epithelial cells in early tumor stage are usually sensitive to TGF-β-elicited growth inhibition (Akhurst and Derynck, 2001; Moustakas et al., 2001; Siegel and Massagué, 2003). However, during malignant tumor progression, epithelium-derived tumor cells often escape TGF-β-induced growth control, and once this has occurred, TGF-β promotes tumor progression resulting in increased tumor cell invasion and metastasis (Akhurst and Derynck, 2001). This is often associated with oncogenic Ras-signaling (Janda et al., 2002; Oft et al., 2002). In the present study, TGF-β had no effect on proliferation of head and neck SCC cells in culture or on the levels of cyclin-dependent kinase inhibitors p15(Ink4b), p21(Cip1/Waf1), and p27(Kip1). In this context, it should also be noted that adenoviral expression of p53 in these SCC cell lines results in induction in p21 expression, indicating that p21 gene in these cells is functional (Ala-aho et al., 2002a). In contrast, the levels of p15 and p21 were markedly increased in normal human keratinocytes upon TGF-β stimulation. These results are in accordance with previous reports showing that keratinocytes are responsive to the growth inhibitory effects of TGF-β (Reynisdottir et al., 1995; Hu et al., 1998; Pardali et al., 2000). Altogether, our results indicate that although TGF-β promotes invasion of head and neck SCC cells via activation of Smad signaling, these cells are resistant to growth inhibitory effects of TGF-β. These results are supported by previous observations demonstrating constitutive Smad activation in melanoma cells not responsive to growth inhibitory effect of TGF-β (Rodeck et al., 1999).
The results of our experiments using an in vivo model of cancer cell invasion showed that cutaneous SCC cell implantation and subsequent growth of xenografts in vivo in SCID mice was significantly suppressed by adenoviral overexpression of Smad7 in SCC cells, as compared to LacZ adenovirus-transduced control cells. Interestingly, another experiment performed with an oral SCC cell line (UT-SCC-2) showed that adenoviral expression of Smad7 entirely inhibited implantation and growth of these cells in SCID mice. It is likely that the inhibition of growth of tumors established by Smad7-transduced cells was associated with reduced MMP-13 expression, since we have previously noted that specific inhibition of MMP-13 expression inhibits the implantation and growth of cutaneous SCC xenografts (Ala-aho et al., 2004). However, as inhibition of Smad signaling also suppressed MMP-1 production, it is possible that both these proteinases contribute together to implantation and subsequent growth of cutaneous SCC xenografts. Moreover, the amount of Ki67-positive cells was lower in tumors derived from Smad7-transduced UT-SCC-7 cells, indicating that inhibition of tumor growth was associated with reduced proliferation of tumor cells. As TGF-β and Smad signaling have no effect on proliferation of SCC cells, it is likely that the effect of Smad7 is mediated via suppression of MMP-13 production, which results in impaired ability of SCC cells to invade the dermal tissue and initiate tumor growth.
Our observations are supported by recent reports implicating Smad cascade both in tumor suppressive and tumor-promoting effects of TGF-β. Inhibition of Smad signaling by stable overexpression of Smad7 inhibits the growth of human melanoma xenografts in nude mice, and is associated with reduced MMP production (Javelaud et al., 2005). Furthermore, Smad3 overexpression has prometastatic effects on malignant breast cancer cells (Tian et al., 2003), and a Smad-binding defective mutant of TβRI enhances their tumorigenesis (Tian et al., 2004). Studies with Smad3 knock-out mice demonstrated that these mice are resistant to chemical cutaneous carcinogenesis due to inhibition of keratinocyte proliferation and reduced skin inflammation (Li et al., 2004). In addition, a recent study provides evidence for the role of TGF-β signaling in promoting the growth and invasion of glioma cells (Uhl et al., 2004). Our results imply that Smad3 mediates the tumor-promoting activities of TGF-β on head and neck SCC cells by regulating the expression of potent matrix-degrading proteinases, MMP-13 and MMP-1, and thereby promoting the collagenolytic and invasive capacity of these malignant cells. However, the role of Smad2 cannot be excluded, since there is also evidence for the specific role of Smad2 in mediating metastasis. Oft et al. (2002) have demonstrated that activities of Smad2 and Ras induce invasiveness of tumor cells resulting in progression from a differentiated squamous carcinoma to a motile invasive stage.
In conclusion, we demonstrate that human head and neck SCC cells are responsive to proinvasive effect of TGF-β, although they are refractory to the growth inhibitory effects of TGF-β. Our results also show that TGF-β promotes the invasion of these cells by upregulating the expression of MMP-13 via Smad pathway, and that inhibition of Smad signaling and subsequent inhibition of MMP-13 expression inhibits the first step in metastasis, that is implantation of cutaneous SCC cells, resulting in marked inhibition in growth of the tumors. Together, these results provide evidence for the crucial role of Smad signaling in promoting the invasive phenotype of human SCC cells by upregulating their MMP-13 expression, suggesting ALK5-mediated Smad signaling as a novel and specific target for anti-invasive therapy of head and neck SCCs.
Materials and methods
Cell cultures and reagents
Human SCC cell lines were established from primary SCCs of the oral cavity (UT-SCC-2), larynx (UT-SCC-8), skin (UT-SCC-12A), and tongue (UT-SCC-15), and from metastasis of cutaneous SCC (UT-SCC-7) (Lansdorf et al., 1999). HaCaT cells, a spontaneusly immortalized nontumorigenic adult human epidermal keratinocyte cell line (Boukamp et al., 1990), were kindly provided by Dr Norbert Fusenig (German Cancer Research Center, Heidelberg, Germany). The cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM; Sigma, St Louis, MO, USA) supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine, 100 IU/ml penicillin-G, 100 μg/ml streptomycin, and non-essential amino acids. Normal human epidermal keratinocytes were established from skin samples as previously described (Boyce and Ham, 1983). Keratinocytes were maintained in a serum-free basal keratinocyte growth medium (KBM-2) (Cambrex, NJ, USA). Human recombinant TGF-β1 was purchased from Sigma (St Louis, MO, USA), and p38 MAPK inhibitor SB203580 from Calbiochem (San Diego, CA, USA).
Infection of cells with recombinant adenoviruses
The construction of empty control virus RAdpCA3 and adenoviruses for HA-tagged Smad2, Smad3, and Smad4 (RAdSmad2, RAdSmad3, and RAdSmad4, respectively) has been described before (Leivonen et al., 2002). Adenovirus for antisense ribozyme against MMP-13 (RAdMMP-13AsRZ) has been described before (Ala-aho et al., 2004). Adenoviruses for Smad3DN (RAdSmad3DN) (Pardali et al., 2000), Smad7 (RAdSmad7) (Fujii et al., 1999), constitutively active ALK5 (RAdCA-ALK5) (Fujii et al., 1999), and kinase-defective ALK5 (RAdKD-ALK5) (Fujii et al., 1999) were kindly provided by Dr Aristidis Moustakas (Ludwig Institute for Cancer Research, Uppsala, Sweden). Recombinant adenovirus RAdLacZ, which contains the Escherichia coli β-galactosidase gene under the control of the CMV IE promoter, and empty control virus RAd66 were kindly provided by Dr Gavin WG Wilkinson (University of Cardiff, UK) (Wilkinson and Akrigg, 1992).
Head and neck SCC cells were infected as described previously (Ala-aho et al., 2002a) with appropriate MOI (from 400 to 2000) to obtain nearly 100% transduction efficiency, and incubated for 6 h in DMEM containing 0.5% FCS. Thereafter, the medium was changed, and incubations continued for 24 or 48 h.
Immunoblotting and antibodies
Immunoblottings were performed as described previously (Leivonen et al., 2002). Polyclonal Smad2 and Smad3 antibodies were from Zymed Laboratories Inc. (San Francisco, CA, USA), mouse monoclonal anti-human MMP-13 antibody (181–15A12) from Calbiochem (San Diego, CA, USA), and polyclonal anti-TIMP-1 from Chemicon International Inc. (Temecula, CA, USA). Polyclonal rabbit antiserum against human MMP-1 was a kind gift from Dr Henning Birkedal-Hansen (NIDCR, National Institutes of Health, Bethesda, MD, USA). Antiserums against phospho-Smad2 (PS2) and phospho-Smad1 (PS1), which shows crossreactivity with phosphorylated Smad3 (Piek et al., 1999; Dooley et al., 2001), were kind gifts from Dr Aristidis Moustakas (Ludwig Institute for Cancer Research, Uppsala, Sweden). Polyclonal antibodies for phospho-p38, p38, phospho-CREB, and CREB were from Cell Signaling Technology (Beverly, MA, USA), and monoclonal antibodies for p15(Ink4B), p21(Waf1/Cip1), and p27(Kip1) from Pharmingen (San Diego, CA, USA). The blots were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech, UK).
To examine the nuclear translocation of HA-Smads produced by adenoviruses, SCC cells growing on coverslips were infected with RAdSmad2, RAdSmad3, and RAdSmad4, and with control virus RAd66. Thereafter, the cells were incubated in DMEM without FCS for 24 h, treated with TGF-β1 (5 ng/ml) for 1.5 h, fixed with methanol at −20°C for 6 min, and stained with rat monoclonal anti-HA antibody (clone 3F10, Roche, Mannheim, Germany) using rhodamine-conjugated anti-rat secondary antibody (Calbiochem). Nuclei of the same cells were counterstained with Hoechst 33342.
TGF-β activity assay
Mink lung epithelial cells stably transfected with a fragment of PAI-1 promoter fused to luciferase gene were kindly provided by Dr DB Rifkin (New York University School of Medicine, New York, NY, USA). SCC cells and HaCaT cells were maintained in serum-free growth media for 24 h. Thereafter, the conditioned media were analyzed for the secretion of TGF-β by using the above-mentioned TGF-β-responsive reporter cells that produce luciferase activity in response to TGF-β (Abe et al., 1994). For dissecting the secretion of different TGF-β isoforms, SCC cell-conditioned media were incubated with the reporter cells in the presence of specific neutralizing antibodies for TGF-β1, -β2, and -β3. The amount of active TGF-β was analyzed directly from conditioned media, and the amount of total TGF-β, from heat-treated conditioned media. In coculture assays, SCC cells were cultured with the reporter cells for 24 h, after which the conditioned media were analyzed for luciferase activity.
Northern blot hybridizations
Northern blot hybridizations were performed as described previously (Leivonen et al., 2002). For hybridizations, fragments covering the coding region and part of the 3′-untranslated region of human MMP-13 cDNA (altogether 1.9-kb) (Johansson et al., 1997c), a 2.0-kb human MMP-1 cDNA (Goldberg et al., 1986), human PAI-1 cDNA (Keski-Oja et al., 1988), and a 1.3-kb rat GAPDH cDNA (Fort et al., 1985) were used. Specific hybridization was visualized with autoradiography and quantitated by densitometric scanning.
Cell culture inserts with 0.8 μm pore size (Falcon 3097, Beckton Dickinson) were precoated either with 25 μg Matrigel (Becton Dickinson) or with 1 mm thick type I collagen gel (1 mg/ml) bovine dermal collagen (Cellon Bovine Dermal Collagen, Cellon, Strassbourg, France, as described previously) (Ala-aho et al., 2002b). For invasion assays, SCC cells were infected with adenoviruses and incubated for 18 h with or without TGF-β (5 ng/ml). Thereafter, the cells (2 × 105/chamber) suspended in DMEM containing 0.1% BSA were seeded on top of the gel in the upper chamber in a final volume of 200 μl, and TGF-β (5 ng/ml) was added. DMEM containing 10% FCS was used as a chemoattractant in the lower chambers. After 24 h incubation, cells in the upper chambers were removed with a cotton bud, and the invaded cells on the lower surface were fixed with 2% paraformaldehyde in PBS for 10 min, stained with 0.1% crystal violet, and counted under light microscope. Statistical significance was determined by Student's t-test.
Analysis of MMP-13 activity
The MMP-13 activity assays were performed on Delfia® anti-mouse IgG Microtitration strips (Wallac, Perkin Elmer Life and Analytical Sciences, Turku, Finland). Mouse monoclonal anti-MMP-13 (Ab-1, Calbiochem, San Diego, CA, USA) was diluted in Delfia® Assay buffer (1.5 ng/μl) and allowed to bind to the wells overnight at +4°C. Thereafter, the wells were washed twice with reaction buffer (50 mM Tris, pH 7.5; 0.2 mM NaCl; 10 mM CaCl2; 0.05% Brij-35). SCC cell-conditioned media were added to the wells and incubated for 1 h at RT, after which the wells were washed three times with the reaction buffer. To activate MMP-13, APMA (1 mM) was added and the plates were incubated for 30 min at +37°C. Europium-labeled MMP-13 peptide substrate (KKGC(Eu)GPLALYG-Dabcyl-Hex) diluted in water (4 μ M) was added to the reaction and fluorescence was measured with VICTOR multilabel counter (Perkin Elmer Life and Analytical Sciences, Wallac, Turku, Finland) at different time points.
Determination of cell viability and proliferation
To determine the effect of TGF-β on SCC cell viability and proliferation, cells (4 × 103) were seeded on 96-well plates and incubated for 24 h in DMEM containing 10% FCS. Thereafter, the medium was replaced with DMEM containing 0.5% FCS, and incubations continued for additional 18 h. Subsequently, TGF-β (5 ng/ml) was added and cells were incubated for different periods of time, as indicated. The cell viability was determined by CellTiter 96® Assay (Promega, Madison, WI, USA) according to the manufacturer's instructions. Cell proliferation was determined by BrdU (5-bromo-2′-deoxyuridine) Cell Proliferation Assay ELISA (Roche, Mannheim, Germany).
Growth of human SCC xenografts in SCID/SCID mice
All experiments with mice were performed according to institutional guidelines and with permission of the animal test review board of the University of Turku, Finland. Severe combined immunodeficiency (SCID/SCID) male mice (6 weeks old) were used. UT-SCC-7 and UT-SCC-2 cells were infected with RAdLacZ and RAdSmad7 as described above at MOI 700, and incubated for 6 h. Thereafter, fresh medium was added and the cells incubated for 18 h. The cells were detached with trypsin, counted, and 5 × 106 cells in 100 μl of PBS were injected subcutaneously in the back of SCID mice. Each experimental group contained five to six mice. Tumor size was measured once a week, and tumor volume was calculated as width2 × length × (pi/6).
Tumors were fixed overnight in phosphate-buffered 10% formalin and embedded in paraffin. Serial sections (5 μm) were deparaffinized and processed for immunohistochemistry with citrate buffer in a microwave oven. Proliferating cells were detected by Ki67 immunostaining using monoclonal antibody for human Ki67 (MIB-1: DAKO, Denmark). Negative controls were incubated without the primary antibody. Mayer's hematoxylin was used as counterstain. The area of Ki67-positive cells was determined in four distinct fields at × 4 maginification from all sections using Soft Imaging System's analySIS® program.
Abe M, Harpel JG, Metz CN, Nunes I, Loskutoff DJ, Rifkin DB . (1994). Anal Biochem 216: 276–284.
Airola K, Johansson N, Kariniemi AL, Kähäri V-M, Saarialho-Kere UK . (1997). J Invest Dermatol 109: 225–231.
Airola K, Karonen T, Vaalamo M, Lehti K, Lohi J, Kariniemi AL et al. (1999). Br J Cancer 80: 733–743.
Akhurst RJ, Derynck R . (2001). Trends Cell Biol 11: S44–S51.
Ala-aho R, Ahonen M, George SJ, Heikkilä J, Grénman R, Kallajoki M et al. (2004). Oncogene 23: 5111–5123.
Ala-aho R, Grénman R, Seth P, Kähäri V-M . (2002a). Oncogene 21: 1187–1195.
Ala-aho R, Johansson N, Baker AH, Kähäri V-M . (2002b). Int J Cancer 97: 283–289.
Ala-aho R, Johansson N, Grénman R, Fusenig NE, López-Otín C, Kähäri V-M . (2000). Oncogene 19: 248–257.
Ala-aho R, Kähäri V-M . (2005). Biochimie 87: 273–286.
Boukamp P, Stanbridge EJ, Foo DY, Cerutti PA, Fusenig NE . (1990). Cancer Res 50: 2840–2847.
Boyce ST, Ham RG . (1983). J Invest Dermatol 81: 33s–40s.
Brown JD, DiChiara MR, Anderson KR, Gimbrone Jr MA, Topper JN . (1999). J Biol Chem 274: 8797–8805.
Cazorla M, Hernandez L, Nadal A, Balbin M, Lopez JM, Vizoso F et al. (1998). J Pathol 186: 144–150.
Datto MB, Yu Y, Wang XF . (1995). J Biol Chem 270: 28623–28628.
Dennler S, Prunier C, Ferrand N, Gauthier JM, Atfi A . (2000). J Biol Chem 275: 28858–28865.
Derynck R, Akhurst RJ, Balmain A . (2001). Nat Genet 29: 117–129.
Dooley S, Delvoux B, Streckert M, Bonzel L, Stopa M, ten Dijke P et al. (2001). FEBS Lett 502: 4–10.
Folgueras AR, Pendas AM, Sanchez LM, López-Otín C . (2004). Int J Dev Biol 48: 411–424.
Fort P, Marty L, Piechaczyk M, el Sabrouty S, Dani C, Jeanteur P et al. (1985). Nucl Acids Res 13: 1431–1442.
Freije JM, Diez-Itza I, Balbin M, Sanchez LM, Blasco R, Tolivia J et al. (1994). J Biol Chem 269: 16766–16773.
Fujii M, Takeda K, Imamura T, Aoki H, Sampath TK, Enomoto S et al. (1999). Mol, Biol Cell 10: 3801–3813.
Goldberg GI, Wilhelm SM, Kronberger A, Bauer EA, Grant GA, Eisen AZ . (1986). J Biol Chem 261: 6600–6605.
Hanafusa H, Ninomiya-Tsuji J, Masuyama N, Nishita M, Fujisawa J, Shibuya H et al. (1999). J Biol Chem 274: 27161–27167.
Hu PP, Datto MB, Wang XF . (1998). Endocr Rev 19: 349–363.
Janda E, Lehmann K, Killisch I, Jechlinger M, Herzig M, Downward J et al. (2002). J Cell Biol 156: 299–313.
Javelaud D, Delmas V, Moller M, Sextius P, Andre J, Menashi S et al. (2005). Oncogene 24: 7624–7629.
Johansson N, Airola K, Grénman R, Kariniemi AL, Saarialho-Kere U, Kähäri V-M . (1997a). Am J Pathol 151: 499–508.
Johansson N, Ala-aho R, Uitto V, Grénman R, Fusenig NE, López-Otín C et al. (2000). J Cell Sci 113: 227–235.
Johansson N, Saarialho-Kere U, Airola K, Herva R, Nissinen L, Westermarck J et al. (1997b). Dev Dyn 208: 387–395.
Johansson N, Vaalamo M, Grénman S, Hietanen S, Klemi P, Saarialho-Kere U et al. (1999). Am J Pathol 154: 469–480.
Johansson N, Westermarck J, Leppä S, Häkkinen L, Koivisto L, López-Otín C et al. (1997c). Cell Growth Differ 8: 243–250.
Keski-Oja J, Raghow R, Sawdey M, Loskutoff DJ, Postlethwaite AE, Kang AH et al. (1988). J Biol Chem 263: 3111–3115.
Knäuper V, López-Otín C, Smith B, Knight G, Murphy G . (1996). J Biol Chem 271: 1544–1550.
Lansdorf CD, Grénman R, Bier H, Somers KD, Kim SY, Whiteside TL et al. (eds) (1999). Human Cell Culture, vol. 2, Cancer Cell Lines, Part 2. Kluwer Academic Press: Dordrecht (Holland).
Leivonen S-K, Chantry A, Häkkinen L, Han J, Kähäri V-M . (2002). J Biol Chem 277: 46338–46346.
Li AG, Lu SL, Zhang MX, Deng C, Wang XJ . (2004). Cancer Res 64: 7836–7845.
Li CY, Suardet L, Little JB . (1995). J Biol Chem 270: 4971–4974.
Massague J, Wotton D . (2000). EMBO J 19: 1745–1754.
Moustakas A, Souchelnytskyi S, Heldin CH . (2001). J Cell Sci 114: 4359–4369.
Nakao A, Afrakhte M, Moren A, Nakayama T, Christian JL, Heuchel R et al. (1997). Nature 389: 631–635.
Nikkola J, Vihinen P, Vlaykova T, Hahka-Kemppinen M, Kähäri V-M, Pyrhönen S . (2001). Melanoma Res 11: 157–166.
Oft M, Akhurst RJ, Balmain A . (2002). Nat Cell Biol 4: 487–494.
Pardali K, Kurisaki A, Moren A, ten Dijke P, Kardassis D, Moustakas A . (2000). J Biol Chem 275: 29244–29256.
Piek E, Westermark U, Kastemar M, Heldin CH, van Zoelen EJ, Nister M et al. (1999). Int J Cancer 80: 756–763.
Ravanti L, Häkkinen L, Larjava H, Saarialho-Kere U, Foschi M, Han J et al. (1999). J Biol Chem 274: 37292–37300.
Ravanti L, Toriseva M, Penttinen R, Crombleholme T, Foschi M, Han J et al. (2001). FASEB J 15: 1098–1100.
Reynisdottir I, Polyak K, Iavarone A, Massagué J . (1995). Genes Dev 9: 1831–1845.
Rodeck U, Nishiyama T, Mauviel A . (1999). Cancer Res 59: 547–550.
Sano Y, Harada J, Tashiro S, Gotoh-Mandeville R, Maekawa T, Ishii S . (1999). J Biol Chem 274: 8949–8957.
Siegel PM, Massagué J . (2003). Nat Rev Cancer 3: 807–821.
Ståhle-Bäckdahl M, Sandstedt B, Bruce K, Lindahl A, Jimenez M, Vega J et al. (1997). Lab Invest 76: 717–728.
Stetler-Stevenson WG, Yu AE . (2001). Semin Cancer Biol 11: 143–152.
Takekawa M, Tatebayashi K, Itoh F, Adachi M, Imai K, Saito H . (2002). EMBO J 21: 6473–6482.
Tian F, Byfield SD, Parks WT, Stuelten CH, Nemani D, Zhang YE et al. (2004). Cancer Res 64: 4523–4530.
Tian F, DaCosta Byfield S, Parks WT, Yoo S, Felici A, Tang B et al. (2003). Cancer Res 63: 8284–8292.
Uhl M, Aulwurm S, Wischhusen J, Weiler M, Ma JY, Almirez R et al. (2004). Cancer Res 64: 7954–7961.
Ungefroren H, Lenschow W, Chen WB, Fändrich F, Kalthoff H . (2003). J Biol Chem 278: 11041–11049.
Wakefield LM, Roberts AB . (2002). Curr Opin Genet Dev 12: 22–29.
Wilkinson GW, Akrigg A . (1992). Nucl Acids Res 20: 2233–2239.
Visse R, Nagase H . (2003). Circ Res 92: 827–839.
Yamaguchi K, Shirakabe K, Shibuya H, Irie K, Oishi I, Ueno N et al. (1995). Science 270: 2008–2011.
The expert technical assistance of Sari Pitkänen, Johanna Markola, and Marjo Hakkarainen is gratefully acknowledged. We also thank Drs E Bauer, J Keski-Oja and P Fort for cDNAs and Dr Jarkko Karvinen (Wallac, Perkin Elmer Life and Analytical Sciences, Turku, Finland) for MMP-13 activity assay. This study was supported by grants from the Academy of Finland (Project 45996), Sigrid Jusélius Foundation, the Cancer Research Foundation of Finland, Turku University Central Hospital (Project 13336), by European Union Framework Programme 6 (CANCERDEGRADOME, LSHC-CT-2003-503297), and by a personal grant to S-KL from Research and Science Foundation of Farmos.
About this article
Cite this article
Leivonen, S., Ala-aho, R., Koli, K. et al. Activation of Smad signaling enhances collagenase-3 (MMP-13) expression and invasion of head and neck squamous carcinoma cells. Oncogene 25, 2588–2600 (2006). https://doi.org/10.1038/sj.onc.1209291
- TGF-β; Smad
- matrix metalloproteinase
- squamous cell carcinoma
H-Ras activation and fibroblast-induced TGF-β signaling promote laminin-332 accumulation and invasion in cutaneous squamous cell carcinoma
Matrix Biology (2020)
European Heart Journal (2019)
Long noncoding RNA TM1P3 is involved in osteoarthritis by mediating chondrocyte extracellular matrix degradation
Journal of Cellular Biochemistry (2019)
Reduced USP33 expression in gastric cancer decreases inhibitory effects of Slit2‐Robo1 signalling on cell migration and EMT
Cell Proliferation (2019)
JCI Insight (2018)