Phospho-specific Smad3 signaling

Abstract

Members of the TGFβ superfamily are known to exert a myriad of physiologic and pathologic growth controlling influences on mammary development and oncogenesis. In epithelial cells, TGFβ signaling inhibits cell growth through cytostatic and pro-apoptotic activities but can also induce cancer cell EMT and, thus, has a dichotomous role in breast cancer biology. Mechanisms governing this switch are the subject of active investigation. Smad3 is a critical intracellular mediator of TGFβ signaling regulated through phosphorylation by the TGFβ receptor complex at the C terminus. Smad3 is also a substrate for several other kinases that phosphorylate additional sites within the Smad protein. This discovery has expanded the understanding of the significance and complexity of TGFβ signaling through Smads. This review highlights recent advances revealing the critical role of phospho-specific Smad3 in malignancy and illustrates the potential prognostic and therapeutic impact of Smad3 phospho-isoforms in breast cancer.

Introduction

Every year approximately 44,000 women die of breast cancer in the United States, and after lung cancer, breast cancer is the most common cause of cancer death in women.¹ Members of the transforming growth factorβ (TGFβ) superfamily of growth factors have been associated with mammary gland development and breast oncogenesis and are known to exert a myriad of both normalizing and pathologic growth controlling influences on mammary physiology.^2-7 TGFβ signaling is transduced primarily through the activation of cell surface serine-threonine kinase receptors and downstream intracellular signaling proteins, the Smads. In epithelial cells, TGFβ inhibits cell growth through its cytostatic and pro-apoptotic activities but can also induce mammary cancer cell epithelial-to-mesenchymal transition (EMT); thus, TGFβ can have a dichotomous role in breast cancer biology. The dual roles of TGFβ action appear to be cancer stage-specific, with the growth factor functioning as a powerful tumor suppressor in early stage breast cancer and as an oncogene in aggressive, advanced stage mammary tumors. The mechanisms governing this switch are the subject of active investigation. Importantly, Smads serve as the substrates for several factors outside the canonical TGFβ pathway. The finding that Smads are phosphorylated and regulated by kinases other than those activated by the TGFβ superfamily has expanded our understanding of both the significance and complexity of TGFβ signaling, including crosstalk with other key signaling pathways, which impacts both tissue homeostasis and carcinogenesis.⁸

Our previous work and that of others has revealed the effect of cyclins/cyclin-dependent kinases (CDKs) on Smad3 signaling and cell cycle progression in breast cancer cells and mouse embryonic fibroblasts (MEF).^8-10 Elegant work has also shown the effect of Ras-mediated c-Jun N-terminal kinase (JNK) hyperactivity on Smad action and consequent EMT-like gene expression and cell invasion.¹¹ This review highlights recent advances describing the critical role of phospho-specific Smad3 in the progression of malignancies, including breast cancer.

TGFβ and Smad Signaling

Members of the TGFβ superfamily of growth factors share significant structural and functional homology, and several of these growth factors also play crucial roles in mammary gland physiology.¹² The biological actions of TGFβ superfamily members are dynamic and contribute to a wide variety of cellular processes, including proliferation, differentiation, motility, adhesion and apoptosis.¹³ There are over 40 members of this superfamily, including the TGFβ isoforms 1–3, activin and its structural homolog inhibin and bone morphogenic proteins (BMPs).¹⁴ TGFβ is the prototypic member of the TGFβ superfamily, mediating a number of biological effects on different cell types. Due to this pleiotropism, TGFβ activity is tightly regulated through several mechanisms, including intracellular factors, cell-cell interactions and crosstalk with other signaling pathways. Through canonical signaling, TGFβ exerts its action by binding to and inducing formation of cell surface receptor complexes consisting of type I (TβRI) and type II (TβRII) serine-threonine kinase receptors. Upon ligand binding, TβRII phosphorylates and activates TβRI. The resulting heteromeric complex facilitates the phosphorylation and subsequent activation of the intracellular Smad pathway (Fig. 1).

Figure 1. Canonical TGFβ signaling. TβRII phosphorylates TβRI in response to TGFβ binding. Activated TβRI phosphorylates Smad3 at its C-terminus, which releases it from the anchoring protein SARA and permits formation of a Smad3-Smad4 complex. This heteromeric complex then translocates to the nucleus, associates with DNA-binding co-activators and co-repressors and regulates the expression of target genes. I-Smads antagonize TGFβ signaling by preventing phosphorylation of Smad3 and thus act as a negative feedback mechanism. PPM1A is a phosphatase that can dephosphorylate Smad3 phosphorylated at its C-terminal SXS motif.

The Smads make up a group of intracellular proteins that have the critical role of transmitting TGFβ superfamily ligand signals to the nucleus.¹⁵ The Smads are categorized into three subgroups, the regulatory Smads (R-Smads), the common Smads (co-Smads) and the inhibitory Smads (I-Smads). In the TGFβ signaling pathway, the anchor protein SARA (Smad anchor for receptor activation) recruits the cytoplasmic signal transduction proteins, the R-Smads, Smad2 and Smad3, to the TβRI kinase domain at the cell surface, resulting in the phosphorylation of serine residues in the C-terminal SSXS motif. Phosphorylation of R-Smads releases them from SARA and into the cytoplasm. Activated Smad2 and Smad3 dimerize with co-Smad4, and these complexes translocate into the cell nucleus where they function as transcription factors to regulate the activity of specific promoters.

Smad2 and Smad3 are highly conserved proteins, with 83.9% amino acid sequence identity.¹⁶ Both proteins have two conserved domains: mad homology (MH)1 and MH2. The MH1 domain is responsible for DNA binding, whereas the MH2 domain interacts with transcriptional co-activators and co-repressors. The major structural discrepancy between Smad2 and Smad3 is in the MH1 domain, where Smad2 contains two short peptide inserts, amino acids 21–30 and 79–108, which impose steric constraints that prevent Smad2 from binding to DNA.¹⁶ Thus, Smad2-Smad4 complexes interact with DNA-binding proteins such as FAST2 (Forkhead activin signal transducer-2) to regulate downstream transcriptional responses. Although Smad3 can bind DNA directly, this Smad also depends on cooperation with DNA binding partners to direct the Smad3-Smad4 complex to specific target genes and to modulate transcriptional regulation. When the Smads successfully interact with target promoters, recruitment of co-activators or co-repressors can positively or negatively affect transcriptional activity of target genes.^15,17 Because TGFβ can stimulate the same canonical pathway—involving TβRII, TβRI and Smads 2, 3 and 4—to effect a variety of cellular responses, it is essentially the specific set of DNA binding Smad cofactors expressed in a cell that determines the response of that cell to TGFβ. Furthermore, in vitro studies have shown that Smad2 and Smad3 have unique functions in mammary epithelium and oncogenesis, in which Smad3 and not Smad2 is critical to inducing TGFβ-mediated apoptosis, cell cycle arrest and EMT.^18,19

Smad3-Related Cell Cycle Events

In epithelial cells, one of the most extensively studied cellular events downstream of canonical TGFβ signaling is growth inhibition. Smad3 is involved in cell cycle arrest through two primary mechanisms (Fig. 2). The first involves the association of Smad3 with cofactors E2F4/5 and p107 and then binding to a Smad-E2F site on the promoter of the mitogen c-myc, resulting in c-myc repression.²⁰ The second mechanism involves a complex including Smad3 with the transcription factor Sp1, which regulates transcription of the CDK inhibitors (cdkis) p15 and p21.^21,22 These endogenous cdkis repress the CDKs that drive the G₁ phase of the cell cycle. In breast epithelial cells, exposure to TGFβ increases expression of the cdk4/6i p15^Ink4b.²³ In addition, p27^Kip1 is a cdk2i that binds to CDK4 and CDK6 but is displaced upon p15 binding to CDK4/6 complexes.²⁴ Thus, TGFβ results in G₁ cell cycle arrest through p15-mediated CDK4/6 inhibition. Interestingly, overexpression of c-myc inhibits Smad-dependent transcription of p15 and p21 to overcome the cell cycle blockade, demonstrating the intersection of these two Smad3-dependent mechanisms of cell cycle control.²⁵ Studies in mammary epithelial cells also reveal how TGFβ acts to repress the CDK activator phosphatase Cdc25A.²⁶ Decreased levels of nuclear Smad3 have been correlated with estrogen receptor-negative breast cancers, overall larger tumor size and higher tumor grade.²⁷ Collectively, these data highlight the critical role that Smad3 plays in TGFβ-mediated growth inhibition. Accordingly, the loss of Smad3 function could allow for uncontrolled cell growth, a defining characteristic of cancer cell behavior.

Figure 2. TGFβ/Smad3 signaling and the cell cycle. CyclinD/CDK4/6 and Cyclin E/CDK2 mediate the transition of the cell cycle from the G₁ to S phase by phosphorylating Rb and thus preventing it from sequestering E2F. E2F is then free to drive the transcription of G₁/S phase transition proteins. TGFβ/Smad3 signaling promotes cell cycle arrest through its induction of p15 and p21, in conjunction with Sp1 and repression of c-myc, in conjunction with E2F4/5 and p107. Ras, in cooperation with Pin1, can increase the transcriptional activity of cyclin D, inducing tumor promotion.

Non-Canonical Smad3 Phosphorylation

Cyclin/CDK regulation.

Although Smad3 is activated through C-terminal phosphorylation by TβRI, phosphorylation by various intracellular kinases also regulates Smad3 action. The Smad3 linker region contains several serine/threonine sites that are phosphorylated by proline-directed kinases, such as CDK2/4, JNK, extracellular signal-regulated kinases (ERK)1/2, p38 mitogen-activated protein kinase (MAPK) and glycogen synthase kinase 3β (GSK3β) (Fig. 3).^8,28-30 Smad3 can be regulated non-canonically after pSmad3C enters the nucleus, where CDKs can phosphorylate Smad3, allowing the formation of pSmad3L/C. CDKs are the catalytic partners of cyclins, and our laboratory has been specifically interested in the impact of cyclin overexpression on TGFβ/Smad3 signaling. Critically, when Matsuura et al. reported that Smad3 was a physiological substrate of CDK4/2, an additional link between the cell cycle and TGFβ pathway was established, and a potential mechanism by which TGFβ resistance occurs in malignancy emerged.⁸ Specifically, mutation of Smad3 CDK phosphorylation sites in MEFs increased Smad3 transcriptional activity, leading to higher expression of p15 and decreased expression of c-myc.⁸ The same study also determined that in the nucleus, CDK2/4-mediated inhibitory phosphorylation of Smad3 occurs mostly at Thr8, Thr179 and Ser213 sites in vivo in addition to the Ser203 and Ser207 sites in vitro.⁸

Figure 3. Smad3 phosphorylation sites. Smad3 is phosphorylated by TβRI at the C-terminus and by various cancer-associated kinases at indicated sites. The MH1 domain is responsible for DNA binding, while the MH2 domain mediates interactions with co-activators and co-repressors.

Cyclin D is a key cell cycle protein that is overexpressed at the mRNA and protein levels in up to 50% of primary breast cancers.³¹ Amplification, translocation and overexpression of cyclin D1 are mechanisms by which cancer cells can obtain high levels of CDK activity.³² Recent evidence also suggests that an increase in stable cyclin D1, through aberrant nuclear export and proteolysis, plays a key role in accelerating tumor growth in a transgenic mouse model of mammary tumorigenesis.³³ Extensive work points toward cyclin D functioning as an oncogene and potential therapeutic target in many types of cancer, including breast cancer.³⁴ Cyclin/CDK complexes permit cells to transition from G₁- to S-phase of the cell cycle (Fig. 2). The activities of cyclin/CDK complexes are also modulated by the binding of cdkis including p15, p16, p21 and p27, which either sequester CDKs or bind and inhibit cyclin/CDK complexes. Specifically, cyclin D forms a complex with CDK4/CDK6, which can then phosphorylate members of the retinoblastoma (Rb) family of proteins.³⁵ Cyclin E/CDK2 can also phosphorylate Rb and functions in cell cycle progression.³⁵ Cyclin E is often overexpressed in breast cancer³¹ and is associated with triple-negative disease³⁶ and poor prognosis.³⁷ Hyperphosphorylation of Rb inhibits its ability to sequester members of the E2F transcription factor family, which are then free to drive the transcription of genes necessary for G₁/S-phase transition and S-phase progression. Thus, cyclin overexpression contributes to the loss of cell cycle control and to the potential for cellular transformation and primary tumor growth. Additionally, cyclin overexpression has been shown to positively correlate with disease progression and metastasis.^38,39 In cyclin-overexpressing cancers, high levels of CDK activity may cooperate with TGFβ-mediated signaling events to facilitate pro-tumorigenic outcomes.

Our previous work has shown that cyclin D- and cyclin E-overexpressing breast cancer cells have elevated CDK4 and CDK2 activity, respectively, and also express higher levels of the Smad3-regulated oncogene c-myc and lower levels of the tumor suppressor p15.^9,10 Transfection of the Smad3 protein mutated at various CDK4/2 phosphorylation sites was shown to restore Smad3 activity, resulting in higher mRNA levels of cdki p15 and lower levels of c-myc.^9,10 Treatment with a CDK4 or CDK2 inhibitor also increased Smad3 transcriptional activity in cyclin D- or cyclin E-overexpressing breast cancer cells. Current studies with CDK inhibitors demonstrate the ability of these predominantly cytostatic agents to also decrease cell migration in breast cancer cells (Jeruss Laboratory, unpublished data). These data suggest that pharmacologic CDK4/2 inhibition could serve as a targeted treatment strategy for patients with cyclin overexpressing breast cancer. Overall, the relative phosphorylation of canonical vs. linker region sites in the Smad3 protein has potential prognostic, as well as predictive capacity, with a therapeutic endpoint.

Ras-associated regulation.

Additionally, Smad3 can be regulated non-canonically by cytoplasmic kinases. Phosphorylation of Smad3 by MAPKs can inhibit canonical TβRI/Smad phosphorylation, while facilitating the formation of linker phosphorylated Smad3 (pSmad3L).^11,29,40-42 pSmad3L can associate with Smad4 and enter the nucleus in the same manner as C-terminal phosphorylated Smad3 (pSmad3C), and yet it inhibits the cytostatic events downstream of pSmad3C.²⁹ MAPKs are downstream of Ras and mutations that lead to constitutively activated Ras occur with relative frequency in pancreatic (90%), colon (50%) and thyroid (50%) cancers, yet ras mutations are comparatively infrequent in breast cancer (5%).⁴³ Simultaneously, Ras may be pathologically activated by overexpression of growth factor receptors that signal in conjunction with this protein. Specifically, Ras activation correlates with epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor (HER)2 overexpression in human breast cancer tissue samples.⁴⁴ Aggressive breast cancer subtypes that overexpress EGFR or HER2 and signal through Ras may provide a mechanism for Ras-mediated linker region Smad3 phosphorylation and the promotion of oncogenic TGFβ signaling in breast malignancy.

In a transformed mouse mammary epithelial cell line, Ras-activated ERK was shown to mediate phosphorylation of Smad3, predominately at linker region sites serine (Ser)204, Ser208 and threonine (Thr)179, resulting in cytoplasmic retention and consequent repression of canonical TGFβ signaling.²⁸ In addition, hyperactive Ras-expressing epithelial cells demonstrated lower TβRI-dependent pSmad3C tumor suppression and elevated JNK-dependent pSmad3L-mediated oncogenic activities, such as invasion.¹¹ Inhibiting JNK or mutating Smad3 linker phosphorylation sites restored tumor-suppressive TGFβ signaling and inhibited cell invasion. Similarly, in hepatic cells, TGFβ tumor suppressive action was mediated by pSmad3C, whereas oncogenic activities such as invasion and proliferation were mediated by pSmad3L.⁴⁵ Considering the direct role of Ras/MAPKs in promoting TGFβ oncogenic signaling through the linker phosphorylation of Smad3, pursuing studies that delineate the functional consequences of crosstalk between these two pathways will be critical to the development of effective novel therapies.

Cell cycle progression requires stimulation by extracellular growth factors until the late G₁ phase, which is referred to as the restriction point. After this point, cells are refractory to both mitogenic and anti-mitogenic signals until the next G₁ phase. Ras plays a critical role in mitogenic signaling, linking growth factor receptors to activation of protein kinase cascades, the most prominent being the MAPK pathway.⁴⁶ The MAPK signaling cascade regulates vital cellular processes related to gene expression, cell proliferation, cell motility and apoptosis.⁴⁷ Prior work suggest that Ras function is required for cell cycle progression through the G₁ to S-phase transition.^48,49 Importantly, growth-factor-induced regulation of cyclin D transcription as well as stabilization of the cyclin D protein and its assembly with CDK4/6 are regulated primarily through Ras-dependent pathways,^50-53 and cyclin D/CDK complexes are first detected after mitogenic stimulation of quiescent cells (Fig. 2).⁵⁴ In addition, cyclin D-deficient mice are resistant to breast cancers induced by the neu and ras oncogenes.⁵⁵ These findings contextualize how Ras/MAPKs, cyclin/CDKs and TGFβ/Smads pathways can crosstalk at many levels, predominately resulting in the potential for positive feedback of uncontrolled cell growth and oncogenic Smad signaling in breast cancer. Smad3 acting as a substrate for CDKs and MAPKs may represent a critical epigenetic mechanism by which Smad3 tumor suppressive activity is decreased.^56,57

Additional Mechanisms of Smad3 Regulation

In addition to both canonical and non-canonical phosphorylation, Smad3 activity is also controlled by I-Smads, Smads6/7. I-Smads are regulated by TGFβ signaling and compete with Smads2/3 for TGFβ receptor binding, which results in reduced canonical Smad3 phosphorylation.^58-60 Active Smad signaling is also impacted by ubiquitin-mediated degradation, with the linker region being the primary site of ubiquitination. The most studied ubiquitin ligase that binds to Smads2/3 is Smurf2, which binds to linker region PY motifs.^17,59 NEDD4L, an E3 ubiquitin ligase, also binds to phosphorylated linker regions of R-Smads.⁶¹ Dephosphorylation of Smads can also terminate Smad signaling by decreasing the ability of R-Smads to bind to Smad4 and to co-activators, co-repressors and other transcription factors. C-terminal Smad2/3 specific phosphatase, PPM1A is a Smad2/3-SXS-motif-directed phosphatase that was shown to abolish Smad2/3 phosphorylation despite the presence of constitutively active TGFβRI.⁶² Conversely, shRNA-mediated knockdown of PPM1A increased Smad2/3 C-terminal phosphorylation.⁶² Additional studies have suggested the existence of a linker region-specific Smad2/3 phosphatases, such as SCP1‑3.^63,64 Taken together, the phosphorylation status of Smad3 may not only impact downstream signaling but also Smad3 inactivation through degradation or dephosphorylation. These regulatory mechanisms require further investigation in the context of pSmad3L signaling.^{22,39,42-46,49,60}

Cofactors/Transcription Factors

R-Smads interact with various transcription factors, co-activators and co-repressors to induce cell type- and cell context-specific gene expression profiles. Certain co-factors have been shown to facilitate TGFβ-induced tumor suppressive or oncogenic transcriptional profiles. Casitas B-lineage lymphoma (CBL), a multifunctional adaptor protein and E3 ubiquitin ligase that is overexpressed in primary human breast cancer tissue samples, inhibits TGFβ-mediated growth arrest. This occurs through direct binding with Smad3 and not Smad2 in a C-terminal phosphorylation-dependent manner, preventing Smad3 from forming a complex with Smad4.⁶⁵ Recently, the transcriptional co-regulator SKI was found to promote linker phosphorylation of Smad3 and TGFβ-induced oncogenesis in melanoma cells.⁶⁶ In invasive breast cancer tissue samples, increased cytoplasmic SKI expression is inversely associated with tumor size, stage and lymph node status and positively associated with longer disease-free survival.⁶⁷ Research on the consequences of SKI action on Smad3 linker phosphorylation is actively being pursued in breast cancer cells.

Pin1 is an enzyme that plays a key role in regulating protein function through isomerization of phospho-serine/threonine-proline motifs, and Pin1 overexpression has been identified in breast cancer.^68,69 Pin1 was also shown to interact with Smad2/3 in a TGFβ-dependent manner, with the phosphorylated T179 residue in the Smad3 linker region acting as a binding site for Pin1.⁷⁰ In our work, T179 also emerged as the most critical of the five CDK4/2 phosphorylation sites on Smad3 in cyclin D-overexpressing breast cancer cells, as mutation of this site resulted in the greatest restoration of Smad3 reporter activity.¹⁰ Additionally, knockdown of Pin1 in human PC3 prostate cancer cells and MDA-MB-231 breast cancer cells resulted in inhibition of TGFβ-mediated migration and invasion.⁷⁰ Interestingly, Pin1 has been shown to cooperate with Ras signaling to increase cyclin D transcription in breast cancer, which potentially introduces another positive feedback mechanism driving oncogenic TGFβ signaling.⁶⁸

Many critical EMT-promoting transcription factors, such as Snail1 and Twist, also interact with Smads to form EMT-promoting Smad complexes that can either repress epithelial genes or activate mesenchymal genes.⁷¹ Further work will determine if these interactions are regulated by Smad phospho-isoforms and how these relationships affect normal and malignant downstream signaling events. Together, these findings indicate the potentially significant role that phospho-isoforms of Smads may play in mediating the specificity of TGFβ signaling through interaction with various key co-factors and transcription factors.

Role of Smad3 in Pro-Metastatic Events

EMT is an indispensable process associated with normal tissue development and organogenesis. However, inappropriate reactivation of EMT is a hallmark of cancer cell invasion and metastasis. As stated, TGFβ can stimulate tumor progression and metastasis by inducing EMT^72,73 in a process that is Smad3-dependent.¹⁹ During EMT, cells lose their epithelial characteristics, including cell adhesion and polarity, and acquire a mesenchymal morphology and the ability to migrate. During this process, cells downregulate the expression of epithelial markers and upregulate the expression of mesenchymal markers.⁷⁴ Smad3 has the capacity to regulate many key EMT-associated transcription factors, including Snail1/2, ZEB1/2 and Twist1.⁷¹ We hypothesize that the availability of different Smad3-binding proteins contributes to TGFβ-mediated induction of EMT. Recently, TGFβ signaling, in conjunction with the Wnt pathway, was linked to EMT and maintenance of mammary epithelial cells in a mesenchymal state both in an autocrine and paracrine manner.⁷⁵ Additionally, increased plasma TGFβ levels in breast cancer patients have been associated with poor outcomes.⁷⁶ These findings contextualize how oncogenes such as cyclins and Ras promote pro-metastatic Smad3 signaling effects in breast cancer cells, likely through the induction of EMT.

Further data supports the role of pSmad3L in the mesenchymal cellular phenotype and the induction of EMT markers. Ras-associated activation of JNK leads to pSmad3L-mediated signaling, which increases extracellular matrix (ECM) deposition and invasion by mesenchymal liver cells.^77,78 pSmad3L is also correlated with high expression of plasminogen activator inhibitor-1 (PAI-1), matrix metalloproteinases (MMP)-1, MMP-2 and MMP-9, known markers of EMT. The expression of these EMT markers is abrogated by JNK inhibition.¹¹ PAI-1 affects cell migration by inducing ECM proteolysis, while MMPs mediate cell invasion. In a colorectal adenocarcinoma model, expression of pSmad3L is increased in invasive cancer cells, while expression of pSmad3C is decreased when compared with normal colorectal epithelial cells.⁷⁹ Together, these studies provide insight into the mechanism by which Smad3 signaling can switch from the induction of cell cycle arrest to the promotion of EMT (Fig. 4).

Figure 4. Model for dichotomous Smad3 signaling. Differential phosphorylation of Smad3 results in the induction of TGFβ-mediated tumor suppressive genes in early-stage breast cancer cells or oncogenic genes in late-stage breast cancer cells. Nuclear kinases CDKs 2/4 phosphorylate Smad3 after pSmad3C translocates into the nucleus, while cytoplasmic MAPKs can phosphorylate Smad3 inducing nuclear activation independent of TGβRI phosphorylation.

Significance for Discovery of New Treatment Strategies

Although the importance of Smad expression and activity has been shown in several types of cancer—including leukemia, pancreas, colon, prostate and breast cancer—thus far, few studies have investigated a relationship between expression of Smad phospho-isoforms and cancer progression.^80-83 Work has shown that patients with chronic hepatitis B can develop hepatocellular carcinoma (HCC) depending on their pSmad3L/pSmad3C expression ratio, with high levels of pSmad3L expression predicting the development of HCC.⁴⁰ In addition, a recent study identified pSmad3C and pSmad3L (specific to the S213 site) as potential biomarkers to delineate patients at a high risk for recurrence of HCC after curative hepatectomy.⁸⁴ Currently, small-molecule inhibitors that target various aspects of TGFβ signaling have entered clinical trials in glioma, melanoma and breast cancer. These include TβR kinase inhibitors, antibodies that block binding of TGFβ ligands to their respective receptors, and anti-sense oligonucleotides that target TGFβ mRNA for degradation.⁸⁵ However, such strategies indiscriminately block all TGFβ signaling effects, both tumor suppressive and oncogenic. In addition, the success of targeting TGFβ signaling has been limited by indiscriminate pleiotropic effects. Therefore, targeting non-canonical Smad phosphorylation pathways in breast cancer may result in selected disruption of oncogenic TGFβ signaling, rendering this strategy a promising therapeutic approach. As pSmad3L and pSmad3L/C have been associated with a more invasive phenotype, pharmacological inhibition of linker phosphorylation may help to repress breast cancer oncogenesis. A clearer understanding of the roles of individual kinases in promoting TGFβ-directed tumorigenesis will allow for the development of more selective and, thus, beneficial treatment strategies for patients with cancers that overexpress certain biomarkers, such as cyclins D or E and Ras. Inhibitors against most kinases determined to phosphorylate Smad3 have been derived and are at different stages of development. Flavopiridol (a pan-CDK inhibitor) and R-roscovitine (CDK2 inhibitor) are two CDK inhibitors involved in clinical trials; however, they have been tested in breast cancer patients indiscriminate of the molecular profile of the tumors (Table 1). We expect that through selective in vivo and human tissue studies, patient subgroups will be identified that will benefit most from CDK inhibitor targeted therapy, resulting in improved patient outcomes.

Table 1. Potential drug candidates that promote TGFβ-mediated tumor suppression.

Name of Drug	Protein Target	Comments
Tipifarnib	Ras	• Tipifarnib + fulvestrant in MBC patients with no prior therapy, 51.6% CB rate • Tipifarnib + doxorubicin + cyclophosphamide as neoadjuvant therapy, 33% PR
Flavopiridol	CDK2/4/6/9	• > 50 clinical trials • Ineffective as monotherapy, but interacts synergistically with other drugs • Patients are those with advanced solid tumors (pancreatic, breast, ovarian) • Partial response in 29% breast cancer patients • Conclude that treatment with weekly, sequential doses of docetaxel then flavopiridol is safe and effective
R-roscovitine	CDK2	• Currently in Phase 3 clinical trails • Effective in breast cancer cells in vitro and in vivo • To date, best results in patients with nasopharyngeal cancer and NSCLC
Dinaciclib	CDK1/2/5/9	• Has acceptable safety and tolerability in patients with advanced, previously treated ER+ breast cancer • Showed antitumor potential as a single agent, needs to be tested in combination with other drugs

MBC = metastatic breast cancer; CB = clinical benefit; PR = partial response; NSCLC = non-small cell lung cancer; ER = estrogen receptor.

Conclusion

The role of TGFβ in malignancy is dynamic, influencing both tumor suppressive and oncogenic events. Prior work has shown the importance of canonical Smad3 signaling in the control of cell cycle progression for early stage breast cancers. More recent work has shown that non-canonical phosphorylation of Smad3, mediated by various factors, including CDKs and Ras, influences tumor promotion through cellular changes, such as EMT, that facilitate subsequent cell migration and invasion. Both the presence or absence of nuclear Smad3 and the relative ratio of non-canonical linker region phosphorylation in various malignancies may become a contributory marker of prognosis, and this is the subject of active study in our laboratory. Taken together, these findings are exciting, as they build on prior work focusing on cancer cell specificity, co-activators and co-repressors and allow for greater comprehension of the broad-spectrum influences of the TGFβ superfamily.

Future studies will elucidate the impact of phospho-specific Smad3 signaling in the context of other TGFβ family members, including activin and Nodal, as well as other kinases and cancer models. Additionally, determining the differential function of Smad2 and Smad3 may be key to understanding the role of Smad signaling during carcinogenesis and may provide even greater insight into the specificity of pro-tumorigenic Smad signaling in cancer cells. Future studies will also incorporate the critical role that Smad3 binding transcription factors play in determining cell type-specific effects of TGFβ signaling.⁸⁶ Further mechanistic understanding of differential kinase activity leading to Smad phosphorylation and the role of non-canonical signaling in the pathogenesis of breast cancer will lead to the identification of novel prognostic markers and therapeutic strategies. This effort will thereby add to the armamentarium available to treat malignancy and improve disease outcomes.

Glossary

Abbreviations:

TGFβ

transforming growth factor beta

EMT

epithelial-to-mesenchymal transition

CDK

cyclin-dependent kinase

JNK

c-Jun N-terminal kinase

TβRI

TGFβ receptor I

TβRII

TGFβ receptor II

R-Smads

receptor-Smads

co-Smads

common Smads

I-Smads

inhibitory Smads

SARA

Smad anchor for receptor activation

MH

mad homology

ERK

extracellular-signal regulated kinase

MAPK

mitogen-activated protein kinase

GSK3β

glycogen synthase kinase 3 beta

Ser

serine

Thr

threonine

pSmad3L

linker phosphorylated Smad3

pSmad2L

linker phosphorylated Smad2

pSmad3C

C-terminal phosphorylated Smad3

HGF

human growth factor

EGF

epidermal growth factor

Rb

retinoblastoma

MEF

mouse embryonic fibroblast

ECM

extracellular matrix

PAI-1

plasminogen activator inhibitor 1

MMP

matrix metalloproteinase

CDKi

CDK inhibitor

HCC

hepatocellular carcinoma

EGFR

epidermal growth factor receptor

HER2

human epidermal growth factor receptor 2