|Home | About | Journals | Submit | Contact Us | Français|
We previously identified cystatin C (CystC) as a novel antagonist of transforming growth factor β (TGF-β) signaling in normal and malignant cells. However, whether the anti-TGF-β activities of CystC can be translated to preclinical animal models of breast cancer growth and metastasis remains unproven. Assessing the preclinical efficacy of CystC was accomplished using metastatic 4T1 breast cancer cells, whose oncogenic responses to TGF-β were inhibited both in vitro and in vivo. Indeed, we observed CystC to prevent TGF-β from stimulating the growth and pulmonary metastasis of 4T1 tumors in mice in part by reducing the extent of Smad2, p38 mitogen-activated protein kinase, and extracellular signal-regulated kinase 1/2 phosphorylation present in 4T1 tumors. We also found CystC to significantly antagonize angiogenesis in developing 4T1 tumors, suggesting a novel role for CystC in uncoupling TGF-β signaling in endothelial cells (ECs). Accordingly, CystC dramatically reduced murine and human EC responsiveness to TGF-β, including their ability to regulate the expression of 1) TGF-β signaling components, 2) inhibitor of differentiation (ID) family members, and 3) matrix metalloproteinases and their inhibitors (TIMPs) and to undergo cell invasion and angiogenic sprouting stimulated by TGF-β. Importantly, CystC prevented TGF-β from stimulating vessel development in Matrigel plugs implanted into genetically normal mice. Collectively, our findings provide the first preclinical evidence that CystC is efficacious in preventing breast cancer progression and angiogenesis stimulated by the oncogenic TGF-β signaling system and suggest that CystC-based chemotherapeutics possesses translational efficacy to one day treat and improve the clinical course of late-stage breast cancers.
Transforming growth factor β (TGF-β) is a multifunctional cytokine that governs a variety of diverse cellular processes in virtually all cell types and tissues, including their ability to differentiate, migrate, and proliferate and to undergo programmed cell death [1,2]. These activities of TGF-β are particularly pronounced in the mammary gland, wherein TGF-β regulates all stages of gland development and differentiation and potently suppresses mammary tumorigenesis [3,4]. Quite paradoxically, mammary tumorigenesis converts the functions of TGF-β in mammary epithelial cells (MECs) from that of a tumor suppressor to that of a tumor promoter. Once converted, TGF-β stimulation of late-stage malignant MECs promotes their invasion and metastasis, which remain the most lethal facets of developing and progressing mammary tumors [1,3,4]. At present, the precise nature of the cellular, genetic, and epigenetic events that enable mammary tumorigenesis to elicit oncogenic signaling and metastasis stimulated by TGF-β remains incompletely understood. However, it is known that the efficient dissemination of metastatic breast cancer cells requires tumor angiogenesis, which facilitates mammary tumor growth and survival, and provides a route for their eventual metastatic spread [5,6]. Like invasion and metastasis, angiogenesis is also a hallmark of tumorigenesis whose activation and resolution are governed by TGF-β [7,8]. The duality of TGF-β in regulating angiogenesis has been attributed to its activation in endothelial cells (ECs) of two distinct type I receptors, namely, 1) activin receptor-like kinase 1 (ALK-1), which mediates angiogenesis by stimulating Smad1/5/8, and 2) TGF-β type I receptor (TβR-I; ALK-5), which mediates angiostasis by stimulating Smad2/3 [7–9]. Unfortunately, how ECs ultimately distinguish between the proangiogenic and antiangiogenic signals induced by TGF-β remains to be fully elucidated. Despite these knowledge gaps, it does stand to reason that developing chemotherapeutic interventions capable of preventing TGF-β from stimulating mammary tumor metastasis and angiogenesis will offer new inroads into improving the clinical course of breast cancer patients.
Cystatin C (CystC) is a ubiquitously expressed secretory protein that inactivates members of the cathepsin family of cysteine proteases [10,11]. We identified CystC as a novel TGF-β type II receptor (TβR-II) antagonist that prevents its ability to bind TGF-β and, consequently, inhibits TGF-β signaling in normal and malignant cells [12,13]. Indeed, CystC administration inhibits TGF-β stimulation of gene expression, of epithelial-mesenchymal transition, and of cell proliferation and morphologic transformation [12,13], suggesting that CystC possesses therapeutic potential in alleviating oncogenic signaling stimulated by TGF-β. Despite these interesting in vitro findings [12,13], a potential role for CystC in suppressing oncogenic TGF-β signaling in malignant metastatic mammary tumors in vivo remains unexplored, as does the potential for CystC to inhibit TGF-β signaling in ECs and their ability to undergo angiogenesis. As such, the aims of this study were to determine whether CystC was efficacious in inhibiting the ability of TGF-β to 1) stimulate the growth and pulmonary metastasis of mammary tumors in mice and 2) provoke EC activities coupled to angiogenesis both in vitro and in vivo.
Control (i.e., green fluorescent protein [GFP]), CystC-, and Δ14CystC-expressing NMuMG cells were described and characterized previously . Malignant metastatic 4T1 cells were purchased from ATCC (Manassas, VA) and were cultured as described previously , whereas murine brain microvascularMB114 ECs were generously supplied by Dr. Michael Hart (University of Wisconsin, Madison,WI) and were cultured as described previously . Human umbilical vein ECs (HUVECs; passages 3–6) were purchased from Cambrex Corporation (East Rutherford, NJ) and were cultured in endothelial cell growth media-2 media supplemented with EC growth factors according to the supplier's instructions. Bicistronic retroviral vectors encoding for human CystC or Δ14CystC were described previously . CystC and Δ14CystC retroviral supernatants were infected into 4T1 carcinoma or MB114 ECs, which subsequently were sorted, collected, and expanded to yield stable polyclonal populations of transgene expressing cells (≥95% GFP-positive) as described previously [12,13].
The construction of bacterial expression vectors encoding for human CystC or Δ14CystC fused to the C-terminus of glutathione S-transferase (GST), together with the subsequent purification of their recombinant protein products from transformed Escherichia coli, was performed as described [12,13].
The effect of CystC or Δ14CystC expression on various TGF-β-stimulated activities in 4T1, MB114, or HUVEC cells was determined as follows: 1) Smad2/3 and Smad1/5/8 phosphorylation was monitored by immunoblot analysis with phosphospecific antibodies (Cell Signaling, Beverly, MA) as described , 2) cell invasion was induced by 2% serum in a modified Boyden chamber coated with Matrigel matrices (diluted 1:50 in serum-free media) as described [12,13], 3) synthetic gene expression was assessed in p3TP- and pSBE-luciferase reporter gene assays as described [12,13], 4) anchorage-independent cell growth during a 10-day period as described previously , 5) angiogenic sprouting in rat tail collagen matrices as described , and 6) endogenous transcript expression was measured using semiquantitative real-time polymerase chain reaction (PCR) as described previously . The oligonucleotide primer pairs used to analyze the expression of MT1-matrix metalloproteinase (MMP), MT2-MMP, MT3-MMP, MMP2, MMP3, MMP9, tissue inhibitor of metalloproteinase 1 (TIMP1), TIMP2, and thrombospondin 1 were described previously . Additional oligonucleotide primer pairs used detecting the expression of PAI-1, ALK-1, ALK-5, Smad1, Smad2, Smad3, Smad5, ID1, ID2, and ID3 are provided in Table W1. Additional detailed descriptions of these procedures are provided in Supplementary Materials and Methods.
Control (i.e., GFP)-, CystC-, or Δ14CystC-expressing 4T1 cells were resuspended in sterile PBS and subsequently were injected (12,000 cells per mouse) orthotopically into the mammary fat pads of 6-week-old female syngeneic Balb/C mice (three mice per condition; Jackson Labs, Bar Harbor, ME). Mice were monitored daily, and primary tumors were measured with digital calipers (Fisher Scientific, Pittsburgh, PA) every other day beginning on day 10 after inoculation. Tumor volumes were calculated using the following equation: tumor volume = x2y0.5, where x is the tumor width and y is the tumor length. Thirty days after inoculation, the mice were killed, and their primary tumors were excised, weighed, and processed for histopathologic analysis in the Pathology Core, University of Colorado Cancer Center. At the time of necropsy, the lungs were removed, minced, and digested proteolytically in PBS supplemented with 1 mg of Blendzyme (Roche Applied Science, Indianapolis, IN). Enzymatic reactions were allowed to proceed for 3 hours at 37°C under continuous rotation and, subsequently, were filtered through 70-µm nylon cell strainers. The resulting single-cell suspensions were washed twice in PBS before culturing the cells (1 x 106 cells per plate) onto 10-cm plates in Dulbecco's modified Eagle's medium/10% FBS media supplemented with 60 µM 6-thioguanine to select for metastatic 4T1 cells, which are resistant to 6-thioguanine treatment. After 14 days of growth in selection media, the resulting metastatic foci were fixed in 10% MeOH/10% acetic acid and stained with crystal violet. Finally, serial histologic sections of control (i.e., GFP)-, CystC-, or Δ14CystC-expressing 4T1 tumors that were stained with phosphospecific antibodies against Smad2 (1:50 dilution; Cell Signaling Technology, Danvers, MA), p38 mitogen-activated protein kinase (MAPK, 1:100 dilution; Cell Signaling Technology), and extra-cellular signal-regulated kinase 1/2 (ERK1/2, 1:100 dilution; Cell Signaling Technology), with antibodies against Ki-67 (1:300 dilution; BD Biosciences, San Jose, CA), with antibodies against CD31 (1:400 dilution; Dako, Denmark), with Masson's trichrome according to the manufacturer's recommendations (Sigma, St. Louis, MO), and with hematoxylin as previously described .
All animal studies were performed three times in their entirety and were performed according to animal protocol procedures approved by the Institutional Animal Care and Use Committee of University of Colorado Denver.
The effect of CystC and Δ14CystC on TGF-β-stimulated angiogenesis in vivo was investigated using the Matrigel implantation essentially as described previously . Briefly, 4- to 6-week-old C57BL/6 female mice were injected subcutaneously in the ventral groin area with Matrigel (500 µl per injection) supplemented with diluent (PBS), or with basic fibroblast growth factor (bFGF, 300 ng/ml; R&D, Minneapolis, MN) together with TGF-β1 (5 ng/ml), and recombinant (50 µg/ml) GST, GST-CystC, or GST-Δ14CystC. Ten days after implantation, mice were killed, and their Matrigel plugs were removed, fixed in 10% formalin, and sectioned in the Pathology Core, University of Colorado Cancer Center. Afterward, the sections were stained using the Masson's trichrome procedure to visualize infiltrating vessels, which were quantified under a light microscope by determining the average number of vessels present in 10 independent fields per slide on three independent slides. Three mice were used for each experimental condition, and this experiment was performed three times in its entirety. All animal studies were performed according to protocol procedures approved by the Institutional Animal Care and Use Committee of University of Colorado Denver.
We previously established CystC as a novel antagonist of oncogenic TGF-β signaling in a variety of normal and malignant cells, including murine and human MECs [12,13]. Whether these in vitro inhibitory activities of CystC could be translated to in vivo models of breast cancer growth and metastasis stimulated by TGF-β remains unknown. To address this important question, we infected malignant metastatic murine 4T1 breast cancer cells with murine ecotropic retroviruses encoding for either control (i.e., GFP), CystC, or Δ14CystC, which lacks the cysteine protease inhibitor signature (i.e., residues 80–93) and thus is incompetent to inactivate cathepsin proteolytic activity but remains competent to antagonize TGF-β signaling . We chose to study 4T1 breast cancer cells for two major reasons. First, the injection of human breast cancer cells into mice requires the use of immunocompromised animals, which can severely limit the interpretation of measured tumor behavior because of the absence of immunosurveillance in the animal. Second, TGF-β is a potent immunosuppressive agent that plays a critical role in maintaining immune system tolerance to self-antigens and in initiating and resolving inflammatory reactions. Moreover, the immunosuppressive activities of TGF-β can contribute to cancer progression in part by inhibiting immunosurveillance mediated by infiltrating lymphocytes. Our use of 4T1 cells circumvents these limitations and is bolstered further by recent findings from our laboratory [17,18] and from others' [19–21], establishing 4T1 cells as an important late-stage model of TGF-β-responsive breast cancer. Figure W1A shows that 4T1 cells transduced with CystC-based retroviruses readily secreted recombinant CystC or Δ14CystC proteins into the media, whereas those transduced with control (i.e., GFP) retrovirus expressed only low levels of endogenous CystC expression. Similar to what we observed previously in NMuMG and MDA-MB-231 cells , we found that the expression of either CystC or Δ14CystC to significantly inhibit TGF-β stimulation of 4T1 cell 1) invasion (Figure W1A), 2) p3TP- and pSBE-luciferase reporter gene expression (Figure W1B), and 3) anchorage-independent growth (Figure W1C). Collectively, these findings confirmed that CystC can suppress the oncogenic activities of TGF-β in malignant metastatic 4T1 cells.
We next analyzed whether CystC and Δ14CystC were able to antagonize the ability of TGF-β to promote the growth and pulmonary metastasis of 4T1 tumors produced in syngeneic Balb/C mice. Consistent with these in vitro findings, the growth (Figure 1A), weight (Figure 1B), and proliferative index (Figure 1C) of 4T1 tumors in Balb/C mice were inhibited significantly by their expression of either CystC or Δ14CystC. In addition, 4T1 cells are resistant to the cytotoxic activities of 6-thioguanine, and as such, metastatic 4T1 cells can be isolated by clonogenic assay of single-cell suspensions derived from the lungs of tumor bearing mice . As shown in Figure 1D, the expression of CystC or Δ14CystC significantly reduced the pulmonary metastasis of 4T1 tumors produced in Balb/C mice. We also performed immunohistochemistry on primary tumor sections to monitor the activation status of Smad2, p38 MAPK, and ERK1/2. Figure 2 shows that the phosphorylation and activation of all three TGF-β effectors were inhibited significantly in 4T1 tumors expressing either CystC or Δ14CystC compared with their control counterparts. Recently, elevated expression of ID1 and ID3 were associated with a pulmonary metastatic gene signature and with the proliferation of newly established pulmonary micrometastases . We therefore compared the changes in ID family member expression regulated by TGF-β in normal NMuMG and malignant metastatic 4T1 cells and, in doing so, found that TGF-β significantly suppressed the expression of ID1, ID2, and ID3 in NMuMG cells (Figure W2, A–C). In stark contrast, 4T1 cells readily upregulated their expression of ID1 and ID3 when treated with TGF-β (Figure W2, A–C), consistent with a role of these IDs in mediating pulmonary metastasis by TGF-β. More importantly, expression of CystC or Δ14CystC significantly reduced ID family member expression regulated by TGF-β (Figure W2D), suggesting that the ability of CystC to suppress breast cancer metastasis stimulated by TGF-β transpires in part through diminished ID1 and ID3 expression in metastatic MECs. Taken together, these findings demonstrate the efficacy of CystC to target and inhibit the oncogenic activities of TGF-β in late-stage mammary tumors produced in mice.
Angiogenesis is essential to tumor progression through its ability to provide developing neoplasms a supply of nutrients, a means for waste removal, and a route for metastatic cell dissemination to secondary organ sites . TGF-β plays critical roles in regulating both the activation and the resolution phases of angiogenesis [7,23–25], and as such, we analyzed the effect of CystC and Δ14CystC expression on the degree of angiogenesis in 4T1 tumors produced in mice. To do so, control (i.e., GFP)-, CystC-, and Δ14CystC-expressing 4T1 tumor sections were treated with Masson's trichrome, which stains collagen fibers blue and erythrocytes red. Figure 3A shows that control 4T1 tumors contained well-developed vessels that were filled with erythrocytes, particularly along the tumor edges. This suggests that 4T1 tumors indeed formed functional capillaries during their development and progression within the mammary fat pads of Balb/C mice, which contrasts sharply with the poorly formed and sporadically distributed vessels detected in CystC- or Δ14CystC-expressing 4T1 tumors (Figure 3A). We quantified the differences in tumor angiogenesis by staining these tumor slices with the EC marker, CD31, which showed that CystC and Δ14CystC both significantly reduced the microvessel densities in 4T1 tumors compared with their control counterparts (Figure 3B). Taken together, these findings provide the first demonstration that CystC possesses potent angiostatic activity against mammary tumors. Moreover, because Δ14CystC lacks inhibitory activity against cathepsins , but retains full antagonistic activity against TGF-β, these findings also suggest that CystC may limit vessel development in tumors by uncoupling TGF-β from angiogenesis activation.
As an initial step in addressing the previously mentioned hypothesis, we examined whether quiescent murine brain MB114 microvascular cells upregulate their production of CystC when stimulated with TGF-β. Figure W3 shows that treating MB114 cells with TGF-β stimulated the production of CystC transcripts and protein, thereby establishing CystC as a novel gene target of TGF-β in ECs as well as a potential regulator of EC response to TGF-β. We tested the latter supposition by infecting MB114 cells with retroviruses encoding control (i.e., GFP), CystC, or Δ14CystC (Figure W4A) and subsequently monitoring their responses to TGF-β. Similar to their effects on TGF-β signaling in normal and malignant MECs (Figures W1 and W2) , the expression of either CystC or Δ14CystC significantly inhibited TGF-β stimulation of MB114 cell invasion (Figure W4A) and luciferase reporter gene expression (Figure W4B). Similar inhibitory actions on TGF-β-stimulated cell invasion and reporter gene expression were observed after administration of either recombinant CystC or Δ14CystC to MB114 cells (Figure W4, C and D).
We also determined the consequences of CystC expression in regulating endogenous transcript production in quiescent and TGF-β-stimulated MB114 cells. In doing so, we limited our analyses to MB114 cells that expressed Δ14CystC because this CystC derivative is fully competent in antagonizing TGF-β signaling but has lost its ability in inhibiting cathepsin proteolytic activity [12,13]. Thus, differences in gene expression detected by semiquantitative real-time PCR likely reflect altered activation of the TGF-β signaling system as opposed to a combination of altered activation of the TGF-β and cathepsin signaling systems. Figure W5 shows that Δ14CystC selectively inhibited the expression of effector molecules operant in mediating the physiological activities of TGF-β, including that of ALK-5, Smad2, and Smad3 (Figure W5A). Moreover, these same cellular conditions reversed the ability of TGF-β to repress the expression of ID1 and ID2 in MB114 cells (Figure W5B). Finally, we observed Δ14CystC to significantly augment the expression of MT1-MMP induced by TGF while simultaneously attenuating that of MT3-MMP in these same cells (Figure W5C). Along these lines, we detected a trend toward diminished coupling of TGF-β to the expression of MMPs 2, 3, and 9 in Δ14CystC-expressing MB114 cells, findings consistent with the anti-invasive activity possessed by Δ14CystC (Figure W5C).
We also addressed whether the anti-TGF-β activities of CystC were unique to MB114 cells or were instead a more generalized phenomenon in ECs. As expected, administration of recombinant CystC to MB114 cells specifically inhibited the ability of TGF-β to stimulate the phosphorylation of Smad2/3 but failed to alter its coupling to the activation of Smad1/5/8 in MB114 cells (Figure 4A). Repeating this experiment in human HUVEC cells demonstrated that CystC antagonized the phosphorylation of Smad2/3 stimulated by TGF-β and reduced its activation of Smad1/5/8 (Figure 4B). Thus, CystC also inhibited TGF-β signaling in human ECs, which significantly impaired their ability to induce luciferase reporter gene expression (Figure 4C) and undergo invasion when stimulated with TGF-β (Figure 4D). Collectively, these findings demonstrate that the anti-TGF-β activities of CystC can be extended to include a generalized angiostatic function in ECs.
We demonstrated previously that overlaying quiescent MB114 cell monolayers with FBS-supplemented rat tail collagen induces their formation of angiogenic sprouts that ultimately give rise to the generation of three-dimensional capillary-like networks [15,16]. We used this assay here and found TGF-β to possess significant angiogenic activity in control MB114 cells but not in their CystC- or Δ14CystC-expressing counterparts (Figure 5, A and B). Similar inhibitory actions on TGF-β-stimulated angiogenic sprouting were observed after administration of either recombinant CystC or Δ14CystC to MB114 cells (Figure 6C). Thus, these analyses identify TGF-β as a proangiogenic factor for MB114 cells, whose ability to initiate tubulogenesis in response to TGF-β is neutralized by CystC and its analog, Δ14CystC. These findings also suggest that CystC administration may serve to alleviate the angiogenic activities of TGF-β in vivo. We tested this hypothesis by using the Matrigel plug implantation assay, which monitors the ability of various angiogenic agents to alter the neovascularization of Matrigel plugs implanted subcutaneously into mice. Figure 6 shows that bFGF stimulated significant neovascularization of implanted Matrigel plugs and that this angiogenic response was further enhanced by TGF-β. Whereas inclusion of recombinant GST into the Matrigel mixtures failed to alter vessel development in mice, addition of either recombinant CystC or Δ14CystC significantly impaired the development and infiltration of vessels into implanted Matrigel plugs stimulated by the combination of bFGF and TGF-β (Figure 6). Collectively, these findings demonstrate that the ability of TGF-β to stimulate angiogenesis in genetically normal mice can be targeted chemotherapeutically by administration of CystC.
The uncoupling of TGF-β from cytostasis often serves as a prelude for its ability to stimulate the eventual growth, invasion, and metastasis of developing and progressing mammary tumors. The oncogenic character of TGF-β in late-stage cancers has generated considerable effort to discover novel chemotherapeutics capable of inactivating the tumor promoting activities of TGF-β in patients with late-stage, metastatic disease, particularly in breast cancer patients. Included in this growing list of potential TGF-β chemotherapeutics are 1) large-molecule TGF-β antagonists, such as monoclonal TGF-β antibodies, soluble TβR-II:Fc fusion proteins, and antisense oligonucleotides, and 2) small-molecule TGF-β antagonists that target the ATP-binding sites of TβR-I and TβR-II [1,26,27]. In general, the use of these TGF-β-targeted therapies in preclinical animal models has yielded mixed results within specific tumor microenvironments, wherein they either alleviate or exacerbate disease development in a manner predicted by the TGF-β paradox . Thus, there remains a significant scientific and clinical need for the development of safe and effective anti-TGF-β chemotherapies that minimize off-target adverse reactions.
We identified CystC as a TGF-β gene target in fibroblasts and as a molecule whose expression is downregulated significantly in 44% of human cancers . More importantly, we established CystC and its engineered derivative, Δ14CystC, as novel TβR-II antagonists that inhibit oncogenic TGF-β signaling in normal and malignant MECs [12,13]. Despite these interesting findings, the question remained as to whether the in vitro anti-TGF-β activities could be recapitulated in and translated to in vivo models of breast cancer regulated by TGF-β and as to whether elevated CystC concentrations within these same mammary tumor microenvironments might also target TGF-β signaling in various stromal components, particularly ECs and their initiation of tumor angiogenesis. We show for the first time that delivering CystC to late-stage aggressive mammary tumors indeed diminish their capacity to grow and metastasize in genetically normal mice (Figure 1). Besides its ability to antagonize TGF-β signaling directly in breast carcinoma cells (Figure 2), CystC also was shown for the first time to 1) limit mammary tumorigenesis by inhibiting angiogenesis (Figure 3) and 2) uncouple TGF-β from stimulating angiogenic activities in ECs both in vitro (Figures 4 and and5)5) and in vivo (Figure 6). It should be noted that the role of cysteine proteases in promoting tumorigenesis has been recognized for decades  and that inhibiting cysteine protease activity can prevent tumor angiogenesis, invasion, and extracellular matrix degradation [30–33]. Interestingly, we find CystC and Δ14CystC, which lack anti-cysteine protease activity , to be equipotent in suppressing mammary tumor growth, metastasis, and angiogenesis in mice, which suggests that the primary antitumor activity of CystC transpires in part by antagonizing the ability of TGF-β to bind TβR-II . However, although our findings clearly support an anti-TGF-β activity as the major mode of action whereby CystC suppresses mammary tumor growth and progression in mice, they do not rule out or exclude alternative mechanisms that facilitate the ability of CystC to inhibit tumorigenesis. Indeed, we recently observed CystC administration to impact the response of carcinoma and ECs to bone morphogenic proteins 6 and 7 (M. Tian and W. P. Schiemann, unpublished observations). Thus, given the ability of the distant CystC relative, fetuin, to inhibit TGF-β and bone morphogenic protein signaling [34,35], it remains plausible that CystC functions as a general regulator of the activities and functions of TGF-β superfamily members. Similarly, it also remains possible that CystC inhibits tumorigenesis through additional protease-independent mechanisms reminiscent of those attributed to TIMP-3, which antagonizes angiogenesis by preventing vascular endothelial growth factor (VEGF) binding to VEGFR2 . Indeed, our broad-spectrum delivery of CystC and Δ14CystC to mammary tumor microenvironments likely mediates additional TGF-β-specific and -nonspecific activities in tumor-associated stroma and infiltrating immune cells, whose altered function and behavior in response to CystC administration clearly warrants examination. Finally, we recently observed that administration of small-molecule TβR-I antagonists induce a partial epithelial-mesenchymal transition in normal MECs that diminishes their epithelial character and morphology (T. M. Allington and W. P. Schiemann, unpublished observations). Importantly, these untoward morphologic and phenotypic adverse effects are not detected in normal MECs treated with either CystC or Δ14CystC , suggesting that elevated CystC concentrations seem to be well tolerated by normal MECs and mammary tissues, which further supports the potential development and application of CystC in clinical settings for patients with advanced breast cancer.
TGF-β expression is frequently upregulated in developing and progressing tumors, and this event has been associated with increasing tumor severity and grade [1,2]. Elevated TGF-β expression also plays a prominent role in directing autocrine and paracrine signaling networks within tumor microenvironments, which ultimately promotes tumor growth, invasion, and metastasis [1,37]. In addition, TGF-β also can instruct carcinoma cells to remodel their extracellular matrix in a manner that favors tumor invasion and metastasis [38,39]. Along these lines, elevated expression of ID1 has been associated with the ability of breast cancer cells to invade and metastasize to the lung , whereas ID1 and ID3 are necessary for EC progenitors to contribute to tumor angiogenesis [40,41]. Interestingly, TGF-β has been reported to be an important regulator of ID family member expression [42–44]. For instance, TGF-β was observed to chronically repress the expression of ID2 and ID3 in normal and malignant epithelial cells  but was found to induce the expression of ID1 and ID3 in fibroblasts . Thus, TGF-β regulates ID family member expression in a cell- and context-specific manner. Consistent with this notion, we observed TGF-β to significantly repress the expression of ID1, ID2, and ID3 in normal MECs, a finding that contrasted sharply with the significant induction of ID1 and ID3 expression in malignant metastatic 4T1 cells stimulated with TGF-β (Figure W2). Importantly, the expression of either CystC or Δ14CystC in 4T1 cells significantly inhibited their expression of ID1 to ID3 in response to TGF-β (Figure W2), thereby identifying a potentially important mechanism to uncouple TGF-β from ID family member expression and, consequently, to limit the malignancy of late-stage breast cancers. Future studies clearly need to explore more thoroughly this idea and the potential for CystC to uncouple TGF-β from the regulation of epithelial-mesenchymal transition-generated “stemness” in malignant metastatic MECs.
In addition to its ability to stimulate carcinoma invasion and metastasis, TGF-β has also been linked to both the activation and resolution phases of angiogenesis, presumably through its differential activation of ALK-1 versus ALK-5 in ECs [7,8,23]. For instance, activation of ALK-5 by TGF-β stimulates Smad2/3 and the production of PAI-1 and fibronectin, which collectively promote angiostasis and vessel maturation [42,47–49]. In contrast, activation of ALK-1 by TGF-β stimulates Smad1/5/8 and the expression of angiogenic genes, such as Id1 and interleukin 1 receptor-like 1 [42,47–49], which collectively promote angiogenesis activation. Moreover, ALK-1 signaling stimulated by TGF-β requires this cytokine to initially activate TβR-II and ALK-5, which then recruit and activate ALK-1 after its association with TβR-II:ALK-5:TGF-β ternary complexes . Thus, activation of ALK-1 and the induction of angiogenesis by TGF-β must first proceed through its assembly of angiostatic TGF-β receptor complexes (i.e., TβR-II:ALK-5). At present, the molecular mechanisms that initially exclude and then recruit ALK-1 to angiostatic TGF-β receptor complexes remain unknown but may reflect a delicate balance between TGF-β and other angiogenic factors located within tumor microenvironments. Indeed, low TGF-β concentrations enhance the ability of bFGF and VEGF to stimulate EC proliferation and angiogenic sprouting, whereas high TGF-β concentrations inhibit these angiogenic activities [25,50]. Along these lines, the proangiogenic functions of TGF-β also have been linked to its ability to regulate the expression and/or activities of other angiogenic factors, such as bFGF and VEGF . It is interesting to note that inclusion of TGF-β to Matrigel plugs implanted into mice only promoted angiogenesis and vessel development in the presence of bFGF and its ability to create a proangiogenic microenvironment (data not shown). Thus, it is plausible that the recruitment of ALK-1 to angiostatic TGF-β receptor complexes may first require the stimulation of accessory angiogenic signals or proteins within activated EC micro-environments. The existence of such a scenario seems unlikely for several reasons. First, CystC administration had no appreciable effect on the coupling of TGF-β to the activation of Smad1/5/8 in MB114 cells (Figure 4), despite the overall effectiveness of CystC to antagonize angiogenesis stimulated by TGF-β. Moreover, we observed CystC administration to selectively downregulate the expression signaling molecules targeted by ALK-5, not those targeted by ALK-1 (Figure W5), which is consistent with the presence of distinct ALK-5 and ALK-1 signaling systems in ECs. Second, our findings suggest that the activation of ALK-1 by TGF-β is dissociated from its stimulation of TβR-II:ALK-5 signaling complexes. This notion is supported by several recent studies showing that 1) ALK-1 and ALK-5 are expressed in a discordant and nonoverlapping manner in blood vessels  and 2) ALK-5 and TβR-II activity are not required for ALK-1 function in ECs . Finally, crystal structure analyses of TGF-β3 bound to the extracellular domains of TβR-I and TβR-II indicate that the molecular interactions necessary for ALK-1 binding are distinct from those operant in mediating the binding of ALK-5, making it unlikely that ALK-1 binds directly to or is readily recruited into active TβR-II:ALK-5 complexes . In addition, there remains considerable debate as to whether ALK-1 signaling does in fact mediate angiogenesis activation and, conversely, whether ALK-5 signaling does in fact mediate angiogenesis resolution [42,47,51,54,55]. Our findings in MB114 cells are consistent with the notion that ALK-5 couples directly to angiogenesis activation, an event that is readily antagonized by CystC administration. Future studies analyzing the angiogenic activities of TGF-β in ALK-1- and ALK-5-deficient MB114 cells clearly are warranted to address this issue, as are studies aimed at determining whether the activation of ALK-1 by TGF-β requires additional accessory receptor molecules or angiogenic factors.
The expression and secretion of recombinant CystC and Δ14CystC proteins by infected 4T1 and MB114 cells was monitored by immunoblotting conditioned medium with anti-CystC antibodies as described [1,2]. The effect of recombinant GST, CystC, or Δ14CystC on the ability of TGF-β to stimulate the phosphorylation of Smad2 or Smad1/5/8 phosphorylation was determined by allowing MB114 or HUVEC cells (100,000 cells per well) to adhere overnight to 24-well plates. The following morning, the cells were washed twice in ice-cold PBS and incubated in serum-free medium supplemented with 10 µg/ml of recombinant GST, CystC, or Δ14CystC for 2 hours at 37°C. Afterward, the cells were stimulated with TGF-β1 (2.5 ng/ml) for 30 minutes at 37°C and, subsequently, were subjected to phospho-Smad2 (1:500 dilution; Cell Signaling Technology, Danvers, MA) or phospho-Smad1/5/8 (1:500 dilution; Cell Signaling Technology). Differences in protein loading were monitored by reprobing stripped membranes with antibodies against β-actin (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA).
The effect CystC and Δ14CystC on the invasion of 4T1 or MB114 cells was determined as described previously [1,2]. Briefly, upper chambers were coated with 100 µl of diluted Matrigel (1:50 in serum-free media; BD Biosciences, San Jose, CA), which was evaporated to dryness overnight at room temperature. The following morning, the Matrigel mixtures were rehydrated and, subsequently, were cultured with control (i.e., GFP)-, CystC-, or Δ14CystC-expressing 4T1 or MB114 cells at a density of 100,000 cells per chamber. Cellular invasion was stimulated by the addition of 2% serum to the lower chambers and by the addition of TGF-β1 (5 ng/ml) to the upper chambers as indicated. Forty-eight hours later, the cells were washed in ice-cold PBS and immediately fixed for 15 minutes with 95% ethanol. Cells remaining in the upper chambers were removed with a cotton swab, whereas those remaining in the lower chamber were stained with crystal violet. The extent of cellular invasion was quantified using the NIH Image J software package to measure the optical density of the stained cells.
In some experiments, the effect of recombinant GST, CystC, and GST-Δ14CystC on MB114 and HUVEC cell invasion was examined. To do so, MB114 or HUVEC cells (100,000 cells per chamber) were allowed to invade through Matrigel in the absence or presence of 10 µg/ml of recombinant GST, GST-CystC, or GST-Δ14CystC, together with or without TGF-β1 (5 ng/ml) as indicated. All subsequent procedures were performed as described previously.
Analysis of luciferase activity driven by the synthetic p3TP or pSBE was performed as described previously [1,2]. Briefly, control (i.e., GFP)-, CystC-, or Δ14CystC-expressing 4T1 or MB114 cells were plated in 24-well plates at density of 30,000 cells per well and allowed to adhere overnight. The cells were transiently transfected the following day by overnight exposure to LT1 liposomes (Mirus, Madison, WI) that contained 300 ng/ml luciferase reporter complementary DNA (cDNA) and 100 ng/ml pCMV-β-gal cDNA, which was used as an internal control for transfection efficiency. Afterward, the cells were washed twice with PBS and stimulated overnight in serum-free medium supplemented with varying concentrations of TGF-β1. On completion of agonist stimulations, luciferase and β-gal activities contained in detergent-solubilized cell extracts were determined.
In some experiments, the effect of recombinant GST, GST-CystC, or GST-Δ14CystC in altering luciferase expression in MB114 or HUVEC cells was determined. To do so, MB114 or HUVEC cells were transiently transfected as previously mentioned and, subsequently, were stimulated with varying concentrations of TGF-β1 in the absence or presence of recombinant (10 µg/ml) GST, GST-CystC, or GST-Δ14CystC. All subsequent procedures were performed as described previously.
The growth of 4T1 cells in soft agar was performed according to the procedures described in . Briefly, triplicate cultures of control (i.e.,GFP)-, CystC-, or Δ14CystC-expressing 4T1 cells (10,000 cells per plate) were grown in 0.3% agar on a cushion of 0.6%agar in 35-mm plates. 4T1 cell growth in the absence or presence of TGF-β1 (5 ng/ml) was allowed to proceed for 10 days, wherein the number of colonies formed was quantified under a light microscope in 10 independent fields.
The ability of MB114 cells to form angiogenic sprouts in collagen matrices was performed as described previously . Briefly, control (i.e., GFP)-, CystC-, or Δ14CystC-expressing MB114 cells were cultured onto six-well plates at a density of 300,000 cells/well and, subsequently, were overlaid with 2 ml per well of solidified rat tail collagen, which was supplemented with 10% FBS to initiate angiogenic sprouting by quiescent MB114 cells that also were treated with or without TGF-β1 (5 ng/ml). In some experiments, recombinant (10 µg/ml) GST, GST-CystC, or GST-Δ14CystC proteins were mixed with rat tail collagen to assess their effects on TGF-β stimulation of angiogenic sprouting by quiescent MB114 cells. Angiogenic sprouting reactions were allowed to proceed for 7 days, at which point the number of invading sprouts was quantified under a light microscope by determining the average number of sprouts present in 10 independent fields per well.
The effect of CystC and Δ14CystC on the ability of TGF-β to regulate gene expression in either NMuMG, 4T1, or MB114 cells was assessed by semiquantitative real-time PCR as described previously . Briefly, control (i.e., GFP)-, CystC-, or Δ14CystC-expressing cells were cultured onto six-well plates at a density of 300,000 cells per well and allowed to adhere overnight. The following morning, the cells were washed in PBS and cultured in serum-free medium supplemented with or without TGF-β1 (5 ng/ml) for varying times at 37°C. Afterward, total RNA was isolated using the RNeasy Plus Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's recommendations. cDNA were synthesized by iScript reverse transcription (Bio-Rad, Hercules, CA), which were then diluted 10-fold in H2O and used in semiquantitative real-time PCR reactions (25 µl) that used the SYBR Green system (Bio-Rad) supplemented with 5 µl of diluted cDNA and 0.1 µM of oligonucleotide pairs listed below. Polymerase chain reactions were performed and analyzed on a Bio-Rad Mini-Opticon detection system, and differences in RNA concentrations were controlled by normalizing individual gene signals to their corresponding β-actin or GAPDH RNA signals. The oligonucleotide primer pairs used to analyze the expression of MT1-MMP, MT2-MMP, MT3-MMP, MMP2, MMP3, MMP9, TIMP1, TIMP2, and thrombospondin 1 were described previously . Additional oligonucleotide primer pair sequences are provided in Table W1.
Statistical values were defined using an unpaired Student's t test, where P < .05 was considered significant. P values for all experiments analyzed are indicated.
Members of the Schiemann Laboratory are thanked for critical comments and reading of the manuscript.