|Home | About | Journals | Submit | Contact Us | Français|
Plasmin is shown to play a crucial role in many pathophysiological processes, primarily through its ability to degrade extracellular matrix and/or mobilizing growth factors that are sequestered in the extracellular matrix. CCN1 is a matricellular protein whose expression is upregulated in cancer and various vascular diseases. The present study was undertaken to investigate whether plasmin liberates CCN1 from the extracellular matrix and whether the released growth factor modulates endothelial cell migration. Treatment of breast carcinoma cells (MDA-MB-231) with plasmin released a truncated form of CCN1 (28 kDa) into the overlying media. Experiments with recombinant CCN1 confirmed that plasmin effectively cleaves CCN1. Thrombin and other clotting/fibrinolytic proteases are ineffective in cleaving CCN1. Further studies revealed that the conditioned media of plasmin-treated carcinoma cells support endothelial cell migration and antibodies specific to CCN1 blocked this enhancing effect. These data were the first to show that plasmin can liberate a pluripotent matrix signaling protein, CCN1, from the ECM. Since both CCN1 and the components of the plasmin generation system are present in tumor cells and a variety of other cells, the proteolysis of CCN1 by plasmin may play a role in many pathophysiological processes, including tumor cell-mediated angiogenesis.
The plasminogen activator/plasmin system plays an important role, in addition to fibrinolysis, in many pathophysiological processes, including cancer, atherosclerosis, vascular remodeling and wound healing (1–4). Plasmin is shown to affect various cellular processes primarily through the proteolysis of the extracellular matrix (ECM), either directly or indirectly, via activation of matrix metalloproteinases (1, 2). Plasmin may also play a role in tissue remodeling and angiogenesis by activating/liberating growth factors, such as TGFß-1 (5, 6) and bFGF (7, 8), from the extracellular matrix. Recent studies suggest that plasmin also stimulates various signaling pathways and induces expression of a number of genes that play an important role in inflammation (9, 10). Our recent studies (11) show that plasmin upregulates the expression of CCN1 (cysteine-rich 61, CYR61) in fibroblasts, a matricellular protein that is capable of regulating various cellular functions (12).
CCN1 belongs to a novel family of growth regulators, often referred as CCN (cysteine-rich 61, connective tissue growth factor, nephroblastoma overexpressed) family (13). CCN1 is a secreted protein and associates with the extracellular matrix and cell surfaces (14). Purified CCN1 is shown to support cell adhesion, migration and proliferation of endothelial cells and fibroblasts (15–19). CCN1 acts as a ligand to multiple integrin receptors, and the cellular activities of CCN1 are shown to be mediated in part through the interaction with integrins and cell surface heparan sulfate proteoglycans in a cell type and context-specific manner (15, 19, 20). Overexpression of CCN1 in tumor cells is shown to promote tumor growth and vascularization (21). Furthermore, recent studies show that CCN1 is highly expressed in breast carcinoma (22, 23) and gliomas (24, 25), and the level of expression is positively correlated with clinical and pathological parameters of cancer, suggesting that CCN1 plays a role in the pathogenesis of cancer. Although stimulatory effects of CCN1 on cell proliferation, migration and survival via its interaction with various integrins are thought to be responsible for its role in tumorigenesis (26), at present, it is unclear how CCN1 sequestered in ECM could interact with its potential receptors or regulatory proteins.
CCN proteins are multimodular proteins. Except for CCN5/WISP-2, all other CCN proteins contain four structural modules (26). The current view is that each of the four modules can act both independently and interdependently to elicit various cellular responses (26). If so, the production of truncated isoforms of the CCN proteins in pathophysiological conditions could play a critical role in the modulation of their biological activity. Evidence had been provided for the presence of naturally occurring truncated forms of CCN2 and CCN3 that are biologically active (27, 28). The molecular mechanisms underlying the production of these secreted truncated forms are presently unknown.
Since coagulation and fibrinolytic pathways are frequently activated in many pathophysiological conditions and are thought to contribute to pathogenesis of various diseases, including tumor metastasis (4), we have investigated in the present study whether proteases involved in clotting and fibrinolysis can generate biologically active truncated forms of CCN1. The present study reveals that plasmin rapidly cleaves CCN1 and releases the truncated form of CCN1 from breast carcinoma cells. The truncated form of CCN1 released from the carcinoma cells is shown to promote endothelial cell migration. These studies raise a novel possibility by which plasmin can play a role in angiogenesis and other cellular processes.
MDA-MB-231 cells were obtained from ATCC (Rockville, MD) and cultured at 37°C and 5% CO2 in Dulbecco’s modified Eagle medium (DMEM) containing high glucose and supplemented with 1% glutamine, 1% penicillin/streptomycin and 10% heat-inactivated fetal bovine serum. Primary cultures of human umbilical vein endothelial cells (HUVEC) were obtained from Clonetics (San Diego, CA) and cultured in F12K medium supplemented with 1% glutamine, 1% penicillin/streptomycin, 10% fetal bovine serum, 15 U/ml heparin and 30 μg/ml endothelial cell growth supplement. Endothelial cells, passages between 3 to 7, were used in the present study.
Plasmin, thrombin and other coagulant proteases (purified from human plasma) were obtained from either Enzyme Research Laboratories (South Bend, IN) or Haematologic Technologies Inc (Essex Junction, VT). Recombinant human factor VIIa was generously provided by Novo Nordisk (Gentofte, Denmark). D-Phe-L-Phe-L-Arg chloromethyl ketone (PPACK) was obtained from Calbiochem, San Diego, CA. Recombinant human CCN1 was expressed in High Five insect cells that were stably transfected with a CCN1 expression plasmid containing V5 and His tag (constructed in pIZ/V5-His vector, Invitrogen, Carlsbad, CA). The recombinant CCN1 from the conditioned media was purified essentially as described earlier using SP-Sepharose chromatography (16). The recombinant protein was further purified by affinity chromatography on a nickel column to obtain an apparent homogenous protein. The purified CCN1 was used to immunize a rabbit to raise polyclonal antibodies against CCN1. The antiserum was subjected to DEAE Affi Gel-Blue (Bio-Rad) chromatography to isolate IgG fraction. CCN1 polyclonal antibodies for epitopes corresponding to internal region of CCN1 (aa 163–240) (H-78) and C-terminus (C-20) were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).
Purified recombinant CCN1 (~ 10 μg/ml) was treated with varying concentrations of plasmin, thrombin, VIIa/TF or other proteases. At different time intervals, aliquots were removed and added to SDS-PAGE sample buffer to stop the reaction. The samples were subjected to SDS-PAGE, followed by immunoblot analysis with CCN1 antibodies, H-78 and C-20.
Confluent monolayers of MDA-MB-231 cells cultured in 12-well plate were washed with serum-free DMEM medium (SFM) and allowed to stabilize in the SFM for 2 h before they were subjected to treatments. The monolayers were treated with control SFM (250 μl) or the SFM containing plasmin or thrombin (10 nM). At varying times, the overlying cell medium was removed (to which PPACK was added immediately to a final concentration of 1 μM) and the cells were harvested in 200 μl cell lysis buffer (20 mM Tris-HCl, 150 mM NaCl pH 7.5 containing the inhibitory cocktail, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, and 1 μM PPACK). Fifteen μl of samples were loaded on SDS-PAGE, followed by immunoblot analysis with CCN1 or ß-actin antibodies. To collect conditioned cell media for cell migration and other studies, cells cultured in T-75 flasks were treated with 4 ml of control SFM or the SFM containing 10 nM of plasmin or thrombin. In parallel, plasmin was added to a culture flask containing no cells. At the end of 2 h treatment, the overlying cell media was removed and then treated with PPACK (1 μM) for 90 min at room temperature to inhibit the proteolytic activity of plasmin and thrombin. The conditioned media was then dialyzed overnight against buffer A (10 mM Hepes, 0.15 M NaCl, pH 7.5 containing 4 mM KCl and 11 mM glucose) and passed through 0.22 μm filter before they were used for cell migration studies.
Anti-CCN1 IgG (raised against full-length recombinant CCN1) was coupled to Affi-Gel-15 (5 mg IgG/ml beads) using 0.1 M Hepes, pH 7.5 as a coupling buffer by following a protocol suggested by the vendor (Bio-Rad, Richmond, CA). The conditioned media (3 ml) collected from the plasmin-treated cells was incubated with anti-CCN1 agarose beads (0.5 ml) for 4 h at 4°C. Then, the beads were loaded into a column and the flow through (unbound material) was collected. The column was washed with 8 ml of 10 mM Hepes, 0.15 mM NaCl, pH 7.5, and the bound material was eluted with 0.1 M glycine, pH 2.4. The pH of eluted fractions was adjusted to 7.5 by adding 1.0 M Tris.HCl. Immunoblot analysis revealed that the flow through was devoid of the 28 kDa CCN1 fragment, and the fragment was recovered in eluted fractions (2 to 3 fractions, 1 ml each). These fractions were pooled and dialyzed overnight against buffer A. A similar procedure was followed for adsorption with heparin-agarose beads and the bound material was eluted with a high salt buffer (buffer A containing 1.0 M NaCl). In this case, the CCN1 peptide fragment was present in the flow through and not in the eluted fractions.
A Transwell system (8 μm pore size, polycarbonate filter, 6.5 mm diameter) was used to evaluate cell migration. Upper chamber was coated with fibronectin (40 μg/ml) for 2 min and then air-dried. HUVEC (50,000 cells in 100 μl SFM) were added to the upper well and 100 μl of conditioned media (plus 400 μl of SFM) was added to the lower chamber. At the end of 6 h incubation at 37°C/5% CO2, cells on the top-side of the membrane were removed by swiping with a damp cotton swab. The membrane was rinsed once with distilled water and fixed/stained with hematoxylin (Hema3 staining kit, Fisher Scientific) according to the instructions provided with the kit. The number of cells that had migrated was determined by counting number of cells on the underside of membrane (in five randomly selected fields) at × 20 magnification using a microscope equipped with a grid in its eyepiece.
Our recent studies in fibroblasts showed that both plasmin and thrombin induced CCN1 mRNA expression (11, 29), but the CCN1 antigen was found only in cell lysates of thrombin-stimulated cells (30). These data suggested that plasmin might be liberating/cleaving newly synthesized CCN1, which associates with cells/ECM. To test this possibility, we investigated whether plasmin cleaves CCN1. Recombinant CCN1 was treated with plasmin, thrombin or other clotting/fibrinolytic proteases for varying times and the aliquots were analyzed by immunoblot analysis using CCN1 antibodies that identify different epitopes on the CCN1. As shown in Fig.1, plasmin (1 nM) rapidly cleaved CCN1, resulting in two peptide fragments initially, corresponding to mol. wts. of 28 and 21 kDa. H-78 antibody, which recognizes an internal epitope of CCN1, detected the 28 kDa fragment (Fig. 1B) whereas C-20 antibody (C-terminus epitope antibody) recognized the 21 kDa fragment (Fig. 1D). These data suggest that the 28 kDa fragment corresponds to the N-terminus portion of CCN1 and the 21 kDa fragment corresponds to the C-terminus portion. Upon prolonged incubation with plasmin, the 21 kDa fragment was further cleaved into a 14.5 kDa fragment (recognized by C-20 antibody). This probably reflects plasmin cleavage of the C-terminal 21 kDa peptide fragment at the junction of CCN1 and V5/His tag (expected mol.wt. of the tag, 6 kDa) or at very near of it. In contrast to plasmin, thrombin was ineffective in cleaving CCN1 (Fig. 1C). However, prolonged incubation of CCN1 with high concentrations of thrombin resulted in a similar, but partial cleavage of CCN1 (data not shown). In addition to plasmin and thrombin, we also investigated the effect of other proteases (at concentrations of 1 and 10 nM) on CCN1 proteolysis. Factor VIIa/tissue factor, activated protein C, urokinase-type plasminogen activator, tissue-type plasminogen activator failed to cleave CCN1. Factor Xa and trypsin cleaved CCN1, but they are inefficient compared to plasmin (data not shown).
Next, we investigated whether plasmin cleaves and liberates CCN1 that is associated with cells/ECM. For these studies, we used MDA-MB-231 breast carcinoma cells, which constitutively express CCN1. MDA-MB-231 cells were treated with a control serum-free medium or the serum-free medium supplemented with plasmin. Immunoblot analysis of cell lysates and the overlying conditioned media obtained at varying times showed that plasmin released the 28 kDa CCN1 fragment into overlying conditioned media with a corresponding decrease in the cell associated CCN1 (Fig. 2). Similar results were observed with other breast carcinoma cells, MCF-7 and MDA-MB-435 (data not shown). However, in contrast to the data obtained with plasmin cleavage products of recombinant CCN1, C-20 antibody failed to detect the C-terminus CCN1 peptide fragment either in the conditioned media or cell lysates of tumor cells exposed to plasmin. The reason for the failure of C-20 antibodies to detect the carboxy peptide fragment probably could be due to low avidity of these antibodies to CCN1, particularly to the peptide fragment. Alternatively, the C-terminal fragment is proteolytically degraded further by plasmin, and thus not detected by C-20 antibodies. Plasmin-induced proteolysis of CCN1 in MDA-MB-231 cells was not mediated through plasmin activation of matrix metalloproteases since BM6001 and GM1489, potent and broad range inhibitors of matrix metalloproteases (MMP1, MMP2, MMP3, MMP8 and MMP9) failed to inhibit plasmin cleavage of CCN1 (data not shown).
To determine pathophysiological relevance of plasmin release of CCN1, we next investigated whether exposure of MDA-MB-231 cells to plasmin releases a soluble factor that is capable of promoting endothelial cell migration. As shown in Fig. 3A, the conditioned media obtained from the plasmin-treated cells increased the endothelial cell migration by 3 to 4-fold compared to the basal endothelial cell migration observed in the absence of a stimulant. In contrast, the conditioned media from MDA-MB-231 cells that were treated with the serum-free medium or thrombin failed to promote endothelial cell migration. The increased endothelial cell migration observed with the conditioned media of plasmin-treated cells could not be due to traces of plasmin leftover in the conditioned media since PPACK, an active site inhibitor that was used to inactivate plasmin protease activity, was shown to completely inhibit the protease activity of plasmin (data not shown). More importantly, plasmin that was subjected to a similar procedure (except exposing it to the cells) failed to promote the endothelial cell migration.
Since plasmin has been shown to release or activate growth factors, such as bFGF (7, 8) and TGFß (5, 6), which could have been sequestered on ECM or cell surfaces, we first investigated whether neutralizing antibodies against bFGF and TGFß impaired the ability of the conditioned media of plasmin-treated cells to support the endothelial cell migration. Both bFGF and TGFß neutralizing antibodies had only minimal and insignificant effect on the endothelial cell migration supported by the conditioned media of plasmin-treated cells (Fig. 3B), suggesting that it is unlikely that the release of these growth factors are responsible for the increased endothelial cell migration observed with the conditioned medium. Similarly, anti-IL-8 antibodies also had no effect on the cell migration.
Next, to determine whether CCN1 released into the conditioned media upon the exposure of breast carcinoma cells to plasmin is responsible for the endothelial cell migration, we investigated the effect of anti-CCN1 antibodies on the endothelial cell migration supported by the conditioned media of plasmin-treated cells. As shown in Fig. 4A, CCN1 antibodies fully attenuated the increased endothelial cell migration observed with the conditioned media of plasmin-treated cells. As expected, control IgG had no effect on the endothelial cell migration. In additional experiments, removal of CCN1 from the conditioned media of plasmin treated cells (by passing through anti-CCN1 antibody column) abolished the ability of the conditioned media to promote the cell migration (Fig. 4B). Consistent with this observation, the material eluted from CCN1 antibody column promoted the cell migration.
Since we were unable to detect the C-terminal fragment of CCN1 by immunoblot analysis either in the conditioned media or cell lysates, we cannot completely rule out the presence of the C-terminal derived fragment(s) or its contribution to endothelial cell migration in the conditioned media derived from plasmin-treated cells. Earlier studies showed that CCN1 binds to heparin tightly and the heparin binding sites are localized within the carboxy terminus module of the molecule (19). We have exploited this property to determine whether the 28 kDa fragment of CCN1, which lacks the carboxy terminus module or the C-terminus fragment of CCN1 was responsible for the cell migration observed in the conditioned media of plasmin-treated cells. The conditioned media of plasmin-treated cells was incubated with heparin-agarose beads, and the beads were washed and eluted with 1.0 M NaCl. Both the unbound and the eluted materials (after the dialysis) were evaluated for their ability to support the cell migration and for the presence of the 28 kDa fragment (by immunoblot analysis). The data revealed that the flow through material and not the eluted material supported the cell migration (Fig. 4C). A small difference in cell migration observed between the starting material and the flow through material was not statistically significant (p = 0.42). Immunoblot analysis revealed that the 28 kDa CCN1 fragment was found exclusively in the flow through material (data not shown). These data suggest that the aminoterminal fragment (28 kDa) is responsible for the cell migration observed in the conditioned media.
To further strengthen the data that the CCN1 fragment generated by plasmin is capable of enhancing the endothelial cell migration, we compared the migrating enhancing ability of plasmin-cleaved CCN1 with the full-length recombinant CCN1. As shown in Fig. 5, both the full-length recombinant CCN1 and the plasmin-cleaved CCN1 increased the endothelial cell migration in a dose-dependent manner. The increase was evident with as low as 200 ng/ml CCN1, and reached statistical significance at 1 μg/ml. Small differences between the full-length and the plasmin-cleaved recombinant CCN1 were not statistically significant. These data show that the migrating ability of CCN1 was retained upon proteolysis by plasmin.
The plasminogen activator/plasmin system is thought to play an important role, in addition to clot lysis, in a number of pathophysiological processes such as inflammation, wound healing, vascular remodeling, tumor invasion and metastasis by regulating various cellular activities involved in these processes (1, 4, 31, 32). In tumor metastasis, the system appears to be involved not only in tumor invasion but also in tumor angiogenesis (32–34). Although uPA, uPAR and PAI-1 were shown to regulate tumor cell migration and angiogenesis by plasmin-independent mechanisms (32, 33, 35), plasmin appears to play a crucial role in facilitating these processes by breaking anatomical barriers by degrading proteins in basement membranes and ECM directly or indirectly via activation of other proteases (36). Hydrolysis of matrix components by plasmin may lead to the release of biologically active fragments of the matrix and/or release of matrix-bound growth factors. This would facilitate the diffusion of biologically active peptide fragments and growth factors to more distant sites. Earlier studies showed that plasmin released or cleaved growth factors, such as bFGF (7, 8) and TGFß-1 (5, 6), from the extracellular matrix. In the present study, we show that plasmin also cleaves a pluripotent matricellular signaling protein, CCN1, and liberates a biologically active peptide fragment of CCN1 from tumor cells/matrix that could support endothelial cell migration.
CCN1 is a novel ECM-associated signaling protein, which was shown to influence not only tumor angiogenesis but also a myriad of cellular functions that contributes to many pathophysiological conditions (12, 21). CCN1 is composed of four discrete structural domains, insulin-like growth factor binding protein-like module (IGFBP), Von Willebrand factor type C-like module (VWC), thrombospondin 1-like module (TSP1) and cystein-knot containing family of growth regulator-like module (CT). A variable region connects domains I and II with III and IV. TSP1 repeat (WSxCSxxCG) is thought to be involved in binding to matrix macromolecules, particularly to sulfated glycoconjugates (13). CT domain contains two heparin binding motifs (27) that may participate in interaction with ECM (28). Our present data show that plasmin cleaves CCN1 in the thrombospondin module region, probably at R250 (Val251) or R259 (Lys260) since the ~28 kDa peptide fragment of the cleavage products was detected by the antibody that recognizes the variable region (163 to 240 aa) and not the C-terminus. Thus, the 28 kDa fragment represents the N-terminus portion of the protein. Multiple attempts of N-terminal sequencing of the 21 kDa C-terminus peptide fragment (using ~100 picomoles or more of the fragment) to identify the precise cleavage site were unsuccessful.
The observation that the 28kDa fragment of CCN1 was found in overlying media and not in cell lysates suggests that the binding motifs that participate in CCN1 binding to cell surfaces or ECM lie out side of this region, i.e., in the C-terminus portion of TSP1 domain and/or in the CT domain. At present, it is unclear whether the C-terminal fragment of CCN1 is still attached to the ECM/cell surfaces or released into the overlying media since we were unable to detect this band, either in the cell lysate or in the overlying media, on immunoblot analysis with the C-20 antibody.
CCN1 is a member of the newly established CCN family that plays an important regulatory role in development, wound healing, vascular diseases and cancer (12, 26, 37, 38) CCN1 is a growth factor-inducible immediate-early gene expressed at very low levels in various cells but rapidly induced by growth factors, proteases and other stimuli (11, 29, 39–41). Upon synthesis, the CCN1 protein is secreted and associates with cell surfaces and ECM (14). CCN1 has been shown to support cell adhesion, cell migration and augment growth factor-induced cell proliferation in various cell types through interactions with integrins and cell surface heparan sulfate proteoglycans (HSPGs) in cell type and context-specific manner (12, 15, 18–20, 42, 43). In endothelial cells, CCN1 was shown to induce cell migration through interactions with integrin αvβ3 (43). Recent studies identified a novel 20-residue sequence in the VWC domain of CCN1 (residues 116–135) as a functional binding site for integrin αvβ3 (44). Our present observation that the N-terminal fragment of plasmin-cleaved CCN1 supports endothelial cell migration fits with the above finding.
Recent studies showed that CCN1 is highly expressed in breast cancers and elevated levels of this protein in primary breast cancer is associated with more advanced disease (23, 24, 45). Tsai et al. (22) examined the expression of CCN1 in many human breast cancer cell lines and found that the expression of CCN1 is strongly correlated with the ability of breast cancer cells to invade in vitro and metastasize in vivo. Furthermore, overexpression of CCN1 in MCF-12A normal breast cells was shown to induce tumor formation and vascularization in nude mice (46). Similarly, expression of CCN1 cDNA under the regulation of constitutive promoter in RF-1 gastric adenocarcinoma cells significantly enhanced the tumor growth and vascularization (21). Since CCN1 protein has been shown to influence many cellular activities, it is possible that CCN1 could contribute to tumor growth by multiple mechanisms (26). However, the ability of CCN1 to promote angiogenesis is thought to be responsible primarily for tumor growth and vascularization (21). Angiogenesis requires the coordinated execution of a series of cellular processes, starting from the degradation of basement membranes and ECM surrounding the parent vessel to proliferation and migration of endothelial cells from the parent vessel toward an angiogenic signal to form new capillary sprouts. A number of studies showed that tumor cells, including breast cancer cells, express plasminogen activator uPA and tPA, and specific receptors for uPA and plasminogen (see rev (33)). In agreement with this, primary breast cancers were shown to express active enzymes capable of catalyzing plasmin formation (47). Hyperpermeability of local microvasculature associated with solid tumors (48) would allow the circulating plasminogen to enter into tumor tissues, which could be readily converted to plasmin on the tumor cell surface. Plasmin, by virtue of its ability to degrade ECM proteins and activate MMPs, can facilitate many steps involved in tumor growth and invasion. The ability of plasmin to liberate the biologically active CCN1 fragment adds a novel step to this complex process.
Plasmin cleavage of CCN1 may have wider biological implications because it permits the CCN1 partition into the soluble phase, rather than into the insoluble matrix, and therefore allows the CCN1 to diffuse freely within the tissue and interact with its plasma membrane receptors on various cell types. The 28 kDa fragment could act as either agonist or antagonist for full length CCN1 in a cell and context-specific fashion since CCN1 interacts with specific integrins in specific cell types, either requiring the CT domain or independent of the CT domain (12, 20, 42, 43). Since a number of pathological conditions are associated with upregulation of CCN1 (45, 49–51) and many cell types express the plasminogen activator/plasmin system, which could be upregulated further by pathological conditions (33, 52–54), it is likely that the truncated CCN1 fragments would be generated in vivo. If so, the 28 kDa fragment of CCN1 may serve as a marker for pathogenesis of disease, particularly cancer and cardiovascular diseases. Although shorter isoforms for other proteins in the CCN family have been detected in biological fluids (26, 27, 55), we are not aware of any published studies that examined the presence of truncated form of CCN1 in biological fluids either in normal or diseased conditions. Thus it will be interesting to test in the future whether the truncated form of CCN1 can be found in plasma or biological fluids of cancer patients and patients with cardiovascular diseases, and whether its levels correlate with disease parameters. It is also of interest to see whether plasmin cleaves other members of the CCN family. In this context, it may be relevant to note that there is 60 to 70% protein sequence homology among various member of CCN family in the region where plasmin cleaves the CCN1 (between aa 243 to 267 in CCN1) and the putative plasmin cleavage site R250↓Val251 (of CCN1) is preserved in 5 of 6 members of the CCN family (26).
We acknowledge the contribution of Veena Pappan in culturing cells.
Grant Support: This work was supported by NIH grant HL 65500.