β-Catenin is markedly induced in a murine model of an arteriovenous fistula: the effect of metalloproteinase inhibition

Abstract

Neointimal hyperplasia contributes to failure of hemodialysis arteriovenous fistulas (AVFs). Increased expression of matrix metalloproteinase (MMP)-9 occurs in AVFs, and MMP-9 is implicated in neointimal hyperplasia and vascular injury. Recent studies demonstrate that MMP-9, by degrading N-cadherin, leads to increased expression of β-catenin and β-catenin-dependent proliferation of smooth muscle cells. The present study examined this pathway in the venous limb of a murine AVF model. Western analyses demonstrate that, in this model, there is diminished expression of N-cadherin accompanied by increased expression of β-catenin, c-Myc, and proliferating cell nuclear antigen (PCNA). By immunohistochemistry, β-catenin and c-Myc localized to proliferating smooth muscle cells in the venous limb of the AVF. Increased expression of β-catenin was accompanied by augmented expression of phosphorylated (p)-glycogen synthase kinase (GSK)-3β, GSK-3β, and integrin-linked kinase. The administration of doxycycline suppressed MMP-9 expression but did not reduce venous histological injury in the AVF, or increase AVF patency assessed 6 wk after its creation. Doxycycline did not influence expression of β-catenin, c-Myc, GSK-3β, or integrin-linked kinase. Thus, in this vascular injury model, the upregulation of β-catenin cannot be readily attributed to MMP-9 upregulation; increased β-catenin expression may reflect either the upregulation of p-GSK-3β, GSK-3β, or integrin-linked kinase. This study provides the first exploration of β-catenin in an AVF, demonstrating substantial upregulation of this mitogenic signaling molecule and uncovering possible mechanisms that may account for such upregulation.

vascular access dysfunction impairs the health of patients with end-stage kidney disease who are maintained on chronic hemodialysis; such access dysfunction is also costly, since it imposes more than a billion dollars each year in health care costs (1, 2, 20, 35). The desired vascular access is the arteriovenous fistula (AVF), but, even for this access, the outlook as regard to functionality is decidedly poor: some 50% of all AVFs never develop adequately such that they can be used for chronic hemodialysis, and 25% of once usable AVFs will no longer be functional after two years (1, 2, 20, 35). There are thus increasing attempts to identify cellular processes and participants involved in AVF dysfunction such that new preventive and therapeutic approaches can be directed to this challenging and major clinical issue (1, 2, 20, 35).

A fundamental mechanism contributing to AVF dysfunction is neointimal hyperplasia occurring at the venous limb or the juxta-anastomotic regions of the AVF (1, 24, 34). Neointimal hyperplasia progressively encroaches upon the vascular lumen, diminishes blood flow, and promotes thrombus formation. While the basis for neointimal hyperplasia is unresolved, essential features of this lesion include proliferation and migration of smooth muscle cells and the phenotypic switch of smooth muscle cells from a contractile to a proliferative and synthetic phenotype (1, 24, 34).

A recently described microsurgical murine AVF model recapitulates the salient features of failing human hemodialysis AVFs, including proliferation of smooth muscle cells (23). The basis for such proliferation in this model and in human AVFs is currently unknown. We have demonstrated that, in this model, neointimal hyperplasia and patency rates were exacerbated by the deficiency of heme oxygenase-1 (HO-1) (23). In exploring the basis for this effect, we observed that such HO-1 deficiency failed to increase the already augmented expression of assorted mitogenic peptide growth factors but did increase the expression and activity of MMP-9 (23). Such effects on MMP-9 were of interest because of the following: MMP-9 expression is increased in rodent AVF models and in human AVFs (7, 8, 12, 23); genetic polymorphisms in certain MMPs, including MMP-9, are attended by altered patency of hemodialysis AVFs (26); increased MMP-9 expression can be vasculopathic (32); and, along with its other effects, MMP-9 is recognized as a mitogenic stimulus for smooth muscle cells (9).

The basis for the proliferative effects of MMP-9, however, is currently unresolved. Recent observations have suggested that the mitogenic actions of MMP-9 are due to the destabilizing effects of MMP-9 on the N-cadherin/β-catenin complex and the attendant increase in cellular levels of β-catenin, the latter recognized as a potent mitogenic signaling molecule (11, 16, 17, 37). In healthy smooth muscle cells, β-catenin is part of a complex connecting the actin cytoskeleton to N-cadherin; N-cadherin is a transmembrane protein that enables smooth muscle cells to interact. It has been suggested that increased MMP-9 expression causes proteolytic cleavage of N-cadherin, thereby destabilizing the complex of proteins associated with β-catenin; in turn, this leads to increased cellular levels of β-catenin and β-catenin-driven cell proliferation (11, 16, 17, 37).

The present study examined this hypothesis (namely: ↑MMP-9 → ↓N-cadherin → ↑β-catenin) as a mechanism for venous neointimal hyperplasia and injury in this AVF model. We thus assessed expression of these and other relevant proteins, and we employed doxycycline, a widely utilized matrix metalloproteinase (MMP)-9 inhibitor (4, 10, 14, 41), to inhibit MMP-9 in vivo. In this model, the assessment of standardized protein expression based on Western blots that are subjected to statistical analysis is quite challenging for at least two reasons. First, the amount of available venous tissue in this microsurgical murine AVF model is relatively limited; second, in our initial studies, we observed that several “housekeeping” proteins conventionally used for standardization of Western analysis are themselves substantially induced in this model, a finding likely attributable to the remarkable growth and remodeling occurring in the vein in the AVF. In preparatory studies, we first addressed and resolved these issues such that, in adequately sized groups of mice with AVFs and intact veins, analysis of standardized protein expression could be undertaken.

The current study thus examined issues in the pathobiology of AVF dysfunction that, to date, have not been investigated: specifically, whether increased MMP-9 expression that occurs in the AVF is accompanied by N-cadherin degradation and increased β-catenin expression; the pathogenetic role of MMP-9 in this model; whether β-catenin, a newly recognized vasculopathic molecule, merits attention as a possible contributor to AVF dysfunction; and, finally, whether the administration of doxycycline may provide a therapeutic approach for failing AVFs.

METHODS

Murine AVF model.

All studies were approved by the Institutional Animal Care and Use Committee of Mayo Clinic and performed in accordance with National Institutes of Health guidelines. Male C57BL/6J mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and used in an age range of 11–17 wk for all studies. An AVF was constructed, as described in our previous studies, by an end-to-side anastomosis between the right common carotid artery and external jugular vein (23). As in these previous studies, the contralateral jugular vein was employed as a control. In some studies, mice were treated with doxycycline (240 mg/l, catalog no. D9891; Sigma-Aldrich, St. Louis, MO) in their drinking water, commenced 1 day before the construction of the AVF, and maintained for the duration of the study (10). The doxycycline solution was made fresh every other day. After the construction of the AVF (1 and 6 wk), veins were harvested for Western analysis or histological/immunohistochemical processing, the latter performed on perfusion-fixed veins.

Immunohistochemical analysis.

For immunohistochemical staining, formalin-fixed, paraffin-embedded tissue sections were deparaffinized, and antigen retrieval was carried out with 1 mM EDTA, pH 8.0, in a 98°C steamer for 30–40 min. Slides were then treated with 3% H₂O₂ to inactivate endogenous peroxidase followed by incubation with Rodent Block M for 30 min (catalog no. RBM961L; Biocare Medical, Concord, CA). Primary rabbit polyclonal antibodies to c-Myc and β-catenin (catalog nos. sc-764 and sc-1496R, respectively; Santa Cruz Biotechnology, Santa Cruz, CA) were applied for 60 min at room temperature. Visualization was carried out using Rabbit on Rodent HRP Polymer (catalog no. RMR622G; Biocare Medical) with diaminobenzidine as the substrate followed by counterstaining with hematoxylin.

Western analysis.

Western blot analysis was performed on murine venous tissue as described in our prior studies (30, 36). Briefly, control or AVF external jugular veins were pooled (2/extraction) and homogenized in RIPA buffer containing protease and phosphatase inhibitors (Halt; Thermo Scientific, Rockford, IL). Lysate proteins (5–15 μg) were separated on 10% Tris·HCl Criterion gels (Bio-Rad Laboratories, Hercules, CA) and transferred to polyvinylidene difluoride membranes. After being blocked with 5% nonfat milk in TBS-Tween 20, membranes were incubated overnight at 4°C with primary antibodies. Antibodies directed against α-tubulin (catalog no. 2125), β-tubulin (catalog no. 2128), proliferating cell nuclear antigen (PCNA) (catalog no. 2586), N-cadherin (catalog no. 4061), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (catalog no. 2118), c-Myc (catalog no. 9402), and phospho (p)-glycogen synthase kinase (GSK)-3β (catalog no. 9336) were obtained from Cell Signaling Technology (Danvers, MA). Antibodies directed against β-catenin (catalog no. 610153), GSK-3β (catalog no. 610201), actin (catalog no. 612656), and integrin-linked kinase (ILK, catalog no. 611802) were from BD Transduction Laboratories (San Jose, CA). The antibody directed against MMP-9 (catalog no. AF909) was from R&D Systems (Minneapolis, MN). Horseradish peroxidase-conjugated goat anti-rabbit, goat anti-mouse, or rat anti-goat secondary antibodies were used, as appropriate, and detection was achieved with enhanced chemiluminescence (HyGLO; Denville Scientific, Metuchen, NJ). Protein bands were captured, and their relative intensities were measured with scanning densitometry (GS800 densitometer; Bio-Rad).

Statistics.

Results are expressed as means ± SE and considered statistically significant for P < 0.05. Student's t-test and the Mann-Whitney test for parametric and nonparametric data, respectively, were employed as appropriate.

RESULTS

In our initial studies, we observed that a number of conventionally employed housekeeping proteins were induced in this model, including β-actin (Fig. 1A), β-tubulin (Fig. 1A), and α-tubulin (Fig. 1B). In the AVF, veins undergo substantial remodeling and growth responses when exposed to the increased luminal hydraulic pressures and blood flow rates following the creation of the AVF. It is thus not unexpected that cytoskeletal proteins (such as β-actin, β-tubulin, and α-tubulin) are induced in the venous wall in the AVF. However, expression of GAPDH was not increased in this model (Fig. 1, A and B) and thus was used as the housekeeping protein in all Western analyses; densitometric readings reported in all subsequent Western analyses represent values corrected for accompanying expression of GAPDH.

Fig. 1.

Western analysis of housekeeping protein expression in the venous limb of the murine arteriovenous fistula (AVF) 1 wk after the creation of the AVF. A: equal amounts of lysate protein from venous segments of control and AVF veins were immunoblotted for β-actin (top), β-tubulin (middle), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (bottom), and the mean densitometric readings are presented below each group. B: equal amounts of lysate protein from venous segments of control and AVF veins were immunoblotted for α-tubulin (top) and GAPDH (bottom) with the mean densitometric readings presented below each group. For these and all subsequent Western analyses, each lane represents protein extract pooled from veins prepared from two mice for that condition.

Expression of N-cadherin in the venous limb of the AVF model was significantly reduced compared with such expression in the contralateral, control vein (Fig. 2); in the AVF, two fragments were observed on Western analysis, thereby indicating cleavage of the protein. Accompanying this reduction in expression of N-cadherin was increased expression of β-catenin in the venous limb of the AVF (Fig. 3), increased expression of the β-catenin-dependent, proliferative, proto-oncogene c-Myc (Fig. 4), and increased expression of a marker of cell proliferation, PCNA (Fig. 5).

Fig. 2.

Western analysis of N-cadherin protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from control and AVF veins was immunoblotted for N-cadherin. Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analysis. Expression of intact N-cadherin was significantly reduced in AVF veins compared with control veins (P < 0.05). The full-length N-cadherin protein band is denoted by an arrow.

Fig. 3.

Western analysis of β-catenin protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from control and AVF veins was immunoblotted for β-catenin. Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analysis. β-Catenin expression was significantly higher in AVF veins compared with control veins (P < 0.05).

Fig. 4.

Western analysis of c-Myc protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from control and AVF veins was immunoblotted for c-Myc. Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analysis. Expression of c-Myc protein was significantly increased in AVF veins compared with control veins (P < 0.05).

Fig. 5.

Western analysis of proliferating cell nuclear antigen (PCNA) protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from control and AVF veins was immunoblotted for PCNA. Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analysis. Expression of PCNA protein was significantly increased in AVF veins compared with control veins (P < 0.05).

These findings on Western analysis regarding increased expression of β-catenin and c-Myc in the AVF were further evaluated by immunohistochemistry. Increased expression of β-catenin and c-Myc was detected in smooth muscle cells in the venous limb, as shown in Fig. 6. Thus increased expression of MMP-9, as shown previously in this murine AVF model, is accompanied by decreased expression of N-cadherin and increased expression of β-catenin, c-Myc, and PCNA.

Fig. 6.

Immunohistochemical analysis of β-catenin and c-Myc protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Immunohistochemical localization of β-catenin protein in control vein (A) and the venous segment of the AVF (C) is shown. Similarly, localization of c-Myc protein in the control vein (B) and the venous segment of the AVF (D) is shown. Negative controls consisting of incubations with nonimmune serum did not show staining (data not shown). Original magnification ×200 for A–D.

To determine the functional significance of increased expression of MMP-9, we examined the effects of doxycycline in this model. Among the recognized effects of doxycycline is its capacity to suppress expression of MMP-9; administration of doxycycline is widely used as a method for inhibiting MMP-9 in rodent studies, evincing beneficial effects in models of vascular injury (10). Mice subjected to an AVF were administered either doxycycline in the drinking water or plain tap water, and, after 6 wk, patency of the AVF and histological injury were assessed. In this AVF model in mice receiving tap water, and at 6 wk after the creation of the AVF, the venous wall is variably affected by neointimal hyperplasia and intraluminal clot formation, both of which, if sufficiently severe, can lead to the loss of the patency of the AVF. Figure 7 illustrates the range of neointimal hyperplasia lesions observed in this model at 6 wk, including those that are mild and focal, moderate, and severe and circumferential; also shown is the presence of an intraluminal thrombus. After 6 wk of administration of doxycycline, there were no significant differences in the percent patency of the AVFs in mice maintained on plain tap water and those receiving doxycycline: patency rates of the AVFs in these groups, respectively, were 11 out of 19 (58%) and 15 out of 18 (83%), P = not significant. Histopathological evaluation revealed a similar range and degree of severity of neointimal hyperplasia and thrombus formation in the AVFs in both groups; Figure 8 illustrates comparably severe neointimal hyperplasia and clot formation in the AVF in mice receiving either plain tap water or doxycycline-containing water.

Fig. 7.

Representative histological lesions in the venous limb of the murine AVF in mice maintained on tap water and assessed 6 wk after the creation of the AVF. Histological sections of the venous limb were stained with trichrome blue and illustrate venous neointimal hyperplasia that is mild and focal (A), moderate and accompanied by an intraluminal thrombus (B), and severe and circumferential (C). The original magnification for A–C was ×100. D: a higher-powered view (original magnification ×200) of the lesion shown in B demonstrating the features of neointimal hyperplasia, including whorls of smooth muscle cells with associated extracellular matrix deposition.

Fig. 8.

Examination of the effects of doxycycline on venous histology in the murine AVF 6 wk after the creation of the AVF. Histological sections demonstrating appearance of the venous limb of the AVF in untreated (A) and doxycycline-treated (B) mice. Original magnification ×200. Tissue sections were stained with hematoxylin and eosin.

We also examined the effect of doxycycline on the expression of MMP-9, N-cadherin, β-catenin, and c-Myc in the AVF. As expected, the administration of doxycycline achieved a significant suppression of MMP-9 (Fig. 9). However, the expression of N-cadherin, β-catenin, and c-Myc in the AVF was not significantly altered following the administration of doxycycline (Fig. 10). Thus, although doxycycline substantially suppressed expression of MMP-9, it failed to change expression of N-cadherin, β-catenin, and c-Myc in the AVF.

Fig. 9.

Western analysis of the effect of doxycycline on matrix metalloproteinase (MMP)-9 protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from the venous segment of the AVF in mice with or without doxycycline treatment was immunoblotted for pro-MMP-9. Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analysis. Doxycycline treatment significantly decreased pro-MMP-9 protein expression in AVF veins (P < 0.05).

Fig. 10.

The effect of doxycycline on N-cadherin, β-catenin, and c-Myc protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from the venous segment of the AVF in mice with or without doxycycline treatment was immunoblotted for N-cadherin (top), β-catenin (middle), and c-Myc (bottom). Equivalency of protein loading was assessed by immunoblotting for GAPDH. Individual and mean standardized densitometric readings are provided below the Western analyses. No significant differences were observed in the expression of any of these proteins in the AVF veins with doxycycline treatment compared with untreated AVF veins [P = not significant (NS)]. The full length N-cadherin protein band is denoted by an arrow.

Because the upregulation of β-catenin persisted despite marked reduction in expression of MMP-9, we explored MMP-9-independent mechanisms that may account for the upregulation of β-catenin. Phosphorylation of GSK-3β is a major mechanism that leads to increased cellular expression of β-catenin, and β-catenin-dependent cell proliferation. When GSK-3β is phosphorylated at the serine-9 residue, GSK-3β can no longer phosphorylate β-catenin; phosphorylation of β-catenin directs β-catenin for degradation by the ubiquitin pathway and thus represents a mechanism that maintains low cellular levels of β-catenin. In the venous limb of the AVF, we observed markedly increased expression of p-GSK-3β (Fig. 11).

Fig. 11.

Western analysis of phosphorylated (p)-glycogen synthase kinase (GSK)-3β and GSK-3β protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from control and AVF veins was immunoblotted for p-GSK-3β (top) and GSK-3β (bottom). Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analyses. Expression of both p-GSK-3β and GSK-3β was significantly increased in AVF veins compared with control veins (P < 0.05).

The cellular effects of GSK-3β can be influenced by several mechanisms that include Wnt-dependent pathways, mitogenic growth factors, and ILK, all of which can increase cellular levels of β-catenin (5, 6, 15, 19, 27). The members of the Wnt family of ligands are important inducers of the canonical β-catenin pathway, but exploration of this family of ligands in the AVF is beyond the scope of the present investigation. Our prior studies demonstrate that several mitogenic growth factors, including transforming growth factor (TGF)-β1, are upregulated in this model (23), and some of these growth factors, such as TGF-β1, may signal through the ILK pathway (15, 25, 27). Because increased ILK is incriminated in proliferation of smooth muscle cells and vascular thickening and injury, as shown in many, but not all, studies (16, 18, 22), and because ILK can interface between integrins and growth factor receptors (15, 27), we thus examined the expression of ILK. In the venous limb of the AVF, we observed markedly increased expression of ILK (Fig. 12).

Fig. 12.

Western analysis of integrin-linked kinase (ILK) protein expression in the venous limb of the murine AVF 1 wk after the creation of the AVF. Protein extracted from control and AVF veins was immunoblotted for ILK. Equivalency of protein loading was assessed by immunoblotting for GAPDH, and individual and mean standardized densitometric readings are provided below the Western analysis. ILK protein expression was significantly increased in AVF veins compared with control veins (P < 0.05).

Increased GSK-3β, per se, has recently been shown to promote Wnt signaling (6, 40, 42, 43) and may thereby lead to increased cellular levels of β-catenin; in the AVF, we observed increased expression of GSK-3β (Fig. 11). Finally, doxycycline, which failed to lessen closure or histological injury in the AVF, did not affect expression of ILK, p-GSK-3β, or GSK-3β in the AVF (data not shown).

DISCUSSION

To the best of our knowledge, the present study is one of the few that explores β-catenin expression in venous disease, in general, and the only one that examines such expression in venous injury in an AVF model. An important structural component of the protein complex linking the actin cytoskeleton to cadherins, β-catenin is now recognized as a fundamental signaling molecule involved in processes that include cell differentiation, proliferation, migration, and survival (5, 15, 21, 27). Indeed, increased or aberrant expression of β-catenin is strongly implicated in the pathogenesis of cancer and fibroproliferative diseases and is now increasingly studied in cardiovascular disease (3, 11, 19, 29, 38, 39).

Cellular levels of β-catenin may be increased in several ways, one of which involves MMP-9-induced cleavage of N-cadherin (the main cadherin expressed by smooth muscle cells) and the attendant destabilization of the β-catenin-containing protein complex; such increased cellular expression of β-catenin can stimulate proliferation of smooth muscle cells (16, 17, 33, 37). Because we previously demonstrated that MMP-9 is significantly induced in this AVF model (23), our current studies explored the expression of N-cadherin and β-catenin. We found that the expression of N-cadherin was significantly diminished, with the protein appearing degraded into fragments, and was accompanied by prominent induction of β-catenin, the β-catenin-inducible proto-oncogene c-Myc, and the cell proliferation marker PCNA.

To determine whether this increased expression of β-catenin resulted from increased MMP-9 in this model, we examined the effect of inhibiting MMP-9 by the chronic administration of doxycycline. This agent effectively suppresses MMP-9 and has been shown to exert assorted vasoprotective effects, including the reduction of smooth muscle cell proliferation and migration in rodent models of arterial neointimal hyperplasia (4); the reduction of neointimal hyperplasia in a human vein graft stenosis model (28); the reduction in aneurysm formation in hemodialysis vascular access (14); and the reduction in aortic aneurysm formation in murine models (41). In our current studies, we confirmed marked suppression of MMP-9 expression in the venous limb of the AVF by doxycycline. Such suppression of MMP-9 by doxycycline failed to evince a beneficial effect in this AVF model, at least at the time point studied, thereby suggesting that the venous histological injury that occurs in this model cannot be readily attributed to the upregulation of MMP-9. In these studies, β-catenin expression was unaltered by doxycycline, thereby leaving open the possibility that the persisting upregulation of β-catenin (despite suppression of MMP-9 expression by doxycycline) may account for a lack of an effect of doxycycline on neointimal hyperplasia.

Our findings thus question the contribution of MMP-9 to venous injury in experimental AVFs and dysfunctional human AVFs. Interestingly, observations that just appeared in the literature demonstrate that genetic polymorphisms in MMP-9, which are attended by reduced activity, are associated with reduced patency of hemodialysis AVFs (26); these findings led to the speculation that reduced MMP-9 activity, by impairing proteolytic degradation of extracellular matrix, promotes matrix buildup, access stenosis and thrombosis, and attendant AVF failure (26). It is thus possible that the presumptive vasculopathic effects of MMP-9 may be offset by its salutary effects such as the degradation of extracellular matrix.

The finding that MMP-9 can be markedly suppressed, but β-catenin expression remains unaltered when doxycycline is administered, suggests that β-catenin upregulation in this model is not driven by MMP-9. This lack of dependency led us to consider other mechanisms that may underlie the upregulation of β-catenin. A critical governor of cellular levels of β-catenin is GSK-3β (5, 15, 27). In quiescent cells, GSK-3β phosphorylates β-catenin, and such phosphorylation of β-catenin leads to its degradation by the ubiquitin pathway; cytosolic levels of β-catenin are thus maintained at relatively low levels such that, in quiescent cells, nuclear translocation of β-catenin and β-catenin-dependent gene transcription do not occur. Phosphorylation of GSK-3β at the serine-9 residue, however, vitiates the ability of GSK-3β to phosphorylate β-catenin and thereby consign β-catenin for proteosomal degradation. Thus, when GSK-3β is increasingly phosphorylated, increasing amounts of β-catenin can escape degradation by the proteosome, enter the nucleus, and elicit genes that induce cell proliferation. In the AVF, the expression of p-GSK-3β was significantly increased; such increased expression of p-GSK-3β may contribute to increased expression of β-catenin.

Phosphorylation of GSK-3β can occur through several mechanisms, one of which is ILK, a kinase known to promote the proliferation of smooth muscle cells via β-catenin-dependent mechanisms (5, 15, 27). Our present study demonstrates that ILK is significantly induced in this model, thereby raising the possibility that this kinase may serve as a mechanism for phosphorylating GSK-3β. Interestingly, ILK can also promote β-catenin signaling by GSK-3β-independent mechanisms (31), thereby providing another linkage between ILK and β-catenin we observed in this model.

Increased cellular content of β-catenin and β-catenin-dependent gene transcription is a major downstream effect of the canonical Wnt pathway. The Wnt pathway is an important signaling pathway that controls tissue development, growth, and injury, and is instigated by a number of Wnt ligands (5, 15, 27, 29, 38); for example, in the kidney, the Wnt/β-catenin pathway was recently implicated in podocyte injury, epithelial-mesenchymal transition, and renal fibrosis (13, 21, 25, 27). Interactions of Wnt ligands with a receptor complex consisting of a Frizzled protein and low-density lipid receptor protein 5/6 initiate this signaling pathway, which culminates in increased cellular expression of β-catenin and increased β-catenin-dependent gene transcription (5, 15, 27, 29, 38). Studied for a number of years in cancer biology and fibroproliferative disorders, the Wnt signaling pathway is now increasingly examined in the pathogenesis of vascular injury (29, 38). This system, as yet unexplored in models of vascular access dysfunction, may account for increased β-catenin we observed in the AVF. Relevant to Wnt signaling in this model is the increased expression of GSK-3β we observed in the AVF (Fig. 11): recent persuasive evidence indicates that increased expression of membrane-associated GSK-3β promotes Wnt signaling (6, 40, 42, 43). It is thus possible that such increased expression of GSK-3β may promote Wnt signaling and in turn increased β-catenin expression.

In summary, we demonstrate that β-catenin, a critical signaling molecule in diverse biological processes, is substantially induced in a murine model of an AVF. Accompanying this upregulation of β-catenin are “downstream” events that include increased expression of proteins such as c-Myc and PCNA that are involved in proliferative responses. We also examined “upstream” events which may be involved in mediating the upregulation of β-catenin. We demonstrate that increased expression of β-catenin is not driven by increased MMP-9, the latter representing a molecule broadly induced in rodent AVF models and human AVFs. The present observations thus uncover in the venous segment of an AVF the recruitment of an important and topical signaling molecule not previously recognized or explored in this disease model, and one that may contribute to neointimal hyperplasia and other processes that compromise the longevity of an AVF. Finally, these studies set the stage for the examination of the Wnt system in the AVF.

GRANTS

These studies were supported by National Institutes of Health Grants DK-70124, DK-47060, and HL-55552 (K. A. Nath, Z. S. Katusic, and J. P. Grande).

DISCLOSURES

No conflicts of interest are declared by the authors.