|Home | About | Journals | Submit | Contact Us | Français|
Bone morphogenetic protein 7 (BMP7) stimulates renal branching morphogenesis via p38 mitogen-activated protein kinase (p38MAPK) and activating transcription factor 2 (ATF-2) (M. C. Hu, D. Wasserman, S. Hartwig, and N. D. Rosenblum, J. Biol. Chem. 279:12051-12059, 2004). Here, we demonstrate a novel role for integrin-linked kinase (ILK) in mediating renal epithelial cell morphogenesis in embryonic kidney explants and identify p38MAPK as a target of ILK signaling in a cell culture model of renal epithelial morphogenesis. The spatial and temporal expression of ILK in embryonic mouse kidney cells suggested a role in branching morphogenesis. Adenovirus-mediated expression of ILK stimulated and expression of a dominant negative ILK mutant inhibited ureteric bud branching in embryonic mouse kidney explants. BMP7 increased ILK kinase activity in inner medullary collecting duct 3 (IMCD-3) cells, and adenovirus-mediated expression of ILK increased IMCD-3 cell morphogenesis in a three-dimensional culture model. In contrast, treatment with a small molecule ILK inhibitor or expression of a dominant negative-acting ILK (ILKE359K) inhibited epithelial cell morphogenesis. Further, expression of ILKE359K abrogated BMP7-dependent stimulation. To investigate the role of ILK in BMP7 signaling, we showed that ILK overexpression increased basal and BMP7-induced levels of phospho-p38MAPK and phospho-ATF-2. Consistent with its inhibitory effects on IMCD-3 cell morphogenesis, expression of ILKE359K blocked BMP7-dependent increases in phospho-p38MAPK and phospho-ATF-2. Inhibition of p38MAPK activity with the specific inhibitor, SB203580, failed to inhibit BMP7-dependent stimulation of ILK activity, suggesting that ILK functions upstream of p38MAPK during BMP7 signaling. We conclude that ILK functions in a BMP7/p38MAPK/ATF-2 signaling pathway and stimulates epithelial cell morphogenesis.
Renal branching morphogenesis (RBM), is defined as growth and branching of the ureteric bud (UB) and its daughter collecting ducts. RBM is dependent on inductive tissue interactions with the surrounding metanephric mesenchyme, or blastema, which secretes growth factors that induce ureteric bud invasion and branching (38). Reciprocal signals are elaborated by the epithelial bud, thereby providing survival and differentiation signals for cells of the nephrogenic mesenchyme. One such signal implicated in RBM is bone morphogenetic protein 7 (BMP7) (11, 31), a member of the transforming growth factor β superfamily. Expression of BMP7 in the ureteric bud and metanephric blastema is consistent with a physiological role during RBM (12, 39); moreover, BMP7 null mice display profound defects of RBM (11, 30).
While the presence of reciprocal tissue interactions limits the ability to interpret the primary effects of BMP7 in vivo, we have identified BMP7 functions in the inner medullary collecting duct 3 (IMCD-3) cell culture model of RBM (37, 39). IMCD-3 cells form tubule progenitors (i.e., morphogenesis) within 48 h of being cultured in collagen gels and respond to serum or purified growth factors in a manner identical to branching observed in embryonic kidney explants (3, 39). BMP7 acts to control IMCD-3 cell morphogenesis in a complex manner. High doses (>0.5 nM) inhibit morphogenesis in an Smad1-dependent manner, whereas low doses (<0.5 nM) stimulate progenitor formation in a Smad1-independent manner via effects on cell proliferation (37, 39). We have also demonstrated that stimulatory doses of BMP7 activate p38 mitogen-activated protein kinase (p38MAPK) and its target, activating transcription factor 2 (ATF-2), and that inhibiting p38MAPK activity blocks IMCD-3 morphogenesis (22). Our demonstration that Smad1 activity and p38MAPK/ATF-2 activation are inversely related suggests a cellular mechanism that integrates stimulatory and inhibitory BMP7 signal transduction pathways.
Integrins mediate essential cell-extracellular matrix (ECM) interactions during mammalian development but also transduce signals regulating cell proliferation, differentiation, and survival. In addition to growth factors, integrins contribute to epithelial-mesenchymal interactions during kidney organogenesis. Genetic ablation of the α8 integrin precludes formation of functional α8β1 integrin in the mesenteric mesenchyme, resulting in severe defects of ureteric bud branching (34). The integrin-linked kinase (ILK) is an intracellular protein serine/threonine kinase that coordinates signaling by integrins and growth factors (6-8, 19), including insulin-like growth factor 1 (IGF-1) (29, 32), nerve growth factor (33), platelet-derived growth factor (2), and vascular endothelial growth factor (24, 44), in a variety of cell types. ILK binds directly to the cytoplasmic domains of β integrin subunits (19). Additional protein interactions may function to physically link ILK with receptor tyrosine kinases (RTKs). ILK binds to PINCH, a LIM-only adaptor protein (46, 49). PINCH binds to another adaptor protein, Nck2, an SH2/SH3-containing protein that associates with ligand-activated epidermal growth factor (EGF) and platelet-derived growth factor receptors (47). In addition, ligand activation of these RTKs stimulates phosphoinositide 3′-OH kinase (PI 3-K) activity, and genetic or pharmacologic inhibition of PI 3-K abolishes both RTK- and integrin-mediated ILK activation (6, 7). The mechanism of ILK activation involves the major lipid product of PI 3-K activity, phosphoinositol-3,4,5-phosphate, which likely activates ILK by binding to the pleckstrin homology-like domain (9). Thus, PI 3-K dependent ILK activity exerts contextual effects governing epithelial-mesenchymal transition (43), myoblast differentiation, neurite outgrowth, and vascular morphogenesis.
BMP7 and α8 integrin regulate RBM in vivo, suggesting integrated growth factor and cell adhesion signaling during collecting duct morphogenesis. Here, we report that ILK displays an expression pattern in the embryonic mouse kidney that is consistent with a role in RBM. Infection of embryonic mouse kidney explants with adenovirus (Ad) expressing a dominant negative mutant of ILK markedly impaired formation of collecting ducts, further implicating ILK signaling in RBM. We therefore investigated the role of ILK in mediating BMP7-dependent renal epithelial cell morphogenesis, using the IMCD-3 model system. Treatment of IMCD-3 cells with stimulatory concentrations of BMP7 rapidly (<15 min) induced activation of ILK. Moreover, adenovirus-mediated expression of ILK markedly stimulated the formation of tubule progenitors and increased levels of phospho-p38MAPK and phospho-ATF-2. Conversely, a small molecule inhibitor of ILK, KP-392, abrogated IMCD-3 morphogenesis in three-dimensional cultures. In addition, infection with a dominant negative ILK mutant blocked BMP7-induced morphogenesis and inhibited the phosphorylation of p38MAPK and ATF-2. Our results identify a novel BMP7/ILK/p38MAPK/ATF-2 signaling pathway controlling epithelial cell morphogenesis.
Embryonic day 13.5 (E13.5) mouse kidney sections were processed for immunohistochemistry as described previously by us, with slight modifications (21). Paraffin-embedded sections were deparaffinized and heated in 10 mM citrate (pH 6.0) with a microwave oven for antigen retrieval. After quenching endogenous peroxidase activity with H2O2 for 10 min, sections were incubated overnight at 4°C with ILK primary antibody (Upstate Biotechnology, Inc.; catalogue no. 06-592) at a 1:35 dilution.
Embryonic kidneys were surgically dissected from E13.5 pregnant mice and cultured as previously described (22). To evaluate the effect of ILK on RBM, adenoviruses expressing green fluorescent protein (GFP), ILK, or a dominant negative variant of ILK (28) were added to culture medium at 106, 105, and 104 PFU/ml, respectively, for 5 days. To verify infection efficiencies, fluorescent photos were taken just before explants were fixed for Dolichos biflorus agglutinin (DBA) staining. Selective staining of ureteric bud-derived structures in the whole-mount kidney specimens was achieved with fluorescein isothiocyanate-conjugated DBA (20 μg/ml) (Vector Laboratories, Burlington, Ontario, Canada) as previously described (16).
The IMCD-3 cell line is derived from the terminal inner medullary collecting duct of the simian virus 40 transgenic mouse. The IMCD-3 cell line retains several differentiated characteristics of the nephron, as previously described (40), and has been used as an in vitro model of collecting duct morphogenesis. IMCD-3 cells were grown in monolayers and maintained in Dulbecco's modified Eagle's medium-Ham's F12 medium (DMEM-F12) supplemented with 5% fetal bovine serum (Sigma), penicillin (100 U/ml), and streptomycin (100 U/ml) in 5% CO2 at 37°C.
To assay IMCD-3 cell morphogenesis, cells were suspended in type I collagen gels in 96-well tissue culture plates as previously described (39). Collagen gels were prepared on ice by mixing 4 μl of 1 M N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (Sigma), 8 μl of 1 M NaHCO3, 40 μl of DMEM-F12, 200 μl of 3.5-mg/ml rat type I collagen (Collaborative Biomedical Products), and 50,000 IMCD-3 cells. Aliquots (each, 50 μl/well) were seeded in 96-well culture plates. Gels were solidified at 37°C, and then 100 μl of medium containing 5% fetal bovine serum was added to each well. Cultures were maintained at 37°C in 5% CO2. BMP7 and epidermal growth factor stock solutions were prepared in DMEM-F12 and added at the indicated concentrations to the 5% fetal bovine serum-containing culture medium of newly established cultures of IMCD-3 cells. Cells were cultured in monolayers for biochemical studies or embedded in collagen gels for morphogenetic studies. After 48 h, gels were fixed in 4% formaldehyde in phosphate-buffered saline for 10 min at room temperature. Fixed gels were then washed four times in phosphate-buffered saline and imaged by differential interference contrast microscopy. Eight microscopic fields of equivalent dimensions were randomly selected and photographed at 100× magnification. Morphogenesis was quantified by counting the number of elongated, linear structures in these fields, presented as the number of independent linear structures per area of standard dimension. Assays were standardized for the number of cells seeded into the each collagen gel, as we have previously published (37, 39). The statistical significances between treatment differences were determined by a two-tailed, nonpaired Student's t test, using Prism software, version 3.0 (GraphPad Software, Inc.).
ILK immune complex kinase assays were carried out as described previously (19, 29). Protein concentrations were determined by Bradford assays (Bio-Rad, Richmond, Calif.). Cell lysates (0.25 to 1.0 mg of protein) were immunoprecipitated with 1 μg of affinity purified rabbit anti-ILK (Upstate Biotechnology, Inc.; catalogue no. 06-592) overnight at 4°C with rotation. Protein A-Sepharose (Sigma), preswollen in NP-40 lysis buffer (150 mM NaCl, 1% [vol/vol] NP-40, 0.5% [wt/vol] sodium deoxycholate, 50 mM HEPES [pH 7.4], 1 μg of leupeptin/ml, 1 μg of aprotinin/ml, 3 mM phenylmethylsulfonyl fluoride) was added for 2 h at 4°C to capture the antibodies. After two washes with NP-40 lysis buffer and two washes with kinase wash buffer (10 mM MgCl2, 10 mM MnCl2, 50 mM HEPES [pH 7.5], 0.1 mM sodium orthovanadate, 1 mM dithiothreitol), assays were performed directly on the protein A beads in a 25-μl reaction volume containing 10 mM MgCl2, 10 mM MnCl2, 50 mM HEPES (pH 7.5), 0.1 mM sodium orthovanadate, 2 mM sodium fluoride, 5 μCi of γ-32P (Amersham, Piscataway, N.J.) and 2.5 μg of myelin basic protein (MBP) as substrate (Upstate Biotechnology, Inc.). Incubation was for 30 min at 30°C. The reaction was stopped with 10 μl of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) nonreducing stop buffer and heated for 5 min at 95°C. Phosphorylated MBP bands were visualized by autoradiography of dried SDS-10% PAGE gels, followed by quantitation in a PhosphorImager (Molecular Dynamics).
Adenoviruses expressing either wild-type or dominant negative ILK were constructed for expression in kidney explants and in IMCD-3 cells. Briefly, the full-length ILK coding sequence or the dominant negative point mutant, E359K, including a C-terminal myc-His epitope tag, was amplified from a pcDNA3.1 construct. EcoRV was used to digest the ILK-myc-His expression fragments for subcloning into pAd Trac (pAd Easy kit; Clontech). Bicistronic GFP and ILK expression allowed infection efficiencies to be confirmed visually by fluorescent microscopy. Viruses were amplified and CsCl gradient purified and titers were determined on HEK293 cells as described previously (28).
For analysis of protein levels by Western blotting, SDS-PAGE gels were transferred to polyvinyldifluoride membranes (Immobilon-P; Millipore, Bedford, Mass.) in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol). Membranes were blocked in 5% milk in TBST (20 mM Tris [pH 7.5], 500 mM NaCl, 0.1% Tween 20). Affinity-purified primary polyclonal or monoclonal antibodies were used at a concentration of 1 μg/ml in Tris-buffered saline with 1% (wt/vol) bovine serum albumin (Fraction V; Sigma). Secondary antibodies were goat anti-mouse or anti-rabbit coupled to peroxidase, used at a dilution of 1/20,000 in TBST. Protein bands were visualized by chemiluminescence (ECL; Amersham) and exposure to Kodak X-Omat film, and signals were quantified with Kodak ID digital imaging software, version 2.0.2. Antibodies specific to phospho-glycogen synthase kinase 3β (phospho-GSK3β) (pSer9), protein kinase B (PKB) (Ser473), p38MAPK, ATF-2, phospho-p38MAPK (Thr180/182), and phospho-ATF-2 (Thr69/71) were purchased from Cell Signaling Technology, and antibodies recognizing total GSK3α/β and PKB were obtained from Transduction Labs. A monoclonal antibody to the c-myc epitope was purchased from Santa Cruz Biotechnology.
Immunohistochemical analysis of embryonic mouse kidney revealed strong ILK immunoreactivity in the uretric bud and metanephric mesenchyme of E13.5 kidneys (Fig. (Fig.1).1). By E17.5, ureteric bud cells have differentiated into collecting duct cells in the cortex and medulla. At this stage as well, ILK is strongly expressed throughout the collecting duct system. Thus, the temporal and spatial pattern of ILK in embryonic kidney is consistent with a functional role for ILK during RBM. To test whether ILK activity plays a role in renal morphogenesis, we infected embryonic mouse kidney explants with adenoviruses carrying wild-type ILK (Ad-ILKWT), or a dominant negative ILK mutant (Ad-ILKE359K). We confirmed that control, Ad-ILKWT, and Ad-ILKE359K viruses infected the kidney explants efficiently, judging by the expression of virally encoded GFP in whole-mount explants (Fig. (Fig.2A).2A). Interestingly, we noted that Ad-ILKWT-infected kidneys appeared slightly larger and those infected with Ad-ILKE359K appeared smaller than control virus (AdCONTROL)-infected explants. This is consistent with a role for ILK in regulating cell proliferation in the intact kidney. After 5 days in culture, formation of collecting ducts was readily apparent in the control virus-infected kidneys. Infection with the Ad-ILKWT virus stimulated an increase in the number of UB branches formed, and UB formation was markedly inhibited in kidneys that had been infected with Ad-ILKE359K (Fig. (Fig.2B).2B). These expression and functional data suggest an important role for ILK signaling in RBM.
The cellular complexity and dynamic epithelial-mesenchymal interactions in the developing kidney make it difficult to discriminate direct and indirect effects of growth factors. To directly examine a role of ILK in renal cell morphogenesis, we first tested whether BMP7 induced ILK activity in IMCD-3 cell cultures. We used EGF as a positive control for ILK stimulation and induction of tubule progenitors, since it potently induces ILK activity in other epithelial cells (unpublished data). Growth factor-dependent morphogenesis was quantified by culturing the IMCD-3 cells in three-dimensional collagen gels containing 5% serum, 0.25 nM BMP7, or 20 ng of EGF/ml. In this morphogenetic assay system, IMCD-3 cells forms structures at 48 h comprising two or more cells, with cellular processes that form the basis for multicellular branches, which are observed after 5 to 7 days in culture. Again, both these growth factors enhanced IMCD-3 morphogenesis significantly over serum-induced (control) levels. BMP7 induced a 1.4-fold increase, and EGF induced a 2-fold increase, over serum-induced levels (Fig. (Fig.3A).3A). We then assayed for growth factor activation of ILK, using immune complex kinase assays. BMP7 (0.25 nM) and EGF (20 ng/ml) induced ILK activity to from five- to sixfold over unstimulated levels (Fig. (Fig.3B).3B). Thus, ILK activation is an early event during BMP7-dependent IMCD-3 cell morphogenesis, and these results therefore raised the question of whether ILK acts to stimulate morphogenesis. We directly tested whether ILK promotes morphogenesis by infecting IMCD-3 cells with Ad-ILKWT. Infection with Ad-ILKWT stimulated tubule progenitor formation 2.8-fold relative to serum-induced levels (Fig. (Fig.4),4), indicating that ILK is a positive mediator of IMCD-3 morphogenesis. We also noted an increase in cell numbers in Ad-ILKWT-infected cultures; thus, like BMP7, ILK controls both proliferation and morphogenesis of renal epithelial cells. To test whether morphogenesis requires ILK activity, we inhibited ILK by infecting IMCD-3 cells with adenovirus expressing a dominant negative ILK mutant, ILKE359K. We (28, 32) and others (33, 45) have shown that the ILKE359K mutant exerts dominant inhibition of growth factor- or ECM-induced ILK activity in a number of cell lines. Infection of IMCD-3 cells with the Ad-ILKE359K virus inhibited tubule progenitor formation by approximately twofold over serum control levels (Fig. (Fig.4).4). Thus, ILK is sufficient to induce IMCD-3 morphogenesis, and inhibition of ILK blocks this response.
We next tested whether ILK mediates BMP7-dependent morphogenesis by quantifying BMP7-induced formation of tubule progenitors in Ad-ILKE359K-infected cultures. BMP7-induced morphogenesis in Ad-ILKE359K-infected cells was decreased by approximately sixfold, relative to that of AdCONTROL virus-infected cells. Similarly, EGF-stimulated morphogenesis was decreased about threefold by Ad-ILKE359K, relative to that of control virus-infected cultures (Fig. (Fig.5A).5A). These results demonstrate that ILK activity mediates morphogenetic signaling by BMP7.
As a second, independent method of inhibiting ILK we used KP-392, a small molecule that selectively inhibits ILK activity (18). IMCD-3 cells were cultured in collagen gels supplemented with 5% serum or 20 ng of EGF/ml, with or without 5 μM KP-392 (50). After 48 h, morphogenesis was visually quantified (Fig. (Fig.5B).5B). Serum stimulated morphogenesis was inhibited by about 2 fold, and EGF stimulation was inhibited by 4.5 fold in KP-392-pretreated cells. We could not determine an effect of KP-392 on BMP7-dependent morphogenesis, since treatment with a dimethyl sulfoxide vehicle at the appropriate control concentration abrogated BMP7 activity. These results, coupled with those showing stimulation of tubule progenitors by Ad-ILK, indicate that ILK activity plays a role in IMCD-3 morphogenesis.
ILK activation, by growth factors or integrin-mediated cell adhesion, is dependent on the activity of PI 3-K. Thus, genetic or pharmacologic inhibition of PI 3-K signaling blocks activation of ILK by a variety of stimuli in epithelial cells, platelets, and myoblasts (9, 13, 32, 35, 36). To test for the involvement of PI 3-K activity in the BMP7 stimulation of ILK, we used the selective PI 3-K inhibitor LY294002. We pretreated IMCD-3, cultured in collagen gels, with LY294002 and then quantified induction of tubule progenitors by BMP7 and EGF. LY294002 pretreatment inhibited BMP7 induction of IMCD-3 tubule progenitors by 2.3 fold and inhibited EGF induction by 3.3 fold (Fig. (Fig.6A),6A), suggesting that PI 3-K activity is required in BMP7- and ILK-dependent morphogenesis. We therefore tested whether LY294002 inhibited BMP7-dependent ILK activation by two independent assays of ILK activity. LY294003 effectively suppressed BMP7-activated GSK3β Ser9 phosphorylation, a known cellular target of ILK (Fig. (Fig.6B).6B). Similarly, pretreatment of IMCD-3 cells with LY294002 inhibited ILK activity as measured by ILK immune complex kinase assays (Fig. (Fig.6B).6B). Our results indicate that BMP7 stimulation of both morphogenesis and ILK activity requires PI 3-K activity.
We recently reported that stimulatory doses of BMP7 activate p38MAPK in IMCD-3 cells and that the selective p38MAPK inhibitor, SB203580, blocks BMP7-dependent morphogenesis (22). In the light of our present results showing ILK-dependent IMCD-3 morphogenesis, we examined whether ILK stimulates p38MAPK activity and whether blocking ILK signaling with a dominant negative ILK mutant blocks stimulation of p38MAPK by BMP7. Ligand-induced phosphorylation on Thr180 and Tyr182 activates p38MAPK. To test for ILK-dependent stimulation of p38MAPK/ATF-2, we assayed levels of phospho-p38MAPK in IMCD-3 cultures infected with AdCONTROL, Ad-ILKWT, or the dominant negative Ad-ILKE359K mutant virus (Fig. (Fig.7).7). Forty-eight hours after infections, cells were harvested and analyzed by Western blotting with phosphospecific antibodies to p38MAPK (Thr180/Tyr182). Cells were treated for 1 h with BMP7 to induce p38MAPK phosphorylation (22). As a positive control, we assessed ILK-dependent GSK3β Ser9 phosphorylation in the same lysates. Phospho-p38MAPK levels were increased in Ad-ILKWT- but not Ad-ILKE359K- or AdCONTROL-infected cultures. Infection with Ad-ILK also stimulated phosphorylation of GSK3β on Ser9, whereas cells infected with AdCONTROL or Ad-ILKE359K showed no increase in pSer9 levels (not shown). BMP7 also stimulated p38MAPK phosphorylation in control, but not in Ad-ILKE359K infected cells. As a positive control, phosphorylation of the known ILK target, PKB Ser473, was stimulated by Ad-ILKWT and blocked by Ad-ILKE359k (Fig. (Fig.7A).7A). These data indicate that ILK can activate p38MAPK independently of ligand and that induction of p38MAPK by BMP7 is ILK dependent.
Activating phosphorylation of the transcription factor, ATF-2, on Thr71 is mediated by p38MAPK. Our previous work showed that BMP7 stimulates ATF-2 phosphorylation in IMCD-3 cells and that pretreatment of cells with the p38MAPK inhibitor SB203580 blocks ATF-2 phosphorylation (22). We treated IMCD-3 cells with BMP7 for 15 and 60 min and found that although ATF-2 phosphorylation was evident at 60 min, it was not appreciably induced at 15 min (data not shown). We reasoned that BMP7 signaling at this relatively early time point would be more robust in ILK-expressing cells and therefore tested whether BMP7-dependent ATF-2 phosphorylation was affected by increased ILK expression. IMCD-3 cells infected with either Ad-ILKWT or Ad-ILKE359K viruses were treated for 15 min with BMP7 and assayed for ATF-2 phosphorylation. Ad-ILKWT but not Ad-ILKE359K stimulated ATF-2 Thr71 phosphorylation (Fig. (Fig.7B),7B), similar to p38MAPK (Fig. (Fig.7A).7A). BMP7 treatment for 15 min potentiated ATF-2 phosphorylation in Ad-ILKWT- but not in Ad-ILKE359K-infected cells (Fig. (Fig.7B).7B). Thus, increased ILK expression accelerated BMP7-induced ATF-2 phosphorylation, further indicating that ILK acts in the BMP7/p38/ATF-2 signaling axis.
Our data suggested that ILK lies upstream of p38MAPK in the BMP7 morphogenetic pathway; therefore, we tested the effects of SB203580 p38MAPK inhibitor on BMP7-induced ILK activity. Cells were pretreated for 60 min with 0 or 10 μM SB203580 and subsequently treated for 15 min with 0.25 nM BMP7. As a measure of ILK activation, we assayed phospho-GSK3β (pSer9) levels. SB203580 pretreatment did not inhibit BMP7-induced Ser9 phosphorylation (Fig. (Fig.7C).7C). These lysates were also subjected to ILK immune complex kinase assays (28, 29), which showed lack of inhibition of BMP7-induced ILK activity by SB203580, whereas LY294002 inhibited ILK activity (Fig. (Fig.7D).7D). These results place p38MAPK downstream of ILK activation by BMP7. Together with the results showing inhibition of BMP7-induced p38MAPK phosphorylation and morphogenesis by ILKE359K, these data identify a novel BMP7/ILK/p38MAPK/ATF-2 signaling axis mediating epithelial cell morphogenesis.
The results presented here implicate ILK as an effector of BMP7-dependent epithelial cell morphogenesis. We find that ILK is abundantly expressed in medullary collecting duct epithelium of E13.5 and E17.5 mouse kidneys, indicating it could mediate RBM. Such a role for ILK is further supported by our observation that expressing dominant negative ILK inhibits formation of ureteric bud branches in embryonic kidney explants. We have shown a direct effect of ILK in mediating renal epithelial cell morphogenesis. BMP7 induces ILK activity during in vitro morphogenesis of IMCD-3 collecting duct cells. Inhibition of ILK by either a small molecule or a dominant negative ILK mutant blocks IMCD-3 tubule progenitor formation. Consistent with this, adenovirus-mediated expression of ILK promotes IMCD-3 morphogenesis, suggesting that ILK is sufficient for induction of epithelial cell morphogenesis. Furthermore, our data showed that ILK activity stimulates phosphorylation of p38MAPK and ATF-2 during BMP7-dependent morphogenesis and that a dominant negative ILK mutant efficiently blocks p38MAPK/ATF-2 activation. Inhibition of p38MAPK activity does not affect BMP7-dependent activation of ILK, consistent with ILK acting upstream of p38MAPK in this morphogenetic pathway.
Using both embryonic kidney explants and the IMCD-3 cell culture model, we have previously shown opposite, dose-dependent effects of BMP7 in the regulation of epithelial cell morphogenesis (39). Low doses (<0.5 nM) stimulate and high doses (>0.5 nM) inhibit formation of IMCD-3 tubule progenitors. Interestingly, induction of ILK activity in our hands is maximal at the stimulatory dose (0.25 nM) of BMP7 (data not shown). We previously reported that inhibitory doses of BMP7 induce phosphorylation of Smad1 and formation of Smad1/Smad4 protein complexes and that a Smad1 dominant negative mutant selectively blocks inhibitory signaling (21, 37). Interestingly, the dominant negative Smad1 mutant potentiates BMP7 activation of p38MAPK/ATF-2 (22), suggesting that Smad1 acts to restrict ILK-p38MAPK signaling. We have previously shown that stimulatory and inhibitory pathways function in parallel in IMCD-3 cells (17); thus, it is likely that dose-dependent BMP7 signals are integrated at a point downstream of ILK-p38MAPK. However, a number of morphogenetic growth factors work through ILK in IMCD-3 cells, including hepatocyte growth factor (HGF) (our unpublished data), indicating that ILK is a point of stimulatory signal convergence. We do not know if ILK regulates RBM in vivo, as ILK null embryos die at E4.5, well before the onset of kidney development (42). Based on the results presented here, we speculate that developmentally regulated expression of morphogens, such as BMP7 and/or HGF (5), determines ILK-dependent ureteric bud branching during kidney organogenesis.
Our data place ILK upstream of p38MAPK in the BMP7 stimulatory pathway and suggest that BMP7 activates ILK in a PI 3-K-dependent manner. As discussed above, PI 3-K activity links a diverse complement of cell surface growth factor receptors and integrins to ILK signal transduction (6, 7). Both in vitro and in vivo studies indicate that activin-like ALK2 and ALK3 receptors mediate inhibitory BMP signaling in kidney epithelia (17, 21, 38); however, we do not know if stimulatory signaling is downstream of these ALKs or is mediated by a distinct receptor. Critical questions, thus, relate to the identify of the BMP7 stimulatory receptor and of the molecules that link it to PI 3-K activation. In this context, it will be important to determine whether the signaling adaptor, Nck2, plays a role in BMP7-ILK signaling.
Although ILK signaling has largely been studied in terms of ILK-PKB or ILK-GSK3β interactions, studies have highlighted the importance of the ILK/p38MAPK signaling axis in regulating cell behavior. Indeed, the ILK/p38MAPK pathway is implicated in different morphogenetic events, since (as with IMCD-3 morphogenesis) inhibition of either kinase activity blocks induction of neurite outgrowth in mouse neuroblastoma cells (23). D'Amico et al. also showed that point mutation of an ATF-2 binding site or expression of a dominant negative ATF-2 mutant abolishes ILK-dependent transcriptional activation of the cyclin D1 gene in mammary epithelial cells (4). We note that BMP7 induction of p38MAPK and ATF-2 follows delayed kinetics relative to activation of ILK, suggesting indirect stimulation of p38MAPK by ILK. Thus, ILK regulation of p38MAPK/ATF-2 mediates diverse, context-dependent developmental processes.
The overlapping expression patterns of BMP7 (12, 14, 15, 31), β1 integrin (1, 27), p38MAPK (20, 22), and ILK (Fig. (Fig.1)1) in embryonic kidneys suggest these molecules interact to regulate epithelial-mesenchymal interactions. Integrin-mediated UB cell-ECM interactions regulate branching in vitro and in vivo (10, 25, 51), and the requirement for a three-dimensional collagen matrix for IMCD-3 cell morphogenesis also implicates β1 integrin signaling in this response. Functional integrity of β1 integrins is required for proper development of the ureteric bud and elaboration of the collecting duct system. Genetic ablation of the α8 subunit in mice abolishes expression of α8β1 integrin, leading to profound defects of RBM (34). Loss of α2β1 integrin expression inhibits both collagen interaction and HGF-stimulated branching morphogenesis of MDCK kidney epithelial cells (41). Accordingly, blocking β1 integrin function markedly inhibits branching of primary UB cells in three-dimensional collagen cultures and UB development in embryonic kidney explants (51). These results indicate the importance of β1 integrin signaling in UB development, and our results reveal a key function of ILK in coordinating integrin and growth factor signaling during mammalian nephrogenesis. Interestingly, blocking β1 integrin function suppresses ductal branching in the mammary epithelium of mice and inhibits HGF-induced branching of mammary epithelial cells in vitro (26). Moreover, expression of ILK in the mammary epithelium of transgenic mice induces formation of ductal branching structures (48), indicating that ILK signaling is likely to be of broad relevance in mediating epithelial branching during mammalian development.
We acknowledge the expert technical assistance of Yunkai Yu, Perry Mongroo for quantifying protein expression data, and S. Dedhar (University of British Columbia) for the gift of KP-392.
S.H. is the recipient of a studentship from the Research Training Committee of the Hospital for Sick Children. This work was supported by grants from the Canadian Institutes of Health Research (CIHR) (to N.D.R. and G.E.H.), and the National Cancer Institute of Canada (to G.E.H., with funds from the Terry Fox Run). G.E.H. was a Scholar of the CIHR.