β1 integrin is necessary for ureteric bud branching morphogenesis and maintenance of collecting duct structural integrity

Formation, growth and branching morphogenesis of the collecting system of the kidney require interactions between the ureteric bud (UB) and the metanephric mesenchyme (MM). The UB ultimately forms the multibranched collecting system within the kidney, as well as the ureter and the bladder trigone. Development of the collecting system initially involves many iterations of branching morphogenesis followed by a period in which kidney growth predominates. This complex developmental process is dependent on numerous factors, including growth-factor-dependent cell signaling induced by glial cell line derived neurotrophic factor (GDNF) and fibroblast growth factors (FGFs) as well as interactions between cells and extracellular matrix(ECM) components (Dressler,2006).

Integrins are cell-surface receptors that mediate the interactions between cells and ECM. They consist of non-covalently bound α and βsubunits that combine in a restricted manner to form specific αβdimers. In mammals there are 18α and 8β subunits that form more than 20 different dimers, each of which exhibits different ligand binding and signaling properties (Hynes,2002). β1 is the most abundantly expressed β subunit and is found in almost all cell types in the body, including the kidney(Hynes, 2002; Kreidberg and Symons, 2000). Integrins α3β1, α6β1 and α6β4 are the major laminin-binding receptors, whereas the predominant collagen receptors are integrins α1β1and α2β1. Although they are primarily thought of as anchoring molecules, integrins play a crucial role in cell adhesion, migration, proliferation and apoptosis by transducing signals through their cytoplasmic tails following ligand binding(Moser et al., 2009; Pozzi and Zent, 2003). Thus,integrins have the potential to play important roles in organ morphogenesis by modulating cell growth, motility and shape(Hynes, 2002; Schwartz and Ginsberg,2002).

The role of β1 integrins in kidney collecting system development is unclear. Among the different integrin α-null mice generated, only integrin α3- and α8-null mice show a severe collecting system phenotype (Kreidberg et al.,1996; Muller et al.,1997). Integrin α8-null mice have defects in UB growth and branching as well as recruitment of mesenchymal cells into epithelial structures, because integrin α8β1 expression is induced in mesenchymal cells upon contact with its specific ligand, nephronectin, which is expressed by the UB (Linton et al.,2007). Mice lacking the integrin α3 subunit have a reduced number of collecting ducts (CDs) in the papilla, suggesting decreased branching morphogenesis of the UB(Kreidberg et al., 1996). When the integrin α3 subunit was specifically deleted in the UB, the kidney papillae were either absent or abnormal(Liu et al., 2009); however,the rest of the collecting system of the kidney was unaffected. Interestingly,integrin α6-null mice do not display a collecting system phenotype(Georges-Labouesse et al.,1996), although α3/α6 double-deficient mice fail to develop ureters (De Arcangelis et al.,1999).

In contrast to the in vivo data, in organ and cell culture models, α6 integrins (i.e. α6β1 and α6β4) have been found to be important for UB branching morphogenesis, and blocking the α6 subunit alone or in combination with the α3 subunit affects UB branching morphogenesis. Thus, the relative importance of integrins α6β1 andα6β4 in UB branching is currently uncertain(Zent et al., 2001).

To determine whether β1 integrins other than α3β1 play a role in UB development in vivo, we selectively deleted β1 integrin in the UB at two time points. A severe branching phenotype was observed when β1 integrin was deleted at E10.5, at the time that UB branching morphogenesis is initiated. By contrast, when β1 integrin was deleted in collecting ducts at E18.5, development was normal; however, severe collecting system injury was observed in adult animals following ureteric obstruction. We found that canonical signaling pathways activated by FGFs require β1 integrin expression in CD cells in vitro and, most importantly, activation of the same pathways was decreased in β1-null UBs in vivo, despite the normal activation state of the FGF receptor. Thus in addition to their well characterized roles in adhesion and migration, β1 integrins play a crucial role in transducing growth-factor-dependent signals required for UB branching morphogenesis and maintaining collecting tubule integrity following injury.

All experiments were approved by the Vanderbilt University Institutional Animal Use and Care Committee. Integrin β1^flox/flox mice(generous gift of Dr E. Fuchs, Howard Hughes Medical Institute, The Rockefeller University) (Raghavan et al.,2000) and integrin β1^flox/flox mice, in which a promoterless lacZ reporter gene was introduced after the downstream loxP site (Brakebusch et al.,2000) were crossed with the HoxB7Cre mice (generous gift of Dr A. McMahon) (Kobayashi et al.,2005) or Aqp2Cre mice (Ahn et al., 2004; Stricklett et al.,1999b). Mice were a F4-F6 generation toward the C56/Black6 background. Aged-matched littermates homozygous for the floxed integrinβ1 gene, but lacking Cre (β1^flox/flox mice), were used as controls.

For morphological and immunohistochemical analysis, kidneys were removed at different stages of development and were: (1) fixed in 4% formaldehyde and embedded in paraffin; (2) embedded in OCT compound without fixation and stored at -80°C until use; or (3) fixed in 2.5% glutaraldehyde, post-fixed in OsO₄, dehydrated in ethanol and embedded in resin. Paraffin tissue sections were stained with either hematoxylin and eosin or periodic acid-Schiff (PAS) for morphological evaluation by light microscopy. For electron microscopy, ultrastructural assessments of thin kidney sections were performed using a Morgagni transmission electron microscope (FEI, Eindhoven,The Netherlands).

Embryonic kidneys were isolated from E12.5 mice and placed on the top of transwell filters and cultured (37°C and 5% CO₂/100% humidity)in DMEM/F12 media supplemented with 10% FBS. After 3 days, the kidneys were fixed in 4% paraformaldehyde and stained with fluorescein-conjugated E-cadherin antibodies (BD Transduction Laboratories, Lexington, KY, USA), as described (Zent et al., 2001). The number of branching structures was counted and quantified as number of branches/kidney. Ten kidneys per phenotype were analyzed.

CD cells were isolated from β1^flox/flox mice following the methodology described by Husted et al.(Husted et al., 1988), andβ1 was deleted by infecting the cells with an adenocre virus in vitro. To verify adequate deletion of β1 integrin, the cells were subjected to flow cytometry as described below.

β1^flox/flox CD cells and β1^-/- cells were incubated with anti-mouse β1, β4, α1, α2, α6 andαv integrin antibodies (BD Transduction Laboratories), followed by FITC-conjugated secondary antibodies. Expression levels of different integrins in these two cell lines were detected by flow cytometry.

Cell adhesion assays were performed in 96 well plates, as described(Chen et al., 2004). Briefly plates were coated with different concentrations of ECM components and blocked with BSA. In each well, 1×10⁵ cells were placed in serum-free DMEM for 60 minutes; non-adherent cells were removed and the remaining cells were fixed, stained with Crystal Violet, solubilized and the optical density of the cell lysates was read at 540 nm. Four independent experiments were performed in triplicate.

Cell migration was assayed as previously described(Chen et al., 2004). Briefly,transwells with 8 μm pores were coated with different ECM components and 1×10⁵ cells were added to the upper well in serum-free medium. The cells that migrated through the filter after 4 hours were counted. Three random fields were analyzed per treatment. Four independent experiments were performed in triplicate.

For each well, 5×10³ cells were plated in 96-well plates on different ECM components and maintained in DMEM (10% FBS). After 12 hours,the cells were incubated in DMEM (2% FBS) for 24 hours and then pulsed with 1μCi/well [³H] thymidine (PerkinElmer Life Sciences). Twenty-four hours later, the cells were solubilized and radioactivity was measured using a scintillation counter. For manual counting assays, 6×10⁴cells were plated per plate on collagen or vitronectin and maintained in a serum-free environment. Cells were counted 48 hours after FGF (10 ng/ml)administration.

Cells were grown on transwells consisting of polyvinylpyrolidone-free polycarbonate filters with 0.4 μm pores. After reaching confluency, cells were fixed in 4% formaldehyde and incubated with anti-ZO-1 (1:200; BD Transduction Laboratories, Lexington, KY, USA) antibodies followed by the appropriate FITC-conjugated secondary antibody. Chamber slides were mounted and viewed using a confocal microscope.

Cells were plated onto slides coated with Collagen I (10 μg/ml) or Vitronectin (10 μg/ml) for 30 minutes, after which FGF (10 ng/ml) or GDNF(10 ng/ml) was added. Forty-five minutes later cells were fixed, permeabilized and exposed to rhodamine phallodin (1:5000). For analysis of phosphorylated and total FGFR1 cells were incubated with anti-pFGFR1 (Y654; Abcam, Cambridge,MA, USA) and anti-FGFR1 antibody (Cell Signaling, Beverly, MA, USA) for 12 hours at 4°C, followed by incubation with a FITC-conjugated secondary antibody for 2 hours, and visualized under a microscope.

The CD cells were trypsinized, washed, suspended in serum-free DMEM and then plated on Collagen I or Vitronectin (10 μg/ml) for 45 minutes. Growth factors (FGF or GDNF, 10 ng/ml) were added to the medium and the cells were lysed at different time points following growth-factor stimulation. For analysis on kidney tissues, the medullas were removed and lysed with RIPA buffer. Lysates were clarified by centrifugation and 30 μg total protein was electrophoresed onto an 10% SDS-PAGE and subsequently transferred to nitrocellulose membranes. Membranes were blocked in 5% milk/TBS Tween and then incubated with the different primary antibodies followed by the appropriate HRP-conjugated secondary antibodies. Immunoreactive bands were identified using enhanced chemiluminescence according to the manufacturer's instructions. AKT (1:1000, 9272), pAkt (1:1000, 9271S), ERK (1:1000, 9102), pERK (1:1000,9101S), p38 MAPK (1:1000, 9212), pp38 MAPK (1:1000, 9211S) antibodies were purchased from Cell Signaling. FAK (sc-558) and pFAK (sc-16662-R) antibodies were purchased from Santa Cruz Biotechnology.

For immunofluorescence, 5 μm frozen kidney sections were fixed in 4%paraformaldehyde for 10 minutes at room temperature, blocked with 1% BSA in PBS and then stained with the following primary antibodies: anti-E-cadherin(1:200); anti-FGF2 (1:100, Upstate 05-118), anti-FGFR1 (1:100, Santa Cruz,sc121) and anti-FGFR2 (1:100, Santa Cruz, sc122). Detection of bound primary antibodies was accomplished with Alexa Fluor 647 goat anti-rabbit IgG or Alexa Fluor 488 anti-mouse IgG (Molecular Probes). Slides were then analyzed under an epifluorescence microscope.

The right ureters of 6-week-old Aqp2Cre; β1^flox/flox mice and β1^flox/flox mice were ligated and mice were sacrificed 5,7 and 10 days after the surgery. Kidneys were removed and fixed in 4%formaldehyde and embedded in paraffin. The degree of tubular injury was then scored by the following criteria. Histological changes consistent with tubular necrosis were quantified by calculating the percentage of tubules with cell necrosis, loss of brush border, cast formation or tubular dilatation as follows. Degree of injury was scored as 0 (none), 1 (1-10%), 2 (11-25%), 3(26-45%), 4 (46-75%) and 5 (76-100%). At least 10 fields were analyzed for each slide. Degree of interstitial fibrosis and inflammatory infiltrate were judged using the same scale.

Student's t-test was used for comparisons between two groups, and analysis of variance using Sigma Stat software was used for statistical differences between multiple groups. P<0.05 was considered statistically significant.

To define the role of β1 integrin in the developing UB, we crossed HoxB7Cre mice, which express Cre in the Wolffian duct and UB from E10.5, with integrin β1^flox/flox mice, in which a promoterless lacZ reporter gene was introduced after the downstream loxP site(Brakebusch et al., 2000). As demonstrated in Fig. 1A, strongβ-galactosidase staining was evident in the collecting system of the kidney. We confirmed β1 integrin deletion by performing immunoblotting on isolated papillae of newborn HoxB7Cre;β1^flox/flox mice with an antibody directed at the integrin β1 subunit(Fig. 1A). Although these mice were born in the normal Mendelian ratio, all the HoxB7Cre;β1^flox/flox mice died between 4 and 6 weeks of age. The kidneys of 6-week-old mice were smaller and slightly cystic compared with those isolated from β1^flox/flox mice(Fig. 1B,C). On microscopic examination, HoxB7Cre;β1^flox/flox kidneys had far fewer nephrons than β1^flox/flox kidneys. Moreover, there were many more dilated tubules in both the cortex and medulla, and there was marked interstitial fibrosis (Fig. 1D-I). The most severe tubular dilatation was seen in the CDs located in both the medulla and the cortex(Fig. 1D-I).

To further define the defects present in these mice, studies were performed on embryonic kidneys from embryonic day 11.5 (E11.5) until birth. At every stage analyzed a branching morphogenesis defect was clearly present in HoxB7Cre;β1^flox/flox mice. At day E11.5 the UB was less branched and smaller in HoxB7Cre;β1^flox/flox mice(Fig. 2A,B). Kidneys of HoxB7Cre;β1^flox/flox mice at E13.5(Fig. 2C,D) as well as E15.5(Fig. 2E,F) were significantly smaller, with a decreased number of UB branches and nephrons. The branching phenotype observed in HoxB7Cre;β1^flox/flox mice was verified by performing in vitro cultures of E12.5 embryonic kidney. After 48 hours in culture, HoxB7Cre;β1^flox/flox E12.5 kidneys formed approximately five branches per structure relative to the 50 detected in theβ1^flox/flox kidneys (Fig. 2G-I).

Kidneys of newborn HoxB7Cre;β1^flox/flox were also smaller and had fewer collecting ducts than β1^flox/flox mice(Fig. 2J,K). Interestingly the tubules, although somewhat disorganized, did not appear to have abnormalities with respect to polarity (Fig. 2L,M). This observation was confirmed when the newborn kidneys were stained with antibodies directed against E-cadherin (Cdh1 - Mouse Genome Informatics) (Fig. 2N,O) and ZO1 (Tjp1 - Mouse Genome Informatics) (data not shown). Electron microscopy on non-perfused newborn kidneys also established that cells forming the tubules were polarized and that the tubular basement membranes were normal, despite the finding that the tubular lumens were dilated(Fig. 2P,Q). A significant decrease in nephrons was also present in newborn HoxB7Cre;β1^flox/flox mice, as there were 4.2 versus 1.6 glomeruli/mm² of cortex in the HoxB7Cre;β1^flox/floxand β1^flox/flox mice, respectively(Fig. 2R). The difference in absolute glomerular number between HoxB7Cre;β1^flox/flox andβ1^flox/flox mice is even bigger when one considers that kidneys of the HoxB7Cre;β1^flox/flox are much smaller than those of β1^flox/flox mice.

Taken together, these data demonstrate that deleting β1 integrin in the UB of the kidney at the time of initiation of branching morphogenesis results in a profound branching phenotype with fewer nephrons and a severely dysmorphic dysplastic collecting system.

To determine the role of β1 integrin in CD development, we utilized the aquaporin 2 Cre (Aqp2Cre) mouse to delete β1 in CDs at E18.5 (R. Nelson and D.E.K., unpublished) (Ahn et al., 2004; Stricklett et al.,1999a; Zhang et al.,2005). Gene deletion in the Aqp2Cre; β1^flox/floxmice was confirmed by immunoblots on isolated papillae of 6-week-old mice(Fig. 3A). These mice were born in the normal Mendelian ratio and lived a normal life span. Despite intensive investigation, no gross or microscopical abnormalities were found at any age in these mice (Fig. 3B-G). Owing to this surprising lack of phenotype, we investigated the timing of theβ1 integrin subunit deletion and found that it was still expressed in the CDs at postnatal day 21 (P21) (data not shown), suggesting that β1 integrin was still present when development was complete.

These mice provided a model to determine whether β1 integrin played a protective role to obstructive injury of the collecting system. When 6-week-old β1^flox/flox and Aqp2Cre;β1^flox/flox mice were subjected to unilateral ureteric obstruction, there was markedly increased tubular dilatation and flattening of tubular epithelial cells in the Aqp2Cre; β1^flox/flox mice at both 5 (Fig. 3H-K) and 10 (data not shown) days following injury. When tubular injury was scored at day 5, it was significantly worse in the Aqp2Cre; β1^flox/flox than theβ1^flox/flox mice (4.8±0.6 versus 1.8±0.3 P<0.01). These mice also had significantly more Trichrome-Blue-positive matrix (Fig. 3L,M), consistent with interstitial fibrosis. Moreover, injured Aqp2Cre; β1^flox/flox mice had more tubular apoptosis, as shown by TUNEL staining (Fig. 3N-P)and verified by increased activation of caspase-3(Fig. 3Q). Thus, deletingβ1 integrins at late stages of UB development renders mice susceptible to severe renal injury following ureteric obstruction.

We next isolated CD cells from 3-week-old β1^flox/flox mice and deleted β1 in vitro, utilizing adeno-cre virus to define its role in CD cell function. CD cells are an ideal in vitro model to study integrin-dependent functions in both adult and embryonic tubular epithelial cells derived from the UB (Chen et al.,2004). Successful deletion of β1 in CD(β1^-/-) cells was verified by flow cytometry utilizing an antibody directed at the extracellular domain of the mouse β1 integrin(Fig. 4A). Similarly, all theα subunits that heterodimerize exclusively with β1 were not expressed on the β1^-/- cells (not shown). No difference inβ4 (data not shown) or αv expression(Fig. 4A) were observed between the two cell types. When cell adhesion to various matrices was investigated,β1^-/- CD cells adhered poorly to collagen I and laminin 1(Fig. 4B). By contrast,adhesion to fibronectin (ligand for both integrin α5β1 andαv-containing integrins) was only partially impaired, whereas binding to vitronectin (ligand for αvβ3 and αvβ5 integrins) was unaffected. Similar results were observed for cell migration(Fig. 4C).β1^-/- CD cells also proliferated significantly less thanβ1^flox/flox cells when plated on β1 integrin-dependent substrates (Fig. 4D), an effect that was not observed on fibronectin, vitronectin or plastic substrata. To determine the role of β1 integrin in the regulation of CD cell polarity,cells were grown on transwells till confluent and stained with ZO-1(Fig. 4E) and E-cadherin (data not shown) antibodies. Confocal microscopy revealed no difference in ZO-1 or E-cadherin localization on z-sectioning between the two cell lines. Thus, deleting integrin β1 from CD cells results in decreased cell adhesion, migration and proliferation on β1-integrin dependent substrates; however, it does not significantly alter the ability of the cells to polarize.

UB branching morphogenesis in vivo is regulated by crucial growth factors,including GDNF and FGF family members. As the major UB developmental defect in the HoxB7Cre;β1^flox/flox mice occurred during the rapid branching phase known to be dependent on these growth factors, we investigated the role of β1 integrin in GDNF-as well FGF2- and 10-mediated signaling in vitro. As GDNF and FGF2 can induce cell spreading(Klint et al., 1999; Murakami et al., 1999), we determined the requirement of β1 integrin for growth-factor-mediated CD cell spreading on collagen I or laminin I (data not shown). Unlikeβ1^flox/flox, β1^-/- CD cells adhered poorly when plated on collagen I, and the few that did adhere were unable to spread(Fig. 5A). Addition of GDNF,FGF2, FGF10 (Fig. 5A) or HGF(data not shown) induced much less spreading of β1^-/- cells compared with β1^flox/flox CD cells. By contrast, bothβ1^-/- and β1^flox/flox CD cells spread equally when plated on vitronectin, and this effect was increased to a comparable degree in both genotypes following the addition of GDNF(Fig. 5A).

To determine the requirement of β1 integrin for GDNF, FGF2, FGF10 and HGF signaling on collagen I, β1^flox/flox andβ1^-/- CD cells where placed on collagen I for 2 hours, after which they were stimulated with the growth factor. Compared withβ1^flox/flox CD cells, FAK, ERK, p38MAPK and Akt phosphorylation was markedly decreased in the β1^-/- CD cells in response to GDNF, FGF2, FGF10 (Fig. 5B-D) and HGF (data not shown). As expected, no difference in GDNF-dependent signaling was observed between the β1^flox/floxand β1^-/- CD cells plated on vitronectin(Fig. 5E). These results suggest that β1 integrin expression is required for CD cells to spread and mediate signaling induced by GDNF, FGF2, FGF10 and HGF when plated onβ1 integrin-dependent substrates.

As β1^-/- CD cells showed a significant decrease in activation of signaling pathways stimulated by GDNF and FGFs, we assessed the activation status of the same pathways in the developing collecting system in vivo. Western blot analysis of isolated medullas of newborn mice showed drastically decreased activation of FAK, ERK, p38 MAPK and Akt in HoxB7Cre;β1^flox/flox compared with β1^flox/floxmice (Fig. 6A). As expected,the decreased signaling activation correlated with decreased cellular proliferation in the UB, which was particularly evident at E15.5, when the proliferation in β1^flox/flox mice was 10-fold greater than that of HoxB7Cre;β1^flox/flox mice(Fig. 6B). Similarly, there was decreased proliferation in the collecting ducts of newborn HoxB7Cre;β1^flox/flox relative to theβ1^flox/flox mice (Fig. 6B). As FGFs play an important role in mediating proliferation and branching of the UB at E15 (Bates,2007; Bush et al.,2006), FGF2, FGFR1 and FGFR2 expression were determined in renal tubules at this time point. No differences in FGF2 and FGFR expression were observed between HoxB7Cre;β1^flox/flox andβ1^flox/flox kidneys (Fig. 6C), suggesting that decreased proliferation and signaling were not due to decreased expression of the growth factors or their receptors.

To investigate whether β1 expression might alter FGF receptor localization and/or activation, β1^-/- andβ1^flox/flox CD cells were plated on collagen I and subsequently stimulated with FGF2. Equal amounts of FGFR1 surface expression and phosphorylation were observed by immunofluorescence in bothβ1^-/- and β1^flox/flox CD cells, despite the inability of the β1^-/- cells to spread on this matrix(Fig. 7A). Immunoblot analysis confirmed comparable FGFR1 phosphorylation and expression inβ1^-/- and β1^flox/flox CD cells plated on collagen (Fig. 7A). As expected, localization and phosphorylation of FGFR1 was the same in theβ1^-/- and β1^flox/flox CD cells plated on vitronectin (Fig. 7B). When the proliferative response of β1^-/- andβ1^flox/flox CD cells to FGF2 was determined,β1^-/- CD cells plated on collagen I did not proliferate,whereas both cell populations responded to this growth factor when grown on vitronectin (Fig. 7C). Thus,although the FGFR1 is expressed on the cell surface of β1^-/-CD cells and can be activated by FGF2, ligation of β1 integrin is required for this growth factor receptor to signal normally and induce cell proliferation.

To determine whether similar findings were seen in vivo, immunoblotting was performed on renal papillae isolated from newborn β1^flox/floxand HoxB7Cre;β1^flox/flox kidneys. Similar to the in vitro results, FGFR1 was phosphorylated equally in β1^flox/flox and HoxB7Cre;β1^flox/flox mice(Fig. 7D), verifying that UBs lacking β1 integrins have severe abnormalities in branching morphogenesis, without alterations in the levels of growth factor receptor phosphorylation/activation.

The complex process of UB development requires both regulated growth factor signaling and cell-ECM interactions. There is in vitro and in vivo evidence that the laminin receptor, integrin α3β1, is required for collecting system development (Kreidberg et al., 1996; Liu et al.,2009), whereas requirement of the α6 integrin subunit has only been demonstrated in vitro (Zent et al., 2001). The collagen receptors, integrins α1β1 andα2β1, do not play a major role in this process(Zent et al., 2001) (R. Zent and A. Pozzi, unpublished data), and the role of the RGD-binding receptorsα5β1 and αv integrins is unknown. In this study we demonstrate that deleting β1 integrin in the UB at E10.5 results in severe branching morphogenesis abnormalities and decreased nephron formation,with death of the mice by 4 to 6 weeks of age. These results confirm the recent observation that deleting β1 integrin in the developing UB results in small kidneys with hypoplastic renal medullary collecting ducts(Wu et al., 2009). The abnormality in branching morphogenesis was significantly worse in mice in which β1 integrin rather than α3 was specifically deleted in the UB(Liu et al., 2009), suggesting that other αβ1 integrins play a role in this process. This was exemplified by the concomitant decrease in nephron number in the Hoxb7Cre;β1^flox/flox kidneys, which was not seen when α3 was either deleted in the whole mouse or specifically in the UB(Kreidberg et al., 1996; Liu et al., 2009). The only other branching organ in which α3, α6 and β1 integrin subunits have been deleted is in the mammary gland(Klinowska et al., 2001; Li et al., 2005; Naylor et al., 2005; Taddei et al., 2008). Mammary gland development was normal when the α3 or α6 integrin subunits were deleted, and β1 deletion resulted in a loss of epithelial integrity and displacement of cells from the basement membrane, but no branching phenotype was evident. Similar to the kidney phenotype, there were no abnormalities in epithelial cell polarity or basement membrane deposition.

When β1 integrin was deleted in the collecting ducts at E18.5 using the Aqp2Cre mouse, no developmental phenotype was observed. This probably occurs because sufficient β1 integrin is expressed in the collecting ducts until P21, when normal development is complete, and suggests that early but not late loss of β1 integrin expression regulates terminal differentiation and collecting duct function(Wu et al., 2009). The lack of phenotype in the Aqp2Cre; β1^flox/flox mice could be explained by the fact that either β1 was inefficiently deleted or its turnover is slow. We would suggest that the latter explanation is more likely, as deleting this matrix receptor rendered the adult kidney more susceptible to injury following ureteric obstruction. Thus β1 integrins are required to maintain structural integrity when the collecting system is subjected to the increased hydrostatic pressure induced by tying the ureter. A similar phenomenon was observed when β1 integrin was deleted in podocytes, where glomerular morphogenesis occurred relatively normally until birth, but the glomeruli underwent rapid destruction when intraglomerular pressure was increased with the advent of glomerular filtration(Kanasaki et al., 2008; Pozzi et al., 2008).

Utilizing cultured CD cells, we found that deleting β1 integrin significantly impaired canonical signaling pathways activated by FGFs and GDNF, both of which are known to be important in UB branching(Durbec et al., 1996; Pachnis et al., 1993; Zhao et al., 2004). Similar alterations in signaling pathways were seen in the HoxB7Cre;β1^flox/flox mice, despite normal expression levels of FGF2 and the FGF receptors 1 and 2. As the phenotype of the HoxB7Cre;β1^flox/flox kidney is primarily that of decreased UB branching morphogenesis, which phenocopies many of the features seen in the GDNF, Ret (the GDNF receptor) and FGFR2 null mice(Durbec et al., 1996; Pachnis et al., 1993; Zhao et al., 2004), we attribute the phenotype, at least in part to abnormalities in growth factor-dependent signaling. To the best of our knowledge, this is the first illustration that β1 integrin mediates growth factor-dependent branching morphogenesis during a developmental process.

Growth factor receptors and integrin-stimulated pathways are very similar and undergo crosstalk at many levels. Growth factors can directly bind to integrins (Mori et al., 2008),integrins can directly interact with receptor tyrosine kinases(Cascone et al., 2005; Wang et al., 2008) and integrin and receptor tyrosine kinase signaling pathways intersect at various levels downstream of the respective receptors(Parise et al., 2000). We demonstrate the requirement of β1 integrins for FGF2 and FGF10 signaling in polarized epithelial cells, despite normal FGFR1 activation and expression on the cell membrane. FGFR1 phosphorylation is also normal in HoxB7Cre;β1^flox/flox and β1^flox/flox mice despite a severe branching phenotype, suggesting that β1 integrins are not required for normal phosphorylation of the receptor in vivo. Thus in the context of UB development we propose that β1 is required for the propagation of the signals downstream of receptors like the FGF receptors rather than for receptor phosphorylation per se.

In conclusion, our data demonstrate that β1 integrins play a role in mediating growth-factor-dependent branching morphogenesis in the developing UB. This novel role for β1 integrins in developmental processes is in addition to their well-defined function in maintaining structural integrity of tubules by promoting adhesion to basement membranes.