angiotensin ii type 1 receptor–egf receptor cross-talk...

Angiotensin II Type 1 Receptor–EGF Receptor Cross-TalkRegulates Ureteric Bud Branching Morphogenesis

Ihor V. Yosypiv, Mercedes Schroeder, and Samir S. El-DahrSection of Pediatric Nephrology, Department of Pediatrics, Tulane University Health Sciences Center,New Orleans, Louisiana

Angiotensinogen-, angiotensin-converting enzyme–, and angiotensin II (Ang II) type 1 receptor (AT1R)–deficient mice exhibita dilated renal pelvis (hydronephrosis) and a small papilla. These abnormalities have been attributed to impaired develop-ment of the ureteral and pelvic smooth muscle. Defects in the growth and branching of the ureteric bud (UB), which gives riseto the collecting system, have not been examined carefully. This study tested the hypothesis that Ang II stimulates UB growthand branching in the intact metanephros. Immunohistochemistry demonstrated that embryonic mouse kidneys express AT1Rin the UB and its branches. Embryonic day 11.5 metanephroi were microdissected from Hoxb7–green fluorescence proteinmice and grown for 48 h in serum-free medium in the presence or absence of Ang II. The number of green fluorescenceprotein–positive UB branch points (BP) and tips was monitored in each explant at 24 and 48 h. Ang II increased the numberof UB tips and BP at 24 h (tips: 24.3 � 1.1 versus 18.3 � 0.7, P < 0.01; BP: 14.4 � 0.6 versus 11.7 � 0.6, P < 0.01) and 48 h (tips:30.2 � 1.3 versus 22.9 � 0.8, P < 0.01; BP: 21.3 � 0.9 versus 15.7 � 0.6, P < 0.01) compared with control. In contrast, treatmentof metanephroi with the AT1R antagonist candesartan inhibited UB branching, decreasing the number of UB tips and BP.Similarly, inhibition of EGF receptor (EGFR) tyrosine kinase activity abrogated Ang II–stimulated UB branching. A cross-talkbetween the renin-angiotensin system and EGFR signaling was elicited at the cellular level by the ability of Ang II to inducetyrosine phosphorylation of EGFR in UB cells and through abrogation of Ang II–induced UB cell branching using an EGFRtyrosine kinase inhibitor. These data demonstrate that Ang II, acting via the AT1R, stimulates UB branching morphogenesis.This process depends on tyrosine phosphorylation of the EGFR. Cooperation of AT1R and EGFR signaling therefore isimportant in the development of the renal collecting system.

J Am Soc Nephrol 17: 1005–1014, 2006. doi: 10.1681/ASN.2005080803

T he metanephros develops via reciprocal inductive in-teractions between the ureteric bud (UB) and the met-anephrogenic mesenchyme (MM) (1,2). Signals from

the MM induce UB outgrowth from the nephric duct and itselongation and entrance into the mesenchyme followed byrepetitive branching to form the renal collecting system (ureter,pelvis, calyces, and collecting ducts). In turn, emerging UB tipsinduce surrounding mesenchymal cells to condense, aggregate,undergo mesenchymal-to-epithelial transition, and formnephrons (from the glomerulus to the distal tubule). Therefore,UB branching morphogenesis is critical in determining nephronnumber, and subtle defects in the efficiency and/or accuracy ofthis process potentially can have profound effects on the properdevelopment of the metanephric kidney. Decreased nephronendowment is linked to hypertension and eventual progressionto chronic renal failure (3,4). In addition, aberrant UB branchingmorphogenesis causes renal dysplasia, the leading cause ofchronic renal failure in human infants.

Genetic inactivation of the renin-angiotensin system (RAS)

genes in mice causes abnormalities in the development of theureter, renal pelvis, and papilla (5–9). Angiotensinogen-, angio-tensin-converting enzyme (ACE)-, or angiotensin II (Ang II)type 1 receptor (AT1R)-deficient mice exhibit pelvic dilation(hydronephrosis) and a small papilla mimicking urinary tractobstruction. Elegant studies from Ichikawa’s laboratory havesuggested that absence of AT1R signaling in ureteral smoothmuscle cells impairs pelvic-ureteral smooth muscle develop-ment and peristalsis (9). Mutations in the AT2R gene in miceand humans are associated with increased incidence of lowerurinary tract anomalies, including double ureters and vesi-coureteral reflux (10). These findings indicate that UB growthand development are a target for Ang II actions.

Work that has performed by several laboratories, includingours, has revealed that the fetal kidney expresses a local RAS.Quantitative analysis of murine Ang II receptors AT1R andAT2R gene expression indicate that AT1R undergo a progres-sive increase during fetal and neonatal life, whereas AT2R arehigh in the fetus and decline significantly with maturation (11).Immunolocalization studies demonstrated that AT1R and AT2Rare expressed in the UB and its early branches of embryonicday 12 (E12) mouse embryos, whereas angiotensinogen andrenin are expressed in the stromal mesenchymal cells thatsurround the UB branches (12–14). The localization of RAScomponents in the stroma and UB provided the initial sugges-tion that paracrine signaling through RAS-AT receptors may

Received August 1, 2005. Accepted January 17, 2006.

Published online ahead of print. Publication date available at www.jasn.org.

Address correspondence to: Dr. Ihor V. Yosypiv, Department of Pediatrics, TulaneUniversity Health Sciences Center, SL-37, 1430 Tulane Avenue, New Orleans, LA70112. Phone: 504-988-5377; Fax: 504-988-1852; E-mail: [email protected]

Copyright © 2006 by the American Society of Nephrology ISSN: 1046-6673/1704-1005

regulate the early stages of kidney development. In vitro studieshave shown further that UB cells that are isolated from E11.5mouse embryos express AT1R that are functionally coupled tothe mitogen-activated protein kinase (MAPK) pathway (15). UBcells that are seeded in collagen gels form cellular cords andtubular structures, and treatment with Ang II increases thenumber and the complexity of these cellular branches (12). AngII–induced UB cell branching is mediated in part via the AT1Rbecause it is abrogated by the specific AT1R antagonist cande-sartan. However, there is no evidence that Ang II can directlystimulate UB growth and branching in the intact metanephros.

In this work, we tested the hypothesis that Ang II stimulatesUB branching morphogenesis in the intact metanephric kidneycultured in vitro. Moreover, given the known interactions be-tween AT1R and EGF receptor (EGFR) signaling pathways inother systems (16,17), we examined the cross-talk between AngII and the EGFR in Ang II–induced UB branching morphogen-esis. The results demonstrate that Ang II, when applied directlyon the metanephric organ culture, stimulates the UB branchingmorphogenesis program. Furthermore, the stimulatory effectsof Ang II on metanephric UB branching and tubulogenesis aremediated via an AT1R-EGFR signaling pathway.

Materials and MethodsImmunohistochemistry for AT1R

We previously demonstrated that AT1R protein is expressed in theUB as early as E12 of murine development (12). In this study, weexamined the expression of AT1R in mouse kidneys on E14 to assessfurther its distribution along the UB branches and collecting ducts.Immunostaining was performed by the immunoperoxidase techniqueusing the Vectastain Elite kit (Vector Laboratories, Burlingame, CA) asdescribed previously (12). A polyclonal rabbit AT1R antibody directedagainst the NH2-terminal domain of the human receptor (identicalsequence in the mouse; N-10, sc-1173; Santa Cruz Biotechnology, SantaCruz, CA) was used at concentration of 1:200. This antibody detectsboth AT1A and AT1B receptor subtypes. The specificity of immunostain-ing was assessed by preadsorption of the primary antibody with a �100excess of its immunogenic peptide overnight at 4°C.

Cell ImmunostainingFor examination of the effect of Ang II on subcellular localization and

tyrosine phosphorylation of EGFR in UB cells, UB cells that were grown oncoverslips were incubated with medium alone (control), Ang II (10�6 M),or EGF (40 ng/ml) for 30 min. After washings, the cells were incubatedwith FITC-conjugated anti–phospho-EGFR antibodies (Tyr-1173, sc-12351,Santa Cruz Biotechnology, Santa Cruz, CA; or Tyr-845, 2231, Cell Signal-ing, Danvers, MA) and propidium iodide to stain the nuclei.

Cell Lysate Extraction, Immunoprecipitation, andImmunoblotting

Mouse UB cells (provided by Dr. Jonathan Barasch, Columbia Uni-versity, New York, NY) initially were obtained from microdissectedureteric buds of E11.5 mouse embryo that was transgenic for simianvirus 40 large T antigen (Immorto-mouse) (18). UB cells were main-tained in DMEM/F12 medium supplemented with 10% FBS at 37°C ina humidified 5% CO2 atmosphere. The cells were cultured in serum-free medium for 48 h before experiments. The cells were used atpassages 10 to 15. Cell stimulation with Ang II (10�5 to 10�7 M) or EGF(40 ng/ml) was carried out at 37°C in serum-free medium. Western blot

analysis was performed as described previously (12). After stimulation,the cells were washed in PBS and homogenized in cold lysis buffer (50mM Tris/HCl [pH 8.0], 150 mM NaCl, 0.02% sodium azide, 1% NonidetP-40, and 0.5% deoxycholate) that contained a cocktail of enzymeinhibitors (0.1 mg/ml PMSF, 0.001 mg/ml aprotinin, 0.001 mg/mlleupeptin, and 0.01 mg/ml Na3VO4). The samples were centrifuged for10 min at 14,000 � g at 4°C, and the supernatants that containedproteins (40 �g/sample) were resolved on 10% Tris-glycine gels andtransferred to nitrocellulose membranes.

For immunoprecipitation of endogenous EGFR, 2 mg of cell lysateswas incubated with 8 �g of anti-EGFR antibody (SC-03; Santa Cruz)overnight at 4°C. The cells were lysed in a lysis buffer (9803; CellSignaling) that contained phosphatase (P5726; Sigma) and protease(P8340; Sigma) inhibitor cocktails and PMSF (0.1 mg/ml). The lysateswere incubated with protein A/G PLUS-Agarose beads (sc-2003; SantaCruz; 60 �l of slurry) for 3 h. Samples were centrifuged at 13,000 rpmfor 30 s at 4°C, and the beads were washed with lysis buffer, solubilizedin sample buffer (7722; Cell Signaling), and subjected to immunoblot-ting (0.15 mg/lane). The adequacy of transfer was assessed by Pon-ceau-S staining of the membranes. Nonspecific binding was blocked byincubating the membranes with PBS that contained 0.2% Tween and 3%BSA overnight at 4°C. The membranes were incubated with the follow-ing primary antibodies: (1) A polyclonal rabbit antibody directedagainst phosphorylated Tyr-1173 in the carboxy terminal of EGFR(sc-12351; Santa Cruz); this antibody was used for immunoblotting at aconcentration of 1:500; and (2) membranes with immunoprecipitatedEGFR were probed with anti-phosphotyrosine mAb (4G10; UpstateBiotechnology, Lake Placid, NY). After three washes in PBS/Tween, thenitrocellulose membrane was exposed to the secondary antibody for 1 hat room temperature. Immunoreactive bands were visualized using theenhanced chemiluminescence detection system (ECL; Amersham, Pitts-burgh, PA). The blots then were exposed to Hyperfilm-ECL films.

In Vitro TubulogenesisUB cells that are cultured in gels form processes and branching

tubular structures when exposed to various growth factors, providinga convenient experimental system to analyze mechanisms of epithelialbranching morphogenesis (19,20). A principal advantage of this ap-proach is the ability to examine the direct effect of a specific factor onUB cell growth that is independent from confounding soluble factorsthat are released by the mesenchyme. The in vitro tubulogenesis assaywas performed as described previously (12). Briefly, the cells were usedat passages 10 to 15. The cells were trypsinized and suspended in typeI rat-tail collagen (Upstate Biotechnology), 10� DMEM, and 200 mMHEPES (pH 8.0) in 8:1:1 ratio at 150 � 103 cell/ml. Subsequently, thesuspension was dispensed in aliquots into a 96-well culture plate (0.1ml/well). After gelation, 0.1 ml/well DMEM/F12 medium with 0.5%FBS or medium with 0.5% FBS that contained (1) Ang II (10�5 or 10�6

M; Sigma), (2) Ang II and AT1R antagonist candesartan (10�6 M;Sigma), (3) Ang II and EGFR tyrosine kinase inhibitor (AG1478, 0.3 �M;Calbiochem, San Diego, CA), (4) EGF (40 ng/ml; Upstate Biotechnol-ogy), alone or combined with AG1478, 0.3 �M, was added to each wellon top of the collagen gel. Candesartan and AG1478 were added on topof the gels 30 min before Ang II or EGF were added. After a 24-hincubation period at 37°C and 5% CO2, 40 individual cells/well werescored for the presence of cell processes directly from the plates usingan Olympus IX70 inverted phase-contrast microscope (Olympus,Melville, NY), and the average number of processes per cell was cal-culated. Each condition was set up in triplicate (n � 4 separate exper-iments). Images were acquired directly from the plates via an OlympusMagnaFire FW camera and processed with Adobe PhotoShop 7.0(Adobe Systems, Mountain View, CA).

1006 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 1005–1014, 2006

Metanephric Organ CultureTo determine the role of Ang II in early stages of UB growth and

development in the intact kidney, we used metanephroi that were isolatedfrom Hoxb7–green fluorescence protein (GFP) transgenic mice (a gift ofDr. Frank Costantini, Columbia University). The kidneys of Hoxb7–GFPtransgenic mice express GFP exclusively in the UB (21). This allows con-tinuous monitoring of UB growth in real time. Embryos were dissectedaseptically from the surrounding tissues on E11.5, and the metanephroiwere isolated. The day when the vaginal plug was observed was consid-ered to be E0.5. GFP-positive metanephroi were photographed immedi-ately after isolation (time 0) and were grown on air-fluid interface onpolycarbonate transwell filters (0.5 �m; Corning Costar, Acton, MA) thatwere inserted into six-well plates that contained DMEM/F12 medium(Life Technologies BRL, Grand Island, NY) alone or in the presence of AngII (10�5 M) alone or Ang II combined with AG1478 (0.3 �M) for 48 h at37°C and 5% CO2. The media that did or did not contain Ang II, cande-sartan, or AG1478 were changed every 24 h. Our recent findings demon-strate that 10�5 M Ang II stimulates cell process formation in UB cells (12).Importantly, endogenous Ang II contents are higher in the embryonic thanin the adult mouse kidney (22). Therefore, higher kidney tissue Ang IIconcentrations are physiologically relevant during embryonic kidney de-velopment. Therefore, we used 10�5 M Ang II in this study. A higher doseof Ang II was chosen also because Ang II has a short half-life as a result ofdegradation by tissue proteases (see below). AG1478 (0.3 �M) and domi-nant-negative EGFR mutant have been shown to inhibit Ang II–inducedEGFR tyrosine phosphorylation in vascular smooth muscle cells (16).Therefore, this concentration of AG was used in our study. Consideringthe variability in the initial kidney size and in UB branching patternsamong littermates (23), the effect of treatments on UB branching wasstudied in the kidneys that were obtained from the same fetus (left kidneywas incubated with medium and right kidney with Ang II, or left kidneywas incubated with Ang II�AG1478 and right kidney with Ang II).Sequential images of branching UB were obtained at 24 and 48 h ofincubation directly from the inserts via an Olympus IX70 inverted phase-contrast microscope and Olympus MagnaFire FW camera and processedwith Adobe PhotoShop 7.0. Quantitative assessment of UB branching ineach treatment group was performed by counting the number of UB tipsand branch points (BP) at time 0 and at 24 and 48 h.

For examination of the role of AT1R in UB branching, E11.5 meta-nephroi that were obtained from wild-type C57B/6J mice were grownas described above in the presence of medium alone (control) or me-dium � AT1R antagonist candesartan (10�6 M) for 48 h. Candesartan at10�6 M blocks predominantly AT1R (24). Also, 10�6 M candesartanabrogated Ang II–induced UB cell branching in collagen matrix gels(12). Therefore, this concentration of candesartan was used in ourstudy. After 48 h in culture, kidney explants were stained with anti-pancytokeratin antibody (C2562, 1:200; Sigma) to label the UB as de-scribed above, and the number of UB tips and BP was comparedbetween the two treatment groups. To determine the degradation rateof exogenous Ang II present in the culture media, we measured Ang IIcontents in the media at time 0 and at 6, 12, and 24 h by RIA with theuse of rabbit anti-Ang II antiserum as described (25).

Statistical AnalysesDifferences among the treatment groups in process-per-cell number

were analyzed by one-way ANOVA. Differences in EGFR phosphory-lation were analyzed by Mann-Whitney rank sum test. Differences inthe number of UB tips and BP in Ang II–, candesartan-, or AG1478-treated metanephroi were analyzed by t test. P � 0.05 was consideredstatistically significant.

ResultsExpression of AT1R in the UB Tree In Vivo

On E14, AT1 immunoreactivity is abundant in the UB, UBstalks and branches, and ampullae (Figure 1A). AT1R is ex-pressed on both luminal and basolateral aspects of UB branches(Figure 1B). Glomeruli and the surrounding mesenchyme alsoexpress AT1R, albeit at relatively lower levels than the UBbranches (Figure 1A). Control sections that were incubatedwith the primary antibody preadsorbed with its immunogenshowed a marked decrease in staining (Figure 1C).

Effect of Ang II on UB Branching Morphogenesis in the ExVivo Cultured Intact Kidney

In a previous study, we demonstrated that Ang II can initiate abranching morphogenesis-like program in UB cells (12). In thisstudy, we examined whether Ang II can induce similar effects ina physiologically more relevant system, in which mesenchymal–epithelial interactions are still maintained. Metanephric kidneysthat are grown ex vivo proceed through the stages of UB branchingand nephrogenesis over several days and closely recapitulate kid-ney development in vivo (23). To monitor the effects of Ang II onthe sequential changes in UB morphogenetic events, we usedHoxb7-GFP transgenic mice (26), which express GFP exclusivelyin the UB and its derived epithelial structures.

Quantitative analysis of UB branching morphogenesis in-cluded counting the number of UB tips and BP. A tip is definedas the distal end of terminal UB branch. BP was defined as thepoint at which three or more segments connect. To measure thetime-dependent changes in UB branching, we analyzed threeimages of each kidney: Immediately after dissection (time 0)and 24 and 48 h thereafter (Figure 2). The number of UB tipsand BP increased progressively during the observation period.As described previously (26), three modes of branching wereobserved: Bifid, trifid, and lateral. Bifid branching results intwo new segments joined to the parental segment at a singlepoint. Trifid branching involved three new segments joined tothe parental segment. Lateral branching involved the out-growth of a new segment from the side of an existing segment.

Ang II increased the number of UB tips and BP at 24 h (tips:24.3 � 1.1 versus 18.3 � 0.7, P � 0.01; BP: 14.4 � 0.6 versus 11.7 �

0.6, P � 0.01) and 48 h (tips: 30.2 � 1.3 versus 22.9 � 0.8, P �

0.01; BP: 21.3 � 0.9 versus 15.7 � 0.6, P � 0.01) compared withcontrol (Figure 2). During the 24-h incubation period, Ang IIcontents in the culture media (fmol/ml) decreased by approx-imately two-fold (time 0 2.4 � 0.27 � 106, at 6 h 1.9 � 0.19 � 106,12 h 1.7 � 0.23 � 106, 24 h 1.1 � 0.28 � 106) presumably as aresult of internalization of Ang II and/or its degradation bytissue proteases.

Role of Endogenous Ang II/AT1RTo examine the role of endogenous Ang II and AT1R in UB

branching in the metanephros, we used the AT1R antagonistcandesartan. Treatment of E11.5 metanephroi with candesartan(10�6 M) for 48 h decreased the number of UB tips and BPcompared with control (tips: 40 � 2 versus 52 � 2, P � 0.01; BP:29 � 2 versus 38 � 2, P � 0.01; n � 11/treatment group; Figure3). These findings are consistent with the results that were

J Am Soc Nephrol 17: 1005–1014, 2006 Angiotensin II in Kidney Development 1007

Figure 1. Immunolocalization of angiotensin II (Ang II) type 1 receptor (AT1R) protein in the fetal kidney on embryonic day 14(E14). AT1R is detectable using antibody concentrations of 1:200. (A) On E14, AT1R is abundantly expressed in the ureteric bud(UB) branches and to a lesser extent in glomeruli and the surrounding mesenchyme. Amp, UB ampulla. (B) The UB is positive forAT1R immunostaining on both apical and basolateral membranes. (C) Control section incubated after preadsorption of theprimary antibody with its immunogen demonstrates no specific staining.

Figure 2. Effect of medium alone (control) and Ang II on UB branching in Hoxb7–green fluorescence protein (GFP) metanephroithat were grown ex vivo for 48 h. (A) Representative sequential images of branching UB. Absolute number of UB tips and branchpoints (BP) is shown. (B) Bar graph shows the effect of medium or Ang II on the number of UB tips and BP at 24 and 48 h.


obtained in UB cells demonstrating that Ang II–induced effectson UB cell branching are mediated in part via AT1R (12). Col-lectively, these findings indicate that the stimulatory effect ofAng II on branching of UB cells that are cultured in three-dimensional gels is a physiologically relevant event because itcan be recapitulated in the intact organ culture.

Role of EGFRWe next examined the role of EGFR in Ang II–induced UB

branching morphogenesis in metanephric organ culture. Treat-ment with Ang II�AG1478 was associated with a significantlylower number of UB tips and BP at 24 h (tips: 17.2 � 0.9 versus24.5 � 1.2, P � 0.01; BP: 12.5 � 0.6 versus 18.5 � 0.8, P � 0.01)and 48 h (tips: 24.4 � 1.1 versus 34.8 � 2.2, P � 0.01; BP: 15.8 �

0.8 versus 24.1 � 1.1, P � 0.01) compared with Ang II alone(Figure 4). These findings are consistent with the results thatwere obtained in UB cells (see below) demonstrating that ty-rosine phosphorylation of EGFR is a key event in Ang II–mediated signaling to stimulate UB branching.

Effect of Ang II on UB Cell Branching In Vitro: Role of EGFRTo determine the functional importance of tyrosine phosphor-

ylation of EGFR elicited by Ang II in UB cell branching, weexamined the effect of the EGFR-specific tyrosine kinase inhibitorAG1478 on cell process formation in Ang II–treated UB cells thatwere cultured in three-dimensional collagen gels. Figure 5, Athrough F, shows phase-contrast photographs depicting the ef-fects of Ang II and EGF (positive control) on cell process formationand branching of UB cells 24 h after cells were plated into collagen

gels. Ang II or EGF caused extensive changes in cell processformation. In the presence of AG1478, Ang II– or EGF-treated cellsformed fewer and shorter branches as compared with cells thatwere exposed to Ang II or EGF alone (Figure 5).

Quantitative analysis revealed that Ang II or EGF increasedthe number of processes per cell compared with untreatedcontrol (Figure 5G). The data are presented as percentage ofcontrol (media � 0.5% FCS � 100%). EGF induced a 2.7-foldincrease in the number of processes (270 � 27 versus 100 � 16%;P � 0.001; n � 5 experiments). The effects of EGF on UB cellswere completely prevented by pretreatment with AG1478 (59 �

11 versus 270 � 27%; P � 0.001). At 10�5 and 10�6 M, Ang IIincreased the number of processes compared with control(194 � 28 and 185 � 28 versus 100 � 16%, respectively; P �

0.05). These effects were abrogated by pretreatment with eitherAG1478 (Ang II 10�5 M: 71 � 12 versus 194 � 28%, P � 0.01;Ang II 10�6 M: 47 � 11 versus 185 � 28%, P � 0.01) or the AT1Rantagonist candesartan (Ang II 10�5 M: 72 � 11 versus 194 �

28%, P � 0.01; Ang II 10�6 M: 73 � 14 versus 185 � 28%, P �

0.05). AG1478 alone did not have any effect on UB cell branch-ing compared with control (90 � 15 versus 100 � 16%).

Effect of Ang II on Tyrosine Phosphorylation of EGFR inUB Cells

To determine the role of EGFR in Ang II signal transduction,we first investigated the ability of Ang II to induce tyrosinephosphorylation of EGFR in UB cells. We used antibody di-rected against phosphorylated Tyr-1173 of the carboxy terminalof EGFR. The carboxy terminal tyrosine residue Tyr-1173 on

Figure 3. Effect of medium alone (control) or candesartan on UB branching in mouse metanephroi that were grown ex vivo for 48 h.After 48 h in culture, kidney explants were stained with anti-pancytokeratin antibody to label the UB, and the number of UB tipsand BP was compared between the two treatment groups. (A) Representative images of branching UB. Absolute number of UBtips and BP is shown. (B) Bar graphs show the effect of medium or candesartan on the number of UB tips and BP.


EGFR is the major site of autophosphorylation, which occurs as aresult of EGF binding (27). In addition, we used monoclonalanti-phosphotyrosine antibody. Positive controls included EGF-treated UB cells and cell extracts from EGF-stimulated A431 hu-man carcinoma cells that contained highly phosphorylated EGFR.As expected, EGF stimulated a robust five-fold increase in EGFRtyrosine phosphorylation (Figure 6). In comparison, Ang II stim-ulated a 1.64-fold increase in tyrosine phosphorylation of endog-enous EGFR compared with control (Ang II: 100 � 12 versus 164 �

15, P � 0.05; EGF: 100 � 12 versus 522 � 45, P � 0.001; n � 3;Figure 6, A, C, and E). In addition, Ang II caused a dose-depen-dent increase in tyrosine phosphorylation of EGFR during immu-noblotting with anti-phosphotyrosine mAb (Figure 6, B and D).These results indicate that Ang II activates the tyrosine phosphor-ylation of the EGFR in UB cells.

We next investigated the effect of Ang II on the nuclearlocalization of the phosphorylated EGFR in UB cells usinganti–phosphotyrosine-EGFR antibodies directed against twodifferent phosphotyrosine residues in carboxyl terminal ofEGFR, Tyr-1173 and Tyr-845. Tyr-845 resides in the activationloop of the EGFR kinase domain (16). Phosphorylation of Tyr-845 stabilizes the activation loop and is required for the mito-genic function of the EGFR (28). Treatment of UB cells with AngII or EGF caused an increase in tyrosine phosphorylation of thenuclear EGFR (Figure 6F). Nuclear EGFR may act as transcrip-tion factor to activate promitogenic genes (29).

DiscussionThe findings of this study demonstrate that Ang II stimulates

UB branching morphogenesis in the intact metanephric kidneycultured in vitro and that tyrosine phosphorylation of the EGFR

is a critical step in the signal transduction pathway down-stream of AT1R leading to UB branching. Genetic inactivationof the RAS genes in mice (angiotensinogen, renin, ACE, andAT1R) causes thinning of the medulla, hypoplastic papilla, andhydronephrosis (5–9). Also, homozygous mutation of AT2Rleads to ectopic UB budding, duplicated collecting system, andhydronephrosis (10). It is currently believed that the smallpapilla and hydronephrosis in RAS-mutant mice are the resultof defective formation and function of the smooth muscle layer inthe pelvic and ureteral wall, mimicking the findings seen in uri-nary obstruction (22). However, to date, the hypothesis that AngII can stimulate UB growth and branching directly and thereforepromote formation of the collecting system has not been testedrigorously. We previously provided two compelling lines of evi-dence to support this hypothesis. First, angiotensinogen and reninare expressed around the UB and its main branches, whereas ATreceptors are expressed on the basolateral and lumenal aspects ofthe UB branches. Second, Ang II acting via AT1R stimulatesbranching morphogenesis in UB cells that are grown in three-dimensional collagen matrix gels (12).

These findings extend our previous data by showing thatAng II stimulates the growth of the metanephric explant as aresult of enhanced UB branching and division into daughtertips. The mechanisms by which Ang II mediates its growth-promoting effects on UB branching are not yet known. How-ever, there are several possibilities. By stimulating the Gq-coupled AT1R, Ang II activates phospholipase C� (PLC�) andinositol triphosphate production, ultimately leading to in-creases in intracellular calcium and protein kinase C activation(30). Activation of the extracellular signal–regulated kinase–MAPK pathway by protein kinase C stimulates transcription of

Figure 4. Effect of Ang II alone or Ang II with EGF receptor (EGFR)-specific tyrosine kinase inhibitor AG1478 (AG) on UBbranching in Hoxb7-GFP metanephroi that were grown ex vivo for 48 h. (A) Representative sequential images of branching UB.Absolute number of UB tips and BP is shown. (B) Bar graph shows the effect of AG combined with Ang II or Ang II alone on thenumber of UB tips and BP.


cell-cycle progression genes, such as cyclin D1, through activa-tion of the transcription factor AP-1 (31). In another pathway,AT1R activates intracellular tyrosine kinases, such as Src, whichin turn activate the EGFR in a ligand-independent manner (32).Alternatively, Ang II may stimulate the proteolytic cleavage ofpro–heparin-binding epidermal growth factor (HBEGF) at thecell membrane (33).

In this study, we investigated the contribution of EGFR acti-vation to Ang II–induced stimulation of UB branching. EGFR isexpressed in renal collecting ducts (34), and its stimulationenhances branching morphogenesis in murine inner medullarycollecting duct cells in vitro (15,35). EGFR-mutant mice haveabnormal collecting ducts (36). It is interesting that reducedEGF mRNA levels are observed in the renal papilla of angio-tensinogen and AT1 null mice (37). In addition to EGF andEGF-like ligands (HBEGF or amphiregulin), EGFR is activatedby G protein–coupled receptors, including AT1R. For example,activation of AT1R can induce tyrosine phosphorylation andactivation of the EGFR in cultured aortic smooth muscle and

COS-7 cells (16,17). Consistent with these findings, we reporthere that Ang II enhances tyrosine phosphorylation of EGFR inUB cells and that inhibition of EGFR tyrosine kinase activityabrogates Ang II–induced branching in UB cells that are grownin three-dimensional gels. Of course, a more important ques-tion is whether the AT1R–EGFR signaling pathway contributesto the growth-promoting effects of Ang II in vivo. To addressthis question, we performed a real-time examination of UBgrowth in metanephric explants that expressed GFP in theUB and its derivatives. Exogenously added Ang II enhancedUB branching at 24 and 48 h. Furthermore, pretreatment ofmetanephric explants with an EGFR tyrosine kinase inhibitorprevented Ang II–induced UB branching. Collectively, thesefindings support the hypothesis that cooperation of AT1Rand EGFR signaling promotes the growth and developmentof the renal collecting system.

The signaling events that link Ang II receptors to EGFRtransactivation and UB branching morphogenesis remain to bedetermined. Several signaling pathways, including MAPK and

Figure 5. Effect of Ang II and EGF in UB cell branching in three-dimensional collagen gels. (A) Control (medium � 0.5% FBS). (B)Ang II. (C) Ang II with candesartan. (D) EGF. (E) EGF with AG1478. (F) AG1478 alone. (G) Effect of Ang II, EGF, and AG1478 aloneor combined with candesartan or AG1478 on cell process formation in UB cells 24 h after plating. Values are presented aspercentage change relative to control (medium � 0.5% FBS) with control equaling 100%. *P � 0.05 versus other groups.


phosphatidylinositol-3 kinase (PI3K)/Akt, have been shown tomediate the effects of AT1 on cell proliferation and hypertrophyin vascular smooth muscle, inner medullary collecting ductcells, and renal mesangial cells (15,38,39). In this regard, phar-macologic inhibition of extracellular signal–regulated kinase1/2 or PI3K attenuates EGF-induced renal epithelial morpho-genesis in vitro (15). Furthermore, inhibition of PI3K blocksglobal cell-line–derived neurotrophic factor (GDNF)—depen-dent UB branching in the metanephric kidney (40). Therefore,one of the possible mechanisms that lead to Ang II–stimulatedEGFR signaling may involve EGFR-mediated stimulation ofMAP kinase and PI3K/Akt pathways. It is interesting that EGFinduces phosphorylation of AT1 and leads to formation of amultireceptor complex that contains AT1 and activated EGFR.Phosphorylation of AT1 is prevented by AG1478 (41). Thesefindings suggest that cross-talk between the RAS and EGFR isbidirectional. A schematic summary of potential signaling

pathways that are involved in AT1R-EGFR cross-talk in UBbranching is presented in Figure 7.

ConclusionOur study demonstrates that UB-derived epithelial cells ex-

press AT1R and that Ang II stimulates UB branching morpho-genesis, a process that depends on tyrosine phosphorylation ofthe EGFR. Cooperation of tyrosine kinase and G protein–cou-pled receptor therefore is important in the control of normalkidney development.

AcknowledgmentsThis work was supported by National Institutes of Health grants P20

RR17659, DK-56264, and DK-62250.UB cells were a kind gift from Dr. Jonathan Barash (Columbia Uni-

versity). Hoxb7-GFP mice were a kind gift from Dr. Frank Costantini(Columbia University).

Figure 6. Ang II and EGF induce tyrosine phosphorylation of EGFR in UB cells. Quiescent mouse UB cells were treated or not with AngII (10�5 to 10�7 M) or EGF for 5 min. Cell lysates were subjected either to immunoblotting with anti-phosphotyrosine EGFR antibody(A) or to immunoprecipitation with anti-EGFR antibody followed by immunoblotting with anti-phosphotyrosine mAb (B). Afterstripping, membrane “A” was reprobed with �-actin antibody (C) and membrane “B” with anti-EGFR antibody (D). (E and F) Bargraphs represent densitometric analysis of the results that were obtained in three independent experiments using anti-phosphotyrosineEGFR antibody (“A”; E) and anti-phosphotyrosine mAb (“B”; F). Bars represent ratio of EGFR protein/�-actin protein (“A”) or ratio ofphosphorylated/total EGFR (“B”). Positive control, EGF-stimulated A431 cell lysate (Santa Cruz). *P � 0.01 versus other groups (F). (G)UB cells that were grown on coverslips were incubated with media (control), Ang II (10�6 M), or EGF (40 ng/ml) for 30 min and thenimmunostained with two different FITC-conjugated anti-phosphotyrosine EGFR antibodies (green) and propidium iodide (red). Theyellow signals indicate localization of tyrosine phosphorylated EGFR in the nucleus.


References1. Aufderheide E, Chiquet-Ehrismann R, Ekblom P: Epitheli-

al-mesenchymal interactions in the developing kidney leadto expression of tenascin in the mesenchyme. J Cell Biol 105:599–608, 1987

2. Ekblom P: Developmentally regulated conversion of mes-enchyme to epithelium. FASEB J 3: 2141–2150, 1989

3. Brenner BM, Garcia DL, Anderson S: Glomeruli and bloodpressure. Less of one, more the other? Am J Hypertens 1:335–347, 1988

4. Lisle SJ, Lewis RM, Petry CJ, Ozanne SE, Hales CN, For-head AJ: Effect of maternal iron restriction during preg-nancy on renal morphology in the adult rat offspring. Br JNutr 90: 33–39, 2003

5. Nagata M, Tanimoto K, Fukamizu A, Kon Y, Sugiyama F,Yagami K, Murakami K, Watanabe T: Nephrogenesis andrenovascular development in angiotensinogen-deficientmice. Lab Invest 75: 745–753, 1996

6. Takahashi N, Lopez ML, Cowhig JE Jr, Taylor MA, HatadaT, Riggs E, Lee G, Gomez RA, Kim HS, Smithies O: Ren1chomozygous null mice are hypotensive and polyuric, butheterozygotes are indistinguishable from wild-type. J AmSoc Nephrol 16: 125–132, 2005

7. Esther CR Jr, Howard TE, Marino EM, Goddard JM,Capecchi MR, Bernstein KE: Mice lacking angiotensin-con-

verting enzyme have low blood pressure, renal pathology,and reduced male fertility. Lab Invest 7: 953–965, 1996

8. Oliverio MI, Kim HS, Ito M, Le T, Audoly L, Best CF, HillerS, Kluckman K, Maeda N, Smithies O, Coffman TM: Re-duced growth, abnormal kidney structure, and type 2(AT2) angiotensin receptor-mediated blood pressure regu-lation in mice lacking both AT1A and AT1B receptors forangiotensin II. Proc Natl Acad Sci U S A 95: 15496–15501,1998

9. Tsuchida S, Matsusaka T, Chen X, Okubo S, Niimura F,Nishimura H, Fogo A, Utsunomiya H, Inagami T, IchikawaI: Murine double nullizygotes of the angiotensin type 1Aand 1B receptor genes duplicate severe abnormal pheno-types of angiotensinogen nullizygotes. J Clin Invest 101:755–760, 1998

10. Oshima K, Miyazaki Y, Brock JW, Adams MC, Ichikawa I,Pope JC 4th: Angiotensin type II receptor expression andureteral budding. J Urol 166: 1848–1852, 2001

11. Garcia-Villalba P, Denkers ND, Wittwer CT, Hoff C, Nel-son RD, Mauch TJ: Real-time PCR quantification of AT1and AT2 angiotensin receptor mRNA expression in thedeveloping rat kidney. Exp Nephrol 94: e154–e159, 2003

12. Iosipiv IV, Schroeder M: A role for angiotensin II AT1receptors in ureteric bud cell branching. Am J Physiol 285:F199–F207, 2003

Figure 7. Proposed signaling pathways involved in mediating the effects of Ang II on UB branching morphogenesis. AT1-mediatedactivation of multiple signaling pathways involving Jak2/STAT, Ras/extracellular signal–regulated kinase 1/2 (ERK1/2), phos-phatidylinositol-3 kinase/Akt, and EGF transactivation may stimulate UB cell proliferation, migration, tubulogenesis, andsurvival. Upon ligand binding, EGFR dimerize and autophosphorylate several tyrosine residues to generate docking sites forsignaling and adaptor proteins. The mechanisms of EGFR activation by AT1 and AT1-dependent effects on UB branching mayinvolve phosphorylation of Tyr-319 on carboxyl terminal of AT1; activation of tyrosine kinases Src and Pyk2; and recruitment ofSHP2, Shc, Grb2, and SOS with subsequent activation of the renin-angiotensin system. AT1 also can induce autocrine generationof Ang II from angiotensinogen (AGT) via Jak2/STAT3/STAT6 as shown in cardiomyocytes (42). This creates an autocrine loopthat amplifies the effects of Ang II on UB branching/stromal signaling. AT2-dependent activation of MAPK phosphatase(MAPKP) leads to ERK1/2 inactivation and may induce death-related signaling. The growth-promoting effects of AT2 may bemediated by Jak2/STAT-dependent activation of Pax2 in the UB. The ultimate effect of Ang II on UB branching and EGFRtransactivation may depend on the balance between AT1- and AT2-mediated actions.


13. Iosipiv IV: Cellular expression of the angiotensin type 2receptor (AT2) during murine organogenesis [Abstract].J Invest Med 50: 134A, 2002

14. Yosypiv IV, Schroeder M: Role of angiotensin type 2 (AT2)receptor in ureteric bud cell branching morphogenesis invitro [Abstract]. J Am Soc Nephrol 15: 419A, 2004

15. Karihaloo A, O’Rourke DA, Nickel C, Spokes K, CantleyLG: Differential MAPK pathways utilized for HGF- andEGF-dependent renal epithelial morphogenesis. J BiolChem 276: 9166–9173, 2001

16. Voisin L, Foisy S, Giasson E, Lambert C, Moreau P, Me-loche S: EGF receptor transactivation is obligatory for pro-tein synthesis stimulation by G protein-coupled receptors.Am J Physiol Cell Physiol 283: C446–C455, 2002

17. Seta K, Sadoshima J: Phosphorylation of tyrosine 319 of theangiotensin II type 1 receptor mediates angiotensin II-induced trans-activation of the epidermal growth factorreceptor. J Biol Chem 278: 9019–9026, 2003

18. Barasch J, Pressler L, Connor J, Malik A: A ureteric bud cellline induces nephrogenesis in two steps by distinct signals.Am J Physiol 271: F50–F61, 1996

19. Piscione TD, Phan T, Rosenblum ND: BMP7 controls col-lecting tubule cell proliferation and apoptosis via Smad1-dependent and -independent pathways. Am J Physiol 280:F19–F33, 2001

20. Zent R, Bush KT, Pohl ML, Quaranta V, Koshikawa N,Wang Z, Kreidberg JA, Sakurai H, Stuart RO, Nigam SK:Involvement of laminin binding integrins and laminin-5 inbranching morphogenesis of the ureteric bud during kid-ney development. Dev Biol 238: 289–302, 2001

21. Srinivas S, Goldberg MR, Watanabe T, Dagati V, Al-Awqati Q, Costantini F: Expression of green fluorescentprotein in the ureteric bud of transgenic mice: A new toolfor the analysis of the ureteric bud morphogenesis. DevGenet 24: 241–251, 1999

22. Miyazaki Y, Tsuchida S, Fogo A, Ichikawa I: The renallesions that develop in neonatal mice during angiotensininhibition mimic obstructive nephropathy. Kidney Int 55:1683–1695, 1999

23. Gupta IR, Lapointe M, Yu OH: Morphogenesis duringmouse embryonic kidney explant culture. Kidney Int 63:365–376, 2003

24. Berellini G, Gruciani G, Mannhold R: Pharmacophore,drug metabolism, and pharmacokinetics models on non-peptide AT1, AT2, and AT1/AT2 angiotensin II receptorantagonists. J Med Chem 48: 4389–4399, 2005

25. Yosipiv IV, El-Dahr SS: Activation of angiotensin-generat-ing systems in the developing rat kidney. Hypertension 27:281–286, 1996

26. Watanabe T, Costantini F: Real-time analysis of uretericbud branching morphogenesis in vitro. Dev Biol 271: 98–108, 2004

27. Sakaguchi K, Okabayashi Y, Kido Y, Kimura S, MatsumuraY, Inushima K, Kasuga M: Shc phosphotyrosine-bindingdomain dominantly interacts with epidermal growth fac-tor receptors and mediates Ras activation in intact cells.Mol Endocrinol 12: 536–543, 1998

28. Tice DA, Biscardi JS, Nickles AL, Parsons SJ: Mechanism ofbiological synergy between cellular Src and epidermalgrowth factor receptor. Proc Natl Acad Sci U S A 96: 1415–1420, 1999

29. Lin SY, Makino K, Xia W, Matin A, Wen Y, Kwong KY,

Bourguignon L, Hung MC: Nuclear localization of EGFreceptor and its potential new role as a transcription factor.Nat Cell Biol 3: 802–808, 2001

30. Berry C, Touyz R, Dominiczak AF, Webb RC, Johns DG:Angiotensin receptors: Signaling, vascular pathophysiol-ogy, and interactions with ceramide. Am J Physiol 281:H2337–H2365, 2001

31. Watanabe G, Lee RJ, Albanese C, Rainey WE, Batle D,Pestell RG: Angiotensin II activation of cyclinD1-depen-dent kinase activity. J Biol Chem 271: 22570–22577, 1996

32. Eguchi S, Numaguchi K, Iwasaki H, Matsumoto T, Ya-makawa T, Utsunomiya H, Motley ED, Kawakatsu H,Owada KM, Hirata Y, Marumo F, Inagami T: Calcium-dependent epidermal growth factor receptor transactiva-tion mediates the angiotensin II-induced mitogen-acti-vated protein kinase activation in vascular smooth musclecells. J Biol Chem 273: 8890–8896, 1998

33. Schafer B, Gschwind A, Ullrich A: Multiple G-protein-coupled receptor signals converge on the epidermalgrowth factor receptor to promote migration and invasion.Oncogene 23: 991–999, 2004

34. Bernardini N, Bianchi F, Lupetti M, Dolfi A: Immunohisto-chemical localization of the epidermal growth factor, trans-forming growth factor alpha, and their receptor in the humanmesonephros and metanephros. Dev Dyn 206: 231–238, 1996

35. Sakurai H, Tsukamoto T, Kjelsberg CA, Cantley LG,Nigam SK: EGF receptor ligands are a large fraction of invitro branching morphogens secreted by embryonic kid-ney. Am J Physiol 273: F463–F472, 1997

36. Threadgill DW, Dlugosz AA, Hansen LA, Tennenbaum T,Lichti U, Yee D, LaMantia C, Mourton T, Herrup K, HarrisRC: Targeted disruption of mouse EGF receptor: Effect ofgenetic background on mutant phenotype. Science 269:230–234, 1995

37. Niimura F, Labosky PA, Kakuchi J, Okubo S, Yoshida H,Oikawa T, Ichiki T, Naftilan AJ, Fogo A, Inagami T: Genetargeting in mice reveals a requirement for angiotensin in thedevelopment and maintenance of kidney morphology andgrowth factor regulation. J Clin Invest 96: 2947–2954, 1995

38. Huwiler A, Stabel S, Fabbro D, Pfeilschifter J: Platelet-derived growth factor and angiotensin II stimulate themitogen-activated protein kinase cascade in renal mesan-gial cells: Comparison of hypertrophic and hyperplasticagonists. Biochem J 305: 777–784, 1995

39. Saward L, Zahradka P: Angiotensin II activates phospha-tidylinositol 3-kinase in vascular smooth muscle cells. CircRes 81: 249–257, 1997

40. Tang MJ, Cai Y, Tsai SJ, Wang YK, Dressler GR: Uretericbud outgrowth in response to RET activation is mediatedby phosphatidylinositol 3-kinase. Dev Biol 243: 128–136,2002

41. Olivares-Reyes JA, Shah BH, Hernandez-Aranda J, Garcia-Gaballero A, Farshori MP, Gatcia-Sainz JA, Catt KJ: Ago-nist-induced interactions between angiotensin AT1 andepidermal growth factor receptors. Mol Pharmacol 68: 356–364, 2005

42. Fukizawa J, Booz GW, Hunt RA, Shimizu N, Karoor V,Baker KM, Dostal DE: Cardiotrophin-1 increases angio-tensinogen mRNA in rat cardiac myocytes through STAT3:An autocrine loop for hypertrophy. Hypertension 35: 1191–1196, 2000


angiotensin ii type 1 receptor–egf receptor cross-talk...

Documents