Expression of Angiopoietin-1, Angiopoietin-2, and the Tie-2 Receptor Tyrosine Kinase during

Expression of Angiopoietin-1, Angiopoietin-2, and the Tie-2Receptor Tyrosine Kinase during Mouse Kidney Maturation

HAI TAO YUAN,* CHITRA SURI, † GEORGE D. YANCOPOULOS,† andADRIAN S. WOOLF**Nephrourology Unit, Institute of Child Health, University College London Medical School, London, UnitedKingdom; and†Regeneron Pharmaceuticals, Inc., Tarrytown, New York.

Abstract.The Tie-2 receptor tyrosine kinase transduces embry-onic endothelial differentiation, with Angiopoietin-1 (Ang-1)acting as a stimulatory ligand and Ang-2 postulated to be anaturally occurring inhibitor. Expression of these genes wassought during mouse kidney maturation from the onset ofglomerulogenesis (embryonic day 14 [E14]) to the end ofnephron formation (2 wk postnatal [P2]), and during medullarymaturation into adulthood (P8). Using Northern and slot blot-ting of RNA extracted from whole organs, these three geneswere expressed throughout the experimental period with peaklevels at P2 to P3. Byin situ hybridization analysis at E18, P1,and P3, Ang-1 mRNA was found to localize to condensingrenal mesenchymal cells, proximal tubules, and glomeruli in

addition to maturing tubules of the outer medulla. In contrast,Ang-2 transcripts were more spatially restricted, being detectedonly in differentiating outer medullary tubules and the vasarecta bundle area. Usingin situ hybridization and immunohis-tochemistry, Tie-2 was detected in capillaries of the nephro-genic cortex, glomerular tufts, cortical interstitium, and me-dulla including vessels in the vasa recta. Using Westernblotting of protein extracted from whole organs, Tie-2 proteinwas detected between E14 and P8 with tyrosine phosphory-lated Tie-2 evident from E18. These data are consistent withthe hypothesis that Tie-2 has roles in maturation of bothglomeruli and vasa rectae.

Each adult kidney receives 10% of the cardiac output, a highblood flow supplying glomerular capillaries, which ultrafilterplasma, and also peritubular capillaries, including the vasarectae, which contribute to the countercurrent system (1,2).The mouse metanephros forms on embryonic day 11 (E11)when the ureteric bud penetrates avascular renal mesenchyme(3,4). The bud branches into collecting ducts, and mesenchymedifferentiates into nephrons (glomerular, proximal tubule, andloop of Henle epithelia). Capillaries are prominent in meta-nephric stroma at E12 and the first vascular glomeruli form atE14 (4). In mice, 80% of glomeruli form after birth, a processcompleted by 2 wk postnatal (5). Vasa rectae mature alongsideloops of Henle, which grow into the medulla even afternephron generation has ended (6). The focus of study is cur-rently on the molecules that direct differentiation of thesemicrocirculations (4,7).

Various classes of molecule are implicated in embryonicendothelial development, including transcription factors (8)and cell adhesion molecules (9). Importantly, phenotypes ofnull mutant mice indicate that the sequential expression ofspecific growth factors and their receptor tyrosine kinases

(RTK) determine survival, proliferation, differentiation, andmorphogenesis of endotheliain vivo (10,11). One RTK groupincludes vascular endothelial growth factor receptor-1(VEGFR-1/Flt-1) and VEGFR-2 (Flk-1/KDR), which are, re-spectively, expressed from mouse E7.5 and E7.0 in yolk sacblood islands and endothelial precursors of the aorta and heart(12). VEGFR-2 is required for endothelial formation, whereasVEGFR-1 modulates vessel assembly (13,14). Vascular endo-thelial growth factor (VEGF), VEGFR-1, and VEGFR-2 areexpressed by uninduced mouse metanephric mesenchyme (15)and later in murine (15–17) and human (18) kidney develop-ment. Functional experiments implicate VEGF in glomerulo-genesisin vivo (19) and in hypoxia-driven metanephric endo-thelial proliferationin vitro (20).

Another class of endothelial RTK is called Tie (tyrosinekinase that containsimmunoglobulin-like loops andepidermalgrowth factor similar domains) (21,22). The onset of Tie-1 andTie-2 (previously called Tek) expression postdate VEGFR-2but precede von Willebrand factor (12,23,24). Tie-1 is anorphan RTK, and null mutant mice die in late gestation withimpaired vessel integrity (25,26). The gene is expressed inmouse metanephric mesenchyme from the inception of nephro-genesis and in interstitial and glomerular capillaries at laterstages of development (15).

Angiopoietin-1 (Ang-1) is a ligand for the RTK called Tie-2(27). Although Ang-1 does not cause significant endothelialproliferation in culture (27), it does induce sproutingin vitro,an activity synergizing with VEGF (28). Ang-1 (29) and Tie-2(25,30) null mutant mice die by E12 with a homogenization ofvessel caliber and poor branching. In addition, endothelia were

Received June 29, 1998. Accepted February 3, 1999.Correspondence to Dr. Hai Tao Yuan, Nephrourology Unit, Institute of ChildHealth, University College London Medical School, 30 Guilford Street, Lon-don WC1EN 1EH, United Kingdom. Phone: 0171 242 9789; Fax: 0171 9160011; E-mail: [email protected]

1046-6673/1008-1722Journal of the American Society of NephrologyCopyright © 1999 by the American Society of Nephrology

J Am Soc Nephrol 10: 1722–1736, 1999

separated from sparse pericytes and growth-retarded myocar-dium. It was speculated that Tie-2 signaling caused endothelialstabilization with secondary maturational effects on differen-tiating pericytes and cardiomyocytes, effects elicited by endo-thelial-derived molecules including platelet-derived growthfactor-BB, heparin-binding epidermal growth factor, and neu-regulin (31). Ang-2, a homologue of Ang-1, binds Tie-2 butdoes not cause tyrosine phosphorylation in cultured endothelia(32). In endothelia, Ang-2 antagonizes Ang-1-induced Tie-2phosphorylation, while Ang-2 overexpressionin vivo disruptsembryonic vessel formation and causes cardiac defects resem-bling Tie-2 and Ang-1 null mutants (32). Furthermore, Wit-zenbichleret al. (33) found that Ang-1 was chemotactic forendothelial cells and that Ang-2 blocked this effect. Hence,Ang-2 may be a naturally occurring antagonist of Ang-1.

Evidence is emerging about the sites of angiopoietin expres-sion. Daviset al. (27) localized Ang-1 mRNA in embryonicmice to cells near developing endothelia, which themselvesexpressed Tie-2, with a similar pattern in embryonic myocar-dium and endocardium. Maisonpierreet al. (32) reported thatAng-2 transcripts localized to areas in mouse embryos “likelyto be endothelial cells or closely associated pericytes” with anexpression pattern more spatially restricted than Ang-1. It hasalso been reported that subsets of endothelial cells expressAng-2 in vitro, whereas smooth muscle cells express Ang-1and Ang-2 (33,34). These reports are consistent with the para-crine modulation of Tie-2 by both Ang-1 and Ang-2, with theadditional possibility of an autocrine action of Ang-2 on en-dothelia.

To date, angiopoietin expression during renal ontogeny hasnot been investigated. The experiments in our current studyprovide novel information on Ang-1, Ang-2, and Tie-2 expres-sion during mouse kidney maturation.

Materials and MethodsReagents

Reagents were obtained from Sigma Chemical Co. unless other-wise specified.

TissuesWe studied a normal mouse strain (CD1) with the day of the

vaginal plug designated E0. In these mice, the metanephros forms onE11 and the first glomeruli with primitive capillary loops forms onE14 (4). New layers of nephrons continue to be generated postnatallyfor 1 to 2 wk. The ages examined in the study were E14, E15, E16,E17, E18 and postnatal weeks 1, 2, 3, 4, and 8 (P1, P2, P3, P4, andP8).

Northern and Slot Blotting for Angiopoietin and TieGenes

Total RNA was isolated with Tri-Reagent from kidneys at 10 timepoints (E14, E15, E16, E17, E18, P1, P2, P3, P4, and P8). Separatepools of 10 to 20 embryonic kidneys were used for the prenatal stages,and a single organ was used for each postnatal time point, with thewhole experiment performed in triplicate. For Northern blotting, 20mg of total RNA was electrophoresed in 1% formaldehyde-denaturedagarose gel in 13 3-[N-morpholino]propanesulfonic acid buffer,transferred onto Hybond-N membrane (Amersham Pharmacia Bio-

tech, Buckinghamshire, United Kingdom), and fixed with UV-Stratal-inker (Stratagene, La Jolla, CA). For slot blotting, 10mg of denaturedtotal RNA (in a total volume of 200ml) was transferred onto pre-wetHybond-N membrane using Bio-Dot SF apparatus (Bio-Rad, Hert-fordshire, United Kingdom) and fixed with UV-Stratalinker. Plasmidswith mouse cDNA inserts were: Tie-2 (1.2 kb; kindly provided by W.Risau, Max-Planck-Institute, Bad Nauheim, Germany), Tie-1 (238 bp;isolated from metanephric cDNA, with identity confirmed by se-quencing; (15)), as well as Ang-1 (560 bp) and Ang-2 (680 bp)(Regeneron Pharmaceuticals, Tarrytown, NY). Inserts were isolatedafter digestion with appropriate restriction enzymes, and randomprimer labeling was performed with Prime-a-Gene labeling system(Promega). Unincorporated labeled-dCTP was removed by using apush-column (Stratagene). Blots were prehybridized with Quick-Hybsolution (Stratagene) at 65°C for 30 min and hybridized with specificprobes at 65°C for 2 h. After hybridization, the filters were washedtwice with 23 SSC at room temperature for 30 min and once with0.13 SSC/0.1% sodium dodecyl sulfate (SDS) at 65°C for 30 min.X-films were exposed to filters for 12 to 48 h at270°C. The equalityof loading for Northern blotting was confirmed by visualizing 28S and18S rRNA. For slot blotting, 28S rRNA signals were obtained byhybridizing the blots with 28S oligonucleotide (10 pmol/ml) as de-scribed (35). The signal intensity relative to 28S rRNA of each samplewas measured with an image analysis program (Phoretix 1D, New-castle upon Tyne, United Kingdom), and the ratio of target signal to28S rRNA of E14 was used as arbitrary value 1. The slot blottingresults were expressed as the mean and SD of the three assaysperformed at each time point. Because of the probable differentaffinities of the angiopoietin and Tie probes to their target mRNA, aswell as the variation in labeling efficiency of different probes, it wasnot possible to compare the absolute amounts of transcripts betweengenes.

Tie-2 Immunoprecipitation and Western Blotting forTie-2 and Phosphotyrosine

We currently have no antibodies available for Ang-1 and Ang-2and therefore these experiments were restricted to Tie-2 and thephosphorylated RTK. Separate pools of 10 to 20 embryonic kidneyswere used for the prenatal stages (E14, E16, E17, and E18), and asingle organ was used for each postnatal time point (P1, P2, P3, P4,and P8), with the whole experiment performed in triplicate. Kidneyswere homogenized in radioimmunoprecipitation assay buffer (30ml/ml of 2.2 mg/ml aprotinin; 10ml/ml of 10 mg/ml phenylmethyl-sulfonyl fluoride; 10ml/ml of 100 mM sodium orthovanadate) at 4°C,and the supernatants were collected by centrifugation at 13,000 rpmfor 30 min. Supernatants were used for protein determination (BCAprotein assay; Pierce, Rockford, IL). To 0.2-mg protein samples (in0.5 ml), 0.2mg of rabbit antibody raised against amino acids 1103–1122 in the carboxy terminus of human Tie-2 (Santa Cruz Biotech-nology, Santa Cruz, CA) was added and incubated at 4°C for 2 h.After this, 8ml of Protein A-agarose (Santa Cruz Biotechnology) wasadded and incubated at 4°C with agitation for at least 4 h. Beads werewashed with radioimmunoprecipitation assay buffer four times andcollected by centrifugation at 2500 rpm for 5 min at 4°C. After thefinal wash, the supernatant was aspirated and discarded, and the pelletwas resuspended in 30ml of 13 electrophoresis buffer. Samples wereboiled for 3 min and separated on an SDS-8% polyacrylamide gelelectrophoresis gel. Proteins were transferred to nitrocellulose mem-branes (Amersham Pharmacia Biotech) by electroblotting (Bio-Rad).Blots were blocked overnight at 4°C with 5% (wt/vol) fat-free milkpowder, 0.3% (vol/vol) Tween 20 in phosphate-buffered saline, and

J Am Soc Nephrol 10: 1722–1736, 1999 Ang-1, Ang-2, and the Tie-2 Receptor 1723

then incubated with rabbit anti-human Tie-2 antibody (1:1000 inblocking solution) for 2 h at4°C. After washing in blocking solution,blots were incubated for 30 min with horseradish peroxidase-conju-gated second antibodies diluted 1:1500 in blocking solution. Blotswere washed three times with blocking solution and once with phos-phate-buffered saline. Immunocomplexes were developed using anenhanced horseradish peroxidase/luminol chemiluminescence reagent(Du Pont New England Nuclear, Boston, MA). Proteins were sizedwith Rainbow markers (Amersham Pharmacia Biotech). After visu-alizing the immobilized proteins, the antibody complex was strippedby rocking the blot in 200 mM glycine, 200 mM NaCl, pH 2.3, for 2 hat room temperature. The blot was then reprobed with antiphospho-tyrosine antibody (PY-99, Santa Cruz) using the same proceduredescribed above.

In Situ Hybridization for Ang-1, Ang-2, and Tie-2Kidney sections were analyzed in longitudinal and cross-sectional

planes in organs harvested at E18, P1, and P3 from three animals ateach stage. Mouse Tie-2, Ang-1, and Ang-2 cDNA plasmids werelinearized with restriction enzymes, and sense and antisense uridinetriphosphate-digoxigenin-labeled riboprobes were prepared using lin-earized plasmid cDNA as template, the appropriate RNA polymerase,and the conditions recommended in the Dig RNA labeling kit (Boehr-inger Mannheim, Sussex, United Kingdom). Tie-2 transcript wassubjected to a limited alkaline hydrolysis to produce a 400-bp probe.In situ hybridization was performed as described (36) with minormodifications. Paraffin-embedded tissue sections (7mm) were dew-axed by treatment with Histoclear (DiaMed, Windham, ME), treatedwith proteinase K (20mg/ml) at 37°C for 10 min, and post-fixed in 4%paraformaldehyde. Sections were covered with 50ml of prehybrid-ization mix (50% vol/vol formamide, 53 SSC, 13 Denhardt’s re-agent, heat-denatured salmon sperm DNA 0.1 mg/ml, and 10% wt/voldextran sulfate) for 30 min at 65°C, followed by 50ml of the samemixture containing the digoxigenin-labeled riboprobe. A glass cover-slip was applied and hybridization was allowed to occur at 65°Covernight. Sections were washed at 65°C with 25% formamide in 23SSC for 1 h, 13 SSC and 0.1% SDS for 30 min, and 0.13 SSC and0.1% SDS for 30 min. Hybridized probe was detected by incubationwith antidigoxigenin antibody (1:1000) conjugated to alkaline phos-phatase, followed by the chromogen solution, nitroblue tetrazolium,and 5-bromo-4-chloro-3-indolylphosphate toluidine. This techniqueproduced a blue/purplein situ hybridization signal. Slides werewashed and mounted with Citifluor (Chemical Labs, London, UnitedKingdom). In some samples, counterstaining with eosin was per-formed to facilitate cellular localization of thein situ hybridizationsignal; however, this was only feasible when the signal was intensebecause counterstaining was found to obscure weakerin situ signals.Controls run in parallel with each experiment included tissue sectionsthat were incubated in hybridization mix without riboprobe added orhybridized in an identical manner with digoxigenin-labeled senseriboprobe. It should be noted that the preparation of sections for theinsitu technique resulted in tissue details that were somewhat lessdistinct compared with the other histologic techniques described inthis article.

Immunohistochemistry for Tie-2, CD34, and Tamm–Horsfall Glycoprotein

Organs were fixed in 4% paraformaldehyde, and 8-mm paraffinsections were treated with trypsin (1 mg/ml) for 10 min at 37°C.Endogenous peroxidase was quenched with 3% H2O2 in methanol for30 min at room temperature, and sections were blocked in 10% goat

serum with 0.1% Tween 20. They were reacted with rabbit anti-human Tie-2 (1:2000; Santa Cruz), a rat anti-mouse monoclonalantibody to the endothelial molecule CD34 (37) (1:50; Pharmingen,San Diego, CA), and a sheep anti-human antibody to the loop ofHenle molecule Tamm–Horsfall glycoprotein (38) (1:50; Europa Bio-products, Ltd., Cambridge, United Kingdom). Bound antibodies weredetected with a streptavidin-biotin peroxidase system (ABC kit;DAKO, High Wycombe, United Kingdom). This procedure produceda brown positive signal. Controls included omission of the primaryantibodies and preincubation of Tie-2 antibody with Tie-2 peptide(Santa Cruz). Some sections were counterstained with hematoxylin,eosin, or periodic acid-Schiff to enhance, respectively, cell nuclei,cytoplasm, or proximal tubule brush border.

ResultsNorthern and Slot Blotting

Figure 1A depicts a Northern blot of gene expression at 10stages from E14 to P8. As assessed by ethidium bromidestaining for 28S and 18S rRNA, the loading of all lanes wassimilar. Major transcripts were observed for Ang-1 (4.1 kb),Ang-2 (2.8 kb), Tie-1 (4.2 kb), and Tie-2 (4.5 kb) at sizessimilar to those reported in other studies (32,39), with the fourgenes expressed at all times points studied. Figure 1B shows adensitometric analysis of gene expression measured by slotblotting after standardization to 28S rRNA. Based on the meanstandardized levels of three separate experiments, the transcriptlevels peaked at P2 to P3 for Ang-1, Ang-2, and Tie-2, and atP2 for Tie-1. Thereafter, the levels of mRNA of all four genesdecreased. The relative rise of transcript levels for Ang-1,Tie-1, and Tie-2 between E14 and P2 was approximatelytwofold but was 25-fold for Ang-2. No direct comparisonscould be made between mRNA levels of these genes becausethe signals generated in slot blotting are determined not only bytranscript levels, but also by the individual hybridization effi-ciency and radioactive labeling of each probe.

Western Blotting for Tie-2 and PhosphotyrosineFigure 2 depicts Western blotting for Tie-2 and the tyrosine

phosphorylated protein after immunoprecipitation with an anti-Tie-2 antibody. Antibody to Tie-2 detected a band at 140 kD,as described for other tissues (24), at all stages examined (E14,E16, E17, E18, P1, P2, P3, P4, and P8). To investigate theinvivo activation of Tie-2, we stripped and reprobed the sameblots with a phosphotyrosine antibody, and a definite band fortyrosine phosphorylated Tie-2 was detected from E18 onward.Although no attempt was made to formally measure the inten-sity of bands, the patterns described above were the same inthree separate experiments using proteins extracted from poolsof E14 to E18 metanephroi or single organs at postnatal stages.

Overview of in Situ HybridizationAnalyses were performed to localize Ang-1, Ang-2, and

Tie-2 mRNA at E18, P1, and P3. Figures 3 through 8 depictresults that are representative of three sets of experiments.Sense ribroprobes for the genes generated low backgroundsignals, which were used to assess the significance of signalson adjacent tissue sections that were generated by the appro-priate antisense probes. Because of the numerous ages and

1724 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 1722–1736, 1999

tissue regions analyzed, we have shown select, but represen-tative, data for Ang-1, Ang-2, and Tie-2 sense probes. Table 1summarizes the main sites of gene expression, which aredescribed in detail below.

In Situ Hybridization at E18At E18, the outer cortex of the kidney contained condensing

mesenchymal cells adjacent to ureteric bud branch tips andnephron precursors including vesicles and primitive glomeruli

Figure 1.Northern and slot blotting. Time points studied were embryonic day 14 (E14) to E18 and postnatal week 1(P1) to P8. (A) The toppanels show Northern blotting. Note the presence of major transcripts for the four genes at each stage studied (angiopoietin-1 [Ang-1] at 4.1kb, Ang-2 at 2.8 kb, [Tie-1] at 4.2 kb, and Tie-2 at 4.5 kb). The lowest panel in A shows 28S and 18S rRNA in an ethidium bromide-stainedagarose gel. Note the approximate equality of loading. (B) Densitometric analysis of mRNA for Ang-1, Ang-2, Tie-1, and Tie-2 after beingstandardized to 28S rRNA as assessed by slot blotting. Each value represents the mean of three measurements and the bar is the SD. The meanderived value for each gene at the E14 stage was arbitrarily designated as “1.” Note the approximately twofold increase in signal for Ang-1,Ang-2, and Tie-2 between E14 and P2, whereas the signal for Ang-2 increased approximately 25-fold in the same period. No direct comparisonscould be made between mRNA levels of these genes because the signals generated in slot blotting were determined not only by absolutetranscript levels, but also by individual hybridization efficiency and radioactive labeling of the probe for each gene.


(Figure 3A). Ang-1 mRNA was expressed by condensing renalmesenchyme, whereas ureteric bud branch tips did not expresssignificant levels of this transcript (Figure 3, B and C). A faintsignal for Ang-1 also localized over primitive glomeruli andproximal tubules deeper in the cortex (not shown). No specificsignal was observed in the nephrogenic zone with the antisenseAng-2 probe (Figure 3D) when compared with the Ang-2 senseprobe control (not shown). Tie-2 transcripts were identified incells located between primitive tubules, in configurations con-sistent with differentiating capillaries (Figure 3E). Using thecurrent methodology, Tie-2 mRNA could not be detected overprimitive glomeruli (Figure 3E). When sections of the neph-rogenic zone were reacted with CD34 antibody, weak positivesignals were detected between developing tubules (Figure 3F).At E18, the outer medulla contained loose associations of loopsof Henle growing toward the inner medulla. Thick ascendinglimbs immunostained for Tamm–Horsfall glycoprotein, as didthe terminal portion of descending limbs (Figure 4, A and B).These loops were flanked by collecting ducts and CD34-positive capillaries (Figure 4, C and D). The diameter of all ofthese tubules was relatively uniform compared with their ma-ture derivatives. Ang-1 mRNA was expressed by most tubulesin this zone (Figure 4E), whereas Ang-2 transcripts wererestricted to a subset of structures (Figure 4F). Tie-2 mRNAlocalized to capillaries aligned in parallel to tubules (Figure4G). Sense control for Tie-2 are shown in Figure 4H. Nosignificant signal was found in sense controls for Ang-1 andAng-2 (not shown).

In Situ Hybridization at P1By P1, the nephrogenic zone was no longer prominent, and

cortex and medullary structures had become more distinct withaggregations of loops of Henle noted in the outer medulla. Inthe deep cortex, Ang-1 transcripts were prominent in proximaltubules (Figure 5, A and B) with no significant signal gener-ated by the Ang-1 sense probe (data not shown). A weak signalfor Ang-1 was also noted in cells in the outermost cortex and

glomeruli (not shown). Usingin situ hybridization, Ang-2mRNA was not detected in the epithelia of the cortex (Figure5C), while signals for Tie-2 were found in interstitial capillar-ies (Figure 5D) and in the endothelial lining of cortical arteries(Figure 5D). Figure 6 depicts the outer medulla at P1. In Figure6A, loops of Henle were evident, often with their “U” turns inthis region. Ascending limbs stained for Tamm–Horsfall gly-coprotein, as did the terminal portion of the thinner, descendinglimbs. The region was rich in CD34-positive capillaries alignedin parallel and at right angles to tubules (Figure 6B). A weaksignal for Ang-1 mRNA was detected in diverse tubules in thisregion (Figure 6, C and D), while an intense signal for Ang-2was restricted to the thinnest of these structures (Figure 6, Eand F), which were therefore likely to represent thin limbs ofloops of Henle derived from the outermost nephrons or matur-ing vasa rectae. Tie-2 transcripts were expressed in capillariesrunning parallel to the tubules (Figure 6, G and H) and by innermedullary capillaries (not shown).

In Situ Hybridization at P3In the P3 cortex, Ang-1 was detected in glomerular tufts

with weaker signals in Bowman’s capsules and proximal tu-bules (Figure 7, A and B). Although all cells of the glomerulusappeared positiveversusthe Ang-1 sense probe (Figure 7A),the highest signal was localized to cells at the periphery of thetuft, where podocytes reside. As for E18 and P1, Ang-2 ex-pression could not be detected in epithelia of the P3 cortex (notshown). Tie-2 transcripts were detected in cells inside glomer-ular tufts, where endothelia and mesangial cells reside (Figure7, C and D). For comparison, CD34 immunostaining of endo-thelial cells within glomerular tufts is depicted in Figure 7E.Tie-2 transcripts were also detected in arterial endothelia (Fig-ure 7F). In the outer medulla, the mature arrangement of thevasa recta was evident at this stage (Figure 8). Ascending limbsof loops of Henle, collecting ducts, and interbundle capillariessurrounded vasa recta bundles. In rodents, these bundles con-tain three structures (1,2): (1) descending vasa recta capillaries

Figure 2. Western blotting of Tie-2. Western blotting with anti-Tie-2 (top panel) and anti-phosphotyrosine (bottom panel) antibodies afterimmunoprecipitation for Tie-2. The stages examined were E14, E16, E17, E18, P1, P2, P3, P4, and P8, and the results shown are representativeof three separate experiments, using protein extracted from different pools of metanephroi or postnatal organs. Antibody to Tie-2 detected aband at 140 kD at all stages, and a definite band for tyrosine phosphorylated Tie-2 was visualized from E18 onward. Size markers for 160 kDand 105 kD are indicated.


comprised of endothelium and surrounding pericytes; (2) as-cending vasa recta capillaries comprised of highly fenestratedendothelia lacking pericytes; and (3) descending thin limbs ofloops of Henle derived from short-looped nephrons that incor-

porate into the outer part of the vasa recta bundle. Figure 8Ashows Tamm–Horsfall immunostained tubules surrounding thevasa recta bundle region, while Figure 8B shows CD34 immu-nostained capillaries in the vasa recta and in an interbundle

Figure 3. Nephrogenic cortex at E18. Panels A and F were counterstained with hematoxylin, but Panels B through E, representingin situhybridization experiments, were not counterstained. (A) The nephrogenic cortex contained a rim of condensing mesenchymal cells (arrow)adjacent to ureteric bud branches (u) and nephron precursors including vesicles and primitive glomeruli (g). (B) Ang-1 mRNA was expressedby condensing mesenchyme cells (arrow) and was expressed faintly in early glomeruli (g), whereas ureteric bud branches (u) were negative.Red blood cells are indicated by arrowheads. (C) No significant signal was detected with the Ang-1 sense probe. (D) Hybridization with theAng-2 antisense probe did not produce a signal that significantly differed from the Ang-2 sense control (not shown). (E) Tie-2 transcripts werenoted in cells (arrows) located between differentiating epithelia, in configurations consistent with cells of the capillary lineage. Tie-2 transcriptswere not detected in primitive glomeruli (g) using this methodology. (F) Using an antibody to the endothelial marker CD34, a faint brown signalwas detected in cells (arrows) around developing epithelia. Primitive glomeruli (g) lacked CD34 expression. Bars, 12mm in all frames.


Figure 4.Outer medulla at E18. Panels A, C, E, F, and H show cross sections, whereas Panels B, D, and G are longitudinal sections with themedulla on the right. Panels A through D were counterstained with hematoxylin, whereas Panels E through H were not counterstained so thatthe blue/purple color of thein situ hybridization signal was not obscured. (A and B) The region is comprised of loose associations of loops


plexus. The antisense probe for Ang-1 produced a barelysignificant signal within the bundle in cells closely associatedwith capillaries (Figure 8, C and D).In situ hybridization forAng-2 showed strong signals in the bundle region in cellsclosely associated with vasa recta capillaries (Figure 8E) and inthin tubules on the periphery of the bundle (Figures 8, E and F),which may represent descending thin loops of Henle. Strongsignals for Tie-2 transcripts were detected closely associatedwith capillaries of the vasa recta, but Tie-2 transcripts weresparse or absent in the interbundle region (Figure 8, G and H).

Immunohistochemistry for Tie-2At E18, linear Tie-2 immunostaining was noted in an inter-

stitial distribution in the nephrogenic cortex and in cores ofimmature glomeruli (Figure 9A). It is probable that Tie-2

immunohistochemistry was more sensitive thanin situ hybrid-ization because the latter method did not detect Tie-2 in prim-itive glomeruli (Figure 3E). Tie-2-positive, spindle-shapedcells were detected in the E18 fetal medulla. Some appeared ina loose network and may have represented developing capillaryplexi (Figure 9B). In the P1 outer medulla, Tie-2 immunostain-ing was detected in capillaries aligned alongside tubules (datanot shown). In the P3 cortex, Tie-2 protein was detected in thecores of glomerular tufts, where capillaries and mesangial cellsare located, and in capillaries between tubules (Figure 9C),with no significant staining when the antibody was preab-sorbed with immunising peptide (Figure 9D). At P3, Tie-2immunostaining was prominent in capillaries of the vasa rectaein the outer medulla as shown in longitudinal section (Figure9E).

of Henle growing toward the inner medulla flanked by collecting ducts. The diameter of these tubules was relatively similar. Thick ascendinglimbs of loops of Henle immunostained for Tamm–Horsfall glycoprotein, as did terminal portions of descending limbs (arrowhead in Panel B).(C and D) CD34-immunostained capillaries (arrows) were mostly aligned in parallel with structures. (E) Ang-1 mRNA was expressed by mosttubules traversing this zone, whereas Ang-2 transcripts appeared restricted to a subset of structures (arrows) depicted in Panel F. (G) Tie-2mRNA was located in capillaries (arrows) parallel to tubules. (H) Sense control for Tie-2 showed no significant signal. Ang-1 and Ang-2 senseprobes also showed no significant signal (not shown). Bars, 18mm in all frames.

Figure 5.Kidney cortex at P1. Panel A was counterstained with hematoxylin and periodic acid-Schiff (PAS), whereas Panels B through D werenot counterstained. (A) PAS-positive proximal tubules (p) are prominent in the P1 cortex. A vessel is indicated (v). (B) Ang-1 antisense probedetected signal over proximal tubules. (C) No significant signal over proximal tubules with the Ang-2 antisense probe. (D) The Tie-2 antisenseprobe detected signal in capillaries (arrows) in the interstitium around proximal tubules and in the endothelium of a nearby vessel. Bars, 12mm in all frames.


Figure 6.Outer medulla at P1. Panels A, B, D, F, and H are longitudinal sections, whereas Panels C, E, and G are cross sections. Panels Aand B were counterstained with hematoxylin, whereas Panels C through H, thein situ hybridization studies, were not counterstained. (A) Noteloops of Henle, often with their “U” turns in this region. Ascending limbs immunostained for Tamm–Horsfall glycoprotein, as did the terminalportion of the thinner, descending limbs (arrowheads). (B) CD34-immunostained (brown) capillaries aligned parallel and at right angles totubules. (C and D) A weak signal for Ang-1 mRNA was detected in diverse tubules in this region. (E and F) An intense signal for Ang-2 waslocalized to the thinnest of these tubules (arrowheads), most likely to represent thin limbs of loops of Henle or maturing vasa rectae. (G andH) Tie-2 transcripts were expressed in capillaries (arrows), which were aligned in parallel to the tubules. Bars, 36mm in all frames.


DiscussionBefore this report, only fragmentary information was avail-

able regarding Tie-2 expression in renal ontogeny, and no

information was available for the angiopoietins. In 1994, Lan-delset al. (40) presented preliminary data reporting the detec-tion of Tie-2 transcripts after amplification and sequencing of

Figure 7.Kidney cortex at P3. Panel C was counterstained with eosin, and Panel E was counterstained with hematoxylin, whereas Panels A,B, D, and F were not counterstained. (A) Ang-1 transcripts were detected in glomeruli and proximal tubules. Although all cells of theglomerulus appeared positive, thein situ hybridization signal was especially strong in the outermost cells of the tuft (arrowheads), which mostlikely were podocytes. (B) No significant signal was generated with the Ang-1 sense probe. (C) Tie-2 transcripts were detected in cells (arrows)within glomerular tufts in this section, which was counterstained with eosin. (D) No significant signal was detected with the Tie-2 sense probe.(E) CD34 immunostaining of capillary endothelia in glomerular tufts. (F) Tie-2 mRNA restricted to the endothelium (arrows) of a renal corticalartery. Bars, 12mm in all frames.


Figure 8.Outer medulla at P3. Panels A, C, E, and G are cross sections, whereas Panels B, D, F, and H are longitudinal sections. Panels Aand B have been counterstained with hematoxylin, and Panels E, G, and H were counterstained with eosin, whereas Panels C, D, and F werenot counterstained. (A) In the outer medulla, the mature arrangement of the vasa recta was evident at P3. Tamm–Horsfall-immunostained(brown) ascending limbs of loops of Henle (h) surrounding the vasa recta bundle (vb). Collecting ducts (c) were noted in the interbundle region,and presumed thin limbs of loops of Henle (arrowhead) were detected on the periphery of the vasa recta bundles. (B) CD34 brown-immunostained capillaries in the vasa recta bundle and in the interbundle plexus. (C and D) Ang-1 antisense probe shows a very weak signalin the vascular bundle in cells (arrows) close to capillaries. (E and F) Ang-2 antisense probe produced an intense signal in vasa rectae both inpresumed thin limbs of loops of Henle on the periphery of the bundles (arrowheads in E) and in cells (arrows in E) closely associated with thevasa rectae capillaries themselves. (G and H) Tie-2 transcripts were detected in cells (arrows) in close association with capillaries of the vasarecta. Bars, 12mm in all frames.


mouse E12 metanephric cDNA using degenerate primers totyrosine kinase domains. Using Western blotting, Keeet al.(41) detected Tie-2 protein in perinatal rat kidney with levelsvariably downregulated in the adults, while Wonget al. (42)immunolocalized Tie-2 to postnatal rat glomeruli. Our currentdata provide a more comprehensive picture of Tie-2 expressionin murine renal maturation and constitute the first report ofangiopoietins in renal ontogeny. On the basis of these results,we hypothesize that the angiopoietins and Tie-2 may have rolesin differentiation of glomeruli and vasa rectae.

By examining 10 time points from E14 to P8 using Northernblotting, we found that Ang-1, Ang-2, and Tie-2 were ex-pressed during mouse kidney maturation and, as assessed byquantitative slot blotting, maximal levels of these mRNA oc-curred at P2 to P3. We noted that the relative increase in Ang-1and Tie-2 levels between E14 and P2/P3 was approximatelytwofold, whereas Ang-2 increased approximately 25-fold. Thistemporal pattern is consistent with the concept that Ang-2 actsas a “brake” to fine-tune Tie-2 signaling, as outlined in theintroductory remarks. Using Western blotting, we demon-strated that Tie-2 protein was expressed through the studyperiod, and when we studied renal Tie-2 phosphorylation wefound that a definite signal could be detected from E18 on-ward.

However, speculation about angiopoietin/Tie-2 signalingbased on analyses of RNA and protein derived from wholeorgans may not be extrapolated to specific patterns of geneexpression or receptor tyrosine kinase phosphorylation, whichmight occur in different regions within the organ. Therefore,we investigated thein situ expression of Ang-1, Ang-2, andTie-2 during renal maturation by seeking gene transcripts and,additionally, Tie-2 protein at E18, P1, and P3. We founddifferent tissue expression patterns for Ang-1, Ang-2, andTie-2. Here, we summarize and discuss gene expression in thecortex and medulla separately.

During cortical development, Ang-1 mRNA first localizedto condensing renal mesenchyme and thereafter, at P1–3, tomaturing proximal tubules and glomeruli where expressionwas prominent in the outer cells of the tuft,i.e., presumptivepodocytes. Of note, VEGF is also expressed by podocytes (43),

where it may maintain fenestrations in adjacent capillaries. Anin situ hybridization signal for Ang-2 could not be detected inepithelia of the renal cortex at any stage examined. Transcriptsfor Tie-2 were initially detected in interstitial areas betweennascent epithelia of the nephrogenic cortex in locations whereCD34 immunostaining was faint or absent. These observationsare consistent with a hypothesis that Tie-2 is expressed earlierin the endothelial lineage than CD34. As the cortex matured,Tie-2 mRNA was detected inside glomerular tufts and inlocations expected of interstitial capillaries. The patterns ofTie-2 transcripts as assessed byin situ hybridization weresimilar to those of Tie-2 protein, as assessed by immunohisto-chemistry. However, the latter method may be more sensitivebecause it detected Tie-2 in the cores of primitive glomeruli atE18, where in situ hybridization was negative. The exactglomerular cells that express Tie-2 will require clarification byimmunoelectron microscopy, but we suggest that they mayinclude both endothelia and mesangial cells. In this respect, itis notable that mesangial cellsin vitro have been found toexpress certain RTK in common with endothelia includingVEGFR-1 (44) and the hepatocyte growth factor receptor Met(45).

All tubules that traversed the E18 outer medulla expressedAng-1 transcripts, and a less intense signal was detected intubules in this region at P1. By P3, only a very faint signal forAng-1 could be detected in the vasa recta bundle. At E18,Ang-2 mRNA was expressed in a subset of outer medullarystructures of uncertain designation. However, by P1, Ang-2expression was clearly localized to the thinnest of structuresthat traversed this region, strongly suggesting that they weredescending loops of Henle or maturing vasa rectae. The im-pression was strengthened by the pattern at P3 when Ang-2was expressed by thin tubules at the periphery of vasa rectabundles, where descending limbs reside in rodents (1,2). Atthis stage, Ang-2 was also expressed by cells within the bundle,which were closely associated with the vasa recta capillaries.We consider that these cells were most likely endothelia orsupporting pericytes, although the resolution of the currenttechnique precludes definitive cellular assignment. As assessedby in situ hybridization and immunohistochemistry, Tie-2-

Table 1. Locations ofin situ hybridization signals for Ang-1, Ang-2, and Tie-2 at E18, P1, and P3a

Time Ang-1 Ang-2 Tie-2

E18cortex Condensing mesenchyme Not detected Interstitial capillariesmedulla Outer medulla tubules Outer medulla tubules Interstitial capillaries

P1cortex Proximal tubules Not detected Interstitial capillaries, arterial endotheliummedulla Outer medulla tubules Outer medulla tubules Interstitial capillaries

P3cortex Glomeruli, proximal tubules Not detected Glomeruli, arterial endotheliummedulla Vasa recta bundle* Vasa recta bundle Vasa recta bundle

a The asterisk indicates a very weak signal. Ang-1, angiopoietin-1; Tie-2, tyrosine kinase that contains immunoglobulin-like loops andepidermal growth factor similar domains; E18, embryonic day 18; P1, postnatal week 1.


expressing cells were noted in the outer medulla. At E18 andP1, these cells were located in capillaries aligned parallel to thetubules. By P3, Tie-2 transcripts had become localized to thevasa recta capillaries. Of note, Tie-2 expression appeared ab-sent from the interbundle region despite the presence of aCD34-expressing capillary plexus.

Taken together, our descriptive data are consistent with thehypothesis that Ang-1 may have multiple roles in the matura-tion of the renal microcirculations. We speculate that renalmesenchymal cell-derived growth factor may be chemotacticfor sprouting endothelia in the nephrogenic zone. Second, theexpression of Ang-1 by proximal tubules may by important for

the maturation of cortical interstitial capillaries that expressTie-2. Furthermore, since Ang-1 and Tie-2 are expressed inglomeruli, the factor may be important in the formation and/ormaintenance of glomerular capillaries. Sincein situ hybridiza-tion did not detect Ang-2 in forming glomeruli at any stageexamined, a role cannot be ascribed to this factor in glomeru-logenesis. It is possible, however, that more sensitive tech-niques might be required to detect Ang-2 transcripts in glo-meruli or other locations in the cortex. Our results are alsoconsistent with a theory that Tie-2 signaling is important forthe development of the vasa recta capillaries. In particular, itwas striking that the intensity of outer medullain situ hybrid-

Figure 9. Immunohistochemistry for Tie-2. All sections were counterstained with hematoxylin. Note that in some sections, renal tubulesgenerate a weak, nonspecific, background color. (A) At E18, linear Tie-2 immunostaining (brown) was noted in an interstitial distribution inthe nephrogenic cortex (arrows) and in the core of immature glomeruli (g). (B) Networks (arrow and x) of Tie-2-positive, spindle-shaped cellswere detected in the E18 fetal medulla. These may represent developing capillary plexi. (C) In the P3 cortex, Tie-2 protein was detected in thecores of glomerular tufts (g) and in capillaries (arrows) between tubules, with no specific staining pattern evident when the Tie-2 antibody waspreabsorbed with immunizing peptide (Panel D). (E) Tie-2 immunostaining was prominent in capillaries (arrows) of the vasa rectae in the P3outer medulla as shown in longitudinal section. Bars, 12mm in all frames.


ization signal for Ang-1 fell between E18 to P3, whereas thesignal for Ang-2 was strong at P3. Therefore, if Ang-2 acts asan antagonist of Ang-1 as outlined in the introductory remarks,we speculate that a changing balance of Ang-1 and Ang-2 mayinitially stimulate, and then terminate, Tie-2 signaling to mod-ulate construction of the vasa recta microcirculation.

Although our results lead to speculation regarding the rolesof the angiopoietins and Tie-2 in differentiation of the renalvasculature, definitive proof of their roles awaits functionaldata. Unfortunately, the relatively early embryonic death ofAng-1 and Tie-2 null mutant mice precludes analyses of thepotential roles of these genes during metanephrogenesis. In thefuture, an analysis of the kidneys of chimeric mice constitutedfrom null and wild-type cells, as has been described for Tie-1(46), may be useful. In addition, further work is clearly neededto fully document the potential expression of the angiopoietinsand Tie-2 in the early nephrogenic period, between E11 andE14.

Our current study further indicates that Tie-1 mRNA isexpressed from E14, with levels peaking at P2 and expressiondownregulated thereafter. This is a pattern similar to Tie-2.Examination of Tie-1lcz/Tie-1lczn2 chimeras have suggested arole for Tie-1 in metanephric vessel formation since mutantcells did not significantly contribute to renal vasculature (46).Loughnaet al. (15) detected Tie-1 mRNA in renal mesen-chyme from E10.5 to E15, and the fate of Tie-1-expressingcells was studied by transplantation of E11 Tie-1/LacZ organsinto the neonatal nephrogenic cortex of wild-type mice (15), asite that facilitates growth of precursors into filtering glomeruli(47). Transgene-expressing glomerular and interstitial capillar-ies developed in these transplants, demonstrating that at leastpart of the renal vasculature can formin situ (15). Furthermore,in organ culture of E13 Tie-1/LacZ mouse metanephroi,Loughnaet al. found that reporter gene expression was up-regulated in hypoxia (48). Hypoxia has also been reported toupregulate VEGFR-1 and VEGFR-2 in murine metanephricorgan culture (20,48), and it would be interesting to study theregulation of Tie-2 expression in this milieu. In this respect, itis known that endothelial Per-AHR-ARNT-Sim domain pro-tein 1, an endothelial-specific transcription factor activated byhypoxia, can upregulate Tie-2 (49).

Of note, a human genetic disease has illuminated the func-tion of Tie-2. Vikkulaet al. (50) investigated two families withdominant inheritance of vascular malformations composed of arelative excess of endothelial cellsversuspericytes. They iden-tified a missense mutation of Tie-2 on chromosome 9p causingan arginine to tryptophan substitution in the kinase domain.This increased Tie-2 tyrosine phosphorylationin vitro (50).Human Ang-1 and Ang-2 genes respectively localize to 8q22and 8p23, but mutations have yet to be implicated in disease.Other clues regarding pathogenetic roles of Tie-2 derive fromstudies of healing skin wounds (42), where Tie-2 expressionwas upregulated in neovessels. Capillary microcirculations arealso implicated in kidney disease pathogenesis (7) and, in thefuture, it will be interesting to study the expression patterns ofthe angiopoietin and the Tie genes in these disorders.

AcknowledgmentsThis work was supported by British Heart Foundation Project

Grant 96120 and the Kidney Research Aid Fund.

References1. Tisher CC, Madsen KM: Anatomy of the kidney. In:Brenner

and Rector’s The Kidney, 5th Ed., edited by Brenner BM, Phil-adelphia, Saunders, 1996, pp 3–71

2. Dworkin LD, Brenner LD: The renal circulations. In:Brennerand Rector’s The Kidney, 5th Ed., edited by Brenner BM, Phil-adelphia, Saunders, 1996, pp 247–285

3. Woolf AS: Embryology of the kidney. In:Pediatric Nephrology,4th Ed., edited by Barratt TM, Avner A, Harmon W, Baltimore,Lippincott Williams & Wilkins, 1999, pp 1–19

4. Loughna S, Landels E, Woolf AS: Growth factor control ofdeveloping kidney endothelial cells.Exp Nephrol4: 112–118,1996

5. Gilbert T, Lelievre-Pegorier M, Malienou R, Meulemans A,Merlet-Benichou C: Effects of prenatal and postnatal exposure togentamycin on renal differentiation in the rat.Toxicology43:301–313, 1987

6. McCausland JE, Ryan GB, Alcorn D: Angiotensin convertingenzyme inhibition in the postnatal rat results in decreased cellproliferation in the renal outer medulla.Clin Exp PharmacolPhysiol23: 552–554, 1996

7. Takahashi T, Huynh-Do U, Daniel TO: Renal microvascularassembly and repair: Power and promise of molecular definition.Kidney Int53: 826–835, 1998

8. Visvader JE, Fujiwara Y, Orkin SH: Unsuspected role for theT-cell leukemia protein SCL/tal-1 in vascular development.Genes Dev12: 473–479, 1998

9. Brooks PC, Clark RAF, Cheresh DA: Requirement of vascularintegrin avb3 for angiogenesis.Science264: 569–571, 1994

10. Mustonen T, Alitalo K: Endothelial receptor tyrosine kinasesinvolved in angiogenesis.J Cell Biol 129: 895–898, 1995

11. Risau W: Mechanisms of angiogenesis.Nature 387: 671–674,1997

12. Dumont DJ, Fong G-H, Puri MC, Gradwohl G, Alitalo K, Bre-itman ML: Vascularisation of the mouse embryo: A study offlk-1, tek, tie, and vascular endothelial growth factor expressionduring development.Dev Dyn203: 80–92, 1995

13. Shalaby F, Rossant J, Yamaguchi TP, Gerstenstein M, Wu X-F,Breitman ML, Schuh AC: Failure of blood island formation andvasculogenesis in Flk-1-deficient mice.Nature376: 62–66, 1995

14. Fong G-H, Rossant J, Gerstenstein M, Breitman ML: Role of theflt-1 receptor tyrosine kinase in regulating the assembly of vas-cular endothelium.Nature376: 66–74, 1995

15. Loughna S, Hardman P, Landels E, Jussila L, Alitalo K, WoolfAS: A molecular and genetic analysis of renal glomerular cap-illary development.Angiogenesis1: 84–101, 1997

16. Breier G, Albrecht U, Sterrer S, Risau W: Expression ofvascular endothelial growth factor during embryonic angio-genesis and endothelial cell differentiation.Development114:521–532, 1992

17. Robert B, St. John PL, Hyink DP, Abrahamson DR: Evidencethat embryonic kidney cells expressing flk-1 are intrinsic, vas-culogenic angioblasts.Am J Physiol271: F744–F753, 1996

18. Simon M, Grone H-J, Johren O, Kullmer J, Plate KH, Risau W,Fuchs E: Expression of vascular endothelial growth factor and itsreceptors in human renal ontogenesis and in adult kidney.Am JPhysiol268: F240–F250, 1995


19. Kiamoto Y, Tokunaga H, Tomita K: Vascular endothelial growthfactor is an essential molecule for mouse kidney development:Glomerulogenesis and nephrogenesis.J Clin Invest99: 2351–2357, 1997

20. Tufro-McReddie A, Norwood VF, Aylor KW, Botkin SJ, CareyRM, Gomez RA: Oxygen regulates vascular endothelial growthfactor-mediated vasculogenesis and tubulogenesis.Dev Biol183:139–149, 1997

21. Sato TN, Qin Y, Kozak CA, Andus KL: Tie-1 and tie-2 defineanother class of putative receptor tyrosine kinase genes ex-pressed in early embryonic vascular system.Proc Natl Acad SciUSA90: 9355–9358, 1993

22. Maisonpierre PC, Goldfarb M, Yancopoulos GD, Gao GX: Dis-tinct rat genes with related profiles of expression define a Tiereceptor tyrosine kinase family.Oncogene8: 1631–1637, 1993

23. Korhonen J, Polvi A, Partanen J, Alitalo K: The mousetiereceptor tyrosine kinase gene: Expression during embryonic an-giogenesis.Oncogene9: 395–403, 1994

24. Dumont DJ, Gradwohl GJ, Fong G-H, Auerbach R, BreitmanML: The endothelial-specific receptor tyrosine kinase, tek, is amember of a new subfamily of receptors.Oncogene8: 1293–1301, 1993

25. Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, FujiwaraY, Gendron-Maguire M, Gridley T, Wolburg H, Risau W, Qin Y:Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 inblood vessel formation.Nature376: 70–74, 1995

26. Puri MC, Rossant J, Alitalo K, Bernstein A, Partanen J: Thereceptor tyrosine kinase TIE is required for integrity and survivalof vascular endothelial cells.EMBO J14: 5884–5891, 1995

27. Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, JainV, Ryan TE, Bruno J, Radjiejewski C, Maisonpierre PC, Yan-copoulos GD: Isolation of angiopoietin-1, a ligand for the TIE2receptor, by secretion trap expression cloning.Cell 87: 1161–1170, 1996

28. Koblizek TI, Weiss C, Yancopoulos GD, Deutsch U, Risau W:Angiopoietin-1 induces sprouting angiogenesis in vitro.CurrBiol 8: 529–532, 1998

29. Suri C, Jones PF, Patan S, Bartunkova S, Maisonpierre PC, DavisS, Sato TN, Yancopoulos GD: Requisite role of Angiopoietin-1,a ligand for the TIE2 receptor, during embryonic angiogenesis.Cell 87: 1171–1180, 1996

30. Dumont DJ, Gradwohl G, Fonh G-H, Puri MC, Gerstenstein M,Auerbach A, Breitman ML: Dominant-negative and targetednull-mutations in the endothelial receptor tyrosine kinase, tek,reveal a critical role in vasculogenesis of the embryo.Genes Dev8: 1897–1909, 1994

31. Folkman J, D’Amore PA: Blood vessel formation: What is itsmolecular basis?Cell 87: 1153–1155, 1996

32. Maisonpierre PC, Suri C, Jones PF, Bartunkova S, Wiegand SJ,Radziejewski C, Compton D, McClain J, Aldrich TH, Papado-poulos N, Daly TJ, Davis S, Sato TN, Yancopoulos GD: Angio-poietin-2, a natural antagonist for Tie2 that disrupts in vivoangiogenesis.Science277: 55–60, 1997

33. Witzenbichler B, Maisonpierre PC, Jones P, Yancopoulos GD,Isner JM: Chemotactic properties of angiopoietin-1 and -2, li-gands for the endothelial-specific receptor tyrosine kinase Tie2.J Biol Chem273: 18514–18521, 1998

34. Mandriota SJ, Pepper MS: Regulation of angiopoietin-2 mRNAlevels in bovine microvascular endothelial cells by cytokines andhypoxia.Circ Res19: 852–859, 1998

35. Barbu V, Dautry F: Northern blot normalization with 28S rRNAoligonucleotide probe [Abstract].Nucleic Acids Res17: 7115,1989

36. Yuan HT, Gowan S, Kelly FJ, Bingle CD: Cloning of guinea pigsurfactant protein A defines a distinct cellular distribution patternwithin the lung.Am J Physiol273: L900–L906, 1997

37. Baumhueter S, Dybdal N, Kyle C, Lasky LA: Global vascularexpression of murine CD34, a sialomucin-like endothelial ligandfor L-selectin.Blood 84: 2554–2565, 1994

38. Hoyer JR, Resnick JS, Michael AF, Vernier RL: Ontogeny ofTamm-Horsfall urinary glycoprotein.Lab Invest30: 757–761,1974

39. Iwama A, Hamaguchi I, Hashiyama M, Murayama Y, YasungaK, Suda T: Molecular cloning and characterisation of mouseTIEandTEK receptor tyrosine kinase genes and their expression inhematopoietic stem cells.Biochem Biophys Res Commun195:301–309, 1993

40. Landels E, Proschel C, Noble MD, Jat PS, Fine LG, Woolf AS:Receptor tyrosine kinases expressed in early nephrogenesis [Ab-stract].J Am Soc Nephrol5: 245, 1995

41. Kee N, McTavish AJ, Papilon J, Cybulsky AV: Receptor ty-rosine kinases in perinatal developing rat kidney.Kidney Int52:309–317, 1997

42. Wong AL, Haroon ZA, Werner S, Dewhirst MW, Greenberg CS,Peters KG: Tie2 expression and phosphorylation in angiogenicand quiescent adult tissues.Circ Res81: 567–574, 1997

43. Brown LF, Berse B, Tognazzi K, Manseau EJ, Van de Water L,Senger DR, Dvorak HF, Rosen S: Vascular permeability factormRNA and protein expression in human kidney.Kidney Int42:1457–1461, 1992

44. Takahashi T, Shirsawa T, Miyake K, Yahagi Y, Maruyama N,Kasahara N, Kawamura T, Matsumura O, Mitarai T, Sakai O:Protein tyrosine kinases expressed in glomeruli and culturedglomerular cells: Flt-1 and VEGF expression in renal mesangialcells.Biochem Biophys Res Commun209: 218–226, 1995

45. Kolatsi-Joannou M, Woolf AS, Hardman P, Gordge M, WhiteSJ, Henderson R: The hepatocyte growth factor/scatter factorreceptor,met, transduces a morphogenetic signal in renal glo-merular fibromuscular mesangial cells.J Cell Sci 108: 3703–3714, 1995

46. Partanen J, Puri MC, Schwartz L, Fischer KD, Bernstein A,Rossant J: Cell autonomous functions of the receptor tyrosinekinase TIE in late phases of angiogenic capillary growth andendothelial cell survival during murine development.Develop-ment122: 3013–3021, 1996

47. Woolf AS, Palmer SJ, Snow ML, Fine LG: Creation of a func-tioning chimeric mammalian kidney.Kidney Int 38: 991–997,1990

48. Loughna S, Yuan H-T, Woolf AS: Effects of oxygen on vascularpatterning inTie1/LacZmetanephric kidneys in vitro.BiochemBiophys Res Commun247: 361–366, 1998

49. Tian H, McKnight SL, Russell DW: Endothelial PAS domainprotein 1 (EPAS1), a transcription factor selectively expressed inendothelial cells.Genes Dev11: 72–82, 1997

50. Vikkula M, Boon LM, Carraway III KL, Calvert JT, DiamontiAJ, Goumnerov B, Pasyk KA, Marchuk DA, Warman ML,Cantley LC, Muliken JB, Olsen BR: Vascular dysmorphogenesiscaused by an activating mutation in the receptor tyrosine kinaseTIE2. Cell 87: 1181–1190, 1996


Expression of Angiopoietin-1, Angiopoietin-2, and the Tie-2 Receptor Tyrosine Kinase during

Documents