Kidney development and gene expression in the HIF2α knockout mouse

Authors


Abstract

The hypoxia-inducible transcription factor-2 (HIF2), a heterodimer composed of HIF2α and HIF1β subunits, drives expression of genes essential for vascularization, including vascular endothelial growth factor (VEGF) and VEGF receptor-2 (VEGFR-2, Flk-1). Here, we used a HIF2α/LacZ transgenic mouse to define patterns of HIF2α transcription during kidney development and maturation. Our results from embryonic heterozygotes showed HIF2α/LacZ expression by apparently all renal endothelial cells. At 4 weeks of age, glomerular mesangial and vascular smooth muscle cells were also positive together with endothelial cells. These expression patterns were confirmed by electron microscopy using Bluo-gal as a β-galactosidase substrate. Small numbers of glomerular and tubular epithelial cells were also positive at all stages examined. Light and electron microscopic examination of kidneys from HIF2α null embryos showed no defects in renal vascular development or nephrogenesis. Similarly, the same amounts of Flk-1 protein were seen on Western blots of kidney extracts from homozygous and heterozygous HIF2α mutants. To examine responsiveness of HIF2α null kidneys to hypoxia, embryonic day 13.5 metanephroi were cultured in room air or in mild (5% O2) hypoxia. For both heterozygous and null samples, VEGF mRNA levels doubled when metanephroi were cultured in mild hypoxia. Anterior chamber grafts of embryonic HIF2α knockouts were morphologically indistinguishable from heterozygous grafts. Endothelial markers, platelet endothelial cell adhesion molecule and BsLB4, as well as glomerular epithelial markers, GLEPP1 and WT-1, were all expressed appropriately. Finally, we undertook quantitative real-time polymerase chain reaction of kidneys from HIF2α null embryos and wild-type siblings and found no compensatory up-regulation of HIF1α or -3α. Our results show that, although HIF2α was widely transcribed by kidney endothelium and vascular smooth muscle, knockouts displayed no detectable deficits in vessel development or VEGF or Flk-1 expression. Developmental Dynamics 236:1115–1125, 2007. © 2007 Wiley-Liss, Inc.

INTRODUCTION

One of the key molecules regulating blood vessel development is vascular endothelial growth factor (VEGF; Gerber et al.,1999; Ferrara,2001; Rossant and Howard,2002). VEGF and two VEGF tyrosine kinase receptors, Flk1 and Flt1, are highly expressed during early kidney development (Robert and Abrahamson,2001). In the developing glomerulus, VEGF signaling from immature podocytes is thought to recruit Flk1-expressing angioblasts to the vascular cleft (Robert et al.,1998). Expression of VEGF by podocytes and its receptors by glomerular endothelial cells continues in adult kidney (Simon et al.,1995; Robert et al.,1998), perhaps to maintain the highly specialized endothelium of the capillary loops. Mice with podocyte-specific deletions of VEGF, and those overexpressing VEGF through podocyte-specific promoters, have severe defects in kidney function, leading to early death (Eremina et al.,2003). VEGF is also expressed in developing distal tubules (Kitamoto et al.,1997) and in developing and mature collecting ducts (Simon et al.,1995). Some extraglomerular capillary endothelia also express VEGF receptors during development and in maturity (Robert et al.,1998). Injection of anti-VEGF antibodies or soluble Flt1 causes glomerular endothelial cell detachment and hypertrophy, leading to proteinuria (Sugimoto et al.,2003). Mice lacking VEGF isoforms 164 and 188 show defects in renal arterial branching and glomerular angiogenesis (Mattot et al.,2002), suggesting that VEGF signaling is critical for renal vascular development and maintenance.

Hypoxia, and the hypoxia-inducible transcription factors (HIFs), play key roles in the expression of VEGF and VEGF receptors (Levy et al.,1995; Liu et al.,1995; Gerber et al.,1997; Kappel et al.,1999). HIFs are heterodimeric transcription factors of the basic helix–loop–helix–PAS (bHLH-PAS) superfamily that induce gene expression in response to a wide variety of environmental stimuli, including hypoxia, xenobiotic exposure, and circadian rhythms (Crews,1998; Gu et al.,2000). The transcriptional activity of the alpha/beta heterodimeric HIFs are finely controlled by the stability of the alpha subunits in response to intracellular oxygen concentration. Three such HIFα subunits have been described to date (Wang and Semenza,1995; Flamme et al.,1997; Tian et al.,1997; Gu et al.,1998; Hara et al.,2001), and all have similar biochemical properties (Jiang et al.,1996; O'Rourke et al.,1999; Maxwell and Ratcliffe,2002). In the normoxic state, HIFα proteins are hydroxylated on a conserved proline residue, which promotes an interaction with the VHL ubiquitin ligase complex (Ivan et al.,2001; Jaakkola et al.,2001). HIFα subunits are then polyubiquinated and destroyed by the 26S proteasome. In hypoxia, proline hydroxylation and subsequent proteasomal degradation are avoided, allowing HIFα proteins to translocate to the nucleus, bind the HIFβ subunit, and induce transcription of target genes. An oxygen-dependent hydroxylation of HIFs also regulates gene transactivation properties. In hypoxia, hydroxylation of a conserved HIFα asparagine residue is blocked, which allows HIF to recruit p300/CBP to the transcriptional apparatus (Lando et al.,2002).

What role(s) HIFs play in kidney development has not yet been determined (Haase,2006). In developing kidney, where VEGF signaling is critical for normal glomerular differentiation, we observed HIF1α and -2α mRNA expression in the nephrogenic zone of the outer cortex (Freeburg et al.,2003). Although originally thought to be confined to endothelial cells (Tian et al.,1997), HIF2α expression has also been localized to immature podocytes, suggesting that it may play a role in glomerular capillary development (Freeburg et al.,2003; Bernhardt et al.,2006). In addition, HIF2α protein is stably expressed in avascular embryonic kidney explants, and is up-regulated by hypoxia, which also causes an increase in VEGF mRNA (Freeburg et al.,2003).

Three different HIF2α knockout mice have been generated, and all show mid-gestational lethal phenotypes (Tian et al.,1998; Peng et al.,2000; Compernolle et al.,2002). In the first knockout analyzed, the authors described HIF2α expression in vascular endothelium and also in the embryonic Organ of Zuckerkandl, which is a site of catecholamine biosynthesis. Moreover, they documented severe bradycardia and decreased noradrenaline levels in embryonic day (E) 12.5 null mutants and no viable embryos beyond E15.5. Supplementing pregnant dams with the catecholamine precursor L-threo-3,4-dihydroxyphenylserine (DOPS) allowed ∼40% of knockout mice to survive only until birth (Tian et al.,1998), implicating an essential role for HIF2α in catecholamine homeostasis. Whether deletion of HIF2α affects kidney development specifically has not been studied in detail previously. Here, we closely examined HIF2α expression patterns in developing kidney in vivo after DOPS administration, determined whether an absence of HIF2α affected renal vascular development or expression of candidate HIF target genes, and examined effects of DOPS withdrawal on kidney development in metanephric organ culture and by grafting mutant metanephroi into anterior eye chambers.

RESULTS AND DISCUSSION

Analysis of HIF2α/LacZ Expression in Developing and Mature Kidney

Kidneys from E13.5 HIF2α/LacZ heterozygotes were removed and subjected to whole-mount X-gal histochemistry. As shown in Figure 1A, a prominent network of HIF2α/LacZ-positive cells was observed, which resembled a vascular pattern. Sections of kidneys showed that, in outer cortical, nephrogenic zones, cells occupying the vascular clefts of comma- and S-shaped nephric figures were blue, as were the newly formed capillaries in early capillary loop stage glomeruli (Fig. 1B). Additionally, individual cells scattered in the metanephric mesenchyme not associated with glomerular capillaries also expressed the LacZ transgene (Fig. 1B). After X-gal histochemistry, slides were also labeled with the endothelial-specific lectin BsLB4-fluorescein (Fig. 1C). There was complete overlap of the blue and green signals (Fig. 1D), suggesting that all HIF2α/LacZ-positive cells in nephrogenic zones were endothelial cells or their progenitors. In addition to vascular endothelial cells, some podocytes in immature and maturing glomeruli of newborn kidney also expressed HIF2α/LacZ (Fig. 1E–G). By 4 weeks of age, HIF2α/LacZ expression was widely expressed by most or all renal endothelial cells, including those of glomerular and peritubular capillaries (Fig. 1H). In addition to endothelium, vascular smooth muscle cells in arteries and arterioles in kidney also expressed HIF2α/LacZ (Fig. 1H,I). Rare expression of HIF2α/LacZ by individual tubular epithelial cells was also occasionally observed (Fig. 1J).

Figure 1.

Hypoxia-inducible transcription factor (HIF) 2α expression in kidney. A: Embryonic day (E) 13.5 HIF2α+/− metanephros stained for β-galactosidase histochemistry shows a branching, vessel-like pattern. B: E14 HIF2α−/− kidney showing β-galactosidase reaction product in glomerular endothelial cells migrating into the vascular cleft (VC), and in capillary loop stage glomeruli (G). Also, small arrows denote individual cells in the metanephric mesenchyme containing β-galactosidase. C: The same slide in B was also labeled with BsLB4, an endothelial marker. D: Colocalization of β-galactosidase and BsLB4 shows complete overlap. E: Some visceral epithelial cells (developing podocytes) in early glomeruli express HIF2α (*). F: Intense vascular expression (arrowhead) of HIF2α is seen in newborn kidney, and within glomeruli (G). G: (Higher magnification of boxed area in F): Some podocytes express HIF2α (arrows), whereas others do not (arrowheads). H: Frozen section of a 4-week HIF2α+/− mouse showing intense β-galactosidase product in an endothelial pattern, and in smooth muscle cells of arteries (arrowheads). I: Frozen section of a 4-week HIF2α+/− mouse showing a small artery. Reaction product is seen in both endothelial cells (arrow) and smooth muscle cells (arrowheads). J: Localization of HIF2α/LacZ to peritubular capillary endothelial cell (arrow) and (rarely) to a tubular epithelial cell (arrowhead).

To examine the expression of HIF2α at the ultrastructural level, we processed kidney tissue from heterozygotes with an electron-dense LacZ substrate, Bluo-gal, and carried out electron microscopy. HIF2α expression localized specifically to endothelial cells of capillary loop stage glomeruli at E14.5 (Fig. 2A). At 4 weeks of age, both glomerular endothelial cells as well as mesangial cells expressed HIF2α/LacZ (Fig. 2B).

Figure 2.

High resolution view of two stages of glomerular development showing hypoxia-inducible transcription factor (HIF) 2α expression in endothelial cells and mesangial cells. A: Photomicrograph showing electron-dense Bluo-gal substrate reaction product (arrows) in endothelial cells of an early capillary loop stage HIF2α+/− glomerulus. B: Four-week-old glomerulus showing both mesangial (Me) and endothelial (En) cells contain the Bluo-gal reaction product. Po, podocytes.

Prior work in our lab with a now obsolete antibody showed nuclear HIF2α protein in developing glomerular cells, including podocytes, as well as in the vasculature (Freeburg et al.,2003). In contrast, our findings presented here show that HIF2α/LacZ expression occurred predominantly in endothelial cells and vascular smooth muscle with only rare reporter gene expression by renal epithelial cells. Although LacZ is a convenient marker for promoter activity, it certainly may not accurately reflect stable HIF2α protein expression. Alternatively, the antibody used in our previous study may not have been solely specific for HIF2α. Beyond these caveats, other factors may also explain the different immunolocalization and reporter gene expression patterns. For example, HIF2α mRNA may only be expressed transiently in podocytes, or the HIF2α/LacZ transcript may be highly unstable (or degraded selectively) in podocytes but not in endothelial and vascular smooth muscle cells. In separate studies, others have shown that the adult rat kidney is largely devoid of HIF2α immunolabeling. However, if rats are exposed to hypoxic conditions, abundant HIF2α labeling is seen in glomeruli and in endothelial cells occupying the inner and outer medulla but not papilla (Rosenberger et al.,2002). Therefore, perhaps the high levels of HIF2α/LacZ reporter gene expression we observe abundantly in kidney endothelial cells helps ensure that appropriate amounts of HIF2α protein can be promptly available in the event of hypoxia.

Normal Development of HIF2α-Deficient Kidneys

In heterozygous matings and without DOPS supplementation, we observed an average litter size of 4.9 live pups/litter. Genotype analysis of these newborns revealed a non-Mendelian distribution, with only two knockout pups born (74/116 heterozygous, 63.8%; 40/116 wild-type, 34.8%; and 2/116 knockout, 1.7%). When pregnant dams were supplemented daily with DOPS beginning at E0.5 and embryos were allowed to develop to E13.5, the number of embryos averaged 7.6, and a Mendelian distribution of genotypes was observed at this gestational age (39/76 heterozygous, 51.3%; 19/76 wild-type, 25.0%; 18/76 knockout, 23.7%). At later gestational stages (E17.5–E18.5), and despite the continued, daily administration of DOPS, increased losses occurred in null embryos, with only ∼50% of the expected ratio obtained. This lethality is consistent with prior findings with this knockout mouse (Tian et al.,1998).

HIF2α knockout embryos were grossly indistinguishable from heterozygous and wild-type counterparts and none were edematous. Light microscopic examination of kidneys from all 11 HIF2α null embryos examined (2 at E12, 2 at E13, 1 at E14, 4 at E16, 1 at E17, 1 at E18) showed no obvious defects in either renal vascular development or in nephrogenesis. In early nephric figures, condensing mesenchyme was observed at the tips of branching ureteric buds (Fig. 3A). Comma- and S-shaped nephrons were clearly visible and appeared morphologically normal, with endothelial cells migrating into the vascular cleft in the usual pattern (Fig. 3B, arrow). Red blood cells were seen within early capillary loop stage glomeruli (Fig. 3C, arrow) as well as in more mature loops of glomeruli at E18.5 (Fig. 3D, arrow), indicating glomerular capillary continuity with the systemic vasculature. Similarly, in frozen sections developed for LacZ activity, no obvious defects could be seen in renal vascular patterning: vascular clefts contained LacZ-positive endothelial cells (Fig. 3E, arrow), and capillary loop stage and maturing stage glomeruli contained endothelial cells in the usual pattern (Fig. 3F,G). Tubule morphology also appeared entirely normal, and there were no tubule dilatations or proteinaceous casts.

Figure 3.

Normal glomerular capillary development in hypoxia-inducible transcription factor (HIF) 2α−/− mice. AD: Toluidine blue staining of thin sections from an embryonic day (E) 18.5 HIF2α−/− kidney. A: Ureteric bud (UB) and condensing mesenchyme (CM) appear normal. B: Endothelial cells are migrating into the vascular cleft of S-shaped bodies (arrow). C,D: Red blood cells are seen in capillary loop stage glomeruli (arrow in C), and open capillary loops with red blood cells are readily apparent in mature glomeruli (arrow in D). EG: Frozen sections from E14 HIF2α−/− mutants show β-galactosidase product in endothelial cells migrating into the vascular cleft (arrow) of comma-shaped nephric figures (E). F: Capillary loop stage glomeruli contain β-galactosidase-positive endothelial cells. Afferent and efferent arterioles seen entering the glomerulus also contain reaction product (arrowhead). G: The arrowhead depicts the open capillary loop within a mature glomerulus. H: Ultrastructure of E18.5 HIF2α−/− capillary loop shows normal podocyte (Po) foot process interdigitation and slit diaphragms, glomerular basement membrane, and fenestrated endothelium (arrow).

The findings seen in the light microscope were supported by electron microscopy of HIF2α null kidneys. In maturing glomeruli at E18, capillary loops contained erythrocytes and showed fenestrated endothelium, morphologically normal basement membrane, and fully interdigitated podocyte foot processes with epithelial slit diaphragms spanning the filtration slits (Fig. 3H), all of which were indistinguishable from wild-type glomeruli.

Amounts and Distribution of Flk-1 Protein Appear Unaltered in HIF2α Knockouts

Because the knockout had no prominent kidney phenotype, we next considered possible downstream targets of HIF2α transcriptional activity. The receptor tyrosine kinase Flk-1 is a likely candidate given its renal expression restricted to endothelial cells and the ability of HIF2α to activate the Flk-1 promoter (Elvert et al.,2003). Whole kidney lysates from individual E13.5 knockout or wild-type kidneys underwent sodium dodecyl sulfate (SDS) -polyacrylamide gel electrophoresis and then Western blotting to detect Flk-1 protein. A high molecular weight band representing Flk-1, at ∼200 kDa, was seen in both the wild-type and knockout samples (Fig. 4A, upper panel). To verify equal loading, blots were stripped and re-probed with the ubiquitously expressed HIF1β subunit (Fig. 4A, lower panel). Relative Flk-1 levels were measured by densitometry on blots from three knockouts versus three wild-types, representing animals from two separate E13.5 litters. As shown in Figure 4B, no significant differences in levels of Flk-1 protein expression were seen.

Figure 4.

Flk-1 and GLEPP1 expression appear normal in hypoxia-inducible transcription factor (HIF) 2α−/− mouse kidney. A: Western blot of HIF2α wild-type (+/+) or knockout (−/−) kidney lysates at embryonic day (E) 13.5 was probed with anti–Flk-1, then stripped and re-probed with anti-HIF1β. B: Quantification of Flk-1 from two separate litters of HIF2α mice showing no difference in the levels of Flk-1 in whole kidney lysates (n = 3 for each genotype). C: Double labeling with anti–Flk-1 (red) and anti-GLEPP1 (green) in a HIF2α+/− mouse shows podocyte expression of GLEPP1 and endothelial labeling of Flk-1 as expected. D: Both GLEPP1 and Flk-1 expression appear normal in the HIF2α−/− glomeruli.

We also evaluated the tissue distribution of Flk-1 in HIF2α heterozygotes and knockouts by immunofluorescence microscopy. Frozen sections were labeled with antibodies to Flk-1 (Fig. 4C,D, red) and the podocyte-specific gene GLEPP1 (Fig. 4C,D, green). Glomeruli from heterozygous and knockout mice both expressed Flk-1 and GLEPP1 in the endothelial and podocyte compartments, respectively, and in the same distribution patterns and apparent intensities.

VEGF mRNA Up-Regulation in HIF2α Knockout Metanephric Organ Cultures

To evaluate responsiveness of HIF2α null kidneys to hypoxia, metanephroi were dissected from E13.5 knockouts and heterozygotes (obtained from dams supplemented daily with DOPS) and maintained in organ culture at room air (∼20% oxygen) and mild hypoxia (5% oxygen) without DOPS for 5 days. Total RNAs were isolated and VEGF mRNA levels quantified using real-time polymerase chain reaction (PCR), as described in the Experimental Procedures section. As observed previously for mRNA obtained from wild-type mice (Freeburg et al.,2003), VEGF mRNA levels from both heterozygous and HIF2α knockouts were significantly greater in kidneys cultured under hypoxic conditions (Fig. 5), representing an approximately twofold increase over that seen when kidneys were cultured in room air.

Figure 5.

Vascular endothelial growth factor (VEGF) mRNA increases under the low oxygen condition in cultured embryonic hypoxia-inducible transcription factor (HIF) 2α−/− kidneys. Relative VEGF mRNA levels were measured from embryonic day (E) 13.5 kidneys cultured for 5 days in either room air (20% O2) or mild hypoxia (5% O2) from HIF2α+/− (dark grey) or HIF2α −/− (light grey) mice (n = 3 for each genotype). Both genotypes show a twofold increase in the abundance of VEGF mRNA in mild hypoxia.

Anterior Chamber Grafts of HIF2α Knockout Metanephroi Develop Normally

Because microvessel formation and glomerular endothelial differentiation fail to occur under typical metanephric organ culture conditions, we transplanted HIF2α knockout and heterozygous metanephroi into anterior eye chambers of adult hosts, a site that normally supports the growth and development of endothelial cells from graft-derived progenitors (Robert et al.,1998). Hosts were not supplemented with DOPS. Five days after transplantation, grafts from knockouts were morphologically indistinguishable from heterozygotes and robust HIF2α/LacZ was expressed in a vascular pattern (Fig. 6A,B). As assessed by immunolabeling for the α3α4α5 chains of type IV collagen, the endothelial-specific markers platelet endothelial cell adhesion molecule (PECAM) and BsLB4, and the podocyte differentiation markers GLEPP-1 and WT1, every aspect of glomerular development appeared normal in grafts of HIF2α knockout kidneys (Fig. 6C–O).

Figure 6.

Hypoxia-inducible transcription factor (HIF) 2α−/− kidney development in anterior chamber grafts proceeds normally. A: Frozen section of a grafted HIF2α −/− E13.5 kidney shows widespread vasculature development following β-galactosidase histochemistry. B,C: Serial frozen sections from a HIF2α −/− graft were either reacted for β-galactosidase (B) or labeled (C) with the podocyte marker anti–WT-1 (green) and the endothelial marker anti- platelet endothelial cell adhesion molecule (PECAM, red). Both labels show normal distribution patterns. DI: Double labeling of frozen sections with antibodies to the basement membrane-specific type IV collagen (ColIV(α3, α4, α5), green) and the podocyte marker GLEPP1 (red). JO: Double labeling of frozen sections with antibodies to the endothelial marker BsLB4 (green) and the podocyte marker WT-1 (red). The same expression pattern is seen for both heterozygous HIF2α grafts (D–F and J–L) and knockout HIF2α grafts (G–I and M–O). Merged images are shown in F, I, L, and O.

Finally, in two rare instances, we obtained female HIF2α knockout mice that survived birth. One died at 3 weeks of age and was less than half the size of wild-type and heterozygous littermates (4.26 g vs. 11.0 g), probably reflecting a catecholamine deficit. However, light and electron microscopic analyses of kidneys from this mouse showed entirely normal tissue architecture (not shown). The second mouse lived to 7 weeks of age and was also smaller than heterozygous littermates (10.9 g vs. 19.2 g). Urinalysis conducted at 6 weeks of age showed a 24-hr albumin excretion rate of 19 μg, which was somewhat greater than a heterozygous sibling (11 μg), but within normal values for control, nondiabetic C57Bl/6 and 129Sv mice (Gurley et al.,2006).

Taken together, our findings indicate that there may be no developmental role for HIF2α in formation of the renal vasculature and glomeruli. Although DOPS administration is efficacious to overcome deficits in catecholamine synthesis, the supplement did not rescue every null mutant, and, as reported earlier, less than half the expected number of knockout mice are born despite DOPS supplementation (Tian et al.,1998). This finding suggests that HIF2α may be required for transcriptional regulation in different cell types. On a hybrid background, a small proportion of HIF2α knockout mice are viable without DOPS supplementation, but succumb to a multiorgan pathology stemming from dysregulation of oxidative stress pathways in mitochondria (Scortegagna et al.,2003a). These mice also had decreases in all blood lineages, and transplantation of wild-type bone marrow into irradiated HIF2α knockout hosts revealed a deficit in the bone marrow microenvironment or a systemic effect on hematopoiesis (Scortegagna et al.,2003b). In another HIF2α knockout model, there was decreased VEGF production by alveolar type II pneumocytes, leading to impaired lung development and neonatal death (Compernolle et al.,2002).

The apparently normal vascular development observed in HIF2α knockout kidney raised the likelihood of compensation by other HIFα family members. Indeed, HIF1α and -2α share 48% sequence identity and both bind and activate transcription from the hypoxia response element (HRE) found in the VEGF and erythropoietin promoters (Ema et al.,1997; Tian et al.,1997). To test for possible up-regulation of HIF1α or -3α in the HIF2α knockouts, we carried out quantitative, real-time PCR on RNAs obtained from E13.5 kidneys (from two wild-type embryos and from two knockouts), and from whole embryos (two wild-types, three knockouts). In kidney mRNAs, the wild-type:knockout expression ratios were closely similar for both HIF1α (1:0.9) and HIF3α (1:0.9). Similarly, in HIF2α null embryos, there was no up-regulation of HIF1α (1:0.7) or HIF3α (1:0.8), and no significant changes in expression of Flk1 (1:0.9) or VEGF mRNAs (1:1) were observed. Much evidence suggests nonredundant roles for HIF1α and -2α. For example, glycolytic genes are not induced by hypoxia in a renal cell carcinoma cell line that only expresses HIF2α. However, glycolytic genes are expressed in the RCC-4 line, which expresses both HIF1α and -2α (Hu et al.,2003). HIF1α−/− embryonic stem cells lack induction of all glycolytic enzymes typical to a hypoxic response, suggesting that HIF2α cannot compensate for loss of HIF1α in this scenario (Iyer et al.,1998; Ryan et al.,1998). Despite stable expression, HIF2α is transcriptionally inactive in embryonic mouse fibroblast cell lines (Park et al.,2003). Furthermore, in A293 cell transient transfections, only HIF2α (but not -1α) is capable of activating an Flk-1 promoter construct (Kappel et al.,1999). Thus, even though HIF1α and -2α are biochemically similar, they most likely induce either different genes and/or the same genes in different cell types.

Despite biochemical evidence for HIF2α induction of Flk-1 (Kappel et al.,1999; Elvert et al.,2003), we did not observe decreases in renal Flk-1 levels in HIF2α knockouts, either by quantitative reverse transcriptase (RT) -PCR, Western blotting, or immunofluorescence. Thus, regulation of Flk-1 in developing kidney endothelial cells in vivo must occur independent of HIF2α. Additionally, when E13.5 HIF2α knockout kidneys were cultured in hypoxia, levels of VEGF mRNA increased significantly over that seen in normoxia, and in parallel with that seen in metanephric cultures from heterozygotes. Again, this finding suggests that HIF2α alone may not mediate the hypoxia-inducible response of the VEGF gene, at least under the organ culture conditions used here. We, therefore, conclude that, despite its widespread expression throughout the developing kidney vasculature, deletion of HIF2α does not affect vascular patterning, the expression of VEGF or Flk1, or nephrogenesis. Nevertheless, we cannot exclude a role for HIF2α in renal vascular biology later in life and/or under physiological stress. Future studies with HIF2α mutants may help illuminate these possibilities.

EXPERIMENTAL PROCEDURES

Transgenic Mice

Epas1tm1Rus mice were purchased from the Jackson Laboratory (Bar Harbor, ME), and the colony was established and maintained by intercrossing heterozygotes (HIF2αLacZ/+; Tian et al.,1998). The following three primers were used to identify wild-type HIF2α+/+, heterozygous HIF2αLacZ/+, or homozygous null HIF2αLacZ/LacZ mice by PCR of tail genomic DNA: 5′-GGAGGCTTTGTCCAGGTGGGAGCTCACACTGTG-3′, 5′-GGTGCGGGCCTCTTCGCTATTA-3′, 5′-GTGGGTCACTACCGCGAGTGTGAATGG-3′. To obtain homozygous mutants, the drinking water of pregnant females was supplemented with 3 mg/ml of the catecholamine precursor DOPS (Sigma Chemical, St. Louis, MO) in 2 mg/ml ascorbic acid. The solution was made fresh daily and administered throughout gestation.

β-Galactosidase Histochemistry

Heterozygous or timed-bred pregnant mice were anesthetized with halothane and killed by cervical dislocation. For analysis of metanephroi, kidneys were microdissected from embryos and nonrenal tissues were used for PCR genotyping. Dissected kidneys were fixed overnight in 0.2% paraformaldehyde in 0.1 M PIPES buffer (piperazine-N,N′-bis[2-ethanesulfonic acid]), pH 6.9, at 4°C. Tissues were washed three times in 2 mM MgCl2 in phosphate buffered saline (PBS) and then cryopreserved in 30% sucrose in 2 mM MgCl2 in PBS overnight at 4°C. Kidneys were then placed in optimal cutting temperature compound (OCT; Miles, Elkhart, IN) and frozen in isopentane in an acetone/dry ice bath. Cryostat sections were cut at a thickness of 8 μm, air-dried, and then fixed again in 0.2% paraformaldehyde in 0.1 M PIPES on ice, pH 6.9, for 10 min. Slides were washed three times in PBS with 2 mM MgCl2 and then incubated in detergent rinse (0.1 M phosphate buffer, pH 7.3, 2 mM MgCl2, 0.01% sodium deoxycholate, 0.02% Nonidet P-40) for 10 min on ice. After detergent rinse, slides were placed in color development solution (detergent rinse containing 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 20 mM Tris, pH 7.3, and 1 mg/ml 5-bromo-4- chloro-3-indolyl-β-D-galactopyranoside [X-gal; Sigma Chemical]) overnight at 37°C in the dark. Sections were then washed three times on ice with PBS with 2 mM MgCl2, post-fixed in 4% paraformaldehyde in PBS, dehydrated through ethanol gradients, and permanently mounted in Permount (Fisher Scientific, Pittsburgh, PA). Additionally, some sections were post-fixed in 0.2% paraformaldehyde after color development and labeled with Bandeiraea simplicifolia BS-L lectin (BsLB4; Sigma Chemical) at a 1:50 dilution for 1 hr at room temperature, followed by mounting with Prolong antifade mounting medium (Molecular Probes, Eugene, OR).

Whole-Mount β-Galactosidase Histochemistry

E13.5 kidneys were removed from DOPS-supplemented females and fixed overnight in 0.2% paraformaldehyde in 0.1 M PIPES buffer, pH 6.9, at 4°C. Tissues were washed twice in 2 mM MgCl2 in PBS, followed by detergent rinse for 10 min. Kidneys were placed in color development solution overnight at 37°C in the dark. Kidneys were rinsed twice in 2 mM MgCl2 in PBS, and re-fixed with 4% paraformaldehyde in PBS for 30 min on ice.

Electron Microscopy

Kidney cortices were minced to 1-mm cubes, then fixed for 2 hr on ice with 2% paraformaldehyde/0.4% glutaraldehyde in PBS containing 2 mM MgCl2. Tissue was washed three times in PBS containing 2 mM MgCl2, and incubated at 37°C overnight with Bluo-gal staining solution (20 mM potassium ferricyanide/ferrocyanide, 2 mM MgCl2, 2 mM Bluo-gal [Invitrogen, Carlsbad, CA], pH 7.3). After washing three times with buffer, samples were re-fixed with Karnovsky's fixative for 30 min, washed three times with 0.1 M sodium-cacodylate in 3.5% sucrose, pH 7.3, and then post-fixed for 1.5 hr with Palade's OsO4. Tissues were dehydrated through graded ethanol and propylene oxide and embedded in Polybed 812 with overnight polymerization at 60°C. Ultrathin sections were stained with 4% uranyl acetate and Reynold's lead citrate. Other tissues were analyzed using routine transmission electron microscopic methods as described previously (Abrahamson and St. John,1992).

Western Blotting

Single metanephroi from E13.5 mice were dissected and immediately homogenized in 75 μl of ice-cold RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with a protease inhibitor cocktail (Freeburg et al.,2003). Lysates were incubated on ice for 10 min, followed by two freeze/thaw cycles. Cellular debris was cleared by centrifugation (11,000 × g, 15 min, 4°C). Total protein was quantified using the DC Protein Assay (Bio-Rad, Hercules, CA). Equal amounts of protein were loaded onto precast gradient gels (Bio-Rad), and, after electrophoresis, proteins were transferred to nitrocellulose. Membranes were probed with a 1:100 dilution of rabbit anti–Flk-1 antibody (C-20, Santa Cruz, Santa Cruz, CA) followed by donkey anti-rabbit immunoglobulin G–horseradish peroxidase (IgG-HRP) at 1:1,000 (Amersham). After ECL, membranes were stripped (Western Blot Recycling Kit, Alpha Diagnostic, San Antonio, TX) and re-probed with a 1:100 dilution of goat anti-HIF1β (C-19, Santa Cruz) followed by rabbit anti-goat IgG-HRP at 1:10,000 (Sigma). Relative Flk-1 intensity was calculated from densitometry values obtained using the Bio-Rad ChemiDoc XRS gel documentation system.

Immunofluorescence

Kidneys were fixed in freshly prepared 0.2% paraformaldehyde (in PBS containing 0.1 M PIPES, pH 6.9, 2 mM MgCl2), cryoprotected in 30% sucrose, and frozen in Tissue-Tek OCT embedding compound. Frozen serial sections, 8 μm thick, were postfixed in 0.2% paraformaldehyde, rinsed in PBS, then permeabilized in 0.5% Triton X-100 in PBS for 5 min, followed by an additional rinse with PBS. Double labeling was carried out with a 1:20 dilution of rat anti-Flk-1 (BD Pharmingen, San Diego, CA) and 10 μg/ml rabbit anti-GLEPP1 (a kind gift from Roger Wiggins, University of Michigan [Wang et al.,2000]); a 1:50 dilution of rabbit anti-WT1 (Santa Cruz) and 50 μg/ml rat anti-PECAM (Serotec, Raleigh, NC); or with 10 μg/ml rabbit anti-GLEPP1 and 100 μg/ml mouse anti-collagen type IV (α3,α4,α5 chains; Heidet et al.,2003). The appropriate secondary antibodies (Alexa 488 and 594 conjugates, Molecular Probes) were applied after PBS rinses. In some cases, fluorescein-conjugated BsLB4 lectin (Sigma) was applied during the secondary antibody incubation. Slides were mounted in Prolong Gold plus DAPI (Molecular Probes). Fluorescent digital images were obtained using a Zeiss LSM510 confocal microscope.

Metanephric Kidney Culture

E13.5 kidneys were isolated from embryos taken from DOPS-supplemented HIF2α+/− females and placed into transwell culture with medium (DMEM with 10% fetal bovine serum, 1% Pen-Strep and 1% matrigel [BD Biosciences, Palo Alto, CA]). Genotyping was performed on embryonic tail tissue. Paired kidneys were placed in separate dishes either at 20% O2 (room air) or 5% O2 (mild hypoxia) for 5 days. Medium was exchanged on the second day of culture. Total RNA was isolated from individual kidneys using the RNAqueous-Micro spin kit (Ambion, Austin, TX). Samples were diluted to 150 ng/μl and amplified using QuantiTect SYBR Green RT-PCR kit (Qiagen, Valencia, CA) with the following primers: VEGF forward 5′-GGAGATCCTTCGAGGAGCACTT-3′ and reverse 5′-GGCGATTTAGCAGCAGATATAAGAA-3′; cyclophilin forward 5′-CAGACGCCACTGTCGCTTT-3′ and reverse 5′-TGTCTTTGGAACTTTGTCTGCAA-3′ (Shih et al.,2002). Real-time PCR was performed using an iCycler (Bio-Rad). The primer sets were validated for efficiency for the comparative Ct method using standard curve analysis (Livak and Schmittgen,2001). PCR products were analyzed on agarose gels to confirm size, and melt curve analysis was performed to reveal a single amplified product.

Anterior Chamber Grafts

Adult CD1 hosts were anesthetized with a ketamine–xylazine combination (100 mg and 15 mg, per kg body weight). Tropicamide was applied to one eye to dilate the iris, 0.5% teracaine was supplied as corneal anesthetic, and the cornea was pierced with a 27-guage needle, and the incision was extended ∼2 mm using Vannas scissors. E13.5 kidneys (from DOPS-supplemented embryos obtained from timed-bred HIF2α heterozygous matings) were transferred through the corneal incision, and positioned over the host iris as described before (Robert et al.,1998). Ophthalmic antibiotic ointment was applied to the eye, and hosts were allowed to recover. Seven days later, hosts were anesthetized by halothane and grafted kidneys were harvested. Grafts were fixed in cold 0.2% paraformaldehyde (in PBS containing 0.1 M PIPES, pH 6.9, 2 mM MgCl2), cryoprotected in 30% sucrose, and frozen in Tissue-Tek OCT embedding compound. Serial 8-micron sections were treated with the various antibody combinations for immunofluorescence and for β-galactosidase histochemistry as described above.

Quantitative Real-Time RT-PCR

Total RNA was isolated from individual kidneys or whole embryos using the RNeasy Micro Kit (Qiagen). Samples were diluted to 10 ng/μl and amplified with the following primers: VEGF forward 5′-GGAGATCCTTCGAGGAGCACTT-3′ and reverse 5′- GGCGATTTAGCAGCAGATATAAGAA-3′; cyclophilin forward 5′- CAGACGCCACTGTCGCTTT-3′ and reverse 5′-TGTCTTTGGAACTTTGTCTGCAA-3′ (Shih et al.,2002), Superarray RT2 PCR Primer Sets for mouse HIF1α (PPM03799A), mouse HIF3α (PPM05268A), and Mouse Kdr (Flk1; PPM03057A; SuperArray, Inc., Bethesda, MD). Real time RT-PCR was carried out on an iCycler (Bio-Rad) with a one-step RT-PCR kit (Qiagen, catalog no. 204245), which uses the double-stranded DNA-binding dye SYBR Green. The primer sets were validated for efficacy using the comparative Ct method and standard curve analysis (Livak and Schmittgen,2001). PCR products were analyzed on agarose gels to confirm size, and melt curve analysis was performed to reveal a single amplified product. Normalization was performed by dividing each value calculated for a specific gene by the value of the housekeeping gene (cyclophilin).

Urinalysis

Mice were weighed and maintained in metabolic cages to collect 24-hr urines. Albumin excretion was measured using the Albuwell murine microalbuminuria assay system, following manufacturer's instructions (Exocell, Inc., Philadelphia, PA).

Acknowledgements

This manuscript is dedicated to the memory of our friend and colleague, Dr. Paul B. Freeburg. We thank Dr. Roger Wiggins, University of Michigan, for providing the GLEPP1 antibody. Confocal images were acquired at the KUMC Confocal Imaging Facility, supported by the Kansas IDeA Network of Biomedical Research Excellence (grant no. RR016475), and we thank Dr. Elizabeth Petroske and Eileen Roach for technical assistance.

Ancillary