Islet1 Deletion Causes Kidney Agenesis and Hydroureter Resembling CAKUT

Isl1 Is Transiently Expressed in the Mesenchyme Surrounding the Ureteric Bud Stalk

First, we examined the expression patterns of Isl1 during kidney development by in situ hybridization. At E10.5, Isl1 was expressed in tissues surrounding the cloaca but was excluded from the kidney primordia, comprising the metanephric mesenchyme and nephric duct (Figure 1A). At E11.5, a similar expression pattern was observed, in that Isl1 expression was detected in the ureteric mesenchyme surrounding the ureteric bud stalk and cloaca but was not expressed in the metanephric mesenchyme or ureteric bud epithelia (Figure 1B). Subsequently, Isl1 expression was rapidly downregulated at E12.5 (Figure 1C), and no Isl1 expression was observed in the kidney proper or the ureter at E14.5 (Figure 1D). These observations were confirmed by whole-mount in situ hybridization, showing no Isl1 signals in the kidney or ureter at E14.5 (Figure 1E). Staining with an anti-Isl1 antibody confirmed these observations, including the negative expression in the ureter at E14.5 (Figure 1F). Nuclear Isl1 staining was observed weakly in the lateral and intermediate mesoderm at E9.5 and robustly in the ureteric mesenchyme at E11.5. (Figure 1, G and H).

Next, the Cre recombinase activity of Hoxb6Cre mice was examined. This mouse strain is often used for deletion of genes in lateral mesoderm-derived tissues, such as the limb buds.¹⁰ To examine the Cre activity in the developing kidney, we crossed Hoxb6Cre mice with a reporter strain, in which the CAG promoter, stop sequences flanked by loxP sites, and the tandem dimer Tomato (tdTomato) are inserted into the Rosa26 locus.¹¹ Owing to the presence of the potent CAG promoter, this mouse strain serves as a very sensitive reporter for Cre recombinase activity. When these mice were crossed with Hoxb6Cre mice, the tdTomato signals were detected mainly in the lateral mesoderm and weakly in the intermediate mesoderm at E9.5 (Figure 1I), which is consistent with the reported Cre activity.¹⁰ At E11.5, Cre activity was broadly detected, except for the ureteric bud epithelia (Figure 1J). Although the activity in the metanephric mesenchyme was mosaic, the regions with Cre activity cover all the Isl1-positive populations described above. Thus Hoxb6Cre mice are useful for efficient deletion of Isl1 during kidney development. The non-nuclear signal in the ureteric epithelium in Figure 1H, which is not consistent with the data obtained by in situ hybridization, could result from background staining of the antibody. In addition, the cryptic Isl1 expression in the ureteric epithelium does not affect our analysis because Hobx6Cre does not delete genes in this population.

Isl1 Deletion Causes Kidney Agenesis or Hydroureter: A CAKUT-Like Phenotype

All the Hoxb6Cre;Isl1^flox/flox mice died shortly after birth (n=11), while Isl1^flox/flox, Isl1^flox/+, and Hoxb6Cre;Isl1^flox/+ mice were apparently normal (n=9, n=13, and n=21, respectively). Regarding the 22 presumptive urogenital systems in the 11 Hoxb6Cre;Isl1^flox/flox mutant mice, 11 (50%) showed complete kidney agenesis (Figure 2, A and B), 10 (45.5%) exhibited hydronephrosis and hydroureter (Figure 2C), and one (4.5%) had normal-sized kidneys. At E16.5, 50% of the kidneys examined (three of six) were absent in Hoxb6Cre;Isl1^flox/flox mice, and dilatation of the pelvis and ureter all along the tract was observed in the remaining three kidneys (Figure 2, D and E). The mutant ureter had a thinner wall but retained the smooth muscle layer, which was confirmed by immunostaining for smooth muscle actin (Figure 2, F and G). The mutant smooth muscle layer was also positive for Sm22α, a maturation marker for this ureteric mesenchyme-derived lineage (Figure 2, H–K). Myocardin, another differentiation marker expressed earlier than Sm22α, was also detected in the Isl1 mutants (Figure 2, L and M), indicating that maturation of the ureteral mesenchyme was unaffected in the absence of Isl1. Therefore, the hydroureter is unlikely to result from developmental impairment of the ureteric mesenchyme and could be caused by other mechanisms, such as mechanical obstruction in the lower urinary tract.

The Ureter Is Not Properly Connected to the Urinary Bladder in the Absence of Isl1

We then focused on Isl1 expression in the lower urinary tract. Isl1 was robustly expressed along the nephric duct, at the base of the ureteric stalk, and in the genital tubercle at E11.5 (Figure 3A). This expression was retained at E12.5 when the expression in the kidney proper was significantly reduced (Figure 3B). At E14.5, Isl1 was still expressed, albeit weakly, at the ureterovesical junction (Figure 3C), as well as in the region along the urethra (Figure 3D).

Ink injected into the pelvis filled up the dilated pelvis and ureters, but did not reach the urinary bladder in any of the newborn mutants (n=4), suggesting the existence of urinary tract obstruction (Figure 3, E–G). We also observed duplicated ureters in one mutant, as shown by negative ink flow in the second ureter (Figure 3G). Serial sectioning of the ureter-bladder (ureterovesical) junctions of these samples, as well as those at E14.5, revealed that the ureters failed to reach the urinary bladder and ended blindly in nine of 11 Isl1 mutant mice examined (Figure 3, H and I). In addition, we found impaired formation of the urethra in a subset of the mutants (three of nine), all of which had the abnormal ureterovesical junctions (Figure 3, J and K). Therefore, the hydroureter in the Isl1 mutants is mainly caused by impaired connection of the ureter to the bladder.

However, at E14.5, when no hydronephrosis became apparent because of the negligible urine production, the kidney size ranged from normal to complete agenesis among 10 kidneys examined (two agenesis, four hypoplasia, four normal size; see Figure 3, L and M, for hypoplasia), suggesting that the development of the kidney proper was also impaired independently of the urinary tract obstruction. Because we did not detect any defects in metanephric mesenchymal genes, such as Six2 or Cited1, in the mutant kidneys, we searched for additional abnormalities.

Ectopic or Accessory Ureteric Buds Are Formed in the Absence of Isl1

At E11.5, the ureteric buds became elongated and invaded into the metanephric mesenchyme in the wild-type kidneys (Figure 4A). In contrast, the mutant kidneys showed variable phenotypes at this step. Overall, 50% of the kidneys examined (6 of 12) showed impaired elongation of the ureteric buds (Figure 4B), while the remaining buds reached the metanephric mesenchyme. In the former, apoptosis was observed in the ureteric mesenchyme and the ureteric bud epithelia (Figure 4, C and D). Isl1 was expressed in the ureteric mesenchyme surrounding the base of the ureteric bud stalk and the nephric duct in the wild-type kidneys (Figure 4E; see also Figures 1 and 3A), which was abolished in the mutant kidneys (Figure 4F). However, some mutant kidneys showed residual Isl1 expression, as evaluated by immunostaining, probably owing to the mosaic deletion of Hoxb6Cre as reported previously,⁹ which could partly explain the variable kidney sizes at E14.5. If the kidney is formed to some extent, hydronephrosis and/or hydroureter could be a prominent phenotype owing to the obstruction at the ureterovesical junction.

Tbx18 and Sox9 were expressed in the ureteric mesenchyme at this stage, and their expression levels were unaffected in the mutant kidneys, although the expression domains were slightly narrower (Figure 5, A–D). We also found that these two genes were expressed in the ureteric mesenchyme close (or proximal) to the metanephric mesenchyme, while Isl1 was distributed in the distant (or distal) domains at the base of the ureteric stalks and around the nephric duct, although these two domains overlapped to some extent. We confirmed these observations by whole-mount in situ hybridization (Figure 5, E–H).

Serial sectioning of the mutant ureteric buds suggested that the direction of the buds might be altered (see Figures 4B and 5D). Therefore, we visualized the ureter branching by crossing the Isl1 mutants with Hoxb7GFP mice. At E12.5, while the control ureteric stalks were connected close to the urogenital sinus, the mutant ureteric bud branched poorly and was still connected to the common nephric duct (Figure 4, G and H). The mutant ureter on the other side failed to bud. At E11.5, the ureteric bud elongated from the nephric duct and branched into the metanephric mesenchyme in the control mice (Figure 4I). In the absence of Isl1, however, the ureteric budding occurred ectopically, as shown by a longer common nephric duct, in three kidneys examined (Figure 4J). We also found an accessory ureteric bud in one mutant kidney (Figure 4K). Whole-mount in situ hybridization of Pax2 using additional samples confirmed these observations (Figure 4, L–N). Seven of nine kidneys examined at E11.5 showed cranially positioned and poorly branched ureteric buds (Figure 4M), while two exhibited accessory buds (Figure 4N). These defects (i.e., ectopically positioned or accessory ureteric buds) could lead to misconnection of the ureterovesical junction or duplicated ureters and eventually hydroureter. The subsequent poor branching of the ureteric buds is likely to cause the kidney agenesis/hypoplasia observed in the Isl1 mutants.

Bmp4 Expression Is Reduced in the Absence of Isl1

At E10.5, a stage before the ureteric bud is formed, a bulge was observed in the nephric duct in the presumptive metanephric region of the control mice (Figure 6A), which was not apparent in the Isl1 mutants (Figure 6B). The size of the Pax2-positive metanephric mesenchyme in the Isl1 mutants was similar to that in the control mice (Figure 6, C and D).

Mice heterozygous for Bmp4 showed variable kidney sizes and hydroureter resulting from ectopic or accessory ureteric buds,² similar to the phenotypes of the Isl1 mutant mice. At E10.5, Bmp4 was expressed along the nephric duct and around the metanephric mesenchyme (Figure 6E). Bmp4 expression, especially in the presumptive metanephric region, was significantly reduced in all of the Isl1 mutant kidneys examined (n=4) (Figure 6F). Similarly, we observed Bmp4 reduction in all of the Isl1 mutants examined at E11.5 (n=8), while Bmp4 was expressed along the nephric duct and around the ureteric bud stalk in the control mice (Figure 1, G and H). This reduction is consistent with the similar CAKUT-like phenotypes in both Isl1 and Bmp4 mutant mice.

However, BMP4 supplementation to the organ cultures starting at E10.5 did not significantly rescue the Isl1 mutant phenotypes, which are variable in nature (Supplementary Figure 1). BMP4 addition inhibited ureteric budding and metanephros development even in control kidneys as reported previously,² while lower concentrations had minimal effects on the Isl1 mutants. It is possible that strict spatiotemporal regulation of BMP activity is necessary for proper ureteric budding, and this cannot be mimicked faithfully by exogenous BMP4 addition. Indeed, there are no reports of successful rescue of the kidney phenotypes even in Bmp4 heterozygotes. Although Bmp4-independent mechanisms could also be involved, the phenotypic similarities between the Isl1 and Bmp4 mutants, and Bmp4 reduction in the Isl1 mutants, suggest that Isl1 is essential for urogenital development, at least partly by regulating Bmp4.

Generation of Mutant Mice

Isl1^flox/flox and Isl1Cre mice were described previously.^16,17 Hoxb6Cre and BMP^flox/flox mice were kindly provided by Drs. M. Kuehn (National Cancer Institute, National Institutes of Health, Bethesda, MD).¹⁰ Hoxb7GFP mice, which were generated by Dr. F. Costantini (Columbia University, New York, NY),¹⁸ were obtained from the Jackson Laboratory. All animal experiments were performed in accordance with our institutional guidelines and ethical review committees. Mice carrying the Cre allele were genotyped with forward primer Cre 1 (5′-AGG TTC GTT CAC TCA TGG A-3′) and reverse primer Cre 2 (5′-TCG ACC AGT TTA GTT ACC C-3′). PCR amplifications were performed under identical conditions using GoTaq DNA polymerase (Promega), with denaturation at 95°C for 5 minutes, 35 cycles of 95°C for 30 seconds, 58°C for 60 seconds, and 72°C for 30 seconds, and a final extension at 72°C for 7 minutes. The PCR products were analyzed by electrophoresis in a 1.2% agarose gel and visualized by ethidium bromide staining.

In situ Hybridization and Immunostaining

Histologic examinations were performed as described previously.^19,20 Mice were fixed in 10% formalin, embedded in paraffin, and cut into 6-μm sections. In situ hybridization was performed using an automated Discovery System (Roche) according to the manufacturer’s protocols. The probe for Bmp4 was a kind gift from Dr. N. Ueno (National Institute for Basic Biology, Japan). Templates for other probes were generated by RT-PCR and sequenced. Immunostaining was carried out automatically using a BlueMap Kit (Roche) and the Discovery System. The following primary antibodies were used: anti-Isl1 (Abcam); anti–smooth muscle actin (Abcam); anti-Sm22α (Abcam); anti-pSmad1/5/8 (Cell Signaling Technology); anti-Sox9 (Millipore); anti–cleaved caspase 3 (Cell Signaling Technology); anti-RFP (Rockland); and anti–cytokeratin 8 (Developmental Studies Hybridoma Bank). For whole-mount in situ hybridization, samples were fixed with 4% paraformaldehyde and processed using an automated InsituPro VS (Intavis AG) according to the manufacturer’s protocol.