High incidence of vesico-ureteral reflux in mice with deletion of fibroblast growth factor receptor 2 in kidney mesenchyme

Fgfr2^Mes−/− conditional knockout mice

The Institutional Animal Care and Use Committee at the Research Institute at Nationwide Children's Hospital approved mouse handling in this study. The conditional knockout mice have been described previously.^{11, 12} Briefly, we crossed transgenic mice with the Pax3 promoter driving cre recombinase expression in metanephric mesenchyme (Pax3cre^Tg) (gift from Dr. Jon Epstein, University of Pennsylvania) with mice with lox-p sites flanking a critical portion of Fgfr2 (Fgfr2^Lox) (gift from Dr. David Ornitz, Washington University). Mice hemizygous for Pax3cre^Tg and homozygous for Fgfr2^Lox (Pax3cre^Tg/+Fgfr2^Lox/Lox) have deletion of Fgfr2 in metanephric mesenchyme (Fgfr2^Mes−/−). Mice with other alleles are phenotypically normal and used as controls.

Genotyping

As described, we extracted DNA from tail clippings or embryonic tissues for PCR genotyping.^11,12 Genotyping primers are included in Table 1.

Common Nephric Duct (CND) Lengths and Apoptosis assays

E11.5 mutant and littermate control embryos were assayed for whole mount in situ hybridization as described utilizing digoxigenin-UTP-labeled RNA probes against Ret.¹²Ret labeled embryos were positioned in identical orientation and images were captured via an AxioCam digital camera (Zeiss, Thornwood, NY) mounted on a DM LB microscope (Leica, Wetzlar, Germany). CND length (caudal end of the nephric duct at the cloaca to the cranial edge of the ureteric bud stalk origin) was measured using ImageJ software (Wayne Rasband, NIH, USA).

To examine apoptosis in CNDs, we methanol fixed E11.5 mutant (n=10) and control littermate (n=10) embryos, followed by whole mount immunofluorescence with antibodies against pancytokeratins (Sigma, cat # C9687; to detect CNDs) and activated caspase-3 (Invitrogen, cat # G748A; to detect apoptosis), as described.¹¹ We detected pancytokeratins with anti-mouse Alex Fluor 488 (Jackson ImmunoResearch, cat# 115-485-003) and caspase with anti Rabbit Cy3 (Jackson ImmunoResearch, cat #711-165-152).

Cystograms

Cystograms were performed in postnatal day 1 (P1) mice to assess for VUR, with the operator blinded to genotypes. After a ventral midline incision, urinary tracts were isolated. The bladder was punctured with a 30-gauge needle, connected via tubing to a 60 mL syringe filled with 1% methylene blue in PBS. The syringe was raised from 30 cm above the microscope stage to 150 cm at 6 cm/sec. The height when reflux was visualized and dye passed through the urethra was noted. VUR was graded according to international classifications.¹³ The experiment was stopped after the syringe was held at 150 cm for 10 seconds and no reflux occurred. Among 21 Fgfr2^Mes−/− and 51 control mice used for cystograms, 2 mutants and 1 control were not assessed due to technical difficulties (bladder punctured completely through). Kidney long axes were measured in the first consecutive 15 Fgfr2^Mes−/− and 45 control mice with successful cystograms using ImageJ. Gravity filling mouse bladders to assess VUR has been validated.¹⁴

3-D Reconstructions

Embryonic Fgfr2^Mes−/− mice have abnormally positioned ureteric buds

While our previous work demonstrated that some Fgfr2^Mes−/− mice initially had 2 ureteric buds¹¹, it was unclear if single ureteric buds in mutants were positioned abnormally. We measured the common nephric duct, the nephric duct distal to ureteric bud stalks, by performing whole mount in situ hybridization on E11.5 mutants and controls for Ret (Figure 1). Controls had a mean (± standard deviation) CND length of 19.9 +/− 2.38 μm (n=10 CNDs). The mean CND length of all mutant Fgfr2^Mes−/− mice including those with multiple ureteric buds (measured from the most anterior bud), was 27.0 +/− 5.31 μm (n=10) (p=0.0012 compared with controls). Among Fgfr2^Mes−/− mice with single ureteric buds, mean CND length was 24.4 +/− 3.2 μm (n=7) (p=0.004 versus controls).

To ensure that differences in CND length were not from apoptosis, we performed whole mount immunofluorescence in E11.5 embryos for activated caspase-3 and pancytokeratins. As shown (Figure 2), mutants and controls stain similarly for caspase 3 along CNDs, indicating similar apoptotic rates. Taken with Ret expression, as a group Fgfr2^Mes−/− mice have longer CNDs and thus cranially shifted ureteric buds (even when only one bud is present).

E15.5 Fgfr2^Mes−/− embryos have low/caudal ureteral insertions into the bladder

To determine whether high/cranial ureteric bud position results in ureteral insertion abnormalities, we performed 3D reconstructive imaging on E15.5 Fgfr2^Mes−/− and controls through the UVJ (supplemental movies 1 and 2). As shown (Figure 3), still images of the external/posterior bladder surface show that mutant ureters appear to insert low/caudal compared with controls. As shown (Table 2), mutant external ureteral angles are more acute than controls (mutant: 20.2 ± 6.7°; control: 27.1 ± 3.3°, p = 0.048). Two mutants had ureteral angles on right and left that were each > 2 standard deviations smaller than control mean angles. Smaller angles are consistent with mutant ureters inserting more caudal along the bladder.

In interrogating position of internal ureteral orifices in the bladder (Figure 3, Table 2), mutants had smaller mean lengths from the ureteral orifice to the bladder neck than controls (mutants: 162 ± 16.2 μm; controls: 247 ± 17.3 μm, p = 0.0000048). Mean mutant interureteral length also trended toward being smaller than control length, but did not reach statistical significance (mutants: 227 ± 27.5 μm; controls: 296 ± 37.4 μm, p = 0.063). Thus, it appears that Fgfr2^Mes−/− ureters empty in a more caudal position (and closer together) into the bladder versus controls.

Fgfr2^Mes−/− newborn mice have high rates of VUR

To test whether abnormal ureteric bud position predisposed Fgfr2^Mes−/− mice to VUR, we performed cystograms on newborn P1 mice (Figure 4). Among 49 controls, only 2 (4.0%) displayed unilateral VUR. In contrast, 9 out of 19 (47.4%) Fgfr2^Mes−/− mutants had unilateral or bilateral VUR (p=0.00006 versus controls). There were 10 total mutant kidneys that refluxed (1 bilateral, 3 unilateral left-sided, and 5 unilateral right-sided). Thus, shortly after birth, a significant number of Fgfr2^Mes−/− mice had VUR compared with controls.

In examining gross morphology of Fgfr2^Mes−/− newborn kidneys after cystograms, none had obvious ureteral obstruction, duplex collecting systems, or renal agenesis. Mean newborn mutant kidney long axis (including those with and without VUR) was 10.7% smaller than mean control long axis (mutant: 0.29 ± 0.044 cm; control: 0.33 ± 0.044 cm, p = 0.00008). Fgfr2^Mes−/− kidneys with VUR (n = 7) and without VUR (n = 23), however, had no statistical differences in mean long axis (VUR: 0.29 ± 0.27 cm; no VUR: 0.29 ± 0.31 cm, p = 0.60). Furthermore, while 4 out of 10 (40%) kidneys with reflux were one standard deviation or smaller than mean control length compared with 7 out of 30 (23.3%) mutant kidneys without VUR, the differences did not reach statistical significance (p = 0.42). Thus, while newborn Fgfr2^Mes−/− kidneys are sometimes smaller than controls, there do not appear to be differences between mutants with or without reflux.

Fgfr2^Mes−/− mutants with VUR have abnormal ureteral insertions in the bladder

To examine ureteral insertion into the bladder, we performed 3-D reconstruction of the UVJ in a subset of P1 Fgfr2^Mes−/− (n=3) with VUR and controls without VUR (n=3). As shown (Figure 5, supplemental movies 3,4), control mice without VUR have ureteral insertions that are distant from the bladder neck and that insert parallel to the transverse plane through the bladder dome (Figure 5A, supplemental movie 3). In contrast, a mutant mouse with right sided, unilateral VUR, has a right ureter that inserts low/caudal on the external bladder wall, near the bladder neck (Figure 5B, supplemental movie 4). Furthermore the interureteral plane in the Fgfr2^Mes−/− mouse deviates significantly from the transverse plane of the bladder dome.

To quantify differences between mutants with VUR and controls, we generated triangles formed by the external ureteral insertion points and bladder neck (Figure 5 C, D) and internal triangles from ureter orifices and bladder neck and intravesicular tunnel lengths (Table 3). Fgfr2^Mes−/− mice had aberrant increases in external angles formed at the ureteral insertion on the side that had reflux (in all cases, mutants had right sided VUR). Mutants had a right sided mean angle of 56.0 +/− 6.1° versus 40.8 +/− 3.2° in the controls (p=0.0075). In addition, the mean of the right (refluxing) angles were statistically different than the means of the left (non-refluxing) angles (56.0 +/− 6.1° versus 32.1 +/− 7.5°; p=0.013) in mutants. In contrast control exterior right and left sided angles were not statistically different (p=0.354). Angles formed at internal ureteral orifices were not different between mutants and controls or between mutant right and left sides. There were also no differences between intravesicular tunnel lengths of refluxing mutant ureters and controls or non-refluxing mutant ureters. Thus Fgfr2^Mes−/− mutants appear to have refluxing ureters that insert abnormally into the bladder.

Fgfr2^Mes−/− mutants have decreased Fgfr2 expression in kidneys

We then determined whether differences in Fgfr2 expression might contribute to the high incidence of VUR in Fgfr2^Mes−/− mice versus controls. As shown (Figure 6), Fgfr2 is expressed in ureteric tips of both mutant and control kidneys, but relatively absent in mesenchymal tissues between mutant ureteric tips compared with controls. Relative Fgfr2 signal in ureters and bladders was comparable in mutants and controls by in situ hybridization (not shown). Quantitative real time PCR confirmed less Fgfr2 mRNA in mutant kidneys (88.5% compared with controls, p = 0.0017). In contrast there were no statistical differences between mutant ureters (100.4% compared with controls, p = 0.93) and although bladders trended toward a difference (90.5% compared with controls, p = 0.066). Thus, there is less Fgfr2 in mutant kidneys, but similar levels in ureters and bladders compared with controls.