Vesicoureteric reflux and reflux nephropathy: from mouse models to childhood disease
- First Online:
- Cite this article as:
- Fillion, M., Watt, C.L. & Gupta, I.R. Pediatr Nephrol (2014) 29: 757. doi:10.1007/s00467-014-2761-3
- 534 Views
Vesicoureteric reflux (VUR) is a common congenital urinary tract defect that predisposes children to recurrent kidney infections. Kidney infections can result in renal scarring or reflux nephropathy defined by the presence of chronic tubulo-interstitial inflammation and fibrosis that is a frequent cause of end-stage renal failure. The discovery of mouse models with VUR and with reflux nephropathy has provided new opportunities to understand the pathogenesis of these conditions and may provide insight on the genes and the associated phenotypes that need to be examined in human studies.
KeywordsVesicoureteric refluxMouse modelsReflux nephropathyChildren
Congenital abnormalities of the kidney and urinary tract, or CAKUT, constitute a spectrum of kidney and urinary tract disorders, the most common of which is vesicoureteric reflux (VUR). VUR is a urinary tract defect in which an abnormal uretero-vesical junction (UVJ) promotes the retrograde flow of urine from the bladder to the kidneys. VUR has been reported to affect anywhere from 1 to 30 % of the general population . This variation in VUR prevalence is likely due to differences in study populations, including ethnicity and age. In many cases, VUR resolves during childhood, possibly due to ureter elongation and bladder growth, both of which lead to a better functioning UVJ . Establishing the diagnosis and incidence of VUR is also complicated by the fact that the test required for diagnosis, the voiding cystourethrogram (VCUG), is difficult to undergo such that the child and/or the parent may not consent to the test. The VCUG requires catheterization of the urethra, followed by filling of the bladder with radio-opaque dye, and imaging with fluoroscopy to view the urinary tract while the patient voids. VUR is graded using a 5-point scale as mild (grades I–II), moderate (grade III), or severe (grade IV and V) based on the degree of retrograde filling and the extent of ureteral dilatation observed from the VCUG . Other imaging modalities, such as ultrasound, radionucleotide cystogram, and magnetic resonance imaging, have been used to limit radiation exposure; however, the VCUG remains the gold standard for diagnosis and for grading .
Roughly 30 % of children who develop a urinary tract infection are subsequently diagnosed with VUR . Urinary tract infections that include the kidney (pyelonephritis) can lead to renal scarring (reflux nephropathy) and result in renin-mediated hypertension and even renal failure. In fact, reflux nephropathy is reported as the fourth and seventh most frequent cause of end-stage renal failure in pediatric and adult databases, respectively [4, 5]. While it is postulated that children with reflux nephropathy have more malformed UVJs, co-existing congenital renal malformations, more severe kidney infections, and/or an inadequate host immune response to infection, this has not been clearly demonstrated in the literature. Children with VUR continue to be treated with surgical reconstruction of the UVJ and/or antibiotic prophylaxis to prevent urinary tract infections despite the lack of evidence that these treatments have been beneficial in decreasing the incidence of renal failure from reflux nephropathy .
The ureteric bud: the embryonic origin of the ureter
To understand how the refluxing UVJ arises, the stages of urinary tract development need to be reviewed. The intermediate mesoderm on each side of the developing embryo gives rise to the nephric duct, an epithelial tube that extends along the length of the embryo and forms the future kidney, ureter, and gonad. During the 5th week of gestation in humans, a swelling, defined as the ureteric bud, emerges from the nephric duct. The distal portion of the ureteric bud, that is attached to the nephric duct, elongates to form the ureter, while the proximal portion undergoes branching morphogenesis to generate the collecting duct network of the adult kidney. Each ureteric bud tip induces the adjacent mesenchyme to epithelialize and form a nephron, the filtration unit of the kidney. Nephrogenesis begins at week 6 and is completed by the 36th week of gestation . Kidney and urinary tract development occurs similarly in the mouse, but over a 20 to 21-day gestation period: the ureteric bud emerges from the nephric duct at embryonic day 10.5 (E10.5), and nephrogenesis is completed approximately 2 weeks after birth. Initially, the ureter is joined via the common nephric duct to the urogenital sinus that will give rise to the future bladder. As development progresses, the common nephric duct undergoes apoptosis, and each ureter undergoes a vertical and lateral movement that establishes an independent entry point into the bladder and allows for the formation of the UVJ at around E14 in the mouse and at approximately E47 in humans [8, 9].
Ureteric bud outgrowth is induced when the ligand glial-derived neurotrophic factor (GDNF), which is expressed in the metanephric mesenchyme, binds to the RET tyrosine kinase receptor and the GFRA1 co-receptor, both expressed by the ureteric bud . Transcription factors, such as PAX2, GATA3, EYA1, SIX1, SALL1, and HOX11, positively regulate GDNF expression . The secreted glycoproteins of the WNT family are also important for ureteric bud outgrowth. WNT11 is required for maintaining normal expression of GDNF and together with the GDNF/RET pathway, it controls ureteric bud branching . Other factors, such as fibroblast growth factor receptor 2 (FGFR2), promote ureteric bud induction, but independently of the GDNF/RET pathway . GDNF/RET activity is negatively regulated by FOXC1/FOXC2, SLIT2-ROBO2, and SPRY1. The deletion of any of these genes results in the formation of multiple ureteric buds and urogenital malformations in mice [14–16]. Transforming growth factor beta 1 (TGFβ1) and bone morphogenetic protein 4 (BMP4) also inhibit the GDNF/RET pathway and restrict the location of the emerging ureteric bud . Studies in humans and mice have demonstrated an association between an abnormally positioned ureteric bud and CAKUT phenotypes, including VUR. Mackie and Stephens hypothesized that if the ureteric bud emerges more caudally along the nephric duct, the ureter may insert in the bladder prematurely and be positioned ectopically, thus resulting in VUR . If the ureteric bud emerges more cranially, the ureter may fail to interact with the primitive bladder and to separate from the nephric duct at the correct time, resulting in urinary tract obstruction . Many of the mouse models of VUR exhibit caudally positioned ureteric buds when compared to their wildtype littermates . The emergence of the ureteric bud is a critical step in urinary tract development, therefore many of the genes implicated in ureteric bud outgrowth have been examined as candidate genes in human studies and in mouse models of VUR; these will be discussed later in this review.
The uretero-vesical junction
A large number of molecules have been implicated in the cytodifferentiation of ureteric mesenchyme, including ID2, Calcineurin (CD), SHH, DLGH1, TSHZ3, BMP4, TBX18, SOX9, and CTNNB1[21–28]. Knockouts for all of these genes lead to hydronephrosis with variable severity of the hydroureter in the mouse. The transcription factor TBX18 is required for specifying mesenchymal precursor cells to a ureteric mesenchyme fate and for their differentiation into smooth muscle cells . In Tbx18−/− mice, the ureteric mesenchyme fails to condense around the distal ureter, and this results in a lack of smooth muscle formation . BMP4 inhibits budding of the distal ureteric bud and directs the ureter to elongate. BMP4 also promotes smooth muscle cell differentiation. The transcription factors TSHZ3 and SOX9 act as downstream mediators of BMP4 signaling in the ureteric mesenchyme and upregulate expression of myocardin (MYOCD), a key regulator of smooth muscle cell differentiation .
A functional urinary tract transmits peristaltic waves to propel urine from the kidneys to the bladder. These waves prevent the reflux of urine retrogradely from the bladder to the kidneys. Peristalsis is initiated by specialized smooth muscle cells at the pelvic-ureteric junction. Hyperpolarization-activated cation 3 channels (HCN3) are highly expressed in the newborn mouse pelvic-ureteric junction and appear to be critical for propagating the electrical signal that drives peristaltic waves from the proximal to the distal urinary tract .
Towards the distal end of the urinary tract, at the UVJ, there is a functional flap valve that prevents the retrograde flow of urine from the bladder back to the kidneys. The common nephric duct, which initially connects the ureter and the nephric duct, undergoes apoptosis at E12, thus allowing for the ureter to enter the bladder and form the UVJ. Indeed, defects in apoptosis of the common nephric duct, as observed in Raldh2 deficiency and in LAR family receptor type S and F (Ptprs and Ptprf) deficiency, lead to failure of separation of the nephric duct and ureter and urinary tract obstruction in mice [32, 33]. These steps of ureter maturation and insertion into the bladder have been shown to be regulated by a GATA3–RALDH2–RET molecular network .
The portion of the ureter within the bladder, the intravesical ureter (IVU), is enveloped by muscle fibers that interface with the bladder musculature to form the UVJ. During voiding, when there is an increase in intravesical pressure, the bladder and ureteral musculature compress the intravesical ureter and together create a functional flap valve that prevents the reflux of urine. These tensile forces are predominantly generated by the trigone, a triangular region between the two ureters and the urethra, which is predominantly made up of bladder muscle. Through linage tracing experiments in mice in which Rarb2 was conditionally deleted from ureteric mesenchyme, it is known that the trigone is derived mostly from bladder muscle with a small contribution from ureteral smooth muscle at its lateral margins .
Structural abnormalities of the UVJ that have been reported in patients with VUR include short intravesical ureters, ectopic insertion of the ureters in the bladder, defects in the smooth muscle layer of the ureter and/or the bladder, and/or a dilated ureteral orifice . Many of the mouse models of VUR also demonstrate a short intravesical ureter . Defects in the extracellular matrix, which provides support for this anti-reflux mechanism, may also compromise the trigone’s ability to compress the IVU. These defects of the UVJ may arise with or without a concurrent renal malformation, depending on when and how they arise during development.
VUR is heritable and genetically heterogeneous
Twin studies have demonstrated that VUR occurs more frequently in monozygotic than dizygotic twins, suggesting it is a heritable disorder . A variety of inheritance patterns have been observed in families with VUR, including monogenic dominant, monogenic recessive, x-linked, and complex patterns .
Various approaches have been employed to identify candidate genes that confer VUR susceptibility, including the genetic analysis of rare patients with VUR and chromosomal abnormalities, whole-genome mapping studies in families with multiple affected individuals, and direct sequencing of candidate genes. These approaches have led to the discovery of many loci and genes that are associated with VUR, leading to the conclusion that VUR is genetically heterogeneous. VUR is also phenotypically heterogeneous: some of the genes function during development at the time of emergence of the ureteric bud (e.g., GATA3, AGTR2, RET, PAX2, ROBO2), another group functions in the uroepithelium to maintain its impermeable state (e.g., UPK2 and UPK3), others appear to be important for the host immune response [e.g., genes within the human leukocyte antigen (HLA) complex], and finally another group appears to be involved in fibrosis and repair (e.g., ACE, TGFB1) [38–46].
Thus far, human studies have not identified one major gene that predisposes to familial VUR. This may be due to the fact that a given gene mutation may give rise to a range of phenotypes within the CAKUT spectrum, such that within an affected family, some individuals may demonstrate VUR while others may have other CAKUT features. Another explanation for the failure to identify one major gene is that VUR is genetically heterogeneous. Each family may have their own “private mutation” such that any given gene may have a limited effect size overall. The ROBO2 gene encodes a transmembrane receptor for the SLIT ligand and plays a key role in the formation of the ureteric bud. Mutations in ROBO2 have been demonstrated to segregate with VUR in humans [47, 48]. Indeed, Bertoli-Avella et al.  suggested that heterozygous mutations in ROBO2 could account for 5 % of familial VUR. However, Dobson et al.  were unable to validate these findings in their cohort of Irish patients with VUR. SOX17 is an HMG box transcription factor and WNT signaling antagonist that is expressed in the ureteric bud and the metanephric mesenchyme during kidney development . Heterozygous mutations in SOX17 were identified in two families and in 4/178 sporadic cases of VUR . Most recently, DSTYK, a dual serine–threonine and tyrosine protein kinase that functions downstream of FGF signaling, has been associated with CAKUT phenotypes, including VUR. Sanna-Cherchi et al. screened 311 unrelated patients with congenital kidney or urinary tract malformations and identified heterozygous mutations in seven of them (2.3 %) .
Many of the genes identified in humans have been functionally validated through the creation of transgenic and knockout mouse lines that have been found to have CAKUT phenotypes, including VUR. Similarly, many susceptibility loci have been reported in humans [17, 52–57], however, few of these loci have been validated when studied in other cohorts , which may be due to the genetic and phenotypic heterogeneity of VUR. At least one of these loci at chromosome 2p24-25 may have a mouse counterpart. This locus has been found to be orthologous to the murine Vurm1 locus that will be discussed in the following section [55, 59].
Animal models of VUR
To understand the biology of a condition or disease, animal models are indispensable. While porcine models of surgically induced VUR do exist and have been used, pigs are not the easiest animal model to work with due to their large size and cost of housing. Mice, on the other hand, are small, easy to house, and have short generation times. The mouse is a good model for the study of human disease since organ structure is similar between the two species, most human genes have homologs in the mouse, and synteny, the conserved gene order in portions of the genome, is shared between the two species. Most importantly, advances in genetics have established the mouse as a powerful model organism to investigate the role of particular genes through the manipulation of its genome.
There are at least 15 mouse models of VUR representing the following phenotypes: VUR with renal malformation, VUR with renal malformation and duplex systems, VUR with abnormal bladder phenotypes, and VUR with normally formed kidneys . Our laboratory has developed a method to diagnose VUR in the mouse . The bladder is punctured with a needle and then dye is injected into the bladder using a hydrostatic pressure gradient. VUR is diagnosed when dye is observed to flow back towards the kidneys. The severity of VUR can be quantified based on the amount of applied hydrostatic pressure when reflux is observed. A UVJ that refluxes at a low pressure is more abnormal than one that refluxes at a high pressure. Using this method, our laboratory has identified many mouse models with VUR, including several inbred mouse lines .
Mouse models of VUR with renal malformation
The HoxB7/Ret+/− mouse model has sustained ectopic expression of the RET receptor tyrosine kinase such that RET is expressed throughout the ureteric bud, the collecting ducts, and the ureter throughout development compared to its normal expression which is restricted to the ureteric bud and its tips . This mouse model exhibits VUR, short intravesical ureters, and renal dysplasia . The kidney and urinary tract defects appear to originate early in development secondary to a ureteric bud that emerges from a more caudal location along the nephric duct . A study of French Canadian families with VUR identified a polymorphism in RET that is predicted to function as a gain-of-function mutation, somewhat similar to overexpression of the RET gene in the Hoxb7/Ret+/− mouse .
PAX2 is a paired-box transcription factor involved in regulating ureteric bud outgrowth and ureter maturation . Humans with dominantly inherited Renal-Coloboma Syndrome (RCS) frequently have a one-base insertion of a guanosine in the second exon of PAX2 that leads to a truncated and non-functional protein . The Pax21Neu+/− mouse model carries the identical mutation in Pax2 and completely phenocopies all of the features observed in humans with RCS, including optic colobomas, renal malformations, and VUR [9, 65].
Mouse models of VUR, renal malformations, and duplex systems
Mice with a conditional deletion of the receptor tyrosine kinase Fgfr2 in the metanephric mesenchyme (Fgfr2Mes−/−) have duplex systems from the emergence of two ureteric buds from one nephric duct. In addition to duplex systems, these mice exhibit renal agenesis, ureter obstruction, and VUR [13, 66]. When Fgfr2 is conditionally deleted in the tailbud mesenchyme-derived stroma (Fgfr2ST−/−) that lies between the nephric duct and metanephric mesenchyme, the embryos display aberrant ureteric budding such that the ureteric bud emerges from a more cranial or a more caudal location along the nephric duct. In turn, this leads to an abnormal site of ureter insertion in the bladder, shorter intravesical ureters, and VUR . In contrast, when Fgfr2 is conditionally deleted from the ureteric bud, the embryos predominantly exhibit a renal phenotype with less ureteric bud branching and small kidneys . Mutations in exon 3a or 3c of FGFR1/FGFR2 are known to result in Pfeiffer syndrome, a dominantly inherited syndrome characterized by several organ defects in addition to CAKUT phenotypes such as hydronephrosis, pelvic kidney, and high-grade VUR .
Another mouse model that exhibits VUR with duplex urinary tract systems and hydronephrosis is the Robo2del5/del5-Robo2del5/flox mouse. These mosaic mutant mice were generated using a Cre–Lox system since Robo2 null mice die from renal failure in the first 48 h of life . Importantly, mutations in ROBO2 have been identified in sporadic and familial forms of VUR [47, 48] .
Other mouse models that exhibit this phenotype are seen when the angiotensin type II receptor is knocked out in the Agtr2−/− mouse and when the homeobox transcription factor, LIM1, is conditionally deleted in the nephric duct [39, 71].
Mouse models of VUR with abnormal bladder phenotypes
The uroplakin proteins are found in plaques covering the umbrella cell layer of the uroepithelium that lines the urinary tract. They maintain the impermeability of the uroepithelium and prevent it from rupturing during bladder distension; they also play a role in preventing bacterial adherence . The ablation of the Uroplakin 3 gene (Upk3) in the mouse was observed to result in a leaky uroepithelium that was devoid of umbrella cells . These mice also have hydronephrosis, enlarged ureteral orifices, and VUR . A similar phenotype was found when the Upk2 gene was ablated in the mouse . Humans with polymorphisms in UPK2 or UPK3 have been found to have VUR, renal malformations, and renal failure [42, 43].
Mouse models of VUR with normally formed kidneys
The Vurm1 locus is large, containing 168 genetic elements and 104 known protein-coding genes, therefore, congenic lines are currently being established to identify the Vurm1 gene . Congenic mouse lines are generated by backcrossing an affected inbred mouse strain to a background strain without the phenotype for at least ten generations to obtain a locus of interest from the donor-affected strain on the background of a non-affected strain and vice versa. These two congenic mouse lines are then intercrossed to create subcongenic lines in which the locus of interest inherited from the affected mouse line is fragmented due to recombination with the allele from the non-affected strain. The creation of congenic and sub-congenic mouse lines is crucial to identify the causative gene in Vurm1.
VUR and reflux nephropathy
Vesicoureteric reflux can lead to renal scarring, known as reflux nephropathy (RN), which is characterized histologically by chronic tubulointerstitial inflammation and the presence of fibrotic scars. RN accounts for up to 25 % of end-stage renal disease (ESRD) in children . The gold standard to establish the diagnosis is from a renal biopsy, however, this procedure is invasive and not always performed. RN is more typically diagnosed via a dimercaptosuccinic acid (DMSA) scan which involves the intravenous injection of DMSA that is taken up and filtered by the kidneys. When a region of reduced uptake is observed, this can represent a congenital absence of nephrons or an acquired loss secondary to a renal scar; therefore, the DMSA scan has some limitations in defining and understanding RN [74, 75].
There are two main theories that have been put forth to explain how VUR progresses to RN. In the first, RN arises from the reflux of infected urine from the bladder to the kidneys. This leads to a kidney infection which triggers the innate immune response and the influx of inflammatory cells. Others believe that the reflux of even sterile urine can cause an increase in intrarenal pressure that can progress to chronic fibrosis. Finally, others postulate that cases of RN that progress to ESRD likely developed in the presence of a congenitally malformed kidney . The pig model has been most frequently used to examine the relationship between VUR and RN although VUR must be surgically induced since these animals do not reflux spontaneously. The results from this model are contradictory, since some authors report that the reflux of sterile urine is sufficient to induce RN while others report that a bacterial kidney infection is also required [77, 78].
Mouse model of reflux nephropathy
We examined the roles of VUR and kidney infection in the development of RN using the refluxing C3H/HeJ and C3H/HeN inbred mouse lines and compared these to the non-refluxing C57Bl/6 J line. The C3H/HeJ and C3H/HeN mouse lines have identical genomes except that the HeJ substrain has a defect in the innate immune system from a spontaneous mutation in the Toll-like receptor 4 gene (Tlr4). TLR4 is a receptor expressed on uroepithelial cells that detects lipopolysaccharide, a protein expressed by Gram-negative bacteria such as uropathogenic Escherichia coli (UPEC). When bacteria invade the urinary tract, they are initially detected by the uroepithelial cells of the urinary tract via TLR4. TLR4 and other proteins in the innate immune response activate signaling cascades that promote the upregulation of proinflammatory cytokines, such as interleukin (IL)-6 and IL-8. IL-8 stimulates the release of neutrophils and other immune cells from the circulation. Neutrophils then migrate to the site of infection and destroy bacteria by phagocytosis, thus allowing bacteria to be cleared.
In the absence of induced urinary tract infection, we noted no renal abnormalities in C3H/HeN mice, suggesting that the reflux of sterile urine is not sufficient to induce RN in this model . The innoculation of the bladders of C3H/HeN and non-refluxing C57Bl/6 J mice with 108 colony forming units (CFU) of UPEC resulted in kidney infections in HeN but not B6 mice. VUR initiated and sustained the kidney infection since the HeN mice still had higher renal bacterial counts at 2 and 6 days post-inoculation, while the C57Bl/6 J mice had no infection. HeN mice with kidney infections exhibited histological evidence of RN, including an increase in neutrophils and collagen deposition in the kidney. VUR thus promotes and sustains kidney infection, and this leads to the pathological changes observed in RN .
Human studies of VUR and the future
The investigation of VUR in mouse models has shed light on the mechanistic roles of several genes and may provide insight on how human cohorts should be classified and studied in the future. We believe that patients with primary VUR need to be classified into those with and those without renal malformation using ultrasound measurements of kidney length and DMSA scans performed soon after the diagnosis of VUR is made. This is the first step in understanding the role of congenital renal malformations in the progression of VUR to RN and to ESRD. We further believe that prospective long-term follow-up of the cohorts thus far recruited is essential to understand the role of infection in the progression of VUR to RN. Once we have better long-term follow-up data, we may be able to identify specific VUR phenotypes with deleterious outcomes and thus establish a screening method to identify patients at risk. Until we begin to answer some of these natural history questions, we will remain uninformed as to who is at risk for RN and ESRD and who is not at risk.
This work was supported by a grant from the Canadian Institute of Health Research to IRG. IRG holds a scholarship award from the Fonds de la Recherche en Santé du Québec.