Novel Insights into the Pathogenesis of Monogenic Congenital Anomalies of the Kidney and Urinary Tract

Abstract

Congenital anomalies of the kidneys and urinary tract (CAKUT) comprise a large spectrum of congenital malformations ranging from severe manifestations, such as renal agenesis, to potentially milder conditions, such as vesicoureteral reflux. CAKUT causes approximately 40% of ESRD that manifests within the first three decades of life. Several lines of evidence indicate that CAKUT is often caused by recessive or dominant mutations in single (monogenic) genes. To date, approximately 40 monogenic genes are known to cause CAKUT if mutated, explaining 5%–20% of patients. However, hundreds of different monogenic CAKUT genes probably exist. The discovery of novel CAKUT-causing genes remains challenging because of this pronounced heterogeneity, variable expressivity, and incomplete penetrance. We here give an overview of known genetic causes for human CAKUT and shed light on distinct renal morphogenetic pathways that were identified as relevant for CAKUT in mice and humans.

Congenital anomalies of the kidneys and urinary tract (CAKUT) are common malformations with a potentially severe effect on health. CAKUT causes about 40% of cases of ESRD in patients who develop it within the first three decades of life.^1,2 Manifestations of the CAKUT spectrum account for 20%–30% of congenital malformations^3–7 and constitute a frequent cause of birth defects (approximately three to six per 1000 live births).^4,8,9 CAKUT may occur either as an isolated condition or along with extrarenal manifestations as part of a syndromic disorder.^10–13 To date, >200 clinical syndromes have been described that comprise features of CAKUT as part of their distinct phenotype.¹⁴

The term CAKUT summarizes a large variety of diverse congenital malformations that range from conditions of the upper urinary tract (e.g., renal agenesis and renal hypodysplasia) to phenotypes primarily affecting the lower urinary tract, such as vesicoureteral reflux (VUR), ureterovesical junction obstruction (UVJO), and posterior urethral valves (PUVs).^3,8,11 In addition, abnormalities of kidney shape or anatomic position, such as horseshoe kidney or pelvic kidney, may occur. Malformations, such as UVJO, PUV, and VUR, may lead to anatomic (UVJO and PUV) or functional (VUR) stasis of urine flow, thereby increasing the risk of urinary tract infections, which may result in renal scarring and potentially, CKD (e.g., from reflux nephropathy). It is important to note that, although different manifestations from within the CAKUT spectrum may coexist in the same individual,^3,15 unilateral CAKUT (e.g., renal agenesis) can also exist in the presence of an entirely unblemished contralateral urinary system.¹⁰

The diverse manifestations of CAKUT phenotypes are thought to result from disturbances at any point in renal morphogenesis (Figure 1).^15,16 Most importantly, imbalances in the communication between the metanephric mesenchyme (MM) and the ureteric bud (UB) are believed to be central to the pathogenesis of CAKUT phenotypes (Figure 1, C–E).¹⁵ The developmental origin of PUVs, however, likely differs from other CAKUT manifestations and remains poorly understood.^3,17,18

Figure 1.

Development of the kidneys and urinary tract. (A) The bilateral nephric ducts (NDs; alternatively mesonephric ducts or Wolffian ducts) and the nephric cords (NCs) are the precursor structures of the adult urinary system. Both originate from the embryonic intermediate mesoderm. The cells of the ND undergo an early mesenchymal to epithelial transition and assemble into epithelial tube–like structures. The associated NC retains characteristics of mesenchymal tissue.^16,62 (B) As the embryo develops, the ND elongates caudally. At approximately E9.5, the most caudal portion of the ND fuses with the cloacal epithelium (Cl). The cloaca is the embryonic precursor of the bladder, and it is derived from cloacal endoderm.^16,66 (C) The NC reorganizes and forms a morphologically distinct domain: the MM. Renal morphogenesis is initiated and maintained by reciprocal interactions between the epithelial ND and the MM. In mice, at embryonic day approximately E10–E10.5, signals from the MM induce the formation of a circumscribed, broad swelling of the ND at the level of the MM.^15,16 (D) At E10.5 in mice and around the fifth week of human gestation, the UB emerges from the swollen portion of the ND and grows dorsally toward the MM.^4,15,16,74 The caudal part of the ND, which is located between the UB and the insertion into the Cl, is referred to as the common nephric duct (CND). (E) Stimulated by MM-derived signals, the UB begins to branch repeatedly (branching morphogenesis) at approximately E11.5. Through continuous reciprocal induction, the MM is important for promoting and maintaining branching events of the UB. The UB branching continues for approximately 9–13 cycles (mice) and then slows down after approximately E15.5. Via branching morphogenesis, the UB gives rise to the renal collecting system consisting of collecting ducts and renal pelvis as well as the ureter. Reciprocally, signals from the UB also support development of MM cells. The MM that is in closest proximity to the UB tips condenses and forms the so-called cap mesenchyme (CM). Stimulated by signals from the UB, the CM undergoes a mesenchymal to epithelial transition. The epithelial cell population subsequently gives rise to structures of the nephron (glomerulus, the proximal tubule, and the distal tubule). Modified from refs. 16 and 62, with permission.

We have previously hypothesized that a high fraction of human CAKUT may be caused by single-gene defects. Monogenic diseases, also known as Mendelian disorders, are caused by mutations in a single causative gene.¹⁰ Supporting evidence for a monogenic etiology in the case of CAKUT comes from (1) the familial occurrence of CAKUT phenotypes (approximately 10%–15% of patients with CAKUT cases are familial), (2) the existence of monogenic syndromes with a distinct syndromic phenotype that includes manifestations from within the CAKUT spectrum (>200 to date), (3) the existence of monogenic mouse models that exhibit CAKUT (>180 to date), and (4) the fact that the development of the kidney and urinary tract is governed by distinct developmental genes.^10,11,19,20

Approximately 40 different monogenic causes for human CAKUT (25 dominant and 15 recessive) have so far been identified (Table 1).^{12,13,21–51} Many established CAKUT genes encode transcription factors and follow a dominant pattern of inheritance.^3,10 Some of the genes listed in Table 1, however, remain questionable in regard to their pathogenic effect and will require more support through functional studies in the future.^52,53

Table 1.

Fifteen recessive and 25 dominant genes that represent monogenic causes of human CAKUT if mutated

Gene	Protein	Refs.
Autosomal recessive
ACE	Angiotensin I–converting enzyme	38
AGT	Angiotensinogen	38
AGTR1	Angiotensin II receptor, type 1	38
CHRM3	Muscarinic acetylcholine receptor M3	39
FRAS1	ECM protein FRAS1	42
FREM1	FRAS1-related ECM protein 1	42
FREM2	FRAS1-related ECM protein 2	42
GRIP1	Glutamate receptor interacting protein 1	42
HPSE2	Heparanase 2 (inactive)	43
ITGA8	Integrin-α8	44
LRIG2	Leucine-rich repeats and Ig-like domains 2	45
REN	Renin	38
TRAP1	Heat shock protein 75 (also known as TNF receptor–associated protein 1)	46
Autosomal dominant
BMP4	Bone morphogenic protein 4	21
CHD1L	Chromodomain helicase DNA binding protein 1 like	22
CRKL	CRK-like proto-oncogene, adaptor protein	23
DSTYK	Dual serine/threonine and tyrosine protein kinase	24
EYA1	Eyes absent homolog 1	25
GATA3	GATA binding protein 3	26,188
HNF1B	HNF homeobox B	13
MUC1	Mucin 1	27
NRIP1	Nuclear receptor interacting protein 1	140
PAX2	Paired box 2	12
PBX1	PBX homeobox 1	53
RET	Proto-oncogene tyrosine-protein kinase receptor Ret	28
ROBO2	Roundabout, axon guidance receptor, homolog 2 (Drosophila)	29,189
SALL1	Sal-like protein 1 (also known as spalt-like transcription factor 1)	49
SIX2	SIX homeobox 2	21
SIX5	SIX homeobox 5	51
SLIT2	Slit homolog 2	29
SOX17	Transcription factor SIX-17	30
SRGAP1	SLIT-ROBO Rho GTPase activating protein 1	29
TBX18	T-box transcription factor TBX18	31
TNXB	Tenascin XB	32
UMOD	Uromodulin	33
UPK3A	Uroplakin 3A	34
WNT4	Protein Wnt-4	35–37
X-linked recessive
KAL1	Anosmin 1	47

Currently, at most, up to 18% of CAKUT cases can be explained by these established monogenic causes (Table 1).^3,10 The percentage rate of molecularly “solved” cases, of note, is hereby likely dependent on the corresponding selection of patients and controls. Approximately 10%–15% of cases have additionally been described to occur due to copy number variations that predominantly involve either the HNFB1 locus on chromosome 17 or the Di-George/velocardiofacial syndrome region on chromosome 22.⁵⁴

It is likely that hundreds of additional monogenic causes of human CAKUT have yet to be identified. Apart from the distinct heterogeneity, with 40 genes identified so far (Table 1), the identification of novel CAKUT-causing genes is complicated by variable expressivity and incomplete penetrance.^10,11 These features are frequently encountered in dominant diseases and can result in the phenomenon that either individuals carrying a mutation in a CAKUT gene can present with a phenotype that differs from the phenotypic manifestation of other individuals with an identical mutation (variable expressivity) or alternatively, an individual carrying a mutation does not exhibit a CAKUT phenotype at all (incomplete penetrance).¹⁰

One theory proposed by Ichikawa et al.¹⁵ suggests that this phenotypic variability may occur as a result of (1) stochastic spatiotemporal differences during key events of renal morphogenesis, (2) gene dosage effects, and (3) redundancy of genes that belong to functionally related gene families.^10,20 This hypothesis thereby provides an explanation for both intraindividual as well as interindividual (i.e., variable expressivity) differences in CAKUT phenotypes. Alternative hypotheses ascribe more importance to epigenetic or environmental influences (e.g., effects of malnutrition, vitamins, or drugs).^3,55 Although epigenetic and environmental factors potentially provide a sufficient explanation for interindividual differences in CAKUT manifestations, they are less applicable in the context of intraindividual (left-right) dissimilarities. Particularly, environmental factors are extremely interesting to discover, because they may allow for a specific prophylaxis of CAKUT. However, despite large studies, to date, only little success has been achieved regarding the search for environmental factors with significant effect on the pathogenesis of CAKUT.^3,56–60

Complex Genetics in CAKUT

Nowadays, modern study designs increasingly yield potential polygenic disease loci for a wide variety of human disease conditions.⁶¹ Despite the complicated etiology of CAKUT, which suggests an interplay of (epi-)genetic and environmental factors,^3,62 to date, there is little evidence for the simultaneous contribution of more than one gene (di-, oligo-, or polygenic effects; i.e., complex genetics) to the pathogenesis of human CAKUT.⁶³ Similar to monogenic research, the identification of polygenic causes of CAKUT is likely complicated by the vast heterogeneity of CAKUT. This results in very few affected patients with shared disease loci essential for the unequivocal identification of novel genetic regions. General awareness of the possibility of complex inheritance patterns in CAKUT is necessary in the genetic evaluation of affected individuals. Polygenic disease hypotheses, however, remain challenging to establish, test, and prove in human CAKUT.

Relevance of Mouse Candidate Genes

One of the reasons why it was hypothesized that a large fraction of human CAKUT may be of monogenic origin was the presence of a large number of monogenic mouse models of CAKUT; >180 monogenic causes of murine CAKUT have been described to date (Supplemental Table 1). These genes constitute promising candidates in the search for (novel) genetic causes of human CAKUT. In fact, many of the 40 established monogenic causes of human CAKUT (e.g., FRAS1, FREM1, FREM2, GRIP1, bone morphogenic protein 4 [BMP4], SIX2, Ret Proto-Oncogene [RET], SALL1, and UPK3A) were initially derived as candidate genes from observations in mouse models of CAKUT and subsequently screened for their prevalence in human disease cohorts.^{21,28,34,42,49} Controlled conditions in the evaluation of genetic mouse models allow for a careful assessment of subtle manifestations of underlying genotypes (e.g., mild phenotypes in single heterozygous carriers of recessive CAKUT genes). Mouse models furthermore provide the opportunity for a controlled evaluation of the contributions of multiple genes (i.e., complex inheritance) to the development of CAKUT. One example is the effect of the loss-of-function of the entire Hox11 paralogous group on renal morphogenesis in mice (complete loss of metanephric kidney induction) in comparison with an isolated loss-of-function of individual or two components of the paralogous group (i.e., Hoxa11, Hoxd11, and Hoxc11; no discernible kidney abnormalities or renal hypoplasia, respectively).⁶⁴

The insights from mouse models, however, do not always directly translate to human genetics. Explanations for this discrepancy include potential species-specific differences in the development of the kidneys and urinary tract, alterations in the required gene dosage during developmental steps, functional compensation by redundant genes, incomplete penetrance, and likely, merely the rarity of human cases with mutations in a particular gene (due to distinct heterogeneity).

Molecular Pathways of Renal Morphogenesis

Despite challenges, mouse models of CAKUT have rendered it possible to identify many developmental pathways as relevant for renal morphogenesis and pathophysiologic events that underlie CAKUT.^3,10,16,48 Most prominently, the glial cell–derived neurotrophic factor (Gdnf)-glial cell–derived neurotrophic factor family receptor α1 (Gfra1)-Ret pathway plays a decisive role in renal morphogenesis in general and specifically, the intricate regulation of the crosstalk between the UB and the MM (Figure 1, C–E).^4,15,62,65 RET is a receptor tyrosine-kinase that is involved in a multitude of developmental processes.^65,66 In the developing kidney, RET is predominantly expressed in nephric duct–derived structures (including UB) (Figure 1).^65,67–69 During renal morphogenesis, RET requires the ligand GDNF as well as the coreceptor GFRA1 for its activation.^{65,66,70–73}

Given the important role of RET signaling for renal morphogenesis, it is not surprising that many of the established causes of murine and human CAKUT (including PAX2, GATA3, EYA1, and SALL1) have been reported to function upstream and downstream of the GDNF-GFRA-RET regulatory pathway (summarized in Supplemental Table 2).^{4,16,62,65,74} Mechanistic details on the RET signaling pathway are beyond the scope of this review and have been reviewed in detail recently.^{4,16,62,65,74,75}

However, various additional morphogenetic pathways have previously been implicated to be essential for proper renal morphogenesis. These signaling pathways include the BMP (TGFβ), NOTCH, Hedgehog-GLI, and FGF-related cascades as well as canonical WNT/β-catenin signaling that converge and partially modulate one another during the development of the kidneys and the urinary tract.^76–78

In this review, we will place special emphasis on two pathways that appear interweaved with GDNF-RET signaling and are not commonly discussed in the context of CAKUT. These are (1) the role of extracellular matrix (ECM) proteins (Supplemental Table 3) at the epithelial-mesenchymal interface between UB and the MM (Figure 2), and (2) vitamin A/retinoic acid (RA) signaling (Figure 3). This latter pathway also constitutes the first molecular pathway potentially amenable to treatment. We furthermore briefly describe (3) the BMP signaling cascade as an example for a well established molecular pathway in renal morphogenesis (Figure 4, Supplemental Table 4). The BMP pathway represents an additional example of a signaling cascade, the genes and proteins of which have recently been identified as genetic causes of human and/or murine CAKUT.^{21,42,79–95}

Figure 2.

ECM proteins cause murine and human CAKUT. Schematic of the interface between the UB and the MM in renal development. Proteins encoded by genes that, if mutated, cause monogenic CAKUT in humans are highlighted in yellow. Red frames indicate proteins encoded by genes mutated in CAKUT in mice. (Left panel) FRAS1 and FREM2 localize in epithelial cells of the UB as transmembrane proteins with intracytoplasmic tail regions.^41,98,103 The interaction with PDZ domains of the intracellular protein GRIP1 is essential for targeting to the basal surface of the UB cells and shedding of FRAS1 from the membrane.^98,107 FREM1 is produced by the MM and secreted into the extracellular space.^106,107 FRAS1, FREM2, and FREM1 assemble at the epithelial-mesenchymal interface to form the FC.¹⁰⁷ Mutations in the genes encoding these proteins cause CAKUT (Supplemental Figure 2). Also, mutations in GRIP1 result in human and murine CAKUT. Nephronectin (NPNT) functions as an adaptor to interconnect the ternary FC with the ITGA8/integrin-β1 heterodimer on the surface of MM cells. ITGA8/integrin-β1 signaling leads to an increased expression of GDNF by the MM, thereby promoting renal morphogenesis.¹⁸³ Mutations in ITGA8 in humans and mice cause CAKUT.⁴⁴ Loss of FC integrity results in a significant decrease in GDNF expression in the MM, thereby hampering the interaction between the UB and the MM and consequentially, impeding renal morphogenesis.¹⁸³ (Right panel) Laminins are cross-shaped ECM molecules that consist of three distinct chains (α, β, and γ).¹⁰¹ Laminins simultaneously interact with cell surface receptors (integrins), HSPG, collagen, and nidogen 1 (see box).¹⁸⁴ Laminins that contain γ1-chains form high-affinity interactions with nidogen 1.¹⁰¹ Fourteen distinct laminin isoforms have been identified to date, of which ten possess the laminin-γ1 chain.^101,185 A homozygous ablation of the nidogen binding site in laminin-γ1 has been shown to result in severe CAKUT phenotypes (including bilateral renal agenesis) in mice.¹⁰¹ Despite its important role in ECM assembly, targeted inactivation of nidogen 1 did not result in a CAKUT phenotype.¹⁰¹ Both nidogen 1 and Laminin are known interactors of HSPG. HPSE is one of the few enzymes with the ability to break down HSPG.¹¹⁵ Knockdown of neither HSPG nor HPSE has been reported to cause CAKUT phenotypes in mice.^{115,120,186,187} However, recessive mutations in HPSE2 have been identified in human patients with urofacial syndrome,^43,93 with a similar phenotype in mice.⁴³ Although HPSE2 has no detectable enzymatic activity itself, it has been shown to function as an endogenous inhibitor of HPSE (Supplemental Figure 3).¹¹⁶ Modified from refs. 107 and 184, with permission.

Figure 3.

RA signaling plays a central role for the development of the kidneys and urinary tract. Genes encoding the proteins highlighted by red frames have been implicated in murine cases of syndromic CAKUT (Supplemental Table 1). Genes highlighted in yellow, if mutated, constitute causes of human CAKUT. Mutations in genes encoding multiple proteins involved in intracellular RA processing and signaling have previously been identified to cause syndromic CAKUT in mice (Retinol Dehydrogenase 10 [Rdh10], Aldehyde Dehydrogenase 1 Family Member A2 [Aldh1a2], and Cytochrome P450 Family 26 [Cyp26], Rxr, Rarα, and Rarβ).^{138,150–152} RDH10¹⁵⁰ catalyzes the reversible reaction from retinol to retinal. Retinal is then converted into the active metabolite all-trans RA via an irreversible reaction catalyzed by the enzyme ALD1A2.¹³⁸ RA then either enters the nucleus or is metabolized via enzymes of the CYP26 family (e.g., CYP26A1)^151,152 in the endoplasmic reticulum. In the nucleus, RA derivatives can serve as a ligand for two receptors: RXR (9-cis-RA and rexinoids) and RAR (all-trans-RA).^123,142 Both receptor proteins have at least three subtypes and isoforms each. RXR and RAR heterodimers bind to retinoic acid–responsive elements (RAREs) of the nuclear DNA.^121,141 RAREs are predominantly located in promoter regions of target genes. The presence of RA recruits coactivators and leads to enhanced binding to RAREs and expression of target genes.^121,141 The transcriptional control of downstream genes is further modified by the binding of additional coregulators (coactivators and corepressors, including NRIP1) to RXR and RAR.^121,141,153 We recently identified mutations in the corepressor NRIP1 (highlighted in yellow) as causing CAKUT in humans.¹⁴⁰ NRIP1 may interact with both RXRs and RARs.¹⁵³

Figure 4.

Members of the BMP signaling cascade with a role in CAKUT. Proteins encoded by genes that, if mutated, cause murine CAKUT are outlined in red, and proteins highlighted in yellow constitute causes of isolated and/or syndromic manifestations of human CAKUT (Supplemental Table 4). BMPs act as ligands for two classes of transmembrane serine-threonine-kinase receptors on the surface of the corresponding effector cells (e.g., bone morphogenic protein receptor type 1A [BMPR1A] and BMPR2).¹⁵⁵ After activated, these receptors initiate intracellular signaling cascades, including canonical SMAD signaling and noncanonical signaling (e.g., via mitogen-activated protein kinase [MAPK]), which eventually result in increased expression of distinct target genes in the cell nucleus.^175,176 The cellular consequences of BMP signaling depend on the location and local concentration of the BMP ligand.¹⁷⁷ The availability of ligands for receptor binding is regulated by diverse intra- and extracellular antagonists and agonists of BMP and includes, for example, Gremlin 1 (GREM1), Follistatin (FST), and bone morphogenic protein binding to the endothelial regulator (BMPER).^178–180 Please note that, BMPR2, although depicted in the figure, has not been implicated in the pathogenesis of CAKUT to date. CRIM1, cysteine-rich transmembrane bone morphogenic protein regulator 1; CTDNEP1, CTD nuclear envelope phosphatase 1; GPC3, Glypican 3.

CAKUT and ECM Proteins

The ECM is a highly complex, three-dimensional network of proteins and proteoglycans.^96,97 It is an essential component of all tissues and synthesized by embryonic cells starting at the earliest stages of development.^96,97 Over the past decades, the understanding of the functional roles of the ECM has changed dramatically. Although the ECM was initially believed to merely provide structural support,^97–99 the current understanding goes far beyond simple adhesion and space-filling properties and includes regulatory roles in cell-cell and cell matrix interactions.⁹⁶

The pathophysiologic inter-relation of ECM components and renal morphogenesis has long been known from knockout animal models that display phenotypic features of the CAKUT spectrum (summarized in Supplemental Table 3). The relevance of ECM proteins for isolated human CAKUT, however, first came to attention when mutations in genes encoding members of the Fraser complex (FC) as well as the associated protein integrin-α8 (ITGA8) were identified in a significant proportion of patients with isolated human CAKUT phenotypes.^42,44 Figure 2 depicts selected ECM components that have previously been shown to be causative of human (highlighted in yellow in Figure 2) as well as murine CAKUT (outlined in red in Figure 2).^{41–44,93,100–105}

The FC is a ternary protein complex consisting of FRAS1, FREM2, and FREM1 (Figure 2).¹⁰⁶ Although not participating in the complex formation itself, GRIP1 is a cytosolic protein essential for trafficking and proper targeting of the transmembrane protein FRAS1 to the basolateral surface of the UB cells.¹⁰⁴ After they are assembled at the epithelial-mesenchymal interface, the members of the FC interact with nephronectin (NPNT), which serves as an adaptor for ITGA8 and integrin-β1 that are expressed by the cells of the MM (Figure 2).^105,107,108 ITGA8 is an important upstream activator of GDNF, which, via interaction with RET, promotes renal organogenesis.^105,107

Genes encoding members of the FC have previously only been known to be causative of the severe syndromic phenotype Fraser syndrome, a rare syndromic condition with phenotypic features including CAKUT as well as cryptophthalmos and cutaneous syndactyly (FRAS1, FREM1, and GRIP1).^{41,103,104,109,110} FREM1 mutations are also an established genetic cause for the phenotypically related conditions Manitoba-oculo-tricho-anal syndrome (OMIM 248450) and Bifid nose with or without anorectal and renal anomalies syndrome (OMIM 608990) (Supplemental Figures 1 and 2).¹¹¹

Interestingly, the disease-causing mutations in FRAS1, FREM1, FREM2, and GRIP1 (Figure 2) in patients with FS or Manitoba-oculo-tricho-anal syndrome are almost exclusively recessive truncating variants^{41,42,103,110–112} (Supplemental Figure 2). Truncating mutations result in an abrogation and thereby, loss-of-function of the corresponding protein. Despite the interaction and a structural similarity, members of the FC do not exhibit functional redundancy, and an abrogation of any member of the FC and/or GRIP1 results in a failure of the complex to assemble.^98,104,107 A consecutive epithelial-mesenchymal detachment is considered causative for distinct manifestations of FS (e.g., skin blebbing resulting in cutaneous syndactyly and cryptophthalmos).⁴¹

Conversely, we recently found that missense mutations in FRAS1, FREM1, FREM2, or GRIP1 result in isolated CAKUT phenotypes (Figure 2).⁴² These findings thereby indicate, that, in the case of the FC genes, the disease manifestation can vary from severe syndromic phenotypes on the basis of truncating mutations to isolated CAKUT on the basis of hypomorphic mutants (e.g., missense) (Supplemental Figure 2).

Another subgroup of genes that cause (predominantly murine) CAKUT, if mutated, coalesces around the proteoglycan component of the ECM (Supplemental Figure 3, Supplemental Table 3). Proteoglycans consist of a protein core coupled to unique, variably sulfated glycosaminoglycan chains (Supplemental Figure 3).^97,113 Although there are various classes of glycosaminoglycan chains, heparan sulfate (HS) has previously been implicated to have the largest functional relevance within the developing kidney.¹¹³

The synthesis of heparan-sulfate proteoglycans (HSPGs) is very complicated and involves a large variety of enzymes (Supplemental Figure 3).¹¹⁴ In contrast, the breakdown is rather simple and predominantly carried out by the enzyme heparanase (HPSE) (Figure 2, Supplemental Figure 3).¹¹⁵ The protein HPSE2 is an inactive enzyme, but it has been reported to function as an endogenous regulator of HPSE activity.¹¹⁶ Recessive mutations in HPSE2 have been identified in humans and mice with urofacial syndrome (OMIM 236730), which includes manifestations from the CAKUT spectrum within its syndromic phenotype.^43,93,100

Overall, the multitude of murine and human disease phenotypes that occur as result of disturbances in HSPG biosynthesis provides important insight into the mechanisms involved (Supplemental Figure 3, Supplemental Table 3).^117,118 However, although for some HS-relevant enzymes, there are strong loss-of-function phenotypes in animal models, mouse models for other HS-related ECM components (e.g., mouse models with a knockout of HPSE function) do not result in overt defects, thereby suggesting a redundancy of certain HS enzymes and HSPGs in the ECM that allow for a functional compensation.^96,119,120 This hypothesis, however, remains to be validated by additional functional studies in the future.

CAKUT and RA Signaling

Vitamin A is a fat-soluble compound that is derived from dietary sources.¹²¹ Vitamin A and its active metabolite, RA,^121–123 play important roles in the regulation of a wide range of biologic processes, including differentiation, proliferation, and apoptosis.^{121,123–130} While exerting essential roles in adult tissues (e.g., vision, fertility, and immune response),^131–133 vitamin A is also a well known teratogen.^{121,123,131,134} Interestingly, both the deficiency and the excess of vitamin A during pregnancy cause similar forms of an embryopathy known as vitamin A–deficient (VAD) syndrome.^121,131,135 VAD syndrome affects many structures, including the ocular, cardiac, respiratory, and urinary systems, and is a cause of CAKUT.^123,135,136 It is, therefore, recommended to carefully monitor vitamin A levels during pregnancy.¹²¹

Over the past decades, efforts, mainly led by Mendelsohn et al.,^{66,122,130,137–139} have broadened the understanding of the molecular mechanisms that underlie the well established interplay of vitamin A/RA signaling and renal morphogenesis in mice. Research by Chia et al.⁶⁶ and Batourina et al.¹²² (among others) convincingly showed that RA is required for two key events during embryonic development of the urinary tract: the insertion of the nephric duct into the cloaca (approximately E9.5) (Figure 1B) as well as the branching morphogenesis of the UB (approximately E11.5–E15.5) (Figure 1, D and E). In both scenarios, the mechanism by which vitamin A/RA mainly exerts its effects seems to be due to control of expression levels of a key player of renal morphogenesis, Ret.^66,122

Although the pathophysiologic relevance of vitamin A for embryologic development, including processes of renal morphogenesis, has long been known, very recently for the first time, a heterozygous mutation in a gene related to vitamin A function (Nuclear Receptor Interacting Protein 1 [NRIP1]) was discovered in one large kindred with isolated human CAKUT.¹⁴⁰

RA plays an important role in the transcriptional regulation of developmentally relevant target genes.^66,121 Retinoid X Receptor-α (RXR) and Retinoic Acid Receptor (RAR; are nuclear receptors that bind RA each has three subtypes, α, β, and γ, with different isoforms).^{121,141–143} RXR (5′) and RAR (3′) assemble to heterodimers and bind to RA-responsive elements in the DNA (Figure 3),^{121,123,141,142,144,145} thereby controlling transcription of functionally diverse target genes.^121,141 The specific subtype and combination of RXRs-RARs thereby most likely permit a certain tissue selectivity, while at the same time, sustaining the pleiotropy of RA-mediated effects.¹²³ RXR and RAR can additionally bind coregulators (coactivators and corepressors), which adds another level of complexity to the transcriptional regulation of RA-dependent genes in an RA-dependent manner.^{121,131,141,142,146–149} A potential contribution to the tissue specificity of RA-mediated effects via tissue-specific expression patterns of coregulators is likely.

Mutations in genes encoding enzymes that are involved in intracellular processing of RA, such as Retinol Dehydrogenase 10,¹⁵⁰ Aldehyde Dehydrogenase 1 Family Member A2,¹³⁸ and Cytochrome P450 Family 26 Subfamily A Member 1,^151,152 as well as double mutants of specific isoforms within the RXR (Rxrα) and RAR family of proteins (Rarα and Rarβ)¹²³ (Figure 3) have previously been shown to result in multisystemic phenotypes in mice that resemble VAD syndrome and present manifestations from within the CAKUT spectrum.

NRIP1 is a member of a group of ligand-inducible transcription factors, and it is involved in the coregulation of variety of biologic processes (e.g., estrogen and thyroid hormone signaling), including RA-mediated pathways.^153,154 NRIP1 has been shown to have the capacity to interact with both RAR and RXR (Figure 3).¹⁵³ Despite the large variety of biologic functions of NRIP1, in the examined family with seven affected individuals, the identified mutation resulted in an isolated CAKUT phenotype without any extrarenal manifestations.¹⁴⁰

To the knowledge of the authors, this is the first time that a protein involved in RA signaling did not result in a syndromic CAKUT phenotype but rather, resulted in a selective CAKUT phenotype. This could potentially point toward a role for this coregulator in a select group of tissues as opposed to Retinol Dehydrogenase 10, Aldehyde Dehydrogenase 1 Family Member A2, and Cytochrome P450 Family 26 Subfamily A Member 1, RARs, and RXRs, which exhibit very broad expression patterns. Although this would have to be further elucidated, the possibility exists that, in the future, more regulators of RA-dependent transcriptional regulation may be identified in patients with cases with isolated features from within the syndromic VAD phenotype.

NRIP1 mutations as a cause of CAKUT are of particular interest, because they represent the first form of CAKUT that could be amenable to prevention treatment. It will have to be tested whether control of RA metabolism may be applicable in pregnant women with a family history of NRIP1-related cases of CAKUT. NRIP1 mutations as genetic cause of human CAKUT are very rare, however, with only one case of familial CAKUT to date.¹⁴⁰

CAKUT and Signaling through BMPs

Bone morphogenic proteins (BMPs) constitute the largest family within the TGFβ superfamily of molecules.¹⁵⁵ BMPs were originally identified as bone-derived proteins capable of inducing ectopic bone formation in vivo.^156,157 Nowadays, the BMP pathway is renowned for its importance in general embryogenesis, with diverse roles depending on the spatial and temporal context.⁷⁸ A large number of in vitro and in vivo studies underscore a crucial role of BMP signaling for renal morphogenesis in animal models (Figure 4, Supplemental Table 4).^{79–92,158–160} Also, the development of the human kidney and urinary tract seems to depend on intact BMP signaling. This is indicated by the finding of mutations in genes inter-related with BMP signaling in patients with isolated (BMP4²¹ and GREM1⁴²) or syndromic (BMPER,⁹⁴ BMP4,⁹⁵ and GPC3¹⁶¹) manifestations of human CAKUT.

Details of the importance of BMP signaling for renal morphogenesis have been previously reviewed in detail^{78,155,162–164} and are continually being revealed. Experimental evidence suggests that the contribution of BMPs to renal morphogenesis is very complex. Research in the field of BMP signaling is generally challenged by the early embryonic lethality that often results from loss-of-function mutations in animal models of BMP-related genes (Supplemental Table 4). This complicates the investigation of respective contributions to the development of the kidneys and the urinary tract.^85,165–167 Spatiotemporal differences in expression patterns of different components of the BMP signaling cascade indicate that proteins from the BMP family perform various functions for both UB and mesenchymal cells during development of the murine kidney.^{79,84,89,168–171} Tight regulation of these proteins seems to be critical in these processes. This is, for instance, exemplified by the finding of CAKUT phenotypes resulting from both mice with null mutations in distinct BMP genes as well as mice with a lack of inhibition of the respective BMP.^78,84 However, it was shown that BMPs may be able to substitute one another during murine renal morphogenesis as previously shown for Bmp7 and Bmp4.¹⁷²

Mechanisms of BMP signaling in the context of renal development are briefly outlined in Figure 4. Figure 4 depicts components of the signaling cascade that have previously been reported to result in manifestations of murine CAKUT if mutated (outlined in red in Figure 4).^{79–92,158–160} Some of these genes/proteins have additionally been identified in human forms of isolated (BMP4²¹ and GREM1⁴²) or syndromic (BMPER,⁹⁴ GPC3,¹⁶¹ and BMP4⁹⁵) CAKUT (highlighted in yellow in Figure 4).

In general, BMP ligands induce cellular responses by forming complexes with two major types of diverse membrane-bound serine-threonine kinase receptors: type 1 BMP receptors and type 2 BMP receptors.¹⁵⁵ Both types 1 and 2 receptors are required for subsequent transduction of the BMP signal.^173,174 The cellular responses, however, seem to be defined by the type 1 receptor.¹⁵⁵ Downstream of ligand-receptor binding, BMPs activate either a canonical or a noncanonical signaling cascade through phosphorylation of downstream effectors. The canonical BMP pathway transmits signals via Smad proteins.^175,176 Noncanonical BMP signaling includes the activation of the mitogen-activated protein kinase family of signaling molecules.¹⁷⁵ Both signaling cascades eventually result in the increased transcription of corresponding target genes in the nucleus. The location and local concentration of the BMP ligand hereby determine the exact cellular consequences of BMP signaling.¹⁷⁷ Intra- and extracellular agonists (e.g., BMPER) and antagonists (e.g., GREM1 and FST) of BMP tightly regulate the level of ligand-receptor interaction at the ligand or (more rarely) the receptor level,^178–180 and they play an important role in renal development.

Conclusion

In conclusion, a long-standing hypothesis has recently found strong confirmation; it stated that CAKUT, to a large part, is caused by single-gene mutations in a high number of different monogenic CAKUT genes, involving different genes in different patients.

Whole-exome sequencing (WES) has facilitated discovery of these genes. Even the subhypotheses found confirmation, stating that monogenic CAKUT genes would often represent candidate genes from mouse models of CAKUT, renal developmental genes, or genes known to cause human clinical syndromes that have a CAKUT component.

The fraction of approximately 17% of patients with CAKUT, in whom a monogenic cause can be detected by WES, is lower than that in other causes of early-onset CKD, being approximately 30% in steroid-resistant nephrotic syndrome¹⁸¹ and 70% in renal cystic diseases.¹⁸² Monogenic CAKUT genes are more difficult to identify due to the high contribution of autosomal dominant renal developmental genes, which often exhibit features of incomplete penetrance (lack of genotype-phenotype correlation) and variable expressivity (differing organ involvement). Previous experience, however, strongly suggests that, before complex genetic mechanisms have to be evoked, such as polygenic inheritance or epigenetic mechanisms, hundreds of monogenic causative genes for CAKUT genes will still be identified.

In this context, the ability to represent a patient’s exact mutation in an animal model of CAKUT (e.g., mice or Xenopus) using CRISPR/Cas knockin techniques has strong potential to conclusively define genotype-phenotype relations for patients with CAKUT.

Identification of monogenic causes of CAKUT by WES will rapidly become common practice in pediatric nephrology and urology clinics due to reduction in cost for WES and continuous improvement of algorithms that define deleteriousness of mutant alleles.

Discovery of additional monogenic CAKUT genes will strongly aide our understanding of the complex pathogenesis of CAKUT. In addition, molecular genetic diagnostics using WES may offer important advances to clinical management for patients with CAKUT, who represent about 50% of all patients with ESRD before age 25 years old, because causative mutations will soon allow for prognostic conclusions, such as outcomes of reconstructive surgery, rate of progression into renal failure, and the likelihood for the development of extrarenal complications.

Disclosures

None.