The present study (
n = 69) included isolated and syndromic CAKUT cases representing the full spectrum of CAKUT phenotypes. Although, no pathogenic or likely pathogenic variants were identified, it is important to report VUS as these variants could be reclassified as pathogenic or likely pathogenic if additional patients with same phenotype and same variants are reported in future. After determining the pathogenic score of the VUS variants using Karbassiet.al. Scoring system, 10 VUS variants which are potentially damaging were identified in 5 genes (
FRAS1, TNXB,
FREM2,
SIX5 and CHD1L). Interestingly two cases with different CAKUT phenotypes (case 51 with PUV and case 71 with L Duplex Ureter system) showed single gene variant (
TNXB) gene at the same position although the variants reported were different (c.857delA and c.856_857insG respectively). Previous Genome Wide Linkage Study (GWLS) and Whole Exome Sequencing (WES) study in a large kindred with VUR identified heterozygous pathogenic variants in
TNXB as a cause of dominantly inherited VUR and joint hypermobility [
25]. In the present study variant in
TNXB gene was observed in duplex ureter system phenotype with reflux in lower moiety. The patient with
CHD1L variant showed a similar phenotype of PUV that has been reported previously in a study conducted by Hwang et al. Authors identified variant in
CHD1L gene in five unrelated families (0.76%) out of 650 families of diverse ethnic background. The phenotypes of individuals carrying the
CHD1L gene variant is very heterogenous: bilateral kidney malrotation, right renal dysplasia, horseshoe kidneys, right duplex collecting system, right MCDK, left UVJO and PUV [
26]. Variants in two different genes may be due to the fact that kidney development is complex involving the interaction of distinct pathways. It is likely that defects in different genes involved in these pathways may result in similar phenotypes [
11,
27‐
31]. These two variants identified in
CHD1L and
TNXB met most of the criteria of pathogenicity but due to lack of data for the loss of function, these variants were further classified as variants of unknown significance. The findings in the current study supports the observations from previous studies that phenotypic classification alone may not be useful to predict the primary genetic defect in CAKUT [
32]. For
TNXB and
FRAS1 gene variants, the observations were similar to that of previous studies; except for PUJO (
FRAS1) and Horse-shoe kidney (
TNXB), which were observed only in our cohort [
25,
26,
33‐
35]. Three novel missense variants were identified in
PAX2, RET and
SALL1 gene. Variant in
PAX2 [p.Ser305Leu] was identified in a child (CT 14) with bilateral VUR and unilateral hypo dysplastic kidney.
PAX2 belonging to the GDNF-RET signaling pathway plays a major role in morphogenesis and development of kidney and urinary tract [
36].
PAX2 related disorder have a variable clinical presentation with renal and ophthalmological abnormalities (60–70%), and sometimes with other abnormalities like hearing loss. On screening, child with the novel missense variant had no abnormality in the eye. In a Japanese cohort, of the seven patients with deleterious variants in
PAX2 gene who progressed to renal failure in childhood, six had truncating variants [
37]. Like
PAX2,
SALL1 also belongs to GDNF-RET signaling pathway. A missense variant (p.Phe447Tyr) was identified in a child (CT 88) with bilateral VUR. Although heterozygous variants of
SALL1 cause Townes-Brocks syndrome, comprising of facial dysmorphism, limb defects, kidney and urinary tract abnormalities and anorectal malformation, isolated kidney and urinary tract abnormalities have also been reported [
38]. Targeted sequencing of 7 known CAKUT-causing dominant genes in 749 individuals identified novel missense variants in
SALL1 in 6 patients, of which 4 had unilateral or bilateral VUR [
26]. In the same study, deleterious variants in
RET gene was identified in three families with different abnormalities of kidney and urinary tract suggesting phenotypic heterogeneity. In this study,
RET variant (p.Gly533Ser) was identified in child (CT 73) with unilateral vesico-ureteric junction obstruction. Nevertheless, information about frequency of
RET as one of the gene associated with CAKUT is conflicting [
39,
40].
The diagnostic yield using targeted panel sequencing in CAKUT is variable with yield not impacted by number of genes included in the panel. Some studies reporting low diagnostic range (2–15%) [
1,
26,
41], while in few other studies, pathogenic variants were identified in 6–20% of patients with CAKUT by analyzing 5–30 genes [
11,
34,
35,
42‐
45].Study in a Dutch population also reported low diagnostic rate (3%), where they sequenced 208 candidate genes in 453 subjects representing the full spectrum of CAKUT phenotypes [
44]. However in another cohort, a higher molecular diagnosis rate (18%) was reported in a panel sequencing study of 330 candidate genes in a cohort of 204 unrelated CAKUT patients [
43]. In the current study many variants were classified as VUS, mainly as they were novel, lacking functional evidence, and also due to lack of evidence for segregation with the disease. In addition, a major barrier to variant calling is the absence of database specific to Indian population for allele frequencies. Larger multi-center study will also help in identifying common variants in children with CAKUT.
These findings indicate that the major limitation in variant classification is absence of functional data and genotype data from diverse ethnic background. With increasing NGS sequencing in clinical setting, it is likely that these VUS may be later classified as pathogenic or likely pathogenic. Previous studies reported
PAX2 and
HNF1B as the most frequently mutated genes in the European and American population with CAKUT. These studies predominantly consisted of children with renal hypodysplasia (RHD) [
46‐
50]. This is in contrast to the present study, wherein children were non-syndromic (92%) with lower urinary tract obstruction phenotypes -VUR (33.3%) and PUV (39.1%).
HNF1B pathogenic variants are more frequently associated with hypo dysplasia and cystic kidneys and are rare in isolated lower urinary tract defects such as PUV and VUR, which were the most common diagnosis in the present study [
46,
50]. Hence it is not surprising that we did not find any deleterious variants in
PAX2 or
HNF1B.
The difference in the diagnostic yield in current study when compared to other studies could be attributed due to the approach using targeted panel as well as due to differences in cohort size, type of CAKUT that constituted the cohort and, the family history of patients, and also the genes included for screening. For example, high diagnostic yield was seen in cohort of severe CAKUT phenotypes, syndromic cases, patients from consanguineous families and familial cases [
51,
52].This also addresses the need for familial testing as it helps in classification of variant into pathogenic or benign. But due to the unavailability of the samples, sequencing or testing of parents were not performed. It is possible that the deleterious variants in genes in targeted panel are rare in Indian population (less than 1% of observed cases with CAKUT) or larger cohorts screening is essential to confirm the observation. Besides monogenic cause, three large cohort studies identified that 4.5–16.6% of patients with CAKUT carry pathogenic CNV [
9,
10,
22,
23,
42,
53‐
55] while targeted panel can identify single exon CNVs, multi exon CNVs are better identified using Whole Exome/genome sequencing panels. It is likely that CNVs in gene(s) or pathogenic variants in the noncoding regions not targeted by our customized panel could be responsible for the disease in the cohort. It is also known that epigenetic factors and non-genetic environmental factors also contribute to the occurrence of CAKUT [
56]. This underlines that yet to be uncovered genetic complexity and environmental factors may be involved in the development of the disease. Hence, using other genetic testing approaches like microarray and clinical or whole exome sequencing may help decipher the genetic spectrum of CAKUT. WES studies reported diagnostic yield of 11–14% in CAKUT patients and identified novel gene variants [
1,
56‐
60] which is higher than that obtained with targeted panel [
8,
53‐
55]. Exome sequencing have distinct advantages compared to targeted panel since it covers a larger number of genes as well identify other types of variants like CNVs but is more expensive compared to targeted sequencing. Targeted panels are phenotype-driven panels, featuring few hundred genes associated with kidney diseases. Its use in CAKUT may be limited due to significant phenotypic heterogeneity of CAKUT resulting in exclusion of genes associated with CAKUT of particular phenotype. It will be more useful to use exome sequencing based approach as many NGS-based targeted gene sets are built on an exome backbone. This allows the laboratory to use the same process for library preparation and sequencing leading to lesser cost. Such an approach not only facilitates targeted gene analysis but also in searching variants in remaining genes of clinical exome if targeted gene analysis fails to identify causal variants The low yield obtained using a targeted panel in this study and the need to frequently update the targeted panel for the newer discovered/novel CAKUT associated genes makes clinical exome sequencing a more favorable approach.