Mammalian kidneys fulfill a tremendous job by filtrating enormous amounts of blood without loss of essential components such as proteins in the urine. The glomerular filtration barrier consisting of three major layers [podocyte foot processes and the slit diaphragm, the glomerular basement membrane (GBM), and glomerular endothelial cells (GEC)] hereby ensures that proteins larger than 60 kD remain within the blood vessels. Smaller filtrated proteins are subsequently largely re-absorbed within the tubular system. Failure of this fine-tuned system results in proteinuria. Loss of large amounts of protein in the urine can result in nephrotic syndrome, characterized by proteinuria of more than >3.5 g per 1.73 m² body surface area per day, hypoproteinemia (serum albumin is reduced to <2.5 g/dl), hyperlipoproteinemia, and edema. Acute complications include thrombosis, pleural effusions and ascites. Over time, growth restriction in children as well as progressive loss of podocytes with subsequent decline in renal function become evident (1). Two main clinical entities of nephrotic syndrome are distinguished: steroid sensitive nephrotic syndrome (SSNS) and steroid resistant nephrotic syndrome (SRNS). Steroid resistant nephrotic syndrome is defined as persistent nephrotic-range proteinuria after 4 weeks of oral prednisone at 60 mg/m² per day (2). Steroid resistant nephrotic syndrome is a rare condition, however represents one of the most common causes of childhood-onset kidney failure, occurring as frequently as 1:10,000 in some populations (3). Histopathological findings are most often focal segmental glomerulosclerosis (FSGS) (4). Disease-causing variants in over 60 genes have been described to date to cause monogenic forms of SRNS, explaining approximately one-third of all SRNS cases with an age of onset <25 years (5). Many disease-associated genes encode for proteins essential for proper podocyte function. Most commonly, disease-causing variants affecting the slit diaphragm proteins Nephrin (NPHS1) and Podocin (NPHS2) or the transcription factor WT1 or LAMB2 are identified (6). Hereditary SRNS usually does not respond to immunosuppressive treatment and, in contrast to SSNS, has a poor prognosis with regards to renal survival: 10 years after onset of proteinuria, SSNS patients experience end-stage kidney failure (ESKF) in <5% of cases while ESKF occurs in 50% of individuals with sporadic SRNS and 80% of genetically proven SRNS cases (4, 7). However, disease re-occurrence after kidney transplantation is less likely in most hereditary SRNS cases compared to sporadic SRNS cases (8). Identification of an underlying genetic cause in SRNS individuals therefore has direct therapeutic consequences. Childhood-onset SRNS occurs more frequently in consanguineous populations as it is often autosomal recessively inherited (9, 10).

In this study, we describe the genetic findings using exome sequencing (ES) in 29 Iranian children with SRNS from consanguineous parents with as well as one Iranian parent pair where only insufficient DNA of the affected child was available.

Molecular genetic diagnostics were undertaken for 30 Iranian unrelated consanguineous families with at least one child clinically diagnosed with SRNS. In one family (sample 3), insufficient DNA of the affected child was available for genetic testing by ES, therefore both parents were tested by ES instead, followed by Sanger sequencing of the child.

In total, we identified an underlying genetic cause in 21/30 families and a variant of uncertain significance (VUS) in one family (diagnostic yield: 73%) (Table 1). All identified disease-causing variants locate to genes known to cause monogenic SRNS when dysfunctional with the exception of one case carrying a variant in a nephronophthisis gene (NPHP1) instead. The gene most frequently affected was NPHS1 (MIM# 602716; nine families: 38% of solved families and 30% of all families), followed by NPHS2 (MIM# 604766; six families: 29% of solved families, 20% of all families). Two cases were found to be caused by WT1 variants (MIM# 607102; 10% of solved families, 7% of all families), while variants in NUP205 (MIM# 614352), COQ6 (MIM# 614647), ARHGDIA (MIM# 601925), and SGPL1 (MIM# 603729) were identified in single families (5% of solved families, 3% of all families). In addition, we identified a disease-causing deletion in NPHP1 [MIM# 607100; gene previously associated with nephronophthisis (NPHP)] in one family (5% of solved families, 3% of all families; Figure 1).

In total, we detected seven variants not previously reported in ClinVar or HGMD, including four different NPHS1 variants (three variants classified as disease-causing and one VUS) and one novel variant in ARHGDIA, SGPL1, and NPHP1 each (all classified as disease-causing). Amongst the known SRNS disease alleles, we identified NPHS2 c.353C>T (p.Pro118Leu) as well as NPHS1 c.3523_3524del (p.Leu1175Valfs^*2) in two unrelated families each. Three of the four novel NPHS1 variants as well as variants in ARHGDIA, SGPL1, and NPHP1 all represent novel disease-causing variants as defined by ACMG criteria.

All identified variants except the two variants identified in WT1 were present in a homozygous state in affected children. Parents were heterozygous carriers of the respective variant. Nine variants represent stop or frameshift variants in NPHS1, NPHS2, and NPHP1 (45%), one allele was a −1 canonical splice site variant in ARHGDIA, one a +2 splice site variant in NPHS1 classified as likely pathogenic according to ACMG criteria, one small deletion causing a start loss in SGPL1, and the remaining variants were missense variants. All but 1 variant (NPHS1 c.1757G>A, p.Arg586Lys, rs730880174; classified as VUS) were classified as disease-causing according to the recommendations of ACMG. Interestingly, rs730880174 in NPHS1 affecting likewise amino acid 586 but changing it into a glycine instead of a lysine has been reported as pathogenic in ClinVar. Likewise, NPHP1 c.386del (p.Ser129Metfs^*53) affects the same amino acid position as rs757139057 (c.385_386del), reported pathogenic in ClinVar.

Both heterozygous missense variants we identified in WT1 have been reported as pathogenic before with WT1 c.1315C>T, p.Arg439Cys identified in Denys-Drash syndrome and Meacham syndrome as well as in a Japanese patient with SRNS (15). Variant information for all identified variants and localization of variants in COQ6, NUP205, ARHGDIA, and SGPL1 on protein level are shown in Table 1 and Figures 1, 2. A graphical summary of subcellular localisations and functions of proteins encoded for by genes in which we detected disease-causing variants in this study is shown in Figure 3.

We identified disease-causing variants according to ACMG criteria in genes previously reported to cause childhood SRNS in 21 out of 30 families (70%) and a VUS in one out of 30 families (3%), exceeding the diagnostic yield reported in other larger NGS studies by far (4, 16–18) (Table 2). However, in contrast to outbred study population described in those studies, all of the families referred to us for genetic diagnostics were consanguineous. In addition, it has been previously shown that the diagnostic yield in SRNS is highest in children with congenital nephrotic syndrome and significantly lower with increasing age at first manifestation during the first 6 years of life (18). In line with this, average age at first manifestation in our cohort was 17.5 months and all of our genetically solved cases were below 5 years of age when showing first SRNS symptoms while two of the genetically unsolved cases were 7 years old (Tables 1, 2).

Congenital cases are often caused by disease-causing variants in NPHS1 and we identified such variants in two cases as well as disease-causing variants in SGPL1 or ARHGDIA in one case, respectively. Age of onset was overall lower in NPHS1 cases compared to NPHS2 cases, as expected. In total, we most frequently detected biallelic variants in NPHS1 (38% of solved families and 30% of all families) followed by NPHS2 (29% of solved families and 20% of all families), in accordance with previous findings that NPHS1 and NPHS2 variants are amongst the most common underlying genetic causes in childhood SRNS in non-Iranian populations (17). This is also in line with findings reported by Basiratnia et al. (20) who identified disease-causing variants in NPHS2 in 15 out of 49 Iranian SRNS cases (31%) and 57% of familiar Iranian SRNS cases. Three of these 15 cases (20%) carried the NPHPS2 p.Pro118Leu variant homozygously while we detected this variant in 2 of 6 cases (30% of all our NPHS2 cases), indicating it represents a fairly common Iranian disease allele. We did not detect the previously described Iranian NPHS2 founder variant p.Arg238Ser reported by Basiratnia et al. (20). Interestingly, the proportion of childhood SRNS cases resulting from biallelic disease-causing NPHS2 variants seems fairly similar across many populations such as European populations [12–19% in sporadic cases (21, 22), 26% in familiar cases (23)], and 15% in Indian cases (24), while disease-causing NPHS2 variants are less frequently identified in Chinese, Japanese, or South African probands with SRNS (3–8%) (19, 25–29).

Although we were able to genetically solve 25% of cases by starting with Sanger sequencing of NPHS2, in the days of broad availability of NGS, performing NGS panel or ES as primary diagnostic step seems more efficient, given the overall high genetic variability in SRNS. Using ES in NPHS2 negative cases, we identified multiple biallelic disease-causing NPHS1 variants, two heterozygous disease-causing WT1 variants as well as biallelic disease-causing variants in COQ6, ARHGDIA, NUP205, and SGPL. Disease-causing variants in COQ6 represent a very rare cause of SRNS with until completion of this study less than 20 disease-causing alleles reported to date in the literature (30, 31) (Figure 2). Very recently, Drovandi et al. reported additional COQ6 cases as part of a large cohort of 116 individuals with primary coenzyme Q10 deficiency in which CoQ10 replacement had beneficial effects regarding disease progression (32, 33). Disease-causing variants in ARHGDIA likewise represent a rare cause of SRNS (Figure 2). ARHGDIA (Rho GDP-Dissociation Inhibitor Alpha) is expressed in podocytes playing an important role in maintaining Rho-GTPases in their inactive state and loss of function subsequently disturbs the actin-cytoskeleton arrangement, resulting in early-onset proteinuria and progressive loss of renal function in humans and mice (34–36). In line with findings in cases previously reported in the literature, our case presented with congenital nephrotic syndrome, confirming a crucial role of ARHGDIA for podocyte integrity. We further identified a homozygous small deletion encompassing exon 2 (the first coding exon) resulting in a start-loss in SGPL1 in a case with syndromic congenital SRNS. SGPL1 encodes sphingosine-1-phosphate lyase-1, a ubiquitously expressed enzyme located at the endoplasmic reticulum. Loss of function results in syndromic SRNS and adrenal insufficiency in humans with 20 disease alleles published to date (37, 38) (Figure 2). We also detected a homozygous NUP205 missense variant, previously reported as disease-causing, in a case with disease onset at 12 months of age. NUP205 encodes for a nuclear pore complex protein and a different homozygous disease-causing missense variant in NUP205 has been previously reported in a Turkish sib-pair affected by SRNS (39). Additionally, NUP205 loss of function has been previously associated with left-right patterning defects in humans (40) (Figure 2) as well as ciliogenesis defects in a frog knockdown model system (41). Our case did not exhibit any overt laterality disturbances. A summary of subcellular localisations and functions of proteins encoded for by genes in which variants have been identified in this cohort is shown in Figure 3.

Last, we identified a phenocopy in a case presenting clinically with SRNS where we identified a homozygous disease-causing NPHP1 frameshift variant. NPHP1 loss of function is a frequent cause of NPHP, with a deletion of NPHP1 representing the most frequent disease allele (42). Interestingly, Kirsty et al. have previously published a Filipino family with nephrotic syndrome but genetic diagnostics revealing disease-causing variants in NPHP4 associated with NPHP (43). While proteinuria occurs frequently in NPHP, usually it is not of the nephrotic range and instead of a glomerular protein pattern observed in SRNS, proteinuria in NPHP rather shows a tubular pattern. For our case, unfortunately no clinical reports stating what type of proteins were lost in the urine are available. It therefore remains unclear if the individual suffered from tubular or glomerular proteinuria. We cannot exclude that in addition to the genetically established NPHP, the individual independently also suffered from SRNS of another origin, including a non-monogenic form of SRNS.

Overall, ES was very efficient in our cohort of consanguineous early-onset SRNS cases to establish a genetic diagnosis. Nevertheless, several cases remain unsolved, putatively due to variants outside of the coding regions, warranting genome sequencing (with RNA sequencing) or due to the presence of a polygenic or non-genetic mechanism.

The datasets for this article are not publicly available due to concerns regarding participant/patient anonymity. Requests to access the datasets should be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Freiburg University ethics commission, votum 122/20 and samples processed under the Radboudumc Diagnostics Innovation programme (CMO2006-048) to establish a genetic diagnosis. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

MN, AR, EK, SS, TB, AS, AR, PN, SB, MK, and AA recruited the probands and/or were involved in their clinical care. MN, KMR, TM, RB, IS, SS, AK, RB, and MS conducted data analysis. MN and MS conceived the study. MS drafted the manuscript. MN and JH supervised the study. All authors read and approved the final version of the manuscript.

IS acknowledges funding from the Freiburg University Medical Faculty Hospital Berta-Ottenstein Clinical fellowship programme. MS acknowledges funding form the European Research Council (ERC): ERC starting grant TREATCilia (grant agreement no. 716344) and MS and AK received funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—Project-ID 431984000—SFB 1453 (CRC Nephgen). MS, AK, and RB acknowledge funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—Excellence Initiative CIBSS—EXC-2189—Project ID 390939984.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.