Introduction

Nephrotic syndrome (NS) is a renal disease characterized by heavy proteinuria, hypoalbuminemia, edema and hyperlipidemia.1 Urinary losses of macromolecules such as albumin reflect a dysfunction of the highly permselective glomerular filtration barrier.2 The glomerular filtration barrier structure consists of podocyte foot-process, glomerular basement membrane, endothelial fenestration and the slit diaphragm. The identification of mutations leading to defects in proteins highly expressed in the podocyte and slit diaphragm has helped to unravel the basis of glomerular filtration barrier physiology and pathophysiology.3 Kidney biopsies in NS patients may show nonspecific changes such as minimal change, as well as focal segmental glomerulosclerosis (FSGS) and diffuse mesangial sclerosis.4 A molecular genetic diagnosis is important for making treatment decisions including suitability for renal transplantation and to enable screening of other family members at risk of disease.

NS may be given a series of descriptive labels dependent on their age of presentation. Congenital NS (CNS) manifests in utero or during the first 3 months of life;5 infantile NS has an onset between 3 months and 1 year of age5 and childhood steroid-resistant NS (SRNS) may be defined as no urinary remission within 4 weeks of prednisone therapy 60 mg m–2 day–1.6 From a molecular genetics standpoint, NS may be grouped into several types, which are briefly reviewed here.

NS type 1(NPHS1) is an autosomal recessive disorder characterized typically by CNS and often followed by a rapid progression to end-stage renal disease (ESRD).7 Affected children are usually born prematurely, with the mother having a large placenta. Hypoalbuminemia, hyperlipidemia, abdominal distention and edema appear soon after birth. Mutations in NPHS1 account for ∼50% of all cases of CNS.5, 8 Electron microscopic studies of the affected kidney in murine models show effacement of the podocytes and absence of the slit diaphragm.9 The Finmajor and Finminor mutations in NPHS1 are seen in the majority of Finnish patients with CNS.7 Although NPHS1 mutations are considered the major genetic cause of CNS, it has also been shown CNS may be caused by mutations in other genes.10

NS type 2 (NPHS2) is an autosomal recessive disorder typically characterized by childhood SRNS, and may progress to end-stage renal disease (ESRD).11 Kidney biopsies may show nonspecific histologic changes such as minimal change, FSGS or diffuse mesangial proliferation. Some patients show a later onset of the disorder.12 Mutations in NPHS2 account for 42% of familial and 10% of sporadic cases of childhood SRNS and have also been found in 39% of patients with CNS.2

NS type 3 (NPHS3) has been described in patients with infantile NS and childhood SRNS whom show histological changes of diffuse mesangial sclerosis on renal biopsy. Mutations in PLCE1 underlie this condition.13, 14

NS type 4 (NPHS4) is secondary to WT1 mutations, encoding the transcription factor Wilms’ tumor-suppressor gene 1. As well as causing Wilms’ tumor, Frasier syndrome and Denys Drash syndrome, WT1 mutations may cause NS and progression to renal failure. Histologically, the picture is usually diffuse mesangial sclerosis, but renal biopsies may show FSGS.15 WT1 mutations may cause SRNS in infants and children without features of Frasier syndrome16, 17 and have been reported in up to 9% of such cases. Genetic screening of other cohorts, including African-American children with FSGS and SRNS18 and Japanese children with CNS19 did not detect WT1 mutations.

NS type 5 (NPHS5) is characterized by CNS and very early-onset progressive renal failure. Some patients have associated ocular abnormalities including myopia, nystagmus and strabismus.20 Mutations in LAMB2 underlie this type of NS. Severe eye disease, NS and developmental delay in association with LAMB2 mutations is known as Pierson syndrome.21 Milder phenotypes with isolated renal disease have been described with homozygous and compound heterozygous missense mutations in LAMB2.20

NS type 6 (NPHS6) was recently described in two consanguineous families with childhood SRNS.22 The implicated gene is known as PTPRO.

Heterozygous mutations in CD2AP typically cause adult onset NS, with an autosomal dominant pattern of inheritance and lead to the histological diagnosis of FSGS.23 However, given its known interaction with nephrin24 and podocin25 it may also account for cases of childhood SRNS. Indeed, a homozygous CD2AP mutation in a child presenting at the age of 10 months has recently been described.26 In addition, a heterozygous CD2AP variant, together with a NPHS2 variant, has been found in a child presenting at 3 and a half years of age with SRNS and biopsy proven FSGS.27

Additional genes, some only very recently described, associated with NS include ACTN4,23 TRPC6,28, 29 INF2,30 and MYO1E.31 Recently, Nei endonuclease VIII-like 1 (NEIL1), which encodes a base-excision DNA repair enzyme and was postulated as a candidate gene for NS in a single consanguineous family.32 Its role in NS remains uncertain. Occasionally, NS may be part of more complex syndromes and examples include Nail-Patella syndrome33 and Galloway–Mowat syndrome.34

In this study, we screened for mutations in nine genes implicated in inherited NS in a Saudi Arabian population with either CNS, infantile NS or childhood SRNS. Such a study has never been conducted for this part of the world. The Saudi Arabian population has a tribal structure and the overall rate of consanguineous marriage is reported to be over 55%, with regional variations.35, 36 In such a population, the identification of mutations in known recessive disease genes is an important consideration and we were interested in our ability to detect a molecular genetic cause of NS. Identification of a molecular genetic cause of NS allows both a definitive diagnosis and improved clinical management of the patient and at risk relatives. It is noteworthy that the Saudi population is at high risk of renal failure, with 133 incident cases per million population per year that require renal replacement therapy.37, 38

We identified 62 cases, representing 49 families with CNS, infantile NS or childhood SRNS and undertook mutational analysis in known and candidate NS genes. We found a high rate of mutations in this cohort, solving 51% of cases. Direct sequencing identified novel and likely pathogenic genetic variants in NPHS1, NPHS2, MYO1E and PLCE1, increasing the known spectrum of mutations in these genes.

Materials and methods

Study cohort

This study has been approved by the research advisory council of King Faisal Specialist Hospital, Riyadh, Saudi Arabia (RAC#2050 045). Following informed consent, DNA was extracted from peripheral blood cells using the Gentra Systems PUREGENE DNA Isolation kit. A total of 62 samples from 49 different families were obtained. Altogether, 25 samples were obtained from 12 families with evidence of familial NS and 37 samples were obtained from families with a single affected individual with NS. Clinical phenotypes included patients with CNS, infantile NS and childhood SRNS. Consanguineous marriages were noted. Clinical data and biopsy reports were reviewed where available.

Mutation analysis

Mutational screening was undertaken of known genes implicated in NS; NPHS1 (RefSeq NM_004646.3), NPHS2 (RefSeq NM_014625.2), LAMB2 (RefSeq NM_002292.3), PLCE1 (RefSeq NM_016341.3), CD2AP (RefSeq NM_012120.2), MYO1E (RefSeq NM_004998.3), WT1 (RefSeq NM_000378.4), PTPRO (RefSeq NM_030667.2) and NEIL1 (RefSeq NM_001256552.1). Direct sequencing of all coding exons and exon–intron boundaries was performed. Oligonucleotide primers for PCR amplification of genomic DNA were designed using Primer3 software (http://frodo.wi.mit.edu/) and synthesized by Metabion International AG (Munich, Germany). Primer sequences are available on request. PCR was performed in a final volume of 25 μl containing approximately 20 ng of genomic DNA and Qiagen (Manchester, UK) master mix kit (including 1X PCR buffer, 100 μmol l–1 dNTP, pair, and 1 U per reaction HotStar Taq polymerase) and 0.5 μmol l–1 primer. PCR products were treated with the Agencourt AMPure PCR purification system (Agencourt Bioscience Corporation, Beverly, MA, USA). PCR products were sequenced using BigDye Terminator Cycle Sequencing kit (PE Applied Biosystems, Beverly, MA, USA) as described by the manufacturer. Sequences were analyzed using Mutation Surveyor software Version 3.24 (SoftGenetics LLC, State College, PA, USA) and SeqMan II software 6.1 (DNAStar, Madison, WI, USA).

Computational analyses of novel missense mutations were performed with PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), Sorting Tolerant From Intolerant (SIFT) (http://sift.jcvi.org/) and SNPs3D (single nucleotide polymorphism resource found at http://www.snps3d.org/). PolyPhen-2 scores range from 0 to 1, the higher the score the more damaging the amino-acid substitution. SIFT scores range from 0 to 1. The amino-acid substitution is predicted damaging is the score is ⩽0.05, and tolerated if the score is >0.05. For SNPs3D, a positive score indicates a variant classified as non-deleterious, and a negative score indicates a deleterious case.

To assess splicing effects we used the GeneSplicer software (http://www.cbcb.umd.edu/software/GeneSplicer/gene_spl.shtml). In addition to database searches, a control DNA panel from 175 individuals from a Saudi Arabian population was used to screen for all novel sequence variants.

Human MYOIE (accession number NP_004989) was modelled using the crystal structure of myosin-IE from Dictyostelium discoideum (protein data bank accession code, 1LKX).39 The Modeller program40 was used to create a three-dimensional model of MYOIE using the amino-acid sequence alignment generated through HHPred.41 Figures were prepared using PyMOL (http://www.pymol.org/).

Results

Sixty-two individual cases affected with CNS, infantile NS, childhood SRNS representing 49 families from the Arabian peninsula (Table 1) were screened for mutations in the following genes: NPHS1, NPHS2, LAMB2, PLCE1, MYO1E, WT1, PTPRO, NEIL1 and CD2AP. We identified mutations known NS genes in 9 out of 12 families (75%) where there was reported to be >1 affected member with NS, suggesting an inherited cause (Table 1). A molecular genetic diagnosis was obtained in 43% (16 out of 37) of families with a single affected member with NS. All families in whom we detected mutations had a history of consanguinity, most commonly first cousin marriages. Overall, in this population, by screening nine genes implicated in inherited NS we established a likely molecular genetic cause in 51% of families (Figure 1, Tables 1 and 2). No mutations were found in WT1 following screening of exons 8 and 9, NEIL1 or PTPRO by screening all coding exons. In this Saudi Arabian cohort, mutations in the NS genes NPHS2, NPHS1 and PLCE1 account for 22%, 12% and 8% of cases, respectively, of NS (Figure 1). Mutations in NPHS2 represent over two-fifths of genetically proven NS in this population.

Table 1 Clinical summary of Saudi Arabian nephrotic syndrome cases
Full size table
Figure 1
figure 1

Molecular analysis of cohort of 49 families with congenital nephrotic syndrome (CNS) and steroid-resistant nephrotic syndrome (SRNS). In all, 51% of families (25 out of 49) had a positive molecular diagnosis in 1 of 5 genes, including NPHS1 (6 of 49), NPHS2 (11 of 49), PLCE1 (4 of 49), MYO1E (3 of 49) and CD2AP (1 of 49).

Full size image
Table 2 Cases with mutations identified in disease-associated genes
Full size table

The unsolved families (Table 1) had a reduced incidence of consanguinity, making the likelihood of sporadic NS higher in this group. Comparing the clinical phenotypes of the groups of NS patients who had a molecular genetic diagnosis compared with those that remain unsolved, there were similar rates of end-stage renal disease (ESRD) but CNS was a much more common finding in the solved cohort. There was a wide spread of age of onset and histological phenotypes in both groups of patients.

Previously published data from a world-wide cohort suggested that ∼50% of CNS is caused by mutations in NPHS1.8 In our cohort, mutations in NPHS1 accounted for five out of eight cases (62.5%) of CNS, which had a confirmed molecular genetic diagnosis.

Of the novel homozygous sequence variants in coding regions (Table 3 and Figure 2), we identified three novel nonsense mutations, a donor splice site mutation with a high likelihood of aberrant splicing and two missense mutations. In silico predictions suggested that these missense changes are pathogenic (Table 3). A detailed analysis of the results for each gene is given below.

Table 3 Novel genetic variants detected in NPHS1, NPHS2, PLCE1 and MYO1E and in silico analysis of pathogenicity
Full size table
Figure 2
figure 2

Sequencing chromatograms of novel genetic variants in patients with nephrotic syndrome (NS) from Saudi Arabia. Sequencing chromatograms are shown for novel sequence variants detected in NPHS1, NPHS2, PLCE1, LAMB2, MYO1E and NEIL1 in patients with NS. Sequence chromatograms are annotated with nucleotides, and reference nucleotide sequences shown below. The reading frame is indicated by black bars and amino-acid translations are given for exonic regions where appropriate. Exon–intron boundaries are marked with a vertical dashed line where appropriate. Likely pathogenic changes are colored red, whereas those of unknown or benign pathogencity are colored green. A full color version of this figure is available at the Journal of Human Genetics journal online.

Full size image

NPHS1

In this gene, we found six families carrying known74243 or predicted pathogenic mutations. Five of the mutations were found in the homozygous state, and one (S22) was a compound heterozygous mutation involving two previously reported variants P368L and R460Q43 (Table 2). Here, segregation of the mutation from each parent was confirmed (data not shown). None of our cohort had Finmajor (nt121delCT, L41fsX91) or Finminor (c.3325C>T, R1109X) mutations.7 The affected patient from family S3 had a homozygous frame shift mutation in NPHS1 combined with a novel heterozygous NPHS2 (c.761G>T; p.C254F) allele (Tables 2 and 4). A p.P264R missense change was found in NPHS1 in two families (Table 5). This allele has previously been identified as contributing to a compound heterozygous mutation in NPHS1 in cases of severe CNS.44, 45 It has also been noted in cases of tri-allelic inheritance of steroid-resistant FSGS.27 The P264R variant has a minor allele frequency of 1% (rs34982899, dbSNP 1000 Genomes) and its pathogenicity is not proven. Indeed it has recently been classified as a polymorphism.46

Table 4 Novel genetic variants of unknown significance and polymorphisms detected in NPHS1, NPHS2, LAMB2, NEIL1 and PLCE1
Full size table
Table 5 Non-pathogenic sequence variants/polymorphisms identified in known nephrotic syndrome genes
Full size table

One novel homozygous sequence variant was detected in the NPHS1 gene (S37 Tables 2 and 3). This sequence variant is predicted to be a splice site mutation affecting the 5′-splice donor site (c.1627+2T>G) predicting exon skipping of exon 12. Another novel heterozygous NPHS1 variant c.528T>C (p.S176S) is in a highly conserved region, and may affect splicing at the acceptor site, given that the 528T nucleotide is the second nucleotide within exon 5 (Table 4). This variant was found in a heterozygous state in two patients from the same family (F22) in combination with a heterozygous missense change in LAMB2 (c.1982G>A; p.R561Q; Table 4).

Another novel heterozygous NPHS1 variant c.840+6G>A was found in five families, two of whom also had a homozygous mutation p.Q1020X in PLCE1 gene (Tables 2 and 4). This NPHS1 c.840+6G>A variant was also found in 1.15% of our normal controls, suggesting it is unlikely to be pathogenic in isolation (Table 4).

Of the six families with homozygous or compound heterozygous NPHS1 mutations, five presented with CNS, whilst the other had a milder phenotypes, consistent with previous reports.47

NPHS2

Among our NS cohort there were 11 families with NPHS2 mutations (Tables 1 and 2). All mutations in affected individuals were in a homozygous state consistent with known parental consanguinity. Mutations detected in NPHS2 accounted for 44% of all the mutations detected in our cohort.

Four of the NPHS2 mutations had previously been reported48, 49, 50 and two novel homozygous NPHS2 mutations were detected. These included a nonsense mutation c.115C>T (p.Q39X), which is the most premature truncating mutation reported so far in NPHS2 and a missense mutation c.413G>C (p.R138P) in patients with SRNS. Of note, this arginine residue at position 138 (R138) of podocin has been the residue mutated in other cohorts, specifically p.R138Q and p.R138X.48 Both mutations, R138Q and R138X, were reported in SRNS patients. Indeed R138Q has been described as a founder mutation in European populations.48 Huber and et al.51 demonstrated that this mutation causes a failure to recruit nephrin into lipid rafts, providing important insights into the pathogenesis of this NPHS2 mutation.

A novel heterozygous NPHS2 variant C254F was detected in a patient homozygous for a c.515delCCA mutation in NPHS1 gene (Tables 2 and 4). The variant C254F was not predicted to be definitely pathogenic by in silico models (Table 4) but neither was it found in our normal controls. This suggests that C254F could be a modifier allele.

Whilst screening our control population for novel NPHS2 variants, we observed a NPHS2 variant P20L (data not shown) in 2.32% of Saudi Arabian controls. This variant is of unknown functional significance52 and is rare in control European populations53 and was not associated with a NS disease phenotype, suggesting that this change should be considered as a rare polymorphism.

PLCE1

A novel homozygous mutation in PLCE1 (Q1020X) was detected in four families in our cohort, with a broad range of phenotypes (Tables 1 and 2). All affected individuals had the identical mutation, and segregation was confirmed in two of the families (F5 and F15, data not shown). This truncating mutation is predicted to have a severe affect on the PLCE1 protein with loss of the PLC catalytic domain, the protein kinase C conserved region and the Ras association domain.13 Interestingly, each of the Q1020X mutations in exon 7 of PLCE1 gene was associated with the variant (c.4665+52G>C) in intron 18 (Table 4).

Affected and unaffected daughters in a family of African ancestry with SRNS (F17) were found to be homozygous for a rare variant (K2173R) in PLCE1 (Table 4). We did not find this variant in our Saudi Arabian control cohort. The affected F17 member was also heterozygous for novel variants c.840+6G>A in NPHS1 gene and R644H in LAMB2 gene (Table 4).

MYO1E

We detected novel homozygous MYO1E mutations in three families (Tables 1 and 2). All families presented with SRNS. In the first two families, renal biopsy revealed FSGS in one case, progressing to end-stage renal disease by 10 years of age and minimal change disease, with preserved renal function in the other. The affected patients both had a Y47X nonsense mutation, segregating from each parent (data not shown). The Y47X is a severe truncating mutation, with a predicted loss of all functional domains of the MYO1E protein, including the motor-head domain, the calmodulin binding IQ domain and the tail domain.31 The affected female patient in family S17 presented at 8 years of age with SRNS and FSGS on renal biopsy and progressed to end-stage renal disease after 4 years (Table 1). She was found to have a T119I homozygous mutation (Table 2 and Figure 2). The threonine residue at position 119 is conserved throughout vertebrates (including zebrafish) and in the myosin-1d homolog of the amoeba Dictyostelium discoideum where it is the terminal residue of the P-loop motif (Figure 3).

Figure 3
figure 3

A homology model of MYO1E showing conservation of the P-loop motif. (a) Homology model of human MYO1E generated using the crystal structure of Dictyostelium discoideum myosin-IE (PDB 1LKX) bound to MgADP.VO4.39 The magnesium ion is shown as a magenta sphere. (b) Close-up of the nucleotide-binding domain of human MYOIE. The P-loop (112GESGAGKT119), together with the purine-binding loop, switch-1 and switch-2 (data not shown), comprise the nucleotide-binding site. The side-chain hydroxyl oxygen of T119 is shown in red. (c) Alignment of human MYOIE with various myosin-IE orthologues. The conserved P-loop threonine residue at position 119 of MYOIE (shown in b) is highlighted. A full color version of this figure is available at the Journal of Human Genetics journal online.

Full size image

CD2AP

A single patient presenting with CNS from family F23 in our NS cohort was found to have two known heterozygous variants in CD2AP (Tables 1 and 2). The first F220L is a variant of unknown significance (rs139926926), in a residue conserved to Danio rerio and was not detected in 350 healthy Saudi Arabian alleles. This variant was combined with T374A, a likely pathogenic mutation.54 The heterozygous T374A mutation in CD2AP was previously reported in a 2-year-old child presenting with SRNS and histological features of FSGS, and the missense change disrupts a proline-rich domain important for protein–protein interactions.54

In addition, the CD2AP variant K301M was found as a heterozygous allele in three of our patients; one was associated with homozygous Q1020X mutation in PLCE1 gene (F5), other two (S14 and S16) were detected alone (Table 5). Although K301M has been reported by Gigante et al.54 previously as a pathogenic mutation, we also found this variant in 2.79% of our normal controls suggesting that this variant is a polymorphism. The previously reported CD2AP variant p.K301M was found a 23-year-old Italian female presenting with SRNS and FSGS.54 The same mutation was found in her 6-year-old child who was reported as phenotypically normal. Functional studies showed that at K301M variant had a defective CD2–CD2AP interaction.54 The high frequency of the K301M variant in the Saudi population would certainly point away from this heterozygous variant alone being disease causing.

We speculate that CD2AP variants in the heterozygous state may increase susceptibility toward glomerular damage and proteinuria but until large-scale studies are conducted to elucidate the role of these variants, they remain of uncertain pathogenicity.

LAMB2

We found novel heterozygous LAMB2 sequence variants in five families from our cohort (Table 4). The affected patient in family S18 had a L465P heterozygous variant in LAMB2 gene together with an A1765T polymorphism in LAMB2;55 (Tables 4 and 5). The L465P missense change is novel, and was absent from control samples, but predicted to be benign using in silico testing (Table 4).

In another family (F22), two affected patients (brothers) with FSGS had a single heterozygous R561Q variant in LAMB2. This novel variation of unknown severity was present in 0.27% of healthy controls and was in combination with a novel heterozygous NPHS1 S176S variant, predicted to have some deleterious impact on splicing (Table 4).

Additional novel (and likely polymorphic) variants were detected in the LAMB2 gene (Table 4; Figure 2). A heterozygous L1258V missense variant was detected in two families and also detected in 0.55% of our normal controls. The variant R644H was detected in one allele of our patients and in 0.82% of normal controls.

Family S39 had a known heterozygous variant in LAMB2 (c.3443G>A; p.R1148H) in association with a heterozygous P264R NPHS1 polymorphism (Table 5). This R1148H LAMB2 variant was found also in 0.82% of our normal controls (Table 5).

Discussion

This study is the first to describe the molecular basis of NS in a Saudi Arabian population. Despite finding a high rate of mutations in known NS genes within our cohort (51%), these findings suggest there are other novel genetic causes of NS yet to be discovered. The genetic heterogeneity underlying NS, even in this highly consanguineous population, is evident. It is noteworthy that all patients with identified mutations in NS-associated genes were from consanguineous marriages and most of the pathogenic mutations identified were in the homozygous state. There were two families (S22 and F23) where we identified compound heterozygous mutations in NPHS1 and CD2AP, respectively. Thus, reliance on screening genes in homozygous regions alone in known consanguineous families may miss compound heterozygous changes in relevant genes. Our experience is not unique. Using a homozygosity mapping approach in 12 families from different backgrounds with CNS, Schoeb et al.8 solved just five families by detecting homozygous mutations in NPHS1. A more systematic search, including all known NS genes, is therefore important for achieving a high mutation detection rate.

Although large deletions of genomic DNA have not previously reported for NPHS1, NPHS2, LAMB2, PLCE1, CD2AP, MYO1E and PTPRO, additional genetic analysis specifically looking for these may improve mutation detection rates. Large deletions in WT1 have been previously reported,56 but were exclusively associated with a Wilms’ tumor phenotype rather than NS.

A strategy of targeted gene sequencing for patients manifesting NS in the first year of life in a world-wide study has previously been reported, and noted that two-thirds of patients could be explained by mutations in one of four genes (NPHS1, NPHS2, WT1 or LAMB2).5 In the modern era of whole-exome sequencing and targeted gene capture and sequencing, these approaches will allow mutations in known NS genes and novel NS genes to be detected with greater efficiency.

In our study, which included families with CNS, infantile NS and childhood SRNS, following molecular analysis of nine known NS genes, mutations were detected in around 51% of cases, which is comparable to the mutation detection rate in SRNS by other groups.57 Cohorts of more restricted phenotypes have identified mutations ∼80% of cases (non-Finnish CNS) by a systematic screen of implicated genes.46 Our mutation detection rate was, as one might predict, higher in patients where there was evidence of familial disease (>1 affected member; 75%) than in single individuals with NS (43%).

Mutations in NPHS2 gene are the most frequent identified genetic cause of NS in our Saudi Arabian cohort, accounting for 44% of all pathogenic mutations detected, in comparison with ∼40% of a European cohort,5 and ∼30% in a large Turkish study.58 Common NPHS2 mutations in our cohort included missense mutations V260E and V180M. All cases with mutations in NPHS2 presented with childhood SRNS, except one family (F3) that had CNS in association with a NPHS2 R168H mutation. All mutations detected in NPHS2 were homozygous, consistent with parental consanguinity in these cases.

NPHS1 mutations were detected, as expected, mainly in patients presenting with CNS patients. One case presented with childhood SRNS.

In PLCE1 gene, we found one mutation common to four families with NS patients. Each of these families were descendants from large Saudi Arabian tribes. In Saudi Arabia, eight tribes account for around 10% of the country’s population. The association of the Q1020X in PLCE1 with the intronic variant (c.4665+52G>C) in PLCE1 in each of these cases may indicate a founder effect (Table 4).

The original description of MYO1E mutations identified just two consanguineous families with missense mutations A159P and nonsense mutation Y695X, originating from Italy and Turkey, respectively.31 No mutations in MYO1E were identified in a screen of sporadic cases of FSGS,31 suggesting that MYO1E is an uncommon cause of NS, but its true incidence, especially in populations of high consanguinity has yet to be determined There have been no other reports of MYO1E mutations to date. Here, by identifying two novel MYO1E mutations in an Arabic consanguineous population, we confirm the pathogenicity of MYO1E in NS and expand the spectrum of mutations in this rare cause of NS.

Most of the mutations previously reported in the CD2AP gene have been heterozygous changes, except in one case with biopsy proven FSGS where a homozygous mutation was found.26 In our cohort, we identified a single patient from a consanguineous family harboring a compound heterozygous mutation (F220L/T374A) in the CD2AP gene. In a previously published cohort of 35 families with SRNS in whom NPHS1, NPHS2 and PLCE1 mutations had been previously excluded, no CD2AP mutations were identified, confirming their rarity.59

We found one case (S3), presenting as SRNS, where a previously reported homozygous mutation (c.515delCCA) in NPHS142 was associated with a heterozygous missense change (c.761G>T; C254F) in NPHS2. This emphasizes the need to screen multiple NS-associated genes, even within a consanguineous pedigree, to determine modifier gene effects. The presence of triallelism in NS has been noted before44, 45, 60 and additional alleles in NS genes may modify the renal phenotype and the clinical presentation and course. Weber et al.49 described a child with CNS in whom a combination of a heterozygous de novo splice mutation in NPHS1 and a homozygous NPHS2 R138Q mutation was detected.

We were extremely careful not to define all the variants we have detected as pathogenic mutations, although many of them were found in <1% of our normal controls. With each new variant, we carefully checked for the presence of additional mutations (in the same and other NS-associated genes) and confirmed the variant co-segregated, where parental samples were available, with clinical status within the family. As an example, the P264R variant we observed in NPHS1 (Table 5) has previously been implicated as pathogenic in cases of severe CNS,44 however, in a more recent paper it has been classified as a polymorphism.46 Often novel (heterozygous) sequence variants were also detected in our ethnically matched control population (Table 4).

We also note that the high rate of consanguineous marriage may make it difficult to estimate pathogenicity of novel mutations without segregation and extended family analysis. However, in silico tools have been used to identify pathogenicity of each of our novel mutations (Table 3).

In families with sporadic NS, the cause of the NS may not be genetic, and screening sporadic cases for known disease genes typically yields fewer results.58 Causes of sporadic NS are likely to be a combination of environmental factors together with genetic susceptibility factors.

The molecular genetic diagnosis of NS remains a vital aid to the clinical management of families with NS. It allows for the appropriate long-term management to be undertaken, genetic counseling to be undertaken and where necessary, screening of siblings and other at risk family members. Decisions regarding immunosuppression and transplantation are aided by a molecular genetic diagnosis. It is well recognized that children with causative mutations in NS genes do not respond well to treatment with corticosteroids and other immunosuppressants.5

In conclusion, in a Saudi Arabian cohort, 51% of families with NS were explained by mutations on five known NS genes (NPHS1, NPHS2, PLCE1, MYO1E and CD2AP). NPHS2 gene mutations were the most common molecular genetic cause of NS in this cohort. Unsolved patients from consanguineous families suggest additional novel genetic causes of NS are likely. The genetic heterogeneity of NS, suggests that screening strategies should continue to include multiple NS genes, including rare and recently discovered genetic causes, to allow a high yield of molecular genetic diagnoses. This will then lead to improvements in both precise diagnosis and clinical management.