Hereditary nephrotic syndrome: a systematic approach for genetic testing and a review of associated podocyte gene mutations

Congenital nephrotic syndrome

Congenital nephrotic syndrome has been arbitrarily defined as the occurrence of NS in patients <3 months of age, although congenital actually means that the disease is present from birth. The most common type of CNS is congenital nephrotic syndrome of the Finnish type (CNF), a recessively inherited disorder characterized by massive proteinuria detectable at birth, a large placenta, marked edema and characteristic radial dilatations of the proximal tubules. These histological lesions are detected more frequently after 3 months of age, but they have also been identified in fetuses [38–41]. They are also considered to be an inconstant feature. Mesangial expansion and capillary obliteration are also evident in a significant proportion of CNF cases [42, 43]. It has been shown that hypoperfusion of glomerular and tubulointerstitial capillaries and rarefaction of the latter may explain the rapid development of fibrosis in these patients [41]. The incidence of CNF in Finland has been estimated to be 1 in 8,200 live births [44], and a high incidence has also been reported among the Old Order Mennonites in Lancaster, Pennsylvania [45]. NPHS1 has been identified as the major gene involved in CNF [11] with the Fin-major (p.L41fsX91) and Fin-minor (p.R1109X) mutations accounting for 78 and 16% of the mutated alleles among Finnish patients, respectively [11]. The Fin-major and Fin-minor mutations have rarely been found in other ethnic groups. Among non-Finnish cases with a CNF phenotype, the NPHS1 mutation detection rate approaches 66% [46]. To date, more than 140 different NPHS1 mutations have been identified, comprising nonsense, missense, frameshift insertion/deletion and splice-site mutations.

It has now become clear that not all NPHS1 mutations cause severe CNS or a severe clinical course. In one study involving patients bearing the p.R1160X NPHS1 mutation in the homozygous state (resulting in a truncated protein lacking the C-terminal 82 amino acids implicated in the interaction with podocin), about half of the patients had a milder phenotype, in that although they presented with severe NS in the first 3 months of life, their further renal course was relatively benign with spontaneous partial or complete remission in childhood [47]. Histological findings, where available, were consistent with CNF. The clinical variability was apparently influenced by gender, as the majority of the mildly affected cases were female.

Although NPHS1 is the main gene that has been identified in patients presenting NS in the first 3 months of life, it has also been shown that CNS may be caused by mutations in several other genes, including NPHS2. Mutations in this latter gene have been detected in patients with a classically severe CNF phenotype [47] and have been shown to account for up to 51% of all mutations in Central European patients with CNS [10]. It has to be pointed out that the median age at onset of CNS in the latter study was relatively late (median age 4 weeks), which may induce a bias toward a greater prevalence of NPHS2 mutations instead of NPHS1 mutations. Indeed, in our large worldwide cohort of CNS patients, which mostly presented proteinuria in the first few days of life, NPHS1 mutations were detected in more than half of all cases (unpublished data). Mutations in the PLCE1 gene [16, 48–50], and the WT1 gene [48, 51–55] have also been detected in patients presenting isolated CNS with DMS on renal histology; mutations in these two genes will be discussed in more detail in the next section.

These observations suggest that, for patients presenting nonsyndromic CNS, the NPHS1 gene should be tested first in those presenting NS shortly after birth as well as in those with typical proximal tubular radial dilatation. Molecular analysis of NPHS2 should be the next step whenever mutations in NPHS1 are not detected. Patients presenting later in the congenital period (particularly if renal biopsy shows FSGS or minimal glomerular changes) should probably be initially screened for NPHS2 mutations, followed by NPHS1. In cases for which renal histological findings are available and DMS is determined, genetic testing of the WT1 and PLCE1 genes should initially be performed.

Infantile nephrotic syndrome and childhood nephrotic syndrome

The term infantile NS has been proposed for patients that develop NS between the ages of 4 and 12 months. NPHS2 mutations are responsible for most of these cases [12], and they have also been found in a significant proportion of patients with childhood-onset SRNS. Mutations in this gene occur in about 40% of familial and 6–17% of sporadic SRNS cases (Table 1) [55–59]; patients typically present NS from birth to 6 years of age and reach ESKD before the end of their first decade of life [10, 55–57, 59, 60]. Renal histology of such patients reveals either minimal glomerular changes (if biopsied early) or FSGS. The identification of mutations is important as it may enable the treating physicians to avoid (or discontinue) prescribing immunosuppressive therapies for these patients, sparing them the significant side-effects associated with these drugs. Indeed, considering that access to genetic testing is now easier than it has been in the past and that sequencing of the NPHS2 gene is relatively fast as it comprises only eight exons, screening for NPHS2 mutations in patients presenting the renal phenotype described above should be performed prior to the initiation of additional—potentially deleterious—therapy. To date, more than 100 pathogenic NPHS2 mutations and 25 variants of unknown significance have been reported, including a full spectrum of protein-truncating nonsense and frameshift mutations, splice-site variants and missense changes, involving all coding exons. Patients with frameshift, nonsense or the homozygous p.R138Q missense mutations manifest symptoms at a significantly earlier age [57, 59]. The p.R138Q mutant, which accounts for up to 32% of all mutant alleles, is retained in the endoplasmic reticulum (ER), where it essentially functions as a null allele and fails to recruit nephrin to lipid rafts [61, 62]; this may explain the phenotype severity associated with this mutation. In contrast, the p.R229Q variant has been shown to lead to late-onset NS [37, 63], when found in association with one pathogenic NPHS2 mutation. This variant represents the most frequently reported nonsynonymous NPHS2 variant in Caucasians and is particularly common among Europeans, in whom the observed frequency of heterozygotes ranges from 0.03 to 0.13 [55, 59, 63–66]. In vitro studies have demonstrated decreased binding of the p.R229Q mutant protein to nephrin, providing a likely explanation for its pathogenic role [63].

It has recently been shown that mutations in NPHS1 also account for a nonnegligible proportion of infantile and childhood-onset SRNS cases. Two studies found NPHS1 mutations in 7–14% of the patients presenting SRNS at least 3 months after birth [age at onset of NS in mutated patients 0.5–8 years (mean 3 years) and 0.7–27 years (mean 8 years), respectively] [36, 67] with minimal glomerular changes, FSGS or mesangioproliferative lesions on renal biopsies. These percentages are, however, overestimated as these studies included patients for which mutations in the NPHS2 gene were excluded. The presence of at least one “mild” mutation likely explains the later onset and milder course of the disease among these cases. As such, missense mutants retaining their abilities to traffic in the cell properly, a splice-site mutation allowing some correct splicing and a protein-truncating mutation involving only the very C-terminal end of the protein may be designated as “mild” because partial function of nephrin is maintained. Similarly, NS with spontaneous partial remissions and repeated relapses concurrently with respiratory infections has been described in two siblings bearing nephrin mutations [68]. Both infants, who presented NS at birth and at 10 months of age, respectively, were compound heterozygous for the p.C265R and p.V822M mutations with minimal glomerular changes observed on the renal biopsy. The p.C625R mutant protein was predominantly trapped within the ER, while the p.V822M variant protein reached the plasma membrane, probably explaining the milder phenotype seen in these patients. Modifier genes or environmental factors may also play a role in renal phenotype variability, as the same two NPHS1 mutations (p.R827X and p.R976S) have been identified in one adult patient diagnosed with FSGS at 27 years of age with unimpaired renal function after 2 years of follow-up [67] and in one patient with infantile-onset NS [36].

In addition to NPHS2 and NPHS1, PLCE1 is involved in some infantile and childhood-onset SRNS cases and is the main gene causing DMS. PLCE1 mutations have been detected in 28.6% of families with isolated DMS [48], with the clinical onset of reported cases of DMS varying from few days of life to 4 years of age [16, 48–50] and all patients having truncating mutations. The results of one study suggested that homozygous PLCE1 missense mutation (found in only two siblings) may lead to a milder phenotype of FSGS with a relatively late age at onset of proteinuria (in this study, 8.8 years and 2.0 years, respectively). However, in our cohort, truncating or missense mutations were detected in both DMS and FSGS patients, leading to a similar renal evolution (in press). Nevertheless, PLCE1 mutations remain an infrequent cause of FSGS: a Dutch study did not find PLCE1 mutations in 19 cases of childhood-onset FSGS [69] nor were mutations in this gene found in 69 families (median age of disease onset 26 years, range 1–66 years) with idiopathic or hereditary FSGS [70].

WT1 mutations may account for about 9% of patients with nonfamilial isolated SRNS [54], and they have been identified in patients with isolated DMS, with a clinical onset varying from a few days of life up to 2 years of age, as well as in isolated FSGS (1–14 years of age) [48, 51, 54, 71–74]. Almost all cases were those of phenotypically female patients, and the mutations occurred mainly in exons 8 and 9, which code for zinc finger domains 2 and 3, respectively [54].

Based on these observations, NPHS2 followed by NPHS1 remain the first genes to be tested in nonsyndromic patients presenting SRNS associated with minimal glomerular changes/FSGS in the infantile or childhood period. In the remaining patients with the same histological lesions, genetic testing for WT1 mutations (exons 8 and 9 in phenotypically female patients) should be performed, while screening for PLCE1 mutations may be considered in some cases (mainly in familial cases). However, the probability of identifying a PLCE1 mutation in this patient category is low, and the molecular analysis of this large 33-exon gene is expensive and time-consuming; thus, these factors should be taken into account in the decision on genetic screening. In cases of isolated DMS, PLCE1 is the most frequently involved gene but, for cost-effective reasons, genetic testing of exons 8 and 9 of WT1 (especially in phenotypically female patients) may be performed prior to PLCE1.

Mutations in the CD2AP, ACTN4 and TRPC6 genes have been anecdotically reported in young patients, precluding any suggestion of systematic mutational screening of these genes in children—unless there is an autosomal dominant familial history of FSGS, which would warrant the molecular analysis of the two latter genes. To date, the clear pathogenic implication of CD2AP in NS has been shown in only one study: a homozygous mutation (p.R612X) was identified in a 10-month-old nephrotic child who presented global glomerular sclerosis on renal biopsy; the truncated protein displayed a dramatic reduction of actin binding efficiency in vitro [75]. Heterozygous expression of the CD2AP mutation in both parents did not lead to any kidney pathology. Nevertheless, the insignificant role of recessive CD2AP mutations has been emphasized in a recent study that did not find homozygous CD2AP mutations in a cohort of 42 children (35 families) with SRNS for whom the NPHS1, NPHS2, PLCE1 and WT1 mutations had been previously excluded [76]. On the other hand, carrying a heterozygous mutation in the CD2AP gene has been reported to be a predisposing factor towards developing FSGS. To date, five different CD2AP heterozygous mutations have been identified in pediatric (and adult) FSGS patients [14, 69, 77], associated with reduced expression of CD2AP or defective CD2–CD2AP interaction in lymphocytes and down-regulation of CD2AP, nephrin and podocin glomerular expression on kidney biopsies. Unfortunately, as complete segregation data are not available in all of these cases, neither penetrance nor inheritance can be assessed with certitude. In addition, heterozygous CD2AP mutations have been found in clinically unaffected patients. Therefore, there is still some doubt surrounding the direct causal link between heterozygous CD2AP mutations and FSGS; it is possible that CD2AP haploinsufficiency has a role in the susceptibility to develop glomerular disorders instead [14]. Mutations in ACTN4 and TRPC6 have been implicated in the rare autosomal dominant FSGS and account for approximately 4 and 6% of familial FSGS, respectively [78, 79]. Patients with ACTN4 mutations usually present proteinuria in their teenage years or later, with a slow progression to ESKD in their fifth decade of life [13, 80]. Similarly, disease onset in patients with TRPC6 mutations has initially been reported to vary between 17 and 57 years [15, 81]. However, de novo ACTN4 mutations have been reported in three children (two families) from 3 to 5 years of age who presented rapid progression to chronic kidney disease/ESKD [78, 82]. We have detected a de novo TRPC6 missense mutation in one patient that presented NS secondary to FSGS at 6.5 years of age and reached ESKD few months later (unpublished data), while TRPC6 missense mutations have also recently been identified in two FSGS patients with disease onset at 7 and 9 years of age, respectively; the first patient showed a partial response to cyclosporine A and mycophenolate mofetil (MMF), and the mutation was also found in her asymptomatic 40-year old father [79, 83]. Therefore, similarly to CD2AP heterozygous mutations, TRPC6 may contribute to glomerular disease in a multi-hit setting. Finally, mutations in INF2 have also been found very recently in 11 families presenting moderate proteinuria and FSGS lesions in early adolescence or adulthood, thereby explaining about 12% of familial FSGS cases with an apparent autosomal dominant inheritance [17]. If the significant role of INF2 as a genetic cause of NS is confirmed by further studies, screening for mutations in this gene will also need to be recommended in the near future in patients with an autosomal dominant familial history of FSGS.

A summary of the mutational screening approach suggested in patients with nonsyndromic SRNS is given in Fig. 1.

Genetic counselling

Genetic counselling should be offered to any patient and his/her family affected with an inherited disorder. Information that needs to be shared will vary according to the mutations found. Therefore, clinicians should be aware of the specific aspects that should be discussed with patients, particularly in cases involving mutations in the NPHS2, WT1 and PLCE1 genes.

In SRNS patients with NPHS2 mutations who are considering having children, genetic testing for the p.R229Q variant, which is found in a high frequency in some populations, should be proposed to their asymptomatic spouses. Indeed, if the p.R229Q variant is identified in the spouse, there is up to a 50% risk of disease (a juvenile or adult-onset form of SRNS) transmission to their progeny. In addition, heterozygous carriers, such as patient siblings, also have a non-negligible risk of disease transmission to their children if the spouse carries the p.R229Q variant.

Special considerations also apply for patients bearing mutations in the WT1 gene. Although most of the WT1 mutations are de novo, XX female patients with isolated DMS/FSGS secondary to the WT1 mutation have normal genital development and may become pregnant; therefore, these patients have a 50% risk of transmitting the mutated gene to their children. The phenotype of the progeny will depend on its karyotype; a 46 XY child may develop complete Denys–Drash syndrome or Frasier syndrome, while a 46 XX child will not present ambiguous genitalia. A case report well illustrates the clinical implications of familial WT1 mutation transmission: the index case’s mother had typical FSGS (clinical onset at 6 years of age) with normal external and internal genitalia, while the index case presented with early DMS and XY pseudohermaphroditism [73]. Therefore, a female bearing a WT1 mutation and planning to be pregnant should be informed that her children may develop a phenotype significantly different (and more severe) from her phenotype.

Variability in the renal phenotype should also be discussed with families in which PLCE1 mutations are present. Indeed, it has been suggested that PLCE1 mutations are not always sufficient to cause DMS [49] as a same homozygous truncating mutation has been found in an asymptomatic father and his affected children with DMS. We have also identified three asymptomatic adults from three unrelated families bearing homozygous mutations with at least one affected sibling haploidentical to the unaffected cases (in press). We may speculate that modifier genes or environmental factors play a role in the renal phenotype variability observed in individuals bearing PLCE1 mutations. These observations need to be considered in the ensuing discussion with the patients on renal prognosis.

Finally, prenatal diagnosis should be offered to families with a known risk for severe NS, such as CNS, and to patients for whom elevated measurements of alpha-fetoprotein (AFP) in the maternal serum and amniotic fluid have been detected. In the first scenario, it has to be stressed that although results can be obtained relatively quickly when the causative mutations have previously been identified in the family, prenatal genetic testing may be more time-consuming in other cases. This highlights the importance of referring these families to a geneticist—ideally before conception or, if not possible, very early during pregnancy. An early genetic consultation may not only provide the background for a discussion of the risk of disease transmission and ethical questions related to medical termination of pregnancy (MTP), it may also allow the appropriate diagnostic procedures (trophoblastic biopsy or amniocentesis) to be planned within an adequate time-frame. In the second scenario involving families in which elevated AFP have been unexpectedly found, genetic counselling is also required as this finding may suggest CNS secondary to nephrin mutations (assuming that anomalies such as fetal anencephaly or omphalocoele have been ruled out). However, this diagnosis needs to be confirmed by mutation analysis prior to considering MTP as AFP elevations have been observed in NPHS1 heterozygous fetal carriers and in a fetus with Denys–Drash syndrome [84, 85].