Main

Primary SRNS is a clinically and genetically heterogeneous disease. Histology phenotypes include focal segmental glomerulosclerosis (FSGS) in 80%, minimal change nephrotic syndrome (MCNS) in 20%, and, rarely, diffuse mesangial sclerosis (DMS). Advances have been made in the understanding of SRNS pathogenesis through positional cloning of genes causing nephrotic syndrome (NPHS1, NPHS2, α-actinin-4, and CD2AP). The encoded proteins nephrin, podocin, α-actinin-4, and CD2AP are integral components of the slit membrane of the glomerular podocyte. Whereas NPHS1 mutations seem to exclusively cause congenital nephrotic syndrome, NPHS2 mutations have been identified in congenital, childhood and adult onset SRNS (1–5) and accounted for 26% of all SRNS patients in one study (3) and 10% of sporadic and 43% of familial cases in a second (5). Several small studies and case reports have been published on WT1 mutations in patients with isolated SRNS (6–11) or associated with urogenital malformations (8,12–17). WT1 mutations occurred mainly in exons 8 and 9, which code for zinc finger domains 2 and 3, respectively. A possible connection between the structural protein nephrin and the transcription factor WT1 is the finding that WT1 transcriptionally activates the NPHS1 promoter, resulting in up-regulation of nephrin mRNA (18,19). WT1 was originally identified as the gene responsible for the childhood kidney cancer Wilms' tumor (WT) (20,21). Denys-Drash syndrome (DDS) (22,23) and Frasier syndrome (FS) (24) were then shown to be caused by WT1 mutations in patients with an XY karyotype (13,25). Recently, our group conducted a study on 115 sporadic SRNS and 110 sporadic steroid-sensitive nephrotic syndrome (SSNS) cases, which focused on WT1 exons 6–9 (26). Only mutations in exons 8 and 9 were identified in 6% of individuals with isolated SRNS and predominantly an XX karyotype. For WT, Royer-Pokora et al. (27) recently published a study on 117 children with WT1 germline mutations located in all 10 exons of the gene. To address the question whether mutations associated with SRNS also occur throughout WT1, we screened a worldwide cohort of 167 patients (121 new; 46 previously examined for exons 6–9 by us) for all 10 exons of WT1 using MCHA and direct sequencing.

MATERIALS AND METHODS

Materials.

Blood samples, clinical data, and informed consent were obtained from patients and their parents. Genomic DNA was isolated from blood samples using standard methods (28). Ethics approval was obtained from the ethics committee of the University of Freiburg (Freiburg, Germany) and the Internal Review Board of the University of Michigan (Ann Arbor, MI). The diagnosis of SRNS was established by pediatric nephrologists at different centers according to published criteria (29). For clinical evaluation a standard questionnaire was used (4). Forms are accessible at www.renalgenes.org. Patients with a familial form of nephrotic syndrome or disease causing mutations in the NPHS2 or NPHS1 gene were not included. Minimal clinical diagnostic criteria to define SRNS were proteinuria, histologic diagnosis of chronic glomerulonephritis, and lack of response to standard steroid therapy. Additional non-exclusive features noted were age at onset of proteinuria and extrarenal manifestations, e.g. genitourinary tract malformations. Response to steroid treatment was defined according to the International Study of Kidney Disease in Children and Arbeitsgemeinschaft für Pädiatrische Nephrologie guidelines (29). All patients in this study with primary SRNS had an age of onset of 18 y or younger. Of these 167 patients, results have been partially published by Ruf et al. (26) for 46 patients. Patients' ethnic background was central European, Turkish, African-American, Hispanic, or Asian.

Heteroduplex analysis.

All patients underwent exon PCR and direct sequencing of exons 8 and 9 of the WT1 gene as described in Ruf et al. (26). A group of 167 patients were analyzed by MCHA according to the recommendations in the original article (30) for exons 2–10 of WT1. For primer sequences, fragment length, and fluorescent labeling, see Table 1. Products of three exons with different fragment lengths and fluorescent labeling were pooled for every individual yielding three panels (exons 2, 4, and 7; exons 8, 9, and 10; exons 3, 5, and 6). Samples were run on a MegaBACE 1000 DNA Analysis System (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) and evaluated with Genetic Profiler software version 1.1 (Molecular Dynamics, Sunnyvale, CA).

Table 1 Primer sequences used for MCHA and direct sequencing
Full size table

Direct sequencing. Due to its high GC content (70%) and product length, exon 1 was considered unsuitable for MCHA. Therefore, this exon was evaluated by direct sequencing of both strands for all 167 patients, as were exons 8 and 9. Also, samples that yielded an aberrant peak pattern or insufficient peak height in MCHA were amplified and analyzed by direct sequencing of the forward strand using the primers in Table 1 (without fluorescent label). Sequences were analyzed using Sequencher version 4.1.4 software (GeneCode, Ann Arbor, MI).

RESULTS

Mutations.

In total, we detected five different WT1 mutations in six new patients and two different mutations in the previously published group of WT1 patients (Table 2) (26). All seven different mutations cause changes in the running pattern of the fragments of exons 8 and 9 and were detected by MCHA in all 12 patients. One novel sequence variant was found to cause an amino acid change in exon 9 (P382S) and will be discussed separately. We also detected 14 different WT1 polymorphisms (see Table 3) of which seven were located in exon 1, three in exon 3, and one each in exons 2, 5, 7, and 8. Polymorphisms in exons 7 and 8 were not detected by MCHA analysis and were only found by sequencing. Nine variants are known SNP, according to the UCSC database May 2004 freeze (http://genome.ucsc.edu), five nonpathogenic sequence variants are not described there and are either located in the intron or cause exonic base changes without consequence at the amino acid level (see Table 3). In WT1 exon 8 we detected two heterozygous missense mutations: c.1079G>A C360Y in patient A201 and c.1119C>A H373Q in patient A644 (Table 2). C360Y has been described previously in a female patient with a diagnosis of DDS (12). H373Q has been recently published for a patient with WT (27). By MCHA, both samples showed shouldered peak patterns as expected (Fig. 1). In WT1 exon 9, five different heterozygous mutations were detected in 10 patients from different families (see Table 2). Mutation H401R in patient A133 is not previously described, but a different mutation affecting the same codon (H401Y) has been identified in a female patient with DDS nephropathy and unilateral WT (31). This base change was not found in 90 healthy controls from central Europe. The point mutation c.1162T>C C388R in patient F1031 has been published by Ruf et al. (26). c.1180C>T R394W was first described by Pelletier et al. (25) and has been found frequently in patients with DDS and WT (8,11,32–34). All three point mutations affect the third zinc-finger domain (amino acids 383–405). In the other seven patients, we identified two different splice site mutations: three patients were heterozygous for IVS9 + 4C>T and four patients were heterozygous for IVS9 + 5G>A. IVS9 + 4C>T was first described by Barbaux et al. (13) in a patient with FS and an XY karyotype. IVS9 + 5G>A has been associated with DDS nephropathy, WT, or gonadoblastoma (11,35,36).

Table 2 Clinical data and WT1 mutations of 121 new and 46 previously published patients with SRNS
Full size table
Table 3 Innocuous Polymorphisms of WT1 in a worldwide cohort of 164 SRNS patients
Full size table
Figure 1
figure 1

Examples of MCHA results exon 8. A201 II-1 (A201_II_1) and A644 carry heterozygous point mutations causing a shoulder pattern in MCHA. Individual identifiers are located in the left upper corners. The horizontal axis denotes scan points, the vertical axis denotes the relative peak intensity. Because no length standard was used, scan points cannot be compared between different samples. F1429 is a control individual yielding the wild type MCHA pattern for comparison.

Full size image

All mutations were detected by MCHA and exhibited running patterns distinct from the wild-type sequence. The running patterns were characteristic for specific mutations: The samples of the three patients with IVS9 + 4C>T showed a double peak, whereas the peaks of the four patients with IVS9 + 5G>A all had shouldered peaks.

New sequence variant P382S in family F1351 in WT1 exon 9.

Sequencing of exon 9 in patient F1351 II-1, a girl from Montenegro with isolated SRNS, showed a not previously described heterozygous base exchange, c.1144C>T, leading to the nonconservative substitution of a proline for a serine at position 382 of the WT1 protein (Table 2). This base change was not found in any of the other 163 patients nor in 90 healthy controls from central Europe. This is not a de novo mutation inasmuch as it was also present in her mother who suffers from end-stage renal disease of unknown primary cause and a healthy 20-y-old sister (Fig. 2A). The proline to serine substitution is from a hydrophobic side chain to a noncharged polar side chain. This proline is highly conserved in mouse, rat, Danio rerio, Drosophila melanogaster, Caenorhabditis elegans, and Saccharomyces cerevisiae and is located directly next to zinc finger domain 3 (amino acids 383-405, Swiss-Prot entry P19544) (Fig. 2B).

Figure 2
figure 2

Sequence variant of unknown significance in F1351. (A) Pedigree of F1351 and sequencing results. Squares denote males, circles females; filled symbols indicate renal disease; arrows denote individuals included in MCHA. (B) Conservation of P382 (arrow) in the Wt1 protein of different species. The conserved proline is located C-terminal to the first amino acid of zinc finger domain 3 (F383).

Full size image

No known SNP are located within the amplicon sequence for exon 9 according to UCSC and no other polymorphisms were detected in exon 9 by sequencing or MCHA.

Clinical data.

Of the 167 patients included in this study, all had SRNS. Data on 46 patients have been published previously (26). A total of 15 patients with WT1 mutations in exons 8 and 9 were identified by direct sequencing. Eight of these patients had been identified and published by Ruf et al. (26). A new group of 121 patients (59 female, 62 male), with nonfamilial SRNS was identified from our database. Informed consent had been given. In this group, seven (5.8%) new patients with WT1 mutations in exons 8 and 9 were identified, 6/59 female (10.2%) and one male patient with genitourinary tract malformations (see Table 2). In total, nine patients out of a group of 229 patients with nonfamilial isolated SRNS were identified with WT1 mutations (3.9%). Including patients with genitourinary tract malformations and WT increases the rate to 15 out of 235 patients (6.3%).

MCHA for exons 2–10 was performed with 164 samples, of which 38 samples belonged to patients who had been included in Ruf et al. (26) and who did not have any mutation in WT1 exons 7–9, and 114 new SRNS patients.

One patient, F1351 II-1, with a previously unpublished sequence change in exon 9 was also included in this study although her mother was reported to suffer from end stage renal failure of unknown pathogenesis (see Fig. 2A and Table 2). The other 151 patients did not have any affected family members. None of the patients without pathogenic mutations in exon 8 or 9 exhibited mutations in other exons.

For the evaluation of the clinical data, three additional patients (A651, F734, F1073) with WT1 mutations who were not available for MCHA were included (see Table 2). In the group of 15 patients with WT1 mutations (WT1+), ten patients had isolated SRNS without genitourinary tract malformations and without diagnosis of WT at the time of enrollment in the study. All 10 patients were phenotypically female, 7/10 exhibited FSGS on kidney biopsy. Data on 8/15 patients has been published previously by Ruf et al. (26). Seven patients with WT1 mutations were part of 120 new patients collected worldwide (see Table 2). Only three patients with a male phenotype and WT1 mutations were identified, leading to a female:male ratio of 4:1 in the WT1+ group. In the group without WT1 mutations (WT1−), the gender distribution was 1:1.2. The median age of onset of proteinuria was 5.8 y compared with 4.7 y in the WT1+ group. The percentage of patients in the WT1− group who developed ESRD was 26% (40/152) compared with 53% (8/15) in the WT1+ group. Median age at ESRD was 13.7 y in the WT1− group versus 8.1 y in the WT1+ group, but with a shorter time interval between clinical diagnosis of proteinuria and ESRD. Histology phenotypes were comparable in both groups (see Table 4).

Table 4 Clinical data on 167 SRNS patients
Full size table

DISCUSSION

In this study, we addressed the question whether a significant number of patients in a worldwide cohort of 152 with isolated SRNS and no mutations in nephrin (NPHS1) or podocin (NPHS2) have WT1 mutations outside the known “hotspot” regions in exon 8 and 9. We identified six different WT1 mutations in seven patients of a worldwide cohort of 120 new patients with isolated SRNS. In addition, we identified a novel base change (P388S) at the border of the third zinc-finger domain of WT1. We demonstrate that mutations in this SRNS cohort and 44 additional patients previously examined for exons 6–9 (26) are restricted to WT1 exons 8 and 9.

Within our on-going study project, sequencing of exons 8 and 9 in patients with SRNS lead to the identification of 15 patients with mutations in exons 8 and 9. Ten patients suffered from isolated nephrotic syndrome and had a female phenotype (see Table 2 and ref. 26). If numbers in Royer-Pokora's study in patients with Wilms' tumor were equivalent to patients with isolated SRNS, we would have expected to identify 8 patients with WT1 mutations in exons 1–7 or 10 (equivalent to 45%) in addition to our 10 patients with isolated SRNS that had mutations in exon 8 and 9 (equivalent to 55%). However, direct sequencing and MCHA did not detect any pathogenic mutation in exons 1–7 or exon 10 in 151 patients with isolated SRNS.

One new sequence variant of unknown significance and five new sequence variants.

One base exchange not previously described, c.1144C>T P382S, was identified. Although determining its significance is difficult, several factors favor the notion that it might cause nephrotic syndrome in this family: 1) It was not found in 90 healthy controls from central Europe. 2) It segregates with the affected mother I-2 and daughter II-1 (Fig. 2A). 3) It affects an amino acid highly conserved in evolution. Although not part of a zinc-finger domain, the change from a nonpolar hydrophobic amino acid to an uncharged polar amino acid and its location immediately N-terminal to zinc-finger domain 3 suggests that it might disrupt the structure and function of WT1 transcriptional activity.

Arguments that this may be an innocuous polymorphism are as follows: 1) The so-far healthy 20-y-old sister II-2 carries the same heterozygous base change (Fig. 2A) but future onset of nephrotic syndrome cannot be excluded. 2) Phenotypic data of the mother is scarce and a cause of ESRD other than primary SRNS cannot be excluded. She presented with renal insufficiency at age 30 y leading to end-stage renal failure 1 y later. 3) Whereas the great majority of described mutations in WT1 are de novo mutations, this change could be another rare example of a transmitted WT1 mutation. There are few reports on familial cases of WT1 mutations (37–39). A mother with isolated proteinuria and preserved renal function carried IVS9 + 5G>A heterozygously and transmitted it to her daughter who was diagnosed with nephrotic syndrome at the age of 9 mo (39). In the other two case reports, siblings with WT inherited heterozygous truncating mutations from phenotypically unaffected parents (37,38). Functional studies will be necessary to finally determine the significance of this new base change c.1144C>T. The other five base changes not previously described were classified as nonpathogenic, because of their localization in intronic sequence or because of their silent character. Also identified were eight known SNP (see Table 3).

Genotype-phenotype correlations.

The predominant phenotype on kidney biopsy in our patient population was FSGS. In fact, clinical data on 7/120 patients from this study and 8/125 patients from a previous study (see Table 2) shows that the histology phenotype of patients with WT1 mutations is not DMS. Age of onset of proteinuria and/or edema was highly variable with congenital cases of SRNS (A651 and F1031), as well as in one juvenile patient (F953), even between patients with the same mutation. WT1 mutations lead to ESRD in 7/15 patients after a variable period of time. Whereas A651 suffered renal failure at onset and died shortly thereafter of sepsis, F963 developed ESRD 11 y after onset of proteinuria. Of the five children that received a kidney transplant, none suffered recurrence of nephrotic syndrome with less than 1 y follow-up in three patients. Patients F921 and F999 lost the transplanted kidney after 13 and 7 y respectively due to chronic rejection (Table 2). Although clinical data of 15 patients is not sufficient to draw genotype-phenotype correlations, it already indicates, given the heterogeneity of the patient group in this study, that it may be difficult to demonstrate clear cut genotype/phenotype correlations for WT1 mutations.

Examination of the gender distribution in the WT1+ group shows that all patients with isolated SRNS were phenotypically female, whereas males exhibit additional genitourinary tract malformations (see Table 2). In contrast, none of the male patients in the WT1− group had a history of genitourinary tract malformations. One might argue that mainly females have isolated SRNS caused by WT1 mutations and that therefore the selection of SRNS patients in our study has a bias toward the male gender. However, case reports on isolated SRNS in patients with WT1 mutations are rare (6–10,39) and few cases of phenotypically male patients without genitourinary tract malformations have been published (9,11).

Sensitivity of MCHA.

In the first publication of MCHA on a MegaBACE, Hoskins et al. (30) achieved a sensitivity of 100%. In our study, sensitivity of MCHA was 91%. The known SNP A903G (dbSNP rs16754) in codon R301 was not detected by MCHA, although it is located 72 bases away from the end of amplicon 7. It is embedded in a GC-rich sequence that could have interfered with its detection (30). In Royer-Pokora's study on patients with WT, 12.9% of all mutations within the WT1 gene (12/93) were located in exon 7 and intron 7, respectively (27). Three patients with WT had the truncating mutation c.901C>T R301X. It is not excluded that not only the innocuous silent polymorphism c.903A>G in exon 7 but also pathogenic mutations affecting the same codon were missed in our patient population.

However, in Royer-Pokora's study (27), 16 of 93 patients with WT1 mutations had heterozygous deletion or insertion mutations that constitutes a rate of 17%. As MCHA has a high sensitivity for insertions and deletions it is unlikely that any deletion or insertion mutations escaped the detection by MCHA in our patient group.

In this study, WT1 mutations in patients with isolated SRNS were detected only in exons 8 and 9. Mutations of WT1 outside exons 8 and 9 can therefore only account for a very small proportion of isolated nephrotic syndrome, if at all. In conclusion, we are convinced that a sensitivity of 91% is extremely high for the rapid and inexpensive screening technique MCHA in comparison to single-stranded conformation polymorphism (SSCP), denaturing HPLC (DHPLC), and other standard methods (30). MCHA is therefore sufficient when screening for mutations in a large group of patients. It should also be sufficient to screen patients for this small group of patients in whom SRNS is caused by WT1 mutations in exons 8 and 9 and allow for directed clinical follow-up and therapy.