NPHS2 Mutations: A Closer Look to Latin American Countries

Nephrotic syndrome is one of the most common kidney pathologies in childhood, being characterized by proteinuria, edema, and hypoalbuminemia. In clinical practice, it is divided into two categories based on the response to steroid therapy: steroid-sensitive and steroid resistant. Inherited impairments of proteins located in the glomerular filtration barrier have been identified as important causes of nephrotic syndrome, with one of these being podocin, coded by NPHS2 gene. NPHS2 mutations are the most frequent genetic cause of steroid resistant nephrotic syndrome. The aim of this review is to update the list of NPHS2 mutations reported between June 2013 and February 2017, with a closer look to mutations occurring in Latin American countries.


Introduction
In the high-throughput sequencing era, new candidate genes associated with monogenic and genetic heterogeneous diseases such as steroid resistant nephrotic syndrome (SRNS) are piling up [1][2][3][4][5][6]. However, mutations in the three main genes (NPHS1, NPHS2, and exons 8 and 9 of WT1 gene) are still the most frequent molecular cause of SRNS in childhood and adolescence. More than 200 NPHS2 (OMIM * 604766) gene mutations are registered in HGMD Professional 2017.1 (http://www.hgmd.cf.ac.uk) and 127 in the HGMD Public 2017.1 associated with familial and sporadic forms of SRNS.
SRNS is one of the most common kidney pathologies in childhood, being characterized by proteinuria, edema, and hypoalbuminemia. The most frequent renal histological feature associated with SRNS is focal segmental glomerulosclerosis (FSGS). Almost 40% of SRNS/FSGS children develop end-stage renal disease (ESRD) before adulthood and may receive a kidney transplant, with a 10 to 50% risk of recurrence of FSGS in the allograft kidney [7][8][9]. Although the pathogenesis of NS is not yet completely understood, much has been learnt about the glomerular filtration barrier (GFB) which is composed of three layers: the fenestrated capillary endothelial cells; the glomerular basement membrane (GBM); and the podocytes, specialized cells with interdigitating foot processes that are interconnected to form a slit diaphragm (SD) membrane, a multiprotein signaling complex that controls the ultrafiltration in this dynamic structure [10]. Nephrotic protein leakage may occur as a result from damage in one of these GFB components [11], although functional pathways specifically in the podocyte have revealed this cell as the key component of the pathogenesis of SRNS [5,6,12,13].
More than 50 genes have been identified so far, associated with SRNS of congenital (0-3 months), infantile (4-12 months), childhood/adolescence (1-18 years), or adult onset [6]. In 2007, Hinkes et al. [14] [5] screened 27 genes associated with SRNS in 1783 families from an international cohort study (107 members from 30 countries). The eight larger contributing centers were Germany, Switzerland, Turkey, Egypt, Saudi Arabia, Los Angeles, Ann Arbor, and India but many other centers participated, including Argentina representing South America with 16 families. The main conclusion was that a monogenic cause was detected in 29.5% of the SRNS cases (0-25 years) in one of the 27 genes analyzed. After a detailed analysis of mutation distribution by gene and by age of onset, NPHS2 mutations were the most frequent (5.7% to 12.7%) in patients with SRNS onset between 1 and 18 years old.
Mutations in NPHS2 gene, located at 1q25-31, are the most common cause of SRNS in childhood and were first described by Boute et al. (2000) [15]. NPHS2 coding region encompasses 1,149 bp, has 8 exons, and encodes a 383-aminoacid protein with 42 kD, called podocin, which is expressed in fetal and mature kidney [15]. Podocin is predicted to have a hairpin-like structure, with both C-and N-terminal domains facing cytosol and one short transmembrane domain [16]. In addition to its role in anchoring nephrin and CD2AP (OMIM * 604241) to the SD, podocin forms homoligomer complexes that bind with cholesterol in lipid rafts, where it may act as a scaffolding/targeting/signaling protein [17][18][19]. NPHS2 mutations initially found in autosomal-recessive inheritance familial cases [15] and further in sporadic SRNS cases [20] as well represent 40% and 6-17% of these cases, respectively [12,[21][22][23]. In 2013, Bouchireb et al. [24], in a detailed and complete review, presented a list of NPHS2 mutations, polymorphisms, and variants of unknown significance published from October 1999 to September 2013. In that review they identified 25 novel pathogenic mutations in addition to the 101 already registered in the mutation database at that time. The mutations were distributed along the entire NPHS2 gene, with no preferential hotspot.
The aim of this review is to update the list of NPHS2 mutations reported in the last few years around the world, with a closer look to mutations occurring in Latin American countries.

NPHS2 Mutations Overview
We searched for articles reporting NPHS2 mutations associated with SRNS in childhood and adolescence that were published from June 2013 to February 2017. The key words "NPHS2", "NPHS2 mutations", "podocin" and "steroid resistant nephrotic syndrome genetics" were used in PubMed databank. We further looked for variants/mutations that were not annotated in open-access databases such as public HGMD (http://www.hgmd.cf.ac.uk/ac/index.php) and GnomeAD Browser (http://gnomad.broadinstitute.org) or in Leiden Open Variation Database (https://www.lovd.nl/ NPHS2). For nonsynonymous variants, we classified them as deleterious or benign according to in silico prediction tools (PolyPhen-2 and SIFT) [25,26]. For splicing variants, we performed splice-site prediction by BDGP neural network [27].
Thirty-nine variants, among them 25 missenses, four nonsenses, three splice-sites, four frameshifts, and three in  [43]. Ten out of those 39 mutations were unique and had not been annotated in public HGMD (http://www.hgmd.cf.ac .uk/ac/index.php) or in GnomeAD Browser (http://gnomad .broadinstitute.org) or in the Leiden Open Variation Database (https://www.lovd.nl/NPHS2): six were missenses, three were located in splice-site regions, and two were frameshifts (Table 1). We searched GnomaAD Browser for all the variants compiled in Table 1 as well as for other variants such those reported in Mexico (p.Leu142Pro), Chile and Brazil (p.Ala284Val), and Brazil (p.Val260Glu), observing frequency, racial ethnicity, and geographical provenience. Only p.Ala284Val and p.Val260Glu were registered at GnomaAD Browser, but no allele counted was from Latin America population for both of them.

Missense Mutations
Four out of the five missenses were described in homozygosity in South Indian SRNS patients (p.Ser46Pro, p.Leu167Pro, p.Pro175Ser, and p.Pro316Ser) [29,30]. The fifth missense, p.Leu139Arg, was identified in two Mexican children with NS, one SRNS, and one SSNS [41]. As in silico predictions were not performed for those variants in their original publications, we investigated their pathogenicity by predictive tools available, such as SIFT and PolyPhen-2. Both p.Leu139Arg and p.Pro316Ser variants were predicted as damaging by SIFT and PolyPhen-2 (Table 1). At the moment, these five missense variants should be considered as variants of unknown significance and only after proper functional studies they can be associated with SRNS.

Splice-Site Mutations
Three splice-site mutations that were not registered in any of the three searched databases have been identified ( Table 1). The homozygous splice-site mutation c.451+3A>T whose effect on podocin protein was evaluated by renal mRNA analysis demonstrated exon 3 skipping that led to a premature termination codon (p.Val128Phefs * 28). This mutation was originally identified by Benetti et al. [32] in an Italian girl with SRNS. Either c.535-1G>A or c.738+2T>C were described by Wang et al. [28] in compound heterozygosis with another known NPHS2 mutation in two nonrelated SRNS Chinese children. They evaluated the conservation of variant sites using PhyloP Primates tool that resulted in scores of 4.481 and 3.839 for c.535-1G>A and c.738+2T>C, respectively, indicating high degree of conservation at these sites. They also evaluated those variants using Mutation Taster that classified  them as damaging. Nevertheless, they did not use the BDGP Splice Site Prediction by Neural Network in silico tool to predict the splicing recognition sites; therefore, we performed this analysis and the resulting prediction is shown in Table 1.

Frameshift Mutations
Two small deletions were described: one (p.Ser329 = fs * 14) was found in heterozygosis in five SRNS individuals from the same Finnish family, with early-onset, slow progression, and dominant inheritance pattern [38]; the other (p.Lis239Argfs * 13) was identified by our group in two Brazilian sisters with early-onset SRNS in association with the p.Val260Glu missense [43].

NPHS2 Mutations in the World with a Closer Look to Latin America
Population studies from different countries, mainly from Europe, South Asia, and North America, revealed that the prevalence of NPHS2 mutations in children with SRNS may vary according to ethnicity. It appears to be frequent among Americans and Turkish [21,44] (26% and 24.7%, resp.) but not as frequent among Greek [45], Chinese [28,46], Indian [47], Japanese [40,48], Pakistani [49], and Korean [50] patients (9%, 4.3%, 4%, 4%, 3.4%, and 0%, resp.). Recently a large multicentric study was performed with samples from 1783 SRNS families from eight contributing centers in which twenty-seven SRNS associated genes were sequenced. Disease-causing mutations were identified in different genes; however, mutations in NPHS2 were more frequent [5,13].
Some of them, with a high frequency in particular geographical regions, are considered as founder alleles for NPHS2: p.Arg138Gln and p.Gly140Aspfs * 41 are predominant in Europe; p.Pro118Leu in Turkey; p.Val180Met in North Africa; p.Arg138 * in Israel and Arabian countries; p.Val260Glu in Oman, Arabia; and p.Met1? and Asn199Lysfs * 14 in Egypt [5,15]. The contribution of Latin American countries to genetic studies in SRNS is scarce. Searches for NPHS2 mutations had been performed mainly in three countries: Mexico, Chile, and Brazil ( Figure 1). In Mexico, only the 3rd exon of NPHS2 was sequenced in eight SRNS and five SSNS children [41]. The heterozygous p.Leu139Arg variant was identified in two patients, one SRNS and one SSNS; therefore, it was considered as a variant of unknown significance. In Chile, Azocar et al. [42] performed a molecular study in SRNS children and found NPHS2 mutations in 21%. The mutations identified were homozygosis for p.Pro341Ser in one patient and compound heterozygosis for p.Arg229Gln and p.Ala284Val in six patients [42,51]. In Brazil, our group performed the molecular analysis of NPHS2 in 27 SRNS children and identified disease-causing mutations in 14  et al. [52] also carried the [p.Ala284Val];[p.Arg229Gln] association, with one-half presenting the adult onset form of the disease. The p.Val260Glu variant is worth mentioning in this group, which is already considered as a founder allele in Oman, Arabia [5], and also identified in one of our familial cases. We are not aware of an Arabian ancestrality of this family, but given the highly miscegenated nature of the Brazilian population, we cannot exclude this possibility.
This review aimed to give a new perspective to Nephrotic Syndrome in Latin American countries, emphasizing the importance of implementing the molecular evaluation of NS, especially investigating mutations on those genes more frequently associated with SRNS in this region. The molecular characterization of SRNS in childhood and adolescence is relevant to guide further treatment, since patients bearing NPHS2 mutations may be spared of the undesired side effects of corticosteroids. Additionally, living donor transplantation might be considered since SRNS patients with homozygous or compound heterozygous mutations in NPHS2 have reduced risks for recurrence of FSGS after renal transplant compared with children without mutations.

Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.