Whole genome sequencing identifies monogenic disease in 56.1% of families with early-onset steroid-resistant nephrotic syndrome

Abstract

Genetic causes of steroid-resistant-nephrotic-syndrome (SRNS) represent a rapidly growing number of monogenic diseases. The reported diagnostic yield of various studies applying genetic panels and exome-sequencing to diagnose SRNS is usually < 30%. We performed genome-sequencing in a cohort of Egyptian SRNS patients. We recruited 47 SRNS patients belonging to 41 unrelated families [28 males/19 females; median (range): 6 (0.5-22 years)]. We established a pipeline for genome sequencing, bioinformatics analysis, variant curation and protein modeling at the Egypt Center for Research and Regenerative Medicine (ECRRM). Disease-causing variants were detected in 27/47 patients (57.4%) belonging to 23/41 families (56.1%), including nine novel variants in NPHS1, NPHS2, COL4A3, MYO1E, NUP93, PLCE1, PODXL, SMARCAL1 and WT1. Novel variants were confirmed by Sanger sequencing and were segregated in families of affected patients. NPHS2 was the most common causative gene in 8/23 (34.8%) of confirmed families, followed by NPHS1, WT1, and SMARCAL1 in 2/23 families (8.7%) each. All detected missense variants were evaluated through protein modeling and were predicted deleterious. Our study expanded the spectrum of SRNS disease-causing variants and revealed a monogenic cause in 56.1% of investigated families. In our cohort, no deep intronic or regulatory variants were detected by genome-sequencing. Pursuing genetic diagnosis in SRNS patients is crucial to inform clinical decision making, genetic counseling, transplantation strategy and prenatal diagnosis thus improving clinical outcome of affected patients.

Fig. 1

Representation of family segregation by Sanger sequencing for novel variants detected in the study. Both parents were available for sequencing in all families except two families: the father of family 19, who passed away and had a history of chronic renal failure and renal transplantation, and the father of family 26, who was healthy but unavailable for sampling. Red arrows indicate the base change location. Het, heterozygous; hom, homozygous. Variants in MYO1E (family 1), WT1 (family 19), NPHS2 (family 26) and NPHS1 (family 28) are sequenced in the reverse strand

Fig. 2

Representation of protein modeling AA changes of the ten detected missense variants in seven genes in the study. A NPHS2, B NPHS1, C LMX1B, D NUP93, E ALOX12B, F CD2 AP, G WT1.This figure illustrates the 3D protein models highlighting the structural impact of missense variants identified in the study. Each panel A–G corresponds to a specific gene and its associated variant(s), with key amino acid residues labeled to show their positions and interactions within the protein structure. Red and yellow ribbons represent the protein backbone, while green and white surfaces indicate the surrounding protein environment. Blue and purple sticks highlight the side chains of the mutated residues and their interacting partners. Arrows point to the specific amino acid changes, with zoomed-in insets providing a closer view of the local structural environment. These interactions are critical because changes in amino acid properties such as charge, size, or hydrophobicity can disrupt local protein folding, stability, or binding interfaces. For example: NPHS2 (Panel A): The p.(Arg168 Cys) substitution replaces a positively charged arginine with cysteine, potentially disrupting ionic interactions or forming aberrant disulfide bonds (Cys168). NUP93 (Panel D): The p.(Tyr185 Cys) variant replaces a bulky tyrosine with cysteine, likely perturbing the hydrophobic core or introducing a disulfide bond (inset). CD2 AP (Panel F): p.(Lys301Met) replaces a positively charged lysine with methionine, potentially impairing electrostatic interactions critical for cytoskeletal binding. WT1 (Panel G): p.(Glu234 Arg) inverts charge in a zinc-finger DNA-binding domain, suggesting altered transcriptional activity