#1346 Population modelling depicts the mutational burden of NPHS2 (podocin) nephropathy and reveals an undiagnosed later onset genetic cohort

Abstract Background and Aims A better understanding of the disease burden is essential to encourage development of therapies for monogenic nephrotic syndrome and improve genetic screening strategies. Here, population modelling was used to estimate pathogenic genotype frequencies in NPHS2 (encoding podocin), the most common gene implicated in childhood nephrotic syndrome, across the US, UK, Europe and Japan. These were then compared to real-world and study data. Method Over 1.2 million genomes from five databases (NIH TopMED, NIH ‘All of us’, UK Biobank, gnomAD, TogoVAR) were included for population modelling. Pathogenic variants were assigned if in agreement with ClinVAR, according to ACMG criteria, or if these were pathogenic based on predictions from 4 independent in silico algorithms. Genotype frequencies were estimated assuming Hardy-Weinberg equilibrium. Genotypes were assigned to phenotypic groups of slow, intermediate or rapidly progressing disease. Modelled genotype frequencies were compared to cohorts from Natera Renasight, a 385 gene panel used for clinical renal genetic testing in the US, and the DUPLEX study (NCT03493685), which recruited adults and children with focal segmental glomerulosclerosis (FSGS). Results Equilibrium modelling confirmed high frequency of p.R138Q in Northwestern Europe associated with the traditionally expected early onset disease (Fig. 1A). By contrast an unexpectedly high proportion (77%) of the cases in the US were attributed to p.R229Q compound heterozygosity, associated with adult onset and slow-progression (Fig. 1B). Half of the cases are modelled to occur in trans with the p.A284V variant (Fig. 1C), which is prevalent among individuals with admixed American genetic ancestry (previously Latin American) (Fig. 1C). 92,984 US patients were sequenced on Natera's Renasight Panel and 355 DUPLEX FSGS patients were genotyped. 67 NPHS2 patients (17 homozygous, 50 compound heterozygotes) were identified on the Natera panel, with the most common genotypes being p.R229Q/p.A284V (37%) and p.R229Q/other (21%) (Fig. 2A). 15 NPHS2 (1 homozygous, 14 compound heterozygotes) patients were identified in DUPLEX, with the most common pathogenic genotypes being p.R229Q/other (47%), followed by p.R229Q/p.A284V (33%) (Fig. 2B). The Natera and DUPLEX data confirm the trends in modelled genotype frequencies, but Natera had a smaller proportion of p.R229Q patients (58%) than the model and was identified to be in disequilibrium (chi-square test, P < 0.0001) (Fig. 2A). Conclusion Here, population modelling based on large human genetic diversity cohorts was used to estimate the frequency of pathogenic NPHS2 variants across US, UK, Europe and Japan. These results reveal a larger than expected p.R229Q mutational burden, associated with late onset disease, confirmed with patient data. In particular in the US, but not in Japan, a specific late-onset subgroup of patients due to NPHS2 p.R229Q compound heterozygosity is likely largely undiagnosed, with implications for genetic screening in adult patients, avoidance of unnecessary immunosuppression, and development of targeted therapies for NPHS2 driven nephropathy.