Multiple rare genetic variants co‐segregating with familial IgA nephropathy all act within a single immune‐related network

Abstract Background IgA nephropathy (IgAN) is a common complex disease with a strong genetic involvement. We aimed to identify novel, rare, highly penetrant risk variants combining family‐based linkage analysis with whole‐exome sequencing (WES). Methods Linkage analysis of 16 kindreds of South Italian ancestry was performed using an ‘affected‐only’ strategy. Eight most informative trios composed of two familial cases and an intrafamilial control were selected for WES. High‐priority variants in linked regions were identified and validated using Sanger sequencing. Custom TaqMan assays were designed and carried out in the 16 kindreds and an independent cohort of 240 IgAN patients and 113 control subjects. Results We found suggestive linkage signals in 12 loci. After sequential filtering and validation of WES data, we identified 24 private or extremely rare (MAF <0.0003) linked variants segregating with IgAN status. These were present within coding or regulatory regions of 23 genes that merged into a common functional network. The genes were interconnected by AKT, CTNNB1, NFKB, MYC and UBC, key modulators of WNT/β‐catenin and PI3K/Akt pathways, which are implicated in IgAN pathogenesis. Overlaying publicly available expression data, genes/proteins with expression notably altered in IgAN were included in this immune‐related network. In particular, the network included the glucocorticoid receptor gene, NR3C1, which is the target of corticosteroid therapy routinely used in the treatment of IgAN. Conclusion Our findings suggest that disease susceptibility could be influenced by multiple rare variants acting in a common network that could provide the starting point for the identification of potential drug targets for personalized therapy.

In Family 4 we found that out of the three validated variants (IFNA21 9:g21165905c>t; ATRAID 2:g27439820a>g; RPUSD3 3:g9880772t>c). ATRAID and RPUSD3 segregated perfectly family members with IgAN and with persistent microscopic hematuria in family members 215,202,198. The IFNA21 variation, on the other hand, was present in only one of the family members with microscopic hematuria (202).
In family 7, we evaluated the segregation of the variant 1:g25894878c>g within the LDLRAP1 gene. This variant was absent in family members that had normal urinalysis (579,2073,1202) and was carried by a family member who presented some episodes of microscopic hematuria (581).
For family 206, the variants THRA 17:g38233146c>t, UBE2G1 17:g4173166g>a and CDC27 17:g45197967a>g that segregated with the affection status also segregated with an individual characterized by persistent microscopic hematuria (1858), while only UBE2G1 gene variant segregated with a second individual characterized by persistent microscopic hematuria.
In family 385, the CAMKD2 4:g114374628t>a and the CHD5 1:g6162250g>gac variant was evaluated for segregation in the extended family. This variant was present in family members with persistent microscopic hematuria and/or proteinuria and in the subject 2162, an obligate carrier with a normal phenotype.
For family 483, we evaluated the co segregating variant 3:g5259973a>g within the EDEM1 gene. This variant was absent in family members who had normal urinalysis apart from the obligate carrier 2542 with normal urinalysis.

Sample donors
The genome-wide linkage analysis involved 34 biopsy-proven familial IgAN patients and 112 relatives from 16 Italian kindreds of South Italian ancestry (Table 1, Figure S1). moderate, low and modifier) and the functional impact of coding variants was also predicted. Low impact variants were predicted by snpEFF and filtered out as they were synonymous coding and "assumed to be mostly harmless or unlikely to change protein behaviour" as decribed in manual (http://snpeff.sourceforge.net/SnpEff_manual.html).
Sequence data were filtered against multiple databases, using annovar  Figure 1. Sixteen multiplex families included in the linkage study, the red bars represent 146 genotyped subjects. Squares and circles represent males and females, respectively; arrows indicate probands and the slash indicate deceased individuals. Filled and unfilled symbols indicate IgAN affected and unaffected individuals, respectively. The symbols with a dot indicate individuals who have not received or have discordant urinalysis. Symbols with a vertical line indicate individuals with documented urinary abnormalities (persistent microscopic hematuria and/or proteinuria). Families 1, 3, 4, were present in the first linkage study perfomed by Gharavi AG et al (Nat Genet 2000;26: 354-7). Families 3,4,7,15,16,206,343,344,385,386,393 were also present in our previous linkage study perfomed by Bisceglia L et al ( Am J Hum Genet 2006;79: 1130-4). Families 14,17,36,483 were new and added exclusively for this study.

Supplementary Figure 2.
Plot of LOD score statistics from NPL analysis for the chromosomes in which the score exceeded the 1.5 level.

Supplementary Figure 3. Pedigrees included in the exome sequencing study
Eight pedigrees were included in the exome sequencing study. Squares and circles represent males and females, respectively; the arrows indicate probands and the slashes indicate deceased individuals. The two variants predicted within the BCLAF1 actually determined by the presence of a deletion in a r by a G/A substitution). Heterozygous deletion of the BCLAF1 gene. Deletion detected in the flipped reverse sequence of two affected individuals (552 ,156 ) and absent in the intra BCLAF1 gene were deletion in a repeated deletion found in the in the flipped reverse and absent in the intra-familal control Supplementary Figure 7. unrelated HBD. We found that region of the BCLAF1 gene (control1 , Control2, control3) for this reason we decided to exclude this variation from further analysis . We evaluated BCLAF1 complex genomic region in six unrelated HBD. We found that the heterozygous deletion in the repeated (AAAAAC)n BCLAF1 gene was also found in three unrelated healthy blood and absent in another three unrelated healthy blood for this reason we decided to exclude this variation from further analysis complex genomic region in six repeated (AAAAAC)n ealthy blood donors unrelated healthy blood donors , for this reason we decided to exclude this variation from further analysis. Ref = reference allele; Alt = alternate allele; CADD SCORE: PHRED-like (-10*log10(rank/total)) scaled C-score ranking a variant relative to all possible substitutions of the human genome (8.6x10^9)