Whole genome sequence analysis identifies a PAX2 mutation to establish a correct diagnosis for a syndromic form of hyperuricemia

Hereditary hyperuricemia may occur as part of a syndromic disorder or as an isolated nonsyndromic disease, and over 20 causative genes have been identified. Here, we report the use of whole genome sequencing (WGS) to establish a diagnosis in a family in which individuals were affected with gout, hyperuricemia associated with reduced fractional excretion of uric acid, chronic kidney disease (CKD), and secondary hyperparathyroidism, that are consistent with familial juvenile hyperuricemic nephropathy (FJHN). However, single gene testing had not detected mutations in the uromodulin (UMOD) or renin (REN) genes, which cause approximately 30-90% of FJHN. WGS was therefore undertaken, and this identified a heterozygous c.226G>C (p.Gly76Arg) missense variant in the paired box gene 2 (PAX2) gene, which co-segregated with renal tubulopathy in the family. PAX2 mutations are associated with renal coloboma syndrome (RCS), which is characterized by abnormalities in renal structure and function, and anomalies of the optic nerve. Ophthalmological examination in two adult brothers affected with hyperuricemia, gout, and CKD revealed the presence of optic disc pits, consistent with optic nerve coloboma, thereby revising the diagnosis from FJHN to RCS. Thus, our results demonstrate the utility of WGS analysis in establishing the correct diagnosis in disorders with multiple etiologies.

Here, we report a kindred considered to have FJHN on the basis of hyperuricemia, gout, reduced FEUA, and CKD, but in whom Sanger DNA sequence analysis had not detected mutations of UMOD or REN, which account for approximately 30-90% of cases.
However, whole genome sequence (WGS) analysis unexpectedly revealed that a mutation of the paired box 2 (PAX2) gene was the likely cause of FJHN in this kindred, which prompted clinical reassessment of the family.

| Editorial policies and ethical considerations
Informed consent and venous blood samples were obtained from nine available members (comprising five affected and four unaffected members) of the family with suspected FJHN, using protocols approved by the Multicentre Research Ethics Committee (UK) (MREC/02/2/93), and local ethics committees (Austria).

| Patients and clinical findings
The proband (Figure 1a, individual II.1), a 57-year-old man, presented with hyperuricemia with reduced FEUA at 32 years of age, and later developed CKD and secondary hyperparathyroidism (Table 1), consistent with FJHN. Histological analysis of a single glomerulus from a kidney biopsy taken at the age of 32 years was suggestive of glomerulonephritis but was considered inconclusive as other glomerula were not present among the biopsy sections to confirm this finding. Electron microscopy of the single glomerulus showed that it was abnormal with segmental lobe collapse, basal membrane ruptures, and segmental sclerosis with numerous tubular-reticular structures. At 53 years of age he had an elevated serum creatinine of 4.2 mg/dl [normal range (NR) = 0.5-1.2 mg/dl], proteinuria of 2,500 mg/g creatinine (NR <110 mg/g), albuminuria of 1,655 mg/g creatinine (NR <3 mg/g), and a reduced FEUA of 4.5% (NR = 7.5 ± 1.8%). He was treated with ramipril 5 mg/day, calcitriol 0.25 μg/day, cholecalciferol 12,000 IU/ week, allopurinol 100 mg/day, and bicarbonate 2,500 mg/day. Two years later peritoneal dialysis was started due to end-stage kidney disease. The proband's brother (individual II.2) was also affected, and presented at the age of 44 years with gout. Clinical evaluation revealed: renal insufficiency with elevated serum creatinine of 1.8 mg/dl; recurrent attacks of gout, hyperuricemia and a reduced FEUA of 4.7%; and proteinuria and albuminuria of 740 and 323 mg/g creatinine, respectively (Table 1). He was treated with ramipril 5 mg and allopurinol 150 mg/day. The proband's father (individual I.1) had chronic renal failure, with serum creatinine of 1.3 mg/dl, and proteinuria of 1,000 mg/g creatinine ( Table 1). The proband's younger brother (individual II.4) had mild albuminuria of 34 mg/g creatinine, and his niece (individual III.3) had albuminuria of 689 mg/g creatinine and proteinuria of 910 mg/g creatinine ( Table 1). The albuminuria observed in patients II.1, II.2, and III.3 was considerably higher than that reported previously in other patients with FJHN (Eckardt et al., 2015;Lee, Kim, Oh, Noh, & Lee, 2010). Mutational analysis of the UMOD and REN genes using leukocyte DNA from the proband did not detect any abnormalities. abnormalities, and also an absence of abnormalities within the SEC61A1 and HNF-1β genes that have been reported to be associated with FJHN. Futhermore, copy number variants (CNVs) were not identified in these four genes, and an examination of all rare (allele frequency <3%) variants in these genes also did not reveal any deleterious alleles to be shared by the two affected brothers, II.1 and II.2 (Table S1). CNVs in three other genes (LINC01060, NRG3, and PMM2) were found (Table S2), but were not further investigated as they were highly unlikely to be causative of the phenotypic abnormalities. However, WGS analysis identified a heterozygous G-to-C T A B L E 1 Clinical details of affected and unaffected members of the kindred with chronic kidney disease (CKD) Proteinuria (mg/g creatinine) (NR <110 mg/g) 1,000 2,500

| DISCUSSION
Our study reports a kindred affected with CKD, reduced FEUA, hyperuricemia, and gout, which were consistent with a diagnosis of FJHN.
However, the kindred did not have UMOD, REN, SEC61A1, or HNF-1β gene mutations, which collectively are associated with approximately 30-90% of FJHN cases, but instead had a missense mutation (p.Gly76Arg) of PAX2, whose abnormalities are more commonly associated with RCS. Indeed, ophthalmic examination, prompted after the identification of the PAX2 mutation by WGS, identified optic nerve abnormalities consistent with RCS, in all four affected family members that were available for ophthalmic assessments (Figure 1a-e).
RCS is characterized by abnormalities in renal structure and function in greater than 90% of patients, ophthalmological anomalies in greater than 75% of patients, and hearing loss in less than 10% of patients (Bower et al., 2012). The most common renal findings are renal hypodysplasia, vesicoureteral reflux (VUR), renal cysts, and multicystic dysplastic kidneys, which occur in 65%, 15%, <10%, and 5% of patients, respectively. Renal failure is reported in approximately 15% of cases, while CKD stage 5 requiring a kidney transplant is common and has a range of onset from birth to greater than 75 years of age (Bower et al., 2012). The ophthalmoscopic findings include optic nerve coloboma, optic disc dysplasia, excavation of the optic disc or optic disc "pits," morning glory anomaly, and hypoplastic optic discs, which occur in 50%, >10%, <10%, 5%, and <5% of patients, respectively (Bower et al., 2012). Retinal, macular, and lens abnormalities have also been reported in some patients (Bower et al., 2012). PAX2 is expressed in other tissues (e.g., cerebellum, hypothalamus otic vesicle, genitourinary tract, and pancreas), and additional features of RCS include CNS anomalies, intellectual disability and elevated pancreatic amylase (Bower et al., 2012).
A frameshift deletion of PAX2 in a family with optic nerve colobomas, renal hypoplasia and VUR (Sanyanusin et al., 1995) represents the first reported single gene defect causation of congenital anomalies of the kidney and urinary tract (CAKUT). Subsequently, larger patient cohort studies confirmed PAX2 mutations as an important cause of syndromic CAKUT and the establishment of RCS as a separate disease entity (Madariaga et al., 2013;Rossanti et al., 2020;Thomas et al., 2011;Weber et al., 2006). PAX2 is a member of the paired box (PAX) family of transcriptional regulatory genes with nine members described in humans. The majority of PAX2 pathogenic mutations are located in the paired domain (comprising a conserved 128 amino acid region) that has DNA binding properties encoded by exons 2-4 (Bower et al., 2012;Eccles et al., 2002). However, evidence from an international consortium of three laboratories collecting data on PAX2 mutations in RCS patients reported that there are no clear genotype/ phenotype correlations, and variable types of PAX2 mutation (missense, frameshifts, splice sites, and deletions) located across 10 of the 12 PAX2 exons can lead to similar phenotypes, while the same mutation within members of the same family can have variable penetrance and manifestations of RCS (Bower et al., 2012). This large intrafamily variability in RCS suggests that factors other than PAX2 may play a role in clinical penetrance (Bower et al., 2012). PAX2 mutations are found in approximately 50% of RCS/PAPRS, thereby suggesting that other abnormalities of genes may be involved in the etiology of this disorder (Dureau et al., 2001;Okumura et al., 2015).
The presence of optic disc pits and dysplastic papilla in the family reported here (Figure 1a,e) is a distinguishing feature confirming RCS from FJHN given that CKD is common to both. This family also has reduced FEUA, hyperuricemia and gout that are commonly found in FJHN. Such occurrence of RCS with hyperuricemia and gout, has been previously reported in only two unrelated families (Megaw et al., 2013).
One family, which had a PAX2 frameshift mutation [c.567_568dup (p. Ile190ArgfsX85)] in exon 5, consisted of five affected males from three generations; all the five affected males suffered from hyperuricemia and/or gout and the proband also suffered from diabetes mellitus and cryptorchidism, which have not previously been associated with RCS (Megaw et al., 2013). RVT is a Wellcome Trust Investigator and NIHR Senior Investigator.

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS
Mark Stevenson: designed this study, acquired data, analyzed and interpreted data, wrote the first draft of the manuscript; Karl Lhotta: designed this study, acquired data, analyzed and interpreted data; Rajesh V. Thakker: designed this study, analyzed and interpreted data, wrote the first draft of the manuscript; Silvia Reichart: acquired data; Charlotte Philpott: acquired data; Kate E Lines: acquired data; Caroline M Gorvin: acquired data; OxClinWGS: provided bioinformatic analysis of WGS data; Alistair T. Pagnamenta: provided bioinformatic analysis of WGS data, analyzed and interpreted data; Jenny C. Taylor: provided bioinformatic analysis of WGS data, analyzed and interpreted data. All coauthors participated in the preparation of the manuscript by reading and commenting on the draft prior to submission.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author on reasonable request.