Molecular diagnostic results of a nephropathy gene panel in patients with suspected hereditary kidney disease

Abstract

Objective

Clinical diagnosis of hereditary kidney disease can be difficult because of its rarity and severe phenotypic variability. Identifying mutated causative genes can provide diagnostic and prognostic information. In this study, we report the clinical application and outcome of a next-generation sequencing–based, targeted multi-gene panel test for the genetic diagnosis of patients with hereditary kidney disease.

Methods

A total of 145 patients evaluated for hereditary kidney disease who underwent a nephropathy panel with 44 different genes were retrospectively reviewed and included in the study.

Results

Genetic diagnosis of other hereditary kidney diseases, particularly autosomal dominant polycystic kidney disease, was made in 48% of patients. The nephropathy panel changed the preliminary diagnosis in 6% of patients. The variants in 18 (12%) patients had not been previously reported in the literature.

Conclusion

This study demonstrates the utility of the nephropathy panel in identifying patients diagnosed with hereditary kidney disease who are referred for genetic testing. A contribution was made to the variant spectrum of genes associated with hereditary kidney disease.

Introduction

Hereditary kidney disease is the leading cause of chronic kidney disease and accounts for at least 10% of end-stage renal disease (ESRD) cases in Europe. The most common hereditary kidney diseases are cystic and glomerular nephropathies.¹ The clinical and genetic heterogeneity of hereditary kidney diseases poses a major diagnostic challenge. Accurate knowledge of the genetic cause of a disease is becoming increasingly important for treatment, management, prognosis, and genetic counseling. For example, Alport syndrome can be treated with angiotensin-converting enzyme (ACE) inhibition to reduce proteinuria and delay deterioration of renal function. In hereditary podocytopathy causing proteinuria, rapid renal transplantation is recommended because ACE inhibition and immunosuppression are less likely to reduce proteinuria.^2,3

One of the benefits of genetic diagnosis is that it can determine the cause of the disease and shorten the diagnostic pathway, regardless of the disease stage, unlike invasive renal biopsies that cannot provide a diagnosis for very early or late stages of the disease.⁴ Due to recent advances in genetic diagnostics, the known number of genes causing various diseases is increasing. This trend can also be seen in the field of nephrology; genes have been discovered that cause various diseases such as renal tubular disorders and hereditary nephrotic syndrome. In recent years, many next-generation sequencing (NGS) diagnostic panels containing these genes have been developed.⁵

This study presents information on genes and variants identified by the nephropathy gene panel in a cohort of 145 index cases with clinically suspected hereditary kidney disease.

Materials and Methods

Patient Cohort

A total of 145 patients with a diagnosis of hereditary kidney disease who presented to the Medical Genetics Department of Bursa Yuksek Ihtisas Training and Research Hospital between 2015 and 2019 were retrospectively evaluated. The patients included those diagnosed with polycystic kidney disease, renal tubular acidosis, nephropathy, chronic renal failure (CRF), Bartter syndrome, Gitelman syndrome, Alport syndrome, focal segmental glomerulosclerosis (FSGS), nephrocalcinosis, and nephrogenic diabetes insipidus (DI). Informed consent for genetic testing was obtained from all patients or their legal guardians. The local ethics committee granted approval for the study (2011-Clinical Research Ethics Committee [CREC]-25 2019/08-01).

Method

An NGS platform (NextSeq 500, Illumina) was used for this study. The nephropathy panel (Nephropathies Solution, Sophia) was studied in patients, and all steps were performed according to the original manufacturer’s protocol. The custom panel, Nephropathies Solution, is a capture-based target enrichment kit. Paired-end sequencing was performed on an Illumina NextSeq 500 system with a read length of 150 × 2. Base calling and image analysis were conducted using Illumina’s Real-Time Analysis software. All bioinformatics analysis was performed on the Sophia DDMTM platform, which includes algorithms for alignment, calling single nucleotide variants (SNVs), and small insertions/deletions (indels) (Pepper), calling copy number variations (Muskat), and functional annotation (Moka). Raw reads were aligned to the human reference genome (GRCh37/hg19). The nephropathy pipeline applies default filters related to low coverage, where the cutoff is 50×, whereas the variant fraction cutoff for SNV is 20% and 15% for indels; a cutoff for homopolymer regions with a length of 10 bp was also applied. For variant calling, a minimum coverage of 50× is recommended. The raw data obtained were filtered and analyzed using the appropriate program (Sophia DDM). Considering the demographic characteristics, clinical findings, and family history of the patients, variants that could be significant were determined. These significant variants, which were detected during the analysis of the nephropathy panel and could be associated with any disease, were evaluated using the Human Gene Mutation Database (HGMD).⁶ This allowed determination of whether the change had been reported in the literature and, if so, to which disease it was associated. For alterations not reported in the literature, classification by American College of Medical Genetics and Genomics (ACMG) criteria and frequency in population studies (gnomAD, Genome Aggregation Database) were determined using the Varsome Analysis Program (https://varsome.com/).^7,8 This panel includes many disease-associated genes, including polycystic kidney disease, Bartter syndrome, Gitelman syndrome, Alport syndrome, FSGS, nephrocalcinosis, and nephrogenic DI. The gene content of the panel is shown in TABLE 1. All coding exons and exon-intron boundaries (±15 bp) of 44 genes were sequenced. Missense, nonsense, frameshift, splice variants, and short indels were detected by software. Variants in deep intronic regions and gene rearrangements (gene fusions or inversions) could not be detected. Large insertions or duplications that are larger than ~1/3 read length might be missed.

Results

The demographic characteristics of the 145 patients with hereditary kidney disease included in the study are given in TABLE 2. Consanguinity was most common in patients with Gitelman and Bartter syndromes.

According to the ACMG classification, only pathogenic and likely pathogenic variants were considered in the evaluation of diagnostic yield. Variants with a low minor allele frequency according to the gnomAD and variants of uncertain significance (VUS) according to ACMG that are thought to be clinically relevant are predicted to be pathogenic variants in in silico prediction tools and are also listed in TABLE 3.^7,8

Causal variants (pathogenic and possibly pathogenic) were detected in 70 of the 145 patients who underwent the nephropathy gene panel. Detailed information on the variants is presented in TABLE 4. Nineteen of these variants had not been previously described in the literature. A causal variant was detected in 37 (58.7%) of 63 patients referred with a clinical diagnosis of polycystic kidney disease. Causal variants were detected in 6 of 9 patients (66.6%) with a clinical diagnosis of Bartter syndrome, 8 of 11 patients (72.7%) with a clinical diagnosis of Gitelman syndrome, 5 of 11 patients (45.4%) with a clinical diagnosis of Alport syndrome, 2 of 3 patients (66.6%) with a clinical diagnosis of nephrogenic DI, 4 of 8 (50%) patients with a clinical diagnosis of CRF, 3 of 17 patients (17%) with a clinical diagnosis of nephrotic syndrome, 1 of 4 patients (25%) with a clinical diagnosis of nephrocalcinosis, and 1 of 7 patients (14.2%) with a clinical diagnosis of FSGS. No clinically causal variant was identified in patients referred with a clinical diagnosis of renal tubular acidosis. Causal variants were found in the TTC21B gene in 2 patients with CRF and dysmorphic findings, in the PHEX gene in 1 patient with hypophosphatemic rickets, and in the WT1 gene in 1 patient with Denys-Drash syndrome.

Pathogenic variants leading to different phenotypes incompatible with the preliminary clinical diagnosis were detected in the CLCN5 and COL4A4 genes in 2 patients with polycystic kidney disease and in the SLC12A3 and OCRL genes in 2 patients with a preliminary diagnosis of Bartter syndrome, COL4A4 gene in 1 patient with a preliminary diagnosis of nephrotic syndrome, HNF1B and CLCNKB genes in 2 patients with a preliminary diagnosis of Gitelman syndrome, and SLC12A1 and COL4A5 genes in 2 patients with a preliminary diagnosis of FSGS. These patient rates were reported in TABLE 2 as “variant incompatible with the preliminary diagnosis” in different disease categories.

A VUS compatible with the preliminary diagnosis was detected in 6 patients with polycystic kidney disease. A homozygous VUS in the CRB2 gene compatible with the clinical diagnosis and a homozygous VUS in the SLC5A2 gene incompatible with the clinical diagnosis were detected in 2 patients with a preliminary diagnosis of nephrotic syndrome. One patient with a preliminary diagnosis of Bartter syndrome was found to have a homozygous VUS in the SLC12A1 gene. In addition, a VUS in the COL4A3 gene was shown in a patient with a preliminary diagnosis of FSGS, which was different from the preliminary clinical diagnosis.

Discussion

As a result of the nephropathy gene panel study, which included 44 genes in 145 patients, most of whom were diagnosed with polycystic kidney disease but also with renal tubular acidosis, nephropathy, CRF, Bartter syndrome, Gitelman syndrome, Alport syndrome, FSGS, nephrocalcinosis, and nephrogenic DI, the overall genetic diagnosis rate was 48%.

In the polycystic kidney disease group, representing the largest number of patients in the cohort, family history was positive in 38% of patients, and a diagnostic yield of 58.7% was obtained. As in the literature, variants were mostly found in the PKD1 or PKD2 genes.^9,10 The diagnostic yield was 72.7% in Gitelman syndrome, 66.6% in Bartter syndrome, and 45.4% in Alport syndrome. Such a high diagnostic yield could not be achieved in patient groups referred to other nephropathy clinics. This is because many genes associated with hereditary kidney disease have been identified in recent years, and the gene panel used at the time of the study was inadequate to detect other hereditary nephropathies. In a recent study, a diagnostic yield of 59% was achieved by applying a 127-gene nephropathy panel to a small cohort of 56 families.¹¹ In another study, a genetic diagnostic yield of 43% was obtained after examining a panel of 207 genes in 135 families.¹² Comparing the overall diagnostic yield in our cohort to these studies, the ratios are close.

In addition, it is hypothesized that many of the VUS representing 11% of the cohort that were not included in our genetic diagnostic statistics and that were also mentioned in TABLE 3 may have a phenotypic effect. However, data such as functional studies, family segregation analyses, and identification of other unrelated affected individuals with the same genetic variants may change the classification of these variants in the future. Moreover, a potential disadvantage of NGS is its limited ability to detect copy number variations and we may have missed these copy number variations in this cohort. Therefore, the true rate of genetic diagnoses is likely higher than we currently report. The variants in 18 patients diagnosed with genetic disease in our cohort have not been previously reported in the literature or HGMD Professional. These variants in AQP2, CLCN5, HNF1B, OCRL, PAX2, PKD1, PKHD1, SLC12A1, SLC12A3, TTC21B genes were pathogenic and likely pathogenic variants according to ACMG classification, most of them in the PKD1 gene. These novel variants included 4 nonsense, 5 frameshift, 3 splice region and 6 missense variants. All novel variants were considered disease-causing by at least 1 prediction tool. Functional studies on these variants data may help to ascertain the role of novel variants in disease development. These variants will contribute to the spectrum of variants in the literature. Genes with novel variants have been previously associated with several known hereditary kidney diseases in OMIM. TABLE 4 provides information on 70 patients diagnosed with genetic disease and on variants. The novel variants are listed together in TABLE 5.

Pathogenic and likely pathogenic variants compatible with the preliminary clinical diagnosis were detected in 61 of 70 patients diagnosed as a result of the nephropathy gene panel study. However, variants different from the preliminary clinical diagnosis were identified in 9 patients. After retrospective reevaluation of these 9 patients, the preliminary clinical diagnoses were changed in agreement with the detected pathogenic variants. This condition, referred to as phenocopy in some studies in the literature, was frequently observed in groups of patients with hereditary kidney disease. In particular, the phenotypic overlap between Bartter and Gitelman syndromes is well known.^13,14 Our cohort’s diagnosis changed from Bartter to Gitelman and vice versa in one patient each. In addition, causal variants of Bartter syndrome were found in 2 of the referred patients with clinical diagnoses of FSGS and nephrocalcinosis. Another study reported that cohorts of patients with clinically diagnosed FSGS had varying rates of pathogenic variants in the Alport syndrome genes COL4A3, COL4A4, and COL4A5 but not in FSGS-related genes.^15,16 In our cohort, a pathogenic variant in the COL4A5 gene was detected in a patient with a clinical diagnosis of FSGS.

In one study, adults awaiting renal transplantation for ESRD of unclear cause were found to have a genetic cause in 12% of cases by renal gene panel study. Since the primary etiology may affect graft survival by recurrence or rejection, it is very important in terms of renal transplantation to know the underlying renal disease for the treatment of ESRD; therefore, genetic diagnosis is very important.¹⁷ In the study cohort, genetic diagnosis was made in 4 patients with ESRD and the etiology was determined. In this way, this study helped to improve these patients’ management before and after transplantation and estimate the risk of kidney disease in relatives.

Our study presents the results of diagnostic tests for hereditary kidney disease that can help clinicians in test selection for appropriate patients and in counseling patients about the possibility of obtaining a positive result with genetic testing. Hereditary kidney disease gene panels enable differentiation of complex phenotypes and classification of patients and facilitate patient management by improving our understanding of the relevant phenotype. The gene panel used is a noninvasive, cost-effective tool for genetic diagnosis of hereditary kidney diseases. In the future, the addition of other genes related to hereditary kidney diseases will enable a higher diagnosis rate and thus more personalized treatment.

Abbreviations:

Abbreviations:
 
• ESRD

  end-stage renal disease

 
• ACE

  angiotensin-converting enzyme

 
• NGS

  next-generation sequencing

 
• CRF

  chronic renal failure

 
• FSGS

  focal segmental glomerulosclerosis

 
• DI

  diabetes insipidus

 
• SNVs

  single nucleotide variants

 
• indels

  insertions/deletions

 
• HGMD

  Human Gene Mutation Database

 
• ACMG

  American College of Medical Genetics and Genomics

 
• gnomAD

  Genome Aggregation Database

 
• VUS

  variant of uncertain significance

Conflict of Interest Disclosure

The authors have nothing to disclose.

REFERENCES