Genetic Testing for Kidney Disease of Unknown Etiology

Abstract

In many cases of CKD, the cause of disease remains unknown despite a thorough nephrological workup. Genetic testing has revolutionized many areas of medicine, and promises to empower diagnosis and targeted management of such cases of kidney disease of unknown etiology. Recent studies using genetic testing have demonstrated that Mendelian etiologies account for approximately 20% of cases of kidney disease of unknown etiology. While genetic testing has significant benefits, including tailoring of therapy, informing targeted workup, detecting extrarenal disease, counseling patients and families, and redirecting care, it also has important limitations and risks that must be considered.

Introduction

Chronic kidney disease (CKD) is one of the most prevalent diseases worldwide, affecting approximately one in ten adults.^1,2 However, as distinct CKD subtypes can have overlapping or ambiguous clinical presentations, or as many patients may present with late disease, achieving a specific diagnosis with can be challenging. Following the onset of symptomatic disease or detection by screening, individuals typically undergo traditional diagnostic workup, including history and physical examination, and biochemical testing, renal imaging, or renal biopsy.³ Yet, in many cases an etiology remains elusive. These cases we will refer to as kidney disease of unknown etiology (KDUE). Individuals with KDUE account for approximately 1 in 10 adults with end-stage renal disease (ESRD).^4–6 Genetic testing has revolutionized the diagnosis of many diseases, including CKD. Known Mendelian forms of CKD account for approximately 70% of pediatric and 10% of adult cases.^7–12 However, as these disorders often have high genetic and phenotypic heterogeneity, they can be difficult to identify using traditional diagnosis. Recent studies indicate that genetic testing can diagnose the cause in approximately 10–40% of cases of KDUE.^13–18

In the following review, we discuss the role of genetic testing for KDUE. We review the epidemiology of KDUE, focusing on Mendelian etiologies. We then discuss the value of a genetic diagnosis, assess the utility of different testing modalities, and offer a potential diagnostic approach. Using case-based vignettes, we illustrate how genetic diagnostics can improve management, yet may also result in clinical dilemmas. We conclude by examining key challenges, and discuss opportunities for future research.

Epidemiology of KDUE

CKD affects more than one in ten adults in the United States,^4,19 and it is one of the leading causes of morbidity and mortality worldwide.^20,21 As CKD can result from many causes, which overlap in presentation, pinpointing an etiology can be difficult with traditional diagnostics. Thus, KDUE accounts for one in ten adult ESRD cases.^4–6 Mendelian nephropathies account for approximately 10% of adult CKD and 20% of pediatric CKD.^7,22,23 Review of recent massively parallel sequencing-based (MPS; also referred to as next generation sequencing) studies, using exome sequencing (ES) or targeted gene panel analysis, shows that genetic etiologies contribute significantly to KDUE.^13–18 Altogether, these studies assessed 443 individuals with KDUE and diagnosed Mendelian nephropathies in 97 cases (22%). While monogenic glomerulopathies were most common, driven largely by variants in type IV collagen genes, diagnostic findings spanned clinical disease subtypes (Figure 1; the complete list of diagnoses are listed in Supplemental Table 1). Furthermore, 29 of the 47 (62%) genetic diagnoses were unique to a single individual across all studies, illustrating the high genetic and phenotypic heterogeneity of KDUE.

Figure 1. Diagnoses in Individuals with KDUE.

Recent genome-wide investigations identified 47 distinct genetic etiologies of CKD within the categories listed within the inner disc. The genes identified within each category are represented in the outer ring, with the number of patients affected in parentheses. Of the 47 genetic diagnoses made, 29 (62%) represented singleton diagnoses, unique to one patient. Of note, collagen IV genes accounted for the majority of glomerular disease, and nearly a third of all diagnoses.

However, these findings should be interpreted with key caveats. Importantly, the majority of investigations assessed individuals, in which the cause of CKD remained unknown following a standard nephrological workup; in many cases, these studies only included individuals with features consistent with hereditary nephropathy, such as early-onset or familial disease. Hence the diagnostic rate may be considerably lower in unselected CKD populations (e.g. as in a primary care practice, or in creatinine clearance-based epidemiological studies). In addition, as these studies analyzed single-nucleotide variants (SNVs), individuals with diagnostic copy number variants (CNVs) for genomic disorders would have gone undetected. Verbitsky, et al. found that CNVs contribute to a significant portion of all-cause pediatric CKD,^24,25 but further research is needed to determine the contribution of genomic disorders to KDUE. Nonetheless, the high detection rate observed by Groopman, et al, which represented the largest and most clinically diverse all-cause nephropathy cohort reported to date, supports that monogenic etiologies contribute meaningfully to KDUE.¹⁴ Interestingly, Groopman, et al. found that the diagnostic rate did not differ significantly by race or ethnicity, demonstrating that Mendelian forms of CKD can be detected across diverse populations. Higher diagnostic rates were found among individuals with a positive family history of CKD and in those with congenital CKD, suggesting these features may be especially enriched for Mendelian diseases. Similarly, Mann, et al. found that positive family history, extra-renal manifestations and consanguinity predicted higher diagnostic rates.¹⁶ And, other studies have reported that early age of onset is also predictive of a higher diagnostic rate.^26,27

Genetic Testing Modalities

Genetic testing includes karyotypes, chromosomal microarray (CMA), targeted sequencing, panel sequencing, ES and genome sequencing (GS). The characteristics of commonly used tests for CKD are outlined in Table 1. Karyotypes and CMA are genome-wide tests for structural variants, including translocations, CNVs or chromosomal aneuploidies.^{36, 37} The karyotype directly visualizes the genome at a low level of resolution, with high sensitivity for aneuploidy, translocations and large inversions, but without the resolution to detect CNVs smaller than 1–2 mB.³⁸ CMAs hybridize a genome being tested to an array of oligonucleotide or single nucleotide polymorphism (SNP) probes. CMAs can be designed to detect CNVs at the exon-level, or as small as 50–200 kB, depending on the probe density of an array.^28,37,38 However, CMA cannot detect balanced translocations or inversions.

Table 1.

Genetic Techniques for Testing CKD

Technique	Description	Benefits	Drawbacks
Sanger Sequencing	• PCR-based, single nucleotide sequencing of a targeted locus, less than 1 kB; or of multiple loci simultaneously • Used for confirmatory testing of NGS results • Used for testing when a specific disease associated when a known locus, or loci are suspected	• Capable of detecting SNVs and short indels • Capable of testing for a microdeletions smaller than limit of CMA testing by using primers flanking a CNV hotspot • Rapid testing and analysis possible • Near perfect accuracy for variants within tested loci • Avoids secondary and incidental findings	• Scope of testing limited to less than 1 kB per sequence • Low throughput limits ability to test multiple variants • Cannot detect most structural variation
CMA	• Genome-wide survey of copy number capable of CNVs greater than 200–400 kb • Used when a genomic disorder is clinically suspected, often in combination with karyotype	• High resolution for CNVs • Unbiased, genome-wide assay	• Cannot detect SNVs and indels28 • Decreased sensitivity within repetitive regions and pseudogenes28 • Cannot detect balanced structural changes
MPS targeted panel	• Massive parallel sequencing in which DNA from many loci is isolated and sequenced • Panels tailored to sequence portions of genes known to be associated with specific diagnoses	• High sensitivity to a variation within a broad region of coverage • Significant reduction in data storage and annotation requirements compared to ES and GS • Useful when secondary and incidental findings are not desired	• Panel excludes novel and rare variants outside of region of coverage • Limited potential for future reclassification of variants
ES	• Massive parallel sequencing of nearly all protein-coding portions of genes • Provides unbiased survey that can detect most known disease-causing SNVs	• High sensitivity screen for exonic SNVs • Data can be reanalyzed periodically as new sequence information becomes available	• Limited detection of indels and CNVs • May not target all CKD genes29 • Highly resource intensive requiring expensive equipment, time consuming data interpretation and expert analysis30 • Can produce undesired incidental and secondary findings31 • Limited detection within repetitive and CG-rich regions32,33 • Platform-specific artifacts can be introduced into sequence data • Not uniformly covered and reimbursed by health insurance34
GS	• Massive parallel sequencing of near entirety of a genome • Provides unbiased survey that can detect most known disease-causing SNVs	• Sensitive to SNVs, including indels and intronic variants • Capable of sequencing genes with high homology to other loci • Capable of detecting genomic disorders • Decreased artifact in sequence data compared to ES • Data can be reanalyzed periodically as new sequence information becomes available	• Unclear significance of non-coding variants35 • Limited detection within repetitive and CG-rich regions32 • Highly resource intensive, particularly regarding longterm data storage • Can produce undesired incidental and secondary findings • Not uniformly covered and reimbursed by health insurance34

Targeted Sanger sequencing is highly sensitive and specific for small variants (less than 1 kB) in known genes. It represents the highest resolution, and therefore the narrowest scope, of genetic testing, and is only performed when a variation at a specific locus is suspected.

MPS has emerged as a powerful tool for clinical diagnosis.³⁹ MPS includes targeted capture of selected genes associated with a phenotype (targeted panel), or genome-wide approaches (ES or GS). In GS, all DNA, including (protein-encoding) exons (non-coding) introns, is ligated and sequenced, whereas in ES only the exons, the approximately 1% of protein-coding DNA, is selectively hybridized and sequenced.³⁹

Each type of MPS has advantages and disadvantages. ES and GS are unbiased, genome-wide tests, while targeted panels test only genes deemed to be associated with the phenotype of interest (e.g. glomerular disease or congenital kidney disease). However, CKD caused by Mendelian disease exhibits broad phenotypic and genetic heterogeneity, and can therefore be difficult to diagnose by panels. To retrospectively compare the diagnostic yield of ES versus targeted panels, Groopman, et al. selected 281 patients with KDUE who had positive ES diagnoses, and compared this yield to what would have been obtained from the broadest possible phenotype-driven panels for glomerulopathy, cystic/congenital defects and tubulointerstitial disorders, each virtually reconstructed from the union of three commercially available panel tests.¹⁴ They found that while 17% of cases of KDUE were diagnosed by ES, only 3–9% of such cases would have been diagnosed by targeted panels. Thus, ES had a superior performance even in comparison to the theoretically largest panel tests. Similarly, ES and GS have both been shown to have higher diagnostic yields than targeted panels in children with suspected genetic diseases, and in children with epilepsy.^40–42 However, ES and GS, by surveying across the entire genome, are more likely to identify variants of unknown significance, and incidental or secondary findings.⁴³ The use of population sequence data and stringent variant curation is thus essential for accurate genetic diagnosis from ES and GS.^44–46 GS has greater sensitivity to several classes of variants, including deep intronic variants, deletions/duplications, transposable elements, and CNVs.⁴⁷ While GS may theoretically have greater diagnostic sensitivity, the proportion of ES-negative cases of KDUE that could be resolved by GS has not been systematically investigated. All forms of MPS have limited coverage of GC-rich, repetitive and homologous DNA regions.³⁹

Diagnostic Approach

Rationale

Identifying a specific disease etiology is essential for effective management of CKD.³ Detection of a syndromic illness with extrarenal manifestations.^22,48 Such extrarenal manifestations can be easily mistaken for secondary complications of nephropathy.^49,50 Therefore, genetic testing of KDUE, particularly in children, can help predict, prevent and treat extrarenal disease. Genetic testing can enable tailored therapy, such as avoidance of corticosteroids, and treatment with coenzyme Q10, in steroid-resistant nephrotic syndrome.⁵¹ Genetic testing is of particular value for ESRD. Ottlewski, et al. recently demonstrated that targeted MPS panel testing identified ESRD etiology in 25% of adults awaiting transplantation.¹⁵ Additionally, 65% of the cohort in which Groopman, et al. demonstrated the diagnostic utility of ES, had ESRD.¹⁴ This population often cannot undergo renal biopsy, and genetic diagnosis can predict graft survival. Specifically, ESRD caused by genetic podocytopathies is unlikely to recur, whereas ESRD caused by atypical hemolytic uremic syndrome or hyperoxaluria, can recur following transplant.^15,52,53 In addition, a genetic diagnosis can identify unaffected relatives, eligible for kidney donation. Finally, diagnosing the cause of CKD can end diagnostic odyssey, and thereby alleviate the stress felt patients and families grappling with illness of unknown cause.

Diagnostic Algorithm

We propose the following approach, based on a review of current literature and our practical experience (Figure 2). This approach assumes individuals have already undergone an initial nephrological workup, including biochemical and serologic testing, imaging of the kidneys, and renal biopsy if indicated. There are two exceptions, in which genetic testing may be performed prior to completion, or as part of the initial workup: 1) targeted genetic testing should be performed in cases highly suggestive of a specific genetic diagnosis; 2) genetic testing should be considered when kidney biopsy poses a greater risk, such as in the setting of bleeding disorders, pregnancy or solitary kidney. Excluding these exceptions, and following a negative or inconclusive initial workup, a patient is considered to have KDUE and may then be stratified according to the probability of genetic disease. We consider higher probability patients with the following risk factors: early onset disease (under 40 years of age); a positive family history of CKD; consanguinity; extrarenal anomalies; cystic disease; or congenital nephropathy. Without any one of these features, the choice to pursue genetic testing may be left to the discretion of the patient, their family, and physician, with consideration of personal preferences, the cost of testing, and insurance coverage and reimbursement.

Figure 2. Suggested algorithm for genetic testing of KDUE.

Patients in which a genetic disease is suspected can be tested as outlined. Secondary analyses may clarify negative results or VUS. Cases without positive results should be periodically reviewed and reanalyzed.

* CNV analysis is possible with exome data and is offered by some commercial testing laboratories, although CMA generally has higher sensitivity.</p/> ** ES may miss some SNVs, exons or entire genes due to variability in capture and coverage. ES may also miss CNVs, moderate sized deletions/duplications, retrotransposition, and deep intronic splice variants.

Prior to testing, appropriate genetic counseling must take place. Ideally, this discussion includes a genetic counsellor, however, it can also be conducted by a nephrologist, clinical geneticist or primary physician with experience and expertise in genetic testing. The pre-test counseling must include discussion of benefits and drawbacks of genetic testing by modality (Table 2). Patients should be counseled that genome-wide approaches, such as ES, have higher likelihood of indeterminate and secondary results, which are less likely in panel testing. Once the decision has been made to proceed with testing, we suggest the following approach. If CKD is present with extrarenal manifestations, particularly congenital disease, or in the case of pediatric onset of CKD, we suggest using CMA and ES, as first-line tests for detecting genetic and genomic disorders. Note that some commercial testing laboratories perform CNV analysis with exome data, although the diagnostic yield for CNVs may be inferior to CMA. For isolated renal disease in an adult patient, we recommend ES. We recommend genome-wide approaches given their increased sensitivity for diagnosing CKD, especially for individuals with nondiagnostic presentations, compared to targeted panels.¹⁴ For cases in which secondary or incidental results are undesired, clinicians may consider targeted analysis restricted to genes known to cause CKD.⁵⁴ (Our group developed such a gene list,¹³ a periodically updated version of which can be found at http://www.columbiamedicine.org/divisions/gharavi/.) If targeted analysis is negative, the clinician may consider broadening to an unrestricted analysis. Results of genetic testing should be interpreted and analyzed by specialists with experience in hereditary nephropathy, using current guidelines for diagnostic sequence interpretation.^44,55

Table 2.

Pre-Test Genetic Counseling

Topics to Address
• Explanation of the basic principles of heredity, genes, genetic disease and testing
• Potential benefits to patient and family with a genetic diagnosis
• Likelihood of results, and explanation of positive, negative and indeterminate results
• Likelihood of secondary or incidental results
• Cost of testing
• Potential for findings regarding paternity
• Potential for discrimination by community, employer or insurer, and existing legal protections
• Explanation of broad benefits and drawbacks by testing modality (Table 1), with focus on the tradeoffs involved with genome-wide approaches (i.e. increased diagnostic sensitivity versus increased potential for nonspecific findings, and increased costs)
• Implications for insurability

Perinatal Testing

Fetuses and neonates with KDUE represent a unique clinical challenge. Congenital kidney disease is associated with very high morbidity and mortality,⁵⁶ and can present ambiguously, such that achieving a specific diagnosis can be difficult. Relevantly, the diagnostic yield of MPS for congenital forms of KDUE is similar to that for neurodevelopmental disease,⁵⁷ and for children with suspected monogenic disorders,⁵⁸ where it is a first-line diagnostic test. Furthermore, prenatal ES in the presence of fetal renal anomalies, can provide an actionable genetic diagnosis in a significant number of cases.^59,60 Given the potential for a genetic diagnosis to aide in clinical decision-making, including redirection of care, and to reduce medical costs,⁶¹ we advocate for expedited ES and CMA or GS as a first-line test for fetal and neonatal kidney disease.

Interpretation of Results

If genetic testing leads to a positive result, it must be considered whether the genetic diagnosis explains the presenting disease. If the diagnosis is consistent with the presentation, genetic counseling should be performed with the patient or their family, and medical therapy should be tailored accordingly. If the genetic results are discordant with the clinical presentation (e.g. genetic diagnosis of polycystic kidney disease in a patient with no renal cysts), a review of the molecular and clinical data with the clinical geneticist or a nephrogenetics expert is recommended. If a high index of suspicion for a genetic diagnosis remains, or in the case of negative results or variants of unknown significance (VUS), the following steps should be taken. The blind spots and limitations of ES (see Table 1) should be considered to assure that relevant nephropathy genes or exons have been adequately captured and covered. For example, some CKD genes, such as GREB1L, are not targeted by all ES capture kits.²⁹ ES will also miss deep intronic splice variants, moderate sized deletions/duplications, and gene-disrupting transposable elements. Additionally, secondary analyses should be performed to detect potential mosaicism (a situation in which postzygotic mutations result in two or more populations of cells with different genotypes in the same individual). Finally, ES has lower analytic performance than GS, and may therefore miss some coding variants detectable by GS. If such possibilities are suspected, GS may be considered. Finally, if access to blood or kidney tissue is available, transcriptome sequencing may be considered as an adjunct for detection of splice altering intronic variants.^62,63

In addition to considering technical limitations of ES, reported VUS in genes associated with CKD may be reviewed for concordance with the patient’s phenotype. For example, if ES reveals a VUS in CLCN5 consistent with Dent disease, the patient might be tested specifically for low molecular weight proteinuria and hypercalciuria.⁶⁴ Finally, in cases without a positive diagnosis, we suggest periodic review as new genetic diagnoses and variant classifications become available.^65–67

Choice of ES versus GS

Our proposed framework favors ES over GS for several reasons. ES generally requires less cost to process, uses less data storage, and is easier to interpret than GS. While GS provides significantly greater detection of CNVs and noncoding variants, it has only marginally improved diagnostic utility because most diagnostic variants are found in coding regions also captured by ES.^47,68 The contribution of noncoding variants to KDUE requires further investigation. Studies in other fields show that analysis of noncoding regions remains challenging and there is only a marginal diagnostic improvement with GS over ES in diverse types of disease.^29,35,69,73 However, GS typically provides more uniform sequence coverage than ES, including of highly homologous regions, making it superior for such CKD-associated genes as PKD1.³² These blind spots of ES may thus miss dominant alleles, or miss recessive second alleles in individuals with compound heterozygous genotypes. Finally, it should be considered that both GS and ES do not cover several critical CKD-causing regions, such as the variable number tandem repeat region in MUC1, which require targeted testing.³³

Limitations and Drawbacks of Genetic Testing

Indeterminate results, reported as VUS, can be challenging to patients and caregivers. This risk can be mitigated with genetic counseling.⁷⁴ The sensitivity and specificity of genome-wide testing may improve, as more data regarding variants and control populations accrue. Presently, it is difficult to interpret whether nondiagnostic findings (e.g., VUS or negative results) truly indicate that an individual does not have a genetic cause of disease. Moreover, putatively pathogenic variants for Mendelian CKD inexplicative of patient’s known renal disease have been found in approximately 1% of patients with CKD,¹⁴ and in approximately 1% of unselected self-declared healthy adults.⁷⁵ There exists a major challenge for diagnosticians approaching patients with KDUE to determine whether identified variants do in fact explain disease. Periodic review of results will help reclassify such variants. Initiatives such as ClinGen are now also developing kidney specific workgroups to facilitate such efforts.⁷⁶

Genetic testing in CKD also presents logistical challenges. The cost of MPS testing has dropped, but ranges from a few hundred to thousands of dollars (United States) per test.³⁰ In the United States, insurance coverage and reimbursement of MPS testing continues to evolve.³⁴ The Centers for Medicare and Medicaid recently announced coverage of MPS testing in cases of advanced cancer.⁷⁷ This represented the broadest use of MPS to be routinely reimbursed. Iglesias, et al. found that insurance providers covered the majority of ES at a single tertiary center, mostly for children with developmental delays, birth defects or seizure disorder, and mostly with private insurance.⁷⁸ There are several research initiatives, such as the Undiagnosed Disease Network,⁷⁹ which provide genome-wide testing free of cost, but these are not available to most individuals. Worldwide, there is considerable heterogeneity in the availability of genetic testing, with differing access to geneticists, genetic counsellors, and technology, and differing insurance coverage and reimbursement.^80–83 While there are focused genomic medicine initiatives underway across Europe, the Middle East, Asia, and elsewhere,⁸⁰ genetic testing for KDUE is not yet routinely available to most people worldwide. Large prospective studies of the clinical value of genetic testing for CKD are needed to inform coverage and reimbursement policies.

Positive results, while clinically useful, can also pose new challenges to patients, families and clinicians. For example, individuals with the classical 22q11.2 deletion have a 75% chance of manifesting congenital heart disease, 75% chance of immunodeficiency, and a 30% chance of manifesting congenital anomalies of the kidneys and urinary tract (CAKUT).^84,85 Long term sequalae include increased risk of Parkinson’s disease and neuropsychiatric disease.^86–88 Diagnosing an individual, or a fetus, with a 22q11.2 deletion, may initiate extensive screening, counseling, and difficult decision-making. Genetic results also confer risk to patients and their families with regards to discrimination, health insurance and life insurance. The United States Genomic Information Nondiscrimination Act (GINA) of 2008 provides important protections against discrimination or providing health insurance on the basis of a genetic diagnosis. However, the law has important exceptions: it does not apply to members of the United States armed services, and does not protect life insurance.⁸⁹ Patients should be appropriately counselled regarding these risks prior to genetic testing.

Finally, genome-wide testing such as ES or GS has the potential to yield medically actionable secondary findings, in genes unrelated to the primary indication for testing. The American College of Medical Genetics (ACMG) designated 59 genes associated with medically actionable disease, including cancer and cardiomyopathy, to be assessed in patients undergoing genome-wide testing.³¹ Patients undergoing ES and GS, must be properly counselled and given the choice of being informed if such secondary findings are discovered. Interestingly, even though they may not explain the etiology of kidney disease, some of the findings in the ACMG 59 genes may be relevant to management of nephropathy.¹⁴ There are significant ethical challenges around secondary findings in children. The benefits of informing children or their families of secondary findings, which usually do not cause disease until adulthood, are unclear. The American Academy of Pediatrics therefore recommends deferring until adulthood testing for adult-onset disease.⁹⁰ Furthermore, challenges can arise when the preferences of children and adolescents differ from their parents regarding disclosure of secondary findings.

Clinical Vignettes

The following fictionalized cases demonstrate our suggested approach to genetic testing for KDUE, and the utility and challenges of genetic testing in this population.

Case 1

PK, a 9-year-old boy presents with low blood pressure and dizziness on repeat examinations. His mother has a history of diabetes and gout. He is short (6th percentile), and has been struggling in school. But he is otherwise well appearing, and asymptomatic. Laboratory testing demonstrates stage III CKD. Renal ultrasound demonstrates bilateral cysts. Targeted sequencing of PKD1 and PKD2 finds no diagnostic variants. CMA and ES were performed. ES found no diagnostic variants; however, CMA demonstrates a 1.2 Mb deletion in chromosome 17q12. Testing of PK’s parents demonstrates that the deletion is shared by his mother, and on follow up examination, she is found to also have bilateral renal cysts, albeit with normal renal function. The 17q12 deletion syndrome, also known as renal cysts and diabetes (RCAD; OMIM #137920) syndrome is characterized by bilateral cystic renal disease, progression to ESRD, hypomagnesemia, hyperuricemia and gout, maturity onset diabetes, short stature, epilepsy and developmental delays, and less commonly disease of the liver, eyes, ears, gonads, or other organ systems, however the presentation is highly variable.^91,92 Recent studies have established that the 17q12 deletion is one of the most prevalent CNVs in children with CKD, including those with CAKUT.^24,50

The variable penetrance and expressivity of 17q12 deletion syndrome presents a significant challenge in genetic counselling. Ultimately, PK is referred for formal audiology and neurodevelopmental testing given his difficulty in school, and he is referred to a cardiologist and ophthalmologist for evaluation for congenital heart and ophthalmologic disease. PK’s primary physician monitors his serum HbA1 c and liver transaminases to monitor for signs of diabetes and liver disease.

PK’s story illustrates the utility of CMA in pediatric CKD, and the challenges created when a genetic diagnosis includes variably penetrant manifestations. A single genetic diagnosis manifested as CKD and developmental delays in PK, and as diabetes, gout and renal cysts in his mother. Furthermore, extrarenal manifestations can be mistaken for sequelae of renal disease. For example, developmental delay in children with CKD, which was previously attributed to the burden of chronic illness, have been linked to CNVs that may cause neurodevelopmental as well as renal disease.⁴⁹ Moreover, the genetic findings have meaningful implications for choice of therapy, as individuals with RCAD are at increased risk of post-transplant diabetes, informing choice of post-transplant immunosuppression.

Case 2

DN is a 42-year-old woman who had been followed for CKD by her primary physician and nephrologist for nearly two decades. She was found to have nephrotic-range proteinuria incidentally by urinalysis; since then, her GFR and blood pressure gradually worsened. Renal biopsy showed non-specific focal segmental glomerulosclerosis (FSGS). She had no family history of CKD. Given DN’s presentation of KDUE, her nephrologist referred DN for genetic testing. ES found her to be compound heterozygous for the p.R229Q and p.E310A missense variants in NPHS2. The genetic analyst noted that although previously reported in FSGS cases and predicted damaging by the in-silico score PolyPhen-2,⁹³ these variants were present at appreciable frequencies in population control databases, and thus classified them as VUS.⁴⁴ DN’s providers consulted with a geneticist with expertise in CKD who diagnosed these variants as the cause of CKD. Podocin is a slit diaphragm component encoded by NPHS2, in which mutations cause steroid-resistant nephrotic syndrome.^94,95 While not disease causal in the homozygous state or in cis with each other, when inherited in trans the p.R229Q and p.E310A variants have been shown to lead to polymerization and mislocalization of podocin, thereby causing disruption of the glomerular basement membrane and FSGS.⁹⁶ DN’s nephrologist also consulted with a genetic counsellor. Given the prevalence of p.R229Q alleles in certain populations,⁹⁶ DN was counseled regarding the risk of having children with CKD should a future partner be a carrier, as well as of the potential benefit to her family for NPHS2 testing.

This case highlights the challenge of diagnosing a disease with a complex genetic mechanism, and the utility of variant reclassification with expert guidance, reinforcing the value of disease- specific knowledge for accurate genetic diagnosis. This case also highlights how diseases traditionally thought of as pediatric-onset, are being increasingly found in adults.

Case 3

BG was a late preterm infant, with a history of oligohydramnios, and respiratory failure requiring intubation shortly after delivery. She had bilateral echogenic kidneys, and developed ESRD in the second week of life requiring peritoneal dialysis. CMA was negative; however, trio ES revealed she was heterozygous for a previously reported missense variant in WT1. The variant was absent in population control databases and was confirmed to have occurred de novo on parental testing. WT1 encodes a transcriptional factor required for normal genitourinary development. Heterozygous WT1 variants cause Denys-Drash syndrome which presents as congenital nephrotic syndrome and renal failure. Mutations in WT1 can also cause CAKUT and Wilms tumor.^97–100

The parents of BG were counselled that the variant in WT1 was the likely cause of her kidney disease, and that as a de novo mutation it was very unlikely, but not impossible, to occur in a subsequent pregnancy. BG’s course was further complicated by peritonitis, sepsis and persistent respiratory failure. Upon on learning the genetic diagnosis, associated with high mortality, her parents chose to redirect care towards goals of comfort, and she died peacefully. BG’s story displays the value of ES in diagnosing the cause of congenital KDUE, and how genetic diagnosis can facilitate end-of-life decision making in this setting.

Case 4

AS is a 35-year-old woman with hematuria and nephrotic range proteinuria. Her father and paternal grandmother also have histories of hematuria, and ESRD in their 5th and 6th decades of life, respectively. She has normal hearing, and no other appreciable extrarenal manifestations of disease. AS underwent a renal biopsy as part of her workup that demonstrated FSGS, not otherwise specified based on the Columbia Calssification.¹⁰¹ No abnormalities of the glomerular basement membrane were noted. Her nephrologist considered a course of corticosteroids as a first-line treatment, but given AS’s strong family history of CKD, she was first tested using ES and CMA. ES demonstrated a heterozygous missense variant, p.G148V, in COL4A3, that was absent in control populations and predicted to be deleterious. Furthermore, the variant was described in a case series of hereditary glomerulopathy.¹⁰² COL4A3 is one of collagen IV genes, which contribute to the glomerular basement membrane. Variation in these genes causes type IV collagen-associated nephropathy, which includes Alport syndrome and thin basement membrane disease.¹⁰³ The classical presentation of type IV collagen-associated nephropathy includes hematuria in childhood, sensorineural hearing loss, ophthalmologic disease, and progression to ESRD that is resistant to corticosteroids.¹⁰³ However, recent studies have found that individuals with type IV collagen-associated nephropathy present with highly variable phenotypes.^14,103 Indeed, collagen IV variants account for the majority of monogenic glomerulopathies identified in individuals with KDUE resolved by genetic testing (Figure 1).

Given this diagnosis, AS’s nephrologist did not prescribe corticosteroids. AS was counselled regarding the risk of progression to ESRD, the risk of CKD in future children, as well as the benefit of genetic testing for the rest of her family. AS was also presented with the option of enrolling into one the new clinical trials for Alport syndrome listed in clinical trials.gov.

This case illustrates several important points. Monogenic glomerular disease may often present with nonspecific findings including FSGS. Monogenic CKD can present with variable presentations that differ from classical descriptions. Genetic resolution of KDUE can lead to tailored medical therapy, and the avoidance corticosteroids. Finally, traditional diagnostic tools cannot diagnose the cause of such monogenic forms of CKD.

Conclusions

Genetic testing in nephrology has dramatically expanded over recent years, and promises to expand our understanding of the molecular pathogenesis of kidney disease. Lack of a clear diagnostic etiology in KDUE impedes accurate prognosis and targeted management. Importantly, genetic studies to date show that monogenic causes account for a significant proportion of KDUE, highlighting the potential of genetic diagnostics to provide personalized care for these individuals, including ending the diagnostic odyssey, predicting disease progression, anticipating extrarenal disease, and tailoring therapy. However, genetic testing includes challenges, including accurate variant interpretation and cost-efficacy. Prospective research into the diagnostic and clinical utility of genetic testing, particularly in different patient populations (e.g. transplant donors/recipients and neonates), the optimal frequency of reanalysis, and its cost efficacy in diverse clinical contexts (Box Item) will help to address these challenges. Additionally, studies of genetic testing in CKD have primarily focused on analysis of existing, biobanked cohorts, which are prone to selection bias. Prospective studies of genetic testing in CKD and development of quantitative phenotype risk scores, such as those developed for other Mendelian diseases,¹⁰⁴ would be of great value to definitively establish clinical utility, and to clearly establish which risk factors predict genetic disease. Prospective analyses comparing the clinical benefit, cost effectiveness and resource utilization of ES, GS and targeted panels would be of particular usefulness. Such efforts are needed to provide best practice guidelines for genetic testing, and to achieve precision medicine for patients with KDUE.

Box Item: Financial Costs and Benefits of MPS Testing.

While the financial costs of ES and GS have dropped significantly in recent years, they range from hundreds to thousands of dollars (United States), and are not uniformly reimbursed by insurers. However, this must be balanced against the significant costs of care of individuals with CKD.¹⁰⁵ Individuals with KDUE in particular may undergo extensive traditional testing and treatments. The potential benefits with genetic diagnosis include accurate risk assessment, tailored medical therapy, prevention of future disease with genetic planning, redirection of care, to name a few. The use of genetic diagnosis has been explored in other populations,^61,106 yet remains poorly understood in the case of CKD. Given the increasing costs of care related to CKD, prospective financial analyses are urgently needed to evaluate genetic testing in KDUE.

Supplementary Material

2

Supplemental Table 1. Genetic Diagnoses in Studies of KDUE