Genetic testing in steroid-resistant nephrotic syndrome: why, who, when and how?

Congenital NS (CNS), which presents within the first 3 months of life, is commonly associated with causative mutations. Indeed, mutations have been identified in 75–100% of cases of CNS [8, 10, 65, 66]. Causative mutations appear to largely occur in one of five genes (NPHS1, NPHS2, WT1, LAMB2 and PLCE1). NPHS1, encoding nephrin, is the main gene implicated in CNS, and mutation is responsible for the autosomal recessive Finnish type (CNF), which typically has a severe phenotype with massive proteinuria and rapid progression to ESRD [11]. However, the NPHS1 mutation detection rate remains high amongst non-Finnish cases of CNS [8, 10, 65]. Mutations in the NPHS2 gene, encoding podocin, are also responsible for a significant number of CNS cases, and the phenotype varies from the severe CNF presentation to milder disease with onset of proteinuria occurring later than in those with NPHS1 mutations [4, 66, 67]. Mutations in the PLCE1, WT1 and LAMB2 genes have also been detected in patients presenting with isolated CNS; mutations in these genes will be discussed in more detail in the next section.

Monogenic NS, which presents in infancy (from 4 to 12 months of life) and during childhood, is most commonly attributed to mutations in the NPHS2 gene, encoding podocin [8, 68]. There is a recognised genotype to phenotype correlation that explains this phenotypic variability [12, 66, 69]. Recent whole-exome sequencing performed on a paediatric cohort of SRNS patients revealed the mean age of onset associated with NPHS2 mutations to be approximately 6 years [8]. That said, it is clearly important to consider NPHS2 mutation as a cause for SRNS in a child of any age and, conversely, to expect a low rate of NPHS2 mutations in certain ethnic groups, namely Chinese, Japanese and Korean [10, 70]. Additionally, NPHS1 mutations have been identified in infants and children presenting with SRNS, with the rarer hypomorphic mutations being associated with a milder late-onset phenotype [8, 71, 72]. Thus, mutations in this gene should be considered in all paediatric age groups.

Mutations in PLCE1 (encoding phospholipase C epsilon-1) typically cause isolated DMS, with patients presenting with severe, early-onset SRNS and rapid progression to ESRD [13]. Again, significant clinical heterogeneity exists, with PLCE1 mutations manifesting from birth and throughout childhood, with both FSGS or DMS found histologically [73, 74].

WT1 encodes Wilms’ tumour 1, a transcription factor and key kidney development gene. Mutations in WT1 cause isolated and syndromic SRNS, which will be discussed later in this review. Previous studies have estimated WT1 mutations to account for approximately 6% of sporadic SRNS cases in childhood, manifesting at any age, depending on the underlying mutation [8, 66, 75, 76]. Genotype–phenotype correlations have been drawn from specific WT1 mutations. Of these, certain mutations cause early-onset, severe disease with DMS histologically, or alternatively, late-onset SRNS with FSGS histologically and slower progression to ESRD [75]. Interestingly, isolated SRNS may result from a wide range of WT1 sequence variations, which are associated with variable expression and incomplete penetrance [75]. It is important to stress that WT1-related nephropathy may be an isolated disease with no associated comorbidities or it may be an initial symptom of syndromic SRNS with extra-renal features manifesting later.

Although mutations in TRPC6 and ACTN4 are typically associated with autosomal dominant late-onset disease, as discussed below, there are reports of both infantile and childhood-onset SRNS caused by mutations in these genes.

Autosomal dominant SRNS typically presents later in life, in adolescence or adulthood, and has significant phenotypic variability. The overall mutation detection rate remains substantial, approaching 25% in adolescence and 12% in adulthood, but it is lower than cases presenting earlier in childhood [8, 65]. The main genes implicated in late-onset SRNS, which presents in adolescence, include NPHS2, TRPC6, INF2 and ACTN4 [31, 65, 68, 77, 78]. TRPC6, encodes a transient receptor potential cation channel and was originally identified as causing autosomal dominant FSGS, presenting in adolescence and early adulthood and showing relatively rapid progression to ESRD. Since then, TRPC6 mutations have been implicated in childhood-onset FSGS, and even SRNS presenting within the first year of life, with variable disease severity [8, 10, 77, 79, 80]. Similarly, mutations in ACTN4, which encodes alpha-actinin 4, typically cause late-onset FSGS with slow progression to ESRD [29, 81], but mutations in this gene have been reported in children presenting with SRNS and rapid progression to ESRD [8, 82]. Mutations in INF2, encoding inverted formin 2, were originally identified in patients with autosomal dominant SRNS, with age of onset ranging from adolescence and throughout adulthood [31, 83]. Although INF2 mutations typically result in isolated FSGS, they have also been detected in a subgroup of patients with associated Charcot–Marie–Tooth neuropathy [84]. Specific NPHS2 mutations may manifest late, in adolescence or adulthood; the common variant R229Q may result in late-onset SRNS in compound heterozygotes with specific second mutations [85].

Syndromic SRNS is associated with extra-renal manifestations and most commonly occurs due to mutations in genes encoding nuclear proteins (WT1, LMX1B, SMARCL1, WDR73), glomerular basement membrane and adhesion components (LAMB2, ITGA3, ITGB4), actin cytoskeleton components (MYH9) and lysosomal (SCARB2) and mitochondrial proteins (COQ2, COQ6, PDSS2, MTTL1, ADCK4). Table 2 lists the major extra-renal manifestations associated with gene defects causing syndromic SRNS; if extra-renal manifestations are present, it is highly likely that a causative mutation will be identified. A full description of syndromic SRNS is beyond the scope of this review but readers are directed to several detailed reviews for a thorough discussion [5, 20, 21, 44, 45, 50, 55, 84, 86, 87].

WT1 mutations are associated with a spectrum of significant extra-renal manifestations, including urogenital abnormalities and malignancy. Mutations in the KTS (splice insertion) site are associated with Frasier syndrome, characterised by childhood-onset SRNS, histologically characterised FSGS, male-to-female sex reversal and increased risk of gonadoblastoma [18, 75]. Missense mutations in exons 8 and 9 (affecting the zinc finger domains) are associated with Denys–Drash syndrome, characterised by infantile-onset SRNS, histologically characterised DMS, sex reversal, gonadoblastoma and Wilms’ tumour [75]. Large genomic rearrangements disrupting WT1 and the neighbouring PAX6 result in WAGR syndrome (Wilms’ tumour, aniridia, genito-urinary abnormalities and mental retardation) [75]. Although these genotype–phenotype correlations have been clearly described, it is important to note that there remains significant phenotypic variability with respect to extra-renal manifestations of WT1-associated disease. Furthermore, histopathological heterogeneity is noted even amongst carriers of the same genetic abnormality [75].

LAMB2 and LMX1B mutations typically cause Pierson syndrome and Nail–Patella syndrome, respectively, but have also been identified in isolated congenital, infantile and childhood-onset SRNS [88, 89], and mutation in these genes should therefore be considered to be causative in these age groups.

Isolated SRNS may also be seen, although rarely, in certain mitochondrial cytopathies, including MELAS (mitochondrial myopathy, encephalopathy, lactic acidosis and stroke-like episodes) caused by MTTL1 mutations [54] and coenzyme Q₁₀ deficiency [90]. Coenzyme Q₁₀ deficiency due to mutations in COQ2 and COQ6 may cause an isolated or syndromic nephropathy [51]. Mutations in PDSS2 cause Leigh syndrome but may also cause isolated SRNS [52].

ADCK4 mutations have been found to cause isolated SRNS with histologically characterised FSGS, manifesting from infancy through to early adulthood [53]. In contrast to previous studies [9], in a large multicentre cohort of Chinese paediatric patients with SRNS, ADCK4 was found to be the most commonly mutated causative gene, responsible for SRNS presenting as early as the congenital period, but most frequently during childhood [10].

Interestingly, there have been reports of patients with coenzyme Q₁₀ deficiency and ADCK4 mutation whereby interventional treatment with COQ₁₀ supplementation has been shown to modify disease progression [53, 90].

Several studies highlight that monogenic SRNS is largely steroid resistant, irrespective of the causative mutation [6, 69, 91]. Subsequent management of such patients often includes various immunosuppressive agents, all with associated adverse side effects and often limited clinical benefit [6, 91]. Indeed, in a recent paediatric cohort of sporadic SRNS, none of those with an identified causative mutation responded to immunosuppressive agents, compared to almost 60% of those without a mutation [6]. However, there are reports in the literature of a partial response to steroids, ciclosporin therapy or calcineurin inhibitors in patients with specific WT1, NPHS2, PLCE1 and TRPC6 mutations or mutations in the regulators of Rho-like GTPase (ARHGDIA, KANK 1, KANK 2 and KANK 3) [13, 92‐94]. Nonetheless, given that responsiveness to immunosuppression is unlikely in monogenic SRNS, upon receiving a positive genetic test result, clinicians should consider weaning immunosuppression if no clinical benefit has been demonstrated, thus protecting patients from the potentially devastating side effects of such treatments. It is important to note that there have been reports of patients with monogenic SRNS partially responding to secondary immunosuppression [13, 92‐94]; in such cases, it would be important to continue therapy whilst clinical improvement continues.

The discovery of rare monogenic causes of SRNS have revealed a small but significant cohort whose disease may be amenable to specific interventional treatment, thereby avoiding lengthy immunosuppression and delaying progression to ESRD. Patients with disease-causing mutations in genes encoding enzymes of the coenzyme Q₁₀ pathway (COQ2, COQ6 and ADCK4) and in the CUBN gene may respond to treatment with coenzyme Q₁₀ and vitamin B12, respectively. Likewise, patients with ARHGDIA mutations, through modulation of Rac I–mineralocorticoid interactions, could theoretically respond to eplenerone (a mineralocorticoid-receptor antagonist) [36].

As highlighted in Table 2, many syndromic forms of SRNS have associated medical problems that may benefit from early recognition and management. An example is the WT1 mutation, which can predispose to malignancy, and the detection of such mutations should trigger monitoring for associated Wilms’ tumour and gonadoblastoma. Given the latter entity is largely associated with sex reversal, a karyotype analysis should also be performed, especially in phenotypically female patients presenting with SRNS and primary amenorrhoea.

Renal histology has historically been utilised as a key diagnostic and prognostic criterion for children with SRNS, but emerging evidence reveals significant histological heterogeneity amongst monogenic causes of SRNS, demonstrating that biopsy findings may not correlate with genetic results [8, 95]. Furthermore, there does not appear to be a notable difference in the frequency of histological lesions found in patients with or without a recognised genetic cause [6]. For these reasons, in cases of primary SRNS, rapid genetic testing has the potential to obviate the need for renal biopsy for diagnostic purposes and serves as a less invasive diagnostic method; this is of particular significance in the younger SRNS cohort, for whom a genetic aetiology is more likely. In the event of rapid genetic testing being inaccessible, renal histology may direct clinicians towards the most likely “culprit” gene; for example, if DMS is detected in an infant presenting with SRNS, it would be prudent to perform mutational analysis on certain genes (LAMB2, WT1, NPHS1, PLCE1) preferentially over others. However, clinical indications for renal biopsy do remain, such as atypical features suggestive of lupus nephritis, with histology providing useful information. Additionally, a histological diagnosis enables phenotype patterns to become better established and is therefore useful from a clinical research perspective.

There is a high risk of progression to ESRD in monogenic SRNS, with many patients requiring renal transplantation [96]. Several studies have highlighted a low risk of disease reoccurrence post-transplant when a genetic aetiology has been confirmed [6, 8, 68, 69, 96, 97]. Conversely, there is a high risk of post-transplant disease reoccurrence in the idiopathic group; this is postulated to be caused by circulating factors [98, 99]. Given that the likelihood of post-transplant reoccurrence is minimal for genetic SRNS; after excluding the mutation in parents, a parental transplant can be planned, although in our experience this typically follows bilateral nephrectomy and an interval of time on dialysis. Post-transplantation prognosis is improved with living donor transplantation, which is associated with prolonged graft survival and decreased rejection rates as compared to deceased donor kidney transplantation.

Genetic counselling prior to testing should ensure that families are informed regarding potential outcomes and limitations of the chosen genetic test, including the discovery of variants of unknown significance and the potential for incidental findings. Possible benefits, including changes to medical management, should be discussed, as well as potential harms, including privacy, legal and social implication. These subjects are covered in the next section.

Making a molecular diagnosis has important implications for a family. It may enable accurate discussion of recurrence risk in future children and potential identification of pre-symptomatic individuals at risk [4]. Early referral to a clinical genetics service can facilitate identification of individuals at risk and genetic testing of family members as well as counselling in terms of family planning, prenatal diagnosis and pre-implantation genetic diagnosis. Genetic screening in unaffected family members may be additionally important when planning a living related donor (LRD) renal transplant, especially in the case of autosomal dominant disease. When inheritance of a severe disease-causing mutation is likely, pre-symptomatic testing for proteinuria and genotyping at birth are avenues that should be discussed and offered to affected families. Ethical considerations of causing potentially unnecessary anxiety should be addressed, and families should be counselled on the benefits and risks of such an approach; that is, the benefits of providing a timely genetic diagnosis, and therefore active clinical management, versus the risks of testing for a disease which may never manifest or be mild in the event of incomplete penetrance or variable expression. Prenatal diagnosis could be offered in families with a known risk of severe NS, such as CNS, or in cases where elevated alpha-fetoprotein levels have been detected in maternal serum or amniotic fluid. It can be used to allow the family to make an informed decision about continuing a pregnancy or to allow preparation both by the family and medical professionals for the birth of an affected child.

In addition to the ethical and emotional considerations that must be addressed during genetic counselling, there are several other risks to the patient and their family who are considering genetic testing or receiving a genetic diagnosis. In the UK, genetic testing is usually paid for by the National Health Service, and in countries with a privatised medical system, health insurance policies often cover the cost of genetic testing performed at the request of a doctor. However, this is not always the case, and in some situations the significant cost, which can be over US$2000, must be covered by the family. Furthermore, upon receiving a genetic diagnosis, the fear of insurance discrimination and the associated costs of enhanced insurance premiums represent a significant emotional and financial burden. Although there is legislation in place which protects those with a genetic disease from discrimination by health insurers, this does not always extend to protect patients from employment discrimination or the amplified costs of life, disability and long-term care insurance [100]. As part of the genetic counselling process, these issues should be discussed with affected families and informed consent obtained prior to genetic testing. A thorough discussion regarding the barriers to genetic testing in public health is beyond the scope of this review, but readers are directed to several useful articles for further information [100‐104].

When accessible and affordable, mutational screening should be performed in all children presenting with primary SRNS. Even in young adults, the likelihood of detecting a causative mutation remains substantial, and when cost allows, mutational screening should also be offered to this cohort. However, when such an inclusive approach is not possible, there are certain indications in which a genetic cause for SRNS becomes more likely, and mutational screening should be performed as a priority. Given that the likelihood of detecting a causative mutation is inversely related to age of disease onset [105], mutational screening becomes increasingly important the earlier the disease manifests. Genotype–phenotype correlations clearly demonstrate that mutations in recessive genes are more frequently implicated in early-onset disease and that mutations in dominant genes are more frequently implicated in adult-onset disease. Although there may not be an obvious family history in early-onset disease, a positive family history in any age group indicates that monogenic SRNS is likely and should trigger mutational screening. Additionally, the likelihood of finding causative recessive mutations correlates directly with the degree of consanguinity [80]; thus, a history of consanguinity should prompt mutational screening. Finally, the presence of extra-renal manifestations suggestive of an underlying genetic syndrome (Table 2) makes screening of associated genes advisable.

The clinical indications for genetic testing in SRNS can be summarized as follows: