Background

Hereditary hematuric diseases comprise a genetically and clinically heterogeneous group of conditions, a common feature of which is microscopic hematuria (MH) since early childhood. The most frequent pathological entity is thin basement membrane nephropathy (TBMN) [1, 2], with an estimated population prevalence of 0.3–1% [3, 4]. About 40–50% of familial TBMN are explained by heterozygous mutations in the COL4A3 and COL4A4 genes, which encode for the α3 and α4 chains of collagen IV, the most abundant component of GBM [5]. A much more severe and progressive glomerulopathy, which also presents with MH since childhood is Alport syndrome, caused by either mutations in the COL4A3/A4 genes or mutations in the X-linked COL4A5 gene [2, 6]. Mutations in CFHR5, a gene playing a role in the regulation of the alternative pathway of complement, in FN1 (fibronectin 1), or in MYH9 (heavy chain of myosin 9), comprise rarer genetic causes of hereditary hematurias [1, 6,7,8].

Although TBMN, caused by COL4A mutations, was considered for many years to be a benign condition accompanied by excellent prognosis on long follow-up, several publications, including ours as early as 2007, convincingly showed that this is not the case [9,10,11]. This is based on results showing that a variable subset of patients will develop proteinuria and focal segmental glomerulosclerosis (FSGS) and progress to chronic renal failure, even end-stage renal disease (ESRD) [12]. In a Cypriot cohort, up to 30% of TBMN patients with known mutations reached ESRD, at an average age of 56-years. To date, these results have been verified by multiple groups [13,14,15,16,17]. The exact mechanisms for this adverse outcome remain unknown but the role of genetic modifiers has been implicated [18,19,20,21,22]. A significant feature of heterozygous COL4A mutations is the broad phenotypic heterogeneity, while the clinical outcome is at times better described as later-onset Alport-related nephropathy (LOAN) [23]. Some authors also use the diagnosis of autosomal dominant Alport syndrome.

Based on available data there must be more genes that confer FMH when mutated [11, 24]. Whole-exome sequencing (WES) is one of the best modern genetic approaches for gene discovery. Here we report the analysis of 6 families by (WES) and the finding of two variants co-inherited in two genes, the COL4A5 and the LAMA5 genes. Invoking this digenic inheritance can better explain the spectrum of symptoms observed in some patients, than the variant at each one locus alone [25].

Methods

Six families segregating FMH in at least two generations (autosomal dominant inheritance) were thoroughly studied. Some patients also had proteinuria, renal impairment or ESRD. Renal biopsy showed FSGS and TBMN in two of six families (Families CY5372 and CY5381). At first, the index patient of each family underwent Next Generation Sequencing (NGS) that included the parallel analysis of a 5-gene panel, COL4A3, COL4A4, COL4A5, CFHR5, FN1 (Ion Torrent, PGM, Life Sciences) [11]. Subsequently, WES was performed in the index patients plus one unrelated healthy subject by Macrogen (Kyoto, Japan), using the Illumina HiSeq platform (San Diego, CA, USA). The analysis included all exons, splice junctions and 5’-UTR/3’-UTR. Mean DNA fragment size was 101 bp. In total, 51.189.318 bases of genomic DNA were selected. The filtering algorithm of the data (Fig. 1) gave special attention to glomerulus specific genes [26]. Selected variants were validated by Sanger DNA re-sequencing using the ABI BigDye Terminator v1.1 Cycle Sequencing Kit and the ABI PRISM 3130xl genetic analyzer.

Fig. 1
figure 1

Hierarchical filtering that we followed for mining the Whole Exome Sequencing data

Full size image

The study was approved by the Cyprus National Bioethics Committee and all participants gave a signed consent.

Results

WES analysis

NGS analysis for the 5-gene panel was negative. In family CY5500 (Table 1) a variant in the COL4A5 gene, p.Asp654Tyr, was not considered as pathogenic and was overlooked as it did not involve a conserved glycine residue and was thought inadequate to explain the phenotype. WES resulted to a mean of 9000 non-synonymous SNPs and indels in coding regions, per individual (Additional file 1 After applying the analytical filtering algorithm (Fig. 1), we excluded about ~ 8500 variants and ended with a list of 549 candidate variants. When checked for any relation with the glomerular function, for animal models, or functional studies available in the literature, seven variants were considered of high risk, confirmed by Sanger re-sequencing (Tables 2 and 3). Two variants were predicted by five of six software to be damaging/disease causing. Specifically, variant COL4A5-p.Asp654Tyr was not found in 54 samples of the Cypriot general population nor in 305 samples with FMH tested at our setting. Also this variant was never reported in the ExAC genomes database [27]. Variant LAMA5-p.Pro1243Leu was neither found in 81 healthy individuals of the Cypriot general population nor was it present in any of 159 TBMN patients’ DNA. Also, according to the ExAC genomes database, it was found in European (non-Finish) populations with a frequency of ~ 1/21,123.

Table 1 Clinical data of CY5500 family
Full size table
Table 2 Effect prediction of the studied variants and general population information
Full size table
Table 3 Direct DNA re-sequencing results (AD: autosomal dominant). Variants not following AD inheritance have not been studied further
Full size table

Family CY5500

Both variants on LAMA5 and COL4A5 genes are co-inherited by all four patients in family CY5500. The LAMA5 variant has been inherited by male UCY2067 that he is presently healthy. There is oral information that the grandfather (generation I) was suffering by chronic kidney disease. Below there is analytical clinical description for all mutation carriers of family CY5500 (Fig. 2).

Fig. 2
figure 2

Pedigree of family CY5500. For family members with a UCY code, there is available a DNA sample. Four patients of the family carry both variants (patients UCY2075 and UCY2069 are actually hemizygous for the COL4A5 variant). Family member UCY2067 carries only the LAMA5 variant in heterozygosity, having normal kidney function and being negative for microscopic hematuria, aged 27

Full size image

UCY2069

CA is a 57-yr-old male, diagnosed with, and treated for, hypertension at the age of 36-yrs, associated with MH, minimal proteinuria and impaired renal function (MDRD eGFR = 66.0 ml/min/1.73m2). At the age of 54-yrs he developed heavy proteinuria of nephrotic range. Despite treatment with steroids and cyclosporine, his kidney function deteriorated and he developed ESRD at the age of 57-yrs. At the age of 54-yrs he had a kidney biopsy showing FSGS. The GBM was folded and appeared thick in many areas, while the podocytes appeared vacuolated and with marked fusion. Renal ultrasound showed multiple renal cortical cysts bilaterally, well before he reached ESRD (Fig. 3).

Fig. 3
figure 3

Ultra-sound of the kidneys of patient CA (UCY2069 on Fig. 1), aged 57-yrs, presently on ESRD. Both Kidneys are smaller than normal, measuring Rt kidney 8.88 cm in long diameter and the Lt kidney 9.51 cm, with increased echogenicity and cortical thinning, cortex measuring about 0.88 cm. There is no polycystic kidney disease (pelvocalyceal dilatation). Multiple cortical cysts are noted in both kidneys, about 6–8 cysts in each kidney, the bigger to the Rt about 4.5 cm and to the Lt about 3.6 cm. The cysts however have been present at least since 4-yrs, according to the patient’s medical record, at the time his renal function was much better (MDRD 50 ml/min)

Full size image

UCY4041

SCA, the daughter of patient CA, is a 25-yr-old girl with MH, borderline proteinuria, normal blood pressure and normal renal function (MDRD eGFR = 130 ml/min/1.73m2). A recent kidney biopsy showed FSGS. Mean width of the GBM was 218.73 nm, thinner than normal. Two measurements showed thickening alternating with thinning. In the most part the podocyte processes were maintained normal but there were a few segments with effacement. Folding of the GBM was observed in a percentage of 5% and in two areas thickening was observed. There were no clear pathognomonic features for Alport syndrome. Electron microscopy showed some deposits which most probably were IgM precipitates, recognized by immunofluorescent microscopy, localized in the periphery and the mesangium. Renal ultrasound was negative for renal cysts.

UCY2075

Male patient SA, brother of CA, aged 60-yrs, was diagnosed with heavy proteinuria, in the nephrotic range, associated with MH and renal impairment (MDRD eGFR = 43.0 ml/min/1.73m2) at the age of 51-yrs. At age 57-yrs he developed ESRD and had a cadaveric kidney transplantation. At the age of 51-yrs he had a kidney biopsy that revealed secondary FSGS, which at the time was attributed to pre-existing arterial hypertension and renal function impairment due to proteinuria. Electron microscope analysis is not available. Renal ultrasound showed multiple renal cortical cysts.

UCY2074

CSA, the daughter of patient SA, is a 34-yr-old girl, with MH and proteinuria. Her kidney function was normal at the age of 29-yrs, taking corticosteroids therapy (MDRD eGFR = 118.0 ml/min/1.73m2). She had a kidney biopsy at the age of 24-yrs, due to MH and proteinuria, revealing FSGS. Renal ultrasound was not available.

UCY2067

EA, the son of CA, aged 27, has inherited only the variant LAMA5-p.Pro1243Leu, and he is presently healthy.

Discussion

Miner et al. [28] described Lama5 (laminin alpha 5 chain) in 1995, member of the vertebrate subfamily of laminin chains, widely expressed in adult tissues, with highest levels in lung, heart, and kidney. In mature glomeruli, collagen α1α2α1(IV) and laminin α1β1γ1 (LN-111) interacting networks are replaced by collagen α3α4α5(IV), laminin α5β1γ1 (LN-511), and laminin α5β2γ1 (LN-521), as glomerular capillary loops expand [29]. It is documented that the maintenance of glomerular filtration barrier integrity requires laminin alpha5 [30]. Despite the fact that Lama5 mouse knockouts are fatal [31], mice with a hypomorphic Lama5 mutation (Lama5neo) that reduces laminin α5 expression, exhibit proteinuria, hematuria and cystic kidneys [32]. Podocyte-specific inactivation of Lama5 in mice, resulted in varying degrees of proteinuria and rates of progression to nephrotic syndrome. The GBM of the proteinuric mice appeared thickened and “moth-eaten,” and podocyte foot processes were effaced [30]. These facts, added to the population data mentioned above, enhance the likelihood that DNA variant LAMA5-p.Pro1243Leu is a mutation, probably a hypomorphic one, since the patients we identified exhibit a great phenotypic similarity with these animal models, presenting with hematuria, proteinuria, renal impairment, thinning and thickening of the GBM, podocyte foot processes effacement and multiple renal cysts.

In the absence of proper functional experiments it is impossible to know how exactly this Lama5 variant affects GBM structure and integrity. We can speculate however that the interaction of the laminin network with the collagen IV α3α4α5 network might be disturbed, resulting in distortion of the mature GBM meshwork and the glomerular filtration barrier. It is also reasonable to suspect an additive effect regarding the damage of GBM in the presence of a defective alpha 5 (Col-IV) chain, where if a dosage threshold is exceeded then a pathogenic phenotype appears. Even though analytical functional experiments are beyond the scope of this paper, we propose that the phenotype is better described by taking into consideration the variants at both loci than each one on its own [25].

It is equally worth mentioning that this is the third report presenting data supporting a probable contribution of the LAMA5 gene to human disease. In a previous report the authors found novel LAMA5 mutations, predicted to be deleterious, in three out of five patients with FSGS or presumed FSGS, also co-inherited with mutations in other genes [33]. Similar results were reported by another study [13]. No renal cysts were described in those patients. An additional argument in support of our hypothesis is the fact that laminin alpha 5 is an important protein of the glomerulus and an interacting partner of collagen IV, while animal models and patients exhibit similar phenotypes with our patients. While the evidence presented is valid, more studies are needed and more patients must be studied before this link is unequivocally documented. Probably, the easiest way is screening for mutations or for modifier variants in the LAMA5 gene in multiple FMH and Alport syndrome families. Confirmation of these results will be of high clinical importance, expanding the spectrum of genes involved in glomerular pathologies. Finally, in recent years other examples have been reported of patients with COL4A5 mutations and Alport syndrome, and phenotypes exacerbated by co-inheritance of mutations in other genes. These include mutations in other collagen IV genes (COL4A3/A4) and in the MYO1E gene [34, 35]. It should not escape our attention that high throughput analysis, including NGS technologies are enabling us to detect and connect clearly pathogenic or hypomorphic mutations in different genes and thus elucidating previously complex phenotypes. Often times though, it is difficult to discriminate cases of true non-Mendelian digenic inheritance from monogenic Mendelian inheritance of phenotypes exacerbated by serendipitous co-inheriting of a second non-allelic mutation or from the effect of otherwise neutral genetic modifiers, or even from the serendipitous co-inheritance of two independent clinical entities.

Conclusions

In conclusion, we found evidence that digenic inheritance in patients of one family presenting with FMH, renal failure/ESRD, FSGS and cystic kidneys, can more clearly explain the phenotypic spectrum than one gene alone. The suspected genes are COL4A5 and LAMA5. Existing animal models and two previous reports for rare mutations in LAMA5 gene support this observation. These results need further confirmation, as LAMA5 could explain a percentage of patients with complex phenotypes.