A further genetic cause of thin basement membrane nephropathy Free

Thin basement membrane nephropathy affects at least 1% of the population and is characterized clinically by persistent isolated glomerular haematuria but normal blood pressure and renal function [1]. There have been few advances in our understanding of the genetics of thin basement membrane nephropathy since it was first interpreted as the carrier state for autosomal recessive Alport syndrome [2, 3]. However this is set to change with the widespread use of whole exome sequencing (WES) as has been demonstrated here with the article by Gale et al. [4], which highlights the usefulness and limitations of WES to identify novel causative genes.

Thin basement membrane nephropathy was the diagnosis in a young woman referred by her family doctor for investigation of incidental but persistent haematuria [1]. In such cases, a renal biopsy typically demonstrates nearly normal histology but a uniformly thinned glomerular basement membrane (GBM) on electron microscopy. The prognosis is usually excellent. Thin membrane nephropathy is found more often in women than in men, possibly because the normal male GBM is thicker and mutations do not result in sufficient thinning to allow the transient GBM breaks and red blood cell loss. Proteinuria appears to be uncommon, as the membrane breaks are small and transient and most leaked urinary protein is reabsorbed by the tubules.

The clinical picture in thin membrane nephropathy contrasts with the young man with recurrent episodes of macroscopic haematuria coincident with mucosal infections who also has proteinuria and hypertension. His renal biopsy usually demonstrates IgA glomerulonephritis and he has a 30% chance of developing end-stage renal failure.

A family history is common with thin basement membrane nephropathy, and at least two-thirds of affected individuals have another affected relative [1]. The serendipitous observation of a thinned GBM in the parents of a person with autosomal recessive Alport syndrome first suggested that individuals with thin membrane nephropathy might be heterozygous for mutations; that is, the compound heterozygous or homozygous form resulted in recessive Alport disease [2]. This was confirmed when 40% of a hospital cohort of families with thin basement membrane nephropathy demonstrated linkage to the COL4A3/COL4A4 locus affected in autosomal recessive Alport syndrome [5] (20% of families actually had X-linked Alport syndrome). Subsequently, the same heterozygous COL4A3 and COL4A4 mutations were detected in thin basement membrane nephropathy as had been demonstrated previously in Alport syndrome, which was further evidence for thin membrane nephropathy being the recessive Alport carrier state [6].

Why do some families with thin membrane nephropathy have haematuria not linked to the COL4A3/COL4A4 locus? Mutations may still affect the COL4A3 and COL4A4 genes when no linkage is demonstrated. Demonstrating haematuria may be problematic. Affected men often have none, and haematuria may be transient or incidental. Could the lack of linkage be due to another major gene locus for thin membrane nephropathy? Probably not. There is no other gene locus known for autosomal recessive Alport syndrome, and there is no description of likely recessive disease (female, early onset renal failure, lenticonus, consanguinity [7–9]) where mutations could not be identified in the COL4A3 or COL4A4 genes. Furthermore, the population prevalence of thin basement membrane nephropathy and autosomal recessive Alport syndrome are consistent with a single major locus. If 1% of the population has thin membrane nephropathy, then 1/100 × 1/100 × ¼ (1/40 000) will have autosomal recessive Alport syndrome [9], which they do.

However, mutations in other genes may still be responsible for thin basement membrane nephropathy. Genes for structural GBM proteins are most likely to be affected. The GBM is 70% collagen IV by dry weight. The major collagen IV networks in the GBM are α1α1α2 and α3α4α5 [10]. Collagen IV α1 and α2 are produced first in the embryo, but a switch occurs in infancy to the α3α4α5 heterotrimer in the highly specialized membranes of the glomerulus, cochlea and retina. Collagen α1 and α2 persist in the adult GBM, but mainly in small amounts in the subendothelium. The next most abundant GBM, protein is laminin, and again the laminin subtypes switch from laminin α1β1γ1 in the embryo to laminin α5β1γ1 and α5β2γ1 in the mature GBM [11].

In this issue, Gale et al. suggest that COL4A1 collagen IV α1 chains mutations can also cause thin membrane nephropathy. For many years, mutations in COL4A1 or COL4A2 (corresponding to the collagen IV α2 chains) were thought likely to be lethal since the corresponding proteins were so abundant in the embryo and in adult vascular membranes. However, early reports of COL4A1 mutations indicated associations with porencephaly [Online Mendelian Inheritance in Man (OMIM) 175780] [12], hereditary angiopathy with nephropathy, aneurysms and muscle cramps (HANAC; OMIM 61773 [13]), cerebral small vessel disease (OMIM 607595) [14], retinal vessel tortuosity (OMIM 180000) [15] and congenital cataracts [16]. COL4A1 mutations causing these phenotypes often affected an integrin-binding site in the collagen IV α1 collagenous domain. Gale’s article suggests that COL4A1 mutations affecting other locations of the collagen α1 chain result in a different phenotype, with GBM thinning, cysts and sometimes renal failure [17].

We too have studied individuals with thin membrane nephropathy for mutations in novel genes. We examined a cohort of 23 individuals where linkage to COL4A3/COL4A4 was excluded for mutations in COL4A1, COL4A2 or LAMC1 (for laminin γ1 chain), but were unable to identify any pathogenic variants (KW Zhang and S Tonna, PhD theses, University of Melbourne). The search was difficult because of the mutation screening methods used at the time (single-strand conformational polymorphism, Sanger sequencing) and the size of the genes. There was also the issue of confirming the pathogenicity of DNA variants. This is easier now with the availability of the multiple normal variant databases (ExAc, dbSNP, LOVD, etc.) and in silico predictive algorithms for pathogenicity (SIFT, Mutation Assessor, PANTHER, PolyPhen-2, etc.). Recent guidelines for assessing pathogenicity have been published [18]. Gale et al. [4] were fortunate to have a large family to study, but haematuria remains difficult to use to define affected status.

Evidence is emerging of mutations in other genes affecting the GBM collagen IV α3 and α4 chains if not actually causing thin membrane nephropathy. Thus, mutations in the LMX1β transcription factor result in the nail-patella syndrome (MIM 169000) with haematuria and proteinuria [19]. LMX1β is a transcription factor for collagen IV α3 and α4, and mutations reduce transcription [20]. The GBM is often thickened with excess, ectopic collagen III. Mutations in other collagen IV transcription factors such as AP2 may affect collagen IV α3 and α4 chain expression and GBM thickness as well [21]. Currently there is no evidence for laminin gene mutations causing thinning. However, bi-allelic mutations, for example in NPHS2 [22] or MYO1E [23] as well as in COL4A3 or COL4A4, appear to worsen disease.

The field of mutation detection and gene discovery has been revolutionized by WES. We used WES to re-examine some members of our cohort where thin membrane nephropathy was not linked to the COL4A3/COL4A4 locus and identified previously unidentified COL4A3, COL4A4 and COL4A5 pathogenic variants.

Gale et al.'s [4] article suggests that membranes in other tissues are also affected in thin basement membrane nephropathy. The collagen IV α3, α4 and α5 chains are found in basement membranes in the glomerulus and distal renal tubules, but also in the alveoli, thyroid, breast, liver and eye (cornea, lens capsule and retina) [24]. Only GBM thickness has been studied in these subjects, and it is not known if COL4A3 or COL4A4 mutations result in thinning of other membranes and whether any thinning is clinically significant. The finding of cysts is consistent with the phenotype reported previously for COL4A1 mutations [25].

Gale et al.’s [4] article also highlights the varying effects of different mutation types. We know already from osteogenesis imperfecta and X-linked Alport syndrome that some collagen mutations (rearrangements, deletions and nonsense changes) are associated with a more severe phenotype than most missense variants [26, 27]. The finding of a frameshift mutation that almost always results in a downstream nonsense mutation, in COL4A1, is consistent with a severe phenotype and is more likely to be pathogenic than many non-Gly missense variants. Many variants that result in a nonsense change produce nonsense-mediated decay, and the absence of the α1 chain will prevent normal collagen IV heterotrimer assembly.

In summary, this article suggests that mutations in genes for other GBM structural proteins also cause thin basement membrane nephropathy. WES is likely to identify further genes.

CONFLICT OF INTEREST STATEMENT

The results presented in this article have not been published previously in whole or part except where indicated. The author has had no involvement that might raise the question of bias in the work reported or in the conclusions, implications or opinions stated.

(See related article by Gale et al. A novel COL4A1 frameshift mutation in familial kidney disease: the importance of the C-terminal NC1 domain of type IV collagen. Nephrol Dial Transplant 2016; 31: 1908–1914)

REFERENCES