The identification of CFHR5 nephropathy

Introduction

In 2006 I was working as a registrar in renal medicine at Hammersmith Hospital. One of my colleagues, Patrick Maxwell, was interested in investigating families in which multiple individuals had unexplained renal disease. He felt that hunting for the genes underlying their conditions was worthwhile, feasible and would make an excellent project for a clinical researcher: recent technical advances in genetics – principally the ability to genotype large numbers of markers across the genome relatively easily – have made linkage studies more widely accessible to researchers, so with Patrick Maxwell's help and supervision I embarked on the study of one such family.

Unusual glomerulonephritis

The family live in West London and have a highly unusual form of glomerulonephritis. The index case (patient A, Figure 1a) had been referred some years earlier when he had presented to his GP at the age of 17 years with headaches. He also reported recurrent episodes of macroscopic haematuria (visible blood in the urine) which occurred at a frequency of approximately once each year, on each occasion within 1 or 2 days of the onset of symptoms of upper respiratory tract infection – a pattern termed ‘synpharyngitic macroscopic haematuria’. Physical examination was normal with the exception of elevated blood pressure and a urine dipstick test revealed microscopic haematuria. The headaches resolved on treatment of the high blood pressure, and blood tests (including kidney function, serum complement C3 and C4 and the autoimmune screen) were all normal. His nephrologist's clinical impression was that this was likely to be IgA nephropathy and he organized a kidney biopsy which took place at St Mary's Hospital in London. IgA nephropathy is the most common glomerulonephritis worldwide¹ and is classically associated with microscopic and synpharyngitic macroscopic haematuria – often with progressive renal dysfunction. IgA nephropathy is diagnosed by renal biopsy which shows deposition of immunoglobulin A (but not other immunoglobulins) in the kidney.

Figure 1.

Family trees for patient A and patient B

Surprisingly, the biopsy did not show IgA nephropathy. Although the light microscopic appearances showed evidence of inflammation sometimes seen in IgA nephropathy, the stain for immunoglobulin A was negative and, in addition, there was no staining for other types of immunoglobulin. Instead there was isolated complement C3 deposited in the glomerulus. The complement cascade is commonly activated by immunoglobulins and diseases in which excess immunoglobulins are generated (such as chronic infections or systemic lupus erythematosus) are often associated with complement C3 deposited in the kidney alongside immunoglobulins. Diseases in which C3 is deposited without immunoglobulins are very rare and are often associated with consumption of circulating complement owing to a systemic defect of complement regulation.

Terry Cook, the histopathologist, regarded the biopsy as highly unusual and inferred that something other than immunoglobulins was causing complement to be deposited in the kidney – although what this might be was not obvious. This histological pattern is now termed C3 glomerulonephritis (C3GN) – reflecting that the primary abnormality is C3 deposition in the glomerulus – and has been associated with acquired or inherited abnormalities of complement alternative pathway regulation.²

The mother of patient A, who was originally from Cyprus, had also undergone a kidney biopsy which showed essentially similar appearances – and this strongly suggested a genetic cause for the disease. She reported a distant relative (now deceased) who had emigrated from Cyprus to the UK several decades previously and had undergone a renal transplant at Charing Hospital. She also told me that this patient's daughter had seen a kidney doctor. Professor Cook reviewed the original kidney biopsy specimens from both these individuals (performed in the late 1970s and early 1990s, respectively) and observed that they both demonstrated the features of C3GN. He also had what turned out to be a crucial insight when he recalled having seen this pattern in a fifth patient (patient B) who also had a Greek-sounding name, raising the question of whether this individual might be a distant relative of the family.

Family history

After having gained ethical approval for undertaking research into families with genetic kidney disease, I first interviewed patient A's mother. She reported that her family was from a village called Gerakies which is in the Troodos mountains of Cyprus. She was not related to patient B as far as she knew but invited me to Cyprus to meet the rest of her family who still lived there in order to screen them for evidence of the disease.

I also contacted patient B who was living in Nicosia in Cyprus and arranged to meet him. Like patient A, he reported frequent previous episodes of macroscopic haematuria which coincided with infections – particularly of the respiratory tract. He also reported that his serum creatinine tended to rise with each episode. In addition he knew of two male relatives on his mother's side who had died from kidney failure (Figure 1b) and that this part of his family was from the village of Kalopanagiotis, which is situated less than 5 km from Gerakies in the Troodos Mountains (Figure 2). In addition, several of his female relatives exhibited microscopic haematuria, although none of them (including his mother, then in her 70s) had either renal impairment or hypertension. No other people in his family had undergone a kidney biopsy.

Figure 2.

Satellite image showing the location of Gerakies and Kalopanagiotis (arrow) and locations of reported ancestry of other affected patients and families (circles) (in colour online)

It seemed likely that the kidney disease was caused by a mutation in a complement-regulating gene and had been inherited in both families from a common affected ancestor who had resided in the Troodos Mountains.

Candidate gene analysis and linkage study

Matthew Pickering, a rheumatologist at Hammersmith Hospital and a world expert in complement biology, played a vital role in investigating the family – initially by excluding defects of known complement-regulating genes. In collaboration with Veronique Fremeaux-Bacchi (in Paris) and Santiago Rodrigues de Cordoba (in Madrid) the genes Complement Factor H (CFH), Complement Factor B, Complement Factor I and Complement C3 (which are all known to be important in regulating complement) had been sequenced, but no mutations which might have caused the disease were identified.

I next undertook a genome-wide linkage study. This was done by genotyping several thousand single nucleotide polymorphisms (SNPs) across the genome and using the computer program GENEHUNTER to identify which parts of the genome were transmitted with the renal disease within the families. This experiment demonstrated that a locus on chromosome 1 spanning the gene Complement Factor H and its 5 homologues, Complement Factor H Related genes 1-5 (CFHR1-5), was linked to the disease in both families. Furthermore, the haplotype (sequence of consecutive SNP alleles along the chromosome) across these genes was identical in both families – a situation which could only arise if this region had been inherited in both families from a single common ancestor (a ‘founder effect’). Importantly, nowhere else in the genome was shared by all affected individuals in both families.

Further studies, including standard sequencing of CFHR genes 1 and 5 within the locus, drew a blank so the next step was to test for copy number variation across the locus. Copy number variation is where instead of the normal two copies of a gene or part of the gene an individual has more or fewer copies owing to a deletion or duplication event in the past. A multiplex ligation-dependent probe amplification assay demonstrated an internal duplication within CFHR5 resulting in extra copies of exons 2 and 3 (but not exon 1) in affected individuals – a change we would later show produces a larger version of the protein which is present in the blood of people with the disease. This change did not show up in standard sequencing (where each exon is sequenced individually) because the sequences of the extra copies of the exons were normal. The duplication was confirmed and its full extent (which turned out to be 6300 base-pairs) determined by Southern blotting and this allowed selective amplification and sequencing of the insertion point of duplication (which lay between exons 3 and 4 of the gene, Figure 3). Knowledge of this sequence allowed design of a one-step assay to detect the duplication in any DNA sample. This assay was used to confirm that the duplication is not present among 100 UK controls. When 1015 Cypriot controls were tested the duplication was found in a single individual – proving that the duplication is not simply a normal variant which is common in Cypriots.³

Figure 3.

Duplication of a 6300 base-pair (bp) region (double-headed arrow) results in extra copies of exons 2 and 3 of the gene CFHR5 (shown in red) (in colour online)

An endemic disease

The existence of a rare mutation flanked by the same haplotype is evidence of a founder effect, where a mutation is inherited from a common ancestor by members of a community. The size of the flanking shared haplotype was used to estimate that it was most likely that 10 generations had elapsed since the most recent common affected ancestor of both families. This estimate is very approximate because it is based on the observed distribution of recombinations which actually occur stochastically. Nonetheless, the fact that the families were not aware of being related to each other was independent evidence that the mutation had occurred at least several generations ago. Together with the observation that the mutation had turned up in a random sample of 1015 Cypriots, this raised the obvious question of whether additional affected patients or families might exist within the Cypriot community.

I therefore contacted Alkis Pierides, an eminent Cypriot nephrologist, to ask firstly whether he had previously encountered other patients with the clinical features which characterised the disease in the families, and secondly whether he had any patients with unexplained kidney disease for whom these findings might be relevant. He responded with great enthusiasm and together with his colleague Constantinos Deltas, who is an expert in genetic kidney disease in Cyprus, they tested DNA from patients and families with unexplained kidney disease. I also tested additional patients from North London (where there is a large Cypriot community) and within 6 months between us we had identified over 100 patients from 17 ostensibly unrelated families, all with this disease. All the patients had Cypriot ancestry, but not all could trace this to the Troodos Mountains region (Figure 3). Together these findings demonstrated that ‘CFHR5 nephropathy’ (as we termed the disease) is a common cause of kidney disease in Cypriots – accounting for approximately 5% of previously unexplained renal failure in this community.^3,4

Clinical observations

The identification of such a large cohort of patients who share the same known molecular defect has allowed very rapid increase in understanding of the clinical course of the disease. This has demonstrated that the risk of developing kidney failure is much higher in men compared with women (80% compared with 20% of those over 50 years old). Clinical observation over time has suggested a strong link between systemic infection and acute renal dysfunction and that conventional immunosuppression had provided little obvious benefit. Several individuals with the disease have undergone renal transplantation with excellent overall results, although the disease can recur following transplantation.^4,5 This last observation proves that the kidney disease results from a circulating abnormality, rather than a defect of a locally produced factor and suggests that systemic therapies which modulate complement activation might alter the natural history of the disease.

Further work

Further work aims to understand how the molecular change in CFHR5 nephropathy leads to renal disease. Initial experiments performed by Elena Goicoechea de Jorge in Matthew Pickering's laboratory have shown that the mutant protein has altered function when compared with the wild type protein, but how this leads to kidney disease is still not understood.³ We also hope to understand what the best treatment for patients with the disease is, and understanding why most women (and a few men) are protected from kidney failure may provide an important clue.

The observation that a mutation in CFHR5 can lead to a kidney disease raises the question of what role the CHFR5 protein usually plays in protecting the kidney and whether administration of supplementary CFHR5 could be a new therapy to protect the kidney in other diseases (such as IgA nephropathy and lupus nephritis) in which complement is deposited in the kidneys.

DECLARATIONS

Competing interests

None declared

Funding

This paper was funded by the Medical Research Council (MRC)

Ethical approval

This study received approval from the Hammersmith and Queen Charlotte's and Chelsea Research Ethics Committee (Ref: 06/Q0406/151)

Guarantor

DPG

Contributorship

DPG is the sole contributor

Acknowledgements

I would like to thank the patients and their families for their trust, encouragement and hospitality; Patrick Maxwell for his expertise and support; all my other collaborators for their invaluable help in this investigation and the MRC for funding the study. This paper is the winner of the RSM Grand Rounds Competition 2010