CFH-CFHR1 hybrid genes in two cases of atypical hemolytic uremic syndrome

Atypical hemolytic uremic syndrome (aHUS) is a rare complement-mediated disease that manifests as the triad of thrombotic microangiopathy. We identified two aHUS patients with neither anti-complement factor H (CFH) antibodies nor causative variants of seven aHUS-related genes (CFH, CFI, CFB, C3, MCP, THBD, and DGKE); however, their plasma showed increased levels of hemolysis by hemolytic assay, which strongly suggests CFH-related abnormalities. Using a copy number variation (CNV) analysis of the CFH/CFHR gene cluster, we identified CFH-CFHR1 hybrid genes in these patients. We verified the absence of aHUS-related abnormal CNVs of the CFH gene in control genomes of 2036 individuals in the general population, which suggests that pathogenicity is related to these hybrid genes. Our study emphasizes that, for patients suspected of having aHUS, it is important to perform an integrated analysis based on a clinical examination, functional analysis, and detailed genetic investigation.

members of her family (I-1, III-1, and III-3) was deduced from a hemolytic assay using sheep red blood cells [14] (a functional assay to detect complement dysregulation mainly associated to CFH). Patient-2 ( Fig. 2A) was a 55-year-old male who presented with TMA and was successfully treated by plasma exchange. Similar to patient-1, the hemolytic assay for patient-2 also showed increased hemolysis indicating CFH-related abnormalities (Fig. 2B). However, a screening of aHUS-related genes (CFH, CFI, CFB, C3, MCP, THBD, and DGKE) did not identify any pathogenic variants, and autoantibodies against CFH were absent.
Two unusual copy number variations (CNVs) were observed in the CFH/CFHR gene cluster using a multiplex ligation-dependent probe analysis. In patient-1 and her family members, we observed a heterozygous deletion extending from CFH exon 23 to its downstream sequence and a heterozygous duplication extending from CFHR1 exon 5 to 6 ( Fig. 1B). In patient-2, an unusual heterozygous deletion was observed extending from CFH exon 23 to CFHR1 exon 5 (Fig. 2C).
Regarding the family of patient-1, long-range PCR confirmed the existence of a CFH-CFHR1 hybrid gene (Fig. 1C) and revealed a breakpoint within a 136-bp region, which was homologous between CFH intron 21 and CFHR1 intron 4 (Fig. 1E). cDNA sequencing for patient-1 confirmed the presence of mRNA transcripted from the CFH-CFHR1 hybrid gene (Fig. 1D). The hybrid gene was in-frame and composed of 23 coding exons, the first 21 of which were derived from CFH exons 1-21 and terminal 2 from CFHR1 exons 5-6 ( Fig. 1F). The protein product is expected to be an in-frame 20-SCR protein where SCRs 1 to 18 are derived from CFH and SCR 19 to 20 from CFHR1 (Fig. 1G). Regarding patient-2, a CFH-CFHR1 hybrid gene was also predicted to be inframe based on the genomic sequence, and its breakpoint was within a 41-bp region, which was homologous between CFH intron 22 and CFHR1 intron 5 (Figs. 2D, E). The length of the deletion was 84.6 kb, extending from CFH intron 22 to CFHR1 intron 5.
Since all the aHUS-related hybrid genes showed an abnormal copy number for the CFH gene C-terminus, we evaluated the copy number of the CFH/CFHR gene cluster in control genomes of 2036 individuals in the general population from the Tohoku Medical Megabank Project [15], using whole-genome sequencing data (described in copy number analysis section in Supplementary Notes). Although some types of CNVs were detected for CFHR1 to CFHR5, the copy number of CFH was normal in all individuals (Supplementary Fig. 1; Supplementary Table 1). Whole genome analysis was also performed on the two aHUS cases reported here, and the copy number changes shown by MLPA were detected by whole genome analysis as well ( Supplementary Fig. 2).
To date, six types of hybrid genes were reported in the CFH/ CFHR gene cluster (Fig. 3) [8][9][10][11][12][13]. In four of these, C-terminaldeficient CFH proteins were fused to the C-terminus of the CFHR protein ("CFH-CFHR protein") [8][9][10][11], similar to our cases. Because the impaired function of the CFH C-terminus, i.e., the SCR 19-20 surface binding domains [16], results in reduced recognition of and binding to host cell surfaces, the CFH-CFHR hybrid proteins devoid of this domain can cause over-activation of the complement system on cell surfaces. On the other hand, the other two hybrid genes consisted of a C-terminal-deficient CFHR protein and C-terminus of the CFH protein ("CFHR-CFH protein") [12,13]. Since the CFHR-CFH proteins can bind to cell surfaces via the C-terminus of the CFH protein, they will compete with and reduce the binding of normal CFH proteins.
Notably, the CFH-CFHR1 hybrid gene found in the family of patient-1 exhibited a reduced penetrance, which is consistent with past reports [17]. Penetrance rates of aHUS patients with the CFH mutation was reported to be 59% [18], and those with the CFH-CFHR hybrid gene showed rates of 33% [10]. The family members of patient-2 denied consent and could not be analyzed in this study. It is possible that his is a sporadic case caused by a de novo genomic rearrangement [9,11].
In considering the pathogenicity of ultra-rare diseases, including aHUS, the frequency of each variation is expected to have a crucial role. Since no abnormal CNVs of the CFH gene occurred in 2 036 healthy racially-matched individuals, the variants described in this study demonstrate moderate evidence of pathogenicity (PM2) according to the American College of Medical Genetics [19]. In addition, the same type of hybrid protein has already been reported to be pathogenic in other aHUS cases [8,9], providing strong evidence of pathogenicity (PS3). The hybrid genes presented here are classified as "likely pathogenic" due to combination of PS3 and PM2.
This study also had a limitation. We only demonstrated the breakpoints that make up the hybrid genes and did not show the breakpoints within introns. Sequences between CFH-CFHR3 and  F Schematic presentation of the mRNA for the CFH-CFHR1 hybrid gene and chromatogram demonstrating its breakpoint. Blue and red lines represent CFH-specific and CFHR1-specific single nucleotide polymorphisms, respectively. The gray area represents the homologous sequence between CFH and CFHR1. G Putative structure of the CFH-CFHR1 hybrid protein. Blue and red circles denote SCRs that originated from CFH and CFHR1, respectively sequences between CFHR1-CFHR4 are highly homologous over a 30-kbp stretch [7]. Considering this, and the fact that the copy number of CFHR3 was shown to be normal by MLPA, we speculate that the 3' side of CFH was replaced by the 3' side of CFHR1; however, as the sequence was restored within the highly homologous intron portion, the copy number of CFHR3 was normal, although we were not able to show this because of the difficulty in designing appropriate primers. In summary, we detected novel breakpoints that cause CFH-CFHR1 hybrid genes in two aHUS cases through CNV and functional analyses, and also confirmed the absence of CNVs of the CFH gene in the general population. Our study emphasizes that the combination of functional and genetic analyses can more accurately elucidate the genetic origin and, ultimately, the biomolecular mechanisms of aHUS.