Effect of heterozygous pathogenic COL4A3 or COL4A4 variants on patients with X‐linked Alport syndrome

Abstract Background Alport syndrome is an inherited renal disease caused by mutations in COL4A3, COL4A4, or COL4A5 genes. Coexisting mutations in either two of the three genes in Alport patients have been reported recently. However, the effect of heterozygous mutations in COL4A3 or COL4A4 genes in X‐linked Alport syndrome (XLAS) patients is unclear. Methods Using targeted next‐generation sequencing, six unrelated Chinese children were identified to have a combination of a pathogenic variant in COL4A5 and a heterozygous mutation in COL4A3 or COL4A4. They were three males and three females. Another three XLAS males each with only one pathogenic variant in COL4A5 were included. The clinical data were analyzed and compared between the males in two groups (group 1, males with a pathogenic variant in COL4A5 and a heterozygous pathogenic variant in COL4A3 or COL4A4; group 2, males with only one pathogenic variant in COL4A5). Results Patients with XLAS who also had heterozygous pathogenic COL4A3 or COL4A4 variants accounted for 1% of Alport syndrome. In this study, three children showed coexisting pathogenic variants in COL4A5 and COL4A3. Two children showed pathogenic variants in COL4A5 and COL4A4. One child had pathogenic variants in the three COL4A3‐5 genes, in which the pathogenic variant in COL4A5 was de novo and the pathogenic variants in COL4A4 and COL4A3 were inherited independently (in trans). The site and type of mutations in COL4A5 were similar between the two groups. It was revealed that males in group 1 presented more severe proteinuria than males in group 2 (p < 0.05). Conclusion The present study provides further evidence for complicated genotype in Alport syndrome. For the first time, we reported a case with three pathogenic variants in COL4A5, COL4A3, and COL4A4 genes. Moreover, we found that heterozygous pathogenic COL4A3 or COL4A4 variants are likely to make XLAS disease more serious.

However, the phenotype of individuals with heterozygous mutations in COL4A3 or COL4A4 genes varies widely. In our previous study, parents of ARAS children who were carriers of heterozygous mutations in COL4A3 or COL4A4 genes, 53% of them had normal urinalysis, 31% had hematuria, and 16% had hematuria and proteinuria . The diseases associated with heterozygous mutations in COL4A3 or COL4A4 genes are thin basement membrane nephropathy and autosomal dominant Alport syndrome (Fallerini et al., 2014;Kamiyoshi et al., 2016;Savige et al., 2003).
To date, the effect of heterozygous mutations in COL4A3 or COL4A4 genes in XLAS patients is unclear. It is important to know whether it would make the XLAS disease worse. Here, we reported six unrelated Chinese children with XLAS who were also detected with heterozygous mutations in COL4A3 or COL4A4 genes. Our study aimed to provide more information on clinical assessment and genetic counseling for Alport syndrome.

| Patients and families
Patients diagnosed or suspected of Alport syndrome in the department of Pediatrics, Peking University First Hospital from 2014 to 2017 were screened for mutations in the COL4A3, COL4A4, COL4A5, and COL4A6 genes. The clinical diagnosis criteria included ① glomerular hematuria, proteinuria, or renal failure, ② family history of Alport syndrome or renal failure without other exact disease, ③ lack or discontinuous staining of α5 (IV) chain in epidermal basement membrane (EBM), or in GBM, ④ the GBM lesions under electron microscopy (irregular thinning, thickening with splitting, and lamellation), ⑤ one pathogenic mutation in COL4A5 or two pathogenic mutations in COL4A3 or COL4A4 genes. Alport syndrome was suspected for individuals with criteria ① and ②. Alport syndrome was diagnosed for individuals with criteria ① ② and one of ③ ④ ⑤.
Six unrelated XLAS children were found to have pathogenic variants in COL4A5 gene and also have heterozygous pathogenic variants in COL4A3 or COL4A4 genes. They were three males and three females. Then another group of XLAS patients was enrolled in this study according to the following criteria: ① with only one pathogenic variant in COL4A5 gene, ② with frameshift variant in exon 37, large deletion in exon 42, or glycine substitution in exon 25 of COL4A5 gene, ③ male. Finally, there were three males of XLAS enrolled. Therefore, there were two groups of male patients (group 1, males with a pathogenic variant in COL4A5 and a heterozygous pathogenic variant in COL4A3 or COL4A4; group 2, males with only one pathogenic variant in COL4A5). For all subjects, clinical data including gender, age of disease onset, initial symptoms, extrarenal manifestations, results of skin biopsy and renal biopsy, family history, therapy, kidney function at latest follow-up were collected.

| Ethical compliance
The Ethical committee of Peking University First Hospital approved the project, and informed consent was obtained from the probands and their family members.

| Genomic DNA and targeted NGS
A sample of peripheral blood in EDTA tubes was collected from probands and all available family members. Gnomic DNA was isolated from the blood samples using a FlexiGene DNA Kit (Qiagen) according to the manufacturer's protocol. Gnomic DNA from probands was detected for mutations in COL4A3-6 genes. Targeted next-generation sequencing (NGS) was performed by BGI-Tianjin, China as published previously (Wang et al., 2017). The pathogenicity of variants identified was based on meeting at least one of the following criteria: (a) truncating mutations (nonsense, consensus splice site ±1 or 2 nucleotide, large deletion, and frameshift), (b) variants previously described as disease causing in a patient with a similar phenotype in the website (HGMD, LOVD, and Clin Var), (c) Glycine missense variants in the intermediate collagenous domains (except p.Gly624Asp in COL4A5), or (d) novel non-Glycine substitutions absent or at very low frequency in large population cohorts (1,000 genomes; ExAC, gnomAD), in domain high evolutionary conservation, and more than two prediction scores classified the allele as disease causing (SIFT, Mutation Taster, Polyphen 2) (Richards et al., 2015;Savige et al., 2018). The pathogenic mutations identified with NGS in probands were confirmed by Sanger sequencing or qPCR (real-time quantitative PCR) for large deletion mutations. Gnomic DNA from family members was analyzed by Sanger sequencing or qPCR to determine whether they had the same mutations as the probands.

| Statistical analysis
Statistical analysis was performed by SPSS 19.0. The differences of proteinuria level, age, and time of therapy in two groups were compared using unpaired t test. If the p value < 0.05, the differences were considered to be statistically significant.

| RESULTS
In this study, among 417 patients diagnosed or suspected of Alport syndrome in our department during 2014-2017, six were identified with pathogenic variants in COL4A5 plus heterozygous pathogenic variants in COL4A3 or COL4A4, which accounted for 1%. They were three males (proband 1, 2, 3) and three females (proband 4, 5, 6). Another three males (proband 7, 8, 9) with only one pathogenic variant in COL4A5 were included. The sites and types of pathogenic variants in COL4A5 gene were similar between the two groups of males. Two frameshift mutations in exon 37 p.G1110Afs*45 and p.G1098Vfs*54 in COL4A5 gene were detected in proband 1 and proband 7, respectively. The same large deletion mutation (deletion of exon 42) in COL4A5 gene was detected in both proband 2 and proband 8. Two missense mutations p.G594R and p.Gly644Val in exon 25 in COL4A5 gene were detected in proband 3 and proband 9, respectively. The pathogenic variants identified in these children were shown in Table 1.
The clinical manifestations were compared between the two groups of XLAS males (Table 2). All the boys in the two groups had initial symptoms of hematuria and proteinuria, and positive family history of kidney disease. The onset age of the disease, and hearing loss rate and age had no difference between the two groups. Kidney biopsy information was available in four boys. Skin biopsy and α5 (IV) chain staining in the EBM were underwent in three boys and two of their mothers. Negative staining of collagen IV α5 chain was identified in either EBM or GBM in five boys. All the six boys had angiotensin-converting enzyme inhibitor (ACEI) and angiotensin receptor blocker (ARB) for treatment after diagnosis. The amount of ACEI and ARB for each patient was 0.25 mg −1 kg −1 d and 0.5 mg −1 kg −1 d, respectively. The length of time for therapy and the age at latest follow-up had no difference between the two groups. All the boys had normal renal function at latest follow-up. However, the proteinuria during follow-up was more severe in three boys (proband 1, 2, 3) with a combination of one pathogenic variant in COL4A5 and one heterozygous pathogenic variant in COL4A3 or COL4A4. And the difference between the two groups was statistically significant (p < 0.05).
In the three females (proband 4, 5, 6), the initial symptom was hematuria and the onset age was from 5 months to 3.6 years ( Table 3). During follow-up, proteinuria was presented in proband 6 at 5 years and was 0.207 g/d at 8 years under ACE inhibitor therapy. Proband 4 was a 3.6-year-old girl with three pathogenic variants in COL4A5, COL4A3, and COL4A4. She had no family history. Hearing was not detected. Renal biopsy at 3.6 years showed thinning and thickening of GBM. She was presented with hematuria and proteinuria at last follow-up (3.8 years).
Family segregation analysis was performed in all available family members. In proband 1 and proband 2, only their mothers' samples were available and it revealed the pathogenic variants in COL4A5 were inherited from the mother, but the pathogenic variants in COL4A3 were not from the mother. In proband 3 and proband 6, it revealed that both of the pathogenic variants in COL4A5 and COL4A4 genes were inherited from the mother. And the pathogenic variants in COL4A5 gene were de novo in proband 4 and proband 5. In addition, the two heterozygous pathogenic variants in COL4A3 and COL4A4 genes in proband 4 were inherited in trans (on opposite chromosomes) from the mother and the father, respectively. The pedigrees of the six families were shown in Figure 1.

| DISCUSSION
The present study provides further evidence for complicated genotype in Alport syndrome. About 1% of patients diagnosed or suspected for Alport syndrome had mutations in more than one of the COL4A3-5 genes. For the first time, we reported a case with three pathogenic variants in COL4A5, COL4A3, and COL4A4 genes. Moreover, we found initially XLAS males with an additional heterozygous pathogenic COL4A3 or COL4A4 variant presented heavier proteinuria than the XLAS males with only one pathogenic variant in COL4A5.
So far, only two males with pathogenic variants in COL4A5 and COL4A4 genes have been reported (Fallerini et al., 2017;Mencarelli et al., 2015). One was 26 years old with variant p.Gly1348Arg in COL4A5 and heterozygous  Their age was 9, 45, and 54 years old. The 9-year-old female was presented with hematuria and proteinuria. The 45year-old female was presented with hematuria, proteinuria, and hearing loss. The 54-year-old female had ESRD at the age of 44. Therefore, these reported data from adult patients were consistent with our results that an additional heterozygous pathogenic COL4A3 or COL4A4 variant would make XLAS disease worse. There were patients reported to have a combination of pathogenic variants in COL4A3 and COL4A4 (Fallerini et al., 2017;Kashtan et al., 2018;Mencarelli et al., 2015). In some cases the variants were on the same homologous chromosome (in cis). But in other cases the variants were inherited independently (in trans), like proband 4 in our study. Learning from the reported data, individuals with two heterozygous pathogenic variants in COL4A3 and COL4A4 in trans had more severe phenotype than those with a single heterozygous pathogenic variants, but had less severe phenotype than ARAS. However, there was no reference about XLAS patients combined with two heterozygous pathogenic variants in COL4A3 and COL4A4 in trans. The proband 4, a 3.6year-old girl, reported here is the first case with this kind of complicated genotype. It is unknown whether her phenotype would be more severe than females with XLAS or ARAS. The long-term follow-up and more cases in the future might let us know more.
Besides, we should be alert that the three XLAS males with heterozygous pathogenic COL4A3 or COL4A4 variant in this study showed negative staining of α5 (IV) chains in EBM or GBM. A clinical diagnose of XLAS is definite Savige et al., 2013;Wang et al., 2012;Wei et al., 2006). However, the genotype may be various, which would affect the risk of inheritance in families (Artuso et al., 2012;Kashtan et al., 2018). Therefore it is necessary to test all the three genes COL4A3-5 by NGS in patients suspected for Alport syndrome, not only meaningful for diagnosis, but also for the genetic counseling (Gross et al., 2017;Kashtan et al., 2018;Savige, Colville, et al., 2016).
In conclusion, we reported six XLAS children with heterozygous pathogenic COL4A3 or COL4A4 variants, which accounted for 1% of Alport syndrome. One of them was the first case with three pathogenic variants in COL4A5, COL4A3, and COL4A4 genes. Our data revealed an additional heterozygous pathogenic COL4A3 or COL4A4 variant would make XLAS males suffering from more severe proteinuria. It suggested that another genetic Note. EBM: epidermal basement membrane; GBM: glomerular basement membrane; ACEI: angiotensin-converting enzyme inhibitor; Ccr: creatinine clearance rate; Scr: serum creatinine concentration; ND: no data. a urinary protein creatinine ratio 0.32 g/g.
hit from COL4A3 or COL4A4 might make the XLAS disease worse.