Detection of Splicing Abnormalities and Genotype-Phenotype

CLINICAL RESEARCH www.jasn.org

Detection of Splicing Abnormalities and Genotype- Phenotype Correlation in X-linked Alport Syndrome

Tomoko Horinouchi,1 Kandai Nozu,1 Tomohiko Yamamura,1 Shogo Minamikawa,1 Takashi Omori,2 Keita Nakanishi,1 Junya Fujimura,1 Akira Ashida,3 Mineaki Kitamura,4 Mitsuhiro Kawano,5 Wataru Shimabukuro,6 Chizuko Kitabayashi,7 Aya Imafuku,8 Keiichi Tamagaki ,9 Koichi Kamei,10 Kenjirou Okamoto,11 Shuichiro Fujinaga,12 Masafumi Oka,13 Toru Igarashi,14 Akinori Miyazono,15 Emi Sawanobori ,16 Rika Fujimaru,17 Koichi Nakanishi,18 Yuko Shima,19 Masafumi Matsuo,20 Ming Juan Ye,1 Yoshimi Nozu,1 Naoya Morisada,1 Hiroshi Kaito,1 and Kazumoto Iijima1

Due to the number of contributing authors, the afﬁliations are listed at the end of this article.

ABSTRACT Background X-linked Alport syndrome (XLAS) is a progressive hereditary nephropathy caused by mutations in the COL4A5 gene. Genotype-phenotype correlation in male XLAS is relatively well established; relative to truncating mutations, nontruncating mutations exhibit milder phenotypes. However, transcript comparison between XLAS cases with splicing abnormalities that result in a premature stop codon and those with nontruncating splicing abnormalities has not been reported, mainly because transcript analysis is not routinely conducted in patients with XLAS. Methods We examined transcript expression for all patients with suspected splicing abnormalities who were treated at one hospital between January of 2006 and July of 2017. Additionally, we recruited 46 males from 29 families with splicing abnormalities to examine genotype-phenotype correlation in patients with truncating (n=21, from 14 families) and nontruncating (n=25, from 15 families) mutations at the transcript level. Results We detected 41 XLAS families with abnormal splicing patterns and described novel XLAS atypical splicing patterns (n=14) other than exon skipping caused by point mutations in the splice consensus sequence. The median age for developing ESRD was 20 years (95% confidence interval, 14 to 23 years) among patients with truncating mutations and 29 years (95% confidence interval, 25 to 40 years) among patients with nontruncating mutations (P=0.001). Conclusions We report unpredictable atypical splicing in the COL4A5 gene in male patients with XLAS and reveal that renal prognosis differs significantly for patients with truncating versus nontruncating splicing abnormalities. Our results suggest that splicing modulation should be explored as a therapy for XLAS with truncating mutations.

J Am Soc Nephrol 29: 2244–2254, 2018. doi: https://doi.org/10.1681/ASN.2018030228

Alport syndrome is an inherited type IV collagen disease Received March 1, 2018. Accepted May 23, 2018. that causes kidney disorder and usually develops into Published online ahead of print. Publication date available at ESRD, accompanied by sensorineural hearing loss and www.jasn.org. ocular abnormalities. X-linked Alport syndrome Correspondence: Dr. Kandai Nozu, Department of Pediatrics, (XLAS) accounts for approximately 85% of cases of Kobe University Graduate School of Medicine, 7-5-1 Kusunoki- Alport syndrome; notably pathogenic variants can be cho, Chuo, Kobe, Hyogo 6500017, Japan. Email: [email protected] detected in the COL4A5 (NM: 000495.4) gene, which u.ac.jp encodes the a5 chain of type IV collagen [a5(IV)].1 Copyright © 2018 by the American Society of Nephrology

2244 ISSN : 1046-6673/2908-2244 JAmSocNephrol29: 2244–2254, 2018 www.jasn.org CLINICAL RESEARCH

The genotype-phenotype correlation in male XLAS is relatively Significance Statement well established; missense mutations (i.e., nontruncating mutations) exhibit milder phenotypes compared with truncating mu- X-linked Alport syndrome (XLAS) is a progressive hereditary ne- tations.2–4 Jais et al.2 reported that, by the age of 30, the probability phropathy caused by mutations in the COL4A5 gene. Previous of developing ESRD was 90% in cases that included large re- studies have shown genotype-phenotype correlations, but this study is the first to demonstrate that splicing mutations that create a arrangements and small frameshift mutations, but 50% in cases premature stop codon and a truncated transcript are associated with missense mutations. Bekheirnia et al.3 also reported that the with worse prognosis. The investigators also find that, for splicing average age at ESRD onset was 22 years for those with large or small mutations, transcriptional analysis is necessary to accurately esti- deletions, 25 years for those with truncating mutations, and 37 mate renal prognosis and in silico predictive tools are not often fi years for those with missense mutations. Regarding splice site mu- useful. The ndings suggest consideration of therapeutic approaches to changing truncating mutation into nontruncating 2 tations, Jais et al. analyzed 29 families with mutations in consensus transcription for XLAS. splice sites, revealing that 70% of those developed ESRD by 30 years of age. In addition, Bekheirnia et al.3 studied 24 families with consensus splice sites and reported that the average age of onset of Among these, 71 (25%) exhibited truncating variants, 159 ESRD was 28 years old. These data suggest that cases with splice site (57%) exhibited nontruncating variants, and 49 (18%) exhibi- variants tend to show moderate severity (i.e., between the severities ted splicing variants. Among the 49 families with splicing var- observed with missense and nonsense mutations). However, in iants, eight families were excluded for the following reasons: these studies, transcript analysiswasnotconducted.Therefore,a possession of COL4A5 variants with somatic mosaic in a male comparison between truncating and nontruncating mutations (i.e., patient, one case; the 47,XXY karyotype, two cases; merging between cases where the number of deleted nucleotide numbers is with COL4A4 variant, one case; merging with missense or is not a multiple of 3) at the transcript level, to determine splicing COL4A5 variant, one case; merging with membranoprolifera- abnormalities in the COL4A5 gene, has not yet been performed. In tive GN, one case; and impossible to analyze mRNA, two cases. some other inherited diseases, in-frame splice site variants exhibit We included the remaining 41 families and described 14 pa- milder phenotypes relative to those caused by truncating mutations tients with atypical splicing patterns; other than exon skipping (e.g., Becker muscular dystrophy and Duchenne muscular dystro- caused by point mutations in the splice consensus sequence, phy [DMD]).5 In XLAS, it is critical to know the difference between these patterns had never been reported. In addition, to examine truncating transcripts, which result in a premature stop codon, and the genotype-phenotype correlation of male patients with nontruncating transcripts, in order to estimate renal prognosis, XLAS, we excluded 11 families with only female patients and perform genetic counseling, and develop further treatment strate- one family with exon 49 skipping because exons 49–51 are gies (e.g., exon skipping therapy, which is already approved by the known as a noncollagenous domain and the skipping of exon US Food and Drug Administration [FDA] for DMD). 49 may result in a severe phenotype. Ultimately, we included 29 However, it is very difficult to make a reliable prediction of families with 46 male patients; we divided these patients into splicing patternswithouttranscriptionalanalysis.Thusfar,wehave two groups with truncating or nontruncating variants at the detected 41 families with splicing abnormalities, including typical transcript level, then compared their clinical severity of disease. (n=16) and atypical (n=25) splicing patterns. Among them, we In our analysis, we included family members who had not un- have previously reported on 11 families with atypical splicing dergone their own DNA analysis but exhibited obvious urine patterns6; since then, we have analyzed an additional 14 families. abnormalities or renal dysfunction. Most patients were fol- In this study, we report on patients with atypical splicing lowed in a variety of local hospitals throughout Japan. Blood patterns (n=14) and examine the genotype-phenotype corre- samples and data were sent to our laboratory after acceptance lation in male patients with XLAS with variants in the COL4A5 of the request for mutational analysis. When we have detected gene that cause aberrant splicing, including those typical and variants that may possibly affect RNA splice processing, we atypical splicing patterns that have been proven by transcript have routinely analyzed the transcripts to confirm the muta- analysis (46 male patients from 29 families). tion-induced splicing abnormalities. In this study, we selected only patients with proven splicing variants.

METHODS Mutational Analyses Mutational analysis of COL4A5 was performed by several meth- Ethical Consideration ods: (1) targeted next-generation sequencing using a custom All procedures were reviewed and approved by the Institutional disease panel, including COL4A3, COL4A4,andCOL4A5 genes; Review Board of Kobe University School of Medicine. In- (2) conventional direct sequencing using the Sanger method for formed consent was obtained from patients or their parents. all exons and exon-intron boundaries; (3) multiplex ligation- dependent probe ampliﬁcation to detect copy-number varia- Participants and Inclusion Criteria tions; and (4) RT-PCR of mRNA and direct sequencing to detect A total of 279 families were genetically deﬁned as having XLAS abnormal splicing. We initially performed methods (1)or(2); if between January of 2006 and July of 2017 at Kobe University. no mutations were detected, we then performed methods (3)

J Am Soc Nephrol 29: 2244–2254, 2018 Splicing Abnormalities in XLAS 2245 CLINICAL RESEARCH www.jasn.org and/or (4). In addition, when we detected a suspected splicing procedure has been described previously.10–12 A mixture of site variant with methods (1)or(2), we also performed method FITC-conjugated rat mAb for the human a5(IV) chain (4). Genomic DNA was isolated from patient peripheral blood (H53) and Texas red–conjugated rat mAb for the human a2 leukocytes using the Quick Gene Mini 80 System (Kurabo In- (IV) chain (H25) was purchased from Shigei Medical Research dustries Ltd., Tokyo, Japan), according to the manufacturer’s Institute (Okayama, Japan). The epitopes were EAIQP at po- instructions. Amplification of all 51 specificexonsofCOL4A5, sitions 675–679 of the a2(IV) chain and IDVEF at positions next-generation sequencing, and multiplex ligation-dependent 251–255 of the a5(IV) chain. probe amplification were conducted as described previously.6–8 RNA from leukocytes was isolated using RNAlater RNA Stabili- Statistical Analyses zation Reagent (Qiagen Inc., Chatsworth, CA), then reverse All calculations were performed using standard statistical soft- transcribed into cDNA using the EcoDry Kit (Clontech Labora- ware (SAS version 9.4 for Windows; SAS, Cary, NC). The oc- tories, Inc., Palo Alto, CA). In all patients in this study, mRNA currence of events (age of developing ESRD) was analyzed via was extracted from peripheral blood leukocytes. the Kaplan–Meier method. To calculate P value, we used a shared frailty model.13 Under the model, the correlation In Vitro Splicing Assay within a frailty was modeled by shared frailty, following a For patient A422, we conducted a “minigene” in vitro splicing log normal distribution. We considered an association to be assay because we detected a remarkably abnormal splicing significant when the P value was ,0.05. pattern of exon skipping in an exonic variant. To create hybrid minigene constructs, we used the H492 vector, on the basis of the pcDNA 3.0 mammalian expression vector (Invitrogen, RESULTS Carlsbad, CA) (Supplemental Figure 1A) that we developed previously.9 We amplified genomic DNA from the patient and Abnormal Splicing Patterns Detected in Our Study from a sample of control peripheral leukocytes, using primers In total, 41 families were included in the study. Among these 41 for COL4A5 intron 8 to intron 13 that contained additional families, 11 exhibit splicing variants that we have already re- restriction sites. This enabled cloning of the PCR products ported, along with corresponding transcript analysis.6,7,14 into the multiple cloning site of the vector, located within an Furthermore, for two families (A43, A302), transcript analysis intron between exons A and B. The forward primer contained results have been reported by another group.15 These cases an NheI site (GCTAGC): (59-GCAGCTAGCCAGTGTACTC- were included in this study for purposes of genotype- TGGCCACTTCC-39), and the reverse primer contained a phenotype correlation analysis. BamHI site (GGATCC): (59-CGTGGATCCTGTTTGCAAGA- Here, we reveal the transcript analysis results of 14 families TAAAATAAGACAGTG-39). PCR products and the H492 vec- whose transcripts had never been reported and who showed tor were digested by NheIandBamHI, then ligated together to atypical splicing patterns, not the mere exon skipping by var- create both wild-type and variant hybrid (NM_000495.4: iation in the consensus AG-GT sequence (Figure 1, for larger c.548dupG) minigenes. The hybrid minigenes were checked view see Supplemental Figure 2). by sequencing, then transfected into HEK293Tand HeLa cells Of these 41 families, 32 families (78%) exhibited splice site via Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, variants (19 families possess variants at the consensus sequence MA). Twenty-four hours later, total RNA was extracted from of either the AG or GT site; the remaining 13 possess variants the cells using the RNeasy Plus Mini Kit (QIAGEN GmbH, out of these sites) (Table1) and five families (12%) exhibited Hilden, Germany). Two micrograms of total RNA was subjec- deep intronic variants creating novel exons; the remaining ted to reverse transcription using the EcoDry Kit, then PCR four families (10%) exhibited exonic variants, one of which was performed with a forward primer corresponding to a seg- created an aberrant splice site, whereas the other three induced ment upstream of exon A and a reverse primer complemen- exon skipping. Twenty-two families (54%) exhibited exon tary to a segment downstream of exon B, as previously skipping, 13 families (32%) revealed production of a new described (Supplemental Figure 1A).9 PCR products were an- splice site, five families (12%) revealed production of a cryptic alyzed by electrophoresis on a 1.2% agarose gel, followed by exon, and one family (2%) revealed production of a large exon direct sequencing. (Figure 1, Supplemental Figure 2, Table 1).

In Silico Splicing Assay In Vitro Splicing Assay We analyzed the strength of the splicing domain of all 29 fam- One patient (A422) showed an atypical splicing abnormality of ilies by in silico analysis, using the Human Splicing Finder exon 10 skipping by exonic variation (c.548dupG, Figure 1N, (http://www.umd.be/HSF3/). Supplemental Figure 2N). To clearly assess the probability, we performed an in vitro splicing assay. Electrophoresis of cDNA Immunohistochemical Analyses from the patient and from a wild-type fragment showed dou- Immunohistochemical analyses were performed with frozen ble and triple bands, respectively (Supplemental Figure 1B). sections of kidney or skin tissue. The immunohistochemical Although some bands revealed an absence of exons 9 and 10

2246 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 2244–2254, 2018 www.jasn.org CLINICAL RESEARCH

Figure 1. Continued.

J Am Soc Nephrol 29: 2244–2254, 2018 Splicing Abnormalities in XLAS 2247 CLINICAL RESEARCH www.jasn.org

Figure 1. Mutations and their consequences (the same figures are shown in Supplemental Figure 2 with larger scales). Upper panels show schemas of aberrant splicing (red lines). Normal splicing is indicated by black lines. The original and new splice sites and flanking sequences are shown below. Patients’ flanking genomic DNA and cDNA sequences are shown in the lower panels. (A) Patient ID A196. IVS23–1G.A eliminated the splice acceptor site of intron 23 to activate a new splice site, one nucleotide downstream. (B) Patient ID A333. IVS49+1 G.A disrupted the splicing donor site of intron 49, resulting in an intron 49 insertion, which creates a transcript with a 345-bp insertion. (C) Patient ID A424. IVS 6–1G.A altered the splice acceptor site of intron 6 one nucleotide downstream, which creates a transcript with a 1-bp deletion. (D) Patient ID A231, A258, A298. IVS35–4A.G altered the splice acceptor site of intron 35 three nucleotides upstream, which creates a transcript with a 3-bp insertion. (E) Patient ID A247. IVS 12+5 G.A disrupted the splice donor site of intron 12, resulting in exon 12 skipping, which creates a transcript with a 42-bp deletion. (F) Patient ID A299. IVS29+3 A. G disrupted the splice donor site of intron 29, resulting in exon 29 skipping, which creates a transcript with a 151-bp deletion. (G) Patient ID A323. IVS40–9C.G altered the splice acceptor site of intron 40 nine nucleotides upstream, which creates a transcript with a 9-bp insertion. (H) Patient ID A371. IVS 18+3_6 del AAGT disrupted the splice donor site of intron 18, resulting in exon 18 skipping, which creates a transcript with a 42-bp deletion. (I) Patient ID A384. IVS48–11A.G altered the splice acceptor site of intron 48 ten nucleotides upstream, which creates a transcript with a 10-bp insertion. (J) Patient ID A402. IVS27+4 del T disrupted the splice donor site of intron 27, resulting in exon 27 skipping, which creates a transcript with a 105-bp deletion. (K) Patients ID A452. IVS29+5 G.A

2248 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 2244–2254, 2018 www.jasn.org CLINICAL RESEARCH even in the wild-type fragments, a direct connection between Age at onset of ESRD differed signiﬁcantly between these two exon 9 and exon 11 was observed only within the sequence of groups. The median time (95% conﬁdence interval) from birth the patient (Supplemental Figure 1C). This indicates that to ESRD was as follows: truncating, 20 years (14 to 23 years); c.548dupG led to the skipping of exon 10, which was suppor- nontruncating, 29 years (25 to 40 years; P=0.001; Figure 2); over- ted by both in vivo and in vitro data. all, 25 years (21 to 34 years) (Table 2).

In Silico Splicing Assay Among 41 families, 19 were completely compatible with the DISCUSSION analyticresult of theHumanSplicing Finder(http://www.umd. be/HSF3/), including complete exon skipping and deep in- This is our second report of a case series for male patients with tronic variants; 17 of the remaining families yielded data XLAS with aberrant splicing; concurrently, it is the first report that enabled prediction of splicing defects, but could not of a comparative analysis of male patients with XLAS with predict a precise novel splice site; and the final five families aberrant splicing, thereby enabling distinction between trun- yielded data that were not sufficient for the prediction of splic- cating and nontruncating transcriptional variants.6 ing abnormalities (Supplemental Table 1). Genotype-phenotype shows a strong correlation in male XLAS; missense mutations result in milder phenotypes, rela- Immunohistochemical Analyses tive to truncating mutations such as nonsense mutations and The results of immunohistochemical analysis for a5(IV) col- deletions/insertions.2–4 However, all of the prior studies have lagen are shown in Supplemental Table 3. Of 41 total patients, grouped splice site variants into a single group without char- a5(IV) collagen staining was performed in 31 patients. acterizing the specific patterns of their transcriptional varia- Among them, 17 were male patients and these data are shown tions. Moreover, the effects of splicing mutations that result in Supplemental Table 4. In the nontruncating group (n=7), in a premature stop codon, designated “truncating transcript,” three of seven patients were positive for a5(IV) staining or nontruncating mutations caused by abnormal splicing, des- (43%). In contrast, none showed a5(IV) positivity in the trun- ignated “nontruncating transcript,” have yet to be elucidated. cating group (n=10). All three patients who were a5(IV) pos- Jais et al.2 reported that splice site variants confer a 70% prob- itive belonged to the nontruncating group (100%). These ability for development of ESRD at age 30 years. Bekheirnia results indicate that a5(IV) positivity in patients with splicing et al.3 reported that the average age of onset of ESRD was 28 variants at the genomic DNA level can predict nontruncating years for patients with splice site variants. However, our study transcript variants and later onset of ESRD. clearly showed that patients with disease caused by aberrant splicing exhibit different prognoses, depending on whether Patients Included in the Genotype-Phenotype their mutations are truncating or nontruncating transcript. Correlation Analysis To most accurately predict kidney prognosis, it is important During genotype-phenotype correlation analysis, we excluded to conduct transcriptional analysis when detecting either one patient (A333) with an insertion in intron 49 that resulted splice site mutations or aberrant-splicing mutations. in the creation of a stop codon; this was because exons 49–51 Transcriptional analysis is frequently difficult to conduct are known as a noncollagenous domain, the lack of which may because transcripts are not stable within tissues. An in silico cause a severe phenotype. Thus, we specifically limited the analysis tool to predict the effect of transcriptional variants, analysis to families with variants in the collagenous domain. such as the Human Splicing Finder (http://www.umd.be/ Therefore, we included 46 male patients, from 29 families HSF3/), may aid in prediction of the disruption of the original who exhibited 27 variants, in our genotype-phenotype corre- splice site. However, our study revealed that in silico analysis lation analysis (Figure 2, Supplemental Table 2). may not precisely identify the aberrant splice site and abnormal transcript. Indeed, 19 families were compatible with the in Genotype-Phenotype Correlation Analysis silico analysis in our study; however, for 22 families, the anal- Twenty-oneparticipantsfrom14familiesrevealedtruncatingtran- ysis tool could not predict the exact splicing sites (Supplemen- scriptional variants and 25 participants from 15 families revealed tal Table 1). Furthermore, the presence of a consensus splicing nontruncating transcriptional variants (Supplemental Table 2). variant at the AG-GTsite is usually suspected to result in exon

disrupted the splice donor site of intron 29, resulting in exon 29 skipping, which creates a transcript with a 151-bp deletion. (L) Patient ID A329. IVS21–367 C.T produced a new splice donor site, resulting in a cryptic exon activation between exons 21 and 22 and creating a transcript with a 93-bp insertion. (M) Patient ID A375. Mutation in last nucleotide of exon 25, C1948 G.T, disrupted the splice donor site of intron 25, resulting in exon 25 skipping, which creates a transcript with a 169-bp deletion. (N) Patient ID A422. Mutation of the second nucleotide of exon 10, C548 dup G, disrupted the splicing acceptor site of intron 9, resulting in exon 10 skipping, which creates a transcript with a 63-bp deletion. gDNA, genomic DNA.

J Am Soc Nephrol 29: 2244–2254, 2018 Splicing Abnormalities in XLAS 2249 2250

Table 1. Clinical and genetic ﬁndings in this study RESEARCH CLINICAL Previously ESRD Splicing Family Number Sex Effect Mutation Intron/Exon Transcription Reported Figure Age (yr) Variant cDNA Analysis ora fteAeia oit fNephrology of Society American the of Journal Splice consensus sequence mutations A7 Male Truncating 9 c.610–2A.T Intron 10 19-bp deletion Creating a novel Nozu et al.6a splice site A8 Male Truncating 16 c.3247–2A.C Intron 36 127-bp deletion Exon 37 skipping

A17 Female Truncating no (11 yo) c.3455–1G.A Intron 38 1-bp deletion Creating a novel Nozu et al.6a www.jasn.org splice site A27 Male Truncating no (5 yo) c.991–1G.A Intron 17 1-bp deletion Creating a novel Nozu et al.6a splice site A28 Female Nontruncating no (6 yo) c.2147–2A.G Intron 27 18-bp deletion Creating a novel Nozu et al.6a splice site A30 Female Nontruncating no (8 yo) c.990+1G.T Intron 17 54-bp deletion Exon 17 skipping A43 Male Truncating no (2 yo) c.1948+1G.A Intron 25 169-bp deletion Exon 25 skipping Wang et al.15 A91 Female Nontruncating no (3 yo) c.646–2A.G Intron 11 42-bp deletion Exon 12 skipping A97 Female Nontruncating no (18 yo) c.1166–1G.A Intron 19 174-bp deletion Exon 20 skipping A116 Female Nontruncating no (10 yo) c.646–1G.C Intron 11 42-bp deletion Exon 12 skipping A141 Female Truncating no (25 yo) c.3247–2A.G Intron 36 127-bp deletion Exon 37 skipping A196 Male Truncating no (7 yo) c.1588–1G.A Intron 23 1-bp deletion Creating a novel 1A splice site A244 Female Nontruncating no (18 yo) c.4069+1G.A Intron 44 72-bp deletion Exon 44 skipping A276 Male Nontruncating 29 c.892–2A.G Intron 15 45-bp deletion Exon 16 skipping A302 Female Truncating no (10 yo) c.1948+1G.A Intron 25 169-bp deletion Exon 25 skipping Wang et al.15 A326 Male Nontruncating no (12 yo) c.2678–1G.A Intron 31 90-bp deletion Exon 32 skipping A333 Male Truncating 23 c.4803+1G.A Intron 49 345-bp insertion Creating a 1B large exon A365 Male Nontruncating no (3 yo) c.3604+1G.A Intron 40 51-bp deletion Exon 40 skipping A424 Male Truncating no (11 yo) c.385–1G.A Intron 6 1-bp deletion Creating a novel 1C splice site Suspected splice site mutations A121 Male Nontruncating no (25 yo) c.546+2_3insT Intron 9 81-bp deletion Exon 9 skipping Hashimura et al.7a A128 Male Nontruncating no (46 yo) c.2042–18A.G Intron 26 105-bp deletion Exon 27 skipping Nozu et al.6a mScNephrol Soc Am J A158 Female Nontruncating no (14 yo) c.2245–8T.A Intron 28 6-bp insertion Creating a novel Nozu et al.6a splice site A231 Female Nontruncating no (3 yo) c.3107–4A.G Intron 35 3-bp insertion Creating a novel 1D splice site A247 Female Nontruncating no (19 yo) c.687+5G.A Intron 12 42-bp deletion Exon 12 skipping 1E

29: A258 Male Nontruncating no (3 yo) c.3107–4A.G Intron 35 3-bp insertion Creating a novel 1D

2244 splice site A298 Male Nontruncating 25 c.3107–4A.G Intron 35 3-bp insertion Creating a novel 1D – 24 2018 2254, splice site A299 Female Truncating no (2 yo) c.2395+3A.G Intron 29 151-bp deletion Exon 29 skipping 1F www.jasn.org CLINICAL RESEARCH

skipping; however, in our study re- 1I 1J 1L 1K 1H 1G 1N 1M sults, seven of 19 cases with varia- Figure tion in the AG-GTsite did not show mere exon skipping. Thus, in silico a a a a a

a analysis is frequently unpredictable. 6 6 6 6 6

14 Previously, it was reported that et al. et al. et al. et al. et al. male patients with XLAS with pos- et al. a Reported

Previously itive staining for 5(IV) in kidney Fu Nozu Nozu Nozu Nozu cDNA Analysis tissue tended to show milder clinical phenotypes and carried nontruncating mutations (mostly missense mutations).7,16 In our co- hort, three of seven patients carrying nontruncating transcriptional Variant Splicing splice site splice site splice site Creating a Creating a Creating a Creating a novel exon novel exon novel exon novel exon novel exon variants were a5(IV) positive. In contrast, no patients carrying nontruncating transcriptional variants were a5(IV) positive (Supplemen- tal Table 4). Although the sample number was not large, our study revealed the possibility that, for pa- with stop codon with stop codon with stop codon with stop codon tients with splice site variants at the genome level, a5(IV) positivity in patient tissues can predict nontruncating transcript variants. Therefore, immunohistochemical Exon 25 169-bp deletion Exon 25 skipping Exon 15 35-bp deletion Creating a novel Exon 41 186-bp deletion Exon 41 skipping Nozu Exon 10 63-bp deletionanalysis Exon 10 skipping might aid in predicting the prognosis of those patients, po- tentially replacing mRNA analysis G Intron 47 84-bp insertion T Intron 21 93-bp insertion Creating a G Intron 25 106-bp insertion G Intron 47 74-bp insertion T Intron 10 123-bp insertion G Intron 47 10-bp insertion Creating a novel A Intron 29 151-bp deletion Exon 29 skipping that is labor-intensive, expensive, . G Intron 40 9-bp insertion Creating a novel . . . . T A . . T . . . and not readily available in most . 9C 11A 367C – 345A

– countries. – – Recently, our group has reported Mutation Intron/Exon Transcription c.876A c.548dupG c.1948G c.3790G that a splicing minigene assay is valuable in determining the variant’seffectonsplicingintransfec- ted cells, in cases of inherited kidney diseases.9,17 In addition, 18 11

ESRD 18

Age (yr) Malone et al. have recently reported that the minigene assay was effective for detecting splicing abnormality in a patient with XLAS. We also conducted the minigene assay in our patient A422, which revealed an exonic variant that caused the exon skipping. Although some unexpected variants were observed both in the patient sequence and Male Nontruncating 29 MaleMale TruncatingMale Truncating no (12 yo) Nontruncating no (3 yo)Male c.609+875G 22Male c.4510+1754T Truncating Truncating c.1424 MaleFemale Nontruncating Truncating no (12 yo) no (6 c.1032+3_6 yo) delAAGT Intron 18 c.4511 42-bp deletion Exon 18 skipping Male Truncating no (6 yo) c.4511 MaleFemale Nontruncating Truncating no (7 yo) no (14 yo) c.2146+4delT c.2395+5G Intron 27 105-bp deletion Exon 27 skipping Male Nontruncating 24 Male Truncating no (22 yo) c.1948+894C the wild-type fragment, it may be appropriate to assume that exon 10 skipping was caused by the ex- Continued onic variant detected in the patient. Recently, splicing modulation of exon skipping by antisense oli- A323 Female Non-truncating no (18 yo) c.3605 A422 A48 A126 A217 A329 A178 A375 A371 A384 A21 A402 A452 A19 Shows the exact reported case. Table 1. Family Number Sex Effect yo, years old. a Exonic mutations Deep intronic mutations creating cryptic exon gonucleotide has been revealed as

J Am Soc Nephrol 29: 2244–2254, 2018 Splicing Abnormalities in XLAS 2251 CLINICAL RESEARCH www.jasn.org

splicing patterns have been reported6,23; notably, these differences were NS, and these reports largely served to confirm the existence of additional splicing variants. One limitation of our study is that we solely confirmed the splicing variants by mRNA extracted from leukocytes; however, the is- sues associated with this method may be limited. Our study has additional limitations. First, clinical information that might influ- ence the prognosis, such as continuous angiotensin-converting enzyme inhibitor or angiotensin receptor blocker treatment, hypertension, nonsteroidal anti-inflam- matory drug use, and/or incidental glomerular disease, was not fully analyzed. Most of the patients that developed ESRD were not treated by angiotensin-converting enzyme inhibitor or angiotensin receptor Figure 2. Probability of developing ESRD according to mutation types. Solid line indicates patients with truncating splicing mutations (n=21). The median age for de- blocker; however, younger patients have be- veloping ESRD was 20 years. Dash-dots indicate nontruncating splicing mutations gun treatment with those drugs, and we (n=25). The median age for developing ESRD was 29 years. could not include these in our analysis. An- other limitation is that our sample size was relatively small; thus, our study contains particularly valuable in cases with truncating mutations; some familial cases or cases with the same nucleotide variant. hence, the splicing modulation induces a nontruncating In conclusion, this is the first study to provide a comparison transcription and can produce a protein that might partially of the kidney prognosis between truncating and nontruncating compensate for the loss of full-length protein, thereby ame- splicing variants in male patients with XLAS. We showed that liorating the symptoms of the disease.19 Antisense oligonu- nontruncating splicing variants exhibit milder phenotypes cleotide was approved by the FDA as a treatment for DMD, compared with truncating splicing variants. This suggests with the aim of changing truncating mutations into exon that transcriptional analysis is necessary to accurately estimate skipping that may lead to nontruncating mutations.20 Sim- renal prognosis and provide better information during genetic ilar therapeutic strategies may be available in the future for counseling. Our results also present factual evidence support- patients with genetically diagnosed XLAS; importantly, we ing the efficiency of therapeutic approaches that use splicing have shown that transcriptional in-frame variants can lead modulation that change truncating transcription into non- to the milder phenotype in our study, which may serve as truncating transcription within patients with XLAS. factual evidence of the efficiency of exon skipping therapy for XLAS with truncating mutations. In most cases, mRNAextracted from the kidney itself is hard to obtain. Fortunately, peripheral blood leukocytes express a ACKNOWLEDGMENTS sufficient quantity of COL4A5 transcripts and have been used as an alternative source for transcript analysis, in order to We thank Ryan Chastain-Gross from the Edanz Group (www.edan- confirm the pathogenicity of suspected splice site variants.6 zediting.com/ac) for editing a draft of this manuscript. However, alternative splicing could occur in a tissue-specific This study was supported by a grant from the Ministry of Health, manner.21,22 Regarding COL4A5, multiple organ-specific Labour and Welfare of Japan for Research on Rare Intractable Diseases in the Kidney and Urinary Tract (H24-nanchitou [nan]-ippan-041 to K.I.) in the “Research on Measures for Intractable Diseases” Project; by Grants-in-Aid for Scientific Research (KAKENHI) from the Table 2. Distribution of the participants Ministry of Education, Culture, Sports, Science and Technology of Variable Total (Censored Case) ESRD Age, Median (95% CI) Japan (subject ID: 15K09691 to K. Nozu and 17H04189 to K.I.); and Nontruncating 25 (10) 29 (25 to 40) by Japan Agency for Medical Research and Development (AMED) Truncating 21 (11) 20 (14 to 23) under grant number 7930006 to K. Nozu and K.I. Total 46 (21) 25 (21 to 34) T.H. and K.I. designed the studyconcept and wrote the manuscript. 95% CI, 95% confidence interval. K. Nozu interpreted the data and wrote the manuscript. T.Y., S.M.,

2252 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 2244–2254, 2018 www.jasn.org CLINICAL RESEARCH

Keita Nakanishi, J.F., A.A., M. Kitamura, M. Kawano, W.S., C.K., A.I., 10. Sado Y, Kagawa M, Kishiro Y, Sugihara K, Naito I, Seyer JM, et al.: Es- K.K., K.O., S.F., M.O., T.I., A.M., E.S., R.F., Koichi Nakanishi, Y.S., tablishment by the rat lymph node method of epitope-defined M.J.Y., and Y.N. collected and interpreted the data. T.O. interpreted monoclonal antibodies recognizing the six different alpha chains of human type IV collagen. Histochem Cell Biol 104: 267–275, 1995 the data and conducted the statistical analyses. M.M., N.M., and H.K. 11. Naito I, Kawai S, Nomura S, Sado Y, Osawa G; Japanese Alport Net- critically reviewed the manuscript. All authors read and approved the work: Relationship between COL4A5 gene mutation and distribution of final version of the manuscript. type IV collagen in male X-linked Alport syndrome. Kidney Int 50: 304– 311, 1996 12. Nakanishi K, Iijima K, Kuroda N, Inoue Y, Sado Y, Nakamura H, et al.: DISCLOSURES Comparison of alpha5(IV) collagen chain expression in skin with disease severity in women with X-linked Alport syndrome. JAmSocNephrol9: K.I. has received grant support from Daiichi Sankyo Co., Ltd. K.I. has re- 1433–1440, 1998 ceived consulting fees from Takeda Pharmaceutical Company and Kyowa 13. Knox KL, Bajorska A, Feng C, Tang W, Wu P, Tu XM: Survival analysis for Hakko Kirin Co., Ltd. K. Nozu has received lecture fees from Novartis Phar- observational and clustered data: An application for assessing indi- maceuticals Corporation. K.I. and K. Nozu have filed a patent application on vidual and environmental risk factors for suicide. Shanghai Jingshen the development of antisense nucleotides for exon skipping therapy in Alport Yixue 25: 183–194, 2013 syndrome. 14. Fu XJ, Nozu K, Eguchi A, Nozu Y, Morisada N, Shono A, et al.: X-linked Alport syndrome associated with a synonymous p.Gly292Gly mutation alters the splicing donor site of the type IV collagen alpha chain 5 gene. Clin Exp Nephrol 20: 699–702, 2016 REFERENCES 15. Wang F, Wang Y, Ding J, Yang J: Detection of mutations in the COL4A5 gene by analyzing cDNA of skin fibroblasts. Kidney Int 67: 1268–1274, 1. Kashtan CE: Alport syndrome and thin glomerular basement mem- 2005 brane disease. JAmSocNephrol9: 1736–1750, 1998 16. Said SM, Fidler ME, Valeri AM, McCann B, Fiedler W, Cornell LD, et al.: 2. Jais JP, Knebelmann B, Giatras I, De Marchi M, Rizzoni G, Renieri A, Negative staining for COL4A5 correlates with worse prognosis and et al.: X-linked Alport syndrome: Natural history in 195 families and more severe ultrastructural alterations in males with alport syndrome. genotype- phenotype correlations in males. J Am Soc Nephrol 11: 649– Kidney Int Rep 2: 44–52, 2016 657, 2000 17. Nakanishi K, Nozu K, Hiramoto R, Minamikawa S, Yamamura T, 3. Bekheirnia MR, Reed B, Gregory MC, McFann K, Shamshirsaz AA, Fujimura J, et al.: A comparison of splicing assays to detect an intronic Masoumi A, et al.: Genotype-phenotype correlation in X-linked Alport variant of the OCRL gene in Lowe syndrome. Eur J Med Genet 60: 631– syndrome. J Am Soc Nephrol 21: 876–883, 2010 634, 2017 4. Gross O, Netzer KO, Lambrecht R, Seibold S, Weber M: Meta-analysis 18. Malone AF, Funk SD, Alhamad T, Miner JH: Functional assessment of a of genotype-phenotype correlation in X-linked Alport syndrome: Im- novel COL4A5 splice region variant and immunostaining of plucked pact on clinical counselling. Nephrol Dial Transplant 17: 1218–1227, hair follicles as an alternative method of diagnosis in X-linked Alport 2002 syndrome. Pediatr Nephrol 32: 997–1003, 2017 5. Monaco AP, Bertelson CJ, Liechti-Gallati S, Moser H, Kunkel LM: An 19. Havens MA, Hastings ML: Splice-switching antisense oligonucleotides explanation for the phenotypic differences between patients bearing as therapeutic drugs. Nucleic Acids Res 44: 6549–6563, 2016 partial deletions of the DMD locus. Genomics 2: 90–95, 1988 20. Drug Approval Package: Exondys 51 Injection (eteplirsen), https:// 6. Nozu K, Vorechovsky I, Kaito H, Fu XJ, Nakanishi K, Hashimura Y, et al.: www.accessdata.fda.gov/drugsatfda_docs/nda/2016/206488_TOC. X-linked Alport syndrome caused by splicing mutations in COL4A5. cfm, FDA, 2016 Clin J Am Soc Nephrol 9: 1958–1964, 2014 21. Black DL: Mechanisms of alternative pre-messenger RNA splicing. 7. Hashimura Y, Nozu K, Kaito H, Nakanishi K, Fu XJ, Ohtsubo H, et al.: Annu Rev Biochem 72: 291–336, 2003 Milder clinical aspects of X-linked Alport syndrome in men positive for 22. Ramanouskaya TV, Grinev VV: The determinants of alternative RNA the collagen IV a5chain.Kidney Int 85: 1208–1213, 2014 splicing in human cells. Mol Genet Genomics 292: 1175–1195, 2017 8. Yamamura T, Nozu K, Fu XJ, Nozu Y, Ye MJ, Shono A, et al.: Natural 23. Martin PH, Tryggvason K: Two novel alternatively spliced 9-bp exons in history and genotype-phenotype correlation in female X-linked alport the COL4A5 gene. Pediatr Nephrol 16: 41–44, 2001 syndrome. Kidney Int Rep 2: 850–855, 2017 9. Nozu K, Iijima K, Kawai K, Nozu Y, Nishida A, Takeshima Y, et al.: In vivo and in vitro splicing assay of SLC12A1 in an antenatal salt-losing tu- bulopathy patient with an intronic mutation. Hum Genet 126: 533–538, This article contains supplemental material online at http://jasn.asnjournals. 2009 org/lookup/suppl/doi:10.1681/ASN.2018030228/-/DCSupplemental.

AFFILIATIONS

1Department of Pediatrics, Kobe University Graduate School of Medicine, Kobe, Japan; 2Clinical and Translational Research Center, Kobe University Hospital, Kobe, Japan; 3Department of Pediatrics, Osaka Medical College, Osaka, Japan; 4Department of Nephrology, Nagasaki University Hospital, Nagasaki, Japan; 5Department of Rheumatology, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan; 6Department of Pediatrics, Japan Community Health Care Organization Kyushu Hospital, Sapporo, Hokkaido, Japan; 7Department of Nephrology and Hypertension, Osaka City General Hospital, Osaka, Japan; 8Department of Nephrology, Toranomon Hospital, Tokyo, Japan; 9Department of Nephrology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan; 10Division of Nephrology and Rheumatology, National Center for Child Health and Development, Tokyo, Japan; 11Department of Urology, Ehime Prefectural Central Hospital, Ehime, Japan; 12Division of Nephrology, Saitama Children’s Medical Center, Saitama, Japan; 13Department of

J Am Soc Nephrol 29: 2244–2254, 2018 Splicing Abnormalities in XLAS 2253 CLINICAL RESEARCH www.jasn.org

Pediatrics, Faculty of Medicine Saga University, Saga, Japan; 14Department of Pediatrics, Nippon Medical School, Tokyo, Japan; 15Department of Pediatrics, Faculty of Medicine Kagoshima University, Kagoshima, Japan; 16Department of Pediatrics, University of Yamanashi, Yamanashi, Japan; 17Department of Pediatrics, Osaka General Hospital, Osaka, Japan; 18Department of Child Health and Welfare (Pediatrics), Graduate School of Medicine, University of the Ryukyus, Okinawa, Japan; 19Department of Pediatrics, Wakayama Medical University, Wakayama, Japan; and 20Department of Physical Therapy, Faculty of Rehabilitation, Kobe Gakuin University, Kobe, Japan

2254 Journal of the American Society of Nephrology J Am Soc Nephrol 29: 2244–2254, 2018