Systematic Review of Genotype-Phenotype Correlations in Frasier Syndrome

Introduction Frasier syndrome (FS) is a rare inherited kidney disease caused by intron 9 splicing variants of WT1. For wild-type WT1, 2 active splice donor sites in intron 9 cause a mixture of 2 essential transcripts (with or without lysine-threonine-serine [+/KTS or −KTS]), and imbalance of the +KTS/−KTS ratio results in the development of FS. To date, 6 causative intron 9 variants have been identified; however, detailed transcript analysis has not yet been conducted and the genotype-phenotype correlation also remains to be elucidated. Methods We conducted an in vitro minigene splicing assay for 6 reported causative variants and in vivo RNA sequencing to determine the +KTS/−KTS ratio using patients’ samples. We also performed a systematic review of reported FS cases with a description of the renal phenotype. Results The in vitro assay revealed that although all mutant alleles produced −KTS transcripts only, the wild-type allele produced both +KTS and −KTS transcripts at a 1:1 ratio. In vivo RNA sequencing showed that patients’ samples with all heterozygous variants produced similar ratios of +KTS to −KTS (1:3.2−1:3.5) and wild-type kidney showed almost a 1:1 ratio (1:0.85). A systematic review of 126 cases clarified that the median age of developing ESKD was 16 years in all FS patients, and there were no statistically significant differences between the genotypes or sex chromosome karyotypes in terms of the renal survival period. Conclusion Our study suggested no differences in splicing pattern or renal survival period among reported intron 9 variants causative of FS.

F rasier syndrome is a rare inherited disease characterized by steroid-resistant nephrotic syndrome with proteinuria that begins in early childhood and progressively worsens with age. It is also associated with a high risk of developing ESKD, gonadal tumors, and male pseudohermaphroditism (female external genitalia with XY sex chromosomes). It is caused by splice donor site variants in intron 9 of the Wilms' tumor 1 (WT1 [NM_024426.6]) gene, with 6 variants identified thus far (c.1432þ1 G>A, c.1432þ2 T>C, c.1432þ4 C>T, c.1432þ5 G>A, c.1432þ5 G>T, and c.1432þ6 T>A). [1][2][3][4] WT1 is composed of 10 exons and encodes a transcription factor with 4 zinc finger motifs; this transcription factor plays an important role in early gonadal and renal development by regulating the sex-determining region Y. 5,6 Two active splice donor sites of intron 9 result in a mixture of 2 transcripts with or without 3 amino acids (þKTS or ÀKTS) between zinc fingers 3 and 4. 7 Although the roles of þKTS and ÀKTS isoforms have not yet been completely clarified, it has been proposed that the ÀKTS isoform may function mostly as a transcriptional regulator, whereas the þKTS isoform acts predominantly at the posttranscriptional level. 8,9 Furthermore, studies in animal models have also shown that these isoforms are essential for survival. Hammes et al. 10 introduced the heterozygous variant into intron 9 of Wt1 in mice using a Cre-loxP strategy. These model mice showed a reduction of þKTS level and developed glomerulosclerosis, representing a model of FS. In contrast, the mice with homozygosity of this variant showed a complete lack of the þKTS isoform and died within 24 hours after birth due to kidney defects. 10 It is reported that the normal range of the WT1 þ/ ÀKTS isoform ratio range is between approximately 5:2 and 1:1 (average 3:2), whereas intron 9 variants lead to a shift in this ratio to between 1:2 and 1:5 (average 1:2) in FS (Table 1), 3 which indicates marked variability. However, all of these studies quantified the transcripts using a semiquantitative method based on agarose gel electrophoresis and need to be reexamined by quantitative analysis using RNA sequencing. Furthermore, in vitro assays on the role of intron 9 variants have been performed in 3 studies; however, these studies were limited to only 4 of the 6 known variants (c.1432þ2 T>C, c.1432þ4 C>T, c.1432þ5 G>A, and c.1432þ5 G>T). 2,3,11,12 Here, we conducted an in vitro splicing assay to measure the þ/ÀKTS ratio for all reported intronic variants (c.1432þ1 G>A, c.1432þ2 T>C, c.1432þ4 C>T, c.1432þ5 G>A, c.1432þ5 G>T, and c.1432þ6 T>A) and a variant newly identified in our study (c.1432þ3 G>T). In addition, in vivo RNA sequencing using patients' kidney-derived samples was conducted for 2 major variants (c.1432þ4 C>T, c.1432þ5 G>A) and a newly identified variant (c.1432þ3 G>T). Furthermore, we conducted a systematic review of previously reported cases of FS. With these analyses, we aimed to uncover genotype-phenotype correlations in FS and to determine differences in splicing pattern and clinical phenotype, mainly renal survival period, among the variants.

In Vitro Splicing Assay
For in vitro splicing experiments, hybrid minigene constructs were created by inserting a test sequence comprising exon 9 of WT1 and its flanking introns into a multiple cloning site within the intron between exons A and B of the minigene construct (H492) built into the pcDNA 3.0 mammalian expression vector (Thermo Fisher Scientific, Waltham, MA) as reported previously 13,14 (Supplementary Figure S1). The test sequence was obtained by amplifying the control sample by polymerase chain reaction using primers that corresponded to introns 8 and 9 (Supplementary Table S1). These primers included NheI and BamHI restriction enzyme recognition sites, respectively. The amplified products were digested with NheI and BamHI (New England Biolabs, Ipswich, MA) and inserted into the minigene that had been digested with the same restriction enzymes. Furthermore, constructs containing each of the reported intronic variants were generated by mutagenesis (the primers used are listed in Supplementary Table S2).
Using this method, we constructed both wild-type and mutant hybrid minigenes that carried exon 9 and either the wild-type or a mutated (c.1432þ1 G>A, c.1432þ2 T>C, c.1432þ3 G>T, c.1432þ4 C>T, c.1432þ5 G>A, c.1432þ5 G>T, and c.1432þ6 T>A) intron 9 sequence. After the construct sequences had been confirmed by Sanger sequencing, the hybrid minigenes were transfected into HeLa and HEK293T cells with Lipofectamine 3000 (Thermo Fisher Scientific). These 2 cell types were chosen because both of them showed high efficiency for in vitro transfection and expressed sufficient levels of transfected mRNA in previous analysis for genes expressed predominantly in the kidney. 15,16 Cells were harvested 24 hours after transfection, and total RNA was extracted with a RNeasy Plus Mini Kit (Qiagen, Hilden, Germany). One microgram of total RNA was subjected to reverse transcription with the RNA to cDNA EcoDry Premix (Doubled Primed; Takara Bio, Otsu, Japan), and polymerase chain reaction was subsequently performed using a forward primer corresponding to a segment of the upstream exon A and a reverse primer complementary to a segment of the downstream exon B as previously described. 13 Polymerase chain reaction products were analyzed with the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA) and direct sequencing (3130 Genetic Analyzer, Thermo Fisher Scientific).
In Vivo mRNA Analysis (Targeted RNA Sequencing) To clarify the þKTS/ÀKTS mRNA ratio for in vivo samples, we analyzed patients' samples for the most common variants (c.1432þ4 C>T and c.1432þ5 G>A). Furthermore, we analyzed the sample with the novel variant c.1432þ3 G>T to confirm splicing abnormality. Total RNA was extracted from urine-derived cells (þ3, þ4) and kidney (þ5) using the RNeasy Plus Mini Kit and ISOGEN RNA extraction reagent (Nippon Gene Co., Toyama, Japan), respectively. Samples of other variants were not available because of their rarity. We also obtained the wild-type kidney total RNA ( . þKTS/ÀKTS ratios were calculated from the average read depth in the last 9 bases of exon 9 for each þKTS/ÀKTS mRNA (Supplementary Figure S1).

Systematic Review
A systematic search of the PubMed database (https:// www.ncbi.nlm.nih.gov/pubmed/) was performed, and only those studies published in English and before December 2020 were included. The following search terms were used: "Frasier syndrome" or "Nephrotic syndrome AND WT1." An additional search was conducted by selecting articles that included genotype and clinical data relevant to the current study, such as age of onset of nephrotic syndrome or ESKD, sex chromosomes, external genitalia, or tumor, from the reference sections of review articles. We constructed renal survival curves by Kaplan-Meier analysis to determine the prognosis and median age of onset of ESKD for each variant, and we determined that as the median age of onset of ESKD became younger, it increased in severity. Furthermore, we summarized the phenotype data reported for each variant including clinical descriptions relating to the external genitalia and sex chromosomes. In this study, patients were classified into the following 3 groups depending on their sex chromosome karyotype and the presence of disorder of sexual development (DSD): XX female (all patients with XX karyotype), XY female (XY karyotype with DSD), and XY male (XY karyotype without DSD). The definition of nephrotic syndrome in this study followed the description of each case report, and many of those reports did not include the detailed definition of nephrotic syndrome. Gonadal abnormality was defined as a mismatch between the sex of the chromosomes and the external genital phenotype.

Statistical Analysis
All calculations were performed using standard statistical software (JMP for Mac, version 14; SAS Institute, Cary, NC). The occurrence of events (renal survival period) was examined using the Kaplan-Meier and logrank tests. Associations were considered to be statistically significant when P values were < 0.05.

In Vitro Splicing Assay
We conducted an in vitro splicing assay of wild-type and mutant minigene constructs to determine the splicing patterns resulting from the 6 intron 9 variants reported to date and a variant that we discovered (c.1432þ3 G>T) associated with FS. Polymerase chain reaction products from the wild-type construct produced 2 bands following capillary electrophoresis ( Figure 1a). Semiquantitative analysis with a bioanalyzer revealed that the molecular ratio of these 2 bands was approximately 1:1 (39.6 nmol/l:40.5 nmol/l). Sequencing of the amplicons from the wild-type construct indicated the presence of both þKTS and ÀKTS transcripts, whereas all mutant constructs produced only ÀKTS transcripts ( Figure 1).

In Vivo mRNA Analysis
We conducted an in vivo mRNA analysis for 3 variants that were available for analysis: 2 variants from urinederived cells (heterozygotes of 1432þ3 G>T and c.1432 þ4 C>T) and 1 from a kidney biopsy sample (heterozygote of c.1432 þ5G>A), along with the wildtype kidney sample. The average depth of sequencing in the target region ranged from 394 to 1639 (Table 1). Our analysis revealed that the þKTS/ÀKTS ratio was 1:0.85 in a wild-type human kidney but 1:3.2 in c.1432þ4 C>T and 1:3.5 in c.1432þ3 G>T and c.1432þ5 G>A ( the newly detected variant c.1432þ3 G>T showed almost the same þ/ÀKTS ratio in RNA sequencing.

Systematic Review
Genetic and clinical data were obtained from a total of 126 cases: 115 cases described in previous reports 6,11,12, and 11 cases diagnosed at our institution (    Table S2). The results showed that 85% of patients who were XY female and none with XX chromosomes developed a gonadal tumor. Interestingly, 40% (2 in 5) of patients who were XY male developed a gonadal tumor, both of whom had been diagnosed with bilateral cryptorchidism. 36,45 Regarding renal symptoms, among the 126 cases, there were descriptions of the age of onset of obvious proteinuria or nephrotic syndrome and kidney function for 102 and 116 cases, respectively. The results showed that the median age of onset of proteinuria was 4 years and that 58 cases had progressed to ESKD. We also created renal survival curves for all variants as well as individually for 2 major intronic variants (c.1432þ4 C>T and c.1432þ5 G>A) using Kaplan-Meier analysis (c.1432þ1 G>A, n ¼ 2; c.1432þ2 T>C, n ¼ 2; c.1432þ3G>T, n ¼ 1; c.1432þ4 C>T, n ¼ 61; c.1432þ5 G>A, n ¼ 47; c.1432þ5 G>T, n ¼ 2; and c.1432þ6 T>A, n ¼ 1) (Figure 2a and b). This analysis revealed that the median age of developing ESKD in the whole cohort was 16 years old. There was no significant difference between the 2 major variants c.1432þ4 C>T (n ¼ 61) and c.1432þ5 G>A (n ¼ 47) in the age of ESKD onset (17 vs. 15 years, P ¼ 0.62, log-rank test; Figure 3a).
We divided the patients according to their sex chromosome karyotypes and analyzed the groups for differences in renal prognosis. No significant difference   in the age of ESKD onset was found between the patients with XX chromosomes (n ¼ 28) and XY chromosomes (n ¼ 84) (29 vs. 16 years, P ¼ 0.09, log-rank test). In addition, to analyze whether the presence of DSD had any effect on renal prognosis, we classified the patients into the following 3 groups: group 1, XX (n ¼ 28); group 2, XY with DSD (n ¼ 74); and group 3, XY without DSD (n ¼ 10). However, there were no significant differences in the renal survival period among these groups (Figure 3b).

DISCUSSION
In this study, we conducted in vitro and in vivo splicing assays for all reported and newly identified variants in intron 9 of WT1 resulting in FS. In addition, we examined the genetic and clinical characteristics of reported patients with FS via a systematic review. Our results confirmed the presence of an abnormal splicing pattern in all variants both in vitro and in vivo. In addition, þKTS/ÀKTS mRNA ratios were equal for all variants evaluated in both analyses. Furthermore, a systematic review showed that there was no difference in renal prognosis according to genotype. Therefore, it was suggested that genotype does not affect renal prognosis in FS caused by WT1 intron 9 variants. In previous studies, transcript analysis was only performed for 4 of the intron 9 variants associated with FS: c.1432þ2 T>C, c.1432þ4 C>T, c.1432þ5 G>A, and c.1432þ5 G>T. These studies revealed that only the ÀKTS transcript is produced from mutant alleles, whereas both þKTS and ÀKTS transcripts are produced from the wild-type allele. 2,12 Here, our in vitro assay revealed identical results for these 4 variants as well as for the other 2 rare variants (c.1432þ1 G>A and c.1432þ6 T>A) and a new variant (c.1432þ3 G>T). Thus, it was confirmed that all intronic variants identified as being associated with FS to date disrupt the original splice donor site, resulting in production of only the ÀKTS transcript.
This study also focused on the production of þKTS/ ÀKTS mRNA in renal tissue of the patients. Previous studies quantified the transcripts in patients by using a semiquantitative method based on agarose gel electrophoresis to determine the ratio of þKTS/ÀKTS transcripts, but the reported ratios show wide variation (Table 1). We quantified this ratio by RNA sequencing, revealing a þKTS/ÀKTS ratio of 1:0.85 in normal kidney and 1:3.2 to 3.5 in patients' samples. Considering the ratio of þKTS/ÀKTS in a normal kidney, it was revealed that the þKTS transcript was not produced at equivalent levels for all analyzed variants. Although the performance of RNA sequencing for only 1 sample in each variant prevented us from examining the differences statistically, together with the in vitro assay, we concluded that all variants causing FS show the same splicing pattern.
In research on rare inherited diseases, large-scale clinical data on the disease including the genotype-   28 Consistent with this, our meta-analysis of data from 83 patients showed no significant difference in the ratio of these patients between variants c.1432þ4 C>T and c.1432þ5 Our study also focused on the influence of sex chromosome karyotypes and sexual development on renal prognosis. Sex differences in phenotype have been well established for various kidney diseases. In general, women are protected against renal diseases compared with men through estrogen's inhibition and androgen's activation of the renin-angiotensin system. 59 Therefore, women show a lower risk of CKD progression than men. 60 In addition, the renal prognosis in women with autosomal dominant polycystic kidney disease is better than that in men. 61 However, women do not always have a lower incidence or severity of renal diseases; for example, women have been reported to have a higher risk of diabetic kidney disease. 62 In this study, the median renal survival period tended to be longer in patients with a sex chromosome pattern of XX than in those with XY, although the difference was not statistically significant (29 vs. 16 years, P ¼ 0.09, log-rank test). In addition, the renal prognosis in patients with XY chromosomes was not influenced by the presence of DSD (the median age of developing ESKD was 16 in DSDþ vs. 15 in DSDÀ, P ¼ 0.84, log-rank test). Taking these findings together, differences in sex chromosome karyotypes and sexual development are not considered to influence the renal prognosis in FS.
Our study has some limitations. First, our analysis of the genotype-phenotype correlation is based on a retrospective systematic review. Thus, we could not collect clinical information regarding treatment or factors that might worsen renal function. Second, some of the reports that we analyzed in this study only focused on renal manifestations, with no detailed description of the sexual phenotype or tumorigenesis. The abnormalities of external genitalia seen in FS range from obvious discrepancies between the sex chromosome karyotypes and external genitalia to more subtle abnormalities. Therefore, the present study model may not be sufficient to clarify the correlation between renal genotype and the phenotype of sexual development in detail. However, we believe that our findings on the genotype and renal phenotype correlation in FS, which is the primary focus of this study, are important for clinicians, patients, and their families. We also conducted in vivo mRNA analysis only for urine-derived cell or kidney samples and not for other tissues such as gonads and blood samples. Therefore, we cannot assess the consistency of transcripts among different tissues. We also could not assess the difference in the þKTS/ÀKTS ratio between males and females. Finally, we did not conduct a sample size calculation because FS is a rare disease and we could only include all previous reported cases with sufficient clinical data. Therefore, the current results might have been affected by the small sample size.
In conclusion, we have confirmed that all variants, including the newly identified c.1432þ3 G>T variant, result in the same abnormal splicing pattern as revealed by in vitro and in vivo transcript analyses. In addition, a large-scale meta-analysis suggested that renal prognosis was not influenced by the location of intron 9 variants, sex chromosomes karyotypes, or DSD. We believe that this study broadens our understanding of the genetic and clinical characteristics of FS.

DISCLOSURE
All the authors declared no competing interests.

AUTHOR CONTRIBUTIONS
YT, TY, and KI designed the study and wrote the manuscript. K Nozu and TH interpreted the data and wrote the manuscript. CN, SI, YA, RR, EO, and NS collected and interpreted the data. ET, KT, TO, and TS collected urine and kidney samples for in vivo mRNA analysis. YS, K Nakanishi, HN, and MM critically reviewed the manuscript. All authors read and approved the final version of the manuscript.

SUPPLEMENTARY MATERIAL
Supplementary File (PDF) Figure S1. Comparison of the results of each patient in targeted RNA sequencing. Table S1. Comparison of renal prognosis in patients with Frasier syndrome between different patterns of sex chromosomes and phenotypes of external genitalia. Table S2. Prevalence of gonadal tumor in patients with Frasier syndrome depending on sex chromosomes karyotypes and phenotype of external genitalia.