Clinica Chimica Acta 481 (2018) 132–138
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Mutation spectrum of hyperphenylalaninemia candidate genes and the genotype-phenotype correlation in the Chinese population
T
Ruifang Wanga,1, Nan Shenb,1, Jun Yea,1, Lianshu Hana, Wenjuan Qiua, Huiwen Zhanga, Lili Lianga, Yu Suna, Yanjie Fana, Lili Wanga, Yu Wanga, Zhuwen Gonga, Huili Liua, ⁎⁎ ⁎ ⁎ Jianguo Wanga, Hui Yana, Nenad Blauc, ,2, Xuefan Gua, ,2, Yongguo Yua, ,2 a
Department of Pediatric Endocrinology and Genetic Metabolism, Shanghai Institute for Pediatric Research, Xinhua Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200092, China b Department of Rehabilitation Medicine, Xinhua Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200092, China c Dietmar-Hopp Metabolic Center, University Children's Hospital, 69120 Heidelberg, Germany
A R T I C LE I N FO
A B S T R A C T
Keywords: Hyperphenylalaninemia Molecular diagnosis Mutation spectrum Genotype-phenotype correlation Allelic phenotype value Genotypic phenotype value
Background: Hyperphenylalaninemia (HPA) is an inherited metabolic disorder that is caused by a deficiency of phenylalanine hydroxylase (PAH) or tetrahydrobiopterin. The prevalence of HPA varies widely around the world. Methods: A spectrum of HPA candidate genes in 1020 Chinese HPA patients was reported. Sanger sequencing, next generation sequencing (NGS), multiplex ligation-dependent probe amplification (MLPA) and quantitative real-time PCR (qRT-PCR) were applied to precisely molecular diagnose HPA patients. The allelic phenotype values (APV) and genotypic phenotype values (GPV) were calculated in PAH-deficient patients based on a recently developed formula. Results: Apart from genetic diagnoses confirmed in 915 HPA patients (89.7%) by Sanger sequencing, pathogenic variants were discovered in another 57 patients (5.6%) through deep detections (NGS, MLPA and qRT-PCR). We identified 196, 42, 10 and 2 variants in PAH, PTS, QDPR and GCH1, respectively. And a total of 47 novel variants were found in these genes. Through the APV and GPV calculations, it was found that the new GPV system was well correlated with metabolic phenotypes in most PAH-deficient patients. Conclusions: More HPA candidate variants were identified using new molecular diagnostic methods. The new APV and GPV system is likely to be highly beneficial for predicting clinical phenotypes for PAH-deficient patients.
1. Introduction
hydroxylase (PAHD, MIM # 261600), which catalyzes the conversion of (Phe) to L-tyrosine (Tyr). The cofactor tetrahydrobiopterin (BH4) and molecular oxygen are required for this enzymatic process [5,6]. Guanosine triphosphate cyclohydrolase I (GTPCH, MIM# 233910), 6-pyruvoyl-tetrahydropterin synthase (PTPS, MIM# 261640), sepiapterin reductase (SR, MIM# 612716), pterin-4alpha-carbinolamine dehydratase (PCD, MIM# 264070) and dihydropteridine reductase (DHPR, MIM# 261630) are essential for the L-phenylalanine
Hyperphenylalaninemia (HPA) is the most common inborn error of amino acid metabolism all over the world. The occurrence of HPA varies widely around the world. The average incidence in Europe is about 1:8000 [1], and in the USA is 1:15,000 [2]. While in mainland China the reported prevalence of this disease is approximately 1:12,000 [3,4]. HPA is mainly caused by a deficiency of hepatic phenylalanine
Abbreviations: HPA, hyperphenylalaninemia; PAH, phenylalanine hydroxylase; BH4, tetrahydrobiopterin; NGS, next generation sequencing; MLPA, multiplex ligation-dependent probe amplification; qRT-PCR, quantitative real-time PCR; APV, allelic phenotype values; GPV, genotypic phenotype values; PAHD, deficiency of hepatic phenylalanine hydroxylase; Phe, Lphenylalanine; Tyr, L-tyrosine; GTPCH, guanosine triphosphate cyclohydrolase I; PTPS, 6-pyruvoyl-tetrahydropterin synthase; SR, sepiapterin reductase; PCD, pterin-4-alpha-carbinolamine dehydratase; DHPR, dihydropteridine reductase; MHP, mild HPA; MPKU, mild phenylketonuria; CPKU, classic phenylketonuria; AV, assigned values; DBS, dried blood spot; ExAC, Exome Aggregation Consortium ⁎ Correspondence to: X. Gu and Y. Yu, Room 801 Sci&EduBldg, 1665 Kongjiang Rd, Yangpu District, Shanghai 200092, China. ⁎⁎ Correspondence to: N. Blau, Dietmar-Hopp Metabolic Center, University Children's Hospital, Im Neuenheimer Feld 669, D-69120 Heidelberg, Germany. E-mail addresses:
[email protected] (N. Blau),
[email protected] (X. Gu),
[email protected] (Y. Yu). 1 The first three authors contributed equally to this work, and should be considered as co-first author. 2 The last three authors contributed equally to this work. https://doi.org/10.1016/j.cca.2018.02.035 Received 16 November 2017; Received in revised form 10 February 2018; Accepted 27 February 2018 Available online 28 February 2018 0009-8981/ © 2018 Published by Elsevier B.V.
Clinica Chimica Acta 481 (2018) 132–138
R. Wang et al.
2.2. Genome sequencing
biosynthesis and regeneration of BH4. Classical forms of BH4 deficiency, caused by mutations in the genes encoding GTPCH, PTPS, PCD and DHPR, can lead to HPA [7,8]. However, in contrast to these forms of BH4 deficiency, SR deficiency may present a phenotype without HPA [9]. All infants should undergo neonatal screening programs in the first few days after birth to detect HPA. Blood Phe concentrations below 120 μmol/L denote normal. If the Phe levels increase to 120–600 μmol/ L, patients are diagnosed with mild HPA (MHP). Those patients with blood Phe concentrations of 600–1200 μmol/L are considered mild phenylketonuria (MPKU), and Phe concentrations above 1200 μmol/L indicate classic phenylketonuria (CPKU) [5,10]. Currently, a combination of different sequencing methods, such as conventional Sanger sequencing and next generation sequencing (NGS) technology, allow physicians to obtain patients' exact genotypes in a short time [11–13]. The genotypes related to HPA, especially PAH genotyping, can be useful for the establishment of a genotype-phenotype correlation and the application of precise therapy for these patients [14]. A correlation between PAH genotypes and metabolic phenotypes has been reported for some PAH mutations [15–17]. Thus, many previous studies have attempted to develop a system that can predict metabolic phenotypes from particular alleles [18,19]. The ability to predict phenotypes contributes to the better characterization of PKU patients. The Guldberg's arbitrary assigned values (AV) system is based on scoring of gene variants and is widely used for the phenotype prediction of PAH variants. It classifies HPA as different metabolic phenotypes using four levels (AV1 = CPKU, AV2 = moderate PKU, AV4 = MPKU and AV8 = MHP) [18]. However, in the Guldberg et al.'s study, researchers enrolled 686 HPA patients from seven European centers. The different methods and criteria used for diagnosis and classification of HPA in different medical centers may have caused to the inconsistencies in the genotypes and phenotypes of some variants. In the present study, a spectrum of gene mutations related to HPA was reported from a comprehensive survey of 1020 HPA patients in mainland China. Sanger sequencing was firstly performed to find causative gene variants in these patients. Then, NGS technology, multiplex ligation-dependent probe amplification (MLPA) and quantitative realtime PCR (qRT-PCR) were used to make further precise molecular diagnoses. Among all 1020 HPA patients, 808 had a PAH deficiency that caused HPA. For patients with PAH variants, we applied a recently developed linear allelic phenotype values (APV) and genotypic phenotype values (GPV) system to predict PAH genotype-phenotype relationship in Chinese PKU population.
Sanger sequencing, NGS, MLPA and qRT-PCR were performed to detect PAH (NM_000277.1), PTS (NM_000317.2), QDPR (NM_000320.2), GCH1 (NM_000161.2) and PCBD1 (NM_000281.3) gene, which encode PAH, PTPS, DHPR, GTPCH and PCD, respectively. These methods for detecting mutations are summarized below briefly. For Sanger sequencing, genomic DNA was extracted from the peripheral blood mononuclear cells of patients by a density centrifugation (Ficoll-Paque) method. The primers used for Sanger sequencing were designed for the PCR amplification of all exons and exon-intron boundaries of the five genes mentioned above using Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, CA, US). The PCR products were purified and sequenced bidirectionally using an ABI 3730xl DNA analyzer (Applied Biosystems, Foster City, CA, USA). NGS was used for patients whose genetic diagnosis of HPA could not be confirmed by Sanger sequencing. A Ion AmpliSeq™ Custom Primer Panel that contained 52 amplicons was designed and ordered. This custom panel can targeted all of the exons (padding: ± 10 bp) of PAH, PTS, GCH1, PCBD1 and QDPR gene, with the exception of 66 bp in exon 1 of GCH1 gene. To construct the libraries, 10–15 ng of each DNA sample was used with an Ion AmpliSeq™ Library Kit v2.0 (Cat. 4,475,345, Life Technologies, CA, USA). Gene mutations were sequenced by the Ion Personal Genome Machine® System (Life Technologies, CA, USA). After sequencing all samples, the data were analyzed on the Ion Torrent server using Torrent Suite v4.2.1 (Life Technologies, CA, USA). First, each sample read was classified according to the special barcode, signal processing, basic calling, read filter and sequence alignment. Second, the generated VCF files for each sample were annotated with the Ion reporter software (https://ionreporter.thermofisher.com). Finally, the annotated results were further analyzed by comparing them with dbSNP (https://www.ncbi.nlm.nih.gov/projects/SNP), 1000 Genomes (http://www.internationalgenome.org/), Exome Aggregation Consortium (ExAC, http://exac.broadinstitute.org/), PAHvdb (http:// www.biopku.org/home/pah.asp) and the PNDdb database (http:// www.biopku.org/home/pnddb.asp). The SALSA MLPA kit P055 PAH (Cat. P055-100R, MRC Holland, Amsterdam, Netherlands) was used to detect large deletions or duplications in patients with only one variant site of PAH or in those with no mutations found. The assay operation was performed according to the manufacturer's protocol. After DNA denaturation, ligation of the probes and multiplex PCR amplification, the PCR products were detected by the ABI 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA, US) using capillary electrophoresis. The raw data were analyzed by Coffalyser data analysis software (MRC Holland, Amsterdam, Netherlands). DNA samples from four healthy individuals were used as normal control samples in each run. Possible deletions in patients with only one mutation site of PTS gene were detected with qRT-PCR. Seven pairs of primers were designed for qRT-PCR using Primer 3.0 software (http://biotools. umassmed.edu/bioapps/primer3_www.cgi) (Supplementary Table 1). Detecting operation was performed on the ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, US).
2. Materials and methods 2.1. Patients and phenotypes A total of 1020 nonconsanguineous Chinese HPA patients (554 males and 466 females) in the pediatric endocrinology clinic of Shanghai Xinhua Hospital were investigated from 2011 to 2015. These patients were identified through a neonatal screening program. The plasma Phe concentrations were measured by tandem mass spectrometry from dried blood spot (DBS) samples before starting treatment. All patients with the maximum pretreatment blood Phe concentrations above 120 μmol/L were enrolled. In addition, urinary pterin analysis using high performance liquid chromatography and a DHPR activity assay [20] on DBS samples were also carried out to distinguish between PAH deficiency and tetrahydrobiopterin deficiency. Informed consents were obtained from all participants or their parents in prior to the commencement of the study. And this research was prospectively reviewed and approved by Ethics Committee of Xinhua Hospital affiliated to Shanghai Jiaotong University School of Medicine (Approval Number XHEC-D-2017-055).
2.3. Pathologic analysis of novel mutations of PAH and PTS genes For novel variants found in our patients, the evolutionary conservation was estimated with Clustal X and the effects of amino acid substitutions were evaluated using three online tools: SIFT (http://sift. jcvi.org), PolyPhen (http://genetics.bwh.harvard.edu/pph/) and Mutation Taster (http://www.mutationtaster.org/).
133
Clinica Chimica Acta 481 (2018) 132–138
R. Wang et al.
splicing defect. The most frequent PAH variant was p.R243Q (17%), followed by p.Y204_T236delinsS (7.4%), p.R241C (7.2%), p.W356Efs*22(5.4%), p.R111* (4.4%), p.Y356* (3.9%), p.R413P (3.8%) and c.442-1G > A (3.7%). These point variants were scattered across the PAH coding sequence except exon 13 (85.9% of all alleles), while only 8.1% distributed in intron regions. Additionally, the distribution of common PAH variants and variant types were quite different in mainland China, compared to Spain [21] and the USA [22] (Fig. 1). In the 204 PTPS deficiency patients, 42 point variants and 2 large gene deletions were detected (Supplementary Table 6). These point variants included 30 missense variants, 7 splice-site variants, 2 nonsense variants, 1 insertion, 1 duplication and 1 deletion. The most common PTS variant was p.P87S (38.2%), followed by c.84-291A > G (11%), p.N52S (8.6%), p.D96N (8.6%) and p.V56 M (5.1%). These variants were distributed equally across all six exons of PTS, while only 15.7% of the PTS mutations were located in the intron region. Moreover, of the 7 DHPR deficiency patients, 4 patients were confirmed to carry homozygous QDPR mutations (p.[V86D];[V86D], p.[R221*]; [R221*], p.[A48V];[A48V] and p.[P172L];[P172L]). In addition, only one HPA patient was identified as a carrier of a compound heterozygous mutation in GCH1 (p.[E56K];[V202I]). No mutations were identified in PCBD1 in this study.
2.4. Calculation of the allelic phenotype values (APV) and genotypic phenotype values (GPV) for PAH variants APV and GPV calculations are based on the formula developed by Children's hospital of Heidelberg university. Briefly, all nonsense, splice, deletion (including large deletion), insertion, indel and missense PAH variants with less than 5% PAH residual in vitro activity (as recorded in the BIOPKU database) were defined as PAH 0-alleles. Functional hemizygotes reported in our patients (with these PAH 0-alleles) were selected for subsequent investigation. All other PAH alleles (non-0-alleles) in the functional hemizygote state were assigned to corresponding metabolic phenotypes and the APV were calculated. The formula is based on the percentage of different phenotypes in the same PAH variant. APV = (%CPKU × 10 + %MPKU × 5 + %MHP × 1) / 100. The GPV were calculated as a mean of APV from the two alleles constituting the genotype. Calculated APV and GPV values are specific for the Chinese PKU population described in this study and may differ from values in the BIOPKU database. 2.5. Statistical analysis A statistical analysis was performed using SPSS Statistics 16.0 software (SPSS Inc., Chicago, USA). The metabolic phenotypes of different genotypes were represented by the concentrations of blood Phe. Correlations between GPV and the average blood Phe levels were described by the fitted curve in order to verify the consistency of GPV and PAH phenotypes.
3.2. Novel variants in HPA candidate genes Twenty-five novel variants in PAH were detected in the present study including 15 missense variants, 3 nonsense variants, 3 splice-site variants, 2 indels, 1 deletion, and 1 duplication (Supplementary Table 7). Twenty-two of these new variants were identified in coding regions, while only three were located in introns. Through in silico analysis of these 15 missense variants, 9 variants were predicted to affect the protein functions and 2 variants were predicted to be benign, and the prediction results of the other 4 variants from different websites were inconsistent. A total of 17 different new PTS variants were discovered, comprising of 11 missense variants, 3 splice-site variants, 2 insertions and 1 nonsense variants (Supplementary Table 8). Five of the missense variants were predicted to destroy protein function, whereas the predictions of the effects of other missense variants on protein function were inconsistent. In addition to three novel PTS variants located in the intron, the other 14 novel variants were identified in the coding regions. Furthermore, four novel QDPR variants and one novel GCH1 variants were also found (Supplementary Table 9). All of the other novel QDPR variants were located in the exon and affected the protein function; however, the prediction the effect of the novel GCH1 variant on protein function was inconsistent on different websites. The conservation degree of the amino acid residues, the classifications of all these novel variants using ACMG/AMP variant interpretation guideline [23] and the predicted effects of these variants on protein function were also shown in Supplementary Tables 7, 8, 9.
3. Results 3.1. Mutation spectrums of HPA candidate genes A total of 1020 HPA patients were included in the study. Based on their clinical symptoms and biochemical tests, 808 patients of them were clinically diagnosed with a PAH deficiency, 204 patients with a PTPS deficiency, 7 patients with a DHPR deficiency and one patient with a GTPCH1 deficiency. The Sanger sequencing results confirmed that the genetic diagnoses in 915 HPA patients (89.7%) were consistent with their clinical diagnoses. However, there were still 105 patients in whom biallelic variants were not found by Sanger sequencing. After deep detection by NGS, MLPA and qRT-PCR, the corresponding pathogenic variants were finally discovered in 57 (5.6%) of these patients (Table 1). Additional biochemical information and genetic testing results are shown in Supplementary Tables 2, 3, 4 as well as Supplementary Fig. 1. As described previously, variants in the PAH gene were found in 808 HPA patients. The spectrum of PAH variants was composed of 196 different point variants (Supplementary Table 5) and included 131 missense variants, 28 splice-site variants, 11 deletions, 20 nonsense variants, 3 duplications, 2 indels and 1 missense variant causing a
Table 1 Summary of diagnosed conditions through biochemical tests and sequential molecular detection methods in 1020 HPA patients. Group/number (frequency)
PAH deficiency BH4 deficiency Total
PAH-D PTPS-D DHPR-D GTPCH-D
Clinical diagnosis
808 (79.2%) 204 (20%) 7 (0.7%) 1 (0.1%) 1,020
Sanger sequencing diagnosis
708 (69.4%) 199 (19.5%) 7 (0.7%) 1 (0.1%) 915 (89.7%)
Inconsistent between biochemical and Sanger sequencing diagnosis
100 (9.8%) 5 (0.5%) 0 0 105 (10.3%)
Further diagnosis NGS
MLPA
qRT-PCR
7(0.7%) 0 0 0 7(0.7%)
48(4.7%) 0 0 0 48(4.7%)
0 2(0.2%) 0 0 2(0.2%)
Undiagnosis
Final genetic diagnosis
45(4.4%) 3(0.3%) 0 0 48(4.7%)
763(74.8%) 201(19.7%) 7(0.7%) 1(0.1%) 972(95.3%)
PAH: Phenylalanine hydroxylase; BH4: tetrahydrobiopterin; PAH-D: Phenylalanine hydroxylase deficiency; PTPS-D: 6-pyruvoyl-tetrahydropterin synthase deficiency; DHPR-D: Dihydropteridine reductase deficiency; GTPCH-D: GTP cyclohydrolase I deficiency; NGS: Next generation sequencing; MLPA: Multiplex ligation-dependent probe amplification; qRT-PCR: Quantitative real-time polymerase chain reaction. cDNA numbering of PAH, PTS, QDPR, GCH1 and PCBD1 according to reference sequence NM_000277.1, NM_000317.2, NM_000320.2, NM_000161.2 and NM_000281.3.
134
Clinica Chimica Acta 481 (2018) 132–138
R. Wang et al.
Fig. 1. The comparison of mutation types of PAH gene in different populations. Panel (A) The proportion of the majority mutations of PAH gene in Chinese, Spanish [21] and American [22] populations. (A-1) Chinese population; (A-2) Spanish population; (A-3) American populations. Panel (B) The proportion of different mutation types of PAH gene in Chinese, Spanish [21] and American [22] populations. (B-1) Chinese population; (B-2) Spanish population; (B-3) American populations. The number of species of each type of mutation was marked in the legend.
Phe restriction and BH4 supplementation [5,10]. However, there were many difficulties limiting the early precise diagnosis of HPA. With the breakthrough of sequencing methods, early molecular genetic diagnosis of HPA became much easier. In this study, among a cohort of 1020 HPA patients, HPA candidate gene variants were detected in 972 patients, yielding a mutation detection rate of 95.3%. A total of 196 different PAH variants were identified, of which 25 were novel variants. In addition, we also found 42 PTS variants, 10 QDPR variants and 2 GCH1 variants among all HPA patients. It is generally known that distributions of common PAH mutations are different among different regions. As mentioned in the result, the most frequent PAH variant in the Chinese population was p.R243Q (17%), which is consistent with the other reports in the literature [24]. However, the most common PAH variants are p.Q355_Y356insGLQ and p.R408W in Spain [21] and USA [22], respectively. The occurrence of p.R408W in particular is common in Europe [25] and USA [22]; however, it is a rare variant in Chinese population. Thus, it is necessary to establish the mutations spectrum of HPA candidate genes in China and to investigate genotype-phenotype correlations for the common gene mutations in Chinese HPA patients. The ability to predict HPA phenotypes accurately based on the genotype has been the goal of many investigations [21,26]. Although many factors, such as the BH4-response, hepatic uptake of dietary Phe and large neutral amino acids (LNAAs) transporters in the blood brain barrier, can affect phenotypic variation in HPA, the HPA candidate genotype especially the PAH genotype is the dominant factor determining HPA phenotype [19,27,28]. One approach based on the Guldberg's AV of PAH variants is used widely to predict metabolic phenotypes in HPA patients and to investigate the PAH genotype-phenotype correlations [18]. The AV-based approach for the prediction of phenotypes in PKU patients has been reported in many other studies [19,27]. However, the different methods and criteria used for diagnosis and classification of HPA in different medical centers have limited the
3.3. Genotype-phenotype correlations in PAH-deficient patients Of the 196 different PAH variants identified in this study, 65 variants (nonsense mutations, deletion/frameshift and splice-site variants) were classified as a PAH 0-allele with less than 5% PAH residual in vitro activity (see Table 1 in Data-in-Brief). The phenotype characteristics of 250 patients, who carry a PAH 0-allele on one of their chromosomes can be considered to be functionally hemizygous. The 65 different variants, in combination with PAH 0-alleles, found in these 250 patients were assigned to one of three categories in which their phenotypes appeared most frequently (Table 2). For 13 variants, the assignment was definite because the phenotype classifications were consistent in two or more functionally hemizygous patients. Thirty-four variants were identified in only one functionally hemizygous patient; thus, their assignments were ambiguous. The other 18 variants were found in two or three types of functionally hemizygous patients, but each was associated with more than one phenotype category (Fig. 2). The APV were calculated for 65 PAH variants and the new scales ranged from 1 to 10 (see Table 2 in Data-in-Brief). In the new APV system, APV = 10 was associated with CPKU and APV = 1 with MHP. A fitted curve was made based on the relationship between APV and the average Phe levels from the above 250 functionally hemizygous patients (using the curve-fitting formula: y = 274.795e0.195x) (Fig. 3). The corresponding GPV of 94 PAH compound heterozygous genotypes in 222 PAH-deficient patients were also calculated in Table 3 in Data-inBrief. It was found that the new GPV system was well correlated with metabolic phenotypes in most PAH-deficient patients (Fig. 4). 4. Discussion As the most common inherited disease of amino acid metabolism disease, HPA can be improved by early interventions, such as dietary 135
Clinica Chimica Acta 481 (2018) 132–138
R. Wang et al.
Table 2 Assignment of 65 mutations of PAH gene to metabolic phenotypes. Characteristics of 250 functionally hemizygous patients (number of patients) CPKU
MPKU
p.R413P(5)c p.R243Q(16)c p.L242F(1)b p.A156P(2)a p.E286K(2)c p.A104V(1)b p.R400K(1)c p.R252Q(3)c p.I65T(1)b p.G239D(1)b p.A434D(1)c p.Y417C(1)b p.R261Q(3)c p.M276R(1)c p.I324N(1)b p.S310F(1)c p.P225A(1)b p.L255V(1)b p.R408W(1)c p.S359L(1)b p.F161S(1)b p.Y204_T236delinsS(34)c
p.R413P(3)c p.S349A(2)a p.R243Q(31)c p.I164V(1)b p.G257V(3)c p.R158W(3)a p.R241C(13)c p.R53H(2)c p.G130D(1)b p.Y206C(2)a p.F331S(3)a p.R408Q(4)c p.E286K(1)c P.E280K(1)c p.R400K(1)c p.R252Q(1)c p.P281L(3)a p.G247V(2)a p.R261Q(2)c p.M276R(1)c p.P147L(1)b p.R400T(1)c
MHP p.S391T(1)c p.S310F(1)c p.S303P(1)b p.V388 M(2)a p.H285Y(1)b p.R408W(2)c p.G312V(1)b p.V291L(1)b p.Y204_T236delinsS(2)c p.R158Q(1)b
p.T418P(2)a p.R243Q(14)c p.V230I(2)a p.G257V(1)c p.R241C(16)c p.H107R(3)a p.P314T(2)a p.R53H(9)c p.L255S(1)b p.F392I(3)a p.R408Q(2)c p.D415N(1)b P.E280K(1)c p.T380M(1)b p.Q375H(1)b p.Q419R(1)b p.V291M(1)b p.A434D(1)c p.A395S(1)b p.D75H(1)b p.R400T(1)c p.S391T(2)c
p.Y387D(1)b p.P314A(1)b p.R169C(1)b p.D338Y(1)b p.M276T(1)b p.C203G(1)b p.A403V(1)b p.L308F(1)b
CPKU: classic PKU; MPKU: mild phenylketonuria; MHP: mild hyperphenylalaninemia. a Definite mutations: phenotype categorizations were consistent in all functionally hemizygous patients (more than one). b Mutations presented in only one functionally hemizygous patients. c Mutations also presented in other phenotype categories, although at lower frequencies.
in the same PAH variant, and expanded the new scales from 1 to 10. This redefined APV could present and predict the severity of allelic phenotypes in more detail. Furthermore, we tested the APV in patients and verified that it conformed well to metabolic phenotypes in PAHdeficient patients. Thus, although the redefined APV was based on a limited number of patients, it still proved to be a useful tool for allelic phenotype prediction with wider scales. To date, only few studies have attempted to explore the prediction of genotypic phenotypes [27,29]. In the present study, we established GPV in the most common PAH compound heterozygous genotypes of PAH deficiency based on the redefined APV system. After verification,
reliability of the Guldberg's AV system. Herein, we attempted to improve this problem by accumulating data from the Shanghai Xinhua Hospital database recorded from 2011 to 2015, which contained information from 1020 patients, including detailed genotypes and metabolic phenotypes. Based on the molecular diagnosis of PAH mutations, we have assigned their metabolic phenotypes and corresponding APV to each of 65 PAH variants. Moreover, the Guldberg's AV system only classifies PKU/HPA into four categories. These classifications are broad, and it is difficult for them to represent the severity of HPA in detail. In this study, we modified the formula developed by Children's hospital of Heidelberg university, based on the percentage of different phenotypes
Fig. 2. Distribution of phenotypes for the 18 mutations of PAH gene that were found in two or three phenotype category in functionally hemizygous patients. Each column represented one type of the observed phenotype: : CPKU; :MPKU and □: MHP. n: number of patients.
136
Clinica Chimica Acta 481 (2018) 132–138
R. Wang et al.
Accession numbers PAH sequence variants c.257G > C (p.R86P), c.199T > G(p.S67A), c.425T > C(p.L142P), c.383T > C(p.L128P), c.641A > G(p.E214G), c.1183G > T(p.A395S), c.156G > T(p.L52F), c.1144T > A(p.F382I), c.932T > G(p.L311R), c.182A > G(p.N61S), c.607T > G(p.C203G), c.1125A > C(p.Q375H), c.1090C > T(p.L364F), c.1235T > A(p.V412D), c.940C > G(p.P314A), c.706+5G > A(IVS6+5G > A), c.1199+19T > C(IVS11+19T > C), c.1200-12del(IVS11-12del), c. 1057G > T(p.E353*), c.609C > A(p.C203*), c.852C > A(p.C284*), c.73_74dup(p.E26Lfs*13), c.59delinsCC(p.Q20Pfs*7), c.23_24delACinsT(p.N8Ifs*30) and c.541_544del(p.E181Kfs*13) reported in this paper have been submitted to the PAHvdb database under accession numbers: PAH 1019, 1020, 1021, 1022, 1023, 1024, 1025, 1026, 1027, 1028, 1030, 1031, 1032, 1033, 1034, 1035, 1036, 1037, 1038, 1039, 1040, 1041, 1042, 1043, 1044, respectively. PTS sequence variants c.34G > C(p.A12P), c.421G > T(p.V141F), c.181G > A(p.G61R), c.62G > A(p.S21N), c.374G > A(p.G125E), c.143G > A(p.G48D), c.193C > T(p.P65S), c.1A > G(p.M1V), c.35C > A(p.A12E), c.280G > C(p.D94H), c.313A > T(p.S105C), c.289_290insCTT(p.V97delinsAL), c.322G > T(p.E108*), c.432dup (p.E145Rfs*18), c.84-6insTTTT(IVS1-6insTTTT), c.84-252G > T(IVS1252G > T) and c.164-37C > T(IVS2-37C > T) reported in this paper have been submitted to the PNDdb database under accession numbers: PNDDB 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, respectively. GCH1 sequence variant c.166G > A (p.E56K) reported in this paper have been submitted to the PNDdb database under accession numbers: PNDDB 276. QDPR sequence variants c.143C > T (p.A48V), c.81dup (p.Q28Afs*14), c.328C > T (p.Q110*) and c.229G > T (p.E77*) reported in this paper have been submitted to the PNDdb database under accession numbers: PNDDB 282, 278, 279, 280, respectively.
Fig. 3. Genotype and phenotype correlation analysis. Curve-fitting: The relationship between the predicted score of each mutation got from 250 functionally hemizygous individuals and the average Phe levels. The X-axis indicated the predicted scores of 65 mutations (from 1 to 10). The Y-axis indicated the average Phe level of each predicted mutation group. From the fitting curve of the 65 mutations, we got the fitting formula: y = 274.795e0.195x(R square = 0.823).
Conflicts of interest
Fig. 4. Observed versus expected metabolic phenotypes in the different severity groups of patients with phenylketonuria (PKU). The X-axis indicated the three phenotypic categories: classic PKU (CPKU) (n = 44), mild PKU (MPKU) (n = 114) and mild HPA (MHP) (n = 64).All the 222 PAHD patients carried two mutations on either allele could get a score in our predicted system. n: number of patients. The Y-axis indicated the number of patients with each observed phenotype in each group. Each column represented one type of the observed phenotype: : CPKU; :MPKU and □:
The authors have no conflict of interest to declare. Author contributions XFG, YGY and N.B. designed and supervised the study. RFW, NS and JY performed majority of the data analysis, wrote the main manuscript and prepared the figures and tables. RFW, YS, YJF, LLW, YW, ZWG, HLL, JGW and HY performed the experiments. XFG, YGY, JY, LSH, WJQ, HWZ and LLL provided patients' data and recruited the patients. N.B. modified the article using scientific English. All authors reviewed the manuscript.
MHP. Predicted phenotype was determined according to the predicted Phe level got from the fitting curve based on the average score of the two mutated alleles of the genotype. The “concordance rate” of CPKU, MPKU and MHP group were 77.3%, 60.5% and 81.3%, respectively.
the new GPV system correlated with metabolic phenotype in PKU patients as well. However, there were inconsistencies between GPV and metabolic phenotypes in a few of cases. We speculated that the large mutational heterogeneity and the limited number of patients caused this inconsistency. The GPV system will still be a good predictor of phenotype prediction and evaluation in patients with a PAH deficiency, even though it requires further improvement. In conclusion, more HPA candidate variants were identified recently using new molecular diagnostic methods, and a more comprehensive variants spectrum of HPA was presented. We believe that our study can provide some of the HPA variant characteristics in the Chinese population. In addition, as powerful new predictors, the redefined APV and GPV systems that were established and tested in this study will benefit the diagnosis and treatment of HPA in clinical settings.
Acknowledgements The authors thank all the patients and their families. We express our gratitude to all of the pediatricians who helped with the study. This work was supported by the “National Natural Science Foundation of China” (No. 81670812, to YYG), the “Innovation Fund for Translational Medicine Plan” from the Shanghai Jiao Tong University School of Medicine (No. 15ZH3003, to YYG), the “Jiaotong University Cross Biomedical Engineering” (No.·YG2017MS72, to YGY), the “Shanghai Municipal Commission of Health and Family Planning” (No. 201740192, to YGY), the “Shanghai Shen Kang Hospital Development Center new frontier technology joint project” (No.·SHDC12017109, to YGY), the National Key Technology R&D Program (2012BAI09B04 to Professor Xuefan Gu), and the Special Basic
137
Clinica Chimica Acta 481 (2018) 132–138
R. Wang et al.
Work of Science and Technology (2014FY110700 to Professor Xuefan Gu).
[16] T. Zhu, J. Ye, L. Han, W. Qiu, H. Zhang, L. Liang, X. Gu, Variations in genotypephenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene 529 (2013) 80–87. [17] A. Daniele, I. Scala, G. Cardillo, C. Pennino, C. Ungaro, M. Sibilio, G. Parenti, L. Esposito, A. Zagari, G. Andria, F. Salvatore, Functional and structural characterization of novel mutations and genotype-phenotype correlation in 51 phenylalanine hydroxylase deficient families from Southern Italy, FEBS J. 276 (2009) 2048–2059. [18] P. Guldberg, F. Rey, J. Zschocke, V. Romano, B. Francois, L. Michiels, K. Ullrich, G.F. Hoffmann, P. Burgard, H. Schmidt, C. Meli, E. Riva, I. Dianzani, A. Ponzone, J. Rey, F. Guttler, A European multicenter study of phenylalanine hydroxylase deficiency: classification of 105 mutations and a general system for genotype-based prediction of metabolic phenotype, Am. J. Hum. Genet. 63 (1998) 71–79. [19] S. Wettstein, J. Underhaug, B. Perez, B.D. Marsden, W.W. Yue, A. Martinez, N. Blau, Linking genotypes database with locus-specific database and genotype-phenotype correlation in phenylketonuria, Eur. J. Hum. Genet. 23 (2015) 302–309. [20] J. Ye, W.J. Qiu, J.D. Zhou, L.S. Han, X.F. Gu, Development of a method for the determination of dihydropteridine reductase and its application, Jian Yan Yi Xue. 21 (2006) 48–51. [21] L. Aldamiz-Echevarria, M. Llarena, M.A. Bueno, J. Dalmau, I. Vitoria, A. FernandezMarmiesse, F. Andrade, J. Blasco, C. Alcalde, D. Gil, M.C. Garcia, D. GonzalezLamuno, M. Ruiz, M.A. Ruiz, L. Pena-Quintana, D. Gonzalez, F. Sanchez-Valverde, L.R. Desviat, B. Perez, M.L. Couce, Molecular epidemiology, genotype-phenotype correlation and BH4 responsiveness in Spanish patients with phenylketonuria, J. Hum. Genet. 61 (2016) 731–744. [22] P. Guldberg, H.L. Levy, W.B. Hanley, R. Koch, R. Matalon, B.M. Rouse, F. Trefz, F. de la Cruz, K.F. Henriksen, F. Guttler, Phenylalanine hydroxylase gene mutations in the United States: report from the Maternal PKU Collaborative Study, Am. J. Hum. Genet. 59 (1996) 84–94. [23] S. Richards, N. Aziz, S. Bale, D. Bick, S. Das, J. Gastier-Foster, W.W. Grody, M. Hegde, E. Lyon, E. Spector, K. Voelkerding, H.L. Rehm, Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology, Genet Med. 17 (2015) 405–424. [24] N. Li, H. Jia, Z. Liu, J. Tao, S. Chen, X. Li, Y. Deng, X. Jin, J. Song, L. Zhang, Y. Liang, W. Wang, J. Zhu, Molecular characterisation of phenylketonuria in a Chinese mainland population using next-generation sequencing, Sci. Rep. 5 (2015) 15769. [25] C. Aulehla-Scholz, H. Heilbronner, Mutational spectrum in German patients with phenylalanine hydroxylase deficiency, Hum. Mutat. 21 (2003) 399–400. [26] E. Jeannesson-Thivisol, F. Feillet, C. Chery, P. Perrin, S.F. Battaglia-Hsu, B. Herbeth, A. Cano, M. Barth, A. Fouilhoux, K. Mention, F. Labarthe, J.B. Arnoux, F. Maillot, C. Lenaerts, C. Dumesnil, K. Wagner, D. Terral, P. Broue, L. de Parscau, C. Gay, A. Kuster, A. Bedu, G. Besson, D. Lamireau, S. Odent, A. Masurel, J.L. Gueant, F. Namour, Genotype-phenotype associations in French patients with phenylketonuria and importance of genotype for full assessment of tetrahydrobiopterin responsiveness, Orphanet J Rare Dis. 10 (2015) 158. [27] L.L. Santos, C.G. Fonseca, A.L. Starling, J.N. Januario, M.J. Aguiar, M.G. Peixoto, M.R. Carvalho, Variations in genotype-phenotype correlations in phenylketonuria patients, Genet. Mol. Res. 9 (2010) 1–8. [28] N. Shen, C. Heintz, C. Thiel, J.G. Okun, G.F. Hoffmann, N. Blau, Co-expression of phenylalanine hydroxylase variants and effects of interallelic complementation on in vitro enzyme activity and genotype-phenotype correlation, Mol. Genet. Metab. 117 (2016) 328–335. [29] A. Acosta, W. Silva Jr., T. Carvalho, M. Gomes, M. Zago, Mutations of the phenylalanine hydroxylase (PAH) gene in Brazilian patients with phenylketonuria, Hum. Mutat. 17 (2001) 122–130.
Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.cca.2018.02.035. References [1] J.G. Loeber, Neonatal screening in Europe; the situation in 2004, J. Inherit. Metab. Dis. 30 (2007) 430–438. [2] National Institutes of Health Consensus Development Conference Statement: phenylketonuria: screening and management, October 16–18, 2000, Pediatrics, 108 2001, pp. 972–982. [3] X. Gu, Z. Wang, J. Ye, L. Han, W. Qiu, Newborn screening in China: phenylketonuria, congenital hypothyroidism and expanded screening, Ann. Acad. Med. Singap. 37 (2008) 107–110. [4] X.T. Shi, J. Cai, Y.Y. Wang, W.J. Tu, W.P. Wang, L.M. Gong, D.W. Wang, Y.T. Ye, S.G. Fang, P.W. Jing, Newborn screening for inborn errors of metabolism in mainland China: 30 years of experience, JIMD Rep. 6 (2012) 79–83. [5] N. Blau, F.J. van Spronsen, H.L. Levy, Phenylketonuria, Lancet 376 (2010) 1417–1427. [6] J.J. Mitchell, Y.J. Trakadis, C.R. Scriver, Phenylalanine hydroxylase deficiency, Genet Med. 13 (2011) 697–707. [7] E.R. Werner, N. Blau, B. Thony, Tetrahydrobiopterin: biochemistry and pathophysiology, Biochem. J. 438 (2011) 397–414. [8] N. Blau, J.B. Hennermann, U. Langenbeck, U. Lichter-Konecki, Diagnosis, classification, and genetics of phenylketonuria and tetrahydrobiopterin (BH4) deficiencies, Mol. Genet. Metab. 104 (Suppl) (2011) S2–9. [9] N. Blau, L. Bonafe, B. Thony, Tetrahydrobiopterin deficiencies without hyperphenylalaninemia: diagnosis and genetics of dopa-responsive dystonia and sepiapterin reductase deficiency, Mol. Genet. Metab. 74 (2001) 172–185. [10] J. Vockley, H.C. Andersson, K.M. Antshel, N.E. Braverman, B.K. Burton, D.M. Frazier, J. Mitchell, W.E. Smith, B.H. Thompson, S.A. Berry, Phenylalanine hydroxylase deficiency: diagnosis and management guideline, Genet Med. 16 (2014) 188–200. [11] D. Trujillano, B. Perez, J. Gonzalez, C. Tornador, R. Navarrete, G. Escaramis, S. Ossowski, L. Armengol, V. Cornejo, L.R. Desviat, M. Ugarte, X. Estivill, Accurate molecular diagnosis of phenylketonuria and tetrahydrobiopterin-deficient hyperphenylalaninemias using high-throughput targeted sequencing, Eur. J. Hum. Genet. 22 (2014) 528–534. [12] Y.Y. Cao, Y.J. Qu, F. Song, T. Zhang, J.L. Bai, Y.W. Jin, H. Wang, Fast clinical molecular diagnosis of hyperphenylalaninemia using next-generation sequencingbased on a custom AmpliSeq panel and Ion Torrent PGM sequencing, Mol. Genet. Metab. 113 (2014) 261–266. [13] Y. Gu, K. Lu, G. Yang, Z. Cen, L. Yu, L. Lin, J. Hao, Z. Yang, J. Peng, S. Cui, J. Huang, Mutation spectrum of six genes in Chinese phenylketonuria patients obtained through next-generation sequencing, PLoS One 9 (2014) e94100. [14] N. Blau, N. Shen, C. Carducci, Molecular genetics and diagnosis of phenylketonuria: state of the art, Expert. Rev. Mol. Diagn. 14 (2014) 655–671. [15] M.A. Bueno, D. Gonzalez-Lamuno, C. Delgado-Pecellin, L. Aldamiz-Echevarria, B. Perez, L.R. Desviat, M.L. Couce, Molecular epidemiology and genotype-phenotype correlation in phenylketonuria patients from South Spain, J. Hum. Genet. 58 (2013) 279–284.
138