Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population

GENE-38871; No. of pages: 8; 4C: Gene xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

Gene journal homepage: www.elsevier.com/locate/gene

Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population☆,☆☆ Tianwen Zhu a, Jun Ye b, Lianshu Han b, Wenjuan Qiu b, Huiwen Zhang b, Lili Liang b, Xuefan Gu b,⁎ a b

Department of Neonatal Medicine, Xin-Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, China Department of Endocrinology and Genetic Metabolism, Xin-Hua Hospital, Shanghai Institute of Pediatric Research Affiliated to Shanghai Jiao Tong University School of Medicine, China

a r t i c l e

i n f o

Article history: Accepted 22 July 2013 Available online xxxx Keywords: PAHD Chinese Han population Phenotype–genotype correlation Complex phenotypes

a b s t r a c t Background: The value of genotyping to predict variant phenotypes in patients with phenylalanine hydroxylase (Pah) deficiency is a matter of debate. However, there exists no comprehensive population relationship study focused on the Han Chinese. Methods: We analyzed genotype–phenotype correlation for 186 different genotypes in 338 unrelated Chinese patients harboring 109 different Pah mutations. Two systems were used in this process. The first was a phenotype prediction system based on arbitrary values (AV) attributed to each mutation. The second was a pair-wise correlation analysis. The observed phenotype for AV analysis was the corresponding metabolic phenotype stratified according to the pretreatment phenylalanine (Phe) value. Results: We found that the observed phenotype matched the predicted phenotype in 54.41% of 272 patients for whom AV information was available; the highest degree of concordance (61.83%) was found in patients with null/null genotypes, whereas the lowest “concordance rate” (32.69%) was observed for patients with expected mild-PKU phenotype. There are repeated inconsistencies for such mutations as R241C, R243Q, R261Q, V388M, V399V, R408Q, A434D and EX6-96ANG which are associated with variable phenotypes in patients with identical genotype. Significant correlations were disclosed between pretreatment Phe values and predicted residual activity (r = − 0.45643, P b 0.0001) or AV sum (r = − 0.59523, P b 0.0001). Conclusion: Our study supports the notion that the Pah mutation genotype is the main determinant of metabolic phenotype in most patients in a particular population, and provided novel insights into the values that underpin the subsequent treatment and the prognosis of PKU in Chinese. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Phenylalanine hydroxylase (Pah [MIM 261600]) deficiency (PAHD) is an autosomal recessive disorder that results in intolerance to the dietary intake of phenylalanine (Mitchell et al., 2011). It is the most common inborn error of amino acid metabolism in Chinese, with an incidence of about 1:11,572 (Zhan et al., 2009), which is similar Abbreviations: AV, arbitrary value; MHP, mild hyperphenylalaninemia; Pah, phenylalanine hydroxylase; PAHD, phenylalanine hydroxylase deficiency; Phe, phenylalanine; PKU, phenylketonuria; PRA, predicted residual PAH activity. ☆ Funding source: This work was supported by grants from the Major Program of Shanghai Committee of Science and Technology (11dz195030), the Shanghai City Health Bureau project (20124104), the National Natural Science Foundation of China (81070700, 81200654), the National Key Technology R&D Program(2012BAI09B04) and the Shanghai Jiao Tong University School of Medicine Fund (11XJ22002). ☆☆ Financial disclosure: All authors have no financial relationships relevant to this article to disclose. ⁎ Corresponding author at: Department of Endocrinology and Genetic Metabolism, XinHua Hospital, Shanghai Institute for Pediatric Research Affiliated to Shanghai Jiao Tong University School of Medicine, Kongjiang Road 1665#, Shanghai 200092, China. Tel.: +86 21 65011012; fax: +86 21 65791316. E-mail address: [email protected] (X. Gu).

to that in Caucasian populations (Zschocke, 2003). The key metabolic feature of this disease is elevated serum concentrations of phenylalanine, leading to the mental retardation with varying degrees if left untreated. So far, more than 500 different mutant alleles, which cause different levels of reduction in the catalytic activity of the enzyme, have been identified at the Pah locus (http://www.pahdb.mcgill.ca/) with wide variation of their frequency and genotypic distribution, generating a wide spectrum of phenotypes ranging in severity from classic phenylketonuria (PKU) to variant PKU or mild hyperphenylalaninemia (MHP). There has been a hope that delineation of genotypes would enable the prediction of variant phenotypes in the case of human genetic disease, which would have added value for prognosis and treatment. Nonetheless, reports on genotype–phenotype relationship in PAHD in some European (Daniele et al., 2007; Groselj et al., 2012; Mallolas et al., 1999) and Oriental populations (Chien et al., 2004; Okano et al., 2011; Qu et al., 2008) often showed no robust correlation due to the high allelic heterogeneity and broad phenotypic variability. Recently, a more formalized system by Guldberg et al. (1998) was developed for estimating genotype–phenotype correlations based on the prediction of the phenotypic impact of each mutation and was adopted in some studies. For some populations, such as Brazilian and

0378-1119/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gene.2013.07.079

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

2

T. Zhu et al. / Gene xxx (2013) xxx–xxx

Lithuanian, in which some particular allele frequencies were highly skewed (Acosta et al., 2001; Kasnauskiene et al., 2003), the system proved to be highly useful, indicating that the efficiency of the method might vary depending on the set of mutations in a specific population. In our previous study of the Chinese Han population (Zhu et al., 2010), 79 different mutations were identified in 212 unrelated patients with 8 mutations accounting for two-thirds of the identified ones, which facilitated investigation of their phenotypic effect. The present study was sought to examine genotype–phenotype correlation for 186 different genotypes in 338 unrelated Chinese patients (the genetic analysis for some of them had been made in our previous study). 2. Materials and methods

Table 1 Classification, by type, of 53 null mutations. Classification of 41 null mutations by type Mutation type

Mutations

Missense (n = 16)

R413Pa, E280Ka, L255Sa, R252Qa, R270Sa, T278Ia, R408Wb, R252Wb, A259Tb, G247Vb, R243Qb, A342Tb, R158Qb, A156Pc, L385Pc, R400Kc R111X, Y356X, Y166X, W326X, R176X, R261X, C375X, E228X, R243X, Q232X, W187X, S411X, Q172X, Q301X, Y414X S70del, 190-194delCACAT, F39del, 1024delG, 541-543delGAG, 540delGGAGG R241NPfs EX6-96ANG, V399V, IVS4-1GNA, IVS6-1GNA, VS7+2TNA, IVS4+3GNC, IVS5+1GNA, IVS4+2TNA, IVS10-14CNG, IVS2+5GNC, IVS7+1GNA, IVS7+5GNA, IVS12+6TNA, IVS12+4ANG, IVS5-2ANG

Protein truncation (n = 15) Deletion (n = 6) Frameshift (n = 1) Splice defective (n = 15)

2.1. Patients and phenotypic classification DNA samples were collected in the Department of Pediatric Endocrinology and Genetic Metabolism, Xin-Hua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. This division treats and provides follow-up for more than two thousand PKU patients with known ethnic origin. 338 unrelated Chinese patients with two presumably causative alterations in the Pah gene from January 2006 through December 2012 were recorded in our database. In all patients, hyperphenylalaninemia (HPA) had been detected by either national screening (57.10%, 193/338) or presence of neurological deterioration at elder age (42.90%, 145/338) with a plasma phenylalanine (Phe) cut-off level of 120 μmol/L. A defect in the synthesis or recycling of tetrahydrobiopterin was excluded by analysis of urinary pterins and dihydropteridine reductase activity in erythrocytes. Metabolic phenotype for each patient was stratified according to the plasma Phe concentration before treatment, and all of which applied in this study were the maximum pretreatment values. Patients were classified as classic PKU (Phe, more than 1200 μmol/L), moderate PKU (Phe, 900 to 1200 μmol/L), mild PKU (Phe, 600 to 900 μmol/L) and MHP who keeps their Phe levels below 600 μmol/L on a free diet. The local Ethics Committee approved this study and informed consents were obtained from the parents of these patients enrolled.

a A null phenotype effect was declared when enzyme activity was demonstrated to be below the level of detection in the in vitro system (typically, b3% or b1% of normal. b From data in literature, although also represented in other categories at lower frequencies. c Identified in at least one patient with classic PKU in the homoallelic state and therefore, may be formally classified as a null mutation.

were excluded from the analyses that depended on in vitro expression information.

2.4. Phenotypic prediction system Mutations were assigned to one of the four-phenotype categories (classic, moderate, mild, and MHP), according to Guldberg et al. (1998). An arbitrary value (AV) was assigned to each mutation: AV = 1 for classical PKU mutation; AV = 2 for moderate PKU mutation; AV = 4 for mild PKU mutation, and AV = 8 for MHP mutation. Phenotypes resulting from a combination of the two mutant alleles were expressed as the sum of the two mutations' AVs. Sixty-six individuals, whose mutations without AV estimates derived from our data (Appendix A) or the literature, were excluded from those analyses that depended on AV information.

2.2. Genotype analysis and mutation nomenclature 2.5. Statistical analysis Genomic DNA was isolated from peripheral blood samples, 13 exons and related intronic boundaries of Pah gene were amplified, and all PCR products were scanned for mutations by direct sequence analysis. Mutations were referred to by their “trivial names”, as registered in the Pah Mutation Analysis Consortium Database. Genotypes were coded in two different ways. Initially, a simple sequential order was used. Thereafter, codes based on the residual activity of each allele were constructed. In this study, the nonsense, frame shift, splicing and those missense mutations with an enzyme activity in vitro less than 3% were given a definition of null Pah activity which were grouped in Table 1. Alleles classified as null or missense with some residual activity, were distributed among three genotype categories: null/null, null/ missense (functionally heterozygous) and missense/missense in the order of increasing predicted residual Pah activity (PRA). 2.3. Data on in vitro expression analysis PRA was recorded from data provided from in vitro experiments using recombinant expressed mutant proteins in eukaryotic cells. Expression data were compiled mainly from the Pah Mutation Analysis Consortium Database (http://www.pahdb.mcgill.ca/) or other published papers (Gersting et al., 2008; Gjetting et al., 2001; Guldberg et al., 1998; Kim et al., 2006; Okano et al., 1998; Pey et al., 2003; Waters et al., 1998). Mean PRA values, the average of the sum of activities of both alleles, were calculated for each genotype. The mutations present in 66 individuals lack expression data. These patients

Statistical analysis was implemented with SPSS 13.0®. Spearman correlation was estimated between the pretreatment Phe levels and mean PRA, AV sum respectively. A significance level of P b 0.05 was considered for all the analyses. The differences of pretreatment Phe concentrations among genotype groups based on PRA were evaluated by one-way ANOVA.

3. Results 3.1. Mutation and phenotype distribution A total of 338 unrelated patients were investigated. 109 different mutations were discovered, including 72 (66.06%) missense mutations, 15 nonsense, 15 splice-site and 7 frame-shift deletions. The eight most prevalent mutations of the 676 alleles were R243Q representing 24.11%, EX6-96ANG with 10.65%, R241C with 7.54%, R111X with 5.47%, IVS4-1GNA with 5.33%, V399V with 4.88%, Y356X with 4.88% and R413P with 4.44%. The first two accounted for one third of the identified mutations, and the next six for another third. The defined mutations were distributed from exon 2 through 7, and in exons 11 and 12. On the basis of individual data on pretreatment Phe levels, the patients were assigned to the four arbitrary metabolic phenotype categories, with 143 (42.31%) as classic PKU, 80 (23.67%) as moderate PKU, 71 (21.01%) as mild PKU and 44 (13.02%) as MHP.

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

T. Zhu et al. / Gene xxx (2013) xxx–xxx

3

Table 2 Predicted versus observed phenotypes in 272 PAH deficiency individuals in Chinese population. Genotypes

No.

AV sum

Predicted phenotype

Observed phenotype

Null/null Y166X/Y166X EX6-96ANG/EX6-96ANG R243Q/R243Q R413P/R413P IVS2+5GNC/IVS4-1GNA EX6-96ANG/R243Q R111X/R243Q R111X/Y356X IVS4-1GNA/R243Q R111X/V399V R243Q/V399V R243Q/R413P R111X/EX6-96ANG R241NPfs/R243Q R243Q/Y356X R111X/R413P IVS4-1GNA/Y356X IVS4-1GNA/R413P EX6-96ANG/Y356X EX6-96ANG/V399V A342T/R413P Y356X/V399V Y356X/R413P S70del/R158Q S70del/IVS4-1GNA S70del/R243Q R111X/Y166X R111X/IVS4-1GNA R111X/IVS7+2TNA R111X/R252Q Y166X/EX6-96ANG Y166X/R243Q R176X/IVS6-1GNA R176X/R243Q IVS4-1GNA/R241NPfs IVS4-1GNA/IVS6-1GNA IVS4+2TNA/Y356X IVS4+3GNC/EX6-96ANG IVS4+3GNC/V399V IVS5+1GNA/R243Q EX6-96ANG/R252W EX6-96ANG/R261X EX6-96ANG/S411X IVS6-1GNA EX6-96ANG IVS6-1GNA/R241NPfs IVS6-1GNA/R243Q IVS6-1GNA/R413P Q232X/R243Q R243Q/R243X R243Q/G247V R243Q/R252Q R243Q/A259T R243Q/W326X R243Q/C375X R261X/Y356X E280K/Y356X W326X/V399V V399V/R413P R408W/R413P R243Q/IVS7+2TNA EX6-96ANG/IVS7+2TNA IVS4-1GN/EX6-96ANG IVS4-1GNA/V399V EX6-96ANG/R413P EX6-96ANG/541-543delGAG S70del/EX6-96ANG R243Q/IVS12+4ANG R243Q/IVS12+6TNA Q301X/IVS12+4ANG S70del/R241NPfs R111X/IVS7+5GNA Y356X/IVS10-14CNG Total (null/null)

1 5 22 3 1 18 9 5 8 4 6 6 5 5 5 3 4 2 2 3 2 6 3 1 1 2 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 3 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 2 1 3 2 1 1 1 1 1 1 1 1 1 186

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c

1mi 3c/1mo/1MHP 14c/6mo/2mi 3c 1c 11c/2mo 1mo(20)/4mi 6c/1mo/2mi 4c/1mo 8c 1c/1mi/2MHP 1c/4mo/1mi 3c/1mo/2mi 4c/1mi 5c 2mo(20)/3c 3c 3c/1mo 2c 1c/1mi 2c/1mo 1c/1mo 6c 3c 1mi 1c 1c/1mo 1MHP 1mo 1mo(20) 1c 1mi 1c 1c 1c 1mo(20) 1MHP 1c 1c/1mo(20) 1mi 1c 1mo 1c 1mo(20) 2c/1mo 1c 1c 1mo(20) c 1c 1mo(20) 1c/1mi 1mo 1c 1c 1mo 1c 1mo(20) 1mi 1mi 1mo/1mi c 1c/1mo/1mi 2mo mo mi(15) c c MHP c c mi c

Inconsistencies (no.) 1 2 8 0 0 7 3 1 0 3 5 3 1 0 2 0 1 0 1 1 1 0 0 1 0 1 1 1 1 0 1 0 0 0 1 1 0 1 1 0 1 0 1 1 0 0 1 0 0 1 1 1 0 0 1 0 1 1 1 2 0 2 2 1 1 0 0 1 0 0 1 0 71 (continued on next page)

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

4

T. Zhu et al. / Gene xxx (2013) xxx–xxx

Table 2 (continued) Genotypes Null/missense E228X/I65T R243Q/R241C EX6-96ANG/R241C IVS4-1GNA/R241C R243Q/A434D R243Q/V388M IVS4-1GNA/A434D S70del/I65T R243Q/R155H V399V/R155H R243Q/F161S V399V/F161S S70del/R261Q R243Q/R261Q S70del/R241C R111X/R241C Y166X/R241C IVS7+2TNA/R241C L255S/R241C EX6-96ANG/Q419R R243Q/Q419R Y356X/A434D E280K/A434D E280K/H170Q R243Q/R408Q R270S/R408Q S70del/V388M R176X/V388M EX6-96ANG/V388M V399V/V388M R53H/R408W R53H/IVS7+2TNA IVS4-1GNA/R261Q IVS4-1GNA/D415N EX6-96ANG/D415N EX6-96ANG/R241H I65T/EX6-96ANG R241C/IVS7+1GNA F55L/1024delG(A342fsdelG) R241C/V399V I65T/R243Q R241C/R413P IVS6-1GNA/R408Q R111X/A434D R241C/L287I R241C/IVS12+6TNA R53H/R243Q Total (null/missense) Missense/missense F161S/R241C I65T/R241H R241C/V388M R241C/R241C Total (missense/missense) Total

No.

AV sum

Predicted phenotype

Observed phenotype

1 13 7 6 3 3 2 1 1 1 1 1 1 2 1 3 1 2 1 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 2 1 1 1 1 1 82

4 5 5 5 3 3 3 4 5 5 3 3 3 3 5 5 5 5 5 5 5 3 3 4 4 4 3 3 3 3 5 5 3 5 5 4 4 5 4 5 4 5 4 3 5 5 5

mo/mi mi mi mi mo mo mo mo/mi mi mi mo mo mo mo mi mi mi mi mi mi mi mo mo mo/mi mo/mi mo/mi mo mo mo mo mi mi mo mi mi mo/mi mo/mi mi mo/mi mi mo/mi mi mo/mi mo mi mi mi

1mo 3mo/4mi/6MHP 4mo/3MHP 6mi 1c/1mo/1mi 3mo 1c/1mi 1mo 1MHP 1mi 1mo 1mi 1MHP 1c/1mi 1MHP 3MHP 1MHP 2mi 1MHP 1MHP 1MHP 1mi 1mi 1c 2MHP 1mi 1mi 1mo 1c 1mi c mo 2c MHP MHP mi mo mi c mi mo 2mi mo mo MHP MHP MHP

1 1 1 1 4 272

6 6 6 8

mi mi mi MHP

1MHP 1MHP MHP 1MHP

Inconsistencies (no.) 0 9 7 0 2 0 2 0 1 0 0 1 0 2 1 3 1 0 1 1 1 1 1 1 2 0 1 0 1 1 1 1 2 1 1 0 0 0 1 0 0 0 0 0 1 1 1 50

1 1 1 0 3 124

AV = arbitrary value; C = classic PKU; Mo = moderate PKU; Mi = mild PKU. A phenotype that occurs at the border between two different classes based on the pretreatment values.

3.2. Classification of mutations and genotypes 109 different Pah mutations, all presumed to be phenotype modifying, were inherited in 186 genotype combinations, from 338 patients in our study. The mutations are classified as missense or null (putative or proved) (Table 1). The majority (n = 304, 89.94%) of patients were heteroallelic, of which 131 were “functionally hemizygous”. 3.2.1. Homoallelic mutant Pah genotypes There were 7 different homoallelic mutant genotypes harbored by 34 individuals. All other things being equal, the corresponding in vivo phenotypes should reflect the effect of the mutation on Pah enzyme activity and thus on the HPA biochemical phenotype. However, two homoallelic mutant genotypes conferred to more than one phenotypes;

R243Q/R243Q was associated with classic or moderate or mild PKU, and EX6-96ANG/EX6-96ANG was associated with classic or moderate PKU or MHP. Accordingly, the homoallelic genotypes are, in general, predictive, but sometimes they reveal an apparent inconsistency (34.38%) in the effect of Pah genotype on phenotype. 3.2.2. “Functionally hemizygous” genotypes Forty-eight different missense Pah mutations were inherited in combination with null mutation, by 130 patients who therefore were considered as “functionally hemizygous”, which would facilitate the assessment of the intrinsic severity of those missense mutations. For 5 mutations (I65T, A156P, S310F, A345T, R400K), the assignment was unambiguous because the phenotype classification was consistent in two or more functionally hemizyous patients. Thirty were found in

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

T. Zhu et al. / Gene xxx (2013) xxx–xxx

only one functionally hemizyous patient each and their assignment to particular phenotype categories may therefore be somewhat uncertain. The remaining 13 (R53H, R155H, R158W, L242F, R400T, G257V, F161S, R261Q, D415N, R241C, A434D, R408Q and V388M) were inconsistent in effect, because each was associated with more than one HPA phenotype. In all cases, the mutation was assigned to the category in which it appeared most often. On the basis of the criteria outlined above, AV estimates and in vitro expression data in the literature or those of which derived from some “functionally hemizygous” patients, 69 of the mutations of our database were assigned to the four-phenotype categories (Appendix A). 3.3. Correlation between genotype and phenotype In our study, 186 patients carried null/null genotypes, 130 carried null/missense genotypes and 22 carried missense/missense genotypes. 3.3.1. Phenotypic prediction system The results of the phenotype prediction based on AVs are summarized in Table 2. Among the 272 patients for whom AV information was available, 148 (54.41%) showed a phenotype in accordance with the AV predicted for their genotype. In this group, 115 patients had genotypes with two null mutations, the most prevalent ones were IVS4-1GNA, R241NPfs, R413P, S70del, R176X and IVS6-1GNA, with consistency in two or more patients. The rest of the group was composed of 32 patients with null/missense genotypes and 1 with two mutations having some residual enzymatic activity. The observed phenotypes in 124 patients were different form the expected phenotypes (Table 2). Seventy-one patients in the null/null genotype group did not show a classic PKU phenotype, and there are 43 genotype–phenotype inconsistencies among those individuals. For eight genotypes, 2 homoallelic and 6 heteroallelic, patients had been recorded in more than two phenotypes. 14 patients homozygous for the R243Q mutation were classified as classic PKU, while 6 patients as moderate and 2 patients as mild. Besides, 17 of 39 patients with genotypes composed of R243Q and a null mutation involving R111X, EX6-96ANG, Y356X and V399V were associated with moderate or mild PKU. In 13 patients, the observed phenotype was more than one category away from the expected phenotype. Also, it is notable that there were repeated inconsistencies for EX6-96ANG, as one homozygote patient assigned to MHP and 4 patients functionally hemizygous of it and a null mutation were associated with mild PKU. Eighty-six patients were assigned to the mild genotype groups, comprising null/missense and missense/missense, of which the observed phenotypes were different from the expected phenotypes in 53 patients. The lowest “concordance rate” (32.69%) was observed for patients with expected mild-PKU phenotype, whereas the correlation was satisfied for the two groups expected to show an MHP phenotype or a phenotype at the border between moderate and mild PKU (Table 3). 3.3.2. Statistical analysis The mean pretreatment Phe concentrations of the severe (null/null) and intermediate (null/missense) genotypes were 1393.66 μmol/L and 911.55 μmol/L, respectively, and were significantly higher than 741.82 μmol/L of the mild genotype (missense/missense) group (F = 50.79, P b 0.05). Spearman correlation analysis revealed a highly significant negative correlation (r = −0.4564, P b 0.0001) between PRA values and pretreatment Phe values. The AV sums also significantly correlated with the pretreatment Phe values (r = −0.5952, P b 0.0001). 4. Discussion Two meta-analyses (Guldberg et al., 1998; Kayaalp et al., 1997) and some molecular epidemiologic studies (Chien et al., 2004; Couce et al., 2013; Daniele et al., 2007; Lee et al., 2004; Mallolas et al., 1999; Okano

5

et al., 1998; Qu et al., 2008; Rivera et al., 2000, 2011; Song et al., 2005) have documented genotype–phenotype correlations in thousands of HPA patients and revealed discordant correlations in a number of individuals when they were related to the predicted effect of the mutant PAH genotype on enzyme function, but there were few similar reports in Chinese of large sample. Herein, we examined genotype–phenotype correlations in 338 PAHD patients harboring 186 different genotypes derived from 109 different mutant alleles with two approaches, arbitrary phenotypic prediction system and pair-wise correlation analysis. Whereas pretreatment Phe values were significantly correlated with PRA and AV sum, we showed that the observed metabolic phenotype is not always consistent with the predicted effect of genotype at the Pah locus. 4.1. Statistical analysis of the relation of genotype to biochemical variables Our study illustrated the use of the in vitro level of Pah activity of individual mutant Pah enzymes to relate the genotype of patients with PAHD to their biochemical phenotype, and demonstrated strong correlations between pretreatment values and PRA or AV sum. In PAHD, the biochemical phenotype groups were described based on two biochemical parameters–phenylalanine levels at diagnosis and dietary tolerance to phenylalanine, and the dietary-tolerance data was given the priority (Desviat et al., 1999; Guldberg et al., 1998). One point of contention here was the comparability of the pretreatment Phe values between neonatal screening and elder age for our data, since the metabolic phenotype was accountable to the exposition to the dietary phenylalanine and the above values were the most welldefined biochemical indexes and the major variables used to classify phenotypes. We therefore re-examined the relation between the pretreatment values and PRA or AV sum, by sub-grouping the patients according to whether they were diagnosed by neonatal screening or in elder age. The relation was not significantly different between these two subgroups (data not shown), suggesting that it was equally valid in subgroups. This was in accordance with the view that any pretreatment serum Phe value was potentially useful in helping diagnosis and prognosis regardless of the age of the patient at the time of testing (Okano et al., 1991). 4.2. Arbitrary value-based phenotypic prediction system The strong correlation, derived from our study, between biochemical variables and AV sum or PRA confirms that predicted Pah activity could reflect enzyme activity in vivo and therefore permit the prediction of the PAHD phenotype. So we used the AV-based method for phenotypic prediction in order to elucidate the possible causes of genotype– phenotype inconsistencies. A consistency for those genotypes composed of null/null mutations (61.83%) was observed, corresponding to the fact that classic PKU comprised the predominant type in our sample (42.31%). Also, a virtually complete association (100%) between genotype and phenotype in the groups of individuals presenting with MHP was discovered, in whom phenotype classification was based solely on pretreatment serum Phe values without complicated by a dietary-therapy regimen. The poorer concordance rates were inspected for patients with expected moderate (36.36%) and mild-PKU (32.692) phenotypes with mutations of in vitro residual activities in the range of 10%–40%. In general, the following are some points to address the relatively high rate of inconsistencies. Firstly, maybe a true lack of correlation between genotype and biochemical phenotype exists, which would be opposed by our findings that a simple genotype–phenotype correlation is present in most patients. Secondly, we should admit that we predict the in vivo Pah activity by analysis of in vitro data for the mutations, which might lead to some genotype misclassifications; we should also admit the inadequacy of

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

6

T. Zhu et al. / Gene xxx (2013) xxx–xxx

Table 3 The observed “concordance rate” for PAH deficiency patients with their expected metabolic phenotypes based on arbitrary phenotypic prediction system. No. with observed phenotypeb PKU AV1 + AV2

Expected phenotypea

Classic

Moderate

Mild

MHP

Total

Perfect matches (%)c

2 3 4 5 and 6 8 9–16 Total

Classic PKU Moderate PKU Moderate/mild PKU Mild PKU Mild PKU/MHP MHP

115* 6 2 1 0 0 124

40 8* 5* 8 0 0 61

25 7 2* 17* 0* 0 51

6 1 2 26 1* 0* 36

186 22 11 52 1 0 272

61.8% 36.4% 63.6% 32.7% 100%

a b c

Determined, for each patient, on the basis of the sum of the AVs of the two Pah mutations (according to the entries in the Appendix A and Table 1. Groups in which the expected phenotype matches the observed phenotype are marked by asterisks (*). Between the observed phenotype and the expected phenotype.

our phenotypic classification for those that occur at the border between two different classes, since on the basis of the analysis of our patients, a continuous spectrum of PAHD phenotypes was observed and some patients have been assigned to the adjacent phenotype classes (e.g. either classic PKU or mild PKU for those who were expected to classify as moderate PKU and moderate PKU or MHP for those who were expected to classify as mild PKU) based on their pretreatment Phe levels. Thirdly, the preferred prevalence of particular mutations or mutation combinations in a specific population might be one of the potential influential factors for the following situations. In our study, we identified a number of cases associated with distinct mutations which may not be explained by above genotypic or phenotypic misclassifications. Notably, R243Q mutation, with an especially high frequency in Chinese patients, appeared either in both outlier classes (e.g. classic PKU and mild PKU) in the homoallelic state or in all four classes in functionally hemizygous and compound heterozygous state, implying that events other than expression of the mutant genotype at the Pah locus itself (Bagheri and Wagner, 2004; Kaufman, 1999; Scriver et al., 1985) contribute to the HPA phenotype in the patient. Pey et al. (2003) found that the amount of the corresponding mutant PAH proteins and their residual activities could be modulated by in vitro experimental conditions for several missense mutations, including R243Q, which associated with a low level of activity (10%) (Wang et al., 1991) in general. Such in vitro modulation has also been described in cystic fibrosis and Parkinson's disease, and the corresponding damaged or abnormal proteins have been demonstrated to rely on elaborated pathways of protein quality control and removal to maintain intracellular protein homeostasis (Bialecka et al., 2009; Nascimento et al., 2011). So we could speculate that the observed phenotypic variability might be due to the inter-individual differences in the components of the cellular quality control system, which include molecular chaperones, the ubiquitin–proteasome system (UPS) and the autophagy–lysosomal pathway (ALP), for which Zurfluh et al. (2008) have classified R243Q in PAH associating with a severe form of PKU in general as a BH4-responsive mutation, which caused reduced stability and accelerated degradation, but BH4 could prevent some mutant PAH from both misfolding and inactivation. Besides, the significant phenotypic dissociation holds identical for the R241C, which was classified as mild mutation and associated predominantly with a mild phenotype (Okano et al., 1994). However, in our sample, 7 of 40 patients with genotypes composed of R241C and any known null allele were associated with moderate PKU, 16 with mild PKU and 17 with MHP. Additionally, the same occurred with V388M (Gamez et al., 2000), which was assigned to the moderate category. One of seven patients bearing V388M and any null allele presented classic PKU, 4 moderate and 2 mild. The above remarkable inconsistencies seemed to argue the dominant observation (Moller et al., 2011) that the milder of two mutations being “quasidominant” and determining the phenotypic outcome. Accordingly, the possibility of interactions between Pah monomers in compound heterozygotes underlying this variability has been rigorously studied in

multiple systems (Velasco-Garcia and Vargas-Martinez, 2012; Waters, 2003; Waters et al., 1998). The R241C mutation (Kim et al., 2006) was observed to occur along the interface region of the regulatory domain, suggesting that upon this mutation, dimer stability is reduced. The V388M mutation (Gamez et al., 2000) was analyzed in a mammalian two-hybrid system and indicated that the misfolded mutant Pah subunits interacted to some extent with wild-type subunits. Other interesting observations were found for the spicing site mutations, such as EX696ANG (Ellingsen et al., 1997) and V399V (Huang et al., 1991), which correspond to a severe phenotype based on their characteristics. From our data, only 3 of 5 individuals homozygous for EX6-96ANG had classic PKU and 25 of 44 patients with genotypes composed of this mutation and any other null one were associated with classic PKU. The V399V mutation is another example for the same situation. Only 10 of 18 patients bearing V399V and any other null allele were observed as classic PKU, suggesting that the effect of this mutation was less severe than previously reported. It was deemed that splice site mutations could cause classic or mild PKU depending on “read through” (i.e. normal splicing may sometimes occur despite the mutation) (Dworniczak et al., 1991; Kole et al., 2004) and the response to in vitro modulation was clearly mutation specific.

5. Conclusions In view of the results obtained in this work, further supports have been provided for the notion that the HPA phenotype at its metabolic level reflects that the allelic variation at the Pah locus is the major determinant nevertheless there is discordance between the mutant Pah genotype and corresponding phenotype in Chinese population. Additional studies on how Pah mutations alter Pah protein integrity and function and how the mutant protein might be modulated by environmental factors remain necessary to bring greater clarity to the genotype–phenotype correlation in our population. We hope to find these answers in research yet to set up.

Conflict of interest All authors have no conflicts of interest to disclose.

Acknowledgments We greatly acknowledge Jing-qiu Ma (Department of Child and Adolescent Healthcare, Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine) in data management and processing. We also thank Ya-fen Zhang, Zhi-wen Gong and Jian-De Zhou (Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine) for help and direction in the laboratory work.

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

T. Zhu et al. / Gene xxx (2013) xxx–xxx

7

Appendix A

Assignment of 109 PAH mutations to metabolic phenotypes. Characteristics of 338 individuals (number of patients) Classic PKU

Moderate PKU b

R243Q(163) EX6-96ANG(72)b R111X(37)a V399V(33)b Y356X(33)a IVS4-1GNA(36)a R413P(30)a IVS6-1GNA(10)a G257V(8)c S70del(10)a Y166X(6)a R241NPfs(9)a W326X(3)a IVS7+2TNA(7)a R176X(3)a E280K(3)a A156P(4)a G247V(2)a L255S(2)a R261X(2)a R252Q(3)a 541-543delGAG(1)a a b c

c

L385P(2) R400K(3) IVS7+1GNA(1)a C375X(1)a A342T(2)a R252W(1)a 1024delG(1)a E228X(1)a IVS7+5GNA(1)a R243X(1)a IVS12+6TNA(2)a Q232X(1)a IVS12+4ANG(2)a R408W(1)a Q301X(1)a R158Q(1)a L287I(1)a A259T(1)a Y414X(1)a 190-194delCACAT(1)a W187X(1)a IVS5-2ANG(1)a F39del(1)a S310F(2)c IVS5+1GNA(1)a Q172X(1)a IVS4+2TNA(1)a IVS10-14CNG(2)a IVS2+5GNC(1)a S411X(1)a R270S(1)a IVS4+3GNC(3)a 540delGGAGG(1)a

b

V388M(9) A434D(8)b F161S(3)b R261Q(5)b

Mild PKU

MHP b

R408Q(8) I65T(5)a R241H(3)a F55L(3)b

Unclassified b

R241C(51) Q419R(2)b R155H(2)b H170Q(1)b R53H(3)a D415N(2)b

L242F(3)G239D(1) C357Y(2)I224T(1) R158W(3)Y154C(1) R400T(5)E56D(1) Y154H(1)P362L(1) L227Q(1)A345T(2) S70P(1)S350Y(1) L367R(1)R157K(2) P275A(1)I324N(1) M276R(1)E183G(1) I406T(1)F121L(1) F392I(1)G247R(2) R157I(1)M276K(1) R400S(2)Y206C(1) H107R(1) C265R(1) Y417C(1) L62V(1) L348P(1) Q375E(1) A322D(1) E76D(1) E286K(1) T380M(1) S391I(1) S349A(1)

G239D(1) I224T(1) Y154C (1) E56D(1) P362L(1) A345T(2) S350Y(1) R157K(2) I324N(1) E183G(1) F121L(1) G247R(2) M276K(1) Y206C(1) C265R(1) L62V(1) Q375E(1) E76D(1) T380M(1) S391I(1) S349A(1)

Unambiguous expression of mutations. Also represented in other phenotype categories, although at lower frequencies. Derived from homozygous patients or at least two functionally hemizygous patients.

References Acosta, A., Silva Jr., W., Carvalho, T., Gomes, M., Zago, M., 2001. Mutations of the phenylalanine hydroxylase (PAH) gene in Brazilian patients with phenylketonuria. Hum. Mutat. 17, 122–130. Bagheri, H.C., Wagner, G.P., 2004. Evolution of dominance in metabolic pathways. Genetics 168, 1713–1735. Bialecka, M., Robowski, P., Honczarenko, K., Roszmann, A., Slawek, J., 2009. Genetic and environmental factors for hyperhomocysteinaemia and its clinical implications in Parkinson's disease. Neurol. Neurochir. Pol. 43, 272–285. Chien, Y.H., et al., 2004. Mutation spectrum in Taiwanese patients with phenylalanine hydroxylase deficiency and a founder effect for the R241C mutation. Hum. Mutat. 23, 206. Couce, M.L., et al., 2013. Molecular epidemiology and BH4-responsiveness in patients with phenylalanine hydroxylase deficiency from Galicia region of Spain. Gene 521, 100–104. Daniele, A., et al., 2007. Molecular epidemiology of phenylalanine hydroxylase deficiency in Southern Italy: a 96% detection rate with ten novel mutations. Ann. Hum. Genet. 71, 185–193. Desviat, L.R., et al., 1999. Genetic and phenotypic aspects of phenylalanine hydroxylase deficiency in Spain: molecular survey by regions. Eur. J. Hum. Genet. 7, 386–392. Dworniczak, B., et al., 1991. Aberrant splicing of phenylalanine hydroxylase mRNA: the major cause for phenylketonuria in parts of southern Europe. Genomics 11, 242–246. Ellingsen, S., Knappskog, P.M., Eiken, H.G., 1997. Phenylketonuria splice mutation (EXON6nt-96A–Ng) masquerading as missense mutation (Y204C). Hum. Mutat. 9, 88–90. Gamez, A., Perez, B., Ugarte, M., Desviat, L.R., 2000. Expression analysis of phenylketonuria mutations. Effect on folding and stability of the phenylalanine hydroxylase protein. J. Biol. Chem. 275, 29737–29742. Gersting, S.W., et al., 2008. Loss of function in phenylketonuria is caused by impaired molecular motions and conformational instability. Am. J. Hum. Genet. 83, 5–17. Gjetting, T., Petersen, M., Guldberg, P., Guttler, F., 2001. In vitro expression of 34 naturally occurring mutant variants of phenylalanine hydroxylase: correlation with metabolic phenotypes and susceptibility toward protein aggregation. Mol. Genet. Metab. 72, 132–143. Groselj, U., et al., 2012. Five novel mutations and two large deletions in a population analysis of the phenylalanine hydroxylase gene. Mol. Genet. Metab. 106, 142–148. Guldberg, P., et al., 1998. 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, 71–79. Huang, S.Z., Ren, Z.R., Zeng, Y.T., Woo, S.L., 1991. PAH 399 GTA (Val)——GTT(Val), a new silent mutation found in the Chinese. Hum. Genet. 86, 305–306. Kasnauskiene, J., Cimbalistiene, L., Kucinskas, V., 2003. Validation of PAH genotypebased predictions of metabolic phenylalanine hydroxylase deficiency phenotype: investigation of PKU/MHP patients from Lithuania. Med. Sci. Monit. 9, CR142–CR146.

Kaufman, S., 1999. A model of human phenylalanine metabolism in normal subjects and in phenylketonuric patients. Proc. Natl. Acad. Sci. U. S. A. 96, 3160–3164. Kayaalp, E., et al., 1997. Human phenylalanine hydroxylase mutations and hyperphenylalaninemia phenotypes: a metanalysis of genotype–phenotype correlations. Am. J. Hum. Genet. 61, 1309–1317. Kim, S.W., et al., 2006. Structural and functional analyses of mutations of the human phenylalanine hydroxylase gene. Clin. Chim. Acta 365, 279–287. Kole, R., Williams, T., Cohen, L., 2004. RNA modulation, repair and remodeling by splice switching oligonucleotides. Acta Biochim. Pol. 51, 373–378. Lee, D.H., et al., 2004. The molecular basis of phenylketonuria in Koreans. J. Hum. Genet. 49, 617–621. Mallolas, J., et al., 1999. Mutational spectrum of phenylalanine hydroxylase deficiency in the population resident in Catalonia: genotype–phenotype correlation. Hum. Genet. 105, 468–473. Mitchell, J.J., Trakadis, Y.J., Scriver, C.R., 2011. Phenylalanine hydroxylase deficiency. Genet. Med. 13, 697–707. Moller, L.B., et al., 2011. Clinical presentation and mutations in Danish patients with Wilson disease. Eur. J. Hum. Genet. 19, 935–941. Nascimento, C.M., et al., 2011. Serum homocysteine and physical exercise in patients with Parkinson's disease. Psychogeriatrics 11, 105–112. Okano, Y., et al., 1991. Molecular basis of phenotypic heterogeneity in phenylketonuria. N. Engl. J. Med. 324, 1232–1238. Okano, Y., et al., 1994. Molecular characterization of phenylketonuric mutations in Japanese by analysis of phenylalanine hydroxylase mRNA from lymphoblasts. Hum. Mol. Genet. 3, 659. Okano, Y., et al., 1998. Molecular characterization of phenylketonuria in Japanese patients. Hum. Genet. 103, 613–618. Okano, Y., Kudo, S., Nishi, Y., Sakaguchi, T., Aso, K., 2011. Molecular characterization of phenylketonuria and tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency in Japan. J. Hum. Genet. 56, 306–312. Pey, A.L., Desviat, L.R., Gamez, A., Ugarte, M., Perez, B., 2003. Phenylketonuria: genotype–phenotype correlations based on expression analysis of structural and functional mutations in PAH. Hum. Mutat. 21, 370–378. Qu, Y.J., et al., 2008. Spectrum of phenylalanine hydroxylase gene mutations and genotype–phenotype correlation in the patients with phenylketonuria in Beijing area of China. Zhonghua Er Ke Za Zhi 46, 115–119. Rivera, I., et al., 2000. The correlation of genotype and phenotype in Portuguese hyperphenylalaninemic patients. Mol. Genet. Metab. 69, 195–203. Rivera, I., et al., 2011. Phenylalanine hydroxylase deficiency: molecular epidemiology and predictable BH4-responsiveness in South Portugal PKU patients. Mol. Genet. Metab. 104, S86–S92 (Suppl.). Scriver, C.R., Gregory, D.M., Sovetts, D., Tissenbaum, G., 1985. Normal plasma free amino acid values in adults: the influence of some common physiological variables. Metabolism 34, 868–873. Song, F., et al., 2005. Phenylketonuria mutations in Northern China. Mol. Genet. Metab. 86 (Suppl. 1), S107–S118.

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079

8

T. Zhu et al. / Gene xxx (2013) xxx–xxx

Velasco-Garcia, R., Vargas-Martinez, R., 2012. The study of protein-protein interactions in bacteria. Can. J. Microbiol. 58, 1241–1257. Wang, T., et al., 1991. Missense mutations prevalent in Orientals with phenylketonuria: molecular characterization and clinical implications. Genomics 10, 449–456. Waters, P.J., 2003. How PAH gene mutations cause hyper-phenylalaninemia and why mechanism matters: insights from in vitro expression. Hum. Mutat. 21, 357–369. Waters, P.J., Parniak, M.A., Nowacki, P., Scriver, C.R., 1998. In vitro expression analysis of mutations in phenylalanine hydroxylase: linking genotype to phenotype and structure to function. Hum. Mutat. 11, 4–17.

Zhan, J.Y., Qin, Y.F., Zhao, Z.Y., 2009. Neonatal screening for congenital hypothyroidism and phenylketonuria in China. World J. Pediatr. 5, 136–139. Zhu, T., et al., 2010. Mutational spectrum of phenylketonuria in the Chinese Han population: a novel insight into the geographic distribution of the common mutations. Pediatr. Res. 67, 280–285. Zschocke, J., 2003. Phenylketonuria mutations in Europe. Hum. Mutat. 21, 345–356. Zurfluh, M.R., et al., 2008. Molecular genetics of tetrahydrobiopterin-responsive phenylalanine hydroxylase deficiency. Hum. Mutat. 29, 167–175.

Please cite this article as: Zhu, T., et al., Variations in genotype–phenotype correlations in phenylalanine hydroxylase deficiency in Chinese Han population, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.07.079