- Email: [email protected]
Genotype-phenotype correlation among Malaysian patients with osteogenesis imperfecta
Clinica Chimica Acta 484 (2018) 141–147
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Genotype-phenotype correlation among Malaysian patients with osteogenesis imperfecta
T
Nadiah Mohd Nawawia, Nalini M. Selveindranb, Rahmah Rasatc, Chow Yock Pinga, ⁎ Zarina Abdul Latiffc, Syed Zulkifli Syed Zakariac, Rahman Jamala, Nor Azian Abdul Murada, ,1, ⁎ Bilkis Banu Abd Azizc, ,1 a
UKM Medical Molecular Biology Institute (UMBI), Jalan Yaacob Latiff, Bandar Tun Razak, 56000 Cheras Kuala Lumpur, Malaysia Department of Pediatrics, Hospital Putrajaya, Jalan P9, Federal Government Administration Centre, Precint 7, 62250 Putrajaya, Malaysia c Department of Pediatrics, UKM Medical Centre, Jalan Yaacob Latiff, Bandar Tun Razak, 56000 Cheras Kuala Lumpur, Malaysia b
A R T I C LE I N FO
A B S T R A C T
Keywords: Brittle bone disease Genotype Osteogenesis imperfecta Phenotype
Background: Osteogenesis imperfecta (OI) is a rare genetic bone disease characterized by bone fragility and low bone mass. OI was mainly caused by genetic mutations in collagen genes, COL1A1 and COL1A2. Nevertheless, new genes have been identified to be causally linked to OI. The clinical features between each OI groups share great similarities and it is sometimes difficult for clinicians to diagnose the disease accurately. Here, we identify the genetic mutations of OI patients from Malaysia and correlate the genetic mutations with the clinical features. Method: Targeted sequencing of fourteen genes panel was performed to identify the mutations in 29 OI patients with type I, III, IV and V disease. The mutations were determined using Ion Torrent Suite software version 5 and variant annotation was conducted using ANNOVAR. The identified mutations were confirmed using Sanger sequencing and in silico analysis was performed to evaluate the effects of the candidate mutations at protein level. Results: Majority of patients had mutations in collagen genes, 48% (n = 14) in COL1A1 and 14% (n = 4) in COL1A2. Type I OI was caused by quantitative mutations in COL1A1 whereas most of type III and IV were due to qualitative mutations in both of the collagen genes. Those with quantitative mutations had milder clinical severity compared to qualitative mutations in terms of dentinogenesis imperfecta (DI), bone deformity and the ability to walk with aid. Furthermore, a few patients (28%, n = 8) had mutations in IFITM5, BMP1, P3H1 and SERPINF1. Conclusion: Majority of our OI patients have mutations in collagen genes, similar to other OI populations worldwide. Genotype-phenotype analysis revealed that qualitative mutations had more severe clinical characteristics compared to quantitative mutations. It is crucial to identify the causative mutations and the clinical severity of OI patients may be predicted based on the types of mutations.
1. Introduction Osteogenesis Imperfecta (OI) is a group of genetic bone disorders primarily characterized by low bone mass, which results in increased bone fractures and disrupted growth [1]. Additional extra-skeletal phenotypes include blue sclera, dentinogenesis imperfecta (DI) and hearing impairment [2]. There is a wide spectrum of clinical features and severity among patients, ranging from mild with minimal or no bone deformities to severe with multiple bone fractures and also perinatal lethality [3]. The variable types of OI occurs in approximately 1 in every 15,000–20,000 births with majority of them being autosomal
⁎
1
dominant inheritance [4]. The classic Sillence classification categorized OI into four numerical groups (OI type I to IV) according to its clinical findings and inheritance patterns [5]. Type I is often the mildest form, with normal or near-normal skeletal features. Type II is lethal in the perinatal period, with multiple intrauterine fractures and skeletal deformities of the limbs. Type III is the severe form with frequent fractures and bone deformities that progress with age. Type IV has various phenotypes from mild to moderate skeletal deformities and varies in height and statures [5]. On the other hand, type V OI was mainly identified by the presence of hyperplastic callus and interosseous membrane [6].
Corresponding authors. E-mail addresses: [email protected] (N.A. Abdul Murad), [email protected] (B.B. Abd Aziz). Bilkis Banu Abd Aziz is the Principal Investigator of the grant and Nor Azian Abdul Murad is responsible for student's supervision.
https://doi.org/10.1016/j.cca.2018.05.048 Received 4 April 2018; Accepted 24 May 2018 Available online 25 May 2018 0009-8981/ © 2018 Elsevier B.V. All rights reserved.
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
LEPRE1, PPIB, SERPINH1, SERPINF1, FKBP10, P3H1, BMP1, TMEM38B, IFITM5, PLOD2 and WNT1. Two primer pools to amplify 320 amplicons were designed using Ion AmpliSeq Designer version 4.48 using the ‘standard DNA’ (225-bp amplicon target sizes) and ‘Gene + UTR’ options (https://www.ampliseq.com). The number of amplicons and base covered for each gene are listed in Supplementary Table 1. A total of 10 ng DNA per sample was used for target enrichment by multiplex PCR and barcoded libraries from 16 samples were pooled in a single tube (Thermo Fisher Scientific, USA). Library pools were subjected to Ion Chef ™ instrument (Thermo Fisher Scientific, USA) for template preparation and enriched Ion Sphere Particles (ISPs) were loaded on Ion 316 v2 BC Chip for sequencing. Potentially pathogenic variants were validated using Sanger sequencing. The candidate mutations were also compared against 667 chromosomes with similar age, gender and ethnicity.
The initial assumption was that this disease was caused by a mutation in either COL1A1 or COL1A2 genes which codes for two collagen type I alpha chains, α1 (1) and α1 (2) [7]. However, with the advancement of genetic testing, mutations in non-collagenous genes have been detected such as LEPRE1, PPIB, SERPINH1, FKBPIO, SP7, IFITM5, BMP1, WNT1, TMEM38B, CRTAP and SERPINF1 genes, all of which are linked to the autosomal recessive OI except for IFITM5 mutation [8]. There are two groups of mutational defects in type I collagen. The first is quantitative mutations due to frameshift, nonsense and splice-site mutations, which introduces a premature termination codon in the coding sequence of COL1A1 allele. The second group is qualitative mutations resulted from the synthesis of collagen molecules that has structural abnormalities, most frequently caused by glycine substitution [9]. The phenotypic severity of OI patients depends on the affected gene, position of mutation, substituted amino acid and the final protein product [10]. However, OI has variable phenotypes and to date, the exact correlation between genotype and phenotype is not fully understood. Hence, we aimed to determine the genetic mutations involved in our local OI patients. Subsequently, the phenotype of these patients will be linked to their genotype and the data obtained may be useful in the clinical diagnosis, genetic counselling and prenatal diagnosis for OI patients
2.5. Sequencing data analysis Data from sequencing run were processed using Ion Torrent Suite software (version 5.0.4; Thermo Fisher Scientific, USA) for base calls, read alignments and variant calling using the reference genomic sequence (hg19). Called variants were annotated using ANNOVAR [12] and the variants were filtered out if the minor allele frequency was equal or higher than 1% in 1000 Genomes Project database and National Heart, Lung, and Brain Institute-Exome Sequencing Project (NHLBI-ESP) with 6500 exomes. Pathogenic variants were predicted using SIFT score (< 0.05) [13], PolyPhen2 score (> 0.4) [14], and MutationTaster [15]. The Integrative Genomics Viewer (IGV) software was used for visualization of protein deleterious effects [16,17]. Novel mutations were identified using Osteogenesis Imperfecta and Ehlers Danlos Syndrome database (https://oi.gene.le.ac.uk/home.php).
2. Materials and methods 2.1. Subjects The study was implemented in compliance with the ethical principles formulated in the declaration of Helsinki and was approved by our local Ethics Committee. Patients granted their informed, written consent to participate. Patients were recruited from the UKM Medical Centre (UKKMC) and Hospital Putrajaya. Written informed consent was given by patients, their parents or legal representatives. The inclusion criteria for OI patients include being Malaysian and with the clinical features of OI.
2.6. Data analysisfor genotype-phenotype correlation Statistical analysis was performed using Statistical Package for Social Sciences (SPSS) version 21. Age differences between the patients and normal was calculated using Independent T Test. Qhi-square was used to determine the gender differences between the patients and normal group. Fisher test was carried out to determine the correlation between gender, OI groups and types of colagen with the patients phenotype. P less than 0.05 was considered as significant.
2.2. Clinical characteristics All 29 patients from 28 families which comprise of 16 males and 13 females were recruited. They underwent clinical and physical examination to evaluate their clinical features and all medical histories were recorded. The clinicians classified the patients into four groups (OI-I, OI-III and OI-IV) according to Sillence classification. OI-V was categorized according to description by Glorieux et al. [6]. The key features observed were blue sclera, bone deformities, degree of disability, dentinogenesis imperfecta (DI) and hearing ability. DI was characterized by the formation of translucent and brittle teeth. For patient's degree of disability, Bleck's 5-point scale was used to assess mobility and walking ability [11].
3. Results 3.1. Clinical characteristics Most of our patients (62.1%) were OI type III, followed by 17.2% of type I and IV respectively. One patient was classified as type V. Table 1 shows the relationship between the clinical characteristics of OI patients and types of OI. None of the patients had OI type II. No group differences were found in the prevalence of blue sclera, however bone deformities were significantly prominent in patients with OI type III as
2.3. Sample preparation Total genomic DNA was extracted from EDTA-preserved peripheral blood using Macherey-Nagel Nucleospin Blood QuickPure kit (Macherey-Nagel, Germany. The Gender validation test was performed for each DNA sample as part of quality assessment. Briefly, PCR was performed using SRY and ATL1 primers and assessed by agarose gel electrophoresis.
Table 1 Phenotypic characteristics among OI patients.
2.4. DNA sequencing Targeted sequencing was performed using Ion PGM Semiconductor ™ (Thermo Fisher Scientific, USA) sequencer with a total of 14 genes with 214 exons selected for the gene panel. The selected genes are associated with osteogenesis imperfecta: COL1A1, COL1A2, CRTAP, 142
Characteristics, n (%)/ Types of OI, n (%)
OI type I 5 (17.2)
OI type III 18 (62.2)
OI type IV 5 (17.2)
OI type V1 (3.4)
p value
COL1A1/COL1A2 Gender (Male/Female) Blue sclera DI Hearing loss Bone deformity Walk with aid
5/0 3/2 5 (17.2) 3 (14.3) 1 (50) 1 (4.5) –
6/3 9/9 18 (62.2) 15 (71.4) – 18 (81.8) 12 (80)
3/1 3/2 5 (17.2) 3 (14.3) 1 (50) 2 (9.1) 2 (13.3)
– 0/1 1 (3.4) – – 1 (4.5) 1 (6.7)
0.392 1.000 – 0.145 0.623 < 0.001 0.028
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
Table 2 Mutational analysis of 14 patients in COL1A1. Individual
OI types
Exon number
Nucleotide change
Protein change
Type of mutation
Novelty
Population
OI 1
III
36
c.2461G > A
p. Gly821Ser
Missense
–
OI 4 OI 5 & OI 20
III III
19 44
c.1273G > A c.3226G > A
p. Gly425Ser p. Gly1076Ser
Missense Missense
– –
OI OI OI OI OI OI
I I III III I IV
37 28 24 31 15 23
c.2560 - 1G > A c.1900C > T c.1631delC c.2028 + 1G > A c.994G > A c.1588G > A
p. Gly332Arg p. Gly530Ser
Splice site Nonsense Frameshift Splice site Missense Missense
– Yes Yes – – –
USA [18], Denmark [19], China [20], Vietnam [21], Sweden [10], South Korea [22], Taiwan [23], Italy [24,25], Poland (Unpublished data), Belgium (Unpublished data) USA [18,26], Denmark [19], Taiwan [23], Greece [27] USA [18], Canada [28], Denmark [19], Sweden [10], China [20], Vietnam [21], Egypt [29], Vietnam [21] USA [18]
III III III
31 32 8
c.2029 - 1G > A c.2155G > A c.599G > T
p. Gly719Ser p. Gly200Val
Splice site Missense Missense
Yes – –
6 7 15 16 17 19
OI 23 & OI 24 OI 25 OI 27
p. Glu634X p. Pro544fs
Belgium (Unpublished data) USA [18], Sweden [10], Canada [28] USA [18], Canada (Unpublished data), Denmark (Unpublished data), Australia [30] USA [18,31], Sweden [10], China [20], India [32] UK [33], Poland [34]
opposed to qualitative mutations (p < .001). Those with quantitative mutations also had independent walking ability, whereas those with qualitative mutations had to walk with aid. In addition, those with quantitative mutations showed lower presence of DI compared to patients with qualitative mutations (28.6% vs 71.4%). Besides that, mutations in α1 (1) were more frequently identified and the patients showed independent walking ability compared to α1 (2) (p < .05). The phenotypic variation with relation to the position of mutation was not significant due insufficient sample size. However, in the patient with mutation affecting amino terminal of Gly22 (p.G200 V) in α1 (1) showed no presence of DI (Fig. 1), whereas two patients with mutations at the carboxyl terminal of Gly898 (p.G1076S) were DI positive. Despite the wide gap of mutation distribution in both of α (1) chains, one patient with OI type III had mutation (Gly568) in one of the eight lethal clusters described by Marini et al. [18] and no mutations were listed among the lethal regions of α1 (1).
compared to both type I and IV (p < 0.01). DI was identified in 71.4% of OI type III patients and 14.3% in both of type I and IV respectively. Patients with OI type I had independent walking ability whereas 66.7% of type III and 40% of type IV patients had to walk with aid. Only two patients of OI type I and III had hearing impairment. Table 5 shows correlation between the clinical characteristics and types of mutation. 3.2. Mutational analysis A total of 22 variants were identified in 26 patients and most variants were found in COL1A1 and COL1A2 genes. There were 12 COL1A1 mutations and 4 COL1A2 mutations in which 3 variants were novel in COL1A1 and 3 novel variants in COL1A2 (Tables 2 and 3). Eleven patients had glycine substitution within the triple helix domain. Eight individuals had glycine substitution to serine and one patient had substitution to arginine, aspartic acid and valine respectively. Six patients had quantitative mutations caused by frameshift, stopgain or splice site variant. Six missense mutations have also been detected in IFITM5, BMP1, P3H1 and SERPINF1 genes (Table 4). Five of the variants were novel. All patients were non consanguineous expect for a pair of twins (OI 23 and OI 24) who shared the same splice site mutations in COL1A1. Individual OI 9 was found to have double mutations in COL1A2 (p.R978H and p.G400D). Three patients had no significant mutations among the genes in the panel.
4. Discussion We have identified 16 (73%) mutations in COL1A1 and COL1A2 in 18 OI patients from 17 families. Previous studies showed 51 to 60% of OI patients had mutations in these two genes [21,23,46]. These results support the early statement that majority of OI patients had mutations in collagen genes, which caused the autosomal dominant OI [18]. Mutation p.Gly1027Ser was identified in two of our patients with OI type III. This mutation was postulated to be a hotspot mutation in COL1A1 instead of founder effect, because it has been found as de novo mutation in other populations [10,20,21,29,47]. Previously, Takagi et al., demonstrated a severe OI type II patient due to double glycine substitutions in COL1A2 (p.Gly208Glu and p.Gly235Asp) [48]. In contrast, our patient with double mutations in COL1A2 only had one glycine substitution within the helical domain and was diagnosed as OI type III. Similar with other published reports, our study also showed that
3.3. Genotype-phenotype correlation There was a significant association between the types of mutation in alpha 1 chains with the types of OI (p < .01). Around 80% of patients with OI type I and 50% of OI type IV had quantitative mutations in α1 (1) chains while none occurred in α1 (2) chain. On the other hand, qualitative mutations were identified in all type III and two of type IV patients in both of COL1A1 and COL1A2. Individuals with quantitative mutations had milder clinical severity, with no bone deformities as Table 3 Mutational analysis of 4 patients in COL1A2. Individual
OI types
Exon number
Nucleotide change
Protein change
Type of mutation
Novelty
OI 3 OI 8
III III
33 46
c.1972G > A c.3034G > A
p. Gly1972Ser p. Gly1012Ser
Missense Missense
Yes –
OI 9
III
OI 12
IV
44 22 44
c.2933G > A c.1199G > A c.2933G > A
p. Arg978His p. Gly400Asp p. Arg978His
Missense Missense Missense
Yes Yes Yes
143
Population
USA [18], Sweden [10], South Korea [22], India [32], Vietnam [21], Italy [35], Finland [36]
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
Table 4 Mutational analysis of 8 patients in other genes. Individual
OI types
Genes
Exon number
Nucleotide changes
OI 2
III
IFITM5
UTR
c.-14C > T
OI OI OI OI OI OI OI
IV III III III III V III
P3H1 P3H1 SERPINF1 BMP1 IFITM5 IFITM5 P3H1
1 1 5 17 UTR UTR 13
c.26 T > C c.26 T > C c.620 T > G c.2350C > G c.-1_0insC c.-14C > T c.1838 + 2 T > C
10 14 18 21 26 28 29
Protein changes
p. p. p. p.
Lys9Pro Lys9Pro Lys207Arg Arg784Gly
Type of mutation
Novelty
Population
Missense
–
South Korea [37], Germany [38], USA [39], Canada [40], Japan [41,42], Australia [43], China [44], Spain [45]
Missense Missense Missense Missense Frameshift Missense Splice site
Yes Yes Yes Yes Yes – Yes
from frequent bone fractures and repeated bone modelling particularly at the long bones [2]. This affects growth and stature of the patients as well as their functional and mobility status. Our study showed that 63.6% (p < .05) of patients had bone deformities at the mean age of 9 years old and had to walk with aid. Previous studies showed that bone deformities were more common in patients with qualitative mutations than quantitative mutations and this is consistent with our results [23]. Similarly, our patients with quantitative mutations also showed independent walking ability as opposed to qualitative mutations that had to walk with aid. Hald et al., speculated that in comparison with qualitative mutations, OI patients with quantitative defects had normal protein structure despite collagen insufficiency in the bone structure [50]. This allow for bone mineralization to occur and only caused fewer fractures than qualitative defects [50]. Blue sclera is one of distinctive clinical feature in OI and commonly observed in patients with OI type I [51]. In patients with OI type III and IV, blue sclera may appear at birth, however the bluish colour fades with increasing age [52]. The mean age of OI patients in this study was relatively young (9 years old, ± 5.66), hence the presence of blue sclera in all the patients. DI is also another key feature in OI and has been identified to occur frequently in OI type III and less common in type I [53,54]. Our results were consistent with this finding, in which 71.4% of OI type III were present with DI, followed by 14.3% of type I and IV respectively. Similarly, Lindahl et al., also reported that DI was more common in severe groups (OI type III and IV) compared to mild OI (type I) caused by qualitative defects in collagen chains [10]. Those with qualitative mutations were at higher risk of having conspicuous signs of DI because the structurally abnormal collagen affects tooth germ development in predentin during mineralization process [55]. In addition, our study also showed 81% of patients with DI also had bone deformities. In this study, most qualitative mutations involve glycine
Table 5 Relationship between clinical characteristics and types of mutation. Characteristics
Quantitative mutation (n = 6)
Qualitative mutation (n = 12)
p value
OI type (I/III/IV) COL1A1/COL1A2 Gender (Male/female) Blue sclera Dentinogenesis imperfecta Hearing loss Bone deformity Walk with aid
4/0/2 6/0 4/2 33.3% 28.6%
1/9/2 8/4 7/5 66.7% 71.4%
0.003 0.245 1.000 – 0.569
100% – –
– 100% 100%
0.33 < 0.001 0.114
mutations in COL1A1 occurred three times more frequent than COL1A2 [20,23]. Mutations affecting α1 (1) are presumed to have more severe clinical features, because collagen type I consists of triple helixes made of two α1 (1) chains and one α1 (2) chain [7]. In this study, however, with the exception of walking ability, mutations in either chains showed no difference in clinical characteristics as reported by Rauch et al. [49]. The difference in walking ability between α1 (1) and α1 (2) mutation groups occurred owing to the difference in walking ability between quantitative and qualitative mutations. We found that quantitative mutations only occurred in α1 (1), whereas qualitative mutations involved both of the α1 chains. Our results are consistent with those of previous studies on Swedish and Canadian populations [10,49]. In comparison with quantitative mutations, qualitative mutations were reported in both OI type III and IV while none occurred in OI type I. This finding indicates that patients with qualitative mutations had more severe phenotypes. Bone deformities in OI often occurred at birth and usually resulted
Fig. 1. Mutations distribution in α1 chains and their relationship with the presence of dentinogenesis imperfecta (DI) and bone deformity (BD). 144
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
substitution within the triplet Gly-X-Y sequence of α1 helical domain. Out of 10 glycine mutations, 70% (n = 7) had glycine to serine substitution and 10% (n = 1) had substitution to aspartic acid, arginine and valine respectively. Our results are concordant with a study by an OI consortium that reported helical mutations occurred more in COL1A1 than COL1A2 [18]. The mutation consortium also reported that glycine was frequently substituted to serine in both COL1A1 (38.9%) and COL1A2 (44%) [18]. Furthermore, previous studies have found that apart from serine, substitution to cysteine, aspartic acid, valine and arginine were also commonly detected [20,21,49]. However, the clinical characteristics of our patients remain similar despite the difference in the substituting amino acid in α1 (1) or α1 (2). Rauch et al. and Lindahl et al., however, showed that serine substitution in α1 (1) had more severe phenotypes compared to α1 (2) [10,49]. Our findings also support earlier observations that nonlethal substitutions by branched or charged amino acid residues (aspartic acid, valine, arginine) were frequently identified at the amino-terminal portion of α1 (1) helical domain, whereas substitutions in other parts of the chain resulted in lethal outcomes [18,26]. In addition, substitutions by branched or charged amino acid within amino terminal were reported to have milder phenotypic effects, as indicated in our study that patient with arginine and valine substitution were able to walk independently than those with serine substitutions [49]. Theoretically, glycine substitutions towards the carboxyl terminal of the α1 chains were clinically more severe than amino terminal [56]. This is consistent with the type I collagen protein model that formation of helix conformation begins from carbon terminal and move towards the amino terminal end [57]. Mutation in the carboxyl terminal will be more disruptive to collagen folding mechanism than mutations in the amino terminal [56]. Previous studies demonstrated that glycine substitution within amino terminal of α1 (1) and α1 (2) chains had negative correlation with DI, whereas substitution towards carboxyl terminal were present with DI [10,49]. In contrast, this study found that patient with glycine substitution within amino terminal of α1 (1) had no presence of DI, while those with mutations in carboxyl terminal were DI positive. However, no such association was found in α1 (2) and this is probably due to insufficient sample size. Nevertheless, patients sharing the same mutation in α1 (1) were concordant for this phenotype but not in α1 (2). This suggests that the position of glycine mutations is predictive of DI outcome. Our results are similar with the mutation report by Marini et al. that described some mutations between α1 (1) and α1 (2) regions had lethal consequences [18]. However, one patient with OI type III in our cohort had mutations in one of the lethal clusters in α1 (2). Hence, we implied that although these lethal regions were considerably useful to determine the function of different regions in α1 chains, it should be noted that mutations within these areas may even be suited for long term survival or mild phenotype. One of our patients was clinically diagnosed as OI type V and revealed to have c.-14C > T mutation in IFITM5. Previous studies identified this mutation to cause autosomal dominant OI type V, which has distinct clinical feature of hyperplastic callus [37,38]. However, one other patient with OI type III in our study also had the same mutation. This is similar to earlier report that identified mutation c.14C > T in patients without typical features of OI type V [6]. Lazarus et al. in their study found that this mutation caused the addition of five amino acids to the N-terminal and consequently produced longer mRNA with increased expression of Bril [43]. In addition to the recurrence of c.-14C > T mutations in IFITM5, we also identified novel mutation c.1_0insC in the same gene. Based on our analysis using Human Splicing Finder version 3.0 (http://www.umd.be/HSF3/HSF.shtml), mutation c.-1_0insC was predicted to introduce a new branch point during mRNA splicing, however no alternative splicing effect was speculated. Apart from collagen genes, our patients also had novel mutations in genes that were previously known to cause autosomal recessive OI.
Mutations in P3H1 were identified in 3 of our patients with no familial relationship. Two of them shared the same mutation and both had moderate and severe phenotypes. Mutation p.L9P is the first missense mutation identified within position 1 to 21 of P3H1 signalling domain. Previous studies described mutations in P3H1 as the causative mutation for OI type VIII [58]. Based on OI mutations database, most of these mutations were nonsense, frameshift and splice site mutations that introduced premature termination codon, thus removal of affected mRNA through nonsense mediated decay (NMD) [59]. These activities lead to truncating effects on P3H1 protein or the loss of P3H1 functional domain [59]. On the other hand, our study also determined mutation in BMP1 in OI type III patient and to the best of our knowledge, mutation p.R784G is the first mutation identified in exon 17 of BMP1 gene. Mutations in BMP1 delayed the processing of type I procollagen Cterminal propeptide, hence disrupt the formation of mature collagen fibrils [60]. Nevertheless, Sangsin et al. reported that mutation in BMP1 displayed wide clinical severity, of which depends on the affected protein [61]. He further postulated that mutations involving both BMP1 and mTLD activities will have more severe phenotypes compared to BMP1 alone [61]. Our study also found a homozygous mutation (p.L207R) in SERPINF1 which has been reported previously in dbSNP database (https://www.ncbi.nlm.nih.gov/projects/SNP/), however the clinical significance of this mutations was not yet available. Nevertheless, our in silico analysis predicted this mutation to have protein damaging effects and in fact, p.L207R was the first missense mutation identified in SERPINF1 gene in the OI mutations database. Previously, reports on mutations in SERPINF1 were mostly frameshift and nonsense mutation that resulted in OI type VI [62,63]. These mutations not only caused the loss of PEDF protein function, but also impairs the production of circulating PEDF [64]. Recently, researchers suggested that the concentration of PEDF in serum to be a potential marker to evaluate patients with OI type VI [65]. Despite sufficient sequencing coverage of 400× for each sample, we failed to identify mutations in 3 of our patients. This probably occurred because the patients carry mutations that were not targeted in this study. For instance, MBTPS2, a gene that encodes S2P, a protein enzyme responsible for cleavage of intramembrane proteins and associated with X-linked OI was not included during the time of our panel design [66]. Those patients with undetected mutations also may carry large gene deletions that we incapable to identify using current techniques. Other possible reason is due to poor coverage in certain amplicons due to high GC content. Hence, one way to overcome these limitations is by using other techniques such as multiplex-ligation dependent probe amplification (MLPA) or comparative genome hybridization [67].
5. Conclusion We have successfully identified 22 variants including 9 novel variants in 26 OI patients. Overall, our mutation results are similar to other OI populations, with majority in our cohort showing mutations in collagen genes. Genotype-phenotype analysis confirmed earlier findings that qualitative mutations had more severe clinical characteristics compared to quantitative mutations. Presence of DI also associated with position of glycine mutation concerning C– or N-terminal portion of type 1 collagen chains. Although small in numbers, we have identified a few patients with mutations in non-collagen genes that displayed either moderate or severe phenotypes.
Competing interest The authors declare that they have no competing interest.
145
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
Acknowledgment
(1997) 711–718. [20] Z.-L. Zhang, H. Zhang, Y.-H. Ke, H. Yue, W.-J. Xiao, J.-B. Yu, et al., The identification of novel mutations in COL1A1, COL1A2, and LEPRE1 genes in Chinese patients with osteogenesis imperfecta, J. Bone Miner. Metab. 30 (1) (2012) 69–77. [21] B.H. Duy, L. Zhytnik, K. Maasalu, I. Kändla, E. Prans, E. Reimann, et al., Mutation analysis of the COL1A1 and COL1A2 genes in Vietnamese patients with osteogenesis imperfecta, Human Genom. 10 (1) (2016) 1. [22] K.-S. Lee, H.-R. Song, T.-J. Cho, H.J. Kim, T.-M. Lee, H.-S. Jin, et al., Mutational spectrum of type I collagen genes in Korean patients with osteogenesis imperfecta, Hum. Mutat. 27 (6) (2006) 599. [23] H.-Y. Lin, C.-K. Chuang, Y.-N. Su, M.-R. Chen, H.-C. Chiu, D.-M. Niu, et al., Genotype and phenotype analysis of Taiwanese patients with osteogenesis imperfecta, Orphanet J. Rare Dis. 10 (2015) 152. [24] A. Fuccio, M. Lorio, F. Amato, A. Elce, R. Ingino, M. Filocamo, et al., A novel DHPLC-based procedure for the analysis of COL1A1 and COL1A2 mutations in osteogenesis imperfecta, J. Mol. Diagnostics 13 (6) (2011) 648–656. [25] G. Venturi, E. Tedeschi, M. Mottes, M. Valli, M. Camilot, S. Viglio, et al., Osteogenesis imperfecta: clinical, biochemical and molecular findings, Clin. Genet. 70 (2) (2006) 131–139. [26] D.L. Bodian, T.-F. Chan, A. Poon, U. Schwarze, K. Yang, P.H. Byers, et al., Mutation and polymorphism spectrum in osteogenesis imperfecta type II: implications for genotype-phenotype relationships, Hum. Mol. Genet. 18 (3) (2009) 463–471. [27] A.E. Konstantinidou, G. Agrogiannis, S. Sifakis, A. Karantanas, V. Harakoglou, P. Kaminopetros, et al., Genetic skeletal disorders of the fetus and infant: pathologic and molecular findings in a series of 41 cases, Birth Defects Res. Part A 85 (10) (2009) 811–821. [28] F. Rauch, L. Lalic, F.H. Glorieux, P. Moffatt, P. Roughley, Targeted sequencing of a pediatric metabolic bone gene panel using a desktop semiconductor next-generation sequencer, Calcif. Tissue Int. 95 (4) (2014) 323–331. [29] J.A. Caparros-Martin, M.S. Aglan, S. Temtamy, G.A. Otaify, M. Valencia, J. Nevado, et al., Molecular spectrum and differential diagnosis in patients referred with sporadic or autosomal recessive osteogenesis imperfecta, Mol. Genet. & Genomic Med. 5 (1) (2017) 28–39. [30] J.F. Bateman, I. Moeller, M. Hannagan, D. Chan, G.C. William, Characterization of three osteogenesis imperfecta collagen alpha 1(I) glycine to serine mutations demonstrating a position-dependent gradient of phenotypic severity, Biochem. J. 288 (Pt 1) (1992) 131–135. [31] K. Mackay, P.H. Byers, R. Dalgleish, An RT-PCR-SSCP screening strategy for detection of mutations in the gene encoding the α1 chain of type I collagen: application to four patients with osteogenesis imperfecta, Hum. Mol. Genet. 2 (8) (1993) 1155–1160. [32] J. Stephen, A. Shukla, A. Dalal, K.M. Girisha, H. Shah, N. Gupta, et al., Mutation spectrum of COL1A1 and COL1A2 genes in Indian patients with osteogenesis imperfecta, Am. J. Med. Genet. A 164 (6) (2014) 1482–1489. [33] M. Balasubramanian, G.J. Sobey, B.E. Wagner, L.C. Peres, J. Bowen, J. Bexon, et al., Osteogenesis imperfecta: ultrastructural and histological findings on examination of skin revealing novel insights into genotype-phenotype correlation, Ultrastruct. Pathol. 40 (2) (2016) 71–76. [34] J. Witecka, M. Aleksandra, Kruczek A. Auguściak-Duma, A. Szydlo, M. Lesiak, et al., Two novel COL1A1 mutations in patients with osteogenesis imperfecta (OI) affect the stability of the collagen type I triple-helix, J. Appl. Genet. 49 (3) (2008) 283–295. [35] A. Forlino, E. D'Amatto, M. Valli, G. Camera, E. Hopkins, J.C. Marini, et al., Phenotypic comparison of an osteogenesis imperfecta type IV Proband with ade novo α2(I) Gly922 → Ser substitution in type I collagen and an unrelated patient with an identical mutation, Biochem. Mol. Med. 62 (1) (1997) 26–35. [36] H. Hartikka, K. Kuurila, J. Körkkö, I. Kaitila, R. Grénman, S. Pynnonen, et al., Lack of correlation between the type of COL1A1 or COL1A2 mutation and hearing loss in osteogenesis imperfecta patients, Hum. Mutat. 24 (2) (2004) 147–154. [37] T.-J. Cho, K.-E. Lee, S.-K. Lee, S.J. Song, K.J. Kim, D. Jeon, et al., A single recurrent mutation in the 5′-UTR of IFITM5 causes osteogenesis imperfecta type V, Am. J. Hum. Genet. 91 (2) (2012) 343–348. [38] O. Semler, L. Garbes, K. Keupp, D. Swan, K. Zimmermann, J. Becker, et al., A mutation in the 5′-UTR of IFITM5 creates an in-frame start codon and causes autosomal-dominant osteogenesis imperfecta type V with hyperplastic callus, Am. J. Hum. Genet. 91 (2) (2012) 349–357. [39] M. Grover, P.M. Campeau, C.D. Lietman, J.T. Lu, R.A. Gibbs, A.E. Schlesinger, et al., Osteogenesis imperfecta without features of type V caused by a mutation in the IFITM5 gene, J. Bone Miner. Res. 28 (11) (2013) 2333–2337. [40] F. Rauch, P. Moffatt, M. Cheung, P. Roughley, L. Lalic, A.M. Lund, et al., Osteogenesis imperfecta type V: marked phenotypic variability despite the presence of the IFITM5 c.− 14C > T mutation in all patients, J. Med. Genet. 50 (1) (2013) 21–24. [41] O.-H. Kim, D.-K. Jin, K. Kosaki, J.-W. Kim, S.-Y. Cho, W.J. Yoo, et al., Osteogenesis imperfecta type V: clinical and radiographic manifestations in mutation confirmed patients, Am. J. Med. Genet. A 161 (8) (2013) 1972–1979. [42] M. Takagi, S. Sato, K. Hara, C. Tani, O. Miyazaki, G. Nishimura, et al., A recurrent mutation in the 5’-UTR of IFITM5 causes osteogenesis imperfecta type V, Am. J. Med. Genet. A (2013), http://dx.doi.org/10.1002/ajmg.a.36025. [43] S. Lazarus, A.M. Mclnerney-Leo, F.A. McKenzie, G. Baynam, S. Broley, B.V. Cavan, et al., The IFITM5 mutation c.-14C > T results in an elongated transcript expressed in human bone; and causes varying phenotypic severity of osteogenesis imperfecta type V, BMC Musculoskelet. Disord. 15 (1) (2014) 107. [44] Z. Zhang, M. Li, J.-W. He, W.-Z. Fu, C.-Q. Zhang, Z.-L. Zhang, Phenotype and genotype analysis of Chinese patients with osteogenesis imperfecta type V, PLoS One (2013), http://dx.doi.org/10.1371/journal.pone.0072337.
We wish to thank the patients and their family for participating in this study. We thank the Ministry of Higher Education of Malaysia for the support of funding. Funding This study was supported by research grant from Ministry of Higher Education of Malaysia, (FRGS/1/2015/SKK/08/UKM/02/1). Author's contributions NMN was responsible in conducting the experiment and writing the manuscript. NMS and RR participated in samples and patient's clinical data collection. SZ, ZAL and BB contributed in development of protocol and analytical framework. Together with NAAM and RJ, they supervised the study design and execution, as well as funding attainment. CYP helped NMN for the data analysis of targeted sequencing. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.cca.2018.05.048. References [1] J.C. Marini, A. Forlino, H.P. Bächinger, N.J. Bishop, P.H. Byers, A.D. Paepe, et al., Osteogenesis imperfecta, Nat. Rev. Disease Primers. 3 (2017) 17052. [2] F. Van Dijk, D. Sillence, Osteogenesis imperfecta: clinical diagnosis, nomenclature and severity assessment, Am. J. Med. Genet. A 164 (6) (2014) 1470–1481. [3] J.L. Shaker, C. Albert, J. Fritz, G. Harris, Recent developments in osteogenesis imperfecta, F1000Res. (2015), http://dx.doi.org/10.12688/f1000research.6398.1. [4] A. Forlino, W.A. Cabral, A.M. Barnes, J.C. Marini, New perspectives on osteogenesis imperfecta, Nat. Rev. Endocrinol. 7 (9) (2011) 540–557. [5] D. Sillence, A. Senn, D. Danks, Genetic heterogeneity in osteogenesis imperfecta, J. Med. Genet. 16 (2) (1979) 101–116. [6] F.H. Glorieux, F. Rauch, H. Plotkin, L. Ward, R. Traverse, P. Roughley, et al., Type V osteogenesis imperfecta: a new form of brittle bone disease, J. Bone Miner. Res. 15 (9) (2000) 1650–1658. [7] D.J. Prockop, C.D. Constantinou, K.E. Dombrowski, Y. Hojima, K.E. Kadler, H. Kuivaniemi, et al., Type I procollagen: the gene-protein system that harbors most of the mutations causing osteogenesis imperfecta and probably more common heritable disorders of connective tissue, Am. J. Med. Genet. A 34 (1) (1989) 60–67. [8] E.R. Valadares, T.B. Carneiro, P.M. Santos, A.C. Oliveira, B. Zabel, What is new in genetics and osteogenesis imperfecta classification? J. Pediatr. 90 (6) (2014) 536–541. [9] H. Kang, J.C. Marini, Osteogenesis imperfecta: new genes reveal novel mechanisms in bone dysplasia, Transl. Res. 181 (2017) 27–48. [10] K. Lindahl, E. Astrom, C.J. Rubin, G. Grigelioniene, B. Malmgren, O. Ljunggren, et al., Genetic epidemiology, prevalence, and genotype-phenotype correlations in the Swedish population with osteogenesis imperfecta, Eur. J. Hum. Genet. 23 (8) (2015) 1042–1050. [11] E.E. Bleck, Nonoperative treatment of osteogenesis imperfecta: orthotic and mobility management, Clin. Orthop. Relat. Res. (159) (1981) 111–122. [12] K. Wang, M. Li, H. Hakonarson, ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data, Nucleic Acids Res. (2010), http://dx. doi.org/10.1093/nar/gkq603. [13] P. Kumar, S. Henikoff, P.C. Ng, Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm, Nat. Protoc. 4 (2009) 1073. [14] I. Adzhubei, D.M. Jordan, S.R. Sunyaev, Predicting functional effect of human missense mutations using PolyPhen-2, Current Protocols Hum. Genet. (2013), http://dx.doi.org/10.1002/0471142905.hg0720s76. [15] J.M. Schwarz, D.N. Cooper, M. Schuelke, D. Seelow, MutationTaster2: mutation prediction for the deep-sequencing age, Nat. Methods 11 (2014) 361. [16] J.T. Robinson, H. Thorvaldsdóttir, W. Winckler, M. Guttman, E.S. Lander, G. Getz, J.P. Mesirov, Integrative genomics viewer, Nat. Biotechnol. 29 (2011) 24. [17] H. Thorvaldsdóttir, J.T. Robinson, J.P. Mesirov, Integrative genomics viewer (IGV): high-performance genomics data visualization and exploration, Brief. Bioinform. 14 (2) (2013) 178–192. [18] J.C. Marini, A. Forlino, W.A. Cabral, A.M. Barnes, J.D. San Antonio, S. Milgrom, J.C. Hyland, et al., Consortium for osteogenesis imperfecta mutations in the helical domain of type I collagen: regions rich in lethal mutations align with collagen binding sites for integrins and proteoglycans, Hum. Mutat. 28 (3) (2007) 209–221. [19] A.M. Lund, A.C. Nicholls, M. Schwartz, S. Skovby, Parental mosaicism and autosomal dominant mutations causing structural abnormalities of collagen I are frequent in families with osteogenesis imperfecta type III/IV, Acta Paediatr. 86 (7)
146
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
mutations in the collagen type I genes, Oral Dis. 23 (1) (2017) 42–49. [56] P.H. Byers, Brittle bones - fragile molecules: disorders of collagen gene structure and expression, Trends Genet. 6 (1990) 293–299. [57] J.A. Engel, D.J. Prockop, The zipper-like folding of collagen triple helices and the effects of mutations that disrupt the zipper, Annu. Rev. Biophys. Biophys. Chem. 20 (1) (1991) 137–152. [58] J.C. Marini, W.A. Cabral, M.A. Barnes, Null mutations in LEPRE1 and CRTAP cause severe recessive osteogenesis imperfecta, Cell Tissue Res. 339 (1) (2010) 59–70. [59] Y. Huang, L. Mei, W. Lv, H. Li, R. Zhang, Q. Pan, et al., Targeted exome sequencing identifies novel compound heterozygous mutations in P3H1 in a fetus with osteogenesis imperfecta type VIII, Clin. Chim. Acta 464 (2017) 170–175. [60] R.J. Fernandes, S. Hirohata, J.M. Engle, A. Colige, D.H. Cohn, D.R. Eyre, et al., Procollagen II amino Propeptide processing by ADAMTS-3: insights on Dermatosparaxis, J. Biol. Chem. 276 (34) (2001) 31502–31509. [61] A. Sangsin, C. Kuptanon, C. Srichomthong, M. Pongpanich, K. Suphapeetiporn, V. Shotelersuk, Two novel compound heterozygous BMP1 mutations in a patient with osteogenesis imperfecta: a case report, BMC Medical Genet. 18 (1) (2017) 25. [62] H. Al-Jallad, T. Palomo, P. Roughley, F.H. Glorieux, M.D. McKee, P. Moffatt, et al., The effect of SERPINF1 in-frame mutations in osteogenesis imperfecta type VI, Bone 76 (Suppl C) (2015) 115–120. [63] J. Becker, O. Semler, C. Gilissen, Y. Li, H.J. Bolz, C. Giunta, et al., Exome sequencing identifies truncating mutations in human SERPINF1 in autosomal-recessive osteogenesis imperfecta, Am. J. Hum. Genet. 88 (3) (2011) 362–371. [64] J.-Y. Wang, Y. Liu, L.-J. Song, F. Lv, X.-J. Xu, A. San, et al., Novel mutations in SERPINF1 result in rare osteogenesis imperfecta type VI, Calcif. Tissue Int. (2016), http://dx.doi.org/10.1007/s00223-016-0201-z. [65] F. Rauch, A. Husseini, P. Roughley, F.H. Glorieux, P. Moffatt, Lack of circulating pigment epithelium-derived factor is a marker of osteogenesis imperfecta type VI, J. Clin. Endocrinol. & Metab. 97 (8) (2012) E1550–E1556. [66] U. Lindert, W.A. Cabral, S. Ausavarat, S. Tongkobpetch, K. Ludin, A.M. Barnes, et al., MBTPS2 mutations cause defective regulated intramembrane proteolysis in Xlinked osteogenesis imperfecta, Nat. Commun. (2016), http://dx.doi.org/10.1038/ ncomms11920. [67] Y. Liu, Ma D. Asan, F. Lv, X. Xu, J. Wang, et al., Gene mutation spectrum and genotype-phenotype correlation in a cohort of Chinese osteogenesis imperfecta patients revealed by targeted next generation sequencing, Osteoporos. Int. (2017), http://dx.doi.org/10.1007/s00198-017-4143-8.
[45] E. Guillén-Navarro, M.J. Ballesta-Martínez, M. Valencia, A.M. Bueno, V. MartínezGlez, V. Lopez-Gonzalez, et al., Two mutations in IFITM5 causing distinct forms of osteogenesis imperfecta, Am. J. Med. Genet. A 164 (5) (2014) 1136–1142. [46] R.M. Patel, S.C.S. Nagamani, D. Cuthbertson, P.M. Campeau, J.P. Krischer, J.R. Shapiro, et al., A cross-sectional multicenter study of osteogenesis imperfecta in North America - results from the linked clinical research centers, Clin. Genet. 87 (2) (2015) 133–140. [47] O. Essawi, S. Symoens, M. Fannana, M. Darwish, M. Farraj, A. Willaert, et al., Genetic analysis of osteogenesis imperfecta in the Palestinian population: molecular screening of 49 affected families, Mol. Genet.& Genomic Med. (2017), http://dx. doi.org/10.1002/mgg3.331. [48] M. Takagi, H. Shinohara, S. Narumi, G. Nishimura, Y. Hasegawa, T. Hasegawa, Severe osteogenesis imperfecta caused by double glycine substitutions near the amino-terminal triple helical region in COL1A2, Am. J. Med. Genet. A 167 (7) (2015) 1627–1631. [49] F. Rauch, L. Lalic, P. Roughley, F.H. Glorieux, Genotype-phenotype correlations in nonlethal osteogenesis imperfecta caused by mutations in the helical domain of collagen type I, Eur. J. Hum. Genet. 18 (6) (2010) 642–647. [50] J.D. Hald, L. Folkestad, T. Harslof, A.M. Lund, M. Duno, J.B. Jensen, et al., Skeletal phenotypes in adult patients with osteogenesis imperfecta—correlations with COL1A1/COL1A2 genotype and collagen structure, Osteoporos. Int. (2016), http:// dx.doi.org/10.1007/s00198-016-3653-0. [51] G. Bardai, P. Moffatt, F.H. Glorieux, F. Rauch, DNA sequence analysis in 598 individuals with a clinical diagnosis of osteogenesis imperfecta: diagnostic yield and mutation spectrum, Osteoporos. Int. (2016), http://dx.doi.org/10.1007/s00198016-3709-1. [52] D. Sillence, B. Butler, M. Latham, K. Barlow, Natural history of blue sclerae in osteogenesis imperfecta, Am. J. Med. Genet. 45 (2) (1993) 183–186. [53] K. Andersson, G. Dahllof, K. Lindahl, A. Kindmark, G. Grigelioniene, A. Astrom, et al., Mutations in COL1A1 and COL1A2 and dental aberrations in children and adolescents with osteogenesis imperfecta – a retrospective cohort study, PLoS One 12 (5) (2017) e0176466. [54] R. Sæves, L.L. Wekre, E. Ambjornsen, S. Axelsson, H. Nordgarden, K. Storhaug, Oral findings in adults with osteogenesis imperfecta, Special Care in Dentistry 29 (2) (2009) 102–108. [55] B. Malmgren, K. Andersson, K. Lindahl, A. Kindmark, G. Grigelioniene, V. Zachariadis, et al., Tooth agenesis in osteogenesis imperfecta related to
147
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Genotype-phenotype correlation among Malaysian patients with osteogenesis imperfecta
T
Nadiah Mohd Nawawia, Nalini M. Selveindranb, Rahmah Rasatc, Chow Yock Pinga, ⁎ Zarina Abdul Latiffc, Syed Zulkifli Syed Zakariac, Rahman Jamala, Nor Azian Abdul Murada, ,1, ⁎ Bilkis Banu Abd Azizc, ,1 a
UKM Medical Molecular Biology Institute (UMBI), Jalan Yaacob Latiff, Bandar Tun Razak, 56000 Cheras Kuala Lumpur, Malaysia Department of Pediatrics, Hospital Putrajaya, Jalan P9, Federal Government Administration Centre, Precint 7, 62250 Putrajaya, Malaysia c Department of Pediatrics, UKM Medical Centre, Jalan Yaacob Latiff, Bandar Tun Razak, 56000 Cheras Kuala Lumpur, Malaysia b
A R T I C LE I N FO
A B S T R A C T
Keywords: Brittle bone disease Genotype Osteogenesis imperfecta Phenotype
Background: Osteogenesis imperfecta (OI) is a rare genetic bone disease characterized by bone fragility and low bone mass. OI was mainly caused by genetic mutations in collagen genes, COL1A1 and COL1A2. Nevertheless, new genes have been identified to be causally linked to OI. The clinical features between each OI groups share great similarities and it is sometimes difficult for clinicians to diagnose the disease accurately. Here, we identify the genetic mutations of OI patients from Malaysia and correlate the genetic mutations with the clinical features. Method: Targeted sequencing of fourteen genes panel was performed to identify the mutations in 29 OI patients with type I, III, IV and V disease. The mutations were determined using Ion Torrent Suite software version 5 and variant annotation was conducted using ANNOVAR. The identified mutations were confirmed using Sanger sequencing and in silico analysis was performed to evaluate the effects of the candidate mutations at protein level. Results: Majority of patients had mutations in collagen genes, 48% (n = 14) in COL1A1 and 14% (n = 4) in COL1A2. Type I OI was caused by quantitative mutations in COL1A1 whereas most of type III and IV were due to qualitative mutations in both of the collagen genes. Those with quantitative mutations had milder clinical severity compared to qualitative mutations in terms of dentinogenesis imperfecta (DI), bone deformity and the ability to walk with aid. Furthermore, a few patients (28%, n = 8) had mutations in IFITM5, BMP1, P3H1 and SERPINF1. Conclusion: Majority of our OI patients have mutations in collagen genes, similar to other OI populations worldwide. Genotype-phenotype analysis revealed that qualitative mutations had more severe clinical characteristics compared to quantitative mutations. It is crucial to identify the causative mutations and the clinical severity of OI patients may be predicted based on the types of mutations.
1. Introduction Osteogenesis Imperfecta (OI) is a group of genetic bone disorders primarily characterized by low bone mass, which results in increased bone fractures and disrupted growth [1]. Additional extra-skeletal phenotypes include blue sclera, dentinogenesis imperfecta (DI) and hearing impairment [2]. There is a wide spectrum of clinical features and severity among patients, ranging from mild with minimal or no bone deformities to severe with multiple bone fractures and also perinatal lethality [3]. The variable types of OI occurs in approximately 1 in every 15,000–20,000 births with majority of them being autosomal
⁎
1
dominant inheritance [4]. The classic Sillence classification categorized OI into four numerical groups (OI type I to IV) according to its clinical findings and inheritance patterns [5]. Type I is often the mildest form, with normal or near-normal skeletal features. Type II is lethal in the perinatal period, with multiple intrauterine fractures and skeletal deformities of the limbs. Type III is the severe form with frequent fractures and bone deformities that progress with age. Type IV has various phenotypes from mild to moderate skeletal deformities and varies in height and statures [5]. On the other hand, type V OI was mainly identified by the presence of hyperplastic callus and interosseous membrane [6].
Corresponding authors. E-mail addresses: [email protected] (N.A. Abdul Murad), [email protected] (B.B. Abd Aziz). Bilkis Banu Abd Aziz is the Principal Investigator of the grant and Nor Azian Abdul Murad is responsible for student's supervision.
https://doi.org/10.1016/j.cca.2018.05.048 Received 4 April 2018; Accepted 24 May 2018 Available online 25 May 2018 0009-8981/ © 2018 Elsevier B.V. All rights reserved.
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
LEPRE1, PPIB, SERPINH1, SERPINF1, FKBP10, P3H1, BMP1, TMEM38B, IFITM5, PLOD2 and WNT1. Two primer pools to amplify 320 amplicons were designed using Ion AmpliSeq Designer version 4.48 using the ‘standard DNA’ (225-bp amplicon target sizes) and ‘Gene + UTR’ options (https://www.ampliseq.com). The number of amplicons and base covered for each gene are listed in Supplementary Table 1. A total of 10 ng DNA per sample was used for target enrichment by multiplex PCR and barcoded libraries from 16 samples were pooled in a single tube (Thermo Fisher Scientific, USA). Library pools were subjected to Ion Chef ™ instrument (Thermo Fisher Scientific, USA) for template preparation and enriched Ion Sphere Particles (ISPs) were loaded on Ion 316 v2 BC Chip for sequencing. Potentially pathogenic variants were validated using Sanger sequencing. The candidate mutations were also compared against 667 chromosomes with similar age, gender and ethnicity.
The initial assumption was that this disease was caused by a mutation in either COL1A1 or COL1A2 genes which codes for two collagen type I alpha chains, α1 (1) and α1 (2) [7]. However, with the advancement of genetic testing, mutations in non-collagenous genes have been detected such as LEPRE1, PPIB, SERPINH1, FKBPIO, SP7, IFITM5, BMP1, WNT1, TMEM38B, CRTAP and SERPINF1 genes, all of which are linked to the autosomal recessive OI except for IFITM5 mutation [8]. There are two groups of mutational defects in type I collagen. The first is quantitative mutations due to frameshift, nonsense and splice-site mutations, which introduces a premature termination codon in the coding sequence of COL1A1 allele. The second group is qualitative mutations resulted from the synthesis of collagen molecules that has structural abnormalities, most frequently caused by glycine substitution [9]. The phenotypic severity of OI patients depends on the affected gene, position of mutation, substituted amino acid and the final protein product [10]. However, OI has variable phenotypes and to date, the exact correlation between genotype and phenotype is not fully understood. Hence, we aimed to determine the genetic mutations involved in our local OI patients. Subsequently, the phenotype of these patients will be linked to their genotype and the data obtained may be useful in the clinical diagnosis, genetic counselling and prenatal diagnosis for OI patients
2.5. Sequencing data analysis Data from sequencing run were processed using Ion Torrent Suite software (version 5.0.4; Thermo Fisher Scientific, USA) for base calls, read alignments and variant calling using the reference genomic sequence (hg19). Called variants were annotated using ANNOVAR [12] and the variants were filtered out if the minor allele frequency was equal or higher than 1% in 1000 Genomes Project database and National Heart, Lung, and Brain Institute-Exome Sequencing Project (NHLBI-ESP) with 6500 exomes. Pathogenic variants were predicted using SIFT score (< 0.05) [13], PolyPhen2 score (> 0.4) [14], and MutationTaster [15]. The Integrative Genomics Viewer (IGV) software was used for visualization of protein deleterious effects [16,17]. Novel mutations were identified using Osteogenesis Imperfecta and Ehlers Danlos Syndrome database (https://oi.gene.le.ac.uk/home.php).
2. Materials and methods 2.1. Subjects The study was implemented in compliance with the ethical principles formulated in the declaration of Helsinki and was approved by our local Ethics Committee. Patients granted their informed, written consent to participate. Patients were recruited from the UKM Medical Centre (UKKMC) and Hospital Putrajaya. Written informed consent was given by patients, their parents or legal representatives. The inclusion criteria for OI patients include being Malaysian and with the clinical features of OI.
2.6. Data analysisfor genotype-phenotype correlation Statistical analysis was performed using Statistical Package for Social Sciences (SPSS) version 21. Age differences between the patients and normal was calculated using Independent T Test. Qhi-square was used to determine the gender differences between the patients and normal group. Fisher test was carried out to determine the correlation between gender, OI groups and types of colagen with the patients phenotype. P less than 0.05 was considered as significant.
2.2. Clinical characteristics All 29 patients from 28 families which comprise of 16 males and 13 females were recruited. They underwent clinical and physical examination to evaluate their clinical features and all medical histories were recorded. The clinicians classified the patients into four groups (OI-I, OI-III and OI-IV) according to Sillence classification. OI-V was categorized according to description by Glorieux et al. [6]. The key features observed were blue sclera, bone deformities, degree of disability, dentinogenesis imperfecta (DI) and hearing ability. DI was characterized by the formation of translucent and brittle teeth. For patient's degree of disability, Bleck's 5-point scale was used to assess mobility and walking ability [11].
3. Results 3.1. Clinical characteristics Most of our patients (62.1%) were OI type III, followed by 17.2% of type I and IV respectively. One patient was classified as type V. Table 1 shows the relationship between the clinical characteristics of OI patients and types of OI. None of the patients had OI type II. No group differences were found in the prevalence of blue sclera, however bone deformities were significantly prominent in patients with OI type III as
2.3. Sample preparation Total genomic DNA was extracted from EDTA-preserved peripheral blood using Macherey-Nagel Nucleospin Blood QuickPure kit (Macherey-Nagel, Germany. The Gender validation test was performed for each DNA sample as part of quality assessment. Briefly, PCR was performed using SRY and ATL1 primers and assessed by agarose gel electrophoresis.
Table 1 Phenotypic characteristics among OI patients.
2.4. DNA sequencing Targeted sequencing was performed using Ion PGM Semiconductor ™ (Thermo Fisher Scientific, USA) sequencer with a total of 14 genes with 214 exons selected for the gene panel. The selected genes are associated with osteogenesis imperfecta: COL1A1, COL1A2, CRTAP, 142
Characteristics, n (%)/ Types of OI, n (%)
OI type I 5 (17.2)
OI type III 18 (62.2)
OI type IV 5 (17.2)
OI type V1 (3.4)
p value
COL1A1/COL1A2 Gender (Male/Female) Blue sclera DI Hearing loss Bone deformity Walk with aid
5/0 3/2 5 (17.2) 3 (14.3) 1 (50) 1 (4.5) –
6/3 9/9 18 (62.2) 15 (71.4) – 18 (81.8) 12 (80)
3/1 3/2 5 (17.2) 3 (14.3) 1 (50) 2 (9.1) 2 (13.3)
– 0/1 1 (3.4) – – 1 (4.5) 1 (6.7)
0.392 1.000 – 0.145 0.623 < 0.001 0.028
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
Table 2 Mutational analysis of 14 patients in COL1A1. Individual
OI types
Exon number
Nucleotide change
Protein change
Type of mutation
Novelty
Population
OI 1
III
36
c.2461G > A
p. Gly821Ser
Missense
–
OI 4 OI 5 & OI 20
III III
19 44
c.1273G > A c.3226G > A
p. Gly425Ser p. Gly1076Ser
Missense Missense
– –
OI OI OI OI OI OI
I I III III I IV
37 28 24 31 15 23
c.2560 - 1G > A c.1900C > T c.1631delC c.2028 + 1G > A c.994G > A c.1588G > A
p. Gly332Arg p. Gly530Ser
Splice site Nonsense Frameshift Splice site Missense Missense
– Yes Yes – – –
USA [18], Denmark [19], China [20], Vietnam [21], Sweden [10], South Korea [22], Taiwan [23], Italy [24,25], Poland (Unpublished data), Belgium (Unpublished data) USA [18,26], Denmark [19], Taiwan [23], Greece [27] USA [18], Canada [28], Denmark [19], Sweden [10], China [20], Vietnam [21], Egypt [29], Vietnam [21] USA [18]
III III III
31 32 8
c.2029 - 1G > A c.2155G > A c.599G > T
p. Gly719Ser p. Gly200Val
Splice site Missense Missense
Yes – –
6 7 15 16 17 19
OI 23 & OI 24 OI 25 OI 27
p. Glu634X p. Pro544fs
Belgium (Unpublished data) USA [18], Sweden [10], Canada [28] USA [18], Canada (Unpublished data), Denmark (Unpublished data), Australia [30] USA [18,31], Sweden [10], China [20], India [32] UK [33], Poland [34]
opposed to qualitative mutations (p < .001). Those with quantitative mutations also had independent walking ability, whereas those with qualitative mutations had to walk with aid. In addition, those with quantitative mutations showed lower presence of DI compared to patients with qualitative mutations (28.6% vs 71.4%). Besides that, mutations in α1 (1) were more frequently identified and the patients showed independent walking ability compared to α1 (2) (p < .05). The phenotypic variation with relation to the position of mutation was not significant due insufficient sample size. However, in the patient with mutation affecting amino terminal of Gly22 (p.G200 V) in α1 (1) showed no presence of DI (Fig. 1), whereas two patients with mutations at the carboxyl terminal of Gly898 (p.G1076S) were DI positive. Despite the wide gap of mutation distribution in both of α (1) chains, one patient with OI type III had mutation (Gly568) in one of the eight lethal clusters described by Marini et al. [18] and no mutations were listed among the lethal regions of α1 (1).
compared to both type I and IV (p < 0.01). DI was identified in 71.4% of OI type III patients and 14.3% in both of type I and IV respectively. Patients with OI type I had independent walking ability whereas 66.7% of type III and 40% of type IV patients had to walk with aid. Only two patients of OI type I and III had hearing impairment. Table 5 shows correlation between the clinical characteristics and types of mutation. 3.2. Mutational analysis A total of 22 variants were identified in 26 patients and most variants were found in COL1A1 and COL1A2 genes. There were 12 COL1A1 mutations and 4 COL1A2 mutations in which 3 variants were novel in COL1A1 and 3 novel variants in COL1A2 (Tables 2 and 3). Eleven patients had glycine substitution within the triple helix domain. Eight individuals had glycine substitution to serine and one patient had substitution to arginine, aspartic acid and valine respectively. Six patients had quantitative mutations caused by frameshift, stopgain or splice site variant. Six missense mutations have also been detected in IFITM5, BMP1, P3H1 and SERPINF1 genes (Table 4). Five of the variants were novel. All patients were non consanguineous expect for a pair of twins (OI 23 and OI 24) who shared the same splice site mutations in COL1A1. Individual OI 9 was found to have double mutations in COL1A2 (p.R978H and p.G400D). Three patients had no significant mutations among the genes in the panel.
4. Discussion We have identified 16 (73%) mutations in COL1A1 and COL1A2 in 18 OI patients from 17 families. Previous studies showed 51 to 60% of OI patients had mutations in these two genes [21,23,46]. These results support the early statement that majority of OI patients had mutations in collagen genes, which caused the autosomal dominant OI [18]. Mutation p.Gly1027Ser was identified in two of our patients with OI type III. This mutation was postulated to be a hotspot mutation in COL1A1 instead of founder effect, because it has been found as de novo mutation in other populations [10,20,21,29,47]. Previously, Takagi et al., demonstrated a severe OI type II patient due to double glycine substitutions in COL1A2 (p.Gly208Glu and p.Gly235Asp) [48]. In contrast, our patient with double mutations in COL1A2 only had one glycine substitution within the helical domain and was diagnosed as OI type III. Similar with other published reports, our study also showed that
3.3. Genotype-phenotype correlation There was a significant association between the types of mutation in alpha 1 chains with the types of OI (p < .01). Around 80% of patients with OI type I and 50% of OI type IV had quantitative mutations in α1 (1) chains while none occurred in α1 (2) chain. On the other hand, qualitative mutations were identified in all type III and two of type IV patients in both of COL1A1 and COL1A2. Individuals with quantitative mutations had milder clinical severity, with no bone deformities as Table 3 Mutational analysis of 4 patients in COL1A2. Individual
OI types
Exon number
Nucleotide change
Protein change
Type of mutation
Novelty
OI 3 OI 8
III III
33 46
c.1972G > A c.3034G > A
p. Gly1972Ser p. Gly1012Ser
Missense Missense
Yes –
OI 9
III
OI 12
IV
44 22 44
c.2933G > A c.1199G > A c.2933G > A
p. Arg978His p. Gly400Asp p. Arg978His
Missense Missense Missense
Yes Yes Yes
143
Population
USA [18], Sweden [10], South Korea [22], India [32], Vietnam [21], Italy [35], Finland [36]
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
Table 4 Mutational analysis of 8 patients in other genes. Individual
OI types
Genes
Exon number
Nucleotide changes
OI 2
III
IFITM5
UTR
c.-14C > T
OI OI OI OI OI OI OI
IV III III III III V III
P3H1 P3H1 SERPINF1 BMP1 IFITM5 IFITM5 P3H1
1 1 5 17 UTR UTR 13
c.26 T > C c.26 T > C c.620 T > G c.2350C > G c.-1_0insC c.-14C > T c.1838 + 2 T > C
10 14 18 21 26 28 29
Protein changes
p. p. p. p.
Lys9Pro Lys9Pro Lys207Arg Arg784Gly
Type of mutation
Novelty
Population
Missense
–
South Korea [37], Germany [38], USA [39], Canada [40], Japan [41,42], Australia [43], China [44], Spain [45]
Missense Missense Missense Missense Frameshift Missense Splice site
Yes Yes Yes Yes Yes – Yes
from frequent bone fractures and repeated bone modelling particularly at the long bones [2]. This affects growth and stature of the patients as well as their functional and mobility status. Our study showed that 63.6% (p < .05) of patients had bone deformities at the mean age of 9 years old and had to walk with aid. Previous studies showed that bone deformities were more common in patients with qualitative mutations than quantitative mutations and this is consistent with our results [23]. Similarly, our patients with quantitative mutations also showed independent walking ability as opposed to qualitative mutations that had to walk with aid. Hald et al., speculated that in comparison with qualitative mutations, OI patients with quantitative defects had normal protein structure despite collagen insufficiency in the bone structure [50]. This allow for bone mineralization to occur and only caused fewer fractures than qualitative defects [50]. Blue sclera is one of distinctive clinical feature in OI and commonly observed in patients with OI type I [51]. In patients with OI type III and IV, blue sclera may appear at birth, however the bluish colour fades with increasing age [52]. The mean age of OI patients in this study was relatively young (9 years old, ± 5.66), hence the presence of blue sclera in all the patients. DI is also another key feature in OI and has been identified to occur frequently in OI type III and less common in type I [53,54]. Our results were consistent with this finding, in which 71.4% of OI type III were present with DI, followed by 14.3% of type I and IV respectively. Similarly, Lindahl et al., also reported that DI was more common in severe groups (OI type III and IV) compared to mild OI (type I) caused by qualitative defects in collagen chains [10]. Those with qualitative mutations were at higher risk of having conspicuous signs of DI because the structurally abnormal collagen affects tooth germ development in predentin during mineralization process [55]. In addition, our study also showed 81% of patients with DI also had bone deformities. In this study, most qualitative mutations involve glycine
Table 5 Relationship between clinical characteristics and types of mutation. Characteristics
Quantitative mutation (n = 6)
Qualitative mutation (n = 12)
p value
OI type (I/III/IV) COL1A1/COL1A2 Gender (Male/female) Blue sclera Dentinogenesis imperfecta Hearing loss Bone deformity Walk with aid
4/0/2 6/0 4/2 33.3% 28.6%
1/9/2 8/4 7/5 66.7% 71.4%
0.003 0.245 1.000 – 0.569
100% – –
– 100% 100%
0.33 < 0.001 0.114
mutations in COL1A1 occurred three times more frequent than COL1A2 [20,23]. Mutations affecting α1 (1) are presumed to have more severe clinical features, because collagen type I consists of triple helixes made of two α1 (1) chains and one α1 (2) chain [7]. In this study, however, with the exception of walking ability, mutations in either chains showed no difference in clinical characteristics as reported by Rauch et al. [49]. The difference in walking ability between α1 (1) and α1 (2) mutation groups occurred owing to the difference in walking ability between quantitative and qualitative mutations. We found that quantitative mutations only occurred in α1 (1), whereas qualitative mutations involved both of the α1 chains. Our results are consistent with those of previous studies on Swedish and Canadian populations [10,49]. In comparison with quantitative mutations, qualitative mutations were reported in both OI type III and IV while none occurred in OI type I. This finding indicates that patients with qualitative mutations had more severe phenotypes. Bone deformities in OI often occurred at birth and usually resulted
Fig. 1. Mutations distribution in α1 chains and their relationship with the presence of dentinogenesis imperfecta (DI) and bone deformity (BD). 144
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
substitution within the triplet Gly-X-Y sequence of α1 helical domain. Out of 10 glycine mutations, 70% (n = 7) had glycine to serine substitution and 10% (n = 1) had substitution to aspartic acid, arginine and valine respectively. Our results are concordant with a study by an OI consortium that reported helical mutations occurred more in COL1A1 than COL1A2 [18]. The mutation consortium also reported that glycine was frequently substituted to serine in both COL1A1 (38.9%) and COL1A2 (44%) [18]. Furthermore, previous studies have found that apart from serine, substitution to cysteine, aspartic acid, valine and arginine were also commonly detected [20,21,49]. However, the clinical characteristics of our patients remain similar despite the difference in the substituting amino acid in α1 (1) or α1 (2). Rauch et al. and Lindahl et al., however, showed that serine substitution in α1 (1) had more severe phenotypes compared to α1 (2) [10,49]. Our findings also support earlier observations that nonlethal substitutions by branched or charged amino acid residues (aspartic acid, valine, arginine) were frequently identified at the amino-terminal portion of α1 (1) helical domain, whereas substitutions in other parts of the chain resulted in lethal outcomes [18,26]. In addition, substitutions by branched or charged amino acid within amino terminal were reported to have milder phenotypic effects, as indicated in our study that patient with arginine and valine substitution were able to walk independently than those with serine substitutions [49]. Theoretically, glycine substitutions towards the carboxyl terminal of the α1 chains were clinically more severe than amino terminal [56]. This is consistent with the type I collagen protein model that formation of helix conformation begins from carbon terminal and move towards the amino terminal end [57]. Mutation in the carboxyl terminal will be more disruptive to collagen folding mechanism than mutations in the amino terminal [56]. Previous studies demonstrated that glycine substitution within amino terminal of α1 (1) and α1 (2) chains had negative correlation with DI, whereas substitution towards carboxyl terminal were present with DI [10,49]. In contrast, this study found that patient with glycine substitution within amino terminal of α1 (1) had no presence of DI, while those with mutations in carboxyl terminal were DI positive. However, no such association was found in α1 (2) and this is probably due to insufficient sample size. Nevertheless, patients sharing the same mutation in α1 (1) were concordant for this phenotype but not in α1 (2). This suggests that the position of glycine mutations is predictive of DI outcome. Our results are similar with the mutation report by Marini et al. that described some mutations between α1 (1) and α1 (2) regions had lethal consequences [18]. However, one patient with OI type III in our cohort had mutations in one of the lethal clusters in α1 (2). Hence, we implied that although these lethal regions were considerably useful to determine the function of different regions in α1 chains, it should be noted that mutations within these areas may even be suited for long term survival or mild phenotype. One of our patients was clinically diagnosed as OI type V and revealed to have c.-14C > T mutation in IFITM5. Previous studies identified this mutation to cause autosomal dominant OI type V, which has distinct clinical feature of hyperplastic callus [37,38]. However, one other patient with OI type III in our study also had the same mutation. This is similar to earlier report that identified mutation c.14C > T in patients without typical features of OI type V [6]. Lazarus et al. in their study found that this mutation caused the addition of five amino acids to the N-terminal and consequently produced longer mRNA with increased expression of Bril [43]. In addition to the recurrence of c.-14C > T mutations in IFITM5, we also identified novel mutation c.1_0insC in the same gene. Based on our analysis using Human Splicing Finder version 3.0 (http://www.umd.be/HSF3/HSF.shtml), mutation c.-1_0insC was predicted to introduce a new branch point during mRNA splicing, however no alternative splicing effect was speculated. Apart from collagen genes, our patients also had novel mutations in genes that were previously known to cause autosomal recessive OI.
Mutations in P3H1 were identified in 3 of our patients with no familial relationship. Two of them shared the same mutation and both had moderate and severe phenotypes. Mutation p.L9P is the first missense mutation identified within position 1 to 21 of P3H1 signalling domain. Previous studies described mutations in P3H1 as the causative mutation for OI type VIII [58]. Based on OI mutations database, most of these mutations were nonsense, frameshift and splice site mutations that introduced premature termination codon, thus removal of affected mRNA through nonsense mediated decay (NMD) [59]. These activities lead to truncating effects on P3H1 protein or the loss of P3H1 functional domain [59]. On the other hand, our study also determined mutation in BMP1 in OI type III patient and to the best of our knowledge, mutation p.R784G is the first mutation identified in exon 17 of BMP1 gene. Mutations in BMP1 delayed the processing of type I procollagen Cterminal propeptide, hence disrupt the formation of mature collagen fibrils [60]. Nevertheless, Sangsin et al. reported that mutation in BMP1 displayed wide clinical severity, of which depends on the affected protein [61]. He further postulated that mutations involving both BMP1 and mTLD activities will have more severe phenotypes compared to BMP1 alone [61]. Our study also found a homozygous mutation (p.L207R) in SERPINF1 which has been reported previously in dbSNP database (https://www.ncbi.nlm.nih.gov/projects/SNP/), however the clinical significance of this mutations was not yet available. Nevertheless, our in silico analysis predicted this mutation to have protein damaging effects and in fact, p.L207R was the first missense mutation identified in SERPINF1 gene in the OI mutations database. Previously, reports on mutations in SERPINF1 were mostly frameshift and nonsense mutation that resulted in OI type VI [62,63]. These mutations not only caused the loss of PEDF protein function, but also impairs the production of circulating PEDF [64]. Recently, researchers suggested that the concentration of PEDF in serum to be a potential marker to evaluate patients with OI type VI [65]. Despite sufficient sequencing coverage of 400× for each sample, we failed to identify mutations in 3 of our patients. This probably occurred because the patients carry mutations that were not targeted in this study. For instance, MBTPS2, a gene that encodes S2P, a protein enzyme responsible for cleavage of intramembrane proteins and associated with X-linked OI was not included during the time of our panel design [66]. Those patients with undetected mutations also may carry large gene deletions that we incapable to identify using current techniques. Other possible reason is due to poor coverage in certain amplicons due to high GC content. Hence, one way to overcome these limitations is by using other techniques such as multiplex-ligation dependent probe amplification (MLPA) or comparative genome hybridization [67].
5. Conclusion We have successfully identified 22 variants including 9 novel variants in 26 OI patients. Overall, our mutation results are similar to other OI populations, with majority in our cohort showing mutations in collagen genes. Genotype-phenotype analysis confirmed earlier findings that qualitative mutations had more severe clinical characteristics compared to quantitative mutations. Presence of DI also associated with position of glycine mutation concerning C– or N-terminal portion of type 1 collagen chains. Although small in numbers, we have identified a few patients with mutations in non-collagen genes that displayed either moderate or severe phenotypes.
Competing interest The authors declare that they have no competing interest.
145
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
Acknowledgment
(1997) 711–718. [20] Z.-L. Zhang, H. Zhang, Y.-H. Ke, H. Yue, W.-J. Xiao, J.-B. Yu, et al., The identification of novel mutations in COL1A1, COL1A2, and LEPRE1 genes in Chinese patients with osteogenesis imperfecta, J. Bone Miner. Metab. 30 (1) (2012) 69–77. [21] B.H. Duy, L. Zhytnik, K. Maasalu, I. Kändla, E. Prans, E. Reimann, et al., Mutation analysis of the COL1A1 and COL1A2 genes in Vietnamese patients with osteogenesis imperfecta, Human Genom. 10 (1) (2016) 1. [22] K.-S. Lee, H.-R. Song, T.-J. Cho, H.J. Kim, T.-M. Lee, H.-S. Jin, et al., Mutational spectrum of type I collagen genes in Korean patients with osteogenesis imperfecta, Hum. Mutat. 27 (6) (2006) 599. [23] H.-Y. Lin, C.-K. Chuang, Y.-N. Su, M.-R. Chen, H.-C. Chiu, D.-M. Niu, et al., Genotype and phenotype analysis of Taiwanese patients with osteogenesis imperfecta, Orphanet J. Rare Dis. 10 (2015) 152. [24] A. Fuccio, M. Lorio, F. Amato, A. Elce, R. Ingino, M. Filocamo, et al., A novel DHPLC-based procedure for the analysis of COL1A1 and COL1A2 mutations in osteogenesis imperfecta, J. Mol. Diagnostics 13 (6) (2011) 648–656. [25] G. Venturi, E. Tedeschi, M. Mottes, M. Valli, M. Camilot, S. Viglio, et al., Osteogenesis imperfecta: clinical, biochemical and molecular findings, Clin. Genet. 70 (2) (2006) 131–139. [26] D.L. Bodian, T.-F. Chan, A. Poon, U. Schwarze, K. Yang, P.H. Byers, et al., Mutation and polymorphism spectrum in osteogenesis imperfecta type II: implications for genotype-phenotype relationships, Hum. Mol. Genet. 18 (3) (2009) 463–471. [27] A.E. Konstantinidou, G. Agrogiannis, S. Sifakis, A. Karantanas, V. Harakoglou, P. Kaminopetros, et al., Genetic skeletal disorders of the fetus and infant: pathologic and molecular findings in a series of 41 cases, Birth Defects Res. Part A 85 (10) (2009) 811–821. [28] F. Rauch, L. Lalic, F.H. Glorieux, P. Moffatt, P. Roughley, Targeted sequencing of a pediatric metabolic bone gene panel using a desktop semiconductor next-generation sequencer, Calcif. Tissue Int. 95 (4) (2014) 323–331. [29] J.A. Caparros-Martin, M.S. Aglan, S. Temtamy, G.A. Otaify, M. Valencia, J. Nevado, et al., Molecular spectrum and differential diagnosis in patients referred with sporadic or autosomal recessive osteogenesis imperfecta, Mol. Genet. & Genomic Med. 5 (1) (2017) 28–39. [30] J.F. Bateman, I. Moeller, M. Hannagan, D. Chan, G.C. William, Characterization of three osteogenesis imperfecta collagen alpha 1(I) glycine to serine mutations demonstrating a position-dependent gradient of phenotypic severity, Biochem. J. 288 (Pt 1) (1992) 131–135. [31] K. Mackay, P.H. Byers, R. Dalgleish, An RT-PCR-SSCP screening strategy for detection of mutations in the gene encoding the α1 chain of type I collagen: application to four patients with osteogenesis imperfecta, Hum. Mol. Genet. 2 (8) (1993) 1155–1160. [32] J. Stephen, A. Shukla, A. Dalal, K.M. Girisha, H. Shah, N. Gupta, et al., Mutation spectrum of COL1A1 and COL1A2 genes in Indian patients with osteogenesis imperfecta, Am. J. Med. Genet. A 164 (6) (2014) 1482–1489. [33] M. Balasubramanian, G.J. Sobey, B.E. Wagner, L.C. Peres, J. Bowen, J. Bexon, et al., Osteogenesis imperfecta: ultrastructural and histological findings on examination of skin revealing novel insights into genotype-phenotype correlation, Ultrastruct. Pathol. 40 (2) (2016) 71–76. [34] J. Witecka, M. Aleksandra, Kruczek A. Auguściak-Duma, A. Szydlo, M. Lesiak, et al., Two novel COL1A1 mutations in patients with osteogenesis imperfecta (OI) affect the stability of the collagen type I triple-helix, J. Appl. Genet. 49 (3) (2008) 283–295. [35] A. Forlino, E. D'Amatto, M. Valli, G. Camera, E. Hopkins, J.C. Marini, et al., Phenotypic comparison of an osteogenesis imperfecta type IV Proband with ade novo α2(I) Gly922 → Ser substitution in type I collagen and an unrelated patient with an identical mutation, Biochem. Mol. Med. 62 (1) (1997) 26–35. [36] H. Hartikka, K. Kuurila, J. Körkkö, I. Kaitila, R. Grénman, S. Pynnonen, et al., Lack of correlation between the type of COL1A1 or COL1A2 mutation and hearing loss in osteogenesis imperfecta patients, Hum. Mutat. 24 (2) (2004) 147–154. [37] T.-J. Cho, K.-E. Lee, S.-K. Lee, S.J. Song, K.J. Kim, D. Jeon, et al., A single recurrent mutation in the 5′-UTR of IFITM5 causes osteogenesis imperfecta type V, Am. J. Hum. Genet. 91 (2) (2012) 343–348. [38] O. Semler, L. Garbes, K. Keupp, D. Swan, K. Zimmermann, J. Becker, et al., A mutation in the 5′-UTR of IFITM5 creates an in-frame start codon and causes autosomal-dominant osteogenesis imperfecta type V with hyperplastic callus, Am. J. Hum. Genet. 91 (2) (2012) 349–357. [39] M. Grover, P.M. Campeau, C.D. Lietman, J.T. Lu, R.A. Gibbs, A.E. Schlesinger, et al., Osteogenesis imperfecta without features of type V caused by a mutation in the IFITM5 gene, J. Bone Miner. Res. 28 (11) (2013) 2333–2337. [40] F. Rauch, P. Moffatt, M. Cheung, P. Roughley, L. Lalic, A.M. Lund, et al., Osteogenesis imperfecta type V: marked phenotypic variability despite the presence of the IFITM5 c.− 14C > T mutation in all patients, J. Med. Genet. 50 (1) (2013) 21–24. [41] O.-H. Kim, D.-K. Jin, K. Kosaki, J.-W. Kim, S.-Y. Cho, W.J. Yoo, et al., Osteogenesis imperfecta type V: clinical and radiographic manifestations in mutation confirmed patients, Am. J. Med. Genet. A 161 (8) (2013) 1972–1979. [42] M. Takagi, S. Sato, K. Hara, C. Tani, O. Miyazaki, G. Nishimura, et al., A recurrent mutation in the 5’-UTR of IFITM5 causes osteogenesis imperfecta type V, Am. J. Med. Genet. A (2013), http://dx.doi.org/10.1002/ajmg.a.36025. [43] S. Lazarus, A.M. Mclnerney-Leo, F.A. McKenzie, G. Baynam, S. Broley, B.V. Cavan, et al., The IFITM5 mutation c.-14C > T results in an elongated transcript expressed in human bone; and causes varying phenotypic severity of osteogenesis imperfecta type V, BMC Musculoskelet. Disord. 15 (1) (2014) 107. [44] Z. Zhang, M. Li, J.-W. He, W.-Z. Fu, C.-Q. Zhang, Z.-L. Zhang, Phenotype and genotype analysis of Chinese patients with osteogenesis imperfecta type V, PLoS One (2013), http://dx.doi.org/10.1371/journal.pone.0072337.
We wish to thank the patients and their family for participating in this study. We thank the Ministry of Higher Education of Malaysia for the support of funding. Funding This study was supported by research grant from Ministry of Higher Education of Malaysia, (FRGS/1/2015/SKK/08/UKM/02/1). Author's contributions NMN was responsible in conducting the experiment and writing the manuscript. NMS and RR participated in samples and patient's clinical data collection. SZ, ZAL and BB contributed in development of protocol and analytical framework. Together with NAAM and RJ, they supervised the study design and execution, as well as funding attainment. CYP helped NMN for the data analysis of targeted sequencing. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.cca.2018.05.048. References [1] J.C. Marini, A. Forlino, H.P. Bächinger, N.J. Bishop, P.H. Byers, A.D. Paepe, et al., Osteogenesis imperfecta, Nat. Rev. Disease Primers. 3 (2017) 17052. [2] F. Van Dijk, D. Sillence, Osteogenesis imperfecta: clinical diagnosis, nomenclature and severity assessment, Am. J. Med. Genet. A 164 (6) (2014) 1470–1481. [3] J.L. Shaker, C. Albert, J. Fritz, G. Harris, Recent developments in osteogenesis imperfecta, F1000Res. (2015), http://dx.doi.org/10.12688/f1000research.6398.1. [4] A. Forlino, W.A. Cabral, A.M. Barnes, J.C. Marini, New perspectives on osteogenesis imperfecta, Nat. Rev. Endocrinol. 7 (9) (2011) 540–557. [5] D. Sillence, A. Senn, D. Danks, Genetic heterogeneity in osteogenesis imperfecta, J. Med. Genet. 16 (2) (1979) 101–116. [6] F.H. Glorieux, F. Rauch, H. Plotkin, L. Ward, R. Traverse, P. Roughley, et al., Type V osteogenesis imperfecta: a new form of brittle bone disease, J. Bone Miner. Res. 15 (9) (2000) 1650–1658. [7] D.J. Prockop, C.D. Constantinou, K.E. Dombrowski, Y. Hojima, K.E. Kadler, H. Kuivaniemi, et al., Type I procollagen: the gene-protein system that harbors most of the mutations causing osteogenesis imperfecta and probably more common heritable disorders of connective tissue, Am. J. Med. Genet. A 34 (1) (1989) 60–67. [8] E.R. Valadares, T.B. Carneiro, P.M. Santos, A.C. Oliveira, B. Zabel, What is new in genetics and osteogenesis imperfecta classification? J. Pediatr. 90 (6) (2014) 536–541. [9] H. Kang, J.C. Marini, Osteogenesis imperfecta: new genes reveal novel mechanisms in bone dysplasia, Transl. Res. 181 (2017) 27–48. [10] K. Lindahl, E. Astrom, C.J. Rubin, G. Grigelioniene, B. Malmgren, O. Ljunggren, et al., Genetic epidemiology, prevalence, and genotype-phenotype correlations in the Swedish population with osteogenesis imperfecta, Eur. J. Hum. Genet. 23 (8) (2015) 1042–1050. [11] E.E. Bleck, Nonoperative treatment of osteogenesis imperfecta: orthotic and mobility management, Clin. Orthop. Relat. Res. (159) (1981) 111–122. [12] K. Wang, M. Li, H. Hakonarson, ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data, Nucleic Acids Res. (2010), http://dx. doi.org/10.1093/nar/gkq603. [13] P. Kumar, S. Henikoff, P.C. Ng, Predicting the effects of coding non-synonymous variants on protein function using the SIFT algorithm, Nat. Protoc. 4 (2009) 1073. [14] I. Adzhubei, D.M. Jordan, S.R. Sunyaev, Predicting functional effect of human missense mutations using PolyPhen-2, Current Protocols Hum. Genet. (2013), http://dx.doi.org/10.1002/0471142905.hg0720s76. [15] J.M. Schwarz, D.N. Cooper, M. Schuelke, D. Seelow, MutationTaster2: mutation prediction for the deep-sequencing age, Nat. Methods 11 (2014) 361. [16] J.T. Robinson, H. Thorvaldsdóttir, W. Winckler, M. Guttman, E.S. Lander, G. Getz, J.P. Mesirov, Integrative genomics viewer, Nat. Biotechnol. 29 (2011) 24. [17] H. Thorvaldsdóttir, J.T. Robinson, J.P. Mesirov, Integrative genomics viewer (IGV): high-performance genomics data visualization and exploration, Brief. Bioinform. 14 (2) (2013) 178–192. [18] J.C. Marini, A. Forlino, W.A. Cabral, A.M. Barnes, J.D. San Antonio, S. Milgrom, J.C. Hyland, et al., Consortium for osteogenesis imperfecta mutations in the helical domain of type I collagen: regions rich in lethal mutations align with collagen binding sites for integrins and proteoglycans, Hum. Mutat. 28 (3) (2007) 209–221. [19] A.M. Lund, A.C. Nicholls, M. Schwartz, S. Skovby, Parental mosaicism and autosomal dominant mutations causing structural abnormalities of collagen I are frequent in families with osteogenesis imperfecta type III/IV, Acta Paediatr. 86 (7)
146
Clinica Chimica Acta 484 (2018) 141–147
N. Mohd Nawawi et al.
mutations in the collagen type I genes, Oral Dis. 23 (1) (2017) 42–49. [56] P.H. Byers, Brittle bones - fragile molecules: disorders of collagen gene structure and expression, Trends Genet. 6 (1990) 293–299. [57] J.A. Engel, D.J. Prockop, The zipper-like folding of collagen triple helices and the effects of mutations that disrupt the zipper, Annu. Rev. Biophys. Biophys. Chem. 20 (1) (1991) 137–152. [58] J.C. Marini, W.A. Cabral, M.A. Barnes, Null mutations in LEPRE1 and CRTAP cause severe recessive osteogenesis imperfecta, Cell Tissue Res. 339 (1) (2010) 59–70. [59] Y. Huang, L. Mei, W. Lv, H. Li, R. Zhang, Q. Pan, et al., Targeted exome sequencing identifies novel compound heterozygous mutations in P3H1 in a fetus with osteogenesis imperfecta type VIII, Clin. Chim. Acta 464 (2017) 170–175. [60] R.J. Fernandes, S. Hirohata, J.M. Engle, A. Colige, D.H. Cohn, D.R. Eyre, et al., Procollagen II amino Propeptide processing by ADAMTS-3: insights on Dermatosparaxis, J. Biol. Chem. 276 (34) (2001) 31502–31509. [61] A. Sangsin, C. Kuptanon, C. Srichomthong, M. Pongpanich, K. Suphapeetiporn, V. Shotelersuk, Two novel compound heterozygous BMP1 mutations in a patient with osteogenesis imperfecta: a case report, BMC Medical Genet. 18 (1) (2017) 25. [62] H. Al-Jallad, T. Palomo, P. Roughley, F.H. Glorieux, M.D. McKee, P. Moffatt, et al., The effect of SERPINF1 in-frame mutations in osteogenesis imperfecta type VI, Bone 76 (Suppl C) (2015) 115–120. [63] J. Becker, O. Semler, C. Gilissen, Y. Li, H.J. Bolz, C. Giunta, et al., Exome sequencing identifies truncating mutations in human SERPINF1 in autosomal-recessive osteogenesis imperfecta, Am. J. Hum. Genet. 88 (3) (2011) 362–371. [64] J.-Y. Wang, Y. Liu, L.-J. Song, F. Lv, X.-J. Xu, A. San, et al., Novel mutations in SERPINF1 result in rare osteogenesis imperfecta type VI, Calcif. Tissue Int. (2016), http://dx.doi.org/10.1007/s00223-016-0201-z. [65] F. Rauch, A. Husseini, P. Roughley, F.H. Glorieux, P. Moffatt, Lack of circulating pigment epithelium-derived factor is a marker of osteogenesis imperfecta type VI, J. Clin. Endocrinol. & Metab. 97 (8) (2012) E1550–E1556. [66] U. Lindert, W.A. Cabral, S. Ausavarat, S. Tongkobpetch, K. Ludin, A.M. Barnes, et al., MBTPS2 mutations cause defective regulated intramembrane proteolysis in Xlinked osteogenesis imperfecta, Nat. Commun. (2016), http://dx.doi.org/10.1038/ ncomms11920. [67] Y. Liu, Ma D. Asan, F. Lv, X. Xu, J. Wang, et al., Gene mutation spectrum and genotype-phenotype correlation in a cohort of Chinese osteogenesis imperfecta patients revealed by targeted next generation sequencing, Osteoporos. Int. (2017), http://dx.doi.org/10.1007/s00198-017-4143-8.
[45] E. Guillén-Navarro, M.J. Ballesta-Martínez, M. Valencia, A.M. Bueno, V. MartínezGlez, V. Lopez-Gonzalez, et al., Two mutations in IFITM5 causing distinct forms of osteogenesis imperfecta, Am. J. Med. Genet. A 164 (5) (2014) 1136–1142. [46] R.M. Patel, S.C.S. Nagamani, D. Cuthbertson, P.M. Campeau, J.P. Krischer, J.R. Shapiro, et al., A cross-sectional multicenter study of osteogenesis imperfecta in North America - results from the linked clinical research centers, Clin. Genet. 87 (2) (2015) 133–140. [47] O. Essawi, S. Symoens, M. Fannana, M. Darwish, M. Farraj, A. Willaert, et al., Genetic analysis of osteogenesis imperfecta in the Palestinian population: molecular screening of 49 affected families, Mol. Genet.& Genomic Med. (2017), http://dx. doi.org/10.1002/mgg3.331. [48] M. Takagi, H. Shinohara, S. Narumi, G. Nishimura, Y. Hasegawa, T. Hasegawa, Severe osteogenesis imperfecta caused by double glycine substitutions near the amino-terminal triple helical region in COL1A2, Am. J. Med. Genet. A 167 (7) (2015) 1627–1631. [49] F. Rauch, L. Lalic, P. Roughley, F.H. Glorieux, Genotype-phenotype correlations in nonlethal osteogenesis imperfecta caused by mutations in the helical domain of collagen type I, Eur. J. Hum. Genet. 18 (6) (2010) 642–647. [50] J.D. Hald, L. Folkestad, T. Harslof, A.M. Lund, M. Duno, J.B. Jensen, et al., Skeletal phenotypes in adult patients with osteogenesis imperfecta—correlations with COL1A1/COL1A2 genotype and collagen structure, Osteoporos. Int. (2016), http:// dx.doi.org/10.1007/s00198-016-3653-0. [51] G. Bardai, P. Moffatt, F.H. Glorieux, F. Rauch, DNA sequence analysis in 598 individuals with a clinical diagnosis of osteogenesis imperfecta: diagnostic yield and mutation spectrum, Osteoporos. Int. (2016), http://dx.doi.org/10.1007/s00198016-3709-1. [52] D. Sillence, B. Butler, M. Latham, K. Barlow, Natural history of blue sclerae in osteogenesis imperfecta, Am. J. Med. Genet. 45 (2) (1993) 183–186. [53] K. Andersson, G. Dahllof, K. Lindahl, A. Kindmark, G. Grigelioniene, A. Astrom, et al., Mutations in COL1A1 and COL1A2 and dental aberrations in children and adolescents with osteogenesis imperfecta – a retrospective cohort study, PLoS One 12 (5) (2017) e0176466. [54] R. Sæves, L.L. Wekre, E. Ambjornsen, S. Axelsson, H. Nordgarden, K. Storhaug, Oral findings in adults with osteogenesis imperfecta, Special Care in Dentistry 29 (2) (2009) 102–108. [55] B. Malmgren, K. Andersson, K. Lindahl, A. Kindmark, G. Grigelioniene, V. Zachariadis, et al., Tooth agenesis in osteogenesis imperfecta related to
147