Background

Multiple osteochondromas (MO) is a relatively rare autosomal dominant skeletal disorder characterized by the formation of multiple osteochondromas and skeletal deformities, including limb length discrepancy, bowing deformities of the forearms, valgus deformity of the lower extremities, and scoliosis [1–3]. In Western countries, the prevalence of MO in the general population is one in every 50,000 individuals, and men tend to be affected more frequently than women [4, 5]. Osteochondroma is a benign bone tumor exhibiting cartilage-capped bone growth that typically originates from the metaphysis of long bones or surface of flat bones. Patients commonly feel pain or irritation of the tissues due to osteochondroma, and some patients may undergo multiple surgeries during their life in an attempt to relieve the symptoms of this disorder [6]. Malignant transformation of osteochondroma toward chondrosarcoma is a serious complication in MO and occurs in 0.38–7.0 % of patients [6–9].

The exostosin-1 (EXT1) and exostosin-2 (EXT2) genes, which encode heparin sulfate glycosyltransferases, are major causative genes in MO [10, 11]. EXT1 is located on chromosome 8q23-q24 [12], and EXT2 is located on chromosome 11p11-p12 [13]; these genes are essential for heparan sulfate chain elongation. Approximately 90 % of patients with MO harbor EXT1 or EXT2 germline mutations; however, the genetic background of patients with MO is heterogeneous. Several reports have described mutational variations and novel mutations in EXT1 and EXT2 genes in patients with MO in several different countries [14–17], and the distribution of mutations in the EXT1 and EXT2 genes has been shown to vary. For example, in Spanish patients with MO, 74 % had mutations in EXT1, and 21 % had mutations in EXT2 [14]. In contrast, in Polish patients with MO, 54.6 % had mutations in EXT1 and 30.3 % had mutations in EXT2 [15]. Most of these mutations are inactivating mutations, including nonsense, frameshift, and splice-site mutations [18]. However, only one study has described variations in genotypes for Japanese patients with MO [16].

Therefore, in this study, we sought to determine genetic backgrounds and mutational patterns in 71 Japanese families with MO; this report describes the genetic diagnostic results of the largest Japanese cohort of MO patients presented to date and identified several novel mutations in the EXT1 and EXT2 genes in MO.

Results

Identification of 22 novel genetic lesions

In this study, all exons and intron/exon junctions in EXT1 and EXT2 were sequenced in 112 patients with MO from 71 families. Eighty (71.4 %) patients harbored 52 mutations in either EXT1 or EXT2. All EXT1 and EXT2 mutations are shown in Table 1. Twenty-nine families (40.8 %) had mutations in EXT1, and 15 families (21.1 %) had mutations in EXT2. Interestingly, three families (4.2 %) had mutations in both EXT1 and EXT2. Twenty-four families (33.8 %) did not have mutations in EXT1 or EXT2. The distribution of mutations was as follows: 40.4 % (21/52) of patients had missense mutations, 30.8 % (16/52) of patients had frameshift mutations, 21.2 % (11/52) of patients had nonsense mutations, 5.8 % (3/52) of patients had splicing mutations, and 1.9 % (1/52) of patients had insertions. Of all 52 mutations, 22 mutations (42.3 %) were novel mutations that had not been registered in the Multiple Osteochondroma Mutation Database (MOdb) (http://medgen.ua.ac.be/LOVDv.2.0/home.php) [18]. Of these mutations, 17 mutations (77.3 %) were identified in the EXT1 gene, while five mutations (22.7 %) were found in the EXT2 gene. The distribution of novel mutations was as follows: 27.3 % (6/22) of patients had missense mutations, 31.8 % (7/22) of patients had frameshift mutations, 22.7 % (5/22) of patients had nonsense mutations, 9.1 % (2/22) of patients had splicing mutations, and 4.5 % (1/22) of patients had insertions.

Table 1 Ext1 and Ext2 mutations in Japanese MO families
Full size table

Characteristic genome mutations in five families with MO

Interestingly, five families with MO showed unique genotypes (MO-11, -25, -44, -47, and -59) as illustrated in Fig. 1. In the family with MO-11, one MO patient harbored missense mutations in both exon 2 of EXT1 and exon 2 of EXT2. Moreover, parents and children in the families with MO-25, -44, and -47 showed different genotypes. In the family with MO-59, a patient had a double missense mutation in EXT1.

Fig. 1
figure 1

Characteristic mutations and hereditary types in Japanese families with MO. Black mark represents patient with MO. White mark represents healthy person. NA: DNA not available. Written informed consents to publish were obtained from each participants described in this figure before study participation

Full size image

Discussion

In this study, we evaluated the presence and features of EXT1 and EXT2 mutations in Japanese families with MO. Our data demonstrated that 29 families (40.8 %) had mutations in EXT1, and 15 families (21.1 %) had mutations in EXT2. Moreover, three families (4.2 %) had mutations in both EXT1 and EXT2, and 24 families (33.8 %) did not have mutations in either EXT1 or EXT2. Of the 52 mutations observed in this study, 34 mutations were identified in EXT1, and 18 mutations were identified in EXT2. Of the 52 mutations 22 novel mutations were identified. Thus, the data presented herein provides important insights into the genetic causes of MO in Japanese families.

MO is an autosomal dominant disorder, and germline and heterozygous mutations conferring loss of function in the EXT1 and EXT2 genes are main causes of MO. Mutational variations in EXT1 and EXT2 are continuously being reported; as of January 2015, 432 mutations in EXT1 and 223 mutations in EXT2 were registered in the MOdb (http://medgen.ua.ac.be/LOVDv.2.0/home.php) [18]. Several studies have described the mutational variations in EXT1 and EXT2 in European countries and Asia. For example, in Spanish patients with MO, 74 % were found to have mutations in EXT1, and 21 % were found to have mutations in EXT2 [14]. Additionally, in Polish patients with MO, 54.6 and 30.3 % were found to have mutations in EXT1 and EXT2, respectively [15]. In an Italian cohort, 69 % of patients were found to have mutations in EXT1, and 27 % of patients were found to have mutations in EXT2. In a previous study of Japanese families with MO, 17 (40 %) of the 23 families had a mutation in EXT1, and six (14 %) of the 23 families had a mutation in EXT2 [16]. In contrast, in Chinese families with MO, 13.9 and 33.3 % of the 36 families were found to have mutations in EXT1 and EXT2, respectively [17]. In most studies, the prevalence of EXT1 mutations has been reported to be higher than that of EXT2 mutations. Similarly, in our current analysis, a greater proportion of EXT1 mutations was observed (EXT1: 40.8 %, EXT2: 21.1 %). However, mutations in these genes were not identified in 24 families (33.8 %) with MO; this percentage was relatively high compared with that in European countries, where the proportion of patients without mutations in EXT1 and EXT2 has been shown to range from 4 to 24 % (Fig. 2a) [9, 14, 19, 20]. Further studies are needed to examine this finding such as MLPA assays because the families with MO harboring no mutations in this study might include deletion mutation. While the EXT family also includes three EXT-like genes (i.e., EXTL1, EXTL2, and EXTL3) [21–23], no reports have described the presence of gene mutations in these three EXT-like genes in families with MO. Therefore, further studies are needed to determine whether these three genes may be causative genes in families with MO who do not harbor mutations in EXT1 and EXT2 genes.

Fig. 2
figure 2

Comparison of mutation frequencies. a The proportions of EXT1 and EXT2 mutations. b The proportion of missense mutations

Full size image

In MO, the most common mutation types in the EXT1 and EXT2 genes are inactivating mutations, such as frameshift, nonsense, and splice-site mutations [24]. Similarly, in data reported in MOdb from 2009, approximately 80 % of mutations in EXT1 and 77 % of mutations in EXT2 were found to be inactivating mutations (EXT1: frameshift 44 %, nonsense 24 %, splice-site 11 %; EXT2: frameshift 42 %, nonsense 22 %, splice-site 13 %) [18]. In Spanish patients with MO, the prevalence of inactivation mutations was reported to be 79.5 % [14]. In our study, 52 mutations were found in EXT1 and EXT2, and 59.6 % (31/52) of these mutations were inactivating mutations (with 30.8, 23.1, and 5.8 % of mutations being frameshift, nonsense, and splice-site mutations, respectively). The proportion of missense mutation was approximately 38.5 %, which was relatively higher in this study than in previous reports (Fig. 2b) [14, 18]. The differences in gene mutations between patients with MO in Japan and other countries may be related to the differences in the prevalence rates of MO or the severity of skeletal abnormalities, including scoliosis, in the various countries. Further studies are required to determine the phenotype-genotype relationships in Japanese patients with MO.

In the present study, approximately 7.0 % (5/71) of families with MO showed characteristic genotypes, e.g. one patient bearing two mutations and a parent and child bearing different mutations as shown in Fig. 1. Especially, in MO-25, MO-44 and MO-47, there was one affected parent with MO and one (or two) affected child, but genotypes of child differed from that of the parents. It was reported that approximately 10 % of patients with MO exhibited de novo mutations [18]; thus, the difference of genotype between parent and child in the three families may be caused by de novo mutations although this scenario may be very unlikely. Therefore, in the future analysis, sequence or genotype might be necessary for the three families to reconfirm these results.

In MO-59, there were two mutations in one MO patient and one was a novel missense mutation. Either mutation might obtain the possibility of it being a non-pathogenic variant; however we evaluated only one genome in MO-59 families, so it is unclear whether mutation is a non-pathogenic variant.

This study has a limitation. In this study, we performed the mutational analysis in EXT1 and EXT2 for Japanese families with MO with using direct sequencing, but MLPA analysis has not been performed for the families with MO harboring no mutations. Thus, the families with MO harboring gene deletions might not be detected and unknown rate of EXT1and EXT2 mutation potentially are high in this study. Further analysis will be necessary, and now we are planning to perform whole-genome sequencing with using next-generation sequencing technology for the families with MO harboring no mutations. In addition, in MO-25, MO-44 and MO-47, re-sequence or genotype would be performed because of unlikely hereditary form.

Conclusions

In this study, we evaluated and characterized mutations in the EXT1 and EXT2 genes in 71 Japanese families with MO. A total of 52 mutations in EXT1 and EXT2 were identified, with 22 of these mutations being reported here for the first time. Additionally, we identified several characteristics of gene mutations in EXT1 and EXT2. Approximately 60 % of Japanese families with MO had inactivating mutations in EXT1 and EXT2. Interestingly, these results differed somewhat from those from other countries and represented the variety of genotype in MO. Further studies are needed to determine the reasons for these differences. This study provides important insights into our understanding of the genetic features of MO in Japanese individuals.

Methods

Study design and ethical approval

In this study, we performed a multicenter study at Gifu University, National Center for Child Health and Development, and Kyusyu University. Ethics Committee of Gifu University (Approval No. 22-221) approved all procedures, and all participants obtained written informed consent before any research procedures. In case of the participant under the age of 16 year old, written informed consents (child assent and parental consent) were obtained. In addition, written informed consents to publish were obtained from all patients before study participation, and in the case of the participant under the age of 16 year old, written informed consents to publish (child assent and parental consent) were obtained.

Patients and clinical studies

From April 2010 to September 2014, patients with MO were recruited for genetic testing of the EXT1 and EXT2 genes. A total of 116 patients (51 women and 65 men) from 74 families with MO were recruited. Clinical diagnosis was performed based on accurate family histories and physical examinations of the patients, including palpation tests for osteochondromas or joint deformities. Ethics Committee of Gifu University approved all procedures, and all participants obtained written informed consents before all procedures.

Mutation analysis

Genomic DNA was isolated from peripheral blood leucocytes of all patients with MO using a Wizard Genomic DNA purification kit (Promega, Madison, WI, USA). All exons and exon/intron junctions in the EXT1 and EXT2 genes (GenBank accession numbers NM_000127.2 and NM_207122.1) were amplified by PCR. After confirming amplification of the DNA fragments by agarose gel electrophoresis and purifying the amplified DNA fragments using a Wizard SV gel and PCR Clean-up System (Promega), the amplified DNA fragments were directly sequenced using a BigDye Terminator v1.1Cycle Sequencing kit (ABI). Sequence analyses were then performed with an ABI PRISM 3100 Genetic Analyzer (ABI). Mutations in EXT1 and EXT2 were evaluated by comparing DNA sequences of normal EXT1 and EXT2 genes with the obtained sequences using Sequencher software (Hitachi Software Engineering Co., Ltd., Tokyo, Japan). Primer sequences are shown in Additional file 1: Table S1. The detected mutations in EXT1 and EXT2 were examined to determine whether they had been reported previously by consulting the MOdb (http://medgen.ua.ac.be/LOVDv.2.0/home.php) [18].

In four patients of three families (MO-62, 63, 64), DNA sequence analysis could not be performed because EXT1 and EXT2 genes were not amplified by polymerase chain reaction (PCR). Finally, 112 patients of 71 MO families (48 women and 64 men), DNA sequence analysis was performed. All procedures were approved by Ethics Committee of Gifu University, and all participants obtained written informed consents before all procedures.

Availability of data and materials

The Datasets used in this paper can be found at http://medgen.ua.ac.be/LOVDv.2.0/home.php [18]. All supporting data are included in the manuscript as well as additional files in the supplementary section.