Introduction

Hereditary Multiple Exostoses (HME) or Multiple Osteochondromas (MO) is a rare autosomal-dominant pediatric disorder with an incidence of about 1 in 50,000 individuals and male-to-female ratio of about 1.5:1 [1, 2]. The disease is characterized by the development of two or more cartilage capped bony outgrowths within perichondrium in long bones and ribs, which can cause a variety of orthopedic deformities such as disproportionate short stature, shortened forearms, and unequal limb length. Although it is generally a benign skeletal tumor, 2.8% (0.5–5%) of patients undergo malignant transformation towards life-threatening chondrosarcomas or osteosarcomas due to their typical resistance to chemo- or radiation therapy [3, 4].

Germline heterozygous loss-of-function mutations in the EXT1 (exostosin-1, located on chromosome 8q23-q24) or EXT2 (exostosin-2, located on chromosome 11p11-p12) tumor suppressor genes are responsible for over 70–95% of HME cases [5, 6]. There are 566 EXT1 and 278 EXT2 mutations reported in the literature (HGMD database). The majority of these mutations (79% in EXT1 and 75% in EXT2) are frameshift, nonsense, and splice-site mutations, resulting in truncated proteins [5]. About 65% of the mutations occur in EXT1 and 25% in EXT2. In about 10–15% of HME cases, genomic alterations cannot be detected by the conventional method due to alterations such as intronic deletions, translocations or somatic mosaicism [7, 8]. The involvement of other genes or the putative EXT3 gene on chromosome 19 still needs investigation.

The genetic defects causing HME have not been systematically investigated in the Arab population. In the present study, we performed molecular analysis of 22 patients from 17 unrelated Saudi families with HME. EXT1 or EXT2 mutations were identified in 77% of patients (13/17) including six novel mutations.

Subjects and methods

Patients

Seventeen Saudi families with HME were investigated (Fig. 1 and Table 1). The inclusion criteria were two or more exostoses diagnosed upon physical and radiographic examinations. Disease severity was divided into 3 classes based on the presence of skeletal deformities and functional limitations using the following criteria: Class I: no deformities and no functional limitations [A ≤ 5 sites with osteochondromas, B > 5 sites with osteochondromas]; Class II: deformities and no functional limitations [A ≤ 5 sites with deformities, B > 5 sites with deformities]; and Class III: deformities and functional limitations [A functional limitation of 1 site, B functional limitation of > 1 site] [9]. Blood samples were obtained from patients and available relatives for genomic DNA extraction after informed consent. The study was approved by the Ethics Committee of King Faisal Specialist Hospital and Research Centre (RAC # 2170 027). Written consent was obtained from the patients or guardian of the patients before enrollment.

Fig. 1
figure 1

Radiology of patients with osteochondromas. Patient#1 has an osteochondroma at left hip joint; Patient #15 has an osteochondroma at right proximal humerus; Patient#18 has an osteochondroma at left distal radius; and Patient # 21 has a right pelvic osteochondroma with malignant transformation. Osteochondroma is indicated by an arrow

Full size image
Table 1 Genetic defects in 17 families with hereditary multiple Exostoses
Full size table

Genomic DNA isolation

Genomic DNA from peripheral blood leukocytes was extracted as described previously [10].

DNA amplification and sequencing

DNA samples were analyzed for mutations in all the coding exons and intron–exon boundaries of EXT1 and EXT2 genes by polymerase chain reaction (PCR) and sequencing analysis. PCR primers and conditions were described previously and listed in Table 2 [11]. The resulting PCR products were directly sequenced with BigDye Terminator 3.1 Cycle Sequencing kit using an automated ABI PRISM 3700 sequencer (Applied Biosystems; Life Technologies, Foster City, CA).

Table 2 EXT1 and EXT2 primer sequences and PCR conditions
Full size table

Analysis of copy number variation

Copy number variation in genomic DNA was analyzed by MLPA (Multiplex Ligation-dependent Probe Amplification) analysis as described previously [12].

Results

EXT1 and EXT2 mutations were identified in 13 out of 17 (77%) unrelated patients and 18 of total 22 patients (82%) (Table 1). Among them, 7 were EXT1 mutations including 1 recurrent mutation in one related family member (35%, 6/17 unrelated patients or 32%, 7/22 total patients); 11 were EXT2 mutations including 4 recurrent mutations from 4 family members (41%, 7/17 unrelated patients, or 50%, 11/22 total patients) (Table 1). Among 13 different mutations, 7 were previously reported mutations (Table 1, Fig. 2) and 6 were novel mutations (Fig. 3). Three novel mutations occurred in the EXT1: c.739G > T (p.E247*), c.1319delG (p.R440Lfs*4), and c.1786delA (p.S596Afs*25) and 3 in the EXT2: c.541delG (p.D181Ifs*89), c.583delG (p.G195Vfs*75) and a gross homozygous deletion of approximately 10 kb including promoter and exon 1 (Table 1, Fig. 3). In the patient with the homozygous deletion, we were able to amplify exon 2 to 14 successfully, but could not amplify exon 1 and its 5′ untranslated region of about 10 kb, indicating a 10 kb deletion of exon 1 and the promoter region. Five patients from unrelated families were found to have mutations without any family history of the disease and these mutations were thus de novo mutations (29%, 5/17). Interestingly, 4 of them were also novel mutations (Table 1). MLPA analysis was performed to detect large deletions in the patients who had no mutation detected by PCR-sequencing analysis. One large heterozygous deletion involving exons 2–11 was detected (Table 1). Among 13 different mutations, 6 were single nucleotide deletions, 3 were nonsense mutations, 1 missense mutation, 1 splice donor site mutation, and 2 large deletions. Therefore, all the mutations except for one missense mutation (92%, 12/13) are predicted to result in frameshift and truncated proteins devoid of enzymatic activity.

Fig. 2
figure 2

Sequence analysis of EXT1 and EXT2 in the patients with hereditary multiple exostoses. Representative electropherograms of previously reported EXT1and EXT2 mutations are shown. Heterozygous mutations are present in the patients and affected family members except for the affected mother (patient#12) in Family 9 who carries a homozygous mutation whereas her daughter (patient#13) has a heterozygous mutation. The mutation is indicated by an arrow

Full size image
Fig. 3
figure 3

Detection of novel EXT1 and EXT2 mutations. a Sequence analysis of EXT1 and EXT2 in the patients with hereditary multiple exostoses. Representative electropherograms of 5 novel EXT1 and EXT2 mutations are shown. They are also de novo mutations except for c.541delG (p.D181Ifs*89) in Family 6. Heterozygous mutations are present only in the patients. The mutation is indicated by an arrow. b Agarose gel analysis of a homozygous EXT2 exon 1 deletion. PCR products were run in a 1.3% agarose gel. Exon 1 was not amplified from patient #11 whereas the remaining exons 2–14 were amplified (only exon 2 amplification was shown)

Full size image

Compared to the patients with EXT2 mutations, most patients with EXT1 mutations had more severe phenotype and required surgery. Germline homozygous EXT2 mutations were identified in two patients (patient #11 and 12 in Table 1) who presented only mild asymptomatic disease and no clinical intervention was required. Furthermore, significant heterogeneity in clinical presentations were demonstrated among family members carrying the same mutations. For example as shown in Table 1, patient#12 carried a homozygous EXT2 c.540G > A mutation with only mild asymptomatic disease whereas her daughter (patient#13) had a heterozygous EXT2 c.540G > A mutation and required multiple operations to remove exostosis and correct bone deformity.

Discussion

In the present study, we have studied EXT1 and EXT2 mutation spectrum in 22 patients from 17 unrelated Saudi families. Disease-causing mutations are identified in 77% of patients (13/17) including 6 novel mutations. The frequency of EXT1 mutation is lower than EXT2: 35% (6/17) for EXT1 and 41% (7/17) for EXT2. Twenty-nine percent of patients (5/17) have de novo mutations, which account for 39% (5/13) of mutations identified.

EXT1 and EXT2 encode for 746 and 718 amino acids glycosyltransferases, respectively, that are involved in the chain elongation step of heparan sulfate biosynthesis in the cell’s Golgi apparatus [13,14,15]. Heparan sulfate is an essential component of cell surface and matrix-associated proteoglycans, which function by interacting with key heparin sulfate-binding proteins such as bone morphogenetic proteins (BMPs), fibroblast growth factor (FGF), Hedgehog and Wnt signaling proteins to regulate skeletal growth and morphogenesis [16, 17]. The growth plate of long bones is known to contain large amounts of heparan sulfate proteoglycans, such as syndecan, glypican and perlecan during cartilage development [18]. The glycosyltransferases are ubiquitously expressed type II transmembrane glycoproteins with transmembrane domain at the N-terminal end, an exostosin interaction domain in the center and a catalytic domain at the C-terminal end. EXT1 and EXT2 form a hetero-oligomeric complex in vivo that leads to accumulation of both proteins in the Golgi apparatus. The Golgi-localized EXT1/EXT2 complex possesses substantially higher glycosyltransferase activity than EXT1 or EXT2 alone, suggesting that the hetero-oligomeric complex is the biological form of the enzyme for heparan sulfate biosynthesis and explains mutations in either EXT1 or EXT2 gene would result in the loss of enzymatic activity and disease development [19,20,21].

HME is a rare childhood-onset skeletal disease caused by germline mutations in the tumor suppressor gene EXT1 or EXT2. Most HME patients carry a germline heterozygous loss-of-function mutation in the EXT1 or EXT2 and display a 50% reduction of systemic heparin sulfate [22]. It is generally believed that exostosis formation and associated defects, such as growth retardation and skeletal deformities, require loss-of-heterozygosity or a second hit in the affected cells [23, 24]. Mice with single heterozygous deletion of Ext1± or Ext2± are normal. Compound heterozygous Ext1+/−; Ext2+/− deletion mice and conditional Ext1 knockout mice display multiple osteochondromas and closely resemble human HME [25,26,27]. However, a second hit in the EXT1 or EXT2 gene are not common in most cases (more than 60%), suggesting that mechanisms other than EXT genetic alterations may play a role in the disease development [28, 29]. In our patients, homozygous germline EXT2 mutations were detected in two patients (patient #11 and 12 (Table 1, Fig. 2 and 3b). To our knowledge, homozygous germline EXT1/EXT2 mutations have not been reported in the literature. Interestingly, the presence of homozygous germline EXT2 mutations does not associated with severity of the disease since both patients have mild asymptomatic disease. Furthermore, no significant difference in clinical presentations or disease progression is found between patients with mutation and those without mutation. In fact, significant heterogeneity in disease development and progression are observed among patients with or without mutations. This is even demonstrated among family members carrying the same mutations, indicating epigenetic and/or environmental factors may contribute to the disease development and progression.

It has been reported that EXT1 mutation is more common (about 65%) than EXT2 (about 30%) and its protein is less tolerant to the damaging mutations [5, 30]. This may explain EXT1 mutations usually result in more severe disease phenotype. Indeed, most of our patients with EXT1 mutations have more severe phenotype and require surgery. In contrast to the higher EXT1 mutation rate reported in the literature, the frequency of EXT1 mutation appears to be lower than EXT2 in our current study. It remains to be determined whether this is due to small sample size or population-specific.

The most common type of mutations in the EXT1 and EXT2 genes are inactivating mutations, such as frameshift, nonsense, and splice-site mutations [6, 31, 32]. Based on the HGMD® Professional 2020.1 (Accessed on August 10, 2020), approximately 79% EXT1 mutations and 75% EXT2 mutations are inactivating mutations: frameshift 47% (268/566), nonsense 22% (123/566), splice-site 10% (58/565) in the EXT1; frameshift 43% (119/278), nonsense 22% (60/278), splice-site 10% (29/278) in the EXT2. The remaining EXT1 mutations are missense (12%, 68/566), gross deletions (7%, 40/566), and complex rearrangements (1%, 7/566) whereas remaining EXT2 mutations are missense (14%, 40/278) and gross deletions (9%, 26/278). In our current study of Saudi patients, the overall frequency of inactivating EXT1 and EXT2 mutations is 92% (12/13): frameshift 46% (6/13), nonsense 23% (3/13), splice-site 8% (1/13), gross deletion (15%, 2/13), which is higher than the overall rate documented in the HGMD (78%, 657/843). This is probably due to small sample size in our study. All of these mutations (92%, 11/12) are predicted to result in truncated proteins devoid of enzymatic activity. Four patients were not found to have EXT1/EXT2 mutations (Patient# 3, 16, 17, 18). Although HME may be confused with enchondroma which is a benign cartilage tumor, enchondroma often affects the cartilage that lines the inside of long bones in the hands and feet. The clinical and radiographic features of our patients (multiple bony outgrowths on the external surface in the metaphysis of long bones) do not support the diagnosis of enchondromas. The involvement of additional genes other than EXT1/EXT2 or other mechanisms may contribute to the disease development [7, 8].

De novo EXT1 and EXT2 mutations have been reported to account for approximately 10% of patients [5, 33]. However, higher frequency are reported in other populations: Polish (21%) [34], English (33%) [35], and Chinese (30%) [36]. The high de novo mutation rate in the Saudi patients (29%) indicates that family history should not be relied upon heavily in the diagnosis of the disease.

Conclusions

We have investigated genetic defects of EXT1 and EXT2 in the Saudi HME patients. EXT1 and EXT2 mutations are detected in 77% of patients. De novo EXT1 and EXT2 mutations are common. The current study further expands the mutation spectrum of HME.