Genetic analysis of seven pateints with Hereditary Multiple Osteochondromas (HMO)

Abstract

Background: HMO (Hereditary Multiple Osteochondroma), an uncommon autosomal dominant disorder, is characterized by the development of multiple osteochondromas, which are nonmalignant cartilage-capped bone tumors growing outwards from long bone metaphyses. Methods: The present work retrospectively analyzed seven children with HMO who were enrolled for routine clinical diagnosis and treatment, including X-ray examination. Subsequent genetic detection was carried out using whole exome sequencing (WES). In addition, this work applied Sanger sequencing to be the validation approach. Moreover, this work also examined amino acid (AA) evolutionary conservatism under the influence of certain missense variants. Results: The clinical indications of all seven patients and their family members were thoroughly indexed. WES identified diagnostic variants in the EXT1 or EXT2 gene in these patients. In these variants, four were reported for the first time, namely EXT1: c.1285-2A>T, EXT2: c.1139delT, EXT1: c.203G>A, and EXT1: c.1645_1673del. Familial validation revealed that three of the variants were hereditary, while the other four were de novo, which was consistent with the phenotype in each case. Conclusion: Our results expanded HMO variation spectrum, and laid certain foundations for the precise counseling of those affected families.

Introduction

Hereditary multiple osteochondroma (HMO, MIM #133700/133701), called hereditary multiple exostoses (HME), as well multiple cartilaginous exostoses, or diaphyseal aclasis, is a rare monogenic skeletal disease with the feature of the development of multiple osteochondromas, which are nonmalignant cartilage-capped bone tumors growing outwards from long bone metaphyses [1]. HMO patients usually have persistent pain as a result of compression on the surrounding tissues, blood circulation disturbance, sometimes due to spinal compression. Nonetheless, malignant development of chondrosarcoma is its serious complication, which makes up around 3.9% of all HMO cases [2]. The exact incidence of HMO is not clear due to the presence of asymptomatic patients, yet it is estimated to be about 1/50,000 in western ethnicities [3-5]. Mordenti et al. classified HMOs in 3 categories based on whether deformity or functional limitation occurred [6].

EXT1 (chromosome 8q24.11; MIM no. *608177) and EXT2 (chromosome 11p11.2; MIM no. *608210) genes, encode the Golgi-associated glycosyltransferases, plus exostosin 1 and 2, which are involved with the heparan sulfate (HS) biosynthesis. Those two genes are the well-defined causative genes for HMO [1] in approximately 90% of patients, while another locus with potential pathogenicity has been postulated [7]. It is demonstrated that heparan sulfate proteoglycans (HSPGs), which are constituted by HS chains and core protein, are related to different physiological and developmental activities in the extracellular matrix (ECM) or on cell surface [8], most importantly, the interaction of them with certain key signaling factors in osteogenic differentiation and bone homeostasis [9]. The fact that mutations in EXT1 and EXT2 genes can cause HMO was first identified by Ahn et al. in 1995 and Wuyts et al. in 1996, respectively [10,11]. So far, over eight hundred pathogenic variants in these two genes which are found to be responsible for HMO have been reported (Human Gene Mutation Database, HGMD; http://www.hgmd.cf.ac.uk/). These variants include a variety of types, of which small insertions and missense mutations together account for more than half [2]. However, there are still ~10% of HMO patients without a clear molecular diagnosis, which makes their consultation and precise management challenging [12].

In the current study, seven families with HMO patients were enrolled. A comprehensive clinical evaluation including X-ray imaging and postoperative pathological examination was conducted; subsequently, this work was conducted with genetic analysis by whole exome sequencing (WES), followed by validation by Sanger sequencing to identify the causative variations.

Materials and methods

Subjects

The present work retrospectively enrolled seven pediatric cases presenting with HMO who were recruited in our centers between Jan 2014 and Dec 2021. A thorough clinical evaluation, X-ray imaging and post-operative routine pathological examination on each proband was carried out. Expanded examination on the family members with symptoms was conducted according to the needs of demanding families. Subsequently, the peripheral blood samples for genetic testing were taken in patients, corresponding parents and symptomatic family members.

The present work gained approval from the Ethics Committee of Shijiazhuang Obstetrics and Gynecology Hospital (No. 2021 0095). Each subject provided informed consent for participation. Each procedure was carried out following the Declaration of Helsinki 1964 together with the subsequent amendments or relevant ethical standards.

DNA extraction

QIAamp DNA Blood Mini-Kit (Qiagen Sciences, USA) was utilized for extracting genomic DNA (gDNA). Thereafter, 1% agarose gels were utilized to validate DNA quality by using Qbit DNA Assay Kit from Qubit 2.0 Flurometer (Life Technologies, CA, USA).

Whole exome sequencing (WES)

WES was carried out as previously described [13-15]. In brief, Sure Select Human Exon Sequence Capture Kit (Agilent, USA) was employed to enrich exonic sequences. Illumina DNA Standards and Primer PremixKit (Kapa Biosystems, USA) was adopted for quantifying sequencing libraries, followed by massive parallel-sequencing with Illumina Novaseq6000 platform. Afterwards, low-quality readings were sequenced and screened, whereas high-quality reads (Q30 percentage > 89%) were later compared with human genome reference sequence [hg19]. Those suspected pathogenic variants (https://software.broadinstitute.org/gatk/) were identified by GATK software. NCBI Reference Sequence using Chromas v2.33 was adopted for aligning sequences against NCBI to identify variations. For those variants identified, their pathogenicity was evaluated based on guideline released by the American Association of Medical Genetics and Genomics (ACMG) [16] with reference to several databases (1000g2015aug_eas, https://www.internationalgenome.org/; ExAC_EAS, http://exac.broadinstitute.org; gnomAD_exome_EAS, http://gnomad.broadinstitute.org/); HGMD: Human Gene Mutation Database (Professional Version 2019.4) with the Enliven^® Variants Annotation Interpretation system (Berry Genomics, China).

Variant validation

By adopting ABI 3730 Automated Sequencer (Applied Biosystems, USA), Sanger sequencing was conducted for validating suspected diagnostic variants in line with specific instructions. MEGA7 (http://www.megasoftware.net) was utilized to examine amino acid (AA) evolutionary conservatism under the influence of certain missense variants by the use of default parameters.

Results

Clinical manifestations and family history

The key clinical information of each case was included in Table 1. The characteristic imaging manifestations were shown in Figures 1 and 2. Pedigree diagram for every family member indicating the relationship was demonstrated in Figure 3. According to our survey, three cases of the seven had family history of HMOs (Case 3, 4 and 5; see Figure 3), while the other four were sporadic. Like general HMO cases, in most of our patients the lesions were concentrated into long bone epiphyses, in particular lower extremities, which were also where the resections were carried out. Yet, there were two cases with additional rare onset sites, which were the right scapula in Case 2 and the left ring finger in Case 5 (Figures 1 and 2, red circles).

Table 1.

Clinical information of the seven patients with HMO

Patient No.	Gender*	Age at diagnosis	Brief clinical history of the proband in each family*	Family history and related issues of symptomatic members
1	M	6	Multiple, gradually enlarged phymas appeared initially on the extremities; Diagnosis was made, and subsequent resections of OC were performed for six times from 2014 to 2020.	None family history
2	M	8	Two previous HMO resections; Multiple OCs at the joints of the extremities; Chronic pain; Bilateral forearms varus deformity, both hips adduction flexion limited; 3 HMO resections in our center from 2019 to 2021.	None family history
3	M	4	Initial OCs in the right upper extremity; Progressive local pain; X-ray revealed HMOs in extremities; 1 resection at right distal ulna and bilateral distal femurs performed in Sept 2021.	Maternal HMO; 1 excision performed on the mother at 11 years old
4	F	3.5	Multiple OCs; 1st resection at right proximal tibia and distal fibula in May 2017; 2021, Left limping; X-ray revealed HMOs in lower extremities; Left common peroneal nerve injury by EMG; 2nd resection conducted in Jun 2021.	Familial HMO in father and paternal grandma; 1 excision performed on the father at 10 years old
5	M	6	Abnormal walking and right knee valgus for 2 years; X-ray revealed OCs in bilateral lower extremities; Resection conducted in Jul 2021.	Father and younger brother with HMOs; 1 excision performed on the brother in Oct 2021
6	M	7	Multiple OCs; Local pain; X-ray revealed OCs in lower extremities; 2 resections performed from 2018 to 2020.	None family history
7	M	6.5	Aged 30; Having undergone 3 OCs resections from 2000 to 2007.	None family history

^*M, male; F, female; OC: osteochondroma.

Figure 1.

Clinical manifestations of the HMO patients (Part I). Case 1: the X-ray images of the patient’s hip (A), tibiofibulas (B), knees (C) and right forearm (D) showing multiple exostoses. Case 2: the X-ray images of the patient’s both humerus (E, F), knees (G), and left scapula (H) with multiple exostoses. (The red circle indicates rare lesion) Case 3: the X-ray images of the patient’s hip (I), tibiofibulas and knees (J), both arms (K, L) showing multiple exostoses at the metaphysis of long bones.

Figure 2.

Clinical manifestations of the HMO patients (Part II). Case 4: the X-ray images of the patient’s hip (A), tibiofibulas and knees (B, front; C, side) with exostoses at the metaphysis regions. Case 5: the X-ray images of the patient’s stem (D, scoliosis), hip (E), knees (F), tibiofibulas (G) and left palm (H) showing multiple exostoses. (The red circle indicates rare lesion) Case 6: the X-ray images of the patient’s upper arms (I, J), tibiofibulas and knees (K) and knees (L, M) showing exostoses at the metaphysis regions. Case 7: the X-ray images of the patient’s hip (N), tibiofibulas (O), knees (P, backwards), left knee (Q, side) and left forearm (R) with multiple exostoses.

Figure 3.

The pedigree diagrams, genetic variants, and the corresponding carrying status in each family. Case 1: the pedigree diagram (A), and the Sanger peak image of EXT1:c.1285-2A>T variant (B). Case 2: the pedigree diagram (C), and the Sanger peak image of EXT2: c.1139delT variant (D). Case 3: the pedigree diagram (E) and bam image of EXT1: c.1016G>A variant (F); (G) the evolutionary conservatism of the EXT1: G339 residue among species. Case 4: the pedigree diagram (H), and bam image of EXT1: c.203G>A variant (I). Case 5: the pedigree diagram (J), and Sanger peak image of EXT1: c.963-1G>T variant (K). Case 6: the pedigree diagram (L), and Sanger peak image of EXT1: c.1469delT variant (M). Case 7: the pedigree diagram (N), and Sanger peak image of EXT1: c.1645_1673del variant (O). Black arrows indicate probands, and red arrows or boxes indicate mutation sites.

Genetic findings

WES identified the diagnostic variations in all seven cases. Figure 3 shows the peak diagram (Sanger sequencing result) or Bam result (next-generation sequencing) for every variant analyzed in the present work. Altogether 6 cases carried heterozygous variants of EXT1, whereas another case had variation in EXT2. To be specific, Patient 1 had the splicing site variant, EXT1: c.1285-2A>T; Patient 2 carried a frameshift variant, EXT2: c.1139delT (p.I380Tfs*30); Patient 3 carried a missense variant, EXT1: c.1016G>A (p.G339D); Patient 4 carried a nonsense variant, EXT1: c.203G>A (p.W68*); Patient 5 carried a splicing site variant, EXT1: c.963-1G>T; Patient 6 carried a frameshift variant, EXT1: c.1469delT (p.L490Rfs*9); and Patient 7 carried a frameshift-causing coding sequence deletion, EXT1: c.1645_1673del (p.F550Rfs*6). The variant details were included in Table 2. Notably, four were reported for the first time, namely EXT1: c.1285-2A>T, EXT2: c.1139delT, EXT1: c.203G>A, and EXT1: c.1645_1673del.

Table 2.

Information of the identified variations in the seven patients

Patient No.	Gene (Transcript version)	Genomic alteration	Peptide alteration	Frequency in three databases*	Prediction by REVEL*	HGMD*/ClinVar* rating	Pathogenicity rating* (Evidences)
1	EXT1 (NM_000127)	c.1285-2A>T	exon skipping?	0; 0; 0	-	Not indexed	Pathogenic (PVS+PS2+PM2+PM5)
2	EXT2 (NM_207122)	c.1139delT	p.I380Tfs*30	0; 0; 0	-	Not indexed	Pathogenic (PVS1+PS2+PM2)
3	EXT1 (NM_000127)	c.1016G>A	p.G339D	0; 0; 0	P	DM/P	Pathogenic (PS4+PM1+PM2+PM5+PP3)
4	EXT1 (NM_000127)	c.203G>A	p.W68*	0; 0; 0	-	Not indexed/P	Likely pathogenic (PVS1+PM2)
5	EXT1 (NM_000127)	c.963-1G>T	exon skipping?	0; 0; 0	-	DM/Not indexed	Pathogenic (PVS1+PS4+PM2)
6	EXT1 (NM_000127)	c.1469delT	p.L490Rfs*9	0; 0; 0	-	DM/P	Pathogenic (PVS1+PS2+PM2)
7	EXT1 (NM_000127)	c.1645_1673delCGTTTTCTGCCCTACGACAACATCATCAC	p.F550Rfs*6	0; 0; 0	-	Not indexed	Pathogenic (PVS1+PS2+PM2)

^*Three databases: 1000g2015aug_eas, https://www.internationalgenome.org/; ExAC_EAS, http://exac.broadinstitute.org; gnomAD_exome_EAS, http://gnomad.broadinstitute.org/; REVEL: A comprehensive tool for predicting pathogenicity of missense variation [PMID: 27666373]; HGMD: Human Gene Mutation Database (Professional Version 2019.4, http://www.hgmd.cf.ac.uk/ac/index.php); ClinVar: https://www.ncbi.nlm.nih.gov/clinvar/; DM: Disease causing mutation; P: Pathogenic; VUS: Variant of uncertain significance; B: benign; Pathogenicity rating: Based on the common guideline issued by the American Association of Medical Genetics and Genomics (ACMG) (Ref. 15; Richards et al., 2015).

According to verification results, co-separation of the variant and phenotype was found in all three familial HMO cases, while all the sporadic patients harbored de novo variants; and the specific carrying status was also shown in the pedigree diagrams in Figure 3. Additionally, Figure 3G indicated that the amino acid residue, EXT1: G339 affected by EXT1: c.1016G>A in Case 3, was maintained evolutionarily across species.

Discussion

In 1814, Alexis Boyer first depicted HMO [17]. HMO has different clinical presentations, such as limited range of motion (ROM), chronic pain syndromes, short stature, limb malformation, scoliosis or neurovascular change [18]. HMOs can also occur in bones during endochondral ossification, usually the long bones [2]. Comparatively, HMO at birth is uncommon, but it can occur in children and adolescents, and it stops growing after the maturation of epiphyseal plates.

In general, de novo pathogenic variants account for the cause of around 10% of HMO cases affected [12,21,22]. Yet in our study, de novo mutations occurred in three of the seven cases, and this higher rate may be related to the proportion bias caused by our small sample size. Apart from long bone epiphyses, particularly within lower extremities, less common on-set sites for HMO include the hands, ribs, scapula, vertebrae and pelvis [23]. The clinical indications of our patients were generally consistent with usual HMOs, but some relatively rare lesions also appeared, such as scapula involvement of Case 2, and Case 5 had a chondroma on his left ring finger. Current pathological results and follow-up showed that our patients did not have the rare malignant transformation, but this risk should not be ignored during the lifetime of all patients.

In this study, six patients carried EXT1 variations with various types, including one nosense, one missense, two splicing-site, one frameshift and one microdeletion leading to pre-termination of protein translation. The other patient, Case 2, carried a frameshift variant in EXT2. According to the ACMG guideline, they were all interpreted as “pathogenic” or “likely pathogenic” rating level (Table 2). Additionally, the evolutionary conservatism of EXT1: p.G339 residue affected by the c.1016G>A (p.G339D) variant strongly supports its pathogenicity. To some extent, this reflects the omnipotence of WES detection technology in identifying causative variations, as in our previous study [24]. Four of the seven variants identified in our study were novelly reported, which expands the mutation spectrum of HMO. In terms of the other three known pathogenic variations, they have been reported with association to the HMO phenotype at least one time [12,25-29]. For each familial HMO case (Case 3, 4, and 5), the couples still have a risk up to 50% of being affected in future pregnancies, so genetic counseling and interventional fertility options, like prenatal or pre-implantation diagnosis, should be considered [30]. Actually, based on our follow-up, two families (Case 4 and 5) have already decided on a reproductive intervention in their next pregnancies, i.e., pre-implantation diagnosis (PGD), and had already entered the cycle. For the other four sporadic cases, there may be parental gonad mosaicism, and the probability of subsequent pregnancy involvement may be about 1%, so it should not be taken lightly [29].

HS chains, which are composed of proteoglycans, have diverse physiological and developmental activities [31]. The HS backbone is formed under the action of 2 glycosyltransferases coded via genes EXT1 and EXT2 [32]. HS chains can be attached to the “core protein” to form one type of proteoglycans called HSPG. HSPGs can be exposed onto cell membrane surface, the pericellular and the ECM, which play key roles in the signal transduction of many molecules [33]. Each signaling protein contains a certain HS-binding domain for regulating skeletal growth and development [9,34]. Those critical signaling proteins include bone morphogenetic proteins (BMPs), fibroblast growth factor (FGF), Hedgehog and Wnt signaling proteins [9,31]. When the above pathways are severely interfered with, it will induce osteochondroma development [18].

Furthermore, one “two-hit” model adopted for developing exostosis has long been set up and elucidated, and the germline mutation and somatic mutation together cause EXT1/EXT2 functional losses or the later carcinogenesis [8,35,36]. The potential second hits are loss-of-heterozygosity (LOH), mutations or aneuploidy of additional genes [36-38]. As for second-hit somatic mutations, together with the subsequent second gene copy’s random inactivation, the synthesis of functional HS is terminated in cells. Consequently, EXT1/EXT2 gene mutations induce seriously impaired biosynthesis and elongation of HS chains, as presented by reduced HS level within ECM and surrounding cells [39]. HS impairs chondrocyte growth and differentiation factors, including IHH or those related to osteogenesis like FGF, BMPs, and Wnt signaling proteins [9,40-42]. Therefore, abnormal bone growth may take place since HS chains completely lose or are shortened by chondrocyte growth and differentiation.

This work is mainly limited by the small sample size as well as lack of in-depth functional experiments. The genotype-phenotype association cannot be fully established and analyzed. Further research based on the “second hits” theory may hold promise for the treatment of HMO.

In summary, this study recruited seven HMO cases, conducted a thorough clinical evaluation, and performed a definitive genetic diagnosis on them. Our results expanded HMO variation spectrum and laid a certain foundation for counseling and reproduction for affected families. Further studies are still required to better understand the nosogenesis and to facilitate to develop new therapeutic drugs on HMO.

Disclosure of conflict of interest

None.