Somatic Mosaic Mutations in IDH1 or IDH2 in Solitary and Ollier Disease Enchondromas and Chondrosarcomas

An ESUN Study Report
By Judith V.M.G. Bovée, MD, PhD and Jan Oosting, MD, PhD

Introduction

Chondrosarcoma is the second most frequent malignant bone tumor that forms cartilage. Adults in the third to sixth decade are affected.¹ Chondrosarcoma is unique amongst mesenchymal tumors since it has benign precursor lesions, may progress from low- to high grade malignant and may occur in a syndrome. Our previous studies have shown that cartilage tumors arising at the surface of bone should be regarded separately from those arising in the medulla (Table 1),² although their histology is very similar. Central chondrosarcoma is most common^1,3 and may be secondary to an enchondroma.⁴ Enchondromas usually present in the long bones in the third or fourth decade.⁴ Malignant transformation towards central chondrosarcoma occurs in <1% of solitary enchondromas.

While most enchondromas are solitary, some patients demonstrate multiple enchondromas with a marked unilateral predominance particularly affecting the limbs,⁵ known as Ollier disease or Maffucci syndrome, the latter of which combines multiple enchondromas with soft tissue vascular lesions. The risk of malignant transformation is increased (10-35%).⁵ In 2002 a mutation was reported in the Parathyroid Hormone Receptor type I (PTHR1) gene in 2 of 6 patients with Ollier disease.⁶ However, in a large multi-institutional series of 31 patients we failed to detect any mutations.⁷ Moreover, three additional mutations were found in PTHR1 in tumors from 3 of 28 Ollier patients reducing the function of PTHR1 to ~70%.⁸ Therefore, PTHR1 may contribute to Ollier disease in 8% of the cases, but is probably not causative.

Approximately 15%³ of chondrosarcomas (secondary (peripheral) chondrosarcoma) are located at the surface of bone and result from malignant transformation of osteochondroma.¹ Osteochondroma is a mostly asymptomatic cartilage capped bony outgrowth representing ~35% of all primary benign bone tumors.⁹ Malignant transformation towards peripheral chondrosarcoma is low (<1% of cases). Although peripheral chondrosarcoma is far less common, many of the active signalling pathways in chondrosarcoma have been identified after the initial genetic event was elucidated. Osteochondromas can occur in multiple osteochondromas (MO) syndrome, an autosomal dominant disorder characterized by mutations in EXT1 or EXT2. The EXT gene products are involved in heparan sulphate biosynthesis, essential for the diffusion of hedgehog proteins. Both enchondromas and osteochondromas were shown to have active hedgehog signalling.^10,11 However, EXT is not involved in central chondrosarcoma.¹²

Purpose and Aims

We expected that we could further unravel the initiating event for enchondroma and central chondrosarcoma by studying Ollier disease as a model, similar to Multiple Osteochondromas elucidating many of the pathways involved in peripheral chondrosarcoma (Table 1).

Several genetic screens have been performed in Ollier disease and Maffucci syndrome in order to find a causative gene.6-8,13-14 Mutations in PTH1R, involved in enchondral bone formation, have been found in ~ 8% of patients with Ollier disease,6-8 but not in patients with Maffucci syndrome.14

Patients with Ollier disease and Maffucci syndrome have an increased incidence of gliomas.^15,16 Mutations in IDH1 and, more rarely IDH2, have been reported in (predominantly lower grade) gliomas and in secondary glioblastomas.^17,18 Interestingly, IDH1 and IDH2 mutations were recently also reported in solitary central and periosteal enchondromas and chondrosarcomas, including few tumors from patients with enchondromatosis.¹⁹ IDH catalyzes the oxidative decarboxylation of isocitrate to α-ketoglutarate (α-KG) and reduces NAD(P+) to NAD(P)H. Both isoforms are involved in lipid metabolism and in the Krebs cycle, with IDH1 being localized in the cytoplasm and IDH2 in the mitochondria.²⁰ Mutations in IDH1 and IDH2 in gliomas are mutually exclusive with rare exceptions, suggesting that a mutation in either of these two isoforms is sufficient to confer growth advantage and cell survival.²¹  Based on the increased incidence of gliomas and on the recently reported occurrence of IDH1 or IDH2 mutations in solitary as well as enchondromatosis related enchondromas and chondrosarcomas,¹⁹ we assessed whether IDH1 or IDH2 mutations may cause enchondroma and spindle cell hemangioma formation in Ollier disease and Maffucci syndrome. Since these disorders are not inherited and the enchondromas are often unilateral, we further hypothesized that mutations may occur in a somatic mosaic fashion.

Results

We collected one of the largest reported series worldwide: fresh frozen tumor tissues (n = 60) of 44 patients with multiple cartilage tumors (36 patients with Ollier disease and 8 patients with Maffucci syndrome) were collected from EuroBoNet consortium¹³ and the Laboratory of Human Molecular Genetics at the de Duve Institute, UCL (Brussels, Belgium). In addition, paraffin embedded tumor tissues (n = 15) from 12 patients were obtained from the files of the Children’s Hospital (Boston, USA) and the EMSOS consortium.

Somatic mosaic IDH1 and IDH2 mutations in Ollier disease and Maffucci syndrome

Sequence analysis of hotspot positions in IDH1, IDH2 using lesional tissue from 43 patients with Ollier disease revealed that heterozygous R132C IDH1, R132H IDH1 or R172S IDH2 mutations were present in 33 patients (78%).²² In Maffucci syndrome, 7 out of 13 patients (54%) carried R132C IDH1 mutations.²² Mutations were absent in DNA from patients’ blood, muscle or saliva. Using a more sensitive approach (hydrolysis probes assay capable of detecting as low as 1% of mutant allele)^23,24 for the detection of R132C and R132H IDH1 mutations we identified an additional 7 tumors with lower levels of R132C or R132H IDH1 mutations. In total, 35 out of 43 (81%) and 10 of 13 (77%) patients with Ollier disease and Maffucci syndrome, respectively, showed IDH1 or IDH2 mutations. When tumors of different anatomic sites were available for a patient, these demonstrated the same mutation.

Using immunohistochemistry to detect the R132H mutation we demonstrated a mixture of normal and neoplastic (mutated) cells within the tumor (intraneoplastic mosaicism) as well as a very low level of mutated cells in normal tissue (somatic mosaicism).

IDH1 or IDH2 mutations in solitary central, periosteal and dedifferentiated cartilaginous tumors and cell lines

Heterozygous IDH1 or IDH2 mutations were found in some of the solitary cartilage tumors examined using Sanger sequencing, including 40 of 101 (40%) solitary central tumors, 7 of 13 (54%) dedifferentiated chondrosarcomas and 3 of 3 periosteal chondrosarcomas. The R132C mutation was most frequent. Four cell lines derived from solitary central chondrosarcomas also showed mutations. IDH1 or IDH2 mutations were more frequently found in solitary central tumors located in hands and feet (11 out of 14) versus those located in long and flat bones (28 out of 84) (p=0.006, Pearson Chi-Square test), which was also reported previously.¹⁹ This correlation was absent in Ollier disease (20 out of 22 versus 28 out of 34, p=0.5, Pearson Chi-Square test).  While in gliomas, similar mutations in IDH1 or IDH2 predict a favorable outcome,²⁵ we found no significant prognostic value of these mutations for the patients with solitary central cartilaginous tumors. In multivariate survival analysis, IDH1 or IDH2 mutations remained insignificant for patient survival.

Hypermethylation in enchondromas with IDH1 mutations

Since in other tumor types the presence of an IDH1 mutation is strongly associated with hypermethylation,^26,27 we assessed whether there was a difference in methylation pattern of enchondromas with (n = 8) and without (n = 4) IDH1 mutations detactable at Sanger sequencing. Enchondromas with IDH mutations were hypermethylated.²²

Summary and future directions

We examined the presence of IDH1 and IDH2 mutations in patients with Ollier disease and Maffucci syndrome because they have an increased incidence of gliomas, which are associated with IDH1 or IDH2 mutations, and because mutations in these genes have recently been reported in solitary and enchondromatosis related enchondromas and central chondrosarcomas.¹⁹ We found that 81% of patients with Ollier disease and 77% of patients with Maffucci syndrome carry IDH1 or rarely, IDH2 mutations in their tumors. Furthermore, we found evidence of somatic mosaicism within enchondromas and in surrounding tissues.

The high mutation frequency in enchondromas and the fact that they are early events suggest a causal rather than a bystander role for IDH1 or IDH2 mutations in tumorigenesis. In gliomas, mutant IDH1 or IDH2 leads to gain of function by producing 2-hydroxyglutarate (2HG), a structural analogue of α-KG, and by ultimately reducing α-KG production.²⁸ In cartilage tumors increased D2HG production as a consequence of IDH mutations was shown, which was published simultaneously with our study.²⁹ We show that in enchondromas IDH1 mutations are associated with hypermethylation.

Future studies should reveal whether chondrosarcomas arising in enchondromas are still dependent on IDH mutations, which would mean that IDH could be a target for treatment, or that other genetic alterations are required for chondrosarcoma development in enchondroma. By elucidating the initiating genetic event driving chondrosarcoma development the pathways crucial for tumor growth may be identified and may serve as target for preventive or therapeutic strategies against chondrosarcoma. Interestingly, four of eight solitary chondrosarcoma cell lines carry different types of mutations in IDH1 or IDH2. Cell lines with IDH1 or IDH2 mutations are not widely available. This provides us with an excellent model to study dependency on IDH mutations. Moreover, within the current project we started whole exome sequencing of five chondrosarcomas arising in patients with Ollier disease, known to harbour somatic mosaic IDH1 mutations, to identify additional genetic changes that may cause or contribute to chondrosarcoma development. The whole exome sequencing results are currently being analyzed. The initial analysis shows that the quality of the sequencing is good. The IDH1 mutation has been recovered in all cases. In all tumors approximately 2000 somatic mutations have been identified. Thirteen genes contain mutations in all tumors and 57 genes contain mutations in five of the tumors.  The next six months will be used to identify recurring changes based on genomic position, gene, or functional pathway.

V9N3 ESUN Copyright © 2012 Liddy Shriver Sarcoma Initiative.

Deep Exome Sequencing to Identify the Gene Causing Solitary and Ollier Chondrosarcomas

By Judith V.M.G. Bovée, MD, PhD and Jan Oosting, MD, PhD

Background

Chondrosarcoma is the second most frequent malignant bone tumour that forms cartilage. Adults in the third to sixth decade are affected (1). Chondrosarcoma is unique amongst mesenchymal tumours since it has benign precursor lesions, may progress from low- to high grade malignant and may occur in a syndrome. Our previous studies have shown that cartilage tumours arising at the surface of bone should be regarded separately from those arising in the medulla (2) (Table 1), although their histology is very similar. Central chondrosarcoma is most common (1;3) and may be secondary to an enchondroma (4). Enchondromas usually present in the long bones in the third or fourth decade (4). Malignant transformation towards central chondrosarcoma occurs in <1% of solitary enchondromas.

While most enchondromas are solitary, some patients demonstrate multiple enchondromas with a marked unilateral predominance particularly affecting the limbs (5;6), known as Ollier disease or Maffucci syndrome, the latter of which combines multiple enchondromas with soft tissue vascular lesions. The risk of malignant transformation is increased (10-35%) (5). In 2002 a mutation was reported in the Parathyroid Hormone Receptor type I (PTHR1) gene in 2 of 6 patients with Ollier disease (7). However, in a large multi-institutional series of 31 patients we failed to detect any mutations (8). Moreover, three additional mutations were found in PTHR1 in tumors from 3 of 28 Ollier patients reducing the function of PTHR1 to ~70% (9). Therefore, PTHR1 may contribute to Ollier disease in 8% of the cases, but is probably not causative.

Approximately 15% (3) of chondrosarcomas (secondary (peripheral) chondrosarcoma) are located at the surface of bone and result from malignant transformation of osteochondroma (1). Osteochondroma is a mostly asymptomatic cartilage capped bony outgrowth representing ~35% of all primary benign bone tumours (10). Malignant transformation towards peripheral chondrosarcoma is low (<1% of cases). Although peripheral chondrosarcoma is far less common, many of the active signalling pathways in chondrosarcoma have been identified after the initial genetic event was elucidated. Osteochondromas can occur in multiple osteochondromas (MO) syndrome, an autosomal dominant disorder characterized by mutations in EXT1 or EXT2. The EXT gene products are involved in heparan sulphate biosynthesis, essential for the diffusion of hedgehog proteins. Both enchondromas and osteochondromas were shown to have active hedgehog signalling (11;12). However, EXT is not involved in central chondrosarcoma (13). We expect to further unravel the initiating event for enchondroma and central chondrosarcoma by studying Ollier disease as a model, similar to Multiple Osteochondromas elucidating many of the pathways involved in peripheral chondrosarcoma (table 1).

Previous Studies by our Group

In the past, we could not detect any significant differences between solitary and Ollier disease related chondrosarcomas at the molecular level in a pilot study using cDNA microarray and high-resolution array-CGH (14;15). Molecules involved in IHH/PTHLH signalling are expressed at similar levels (8;16). Therefore, solitary and Ollier disease-related chondrosarcomas seem similar justifying the study of Ollier disease as a model to understand central chondrosarcoma development. To study the genetic background of Ollier disease we performed SNP analysis using Affymetrix SNP6.0 on 15 enchondromas and 24 chondrosarcomas of different grades from 30 Ollier patients and normal DNA from 12 Ollier patients for paired comparison collected within the EuroBoNeT Network of Excellence (17). We studied tumour tissue since we hypothesized that Ollier disease is a mosaic condition, since it affects multiple bones with an often unilateral predominance. All samples were divided into three groups: normals, enchondromas and chondrosarcomas. Non-recurrent EC specific copy number alterations were found at FAM86D, PRKG1andANKS1B. LOH with copy number loss of chromosome 6 was found in two ECs from two unrelated Ollier patients. One of these patients also had LOH at chromosome 3. However, no common genomic alterations were found for all ECs. Using an integration approach of SNP and expression array we identified loss as well as down regulation of POU5F1 and gain as well as up regulation of NIPBL. None of these candidate regions were affected in more than two Ollier patients suggesting these changes to be random secondary events in EC development. An increased number of genetic alterations and LOH were found in Ollier CS which mainly involves chromosomes 9p, 6q, 5q and 3p. In summary the absence of common copy number variations or loss of heterozygosity suggests that instead point mutations or epigenetic mechanisms seem to play a role in the origin of Ollier disease (17). Mutation analysis revealed absence of the reported G121E, A122T, R150C and R255H variations in PTHR1 in our series.

Hypotheses

1. Ollier disease is a mosaic condition.

The fact that in most patients only certain parts of the body are affected and that the disease is not inherited from the parents suggest that the as far unknown etiological factor affects the limb buds during early foetal life. An early postzygotic mutation resulting in asymmetric involvement of skeletal structures can be expected, as was shown for polyostotic fibrous dysplasia (18). This putative mosaicism indicates that the cause of Ollier disease is to be found by studying tumour tissue.

2. Ollier disease is caused by a point mutation.

Our preliminary results show absence of loss of heterozygosity in enchondromas from Ollier disease which suggests that the classical tumour suppressor model as found in Multiple Osteochondromas (19) does not apply to Ollier disease. Therefore, similar to fibrous dysplasia a single somatic point mutation probably causes the disease.

3. The same gene(s) is / are responsible for Ollier disease-related and solitary cartilage tumours.

We showed Ollier disease-related and solitary tumours to be similar at the DNA (20) and expression level (8;14;16). We therefore expect that the same gene is involved, in which the mutation in solitary chondrosarcomas occurs somatically, later in life. Thus, genes involved in central chondrosarcomagenesis are expected to be found by studying Ollier disease.

Purpose

The aim of the proposed studies is to identify the gene(s) causing solitary as well as Ollier disease related enchondroma and chondrosarcoma.

Research Plan and Experimental Design

We will start by deep exome sequencing of at least 3 Ollier related hand enchondromas and patient matched normal tissue, followed by three more pairs (see below). Normal DNA was isolated from blood or saliva, obtained after informed consent. Samples were coded and all procedures were performed according to the ethical guidelines "Code for Proper Secondary Use of Human Tissue in The Netherlands" (Dutch Federation of medical Scientific Societies). Patients are very well characterized (see preliminary results). We choose to start with hand enchondromas to have a homogeneous group, since our preliminary data show a difference in clinical behaviour in addition to minor differences in copy number alterations between small – and long bones enchondromas. Since we expect Ollier disease to be a mosaic condition, comparable to polyostotic fibrous dysplasia, an early somatic mutation is expected with mutated cells spread throughout the body. Tumor tissue however, is expected to be enriched for the cells containing the mutation. A coverage of 30-50 deep is needed to find variants in the mosaic condition (21). This is why exome sequencing is favored over whole genome sequencing.

This technique requires 1 microgram of DNA which is available for most of the tumours previously analyzed by SNP-array (see preliminary results). It is estimated that within a tumor several hundred changes occur in the DNA (22). Results of the tumour will therefore be compared to germline DNA from the same patient. In addition, it is necessary to combine the results of a number of patients or tumors to identify the driver changes and filter out bystander changes. Therefore, after the initial bio-informatical analysis of the 3 tumor-normal pairs an evaluation is made. If the candidate list is already very short, a set of 3 additional long bone enchondromas will be analyzed to find whether the difference in biological behavior can be attributed to genetic differences. If the candidate list is still long, 3 more hand tumors will be analyzed in order to strengthen the evidence. Since we hypothesize the disease to be germline mosaic, the inclusion of surrounding normal tissue will in this case be considered, if available.

The experiments will be performed at the Leiden Genome Technology Centre (LGTC) using the Illumina Genome Analyzer II or HiSeq 2000. Candidate genes will be verified using DNA of the larger cohort that is available and well documented (frozen tissue n=~40, paraffin n=~70) see preliminary results), as well as of our series of solitary central chondrosarcomas (13;23). In addition, expression of the (candidate) gene(s) can be studied using available tissue microarrays (TMA) for Ollier tumours. A tissue microarray was previously constructed containing 66 tumours from 44 patients with Ollier disease, 24 solitary tumours, and 12 tumours from 7 patients with Maffucci syndrome. In addition, recently constructed TMAs containing 100 solitary conventional central chondrosarcomas can be used, in case antibodies are available for immunohistochemistry. This will be the starting point for additional functional studies.

Impact and Clinical Relevance

Identification of the gene causing Ollier disease is expected to also play a role in the origin of the most common chondrosarcoma subtype, conventional central chondrosarcoma. By elucidating the initial causative event we will better understand central chondrosarcoma development, similar to recent advances for EXT and its role in peripheral cartilaginous tumours (24). In addition, identification of the exact pathways causing the disease may provide new targets for preventive (in case of Ollier disease) or therapeutic (in case of conventional chondrosarcoma) strategies. Since chondrosarcomas are resistant to conventional chemo- and radiotherapy there is nothing curative to offer patients with irresectable or metastatic disease, and new treatment strategies are therefore urgently required.

Conclusion

Using a whole exome sequencing approach of enchondromas in patients affected by Ollier disease, we intend to elucidate the underlying genetic event causing the disease which we expect to be mosaic. Since Ollier related and solitary chondrosarcomas share many similarities, it is expected that the same gene underlies the development of solitary central chondrosarcoma. By elucidating the initiating genetic event driving chondrosarcoma development the pathways crucial for tumour growth may be identified and may serve as target for preventive or therapeutic strategies against chondrosarcoma.

V8N1 ESUN Copyright © 2011 Liddy Shriver Sarcoma Initiative.

Study Funding

The Liddy Shriver Sarcoma Initiative announced the funding of this grant in February 2011. The study was made possible by generous donations made in honor of Amy Kropp, Gavin Kiener, Brendon Martin, Mary Elizabeth Weigand, Chelsea Byers, Elizabeth Munroz, Sara Alan and Carly Laverty; and by generous donations made in memory of Kenny Allen, Scott Stafford, Tom Hand, and Lorelynn Roat.

Copyright © 2011 Liddy Shriver Sarcoma Initiative.