Ollier disease and Maffucci syndrome are caused by somatic mosaic mutations of IDH1 and IDH2

Abstract

Ollier disease and Maffucci syndrome are characterized by multiple central cartilaginous tumors that are accompanied by soft tissue hemangiomas in Maffucci syndrome. We show that in 37 of 40 individuals with these syndromes, at least one tumor has a mutation in isocitrate dehydrogenase 1 (IDH1) or in IDH2, 65% of which result in a R132C substitution in the protein. In 18 of 19 individuals with more than one tumor analyzed, all tumors from a given individual shared the same IDH1 mutation affecting Arg132. In 2 of 12 subjects, a low level of mutated DNA was identified in non-neoplastic tissue. The levels of the metabolite 2HG were measured in a series of central cartilaginous and vascular tumors, including samples from syndromic and nonsyndromic subjects, and these levels correlated strongly with the presence of IDH1 mutations. The findings are compatible with a model in which IDH1 or IDH2 mutations represent early post-zygotic occurrences in individuals with these syndromes. [PUBLICATION ABSTRACT]

Ollier disease and Maffucci syndrome are characterized by multiple central cartilaginous tumors that are accompanied by soft tissue hemangiomas in Maffucci syndrome. We show that in 37 of 40 individuals with these syndromes, at least one tumor has a mutation in isocitrate dehydrogenase 1 (IDH1) or in IDH2, 65% of which result in a R132C substitution in the protein. In 18 of 19 individuals with more than one tumor analyzed, all tumors from a given individual shared the same IDH1 mutation affecting Arg132. In 2 of 12 subjects, a low level of mutated DNA was identified in non-neoplastic tissue. The levels of the metabolite 2HG were measured in a series of central cartilaginous and vascular tumors, including samples from syndromic and nonsyndromic subjects, and these levels correlated strongly with the presence of IDH1 mutations. The findings are compatible with a model in which IDH1 or IDH2 mutations represent early post-zygotic occurrences in individuals with these syndromes.

Ollier disease (enchondromatosis) and Maffucci syndrome are rare, classically nonfamilial conditions that are characterized by multiple intramedullary (central) cartilaginous tumors. In Maffucci syndrome, tumors also occur as soft tissue hemangiomas, which are absent in Ollier disease. Both conditions generally present during childhood and, depending on the extent of skeletal involvement, can result in severe deformity of the affected tissues (Fig. 1)1. The cartilaginous tumors arise close to the growth plate and with physiological growth and increased age frequently extend along the diaphysis as clusters of tumors. The similarities of central cartilaginous tumors and growthplate tissue at a molecular level suggest that the tumors evolve as a consequence of a growth or differentiation disorder during physiological endochondral ossification2. Malignant transformation of chondromas and the development of secondary neoplasms, mainly gliomas and occasionally acute myeloid leukemia (AML), have been reported in Ollier disease and Maffucci syndrome, but the risk of these is unknown, and reports in the literature are biased toward cases with severe phenotypes3.

Mutations in IDH1 (NM_005896.2) and IDH2 (NM_002168.2) are found in up to 85% of high-grade gliomas and secondary glioblastomas, in ~16% of AMLs and occasionally in other neoplasms4. It has been reported that the mutant enzyme has neomorphic activity, catalyzing the reduction of α-ketoglutarate (α-KG) to D-2-hydroxyglutarate (2HG)5, which is an oncometabolite affecting the activity of α-KG-dependent dioxygenases. These events affect a number of cellular responses, including those controlling the epigenetic status of tumor cells, as demonstrated by global DNA hypermethylation in gliomas and AMLs harboring IDH1 and IDH2 mutations6,7.

We recently reported that IDH1 or IDH2 mutations, including those causing R132C (c.394C>T; NM_005896.2), R132S (c.394C>A; NM_005896.2), R132G (c.394C>G; NM_005896.2), R132H (c.395G>A; NM_005896.2), R132L (c.395G>T; NM_005896.2) and R172S (c.516G>C; NM_002168.2) amino acid substitutions, occur in 56-70% of central and periosteal cartilaginous tumors, 90% of which were solitary8. Previously, mutations in PTH1R (encoding parathyroid hormone receptor 1) were reported in up to 10% of individuals with Ollier disease2,9. In this study, we tested whether the cartilaginous and vascular tumors in subjects with Ollier disease or Maffucci syndrome result from early post-zygotic somatic events in IDH1 and IDH2, which would explain the mosaic pattern of disease distribution.

Seventy-four tumors from 40 individuals (32 with Ollier disease, 8 with Maffucci syndrome) were analyzed for mutations in IDH1 (altering Arg132) and IDH2 (altering Arg140 and Arg172) (Supplementary Table 1). A large proportion (90.5%) of the tumors harbored one of these mutations: 62/68 cartilaginous tumors had an IDH1 mutation and 1 tumor had an IDH2 mutation (without a distinguishing phenotype). Samples from multiple tumors (range 2-6, mean 2.8 per subject) were available from 19 of the 40 subjects. Of these subjects, 15 individuals were found to have the same mutation in each of their tumor samples that were examined (Table 1). Tumors from three individuals (two tumors each) had wild-type IDH1 and IDH2 sequences (subjects 2, 10 and 25). One tumor in subject 21 harbored a mutation causing an R132S substitution, whereas the second tumor had wild-type sequences. Sequencing all coding regions of IDH1, IDH2 and PTH1R for the three tumors with wild-type sequences for IDH1 and IDH2 from which frozen tissue was available for such analysis did not reveal mutations. The rare IDH1, IDH2 and PTH1R mutations previously reported in these genes were not detected in the remaining four paraffin-embedded tumors with wild-type sequences (Supplementary Table 1). Additional sequencing of 12 of the 62 tumors with IDH1 mutations did not reveal common PTH1R mutations.

Given the frequency of somatic mutations and wild-type sequences in the tumors (Table 2), the probability that a concordant pattern of sequences encoding an Arg132 substitution would occur by chance in tumor sets from 18 of 19 individuals is less than 1/100,000 (P < 0.00001). This finding argues in favor of the mutations representing early post-zygotic events that result in a mosaic pattern of disease distribution, and this idea is supported by the presence of mutations in low-grade tumors and in tumors of two cell lineages and by the occurrence of an IDH1 mutations encoding the same Arg132 alteration in six tumors from a single individual (one hemangioma, five cartilaginous tumors). Detection of a mutation in the glioma and in cartilaginous tumors in subject 30 with Maffucci syndrome, the only individual with a brain tumor (Fig. 1a), would have strengthened the evidence for somatic mosaicism. However, this analysis was not possible, as only a hemangioma (carrying an R132C substitution) was available for testing.

Mutations in a variety of somatic and sometimes germline tissues are observed in somatic mosaicism. Because the individuals in this study had no relevant family history of Ollier disease or Maffucci syndrome when such history was available (Supplementary Table 1), the detection of the same mutations in both nonlesional tissue and tumors from 2 of 12 individuals (subject 1, bone marrow; subject 6, blood), using a custom-made Taqman assay and MassARRAY but not by capillary sequencing, was taken as evidence that Ollier disease and Maffucci syndrome represent somatic mosaic disorders (Supplementary Table 2). The inability to detect mutations in nonlesional tissue in 10 of the 12 individuals may potentially be explained by there being extremely low numbers of mutant cells present in the hematopoietic cells analyzed, by the derivation of hematopoietic cells from a different origin than the cell that acquired the genetic alteration in early development or by the possible occurrence of mutations at a stage in development when the myeloid lineage could be affected by elevated 2HG levels. In this last scenario, the mutation in IDH1 or IDH2 may have been selected against, thereby contributing to restriction of the mutation to precursor cells, including chondrocyte precursors, in the majority of individuals with the syndromes. Analysis of a wide variety of nonneoplastic tissue will resolve this question, but the rarity of the disease makes this undertaking difficult.

As previously reported for gliomas and AML6, we found a strong correlation between the presence of mutations and high levels of 2HG and between the absence of mutations and low 2HG levels in a series of central cartilaginous tumors and one hemangioma derived from subjects with Ollier disease or Maffucci syndrome and from those solitary neoplasms (P < 0.0001) (Fig. 2, Table 1 and Supplementary Table 3). However, it was notable that two tumors with wild-type sequences from a subject with Maffucci syndrome had high levels of 2HG, suggesting that in rare cases there may be mutations in other genes, such as d- and l-2-hydroxyglutarate dehydrogenases, malate dehydrogenase and possibly others, that result in high levels of 2HG10. An alternative interpretation is that only a small fraction of tumorbearing cells exist in these neoplasms with wild-type sequences11, but this theory is difficult to reconcile with the elevated 2HG levels in tumors from subject 10 (Fig. 2). In contrast, the third tumor with wild-type sequences, from a subject with multiple tumors (subject 2), had low levels of 2HG, and therefore it may be that a large population of cells lacking the mutation can mask the effects of IDH1 or IDH2 mutations (Fig. 2, Table 1 and Supplementary Table 3).

The finding that one subject had a cartilaginous tumor with an IDH1 mutation encoding an R132S substitution and another tumor that had the wild-type IDH1 sequence raises the possibility that circulating 2HG derived from the mutant tumor can stimulate the growth of second-site tumors in a paracrine manner. However, low cellularity in the paraffin-embedded material is a likely alternative explanation for this observation.

In summary, we have demonstrated that IDH1 mutations are common events in tumors in both Ollier disease and Maffucci syndrome, providing genetic evidence that the diseases are related. We have also confirmed that PTH1R mutations are exceptionally rare in these syndromes3. Evidence is provided for a model in which IDH1 mutations are early post-zygotic events in individuals with these syndromes, implying that the mutations are required for tumorigenesis. Our findings also highlight the phenotypic spectra seen in Ollier disease and Maffucci syndrome, underscoring the observation that hands and feet are not always involved (Supplementary Table 1).

This study sets the stage for further investigation into the pathogenesis of Ollier disease and Maffucci syndrome. Future experiments should clarify whether 2HG mediates widespread epigenetic control of gene expression in a subset of cartilaginous and vascular tumors as it does in other tumor types6, and exome analysis should not only reveal the mutations that bring about high levels of 2HG in the absence of IDH1 or IDH2 mutations in these tumors but also identify other genetic events involved in these neoplasms. It will be of interest to discover the functional links between IDH1 and IDH2 and genes involved in the development of cartilaginous tumors in other syndromes (EXT1, EXT2, PTH1R, PTPN11 and ACP5)3.

METHODS

Methods and any associated references are available in the online version of the paper at http://www.nature.com/naturegenetics/.

Note: Supplementary information is available on the Nature Genetics website.

ACKNOWLEDGMENTS

We are grateful to the individuals who participated in the research and to the clinicians and support staff in the London Sarcoma Service involved in their care. The research was funded by Skeletal Cancer Action Trust (SCAT) UK, The Bone Cancer Research Trust UK and The Wellcome Trust (WT077012). This research was part of the Royal National Othopaedic Hospital Musculoskeletal Research Programme and Biobank and the University College London Hospital and University College London Comprehensive Biomedical Research Programme.

AUTHOR CONTRIBUTIONS

A.M.F., M.F.A. and A.F. conceived of the project. A.M.F., M.F.A., A.F., D.H. and N.P. planned the experiments. D.H., M.E., N.P., F. Berisha, S.L., C.L.G. and R.E.G. performed the experiments. F. Bonar, R.T., S.M., A.M.F., M.F.A. and W.A. reviewed the histopathology and selected and provided the samples. V.R.F. and K.S.S. performed 2HG measurements. A.M.F., M.F.A. and S.D. wrote the manuscript. P.C. performed the statistical analysis. All authors reviewed the manuscript.

COMPETING FINANCIAL INTERESTS

The authors declare no competing financial interests.

Published online at http://www.nature.com/naturegenetics/.

Reprints and permissions information is available online at http://www.nature.com/ reprints/index.html.

Sidebar

ONLINE METHODS

Samples. Tissue samples were obtained from the archives of the Royal National Orthopaedic Hospital, UK, Douglass Hanly Moir Pathology, Sydney, Australia and the School of Medicine, The University of Sydney. Approval for these studies was obtained from the Institutional Review Boards of each of these institutions. Informed written consent was obtained from all participants. Each tumor sample was obtained from a different bone or soft tissue lesion (hemangioma).

DNA was extracted from tumor-rich areas of frozen and or paraffin- embedded tumor samples and from non-neoplastic tissue from the same subjects, when these were available. Tumor and non-neoplastic paraffin-embedded samples were analyzed for mutations in IDH1 encoding the Arg132 substitution by at least two of the following techniques: capillary Sanger sequencing, TaqMan, PCR followed by restriction digestion and high-throughput MassARRAY assay using the Sequenom platform as previously described8. The mutations in IDH2 encoding the Arg172 and Arg140 alterations and mutations in PTH1R were analyzed by capillary Sanger sequencing (see Supplementary Table 1 for details). Non-neoplastic frozen tissue and blood were analyzed by TaqMan assay and capillary Sanger sequencing. The primer sequences used in these assays are provided in Supplementary Table 4.

Immunohistochemistry. Immunohistochemistry was performed using the Leica Bond-Max Autostainer (Leica Microsystems) with an antibody recognizing IDH1 R132H (clone H09, Dianova; 1:100 dilution, 20 min of citrate buffer pre-treatment).

Capillary Sanger sequencing. A touchdown PCR protocol was employed to amplify the regions of IDH1 and IDH2 that would contain mutations (as previously described)8, and PCR products were purified using the DNA Clean & Concentrator kit (Zymo Research). DNA sequencing was performed at Scientific Support Services, University College London (London's Global University). Sequencing reactions were run using BigDye terminator v1.1 chemistry (Applied Biosystems) and were analyzed on an ABI 3730XL DNA Analyzer (Applied Biosystems).

Measurement of 2HG levels. Frozen tumor tissue was weighed and homogenized in ice-cold 80% (vol/vol) methanol using 100 mg of tissue per 1 ml of buffer. A Covaris Adaptive Focused Acoustics disruptor was used to generate tissue homogenates that were subsequently centrifuged at 4 °C for 15 min. Supernatants were diluted 1:1 with 100 µl deionized water containing an internal standard. All samples were then analyzed by liquid chromatography-mass spectrometry as previously described5.

Custom TaqMan assays. Genotyping to detect mutations encoding the R132C substitution was performed using Custom TaqMan SNP Genotyping Assays, which consisted of a mix of unlabeled PCR primers (sequences provided in Supplementary Table 4) and the TaqMan minor groove-binding group (MGB) probe (FAM and VIC dye-labeled; Applied Biosystems). PCR was performed using 12.5 µl of TaqMan Universal Master Mix, No AmpErase UNG (Applied Biosystems), 50 ng of DNA, 0.625 µl of Custom TaqMan SNP Genotyping Assay mix (40× stock solution) and 10.875 µl of sterile water to give a final reaction volume of 25 µl. Thermal cycling was performed as follows: initial denaturing at 95 °C for 10 min and 40 cycles at 92 °C for 15 s and 60 °C for 1 min.

Statistical analysis. To test whether the concordance of mutations within a sample was greater than expected by chance, we performed a permutation test in which the sample IDs and mutations were randomly sorted into 100,000 iterations. In this way, we determined the null distribution of the number of subjects showing concordant mutation status, taking into account the differing number of samples per subject and the differing prevalence of observed mutations.

References

1. Unni, K.K. Cartilaginous lesions of bone. J. Orthop. Sci. 6, 457-472 (2001).

2. Hopyan, S. et al. A mutant PTH/PTHrP type I receptor in enchondromatosis. Nat. Genet. 30, 306-310 (2002).

3. Pansuriya, T.C., Kroon, H.M. & Bovee, J.V. Enchondromatosis: insights on the different subtypes. Int. J. Clin. Exp. Pathol. 3, 557-569 (2010).

4. Dang, L., Jin, S. & Su, S.M. IDH mutations in glioma and acute myeloid leukemia. Trends Mol. Med. 16, 387-397 (2010).

5. Dang, L. et al. Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739-744 (2009).

6. Figueroa, M.E. et al. Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell 18, 553-567 (2010).

7. Noushmehr, H. et al. Identification of a CpG island methylator phenotype that defines a distinct subgroup of glioma. Cancer Cell 17, 510-522 (2010).

8. Amary, M.F. et al. IDH1 and IDH2 mutations are frequent events in central chondrosarcoma and central and periosteal chondromas but not in other mesenchymal tumours. J. Pathol. 224, 334-343 (2011).

9. Pansuriya, T.C. et al. Genome-wide analysis of Ollier disease: is it all in the genes? Orphanet J. Rare Dis. 6, 2 (2011).

10. Krell, D. et al. Screen for IDH1, IDH2, IDH3, D2HGDH and L2HGDH mutations in glioblastoma. PLoS ONE 6, e19868 (2011).

11. Preusser, M. et al. Value and limitations of immunohistochemistry and gene sequencing for detection of the IDH1-R132H mutation in diffuse glioma biopsy specimens. J. Neuropathol. Exp. Neurol. 70, 715-723 (2011).