Genomic DNA was isolated from the enchondroma tissues in right-hand finger bones and peripheral blood of the patient. Exome was captured and sequenced to 100× by BGI (Shenzhen Guangdong, China) using the Illumina HiSeq™ 2000 Sequencing Systems. The resulting sequences were first compared with published reference sequences (https://www.globus.org/) to exclude normal polymorphisms. Variants between enchondromas and peripheral blood DNA were analyzed by BGI to characterize somatic mutations with reads above 10% in the 100× readouts (22, 23). The mutations and frequencies were confirmed by randomly sequencing 21 PCR clones with a pMD^®18-T vector (Takara Biotechnology, Dalian, China).

ATDC5 and HUVECs cells were purchased from the Chinese Academy of Cell Bank. ATDC5 cells were cultured with DMEM/F12 medium at 37°C with 5% CO₂ in a humidified incubator. For chondrocyte induction, ATDC5 cells were cultured with CDM (chondrogenic differentiation medium) (Cyagen Biosciences Inc., China) at 37°C with 5% CO₂ in a humidified incubator. HUVECs were cultured in 2% FBS EBM-2 endothelial cell basic medium at 37°C for 24 h under a 5% CO₂ atmosphere.

The coding sequence of the target gene (IDH1, IDH1 with R132C mutation, ERC2 and ERC2 with L309I mutation) was PCR amplified from the GV287 target gene using primers with AgeI/AgeI overhangs and cloned into pTZ58 (Fermentas, Vilnus, Lithuania). The AgeI/AgeI fragment was then subcloned into pUbi (AgeI/AgeI) and pEGFP-C1 (Clontech, Mountain View, CA) (AgeI/AgeI) to generate Ubi-GENE-3FLAG-SV40-EGFP encoding plasmids, respectively. To produce lentivirus, the pBABE-puro plasmids were coinfected along with the helper plasmids into 293T cells, and the medium was harvested 36 h and 72 h after infection (24). ATDC5 or HUVEC infection was performed by incubating the cells in virus-enriched medium for 12 h, which included 4 μg/ml polybrene. Transduced cells were identified for EGFP expression under a fluorescence microscope.

The cells (mouse chondrogenic cell line ATDC5 or HUVECs) were seeded in 96-well plates at a density of 2×10⁴ cells/well. For chondrocyte induction, 10 μg/ml bovine insulin (Wako Pure Chemical, Osaka, Japan) was added after 12 h of culture under a 5% CO₂ atmosphere (21). After an additional 24 h of culture, 10 μl of WST-8 mixture (Dojindo, Shanghai, China) was added to each well and cultured for another 2 h. After washing, the cells were subjected to the measurement of absorbance under a microplate reader at a wavelength of 450 nm.

ATDC5 cells were seeded in a 24-well plate at a density of 8×10⁴ cells/well, and the cells were induced into chondrocytes as described above. A scratch was next created using a sterile yellow tip, the detached cells were removed, and the scratches were monitored for 48 h under culture conditions without any bovine insulin. Each set of experiments was performed in triplicate, and photographs were taken at the indicated time points.

ATDC5 cells or HUVECs were placed on the upper layer of culture medium inserted with a permeable membrane, and BD Matrigel (BD Biosciences, San Jose, CA) was placed below the cell permeable membrane. After 12 h of incubation, the ATDC5 cells that migrated through the membrane were stained with crystal violet solution and then counted under a light microscope, while the migrated HUVECs were analyzed under a fluorescence microscope.

The transduced ATDC5 cells were first induced to chondrocytes as described earlier and then synchronized to the G₀/G₁ phase by 24 h of serum starvation, and 10% FBS was added to the cultures. After another 24 h of culture, the cells were harvested for cell cycle analysis. Briefly, after washing, the cells were stained with propidium iodide, and the cell cycle distribution for each culture was analyzed by flow cytometry.

To examine tube formation, growth factor-reduced Matrigel (BD Bioscience, San Jose, CA) was placed in 96-well tissue culture plates (100 μL/well) and allowed to form a gel at 37°C for at least 30 min. HUVECs (2 x 10⁴ cells) after 24 h of transduction were added into each well and incubated in 2% FBS EBM-2 endothelial cell basic medium at 37°C for 24 h under a 5% CO₂ atmosphere. Endothelial tubes were then examined under a fluorescence microscope by inspecting the overall tube length and branch points.

The transduced HUVECs were seeded in a 24-well plate at a density of 2×10⁴ cells/well. After 12 h of culture, cell-permeant acetoxymethyl (AM) esters of X-Rhod-1 were loaded into the cultures. After another 4 h of culture, the cells were subjected to analysis of intracellular calcium concentration under an Olympus IX73 fluorescence microscope at 550 nm excitation and 600 nm emission. Images were taken at 100X magnification.

For pairwise comparisons, the data were analyzed using Student’s t-test. A comparison between multiple experimental groups was accomplished by one-way ANOVA using SPS 11.5 for windows. All experiments were conducted with at least 3 independent replications. All data are presented as the mean ± SEM. In both cases, p < 0.05 was considered to be statistically significant.

The patient here we report is a 26-year-old young woman who had typical multiple enchondromas but had no family members with disease history. Physical exams revealed that her 3-year-old daughter was also normal. The patient had abnormal protrusion in her right-hand fingers when she was 1 year old. Unfortunately, no medical examination was conducted at that time because of her family’s financial problems. The patient presented when she was 26 years old; by then, all of her limbs could still move but were affected by deformities. The right hand had dangerous tumor-like deformities, fingers in this hand displayed deformity shape of the visible nodular and soft spherical bulge in light blue ( Figure 1A ), and the right upper arm was shortening ( Figure 1B ), but her left hand showed healthy appearance ( Figure 1C ). The right knee showed strong varus, and the left knee displayed valgus; both feet and ankles displayed deformities (data not shown). Palpable subcutaneous nodules could be found around the bones of these deformity areas, but no tenderness was characterized.

Ultrasound examinations revealed that her liver, gallbladder, spleen, pancreas, pancreatic duct, kidneys, and bladder were healthy. MR examinations showed that the phalanges of the fingers are circular or ovoid, with multi-locular, well-defined chondromatoid lesions and the soft tissue around the phalanx of the finger presents dilatative, well-bounded hyperintensity on T2WI.( Figure 1D ). X-ray examinations demonstrated that the bones of the right palm, fingers of the right hand, right ulna, both radius ( Figure 1B ), humerus, scapulas, head of the right seventh rib ( Figure 1E ) and metatarsals of both feet ( Figure 1F ) were characterized by irregular swelling and morphological abnormalities. Cortical bones of these parts were markedly thinned, and multiple cystic-like lucent and dotted calcification areas can be noted around these parts. Bones in the cranial and maxillofacial regions, however, had no apparent abnormalities ( Figure 1G ). These asymmetrically distributed enchondroma bone destruction and surrounding soft tissue hemangioma malformations are typical imaging features of Maffucci’s syndrome.

These lesions seriously affected the function of the patient’s hands, and the patient was very concerned that these lesions may be malignant. In order to further clarify the disease and nature, biopsy was necessary. Histological analysis indicated typical enchondroma changes in sections originating from right finger bones ( Figures 2A, B ), and cavernous hemangioma changes were noted in the enchondroma tissues from the same finger ( Figures 2C, D ).

According to these typical clinical manifestations, radiological data, and histological analysis by the authoritative pathologist, the diagnosis of Maffucci’s syndrome was clear.

Exome sequencing was next conducted using the patient’s DNA samples originating from her peripheral blood and enchondroma tissues with 100x coverage. Comparative analysis of her enchondroma exome sequences with the sequences of 1000 standard human specimens in the public database Globus (https://www.globus.org/) revealed more than 65,000 variations, more than 99% of which are likely DNA polymorphisms between individuals. We thus first excluded those common polymorphisms identified through the public database and then aligned her enchondroma exome sequences with her peripheral blood exome sequences, via which we identified 90 mutations that cause amino acid changes. In general, somatic mutations in the enchondroma tissues are likely mosaic because of normal cell contamination or very low frequencies (only a proportion of tumor cells carry the same somatic mutation). By exclusion of those possibilities, the differences were limited to two mutations: the C394T mutation of IDH1 (NM_005896) in exon 4, which causes arginine to cysteine at position 132 (R132C), and the frequency of this mutation is 38.7% in total sequence readouts; and the C925A mutation of ERC2 (NM_015576) in exon 3, which mutates leucine to isoleucine at position 309 (L309I), and the ratio of this mutation is 23.8% in total sequence readouts.

To confirm the above sequencing data, we next PCR amplified the two regions (IDH1-c. C394T and ECR2-c. C925A) from her peripheral blood DNA and enchondroma DNA, respectively. The resulting PCR products were cloned into a TA vector, followed by sequencing analysis of 21 randomly selected clones. Indeed, these mutations were absent in the blood DNA, while approximately 33.3% of the clones were positive for the IDH1 R132C mutation ( Figure 3A ), and 19.0% of the clones were positive for the ECR2 L309I mutation ( Figure 3B ). To exclude the possibility that those mutations are also present in healthy individuals, we genotyped 500 healthy individuals, and none of the subjects detected those two mutations.

The foregoing case of Maffucci’s syndrome demonstrates that there may be two gene mutations of IDH1 and ERC2 in Maffucci’s syndrome. To demonstrate additional evidence of these two mutations in disease pathobiology, we genotyped 69 patients with Ollier disease using DNAs extracted from enchondroma tissues after dissection of embedded paraffin blocks. Excitingly, the IDH1 R132C mutation was detected in most of the patients with Ollier disease (51/69 patients), and the mean frequency of this mutation was 63.3% in total sequence readouts, but the ECR2 L309I mutation was absent in all of the patients with Ollier disease ( Figure 4 ). Unfortunately, we were unable to identify additional patients with Maffucci’s syndrome for the analysis of the ECR2 L309I mutation.

Given that the ECR2 L309I mutation is absent in patients with Ollier disease (with multiple chondromas only), while Maffucci’s syndrome is characterized by the coexistence of multiple chondromas and hemangiomas, we thus hypothesized that the IDH1 R132C mutation causes multiple chondromas, while the ERC2 L309I mutation is responsible for the development of hemangiomas. To test this hypothesis, we conducted studies in chondrocytes, in which we induced a mouse chondrogenic cell line, ATDC5, into chondrocytes after transduction with lentiviruses expressing wild-type IDH1 (IDH1-wt) or the R132C mutant (IDH1-mu) (22, 23). As expected, chondrocytes transduced with the IDH1-mu viruses exhibited significantly higher migration capacity ( Figure 5 ) and proliferation capability ( Figure 6A ). Cell cycle analysis revealed that the R132C mutation significantly promoted the G1-S phase transition ( Figures 6B, C ).

To demonstrate the impact of the ERC2 L309I mutation on the development of hemangiomas, we checked its role in angiogenesis, as hemangiomas are characterized by excessive vessel formation. For this purpose, we transduced HUVECs with lentiviruses expressing wild-type ERC2 (ERC2-wt) and the L309I mutant (ERC2-mu). Interestingly, the transduction of ERC2-mu viruses potently enhanced the capacity of HUVECs for angiogenesis, as evidenced by the higher capability for proliferation ( Figure 7A ), tubular formation ( Figures 7B, C ) and migration ( Figures 7D, E ). Collectively, these data suggest that the IDH1 R132C mutation combines with the ERC2 L309I mutation to cause the development of Maffucci’s syndrome.

Given that CAST/ERC2 has been noted to modulate neurotransmitter release in nerve terminals by regulating intracellular Ca²⁺ concentrations (24), we then examined the effect of the L309I mutation of ERC2 on the intracellular calcium concentration in HUVECs by staining with Fura-2/AM (Invitrogen, OR, USA), a dye used to measure intracellular free calcium. HUVECs transduced with ERC2-mu displayed significantly higher intracellular free calcium concentrations than cells transduced with ERC2-wt ( Figure 8 ).