Methylation and copy number profiling: emerging tools to differentiate osteoblastoma from malignant mimics?

Rearrangements of the transcription factors FOS and FOSB have recently been identified as the genetic driver event underlying osteoid osteoma and osteoblastoma. Nuclear overexpression of FOS and FOSB have since then emerged as a reliable surrogate marker despite limitations in specificity and sensitivity. Indeed, osteosarcoma can infrequently show nuclear FOS expression and a small fraction of osteoblastomas seem to arise independent of FOS/FOSB rearrangements. Acid decalcification and tissue preservation are additional factors that can negatively influence immunohistochemical testing and make diagnostic decision-making challenging in individual cases. Particularly aggressive appearing osteoblastomas, also referred to as epithelioid osteoblastomas, and osteoblastoma-like osteosarcoma can be difficult to distinguish, underlining the need for additional markers to support the diagnosis. Methylation and copy number profiling, a technique well established for the classification of brain tumors, might fill this gap. Here, we set out to comprehensively characterize a series of 77 osteoblastomas by immunohistochemistry, fluorescence in-situ hybridization as well as copy number and methylation profiling and compared our findings to histologic mimics. Our results show that osteoblastomas are uniformly characterized by flat copy number profiles that can add certainty in reaching the correct diagnosis. The methylation cluster formed by osteoblastomas, however, so far lacks specificity and can be misleading in individual cases.

Ligustilide inhibits the proliferation of human osteoblastoma MG63 cells through the TLR4-ERK pathway.

OBJECTIVE: To study the proapoptotic effect of ligustilide on osteoblastoma (OS) and the relative related molecular mechanism. METHODS AND MATERIALS: An MTT was used to examine the proliferation of OS cells, and Flow cytometry was used to analyze apoptosis and the cell cycle. Western blotting was used to detect the signaling pathway of apoptosis, and immunohistochemical (IH) staining was used to detect the apoptosis status of OS cells. A TLR4 inhibitor was used to study the effect of ligustilide on OS. RESULTS: Ligustilide inhibited OS cell proliferation but had no inhibitory effect on normal bone marrow cells. Flow cytometry results showed that ligustilide induced apoptosis in OS cells, and the cell cycle was arrested at the M/G2 phase. Western blot results showed that ERK, P53, P21, Caspase 9, Caspase 8 and Caspase 3 were all activated; cytochrome C and Bax increased; and Bcl-2 decreased when OS was treated with ligustilide. When an ERK or Caspase inhibitor was added to the culture medium, the apoptosis of OS cells decreased to some degree. When OS cells were pretreated with CLI-095, which is a TLR4 inhibitor, the percentage of apoptotic cells and cell cycle arrest were both reversed. IH results also showed that ligustilide induced apoptosis in OS cells, and the effect was blocked by the TLR4 inhibitor. CONCLUSION: Ligustilide selectively inhibited the proliferation of OS cells by inducing apoptosis, which possibly included endogenous and exogenous apoptosis through TLR4.

The one-in-all diagnostic value of 99mTc-MDP bone scan combining with single-photon emission tomography (SPECT)/CT imaging in spinal osteoblastoma.

BACKGROUND: Osteoblastoma (OB) is an intermediate lesion, which makes the accurate preoperative diagnosis very important. 99mTc-methylene diphosphonate (99mTc-MDP) bone scan and SPECT/CT imaging were evaluated for their diagnostic value in spinal OB. METHODS: This study was a retrospective analysis of patients with spinal OB lesions confirmed by pathology and diagnosed with bone scan and SPECT/CT for preoperative diagnosis from January 2008 to December 2018. The uptake levels of OB on planar bone scan were divided into low, medium, and high groups by visual assessment referring to the uptake of the normal rib, spine, and bladder. X-ray, CT, MRI, bone scan, and SPECT/CT imaging of the patients were analyzed for characteristics summary. RESULTS: Twenty-five patients were diagnosed for spinal OB (17 males and 8 females with a proportion of 2.1:1), and the average age was 26.8 ± 10.8 years (range 5~59). There were 8 lesions located in the cervical, 6 in the thoracic, and 11 in the lumbar vertebrae. Twenty-four lesions involved posterior elements, especially the pedicles (14/25). Symptoms were predominantly painful with a duration of 18.3 ± 13.9 months (range 0.5~60 months). The lesion size ranged from 9 to 35 mm. All the lesions were low to high uptake in the planar bone scan, and the percentages of low to high levels were 1 (4%), 8 (32%), and 16 (64%) cases. CONCLUSIONS: Spinal OB mainly involved the posterior area, and elderly patients should be considered as well. SPECT/CT combined the characteristics of bone uptake and anatomical features of bone tumors, proving its one-in-all diagnostic value for spinal OB and other osteogenic tumors.

FOS-ANKH and FOS-RUNX2 Fusion Genes in Osteoblastoma.

BACKGROUND/AIM: Osteoblastoma is a rare benign tumor of the bones in which recurrent rearrangements of FOS have been found. Our aim was to investigate two osteoblastomas for possible genetic aberrations. MATERIALS AND METHODS: Cytogenetic, RNA sequencing, and molecular analyses were performed. RESULTS: A FOS-ANKH transcript was found in the first tumor, whereas a FOS-RUNX2 was detected in the second. Exon 4 of FOS fused with sequences either from intron 1 of ANKH or intron 5 of RUNX2. The fusion events introduced a stop codon and removed sequences involved in the regulation of FOS. CONCLUSION: Rearrangements and fusions of FOS show similarities with those of HMGA2 (a feature of leiomyomas and lipomas) and CSF1 (tenosynovial giant cell tumors). The replacement of a 3'-untranslated region, controlling the gene's expression, by a new sequence is thus a common pathogenetic theme shared by FOS, HMGA2, and CSF1 in many benign connective tissue tumors.

The effect of strontium incorporation into sol-gel biomaterials on their protein adsorption and cell interactions.

It is known strontium can both inhibit the osteoclast formation and stimulate the osteoblast maturation, so biomaterials containing this element can favour bone structure stabilisation. The addition of Sr to biomaterials could affect their interactions with proteins and cells. Here, a silica-hybrid sol-gel network doped with different amounts of SrCl2 and applied as coatings on titanium discs was examined. in vitro analysis was performed to determine the potential effect of Sr in the coatings, showing enhanced gene expression of osteogenic markers (alkaline phosphatase and transforming growth factor-ß) in MC3T3-E1 incubated with Sr-doped biomaterials. The examination of inflammatory markers (tumour necrosis factor-α and interleukin 10) in RAW 264.7 macrophages revealed an anti-inflammatory potential of these materials. Proteins adsorbed onto the coatings incubated with human serum (3 h at 37 °C) were also analysed; mass spectrometry was used to characterise the proteins adhering to materials with different Sr content. Adding Sr to the coatings increased their affinity to APOE and VTNC proteins (associated with anti-inflammatory and osteogenic functions). Moreover, the proteins involved in coagulation processes, such as prothrombin, were more abundant on the coatings containing Sr than on the base sol-gel surfaces. Correlations between gene expression and proteomic results were also examined.

Recurrent rearrangements of FOS and FOSB define osteoblastoma.

The transcription factor FOS has long been implicated in the pathogenesis of bone tumours, following the discovery that the viral homologue, v-fos, caused osteosarcoma in laboratory mice. However, mutations of FOS have not been found in human bone-forming tumours. Here, we report recurrent rearrangement of FOS and its paralogue, FOSB, in the most common benign tumours of bone, osteoblastoma and osteoid osteoma. Combining whole-genome DNA and RNA sequences, we find rearrangement of FOS in five tumours and of FOSB in one tumour. Extending our findings into a cohort of 55 cases, using FISH and immunohistochemistry, provide evidence of ubiquitous mutation of FOS or FOSB in osteoblastoma and osteoid osteoma. Overall, our findings reveal a human bone tumour defined by mutations of FOS and FOSB.

microRNA-135b expression silences Ppm1e to provoke AMPK activation and inhibit osteoblastoma cell proliferation.

Forced-activation of AMP-activated protein kinase (AMPK) can possibly inhibit osteoblastoma cells. Here, we aim to provoke AMPK activation via microRNA silencing its phosphatase Ppm1e (protein phosphatase Mg2+/Mn2+-dependent 1e). We showed that microRNA-135b-5p ("miR-135b-5p"), the anti-Ppm1e microRNA, was significantly downregulated in human osteoblastoma tissues. It was correlated with Ppm1e upregulation and AMPKα1 de-phosphorylation. Forced-expression of miR-135b-5p in human osteoblastoma cells (MG-63 and U2OS lines) silenced Ppm1e, and induced a profound AMPKα1 phosphorylation (at Thr-172). Osteoblastoma cell proliferation was inhibited after miR-135b-5p expression. Intriguingly, Ppm1e shRNA knockdown similarly induced AMPKα1 phosphorylation, causing osteoblastoma cell proliferation. Reversely, AMPKα1 shRNA knockdown or dominant negative mutation almost abolished miR-135b-5p's actions in osteoblastoma cells. Further in vivo studies demonstrated that U2OS tumor growth in mice was dramatically inhibited after expressing miR-135b-5p or Ppm1e shRNA. Together, our results suggest that miR-135b-induced Ppm1e silence induces AMPK activation to inhibit osteoblastoma cell proliferation.

Sclerostin expression in skeletal sarcomas.

Sclerostin (SOST) is an extracellular Wnt signaling antagonist which negatively regulates bone mass. Despite this, the expression and function of SOST in skeletal tumors remain poorly described. Here, we first describe the immunohistochemical staining pattern of SOST across benign and malignant skeletal tumors with bone or cartilage matrix (n=68 primary tumors). Next, relative SOST expression was compared to markers of Wnt signaling activity and osteogenic differentiation across human osteosarcoma (OS) cell lines (n=7 cell lines examined). Results showed immunohistochemical detection of SOST in most bone-forming tumors (90.2%; 46/51) and all cartilage-forming tumors (100%; 17/17). Among OSs, variable intensity and distribution of SOST expression were observed, which highly correlated with the presence and degree of neoplastic bone. Patchy SOST expression was observed in cartilage-forming tumors, which did not distinguish between benign and malignant tumors or correlate with regional morphologic characteristics. Finally, SOST expression varied widely between OS cell lines, with more than 97-fold variation. Among OS cell lines, SOST expression positively correlated with the marker of osteogenic differentiation alkaline phosphatase and did not correlate well with markers of Wnt/ß-catenin signaling activity. In summary, SOST is frequently expressed in skeletal bone- and cartilage-forming tumors. The strong spatial correlation with bone formation and the in vitro expression patterns are in line with the known functions of SOST in nonneoplastic bone, as a feedback inhibitor on osteogenic differentiation. With anti-SOST as a potential therapy for osteoporosis in the near future, its basic biologic and phenotypic consequences in skeletal tumors should not be overlooked.

Bufotalin-induced apoptosis in osteoblastoma cells is associated with endoplasmic reticulum stress activation.

The search for novel and more efficient chemo-agents against malignant osteoblastoma is important. In this study, we examined the potential anti-osteoblastoma function of bufotalin, and studied the underlying mechanisms. Our results showed that bufotalin induced osteoblastoma cell death and apoptosis in dose- and time-dependent manners. Further, bufotalin induced endoplasmic reticulum (ER) stress activation in osteoblastoma cells, the latter was detected by the induction of C/EBP homologous protein (CHOP), phosphorylation of inositol-requiring enzyme 1 (IRE1) and PKR-like endoplasmic reticulum kinase (PERK), as well as caspase-12 activation. Conversely, the ER stress inhibitor salubrinal, the caspase-12 inhibitor z-ATAD-fmk as well as CHOP depletion by shRNA significantly inhibited bufotalin-induced osteoblastoma cell death and apoptosis. Finally, by using a mice xenograft model, we demonstrated that bufotalin inhibited U2OS osteoblastoma cell growth in vivo. In summary, our results suggest that ER stress contributes to bufotalin-induced apoptosis in osteoblastoma cells. Bufotalin might be investigated as a novel anti-osteoblastoma agent.

ß-catenin is a valuable marker for differential diagnosis of osteoblastoma and osteosarcoma.

Osteoblastoma (OB) and osteosarcoma (OS) are 2 bone tumors that predominantly affect young adults. The clinical management of OS differs significantly from that of OB, and thus, accurate diagnosis of OB and OS is critical in determining appropriate treatment modality. However, in certain cases, OS significantly overlaps with OB in clinical and radiographic characteristics, and therefore, the differential diagnosis of OB and OS can be difficult, especially when biopsy material is insufficient. To date, there have been few reports on markers for differential diagnosis of OB and OS. We have previously shown that the Wnt/ß-catenin pathway is inactivated in OS. In this study, we aimed to investigate whether the cellular distribution pattern of ß-catenin is a potential marker for the differential diagnosis of OB and OS. Immunohistochemical staining was studied in 17 OB samples (21 biopsies; 17 primary and 4 recurrent) and 37 OS samples with complete follow-up information. Moderate-to-strong nuclear ß-catenin staining was found in all OB specimens (17/17). In contrast, positive staining of ß-catenin was found in the cytoplasm and/or membrane but not the nucleus in all 32 cases of nonchondroblastic OS (32/32) and the classic OS component in chondroblastic OS (5/5). The only positive nuclear ß-catenin staining detected in OS biopsies was in chondroblastic OS cells (5/5). In summary, our results indicate that, in addition to conventional histopathologic evaluation, cellular distribution of ß-catenin may be used as a valuable marker in the differential diagnosis of OB and OS. Nuclear ß-catenin staining strongly suggests OB, whereas cytoplasmic/membranous staining of ß-catenin suggests OS.

Salinomycin activates AMP-activated protein kinase-dependent autophagy in cultured osteoblastoma cells: a negative regulator against cell apoptosis.

BACKGROUND: The malignant osteoblastoma has poor prognosis, thus the search for novel and more efficient chemo-agents against this disease is urgent. Salinomycin induces broad anti-cancer effects both in vivo and in vitro, however, its role in osteoblastoma is still not clear. KEY FINDINGS: Salinomycin induced both apoptosis and autophagy in cultured U2OS and MG-63 osteoblastoma cells. Inhibition of autophagy by 3-methyladenine (3-MA), or by RNA interference (RNAi) of light chain 3B (LC3B), enhanced salinomycin-induced cytotoxicity and apoptosis. Salinomycin induced a profound AMP-activated protein kinase (AMPK) activation, which was required for autophagy induction. AMPK inhibition by compound C, or by AMPKα RNAi prevented salinomycin-induced autophagy activation, while facilitating cancer cell death and apoptosis. On the other hand, the AMPK agonist AICAR promoted autophagy activation in U2OS cells. Salinomycin-induced AMPK activation was dependent on reactive oxygen species (ROS) production in osteoblastoma cells. Antioxidant n-acetyl cysteine (NAC) significantly inhibited salinomycin-induced AMPK activation and autophagy induction. CONCLUSIONS: Salinomycin activates AMPK-dependent autophagy in osteoblastoma cells, which serves as a negative regulator against cell apoptosis. AMPK-autophagy inhibition might be a novel strategy to sensitize salinomycin's effect in cancer cells.

Recurrent chromosome 22 deletions in osteoblastoma affect inhibitors of the Wnt/beta-catenin signaling pathway.

Osteoblastoma is a bone forming tumor with histological features highly similar to osteoid osteoma; the discrimination between the tumor types is based on size and growth pattern. The vast majority of osteoblastomas are benign but there is a group of so-called aggressive osteoblastomas that can be diagnostically challenging at the histopathological level. The genetic aberrations required for osteoblastoma development are not known and no genetic difference between conventional and aggressive osteoblastoma has been reported. In order to identify recurrent genomic aberrations of importance for tumor development we applied cytogenetic and/or SNP array analyses on nine conventional and two aggressive osteoblastomas. The conventional osteoblastomas showed few or no acquired genetic aberrations while the aggressive tumors displayed heavily rearranged genomes. In one of the aggressive osteoblastomas, three neighboring regions in chromosome band 22q12 were homozygously deleted. Hemizygous deletions of these regions were found in two additional cases, one aggressive and one conventional. In total, 10 genes were recurrently and homozygously lost in osteoblastoma. Four of them are functionally involved in regulating osteogenesis and/or tumorigenesis. MN1 and NF2 have previously been implicated in the development of leukemia and solid tumors, and ZNRF3 and KREMEN1 are inhibitors of the Wnt/beta-catenin signaling pathway. In line with deletions of the latter two genes, high beta-catenin protein expression has previously been reported in osteoblastoma and aberrations affecting the Wnt/beta-catenin pathway have been found in other bone lesions, including osteoma and osteosarcoma.

Osteoblastoma is a metabolically active benign bone tumor on 18F-FDG PET imaging.

We describe a case of a 9-y-old girl who on (18)F-FDG PET imaging was found to have a highly metabolically active sacral tumor with an average standarized uptake value of 6.2. The tumor was proven to be osteoblastoma by pathologic examination. Osteoblastoma is a relatively rare benign primary bone tumor and occurs predominantly in patients younger than 20 y. The most common area of involvement is the spine. Osteoblastoma has been reported to be metabolically active on (18)F-FDG PET imaging, with an average standarized uptake value of 3.2, which renders (18)F-FDG PET imaging unable to differentiate benign from malignant primary bone tumors. To our knowledge, only 5 cases of osteoblastoma evaluated by (18)F-FDG PET imaging have been reported in the literature; all were metabolically active on (18)F-FDG PET imaging. The objective of this case report is to show that a metabolically active primary bone tumor on (18)F-FDG PET imaging might be benign and not necessarily malignant.

Osteoid osteoma and osteoblastoma: novel histological and immunohistochemical observations as evidence for a single entity.

AIMS: Osteoid osteoma and osteoblastoma have, in the past, been variously regarded as both similar and distinct entities. Currently, WHO classifies these tumours separately. We compared archetypal cases to identify novel histomorphological and immunohistochemical features attempting to clarify their mutual relationship. METHODS AND RESULTS: 10 osteoid osteomas and 20 osteoblastomas (10 spinal and 10 non-spinal) were retrieved and reviewed clinically, radiologically and histologically. Immunohistochemistry was performed for: desmin, SMA, neurofilament, S100, vimentin, PGP9.5, GFAP, EMA, caldesmon, CD34, broad-spectrum cytokeratins, claudin-1. We identified features, common to both osteoid osteoma and osteoblastoma, namely, areas of lesional non-osteoblastic stroma and the presence of scattered, large cells with smudged/degenerate nuclei. Immunohistochemically, we confirmed the innervated status of osteoid osteomas, and found that osteoblastomas were similarly innervated. The non-osteoblastic lesional stroma was distinctive owing to expression of EMA and NSE by the mesenchymal spindle cells and expression of desmin, PGP9.5 and S100 by the scattered, large cells with 'smudged' nuclei. CONCLUSIONS: Both osteoid osteoma and osteoblastoma are innervated bone-forming lesions which share novel histomorphological and immunohistochemical features supporting the view that separate classification is unjustified, and we offer a pathogenetic explanation for their apparent clinical and radiological variance.

SATB2 is a novel marker of osteoblastic differentiation in bone and soft tissue tumours.

AIMS: Diagnosing osteosarcoma can be challenging, as osteoid deposition is often limited in extent, and hyalinized stroma may closely mimic osteoid. SATB2 is a nuclear protein that plays a critical role in osteoblast lineage commitment. The aim of this study was to examine SATB2 expression in osteosarcomas and other bone and soft tissue tumours, to evaluate its diagnostic utility. METHODS AND RESULTS: Whole sections of 215 tumours were evaluated, including 52 osteosarcomas (43 of skeletal origin; nine extraskeletal), 86 other bone tumours, and 77 other soft tissue tumours. All skeletal osteosarcomas, osteoblastomas, osteoid osteomas, and fibrous dysplasias, eight (89%) extraskeletal osteosarcomas, five (83%) giant cell tumours and three (50%) chondromyxoid fibromas showed nuclear immunoreactivity for SATB2. Staining in other bone and soft tissue tumours was predominantly limited to areas of heterologous osteoblastic differentiation. Focal weak staining was identified in one (9%) unclassified pleomorphic sarcoma and one (13%) monophasic synovial sarcoma. SATB2 was negative in all soft tissue tumours with prominent sclerotic stromal collagen. CONCLUSIONS: SATB2 is a marker of osteoblastic differentiation in benign and malignant mesenchymal tumours. Although SATB2 is not specific for osteosarcoma, it has the potential to be a useful adjunct in some settings, particularly in the distinction between hyalinized collagen and osteoid.

Role of immunohistochemical cyclo-oxygenase-2 (COX-2) and osteocalcin in differentiating between osteoblastomas and osteosarcomas.

BACKGROUND/AIMS: Differential diagnosis between aggressive osteoblastoma and low grade osteosarcoma may be very difficult or even impossible on a small biopsy. This study was designed to assess the usefulness of immunoexpression of COX-2 and osteocalcin in the differential diagnosis of the two tumour types. METHODS: Immunostaining of COX 2 and osteocalcin were studied in 9 osteoblastomas and 30 osteosarcomas. RESULTS: All osteoblastomas and 11/20 (55%) high-grade osteosarcomas showed COX-2 immunoreactivity. All low grade osteosarcomas were COX-2 negative. COX-2 was significantly higher (p<0.002) in osteoblastomas 9/9 (100%) than in osteosarcomas 13/30 (43%) and in aggressive osteoblastomas versus low grade osteosarcomas (p<0.01). Osteocalcin was found in tumour cells of all osteosarcomas and osteoblastomas and in the osteoid matrix of 84% of osteosarcomas and 78% of osteoblastomas. Strong osteocalcin was significantly higher (p<0.02) in osteoblastomas (78%) than in osteosarcomas (27%). CONCLUSION: COX-2 is a valuable marker in distinction between osteosarcoma and osteoblastoma. Negative COX-2 could confirm the diagnosis of low grade osteosarcoma versus aggressive osteoblastoma. Intensity and distribution of osteocalcin may indicate the degree of osteoblastic differentiation.

Nuclear p63 expression in osteoblastic tumors.

Expression of the p63 tumor suppressor protein has been reported in the mononuclear stromal cells of giant cell tumor of the bone, which may represent osteoblast-precursor cells. Only a limited number of osteoblastic tumors have been studied for p63 expression thus far. We therefore examined whether p63 may serve as a marker for osteoblastic differentiation in osteosarcomas or as a differential diagnostic marker to distinguish osteoblastoma from osteosarcoma. Immunohistochemical stains for p63 were performed on a tissue microarray containing 71 chemotherapy naïve biopsy samples of osteosarcoma, 21 whole sections of osteosarcoma, and 8 osteoblastomas. Nuclear p63 was detected in seven of eight osteoblastomas but was restricted to stromal cells within primitive, immature-appearing areas of osteoid deposition. Although only 7 of 71 (10 %) biopsy samples of osteosarcoma represented on the tissue microarray were positive for p63, 7 of 21 (33 %) osteosarcomas were positive when whole tissue sections were evaluated. Although p63 is detected in most osteoblastomas, it is also observed in a significant subset of osteosarcomas, severely limiting its utility in distinguishing between benign and malignant osteoblastic tumors. The relatively low prevalence of p63 expression in osteosarcoma would also seem to preclude its use as a marker of osteoblastic differentiation in skeletal sarcomas.

CD138 (syndecan-1) expression in bone-forming tumors.

CD138 (syndecan-1), a cell surface proteoglycan, is sensitive and specific for plasmacytic differentiation in hematologic disorders. Expression of CD138 has been observed in a majority of epithelial neoplasms and, rarely, soft tissue tumors. However, its expression in bone tumors has not been evaluated. We studied CD138 expression in 27 osteosarcomas, 12 benign bone-forming tumors (osteoid osteoma and osteoblastoma), and 17 reactive bone cases. CD138 expression was also evaluated in a tissue microarray (TMA) constructed from 24 osteosarcomas, 24 chondrosarcomas, 12 giant cell tumors of bone, and 9 normal bone samples. Membranous expression of CD138 was found in an average of 31% of osteosarcoma cases (16/51; 14/27 [52%] in in-house cases; 2/24 [8%] in TMA cases) and in 83% of osteoid osteoma/osteoblastoma cases (10/12). Subsequent immunoglobulin κ and λ stains were negative in the CD138+ cases. All cases of chondrosarcoma, giant cell tumor of bone, and normal/reactive bone tested were nonreactive with anti-CD138. Our results show that CD138 reactivity for neoplastic cells in bone is not a definitive marker for plasmacytic origin, and caution is required to interpret CD138+ cells from a bony lesion for which a hematologic etiology has not been established.

[Expression and significance of N-cadherin and ß-catenin protein in osteosarcoma].

OBJECTIVE: To evaluate the expression of N-cadherin and ß-catenin protein and their relationship with clinicopathological characteristics of osteosarcoma. METHODS: The expressions of N-cadherin and ß-catenin at protein level were detected by immunohistochemical staining in 54 cases of osteosarcoma, 11 cases of osteoid osteoma, 7 cases of osteoblastoma and 8 cases of newly formed bone in nonmalignant bone diseases. The relationship between the two indexes and clinicopathological characteristics of osteosarcoma was analyzed. RESULTS: In newly formed bone, osteoblastoma, osteoid osteoma and osteosarcoma, the positive expression rate of N-cadherin protein was 75.0%, 71.4%, 63.6% and 35.2%, respectively. The positive expression rate of N-cadherin protein in osteosarcoma was significantly lower than that in osteoid osteoma, osteoblastoma and newly formed bone in nonmalignant bone diseases (P = 0.035). The positive expression rate of N-cadherin protein in osteosarcoma cases with pulmonary metastasis was lower than that in cases without (21.7% vs. 56.3%, P = 0.027). The positive expression rate of N-cadherin protein in osteosarcoma cases died in two years was lower than that in cases lived for more than two years (18.2% vs. 50.0%, P = 0.024). In newly formed bone, osteoblastoma, osteoid osteoma and osteosarcoma, the aberrant expression rate of ß-catenin protein was 12.5%, 28.6%, 27.3% and 66.7%, respectively. The aberrant expression rate of ß-catenin protein in osteosarcoma was significantly higher than that in osteoid osteoma, osteoblastoma and newly formed bone (P = 0.002). Aberrant expression rate of ß-catenin in osteosarcoma cases with pulmonary metastasis was higher than that without (82.6% vs. 43.8%, P = 0.011). An inverse correlation was found between the aberrant expression of ß-catenin and N-cadherin expression in osteosarcoma(r = -0.302, P = 0.027). CONCLUSION: The positive expression rate of N-cadherin is decreased in osteosarcoma while aberrant expression rate of ß-catenin increased. The expression of N-cadherin protein is closely correlated with the metastasis and prognosis of osteosarcoma, but the expression of ß-catenin protein is merely correlated with the metastasis of osteosarcoma.