| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |




From the Department of Pathology,* Brigham and Womens Hospital, Boston; the Department of Laboratory Medicine and Pathology,
Mayo Clinic, Rochester, Minnesota; the Department of Pathology,
Childrens Hospital Medical Center, Boston; the Departments of Orthopedic Surgery,
Beth Israel Deaconess Medical Center and Childrens Hospital Medical Center, Boston; the Department of Pathology,¶ Massachusetts General Hospital, Boston; and the Department of Pediatric Oncology,|| Dana-Farber Cancer Institute, Boston, Massachusetts
| Abstract |
|---|
|
|
|---|
The purported reactive nature of ABC was refuted by the work of Panoutsakopoulos et al,9 which demonstrated chromosomal translocation t(16;17)(q22;p13) as a recurrent cytogenetic abnormality in primary ABC. Subsequently, other groups confirmed 17p13 rearrangement as a frequent cytogenetic aberration in primary ABC.10-15 Recently, we extended these studies by showing that the t(16;17)(q22;p13) fuses the promoter region of the osteoblast cadherin 11 gene (CDH11) on chromosome 16q22 to the entire coding sequence of the ubiquitin protease USP6 gene (also known as TRE2 or TRE17) on chromosome 17p13, suggesting that the pathogenesis of many primary ABC involves up-regulation of USP6 transcription.16
In this study we determined the frequency and the clinicopathologic consequences of CDH11 and USP6 genetic abnormalities in 52 primary ABC, and we also used these molecular markers to evaluate the pathogenetic relationship or lack thereof between primary and secondary ABC. Further, we identify the neoplastic cell in primary ABC by determining which cell components contained the genetic rearrangements.
| Materials and Methods |
|---|
|
|
|---|
Fifty-two primary ABC and 17 secondary ABC were histologically characterized according to established criteria.8 Metaphase preparations were available from eight of these cases, including two for which the cytogenetic findings have been reported previously (cases 19 and 29).16 Frozen tissue was available from 16 cases.
Fluorescence in Situ Hybridization (FISH)
BAC clones were obtained from Childrens Hospital Oakland Research Institute (Oakland, CA) and Research Genetics (Huntsville, AL). BAC minicontigs were assembled based on genomic mapping and sequence data from the Human Genome Working Draft. Minicontigs telomeric and centromeric to the USP6 locus were USP6. T (BACs RP11124C16, RP11111I16, and RP11177H5) and USP6. C (CTD-2367F23 and RP11457I18), respectively. Those telomeric and centromeric to CDH11 were CDH11. T (RP11137A18, RP11631H23, and RP11351A20) and CDH11. C (RP11615M9, RP11730A21, and RP1176J1), respectively. BAC DNA isolations and labeling were performed as described previously.16
FISH was performed using 4-µm paraffin-embedded tissue sections which were deparaffinized in xylene (3 x 10 minutes), dehydrated twice in 100% ethanol for 2 minutes, and treated with 100 mmol/L Tris and 50 mmol/L EDTA (pH 7.0) for 15 minutes at 93°C. Tissue sections were then rinsed once in 1X PBS and protein digested with Digest All-3 (Zymed, San Francisco, CA). After briefly washing in 1X PBS, the slides were sequentially dehydrated in alcohol (70%, 85%, 95%, and 100%), and air-dried for 1 hour at room temperature. Tissue sections were denaturated at 75°C for 2 minutes and BAC probe hybridization was carried out overnight in a humidified chamber at 37°C. Tissue sections were then washed in 0.5X SSC for 5 minutes at 73°C and treated with CAS block (Zymed) for 10 minutes. Probe detection was performed using FITC-anti-digoxigenin (1:500) and Alexa Fluor 594-streptavidin (1:500) (Molecular Probe, Eugene, OR)17 for 30 minutes. Slides were then mounted in VECTASHIELD mounting medium with 1.5 µg/ml of 4',6-diamidino-2-phenylindole (DAPI). ABC were scored as positive for gene rearrangement if more than 5% of cells showed splitting apart of the flanking FISH probes. If fewer than 20% of cells showed rearrangement, the findings were corroborated by repeating the FISH assay in another section from the same paraffin block.
RNA Isolation and RT-PCR
RNA was isolated from frozen tumors after mechanical homogenization and overnight incubation in Trizol (Invitrogen, Carlsbad, CA) at 4°C. RNA reverse transcription into cDNA was performed using the GeneAmp RNA PCR Kit (Applied Biosystems, Foster City, CA) for 2 hours at 42°C using random hexamers. RNA isolation from paraffin sections were performed according to a previously described protocol.18
PCR reactions were performed using the Takara Ex Taq kit with the following parameters for 35 cycles: denaturation at 94°C for 30 seconds, annealing at 65°C for 30 seconds, and extension at 72°C for 1 minute. PCR primers for RT-PCR evaluation of the CDH11-USP6 fusion oncogene were CDH11+71F (5'-CGCCGCTGACTTGTGAAT-3') and USP6+1781R (5'-CTCGGTGTCCCTTGTCATACTT-3'). Evaluation of low-abundance fusion transcripts in tumors where the first round PCR was negative was performed by nested PCR using primers CDH11+83F (5'-GTGAATGGGACCGGGACT-3') and USP6+1736R (5'-CAGGAGCGGAAGGACATACTTA-3') at the same cycling parameters as above for 25 cycles.
Statistical Analysis
Statistical analyses were performed with the S-PLUS 6.0 software package (Insightful Corp., 2001). The Fisher exact test was used to evaluate associations among categorical variables. The Mann-Whitney U-test was used to evaluate associations between categorical and continuous variables. Recurrence-free survival was calculated using the Kaplan-Meier product-limit method, and univariate survival analyses were calculated using the log-rank test. Multivariate survival analyses were performed using the Cox multivariate regression model. Proportionality assumption was evaluated using Kaplan-Meier and log-minus-log survival curves. Reported P values were two-sided and statistical significance was set at P
0.05.
| Results |
|---|
|
|
|---|
Clinical features are summarized in Table 1
. Median patient age at diagnosis was 14 years (range, 2 to 42 years), and the genders were equally represented (27 females and 25 males). The most frequent primary sites were tibia (n = 11), femur (n = 8), fibula (n = 7), vertebra (n = 5), and humerus (n = 4). Thirty-nine tumors arose in peripheral locations, whereas 13 were central. Median tumor size was 4 cm (range, 1 to 8 cm). Treatment information was available for 49 patients, of whom 42 (86%) were treated by curettage and seven by local excision. No patient received radiotherapy. Clinical follow-up was available for 44 patients, of whom 17 (39%) had a local recurrence after a median interval of 35 months (range, 2 to 35 months). Two patients had more than one recurrence.
|
Molecular Cytogenetics and Molecular Genetics of Primary and Secondary ABC
FISH analyses in 52 primary ABC showed that 36 (69%) had CDH11 and/or USP6 locus rearrangement (Table 2)
. Of these, 10 cases (28%) exhibited rearrangement of both loci, and a fusion transcript CDH11-USP6 was confirmed by RT-PCR in each of these cases. Notably, 23 ABC featured USP6 rearrangement without associated CDH11 rearrangement, consistent with variant USP6 activation mechanisms (Table 2)
. Only three ABC had CDH11 rearrangement without associated USP6 rearrangement (Table 2)
. CDH11-USP6 RT-PCR validations were performed in a representative group of the ABC with FISH rearrangements of USP6 only, CDH11 only, or neither USP6 or CDH11, and each of these tumors lacked the CDH11-USP6 fusion transcript (Table 2)
. Generally, the CDH11 and USP6 rearrangements appeared to be genomically balanced, although two ABCs had an unbalanced USP6 rearrangement with loss of chromosomal material telomeric to USP6, as evidenced by deletion of the FISH probe in that region. Another ABC had an unbalanced rearrangement of the CDH11 locus with loss of chromosome material centromeric to CDH11. Among seven solid variants of ABC, six cases had CDH11 or USP6 loci rearrangements (Table 2)
.
|
|
In primary ABC (Figure 1A)
with USP6 or CDH11 rearrangement, the percentage of abnormal cells varied from 7% to 82% (median, 27%). These cells were usually spindled, wavy, or oval (Figure 1B)
, and were indistinguishable morphologically from surrounding spindle cells that lacked USP6 rearrangement. The cytogenetically abnormal cells were generally scattered diffusely throughout the lesion and were often grouped in small or large clusters adjacent to multinucleated giant cells (Figure 1C)
. USP6 and CDH11 rearrangement were never seen in the multinucleated giant cells, inflammatory cells, endothelial cells, and in the metaplastic bone-associated osteoblasts, corroborating their non-neoplastic nature. In addition, ABC cyst wall lining cells did not have USP6 or CDH11 locus rearrangement. Ki-67 antigen (MIB-1) median labeling index in ABC was 7% (range, 1 to 30%), and there was no association between the presence of USP6 locus rearrangement and MIB1 labeling index (P = 0.47).
Clinicopathologic Correlations of USP6 and CDH11 Rearrangements in ABC
Clinical correlates were compared in primary ABC with molecular aberrations (USP6 or CDH11 rearrangement) versus those lacking demonstrable USP6 or CDH11 alterations. Patients with ABC with and without the molecular aberrations did not differ in age at diagnosis (P = 0.27), gender (P = 0.38), tumor size (P = 0.99), or location (central versus peripheral; P = 0.30). In addition, there was no difference in local recurrence-free survival after a median follow-up of 28 months (range, 1 to 108 months) (Figure 2
and Table 3
). Cox multivariate regression analysis confirmed these findings and showed that only central tumor location was associated with an increased risk for local recurrence after adjusting for age, gender, tumor size, type of treatment (curettage or complete excision), and molecular aberrations (Table 3)
.
|
|
| Discussion |
|---|
|
|
|---|
Histologically, ABC is characterized by hemorrhagic cystic and cavernous spaces of varying size surrounded by fibrous septa or more solid areas composed of mitotically active spindle or ovoid cells intermixed with inflammatory cells, scattered osteoclast-like multinucleated giant cells, extravasated red blood cells, and numerous capillaries. The histological growth is reminiscent of early granulation tissue. Trabecular osteoid formation with active osteoblasts and a characteristic matrix calcification with chondroid features are also frequently seen. Some tumors exhibit minimal or no cystic component, and these lesions have been designated as "solid variant of ABC" and "giant cell reparative granuloma". The latter term is particularly used for lesions that occur in craniofacial bones and the small bones of the hand and feet.1,8,19,22-25
The pathogenesis of ABC has been a subject of debate since its original description.1 However, identification of recurrent chromosomal translocation t(16;17)(q22;p13) by Panoutsakopoulos et al9 strongly indicated a neoplastic pathogenesis in at least a subset of primary ABC. These findings were confirmed by additional cytogenetic reports,10-15 which also showed similar cytogenetic abnormalities in the solid variant of ABC and in soft tissue ABC.10,26 Furthermore, these studies showed that chromosome 17p13 is the most frequently rearranged chromosomal region in ABC.
Recently, we demonstrated that the chromosomal translocation t(16;17)(q22;p13) fuses the promoter region of the osteoblast cadherin 11 gene (CDH11) on chromosome 16q22 to the entire coding sequence of the ubiquitin protease USP6 gene on chromosome 17p13.16 We also found that CDH11 and USP6 loci were rearranged in ABCs with alternate translocations involving either chromosome 16q22 or 17p13. These studies suggested that the pathogenesis of some primary ABC involves transcriptional up-regulation of USP6. The transforming mechanisms of USP6 remain to be determined but could involve known USP6 roles in regulation of actin remodeling through interactions with the Rho GTPases Cdc42 and Rac1.27
In the present studies, we demonstrate CDH11 or USP6 rearrangements in 69% of primary ABC. Notably, 44% (23 of 52) had variant USP6 rearrangements, in the absence of CDH11 rearrangement. These findings suggest that USP6 could be the most prevalent fusion partner in ABC, with variant translocations providing alternative oncogenic mechanisms for USP6 transcriptional up-regulation. The identification of three cases in which only the CDH11 locus was rearranged, without evidence of USP6 rearrangement, also suggests that genes other than USP6 are up-regulated by juxtaposition with CDH11 in ABC.
The possibility of a genetic relationship between primary and secondary ABC was suggested recently by the finding of chromosome band 16q22 and 17p13 rearrangements in a giant cell tumor with secondary ABC.21 Nonetheless, in contrast to the situation in primary ABC, we found no evidence of CDH11 or USP6 rearrangements in secondary ABC associated with giant cell tumor, chondroblastoma, osteoblastoma, and fibrous dysplasia. Although chromosome 16 and 17 abnormalities have been found infrequently in giant cell tumors,21,28,29 it is unclear whether any of these involve CDH11 or USP6 loci. Based on the present evidence, we conclude that secondary ABC and primary ABC have a different pathogenesis. Secondary ABC, presumably, is a common endpoint of differentiation or a non-specific morphological pattern in various non-ABC tumors. Despite being a morphological mimic of primary ABC, secondary ABC likely has varied genetic features, corresponding to the specific bone tumors with which it is associated.
Presence of USP6 and CDH11 rearrangements did not correlate with ABC clinicopathological variables, nor, in univariate and multivariate analyses, with recurrence-free survival. The only variable that correlated with recurrence-free survival was tumor location, with central ABC being at greater risk for recurrence. However, we cannot exclude the possibility that specific USP6 or CDH11 rearrangement mechanisms are predictive of recurrence. Therefore, it would be worthwhile to evaluate relationships between molecular aberrations and ABC recurrence further, ideally in the context of a larger-scale and prospective clinical study.
The paraffin-section FISH analyses in this study enabled evaluation of ABC cell types which had the USP6 and CDH11 rearrangements. All cells with USP6 or CDH11 rearrangements were spindle or ovoid, and were diffusely admixed within the other ABC cell components, as well as being clustered around osteoclast-like multinucleated giant cells. The lack of cytogenetic aberrations in the inflammatory cells, endothelial cells, metaplastic bone-associated osteoblasts, and the multinucleated osteoclast-like giant cells supports their reactive nature. Conceivably, they are involved in a host response to the neoplastic ABC cells. It is also possible that one or more of the non-neoplastic cellular constituents release factors that stimulate division of the neoplastic ABC cells. The frequent clustering of neoplastic cells in the vicinity of multinucleated giant cells suggests a role for these cells in the neoplastic proliferation. Notably, the ABC cells with USP6 or CDH11 rearrangements were generally admixed with morphologically indistinguishable spindle cells lacking such rearrangements. We hypothesize that the latter group is composed by reactive fibroblasts/myofibroblasts, but it is also possible that some of these are committed ABC neoplastic cells at an earlier point in transformation, before acquisition of the USP6 or CDH11 rearrangements. Although the majority of cells in some ABC had USP6 or CDH11 rearrangements, there were also tumors in which fewer than 10% of cells had rearrangements (Table 2)
. The small numbers of apparently rearranged cells were not false-positive findings, as they were reproducibly present in these tumors when the FISH was repeated, and when scored by independent observers. Furthermore, RT-PCR reactions corroborated these molecular cytogenetic results showing the presence of CDH11-USP6 fusion transcripts in these tumors. In all, these findings suggest that the neoplastic ABC component induces a vigorous, reactive, host response mimicking young granulation tissue, including inflammatory, myofibroblastic, osteoclast-like giant cells, osteoblastic cells, and numeous capillaries, therefore accounting for the historical perception that these are largely reactive lesions.
In conclusion, primary ABC is a mesenchymal neoplastic disease characterized by a spindle cell proliferation exhibiting USP6 or CDH11 rearrangements in approximately two thirds of the cases. Secondary ABC is a morphological mimic of primary ABC, but lacks the hallmark USP6 or CDH11 rearrangements of primary ABC, and likely represents a common endpoint of differentiation in various non-ABC bone tumors.
| Footnotes |
|---|
Supported by Mayo Clinic and Mayo Clinic Foundation.
Accepted for publication August 3, 2004.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. M. Oliveira, M. M. Chou, A. R. Perez-Atayde, and A. E. Rosenberg Aneurysmal Bone Cyst: A Neoplasm Driven by Upregulation of the USP6 Oncogene J. Clin. Oncol., January 1, 2006; 24(1): e1 - e1. [Full Text] [PDF] |
||||
![]() |
C. Shen, Y. Ye, S. E. Robertson, A. W. Lau, D.-O. D. Mak, and M. M. Chou Calcium/Calmodulin Regulates Ubiquitination of the Ubiquitin-specific Protease TRE17/USP6 J. Biol. Chem., October 28, 2005; 280(43): 35967 - 35973. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. J. Mankin, F. J. Hornicek, E. Ortiz-Cruz, J. Villafuerte, and M. C. Gebhardt Aneurysmal Bone Cyst: A Review of 150 Patients J. Clin. Oncol., September 20, 2005; 23(27): 6756 - 6762. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |