• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Genetic Analysis of Fibrosarcoma of Bone, a Rare Tumour Entity Closely Related to Osteosarcoma and Malignant Fibrous Histiocytoma of Bone

Posted on: Thursday, 9 December 2004, 03:00 CST

Fibrosarcoma of bone; Comparative genomic hybridisation; Microarray; Real-time quantitative PCR; Chromosome 22; Platelet- derived growth factor beta; Immunohistochemistry

Fibrosarcoma (FS) of bone is an extremely rare and genetically uncharacterised malignant tumour arising in the skeleton. On the basis of clinicopathologic features it appears to be closely related to either fibroblastic osteosarcoma (OS) or malignant fibrous histiocytoma (MFH) of bone. In this study, 27 decalcified, paraffin- embedded FS of bone were collected for genetic and immunohistochemical characterisation. Good quality DNA, suitable for genetic analyses, was isolated from nine cases (7 primary tumours, 1 local recurrence, and 1 lung metastasis), which were analysed by comparative genomic hybridisation (CGH) on chromosomes and DNA microarrays. DNA sequence copy number changes were found in five out of seven primary tumours (72%), as well as in both, the local recurrence and the metastatic lesion, by CGH on chromosomes. The most frequent aberration was gain of the chromosomal region 22q, which was present in four out of the five primary tumours with genetic changes, in the local recurrence and, as the sole genetic aberration, in the lung metastasis. DNA microarray analysis showed that gain of the platelet-derived growth factor beta (PDGF-B) gene (located at 22q12.3-q13.1) was the most frequent gene imbalance, which was present in three out of the five analysed tumours. In these three cases, real-time PCR revealed a 2.1- to 2.7-fold increase of PDGF-B gene copy numbers. By immunohistochemistry, a positive reaction for B-chain-containing PDGF proteins was revealed in all the cases showing gain of 22q. A more extensive immunohistochemical analysis identified the presence of PDGF-B proteins in 8/20 primary FS of bone (40%), 3/3 lung metastases and in 1/2 local recurrences. A simultaneous positive reaction for PDGF- B proteins and PDGF receptors was found in two third of PDGF-B- positive cases (8/12). Taken together, the genetic and immunohistochemical data indicate that over-representation of the chromosomal region 22q, including particularly the PDGF-B gene, may be important for the pathogenesis of FS of bone. Our results also demonstrate that CGH on chromosomes and DNA microarrays are suitable for the genetic characterisation of decalcified, paraffin-embedded tumour tissue samples and may facilitate, combined with other techniques, the rapid acquisition of data providing insight into the molecular genetic and biologic basis of rare bone sarcomas. Moreover, these findings suggest the possible presence of an autocrine loop in FS of bone, which might be taken into account for planning innovative therapeutic strategies for patients unresponsive to conventional treatments.

Abbreviations. CGH Comparative genomic hybridisation. - DAPI 4',6- Diamidino-2-phenylindole - FS Fibrosarcoma. - MFH Malignant fibrous histiocytoma. - OS osteosarcoma. - PDGF-B Platelet-derived growth factor beta. - SSC Saline-sodium citrate. - T/R Test-to-reference. - TRM T/R mass normalised.

Introduction

Fibrosarcoma (FS) of hone is an extremely rare primary malignant tumour arising in the skeleton, which accounts for less than 5% of all primary malignant hone neoplasms (Campanacci, 1999; Dorfman and Czerniak, 1995: Kahn and Vigorila, 2002). This rare tumour usually arises in the metaphysis of long hones, especially near the knee, in patients from the second to the seventh decade. It is now commonly considered as a distinct tumour entity, which was formerly included in the category of high-grade osteosarcoma (OS) or malignant fibrous histiocytoma (MFH) of bone (Campanaeei, 1999; Dorfman and Czerniak, 1995; Kahn and Vigorila, 2002; Papagelopoulos et al., 2002, Taconis and Mulder. 1984).

Histologically, FS of bone is characterised by the presence of spindle cells, usually arranged in bundles, in association with a fibrous intercellular substance in which fine collagen fibres are often arranged in a so-called herring-bone pattern, without any evidence of osteoid or cartilage formation. The amount of fibres, atypical nuclei and number of milotic figures may considerably vary among different eases and are taken into account for tumour grading. These histological features are not exclusive for FS of bone, and sometimes it is difficult to distinguish this tumour from fibroblastic OS with a poor osteoid production or from MF1H of bone (Capanna et al., 1984; Campanaeei. 1999: Dorfman and Czerniak, 1995; Kahn and Vigorita, 2002).

FS of hone may be graded according to increasing malignancy and decreasing differentiation. High-grade FS of bone is an aggressive neoplasm that develops metastases, most commonly located in the lung, in 55-65% of patients and with a 10-year survival rate not exceeding40% (Kahn and Vigorita, 2002; Papagelopoulos et al., 2002). Treatment consists of wide surgical excision or amputation, often associated with adjuvant or neoadjuvant chemotherapy regimens, which are similar or identical to those used for high-grade OS (Campanacci, 1999: Dorfman and Czerniak, 1995).

Although analyses of genetic aberrations have been performed for both high-grade OS and MFH of bone (Knuutila et al., 1998) (for a summary of the available CGFI data on bone sarcomas, see the web site: http://www.helsinki.fi/cmg/cgh_data.html), no data have been reported so far regarding the genetic characteristics of FS of hone. In the present study, complementary CGH-hased and molecular biologic techniques have been used to screen for chromosomal imbalances and gene copy number changes in FS of hone, in order to better characterise this neoplasm in relation to the other more common malignant bone sarcomas.

Materials and methods

Samples

Among the patients referred to the Rizzoli Institute in the last 20 years, 25 cases were diagnosed as FS of bone and were considered for this study. Two additional patients (cases 4 and 5) were from the Department of Pathology and Medical Genetics of the Haartman Institute and Helsinki University Central Hospital (Finland). In a recent review of all these cases, histological diagnosis and grade were confirmed on hemaloxylin/eosin-stained slides by following conventional criteria (Campanacci, 1999; Dorfman and C/erniak, 1995; Kahn and Vigorita, 2002). Differential diagnosis between FS and MFH of bone was based on the presence of a herring-bone pattern of growth with a mild cellular pleomorphism and rare giant cells in the first one. compared to the presence of a well-defined sloriform pattern of growth together with a marked cellular pleomorphism and the presence of large round cells with vacuolated cytoplasm and bizarre, polynucleated giant cells in the latter. Differential diagnosis with fibrohlastic OS was mostly based on the absence of ostcoid production in FS of bone.

Formalin- or ethanol-fixed, decalcified, paraffin-embedded tissue biopsies were available for all 27 cases, which included 22 primary tumours, three lung metastases, and two local recurrences. For each case, the estimation of the relative amount of tumour cells was determined on haematoxylin/cosin-stained tissue sections in order to exclude nonrepresentative specimens. All cases showed amounts of tumour cells ranging from 75% to 95% and were considered representative.

DNA isolation

Genomic DNA for genetic analyses was isolated from ten to twenty 8 m-thick paraffin sections from each case. Sections were de-waxed with xylenc at 56 C for 30 -60 min and then dehydrated with absolute ethanol at room temperature. After enzymatic digestion with collagenase type I (Sigma-Aldrich, St. Louis, MO;2 mg/ml phosphate- buffered saline at 37 C for 30 - 90 min), tissue fragments were incubated with the DNAzol reagent (Invitrogen Italia, Milan, Italy) and DNA was isolated according to the manufacturer's procedure. DNA concentration was determined by spectropholometry and the quality of the DNA was evaluated by agarosc gel eleclrophorcsis. A good quality DNA suitable for genetic analyses was obtained in 9/27 cases (33%). which were consecutively numbered from 1 to 9 and were analysed by CGH on chromosomes.

Comparative genomic hybridisation (CGH) on chromosomes and microarrays

CGH on chromosomes was performed as previously described (Hl- Rifai et al., 1997). Oenomic DNA from each tumour sample was used as test DNA, whereas DNA isolated from sex-matched, peripheral blood of healthy males or females was used as normal reference DNA. By using nick translation, test DNA was labelled with fluorcscein and normal reference DNA with Texas Red. Equal amounts, corresponding to 1 g of labelled DNA, were co-precipitated together with 10 g of human Cot- 1 DNA (Oibco BRL, Oaithershurg, MD). DNA samples were cohybridised at 37 C for 48 hours on metaphase chromosomes obtained from human normal lymphocytes, which had previously been denatured with 70% formamide/2 saline-sodium citrate (SSC) at 66 C for 2 min. After post-hybridisation washes with 50% formamidc/2 SSC (three times for 10 min, each at 45 C ), 2 SSC (two times for 10 min, each at 45 C), 0.1 SSC (10 min at 45 C), and 2 SSC, 0.1 MNaH^sub 2^PO^sub 4^/0.1 M Na^sub 2^HPO^sub 4^ buffer and distilled water (each for 10 min at room temperature), samples were counterstained with 4',6-d\iamidino- 2-phcnylindole (DAPI; 0.15 g/ml antifade solution). The hybridisation signals were analysed by using a Zeiss fluorescence microscope and the ISIS digital image analysis system (MetaSystems GmbH, Altlussheim, Germany). Three-colour images (green for the tumour DNA, red for the normal reference DNA, and blue for DAPI counterstaining) were obtained from each metaphase and average ratio profiles were determined from the analysis of at least ten metaphases for each specimen. Chromosomal regions were considered to be over-represented when the green-to-red ratio was ≥ 1.17 (gains) or ≥ 1.5 (high-level amplifications), and under- represented when the green-to-red ratio was ≤ 0.85 (losses).

Gene copy number changes were determined by CGH on microarrays with the AmpliOne microchip (Abbott-Vysis, Downers Grove, IL), as previously described (Hattinger et al., 2003). The AmpliOne microchip contains 177 spots specific for 59 oncogenes and tumour suppressor genes, which have been reported to be frequently amplified in human cancers (Daigo et al., 2001; Mao et al., 2002). Tumour DNA (test DNA) from five of the nine cases analysed by CGH on chromosomes (Cases 1, 2, 3, 8, and 9) was labeled in green with AlexaFluor 488-5-dUTP (Molecular Probes, Leiden, The Netherlands) and sex-matched human normal DNA (reference DNA) was labeled in red with AlexaFluor 594-5-dUTP (Molecular Probes) by nick translation (Microarray Nick Translation Kit, Abbott-Vysis). Alter labeling, DNA samples were electrophoresed on a 1.5% agarose gel to ensure that the fragment length ranged between 50 and 500 bp. For each microarray experiment, 500 ng of test and reference DNAs were mixed with 25 l of microarray hybridisation buffer (AmpliOnc Microarray Kit v.0.8, Abbott-Vysis), denatured at 80 C for 10 min and hybridised at 37 C for 72 hours on the AmpliOnc microchip in a 50%/ 2 SSC atmosphere. After post-hybridisation washes with 50% formamide/ 2 SSC (three times for 10 min, each at 40C), 1 SSC (4 times for 5 min, each at room temperature), and distilled water (5-10 seconds at room temperature), microarrays were counterstained with DAPI IV (AmpliOnc Microarray Kit v.0.8, Abbott-Vysis). For each microarray, three images (blue, green and red) were captured with the GcnoSensor Reader System and analysed with the GcnoSensor Reader Software (both from Abbott-Vysis). For each target, the GenoSensor Reader Software provided the mean of normalised test-to-reference ratios of the included spots as calculated by the mass method (T/R mass normalised, from now on referred to as TRM). In control experiments, normal male DNA was co-hybridised with normal female DNA in order to determine the cut-off values for dcletions/under-reprcsentations and gains/amplifications. Targets with TRM values ≥ 1.25 were interpreted as gained, and targets with TRM values ≤ 0.75 were interpreted as deleted. For each spot, the GenoSensor Reader Software also provided the Pearson's coefficient of correlation p between test and reference pixel fluorescence intensities. Target spots with p < 0.8, indicating poor quality of hybridisation, were excluded from the final evaluation. The final evaluation included only targets for which at least two spots with p ≥ 0.8 and a mass ratio coefficient of variation ≤ 10% were available.

Real-time quantitative PCR

A real-time fluorescence detection method based on the 5' exonucleasc activity of the Taq polymerasc was used to determine the relative copy number of the PDGF-B gene. In addition to the sense and antisense primers, a non-extendable probe with a 5' fluorescent reporter dye (6-FAM) and a 3' quencher dye (TAMRA), hybridising to the target sequence downstream of the sense primer, was used. During the extension phase, the Taq polymerase hydrolysed the probe, which, as a result, generated a fluorescence signal being directly proportional to the amount of PCR product synthesised. This fluorescence signal was monitored online using the laser detector of the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City CA).

The PCR amplification was performed using a 96-well plate with a final reaction mixture (25 l) containing 800 nM of each primer, 100 nM probe, and 1TaqMan Universal Master Mix (Applied Biosystems, Foster City, CA). In order to work under the same cycling conditions (95C for 10 min followed by 40 cycles at 95C for 15 s and 60C for 1 min), TaqMan probes and primers for the quantitative detection of human β-actin (reference gene) and human PDGF-B (target gene) were designed by using Primer Express software (Applied Biosystems), generating products with sizes of 62 bp (β-actin) and 81 bp (PDGF-B). Primer sequences were as follows: PDGF-B forward: 5'- ATGGTGT-CAGAGGGAGGATAAACC; PDGF-B reverse: 5'-AGCGGAG- GACTTTGGGAAAT; PDGF-B probe: 6-FAM5'-CAGGGAGG-CAACACTGCTGTCCACAT- 3'TAMRA; β-actin forward: 5'-AGCGCGGCTACAGCTTCA; β-actin reverse: 5'-TTCTCCTTAATGTCACGCACGA; β-actin probe: 6-FAM5'- CACCACGGCCGAGCGGGA-3'TAMRA.

Quantification was performed by using both the standard curve method and the comparative CT method. Standard curves were constructed in each PCR run. Serial dilutions of normal human genomic DNA were used as calibrator and as template for the standard curve. After normalisation with the reference gene, the PDGF-B copy number was calculated by dividing these normalised values by the copy number of the normal human genomic DNA (which had been normalised to the correspondent calibrator). It was also possible to estimate the PDGF-B copy number on the basis of the C^sub T^ values calculated as 2^sup -&916;CT^, where ΔC^sub T^-C^sub 1^ target gene - C^sub T^ reference gene, and &916;C^sub T^ = ΔC^sub T^- sample ΔC^sub T^ calibrator.

Immunohistochemistry

Immunohistochemistry was performed on all the 25 FS of hone from the Rizzoli Institute, as well as on 10 fibroblastic OS and 5 MFH of hone. Immunohistochemical analyses focused on non-collagenous proteins of bone, on B-chain-containing human PDGF (PDGF-BB or PDGF- AB) proteins and on PDGF receptors. Non-collagenous proteins of bone were studied to hettcr define the histological characteristics of FS of bone in relation to fibroblastic OS or MFH of hone. PDGF-B proteins and PDGF receptors were selected to be studied by immunohislochcmistry on the basis of microarray analyses, which identified gain of the PDGF-B gene as the most recurrent gene aberration.

By using an avidin-biotin-peroxidasc complex method, immunohistochemistry was carried out with rabbit polyclonal antibodies against osteonectin (LF-bONII), osteopontin (LF-19), and osteocalcin (LF-32), as previously described (Serra et al., 1996). Before immunostaining, samples were pre-treated with pepsin (Biomeda, Foster City, CA) for 10 min at 40C in order to retrieve the antigens. Primary antibodies were incubated overnight at 4 C at the following dilutions: 1:150 for LF-bONII, 1:100 for LF-19. and 1:200 for LF-32.

For immunodetection of PDGF-B proteins and PDGF receptors, antigen retrieval was not performed and sections were incubated for 48 h at 4C with the Sis1 monoclonal antibody (specific for PDGF-B proteins; BD Pharmingen, San Diego, CA) diluted 1:10; the anti- human PDGF receptor β-subunit (1:200 dilution rate) and the anti-human PDGF receptor α-subunit (1:100 dilution rate) monoclonal antibodies (both from Genzyme, Cambridge, MA). Tissue sections of a PDGF-B-and PDGF receptor-positive breast cancer were used as positive controls.

Development of all immunoreactions was obtained with diaminobcnzidine, and nuclei were counterstained with Gill's haematoxylin. For all antigens, immunostaining was scored on a scale ranging from one plus to three plus, according to the percentage of positive cells. Cases with less than 10% positive cells were classified as + ;cases with positive cells ranging between 10% and 50% were classified as ++ ; cases with more than 50% positive cells were classified as + + +. For non-collagenous proteins of bone, a similar score system was used to identify also the immunohistochemical reaction of the extracellular matrix (where present). According to the intensity of the extracellular matrix immunostaining, cases were classified as + (weakly positive), + + (moderately positive), or ++ + (strongly positive). All immunohistochemical stainings were evaluated by an independent investigator, without knowing the genetic data in order to avoid any possible bias in data interpretation.

Results

Expression of non-collagenous proteins of bone in fibroblastic OS, FS of bone and MFH

As shown in Table 1, immunohistochemical analysis of non- collagenous proteins revealed a positive staining for osteonectin, osteopontin, and osteocalcin, both in cells and in extracellular matrix, in all fibroblastic OS (Fig. 1A). MFH of bone was completely negative for all these non-collagenous proteins, with the only exception of one case, which showed a weak reaction for osteonectin in a few cells (Table 1 and Fig. 1B).

Compared to fibroblastic OS, FS of bone samples showed cither an invariably weaker reaction for all the non-collagenous proteins (Table 1 and Fig. 1C) or were completely negative. Extracellular matrix was either very scarce or completely absent: when it was present, it did not show any positive immunoreaction, confirming its non-ostcoid nature. Both, the local recurrence and the two lung metastases showed only a weak reaction for osteonectin (Table 1).

Table 1. Immunohistochemical staining for non-collagenous proteins of bone.

Table 2. DNA sequence copy number changes detected by CGH on chromosomes in nine fibrosarcomas of bone.

Genomic imbalances and gene copy number changes in FS of bone revealed by CGH-based techniques

Among the 27 FS of bone initially included in this study, a good- quality DNA suitable for genetic analyses was obtained in nine cases (\13%), including seven primary tumours (Cases 1 to 7), one local recurrence (Case 8) and one lung metastasis (Case 9). Results of CGH on chromosomes are summarised in Table 2. DNA sequence copy number changes were detected in five out of seven primary tumours (72%), as well as in the local recurrence and the mctastatic lesion. In primary tumours, the mean number of changes per sample was 6.4 (range 3-11). Losses were more frequent than gains. High-level amplifications were detected in two primary tumours (Cases 2 and 5) at the chromosomal regions 17p11.2, 17q24 and 22q11.2-q12. The most frequent DNA sequence copy number change was gain of 22q, which was present in four of the five rearranged primary tumours, in the local recurrence and as the sole genetic aberration in the lung metastasis (Fig. 2). The second most common alteration was gain of the chromosomal region 8q24.1-qter, which was present in two primary tumours and the local recurrence (Fig. 2).

In order to obtain a gene copy number change profiling, DNAs from 5 of the FS analysed by CGH on chromosomes, including three primary tumours (Cases 1, 2, and 3), the local recurrence (Case 8), and the lung metastasis (Case 9), were hybridised to the AmpliOnc microchip. As shown in Figure 2, microarray analyses identified only gene copy number gains affecting six different genes: PDGF-B (22q12.3-q13.1) (Cases 2, 8 and 9), FGR (1p3n. 1-p36.2) (Case 1), FES (15q26.1) (Case 3), MYC (8q24.12-q24.13) (Case 8), and YES1 (18p11.3) and BCR (22q11.21) (Case 9). Four of these genes map to chromosomal regions identified as gained by CGH on chromosomes (Fig. 2). The most common aberration was the increase of PDGF-B gene copy number, which was present in three out of 5 cases, including one primary tumour (Case 2, TRM = 1.42), the local recurrence (Case 8, TRM = 1.28), and the lung metastasis (Case 9, TRM = 1.27). It is worthwhile noting that these three cases also displayed the major levels of 22q gain by CGH on chromosomes (Fig. 3). Co-amplification of BCR, the second gene of 22q present on the AmpliOnc microchip, was found only in the lung metastasis (Case 9, TRM = 1.51). On the basis of these results, further analyses focused on the assessment of FDGF-B gene copy number changes by real-time PCR and immunodctection of PDGF-B proteins and PGDF receptors on paraffin-embedded tumour tissue sections.

Fig. 1. Immunostaining for osteonectin (A-C), PDGF-B (D, E) and PDGF receptors α (F) and β (G). A positive reaction for osteonectin was seen in fibroblastic osteosarcoma (A), malignant fibrous histiocytoma of bone (B), and fibrosarcoma of bone (C). Arrows in (B) and (C) indicate cells with a weak positive immunostaining. PDGF-B as detected in a breast cancer sample that was used as positive control (D) and in one representative case of fibrosarcoma of hone (E), which showed also a simultaneous positive reaction for both PDGF receptor α (F) and PDGF receptor β (G). Positive reaction for PDGF receptors in the breast cancer sample was similar to that for PDGF-B shown in (D). All images were taken at 200 magnification.

Fig. 2. Summary of DNA sequence and gene copy number changes detected by CGH on chromosomes and on microarrays in fibrosarcomas of bone. Vertical lines on the right side of chromosome diagrams identify gains and thick lines indicate high-level amplifications; vertical lines on the left side represent losses. Genes found to be gained by CGH on microarrays are represented by circles. Solid gray lines and circles represent primary tumours; solid black line and circle the lung metastasis: dashed-light gray lines and circles the local recurrence.

PDGF-B gene copy number status and protein expression determined by real-time quantitative PCR and immunohistochemistry in FS of bone

By using paraffin-embedded samples, only DNA and tumour tissue sections were available for the assessment of PDGF-B gene copy number and PDGF-B proteins and PDGF receptor levels. An increase of PDGF-B gene copy number greater than two-fold compared to the gene copy number of normal human DNA was found in case 2 (2.6-fold), Case 8 (2.1 -fold) and Gase 9 (2.7-fold). These data positively correlated with the results obtained by the CGH-based techniques (Table 3). In fact, these three cases also showed gain of the PDGF- B gene by microarray analysis and of the chromosomal region 22q by CGH on chromosomes. On the other hand, no PDGF-B gene copy number changes were observed in case 1 and case 3, which showed gain of 22q by CGH on chromosomes only, but no gain of the PDGF-B gene by microarray analysis (Table 3).

Using immunohistochemistry, PDGF-B proteins were detected in the five cases that also showed gains of 22q by CGH on chromosomes (Fig. 1E; Table 3). However, it is worthwhile noting that the three cases with a higher ratio of PDGF-B-positive cells (Cases 2, 8, and 9) were those in which gain of the PDGF-B gene was revealed by both microarray and real-time PCR analyses (Table 3). Immunohistochemistry for PDGF-B proteins and receptors was also extended to all the other FS of hone from the Rizzoli Institute, which were not included in the genetic analyses. Considering the whole scries analysed by immunohistochemistry, PDGF-B proteins were found in 12/25 cases (48%), including 8/20 primary tumours (40%), 3/ 3 lung metastases and 1/2 local recurrences (Table 4).

PDGF receptors (Fig. 1F, G) were found in smaller proportions of cases. In particular, 5/25 cases (20%) were positive for the PDGF receptor α and 7/25 cases (28%) were positive for the PDGF receptor β. A positive reaction for PDGF receptors was frequently associated with a simultaneous positive reaction for PDGF- B proteins (Fig. 4), thus suggesting the possible presence of a PDGF- mediatcd autocrinc loop. In fact, among the PDGF-B-positive samples, 2/12 cases (17%) simultaneously expressed the PDGF receptor α subunit, 4/12 cases (33%) simultaneously expressed the PDGF receptor β subunit, and 2/12 cases (17%) simultaneously expressed both subunits of PDGF receptor (Fig. 4). Only one case resulted as positive for both PDGF receptor α and β subunits. without any evidence of PDGFB immunostaining (Fig. 4).

Fig. 3. Mean grcen-to-red ratio profiles for chromosome 22 in the five fibrosarcomas of bone that were analysed by both CGH on chromosomes and microarrays. The threshold values for losses (0.85) and gains (1.17) are indicated to the left and the right of the baseline (1.0). The names of the genes, which were found to be gained by microarray analysis, are shown at their corresponding chromosomal location.

Table 3. Combined analysis of 22q and PDGF-B gene and protein status in nine fibrosarcomas of bone.

PDGF-B immunohistochemistry in OS and MFH of bone

Immunohistochemistry for PDGF-B proteins and PDGF receptors was also extended to the 10 fibroblastic OS and 5 MFH of bone considered in this study. PDGF-B proteins were found in 4/10 fibroblastic OS (40%), three of which were simultaneously positive for both PDGF receptors, whereas the fourth case was PDGF receptor-negative. Among MFH of bone, 1/5 cases (20%) was simultaneously positive for PDGF-B and for both PDGF α and β receptors, whereas the other four cases were completely negative (data not shown).

Discussion

In the past, some authors denied the existence of FS arising in bone and contended that such a lesion was an undifferentiated osteogenic sarcoma. However, more recent studies have indicated that FS of bone should be considered a distinct clinicopathologic entity, which can arise as a primary tumour of the skeleton in either medullary (central) or periosteal (peripheral) locations (Campanacci, 1999; Dorfman and Czerniak, 1995; Kahn and Vigorita, 2002; Papagelopoulos et al., 2002). Our data further support this latter assumption. In fact, immunohistochemical evaluation of non- collagenous proteins of bone confirmed the previously suggested differential immunostaining of FS of bone compared to OS or MFH of bone (Serra et al., 1996), further indicating that they arc similar, closely related, but distinct bone neoplasms, and that immunohistochemical evaluation of osteonectin, osteopontin and osteocalcin might be considered a useful tool for their differential diagnosis.

Differently from MFH of bone or high-grade OS, no biologic or genetic studies have been performed so far for FS of bone because of its rarity, the absence of adequate experimental models, and the lack of techniques that allowed genetic analyses on archived material. Therefore, aim of the present study was the genetic characterisation of this tumour by applying complementary CGH-based and molecular biologic techniques on archived clinical tissue samples. We were able to collect 27 cases of well documented FS of bone, which is a remarkable number of cases when considering the rarity of this neoplasm (at least ten times less frequent than osteosarcoma in the Rizzoli-Institutc experience) (Campanacci, 1999).

Fig. 4. Distribution of PDGF-B and PDGF receptors as revealed by immunohistochemislry in 13 fibrosarcomas of bone, which were positive to at least one of these molecules.

Table 4. Immunohistochemical detection of PDGF-B proteins and PDGF receptors in 25 fibrosarcomas of bone. Data are presented as: positive cases/total cases (%).

As the first step of the study, a screening for chromosomal imbalances and for oncogenc and tumour suppressor gene copy number changes was performed by using CGH-based techniques. The protocols described herein enabled us to isolate a good quality DNA suitable for CGH analyses in one third of the cases, which has to be considered a very good result, considering that only decalcified, paraffin-embedded tumour specimens were available. The possibility to apply these techniques also to decalcified, paraffin-embedded tumour tissue samples may significantly expand the genetic characterisation of bone neoplasms in \the future. These studies have been limited in the past because of the lack of adequate techniques for DNA isolation and genetic analyses suitable for these tumour specimens. This fact becomes even more relevant for the study of very rare bone tumours, for which only archived tissue samples are usually available.

In this study, CGH on chromosomes was used to screen for DMA sequence copy number changes, in order to identify chromosomal regions which were most commonly gained or lost in FS of bone. A low number of DNA sequence copy number changes per tumour was revealed by CGH on chromosomes, which was remarkably lower compared to the results reported so far for high-grade OS and MFH of bone (Knuutila et al., 1998; Tarkkanen et al., 1995, 1999). These results suggest that specific genetic events, rather than the number of genetic aberrations, may be important for the pathogenesis of this neoplasm, which appears to have a higher genetic stability compared to OS and MFH of bone. Moreover, in FS of bone, CGH on chromosomes revealed a high frequency of 22q gain, which was present in four out of five rearranged primary tumours, in the local recurrence and, as the sole genetic aberration, in the lung metastasis. This high incidence of gain of the chromosomal region 22q appears to be a distinctive feature that distinguishes FS of bone from high-grade OS or MFH of bone, in which this genetic aberration has been reported to be only rarely present (Knuutila et al., 199R; Tarkkanen et al., 1999).

The frequent involvement of 22q gain indicates that overrepresentation of gene(s) located at this chromosomal region may be important for the pathogenesis of FS of bone. This assumption was confirmed by the microarray and real-time PCR results, which indicated that gain of the PDGF-B gene (located at 22q12.3-q131) was the most common gene aberration in this neoplasm. Moreover, immunohistochemistry showed that PDGF-B proteins were present in a relevant proportion (48%) of FS of bone, and that higher levels of PDGF-B proteins appear to be associated with gain of the PDGF-E gene. Co-amplification of BCR, the second gene of 22q present on the AmpliOnc microchip, was found in one out of the five cases analysed by microarrays, suggesting that a high-resolution microarray specific for chromosome 22 might reveal additional genes gained in FS of bone.

PDGF-B, the cellular equivalent of the v-sis oncogene, exhibits transforming activity and is a potent mitogen for a number of cell types (Jin et al., 1994). Although its role in oncogenic processes is still not fully understood, the PDGF-B gene has been reported to be involved in the histogenesis of different sarcomas. In particular, the PDGF-B gene has been demonstrated to play an important role in the development of superficial adult soft-tissue FS, which is the soft-tissue counterpart of FS of bone and a rare variant of adult FS sharing several features with dermatofibrosarcoma protuberans (DFSP) (Sheng ct al., 2001), as well as in other proliferative disorders of fibroblastic origin (Simon et al., 1997; Smits et al., 1992; Taniuchi et al., 1997). Expression of the PDGF-B gene has also been reported in several experimental models, including OS (Abdiu et al., 1999; Benini et al., 1999) and MFH of bone (Abdiu et al., 1998) cell lines. Although fusion transcripts of the COL1A1 and PDGFE genes have been identified in superficial adult soft-tissue FS (Sheng ct al., 2001) and DFSP (Simon et al., 1997), no data have been reported so far about the involvement of the PDGF-B gene in the pathogenesis of fibromatous tumours of the skeleton.

Our data indicate that over-representation of the PDGF-B gene and, probably, of other genes mapping to 22q could play a key role in the pathogenesis of FS of bone. Unfortunately, no adequate experimental models are available to confirm and better delineate the importance of this marker in FS of bone and to evaluate whether a PDGF autocrine loop may exist also in this neoplasm, as it has been reported for other bone tumours including OS (McGary et al., 2002; Sulzbacher et al., 2000, 2003), MFH of bone (Abdiu et al., 1998), and Ewing's sarcoma (Uren ct al., 2003). However, our immunohistochemical findings revealed the simultaneous presence of both PDGF-B and PDGF receptors in about one third of the cases, suggesting the possible presence of a PDGF-mediated autocrine loop also in this neoplasm. Therefore, although additional studies are required to validate the role of PDGF-B in FS of bone, these findings suggest that new therapies targeting PDGF/PDGF receptors with drugs like STI571 or CGP57148B might be considered a novel treatment strategy for FS of bone patients, who are unresponsive to conventional therapies.

Acknowledgements. The authors would like to thank Dr. Tom Bhling (Department of Pathology and Medical Genetics, Haartman Institute and Helsinki University Central Hospital, Finland) for having provided two cases of fibrosarcoma of bone; Dr. Alba Balladelli for the spelling correction of this manuscript; Ahbotl-Vysis and Olympus- Italy for the technical support. - This study was supported by grants from: Associazione Italiana per la Ricerca sul Cancro (A.I.R.C.); Istituti Ortopedici Rizzoli (Ricerca Corrente); Italian Ministry of Health (Ricerca Finalizzata); Italian National Council for Research (C.N.R.) and the Paulo Foundation.

References

Abdiu, A., Walz, T. M., Nishikawa, B. K., Wingren, S., Larsson, S. E., Wasteson, A., 1998. Human malignant fibrous histiocytomas in vitro: growth characteristics and their association with expression of mRNA for platelet-derived growth factor, transforming growth factor-alpha and their receptors. Eur. J. Cancer 34, 2094-2100.

Abdiu, A., Wingren, S., Larsson, S. E., Wasteson, A., WaIz, T. M., 1999. Effects of human platelet-derived growth factor-AB on sarcoma growth in vitro and in vivo. Cancer Lett. 141, 39-45.

Benini, S., Baldini, N., Manara, M. C., Chano, T., Serra, M., Rizzi, S., Lollini, P. L., Picci, P., Scotlandi, K., 1999. Redundancy of autocrine loops in human osteosarcoma cells. Int. J. Cancer 80, 581 -588.

Campanacci, M., 1999. Bone and soft tissue tumors. Springer- Verlag, Wien, pp. 149-159.

Capanna, R., Bertoni, R, Bacchini, P., Bacci, G., Guerra, A., Campanacci, M., 1984. Malignant fibrous histiocytoma of bone. The experience at the Rizzoli Institute: report of 90 cases. Cancer 54, 177-187.

Daigo, Y., Chin, S. F., Gorringe, K. L., Bobrow, L. G., Ponder, B. A., Pharoah, P. D., Caldas, C., 2001. Degenerate oligonucleotideprimedpolymerase chain reaction-based array comparative genomic hybridization for extensive amplicon profiling of breast cancers: a new approach for the molecular analysis of paraffin-embedded cancer tissue. Am. J. Pathol. 158, 1623-1631.

Dorfman, H. D., Czerniak, B., 1995. Bone cancers. Cancer 75, 203- 210.

El-Rifai, W., Larramcndy, M. L., Bjorkqvist, A. M., Hemmer, S., Knuutila, S., 1997. Optimization of comparative genomic hybridization using fluorochrome conjugated to dCTP and dUTP nucleolidcs. Lab. Invest. 77, 699-700.

Hattinger, C. M., Revertcr-Branchat, G., Remondini, D., Castellani, G. C, Benini, S., Pascllo, M., Manara, M. C, Scotlandi, K., Picci, P., Serra, M., 2003. Genomic imbalances associated with methotrexate resistance in human osteosarcoma cell lines detected by comparative genomic hybridization-based techniques. Eur. J. Cell Biol. 82, 483-493.

Jin, H. M., Robinson, D. E, Liang, Y., Fahl, W. E., 1994. SIS/ PDGF-B promoter isolation and characterization of regulatory elements necessary for basal expression of the SIS/PDGF-B gene in U2- OS osteosarcoma cells. J. Biol. Chem. 269, 28648-28654.

Kahn, L. B., Vigorita, V., 2002. Fibrosarcoma of hone, in: Flctcher, C. D.M., Unni, K. (Eds.), World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Soft Tissue and Bone. IARC Press, Lyon, pp. 289-290.

Knuutila, S., Bjorkqvist, A. M., Autio, K., Tarkkanen, M., Wolf, M., Monni, O., Szymanska, J., Larramendy, M. L., Tapper, J., Pere, H., El-Rifai, W., Hemmer, S., Wasenius, V. M., Vidgrcn, V., Zhu, Y., 1998. DNA copy number amplifications in human neoplasms: review of comparative genomic hybridization studies. Am. J. Pathol. 152, 1107- 1123.

Mao, X., Lillington, D., Child, F., Russel-Jones, R., Young, B., Whitlaker, S., 2002. Comparative genomic hybridization analysis of primary cutaneous B-ccll lymphomas: Identification of common genomic alterations in disease pathogencsis. Genes Chromosomes Cancer 35, 144-155.

McGary, E. C, Weber, K., Mills, L., Doucet, M., Lewis, V, Lev, D. C, Fidler, I. J., Bar-Eli, M., 2002. Inhibition of platelet-derived growth factor-mediated proliferation of osteosarcoma cells by the novel tyrosine kinase inhibitor STI571. Clin. Cancer Res. 8, 3584- 3591.

Papagelopoulos, P. J., Galanis, E. C., Trantafyllidis, P., Boscainos, P. J., Sim, F. H., Unni, K. K., 2002. Clinicopathologic features, diagnosis, and treatment of fibrosarcoma of bone. Am. J. Orthop. 31, 253-257.

Serra, M., Scotlandi, K., Sollazzo, M. R., Sarti, M., Maurici, D., Benini, S., Picci, P., Bertoni, F., Baldini, N., 1996. Value of immunohistochemical detection of noncollagenous proteins of bone for the diagnosis of bone tumours. Int. J. Oncol. 9, 257-261.

Sheng, W. Q., Hashimoto, H., Okamoto, S., Ishida, T., Meis- Kindblom, J. M., Kindblom, L. G., Hisaoka, M., 2001. Expression of COL1A1-PDGFB fusion transcripts in superficial adult fibrosarcoma suggests a close relationship to dcrmatofibrosarcoma protubcrans. J. Pathol. 194, 88-94.

Simon, M. P., Pedeutour, F., Sirvent, N., Grosgeorgc, J., Minoletti, F., Coindrc, J. M., Terrier-Lacombe, M. J., Mandahl, N., Graver, R. D., Blin, N., Sozzi, G., Turc-Carel, C, O'Brien, K. P., Kedra, D., Fransson, I., Guilbaud, C, Dumanski, J. P., 1997. Deregulation of the platelet-derived growth factor B-chain gene via fusion with collagen gene COLlAl in dermatofibrosarcoma protubcrans and giant-cell fibroblas\toma. Nat. Genet. 15, 95-98.

Smits, A., Funa, K., Vassbotn, F. S., Beausang-Linder, M., af Ekenstam, F., Heldin, C. H., Westermark, B., Nistcr, M., 1992. Expression of platelet-derived growth factor and its receptors in prolifcralive disorders of fibroblastic origin. Am. J. Pathol. 140, 639-648.

Sulzbacher, I., Traxler, M., Mosberger, L, Lang, S., Chott, A., 2000. Platelet-derived growth factor-AA and -alpha receptor expression suggests an autocrine and/or paracrine loop in osteosarcoma. Mod. Pathol. 13, 632-637.

Sulzbacher, I., Birner, P., Trieb, K., Traxler, M., Lang, S., Chott, A., 2003. Expression of platelet-derived growth factor-AA is associated with tumor progression in osteosarcoma. Mod. Pathol. 16, 66-71.

Taconis, W. K.,Mulder, J. D., 1984. Fibrosarcoma and malignant fibrous hisliocytoma of long bones: radiographie features and grading. Skeletal Radiol. 11, 237-245.

Taniuchi, K., Yamada, Y, Nonomura, A., Takehara, K., 1997. Immunohistochcmical analysis of platelet-derived growth factor and its receptors in fibrohistiocytic tumors. J. Cutan. Pathol. 24, 393- 397.

Tarkkanen, M., Karhu, R., Kallioncmi, A., Elomaa, I., Kivioja, A. H., Nevalainen, J., Bhling, T, Karaharju, E., Hyytinen, E., Knuutila, S., Kallioncmi, O.-P, 1995. Gains and losses of DNA sequences in osteosarcoma by comparative genomic hybridization. Cancer Res. 55, 1334-1338.

Tarkkanen, M., Elomaa, I., Blomqvist, C., Kivioja, A. H., Kellokumpu-Lehtinen, P., Bohling, T, Valle, I, Knuutila, S., 1999. DNA sequence copy number increase at 8q: a potential new prognostic marker in high-grade osteosarcoma. Int. J. Cancer 84, 114-121.

Uren, A., Merchant, M. S., Sun, C. J., Vitolo, M. I., Sun, Y, Tsokos, M., Illei, P. B., Ladanyi, M., Passaniti, A., Mackall, C, Toretsky, J. A., 2003. Bcta-platelet-derived growth factor receptor mediates motility and growth of Ewing's sarcoma cells. Oncogene 22, 2334-2342.

Claudia Maria Hattinger1)a, Maija Tarkkanen(b), c, Stefania Benini(a), Michela Pasello(a), Giuseppina Stoico(a), Patrizia Bacchini(d), Sakari Knuutila(b0, Katia Scotlandi(a), Piero Picci(a), Massimo Serra(a)

a Laboratorio di Ricerca Oncologica, Istituti Ortopedici Rizzoli, Bologna, Italy

b Departments of Pathology and Medical Genetics, Haartman Institute and HUSLAB, University of Helsinki and Helsinki University Central Hospital, Finland

c Department of Oncology, Helsinki University Central Hospital, Helsinki, Finland

d Servizio di Anatomia Patologica, Istituti Ortopedici Rizzoli, Bologna, Italy

Received May 18, 2004

Received in revised version July 13, 2004

Accepted July 14, 2004

1) Corresponding author: Dr. Claudia M. Hattingcr, Laboratorio di Riccrca Oncologica, Istituti Ortopcdici Rizzoli, Via di Barbiano 1/ 10, 1-40136 Bologna, Italy, e-mail: claudia.hattinger@ior.it, Fax: + 39051 6366761.

Copyright Urban & Fischer Verlag Sep 2004

Source: European Journal of Cell Biology

More News in this Category