Purpose: Angiomatoid fibrous histiocytoma (AFH) is a low-grade mesenchymal neoplasm which usually occurs in children and adolescents. Either FUS-ATF1 or EWSR1-ATF1 have been detected in the few cases published, pointing to the interchangeable role of FUS and EWSR1 in this entity. EWSR1-ATF1 also represents the most frequent genetic alteration in clear cell sarcoma, suggesting the existence of a molecular homology between these two histotypes. We investigated the presence of EWSR1-CREB1, recently found in gastrointestinal clear cell sarcoma, and FUS-CREB1, as well as the already reported FUS-ATF1 and EWSR1-ATF1 in a series of AFH.

Experimental Design: Fourteen cases were analyzed by fluorescence in situ hybridization (FISH) on paraffin-embedded tissue sections, using a commercial EWSR1 probe and custom-designed probes for FUS, ATF1, and CREB1. In two cases, four-color FISH was also done. Reverse transcription-PCR for the four hypothetical fusion genes was done in one case, for which frozen material was available.

Results: Thirteen cases showed rearrangements of both EWSR1 and CREB1, whereas one case showed the rearrangement of both EWSR1 and ATF1. Four-color FISH confirmed the results in two selected cases. Reverse transcription-PCR showed EWSR1-CREB1 transcript in the case analyzed.

Conclusion: We identified the presence of either EWSR1-CREB1 or EWSR1-ATF1 in all the cases, strengthening the concept of chromosomal promiscuity between AFH and clear cell sarcoma. Either the occurrence of a second unknown tumor-specific molecular event or, perhaps more likely, divergent differentiation programs of the putatively distinct precursor cells of AFH and clear cell sarcoma might be invoked in order to explain the two different phenotypes.

Angiomatoid fibrous histiocytoma (AFH) represents a low-grade mesenchymal neoplasm of uncertain differentiation (1), which usually occurs in the extremities of children and young adults. Morphologically, it is a multinodular proliferation of bland spindle to ovoid eosinophilic cells, sometimes lining pseudoangiomatoid spaces and surrounded by a thick fibrous pseudocapsule, often featuring a prominent lymphoplasmacytic infiltrate. Immunohistochemically, the most relevant finding is represented by the expression of desmin in ∼40% of cases, suggesting myogenic differentiation (2, 3). The recurrence rate is between 2% and 11%, and the metastatic rate is <1%. In the context of an uncommon tumor type with a very low rate of metastasis, it is unsurprising that no histologic or immunophenotypic features have thus far been identified to predict patients' outcome.

Until recently, only five cases of AFH studied by cytogenetics (4, 5) and/or molecular genetics (47) have been published in detail. Two of them were characterized by the expression of an FUS-ATF1 fusion gene (4, 6), resulting from a t(12;16) (q13;p11) (ref. 4), whereas in the remaining three cases, a EWSR1-ATF1 fusion gene was detected (5, 7), resulting from a t(12;22)(q13;q12) (ref. 5). EWSR1 and FUS are members of the TET family, characterized by COOH-terminal RNA-binding domains, which are believed to act as adaptors between transcription and RNA processing (8). The ATF1 gene encodes a member of the CREB-ATF basic leucine-zipper family of transcription factors, which regulate cell gene expression via homodimeric or heterodimeric DNA-binding to cyclic AMP response elements (9). As a consequence of the fusion of the NH2 terminus of FUS/EWSR1 with the COOH terminus of ATF1 and the replacement of FUS/EWSR1 RNA-binding domain by ATF1 DNA-binding domain, FUS/EWSR1-ATF1 chimeric proteins might activate transcription independently of cyclic AMP induction.

Interestingly, both FUS and EWSR1 are involved in recombinations as 5′ partners in multiple different tumor types and seem to be functionally interchangeable at least in some of them. In fact, besides the fusion with ATF in AFH, both of them rearrange with ERG in Ewing's sarcoma (10, 11) and with DDIT3 in myxoid liposarcoma (12, 13). Furthermore, four different types of EWSR1-ATF1 transcript, only three of which were in-frame, including one identical to that reported in AFH (5, 7), have been detected in clear cell sarcoma of soft tissue (14, 15), suggesting the existence of shared genetic mechanisms between AFH and clear cell sarcoma. The recent description of a novel fusion gene, i.e., EWSR1-CREB1, in gastrointestinal clear cell sarcoma, raises the possibility that ATF1 and CREB1 might also be functionally interchangeable (16).

The genetic parallelism between AFH and clear cell sarcoma, and the functional homology of EWSR1 and FUS, represent a rationale to investigate, in addition to the already reported fusion genes (i.e., FUS-ATF1 and EWSR1-ATF1), the presence of EWSR1-CREB1 and FUS-CREB1 in AFH. In fact, recent data presented in abstract form earlier this year (17, 18), and now very recently published in full (19), have shown that FUS-ATF1 and EWSR1-ATF1 characterize only a subgroup of AFH and that the most frequent event in AFH indeed seems to be an EWSR1-CREB1 rearrangement.

In this study, we assessed the prevalence of FUS/EWSR1-ATF1/CREB1 fusion genes as detected by fluorescence in situ hybridization (FISH) in a series of 14 cases of AFH. We also investigated the presence of all four possible fusion transcripts by reverse transcription-PCR (RT-PCR) in one case for which frozen material was available.

Patients and tumor samples

Twenty-two cases of AFH with paraffin blocks were retrieved from the consultation files of two of the authors (C.D.M. Fletcher and A.P. Dei Tos) and from the Pathology Department of the Catholic University of Leuven (Leuven, Belgium). Among these, 14 cases, in which adequate tumor tissue for FISH analysis was available, were selected for our study. In addition, frozen tissue was available in one case. Six patients were female and eight were male. Patients' ages ranged from 4 to 37, with a median of 14. Eight tumors were located in the lower extremities, four in the upper extremities and two in the trunk. Clinical information about the depth of the tumors were available in all the cases but one. Four tumors were deep-seated, including three tumors located below the fascia and one tumor located in the prepatellar region, whereas nine tumors were subcutaneous. Patients' clinical data are summarized in Table 1. Morphologically, all lesions showed typical features of AFH, being composed of multinodular proliferation of spindle and ovoid palely eosinophilic cells with a syncytial appearance (Fig. 1A and B). In five cases, prominent pseudovascular spaces filled with RBC were observed (Fig. 1C). In four cases, a peripheral lymphoplasmacytic infiltrate and stromal fibrosis were prominent. Cellular pleomorphism was diffuse in one (Fig. 1D) and focal in three cases.

Table 1.

Clinical data, FISH, and RT-PCR results

Case nos.Sex/ageSite/depthTwo-color FISH
Four-color FISHRT-PCR
EWSR1FUSATF1CREB1
F/12 Buttock/subcutaneous nnR nnR — — 
M/22 Thigh/subcutaneous nnR nnR EWSR1-CREB1 — 
F/16 Knee/subcutaneous nnR nnR — — 
M/10 Thigh/subcutaneous nnR nnR — — 
F/30 Arm/subcutaneous nnR nnR — — 
F/10 Arm/subfascial nnR nnR — — 
M/7 Thigh/subcutaneous nnR nnR — — 
F/26 Forearm/subfascial nnR nnR EWSR1-ATF1 — 
M/4 Forearm/subcutaneous nnR nnR — — 
10 F/4 Back/subcutaneous R(pl) nnR(pl) nnR(pl) R(pl) — — 
11 M/18 Thigh/NA nnR nnR — — 
12 M/37 Thigh/subcutaneous nnR nnR — — 
13 M/7 Thigh/prepatellar nnR nnR — — 
14 M/35 Thigh/subfascial nnR nnR — EWSR1-CREB1 
Case nos.Sex/ageSite/depthTwo-color FISH
Four-color FISHRT-PCR
EWSR1FUSATF1CREB1
F/12 Buttock/subcutaneous nnR nnR — — 
M/22 Thigh/subcutaneous nnR nnR EWSR1-CREB1 — 
F/16 Knee/subcutaneous nnR nnR — — 
M/10 Thigh/subcutaneous nnR nnR — — 
F/30 Arm/subcutaneous nnR nnR — — 
F/10 Arm/subfascial nnR nnR — — 
M/7 Thigh/subcutaneous nnR nnR — — 
F/26 Forearm/subfascial nnR nnR EWSR1-ATF1 — 
M/4 Forearm/subcutaneous nnR nnR — — 
10 F/4 Back/subcutaneous R(pl) nnR(pl) nnR(pl) R(pl) — — 
11 M/18 Thigh/NA nnR nnR — — 
12 M/37 Thigh/subcutaneous nnR nnR — — 
13 M/7 Thigh/prepatellar nnR nnR — — 
14 M/35 Thigh/subfascial nnR nnR — EWSR1-CREB1 

Abbreviations: R, rearranged; nnR, nonrearranged; (pl), polyploid; NA, not available.

Fig. 1.

Morphologic features of AFH. A, a case of AFH showing typical pushing borders and peripheral stromal fibrosis with a lymphoid aggregate (case 7; magnification, ×10). B, a high-power view of the same case in (A), showing the typical spindle and ovoid palely eosinophilic cells (case 7; magnification, ×40). C, another case of AFH characterized by the typical pseudovascular spaces with RBCs (case 10; magnification, ×20). D, the same case in (C) with diffuse nuclear pleomorphism (case 10; magnification, ×20).

Fig. 1.

Morphologic features of AFH. A, a case of AFH showing typical pushing borders and peripheral stromal fibrosis with a lymphoid aggregate (case 7; magnification, ×10). B, a high-power view of the same case in (A), showing the typical spindle and ovoid palely eosinophilic cells (case 7; magnification, ×40). C, another case of AFH characterized by the typical pseudovascular spaces with RBCs (case 10; magnification, ×20). D, the same case in (C) with diffuse nuclear pleomorphism (case 10; magnification, ×20).

Close modal

FISH

Probe selection and labeling. Two-color FISH for EWSR1, FUS, ATF1, and CREB1 was done on each case. To detect possible rearrangements of FUS gene at region 16p11, two bacterial artificial chromosomes (BAC) clones, RP11-196G11, proximal to the breakpoint, and RP11-120K18, distal to the breakpoint, were selected. To investigate ATF1 at region 12q13.13, BAC clones RP11-189H16 proximal to the breakpoint and RP11-407N8 distal to the breakpoint, were chosen. For CREB1 at region 2q33.3, we selected two contiguous fosmid clones distal to the breakpoint, G248P81788B12 and G248P89268A6, which were pooled to enhance the signal, and a BAC clone, RP11-167C7 proximal to the breakpoint. All BAC clones were obtained from Wellcome Trust, Sanger Institute. Fosmid clones were obtained from BACPAC Resources. Detailed data about BAC and fosmid clones are summarized in Table 2. DNA from all the BAC and fosmid clones was extracted according to the High pure plasmid isolation kit (Roche Diagnostics GmbH). For EWSR1, a commercial split-signal EWSR1 probe (SPoT-Light EWS Translocation Probes; Zymed, Invitrogen), which contains a digoxigenin-labeled proximal part and a biotin-labeled distal part of the EWSR1 regions, was used.

Table 2.

Data about the BAC and fosmid clones selected

CloneGenBank IDStart position*End position*Size (kb)Probe regionProx-Dist probes gap (kb)Locus
RP11-196G11 AC135050 30862945 31045174 182.2 FUS (dist) 118.5 16p11.2 
RP11-120K18 AC093520 31163677 31329206 165.5 FUS (prox)  16p11.2 
RP11-189H16 AC078818 49293862 49433792 139.9 ATF1 (prox) 117 12q13.13 
RP11-407N8 AC008121 49550796 49654781 103.9 ATF1 (dist)  12q13.13 
RP11-167C7 AC009226 207861196 207999773 138.5 CREB1 (prox) 180 2q33.3 
G248P81788B12 AC079767 208191367 208230251 40.8 CREB1 (dist)  2q33.3 
G248P89268A6 AC079767 208215933 208254245 38.3 CREB1 (dist)  2q33.3 
CloneGenBank IDStart position*End position*Size (kb)Probe regionProx-Dist probes gap (kb)Locus
RP11-196G11 AC135050 30862945 31045174 182.2 FUS (dist) 118.5 16p11.2 
RP11-120K18 AC093520 31163677 31329206 165.5 FUS (prox)  16p11.2 
RP11-189H16 AC078818 49293862 49433792 139.9 ATF1 (prox) 117 12q13.13 
RP11-407N8 AC008121 49550796 49654781 103.9 ATF1 (dist)  12q13.13 
RP11-167C7 AC009226 207861196 207999773 138.5 CREB1 (prox) 180 2q33.3 
G248P81788B12 AC079767 208191367 208230251 40.8 CREB1 (dist)  2q33.3 
G248P89268A6 AC079767 208215933 208254245 38.3 CREB1 (dist)  2q33.3 

Abbreviations: Prox, proximal to the breakpoint; dist, distal to the breakpoint.

*

According to the Human March 2006 Assembly.

For the two-color FISH, the probes were labeled either with biotin-11-dUTP or digoxigenin-11-dUTP (Boehringer Mannheim) using a nick translation labeling kit (Boehringer Mannheim and Enzo) according to the manufacturer's instructions. All probes were hybridized on metaphase spreads of lymphocytes to confirm chromosomal location and specificity (data not shown).

For four-color FISH, the CREB1 proximal probe RP11-167C7 and the ATF1 proximal probe RP11-189H16 were labeled by nick translation with Cy5, whereas CREB1 distal probes G248P81788B12 and G248P89268A6 and ATF1 distal probe RP11-407N8 were labeled with Cy3. Four-color FISH for EWSR1-CREB1 and for EWSR1-ATF1 was done on two different cases, in which either EWSR1 and CREB1 or EWSR1 and ATF1, respectively, were rearranged by two-color FISH (Fig. 3).

Prehybridization, hybridization, and posthybridization treatment of paraffin-embedded tissue sections. Four-micrometer deparaffinized ethanol-dehydrated tissue sections were boiled in 50 mmol/L of Tris-HCl/2 mmol/L EDTA-buffer at pH 9 with a pressure of 1 atm. for 7 min and rinsed with 2× SSC at room temperature. They were subsequently soaked in preheated 1 mol/L of sodium sulfocyanate at 80°C for 45 min, allowed to cool in bidistillated water and rinsed with 2× SSC. They were then incubated with RNase (100 μg/mL) at 37°C, washed twice in 2× SSC, incubated with 4% pepsin/0.02 mol/L HCl at 37°C for 20 min, washed three times with PBS, and ethanol-dehydrated. The labeled DNA probes, except for EWSR1, were diluted in hybridization mixture (60% formamide, 10% dextran sulfate, and 2× SSC; pH 7) to a final concentration of 1 ng/μL. Eleven microliters of hybridization mixture were added onto the slides and probe and target DNA were denatured for 10 min at 80°C. After overnight hybridization at 37°C, two 5-min washes with 2× SSC/0.1% Tween 20 at 37°C, two 5-min washes with 50% formamide/2× SSC at 44°C, one 5-min wash in 2× SSC/0.1 Tween 20 at 44°C, two 5-min washes with 0.1× SSC at 60°C, and one 5-min wash in 1× TBS/0.05% Tween 20 (TNT) at room temperature were done.

For the two-color FISH, all the probes were detected with an indirect method. Sections were incubated with Cy3-labeled streptavidin (1:500; Sigma) and FITC-labeled mouse antidigoxigenin antibody (1:250; Sigma) at 37°C for 30 min. After three 5-min washes with TNT, sections were incubated with FITC-labeled rabbit anti-mouse antibody (1:1,000; Sigma).

For the four-color FISH, because ATF1 and CREB1 probes were directly labeled, only the EWSR1 commercial probe was revealed with an indirect method. Sections were incubated with FITC-labeled mouse antidigoxigenin antibody (1:250; Sigma) and streptavidin LaserPRO 790 (1:250; Molecular Probes, Invitrogen) for 30 min at 37°C. After three 5-min washes with TNT, they were incubated with FITC-labeled rabbit anti-mouse antibody (1:1,000; Sigma) and biotin-labeled goat anti-streptavidin antibody (1:500; Vector Laboratories) with the same conditions. Sections were subsequently washed again three times with TNT and incubated with streptavidin LaserPRO 790 (1:250; Molecular Probes, Invitrogen) for 30 min at 37°C. All the antibodies used were diluted with 0.5% Boehringer blocking reagent/1× TBS (TND). Either for the two-color FISH or for the four-color FISH, after the last antibody incubation, sections were washed thrice with TNT, ethanol-dehydrated, and counterstained with 4′,6-diamidino-2-phenylindole.

Scoring. FISH evaluation was done by counting at least 100 nuclei for each slide. In the samples which showed an abundant reactive lymphoid infiltrate surrounding the tumor, both normal and tumor regions were scored. Nuclei with an incomplete set of signals were omitted from the score. Signals were considered colocalized when their distance was equal to or smaller than the size of the hybridization signal. The mean percentage of cells with false split signals for FUS, ATF1, EWSR1, and CREB1 probes, calculated by scoring the nonneoplastic tissue at the peripheral areas, was <1%.

For the two-color FISH, only cases with a complete pattern of two separate red and green signals, and one colocalized or fused signal were considered as rearranged, whereas cases with two colocalized or fused signals were considered as nontranslocated. For the four-color FISH, on nontranslocated cells, the simultaneous hybridization of the four differentially labeled probes was expected to result in a colocalization or fusion of the green and red signals for the normal EWSR1 gene and of the magenta and blue signals either for the normal ATF1 or for the normal CREB1. In case of either EWSR1-CREB1 or EWSR1-ATF1 rearrangement, the green signal of the proximal part of EWSR1 was expected to colocalize with the blue signal of the distal part of either CREB1 or ATF1, respectively. Additionally, in case of a balanced translocation, the magenta signal of the proximal part of either CREB1 or ATF1 was expected to colocalize with the red signal of the distal part of EWSR1.

RT-PCR

Frozen tissue was available in one case. RNA was extracted as previously described (20). Two micrograms of total RNA was reverse-transcribed with avian myeloblastosis virus reverse transcriptase (Roche). RT-PCR for FUS-ATF1, FUS-CREB1, EWSR1-ATF1, and EWSR1-CREB1 was done. RT-PCR for each of the nontranslocated genes was also done. A clear cell sarcoma cell line cDNA was used as a positive control for EWSR1-ATF1 translocation. All PCR amplifications were done in a 25 μL reaction volume containing 20 mmol/L of Tris-HC1 (pH 8.0), 1.5 mmol/L of MgC12, 0.2 mmol/L of each deoxynucleotide triphosphate, 2 units of AmpliTaq polymerase (Applied Biosystems), 0.5 μmol/L of each of the forward and reverse primers, and 1 μL of 1:5 diluted, synthesized cDNA. Detailed information about position, length, and direction of the primers used are reported in Table 3. Different primer combinations were used. To detect FUS-ATF1 rearrangement, the primers 1 (4) and 8 (6) and the primers 2 (4) and 9 (4) were used with an annealing temperature of 60°C and 64°C, respectively. To detect FUS-CREB1 rearrangement, the primers 1 (4) and 11 (16) and the primers 2 (4) and 12 (16) were used with an annealing temperature of 60°C and 64°C, respectively. To investigate possible EWSR1-ATF1 and EWSR1-CREB1 rearrangements, the primers 5 (our) and 8 (6) and the primers 5 (our) and 12 (16) were used with an annealing temperature of 60°C. To detect the normal genes, the primers 1 (4) and 3 (4) for FUS, the primers 7 (4) and 9 (4) for ATF1, and the primers 10 (our) and 11 (16) for CREB1 were used with an annealing temperature of 53°C, whereas the primers 7 (4) and 8 (6) for ATF1, the primers 1 (4) and 4 (4) for FUS and the primers 5 (our) and 6 (our) for EWSR1 were used with an annealing temperature of 58°C. All PCR reactions consisted of initial denaturation of 4 min, 35 cycles of 30 s at 94°C, 30 s at the annealing temperature chosen for each primer set and 1 min at 72°C, followed by a final extension of 10 min at 72°C. Ten microliters of the PCR products were analyzed by electrophoresis through 1.5% agarose gels stained with ethidium bromide and photographed. For sequence analysis, the amplified cDNA fragment was purified from the gel, using the Qiagen gel extraction kit (Qiagen) and directly sequenced using the dideoxy procedure with an ABI Prism BigDye terminator cycle sequencing ready reaction kit (PE Applied Biosystems) on the Applied Biosystems model 373A DNA sequencing system.

Table 3.

Primers used

PrimersGeneAccession no.PositionDirectionReference no.
FUS NM_004960 552-575 Forward 
FUS NM_004960 336-358 Forward 
FUS NM_004960 986-1009 Reverse 
FUS NM_004960 1068-1092 Reverse 
EWSR1 NM_005243 762-783 Forward — 
EWSR1 NM_005243 1542-1561 Reverse — 
ATF1 NM_005171 180-200 Forward 4* 
ATF1 NM_005171 649-627 Reverse 
ATF1 NM_005171 918-891 Reverse 4* 
10 CREB1 BC095407 295-314 Forward — 
11 CREB1 BC095407 737-757 Reverse 16 
12 CREB1 BC095407 791-812 Reverse 16 
PrimersGeneAccession no.PositionDirectionReference no.
FUS NM_004960 552-575 Forward 
FUS NM_004960 336-358 Forward 
FUS NM_004960 986-1009 Reverse 
FUS NM_004960 1068-1092 Reverse 
EWSR1 NM_005243 762-783 Forward — 
EWSR1 NM_005243 1542-1561 Reverse — 
ATF1 NM_005171 180-200 Forward 4* 
ATF1 NM_005171 649-627 Reverse 
ATF1 NM_005171 918-891 Reverse 4* 
10 CREB1 BC095407 295-314 Forward — 
11 CREB1 BC095407 737-757 Reverse 16 
12 CREB1 BC095407 791-812 Reverse 16 
*

Slightly modified from the ones previously reported.

FISH. Two-color FISH for FUS, EWSR1, ATF1, and CREB1 genes was done on 14 cases. Thirteen cases showed rearrangement of both EWSR1 and CREB1 genes, indicating the occurrence of a t(2;22) in these cases (Fig. 2A and B). In nine cases, the number of rearranged cells ranged between 50% and 95%. In four cases, which were morphologically characterized by an abundant inflammatory infiltrate, the number of translocated cells ranged between 10% and 20%. Because these cases showed similar split ratios when analyzed by multiple FISH tests using EWSR1 and CREB1 probes in separate reactions, the presence of a small tumor clone diluted in an abundant reactive background was convincingly proved. Moreover, these cases showed a <1% split signal when analyzed by the ATF1 and FUS FISH.

Fig. 2.

Overview of interphase FISH results on formalin-fixed paraffin-embedded tissue. A, a case showing EWSR1 rearrangement by interphase FISH using EWSR1 break-apart probe (Zymed). Proximal and distal EWSR1 regions are hybridized by separate probes (green and red, respectively). The presence of separate green and red signals indicates a rearrangement of the EWSR1 region on chromosome 22 (arrows), whereas colocalized signals represent the normal EWSR1 region on chromosome 22 (arrowhead). B, the same case as in (A) with CREB1 rearrangement. The distal CREB1 region is hybridized to two contiguous fosmid probes, G248P81788B12 and G248P89268A6, labeled with biotin-11-dUTP and revealed by Cy3-labeled streptavidin (red), whereas the proximal CREB1 region is hybridized to the BAC probe RP11-167C7, labeled with digoxigenin-11-dUTP, and revealed by FITC-labeled mouse antidigoxigenin antibody (green). Although separate green and red signals indicate the rearrangement of CREB1 region on chromosome 2 (arrows), colocalized signals represent the normal CREB1 region on chromosome 2 (arrowhead). C, only one case showed ATF1 rearrangement. The proximal ATF1 region is hybridized to the BAC probe RP11-189H16, labeled with digoxigenin-11-dUTP, and revealed by FITC-labeled mouse antidigoxigenin antibody (green), whereas the distal ATF1 region is hybridized to the BAC probe RP11-407N8, labeled with biotin-11-dUTP, and revealed by Cy3-labeled streptavidin (red). Rearrangement of the ATF1 region on chromosome 12 is indicated by the presence of separate red and green signals (arrows), whereas colocalized signals represent normal ATF1 region on chromosome 12 (arrowhead).

Fig. 2.

Overview of interphase FISH results on formalin-fixed paraffin-embedded tissue. A, a case showing EWSR1 rearrangement by interphase FISH using EWSR1 break-apart probe (Zymed). Proximal and distal EWSR1 regions are hybridized by separate probes (green and red, respectively). The presence of separate green and red signals indicates a rearrangement of the EWSR1 region on chromosome 22 (arrows), whereas colocalized signals represent the normal EWSR1 region on chromosome 22 (arrowhead). B, the same case as in (A) with CREB1 rearrangement. The distal CREB1 region is hybridized to two contiguous fosmid probes, G248P81788B12 and G248P89268A6, labeled with biotin-11-dUTP and revealed by Cy3-labeled streptavidin (red), whereas the proximal CREB1 region is hybridized to the BAC probe RP11-167C7, labeled with digoxigenin-11-dUTP, and revealed by FITC-labeled mouse antidigoxigenin antibody (green). Although separate green and red signals indicate the rearrangement of CREB1 region on chromosome 2 (arrows), colocalized signals represent the normal CREB1 region on chromosome 2 (arrowhead). C, only one case showed ATF1 rearrangement. The proximal ATF1 region is hybridized to the BAC probe RP11-189H16, labeled with digoxigenin-11-dUTP, and revealed by FITC-labeled mouse antidigoxigenin antibody (green), whereas the distal ATF1 region is hybridized to the BAC probe RP11-407N8, labeled with biotin-11-dUTP, and revealed by Cy3-labeled streptavidin (red). Rearrangement of the ATF1 region on chromosome 12 is indicated by the presence of separate red and green signals (arrows), whereas colocalized signals represent normal ATF1 region on chromosome 12 (arrowhead).

Close modal

In one case, which was morphologically characterized by a remarkable degree of pleomorphism (Fig. 1D), besides cells with one colocalized green/red signal, one green signal, and one red signal; cells with two colocalized signals, two red signals, and one green signal; and cells with two colocalized signals, two green signal, and one red signal were also seen, suggesting that this was a polyploid case with an unbalanced translocation. When hybridized with FUS and ATF1 probes, this case showed a proportion of nuclei with multiple colocalized signals, confirming that it was polyploid. One of 14 cases showed the rearrangement of both EWSR1 and ATF1 genes in 60% of cells, pointing to the presence of a t(12;22) in this case (Fig. 2C). Notably, none of the cases showed a rearrangement of the FUS region.

Four-color FISH was done on two selected cases, in which the signals colocalized, as expected (Fig. 3A and B), confirming EWSR1-CREB1 rearrangement in one case (Fig. 3C) and EWSR1-ATF1 rearrangement in another case, which had been already detected by two-color FISH. FISH results are summarized in Table 1.

Fig. 3.

A, four-color FISH labeling scheme. EWSR1 commercial probe (Zymed) contains a digoxigenin-labeled proximal part and a biotin-labeled distal part and was revealed by a three-step indirect method, which resulted in a EWSR1 probe showing a FITC-labeled proximal part (green) and a LaserPro-labeled distal part (red). CREB1 proximal probe was directly labeled with Cy5 (magenta), whereas CREB1 distal probes were directly labeled with Cy3 (blue). B, schematic representation of EWSR1-CREB1 rearrangement detected by four-color FISH. Colocalization or fusion of the green signal of the proximal part of EWSR1 with the blue signal of the distal part of CREB1 was expected in the case of EWSR1-CREB1 rearrangement. Due to the balanced nature of the rearrangement, a fusion between the proximal CREB1 (magenta) and distal EWSR1 (red) should be present. C, EWSR1-CREB1 rearrangement detected by four-color FISH. The magenta and blue colocalized signals indicate the normal CREB1 region on chromosome 2, whereas the red and green signals indicate EWSR1 normal region on chromosome 22 (white arrowheads). The green and blue colocalized signals indicate the rearrangement involving proximal EWSR1 region on chromosome 22 and distal CREB1 region on chromosome 2. The colocalization of magenta and red indicates the CREB1-EWSR1 reciprocal fusion gene (blue arrows).

Fig. 3.

A, four-color FISH labeling scheme. EWSR1 commercial probe (Zymed) contains a digoxigenin-labeled proximal part and a biotin-labeled distal part and was revealed by a three-step indirect method, which resulted in a EWSR1 probe showing a FITC-labeled proximal part (green) and a LaserPro-labeled distal part (red). CREB1 proximal probe was directly labeled with Cy5 (magenta), whereas CREB1 distal probes were directly labeled with Cy3 (blue). B, schematic representation of EWSR1-CREB1 rearrangement detected by four-color FISH. Colocalization or fusion of the green signal of the proximal part of EWSR1 with the blue signal of the distal part of CREB1 was expected in the case of EWSR1-CREB1 rearrangement. Due to the balanced nature of the rearrangement, a fusion between the proximal CREB1 (magenta) and distal EWSR1 (red) should be present. C, EWSR1-CREB1 rearrangement detected by four-color FISH. The magenta and blue colocalized signals indicate the normal CREB1 region on chromosome 2, whereas the red and green signals indicate EWSR1 normal region on chromosome 22 (white arrowheads). The green and blue colocalized signals indicate the rearrangement involving proximal EWSR1 region on chromosome 22 and distal CREB1 region on chromosome 2. The colocalization of magenta and red indicates the CREB1-EWSR1 reciprocal fusion gene (blue arrows).

Close modal

RT-PCR. RT-PCR for FUS-ATF1, FUS-CREB1, EWSR1-ATF1, and EWSR1-CREB1 fusion genes was attempted in one case for which frozen material was available. A strong band of 453 bp was obtained (Fig. 4A), by using a forward primer located in exon 6 of EWSR1 and a reverse primer located in exon 7 of CREB1 (Fig. 4B). Direct sequencing of the product confirmed the presence of a EWSR1-CREB1 chimeric transcript with a junction between EWSR1 exon 7 and CREB1 exon 7 (Fig. 4C). None of the other three possible chimera (EWSR1-ATF1, FUS-CREB1, and FUS-ATF1) were identified. PCR for EWSR1-ATF1 showed a band of the expected size only in the clear cell sarcoma cell line used as control. As control reactions, PCR for the normal genes showed bands of the expected sizes: 457 and 525 bp for FUS when primer 1 was used together with primers 3 and 4, respectively, 738 and 469 bp for ATF1 when primer 7 was used with primers 9 and 8, respectively (Fig. 4A), and 462 bp for CREB1 and 799 bp for EWSR1.

Fig. 4.

A, RT-PCR results. RT-PCR was done in one case (case 14) for which frozen tissue was available. Detection of strong bands of the expected size (469 bp) obtained with the use of specific primers for ATF1 normal gene, primers 7 and 8, confirmed the good quality of the cDNA extracted. RT-PCR with specific primers for EWSR1-ATF1 fusion gene, primers 5 and 8, showed a band of the expected size only in the clear cell sarcoma cell line cDNA, used as positive control. RT-PCR with specific primers for EWSR1-CREB1 fusion gene, primers 5 and 12, showed a strong band of 453 bp in case 14. B, schematic representation of EWSR1-CREB1 chimeric transcript. EWSR1-CREB1 chimeric transcript consisted of a junction between EWSR1 exon 7 and CREB1 exon 7. Primer 5 was located in exon 6 of EWSR1 and primer 12 was located in exon 7 of CREB1. C, direct sequencing of the RT-PCR product. The EWSR1-CREB1 fusion involved codon 265 of EWSR1 (NM_005243) and codon 169 of CREB1 (BC095407).

Fig. 4.

A, RT-PCR results. RT-PCR was done in one case (case 14) for which frozen tissue was available. Detection of strong bands of the expected size (469 bp) obtained with the use of specific primers for ATF1 normal gene, primers 7 and 8, confirmed the good quality of the cDNA extracted. RT-PCR with specific primers for EWSR1-ATF1 fusion gene, primers 5 and 8, showed a band of the expected size only in the clear cell sarcoma cell line cDNA, used as positive control. RT-PCR with specific primers for EWSR1-CREB1 fusion gene, primers 5 and 12, showed a strong band of 453 bp in case 14. B, schematic representation of EWSR1-CREB1 chimeric transcript. EWSR1-CREB1 chimeric transcript consisted of a junction between EWSR1 exon 7 and CREB1 exon 7. Primer 5 was located in exon 6 of EWSR1 and primer 12 was located in exon 7 of CREB1. C, direct sequencing of the RT-PCR product. The EWSR1-CREB1 fusion involved codon 265 of EWSR1 (NM_005243) and codon 169 of CREB1 (BC095407).

Close modal

Until recently, five cases of AFH with cytogenetic and molecular data have been published in detail. Two of them were characterized by t(12;16) (q13;p11) with FUS-ATF1 fusion gene (4, 6), whereas the remaining three cases harbored a t(12;22) (q13;q12) with EWSR1-ATF1 fusion gene (5, 7). Because it was known that AFH and clear cell sarcoma share the EWSR1-ATF1 fusion gene (5, 7), and because the EWSR1-CREB1 fusion gene had recently been detected in clear cell sarcoma of the gastrointestinal tract (16), we selected specific probes for FUS, ATF1, and CREB1 genes and used them together with a EWSR1 commercial probe to test 14 cases of formalin-fixed paraffin-embedded AFH for FUS/EWSR1-ATF1/CREB1 rearrangements. In our series, all cases but one showed EWSR1-CREB1 rearrangement, indicating the presence of a t(2;22)(q34;q12). One of the EWSR1-CREB1 rearranged cases, which was characterized by a remarkable degree of nuclear pleomorphism, proved to be a polyploid case, which might bear an unbalanced translocation, suggesting, in some cases, the occurrence of more complex karyotypes. However, this molecular feature does not seem to adversely affect the prognosis because the patient has not shown any sign of relapse 7 years after the excision. Furthermore, we confirmed the presence of EWSR1-CREB1 fusion transcript by RT-PCR in the case for which frozen material was available. The transcript was characterized by the fusion of exon 7 of EWSR1 and exon 7 of CREB1, similarly to the one reported in cases of gastrointestinal clear cell sarcoma (16). Although there were no critical methodologic differences between our study and the previous reports, we could not find any FUS-ATF1 rearrangement and only one case with EWSR1-ATF1 rearrangement, suggesting that these are rare molecular events in AFH. Interestingly, the EWSR1-CREB1 rearrangement was detected in the majority of AFH, therefore confirming and expanding data presented in abstract form earlier this year (17, 18) and now very recently published in full (19). We could not identify any correlation between fusion transcript type and either clinical or pathologic variables, such as patients' age and sex, and the tumors' site, size, or depth in our series. Additionally, no significant clinicopathologic differences were found between the cases in our series and the cases previously published (47). However, given the diagnostic difficulty which AFH quite often poses, due to its morphologic variability, our findings do support a new and expanded role for molecular diagnosis in these tumors.

Nonrandom balanced translocations are generally considered tumor-specific and pathogenetically relevant and therefore are used as diagnostic criteria. Two models have been proposed to explain their supposed specificity (21). In the instructive model, the fusion gene resulting from the translocation would instruct the cell to follow a specific program of differentiation through gene expression changes, with consequent development of a specific tumor phenotype. Alternatively, in the context-dependent model, the specificity would be determined by a selection process for which only one or a limited number of cell lineages, with a particular chromatin status in specific regions and/or with certain survival pathways, would allow the occurrence of a specific translocation and/or would tolerate its biological consequences. However, an increasing number of examples of chromosomal promiscuity have been detected in recent years, for example, the ETV6-NTRK3 fusion gene, which is found in infantile fibrosarcoma, mesoblastic nephroma, acute myeloid leukemia, and secretory breast carcinoma or the ASPSCR1-TFE3 fusion gene, which represents a shared genetic aberration of both alveolar soft part sarcoma and pediatric forms of renal cell carcinoma (22, 23). Previous findings in AFH (5, 7, 18, 19), and our findings, highlight an additional example of chromosomal promiscuity, demonstrating that AFH and clear cell sarcoma, which are two morphologically, immunophenotypically, and clinically different tumor entities, share the same fusion genes, specifically EWSR1-CREB1 and EWSR1-ATF1. The occurrence of other unknown downstream genetic events might explain the two distinct phenotypes. Interestingly, twin studies in childhood leukemia have well documented that tumor-specific chromosomal translocations are not always able to induce overt disease and that a second hit might be required. For instance, acute lymphoblastic leukemia is initiated by the ETV6-RUNX1 fusion gene, but only the subsequent deletion of 12p, with consequent loss of the nonrearranged ETV6 gene, leads to final clinically evident leukemia development (24). This intriguing model does not seem to be applicable to clear cell sarcoma. In fact, given that EWSR1-ATF1 knockdown has a growth-suppressive effect on soft agar colonies of clear cell sarcoma cell lines (25), it is likely that the translocation per se is sufficient for cell transformation in clear cell sarcoma. Although theoretically, the occurrence of additional molecular events could still be invoked to explain the alternative phenotype of AFH, this possibility seems remote. It is more reasonable to hypothesize that the two distinct phenotypes of AFH and clear cell sarcoma might be the result of differentiation programs already present in two different tumor progenitor cells. Among the possible candidate genes involved, M-MITF and SOX10 might be considered. In fact, there is evidence that the M-MITF promoter is transactivated by the EWSR1-ATF1 fusion gene in the presence of SOX10 in clear cell sarcoma cell lines, leading to melanocytic differentiation, tumor cell survival, and proliferation (25), and that both M-MITF and SOX10 are highly expressed in clear cell sarcoma (15, 26). Given that M-MITF and SOX10 were not detected in three cases of AFH tested, which were characterized by an EWSR1-ATF1 fusion gene (5, 7), it is reasonable to think that the EWSR1-ATF1 fusion gene in AFH targets other genes rather than M-MITF. A possible explanation might be found in the putative myoid nature of the AFH precursor cell, which would lack expression of SOX10, which is typically restricted to neural crest–derived cells. However, the scenario becomes more complex when clear cell sarcoma in gastrointestinal locations is considered. In fact, despite the presence of either EWSR1-ATF1 or EWSR1-CREB1, gastrointestinal clear cell sarcomas do not express M-MITF and lacks specific melanocytic markers in the majority of the cases (16). Furthermore, SOX10 levels in one case of gastrointestinal clear cell sarcoma were not significantly different from the levels in non–gastrointestinal clear cell sarcoma by gene expression profiling, suggesting that gastrointestinal and non–gastrointestinal clear cell sarcoma have a common SOX10-expressing neuroectodermal precursor, which in a gastrointestinal location, might have lost the potential to differentiate along the melanocytic lineage (16). In vitro studies with transfection of the EWSR1-ATF1/CREB1 fusion gene into different cell lineages could potentially help to clarify the pathogenetic mechanism.

In this study, we assessed the prevalence of FUS/EWSR1-ATF1/CREB1 fusion genes by interphase FISH in a series of 14 cases of AFH. We concluded that the large majority of AFH are characterized by rearrangement of EWSR1 with one member of the CREB/ATF family of transcription factors, the EWSR1-CREB1 fusion gene being the most frequent molecular abnormality. These data point to the potential diagnostic role of molecular testing in a clinical setting. Furthermore, we verified that the EWSR1-CREB1 transcript is identical to the one which characterizes gastrointestinal clear cell sarcoma, strengthening the molecular homology between AFH and clear cell sarcoma.

Grant support: European Commission (EUROBONET grant no. 018814).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. Roberta Maestro and Daniela Gasparotto from CRO of Aviano, Drs. Salvatore Romeo and Anne Marie Cleton-Jansen from LUMC of Leiden for critical discussion, and Marja van der Burg for technical assistance in performing FISH experiments.

1
Fanburg-Smith JC, Dal Cin P. Angiomatoid fibrous histiocytoma. Fletcher CDM, Unni KK, Mertens F. WHO classification of tumours. Pathology and genetics of tumours of soft tissue and bone. 194–5. IARC Press. Lyon, 2002.
2
Fletcher CD. Angiomatoid “malignant fibrous histiocytoma”: an immunohistochemical study indicative of myoid differentiation.
Hum Pathol
1991
;
22
:
563
–68.
3
Fanburg-Smith JC, Miettinen M. Angiomatoid “malignant” fibrous histiocytoma: a clinicopathologic study of 158 cases and further exploration of the myoid phenotype.
Hum Pathol
1999
;
30
:
1336
–43.
4
Waters BL, Panagopoulos I, and Allen EF. Genetic characterization of angiomatoid fibrous histiocytoma identifies fusion of the FUS and ATF-1 genes induced by a chromosomal translocation involving bands 12q13 and 16p11.
Cancer Genet Cytogenet
2000
;
121
:
109
–16.
5
Hallor KH, Mertens F, Jin Y, Meis-Kindblom JM, Kindblom LG, Behrendtz M, et al. Fusion of the EWSR1 and ATF1 genes without expression of the MITF-M transcript in angiomatoid fibrous histiocytoma.
Genes Chromosomes Cancer
2005
;
44
:
97
–102.
6
Raddaoui E, Donner LR, Panagopoulos I. Fusion of the FUS and ATF1 genes in a large, deep-seated angiomatoid fibrous histiocytoma.
Diagn Mol Pathol
2002
;
11
:
157
–62.
7
Hallor KH, Micci F, Meis-Kindblom JM, et al. Fusion genes in angiomatoid fibrous histiocytoma.
Cancer Lett
2007
;
25
:
158
–63.
8
Law WJ, Cann KL, Hicks GG. TLS, EWS, and TAF15: a model for transcriptional integration of gene expression.
Brief Funct Genomic Proteomic
2006
;
5
:
8
–14.
9
Persengiev SP, Green MR. The role of ATF/CREB family members in cell growth, survival and apoptosis.
Apoptosis
2003
;
8
:
225
–8.
10
Shing DC, McMullan DJ, Roberts P, et al. FUS/ERG gene fusions in Ewing's tumors.
Cancer Res
2003
;
63
:
4568
–76.
11
Kaneko Y, Kobayashi H, Handa M, Satake N, Maseki N. EWS-ERG fusion transcript produced by chromosomal insertion in a Ewing sarcoma.
Genes Chromosomes Cancer
1997
;
18
:
228
–31.
12
Crozat A, Aman P, Mandahl N, Ron D. Fusion of CHOP to a novel RNA-binding protein in human myxoid liposarcoma.
Nature
1993
;
363
:
640
–4.
13
Panagopoulos I, Höglund M, Mertens F, Mandahl N, Mitelman F, Aman P. Fusion of the EWS and CHOP genes in myxoid liposarcoma.
Oncogene
1996
;
12
:
489
–94.
14
Panagopoulos I, Mertens F, Debiec-Rychter M, et al. Molecular genetic characterization of the EWS/ATF1 fusion gene in clear cell sarcoma of tendons and aponeuroses.
Int J Cancer
2002
;
99
:
560
–7.
15
Antonescu CR, Tschernyavsky SJ, Woodruff JM, Jungbluth AA, Brennan MF, Ladanyi M. Molecular diagnosis of clear cell sarcoma: detection of EWS-ATF1 and MITF-M transcripts and histopathological and ultrastructural analysis of 12 cases.
J Mol Diagn
2002
;
4
:
44
–52.
16
Antonescu CR, Nafa K, Segal NH, Dal Cin P, Ladanyi M. EWS-CREB1: a recurrent variant fusion in clear cell sarcoma-association with gastrointestinal location and absence of melanocytic differentiation.
Clin Cancer Res
2006
;
12
:
5356
–62.
17
Terra SBSP, Folpe AL, Weiss SW, et al. Frequency and characterization of ATF1—fusion genes in angiomatoid fibrous histiocytoma [abstract 76]. Mod Pathol 2007;20.
18
Antonescu CR, Dal Cin P, Nafa K, Teot LA, Fletcher CDM, Ladanyi M. EWS-CREB1 is the predominant gene fusion in so-called angiomatoid fibrous histiocytoma [abstract 31]. Mod Pathol 2007;20.
19
Antonescu CR, Dal Cin P, Nafa K, et al. EWSR1-1 is the predominant gene fusion in angiomatoid fibrous histiocytoma.
Genes Chromosomes Cancer
2007
;
46
:
1051
–60.
20
Baelde HJ, Cleton-Jansen AM, van Beerendonk H, Namba M, Bovee JV, Hogendoorn PC. High quality RNA isolation from tumours with low cellularity and high extracellular matrix component for cDNA microarrays: application to chondrosarcoma.
J Clin Pathol
2001
;
54
:
778
–82.
21
Helman LJ, Meltzer P. Mechanisms of sarcoma development.
Nat Rev Cancer
2003
;
3
:
685
–94.
22
Stenman G. Fusion oncogenes and tumor type specificity—insights from salivary gland tumors.
Semin Cancer Biol
2005
;
15
:
224
–35.
23
Lannon CL, Sorensen PH. ETV6-3: a chimeric protein tyrosine kinase with transformation activity in multiple cell lineages.
Semin Cancer Biol
2005
;
15
:
215
–23.
24
Greaves MF, Maia AT, Wiemels JL, Ford AM. Leukemia in twins: lessons in natural history.
Blood
2003
;
102
:
2321
–33.
25
Davis IJ, Kim JJ, Ozsolak F, et al. Oncogenic MITF dysregulation in clear cell sarcoma: defining the MiT family of human cancers.
Cancer Cell
2006
;
9
:
473
–84.
26
Li KK, Goodall J, Goding CR, et al. The melanocyte inducing factor MITF is stably expressed in cell lines from human clear cell sarcoma.
Br J Cancer
2003
;
89
:
1072
–8.