Abstract
Background/Aim: Chimeras involving the high-mobility group AT-hook 2 gene (HMGA2 in 12q14.3) have been found in lipomas and other benign mesenchymal tumors. We report here a fusion of HMGA2 with the nuclear receptor co-repressor 2 gene (NCOR2 in 12q24.31) repeatedly found in tumors of bone and the first cytogenetic investigation of this fusion. Materials and Methods: Six osteoclastic giant cell-rich tumors were investigated using G-banding, RNA sequencing, reverse transcription polymerase chain reaction, Sanger sequencing, and fluorescence in situ hybridization. Results: Four tumors had structural chromosomal aberrations of 12q. The pathogenic variant c.103_104GG>AT (p.Gly35Met) in the H3.3 histone A gene was found in a tumor without 12q aberration. In-frame HMGA2–NCOR2 fusion transcripts were found in all tumors. In two cases, the presence of an HMGA2–NCOR2 fusion gene was confirmed by FISH on metaphase spreads. Conclusion: Our results demonstrate that a subset of osteoclastic giant cell-rich tumors of bone are characterized by an HMGA2–NCOR2 fusion gene.
- Benign bone tumor
- non-ossifying fibroma
- giant cell tumor of bone
- osteoclastic giant cell-rich tumors of bone
- cytogenetics
- chromosomal aberrations
- HMGA2
- NCOR2
- HMGA2–NCOR2 fusion gene
According to the latest edition of the World Health Organization classification of soft tissue and bone tumors, published in 2020, aneurysmal bone cyst, non-ossifying fibroma, and giant cell tumor of bone comprise a group of osteoclastic giant cell-rich tumors (1, 2). Their common feature is that they contain reactive osteoclast-type multinucleated giant cells (1, 2).
Aneurysmal bone cysts are pathogenetically characterized by rearrangements of chromosome band 17p13 targeting the ubiquitin-specific protease 6 gene (USP6 or Tre2) resulting in fusions in which USP6 is the 3’ moiety, but where the promoter of USP6 is replaced by a stronger promoter from the partner fusion gene so that USP6 becomes transcriptionally up-regulated (3-8). The chromosome translocation t(16;17)(q22;p13) is the most frequent cytogenetic aberration reported (21%) in aneurysmal bone cysts with an abnormal karyotype (9). It generates a fusion gene in which the strong promoter of the cadherin 11 gene (CDH11) from 16q21 fuses with the USP6 coding sequence from 17p13 (6, 7) leading to a CDH11–USP6 chimera in which the USP6 gene is transcriptionally up-regulated (3, 4).
In non-ossifying fibromas, activating mutations in the mitogen-activated protein kinase pathway were recently found in 81% of examined tumors (10, 11). Using DNA sequencing methodology, mutually exclusive pathogenic mutations were identified in the KRAS proto-oncogene (in 12p12), fibroblast growth factor receptor 1 gene (FGFR1 in 8p11), and neurofibromin 1 gene (NF1 in 17q11) (10, 11). Giant cell tumors of bone are genetically characterized by pathogenic variations in the 35th codon (exon 2) of the H3.3 histone A gene (H3-3A alias H3F3A) (12-20). The pathogenic variation c.103G >T, p.Gly35Trp (NM_002107.4, NP_002098.1, COSM1732355) is found in most giant cell tumors, where it is detected only in cells with a fibroblast-like appearance, not in giant cells or their precursors (12-18). Other mutations found in the same codon of H3-3A are p.Gly35Leu, p.Gly35Val, p.Gly35Met, p.Gly35Glu, and p.Gly35Arg (12-18). Giant cell tumors of bone without mutation in H3-3A have also been reported (13, 14).
In the present study, we describe six osteoclastic giant cell-rich tumors that carried a fusion between the genes coding for high-mobility group AT-hook 2 (HMGA2) and nuclear receptor co-repressor 2 (NCOR2).
Materials and Methods
Ethics statement. The study was approved by the Regional Ethics Committee (Regional komité for medisinsk forskningsetikk Sør-Øst, Norge, http://helseforskning.etikkom.no). All clinical information has been de-identified.
Tumor description. Table I shows the patients’ gender, age, diagnosis, and tumor location. The patients were three males and three females. The age range was from 16 to 40 years with a median of 28.5 years. The tumors in patients 1 and 2 were diagnosed as benign fibrous histiocytoma of bone. In patients 3 to 6, the tumors were diagnosed as giant cell tumor of bone. Immunohistochemistry showed that all tumors were negative for H3.3 G35W (previously written as G34W).
Clinicopathological data on the tumors with high-mobility group AT-hook 2 and nuclear receptor co-repressor 2 genes (HMGA2–NCOR2) fusion.
Patient 1 was a 37-year-old woman with back pain for 2 years and neurological symptoms. A tumor in the TH9 vertebra was found with computed tomography. Core-needle biopsy showed a fibrous lesion in the bone, with macrophages and scattered giant cells. The diagnosis was benign fibrous histiocytoma of bone (Figure 1A). The lesion was H3.3 G35W IHC-negative (Figure 1B). Examination of the curetted specimen confirmed the diagnosis.
Microscopic examination of osteoclastic giant cell-rich tumor of bone from case 1 (A and B) and case 4 (C and D). A: The tumor tissue consisted of mononuclear cells with eosinophilic cytoplasm and small, round to oval nuclei. Throughout the tumor tissue there were scattered areas of foamy macrophages and a few multinuclear giant cells (the latter not visible in the image), hematoxylin and eosin (H&E), ×20 objective. B: Immunohistochemical examination against H3.3 G35W (previously written as G34W) showing cell nuclei within the tumor tissue that were negative for H3.3 G35W. C: The histological picture was dominated by mononuclear cells without atypia. The nuclei were round to oval, and the cytoplasm had indistinct borders. Throughout the lesion there were areas with foamy macrophages. Very few giant cells were seen, H&E, ×20 objective. D: Immunohistochemical examination showed that nuclei in the tumor tissue were negative for H3.3 G35W.
Patient 2 was a 40-year-old woman. The tumor was located in the pubic bone. It was H3.3 G35W IHC-negative and diagnosed as a benign fibrous histiocytoma of bone.
The tumors of patients 3 and 5 were located in the distal tibia and distal femur, respectively. They were diagnosed as giant cell tumors of bone with components of aneurysmal bone cyst.
Patient 4 was a 16-year-old girl. She had had shoulder pain for 9 months. An expansive lesion in the scapula with bone destruction was found. Examination of a core-needle biopsy showed tumor tissue with scattered giant cells against a background of bland-looking mononuclear cells, scattered macrophages, and hemosiderin deposits (Figure 1C). Despite negative IHC analysis for H3.3 G35W (Figure 1D), a diagnosis of giant cell tumor of bone was reached and the patient received denosumab for 4 months followed by curettage. The histological picture of fibrous tumor tissue with mononuclear cells and scattered giant cells (Figure 1C) was consistent with the diagnosis of a giant cell tumor.
In patient 6, a tumor was found in the L5 vertebra. Although it was H3.3 G35W IHC-negative, a diagnosis of giant cell tumor of bone was reached.
G-Banding and karyotyping. The methods used to investigate the tumors cytogenetically were described elsewhere (21). Briefly, a part of resected tumor was minced with scalpels into 1-2 mm fragments and then enzymatically disaggregated with collagenase II (Worthington, Freehold, NJ, USA). The resulting cells were cultured, harvested, and processed for cytogenetic examination using standard techniques (22). A G-banding pattern of chromosomes was obtained using Wright’s stain (Sigma-Aldrich, St Louis, MO, USA) (22). Metaphases were analyzed and karyograms prepared using the CytoVision computer-assisted karyotyping system (Leica Biosystems, Newcastle upon Tyne, UK). The karyotypes were described according to the International System for Human Cytogenomic Nomenclature (23).
Sanger sequencing for detection of pathogenic variation in exon 2 of the H3-3A gene. Genomic DNA from tumor samples was extracted using a Maxwell RSC Instrument and a Maxwell RSC Tissue DNA Kit (Promega, Madison, WI, USA) and subsequently quantified with Quantus Fluorometer and the QuantiFluor ONE dsDNA System (Promega). Extracted DNA (100-150 ng) was used as polymerase chain reaction (PCR) template for detection of possible pathogenic variation in exon 2 of the H3-3A gene (NCBI reference sequence: NM_002107.4). The primers M13For-H3F3A-Ex2-F2 and M13Rev-H3F3A-Ex2-R2 (Table II), together with the BigDye Direct Cycle Sequencing Kit, were used to perform both PCR and cycle (Sanger) sequencing according to the company’s recommendations (ThermoFisher Scientific, Waltham, MA, USA). The sequences obtained by Sanger sequencing were compared with the NCBI reference sequence NM_002107.4.
Primers used for polymerase chain reaction amplification and Sanger sequencing analyses. M13 forward primer (TGTAAAACGACGGCCAGT) and M13 reverse primer (CAGGAAACAGCTATGACC) sequences are shown in italics.
RNA sequencing. Total RNA was extracted from frozen tumor tissue adjacent to that used for cytogenetic analysis and histological examination using miRNeasy Mini Kit (Qiagen, Hilden, Germany). RNA sequencing was performed on cases 1 and 2 (Table I). For this purpose, 1 μg of total RNA was sent to the Genomics Core Facility at the Norwegian Radium Hospital, Oslo University Hospital for high-throughput paired-end RNA sequencing. A total of 89×106 and 107×106 76-bp-length reads were obtained from the tumors of cases 1 and 2, respectively. FASTQC software was used for quality control of the raw sequence data (available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Fusion transcripts were found using FusionCatcher software (24, 25). The “grep” command was also used to search the fastq files of the sequence data for the HMGA2–NCOR2 fusion sequence. The principle of this approach has been described in detail elsewhere (26-28). The search terms were “AGGAAATGGCAGCAGCAA” and “AGGAAATGGC AGCAACAA” for case 1 and “CCTGCTCAGCAGCAGCAA” and “CCTGCTCAGCAGCAACAA” for case 2.
Reverse transcription (RT)PCR and Sanger sequencing analyses. The primers used for PCR amplification and Sanger sequencing analyses are shown in Table II. The methodologies for cDNA synthesis, RT-PCR amplification, and Sanger sequencing were described elsewhere (29, 30). For the first outer PCR amplification, the primer combination HMGA2-846F1/NCOR2-2127R1 was used. For the second inner PCR, the primer combination HMGA2-925F1/NCOR2-2076R1 was used. The Basic Local Alignment Search Tool (BLAST) was used to compare the sequences obtained by Sanger sequencing with the NCBI reference sequences NM_003483.4 (HMGA2) and NM_006312.6 (NCOR2) (31).
Fluorescence in situ hybridization (FISH). For the detection of HMGA2–NCOR2 fusion gene, a homemade double-fusion FISH probe was used. The BAC probes were purchased from the BACPAC Resource Center operated by BACPAC Genomics, Emeryville, CA, USA (https://bacpacresources.org/) (Table III). The probes for HMGA2 were labelled with fluorescein-12-dCTP (PerkinElmer, Boston, MA, USA) to obtain a green signal. The probes for NCOR2 were labelled with Texas Red-5-dCTP (PerkinElmer) to obtain a red signal. Detailed information on the FISH procedure was given elsewhere (30, 32). Fluorescent signals were captured and analyzed using the CytoVision system (Leica Biosystems).
BAC probes used for fluorescence in situ hybridization experiments in order to detect the high-mobility group AT-hook 2 and nuclear receptor co-repressor 2 (HMGA2–NCOR2) fusion gene. The positions of the HMGA2 and NCOR2 genes are also given.
Results
Karyotyping. Cytogenetic examination revealed chromosomal aberrations in five tumors whereas a normal male karyotype was obtained in both samples from case 6 (Table I). The tumors had pseudo- or near-diploid karyotypes. The dominating feature was an aberration of the q arm of chromosome 12 found in four cases (Table I, Figure 2). In case 1, G-banding analysis detected two cytogenetically unrelated clones: One had the chromosome abnormalities add(12)(q24) and del(20)(q11) whereas the second, composite clone had one extra copy of chromosomes 7 and 18. In case 2, a single cytogenetic clone was found with structural aberrations der(6)(6pter->6q21::12q13->12q?::?::12q24->12q14:) and add(12)(q13). In the third case, a balanced translocation t(12;12)(q13~15;q24) was found as the only abnormality. In the fourth tumor, three related clones were found connected by the presence of a der(2)(2pter->2q37::12q24->12q14:) and a del(12)(q14~15). Finally, the tumor of the fifth case had monosomy 4 and a ring chromosome.
G-Banding analyses of osteoclastic giant cell-rich tumors of bone. Partial karyotypes of tumors (cases 1-4) with 12q anomalies are presented. From case 1, the abnormal chromosome add(12)(q24) together with the normal chromosome 12 are shown. From case 2, the two abnormal chromosomes, der(6)(6pter->6q21::12q13->12q?::?::12q24->12q14:) and add(12)(q13), are shown together with normal chromosomes 6 and 12. From case 3, the two derivative chromosome 12 from the translocation t(12;12)(q13~15;q24) are shown. From case 4, the two abnormal chromosomes, der(2)(2pter->2q37::12q24->12q14:) and del(12)(q14~15), are shown together with the normal chromosomes 2 and 12. Arrows indicate breakpoints.
Sanger sequencing of exon 2 of H3-3A. None of the tumors had the pathogenic variation c.103G >T, p.Gly35Trp (NM_002107.4, NP_002098.1, COSM1732355). In the tumor sample from patient 5, the pathogenic variant c.103_104GG>AT (p.Gly35Met) was found (Figure 3).
Pathogenic variation in exon 2 of the H3-3A gene. Partial sequence chromatograms of DNA amplified fragments from case 5 showing the presence of the pathogenic variant c.103_104GG>AT (p.Gly35Met) in the tumor.
RNA sequencing. Analysis of the sequencing data with FusionCatcher for case 1 detected an in-frame fusion of exon 3 of HMGA2 from 12q14 (NCBI reference sequences NM_003483.4) with exon 16 of NCOR2 from 12q24 (reference sequence NM_006312.6): AGCCACTGGAGAAAAACGG CCAAGAGGCAGACCTAGGAAATGG*CAGCAGCAACA ACAGCAGCAGCAGCAGCAGCAGCAGCAGCAGC. Examining the fastq files of the sequence data using the “grep” command and the search term “AGGAAATGGCAGCA GCAA”, nine sequence reads with the above-mentioned fusion point were found (Table IV), whereas use of the “grep” search with “AGGAAATGGCAGCAACAA” yielded 15 sequence reads with this fusion point. The obtained results agreed with the Sanger sequencing chromatograms, where an AGGAAATGG-CAGCARCAACAR sequence was found at the junction (see below). In case 2, FusionCatcher detected an in-frame fusion of exon 4 of HMGA2 with exon 16 of NCOR2: TAGGAAATGGCCACAACAAGTTGTTCAGAAGAAGCCT GCTCAG*CAGCAGCAACAACAGCAGCAGCAGCAGCA GCAGCAGCAGCAGC. Using the “grep” command and the search term “CCTGCTCAGCAGCAGCAA”, 17 sequence reads with the above-mentioned fusion point were found (Table IV), whereas the “grep” search with term “CCTGCTCAGCAG CAACAA” yielded 16 reads with this fusion point. The results agreed with the Sanger sequencing chromatograms which showed an CCTGCTCAG-CAGCARCAACAR sequence at the junction (see below).
Confirmation of fusion junction of the high-mobility group AT-hook 2 and nuclear receptor co-repressor 2 (HMGA2–NCOR2) chimeric transcript using the “grep” command line utility with the search term “AGGAAATGGCAGCAGCAA” and “AGGAAATGGCAGCAACAA” for case 1 and “CCTGCTCAGCAGCAGCAA” and “CCTGCTCAGCAGCACAA” for case 2. The data show that there was an in-frame insertion/deletion of a CAG triplet in exon 16 of NCOR2 at the junction of fusion transcripts. The sequence of NCOR2 is shown in bold italics.
Thus, for cases 1 and 2, FusionCatcher, examination of fastq files of sequence data with the “grep” command, and RT-PCR/Sanger sequencing showed that there was an in-frame insertion/deletion of a CAG triplet in exon 16 of NCOR2 at the junction of fusion transcripts.
RT-PCR and Sanger sequencing analyses. In cases 1 and 2, RT-PCR with the outer HMGA2-846F1/NCOR2-2127R1 primer combination amplified a 412-bp and a 445-bp cDNA fragment, respectively (Figure 4A). Nested PCR with the inner primer combination HMGA2-925F1/NCOR2-2076R1 amplified 284-bp (case 1) and 317-bp (case 2) cDNA fragments (Figure 4B). Sanger sequencing of the amplified fragments showed that they were HMGA2–NCOR2 chimeric cDNA fragments in which exon 3 of HMGA2 was fused to exon 16 of NCOR2 (case 1) or exon 4 of HMGA2 fused to exon 16 of NCOR2 (case 2) (Figure 4C), thus confirming the results obtained by RNA sequencing.
Reverse transcription polymerase chain reaction (PCR) and Sanger sequencing of the osteoclastic giant cell-rich tumors of bone. A: Gel electrophoresis showing amplified high-mobility group AT-hook 2 and nuclear receptor co-repressor 2 fusion (HMGA2–NCOR2) cDNA fragments using the outer primer combination HMGA2-846F1/NCOR2-2127R1 for cases 1 (lane 1) and 2 (lane 2). B; Gel electrophoresis showing amplified HMGA2–NCOR2 cDNA fragments with nested PCR using the inner primer combination HMGA2-925F1/NCOR2-2076R1 for cases 1 (lane 3) and 2 (lane 4). M: GeneRuler 1 kb Plus DNA ladder (ThermoFisher Scientific). C: Partial sequence chromatograms of cDNA amplified fragments from cases 1 and 6, case 2, and cases 3-5 showing the junction between HMGA2 and NCOR2. In cases 1 and 6, exon 3 of HMGA2 was fused to exon 16 of NCOR2. In cases 2 to 5, exon 4 of HMGA2 was fused to exon 16 of NCOR2. In cases 1, 2, and 6, there was an in-frame insertion/deletion of a CAG triplet in exon 16 of NCOR2 at the junction of the fusion transcripts. For this reason, the NCOR2 sequence at the junction is CAGCARCAACARCAGCAGCA. The letter R in exon 16 of NCOR2 indicates nucleotide G or A.
In cases 3, 4, and 5, nested PCR amplified a 317-bp cDNA fragment (data not shown). Sanger sequencing of these fragments detected fusion between HMGA2 exon 4 and NCOR2 exon 16 (Table I, Figure 4C). In case 6, nested PCR amplified a 284-bp cDNA fragment in which exon 3 of HMGA2 was fused to exon 16 of NCOR2. At the junction of fusion transcript an in-frame insertion/deletion of a CAG triplet in exon 16 of NCOR2 was found, similar to cases 1 and 2 (Table I, Figure 4C).
FISH analyses. Metaphase spreads from cases 2, 4, and 5 (sample 5c) were available for FISH analysis with a homemade double HMGA2–NCOR2 fusion probe consisting of green-labeled HMGA2 (12q14) and red-labeled NCOR2 probes (12q24) (Table III, Figure 5A). For case 2, a yellow signal for HMGA2–NCOR2 fusion as well as a green signal for HMGA2 were found on the der(6)(6pter->6q21::12q13->12q?::?::12q24->12q14:) (Figure 5B). Green signals for HMGA2 were seen also on chromosomes 12 and add(12)(q13), suggesting the presence of at least one intact (normal) HMGA2 locus in the abnormal metaphases. A red signal for NCOR2 was found on chromosome 12 (Figure 5B).
Fluorescence in situ hybridization (FISH) analysis of osteoclastic giant cell-rich tumors of bone using double-fusion FISH probes. A: The position and the orientation of transcription (5’->3’) of the high-mobility group AT-hook 2 (HMGA2; green label) and nuclear receptor co-repressor 2 (NCOR2; red label) genes are shown on the ideogram of chromosome 12. B: FISH results on metaphase spreads from cases 2, 4, and 5. The probes for HMGA2 were labelled with fluorescein-12-dCTP to obtain green signals. The probes for NCOR2 were labelled with Texas Red-5-dCTP to obtain a red signal. Case 2: A yellow signal for HMGA2–NCOR2 fusion together with a green signal for HMGA2 were detected on the der(6)(6pter->6q21::12q13->12q?::?::12q24->12q14:). Green signals for HMGA2 were found on chromosomes 12 and add(12)(q13). A red signal for NCOR2 was found on chromosome 12. Case 4: Two fusion signals were found on der(2)(2pter->2q37::12q24->12q14:) together with a green signal for HMGA2 and a red signal for NCOR2 on chromosome 12. Case 5: A metaphase spread with ring chromosome showing green (HMGA2) and red (NCOR2) signals on the two chromosomes 12 indicating that the HMGA2–NCOR2 fusion must have been formed in another, undetected cytogenetic clone.
In case 4, two fusion signals were found on der(2)(2pter->2q37::12q24->12q14~15) together with a green signal for HMGA2 and a red signal for NCOR2 on chromosome 12 (Figure 5B). In case 5, FISH experiments on metaphase spreads with ‒4 and +r showed normal patterns, in other words, green (HMGA2) and red (NCOR2) signals on the two chromosomes 12 (Figure 5B).
Discussion
We describe an HMGA2–NCOR2 fusion transcript found in six osteoclastic giant cell-rich bone tumors in which immunohistochemistry against H3.3 G35W showed that all of them were negative. In one tumor (patient 5), the extremely rare pathogenic variant c.103_104GG>AT (p.Gly35Met) of the H3-3A gene was found. This tumor had two cytogenetically unrelated clones: one with monosomy 4 and a ring chromosome with normal patterns for the HMGA2 and NCOR2 genes, the other with an HMGA2–NCOR2 fusion. Thus, we do not know if the p.Gly35Met of the H3-3A gene and HMGA2– NCOR2 fusion occurred in the same clone or not.
In the tumor of cases 1 and 6, the fusion occurred between exons 3 of HMGA2 and 16 of NCOR2, whereas in the other tumors, exon 4 of HMGA2 was found to be fused to exon 16 of NCOR2. Cytogenetically, four tumors had visible rearrangements of 12q, with one of them (case 3) showing a translocation between the two copies of chromosome 12, t(12;12)(q13~15;q24). Since HMGA2 (in 12q14) is transcribed from centromere to telomere whereas NCOR2 (in 12q24) is transcribed from telomere to centromere (Figure 4A), neither a simple balanced 12q14;12q24-translocation nor an interstitial deletion del(12)(q14q24) is sufficient to generate the HMGA2–NCOR2 fusion gene. We therefore believe that the seemingly solitary t(12;12)(q13~15;q24) of case 3 must be accompanied by additional submicroscopic rearrangement(s), probably an inversion, to enable the generation of a functionally competent HMGA2–NCOR2 fusion on one of the derivative chromosomes 12. A t(12;12)(q14;q24) was reported in a uterine leiomyoma with a complex karyotype (33).
A paracentric inversion, inv(12)(q14q24), with break in intron 3 or 4 of HMGA2 and a break upstream of exon 16 of NCOR2, would move part of the 3’-end partner gene NCOR2 from 12q24 to 12q14 into the HMGA2 locus where the fusion then takes place. The FISH results from cases 2 and 4 indicate that such an inversion is accompanied by other chromosomal aberrations, including translocations. In case 2, which had a der(6)(6pter->6q21::12q13->12q?::?::12q24->12q14:), the fusion signal (yellow) was proximal whereas a green HMGA2 signal was distal on the der(6). Finally in case 4, a paracentric inversion between 12q14 and 12q24 chromosomal bands, with breakpoints in the HMGA2 and NCOR2 loci and translocation of this part of chromosome 12 onto a chromosome 2, may explain the finding of a der(2)(2pter->2q37::12q24->12q14:) with two FISH fusion signals on it.
Paracentric inversions inv(12)(q14q24) affecting the HMGA2 gene have been reported in chondroid hamartomas, endometrial polyps, and lipomas (34-39). Fusion of HMGA2 with aldehyde dehydrogenase 2 family member gene (ALDH2) was also reported, in a uterine leiomyoma, as the result of a cryptic inv(12)(q14q24) (34).
Alternatively, more complex chromosomal aberrations involving double breaks upstream and in intron 3 or 4 of HMGA2 together with breaks in the 5’-end region of NCOR2 would generate an HMGA2–NCOR2 fusion gene at 12q24. There is no cytogenetic evidence at hand, however, pointing to the actual existence of these potential cytogenetic ways of obtaining HMGA2 rearrangements in connective tissue tumors.
This notwithstanding, many chromosomal rearrangements, such as balanced and unbalanced translocations, inversions, insertions, and deletions involving band 12q14 and targeting the HMGA2 gene have repeatedly been reported in lipomas and other benign neoplasms of connective tissues (40). In most cases, HMGA2 fuses out-of-frame with the 3’-end partner gene or with intergenic sequences (32, 34, 36, 41-53). In these fusions, the part of the HMGA2 gene coding for the AT-hook domains, i.e., exons 1 to 3, is separated from the 3’-untranslated region which regulates HMGA2 transcription, resulting in expression and translation of a tumorigenic, truncated form of HMGA2 (54-60). Thus, the addition of ectopic fusion sequences does not seem to have any significant impact compared with the tumorigenic properties of truncated HMGA2 (61).
In some cases, the chromosomal rearrangements result in an HMGA2 fusion with in-frame transcripts coding for chimeric HMGA2 proteins. The chromosomal translocation t(3;12)(q28;q14) repeatedly found in lipomas (especially) but also chondroid hamartomas and chondromas, fuses in-frame HMGA2 with a gene whose official full name is LIM domain-containing preferred translocation partner in lipoma (LPP) (47, 62-68). The resulting chimera codes for an HMGA2– LPP protein containing the AT-hook domains of HMGA2 and the LIM domains of the LPP protein (47, 62-70).
Fusion of HMGA2 with epidermal growth factor receptor gene (EGFR) from 7p11 gives rise to a fusion gene which encodes a chimeric HMGA2–EGFR protein; this was reported in a single case of glioblastoma (71). An in-frame fusion transcript of HMGA2 with Yes1-associated transcriptional regulator gene (YAP1) from 11q22 was reported in an aggressive angiomyxoma (72).
Each of the above-mentioned partners is involved in different cell functions and all have been implicated in tumor development. LPP belongs to the zyxin family of LIM domain proteins, has focal adhesion capacity, is a transcription activator, interacts with various proteins through its numerous protein-protein interactions domains, and is known to play a role in many biological and cellular processes (69, 73-75). In a secondary acute myeloid leukemia with a t(3;11)(q28;q23) chromosomal translocation, LPP fused to lysine methyltransferase 2A (KMT2A, formerly MLL) generating a KMT2A–LPP fusion gene which codes for a chimeric protein containing the AT hook of KMT2A and the LIM domains of LPP (76). LPP was found to be involved in tumor cell migration, invasion, and metastasis (75). EGFR is a transmembrane glycoprotein, a member of the protein kinase superfamily, and a receptor for members of the EGF family (77). Mutation and amplification/overexpression of EGFR and its protein have been reported in lung cancer and brain tumors (77). YAP1 is a transcriptional co-activator which affects the Hippo signaling pathway (78-80). Mutations, amplifications, and fusion genes of YAP1 have been reported in various types of cancer (78-80).
Tumorigenic properties were found for the HMGA2–LPP and HMGA2–EGFR fusion proteins, whereas functional studies have not been performed to test for effects of the HMGA2–YAP1 chimera (54, 70-72). HMGA2–LPP transforms 3T3 mouse embryonic fibroblasts (54). HMGA2– LPP was also found to act as a transcription activator stimulated by co-expression of HMGA2 (70). Furthermore, it was shown that through the AT-hook DNA-binding domains of HMGA2, HMGA2–LPP transactivates the promoter of collagen type XI alpha 2 chain (COL11A2) gene and, consequently, promotes chondrogenesis (81).
The HMGA2–EGFR chimeric protein, which contains the AT-hook domains of HMGA2 and transmembrane and kinase domains of EGFR, showed transforming potential, significantly enhancing colony formation in soft agar (71). Implantation of cells expressing HMGA2–EGFR into mice resulted in accelerated tumor growth (71).
The HMGA2–NCOR2 chimeric gene we describe here is the fourth fusion coding for a chimeric protein. Based on the HMGA2 reference sequences NM_003483.4/NP_003474.1 and the NCOR2 reference sequences NM_006312.6/NP_006303.4, the HMGA2–NCOR2 fusion transcripts would code for chimeric proteins composed of the first 83 amino acid residues of HMGA2 (or the first 93 amino acids encoded by the HMGA2 exon 4-NCOR2 exon 16 fusion) and the last 2021 amino acids of NCOR2 protein (amino acids 494-2514 in NP_006303.4). The chimeric HMGA2–NCOR2 protein thus contains the three regions of HMGA2 (AT-hooks) which bind to the minor groove of adenine-thymine (AT)-rich DNA: AT-hook 1 between amino acid residues 24-34, AT-hook 2 between amino acid residues 44-54, and AT-hook 3 between amino acid residues 71-82 (82-84). It would also contain all functional and interaction domains of NCOR2 except the first N-terminal repressor domain (RD1) and the deacetylase activation domain (absence of NCOR2 amino acid residues 1-493). Thus, it contains the histone interaction domain, the RD2 and RD3 repressor domains, the two nuclear receptor interaction domains, and RID1 and RID2 (85-87).
While we were working on the present project, a publication appeared reporting the finding of an HMGA2– NCOR2 fusion gene in six giant cell-rich soft tissue tumors which expressed low- to high-molecular-weight keratins (the tumor cells stained positively with the immunohistochemical marker cytokeratin AE1/AE3, which is a mixture of two different clones of monoclonal antibodies to cytokeratin AE1 and AE3) (88). In four of the tumors, exon 3 of HMGA2 fused to exon 16 of NCOR2, i.e., the fusion transcript was identical to that found by us in case 1. In the other two tumors, exon 3 of HMGA2 fused in-frame to exon 20 of NCOR2 (88). Thus, both studies, the present and that published by Agaimy and coworkers (88), showed HMGA2– NCOR2 fusion genes coding for similar chimeric proteins and that a group of giant cell-rich tumors of bone or soft tissue are characterized by the HMGA2–NCOR2 fusion gene. Our study, furthermore, presents the first cytogenetic evidence, also obtained by FISH analyses, of this fusion.
An in-frame HMGA2–NCOR2 fusion transcript was also reported in a tenosynovial giant-cell tumor without any information on the fusion junction (89). However, the data indicate that the HMGA2–NCOR2 pathogenetic pathway might also be found in other types of neoplasia.
The tumors with HMGA2–NCOR2 of the present study belonged to the group of osteoclastic giant cell-rich tumors; they were initially diagnosed as giant cell tumors of bone negative for H3.3 G35W (1, 2) and difficult to distinguish from non-ossifying fibromas. We believe that the HMGA2– NCOR2 fusion transcript can be used as a differential diagnostic marker useful in subdividing various osteoclastic giant cell-rich tumors.
Acknowledgements
This work was supported by grants from Radiumhospitalets Legater.
Footnotes
Authors’ Contributions
IP designed and supervised the research, performed molecular genetic experiments and bioinformatics analysis, and wrote the article. LG performed cytogenetic analysis. KA performed cytogenetic analysis, molecular genetic experiments, and evaluated the data. ML-I performed pathological examination. IL performed pathological examination. SH evaluated the cytogenetic data and assisted in writing the article. All authors read and approved of the final article.
Conflicts of Interest
The Authors declare that they have no potential conflicts of interest in regard to this study.
- Received November 10, 2021.
- Revision received December 9, 2021.
- Accepted December 10, 2021.
- Copyright © 2022, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.