Abstract
Atypical lipomatous tumors (ALTs)/well-differentiated liposarcomas represent a distinctive subset of mesenchymal neoplasms featuring mature adipocytic proliferation. These tumors are characterized cytogenetically by the presence of supernumerary ring and/or long marker chromosomes that contain several copies of the chromosomal region 12q13–15, in which the HMGA2 gene is located. Deregulation of the HMGA2 gene is a common molecular alteration implicated in the development of a variety of benign tumors, such as lipomas, uterine leiomyomas, and pulmonary chondroid hamartomas. In this study, we observed HMGA2 overexpression in 7 of 12 ALT primary cell cultures examined. Subsequently, we generated an adenovirus containing the HMGA2 gene in the antisense orientation (Ad-A2as) to study the effect of HMGA2 protein suppression in ALT cells. The infection of six ALT cells, three of which were positive for HMGA2 expression, resulted in growth inhibition coupled with a significant increase in apoptosis. In addition, the growth of the ALT cells negative for HMGA2 expression was not affected by the infection with either the Ad-A2as or the control virus. On the basis of these findings, the targeting of the HMGA2 protein expression may represent a promising approach for treating the well-differentiated liposarcomas resistant to conventional therapies.
INTRODUCTION
ALTs 3 represent a subset of mesenchymal neoplasms featuring mature adipocytic differentiation. Clinically, ALTs tend to arise in the somatic soft tissue of the limbs and in the retroperitoneum. ALTs and WD liposarcoma are synonymous in describing lesions that are identical both in morphology and karyotype. The use of the term “atypical lipomatous tumors” has been introduced by the WHO classification of tumors to emphasize that WD liposarcomas show a high risk of local recurrence that is dependent on tumor location and resectability.
The characteristic cytogenetic feature of ALT/WD liposarcomas is the presence of supernumerary ring and/or giant marker chromosomes. These ring chromosomes contain a consistent amplification of the 12q13-q15 region associated with coamplification of various different chromosome regions and present centromeric alterations, as shown by the lack of specific α satellite sequences (1 , 2) . The 12q13-q15 locus contains several genes implicated in tumorigenesis, such as GLI, GADD153, SAS, CDK4, MDM2, and HMGA2. These genes display different coamplification patterns in ALT/WD liposarcomas (3 , 4) .
HMGA2 (formerly HMGIC) belongs to the HMGA family of proteins, a group of nonhistone nuclear proteins known as architectural transcriptional factors. They function in vivo both as structural components of chromatin as well as auxiliary gene transcription factors (5, 6, 7) . The HMGA family is composed of three known proteins: HMGA1a, HMGA1b, and HMGA2. Whereas the first two are generated from a single functional gene and differ by 11 amino acids (8 , 9) , the HMGA2 is the product of a separate gene (10) .
Alterations in HMGA family structure and/or expression play an important role in a number of benign and malignant tumors (11, 12, 13, 14) . HMGA overexpression was first described in transformed rat thyroid cells and in experimental thyroid tumors (15, 16, 17) , whereas additional studies assessed its role in human malignant neoplasias (11) . Normally, HMGA expression is restricted to embryonic development and is almost undetectable in normal adult tissues (18, 19, 20) . Previous results have shown that HMGA overexpression is a requirement for the cellular transformation of normal rat thyroid cells. In fact, transfection of these cells with a vector carrying either the HMGA2 or the HMGA1 gene in an antisense orientation resulted in the suppression of the HMGA proteins and prevented neoplastic transformation induced by the myeloproliferative sarcoma virus or by the Kirsten murine sarcoma virus (21 , 22) . Moreover, the suppression of the HMGA1 expression by an adenovirus carrying the antisense of the HMGA1 gene (Ad-Yas) induced apoptosis in two human thyroid anaplastic carcinoma cell lines (ARO and FB-1), but not in normal thyroid cells (23) .
Interestingly, rearrangements of the 12q15 chromosomal region, in which the HMGA2 gene is located, are a well established feature of a wide spectrum of human benign tumors, mainly of mesenchymal origin (24, 25, 26) . In the majority of the lipomas, breaks occur within the third intron of the HMGA2 gene, resulting in chimeric transcripts containing exons 1–3 (encoding the AT-hook domains) and ectopic sequences from other genes (27 , 28) . In uterine leiomyomas, instead, most of the 12q15 breakpoints are located 5′ to HMGA2 locus, suggesting that the formation of fusion transcripts involving HMGA2 is uncommon and not required to initiate the tumor growth (29) . A similar mechanism has been suggested in i.v. leiomyomatosis and three pulmonary chondroid hamartomas, all exhibiting two normal chromosomes 12 and a der (14) t(12;14) (q14–15;q24) (30) . A role of HMGA2 gene overexpression in the generation of lipomas has been proposed because transgenic mice overexpressing a truncated or a wild-type HMGA2 gene show a giant phenotype together with a predominantly abdominal/pelvic lipomatosis and high incidence of lipomas (31, 32, 33) . Interestingly, this phenotype is the opposite of that of HMGA2 null mice that conversely show a pigmy phenotype, with a drastic reduction of the fat tissue (18) .
Previous work has shown HMGA2 immuonopositivity in 10 of 12 ALTs and HMGA2 amplification in 3 of 5 ALTs by Southern blot analysis (3) , suggesting that HMGA2 expression may have a role in the generation of ALTs. These results prompted us to target the HMGA2 gene and investigate whether its suppression may affect the growth of the ALT-derived cells.
In our study, we analyzed the expression of HMGA2 in several ALT cells and observed HMGA2 overexpression in 7 of 12 cases. Subsequently, we generated an adenovirus carrying HMGA2 cDNA in the antisense orientation (Ad-A2as) and infected six ALT cells. After Ad-A2as treatment, all of the ALT cells overexpressing HMGA2 showed growth inhibition together with a significant apoptosis, whereas the ALT cells negative for HMGA2 expression were unaffected.
MATERIALS AND METHODS
Cell Cultures.
Primary atypical lipomatous cells were obtained from the Centre of Human Genetics (Leuven, Belgium). All ALTs were included in the Chromosomes and Morphology (CHAMP) collaborative study group (34) . After disaggregation with collagenase, all tumor samples were characterized cytogenetically after short-term culture and then frozen as primary culture (5–7 days). Each ALT exhibited one to three ring chromosomes, the well-established cytogenetic marker of ALTs. Cells were maintained at 37°C in a humidified atmosphere of 5% CO2 in DMEM containing 10% FCS. RNA extraction and viral infection were performed after a few passages in culture.
Generation of the Recombinant Adenoviruses.
Vector construction was performed following the manufacturer’s instructions (Quantum, Montreal, Canada). Briefly, the HMGA2 cDNA was inserted in sense and antisense orientation into pac-CMVpLpa to generate the pac-CMV-HMGA2s and pac-CMV-HMGA2as constructs, respectively. They were cotransfected with pJM17 into the human embryonic kidney 293 cell line (American Type Culture Collection) to generate the Ad5CMV-HMGA2s GFP(Ad-A2s) and the Ad5CMV-HMGA2as GFP (Ad-A2as) viruses. Viral stocks were expanded in 293 cells, which were harvested 36–40 h after infection and lysed. Virus titer was determined by pfu in the 293 cells. Viral stocks were aliquoted and stored at −80°C. The AdCMV-GFP (Ad-GFP) vector (Quantum Biotechnology) was also used as a control.
In Vitro Transduction.
ALT cells were transduced with adenoviral vectors by directly applying the diluted vectors into the growth medium. Cells were transduced with the various viral constructs at different multiplicities of infection. The transduction efficiency was determined by measuring the proportion of GFP-expressing cells after Ad-GFP transduction by fluorescence-activated cell sorting. Direct visualization of the Ad-GFP-transduced cell population by fluorescent cell microscopy was used to confirm the fluorescence-activated cell-sorting data.
RNA Isolation and RT-PCR Analysis.
Total RNA was extracted with the RNAfast Isolation System (Molecular System, San Diego, CA). A previously described strand-specific RT-PCR procedure (21) was used to detect the antisense HMGA2-specific sequences. The following primers were used to amplify the HMGA2 transcript: forward primer exon 1, 5′-CGAGTCCTCTTCGGCAGACTC-3′; forward primer exon 4, 5′-TGTTCAGAAGAAGCCTGCTC-3′; reverse primer exon 5, 5′-CTAGTCCTCTTCGGCAGACTC-3′. For the expression of the HMGA2 antisense construct, the forward primer 5′-AATTCCAACACACTATTGC-3′, overlapping the construct Ad-A2as upstream of the cloned antisense cDNA, and the reverse primer 5′-GGTACCGGTAGAGGCAGTGG-3′, overlapping the 5′ of the HMGA2 gene, were used. Amplification of the glyceraldehyde-3-phosphate dehydrogenase gene with the primers 5′-ACATGTTCCAATATGATTCC-3′ (forward) and 5′-TGGACTCCACGACGTACTCA-3′ (reverse) served as an internal control for the amount of cDNA. To verify that RNA samples were not contaminated by DNA, we obtained negative controls by running PCRs on samples that were not reverse transcribed but otherwise processed identically to the samples used for the experiment.
Immunoblotting Analysis.
Cells were washed once in cold PBS and lysed in a buffer containing 50 mm HEPES (pH 7.5), 1% (v/v) Triton X-100, 50 mm NaCl, 5 mm EGTA, 50 mm NaF, 20 mm sodium PPI, 1 mm sodium vanadate, 2 mm phenylmethylsulfonyl fluoride, and 0.2 mg each of aprotinin and leupeptin per milliliter. Lysates were clarified by centrifugation at 10,000 × g for 15 min, and the supernatant was stored at −70°C. Protein concentration was estimated by a modified Bradford assay (Bio-Rad). Total proteins were separated by 15% SDS-PAGE and transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore). Membranes were blocked with 5% nonfat milk proteins and incubated with polyclonal antibodies raised against the HMGA2 recombinant protein (35) at a dilution of 1:1,000. Bound antibodies were detected by the appropriate secondary antibodies and revealed with the Amersham enhanced chemiluminesce system.
TUNEL Assay.
For the TUNEL assay, we used the In Situ Cell Death Detection kit (Boehringer Mannheim) following the manufacturer’s instructions. Briefly, the air-dried cells were fixed with a freshly prepared paraformaldehyde solution (4% in PBS, pH 7.4) for 30 min at room temperature. The slides were then rinsed with PBS and incubated in permeabilization solution (0.1% Triton X-100 and 0.1% sodium citrate) for 2 min on ice. The slides were then rinsed twice with PBS, and 50 μl of TUNEL reaction mixture were added. After a 30-min incubation at 37°C, substrate solution was added. After 10 min at room temperature, the slides were mounted under glass coverslips and examined under a light microscope.
RESULTS
Analysis of HMGA2 Expression in ALT Cell Cultures.
Twelve primary culture cells derived from atypical lipomatous tumors were analyzed for HMGA2 expression by RT-PCR using primers specific for exons 1 and 5 of the HMGA2 gene. Seven ALT cells (ALT 1, ALT 4, ALT 10, ALT 11, ALT 12, ALT 14, and ALT 16) showed HMGA2 gene amplification, whereas in five ALT cells (ALT 2, ALT 3, ALT 8, ALT 13, and ALT 15) no amplification was observed (Fig. 1A ⇓ and Table 1 ⇓ ). Western blot analysis performed on 10 ALT cells, using antibodies raised against the recombinant HMGA2 protein, resulted in the detection of a Mr 15,000 band, corresponding to the HMGA2 protein (Fig. 1C) ⇓ , consistent with the RT-PCR data.
HMGA2 expression in ALT cells. A, RT-PCR analysis of HMGA2 gene expression using primers specific for the exons 1 and 5 of the HMGA2 gene. cDNAs from ALT cells were amplified with primers specific for exons 1 and 5 of the HMGA2 gene and coamplified with β-actin as an internal control. The PCR products were blotted and subsequently hybridized, first with the radiolabeled HMGA2 cDNA and then with a β-actin probe, as a control for RNA loading. The sources of RNAs are shown. B, RT-PCR analysis of HMGA2 gene expression in ALT cells using primers specific for exons 4 and 5 of the gene. cDNAs from ALT cells were amplified with exons 4 and 5 HMGA2-specific primers and coamplified with β-actin as an internal control. The PCR products were blotted and hybridized as described in A. The sources of RNAs are shown. C, Western blot analysis of the HMGA2 proteins in ALT cells. Normal rat thyroid cells (PC Cl 3) were used as a negative control. PC MPSV cells (those infected with the myeloproliferative sarcoma virus) were used as a positive control. As a control for equal protein loading, the blotted proteins were incubated with γ-tubulin-specific antibodies. The sources of proteins are shown.
Characteristics of atypical lipoma cells
Because it has been shown that the presence of supernumerary ring chromosomes may result in the amplification of only a part of the gene, we decided to investigate HMGA2 expression in the ALT cells using primers specific for exons 4 and 5, the region 3′ to the third large intronic sequence in which most of the breakpoints occur in lipomas. As shown in Fig. 1B ⇓ , three of the ALT cells that were negative for HMGA2 expression using primers on exons 1 and 5 (cells ALT 8, ALT 13, and ALT 15) turned to be positive for exons 4 and 5, whereas ALT 2 and ALT 3 cells showed no amplification. As expected, ALT 1 and ALT 11 cells were confirmed to be positive. This result may be because of a break into the third large intronic sequence, followed by amplification of the 3′ downstream HMGA2 region.
Generation of an Adenovirus Construct Carrying the HMGA2 Gene in an Antisense Orientation.
To investigate the role of the HMGA2 gene in ALT cells, we decided to use an antisense approach that specifically inhibits HMGA2 protein expression. To this purpose, we generated a replication-defective adenovirus carrying the HMGA2 gene in the antisense (Ad-A2as) orientation and a marker protein easily detected, such as GFP (Fig. 2) ⇓ .
Map of the Ad-A2as virus carrying the HMGA2 gene in antisense orientation and GFP.
Infection of ALT Cells with Ad-A2as Suppresses HMGA2 Protein Synthesis.
Three ALT cells overexpressing HMGA2 (ALT1, ALT 4, and ALT 10) and three ALT cells (ALT 2, ALT 3, and ALT 8) that did not show the expression of the entire HMGA2 were infected with the Ad-A2as carrying the HMGA2 sequences in antisense orientation. We first demonstrated by direct fluorescence the expression of the adenoviral construct (data not shown) in the ALT cells infected with Ad-A2as. This result was further confirmed by a strand-specific RT-PCR assay (Fig. 3A) ⇓ . In fact, a specific amplification was observed only in the infected ALT cells and not in the control uninfected cells. Subsequently, Western blot analysis showed a drastic reduction of HMGA2 protein levels in the ALT cells overexpressing the HMGA2 protein after the infection with Ad-A2as (Fig. 3B) ⇓ . Conversely, Ad-GFP infection did not affect HMGA2 levels in either ALT1 or ALT4 or ALT 10 cells (Fig. 3B) ⇓ . No changes in γ-tubulin protein levels were observed in the ALT cells infected with the same adenoviruses, which indicates that an equal amount of protein was loaded and that no nonspecific inhibition of protein synthesis occurred in the infected cells.
Inhibition of HMGA2 protein synthesis in human ALT cells by Ad-A2as. A, Strand-specific RT-PCR assay for the HMGA2 antisense. Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as an internal control. B, Western blot analysis of the HMGA2 proteins in ALT cells infected with Ad-A2as and Ad-GFP at 10 pfu/cell. As a control for equal protein loading, the blotted proteins were incubated with γ-tubulin-specific antibodies. The sources of proteins are shown. CTRL, control.
Inhibitory Growth Effect of Ad-A2as Is Restricted to ALT Cells Overexpressing the HMGA2 Protein.
To study the effect of Ad-A2as on cell growth, we plated the ALT 1, ALT 4, ALT 10, ALT 2, ALT 3, and ALT 8 cells, and the next day we exposed them to 10 and 50 pfu/cell of Ad-A2as or Ad-GFP. Treated cells were harvested daily for cell counting. As shown in Fig. 4 ⇓ , infection of ALT 1, ALT 4, and ALT 10 cells with the Ad-A2as virus significantly reduced cell numbers over the 4 days examined. In contrast, the growth of ALT 2, ALT 3, and ALT 8 cells (data not shown for the last two cells), which do not express the HMGA2 protein, was not affected by adenovirus infection (Fig. 4) ⇓ , even though antisense HMGA2 sequences were expressed in the Ad-A2as-infected ALT 2, ALT 3, and ALT 8 cells (Fig. 3A) ⇓ . No effect on cell growth was achieved by Ad-GFP infection in all of the infected cell lines.
Cell growth of Ad-A2as-infected cells. ALT 1, ALT 2, ALT 4, and ALT 10 cells were plated (50,000/well) and the next day were exposed to 10 or 50 pfu/cell of either Ad-A2as or Ad-GFP and harvested daily for cell counts.
The Ad-A2as Virus Induces Apoptosis in the HMGA2-expressing ALT Cells.
In an attempt to determine the mechanism by which suppression of HMGA2 synthesis inhibits the growth of the neoplastic cells examined, ALT 1 and ALT 10 Ad-A2as-infected cells were analyzed by a TUNEL assay.
This assay (Fig. 5) ⇓ showed a significant positivity for DNA fragmentation in the ALT 1 and ALT 10 (not shown) cells infected with the Ad-A2as virus. Conversely, the number of the ALT 8 and ALT 2 (not shown) cells infected with the same virus showing positive staining was not significantly different from the uninfected or Ad-GFP-infected cells. These results, confirmed also by DNA laddering (data not shown), indicate that the specific suppression of HMGA2 proteins is able to induce apoptosis only in the ALT cells that overexpress HMGA2.
Inhibition of the HMGA2 protein synthesis induces apoptosis in ALT cells. The percentage of ALT 1 and ALT 8 cells positive for apoptosis by the TUNEL assay. Cells were treated with PBS, or infected with Ad-GFP 10 pfu/cell, or infected with Ad-A2as (10 pfu/cell).
DISCUSSION
Many mesenchymal neoplasms are often associated with some nonrandom cytogenetic abnormalities (36) . In the case of ALT, several studies have observed the presence of supernumerary ring and/or long marker chromosomes in 93% of these neoplasms (3 , 37) . Interestingly, chromosome painting by fluorescence in situ hybridization revealed that these marker chromosomes mostly contain DNA sequences from the long arm of chromosome 12 (3) . In particular, the 12q13–15 region is found to be amplified within these aberrant chromosomes. This region contains several genes (such as MDM2, CDK4, and HMGA2) that play a direct role in human tumorigenesis (4 , 38 , 39) . These cytogenetic findings are consistent with previous data that describe the HMGA2 gene as a frequent target of rearrangements because of chromosomal translocations that involve the 12q13–15 region (11, 12, 13, 14) .
In this study, we analyzed HMGA2 gene expression in 12 primary cell cultures derived from ALTs. RT-PCR showed 58% of these tumors to be positive for HMGA2 expression using primers specific for exons 1 and 5 of the gene, whereas 83% of tumors were found to be positive when we used primers specific for exons 4 and 5. This is consistent with the finding that chromosomal rearrangements of the large third intron of the HMGA2 gene often occur in benign mesenchymal tumors (26 , 27) . These data strongly suggest that HMGA2 may have a causal role in the process of ALT carcinogenesis. In fact, enhanced HMGA2 expression leads to transformation and anchorage-independent cell growth in two experimental cell lines (40) . Moreover, transgenic mice overexpressing HMGA2 develop mixed growth hormone cell/prolactin cell pituitary adenomas and natural killer-T cell lymphomas (33 , 41) . In addition, analysis of HMGA2 null mice has implicated the HMGA2 gene in the control of fat cell proliferation and adipocyte homeostasis (18) .
Because we have previously shown that the suppression of HMGA2 prevents the neoplastic transformation of rat thyroid cells, induced by murine retroviruses (21) , we decided to use an antisense approach to study the effect of HMGA2 protein suppression on ALT cells.
We generated an adenovirus carrying the HMGA2 gene in the antisense orientation and subsequently we infected six ALT cells, expressing and not expressing HMGA2, respectively. The result was a strong decrease of HMGA2 protein levels and a significant growth inhibition, only in the ALT cells overexpressing the HMGA2 gene. ALT cells negative for HMGA2 expression were unaffected by adenoviral treatment, and all of the ALT cells were unaffected when treated with the control virus. In an attempt to understand the mechanism by which cell growth is impaired in Ad-A2as-treated cells, we performed TUNEL and genomic DNA laddering (data not shown) assays. Apoptosis seems to account for the death of the cells treated by Ad-A2as.
The mechanism by which the suppression of HMGA2 protein synthesis inhibits neoplastic cell growth has yet to be defined, but several hypotheses may be considered. As suggested by our previous findings, the suppression of HMGA2 is able to prevent the neoplastic transformation of rat thyroid cells by blocking the induction of the activator protein 1 transcription factor that is required for the development of a malignant phenotype (42) . In addition, recent data from our laboratory suggest that HMGA2 overexpression might interfere with the RB-E2F pathway, by direct binding of HMGA2 to the Retinoblastoma (RB) protein, deregulating the G1-S cell cycle checkpoint that controls cell proliferation. 4
The results of our study suggest that targeting HMGA2 gene expression may provide an alternative strategy for the therapy of some ALTs. Because ALT neoplasias are typically found in sites such as the retroperitoneum or mediastinum, locations where a wide surgical resection is hard to achieve, repeated recurrences often occur, leading to patient death. In those cases that are resistant to conventional therapy, novel therapeutic approaches are necessary. In this study, we show that the suppression of HMGA2 proteins has several attractive features. First of all, the delivering of antisense molecules through adenoviral infection is highly efficient in specifically targeting the HMGA2 gene. This adenovirus treatment is able to inhibit cell growth and induce apoptosis in ALT neoplastic cells overexpressing HMGA2. This approach potentially has the advantage of a low or null interference with the function of normal cells because HMGA2 gene expression is restricted to embryogenesis and to growing adipocytes. In fact, the growth inhibition of Ad-A2as-treated ALT cells is limited to the cells overexpressing HMGA2, and no cytotoxic effects were observed in the others. Additional studies are in progress in our laboratory to assess the efficacy of Ad-A2as treatment in vivo and to evaluate a possible synergism with antineoplastic drugs.
Acknowledgments
We thank the Associazione Partenopea per le Ricerche Oncologiche for support. We are grateful to Eric S. Martin for editing the manuscript.
Footnotes
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.
↵1 Supported by grants from the Associazione Italiana Ricerca sul Cancro (Progetto Speciale Oncosoppressori), the Progetto Finalizzato “Biotecnologie” of the Consiglio Nazionale delle Ricerche, the Ministero dell’Università e Ricerca Scientifica e Tecnologica projects “Terapie antineoplastiche innovative” and “Piani di Potenziamento della Rete Scientifica e Tecnologica,” and the Ministero della Sanità.
↵2 To whom requests for reprints should be addressed, at Dipartimento di Biologia e Patologia Cellulare e Molecolare, Facoltà di Medicina e Chirurgia di Napoli, Università degli Studi di Napoli, 80131 Naples, Italy. Phone: 39-081-7463056; Fax: 39-081-7463037; E-mail: afusco{at}napoli.com
↵3 The abbreviations used are: ALT, atypical lipomatous tumor; WD, well-differentiated; HMGA, high-mobility group A; pfu, plaque-forming unit(s); GFP, green fluorescent protein; TUNEL, terminal deoxynucleotidyltransferase-mediated nick end labeling.
↵4 M. Fedele, manuscript in preparation.
- Received May 2, 2003.
- Revision received July 28, 2003.
- Accepted August 8, 2003.
- ©2003 American Association for Cancer Research.
References
Volume 63, Issue 21
