Expression of Reciprocal Fusion Transcripts of the HMGIC and LPP Genes in Parosteal Lipoma

Expression of Reciprocal Fusion Transcripts of the HMGIC and LPP Genes in Parosteal Lipoma

Expression of Reciprocal Fusion Transcripts of the HMGIC and LPP Genes in Parosteal Lipoma Marleen M. R. Petit, Sarah Swarts, Julia A. Bridge, and Wim...

423KB Sizes 0 Downloads 39 Views

Expression of Reciprocal Fusion Transcripts of the HMGIC and LPP Genes in Parosteal Lipoma Marleen M. R. Petit, Sarah Swarts, Julia A. Bridge, and Wim J. M. Van de Ven

ABSTRACT: Parosteal lipomas are rare benign neoplasms of adipose tissue that exhibit a contiguous relationship with the periosteum. These lipomas of the bone share some histopathologic features with their commonly occurring soft tissue counterparts. The latter are well-characterized cytogenetically, primarily by rearrangements involving chromosome region 12q13–q15. In particular, translocations involving 12q13–q15 are prominent, with chromosomal region 3q27–q28 as the most frequent translocation partner. Recently, we established that the genes HMGIC at 12q15 and LPP at 3q27-28 are affected by the 3;12-translocation and demonstrated that, as a direct result, HMGIC/LPP and LPP/HMGIC fusion transcripts are expressed in soft tissue lipomas. In this study, cytogenetic and molecular analyses revealed similar findings in a parosteal lipoma. Specifically, a t(3;12)(q28;q14) was detected cytogenetically in a parosteal lipoma from a 51-year-old female and subsequently confirmed by FISH utilizing a chromosome 3 breakpoint spanning YAC probe and chromosome 12 breakpoint flanking cosmid probes. RTPCR analysis showed expression of HMGIC/LPP and LPP/HMGIC fusion transcripts in this parosteal lipoma; nucleotide sequence analysis revealed that these transcripts are identical to those expressed in soft tissue lipomas characterized by a 3;12-translocation. These findings lend further support to a common histopathogenesis between lipomas of soft tissue and parosteal origin. © Elsevier Science Inc., 1998 INTRODUCTION Lipomas, benign adipose tissue tumors, are the most common tumors of mesenchymal origin in humans. Lipomas most often arise superficially in the soft tissue, but occasionally they are deep-seated [1]. Parosteal lipomas are rare deep-seated tumors comprising less than 0.5% of all lipomas. These lesions exhibit a contiguous relationship with the periosteum; approximately 50% are associated with an underlying osseous reaction [2]. Typically, there is focal cortical hyperostosis involving the underlying bone, which can vary from a small focus of cortical thickening to a large bony excrescence. Cortical erosion and osseous bowing have also been reported but are less common. Virtually any bone can be involved, but the radius and femur are the most prominent. Clinically, a parosteal

From the Laboratory for Molecular Oncology, Center for Human Genetics, University of Leuven and Flanders Interuniversity Institute for Biotechnology (M. M. R. P., W. J. M. VdV.), Leuven, Belgium; and the Departments of Pathology, Microbiology, Pediatrics, and Orthopaedic Surgery, University of Nebraska Medical Center (S. S., J. A. B.), Omaha, Nebraska, USA. Address reprint requests to: Dr. Wim J. M. Van de Ven, University of Leuven, Center for Human Genetics, Herestraat 49, B3000 Leuven, Belgium. Received October 21, 1997; accepted February 12, 1998. Cancer Genet Cytogenet 106:18–23 (1998)  Elsevier Science Inc., 1998 655 Avenue of the Americas, New York, NY 10010

lipoma presents as a slowly enlarging, often asymptomatic mass; however, loss of motor and/or sensory function as a result of the compression or stretching of a nerve by the expanding lipoma may also be seen. The etiology of parosteal lipomas is not known. Because of their intimate relationship to the bone, they are considered lipomas of bone. During the last decade, soft tissue lipomas have been studied extensively by cytogenetic analysis. At least 60% are karyotypically abnormal, with translocations involving chromosome segment 12q13–q15 most frequent [3, 4]. Nearly every chromosome has been described as a translocation partner of 12q13–q15, but chromosome 3 at bands q27–q28 is preferentially involved, representing approximately 25% of all lipomas with a 12q13–q15 rearrangement. Recently we [5], as well as others [6], identified the HMGIC gene at 12q15 to be consistently affected in lipomas and a variety of other benign mesenchymal tumor types characterized by genetic aberrations involving 12q13-q15. We subsequently identified the LPP-gene at 3q27–q28 as the preferred translocation partner gene of HMGIC [7]. HMGI-C is a member of the HMGI family of high mobility group proteins and consists of three DNAbinding domains followed by an acidic tail. The protein binds to the minor groove of the chromosomal DNA [8] and it has been suggested that it may act as an architectural factor [9] in the nuclear scaffold [10] and play a criti-

0165-4608/98/$19.00 PII S0165-4608(98)00038-7

HMGIC and LPP Involvement in a Parosteal Lipoma cal role in the assembly of stereoscopic transcriptional complexes in the nucleus [11]. It is of interest to note that creation of a null mutation in the HMGIC gene in mice causes aberrant growth resulting in the so-called pygmy phenotype [12] and that retrovirally induced neoplastic transformation of rat thyroid cells is suppressed by the inhibition of HMGI-C protein synthesis. As far as the translocation partner gene is concerned, the LPP gene is a member of group 3 of the relatively large LIM protein gene family [13, 14]. The protein is unusually proline-rich in its amino-terminal region and has three LIM domains in its carboxy-terminal region. The function of LPP is not known but database screening revealed that the LPP-encoded protein displays some similarity to the LIM protein zyxin [15, 16], which is found in immunocytochemical studies at sites of cell adhesion [17]. In this study, the chromosomal breakpoints of the 3;12translocation (3q27-28 and 12q14) in a parosteal lipoma of the femur of a 51-year-old female were analyzed using various technologies to evaluate, at the molecular level, the implications of this genetic aberration.

MATERIALS AND METHODS Soft Tissue Lipoma Cell Line and Parosteal Lipoma The origin, chromosome aberrations, and growth conditions of the soft tissue lipoma cell line Li-538/SV40 have been described before [18, 19]. The primary parosteal lipoma was obtained at the University of Nebraska Medical Center, Omaha. The clinical findings and cytogenetic methodologies have been described previously [20]. Briefly, the tumor was removed from a 51-year-old female who noted a slowly enlarging, occasionally painful mass in the posterior aspect of her left thigh approximately 5 months prior to admission. Radiographic studies that included plain films, CT scan, MRI and bone scan revealed a mass of the same intensity as fat displacing the vastus medialis and vastus intermedius juxtaposed to the cortex of the medial aspect of the femur. A 5.0 ⫻ 5.0 ⫻ 10.0 cm well-circumscribed adipose tissue lesion consistent with benign parosteal lipoma was removed in total with the excrescence from the femur. A sterile representative portion of this specimen was mechanically and enzymatically disaggregated and cultured for 4–6 days. Twenty GTG-banded metaphase cells were analyzed following standard harvesting procedures [21]. DNA Probes and Labeling by Nick Translation HMGIC-specific cosmid clones 27E12 and 142H1 were isolated from an arrayed human chromosome 12-specific cosmid library (LL12NCO1) [22] obtained from Lawrence Livermore National Laboratory (Courtesy P. de Jong). CEPH YAC clone 192B10, localized to 3q27–q28, has been described before [7]. The clones are indicated by their microtiter plate addresses. DNA extraction of the YAC and cosmid clones was performed according to standard procedures [23]. Cosmids 142H1 (1 ␮g) and 27E12 (1 ␮g), were labeled in SpectrumRed (Vysis Nick Translation Kit; Vysis, Down-

19 ers Grove, IL) according to the manufacturer’s instructions; the labeling reaction proceeded overnight. YAC 192B10 (1 ␮g) was labeled with biotin using the BioNick labeling system (Life Technologies, Inc., Gaithersburg, MD). The manufacturer’s instructions were followed, with one exception: 1 ␮l (10 units/␮l) of DNA polymerase I (Life Technologies, Inc.) was added to the reaction mixture. The reaction proceeded for 75 minutes. The products were purified by column chromatography. Each probe was precipitated with herring sperm DNA (10 ␮g), 1/10 volume of 3 M sodium acetate, and 2 volumes cold absolute ethanol; the DNA was resuspended in hybrisol VII (18 ␮l, Oncor Inc., Gaithersburg, MD) and unlabeled Cot-1-DNA (2 ␮l, 1 mg/ml, Life Technologies, Inc.). FISH and Image Analysis A cell suspension from the original case was removed from liquid nitrogen, and cultured and harvested as described previously [20]. Briefly, the frozen cells were cultured in RPMI 1640 media and passaged once prior to harvest. Two to four hours prior to harvest, cells were exposed to Colcemid (0.02 ␮g/ml). Following hypotonic treatment, the preparations were fixed three times with methanol:glacial acetic acid (3:1). The coverslips were aged overnight at room temperature prior to use. The metaphase slides were incubated at 37⬚C for 20 minutes in 2 ⫻ SSC and dehydrated in graded ethanol (70%, 80%, 100%) at room temperature. Spectrum Redlabeled cosmid 142H1 (50 ng), Spectrum Red-labeled cosmid 27E12 (50 ng), and biotin-labeled YAC 192B10 (50 ng) were mixed in 7 ␮l hybrisol VII (Oncor Inc.). The probe preparation was dropped onto the metaphase slide and sealed under a coverslip. The probe and target sample were codenatured for 5 minutes in a 75⬚C oven. The sample was hybridized overnight at 37⬚C in humidity. After hybridization, the coverslip was removed and the slide washed three times in 50% formamide/2 ⫻ SSC (pH 7.0) 45⬚C for 2 minutes each. This was followed by a room temperature 1 ⫻ PBD (phosphate-buffered detergent) wash for 2 minutes. The slide was immunohistochemically stained by a layer of fluorescein isothiocyanate (FITC)-avidin (Oncor Inc.) for 5 minutes at 37⬚C in humidity; the slide was washed three times for 2 minutes each, with 1 ⫻ PBD at room temperature. The slide was counterstained with DAPI-Antifade (Oncor Inc.) and coverslipped. Images of each fluorochrome (DAPI, FITC, and Spectrum Red) were collected from each metaphase using a Zeiss axioscope (Zeiss, Thornwood, NY) and a charge-coupled device camera interfaced with a Cytovision imaging system (Applied Imaging International, Ltd.). RNA Isolation and RT-PCR Analysis Total RNA was isolated from lipoma cell line Li-538/SV40 and the primary parosteal lipoma using TRIzol Reagent (Total RNA Isolation Reagent, GIBCO/BRL) according to the supplier’s instructions. RT-PCR experiments (reverse transcriptase reaction combined with a polymerase chain reaction) were performed following the procedure described by Schoenmakers et al. [5]. To detect a possible HMGIC/LPP fusion transcript, cDNA was synthesized us-

20

M. M. R. Petit et al.

ing the LPP-specific primer 729. Thereafter, PCR-amplification was performed with primers 517 and 730, in the first round, and nested primers 518 (biotinylated) and 731, in the second round of PCR. To detect a possible LPP/ HMGIC fusion transcript, cDNA synthesis was started with HMGIC-specific primer 645. Primers 649 and 821 were used in the first round of PCR while, in the second round, nested primers 822 (biotinylated) and 638 were used. The location and nucleotide sequences of the primers are given in Figure 1 and Table 1, respectively. Results were evaluated by electrophoretic analysis of 10 ␮l aliquots of the reaction product using polyacrylamide minigels. Finally, it should be noted that, under the experimental conditions used, the observed mobility of the various PCR products was always somewhat higher than expected, on the basis of the nucleotide sequence data of the PCR products. Nucleotide Sequence Analysis Nucleotide sequences were determined according to the dideoxy chain termination method [24]. RT-PCR fragments were solid-phase sequenced using fluorescein isothiocyanate (FITC)-labeled LPP- or HMGIC-specific primers and the Autoread Sequencing Kit (Pharmacia Biotech), after isolation of the biotinylated strand using Dyna-

beads M-280 Streptavidin (Dynal). Sequencing results were obtained using an Automated Laser Fluorescent (A.L.F.) DNA sequencer (Pharmacia Biotech) and standard 30 cm, 6% HydrolinkR, Long RangeTM gels (AT Biochem). The nucleotide sequences were analyzed using the sequence analysis software Lasergene (DNASTAR, Inc.). All oligonucleotides were purchased from Pharmacia Biotech.

RESULTS The HMGIC Gene Is Affected by the t(3;12) in the Parosteal Lipoma Cytogenetic analysis of short-term cultured cells of the parosteal lipoma revealed the following chromosomal complement: 46,XX,t(3;12)(q28;q14). FISH analysis was performed to determine in more detail the nature of the chromosomal translocation. FISH analysis performed with a mixture of the HMGIC-specific cosmids 142H1 and 27E12 revealed signals on the normal chromosome 12 as well as on the der(3) and der(12). Since cosmid 142H1 contains the first two exons of the HMGIC gene and cosmid 27E12 exons 4 and 5, these results suggested that the HMGIC gene is directly affected by the 3;12-translocation in the parosteal lipoma. Similar FISH analysis utiliz-

Figure 1 FISH analysis of cultured cells from the parosteal lipoma. Metaphase chromosomes were hybridized using a mixture of cosmid clones 142H1 and 27E12 as well as YAC clone 192B10. Hybridization with cosmids 142H1 and 27E12, which are localized to chromosome 12 (142H1 is proximal to 27E12 on 12q13-q15), is visible as red signals; YAC 192B10, which is localized to 3q27–q28, is visible as a green signal. Normal chromosomes 3 and 12 are indicated with an open arrow, the der(3)t(3;12)(q28;14) is shown with an arrowhead, and the derivative 12 is indicated with an open arrowhead. Note the juxtaposition of red and green signals on the derivative chromosomes, indicative of translocation.

21

HMGIC and LPP Involvement in a Parosteal Lipoma Table 1 Sequences of RT-PCR primers Primer HMGIC-specific primers 517 518 638 645 649 LPP-specific primers 729 730 731 821 822

Nucleotide sequence

5⬘CTTCAGCCCAGGGACAAC 5⬘CGCCTCAGAAGAGAGGAC 5⬘ATAGTCCTTTTTAAGGTTATGTGA 5⬘TACAGCAGTTTTTCACTA 5⬘CCTGGGACTGTGAAGGGATTACAA 5⬘GCAAGGTTAATAGCAGA 5⬘GCGTAGTTTGATGGCCTTTA 5⬘GTGTTCGTTCTTCTCAGTGC 5⬘CAGTCACTGGTCCCAAGAAGACCT 5⬘AAGAAGACCTATATCACAGATCCT

ing the chromosome 3 breakpoint spanning probe YAC 192B10 revealed splitting of the chromosome 3 signal. The latter observation raised the possibility that the LPP gene acts as translocation partner of HMGIC in this tumor. The FISH results are illustrated in Figure 1. Expression of HMGIC/LPP and LPP/HMGIC fusion transcripts in the parosteal lipoma To confirm LPP and HMGIC involvement, RNA isolated from the parosteal lipoma was tested by RT-PCR for the presence of fusion transcripts. Selected HMGIC- and LPP-

Figure 2 RT-PCR analysis of mRNA isolated from a parosteal lipoma (lanes 1 and 4) and soft tissue lipoma cell line Li-538/ SV40 (lanes 2 and 5). In control experiments (lanes 3 and 6), total RNA from human placenta was tested. Lanes 1–3: HMGIC/LPPspecific RT-PCR analysis; lanes 4–6: LPP/HMGIC-specific RTPCR analysis. M: molecular weight markers, 2645, 1605, 1198, 676, 517, 460, 396, and 350 bp.

specific primers (Table 1) were used. Analysis of total RNA of the parosteal lipoma for the presence of HMGIC/ LPP transcripts resulted in the identification of a single PCR product (Fig. 2, lane 1). A similar PCR product (Fig. 2, lane 2) was detected with RNA isolated from cell line Li538/SV40, for which expression of an HMGIC/LPP fusion transcript has been previously demonstrated [7]. In RTPCR experiments using primer sets to detect expression of LPP/HMGIC fusion transcripts in the parosteal lipoma tumor cells (Fig. 2, lane 4) and the Li-538/SV40 cell line (Fig. 2, lane 5), PCR products of similar molecular weight were detected. The lipoma cell line is known to express an LPP/HMGIC fusion transcript [7]. The RT-PCR results indicate that, in the parosteal lipoma, transcripts are expressed which consist of HMGIC and LPP sequences and that these are highly similar to those expressed in the Li538/SV40 cell line. To further characterize them, the nucleotide sequence of the PCR products was determined (data not shown). Comparison of these nucleotide sequence data to those of the wild type HMGIC and LPP cDNA sequences revealed that, in the HMGIC/LPP hybrid transcript of the parosteal lipoma, the LPP sequences were fused to HMGIC sequences with a sequence diversion point immediately downstream of HMGIC exon 3. This indicated that the rearrangement in the HMGIC gene had occurred within the large intron (intron 3) of the gene, similarly to the Li-538/ SV40 cell line. The diversion point in the LPP sequences was found immediately downstream of the sequences that encode the first LIM domain, indicating that the rearrangement in the LPP gene had occurred within intron 8 of the gene; this also was similar to the Li-538/SV40 cell line. Evaluation of the nucleotide sequence of the LPP/HMGIC hybrid transcript revealed that it was complementary to that of HMGIC/LPP. In conclusion, the HMGIC/LPP transcript detected by RT-PCR was found to encode the three DNA binding domains of HMGI-C followed by the LIM2 and LIM3 domains of LPP, while the LPP/HMGIC transcript encodes the proline-rich domain and the LIM1 domain of LPP followed by the spacer and acidic tail of HMGI-C (Fig. 3).

DISCUSSION Parosteal lipomas are benign adipose tumors that exhibit an intimate relationship with the adjacent periosteum. According to Fleming et al. [2], the original description of this very uncommon lesion was published in the German literature by Seerig [25] in 1836. D’Arcy Power [26] (in 1888) introduced the term ‘parosteal lipoma’ to describe a fatty tumor connected with the periosteum of the femur. However, due to the intimate relationship of these tumors with the periosteum, many surgeons and pathologists believe that these tumors originate from the periosteum. Therefore some authors prefer to use the term ‘periosteal lipoma’ which implies a definite origin from the periosteum. One of these authors, Edwin Bartlett [27], wrote in his report of two cases of ‘periosteal lipoma’: “At operation the tumor capsule and periosteum merged together so

22

M. M. R. Petit et al.

Figure 3 Schematic representation of the HMGIC/LPP and LPP/HMGIC fusion transcripts found to be expressed in parosteal lipoma. Coding sequences are represented as boxes and the protein domains they encode are indicated. Noncoding sequences are represented as solid lines. The HMGIC/LPP fusion transcript contains the coding sequences for the three DNA binding domains of HMGI-C and the two most carboxy-terminal LIM domains of LPP. The reciprocal LPP/HMGIC fusion transcript contains the coding region for the proline-rich domain and the first LIM domain of LPP followed by the spacer and the acidic tail of HMGI-C. The relative positions of the various primers (numbered boxes), which are listed in Table 1, are given. Abbreviations: DBD, DNA binding domain; AD, acidic domain.

completely that there was no line of cleavage. The tumor had to be separated from the bone by sharp dissection.” However, such a firm attachment to the periosteum is not always observed. Fleming et al. [2] stated in 1962 that the term ‘parosteal lipoma’ was more suitable since this term indicates the continguity with bone rather than a definite origin from the periosteum. Since then, the modern literature tends to favor the term parosteal because the precise site of origin is still not known. The results of the current study provide new insight into this intriguing matter. Detection of a t(3;12)(q28;q14) cytogenetically in a parosteal lipoma raised the possibility that the HMGIC gene at 12q15 and the LPP gene at 3q27– q28 were directly affected in the tumor. Rearrangement of these same genes is well recognized in soft tissue lipomas [5–7]. RT-PCR and nucleotide sequence analysis in the current study confirmed that both genes are also affected in parosteal lipoma. Expression of both HMGIC/LPP and LPP/HMGIC fusion transcripts was detected. The former, which consists of the three DNA binding domains of

HMGI-C and the two most carboxy-terminal LIM domains of LPP, is identical to the HMGIC/LPP fusion transcripts most commonly found in soft tissue lipomas with a 3;12translocation [7]. The reciprocal LPP/HMGIC fusion transcript, which consists of the proline-rich domain and the first LIM domain of LPP followed by the spacer and the acidic tail of HMGIC, is rarely found. In fact, in addition to the parosteal lipoma presented here, Li-538 (from which cell line Li-538/SV40 was derived) is the only other lipoma identified until now, in which such a reciprocal fusion transcript has been detected. These results suggest that the HMGIC/LPP is pathogenetically relevant rather than the reciprocal transcript. Parosteal lipomas are essentially identical in their gross and histologic appearance to soft tissue lipomas, i.e., encapsulated, lobular, yellow soft tissue composed of mature lipocytes with either prominent or minimal amounts of interlobular fibrous connective tissue [28, 29]. In this study, we have demonstrated that these bone and soft tissue counterparts also share similar genetic findings. In conclu-

HMGIC and LPP Involvement in a Parosteal Lipoma

23

sion, these data provide additional support to a histopathogenetic relationship between parosteal lipoma and soft tissue lipoma.

13. Taira M, Evrard J-L, Steinmetz A, Dawid IB (1995): Classification of LIM domains. Trends Genet 11:431–432.

This work was supported in part by the Fonds voor Wetenschappelijk Onderzoek Vlaanderen (FWO), the Geconcerteerde Onderzoekacties 1997-2001,” the ASLK Programma voor Kankeronderzoek, and the Nebraska State Department of Health LB595. The authors would like to thank Reinhilde Thoelen, Ms. Joanne Degenhardt, and Ms. Mari Nelson for their technical assistance and Dr. James Neff for his clinical support. Marleen M.R. Petit is a Research Assistant of the FWO (Kom op tegen Kanker, Belgium).

15. Macalma T, Otte J, Hensler ME, Bockholt SM, Louis HA, Kalff-Suske M, Grzeschik KH, Von der Ahe D, Beckerle MC (1996): Molecular characterization of human zyxin. J Biol Chem 271:31470–31478.

REFERENCES

18. Kazmierczak B, Bartnitzke S, Hartl M, Bullerdiek J (1990): In vitro transformation by the SV40 “early region” of cells from a human benign salivary gland tumor with a 12q13-q15 rearrangement. Cytogenet Cell Genet 53:37–39.

1. Enzinger FM, Weiss SW (1988): Soft Tissue Tumors. 2nd ed. C.V. Mosby, St. Louis. 2. Fleming RJ, Alpert M, Garcia A (1962): Parosteal lipoma. AJR 87:1075–1084. 3. Sreekantaiah C, Leong SPL, Karakousis CP, McGee DL, Rappaport WD, Villar HV, Neal D, Fleming S, Wankel A, Herrington PN, Carmona R, Sandberg AA (1991): Cytogenetic profile of 109 lipomas. Cancer Res 51:422–433. 4. Mandahl N, Höglund M, Mertens F, Rydholm A, Willén H, Brosjö O, Mitelman F (1994): Cytogenetic aberrations in 188 benign and borderline adipose tissue tumors. Genes Chromosom Cancer 9:207–215. 5. Schoenmakers EFPM, Wanschura S, Mols R, Bullerdiek J, Van den Berghe H, Van de Ven WJM (1995): Recurrent rearrangements in the high mobility group protein gene. HMGI-C, in benign mesenchymal tumours. Nat Genet 10:436–444. 6. Ashar HR, Schoenberg Fejzo M, Tkachenko A, Zhou X, Fletcher JA, Weremowicz S, Morton CC, Chada K (1995): Disruption of the architectural factor HMGI-C: DNA-binding AT hook motifs fused in lipomas to distinct transcriptional regulatory domains. Cell 82:57–65. 7. Petit MMR, Mols R, Schoenmakers EFPM, Mandahl N, Van de Ven WJM (1996): LPP, the preferred fusion partner gene of HMGIC in lipomas, is a novel member of the LIM protein gene family. Genomics. 36:118–129. 8. Maher JF, Nathans D (1996): Multivalent DNA-binding properties of the HMG-I proteins. Proc Natl Acad Sci USA 96:6716–6720. 9. Wolffe AP (1994): Architectural transcription factors. Science 264:1100–1101. 10. Saitoh Y, Laemmli UK (1994): Metaphase chromosome structure: bands arise from a differential folding path of the highly AT-rich scaffold. Cell 76:609–622. 11. Tjian R, Maniatis T (1994): Transcriptional activation: a complex puzzle with few easy pieces. Cell 77:5–8. 12. Zhou X, Benson KF, Ashar HR, Chada K (1995): Mutation responsible for the mouse pygmy phenotype in the developmentally regulated factor HMGI-C. Nature 376:771–774.

14. Sánchez-García I, Rabbitts TH: The LIM domain (1994): a new structural motif found in zinc-finger-like proteins. Trends Genet 10:315–320.

16. Zumbrunn J, Trueb B (1996): A zyxin-related protein whose synthesis is reduced in virally transformed fibroblasts. Eur J Biochem 241:657–663. 17. Sadler I, Crawford AW, Michelsen JW, Beckerle MC (1992): Zyxin and cCRP: two interactive LIM domain proteins associated with the cytoskeleton. J Cell Biol 119:1573–1587.

19. Schoenmakers EFPM, Kools PFJ, Mols R, Kazmierczak B, Bartnitzke S, Bullerdiek J, Dal Cin P, De Jong PJ, Van den Berghe H, Van de Ven WJM (1994): Physical mapping of chromosome 12q breakpoints in lipoma, pleomorphic salivary gland adenoma, uterine leiomyoma, and myxoid liposarcoma. Genomics 20:210–222. 20. Bridge JA, DeBoer J, Walker CW, Neff JR (1995): Translocation t(3;12)(q28;q14) in parosteal lipoma. Genes Chromosom Cancer 12:70–72. 21. Bridge JA, Sreekantaiah C, Mouron B, Neff JR, Sandberg AA, Wolsan SR (1991): Clonal chromosomal abnormalities in desmoid tumors. Implications for histopathogenesis. Cancer 9:430–436. 22. Montgomery KT, LeBlanc JM, Tsai P, McNinch JS, Ward DC, deJong PJ, Kucherlapati R, Krauter KS (1993): Characterization of two chromosome 12 cosmid libraries and development of STSs from cosmids mapped by FISH. Genomics 17:682–693. 23. Sambrook J, Fritsch EF, Maniatis T (1989): Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 24. Sanger F, Nicklen S, Couslon A (1977): DNA sequencing with chain terminating inhibitors. Proc Natl Acad Sci USA 74:5463–5467. 25. Seerig W (1836): Geschichte eines sehr grossen Steatoms im Hinterhaupte eines 2 und 1/2 järigen Kindes. Mag Ges Heil: 511–514. 26. Power D’A (1888): A parosteal lipoma, or congenital fatty tumour, connected with the periosteum of the femur. Trans Pathol Soc London 39:270–272. 27. Bartlett EI (1930): Periosteal lipoma—Report of two cases. Arch Surg 21:1015–1022. 28. Goldman AB, DiCarlo EF, Marcove RC (1993): Case report 774. Skel Radiol 22:138–145. 29. Steiner M, Gould AR, Rasmussen J, LaBriola D (1981): Parosteal lipoma of the mandible. Oral Surg 52:61–65.