Main

Well-differentiated liposarcomas/atypical lipomatous tumors are among the most common sarcomas presenting in older adults.1, 2, 3, 4, 8 These tumors often arise in deep tissues as slow growing masses and can occasionally be difficult to distinguish morphologically from lipomas.1, 2, 3, 4, 8 Cytogenetically, well-differentiated liposarcomas/atypical lipomatous tumors are characterized by supernumerary ring and large marker chromosomes. These abnormal chromosomes are mainly composed of amplified genomic sequences derived from chromosome bands 12q13-15, and contain several genes. Among them, MDM2 seems to be the most consistently amplified gene.1, 2, 3, 4 During a high-resolution microarray comparative genomic hybridization experiment on a series of lipomatous neoplasms, we discovered that CPM, a gene that encodes carboxypeptidase M, was consistently amplified in liposarcomas but not in different subtypes of lipoma or normal fat.

In this study, we compared MDM2 and CPM on a series of lipomatous tumors to establish whether CPM could be used clinically to discriminate well-differentiated liposarcoma/atypical lipomatous tumors from lipomas.

Materials and methods

Tumor Samples

After receiving Institutional Review Board approval, formalin-fixed and paraffin-embedded specimens were identified from the archives of the Department of Laboratory Medicine and Pathology at the Mayo Clinic, Rochester, MN, USA. Hematoxylin and eosin (H&E)-stained sections were reviewed in all cases for diagnostic confirmation. The verified diagnostic categories included well-differentiated liposarcoma/atypical lipomatous tumors (n=32), ordinary lipomas (n=63), myxoid liposarcomas (n=7), undifferentiated pleomorphic sarcomas (n=15), pleomorphic lipomas (n=11), sarcomatoic renal cell carcinomas (n=2), angiomyofibroblastoma (n=2), low-grade myxofibrosarcoma (n=2), lipoblastoma (n=2), hibernoma (n=1), nodular fasciitis (n=1), and normal fat samples (n=17) (Tables 1 and 2).

Table 1 Results for MDM2 and CPM
Full size table
Table 2 Results for CPM only (independent set)
Full size table

FISH

All cases were evaluated with laboratory-developed fluorescent in situ hybridization (FISH) probe sets. Bacterial artificial chromosome (BAC) clones spanning the CPM and MDM2 loci, 12q13-15, were obtained from the Children's Hospital Oakland Research Institute (Oakland, CA, USA). BAC clones for CPM are RP11-717F7, RP11-426B12, RP11-630N19, RP11-1104N20, and RP11-1036O8. BAC clones for MDM2 are RP11-61F20, RP11-816C9, RP11-185H13, RP11-450G15. BAC clones spanned the CPM and MDM2 genes and are 729 kb and 765 kb, respectively. The chromosome 12 reference probes for FISH and chromogenic in situ hybridization (CISH) were CEP12 (Abbott Laboratory, North Chicago, IL, USA) and the chromosome 12 pericentromeric gene SYT10 (RP11-8P13, RP11-267D19, RP11-855O3, RP11-88P4, RP11-102G23, RP11-51O12 and RP11-1143D16), respectively. All of the identities for the BAC clones were individually confirmed by PCR and by hybridization on metaphase preparations from the peripheral blood of five normal individuals. Analytical sensitivity and specificity of each probe were calculated and their performance on paraffin-embedded tissue was verified on numerous normal tissue types, including the skeletal muscle, adipose tissue, and others. Amplification was defined as CPM/CEP12 and/or MDM2/CEP12 or CPM/SYT10 ratio ≥3. Normal structures that could be readily identified under the 4′, 6-diamidino-2-phenylindole (DAPI) staining on thin sections, such as the blood vessels and epidermis, were used as internal controls for the cases analyzed. Normal signal patterns were established by scoring 100 cells from each normal tissue.

DNA isolation was carried out using the Qiagen Plasmid Maxi kit (Qiagen, Valencia, CA, USA). DNA was labeled using a nick translation kit (Abbott Laboratory). Interphase molecular cytogenetic studies were carried out using 4-μm paraffin-embedded thin sections that were deparaffinized twice in xylene (15 min pretreatment), dehydrated once in 100% ethanol (5 min), and treated with 10 mmol/l citric acid (10 min, in a humidified microwave). Tissue sections were incubated in warm (37°C) sodium chloride–sodium citrate buffer (2 × SSC) for 5 min. Protein was digested with Digest-All 3 (Invitrogen Corporation, Carlsbad, CA, USA). After a brief wash in phosphate-buffered saline (1 × PBS), slides were sequentially dehydrated in ethanol (70, 80, and 100%) and air dried at room temperature. Tissue sections were denatured at 85°C for 5 min, and BAC probe hybridization was performed overnight in a humidified chamber at 37°C. Tissue sections were washed in 0.1% Nonidet P-40 (NP40) in 2 × SSC at 76°C for 2 min and then washed in the same solution at room temperature for 1 min. Slides were mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA) with 1.5 μg/ml DAPI medium (Vector Laboratories). Tumor samples were considered positive for amplification if a ratio ≥3:1 CPM/CEP12 and/or MDM2/CEP12 was found in >10% of 100 cells. Tumors were evaluated and scored by three independent investigators (MEJ, ARS, and CWR).

CISH

Using the same DNA extracted for FISH probes (CPM, MDM2, and SYT10), DNA was labeled using the DIG DNA-labeling mix (Roche Diagnostics, Mannheim, Germany). In brief, 5 μl of 0.2 μg/ml DNA and 2.5 × random primers were mixed and placed on a thermal cycler at 99°C for 5 min, followed by placing on ice for 1 min. In all, 5 μl 10 × DIG-dNTP (Roche Diagnostics), 19 μl water, and 1 μl Klenow fragment (Roche Diagnostics) were added, gently mixed, and placed on a thermal cycler overnight at 37°C. DNA was then purified as per the manufacturer's instructions using a microspin s-200HR column (GE Healthcare, Fairfield CT, USA) and precipitated by the addition of 1 μl of 20 μg/ml glycogen (Invitrogen Corporation) and 33 × 7.5 M ammonium acetate (Sigma-Aldrich, St Louis, MO, USA) and 2 × 100% ethanol. Samples were precipitated overnight in an ethanol bath at −80°C. Labeled probe was centrifuged at 14 000 r.p.m. for 15 min at 4°C and pellets were resuspended in 100 μl LSI/WCP hybridization buffer (Abbott Laboratory). CISH slides were prepared following the FISH processing procedure until the addition of probe. CISH probe (5 μl) and 2 μl COT-1 (Invitrogen Corporation) were added to the slides and hybridized at 80°C for 5 min, and then held at 37°C overnight. On the following day, coverslips were removed and slides were washed in 76°C 0.5 × SSC for 2 min, followed by a wash in 1 × PBS for 2 min. Slides were visualized by using the Spot-Light CISH Polymer detection kit (Zymed Laboratories, South San Francisco, CA, USA). Briefly, slides were incubated in 3% hydrogen peroxide for 5 min followed by three washes in PBS for 2 min each. Slides were blocked using in CAS-Block (Invitrogen Corporation) for 10 min. Reagent B (mouse-anti DIG) was added and incubated for 60 min followed by three washes in 1 × PBS, followed by the addition of 3′3-diaminobenzidine (DAB), incubated for 5 min, and washed in PBS. Following the addition of DAB, the slides were counterstained with hematoxylin, dehydrated in an ethanol series (70, 85, and 100%), immersed in xylene, and coverslipped using histomount. Tumor samples were considered positive for amplification if a ratio ≥3:1 CPM/SYT10 or MDM2/SYT10 was found in >10% of 100 cells analyzed. Tumors were evaluated and scored by two independent investigators (MEJ and AMO).

Comparative Genomic Hybridization

This study is part of a major genomic analysis of a large series of lipomatous neoplasms, which is currently in progress. Briefly, CPM was discovered using the Human CNV370-Quad BeadChip (Illumina, San Diego, CA, USA), which contains over 370 000 markers. Analysis was by multiple copy number variation (CNV) detection tools, including cnvPartition (Illumina), CGHAnalyzer (open source software from University of Pennsylvania), and Segmentation algorithm in Partek Genomic Suit (Partek Incorporated, St Louis, MO, USA).

Results

FISH and CISH

All cases were found to have detectable and analyzable signals. The data in Tables 1 and 2 depict the results for each tumor category. All 32 well-differentiated liposarcomas/atypical lipomatous tumors were found to have amplification of MDM2 and/or CPM, usually with >20 copies per cell (Tables 1 and 2, and Figure 3). Lipomas, lipoblastomas, and normal adipose tissue samples showed a normal FISH signal pattern with two copies of MDM2 and CPM (Tables 1 and 2, and Figure 1). Pleomorphic lipomas and pleomorphic liposarcomas showed FISH signal patterns consistent with numerical abnormalities of chromosome 12 but without amplification of either CPM or MDM2 (Figure 2). Two myxoid liposarcomas (29%) exhibited patterns consistent with monosomy 12 or loss of the CPM/MDM2 loci (Tables 1 and 2). The remaining cases showed normal FISH signal patterns. Using the same FISH scoring criteria, all 27 cases analyzed by CISH (21 lipomas, 4 liposarcomas, and 2 normal fat samples) were found to have detectable and analyzable signals. CISH produced results similar to FISH, resulting in normal patterns for lipomas and normal fat samples (Figure 1), and amplification for liposarcoma (Figure 3). Pleomorphic lipomas and myxoid liposarcomas showed signal patterns consistent with numerical chromosomal abnormalities of chromosome 12 without amplification, similar to what was seen with FISH (Figure 2).

Figure 1
figure 1

Lipoma. (a) Histological features of a representative case. (b) and (c) FISH images of this case with MDM2 and CPM probes, respectively. Orange signals represent either MDM2 or CPM. Green signals show chromosome 12 centromeres. (d) Normal CISH pattern for CPM in the same case. Brown signals represent normal copies of CPM.

Full size image
Figure 2
figure 2

Pleomorphic lipoma. (a) Histological features of a representative case. (b) and (c) FISH images of this case with MDM2 and CPM probes, respectively. Orange signals represent either MDM2 or CPM. Green signals show chromosome 12 centromeres. (d) CISH image of a CPM probe hybridized to a pleomorphic lipoma. Scattered brown signals represent multiple copies of CPM without amplification (clustering). These images show FISH and CISH signatures consistent with numerical abnormalities of chromosome 12 but no amplification of either MDM2 or CPM (CPM/SYT10 and MDM2/CEP12 ratios=1).

Full size image
Figure 3
figure 3

Well-differentiated liposarcoma/atypical lipomatous tumor. (a) Histological features of a representative case. (b) and (c) FISH images of this case show MDM2 and CPM amplification, respectively. Orange signals represent either MDM2 or CPM. Green signals show chromosome 12 centromeres. (d) CISH image of CPM probe hybridized to this case; clusters of brown signals represent CPM amplification. These images show FISH and CISH signatures consistent with the amplification of these genes (CPM/SYT10 and MDM2/CEP12 ratios ≥3).

Full size image

Discussion

Well-differentiated liposarcomas/atypical lipomatous tumors with minimal or no cytological atypia can be difficult to distinguish from lipomas and fibrolipomas.1, 3, 5, 6, 7, 8, 9 However, the identification of ring and giant rod chromosomes by cytogenetic analysis supports the diagnosis of well-differentiated liposarcomas/atypical lipomatous tumor. These abnormal chromosomes are mainly composed of amplified genomic sequences derived from chromosome bands 12q13-15 and contain several genes, including MDM2, CDK4 (SAS), TSPAN31, HMGA2, among others.8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 MDM2 seems to be amplified in almost all cases of well-differentiated liposarcoma/atypical lipomatous tumors, especially when molecular cytogenetic analysis is carried out.1, 9 In addition, MDM2 amplification has been described in up to 25% of other sarcomas.1, 2, 3

As a part of large genomic study of lipomatous neoplasms, we found CPM to be consistently amplified well-differentiated liposarcomas/atypical lipomatous tumors in a CNV screening array CGH.22 Our experiments showed co-amplification of MDM2 and CPM in all well-differentiated liposarcoma/atypical lipomatous tumors evaluated in an initial test set. An independent group of well-differentiated liposarcoma/atypical lipomatous tumors was then tested only for CPM. Amplification of this gene was observed in all instances. All other tumors were negative for CPM amplification. These results correlate with previous studies that showed consistent amplification of MDM2 in well-differentiated liposarcoma/atypical lipomatous tumors, but not in lipomas.1, 9, 22 FISH results were also corroborated by a CISH assay in 27 of these tumors. Both FISH and CISH are closely related techniques resulting in the ability to enumerate chromosome and/or gene status. FISH requires a fluorescent microscope, trained technologists, and the signals of slides fade over time. In the other hand, CISH signals can be read out by any pathologist in a bright field microscope, and the findings can be directly correlated with specific morphological details of the tumor. However, CISH signals may be more difficult to be visualized and individually enumerated, especially if one is dealing with low levels of gene amplification. The final decision on which technique to use will depend on the comfort level and familiarity of the performing laboratory with each one of them.

CPM spans ∼113 kb of genomic DNA, contains 11 exons, and encodes for a membrane-bound zinc-dependent protease that cleaves C-terminal basic residues from peptides and proteins. CPM is located 11 kb downstream from MDM2 and the encoded protein has been implicated in many functions, such as adipose tissue differentiation, osteogenesis, inflammation, and coagulation.23 CPM is expressed at high levels in many tissues and cell types, including the pulmonary membranes, placenta, and kidney.23 Gene expression profiling has shown CPM transcriptional upregulation in response to specific external stimuli, which may be used as an indicator of body's response to infection or disease.23 Recently, CPM has been found to be upregulated in soft tissue sarcomas, such as synovial sarcoma, gastrointestinal stromal tumors, and dedifferentiated/pleomorphic liposarcomas.23 However, whether CPM has an important oncogenic role in those tumors or in liposarcomas is still unknown.

In summary, we have found that CPM is consistently amplified in well-differentiated liposarcoma/atypical lipomatous tumors but not in ordinary or pleomorphic lipomas. These results offer another avenue for the investigation of the biology of well-differentiated liposarcomas and suggest that CPM could be used as an alternative and novel diagnostic tool for these tumors.