Context.—MDM2 is known to be abnormally upregulated in a variety of human neoplasms, secondary to gene amplification. Assessment of MDM2 amplification is most useful clinically for separating lipomas (nonamplified) from atypical lipomatous neoplasms or well-differentiated liposarcomas (amplified). MDM2 amplification occurs in approximately 7% of all human neoplasms. In this study, we sought to determine the utility of MDM2 amplification for the separation of benign (schwannomas) and malignant peripheral nerve sheath tumors (MPNSTs). The expression of p53 was correlated with MDM2 amplification because early studies have indicated that MDM2 is rarely amplified in MPNSTs that express p53.
Objectives.—To determine the percentage of MPNSTs with MDM2 amplification and the specificity of MDM2 amplification for malignancy in nerve sheath tumors.
Design.—Fifteen MPNSTs, 14 neurofibromas, and 15 schwannomas were obtained from the files of the Department of Pathology. These cases underwent fluorescent in situ hybridization analysis for the presence of MDM2 amplification. Assessments were also made for cellularity (low or high), percentage of cells staining positively for p53 and MDM2 protein, and percentage of cells staining with MIB-1.
Results.—Of 15 MPNSTs, 3 (20%) demonstrated amplification of the MDM2 gene. No neurofibromas or schwannomas demonstrated MDM2 amplification. All 3 MDM2-amplified MPNSTs were positive for p53. Correlation of MDM2 amplification status and p53 immunoreactivity was statistically significant (P = .004).
Conclusions.—The low frequency (20%) of MDM2 amplification in our series of MPNSTs demonstrates that MDM2 fluorescent in situ hybridization has limited diagnostic value for the separation of benign schwannomas and MPNSTs. Our study demonstrated a positive correlation (P = .004) between MDM2 amplification and p53 expression.
The MDM2 oncogene has been shown to be upregulated in a number of human neoplasms. This upregulation is the result of a variety of mechanisms, including gene amplification, increased transcription, and increased translation.1 MDM2 upregulation has several implications for both diagnostic surgical pathology and for understanding basic oncogenic processes in a number of human malignancies. Upregulation of MDM2 can result in p53 inactivation and represents a separate mechanism in addition to p53 mutation for inhibition of p53 function. Such p53 inactivation can lead to uncontrolled cell proliferation and tumor growth.
Although MDM2 gene amplification has been shown to occur in up to 7% of all human tumors, MDM2 in soft tissue sarcomas is more often upregulated (20%).1 ,MDM2 amplification and overexpression have been most diagnostically useful in separating lipomas (MDM2 nonamplified) from well-differentiated liposarcomas/atypical lipomatous tumors (MDM2 amplified).2,3 Prior reports on the percentage of malignant peripheral nerve sheath tumors (MPNST) showing MDM2 overexpression or amplification have documented variable findings.4–9 Although Kindblom et al4 did not find MDM2 amplification in their series of p53-positive MPNSTs, Flørenes et al5 reported that 50% of MPNSTs were amplified for MDM2. Immunohistochemical studies of MDM2 protein expression in MPNSTs indicate variable findings as well. Szadowska et al6 reported that 28.1% of MPNSTs were positive for MDM2 by immunohistochemistry, whereas Watanabe et al7 reported that 67.3% of MPNSTs were immunohistochemically positive for MDM2. However, neither of these latter 2 studies directly investigated the amplification status of MDM2 in MPNSTs.
The p53 and MDM2 proteins function in the same cell proliferation pathway and have directly antagonistic functions. Thus, some authors1 have hypothesized that neoplasms would be either p53 mutant or MDM2 amplified. Most immunohistochemical data for p53 and MDM2 protein expression support this premise, with only 7.9% of tested soft tissue sarcomas expressing both p53 and MDM2.6 In a study of 115 soft tissue sarcomas, the p53−/MDM2+ phenotype predominated (20% of cases) with only 1.8% of tumors being p53+/MDM2−. Because p53 expression correlates with high histologic grade and proliferation index, p53 mutation was assumed to be a late event.7 ,MDM2 amplification was assumed to be a relatively early event because it did not correlate with either grade or proliferation index.7
Given the variable reports of MDM2 amplification and p53 dysregulation in MPNSTs, we studied a series of 15 MPNSTs, 14 neurofibromas, and 15 control schwannomas. The percentage of MPNSTs demonstrating MDM2 amplification was determined and correlated with p53 protein expression, MIB-1 proliferation index, and tumor cellularity. Herein, we report the results of that study.
MATERIALS AND METHODS
Sources of Tissue
The surgical pathology files at the University of Utah (Salt Lake City) were searched for the period 2000 to 2010 for MPNST cases. Fifteen cases were selected and reviewed to confirm the original diagnoses. All cases of MPNST selected were characterized by at least 1 of the 3 following criteria: (1) arose grossly in association with a nerve, (2) arose in a patient with neurofibromatosis, or (3) was immunohistochemically positive for S100 protein and protein gene product (PGP) 9.5. In addition, 15 schwannoma cases and 14 neurofibromas were also selected for study purposes. The use of de-identified, human tissue for this study was approved by the University of Utah Institutional Review Board (IRB 24487).
Immunohistochemistry
Immunohistochemical staining for p53, MDM2, and Ki-67 was performed with antibodies (p53, clone DO-7, Ki-67, clone Mib-1) from Dako (Carpinteria, California) and clone SMP14 from Abcam (Cambridge, Massachusetts) for MDM2. Automated immunohistochemical staining was performed on 4-µm-thick sections of formalin-fixed, paraffin-embedded tissues. Sections were air-dried and then melted in a 60°C oven for 30 minutes. Slides were loaded onto the Ventana (Tucson, Arizona) ULTRA automated staining instrument where deparaffinization with EZ Prep solution and antigen-retrieval with CC1 buffer (Ventana; Ki-67, 64 minutes at 100°C; p53, 36 minutes at 95°C) were performed. Primary antibody was applied at 37°C for 1 hour (Ki-67, 1∶100), 80 minutes (MDM2, 1∶50), or 24 minutes (p53, 1∶500). Nonspecific staining was decreased, and signal amplification was performed with the Avidin/Biotin Blocking and Amplification kits (Ventana), respectively. The slides were detected using the IVIEW DAB (3-3′-diaminobenzidine) detection kit (Ventana), which is a biotinylated goat anti-mouse/anti-rabbit secondary antibody, streptavidin–horseradish-peroxidase system, with 3-3′-diaminobenzidine as the chromogen. Tissue sections were counterstained with hematoxylin for 8 minutes, removed from the immunostainer, and placed in a deionized water–Dawn (dishwashing liquid soap) mixture. Sections were gently washed in this mixture to remove any coverslip oil applied by the automated instrument. The slides were then gently rinsed in deionized water until all of the wash mixture was removed. Sections were placed in iodine for 30 seconds to remove any precipitates from fixation and dipped in sodium thiosulfate to clear the iodine. Slides were dehydrated in graded alcohols (70%, 95%, and 100%), cleared in xylene, and then covered with a coverslip. Results from the p53 immunohistochemistry were defined as high-level staining at 20% or greater. High MIB-1 staining was defined as greater than 15% of nuclei staining.
MDM2 Fluorescent In Situ Hybridization
MDM2 gene amplification was detected by fluorescent in situ hybridization (FISH) with the Repeat-Free Poseidon MDM2 (12q15) and SE12 probes from Veridex (Raritan, New Jersey). The manufacturer's recommended procedures for formalin-fixed, paraffin-embedded tissue were followed, with the exception that the posthybridization washes used 2× standard saline citrate with 0.3% Igepal (Sigma-Aldrich Co., St Louis, Missouri). The FISH-positive cases were defined by a MDM2 to SE12 ratio greater than 2.0.
Statistics
The Fisher exact test was used to correlate MDM2 gene amplification with p53 immunohistochemistry, MIB-1 immunohistochemistry, and tumor cellularity.
RESULTS
Table 1 lists the sites of origin for the 15 MPNSTs and their relationship with neurofibromatosis. No cases of MPNST appeared to arise in association with a neurofibroma. Four MPNSTs were low grade (all nonamplified), and 11 were high grade. The 15 MPNSTs, 15 schwannomas, and 14 neurofibromas were evaluated for MDM2 gene amplification by FISH. Three of 15 MPNSTs (20%) demonstrated MDM2 amplification (Table 1). None of the neurofibromas and schwannomas, including a cellular schwannoma, was characterized by MDM2 amplification. Figure 1, A and B, shows tumors with MDM2 amplification (MDM2 to SE12 ratio = 16.5) and nonamplification (MDM2 to SE12 ratio = 1.0), respectively. Immunohistochemistry demonstrated strong MDM2 expression in 4 cases and weak expression in a fifth (Table 1). One case strongly expressing MDM2 protein was not MDM2 amplified. The weakly MDM2 overexpressing case was also not amplified. The remaining 3 MDM2 overexpressing MPNSTs were amplified for MDM2. No schwannomas or neurofibromas overexpressed MDM2.
Malignant peripheral nerve sheath tumors, demonstrating amplification (A) and nonamplification (B) of the MDM2 gene (fluorescent in situ hybridization, original magnifications ×1000).
Figure 2. Malignant peripheral nerve sheath tumors with low (A) or high (B) expression of p53 protein (immunohistochemistry, original magnifications ×200).
Figure 3. Malignant peripheral nerve sheath tumors with low (A) or high (B) MIB-1 proliferation index (immunohistochemistry, original magnifications ×200).
Malignant peripheral nerve sheath tumors, demonstrating amplification (A) and nonamplification (B) of the MDM2 gene (fluorescent in situ hybridization, original magnifications ×1000).
Figure 2. Malignant peripheral nerve sheath tumors with low (A) or high (B) expression of p53 protein (immunohistochemistry, original magnifications ×200).
Figure 3. Malignant peripheral nerve sheath tumors with low (A) or high (B) MIB-1 proliferation index (immunohistochemistry, original magnifications ×200).
High positivity for p53 protein expression (>20% of cells positive) was observed in 4 of 15 MPNSTs (27%) (Table 1). MDM2 amplification correlated with p53 protein expression (P = .004) (Table 2) because all 3 MDM2-amplified MPNSTs (100%) showed high expression of p53 protein (55%, 36%, 23%). Figure 2, A and B, shows examples of MPNSTs with low and high expression of p53 protein, respectively. Most MPNSTs were either MDM2−/p53− (73%) or MDM2+/p53+ (20%) (Table 3). No examples of MDM2+/p53− MPNSTs were present in our series. None of the tested schwannomas demonstrated high p53 protein expression (data not shown).
The MIB-1 immunohistochemistry revealed high levels (>15%) of MIB-1 staining in 10 of 15 MPNSTs (66.7%) (Table 1). Figure 3, A and B, shows examples of MPNSTs with low and high MIB-1 proliferation indexes, respectively. All 3 MDM2-amplified MPNSTs (100%) were associated with a high proliferation rate but this did not reach statistical significance (P = .37) (Table 2). Of 12 MPNSTs characterized by MDM2 nonamplification, 7 (58.3%) had a high MIB-1 proliferation index (Table 3). Of those 7 MDM2−/MIB-1+ cases, only 1 (14%) had high p53 protein expression. With the exception of the cellular schwannoma, none of the schwannomas demonstrated a high MIB-1 proliferation index (data not shown).
Tumor cellularity was high in 11 of 15 MPNSTs (73%) (Table 1). All 3 MDM2-amplified MPNSTs (100%) had high cellularity, but the association was not statistically significant (P = .34) (Table 2). Of the 11 MPNST cases with high cellularity, 10 (91%) had a high MIB-1 proliferation index, and 4 (36%) were p53+.
COMMENT
MDM2 amplification occurs in approximately 20% of soft tissue tumors.1 Flørenes et al5 found 50% of MPNST to be amplified for MDM2, but Tanas et al9 failed to find MDM2 amplified in the 7 MPNSTs they studied. Because MDM2 amplification and protein expression occur predominately in sarcomas, rather than benign mesenchymal neoplasms, demonstration of either MDM2 gene amplification or protein expression has been used to separate some well-differentiated sarcomas from their benign counterparts.2,3 Given that MDM2 amplification has been reported in up to 50% of MPNSTs,6 we sought to determine whether MDM2 FISH might be diagnostically useful in separating benign nerve sheath tumors, including neurofibromas and cellular schwannomas, from MPNSTs. In our series of 15 MPNSTs, 14 neurofibromas, and 15 schwannomas, 3 MPNSTs (20%) and no neurofibromas or schwannomas demonstrated MDM2 amplification by FISH. All 3 of the MDM2-amplified cases (100%) also overexpressed MDM2 as did 2 nonamplified MPNSTs. Interestingly, both MDM2-expressing but nonamplified MPNSTs demonstrated low MIB-1 and p53 levels. The clinical utility and diagnostic value of MDM2 amplification testing for the separation of benign and malignant nerve sheath tumors appears limited by the relatively low percentage of MPNSTs that are amplified. Even with the higher percentage of MDM2 amplification reported by others (50%),5 a negative result would not exclude the diagnosis of MPNST.
MDM2 acts as an antagonist to p53 mainly by targeting it for degradation. When MDM2 is overexpressed by amplification or another mechanism, the function of p53 is inhibited and cell proliferation is increased. Thus, some authors1 have suggested that tumors showing MDM2 amplification would not also demonstrate p53 mutations or p53 protein overexpression. Immunohistochemical data for p53 and MDM2 protein expression support this hypothesis, with only 7.9% of tested soft tissue sarcomas expressing both p53 and MDM2 protein.6 In a separate study by Kindblom et al,4 none (0%) of the p53-protein–positive tumors tested were found to be MDM2 amplified. In that study,4 only 2 of 26 MPNSTs (8%) were analyzed for MDM2 amplification, and the analysis was performed by Southern blot methodology limiting its validity. Our findings are in contrast with those prior studies. All 3 of our MDM2-amplified cases (100%) expressed high levels of p53 protein. The correlation between MDM2 gene amplification and p53 expression was statistically significant (P = .004). As would be expected, those 3 MDM2-amplified cases also demonstrated high MIB-1 proliferation indices (45%, 26%, and 41%) and high cellularity, although the associations did not reach statistical significance. Thus, p53 overexpression can coexist with MDM2 amplification.
Prior studies have shown that the most common combination of p53 and MDM2 reactivity in soft tissue sarcomas is p53−/MDM2+ (20%).6 Our findings are different, with no MPNSTs showing the pattern of p53−/MDM2+. Our most common pattern of expression (73%) was p53−/MDM2−. Of those 11 cases of p53−/MDM2−, 6 (55%) had a high MIB-1 proliferation index (Table 3). Cellular proliferation in these cases is likely driven by a mechanism that does not involve p53 or MDM2.
The pattern of p53+/MDM2− appears to be rare in soft tissue sarcomas because it was reported in only 1.8% of tumors in the Szadowska et al6 study. Our results are similar with only 1 MPNST case (7%) demonstrating the pattern of p53+/MDM2−. That 1 case also had a high MIB-1 proliferation index (70%). Inhibition of p53 in this case is likely due to p53 gene mutation, which results in excess expression of a dysfunctional p53 protein.
In our study, only 20% of MPNSTs demonstrated MDM2 amplification, limiting the value of MDM2 amplification testing for the separation of benign from malignant nerve sheath tumors. Unlike prior studies, all our MDM2-amplified cases highly expressed p53 protein. Thus, the presence of MDM2 amplification does not exclude the possibility of p53 gene mutation in the same neoplasm, and p53 protein expression does not exclude MDM2 amplification.