Abstract
Background
Undifferentiated pleomorphic sarcomas (UPS) present a diagnostic and therapeutic challenge. Identification of prognostic molecular markers is required for the discovery of novel treatment approaches. The aim of this study was to correlate clinicopathologic variables, expression of tyrosine kinase receptors, and markers of cell cycle progression and survival with oncologic outcomes.
Methods
A tissue microarray containing 208 primary UPS samples was analyzed by immunohistochemistry for protein markers and in situ hybridization for microRNA. Staining results were correlated with clinicopathologic features and oncologic outcomes. Univariate and multivariate analyses were conducted to assess associations between expression of protein markers, mi-RNA and outcome.
Results
At a median follow-up of 3.9 years (9 years for survivors), 5-year disease-specific survival (DSS) was 63%. Clinical variables associated with improved DSS included age < 61 years, tumor size < 10 cm, margin-negative resection and sporadic-tumor status. At the protein level, loss of cyclin D1 (p=0.06), pEGFR (p=0.023), pIGF-1R (p=0.022), and PTEN (p<0.001) and overexpression of AXL (p=0.015) were associated with reduced DSS on univariate analysis. Ki67, PCNA and pEGFR were more highly expressed in sporadic UPS than radiation-associated (RA-UPS) while RA-UPS samples expressed higher levels of both phosphorylated and total IGF-1R.
Discussion
Cyclin D1, AXL and PTEN are associated with cancer-specific outcomes and warrant further investigation in UPS. The differences in protein expression in sporadic versus RA-UPS may indicate that the activated molecular signaling nodes may be different for each specific histology and could also explain the aggressive phenotype seen in RA-UPS when compared to the sporadic lesions.
Keywords: Undifferentiated pleomorphic sarcoma, UPS, MFH, Radiation
Introduction
Soft tissue sarcomas comprise a large, heterogeneous group of rare tumors, originating from varied mesenchymal cell types. Undifferentiated pleomorphic sarcoma (UPS) presents a unique diagnostic challenge, as there is no clear origin of histogenesis. These tumors are largely morphologically high grade with a complex karyotype.1
Margin-negative surgical resection remains the mainstay of treatment. Despite aggressive surgery, local recurrences occur in 13–42% of patients and nearly 31–35% will develop metastatic disease.2–4 Although the addition of chemotherapy has shown some benefit in meta-analyses, these responses are marginal at best, highlighting the need for improved therapeutic options for UPS patients.5,6 The use of radiation therapy is associated with reduced local recurrences in high grade sarcomas; however, approximately 3–5% of UPS arise in a prior site of therapeutic radiation for an unrelated malignancy and are termed radiation-associated UPS (RA-UPS), making radiation therapy controversial in this subset of the patients due to the comorbidities of re-irradiation 7,8. Therefore, a greater understanding of the molecular biology of UPS, both sporadic and radiation-associated, is necessary to facilitate the discovery of novel treatment approaches in the future.
Advanced age, large tumors, and radiation-associated tumors have been associated with reduced survival in patients with UPS.4 While recent studies have described genomic rearrangements and gene expression profiles of UPS tumors, the expression patterns and prognostic potential of specific proteins and microRNAs (miRNAs) have not yet been extensively investigated. 9–12 Previously, Tomita et al 13 found that high levels of activated (phosphorylated) AKT (pAKT), a key step in the phosphatidylinositiol-3-kinase/mammalian target of rapamycin (PI3K/mTOR) pathway, correlated with poorer overall (OS) and disease-specific survival (DSS) outcomes. Similarly, in previous studies, we found that a subset of UPS tumors expressed elevated pAKT levels which was associated with poorer DSS outcomes.14
Recent evidence suggests that miRNA-mediated gene regulation interconnects with the AKT pathway, forming an AKT–microRNA (miRNA) regulatory network. In this network, the miRNA and PI3K/AKT/mTOR pathway work together to exert their cellular functions. Thus, to better understand these interactions we evaluated some miRNAs that have been associated to the PI3K/AKT/mTOR network and have been found to be relevant prognostic indicators of survival in other sarcomas15 and UPS16–19.
Therefore, in the current study, the protein expression levels of the PI3K/mTOR pathway components were examined, including downstream effectors and upstream activating receptor tyrosine kinases (RTKs) as well as miRNAs for their association with patient survival outcomes. Furthermore, we evaluated the potential prognostic associations of several miRNAs indicated to be differentially expressed in UPS.
Methods
Clinical database and tissue microarray construction
With the approval of the Institutional Review Board (IRB) of The University of Texas MD Anderson Cancer Center (UTMDACC), the UTMDACC Pathology Archive was searched for formalin-fixed, paraffin embedded (FFPE) sporadic and radiation-associated UPS specimens. UPS diagnoses were confirmed by a UTMDACC sarcoma pathologist and in the context of multidisciplinary tumor board review. A tissue microarray (TMA) consisting of 168 sporadic and 60 RA-UPS tumor samples was constructed as previously described20, with duplicate 1.2 mm cores included from each tissue block. The TMA was used for protein and miRNA analysis.14,21 MDM2 immunohistochemistry or fluorescence in situ hybridization was performed on all samples prior to inclusion in this study to exclude liposarcoma histology. A clinical database was populated by collecting clinicopathologic variables including patient age and gender; tumor size, grade, and location; and the use of chemo- or radiotherapy. Local recurrence was considered as any recurrence at the primary site without concomitant metastasis. For the outcomes analysis, only primary tumor tissue with sufficient clinical history were included (total cohort: n = 208; RA-UPS: n = 35; sporadic UPS: n = 173). Therefore, nine (4%) primary samples lacking adequate corresponding clinical history, 8 (4%) recurrent samples, and 3 (1%) metastatic samples were excluded.
Immunohistochemistry and in situ hybridization
Immunohistochemical (IHC) studies were performed using commercially available antibodies, following standard automated and manual protocols. Immunostaining for pAKT, AKT, pS6RP, S6RP, and p4EBP1 was performed by the histology core facility at The Virginia Harris Cockrell Cancer Research Center at Science Park histology core. Immunohistochemistry for Ki67, PCNA, Cyclin D1, CD31, p53, c-KIT, pEGFR, IGF-1R, PTEN, PDGFRα, and PDGFRβ was performed by the UTMDACC Research Histopathology Facility. 14 In situ hybridization was performed for miR-1, miR-133a, miR-182 and miR-183 by the RNA Center at UTMDACC. 22
Immunostaining for pIGF-1R, pMEK, MEK, pMET, and AXL was conducted in our laboratory as follows. Tumor specimens were deparaffinized in xylene and rehydrated using a graded ethanol series. Sections were subjected to antigen retrieval at 100°C for 45 minutes in Tris-EDTA, pH 8 and endogenous peroxidase blocking in 3% H2O2 in PBS for 12 minutes. Primary antibody incubation took place at 4°C overnight using pIGF-1R rabbit monoclonal antibody #3024 diluted 1:25, pMEK rabbit monoclonal antibody #2338 diluted 1:50, MEK rabbit polyclonal primary antibody #9122 diluted 1:50, pMET rabbit monoclonal antibody #3077 diluted 1:100, or AXL rabbit monoclonal antibody #4566 diluted 1:100 (Cell Signaling Technology). Primary antibody was visualized using the 4plus two-step HRP detection system (Biocare Medical) and diaminobenzidine. Sections were counterstained with hematoxylin and mounted with Permount (Fisher).
Scoring was performed by 2 independent sarcoma pathologists (GAA and WLW). The percentage of tumor cells expressing the marker was recorded and an intensity score was assigned (0 = no expression; 1 = low expression; 2 = intermediate expression; 3 = high expression). Cutoffs for survival analysis stratification were assigned based on median expression tendency. A negative or positive expression classification was assigned to markers (Ki67, Cyclin D1, p53, AXL, c-kit, pEGFR, IGF-1R, PTEN, pS6RP, and p4EBP1) with low median expression (0–30% of cells positive for stain); samples were considered positive if ≥ 10% of cells were stained. A negative/low or high expression classification was assigned to markers (PCNA, CD31, pIGF-1R, pMET, MET, pAKT, AKT, S6RP, PDGFRa, PDGFRb, pMEK, MEK, miR-1, miR-133a, miR-183, and miR-182) with moderate to high median expression (40–100% of cells positive for stain); samples were classified as negative/low expression (≤60% of cells positive for stain) or high expression (>60% of cells positive). Studies have shown that PI3K pathway-associated kinases such as AKT translocate within the cell and that subcellular localization often dictates protein function adding complexity to the functional repertoire of the PI3K pathway23; therefore, we make a distinction between nuclear and cytoplasmic staining.
Statistics
Fisher’s exact test was used to differentiate protein and mi-RNA expression between sporadic-UPS and RA-UPS. Univariate Kaplan and Meier methods were used to estimate overall survival (OS), disease-specific survival (DSS), local recurrence-free survival (LRFS), and metastasis-free survival (MFS). OS and DSS time were measured from date of pathological confirmation of UPS to date of death owing to any cause or date of sarcoma-specific death, respectively. LRFS and MFS time was measured from the date of surgical resection to time of pathological or radiographic determination of recurrent or metastatic disease, respectively. Multivariate Cox proportional hazard regression models were constructed to assess associations between marker expression, clinical variables, and survival outcomes. Variables with a p value of < 0.1 in univariate analysis were fitted to multivariate models. A two-tailed p value of < 0.05 was accepted as statistically significant. All analyses were performed using SPSS version 22.0 (IBM Corp., Armonk, NY).
Results
UPS patient and tumor characteristics
Patient characteristics and tumor variables corresponding to the 208 primary UPS samples (patients, n = 208) are summarized in Table 1. Median age at diagnosis was 64 years and 63% of the cohort was comprised of males. Primary sites of origin included the extremities (73%), trunk or retroperitoneum (24%), and head and neck (3%). The median tumor size was 7.5 cm and the majority of tumors were high grade (76%). Sixty-five (31%) patients were treated with systemic chemotherapy and 118 (57%) received radiation therapy. As we have previously reported differences in clinical outcomes among patients with sporadic versus RA-UPS4, we also analyzed the current data within the context of disease setting (Table 1).
Table 1.
Clinicopathologic Characteristics of Patients with Undifferentiated Pleomorphic Sarcoma by Radiation-Associated Status
| Variable | Entire Cohort | Sporadic | RAS |
|---|---|---|---|
| Entire Cohort | 208 (100.0%) | 173 (83.2%) | 35 (16.8%) |
| Gender | |||
| Male | 131 (62.9%) | 114 (65.9%) | 17 (48.6%) |
| Female | 77 (37.0%) | 59 (34.1%) | 18 (51.4%) |
| Age | |||
| Median (range) | 64.0 (17–87) | 65.0 (22–87) | 62.0 (17–87) |
| Tumor size, cm | |||
| Median (range) | 7.5 (0.9–35.0) | 8.0 (1.0–35.0) | 6.0 (0.9–30.0) |
| Grade | |||
| Low | 6 (2.8%) | 4 (2.3%) | 2 (5.7%) |
| Intermediate | 38 (18.2%) | 38 (21.9%) | 0 (0.0%) |
| High | 157 (75.5%) | 130 (75.1%) | 27 (77.1%) |
| Unknown | 7 (3.4%) | 1 (0.6%) | 6 (17.1%) |
| Location | |||
| Head & Neck | 6 (2.9%) | 0 (0.0%) | 6 (17.1%) |
| Trunk/Retroperitoneum | 50 (24.0%) | 30 (17.3%) | 20 (57.1%) |
| Extremity | 152 (73.1%) | 143 (82.7%) | 9 (25.7%) |
| Depth | |||
| Superficial | 28 (13.5%) | 25 (14.5%) | 3 (8.6%) |
| Deep to investing fascia | 180 (86.5%) | 148 (85.5%) | 32 (91.4%) |
| Therapy received | |||
| Chemotherapy | 65 (31.2%) | 57 (32.9%) | 8 (22.9%) |
| Radiation | 118 (56.7%) | 99 (57.2%) | 19 (54.3%) |
| Resection Margins | |||
| R0 | 145 (69.7%) | 119 (68.8%) | 26 (74.3%) |
| R1 | 57 (27.4%) | 51 (29.5%) | 6 (17.1%) |
| R2 | 3 (1.4%) | 1 (0.6%) | 2 (5.7%) |
| Unknown | 3 (1.4%) | 2 (1.2%) | 1 (2.9%) |
Protein and miRNA expression in UPS and association with survival outcomes
After a median follow-up period of 3.9 years (range, 0.1–20.75 years; for survivors: 9 years, range 0.2–20.7), local recurrence following resection was observed in 27% (n = 56) of patients. Distant metastasis occurred in 63 (30%) patients. Median time to local and distant recurrence was 0.9 years (range, 0.1–13.7 years) and 0.8 years (range, 0.1–16.3 years), respectively. The local recurrence rate was higher in RA-UPS compared to sporadic UPS (43% versus 24%, respectively; p = 0.019). The metastasis rate was not significantly different between RA-UPS and sporadic UPS (34% versus 29%, respectively; p = 0.353). Univariate and multivariate analysis of clinical and protein markers associated with reduced risk of local recurrence included sporadic tumors, margin-negative resection and retained PTEN expression (Table 2). Loss of PTEN and tumors > 10cm were associated with increased distant recurrence on univariate and multivariate analysis (Table 2).
Table 2.
Univariate and Multivariate Analysis for Overall and Disease-specific Survival in Patients with Undifferentiated Pleomorphic Sarcoma
| Overall Survival | Disease-specific survival | ||||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| Covariate, (indicator ) | Univariable | Multivariable (n = 123)† | Univariable | Multivariable (n = 129) † | |||||
|
| |||||||||
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | ||
| Clinical markers | Age, (<61) v. ≥61 | 0.36 (0.23–0.57) | <0.001* | 0.17 (0.08–0.38) | <0.001* | 0.41 (0.24–0.67) | 0.001* | 0.22 (0.08–0.55) | 0.001* |
| Sex, male v. female | 1.13 (0.74–1.69) | 0.555 | 1.21 (0.75–1.98) | 0.430 | |||||
| Setting, sporadic v. RAS | 0.44 (0.28–0.69) | <0.001* | 0.36 (0.16–0.79) | 0.011* | 0.37 (0.22–0.63) | <0.001* | 0.19 (0.08–0.46) | <0.001* | |
| Size, <10 cm v. ≥10 cm | 0.51 (0.34–0.76) | 0.001* | 0.45 (0.24–0.84) | 0.012* | 0.48 (0.30–0.77) | 0.002* | 0.33 (0.16–0.71) | 0.004* | |
| Depth, superficial v. deep | 0.78 (0.43–1.42) | 0.415 | 0.63 (0.29–1.84) | 0.253 | |||||
| Grade, low-intermediate v. high | 0.22 (0.31–1.58) | 0.132 | 0.29 (0.04–2.06) | 0.285 | |||||
| Chemotherapy, no v. yes | 0.92 (0.59–1.45) | 0.732 | 0.93 (0.55–1.54) | 0.763 | |||||
| Radiation, no v. yes | 1.04 (0.70–1.54) | 0.830 | 1.04 (0.59–1.69) | 0.851 | |||||
| Margins, negative v. positive | 0.61 (0.40–0.91) | 0.016* | 0.24 (0.24–0.84) | <0.001* | 0.62 (0.38–1.00) | 0.051* | 0.26 (0.11–0.62) | 0.002* | |
|
| |||||||||
| Protein markers | Ki67, negative v. positive | 0.75 (0.48–1.19) | 0.228 | 0.90 (0.52–1.56) | 0.904 | ||||
| Cyclin D1, negative v. positive | 1.40 (0.89–2.21) | 0.145 | 1.73 (0.98–3.05) | 0.060** | 2.16 (0.96–4.84) | 0.063 | |||
| CD31, negative-low v. high | 1.26 (0.51–3.14) | 0.617 | 1.10 (0.39–3.07) | 0.850 | |||||
| p53, negative v. positive | 0.81 (0.52–1.26) | 0.350 | 0.86 (0.51–1.45) | 0.562 | |||||
| AXL Receptor, negative v. positive | 0.66 (0.40–1.06) | 0.085** | 0.91 (0.46–1.83) | 0.797 | 0.46 (0.25–0.86) | 0.015* | 0.50 (0.21–1.19) | 0.119 | |
| c-kit, negative v. positive | 1.38 (0.66–2.87) | 0.391 | 1.43 (0.57–3.58) | 0.447 | |||||
| pEGFR, negative v. positive | 0.69 (0.35–1.35) | 0.278 | 1.94 (1.09–3.43) | 0.023* | 0.58 (0.26–1.30) | 0.186 | |||
| pIGF1R, negative-low v. high | 0.92 (0.47–1.78) | 0.804 | 0.49 (0.27–0.90) | 0.022* | 0.59 (0.27–1.33) | 0.208 | |||
| IGF1R, negative v. positive | 0.67 (0.41–1.07) | 0.091* | 0.87 (0.42–1.79) | 0.706 | 0.65 (0.37–1.13) | 0.130 | |||
| pMET, negative-low v. high | 0.97 (0.38–2.51) | 0.955 | 0.88 (0.30–2.55) | 0.814 | |||||
| PTEN, negative v retained | 7.70 (3.47–17.1) | <0.001* | 6.73 (2.48–18.3) | <0.001* | 10.3 (4.55–23.4) | <0.001* | 20.8 (6.34–68.2) | <0.001* | |
| pAKT (Cytoplasmic), negative-low v. high | 0.83 (0.53–1.32) | 0.434 | 0.74 (0.43–1.26) | 0.261 | |||||
| pAKT (Nuclear), negative-low v. high | 0.86 (0.55–1.35) | 0.520 | 1.05 (0.62–1.78) | 0.854 | |||||
| AKT (Cytoplasmic), negative-low v. high | 1.04 (0.65–1.67) | 0.861 | 1.15 (0.64–2.05) | 0.629 | |||||
| AKT (Nuclear), negative-low v. high | 0.61 (0.39–0.96) | 0.034* | 0.78 (0.39–1.54) | 0.468 | 0.83 (0.47–1.45) | 0.827 | |||
| pS6RP, negative v positive | 1.23 (0.77–1.96) | 0.380 | 1.58 (0.88–2.86) | 0.128 | |||||
| S6RP, negative-low v. high | 0.78 (0.46–1.35) | 0.379 | 0.67 (0.35–1.31) | 0.241 | |||||
| P4EBP1 (Cytoplasmic), negative v positive | 0.97 (0.62–1.50) | 0.879 | 1.04 (0.61–1.76) | 0.890 | |||||
|
| |||||||||
| miRNA markers | miR-1, negative-low v. high | 0.83 (0.47–1.49) | 0.537 | 0.76 (0.38–1.52) | 0.437 | ||||
| miR-133a, negative-low v. high | 0.69 (0.39–1.23) | 0.213 | 0.82 (0.39–1.69) | 0.585 | |||||
| miR-183, negative-low v. high | 0.79 (0.45–1.37) | 0.397 | 0.89 (0.46–1.74) | 0.736 | |||||
| miR-182, negative-low v. high | 0.89 (0.42–1.89) | 0.759 | 0.92 (0.36–2.38) | 0.863 | |||||
Significant at level of p < 0.05
Significant at level of p <0.1
Patients without expression scores for all molecular variables included in the multivariate analysis are excluded
Five-year overall survival was 54.7% and disease-specific survival was 63% (data not shown). Univariate analysis of clinical factors for DSS resulted in the identification of age < 61 years at diagnosis, tumor size < 10 cm, margin-negative resection and sporadic tumors were associated with improved DSS (Table 3; Fig. 1 A–C). Expression of cytoplasmic AXL and loss of PTEN were also associated with reduced DSS [hazard ratio (HR) = 0.46 95% confidence interval (CI), 0.25–0.86; p= 0.015 and HR =10.3; CI, 4.55–23.4; p= 0.015, respectively] (Fig. 1 D–F). Although not statistically significant, we observed that tumors expressing high levels of cyclin D were associated with a better survival (5-year DSS 78% versus 58%, p = 0.060). When evaluated in a multivariate model, only age < 61 years at diagnosis, tumor size < 10 cm, sporadic tumors, and retained PTEN were associated with improved DSS (Table 3). Previously, we evaluated miRNA expression in several UPS cell lines and found that expression of miR-1 and miR-133a was the most significantly decreased while the expression of miR-182 and miR-183 was strongly increased when compared to the human mesenchymal stem cell control (data not shown). However, in this study we did not observe any prognostic significance for the four miRNAs.
Table 3.
Univariate and Multivariate Analysis for Local Recurrence-free and Metastasis-free Survival in Patients with Undifferentiated Pleomorphic Sarcoma
| Local recurrence-free survival | Metastasis-free survival | ||||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| Covariate, (indicator ) | Univariable | Multivariable (n = 145)† | Univariable | Multivariable (n = 175)† | |||||
|
| |||||||||
| HR (95%CI) | P-value | HR (95%CI) | P-value | HR (95%CI) | P-value | HR (95%CI) | P-value | ||
| Clinical markers | Age, (<61) v. ≥61 | 0.78 (0.46–1.34) | 0.372 | 0.59 (0.35–1.00) | 0.051** | 0.53 (0.28–1.00) | 0.052 | ||
| Setting, sporadic v. RAS | 0.38 (0.21–0.69) | 0.002* | 0.29 (0.15–0.56) | <0.001 | 0.75 (0.40–1.41) | 0.378 | |||
| Size, <10 cm v. ≥10 cm | 0.68 (0.39–1.2) | 0.149 | 0.41 (0.25–0.68) | 0.001* | 0.36 (0.19–0.66) | 0.001 | |||
| Depth, superficial v. deep | 0.81 (0.37–1.79) | 0.600 | 0.61 (0.26–1.42) | 0.250 | |||||
| Grade, low-intermediate v. high | 0.91 (0.22–3.77) | 0.898 | 0.31 (0.04–2.24) | 0.245 | |||||
| Chemotherapy, no v. yes | 0.87 (0.49–1.52) | 0.619 | 1.18 (0.67–2.05) | 0.569 | |||||
| Radiation, no v. yes | 0.78 (0.46–1.35) | 0.782 | 1.23 (0.75–2.02) | 0.409 | |||||
| Margins, negative v. positive | 0.51 (0.29–0.87) | 0.013* | 0.45 (0.24–0.87) | 0.018 | 0.71 (0.42–1.18) | 0.187 | |||
|
| |||||||||
| Protein markers | Ki67, negative v. positive | 1.25 (0.67–2.35) | 0.480 | 0.88 (0.49–1.55) | 0.653 | ||||
| Cyclin D1, negative v. positive | 0.97 (0.53–1.79) | 0.928 | 1.19 (0.67–2.11) | 0.549 | |||||
| CD31, negative-low v. high | 1.19 (0.37–3.87) | 0.773 | 0.62 (0.24–1.58) | 0.317 | |||||
| p53, negative v. positive | 0.86 (0.48–1.56) | 0.624 | 1.02 (0.58–1.81) | 0.936 | |||||
| AXL Receptor, negative v. positive | 0.71 (0.37–1.37) | 0.309 | 0.58 (0.31–1.11) | 0.100 | |||||
| c-kit, negative v. positive | 1.17 (0.46–2.98) | 0.741 | 1.12 (0.45–2.84) | 0.803 | |||||
| pEGFR, negative v. positive | 0.63 (0.32–1.20) | 0.161 | 1.23 (0.64–2.39) | 0.534 | |||||
| pIGF1R, negative-low v. high | 0.60 (0.32–1.15) | 0.126 | 0.86 (0.43–1.71) | 0.660 | |||||
| IGF1R, negative v. positive | 0.57 (0.30 1.09) | 0.092** | 0.72 (0.37–1.39) | 0.329 | 0.80 (0.43–1.49) | 0.482 | |||
| pMET, negative-low v. high | 1.66 (0.39–7.12) | 0.492 | 3.00 (0.39–22.8) | 0.287 | |||||
| PTEN, negative v retained | 6.25 (2.18–17.9) | 0.001* | 5.29 (1.77–15.9) | 0.003 | 8.68 (3.33–22.7) | <0.001* | 7.98 (2.89–22.0) | <0.001 | |
| pAKT (Cytoplasmic), negative-low v. high | 0.89 (0.48–1.68) | 0.732 | 0.64 (0.36–1.11) | 0.636 | |||||
| pAKT (Nuclear), negative-low v. high | 0.85 (0.46–1.56) | 0.603 | 0.96 (0.55–1.67) | 0.878 | |||||
| AKT (Cytoplasmic), negative-low v. high | 1.56 (0.80–3.04) | 0.192 | 1.00 (0.56–1.81) | 0.982 | |||||
| AKT (Nuclear), negative-low v. high | 0.91 (0.48–1.70) | 0.909 | 0.76 (0.43–1.36) | 0.357 | |||||
| pS6RP, negative v positive | 1.54 (0.80–2.95) | 0.193 | 1.24 (0.69–2.26) | 0.472 | |||||
| S6RP, negative-low v. high | 0.77 (0.37–1.60) | 0.479 | 0.83 (0.42–1.62) | 0.577 | |||||
| P4EBP1 (Cytoplasmic), negative v positive | 0.92 (0.51–1.69) | 0.798 | 1.04 (0.59–1.82) | 0.883 | |||||
|
| |||||||||
| miRNA markers | miR-1, negative-low v. high | 1.87 (0.83–4.21) | 0.132 | 0.79 (0.36–1.75) | 0.572 | ||||
| miR-133a, negative-low v. high | 0.75 (0.33–1.66) | 0.472 | 1.12 (0.46–2.69) | 0.804 | |||||
| miR-183, negative-low v. high | 1.14 (0.52–2.49) | 0.740 | 0.65 (0.31–1.35) | 0.246 | |||||
| miR-182, negative-low v. high | 1.00 (0.34–2.95) | 0.992 | 2.36 (0.56–9.93) | 0.242 | |||||
Significant at level of p < 0.05
Significant at level of p <0.1
Patients without expression scores for all molecular variables included in the multivariate analysis are excluded
Figure 1.
Univariable Kaplan-Meier survival curves of disease-specific survival stratified by (A) age at diagnosis (p = 0.001; HR 0.41, 95% CI 0.24–0.67), (B) disease setting (p < 0.001; HR 0.37, 95% CI 0.22–0.63), (C) index tumor size (p = 0.002; HR 0.48, 95% CI 0.30–0.77), (D) cyclin D1 expression (p = 0.060; HR 1.73, 95% CI 0.98–3.05), (E) AXL expression (p =0.015; HR 0.46, 95% CI 0.25–0.86), and (F) PTEN expression (p <0.001; HR 10.3, 95% CI 4.55–23.4).
Although RA-UPS is associated with poorer survival outcomes than sporadic UPS4, currently no molecular distinctions between the two disease settings have been identified. Therefore, we examined protein and miRNA expression patterns separately. Both Ki67 and PCNA, markers of cell proliferation, were more highly expressed in sporadic UPS than RA-UPS (Table 4). In addition, more sporadic UPS samples were positive for pEGFR expression, while RA-UPS samples expressed higher levels of both phosphorylated and total IGF-1R when compared to the sporadic lesions.
Table 4.
Protein and miRNA Expression by Radiation-Associated Status.
| Variable | Overall Median (range) | Sporadic | Radiation-Associated | P-value | ||||
|---|---|---|---|---|---|---|---|---|
|
| ||||||||
| N | Negative | Positive⋄ | N | Negative | Positive | |||
| Ki67 | 10.0 (0.0–80.0) | 131 | 39.7% | 60.3% | 29 | 72.4% | 27.6% | 0.001 |
| Cyclin D1 | 0.0 (0.0–50.0) | 137 | 58.4% | 41.6% | 30 | 50.0% | 50.0% | 0.261 |
| p53 | 10.0 (0.0–90.0) | 137 | 47.4% | 52.6% | 39 | 37.9% | 62.1% | 0.234 |
| AXL (Cytoplasmic) | 10.0 (0.0–90.0) | 114 | 50.0% | 50.0% | 22 | 36.4% | 63.6% | 0.174 |
| c-kit | 0.0 (0.0–70.0) | 141 | 87.2% | 12.8% | 31 | 93.5% | 6.5% | 0.257 |
| pEGFR | 50.0 (0.0–100.0) | 131 | 19.1% | 80.9% | 31 | 54.8% | 45.2% | <0.001 |
| IGF1R | 50.0 (0.0–100.0) | 120 | 75.0% | 25.0% | 30 | 46.7% | 53.3% | 0.003* |
| PTEN | 2 (0–3)* | 144 | 3.5% | 96.5% | 31 | 6.5% | 93.5% | 0.360 |
| pS6RP | 10.0 (0.0–90.0) | 142 | 63.4% | 36.6% | 31 | 61.3% | 38.7% | 0.490 |
| p4EBP1 (Cytoplasmic) | 5.0 (0.0–90.0) | 140 | 57.1% | 42.9% | 29 | 48.3% | 51.7% | 0.251 |
| N | Negative-low | High** | N | Negative-low | High | |||
| PCNA† | 90.0 (0.0–100.0) | 128 | 1.6% | 98.4% | 29 | 24.1% | 75.9% | <0.001 |
| CD31 | 2 (1–3)* | 132 | 90.9% | 9.1% | 24 | 100.0% | 0.0% | 0.124 |
| pIGF-1R | 50.0 (0–100.0) | 86 | 70.9% | 29.1% | 30 | 20.0% | 80.0% | <0.001 |
| pMET | 30.0 (0.0–70.0) | 35 | 91.4% | 8.6% | 29 | 79.3% | 20.7% | 0.152 |
| MET† | 90.0 (40.0–100.0) | 116 | 0.9% | 99.1% | 29 | 0.0% | 100% | 0.800 |
| pAKT (Cytoplasmic) | 50.0 (0.0–90.0) | 134 | 66.4% | 33.6% | 29 | 51.7% | 48.3% | 0.101 |
| pAKT (Nuclear) | 60.0 (0.0–100.0) | 134 | 46.3% | 53.7% | 29 | 48.3% | 51.7% | 0.502 |
| AKT (Cytoplasmic) | 50.0 (0.0–90.0) | 127 | 63.8% | 36.2% | 30 | 56.7% | 43.3% | 0.301 |
| AKT (Nuclear) | 50.0 (0.0–90.0) | 127 | 67.7% | 32.3% | 30 | 50.0% | 50.0% | 0.055 |
| S6RP | 70.0 (0.0–100.0) | 139 | 25.9% | 74.1% | 31 | 16.1% | 83.9% | 0.181 |
| PDGFRa† | 90.0 (40.0–100.0 | 144 | 0.7% | 99.3% | 31 | 3.2% | 96.8% | 0.324 |
| PDGFRb† | 100.0 (60.0–100.0) | 139 | 0.0% | 100% | 30 | 0.0% | 100% | 1.000 |
| pMEK† | 100.0 (30.0–100.0) | 89 | 3.4% | 96.6% | 30 | 3.3% | 96.7% | 0.737 |
| MEK† | 100.0 (0.0–100.0) | 98 | 4.1% | 95.9% | 29 | 0.0% | 100% | 0.350 |
| miR-1 | 50.0 (0.0–90.0) | 64 | 56.3% | 43.8% | 23 | 52.2% | 47.8% | 0.462 |
| miR-133a | 40.0 (0.0–90.0 | 64 | 67.2% | 32.8% | 25 | 64.0% | 36.0% | 0.481 |
| miR-183 | 40.0 (0.0–90.0) | 74 | 66.2% | 33.8% | 25 | 64.0% | 36.0% | 0.512 |
| miR-182 | 30.0 (0.0–90.0) | 62 | 87.1% | 12.9% | 29 | 72.4% | 27.6% | 0.080 |
Intensity score
Not evaluated in survival analysis due to homogeneously high expression.
Samples were considered positive if there was labeling present in ≥10% of cells.
Samples were considered to have high expression if there was labeling in >60% of cells.
Discussion
There is no clear origin of histogenesis for UPS; therefore, this histological entity represents a unique diagnostic and therapeutic challenge. Surgical resection with negative margins remains the mainstay of treatment for these tumors1,24 Subtype-specific benefit of neoadjuvant or adjuvant therapies has not yet been demonstrated. Therefore, it is critical to not only identify novel targets for therapeutic interventions but evaluate potential biomarkers to predict response to therapy and survival outcomes.4–6,24,25 Previously, we have demonstrated that radiation-associated disease, tumor size, and advanced age are negative clinical prognosticators for UPS.4 However, little is known about the molecular dysregulations in UPS. In the current study, we evaluate the expression profiles of numerous protein biomarkers including components of the PI3K/mTOR and MAPK pathways as well as upstream receptor tyrosine kinases (RTKs) and specific miRNAs in a relatively large cohort of sporadic and radiation-associated UPS tumors. Although immunohistochemistry is typically used to rule out other high grade sarcomas, few studies have specifically examined whether there is a difference between sporadic and radiation-associated lesions. We identified cyclin D1, cytoplasmic AXL and PTEN as potential markers of prognosis for patients with UPS. In addition, different protein expression profiles between sporadic and radiation-associated UPS.
Although RA-UPS is associated with worse clinical outcomes, the molecular mechanisms governing these differences are currently unknown.4,26 In this study, we demonstrate differences in protein expression between sporadic and RA-UPS. Positive pEGFR expression was more readily detected in sporadic UPS, while phosphorylated and total IGF-1R were more highly expressed in RA-UPS. While preclinical data have supported the development of anti-IGF-1R therapies, the majority of clinical trials evaluating these therapeutic interventions in large, unselected patient populations have reported low efficacy.27,28 The identification and validation of biomarkers, such as pIGF-1R and other members of the IGF pathway, suggests that there may be subgroups that would derive clinical benefit from anti-IGF-1R therapy. The tumor suppressor PTEN, a negative regulator of the PI3K/mTOR pathway, is frequently inactivated in human cancer.29,30 In addition, loss of function is associated with increased pathway activity, advanced disease, and worse patient outcome in numerous cancer types.29 PTEN was the only biomarker significantly associated with recurrence and survival outcomes in our cohort. Previous studies have described decreased PTEN expression in pleomorphic sarcomas, including UPS, as a result of the loss of chromosome 10q in some samples.9,31 However, only 4% of patients demonstrated loss of PTEN suggesting that activation of the AKT/mTOR pathway occurs independently of PTEN protein loss in the majority of this subset of tumors.
Oncogenic PI3K/mTOR pathway activation can also occur through the inappropriate or unregulated activation of upstream RTKs. In our study, we found overexpression of the RTK, AXL. Others have shown than activation of AXL can promote multiple tumorigenic responses including cell survival, proliferation, migration, and adhesion.32–34 Overexpression of AXL has been associated with worse prognosis in many cancer types, including Ewing’s and osteosarcoma.32,35–37 Anti-AXL monoclonal antibodies have shown promise in preclinical models of Ewing’s sarcoma 33,35 and one such therapy is currently being investigated in a phase II trial in lung cancer patients.38 In the current study, overexpression of cytoplasmic AXL was associated with worse DSS (5-yr: 75% versus 50%, p = 0.015) in univariate analysis, which supports the further consideration of anti-AXL therapy in UPS patients.
There are several limitations to the current study, namely the retrospective nature of the analysis, variable use of chemotherapy and radiation therapy, and the limited number of UPS samples available for evaluation. However, to our knowledge, this is one of the largest UPS-centric retrospective studies examining potential biomarkers to date. Here we confirmed that the clinical variables of age, tumor size, negative margins, and radiation-associated disease indicate increased recurrence and poor survival outcomes. Additionally, we demonstrated the potential of PTEN loss associated with worsened patient outcomes. Furthermore, we propose that pIGF-1R expression may differentiate disease setting and may be a useful therapeutic target in UPS.
In conclusion, cyclin D1, AXL and PTEN are associated with cancer-specific outcomes and warrant further investigation in UPS. Differences in protein expression in sporadic versus RA-UPS may explain biologic differences in oncologic outcomes.
Synopsis.
Sporadic and radiation-associated undifferentiated pleomorphic sarcomas (UPS) demonstrate differences in protein expression. PTEN status is associated with disease-specific survival in UPS.
Acknowledgments
Funding for this research was provided in part by a NIH/NCI K08CA160443 (to KET), the Sally M. Kingsbury Sarcoma Research Foundation (to KET), the Jay Vernon Jackson Fund (to KET), Michael and Mary Kay Poulos (to KET), the Marty Lindley Foundation (supporting KLW and DRI), and the Amschwand Sarcoma Cancer Foundation (supporting CDM). None of the funding organizations had a role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
Footnotes
Disclosures: We have no conflicts of interest to declare. Data in this report was previously presented as an oral presentation at the SSO 2015 Annual Cancer Symposium, Houston, TX on March 26, 2015.
References
- 1.WHO. Classification of Tumors of Soft Tissue and Bone. 2013. [Google Scholar]
- 2.Delisca GO, Mesko NW, Alamanda VK, et al. MFH and high-grade undifferentiated pleomorphic sarcoma-what’s in a name? Journal of surgical oncology. 2015 Feb;111(2):173–177. doi: 10.1002/jso.23787. [DOI] [PubMed] [Google Scholar]
- 3.Mullen JT, Kobayashi W, Wang JJ, et al. Long-term follow-up of patients treated with neoadjuvant chemotherapy and radiotherapy for large, extremity soft tissue sarcomas. Cancer. 2012 Aug 1;118(15):3758–3765. doi: 10.1002/cncr.26696. [DOI] [PubMed] [Google Scholar]
- 4.Dineen SP, Roland CL, Feig R, et al. Radiation-Associated Undifferentiated Pleomorphic Sarcoma is Associated with Worse Clinical Outcomes than Sporadic Lesions. Annals of surgical oncology. 2015 Mar 6; doi: 10.1245/s10434-015-4453-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Pervaiz N, Colterjohn N, Farrokhyar F, Tozer R, Figueredo A, Ghert M. A systematic meta-analysis of randomized controlled trials of adjuvant chemotherapy for localized resectable soft-tissue sarcoma. Cancer. 2008 Aug 1;113(3):573–581. doi: 10.1002/cncr.23592. [DOI] [PubMed] [Google Scholar]
- 6.Adjuvant chemotherapy for localised resectable soft-tissue sarcoma of adults: meta-analysis of individual data. Sarcoma Meta-analysis Collaboration. Lancet. 1997 Dec 6;350(9092):1647–1654. [PubMed] [Google Scholar]
- 7.Beane JD, Yang JC, White D, Steinberg SM, Rosenberg SA, Rudloff U. Efficacy of adjuvant radiation therapy in the treatment of soft tissue sarcoma of the extremity: 20-year follow-up of a randomized prospective trial. Annals of surgical oncology. 2014 Aug;21(8):2484–2489. doi: 10.1245/s10434-014-3732-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Yang JC, Chang AE, Baker AR, et al. Randomized prospective study of the benefit of adjuvant radiation therapy in the treatment of soft tissue sarcomas of the extremity. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 1998 Jan;16(1):197–203. doi: 10.1200/JCO.1998.16.1.197. [DOI] [PubMed] [Google Scholar]
- 9.Silveira SM, Villacis RA, Marchi FA, et al. Genomic signatures predict poor outcome in undifferentiated pleomorphic sarcomas and leiomyosarcomas. PloS one. 2013;8(6):e67643. doi: 10.1371/journal.pone.0067643. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Villacis RA, Silveira SM, Barros-Filho MC, et al. Gene expression profiling in leiomyosarcomas and undifferentiated pleomorphic sarcomas: SRC as a new diagnostic marker. PloS one. 2014;9(7):e102281. doi: 10.1371/journal.pone.0102281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Takahashi A, Nakayama R, Ishibashi N, et al. Analysis of gene expression profiles of soft tissue sarcoma using a combination of knowledge-based filtering with integration of multiple statistics. PloS one. 2014;9(9):e106801. doi: 10.1371/journal.pone.0106801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ruping K, Altendorf-Hofmann A, Chen Y, et al. High IGF2 and FGFR3 are associated with tumour progression in undifferentiated pleomorphic sarcomas, but EGFR and FGFR3 mutations are a rare event. Journal of cancer research and clinical oncology. 2014 Aug;140(8):1315–1322. doi: 10.1007/s00432-014-1700-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tomita Y, Morooka T, Hoshida Y, et al. Prognostic significance of activated AKT expression in soft-tissue sarcoma. Clinical cancer research : an official journal of the American Association for Cancer Research. 2006 May 15;12(10):3070–3077. doi: 10.1158/1078-0432.CCR-05-1732. [DOI] [PubMed] [Google Scholar]
- 14.Lahat G, Zhang P, Zhu QS, et al. The expression of c-Met pathway components in unclassified pleomorphic sarcoma/malignant fibrous histiocytoma (UPS/MFH): a tissue microarray study. Histopathology. 2011 Sep;59(3):556–561. doi: 10.1111/j.1365-2559.2011.03946.x. [DOI] [PubMed] [Google Scholar]
- 15.Nakatani F, Ferracin M, Manara MC, et al. miR-34a predicts survival of Ewing’s sarcoma patients and directly influences cell chemo-sensitivity and malignancy. The Journal of pathology. 2012 Apr;226(5):796–805. doi: 10.1002/path.3007. [DOI] [PubMed] [Google Scholar]
- 16.Sarver AL, Li L, Subramanian S. MicroRNA miR-183 functions as an oncogene by targeting the transcription factor EGR1 and promoting tumor cell migration. Cancer research. 2010 Dec 1;70(23):9570–9580. doi: 10.1158/0008-5472.CAN-10-2074. [DOI] [PubMed] [Google Scholar]
- 17.Yu H, Lu Y, Li Z, Wang Q. microRNA-133: expression, function and therapeutic potential in muscle diseases and cancer. Current drug targets. 2014;15(9):817–828. doi: 10.2174/1389450115666140627104151. [DOI] [PubMed] [Google Scholar]
- 18.Sachdeva M, Mito JK, Lee CL, et al. MicroRNA-182 drives metastasis of primary sarcomas by targeting multiple genes. The Journal of clinical investigation. 2014 Oct;124(10):4305–4319. doi: 10.1172/JCI77116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wong P, Hui A, Su J, et al. Prognostic microRNAs modulate the RHO adhesion pathway: A potential therapeutic target in undifferentiated pleomorphic sarcomas. Oncotarget. 2015 Apr 23; doi: 10.18632/oncotarget.3926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lusby K, Savannah KB, Demicco EG, et al. Uterine Leiomyosarcoma Management, Outcome, and Associated Molecular Biomarkers: A Single Institution’s Experience. Annals of surgical oncology. 2013 Jan 20; doi: 10.1245/s10434-012-2834-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lahat G, Tuvin D, Wei C, et al. Molecular prognosticators of complex karyotype soft tissue sarcoma outcome: a tissue microarray-based study. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 2010 May;21(5):1112–1120. doi: 10.1093/annonc/mdp459. [DOI] [PubMed] [Google Scholar]
- 22.Jorgensen S, Baker A, Moller S, Nielsen BS. Robust one-day in situ hybridization protocol for detection of microRNAs in paraffin samples using LNA probes. Methods. 2010 Dec;52(4):375–381. doi: 10.1016/j.ymeth.2010.07.002. [DOI] [PubMed] [Google Scholar]
- 23.Davis WJ, Lehmann PZ, Li W. Nuclear PI3K signaling in cell growth and tumorigenesis. Frontiers in cell and developmental biology. 2015;3:24. doi: 10.3389/fcell.2015.00024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Brennan MF, Antonescu CR, Maki RG. Management of Soft Tissue Sarcoma. New York: Springer; 2013. [Google Scholar]
- 25.Lehnhardt M, Daigeler A, Homann HH, et al. MFH revisited: outcome after surgical treatment of undifferentiated pleomorphic or not otherwise specified (NOS) sarcomas of the extremities -- an analysis of 140 patients. Langenbeck’s archives of surgery / Deutsche Gesellschaft fur Chirurgie. 2009 Mar;394(2):313–320. doi: 10.1007/s00423-008-0368-5. [DOI] [PubMed] [Google Scholar]
- 26.Gladdy RA, Qin LX, Moraco N, et al. Do radiation-associated soft tissue sarcomas have the same prognosis as sporadic soft tissue sarcomas? Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2010 Apr 20;28(12):2064–2069. doi: 10.1200/JCO.2009.25.1728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yee D. A tale of two receptors: insulin and insulin-like growth factor signaling in cancer. Clinical cancer research : an official journal of the American Association for Cancer Research. 2015 Feb 15;21(4):667–669. doi: 10.1158/1078-0432.CCR-14-2056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Guha M. Anticancer IGF1R classes take more knocks. Nature reviews. Drug discovery. 2013 Apr;12(4):250. doi: 10.1038/nrd3992. [DOI] [PubMed] [Google Scholar]
- 29.Chu. PTEN regulatory functions in tumor suppression and cell biology. Chu; 2004. [PubMed] [Google Scholar]
- 30.Niemeyer BF, Parrish JK, Spoelstra NS, Joyal T, Richer JK, Jedlicka P. Variable expression of PIK3R3 and PTEN in Ewing Sarcoma impacts oncogenic phenotypes. PloS one. 2015;10(1):e0116895. doi: 10.1371/journal.pone.0116895. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Gibault L, Perot G, Chibon F, et al. New insights in sarcoma oncogenesis: a comprehensive analysis of a large series of 160 soft tissue sarcomas with complex genomics. The Journal of pathology. 2011 Jan;223(1):64–71. doi: 10.1002/path.2787. [DOI] [PubMed] [Google Scholar]
- 32.Han J, Tian R, Yong B, et al. Gas6/Axl mediates tumor cell apoptosis, migration and invasion and predicts the clinical outcome of osteosarcoma patients. Biochemical and biophysical research communications. 2013 Jun 7;435(3):493–500. doi: 10.1016/j.bbrc.2013.05.019. [DOI] [PubMed] [Google Scholar]
- 33.Liu R, Gong M, Li X, et al. Induction, regulation, and biologic function of Axl receptor tyrosine kinase in Kaposi sarcoma. Blood. 2010 Jul 15;116(2):297–305. doi: 10.1182/blood-2009-12-257154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hutterer M, Knyazev P, Abate A, et al. Axl and growth arrest-specific gene 6 are frequently overexpressed in human gliomas and predict poor prognosis in patients with glioblastoma multiforme. Clinical cancer research : an official journal of the American Association for Cancer Research. 2008 Jan 1;14(1):130–138. doi: 10.1158/1078-0432.CCR-07-0862. [DOI] [PubMed] [Google Scholar]
- 35.Fleuren ED, Hillebrandt-Roeffen MH, Flucke UE, et al. The role of AXL and the in vitro activity of the receptor tyrosine kinase inhibitor BGB324 in Ewing sarcoma. Oncotarget. 2014 Dec 30;5(24):12753–12768. doi: 10.18632/oncotarget.2648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Lee CH, Yen CY, Liu SY, et al. Axl is a prognostic marker in oral squamous cell carcinoma. Annals of surgical oncology. 2012 Jul;19( Suppl 3):S500–508. doi: 10.1245/s10434-011-1985-8. [DOI] [PubMed] [Google Scholar]
- 37.Chen PX, Li QY, Yang Z. Axl and prostasin are biomarkers for prognosis of ovarian adenocarcinoma. Annals of diagnostic pathology. 2013 Oct;17(5):425–429. doi: 10.1016/j.anndiagpath.2013.01.005. [DOI] [PubMed] [Google Scholar]
- 38.Sheridan C. First Axl inhibitor enters clinical trials. Nature biotechnology. 2013 Sep;31(9):775–776. doi: 10.1038/nbt0913-775a. [DOI] [PubMed] [Google Scholar]