Soft tissue sarcomas: evaluation of biomarkers and clinical outcomes

Background: Soft tissue sarcomas (STS) are a rare, heterogeneous and complex group of tumors of mesenchymal origin. A great proportion of patients (pts) with high-risk STS eventually develop metastatic disease, and pts with advanced disease have a median overall survival (OS) of about 12 months [Judson I et al, Lancet Oncol 2014]. STS have a tendency to metastasize to lungs, but may also relapse in other distant organs. Systemic chemotherapy (CT) comprising anthracycline therapy remains to date the standard reference regimen. Among the over 50 different histological types known, leiomyosarcoma (LMS), undifferentiated pleomorphic sarcoma (UPS) and myxofibrosarcoma (MFS) contain complex genomes characterized by a multitude of rearrangements, amplifications, and deletions. Only few diagnostic and prognostic markers exist, and the accurate diagnosis and prediction of the clinical behaviour of many of these tumors remain a challenge. Some biomarkers, specifically markers related to Epithelial to Mesenchymal transition (EMT) and its reverse process (MET) and microRNAs (miRNAs) may be useful to identify a possible signature with prognostic and diagnostic value, and could also elucidate new possible pathogenic mechanisms. The aims of this study were to describe efficacy and toxicity of CT for advanced STS in a cohort of unselected pts treated at Istituto Oncologico Veneto (IOV), and to evaluate the significance and the prognostic value of the expression of markers linked to EMT/MET processes, related miRNAs and myo-miRNAs in tumors samples. Patients and methods: Medical records of pts with advanced STS treated at Istituto Oncologico Veneto from January 2010 to December 2015 were reviewed and clinical data retrieved. Vital status was recorded as of September 30th 2016. OS was estimated with Kaplan-Meier method, and univariate analysis for OS was performed with Cox regression. Tumor tissues from pts affected by STS referred to the Istituto Oncologico Veneto, Padova, were either retrieved from the Tissue Biobank of the Department of Surgical, Oncological and Gastroenterological Sciences, University of Padova, or freshly received as tru-cut biopsy or surgical specimen. All the specimens were reviewed by an expert pathologist for confirmation of representativeness of the samples. Samples were collected from November 2014 (date of Ethical Approval) to September 2016. Eligible histological types were LMS, UPS, and MFS. Total RNA, enriched in low molecular weight molecules, was extracted using NucleoSpin miRNA columns (MN GmbH & Co, Germany) from frozen or fresh tissue samples. Markers linked to EMT/MET were analyzed by qRT-PCR using primers specific for epithelial markers (E-cadherin and ZO-1), and mesenchymal markers (Snail, Slug, Vimentin, Zeb−1, Zeb−2, and N-cadherin). Expression of alpha-SMA was also evaluated along with Periostin. The expression of specific miRNAs linked to EMT (miR-100-5p-5p), to MET (miR200b-3p, miR30b-5p and miR30c-5p) and myo-miRNAs (miR1, miR133a-3p andmiR133b) was measured. The study was approved by the institutional Ethics Committee and conducted in accordance with the Helsinki Declaration and Good Clinical Practice guidelines. Written informed consent was obtained from all pts. Results: Clinical evaluation of pts referred to Medical Oncology Unit 1 – Istituto Oncologico Veneto, IOV - IRCCS. Overall 405 pts were eligible, of whom 51.4% had advanced or metastatic disease. One hundred-sixty seven patients were taken in charge at Istituto Oncologico Veneto. Median age was 61 years (range 16-89); 87 patients (52.1%) were female, and 54 (32.3%) were aged ≥ 70 years. Out of these, 37 patients (19.8%) did not receive chemotherapy. The prevalent histological types were LPS (24%), LMS (19.8%) and UPS (11.4%). Most pts had ECOG PS 0-1 (71.8%). Globally, median OS was 17.7 months (95%CI 13.59-21.75) and median PFS was 10 months (95%CI 7.9-12.0). Patients receiving only first-line CT were 57 (34.1%); 43 patients (25.7%) received more than two lines of CT. First-line CT regimens were anthracycline-based in 81 patients (62.3%). Median OS of patients treated with CT was 19.9 months (95%CI 16.4–23.5), while that of untreated patients was 3.3 months (95%CI 1.3-5.2), p<0.001. At univariate analysis ECOG PS, anemia and lymphopenia were associated with prognosis. Patients with PS 0 had better median OS compared to those with PS 1 and PS 2-3, respectively 23.5 months (95%CI 18.5–28.5), 15.8 months (95%CI 18.5–28.5), 6.6 months (95%CI 1.0-12.3), p<0,001. Baseline low lymphocyte count was associated with worse survival, with median OS of12.5 months (95%CI 5.4-19.5) compared to patients with neutrophil/lymphocyte ratio <3, who had median OS of 22.8 months (95%CI 15.2-30.4), p=0.005. Baseline anemia was also associated with worse survival, with median OS for anemic patients (Hb<12 g/dL) being 9 months (95%CI 3.2-14.8) and median OS for non-anemic patients being 20.9 months (95%CI 16.6-25.4). No difference in OS was seen according to age, gender, chemotherapy regimen, BMI, albumin levels, or LDH levels. Grade ≥3 toxicities occurred in 59 patients (45.4%), and 23 patients (17.7%) required hospitalization for toxicity management. Molecular analysis on samples Overall, 55 STS samples were fully characterized, specifically 28 LMS, of which 10 low/intermediate grade (LG) and 18 high grade (HG) LMS; 13 myxoid sarcomas (of which 8 MFS/ and 5 low-grade FMS); and 14 UPS. Correlation of EMT-related markers and miRNA with histological type and grade was assessed. The samples were analysed for the expression of E-cadherin and ZO-1, as epithelial markers, and of Slug, Vimentin, Snail, ZEB-1, ZEB-2, N-cadherin, and Periostin, as markers related to a mesenchymal status. E-cadherin expression was not found in all analysed STS. Alpha-SMA was significantly expressed in LMS compared to UPS and MFS/FMS (p<0.001), and not significantly different in UPS compared to MFS/FMS. ZEB-1 and ZEB-2 expression was significantly higher in LMS compared to MFS/FMS (p<0.001) and UPS (p=0.001 for ZEB-1, p=0.003 for ZEB-2), whereas no differential expression was measured between UPS and MFS/FMS. Interestingly, ZEB-1 and ZEB-2 were differentially expressed in HG and LG LMS (p=0.038 for ZEB-1, p=0.048 for ZEB-2). Also N-cadherin expression was significantly higher in LMS compared to MFS/FMS (p=0.006) and UPS (p=0.028), whereas no differential expression was measured according to the grading. As for periostin, this was found to be higher in LMS compared to MFS/SFM (p=0.002), and in UPS compared to MFS/FMS (p=0.005); no difference was observed between LMS and UPS. miR-1, miR-133a-3p and miR-133b (“myo-miRNAs”) expression was found to be significantly higher in LMS compared to MFS and UPS (p=0.002), though no difference was observed between HG and LG LMS. All other analysed miRNAs did not show a different expression in the three histological subtypes, nor it was different according to grade., with the exception of miR-100-5p, which was found to be significantly over-expressed in LMS compared to MFS/FMS (p=0.02). An inverse correlation between Slug and myo-miRNAs expression was observed. Also, a direct correlation between ZEB family members, and an inverse correlation between ZEB-1 and miR-200b was observed. In univariate analysis, high ZEB-1 (≥0.4) was correlated with worse OS (2.3 months, 95%CI 0.9-3.4) vs low ZEB-1 (8.6 months, 1.5 – n.r.), p=0.058. Similarly, high ZEB-2 (≥0.9) was correlated with worse OS (2.2 months, 95%CI 0.9-32.7) vs low ZEB-2 (8.6 months, 95%CI 1.5 – n.r.), p=0.052. High Periostin was also correlated with worse OS (2.2 months, 95%CI 1.2-2.5) vs low Periostin (8.6 months, 95%CI 1.3 – n.r.), p=0.028. In multivariate analysis, grade and periostin and ZEB-1 levels confirmed to be associated with overall survival. All the other analysed EMT-related markers, as well as all analysed miRNAs, were not correlated with OS nor with PFS. Conclusions: Our clinical data confirm that there is a benefit for pts that have been treated for advanced disease compared to those not receiving active treatment. There remains a need for novel effective therapies in metastatic STS, particularly for pts with certain chemo-resistant subtypes. The analysis on tumor samples highlighted that a “myo-miRNA” signature may serve as potential confirmatory markers in LMS samples with difficult/controversial histological findings Moreover, some biomarkers linked to the mesenchymal phenotype (i.e. ZEB-1, ZEB-2, and Periostin) may have prognostic value. In light of the findings from this study, we have planned to proceed with validation of such results in a larger sample and to provide a correlation with immunoistochemical staining. Also, we are planning to verify whether circulating Periostin and N-cadherin may be correlated with outcomes and response to therapy in STS.