Targeting abundant survivin expression in liposarcoma: subtype dependent therapy responses to YM155 treatment

The study was approved by the local ethics committee at the Medical Faculty of the Heinrich-Heine-University Duesseldorf, Germany (institutional board review no. 3821), and carried out in accordance with good clinical practice and the Declaration of Helsinki (World Medical Association 2013). Each of the patients underwent tumour resection for primary LPS at the University Hospital Duesseldorf, Germany, in between 2001 and 2014. The tumours were staged and graded by pathologists according to the 8th edition of the TNM-classification recommended by the International Union Against Cancer (UICC) and World Health Organization (WHO). Tumours in which pathological staging was based on older TNM editions where re-classified according to the 8th edition. Follow-up data was collected until all experiments had been completed and overall survival was determined as the period from the date of surgery until the date of the last follow up or death of any cause. Patients with stage I-IV disease independently of the tumour localisation, neoadjuvant therapy and microscopic resection margin status who received surgery with curative intend were included in this study. Patients who had received only palliative chemotherapy after histological confirmation of LPS, who had deceased perioperatively within 30 days after surgery, or who had been lost to follow up were excluded from the study. Clinicopathological details were collected retrospectively from original pathological reports and patient case files. Follow-up data were retrieved before reviewing the experimental results.

Five tissue microarrays (TMA) were constructed from formalin-fixed and paraffin-embedded specimens comprising 49 samples of primary liposarcoma complemented by 15 lipomas, 13 samples from normal fatty tissue, and samples from other organs serving as positive controls (Packeisen et al. 2003), all of which originated from the Institute of Pathology, University Hospital Duesseldorf, Germany, where all cases had been reviewed by board-certified pathologists. A TMA contained two cylindrical specimens for each tumour sample from a donor block, five extracts from lipoma samples, five cylinders from normal fatty tissue, and three specimens deriving from other tissue types.

After preparing tissue sections with a thickness of 4 µm from the TMA, immunohistochemical staining was performed using the ZytoChem Plus HRP-DAB Kit (Zytomed Systems, Berlin, Germany) as described previously (Werner et al. 2016). In brief, after deparaffinisation and rehydration, epitope demasking was carried out at 95°C for 30ʹ using a 3% trisodium citrate dihydrate buffer equilibrated at pH 6.0, followed by cooling for 20’ to room temperature. Incubation of the tissue sections in 3% H₂O₂ phosphate-buffered saline (PBS, pH 7.4) for 10’ blocked endogenous peroxidase before the slides were rinsed three times for 2’ in PBS with 0.1% Tween-20 (Sigma-Aldrich, St. Louis, MO, USA). After blocking reagent was added to the sections for 10’ to block unspecific binding sites minimising background staining, the slides were washed in PBS with 0.1% Tween-20. Incubation with the rabbit primary polyclonal anti-survivin antibody (NB500-201; 1:750 dilution; Novus, Littleton, CO, USA), took 60’ at room temperature. Isotype controls with rabbit immunoglobulin fraction (Code X0903; 1:1,000 dilution; Dako, Glostrup, Denmark) served as negative controls. After triple rinsing the slides in PBS with 0.1% Tween-20, the sections were incubated with biotinylated secondary antibody and streptavidin-HRP conjugate, before 3,30-diaminobenzidine high contrast was added for 10’ in darkness resulting in epitope visualisation. Finally, tissue sections were counterstained with Mayer’s haematoxylin. Human colonic and tonsillar tissue specimens had been positively pretested for survivin expression served as positive controls.

Survivin staining intensities and percentage of chromogen positive cells were scored by two independent investigators according to the immunoreactivity score (IRS) reported by Remmele and Stegner (1987) without knowledge of histopathological parameters or patient survival outcome. Both investigators were experienced in the visual assessment and evaluation of the IRS. Differing ratings resulted in a re-examination of the respective samples by both investigators until a consensual scoring was reached.

While the cell line Lipo-DUE1 was cultivated in RPMI 1640 Medium GlutaMax™ as previously described (Mersch et al. 2016), DDLPS cell line Lipo246A and PLS cell line PLS-1 kindly provided by Dina Lev (MTA No. MT2012-10,265) were maintained in DMEM 1× GlutaMax™ (both obtained from Gibco Life Technologies, Carlsbad, CA, USA). Both media were supplemented with 10% heat inactivated bovine FCS (Gibco Life Technologies, Carlsbad, CA, USA), penicillin, and streptomycin (both obtained from Biochrom GmbH, Berlin, Germany) to be kept in an atmosphere with 5% CO₂ at 37°C. Cells were passaged routinely within seven days at a confluence of 80% by trypsinisation with 0.05% Trypsin/EDTA (Gibco Life Technologies, Carlsbad, CA, USA) and after washing with PBS (Gibco Life Technologies, Carlsbad, CA, USA).

Total RNA was extracted from PBS washed LPS cells by RNeasy Mini Kits (Qiagen GmbH, Hilden, Germany) and concentrations were determined with the Infinite^® M200 microplate reader (Tecan Group Ltd., Mannedorf, Switzerland). After 5 µg of total RNA per sample were incubated with 0.5 µg Oligo(dT)18 Primer (Thermo Fisher Scientific, Waltham, MA, USA) for 10ʹ at 65 °C, cDNA was synthesized in 7 µl from a master mix of 0.5 µl Transcriptor Reverse Transcriptase, 4 µl Transcriptor RT Reaction Buffer 5× concentrated, 5 µl Protector RNase Inhibitor, and 2 µl dNTP Mix (Roche Diagnostics GmbH, Mannheim, Germany) for 30ʹ at 55 °C before the reverse transcriptase was inactivated for 5ʹ at 85 °C. After adjusting the cDNA concentrations to 2.5 ng/µl, triplicates from 2.5 µl of cDNA were mixed each with 12.5 µl FastStart TaqMan^® Probe Master (Roche Diagnostics GmbH, Mannheim, Germany), 0.25 µl probe solution (probe 11 and probe 60) from the Human Universal Probe Library Set (Roche Diagnostics GmbH, Mannheim, Germany) and 0.25 µl of forward (survivin: 5ʹ GCC CAG TGT TTC TTC TGC TT 3ʹ; GAPDH: 5ʹ GCC CAG TGT TTC TTC TGC TT 3ʹ) and reverse primers (survivin: 5ʹ AAC CGG ACG AAT GCT TTT TA 3ʹ; GAPDH: 5ʹ GCC CAA TAC GAC CAA ATC C 3ʹ) for quantitative real-time-PCR (survivin primers: Eurofins Scientific, Luxemburg, Luxemburg; GAPDH primers: Roche Diagnostics GmbH, Mannheim, Germany). GAPDH was used as internal reference gene. qPCR runs were conducted with the Chromo4 detector on a Dyad Disciple thermal cycler (Bio-Rad Laboratories Inc., Hercules, CA, USA) with 95 °C for 10ʹ, followed by 40 cycles of denaturation at 95 °C for 15ʹʹ and annealing and extension at 60 °C for 1ʹ. RNA expression values were calculated in relation to GAPDH and qPCR Human Reference Total RNA (Stratagene, La Jolla, CA, USA) in accordance with the 2^−ΔΔCT-method as published by Livak and Schmittgen (2001).

LPS cells were prepared for FACS analyses using the Molecular Probes FITC Annexin V/Dead Cell Apoptosis Kit (Thermo Fisher Scientific, Waltham, MA, USA). After treatment with YM155 (Selleck Chemicals LLC, Houston, TX, USA) in three different concentrations (30 nM; 100 nM; 300 nM) for 48 h, 1 × 10⁶ cells were washed in PBS and transferred to FACS vials (BD Biosciences, San Jose, CA, USA) before passing through the flow cytometry process in the BD FACSCanto™ device (BD Biosciences, San Jose, CA, USA) according to the manufacturer’s protocol. Cells were gated depending on the detected FITC-annexin and propidium iodide intensities, among which the double-positive cells were attributed to the apoptotic cell faction.

MTS assays for cell viability were analysed in 96‑well culture plates with 1 × 10⁴ LPS cells seeded per well. After 24 h of cultivation as described above, cells were treated with various compound concentrations (30 nM; 100 nM; 300 nM) of YM155 (Selleck Chemicals LLC, Houston, TX, USA), doxorubicin (AppliChem GmbH, Darmstadt, Germany), etoposide (Merck KGaA, Darmstadt, Germany), or dimethyl sulfoxide (DMSO; Gibco Life Technologies, Carlsbad, CA, USA) in a minimum of three wells for 96 h, respectively. The CellTiter 96^® AQueous Non‑Radioactive Cell Proliferation Assay (Promega Corporation, Madison, WI, USA) was used to measure cell viability. All experiments were performed in triplicates and the mean IC₅₀ was obtained based on the results of three independent experiments. For combinated treatment assays, the fractional products (FP) were determined as described by Webb (Webb 1963) with FP values < 1 representing synergistic effects, values = 1 additive effects, and values > 1 antagonistic effects.

Cell proliferation was determined by BrdU incorporation using a cell proliferation ELISA BrdU assay (Roche Diagnostics GmbH, Mannheim, Germany). Both assays were conducted according to the manufacturers’ protocols. Absorbances were measured with the Infinite^® M200 microplate reader (Tecan Group Ltd., Mannedorf, Switzerland), whereby absorbance values of YM155 treated cells were recorded as proportional to the absorbance of the corresponding DMSO treated control cells.

1 × 10⁵ cultivated cells were harvested, washed in PBS, and transferred to 25 cm² cell culture flasks for 24 h before being treated with YM155 (Selleck Chemicals LLC, Houston, TX, USA) or DMSO (Gibco Life Technologies, Carlsbad, CA, USA) for another 24 h. Then, cells were lysed in RIPA lysis buffer (Merck KGaA, Darmstadt, Germany) and incubated with protease inhibitor mix (cOmplete, Roche Diagnostics, Basel, Switzerland). Protein lysates were separated on SDS‑PAGE gels and blotted to nitrocellulose membranes (Thermo Fisher, Waltham, MA, USA) which were blocked with TBS‑T buffer containing 5% soluble nonfat dry milk (Nestlé, Vevey, Switzerland). After incubation with anti-survivin primary antibody (NB500-201; 1:1,000 dilution; Novus, Littleton, CO, USA) for 16 h at 4 °C and rinsing in TBS-T buffer, anti-rabbit IgG secondary antibody (HRP-linked Antibody #7074; 1:1,000 dilution; Cell Signalling Technology, London, UK) was added and incubated with 1.3 µl of Precision Protein™ StrepTactin-HRP (Bio-Rad Laboratories, Inc., Hercules, CA, USA) for 1 h. GAPDH as a loading control was detected by primary mouse anti-GAPDH antibody (Clone 6C5; 1:5,000 dilution; Abcam, Cambridge, UK) and marked by goat anti-mouse IgG (H&L (HRP); 1:5,000 dilution; Abcam, Cambridge, UK). Membranes were washed again in TBS-T buffer and developed with the Clarity Max™ Western ECL Substrate (Bio-Rad Laboratories, Inc., Hercules, CA, USA) and visualised with the VersaDoc Imaging System (Bio-Rad Laboratories GmbH, Munich, Germany). One representative western blot was selected for presentation after experiments were repeated thrice.

Survivin protein expression in immunohistochemical stained TMAs was assessed by immunoreactivity scores (IRS) according to Remmele and Stegner (1987) and categorized into high (IRS ≥ 3) or low (IRS < 3) levels of expression according to the median IRS for survivin expression in all investigated LPS tissue samples. Correlations between non-parametrical data sets were analysed using the paired t-test, Mann‑Whitney-U test, Kruskal-Wallis test, or Dunn-Bonferroni test as indicated.The Fisher’s exact test, Cramér’s V, or, whenever appropriate, the Chi-square test were applied for categorical data. The correlation of numerical data with clinicopathological variables were examined applying the Mann‑Whitney-U test. Kaplan‑Meier curves were compiled and analysed using the log‑rank test (Mantel-Cox). Variables with a p value < 0.05 by univariate analysis were included in a multivariate Cox regression model using a backward selection. Computed analyses were conducted employing GraphPad Prism for Windows (version 5; GraphPad Software Inc., La Jolla, CA, USA), Microsoft Excel (version 14; Microsoft Corp., Redmond, WA, USA) and SPSS statistics for Windows (version 17.0; SPSS Inc., Chicago, IL, USA). A p < 0.05 was defined to indicate a statistically significant difference. Cramér’s V values of < 0.2 were interpreted as weak correlations, values ≥ 0.2–0.5 as moderate, and values > 0.5 as strong correlations.

According to our selection criteria, 49 samples from patients with primary LPS were included in the present study after undergoing surgical resection at our department between 2001 and 2014. The baseline characteristics of the included patients are summarized in Table 1. Patients’ median overall survival (OS) time was 41 months (range 4–146 months) resulting in a 5-year OS of 52.6%. At the end of follow-up 31 (63.3%) patients were still alive with a median follow-up time of 68 months (range 6–146 months).

Assessing the clinicopathologic parameters statistically, LPS affecting the deep soft tissue in the abdomen and retroperitoneum, which was to be distinguished from LPS localised within the superficial soft tissue of the head, extremities, or thorax, correlated with LPS subtypes PLS/DDLPS (p = 0.037; Cramér’s V = 0.417), higher tumour grade (p = 0.037; Cramér’s V = 0.417) and larger primary tumour size (T3/4) (p = 0.016; Cramér’s V = 0.344).

After immunohistochemical staining, the IRS for survivin was assessed and the median IRS of 3 was defined as a cut-off to differentiate high from low expression. In contrast to the distinct LPS subtypes, survivin expression was undetectable in 13 fatty tissue samples and in 15 lipomas (Fig. 1A–F) [p < 0.001; Cramérs V = 0.655].

Whereas 83.3% of WDLPS samples demonstrated a low survivin expression, the remaining LPS subtypes (DDLPS 100%, MLPS 94.1%, PLS 100%) presented with exclusively or predominantly high expression (Table 2). Of note, high survivin expression levels also significantly correlated with higher grading (G2-3). In addition, when comparing the IRS as numeric variable across groups for each clinicopathological variable, we confirmed the association between survivin expression levels and LPS subtype or tumour grade (Fig. 1G–J).

Next, we investigated whether survivin expression and clinicopathological variables were associated with patients’ overall survival. Thereby, we created Kaplan–Meier survival curves and performed survival analysis using log rank analysis as well as a Cox regression model (Fig. 2, Table 3).

Accordingly, univariate analysis revealed that high survivin expression (IRS ≥ 3) correlated with a shorter overall survival for patients with high expression levels of survivin in the tumour (HR 5.307, CI 1.215–23.192, p = 0.027) (Table 3). In addition, LPS subtype was identified as a prognostic parameter (HR 1.960, CI 1.289–2.980, p = 0.002). Moreover, high grade tumours (G2/3) (HR 6.389, CI 1.825–22.364, p = 0.004) and advanced T stages (T3-4) (HR 4.063, CI 1.437–11.490, p = 0.008) were significantly associated with patients’ prognosis.

Multivariate analysis finally identified the primary tumour depth (T stage) (HR 4.391, CI 1.479–13.036, p = 0.008) and tumour subtype (HR 2.257, 1.380–3.691, p = 0.001) as independent prognostic markers for the assessed cohort of primary LPS (Table 3).

To explore the biological role of survivin in LPS, we first analysed the base line expression in liposarcoma cell lines Lipo-DUE1 (DDLPS), Lipo246A (DDLPS), and.

PLS-1 (PLS) by quantitative RT-PCR and western blot (Fig. 3A, B). While Lipo-DUE1 cells showed a significantly higher RNA expression with a mean 2^−ΔΔCT of 5.86 ± 2.47 (p = 0.0001; Kruskal–Wallis test), the protein expression in the cell line was comparatively weak. In contrast, PLS-1 as well as Lipo246A cells exhibited lower mean RNA expression levels of 0.73 ± 0.32 and an intermediate RNA-level amounting to 2.63 ± 1.22, respectively. At the same time, Lipo246A and more so PLS-1 protein levels exceeded the weak expression of survivin in Lipo-DUE1 cells.

To further elucidate the effect of a chemical inhibition of survivin, we incubated LPS cell lines with increasing concentrations of the small molecule antagonist YM155 for 96 h and measured cell viability by performing MTS assays (Fig. 3C). Incubation with YM155 resulted in a significant reduction of Lipo-DUE1, Lipo246A and PLS-1 cells in a dose-dependent manner with an IC50 of 0.15 µM, 0.16 µM, and 0.03 µM, respectively. Of note, a decrease in survivin protein levels became evident only in the more sensitive PLS-1 cells at 1 µM (Fig. 3D). To further assess the pro-apoptotic potency of survivin small molecule antagonist YM155, we again incubated LPS cell lines with increasing concentrations of YM155 and determined 24 h later the fraction of apoptotic cells by Annexin V/PI-staining and FACS (Fig. 3E). While Lipo-DUE1 cells showed an increase of apoptotic cells significantly correlating with the amount of administered YM155 reaching 17.3% at 100 nM and 54.8% at a 300 nM concentration (Kruskal-Wallis test, p = 0.0237; Dunn-Bonferroni test, p < 0.05), the treatment of Lipo246A cells with YM155 did not substantially affect the fraction of Annexin V and PI positive cells. In PLS-1 cells again, the increasing concentrations of YM155 were associated with higher proportions of apoptotic cells (Kruskal-Wallis test, p = 0.0156; Dunn-Bonferroni test, p < 0.01).

Next, we tested the response of LPS cell lines to the clinically established cytotoxic single-agents doxorubicin and etoposide solitarily as well as in combination with YM155 (Fig. 4).

Both, doxorubicin and etoposide induced a significant decrease in cell viability of all investigated cell lines, whereby significantly higher concentrations were necessary for the latter (Fig. 4A, B). Of note, combinational treatment of doxorubicin (10 µM) with increasing concentrations of YM155 (10 nM; 30 nM; 100 nM) for 72 h demonstrated only in Lipo-DUE1 cells a synergistic effect for 10 nM (FP = 0.64) of YM155 when administered together with doxorubicin (Fig. 4C, Supplementary Table 1). However, in Lipo-DUE1 and PLS cells treated with 10 µM of etoposide, we observed a synergistic effect when combined with YM155 at nanomolar concentrations. (Fig. 4D, Supplementary Table 1).