Transcriptomic and proteomic insight into the effects of a defined European mistletoe extract in Ewing sarcoma cells reveals cellular stress responses

Viscum and TT extracts were prepared from Viscum album L. harvested from apple trees (malus) as previously described [36] by the Birken AG (Niefern-Oeschelbronn, Germany), who kindly provided the lyophilized viscum and TT extracts for this study. Intact mistletoe lectin I (ML-I) was analysed by ELISA in viscum extract [37]. Oleanolic and betulinic acid were quantified, as a measure of extract activity, using GC-FID [18]. Lyophilized viscum extract was reconstituted in PBS (Gibco® Life Technologies, Darmstadt, Germany) to a final concentration of 2 μg/mL intact ML-I and <1 μg/mL viscotoxins. Lyophilized TT extract (containing cyclodextrins) was reconstituted in phosphate-buffered saline to a final concentration of 4000 μg/mL oleanolic and 350 μg/mL betulinic acid.

Human Ewing sarcoma cell lines were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). The TC-71 and MHH-ES-1 cell lines were maintained in Iscove’s Modified Dulbecco’s Medium (Gibco Life Technologies) and RPMI 1640 base medium both supplemented with L-glutamine (Gibco Life Technologies), respectively. Base media were supplemented with 10% heat-inactivated FBS (Biochrom, Berlin, Germany), 100 U/mL penicillin and 100 μg/mL streptomycin (Biochrom). For assays, TC-71 cells were seeded into 6-well (2 × 10⁵) or 12-well (1 × 10⁵) microtiter plates or 25 cm² (5 × 10⁵) or 75 cm² (1.5 × 10⁶) cell culture flasks depending on experimental set-up. MHH-ES-1 cells were seeded into 6-well (4 × 10⁵) or 12-well (2 × 10⁵) microtiter plates. Cells were cultured 24 h to allow cell attachment, and treated 24 h with Viscum album L. extracts added to culture media. Viscum, TT and viscumTT concentrations were assessed by dose-effect-curves of apoptosis measurements as previously described [38].

TC-71 cells were incubated with increasing concentrations of the extracts for 24 h. RNA was isolated using the NucleoSpin® RNA Kit according to the manufacturer’s protocol (Macherey-Nagel, Düren, Germany) in five independent experiments. Purity and concentration was determined by OD260/280 on the NanoDrop™ 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA).

TC-71 cells were treated once with ~IC₅₀ extract concentrations (viscum 2 ng/mL ML-I, TT 50 μg/mL oleanolic acid, viscumTT 1 ng/mL ML-I + 10 μg/mL oleanolic acid) for 24 h. After total RNA isolation from treated and control cells, Illumina TruSeq RNA sample preparation including a polyA selection step via oligo-dT beads was used to isolate mRNA and generate cDNA libraries. Samples were sequenced by paired-end mRNA sequencing on an Illumina HiSeq 2500 (50 bp reads, n = 1). Reads were mapped uniquely to human genome hg19 using CLC genomics software. Normalisation and identification of differentially expressed genes was performed using DE-Seq software [39] (Bioconductor open source software) to calculate fold-change relative to untreated control cells, false discovery rate (FDR) [40] and p value using the negative binomial distribution, with p ≤ 0.05 considered as significant. The heat map and Venn diagram illustrating differential gene expression as log₂-fold change relative to untreated control cells was performed by using the R software package (http://​www.​r-project.​org/). Gene enrichment and functional annotation analysis was performed with the top differentially expressed genes (p ≤ 0.01) using DAVID Bioinformatics Resources 6.7 NIAID/NIH (http://​david.​abcc.​ncifcrf.​gov/).

Using 2 μg RNA from treated or control TC-71 cells, cDNA was synthesised using the High Capacity RNA-to-cDNA Kit according to the manufacturer’s protocol (Applied Biosystems, Waltham, MA, USA). To confirm mRNA sequencing results, gene expression was measured by real-time PCR on the StepOnePlus™ System in 96-well fast plates under standard conditions (10 min, 95 °C; 15 s, 95 °C and 60 s, 60 °C, 40×) using Power SYBR Green Master Mix including ROX as a passive reference (Applied Biosystems). The quantitative PCR reactions were 20 μl total volume containing 5 ng cDNA and 500 nM primers, which were predesigned and purchased from Integrated DNA Technologies (IDT, Leuven, Belgium): DDIT3: Hs.PT.58.14610020; JUN: Hs.PT.58.25094714.g; MAP2K6: Hs.PT.58.3312889; GAPDH: Hs.PT.39a.22214836. Primer efficiency was determined to be 90–100%. Expression was normalised using GAPDH expression. The relative expression of genes was calculated by ∆∆CT method: ΔΔCT = (CT(target,untreated) – CT(ref,untreated)) − (CT(target,treated) – CT(ref,treated)); fold-change =2^^{(−∆∆CT)} [41].

TC-71 cells were grown in 25 cm² cell culture flasks in triplicate and treated with viscum, TT or viscumTT extracts at ~ IC₅₀ concentrations (viscum 2 ng/mL ML-I, TT 50 μg/mL oleanolic acid, viscumTT 1 ng/mL ML-I + 10 μg/mL oleanolic acid) for 24 h. Cells were harvested and lysed under denaturing conditions in a buffer containing 6 M guanidinium chloride, 10 mM tris(2-carboxyethyl)phosphine, 40 mM chloroacetamide and 100 mM Tris pH 8.5. Lysates were sonicated and boiled at 95 °C for 5 min. Lysates were diluted 1:10 in 10% acetonitrile in 25 mM Tris, pH 8.5, and 2% of each total lysate volume was digested with 1 μg trypsin at 37 °C overnight. Peptides were acidified by adding formic acid to a final concentration of 1%, desalted using C18 StageTips (Thermo Scientific, Waltham, MA, USA) and lyophilised. Pellets were dissolved in 5% acetonitrile and 2% formic acid and one quarter of the digest was injected for nanoflow reverse-phase liquid chromatography (Dionex Ultimate 3000, Thermo Scientific) coupled online to a Thermo Scientific Q-Exactive Plus Orbitrap mass spectrometer (nanoLC-MS/MS). Briefly, the LC separation was performed using a PicoFrit analytical column (75 μm ID × 40 cm long, 15 μm Tip ID (New Objectives, Woburn, MA) packed in-house with 2.1 μm C18 resin (Reprosil-AQ Pur, Dr. Maisch, Ammerbuch-Entringen, Germany) under 50 C controlled temperature. Peptides were eluted using a nonlinear gradient from 2 to 40% solvent B in solvent A over 180 min at 266 nL/min flow rate. Solvent A was 0.1% formic acid and solvent B was 79.9% acetonitrile, 20% H₂O, 0.1% formic acid). Nanoelectrospray was generated by applying 3 kV. A cycle of one full Fourier transformation scan mass spectrum (300–1700 m/z, resolution of 35,000 at m/z 200) was followed by 12 data-dependent MS/MS scans with normalised collision energy of 25 eV. A dynamic exclusion window of 30 s was used to avoid repeated sequencing of the same peptides. MS data were analysed by MaxQuant (v1.5.0.0) [42]. The algorithm MaxLFQ [43], which is integrated into MaxQuant, was used for label free quantification (LFQ). Peptides were searched against the human proteome UniProtKB database released in 11/2014 with 88,717 entries, released in 11/2014, using an FDR of ≤0.01 for proteins and peptides with a minimum length of seven amino acids. A maximum of two missed cleavages in the tryptic digest was allowed. Cysteine carbamidomethylation was set as a fixed modification, while N-terminal acetylation and methionine oxidation were set as variable modifications. Principal component analysis (PCA, Fig. 3b) and a heat map of the replicates (Fig. 3a) were carried out using the Perseus software (v1.5.1.6). Gene set enrichment analysis (GSEA, v2.1.0) [44] was applied to assess a priori defined protein sets with statistically significant expression differences between treated and control cells. We used GSEA standard settings, except that minimum size exclusion was set to five and KEGG v2.1.0 as well as reactome v5.0 was used as the gene set database. The String software tool (v10) was used to visualise protein-protein interaction networks of proteins up- or downregulated at least 2-fold [45]. Protein nodes that were not integrated into the protein-protein interaction network were removed.

TC-71 and MHH-ES-1 cells were incubated with viscum, TT or viscumTT in increasing concentrations for 24 h. The cells were washed twice with phosphate-buffered saline and incubated in Lysis Buffer 17 (R&D systems, Minneapolis, MN) containing protease inhibitors (complete Protease Inhibitor Cocktail Tablets, Roche Diagnostics GmbH) to obtain cell lysates. Protein concentration was determined using Bradford solution (Bio-Rad, Munich, Germany). TC-71 and MHH-ES-1 cell lysates (30 μg protein/lane) were separated on SDS-PAGE, transferred to nitrocellulose membranes (Bio-Rad system) and blocked with 5% non-fat milk in 50 mM Tris-buffered saline containing 0.05% Tween-20 (TBST) for 1 h at room temperature. Blots were incubated overnight at 4 °C in TBST containing 5% BSA and primary antibody, washed thrice in TBST and incubated 1 h with HRP-conjugated secondary antibodies (anti-rabbit and anti-mouse, Bio-Rad) then visualized by ECL (Thermo Scientific) on a Molecular Imager ChemiDoc (Bio-Rad). Primary antibodies were directed against p-MAPK14 (Thr180/Tyr182, #9211 Cell Signaling Technology, Danvers, MA, USA), LC3B (#2775 Cell Signaling Technology), EIF2AK3 (#3192, Cell Signaling Technology), p-MAPK8 (sc-6254, Santa Cruz biotechnology, Dallas, TX, USA), HSPA5 (#G8918, Sigma-Aldrich) and ß-actin conjugated directly to peroxidase (#A3854, Sigma-Aldrich).

To measure the impact of TLR4 signalling, MAPK14 and MAPK8 activation or oxidative stress on apoptotic induction, TC-71 cells were pre-incubated with specific inhibitors for 1 h followed by an incubation with ~ IC₅₀ extract concentrations for 24 h. Specific inhibitor treatment included 5–50 μM SB203580 (Cell Signaling Technology), 1 μM – 25 μM SP600125 (Sigma-Aldrich), 1–10 mM N-acetylcysteine (NAC, Sigma-Aldrich), 0.1–10 μg mL⁻¹ LPS-RS (InvivoGen, San Diego, CA, USA). DMSO (Sigma-Aldrich) or PBS was added to extracts as solvent control, depending on the inhibitor diluent. After treatment, cells were washed twice with PBS, resuspended in 100 μl binding buffer and stained with APC-conjugated annexin V (BD Biosciences, Franklin Lakes, NJ, USA), according to the manufacturer’s protocol, then counterstained with 1 μl 1 mg/mL propidium iodide (Sigma Aldrich). Cells were analysed by flow cytometry (FACSCalibur, Becton Dickinson, Heidelberg, Germany), and the data were evaluated using FlowJo Software (TreeStar, Ashland, OR, USA).

All qPCR experiments were performed in triplicate and experiments were repeated four times. Results are presented as mean values ± standard error of the mean (SEM). Inhibitor treatment and measurement of apoptosis was repeated thrice independently. The results are also presented as mean values ± SEM. Western blots were performed in three independent experiments. mRNA sequencing was performed in one experiment, while proteomic profiling was performed in triplicate. The cut-off for significantly deregulated pathways was set to p ≤ 0.01 for mRNAseq and p ≤ 0.05 and FDR ≤ 0.25 for proteomic profiling. Principal component analysis was performed to compare differences within the triplicates used for proteomic profiling. Protein-protein interaction networks of proteins up- or downregulated were regarded as significant by at least 2-fold change with a confidence level of 0.7 [45].

Firstly, to explore global changes asserted by mistletoe extracts in the Ewing sarcoma cell transcriptome, we treated TC-71 cells for 24 h with viscum, TT or viscumTT in ~ IC₅₀ concentrations in reference to untreated control cells and performed one mRNA sequencing experiment. mRNA sequencing detected >62,400 transcripts belonging to >17,300 genes. Treatment with the extracts had a massive impact on the transcriptome of TC-71 cells, as shown in the heat map (hierarchical clustering) of log₂-transformed fold-changes in gene expression (Fig. 1). Interestingly, treatment with viscumTT or viscum produced similar transcriptomic patterns of differentially expressed genes and resulted in a differential expression of >1000 genes, most of which were upregulated relative to the untreated control cells (Fig. 1). TT treatment displayed a different transcriptomic pattern and had a lesser impact on the TC-71 transcriptome, causing the differential expression of only 249 genes (Fig. 1). An overview of the differentially expressed genes is also given in a Venn diagram (Additional file 1: Figure S1).

Several components of immune and cellular stress responses as well as cell survival and death pathways were within the genes deregulated by viscumTT, viscum or TT treatment. The top 40 genes (in terms of p ≤ 0.05) regulated by viscumTT, viscum or TT treatment of TC-71 cells are shown in Additional file 2: Tables S1, Additional file 3: Table S2 and Additional file 4: Table S3, respectively. For instance, JUN, DDIT3 and CXCL8 (formerly IL8) were upregulated and MAP2K6 was downregulated by either viscumTT or viscum treatment within the top 40 deregulated genes, while KLHDC7B, NUPR1 and CYTH3 (formerly GPR1) were upregulated by TT treatment. JUN and DDIT3 upregulation and MAP2K6 downregulation were confirmed by qPCR (Fig. 2).

To look more deeply into pathways affected at the transcript level by mistletoe extract treatment, we preformed gene enrichment and functional annotation analysis using the DAVID bioinformatics tool. Since mRNA-sequencing experiment was performed only once, we chose a strict cut-off using exclusively genes with p ≤ 0.01 for differential expression changes from the control cells. ViscumTT and viscum treatment impacted functional annotation clusters for the positive regulation of cell death, response to reactive oxygen species/oxidative stress and MAPK signalling, while TT affected clusters for Toll-like receptor signalling, positive regulation of cell death and inflammatory response (Table 1). Taken together, the combined viscumTT extract as well as both single extracts altered expression of many genes, mostly upregulating them, in the TC-71 Ewing sarcoma cell line. ViscumTT or viscum treatment of TC-71 cells indicates a transcriptomic upregulation of cell stress and MAPK signalling genes, while TT appears to involve inflammatory response/TLR signalling, with both extract components transcriptionally exerting a pro-apoptotic effect on TC-71 Ewing sarcoma cells.

We next extended our investigations of mistletoe extract component effects to changes exerted on the TC-71 Ewing sarcoma cell proteome using LC − MS/MS technology. We examined the same temporal and dose response window (24 h and ~ IC₅₀ concentrations) as for transcriptomic changes. To get more reliable results, TC-71 cells were treated in three independent experiments with viscumTT, viscum or TT in reference to untreated control cells followed by subsequent analysis by LC − MS/MS. More than 3 × 10⁴ MS spectra were identified in each sample, which could be mapped to 3554 proteins in total. The ion intensities were log₂-transformed and normalised using the z-score by Perseus software and then plotted as a heat map (hierarchical clustering) to display the relative protein expression levels in the different TC-71 cell treatments and untreated control cells. Similar to the transcriptomic results, the heat map displayed an impressive alteration of the proteome after treatment with viscumTT or viscum compared to control cells, whereas TT treatment showed fewer changes (Fig. 3a). Principal component analysis verified good compliance between the experimental replicates (Fig. 3b). Therefore, LFQ ion intensities from the experimental replicates were averaged for pathway analyses, only valid values were used. The top 40 differentially expressed proteins (in terms of p ≤ 0.05) are presented in Additional file 5: Tables S4, Additional file 6: Table S5 and Additional file 7: Table S6. To look more deeply into pathways affected at the protein level by mistletoe extract treatment, we preformed gene set enrichment analysis using the GSEA bioinformatics tool. Downregulated proteins after treatment with viscumTT, viscum or TT related to ribosome/translation and transcription (Table 2). Protein-protein interaction network analyses of proteins downregulated by at least 2-fold revealed for viscumTT or viscum treatment proteins involved in the ribosome and spliceosome, whereas TT treatment was not linked with any functional protein networks (Fig. 3c). Gene set enrichment analysis also revealed that viscumTT, viscum or TT treatment upregulated proteins linked with aminoacyl-tRNA biosynthesis and the proteasome. Additionally, treatment with viscumTT or viscum also upregulated proteins associated with the regulation of apoptosis, protein folding, PERK-regulated gene expression, MAP kinase activation downstream of the toll-like receptor and the immune system (Table 3). Protein-protein interaction network analyses of proteins upregulated at least 2-fold by viscumTT treatment implicated both the protein processing network in the endoplasmic reticulum (ER) and the proteasome, while viscum treatment implicated aminoacyl-tRNA synthesis and the proteasome (Fig. 3c). Our data indicate that treatment with viscum, TT or the combined viscumTT extracts results in a reduction of proteins related to the ribosome and spliceosome, while increasing proteins associated with the proteasome. Treatment with either viscumTT or viscum also triggers protein upregulation associated with apoptosis, protein folding and MAPK signalling.

We next aimed to analyse the activation of cellular stress responses upon mistletoe extract treatment using western blotting to validate corresponding hints from transcriptomic and proteomic results. Activation of stress-mediated MAPK signalling by viscumTT and viscum was confirmed by increased phosphorylation of MAPK8 (formerly JNK) and MAPK14 (formerly p38/MAPK) in TC-71 and MHH-ES-1 cells 24 h after treatment with the extracts (Fig. 4a). All mistletoe extracts upregulated expression of the HSPA5 ER-chaperone protein, indicating response to cellular stress (e.g. oxidative stress or ER stress) by activation of the unfolded protein response (Fig. 4a). ViscumTT and viscum also reduced levels of the EIF2AK3 ER-stress sensor protein, whereas TT slightly increased EIF2AK3 levels (Fig. 4a). Since ER stress is linked to autophagy [46, 47], we next checked LC3B expression, whose conversion from LC3B-I to LC3B-II is a marker for autophagy. ViscumTT and TT increased LC3B-II levels in TC-71 cells, whereas only TT increased LC3B-II levels in MHH-ES-1 cells (Fig. 4a).

In order to further analyse the role of stress-mediated MAPK or TLR signalling and oxidative stress in mistletoe-mediated apoptosis of Ewing sarcoma cells, TC-71 cells were treated with viscumTT, viscum or TT for 24 h in the absence or presence of the SB203580 MAPK14 inhibitor, the SP600125 MAPK8 inhibitor, the LPS-RS TLR4 inhibitor or the antioxidant NAC. 5 μM MAPK8 and 10 μM MAPK14 inhibitor treatment was not able to prevent apoptotic induction by mistletoe extracts. Western blot analyses of TC-71 cell lysates after inhibitor treatment revealed no reduction in MAPK8 phosphorylation and, unexpectedly, enhanced MAPK14 phosphorylation (Fig. 4b). Higher MAPK8 and MAPK14 inhibitor concentrations (up to 50 μM SB203580 and 25 μM SP600125), as well as higher concentrations of the other used inhibitors (Table 3), all increased apoptosis in TC-71 cells. Pre-treatment with the antioxidant NAC (5 mM), however, reduced apoptosis by 21% after viscum treatment and ~13% after viscumTT treatment (Fig. 4c). LPS-RS (0.1 μg/mL), which binds TLR4 but prevents receptor activation, reduced TT-mediated apoptosis by ~20% (Fig. 4c). Taken together, these data indicate that viscumTT, viscum and TT activate the unfolded protein response. TT extract treatment suggests the activation of TLR4 signalling and autophagy, while viscumTT and viscum indicate the induction of oxidative stress and stress-mediated MAPK signalling. Since mistletoe extracts appear to trigger multiple pathways sensitizing Ewing cells to apoptosis, inhibiting single pathways remained insufficient to prevent apoptosis in combination with extract treatment.