Synergistic Antitumour Properties of viscumTT in Alveolar Rhabdomyosarcoma

Aqueous mistletoe extracts from the European mistletoe (Viscum album) contain mainly mistletoe lectins and viscotoxins as cytotoxic compounds. Lipophilic triterpene acids, which do not occur in conventional mistletoe preparations, were solubilised with β-cyclodextrins. The combination of an aqueous extract (viscum) and a triterpene-containing extract (TT) recreated a whole mistletoe extract (viscumTT). These extracts were tested on rhabdomyosarcoma in vitro, ex vivo, and in vivo with regard to anticancer effects. Viscum and viscumTT inhibited cell proliferation and induced apoptosis effectively in a dose-dependent manner in vitro and ex vivo, whereas TT showed only moderate inhibitory effects. viscumTT proved to be more effective than the single extracts and displayed a synergistic effect in vitro and a stronger effect in vivo. viscumTT induced apoptosis via the extrinsic and intrinsic pathways, evidenced by the loss of mitochondrial membrane potential and activation of CASP8 and CASP9. CASP10 inhibitor inhibited apoptosis effectively, emphasising the importance of CASP10 in viscumTT-induced apoptosis. Additionally, viscumTT changed the ratio of apoptosis-associated proteins by downregulation of antiapoptotic proteins such as XIAP and BIRC5, thus shifting the balance towards apoptosis. viscumTT effectively reduced tumour volume in patient-derived xenografts in vivo and may be considered a promising substance for rhabdomyosarcoma therapy.

Lipophilic triterpene acids, such as oleanolic acid (OA) and betulinic acid (BA), represent another group of mistletoe-derived substances with known cytotoxic effects. Previous studies on neuroblastoma [30,31], osteosarcoma [32], acute leukaemia [33], breast cancer [34], and gastric cancer [35] showed that OA and its derivatives inhibit cell proliferation and induce apoptosis effectively in vitro. Moreover, in vivo experiments have shown cytotoxic effects of OA on osteosarcoma [36], gallbladder carcinoma [37], hepatocellular carcinoma [38], and pancreatic carcinoma [39]. Due to their low solubility, triterpene acids do not occur in significant amounts in commercial mistletoe extracts [40]. Using cyclodextrins, it was possible to solubilise triterpene acids and therefore to overcome their loss in standardised aqueous extracts.
Previous studies showed the efficacy of a combined mistletoe extract produced by addition of solubilised triterpene acids (TT) to aqueous mistletoe extracts (viscum), creating a whole plant extract (viscumTT). This whole plant extract effectively induced apoptosis in acute myeloid and lymphoblastic leukaemia [41,42], Ewing-sarcoma [43], and osteosarcoma [44] and inhibited tumour growth in murine melanoma in vivo [45]. The aim of this study was to analyse the effect of the whole mistletoe extract, viscumTT, on rhabdomyosarcoma in vitro, ex vivo, and in vivo.

2.2.
Viscum album Extracts. Viscum album extracts were kindly provided by Birken AG (Niefern-Öschelbronn, Germany) and prepared from Viscum album harvested from apple trees (Malus domestica Borkh.) as previously described [41,42,46]. Intact mistletoe lectin-I (A + B chain) was quantified by ELISA in the viscum extract as described before [47]. OA and BA were quantified in the TT extract by gas chromatography-flame ionisation detector (GC-FID) and external calibration with OA as reference substance [40]. Both extracts, viscum and TT, were reconstituted in PBS to a final concentration of 1 μg/mL intact ML-I and <1 μg/mL VT in viscum and 4000 μg/mL OA and 172 μg/mL BA in TT (Table 1).
2.3. Cell Culture. Two human ARMS cell lines were used in this study. RH-30 cells were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany) and RMS-13 cells from the American Type Culture Collection (ATCC, Manassas, VA, USA). Both cell lines were maintained in RPMI 1640 medium with L-glutamine, supplemented with 10% heatinactivated FCS, 100 U/mL penicillin, and 100 μg/mL streptomycin. For experimental settings, 2 × 10 5 /mL cells were seeded onto 6-well plates with a final volume of 2 mL/well. Cell numbers were adjusted to sizes of well plates according to experimental settings. Cells were cultured 24 h to allow cell attachment, then viscum, TT, and viscumTT were added in defined concentrations for another 24 h.

Determination of Cell Proliferation and Measurement of
Early Cytotoxicity. Cell counts were determined with the CASY® Cell Counter and Analyser System of Schärfe System GmbH (Reutlingen, Germany). Alterations of cell proliferation were indicated as the percentage of untreated control cells (100% viability). To rule out early cytotoxicity, lactate dehydrogenase (LDH) release was measured after 2 h of incubation with viscum, TT, and viscumTT using the Cytotoxicity Detection Kit (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions.

Assessment of Mitochondrial Membrane Potential (ΔΨ m ).
Changes in mitochondrial membrane potential (ΔΨ m ) were detected by JC-1 staining and flow cytometry, according to the manufacturer's instructions. Subsequently, results were evaluated with FlowJo Software. 50 μM depolarising CCCP was used as a reference control.
2.8. Caspase Inhibitor Assays. Cells were pretreated for one hour with 50-100 μM caspase inhibitors, before 24 h incubation with viscum (2.5-5 ng/mL) and viscumTT (2.5-5 ng/mL viscum + 35-40 μg/mL TT). DMSO was added as a solvent control. Induction of apoptosis was determined by annexin V/PI staining and flow cytometry analysis as described above. In this experimental setting, we used, in addition to a pan-caspase inhibitor (Z-VAD-FMK), inhibitors of CASP8 (Z-IETD-FMK), CASP9 (Z-LEHD-FMK), and caspase-10 (CASP10; Z-AEVD-FMK) (R&D Systems, Minneapolis, MN, USA). , or a combination thereof (viscumTT). Body weight was measured before each treatment, and mice were carefully monitored for health and symptoms of toxicity. The animals were sacrificed by cervical dislocation at the end of the experiment or if the mice were moribund (tumour volume > 1.2 cm 3 or >10% body weight loss). The animal experiments were carried out in accordance with the German legislation on the care and use of laboratory animals and in accordance with the United Kingdom Coordinating Committee on Cancer Research Guidelines for the Welfare of animals in Experimental Neoplasia to minimise suffering. Approval for the study was obtained from the Regional Office for Health and Social Affairs (LaGeSo, approval G-0030/15).
2.12. Statistics. All in vitro experiments were performed in independent experiments at least three times, for which means ± standard error are displayed in bar graphs. Ex vivo experiments were carried out only in sets of two independent experiments due to low cell growth. To assess whether the effect of viscumTT was synergistic, Webb's fractional product ( * Fp) [48] was calculated as described before [42]. Values greater than one ( * Fp > 1) display a synergistic effect. Twoway ANOVA and Bonferroni post hoc tests were applied to determine differences between mouse xenograft treatment groups. All results with p ≤ 0 05 were considered significant.

viscumTT Inhibits Cell Proliferation and Induces
Apoptosis Synergistically In Vitro. To assess the antiproliferative effects of mistletoe extracts, RMS-13 and RH-30 cells were incubated with increasing concentrations of viscum, TT, and viscumTT. Both, viscum and viscumTT, effectively inhibited proliferation in a dose-dependent manner in RMS-13 cells, whereas TT showed only moderate inhibitory effects (Figure 1(a)). RH-30 cells were more resistant towards treatment with viscum or TT, while viscumTT showed a strong inhibitory effect. Notably, treatment with viscumTT led to a synergistic inhibitory effect on cell proliferation in both cell lines in higher concentrations ( * Fp > 1).
To determine the induction of apoptosis, RMS-13 and RH-30 cells were analysed by annexin V/PI staining  and FACS analysis after 24 h of incubation with viscum, TT, and viscumTT. Neither viscum nor TT affected apoptosis in RH-30 cells; however, viscumTT induced apoptosis at high concentration ranges (Figure 1(a)). In comparison, RMS-13 cells reacted more sensitively to incubation with viscum and viscumTT, but treatment with TT also proved to be ineffective. viscumTT again displayed a synergistic effect on apoptosis induction in both  cell lines at high concentration ranges. Detection of PARP cleavage by Western blot analysis confirmed apoptosis induction (Figure 1(c)).
To exclude an unwanted early cytotoxic effect via necrosis, LDH-release was measured after 2 h of incubation with viscum, TT, and viscumTT in RMS-13 cells. As shown in Figure 1(b), no significant LDH-release was detected; hence, a cytotoxic effect by necrosis could be excluded.

viscumTT Displays a Synergistic Effect on Mitochondrial Membrane Depolarisation and Caspase Activation In Vitro.
To evaluate the underlying mechanisms of viscum-, TT-, and viscumTT-induced cytotoxicity more closely, we analysed the involvement of mitochondria and caspases in apoptosis. For this purpose, RMS-13 cells were treated as described above and ΔΨ m loss was detected by JC-1 staining and FACS analysis. Treatment with viscum, and to an even more pronounced extent with viscumTT, led to a dosedependent loss of ΔΨ m , whereas TT showed no significant effect (Figure 2(a)).
Caspases play an essential role in apoptosis, since they are initiators (CASP9, CASP8) as well as effectors (CASP3). After 24 h treatment with viscum, TT, and viscumTT, caspase activation was assessed by active caspase assays and FACS analysis in RMS-13 cells. Viscum as well as viscumTT activated CASP3, CASP8, and CASP9 in a dose-dependent manner (Figure 2(b)). viscumTT displayed a synergistic effect on mitochondrial membrane depolarisation and caspase activation, confirming observations made in the previous experiments.
To further determine caspase dependency in apoptosis induction, caspase inhibitor assays were performed and inhibition of apoptosis was estimated from untreated control cells. Interestingly, caspase inhibitors were more effective in preventing apoptosis in cells treated with viscum rather than viscumTT. The pan-caspase inhibitor achieved reduction of apoptosis by up to 73% in viscum-and 46% in viscumTTtreated cells (Figure 3). Inhibition of the extrinsic pathway by CASP10 and CASP8 inhibitors prevented apoptosis by up to 75% and 60%, respectively, after treatment with viscum and by 50% and 30%, respectively, in viscumTT-treated cells. Inhibition of the intrinsic apoptosis signalling pathway by CASP9 inhibitor was also effective in viscum-treated cells, but showed no effect after incubation with viscumTT. On the contrary, in increasing concentrations, CASP9 inhibitor enhanced viscumTT-induced apoptosis.
These results suggest that in addition to the intrinsic signalling pathway the extrinsic pathway plays a role in viscum and viscumTT-induced caspase-dependent apoptosis.

Viscum album Extracts Alter the Expression of Antiapoptotic Proteins.
To further elucidate the underlying molecular mechanisms of viscum-, TT-, and viscumTTinduced apoptosis, we analysed protein expression of apoptosis-associated proteins by Western blot analysis after 24 h treatment with mistletoe extracts. Viscum, TT, and viscumTT suppressed the expression of the inhibitor of apoptosis protein (IAP) family members XIAP and BIRC5. Further, incubation with viscum and viscumTT led to downregulation of BCL2 family members BCL2, BCL2L1, and MCL1 (Figure 4).

viscumTT Inhibits Proliferation and Induces Apoptosis Ex
Vivo. To verify in vitro results, we next examined the effects of viscum, TT, and viscumTT in primary RMS tumour cells (ex vivo). For this purpose, ex vivo cultures of primary cells of three different patients were treated with TT, viscum, and viscumTT in increasing concentrations for 48 h (patient no. 1) and 24 h (patient no. 2; patient no. 3), respectively. Patient no. 1 was incubated with lower concentrations of viscum, TT, and viscumTT than patient no. 2 and patient no. 3, but we changed the time course and extended the incubation time to 48 h. Viscum and to a more pronounced extent viscumTT effectively induced apoptosis and decreased ΔΨ m , while TT showed no cytotoxic effect ex vivo ( Figure 5(a)). Patient no. 2 showed only a moderate increase of apoptotic cells after treatment with viscumTT, while neither TT nor viscum inhibited cell proliferation or induced apoptosis effectively after 24 h incubation ( Figure 5(b)). After 24 h incubation with viscum and viscumTT, we were able to detect a concentration-dependent inhibition of proliferation as well as induction of apoptosis in patient no. 3 ( Figure 5(c)).

viscumTT Effectively Inhibits Tumour Growth In Vivo.
To validate our in vitro and ex vivo findings, we established two patient-derived xenografts (PDX) to examine the effects of viscum, TT, and viscumTT in vivo. For this purpose, we focused on two different study designs. In PDX no. 1, we examined the therapeutic effect of i.t.-administered vis-cumTT compared to viscum, TT, and doxorubicin. In PDX no. 2, however, we investigated the therapeutic effectiveness of i.t. and i.v. application. Therefore, NMRI-nu/nu mice were treated with viscum, TT, and viscumTT (i.v. and/or i.t.). Control groups received the vehicle cyclodextrin i.v. or i.t. Extracts were administered every two to three days in increasing concentrations, and each concentration was given twice. During treatment, the mice showed no significant loss of body weight, and the administered extracts were well tolerated. The mean tumour volumes were calculated ( Figure 6). In PDX no. 1, i.t. administration of viscum and TT reduced tumour volume effectively and viscumTT led to a significant inhibition of tumour volume in comparison to the control group (Figure 6(a)), whereas doxorubicin as a standard therapeutic drug had no effect. In PDX no. 2, viscumTT i.t. was more effective than viscumTT i.v. and both application forms inhibited tumour volume significantly compared to control mice ( Figure 6(b)).

Discussion
This study shows that viscum and viscumTT exerted strong antiproliferative and apoptosis-inducing effects on RMS in vitro and ex vivo, whereas TT showed only moderate inhibitory effects. RH-30 cells were more resistant to viscum treatment than RMS-13 cells. It is known that cytotoxic effects of ML, the main component of the viscum extract, differ in cell lines, probably due to differences in expression of ML binding sites on the cell surface [49,50]. The marker substance of the TT extract OA, which exert cytotoxic effects in various cancer cell lines [37,46,51], did not induce apoptosis in either RMS cell line effectively, which might be due to low OA concentrations. The whole mistletoe extract, viscumTT, proved to be more potent than the single extracts and displayed a synergistic effect on apoptosis induction in RMS cells in vitro and ex vivo. This result is in line with the assumption that whole plant extracts can be more effective than their single constituents [52][53][54]. Synergistic effects of viscumTT were already recently shown in other paediatric tumour cell lines [41][42][43][44].
Viscum and viscumTT triggered apoptosis in RMS via the extrinsic as well as the intrinsic apoptosis signalling pathway, evidenced by mitochondrial membrane depolarisation and activation of CASP8 and CASP9 (Figure 7). This is in line with results from several studies, which reported that ML as well as OA induced apoptosis by affecting different apoptosis pathways. Aqueous mistletoe extracts as well as purified ML proved to exert cytotoxic effects in various cancer cell lines either by the intrinsic [21,25,55] or additionally by the extrinsic [41,56] apoptosis signalling pathway. OA and its derivatives are also known to invoke both the extrinsic [57][58][59][60][61][62] as well as the intrinsic [37][38][39] signalling pathway. Furthermore, BA could induce apoptosis via the mitochondrial signalling pathway in RMS [63] and neuroblastoma [64]. Additionally, BA and OA have been shown to interact directly with mitochondrial membranes [64][65][66].
Caspase inhibitory assays showed that viscum-and viscumTT-induced apoptosis was caspase-dependent. Interestingly, caspase inhibitors were more effective in preventing apoptosis in cells treated with viscum rather than viscumTT. Addition of triterpene acids to aqueous mistletoe extracts seems to activate as yet unknown mechanisms, which enhance apoptosis induction caspase independently.
Caspase-independent apoptosis induction by OA was also observed by Konopleva et al. [66].
Inhibition of the extrinsic pathway by CASP8 and CASP10 inhibitors reduced apoptosis significantly. Notably, the CASP10 inhibitor was more effective in reducing apoptosis than the pan-caspase inhibitor, underlining the dependency of viscum-and viscumTT-mediated apoptosis induction on CASP10. CASP10 is a homologue of CASP8 and activated through the receptor-mediated pathway.  However, some studies indicate involvement of CASP10 in apoptosis induction by cytotoxic agents in a FADDdependent but receptor-independent manner [67,68]. Interestingly, the CASP9 inhibitor was not only ineffective in viscumTT-treated cells, but enhanced vis-cumTT-induced cell death. This result is in line with the previous results, which reported augmented apoptosis after treatment with inhibitors of CASP9 in combination with oridonin [69]. Also, CASP9 inhibition is suspected to block autophagic flux and thus to promote caspase-independent cell death [69,70]. However, caspase inhibitor assays suggest that CASP9 does not act as an initiator but merely as an effector caspase in an amplification loop and is activated by CASP8 or CASP10 similar to TNF-α, which induced apoptosis via activation of CASP9 by CASP8 [71,72]. Additionally, the mitochondrial and receptor-mediated pathways are crosslinked by the cleavage of BH3 interacting domain death agonist (BID) into truncated BID (tBID) by CASP8 or CASP10. Also, cotreatment with the OA derivative C-28 methyl ester of 2-cyano-3,12- dioxoolean-1,9-dien-28-oic acid (CDDO-Me) and CASP8 inhibitor CASP9 activity was decreased, suggesting that CASP9 activation is downstream of CASP8 [73].
To illuminate the underlying mechanisms of viscum-, TT-, and viscumTT-induced apoptosis further, apoptosisassociated proteins were analysed by Western blot analyses to detect altered expression. The results showed that viscum and even more prominently viscumTT shifted the balance towards apoptosis by changing the ratio of pro-and antiapoptotic proteins.
After treatment with viscum and viscumTT, downregulation of the antiapoptotic BCL2 family members BCL2 and BCL2L1 was detected. Overexpression of BCL2 family members is often found in cancers and contributes to apoptosis resistance. Genetic mutations, for example, gli-amplification and a positive PAX3/FOXO1 fusion status, and increased activity of transcription factor signal transducer and activator of transcription 3 (STAT3) lead to overexpression of BCL2 and BCL2L1 in RMS [74]. Downregulation of BCL2 and BCL2L1 was found after treatment with viscum [42] and OA [75,76]. MCL1, another antiapoptotic BCL2 family member, which is upregulated in RMS [77], is also downregulated by viscum and viscumTT. This is in line with the previous results of decreased expression of MCL1 after treatment with ML [21,78] and viscumTT [43]. Furthermore, Ryu et al. have shown inhibition of the STAT3 pathway with downregulation of BCL2L1, BIRC5, and MCL1 after incubation with CDDO-Me [79]. Moreover, viscum and viscumTT effectively reduced expression of IAP family members BIRC5 and XIAP. IAP family members prevent apoptosis by directly inhibiting caspases or the assembly of proapoptotic complexes. BIRC5 is often highly expressed in malignant cells but not in differentiated tissues [80], and high expression levels are associated with poor prognosis [81] thus appointing it an attractive target in tumour therapy. RMS expresses high levels of BIRC5, and Caldas et al. were able to effectively reduce tumour growth by blocking BIRC5 in vivo [82]. OA treatment was able to effectively reduce BIRC5 expression in lung cancer [83], ovarian cell carcinoma [75,84], and leukaemia [76] as well as XIAP expression in hepatocellular carcinoma [85]. Furthermore, we were able to show that viscum, TT, and viscumTT led to downregulation of BIRC5 and XIAP in AML [41], Ewing-sarcoma [43], and osteosarcoma [44].
Treatment with viscumTT seems to be more effective compared to the single compounds ex vivo and in vivo. Furthermore, viscumTT was more effective in reducing tumour volume than i.v. doxorubicin treatment in PDX no. 1. In addition, comparison of i.v. and i.t. treatment (PDX no. 2) showed that i.t. treatment is more effective. On the other hand, viscumTT can be administered in higher concentrations i.v. than i.t. Notably, viscumTT was more effective in vivo than in vitro. While viscumTT showed only moderate apoptosis-inducing properties ex vivo, tumour volume was effectively reduced in the xenograft derived from the same patient. Since mistletoe extracts are metabolised by mice, these in vitro results cannot be directly extrapolated to in vivo settings. However, viscumTT shows promising antitumour activity in ARMS.

Conclusions
The results show the high potential of viscumTT to induce apoptosis in ARMS without limiting toxicity. viscumTT is more effective than the single compounds and leads to a synergistic effect on apoptosis induction in vitro as well as ex vivo. In vivo viscumTT treatment resulted in an effective reduction of tumour volume compared to controls. In conclusion, addition of triterpene acids to aqueous mistletoe extracts enhances their cytotoxic effects in ARMS through an as yet unknown mechanism. However, viscumTT is a promising novel treatment approach for ARMS. Furthermore, ML exerts immunomodulating effects [86,87]. The antiproliferative and apoptosis-inducing properties of viscumTT combined with the immunomodulating actions of ML provide an attractive research target not only for cancer therapy, but also for other research purposes such as rheumatic and cardiovascular diseases.

Ethical Approval
Written informed consent was obtained from the patient's parents and/or legal guardians in accordance with the Declaration of Helsinki, approved by the local ethics committee of Charité-Universitätsmedizin Berlin.

Conflicts of Interest
The authors have the following interests: J. Rolff is employed by EPO GmbH. Sebastian Jäger is employed by Birken AG. A patent, EP 1852105 A3, has been filed by this company. The patent affects commercial reproduction and use of triterpene containing plant extracts like viscumTT, while noncommercial use for research is not affected. There are no further patents, products in development, or marketed products to declare.