J Cancer Res Clin Oncol DOI 10.1007/s00432-014-1683-6

Original Article – Cancer Research

Re‑evaluation of cytostatic therapies for meningiomas in vitro Annette Wilisch‑Neumann · Doreen Pachow · Maren Wallesch · Astrid Petermann · Frank D. Böhmer · Elmar Kirches · Christian Mawrin 

Received: 7 April 2014 / Accepted: 10 April 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  Purpose The purpose was to re-evaluate in cell culture models the therapeutic usefulness of some discussed chemotherapies or targeted therapies for meningiomas with a special emphasis on the role of the neurofibromatosis type 2 (NF2) tumor suppressor, which had been neglected so far. In addition, the study intended to evaluate a potential benefit from a treatment with drugs which are well established in other fields of medicine and have been linked recently with tumor disease by epidemiological studies. Methods  Meningioma cell lines corresponding to various subtypes and pairs of syngenic meningioma cell lines with or without shRNA-induced NF2 knockdown were analyzed for their dose-dependent response to the drugs in microtiter tetrazolium assays, BrdU assays and for selected cases in ELISAs measuring nucleosome liberation to specifically separate cell death from pure inhibition of cell proliferation. Results  We confirmed a moderate efficacy of hydroxyurea (HU) in clinically relevant concentrations. Under appropriate dosing, we neither detected major responses to the alkylating compound temozolomide nor to various drugs targeting membrane receptors or enzymes (tamoxifen, erlotinib, mifepristone, losartan, metformin and verapamil). Only concentrations far beyond achievable serum levels generated significant effects with the exception of losartan, A. Wilisch‑Neumann · D. Pachow · M. Wallesch · E. Kirches · C. Mawrin (*)  Department of Neuropathology, Otto-von-Guericke University, Leipziger Str. 44, 39120 Magdeburg, Germany e-mail: [email protected] A. Petermann · F. D. Böhmer  Institute of Molecular Cell Biology, CMB, Jena University Hospital, Jena, Germany

which showed no effects at all. Chemosensitivity varied markedly among meningioma cell lines. Importantly, cells with NF2 loss exhibited a significantly higher induction of cell death by HU. Conclusions Alternative chemotherapeutic or targeted approaches besides HU have still to be evaluated in further studies, and the role of NF2 must be taken into account. Keywords  Meningioma · Chemotherapy · NF2 · Radiosensitization

Introduction Meningiomas are among the most common intracranial neoplasms of adulthood. At present, recurring and highgrade tumors are insufficiently treated, because an established chemotherapeutic option for these cases is missing. Specifically, for the quite common recurrences of incompletely resected benign tumors (WHO grade I), chemotherapy should have to be a mild one without major adverse effects, i.e., potentially a targeted therapy directed against a single receptor or enzyme. Numerous strong cytotoxic agents have been tested in refractory cases with disappointing results (Chamberlain and Blumenthal 2004; Kyritsis 1996). The failure of alkylating agents, e.g., temozolomide, may be explained by high levels of the protein MGMT (methyl-guanine-DNA-methyltransferase) in meningiomas which removes alkyl residues from the DNA (de Robles et al. 2008). Thus far, only the ribonucleotide-reductase inhibitor hydroxyurea (HU), which strongly reduces DNA synthesis of tumor cells, has been applied with some success and evaluated in several clinical studies (Madaan et al. 2012). Already in 1997, Schrell and colleagues reported

13



a reduction in tumor mass by HU in three patients with unresectable benign meningiomas and stable disease for 2 years in a patient with a malignant meningioma, which had recurred previously multiple times. The authors validated the anti-tumor efficacy of the drug by demonstrating growth inhibition, cell cycle arrest and apoptosis of meningioma cells in vitro and in vivo (Schrell et al. 1997). In a larger study including 21 patients, a moderate efficacy of the same treatment scheme (20 mg/kg daily) was confirmed, leading to the suggestion of using HU if tumors cannot be removed completely and/or are resistant to irradiation (Newton 2007). A combination of HU and radiotherapy resulted in stable disease in a majority of cases (Hahn et al. 2005). HU also showed modest activity if combined with the receptor tyrosine kinase (RTK) inhibitor imatinib (Reardon et al. 2012). However, not all studies supported a benefit from HU therapy in recurrent non-malignant meningiomas (Loven et al. 2004). Recently, two larger retrospective studies of Chamberlain and colleagues suggested no or very limited activity of HU, irrespective of the WHO grade of treated tumors, putting a beneficial effect of HU into question (Chamberlain 2012). Hormonal agents, such as the estrogen receptor (ER) modulator tamoxifen and the progesterone receptor (PR) inhibitor mifepristone, have been evaluated to a lesser extent and were shown to have only low activity (Goodwin et al. 1993; Grunberg et al. 2006). The overall low activity of ER modulators was usually explained by the small percentage of meningiomas expressing ER (about 10 %), while two-thirds of the tumors are positive for PR. It has been speculated that the disappointing outcome of mifepristone treatment may relate in part to the fact that most of these studies contained mainly meningiomas of higher grade, known to exhibit only low-level expression of PR (Eichler et al. 2010). Some small clinical studies have explored therapies targeting RTKs, which are important for tumor cell proliferation, such as platelet-derived growth factor receptor (PDGFR) and epidermal growth factor receptor (EGFR) or for neovascularisation, i.e., vascular endothelial growth factor receptor (VEGFR). However, first results for targeting PDGFR with imatinib (Adams et al. 2013) or EGFR with erlotinib and gefitinib (Norden et al. 2009b) were disappointing. Thus, it appears necessary to establish better chemotherapy strategies for the treatment of aggressive high-grade, non-resectable or otherwise therapy-refractory meningiomas, respectively. It should be further emphasized that except for HU, no major efforts have been undertaken to distinguish the responses of high- and low-grade meningiomas. In other cancer entities, the absence or presence of tumor suppressors has been shown to affect treatment responses (Garnett et al. 2012), and large-scale pharmacogenomic studies often detected unexpected correlations between genotypes and chemosensitivity (Garnett et al.

13

J Cancer Res Clin Oncol

2012). However, this issue has not yet been addressed for meningiomas. Notably, the status of the neurofibromatosis type 2 (NF2) tumor suppressor gene, which is frequently altered in meningiomas, has not been taken into account as a potential determinant for meningioma chemosensitivity. The NF2 status is of significant clinical interest, since preservation or loss of this tumor suppressor divide patients into two groups of similar size and NF2 is connected to some intended targeted therapies at the molecular level. Moreover, activity of several RTKs is regulated by the NF2-encoded protein merlin in meningioma cells (Curto et al. 2007; Fraenzer et al. 2003), as well as several enzyme cascades transmitting RTK signals to the nucleus (Zhou and Hanemann 2012). While in the past single drugs had sporadically been tested with selected meningioma cell lines or primary cultures, we intended to generate more comprehensive data by applying a standard panel of four cell lines, representing different subtypes and grades of meningioma. In the present study, we thus present results of an in vitro chemosensitivity screening of different meningioma cell lines including syngenic cells differing in the expression level of merlin. The drugs tested cover a wide spectrum of modes of action including drugs already used in clinical studies (see above). Moreover, given the fact that meningiomas arise in older patients with a high probability of concomitant cardiovascular diseases, we included the angiotensin II receptor type 1 (AT1) antagonist losartan, which has recently been described to inhibit the growth of breast cancer cells in preclinical models (Chen et al. 2013), and the calcium channel blocker verapamil, which is known to block drug efflux pumps expressed in meningiomas (Demeule et al. 2001; Andersson et al. 2004). We further tested the antidiabetic drug metformin, which seems to exert tumor-inhibiting effects not only at the systemic level by decreasing insulinemia, but also directly in tumor cells by influencing the mammalian target of rapamycin (mTOR) pathway which is activated in meningiomas (Loubiere et al. 2013; Pachow et al. 2013). There is a body of knowledge regarding achievable serum concentrations and tolerability for losartan, verapamil and metformin. Most drugs excreted cytostatic/cytotoxic effects only at concentrations far beyond the reachable plasma levels, with the exception of HU, which was moderately active. NF2/merlindeficiency in syngenic cell line pairs was associated with enhanced responses to HU.

Materials and methods Cell culture The h-tert immortalized benign meningioma cell line BenMen-1 (Puttmann et al. 2005), derived from a

J Cancer Res Clin Oncol

meningothelial meningioma (WHO grade I), was kindly provided by Werner Paulus (Neuropathology, University of Muenster, Germany), while HBL52, derived from a transitional meningioma (WHO grade I), was purchased from Cell Lines Services (Heidelberg, Germany). The line IOMM-Lee, which had originally been derived from an intraosseous malignant meningioma (WHO grade III), was kindly provided by David H. Gutmann (Department of Neurology, Washington University School of Medicine, St. Louis, MO, USA). KT21-MG meningioma cells, derived from a malignant meningioma (WHO grade III) (Tanaka et al. 1989), and SF4068 cells, derived from a benign meningioma (Cuevas et al. 2005), were obtained from Anita Lal (Brain Tumor Research Center, University of California, San Francisco, CA, USA). The syngenic cell lines SF4068-shNF2 and SF4068-shCon were obtained by lentiviral transduction of SF4068 cells with an shRNA construct targeting NF2 or a corresponding control construct (both kindly provided by Dr. Helen Morrison, Leibniz Institute for Age Research, FLI, Jena). Transduced cells were selected by FACS sorting based on a turbo-GFP marker, which was also encoded by the used vectors. The syngenic benign meningioma cell lines Men-shNF2 and Men-shCon, as well as the immortalized syngenic arachnoidal cell lines AC-shNF2 and AC-shCon were kindly provided by Anita Lal. The expected difference in merlin expression between wtNF2 and shNF2 cells was monitored by Western blotting. These syngenic cell lines served as a model to evaluate a potential influence of NF2 status on chemosensitivity, since NF2 knockdown simulates functional loss of the gene in human meningiomas by chromosomal losses, mutation and gene silencing. All cells were cultured in high-glucose DMEM, supplemented with 10 % fetal calf serum (FCS) and penicillin/ streptomycin. The identity of the cell lines was analyzed using the AmpFSTR kit and the software GeneMapper ID v3.2 from Applied Biosystems (ABI, Foster City, CA, USA). While IOMM-Lee and BenMen-1 exhibit similar and strong expression of merlin, KT21-MG cells do not express merlin (Striedinger et al. 2008). Drugs Metformin (Metformin Hydrochloride, stock solution 320 mM in aqua dest.), HU (stock solution 437 mM in DMSO), verapamil (verapamil hydrochloride; stock solution 100 mM in aqua dest.), temozolomide (stock solution 102 mM in DMSO) and mifepristone (stock solution 92 mM in DMSO) were all from Sigma-Aldrich, Taufkirchen, Germany. Other drugs evaluated were losartan (Fluka, Buchs, Switzerland; stock solution 86 mM in DMSO), erlotinib (Tarceva 25 mg from Roche, Basel, Switzerland; stock solution 45 mM in DMSO), and tamoxifen

(Nolvadex 20 mg from AstraZeneca, London, Great Britain; stock solution 4 mM in DMSO). MTT and BrdU assays Cells seeded into microtiter plates (1,000/well for 72-h experiments and 2,000/well for 48-h experiments, exception IOMM-Lee cells: 800/well for 48-h and 1,500/well for 72-h experiments) were allowed to attach for 24 h and then treated for 72 h (verapamil and metformin) or 48 h (all other drugs) with the final drug concentrations indicated in the Figures prior to standard microtiter tetrazolium (MTT) assays, which were performed as described previously (Andrae et al. 2012). The mean optical densities were normalized to controls treated with the pure solvent, and results were expressed as percentage of viable cells for a given drug concentration. To assess the radiosensitizing properties of the drugs, in some experiments X-rays (5 Gy) were applied after the first 24 h of drug application, using a Gulmay-D3225 machine (Gulmay Inc, GA, USA). For each batch of cells, 24 wells were used and a minimum of three repetitions with independent cell batches (passages). For BrdU assays to assess the rate of DNA synthesis, seeding conditions and drug treatment were identical to the above-described MTT assays. BrdU assays were performed using a kit from Roche Applied Sciences (Mannheim, Germany) according to the manufacturer’s instructions, applying a BrdU incubation time of 18 h. Nucleosome liberation assay Since MTT assay does not clearly discriminate between anti-proliferative, cytotoxic and metabolic effects of the drugs, which may all decrease the reduction rate of the tetrazolium compound, for selected drugs/conditions, cell death was monitored by the Cell Death Detection ELISAPlus (Roche Applied Sciences, Mannheim, Germany). The ELISA was performed basically according to the manufacturer’s instructions, but adherent cells in 96-well plates (a minimum of three independent cell batches) were lysed in a volume of only 70 µl in order to increase the concentration of free nucleosomes occurring in the cytoplasm of dying cells. Results were expressed as fold-increase of the ELISA signal, relative to wells treated with the solvent only. They specifically represent cytoplasmic free nucleosomes of dying cells, which were still adherent to the wells after drug exposure. Parallel measurements of the cell culture supernatants were not performed. Methylation‑specific PCR of the MGMT promoter About 0.5 µg genomic DNA of the cell lines was treated with the bisulfite reaction using the EZ DNA Methylation

13

Fig. 1  Effects of the DNA alkylating drug temozolomide on menigioma cell lines. a MTT assay shows a significant growth reduction only in IOMM-Lee cells treated with high temozolomide concentration (20 µM; p