Docetaxel targets aggressive methylation profiles and serves as a radiosensitizer in high-risk meningiomas

Abstract

Background

Meningioma is the most common primary intracranial tumor in adults. A subset of these tumors recur and invade the brain, even after surgery and radiation, resulting in significant disability. There is currently no standard-of-care chemotherapy for meningiomas. As genomic DNA methylation profiling can prognostically stratify these lesions, we sought to determine whether any existing chemotherapies might be effective against meningiomas with high-risk methylation profiles.

Methods

A previously published dataset of meningioma methylation profiles was used to screen for clinically significant CpG methylation events and associated cellular pathways. Based on these results, patient-derived meningioma cell lines were used to test candidate drugs in vitro and in vivo, including efficacy in conjunction with radiotherapy.

Results

We identified 981 genes for which methylation of mapped CpG sites was related to progression-free survival in meningiomas. Associated molecular pathways were cross-referenced with FDA-approved cancer drugs, which nominated Docetaxel as a promising candidate for further preclinical analyses. Docetaxel arrested growth in 17 meningioma cell sources, representing all tumor grades, with a clinically favorable IC₅₀ values ranging from 0.3 nM to 10.7 mM. The inhibitory effects of this medication scaled with tumor doubling time, with maximal benefit in fast-growing lesions. The combination of Docetaxel and radiation therapy increased markers of apoptosis and double-stranded DNA breaks, and extended the survival of mice engrafted with meningioma cells relative to either modality alone.

Conclusions

Global patterns of DNA methylation may be informative for the selection of chemotherapies against meningiomas, and existing drugs may enhance radiation sensitivity in high-risk cases.

Meningioma is the most common primary intracranial neoplasm in adults, with an incidence of nearly 35 000 per year in the United States.¹ The risk of meningiomas increases with age (especially after 65 years), disproportionately strikes women and blacks, and the overall prevalence has been rising over the past 20 years.¹ Meningiomas are classified by the World Health Organization (WHO) into 3 grades, wherein nearly 80% are grade 1, 18% are grade 2, and 2% are grade 3.^1–3 However, this grading scheme does not always accurately predict actual patient outcomes, as many meningiomas recur and cause considerable morbidity, especially if they are located in surgically challenging locations.⁴ Even with radiotherapy, the 5-year recurrence rates for high-grade meningiomas are 50% (grade 2) and 90% (grade 3); while 10-year overall survival is 53% and nearly 0%, respectively.⁵

Because meningiomas arise from the membranes covering the brain, and not the brain itself, they exist outside the blood-brain barrier. Though a subset of these lesions exhibits parenchymal invasion, this typically occurs via contiguous extensions of the main tumor, unlike diffuse gliomas that infiltrate on a single-cell basis. These favorable biological features should make meningiomas amenable to systemic therapies, however, the standard of care remains maximum safe resection, with additional radiotherapy for tumors deemed at higher risk of regrowth due to residual disease and/or WHO grade.⁶ While clinical trials have evaluated a wide range of drugs against meningiomas, including temozolomide, hydroxyurea, irinotecan, octreotide, everolimus, receptor tyrosine kinase inhibitors, angiogenesis inhibitors, and hormone inhibitors, among others, results have been disappointing thus far.^7–23

Notably, among drugs that have been evaluated in previous clinical trials, most were cytostatic (rather than cytotoxic), and most were tested as monotherapies, not concurrent with radiation. Although ongoing trials are evaluating new strategies like histone deacetylase inhibitors and immunotherapy,⁵ many established systemic therapies, including FDA-approved cytotoxic drugs that are known radiosensitizers in other cancers, have not been thoroughly explored in meningiomas.

An emerging development in molecular diagnostics for meningiomas is the use of genomic DNA methylation profiling, which has been shown to improve the prognostic stratification of meningiomas beyond what traditional WHO histologic grading can provide.^24–26 It remains unclear if specific methylation signatures are sufficient to drive aggressive behavior in meningiomas,²⁷ or rather represent a marker for tumors that harbor additional causative molecular features. In either case, the association of individual gene methylation events with recurrence suggests they may play a role in the underlying tumor biology, and could thus be informative in the selection of appropriate targeted treatment regimens. To date, methylation profiling has not been leveraged to identify systemic therapies that might be effective against meningiomas. Herein, we describe the repurposing of chemotherapies against high-risk meningiomas, based on tumor DNA methylation signatures that associate with poor patient outcomes.

Materials and Methods

Analysis of Previously Reported Methylation Data

Raw methylation data files (Illumina 450k) from 497 meningioma patients were sourced from a cohort previously published by Sahm et al. (Supplementary Table S1).²⁴ Four patients were excluded due to corrupted and/or duplicated data; and methylation analysis was completed on 210 cases that included progression-free survival data. A three-step method was applied to filter CpG methylation probes for quality assurance. First, probes with a detection P > .01 were excluded from downstream analyses. Next, single-nucleotide polymorphisms located inside the probe body or at the nucleotide extension were removed from analyses using a minor allele frequency cutoff of 0.05. Finally, through principal component analysis plots of methylation beta estimates, biological sex was determined to be the greatest source of variation in the data, which was expected since meningiomas tend to be more aggressive in men versus women. Thus, probes located on the sex chromosomes were removed and the remaining 463 469 CpG methylation probes were used for statistical analyses. The Bioconductor package, “minfi” version 3.8, was used to preprocess methylation files including normalization using the stratified quantile method. Probe beta-values (ranging 0–1 with higher values indicating higher methylation) were used for downstream analysis. We used the Illumina 450 k human annotation for methylation arrays as part of the “IlluminaHumanMethylation450kanno.ilmn12.hg”, version 3.8, package in R. All statistical analyses were performed in R, version 3.5.1, on Quest, Northwestern University’s high-performance computing cluster.

The primary outcome was progression-free survival (PFS), defined as the time from surgery to the date of progression (recurrence) in months. A Cox proportional hazards model was used to determine the association between PFS and individual CpG methylation sites (probes). While current literature has determined that age may be a confounder of the association between methylation and PFS, Cox models were not adjusted for age or other clinical variables, as this data was unavailable for many cases. To combat the multiple testing burden associated with high-dimensional data analysis, P values from the per-CpG-site Cox models were corrected for multiple comparisons according to the Benjamini and Hochberg false discovery rate (FDR) approach. FDR-adjusted P < .05 was considered statistically significant.

Pathway analysis was conducted using Reactome (version 68),²⁸ based on a gene list generated from the statistically significant markers identified in the survival analysis results (Supplementary Table S2). Reactome uses a hypergeometric distribution for significance estimation,²⁹ adjusting for multiple testing using the Benjamini and Hochberg FDR approach with a significance threshold of 0.05.²⁹ Based on the results of these analyses, cancer-specific drugs were identified using the Reactome FI app in Cytoscape, version 3.7.1,³⁰ which displays medications that target specific pathways. We focused specifically on pathways identified using the Reactome database.

Analysis of In-House DNA Methylation Data

Genome-wide DNA methylation profiles were acquired for all primary patient cell cultures that underwent therapeutic evaluation, as well as the cell lines IOMM-Lee and CH157. Data was collected using the Infinium MethylationEPIC BeadChip kit (Illumina), which assesses methylation intensities at over 850 000 sites across the genome. Raw intensity files (*.idat) were processed to obtain ratios of methylated probe intensities and total probe intensities (beta-values) using the Bioconductor package, “minfi” version 3.8. Data was normalized preprocessed, and filtered in a similar manner as the Sahm et al. cohort, as described above. Methylation status was interrogated at sites found to be associated with PFS in the Sahm cohort, and using these probes, samples underwent clustering based on Euclidian distance. The DNA methylation subgroup was identified according to a recently published paradigm (see Table 1).²⁶

RNA-Sequencing Analysis

Primary meningioma cell cultures underwent transcriptional analysis based on RNA-sequencing, and data was also collected for the commercial cell lines CH157 and IOMM-Lee. RNA and DNA were isolated using the All Prep RNA/DNA Mini kit (Qiagen; cat #80204), after tissue lysis and homogenization steps (extracted DNA was used for methylation analysis, as described above). Sample preparation and sequencing were performed at the Northwestern University NUSeq Core Facility. RNA was quantified using a Qubit fluorometer, and fragment size was assessed using an Agilent Bioanalyzer 2100. The TruSeq Stranded Total RNA Library Preparation Kit (Illumina) was used to prepare sequencing libraries from total RNA. This included Ribo-Zero rRNA Removal, cDNA synthesis, 3′ end adenylation, adapter ligation, library PCR amplification, and validation steps. Libraries were then sequenced on the Illumina platform (HiSeq 4000), using single-end, 50 base pair reads. The resulting data underwent analysis on Quest, Northwestern University’s high-performance computing cluster. Adapters were removed and low-quality bases were trimmed. Reads underwent pseudo alignment to Ensembl v96 using Kallisto (v0.46.1),³¹ followed by postprocessing analysis in R using DEseq2.^32,33 Expression of specific transcripts was considered after variance stabilizing transformation was performed. A curated list of 139 genes associated with the “G alpha (s) signaling events” pathway was obtained from Reactome (https://reactome.org/PathwayBrowser), and samples were clustered according to the expression of this list. Genes with no variance across all samples were removed.

Tumor Cell Sources and Cultures

Deidentified patient-derived cell lines were developed from freshly resected meningiomas through the Northwestern Nervous System Tumor Bank, under the auspices of a Northwestern University Institutional Review Board-approved protocol (#00095863) (Table 1). Tissues were minced into small pieces and dissociated with papain (Worthington Biochemical Corporation, #LK003150) as previously described.³⁴ Cells were then cultured in complete media containing 10% fetal bovine serum (FBS) with antibiotic/antimycotic added, and low-passage stocks were cryopreserved. IOMM-Lee and CH157 cells were obtained from the American Type Culture Collection.

Meningioma cell lines were propagated as monolayers on cell culture-treated vessels in a complete medium consisting of Dulbecco’s Modified Eagle Medium: Nutrient Mixture F-12 (DMEM/F-12, Gibco), 10% FBS (Cytiva), and sodium pyruvate (Gibco). Half the medium on each culture was replaced with fresh medium every 3–4 days. Culture passages were performed using Trypsin-EDTA (0.05%, Gibco, #25300-054). All cells were incubated at 37°C in a humidified atmosphere containing 95% O₂ and 5% CO₂. Cells were regularly authenticated by short-tandem repeat analyses.

To enable cells for in vivo bioluminescence imaging (BLI), lentiviral vectors containing firefly luciferase were generated as previously described,³⁵ and were used to transduce CH157 and IOMM-Lee cells. The cells were screened in vitro for transduction efficiency and luciferase activity by treatment with luciferin (d-luciferin potassium salt, 150 mg/kg, Gold Biotechnology), and analyzed for luminescence using a Synergy 2 Microplate Reader (BioTek Instruments).

Drugs and Dose–Response Curves

Cells were dissociated in Trypsin-EDTA (0.05%, Gibco, #25300-054), counted with a Countess II FL automated cell counter (Invitrogen #AMQAF1000), and plated at 1000 cells/well in 96-well tissue culture plates. After a 24-hour incubation, serially diluted drugs were added to each well. Relative cell numbers were determined using CellTiter-Glo 2.0 Assay (Promega, #G9242) against ATP standards (Thermo Scientific, #R0441). Dose–response curves were generated and the half-maximal inhibitor concentration (IC₅₀), 50% of cancer cell elimination in comparison to control (GI₅₀), and maximum efficacy (E_max) were calculated using GraphPad Prism 9 software. In samples for which E_max was > 0.5, a GI₅₀ value was not calculated. Docetaxel was obtained from Alfa Aesar (#J60174MC). Raloxifene hydrochloride was obtained from Cayman Chemical (#10011620).

Quantification of Apoptosis via Annexin V-FITC Staining

IOMM-Lee, CH157, and NU1855 cells were treated with 0, 0.1, 1.0, and 10 nM concentrations of docetaxel, one day after being plated in 6-well plate. After 24 hours, the treated cells were collected and apoptosis was detected using Annexin V-FITC/PI (propidium iodide) staining (ab14085, abcam). Apoptotic cells were analyzed using BD FACSymphony flow cytometer and FlowJo software (Version 10.7.0, FlowJo LLC).

Clonogenic Assays

Cells were seeded into 6-well tissue culture plates and allowed to adhere for 24 hours. Attached cells were treated with various doses of drugs (ranging from 0 nm to 2 nm) 4 hours before irradiation (2 Gy). Radiation was delivered using a ¹³⁷Cs source (Mark I, model 68A irradiator, JL Shepherd & Associates). After 2 weeks of incubation, cells were fixed and stained with a mixture of 6.0% glutaraldehyde and 0.5% crystal violet. The number of colonies with >50 cells was determined. Plating efficiencies were calculated as the percentage of the number of colonies formed to the number of cells seeded. Synergistic effects between drugs and radiotherapy were determined using Combenefit software.³⁶

Foci-Based Gamma-H2AX Assay

Cells were seeded on 8-well Nunc Lab-Tek II chamber slides (Thermo Scientific, #154534) and allowed to adhere for 24 hours. Attached cells were treated with 2 nM Docetaxel 4 hours before irradiation with a ¹³⁷Cs source. Twenty-four hours after treatment, cells were fixed in 4% paraformaldehyde for 40 min, washed with PBS and permeabilized with 0.3% Triton X -100 solution for 30 min. The cells were then incubated with a phospho-histone H2A.X antibody (Ser-139, Cell Signaling Technology #2577) diluted in immunofluorescence antibody dilution buffer (Cell Signaling Technology #12378) overnight at 4°C. Phosphorylation of the Ser-139 residue of the histone variant H2AX, forms “γH2AX”, which is an established marker of DNA damage.³⁷ Subsequently, washed cells were incubated with Rhodamine Red-X (RRX) AffiniPure Goat Anti-Rabbit IgG (Jackson ImmunoResearch Laboratories, #111-295-003) in PBS for 1 hour. After washing with PBS, a coverslip was mounted on each slide with ProLong Gold Antifade Mountant with 4′,6-diamindino-2-phenylindole (Thermo Scientific, # P36931). Slides were imaged using EVOS FL imaging system (Life Technologies, #AMF4300). Number of cells with > 12 foci was determined. The percentage of positive cells was statistically compared between each treatment group (sham + vehicle; sham + Docetaxel; 2 Gy RT+ Vehicle; 2 Gy RT + Docetaxel).

In Vivo Studies

Six-week-old athymic nude female mice (NCr-Foxn1^nu) were purchased from Envigo. All mice were housed under aseptic conditions, which included filtered air and sterilized food, water, bedding, and cages. The Northwestern University Institutional Animal Care and Use Committee approved all animal protocols, and all animal care was in accordance with institution guidelines. CH157 or IOMM-Lee cells (1.5 × 10⁵ cells) were implanted under the skull in the right frontal subdural region approximately 2 mm posterior to the bregma and 1 mm to the right of midline and to a depth of 3 mm from the bone surface. Three days after engraftment, mice were randomized to three treatment groups and vehicle control, as discussed in the Results section. Engrafted mice were imaged 1–2 times weekly by BLI, as previously described.³⁸ BLI was performed by subcutaneous injection with 150 mg/kg d-luciferin potassium (Gold Biotechnology) in mice anesthetized via 2.5% isoflurane. Each mouse was imaged 10 minutes after luciferin injection on a Lago imaging system (Spectral Instruments Imaging) with Aura data-acquisition software (Spectral Instruments Imaging). Intracranial signal intensities were quantified within regions of interest defined by the Living Image software. Animals were euthanized by CO₂ asphyxiation, followed by cervical dislocation when they became moribund (>20% weight loss, neurologic symptoms, or evidence of pain/distress), and survival time was recorded. Survival analyses were performed using GraphPad Prism 9 software.

Statistical Analyses

Methylation data were analyzed using R, as described above. All other data were processed in spreadsheets (Microsoft Excel) and analyzed using Prism 8 (GraphPad Software). P values were calculated as described in the figure legends with either Student’s t-test, ANOVA, or log-rank analysis, as appropriate.

Results

Methylation Profile Analysis Nominates Docetaxel and Raloxifene as Drug Candidates for Meningioma

DNA methylation analysis was performed on 210 cases that included PFS data, among a cohort of 497 meningiomas previously published by Sahm et al. (Supplementary Table S1; see methods above).²⁴ Using Cox models, we identified CpG methylation probes that exhibited a significant association with PFS, with either a positive (higher methylation was associated with shorter PFS) or negative (higher methylation was associated with longer PFS) hazard ratio. Each probe was mapped to the nearest gene, and many genes subsequently harbored a mixture of positive and negative hazard ratio sites. We thus filtered this list to include only genes in which at least 40% of all mapped CpG sites were significantly associated with PFS in a consistent direction, either positive or negative (N = 981; Supplementary Table S2).

From this resulting gene set, we next investigated pathway enrichment using the Reactome database.²⁹ Notably, the only significant pathways after FDR-correction were “Olfactory Signaling” and “G Alpha (S) Signaling Events” (Supplementary Table S3), and these terms were considered as candidate terms for drug discovery. The Cytoscape/Reactome FI tool, which draws upon multiple cancer drug databases established prior to 2018,^39,40 was used to elucidate current FDA-approved medications associated with the identified pathways (Supplementary Figure S1). Though we were unable to find candidates targeting “Olfactory Signaling,” several medications were identified related to “G Alpha (S) Signaling Events” (Supplementary Figure S2). Among these drugs, Docetaxel and Raloxifene hydrochloride were found to target multiple downstream molecules that are key to pathway function, and we thus selected these for further preclinical evaluation.

Docetaxel Inhibits Meningioma Growth In Vitro

Based on in silico results, the efficacy of Docetaxel and Raloxifene was tested on 17 patient-derived meningioma cell cultures, including cells developed from 15 in-house meningiomas (Table 1). To compare the molecular background of our cohort to samples reported in the discovery study, we performed DNA methylation profiling, focusing specifically on probes associated with PFS in our initial analysis. Radiographic follow-up was limited among our in-house cases, which prevented a direct correlation of PFS to probe intensity (as done in the Sahm cohort). However, methylation clustering revealed a distinct clade of aggressive meningiomas, which included all cases with hypermitotic signatures (NU02188, NU02098, and NU02187; Table 1), as well as the only grade 3 tumor (NU01855)(Supplementary Figure S3). This suggested that probes used for drug discovery in our initial analysis were also sufficient to identify aggressive cases in our therapeutic testing cohort.

With the administration of Docetaxel and Raloxifene, the IC₅₀, GI₅₀, and E_max were calculated for each culture based on dose-response curves. In most cell lines, IC₅₀ and GI₅₀ were achieved at very low Docetaxel concentrations (IC₅₀ ranging from 0.3 to 10.7 nM, GI₅₀ from 0.96 to 92.62 nM) (Figure 1A-B). However, E_max was higher than 0.5 for six cell lines (NU01855, NU01985, NU01989, NU02003, NU02242, and NU02187), suggesting intertumoral heterogeneity in therapy response. With the exception of a single case (NU02188), all members of the aggressive methylation cluster and an adjacent clade showed robust response to Docetaxel (IC₅₀ < 1.0 nM)(Supplementary Figure S3). Raloxifene was less effective than Docetaxel at inhibiting meningioma growth, with an IC₅₀ ranging from 10.9 µM to 36.3 µM and GI₅₀ values from 12.3 to 37.2 µM (Supplementary Figure S4). While E_max values for Raloxifene were very low, total growth inhibition was only achievable at high drug concentrations, which were not deemed realistic in the clinical setting. Notably, the IC₅₀ of Docetaxel against the meningioma cells were substantially lower relative to the established values for most cancer cell lines in the Genomics of Drug Sensitivity in Cancer datasets (GDSC1 and GDSC2)⁴¹(Figure 1C-D). When grouped by cancer type, the IC₅₀ for the meningioma cells was consistently lower than for other types of cancer, including cancers in which Docetaxel has been approved as chemotherapy.

To understand the differences in response of meningiomas to these drugs, we assessed correlations between GI₅₀ values and doubling times for untreated cells. For each cell culture, doubling time was calculated based on the log phase of its growth curve (Supplementary Figure S5). We found that doubling times significantly correlated with GI₅₀ values for Docetaxel (Spearman r = 0.63, P = .044), but not Raloxifene (Spearman r = 0.36, P = .17) (Figure 1E), consistent with the established propensity of taxanes for dividing cells via disruption of microtubule function. Notably, β3-tubulin has been reported to be a marker of resistance to Docetaxel in prostate, gastric and breast cancers,⁴² but in this study, β3-tubulin expression, as measured by western blot, did not correspond to meningioma cell IC₅₀ or GI₅₀ values (Figure 1F). As docetaxel was selected based on an association of PFS with the “G alpha (s) signaling events” methylation signature, we next investigated transcriptional patterns of genes in this pathway. Expression clustering of these genes using RNA-seq data revealed three groups of cell culture samples (Supplementary Figure S6A), marked by enrichment of low-grade cases (7 of 9; cluster III), higher-grade cases (3 of 5; cluster II), and a final clade that housed both cell lines and a recurrent primary meningioma culture (cluster I). Notably, this latter group exhibited significantly lower E_max of docetaxel treatment than the other 2 groups (P = .004594; Supplementary Figure S6B), suggesting that expression of genes associated with G alpha (s) may predict response to this medication.

Docetaxel Causes Meningioma Cells to Undergo Apoptosis

To understand the mechanism by which Docetaxel inhibits meningioma cell growth, we performed western blots with lysates of cells treated with various concentrations of this drug, analyzing for apoptosis markers. The presence of these signals would suggest a cytotoxic effect of this medication, in contrast to the cytostatic effect observed in many alternative chemotherapies. Docetaxel triggered apoptotic cell death in the commonly used meningioma cell lines CH157 and IOMM-Lee, as demonstrated by the increased level of cleaved PARP and cleaved caspase 3 (Figure 2A). These findings were reproduced in representative patient-derived meningioma cells, which detected elevated cleaved caspase 9 (Figure 2B) as well as cleaved PARP (Figure 2C). To confirm these results, we performed Annexin V-FITC staining to label cells undergoing apoptosis, followed by flow cytometry. At elevated concentrations of Docetaxel we observed increases in early and late apoptotic cell numbers among the tested samples, including IOMM-Lee, CH157, and NU1855 (Supplementary Figure S7). Furthermore, we found that apoptosis events in patient-derived samples were associated with markers of DNA damage (NU02109 and NU02141), and an additive effect persisted even in presence of RT (Supplementary Figure S8). These results suggest that Docetaxel induces apoptosis in meningioma cells, likely via DNA damage, and may result in additional, additive effects with RT.

Docetaxel has Synergistic Effects with Radiotherapy in Meningioma Cells

Meningiomas with aggressive histological features often undergo adjuvant RT after resection, aiming to target the rapidly dividing remnant cells that underlie tumor recurrence. As Docetaxel also preferentially affects mitotically active cells, we next investigated if concomitant use of this drug could synergistically improve the therapeutic effect of radiation. We found that the apoptotic effects of Docetaxel on IOMM-Lee cells were further enhanced by the addition of radiation, in an additive manner, evidenced by elevated levels of cleaved PARP (Figure 3A). Furthermore, when IOMM-Lee and CH157 cells were treated with 2 nM Docetaxel and/ or 2 Gy RT, we observed significant increases in double-stranded DNA damage, demonstrated by an increase in cells with γ-H2AX formation³⁷ (Figure 3B). Notably, combining both treatments produced a significantly stronger effect (Figure 3C). Using Combenefit software, synergy was shown between low concentrations of Docetaxel (1–2 nM) and a single low dose of RT (2 Gy) (Figure 3D).

Clonogenic assays supported these synergistic effects (Figure 3E). IOMM-Lee cells were treated with various doses of Docetaxel, and either 0 or 2 Gy of RT. We observed decreases in colony formation with monotherapeutic Docetaxel at the lowest concentration tested (0.31 nM), and we observed approximately 50% fewer colonies relative to control when a concentration of 1nM was applied (Supplementary Figure S9A). Synergy with RT was detected at Docetaxel concentrations as low as 0.0625 nM (Figure 3F, Supplementary Figure S9B). Notably, we found that treatment with 2 Gy RT and 1 nM Docetaxel decreased colony formation by 90%. This translates to 8.1 pg/ml, which is well below the peak serum level of 1 µg/ml achieved among most subjects receiving 40 mg/m² of Docetaxel in a previous human study.⁴³

Docetaxel Improves Mouse Survival in an Orthotopic Meningioma Model

To evaluate the therapeutic potential of combined Docetaxel and RT in vivo, we developed 2 orthotopic intracranial models using the CH157 and IOMM-Lee cells. These cell lines were selected as most patient-derived meningioma cells do not propagate in vivo in a manner that would be conducive to testing therapeutics. Three days after engraftment, animals were randomized among four groups: (i) vehicle control; (ii) single dose of irradiation (7.5 Gy); (iii) four doses of Docetaxel (7.5 mg/kg) over 2 weeks; (iv) RT + Docetaxel (Figure 4A). This Docetaxel regimen was the maximum tolerated dose in tumor-free nude mice, as indicated by body weight (Supplementary Figure S10). In the IOMM-Lee model, the median survival in untreated mice was 15 days, while those receiving RT or Docetaxel survived a median of 16 days and 17 days respectively. Notably, mice that received both treatments survived significantly longer, with a median duration of 29 days (P < .001 relative to the control; P = .003 relative to RT; P = .007 relative to Docetaxel) (Figure 4B). The tumor size, as measured by BLI at the completion of treatment, indicated the combinatorial strategy had resulted in a trend towards reduced tumor growth (Figure 4C). A similar effect was noted in a second meningioma model using CH157 cells, though in this case, the radiation group fared relatively better. The median survival in untreated mice was 17 days, 26 days for those treated with RT, 18 days for those treated with Docetaxel, and 27 days for the combinatorial treatment (P < .0001 relative to the control; P < .0001 relative to Docetaxel)(Figure 4D). Similar to the IOMM-Lee model, the tumor size as measured by BLI at the completion of treatment also indicated the combinatorial strategy may suppress tumor growth (Figure 4E). Comparison of Kaplan Meier curves among the 2 cell lines suggested relative less synergy in dual treatment among animals injected with CH157 (Figure 4B vs D).

Given the sensitivity of both IOMM-Lee and CH157 to Docetaxel in vitro (Figure 1), it was surprising to observe differing synergistic responses in vivo. To explore this further, the expression of genes known to confer taxane resistance in other cancers were evaluated. Using quantitative PCR, we found that the drug efflux gene ABCB1, and drug metabolism genes CYP1B1 and CYP3A4, were expressed at higher levels in CH157 cells than in IOMM-Lee cells (Figure 4F), whereas TUBB3 (β3-tubulin) and CYP2C8 were higher in IOMM-Lee cells (Supplementary Figure S11). Among the taxane-resistant genes that were differentially expressed between CH157 and IOMM-Lee cells, none were significantly increased among previously reported patient-derived grade 2–3 meningiomas, compared to grade 1 tumors (Supplementary Figure S12).^44–46 In fact, the general pattern seemed to be a downregulation of such genes as tumor grade increased.

Discussion

Meningiomas have a higher incidence than adult-type diffuse gliomas,¹ are more accessible to systemic therapies because they are located outside the blood-brain barrier, and are often supplied by branches of the external carotid arteries. Despite these potential therapeutic advantages, no systemic medications have been identified that demonstrate efficacy against recurrent meningiomas. However, there remain many therapies that have not been tested against these tumors in a preclinical setting. Our data, based on an analysis of methylated genes associated with patient outcomes in meningiomas, indicates that Docetaxel, a taxane that has been FDA approved to treat cancers since 1998,⁴⁷ may be a useful adjuvant therapy in many of these tumors. Our in silico results are corroborated by systematic in vitro and in vivo evaluation, using new and preexisting patient-derived meningioma cells that encompass all WHO grades and methylation subgroups. Finally, we show that this chemotherapy can exert synergistic effects with radiation, and may thus bolster the impact of current standard-of-care approaches for recurrent and/ or aggressive lesions.

Like other taxanes, Docetaxel disrupts microtubule function, thereby functioning as a mitotic inhibitor. This is consistent with our data showing a correlation between GI50 and doubling time in the tested meningioma cell cultures. Docetaxel also has a variety of other anticancer effects, such as inhibition of the anti-apoptotic effects of bcl-2 and bcl-xL,^48,49 and is well known to enhance RT.^50,51 Given that RT is a mainstay of treatment in recurrent and high-risk meningiomas, concurrent administration of radiosensitizing therapeutics (especially cytotoxic therapeutics rather than cytostatic) holds particular promise to improve outcomes of this disease. Yet most preclinical and clinical trials of new drugs against meningiomas have been monotherapeutic, have been cytostatic, and have not been tested in conjunction with RT. Although the side effects of Docetaxel include hair loss, neutropenia, and liver damage, an abbreviated cycle of four doses, concurrent with RT, could minimize these effects while still being efficacious—especially since Docetaxel showed consistently lower IC₅₀ values in meningioma cells relative to other cancers. It may therefore be possible to reduce the overall dose in meningioma patients, although this would require formal testing in a clinical trial.

Docetaxel has been in regular clinical use for over 20 years, but very little has been published on its activity, or the activity of other taxanes, in meningiomas. One in vitro study showed that patient-derived meningioma cells were more sensitive to paclitaxel than cells from adult-type diffuse gliomas,⁵² while another showed that low nanomolar concentrations of paclitaxel-induced apoptosis.⁵³ Neither report tested paclitaxel in conjunction with RT or in vivo, and clinical data have likewise been scarce. In one case report, a grade 1 meningioma was noted to shrink after treatment with the combination of paclitaxel and the VEGF inhibitor bevacizumab.⁵⁴ Notably, these drugs were administered for treatment of breast cancer in this patient, and the meningioma findings were incidental. Our bioinformatics analysis also identified the potential efficacy of Raloxifene, an estrogen agonist, in meningiomas, however in vitro modeling failed to show a robust response. As both docetaxel and raloxifene appeared to be promising candidates based on DNA methylation profiles, it is possible that additional epigenetic, environmental, or genomic factors may conspire to render the latter drug a less favorable therapeutic. Indeed, previous studies investigating hormonal treatments in meningiomas have yielded mixed results, despite the well-established expression of hormonal receptors in these tumors.

Just as dosing regimens, pharmacokinetics, and side effects of Docetaxel are all well-characterized, mechanisms of cancer resistance to Docetaxel (and other taxanes) have also been explored in great detail (reviewed in⁵⁵). Such mechanisms include upregulation of drug-exporting genes like ABCB1, drug-metabolizing, cytochrome P450-encoding genes like CYP1B1, CYP3A4, and CYP2C8, and increased expression of TUBB3, encoding β3-tubulin. Based on our in vitro data, we were not anticipating that either CH157 or IOMM-Lee cells would be resistant to Docetaxel in vivo. However, CH157 did show mild resistance, whereas IOMM-Lee cells did not. This prompted a screening for differential expression of genes known to be associated with taxane resistance. Among the genes screened, ABCB1, CYP1B1, and CYP3A4 mRNA levels were 1-2 orders of magnitude higher in CH157 cells compared to IOMM-Lee cells, whereas ABCC5, ABCC10, CYP2C8, and TUBB3 were either similar in both cell types or higher in IOMM-Lee cells. This may suggest that ABCB1, CYP1B1, and/or CYP3A4 have more taxane-suppressive effects in vivo than in vitro, for reasons that remain unclear, however, this hypothesis is speculative. An area of future investigation could focus on whether these taxane resistance mechanisms are regulated differently within the brain versus in culture, and if so, identify the underlying mechanisms. Moreover, expression of taxane resistance genes was generally lower in patient meningiomas with increasing tumor grade. Thus, while screening for known taxane resistance genes would be a logical, even essential, aspect of any clinical trials of Docetaxel in meningiomas, our data suggest that Docetaxel might nonetheless be a viable therapeutic option for meningiomas in which radiosensitization is needed.

In sum, our data indicate that Docetaxel, which has decades of patient-based experience in other cancers, may have a beneficial role in the treatment of meningiomas that are behaving, or are expected to behave, in a more aggressive fashion. These data also align with the general initiative towards repurposing existing anticancer therapeutics, especially those, like Docetaxel, that now have generic drug status and are therefore more cost-effective.

Funding

National Institutes of Health (R01NS102669, R01NS117104, R01NS118039 to CH, P50CA221747); the Lou and Jean Malnati Brain Tumor Institute at Northwestern.

Acknowledgments

The authors thank the Northwestern Nervous System Tumor Bank, which is supported by the P50CA221747 SPORE for Translational Approaches to Brain Cancer.

Conflicts of interest: None of the authors have any conflicts of interest concerning the research in this manuscript.

Authorship: Conceptualization and funding: CDJ and CH. Data curation: MWY, ANT, WW, DS, LZ, KOS, DU, JW, KM, YW, AB, FS, KA, and JPC. Formal analysis: MWY, ANT, DU, JW, KM, YW, AB, and ABH. Writing original draft: ANT, DS, LZ, KOS, and CH. Editing manuscript: MWY, DS, STM, SS, RVL, AA, ABH, and CH.

References