Radio-sensitizing effect of MEK inhibition in glioblastoma in vitro and in vivo

Glioblastoma (GBM) is an incurable cancer type. New therapeutic options are investigated, including targeting the mitogen-activated protein kinase (MAPK) pathway using MEK inhibitors as radio-sensitizers. In this study, we investigated whether MEK inhibition via PD0325901 leads to radio-sensitization in experimental in vitro and in vivo models of GBM. In vitro, GBM8 multicellular spheroids were irradiated with 3 fractions of 2 Gy, during 5 consecutive days of incubation with either PD0325901 or MEK-162. In vivo, we combined PD0325901 with radiotherapy in the GBM8 orthotopic mouse model, tumor growth was measured weekly by bioluminescence imaging and overall survival and toxicity were assessed. Regrowth and viability of spheroids monitored until day 18, showed that both MEK inhibitors had an in vitro radio-sensitizing effect. In vivo, PD0325901 concentrations were relatively constant throughout multiple brain areas and temporal PD0325901-related adverse events such as dermatitis were observed in 4 out of 14 mice (29%). Mice that were treated with radiation alone or combined with PD0325901 had significantly better survival compared to vehicle (both P < 0.005), however, no significant interaction between PD0325901 MEK inhibition and irradiation was observed. The difference between the radiotherapy-enhancing effect of PD0325901 in vitro and in vivo urges further pharmacodynamic/pharmacokinetic investigation of PD0325901 and possibly other candidate MEK inhibitors.


Introduction
Glioblastoma (GBM) accounts for approximately 70-80% of all primary brain tumors among adults and is the most common and malignant type of glioma (Ostrom et al. 2020).
The annual incidence rate is 3.19 per 100.000 in the United States and several international studies have shown that the incidence rate is increasing (Ostrom et al. 2020;Korja et al. 2019;Lee et al. 2010;Gousias et al. 2009;Philips et al. 2018;Arora et al. 2009). Standard therapy for newly diagnosed patients consists of maximal safe surgical resection followed by radiotherapy and adjuvant and/or concomitant treatment with temozolomide (Stupp et al. 2005). The median survival for patients with GBM, after standard treatment, is 15 months with a relative survival estimate of 7.2% after 5 years (Stupp et al. 2009;Ostrom et al. 2020). The treatment options are limited and resistance to standard therapy accounts for the poor clinical outcome.
There is a great need for a new therapeutic approach next to standard therapy or after treatment failure. Such a new therapy should be aimed to sensitize GBM cells to radiation as radiotherapy is the most important component of the standard treatment (Stupp et al. 2005;Khosla 2016;Lu et al. 2020). A new form of therapy includes targeting the mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK) pathway (Narayan et al. 2018;Zhao and Adjei 2014;Lin et al. 2014;Studebaker et al. 2015;Chung et al. 2009). Recently, Essien et al. showed in vitro that the combination of MEK inhibition and histone deacetylation inhibition with radiation led to a decrease in GBM spheroid formation as well as a decrease in glioblastoma markers CD44, Nestin, and SOX2 (Essien et al. 2022). In non-smallcell lung cancer, the combination of MEK inhibition and mTOR inhibition enhanced the effect of radiation in vitro and in vivo (Kim et al. 2021). Also, Narayan et al. showed that MEK1/2 inhibitor MEK-162 act as a radio-sensitizer on patient-derived glioblastoma cells in vitro and in vivo (Narayan et al. 2018). The radio-sensitizing effect was attributed to a MEK1/2 inhibition-induced decrease in protein levels involved in DNA-damage sensing. Similar results have been observed by other researchers (Chung et al. 2013;Yan et al. 2007;Caunt et al. 2015).
Currently, there are four FDA-approved MEK inhibitors available in the clinic, including Trametinib, Cobimetinib, Selumutinib and Binimetinib (Tran and Cohen 2020). Several clinical trials are ongoing in which the effect of MEK inhibitors is studied, but GBM patients are only included in a small number of trials (e.g., NCT03363217; NCT02070549; NCT00866177; NCT03973918). Moreover, these clinical trials aimed to study the effect of MEK inhibition alone or in combination with another drug but not combined with radiotherapy (Catalanotti et al. 2013;Voon et al. 2022;Perreault et al. 2019). To our knowledge, no clinical trial was performed to study the sensitizing effect of a MEK inhibitor in combination with radiotherapy in GBM patients.
Other researchers previously compared the pharmacodynamics and pharmacokinetics of a panel of MEK inhibitors including Trametinib, Selumitinib, MEK-162 (Binimetinib), Pimasertib, and PD0325901 in the brain (Gooijer et al. 2018). From this panel of MEK inhibitors, PD0325901 stood out as showing the most efficient target inhibition in the brain, which indicated a capability of crossing the blood-brain barrier (BBB), facilitated by being a poor substrate for efflux pumps. Based on these results, PD0325901 seems to be a promising drug to combine with radiation in glioblastoma.
Given these more optimal BBB characteristics of PD0325901 over MEK162, we evaluated the in vitro and in vivo radio-sensitizing effect of the BBB-penetrating MEK-inhibitor PD0325901 alongside our earlier identified radio-sensitizing MEK-inhibitor, MEK162.

Cell line authentication
Authentication was assessed via the Infinium Methylation EPIC array by the Human Genomics Facility of the Genetic Laboratory of the Department of Internal Medicine (Erasmus Medical Center, Rotterdam, the Netherlands) and cells were regularly negatively tested for mycoplasma by www. micro biome. nl.

Chemicals
MEK-162 was purchased from Selleck Chemicals (Houston, TX, USA) and PD0325901 was purchased via MedChem Express (Monmouth, NJ, USA). Both chemicals were dissolved in DMSO and stored in − 80°C as 10 mM stock.

IC50 assays
To obtain a single-cell suspension, neurospheres were dissociated by incubation in Accutase (ThermoFisher Scientific, Waltham, MA, USA) for 5 min at 37°C followed by deactivation of the Accutase with NBM and re-suspended in complete NBM. Cells were plated at a cell density of 3000 cells/well in µClear ® 384-well F-bottom plates (#781,097, Greiner Bio-one, Kremsmünster, Austria). After 24 h spheroid formation, MEK-162 and PD0325901 were added in a serial dilution (5 pM-40 μM), using a Tecan D300e dispenser (Tecan Group Ltd, Männedorf, Switzerland). The cell viability was determined after 96 h, 72 h compound exposure, using CellTiter-Glo 3D luminescent Cell Viability Assay (Promega, Madison, WI, USA) according to the manufacturer's protocol. The relative light units (RLUs) were measured with a Tecan Infinite ® 200 reader using iControl 1.10 software, RLUs were normalized based on the DMSO controls. Bioassays were replicated at least three times.

Cell viability assays
Neurospheres were dissociated as described above. Cells were plated at a cell density of 750 cells/well in cellrepellent 96-well U-bottom plates (#650,970, Greiner Bio-one, Kremsmünster, Austria). These conditions ensured the formation of one spheroid per well during the first 72 h of incubation. After spheroid formation, a serial dilution of MEK-162 or PD0325901 was added via the Tecan D300e dispenser 1 h before irradiation. Radiation of the cell culture plates with drug-treated spheroids was performed at room temperature with a Cobalt-60 source at a dose rate of ~ 150 Gy/h (Gammacell 220 ® , Atomic Energy of Canada, Mississauga, ON, Canada). Irradiation was performed for 3 consecutive days and cells were exposed to the drugs until day 4. Every 3 to 4 days, NBM medium was refreshed and phase-contrast images were automatically taken by Leica DMI3000 microscope (Leica, Rijswijk, the Netherlands). Regrowth of neurospheroids was monitored until day 18 after which cell viability was measured with CellTiter-Glo 3D as described above.

Pharmacokinetics of PD0325901 in vivo
Ethical approval for studies was provided by the Animal Welfare Body (IVD) of the VU and VUMC (in Amsterdam) (AVD114002017841, protocol #841-NCH-18-06). All experiments meet ARRIVE guidelines and were performed under the European Community Council Directive (2010/63/EU) for laboratory animal care and the Dutch Law on animal experimentation. Subject Demographics: Female athymic nude mice, obtained via Envigo (6-8 weeks old; Horst, the Netherlands), were used to assess the blood-brain barrier (BBB) penetration of PD0325901. Randomization: after initial chemiluminescence detection, mice were randomly stratified into the experimental arms. Mice were injected with a PD0325901 dose of 5 mg/kg intravenous (i.v.) via the tail vein. To measure the BBB penetration of PD0325901, 3 mice were sacrificed 5 min after injection, 3 mice 60 min after injection and 3 mice 240 min after injection. Euthanasia Agents: Mice were euthanized with 20% Euthasol ® (AST Farma, Oudewater, the Netherlands), after which brains were collected and stored in liquid nitrogen upon analysis by LC-MS/MS. Previously it was shown in vivo, that PD0325901 can reach sufficient brain concentrations to inhibit phosphorylation of the MEK targets ERK1 and 2 (Gooijer et al. 2018). Power analysis: To estimate the optimal sample size of animal experiments, we used power analysis resulting in a group size of n = 7 mice per arm.

LC-MS/MS
To determine the plasma and brain concentration of PD0325901 in athymic nude mice, an adapted LC-MS/ MS method as previously described by Honeywell et al. was used (Honeywell et al. 2010). Each frozen mouse brain segment was weighted and subdivided into a minimum of 7 sections. These sections were weighed individually before being freeze-dried overnight as previously described by Amir et al. (Avan et al. 2013). Briefly, PD0325901 was extracted from the dried tissue by the addition of 5:1 acetonitrile:water. Each tissue sample was subsequently manually handed homogenized and then centrifuged (14000 g, 10 °C) with 1 µl of the supernatant being directly introduced onto the LC-MS/MS. A Dionex Ultimate 3000 LC with a prodigy ODS3 column, 5 µm, 150 × 3.1 mm (Phenomenex, Utrecht, the Netherlands) was used for chromatography and eluted at 200 μL/min with a tertiary mobile phase consisting of 20 mM ammonium acetate:acetonitrile: methanol. A turbo-spray ionization source and mass spectrometric positive multi-reactionmonitoring-mode (+ MRM) at an ion voltage of + 3500 V and a mass transition of 483.0/249.1 with a collision energy of 47 were used for detection. All other conditions were as reported previously, with PD0325901 eluting after 18.5 min (Honeywell et al. 2010). The absolute amount of PD0325901 was determined in the tissue reconstitution volume and was directly related to the weight of the tissue. To ensure complete extraction of PD0325901, the extracted tissue was dried again and extracted for a second time with acetonitrile:water, the second extraction demonstrated no evidence of PD0325901 indicating that the first extraction was greater than 99% of the total. For estimating the brain concentration, a density of 1.05 g/cm 3 was taken (DiResta et al. 1991).

In vivo validation
6-8-week-old female athymic nude mice were obtained via Envigo and housed according to guidelines and regulations of the VU University (AVD114002017841, protocol #841-NCH-19-10). After one week of acclimatization, 0.5 × 10 6 GBM8 cells which stably express Firefly Luciferase (Fluc) and mCherry, were stereotactically injected into the striatum (coordinates relative to the bregma; x− 2 mm; y0.5 mm; z− 3 mm) as previously described (Narayan et al. 2018;Brahm et al. 2020;Praxinos 2008). Mice received the following analgesia; at least 24 h pre-and postoperative carprofen (0.067 mg/mL) via their drinking water, half an hour before the intracranial surgery subcutaneous injection of Buprenorphine (0.05 mg/kg), and a droplet of Lidocaine (5 mg/mL) was applied onto and spread over the exposed periosteum as local analgesia. Mice were sedated using gas anesthesia (0.6L O 2 and 0.6L compressed air and 2.5% isoflurane). Tumor engraftment and growth were measured by imaging the (Fluc) activity via a CCD camera after i.p. injection of 150 uL D-luciferin (Gold Biotechnology, St. Louis, MO, USA). Tumor engraftment was measured on day 6, mice with incomplete tumor engraftment (< 10 4 RLU) were excluded (6 out of 48, including two arms of an independent study (Brahm et al. 2020)). Mice were stratified into four treatment groups each containing 7 mice based on their Fluc activity by distributing the maximal dynamic range over all groups; Vehicle (PBS; 200 uL/day), PD0325901 (5 mg/kg/ day), radiotherapy (2 Gy/day) and PD0325901/radiotherapy (5 mg/kg/day and 2 Gy/day). PD0325901 was administered via i.v. injection by tail vein. Radiotherapy was applied 1 h after PD0325901 injection as previously described by Lagerweij et al. (Lagerweij et al. 2021). In short, the mice were anesthetized before irradiation via 2% isoflurane. Only the head of the mice was irradiated and the body was shielded from the irradiation via lead. The mice received in a cranial anterior-posterior direction, 6 MV photon irradiation. The irradiation was delivered with a Clinac 2300 C/D (Varian Medical Systems, Palo Alto, CA, USA) at a dose rate of 360 Gy/h. After treatment, tumor growth was evaluated by measuring the Fluc activity at least once a week with a total amount of 15 times as described above. Survival was scored based on humane endpoints, defined as > 15% weight loss in 1-2 days or > 20% weight loss of highest recorded weight or neurological symptoms. When reaching a humane endpoint, mice were euthanized with 20% Euthasol ® (AST Farma, Oudewater, the Netherlands). Statistical analysis of the overall survival was performed using the log-rank (Mantel-Cox) test corrected for multiple testing. The statistical analyses were performed with Graphpad Prism (version 5), and a P value < 0.05 was considered to be statistically significant.

PD0325901 has a higher radio-sensitizing capacity than MEK-162 on GBM8 spheroids
We previously demonstrated that MEK-162 enhances the effect of irradiation in GBM cells in vitro and in vivo in a dose-dependent manner (Narayan et al. 2018). Alternative MEK1/2 inhibitors with properties that allow a better delivery to the brain could improve this effect. From a panel of MEK inhibitors evaluated both in vitro and in vivo, PD0325901 reached the brain most efficiently (Gooijer et al. 2018). Therefore, we investigated whether PD0325901 can enhance radiotherapy in spheroid cultured GBM8 cells similar to MEK-162. The IC50 of PD0325901 (0.7 μM) was lower than that for MEK-162 (3.9 μM) in GBM8 cells in a short-term assay (Fig. 1A, B).
Radiotherapy commonly shows its effect after a week of 3D spheroid expansion (Narayan et al. 2018). Therefore, we performed a long-term assay in which the drug exposure time was kept constant followed by a longer follow-up period, which also mimics the in vivo conditions (Fig. 1C). This longer follow-up after exposure changed the dynamics resulting in sensitivity with PD0325901 IC50 at 4 μM and MEK162 IC50 at 17 μM (Fig. 1D). Importantly, when GBM8 cells were exposed to a serial dilution PD0325901 before 2 Gy irradiation, it was cytotoxic at a lower concentration compared to MEK162 in combination with irradiation (Fig. 1D, PD0325901; 1.4 μM, MEK-162; 4.9 μM). This enhancing effect was also seen in the size of the spheroids (Fig. 1E).

Fig. 1 PD0325901 and MEK-162 radio-sensitize patient-derived GBM cells in vitro in a dose-dependent manner.
A Schematic overview of IC50 assay. GBM8 cells were treated for 72 h to PD0325901 or MEK162. Cell viability is assessed with CellTiterGlo 3D. B IC50 viability curves of GBM8 cells after 72 h treatment with either PD0325901 (red) or MEK-162 (green). Data represented as percentage viability compared to DMSO controls, average + SEM (n = 3). C Schematic overview of the long-term spheroid assay. GBM8 cells grown as one single spheroid treated for 96 h with PD0325901 or MEK162 and/or combined with 2 Gy radiotherapy for 3 consecu-tive days followed up until day 18 at which cell viability is assessed with CellTiterGlo 3D. D IC50 viability curves of GBM8 spheroids treated for 96 h with PD0325901 (light red), MEK-162 (light green), and PD0325901 combined with radiotherapy (3 × 2 Gy, dark red) or MEK162 combined with radiotherapy (3 × 2 Gy, dark green). Viability was measured on day 18, data represented as percentage viability compared to DMSO control, average + SEM (n = 3). E Images of GBM8 spheroids on day 18 treated with MEK162, PD0325901, MEK162 with radiotherapy, or PD0325901 with radiotherapy (n = 3)

Intravenously injected PD0325901 leads to micromolar plasma-brain concentrations
To have a radio-sensitizing effect of PD0325901, the drug has to reach sufficient levels in the brain to inhibit MEK1/2 and its downstream targets. Therefore, we determined the plasma-brain concentration levels in female athymic nude mice by administration of PD0325901 intravenously at a dose of 5 mg/kg ( Fig. 2A). One hour after i.v. administration, the highest PD0325901 brain concentrations were measured, followed by a slow clearance of PD0325901 in the next 4 h (Fig. 2C). Moreover, we analyzed per mouse brain 7 or more samples from different areas of the brain to assess the PD0325901 distribution throughout one mouse brain (Fig. 2B). We measured a similar amount of PD0325901 in the spatially different samples that originated from the same mouse brain (Supplementary Table S1).

Efficacy of PD0325901 in combination with radiotherapy in an orthotopic GBM8 patient-derived glioblastoma in vivo model
To study the radio-sensitizing effect of PD0325901 in vivo, we tested PD0325901 with or without radiotherapy in athymic nude mice bearing the patient-derived glioblastoma GBM8 (Fig. 3A). Orthotopic growth was assessed throughout the experiment by measuring the Fluc activity in the brain (Fig. 3B). Compared to vehicle, we observed a small non-significant difference in the estimated tumor volume in mice treated with radiotherapy alone and after combination with PD0325901 (Fig. 3C). Overall survival was significantly different between both the single-arm irradiation and the combination arm compared to vehicle (P < 0.005, Fig. 3D). No significant difference was observed between irradiation alone and irradiation combined with MEK inhibition (P ≥ 1.0, Fig. 3D). After two days of treatment, 29% (4 out of 14) of mice showed skin lesions on their back which recovered successfully after administration of the antibacterial ointment L-mesitran. These PD0325901-related side effects did not affect the scoring of the animals' condition or weight in the follow-up period (Fig. 3E-H). Hence, the endpoint of the study was not affected by the observed toxicity. Although we confirmed independently that MEK inhibition has a radio-sensitizing effect in vitro, this does not translate to an in vivo effect.

Discussion
In this study, we found a radio-sensitizing effect of PD0325901 in patient-derived glioblastoma spheroid cultures. The effect was seen both in a short-term and longterm assay, where the latter is more representative of the in vivo situation. The in vitro results agree with earlier observed results where we showed that another MEK inhibitor, MEK-162, significantly enhanced the effect of radiation on patient-derived glioblastoma spheroids in vitro and in an orthotopic GBM model in vivo (Narayan et al. 2018).

Fig. 2
Clinically relevant plasma-brain concentration observed after i.v. administration of PD0325901 in athymic nude mice. A Schematic overview. B Spatial orientation of the different segments taken from the mouse brain which were measured on the LC-MS/MS. C PD0325901 clearance of plasma and brain in WT athymic nude mice 5, 60, and 240 min after 5 mg/kg i.v. administration. Plasma concentrations are means ± SEM (n ≥ 2), brain concentrations are means ± SEM (n ≥ 7) Based on our previous study and in vitro results, we reasoned that PD0325901 in combination with irradiation might increase therapeutic efficacy in GBM8 tumor-bearing mice, also given that this drug can reach relevant concentrations in the brain. Hence, to reach therapeutic efficiency, a sufficient intracranial concentration has to be reached, where BBB penetration might be a limiting factor. Accordingly, we assessed the BBB penetration of PD0325901 in athymic nude mice. Gooijer et al. showed that administering PD0325901 i.v. 5 mg/kg in mice would lead to target GBM8-FM-bearing mice received treatment for 5 consecutive days with PD0325901. B Tumor engraftment was confirmed and tumor growth was measured over time by bioluminescence imaging. C Tumor growth per treatment group. Luciferase activity was measure by bioluminescence imaging of GBM8-FM bearing mice, data represented as average per treatment group with SEM (n = 7). D Overall survival analysis of GBM8-FM orthotopic xenografts bearing mice treated with vehicle (black line, n = 7), PD0325901 (light red line, n = 7), radiotherapy (blue line, n = 7) or PD0325901 and radiotherapy (dark red line, n = 7). E-H Weight development of GBM8 orthotopic xenograft-bearing mice over time treated with vehicle (black), PD0325901 (light red), 3 × 2 Gy radiotherapy (blue), or PD0325901 combined with 3 × 2 Gy radiotherapy (dark red). Radiotherapy was applied 1 h after PD0325901 injection as previously described by Lagerweij et al. [32). Significance was determined with the log-rank (Mantel-Cox) test corrected for multiple testing (*P < 0.05) inhibition of ERK1/2 (Gooijer et al. 2018). In our study, after administering 5 mg/kg PD0325901 via i.v., we found a corresponding PD0325901 concentration in the brain of athymic nude mice (Fig. 2C). Moreover, a similar concentration compared to the IC50 concentrations observed in vitro (4 μM) can be reached in vivo in plasma and brain after 1 h (plasma; 4.7 ± 0.9 μM (n = 3), brain; 7.1 ± 1.8 μM (n = 2). Based on these results, we assessed whether 5 mg/kg/day of PD0325901 could enhance the effect of radiotherapy in GBM-bearing mice. Although we were able to confirm the effect of irradiation on tumor growth in glioblastoma (Fig. 3D), PD0325901 did not radio-sensitize the tumor cells as observed in vitro. The combination of PD0325901 and radiotherapy did not significantly improve the overall survival of the mice compared to radiotherapy alone as expected based on our earlier results (Narayan et al. 2018). The current study indeed confirmed that MEK inhibition enhances radiotherapy in vitro, seen for both PD0325901 and MEK-162. This radio-sensitizing effect has a relatively small therapeutic effect range between 1 and 2 μM for PD0325901. This small therapeutic window from in vitro to in vivo could have contributed to the difference in the effect of both inhibitors as seen in vivo. Subtle changes in concentrations due to PD/ PK dynamics could easily fall outside the radio-sensitizing effect range of the drug. Hence, this could explain why, in the present study, we do not observe a similar effect with PD0325901 as seen previously with MEK162 (Narayan et al. 2018). PD0325901-related side effects are already evident after two days of treatment with 5 mg/kg/per day. Therefore, aiming for treatment with a higher dosage of PD0325901 is not advised without taking countermeasures to prevent side effects. Alternatively, different dose schedules or delivery methods could be explored. For instance, instead of just one-week treatment, several cycles of the same doses of radiotherapy in combination with PD0325901 could be tested which would also fit better in clinical practice. To improve drug delivery, Hou et al. demonstrated that RGD peptides in combination with PD0325901 led to enhanced antitumor activity in the subcutaneous U87 tumors in vivo (Hou et al. 2016). Although it can be argued whether such large molecules, like RGD peptides, would be able to reach a sufficient concentration in the brain and subsequently in the tumor. Together, our data indicate that further investigation into PD0325901 pharmacokinetics specifically focusing on the brain and the tumor is required.
In conclusion, our data provide valuable further proof of MEK inhibitors as radio-sensitizers. We confirmed an earlier identified radio-sensitizing effect of MEK inhibition using another compound that has a favorable BBB penetrating profile. Although PD0325901 reaches a sufficient concentration throughout the brain, matching the therapeutic window, we have been unable to show a positive therapeutic interaction in vivo. This indicates that more investigation is required focusing on target engagement in relation to the pharmacokinetic profile of PD0325901 among other candidate MEK inhibitors.