Synergistic activity of mTORC1/2 kinase and MEK inhibitors suppresses pediatric low-grade glioma tumorigenicity and vascularity

Abstract

Background

Pediatric low-grade glioma (pLGG) is the most common childhood brain tumor. Many patients with unresectable or recurrent/refractory tumors have significant lifelong disability. The majority of pLGG have mutations increasing the activity of the Ras/mitogen-activated protein kinase (MAPK) pathway. Activation of mammalian target of rapamycin (mTOR) is also a hallmark of pLGG. We therefore hypothesized that the dual target of rapamycin complexes 1 and 2 (TORC1/2) kinase inhibitor TAK228 would synergize with the mitogen-activated extracellular signal-regulated kinase (MEK) inhibitor trametinib in pLGG.

Methods

We tested TAK228 and trametinib in patient-derived pLGG cell lines harboring drivers of pLGG including BRAF^V600E and neurofibromatosis type 1 loss. We measured cell proliferation, pathway inhibition, cell death, and senescence. Synergy was analyzed via MTS assay using the Chou–Talalay method. In vivo, we tested for overall survival and pathway inhibition and performed immunohistochemistry for proliferation and vascularization. We performed a scratch assay and measured angiogenesis protein activation in human umbilical vein endothelial cells (HUVECs).

Results

TAK228 synergized with trametinib in pLGG at clinically relevant doses in all tested cell lines, suppressing proliferation, inducing apoptosis, and causing senescence in a cell line–dependent manner. Combination treatment increased median survival by 70% and reduced tumor volume compared with monotreatment and control cohorts. Vascularization of tumors decreased as measured by CD31 and CD34. Combination treatment blocked activation of focal adhesion kinase (FAK) and sarcoma proto-oncogene non-receptor tyrosine kinase (SRC) in HUVEC cells and reduced HUVEC migration compared with each drug alone.

Conclusions

The combination of TAK228 and trametinib synergized to suppress the growth of pLGG. These agents synergized to reduce tumor vascularity and endothelial cell growth and migration by blocking activation of FAK and SRC.

Key Points.

1. Combining 2 new drugs taken by mouth leads to decreased pediatric low-grade glioma tumor growth.

2. These drugs block pathways that drive tumor cell growth and blood vessel recruitment.

3. Our data support a clinical trial of mTOR and MEK inhibitors in pediatric low-grade glioma.

Importance of the Study.

Pediatric LGGs show activation of the mTORC1/2 and MAPK pathways. We show that the brain-penetrant TORC1/2 kinase inhibitor TAK228 synergizes with the MEK inhibitor trametinib to suppress pLGG. We also discover synergistic inhibition of the vascularity of LGG in vivo, and mechanistically demonstrate that the 2 pathways converge in endothelial cells on FAK and SRC to suppress tumor angiogenesis. TAK228 is under consideration for pediatric brain tumor trials, and our results suggest TAK228 combination with MEK inhibitors may be efficacious for pLGG.

Pediatric low-grade glioma (pLGG), the most common brain tumor in children, becomes clinically challenging in unresectable tumors or when the tumor recurs or develops resistance to the standard chemotherapies. V-Raf Murine Sarcoma Viral Oncogene Homolog B (BRAF) and neurofibromatosis type 1 (NF1) mutations in pLGG regulate 2 key pathways that govern cell growth: the mitogen-activated protein kinase (MAPK) and mammalian target of rapamycin complexes 1 and 2 (mTORC1/C2) signaling pathways.^1–5

Trametinib is a third-generation, orally available, highly selective allosteric ATP noncompetitive inhibitor of the 2 isoforms of mitogen-activated extracellular signal-regulated kinase 1 (MEK1) and MEK2.⁶ Trametinib has yielded significant improvement in progression-free and overall survival in patients with BRAF^V600E/K mutated melanoma, leading to FDA approval.⁷ Trametinib was used in 6 children with progressive pLGG with a MAPK pathway activation,⁸ and is currently being studied in larger early phase clinical trials in pediatric patients (NCT02124772, NCT02684058, NCT03363217).

The FDA-approved rapalog everolimus inhibits mTORC1 and has activity in pLGG.^9,10 The high-level activation of mTORC2 in pLGG^3,4 led us to investigate the antitumor activity of TAK228, a second-generation mTORC1/2 inhibitor in pLGG. TAK228 is an orally bioavailable, potent, and highly selective ATP-competitive inhibitor of both mTORC1 and mTORC2 subunits. Previously, we found that the mTORC1/2 kinase inhibitor TAK228 was superior to everolimus in aggressive pediatric brain tumors.¹¹ In the present study, we hypothesized that TAK228 in combination with trametinib would synergize in pLGG.

Materials and Methods

Cell Lines

The pediatric pilocytic astrocytoma (PA) cell line Res186 (harboring deletion of phosphatase and tensin homolog [PTEN]) and the diffuse astrocytoma cell line Res259 (harboring gain of platelet derived growth factor receptor α [PDGFRα] and deletion of cyclin-dependent kinase inhibitor 2A [CDKN2A]) were kindly provided by Dr Chris Jones (Institute of Cancer Research, Sutton, UK).¹² These cell lines were originally described by Bobola et al.¹³ No BRAF alterations were detected in Res186 or Res259.¹⁰ Both cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM)/nutrient mixture F12 supplemented with 10% heat-inactivated fetal bovine serum. The JHH_NF1_PA1 cell line was grown in 50% DMEM/F12 supplemented with human insulin and 50% DMEM/F12 conditioned medium derived from NIH 3T3. Derivation and maintenance of the JHH_NF1_PA1 cell line will be described in detail elsewhere (Yuan et al under review). Briefly, a 16-year-old male patient with NF1 and an optic chiasm PA received first carboplatin, then bevacizumab, and later everolimus treatment. A biopsy after treatment with everolimus was used to generate the line. Genetic profiling revealed NF1 mutation, and western blot and PCR analysis confirmed loss of NF1 protein expression and no BRAF alterations.¹⁰ We also used a short-term culture from a BRAF^V600DL LGG tumor isolated from an infant. These cells were grown in the same media as JHH_NF1_PA1. All experiments with these freshly derived cells were conducted between passages 3 and 7. Human umbilical vein endothelial cells (HUVECs) were kindly provided by Dr Malia Edwards. Cells were grown in endothelial cell growth medium (Sigma-Aldrich).

Cells were verified through DNA short tandem repeats.¹⁰ Cell lines were routinely tested for the presence of mycoplasma contamination via PCR.

All in vivo experiments used the patient-derived BT40 xenograft line, which was previously characterized as pediatric PA tumor containing a BRAF^V600E mutation.^14,15 BT40 only grows as serially passaged flank tumors in immunocompromised mice and will not form tumors intracranially or grow in cell culture.

Experimental Agents

TAK228 (#S2811), trametinib (#S2673), and everolimus (#S1120) were purchased from Selleckchem. For in vitro experiments, TAK228, trametinib, and everolimus were dissolved in dimethyl sulfoxide (DMSO). For in vivo studies, TAK228 and trametinib were formulated in 0.5% methylcellulose and 0.2% Tween 80 and everolimus in 30% propylene glycol and 5% Tween 80. Agents were dissolved using a 5-min ultrasonication with degassing function water bath at room temperature (Branson #2510).

Determination of Cell Growth: MTS Assay

Cell titer 96 (Promega) was used to determine growth in viable cell mass after treatment with TAK228 and/or trametinib as described previously.¹⁰

Synergy Calculations

Based on results of assay by MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium), drug synergy assessment was performed using the Chou–Talalay method.¹⁶ Drug concentrations were increased in parallel so that 5 nM TAK228 was tested in combination with 5 nM trametinib and 10 nM TAK228 with 10 nM trametinib and so on. Data analysis was conducted using CompuSyn software (http://www.combosyn.com/).

Proliferation Assay

Proliferation assays were performed by incubating cells with 100 µM 5-bromo-2′-deoxyuridine (BrdU; Sigma-Aldrich) for 4 hours as described previously.¹⁷ Anti-BrdU antibody was used per the manufacturer’s direction (Cell Signaling Technologies #5292) at 1:1000 dilution.

Determination of Migration with Wound Healing Assay

Wound healing assay was performed as described.¹⁸ Briefly, 1 × 10⁵ HUVECs were seeded in each well of a 12-well plate containing 20 ng/mL vascular endothelial growth factor 165 (VEGF₁₆₅). On the next day, the confluent monolayer was wounded with a single pass of a 200-µL pipette tip, medium was removed, and cells were washed twice with phosphate buffered saline (PBS) to remove floating cells. Cells were treated with TAK228 10 nM and/or trametinib 10 nM or DMSO including VEGF₁₆₅ and photographed at interval timepoints.

Western Blot Analysis

Protein extraction, separation, and immunoblotting were performed as previously described.¹⁹ All western blots are representative of 3 independent experiments. For pathway inhibition studies, cells were treated for 1 or 2 hours. Apoptosis was assessed at 24–48 hours, depending on the cell line. See Supplementary Methods for details on antibodies used. Densitometry was performed using ImageJ v1.440 software (https://imagej.nih.gov/ij/).

In Vivo Efficacy Experiments

All studies performed were in compliance with the United States Animal Welfare Act regulations and Public Health Service Policy. The “Principles of Laboratory Animal Care” (NIH publication no. 8623, revised 1985) was followed, with protocol approved by the Johns Hopkins Animal Care and Use Committee. 1 × 10⁶ BT40 tumor cells in 50% DMEM/F12 and 50% Matrigel (Corning), were injected into the flanks of 4- to 6-week-old female nu/nu mice (Charles River). Treatment started after the flank tumors were visible and measurable. Mice were randomized into 4 groups of 4–5 mice and were treated by oral gavage with trametinib (1.5 mg/kg, 5x per week, Monday–Friday), TAK228 (1 mg/kg, 3x per week, Monday, Wednesday, Friday), or the combination of these drugs at the same concentration as their single doses. Control animals received vehicle solution accordingly. Tumor volumes were measured using the following formula: V = (largest tumor dimension) × (smallest tumor dimension) × 0.52. When the overall flank tumor volume per mouse reached the maximum volume declared in the animal protocol, the mouse was sacrificed.

Statistical Analysis

Differences between 3 or more groups were assessed using a one-way ANOVA test with post hoc Dunnett’s or Tukey’s multiple comparison test. All statistical tests were performed using GraphPad Prism 6 software.

Results

TAK228 and Trametinib as Monotherapies Inhibit Cell Growth in pLGG

We tested TAK228 against the following patient-derived pLGG cell lines: JHH_NF1_PA1^NF1-/-, Res186^PTEN-/-, and Res259^CDKN2A-/-. TAK228 significantly inhibited the growth of these cell lines at low nM concentrations (Fig. 1A). TAK228 treatment led to dose-dependent decreases in TORC1 readouts phosphorylated (p)S6, p4E-BP1(threonine 37/46), and TORC2 readout pAKT (serine 473) (Fig. 1B). In all cell lines, inhibition of mTORC1 and mTORC2 led to an increase in MAPK pathway activation detected with phosphorylated extracellular signal-regulated kinase 1 and 2 (pERK1/2) (Fig. 1C).

Fig. 1.

TAK228 inhibits cell growth and decreases mTORC1/C2 activation in pLGG. (A) MTS assays showing the effect of TAK228 in different concentration levels on cell growth after 6 days treatment in Res186, Res259, and JHH_NF1_PA1. In the box whisker plots, the box indicates interquartile range; the center bar, median; whiskers, data range; and dots, data points from 3 independent experiments each with 3 technical replicates; *P < 0.05, ***P < 0.001, ****P < 0.0001 against vehicle by one-way ANOVA. (B) Western blot showing suppression of mTOR targets after TAK228 treatment. β-actin was used as loading control. Numbers above the blot indicate the densitometry compared with vehicle control.

Trametinib at 5 nM inhibited pLGG growth after 6 days of treatment (Supplementary Figure 1A) and suppressed pERK (Supplementary Figure 1B). Trametinib caused an increase in mTORC1 and mTORC2 activation (Supplementary Figure 1C).

Dual Inhibition of mTORC1/C2 and MAPK Pathways Synergistically Inhibits pLGG Growth

Compensatory upregulation of the mTOR or MAPK pathway with monotherapy led us to hypothesize that combination therapy would synergize in pLGG. TAK228 and trametinib inhibited cell growth as monotherapy, but the inhibitory effect was stronger in combination (Fig. 2A). Pediatric LGG cells showed a statistically significant suppression of growth in combination treatment at as low as 1 nM for TAK228 and 1 nM trametinib (Fig. 2B). The Chou–Talalay method revealed a synergistic growth inhibition in clinically relevant doses (Fig. 2C, Supplementary Figure 2A). Combination of higher, nonclinically relevant drug concentrations resulted in a combination index higher than 1, which is an indicator for antagonistic effect of both drugs. Images of Res186 and Res259 cells further demonstrated the inhibitory effect of monotreatment and an increased inhibition with combination treatment (Supplementary Figure 2B). In a short-term culture derived from a pLGG harboring a BRAF^V600DL alteration, combination therapy statistically significantly suppressed growth to a greater extent than monotherapy (Supplementary Figure 3).

Fig. 2.

Combination treatment with TAK228 and trametinib synergistically slows cell growth in pLGG cell lines. (A) Representative images of JHH_NF1_PA1 cells treated with TAK228, trametinib, and combination at 10 nM concentration. (B) MTS cell growth assay shows TAK228 combined with trametinib for Res186, Res259, and JHH_NF1_PA1 cell lines treated with 1–20 nM for 6 days shown as box whisker plots. Box indicates interquartile range; the center bar, median; whiskers, data range; and dots, data points from 3 independent experiments each with 3 technical replicates. *P < 0.05, ***P < 0.001, ****P < 0.0001 against vehicle by one-way ANOVA. (C) Assessment of the degree of synergy between TAK228 and trametinib in Res186, Res259, and JHH_NF1_PA1 using the Chou–Talalay method. Drug interaction was plotted as combination index (CI) on the y-axis and the fraction affected (Fa) on the x-axis. Concentrations used were 1, 5, 10, 20, 40, and 80 nM of each drug.

Combination of TAK228 and Trametinib Suppresses mTOR and MAPK Activation

Combined TAK228 and trametinib at 5 or 10 nM of each drug inhibited pAKT (ser473), pS6, and p4E-BP1 in pLGG cells (Fig. 3A). At ≥5 nM concentrations of TAK228 and trametinib, pERK was 100% inhibited in all lines (Fig. 3B). BRAF^V600DL pLGG cells showed a reduction in mTOR and MAPK pathway activation in combination treatment (Supplementary Figure 4). Combination therapy at clinically achievable low nanomolar concentrations abolished the compensatory upregulation of mTOR and MAPK pathways we observed after monotherapy.

Fig. 3.

TAK228 combined with trametinib decreases mTORC1/C2 and MAPK activity in the low nanomolar range levels in pLGG. Western blot in (A) mTOR targets (B) for the MAPK readout pERK1/2 (thr202/tyr204), and total ERK1/2 in pLGG cell lines. β-actin was used as loading control. Numbers above the blot indicate the densitometry compared with vehicle control.

TAK228 and Trametinib Inhibit Proliferation

To investigate the mechanism of growth inhibition, we performed BrdU proliferation assays. In JHH_NF1_PA1, monotherapy with TAK228 suppressed proliferation at 20 nM and monotherapy with trametinib decreased proliferation at 10 nM. However, the combination significantly reduced cell proliferation at 5 nM and above (Fig. 4A). In Res186, proliferation was significantly reduced in combination treatment at ≥20 nM (Supplementary Figure 5A). No significant reduction in BrdU incorporation was seen in Res259 cells in clinically relevant doses up to 50 nM (Supplementary Figure 5B). In contrast, the total cell number was significantly reduced in Res186 and Res259 cell lines in combination treatment in all tested concentrations, suggesting combination therapy might also act through other mechanisms (Supplementary Figure 6).

Fig. 4.

Dual mTORC1/2 kinase/MEK inhibition suppresses proliferation and induces either apoptosis or senescence in pLGG. (A) Suppression of BrdU incorporation by immunofluorescence following 72-hour drug incubation (left panel). Representative images showing immunofluorescence at 10 nM dose for JHH_NF1_PA1 cell line. Bar = 20 microns. On the right panel, BrdU-positive cells are shown as dot plots showing the median as center bar including ±SD from 3 independent experiments. ***P < 0.001, ****P < 0.0001 against vehicle by one-way ANOVA. (B) Res186, Res259, and JHH_NF1_PA1 cells were treated with TAK228, trametinib, combination, and DMSO (vehicle) at different concentrations for 24 hours and subjected to immunoblot analysis to determine levels of PARP (full-length 116 kDa, cPARP 89 kDa). Numbers above the blot indicate the densitometry compared with vehicle control. (C) JHH_NF1_PA1 cells were treated with vehicle, TAK228, trametinib, and combination (10 nM) for 96 hours, and then expression analysis of senescence-associated β-galactosidase by β-gal staining was performed (C). Bar = 40 microns. (D) After 72-hour incubation, western blot of phosphorylated retinoblastoma (pRB), total RB, and p27. β-actin is a loading control. Numbers above the blot indicate the densitometry compared with vehicle control.

TAK228 and Trametinib Induce Apoptosis More Effectively in Combination

To test if combination treatment was inducing apoptosis, we performed western blot analysis for cleaved poly(ADP-ribose) polymerase (cPARP) (Fig. 4B). In Res186, combination therapy in low doses induced cPARP, while TAK228 did not induce cPARP, and trametinib only induced cPARP at 20 nM and above. In Res259, trametinib monotherapy robustly induced cPARP and the addition of TAK228 did not significantly increase apoptosis. In JHH_NF1_PA1, no cPARP could be detected. In the pLGG BRAF^V600DL-derived short-term cell culture, combination treatment also induced cPARP expression (Supplementary Figure 7).

Combination Treatment with TAK228 and Trametinib Induces Senescence in pLGG

During treatment of JHH_NF1_PA1 cells, the morphology of the cells changed to a senescence-like phenotype, with large flattened cells (Fig. 4C). After 4 days of treatment we stained cells for senescence-associated β-galactosidase (SA-β-gal) enzyme activity. The trametinib- and combination-treated cell culture wells grew less well than TAK228 or control cells, concordant with our MTS findings. Staining for SA-β-gal revealed elevated enzyme activity with trametinib and combination treatment in NF1-driven pLGG cells (Fig. 4C). To confirm induction of senescence, we measured protein expression of p27 via western blot (Fig. 4D). In JHH_NF1_PA1 cells, TAK228 caused a slight increase in p27; however, trametinib increased p27 by 7-fold and combination by 13-fold at 10 nM concentrations. We did not observe an induction of p27 in Res186 (data not shown). Res259 cells had an approximately 2-fold increase in p27 expression with combination therapy. We did not observe an induction of p27 or p21 in the BRAF^V600DL short-term cell culture. Together these data show that mTORC1/2 and MAPK pathway inhibition acts through diverse mechanisms to suppress pLGG growth.

TAK228 and Trametinib Synergize to Suppress BRAF^V600E Glioma Tumor Growth In Vivo

To assess whether treatment with TAK228 in combination with trametinib provides a survival benefit in vivo, we used the patient-derived BT40 tumor cell line. This established pediatric PA xenograft model harbors BRAF^V600E leading to activation of the mTOR and MAPK pathways.^14,15 BT40 also has CDKN2A and TP53 deletions, typical for many World Health Organization grade II pLGG tumors.^20,21 After tumors were measurable, mice were randomized to treatment with vehicle, TAK228, trametinib, or combination. Overall survival was significantly prolonged in the combination cohort compared with monotreatment (P < 0.0001, log-rank test) (Fig. 5A). The median survival for the combination treatment was 36 days compared with 12 days under each monotreatment and 11 days for the vehicle cohort. One mouse in the combination treatment group was sacrificed 22 days post treatment with a smaller xenograft because of an abscess unrelated to tumor injection or treatment and was censored in the statistical analysis. The comparison of tumor volume at day 12 (Fig. 5B) and day 8 (Supplementary Figure 8A) demonstrated that combination-treated tumors were smaller over time compared with monotherapy and vehicle-treated tumors.

Fig. 5.

Dual mTORC1/C2 kinase/MEK inhibition suppresses proliferation and prolongs life span in BT40 pLGG xenografts. (A) Kaplan–Meier plots showing survival of BT40 pLGG xenografts after oral treatment with TAK228 (1 mg/kg; 3x/week; p.o.) combined with trametinib (1.5 mg/kg; 5x/week; p.o.) compared with monotreatment and vehicle controls. Five mice were in each treatment group. Log-rank test was used to determine statistical significance comparing combination group with each of the 3 control groups individually (P < 0.0001). (B) Tumor volume measurement after 12 days treatment. (C) Five independent tumors for each condition were stained for pERK1/2 and validated with the H-score and graded as follows: 1 = weak, 2 = median, or 3 = strong and the percentage of positive cells shown as box whisker plots (range: 0–300). Representative 200X immunohistochemistry images for pERK1/2 including the grade shown in the lower panel. Bar = 40 microns. (D) Tumor tissue was stained for the proliferation marker Ki67. Dot plot shows the median as center bar including SD. Representative Ki67 immunohistochemistry 200X images for each treatment cohort are in the right panel. Bar = 40 microns. (E) Photographs of isolated flank tumors at endpoint of the experiment. Scale bar represents 1 cm. For B–D, one-way ANOVA with post-hoc Dunnett’s test was used to determine statistical significance between vehicle and TAK228 or trametinib or combination treatment; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001.

TAK228 and Trametinib Reduce mTOR and MAPK Pathways and Decrease Proliferation In Vivo

To determine if TAK228, trametinib, or combination treatment influenced the specific pathways, in short-term experiments we evaluated 4 different flank tumors from each group by western blot (Supplementary Figure 8B). Phosphorylated AKT (ser473) was decreased over 98% in TAK228-treated mice and combination treatment resulted in 82% reduction. We saw a 14% increase of pAKT (ser473) in the trametinib-treated cohort compared with vehicle. TAK228 or combination treatment caused a 90% reduction in pS6 expression. Interestingly trametinib caused a 50% reduction in pS6. Immunohistochemistry performed on treated tumors revealed a corresponding reduction in pERK1/2 (Fig. 5C). Trametinib and combination treatment caused a significant decrease in pERK expression compared with the vehicle control. Staining of pERK1/2 (thr202/tyrosine 204) was stronger in the nuclei and weaker in the cytoplasm of trametinib-treated mice. No difference in pERK was noted between vehicle and TAK228-treated mice (Fig. 5C). We observed suppression of pS6 in TAK228- and combination-treated mice, with a greater suppression of pS6 seen in the combination (Supplementary Figure 9).

To evaluate the mechanistic reason for growth inhibition in vivo, we stained tumor tissue for the proliferation marker Ki67. After immunohistochemistry staining for Ki67, we detected a significant reduction under trametinib treatment (P = 0.046) and a much stronger inhibition in combination treatment (P < 0.0001) by one-way ANOVA (Fig. 5D).

Combination Treatment Leads to Decreased Vascularity in BRAF^V600E Mutant pLGG Tumors

At the endpoint for the in vivo experiments, the BT40 transplanted flank tumors were removed from the flanks for further analysis. Flank tumors in the combination cohort were smaller and firmer compared with vehicle-treated cohort (Fig. 5E). The amount of blood vessels was grossly reduced in the combination group compared with vehicle and monotherapy. This led us to investigate changes in vascularity. Combination treatment tumors had a reduction of the endothelial markers cluster of differentiation (CD)31 and CD34 of >90% via western blot compared with control (Fig. 6A). There was also a trend of downregulation of VEGF in combination-treated tumors. To verify these data, we stained tumor slides for CD34 (Supplementary Figure 10A). We found a decrease in CD34+ blood vessels in combination-treated tumors compared with vehicle (P = 0.0476, Fisher’s exact test). We also performed immunohistochemistry for CD31 and found a similar reduction in staining in BT40 tumors with combination therapy (Supplementary Figure 10B).

Fig. 6.

Dual mTORC1/C2 kinase/MEK inhibition suppresses angiogenesis in vivo and inhibits the FAK/SRC complex. (A) Western blot of tumor lysates for endothelial markers with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as loading control. (B) HUVECs were treated with TAK228, trametinib, and combination at 10 nM concentration for 6 hours. DMSO was included as vehicle control. Western blot in the left panel shows decreased activation of FAK and SRC. β-actin is the loading control. Numbers above the blot indicate the densitometry compared with vehicle control.

TAK228 and Trametinib Decrease pFAK and pSrc Activity in Human Endothelial Cells

Our data on BT40 tumors led to the hypothesis that TAK228 and trametinib have an inhibitory effect on endothelial cells and therefore inhibit blood vessel formation. Mono- and combination treatment of HUVECs resulted in a reduction in mTOR and MAPK pathways in HUVECs similar to pLGG cell lines (Supplementary Figure 11). We then evaluated the phosphorylation of focal adhesion kinase (FAK) and sarcoma proto-oncogene non-receptor tyrosine kinase (SRC), two kinases that are critical for endothelial cell growth and migration.^22,23 Trametinib led to a 90% reduction on the MAPK-specific phosphorylation site on FAK (S910) (Fig. 6B). In combination, FAK phosphorylation was 99% reduced. Next, we found that TAK228 and trametinib were each reduced in monotherapy SRC phosphorylation by 40–50% and in combination SRC phosphorylation was reduced by 95% (Fig. 6B). Combination treatment therefore resulted in inactivation of SRC as well FAK. Finally, we evaluated the effect of TAK228 and trametinib on cell movement. We performed a scratch assay to assess the HUVEC migration capacity after treatment. After 48 hours the scratch persisted in the combination treatment, whereas monotreated and vehicle-treated cells had a fully closed scratch (Supplementary Figure 12).

Adverse Effects of Mono- and Combination Treatment

Adverse effects are a major concern for inhibition of key growth pathways in children. TAK228 and trametinib monotreated mice showed no adverse signs. However, 1 out of 4 mice in the combination treatment cohort lost weight during treatment and had skin redness. In this mouse, the growth of tumor volume was the smallest of all combination-treated mice. We interrupted the treatment until the weight normalized (4 days). We continued the treatment and adapted dosing to daily weight changes. This particular mouse survived the longest with the smallest flank tumors. Monitoring of weight of mice not bearing flank tumors and treated with mono- or combination therapy for 28 days showed that combination-treated mice did not gain weight as well as control or monotreated mice but that this difference was not statistically significant (P = 0.07 for combination treatment vs vehicle control by ANOVA; Supplementary Figure 13). After 28 days of treatment, mice were sacrificed and brain, liver, kidney, spleen, heart, and lungs were fixed and processed for histology for a masked toxicology assessment by a board-certified pathologist (C.G.E.). The pathology results were negative for apoptosis, vascular deficits, or necrosis in all organs examined (data not shown).

Combination Treatment with mTORC1 Inhibitor Everolimus with Trametinib Inhibits pS6 but Increases pAKT (ser473) and Has No Effect on p4E-BP1 in pLGG Cell Lines

To compare the dual mTORC1/C2 inhibitor TAK228 directly with the FDA-approved mTORC1 inhibitor everolimus, we conducted an in vitro and short-term in vivo experiment with everolimus. Pediatric LGG cell lines treated with everolimus for 6 hours resulted in an inhibition of pS6 but not p4E-BP1 and pAKT (ser473) as analyzed by western blot (Supplementary Figure 14A). In Res186 and JHH_NF1_PA1 cell lines, everolimus treatment resulted in increased pAKT (ser473) activation compared with the vehicle control (Supplementary Figure 14A). No activation of the MAPK pathway was observed (Supplementary Figure 14B). In vivo in BT40 xenografts, a single dose of TAK228 resulted in a decrease in pAKT, pS6, and p4E-BP1 in mono- and combination treatment compared with vehicle control (Supplementary Figure 15A). Everolimus resulted in a decrease of pS6 in mono- and combination treatment compared with vehicle control, but no suppression of p4E-BP1. Everolimus mono- or combination treatment increased pAKT (ser473) by 1.5- to 2-fold. No changes were observed in pERK expression compared with the vehicle control. Immunohistochemistry showed superior suppression of p4E-BP1 with TAK228 compared with everolimus in mono and combination treatment with trametinib (Supplementary Figure 15B).

Discussion

Activation of MAPK and/or mTOR signaling is among the hallmarks of cancer.²⁴ Between both pathways exists a crosstalk leading to compensatory activation of one of the pathways if the other pathway is inhibited.^25,26 Blocking both pathways simultaneously may represent an important treatment strategy in pLGG. In clinically relevant low nanomolar concentrations, we saw a significant growth and pathway inhibition with TAK228/trametinib combination therapy in our pLGG models. We detected increased apoptosis in Res186 but not in our other 2 cell lines, which instead showed signs of induction of senescence. One explanation for why we did not detect apoptosis in our NF1-driven tumor cell line could be that JHH_NF1_PA1 cells grew in the media which contained the Rho kinase (ROCK) inhibitor Y-27632. ROCK inhibition prevents apoptosis through suppression of caspase-3 activity.²⁷

We found that combination therapy in vivo suppressed mTORC1/2 and MAPK targets, decreased proliferation, and doubled the median survival of BT40-harboring mice. Tumors were decreased in size and were markedly less vascular, with a decrease of CD34-positive blood vessels on histology, as well as VEGF and CD31 expression in the tumor. Combination TAK228/trametinib therapy inhibited the migration of human endothelial cells, decreasing FAK and SRC activation. Rapamycin, everolimus, and trametinib have an anti-angiogenic effect on mature blood vessels in other tumor types.^28,29 The high degree of gadolinium enhancement in pLGG and the success of anti-VEGF therapy such as bevacizumab³⁰ suggest that TAK228/trametinib-mediated inhibition of the abnormal blood supply of pLGG may have additional therapeutic benefits to patients. Our results are concordant with a recently published study showing that TAK228 suppressed bladder cancer vascularization, as evidenced by reduced tumor expression of CD31, CD34, and VEGF.³¹

We found that the FDA-approved mTORC1 inhibitor everolimus suppresses pS6 activation in vitro and in vivo. However, everolimus does not decrease p4E-BP1. The mTORC2 marker pAKT (ser473) also remained activated or increased activation after everolimus treatment. TAK228’s ability to suppress p4E-BP1 and pAKT (ser473) in BT40 was similar to that observed by us in atypical teratoid/rhabdoid tumor orthotopic xenografts.¹¹ We have optimized murine TAK228 dosing based on established once- or thrice-weekly dosing schemes in adult human patients to accentuate tumor penetration and target occupancy, which may explain why our results differ from those reported recently in high-grade glioma.^32–34

Olow et al showed a significant survival increase after everolimus monotreatment compared with the vehicle control in BT40-bearing mice,³⁵ and in an earlier study, we detected a similar increase in survival with everolimus monotreatment.¹⁰ We have not observed a survival increase in TAK228 as a single agent. A possible reason could be our treatment schemata for 3 days per week (Monday, Wednesday, Friday) and TAK228’s short half-life compared with the long half-life of everolimus.^9,33,34,36 Daily dosing schemes exist in adults for TAK228; however, near-constant inhibition of mTORC1/mTORC2 signaling in combination with MEK inhibition could prove highly toxic. We observed some toxicity (weight changes, skin redness) with trametinib/TAK228 combination even when TAK228 was administered thrice weekly. These side effects are expected with MEK inhibitors, and established clinical trial schemas exist for the management of MEK inhibitor skin toxicity.³⁷ Importantly, our in vivo studies allowed the mice to recover with no treatment for 2 days per week. In comparison to studies of everolimus,^10,35 we may have seen a longer survival trend in the TAK228 single agent cohort if we had administered TAK228 for 7 days a week.

TAK228 is in clinical trials in adult patients and available through the National Cancer Institute–Cancer Therapeutic Evaluation Program. Early phase clinical trials are planned for TAK228 in pediatric brain tumors. Our findings are concordant with prior reports that everolimus and a MEK inhibitor synergized in heterotopic models bearing BRAF fusions³⁸ and that MEK or phosphatidylinositol-3 kinase inhibition suppressed optic pathway tumors in Nf1-deficient mice.³⁹ The increased activation of TORC2 in pLGG^3,4 and the robust response of human pLGG cell lines and xenografts to TAK228/trametinib suggests that TAK228 in combination with a MEK inhibitor such as trametinib may be a new therapeutic strategy for children with recurrent/refractory pLGG.

Funding

This work was supported by the Imagine an Answer Foundation, Alex’s Lemonade Stand Foundation, Giant Food Foundation, and a National Cancer Institute core grant to the Johns Hopkins Sidney Kimmel Comprehensive Cancer Center (CA006973).

Conflict of interest statement. The authors declare no conflicts of interest.

Authorship statement. Designed experiments: AA, MY, CGE, ER. Performed experiments: AA, AP, MY, LH. Analyzed data: AA, LH, CGE, ER. Drafted manuscript/edited manuscript: AA, MY, AP, MY, CGE, ER.