cMyc and ERK activity are associated with resistance to ALK inhibitory treatment in glioblastoma

Abstract

Background Anaplastic lymphoma kinase (ALK) is expressed in ~ 60% of glioblastomas and conveys tumorigenic functions. Therefore, ALK inhibitory strategies with alectinib are conceivable for patients with glioblastoma. The aims of this preclini- cal study were to investigate efficacy as well as to understand and potentially overcome primary and acquired resistance mechanisms of alectinib in glioblastoma.

Methods Efficacy of alectinib was analyzed dependent on ALK expression in different glioblastoma initiating cells and after lentiviral knockdown of ALK. Alectinib resistant cells were generated by continuous treatment with increasing alectinib doses over 3 months. M-RNA, phospho-protein and protein regulation were analyzed to decipher relevant pathways associ- ated to treatment or resistance and specifically inhibited to evaluate rational salvage therapies.

Results Alectinib reduced clonogenicity and proliferation and induced apoptosis in ALK expressing glioblastoma initiat- ing cells, whereas cells without ALK expression or after ALK depletion via knockdown showed primary resistance against alectinib. High expression of cMyc and activation of the ERK1/2 pathway conferred resistance against alectinib in ALK expressing glioblastoma cells. Pharmacological inhibition of these pathways by cMyc inhibitor or MEK inhibitor, trametinib, overcame alectinib resistance and re-sensitized resistant cells to continued alectinib treatment. The combination of alectinib with radiotherapy demonstrated synergistic effects in inhibition of clonogenicity in non-resistant and alectinib resistant glioblastoma cells.
Conclusion The data offer rationales for alectinib treatment in ALK expressing glioblastoma and for the use of ALK expres- sion status as potential biomarker for alectinib treatment. In addition, the results propose MEK inhibition or radiotherapy as reasonable salvage treatments after acquired alectinib resistance.

Keywords : Glioblastoma · Anaplastic lymphoma kinase · Alectinib · Biomarker · Treatment resistance

Introduction

Molecularly informed precision treatments are of increas- ing interest in neuro-oncology, but primary and acquired treatment resistance challenges their effective use. Preclin- ical studies deciphering signaling pathways and potential resistance mechanisms of these treatments help to improve patient selection and potential combinatorial approaches. Alectinib is a highly selective, second-generation inhibitor of the tyrosine kinase anaplastic lymphoma kinase (ALK) [1], and represents a clinically approved targeted therapy for ALK-rearranged non-small lung cancers (NSCLC) [2, 3]. Importantly, alectinib is also effective in the treatment of brain metastasis of ALK-rearranged NSCLCs, demonstrating its high brain penetrance [4-6]. Given the presence of molecular lesions, alectinib might also be explored for patients with primary brain tumors.

ALK receptor activity is enhanced by ALK translo- cations [7, 8], ALK mutations [9, 10], ALK amplifica- tion and overexpression [11, 12] or by binding of the ligands pleiotrophin and midkine [12]. Activation of the ALK receptor conveys tumorigenic functions including increased tumor growth, cell invasion and anti-apoptotic effects [12, 13].

In glioblastoma, ALK fusion proteins and ALK muta- tions are rare events with occurrence of only 1% [14, 15]. However, ALK expression was detected in up to 60% of glioblastoma tissue with 20% of glioblastoma demonstrat- ing intensive and 41% moderate ALK immunostaining [16]. Furthermore, ALK expression was more frequently found in younger patients [17] and in glioblastoma com- pared to grade II or III astrocytoma [18-20] and was shown to be associated with a worse prognosis of glioma patients [18, 19]. Several preclinical studies demonstrated that already the presence of ALK expression was sufficient to activate ALK-mediated tumorigenic functions in glioblas- toma. ALK depletion by knockdown or inhibitory single- chain antibodies resulted in reduced tumor growth, inhi- bition of invasion and induction of apoptosis in U87MG glioblastoma cells [20, 21]. These anti-tumorigenic effects were further increased by double knockdowns of ALK and pleiotrophin [22]. In addition, activation of the ALK path- way is involved in neovascularization of glioblastoma [19] and required for self-renewal and tumorigenicity of glio- blastoma stem cells [23]. Only sparse clinical data about ALK inhibitory strategies in glioblastoma show some effi- cacy of the ALK inhibitor crizotinib in one case report of recurrent glioblastoma with ALK and cMET overexpres- sion [24].

Alectinib belongs to the targeted therapies in the ongo- ing NCT Neuro Master Match (N2M2)/NOA-20 trial which intends to translate molecular diagnostics into personal- ized treatments in glioblastoma patients [25, 26] and will thereby provide more insights into the clinical use of ALK inhibitors. The inclusion criteria for an assignment of patients to alectinib treatment require detection of ALK translocations or ALK mutations in the glioblastoma tis- sue. Consequently, alectinib treatment will be restricted to a very small subset of glioblastoma patients. One aim of the present preclinical assessment was therefore to analyze whether ALK expression on its own could be a sufficient biomarker for response to alectinib treatment. Further- more, acquired resistance mechanisms and potential sal- vage strategies to overcome resistance were investigated.

Clonogenicity assay

Clonogenic capacity (“clonogenicity”) was analyzed by spheroid and limiting dilution assays (LDA) as previously described [29, 30].

Proliferation assays

Proliferation was measured with 3H-Thymidine incorpo- ration assays. 1000 cells per well were seeded in 96-well plates, treated as indicated and pulsed after 6 days with 3H-methylthymidine (0.5lCi) (Amersham Radiochemical Centre, UK) for 24 h. Radionuclide uptake was measured by scintillation counting.

Flow cytometry

For analysis of apoptosis, cells were stained 72 h after treat- ment with APC-Annexin V (1:100, eBioscience, Germany) and DAPI (1:250, Invitrogen, Carlsbad, USA) and analyzed with BD-FACS Canto II flow cytometer. Data were pro- cessed with FlowJo software (Treestar) and cells positive for DAPI and Annexin V were quantified.

Quantitative real‑time PCR (qRT‑PCR)

RNA extraction, cDNA synthesis and qRT-PCR were per- formed as previously described [29, 31]. All results were normalized to the housekeeping gene glyceraldehyde- 3-phosphate dehydrogenase (GAPDH). Primer sequences are listed in Supplementary Table 1. For molecular analysis of treatment with trametinib and cMyc inhibitor, higher con- centrations were used for qRT-PCR analysis compared to proliferation and clonogenicity assays as cells were treated only for 24 h compared to 6 days in proliferation or 3 weeks in clonogenicity assays.

ALK sequencing

RNA and cDNA of alectinib resistant S24 and T269 and respective DMSO control cells were extracted as previously described [31]. Fragments of the tyrosine kinase domain including exons 20–28 of the ALK gene were amplified using PCR technique and sequenced with Sanger sequenc- ing (Eurofins genomics, Deutschland). Primer sequences are listed in Supplementary Table 1.

Immunoblot analysis

Whole cell lysates and immunoblot analysis were per- formed as described in supplementary methods. Follow- ing antibodies were used: sheep ALK, mouse cMyc (both R&D-systems, Wiesbaden-Nordenstadt, Germany), rab- bit phospho-ALK (Tyr1604), rabbit phospho-ERK1/2 and rabbit ERK1/2 (all Cell Signaling, Cambridge, UK). Equal protein loading was controlled with mouse α-Tubulin (Sigma Aldrich, USA) or GAPDH (Linaris, Germany) staining.

Microarray and phospho‑proteomic analysis

Microarray and mass spectrometry based proteomic analysis was done by the Genomics and Proteomics Core Facility of the German Cancer Research Center (Heidelberg, Ger- many) using Illumina HumanHT-12v4 Expression Bead Chips or Q-Exactive HF-X mass spectrometers, respectively. For microarray analysis three independent samples of S24 cells treated with 500 nM alectinib or DMSO for 24 h were used for comparison (Gene expression omnibus number: GSE140145). For phospho-proteomics three independent samples of alectinib resistant S24 and T269 and respective DMSO control cells were used for comparison. Data analy- sis was performed with Ingenuity® Pathway Analysis (IPA, Ingenuity Systems Inc., USA).

Statistical analysis

Statistical significance was calculated using Student’s t-test (Excel, Microsoft, USA) or one-sample t-test as appropriate (Graph Pad Software). P-values with p < 0.05 were consid- ered significant. Figures consisted of at least three independ- ent experiments. Only results of immunoblot and qRT-PCR analysis were shown as one representative out of three inde- pendent experiments. ImageJ was used for quantification of immunoblot data. Synergistic effects were analyzed based on at least three independent experiments. Observed inhibi- tion of combination therapy and expected inhibition under independence of the individual therapies were calculated using Bliss’ independence method [32] and were demon- strated together with p-values of over-additive interaction in respective figures to illustrate synergisms. Synergistic effect of the combination therapy was tested in a linear regression model based on log-transformed inhibition measurements with main therapy effects, an interaction term between both therapies and the experiment effect to account for paired measurements as predictors. An over-additive significant interaction was interpreted as synergism. P-values < 0.05 were considered significant.

Results
ALK expression as biomarker for response to alectinib

Expression of ALK protein and mRNA were detected in sev- eral glioma initiating cells (GICs), but not or only to a very low magnitude in glioblastoma cell lines if expression levels were compared to the highly ALK expressing GICs (Fig. 1a, Suppleentary Fig. 1A). Alectinib treatment dose-depend- ently reduced the proliferation of ALK expressing GICs (S24, T1 and T269) starting at concentrations of 100 nM or 500 nM (Fig. 1b). In contrast, GICs without ALK expression (T325 and PJ) demonstrated only minimal responses at high concentrations of 1 µM alectinib. Alectinib treatment had no significant impact on proliferation of glioblastoma cell lines in concentrations up to 1 µM (Fig. 1b).

Further analyses of alectinib effects and resistance mech- anisms were performed in the GICs S24 and T269. These progenitor cell lines were chosen to substantiate the results in two ALK positive cells with different expression levels of ALK and to thereby best reflect the reported magnitude of moderate to intensive ALK expression in glioblastoma tissue [16].

Knockdown of ALK was performed in ALK express- ing S24 and T269 cells and confirmed by immunoblot and qRT-PCR analyses (Fig. 1c, Supplementary Fig. 1B). ALK knockdown cells were significantly less proliferative and clonogenic than respective vector control cells (Fig. 1d). Alectinib treatment was less effective in ALK depleted S24 and T269 cells regarding reduction of clonogenicity com- pared to respective ALK expressing vector control cells (Fig. 1e).

Effects of alectinib treatment in ALK positive glioblastoma cells

Alectinib reduced the expression of phospho-ALK in ALK positive S24 cells shown by immunoblot analysis (Fig. 2a). Inhibition of the ALK pathway by alectinib reduced the pro- liferation (Fig. 2b) and clonogenicity (Fig. 2c) of ALK posi- tive S24 and T269 cells at nanomolar concentrations starting at 250 nM with T269 cells being even more sensitive to alec- tinib treatment than S24 cells. In contrast, proliferation and clonogenicity of ALK negative T325 cells was only reduced by high concentrations of 750 nM (Fig. 2b, c). In accord- ance, alectinib dose-dependently induced apoptosis in ALK positive S24 and T269 cells, but not in T325 cells (Fig. 2d). Effects of combined alectinib treatment and radiotherapy were further investigated in ALK positive S24 and T269 cells. Here, combined treatment of alectinib and radiother- apy showed synergistic effects on inhibition of clonogenicity
in these cells (Fig. 2e).

Effects of alectinib treatment were mediated by inhibition of cMyc

Microarray analysis of S24 cells treated for 24 h with alec- tinib or DMSO control revealed the cMyc pathway as the most relevantly regulated pathway, which was inhibited by alectinib treatment (Supplementary Fig. 2A). Reduction of cMyc mRNA expression by alectinib was confirmed by qRT- PCR in S24 and T269 cells (Fig. 3a). To further evaluate the relevance of this pathway for alectinib mediated treatment effects, cMyc was exogenously overexpressed or specifically inhibited. S24 cells with overexpression of cMyc (Fig. 3b) showed marked resistance against short-term alectinib treat- ment as inhibition of clonogenicity by alectinib treatment was lower in highly cMyc expressing compared to vector control transfected S24 cells (Fig. 3c). On the contrary, combining alectinib treatment with pharmacological cMyc inhibition resulted in synergistic effects on inhibition of clo- nogenicity in S24 wildtype cells (Fig. 3d). These synergistic effects were confirmed in T269 cells (Fig. 3d) and further substantiated by evaluating the inhibition of proliferation with combined treatment in these cells (Supplementary Fig. 2B).

Acquired resistance against alectinib

Alectinib resistant S24 and T269 cells were achieved by continuous treatment with increasing alectinib doses over 3 months. Acquired resistance of alectinib resistant S24 and T269 cells was confirmed by short-term treatment with alectinib at clinically relevant doses (Fig. 4a). Whereas in NSCLC acquired ALK mutations are frequently described as resistance mechanism against ALK inhibitors [33, 34], sequencing of the tyrosine kinase domain of the ALK recep- tor in alectinib resistant S24 and T269 cells did not dem- onstrate any ALK mutations as driver of resistance in our preclinical model.
The cMyc pathway, which was inhibited by short-term alectinib treatment, was induced in alectinib resistant S24 and T269 cells substantiated by a higher cMyc protein and mRNA expression compared to DMSO control cells (Fig. 4b). Similar to the effects in non-resistant cells, phar- macological inhibition of the cMyc pathway demonstrated synergistic activity in combination with short-term alec- tinib treatment on reduction of clonogenicity (Fig. 4c) and proliferation (Supplementary Fig. 3) in alectinib resist- ant S24 and T269 cells. The enhanced inhibition of cMyc pathway by combined treatment compared to single treat- ments was further verified by a stronger reduction of the expression of cMyc downstream targets with combined treatment compared to respective monotherapies (Supple- mentary Fig. 4).

Phospho-proteomic analysis demonstrated an activation of the ERK1/2 pathway in alectinib resistant S24 and T269 cells, which was confirmed by an increased expression of phospho-ERK1/2 in immunoblot analysis (Fig. 4d, Supple- mentary Fig. 5). As short-term alectinib treatment did not affect the ERK pathway, this pathway seems to represent an activated bypass at alectinib resistance. The combination of trametinib, a clinically approved MEK inhibitor, with short- term alectinib treatment resulted in synergistic inhibition of clonogenicity and proliferation in alectinib resistant S24 and T269 cells and thereby re-sensitized alectinib resistant cells to short-term alectinib treatment at reasonable, i.e. clinically relevant, doses (Fig. 4e, Supplementary Fig. 6).

On the molecular level, trametinib reduced the mRNA expression of cMyc in alectinib resistant S24 and T269 cells indicating that trametinib also has an inhibitory effect on the enhanced cMyc pathway (Fig. 5a). Combining the cMyc and ERK inhibitors with alectinib demonstrated further synergis- tic effects compared to respective dual treatments of cMyc or ERK inhibitor and alectinib in alectinib resistant S24 and T269 cells (Fig. 5b).
Due to the demonstrated synergistic effects of radio- therapy and alectinib in S24 and T269 wildtype cells, we further investigated the efficacy of radiotherapy at alectinib resistance. In accordance to the data in non-resistant cells, combined treatment of alectinib and radiotherapy also dem- onstrated synergistic effects in alectinib resistant S24 and T269 cells (Fig. 5c).

Discussion

Alectinib effectively inhibited proliferation as well as clo- nogenicity and induced apoptosis in nanomolar concen- trations in ALK expressing glioblastoma cells, whereas cells without ALK expression were primary resistant against alectinib. The dependency of alectinib efficacy on the expression of its target ALK was further sustained by depletion of ALK via knockdowns. In addition, alectinib treatment was even more effective in T269 cells which had a higher ALK expression compared to S24 cells which express ALK in a moderate magnitude. However, both progenitor cultures demonstrated a significant sensitivity to alectinib treatment in clinically relevant concentrations. Pharmacokinetic studies reported a maximal plasma con- centration (Cmax) of 1.4 µM (± 0.28 µM) after adminis- tration of the recommended dosage of 600 mg alectinib twice daily [35] and the brain to plasma ratio was shown to range from 0.63 to 0.94 in preclinical studies [36]. Based on these data, the concentrations for alectinib treatment applied in this preclinical analysis were below the clini- cally relevant range. Furthermore, the value of ALK path- way inhibition in brain tumor therapy is supported by pre- viously published data demonstrating that ALK mediated tumorigenic effects in glioblastoma cells were abolished after ALK depletion via knockdowns or single-chain anti- bodies [20, 21] which was also confirmed by ALK knock- down experiments in the present study.

Therefore, the presented data provide an important rationale to determine also ALK expression status as bio- marker for response to alectinib treatment and substan- tially support the use of alectinib as targeted therapy in ALK expressing glioblastoma. The latter is further encour- aged by the high brain penetrance of alectinib [4, 5] ensur- ing that alectinib is actually reaching its target also in brain tumors. As around 60% of glioblastoma express ALK [16, 19, 20], alectinib might represent a treatment option for a high proportion of glioblastoma patients. Therefore, the reported results should be further substantiated in ortho- topic glioblastoma mouse models and in clinical trials to translate the results into clinical practice.
Interestingly, ALK expression was significantly higher in several glioma initiating cells than in the glioblas- toma cell lines tested in this study. As ALK is transiently expressed during the development of the central nervous system [18, 37] the data might indicate a role of ALK in gliomagenesis. Furthermore, although the differences in ALK expression status of U87MG cells compared to previously published results cannot fully be explained, the primary glioblastoma cell cultures used in this preclini- cal study might better reflect the glioma biology of glio- blastoma patients. Therefore, our results relevantly extend previous studies about ALK in glioblastoma, which were mostly conducted in the cell line U87MG [20, 21].

Furthermore, investigating the underlying molecular mech- anisms, the cMyc pathway was deciphered to be an important mediator to translate alectinib effects but also primary and acquired resistance. Short-term alectinib treatment inhib- ited cMyc in non-resistant cells and high cMyc expression conferred primary and acquired resistance against alectinib. The involvement of cMyc was further substantiated by phar- macological cMyc inhibition which sensitized non-resistant and resistant S24 and T269 cells to short-term alectinib treat- ment. The role of cMyc in mediating treatment effects of ALK inhibitors was further supported by similar findings in NSCLCs in which overexpression of cMyc reduced sensitivity to alectinib [38] and inhibition of cMyc by RNAi sensitized cells to another ALK inhibitor, critzotinib.[39].

In addition, the activation of the ERK1/2 pathway, which represents one downstream signaling pathway of ALK [12, 40], was identified to confer acquired alectinib resistance. Inhibition of this pathway by trametinib re-sensitized resist- ant cells to alectinib treatment. On the molecular level,
trametinib also reduced the expression of cMyc, which might further enhance the re-sensitizing effects to alectinib (Fig. 5d). These effects were further substantiated by effects of a triple therapy with trametinib, cMyc inhibitor and alec- tinib which led to additional treatment efficacy compared to respective dual treatments in alectinib resistant cells.

In clinical practice treatment is often stopped after emer- gence of resistance and replaced by another salvage therapy. However, the synergistic effects of combined trametinib or cMyc inhibitor and alectinib treatment in alectinib resist- ant cells might suggest that alectinib should be continued and combined with these inhibitors as salvage therapy once resistance has occurred. These re-sensitizing features indi- cate that alectinib resistant cells keep dependency on ALK signaling after resistance emerged. This maintenance of ALK dependency is further supported by the efficacy of combinational treatment approaches with cMyc inhibitors and radiotherapy which demonstrated synergistic effects in both non-resistant and resistant cells and by clinical data in NSCLCs in which ALK inhibitory treatments can be effec- tively switched to another ALK inhibitor after resistance has developed by acquisition of ALK mutations [41].

Furthermore, the results provide a relevant rationale for combining alectinib with radiotherapy at first-line therapy and at recurrent disease in glioblastoma patients.

Conclusions

The presented data provide a rational to determine ALK expression as additional biomarker apart from ALK fusions and ALK mutations for alectinib treatment in glioblastoma and support the use of alectinib for the treatment of ALK expressing glioblastoma which should be further evaluated in orthotopic glioblastoma mouse models and in clinical trials to translate results into clinical practice. The cMyc pathway was identified as an important pathway to mediate alectinib treatment effects and to be involved in acquired resistance to alectinib. In addition, the ERK1/2 pathway was activated at alectinib resistance. Pharmacological inhibition of the activated cMyc and ERK1/2 pathway re-sensitized alectinib resistant cells to short-term alectinib treatment and thereby represent reasonable salvage therapies after resistance against alectinib has emerged. Importantly, the data support the addi- tion of these treatments to continued alectinib therapy due to the observed synergistic effects. In addition, the results provide a preclinical rational for combination of radiotherapy and alectinib as first-line therapy and at recurrent disease. Alectinib belongs to the targeted therapies of the ongoing N2M2 trial [25] providing an ideal opportunity to further investigate the described resistance mechanisms and pro- posed salvage therapies in glioblastoma patients.