The role of GSK3 and its reversal with GSK3 antagonism in everolimus resistance

in Endocrine-Related Cancer

Pancreatic neuroendocrine tumors (panNETs) are often inoperable at diagnosis. The mTORC1 inhibitor everolimus has been approved for the treatment of advanced NETs. However, the regular development of resistance to everolimus limits its clinical efficacy. We established two independent everolimus-resistant panNET (BON1) cell lines (BON1 RR1, BON1 RR2) to find potential mechanisms of resistance. After 24 weeks of permanent exposure to 10 nM everolimus, BON1 RR1 and BON1 RR2 showed stable resistance with cellular survival rates of 96.70% (IC50 = 5200 nM) and 92.30% (IC50 = 2500 nM), respectively. The control cell line showed sensitivity to 10 nM everolimus with cellular survival declining to 54.70% (IC50 = 34 nM). Both resistant cell lines did not regain sensitivity over time and showed persistent stable resistance after a drug holiday of 13 weeks. The mechanisms of resistance in our cell line model included morphological adaptations, G1 cell cycle arrest associated with reduced CDK1(cdc2) expression and decreased autophagy. Cellular migration potential was increased and indirectly linked to c-Met activation. GSK3 was over-activated in association with reduced baseline IRS-1 protein levels. Specific GSK3 inhibition strongly decreased BON1 RR1/RR2 cell survival. The combination of everolimus with the PI3Kα inhibitor BYL719 re-established everolimus sensitivity through GSK3 inhibition and restoration of autophagy. We suggest that GSK3 over-activation combined with decreased baseline IRS-1 protein levels and decreased autophagy may be a crucial feature of everolimus resistance, and hence, a possible therapeutic target.

Abstract

Pancreatic neuroendocrine tumors (panNETs) are often inoperable at diagnosis. The mTORC1 inhibitor everolimus has been approved for the treatment of advanced NETs. However, the regular development of resistance to everolimus limits its clinical efficacy. We established two independent everolimus-resistant panNET (BON1) cell lines (BON1 RR1, BON1 RR2) to find potential mechanisms of resistance. After 24 weeks of permanent exposure to 10 nM everolimus, BON1 RR1 and BON1 RR2 showed stable resistance with cellular survival rates of 96.70% (IC50 = 5200 nM) and 92.30% (IC50 = 2500 nM), respectively. The control cell line showed sensitivity to 10 nM everolimus with cellular survival declining to 54.70% (IC50 = 34 nM). Both resistant cell lines did not regain sensitivity over time and showed persistent stable resistance after a drug holiday of 13 weeks. The mechanisms of resistance in our cell line model included morphological adaptations, G1 cell cycle arrest associated with reduced CDK1(cdc2) expression and decreased autophagy. Cellular migration potential was increased and indirectly linked to c-Met activation. GSK3 was over-activated in association with reduced baseline IRS-1 protein levels. Specific GSK3 inhibition strongly decreased BON1 RR1/RR2 cell survival. The combination of everolimus with the PI3Kα inhibitor BYL719 re-established everolimus sensitivity through GSK3 inhibition and restoration of autophagy. We suggest that GSK3 over-activation combined with decreased baseline IRS-1 protein levels and decreased autophagy may be a crucial feature of everolimus resistance, and hence, a possible therapeutic target.

Introduction

Pancreatic neuroendocrine tumors (panNETs) are a type of gastroenteropancreatic neuroendocrine tumor (GEP-NETs) (Yao et al. 2008). Due to their slow and frequently asymptomatic growth, panNETs are often advanced when diagnosed, such that curative surgery is no longer an option (Briest & Grabowski 2014). For this reason, the 10-year survival for panNETs is only 44%, in spite of various advances in treatment modalities (Ekeblad et al. 2008, Halfdanarson et al. 2008, Auernhammer et al. 2017, Aristizabal Prada & Auernhammer 2018). Common treatments for advanced disease such as biotherapy, chemotherapy, molecular-targeted therapy and peptide radioreceptor therapy remain limited in their efficacy, and novel treatment options are urgently needed (Frilling et al. 2014, Auernhammer et al. 2017, Aristizabal Prada & Auernhammer 2018).

One of the most important receptor tyrosine kinase (RTK)-induced oncogenic signaling cascades in GEP-NETs is the mTORC1-p70S6K signaling pathway, which can be inhibited by the mTORC1 inhibitor everolimus (Fig. 1) (Zaytseva et al. 2012, Wolin 2013, Shipkova et al. 2016). The PI3K–Akt–mTORC1 axis is frequently over-activated in cancers, including panNETs, causing increased cellular survival, proliferation and protein synthesis (Briest & Grabowski 2014, Pavel & Sers 2016). Accordingly, the mTORC1 inhibitor everolimus has shown anti-tumor potential in several clinical phase 3 trials in patients with NETs (RADIANT-2, RADIANT-3, RADIANT-4) (Pavel et al. 2011, Yao et al. 2011, 2016a,b), and has been approved for the treatment of pancreatic, intestinal and pulmonary NETs in many countries (Pavel & Sers 2016, Chan et al. 2017). Unfortunately, only around 5% of patients show partial remission in response to everolimus treatment, although many more patients show stable disease. However, most tumors eventually start to progress even if everolimus treatment is continued (Yao et al. 2011). Thus, secondary resistance seems to occur in most patients, although primary resistance may also be a problem (Wagle et al. 2014, Capozzi et al. 2015, He et al. 2016, Vandamme et al. 2016, Zatelli et al. 2016). Advanced panNET patients are treated with a fixed dose of 10 mg everolimus daily (Nbib510068) (Yao et al. 2013, 2014): the majority develops treatment resistance within 1 year of treatment (Capurso et al. 2012, Yao et al. 2013). Consequently, reversible adaptive resistance to everolimus in panNET cell lines, and combinatory regimes to overcome acquired resistance, have been the object of previous studies (Zitzmann et al. 2010, Passacantilli et al. 2014, Vandamme et al. 2016). However, long-term resistance to everolimus treatment has only been little studied to date (Capurso et al. 2012, Passacantilli et al. 2014, Vandamme et al. 2016), and thus, there is an unmet need to better understand the mechanisms of resistance and to establish predictive biomarkers for resistance to everolimus, and hence, appropriate treatments. Activation of autophagy by mTORC1 inhibitors has been reported as one potential mechanism for developing mTORC1 inhibitor resistance and for promoting tumor cell growth (Avniel-Polak et al. 2016, Avniel et al. 2018). Autophagy seems to play a controversial role in cancer: on the one hand, autophagy might offer a protective mechanism in some cancer cells promoting acquired drug resistance (Sui et al. 2013, Avniel-Polak et al. 2016). On the other hand, autophagy might also induce autophagic cell death, separate from apoptosis (Sui et al. 2013).

Figure 1
Figure 1

Proposed mechanism of action of the mTORC1-inhibitor everolimus. Everolimus inhibits the mTORC1 complex and thereby down-regulates the mTORC1/p70S6K signaling cascade, preventing the cell from cell cycle progression, survival and proliferation. Inhibition of mTORC1/p70S6K leads to disinhibition/activation of IRS-1 and thereby activation of PI3K/Akt, which in turn inhibits GSK3 by phosphorylation. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Novel insights into the mechanisms of resistance in response to everolimus treatment could provide a rationale for dual-targeting approaches in order to overcome such acquired or intrinsic resistance (Capurso et al. 2012, Passacantilli et al. 2014, Capozzi et al. 2015, Vandamme et al. 2016, Avniel et al. 2018). Although the study group around Vandamme et al. treated BON1 and QGP1 cell lines over 24 weeks with increasing doses of everolimus up to 1 µM according to the ‘pulse selection model’ (McDermott et al. 2014), BON1 and QGP1 cells only showed reversible resistance and regained sensitivity to everolimus after a drug holiday of 10–12 weeks (Vandamme et al. 2016).

Therefore, our aim was to establish a clinically relevant panNET cell line model from BON1 cells, demonstrating stable, that is, consistent and reliable resistance to everolimus in order to study the mechanisms of long-term resistance and explore novel combinatory treatment options to overcome such resistance.

Besides exploring PI3K-Akt signaling as a potential pathway of resistance (O’Reilly et al. 2006, Zitzmann et al. 2010, Passacantilli et al. 2014, Vandamme et al. 2016), we also investigated GSK3 signaling. GSK3α/β is a direct effector protein of Akt: activated Akt phosphorylates and inhibits GSK3α/β (Porta et al. 2014). GSK3 has been implicated in metabolism since GSK3 is inactivated by insulin leading to disinhibition/activation of glycogen synthase. Moreover, it seems to play a role in diseases such as diabetes, Alzheimer’s disease and psychiatric diseases, including bipolar disorder and schizophrenia (Maurer et al. 2014). GSK3α/β has shown controversial effects on cell death and survival, including pro- and anti-apoptotic effects (Maurer et al. 2014). Activated GSKβ has also been reported to target cell cycle proteins such as Cyclin D for degradation (Maurer et al. 2014). On the one hand, treatment with different GSK3 inhibitors (CHIR99021, 6-bromoindirubin-3′oxime-, 1-azakenpaullone) and siRNA stimulated proliferation of rat insulinoma INS-1E cells (Mussmann et al. 2007). Conversely, the non-specific GSK3α/β inhibitor lithium chloride inhibited neuroendocrine tumor growth in vitro and in vivo (Greenblatt et al. 2010, Lubner et al. 2011); however, in a clinical phase 2 study with 15 patients with low-grade NETs (8 carcinoid tumors, 5 islet cell tumors), the non-specific GSK3α/β inhibitor lithium chloride caused no objective responses, and pre- and post-treatment tumor biopsies showed no consistent GSK3β inhibitory effects due to insufficient phosphorylation of GSK3β in patients’ serum levels (Lubner et al. 2011). To assess the effects of more specific GSK3α/β inhibition on NETs, we have recently investigated the selective GSK3α/β inhibitor AR-A014418 in pancreatic, pulmonary and midgut NET cells and reported a significant inhibitory effect on cell viability and strong GSK3α/β phosphorylation/inhibition (Aristizabal Prada et al. 2017). Biochemical studies have focused on GSK3β inhibition, while other studies have also shown the importance of GSK3α in drug resistance (McCubrey et al. 2014). Thus, GSK3 appears to play a pleiotropic role in regulating cell proliferation.

We have used our everolimus-resistant cell line model to demonstrate combinatorial treatments that reversed everolimus resistance via GSK3 inhibition in vitro. In due course, other cell lines can be studied to explore whether they also develop stable resistance and to compare the intrinsic mechanisms. In a pilot study, we transferred our everolimus-resistant cell line model to an orthotopic intrapancreatic xenograft mouse model providing the basis for translating the in vitro approach to an in vivo situation.

Materials and methods

Cell culture

The human panNET cell line BON1 (Evers et al. 1991) was kindly obtained from Prof. R Göke (Marburg, Germany). The BON1 cell line was analyzed and certified by the German Biological Resource Centre DSMZ (DSMZ, Braunschweig, Germany) using short tandem repeats (STRs). BON1 cells were grown in DMEM/F12 (1:1) (Life Technologies), supplemented with 10% FBS (Biochrom, Berlin, Germany), 1% penicillin/streptomycin (Life Technologies) and 0.4% amphotericin B (Biochrom) at 5% CO2 and 37°C and tested mycoplasma-free for all experiments reported. BON1 is a permanent cell line derived from a lymph node metastasis of a human carcinoid tumor of the pancreas and presents phenotypic NET characteristics (Grozinsky-Glasberg et al. 2012).

Experimental design

In order to yield everolimus resistance in panNET cells, BON1 cells (passage 13) were treated continuously with a constant low dose of 10 nM everolimus for 24 weeks. A dose of 10 nM everolimus was chosen since the plasma levels of everolimus in patients on daily therapy with 10 mg everolimus range between around 8 nM (C min after 2 weeks of daily exposure to 10 mg everolimus) and 59 nM (C 2h after 12 weeks of daily exposure to 10 mg everolimus) (Budde et al. 2016). Every 3-day medium was exchanged and everolimus treatment was renewed. The concentrations were not increased over time. To determine commonalities in resistance mechanisms in BON1 cells, we have established two independent everolimus-resistant BON1 cell lines (BON1 RR1 and BON1 RR2) in parallel and a control cell line (BON1 Control DMSO) to control for adaptive strategies of long-term in vitro culturing and passaging as such. BON1 is the initial parental cell line, from which the development of resistance started. To obtain a sensitive control cell line (BON1 Control DMSO), BON1 was continuously vehicle treated with a comparable concentration of DMSO (0.00005%) over 24 weeks. We compare our results to the BON1 Control DMSO cell line. In order to assess the stability and reliability of the acquired resistance phenotype, all cells were frozen in liquid nitrogen (minimum for 48 h) and thawed, followed by drug withdrawal (McDermott et al. 2014). Resistance was defined as at least a doubling of the IC50, compared to the sensitive control cell line (BON1 Control DMSO), after treatment with 10 nM everolimus for 144 h (McDermott et al. 2014).

Inhibitors

The mTORC1 inhibitor everolimus (RAD001), the GSK3 inhibitor AR-A014418 and the PI3K inhibitor BYL719 were acquired from Sigma-Aldrich. All substances were dissolved in dimethyl-sulfoxide (DMSO, Sigma-Aldrich) and stored at −20°C. The stock solutions of everolimus, AR-A014418 and Byl719, were 20 mM in DMSO.

Cell viability assessment

BON1 cells were seeded at ~1500 cells/96-well for 24 h in complete medium. Then, the cells were treated with different concentrations of everolimus (from 1 nM to 2 µM), BYL719 (from 1 µM to 10 µM) or AR-A014418 (from 5 µM to 20 µM) alone or they were treated with everolimus (from 1 nM to 10 nM) in combination with BYL719 (from 1 µM to 10 µM) or in combination with AR-A014418 (from 5 µM to 20 µM). The BYL719 concentrations we chose are potentially clinicially relevant since in a clinical study assessing population pharmacokinetics and pharmacodynamics of BYL719 in patients with advanced solid malignancies median BYL719 plasma concentrations between 2 µM and 11 µM have been found with 1-month daily oral therapy with therapeutically relevant doses of 300–450 mg BYL719 2–8 h post dose (De Buck et al. 2014). For everolimus alone, the two final DMSO controls were equal to 1 µM everolimus (0.005% DMSO final concentration) and 2 µM everolimus (0.01% DMSO final concentration), respectively, depending on the experiment. For the higher dose BYL719/everolimus combination treatment, the final DMSO control was equal to 1:2000 (Byl719) + 1:2,000,000 (everolimus) = 0.05005% final DMSO concentration. For the lower dose BYL719/everolimus combination treatment, the final DMSO control was equal to 1:4000 (Byl719) + 1:2,000,000 (everolimus) = 0.02505% final DMSO concentration. For AR-A014418/everolimus combination treatment, the final DMSO control was equal to 1:1000 (AR-A014418) + 1:2,000,000 (everolimus) = 0.10005% final DMSO concentration. The final DMSO concentrations were never higher than 0.1% and had no significant effect on cell viability. The metabolic activity of living cells was assessed by the ‘Cell Titer Blue’ cell viability assay (Promega) after 72 h and 144 h of incubation with the different agents. Cells were incubated for 4 h with ‘Cell Titer Blue’ solution and fluorescence was measured at 560/590 nm using a GLOMAX plate reader (Promega).

Cell cycle analysis by flow cytometric analysis

Cell cycle phase distribution was analyzed (Riccardi & Nicoletti 2006) (BD Accuri C6 Analysis). First, cells were cultured in six-well plates for 24 h, before medium was replaced by antibiotic-free medium and incubated with 10 nM everolimus. Seventy-two hours later, cells were collected, washed and resuspended in 300 μL propidium iodide (Sigma-Aldrich). After 8 h, 20,000 events from each sample were analyzed.

Caspase-3/-7 activity assay

To measure the activity of Caspase-3/-7, we used the Apo-One homogeneous Caspase-3/7-Assay kit (Promega): 10,000 cells were seeded per well, incubated for 24 h without treatment or with respective doses of BYL719 (5–10 µM) or everolimus (10 nM), and Caspase-3/7 activity was assessed according to the manufacturer’s instructions. We compared all results of the BYL719- or everolimus-treated cell lines to the vehicle-treated cell lines. For each experiment, the vehicle represented the highest DMSO concentration present in the respective experiment. The exact DMSO concentrations are detailed under ‘Cell viability assessment’ section.

Protein extraction and Western blotting

The cells were seeded into 10 cm plates and grown for 24 h in complete medium for Western blot analysis. Following medium exchange, cells were incubated with 10 nM everolimus, either alone or in combination with BYL719 or AR-A014418. We compared all results to the vehicle-treated cells. For each experiment the vehicle represented the highest DMSO concentration present in the respective experiment. The exact DMSO concentrations are detailed under ‘Cell viability assessment’ section. The incubation times were up to 72 h. Western blotting was conducted as described previously (Reuther et al. 2016). A cell line screening to determine commonalities between both resistant cell lines (BON1 RR1 and BON1 RR2) in protein expression was performed using the following primary antibodies: pIRS-1 Ser612 (#3203) (1:1000), IRS-1 (#3407) (1:1000), pAkt (Ser473) (#4060) (1:5000), Akt (#2920) (1:5000), pTORC1 (S151) (#3359) (1:1000), TORC1 (#2587) (1:000), pp70S6K (T387/389) (#9205) (1:2000), p70S6K (#9234) (1:5000), pS6 (S240/244) (#5354) (1:50000), S6 (#2217) (1:5000), p4EBP1 (Ser65) (#9451) (1:1000), 4EBP1 (#9644) (1:10000), pCDK1 (Tyr15) (#4539) (1:1000), CDK1 (#9116) (1:2000), pEGFR (Y1068) (#3777) (1:1000), EGFR (#4267) (1:5000), pMet (Y1234/5) (#3077) (1:2000), Met (#3127) (1:5000), pGSK3 (S21/9) (#9331) (1:2000), GSK3 (#5676) (1:10000), LC3-A (#4599) (1:1000), Caspase 3 (#9662) (1:1000), cleaved Caspase 3 Asp175 (#9661) (Cell Signaling) (1:1000), Actin (A5441) (Sigma) (1:50000), Bcl2 (610539) (BD Transduction Laboratories, Franklin Lakes, NJ, USA) (1:1000), CgA (ab68271) (Abcam) (1:100,000).

Cell migration assay

Cells were seeded at densities of 120,000 (BON1) cells per chamber in culture inserts (Ibidi, Munich, Germany). After 24 h, the inserts were removed, and the cells were treated with 10 nM everolimus for 48 h. We compared the four cell lines BON1, BON1 Control DMSO, BON1 RR1 and BON1 RR2 to each other after treatment with 10 nM everolimus. In this case the BON1 Control DMSO cell line functioned as control and was compared to the resistant cell lines. After 0 h and 48 h of incubation with everolimus, photographs of the gap between the two cell layers were taken (Zeiss, Axiovert 135 TV (microscope) and Zeiss, AxioCam MRm (camera)) and analyzed.

Statistical analysis

The results are displayed as mean ± s.d. of at least three independently performed experiments. Each cell viability experiment consisted of at least 6 samples per substance concentration and incubation period. A priori tests considering the normal distribution and homogeneity of variances were performed applying the Kolmogorov–Smirnov test and Levene’s test of the SPSS statistical package SPSS (version 13.0 for Windows, SPSS Inc. (2005)). When parametric criteria were met, an ANOVA comparison of means with a post hoc Tukey test or a two-tailed t-test was performed; when non-parametric criteria were met the Kruskal–Wallis followed by the Mann–Whitney test was performed. When the one group was normalized to 100%, comparisons were performed as one-sample tests.

To identify potential synergistic effects between BYL719 and everolimus on cell viability, we used Linear Mixed Effects Models. BYL719, everolimus (each in three different doses) and the interactions between both were considered as fixed effects, the trial as a random effect. If a significant negative interaction between BYL719 and everolimus on cell survival was found, we concluded a synergistic effect. One model was estimated for each cell line. These computations were done with R 3.3.3 (GNU General Public License; R Foundation, Vienna, Austria). For the analysis of the baseline differences in cell viability and Caspase 3/7 assay, we used linear mixed effects models too. While the cell line was considered as fixed, the trial was considered as random.

Statistical significance was assessed at P < 0.05, besides for the Western blot analysis, where due to the exploratory character and the sample size, significance was assessed at P < 0.1.

Pilot mouse study: establishment of an orthotopic intrapancreatic everolimus-resistant tumor xenograft mouse model

BON1 RR2 (2 × 106 in 20 µL medium) were implanted into the pancreas of 12-week-old female SCID mice (weight 20 g) (n = 4) as described by Soares et al. (2014) with minor changes. This technique was previously established by us to generate the orthotopic BON1 xenograft mouse model. Briefly, SCID mice were anesthetized by isoflurane and intraperitoneal injection of ketamine (70 mg/kg). After local shaving and disinfection, the abdominal area was opened by a 1 cm incision along the Linea Alba. The entire pancreatic body was then exposed on sterile gauze and retracted laterally on saline-soaked cotton tips in such way that enough solid pancreatic tissue was available for cell injection, before it was placed back into the abdominal cavity. Tumor growth was monitored once a week by a preclinical 1T PET/MRI (nanoScan, Mediso, Hungary) using a whole-body coil and T1-weighted gradient-echo and T2-weighted spin echo sequences. The metabolic activity of the engrafted tumor was estimated by 18F-FDG PET/MRI, and somatostatin receptor 2 (SSTR2) levels were assessed by 68Ga-DOTATOC PET/CT (15 MBq injected i.v. respectively). When the tumors had reached a size of ~1000 mm3, the experiment was terminated. Animal experiments were performed in accordance with the national and local guidelines for animal welfare and approved by the ethics committee of the state Berlin. All in vivo experiments were performed at the Berlin Experimental Radionuclide Imaging Center (BERIC), Charité – Universitätsmedizin Berlin, Germany.

Results

Following 24 weeks of permanent exposure to 10 nM everolimus, we have, for the first time, been able to develop two independent stable everolimus-resistant panNET cell lines (BON1 RR1 and BON1 RR2). Both resistant cell lines did not regain sensitivity over time, and showed stable resistance after a drug holiday of 13 weeks. Moreover, both cell lines showed a GSK3 over-activation as a crucial commonality.

Stable resistance

After 24 weeks of permanent everolimus exposure, both cell lines (BON1 RR1 and BON1 RR2) showed strong reliable resistance to everolimus, whereas BON1 Control DMSO cells displayed no resistance (Fig. 2). We compared dose–response curves of the resistant cell lines to the dose–response curves of the non-resistant cell lines with 1 nM–2 µM everolimus and found resistance, i.e. at least a doubling of the IC50 in the resistant cell lines, compared to the sensitive controls, after 144-h everolimus treatment. IC50 calculations showed a 153-fold increase (IC50 = 5200 nM) in BON1 RR1 and a 74-fold increase (IC50 = 2500 nM) in BON1 RR2 cells when compared to the BON1 Control DMSO cells (IC50 = 34 nM) after 144-h everolimus treatment. BON1 Control DMSO cells demonstrated a significant decrease in cellular survival after re-application of 10 nM everolimus for 144 h, whereas BON1 RR1 and BON1 RR2 cells showed a significant decrease in cellular viability only at 500 nM and 60 nM, respectively (Fig. 2). The resistance was evaluated after freezing and thawing the cells again to confirm stable resistance. We performed additional dose–response curves after a drug holiday of 13 weeks and again found persistent stable resistance with an IC50 >2000 nM in BON1 RR1 and BON1 RR2 cells.

Figure 2
Figure 2

Effects of everolimus (1 nM–2 µM) on cell survival of resistant (BON1 RR1 and BON1 RR2) and sensitive (BON1 Control DMSO) cell lines after 144 h of incubation with everolimus. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to BON1 Control DMSO are shown, considering P < 0.05 as significant (*). A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Morphologic changes, cell growth and neuroendocrine identity

Everolimus-sensitive BON1 cells showed an oval shape with fibroblastic-like cellular extensions in some cells and regular cluster formation (Fig. 3A). Morphologic characteristics of everolimus-resistant cells displayed an amorphous, flattened cellular shape with lysosome-like bubbles and clumping conglomerates (Fig. 3A). The viable cell number of the resistant cell lines was slightly but significantly increased compared to the sensitive control cell line after 72 h without treatment (P < 0.001), suggesting more rapid proliferation (Fig. 3B).

Figure 3
Figure 3

Everolimus resistant cells show morphologic changes. (A) Everolimus-resistant cells (BON1 RR1 and BON1 RR2) displayed an amorphous, flattened cellular shape with lysosome-like bubbles and clumping conglomerates. (B) The viable cell number of the resistant cell lines was slightly but significantly increased compared to the sensitive control cell line after 72 h without treatment (P < 0.001), suggesting more rapid proliferation.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

We investigated the neuroendocrine identity of the resistant cell lines by assessing chromogranin A (CgA) by Western blot techology (Supplementary Fig. 1, see section on supplementary data given at the end of this article) and found strong expression of CgA in both resistant cell lines.

Heterogeneity of BON1 RR1 and BON1 RR2

We observed inhibition of mTORC1/p70S6K signaling to a differential extent between BON1 RR1 and BON1 RR2 after re-application of everolimus (Supplementary Fig. 1). By using two independently established everolimus-resistant cell lines, we have attempted to reduce if not totally eliminate the bias due to tumor heterogeneity in response to long-term everolimus treatment.

Everolimus resistance: G1 phase cell cycle arrest associated with reduced baseline expression of CDK1(cdc2) and of the autophagy marker LC3AII

Both resistant cell lines displayed a significant increase in sub-G1 phase and G1 phase in response to everolimus (Fig. 4A). Associated with the sub G1 increase and G1 phase cell cycle arrest, we found a decreased baseline CDK1(cdc2) expression and reduced baseline expression of the autophagy marker LC3AII demonstrating reduced autophagy in BON1 RR1 and BON1 RR2 cells (Figs 4B and 5).

Figure 4
Figure 4

(A) Cell cycle analysis via FACS showed a significant increase in sub-G1 population and G1 phase cell cycle arrest in both resistant cell lines (BON1 RR1 and BON1 RR2) after 72 h everolimus treatment. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to BON1 Control DMSO are shown, considering *P < 0.05; **P < 0.01; ***P < 0.001. (B) LC3A consists of two bands: LC3AI (upper band) and LC3AII (lower band). Baseline LC3AII expression level measured by Western blot analysis was decreased in BON1 RR1 and BON1 RR2, compared to BON1 and BON1 Control DMSO. A representative blot out of three independently performed experiments is shown.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Figure 5
Figure 5

Western blot analysis via densitometry of CDK1(cdc2) showed a baseline decrease in CDK1(cdc2) expression levels in BON1 RR1 and BON1 RR2 cell lines, compared to the sensitive control cells. A representative blot out of three independently performed experiments is shown. For densitometry analysis the arithmetic means and standard deviation of at least three independent experiments are shown. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Everolimus resistance increased baseline apoptosis

Baseline cleaved caspase 3 as a marker of apoptosis was slightly increased in the resistant cell lines, compared to hardly detectable cleaved caspase 3 in the sensitive controls, indicating slightly stronger baseline apoptosis in the resistant cells (Fig. 6). In order to show these results more clearly, we additionally performed the more sensitive Caspase 3/7 assay and found the same results with significantly increased baseline apoptosis, in both resistant cell lines, compared to the sensitive controls (P < 0.01) (Fig. 6). Baseline expression of the anti-apoptotic marker Bcl2 was much lower in the resistant cell lines compared to the sensitive controls and showed a lesser increase in response to BYL719 and everolimus treatment (Fig. 6).

Figure 6
Figure 6

In the resistant cell lines, Western blot analysis showed a baseline decrease in IRS-1 protein levels, an increase in cleaved Caspase 3 and a decrease in anti-apoptotic Bcl-2 as a sign of increased apoptosis, as well as a decrease in LC3AII as a sign of decreased autophagy. LC3A consists of two bands: LC3AI (upper band) and LC3AII (lower band). Caspase 3/7 assay showed significantly increased apoptosis in both resistant cell lines, compared to the sensitive control cell line (P < 0.01).

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Everolimus resistance increased migration potential and baseline c-Met activation

The migration assay reflected a significant increase in the migration potential of BON1 RR1/RR2 cells, compared to the sensitive control cell line, after treatment with 10 nM everolimus (Fig. 7). Both resistant cell lines displayed stronger baseline activation of c-Met when compared to the control cells (Fig. 8).

Figure 7
Figure 7

Everolimus-resistant cells showed a significantly increased migration potential with a significantly smaller migration gap after 48 h of everolimus treatment. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to BON1 Control DMSO are shown, considering *P < 0.05. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Figure 8
Figure 8

Western blot analysis via densitometry of phosphorylated/activated c-Met and total c-Met in resistant (BON1 RR1 and BON1 RR2) and sensitive control cells showed an increase in phosphorylated/activated c-Met in the resistant cell lines. A representative blot out of three independently performed experiments is shown. For densitometry analysis the arithmetic means and standard deviation of at least three independent experiments are shown. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

Everolimus resistance: baseline GSK3α/β (GSK3) over-activation and strong BON1 RR1/RR2 cell viability reduction by specific GSK3 inhibition

As GSK3 has been shown to be implicated in resistance to radio-, chemo- and molecular-targeted therapy (Shimura 2011, Martelli et al. 2014, McCubrey et al. 2014), we examined long-term treatment effects of everolimus on GSK3 (Fig. 9A). GSK3 is inhibited by phosphorylation. BON1 RR1/RR2 cells displayed a pronounced baseline decrease in phosphorylated/inhibited GSK3, and thus, GSK3 over-activation when compared to the sensitive control cell line (Fig. 9A). However, levels of activated and total Akt were similar in resistant and non-resistant cells with and without treatment (Fig. 9A). Akt was phosphorylated/activated in all resistant and sensitive cell lines after 24-h stimulation with everolimus (Fig. 9A).

Figure 9
Figure 9

GSK3 over-activation in the resistant cell lines. (A) Both everolimus-resistant cell lines showed significantly enhanced activation of GSK3α/β (GSK3) compared to the sensitive control cell line (BON1 Control DMSO) and a significantly stronger activation of GSK3β compared to the parental BON1 cells, but stable phosphorylation/activation of AKT after 24 h of incubation with everolimus. A representative blot out of three independently performed experiments is shown. Statistically significantly different results (*) in comparison to BON1 Control DMSO are shown. (B) Single substance treatment of AR-A014418 (10–20 µM) significantly decreased cellular survival in all cell lines tested after 144 h of incubation, although much stronger in BON1 RR1 and BON1 RR2, compared to the sensitive control cell line. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significant differences in comparison to either single substance treatment are shown, considering ***P < 0.001. (C) Western blot analysis showed inhibition of GSK3 and an increase in the autophagy marker LC3AII in response to the selective GSK3α/β inhibitor AR-A014418. LC3A consists of two bands: LC3AI (upper band) and LC3AII (lower band). A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

GSK3 inhibition by the specific GSK3 inhibitor AR-A014418 caused a pronounced decrease in BON1 RR1/RR2 cell viability and – to a much lesser extent – in sensitive control cells (Fig. 9B). In both resistant and in the sensitive control cell line, Western blot analyses showed clear inhibition of GSK3 and an increase in the autophagy marker LC3AII in response to AR-A014418 treatment (Fig. 9C). However, combination treatment of everolimus plus AR-A014418 did not enhance the inhibitory effect of single AR-A014418 treatment on cell viability in resistant cells (Supplementary Fig. 2).

We also performed Western blots on the combination of everolimus plus AR-A014418 and, while single AR-A014418 treatment led to strong GSK3α inhibition in the resistant cell lines, combination treatment caused no stronger GSK3α inhibiton compared to single AR-A014418 treatment (Supplementary Fig. 2); this may help explain the absence of a synergistic effect.

Everolimus resistance: lower baseline IRS-1 protein levels

Interestingly, baseline IRS-1 protein levels showed a decrease in BON1 RR1 and BON1 RR2 cells compared to parental BON1 and BON1 Control DMSO cells (Fig. 6).

For pIRS-1, the differences between the different cell lines were minor according to the Western blot quantification. Western blot quantifications for IRS-1 and pIRS-1 are shown in Supplementary Fig. 3. The effects of BYL719 and everolimus treatment on IRS-1 are described below.

Total mTORC1 showed no clear differences between resistant cells and non-resistant cells (Supplementary Fig. 4). No activated pmTORC1 could be detected in our cell lines.

Dual-targeting approach with the PI3Kα inhibitor BYL719 overcame everolimus resistance and restored sensitivity through GSK3 inhibition, increased autophagy and decreased apoptosis

Combination treatment with the PI3Kα inhibitor BYL719 plus everolimus overcame everolimus resistance and restored treatment sensitivity (Fig. 10A). A significantly stronger decrease in cellular survival after combination treatment compared to single substance treatment could be observed in both resistant cell lines (Fig. 10A). We moreover studied additional dose–response relationships combining low doses of 1–10 nM everolimus with 1–5 µM BYL719 and found significant synergistic effects at almost all drug doses tested in all cell lines, including both resistant cell lines (*P < 0.05) (Supplementary Figs 5, 6, 7 and 8). This synergistic inhibitory effect on cell survival after the BYL719/everolimus combination treatment was associated with inhibition of GSK3, EGFR, p70S6K and 4EBP1 signaling, as well as showing an increase in IRS-1 protein levels in BON1 RR1/RR2 cells (Figs 6 and 10B). In both sensitive cell lines, everolimus alone already led to strong GSK3 phosphorylation/inhibition. In BON1 RR1/RR2 cells, single application of everolimus led to lower or no phosphorylation/inhibition of GSK3, but the combination treatment induced stronger inhibition of GSK3, compared to each drug separately. A different response of IRS-1 protein levels to everolimus and BYL719 treatment between resistant and non-resistant cell lines was observed: while sensitive control cells showed a slight increase in IRS-1 protein levels after everolimus treatment, the resistant cells showed a lesser or no IRS-1 increase (Fig. 6 and Supplementary Fig. 3A). In the non-resistant cells, BYL719 treatment alone was accompanied by increased IRS-1 protein expression, while in the resistant cells a lesser, but nevertheless clear, increase in IRS-1 protein levels was observed in response to BYL719 treatment (Fig. 6 and Supplementary Fig. 3A).

Figure 10
Figure 10

Dual-targeting approach with the PI3Kα inhibitor BYL719 overcame acquired everolimus resistance and restored sensitivity through inhibition of GSK3, EGFR, p70S6K and 4EBP1 signaling. (A) Combination treatment of everolimus plus BYL719 showed a significantly enhanced inhibitory effect on cell viability compared to single substance treatment in all tested cell lines. Cells were incubated with BYL719 (5 µM and 10 µM) and everolimus (10 nM) alone and in combination for 72 h. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to either single substance treatment are shown, considering *P < 0.05; **P < 0.01; ***P < 0.001. (B) Protein levels of components from the EGFR-Akt (pEGFR, EGFR, pAkt, Akt), mTORC1-p70S6K (pp70S6, p70S6K, pS6, S6, p4EBP, 4EBP1) and GSK3 (pGSK3, GSK3) signaling pathway are displayed. Inhibition of GSK3, EGFR, p70S6K and 4EBP1 signaling was observed after combination treatment. A representative blot out of three independently performed experiments is shown. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Citation: Endocrine-Related Cancer 25, 10; 10.1530/ERC-18-0159

The BYL719/everolimus combination treatment showed a strong increase in IRS-1 expression in the non-resistant BON1 and BON1 Control DMSO cells, and a lesser, but still clear, increase in IRS-1 protein levels in the resistant cells (Fig. 6 and Supplementary Fig. 3A).

We studied the differences concerning cell death mechanisms between resistant and non-resistant cell lines in response to BYL719/everolimus treatment: We found that in BON1 RR1/RR2 cells, BYL719/everolimus combination treatment was accompanied by an increase in the autophagy marker LC3AII, compared to each drug separately, indicating an increase in autophagy in response to the BYL719/everolimus combination treatment (Fig. 6).

In the sensitive control cells, the Caspase 3/7 assay showed apoptosis induction at 24 h with BYL719 treatment, but not in BON1 RR1/RR2; everolimus treatment had no influence on apoptosis induction, while the everolimus/BYL719 combination treatment significantly decreased apoptosis in BON1 RR1/RR2 cells compared to each drug separately, but had no significant influence on apoptosis induction in the sensitive control cells (Supplementary Fig. 9). This was consistent with a decrease in cleaved Caspase 3 in the resistant cell lines after BYL719/everolimus combination treatment, compared to treatment with each drug separately.

Transferability of the resistant cell line model to an orthotopic intrapancreatic xenograft mouse model

In an in vivo pilot study, we were able to transfer our resistant cell line model to an orthotopic xenograft SCID mouse model with intrapancreatic tumor growth (n = 4). Supplementary Figure 10A, B and C show an example of the growth of an intrapancreatic BON1 RR2 tumor in a SCID mouse: 14 days after tumor cell injection into the pancreas, a detectable tumor developed and reached a size of ~1000 mm3 around 5 weeks after first detection by MRI (Supplementary Fig. 10A, B and C). Tumor size and tissue morphology were determined by T2w imaging. The resistant tumors showed a similar growth rate and tumor morphology as the native BON1 tumors. When the tumors grew bigger, necrotic areas in the tumor core increased (Supplementary Fig. 10A, B and C) towards the end of the experiment. The glucose uptake (standard uptake value (SUV)) calculated from 18F-FDG PET confirmed that the tumors were metabolically active (Supplementary Fig. 10E). As expected, 68Ga-DOTATOC PET scanning showed no uptake in the resistant tumors, confirming the absence of somatostatin receptor-2 (Supplementary Fig. 10F).

Discussion

Stable resistance and morphologic changes

In our BON1 in vitro model, permanent long-term treatment with a constant low dose of 10 nM everolimus over 24 weeks caused stable resistance to everolimus. In contrast to the study of Vandamme et al., who found a complete reversal of resistance after a drug holiday of 10–12 weeks (Vandamme et al. 2016), we demonstrated stable persistent resistance after a drug holiday of 13 weeks.

Everolimus-resistant BON1 cells have previously been described as developing an elongated shape with fewer cell-to-cell contacts (Passacantilli et al. 2014) or to cause no morphologic changes (Vandamme et al. 2016). We, however, found an amorphous, flattened cellular shape with lysosome-like bubbles and clumping conglomerates as a morphologic feature of everolimus-resistant BON1 cells. The differences in stability of resistance and morphologic appearance between the studies may be explained by experimental design differences: Passacantilli et al. treated panNETs cells permanently only for 8 weeks with 10 nM everolimus and did not confirm stable resistance (Passacantilli et al. 2014); Vandamme et al. followed a pulsed-selection experimental design (McDermott et al. 2014) by permanently augmenting everolimus doses up to 1 µM, which might be less clinically relevant due to the application of high drug doses out of the clinically relevant range (around 8–59 nM) (Budde et al. 2016, Vandamme et al. 2016).

Since BON1 RR1 and BON1 RR2 have both been developed on the same genetic background and BON1 cells have a higher proliferative index than the typical well differentiated neuroendocrine tumor, we aimed to confirm our model in the novel slow growing human pancreatic NET cell line NT-3 (Benten et al. 2018). Unfortunately, due to the slow growing nature of NT-3 cells, these experiments are still ongoing now for more than 12 months. The initial results showed a gradual increase in proliferation over time in the presence of doses up to four times of the IC50 value (4 nM) of the parental cell line, and we observed similar changes in morphology as in resistant BON1 cells with a switch to a flattened and more pitted appearance under continuous everolimus exposure (unpublished results).

Heterogeneity of BON1 RR1 and BON1 RR2

Tumor heterogeneity has been suggested as a possible mechanism of everolimus resistance (Vandamme et al. 2016) and could be the driving force of sub-clone selection of the most resistant phenotype in response to treatment (Marusyk et al. 2014, Vandamme et al. 2016). Our data are compatible with this theory since both resistant cell lines established in parallel do not show a completely identical phenotype with slightly different effects of everolimus on p70S6K phosphorylation in both cell lines. The novelty of our work lies in its ability to elicit common features of both independently established everolimus-resistant cell lines in order to reduce the bias due to tumor heterogeneity.

Everolimus resistance: G1 phase cell cycle arrest, reduced baseline autophagy and increased baseline apoptosis

Consistent with an increase in G1 phase cell cycle arrest and sub-G1 population in the resistant cell lines in response to everolimus, we observed a baseline decrease in CDK1(cdc2) expression. CDK1(cdc2) is a key regulator of the mitotic transition (Nurse 1990) and its inactivation causes exit from mitosis and stabilizes G1 phase (Potapova et al. 2009). The G1 phase cell cycle arrest/sub-G1 increase in the resistant cell lines might be associated with the observed decrease in baseline autophagy. Inhibition of autophagy has shown to be linked to a G1 phase cell cycle arrest, sub-G1 increase and apoptosis induction in different cancer entities (Choi et al. 2013, Liu et al. 2014, Almasi et al. 2018). Nevertheless, the decrease in autophagy seems unexpected in the context of resistance since mTORC1 inhibitor-induced autophagy has previously been described as a potential mechanism of promoting tumor cell resistance: inhibition of autophagy increased everolimus efficacy in NET cells (Avniel-Polak et al. 2016, Avniel et al. 2018). The decrease in the autophagy marker LC3AII observed in the resistant cell lines might be a mechanism of adaption to the permanent induction of autophagy in the presence of the mTORC1 inhibitor everolimus. This might indicate a sensitive balance of the dual effects of autophagy as a protective mechanism for the cancer cell on the one hand and as a mediator of cell death on the other (Sui et al. 2013).

Since apoptosis negatively correlates with autophagy (Liao et al. 2012, Sui et al. 2013, Avniel-Polak et al. 2016, Mejias-Pena et al. 2017), the decrease in autophagy in the resistant cell lines might cause the observed increase in apoptosis. All these effects might be due to GSK3 over-activation. GSK3 over-activation, which is discussed below in detail, has previously been reported to be directly associated with G1 cell cycle arrest, decreased autophagy and increased apoptosis (Li et al. 2011, Giancotti 2014, Maurer et al. 2014, Weikel et al. 2016, Zhang et al. 2016, Aristizabal Prada et al. 2017) and to be a crucial feature of cancer cell resistance to chemotherapy, radiotherapy and targeted therapies (Domoto et al. 2016, McCubrey et al. 2016).

Everolimus resistance increased migration potential and c-Met activation

BON1 RR1 and BON1 RR2 cells displayed significantly increased migration potential in response to everolimus treatment and baseline c-Met over-activation. Temsirolimus-(mTORC1 inhibitor)-resistant renal cell carcinoma cells also showed enhanced migration potential (Juengel et al. 2014). Mis-regulation, overexpression or gene amplification of the RTK c-Met have been identified to play a crucial role in tumor invasion, motility and metastasis formation associated with poor prognosis (Di Renzo et al. 1992, Camp et al. 1999, Chen et al. 1999, Takeo et al. 2001, Ma et al. 2003, Song et al. 2010). However, in previous studies, we demonstrated that specific c-Met inhibition by either knock-down experiments with siRNA or the specific c-Met inhibitor INC280 could neither decrease cellular viability nor cellular migration in NET cells in vitro (Reuther et al. 2016). C-Met activation might be rather indirectly associated with cellular migration potential in NET cells in vitro through activation of alternative pathways, such as the PI3K-Akt-mTORC1 pathway (Engelman et al. 2007).

Everolimus resistance: baseline GSK3 over-activation and decrease in IRS-1 protein levels

A feedback activation of RTKs such as the insulin receptor due to mTORC1 inhibition has previously been reported (Chandarlapaty et al. 2011). The feedback activation of PI3K-Akt signaling by the mTORC1 inhibitor everolimus via disinhibition/activation of IRS-1 has been discussed as a possible cause of resistance in NETs (Fig. 1, red arrows) (O’Reilly et al. 2006, Zitzmann et al. 2010, Vandamme et al. 2016). We have indeed found that pro-survival signaling of the PI3K-Akt pathway through Akt phosphorylation/activation upon everolimus re-application persisted in BON1 RR1 and BON1 RR2, as described previously in long-term everolimus-treated panNET cells (Passacantilli et al. 2014). Therefore, we speculate that there was permanent PI3K/Akt activation in the presence of persisting everolimus treatment during the development of resistance. Constitutively active PI3K has previously been reported to induce a reduction in IRS-1/2 protein levels (Pirola et al. 2003, Wada 2009). Therefore, IRS-1 downregulation observed in the resistant cell lines might be a compensatory escape mechanism in response to chronic PI3K/Akt over-activation, similar to the mechanism of development of insulin resistance (Pirola et al. 2003). Different IRS-1-induced signaling pathways have previously been reported to inhibit GSK3 (Cohen & Frame 2001, Wada 2009, Maurer et al. 2014, Vandamme et al. 2016). Therefore, lower baseline IRS-1 protein levels in the resistant cells may lead to lower GSK3 inhibition and thus stronger GSK3 activation. Since baseline Akt activity or expression was not changed in the resistant cells, compared to the sensitive controls, it is unlikely that GSK3 is over-activated through modulation of the Akt pathway, but more likely through an alternative pathway. These mechanisms need to be investigated in future studies in NETs.

Various studies have demonstrated an association between IRS-1/GSK3α/β deregulation and tumorigenesis (Wada 2009, McCubrey et al. 2016). GSK3 over-activation may promote the development of everolimus resistance, as recently reviewed for GSK3 in the context of resistance to chemotherapy, radiotherapy and targeted therapies (Domoto et al. 2016, McCubrey et al. 2016). Similar effects were observed in transfected MCF-7 breast cancer cells, where a constitutively active GSK3 construct showed resistance to the mTORC1 inhibitor rapamycin (McCubrey et al. 2014).

In order to prove the importance of constitutively-active GSK3 in everolimus-resistant panNET cells, we treated BON1 RR1/RR2 cells with the selective GSK3 inhibitor AR-A014418 and found a striking decrease in cellular viability in BON1 RR1/RR2, and to a lesser extent in sensitive control cells, associated with an increase in the autophagy marker LC3AII. However, unexpectedly, combination treatment of AR-A014418 plus everolimus did not enhance the inhibitory effects of single AR-A014418 treatment. The absence of synergism in the AR-A014418/everolimus combination treatment may be explicable by GSK3α inhibition being similar in response to the combination treatment compared to single treatment with AR-A014418 in the resistant cell lines.

Dual-targeting approach with the PI3Kα inhibitor BYL719 overcame acquired everolimus resistance

In a previous study on different NET cell lines, the selective PI3Kα inhibitor BYL719 showed synergistic anti-proliferative effects in combination with everolimus through strong GSK3 inhibition (Nölting et al. 2017). In contrast to AR-A014418/everolimus combination treatment, the combination of everolimus with BYL719 was synergistic at low therapeutically-relevant doses in both resistant cell lines and restored everolimus sensitivity through more potent GSK3 inhibition, compared to each drug separately; this was associated with IRS-1 up-regulation, increased autophagy associated with decreased apoptosis, and an inhibition of EGFR, 4EBP1 and p70S6K. This is consistent with our previous study showing that selective GSK3 inhibition led to inhibition of EGFR, p70S6K and 4EBP1 and enhanced the autophagy marker LC3AII in BON1 cells (Aristizabal Prada et al. 2017). It has previously been reported that inhibition of GSK3 restored autophagy (Weikel et al. 2016, Zhang et al. 2016, Aristizabal Prada et al. 2017) and re-sensitized different tumor entities to chemotherapy, radiotherapy or targeted therapy (Domoto et al. 2016).

Similar combination approaches using PI3K inhibitors (BEZ235 or BKM120), although without taking GSK3 signaling into account, have previously been shown to overcome reversible everolimus resistance (Passacantilli et al. 2014, Vandamme et al. 2016). BYL719 is currently being investigated in a late phase 3 clinical study (SOLAR-I) for breast cancer (Nbib2437318) and will probably enter clinical use. Moreover, a phase 1b clinical study is currently evaluating the safety of BYL719 plus everolimus in advanced breast cancer, renal cell cancer and pancreatic tumors including panNETs (Nbib2077933). Therefore, we provide novel data on the potentially clinically relevant PI3Kα inhibitor BYL719. GSK3 inhibition by BYL719 might be considered as second-line treatment in everolimus-resistant panNETs if indeed permanent GSK3 over-activation, potentially in combination with decreased IRS-1 protein levels, can be demonstrated as a central mechanism of resistance in vivo.

Conclusions and perspectives

In this study, we established two stable everolimus-resistant panNET (BON1) cell lines. The mechanisms of resistance may include G1 cell cycle arrest, decreased CDK1/cdc2 expression, reduced autophagy, increased apoptosis and migration potential potentially due to permanent GSK3 over-activation compatible with downregulation of IRS-1 protein levels. The dual-targeting approach with the PI3Kα inhibitor BYL719 restored treatment sensitivity through GSK3 inhibition, IRS-1 upregulation and increased autophagy in the resistant cells. In a pilot study, we have already transferred our resistant cell line model to an orthotopic intrapancreatic tumor xenograft mouse model, and this may represent a valid basis to validate our in vitro findings of this first stable everolimus-resistant cell line model and translate it into clinical practice.

Supplementary data

This is linked to the online version of the paper at https://doi.org/10.1530/ERC-18-0159.

Declaration of interest

The authors declare that there is no conflict of interest that would prejudice the impartiality of the research reported. C J A and A B G have received research contracts (Ipsen, Novartis), lecture honoraria (Ipsen, Novartis) and advisory board honoraria (Novartis). J S has received research contracts (Novartis) and lecture honoraria (Novartis, Ipsen). S N has received a research contract (Novartis) and lecture honoraria (Ipsen).

Funding

E T Aristizabal Prada has received a scholarship from FAZIT-Stiftung (http://www.fazit-stiftung.de). This work was supported by a Young Investigator Research Grant (YING) to S N, project number RAFI 029, from Novartis Pharma GmbH, Nürnberg, Germany. This research was supported, in part, by the Intramural Research Program of the NIH, NICHD and by the Deutsche Forschungsgemeinschaft (DFG) (within the CRC/Transregio 205/1 'The Adrenal: Central Relay in Health and Disease').

Acknowledgements

The authors thank Jakob Albrecht (Department of Nuclear Medicine, Charité Berlin) for performing the implantation of everolimus-resistant BON1 cells into SCID mice and Berlin Experimental Radionuclide Imaging Center (BERIC), Charité – Universitätsmedizin Berlin, Germany for in vivo experiments.

References

  • AlmasiSKennedyBEEl-AghilMStereaAMGujarSPartida-SanchezSEl HianiY 2018 TRPM2 channel-mediated regulation of autophagy maintains mitochondrial function and promotes gastric cancer cell survival via the JNK-signaling pathway. Journal of Biological Chemistry 293 36373650. (https://doi.org/10.1074/jbc.M117.817635)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Aristizabal PradaETAuernhammerCJ 2018 Targeted therapy of gastroenteropancreatic neuroendocrine tumours: preclinical strategies and future targets. Endocrine Connections 7 R1R25. (https://doi.org/10.1530/EC-17-0286)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Aristizabal PradaETWeisCOrthMLausekerMSpoettlGMaurerJGrabowskiPGrossmanAAuernhammerCJNöltingS 2017 GSK3alpha/beta: a novel therapeutic target for neuroendocrine tumors? Neuroendocrinology 106 335351. (https://doi.org/10.1159/000481887)

    • PubMed
    • Search Google Scholar
    • Export Citation
  • AuernhammerCJSpitzwegCAngeleMKBoeckSGrossmanANöltingSIlhanHKnoselTMayerleJReinckeM 2017 Advanced neuroendocrine tumours of the small intestine and pancreas: clinical developments, controversies, and future strategies. Lancet Diabetes and Endocrinology 6 404415. (https://doi.org/10.1016/S2213-8587(17)30401-1)

    • Search Google Scholar
    • Export Citation
  • AvnielSLeibowitzGDovinerVGrossDJGrozinsky-GlasbergS 2018 Combining chloroquine with RAD001 inhibits tumor growth in a NEN mouse model. Endocrine-Related Cancer 25 677686. (https://doi.org/10.1530/ERC-18-0121)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Avniel-PolakSLeibowitzGRiahiYGlaserBGrossDJGrozinsky-GlasbergS 2016 Abrogation of autophagy by chloroquine alone or in combination with mTOR inhibitors induces apoptosis in neuroendocrine tumor cells. Neuroendocrinology 103 724737. (https://doi.org/10.1159/000442589)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • BentenDBehrangYUnrauLWeissmannVWolters-EisfeldGBurdak-RothkammSStahlFRAnlaufMGrabowskiPMobsM 2018 Establishment of the first well-differentiated human pancreatic neuroendocrine tumor model. Molecular Cancer Research 16 496507. (https://doi.org/10.1158/1541-7786.MCR-17-0163)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • BriestFGrabowskiP 2014 PI3K-AKT-mTOR-signaling and beyond: the complex network in gastroenteropancreatic neuroendocrine neoplasms. Theranostics 4 336365. (https://doi.org/10.7150/thno.7851)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • BuddeKZonnenbergBAFrostMCheungWUrvaSBrechenmacherTSteinKChenDKingswoodJCBisslerJJ 2016 Pharmacokinetics and pharmacodynamics of everolimus in patients with renal angiomyolipoma and tuberous sclerosis complex or lymphangioleiomyomatosis. British Journal of Clinical Pharmacology 81 958970. (https://doi.org/10.1111/bcp.12834)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • CampRLRimmEBRimmDL 1999 Met expression is associated with poor outcome in patients with axillary lymph node negative breast carcinoma. Cancer 86 22592265. (https://doi.org/10.1002/(SICI)1097-0142(19991201)86:11<2259::AID-CNCR13>3.0.CO;2-2)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • CapozziMCaterinaIDe DivitiisCvon ArxCMaiolinoPTatangeloFCavalcantiEDi GirolamoEIaffaioliRVScalaS 2015 Everolimus and pancreatic neuroendocrine tumors (PNETs): activity, resistance and how to overcome it. International Journal of Surgery 21 (Supplement 1) S89S94. (https://doi.org/10.1016/j.ijsu.2015.06.064)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • CapursoGFendrichVRinzivilloMPanzutoFBartschDKDelle FaveG 2012 Novel molecular targets for the treatment of gastroenteropancreatic endocrine tumors: answers and unsolved problems. International Journal of Molecular Sciences 14 3045. (https://doi.org/10.3390/ijms14010030)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ChanDLSegelovESinghS 2017 Everolimus in the management of metastatic neuroendocrine tumours. Therapeutic Advances in Gastroenterology 10 132141. (https://doi.org/10.1177/1756283X16674660)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ChandarlapatySSawaiAScaltritiMRodrik-OutmezguineVGrbovic-HuezoOSerraVMajumderPKBaselgaJRosenN 2011 AKT inhibition relieves feedback suppression of receptor tyrosine kinase expression and activity. Cancer Cell 19 5871. (https://doi.org/10.1016/j.ccr.2010.10.031)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ChenBKOhtsukiYFurihataMTakeuchiTIwataJLiangSBSonobeH 1999 Overexpression of c-Met protein in human thyroid tumors correlated with lymph node metastasis and clinicopathologic stage. Pathology: Research and Practice 195 427433. (https://doi.org/10.1016/S0344-0338(99)80017-X)

    • Search Google Scholar
    • Export Citation
  • ChoiHSJeongEHLeeTGKimSYKimHRKimCH 2013 Autophagy inhibition with monensin enhances cell cycle arrest and apoptosis induced by mTOR or epidermal growth factor receptor inhibitors in lung cancer cells. Tuberculosis and Respiratory Diseases 75 917. (https://doi.org/10.4046/trd.2013.75.1.9)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • CohenPFrameS 2001 The renaissance of GSK3. Nature Reviews Molecular Cell Biology 2 769776. (https://doi.org/10.1038/35096075)

  • De BuckSSJakabABoehmMBootleDJuricDQuadtCGogginTK 2014 Population pharmacokinetics and pharmacodynamics of BYL719, a phosphoinositide 3-kinase antagonist, in adult patients with advanced solid malignancies. British Journal of Clinical Pharmacology 78 543555. (https://doi.org/10.1111/bcp.12378)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • Di RenzoMFOliveroMFerroSPratMBongarzoneIPilottiSBelfioreACostantinoAVigneriRPierottiMA 1992 Overexpression of the c-MET/HGF receptor gene in human thyroid carcinomas. Oncogene 7 25492553.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • DomotoTPykoIVFurutaTMiyashitaKUeharaMShimasakiTNakadaMMinamotoT 2016 Glycogen synthase kinase-3beta is a pivotal mediator of cancer invasion and resistance to therapy. Cancer Science 107 13631372. (https://doi.org/10.1111/cas.13028)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • EkebladSSkogseidBDunderKObergKErikssonB 2008 Prognostic factors and survival in 324 patients with pancreatic endocrine tumor treated at a single institution. Clinical Cancer Research 14 77987803. (https://doi.org/10.1158/1078-0432.CCR-08-0734)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • EngelmanJAZejnullahuKMitsudomiTSongYHylandCParkJOLindemanNGaleCMZhaoXChristensenJ 2007 MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling. Science 316 10391043. (https://doi.org/10.1126/science.1141478)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • EversBMTownsendCMJrUppJRAllenEHurlbutSCKimSWRajaramanSSinghPReubiJCThompsonJC 1991 Establishment and characterization of a human carcinoid in nude mice and effect of various agents on tumor growth. Gastroenterology 101 303311. (https://doi.org/10.1016/0016-5085(91)90004-5)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • FrillingAModlinIMKiddMRussellCBreitensteinSSalemRKwekkeboomDLauWYKlersyCVilgrainV 2014 Recommendations for management of patients with neuroendocrine liver metastases. Lancet Oncology 15 e8e21. (https://doi.org/10.1016/S1470-2045(13)70362-0)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • GiancottiFG 2014 Deregulation of cell signaling in cancer. FEBS Letters 588 25582570. (https://doi.org/10.1016/j.febslet.2014.02.005)

  • GreenblattDYNdiayeMChenHKunnimalaiyaanM 2010 Lithium inhibits carcinoid cell growth in vitro. American Journal of Translational Research 2 248253.

  • Grozinsky-GlasbergSShimonIRubinfeldH 2012 The role of cell lines in the study of neuroendocrine tumors. Neuroendocrinology 96 173187. (https://doi.org/10.1159/000338793)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • HalfdanarsonTRRabeKGRubinJPetersenGM 2008 Pancreatic neuroendocrine tumors (PNETs): incidence, prognosis and recent trend toward improved survival. Annals of Oncology 19 17271733. (https://doi.org/10.1093/annonc/mdn351)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • HeKChenDRuanHLiXTongJXuXZhangLYuJ 2016 BRAFV600E-dependent Mcl-1 stabilization leads to everolimus resistance in colon cancer cells. Oncotarget 7 4769947710. (https://doi.org/10.18632/oncotarget.10277)

    • PubMed
    • Search Google Scholar
    • Export Citation
  • JuengelEMakarevicJReiterMManiJTsaurIBartschGHaferkampABlahetaRA 2014 Resistance to the mTOR inhibitor temsirolimus alters adhesion and migration behavior of renal cell carcinoma cells through an integrin alpha5- and integrin beta3-dependent mechanism. Neoplasia 16 291300. (https://doi.org/10.1016/j.neo.2014.03.011)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • LiYKongDBaoBAhmadASarkarFH 2011 Induction of cancer cell death by isoflavone: the role of multiple signaling pathways. Nutrients 3 877896. (https://doi.org/10.3390/nu3100877)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • LiaoXSluimerJCWangYSubramanianMBrownKPattisonJSRobbinsJMartinezJTabasI 2012 Macrophage autophagy plays a protective role in advanced atherosclerosis. Cell Metabolism 15 545553. (https://doi.org/10.1016/j.cmet.2012.01.022)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • LiuQShiXZhouXWangDWangLLiC 2014 Effect of autophagy inhibition on cell viability and cell cycle progression in MDAMB231 human breast cancer cells. Molecular Medicine Reports 10 625630. (https://doi.org/10.3892/mmr.2014.2296)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • LubnerSJKunnimalaiyaanMHolenKDNingLNdiayeMLoconteNKMulkerinDLSchelmanWRChenH 2011 A preclinical and clinical study of lithium in low-grade neuroendocrine tumors. Oncologist 16 452457. (https://doi.org/10.1634/theoncologist.2010-0323)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • MaPCKijimaTMaulikGFoxEASattlerMGriffinJDJohnsonBESalgiaR 2003 c-MET mutational analysis in small cell lung cancer: novel juxtamembrane domain mutations regulating cytoskeletal functions. Cancer Research 63 62726281.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • MartelliAMBuontempoFEvangelistiC 2014 GSK-3beta: a key regulator of breast cancer drug resistance. Cell Cycle 13 697698. (https://doi.org/10.4161/cc.28091)

  • MarusykATabassumDPAltrockPMAlmendroVMichorFPolyakK 2014 Non-cell-autonomous driving of tumour growth supports sub-clonal heterogeneity. Nature 514 5458. (https://doi.org/10.1038/nature13556)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • MaurerUPreissFBrauns-SchubertPSchlicherLCharvetC 2014 GSK-3 - at the crossroads of cell death and survival. Journal of Cell Science 127 13691378. (https://doi.org/10.1242/jcs.138057)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • McCubreyJASteelmanLSBertrandFEDavisNMSokoloskyMAbramsSLMontaltoGD’AssoroABLibraMNicolettiF 2014 GSK-3 as potential target for therapeutic intervention in cancer. Oncotarget 5 28812911. (https://doi.org/10.18632/oncotarget.2037)

    • PubMed
    • Search Google Scholar
    • Export Citation
  • McCubreyJARakusDGizakASteelmanLSAbramsSLLertpiriyapongKFitzgeraldTLYangLVMontaltoGCervelloM 2016 Effects of mutations in Wnt/beta-catenin, hedgehog, Notch and PI3K pathways on GSK-3 activity-Diverse effects on cell growth, metabolism and cancer. Biochimica et Biophysica Acta 1863 29422976. (https://doi.org/10.1016/j.bbamcr.2016.09.004)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • McDermottMEustaceAJBusschotsSBreenLCrownJClynesMO’DonovanNStordalB 2014 In vitro development of chemotherapy and targeted therapy drug-resistant cancer cell lines: a practical guide with case studies. Frontiers in Oncology 4 40. (https://doi.org/10.3389/fonc.2014.00040)

    • Search Google Scholar
    • Export Citation
  • Mejias-PenaYEstebanezBRodriguez-MiguelezPFernandez-GonzaloRAlmarMde PazJAGonzalez-GallegoJCuevasMJ 2017 Impact of resistance training on the autophagy-inflammation-apoptosis crosstalk in elderly subjects. Aging 9 408418. (https://doi.org/10.18632/aging.101167)

    • PubMed
    • Search Google Scholar
    • Export Citation
  • MussmannRGeeseMHarderFKegelSAndagULomowABurkUOnichtchoukDDohrmannCAustenM 2007 Inhibition of GSK3 promotes replication and survival of pancreatic beta cells. Journal of Biological Chemistry 282 1203012037. (https://doi.org/10.1074/jbc.M609637200)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NöltingSRentschJFreitagHDetjenKBriestFMobsMWeissmannVSiegmundBAuernhammerCJAristizabal PradaET 2017 The selective PI3Kalpha inhibitor BYL719 as a novel therapeutic option for neuroendocrine tumors: Results from multiple cell line models. PLoS ONE 12 e0182852. (https://doi.org/10.1371/journal.pone.0182852)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • NurseP 1990 Universal control mechanism regulating onset of M-phase. Nature 344 503508. (https://doi.org/10.1038/344503a0)

  • O’ReillyKERojoFSheQBSolitDMillsGBSmithDLaneHHofmannFHicklinDJLudwigDL 2006 mTOR inhibition induces upstream receptor tyrosine kinase signaling and activates Akt. Cancer Research 66 15001508. (https://doi.org/10.1158/0008-5472.CAN-05-2925)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • PassacantilliICapursoGArchibugiLCalabrettaSCaldarolaSLoreniFDelle FaveGSetteC 2014 Combined therapy with RAD001 e BEZ235 overcomes resistance of PET immortalized cell lines to mTOR inhibition. Oncotarget 5 53815391. (https://doi.org/10.18632/oncotarget.2111)

    • Search Google Scholar
    • Export Citation
  • PavelMESersC 2016 WOMEN IN CANCER THEMATIC REVIEW: Systemic therapies in neuroendocrine tumors and novel approaches toward personalized medicine. Endocrine-Related Cancer 23 T135T154. (https://doi.org/10.1530/ERC-16-0370)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • PavelMEHainsworthJDBaudinEPeetersMHorschDWinklerREKlimovskyJLebwohlDJehlVWolinEM 2011 Everolimus plus octreotide long-acting repeatable for the treatment of advanced neuroendocrine tumours associated with carcinoid syndrome (RADIANT-2): a randomised, placebo-controlled, phase 3 study. Lancet 378 20052012. (https://doi.org/10.1016/S0140-6736(11)61742-X)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • PirolaLBonnafousSJohnstonAMChaussadeCPortisFVan ObberghenE 2003 Phosphoinositide 3-kinase-mediated reduction of insulin receptor substrate-1/2 protein expression via different mechanisms contributes to the insulin-induced desensitization of its signaling pathways in L6 muscle cells. Journal of Biological Chemistry 278 1564115651. (https://doi.org/10.1074/jbc.M208984200)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • PortaCPaglinoCMoscaA 2014 Targeting PI3K/Akt/mTOR signaling in cancer. Frontiers in Oncology 4 64. (https://doi.org/10.3389/fonc.2014.00064)

  • PotapovaTADaumJRByrdKSGorbskyGJ 2009 Fine tuning the cell cycle: activation of the Cdk1 inhibitory phosphorylation pathway during mitotic exit. Molecular Biology of the Cell 20 17371748. (https://doi.org/10.1091/mbc.e08-07-0771)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ReutherCHeinzleVSpampattiMVlotidesGde ToniESpottlGMaurerJNoltingSGokeBAuernhammerCJ 2016 Cabozantinib and tivantinib, but not INC280, induce antiproliferative and antimigratory effects in human neuroendocrine tumor cells in vitro: evidence for ‘off-target’ effects not mediated by c-Met inhibition. Neuroendocrinology 103 383401. (https://doi.org/10.1159/000439431)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • RiccardiCNicolettiI 2006 Analysis of apoptosis by propidium iodide staining and flow cytometry. Nature Protocols 1 14581461. (https://doi.org/10.1038/nprot.2006.238)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ShimuraT 2011 Acquired radioresistance of cancer and the AKT/GSK3beta/cyclin D1 overexpression cycle. Journal of Radiation Research 52 539544. (https://doi.org/10.1269/jrr.11098)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ShipkovaMHesselinkDAHoltDWBillaudEMvan GelderTKunickiPKBrunetMBuddeKBartenMJDe SimoneP 2016 Therapeutic drug monitoring of everolimus: a consensus report. Therapeutic Drug Monitoring 38 143169. (https://doi.org/10.1097/FTD.0000000000000260)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • SoaresKCFoleyKOlinoKLeubnerAMayoSCJainAJaffeeESchulickRDYoshimuraKEdilB 2014 A preclinical murine model of hepatic metastases. Journal of Visualized Experiments 51677. (https://doi.org/10.3791/51677)

    • Search Google Scholar
    • Export Citation
  • SongJLiMTretiakovaMSalgiaRCaglePTHusainAN 2010 Expression patterns of PAX5, c-Met, and paxillin in neuroendocrine tumors of the lung. Archives of Pathology and Laboratory Medicine 134 17021705. (https://doi.org/10.1043/2009-0664-OAR1.1)

    • Search Google Scholar
    • Export Citation
  • SuiXChenRWangZHuangZKongNZhangMHanWLouFYangJZhangQ 2013 Autophagy and chemotherapy resistance: a promising therapeutic target for cancer treatment. Cell Death and Disease 4 e838. (https://doi.org/10.1038/cddis.2013.350)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • TakeoSAraiHKusanoNHaradaTFuruyaTKawauchiSOgaAHiranoTYoshidaTOkitaK 2001 Examination of oncogene amplification by genomic DNA microarray in hepatocellular carcinomas: comparison with comparative genomic hybridization analysis. Cancer Genetics and Cytogenetics 130 127132. (https://doi.org/10.1016/S0165-4608(01)00479-4)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • VandammeTBeyensMde BeeckKODoganFvan KoetsveldPMPauwelsPMortierGVangestelCde HerderWVan CampG 2016 Long-term acquired everolimus resistance in pancreatic neuroendocrine tumours can be overcome with novel PI3K-AKT-mTOR inhibitors. British Journal of Cancer 114 650658. (https://doi.org/10.1038/bjc.2016.25)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • WadaA 2009 GSK-3 inhibitors and insulin receptor signaling in health, disease, and therapeutics. Frontiers in Bioscience 14 15581570. (https://doi.org/10.2741/3324)

    • Search Google Scholar
    • Export Citation
  • WagleNGrabinerBCVan AllenEMAmin-MansourATaylor-WeinerARosenbergMGrayNBarlettaJAGuoYSwansonSJ 2014 Response and acquired resistance to everolimus in anaplastic thyroid cancer. New England Journal of Medicine 371 14261433. (https://doi.org/10.1056/NEJMoa1403352)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • WeikelKACacicedoJMRudermanNBIdoY 2016 Knockdown of GSK3beta increases basal autophagy and AMPK signalling in nutrient-laden human aortic endothelial cells. Bioscience Reports 36 e00382. (https://doi.org/10.1042/BSR20160174)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • WolinEM 2013 PI3K/Akt/mTOR pathway inhibitors in the therapy of pancreatic neuroendocrine tumors. Cancer Letters 335 18. (https://doi.org/10.1016/j.canlet.2013.02.016)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • YaoJCHassanMPhanADagohoyCLearyCMaresJEAbdallaEKFlemingJBVautheyJNRashidA 2008 One hundred years after ‘carcinoid’: epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. Journal of Clinical Oncology 26 30633072. (https://doi.org/10.1200/JCO.2007.15.4377)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • YaoJCShahMHItoTBohasCLWolinEMVan CutsemEHobdayTJOkusakaTCapdevilaJde VriesEG 2011 Everolimus for advanced pancreatic neuroendocrine tumors. New England Journal of Medicine 364 514523. (https://doi.org/10.1056/NEJMoa1009290)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • YaoJCPhanATJehlVShahGMeric-BernstamF 2013 Everolimus in advanced pancreatic neuroendocrine tumors: the clinical experience. Cancer Research 73 14491453. (https://doi.org/10.1158/0008-5472.CAN-12-3923)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • YaoJWangJYLiuYWangBLiYXZhangRWangLSLiuL 2014 A randomized phase II study of everolimus for advanced pancreatic neuroendocrine tumors in Chinese patients. Medical Oncology 31 251. (https://doi.org/10.1007/s12032-014-0251-x)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • YaoJCFazioNSinghSBuzzoniRCarnaghiCWolinETomasekJRadererMLahnerHVoiM 2016a Everolimus for the treatment of advanced, non-functional neuroendocrine tumours of the lung or gastrointestinal tract (RADIANT-4): a randomised, placebo-controlled, phase 3 study. Lancet 387 968977. (https://doi.org/10.1016/S0140-6736(15)00817-X)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • YaoJCPavelMLombard-BohasCVan CutsemEVoiMBrandtUHeWChenDCapdevilaJde VriesEG 2016b Everolimus for the treatment of advanced pancreatic neuroendocrine tumors: overall survival and circulating biomarkers from the randomized, phase III RADIANT-3 study. Journal of Clinical Oncology 34 39063913. (https://doi.org/10.1200/JCO.2016.68.0702)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • ZatelliMCFanciulliGMalandrinoPRamundoVFaggianoAColaoAGroupN 2016 Predictive factors of response to mTOR inhibitors in neuroendocrine tumours. Endocrine-Related Cancer 23 R173R183. (https://doi.org/10.1530/ERC-15-0413)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • ZaytsevaYYValentinoJDGulhatiPEversBM 2012 mTOR inhibitors in cancer therapy. Cancer Letters 319 17. (https://doi.org/10.1016/j.canlet.2012.01.005)

  • ZhangCHouBYuSChenQZhangNLiH 2016 HGF alleviates high glucose-induced injury in podocytes by GSK3beta inhibition and autophagy restoration. Biochimica et Biophysica Acta 1863 26902699. (https://doi.org/10.1016/j.bbamcr.2016.08.004)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • ZitzmannKRudenJBrandSGokeBLichtlJSpottlGAuernhammerCJ 2010 Compensatory activation of Akt in response to mTOR and Raf inhibitors – a rationale for dual-targeted therapy approaches in neuroendocrine tumor disease. Cancer Letters 295 100109. (https://doi.org/10.1016/j.canlet.2010.02.018)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Downloadable materials

 

      Society for Endocrinology

Related Articles

Article Information

Metrics

All Time Past Year Past 30 Days
Abstract Views 498 498 5
Full Text Views 1424 321 21
PDF Downloads 103 47 9

Altmetrics

Figures

  • View in gallery

    Proposed mechanism of action of the mTORC1-inhibitor everolimus. Everolimus inhibits the mTORC1 complex and thereby down-regulates the mTORC1/p70S6K signaling cascade, preventing the cell from cell cycle progression, survival and proliferation. Inhibition of mTORC1/p70S6K leads to disinhibition/activation of IRS-1 and thereby activation of PI3K/Akt, which in turn inhibits GSK3 by phosphorylation. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

  • View in gallery

    Effects of everolimus (1 nM–2 µM) on cell survival of resistant (BON1 RR1 and BON1 RR2) and sensitive (BON1 Control DMSO) cell lines after 144 h of incubation with everolimus. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to BON1 Control DMSO are shown, considering P < 0.05 as significant (*). A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

  • View in gallery

    Everolimus resistant cells show morphologic changes. (A) Everolimus-resistant cells (BON1 RR1 and BON1 RR2) displayed an amorphous, flattened cellular shape with lysosome-like bubbles and clumping conglomerates. (B) The viable cell number of the resistant cell lines was slightly but significantly increased compared to the sensitive control cell line after 72 h without treatment (P < 0.001), suggesting more rapid proliferation.

  • View in gallery

    (A) Cell cycle analysis via FACS showed a significant increase in sub-G1 population and G1 phase cell cycle arrest in both resistant cell lines (BON1 RR1 and BON1 RR2) after 72 h everolimus treatment. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to BON1 Control DMSO are shown, considering *P < 0.05; **P < 0.01; ***P < 0.001. (B) LC3A consists of two bands: LC3AI (upper band) and LC3AII (lower band). Baseline LC3AII expression level measured by Western blot analysis was decreased in BON1 RR1 and BON1 RR2, compared to BON1 and BON1 Control DMSO. A representative blot out of three independently performed experiments is shown.

  • View in gallery

    Western blot analysis via densitometry of CDK1(cdc2) showed a baseline decrease in CDK1(cdc2) expression levels in BON1 RR1 and BON1 RR2 cell lines, compared to the sensitive control cells. A representative blot out of three independently performed experiments is shown. For densitometry analysis the arithmetic means and standard deviation of at least three independent experiments are shown. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

  • View in gallery

    In the resistant cell lines, Western blot analysis showed a baseline decrease in IRS-1 protein levels, an increase in cleaved Caspase 3 and a decrease in anti-apoptotic Bcl-2 as a sign of increased apoptosis, as well as a decrease in LC3AII as a sign of decreased autophagy. LC3A consists of two bands: LC3AI (upper band) and LC3AII (lower band). Caspase 3/7 assay showed significantly increased apoptosis in both resistant cell lines, compared to the sensitive control cell line (P < 0.01).

  • View in gallery

    Everolimus-resistant cells showed a significantly increased migration potential with a significantly smaller migration gap after 48 h of everolimus treatment. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to BON1 Control DMSO are shown, considering *P < 0.05. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

  • View in gallery

    Western blot analysis via densitometry of phosphorylated/activated c-Met and total c-Met in resistant (BON1 RR1 and BON1 RR2) and sensitive control cells showed an increase in phosphorylated/activated c-Met in the resistant cell lines. A representative blot out of three independently performed experiments is shown. For densitometry analysis the arithmetic means and standard deviation of at least three independent experiments are shown. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

  • View in gallery

    GSK3 over-activation in the resistant cell lines. (A) Both everolimus-resistant cell lines showed significantly enhanced activation of GSK3α/β (GSK3) compared to the sensitive control cell line (BON1 Control DMSO) and a significantly stronger activation of GSK3β compared to the parental BON1 cells, but stable phosphorylation/activation of AKT after 24 h of incubation with everolimus. A representative blot out of three independently performed experiments is shown. Statistically significantly different results (*) in comparison to BON1 Control DMSO are shown. (B) Single substance treatment of AR-A014418 (10–20 µM) significantly decreased cellular survival in all cell lines tested after 144 h of incubation, although much stronger in BON1 RR1 and BON1 RR2, compared to the sensitive control cell line. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significant differences in comparison to either single substance treatment are shown, considering ***P < 0.001. (C) Western blot analysis showed inhibition of GSK3 and an increase in the autophagy marker LC3AII in response to the selective GSK3α/β inhibitor AR-A014418. LC3A consists of two bands: LC3AI (upper band) and LC3AII (lower band). A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

  • View in gallery

    Dual-targeting approach with the PI3Kα inhibitor BYL719 overcame acquired everolimus resistance and restored sensitivity through inhibition of GSK3, EGFR, p70S6K and 4EBP1 signaling. (A) Combination treatment of everolimus plus BYL719 showed a significantly enhanced inhibitory effect on cell viability compared to single substance treatment in all tested cell lines. Cells were incubated with BYL719 (5 µM and 10 µM) and everolimus (10 nM) alone and in combination for 72 h. The arithmetic means and standard deviation of at least three independent experiments are shown. Statistically significantly different results in comparison to either single substance treatment are shown, considering *P < 0.05; **P < 0.01; ***P < 0.001. (B) Protein levels of components from the EGFR-Akt (pEGFR, EGFR, pAkt, Akt), mTORC1-p70S6K (pp70S6, p70S6K, pS6, S6, p4EBP, 4EBP1) and GSK3 (pGSK3, GSK3) signaling pathway are displayed. Inhibition of GSK3, EGFR, p70S6K and 4EBP1 signaling was observed after combination treatment. A representative blot out of three independently performed experiments is shown. A full colour version of this figure is available at https://doi.org/10.1530/ERC-18-0159.

Cited By

PubMed

Google Scholar