A combinatorial strategy for overcoming primary and acquired resistance of MEK inhibition in colorectal cancer

Introduction

Colorectal cancer (CRC) is a major gastrointestinal malignancy, whose incidence and mortality ranks third in common cancers [1]. The therapeutic methods for CRC include surgery, radiation, traditional chemotherapy, and targeted therapy [2]. Targeted therapy or molecular targeted therapy is one of the primary ways for cancer treatment. Most of the targeted drugs for clinical treatment to CRC have largely relied on EGFR and VEGFR monoclonal antibodies [3,4], including cetuximab, panitumumab and ramucirumab [5–7]. In clinical cancer practice, these targeted agents have achieved certain therapeutic effects. For instance, the overall survival of patients with stage IV colorectal cancer is extended to more than 30 months through targeted therapy [8]. However, inter-tumor and intra-tumor heterogeneity, primary drug-resistance and acquired resistance caused by long-term treatment severely limit the efficacy of targeted therapy [9,10]. The mechanisms of drug resistance are complex, and usually involve mutations in the target gene or activation of the bypass signaling pathway

The Ras-Raf-MEK-ERK (MAPK) and PI3K-AKT are two important signaling pathways that control cell proliferation. Constitutive activation of any of these pathways would result in sustained cell proliferation, leading to tumorigenesis and progression [13–15]. The RAS-MAPK signaling is downstream of EGFR. Clinical research has shown that about 40% of CRC patients have KRAS mutations, and 5%–10% of patients have BRAF mutations, making these patients unresponsive to EGFR monoclonal antibodies [5,6]. This situation has prompted the search of new CRC-specific targeted drugs or more effective combinations to overcome the resistance [16,17].

Trametinib (GSK1120212) is a second generation of MEK inhibitor [18,19] that has been approved by the US Food and Drug Administration for treating metastatic melanoma patients with BRAF V600E/K mutations. This agent can inhibit the proliferation of B-Raf mutant CRC cell line HT-29 and Colo205 with IC50 of 0.48 and 0.52 nM, but there are few reports of Trametinib for CRC treatment [20,21]. Besides, clinical trials have shown that melanoma patients with resistance to trametinib monotherapy will appear within 6–7 months [22]. These data indicate that the resistance to MEK inhibitors may also be a major clinical challenge.

Therefore, to explore the possibility of trametinib in treatment of CRC, in the present study, we screen the sensitivity of CRC cell lines to trametinib, select and set up the primary and acquired resistant CRC cell lines to this agent, and then test the effective combinative drug with trametinib to overcome its resistance in vitro and in vivo. Furthermore, we will also investigate the mechanism of trametinib resistance in vitro.

2. Materials and methods

2.1. Cell culture

The RKO resistant cell line was generated by treating parental cell with increasing concentrations of Trametinib as previously described

[23]. Other human colorectal cancer cell lines were obtained from the Shanghai Institutes for Biological Sciences (Shanghai, China). CW-2 and RKO-R were maintained in a humidified incubator at 37 °C with 5% CO2, and respectively grown in RPMI 1640 (Hyclone, USA) and MEM (Sigma, UK), while SW480 was cultured in L-15 medium (Gibco, USA) at 37 °C with 100% air. The medium was supplemented with 10% FBS and 100 IU ml−1 penicillin/streptomycin. All cell lines were detected for the presence of mycoplasma using Detection Kit (Genechem, China).

2.2. Reagents

All the targeted inhibitors were purchased from Selleck (Houston, USA). The inhibitors were dissolved in DMSO to a final concentration of 20 mmol/L and stored at −20 °C. The stock solution was diluted to working concentrations with culture medium just before each experiment. MTT. Cancer cell lines were seeded in 96-well plates (1–2 × 103 cells per well) and were treated with different concentrations of inhibitors. After 72 h, cell viabilities were measured with the 3-(4,5-dimethylthiazol-2- yl)-2,5-diphenyl- tetrazolium bromide (MTT) assay. Results of IC50 and the synergistic combination were determined according to the method of Chou and Talalay by using the CompuSyn software program [24]. CompuSyn software was used to generate combination index (CI) values, where CI < 1, = 1, and > 1 indicate synergism, additive effect, and antagonism, respectively.

2.3. Colony formation

For the assays of colony formation, cells were seeded into 6-well plates (1 × 103 to 2 × 103 cells per well) and allowed to adhere over 12 h in a regular culture medium. Then cells were cultured in the absence or presence of drugs as indicated in complete medium for 10–15 days, and the medium were replaced every other day. After the culturing, cells were fixed with methanol, stained with 10% Giemsa.

2.4. Xenograft studies

Cells were subcutaneously injected (5 × 106 cells per injection) into BALB/c nude male mice (4–5 weeks old) which were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). When tumors reached an average volume of approximately 100 mm3, mice were randomized into control and three different medicated groups. Trametinib and GSK2126458 were suspended in 1% Cellulose Sodium which was purchased from Chron Chemicals (Chengdu, China). Trametinib was administered through oral gavage daily at 0.3 mg/kg dose for 21 days, and GSK2126458 was orally ad- ministered at 0.3 mg/kg alone or in the indicated combination daily for 21 days. Tumor volume was assessed and calculated using the following formula: Volumes = (Length × Width2)/2. All experiments were per- formed according to the protocols approved by the Animal Care and Use Committee guidelines of Shanxi Medical University.

2.5. Immunohistochemistry

Immunohistochemistry (IHC) was performed on all the tissues of xenograft tumor using biotin-streptavidin HRP detection systems. Briefly, paraffin sections were subjected to deparaffinization and an- tigen retrieval. Then the slides were incubated with Ki67 (1:400), Cleaved caspase-3 (1:200) antibodies (Cell Signaling Technology, USA). Biotinylated secondary antibody (Zhongshan Golden Bridge Biotechnology Co. Ltd., China) was performed according to the manu- facturer’s recommendations.

3. Results

3.1. Screening the cell model for primary trametinib resistance

A panel of human colorectal cancer cell lines with different gene mutation types (Supplementary Table 1) was used to test their sensi- tivity to 30 targeted inhibitors. Relative cell viability and IC50 were measured after treatment of cells with 16 different concentration gra- dients. Results showed that eight out of nine cell lines were sensitive to only one to three targeted drugs (Figs. S1–S9, Supplementary Table 2), which indicates strong resistance characteristic. It is noteworthy that seven out of nine CRC cells are bearing KRAS mutations, but these cell lines showed different responses to inhibitors. For instance, SW480 and SW620 cell line have the same KRAS G12V mutation, but SW620 was sensitive to trametinib. SW480 was tolerant to all targeted inhibitors, including the MEK inhibitor trametinib. Therefore, SW480 and CW- 2 cells were selected as primary trametinib resistance cell model for further investigation of combination therapy. SW480 cells were PI3KCA wild type, and CW-2 cells have P283S mutation.

3.2. Establishment of the cell model of acquired trametinib resistance

An acquired resistance cell model was established, which is resistant to trametinib by long-term treatment of colorectal cancer cell line RKO (designated as RKO-R). MTT assay showed that the parental cell was sensitive to trametinib (IC50 0.01–0.05 μM). However, the RKO-R cell line required higher doses of this inhibitor for partial growth inhibition (IC50 0.5–1.0 μM) (Fig. 1A). We also observed that 0.2 μM trametinib treatment can only inhibit 1/3 colony formation in RKO-R cells in comparison with parental RKO cells (Fig. 1B). Flow cytometry analysis showed that after treatment with 0.1 μM of trametinib, RKO-R cells had a decreased percentage of G0/G1 phase and an increased S phage (P < 0.05) (Fig. 1C). The inverted behavior indicated that trametinib did not retard cell cycle in G0/G1, which showed RKO-R cells were resistant to trametinib's killing.

3.3. Trametinib combination with GSK2126458 inhibits tumor growth in colorectal cancer xenograft models

To verify the synergistic effect of drug combo, we next established SW480 and RKO-R xenografts and treated them with vehicle control, trametinib alone, GSK2126458, or the combination. SW480 and RKO-R xenografts were daily administrated with 0.3 mg/kg trametinib alone or combo with GSK2126458 (0.3 mg/kg) when tumors reached ~100 mm3. Of interest, combined treatment showed a significant an- titumor effect compared with control and single-agent groups in RKO-R xenograft (Fig. 4). We also observed that in SW480 xenograft model, the tumor sizes have less difference between mice receiving trametinib alone and combination groups. This could be attributed to the like- lihood that trametinib exhibited more inhibition effects in mice than in cultivated cells. However, there is still a statistical difference between combination group and trametinib group. Hematoxylin

4. Discussion

We have several findings in this study: First, nine CRC cells showed different responses to 30 targeted inhibitors. One (SW480) of the nine cell lines did not respond to any of the inhibitors. Second, the trame- tinib resistance of SW480, CW-2, and RKO-R cells could be reversed by

combination with PI3K inhibitor GSK2126458. Third, Western blotting analyses indicated that the bypass activation of PI3K-Akt signaling pathway was contributing to the trametinib resistance. This activation was partly due to IGF1R upregulation with trametinib treatment.

Our data showed that none of the 30 targeted inhibitors could block the proliferation of all nine CRC cells. Six of the nine CRC cells bear KRAS mutations at codon 12 and 13, which indicates constitutively activation of MAPK pathway. These cells are supposed to be blocked by trametinib. But only three of them were sensitive to trametinib. KRAS mutation cell line HCT15 and LoVo were resistant to trametinib. However, they were sensitive to GSK2126458, which indicates that the cellular proliferation was driven by PI3K-AKT pathway. As we focused on GSK2126458 plus trametinib to overcome resistance, we did not select these two cell lines for further investigation.

The different responses of CRC cell lines indicate the complex mo- lecular heterogeneity in tumor [27,28]. Activation of tumor compen- satory signaling pathways [29], changes in the tumor microenviron- ment [30], and adaptation of tumors to inhibitors [31] are the main resistance mechanisms to targeted therapy. Recent reports showed that microRNAs are also a factor influencing the acquired drug resistance of cancer cells [32,33]. Resistance to MEK inhibitors has been observed in various human tumors, including lung cancer, colorectal cancer, and ovarian cancer [26,31,34], etc. Considering that resistance to analysis of SW480 cells treated with trametinib (0.05 μM), BMS-754807 (0.05 μM), or their combination for 24 h. Then the expression levels and activation of ERK and AKT were detected by indicated antibodies.

We find that one mechanism of colorectal cancer cells resistance to trametinib is due to the activated PI3K-AKT bypass pathway. This finding can also be verified by the phenomenon that the MAPK signal pathway in RKO-R cells cannot be blocked under long-term MEK in- hibition. This may be due to the bypass activation PI3K-AKT pathway under trametinib treatment. It has been reported that there is certain interaction between the PI3K-AKT and MAPK pathways, which not only have common upstream molecules, but they can all be activated by oncogenic RAS [38]. And some compensatory signals may appear whenone of the two signal pathways is inhibited [39]. It has been shown that MAPK is activated by PI3K via RAS when the downstream target mTOR of AKT is restrained [40]. It is also found that these two pathways are simultaneously activated in many different tumor types including melanoma, prostate cancer, and colorectal cancer, suggesting that they work through complex signaling networks [41,42]. So blocking both pathways at the same time may be an effective approach to reverse the resistance of a single targeted drug. Clinical trials have shown that dual inhibition of these two pathways may have potential advantages than monotherapy, especially in patients with coexisting PI3K genetic al- terations and KRAS/BRAF mutations [43].

To further explain the upstream molecule that activates PI3K-AKT, we have found that MEK inhibition can up-regulate the upstream pro- tein IGF1R of PI3K in SW480 cells. It has been reported that cetuximab resistance is associated with increased expression of insulin-like growth factor 1 receptor (IGF1R) and tyrosine-protein kinase Met (c-MET) in patients with metastatic CRC [44]. To verify IGF1R upregulation is responsible for PI3K-AKT activation, we treated trametinib resistant CRC cells with combined MEK and IGF1R inhibition. As expected, this combo displayed a synergistic effect, but could not completely prevent the activation of p-AKT caused by trametinib. These findings suggest that inhibition of IGF1R sensitizes cells to trametinib, but not as ef- fectively as directly block PI3K-AKT, perhaps in connection with IGF1R regulation of other signaling pathways [45]. Therefore, in subsequent xenograft studies, we still used the combination therapy of trametinib and GSK2126458.

In initial xenograft experiments, mice were treated daily with trametinib at 1mg/kg, a dose used commonly in xenograft studies [19], and GSK2126458 was at the dose of 1mg/kg[46], too. Vehicle and monotherapy were well-tolerated, but mice receiving the combination therapy began to lose weight (> 10%) and lassitude and started to die on the seventh day after administration. Although it has been reported that trametinib was administered to mice even at a dose of 3 mg/kg body weight [25], the mouse allometric equivalent of the maximum tolerated dose (MTD) in humans is about 0.3 mg/kg. Then we ad- ministrated the light weight group of mice with the reduced daily dose of trametinib and GSK2126458 (0.5 mg/kg respectively). Mice with a dose halved lived longer than before; the combination still led to weight loss and death (data not shown). Exploratory dose finding resulted in a tolerable schedule, and we administrated the mice with a daily dose of 0.3 mg/kg trametinib and GSK2126458 respectively, so they showed stable weight and normal behavior. Finally, in the absence of significant side effects, experiments also verified the effect of this combination therapy in vivo. The combination of trametinib and GSK2126458 in- hibited tumor growth in SW480 and RKO-R cell xenografts.

In summary, our study provides a molecular explanation for the primary and acquired resistance of colorectal cancer cells to MEK in- hibitor trametinib. Our findings provide a reasonable rationale for the combination of MEK and PI3K inhibitors in colorectal cancers with existing feedback activation of PI3K. It will be crucial in future studies to explore whether combination therapy with trametinib and GSK2126458 is effective for treatment of colorectal cancer in clinic.