Design, synthesis and antitumor activity of novel thiophene-pyrimidine derivatives as EGFR inhibitors overcoming T790M and L858R/T790M mutations

Highlights

• A series of thiophenpyrimidines derivatives (9a-h, 10a-f, 11a-f, 12a-f, 13a-f) were designed and synthesized.

• Most of the target compounds showed excellent antiproliferative activity against one or several cancer cell lines.

• The pharmacological results indicated that the compound 13a could induce apoptosis of human lung cancer A431 cells.

• The results prompted us that compound 13a may be potential selective EGFR^T790M and EGFR^T790M/L858R inhibitors.

Abstract

Five series of novel thiophene-pyrimidine derivatives (9a-h, 10a-f, 11a-f, 12a-f, 13a-f) have been synthesized and tested for their anti-proliferative activity against several cancer cell lines in which EGF is highly expressed. Most of the target compounds showed excellent activity against one or more cancer cell lines. The most promising compound 13a, of which IC₅₀ values on of cell lines A549 and A431 (4.34 ± 0.60 μM and 3.79 ± 0.57 μM) were similar to the lead compound Olmutinib, showed strong activity and selectivity to EGFR^T790M and EGFR^T790M/L858R. Inhibition data of human normal hepatoma cell line LO2 indicated that most target compounds were less toxic to normal cells and had selective inhibitory effects on cancer cells. In addition, the structure-activity relationship was analyzed and the mechanism of apoptosis induced by the 13a was studied. The results showed that compound 13a induced late apoptosis of A431 cancer cells in a dose-dependent manner.

Introduction

Lung cancer has been the most common cancer disease in the world. Non-small cell lung cancer (NSCLC) accounts for about 80% of lung cancer, and the 5-year survival rate is less than 15% [1]. With the development of molecular biology, researchers have found that EGFR (Epidermal Growth Factor Receptor) is highly expressed in 20–80% of NSCLC patients [2]. EGFR belongs to the receptor tyrosine kinase family, which is mainly located in the cytoplasmic membrane. Activated by EGF and other ligands, it leads to the formation of tyrosine kinase activity in the intracellular domain through dimerization. After activation, it conducts auto-phosphorylation, which will further activate the downstream cell signal transduction pathway and complete the transmembrane signal transduction process [3]. The development of EGFR inhibitors to block this process is expected to effectively inhibit the development of NSCLC. The non-conservation of amino acids leads to EGFR mutations, such as the point mutations L858R on exon 21 and exon 19 deletion, which lead to the activation of downstream pathway of EGFR and uncontrolled cell growth, leading to the occurrence of tumor [4].

The first generation of EGFR TKIs (Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitors) were developed to solve the above-mentioned mutation problems, as shown in Fig. 1. Currently, the first generation on the market are mainly Gefitinib (Iressa), Erlotinib (Tarceva) and Icotinib (Conmana), which have significant clinical response rates in NSCLC patients. However, acquired resistance T790M and T790M/L858R occurred 9–14 months after clinical treatment, deactivating the first-generation inhibitors [[5], [6], [7]]. The mutations increase the affinity of ATP for EGFR, thus increasing the competition between ATP and reversible EGFR TKIs and restoring enzyme activity at physiologically achievable EGFR TKIs concentrations [8]. The second generation of EGFR TKIs mainly includes Afatinib (Xovoltib) and Dacomitinib (Vizimpro), as shown in Fig. 1. These drugs contain electrophilic acrylamide that reacts irreversibly with cysteine CYS797 to form covalent complexes, thus, overcoming resistance mediated by EGFR^T790M or EGFR^T790M/L858R mutation. However, acrylamide in most drugs is so reactive that it binds to cysteine in untargeted proteins, creating toxic side effects. The narrow therapeutic window limits its clinical application [9]. In order to solve the problem of targeted toxicity, several third-generation inhibitors have been developed, such as WZ4002 [10], AZD9291 (Tagrisso) [11], Olmutinib (Olita) [12], etc., as shown in Fig. 1. These inhibitors not only have good anti-tumor activity, but also have good selectivity to EGFR^T790M and EGFR^T790M/L858R kinase. In 2015, the US FDA approved AZD9291 to treat patients with EGFR^T790M mutations who have been treated with second-generation drugs. Another T790M selective EGFR inhibitor Olmutinib (HM61713, BI1482694), was developed by Hanmi Pharmaceuticals and Boehringer Ingelheim. Olmutinib received breakthrough treatment designation in the US in December 2015, and was approved for using in South Korea in May 2016 for locally advanced or metastatic EGFR^T790M positive NSCLC patients. But Olmutinib is not available in more countries because of its toxic side effects [13], such as palmoplantar keratoderma and diarrhea. Based on which, this project analyzes and studies the molecular docking modes of Olmutinib and summarizes the structure-activity relationship of similar drugs to optimize its structure, hoping to obtain a batch of compounds that could overcome its toxic side effects.

There are no reports about the co-crystalline of Olmutinib with protein in the RCSB Protein Data Bank, so we chose EGFR^T790M protein (code: 3IKA) and EGFR^WT protein (code: 4ZAU) to dock with Olmutinib (Fig. 2). After docking Olmutinib with EGFR^T790M protein (Fig. 2A) and EGFR^WT protein (Fig. 2B), the following three points were found, which were similar to the other third generation inhibitors [14]: (1) Aminopyrimidine can form two-dentate hydrogen bonds with the amino acid residue MET793 in the hinged region; (2) Compounds bind to proteins in the form of “U". Side chains containing aniline all point to the solvent region, while the end containing Michael receptor extends into the back pocket (Fig. 2C); (3) Olmutinib binds more tightly to EGFR^T790M, and its hydrogen bond is shorter than EGFR^WT. This may be because the T790M mutant methionine has a larger molecular structure than threonine, causing the compound to move towards the solvent region, leading to a shorter hydrogen bond distance and the formation of hydrogen bond between the acrylamide warhead and the nearby amino acid residue CYS797. Fig. 2D shows the whole binding site of Olmutinib in EGFR^T790M protein. Based on the docking results of Olmutinib and most docking results of the third generation EGFR TKIs, we made a spatial orientation map of Olmutinib in EGFR^T790M protein (Fig. 2C), and modified the compounds on this basis.

On the one hand, Pfizer’s Cheng H [15] and Planken S [16] teams have reported the development process of the third generation EGFR inhibitors PF-06459988 and PF-06747775 (shown in Fig. 1), whose structures are very similar to Olmutinib’s. In their reports, the original compound containing phenylpiperazine had poor permeability. Replacing the phenylpiperazine with N-methylpyrazole group could improve ligand efficiency to a certain extent. This point was used for reference in our design strategy. By changing the length of the side chain and the electrophilicity of the group in the deep solvent region, we explored which type of side chain was effective for the activity of the compounds. On the other hand, according to previous studies [[17], [18], [19]], although covalent inhibitors have advantages over their reversible, noncovalent conjugated counterparts, such as improving biochemical efficiency, prolonging the action time, and avoiding the potential of some drug-resistant mechanisms, when covalent inhibitors are highly reactive and/or lack of specificity, covalent modifications in proteins may participate in immunotoxicity and specific hypersensitivity. In previously published literature [[20], [21], [22]], the introduction of small substituents at the end of the acrylamide side chain can change the activity of the reaction between the acrylamide warhead and the amino acid residue CYS797 in the back pocket. Therefore, we introduced side chains of different lengths at the end of acrylamide to investigate the effect of side chain length and whether the side chain contains branched chains on the activity of the compounds. The design strategy is shown in Fig. 3. However, in this study, the activity of compounds containing N-methylpyrazole group was decreased, possibly because the chain length was too short to allow the compounds’ side chain to penetrate into the solvent zone well. Inspired by WZ4002, we introduced a side chain containing methoxide and planned to extend the side chain appropriately on the basis of N-methylpyrazole group, so the fennel amine structure was adopted. The activity was significantly increased as expected. We also explored the effect of methoxy’s position in anisidine and whether there being a nitrogen in the benzene ring on the activity of the compounds. After discovering that the methoxy group in anisidine has the best activity at the 3 positions of the benzene ring, we considered to introduce the electron-withdrawing group, namely the cyano group, at the 3 positions of the benzene ring to explore the changing rules of the activity of the compound. Finally, compounds 9a-h, 10a-f, 11a-f, 12a-f, 13a-f were obtained.