clinical benefits in patients with NSCLC, especially those harboring the EGFR-activating mutations such as delE746_A750 and
L858R [5]. However, a secondary mutation T790M at the “gatekeeper” position of EGFR leads to drug resistance in much cases
after 10–14 months of gefitinib administration [6,7]. In order to overcome the resistance mechanism correlated with T790M
mutation, a mount of irreversible inhibitors such as rociletinib and osimertinib have been developed, in which a acrylamide
functionality was devised to undergo a Michael addition reaction with Cys797 [8–10] .
In our previous work, we reported a series of hydrazone moiety-bearing aminopyrimidines as selective inhibitors of EGFR
T790M mutant, from which we identified a potent compound (1) (Fig. 1) as the precusor [11]. As a continuous study, we initiated
a further optimization program on this promising series in order to discover compounds with more potent activity. Biological
evaluations led to the identification of compound 7f and 7k which showed significant and selective activity in inhibition of
gefitinib-resistant H1975 cancer cells.
2. Results and discussion
2.1. Chemistry
The synthesis of the target compounds (7a–k) is described in Scheme 1. The intermediate 4 was synthesized via a sequential
coupling of 2,4-dichloro-5-(trifluoromethyl)pyrimidine (2) with tert-butyl (3-aminophenyl)carbamate and hydrazine hydrate using
a modified condition as previously reported [11]. The condensation of intermediate 4 with an appropriate aromatic aldehyde or
ketone in an addition-elimination sequence generated intermediates 5a–k [12,13], which underwent a convenient deprotection
reaction to yield amines 6a–k [14]. Subsequently, amidation of 6 with commercial (E)-4-(dimethylamino)but-2-enoic acid using
HATU as a dehydrant furnished the desired compounds.
2.2. Design of new compounds
At beginning, compound 1 was docked in the active site of EGFR T790M (PDB: 3IKA) [15] to investigate the probable
interactions with kinase using Discovery Studio 3.1. The results reflected that it formed anticipated hydrogen bonds with Met793 in
hinge region. However, the functional acrylamide group adopted a reversed conformation, which prevented it from interacting
well with Cys797. Encouraged by these observations, we initiated a program to develop an additional series of compounds with
more suitable conformation to form desirable covalent bond with Cys797. In binding models of both 1 and WZ4002, a hydrophobic
interaction was observed between Cl atom with Met790. We speculated that the replacement of the Cl atom with a more
hydrophobic CF3 group might increase the interaction and more importantly, it could enforce the compound adopting a
significantly changed conformation as compared with compound 1. To rationalize this hypothesis, a molecular docking was
conducted using compound 7b. Figure 2 showed the putative binding model with EGFR T790M, which demonstrated that
apparent hydrogen-bond was formed between pyrimidine with Met793. Moreover, compound 7b adopted a very favorable
conformation close to Cys797. As a consequence, an additional series of hydrazone moiety-bearing aminopyrimidines were
synthesized and evaluated in both enzymatic and cellular levels with the aim to identify compounds with more potent activity.
2.3. Inhibitory activity against EGFR kinase
The newly synthesized compounds were screened in a double mutant EGFR kinase, namely T790M/L858R, as well as in wild
type (WT) EGFR. The enzymatic assay was carried out using a well-established mobility shift assay, and the results were shown
in Table 1. The biological data clearly demonstrated that all the target compounds suppressed EGFR T790M/L858R in
micromolar range (IC50, 0.75–5.39 μmol/L), and was more potent than gefitinib (positive control; IC50, 11.38 μmol/L). All the
target compounds were not active at 10 μmol/L against wild type EGFR, suggesting a favorable kinase selectivity. Compound 7b
exhibited significant activity against EGFR T790M/L858R, with IC50 value of 1.05 μmol/L, but no improvement was observed as
compared with the precusor 1 (IC50, 1.13 μmol/L). Further modification on terminal benzene ring indicated that electron-donating
groups in this region were better tolerated. Replacement of the fluorine atom with a methyl group led to compound 7d (IC50, 0.88
μmol/L), which was slightly more potent than 7b. Moreover, significantly increased potency could be observed when 7c (Ar =
2,4-(F)2-Ph; IC50, 5.39 μM) was compared with 7e (Ar = 2,4-(CH3O)2-Ph; IC50, 0.77 μmol/L). On the other hand, a substituent at
para-position was beneficial for high activity, as shifting the fluorine atom of compound 7b to ortho-position resulted in a 4.8-fold
decrease in activity (7a; IC50, 5.1 μmol/L). This effect could be explained by the putative binding model as described in Figure 2,
a hydrogen-bond was formed between fluorine atom with Lys716. The introduction of a methyl group in the hydrazone moiety (R1)
was not crucial for inhibition of EGFR T790M/L858R, which indicated that the structure-activity relationship (SAR) in this
region was limited. Interestingly, compound 7k containing a pyridine moiety instead of the benzene ring maintained favorable
potency (IC50, 3.32 μmol/L) as well as selectivity, which highlighted the potential for further optimization.
2.4. Inhibitory activity against H1975 and A549 NSCLC cells
Compounds 7a–k were screened in two type of NSCLC cancer cells including A549 (WT EGFR and K-Ras mutation) and
gefitinib-resistant H1975 (EGFR T790M/L858R). Notably, the attractive inhibitory activity against EGFR T790M/L858R
translated well into efficacy in inhibition of H1975 cancer cells (IC50; 0.2–1.06 μmol/L), except for compound 7b, and was much