Optimization of Brigatinib as New Wild-Type Sparing Inhibitors of EGFR^T790M/C797S Mutants

Abstract

A series of brigatinib derivatives were designed and synthesized as new potent and selective EGFR^T790M/C797S inhibitors. One of the most potent and selective compounds 18k strongly suppressed the EGFR^{L858R/T790M/C797S} and EGFR^{19Del/T790M/C797S} kinases with IC₅₀ values of 0.7 and 3.6 nM, respectively, which were over 54-fold more potent than the lead compound. 18k also demonstrated promising EGFR^T790M/C797S mutant selectivity, and was 94-fold less potent against the wild type EGFR. A cocrystal structure of EGFR^T790M/C797S with a close derivative 18f was solved to provide insight on the inhibitor’s binding mode. Moreover, compound 18k was orally bioavailable and demonstrated highly desirable PK properties, making it a promising lead compound for further structural optimization.

Epidermal growth factor receptor (EGFR) is a well validated target for non-small cell lung cancer (NSCLC) drug discovery.¹ Five small molecular EGFR inhibitors (i.e., gefitinib (1),² erlotinib (2),³ afatinib (3),⁴ dacomitinib (4),⁵ and osimertinib (AZD9291, 5),^6,7Figure 1) have been approved by the U.S. Food and Drug Administration (FDA). The first generation drugs 1 and 2 clinically apply to the treatment of NSCLC patients with EGFR activating mutations (e.g., exon 19 deletion (19 Del) and a Leu⁸⁵⁸ → Arg⁸⁵⁸ point mutation (L858R)). Irreversible compounds 3 and 4 belong to the second generation EGFR inhibitor drugs which achieve significant clinical benefit for metastatic and advanced NSCLC patients harboring “gain of function” mutated EGFRs. Although drugs 3 and 4 exhibit strong suppression on the resistant Thr⁷⁹⁰ → Met⁷⁹⁰ (T790M) mutants in preclinical models, the low clinical maximal tolerated dose (MTD) attributing to the strong wild-type EGFR (EGFR^WT) inhibition limited their application in resistant patients with the EGFR^T790M mutant. Wild-type sparing EGFR^T790M inhibitor 5 is the first FDA approved third generation drug for resistant NSCLC patients with EGFR^T790M mutations, and it was later approved as a first-line treatment for metastatic NSCLC patients with positive EGFR mutations. Following the success of 5, two structurally close derivatives, i.e., almonertinib (6)⁸ and furmonertinib (7),⁹ were approved by National Medical Products Administration of China (NMPA) in 2020 and 2021, respectively. Several other third generation EGFR inhibitors, e.g., lazertinib, rezivertinib, and abivertinib, are in late stages of clinical investigation.¹⁰ We have also developed ASK120067 (8) as a highly potent and selective wild-type sparing EGFR^T790M inhibitor. A New Drug Application (NDA) was filed recently in China based on its highly promising efficacy in advanced and multiple brain metastatic NSCLC patients harboring EGFR^T790M mutations.¹¹

Figure 1.

Chemical structures of approved or literature reported EGFR inhibitors.

Structurally, all of the third generation EGFR inhibitors possess an acrylamide warhead to covalently react with the Cys⁷⁹⁷ residue to achieve improved competitiveness with the high concentration ATP (mM level) in cells. A tertiary C797S (Cys⁷⁹⁷ → Ser⁷⁹⁷) mutation would impair the covalent formation between Cys⁷⁹⁷ and the acrylamide warhead and induce the resistance against the third generation drugs in approximately 20% of the pretreated NSCLC patients.¹² Although the patients with trans EGFR^T790M/C797S mutations in which the T790M and C797S mutations are located on different alleles could be efficiently treated by combinational therapies of a first generation drug and 5, the cis- EGFR^T790M/C797S mutation (mutations are on a same allele) mediated resistance remains an unmet clinical need.^13,14 Therefore, development of the fourth generation EGFR^T790M/C797S inhibitors is highly desirable for overcoming the new mutation mediated resistance.¹⁵

Several classes of EGFR^T790M/C797S inhibitors have been reported. Examples include the traditional ATP competitive inhibitor brigatinib (9) and derivatives,¹⁶ 4-hydroxy-3-(4-aminoquinazolin-2-yl)benzonitrile derivative 10,¹⁷ trisubstituted imidazole derivative 11,¹⁸ 5-methylpyrimidopyridone derivative 12,¹⁹N-(pyrimidin-4-yl)-5H-pyrido[4,3-b]indol-3-amine analogue 13,²⁰ allosteric inhibitors EAI045 (14)²¹ and JBJ-04-125-02 (15),²² and macrocyclic molecules BI4020 (16).²³ To the best of our knowledge, only the brigatinib derivative TQB3804 (17) was advanced to clinical investigation in resistant NSCLC patients with EGFR^T790M/C797S mutations (NCT04128085).^24,25 Herein, we would like to report a straightforward structural optimization of compound 9. The effort eventually yielded the discovery of a new highly potent and selective EGFR^T790M/C797S inhibitor 18k with approximatly 54-fold improved kinase inhibitory potency.

A cocrystal structure of EGFR^WT with inhibitor 9 was recently reported (PDB 7AEM, Figure 2a in stick and line representation and Figure 2b in surface representation).²⁶ The results showed that NH group of the aniline in 9 made a donor interaction with the backbone carbonyl of Met⁷⁹³, while the pyrimidine N accepted a hydrogen bond from the backbone NH of Leu⁷⁹², in the hinge region. An oxygen atom of the dimethylphosphine oxide (DMPO) group formed acceptor interactions with the side chain NH₃ of the conserved catalytic Lys⁷⁴⁵ and the side chain OH of Thr⁸⁵⁴, respectively, whereas the hydrophilic piperazinylpiperdine moiety was oriented to the solvent accessible surface, indicating it might not contribute much to the protein binding. A preliminary computational study suggested it might bind with EGFR^T790M/C797S in a similar mode (Figure 2c). The structural information also suggested that the 2′-methoxyl group might cause potential steric clash with Pro⁷⁹⁴ although it could occupy a small hydrophobic cavity in EGFR to improve the target selectivity. This was consistent with the previous observation that removal of the corresponding methoxyl group resulted in improved EGFR kinase inhibitory potency for the third generation inhibitors.^27,28

Figure 2.

X-ray crystal structure of brigatinib (9) with EGFR^{Thr790/Cys797}(EGFR^WT) kinase domain (PDB 7AEM) shown in cyan and gray in stick and line representation (a) or in surface representation (b). Preliminary computational docking complex of 9 (c) and 18a (d) with EGFR^{Met790/Ser797} (PDB 5ZTO). Hydrogen bonds are indicated by yellow hatched lines to key amino acids.

A preliminary docking study also suggested that a change of the 2′-methoxyl group to 3′- position would increase the EGFR inhibitory activity by avoiding the potential steric clash (Figure 2d). Therefore, a series of brigatinib derivatives with substituents at the 3′- position of 2-aniline group were synthesized by following the reported procedures (Scheme 1).²⁹ Beginning with the commercially available 2-iodoaniline (19) and dimethylphosphine oxide, intermediate 20 was obtained via a palladium-catalyzed coupling reaction. Then a two-step, sequential S_NAr reaction was conducted to give the target compounds 18a–18k. Biological evaluation showed that compound 18a indeed exhibited a 16-fold improved EGFR^{L858R/T790M/C797S} (EGFR^TM1) inhibitory potency with an IC₅₀ value of 2.4 nM compared with 9 (Table 1). In addition , the switch of the substituent position also resulted in EGFR^WT and EGFR^L858R/T790M (EGFR^DM) potency improvement by factors of 5.1 and 21.2, respectively. Notably, compound 18a exhibited an improved wild-type sparing selectivity with an EGFR^WT/EGFR^TM1 ratio of 21.3, while the corresponding value for 9 was 6.8. Further investigation also revealed that the 3′-position could be substituted by a variety of substituents, e.g., −CH₃ (18b), −OCF₃ (18c), −CF₃ (18d), −CF₂H (18e), and −Cl (18f), to demonstrate wild-type sparing strong inhibition against the resistant EGFR^TM1 and EGFR^DM kinases. Particularly, compounds 18b and 18f exhibited the strongest EGFR^TM1 inhibition with IC₅₀ values of 0.7 and 0.6 nM, respectively. The compounds could also potently suppress EGFR^DM kinase with IC₅₀ values of 2.7 and 1.8 nM, but their activities against EGFR^WT were 23- and 21-fold less potent, respectively. Additionally, results showed that a replacement of 5-Cl of the pyrimidine motif in 18f by a Br group (18g) barely led to an obvious effect on the activity and selectivity, while the 5-CF₃ substituted analogue (18h) totally abolished the EGFR inhibitory potency. The investigation also suggested the aniline group at the pyrimidine 2-position could be replaced by a pyridinyl moiety (18i) without causing a significant potency change to the EGFR triple mutated kinase (IC₅₀ = 2.1 nM). Further investigation revealed that 3′,5′-dicholor analogue 18j was 55-fold less potent than 18f against EGFR^TM1 triple mutant, while the 2′,5′-dicholor compound 18k exhibited identical EGFR^TM1 inhibition to that of 18f, with an IC₅₀ value of 0.7 nM. Further evaluation suggested that 18k potently suppressed the EGFR^{19Del/T790M/C797S} mutated kinase, which is a clinically more common resistant mutation, with an IC₅₀ value of 3.6 (±1.3) nM. In addition, 18k was also highly potent against EGFR^L858R/T790M with an IC₅₀ value of 0.7 nM, but its potency against EGFR^WT was 94.1-fold less potent.

Scheme 1. Synthesis of Compounds 18a–18k.

Reagents and conditions: (a) dimethylphosphine oxide, Pd(OAc)₂, xantphos, K₃PO₄, DMF, 120 °C, overnight, 82%; (b) R₁ substituted pyrimidine, K₂CO₃, n-Bu₄NHSO₄, DMF, 65 °C, overnight, 63–84%; (c) R₂ substituted aniline, 2.5 M HCl in EtOH, 2-methoxyethanol, 120 °C, overnight, 37–75%.

Table 1. In Vitro EGFR Inhibition and Antiproliferation Data of Compounds 18a–18k^a.

	substituents		kinase activity IC50 (nM)b
compd	R1	R2	EGFRWT	EGFRDM	EGFRTM1	WT/TM1 ratioc
18a	Cl	2′-H, 3′-OCH3	51.2 ± 21.2	2.4 ± 0.4	2.4 ± 1.0	21.3
18b	Cl	2′-H, 3′-CH3	62.3 ± 11.5	2.7 ± 0.7	0.7 ± 0.3	89.0
18c	Cl	2′-H, 3′-OCF3	157.0 ± 26.4	9.4 ± 3.1	12.9 ± 3.2	12.2
18d	Cl	2′-H, 3′-CF3	538.1 ± 62.7	32.8 ± 3.6	3.5 ± 0.9	153.7
18e	Cl	2′-H, 3′-CF2H	198.7 ± 41.2	13.7 ± 3.3	23.1 ± 1.6	8.6
18f	Cl	2′-H, 3′-Cl	38.3 ± 5.6	1.8 ± 0.4	0.6 ± 0.3	63.8
18g	Br	2′-H, 3′-Cl	113.1 ± 24.2	1.8 ± 0.5	3.3 ± 0.4	34.3
18h	CF3	2′-H, 3′-Cl	>1000	>1000	>1000
18i	Cl	2′-H, 3′-CH3, 5′ = N	337.6 ± 141.6	25.3 ± 6.1	2.1 ± 0.7	160.8
18j	Cl	3′-Cl, 5′-Cl	299.8 ± 169.7	29.4 ± 8.0	33.5 ± 3.5	8.9
18k	Cl	2′-Cl, 5′-Cl	65.9 ± 10.7	0.7 ± 0.1	0.7 ± 0.3	94.1
9	Cl	2′-OCH3, 3′-H	261.3 ± 78.7	50.9 ± 12.8	38.3 ± 7.0	6.8
5			207.5 ± 93.6	2.0 ± 0.1	236.6 ± 96.0	0.88

^aEGFR^WT, EGFR^DM and EGFR^TM1 kinase inhibition was tested by ELISA assay. The data are mean values ± SD from at least three independent experiments.

^bTM1, L858R/T790M/C797S; DM, L858R/T790M.

^cWT/TM1 ratio was calculated by IC₅₀(EGFR^WT)/IC₅₀(EGFR^TM1), which reflected the selectivity of the compounds to EGFR^WT and EGFR^TM1.

The X-ray cocrystal structure of EGFR^T790M/C797S with 18f was solved (PDB 7ER2, Figure 3a). Similar to our prediction, the results showed that the molecule anchored in the hinge region of the ATP-binding pocket with bidentate hydrogen bond interactions between Met⁷⁹³ and 2-aminopyrimidine motif. Interestingly, structural alignment of EGFR-18f complex and EGFR-brigatinib complex (PDB 7AEM) showed that the backbone amido linkages between Thr⁸⁵⁴ and Asp⁸⁵⁵ pointed in opposite directions (Figure 3b). Therefore, instead of forming a hydrogen bond directly with side chain OH of Thr⁸⁵⁴ as exhibited in Figure 2a, the oxygen atom on the DMPO group of 18f fostered water-molecule-mediated hydrogen bond interactions with the backbone carbonyl of Thr⁸⁵⁴ and additionally the side chain NH₃ of Lys⁷⁴⁵, inducing a stable conformation and improving binding affinity toward the protein. In addition, the 3′-Cl in 18f located in an appropriate cavity with no steric clash with any residues, providing a rational explanation for the improved EGFR inhibitory potency. Additionally, the C5-chlorine atom on the pyrimidine scaffold was directed toward the “gatekeeper” residue Met⁷⁹⁰ with a distance of 3.2 Å, indicating that this position was unable to accommodate large hydrophobic groups such as CF₃, which explains the poor potency of 18h.

Figure 3.

(a) X-ray crystal structure of EGFR^T790M/C797S in complex with 18f (PDB 7ER2). Key residues and 18f are shown in magenta by the element in lines and in sticks, respectively. Hydrogen bond interactions are showed as red dotted lines, and the water molecule is indicated by a red dot. (b) Superposition of 18f (PDB 7ER2; compound and protein shown as magenta sticks and lines, respectively) and brigatinib complex (PDB 7AEM; compound and protein shown as cyan sticks and lines, respectively).

Having demonstrated the selectivity of 18k for its direct target EGFR, we next examined its inhibitory activities against a small kinome panel at a concentration of 100 nM. The results in Figure 4 showed that 18k exhibited >30× selectivity in all 28 tyrosine kinases examined (see the Supporting Information for raw inhibition rate). Strikingly, 18k markedly attenuated the inhibitory effect on ALK, the original target of brigatinib, indicating its improved kinase selectivity.

Figure 4.

Kinase inhibitory activity of 18k against 28 different kinases at a concentration of 100 nM. Kinase inhibition was tested by ELISA assay. Inhibition rate are shown as mean ± SD.

Furthermore, the antiproliferative activity of 18k was evaluated in BaF3 cells stably expressing EGFR^{L858R/T790M/C797S} (BaF3-EGFR^TM1) and EGFR^{19Del/T790M/C797S} (BaF3-EGFR^TM2) mutants (Table 2). It was shown that 18k exhibited moderate growth inhibition against the cell models with IC₅₀ values of 0.258 and 0.141 μM, respectively, which were comparable to those of drug 9, whereas its cytotoxic effect on BaF3 parental cells was a bit more potent with an IC₅₀ value of 1.69 μM. Compound 18k also exhibited a 0.60 μM IC₅₀ value against the proliferation of self-constructed osimertinib-resistant PC9 NSCLC cells harboring EGFR^{19Del/T790M/C797S} mutation (PC9-EGFR^TM2). Selectivity was observed against PC9 (EGFR^19Del) and H1299 (EGFR^WT) cell lines, which was within the kinase selectivity margin. Further immunoblotting analysis suggested that the compound dose-dependently suppressed the phosphorylation of EGFR in BaF3-EGFR^TM1, BaF3-EGFR^TM2, as well as PC9-EGFR^TM2 cells, supporting its cellular inhibitory potency on the resistance related EGFR mutations (Figure 5).

Table 2. Antiproliferative Activity of 18k in BaF3, PC9, PC9-EGFR^TM2, and H1299 Cell Lines^a.

	BaF3
IC50 (μM)	EGFRTM1	EGFRTM2	parental cells	PC9	PC9-EGFRTM2	H1299b
18k	0.258 ± 0.167	0.141 ± 0.081	1.691 ± 0.253	1.192 ± 0.727	0.603 ± 0.294	2.115 ± 0.302
brigatinib (9)	0.286 ± 0.176	0.155 ± 0.046	4.191 ± 1.614	0.829 ± 0.148	1.110 ± 0.028	1.049 ± 0.160
AZD9291 (5)	2.893 ± 0.601	2.885 ± 0.689	4.427 ± 1.162	0.011 ± 0.005	3.210 ± 0.853	>10.0

^aBaF3-EGFR^TM1, BaF3-EGFR^TM2, and PC9-EGFR^TM2 cell lines were built by Jian Ding’s laboratory, Shanghai Institute of Materia Medica, Chinese Academy of Sciences. PC9-EGFR^TM2 cell line was constructed by introducing the T790M/C797S mutant into PC9 (EGFR^19Del). Cellular assay was determined using the CCK8 assay.

^bH1299 (EGFR^WT) cellular assay was determined using sulforhodamine B (SRB) assay. The data were shown as mean ± SD from at least three independent experiments. TM1, L858R/T790M/C797S; TM2, 19Del/T790M/C797S.

Figure 5.

Target inhibition potency of 18k in EGFR^C797S mutant BaF3 cells and PC9 cells. Cells were treated with indicated compounds for 2 h, and the activation of EGFR and downstream signaling molecule were evaluated using Western Blot analysis.

The pharmacokinetic (PK) parameters of 18f and 18k were also assessed in Sprague–Dawley (SD) rats (Table 3). Both compounds exhibited good PK properties with high oral bioavailability, decent exposure, and acceptable half-life. For instance, 18k exhibited an oral bioavailability value of 81.7%, an AUC (area under curve) value of 5219 h·ng/mL, and a T_1/2 value of 2.67 h, which were optimal for oral administration for the following in vivo efficacy investigation.

Table 3. Mean PK Parameters of 18f and 18k after po or iv Administration in SD Rats^a.

	18f		18k
parameters	pob	ivc	pob	ivc
T1/2 (h)	2.84 ± 0.11	3.73 ± 1.36	2.16 ± 0.2	2.67 ± 0.45
Tmax (h)	1.50 ± 0.87		2.67 ± 1.15
Cmax (ng/mL)	470 ± 155	345 ± 52	868 ± 181	288 ± 16
AUC(0-t) (h·ng/mL)	2553 ± 704	862 ± 88	4552 ± 1154	929 ± 91
AUC(0-∞) (h·ng/mL)	2563 ± 709	898 ± 57	5219 ± 1536	1060 ± 144
CL (mL/min/kg)		9.30 ± 0.57		7.97 ± 1.17
MRT(0-∞) (h)	4.79 ± 0.49	4.08 ± 1.24	4.37 ± 0.62	3.65 ± 0.56
Vss (mL/kg)		2275 ± 710		1720 ± 40
F (%)	49.3 ± 11.8		81.7 ± 20.7

^an = Three animals per study.

^bDosed at 3 mg/kg.

^cDosed at 0.5 mg/kg.

In summary, a straightforward structure optimization of brigatinib was conducted to yield compound 18k as a new potent and selective EGFR^T790M/C797S inhibitor. The compound strongly suppressed the EGFR^{L858R/T790M/C797S} and EGFR^{19Del/T790M/C797S} kinases with sub-nanomolar IC₅₀ values but was approximately 94-fold less potent against the wild type EGFR. A cocrystal structure of EGFR^T790M/C797S with a close derivative 18f was also solved to provide understanding of the binding mode of the inhibitor. Moreover, compound 18k exhibited cellular activities in all three cell models, BaF3-EGFR^TM1, BaF3-EGFR^TM2, and PC9-EGFR^TM2, comparable to that of brigatinib. Compound 18k is orally bioavailable and demonstrates highly desirable PK properties, making it a promising lead compound for further structural optimization.

Glossary

Abbreviations

EGFR

epidermal growth factor receptor

NSCLC

nonsmall cell lung cancer

19Del

Exon 19 deletion

MTD

maximal tolerated dose

WT

wild type

NMPA

National Medical Products Administration of China

FDA

U.S. Food and Drug Administration

NDA

New Drug Application

DMPO

dimethylphosphine oxide group

TM1

L858R/T790M/C797S

TM2

19Del/T790M/C797S

DM

L858R/T790M

PK

pharmacokinetic

AUC

area under curve

SD rats

Sprague–Dawley rats

ALK

anaplastic lymphoma kinase

CCK8

cell counting kit 8

SRB

sulforhodamine.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00555.

• Chemistry, biological assay, X-ray crystallography data, and copies of NMR spectra and HPLC purity data (PDF)

Author Contributions

^∇ S.L., T.Z., and S.-J.Z. contributed equally to this work.

The authors declare no competing financial interest.