Linking ATP and allosteric sites to achieve superadditive binding with bivalent EGFR kinase inhibitors

The development of bivalent or heterobifunctional molecules, which consist of two functional motifs connected by a linker, represents a promising strategy in drug discovery. This approach offers the potential for enhanced target binding and unique biological effects compared to using monovalent inhibitors. However, designing effective bivalent inhibitors is challenging due to the complex interplay between the linker structure, the orientation of the two functional motifs, and their interaction with the target. Optimizing the linker is crucial because even subtle changes in its structure can dramatically impact the compound's potency. Fragment-based drug discovery (FBDD) frequently employs this strategy, connecting low-molecular-weight building blocks to create high-affinity molecules that bind to distinct sites on a target protein. A desirable outcome in FBDD is superadditivity, where the linked molecule displays significantly higher potency than the sum of its individual components. However, achieving superadditivity is uncommon. Tyrosine kinase inhibitors (TKIs) often provide successful examples of this strategy. Kinase inhibitors commonly bind to the ATP-binding site (orthosteric site), but allosteric inhibitors, which bind to distinct regions of the kinase, have also been developed. The epidermal growth factor receptor (EGFR) is a key target in non-small cell lung cancer (NSCLC), and its kinase domain has become a prime focus for drug development due to its frequent mutation and role in cancer progression. Clinically effective TKIs usually show selectivity for specific activating mutations like L858R and exon 19 deletion, while other mutants develop resistance to these TKIs. Studies have demonstrated synergy in tumor regression when combining ATP-competitive and allosteric EGFR inhibitors, suggesting that targeting both sites may enhance therapeutic outcomes and potentially delay the emergence of drug resistance. The close proximity of the ATP and allosteric binding sites in EGFR has led to the design of bivalent molecules aiming to simultaneously occupy both, potentially leading to superior potency and selectivity. This study focuses on synthesizing and characterizing a series of bivalent EGFR kinase inhibitors that bind simultaneously to both the ATP and allosteric sites, varying the linker structure to investigate its impact on potency and binding mode.

The literature extensively documents the challenges and successes in designing bivalent inhibitors. Studies highlight the importance of linker structure, demonstrating that even small changes can significantly alter the molecule's activity. The concept of superadditivity in fragment-based drug design has been explored, with the general consensus being that achieving this desirable effect is relatively rare. Numerous examples of bivalent molecules targeting various proteins are available, illustrating the broad applicability of this approach. Several studies specifically focus on bivalent kinase inhibitors, particularly EGFR inhibitors, showcasing the potential benefits of targeting multiple sites within the kinase domain. The prior research also demonstrates the synergistic effects of combining ATP-competitive and allosteric inhibitors for EGFR, providing the rationale for designing bivalent molecules that occupy both sites simultaneously. The existing literature on EGFR inhibitors, encompassing both ATP-competitive and allosteric inhibitors, serves as a foundation for understanding the target’s characteristics and informs the design of the bivalent molecules explored in this study.

This research involved the synthesis of a series of bivalent EGFR kinase inhibitors. The study began by selecting established ATP-competitive inhibitors based on trisubstituted imidazole molecules and mutant-selective allosteric 5,10-dihydro-11H-dibenzo[b,e][1,4]diazepin-11-one inhibitors. The researchers synthesized bivalent inhibitors with different linkers, including an N-linked methylene and C-linked amide structures. Detailed synthetic schemes are provided in the supplementary information. Biochemical assays, specifically HTRF-based activity assays, were performed to evaluate the inhibitory potency of the synthesized compounds against purified EGFR kinase domains, including wild-type and mutant variants (L858R, L858R/T790M, and L858R/T790M/C797S). To compare the potency of the bivalent inhibitors with their corresponding monovalent components, the researchers also tested related ATP and allosteric site analogues. X-ray crystallography was employed to determine the co-crystal structures of two selected bivalent inhibitors (N-linked and C-linked) bound to the EGFR(T790M/V948R) kinase domain. This structural analysis revealed the binding modes of the bivalent inhibitors and provided insights into the influence of the linker on the orientation of the functional motifs. Molecular dynamics (MD) simulations were conducted to further investigate the dynamic behavior of the bivalent inhibitors bound to EGFR. These simulations provided additional information on the binding interactions and stability of the complexes, complementing the crystallographic data. Cellular assays, including Western blotting and anti-proliferation assays using various human NSCLC cell lines (H1975, H3255, HCC827), were performed to assess the efficacy of the most potent bivalent inhibitor in a cellular context. Metabolic stability in human liver microsomes was also evaluated to assess the compound's stability in a relevant physiological environment. Computational docking studies were performed to predict the binding mode of the covalent inhibitor.

The study revealed a striking difference in potency between the N-linked and C-linked bivalent EGFR inhibitors. The N-linked inhibitor showed limited potency against all EGFR variants tested, with IC50 values exceeding 1 µM. In contrast, the C-linked inhibitors exhibited significantly higher potency, with IC50 values in the low nanomolar to picomolar range against drug-resistant mutants (L858R/T790M and L858R/T790M/C797S). This difference in potency was approximately 103 to 106-fold compared to the parent molecules. Crystal structures revealed that the N-linked inhibitor bound to EGFR with an "outward" conformation of the allosteric moiety, while the C-linked inhibitor displayed an "inward" conformation. The C-linked inhibitor exhibited several additional interactions with the protein (e.g., H-bonding), which were not possible in the N-linked compound. Molecular dynamics simulations showed the allosteric benzo moiety’s inward conformation in the C-linked inhibitor enhanced hydrophobic interactions and displaced water molecules, contributing to the higher binding affinity. The crystallographic B-factors and MD simulations indicated that the superior potency of the C-linked compounds resulted from several factors attributable to the linker, such as improved mobility and enhanced pocket complementarity. The most potent bivalent inhibitor (C-linked, covalent) effectively suppressed EGFR phosphorylation in various human NSCLC cell lines, including those expressing drug-resistant mutations. This compound also exhibited anti-proliferative activity and was shown to be metabolically stable in human liver microsomes.

The significant potency differences observed between the N-linked and C-linked bivalent EGFR inhibitors highlight the critical role of linker structure in bivalent inhibitor design. The "inward" conformation of the allosteric moiety in the C-linked inhibitor, enabled by the linker, leads to superior binding affinity through increased interactions with the EGFR kinase domain and improved structural complementarity with the allosteric binding pocket. This study underscores the importance of considering the linker's point of connection to the functional motifs in bivalent inhibitor design, as a shift in the attachment point can profoundly influence the binding mode and consequently the compound's potency. This observation provides a new perspective on linker optimization strategies, suggesting that exploring alternative points of connection in early-phase drug discovery may streamline the process and improve outcomes. The results presented have implications for FBDD, emphasizing the need to explore linker variations beyond traditional parameters such as length or functional group modifications. The findings indicate that the C-linked bivalent inhibitor demonstrates therapeutic potential due to its activity against drug-resistant EGFR mutants in various NSCLC cell lines. However, its relatively high molecular weight might affect membrane permeability and limit its effectiveness.

This research demonstrates a significant enhancement in the potency of bivalent EGFR kinase inhibitors by strategically modifying the linker structure. The C-linked amide linker yielded superadditive binding, exceeding the combined potency of the individual parent molecules. Crystal structures and MD simulations revealed the structural basis for this enhanced potency, linked to improved mobility and conformational changes facilitated by the linker in the allosteric binding site. This study emphasizes the importance of considering linker design beyond traditional parameters, particularly the point of connection, to optimize bivalent inhibitor activity. Future studies should focus on improving the medicinal chemistry properties of such compounds to overcome limitations like membrane permeability.

The study focused primarily on a limited set of linker variations, therefore the findings might not be broadly generalizable to all types of bivalent molecules. The high molecular weight of the most potent compound could impact its pharmacokinetic properties, such as membrane permeability. Further investigation is needed to explore the potential impact of this high molecular weight on in vivo efficacy. Finally, while the cell-based assays demonstrate promising activity, in vivo studies are required to confirm the therapeutic potential of these bivalent inhibitors.