Mcl1
July 2023
Myeloid Cell Leukemia 1 (Mcl1)
Myeloid Cell Leukemia 1 (Mcl1) is a protein that plays a crucial role in regulating apoptosis (programmed cell death). Mcl1 promotes the survival of cells by binding to other proteins that induce cell death, blocking their action.
In the context of cancer, Mcl1 has gained significant attention due to its role in promoting tumor cell survival and resistance to various anti-cancer therapies. Overexpression of the Mcl1 gene is frequently observed in many types of cancers, including leukemia, multiple myeloma, lung cancer, and colon cancer. Therefore, Mcl1 has emerged as a promising target for cancer treatment.
Several companies (such as Amgen, Servier, AstraZeneca, Abbvie) are actively investigating the development of Mcl1 inhibitors that can block its function. Several Mcl1 inhibitors entered preclinical and clinical trials, but some of them were interrupted due to safety concerns. Some studies have demonstrated that prolonged inhibition of Mcl-1 induces cardiotoxicity, posing new challenges for the development of safe medications.
A recent paper from Janssen addressed this issue, focusing on the development of molecules with improved toxicity profiles, with particular attention to cardiac safety. Structural considerations supported the discovery of a new oral Mcl1 inhibitor with low cardiotoxicity.
All the small-molecule Mcl-1 inhibitors with a reported 3D structure target the region of the protein that binds its protein partners (BH3 domain), modulating their protein-protein interactions (Image 1A). For instance, most of the known inhibitors bind to Mcl-1 at the level of Arg263, replacing the native bond between Arg263 and Asp12 on the BH3 peptide of the protein partner (Image 1B).
Image 1. A: Representation of protein-protein interactions network between Mcl-1 (in orange, PDB: 5W8F) and Bim protein BH3 peptide (in purple, PDB: 5W8F). The interaction between Arg263 on MCL-1 and Asp12 on the peptide is represented in yellow, and pointed by the white arrow. Protein-protein contacts were calculated with the 3decision® software. B: Overlay of the Mcl-1 structure in complex with the native peptide (in purple, PDB: 5W8F) and with a known ligand, S64315/MIK665 (in blue, PDB: 6YBL). The residues of the binding pocket of the Mcl-1 are represented in blue. The interaction between the ligand and Arg263 is represented in dotted yellow lines, indicated by the orange arrow. The native Arg263 and Asp12 residues are highlighted in white and their interaction is a dashed yellow line, indicated by the white arrow. The pictures were produced with the 3decision® software.
The binding site is rather large and shows different sub-pockets that bind different portions of the ligands (Image 2). Pocket P1 and P2 are hydrophobic pockets, that are essentially always targeted by Mcl1 inhibitors for which co-crystal structures are available. Pocket P3 is mostly solvent-exposed but many of the compounds cocrystallized show interactions with this part of the pocket. Finally, pocket P4 is a small cleft that is not often occupied by Mcl1 inhibitors.
Image 2. BH3 binding domain of MCL-1 (in white, PDB: 6YBL) in complex with known inhibitors: S64315/MIK665 (in orange, PDB: 6YBL), AZD5991 (in brow, PDB: 6FS0) and AM-8621 (in blue, PDB: 6OQC). The Arg263 residue and the different sub-pockets (P1-P4) in the binding site are indicated. The protein-ligand interactions and the molecular surface of the pocket (colored by hydrophobicity) were calculated by the 3decision® software.
To improve potency and physiochemical properties, while minimizing the risk of cardiotoxicity, researchers at Janssen decided to develop a molecule with increased polar character, with the following structural features:
tight contact with the pocket P1 and P2, and with the Arg263 residue
interactions with the pocket P4
Even if the known inhibitors only show hydrophobic groups allocated in lipophilic pocket 2, they decided to explore the use of a polar tetrahydrofuran (THF) ring at this position of the ligand. They also included a fluorine atom in the molecule, which usually improves physiochemical properties. These modifications led to compound 12, which showed promising binding affinity to Mcl-1 (Ki 0.22 nM) and good cellular activity (AC50 191 nM), together with good solubility. From the X-ray crystal structure (PDB: 8G3T), 12 bound very similarly to the previously published inhibitors, and the THF ring was successfully accommodated in pocket P2 (Image 3A), where it formed a water-mediated contact with the protein. The ligand also interacted with the pockets P1 and P3, and also has a contact with Thr266 in pocket P4. From the X-ray structure, they observed the presence of co-crystallized water molecules in the P4 - for instance, the one bridging Thr266 and Gly262 (Image 3B). They reasoned that their displacement by growing the inhibitor could create the opportunity to establish new interactions with this subpocket.
Image 3. X-ray crystal structure of ligand 12 with Mcl1 (in brick red, PDB: 8G3T). A: focus on the pocket P2 and the water-mediated interaction with Ala227. Pockets P1 and P3 are also indicated. B: focus on the pocket P4, with the interaction between the ligand 12 and Thr266, and the water molecules occupying the pocket. Notice the water bridging residues Thr266 and Gly262. Protein-ligand interactions are calculated by the 3decision® software.
To expand the molecule in the pocket P4, they added a substituent at the sulfonimidamide group that engaged with the Thr266 residue. Compound 28, carrying this modification, drastically improved biochemical and cellular activity (Ki 0.011 nM; AC50 44 nM). The 3D structure of the protein-ligand complex (PDB: 8G3W) showed that the substituent engaged with pocket P4, establishing a hydrogen bond with the sidechain of Thr266 and displacing the waters observed in the complex with compound 12. Also, the sulfonimidamide group formed a water-mediated contact with Arg263. The rest of the molecule keeps essentially the same binding mode as compound 12 (Image 4).
Image 4. Superposition of Mcl-1 in complex with ligand 12 (in brick red, PDB: 8G3T) and compound 28 (in white, PDB: 8G3W). Waters that have been displayed by the expansion of the ligand are indicated by the red arrows. The interactions that the compound 28 form with Thr266 and Gly262 are indicated by the white arrows. Protein-ligand interactions are automatically calculated by the 3decision® software.
The compounds produced in this study were all further evaluated for cardiotoxicity, and they displayed low risk or up to 10 µM inhibitor concentration. Since compound 28 had high biochemical and cellular activity, together with excellent in vitro physiochemical properties, it was evaluated in vivo. All the good properties observed in vitro translated nicely in a mouse model.
In this study, structural considerations guided the optimization of a new molecule with improved potency and physiochemical properties, while minimizing the risk of cardiotoxicity. This work paves the way for the development of safer drugs for cancer treatment.
Reference:
Romanov-Michailidis F, Hsiao CC, Urner LM, Jerhaoui S, Surkyn M, Miller B, Vos A, Dominguez Blanco M, Bueters R, Vinken P, Bekkers M, Walker D, Pietrak B, Eyckmans W, Dores-Sousa JL, Joo Koo S, Lento W, Bauser M, Philippar U, Rombouts FJR. Discovery of an Oral, Beyond-Rule-of-Five Mcl-1 Protein-Protein Interaction Modulator with the Potential of Treating Hematological Malignancies. J Med Chem. 2023 May 11;66(9):6122-6148. doi: 10.1021/acs.jmedchem.2c01953. Epub 2023 Apr 28. PMID: 37114951.