MALT1
August 2022
Mucosa-associated lymphoid tissue lymphoma translocation protein 1
MALT1 (mucosa-associated lymphoid tissue lymphoma translocation protein 1) is a signaling modulator playing a central role in the activation of immune pathways and is a validated target for autoimmune diseases and several cancer types. MALT1 was originally believed to have only a scaffolding function in the formation of a multi-protein complex regulating the signaling pathway. Later, it was discovered that it also shows protease activity, making it an attractive pharmacological target.
Many companies are currently investigating MALT1 as a therapeutic target such as Novartis, Janssen, and even molecule modeling company Schrödinger, whose MALT1 inhibitor was recently approved for a human clinical trial.
Here we discuss a study conducted by a Novartis team, who has discovered two classes of molecules, with different inhibition mechanisms. They determined the crystal structures of the ligand-MALT1 complexes, which supported the optimization of the most promising compound. This is a nice example of a structure-driven lead optimization process.
Previous studies on MALT1 have shown that is inactive as a monomer, while the catalytically active form is the dimer. A common therapeutic strategy for proteases is to target the catalytic site in its active conformation, using an irreversibly binding ligand. This is often a short peptidic sequence mimicking the protease natural substrate sequence, with a chemical moiety that covalently binds to the active site. However, these types of molecules have limitations in their therapeutic applicability due to the irreversibility of the binding, their size, and cost, so scientists decided to develop a small molecule inhibitor.
Following a high-throughput screening of the Novartis collection containing around 1 million compounds, two hit series were identified:
one binding to the catalytic site (orthosteric)
one to an allosteric pocket
Both ligands stabilize MALT1 in an inactive conformation, disrupting the catalytically active fold (Image 1).
Image 1. Different conformations of the MALT1 catalytic site: the active conformation (structure in yellow, PDB ID: 3V4O) is disrupted both by the binding of the orthosteric ligand (structure in orange, PDB ID: 6YN8) and the distant allosteric ligand (structure in blue, PDB ID: 6YN9; ligand not shown). E500 is selected as reference residue in the three structures. The image was prepared with 3decision highlight mode: only differences among the structures are displayed.
However, the orthosteric ligand had a weak biochemical activity that did not translate into cellular activity, so it was not further developed. The allosteric ligand (Compound 2), containing a sulfonamide moiety, was highly potent, so it was optimized to improve its efficacy.
From compound 2 to compound 7 (MLT-208) with structure-based insights
The first modification of this allosteric series was to add one carboxylic acid to the terminal furane ring (compound 3) to increase the solubility for biophysical and structural studies. They succeeded in obtaining the structure of MALT1 in complex with compound 3. They observed that the binding is mainly driven by hydrophobic contacts at the interface between the caspase and Ig3 domains.
Interestingly, this allosteric mechanism is led by the flip of the side chain of the tryptophane 580. Compound 3 is filling this hydrophobic sub-pocket by its diethyl phenyl moiety, and that is how W580 pocket inhibitors are stabilizing an inactive conformation.
Image 2. Superposition of the allosteric pocket with ligand bounded (structure in purple, a ligand in orange, PDB ID: 6YN9) and apo inactive structure (structure in white, PDB ID: 3V55). The residue W580 in the inactive conformation is located inside the pocket. The ligand replaces this residue and stabilizes the protein in its inactive conformation. The picture is taken with 3decision - a protein structure repository and data management tool.
Then, a structure-property relationship study was engaged to find an optimal balance between lipophilicity (clogP) and potency (pIC50). They monitored this optimization using the “ligand- lipophilic efficiency” (LLE) parameter (see Table below):
LLE = pIC50 – clogP
Firstly, compound 5 was discovered through a scaffold expansion campaign focusing on the biaryl part.
Secondly, based on the crystal structure, there is a hypothesis that only one of the two ethyl groups is important to keep the active conformation and therefore the other one can be modified. This led to compound 6, which is water soluble and active.
Thirdly, following identifying one metabolic hot spot next to the pyrrolidine nitrogen of compound 6, they blocked it by adding a carbonyl group (Image 3).
Compound Table
| Compound number | MALT1 biochemical IC50 (nM) | IL2 jurkat cellular IC50 (nM) | clogP | LLE | Solubility (µM) |
|---|---|---|---|---|---|
| 2 | 180 | 2000 | 3.8 | 2.9 | <4 |
| 3 | 3700 | - | - | - | - |
| 5 | 38 | 943 | 3.6 | 3.8 | 54 |
| 6 | 10 | 279 | 2.8 | 5.2 | 29 |
| 7 | 15 | 53 | 2.9 | 4.9 | 184 |
Image 3. Lead optimization steps - from compound 2 to compound 7
The compound 7, also called MLT-208, is soluble in water, highly potent, highly selective against a representative panel of proteases and shows promising pharmacokinetic properties in rat.
This is a nice example of a structure-based drug discovery approach. The structural elucidation of ligand-protein complex successfully drove lead optimization, supporting and accelerating the design of new compounds with increased solubility, potency and selectivity.