By Nancy Pomarici

Antibodies are normally produced by our immune system, as a defence against different pathogens, as bacteria or viruses. Their ability of selectively bind the target, called antigen, makes them promising therapeutics, as treatment for cancer, immune disease or other pathologies.

In order to be approved as a therapeutic, a molecule has to respect certain criteria that ensure its drug-like properties, manufacturability and safety profile. The Lipinski’s rule of 5 have been widely applied to assess the drug-likeness of small molecules, but they are not valid for macromolecules, as antibodies. In this case, other properties have to be evaluated to assess the developability of the biomolecules, such as hydrophobicity, viscosity, polyspecificity, self-aggregation and stability. The latter has particular relevance, considering that antibody proteins have to function in harsh extracellular environments, sometimes at high dilution. Therefore, stability is key in the drug development process and worthy to be thoroughly investigated.

From a structural point of view, antibodies are Y-shaped proteins composed by two heavy and two light chains. The two arms of the Y are named antigen binding fragments (Fab) and they interact directly with the antigen. The crystallizable fragment (Fc), instead, is responsible for the interactions with the immune system to trigger its response. Focusing the attention on the Fab, other two sections can be recognized: the two lower domains (CH1-CL), one belonging to the heavy and one to the light chain, are called constant domains, because their sequence and structure is quite conserved between different antibodies. The upper two domains, instead, shape the variable fragment (Fv), since it has much more variability in sequence and structure. The binding happens in the Complementary Determining Region (CDR), which is shaped by six hypervariable loops. The high variability of the structure allows the antibody to bind a wide range of antigens.

Figure 1. Schematic representation of an antibody.

In one of my PhD projects, we investigated the stability of 16 Fab structures, obtained from the different pairings of four heavy chains with four light chains. Starting from their crystal structures, MD simulations have been performed. Conventional Molecular Dynamics simulations (cMD) allow to visualize movements in the interface and sidechain rearrangements, but they are not sufficient to observe the collective motion in the protein. Therefore, enhanced sampling techniques are applied, in which a boost is applied to help the structure to overcome high energy barriers and visit a wider area of the free energy landscape.

Between the several enhanced sampling techniques, umbrella sampling was chosen in order to boost the domain apart and observe their dissociation. An advantage of umbrella sampling is the possibility of obtaining the free energy curve of the process and, therefore, calculate the energy of the depth of the minimum that corresponds to the undissociated (stable) structure. These values are in agreement with the experimentally measured melting temperatures of the systems, as a proof that our simulations well represent the reality.

The analysis of the trajectories can provide us with interesting insights regarding the mechanism of dissociation of the structures and stabilizing interactions in the interface.

First of all, the dissociation process can happen with the same probability in the Fv region or in the lower CH1-CL region. When this happens in the variable region, the interactions that keep the domains together until dissociation are located not only in the framework region, but also in the CDR loops, meaning that these region does not only bind the antigen, but it also has an important role in the stabilization of the structure.

In conclusion, the analysis of the mechanism of dissociation of antibodies fragments has a huge impact in drug development, in order to enhance the stability of the drugs and therefore facilitate their approval to the market.