Immunological peptide design by MD simulation
(collaboration with lab of Prof. Moustafa Gabr)
This project is an ongoing collaboration with the group of Prof. Moustafa Gabr at Cornell University to design next generation immune checkpoint inhibitors.
Background: Cancers are in part immunological diseases, e.g., because most tumor cells display surface ‘neoantigens’, which are normally recognized by immune T-cells (and NK-cells) as foreign. However, T/NK cell responses can be suppressed by various means, such as (i) changes in the tumor microenvironment, (ii) over-expression of immune checkpoint receptors. In turn, we can target these receptors or interactions to resensitize the T-cells, e.g. by designing antibodies to CTLA4, PD1, PDL1, LAG3, TIM3, etc:
Illustration of checkpoint blockade inhibition (CBI):
-Tumor cell overexpresses PD-L1
-PDL/PD-1 interaction ‘turns off’ T-cell
-Abs to PD1 or PDL1 resensitize T-cell
In this project, we used Xray crystal structures and MD simulation to develop peptides to inhibit the TIM-3 immune checkpoint protein, as outlined in the figures below. The robustness of the design process is demonstrated by the good correlation between computed peptide rank and the binding affinity.
Peptide design process by MD simulation.
A. TIM3 bound to the antibody (Ab) M6903 (only the variable Ab domain is shown). TIM3 is in blue, the Ab heavy domain (HC) is in gray, the Ab light chain (LC) is in yellow; the Ab residues that are within 7\AA~of a TIM3 atom are in red, and the CDR2 fragment of the HC (res.~49-60) is in black; B. The first design obtained by connecting Ab fragments proximal to TIM3, as described in the text; the LC and HC fragments are in yellow and red respectively, and a CDR2 fragment of the HC (res. 54-60) is black; C. Ensemble of conformations of design #1 from the first 20ns of MD simulation; colors are as in B; D. Root-mean-square distances (RMSDs) between MD simulation structures and the corresponding initial structures for all designs. Three bars correspond to each design; the average RMSD over the MD trajectory length (indicated in ns above the white bar) is in black, the highest RMSD is in white, and the average RMSD with two standard deviations (2STD) is in gray. The designs are ordered from left to right in increasing RMSD+2STD. E. Initial structure of design #13 bound to TIM3; F. 10 conformations of design #13 taken from the 1 microsecond MD simulation in 100ns time intervals.
Reference:
S. A. Abdel-Rahman, V. Ovchinnikov, and M. T. Gabr. Structure-based rational design of constrained peptides as TIM-3 inhibitors. Med. Chem. Lett., 15(6):806–813, 2024.