Gene therapy: modelling for more effective screening


© Andreus / 123RF Banque d'images

One of the challenges of gene therapy is developing new vectors, molecules capable of transporting “therapeutic tools” to the heart of the cell. Dendrimers are branched macromolecules that could fulfil this vector role. A Franco-Swiss team has demonstrated that modelling can predict the differences observed in behaviour between several similar dendrimers, suggesting that a large number of tests could be conducted virtually, thus making it possible to avoid many laboratory and animal experiments. These studies, which appeared in Nanoscale, are the result of international collaboration including Institut Galien Paris-Sud (CNRS/Université Paris-Sud), Laboratoire de chimie de coordination [Laboratory of Co-ordination Chemistry] (CNRS), Université de Toulouse (UPS/INPT) and Laboratoire de chimie et de biochimiepharmacologiques et toxicologiques [Laboratory of Chemistry and Pharmacological and Toxicological Chemistry Laboratory] (CNRS/Université Paris Descartes).

Gene therapy, which is very promising, consists of sending a “messenger” to the heart of the cell which delivers information to replace a defective gene. This “messenger” can also inhibit the production of proteins responsible for diseases, as in the case of certain lung inflammations (asthma, chronic obstructive pulmonary disease) which could be curbed by a particular nucleic acid: “TNF-Alpha” small interfering RNA.

The difficulty is in transporting this RNA to the heart of the cell, which researchers at Institut Galien Paris-Sud (CNRS, Université Paris-Sud), Laboratoire de chimie de coordination (CNRS), Université de Toulouse (UPS/INPT) and Laboratoire de chimie et de biochimiepharmacologiques et toxicologiques (CNRS/Université Paris Descartes) used two dendrimers for. These arborescent macromolecules, which are positively charged on the surface, should capture negatively-charged RNA via electrostatic interactions, before entering the cell. However, an initial study conducted in 2017 in vitro on cells and in vivo on mice, demonstrated a marked difference in efficacy between two dendrimers which were however very similar.

In order to understand these differences in behaviour better, the researchers turned to molecular modelling, in partnership with Swiss researchers from the Istituto Dalle Molle di studi sull'Intelligenza Artificiale [Dalla Molle Institute of Artificial Intelligence] and Università della Svizzera italiana, in order to study interactions between the RNA and the two dendrimers. These results confirmed that one of the two dendrimers has a stronger affinity for the RNA and additionally showed that this dendrimer binds with two RNAs while the other RNA only captures one. This strong affinity explains and confirms the greater efficacy of one of the two therapies.

The molecular modelling demonstrates its robustness and its benefit. By predicting the dendrimer's degree of affinity for small interfering RNA, it could enable upstream screening of a whole bank of dendrimers to choose the most relevant scenarios, thus reducing the number of experiments to be conducted in vitro or in vivo.