Computationally guided high-throughput design of self-assembling drug nanoparticles.
Daniel Reker1,2,3, Yulia Rybakova1, Ameya R Kirtane1,2
1Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology, Cambridge, MA, USA.
Nature Nanotechnology
|March 26, 2021
Summary
Researchers developed a machine learning platform to rapidly identify self-assembling drug nanoparticles. This approach discovered 100 new nanoformulations with high drug-loading capacities, accelerating therapeutic development.
Area of Science:
- Pharmaceutical Sciences
- Nanotechnology
- Computational Chemistry
Background:
- Nanoformulations enhance drug delivery but often suffer from complex production and low drug loading.
- Co-assembled drug nanoparticles offer high drug-loading capacities (up to 95%) but identifying suitable combinations is challenging.
Purpose of the Study:
- To develop a rapid and scalable method for identifying self-assembling drug nanoparticles.
- To overcome the lack of understanding regarding small-molecule combinations that form effective nanoformulations.
Main Methods:
- Integration of machine learning (ML) with high-throughput experimentation (HTE).
- Screening of 2.1 million drug-excipient pairings (788 drugs, 2,686 excipients).
- Ex vivo and in vivo characterization of selected nanoparticles.
Main Results:
- Identification of 100 self-assembling drug nanoparticles from 2.1 million pairings.
- Successful characterization of sorafenib-glycyrrhizin and terbinafine-taurocholic acid nanoparticles.
- Demonstration of a platform for accelerated nanoformulation discovery.
Conclusions:
- The ML-HTE platform enables rapid, large-scale identification of high drug-loading nanoformulations.
- This approach can significantly accelerate the development of novel therapeutics.
- The findings pave the way for safer, more efficacious nanoformulations across various treatments.
