Mobile Menu

AI used to develop new protein delivery system

Gene therapies are often touted as the future of healthcare, yet delivery of the relevant materials to human cells remains a significant challenge. Today, we’re taking a look at a new technology that combines a natural bacterial system with artificial intelligence and protein engineering in a potentially revolutionary new protein delivery approach.

Pioneering tech

The development of the system, details of which were published this March in Nature, was led by Broad Institute core faculty member Feng Zhang. The CRISPR pioneer is no stranger to ground-breaking inventions, also launching his own start-up company in February of this year. In this latest venture, Zhang and his team have engineered a natural bacterial system to deliver proteins into human cells.

The concept is based on the mechanism by which endosymbiotic bacteria interact with their host using the extracellular contractile injection system (eCIS). The eCIS is a 100-nanometre tube-like structure that can be used to inject and deliver proteins to host cells. Whilst the system has been seen to naturally interact with various non-human cells, such as from mice and insects, it does not appear to have the ability to deliver to human hosts.

Figure 1: Image describing the basic concept of the eCIS. Adapted from Kreitz et al., 2023.

Manipulating the system

To combat the above problem, the team used artificial intelligence tech AlphaFold to manipulate the eCIS structure to allow it to bind to human cells. A radical technology in its own right, AlphaFold operates using a deep learning model to predict protein structures based on amino acid sequence.

After successfully engineering the eCIS to bind to the cells and deliver toxins in vitro, the team then investigated if the material transported by the system could be altered. Further tweaks to the complex facilitated the transport of tailored proteins of the scientists’ choosing, suggesting that the system could be used to successfully deliver protein-based drugs. This finding could be revolutionary for the future of gene therapies.

Where now?

So, is this modified bacterial system the ‘next CRISPR’?

Perhaps the main thing to consider is its potential to fight disease and help patients. For example, when the system was engineered to deliver toxins to leukaemia cells, it killed them with striking efficiency. And further work revealed that delivering modified proteins to EFGR-expressing cancer cells provoked an effective and extremely targeted response. With the potential to deliver gene-editing tools to cells, the system could transform the field of drug delivery and gene therapy as we know it.

Discussing the tech, Zhang stated, ‘Delivery of therapeutic molecules is a major bottleneck for medicine, and we will need a deep bench of options to get these powerful new therapies into the right cells in the body. By learning from how nature transports proteins, we were able to develop a new platform that can help address this gap.’

First author Jospeh Kreitz added, ‘This is a really beautiful example of how protein engineering can alter the biological activity of a natural system. I think it substantiates protein engineering as a useful tool in bioengineering and the development of new therapeutic systems.’

Zhang and his colleagues have also recently determined the structure of the R2 non-long-terminal-repeat retrotransposon – a type of mobile genetic element that uses target-primed reverse transcription to insert itself into a genome – as it initiates the reverse transcription process. This work also forms the basis of a potential new gene editing strategy. Commenting on Twitter, first author Max Wilkinson stated, ‘Maybe this could one day be really useful for gene therapy.’

It is clear that the future of gene therapy relies on the development of effective delivery models, and solving this bottleneck could lead to huge benefit for patients. This recent work highlights exciting advances in the field, so keep an eye out for more.

More on these topics

AI / Drug delivery / Gene Editing