Parasites for therapeutics?

Using Toxoplasma as a vehicle for therapeutic delivery to the brain (Bracha et al, Nature Microbiology 2024)

This is one of those papers where the tech both sounds crazy and I can’t believe no one tried it earlier.

See for yourself:

This is conceptually very cool, but the important question is: (assuming one can overcome the challenges of using a live parasite in humans) is it even something we need? After all, we already have AAV9, a viral vector that delivers DNA cargo to the central nervous system & is used in Novartis’ Zolgensma.

Let’s compare them on a few important features:

Brain distribution

AAV9 can cross the blood brain barrier, but its efficiency of brain delivery is often lackluster with systemic injection. See one example in mouse:

from Hudry, E et al, Gene Therapy (2016)

Compared to the distribution observed in the Toxoplasma gondii study:

from Bracha S et al, Nat Microbiol (2024)

T. gondii might perform a bit better here, but it’s hard to say¹. And as new & improved AAV9 variants are developed, the viral capsid may soon win handily.

Protein v. DNA

A key difference between engineered T. gondii and AAV9 is that they deliver different things: T. gondii secretes protein, AAV9 delivers DNA.

Delivery of DNA is great for gene therapy - it can express a therapeutic protein for years as an episome, and potentially indefinitely if genomically integrated.

So why deliver protein itself? Well, for one, it won’t stick around very long, which enables more control over the levels of therapeutic protein at any given time. That could be useful for diseases of an episodic nature, or developmental diseases with a critical window for treatment.

That said, I can’t think of too many diseases where short-term protein presence is preferred. Rett syndrome, which the authors of this paper suggest as a therapeutic application, could make sense, though an AAV9-based gene therapy for this disease is already in trials.

Carrying capacity

T. gondii, being >100x the diameter of AAV, can hold bigger cargo. While AAV maxes out at a ~4.5kb DNA payload, you can theoretically pack a protein of any size into our parasite. But whether T. gondii can programmably secrete a really big protein into host cells is another question. The authors test payloads of up to 110 kDa in the publication, a size which would also be deliverable via AAV.

There may not be much need for a bigger cargo ship, anyway: Few brain-localized therapeutic targets are larger than AAV’s carrying capacity. Neurexins, some of which are involved in autism and schizophrenia, could be one example, though no one is proposing a neurexin as a gene therapy target as far as I can tell. Further, large cargos like Prime editor can be delivered via AAV by splitting the protein across two capsids.

In summary…

Even if a Toxoplasma strain could be engineered to be lowly immunogenic - enough to re-administered many times - it’s not clear that it offers a major benefit for therapeutic purposes.

But if T. gondii could just act as a brain delivery vehicle for DNA… that’d be something to consider.

1. Images are not directly comparable - one uses antibody staining while the other is GFP, timepoints, dosage and image processing may differ, etc.