Currently, the best inoculation fraction for primate brains is about 2% of the neurons. This is done using injection into the cisterna magna (a small cerebrospinal fluid cavity at the base of the brain) or else intravenous injection. For some diseases this will help. For others, it will not be nearly enough. So, I think about how it might be possible to inoculate most or all of the large brains of humans.

Here’s an interesting paper that suggests a strategy for successfully dosing of neurons in the deep brain. CD898hc single chain antibodies could be made, and their epitope binding region isolated, That binding region could be tested on the surface of a virus capsid to use for targeting to the brain. Would need to test whether a virus capsid would cross the blood brain barrier as well as an antibodies do, but I’d give it decent odds. I think that HSV would be a good vector to work with. AAV is nice, but it’s quite small, and we need a lot more room in the vector if we are going to do base editing.

Here’s a rendering of a rhinovirus that gives some idea of what such an engineered virus might look like. Here, the F’ab region, the functional part of the antibody that has the epitope binding region is attached to the rhinovirus’s cell receptor binding sites. In an engineered virus, those little pigtails would be attached to the virus capsid, and the flattish binding region would be sticking out.

Cold virus with antibody Fabs attached to its binding sites.

https://www.rcsb.org/structure/1rvf – Fab complexed with human rhinovirus.
Now, this cold virus is smaller than the Herpese Simplex Virus (HSV) is. Rhinoviruses (cold virus) are about 30 nanometers in diameter. The inner capsid of herpes simplex virus is 125 nanometers, 183 times the volume of the most common gene therapy vectors. So, it’s about 4 times the diameter, of this example I’m using. The most commonly used virus, Adeno-associated virus (AAV), is 22 nanometers, even smaller than a rhinovirus. Adenovirus is 65-80 nanometers in size.

What is a Fab versus whole antibody?

A whole, idealized (because real antibodies flop around) human IgG1 antibody is shown below. The little molecule on the left is what the epitope binding region (hands of the Y) bind to. A and B are roughly the Fab regions. C is the heavy chain. This C region is what the immune system binds to. Almost all antibodies have two protein chains that form the binding region. But that makes it very hard to use molecular biology to express it. So, a special set are used from one set of animals that have single chain epitope binding regions.

A & B – These ends bind to antigens.
C – The Fc chain. Trim21 binds to this.

Below is a model I tweaked to remove the light chain, leaving only the heavy chain part of the antibody. This is a good approximation of what the single chain antibodies are. As much of that binding region as needed can be used. It’s hard to predict exactly how long is optimum. I’d try shortest possible, cut it mid-site, and then at the base of the Y.