There are a lot of really great things you can accomplish with protein-based therapeutics that are hard to do by any other means. Monoclonal antibodies are a perfect example, and others include antibody-drug conjugates, enzyme replacement therapies, and things like CRISPR and some other gene-editing technologies, where you need to introduce exogenous proteins to do the cut-and-paste work for you in the genome. These things can be quite a different world from the small-molecule drug discovery areas that I’ve spent my career in, but the general concerns apply.
Dosing, for one thing, and pharmacokinetics in general. Oral dosing for most of these things is out of the question, since you’re tossing your carefully engineered protein constructs into a destructive bath of enzymes that are specifically designed to rip everything to shreds in the stomach and intestines. So you’ll be dosing these by injection, but what kind of injection? Subcutaneous? Intramuscular? Intravenous? They each have their particular character, with advantages and disadvantages for each, but at least they all avoid the digestive system. I.v. is going to hit the hardest and fastest, other things being equal, but it has to really be done in a medical setting, while patients can administer their own s.c. doses, if they’re OK with the idea. But that’ll have a different blood-level profile, often a more spread-out one, which might not be a bad thing. But once these things are in the blood stream, how long do they last? There are enzymes in circulation that can do damage to your therapeutic proteins, so you have to watch the half-lives of your various candidates and adjust accordingly.
Monoclonal antibodies are at the far end of the spectrum. They take advantage of the body’s general handling of that protein class (which is to give them an extended pass for circulation), and many of them have circulating half-lives of weeks. That’s often really good news from a drug development standpoint, but it also means that you’d better be sure that there aren’t circumstances that could develop where you’d want your new therapy to suddenly clear out of the way, because that’s not going to happen. You can just stop taking a twice-a-day oral medication, but once a month is another story. There have been several ingenious ideas for built-in “off switches” for long-half-life protein therapies, but I don’t think any of these have reached human usage yet.
One of the biggest worries, though, is immunogenicity. We humans demonstrate constantly that there are proteins and protein derivatives that can raise strong immune responses, and the worrisome part is that this is extremely variable, with an extraordinarily long tail of unusual responders. The examples of rare life-threatening allergies to bee stings or peanuts are well known, but both small molecule drugs and engineered proteins can potentially do the same thing. A lot of low-level but potentially severe side effects fall into this category, for sure. But you have to worry about both ends of this scale, too – the few patients who might have a terrible reaction, and the many who might have enough of a reaction to just sort of neutralize your protein in general. (In case you’re wondering, you can indeed develop antibodies to antibodies, and this can limit the effectiveness of otherwise very useful monoclonals).
That’s illustrated by this new paper, which looks at a candidate enzyme for CRISPR therapy. RfxCas13d is a particular bacterial protein that’s of interest because it’s quite small and very selective, making it a good candidate for therapeutic use. But these authors report that a large number of people may already have an immune response to it. That’s already been demonstrated with earlier CRISPR ideas such as Cas9 enzymes from common bacteria like Staphylococcus aureus or Streptococcus pyogenes, species that every human being on Earth has been exposed to countless times. RfxCas13d is from Ruminococcus flavefaciens though, which as the name indicates is a species from the guts of ruminant mammals, where it helps digest cellulose. Unfortunately, testing in a panel of human plasma samples showed similar antibody and T-cell responses as compared to those other bacterial Cas enzymes.
This wasn’t an expected result – R. flavefaciens almost never colonizes humans, and the hope was that we just wouldn’t recognize it. From reading the paper, it doesn’t seem to be clear yet if this is part of a general recognition of Cas enzymes by our immune system, or whether we humans are just more previously exposed to this one than you’d have thought. The authors recommend careful evaluation of new CRISPR enzyme candidates, further work on decreasing the immunogenicity of their sequences, and caution in advancing RfxCas13d in particular. You would not want to set off an inflammatory response on top of gene therapy, and that’s of particular concern with ideas about expressing the CRISPR system over a longer term. Will we end up using immunosuppressants as part of CRISPR therapy in general? Too early to say, but I’m glad that these issues are being brought to light.
