Targeted protein degradation has been as hot a topic as you could want for the last several years in drug discovery, it’s no wonder: causing a protein to basically disappear from a living cell is a therapeutic mode different from anything we’ve ever investigated. Inhibiting a protein’s function is one thing – but proteins have more going on than just a single active site, which is what a small molecule will generally be able to block for you. All of the other partner interactions are suddenly left dangling by TPD, often to great effect.
And the idea has ramifications that go beyond the first applications. For example, it can be applied to antiviral therapy, which another topic that has received sporadic attention during the last couple of years, here and there, on and off. Probably the first example of this was from 2014, the targeting of the hepatitis B “X protein” that is known to be important in virally-induced liver cancer. That was a bit of an unusual one, in that it wasn’t the “traditional” sort with two small molecules separated by a linker. You know, the kind of protein degrader that Grandma used to make. In that HepB case, two protein domains were fused together, one to target the X protein and the other to mark it for degradation, and the whole thing was made more cell permeable by adding a poly-Arg peptide. It really did seem to work in HepG2 cells, though, causing the X protein to be degraded, and you would expect that to lead to effects both on Hep B infection and on the subsequent risk for hepatocarcinoma.
More recently there was a report of a degrader targeting a key protease of the hepatitis C virus. That’s one target of the drug cocktails for the disease, and in fact this degrader was made by taking such a drug (telaprevir) and attaching it to a ligand to recruit the cereblon-containing protein degradation complex. This did indeed clear the protease enzyme with the expected effects on Hep C infection in cells. Even more interestingly, the degrader molecule was less sensitive to mutations around the telaprevir binding site that are known to confer resistance to the drug: an A156S mutation that had been shown to reduce the covalent binding of the ketoamide “warhead” of the small molecule lowered the small molecule’s efficacy by 10x, but lowered the degrader’s activity by only 3x.
Then there’s this report, on everyone’s favorite coronavirus. It appeared in November of 2020, so the group involved was pretty fast off the mark! They came up with a 23-mer peptide that was targeted to the Spike protein’s receptor binding domain (RBD) by trimming down the binding surface of its endogenous partner (the human ACE2 protein) and attached that to TRIM21, whose use in this fashion was described back in 2010 for common-cold viruses (and blogged about here). That has subsequently been developed into the “Trim-Away” technique for degrading proteins. Now, for that to work in the usual manner it was proposed for (degrading endogenous proteins), you have to micro-inject an antibody for them, in cells that have been engineered to overexpress TRIM21 itself. That TRIM21 then binds to a conserved region on most antibodies, and a domain on it that remains exposed then targets the proteasome and everything gets chewed up. In this case, the team took that RBD-binding peptide and fused the Fc antibody domain to it (the part that TRIM21 binds to so well), so it’s already built to take advantage. They still had to co-express TRIM21 in the infected cells, but when they did that, the proof-of-concept worked. The RBD protein was indeed degraded, and a pseudovirus infection assay showed 60% lower infection rates. That co-expression requirement makes this more of a lab curiosity than any sort of clinical tool, though. Trim-away in general is a really interesting discovery tool, but it’s not really a treatment modality.
There’s new work that takes a different tack. The authors are trying to come up with better attenuated viruses for vaccines, and that (traditionally) has been something of an artisanal art form. You need a virus (ideally) that’s replication-competent (because you’re going to need to manufacture it in cells!), with as close to the same suite of antigen proteins (so as to set off an appropriate immune response) but which itself doesn’t cause any actual disease symptoms. That’s a tiny needle to thread! In this case, the team engineer a proteasome-targeting domain (PTD) into influenza virus – that PTD it contains a sequence that binds to VHL, which is an enzyme in a degradation complex that also has a small-molecule ligand and has been used many times in bifunctional small-molecule degrader experiments. They further engineered a cleavage site into this section which is vulnerable to tobacco etch virus (TEV) protease, with the plan that the virus could then be grown in cells that were built to express that exogenous protease.
The PTD is built onto a key influenza protein (M1). So in normal cells, this new version of the flu will not be replication-competent, because the virus will find one of its important proteins being hauled off to the proteasome for destruction. Indeed, the PTD-engineered virus was tens of thousands of times less able to replicate in normal cells as compared to wild-type virus, and in mice as well. In a ferret model it was merely several hundred times less competent, which is still pretty impressive. But the other viral proteins are not being degraded – they aren’t attached to a PTD, and they’re available to set off an antibody response.
So what you have, in the end, is a remarkably attenuated virus that still resembles the wild type in almost every single way. These hold out the promise of being much more effective as vaccines for that reason. Live-type attenuated viruses may not resemble the live pathogen enough, and “killed” virus preparations often have had their proteins altered too much by the heat or chemical treatment used in those cases. Keep in mind that the current vaccines only raise a single viral protein from the coronavirus (the Spike). And they still work quite well – but perhaps setting off a more varied immune response with what’s essentially a whole virus (albeit one that now can’t replicate) could do even better?
There’s a lot of work to do on this before it’s ready for prime time. The authors looked over several influenze viral proteins before selecting M1 as the one to add the PTD to, but you’ll need to tune this for both wiping out the replication competence in normal cells and for making them replication competent in the TEV protease expressing ones that will be used for production. And of course as you move to other viruses, all of this groundwork will need to be done again. But once you’ve got the system figured out, this could be a really fast and effective way to produce a whole new generation of antiviral vaccines. Let’s hope it works!
