Here’s a new look at the various ways that small molecules can affect a well-known drug target (the estrogen receptor) and it shows us that we’re all going to have to look at these things more carefully than we do.
Now, to be fair, the ER is already fairly complicated, because it’s a nuclear receptor. Those things tend to have rather large binding sites for small-molecule ligands, which is why (in this case) there are so many ligands known, so many unrelated compounds that can have estrogen-like effects, and so on. And for most nuclear receptors, the downstream signaling is quite multivariate – they’re recruiting a whole suite of other proteins for their transcriptional effects, and the effects you get depend on which of these are in the complex. Which depends partly on what sort of ligand is in the binding site, and partly on what the local concentrations of those cofactor proteins are in the first place. This means that the same compound can have totally different effects in different tissues – indeed, this has been demonstrated with the estrogen receptor itself, where tamoxifen behaves completely differently in breast tissue versus the uterus. This is one of the reasons that I would like to climb in a time machine and whack the person who imported the terms “agonist” and “antagonist” into the nuclear receptor field with a board. They’re too simple, and they give people a misleading mental framework for dealing with NR signaling.
But wait, there’s more, as the informercials used to say. This new paper is looking at various ER ligands that have been used to block the receptor’s effects in hormone-dependent breast cancer (a big field of research, naturally). And you have your “antagonists” (grrr), which keep various estrogen-dependent transcriptional events from occurring, but you also have compounds that seem to throw the receptor into a less stable state where it’s more likely to be degraded. As with all protein degradation mechanisms, that opens up a whole new set of effects, because if the protein is totally removed, its other effects that don’t even depend on the ligand-binding site are also taken out, and these can be substantial.
The prototype of these “selective estrogen receptor degrader” compounds is fulvestrant, an unusual-looking molecule with a steroid head group and a long tail containing both a sulfoxide linker and a perfluoro anchor at the far end. It’s important to note that this is not really a “bifunctional degrader” in the sense that people are using the word, because the tail group is not explicitly a ligand for some ubiquitin ligase degradation machinery to be brought into play. This review will put it into context – fulvestrant (and the molecules being developed to follow up on it) have been proposed to work instead by inhibiting receptor dimerization, which causes them to be regarded by the cell’s protein maintenance machinery as defective and marked for degradation. (I should note that there are several different mechanisms proposed for the various “SERDs”, and they may well break down into two broad classes based on whether or not they’re steroid-derived).
And this new paper has more detail to add. It brings in a good deal of evidence to show that fulvestrant actually works most by impairing the mobility of the receptor protein. Remember, these nuclear receptors have to end up, with the cohort of cofactors and complex partners, down there in the nucleus on some DNA sequence or another. In fact, another lesson of this new work is that you can’t find fulvestrant-like molecules just by monitoring degradation of the receptor protein per se: that may be almost a side effect of the real mode of action, which is slowing its ability to get around the nucleus. And if you hadn’t considered that as a key mechanism, you are not alone.
But what this does is open up the possibility of transcription factor dynamics as a new place to find and optimize small-molecule drugs. It’s not going to be particularly easy to do that (you’ll probably need advanced imaging capabilities), but definitely not impossible. And given the subtle, fast, multivariate nature of all these transcriptional complexes in action, there may well be modes of action and mechanistic differences that we’ve never even realized were present. The more we know about such things, the better – targeting transcription in general is an area that needs all the help it can get. Onward!