I’ve written several times here about phase-separated condensates in cells, but now comes a rarity: a paper with some evidence for a therapeutic application. Everyone in the field has been thinking along such lines, naturally, but this is the first small-molecule screen that I’ve seen that tries to tie modifying condensate behavior in that way to a disease state.

In this case, it’s the handling of stress granules in the context of ALS. Stress granules have been the focus of a lot of work in the condensate field, which is natural enough. They’d been recognized for many years in cell biology as structures that appeared and disappeared in response to heat stress, hypoxia, and so on, and recently their liquid-liquid phase-separated nature has been the focus. Like many condensates, stress granules seem to be a mixture of proteins and nuclei acids, with various RNA-binding proteins prominent in their composition. The connection to ALS is via mutations in some RNA-binding proteins (such as FUS and TDP-43_ that have shown up in ALS patients, and it’s believed that these mutants can either allow for more stress granule formation or cause them to be abnormally persistent. Protein deposition is, of course, a hallmark of the disease. It’s been observed in many condensate systems (both in vitro and in vivo) that some phases that start out as liquid droplets can “harden” over time, becoming more gel-like and harder to reverse, and even progressing to insoluble aggregates. It’s quite tempting, then, to see aberrant condensate formation (or aberrant re-formation of a single phase) as related to such disease processes.

This new paper is a multicenter effort with Tony Hyman of Dresden as corresponding author, and it’s worth noting that he’s on the founding team at Dewpoint Therapeutics, who are trying to turn discoveries in condensate biology into actual therapeutics. Now, there are compounds known that seem to modulate condensate behavior, but they’re not too drug-like. 1,6-hexanediol shows up the literature quite often as a general “condensate dissolver”, even to the point of treating living cells, but that’s not a feasible path in a whole animal from a pharmacokinetic and toxicological standpoint. You also would probably not want to just go in and dissolve every condensate in the cell simultaneously! In this case, the team screening 1600 compounds (mostly known bioactives and drugs) to see which ones affected FUS condensate behavior, but did not act as “universal solvents”.

Lipoamide and lipoic acid both came out of the screen as such compounds, testing them in HeLa cells that were engineered with a fluorescent FUS protein at roughly endogenous levels. (That’s important – sometimes to see these condensates people have had to crank up expression, which likely wouldn’t be appropriate for a screening effort). Under normal conditions, FUS is mostly in the nucleus, sometimes as part of yet another membraneless structure, the paraspeckles. But under stress, it exports back to the cytoplasm and helps to form stress granules. In this case, the stress was in the form of arsenate treatment, which messes up oxidative metabolism in general and is known to reliably bring on the stress-granule response. Lipoic acid was able to stop stress granule formation, but did not seem to affect the FUS-driven paraspeckles or other known condensate structures (such as the nucleolus itself).

The follow-up assays are quite interesting. Lipoic acid was shown to inhibit stress granule formation in living C. elegans roundworms, and to ameliorate motor defects in Drosophila flies that have been engineered to express human FUS protein instead of their own homolog. In fact, these flies had an even bigger problem; they were made to express mutant human FUS (P525L or R521C mutations), which are known to cause more severe motility problems. P525L is an already-characterized mutation in humans, and if you express it in iPSC-derived motor neurons, you get defects that are consistent with the clinical phenotype of the disease (among other things, they have a much greater tendency to form stress granules). Lipoic acid treatment prevented the axonal die-back seen in the untreated cells, and also restored axonal transport to wild-type levels, both of which were associated with a corresponding lack of stress granule pathology.

To the best of my knowledge, this is the first study that ties together condensates, small molecule treatment, and a disease state in this manner. Could lipoic acid itself be a treatment for ALS? The paper proposes that idea, but I’ll leave it up to the clinicians. You need pretty high concentrations (tens to hundreds of micromolar), and I’m just not sure if that’s clinically feasible. On the other hand, lipoic acid itself is pretty benign, and I would think that this falls into the “What do you have to lose” category. But you’d also think that you’d want to treat people as early in the disease state as possible, before more irreversible damage has had a chance to pile up – I don’t think lipoic acid (or any such compounds) is going to reverse that, but rather keep further damage from occurring. Drug discovery in this space is going to be quite an experience:

The necessary properties of a compound affecting protein phase separation are likely different to those of a canonical drug binding a well-defined structured protein site. Condensates are formed by protein liquid-liquid phase separation involving many transient interactions , unlike strong enzyme-substrate or protein-protein interactions typically targeted by drugs. Small molecules can interfere with phase separation and alter the properties of phase boundaries (surfactants). For instance, ATP has been identified as a hydrotrope which helps keep proteins soluble. The identification of lipoamide and lipoic acid suggests that further small molecules could be identified that target phase separated compartments.

And the hunt is on for them, especially now that this paper has come out. We’re going to be hearing a lot more about this sort of thing, so we’d better all get used to it. . .