Biomolecular Condensates: A New Phase of Epilepsy Research?

Abstract

Biomolecular condensates are membrane-less compartments that are found in all different types of cells, including neurons. They form by a process called “liquid-liquid phase separation” which happens when it becomes more energetically favorable for macromolecules to cluster, rather than diffuse freely while surrounded by their water-shells. Most research on biomolecular condensates has focused on their role in cellular stress and regulating protein expression, but we are now recognizing their importance in neuronal processes such as synaptic transmission and neurodegeneration. To date, however, little consideration has been given to any role in epileptic pathology. Here we outline what is known about biomolecular condensates, how they form, how they might influence cellular physiology, and how these processes might relate to various open questions in epilepsy research. In particular, we draw attention to their possible role in the cellular regulatory effects of mTOR and how this impacts the control of ionic distribution. We explain how this affects circadian modulation of cortical function, including the susceptibility to seizures.

Advances in fundamental cellular physiology tend to progress through research on simple cell systems. Neurons, on the other hand, are complex beasts, and are not typically the preferred experimental preparation for investigating general cell biology. Research on neurons usually focuses upon phenomena that set them apart: their extended and highly polarized morphology, how they communicate with other neurons, how this generates emergent behavior across populations of neurons, and of course, their excitability, which allows them to flip functional states within an instant. Consequently, neuroscientists are sometimes slower to pick up other interesting research trends in cellular biology. Our sense is that this is very much the case for biomolecular condensates, a truly fascinating and evolving research topic that has yet to really impact upon neuroscientists’ consciousness, judging by how little has been written about it in a neuroscientific context. A PubMed search for “biomolecular condensates neuron” yielded just 85 citations (Sept 2025); for “biomolecular condensates epilepsy,” none. Our aim is to redress this deficit, to explain what is known currently about biomolecular condensates and why this should interest epilepsy researchers.

The recently coined term “biomolecular condensates” ¹ refers to various membraneless compartments that are found seemingly across all biology. Perhaps the most important, in terms of driving this field, are stress granules which form in many cells in response to different noxious stimuli, and P bodies (processing bodies) which are complexes of mRNA and protein that are involved in translational repression.^2,3 Other examples include nucleoli and, in neurons, the postsynaptic density and synaptic release zones, ⁴ to which we will return.

The formation of biomolecular condensates can be understood in terms of a liquid-liquid phase separation, rather like how vinaigrettes separate into water and fat layers. The critical pre-requisite for this to happen is the existence within the cytosol of multivalent macromolecules (multivalent refers to multiple points of attractive interaction), most commonly, proteins or nucleic acids. As their concentration rises, a threshold level is passed above which the attractive interactions between the macromolecules causes a sudden switch from the homogeneous solution in which the molecules are surrounded by a water shell, to a more energetically favorable arrangement in which dense clumps of macromolecules (“condensates”) form, floating within the cytosol (note that concentrating macromolecules within condensates means that within the cytosolic compartment their concentration is diluted). The threshold for this phase separation is influenced by temperature, pressure, the presence of other particles (including other macromolecules and ions), pH (which affects the dissociation of -COOH and NH₂ groups), phosphorylation and other post-translational modifications (eg, sumoylation) that alter attractive forces between molecules, to give a few instances. One can easily see how the propensity to form condensates within cells might thus be dictated by a myriad of systemic factors, both physiological and pathological, such as fever or hypothermia, elevated intracranial pressure, changes in osmotic tension either inside or outside the cell, anoxia, and others. Any of these can trigger rapid reorganization of condensates, creating, dissolving, or fusing them to change their size and number. While this is governed by the laws of physics, we are starting to recognize how evolutionary forces might also be at play. Certain protein sequences, termed “intrinsically disordered regions” (IDRs, also sometimes called “low-complexity regions”), are flexible sequences within proteins that allow them to change shape more easily, and this appears to facilitate condensate formation. ⁵ Their prevalence throughout the genome speaks to an important role and a strong evolutionary selective pressure for these sequences.

It is noteworthy that prion proteins have many such aggregation-prone stretches of amino acids.^5,6 Prion proteins are particularly concentrated at the synapses, ⁷ where the cycling back and forth between condensate and solute has been suggested, recently, to play a key role in synaptic vesicle organization ⁴ coordinated by the presynaptic protein, synapsin, which also spontaneously forms phase-separated condensates. These condensates capture small lipid vesicles, but disassemble upon phosphorylation by calcium/calmodulin-dependent protein kinase II. ⁴ Interestingly, α-synuclein reduces the likelihood of synapsin condensate formation.^8,9 Tau protein, a highly phosphorylated protein with IDRs which is associated with multiple neurodegenerative disorders, also forms presynaptic condensates which regulate vesicle trafficking ¹⁰ ; notably, mutations that alter its ability to undergo phase separation are known to alter synaptic function. ¹¹ All these results point to an important role for biomolecular condensates in synaptic transmission, but one which may also predispose toward neurodegenerative changes.^12,13

Research into biomolecular condensates is a very young field; consequently, our understanding of their functional significance is somewhat patchy, and much remains a matter of speculation. What has been established is that condensates act as microcompartments within cells, concentrating substrates and facilitating biochemical reactions. Serine residues are frequently present within IDRs, leading to hypothesized roles in cellular signaling. For instance, if serines are unphosphorylated at the time a condensate forms, then phosphorylation within the condensate would alter the domain's charge, changing the macromolecular attractive forces, and leading to expulsion of the newly phosphorylated macromolecule from the condensate and back into the dilute phase. ¹⁴ This is exemplified in the response to hypertonic stress, which triggers crowding-induced phase-separation of the WNK (“with no lysine” (K)] protein) kinases, forming biomolecular condensates with SPAK (SPS1-related proline/alanine-rich kinase) and OSR1 (oxidative stress-responsive kinase 1), ¹⁵ facilitating phosphorylation of those 2 molecules which in turn regulate the activity of the 2 cation-chloride co-transporters, NKCC1 and KCC2, also through phosphorylation. ¹⁶

Various excellent review articles discuss at length the role of biomolecular condensates in creating physical and temporal microcompartments to facilitate different biochemical interactions^1,14,17,18; far less consideration has been made of how they might affect the electrotonic properties of neurons and their excitability. Dai et al ¹⁹ point out that if condensates comprise unbalanced cationic and anionic elements, then the principle of charge neutrality will lead to redistribution of other ionic species (viz Na⁺, K⁺, and Cl⁻), creating a potential across the liquid-liquid phase separation, potentially leading to local intracellular potential anisotropy. Ionic redistribution across this phase separation will alter ionic concentrations within the surrounding cytosol, with secondary effects upon the plasmalemmal reversal potentials and also the cytosolic resistivity. The presence of biomolecular condensates will also change the volume of the effective conducting media within the cell. If this is not offset by other changes (eg, by an increase in the conducting ion concentration), then the electrotonic length of neuronal compartments will be increased, which would reduce neuronal excitability in various ways, by reducing synaptic summation, and also making saltatory propagation between excitable zones less likely (between nodes of Ranvier, or between the somatic and dendritic excitable areas^20,21). While we may be confident in our assertion that biomolecular condensation will affect excitability and the electrotonic properties of neurons, the precise effects are, however, very difficult to intuit, because it is possible to envisage condensates that could change local cytosolic ionic concentrations in different directions, ¹⁹ meaning that almost any effect on neuronal behavior is possible. Moreover, systemic effects may impact differentially on cell classes, and the emergent network behavior is impossible to predict. In short, there exist rich pickings for research in this area.

mTOR: Linking Condensates and Pathophysiology?

Central to the topic in hand is mTOR (mammalian target of rapamicin) function, and why “mTOR-opathies” are commonly associated with epilepsy and general neuropathology. mTOR is often termed a master regulator of cellular function, which should alert readers to the likelihood that it may influence pathophysiology in multiple different ways. We want to focus here, however, upon recent studies that indicate a fundamental role to modulate ionic distribution patterns during daily rhythms, which in turn affects network excitability.^22–24

When an animal feeds, following a period of fasting, mTOR is strongly activated. ²⁵ Although mTOR activity is also affected by many other factors (amino acid availability, adenosine tri-phosphate (ATP) availability, extracellular growth factors and mitogens, WNT (“Wingless and Int-1”) signaling, DNA damage and hypoxia), the dominant effect is a large diurnal modulation of mTOR activity, typically being high during the animal's active phase; in mice, this is during the dark phase. In turn, mTOR coordinates protein degradation and synthesis, which may vary as much as 2-fold across the day, and this leads directly to a large mTOR-dependent fluctuation in the concentration of soluble proteins, in phase with the mTOR activity. ²² This molecular crowding during an animal's active phase constitutes a very large osmotic pressure, but this was discovered to be counter-balanced by a substantial redistribution of ions mediated by electroneutral cation-chloride co-transporters in many different cell types, from single cell organisms up to cardiomyocytes. ²² When soluble protein concentration is high, Na⁺, K⁺, and Cl⁻ are all shunted out of cells; when protein concentration is low, these ions redistribute back into the cytoplasm. In this way, the cell volume is kept stable by counter-balancing the osmotic impact of a fluctuating macromolecular concentration with ionic redistribution. ²² This has been termed crowding associated compensatory transport of ions (CACTI). ¹⁷ In nonexcitable cells, osmoregulation thus seems to be prioritized over stability of ionic reversal potentials; balancing osmotic challenges using ionic redistribution seems like a logical solution in these cells, but one might think that the priorities would be different in neurons. Indeed, neuroscience is so in thrall to Nernst that it can come as a surprise to think that the baseline distribution of these ion species might vary at all (they are known to be affected by neuronal activity, of course, which dissipates the various ionic gradients). However, observations by 2 other, independent groups^24,26 confirm this is the case, reporting that baseline intracellular [Cl⁻] in mouse neocortical pyramidal cells almost doubles between day and night, equating approximately to a 20 mV shift in the GABAergic reversal potential. ²⁴ This physiological modulation of E_GABA appears at least as large, if not larger, than has been reported to contribute to ictogenesis.^27–29

The relative disinhibition of pyramidal cells during the active phase is reflected in altered physiological and pathological network excitability, ²⁴ although perhaps not to the extent that might be expected given the size of the Cl⁻ redistribution. The implication is that other controls on neuronal excitability can pick up the slack at times during the daily cycle when synaptic inhibition is weak, like shift workers providing round-the-clock protection. Intriguingly, in parvalbumin-positive interneurons, intracellular [Cl⁻] varies out of phase (high during daytime, and low at night) with the pyramidal cell modulation (low in daytime, high at night ²³ ); the 2 neuronal populations are relatively disinhibited at different times of the day. Given that both populations are exposed to broadly the same synaptic drive,^30,31 this suggests that activity-dependent changes probably do not underlie the chloride redistribution. An alternative explanation offered by this new CACTI hypothesis is that cell-class specific differences in [Cl⁻]_in modulation may arise instead from differences in metabolic controls such as mTOR.

It is interesting to consider these mTOR-dependent changes in ionic distribution in the context of the substantial body of evidence that points to raised [Cl⁻]_in as an important contributory factor in ictogenesis^{24,27–29,32,33} (see also reviews^34–36). Pursuing this line of reasoning, might this also suggest that cells with dysregulated mTOR will therefore have secondary dysregulation of ionic distribution, with consequent effects upon neuronal excitability? Mutations affecting multiple genes within the larger mTOR signaling pathway are associated with severe epilepsy disorders, many of them drug resistant (see recent reviews^37,38). The mTOR pathway is also implicated in glioma-related epilepsy.^39,40 Interestingly, epilepsy-associated mTOR mutations typically increase mTOR activity, ³⁷ leading to cellular hypertrophy and raised protein concentration, while mTOR activation has been shown to directly induce biomolecular condensate formation in single cell models. ⁴¹

Another important aspect of phase separation is that biomolecular condensates form rapidly, and just as quickly may reverse. Given that many of the predisposing conditions are finely balanced, phase separation confers powerful homeostatic properties (even, in certain cases allowing complete desiccation), that could be pushed and pulled in different directions depending upon what needs prioritizing, or what is perturbed, within the system. For instance, Stangherlin et al ²² demonstrated an interesting reciprocal relationship between intracellular ion concentration and macromolecular content, whereby disruption of mTOR influences the regulation of the SLC12A co-transporter family (which includes the sodium-potassium-chloride co-transporter, NKCC1, and the potassium-chloride co-transporter, KCC2), and conversely, perturbations of co-transporter activity affect the activation of mTORC1 by other triggers.

Significance to Epilepsy

With these caveats regarding predictions of what role biomolecular condensates have on neuronal excitability, it is still possible to suggest some interesting lines of investigation, several of which we have already mentioned. For instance, given the apparent importance of raised [Cl⁻]_in in ictogenesis, we need to reconcile this with observations that there appears an even larger physiological modulation of [Cl⁻]_in in different neuronal classes through the daily cycle.^23,24 Clearly this will contribute to the fundamental differences in brain activity and function at different times of day, as well as the well-established circadian fluctuations in many neurological conditions, notably in seizure susceptibility. ⁴² An important research question regards how can the brain tolerate raised [Cl⁻]_in without tipping into a seizure in some circumstances, but not others; how is neuronal excitability restrained at those times, but not others? This information would guide what anti-epileptic treatments should be given at different times, or for epileptic conditions with different circadian patterns.

The out-of-phase modulation of [Cl⁻]_in in different neuronal populations may be pertinent: the time of day when pyramidal cells have high [Cl⁻]_in is also when parvalbumin interneurons are likely to fire less because they have low [Cl⁻]_in and thus are restrained by synaptic inhibition. ²³ In contrast, certain ictogenic perturbations, such as raising extracellular [K⁺] will cause [Cl⁻]_in to rise in both cell classes together. Similarly, network activity dependent changes in [Cl⁻]_in are likely to be felt in all neuronal classes in tandem.

We might further seek to understand whether condensates play any role in sudden brain state transitions, such as seizure initiation and termination, and physiological state transitions too. For instance, a novel experimental manipulation involving the persistent strong activation of the optogenetic chloride pump, halorhodopsin, led to a substantial ionic redistribution of both Cl⁻ and K⁺ ions into neurons, and triggered cortical spreading depression events. ⁴³ Similarly, seizures are also associated with osmotic shifts from extra- to intracellular compartments, leading us to propose that the cellular response to this may link seizure termination and spreading depression. ⁴⁴ The link between these observed patterns of ionic redistribution and the associated state changes remains unclear, but may involve cytosolic changes including biomolecular condensate formation, as has been seen in other cell classes. Another point to consider is that if biomolecular condensate formation and dissolution is a key factor in brain state transitions, then it may be possible to predict when such transitions will happen by assaying the system in new ways.

In conclusion, research into biomolecular condensates and cytosol reorganization has focused mainly on biochemical modulation, but how they influence neuronal behavior has barely been studied. Our intention in this review is to motivate research in this field; while some of the links are somewhat speculative, we suggest that the Venn diagram of epilepsy and biomolecular condensates shows rather more overlap than has registered yet on Pubmed.