Phase Separation in Oncology

This section covers the phase-separation interpretation linked to melatonin in oncology.

The idea comes from the growing literature on biomolecular condensates.

These are liquid-like, membraneless compartments that cells form to concentrate proteins, RNA, and signalling machinery.

The basic picture

Cells can create temporary droplets that act like pop-up workstations.

They help organise transcription, stress responses, RNA handling, and signalling without building a membrane-bound organelle.

In cancer, that flexibility can become a survival advantage.

These droplets are often described as biomolecular condensates.

They are usually liquid-like.

They can fuse, exchange components, and change material state over time.

They do not need a membrane to act like a compartment.

That is what makes them so useful to both healthy cells and tumours.

Why cells use condensates at all

Condensates help cells do three things quickly.

• Concentrate the right molecules in one place

• Exclude molecules that would interfere

• Reverse the structure when the job is done

Examples include stress granules, P-bodies, transcriptional hubs, and DNA-damage response compartments.

What goes wrong in cancer

Condensates that should be temporary can become persistent.

They can also become more gel-like or even more solid over time.

They can then stabilise:

• oncogenic transcription programs

• EMT and cell-state transitions

• stress adaptation under hypoxia, acidity, and therapy pressure

In this view, cancer cells build biophysical sanctuaries that protect malignant programmes.

A simple picture for the cancer problem

In a healthy cell, these structures appear when needed and then dissolve.

In a cancer cell, many of them stay active too long, become too stable, or harden into survival-supporting hubs.

That helps the tumour preserve growth programs, stress adaptation, and resistance behaviour under pressure.

The city-story version

One useful way to picture the paper is as a city under chronic siege.

• Axis I behaves like command rooms that decide which genes stay on

• Axis II behaves like control rooms that drive state changes such as EMT and invasion

• Axis III behaves like bunkers that protect the cell during oxidative, metabolic, and immune stress

In a healthy city, those rooms appear briefly and then shut down.

In a cancer city, they stay active and become fortified.

The three condensate axes

Axis I — Nuclear decision-making condensates

These shape gene expression.

Examples include hubs involving MYC, TP53, EZH2, EP300, KDM1A, SOX9, and SMAD3.

Axis II — State-transition condensates

These shape EMT, plasticity, invasion, and therapy resistance.

Examples include YAP or TAZ, β-catenin, EGFR, TWIST1, vimentin, and NEMO.

Axis III — Stress-response condensates

These help cells survive oxidative, metabolic, and immune stress.

Examples include SQSTM1, KEAP1, BCL2, cGAS, TFEB, TFAM, and USP10.

The three levers melatonin is proposed to affect

Lever I — Redox tuning

Redox state changes how sticky condensate proteins are.

That can help droplets form, stay fluid, or dissolve.

This matters because tumours often maintain an abnormally reduced intracellular state that favours survival-oriented condensates.

Lever II — Multivalent scaffold plasticity

Many condensate proteins behave like weak molecular Velcro.

If melatonin changes their interactions, the scaffold may become less stable.

This includes intrinsically disordered regions and repeated interaction domains that create many weak contacts at once.

Lever III — Electrostatic and proton-balance recalibration

Condensates can create their own micro-environment.

Changes in charge balance, dielectric behaviour, or proton handling may alter how protected those droplets remain.

Why this is being linked to melatonin

Loh et al. 2026 proposed a 26-gene signature at the overlap of melatonin biology and phase-separation biology.

The main idea is that melatonin may not act only through one receptor or one pathway.

It may also act as a broader landscape-level regulator of the physical conditions that let oncogenic condensates persist.

Dose and timing context

The Loh et al. 2026 study is mechanistic and bioinformatic.

It does not prescribe a dose.

At physiological or low pharmacological doses, melatonin behaves mainly as a chronobiotic and systemic antioxidant.

At 10 to 50 mg daily, it begins to overlap the human oncology range where redox, signalling, and adjunctive anti-cancer claims are usually discussed.

Whether the condensate-disrupting effects described in the paper require that range, a higher pharmacological range, or something in between is not yet established in human studies.

Moderate-dose studies do show effects on several genes and pathways linked to the 26-gene signature, including EZH2, YAP or TAZ, and β-catenin.

Caution is warranted before inferring that the phase-separation model justifies very high doses.

That remains speculative.

Circadian timing may still matter, because condensate behaviour and mitochondrial signalling both vary across the day.

Chronotherapy angle

Timing is not a minor side issue here.

Several older animal studies reported that late-day or night-time melatonin inhibited tumour growth more reliably than morning dosing.

Some morning-dosing models were neutral or even unfavourable.

That does not create a fixed universal clock.

It does support the idea that condensates, mitochondrial signalling, and circadian phase should be thought about together.

Supporting evidence: melatonin modulates condensate-linked genes at moderate doses

The following studies demonstrate gene-level effects at doses within or close to the moderate clinical range — supporting the position that the condensate story is not exclusively a high-dose phenomenon.

Caveat for the document

Most of these studies use in vitro pharmacological concentrations, in the μM to low mM range, bathing cells directly.

The EZH2 and Hippo data in particular do not yet have a clean human oral-dose equivalence established.

The β-catenin and EMT data from in vivo mouse models are the strongest for inferring something meaningful happens below the RET-threshold dose range.

The condensate-disruption implication remains a reasonable inference, but it is still inference, not direct measurement of condensate behaviour.

Practical summary

If this model is right, melatonin may pressure tumours in three connected ways.

• It may push redox balance out of the tumour's safe zone.

• It may soften or destabilise condensate scaffolds.

• It may disrupt the protected micro-environment inside those condensates.

At moderate doses, that pressure may show up mostly through signalling and redox recalibration.

At higher pharmacological doses, the same framework may overlap more with mitochondrial stress and RET-related redox shock.

Related pages

• Melatonin in Oncology - Study Notes — the hub page that ties together the mitochondria, dosing, immune, phase-separation, and fibrosis sections

• Fibrotic Drivers and the 26-Gene Signature — how the condensate story overlaps with EMT, fibrosis, YAP/TAZ, β-catenin, and stromal remodelling

• High-Dose Mitochondria, RET, and ROS — the mechanistic RET, ROS, uncoupling, and apoptosis sequence behind the high-dose claim

• Moderate-Dose Immune Effects and Timing — the human oral adjunct literature, Th1 logic, and timing framework

• Dosing, Bioavailability, and Human Scaling — how mouse and cell data translate into estimated human exposure, route limits, and bioavailability caveats

Additional context from Doris Loh

Doris Loh, lead author of Multiaxial Biophysical Control of Oncogenic Phase Separation, now also shares public explainer videos for readers who want more context around the biophysical framework discussed here.

She describes the channel as an exploration of the evolutionary blueprints of life, bridging ancient molecular frameworks with modern thermodynamics and cellular biophysics.

The channel focuses in part on how molecules such as melatonin and ATP may help regulate biological order, resilience, and health across time.

https://www.youtube.com/watch?v=LWJpMfyEgJM

Watch all Doris Loh's YouTube videos

References

1. Loh et al. 2026 — Multiaxial Biophysical Control of Oncogenic Phase Separation

2. Biomolecular condensate background

3. Phase separation and cancer review

4. Melatonin sensitizes esophageal cancer cells to 5-fluorouracil via promotion of apoptosis by regulating EZH2 expression

5. Melatonin Inhibits Glioblastoma Stem-like Cells through Suppression of EZH2–NOTCH1 Signalling Axis

6. Melatonin and its derivative disrupt cancer stem-like phenotypes of NSCLC cells (β-catenin / Oct-4 / Nanog via AKT)

7. Melatonin and andrographolide synergize to inhibit the colospheroid phenotype by targeting Wnt/β-catenin signaling

8. Melatonin reverses nasopharyngeal carcinoma cisplatin resistance by inhibiting Wnt/β-catenin signalling

9. Melatonin mitigates the HDAC7/β-catenin/c-Myc positive feedback loop in esophageal squamous cell carcinoma

10. Melatonin and Hippo Pathway: Is There Existing Cross-Talk?

11. Melatonin in Cancer Treatment: Current Knowledge and Future Opportunities (VIM/EMT review)

12. Melatonin in Cancer Therapy: Lessons From 50 Years of In Vivo Research

13. Interactions of melatonin with various signalling pathways in cancer (broad review)