HA-fragment–driven protein condensates as a proteostatic brake on fibroblast senescence

Dr. Ines Lima· Mar 26, 2026

Mechanism: In aged fibroblasts, LMW-HA fragments trigger protein condensate formation that sequesters inflammatory signaling proteins, acting as a proteostatic brake. Readout: Readout: When condensate capacity is overwhelmed, NF-κB activates, leading to high inflammatory cytokine release and senescence markers; enhancing condensate capacity reverses this, reducing inflammation and senescence.

Hypothesis

Chronic accumulation of low‑molecular‑weight hyaluronan (LMW‑HA) fragments in aged dermis triggers liquid‑liquid phase separation (LLPS) of HA‑binding adaptor proteins (CD44, TLR4‑MyD88, TRAF6) and downstream NF‑κB components, forming amyloid‑like condensates that sequester these signaling hubs. This condensate formation represents a proteostatic containment strategy that limits inflammatory signaling; when condensate capacity is exceeded, fibroblasts transition to a senescent state.

Mechanistic rationale

LMW‑HA acts as a DAMP that engages CD44 and TLR4, recruiting adaptor proteins that possess intrinsically disordered regions and multivalent interaction domains.2
Multivalent HA‑protein interactions promote LLPS, analogous to how RNA drives stress granule formation. The resulting condensates mature into β‑sheet‑rich aggregates detectable by Thioflavin T and filter‑trap assays.
Sequestration of CD44‑TLR4 adaptors within condensates attenuates NF‑κB nuclear translocation, reducing IL‑6, TNF‑α, and IL‑1β output.
When HA‑fragment load overwhelms condensate buffering capacity (e.g., due to severe HAS2 loss and ROS‑mediated HA cleavage), excess signaling escapes condensates, activates NF‑κB, and drives ER‑stress‑linked senescence pathways.3
This model explains the bimodal inflammatory phenotype observed in aged fibroblasts: low‑grade inflammation in cells with robust condensate formation versus deep senescence in those that fail to sequester HA‑signal complexes.

Predictions

Fibroblasts from elderly donors will show higher levels of HA‑dependent protein condensates (Thioflavin T⁺ puncta) correlating with LMW‑HA accumulation and inversely with HAS2 expression.
Disruption of condensate formation (by 1,6‑hexanediol or HSP70 overexpression) will increase NF‑κB activity and cytokine release without affecting HA fragment levels.
Enhancing condensate capacity (via overexpression of an aggregation‑prone, inert scaffold such as poly‑Q‑GFP) will reduce inflammatory signaling and delay senescence markers (p16^INK4a, SA‑β‑gal) despite persistent LMW‑HA.
Conversely, pharmacological inhibition of condensate maturation (e.g., with EGCG) will accelerate senescence in young fibroblasts exposed to exogenous LMW‑HA.

Experimental approach

Condensate detection: Isolate nuclear‑cytoplasmic fractions from young (0‑19 yr) and aged (≥70 yr) human dermal fibroblasts; perform filter‑trap assay and Thioflavin T fluorescence to quantify amyloid‑like aggregates. Co‑stain for CD44, MyD88, and p65 to assess sequestration.
Manipulation: Treat cells with 1,6‑hexanediol (LLPS disruptor), overexpress HSP70 (chaperone that prevents aggregation), or express poly‑Q‑GFP (synthetic condensate promoter). Measure NF‑κB luciferase activity, cytokine ELISA, and senescence readouts.
Rescue experiments: Add exogenous LMW‑HA to young fibroblasts ± condensate modulators; assess whether enhancing condensate formation protects against inflammation‑induced senescence.
Controls: Use HA‑size‑specific antibodies to verify fragment molecular weight; confirm that HAS2 knockdown recapitulates aged phenotype.

Potential outcomes and falsification

If condensate levels do not correlate with donor age or HA fragment size, or if disrupting condensates fails to alter NF‑κB signaling, the hypothesis would be falsified.
If enhancing condensate formation does not mitigate inflammation or senescence despite reduced NF‑κB activity, alternative mechanisms (e.g., direct transcriptional repression) would be implicated.
Conversely, observing a tight inverse relationship between condensate burden and inflammatory output, with manipulation of condensate properties predictably shifting fibroblasts between low‑grade inflammation and senescence, would support the model.

This framework translates the seed idea—that protein aggregation is a last‑ditch ordering effort—into a specific, testable mechanism linking extracellular matrix deterioration to intracellular proteostatic responses in skin aging.

Hypothesis

Mechanistic rationale

LMW‑HA acts as a DAMP that engages CD44 and TLR4, recruiting adaptor proteins that possess intrinsically disordered regions and multivalent interaction domains.2

Multivalent HA‑protein interactions promote LLPS, analogous to how RNA drives stress granule formation. The resulting condensates mature into β‑sheet‑rich aggregates detectable by Thioflavin T and filter‑trap assays.

Sequestration of CD44‑TLR4 adaptors within condensates attenuates NF‑κB nuclear translocation, reducing IL‑6, TNF‑α, and IL‑1β output.

When HA‑fragment load overwhelms condensate buffering capacity (e.g., due to severe HAS2 loss and ROS‑mediated HA cleavage), excess signaling escapes condensates, activates NF‑κB, and drives ER‑stress‑linked senescence pathways.3

This model explains the bimodal inflammatory phenotype observed in aged fibroblasts: low‑grade inflammation in cells with robust condensate formation versus deep senescence in those that fail to sequester HA‑signal complexes.

Predictions

Fibroblasts from elderly donors will show higher levels of HA‑dependent protein condensates (Thioflavin T⁺ puncta) correlating with LMW‑HA accumulation and inversely with HAS2 expression.

Disruption of condensate formation (by 1,6‑hexanediol or HSP70 overexpression) will increase NF‑κB activity and cytokine release without affecting HA fragment levels.

Enhancing condensate capacity (via overexpression of an aggregation‑prone, inert scaffold such as poly‑Q‑GFP) will reduce inflammatory signaling and delay senescence markers (p16^INK4a, SA‑β‑gal) despite persistent LMW‑HA.

Conversely, pharmacological inhibition of condensate maturation (e.g., with EGCG) will accelerate senescence in young fibroblasts exposed to exogenous LMW‑HA.

Experimental approach

Condensate detection: Isolate nuclear‑cytoplasmic fractions from young (0‑19 yr) and aged (≥70 yr) human dermal fibroblasts; perform filter‑trap assay and Thioflavin T fluorescence to quantify amyloid‑like aggregates. Co‑stain for CD44, MyD88, and p65 to assess sequestration.

Manipulation: Treat cells with 1,6‑hexanediol (LLPS disruptor), overexpress HSP70 (chaperone that prevents aggregation), or express poly‑Q‑GFP (synthetic condensate promoter). Measure NF‑κB luciferase activity, cytokine ELISA, and senescence readouts.

Rescue experiments: Add exogenous LMW‑HA to young fibroblasts ± condensate modulators; assess whether enhancing condensate formation protects against inflammation‑induced senescence.

Controls: Use HA‑size‑specific antibodies to verify fragment molecular weight; confirm that HAS2 knockdown recapitulates aged phenotype.

Potential outcomes and falsification

If condensate levels do not correlate with donor age or HA fragment size, or if disrupting condensates fails to alter NF‑κB signaling, the hypothesis would be falsified.

If enhancing condensate formation does not mitigate inflammation or senescence despite reduced NF‑κB activity, alternative mechanisms (e.g., direct transcriptional repression) would be implicated.

Conversely, observing a tight inverse relationship between condensate burden and inflammatory output, with manipulation of condensate properties predictably shifting fibroblasts between low‑grade inflammation and senescence, would support the model.