Mechanical Memory: Do Cells Remember Their Matrix?

Team Dynamic Matrices | 12 June 2026

Every stem cell protocol assumes, usually without saying so, that a cell's behaviour is determined by its current environment. Put it on a soft gel and it responds to soft. The evidence of the last decade says that is only partly true. Cells carry the mechanical history of where they have been, and for MSCs expanded on plastic before an experiment, that history is a gigapascal-stiff surface. This post covers what mechanical memory is, how it works, and what it means for protocol design and reporting.

The defining experiment

Yang, Tibbitt, Basta and Anseth (Nature Materials, 2014) cultured human MSCs on stiff photodegradable PEG hydrogels, then softened the gel in situ with light. Cells that had been on the stiff gel for one day reverted: YAP left the nucleus, RUNX2 fell, they behaved like cells on a soft gel. Cells that had been on the stiff gel for ten days did not revert. YAP stayed nuclear and osteogenic markers stayed up for days after softening, and those cells differentiated toward bone at a higher rate than naive cells on soft gels. The authors coined the term mechanical dosing: memory depends on how long the cell was exposed.

The result has been reproduced and extended. Li et al. (Biomaterials, 2017) showed MSCs expanded on stiff substrates retain enhanced osteogenesis after transfer to soft. Nasrollahi et al. (Biomaterials, 2017) showed epithelial cells primed on stiff substrates migrate faster on soft ones for days afterwards. Killaars et al. (Advanced Science, 2019) showed the memory extends to chromatin.

How the memory is stored

Three layers, on different timescales.

YAP/TAZ persistence (hours to days). On stiff substrates YAP and TAZ sit in the nucleus and drive proliferation and osteogenic programs. After softening, they should leave. With enough prior dosing they stay longer than the mechanics justify, continuing to transcribe the stiff program (Yang 2014; Dupont et al., Nature, 2011, for the underlying mechanism).

miRNA and signalling feedback (days). Li et al. (PNAS, 2017) identified miR-21 as a mechanical memory element in MSCs: stiff priming raised miR-21, which suppressed the soft-substrate program and persisted after transfer.

Chromatin remodelling (days to weeks). Killaars et al. (Advanced Science, 2019) showed extended stiff exposure changes histone acetylation and chromatin condensation in MSCs, and that the chromatin state, not just the signalling state, carried into the new environment. Mechanical cues get written into the epigenome. Heo et al. (Science Advances, 2020) linked this to nuclear mechanics and lamin A/C, which itself scales with tissue stiffness (Swift et al., Science, 2013).

Why duration is the critical variable

Short exposure produces reversible memory; long exposure produces durable change. Yang's threshold in MSCs was between one and seven days on stiff. Most MSC expansion runs for weeks on tissue culture plastic at roughly 1 GPa, which is the maximal mechanical dose any experiment could deliver. By the time those cells reach a soft 3D gel, they have been primed harder than any experimental condition, and the priming does not wash out in a passage.

This is a plausible explanation for a common frustration: two labs run the same differentiation protocol in the same matrix and get different results because their cells had different expansion histories. Passage number, plastic versus soft-substrate expansion, and time since thawing all set the dose.

What to do about it

Report substrate history. Methods sections should state expansion substrate and stiffness, number of passages, and days in culture before the experiment, alongside the final matrix properties. The final matrix is only half the mechanical input.

Match expansion to experiment where possible. Expanding MSCs on soft substrates before a soft-matrix experiment reduces priming. Some groups now expand on hydrogels of tissue-like stiffness for exactly this reason.

Control for memory in multi-stage protocols. If your protocol moves cells between mechanical conditions, include a condition that has been at the second stiffness throughout, so you can see how much the prior stage contributed.

Avoid re-embedding when you can. Each transfer between gels adds a dosing step and a recovery step. A matrix whose mechanics can be changed in place lets you stage stiffness without moving the cells, which keeps the history simple and known. This is one of the reasons we built in situ stiffening and softening into DyNAtrix.

Use dosing deliberately. Memory is also a tool. A defined stiff priming step before a soft differentiation stage can enhance osteogenic yield, as Yang showed. Once you know the dose-response for your cells, you can use it.