Static Matrices: A Historical Accident in Dynamic Cell Culture

Team Dynamic Matrices | 2 September 2025

For two decades, 3D cell culture has run on gels with a single fixed stiffness set at the moment of gelation and held for the whole experiment. That became the standard not because anyone decided tissues are static, but because the materials that were easy to make and characterise in the 1990s and 2000s happened to be static. We think it is worth telling that history plainly, because it explains a default many labs have never questioned.

How the default was set

Polyacrylamide gels entered mechanobiology because Pelham and Wang (PNAS, 1997) needed a substrate whose stiffness could be varied by changing crosslinker ratio while keeping surface chemistry constant. That was a brilliant experimental design choice: it isolated stiffness as a variable. Engler et al. (Cell, 2006) then used the same material to show that MSCs commit to neural, muscle or bone lineages depending on substrate elasticity alone. The result was rigorous and it launched a field.

It also left an unintended assumption behind: that matrix mechanics is a single dial, set at the start. Polyacrylamide is covalently crosslinked, purely elastic, and never changes. Matrigel, the other workhorse, gels once at 37 °C and then has whatever mechanics the lot happened to have. Neither material could do anything else, so nobody designed experiments in which the matrix evolved.

What real tissue does instead

Living ECM is not a fixed elastic solid. It relaxes stress on timescales of seconds to minutes because its crosslinks are physical and reversible (reviewed in Chaudhuri et al., Nature, 2020). It is continuously degraded by cell-secreted proteases and rebuilt by deposition. Its stiffness changes over development, wound healing, fibrosis and tumour progression, sometimes by an order of magnitude. Gjorevski et al. (Nature, 2016) showed why this matters for organoids: intestinal stem cells needed one stiffness to expand and a softer, degrading gel to differentiate. A fixed gel forces you to choose one regime and lose the other.

The elasticity-sensing pathways identified in the early work remain valid. They are the reason dynamic matrices matter: if cells respond to stiffness, they will respond to stiffness changing.

The downstream costs of a static default

• Organoid protocols pick one mechanical regime at gelation and accept that later stages are mismatched.
• Studies that need a change in mechanics physically transfer cells between gels, adding handling steps that damage cells and destroy reproducibility.
• Matrigel's fixed mechanics are also lot-variable mechanics, so "static" in practice means "uncontrolled".
• Purely elastic synthetic gels report a stiffness that sounds physiological while presenting a mechanical environment no tissue has.

Programmable matrices change the baseline

Reversible, exchangeable crosslinks let a gel's mechanics be adjusted after cells are encapsulated. Chaudhuri et al. (Nature Materials, 2016) did this with ionic alginate to tune stress relaxation at fixed stiffness. We did it with DNA: in DyNAtrix, short DNA duplexes crosslink a polyacrylamide backbone, and their sequence and length set the relaxation time over about five orders of magnitude. Adding a strand mid-culture stiffens or softens the gel in place, with no transfer and no disturbance to the tissue (Peng et al., Nature Nanotechnology, 2023).

The point is not that every experiment needs a changing matrix. It is that dynamic behaviour should be the baseline against which "static" is a deliberate simplification, not the other way round. Fixed stiffness was never biologically optimal. It was the limit of what materials could do. That limit no longer applies.