Static Matrices Are a Historical Accident, Not a Design Choice

Team Dynamic Matrices | 2 September 2025

For twenty years, the standard 3D cell culture substrate has been a gel with one number attached to it: a stiffness value, measured once, fixed for the life of the experiment. Researchers have treated that number as a biological parameter worth controlling for, the way you would control temperature or pH. It rarely occurs to anyone to ask why the matrix should be static in the first place. The answer has almost nothing to do with cell biology and almost everything to do with what happened to be easy to manufacture and characterize forty years ago.

Polyacrylamide gels became the workhorse of mechanobiology because they were tunable through crosslinker concentration, optically clear, and simple to functionalize with collagen. Matrigel became the default 3D matrix because it was the only gelling, cell-adhesive material anyone had in a tube, extracted from Engelbreth-Holm-Swarm mouse sarcoma and sold by the bottle. Neither material was chosen because it modeled how the extracellular matrix actually behaves. Both were chosen because they were available, and availability hardened into convention. That convention now shapes what questions the field considers askable.

A landmark result built on a limitation, not despite one

The moment mechanobiology became a discipline in its own right is usually traced to Engler, Sen, Sweeney and Discher's 2006 paper in Cell, which showed that naive mesenchymal stem cells committed to neurogenic, myogenic or osteogenic fates depending on the elasticity of the polyacrylamide substrate they sat on, independent of soluble differentiation factors. It is a genuinely foundational result, and it is also, quietly, a demonstration of what a static-matrix worldview makes visible and what it hides. The experiment worked because stiffness was held constant long enough to isolate its effect. That is good experimental design. But it also trained a generation of biologists to think of matrix mechanics as a dial you set once, rather than a variable the tissue itself is constantly adjusting.

Real extracellular matrix does not hold still. It stress-relaxes under sustained load, remodels through enzymatic degradation and deposition, stiffens with fibrosis and disease, and softens during development and wound resolution. Cells do not read a single elasticity value, they read the rate and history of mechanical change, through pathways like YAP/TAZ nuclear localization that respond to how a substrate behaves over time, not what it measures on a rheometer at t=0. A field that spent two decades building its intuition on fixed-stiffness gels was, in effect, running every experiment with the clock stopped.

What the accident has cost

The clearest evidence that static matrices were a limitation rather than a design choice is what has gone wrong downstream. Matrigel's batch-to-batch variability in protein composition, concentration and stiffness is now a well-documented driver of poor reproducibility across organoid studies, with the same protocol in two labs, or even two lots, producing organoids of different size, morphology and functional maturity. That is not noise around a good model. It is a symptom of building an entire subfield on a substrate nobody can fully define or control.

The cost shows up commercially too. Undefined, animal-derived matrices carry regulatory and manufacturing baggage that static synthetic gels only partly solve, because a synthetic gel that is merely stiffness-tunable still forces researchers to choose a mechanical regime up front and commit to it for the culture's entire life. Organoid maturation, tissue morphogenesis and disease progression are all processes that unfold over time and require different mechanical environments at different stages. A matrix that cannot change with the tissue forces researchers to either accept a mismatch at some point in the culture, or physically transfer cells between different fixed-stiffness gels, introducing exactly the kind of manual handling step that undermines the reproducibility everyone is trying to fix.

Dynamic matrices are not a feature add-on

Once you see static stiffness as a manufacturing legacy rather than a biological requirement, the recent wave of viscoelastic and stress-relaxing hydrogel research reads differently. Work on stiffness-adjustable hydrogels driving organoid maturation through YAP/Notch mechanotransduction, and on reversible, transient-crosslink systems that let researchers dial viscoelasticity up or down mid-culture without disrupting cell-matrix adhesion, is not an incremental improvement on the 2006-era toolkit. It is the correction of a category error. These systems let the matrix follow the tissue's developmental trajectory instead of forcing the tissue to develop around a matrix property that was fixed at the moment of gelation.

This is the premise DyNAtrix® is built on. A programmable matrix whose mechanical and biochemical properties can be reset after cells are already inside it treats dynamism as the baseline condition of a physiological microenvironment, not a specialized feature for labs studying mechanotransduction specifically. Every lab culturing organoids, iPSC-derived tissue or primary cells over multiple weeks is running a process that changes state. The matrix should be able to change state with it.

The field was always headed here

None of this is a rejection of the static-matrix era's findings. The elasticity-sensing pathways Engler and colleagues identified are real, and they are precisely what make dynamic matrices interesting rather than merely convenient. The point is that fixed stiffness was never the biological ideal being approximated. It was the ceiling imposed by what could be poured, crosslinked and characterized with 1980s and 2000s tools. Now that programmable, reversibly tunable materials exist, holding onto static gels as a default is no longer a simplifying assumption, it is an unforced constraint that costs reproducibility, costs biological fidelity, and costs the ability to model processes that actually happen over time. The question worth asking in every grant proposal and product design review going forward is not whether a matrix's stiffness is right, but whether it is even supposed to stay the same.