Isolating EVs from Hydrogel-Conditioned Media

Team Dynamic Matrices | 20 May 2026

Cells in 3D secrete extracellular vesicles that differ from those the same cells secrete on plastic, in number, cargo and size (Thippabhotla, Zhong and He, Scientific Reports, 2019; Rocha et al., Advanced Science, 2019). That makes 3D-conditioned medium the more relevant source for EV studies, and it also makes the isolation harder, because the medium now carries whatever the hydrogel sheds. Standard EV protocols designed for 2D supernatants do not account for that. Here is how we handle it.

What the matrix adds to the medium

Any hydrogel releases something into the medium over a culture: unpolymerised precursor, degradation fragments, leached proteins, and particulates in the tens-to-hundreds of nanometre range. Basement membrane extract sheds protein aggregates and laminin-rich particles; collagen sheds fibril fragments; synthetic gels shed polymer fragments and, if degradable, cleavage products. Many of these overlap with EVs in size and some in density, so they co-isolate and inflate particle counts, dilute purity, and can carry adsorbed RNA and protein that looks like cargo.

Step 1: the hydrogel-only control

Before optimising anything, prepare a gel identical to your culture (same lot, same volume, same medium) with no cells, incubate it on the same schedule, and process its medium through your full EV isolation workflow. Then characterise the output by nanoparticle tracking analysis and protein assay. This tells you what the matrix contributes on its own, whether it sheds particles in the EV size range, and how much of your "EV" signal in cell-containing cultures could be matrix. Run it once per matrix lot and again whenever the protocol changes. It is the single most informative control in the workflow, and almost nobody runs it.

Step 2: choose an isolation method that discriminates against matrix

Our combined approach: extended pre-clearing, concentration by 100 kDa ultrafiltration, then SEC (qEV or equivalent), and a density gradient when purity matters more than yield. Check against the hydrogel-only control at each stage.

Step 3: extended pre-clearing

Standard pre-clearing is 300 g for 10 minutes to remove cells and 2,000 g for 20 minutes for debris. For hydrogel media, add a 10,000 g spin for 30 minutes to remove larger matrix fragments and a 0.22 µm filtration before concentration. For gels that shed sub-200 nm particles, a 0.1 µm filter is worth testing, at the cost of losing larger EVs.

Step 4: collect on a consistent schedule

EV concentration in medium rises with time since the last feed. If replicates are collected at 24, 36 and 48 hours after feeding, the variation in EV yield will exceed most biological effects. Fix the collection interval, use EV-depleted serum or serum-free medium for the conditioning window, and record the volume of medium per gel volume, because the ratio changes the concentration.

Step 5: verify what you isolated

MISEV2023 (Welsh et al., Journal of Extracellular Vesicles, 2024) asks for characterisation on three axes, and for hydrogel-derived samples we treat it as mandatory rather than advisory:

• Particle count and size by NTA or TRPS, compared to the hydrogel-only control
• Protein markers: positive (CD63, CD9, CD81, TSG101, ALIX) and negative (calnexin, GM130, and any matrix protein, for example laminin for BME)
• Morphology by transmission electron microscopy on a subset, to distinguish cup-shaped vesicles from amorphous aggregates

Report the particle-to-protein ratio as a purity metric and the method in full.

A note on which gel you use

Matrix background is easiest to manage when the matrix is defined and does not degrade unless told to. A synthetic gel with a known composition sheds a known thing, which the hydrogel-only control can quantify; a biological extract sheds a lot-variable mixture. DyNAtrix does not release protein and is stable through a culture window, and it can be dissolved on demand with a nuclease, which we use to recover EVs trapped inside the gel at the end of an experiment rather than only those that diffused into the medium. That intra-gel fraction is often larger than the medium fraction for dense cultures.