Dissociating Organoids to Single Cells Without Wrecking Viability

This protocol covers dissociating organoids to a single-cell suspension without losing the cells you are trying to measure: removing the extracellular matrix, choosing and timing the enzymatic step, controlling shear during trituration, and matching the endpoint to what happens next — replating, flow cytometry, or single-cell RNA sequencing. It is written reagent-neutral, and it is explicit about which parts are supported by published data and which are starting points you must titrate yourself.

Decide the Endpoint Before You Start

“Single cells” means different things depending on what follows, and the protocol changes accordingly:

Materials and Reagents

• Established organoid culture in a basement-membrane matrix dome, harvested at the density and day appropriate to the model (see Broutier et al., Nat Protoc 2016 and Driehuis et al., Nat Protoc 2020 for model-specific culture conditions)
• Ice-cold basal medium (advanced DMEM/F12 or equivalent) for matrix removal, or a commercial cell-recovery solution
• Sterile DPBS without Ca²⁺ and Mg²⁺
• Dissociation reagent — a TrypLE™-type recombinant enzyme, an animal-origin-free recombinant trypsin-like enzyme, or Accutase®; see the reagent notes below
• DNase I, typically 10–100 µg/mL final, to prevent DNA-mediated clumping from lysed cells
• ROCK inhibitor Y-27632, 10 µM, for any workflow that replates single cells
• Quench medium — serum-containing medium, a defined inhibitor, or plain medium for dilution-quenched enzymes
• Cell strainers, 37–40 µm (and optionally 70 µm for a first pass)
• Pre-wetted wide-bore and standard P1000 tips; 0.4% trypan blue and a counter
• Refrigerated benchtop centrifuge; ice bucket; Class II cabinet

Protocol: Organoid to Single-Cell Suspension

Volumes assume the organoids from two to four 50 µL matrix domes in a 24-well plate. Scale proportionally.

1. Chill the plate. Aspirate the culture medium. Place the plate on ice and add 0.5–1 mL ice-cold basal medium per well. Basement-membrane matrices depolymerise in the cold; giving the matrix time to liquefy before any enzyme is added is the difference between recovering organoids and recovering gel fragments.
2. Release the domes. Disrupt each dome with a pre-wetted wide-bore P1000, transfer to a pre-wetted, pre-chilled 15 mL tube, and keep on ice. Rinse the wells with a further 0.5 mL cold medium and pool.
3. Wash out the matrix. Centrifuge at 300 × g, 5 minutes, 4 °C. Aspirate the supernatant carefully — the pellet is loose. Resuspend in 1 mL cold basal medium and repeat once. Residual matrix protein is a common cause of an enzymatic step that appears not to work, because the enzyme spends itself on the gel instead of the cells.
4. Optional — mechanically break large organoids first. For large, dense or budded organoids, triturate 5–10 times with a standard (not wide-bore) P1000 before adding enzyme. Fragmenting first shortens the enzymatic exposure needed, which is the exposure that damages cells.
5. Add the dissociation reagent and DNase I. Resuspend the pellet in 0.5–1 mL of pre-warmed dissociation reagent containing DNase I. Use enough volume to keep the fragments freely suspended; more than that just dilutes your recovery.
6. Incubate with intermittent trituration. Place at 37 °C. Every 3–5 minutes, remove the tube, triturate 10–15 times with a P1000, and take a 5 µL aliquot to the microscope. Total incubation for most epithelial organoid models falls in the 5–20 minute range, but this is a range to titrate, not a specification — it varies with organoid size, model, passage, and the day of the culture. Stop as soon as the field is dominated by single cells with only occasional doublets. Chasing the last 5% of doublets costs far more viability than it gains.
7. Quench. Add at least twice the enzyme volume of cold medium. Use serum-containing medium, a defined inhibitor, or — for recombinant trypsin-like enzymes designed for dilution quenching — plain medium, in which case no inhibitor is needed. Confirm against your specific reagent's instructions.
8. Strain. Pass through a 37–40 µm strainer, pre-wetted with medium. For debris-heavy preparations, a 70 µm pre-strain reduces clogging. Rinse the strainer with a small volume of medium to recover retained cells.
9. Wash and resuspend. Centrifuge at 300–400 × g, 5 minutes, 4 °C, aspirate, and resuspend in the buffer the downstream step requires — medium with 10 µM Y-27632 for replating, cold PBS with 0.04% BSA for most droplet scRNA-seq platforms, or your staining buffer for flow.
10. Count and assess. Trypan blue count plus a visual check for debris and doublets. For scRNA-seq, most platforms want >80–90% viability and a clean, debris-free suspension; a viable count alone does not tell you whether the preparation is clean.
11. Proceed immediately. Dissociated epithelial cells deteriorate on the bench. Keep everything cold and move to the next step without a pause.

The Artefact Nobody Sees in a Viability Count

If your endpoint is transcriptomic, the dissociation is part of the experiment, not a preparation step before it. Three findings should shape your protocol:

Warm enzymatic digestion induces a stress transcriptome. Comparing tissue dissociation with collagenase at 37 °C against a cold-active protease at 6 °C across 155,165 cells from patient tumours, xenografts and cell lines, collagenase digestion produced a stress response; the authors derived a core set of 512 heat-shock and stress-response genes, including FOS and JUN, induced by warm collagenase digestion and minimised at 6 °C. Induction was highly conserved across cell types, and cell-type-specific responses were also seen in patient tissue (O'Flanagan et al., Genome Biol 2019).

The artefact is not evenly distributed. Single-cell sequencing showed dissociation-induced gene expression concentrated in particular tissue subpopulations rather than spread uniformly — so it does not cancel out, and it can look like real biological difference between cell types (van den Brink et al., Nat Methods 2017).

Storage choices carry their own bias. A systematic comparison of dissociation and storage in mouse kidney confirmed that digestion on ice avoids the stress response seen at 37 °C, and found that cryopreserving dissociated cells caused a major loss of epithelial cell types, while methanol fixation preserved composition but suffered ambient RNA leakage (Denisenko et al., Genome Biol 2020).

What to do about it:

• Keep the warm enzymatic window as short as the endpoint allows, and everything else cold.
• Dissociate every arm of a comparison with the same reagent, the same temperature and the same duration. A batch effect that tracks your experimental groups is indistinguishable from biology.
• Score the O'Flanagan stress gene set in your data and report it. If it separates your conditions, you have measured your protocol.
• Consider single-nucleus RNA-seq where the epithelium will not survive dissociation intact — noting that nuclear libraries carry their own composition biases, including under-representation of lymphocytes.

Replating After Dissociation

If the point of the dissociation is to passage the line rather than to measure it, two rules change:

• Stop earlier. Small multicellular fragments re-form organoids far more efficiently than true single cells. Complete single-cell dissociation is a requirement of the assay, not of the culture.
• Add a ROCK inhibitor when you do plate single cells. Y-27632 at 10 µM in the plating medium is standard practice for single-cell organoid re-establishment, for the same anoikis-and-contractility reason it is used with pluripotent cells. Single sorted Lgr5⁺ stem cells can initiate crypt–villus organoids under appropriate conditions (Sato et al., Nature 2009), but single-cell plating efficiency is always lower than fragment plating efficiency.

Keep the cells cold and in suspension for as short a time as possible before they are back in matrix.

Troubleshooting

Reagent Choice, and What the Data Covers

Laboratories running xeno-free, GMP-adjacent or clinical-track organoid work replace animal-derived proteases with recombinant, animal-origin-free enzymes to remove TSE/BSE and adventitious-agent risk and to reduce lot-to-lot variability. CellTrypase (c-LEcta GmbH, a Kerry company) is one option in that class: a recombinant trypsin-like serine protease from a Fusarium oxysporum gene expressed in a Bacillus sp. host, ~22 kDa, cleaving after lysine and arginine; ≥95% purity by HPLC; formulated in PBS with 1.1 mM EDTA and sterile filtered; pH 7.1–7.6 and 270–320 mOsm/kg; endotoxin ≤1 EU/mL at 1× and ≤10 EU/mL at 10×; mycoplasma negative; sterility per Ph. Eur. 2.6.1 / USP <71>. It is quenched by dilution in buffer or medium, so no trypsin inhibitor is added to the stream, and it is stable for at least 15 months at 2–8 °C. R&D grade is manufactured under ISO 9001:2015 and GMP grade under EXCiPACT® GMP/GDP, from the same manufacturer.

Related reading: the CellTrypase cell dissociation protocol, the CellTrypase specifications guide, and the single-cell RNA sequencing protocol for the downstream workflow.

Frequently Asked Questions

How long should organoid dissociation take?

For most epithelial organoid models, 5–20 minutes of enzymatic exposure at 37 °C with intermittent trituration, but treat that as a range to titrate rather than a specification. Check a 5 µL aliquot under the microscope every 3–5 minutes and stop at single cells with occasional doublets.

Why is my suspension stringy and clumped?

Free DNA from lysed cells. Include DNase I in the dissociation reagent from the beginning rather than adding it as a rescue, and reduce your trituration force — the stringiness is telling you cells are being destroyed.

Do I need a ROCK inhibitor?

Only if you are replating single cells, where 10 µM Y-27632 in the plating medium is standard. It is unnecessary if you are dissociating for flow cytometry or sequencing, where the cells are not going back into culture.

Does dissociation change my single-cell RNA-seq data?

Yes, measurably. Warm enzymatic digestion induces a conserved set of heat-shock and stress-response genes, defined as a 512-gene core set including FOS and JUN, that is minimised by cold-active protease digestion at 6 °C (O'Flanagan et al. 2019), and the induction is not uniform across cell types (van den Brink et al. 2017). Keep the warm window short, hold it identical across arms, and score the stress genes in your analysis.

Should I remove the matrix before adding enzyme?

Yes. Residual basement-membrane matrix consumes protease and is the most common reason a dissociation appears not to work. Chill the plate, liquefy the dome in cold medium, and wash twice at 4 °C before the enzymatic step.

What viability do I need for a droplet scRNA-seq platform?

Most platforms specify above 80–90% viability with a clean, debris-free suspension. Check the specification for your specific instrument. Note that viability alone does not capture ambient RNA or debris, both of which degrade the run.

Can I freeze dissociated organoid cells and process them later?

Be careful. In a systematic kidney comparison, cryopreserving dissociated cells caused a major loss of epithelial cell types, while methanol fixation preserved composition but introduced ambient RNA leakage (Denisenko et al. 2020). If you must store, validate the storage step on your own model before committing samples to it.

Is an animal-origin-free enzyme suitable for organoids?

Animal-origin-free recombinant enzymes are used as a class in xeno-free and clinical-track organoid work. For any specific product, ask what performance data exists for organoids. For CellTrypase, the manufacturer's measured data covers CHO-K1, HEK 293, MDCK and Vero only; organoids are listed as an intended application with no performance figures attached, so qualify it on your own model before switching.

Reagents and Support

Browse dissociation enzymes in the BioHippo enzymes collection. For help designing a dissociation comparison on your own organoid model, ask a scientist.

For research use only; not for use in diagnostic or therapeutic procedures. The R&D grade of CellTrypase is supplied for research use only. TrypLE™ is a trademark of Thermo Fisher Scientific and Accutase® is a registered trademark of its respective owner; these are referred to here for identification and comparison only, and BioHippo is not affiliated with, endorsed by, or sponsored by those companies. Incubation times, volumes, DNase concentrations and centrifugation settings given here are starting values to be confirmed on your own model, not specifications. Performance figures attributed to c-LEcta PIS CellTrypase v2.0 apply only to the four cell lines named; no performance data is claimed for organoids or any other cell type. Figures 1–4 are illustrative summaries prepared for this article from the cited published sources and the manufacturer's product information sheet; they are not experimental data generated by BioHippo, and schematic cell fields and relative bar heights marked qualitative are not measurements. Peer-reviewed sources are cited inline and indexed in PubMed.