An organoid culture protocol produces a population of three-dimensional structures that vary in size, in the ratio of stem to differentiated cells, and in how far a compound has to diffuse to reach the center. That variability is not a defect. It is what makes organoids more tissue-like than a monolayer, and it is also why an experimental design that works for a 96-well plate of cells often produces uninterpretable data in 3D. Most organoid experiments fail at the planning stage, not the bench — and the three decisions that decide whether the result means anything are which controls you plate, when you treat, and which readout you choose.
The failure modes are consistent. Organoids are treated before they have recovered from passaging, so drug effect and recovery kinetics are confounded. A single well per condition is scored, in a system where well-to-well variance is large. A viability reagent is chosen without checking whether it can penetrate the matrix dome. Conditioned medium changes batch mid-experiment, and a real difference appears that has nothing to do with the treatment. None of these are fixed by better pipetting. They are fixed before anything is plated.
This guide walks the full organoid culture workflow from tissue to data for epithelial and tumor organoids, in the order the decisions actually have to be made. It is a planning guide, not a bench SOP. Day numbers are typical for intestinal and colorectal epithelial organoids; tissue type changes them substantially, so treat the structure as transferable and confirm the intervals against a protocol validated for your tissue. Catalog anchors come from the BioHippo cell line, recombinant protein, antibody and ELISA ranges; every product named below was checked against the live catalog on 3 September 2026.
Five Decisions to Make Before Plating an Organoid Experiment
Work through these before ordering reagents. Each one constrains the choices after it, so taking them out of order usually means re-doing the experiment.
| Decision | What it determines |
|---|---|
| 1. What is the actual question? | Whether you need a viability number, a differentiation phenotype, or a morphology measurement. “Does this compound do anything?” and “does this compound kill stem cells specifically?” require different plates. |
| 2. Which model, and how many lines? | A single organoid line gives you a result about that line. Donor-to-donor variation in patient-derived organoids is substantial, so a claim about a tumor type needs several independent lines, not several wells of one. |
| 3. Niche factors: defined or conditioned? | Recombinant Wnt3a / R-spondin / Noggin give lot-controlled, reportable concentrations. L-WRN conditioned medium is far cheaper at scale but introduces a batch variable you must control for.2 |
| 4. When does treatment start? | Everything about the timeline. This is the single most common design error and is covered in detail below. |
| 5. Destructive or longitudinal readout? | Whether one plate can answer the question or you need a separate plate per timepoint. Lysis-based assays end the well; imaging does not. |
If you cannot answer 1 and 5 in one sentence each, the experiment is not ready to plate.
The Organoid Culture Workflow, Phase by Phase
The timeline below runs from medium preparation a week before seeding to the final readout. Two phases are marked as decision points: ROCK-inhibitor withdrawal and the treatment window. Both are where the timing goes wrong most often.
Day −7 to −1 · Preparation: establish the niche before you need it
If you are using conditioned medium, produce and qualify the batch now. A single large batch that covers the entire experiment removes a variable you otherwise cannot control retrospectively. L-WRN cells secrete Wnt3a, R-spondin-3 and Noggin together, which is why they are widely used as a single source for gastrointestinal organoid medium.2
Why it matters. Conditioned medium potency varies between harvests. Qualify each batch functionally, for example by confirming it supports normal organoid formation efficiency from a reference line, and record the batch ID against every plate. A mid-experiment batch change is not recoverable at analysis.
Day 0 · Establishment: isolate, embed, and seed
Tissue is dissociated to crypts or small cell clusters, resuspended in cold basement membrane matrix, and dispensed as domes. Sato and colleagues showed that single Lgr5+ stem cells build crypt–villus organoids in matrix with EGF, Noggin and R-spondin, without any mesenchymal support.1
Why it matters. Seeding density drives organoid size at every later timepoint, and size drives drug penetration. Count and normalize input rather than splitting by ratio, or your treated and control wells will differ in geometry before you add anything.
Day 0 to +2 · ROCK inhibition: add Y-27632, then take it away
Dissociation triggers apoptosis in single epithelial cells. The ROCK inhibitor Y-27632 suppresses that dissociation-induced death and markedly improves survival and cloning efficiency of dissociated cells.3 It is standard for the first 48–72 hours after seeding or passaging.
Why it matters. Y-27632 is a signaling inhibitor, not an inert supplement. Leaving it in through treatment means every well is under chronic ROCK inhibition, which alters morphology and differentiation. Withdraw it on schedule, and withdraw it identically across every arm.
Day +3 to +10 · Expansion: reach a defined, reproducible state
Feed on a fixed schedule. Organoids expand, and depending on the medium composition either stay in a stem-enriched spheroid state or begin differentiating into budded structures containing multiple lineages.
Why it matters. “Confluent” has no meaning in 3D. Define your starting condition by something measurable — days post-passage plus mean organoid diameter — and hold it constant across experiments. This is what makes two runs comparable months apart.
Day +4 to +7 typical · Treatment window: treat only after recovery is complete
The window opens once organoids have re-formed lumens and resumed steady growth, and closes before they outgrow the dome or exhaust the medium. Treating inside the recovery period is the error that most often produces a large, clean, meaningless effect.
Why it matters. An organoid still recovering from dissociation is already stressed, so a compound applied then measures additive stress, not drug activity. Verify recovery by morphology before dosing, and dose every arm on the same day post-passage.
Treatment +24 h to +14 d · Readout: match the readout to the question and the timescale
Acute cytotoxicity resolves in 24–72 hours. Effects on stem-cell maintenance or differentiation need longer, often a full passage or more, because the phenotype only appears as the population turns over.
Why it matters. A short endpoint on a slow phenotype reads as “no effect.” Before choosing the timepoint, ask how long the biology needs to manifest, then confirm with a pilot time course rather than a single guess.
Organoid Culture Controls: What Goes on the Plate Besides Your Samples
Organoid experiments need controls that monolayer work does not, because the matrix, the conditioned medium, and the three-dimensional geometry each introduce their own artifacts. The four marked essential belong on every plate.
| Control | Priority | Why it is on the plate |
|---|---|---|
| Vehicle control | Essential | Same solvent, same final concentration, same volume as the highest dose arm. DMSO above roughly 0.1% affects many organoid cultures on its own, so a vehicle arm matched to the top dose is what tells you whether the effect is the compound. |
| Matrix-only (no-cell) wells | Essential | Domes with matrix and medium but no organoids. Basement membrane matrix and phenol red both contribute background to luminescent, colorimetric and fluorescent readouts, and that background is not always trivial relative to a small organoid signal. |
| Positive (killing) control | Essential | A known cytotoxic agent at a dose that clears the well. Without it, a flat dose–response cannot be distinguished from an assay that never had a working dynamic range in this format. |
| Plate-position control | Essential | Edge wells evaporate faster and behave differently over multi-day incubations. Either leave the perimeter unused and fill it with PBS, or distribute conditions across positions so that position and treatment are not confounded. |
| Passage-matched comparator | Strongly recommended | Organoids drift with extended passaging. Compare arms at the same passage number, and record it. A control from passage 4 against a treatment at passage 19 is not a clean comparison. |
| Conditioned medium batch control | If using conditioned medium | A reference line carried alongside on each medium batch, scored for formation efficiency. This is what lets you tell a real treatment effect from a medium potency shift when results move between runs. |
| Untreated (no-vehicle) baseline | Recommended | Medium only, no solvent. Distinguishes vehicle toxicity from compound activity, which matters whenever the vehicle arm itself looks depressed relative to expectation. |
| Multiple independent lines | For generalizable claims | Technical replicates measure pipetting. Independent donor lines measure biology. Patient-derived organoids recapitulate donor treatment responses,5 which is precisely why a single line cannot support a claim about a tumor type. |
Choosing Organoid Readouts: Question, Assay, and the 3D Caveat
Every assay below is routine in 2D. What changes in 3D is the matrix, which can trap analytes, block reagent access, and add background. The caveat column is the part that catches people out.
| Your question | Readout | The 3D caveat |
|---|---|---|
| Are they alive, and how many? | ATP-based luminescent viability, in a formulation validated for 3D | Standard 2D viability reagents may not lyse organoids inside an intact dome. Use a 3D-validated formulation, or confirm complete lysis before trusting the numbers. |
| Is the compound killing cells? | LDH release into supernatant | Released LDH can be partly retained within the matrix dome rather than reaching the supernatant, which biases the readout low. Interpret as relative across matched wells, not as absolute cytotoxicity. |
| Apoptosis specifically? | Cleaved caspase-3 by IF or ELISA on lysate | Whole-dome lysate averages across the population and hides whether death is uniform or confined to organoid cores. Imaging preserves that spatial information; lysate does not. |
| Are they still proliferating? | Ki-67 immunostaining | Proliferation in organoids is spatially restricted to specific zones. A whole-well average will miss a redistribution of proliferation that leaves the total unchanged. |
| Is the stem compartment affected? | LGR5 expression; organoid formation efficiency after replating | Formation efficiency on replating is the functional test and is generally more informative than a marker level alone. It costs an extra passage, so build it into the timeline from the start. |
| Has differentiation changed? | Lineage markers (e.g. CDX2 for intestinal identity), morphology scoring | Differentiation state is strongly driven by medium composition. Only compare arms cultured in identical medium, from the same batch. |
| Are they growing? | Brightfield imaging, diameter and count over time | Non-destructive and longitudinal, which is its main advantage. Segmentation must be automated and applied blind, since manual scoring of organoid boundaries is where bias enters most easily. |
Pairing one functional readout with one imaging readout generally answers more than either alone, because it separates “how much signal” from “where the signal is.”
Destructive vs longitudinal readouts: how many plates the question costs
A lysis-based assay ends the well, so every timepoint needs its own plate. Brightfield imaging leaves the well intact, so one plate can be read at 24 h, 72 h, 7 d and 14 d. Decide this before ordering plates and matrix, because the plate count sets the budget.
Six Organoid Timing Errors and What They Look Like in the Data
| Error | How it presents | Fix |
|---|---|---|
| Treating during post-passage recovery | Large effect, poor reproducibility between runs | Confirm morphological recovery before dosing; fix days post-passage across arms |
| Y-27632 left in through treatment | Altered morphology in every arm; muted differentiation phenotypes | Withdraw after 48–72 h, identically in all arms |
| Endpoint too early for the biology | Clean “no effect” on a real phenotype | Pilot time course before committing to a single endpoint |
| Organoids overgrown at dosing | Apparent resistance at high dose; core necrosis in controls | Standardize diameter at treatment start, not just day number |
| Conditioned medium batch change mid-run | Step change between plates that tracks date, not treatment | One qualified batch per experiment; log batch ID per plate |
| Arms dosed on different days | Day-of-treatment confounded with condition | Dose all arms same day; if impossible, block and randomize deliberately |
Organoid Culture Reagents: What BioHippo Stocks for This Workflow
Every product below was verified against the live catalog on 3 September 2026. Pack sizes, species and formulation vary, so confirm on each product page before ordering.
Niche factors (defined, recombinant)
| Product | Supplier | Role in the protocol |
|---|---|---|
| Recombinant Human RSPO1 | Fine Test | Wnt agonist; core stem-cell niche factor |
| Recombinant Human RSPO2 | Fine Test | R-spondin family alternative |
| Recombinant Human WNT3A Protein, N-His | AtaGenix | Wnt pathway activation |
| Recombinant Human NOG/Noggin Protein, C-Fc | AtaGenix | BMP inhibition |
| Recombinant Human EGF Protein, N-GST | AtaGenix | Proliferative drive |
| Recombinant Human FGF-10 | Fine Test | Used in several non-intestinal organoid systems |
Conditioned medium source lines and small molecules
| Product | Supplier | Role in the protocol |
|---|---|---|
| L-WRN cell | Cytion | Secretes Wnt3a, R-spondin-3 and Noggin together2 |
| L-WRN Cells | abm | Alternative L-WRN source |
| L Wnt-3A cell | Cytion | Wnt3a-only conditioned medium |
| Y-27632 | StressMarq | ROCK inhibitor; post-dissociation survival3 |
Tumor organoid–relevant patient-derived lines
Tumor organoids can be derived from resected tissue and biopsies, survive cryopreservation, and retain disease-stage-specific characteristics, which is what makes banked patient-derived lines usable as a starting point rather than requiring fresh tissue for every experiment.4
| Product | Supplier | Role in the protocol |
|---|---|---|
| HROC395Met1 cell | Cytion | Patient-derived colorectal metastasis line |
| HROC450Met1 T0 M1 cell | Cytion | Patient-derived colorectal metastasis line |
Readout reagents
| Product | Supplier | Readout |
|---|---|---|
| LDH Cytotoxicity Assay Kit | BioAssay Systems | Membrane damage / cytotoxicity |
| Human Cleaved CASP3 ELISA Kit | Fine Test | Apoptosis on lysate |
| Anti-Ki67/MKI67 Polyclonal Antibody | AtaGenix | Proliferation by IF/IHC |
| Ki67 Antibody | NSJ Bioreagents | Proliferation, alternative clone |
| CDX2 Antibody | NSJ Bioreagents | Intestinal lineage identity |
| Human LGR5 ELISA Kit | Fine Test | Stem-cell marker quantification |
Browse the full ranges: Proteins & Peptides · Cells · Antibodies · ELISA Kits · Biochemicals. Several reagents this workflow requires, including basement membrane matrix and complete organoid media, are outside the products listed here; contact us and we can help you scope the full list.
Organoid Culture Protocol FAQ
How many replicates does an organoid experiment need?
More than a monolayer experiment, and the useful distinction is between technical and biological. Several domes per condition control pipetting and plate position. Independent passages, and ideally independent donor lines, control biology. Because well-to-well variance in 3D is high, three technical replicates is a practical floor rather than a target, and a claim that generalizes beyond one line needs independent lines regardless of how many wells you ran.
Recombinant niche factors or L-WRN conditioned medium?
Defined recombinant factors when the concentration needs to be known, reported and reproducible, which includes most published dose–response work. L-WRN conditioned medium when you are running at a scale where recombinant factors are the dominant cost. The trade-off is real: conditioned medium is cheaper but adds a batch variable, so if you go that route, qualify each batch functionally and log the batch ID against every plate.
Why withdraw Y-27632 if the organoids look fine on it?
Because it is doing something. Y-27632 protects dissociated cells from apoptosis, which is exactly why it is added after passaging, but ROCK signaling also influences cytoskeletal organization and differentiation. Leaving it in means every arm is under chronic pathway inhibition, and a differentiation phenotype you are trying to detect may be suppressed in controls and treatment alike. Withdraw it on a fixed schedule and apply that schedule identically everywhere.
Can I use my standard 2D viability kit on organoids?
Check lysis before you trust it. Most 2D viability chemistries were validated on monolayers, and an intact organoid inside a matrix dome is a much harder lysis target. Incomplete lysis reads as reduced viability, which looks exactly like a drug effect. Either use a formulation validated for 3D, or run a lysis-completeness check against a known input before relying on the numbers.
How do I know organoids have recovered enough to treat?
Use morphology, not the calendar alone. Re-formed lumens and a return to steady growth indicate recovery; the day number is a planning estimate that varies with tissue, line and seeding density. In practice, run a short pilot on your line, record the day at which recovery is complete, then fix that day for every subsequent experiment so treatment always starts from the same state.
What is the minimum control set for an organoid culture experiment if plate space is tight?
Vehicle matched to the top dose, matrix-only background wells, a positive killing control, and deliberate handling of plate position. Those four separate compound effect from solvent effect, from matrix background, from a dead assay, and from evaporation gradients. Everything else in the controls section improves the experiment; those four are what make it interpretable at all.
Do tumor organoids need the same controls as normal epithelial organoids?
Yes, plus one more. Tumor organoids from different donors respond differently to the same treatment,5 so a claim about a tumor type needs several independent patient-derived lines run through the same plate design. The four essential controls still apply to every line; the extra requirement is line number, not control type.
Scope Your Organoid Reagent List With a Specialist
Tell a BioHippo technical specialist your tissue type, whether you are going defined or conditioned on niche factors, and what your endpoint is. We will map the reagent list against your workflow and help you fill any remaining gaps, including items outside the products listed here. Talk to a specialist or request a quote.
References
- Sato T, Vries RG, Snippert HJ, et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature. 2009;459(7244):262–265. PMID 19329995. doi:10.1038/nature07935
- Miyoshi H, Stappenbeck TS. In vitro expansion and genetic modification of gastrointestinal stem cells in spheroid culture. Nat Protoc. 2013;8(12):2471–2482. PMID 24232249. doi:10.1038/nprot.2013.153
- Watanabe K, Ueno M, Kamiya D, et al. A ROCK inhibitor permits survival of dissociated human embryonic stem cells. Nat Biotechnol. 2007;25(6):681–686. PMID 17529971. doi:10.1038/nbt1310
- Boj SF, Hwang CI, Baker LA, et al. Organoid models of human and mouse ductal pancreatic cancer. Cell. 2015;160(1–2):324–338. PMID 25557080. doi:10.1016/j.cell.2014.12.021
- Vlachogiannis G, Hedayat S, Vatsiou A, et al. Patient-derived organoids model treatment response of metastatic gastrointestinal cancers. Science. 2018;359(6378):920–926. PMID 29472484. doi:10.1126/science.aao2774
Day intervals given are typical for intestinal and colorectal epithelial organoids and vary substantially by tissue type, line and seeding density; confirm against a protocol validated for your system. Product availability, species and formulation were verified against the live BioHippo catalog on 3 September 2026 and are subject to change, so confirm on each product page before ordering. All figures are illustrative schematics, not experimental data. Research use only.
