Organotypic culture assays for murine and human primary and metastatic-site tumors

Abstract

Cancer invasion and metastasis are challenging to study in vivo since they occur deep inside the body over extended time periods. Organotypic 3D culture of fresh tumor tissue enables convenient real-time imaging, genetic and microenvironmental manipulation, and molecular analysis. Here we provide detailed protocols to isolate and culture heterogenous organoids from murine and human primary and metastatic site tumors. The time required to isolate organoids can vary based on the tissue and organ type but typically takes <7 hours. We also describe a suite of assays that model specific aspects of metastasis, including proliferation, survival, invasion, dissemination, and colony formation. We also specify comprehensive protocols for downstream applications of organotypic cultures that will allow users to (1) test the role of specific genes in regulating various cellular processes, (2) distinguish the contributions of several microenvironmental factors, and (3) test the effects of novel therapeutics.

EDITORIAL SUMMARY

This protocol describes the isolation and 3D culture of organoids from fresh murine or human primary and metastatic tumor tissue. It also provides instructions for real-time imaging, genetic and microenvironmental manipulation, and molecular analysis.

INTRODUCTION

Our current understanding of cell and molecular biology is largely derived from the study of cells in conventional 2D culture assays. However, this format has a limited ability to model the spatial relationships within and among cell types that characterizes intact tissue and organ structure. Culture in 2D on rigid substrates can also induce phenotypic changes¹. The advent of diverse 3D culture formats, including both tissue and stem cell organoids, has enabled a renaissance in the study of tissue level processes. Organoids have been used to reconstitute organ structure and function of the breast^2–8, lung^9–11, colon^12,13, pancreas¹⁴, prostate^15,16, liver¹⁷, stomach¹⁸, salivary gland¹⁹, kidney²⁰, and brain²¹. They have also been used to model complex disease states including cystic fibrosis²², Zika virus infections^23,24, and cancer^4,25–30. Tumor-derived organoids have been developed for several organ systems and have been shown to retain the genetic heterogeneity and histopathological characteristics of their originating tumor^28,31. Organoids are also used extensively in pre-clinical settings for drug development and testing since organoids may allow more accurate prediction of the drug responses of patients^28,32,33.

The term organotypic culture is used variously to describe distinct culture methods that attempt to mimic the tissue architectures and/ or behavior in vivo. These organotypic culture methods differ in terms of the cellular input used to initiate the culture, the culture format, and the complexity of cell types and microenvironment^31,34. Organotypic culture provides a histologically realistic and versatile platform to integrate diverse analytical approaches including real-time microscopy, genetics, next-generation sequencing, chromatin immunoprecipitation, and other biochemical or molecular biology applications. Here, we describe the organotypic culture methods that we have successfully used to identify molecular drivers of cancer invasion and metastasis^{4,25,26,30,35,36}.

Introduction to organotypic culture

Here, we briefly introduce three broad categories of 3D organotypic cultures: stem cell organoids, reaggregated spheroid cultures, and heterogenous tissue organoids.

Stem cell organoids.

Stem cell organoids are expanded from individual organ-specific primary stem cells, induced pluripotent stem (iPS) cells, or embryonic stem cells that expand and self-organize into lineage-restricted structures³⁷. These methods have been successfully applied to organoid isolation from normal and cancerous colon¹², pancreas³⁸, kidney^20,39, lung⁴⁰, esophagus⁴¹, and liver¹⁷. Lgr5, a receptor for Wnt ligands, often marks the stem cell population in adult epithelial tissues^12,14,18,39. Therefore, Wnt activators such as Wnt3A or R-spondins are essential to establish and maintain these organoid cultures³⁷. A small minority of cells isolated from a tissue are capable of generating stem cell organoids but do so only after several rounds of cell division.

Reaggregated cultures.

To accelerate organoid formation, 3D culture methods have also utilized reaggregation of primary cells or cell lines to generate spheroids or organoids. When maintained in suspension in the absence of an adherent surface or scaffold, cells will spontaneously assemble into 3D structures that recapitulate key aspects of in vivo tissue architecture^42,43. This approach has been successfully performed in diverse cell types, including kidney⁴⁴ and mammary epithelial cells⁴⁵. Reaggregation is also a convenient method to combine different cell types; for example, stem-cell derived hepatocytes were aggregated with mesenchymal stem cells and endothelial cells to recapitulate functional elements of a liver⁴⁶. A drawback of this method is that key chemical or spatial cues that drive self-organization of a tissue in vitro may be distinct from those in vivo.

Tissue organoids.

The third category of organoids, heterotypic tissue organoids, is the focus of the protocols described in this manuscript. Tissue organoids were first conceptualized by developmental biologists to distinguish the relative contributions of an epithelium and its surrounding mesenchyme towards organ patterning. Consequently, “purely” epithelial tissues were isolated using physical and enzymatic methods in a mesenchyme-free organoid culture model. These organoids are intact fragments of the originating tissue and therefore, maintain its genetic and cellular heterogeneity. When embedded in 3D matrices, these tissue organoids recapitulate the structure and function of the mammary gland^4,25,47, lung⁹, liver^17,48, kidney⁴², colon¹², and the prostate¹⁵. We have previously described our 3D culture methods for studying branching morphogenesis of the murine mammary epithelium⁴⁹. In this protocol, we have extended key principles of organoid generation from normal tissues to primary and metastatic tumors of the breast, liver, and pancreas. Since the basic principles of organoid isolation broadly apply to all the organ systems mentioned above, our troubleshooting suggestions are agnostic to the source of the tissue, unless explicitly stated. Consequently, these methods could readily be extended by users to most, if not all, types of epithelial tumors.

Experimental design

Isolation of tumor organoids.

In the first part of the procedure, we describe the methods for tumor digestion, stromal depletion, and organoid enrichment (Fig. 1, protocol steps 1–22). Briefly, tumors are physically minced and enzymatically digested using a combination of collagenase and trypsin. The composition of the digestion medium depends on the type of tissue; we provide optimized formulations for the digestion of murine or human primary and metastatic mammary tumors, genetically engineered mouse models (GEMM) of hepatocellular carcinoma (HCC) and for patient-derived xenograft (PDX) models for invasive breast cancer and pancreatic adenocarcinoma (PDAC). We use a series of fast and short differential centrifugation steps to deplete single cancer cells or stromal cells and enrich for epithelial organoids. We routinely generate 1–2,000 organoids from clinical samples, 25,000 from PDX breast tumors, and 250,000 from GEMM mammary tumors. Accordingly, we can simultaneously test multiple hypotheses with tissue from the same tumor and query the heterogeneity of responses of different regions of the same tumor. We also describe methods to efficiently transduce organoids with adenoviral and lentiviral particles (protocol steps 23–24), permitting the phenotypic evaluation of molecular changes in tumor organoids. We do not routinely passage or freeze down organoids. Instead, we only use freshly isolated primary tissue organoids for all our assays. We have recovered organoids from ECM and passaged 2–3 times with no apparent ill effect⁴. But we have never frozen and recovered organoids and cannot attest to whether this would change their behavior.

Figure 1: Workflow diagram.

A brief summary of the various assays described in this protocol including the isolation and culture of organoids, and downstream biochemical assays using 3D embedded organoids. An approximate timeline is also provided for each step of the protocol.

3D culture assays.

In the second part of the protocol, we describe 3D culture methods to assay growth, survival, invasion, local dissemination, and colony forming potential of tumor organoids (Fig. 1, protocol steps 29–54). These 3D culture assays also allow convenient variation of the molecular and mechanical properties of the extra-cellular matrix (ECM), the source of cancer cells (e.g. primary tumor vs metastasis), or the signaling environment (e.g. growth factors, chemokines, inhibitors). For example, we find that culture of primary tumor organoids in Matrigel is an ideal assay to measure their proliferative ability, since cancer cells rarely detach from the organoid bulk in this basement membrane-like ECM⁴. In contrast, tumor organoids extend invasion strands and locally disseminate into collagen I^4,25. Users can also adapt these 3D culture methods to precisely dissect features of tumor-stroma crosstalk by adding in specific stromal cells, such as fibroblasts or immune cells, into these cultures.

Biochemical and molecular biology assays.

In the third part of this protocol, we describe methods to adapt 3D cultures assays to downstream biochemical and molecular techniques including immunofluorescence, Western blotting, and flow cytometry (Fig. 1, protocol steps 60–85). We do not anticipate protocols for these downstream applications to change significantly based on the tissue source. We have had considerable success using these protocols for organoids isolated from murine, human, or PDX primary or metastatic-site tumors.

We have routinely compared protein expression from organoids embedded in Matrigel or collagen I and observed no significant or unexpected differences in protein expression levels or in the ability to detect protein phosphorylations. This is despite the two culture methods having slightly different extraction protocols but probably due the inclusion of phosphatase inhibitors in all our digestion and lysis buffers.

Collectively, the methods described in this manuscript can be used to answer fundamental questions in cancer biology and enable the high throughput testing of novel therapeutics. Tumor organoids are an effective model system to precisely define tumor and stroma-specific contributions to metastasis; therefore, the adaptation of these methods to different tumor types can be an effective research strategy.

REAGENTS

• Mice: All transgenic mice (MMTV-PyMT, C3(1)-Tag) described here were female and maintained on an FVB/n background. Breast PDX tumors were obtained from Jackson Laboratories and were passaged subcutaneously in NOD-SCID gamma (NSG) mice.

  CAUTION: Experiments involving rodents must be conducted in compliance with relevant institutional and governmental standards. Our mouse husbandry and procedures are in accordance with protocols approved by the Johns Hopkins School of Medicine Animal Care and Use Committee.

• Human samples: We obtain primary and metastatic human samples from the Cooperative Human Tissue Network in accordance with a study protocol (NA_00077976) that was acknowledged as exempt / not human subjects research by the Johns Hopkins School of Medicine Institutional Review Board.

  CAUTION: Experiments involving human samples must be in compliance with ethics review board regulations and must obtain informed consent, as applicable in their locale.

  CRITICAL After surgical resection, human mammary tumors should be stored in basal DMEM or RPMI based media at 4°C until they are ready to be processed.

Reagents for organoid isolation:

• D-PBS (Sigma, #D8662)

• Bovine serum albumin (BSA) solution (Sigma, #A9576)

• DMEM/ F-12 (Life Technologies, #10565–018)

• DMEM (Sigma, #D6546)

• RPMI-1640 (Life Technologies, #11875–093)

• Williams’ Medium E (Sigma, #W4128)

• Fetal bovine serum (FBS; Sigma, #F0926)

• Gentamycin (Life Technologies, #15750–060)

• Insulin (Sigma, #I9278)

• Collagenase from Clostridium histolyticum (Sigma, #C2139)

• Trypsin (Life Technologies, #27250–018)

• Penicillin streptomycin (Sigma, #P4333)

• GlutaMAX (Life Technologies, #35050–061)

• HEPES buffer (Life Technologies, #15630–080)

• Deoxyribonuclease (DNase; Sigma #D4263)

• Amphotericin B (Life Technologies; #15290–018)

Reagents for embedding tumor organoids in 3D matrix:

• Growth factor reduced Matrigel (BD Biosciences, #354230)

• Rat tail collagen I (BD Biosciences, #354236)

• 10X DMEM, low glucose (Sigma, #D2429)

• Sodium hydroxide (Sigma, #S2770)

CAUTION: Sodium Hydroxide can cause severe skin burns and eye damage. Please handle with appropriate safety gear.

Reagents for growth, invasion, and colony formation assays:

• Basic human fibroblast growth factor (FGF2; Sigma, #F0291)

• Human epidermal growth factor (EGF; Sigma, #E9644)

• DMEM/ F-12 (Life Technologies, #10565–018)

• Penicillin streptomycin (Sigma, #P4333)

• Insulin-Transferrin-Selenium-Ethanolamine (Life Technologies, #51500–056)

• DMEM (Sigma, #D6546)

• GlutaMAX (Life Technologies, #35050–061)

• HEPES buffer (Quality Biological Inc., #118–089-721)

• Bovine serum albumin (BSA; Sigma, #A9576)

• Cholera Toxin (Sigma, #C8052)

CAUTION: Cholera toxin can cause acute toxicity (oral, dermal, inhalation). Please handle with appropriate safety gear.

• Hydrocortisone (Sigma, #H0396)

• Insulin (Sigma, #I9278)

• Fetal bovine serum (FBS; Sigma, #F0926)

• Phosphate buffered saline (PBS, without Ca²⁺, Mg²⁺; Life Technologies, #10010–023)

• TrypLE enzyme (Life Technologies, #12604–013)

Reagents for immunofluorescence in organoids:

• Dulbecco’s phosphate buffered saline (DPBS, with Ca²⁺ and Mg²⁺; Sigma #D8662)

• Paraformaldehyde (Electron Microscopy Sciences, #15714-S)

  CAUTION: Formaldehyde can cause cancer. Please handle with appropriate safety gear.

• Fetal bovine serum (FBS; Sigma, #F0926)

• Bovine serum albumin (BSA; Sigma, #A9576)

• Triton X-100 (Sigma, #X100)

  CAUTION: Triton-X can be corrosive and cause acute toxicity. Please handle with appropriate safety gear.

• Tissue-Tek® O.C.T. compound (VWR, #25608–930)

• Primary and secondary antibodies

Reagents required for protein isolation from 3D embedded tumor organoids:

• cOmplete™ mini protease inhibitor cocktail (Sigma, #11836153001)

• PhosSTOP phosphatase inhibitor cocktail (Sigma, # 4906837001)

• EDTA (Sigma, #EDS)

• Dulbecco’s phosphate buffered saline (DPBS, with Ca²⁺ and Mg²⁺; Sigma #D8662)

• Collagenase from Clostridium histolyticum (Sigma, #C2139)

• Glycerol (Sigma, #G9012)

• 10X RIPA buffer (Millipore, #20–188)

• Sodium dodecyl sulphate, 10% (SDS; ThermoFisher Scientific, #15553027)

• Millipore water

Reagents for preparing organoids for flow cytometry:

• Phosphate buffered saline (PBS, without Ca²⁺, Mg²⁺; Life Technologies, #10010–023)

• TrypLE enzyme (Life Technologies, #12604–013)

• Propidium iodide (ThermoFisher Scientific, #P1304MP)

Reagents for transduction of tumor organoids:

• Adenoviral particles (desirable titer ~10⁸ PFU/ml)

CAUTION: Adenovirus must be used only in BSL-2 laboratory under a certified Biosafety Cabinet. Please handle with appropriate safety gear.

• ViroMag R/L transduction reagent (OZBiosciences, #RL40100)

• Puromycin (Gibco, #A11138–03) or another appropriate antibiotic

• Lentiviral particles (desirable titer >10⁷ IU/ml)

CAUTION: Lentivirus must be used only in BSL-2 laboratory under a certified Biosafety Cabinet. Please handle with appropriate safety gear.

Antibodies for characterization, for example we have used:

• Pan-cytokeratin (Abcam, #ab9377; RRID AB_307222)

• Phospho-histone H3 (PH3; Cell signaling, #9701; RRID AB_331535)

• E-cadherin (Cell Signaling, #3195; RRID AB_2291471)

• Keratin-14 (K14; Biolegend, #905301; RRID AB_2565048)

• GAPDH (Abcam, #ab9435; RRID AB_307275)

• pERK (Cell Signaling, #9101; RRID AB_331646)

• ERK (Cell Signaling, #4695; RRID AB_390779)

• Cleaved caspase 3 (CC3; Cell Signaling, #9664; RRID AB_2070042)

EQUIPMENT

Common equipment:

• Biological safety cabinet (LabGard® ES NU-540 Class II)

• Cell culture incubator maintained at 5% CO₂, 37° C (Thermo Fischer Scientific Heracell™ Vios 160i)

• High-speed centrifuges (Thermo Fischer Scientific Sorvall™ Legend™ XIR)

• Water bath maintained at 37° C (Thermo Fischer Scientific Precision™ Coliform Water Baths)

• Bright-field inverted microscope (Nikon, #TS2-S-SM)

• DIC microscope (Zeiss AxioObserver Z1 and AxioCam MRM, LD Plan-Neofluar 20x/0.4 Korr Ph2 objective)

• Laser-scanning confocal microscope (Zeiss LSM780)

• Temperature controlled bench-top shaker maintained at 37° C (Thermo Scientific MaxQ 4450)

• Laboratory-grade refrigerator maintained at 4° C (VWR, #GDM 49)

• Laboratory-grade freezer maintained at −20° C (Fischer Scientific, #13–986-428F)

• Laboratory-grade freezer maintained at −80° C (Thermo Scientific, Revco UxF)

• Tool sterilizer (Fine science tools, Steri 250)

• Sterile Spencer Ligature scissors (Fine science tools, #14028–10)

• Sterile Iris scissors (Fine science tools, #14060–09)

• Sterile Standard forceps (Fine science tools, #11003–12)

• Sterile Graefe forceps (Fine science tools, #11051–10)

• Sterile Fine-tip Precision forceps (Fine science tools, #11412–11)

• 10 cm Petri dishes (Corning, # 430293)

• Scalpel, #10 blade (Surgical Design Inc.)

• Dissection microscope (Zeiss SteREO Discovery V12 modular stereoscope)

• 0.22 μm sterile filter (Millipore, #SCGP00525)

• Pipette aid (Drummond, #177351)

• Serological pipettes – 5 ml (Genemate, #P2837–5), 10 ml (Falcon, #357551), 25 ml (Genemate, #P2837–25)

• Pasteur pipettes (Argos, #P0900)

• Calibrated pipettes – p10 (Gilman, # F144802), p100 (Gilman, #F123615), p200 (Gilman, #F123601), p1000 (Gilman, #F123602)

• Filter tips – 0.1–10 μl (Genemate, #P1237–10XL), 10–100 μl (Genemate, #P1237–100), 20–200 μl (Genemate, #P1237–200), 100–1000 μl (Genemate, #P1237–1250)

• Conical centrifuge tubes – 15 ml (Falcon, #352097), 50 ml (Falcon, #352098)

• Microcentrifuge tubes, 1.7 ml (Genemate, #C3260–1)

• Heating block (Benchmark Scientific, #BSH1002)

• 24-well black coverslip-bottomed plate (Greiner Bio-One, #662892)

• 96-well black coverslip-bottomed plate (Greiner Bio-One, #655892)

• 10 μm cell strainer (PuriSelect, #43–50010-01)

• 100 μm cell strainer (PuriSelect, #43–50300-01)

• Falcon™ test tube with cell strainer cap (Corning, #352235)

• Ice

• Ice bucket

Equipment required for colony formation assay and/ or flow sorting organoids:

• Flow cytometer (Beckman Coulter MoFlo Cytometer)

• Standard Hemocytometer (Fischer Scientific, #02–671-6)

• Falcon™ test tube with cell strainer cap (Corning, #352235)

• 96-well cell-repellent microplate (Greiner Bio-One, #655970)

Equipment required for immunofluorescence in organoids:

• Plus-gold glass slides (Thermo Scientific, # FT4981GLPLUS)

• Disposable base molds (Fischer Scientific, #22–363-553)

• Precision coverslips, #1.5H (Marienfeld, #0107222)

• Cryostat at a cutting temperature of −27° C (Thermo Fischer, Cryostar™ NX70)

• Staining tray with a black lid (Simport, #M920–2)

• Slide boxes (VWR, #82003–406)

Equipment required for protein isolation from 3D embedded tumor organoids:

• Conical tube shaker maintained at 4° C (Bechmark Scientific, #15090078)

Equipment required for transduction of tumor organoids:

• Magnetic plate (OZ Biosciences, #MF10000)

• 96-well cell-repellent microplate (Greiner Bio-One, #655970)

REAGENT SETUP

Collagenase stock solution:

Dissolve 1 g of collagenase in 10 ml of DMEM-F12. Aliquot (200 μl) and store at −20° C (stable for at least 3 months). Do not refreeze.

CAUTION: Do not filter sterilize collagenase stock solutions as it will clog the filters.

Trypsin stock solution:

Dissolve 1 g of trypsin in 10 ml of DMEM-F12. Aliquot (200 μl) and store at −20° C (stable for at least 3 months). Do not refreeze.

DNase stock solution:

Dissolve 2,000 U DNase in 1 ml of PBS, no Ca²⁺ or Mg²⁺. Aliquot (40 μl) and store at −20° C (stable for at least 3 months). Do not refreeze.

BSA coating solution:

Make 2.5% vol/vol BSA solution in D-PBS. Filter sterilize (0.22 μm). Store at 4° C. The solution can be reused several times for ~6–8 weeks, as long as it is kept sterile.

Digesting solutions: see table 1.

Table 1:

Composition of digesting medium depending on the type and source of tissue

Tissue type	Murine mammary tumors	Murine lung metastases	Murine liver tumors	Human mammary tumors (primary or metastatic)	Breast PDX tumors	Pancreatic PDX tumors
Total volume	30 ml	10ˆ ml	20 ml	20 ml	20 ml	20 ml
Collagenase Stock = 100 mg/ml	2 mg/ml (600/ 200ˆ µl of stock)		2 mg/ml (400 μl of stock)	2 mg/ml (400 μl of stock)	4mg/ml (800 μl of stock)	2mg/ml (400 μl of stock)
Trypsin	2 mg/ml (600/ 200ˆ μl of stock)		2 mg/mL (400 μl of stock)	-	-	-
Antibiotic type	Gentamycin		Gentamycin	Pen-strep	Pen-strep	Pen-strep
Antibiotic concentration	50 μg/ml (30/ 10ˆ μl of 50 mg/ml stock)		50 μg/ml (10 μl of 50 mg/ml stock)	1% vol/vol (200 μl)	1% vol/vol (200 μl)	1% vol/vol (200 μl)
FBS	5% vol/vol (1.5/ 0.5ˆ ml)		5% vol/vol (1 ml)	5% vol/vol (1 ml)	5% vol/vol (1 ml)	5% vol/vol (1 ml)
Insulin Stock = 10 mg/ml	5 μg/ml (15/ 5ˆ μl of stock)		5 μg/ml (10 μl of stock)	5 μg/ml (10 μl of stock)	-	5 μg/ml (10 μl of stock)
GlutaMAX	-		-	1% vol/vol (200 μl)	1% vol/vol (200 μl)	1% vol/vol (200 μl)
HEPES buffer	-		-	-	10 mM (200 μl of 1M stock)	NA
Medium	DMEM/F12		Williams’ medium E	DMEM	RPMI-1640	DMEM
Shaker speed (37°C)	180 rpm	150 rpm	150 rpm	200 rpm	150 rpm	150 rpm
Time	1 h	30 min - 1 h	1 h	30 min to 4 h *	1 h	1 h

^ˆVolumes of reagent for the digestion of murine metastatic tumors.

^*The time to digest the tumor will depend on its size and composition and could take anywhere from 30 minutes – 4 hours. For example, small and adipose-rich tumors typically digest quickly. We recommend checking the progress of the digestion every 20–30 minutes. Due to the smaller tissue size, we typically digest human tumors and murine metastatic tumors until few undigested fragments remain. These larger fragments can be removed using a p1000 pipet.

CRITICAL Prepare fresh before every use and filter sterilize (0.22 μm filter).

Culture mediums: see table 2.

Table 2:

Composition of culture medium depending on type and source of tissue.

Tissue type	Murine mammary tumors	Human mammary tumors	Mammary PDX tumors	Murine liver tumors	Pancreatic PDX tumors
Prepare medium with the following components and store for up to 3 months at 4°C
ITS	1% vol/vol	-	-	1% vol/ vol	1% vol/ vol
GlutaMax	-	1% vol/vol	1% vol/vol	-	1% vol/ vol
Penicillin-streptomycin	1% vol/vol	1% vol/vol	1% vol/vol	1% vol/ vol	-
Insulin	-	5 μg/ml	5 μg/ml	-	-
HEPES	-	10 mM	10 mM	-	-
BSA	-	0.075% vol/ vol	0.075% vol/ vol	-	-
Cholera toxin	-	10 ng/ml	10 ng/ml	-	-
Hydrocortisone	-	0.47 μg/ml	0.47 μg/ml	-	-
FBS	-	-	2% vol/ vol*	1% vol/ vol	10% vol/ vol
Media	DMEM/F12	DMEM	DMEM	DMEM/F12	DMEM/F12
Add the following components to the medium before every use
FGF2	2.5 nM	-	-	-	-
EGF	-	50 ng/mL	-	50 ng/mL	-

^*Amount of FBS may vary based on the PDX tumor. In our experience, most PDX tumors grow well at 2% FBS. Culture medium for PDX-derived organoids should be changed at least every 48 hours to maximize cell survival.

CRITICAL Prepare fresh before every use and filter sterilize (0.22 μm filter).

PBS/ Fungizone wash buffer:

Combine 2% vol/vol penicillin-streptomycin and 2% vol/vol fungizone to a total volume of 100 ml in DPBS. Store at 4° C for up to 3 months.

Triton-X permeabilization solution:

Triton-X (0.5% vol/vol) in D-PBS. Store at room temperature (typically 20–25 ° C) for up to 6 months.

Immunofluorescence blocking buffer:

Combine 10% vol/vol FBS, 1% vol/vol BSA, and 0.2% vol/vol Triton-X in D-PBS. Store at 4° C for up to 3 months.

Immunofluorescence antibody diluent buffer:

Combine 1% vol/vol FBS, 1% vol/vol BSA, and 0.2% vol/vol Triton-X in D-PBS. Store at 4° C for up to 3 months.

7X protease-phosphatase inhibitor stock:

Dissolve 2 tablets each of the cOmplete™ mini protease inhibitor cocktail and PhosSTOP phosphatase inhibitor cocktail in 2.856 ml of distilled water. Make 142.8 μl aliquots and store at −20° C for up to 6 months. Do not refreeze.

PBS-EDTA Matrigel dissolving buffer:

For every milliliter of Matrigel dissolving buffer, combine 142.8 μl of 7X protease-phosphatase stock, 20 μl of 500 mM EDTA, and 837.2 μl of D-PBS. Prepare fresh before every use.

Collagenase gel dissolving buffer:

Collagenase solution (2 mg/ml) in D-PBS. Prepare fresh before every use.

Protein lysis buffer:

For every ml of lysis buffer, combine 142.8 μl of 7X protease-phosphatase stock, 50 μl glycerol, 100 μl 10X RIPA buffer, 10 μl 10% SDS, and 697.2 μl distilled water. Prepare fresh before every use.

PROCEDURES

Part 1: Isolation of tumor organoids

CRITICAL The workflow described in this section is schematized in Fig. 2a,g and 6a,f. CRITICAL If using primary and metastatic tumors of human origin, start directly at step 2 as the starting tissue is already collected.

Figure 2: Isolation of organoids from primary murine and human mammary tumors.

(a) Workflow for isolating tumor organoids from mouse models of metastatic breast cancer. Isolated organoids are embedded in 3D matrices such as Matrigel or collagen I and cultured in the presence of FGF2 as a growth factor.

(b) Representative primary tumor freshly isolated from a 10–12 week-old MMTV-PyMT mouse, and after being minced (b’). Scale bar, 1 cm.

(c-d) After digestion with trypsin and collagenase, organoids are in suspension with stromal cells (c). Stromal cells are depleted, and epithelial organoids are enriched by performing short (3–4 seconds) differential spins (d). Scale bar, 100 μm.

(e) A representative murine tumor organoid. Non-epithelial fragments such as muscle can be observed in the final organoid suspension. Scale bar, 20 μm.

(f) Dotplot depicting the number of tumor organoids isolated from a gram of MMTV-PyMT or C3(1)-Tag tumors. Bar – median.

(g) Workflow for isolating primary tumor organoids from surgical samples from patients with invasive breast cancer. Isolated organoids are embedded in 3D matrices such as Matrigel or collagen I and cultured in the presence of EGF as a growth factor.

(h) Representative surgical specimen received. Yellow arrow marks an adipose-rich region. Scale bar, 1 cm.

(i) Representative micrograph of an organoid suspension isolated from a primary human tumor. Scale bar, 100 μm. A zoomed-in organoid in suspension is represented in the black box. Scale bar, 20 μm.

(j) Undigested stroma (S) can be observed in the final organoid suspension. A Masson Trichome stain shows tumor cells (T) entrapped within stromal regions. Scale bar, 50 μm.

(k) Dotplot depicting the number of tumor organoids isolated from a gram of primary human breast tumors. Bar – median.

Figure 6: Isolation of metastatic organoids from murine and human mammary tumor-derived metastases.

(a) Workflow for isolating organoids from metastatic lesions in mice. Organoids isolated from fluorescently labelled MMTV-PyMT primary tumors were trypsinized to small clusters, injected via the tail of non-fluorescent NSG mice. Lungs from these mice were harvested 4 weeks later. Fluorescently labeled metastases were micro-dissected and digested into metastatic organoids. These metastases are cultured in 3D matrices in the presence of FGF2.

(b) Metastatic organoids derived from MMTV-PyMT tumors form “branched” structures in 3D Matrigel. Scale bar, 100 μm.

(c) Organoids isolated from the MMTV-PyMT primary tumor invade robustly into collagen I. Scale bar, 100 μm.

(d) Metastatic organoids derived from MMTV-PyMT tumors rarely invade into collagen I. This is in contrast to the primary tumor organoids they were derived from. Left: More common non-invasive phenotype. Right: Rare invasive phenotype. Scale bar, 100 μm.

(e) Immunofluorescent images of a non-invasive and invasive metastatic organoids embedded in collagen I. Cells within the organoids were mTomato+ and keratin+. Scale bar, 100 μm.

(f) Workflow for isolating organoids from surgical specimens of metastatic lesions from breast cancer patients. Isolated organoids are cultured in 3D matrices in the presence of EGF as a growth factor.

(g) Representative images of primary human metastatic organoids isolated from various sites. Most of these metastatic organoids remain non-invasive in collagen I. Scale bar, 100 μm.

• 1
  Choose from the following options, depending on the tissue of origin. Collect and process primary mammary tumors, murine hepatocellular carcinoma (HCC) tumors or patient-derived xenograft (PDX) tumors of the breast or pancreas using option (a). Collect and process metastatic tumors using option (b).

  • a
    Murine mammary tumors, murine liver tumors, breast and pancreatic PDX tumors.

TIMING: 30 minutes

• iEuthanize the mice.

CAUTION: Mouse euthanasia should be carried out in accordance with ethically approved standards. We typically perform CO₂ asphyxiation followed by cervical dislocation to euthanize mice.

• iiSterilize the dissecting area with 70% ethanol. Sterilize dissecting tools with 70% ethanol or by using a glass bead sterilizer.
• iiiPlace the mouse face-up. Use standard forceps to lift up the skin right above the groin, and the Spencer Ligature scissors to make an initial incision. Then, extend this incision along the mid-line up to the chin of the mouse. Next, make lateral incisions to pull back the skin and expose the tumors.
• ivUse the dorsal side of the Graefe forcep to separate the tumors from the peritoneum. Use the Graefe forceps and Iris scissors to dissect out the tumors while taking care to avoid the surrounding tissues.
• vTransfer the dissected tumors (Fig. 2b) into a sterile Petri dish.

PAUSE POINT: Freshly dissected (uncut) tumors can be stored in basal media for up to 24 hours at 4° C. Longer storage times may affect organoid yield and viability.

• bMurine lung metastatic tumors.

TIMING: 4 weeks for generation of metastatic tumors (as described in step 59, option B). 2 hours for tissue collection and digestion. Isolation and culture could alternately be performed on spontaneous metastases. In that case, start at step ii.

1. Follow steps (i) to (iii) of option A.

   CAUTION: Mouse euthanasia should be carried out in accordance with ethically approved standards.

2. Make a ventral midline incision into the abdomen to expose the diaphragm muscle.

3. Lift up the xiphoid process by grabbing it with the Graefe forceps and cut the diaphragm by following the thoracic cage to release it from the ribs.

4. While still lifting the xiphoid process, cut the ribs on each side of the sternum to remove the thoracic cage.

   CAUTION: Care must be taken to not damage the lungs while cutting.

5. While grasping the heart with standard forceps, cut the dorsal attachments, and the trachea to remove lungs and heart en bloc.

6. Examine mouse lungs under a fluorescence dissecting microscope. Use a pair of precision forceps to isolate fluorescently labelled tumors from the lungs. Place dissected tumors into a conical tube with 10ml of DMEM-F12.

CRITICAL STEP: Care must be taken to prevent “contamination” with the surrounding lung tissue. Use high magnification and sharp tweezers to isolate metastatic tumors from surrounding non-fluorescent tissue.

Part 1 continued: Tissue digestion

TIMING: 30 minutes to 4 hours.

• 2OPTIONAL: Quickly wash the sample with ~5 ml PBS/fungizone.
• 3In a sterile biosafety cabinet, use a #10 scalpel to mince the tissue ~50–100 times until only small pieces of approximately 1 mm x 1 mm remain (Fig. 2b’). Tumor samples can be very variable in stiffness and fat content.

  CAUTION: Use standard BSL-2 precautions when working with human specimens.

• 4Transfer the minced tumor into a 50 ml conical tube with the appropriate digesting solution (Table 1).

  CAUTION: The volume of digesting solutions has been calculated for ~1 g of murine and PDX tumors, and ~0.5 g human tumors. Volumes should be adapted for differently sized tumors.

• 5Shake the suspension at 37°C on a benchtop shaker using the appropriate speed and time (Table 1).

CRITICAL STEP: Do not over-digest the sample. We recommend that the tissue be monitored closely as the end of the digestion time approaches. The sample should be digested until approximately 1 mm x 1 mm undigested fragments remain. Longer or shorter trypsinization can affect organoid yield and viability.

Part 1 continued: Organoid enrichment and enumeration

TIMING: 45–60 minutes.

CRITICAL The volumes used in the following steps have been calculated for ~1 g of tumor. For smaller tumors, such as human specimens, these volumes can be adjusted to 50–75%.

CRITICAL All steps in section of the protocol can be performed at room temperature, unless otherwise stated.

• 6Pipette a small volume of the digested solution into a Petri dish and check for the presence of organoids. Please see Fig. 2c for an example.
• 7Pellet the digested tissue by spinning at 400 g for 10 minutes at room temperature.
• 8Discard the supernatant.

CRITICAL STEP: If the tumor is adipose-rich (e.g. Fig. 2h), there will be a fatty layer on top of the supernatant. In order to maximize organoid yield, use a BSA-coated pipette to transfer this layer into a fresh BSA-coated 15 ml conical tube. Make up the volume to 10 ml with DMEM-F12. Pipet vigorously to disperse the fat layer into the solution. Pellet organoids by spinning at 400 g for 10 minutes. Continue with step 8. In all cases, Falcon/ Eppendorf tubes can be BSA-coated by quickly transferring an appropriate volume (1 ml for Eppendorf tubes and 10 ml for falcon tubes) of 2.5% BSA solution along the sides of the tubes and then removing the BSA solution as completely as possible. Pipets can be BSA-coated by drawing up and releasing 2.5% BSA solution.

• 9Add 8ml of DMEM-F12 and resuspend the pellet using a BSA-coated pipette. The suspension will now contain organoids attached to stromal cells. Add 80 μl of DNase (2 U/μl). Gently invert the tube several times for 3–5 minutes at room temperature.

CRITICAL STEP: If stromal contamination is observed in cultures (e.g. Fig. 4e⁵), increase the time of incubation with DNase by another couple of minutes. We hypothesize that as adipocytes are abundant in the mammary gland and that they lyse readily, large amounts of extracellular DNA maybe released by the stromal component. In tumors, an additional source of DNA might be necrotic regions. Removal of extracellular DNA results in organoids and single cells that readily separate on differential centrifugation.

Figure 4: 3D assays for growth, invasion, and dissemination of tumor organoids.

(a) Growth of tumor organoids can be measured by embedding in 3D Matrigel and culturing for 4–6 days in the presence of a growth factor.

(b) Organoids from a MMTV-PyMT tumor (b) or primary human tumor (b’) embedded in Matrigel are small and have a rounded morphology (left, Day 0). Over 4–6 days, they grow in size and form large “branched” structures (right, Day 5). Scale bar, 100 μm.

(c) Workflow of an invasion and dissemination assay. Collagen I is prepared by neutralizing rat tail collagen I and allowing collagen to polymerize at 4°C until it turns translucent. Neutralized collagen I is used to make underlays in glass-bottomed plates. Freshly isolated tumor organoids are resuspended in polymerized collagen I and plated at 37° C.

(d) MMTV-PyMT tumor organoids invade collectively into surrounding collagen I (d) while C3(1)-Tag tumor organoids both invade and disseminate (d’). Scale bar, 100μm. Zoom-ins of invasion and disseminated units are in black boxes. Black arrows mark single disseminated cells, while yellow arrows label disseminated cell clusters. Scale bar, 50 μm.

(e) DIC micrographs depicting possible alternate phenotypes of tumor organoids in 3D collagen I. A tumor organoid with rounded or “branched” invasion strands (e1). A tumor organoid with no invasion strands (e2). A tumor organoid that died during the course of the culture (e3). A tumor organoid that made contact with the coverslip. Black box zoom-in depicts the difference between the 3D vs 2D regions of the structure (e4, Scale bar, 50μm). Stromal cells within the organoid extending a protrusion into collagen I which is zoomed into in the white box (e5, Scale bar, 50μm). Scale bar, 100 μm.

(f) Primary human breast tumor organoids exhibit a wide spectrum of phenotypes in collagen I. From left to right: a non-invasive, disseminative, collectively invasive, and both invasive and disseminative organoids. Black arrows mark disseminated units. Scale bar, 100 μm.

• 10Pellet the digested tissue by spinning at 400 g for 10 minutes at room temperature.
• 11Discard the supernatant.
• 12Add 25–30 ml DMEM-F12 to the pellet. Use a BSA-coated pipet to gently mix the contents several times.

CRITICAL STEP: Vigorous pipetting can result in shedding of single cells from tumor organoids.

• 13Allow larger undigested particles to settle at the bottom of the tube (~ 5–10 seconds).
• 14Transfer the supernatant into a fresh BSA-coated tube using a BSA-coated pipet.
• 15Pellet the organoids by spinning at 400 g for 3 seconds at room temperature.

CRITICAL STEP: This step uses differential centrifugation to enrich for epithelial organoids and eliminate single cancer cells and other stromal cells. If recovery of the tumor stromal population is desired, the supernatant can be saved in a 50 ml conical tube after completion of each differential spin.

• 16Aspirate the supernatant.

CRITICAL STEP: Due to the short spin, the pellet will be loosely settled. It is recommended to leave a small volume (~1 ml) of the supernatant behind to prevent loss of organoids.

• 17Gently resuspend the pellet in 10 ml of DMEM-F12 using a BSA-coated pipet.
• 18Repeat steps (15–17) three more times. At this stage, the pellet will largely consist of tumor organoids and single cells should be depleted (Fig. 2d,e,i). In some cases, the organoid suspension may be contaminated with muscle (can be observed in mouse tumor samples, see Fig. 2e). or fibrous stroma (common for human tumor samples, see Fig. 2j).

CRITICAL STEP: Large amounts of fibrous stroma may degrade 3D ECM gels. They can be depleted by filtering the organoid suspension through a 100–300 μm filter.

PAUSE POINT: Isolated organoids can be kept in basal medium for up to 6 hours at 37°C. They can also be kept for up to 24–48 hours at 37° C when plated into cell-repellent plates in the presence of growth factor containing organoid medium.

• 19Resuspend pellet in 10 ml of DMEM-F12.

If you wish to isolate small organoids, which can be desirable for certain downstream applications (Eg: to increase viral transduction efficiency), pellet large organoids (>300 cells; ~300 – 500 μm) by spinning at 100 g for 3 seconds. Transfer the supernatant, which largely consists of small organoids (<200 cells; ~ <100 μm), into a fresh BSA-coated 15 ml conical tube.

• 20Estimate organoid density by transferring a small volume (10–50 μl) of the organoid suspension (well-mixed) into a clean Petri dish, using a BSA-coated tip. CRITICAL STEP The volume of sample used for counting in the next step should vary depending on the organoid density in the original suspension.
• 21Count the number of organoids within the sample under an inverted microscope. Estimate the total organoid yield in the remaining volume.

CRITICAL STEP: This step must be completed as soon as the small volume is aliquoted since organoids can clump together within a couple minutes, making the count unreliable.

• 22Estimate the required volume of organoid suspension for any downstream application and transfer into a BSA-coated Eppendorf tube for storage.

PAUSE POINT Store at 37°C until ready for use. We do not recommend leaving organoids in basal medium for longer than 2 hours.

Part 1 continued: Transduction of tumor organoids (optional)

• 23If you wish to transduce the organoids, follow option A to use adenovirus and option B for lentivirus.

  1. Adenoviral transduction of tumor organoids

     CRITICAL The workflow for transducing primary murine, human, or PDX tumor organoids with adenovirus is depicted in Fig. 3a.

     TIMING: 1 hour to set up the assay with an overnight incubation step.

     1. Isolate small sized (<200–300 cells) tumor organoids.

     2. Estimate organoid density. Pellet the desired number of organoids by spinning at 400 g for 5 minutes at room temperature.

     3. Thaw adenoviral particles on ice for at least 30 minutes.

     4. In a fresh tube, combine appropriate organoid medium for your tissue without FBS and growth factor (Table 2) with adenoviral particles at a concentration of ~25,000 plaque forming units per organoid. This concentration should infect ~50–75% of the cells. We recommend using 2 μl of adenoviral particles (10¹⁰ PFU/ml) to achieve a ~80% recombination efficiency for every 800 organoids.

     5. Resuspend pelleted organoids with the above virus-containing basal organoid medium at a concentration of 800–1000 organoids per 50 μl of media.

     6. Plate 50 μl of the above organoid suspension per well in a cell-repellent 96-well plate.

     7. Incubate the 96-well plate overnight at 37° C, 5% CO₂.

     8. Transfer the organoid suspension into fresh BSA-coated Eppendorf tubes. Pellet organoids by spinning at 400 g for 10 seconds at room temperature to get rid of detached single cells or cell debris.

     9. Resuspend in organoid medium, estimate organoid density and proceed to desired other downstream applications.

  2. Lentiviral transduction of tumor organoids

     CRITICAL The workflow for transducing primary murine, human, or PDX tumor organoids with lentivirus is depicted in Fig. 3c. Lentiviral transduction of organoids requires them to be maintained in suspension for several days. During this time, some cell death is expected. However, organoids recover quite well once they are plated in 3D matrices. Nevertheless, it is essential to always incorporate non-targeting controls which are treated identically to account for these effects.

     TIMING: 5 hours to set up the assay, 24–48 hours for organoid recovery, 72–96 hours for selection.

     CAUTION: Working with lentiviral particles corresponds to biosafety level 2. All procedures need to be performed in accordance with appropriate biosafety guidelines.

     1. Day 0: Transduction. Isolate small sized (<200–300 cells) tumor organoids.

     2. Estimate organoid density. Pellet the desired number of organoids by spinning at 400 g for 5 minutes at room temperature. We recommend using at least 1600 organoids per condition.

     3. Thaw lentiviral particles on ice for at least 30 minutes.

     4. In a non-BSA coated Eppendorf tube, combine magnetic nanoparticles (ViroMag), lentiviral particles, and corresponding organoid medium without FBS and growth factor (Table 2). We recommend using 3 μl ViroMag/ 50 μl volume. We also recommend using lentiviral particles at a target MOI of 8–20. For 800 organoids with an average size of 150 cells/ organoid, we require 12 μl of virus at a titer of 10⁸ TU/ml to attain a MOI = 10. This is according to the formula:

        $$Volumeofvirus(\mu l)=\frac{\#orgs/well\times avg.\#ofcells/org\times desiredMOI\times 1000}{viraltiter\left(\frac{TU}{ml}\right)}$$

     5. Incubate viral particles with ViroMag for 30 minutes at room temperature.

     6. Pellet desired number of organoids for each condition by spinning at 400 g for 5 minutes at room temperature. If helpful, the average number of single cells per organoid specific to individual preparations can be determined by pelleting 200–400 organoids at 400 g, then dissociating them into single cells by resuspension in 200 μl TrypLE followed by 15–20 minutes incubation at 37°C with intermittent vigorous pipetting. Single cells can then be counted using a cell counting chamber.

     7. For every 800 organoids, resuspend in 50 μl of the virus-ViroMag mixture.

     8. Plate 50 μl of the above organoid suspension per well in a cell-repellent 96-well plate.

        CRITICAL POINT: It is essential to include at least one well (800 organoids resuspended in 50 μl of basal organoid medium) without virus. This will be an essential negative control for confirming sensitivity of tumor organoids to antibiotics used for selection.

     9. Incubate the 96-well plate on top of a magnetic plate at 37° C, 5% CO₂ for 90 minutes.

     10. Remove magnetic plate. Incubate the 96-well plate at 37° C, 5% CO₂ overnight.

     11. Day 1: Organoid recovery. Gently add 150 μl of organoid medium to each well. Allow organoids to resettle at the bottom of the well for 30–60 minutes.

     12. Carefully remove ~150 μl of medium from each well. Ensure that there are no organoids in the medium to avoid loss.

     13. Add in 200 μl of pre-warmed growth factor containing organoid medium to each well.

     14. Use BSA-coated pipet tips to disperse the organoids in the well.

     15. Day 3/4: Selection. Carefully remove ~200 μl of medium from each well. Ensure that there are no organoids in the medium to avoid loss.

     16. Add in 200 μl of pre-warmed growth factor containing organoid medium with 1.5–3 μg/ml puromycin to each well. CRITICAL STEP: The concentration of antibiotic used, and the duration of selection may vary depending on the cell type. We recommend determining the dose of antibiotic necessary for the selection by doing a dose response on non-infected cells. This dose must induce 100% cell death in 3–4 days.

     17. Use BSA-coated pipet tips to disperse the organoids in the well. Incubate organoids in a cell culture incubator for 3–4 days.

     18. Day 6/7: Preparation of organoids for downstream application Use BSA-coated pipet tips to transfer organoids into BSA-coated Eppendorf tubes.

     19. To avoid retaining significant amounts of detached single cells, pellet the organoids by spinning at 400 g for 10 seconds at room temperature.

     20. Resuspend in organoid medium, estimate organoid density and proceed to desired downstream application.

Figure 3: Viral transduction of organoids.

(a) Workflow for transducing organoids with adenoviral particles. Organoids are transduced in suspension overnight.

(b) Representative flow cytometry histograms demonstrating the efficiency of adenoviral transduction (adeno-GFP) in murine tumor organoids (>80%).

(c) Workflow for transducing organoids with lentiviral particles. Organoids are transduced in suspension in the presence of magnetic nanoparticles overnight, allowed to recover for another 24 hours in the presence of growth factor containing organoid media. Successfully transduced organoids are selected using an antibiotic for 3–4 days. Representative flow cytometry histograms demonstrating the efficiency of lentiviral transduction in murine tumor organoids. Successfully transduced cells express EGFP. Transduction efficiency at MOI = 10 was found to be ~46% and ~23% in the presence and absence of magnetic nanoparticles respectively.

Part 2: 3D culture assays using tumor organoids

CRITICAL: We recommend placing a coverslip-bottomed plate at 37° C up to 30 minutes before plating organoids.

• 24Pellet desired number of organoids in a BSA-coated Eppendorf tube by spinning at 400 g for 5 minutes at room temperature.

  CRITICAL STEP: Tumor organoids can have significant dependence on paracrine signals regulating, for example, growth or invasion. Therefore, the density of organoids within the 3D ECM gel can be important. We recommend 30–60, 120–150, and 750–1000 organoids per well in a 96-, 24-, and 6-well plate respectively. These numbers are approximately doubled for organoids isolated from PDX tumors since they can be more sensitive to such paracrine signals.

• 25Set a heating block to 37°C.
• 26To carry out an organoid growth assay, follow option A. To undertake an organoid invasion and dissemination assay in 3D collagen I follow option B. To undertake an ex vivo in 3D Matrigel or in vivo tail vein colony formation assay follow option C followed by option D or E respectively.

A. Organoid growth assay: 3D Matrigel

TIMING: 1.5 hours to set up the assays, 5 days for end-point measurements

CRITICAL Matrigel is a basement-membrane rich matrix that primarily consists of laminin, collagen IV, and entactin. For all our assays, we use only growth-factor reduced Matrigel. This gives the flexibility to define growth conditions more precisely for organoids and also significantly limits batch-to-batch variability. The workflow for plating organoids in Matrigel is depicted in Fig. 4a.

1. Matrigel is stored at −20° C and needs to thaw to 4° C. Allow 2–3 hours at 4° C for it to thaw prior to embedding organoids.

   CRITICAL STEP: Matrigel must be kept on ice at all times.

2. Place the Eppendorf tube with pelleted organoids on ice.

3. Gently resuspend organoids with the appropriate amount of Matrigel based on the desired number and size of wells used for culture. We recommend 30–50 μl, 100–120 μl, and 750–1000 μl per well in a 96-, 24-, and 6-well plate respectively. It is imperative to pipet in ~10% excess volume of Matrigel due to significant amounts of pipetting loss. It is also recommended to use cold (stored at 4° C) pipet tips while working with Matrigel to minimize loss.

   CRITICAL STEP: Pipet very gently to avoid creation of bubbles.

4. Incubate the plate at 37°C, 5% CO₂ for 30–60 minutes.

5. Add growth medium appropriate for the tissue type (see Table 2).

6. Return the plate to 37° C, 5% CO₂ for the duration of the assay. We recommend culturing tumor organoids for 5 days.

B. Organoid invasion and dissemination assay: 3D collagen I

TIMING: 1.5 hours to set up the assays, 5 days for end-point measurements

CRITICAL Fibrillar collagen I is a structural protein that is enriched in the stroma of invasive carcinomas. We have previously shown that collagen I induces an invasive program in mammary tumor organoids⁴. For our assays, we use collagen I that is acid-solubilized from rat tails. We neutralize and slowly polymerize this collagen at low temperatures to form a fibrillar matrix. The collagen matrix is sensitive to physical and biochemical changes including pH, temperature, and collagen concentration⁵³. We prepare our collagen at a concentration of ~3 mg/ml and slow-polymerize it at 4° C. The workflow for plating organoids in collagen is depicted in Fig. 4c.

CRITICAL Perform steps i-vii on ice.

1. For a ~3.5 ml collagen I solution, begin by combining together 375 μl of 10X DMEM (with phenol red) and 100 μl of NaOH in a 15 ml conical tube. Mix thoroughly. The solution will be a dark pink color.

2. Gently add in 3 ml of 3–3.5 mg/ml rat-tail collagen I. Due to its high viscosity, it takes several attempts to homogenize the solution to a uniform color.

3. Once a stable color is reached, use small volumes (1–10 μl) of NaOH or collagen I to attain a light pink or salmon pink color corresponding to a pH of ~7. If the solution is dark pink (Fig. 4c’), it is basic and needs to be titrated with collagen I or 10X DMEM. If the solution is yellow (Fig. 4c’), it is acidic and needs to be titrated with NaOH.

   CRITICAL STEP: It is essential to attain the optimal pH for collagen I since it will affect the polymerization time and fiber density.

4. Use the neutralized collagen to make underlays on a coverslip-bottomed imaging plate. We recommend 5 μl, 15 μl, and 75 μl per well in a 96-, 24-, and 6-well plate respectively.

5. Place the imaging plate with the underlays at 37°C for 30–60 minutes prior to plating.

6. Allow the collagen to polymerize on ice for 30–120 minutes until it is cloudy and fibrous. Collagen fibers can be visualized under an inverted microscope.

   CRITICAL STEP: Collagen polymerization times can vary significantly and the polymerization state of the collagen can have a strong effect on cell behavior⁵³. It is essential to check the progress of polymerization every 15–20 minutes. Do not let the collagen over-polymerize, as it will form a highly viscous liquid that is very hard to pipet.

7. Place the Eppendorf tube with pelleted organoids on ice. Place the culture plate with underlays on a 37°C heating plate.

8. Gently resuspend organoids with the appropriate amount of polymerized collagen based on the desired number and size of wells used for culture. We recommend 30–50 μl, 100–120 μl, and 750–1000 μl per well in a 96-, 24-, and 6-well plate respectively. Carefully add organoids resuspended in collagen I to the bottom of a well.

   CRITICAL STEP: Pipet very gently to avoid creation of bubbles.

   CRITICAL STEP: It is important for the plate to be pre-warmed to 37°C so that the organoids are “trapped” in the upper layer and will not migrate along the interface between the underlay and upper layer of collagen.

9. Incubate the plate at 37°C, 5% CO₂ for 30–60 minutes.

10. Add media (see Table 2).

    CRITICAL STEP: High pipetting pressure or speed when adding media can result in the gel detaching from the coverslip.

11. Return the plate to 37°C, 5% CO₂ for the duration of the assay. We recommend culturing tumor organoids for 3–6 days. The number of days required for organoids to invade may vary for each tumor, but invasion strands are visible typically within 48 hours.

C. Dissociation of organoids into small clusters for colony formation assays

CRITICAL This is required prior to Ex vivo in 3D Matrigel (option D) or in vivo tail vein (option E)

TIMING: 3 hours.

CRITICAL Tumors are heterogeneous with phenotypically distinct populations. An essential step of metastasis is tumor cell seeding and outgrowth. We have previously shown that tumor cell clusters are highly efficient metastatic seeds²⁶. In our experience, the ex vivo colony formation assay described below models the survival and proliferative processes involved in expansion from a metastatic seed to a growing metastasis (schematic of assay is depicted in Fig. 5a). These would be the cellular phenotypes required by cancer cells to metastasize when injected via the tail vein of a mouse, an in vivo assay to measure colony formation potential.

Figure 5: 3D colony formation assay to model metastasis formation.

(a) Workflow for a colony formation assay. Freshly isolated murine tumor organoids are trypsizined into a heterogeneous mixture of 1–50 cell clusters. This suspension is flow sorted to isolate single cells or 2–5 cell clusters. If single cells were isolated, they can be reaggregated overnight in a cell-repellent dish to form 2–5 cell clusters. Cell clusters are then embedded in 3D Matrigel and colony formation is assessed 7 days after culturing in the presence of FGF2.

(b) Left: Single cancer cells directly embedded into Matrigel rarely form colonies. Scale bar, 100μm. A representative image of a single cell at the end of 7 days in culture is zoomed into the black box. Scale bar, 10μm. Right: Confocal images of fluorescently labeled single cells when the culture was started (top, Day 0) and at the end of 7 days in culture (bottom, Day 7). Scale bar, 10 μm.

(c-d) Representative images of colonies arising from reaggregated clusters (c) or from flow sorted cell clusters (d). Scale bar, 100μm. Representative colonies are zoomed into in black boxes. Scale bar, 50μm. Confocal images of fluorescently labeled small clusters (top, Day 0) and corresponding colonies arising after culture (bottom, Day 7). Scale bar, 20 μm.

(e) Workflow for tail vein injections to assess colony formation potential in vivo. Fluorescently labeled MMTV-PyMT cancer cells are trypsinized into small clusters and injected into non-fluorescent NSG host mice. Lungs from these mice are harvested after 3–4 weeks and number of metastatic colonies counted using a dissection microscope.

1. Resuspend pelleted organoids in 5–10 ml of PBS without Ca²⁺ or Mg²⁺. Incubate at room temperature for 5 minutes with intermittent mixing while inverting the tube.

2. Pellet the organoids by spinning at 400 g for 5 minutes at room temperature.

3. Resuspend organoids in an appropriate volume of TrypLE. We recommend 1ml of TrypLE for 50–60,000 organoids. Incubate at 37°C for 5–10 minutes.

   CRITICAL STEP: Do not trypsinize organoids all the way to single cells. We recommend examining a small aliquot under an inverted microscope to estimate the extent of trypsinization.

4. Inactivate TrypLE with 10 ml of 5% FBS vol/vol in PBS without Ca²⁺ or Mg²⁺.

5. Pellet tumor cells by spinning at 400 g for 5 minutes at room temperature.

6. Resuspend tumor cells in 4 ml of DMEM/F12.

7. Add 40 μl of DNase (2 U/μl). Gently invert the tube several times for 1–3 minutes at room temperature. Inactivate DNase by adding 6ml of D-PBS.

8. Pellet the tumor cells by spinning at 400 g for 10 minutes at room temperature.

9. Discard the supernatant.

10. Resuspend the cells in 2–4 ml of DMEM-F12. Filter through a 40 μm flow tube.

11. Estimate the number of tumor cells or clusters. Adjust concentration to 10⁶ cells/ ml.

12. Flow sort desired number of cells. To flow sort cell clusters (2–5 cells), use the gating strategy depicted in Fig. 5a (clusters). Flow sort desired number of cell clusters into cold DMEM-F12 and pellet by spinning at 400 g for 5 minutes at room temperature. To flow sort single cells and reaggregate to small clusters use the gating strategy depicted in Fig. 5a (single cells). Flow sort desired number of cells in DMEM-F12 and pellet by spinning at 400 g for 5 minutes at room temperature. Resuspend in growth factor containing organoid medium at a concentration of 3×10⁵ cells/ml. Plate 100 μl of this suspension per well in a cell-repellent 96-well plate. Allow cells to aggregate overnight at 37°C, 5% CO₂. After 12–16 hours, pellet aggregated cells by spinning at 400 g for 5 minutes at room temperature.

CRITICAL STEP: A large nozzle size of a 100 μm or larger should be used to flow sort cell clusters to ensure maximal viability.

CRITICAL STEP: The ability of different cell types to aggregate might vary. If fewer clusters are observed, increase the density of cells per well. Round-bottomed cell-repellent plates can also be used to enhance aggregation.

D). Ex vivo colony formation assay in 3D Matrigel

TIMING: 1.5 hours to setup the assay, 1 week for colony enumeration

1. Remove Matrigel from storage at −20°C and thaw Matrigel to 4° C. Allow 2–3 hours at 4° C for it to thaw prior to plating.

   CRITICAL STEP: Matrigel must be kept on ice at all times.

2. Transfer plate onto a heating block set to 37°C.

3. Pipet appropriate volumes of cells/ clusters (from option C step Xiii) resuspended in Matrigel into each well of the pre-warmed plate. Gentle pipetting perpendicular to the plate will ensure that the matrix does not slide to edges of the wells and remains in the center.

   CRITICAL STEP: The number of cells or clusters plated per well can affect colony formation efficiency, likely due to paracrine effects. We recommend plating 30,000 small cell clusters per 100 μl of Matrigel plated in a 24-well plate.

4. Incubate the plate at 37°C, 5% CO₂ for 30–60 minutes.

5. Add media (see Table 2).

   CRITICAL STEP: High pipetting pressure or speed can result in the gel detaching from the coverslip.

   CRITICAL STEP: If testing efficacy of small molecules, they may be added at the desired concentration while adding media to the wells.

6. Return the plate to 37°C, 5% CO₂ for the duration of the assay. We recommend culturing tumor organoids for 7 days. Colonies can be counted under an inverted microscope after 7 days. We typically count structures > 50–75 μm in diameter (>10 cells) as a successfully formed colony.

E). In vivo colony formation tail vein assay

TIMING: 2 hours to setup the assay, 3–4 weeks for colony enumeration

1. Resuspend cells/ clusters (from option C step Xiii) in D-PBS to a final concentration of 10⁶ cells/ml (only cells in clusters are counted; single cells are ignored). We recommend using fluorescently labeled cells to make counting colonies in the lung much easier.

2. Inject 200 μl of the cell suspension via the tail vein of an immunocompromised, non-fluorescent, 6–10 weeks old, female NOD-SCID gamma (NSG) mouse. The tail vein injection will result in cancer cells seeding the lung.

   CAUTION: Any experiments performed in mice must be in accordance with government and/ or institutional regulations.

3. Allow 3–4 weeks for the formation of metastatic colonies in the lung.

4. Euthanize the mice. We typically perform CO₂ asphyxiation followed by cervical dislocation to euthanize mice. Harvest all four lobes of the lungs and quantify lung colonies under a fluorescence dissecting microscope.

   CAUTION: Mouse euthanasia should be carried out in accordance with ethically approved standards.

Part 3: Biochemical and molecular biology assays using 3D organotypic cultures:

• 27For immunofluorescence staining, follow option A. Organoids embedded in 3D ECM gels need to be extracted prior to some downstream applications. To extract organoids from Matrigel, follow option B. To extract organoids from collagen, follow option C. For protein extraction, follow option D. For flow cytometry analysis, follow option E. The workflow for these is described in Fig. 7c.

Figure 7: Downstream applications of 3D organotypic cultures – immunofluorescence, protein isolation, and FACS.

(a) Workflow for performing immunofluorescence in 3D cultured organoids. After fixation, organoids in 3D gels such as Matrigel and collagen can either be saved for immunofluorescence in-gel or embedded in O.C.T. to cryosection. Permeabilization, blocking, and antibody staining is then performed on organoids in 3D gels or cryosections.

(b) Representative immunofluorescent images of organoids embedded in 3D Matrigel or collagen I. Left to right (green): F-actin, phosphohistone-H3, E-cadherin, pan-cytokeratin, and keratin-14.

(c) Workflow for isolating 3D embedded organoids from Matrigel or collagen I gels prior to protein extraction or FACS. Organoids embedded in Matrigel are isolated using PBS-EDTA, while those in collagen are isolated using PBS-collagenase.

(d) Representative Western blots for various proteins using lysates from organoids embedded in Matrigel (left) and collagen I (right).

(e) Representative histogram demonstrating successful recovery of cells for FACS from organoids embedded in 3D ECM gel. Live/ dead assessment is based on PI staining.

A). Immunofluorescence staining of 3D embedded tumor organoids

TIMING: 10–48 hours

CRITICAL Due to the thickness of 3D matrix gels, antibody accessibility may be limited and cause poor immunofluorescence signal. However, the use of appropriate permeabilization and blocking buffers can enhance signal quality. The workflow for immunofluorescence of embedded organoids is depicted in Fig. 7a.

1. Carefully remove organoid medium from each well.

2. Wash once with D-PBS.

3. To fix Matrigel cultures, pre-warm 2% PFA to 37° C 10 minutes before use and then add 2% PFA to each well and incubate for 10–12 minutes at 37° C. To fix collagen cultures just add 4% PFA to each well and incubate for 10–15 minutes at room temperature. Then carefully remove PFA from each well and wash thrice with D-PBS for 10 minutes each. The washing with PBS can be performed on a benchtop-shaker at low speeds.

   CAUTION: PFA contains formaldehyde which can cause cancer. Please handle with appropriate safety equipment and protocols.

   CRITICAL STEP: Longer incubation times with Matrigel can cause it to completely dissolve. Alternatively, 4% PFA with 0.25% glutaraldehyde can be used to prevent Matrigel de-polymerization.

   PAUSE POINT: Culture plates can be stored with D-PBS in wells at 4° C for up to 2–4 weeks prior to immunofluorescence. If tumor organoids are fluorescent, wrap the plate in aluminum foil.

4. Immunofluorescence of fixed 3D cultures. To stain a whole gel, proceed direct to step Xii. CRITICAL Staining of organoids embedded in 3D gels can work well especially for highly expressed proteins or highly selective antibodies. Alternatively, staining of organoids embedded in 3D gels can also be performed after cryosectioning (steps v – Xi) into thin slices. This is particularly useful for lower quality antibodies or non-abundant proteins.

5. Use a pair of flat-head handling tweezers to detach the gel from the culture plate.

6. Transfer the gel into a plastic, disposable mold pre-lined with a small amount of O.C.T.

7. Freeze the mold for 10–30 minutes at −80° C.

8. Dust off any ice crystals with a soft-bristled paint brush and fill up the mold with O.C.T. such that the entire gel is covered. Freeze the mold at −80° C for at least 12 hours prior to sectioning.

   PAUSE POINT: Frozen molds will last for several years if maintained at −80° C.

9. For cryosectioning, adjust the instrument cutting temperature to −20 to −27° C. Transfer 10–60 μm sections on to plus-charged glass slides using a pair of forceps.

   PAUSE POINT: Slides can be stored at −80° C for several years.

10. To begin antibody staining, thaw slides in a black staining tray (with a small amount of water in the reservoir below the slides) for 2–5 minutes at room temperature.

    CRITICAL STEP: Depending on the organoid density in the original 3D gel and section thickness, it is possible that some sections will not have any organoids. In order to identify sections that contain organoids, we recommend pre-staining organoids with DAPI prior embedding in O.C.T., or examining sectioned slides using an inverted microscope.

11. Dissolve OCT by washing with D-PBS for 30–45 minutes at room temperature.

    CRITICAL STEP: For all steps that require pipetting solutions onto slides with sections, it is recommended to pipet around the section using a p100 or p200, and not directly above the section since it can cause sections to detach.

12. Permeabilize using 0.5% vol/vol Triton-X dissolved in D-PBS.

    CRITICAL STEP Incubation times may vary between in-gel and cryosection staining. For incubation time of this and the following steps, refer to Table 3

13. Remove Triton-X solution. Block with a 10% FBS/ 1% BSA/ 0.1% Triton-X (all percentages are vol/vol) dissolved in D-PBS.

14. Remove blocking solution. Add primary antibody dissolved in 1% FBS/ 1% BSA/ 0.1% Triton-X (all percentages are vol/vol) in D-PBS.

15. Remove antibody diluent. Wash thrice with D-PBS for 10 minutes each at room temperature.

16. Add secondary antibody dissolved in 1% FBS/ 1% BSA/ 0.1% Triton-X (all percentages are vol/vol) in D-PBS.

17. Remove antibody diluent. Wash thrice with D-PBS for 10 minutes each at room temperature.

18. Either store stained plates in D-PBS at 4° C or mount slides with 1–2 drops of mounting medium and a glass coverslip. Let these slides dry at room temperature or 37° C (in the dark). Once dried, they can be stored at 4° C.

Table 3:

Tabular summary of immunofluorescence staining of organoids

Step	Reagents	In-gel	Cryosections
Permeabilization	0.5% Triton X-100 in D-PBS	1 h	30 – 45 min
Blocking	10% FBS, 1% BSA, 0.1% Triton X-100 in D-PBS	3 h	1h
Primary antibody	1% FBS, 1% BSA, 0.1% Triton X-100 in D-PBS	Overnight at 4°C	3 h at RT (or) overnight at 4°C
Secondary antibody	1% FBS, 1% BSA, 0.1% Triton X-100 in D-PBS	3–4 h at RT	1 h at RT

RT – room temperature

PAUSEPOINT We recommend 3D gels stored in D-PBS be imaged within 48 hours and mounted slides within 48–72h.

B). Extracting organoids from Matrigel

TIMING: 30 to 45 minutes

1. Place culture plate on ice and carefully remove organoid medium from each well.

2. Wash once with ice-cold PBS without Ca²⁺ and Mg²⁺.

3. For every 100 μl Matrigel gel, add 2 ml of Matrigel dissolving buffer (see reagent setup for preparation details). Use a p1000 to disrupt the gel.

4. Transfer the contents of each well into a new ice-cold 15 ml conical tube. You may combine contents of multiple wells from the same experimental condition.

5. Incubate for ~10–30 minutes in a tube rotator at 4° C.

   CRITICAL STEP: The Matrigel needs to be completely dissolved before proceeding to the next step. Steps (iii)-(v) can be repeated if residual Matrigel persists. Residual Matrigel will skew protein quantification results. The incubation times should be kept to the minimum to ensure that protein stability and phosphorylation is unaffected. A phosphatase inhibitor can also be added to the PBS-EDTA buffer used to dissolve Matrigel.

6. Pellet organoids by spinning at 400 g for 5 minutes at 4° C.

C). Extracting organoids from collagen I

TIMING: 15 minutes

1. Pre-warm D-PBS and collagenase-PBS digestion buffer to 37° C.

2. Carefully remove organoid medium from each well.

3. Wash once with warm D-PBS.

4. For every 100 μl collagen I gel, add 500 μl of collagenase-PBS digestion buffer. Use a p1000 to disrupt the gel.

5. Shake the plate on a benchtop shaker set to 150 rpm at 37° C for 10–12 minutes.

   CRITICAL STEP: Do not over-digest the sample since it will affect structures of membrane proteins. It can also result in the premature lysis of cells and significant losses in protein yield.

6. Transfer contents of each well into an ice-cold Eppendorf tube. You may combine contents of multiple wells from the same experimental condition.

7. Pellet organoids by spinning at 400 g for 5 minutes at 4°C.

D). Protein isolation from 3D embedded tumor organoids

TIMING: 1 hour

CRITICAL: We routinely compare protein expression from organoids embedded in both Matrigel and collagen I. We have observed no significant or unexpected differences in protein expression levels or in the ability to detect protein phosphorylations in either matrix.

1. Aspirate the supernatant. Wash twice with ice-cold PBS. Transfer contents to an ice-cold Eppendorf tube. Each time, pellet organoids by spinning at 400 g for 5 minutes at 4°C.

2. Resuspend pellet in lysis buffer. The volume of lysis buffer will depend on the size of the pellet. We recommend resuspending ~100 organoids in 20 μl of lysis buffer.

3. Incubate on ice for 15–30 minutes with intermittent mixing (using gentle vortexing or pipetting).

4. Spin at 18000 g for 10 minutes at 4°C.

5. Transfer the supernatant to a fresh ice-cold Eppendorf tube.

PAUSE POINT Store at −80°C until ready for use. Protein lysates are typically stable for several months (may be longer).

E). Fluorescence activated cell sorting (FACS) using 3D embedded organoids.

TIMING: 1 to 2 hours

1. Extract organoids from 3D matrices as described above.

2. Pellet organoids by spinning at 400 g for 5 minutes at 4° C. Remove supernatant.

3. Resuspend pellet in appropriate volume of TrypLE. We recommend resuspending ~200 organoids in 100 μl of TrypLE Pipette gently with p200 several times to disrupt pellet. Incubate at 37° C for 4–7 minutes, pipetting periodically with p200 to assist digestion.

4. Pipette gently 5–7 times to disperse cells. Inactivate TrypLE with 1 ml of 5% FBS or 5% BSA vol/vol in PBS without Ca²⁺ or Mg²⁺.

5. Pellet cells by spinning at 400 g for 10 minutes at 4° C. Resuspend cell pellet in 500 μl PBS.

6. Stain cells according to desired assay protocol. For viability staining, add 1 drop of propidium iodide, gently vortex, and incubate for 5–15 minutes at room temperature.

7. Transfer cell suspension into 5 ml FACS tubes through the 35 μm strainer cap and procced to flow cytometry.

CRITICAL STEP: The FACS tube strainer cap prevents transfer of any residual cell clusters or matrix material that might clog the flow cytometer.

TIMING

Part 1: Isolation of tumor organoids.

Step 1: Tissue collection.

1. Murine mammary tumors, murine liver tumors, breast and pancreatic PDX tumors. 30 minutes

2. Murine lung metastatic tumor. 2 hours

Steps 2 to 5: Tissue digestion. 30 minutes to 4 hours

Steps 6 to 22: Organoid enrichment and enumeration. 45 to 60 minutes

Step 23: Transduction of tumor organoids.

1. Adenoviral transduction of tumor organoids. 1 hour to set up the assay with an overnight incubation step

2. Lentiviral transduction of tumor organoids. 5 hours to set up the assay, 24–48 hours for organoid recovery, 72–96 hours for selection

Part 2: 3D culture assays using tumor organoids.

Step 24 to 26:

1. Organoid growth assay (Matrigel). 1.5 hours and 30 to set up the assays, 5 days for end-point measurements

2. Organoid invasion and dissemination assay (Collagen I). 1.5 hours to set up the assays, 5 days for end-point measurements

3. Dissociation of organoids into small clusters for colony formation assays. 3 hours

4. Ex vivo colony formation assay. 1.5 hours to setup the assay, 1 week for colony enumeration

5. In vivo colony formation tail vein assay. 2 hours to setup the assay, 3–4 weeks for colony enumeration

Part 3: Biochemical and molecular biology assays using 3D organotypic cultures:

Step 27: Biochemical and molecular biology assays

1. Immunofluorescence staining of 3D embedded tumor organoids. 10–48 hours

2. Extracting organoids from Matrigel. 30 to 45 minutes

3. Exctracting organoids from collagen I. 15 minutes

4. Protein isolation from organoids. 1 hour

5. Fluorescence activated cell sorting (FACS) using 3D embedded organoids. 1 to 2 hours

ANTICIPATED RESULTS

The procedures described in this article have been optimized to provide highly reproducible results across experiments. We note that progress in the mincing of the tumor, digestion in enzymes, dispersal by pipetting, and differential centrifugation can all be monitored by visual inspection. Accordingly, it is useful to pay careful attention during the protocol to whether a particular tumor is behaving as expected or might need more or less time in a given step. This is particularly important with human tumors, as they have greater variability in cell and ECM content.

Murine mammary tumors that we have worked with extensively include the luminal MMTV-PyMT model and the basal C3(1)-Tag model. These mice develop late-stage invasive carcinoma around the ages of 12-weeks and 25-weeks respectively^50,51. Our median organoid yield from MMTV-PyMT and C3(1)-Tag is >110,000 and >50,000 organoids/ gram of tumor respectively (Fig. 2f), and the success rate for isolating organoids is 100% (across several hundred tumors). It is important to note that these late stage tumors can weigh >2 g, which typically yields several hundred thousand organoids. This highlights the potential use our assays in high-throughput settings. Our median yield for primary human mammary tumors is 2,628 organoids/ gram of tumor (Fig. 2k), and the success rate is ~86% (across ~80 tumors).

Tumor organoids can be genetically manipulated using adeno- or lentiviral vectors. In order to obtain high transduction efficiencies, smaller organoids (<200–300 cells) should be used since viral particles will have lower penetrance in larger organoids. We observe a 79% (±6.3%, r = 3) transduction efficiency when infected with adenovirus (Fig. 3b). We also recommend using magnetic nanoparticles (e.g. ViroMag) to increase lentivirus transduction efficiencies (26±5.5% and 43±8% transduced cells without and with ViroMag respectively, Fig. 3d). When lentiviral vectors include a fluorescent protein, flow sorting can be used instead of antibiotic selection as described in this protocol.

The phenotype of organoids in 3D matrices is greatly influenced by the composition of the ECM and the tumor itself⁴. We use Matrigel to model the proliferative capacity of the tumors as we rarely observe robust dissemination of primary tumor organoids in this matrix. Organoids from murine or human primary mammary tumors grow several fold and form branched-like structures in Matrigel (Fig. 4b,b’). In contrast, we use collagen I to assay for the invasive and disseminative capacity of primary tumors. MMTV-PyMT tumors primarily invade collectively, while C3(1)-Tag tumors both invade and disseminate into collagen I (Fig. 4d,d’). Dissemination can be as single cells or clusters (Fig. 4d’). Other possible, but less common, phenotypes include organoids exhibiting rounded/ bulky invasion strands (Fig. 4e¹), non-invasive organoids (Fig. 4e²), dead organoids (Fig. 4e³), organoids that make contact with the coverslip-bottom and expand in 2D (Fig. 4e⁴). Interestingly, human tumors exhibit a wide spectrum of invasion morphologies in 3D collagen I. Human mammary tumors can remain non-invasive, can primarily invade or disseminate, or can both invade and disseminate into collagen I (Fig. 4f). Invasion and dissemination in these cases can be mediated by both single cells and clusters of cells. The organoid assays described in this protocol provide a platform to identify the molecular drivers of breast cancer invasion and to search for conserved regulators of invasion across solid cancers.

We use Matrigel embedded tumor cell clusters to model micro-metastatic outgrowth. In our experience, this ex vivo colony formation assay effectively models the survival and proliferative properties of a metastatic seed²⁶ and predicts the results of tail-vein experimental metastasis models³⁰. Single cancer cells rarely form colonies when embedded in Matrigel (Fig. 5b) or when injected into the tail-vein of NSG mice (Fig. 5e’). In contrast, flow sorted cancer cell clusters or reaggregated cancer cells form several, large colonies when embedded in Matrigel (Fig. 5c–d) or when injected into the tail-vein of NSG mice (Fig. 5e’). These results are consistent with our in vivo observation that clusters of cancer cells have a ~100-fold higher efficiency in forming metastases²⁶. Since this assay closely models the micro- to macro-metastasis formation, it could readily be adapted for drug/ shRNA/ CRISPR screens to identify its molecular regulators (Fig. 5e”).

We also show results from isolating organoids from metastatic colonies arising from murine mammary tumors (Fig. 6a). Similar to the primary tumor they were derived from, metastatic organoids remain non-invasive in 3D Matrigel (Fig. 6b). Interestingly, both murine and human metastatic organoids were minimally to moderately invasive in collagen I (Fig. 6c,d,g), suggesting there may be additional regulators of metastatic invasion that are not yet modeled in the assay. We observe that cells in the murine metastatic organoids retain the fluorescent label from their “parent” primary tumor and retain expression of epithelial keratins (Fig. 6e). We also show results from organoids from human metastatic mammary tumors surgically removed from the lung, ovary, femur, spinal cord, soft-tissue, and axillary sites (Fig. 6f,g).

Tumor organoids embedded in 3D ECM gels can be made compatible with most biochemical and molecular biology applications. We describe our protocols for immunofluorescence, protein extraction and FACS using organoids embedded in Matrigel or collagen. Using these protocols, we have successfully detected proteins with different expression levels and different cellular localizations (F-actin, phosphor-histone H3, E-cadherin, pan-cytokeratin, and keratin-14; Fig 7b). In order to prevent protein quantification assays (e.g. Western blotting) from being skewed due to the presence of highly abundant ECM proteins, it is essential that all of the residual Matrigel or collagen I is dissociated away from the organoids. We have successfully detected proteins with different expression levels and phosphorylation states by Western blotting in both Matrigel and collagen I (actin, GAPDH, E-cadherin, keratin-14, phospho-ERK (p42, p44), total ERK, and cleaved caspase 3; Fig. 7d). Finally, we also describe our optimized protocol for isolating single cells from embedded organoids for FACS. These methods have allowed us to maintain over 90% cell viability during FACS (Fig. 7e).

We have successfully adapted our organoid isolation and culture protocols to murine models of HCC, and PDX models of breast and pancreatic cancers. Organoids isolated from murine HCC tumors do not disseminate in Matrigel and extend collective invasion strands in collagen I (Fig. 8a–c). Breast and PDAC PDX tumor organoids frequently maintain non-invasive borders in both Matrigel and collagen I (Fig. 8d–i). However, in rare cases, breast PDX organoids may disseminate into Matrigel, and breast and PDAC PDX organoids may invade into collagen I (Fig. 8e,f,h,i).

Figure 8: Application of methods to multiple model and organ systems.

(a) Transgenic mice were developed to express Myc under the control of liver-specific Liver Activator Protein. LAP drives the expression of a tetracycline transactivator (tTA), which is in the ‘OFF’ state in the presence doxycycline (Dox). Upon withdrawal of Dox, tTA binds to the tetracycline operator (tetO) in the promoter region to drive the expression of Myc. These mice develop localized HCC. Representative primary tumor from a mouse is shown. The tumor is mechanically and enzymatically digested to generate organoids. Organoids are cultured in a 3D matrix in the presence of EGF as a growth factor.

(b) Murine HCC tumor organoids in Matrigel do not disseminate. Scale bar, 100 μm.

(c) Murine HCC tumor organoids in collagen-I displaying collective cell invasion. Scale bar, 100 μm.

(d) Micrograph of a subcutaneously passaged triple negative breast PDX tumor isolated from an NSG mouse. The tumor is mechanically and enzymatically digested to generate organoids. Organoids are cultured in a 3D matrix in the presence of 2% serum.

(e) Organoids from breast PDX tumors commonly grow with non-invasive borders when embedded in 3D Matrigel (left). In rare cases, organoids can have several cells disseminate into the surrounding matrix. Arrows mark disseminated cells (right). Scale bar, 100 μm.

(f) Organoids isolated from a breast PDX tumor typically grows and remains non-invasive in 3D collagen I (left). In rare cases, organoids can invade collectively into the surrounding matrix (right). Scale bar, 100 μm.

(g) Micrograph of a subcutaneously passaged pancreatic adenocarcinoma PDX tumor isolated from a NSG mouse. The tumor is mechanically and enzymatically digested to generate organoids. Organoids are cultured in a 3D matrix in the presence of serum.

(h) PDAC PDX tumor organoids grow with a non-invasive border in 3D Matrigel. Scale bar, 100 μm.

(i) PDAC PDX tumor organoids display a high level of rotational movement and motility but typically remain non-invasive in 3D collagen I (left). In rare cases, organoids can invade collectively or exhibit a single-file invasion strands (right). Scale bar, 100 μm.

In summary, the protocols described here can be used to study various aspects of tumor progression, to test the efficacy of small molecule drugs, or to identify cellular or molecular functions of your protein of interest. These methods can be applied to murine, PDX, or primary patient-derived 3D culture models for breast, pancreatic, and liver cancer. We anticipate that these protocols could be adapted to most solid cancers. When we adapted our breast tumor protocols to liver and pancreatic tumors, we focused on optimizing the extent of mincing, time of digestion, extent of dispersal by pipetting, and the media composition. In most cases, we modeled the media from classic formulations for the embryonic tissue of origin (e.g. embryonic hepatocyte media for liver cancer). Finally, we note that it is critical to use 3D culture as a hypothesis generating platform and validate key conclusions in vivo and in human samples.

Extended Data

Extended Data Figure 1: Variables that affect organoid yield from mammary human tumor organoids.

(a) Organoid yield increased as the protocol was optimized during the course of the study.

(b) Variations in organoid yield based on modifications to the protocol.

Supplementary Material

FACS Gating Strategy

Supplementary Figure 1: Flow cytometry gating strategy

(a) Gating strategy for evaluating the percentage of infected cells based on GFP expression is depicted. Forward and side scatter density plot (FSC-A/SSC-A) is used for identifying the cell population and excluding debris. Then, single-cell population is selected using the forward scatter height versus forward scatter area (FSC-H/FSC-A). Finally, the gate for GFP-negative cells is set-up based on the non-infected cell population. Positive cells are evaluated after lentiviral infection with a plasmid containing GFP.

(b) Gating strategy for evaluating the percentage of viable cells based on PI staining is depicted. After selecting the cell population on the forward and side scatter density plot (FSC-A/SSC-A), doublets are excluded using the forward scatter height versus forward scatter area (FSC-H/FSC-A). The negative population is set-up on non-stained cells and the positive population is validated using fixed cells stained for PI as control.

Table 4.

Troubleshooting

Part	Step	Problem	Possible reason	Solution
I	1) b) (vi)	No metastasis into the lungs	Injection of single cells instead of clusters	Decrease the time of TrypLE incubation, pipet very gently to ensure that small clusters are not completely dissociated into single cells.
	6	No organoids are detected in the supernatant	Sample volume too small to detect any organoids	Count organoids from at least 150 – 200 μL from the supernatant.
			Digesting time too short or too long	Improper digestion of tissue will affect organoid yield. If pieces of tissue bigger than 1x1 mm remain, continue the digestion. If no tissue remains and no organoids are detected, the digesting time was probably too long and the organoids have been dissociated into single cells. It is important to check on the progression of digestion every 15–30 minutes.
	19	Presence of single cells	Pipetting too vigorously	Organoids can break down easily to single cells. Some tumor types can be more sensitive than others to vigorous pipetting. We recommend pipetting up and down very gently in all cases.
			Large amount of tissue	For big tumors we recommend increasing the amount of differential spins (step 17: 4 or 5 times instead of 3) to remove all the single cells.
		Presence of fibrous stroma	Fibrotic tissue	After the differential spins, filter the organoids using a 300 μm filter to remove fibrous stroma.
		Presence of muscle fibers	Improper dissection	When collecting the tissue, be careful to not collect the surrounding muscle.
		Low organoid yield	Tissue collection	Be careful to collect the maximum amount of tumor or metastases without taking any of the surrounding tissue
			Not enough mincing	Mincing the tumor into very small pieces is crucial for a good digestion since it allows even access to the enzymes. We recommend mincing the tissue until only small pieces 1x1 mm remain.
			Aspiration of the organoids during the differential spins	Because of the short nature of differential spins (step 15 to 17), the pellet can be very loose. We recommend leaving 1 mL of supernatant after differential spins. Sudden movements while handling the tube can also disturb the pellet.
			Usage of non BSA-coated pipet and tips	BSA-coating all serological pipets, pipet tips, and tubes is essential for a good organoid yield.
	20	Yield of small organoids too low	Low organoid yield or few small organoids collected	To increase the number of small organoids, we recommend pipetting up and down the big organoids multiple times to break them down. Then repeat step 19.
	23) a, b)	Low percentage of infected cells	Not enough virus or organoids too big	To increase the percentage of infected cells, we recommend increasing the MOI, decreasing organoid size (use only 150 cells organoids). We also noticed that using BSA-coated tips or FBS in the media decrease infection efficiency.
	23) b)	Presence of large aggregates	Organoids were not dispersed while in suspension	We recommend dispersing the organoids every day or every 2 days by pipetting up and down with a BSA-coated tip.
		No live organoids remaining	Low infection rates	See troubleshooting for steps 23/24 above
			Selection started too early	We recommend waiting at least 24 hours or 48 hours before starting the selection.
		Selection not effective in non-infected control	Antibiotic concentration too low	A dose response must be done on the organoids used to determine the effective antibiotic dose for selection.
II	26) a) (iii)	Matrigel too viscous	Matrigel was kept off ice	We recommend thawing Matrigel for 2 hours prior to use at 4°C. Use of pre-chilled pipet tips can also prevent Matrigel loss.
	26) a, b, c)	A large fraction of organoids invading in 2D	Flattening of ECM gel during plating	Place the plate on a heating block at 37° C while plating the organoids and wait at for at least 1–2 minutes before moving the plate. Organoids should be pipetted out very slowly to allow enough time for the ECM to gel.
	26) b) (iii)	Collagen too pink	pH too basic	Add small amount of collagen until the optimal salmon pink color is attained.
		Collagen too yellow	pH too acidic	Add small amount of NaOH until the optimal salmon pink color is attained.
	26) b) (vi)	Collagen polymerizes too fast	Prolonged exposure to basic pH or temperatures > 4 oC	Decrease the amount of time at basic pH while making the collagen solution. Also make sure that collagen is always kept on ice.
	26) b) (xi)	Difficult to visualize the organoids in collagen	Collagen too dense	Polymerization of the collagen at 4°C creates longer, thicker, and sparser collagen fibers than at 37°C. However, very long polymerization times at 4°C can lead to a very dense gel. Try polymerizing the collagen for shorter times. The polymerization status can be verified by looking a 15μL drop of gel under the microscope every 15–20 minutes.
	26) c) (iv)	Too few clusters	Inefficient digestion	The presence of FBS or Ca2+ can reduce the efficiency of TrypLE. Pre-wash the organoids with PBS before adding TrypLE. Pre-warming the TrypLE at 37°C will also help the digestion. Incubation time must be adapted to every tissue. Pipetting up and down during the incubation in TrypLE can help the digestion but may cause more cell death. Progress of the digestion can be verified by looking at a small sample under the microscope every 5 minutes.
		Presence of single cells but no clusters	Over-digestion	Decrease the incubation time with TrypLE. Remember to pipet the solution gently.
	26) c) (xiii)	No clusters after overnight incubation	Cell density too low	Increase the cell density per well. Using round bottomed cell repellent plate can also help the formation of clusters.
	26) d) (v)	No colony after 7 days incubation	Cell density too low	The colony formation assay is highly dependent on cell density. Increase the number of clusters plated per well.
III	27) a) (iii)	Matrigel dissolves during fixation	PFA depolymerization of Matrigel	Adding 0.25% of glutaraldehyde can prevent Matrigel depolymerization during fixation.
	27) a) (iv)	Low immunofluorescence signal	Antibody unable to penetrate in gel	Use protocol 65 B for cryosections.
	27) a) (ix)	Cryosections do not have organoids	Due to the 3D organization of the gel, some sections can lack organoids	Pre-staining organoids with DAPI (in PBS for 15 minutes) before embedding in O.C.T. will help positively identify sections containing organoids.
		Gel organization is lost during sectioning.	Bad sectioning temperature	The cryostat temperature should be adjusted between −20 to −27°C.
	27) b) (v)	Residual Matrigel bound to organoids	Incubation time too short or amount of EDTA in buffer too low	Repeat step 73(iii) until all of the Matrigel is completely dissolved. If any Matrigel remains, the BCA assay will give inaccurate estimates of protein content.
	27) d) (v)	Low protein yield	Low number of organoids	Increase the number of organoids or reduce the amount of lysis buffer
			Organoids not completely lysed	Increase the lysis incubation time to 1h while mixing regularly.
		Smear of protein on the western blot membrane	Protein degradation	Do not incubate organoids in collagenase or EDTA while dissociating organoids from ECM gels. All steps must be performed at 4° C in presence of phosphatase inhibitors.
	27) e) (vi)	Increased number of dead cells	Enzyme digestion too long	Decrease the incubation time with TrypLE and inactivate it with FBS. We recommend TrypLE over trypsin as it is gentler on primary cells.

Box 1: Mouse models of mammary, pancreatic, and hepatocellular carcinoma.

MMTV-PyMT transgenic female mice will form 15–20 mm mammary tumors by ~12–14 weeks of age⁵⁰. C3(1)-Tag transgenic female mice will form 15–20 mm mammary tumors by ~25–32 weeks of age⁵¹. Breast and pancreatic adenocarcinoma (PDAC) PDX tumors are passaged subcutaneously in NOD SCID gamma mice. The mice should be euthanized when the tumors are 15–20 mm and tumor dissected. The murine model for HCC is driven by the expression of c-Myc under the control of liver-activator protein (LAP). LAP also drives the expression of tetracycline transactivator (tTA) which is “OFF” in the presence of doxycycline⁵². These mice should be euthanized when they gain 25% of their body weight after doxycycline withdrawal.

DATA AVAILABILITY

The data mentioned in the protocol is included. Any additional information may be provided by the corresponding author, Andrew J. Ewald, upon request.