Derivation of Epithelial-Only Airway Organoids from Human Pluripotent Stem Cells

Abstract

New protocols to efficiently generate functional airway epithelial organoids from human pluripotent stem cells (PSCs) would represent a major advance towards effective disease modeling, drug screening and cell based therapies for lung disorders. This unit describes an approach using stage-specific signaling pathway manipulation to differentiate cells to proximal airway epithelium via key developmental intermediates. Cells are directed via definitive endoderm (DE) to anterior foregut, and then specified to NKX2-1+ lung epithelial progenitors. These lung progenitors are purified using cell surface marker sorting and replated in defined culture conditions to form three-dimensional, epithelial-only airway organoids. This directed differentiation approach using serum-free, defined media also includes protocols for evaluation of DE induction, intracellular FACS analysis of NKX2-1 specification efficiency and enrichment, and approaches for characterization and expansion of airway organoids. Taken together, this represents an efficient and reproducible approach to generate expandable airway organoids from human PSCs for use in numerous downstream applications.

INTRODUCTION

A major application of pluripotent stem cells (PSCs) is their directed differentiation to diverse organ-specific cell types of interest. Much like normal development, this approach relies on stage-specific developmental signals that lead to sequential restriction of cell fate. For example, lung progenitors derived from PSCs arise via the primitive streak and definitive endoderm, then proceed via anterior foregut endoderm to a primordial lung progenitor, identified by the expression of the transcription factor Nkx2-1+ (Longmire et al., 2012; Green et al., 2011; Huang et al., 2013; Mou et al., 2012). In general, this approach is useful not only as a model with the capacity to reveal cell fate decisions in high resolution and real time, but also for the generation of patient and tissue-specific cell types for disease modeling and drug screening.

Taken together, this unit describes an approach to generate primordial lung progenitors from pluripotent stem cells(Hawkins et al., 2017), purify these progenitors using a cell surface marker sorting strategy(Hawkins et al., 2017), and differentiate these cells further to epithelial-only airway spheres containing proximal lung epithelial cells, including secretory cells and lung basal cells. While these spheres do not contain ciliated cells, this lineage can be differentiated from these spheres using pharmacological Notch inhibition or after transfer to air-liquid interface culture.(McCauley et al., 2017).

DIFFERENTIATION OF HUMAN PLURIPOTENT STEM CELLS (PSCS) TO NKX2-1+ LUNG PROGENITORS VIA DEFINITIVE ENDODERM

This protocol describes an approach to generate and evaluate lung-competent definitive endoderm from PSCs and further specify these cells towards a well-characterized NKX2-1+ lung progenitor stage, from which point they have the potential to differentiate to the major cell types of the lung epithelium(Hawkins et al., 2017; McCauley et al., 2017). The focus of this chapter is the derivation of proximal airway epithelial cells from these NKX2-1+ progenitors via the withdrawal of Wnt signaling activation (McCauley et al., 2017). Although distal, alveolar epithelial cells also can be derived from the NKX2-1+ progenitor pool in the presence of sustained Wnt signaling activation, protocols for their efficient alveolar maturation are still under development, and hence are not the focus of this chapter. In order to generate the proximal airway lineages presented here, PSCs are first differentiated to definitive endoderm, which is subsequently patterned into anterior foregut endoderm via dual-SMAD inhibition. Subsequent activation of Wnt, BMP, and retinoic acid signaling drives differentiation to an NKX2-1+ (also known as TTF-1) lung progenitor fate.

This protocol is broadly applicable to human PSC lines, including embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs), although adaptation from maintenance culture on irradiated feeder layers to feeder-free mTeSR cultures prior to beginning this protocol is highly recommended. Furthermore, high line-to-line variability in specific timing of protocol steps and efficiency of NKX2-1 induction is expected and consequently specific optimization of the protocol for each individual PSC line is highly recommended. Hence, each relevant step details key advice for protocol optimization.

The authors’ experience suggests that 72 hours of definitive endoderm induction followed by 72 hours of anteriorization and then 7–10 days of lung specification is the typical time frame for differentiation of normal BU3 iPSCs or RUES2 ESCs.

The overview of the timeline for this directed differentiation protocol is shown in Figure 1.

Figure 1.

Timeline of directed differentiation to human lung progenitors.

Materials

• Human pluripotent stem cells (e.g. BU3 iPSCs (Kurmann et al., 2015))

• mTeSR1 defined feeder-free medium (StemCell Technologies, cat. no. 05850

• Matrigel human ESC-qualified matrix (Corning, cat. no. 354277)

• DMEM (Gibco cat. no. 11995-065)

• Gentle Cell Dissociation Reagent (StemCell Technologies, cat. no. 07174)

• StemDiff Definitive Endoderm Kit (StemCell Technologies, cat. no. 05110)

• 10 mM Y-27632 (see recipe)

• Complete Serum Free Differentiation Medium (cSFDM, see recipe)

• 10 mM SB431543 (see recipe)

• 2 mM Dorsomorphin (see recipe)

• 3 mM CHIR99021 (see recipe)

• 10 µg/mL recombinant human BMP4 (see recipe)

• 100 µM retinoic acid (see recipe)

• Mouse IgG1, APC conjugated (Life Technologies cat. no. MG105)

• Anti-human CD117, APC conjugated (Invitrogen cat. no. CD11705; also known as anti-C-Kit)

• Mouse IgG2a, PE conjugated (StemCell Technologies, cat. no. 60108PE)

• Anti-human CD184 (CXCR4), PE conjugated (StemCell Technologies cat. no. 60089PE

• FACS buffer (see recipe)

• 1.6% paraformaldehyde, diluted with PBS from 16% stock (Ted Pella, Inc., cat. no. 18505)

• Rabbit monoclonal antibody to TTF-1 (clone EP15847, Abcam, cat. no. ab76013; also known as anti-NKX2-1)

• 1.5 mg/mL AffiniPure Donkey Anti-Rabbit IgG (H+L), AlexaFluor 488 conjugated (Jackson Immunoresearch, cat. no. 711-225-152)

• Sterile 12- or 6-well tissue culture plates

• Humidified 37°C incubator with 5% CO₂

• Sterile 15 mL and 50 mL conical tubes

• Automated cell counting system (e.g. Luna Automated Cell Counter, Logos Biosystems)

• Falcon round bottom polystyrene or polypropylene tubes (dependent on flow cytometer specifications)

• Flow cytometer: able to detect PE, APC, and AlexaFluor 488 fluorescence

Differentiation to definitive endoderm

1. Expand PSCs on tissue culture plates using standard procedure.

   Definitive endoderm induction requires 2 × 10⁶ cells for each well of a 6-well plate of endoderm desired, which is equivalent to approximately one confluent well of a 6-well plate. Expansion of undifferentiated PSCs on mTeSR is highly recommended. Typically the adaptation of a cell line from feeder culture to feeder-free can be accomplished in approximately 2 weeks; however, if cells must be maintained on an irradiated MEF feeder layer, many more cells (approximately 3–4 confluent wells of a 6-well plate per well of definitive endoderm) may be required. Maintenance of PSCs using other feeder-free methods (e.g. E8 or equivalent) has not been thoroughly tested or validated for differentiation in this protocol.

   For cells maintained in mTeSR, one near-confluent well (ready to passage; Figure 2) typically contains sufficient cell numbers to be replated at the appropriate density in 1 well of the same size for definitive endoderm differentiation. See Figure 2 for images of typical density of undifferentiated cells at day −1 and day 0.

2. Prepare as many wells of tissue culture plates (typically one 6-well plate) as desired for definitive endoderm induction by diluting one aliquot of ESC-qualified Matrigel in 25 mL DMEM (final concentration = 1 mg/mL), adding 1 mL/well of a 6 well plate, and allowing to gel for 30–60 minutes at 37°C.

   Follow manufacturer’s directions for aliquoting and subsequent dilution of ESC-qualified Matrigel, as this product exhibits lot-to-lot variability in protein concentration.

   It is recommended to include an extra well of endoderm for evaluation of endoderm induction efficiency.

3. Aspirate medium from the cells, add 1 mL/well (for 6-well plates) Gentle Cell Dissociation Reagent (GCDR) and incubate for 8–10 minutes at 37°C.

   The duration of GCDR treatment required for effective dissociation is highly variable from line-to-line and across experiments. When cells are ready to dissociate, the previously densely packed PSC colonies will become less well-formed and contain visible single cells.

4. Gently pipette up and down with a p1000 to dissociate PSCs from the plate and transfer to a new 15 mL conical. Add an equivalent volume of warm DMEM to the cells to dilute the GCDR.

5. Centrifuge cells at 4°C for 5 minutes at 300 RCF.

6. Aspirate media from cell pellet and resuspend in 1 mL mTeSR1 supplemented with 10 µM Y-27632. Remove a sample of cells (typically 10 µL) and combine with an equal volume for automated cell counting to determine total cell number. After cell counting, add additional mTeSR with Y-27632 to a final concentration of 1×10⁶ cells/mL.

7. Aspirate dilute Matrigel from plate to leave thin layer of polymerized matrix, and plate 2 mL of cell suspension per Matrigel-coated well of a 6-well plate.

8. 24 hours after replating, aspirate media and wash cells with 1 mL DMEM per well. Aspirate DMEM and refeed cells with 2 mL per well of Endoderm Medium 1.

   For 1 mL of Endoderm Medium 1:

   • 980 µL Definitive Endoderm Base (from Definitive Endoderm Kit)
   • 10 µL Supplement A
   • 10 µL Supplement B

   This is Day 0 of differentiation. This is different from the numbering of days indicated in the StemDiff Definitive Endoderm Kit protocol.

   By 24 hours after replating, the plate should be 90–100% confluent and will likely contain many floating dead cells. Any density >70% is generally sufficient to initiate definitive endoderm induction. If cells are under-confluent, they can be left for 1–2 days to proliferate further. See Figure 2 for typical appearance of cells at this “Day 0” stage.

9. Exactly 24 hours after beginning differentiation, aspirate medium and refeed cells with 2 mL per well Endoderm Medium 2.

   For each 1 mL of Endoderm Medium 2:

   • 990 µL Definitive Endoderm Base
   • 10 µL Supplement B

10. Refeed cells with Endoderm Medium 2 again the following day (Day 2).

11. Begin to evaluate definitive endoderm induction efficiency at approximately 60 hours of differentiation.

    There is high line-to-line variability in the timing of optimal induction of definitive endoderm for downstream lung competence. While 72 hours is a typically appropriate duration to generate lung-competent definitive endoderm for most lines, it is highly recommended to perform a careful kinetic (every 6 – 12 hours between 60 – 96 hours) of endoderm induction for any new lines, using FACS analysis of CD117/CXCR4 co-expression and downstream NKX2-1 efficiency as readouts. See Figure 2 for typical appearance of cells at this “Day 3” stage.

Figure 2.

Phase contrast microscopy of cells from undifferentiated PSCs (Day −1) through Day 4 of differentiation (Anteriorization Stage) showing typical morphologies (scale = 500 µm for first row, 200 µm for high magnification images in second row). Spontaneous differentiation in undifferentiated PSCs is indicated with a black circle. It is recommended that cultures have < 5% spontaneous differentiation by eye prior to being used for lung epithelial differentiation experiments.

Evaluation of definitive endoderm efficiency

1. At desired endoderm timepoint (e.g. 72 hours; Day 3), aspirate media and add 1 mL/well of GCDR. Incubate for 2–4 minutes at 37°C. Evaluate dissociation progress periodically by light microscopy.

   Typically, 1 well of endoderm is harvested to evaluate endoderm efficiency and the rest of the endoderm wells are replated for ongoing differentiation, as discussed in the following section.

2. After cells are fully dissociated, gently pipette the cells up and down in the GCDR to generate a single cell suspension and transfer to a new conical. Add in an equivalent volume of empty media (e.g. IMDM). Spin cells for 5 minutes at 4°C and 300 RCF.

3. Resuspend cells in 1 mL of FACS buffer (see recipe) and perform a cell count. Transfer a volume containing 0.5 × 10⁶ cells into each of 5 Eppendorf 1.5mL tubes and spin cells for 5 minutes at 4°C and 300 RCF.

4. Resuspend each pellet in 100 µL FACS buffer. Label tubes for each stain, indicated below, and add the appropriate antibodies:

   • Unstained: No antibody
   • Isotypes: 2 µL mouse IgG1, APC conjugated; 2 µL mouse IgG2a, PE conjugated
   • CD117 only: 2 µL anti-human CD117
   • CXCR4 only: 2 µL anti-human CXCR4
   • Double stain: 2 µL anti-human CD117; 2 µL anti-human CXCR4

5. Gently pipette cells up and down and place on ice in the dark for 30 minutes to stain.

6. After 30 minutes, add 1 mL FACS buffer to each tube and spin cells for 5 minutes at 4°C and 300 RCF.

7. Resuspend each pellet in 500 µL FACS buffer. Evaluate staining using a compatible flow cytometer and standard protocols, including compensation, by gating first for cell size (FSC vs. SSC) and subsequently for expression of CD117 and CXCR4.

   Endoderm score is calculated as the percentage of cells coexpressing CD117 and CXCR4. An endoderm efficiency of 90–100% is expected within 72 – 96 hours of differentiation and is optimal for downstream lung differentiation. See Figure 3 for an example of efficient definitive endoderm induction.

   As previously noted, CD117/CXCR4 coexpression is not the only metric with which definitive endoderm should be evaluated when optimizing a protocol. Typically, the earliest timepoint at which the cells are >90% CD117/CXCR4 double-positive is optimal for downstream lung differentiation. However, it is highly recommended that prior to proceeding with the protocol with new or difficult to differentiate lines, a careful endoderm kinetic is performed to address the relationship between endoderm timing and downstream NKX2-1 induction efficiency(Hawkins et al., 2017).

Figure 3.

Representative flow cytometry dot plots showing typical gating for size and granularity (FSC vs. SSC) and CD117 vs. CXCR4 expression. Efficient endoderm induction after 72 hours from PSCs is demonstrated by staining with isotype controls or CD117 and CXCR4 antibodies.

Replating cells to generate anterior foregut endoderm

1. Once cells have reached the definitive endoderm stage (typically Day 3), prepare anterior foregut endoderm medium (“DS/SB”) by adding 1 µL of 2 mM dorsomorphin and 1 µL of 10 mM SB431540 per 1 mL of cSFDM (final concentration: 2 µM dorsomorphin, 10 µM SB431540). Further supplement enough media for replating with 1 µL/mL Y-27632 (final concentration: 10 µM)

   Enough media for replating the cells and feeding them one time can be prepared at this stage, including approximately 10% excess. For example, one well of endoderm might be replated into 4 wells of anterior foregut endoderm. In this case, prepare 18 mL of DS/SB media, aliquot 8 mL into a separate conical, and add 8 uL Y-27632 for replating. The remaining 10 mL will be used to refeed the cells on the following day.

2. Prepare as many wells of tissue culture plates as required for replating by diluting one aliquot of ESC-qualified Matrigel in 25 mL DMEM (final concentration = 1 mg/mL) and allowing to gel for 30–60 minutes at 37°C.

3. Aspirate media and wash with 1 mL/well DMEM/F-12.

4. Add 1 mL/well GCDR and incubate at room temperature (25°C) for 2–3 minutes.

5. Aspirate GCDR and add 1 mL/well DS/SB + Y-27632. Pipette gently with a p1000 or a 5 mL serological pipette to dislodge monolayer into large, uniformly sized clumps of approximately 10 – 20 cells. Transfer cells to a new conical tube and add DS/SB + Y-27632 to a final volume equivalent to 2 mL/well for the number of wells to be replated.

   Typically 2 – 6 wells are replated for each well of definitive endoderm.

   There is high line-to-line variability on the density of replating for definitive endoderm. A good rule of thumb is to try several splits in the range of 1:2 – 1:6 for initial differentiations of any given line. Cells should be approximately 60 – 70% confluent on the next day. See Figure 2 for typical appearance of cells on Day 4 (24 hours after replating).

   It is critical at this stage to avoid pipetting as much as possible. If necessary, a cell scraper can be used to lift off the cells from the plate in lieu of additional pipetting.

6. Pipetting gently, replate 2 mL/well of cells into each Matrigel-coated well of a 6-well plate.

7. Evaluate cell survival and density after 24 hours. Refeed with fresh DS/SB media.

   Cells should be ~60 – 70% dense after 24 hours. For highly-proliferative lines, a lower density may be desired. If downstream NKX2-1 efficiency is low, replating density can be optimized; typically, a lower density results in higher NKX2-1 efficiency.

   Removal of Y-27632 after 24 hours will result in a change in cell morphology from spiky-appearing cell borders to a more epithelial appearance.

   Typically, the optimal period for anteriorization in DS/SB media is 72 hours, during which time the media only needs to be changed on the second day (24 hours after replating). Variation of duration of anteriorization, using downstream NKX2-1 efficiency as a readout, is another option for optimization of differentiation efficiency across lines.

Specification of NKX2-1+ lung progenitors

1. After anteriorization (typically 72 hours, or Day 6 of total differentiation), prepare lung induction medium without RA (“CB”) by adding 1 µL of 3 mM CHIR99021 and 1 µL of 10 ug/mL rhBMP4 per 1 mL of cSFDM (final concentration: 3 uM CHIR99021, 10 ng/mL rhBMP4). This media can be stored at 4°C for up to one week.

   As cells are typically refed every other day, it is recommended to prepare enough media at a time to refeed all wells of cells three times (e.g. 6–8 mL/well), as this volume will be used up in approximately one week.

2. Supplement 2 mL/well of prepared “CB” medium with 1 µL/mL retinoic acid (Final concentration: 100 nM) to generate “CBRA” medium.

   Retinoic acid is highly light sensitive and this step and all subsequent steps using this media should be performed in low lighting (i.e. with the tissue culture hood lights off).

3. Aspirate DS/SB media and wash each well with 1 mL/well empty medium (e.g. IMDM).

4. Aspirate wash media and add 2 mL/well of CBRA medium.

5. Change CBRA media every other day, or more frequently if necessary.

   Retinoic acid media should be added fresh to prepared CB media each day prior to refeeding. Media will turn yellow if it needs to be replaced more frequently than every other day.

   After 1–2 days in CBRA media, cells will proliferate rapidly and form distinctive round clusters.

   Cells will begin to upregulate NKX2-1 expression by Day 8 of differentiation and are typically evaluated and replated for airway differentiation between Day 14 and 16.

6. After 14 – 16 days of differentiation, evaluate NKX2-1 lung progenitor specification efficiency by immunofluorescent staining for NKX2-1 and FOXA2 expression (see Support Protocol 1) or by intracellular FACS for NKX2-1 expression (see Support Protocol 2).

   For downstream differentiation to airway epithelial organoids, it is critical to evaluate the induction efficiency of NKX2-1 on a per-cell basis. Therefore, it is recommended that either immunocytochemistry or FACS for intracellular NKX2-1 be used to evaluate differentiation efficiency. It is inadequate to estimate differentiation efficiency solely using quantitative real-time PCR analysis. An additional option for evaluating NKX2-1 induction efficiency while learning this protocol is the use of a validated NKX2-1^GFP reporter PSC line (Hawkins et al., 2017; McCauley et al., 2017).

BASIC PROTOCOL 2: Purification of NKX2-1+ lung progenitors via cell sorting on surface proteins CD47 and CD26

Basic Protocol 2 describes an approach to enrich for NKX2-1+ cells derived in Basic Protocol 1 by flow cytometry-based cell sorting. Cells are harvested as a single-cell suspension, immunostained for CD47 and CD26, and sorted for cells expressing high levels of CD47 and low/no levels of CD26. These cells are relatively enriched for expression of NKX2-1 and can be further characterized or replated for downstream differentiation to airway epithelium. This protocol is applicable to cells derived using the approach described in this unit, and has not been further validated to enrich for NKX2-1 expression in lung progenitors derived from PSCs using other approaches.

Materials

• 0.05% Trypsin-EDTA (e.g. Gibco, cat. no 25-300-062)

• Fetal bovine serum (FBS, e.g. Gibco, cat. no. 10082139)

• DMEM (Gibco cat. no. 11995-065)

• Sort buffer (see recipe)

• Mouse monoclonal antibody to CD26, PE conjugated (clone BA5b, Biolegend cat. no. 302705)

• Mouse monoclonal antibody to CD47, PerCP-Cy5 conjugated (clone CC2C6, Biolegend, cat. no. 323110

• Mouse IgG1 isotype, PE conjugated (Biolegend cat. no. 400113)

• Mouse IgG1 isotype, PerCP/Cy5-5 conjugated (Biolegend cat. no. 400149)

• 10 µM calcein blue, AM, in DMSO (ThermoFisher, C1429)

• Sterile 12- or 6-well tissue culture plates

• Humidified 37°C incubator with 5% CO₂

• Sterile 50 mL conical tubes

• Sterile 1.5 mL Eppendorf tubes

• Automated cell counting system (e.g. Luna Automated Cell Counter, Logos Biosystems)

• 5 mL Falcon round bottom polystyrene or polypropylene tubes (dependent on cell sorter specifications)

• High-speed cell sorter (e.g. MoFlo) able to perform two-way cell sorting and detect and compensate PerCP/Cy5-5, calcein blue, and PE staining

• Additional reagents and equipment for Support Protocol 2 may be required to evaluate efficiency of NKX2-1 enrichment

Harvesting and Dissociating Cells For Staining

1. Wash specified cells with DMEM, 1 mL/well of a 6-well plate.

2. Aspirate wash media. Add 1 mL /well of a 6-well plate of 37°C 0.05% trypsin and place in incubator for 14–18 minutes.

   It is critical to ensure that the majority of the dissociation of the cells at this stage is enzymatic. Therefore, one should visualize the cells under a microscope at the end of the expected dissociation time. If cells remain tightly stuck to the plate, return them to the incubator for approximately 3–5 more minutes before proceeding to the next step.

   If the trypsin turns yellow during the dissociation step, which is common when cells are very confluent, add 1–2 more mL of trypsin to each well and return the plate to the incubator.

3. Pipette up and down (gently—2 or 3 times) until cells are fully detached from the plate and transfer to a 50 mL conical.

4. Wash the well with an additional 1 mL of 0.05% trypsin. Add this wash to the conical tube and pipette up and down 1–2 additional times with a p1000 or 5 mL pipette to further dissociate the cell clusters. If many large clumps persist, manually shake, flick, and roll the 50ml conical for 3–5 minutes.

   Do not pipette more than a few times at this stage. There will frequently still be some cells stuck together in large clumps. If there are many of these clumps, continue to pipette a few times more with a 5 mL pipette. However, these clumps will be filtered out during subsequent steps, and while it is important to get enough cells dissociated and not to lose too many during the filtration process, it is also important to dissociate the cells gently. Therefore, one should try to break up as many clumps as possible by continued trypsinization and gentle pipetting, but without resorting to substantial or harsh pipetting to manually disperse the clusters.

5. Prepare “stop medium” by adding 50 mL FBS to 450 mL of DMEM. Add an equivalent volume of stop medium to the cell suspension and pipette up and down gently with a 25 mL pipette to mix the cell suspension.

   For fewer wells of cells (with a consequent lower final volume of the cell suspension), a 5 – 10 mL pipette can be used instead.

6. Fit a 40 µm strainer to a new 50 mL conical tube and gently transfer the cell suspension into this new tube.

7. Spin down cells at 4°C for 5 minutes at 300 × g.

8. Resuspend cells in 1 mL of FACS buffer and perform a cell count.

9. Spin down cells at 4°C for 5 minutes at 300 × g and resuspend the cell pellet in sort buffer at 100 µL per 1 × 10⁶ cells.

Sorting CD47^high, CD26^low Lung Progenitors

• 1
  Aliquot cells into separate 1.5 mL Eppendorf tubes for controls and staining (6 samples total):

  • Sample 1: Unstained
  • Sample 2: Calcein blue only
  • Sample 3: CD47^PerCP/Cy5.5 only
  • Sample 4: CD26^PE only
  • Sample 5: mIgG1^PerCP/Cy5.5, mIgG2^PE (isotype controls)
  • Sample 6: CD47^PerCP/Cy5.5/CD26^PE

  As the final condition (Sample 6, double stained) is the actual staining condition for the cells that will be sorted for downstream applications, the majority of the cells should be aliquoted into Sample 6. For example, if one harvests 10 × 10⁶ cells, 1 × 10⁶ cells should be transferred to each of Samples 1–5 and 5 × 10⁶ cells should be used for Sample 6, the sample to be sorted. Sample 6 staining can be performed in a 15- or 50 mL conical, if sample volumes are too large for a 1.5 mL Eppendorf tube.

  An aliquot of unsorted cells should also be taken at this stage for later use to determine NKX2-1 differentiation efficiency and enrichment by sorting using intracellular flow cytometry (see Support Protocol 2).

  Additionally, if few cells are available for staining, 50 µL of staining volume can be used for the controls (Samples 1–5) with a concordant reduced volume of antibody.

• 2Add 0.5 µL antibody per 100 µL staining volume to the appropriate conditions.

  For example, Samples 1–2 should receive no antibody, Sample 3 should only have the CD47^PerCP/Cy5.5 antibody, etc.

• 3Place the samples on ice and stain the samples for 30 minutes, protected from light.
• 4While cells are staining, prepare collection tubes to capture sorted cells by aliquoting a small volume of sort buffer into each collection tube.

  Collection tubes used depend on the specifications of the cell sorter used. In practice, some typical options (depending on cell number to be sorted) include 1.5 mL Eppendorf tubes (add approximately 200 µL of sort buffer), 4 mL round-bottom Falcon tubes (add approximately 500 µL of sort buffer), or 15 mL conical tubes (recommended; add approximately 1 mL of sort buffer).

• 5After 30 minutes of staining, add 1 mL stop medium to each tube to remove excess antibody and centrifuge at 4°C for 5 minutes at 300 × g.
• 6Aspirate supernatant and re-suspend the cell pellet in 500 µL of sort buffer per 1 × 10⁶ cells.
• 6Add 1 µL/mL of 10 µM calcein blue, AM (final concentration: 10 nM) to each of Samples 2–6.

  Any viability dye that does not interfere with either PE or PerCP-Cy5.5 staining can be used to exclude non-viable cells.

• 7Transfer filtered cell suspension to appropriate round bottom Falcon tubes for the sorting instrument and bring cells to sorter on ice, protected from the light.
• 8Using a flow cytometer, and maintaining sterile conditions, sort CD47 high, CD26 low/negative cells into a collection tube (Figure 4).

  If possible, it is recommended to sort the cells with a trained operator familiar with sorting cells for replating. Sorting efficiency can be optimized, if needed, by diluting cells in a larger volume and filtering multiple times to prevent clumping.

  Cells should be gated first for size to exclude fragmented particles (FSC vs. SSC), then gated for singlets (e.g. FSC-A vs. FSC-H), then for viability (recommended: calcein blue+; Figure 4). Typically, >90% of cells are viable. While the authors of this chapter employ a logarithmic SSC scale and linear FSC scale (Figure 4a), any scale familiar to the user may be effectively employed. Next, cells should be gated for the CD47^high, CD26^low population to enrich for lung progenitors. Typically, >90% of the cells express CD47 at a level higher than isotype controls but only the CD47^high population is enriched for NKX2-1+ cells. The optimal location for the lower boundary of the CD47^high sort gate is just above the intersection of the CD47 and the CD26 clouds. This gating strategy is detailed in Figure 4.

  Typically, the higher the CD47 gate, the better the NKX2-1 enrichment score (enrichment score= presort NKX2-1+ %/post sort NKX2-1+ %). Depending on the application, it might be more important to be stringent with the gating (i.e. fewer cells are sorted but it is more important that they are highly enriched for NKX2-1).

  Enrichment for NKX2-1 expression should be evaluated by intracellular flow cytometry by taking a subsample of cells pre- and post-sorting for fixation in 1.6% paraformaldehyde and subsequent NKX2-1 staining (see details in Support Protocol 2). CD47^low cells should also be sorted as a gating control for this analysis. Alternatively, the protocol can be practiced using a validated NKX2-1^GFP reporter PSC line, in which case all three markers (CD47, CD26, and GFP) can be analyzed simultaneously for calculation of enrichment scores (Hawkins et al., 2017).

Figure 4.

Sorting algorithm for CD47^high, CD26^low lung progenitors. (a) Flow cytometry dot plots showing typical gating strategy for differentiations of variable efficiency. Cells were first be gated for size and granularity (FSC vs. SSC), then live cells (calcein blue+), then CD47 vs. CD26. As depicted, the lower limit of the CD47^high gate should be drawn just above where the CD26⁺ cloud intersects the main population. Following sorting, the indicated NKX2-1% enrichment was evaluated by intracellular NKX2-1 staining of aliquots of either bulk (“presort”) or sorted (CD47high/CD26low; “postsort”) cells. (b) Typical appearance of controls, including isotypes, CD47 single-stain, CD26 single-stain, and CD47/CD26 double staining after appropriate pre-gating (SSC/FSS/Calcein blue) and compensation (using single color stained controls) for bleed through of the CD26-PE fluorescence into the channel measuring CD47-PerCP-Cy5 fluorescence. Note the expression of CD47 over isotype levels in nearly all cells.

BASIC PROTOCOL 3: Generation and passaging of airway epithelial organoids

Basic Protocol 3 describes an approach to establish three-dimensional (3D) culture of purified PSC-derived lung progenitors generated in Basic Protocol 1 and Basic Protocol 2 and drive them towards an airway epithelial phenotype. After specification of the NKX2-1+ lung progenitor identity, withdrawal or inhibition of Wnt activity drives these epithelial lung progenitors towards a proximal airway fate, whereas continued Wnt activation drives distal differentiation(McCauley et al., 2017). Therefore, this protocol relies on the expansion of specified, purified lung progenitors in “low-Wnt,” FGF-containing media to drive differentiation or airway epithelial progenitors. Briefly, cells are resuspended in 3D Matrigel matrix and replated in tissue culture dishes. Culture of these cells in media driving activation of FGF signaling via FGF2 and FGF10 and containing corticosteroids and cyclic-AMP drives cells to form epithelial spheres containing differentiated airway cell types.

Materials

Growth factor-reduced Matrigel matrix (Corning, cat. no. 356234)

Airway Differentiation Medium (see recipe)

2 mg/mL dispase II (see recipe)

0.05% Trypsin-EDTA (e.g. Gibco, cat. no 25-300-062)

Fetal bovine serum (FBS, e.g. Gibco, cat. no. 10082139)

DMEM (Gibco cat. no. 11995-065)

12-well sterile tissue culture plate

Sterile 15 mL conical tubes

Automated cell counting system (e.g. Luna Automated Cell Counter, Logos Biosystems)

Establishing three-dimensional airway epithelial organoid culture

1. Spin down single cells post-sort for 5 minutes at 4°C and 300 × g.

2. Resuspend cells at a concentration of 1000 cells/µL in undiluted Matrigel matrix and replate in 50 – 100 µL drops.

   Prior to resuspending cells, Matrigel should be thawed and kept cold (on ice) to prevent polymerization. For ease of thawing, it is convenient to aliquot the Matrigel to be used in 3D culture into 500 µL – 1 mL aliquots. Take care when resuspending cells not to introduce bubbles into the Matrigel and to efficiently disperse the cell pellet into single cells distributed throughout the Matrigel.

   Smaller Matrigel drops can be used but are not as robust at generating organoids as cells can settle through the Matrigel and attach to the bottom of the plate.

   If more cells are required downstream, cells can also be replated in several small (50 – 100 µL) drops in a 6-well plate or a p100 dish.

   Cells should be plated in one drop per well of a 12-well plate.

3. Allow drops to solidify for 20 minutes at 37°C.

4. After drops have fully polymerized, add Airway Differentiation Medium carefully to wells.

   Add enough media to ensure that drops are fully covered, typically 1 – 2 mL per well of a 12 well plate.

5. Cells will begin to form epithelial spheres (“organoids”) after several days to one week of culture and will continue to proliferate and expand until the drop is filled with cells (see Figure 5).

   After organoid cultures are established, they can be characterized using several standard techniques, including quantitative real-time PCR, whole mount immunofluorescence microscopy of organoids, or embedding for formalin-fixed, paraffin embedded sections (as in McCauley et al., 2017). To generate sufficient material for sectioning, it is important to wait until organoids are of sufficient size for embedding as a large pellet. Organoids can also be resuspended in low-melting temperature agarose prior to fixation, embedding, and sectioning.

   Airway epithelial organoid identity is defined by expression of proximal lung markers, including NKX2-1 and SOX2 as well as by the presence of proximal lung lineages, including cells expressing markers of secretory lineages (NKX2-1+ and SCGB3A2+ or SCGB1A1) and airway basal cells (NKX2-1+P63+KRT5+). The proximalized epithelial airway organoids generated in this chapter can also be differentiated to ciliated cells using pharmacologic inhibition of Notch signaling or can be expanded and differentiated in two-dimensional air-liquid interface culture(McCauley et al., 2017).

Figure 5.

Phase contrast microscopy and immunofluorescence microscopy of typical three-dimensional culture at day 31 (airway organoid stage) after differentiation from single cells in Airway Differentiation Medium from day 15 – 31. Table indicates recommended markers to profile airway markers and cell types (basal and secretory cells) derived in this protocol. Immunofluorescence microscopy is reprinted with permission from McCauley, K. B. et al., 2017. Efficient Derivation of Functional Human Airway Epithelium from Pluripotent Stem Cells via Temporal Regulation of Wnt Signaling. Cell Stem Cell, 20(6), 844–857.e6. http://doi.org/10.1016/j.stem.2017.03.001.

Passaging airway epithelial organoids

1. Aspirate media and add 2 mg/mL dispase to well to cover droplet (typically 1 mL/well) and incubate at 37°C for 30 minutes to 1 hour, until Matrigel is fully dissolved.

   Dislodging the Matrigel pellet with a pipette prior to incubation and gentle pipetting 3–5 times after 10 min can facilitate dissocation.

2. Using a p1000 pipette, transfer dissociated organoids to a new 15-mL conical tube and add an equivalent volume of DMEM.

3. Spin down spheres for 1–2 minutes at 4°C, 300 × g.

   If organoids have not formed a pellet after this time, spin for an additional 1–2 minutes.

   If organoids are particularly large, they can be allowed to settle to the bottom of the conical instead of this centrifugation step. This is particuarily useful if there is a lot of debris in the Matrigel drop, as this will not settle and is aspirated with the supernatant.

4. Aspirate supernatant and add 1 – 2 mL 0.05% trypsin per dissociated drop.

   For example, if 3 drops were originally dissociated, add 3 mL trypsin.

5. Transfer trypsin and cells to a well of a 6-well plate and incubate for 8 – 10 minutes at 37°C.

   Cells will begin to visibly dissociate from the spheres. Allow cells to incubate with trypsin until they are entirely dissociated; they will not survive being mechanically dissociated by pipetting so most of the dissociation should be enzymatic. If cells have not dissociated after 12 minutes in trypsin, collect the cells, spin them down, and resuspend in fresh trypsin for an additional 3 – 4 minutes.

6. While cells are dissociating, prepare “stop medium” by adding 50 mL FBS to 450 mL DMEM.

7. Collect dissociated cells in a new 15 mL conical tube and add an equivalent volume of stop medium.

8. Spin down cells for 5 minutes at 4°C and 300 × g.

9. Resuspend cells in 1 mL DMEM and count using an automated cell counter.

10. Spin down cells and resuspend at a concentration of 1000 cells/µL in undiluted Matrigel matrix and replate in 50 – 100 µL drops, allow drops to solidify for 30 minutes to 1 hour at 37°C, and add Airway Differentiation Medium carefully to wells.

    This protocol also is effective to prepare a viable single cell solution of cells for flow cytometry or sorting from airway organoids. For this approach, spin down cells and resuspend in FACS buffer for staining or other downstream analysis.

Support Protocol 1: Fluorescent immunocytochemistry for endodermal NKX2-1 induction efficiency

Support Protocol 1 describes an approach to characterize efficiency of differentiation of PSCs to lung progenitors in Basic Protocol 1 by immunocytochemistry for NKX2-1 expression. Although less quantitative than other approaches, such as flow cytometry, immunocytochemistry gives a valuable overall impression of the differentiation efficiency of a given line or protocol.

Additional Materials

• 4% paraformaldehyde

• Phosphate buffered saline (PBS)

• 0.3% Triton X-100 in phosphate buffered saline (PBS)

• Normal donkey serum (or other blocking reagent, for example 4% normal goat serum or 1% bovine serum albumin)

• Rabbit anti-TTF1 antibody, clone EP1584Y (Abcam, cat. no. ab76013)

• 1.5 mg/mL AffiniPure Donkey Anti-Rabbit IgG (H+L), AlexaFluor 488 conjugated (Jackson Immunoresearch, cat. no. 711-225-152)

• Hoechst (ThermoFisher H3570)

• 1.5 mL Eppendorf tubes

• Standard fluorescence microscope capable of imaging AlexaFluor 488 and Hoechst staining in a tissue culture plate format

1. Aspirate media from differentiated cells and wash with empty media (e.g. IMDM).

   Staining for NKX2-1 with appropriate controls requires 3 wells of cells.

   Furthermore, it is beneficial to include a negative control cell population for staining, as the NKX2-1 antibody has high levels of nuclear background. Undifferentiated PSCs are typically the most accessible cell type, although other NKX2-1 negative cells can also be used for this purpose. These cells should be harvested and stained using the same approach as the differentiated lung progenitors.

2. Add 1 mL/well of 4% paraformaldehyde and incubate at room temperature (25°C) for 15 minutes.

   Volumes given are for cells cultured in a 6 well plate. The staining protocol can be readily scaled to smaller well sizes.

3. Wash 3 times for 3 minutes each with PBS.

   At this stage, plates can be sealed with parafilm and stored at 4°C for up to 2 weeks with cells in PBS prior to continuing with the staining protocol.

4. Aspirate PBS and add 1 mL/well of 0.3% Triton-X 100 for 10 minutes at 25°C to permeabalize cells. Wash cells 3 times for 3 minutes each with PBS.

5. Prepare blocking solution by diluting normal donkey serum (NDS) in PBS to a final concentration of 4% v/v. Aspirate PBS and add 1 mL/well of 4% NDS. Incubate cells for 1 hour at room temperature.

   For example, add 400 µL of NDS to 9.6 mL of PBS.

   Other blocking reagents, such as 1% BSA, can also be used in lieu of 4% normal donkey serum.

6. Prepare solutions of NKX2-1 antibody in 4% NDS, as follows. Staining requires 500 µL of dilute antibody/well:

   • Unstained: No primary antibody
   • Secondary only: No primary antibody
   • NKX2-1: 1 µL antibody/500 µL 4% NDS

7. Aspirate blocking solution and add each prepared antibody solution to a separate well of prepared cells. Incubate overnight at 4°C or for 1 hour at 37°C.

   Overnight incubation may improve antibody specificity.

8. Wash cells 3 times for 3 minutes each with PBS.

9. Prepare dilute secondary antibody solutions and add the appropriate secondary antibodies to each well:

   • Unstained: No secondary antibody
   • Secondary only: 1 µL donkey anti-rabbit (AlexaFluor 488 conjugated) in 500 µL PBS
   • NKX2-1: 1 µL donkey anti-rabbit (AlexaFluor 488 conjugated) in 500 µL PBS

10. Incubate with appropriate secondary antibody at room temperature for 30 minutes in the dark. After staining, wash 3 times for 3 minutes each with PBS.

11. Counterstain with Hoechst (diluted 1:500 in PBS) for 30 minutes at room temperature in the dark.

12. Wash 3 times for 3 minutes each with PBS.

13. Image cells on a compatible fluorescence microscope for Hoechst (blue) and NKX2-1 (AlexaFluor 488 labeled) nuclear staining. Nuclear staining efficiency can be quantified by standard image analysis software (e.g. ImageJ).

    Typical Day 15 NKX2-1 immunofluorescence is shown in Figure 6.

Figure 6.

Immunofluorescence microscopy for NKX2-1 (magenta) and DAPI (grey) at Day 15 showing typical staining pattern for a) high and b) low efficiency differentiations.

Support Protocol 2: Flow cytometry for intracellular NKX2-1

Support Protocol 2 describes an approach to evaluate the efficiency of differentiation of lung progenitors using Basic Protocol 1 and the enrichment of these cells by cell sorting in Basic Protocol 2. For this latter application, an aliquot of cells should be taken after sorting based on CD47/CD26 gating for evaluation of NKX2-1 enrichment.

This approach relies on the fixation, permeabilization, and intranuclear staining of cells for NKX2-1 protein before evaluation by flow cytometry.

Additional Materials

• 0.05% Trypsin-EDTA (Gibco, cat. no 25-300-062)

• Fetal bovine serum (FBS, e.g. Gibco, cat. no. 10082139)

• 1.6% paraformaldehyde

• 10× Intracellular Staining Permeabilization Wash Buffer (Biolegend, cat. no. 421002)

• Rabbit anti-TTF1 antibody, clone EP1584Y (Abcam, cat. no. ab76013)

• Rabbit IgG, clone EPR25A (Abcam cat. no. 172730)

• 1.5 mg/mL AffiniPure Donkey Anti-Rabbit IgG (H+L), AlexaFluor 488 conjugated (Jackson Immunoresearch, cat. no. 711-225-152)

• 1.5 mL Eppendorf tubes

• 40 µm cell strainer (Corning, cat. no. 431750)

1. Incubate differentiated cells with 0.05% Trypsin-EDTA at 37°C for 10 – 12 minutes. Evaluate single cell dissociation efficiency by light microscopy. While cells are dissociating, prepare “stop medium” by adding 50 mL FBS to 450 mL empty media (e.g. DMEM).

   It is also beneficial to simultaneously harvest undifferentiated PSCs to use as a staining control for FACS at this stage.

2. Pipette gently to transfer cells to a new 50 mL conical. Add stop medium to the conical of cells to inactivate the trypsin.

   It is important to avoid excessive pipetting at this stage. Therefore, cells should be harvested from the plate when almost all adherent cells have lifted off and formed floating clusters of single cells.

   If cells are still in large clumps after being transferred to the new conical tube, cap the tube and shake vigorously side-to-side to dissociate large clusters prior to adding stop medium.

3. Filter cells through a 40 µm cell strainer into a new 50 mL conical and spin down for 5 minutes at 4°C and 300 RCF.

4. Resuspend cells in 1 mL FACS buffer and count using an automated cell counter. Centrifuge for 5 minutes at 4°C and 300 RCF.

   Ideally, at least 1 × 10⁵ cells should be used per staining condition.

   For evaluation of cells that have been sorted using the CD47^high/CD26^low algorithm, begin at step 5 (below) after spinning down the sorted cells for 5 minutes at 4°C and 300 × g. To ensure appropriate gating and scoring of differentiation efficiency, aliquots of sorted CD47^low and presorted cells should also be included as controls for NKX2-1 intracellular FACS.

5. Aspirate FACS buffer and resuspend cells in 1.6% paraformaldehyde. Incubate, gently shaking or rocking, at 37°C for 10 minutes.

6. Prepare staining buffer by diluting 10× Intracellular Staining Permeabilization Wash Buffer in FACS buffer to a final concentration of 1×.

   For example, add 1 mL 10× Intracellular Staining Permeabilization Wash Buffer to 9 mL FACS buffer.

7. Spin down cells for 5 minutes at 4°C and 300 RCF and resuspend in 100 µL staining buffer per 1 × 10⁵ – 1 × 10⁶ cells for a total volume of at least 300 µL. Transfer 100 µL of cells to each of 3 Eppendorf tubes (1.5 mL). Add 900 µL to each tube (final volume: 1 mL), incubate for 5 minutes at room temperature, and spin down cells for 5 minutes at 4°C and 300 RCF.

8. Aspirate buffer and add appropriate primary antibody at the appropriate concentration (diluted in staining buffer) to each tube:

   • Unstained: No primary antibody
   • Isotype: 0.2 µg/mL rabbit IgG (approximately 0.02 µL antibody per 100 µL staining buffer)
   • NKX2-1: 0.2 µg/mL rabbit anti-TTF1 antibody (approximately 0.2 µL antibody per 100 µL staining buffer)

   Antibody concentrations will vary from lot to lot, so it is important to re-calculate the appropriate dilution for each new antibody stock. Similarly, one should note that the isotype and primary antibody stock concentrations are not the same, but the final concentration should be equal.

   Additionally, it may be necessary to make serial dilutions of the primary antibody in staining buffer to determine the appropriate final concentration for robust staining using new antibody lots.

9. Incubate cells with primary antibodies for 30 minutes at 25°C.

10. After 30 minutes of staining, spin down cells for 5 minutes at 4°C and 300 RCF and wash by resuspending in 1 mL staining buffer. Repeat 2 times.

11. Spin down cells for 5 minutes at 4°C and 300 RCF and resuspend in donkey anti-rabbit IgG antibody, AlexaFluor 488 conjugated, diluted 1:500 in staining buffer.

12. Incubate cells in dark in secondary antibody for 30 minutes at 25°C.

13. After 30 minutes of staining, spin down cells for 5 minutes at 4°C and 300 RCF and wash by resuspending in 1 mL staining buffer. Repeat 2 times.

14. Spin down cells for 5 minutes at 4°C and 300 RCF and resuspend in FACS buffer. Transfer to FACS tubes and analyze on a suitable flow cytometer for NKX2-1 (AlexaFluor 488-labeled) expression.

    Note that the isotype controls are typically much lower in fluorescence than the NKX2-1 negative cells that were stained with the antibody. It is typically more effective and informative to draw gates to estimate NKX2-1 induction efficiency using a negative control (e.g. undifferentiated cells) stained with anti-NKX2-1 antibody in the same manner described above). Typical FACS plots and gating strategy are shown in Figure 7.

    When performing flow cytometry to evaluate the outcome of CD47^high/CD26^low sorting, enrichment can be scored by comparing the percentage of NKX2-1+ cells in the post- to pre-sort populations as a ratio. For example, if the pre-sorted population is 45% NKX2-1+ and the post-sorted population is 90% NKX2-1+, then the enrichment score would be: 90/45= 2 or “2-fold enrichment for NKX2-1.”

Figure 7.

Representative plots of intracellular flow cytometry staining for NKX2-1. (a) Dot plots showing typical gating algorithm to evaluate NKX2-1 expression by first gating for size and granularity (FSC vs. SSC) then for NKX2-1 expression relative to a negative control population e.g. CD47^low (depicted) or undifferentiated PSCs). (b) From left to right, dot plots indicate isotype control and NKX2-1-stained unsorted and sorted CD47^high populations. Note the low staining levels in the isotype control relative to the NKX2-1⁻ cloud in the stained, unsorted population.

Reagents and Solutions

Complete Serum Free Differentiation Medium (cSFDM)

375 mL 1× Iscove's Modified Dulbecco's Medium (IMDM, Invitrogen, cat. no. 12200-036)

125 mL Ham’s F12 (e.g. Invitrogen, cat. no. 11765-054)

5 mL B27 with retinoic acid (ThermoFisher, cat. no. 17504044)

2.5 mL N2 (ThermoFisher, cat. no. 17502048)

500 µL ascorbic acid, 50 mg/mL

Ascorbic acid can be prepared by diluting stock powder (e.g. Sigma A4544-25G) in sterile, tissue-culture grade water to a final concenvration of 50 mg/mL

19.5 µL monothioglycerol, 500 µg/mL (Sigma, cat. no. M6145-25ML)

3.75 mL Bovine Albumin Fraction V, 7.5% solution (Gibco, cat. no. 15260-037)

5 mL Glutamax (Gibco, cat. no. 35050)

500 µL primocin (Invivogen, cat. no. ant-pm-2)

Filter sterilize (for example, with Millipore Steritop Sterile Vacuum Bottle-Top Filters, cat. no. SCGPS01RE) and store for up to one month at 4°C, kept away from light.

The following recipes for Y-27632, SB431540, Dorsomorphin, and CHIR99021 are examples of possible dilution factors for each of these compounds. The molecular weight of these chemicals is batch-specific. It is important to recalculate the volume of water added to reach the desired final concentration based on the molecular weight printed on the label.

10 mM Y-27632 (Y)

10 mg Y-27632 dihydrochloride (Tocris, cat. no. 1254)

3.1 mL sterile, tissue-culture grade water

Filter sterilize (for example, with EMD Millipore Sterile Disposable Vacuum Filter Units, cat. no. SCGP00525).

Aliquot and store at −80°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 month.

10 mM SB431543 (SB)

10 mg SB431542 (Tocris, cat. no. 1614)

2.6 mL DMSO

Filter sterilize.

Aliquot and store at −20°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 week.

2 mM Dorsomorphin (DS)

2 mg Dorsomorphin (Stemgent, cat. no. 04-0024)

2.0 mL DMSO

Filter sterilize.

Aliquot and store at −20°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 week.

3 mM CHIR99021 (CHIR)

CHIR99021 (Tocris, cat. no. 4423)

7.1 mL DMSO

Filter sterilize.

Aliquot and store at −20°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 week.

10 µg/mL recombinant human BMP4

50 µg recombinant human BMP-4 protein (R&D Systems cat. no. 314-BP-050)

5 mL 0.1% BSA in 4 mM HCl

Prepare 0.1% BSA in PBS by diluting 13 µL Bovine Albumin Fraction V, 7.5% solution in 1 mL of 4 mM HCl (prepared from HCl stock solution with sterile, tissue culture grade water).

Filter sterilize.

Aliquot and store at −80°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 week.

100 µM retinoic acid (RA)

Prepare 5 mM stock solution:

50 mg retinoic acid (Sigma, cat. no. R2625-50MG)

33.24 mL DMSO

Aliquot and store at −80°C for up to 1 year, kept away from light.

Prepare 100 µM working solution:

1 mL 5 mM stock solution.

49 mL molecular biology grade ethanol

Store at −20°C for up to 1 month, kept away from light.

FACS buffer

49.5 mL 1× phosphate buffered saline

0.5 mL fetal bovine serum (e.g. Gibco, cat. no. 10082139)

Sort buffer

50 mL FACS buffer

50 µL 10 mM Y-27632

10× cAMP/IBMX Stock

50 mL cSFDM base

21.5 mg 8-Bromoadenosine 3’,5’-cyclic monophosphate sodium salt (cAMP, Sigma-Aldrich, cat. no. B7880-100MG)

500 uL 0.1 M IBMX (3-isobutyl-1-methylxanthine, Sigma, cat. no. I5879)

Filter sterilize and store at 4°C for up to 1 month.

250 µg/mL recombinant human FGF2

1 mg recombinant human FGF basic protein (R&D Systems cat. no. 233-FB-025)

4 mL 0.1% BSA

Prepare 0.1% BSA in PBS by diluting 13 µL Bovine Albumin Fraction V, 7.5% solution in 1 mL PBS.

Filter sterilize.

Aliquot and store at −80°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 week.

10 µg/mL recombinant human FGF10

25 µg recombinant human FGF10 protein (R&D Systems cat. no. 345-FG-025)

2.5 mL 0.1% BSA

Prepare 0.1% BSA in PBS by diluting 13 µL Bovine Albumin Fraction V, 7.5% solution in 1 mL PBS.

Filter sterilize.

Aliquot and store at −80°C for up to 1 year.

After thawing individual aliquots, store at 4°C for up to 1 week.

100 µM dexamethasone

Prepare 1 mM stock:

Dexamethasone powder (Sigma, cat. no. D4902-25MG)

63.7 mL molecular biology grade ethanol

Store at −20°C for up to 2 years.

Prepare 100 µM stock:

500 µL 1 mM dexamethasone stock

49.5 mL molecular biology grade ethanol

Aliquot and store at −20°C for up to 1 year. This is the working concentration.

Airway Differentiation Medium

45 mL cSFDM base

5 mL 10× cyclic AMP/IBMX stock

50 µL 250 µg/mL rhFGF2

500 µL 10 µg/mL rhFGF10

25 µL dexamethasone

50 µL 10 mM ROCK inhibitor

Store at 4°C for up to 2 weeks, kept out of light.

2 mg/mL dispase II

100 mg Dispase II, powder (ThermoFisher, cat no. 17105041)

50 mL DMEM

Dissolve and filter sterilize. Store at 4°C for up to 2 weeks or aliquot and freeze at −20°C for up to 6 months.

Commentary

Background information

This procedure builds on years of careful work developing directed differentiation approaches for early and later stages of early lung development, from definitive endoderm to lung specification. Initially, methods were generated to derive definitive endoderm from both mouse and human PSCs, mainly via the activation of nodal signaling and consequent gastrulation and formation of the anterior primitive streak via high-dose Activin A(Kubo, 2004; D'Amour et al., 2005). Additional studies using Bry and Foxa2 reporters have revealed that BMP, Wnt, and nodal signaling are required for and drive the formation of the primitive streak from PSCs(Gadue et al., 2006; Loh et al., 2014; D'Amour et al., 2005). However, continued activation of BMP and Wnt after establishment of the primitive streak result in generation of a more posterior identity, resulting in emergence of mesodermal fate at the expense of definitive endoderm(Loh et al., 2014).

PSC-derived definitive endoderm expresses early endodermal lineage markers (e.g. Foxa2, Sox17), differentiates to multiple endodermal lineages after transplantation in the mouse kidney capsule(Kubo, 2004; D'Amour et al., 2005) and can be driven in vitro with further patterning to more differentiated endodermal lineages, such as hepatocytes(Gouon-Evans et al., 2006). These PSC-derived endodermal progenitors can also be patterned, much like the in vivo early endoderm, to subsets of foregut cells. Relevant to the goal of differentiating lung lineages is the ability to drive the expression of anterior foregut markers via the dual inhibition of BMP and TGFβ signaling(Green et al., 2011). When derived from either mouse or human PSCs, these anterior foregut cells are then able to respond to FGF2 and BMP signaling to differentiate to Nkx2-1+Pax8+ thyroid progenitors(Longmire et al., 2012; Kurmann et al., 2015) or to combinatorial Wnt, retinoic acid, and BMP signaling to differentiate to Nkx2-1+Pax8- lung progenitors (Longmire et al., 2012; Huang et al., 2013; Rankin et al., 2016). Recent work has characterized these NKX2-1+ lung progenitors in detail and demonstrated that they can differentiate to SFTPC+ distal lung epithelium in response to trophic stimulation by defined factors (e.g. CHIR99021, FGF10, KGF, steroids) or co-culture with mouse lung mesenchyme(Bilodeau et al., 2014; Hawkins et al., 2017; Huang et al., 2013). The reproducibility across human PSC lines, relative efficiency, and thorough characterization of this method to generate lung endoderm is the central advantage of the approach described here.

As directed differentiation can be inefficient and heterogeneous, the engineering of reporter cell lines allowing for live-cell sorting of populations of interest is critical to the study of cell fate decisions during directed differentiation. For example, our lab has previously published the use of reporter lines (e.g carrying GFP targeted to the Nkx2-1 locus) for assessing the efficiency of endodermal lung or thyroid lineage specification in mouse (Longmire et al., 2012) or human PSCs(Hawkins et al., 2017). These reporter lines can be generated via a number of gene engineering approaches to introduce both a DNA double-strand break and repair template containing the reporter sequence targeted to a gene locus of interest.

In the absence of reporter lines, cell surface markers can be used as a surrogate to enrich for the population of interest. To identify PSC-derived NKX2-1+ lung epithelial progenitors, two reliable cell surface markers, CPM and CD47, have recently been described(Gotoh et al., 2014; Hawkins et al., 2017).

Finally, the efficient differentiation of purified NKX2-1+ lung progenitors to mature airway epithelial lineages has only recently been described (Konishi et al., 2016; McCauley et al., 2017). The approach described in this protocol builds on the concept of directed differentiation of early progenitors to a defined population subset (e.g. proximal airway epithelium) by manipulation of developmentally-relevant signaling pathways. In this case, the protocol described here represents a method using three-dimensional culture of purified cells in conditions with low-to-no canonical Wnt signaling to drive the formation of airway epithelial spheres. These spheres are “epithelial-only” and contain lung secretory and basal epithelial cells, but no multiciliated cells unless a Notch inhibitor is added (likely due to high levels of Notch signaling as shown in McCauley et al., 2017). While their epithelial-only composition makes these spheres amenable to experiments studying airway epithelial cell biology, additional mesenchymal, vascular, and immune lineages would need to be added to this model in order to develop the multi-lineage models required for future recapitulation of the full complexity of in vivo airway structure and function.

Critical parameters

• Pluripotency of undifferentiated PSCs

  Differentiation efficiency relies on the pluripotency and health of the undifferentiated PSCs prior to replating for definitive endoderm. PSCs should be maintained in standard culture. All cell lines used for differentiation should be evaluated for karyotypic stability, absence of mycoplasma contamination, and pluripotency using standard techniques. Furthermore, a pipette should be used to scrape and discard any undifferentiated contamination prior to passaging cells to initiate differentiation to ensure robust induction of definitive endoderm.

• Line to line variability

  Pluripotent human ESCs are well known to have variable in vitro differentiation potential across lines. In the directed differentiation protocol described here, for example, NKX2-1 specification efficiency across diverse PSC lines can vary from 5 – 95%. With this in mind, it is important to characterize the NKX2-1 specification efficiency of any new PSC line and optimize the parameters that can contribute to NKX2-1 efficiency. In particular, the timing and density of replating of definitive endoderm for anteriorization is a critical step in the efficiency of directed differentiation of PSCs to lung epithelium(Hawkins et al., 2017). In differentiating any new line, one should test the effect of various timings and densities of definitive endoderm on the downstream NKX2-1 induction efficiency; for example, by replating endoderm every 12 hours between 60 and 96 hours. This experiment can furthermore be a useful training exercise that serves to emphasize the importance of careful timing of the endodermal stage on downstream differentiation capacity. One should take care to ensure that the anteriorization duration (e.g. 72 hours) is maintained for each endoderm timepoint.

• Negative controls

  To accurately evaluate the efficiency of NKX2-1 specification and of later airway differentiation, it is critical to include appropriate negative controls for both immunofluorescence and flow cytometry analyses as well as for reverse transcriptase quantitative PCR (RT-qPCR) analysis. Although undifferentiated PSCs are a convenient negative control, it can be difficult to interpret staining patterns in these cells in comparison to differentiated cells. It can therefore be useful to differentiate cells to a non-lung lineage, such as hepatocytes, using standard published protocols. These cells can then be used as negative controls, in particular when one is learning the expected NKX2-1 staining patterns. Comparisons to positive control primary lung cells or tissues is also recommended, particularly when employing RT-qPCR analyses of PSC-derived putative lung lineages.

• Selection of CD47^high/CD26^low gate

  As most cells (both lung and non-lung) generated in the approach described in Basic Protocol 1 express CD47, the gate for CD47^high expression intersects the positive cloud. Enrichment of the NKX2-1+ population is directly dependent on the location of this gate, which is most effective the higher it is on the CD47 channel. Typically, it is recommended to position the lower limit of the gate for the CD47^high population just above where the CD26+ cloud intersects the CD47+ population. Figure 4 provides several representative examples of acceptable CD47 gating based on this strategy.

• Evaluation of lung progenitor differentiation

  This protocol emphasizes the use of NKX2-1 expression as a metric to measure the efficiency of differentiation to the early lung progenitor lineage. However, there are several additional criteria that must be met to satisfy the designation of a cell as a “lung progenitor.” Other than NKX2-1, there are no additional validated markers for early lung progenitors, and indeed NKX2-1 itself is also expressed early in thyroid endoderm and in forebrain neuroectoderm. The ideal approach, therefore, is to rule out the presence of other non-lung NKX2-1+ lineages (thyroid and forebrain). Lung progenitors should be NKX2-1+PAX8-TG- (non-thyroid) and NKX2-1+OTX2-SIX3- (non-forebrain) (Hawkins et al., 2017). To exclude these lineages, expression of these recommended markers should be evaluated by immunofluorescence microscopy or RT-qPCR in Day 15 lung progenitors. Lung progenitors should also be evaluated for their downstream lung competence by differentiation to cell types expressing highly lung-specific markers including 1) SCGB3A2+ proximal airway epithelium(McCauley et al., 2017) or 2) surfactant protein-C (SFTPC+) distal lung epithelium (Hawkins et al., 2017; Huang et al., 2013). Expression of other popular “lung” markers, such as P63, CFTR, mucins, keratins, FOXJ1, or SFTPB in isolation are not lung specific, and in the absence of co-expression of NKX2-1 or in the absence of direct derivation from NKX2-1+ precursors cannot be reliably interpreted as proof of lung lineage.

  While the protocol in this unit has not been found to produce neural or thyroid contaminants (Hawkins et al., 2017; McCauley et al., 2017), it remains critical for researchers to evaluate the expression of these markers and the downstream competence of these progenitors when using this protocol for their research.

• Evaluation of airway epithelial sphere differentiation

  It is critical to evaluate the final, differentiated proximalized lung spheres to ensure their airway identity. For initial experiments, spheres should be evaluated at least two weeks post-sorting. At this stage, the majority of spheres should be comprised of epithelial (EPCAM+) cells with a distinct lumen and the majority of cells should express SOX2(McCauley et al., 2017). Approximately 40–90% of cells should maintain expression of NKX2-1, and these NKX2-1+ cells can be segregated into multiple subsets, such as: SCGB3A2+ secretory cells, and P63+ cells, a portion of which also co-expresses KRT5+ reminiscent of in vivo airway basal cells (McCauley et al., 2017). The identity of NKX2-1 negative cells in these spheres and whether these cells arise from the outgrowth of rare NKX2-1 negative cells present in the day 15 sorted samples remains uncertain to date, and may represent contaminating non-lung endodermal lineages that persist at some level throughout our protocol. Expression of markers can be evaluated by immunofluorescence microscopy or by transcriptomic methods (e.g. RT-qPCR), although the former is preferred as it provides a per-cell metric of the population subsets. When evaluating the identity and differentiation of these airway epithelial spheres, it is important to keep in mind the approach and the goals of the experiment. For example, it may be relevant to characterize the expression pattern and population segregation of a gene of interest in a particular airway disease, such as CFTR for cystic fibrosis.

Troubleshooting

• Low endoderm induction efficiency

  Endoderm induction, as measured by both nuclear expression of FOXA2 and by coexpression of the cell surface markers CD117 and CXCR4, is typically extremely efficient (>90%) using the StemDiff Definitive Endoderm Kit (StemCell Technologies). In general, troubleshooting of inefficient endoderm induction is well-described in the kit protocol, and therefore will not be discussed in detail here. In brief, the most common cause is low pluripotency of starting cultures, which can be addressed by ensuring that the starting cells have less than 5% spontaneous differentiation by eye prior to passaging for differentiation (see Figure 2). Low confluency of starting cultures post-passage for definitive endoderm induction can also contribute to poor induction efficiency.

  One important consideration that is more specific to this particular protocol is that low endoderm efficiency may result from evaluating the induction too early in differentiation; while 72 hours is typically long enough for >90% of cells to be CD117 and CXCR4 double-positive, some lines may require longer differentiation time before one observes efficient endoderm induction by this metric.

• Poor cell survival after endoderm replating

  Cell survival post-passaging after endoderm differentiation (typically Day 3) can be a critical barrier to the success of this protocol. Low cell survival typically indicates that the cells were either dissociated into too-small clumps, typically by excessive pipetting. One can improve cell survival by 1) using a cell scraper to dislodge cells from the confluent monolayer after dissociation by Gentle Cell Dissociation Reagent (StemCell Technologies), 2) pipetting cells as few times as possible (under 4 total is preferable) and 3) using a 5 mL pipette instead of a p1000 (or similar), to reduce shear forces. It can be useful to visualize the cells after replating to observe the initial size of the clumps, which should typically be between 15–20 cells in size. Clumps smaller than this, especially in more sensitive lines, may result in cell death during the anteriorization stage.

• Low NKX2-1 induction efficiency

  Other than cell survival, the most common challenge in this directed differentiation protocol is achieving consistent and efficient downstream induction of NKX2-1 expression. While substantial line-to-line variability (as well as some run-to-run variability within lines) is expected, consistent differentiations resulting in low-to-no differentiation of NKX2-1+ progenitors can result from several common factors:

  First, as has been previously mentioned several times, efficient induction of definitive endoderm is required for effective downstream differentiation of lung progenitors. Furthermore, differentiation is additionally sensitive to the timing of definitive endoderm induction. Therefore, as has been previously recommended, a careful time kinetic of endoderm induction may improve downstream NKX2-1 efficiency(Hawkins et al., 2017).

  The confluency of cells after replating from endoderm to anterior foregut endoderm (typically Day 3) is also critical. If cultures are too confluent on Day 4, many lines can fail to generate appreciable numbers of NKX2-1+ progenitors during the subsequent lung induction stage. See Figure 2 for the typical appearance of cells on Day 4.To address this issue, one can either 1) test multiple replating densities to determine the optimal density for each line, or 2) passage cells using 0.05% Trypsin-EDTA on day 9–10 of differentiation at approximately a 1:3 split ratio.

  Finally, as NKX2-1 induction relies on coordinated and time-sensitive signaling pathway activation or inhibition, it is important that growth factors are carefully diluted and aliquoted to ensure accurate concentrations of each stock, which should be stored at the recommended temperature and should not be kept beyond the recommended duration. As one additional point, it is important to keep track of the lot number of each stock of growth factors aliquoted, in case an issue should arise.

• Low NKX2-1 enrichment from CD47 sorting

  NKX2-1 enrichment is contingent on a minimum threshold of initial differentiation. Differentiation efficiencies of less than 10% NKX2-1+ cells are not amenable to effective enrichment by CD47/CD26 sorting. If the differentiation efficiency is adequate, other possible explanations for low enrichment include inefficient staining. This can be evaluated by comparing the stained cells to the isotype. Almost all stained cells should be CD47+ (brighter staining) compared to the isotype antibody-stained control cells. See Figure 4 for typical staining patterns for various differentiation efficiencies.

  One final possible explanation for low NKX2-1 enrichment is the selection of the CD47 gating parameters. Higher (more stringent) CD47 gating results in increased enrichment for NKX2-1 expression.

• Low cell survival after sorting and replating

  The most common explanation for low cell survival after sorting is high levels of cell death during dissociation, which stresses the surviving cells and extends the time required to sort an adequate population of live cells. To avoid this, cells should be dissociated primarily through enzymatic methods. The most common pitfall of this approach occurs when cells die after dissociation or sorting. The most common explanation for this cell death is over-pipetting of cells during dissociation steps. A good guideline is that >65% of cells should fall within the size gate on FSC vs. SSC and, of these cells, >90% should be viable based on calcein blue or other viability staining (See Figure 4). Numbers lower than these, in particular increased levels of debris on the FSC vs. SSC dot plot, indicate excessive pipetting during dissociation.

  Other potential explanations for low cell survival after sorting include the settings and nozzles used on the sorting instrumentation itself and the experience of the sorter and the investigator replating the cells, which can be remedied by open dialogue about the instrument parameters and continued practice with sorting and replating cells.

Anticipated Results

Using the described approach for directed differentiation, NKX2-1 induction efficiencies of 30 – 50% can be routinely expected from most PSC lines with some expected variability. For example, some lines routinely differentiate at >80% efficiency (e.g. BU3(Hawkins et al., 2017; Kurmann et al., 2015)) and others more typically are <5% NKX2-1+ (e.g. H9 ESCs; also known as WA0- or NIH registry #0062). As discussed, this efficiency can be further optimized by careful design of experiments testing various differentiation parameters (e.g. cell density).

After sorting and replating in three-dimensional culture, sphere formation should be efficient. By 2 weeks post-replating (e.g. total differentiation day 30), the majority of spheres are luminal, EPCAM+ and SOX2+. NKX2-1 expression is typically maintained in 40 – 90% of cells within these spheres. Differentiation to airway epithelial cell types including SCGB3A2+NKX2-1+ secretory cells and P63+KRT5+NKX2-1+ basal cells is easily detected by immunostaining. While this protocol does not reliably produce multiciliated cells, these cells can be generated within the spheres by either pharmacological inhibition of Notch signaling or by transfer of single-cell suspensions prepared from these spheres into 2D cultures on transwell inserts for expansion and differentiation at the air-liquid interface(McCauley et al., 2017). Additionally, the relative ratios of the populations and the rate of emergence of these cell types (e.g. SCGB3A2+ airway progenitors vs. P63+KRT5+NKX2-1+ basal-like cells, etc.) is variable from run-to-run and should be routinely evaluated for each differentiation at the protein or transcriptome level, dependent on the experimental goals.

Experimental Design Considerations

The major important experimental design consideration, aside from the differentiation timing, is planning the number of wells of endoderm with which to begin the differentiation. As cells are passaged many times throughout the protocol, it is often possible to begin differentiation with just a single well of ~1 × 10⁶ cells for definitive endoderm induction (plus one additional well to use to evaluate the definitive endoderm induction and end with many confluent wells by the end of lung specification (for example, 6 wells of approximately 3 × 10⁶ cells each by day 15). The typical differentiation yield is approximately 4 – 18 NKX2-1+ progenitor cells on day 15 per input undifferentiated day 0 PSC. Therefore, to save time and resources, it is important to consider how many cells might be required for the final steps of the experiment and plan the number of wells of definitive endoderm accordingly.

Time considerations

Although this protocol in its entirety is lengthy (~3–5 weeks), many days require only changing media. Therefore, it can be useful to plan the differentiation experiments around the key days of definitive endoderm induction (typically Days −1 to 3), beginning lung specification (Day 6), and harvesting lung progenitors for analysis or downstream applications (Day 15). This lung progenitor timepoint can be varied around the Day 15 timepoint. Typically, one should plan to harvest the lung progenitors sometime between Day 14 and Day 17. After sorting and replating, the cells can be harvested anytime after they have recovered from sorting and formed distinct epithelial spheres, typically 1–2 weeks after sorting.

Related to this, planning an experiment can be challenging because it can be difficult to know when the undifferentiated cells will be ready to passage to begin differentiation. It is important to practice culturing and evaluating undifferentiated cell cultures to be able to anticipate the window when cells will be ready.

From a practical standpoint, the duration of sorting itself can vary depending on the experience of the investigator. With practice and the appropriate flow cytometer, an investigator can dissociate, sort on CD47 and CD26, and replate a 6-well plate of PSC-derived NKX2-1+ lung progenitors in approximately 3 hours.

Significance Statement.

Diseases affecting the epithelium of the lung include diverse pathologies, ranging from mongeneic disorders such as cystic fibrosis to complex acquired illnesses, such as asthma and chronic obstructive pulmonary disease (COPD). In spite of their clinical burden, the pathogenesis of these diverse lung diseases is incompletely understood and treatment options for patients remain limited. One potential option to improve understanding and develop new therapies for these disorders is the establishment of new in vitro models via the differentiation of human pluripotent stem cells to proximal airway epithelial cells using defined developmental signaling pathways. This protocol describes a robust approach for the generation and characterization of proximalized airway epithelial “bronchospheres” from pluripotent stem cells to address this need.