Generation of liver mesenchyme and ductal cell organoid co-culture using cell self-aggregation and droplet microfluidics

Summary Within the peri-portal region of the adult liver, portal fibroblasts exist in close proximity to epithelial ductal/cholangiocyte cells. However, the cellular interactions between them are poorly understood. Here, we provide two co-culture techniques to incorporate liver portal mesenchyme into ductal cell organoids, which recapitulate aspects of their cellular interactions in vitro. We integrate several techniques from mesenchyme isolation and expansion to co-culture by microfluidic cell co-encapsulation or 2D-Matrigel layer. The protocol is easily adaptable to other cells from other organs. For complete information on the generation and use of this protocol, please refer to Cordero-Espinoza et al.1


SUMMARY
Within the peri-portal region of the adult liver, portal fibroblasts exist in close proximity to epithelial ductal/cholangiocyte cells. However, the cellular interactions between them are poorly understood. Here, we provide two co-culture techniques to incorporate liver portal mesenchyme into ductal cell organoids, which recapitulate aspects of their cellular interactions in vitro. We integrate several techniques from mesenchyme isolation and expansion to co-culture by microfluidic cell co-encapsulation or 2D-Matrigel layer. The protocol is easily adaptable to other cells from other organs. For complete information on the generation and use of this protocol, please refer to Cordero-Espinoza et al. 1

BEFORE YOU BEGIN
In Cordero-Espinoza et al., 1 we reported that when primary liver mesenchyme is co-cultured with liver ductal epithelium in 3D classical Matrigel dome, the cells fail to establish cell-cell interactions. The cells segregate in the well with the mesenchyme attaching to the plate bottom and the ductal cells forming organoids inside of the Matrigel (for reference see Figure S4M from ref. 1 ). Therefore, we developed two co-culture methods: a microfluidics -based method and a 2D Matrigel layer coculture method, to enable the cell-cell interactions between both populations in vitro. The microfluidics approach facilitates the cell-cell interactions recapitulating the 3D architecture (See Figure 5 and Figure S5 from ref. 1 ), while the 2D-co-culture method allows more control of the ratios between cell types (See Figure 6 and Figure S6 from ref. 1 ) as well as it is amenable to chemical and siRNA manipulation (See Figure 7 from ref. 1 ).
The protocol below explains the specific steps for co-culturing primary liver ductal organoids with primary liver mesenchymal cells, mainly portal fibroblasts. Both, mesenchymal cells and ductal cells can be co-cultured directly after isolation or after several sub-cultivations (2-5 passages for portal mesenchyme and any passage for ductal organoids). Notably, at the passages tested, the cells do not present any change on their phenotype following culture (for details see Figure 4D-E and Figures S4D-S4F from ref. 1 ). Workflows involving microfluidic cell encapsulation and co-culture protocols in Matrigel are broadly applicable to other cell types.
Cells were isolated from PDGFRa-H2B-GFP mice (B6.129S4-Pdgfra tm11(EGFP)Sor /J) 2 which can be crossed with mTmG mice (Gt(ROSA)26Sor tm4(ACTBtdTomato,-EGFP)Luo /J) 3 for the easier double fluorescence labeling of the mesenchyme. For the PDGFRa-H2B-GFP line, these mice will only be bred as heterozygous, because homozygous mice are not viable. A wild-type mouse is also necessary as a control for sorting in each experiment. The detailed protocol is applicable to adult mice of all ages although the mice used for the experiments presented here are adults between 8-12 weeks of age. From one mouse, the total cell yield is 10 000-40 000 of all cells achieved with this digestion method. Typical total number of cells from this kind of preparation ranges from 10-80 million cells. Typical viability after sorting is >90%.
In addition, this protocol describes production of microfluidic chips made of polydimethylsiloxane (PDMS) by soft lithography. The design of the microfluidic chip (triple inlet flow focusing device) is available as a CAD file from https://openwetware.org/wiki/DropBase:Droplet_3_inlets. Microfluidic chips were produced based on master molds obtained via soft lithography with an SU-8 photoresist on a silica substrate, as described elsewhere, 4,5 in polydimethylsiloxane (PDMS).

Microfluidic chip production
Timing: 24 h For a detailed illustration of the chip production please refer to Figure 1.
a. Mix SYLGARDä 184 Silicone Elastomer and curing reagent (10:1 ratio) in a plastic cup and mix thoroughly ( Figure 1C). b. De-gas the SYLGARDä 184 Silicone Elastomer mixture in a desiccator under vacuum until no air bubbles raise to the surface ( Figures 1D and 1E). c. Pour the de-gassed SYLGARDä 184 Silicone Elastomer mixture onto the SU-8-derived silica master mold ( Figure 1F). d. De-gas the freshly poured SYLGARDä 184 Silicone Elastomer again in a desiccator under vacuum ( Figure 1G).
CRITICAL: Ensure that the SYLGARDä 184 Silicone Elastomer mixture is devoid of bubbles as entrapped air will cause undesirable cavity formation during polymerization.
Note: SYLGARDä 184 Silicone Elastomer can also be cured at 19 C-25 C (room temperature). However, this process will take at least 48 h.
3. After polymerization, use a scalpel to carefully cut the SYLGARDä 184 Silicone Elastomer (PDMS) ( Figures 1I-1K). a. Use a 1 mm disposable biopsy needle to punch the tubing inlet and outlets of the PDMS chip ( Figures 1L and 1M). b. Wash the PDMS chip briefly (5-20 s) by placing it in a 50 mL test tube filled with 2-propanol and vortex it for 10 s. c. Dry the PDMS with compressed air. d. Clean microscopy slides (as many as PDMS chips need to be produced) by removing any dust particles with scotch tape.
CRITICAL: The inlets/outlets should fit with the desired tubing's outer diameter, where the tubing can be fitted with ease whilst sitting there tightly (i.e. resistant to a mild pulling force).  a. Fill a 1 mL plastic syringe with the 1% PFOCTS solution and attach a needle with 5 cm long tubing. 6. Oxygen plasma: treat the PDMS chips and microscopy slides for 12 s with oxygen plasma ( Figure 1O). 7. Directly afterward, flip the PDMS chips by 180 (now with the channels facing downwards again) and gently press them onto the microscopy cover slides ( Figure 1P). 8. Silane treatment: flush channels with the freshly prepared and filtered solution of 1% (v/v) PFOCTS in HFE-7500. 9. Incubate the PFOCTS-filled PDMS chips for 15 min at 65 C (in an oven or a heating plate).
CRITICAL: The placing of PDMS onto the coverslips and the subsequent silane treatment must be done immediately after the plasma treatment. Protocol CRITICAL: prepare immediately before the experiment, cannot be stored and reused in suspension.
CRITICAL: prepare immediately before the experiment, cannot be stored and reused in suspension.
CRITICAL: prepare immediately before the experiment, it is not recommended to keep dissolved. Isolation of the primary cells from the liver tissue using mechanical disruption and enzymatic digestion. This step produces a single cell suspension of various liver cells (adapted from ref. 6 ).
1. Euthanize the mice and dissect the liver; all standard euthanasia methods yield similar results for isolation. a. Wash the livers several times to remove as much blood as possible; stop when PBS is mostly clear.
Note: No gentle washing necessary -vigorous washing is allowed at this step.
b. Transfer livers to a petri dish, remove gallbladder and any peritoneal fat or non-liver pieces of tissue. c. Remove all PBS or transfer to a new petri dish without PBS and chop the liver into pieces with scissors or with 1 razor blade; make sure pieces are very small ( 1-2 mm length and width). d. Transfer the pieces to a new tube containing 10 mL of PBS. 2. Let the pieces settle down by gravity or, alternatively, spin at 120 Relative Centrifugal Force (RCF)/2 min. Discard the supernatant to remove as much blood as possible. 3. Wash the pellet with 40 mL of PBS by inverting the tube several times 4. Repeat step 2. 5. Add 10 mL of collagenase/dispase media mix (top up to approx. 12.5 mL, adjust it to the tissue pellet size). In that step the addition of DNAse is optional. 6. Incubate 10 min at 37 C in a water bath. 7. Pipette up and down with 10 mL serological pipette. 8. Spin at 120 RCF/5 min and discard the supernatant to remove as many blood cells as possible. 9. Resuspend in 25 mL of collagenase/dispase media mix supplemented with DNAse. 10. Incubate 90 min at 37 C in a horizontal shaker at 130 RPM. 11. Check digestion under a brightfield microscope; some ducts and some parenchyma should be present in the mixture.
CRITICAL: In contrast to ductal isolation from previous publication, 6 make sure isolated ducts have some parenchyma present on them and do not appear completely cleared of parenchyma ( Figure 2).  There is a lot of debris in this type of liver digestion. 43. Remove cell doublets by gating as shown in Figure 3B, using a forward-scatter area vs. height plot. 44. Gate on PECy7 negatives (this step removes any remaining blood, immune and endothelium fractions) and APC-EpCAM negatives (this step removes liver ductal cells) -the double negatives are shown in Figure 3C. 45. When sorting ductal cells from the same preparation, gate and sort the EpCAM + cells ( Figure 3C, gate 1). At this step, a typical expected yield is 50 000-100 000 ductal cells/liver, with >90% viability. 46. Gate on SCA1-SB436 + and PDGFRa-GFP + double positives (you can also include SCA1if needed), as shown in Figure 3D. 47. Collect cells in 300 mL of Advanced DMEM/ F12 +++ media. At this step, a typical expected yield is 10 000-40 000 SCA1 + PDGFRa-GFP + mesenchymal cells/liver, with >90% viability. 48. After sorting, spin cells at 500 RCF/5 min. 49. Resuspend cells in 200 mL of MM and plate in 1 well of 96 well tissue culture-treated plates at a ratio of up to 10 000 cells/well. Proceed to step 50 for culturing.

Take off
Note: If the PDGFRa-H2B-GFP mouse model is not available, antibody staining against murine PDGFRa protein is possible.
Note: Here, gating on EpCAM positives allows isolation of the ductal cells as well using the same antibody staining strategy.
Note: Certain FACS-sorting machines have issues with crosstalk between PE (tdTomato channel) and PE-Cy7, as the PE-Cy7 composite fluorophore also has a small emission peak at the spectral height of PE; additionally, PE-Cy7 can be degraded and therefore provide additional ll OPEN ACCESS fluorescence confounding signal of PE; to avoid this, we recommend the Brilliant Violet 650 conjugates to the same monoclonal antibodies (excitation: laser line 405 nm, emission: similar to PE-Cy7).
CRITICAL: The 96-well plate has to be tissue culture treated for attaching cells, e.g. flat bottom Nunclon Delta-Treated plates from ThermoFisher.

Timing: 2 weeks
Expansion and passaging of liver portal fibroblast population in standard cell culture on plastic dishes.
Note: We do not recommend passaging mesenchymal cells described here for more than 2 passages, as we observed change in appearance ( Figures 4A and 4B, passage 3 -more spread cells) as well as an increase in the expression of activated fibroblast marker aSMA, as noted in our previous manuscript (see ref. 1 Figure S4F).

Timing: 1 h
Co-culture method description, which allows combination of liver mesenchyme and ductal organoids in a layer of Matrigel. This co-culture model allows for direct control of respective number of cellcell contacts, which scale proportionally to the number of cells added (i.e., the cell-cell ratio).
CRITICAL: Matrigel will solidify at 19 C-25 C, so work quickly and keep the basement matrix cold throughout the process.
Note: In the classic matrigel dome 3D culture, the 2 cell types co-cultured together segregate, with the mesenchymal cells preferentially attaching to the plate bottom and the ductal cells forming ductal organoids without any mesenchymal contact. For more information, please refer to Figure S4M in ref. 1 .

Note:
The expanded mesenchyme can be used up to passage 2, and ductal organoids until passage 10. Later passages of both cell types have successfully been used to make co-cultures, but we do not recommend this for the mesenchyme, as it shows signs of fibroblastlike activation in prolonged culture. Microfluidic co-encapsulation of liver mesenchyme and ductal cells

Timing: 2 h
Second co-culture method description, which allows combination of liver mesenchyme and ductal organoids in an agarose microgels, and subsequent culture in a 3D matrigel dome. The advantage of using this co-culture method is the better recapitulation of the periportal tract in vitro, however this method does not allow for the control of respective cell numbers in the structure.
84. Prepare a 3% low-melting agarose solution for microfluidic cell co-encapsulation. a. Prepare a 3% (weight/volume) agarose solution in a test tube by weighing e.g., 15 mg agarose for 500 mL PBS. b. To dissolve the agarose, spread the measured-out agarose powder on tube's walls and add an appropriate amount of PBS to ensure homogeneous distribution of the powder in PBS. c. Melt the agarose at 75 C in an orbital thermomixer for 30 min while shaking at 400 RPM. d. Once agarose is melted, cool the agarose solution to 37 C in an orbital thermomixer.
CRITICAL: Low-melting agarose must be used for microfluidic cell encapsulation. Other agarose preparations will solidify when cooled to 37 C. CRITICAL: Agarose must be kept in the thermomixer at 37 C. Cooling it outside of a thermomixer may cause it to cool too quickly, which might cause it to solidify.
Alternatives: In this protocol SeaPrepâ ultra-low melting agarose (LONZA, #50302) is used. However, a variety of different low-melting agarose products are available from different suppliers. If alternatives are being used, the resulting gels may have different biophysical properties.
Note: The tubing recommended here fits with 1 mm inlets/outlets for the microfluidic device used in the ''Before you begin'' step 3a. If necessary, tubing diameter can be adjusted depending on the size of the biopsy needle used to generate the inlets and outlets.
Note: The expanded mesenchyme can be used up to passage 2, and ductal organoids until passage 10. Later passages of both cell types have successfully been used to make co-cultures, but we do not recommend this especially for the mesenchyme, as it shows signs of fibroblast-like activation in prolonged culture.  a. Use HFE-7500 3Mä Novecä (Fluorochem, # 051243) engineered fluid as the carrier oil phase. b. Use Pico-Surf1 surfactant (5%; Sphere Fluidics, #C022) in the above-mentioned carrier oil diluted to 0.3% to stabilize the agarose-in-oil emulsion during encapsulation. c. If surfactant appears milky, filter it though a 0.22 mm filter. 86. Prepare the oil (carrier phase) syringe.
a. Prepare tubing long enough to reach from the syringe (pumps) to the chip. b. Use tweezers to attach the tubing to the needle. c. Fill the oil syringe (2.5 mL, Luer lock glass syringe) with the oil+surfactant solution (step 85: 0.3% Pico-Surf1 in HFE-7500 3Mä Novecä). d. Attach the tubing and needle to the syringe and push the air out, so that the whole tubing is filled with oil. e. Place the syringe in the pump and attach the tubing to the oil inlet of the PDMS chip. 87. Prepare cells for encapsulation.
a. Prepare single cell suspensions of ductal and mesenchymal cells in two different tubes.
Note: for detailed description of preparing a single cell suspension, please refer to step 54 of this protocol for mesenchyme, and step 67 for ductal cells.
b. Pass ductal cell single cell suspension through a 40 mm strainer and subsequently collect at 500 RCF for 5 min. c. Mesenchymal cell suspension does not need to be strained, but if strained use also 40 mm strainer; collect cells at 500 RCF for 5 min. d. Resuspend each of cell types in 50 mL of MM media at a concentration of 0.5-1 3 10 6 cells/ mL. e. Keep cells on ice until mixing with agarose. f. Warm up the cell suspension by placing the cells at 37 C or by holding the tube in your hand before adding the agarose from step 84. g. Slowly add 50 mL of agarose solution to the cells (1:1 ratio) and carefully mix the suspension by pipetting up and down without creating air bubbles. h. The cells are now ready to be loaded into the syringes. Proceed to step 88.
CRITICAL: To achieve the mesenchyme cell concentrations specified above, isolation from several mice is often necessary. A pilot experiment is suggested, to assess how well the primary mesenchyme expands in the in vitro culture.
88. Prepare the cell-laden (aqueous phase) syringes ( Figure 6). a. Cut 2 sets of tubing and attach each to a separate needle ( Figure 6A). b. To see the contents of the tubing better and to avoid losing the tube content, make a loop at the end of the tubing and hold it in your hand so that the tube opening is facing upwards (Figure 6B). c. Fill each syringe with 100 mL of HFE-7500 and attach the needle to the tubing. Make sure that the needle reservoir is also filled with HFE-7500 ( Figure 6C). d. Push out the HFE-7500 with the syringe, leaving the oil only in the tubing ( Figure 6D). e. Create a small air bubble at the end of the tubing by gently pulling the syringe plunger backwards ( Figure 6E). The air bubble will allow monitoring of how far the tubing has been filled with agarose/cell suspension mixture. f. Place the end of the tubing in the agarose/cell suspension mixture and gently fill the tubing by moving the plunger slowly backwards ( Figure 6F). g. Stop filling the tubing just before the air bubble enters the needle. h. Stop filling the tubing before you completely run out of cell-agarose mixture. Avoid air bubbles. i. Connect the tubing to the correct inlets on the PDMS chip. 89. Add tubing for the outlet exiting the chip, and put a 1.5mL test tube on ice to collect the agarose droplets emerging from the tubing -there, they will solidify there and become microgels. 90. Perform microfluidic encapsulation (Figure 7).
a. Switch on all the equipment (computer, pumps, microscope and camera, Figure 7A). b. In the pump software, select the correct syringe for each pump. c. Select the correct flow rate for your experiments; here we used 3 mL/min for the cell syringes and 30 mL/min for the oil syringe ( Figure 7B). d. Modify the flow rates to change droplet size and droplet forming speed in your experiments. e. Run the experiment. f. During the run, watch out for blockages, do not leave the equipment unattended while encapsulating. g. Collect the cell-laden agarose droplets gels in a test tube on ice for solidification and formation of microgels h. Stop all equipment once the syringes are empty of the cell-agarose solutions; the oil syringe with the oil-surfactant can be reused for the next encapsulation if it is not empty.
Note: Depending on microfluidic chip design, other flow rates may be equally suitable and can be experimentally optimized. Optimal flow rates are influenced by the viscosity of the oil and the hydrogel precursor used necessitating empirical adjustment.
91. De-emulsify the microgels. a. At the end of the experiment, wait 5 min for the phases to separate in the microgel collection tube on ice ( Figure 8A). b. Remove as much oil as possible without touching the cell phase with a pipette. It is better to leave some oil rather than disrupt the microgel-laden phase. c. Add 200 mL of MM media to the microgels collection tube.
CRITICAL: Do not pipette up and down, just add media on top of the microgels phase. imaging, seed 20 mL in a well of an IBIDI dish. 94. Leave for 15 min at 37 C and 5% CO 2 incubator for Matrigel solidification 95. Overlay with cell culture media. 96. Monitor the growth of the culture over the next 7-10 days (Figure 9), changing the media every 3 days.
CRITICAL: Matrigel will solidify at 19 C-25 C, so work quickly and keep the basement matrix cold throughout the process.
CRITICAL: Prepare cell culture plates or dishes for seeding by pre-warming them in 37 Cat least 30 min in advance, but the plates can also be pre-warmed a day in advance.

EXPECTED OUTCOMES
Microfluidic chip production Microfluidic chips should be clean and be devoid of any particles in the channels. The PDMS should be well attached to the microscopy slide and not detach upon operation. The inlets/outlets should fit with the desired tubing's outer diameter, where the tubing can be fitted with ease whilst sitting there tightly (i.e., resistant to a mild pulling force). This will prevent the pressure of the flow and any unexpected strains from detaching the tubing. The chip should support a laminar flow of liquid, which can be tested with pure oil.

Cell isolation and sorting
An expected yield for this part of the protocol is from 10 000 to 40 000 mesenchymal cells for one mouse, and 50 000 to 100 000 ductal cells. After seeding, mesenchymal cells are expected to attach within the first 3 days and show fibroblast morphology at passages 0-2, as illustrated in Figure 4.
Liver portal mesenchyme in vitro culture: By day 7-9 after seeding, the mesenchymal primary cells should be ready for passaging to passage 1. By day 14-16 they should be ready for passaging to passage 2. By 3 weeks from isolation, they should have expanded sufficiently for an encapsulation or co-culture experiment. Mesenchyme from 1-3 animals might be necessary for achieving enough cell numbers for subsequent microfluidic co-cultures. The 2D co-culture is more permissive to initial low isolation numbers. For the detailed analysis of gene expression at the time of cell isolation and during subculturing, please refer to Figures 2, Figure 4 and Figure S4 in ref. 1 .
Matrigel 2D layer co-culture: In the first 3 days of the co-culture, and especially in higher ratios of mesenchyme to ductal cells (0.5:1 or above), large cell aggregates can be observed in brightfield and fluorescence microscope ( Figure 5). The fully formed assembled 3D structures can be seen afterward, typically from day 4. At lower ratios of portal mesenchyme (below 0.5:1), formation of cystic ductal organoids should be observed, as shown in the most left panels of Figure 5 (no mesenchyme, or 0.1:1 ratio).

Microfluidic encapsulation and co-culture
At the end of microfluidic encapsulation, formation of two phases should be apparent in the test tube used for collection: an oil phase (bottom) and an aqueous phase containing the cell-laden microgels (top) (Figure 8). The liquid agarose droplets should have solidified into microgels upon collection on ice. After the removal of the majority of oil and subsequent addition of media and PFO, the phases should be visibly separated as an oil and an aqueous phase, with a mixed layer in between the two phases ( Figure 8). Microgels can be viewed on a brightfield or fluorescence microscope, if needed.
After 5-7 days in culture, the ductal organoids should be expanding and forming hollow cysts, as long as they do not have higher than 0.3:1 ratio of the mesenchyme in the structure (homeostatic ratio, as described in 1 ). While the concentration of input cells is the same for both cell types, the encapsulation in the agarose microgel and therefore formation of the co-culture ratio is stochastic,

OPEN ACCESS
occurring in any of the ratios from 0.1:1 to 10:1 (for details structure variety observed, please refer to Figure 5 and Figure S5 in ref. 1 ). If the ratio is higher, collapse of the organoid or the lack of growth is expected (Figure 9, Day 6, panel ii). The agarose microgels can be seen persisting in Matrigel co-culture and do not interfere with growth. In case of substantial organoid growth, the disintegration of agarose can be observed.

LIMITATIONS
In this protocol, we describe co-cultures of primary cells: liver ductal cells and liver portal mesenchyme. However, it is quite challenging to retrieve more than 40 000 mesenchymal cells from one mouse liver -isolation of around 5000 cells is common. It is possible to scale the isolation up, combining several mouse livers, even though this significantly extends the sorting time. Additionally, we do not recommend expanding the mesenchyme more than 2 passages in vitro, as it becomes activated (similar to the activation of fibroblasts).
Another challenge remains controlling the stoichiometry of encapsulation in the microfluidic system, which at present is probabilistic (following a Poisson distribution) and can only be minimally adjusted by increasing or decreasing input cell density.

Potential solution
It is usual to obtain low numbers of mesenchyme, typically 5000 cells, especially when digestion time has not been optimized before. To increase the number of sorted cells, altering the time of tissue digestion under the user's specific conditions is recommended. Additionally, yield can vary between different batches of collagenase and dispase used. Yields above 40 000 cells from a single liver are rare, but possible.

Problem 2
Low co-encapsulation efficiency (steps 84-96). . Co-culture of mesenchyme and ductal cells after microfluidic co-encapsulation Representative images of encapsulated cells embedded in Matrigel at day 0 and day 6 of culture. Representative of organoid structures where contact between ductal cells (red) and portal mesenchyme (green) is established (i, ii, arrows), as well as ductal cell organoid structures with mesenchyme growing in the vicinity but where contact has not been established (iii, iv, arrowheads). BF, bright field. Scale bar, 100 mm.

Potential solution
The cause of this problem is usually low cell density in the starting material, below the 0.5-1 3 10 6 cells/mL, specified in step 87d. Starting with higher number of isolated cells by increasing the number of mouse livers, or by expanding mesenchyme to achieve higher cell density is a potential solution.

Potential solution
Clogging is a very frequent problem. Dust particles that block the microfluidic chip can originate from variety of sources. The most common source of blockage are fragments from the tubing or the chip itself. To avoid this issue, the tubing and the chip should be washed with the oil prior to the experiment. If the suspected source of dust particles is the cell suspension, additional cell washes after TrypLE treatment might be necessary to remove the debris (as described in step 60 for mesenchyme and 72 for ductal cells). If the suspected fraction is agarose, we recommend purchasing a new aliquot. Finally, if the dust particles appear to be originating from oil and/or surfactant, we recommend filtering those through a 0.22 mm filter.

Problem 4
Low or no growth of cell co-cultures (steps 83 and 96).

Potential solution
The most common issue for low co-culture growth is the use of old, expired media -the media should be prepared fresh and kept no longer than one month. The overall high ratios of mesenchyme to ductal cells in the culture can also cause low expansion of the co-culture, as the mesenchyme inhibits ductal growth through direct cell-cell contact. 1 Also, if the overall numbers of mesenchyme in the culture well are low, the ductal cells do not expand well in MM media, which only contains WNT3a but otherwise is devoid of any of the growth factors normally present in the EM (expansion media), even if the ratio of mesenchyme to ductal cells is growth-permitting. Especially in the microfluidic encapsulation and subsequent co-culture, it is recommended in low-yield encapsulations to seed higher number of microgels per Matrigel dome, in order to increase the fraction of mesenchyme in the cell culture well.

Potential solution
If the Matrigel becomes wobbly or unstable in the co-cultures, it can be caused by several factors, namely: lower than usual amount of total protein in the Matrigel, seeding Matrigel on a cold plate, or using cold media to put on top of Matrigel. To avoid these issues, Matrigel of known and constant protein concentration should be used; plates should be pre-warmed at 37 C before seeding cells with Matrigel; media should be pre-warmed at least until 19 C-25 C before pipetting it on the co-culture.

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Meritxell Huch (huch@mpi-cbg.de).

Materials availability
This study did not generate new unique reagents.