Protocol for Tri-culture of hiPSC-Derived Neurons, Astrocytes, and Microglia

Summary This protocol establishes a tri-culture of hiPSC-derived neurons, astrocytes, and microglia for the study of cellular interactions during homeostasis, injury, and disease. This system allows for mechanistic studies that can identify the roles of individual cell types in disease and injury response in a physiologically relevant, all-human system. This protocol utilizes and modifies prior differentiations. Limitations include the prolonged maturation of human astrocytes and neurons and scalability. For complete details on the use and execution of this protocol, please refer to Ryan et al. (2020).

8. Replace stem cell media with standard Stem MACS iPS-Brew. 9. In each of the 6 wells, add 4 mL of the 1 mg/mL stock of polybrene to the 2 mL of Stem MACS media. 10. In 5 of the 6 wells, leaving one as a polybrene-only control add the following amounts of EACH virus: 0.25, 0.5, 1, 2, and 4 mL. 11. Dispose of all excess virus and all pipettes that have touched the media and viruses in 10% bleach or inactivate lentiviral particles according to local regulations. 12. Wait 6 h with cells in 37 C 5% CO 2 incubator, then replace media with fresh Stem MACS iPS-Brew XF media. a. All media and pipette tips must be disposed of in 10% bleach. 13. Evaluate cell survival and perform full media exchanges on the cells after 24 and 48 h. The well exposed to the highest concentration of virus that survived the infection can be taken and expanded. 14. Expand and freeze down at least 50 vials per line (2 million cells per vial) as a master stock.
a. Once cells reach $80% confluency remove Stem MACS and wash 13 with 20 C PBS. b. Aspirate PBS from hiPSCs and apply warm Accutase at half the normal amount of media in the plate (i.e., 1 mL per well of a 6 well plate). c. Incubate in 37 C 5% CO 2 incubator until cells begin lifting from plate (roughly 5 min). d. Apply a volume of 37 C DMEM:F-12 equal to that of the Accutase to inactivate the Accutase. e. Lift cells from the plate by pipetting up and down slowly with a serological pipette. f.
Place cells in an appropriately sized conical tube and spin for 5 min at 162 3 g 20 C. g. Aspirate supernatant. h. Resuspend in Stem MACS by slowly adding media to the tube and then gently pipetting up and down with a 10 mL pipette until the pellet has uniformly gone into suspension. i. Re-plate cells at 1:12. j. Expand to at least 50 wells of 6 well plates. k. Freeze down at least 50 wells by lifting and spinning hiPSCs as just described. l. Count cells by hemocytometer or vicell or equivalent method. m. Resuspend in Freezing media (10% DMSO in Stem MACS) at 1 mL per 2 3 10 6 cells. n. Freeze down 1 mL per vial and store in a Mr. Frosty or equivalent storage container at À80 C freezer for at least 24 h before moving to LN2 for long term storage.
Note: Freezing down 50+ vials will allow sufficient vials to perform a study all from the same batch of hiPSCs. This will prevent issues arising from random mutations that may be introduced during passage as well as contamination that may occur.
17. Replace media the following day without Y27632. a. Continue to feed daily until cells reach $80% confluency (1 to 2 days).
Note: A single passage post thaw is sufficient. a. Remove Stem MACS and wash 13 with 20 C PBS. b. Aspirate PBS from hiPSCs and apply warm Accutase (pre-warmed in 37 C water bath for 15 min) at half the normal amount of media in the plate (i.e., 1 mL per well of a 6 well plate). c. Incubate in 37 C 5% CO 2 incubator until cells begin lifting from plate (roughly 5 min). d. Apply a volume of 37 C DMEM:F-12 equal to that of the Accutase used. e. Wash cells from bottom of plate by pipetting up and down slowly three times with a 5 mL serological pipette. f. Place cells in an appropriately sized conical tube and spin for 5 min at 162 3 g 20 C. g. Resuspend in Stem MACS by slowly adding media to the tube and then gently pipetting up and down with a 10 mL pipette until the pellet has uniformly gone into suspension. Plate at 1. Note: Line-to-line variability can account for great differences in yield. However, a 10 cm plate should yield between 6 million and 12 million NPC-like cells. a. Re-plate cells to a 1:20 Matrigel: DMEM/F12 coated plate at 1 3 10 5 cells/cm 2 . i. Note: seeding density will need to be empirically determined for each line / transduction. However, most cell lines operate well at this density. ii. Note: a 1:20 Matrigel coating increased adherence of neurons through differentiation.
Note: We have found that a small percentage of cells that are not transduced survive puromycin selection. If unchecked, they can take over the plate as they are mitotic. In order to remove these cells, Ara-c is added, which will kill dividing cells. This step will produce a stock of fully differentiated astrocytes that are ready to be added to the triculture.
Note: The FBS, EGF, and FGF will push the NPC-like cells toward an astrocyte fate.
3. Let grow with half media exchanges every 3 days and re-plate via the same method as for hiPSCs 1:2 when it reaches $90% confluency to 1:35 Matrigel-coated plates.
CRITICAL: High density is extremely important throughout differentiation. If the cells are too sparse, they will stop growing completely. This is especially true in the first two stages (high FGF media and low FGF media, up through day 70 of differentiation). Note: Cells will have a pyramidal shape with sparse, short processes ( Figure 2B).

(Day 30)
Exchange half media to low FGF Astrocyte media for second stage maintenance. 5. Continue to feed, let grow, and expand as before now in low FGF Astrocyte media for second stage maintenance up to day 69.
Pause Point: You can freeze down at this stage in low FGF astrocyte media + 10% DMSO.
Note: Cells will have similar pyramidal shape as in the high FGF stage. However, in more sparse cultures, the cells can flatten out ( Figure 2C).
7. Continue to perform half media exchanges in ScienCell astrocyte media, let grow, and expand to day 90. a. At this point, when you re-plate, you do not need to coat with Matrigel.
Pause Point: You can freeze down at this stage in ScienCell astrocyte media + 10% DMSO at 1 million cells per vial. b. Cells are ready after day 90 for tri-culture.
Note: Cells will have abundant processes and take on a more star-shaped astrocyte phenotype ( Figure 2D). a. Greater than 80% of the cells should be positive for nestin, glutamine synthetase and SOX9 and at least 50% of the cells should express thrombospondin-1. b. iAstrocytes can also be tested for calcium wave propagation and glutamate uptake.
Note: glutamate uptake can be observed at this stage, but the iAstrocytes are still immature.
Note: The high concentration of Matrigel is necessary for the health of the mesoendothelial cells that are produced by the differentiation, which produce the non-adherent common myeloid progenitors. a. Remove Stem MACS and wash 13 with 20 C PBS. b. Aspirate media from stem cells and apply warm dispase (incubated in 37 C water bath for 15 min) at half the normal amount of media in the plate (i.e., 1 mL per well of a 6 well plate) Note: Dispase is used in this passaging stage to keep the cells in clumps. The Human Pluripotent Stem Cell Core at CHOP, who developed this method, found this to be effective for performing the differentiation. c. Incubate in 37 C 5% CO 2 incubator until cells begin lifting from plate (roughly 5 min). d. Apply a volume of warm DMEM:F-12 equal to that of the dispase used. e. Wash cell clumps from bottom of plate by pipetting up and down slowly with a 5 mL serological pipette 2-3 times, making sure not to break the clumps. f. Place cells in an appropriately sized conical tube and spin for 5 min at 162 3 g, 20 C. g. Resuspend in Stem MACS by slowly adding media to the tube and then gently pipetting up and down with a 10 mL pipette until the pellet has uniformly gone into suspension. Take a p1000 and slowly break up the colonies 3-5 times so that they are a uniform suspension of small clumps. h. Re-plate with 10 mM Y27632 at 1:15 ratio.
CRITICAL: From days 7 to 9, single cells shed off the adherent layer into the medium and are collected. These are the common myeloid progenitors (CMPS). Collect supernatant by pulling all the media from the wells and placing into a 50 mL tube. Centrifuge at 162 3 g for 5 min at 20 C. Aspirate the supernatant. The cell pellet is often not visible, so leave roughly 500 mL of the supernatant to avoid aspirating the CMPs. Resuspend in 2 mL of SFD with small molecules. Feed the wells with the fresh media while spinning, and then add back the resuspended used media equally across the wells.
Note: There is line-to-line variability on the number of CMPs produced. An efficient differentiation will yield 2 3 10 5 to 5 3 10 5 per well.
Note: Cultures are terminated after day 9 as non-CMPs begin to be produced and reduce the purity.

CMPs are frozen at 1-3 million cells per vial in 90% FBS and 10% DMSO.
Pause Point: Frozen CMPs can be stored in LN2 until ready to start tri-culture.
Assemble the Tri-culture

Timing: 30 days
This is a 30-day protocol that combines three differentiations. It is expected that the iPSC line for the iNeurons has already been transduced with the NGN2 virus, have had stocks frozen down, and have been tested for differentiation yield. It is also expected that the iAstrocyte differentiation is completed to day 90 and are frozen down. Lastly, it is expected that CMPs have been generated before the 30-day protocol begins. This protocol is designed to end on day 21 of the iNeuron differentiation, with the iAstrocytes added on day 5 and iMicroglia added on day 7 of the iNeuron differentiation ( Figure 4A). This can be amended for extended cultures if desired. The iMicroglia differentiation in this section was based on previously published work (Abud et al., 2017). Time commitments per step are dependent on scale, number of lines, and number of conditions in the experiment. It is highly recommended you write out the timepoints of the protocol on a calendar before you begin. b. Feed and expand in ScienCell media up to day 14 of the differentiation. Initial amount that is thawed is experiment dependent based on how many tri-culture wells are needed. 21. (Day 4) Thaw a vial of 2 3 10 6 transduced hiPSCs in preparation for the iNeuron differentiation onto 3 wells of a 1:50 Matrigel-coated 6 well plate. a. Expand the hiPSCs as previously described to 2 1:50 Matrigel-coated 10-cm dishes. b. The goal is to have the hiPSCs at $60% confluency on 2 10-cm dishes at day 9 when the iNeuron differentiation will begin.
CRITICAL: This generally is enough time to have a single vial plated onto 3 wells of a 6 well plate, then expanded to 2 10-cm dishes reach $60% confluency. However, this needs to be

ll
OPEN ACCESS empirically tested, as described above, as each line will have different growth rates and different transduction efficiencies. In addition, this will need to be scaled based on the needs of the experiment.
CRITICAL: Do not use a larger sized well than a 24-well even at the same seeding density. iMicroglia yield will exponentially drop. Smaller wells do work.
Note: iMicroglia can also be plated onto 24 wells plates/glass or smaller with 1:35 Matrigel coating, but it will have a slightly lower yield than the CellBIND counterpart. a. Thaw CMPs in the same manner as hiPSCs. b. Thaw in a 37 C water bath until there is a pea-sized amount of ice left. c. Slowly add thawed CMP solution to 9 mL of warmed RPMI. d. Spin at 162 3 g for 5 min at 20 C. e. Aspirate supernatant and resuspend pellet gently with a 5 mL pipette in iMg media (3.3 3 10 5 cells/mL). Slowly pipette up and down until pellet has fully gone into suspension. f. Plate at 1 mL/well on CellBIND 24-well plate. g. Let cells rest for 48 h (days 0 and 1). h. Half media exchanges will be performed on days 2 and 5, and a full exchange on day 8 of iMicroglia differentiation.

(Day 7) Perform a half media exchange on iMicroglia.
CRITICAL: Cytokines in the iMg media (IL-34, M-CSF, and TGF-b1) are unstable, so they must be added fresh for each media exchange (day 2 of iMicroglia differentiation).
Note: Microglia will be highly ramified by the end of the 11-day differentiation ( Figure 4B).
Note: 8-well chamber slides will have 500 mL of media.
Note: Tri-culture will be combined onto the iNeurons in the chamber slide. Tri-cultures can be combined onto different plate formats, but they have not been rigorously tested by this group. CRITICAL: Resuspend iAstrocytes in ScienCell astrocyte media to 5 3 10 4 cells/100 mL. Survivability in the tri-culture increases if they are re-seeded in a mixed media initially.
Note: The iAstrocytes are added as early as possible to help support and mature the iNeurons, but it must be added after Ara-c treatment is removed, otherwise they will die.
CRITICAL: iMicroglia yield (total number of cells after lifting) can dramatically decrease with larger tubes. Use as small a tube (preferentially 2 mL Eppendorf tubes) as possible to increase yield.
CRITICAL: Resuspend iMicroglia in iMg media to 1 3 10 5 cells/100 mL. Survivability in the tri-culture increases if they are re-seeded in a mixed media initially.
Note: Yield for iMicroglia from CMPs is usually around 50%-60% of total number of CMPs seeded but varies across lines.
Note: iMicroglia can be differentiated beforehand and frozen. However, the yield will be reduced due to cell death during freeze/thaw process. Note: Tri-culture can be extended beyond this timepoint. This group has extended iNrn differentiations to upwards of day 80. However, these were grown on rat astrocytes and were switched over to BrainPhys media (STEMCELL Technologies #05793) starting day 21 of iNeuron differentiation.
Roughly 1 million CMPs, astrocytes, and neurons will produce 5 to 6 tri-cultures in the 8-well chamber slide system due to the 50%-60% yield of microglia from CMPs. By scRNA-seq data (Ryan et al., 2020), with the described seeding densities, the tri-cultures will have a cell ratio of roughly 3.7:1:1 neurons to astrocytes to microglia. These tri-cultures are quiescent and can produce an inflammatory response if exposed to a stimulus such as viral infection. In addition, the cells in the culture exhibit physiological functions including synaptic phagocytosis and cytokine production by the iMicroglia, synapses and action potentials by the iNeurons, and GLT-1 expression and neuronal support by the iAstrocytes (Ryan et al., 2020). Tri-cultures can be extended beyond the 30 days described in this protocol.

LIMITATIONS
Due to the relatively low yields of CMPs and inability to expand the iMicroglia, scaling the tri-culture up may be difficult or at the very least require large investments in time and reagents. Depending on the line, some hiPSC lines can produce less homogenous populations of CD41+CD235+ CMPs. One may need to FACS sort to ensure the non-adherent cells are CMPs. In addition, changes in confluency or Matrigel concentration can reduce yields of CMPs. Tri-cultures are extremely delicate and excessive shaking, moving, or too forceful media exchanges can cause the neurons to lift from the plate.

TROUBLESHOOTING Problem
Transduced iPSCs have not grown to proper confluence by day 4 of the tri-culture differentiation.

POTENTIAL SOLUTION
If the timing is off for growing the NGN2 rTTA transduced cells, it is acceptable to allow the iMicroglia or iAstrocytes to stay in mono-culture, even past day 11 for the iMicroglia. iMicroglia were maintained in mono-culture up to 15 days past D11 with no signs of increased death or cell stress. However, we have not tested the efficacy of the tri-culture if the iAstrocytes or iMicroglia are added later than described. There may be a decrease in maturity of the iNeurons without the iAstrocytes.

PROBLEM
Astrocyte differentiation fails.

POTENTIAL SOLUTION
The iAstrocyte differentiation is very sensitive to density. Too low of a density and the iAstrocytes will flatten out and stop dividing. It is imperative to not split more than 1:2 throughout the differentiation. After day 90, they can be split upwards of 1:5, but this can vary from line to line.

PROBLEM
iMicroglia are lost during collection.

POTENTIAL SOLUTION
When the iMicroglia are lifted from their CellBIND plates, it is imperative to use the smallest tube possible to centrifuge the cells. This may require using less PBS or wash media to quench the Accutase. It is also necessary to use a bucket rotor and not one at a fixed angle, as in the fixed angle centrifuges, the cells will not properly pellet at the bottom of the tube, leading to loss of iMicroglia. CMPs plated to 24 well or smaller work very effectively. Larger plate sizes (i.e., 12 well and 6 well) have a massive drop in iMicroglia yield even accounting for seeding density for the larger surface area. This creates an issue when lifting the cells for the tri-culture as it requires more wells, more Accutase, and more time.

PROBLEM
Low transduction rates for hiPSC lines with NGN2 and rTTA lentiviruses POTENTIAL SOLUTION Some lines have low transduction rates for the NGN2 and rTTA lentiviruses. We have found that using Lentiblast (Oz Biosciences LB00500) has increased efficiency.

POTENTIAL SOLUTION
There will be some line variability for survival during re-plating for each of the cell types. This was most apparent with iNeuron and iAstrocyte differentiations. The growth and survival rates need to be determined empirically. However, the described seeding densities and timelines have worked for at least three lines for each of the three differentiations.

PROBLEM
CMPs do not attach to plate after 1 day.

POTENTIAL SOLUTION
The CMPs attach poorly in wells with large surface areas (larger than a 24 well). Collect the media from the well and re-plate onto a smaller, Matrigel-coated plate or cellBIND plate. Let rest for 48 h and then proceed from the day 2 step with a half media exchange.

POTENTIAL SOLUTION
The iNeurons become more fragile and lift more easily form the plate as they differentiate and grow. If the iNeurons lift, unfortunately they are not salvageable. Be gentle when exchanging media, travel as small a distance with the plates as possible, and do not fully remove media when fixing and performing immunostaining. To compensate for the remaining liquid during washes during fixation or immunostaining, we recommend at least 10-min washes.

PROBLEM
Phenotypes are not present in disease models using this tri-culture.

POTENTIAL SOLUTION
The iAstrocytes and iNeurons have relatively immature phenotypes (Ryan et al., 2020). Extending the cultures or utilizing other differentiations for the individual cell types may result in a more mature phenotype (Nehme et al., 2018;Tchieu et al., 2019;Wang et al., 2013) and thus a more physiologically relevant system depending on what question is being asked.

Lead Contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Stewart Anderson (sande@pennmedicine.upenn.edu).