A simple and flexible high-throughput method for the study of cardiomyocyte proliferation

Cardiac muscle cells lack regenerative capacity in postnatal mammals. A concerted effort has been made in the field to determine regulators of cardiomyocyte proliferation and identify therapeutic strategies to induce division, with the ultimate goal of regenerating heart tissue after a myocardial infarct. We sought to optimize a high throughput screening protocol to facilitate this effort. We developed a straight-forward high throughput screen with simple readouts to identify small molecules that modulate cardiomyocyte proliferation. We identify human induced pluripotent stem cell-derived cardiomyocytes (hiCMs) as a model system for such a screen, as a very small subset of hiCMs have the potential to proliferate. The ability of hiCMs to proliferate is density-dependent, and cell density has no effect on the outcome of proliferation: cytokinesis or binucleation. Screening a compound library revealed many regulators of proliferation and cell death. We provide a comprehensive and flexible screening procedure and cellular phenotype information for each compound. We then provide an example of steps to follow after this screen is performed, using three of the identified small molecules at various concentrations, further implicating their target kinases in cardiomyocyte proliferation. This screening platform is flexible and cost-effective, opening the field of cardiovascular cell biology to laboratories without substantial funding or specialized training, thus diversifying this scientific community.


Structured Illumination Microscopy
SIM imaging and processing ( Figure S1A) was performed on a GE Healthcare DeltaVision OMX equipped with a 60x 1.42 NA oil objective and sCMOS camera.

Fluorescence and live-cell microscopy
Some wide-field fluorescence images (Figures S1B-D) were acquired on a Nikon Eclipse Ti equipped with a Nikon 20x 0.4 NA air objective and a Nikon DS-Qi2 CMOS camera. All other images were acquired on an Incucyte (4647, Essen Biosciences, Ann Arbor, MI) at 4x, 10x, or 20x, at 37°C and 5% CO2.

Data Quantification
For nuclei count over time ( Figures 2D, 4A-D, H, 5I), nuclei were counted using the IncuCyte analysis interface. Briefly, nuclei were thresholded using rolling ball background subtraction for each time point. Data were normalized to controls, normalizing control wells to a 1 fold-change or a 0% nuclear count increase. All values shown are mean +/-SEM over three experiments. Binucleation was manually quantified ( Figures 2E, 5J, S1F) using an antibody for β-catenin and a nuclear marker (NucLight, 4717, Essen Biosciences, Ann Arbor, MI). Cells on the edge of a field of view that had a portion cut off were not included. Ki67-positive percentage was calculated as percentage of nuclei which were positive for Ki67 localization.

Statistics
P-values were calculated using One-Way ANOVAs with the exception of Figure S3F. In Figure  2D, 2E, 5J, repeated measures One-Way ANOVAs were performed. For Figure 5J, a one-way ANOVA without repeated measures was performed. A Dunnett's post-hoc test was performed if the ANOVA p-value was less than 0.05. In Figure S3F, a two-tailed paired student's t-test was performed.

DETAILED PROTOCOL
We used a relatively small library of 429 small molecules for this proof of concept manuscript. In total, we used 445 wells of a 384 well plate ( Figure S1G). This included the library with 16 accompanying DMSO controls. As such, we needed 1.5 million human iPSC-derived cardiomyocytes (hiCM) to plate at a density of ~333 hiCM/mm 2 , that is, 3333 hiCM/well; accounting for pipette variance. This density was determined by calculating the middle density between the densities at which the hiCMs proliferate the least and the most ( Figure 1D).  Immediately before the thaw, aspirate gelatin from all 445 wells in a culture hood and return to 37°C.
 Remove a vial of cells from the vapor phase of liquid nitrogen storage using a large pair of forceps and warm the tube in a 37°C water bath for 3 minutes. Take care to maintain sterility by not submerging the cap of the tube, and hold the tube stationary-NO SWIRLING. The large pair of forceps may be used to hold to tube in the water bath, or a floating tube rack can also be used as long as the cap remains above water.
Note: Follow manufacturer's instructions and/or a differentiation protocol 68 to obtain 1.5 million cardiomyocytes. We used a variation on Cellular Dynamics plating protocol for a 500 µL tube containing approximately 3 million cells. We thawed the hiCMs into wells of a 384-well plate. This plate was used directly for the experiments as we noticed that re-plating hiCMs caused loss of hiCMs due to cell death. We allowed at least one row of empty wells on the sides of the plate. This minimizes contamination and evaporation, which is an important concern when media is not being changed daily.
Note: There are two main companies that provide quality differentiated hiCMs that we have used. We chose Cellular Dynamics because that is what we had frozen in our liquid nitrogen storage at the start of this project. Cellular Dynamics hiCMs are frozen at 30 days of differentiation. We have also used hiCMs from Ncardia to study sarcomere assembly and found them to be similar to those from Cellular Dynamics 70 . However, we have not hiCM for cell from Ncardia for proliferation studies as of yet.
 Remove the cryovial from the water bath, spray it with 70% ethanol, and place it into the culture hood.
 In the cell culture hood, use a 1 mL pipette to slowly remove cells from the cryovial over two seconds and them to a sterile 50 mL conical centrifuge tube (89039-656, VWR International, Suwanee, GA or 82050-348, Greiner, Kremsmünster, Austria) by slowly expelling them over 4 seconds.
 Gently rinse empty cryovial with 1 mL plating medium to recover remaining cells. To do so, tilt the pipette tip so that the media runs down the inner sides of the tube, rotating the pipette tip to rinse the circumference of the tube. Transfer the 1 mL of plating medium with the recovered cells from the cryovial to the 50 mL conical tube drop-wise over 90 seconds while gently swirling the conical tube. Slow addition of medium in this step is critical to minimizing osmotic shock: add approximately one drop per ~3 seconds.
 Slowly add 4.5 mL of plating medium to the 50 mL conical tube. Add each remaining 1 mL of medium drop-wise over 30 seconds per mL. Gently swirl the tube while adding medium.
 Close the cap on the 50 mL conical tube and invert gently 2 times. Do not shake or vortex the cell suspension.
Note: To determine how much plating medium to add, see cell viability calculation of the batch of hiCM purchased. Each 50,000 cells will require a total of 100 µL of medium. For example, a batch with 3 million viable cells will produce 60 x 50,000 cells, so 60 x 100 µL = 6 mL is required in total. After the 1.5 mL (500 µL cells + 1 mL plating medium used to wash cryovial), 4.5 mL remains.
Note: It is important to optimize which plate to use based on the type of image information your experiment requires (e.g., fluorescence and/or phase contrast). For example, Greiner 384-well plates (781182) a have larger font size on the plates, allowing for easy addition of small molecules, and work well for fluorescent imaging as they have no detectable auto-fluorescence ( Figure 2C). However, these plates produce an optical aberration that creates linear patterns, and often have small scratches ( Figure 2C). On the other hand, some plates (e.g., Thermo Fisher 384-well plates-164688) do not produce such patterns and have less. However, these plates are labeled with a smaller font that is difficult to read from afar when the plates are in a laminar flow hood behind glass. 96-well plates or even fewer/bigger wells may be used as well. This increases sample size at the expense of cost and overall throughput.
 Transfer 3 mL of the cell suspension to a new sterile 50 mL conical tube. The remainder of the cell suspension may be used for other purposes such as second screen or other experiments.
 Add 19.5 mL of plating medium to the 3 mL of cell suspension, cap the tube, and invert gently 2 times.
 Pour the entire contents of tube (~22.5 mL) into a sterile reagent reservoir.
 Retrieve the two 384-well plates from incubator and transfer to the culture hood.
 Use the multichannel pipette set to 50 µL to add cells from the reservoir into the wells of the 384-well plates, using the plate map of choice.
 Put the plates into the incubator for at least 5 hours (Note: we have gone up to 7 hours). This allows time for the cells to recover and adhere to the plate so that they are not washed away during the media change.
 During the incubation in Step 15, warm 22.5 mL of maintenance medium to 37°C in a bead bath, water both or incubator.
 5-7 hours post-thaw, retrieve the plates from incubator and put them in the cell culture hood.
 Pour the warm maintenance medium into new sterile reagent reservoir in the cell culture hood.
 With two multichannel pipettes set to 50 µL each, simultaneously remove 50 µL of plating medium from the wells with one hand and add 50 µL of maintenance media from reservoir to the cells with the other hand.
 Put the cells back in the incubator for 2 days.

Maintaining cardiomyocytes: Day 3
Note: Be careful not to touch the bottom of the wells while changing media, and minimize time the cells are without media. These cells are fragile, so all media changes in this protocol are done by hand rather than with a liquid handling system. A makeshift waste container for large quantities of pipette tips can be made out of a tip box lid. If two multichannel pipettes are not available, add new media immediately after removing old media, row-by-row. 24. Pour ~1/2 of the contents of the conical tube containing media into a sterile reagent reservoir in hood, then close the conical and return it to the bead bath or equivalent to maintain temperature and sterility. (Small molecule addition can take several minutes depending on the researcher's pipetting speed, and medium that cools down will be dangerous to the cells) 25. Remove cell plate 1 from the incubator and remove seal from small molecule plate 1 in the hood. 26. Pipette the small molecules in this order (see troubleshooting for addition details on the handling procedure): Note: For our purposes, we obtained two 384-well plates with plate maps as in Figure  S1G from the Vanderbilt University High Throughput Screening Core, with 50 nL of each small molecule from the SelleckChem Kinase Inhibitor Library per well at 10 mM. The plates were sealed and stored for ~6 hours in a desiccator until use. In this way, it is simple to add media to each well of the 'small molecule plate', then transfer media from the small molecule plate to the cell plate, 12 wells at a time using a multichannel pipette. We recommend having the small molecules of interest in a separate plate matching the planned map of the cell plate in order to simplify small molecule addition and minimize contamination. See plate maps in Figure S1G or troubleshooting for more information Note: If transducing a stably-expressing nuclear marker (or transducing or transfecting any other construct), it is possible to do so today. For example, if using Cell Light Nucleus (C10602, Invitrogen, Carlsbad, CA).
a. Add 50 µL of media to the small molecule plate using a multichannel pipettor and gently mix by pipetting up and down (this dilutes the small molecules to 10 µM).
b. Remove 50 µL of media from the corresponding wells of the cell plate using a second multichannel pipette.
c. Transfer 50 µL media from the small molecule plate to the corresponding wells of the cell plate using first multichannel pipette. Discard both sets of pipette tips.

Repeat
Step 26 for plate 2, adding warm media to the reagent reservoir when needed. Take care to avoid cross-contamination by replacing tips at every step. For 445 wells, this protocol will use ~9 pipette tip boxes per drug addition day.
28. Place the cell plates on the microscope of choice and start imaging.

Small molecule addition (Day 8: Repeat Steps 22-28)
29. Warm 22.5 mL of the maintenance medium to 37°C in a bead bath or equivalent.
30. Obtain the new plate of small molecules as previously. 31. Similarly add media to the small molecule plate, remove media from the cell plate, and transfer media from the small molecule plate to the cell plate as previously (Steps 22-28).
32. Place cell plates back on the microscope.

Quantify (Day 11)
Note: We used an automatic microscope that can image up to 6 plates at a time. However, any microscope with an automatic stage and multipoint capabilities could be used. This would limit the number of samples imaged at a time as most microscopes can only image one plate at a time. Alternatively, if live-cell imaging is not available, fixed time points at various hours or days post-small molecule addition and imaging of nuclei with a marker such as DAPI will also facilitate nuclei count.
33. Remove cells from automatic imager 6 days post-small molecule addition 1.
34. Quantitative assessment of cardiomyocyte proliferation. We used the Incucyte automatic analysis interface to do so, but thresholding nuclei in FIJI/ImageJ will also provide a mask to quantify nuclei count over time.  Step 1.

Fix and stain cells (Optional
Note: We recognize this uses a large quantity of antibody, which can be costly. To circumvent this issue, we recommend fixing all wells and keeping the plates in a 4°C fridge until the wells are quantified and any hits are identified. Then, only the relevant wells can be stained to conserve antibody. 15 µL is also the minimum requirement for these steps: most 384-well plate wells can hold up to 100 µL of liquid.  This file shows the raw data for the calculations in Figure 3. The first sheet shows the data for Figure 3A, that is, all of the data for each screen sorted by the average (Columns A-F). The first sheet also shows all wells with a nuclear fold change greater than one sorted by the standard error of the mean (Columns I-N). The second through fifth sheets show the raw nuclear counts for each well in the first, second, and third screens, respectively. Note that the first screen had a different plate map than the rest, and that the conversion between screens can be found in the first sheet of "AllScreenNotes.xlsx". Screens 3 and 4 used a different nuclear marker than screens 1 and 2, so the nuclear fold changes are not comparable until normalized to control (in yellow highlight on each sheet). Screen 1 and Screen 2 had averaged controls.
"DetailedDrugInfo.xlsx" This file, modified from the small molecule file from Selleck Chem, shows detailed information for each of the small molecules in the library. The first column (A) shows the well of the plate that the small molecule was in. The second column (B) shows the small molecule's unique identifier which can also be found in "AllScreenNotes.xlsx" in order to easily compare files. Column C shows the catalog number of the small molecule from Selleck Chem, and Column D shows the target of the small molecule. Columns E-I shows detailed information directly from Selleck Chem on the small molecule, including any known aliases of the compound and other targets.