Scalable 96-well Plate Based iPSC Culture and Production Using a Robotic Liquid Handling System

Development of a plate based robotic stem cell culture platform.

The demand for iPSCs is growing due to their utility in drug development and regenerative medicine. Yet scalable production has focused on suspension culture³² in relatively complex equipment that excludes many researchers and diverges from established adherent culture methods. Rather than drastically alter proven plate based stem cell culture methods, such as switching to a suspension culture or using a microcarrier, we focused on miniaturizing and automating our existing adherent iPSC culture techniques. The first aim was to automate feeding and passaging to normalize the entire culture process. A platform capable of rapid scale up with existing technology was also required. To achieve these goals a method was devised to culture stem cells in a 96-well plate utilizing an automated liquid handling system (Figure 1). The protocols developed here with the liquid handling robot for feeding and passage do not require a centrifugation step or continuous monitoring by a technician. These protocols were developed with two feeder free iPSC lines; one derived from fibroblasts and the other from adipose cells³³.

The computer controlled liquid handling system (Figure 1A) consists of a flatbed that moves front and back. The bed has nine, approximately 5 x 3.3 inch numbered recesses with pressure retaining clips to accept standard cell culture plates, liquid reservoirs and other custom hardware. The pipet head moves left to right (perpendicular to bed movement) as well as up and down. Together with the bed, the pipette head can be programmed to remove or deliver media anywhere on the bed after picking up pipette tips from a refillable rack. If necessary, each of the 96-wells can be kept separate to eliminate cross contamination. In the present configuration, 0 to 200 µl of media can be moved by each tip of the pipet head, but other volume ranges are possible based on tip size.

When passaging stem cells in standard plates, enzymatic or chemical dissociation typically requires incubation at 37 °C. Initial testing confirmed dissociation with a proteolytic and collagenolytic dissociation reagent or an EDTA based reagent performed better at 37 °C than RT both for time to dissociation and homogeneity of colony size produced for our two 96-well iPSC lines (data not shown). To automate the split process and avoid moving plates in and out of an incubator, sloped, temperature controlled ramps that accept and reproducibly position standard culture plates were built (Figure 1A). When the ramps were heated to 37 °C, a proteolytic and collagenolytic dissociation reagent or EDTA based dissociation of the iPSCs from 96-well plates was comparable to plates placed in a 37 °C incubator (data not shown). The sloped ramps were also utilized to enable pipet tips to collect an entire well’s contents (less than 5 uL residual volume) by reaching the lowest point in the well whereas aspiration from a flat plate left a residual volume of approximately 10-15 µl, resulting in cell loss. The sloped ramp also allowed the wells to be washed (triturated) by ejecting media at the top of the sloped well and to collect it at the bottom.

Robotic culture of stem cells in the 96-well plate format; feeding and passaging.

Culture of iPSCs using the robotic liquid handling system has two phases; passage and feeding. When a colony is ready to passage, it is split at a predetermined ratio to seed a new plate which is then fed at regular intervals until it is ready to passage again (Figure 1B). Frozen or non-96-well cultures can be started in the 96-well format and immediately enter the passage/feed cycle. Cells that are not used to maintain the colony, which is the majority of a plate, are harvested for other uses and represent the production component of this system.

Feeding 96-well cultured iPSCs is accomplished when some or all of the media is removed after a period of culture and deposited in a waste reservoir. Then fresh media is aliquoted to the same plate. The robot can use new tips and media troughs to segregate every well for fully customizable feeding, including blending conditioned media with new media.

Typical passage of stem cells requires a method of dissociation and often a centrifugation step. The protocols described here aims to normalize dissociation and eliminate centrifugation. The passage protocol begins when the robot delivers dissociation reagent to 96-well plates mounted on 37 °C sloped ramps. After a preset time, the dissociated cells are collected with a washing action where the user has control over the position, repetition, volume, and trituration rate. Enzyme time, trituration rate and the number of trituration repetitions were sufficient to vary dissociated colony size and homogeneity (Figure 2), but each parameter is adjustable to fit a given cell line’s requirements. This control removes variability introduced by technicians who pipet with varying rates, positions, and repetitions. Once the dissociated cells are collected and pooled in a reservoir with fresh media, they are distributed to a new plate. To avoid centrifugation and seeding problems from substrate degradation or cell death due to enzyme action, dissociation reagent was diluted to the minimum point of acceptable dissociation that resulted in reliable seeding. This technique was effective for a proteolytic and collagenolytic dissociation reagent in mTeSR1 medium³⁴, which has a relatively high protein content, but also effective in low protein media like Essential-8 ³⁵, both of which are compatible with this system (E8 data not shown).

Controlling the seeding density of stem cells after passage is paramount to routine and successful stem cell culture. Seeding too low or high can result in ectopic differentiation and loss of pluripotency in addition to causing irregular passage intervals. Typical stem cell culture relies on ratio splitting where one well of a 6-well plate is used to seed an entire new 6-well plate (a 1:6 split). With the liquid handling robot this ratio can be adjusted to suit the needs of the user (e.g., 1:6, 1:9, 1:12, etc.) and it has the most influence on the passage interval. The robot can harvest less than one column, one entire column (8 wells), or more than one column depending on the user’s needs, which should be determined empirically. The adipose and fibroblast derived iPSCs maintained a regular three day feeding schedule with passage on the fourth day when split 1:12. In this case one column was harvested and used to seed one new 96-well plate, creating a 1:12 split with each cycle (Figure 3A). The remaining 11 wells were then available for downstream applications. To harvest these 11 production wells, the passage protocol was repeated except the cells were deposited into a trough for collection. Alternatively, the cells can be left on the plate and used directly.

To achieve consistent seeding density passage to passage without counting, and maintain a routine splitting schedule, our protocols call for splitting the same predetermined number of columns when the colony is at an ideal density for passage. To help users identify the proper density for passage, a series of images was generated showing a range of densities with an ideal passage density highlighted (Figure 3B). To determine when to passage, a technician examines their plate and compares it to the density image scale. The ideal density to passage was determined from culture of the adipose and fibroblast derived iPSCs and found to be when the space between most colonies was about 25% of adjacent colony diameter. It is important that the colonies be homogenously distributed. This amount of colony separation correlates with the maximum desirable density because one additional feeding cycle would result in colonies touching, which is a trigger for ectopic differentiation.

A necessary protocol developed here addresses how to start a culture in the 96-well format and how to reset a culture after a density problem. When a new culture is started, the best seeding density and number of feeding cycles before passaging are unknown. To ensure a usable density after seeding, the robot distributes existing or thawed cells across a 96-well plate in a serial dilution (Figure 4A). Example results of such a gradient are shown in Figure 4B-E. After several feeding cycles, some columns approach the ideal density to split, at which point a technician uses the robot to seed a uniformly diluted plate to start regular culture. This density gradient protocol was also used to restore regular passage intervals and colony homogeneity to an irregular plate. If passage is delayed or missed, colony density and size become too high, whereas if passage is too early, colony density is too low and individual colonies may become too large without being evenly distributed across the well. The density gradient protocol resolves these issues and allows the user to restart by picking an ideally seeded set of wells.

Two iPSC lines remained pluripotent after 3 months of robotic culture.

Maintenance of stem cell pluripotency can be adversely affected by culture conditions^36,37. To evaluate whether adipose and fibroblast derived iPSC lines (a-iPSC, f-iPSC, respectively) maintained pluripotency when cultured robotically in 96-well plates for an extended period of time, the two lines were cultured on the liquid handling robot as described above for 3 months and then pluripotency markers were examined. For this period both lines were passaged with a proteolytic and collagenolytic dissociation reagent greater than 20 times without centrifugation. Fixed 96-well plates from a-iPSC and f-iPSC lines stained for Oct4 and Nanog exhibited nuclear accumulation of these markers while Ssea-4 was observed on the cell surface (Figure 5A-L). This is consistent with previous reports for pluripotent iPSCs^38–41. Counterstaining with DAPI did not reveal any additional cells that were not also positive for the three pluripotency markers. Chromosomal abnormalities are commonly observed during stem cell culture^42,43 but may not affect the distribution of pluripotency markers like those shown in Figure 5A-L. Karyotype analysis revealed both robotically passaged iPSC lines had 46 normal chromosomes, suggesting the robotic culture method did not introduce chromosomal instability beyond what might normally occur (Figure 5M-N).

To probe established stem cell gene expression markers^1,41,44,45 in the robotically cultured iPSC lines, total RNA was collected in parallel to staining and karyotype analysis. Both 96-well plate iPSC lines had similar expression of pluripotency markers NANOG, POU5F1 and REX1 whereas cardiomyocytes differentiated from each line did not, nor did a separate line of human dermal fibroblasts (HDFs) (Figure 6). The cardiomyocytes derived from both lines were MYL7 positive whereas the HDFs and stem cells did not express MYL7. Both iPSC lines were differentiated into cardiomyocytes with the small molecule, Wnt pathway manipulation technique described recently⁴⁶. When stained, cardiomyocytes from both lines were cTnT positive with clear sarcomere formation (Figure 7). Differentiation into cardiomyocytes in the 96-well format was successful in greater than 80% of the wells (data not shown). Together, these data suggest 3 months and more than 20 passages of robotic culture resulted in chromosomally normal cells exhibiting a transcription program consistent with pluripotency that were also capable of differentiation into cardiomyocytes.

Figure 1. Overview of robotic stem cell culture equipment and typical iPSC culture routine. (A) Robotic liquid handling system shown with typical bed layout for 96-well stem cell culture. New sterile pipet tips (Tips), removable tip waste container (Waste), multi-well autoclavable trough to hold required liquid reagents (Liquids), 96-well plates are positioned and supported on a temperature controlled, sloped ramp (96-well plates), 8 channel robotic pipet head that moves left/right and up/down (Pipet head). Plate lid holding rack (PLHR). Bed position numbers shown below in table. (B) Robotic iPSC culture has two phases: Feeding and Passage. The cycle of cell production starts when a colony plate is ready to passage. The user chooses the desired ratio to passage at and the robot seeds a new group of 96-well plates then fed at regular intervals until ready for passage again. When the plates are ready to passage, most of each plate is not used to seed subsequent colony plates and is available for other uses, thus representing the production phase of the system. Frozen or existing cell lines can be introduced at any time and maintained in the feed/passage cycle. Please click here to view a larger version of this figure.

Figure 2. Robotic liquid handling parameters can be used to control iPSC colony dissociation characteristics. Each representative image shows fibroblast derived iPSCs dissociated after the indicated treatment while still suspended in the 96-well. (Top Row, A–C) Enzyme Time, no trituration. Fibroblast derived iPSCs grown in 96-well plates were subjected to increasing times of proteolytic and collagenolytic dissociation reagent dissociation and no robotic trituration was performed. (Middle Row, D–F) Trituration rate, 3 minutes enzyme. Cells were exposed to three minutes of proteolytic and collagenolytic dissociation reagent treatment and then 175 µl growth media was applied and each well pipetted up and down once at the indicated rate. µl/sec = microliters per second. (Bottom Row, G–I) Trituration Repetitions, 3 minutes enzyme, 160 µl/sec. Cells were exposed to proteolytic and collagenolytic dissociation reagent for 3 minutes, 175 µl growth media was applied, then triturated at 160 µl/sec for the indicated number of repetitions. Please click here to view a larger version of this figure.

Figure 3. Typical adipose and fibroblast iPSC colony maintenance scheme and a series of colony density images to determine when to passage to maintain a regular feeding/passage cycle. (A) Beginning with an adipose or fibroblast 96-well iPSC colony ready to passage, one column of cells was passaged on the robot without centrifugation and diluted to 12 new columns which were fed at specific intervals thereafter until the plate was ready to passage again. For cell production, columns not used to maintain the colony were passaged and collected for subsequent use. Any number of columns may be passaged to control expansion rate. (B) To determine when to passage, a series of density images captured every 24 hr was provided to users. In the top red box (within 72 hr after initial seeding) the colony is not dense enough to passage and should be fed. When the density matches that shown in the green box (at ~96 hr), the colony is at an ideal density for passage. By 120 hr, the colony is too dense to passage and this scenario should be avoided. The latter may be resolved with a seeding density gradient but routine culture at this density is strongly discouraged. White bar equals 200 µM. Please click here to view a larger version of this figure.

Figure 4. 96-well robotic culture system allows users to start cultures from frozen or active cell lines and perform a seeding gradient. (A) A cell solution is prepared by the user from a frozen cell stock or after harvesting an active culture and placed on the robot. The robotic system then performs a serial dilution protocol to distribute the initial cell solution along the 96-well plate with the density gradient distributed by column. (B–E) Examples of cell density from four of the twelve columns from A, 48 hr after the density gradient seeding protocol. White lines equal 225 µm. Please click here to view a larger version of this figure.

Figure 5. Pluripotency is maintained for adipose and fibroblast derived iPSC lines during 96-well robotic culture for 22 (adipose) or 23 (fibroblast) passages (~3 months). (A–L) Adipose and fibroblast derived iPSC cell lines were robotically fed and passaged as described in the text without centrifugation in 96-well plates for 22 or 23 passages then fixed and stained for pluripotency markers Oct4, Nanog, or Ssea4 with DAPI counterstain. White bar equals 100 µm. (M–N) Parallel cultures to those described in A were submitted for karyotype analysis which revealed a normal complement of 46 chromosomes. Please click here to view a larger version of this figure.

Figure 6. Adipose (96-A) and fibroblast (96-F) derived iPSCs cultured for more than 20 passages in the 96-well robotic format maintained pluripotency gene expression and differentiated into cardiomyocytes. Total mRNA was extracted from both robotically cultured iPSC lines as well as cardiomyocytes differentiated from both lines (CM-A=adipose iPSC derived cardiomyocytes, CM-F=fibroblast iPSC derived cardiomyocytes, collected 14 days after differentiation induction) and used for RT-PCR to probe pluripotency markers NANOG, POU5F1, REX1, cardiomyocyte marker MYL7 and loading control GAPDH. Human dermal fibroblasts (HDF) were also collected and used as a non-iPSC, non-cardiomyocyte control. Please click here to view a larger version of this figure.

Figure 7. Fibroblast and adipose derived iPSCs maintained the ability to differentiate into cardiomyocytes after long term 96-well robotic culture. (A–H) Adipose and fibroblast iPSCs cultured in the 96-well robotic format for more than 20 passages were differentiated into cardiomyocytes. 14 days after differentiation induction, 96-well plate bound cardiomyocytes were dissociated and replated at a low density onto glass slides for immunostaining of troponin T (red), F-actin (green) and nuclei (blue). Please click here to view a larger version of this figure.

Figure 8. Scaling up 96-well plate stem cell culture. Scale up of stem cell production depends on the number of plates used to seed the subsequent group of plates. The user can either maintain one production level by passaging the same number of plates at each split opportunity or expand production by seeding more plates. For example, in Cycle 0, one 96-well plate is passaged to twelve new plates (Cycle 1), creating a 12 fold expansion. This rate may be maintained if one of the twelve plates is passaged to a new set of twelve plates (Cycle 1 to 2) or expanded by passaging multiple plates (Cycle 2 to 3). During passage, any plates not used to maintain the colony are available for downstream applications. Therefore, passaging 12 plates routinely could yield as many as 144 plates in two passages: 12 for colony maintenance and 132 for other uses. Please click here to view a larger version of this figure.

Figure 9. Efficiency of 96-well plate based culture for cell production greatly exceeds manual cultivation. As noted in Figure 8, one plate can be passaged to seed twelve plates, which can then be passaged to 144 plates. This rate of expansion can be achieved in two split opportunities. A technician requires about 3.5 minutes to feed one 6-well plate whereas they spend about 30 sec with one 96-well plate to examine it on the microscope and place it on the robot. If the technician is carrying 144 plates, they will spend about 1.2 hr handling plates when feeding with the robot whereas manual culture will require a dedicated 8 hr to feed. Please click here to view a larger version of this figure.