SEARCH

SEARCH BY CITATION

Keywords:

  • Embryonic stem cells;
  • Embryoid body;
  • Differentiation;
  • Suspension culture

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Embryonic stem (ES) cells hold great promise as a robust cell source for cell-based therapies and as a model of early embryonic development. Current experimental methods for differentiation of ES cells via embryoid body (EB) formation are either inherently incapable of larger-scale production or exhibit limited control over cell aggregation during EB formation and subsequent EB agglomeration. This report describes and characterizes a novel method for formation of EBs using rotary orbital motion that simultaneously addresses both concerns. EBs formed under rotary suspension conditions were compared with hanging-drop and static EBs for efficiency of EB formation, cell and EB yield, homogeneity of EB size and shape, and gene expression. A 20-fold enhancement in the number of cells incorporated into primitive EBs in rotary versus static conditions was detected after the first 12 hours, and a fourfold increase in total cell yield was achieved by rotary culture after 7 days. Morphometric analysis of EBs demonstrated formation and maintenance of a more uniform EB population under rotary conditions compared with hanging-drop and static conditions. Quantitative gene expression analysis indicated that rotary EBs differentiated normally, on the basis of expression of ectoderm, endoderm, and mesoderm markers. Increased levels of endoderm gene expression, along with cystic EB formation, indicated by histological examination, suggested that differentiation was accelerated in rotary EBs. Thus, the rotary suspension culture method can produce a highly uniform population of efficiently differentiating EBs in large quantities in a manner that can be easily implemented by basic research laboratories conducting ES cell differentiation studies.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Embryonic stem (ES) cells have the ability to differentiate into all somatic cell types [1, [2]3], enabling them to serve as a potent cell source for tissue engineering and regenerative medicine applications, as well as an in vitro model system to study early embryonic development. Differentiation of mouse and human ES cells can be induced in a variety of manners, but the most common technique is via formation of cell aggregates in nonadherent spheroids referred to as embryoid bodies (EBs) [4, [5]6]. The molecular and cellular morphogenic signals and events within EBs recapitulate many aspects of the developing embryo, including differentiation to cells of endoderm, mesoderm, and ectoderm lineages, similar to gastrulation in vivo [7, 8]. Embryoid bodies generally develop an outer layer of primitive endoderm that envelops a core of differentiating cells that take on a cystic appearance [8]. The precise number and spatial coordination of the various cell-cell interactions involved in EB formation are thought to influence the course of ES cell differentiation, and as a result, controlling cell number and size of EBs may be an important step in directed differentiation strategies.

Two techniques typically used for EB formation are hanging-drop and static suspension culture [4]. The hanging-drop method, in which cells are dispersed in 15–20-μl drops suspended from the lid of a Petri dish, has been used to more precisely control the microenvironment for EB formation and differentiation [9, [10]11]. Since EBs formed in each drop are physically separated, individual EBs are not able to agglomerate, allowing production of a fairly homogeneous final population of EBs. This method, however, is not practical for large-scale applications, as each drop generally yields only one EB, and a typical 100-mm Petri dish can accommodate only approximately 100 drops. Suspension culture, on the other hand, can be used to produce EBs in larger quantities (hundreds to thousands) by simply inoculating ES cells in differentiation medium at a density of 104–106 cells per milliliter on a nonadherent surface, whereupon the cells spontaneously aggregate into spheroids. Whereas this method can be used to produce a larger number of EBs and differentiated cells than hanging drops, it offers little control over the size and shape of EBs and often results in agglomeration of EBs into large, irregular masses. Currently, there are no techniques for laboratory-scale production of EBs that offer a satisfactory compromise between the controlled environment of hanging drops and the ease of producing large quantities of EBs by suspension culture methods.

Recently, spinner flask, slow-turning lateral vessel (STLV), and high-aspect rotating vessel (HARV) bioreactors have been used for larger-scale production of EBs [12, [13], [14], [15], [16], [17]18]. The agitation provided by bioreactors improves circulation of nutrients and cell waste products, but it can also be manipulated to control cell aggregation [19, [20]21]. As a result, certain bioreactors have been shown to produce a controlled environment for EB formation and differentiation. Although such approaches may provide a more uniform differentiation environment capable of sustaining increased EB and differentiated cell yield, bioreactors may not be practical solutions for many researchers examining multiple experimental samples in parallel. For example, the relatively large volume of even the smallest common bioreactors (typically >100 ml per bioreactor) makes conducting screening studies of various medium compositions much more laborious and expensive for researchers studying basic mechanisms of ES cell differentiation. Thus, an improved method for simple, laboratory-scale EB differentiation capable of mimicking the beneficial attributes of larger-volume bioreactors would be useful to many stem cell researchers.

Orbital rotary shakers have been used to create spheroids from cell types such as tumor cells [22, 23], hepatocytes [24], neural stem cells [25], and fetal brain cells [26] since the constant circular motion provided by this simple system generally increases the efficiency of spheroid formation. Multiple cell culture dishes can be accommodated on the rotary platform, easily allowing production of numerous parallel samples and comparison of different experimental parameters. Although horizontal rotation of mouse ES cells has been noted qualitatively to improve the homogeneity and reproducibility of EB formation [27], a quantitative characterization of the effects of rotary culture in comparison with hanging-drop and static suspension EB culture methods has not been previously described. In this study, differences in the formation efficiency, homogeneity, yield, and differentiation of suspension EBs formed by hanging-drop, static, and rotary conditions were directly compared. EBs initiated on a rotary orbital shaker formed more efficiently, demonstrated higher cell and EB yield, and exhibited improved homogeneity of size and shape compared with EBs formed in static suspension conditions. In addition, rotary EBs appeared to differentiate more efficiently than static suspension cultures on the basis of morphological appearance and gene expression profile patterns. These results suggest that rotary suspension culture is a simple method for EB formation and differentiation that is significantly better than hanging-drop and static suspension culture.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Embryonic Stem Cell Culture

Murine embryonic stem cells (D3) [5] (passages 23–32) were maintained in an undifferentiated state on tissue culture dishes coated with 0.67% gelatin in Dulbecco's modified Eagle's medium (Mediatech Inc., Herndon, VA, http://www.cellgro.com), supplemented with 15% fetal bovine serum (HyClone, Logan, UT, http://www.hyclone.com), 2 mM l-glutamine (Mediatech), 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml amphotericin (Mediatech), 1× minimal essential medium nonessential amino acid solution (Mediatech), 0.1 mM 2-mercaptoethanol (Fisher Scientific, Fairlawn, NJ, http://www.fishersci.com), and 103 U/ml leukemia inhibitory factor (LIF) (Chemicon International, Temecula, CA, http://www.chemicon.com). Cells were routinely passaged every 2–3 days prior to reaching 70% confluence.

Embryoid Body Formation

Undifferentiated ES cells were dissociated from monolayer culture with a 0.05% trypsin-EDTA solution (Mediatech) to obtain a single-cell suspension. Suspension cultures of embryoid bodies were initiated by resuspending 4 × 105 cells per milliliter in differentiation medium consisting of undifferentiated growth medium without LIF. Static and rotary suspension cultures were cultivated in sterile 100- × 15-mm nontissue culture-treated polystyrene (i.e., bacteriological grade) Petri dishes (Becton Dickinson Biosciences, San Jose, CA, http://www.bdbiosciences.com) with a final volume of 10 ml of differentiation medium. Rotary suspension culture EBs were initiated by placing dishes on an orbital rotary shaker (Lab-Line Lab Rotator; Barnstead International, Dubuque, IA, http://www.barnsteadthermolyne.com) set at approximately 40 revolutions per minute, except during studies in which rotary speeds of 25 and 55 rpm were also examined. Embryoid bodies were maintained for up to 7 days in suspension, with medium changed every 1–2 days by allowing EBs to sediment in 15-ml centrifuge tubes, aspirating the old medium, and resuspending in fresh medium. Hanging-drop EBs were prepared by suspending 4 × 104 cells per milliliter in differentiation medium in 15-μl drops on the lids of a 100- × 15-mm square Petri dishes, with ∼72 drops per plate. This concentration and this volume (resulting in ∼600 cells per drop) were chosen because lower concentrations resulted in inefficient production of EBs (i.e., many drops do not produce an EB), and higher concentrations often resulted in formation of multiple EBs per drop.

Efficiency of Embryoid Body Formation

The efficiency of formation for static and rotary EBs was determined by quantifying the percentage of total cells that incorporated into EBs, as opposed to cells that remained free in suspension or adhered to the culture dish. After 6 and 12 hours, static and rotary suspension cultures were passed through a 40-μm cell strainer to separate cell aggregates constituting primitive EBs from individual cells. The cell strainer was rinsed five times with 1.0 ml of phosphate-buffered saline (PBS), and the suspension of individual cells that readily passed through the cell strainer was centrifuged at 1,000 rpm for 5 minutes to pellet the cells. The supernatant was aspirated, and the cells were resuspended in 1.0 ml of PBS. EBs trapped in the cell strainer were collected by inverting the strainer over a 50-ml centrifuge tube and rinsing thoroughly with PBS to dislodge the EBs. The EBs were allowed to sediment by gravity, and the supernatant was aspirated. The EB pellet was rinsed once with 5.0 ml of PBS, and the supernatant was again aspirated after the EBs settled. The EBs were trypsinized for 5 minutes at 37°C in 0.05% trypsin-EDTA and neutralized with an equal volume of culture medium. In addition, the cells attached to the culture dish were rinsed with PBS and trypsinized. The three fractions of cells (cells incorporated into EBs, individual cells in suspension, and adherent cells attached to the culture dish) were quantified using a Coulter Multisizer III with a 100-μm aperture (Beckman Coulter Inc., Fullerton, CA, http://www.beckmancoulter.com). The total number of cells was calculated from the sum of the three cell fractions and used to determine the percentage of cells in each fraction.

Yield of Embryoid Bodies and Cells

Embryoid body and differentiating cell yield were assessed by directly counting individual EBs and dissociated cells from EBs in separate parallel samples. Static and rotary EBs were collected by gravity sedimentation, the excess medium was aspirated, and EBs were suspended in 2.0 ml of a 60:40 solution of Isoton II (Beckman Coulter):glycerol (Fisher Scientific). Triplicate samples of this solution were each added to 150 ml of Isoton/glycerol solution and counted using a Coulter Multisizer III equipped with a 2,000-μm aperture. The total number of EBs present was calculated based on this count, taking the appropriate dilution factor into account. The total number of cells within EBs was determined by dissociating the EBs with 0.25% trypsin-EDTA (Mediatech) and counting the cells present with a 100-μm aperture Coulter counter, as described above (higher trypsin concentration was used in this study than in previous studies to more efficiently dissociate EBs).

LIVE/DEAD Assay

Static and rotary EBs were formed as described above, and after 2, 4, and 7 days of differentiation, cell viability was assessed using LIVE/DEAD staining (Molecular Probes Inc., Eugene, OR, http://www.probes.invitrogen.com). Samples were incubated in serum-free, phenol red-free medium containing 1 μM calcein AM and 2 μM ethidium homodimer I at room temperature for 30 minutes. EBs were then washed three times with PBS, resuspended in fresh PBS, transferred to six-well plates, and immediately imaged using fluorescent microscopy.

5-Bromo-2′-deoxyuridine Assay

Cell proliferation was assessed using immunostaining for 5-bromo-2′-deoxyuridine (BrdU) incorporated into cells synthesizing DNA. After 2 and 5 days of suspension culture, static and rotary EBs were pulsed with 10 μM BrdU (Molecular Probes) for 6 hours. EBs were then washed twice with PBS, fixed in 10% formalin for 30 minutes, washed three times in PBS, and resuspended in HistoGel (Richard-Allan Scientific, Kalamazoo, MI, http://www.rallansci.com). Fixed samples were dehydrated in a series of graduated alcohol solutions (70%–100%) and xylene and embedded in paraffin, and 5-μm paraffin sections were cut using a rotary microtome. Sections were deparaffinized, and antigen retrieval was performed by incubating slides in 1.5 N HCl for 15 minutes at 37°C followed by two 5-minute rinses in 0.1 M borax buffer (pH 8.5). Slides were incubated for 30 minutes in blocking buffer (2% goat serum, 0.05% Triton X-100 in PBS), followed by a 2-hour incubation with peroxidase-conjugated anti-BrdU antibody (1:50 dilution; Roche Applied Science, Indianapolis, IN, http://www.roche-applied-science.com). Slides were then washed three times in PBS prior to colorimetric development of diaminobenzidine (DAB) (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Nuclei were counterstained with hematoxylin, and slides were coverslipped.

Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling Assay

Apoptotic cells in static and rotary EBs were detected using a terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay. After 4 and 7 days of suspension culture, static and rotary EBs were fixed in 10% formalin for 30 minutes, washed three times with PBS, and resuspended in HistoGel. Fixed samples were processed and embedded as described above, and 5-μm paraffin sections were cut using a rotary microtome. Apoptotic cells were identified using the DeadEnd colorimetric TUNEL system (Promega, Madison, WI, http://www.promega.com). Briefly, sections were deparaffinized in xylene, rehydrated in graded ethanol washes, and rinsed in PBS. Sections were permeabilized in proteinase K solution and then incubated in buffer containing biotinylated nucleotide mix and recombinant terminal deoxynucleotidyl transferase. A streptavidin horseradish peroxidase solution was then added to bind to biotinylated nucleotides incorporated into the DNA of apoptotic cells. The DAB substrate was then added for colorimetric development to stain apoptotic cell nuclei brown, and cell nuclei were counterstained blue with hematoxylin.

Morphometric Analysis

Phase images of EBs were acquired daily during the course of hanging-drop, static, and rotary culture with a Nikon TE 2000 inverted microscope (Nikon Inc., Melville, NY, http://www.nikonusa.com) with a SpotFLEX camera (Diagnostic Instruments, Inc., Sterling Heights, MI, http://www.diaginc.com). EBs were analyzed using ImageJ image analysis software (http://rsb.info.nih.gov/ij) to measure cross-sectional area and perimeter and calculate circularity (defined as 4π× [Area/Perimeter2]) of static and rotary EB samples for at least 100 EBs from each sample (n = 3). Hanging-drop EBs were analyzed by transferring two plates of hanging-drop EBs (roughly 100 EBs) into suspension immediately before imaging. Histogram plots of area and circularity values were generated with equal numbers of static and rotary EBs divided into 30 equally sized bins (because of the lower number of hanging-drop EBs present, only 12 bins were used for hanging-drop area analysis). The fraction of EBs within each bin was plotted versus the upper bin limit for both area and circularity. Mode absolute deviation (MAD) for area plot analysis was calculated for static and rotary EBs at each day as the sum of the absolute difference between each area value and modal area value divided by the total number of data points for the 95% least deviant values.

Histology

Embryoid bodies were formed as described above, and after 2, 4, and 7 days of culture, samples were harvested for histological examination. Samples were fixed in 10% formalin for 30 minutes and were processed, embedded, and sectioned as described above. Slides were then stained with H&E.

Reverse-Transcription Polymerase Chain Reaction Analysis

RNA was extracted from undifferentiated ES cells on day 0 and from hanging-drop, static, and rotary EBs on days 2, 4, and 7 using the RNeasy Mini kit (Qiagen, Valencia, CA, http://www1.qiagen.com). Reverse transcription for complementary DNA synthesis was performed with 1 μg of RNA per sample using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA, http://www.bio-rad.com), and real-time polymerase chain reaction (PCR) was performed with SYBR Green technology on the MyiQ cycler (Bio-Rad). Forward and reverse primers for Oct-4, Nanog, Brachyury-T, nestin, α-fetoprotein (AFP), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were designed with Beacon Designer software (sequences and conditions are given in supplemental online Table 1) purchased from Invitrogen (Carlsbad, CA, http://www.invitrogen.com) and validated with appropriate controls. Oct-4 and Nanog relative gene expression were quantified as compared with undifferentiated ES cell expression levels using the Pfaffl method of quantification [28]. Brachyury-T, nestin, and AFP concentrations were calculated using standard curves and normalized to GAPDH expression levels.

Statistical Analysis

All experimental samples were analyzed in triplicate, with data presented as mean ± SD. Statistical significance was determined using Student's t tests with a significance level of p < .05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Time Course of Embryoid Body Formation

Static and rotary suspension cultures were initiated with equal numbers of cells (4 × 105 cells per milliliter), and hanging-drop EBs were formed with ∼600 cells per drop (4 × 104 cells per milliliter). Rotary cultures appeared to be forming regular, spherical EBs after 2 days, whereas static EBs appeared smaller and formed less defined cell aggregates (Fig. 1A). At the same time point, EBs had formed in hanging drops and appeared similar in shape to rotary EBs yet slightly larger in diameter (∼250 μm diameter for hanging-drop, compared with ∼190 μm diameter for rotary). By day 4, hanging-drop EBs increased in size (∼500 μm diameter) and remained largely uniform in shape, although some irregularly shaped EBs were present. Static cultures formed small, spherical EBs, as well as larger, irregularly shaped aggregates of individual EBs. Rotary EBs at day 4 were larger than at day 2 (∼250 μm), but they remained regular and spherical with no appearance of agglomeration. Day 7 hanging-drop EBs increased slightly in size from day 4 (∼625 μm), and static cultures showed fewer small spherical EBs than at day 4, as well as many more large agglomerates of multiple EBs. Rotary EBs at day 7 were larger in size than at day 4 (∼290 μm) and remained largely spherical and uniform in appearance. Higher magnification of EBs revealed that hanging-drop and rotary EBs at 4 and 7 days of suspension culture exhibited a smooth, primitive endoderm-looking cell layer at the periphery of most EBs, whereas static EBs displayed a more irregular layer of cells at the edges of the EBs (Fig. 1, insets).

thumbnail image

Figure Figure 1.. Time course of embryoid body formation for hanging-drop, static and rotary suspension conditions. (A): Hanging-drop and rotary EBs formed by day 2 and increased in size at days 4 and 7 while showing minimal signs of agglomeration. Static cultures formed small, irregular cell aggregates at day 2. By day 4, both small, spherical EBs and large agglomerates were present, and by day 7, the majority of static EBs were large, irregular clusters. (B): The numbers of cells contained in embryoid bodies, free in suspension, and adhered to the plate were assessed at 6 and 12 hours, with significantly more cells found in EBs in rotary culture than in static cultures after both 6 and 12 hours. Static cultures exhibited an increased tendency for free cells to begin to adhere to the surface of the plate. (C): Rotary EBs initiated at 25 rpm formed large EBs by day 2, whereas EBs at 40 and 55 rpm rotary conditions formed a greater number of smaller EBs. **, p < .01 compared with static at the same time point. Scale bar = 500 μm.

Download figure to PowerPoint

Influence of Rotary Speed on Embryoid Body Size

To study the effect of rotary speed on EB formation, 4 × 105 cells per milliliter were inoculated in dishes placed on orbital shakers at 25, 40, or 55 rpm and observed using phase-contrast microscopy (Fig. 1C). By day 2, EBs formed on rotary shakers at 25 rpm were quite large, roughly 400 μm in diameter (406 ± 64 μm). In addition, fewer total EBs were present in 25 rpm EBs than at 40 or 55 rpm, as counts of EBs sampled from each population indicated that roughly 10 times more EBs were present at 40 rpm than 25 rpm. After 2 days, EBs formed at 55 rpm were smaller and appeared to have more free cells not incorporated into EBs than either 25- or 40-rpm samples (Fig. 1C). These trends persisted throughout 7 days of suspension culture, with 25-rpm rotary motion producing the largest and fewest EBs, and the 55-rpm shaker yielding the smallest EBs and a higher proportion of free cells in suspension. Because of the relatively large number of EBs obtained compared with 25 rpm, and the more efficient formation of EBs (indicated by fewer individual cells remaining in suspension) compared with 55 rpm, 40 rpm was used as the rotary speed for all of the subsequent characterization experiments for this study.

Efficiency of Embryoid Body Formation

From the preliminary time course studies, it appeared that rotary EBs formed more efficiently than static cultures within the first 1–2 days of suspension. To quantify this more specifically, the fractions of cells incorporated within EBs, remaining in suspension, or attached to the polystyrene dish surface were counted using a Coulter counter after 6 and 12 hours of EB formation. By 6 hours, 69.3% ± 3.2% of cells were forming putative EBs in rotary cultures (40 rpm) compared with only 5.6% ± 3.4% for static EBs (Fig. 1B). After an additional 6 hours of suspension culture (12 hours total), 88.4% ± 4.6% of cells were incorporated into primitive EBs for rotary samples, whereas in static cultures, only 4.5% ± 2.6% of cells were contained within floating EBs. In addition, 79.9% ± 3.3% of the cells adhered to the culture dish under static conditions, whereas no adherent cells could be obtained from the rotary samples. This suggested that whereas free cells in rotary suspension continued to incorporate into EBs during the first 12 hours of culture, static suspension cells became increasingly adherent to the polystyrene surface, thereby delaying or reducing overall EB formation.

Embryoid Body and Cell Yield

In addition to an increase in the efficiency of EB formation, the number of cells and EBs formed under rotary conditions also increased compared with static conditions. This was initially observed in preliminary studies by a considerably larger pellet of rotary EBs collected during refeedings, as well as the need for more frequent medium changes to sustain the rotary cultures properly. To quantify the yield of EBs and total differentiating cells, static and rotary EBs were collected at days 2, 4, and 7 and counted directly or enzymatically dissociated with trypsin to obtain a single-cell suspension prior to counting. The number of EBs present was greater for rotary than static cultures at each of the time points examined (Fig. 2A, inset). At day 2, 8,118 ± 1,183 rotary EBs were present, compared with only 3,459 ± 1,088 static EBs. However, because of the detection limits of the instrument used, cell aggregates less than 50 μm in diameter could not be counted. As a result, many of the small clusters of cells formed by day 2 in static suspension, shown in Figure 1A, were not counted. By day 7, the number of rotary EBs decreased to 6,724 ± 430, and the number of static EBs decreased slightly to 3,255 ± 671. Although EB yield revealed significant differences between the static and rotary cultures, the total number of differentiating cells is a more accurate and meaningful metric of overall yield. At days 2 and 4 of differentiation, modest increases in cell yield were observed in rotary cultures relative to static cultures (Fig. 2A), but by day 7, rotary cultures contained an average of 95.9 ± 5.0 × 106 cells, compared with only 24.4 ± 1.2 × 106 cells for static culture.

thumbnail image

Figure Figure 2.. Time course comparison of cell and embryoid body yield in static and rotary conditions. (A): Total cell yield counted from dissociated EBs indicated that rotary culture yielded significantly more cells than static culture at days 2 and 7 of differentiation. In addition, whole EBs were counted at days 2, 4, and 7, revealing significantly more EBs produced by rotary culture than static culture at days 2 and 7 (inset). *, p < .05; **, p < .01. (B): LIVE/DEAD staining revealed that majority of cells within EBs were viable (green stain) for both static and rotary EBs, whereas static EBs appeared to contain more dead cells (red stain) than rotary EBs (white arrows indicate clusters of dead cells). (C): Apoptotic cells in EBs were labeled brown using a terminal deoxynucleotidyl transferase dUTP nick-end labeling assay and revealed the presence of apoptotic cores at the center of static EBs (black arrows).

Download figure to PowerPoint

Viability of Cells Within Embryoid Bodies

To elucidate the cause of differences in cell yield between static and rotary EBs at later time points, cell viability, proliferation, and apoptosis assays were performed. LIVE/DEAD staining performed at days 2, 4, and 7 indicated that at all three points, both static and rotary EBs contained mostly viable cells, as indicated by prevalent green staining. Staining of day 2 EBs revealed no difference in the presence of dead cells between static and rotary EBs (data not shown), but by day 4, static EBs appeared to contain more clusters of dead cells (red stain, clusters indicated by white arrows) than rotary EBs, which contained sparse, individual dead cells (Fig. 2B). This trend appeared to continue throughout EB differentiation, with dead cells found more prevalently in static EBs than rotary at day 7 (data not shown). BrdU staining was performed to detect proliferating cells within static and rotary EBs after 2 and 5 days of differentiation. Low levels of proliferating cells were detected at each time point, likely due to the relatively long doubling time of cells within EBs, and no differences were noted between static and rotary EBs (data not shown). TUNEL staining for the detection of apoptotic cells revealed no differences between static and rotary EBs at day 4 of suspension culture; however, at day 7, more static EBs contained apoptotic cells than rotary EBs (Fig. 2C). In addition, dense cores of apoptotic cells were found within static EBs (Fig. 2C, arrows), whereas positive staining within rotary EBs was sparser and appeared to constitute less of the total EB area. These data indicate that static EBs show reduced cell viability, perhaps because of a greater number of cells undergoing programmed cell death compared with rotary EBs.

Homogeneity of Embryoid Body Formation

In an attempt to quantify the homogeneity of EB formation, the area and circularity of EBs formed in hanging-drop, static, and rotary cultures were assessed by image analysis methods. Hanging-drop culture resulted in EBs with a fairly narrow area distribution at day 2, followed by considerable increase in cross-sectional area and dramatic broadening of the area distribution at days 4 and 7, indicating deviation over time from an initially homogeneous size distribution (Fig. 3A). Static cultures formed small putative EBs at day 2 and increased in size at days 4 and 7, as shown in Figure 3B. As with hanging-drop EBs, static EBs showed broadening of the area distribution curve over the course of differentiation. By day 7, the distribution exhibited no clear peak, as EBs varied widely from very small-cell clusters to giant agglomerates of EBs. Rotary cultures showed an increase in size with time, yet the peak broadening was much less pronounced than for hanging-drop or static EBs (Fig. 3C). Distribution broadening was quantified by calculating the MAD for each histogram for static and rotary EBs (Fig. 3D). The MAD index was larger in static cultures than in rotary at all time points studied, with differences approaching significant levels by day 4 (p < .06) and statistical significance achieved by day 7 (p < .05).

thumbnail image

Figure Figure 3.. Homogeneity of embryoid body formation. Embryoid bodies formed by hanging-drop, static, and rotary cultures were imaged at days 2, 4, and 7 of suspension culture, and the area (A–D) and circularity (E–H) of EBs were examined. (A–D): Area values for hanging-drop (A), static (B), and rotary (C) EBs were sorted into 30 bins (12 for hanging-drop), and the number of values in each bin was plotted versus upper bin limit (note scale differences for hanging-drop). Rotary EBs showed more uniform distribution than hanging-drop and static EBs, indicated by larger peak values and smaller peak widths. Mode absolute deviation showed greater deviance in static culture than in rotary culture (C). (E–H): Circularity was more uniform for hanging-drop (E) and rotary (G) EBs than static EBs (F), and a significantly higher fraction of rotary EBs had a circularity value greater than 0.8 at all time points studied (H). *, p < .05; **, p < .01. Abbreviation: MAD, mode absolute deviation.

Download figure to PowerPoint

Circularity [4π× (Area/Perimeter2]) was used to indicate the degree to which irregular clusters of EBs formed as opposed to regular, spherical EBs. Hanging-drop EBs showed narrow distributions with sharps peaks close to 0.9 at days 2, 4, and 7, indicating formation and maintenance of spherical EBs (Fig. 3E). Static cultures displayed a very broad circularity distribution at day 2, with no discernable peak present (Fig. 3F). In contrast, day 2 rotary cultures exhibited a sharp peak around 0.9, similar to hanging-drop cultures (Fig. 3G). This was in agreement with the observation that rotary EBs formed earlier (Fig. 1C) and were regular in shape at day 2, whereas static cultures showed irregular multicell aggregates at the same time point (Fig. 1A). By day 4, static EBs showed an increase in circularity, with a small peak present around 0.9. This coincided with the appearance of EBs by day 4 (Fig. 1A). Rotary cultures still displayed a peak around 0.9 that was slightly broadened compared with day 2, indicating slightly less uniform EB circularity. Day 7 distributions appeared similar to day 4 distributions for both rotary and static cultures, but they were slightly broader and negatively shifted, indicating decreased homogeneity of circularity. The fraction of EBs with circularity values greater than 0.8 was calculated for both rotary and static EBs, since 0.8 falls exactly in between circularity values associated with round, individual EBs (0.9) and the merger of two distinct EBs into a larger agglomerate (0.7) (supplemental online Fig. 1). The fraction of rotary EBs with circularity values >0.8 was significantly greater than static cultures at 2 (p < .01), 4 (p < .05), and 7 days (p < .01) (Fig. 3H).

Histology of Embryoid Bodies

Embryoid bodies were collected from hanging-drop, static, and rotary cultures at days 2, 4, and 7 of suspension culture and visually assessed by H&E staining. At day 2, static EBs were small and irregular, in agreement with live phase images in Figure 1A, and appeared to be largely undifferentiated, indicated by a lack of organized cellular structures (Fig. 4). Hanging-drop and rotary EBs appeared slightly larger than static EBs at day 2, but they exhibited a similar undifferentiated morphology. By day 4, hanging-drop EBs increased in size compared with day 2 and appeared to increase in packing density. Static EBs began to agglomerate at day 4, in agreement with phase images (Fig. 1), and showed little sign of differentiated cell morphologies. In contrast, rotary EBs did not appear to agglomerate and contained multiple small cystic structures (Fig. 4, black arrows) in a majority of the EBs. At day 4, more than 70% of rotary EBs contained at least one cyst, compared with less than 8% in static EBs. Day 7 hanging-drop EBs continued to increase in size and also began to display cysts, similar to rotary EBs. By day 7, static EBs began to display cystic structures, with 50% of EBs containing at least one cyst, compared with 56% in rotary. In addition, some EBs contained very large cysts (Fig. 4, red arrow) constituting a majority of the EB cross-sectional area. These structures were more apparent in day 4 rotary EBs than static samples; however, by day 7, static cultures exhibited a percentage of large cystic EBs similar to that of the rotary cultures. Large day 4 and day 7 static EBs began to exhibit necrotic core areas, suggesting possible diffusion limitations negatively impacting cell viability. This morphology was in agreement with cell yield data, which demonstrated diminished cell yield in static EBs, and also previous reports that suggest that EB agglomeration results in a reduction in viable cells [12, 14]. Rotary EBs, on the other hand, did not possess necrotic regions, consistent with results of the viability studies, and the majority of the cells distributed throughout rotary EBs appeared viable.

thumbnail image

Figure Figure 4.. Histology of embryoid bodies formed by hanging-drop, static, and rotary culture. Day 2 static EBs were small and irregular and lacked internal cellular organization, whereas day 2 hanging-drop and rotary EBs appeared to be similar to static, but slightly larger and more regularly shaped. Day 4 hanging-drop EBs increased in size, whereas static EBs showed signs of agglomeration, as indicated by large, irregularly shaped EBs. Day 4 rotary EBs showed more differentiated appearance, with small cysts forming (black arrows). By day 7, hanging-drop EBs were larger and formed cysts; static EBs displayed small cystic cavities, as well as large cysts occupying much of the EB; and rotary EBs exhibited large, completely cystic EBs (red arrow). Magnification, ×20. Scale bar = 100 μm.

Download figure to PowerPoint

Quantitative Gene Expression Analysis

Expression of the pluripotent markers Oct-4 and Nanog, the early mesoderm marker Brachyury-T, the ectoderm marker nestin, and the endoderm marker α-fetoprotein was assessed quantitatively using PCR. The expression of Oct-4 (Fig. 5A) and Nanog (supplemental online Fig. 2) relative to undifferentiated ES cells was similar among hanging-drop, static, and rotary EBs, indicating that rotary culture did not inhibit or delay normal ES cell differentiation. Brachyury-T expression was also similar between static and rotary EBs, with a transient increase in expression at day 4 that subsided by day 7, whereas hanging-drop EBs showed insignificantly higher expression levels at day 4, with similar downregulation by day 7 (Fig. 5B). Nestin expression was similar in EBs under all conditions, with no significant differences noted between groups for a given time point (Fig. 5C). α-Fetoprotein showed minimal expression prior to day 7, at which point hanging-drop, static, and rotary EBs displayed increased levels of the gene (Fig. 5D). The higher levels of AFP expressed by hanging-drop and rotary EBs correlated with the smooth, primitive endoderm layer shown in phase images (Fig. 1A, insets). These data suggest a possible enhancement of differentiation in rotary cultures relative to static. In addition, the expression of genes from all three germ layers indicates that EBs formed by rotary suspension retain their pluripotent differentiation capacity. The pluripotent capacity of rotary EBs is further supported by the observation of beating foci, neurite extensions, and other mature phenotypes after plating EBs on gelatin-coated substrates, similar to that seen in hanging-drop and static suspension EBs.

thumbnail image

Figure Figure 5.. Quantitative polymerase chain reaction (PCR) analysis of gene expression in hanging-drop, static, and rotary EBs. RNA extracted from EBs at days 2, 4, and 7 was compared for hanging-drop, rotary, and static cultures using quantitative PCR. Similar gene expression patterns were found for the pluripotent marker Oct-4 (A) and the mesoderm marker B-T (B), with hanging-drop EBs showing increased expression at day 4. Expression of the ectoderm marker nestin was similar for all groups (C), whereas cells from rotary EBs showed higher AFP expression than static EBs at day 7 (D), suggesting enhanced endoderm differentiation. *, p < .05. Abbreviations: AFP, α-fetoprotein; B-T, Brachyury-T; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

In this study, the effects of orbital rotary culture on the efficiency of EB formation, yield and viability of differentiating cells and EBs, homogeneity of size and shape of EBs, and differentiation were investigated relative to hanging-drop and static suspension culture. Rotary culture increased the efficiency of EB formation, resulting in a 20-fold increase in the number of ES cells incorporated into EBs compared with static culture after 12 hours, and rotary culture also significantly increased the yield of EBs and cells formed after 7 days of differentiation in suspension. EBs formed from rotary suspension also appeared to contain fewer dead and apoptotic cells than those from static suspension culture. Rotary culture produced EBs with greater homogeneity in size than either hanging-drop or static suspension culture, and it also produced an EB population with a homogeneity of shape similar to that of hanging-drop and greater than that of static suspension culture. In addition, EBs formed by the rotary process showed increased levels of endoderm expression compared with static EBs but comparable induction of mesoderm and ectoderm markers and downregulation of pluripotent markers during the course of differentiation. These results indicate that the rotary method of EB formation is an effective means for producing a controlled, uniform population of EBs in bulk suspension while simultaneously increasing the efficiency and yield of differentiating ES cells.

Physical separation techniques, such as hanging-drop cultures, provide a route to control the size and efficiency of initial EB formation and prevent EB agglomeration [29]. Similarly, EBs have been formed in individual conical tubes [30] or 96-well round-bottomed plates [31, 32] with or without centrifugation to promote cell aggregation for EB formation. Although each approach can effectively produce distinct EBs of similar size, individual formation and maintenance of EBs is laborious and inefficient. Thus, physical separation methods may be useful for small-scale studies of EB differentiation, but they are not feasible for larger-scale production of EBs in bulk suspension culture. As an alternative approach, EB formation in agarose [13, 33] and alginate [34] beads has been investigated as a means to “shield” EBs from one another in suspension and thereby prevent agglomeration. However, cell encapsulation methods require additional materials and equipment that may complicate EB production and affect the overall homogeneity of the EB population. In this study, consistent rotary motion alone during suspension culture enabled EBs to form rapidly, efficiently, and homogeneously during the first 2 days of culture. In addition, after EB formation, persistent rotary motion appeared to inhibit EB agglomeration, resulting in a more homogeneous population of EBs throughout the duration of suspension culture.

Continuously mixed culture methods have been used successfully with many cell types, such as hepatocytes [35], tumor cells [36], and neural stem cells [20, 25], to enhance the formation of cell aggregates. Similar strategies to increase the efficiency and scale of EB production from both mouse and human ES cells have been examined more recently with various bioreactors, such as spinner flasks, STLVs, and HARVs [12, 14, 15, 17, 18, 37]. HARVs have been shown to promote formation of large aggregates of EBs and induce cell death, thereby reducing the overall yield of viable cells and creating a more heterogeneous population of EBs [14]. Certain spinner flasks and STLVs, however, have been shown to reduce aggregation of EBs and thus sustain larger numbers of EBs. Interestingly, spinner flasks with bulb-shaped impellers are capable of forming EBs from single-cell suspensions, whereas spinner flasks with paddle-type impellers merely induce massive cell aggregation and EB agglomeration [17]. These observations, along with the efficient EB formation and reduced agglomeration seen in the present study, imply that the specific fluid mechanics profile of mixed culture systems may need to balance the induction of early cell aggregation required to form EBs with inhibition of EB agglomeration once individual EBs have formed.

Imposing flow on cells, particularly in three-dimensional culture systems, provides a number of beneficial in vitro effects, such as increased diffusion of nutrients and oxygen, that promote cell viability. Increasing the yield of viable differentiated cells relative to the starting number of ES cells is important, since the need for fewer input cells to produce EBs reduces the labor and cost of ES cell expansion prior to initiating differentiation protocols. In addition, reduction of EB agglomeration can also increase the yield of viable cells by preventing oversized EB formation [12, 14]. However, the benefits of dynamic culture systems exist only up to a certain threshold, whereupon relatively large shear stresses begin to exert a detrimental effect on cells. For example, as higher rotation speeds, such as 55 rpm, were examined, the cultures contained more free cells in suspension than slower speeds, although it was not clear whether free cells were being sheared off EBs or simply incorporating less into EBs originally. Furthermore, lower rotary speeds may have negative effects on EB viability and yield as well, as massive cell agglomerates have been observed at rotary speeds below 20 rpm, whereas 25 rpm produces roughly 1/10th as many EBs as 40 rpm. In contrast, rotary speeds close to 40 rpm appear to produce a population of EBs that form both efficiently, with few free cells remaining in suspension, and in high yield. Thus, it is likely that a range of values for rotational speed exists that benefits cell viability and overall differentiated cell yield.

The ability to control the size of EBs may be necessary for creating robust, reproducible strategies for ES cell differentiation because cell differentiation is influenced by spatial and temporal patterns of cell-cell interactions, and thus the size of individual EBs formed may affect their differentiation profiles. Hanging-drop size as a function of initial ES cell number has been integrated into differentiation protocols for different cell lineages [9]; however, this method is limited in its ability to effectively control EB size, as small differences in initial cell number within drops may result in large differences in EB size after just 2–3 population doublings, potentially accounting for the broad distributions observed in hanging-drop area at days 4 and 7 in this study (Fig. 3A). Rotation speed and cell density of larger-volume, stirred-culture systems have been shown to influence the size and shape of EBs, but differences in cell differentiation have not been described [14, 17, 37]. Using rotary orbital culture, lower rotary speeds, such as 25 rpm, produced larger EBs (Fig. 1B), whereas higher speeds, such as 55 rpm, yielded much smaller EBs, indicating that an inverse relationship between rotary speed and initial EB formation size exists. Such observations warrant additional studies to more precisely define the relationship between EB size and rotary speed as a means to spatially control the size of EB populations in bulk suspension more accurately and explore the influence of EB size and fluid shear forces on cell differentiation pathways.

In addition to creating a more homogeneous and larger population of EBs in bulk suspension, rotary culture may induce differentiation of ES cells more efficiently than static suspension culture. Mature embryoid bodies are generally considered to be formed from an outer layer of primitive endoderm surrounding an interior of differentiating cells, with the endoderm providing molecular signals necessary for proper differentiation [8, 38]. The morphological appearance of a primitive endoderm surface layer and increased levels of AFP gene expression in rotary EBs appear to indicate enhanced endoderm formation. Furthermore, formation of cystic structures in the interior of EBs is generally considered to be a positive indication of ES cell differentiation. The earlier and more pronounced appearance of cystic structures in rotary EBs than in static EBs suggests that differentiation is induced sooner and could lead to more rapid and efficient differentiation of subsequent cell types dependent on endoderm-derived morphogenic cues for proper differentiation. Additional studies remain ongoing to assess the effects of rotary suspension culture on the differentiation of other cell types, such as cardiomyocytes and neural progenitors, after the suspension phase of EB differentiation (i.e., >7 days) when the EBs are replated onto adherent substrates and permitted to differentiate for longer periods of time.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

In summary, we have shown that a simple rotary shaker can be used to create a dynamic cell culture environment capable of efficiently producing a homogeneous population of EBs with a high overall cell yield. This should assist researchers conducting basic and applied studies of ES cell differentiation by improving the reproducibility and reliability of EB suspension cultures.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

We thank Ross Marklein, Jonathan Kent, Beth Krauth, and Geoffrey Berguig for providing technical assistance and performing data analysis; Dr. Brani Vidakovic for assistance with statistical analysis; and Dr. Taby Ahsan for critical review of the manuscript. R.L.C. was partially supported by a GAANN Fellowship from the Center for Drug Design, Development and Delivery (DoE P200AD60118), and C.Y.S. is supported by a National Science Foundation Graduate Fellowship. This work was supported in part by funds from the Whitaker Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
sc-06-0523-SupplementalData.pdf203KSupplemental Data

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.