H9 hESCs were cultured on irradiated mouse embryonic fibroblast (MEF) as described previously (13, 14). hESCs were fed daily with DMEM/F12 medium supplemented with 20% Knockout serum replacement (Life Technologies, Seoul, Korea), 0.1 mM 2-Mercaptoethanol, 1% Non-essential amino acid, 1 mM glutamine, 100 U/ml penicillin G, 100 μg/ml streptomycin, and 4 ng/ml basic fibroblast growth factor (PeproTech, Rocky Hill, NJ, USA). Feeder-free culture method was performed as described previously (13, 15). Briefly, 10 μM of Y27632 (Sigma-Aldrich, Seoul, Korea) were treated for 1 hour prior to detaching hESCs, the cells were harvested with trypsin-EDTA. After elimination of MEF cells on gelatin coated tissue culture dish, hESCs were cultured in conditioned medium supplemented with 10 μM of Y27632 on tissue culture dish coated with Matrigel (BD Biosciences, Seoul, Korea). For differentiation of hESCs, cells were treated 10⁻⁵ M retinoic acids (RA) in culture medium for 10~14 days. Colo205 cells were maintained following recommendation of ATCC.

Antibody was biotinylated by using DSB-X^TM biotin labeling kit (Life Technologies) according to manufacturer’s procedure.

Cell isolation was carried out by using Dynabeads FlowComp Flexi kit (Life Technologies) according to the manufacturer’s protocol. The cells were detached with dissociation solution for human ES/iPS cells (ReproCELL). To prevent physiological cell damage, dissociation buffer was immediately neutralized by complete medium. The mixture of undifferentiated and RA-treated hESCs were suspended in 500 μl of cold isolation buffer (phosphate buffered saline (PBS), pH 7.4, 5% Knockout serum replacement, 1 μg/ml Doxycyclin, 0.5 mM Mg²⁺, 1 mM Ca²⁺) in Eppendorf tube and incubated with 20 μg of biotinylated 57-C11 for 30 min at 4°C with rolling and tilting. Cells were harvested after addition of 500 μl of cold isolation buffer. Cells (2×10⁶) were then resuspended in 500 μl of cold isolation buffer and incubated for 30 min at 4°C with rolling and tilting after addition of 50 μl (7.5×10⁷ beads) of Dynabeads (Life Technologies). 57-C11-positive hESCs were collected by placing the tube in a magnet for 2~3 min. The supernatant containing 57-C11-negative hESCs was carefully removed in the magnet, and the bead/cell complexes were cleaned up by a total of three washes. Cells were then removed from the magnet by incubation of 1 ml of FlowComp Release Buffer (Life Technologies) for 10 min at room temperature (RT) with rolling and tilting. The supernatant with 57-C11-positive hESCs was carefully transferred to a new tube and put on the magnet again to remove residual beads. The supernatant with 57-C11-positive hESCs was subjected to centrifugation at 600 × g for 10 min, and the bead free-cells were resuspended by stepwise addition of a 1:300 mixture of PBS: Matrigel (BD Bioscience) to a final total volume of 50 μl before injection.

The procedure of teratoma formation experiments was performed as described (16). All procedures were carried out under guidelines approved by the Institutional Review Board (SJU-2015-008) and Institutional Animal Care and Use Committee (IACUC No. SJ-20140801) of Sejong University. Male NOD/SCID mice (8~12 weeks) were used for all experiments. All procedures were carried out in a biological safety cabinet under sterile conditions using sterile materials. Injection of hESCs into testes was done under anesthesia. Mice received the following anesthetics and analgesics: Zoletil 50 (30 mg/kg)/Rompun (9.32 mg/kg). For testis injection, a sterile 1 ml syringe with a 31 G needle was injected into testes with hESC suspension (50 μl). After successful injection, the needle was retracted swiftly, the injection site was pressed with a cotton swab for about 1 min, and then a sterile, resorbable wound pad was placed at the injection site. Teratomas were surgically removed at 12~18 weeks postinjection under sterile conditions. The sizes of teratomas were measured by a ruler.

Flow cytometry analysis was performed described as previous study (10). Briefly, the undifferentiated H9, RA-induced differentiated H9 cells or Colo205 cells were incubated with 20 μg of DSB-X biotin-labeled 57-C11. After isolation process described above, the cells were washed with PBS (pH7.4) supplemented with 1% bovine serum albumin (PBA). The cells were incubated with 57-C11 or biotin-labeled 57-C11 for 30 min, followed by fluorescein isothiocyanate-conjugated anti-mIgG or streptavidin-phycoerythrin, depending on the isotypes of the primary antibodies. After washing with PBA, Propidium iodide-negative cells were analyzed by FACScalibur (BD biosciences).

Previously, we found that 57-C11-reactive E1B-AP5 is closely associated with pluripotent and undifferentiated H9 hESCs and the surface expression of 57-C11-reactive E1B-AP5 is downregulated during the differentiation of H9 cells (11). To test whether 57-C11 is able to specifically eliminate undifferentiated H9 cells from the mixture of undifferentiated and differentiated H9 cell population, in this study, we have developed a 57-C11-mediated magnetic cell separation system, based on the specificity of 57-C11 to undifferentiated H9 cells (11). First, 57-C11 was biotinylated and 57-C11-bound H9 cells were analyzed by flow cytometry analysis with streptavidin-phycoerythrin conjugate. As shown in Fig. 1A, H9 cells (10⁶ cells) were readily detected with biotin-labeled 57-C11 at final concentrations of 10 or 30 μg/ml. As compared with that of 10 μg/ml of biotin-labeled 57-C11, 57-C11 binding to H9 cells were not significantly increased with 30 μg/ml of biotin-labeled 57-C11, indicating that 10 μg/ml of biotin-labeled 57-C11 is enough to analyze one million hESCs. To confirm whether biotin-labeled 57-C11/streptavidin-magnetic beads were able to specifically isolate undifferentiated H9 cells from the mixed cell population, CA142 (IgG1, k), a MAb specific to surface antigen on Mycoplasma hyorhinis, was also biotinylated and used as a control isotype antibody (17). When the mixture of undifferentiated H9 hESCs and Colo205 cells were incubated with 10 or 20 μg/ml of biotin-labeled 57-C11 and subjected to the streptavidin-magnetic beads-associated isolation procedure, 57-C11-positive H9 cells were clearly separated with 20 μg/ml of biotin-labeled 57-C11 (data not shown). To reflect the possible scenario of contaminating hESC in mixed cell populations, the mixture of undifferentiated and differentiated hESCs were incubated with 20 μg/ml of biotin-labeled 57-C11 and subjected to the same isolation procedure (Fig. 2A). 57-C11-positive undifferentiated H9 cells were mostly removed with 20 μg/ml of biotin-labeled 57-C11 while the same H9 cells were not removed with the irrelevant antibody CA142 (data not shown and Fig. 1B). Thus, it seems that 57-C11/streptavidin beads is able to specifically isolate 57-C11-positive undifferentiated hESCs from the mixed cell population.

To examine the applicability and efficiency of this separation strategy, we prepared several pools of undifferentiated and differentiated H9 cells at different ratios (0, 10, 30, 50, or 100% undifferentiated H9 hESCs) and performed the same 57-C11-mediated isolation procedure (Fig. 2A). Approximately 10% of hESCs were recovered from 1×10⁵ cells of differentiated H9 cells after the isolation procedure, suggesting that some cells are isolated from the differentiated H9 cells during the isolation procedure (Fig. 1C). Approximately 91% of hESCs were recovered from 1×10⁵ cells of undifferentiated H9 cells after the isolation procedure (Fig. 1C). Our previous studies showed that approximately 89~96% of undifferentiated H9 cells were 57-C11-positive (11). Therefore, it seems that undifferentiated hESCs are almost completely recovered through the isolation procedure. When undifferentiated and differentiated hESCs were mixed at different ratios, approximately 77~100% of undifferentiated H9 hESCs were recovered after the isolation procedure (Fig. 1C). Thus, the spiking experiments suggest that the recovery rate of 57-C11-positive hESCs from mixed cell populations is approximately 77~100%.

Before we validated whether isolated 57-C11-positive H9 cells were tumorigenic in NOD/SCID mice, we tested how many undifferentiated hESCs were able to form teratomas in NOD/SCID mice. Undifferentiated hESCs rapidly undergo apoptosis after the preparation of single cell suspensions from the colonies. Therefore, more than one million hESCs are recommended to use for teratoma formation in immunodeficient mice (16, 18). When we injected with 1×10⁵, 4×10⁵, 1×10⁶, or 2×10⁶ cells of undifferentiated and RA-differentiated H9 cells into the left and right testes of the same mice, respectively, we found that undifferentiated H9 cells could form teratomas with 4×10⁵ cells under our experimental condition 8~18 weeks after injections while RA-differentiated H9 cells were not able to do (Fig. 3).

The overall scheme of the experiment including magnetic cell separation of 57-C11-positive and -negative H9 hESCs from mixed cell populations and teratoma formation of isolated H9 hESCs in NOD/SCID mice is shown in Fig. 2A. Before the injection of 57-C11-positive and -negative hESCs into NOD/SCID mice, the viability of isolated H9 cells were tested. After the isolation procedure, approximately 85% of 57-C11-negative cells remained alive while only 25% of 57-C11 positive cells remained alive (Fig. 2B). It seems that the releasing processes from magnetic beads are harmful to 57-C11-positive H9 cells. Therefore, the viability of injected hESCs were checked before injection and 1×10⁶ cells of 57-C11-positive and negative hESCs were injected into left and right testes of NOD/SCID mice. Untreated and RA-differentiated H9 cells were also injected into NOD/SCID mice as a control experiment in the same way. When the first set of mice were opened at 12 weeks postinjection, teratoma formation was not observed even with untreated H9 cells (Fig. 4A). When the second set of mice were opened at 18 weeks postinjection, teratoma formation was observed with 57-C11-positive H9 cells or untreated H9 cells whereas it was not observed with 57-C11-negative H9 cells or RA-treated H9 cells (Fig. 4A and 4B). The results suggest that 57-C11-positive hESCs are undifferentiated and pluripotent hESCs whereas 57-C11-negative hESCs are not.