Correlation between PSC differentiation potential and level of CHD7 expression

The potential to differentiate is a critical feature of PSCs used for cell transplantation therapy. Therefore, establishing an assay to evaluate differentiation potential is essential for the maintenance culture of PSCs. EB formation in EB assays is used as a minimum requirement to demonstrate differentiation potential, although EB formation assays may not necessarily guarantee the ability to differentiate into the designated target cells without bias. We used ESC H9 cells in the majority of experiments shown in this study as a representative PSC cell line to minimize the concern of clonal variance in PSC clones that is typically observed among iPSC clones generated from somatic cells with various genetic and epigenetic profiles and with versatile reprogramming methods. H9 cells cultured on VTN-Ncoated dishes with Es8 (Thermo Fisher) medium formed a considerable number of EBs; however, the number of EBs was reduced considerably after the cells were transferred to RFF2 medium and cultured for 15days (3days/passage5). The cells showed an ability to form a comparable number of EBs again when transferred to Es8 and cultured for 24days (3days/passage8 passages), consistent with our previous report using ESC KhES-1 and iPSC PFX#91. The expression level of CHD7 determined by flow cytometry and the copy number of CHD7 measured by ddPCR was higher in cells cultured with Es8 than in cells cultured with RFF2 (Fig.1A). We noted that the cell number scored at day 3 was approximately 3 times higher in cells cultured with Es8 than with RFF2. There was a positive relationship between cell growth rate, CHD7 expression level, and differentiation potential when H9 cells were cultured on VTN-Ncoated dishes and passaged in a single-cell suspension.

The differentiation potential of cells in culture can be altered by culture medium. (A) H9 cells cultured with Essential 8 (Es8) medium on vitronectin-N (VTN)coated dishes were transferred to RFF2 medium, cultured for 15days (3days/passage5 passages), transferred again to Es8 medium, cultured 24days (3days/passage8 passages), and then transferred again to RFF2 medium. Photos of cells in designated culture conditions, with the cell number scored at day 3 after seeding 1.0105 cells (left panels); flow cytometric analysis of CHD7, CHD7 copy numbers from 5ng total RNA at day 3 (middle panels); and photographs of EBs formed by day 14 from cells in each culture condition and numbers of EBs formed (right panels). The results are representative of three independent experiments. (B) H9 cells were cultured either with Es8 or RFF2 on VTN-Ncoated dishes. The loci of copy number variants (CNVs) detected when cells were cultured with Es8 medium (left panels) or RFF2 medium (right panels) are shown. CHD7 expression was determined by flow cytometry (mean values are shown), and CHD7 copy numbers were determined by digital droplet PCR in cells cultured with Es8 or RFF2 medium.

We next explored the mechanisms through which cells had altered CHD7 expression levels and the ability to form EBs by simply changing the culture medium. There were at least two possible explanations for this mechanism. First, cells in culture might exhibit alterations in both CHD7 expression and the resultant differentiation potential because of signals initiated and mediated by certain factors in the medium. Alternatively, CHD7 expression levels might be genetically and epigenetically predetermined in individual cells and might not be regulated or changed by signals triggered by factors in the culture medium. In the latter case, CHD7 expression levels in cultured cells might change if different dominant cell populations were selected based on a growth advantage in a new culture medium. To evaluate these possible mechanisms, cells in the culture were marked by their CNVs so that changes in the dominant cell population could be detected by comparing CNV profiles. H9 cells cultured with Es8 medium were transferred to RFF2 medium and then were placed back in Es8 medium, and the CNV profiles of H9 cells were examined and compared. Notably, the CNV profiles of cells cultured with Es8 medium included CNVs at loci 4q22.1, 8q23.1, 16p11.2, and Xq26.1, whereas cells cultured with RFF2 medium had CNVs at none of these loci. Additionally, cells cultured with RFF2 medium contained CNVs at the specific locus 14q32.33, and these CNVs were not detected in cells cultured with Es8 medium, indicating that the cell population cultured with Es8 medium was different from that cultured with RFF2 medium (Fig.1B). This observation led us to explore the mechanisms through which certain cell populations could be selected to expand under specific culture conditions.

Next, we explored the impact of cell culture medium on the metabolic systems of cultured cells. The major metabolic pathway used by PSCs and cancer cells is the glycolytic pathway7, which is coupled with suppression of mitochondrial activity, as reflected by a low mitochondrial membrane potential (M) and reduced ROS in the mitochondria8,9. We found that the majority of cells cultured with Es8 medium did not show marked ROX staining, which was used to detect ROS produced by mitochondrial activity; the exception was that cells along the rims of colonies did show ROX staining. Furthermore, JC-1 assays showed a suppression of mitochondrial membrane voltage, suggesting that there was no marked mitochondrial activity by day 3 of culture (Fig.2A). In contrast, cells cultured with RFF2 showed marked ROX staining in most cells and an activated mitochondrial membrane potential by the JC-1 assays, suggesting activated mitochondrial function in cells cultured with RFF2 (Fig.2A). RFF2 medium contained high concentrations (approximately 23mg/mL) of protein and various amino acids in addition to moderately high glucose (2.52g/L), which could support mitochondrial function. However, Es8 medium contained high glucose (3.1g/L) and a limited amount of amino acids. Thus, Es8 medium could support the glycolytic pathway and at the same time limit the activation of mitochondrial function. The suppressed mitochondrial membrane voltage of cells cultured with Es8 medium supported this idea. There was a reciprocal relationship between the expression of CHD7 and mitochondrial function when cells were maintained in an undifferentiated state (Fig.2A). Metabolic analysis showed that the RFF2 culture medium contained malate and citrate as a result of activation of the tricarboxylic acid cycle in cells, whereas the Es8 culture medium did not (Fig.2B), consistent with the above argument. Furthermore, 2-aminoadipic acid (2-AAA) was detected in the RFF2 medium but not in the Es8 medium (Fig.2B), indicating that the kynurenine catabolic pathway, which leads to loss of an undifferentiated state and initiation of ectoderm differentiation6, was activated in cells cultured with RFF2. This observation suggested that some cells cultured with RFF2 exhibited activated mitochondrial function and underwent spontaneous differentiation, but could not be maintained in RFF2 as this medium lacked the factors necessary to support differentiated cells, and therefore these cells died. Thus, only undifferentiated cells with mitochondrial activation below the permissible level not to undergo differentiation could be cultured and maintained with the RFF2 medium. A positive correlation between the activation of mitochondrial membrane voltage and the initiation of differentiation, as suggested by the secretion of 2-AAA, was observed during the culture of cells with RFF2. This observation was supported by additional experiments; namely, H9 cells cultured with Es6 medium depleted of basic fibroblast growth factor and transforming growth factor 1 compared with Es8 medium showed both an initiation of ectodermal differentiation, as demonstrated by gene expression profiling using RT-qPCR (Fig.2C, Fig. S1), and an elevated mitochondrial membrane voltage (Fig.2A,C). Thus, there is evidence that the activation of mitochondrial function is coupled with the initiation of differentiation processes. Next, we examined the impact of elevated CHD7 expression levels and the induction of spontaneous differentiation by introducing mCHD7 into undifferentiated cells.

Activation of mitochondrial function is coupled with differentiation. (A) Morphology, CellROX (ROX) immunostaining, CHD7 copy numbers, and mitochondrial membrane voltage (JC-1 assays) in cells cultured with Es8 medium on VTN-Ncoated dishes (Es8/VTN) for 3days (left panels) or with RFF2 medium on VTN-Ncoated dishes (RFF2/VTN) for 3days (right panels) are shown. Mitochondrial membrane voltage was assessed by subtracting baseline electrons (after depolarization) from total electrons (red circle). The percentage of each fraction in the scatter plot of JC-1 assays is shown. (B) H9 cells were cultured with Es8 or RFF2 medium, and culture medium was collected and replaced with fresh medium every day for 3days. 2-Aminoadipic acid (2-AAA), malate, and citrate levels in culture medium were measured using LCMS/MS. The measured values were standardized as the mean area ratio/cell/h for 3days. The average values (n=3) with error bars (SD) are shown in the bar graphs. The results of three independent experiments are shown. (C) Morphology, ROX staining, mitochondrial membrane voltage (JC-1 assays; red circle), and gene expression profiles (RT-qPCR score card panels) of H9 cells cultured with Es8 medium on VTN-Ncoated dishes on day 5 (left panel: starting material for differentiation by Es6 medium) and Es6 medium on VTN-Ncoated dishes on day 5 are shown (right panel). The interpretation of gene expression levels by RT-qPCR is shown in the attached table. The results of three independent experiments are shown.

There was a positive correlation between the level of CHD7 expression in undifferentiated cells and the differentiation potential manifested by the number of EBs formed in the EB formation assay (Fig.1A). Interestingly, mCHD7 induced differentiation of the three germ layers simultaneously, as determined by RT-qPCR in cells cultured with both Es8 and RFF2 media (Fig.3A, Fig. S2), suggesting a positive role of CHD7 in both endodermal and mesodermal differentiation processes as well as in ectodermal development. Furthermore, this suggested that there is an upper permissible level of CHD7 being in an undifferentiated state. Es8 and RFF2 media are designed to support the proliferation of undifferentiated cells, not differentiated cells, and cells that forced to differentiate following the introduction of mCHD7, could not be maintained in these culture media. Consequently, the number of cells to form EBs was markedly reduced after introduction of mCHD7 (Fig.3A). Moreover, the introduction of siCHD7 reduced the differentiation potential of cells cultured with Es8, as reflected by the marked reduction in the number of EBs formed (Fig.3A). The introduction of siCHD7 to cells cultured with RFF2 further reduced the level of CHD7 and naturally led to no or few EBs being generated. These results provided evidence for the observation in Fig.1A, demonstrating that the differentiation potential of undifferentiated cells correlated with CHD7 expression.

CHD7 expression affected the differentiation potential and growth of undifferentiated cells. (A) H9 cells cultured with Es8 on VTN-Ncoated dishes (Es8/VTN, left panels) or with RFF2 on VTN-Ncoated dishes (RFF2/VTN, right panels) were transfected with mock (control), mCHD7, or siCHD7. The morphology, CHD7 copy numbers, gene expression profiles (RT-qPCR), EB morphology, and EB numbers formed at day 14 under different culture conditions are shown. The representative results of three independent experiments are shown. (B) CHD7 expression in H9 cells determined by flow cytometry after cells were transferred from RFF2 to Es8 on VTN-Ncoated dishes at passage 0 (P0), P5, and P7. Cells were cultured for 3days between passages. (C) Fold increase of H9 cells after 48h (upper panel) and CHD7 expression, as determined by RT-qPCR, after transfection of H9 cells with various doses of siCHD7 (lower panel). The average values (n=3) with error bars (SD) are shown in the bar graphs. Representative data from three independent experiments are shown.

It is interesting to note that both the increased expression of mCHD7 and the activation of mitochondrial function induced differentiation. Therefore, there must be a reciprocal relationship between these events in cells in an undifferentiated state. In other words, cells with activated mitochondrial function need to express a limited level of CHD7 to grow in an undifferentiated state at the expense of having a reduced differentiation potential, whereas cells with suppressed mitochondrial function could have relatively high CHD7 levels, enabling these undifferentiated cells to retain differentiation potential. The level of CHD7 that can ensure the differentiation potential of cells varied across cell lines and culture methods, therefore we cannot determine a universal cutoff value for every cell line. However, H9 cells with a CHD7 copy number of less than 2000 copies/5ng total RNA showed a limited differentiation potential when cultured on VTN-Ncoated dishes (Figs. 1B, 2A, 3A).

In the previous sections, we have shown (1) the introduction of mCHD7 induced spontaneous differentiation (Fig.3A), (2) the differentiation process was coupled with the activation of mitochondrial function (Fig.2C), and (3) there was a reciprocal relationship between the CHD7 expression level and the degree of mitochondrial function in undifferentiated cells (Fig.2A). The question is how the CHD7 expression and the degree of mitochondrial function corelated each other. We showed culture medium selected a cell population to grow (Fig.1B), and the activation of mitochondria of cells in culture is directly affected by the formula of culture medium (Fig.2A). While, we could not demonstrate the relationship between formula of the medium and the expression of CHD7, rather the CHD7 expression level in cells as assessed by flow cytometry showed a broad coefficient of variation (CV) just after the culture medium was changed from RFF2 to Es8 (Fig.3B, P0). Then, the level of CHD7 expression came to converge at the highest level during the culture (Fig.3B, P5 and P7). This result suggests that cells with a higher CHD7 expression have a growth advantage and become dominant during the culture. This presumption was manifested by the CHD7 knockdown experiment using siCHD7. This experiment indicated that the level of CHD7 was positively correlated with cell proliferation potential (Fig.3C) and cells with a higher CHD7 expression became dominant due to a growth advantage after a couple of passages. This would explain the observation that the expression of CHD7 reached its highest level during the late passages, as shown in Fig.3B (P7).

In addition to the differentiation potential, the retention of self-renewal potential is a key feature of PSCs. PSCs require cell-to-cell contact to grow and, therefore, PSCs need to form colonies. For the clinical application of PSCs, we must focus on an animal-free cell culture system. Therefore, synthetic ECM was used as the dish-coating material based on regulatory considerations. However, cells on the rims of the 2-dimensional (2-D) colonies lack the signals triggered by cell-to-cell contact at one open end, which is in sharp contrast with the majority of cells located in the middle of the colony that are surrounded by other cells along their cell membrane without interruption. Cells along the rim of the colony have an uneven distribution of molecules and ion flux related to the cell-to-cell contact-mediated signals and undergo uneven segregation in mitosis. This, then, results in a break of the self-renewal state where two identical daughter cells are generated from a mother cell, triggering spontaneous differentiation10,11,12. Indeed, cells on the rims of the colonies were positively stained with anti-superoxide dismutase 2 (SOD2) antibodies (Fig.4A). SOD2 is an enzyme that belongs to the Fe/Mn superoxide dismutase family, which scavenges excess ROS generated as a result of mitochondrial activation. SOD2 gene expression in H9 cells in the culture showed that these cells committed ectoderm and mesoderm differentiation (Fig.4A). Consequently, the population of undifferentiated cells would decrease if the spontaneously differentiated cells were not properly removed from the culture. Notably, the percentage of SOD2-positive cells (4.9%) on day 5 of culture with Es8/L511 was reduced after cells were seeded in single-cell suspensions on VTN-N(0.9%), L521-(2.6%), or L511-(2.8%) coated dishes after 30h (Fig.4B). This suggests that the ability of cells to adhere to the ECM was reduced in differentiated cells compared with undifferentiated cells, and the cell-binding ability of L511 or L521 for differentiated cells was higher than that of VTN-N. Gene expression profiles showed that cells cultured on L511 or L521 were committed to ectoderm and mesoderm differentiation (Fig.4B). Thus, by exploiting the reduced cell adhesion properties of differentiated cells and the less potent cell-binding properties of VTN-N, differentiated cells could be effectively eliminated from the culture at a single-cell level by seeding cells in a single-cell suspension at each passage.

The removal of differentiated cells by seeding on a less adhesive material. (A) H9 cells cultured with Es8 on L511-coated dishes for 5days were stained with anti-SOD2 antibodies (upper left panel), and SOD2-positive (red dots) and SOD2-negative (black dots) cells were sorted (upper right panel) to examine the ectodermal or mesodermal gene expression patterns of each population by RT-qPCR (bottom panel). (B) H9 cells cultured with the conditions described in panel A (total 2.1106 cells, 4.9% SOD2-positive cells) were collected and 5.0104 cells from them were seeded as single-cell suspensions either on L511-, L521-, or VTN-Ncoated dishes and cultured for 30h with Es8. The total cell numbers harvested and the percentages of SOD2-positive cells under different culture conditions are shown. The ectodermal or mesodermal gene expression levels of cells cultured under relevant conditions as determined by RT-qPCR are shown in the lower bar graph. The interpretation of gene expression levels determined by RT-qPCR is shown in the attached table. Representative results from three independent experiments are shown.

In previous sections, we showed data using ESC H9 cells as the standard control PSC clone to avoid possible arguments about iPSC clones having diverse genetic and epigenetic backgrounds. Therefore, there is a strong need to standardize iPSC clones to develop iPSC-based cell therapy. In the previous section, we showed that the differentiation potential of even ESC H9 cells, which have relatively homogenous genetic and epigenetic profiles, could be altered by culture medium (Fig.1) and there is a possibility that we can improve the differentiation potential by optimizing culture conditions. Optimized culture conditions may include the selection of an appropriate culture medium that supports the glycolytic pathway, the seeding of cells as single-cell suspensions during passaging, and the culture of cells on an ECM substrate with a relatively weak cell-binding capacity, such as VTN-N, to minimize the inclusion of differentiated cells in undifferentiated cell cultures and to maintain the self-renewal population for the expansion of cell clones. To verify that culture conditions improved the differentiation potential of established iPSC clones, we cultured the iPSC clones 253G113, 201B75, PFX#9, and SHh#24 and the ESC clone H9 (control) with iPSC medium4 or mTeSR1 and maintained them on feeder cells or on L511- or L521-coated dishes that were transferred to Es8 medium, cultured on VTN-Ncoated dishes, and passaged as single-cell suspensions. The CHD7 expression profile by flow cytometry and the number of EBs formed before and after the transition to Es8/VTN-N culture were measured. Notably, increased levels of CHD7 expression by flow cytometry before and after recloning (Fig.5A) may be a good index for an improved differentiation potential of cells, as manifested by an increase in the number of EBs formed (Fig.5B). The convergence of CHD7 expression by flow cytometry (Fig.5A) may represent a decreased variance in the differentiation potential among iPSCs in a given culture.

Recloning of cells with differentiation potential based on culture conditions. (A) iPSC clones (201B7, PFX#9, SHh#2, or 253G1 cells) or ESC clones (H9 cells) were cultured either on feeder cells or on L511- or L521-coated dishes with iPSC or mTeSR1 medium. Clones were then transferred to Es8 medium and cultured on VTN-Ncoated dishes. The mean and convergence of CHD7 expression of cell clones was determined by flow cytometry before (gray histogram) and after (red histogram) changing culture conditions. Representative results from three independent experiments are shown. (B) Flow cytometric analysis of cell clones for the mean and coefficient of variation (CV) measured before (circle) and after (square) changing culture conditions are plotted on the left panel and the differentiation potential before and after changing the culture conditions was assessed by the number of EBs formed and is shown on the right panel. The data set shown in (B) was generated from the same samples shown in (A).

Although we cannot alter the genetic background of individual cells by changing culture conditions, a cell population with a higher differentiation potential could be selected to grow, or be recloned, by culture conditions that support the glycolytic pathway and by eliminating spontaneously differentiated cells by seeding on an ECM with a less potent cell-binding capability, thus exploiting their reduced adhesive properties. This could also reduce the variability in differentiation potential, especially among iPSC clones.

See the article here:
Improving the differentiation potential of pluripotent stem cells by optimizing culture conditions | Scientific Reports - Nature.com