3.4. Transcriptome Heterogeneity Analyses of Adherent Cultures
In our single-cell transcriptome analysis of suspension cultures, no subpopulations were apparent. This finding differs from that of previous studies, which indicated a strong correlation between genome-wide gene expression and the cell cycle stage of adherent cells (Buettner et al., 2015; Santos, Wernersson, & Jensen, 2015; Sasagawa et al., 2013; Scialdone et al., 2015). Cellular heterogeneity is known to be influenced by the cellular microenvironment in adherent, two-dimensional cultures (Snijder et al., 2009). It is possible that cell cycle–dependent genome-wide gene expression is an adherent culture-specific phenomenon. In this context, the origin of cell cycle–dependent cellular heterogeneity in adherent cultures would not be the product of intrinsic factors but rather the product of the extrinsic milieu, with cellular division altering genome-wide gene expression in cells by changing the microenvironment in adherent cultures. A previous study supports this interpretation; suspended T cells did not show any cell cycle–dependent cellular transcriptome heterogeneity (Mahata et al., 2014).
To further investigate cell cycle–dependent cellular transcriptome heterogeneity, we performed single-cell SMART-Seq as previously reported (Ziegenhain et al., 2017) on 24 G1 cells and 8 G2 cells from adherent cultures. To compare the transcriptomes between cells in different phases of the cell cycle, we performed PCA of all the expression data (Figure 4). The proportion of variance of PC1 was 0.86, while that of PC2 was 0.033. The G2 cells formed a distinct cluster. We, therefore, rejected the null hypothesis that cell cycle–dependent subpopulations do not appear during adherent culture, based on the Bartlett test for the homogeneity of variances in PC1 between G1 and G2 cells (p = 0.000245). Thus, cell cycle–dependent heterogeneity in genome-wide gene expression was found to be a phenomenon specific to adherent cultures of CHO cells.