ALPS enabled photographing each cell before RNA extraction

A iPSC-RPE in vitro shows diverse degree of pigmentation among the cells. B Each cell was picked by ALPS and photographed under a light-microscope. The intensities of red, green and blue were measured. As there was no difference among the three intensities, the intensity of blue wavelength was used to represent the color intensity of each cell (see also Figure 1 –figure supplement 1).

Single cell transcriptome analysis of the cells picked with ALPS.

Two-dimension plotting of the gene expression profile of each cell colored by the cell-types (A), clusters defined by gene expression profiles (B) or color intensities (C), and a violin plot quantitatively showing color intensities of the cells in each cluster (D). A iPSC-RPE cells formed a different cluster from Lonza-RPE. B iPSC-RPE cells were a heterogeneous population consisted of 4 sub-clusters. C, D Either very black or very white cells did not localize to a specific sub-cluster of iPSC-RPE.

Correlation analysis between the expression level of each gene and the color of the iPSC-RPE.

List of the genes with positive (A) and negative (B) correlation to color intensities of in vitro cultured iPSC-RPE cells are shown. Even for the gene with the highest correlation to the color, the correlation value was 0.565. Notably, the melanin-synthesis enzymes TYRP1 and tyrosinase (TYR) both showed only weak correlations (correlation value = 0.458 and 0.434, respectively).

GSEA of iPSC-RPE cells elucidated gene sets correlated with color intensities.

GSEA revealed 15 pathways that had high correlation with the intensities of the color of iPSC-RPE. Among them, there were lysosome-related and complement-related pathways.

The cells picked by ALPS and subjected to single cell transcriptome analysis

Since the red-, green- and blue-intensities were linearly correlated among the iPSC-RPE cells, the intensity of blue wavelength was used to represent the color intensity of each cell in the following analyses.

The expression levels of the genes critical for quality management of iPSC-RPE cells for clinical use are shown by violin plots of each cluster defined in Figure 2B.

Gene enrichment analysis of the 2 major clusters of iPSC-RPE (cluster 0 and 1) defined in Figure 2B did not elucidate qualitatively significant difference between the two.

A minor cluster of iPSC-RPE defined in Figure 2B, which consisted of 32 cells (upper right), showed a stem cell-like feature with higher expression of proliferation marker MIK67 (upper left) and active cell cycle by gene enrichment analysis (lower panel), but was seemingly not a residual of undifferentiated iPSCs as shown by little expression of pluripotency marker LIN28A (upper middle).

Hypothetical model for the interpretation of weak correlation between the expression of TYR and the degree of pigmentation in iPSC-RPE cells.

When cultured in vitro, RPEs become more pigmented at higher cell-density, and return less pigmented when re-plated to lower cell density. The material of the color in RPE is melanin, which is synthesized by the enzymes such as TYRP1 and TYR, although the expressions of TYRP1 and TYR were not strongly correlated with the color of RPE (Figure 3A).

The interpretation for this could be:

i) When RPEs contact each other, the nature as protective cell makes them to synthesize melanin, with upregulation of TYRP1 and TYR.

ii) With sufficient amount of melanin, TYRP1 and TYP become downregulated.

iii) Melanin will be divided among the daughter cells or secreted, which makes the cells less pigmented.

iv) When the amount of melanin becomes lower, TYRP1 and TYR become upregulated again.

This model interprets the colors of iPSC-RPE cells are representing not a stable characteristic but a temporal condition of each cell, which makes it reasonable for the iPSC-RPE cells not having a specific gene expression profile that reflects on the degree of pigmentation (as shown in Figure 2).

Genes of lysosome-related pathway showed correlation with the color of iPSC-RPE cells.

Genes of complement-related pathway showed moderate correlation with the color of iPSC-RPE cells.

Genes related to retinol metabolism had weak correlation with the color of iPSC-RPE cells.

Genes related to melanogenesis had weak correlation with the color of iPSC-RPE cells.