Species-specific developmental timing is maintained by pluripotent stem cells ex utero

Abstract

How species-specific developmental timing is controlled is largely unknown. By following human embryonic stem (ES) cell and mouse epiblast stem (EpiS) cell differentiation through detailed RNA-sequencing time courses, here we show that pluripotent stem cells closely retain in vivo species-specific developmental timing in vitro. In identical neural differentiation conditions in vitro, gene expression profiles are accelerated in mouse EpiS cells compared to human ES cells with relative rates of differentiation closely reflecting the rates of progression through the Carnegie stages in utero. Dynamic Time Warping analysis identified 3389 genes that were regulated more quickly in mouse EpiS cells and identified none that were regulated more quickly in human ES cells. Interestingly, we also find that human ES cells differentiated in teratomas maintain the same rate of differentiation observed in vitro in spite of being grown in a mouse host. These results suggest the existence of a cell autonomous, species-specific developmental clock that pluripotent stem cells maintain even out of context of an intact embryo.

Keywords: Brain development; Developmental time; Differentiation; Embryonic stem cells.

Figure 1. In vitro neural differentiation occurs more quickly in mouse EpiS cells compared to human ES cells

(A) EGFP-H1 and EGFP–mouse EpiS cells were exposed to identical differentiation conditions on day 0 and were fixed and stained with the indicated antibodies at various time points. Samples were imaged on a Nikon confocal A1R microscope (scale bars= 250 μm). Other samples were lysed at regular time intervals and subjected to RNAseq (B). Expression of gene TPMs were scaled from their minimum (0) to maximum (100) values to compare dynamic ranges between mouse and human samples. Classical gene markers of embryonic neuroectoderm (NE), forebrain, neocortex, neurogenesis, and synapse formation are shown. (C) BRN2 expression is shown as an example of a gene identified as accelerated in mouse (red) compared to human cells (black) using DTW analysis by identifying and warping similarly patterned regions (dotted lines). Global DTW was applied to all genes to identify significantly faster genes (p<0.01) in mouse compared to human cells during in vitro neural differentiation (D). (E) The 2,000 most significantly accelerated genes were screened for enriched functional GO terms using the DAVID functional annotation tool. The top 20 terms are shown, and neural cell differentiation-related terms are in bold.

Figure 2. Sample correlations across species reveal that global changes in neural gene expression occur more quickly in mouse compared to human cell differentiation

Pearson correlations of 3,061 neural genes were applied to either human to human samples (A), mouse to mouse samples (B), or human to mouse RNA-seq samples (C).

Figure 3. Neural gene expression patterns in teratomas closely mirrors those observed in vitro in a species-dependent manner

(A) EGFP-positive human ES or mouse EpiS cells were intramuscularly-injected into immunocompromised mice, and teratoma differentiation was followed by isolating EGFP-positive cells by FACS and sequencing transcriptomes. In vitro scaled gene expression trends (blue) were compared with those observed during teratoma formation for mouse (M) and human (H) cells (B–D). Representative genes of the neural tube early forebrain development (B), neurogenesis and neural precursor expansion (C), and GABAergic and Glutamatergic neuron identity and synapse function are shown (D). (E) The top 2,000 DTW-identified accelerated genes in mouse compared to human pluripotent stem cells during in vitro neural differentiation and teratoma formation were enriched for GO terms using the DAVID functional annotation tool. The top 15 neural development-related terms identified during in vitro neural differentiation (blue) were also identified as significantly enriched (p<0.001) in genes accelerated in mouse compared to human teratomas (red).

Figure 4. Correlation of mouse and human gene expression peaks identified by segmentation regression analysis closely mirror CS progression

(A) Gene expression profiles with peaks identified by segmentation regression analysis are indicated over time post differentiation for mouse (blue) and human (orange) cells, and the 10 peaks with the highest amplitudes in human samples are shown in (B) with their mouse counterparts. (C) The gene peaks from B found in human and mouse samples were plotted over time (after adding 6.5 days to mouse samples and 15 days to human samples to transpose to embryonic day equivalents), and were overlaid with in utero Carnegie stages (CS, in black). The linear trendlines and their slopes are shown in their respective colors.