Interspecies chimeric conditions affect the developmental rate of human pluripotent stem cells

Jared Brown, Chris Barry, Matthew T. Schmitz, Cara Argus, Jennifer M. Bolin, Michael P. Schwartz, Amy Van Aartsen, John W. Steill, Scott Swanson, Ron M. Stewart, James A. Thomson, Christina KendziorskiPLOS Computational Biology 2021 · DOI

TL;DR

Human pluripotent stem cells (hPSCs) differentiate much slower than mouse stem cells, hindering regenerative medicine applications. This study shows that co-culturing hPSCs with mouse stem cells significantly accelerates their neural differentiation, mimicking developmental timelines more closely. This finding suggests that species-specific developmental clocks are more plastic than previously thought and opens avenues for faster generation of mature cell types.

What Problem Does This Paper Address?

Generating mature, functional cell types from human pluripotent stem cells (hPSCs) for therapeutic use is a major challenge. hPSCs have an intrinsic "developmental clock" that dictates a slow differentiation rate, mirroring the nine-month human gestation period. This contrasts sharply with mouse stem cells, which differentiate much more rapidly, reflecting a three-week gestation. This disparity necessitates lengthy differentiation protocols for hPSCs, often resulting in immature cell characteristics that limit their clinical utility. The authors investigate whether the presence of rapidly differentiating mouse stem cells can influence and accelerate the developmental pace of hPSCs.

Methods

The core of this study involves a comprehensive RNA-sequencing (RNA-seq) time-course analysis. Human embryonic stem cells (hESCs) and mouse epiblast stem cells (mEpiS) were co-cultured under neural differentiation conditions. Different ratios of human to mouse cells were tested, including a 10% human to 90% mouse ratio (H10) and a 100% human control (H100). Samples were collected at regular intervals over six weeks.

Key analytical techniques included:

  • RNA-seq: To capture genome-wide gene expression profiles.
  • Transcriptome Alignment: A combined human-mouse transcriptome was used to accurately quantify species-specific gene expression from chimeric samples.
  • Segmented Regression (Trendy package): This algorithm was used to model gene expression dynamics over time, identifying trends, up-regulation start times, and peak expression times.
  • Gene Set Enrichment Analysis (GO terms): To identify biological processes and pathways that were significantly affected by the chimeric conditions.
  • Correlation Analysis: To compare in vitro differentiation trajectories with in vivo human fetal brain development data.
  • Principal Component Analysis (PCA): Used to assess dissimilarity between in vitro and in vivo samples.
  • Deconvolution Analysis: To estimate the proportions of different cell types present in the bulk samples over time.

A crucial validation step involved sorting cells by species after co-culture to confirm that observed effects were not due to transcript misalignment.

Datasets

The primary dataset generated for this study is a six-week RNA-seq time course of neural differentiation from:

  • Human embryonic stem cells (hESCs) alone (H100).
  • Co-cultures of hESCs with mouse epiblast stem cells (mEpiS) at various ratios, notably a 10% human:90% mouse (H10) and an intermediate 85% human:15% mouse (H85) ratio.
  • Mouse epiblast stem cells (mEpiS) alone (M100) served as a reference for mouse cell behavior.

The study also utilized publicly available datasets for comparison:

  • BrainSpan Atlas: Human fetal tissue samples from different brain regions and developmental stages (weeks 8, 9, and 12) were used to compare in vitro differentiation with in vivo development.
  • Human Protein Atlas (HPA): Tissue-specific gene expression data was used for correlation analysis.
  • CoDEx dataset: Annotated single-cell sequencing data of the developing human cortex was used for deconvolution analysis.

Key Insights and Ideas

  • Developmental Clocks are Malleable: The most striking insight is that the intrinsic species-specific developmental clock of hPSCs is not fixed and can be significantly influenced by the presence of mouse stem cells. This challenges the long-held notion that these timers are immutable.
  • Acceleration is Specific to Neural Development: The acceleration observed was not a general effect but was specifically enriched for genes involved in neurogenesis, neuron differentiation, maturation, and synapse signaling. This suggests a targeted influence on specific developmental programs.
  • Dose-Dependent Effect: The degree of acceleration was directly proportional to the proportion of mouse cells in the co-culture, indicating a dose-response relationship and suggesting cell-cell signaling as a key mechanism.
  • Bi-directional Influence: Interestingly, mouse cells also showed signs of deceleration when outnumbered by human cells, suggesting that the interspecies influence is not unidirectional.
  • In Vivo Relevance: The accelerated in vitro differentiation of hPSCs in chimeric conditions correlated with in vivo human fetal brain development earlier than control hPSC cultures, lending biological relevance to the findings.

Key Results

  • 929 genes upregulated earlier: In H10 co-cultures, 929 human genes (over 57% of all trending genes) showed earlier upregulation (at least 2 days earlier) compared to H100 controls. These genes were significantly enriched for neurogenesis and synaptic signaling pathways.
  • 535 genes peaked earlier: Similarly, 535 genes peaked earlier in H10 co-cultures, representing over 46% of all peaking genes. These were also enriched for neurodevelopmental processes.
  • Acceleration Factor: The median acceleration factor for the first 16 days of differentiation in H10 compared to H100 was 1.699, indicating a substantial speed-up. This acceleration was dose-dependent, with lower mouse cell proportions (H85) showing less acceleration (median factor of 1.376).
  • Mouse Gene Deceleration: In contrast, when human cells outnumbered mouse cells (H85 vs. M100), mouse genes showed delayed upregulation and peak times, with a median acceleration factor of 0.894 (indicating deceleration).
  • Earlier Correlation with In Vivo Data: Chimeric hPSC cultures (H10 and H85) correlated with in vivo human fetal neocortical samples earlier in their differentiation time course compared to H100 controls, with the H10 culture showing the earliest correlation.

Limitations and Future Work

  • Mechanism of Acceleration: While the study demonstrates that co-culture is sufficient to accelerate differentiation, the precise molecular mechanisms driving this acceleration are not fully elucidated. The authors suggest that cell-cell signaling is likely involved and point to candidate pathways (GPCR signaling, MAPK/ERK, PI3K/AKT) and regulators (NSRF, OCT1) for future investigation.
  • Functional Maturity: The study focused on gene expression timing. Determining the functional maturity and electrophysiological properties of the neurons generated under chimeric conditions remains an important future direction.
  • Xenogeneic Nature: The current protocol involves mouse cells, making it xenogeneic and unsuitable for direct clinical applications. However, the proof-of-concept that the human developmental clock can be accelerated is highly significant.
  • Cell Lineage Outcomes: While neural differentiation was accelerated, the authors acknowledge that some altered cell fate choices might occur due to species-specific differences in brain physiology. Monitoring cell type contributions and organ function in chimeric embryos for organogenesis is highlighted as crucial.
  • Broader Applicability: While neural differentiation was studied, it would be valuable to explore if similar acceleration effects can be achieved for other hPSC differentiation protocols.

Summary updated

BibTeX
@article{brown2021interspecies,
  author = {Jared Brown and Chris Barry and Matthew T. Schmitz and Cara Argus and Jennifer M. Bolin and Michael P. Schwartz and Amy Van Aartsen and John W. Steill and Scott Swanson and Ron M. Stewart and James A. Thomson and Christina Kendziorski},
  title = {{Interspecies chimeric conditions affect the developmental rate of human pluripotent stem cells}},
  journal = {PLOS Computational Biology},
  year = {2021},
  doi = {10.1371/journal.pcbi.1008778},
}
Figure