Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 7 (4), e319

Making Lineage Decisions With Biological Noise: Lessons From the Early Mouse Embryo


Making Lineage Decisions With Biological Noise: Lessons From the Early Mouse Embryo

Claire S Simon et al. Wiley Interdiscip Rev Dev Biol.


Understanding how individual cells make fate decisions that lead to the faithful formation and homeostatic maintenance of tissues is a fundamental goal of contemporary developmental and stem cell biology. Seemingly uniform populations of stem cells and multipotent progenitors display a surprising degree of heterogeneity, primarily originating from the inherent stochastic nature of molecular processes underlying gene expression. Despite this heterogeneity, lineage decisions result in tissues of a defined size and with consistent proportions of differentiated cell types. Using the early mouse embryo as a model we review recent developments that have allowed the quantification of molecular intercellular heterogeneity during cell differentiation. We first discuss the relationship between these heterogeneities and developmental cellular potential. We then review recent theoretical approaches that formalize the mechanisms underlying fate decisions in the inner cell mass of the blastocyst stage embryo. These models build on our extensive knowledge of the genetic control of fate decisions in this system and will become essential tools for a rigorous understanding of the connection between noisy molecular processes and reproducible outcomes at the multicellular level. We conclude by suggesting that cell-to-cell communication provides a mechanism to exploit and buffer intercellular variability in a self-organized process that culminates in the reproducible formation of the mature mammalian blastocyst stage embryo that is ready for implantation into the maternal uterus. This article is categorized under: Gene Expression and Transcriptional Hierarchies > Cellular Differentiation Establishment of Spatial and Temporal Patterns > Regulation of Size, Proportion, and Timing Gene Expression and Transcriptional Hierarchies > Gene Networks and Genomics Gene Expression and Transcriptional Hierarchies > Quantitative Methods and Models.

Keywords: blastocyst; heterogeneity; modeling; noise; specification.


Figure 1
Figure 1. Staging and lineage specification during mouse preimplantation development
(A) Approximate relationship between developmental time in embryonic days from fertilization and cell number in the embryo. This relationship differs slightly between mouse strains and exact conditions of husbandry. Staging by cell number is thus preferable to facilitate comparison between studies. (B) Schematic representation of the morphological and cell fate specification events that convert the single-cell totipotent zygote from fertilization at embryonic day (E) 0.5 to the ~200 cell blastocyst comprising of three spatially distinct lineages prior to implantation at E4.5. The first cell fate choice bifurcates totipotent blastomeres to outer trophectoderm (TE; green) cells and the inner cell mass (ICM; purple). The second fate choice further subdivides the ICM into primitive endoderm (PrE; blue) and pluripotent epiblast (EPI; red), which initially arise interspersed in a seemingly random pattern, prior to sorting into two distinct layers. Zygotic Genome Activation (ZGA).
Figure 2
Figure 2. Gene regulatory networks orchestrating lineage divergence in the inner cell mass (ICM)
Gene regulatory networks within cells of the ICM. (A) At the early blastocyst stage, uncommitted progenitor ICM cells initially co-express the transcription factors Nanog and Gata6, and the FGF receptor Fgfr1, with a subset of ICM cells expressing Fgf4. (B) Epiblast cells (EPI) exclusively express Nanog, and Fgf4, and primitive endoderm (PrE) cells exclusively express Gata6 and Fgfr2. EPI-derived FGF4 signals in a paracrine manner to PrE via both Fgfr1 and Fgfr2, and in an autocrine manner via Fgfr1 in EPI cells.
Figure 3
Figure 3. Properties of candidate genetic circuits executing cell fate decisions in the inner cell mass (ICM)
Cell fate decisions are coordinated by regulatory interactions between the transcription factors Nanog and Gata6. These dynamic networks can be modeled to generate quasi-potential landscapes for each genetic circuit. Stable attractor states are represented as minima on the potential state-space. These stable states correspond to ICM (purple), EPI (red) and PrE (blue). (A) In a mutual repression circuit there are two stable attractor states, EPI and PrE, whereas the ICM is unstable. The circuit results in a bistable fate choice. (B) In a mutual repression and auto-activation circuit, there is tri-stability; three stable attractor states, ICM, EPI and PrE. (C) When cells are coupled through FGF4/ERK signaling to the cell autonomous Nanog/Gata6 mutual repression auto activation circuit, cell fate choice is coordinated across the population to establish appropriate ratios of EPI:PrE. For small numbers of EPI cells, low levels of FGF4 shift the potential landscape in favour of generating more EPI. Conversely, for large numbers of EPI cells, high levels of FGF4 shift the potential landscape in favour of generating more PrE. Modified from Huang, 2013.
Figure 4
Figure 4. Evolution of cell-cell variability in different models of the Epi-vs-PrE fate decision
(A) In GRNs containing a mutual repression motif between NANOG and GATA6, small initial differences in the concentrations of the two factors will be amplified, leading to the emergence of two distinct fates. FGF/ERK signaling can be included in those models (Schroter et al., 2015), but is usually assumed to be equal between cells. (B) In the tristable model by Bessonnard et al. (2014), GATA6 and NANOG are initially expressed from low levels with similar rates, but FGF/ERK signaling is assumed to differ between cells. This drives some cells towards the NANOG-expressing Epi-state, and other cells towards the GATA6-expressing PrE-state. (C) Extension of the Bessonnard-model in De Mot et al., (2016). When considering transcriptional noise in GATA6- and NANOG expression, this can break the symmetry between cells and lead to differential signaling between cells that further locks in fates. However, in some parameter regimes this model leads to unrealistic fate switching events (not shown).
Figure 5
Figure 5. Towards a fully integrated model of ICM lineage specification
Advancement of current models of ICM lineage specification require the integration of (A) FGF/ERK dynamics. Current models are dependent on the assumption that there is heterogeneity between cells in FGF/ERK. Direct quantitation of the intra-cellular FGF/ERK signaling activity over time is required to validate this assumption, and understand how heterogeneities arise and spatially pattern the system. (B) 3D-spatial pattern. At the mid-blastocyst stage DP, EPI and PrE cells are arranged in a seemingly random pattern, and later sort into distinct layers at the late blastocyst stage. Whether the initial pattern is truly random or dependent on the identity of nearest neighbours remains to be investigated. (C) Proliferation and apoptosis. Current models are based on a fixed number of cells, however in the embryo cells proliferate over time (t1 → t2). EPI (red) cells give rise to EPI progeny, PrE (blue) cells give rise to PrE progeny, and ICM double positive (DP) progenitors (purple) may give rise to either DP, EPI or PrE progeny (boxed region). In addition, some cells undergo apoptosis (illustrated by blebbing cell). Generation of a reproducible ratio of PrE:EPI can be fine-tuned by allocating DP cells to either PrE or EPI lineage, but could also be achieved by lineage-specific modulation in the rate of proliferation and/or apoptosis and should be measured and accounted for in future models.

Similar articles

See all similar articles

Cited by 5 PubMed Central articles

Publication types

LinkOut - more resources