(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Global genome decompaction leads to stochastic activation of gene expression as a first step toward fate commitment in human hematopoietic cells [1]

['Romuald Parmentier', 'École Pratique Des Hautes Études', 'Psl Research University', 'St-Antoine Research Center', 'Inserm', 'Ap-Hp', 'Siric Curamus', 'Paris', 'Laëtitia Racine', 'Alice Moussy']

Date: 2022-11 

When human cord blood–derived CD34+ cells are induced to differentiate, they undergo rapid and dynamic morphological and molecular transformations that are critical for fate commitment. In particular, the cells pass through a transitory phase known as “multilineage-primed” state. These cells are characterized by a mixed gene expression profile, different in each cell, with the coexpression of many genes characteristic for concurrent cell lineages. The aim of our study is to understand the mechanisms of the establishment and the exit from this transitory state. We investigated this issue using single-cell RNA sequencing and ATAC-seq. Two phases were detected. The first phase is a rapid and global chromatin decompaction that makes most of the gene promoters in the genome accessible for transcription. It results 24 h later in enhanced and pervasive transcription of the genome leading to the concomitant increase in the cell-to-cell variability of transcriptional profiles. The second phase is the exit from the multilineage-primed phase marked by a slow chromatin closure and a subsequent overall down-regulation of gene transcription. This process is selective and results in the emergence of coherent expression profiles corresponding to distinct cell subpopulations. The typical time scale of these events spans 48 to 72 h. These observations suggest that the nonspecificity of genome decompaction is the condition for the generation of a highly variable multilineage expression profile. The nonspecific phase is followed by specific regulatory actions that stabilize and maintain the activity of key genes, while the rest of the genome becomes repressed again by the chromatin recompaction. Thus, the initiation of differentiation is reminiscent of a constrained optimization process that associates the spontaneous generation of gene expression diversity to subsequent regulatory actions that maintain the activity of some genes, while the rest of the genome sinks back to the repressive closed chromatin state.

To do this, we correlated the dynamic changes of the transcription profiles determined by single-cell RNA sequencing (scRNA-seq) at different time points during the 96-h period following their stimulation with the chromatin profiles during the same period as determined by bulk and single-cell ATAC sequencing (scATAC-seq). The data revealed that a rapid and global nonspecific chromatin decompaction precedes the global up-regulation of gene expression by an unusually long lag of 24 h. Specific regulatory actions may come at the next stage to stabilize and maintain the activity of a subset of genes that allow the cell to better thrive in the changing environment. The remaining part of the genome becomes repressed again as a consequence of the chromatin recompaction.

Frequently considered as a paradigm of cell differentiation in general, hematopoietic cells are widely used as experimental model to study fate commitment. The differentiation of the hematopoietic cells is frequently represented as a series of binary fate decisions under the action of key instructive factors inducing specific changes in the cell and leading to progressively decreasing capacity of self-renewal, proliferation and lineage potential [ 17 , 18 ]. Such a strict hierarchical process must imply tight regulation of the expression of key genes. A number of genes that play a key role in the process and the core gene regulatory network (GRN) of hematopoiesis have been identified [ 19 , 20 ]. The early ideas [ 3 ] about gene regulation acting linearly during differentiation evolved toward the dynamic system view and the conceptual framework of the complex dynamical systems is now applied to the study of the hematopoietic differentiation also [ 21 ]. The concept of stochasticity has also appeared early in the study of hematopoietic differentiation, thanks to the pioneering work by Till [ 22 ]. Single-cell gene expression studies added a new layer to the general picture. They demonstrated that soon after their stimulation for differentiation, multipotent CD34+ cells go through a phase of disordered gene expression called “multilineage-primed” phase characterized by concomitant expression of genes typical for alternative lineages [ 23 – 26 ]. More recent studies confirmed that hematopoietic stem cells (HSCs) gradually acquire lineage characteristics along multiple directions without passing through discrete hierarchically organized and demarcated progenitor populations [ 27 ] and that lineage-restricted cells emerge directly from a “continuum of low-primed undifferentiated hematopoietic stem and progenitor cells” [ 27 ]. It has been shown that this phase is accompanied by instabilities and fluctuations of the cell transcriptome, morphology, and dynamic cell behavior essentially during the first 2 to 3 cell cycles [ 26 , 28 ]. How this quasi-random gene expression pattern is generated remains unclear. Indeed, it is hardly possible to imagine that a different strictly regulated hierarchical processes targeting specific genes could generate a unique mixed gene expression pattern in each cell and subsequently make them to converge to the same defined profile. In order to determine how such a response is produced, we investigated the early chromatin and transcriptional changes during the short initial period of time when the critical fate decision is initiated in CD34+ cells.

Understanding the process of cell differentiation that generates functionally and morphologically different cells with distinct gene expression profiles is one of the major challenges in biology. The way cell differentiation is conceptualized has changed during the last years [ 1 ]. Initially, cell differentiation was considered as a predetermined sequence of molecular and cellular events programmed by the genome. In this classical cause-and-effect paradigm, the new phenotype is induced by the action of specific signals that activate specific genes resulting in a linear deterministic process of cell fate determination and phenotypic differentiation [ 2 , 3 ]. The idea of linear causation has been progressively undermined by the large amount of data provided by the various “omics” approaches that raised the urgent need for generalizable principles [ 4 ]. The introduction of the conceptual arsenal of the dynamical complex system’s field can potentially satisfy this need [ 4 ] and provide an example how mathematics and physics can stimulate thinking in biology [ 5 ]. It is now generally accepted that molecular interactions within the cell, including gene transcription and translation, are fundamentally stochastic [ 6 , 7 ]. First considered as a simple “noise” perturbing the neatly functioning of the deterministic regulatory pathways, now it is becoming likely that the molecular variations are part of the system and play an essential biological role [ 8 ]. This view is further reinforced by the demonstration that molecular fluctuations are not only ubiquitous, but the cell is unable to suppress them by specifically dedicated mechanisms [ 9 ]. The conceptual framework of the complex dynamical systems allows the incorporation of molecular stochasticity and the resulting nonlinear dynamics in the explanatory scheme [ 10 ]. Importantly, the fundamental role of molecular stochasticity in cell differentiation was conjectured long time ago by Kupiec [ 11 – 13 ]. He proposed that cell differentiation can be viewed as a process of selective stabilization of gene expression profiles generated by spontaneous stochastic variation of gene transcription. The initial theory has been further developed [ 14 , 15 ] and now supported by a large body of experimental observations [ 16 ].

Results

Our experimental strategy (Fig 1A) was as follows. First, we evaluated the progression of the human CD34+ cord blood cells toward defined fates after cytokine stimulation using scRNA-seq. This approach allowed us to assess quantitatively the phenotypic heterogeneity and identify subpopulations at each time point within the time window defined by our previous study [26]. Then, we investigated the genome-scale changes of the chromatin structure using whole-cell population-level ATAC-seq (referred to as bulk ATAC-seq). scATAC-seq was used at a critical time point to confirm the conclusions. Finally, we analyzed the data to determine how global chromatin changes are related to global transcription changes. As the hematopoietic system is a very well-studied experimental model and the key individual elements are well known, we focused our analysis on the less known global tendencies rather that individual genes and chromatin elements.

PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Experimental strategy and global gene expression dynamics. (A) CD34+ cells were isolated from human cord blood and cultured in serum-free medium with early acting cytokines. scRNA-seq was used to analyze transcription at 5 h, 24 h, 48 h, 72 h, and 96 h. Concomitantly, at 0 h, 5 h, 24 h, and 48 h, 5,000 living cells were collected for “bulk” ATAC-seq analysis of the DNA accessibility. The 24-h time point was analyzed by scATAC-seq also. (B) Number of detected genes per cell with scRNA-seq. Two donors were analyzed separately, both showed similar dynamics. The exact numbers are indicated in the Results section. Note the rapid increase in the number of genes expressed per cell between 5 h and 24 h and the slow decrease after a plateau between 24 h and 72 h. (C) WGCNA reveals groups of genes with similar dynamic patterns in the average mRNA expression in donor1 and donor2. Details about WGNCA are given in Materials and methods. Note that category 1 reproduces the best overall dynamic pattern observed for genes showing detectable expression in single cells in (B). Category 1 = 5,194 genes (donor1) and 5,518 genes (donor2), category 2 = 3,977 genes (donor1) and 2,602 (donor2), category 3 = 1,089 genes (donor1) and 609 genes (donor2). (Numerical values available in scRNA-seq repository GSE156734: “GSE156734_Spread_MARSseq_data_all_filters_20200728.csv”) scATAC-seq, single-cell ATAC sequencing; scRNA-seq, single-cell RNA sequencing; WGCNA, weighted correlation network analysis. https://doi.org/10.1371/journal.pbio.3001849.g001

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001849

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/