The role of specialized cell cycles during erythroid lineage development: insights from single-cell RNA sequencing

Abstract

Early erythroid progenitors known as CFU-e undergo multiple self-renewal cell cycles. The CFU-e developmental stage ends with the onset of erythroid terminal differentiation (ETD). The transition from CFU-e to ETD is a critical cell fate decision that determines erythropoietic rate. Here we review recent insights into the regulation of this transition, garnered from flow cytometric and single-cell RNA sequencing studies. We find that the CFU-e/ETD transition is a rapid S phase-dependent transcriptional switch. It takes place during an S phase that is much shorter than in preceding or subsequent cycles, as a result of globally faster replication forks. Furthermore, it is preceded by cycles in which G1 becomes gradually shorter. These dramatic cell cycle and S phase remodeling events are directly linked to regulation of the CFU-e/ETD switch. Moreover, regulators of erythropoietic rate exert their effects by modulating cell cycle duration and S phase speed. Glucocorticoids increase erythropoietic rate by inducing the CDK inhibitor p57^KIP2, which slows replication forks, inhibiting the CFU-e/ETD switch. Conversely, erythropoietin promotes induction of ETD by shortening the cycle. S phase shortening was reported during cell fate decisions in non-erythroid lineages, suggesting a fundamentally new developmental role for cell cycle speed.

Keywords: CDK inhibitors; Cell cycle; Erythropoiesis; Erythropoietic stress response; Glucocorticoids; Replication forks.

Figure 1. Principal stages of erythropoiesis

Early erythroid progenitors are defined by their colony forming potential. Erythroblasts during terminal differentiation (ETD) are defined by their morphology. Flow cytometric approaches that make use of FSC and of the cell surface markers CD71 and Ter119 permit prospective isolation of differentiation-stage-specific erythroblasts.

Figure 2. The mouse fetal liver as a developmental model for studying erythropoiesis

At mid-gestation, the FL is largely an erythroid tissue. Subsets S0 and S1 contain all of the FL CFU-e; subsets S2 to S5 contain erythroblasts in terminal differentiation. Adapted from Pop et al 2010 .

Figure 3. Onset of Epo/EpoR dependence at the transition from S0 to S1

Top panel shows mid-gestation (embryonic day 12.5) littermate embryos. The Epor^−/− embryo is pale and has a poorly developed, pale liver. Flow cytometric analysis of the Epor^−/− FL shows developmental arrest at the S0/S1 transition .

Figure 4. The CFU-e/ETD transition is an S-phase dependent transcriptional switch

The CFU-e/ETD transition coincides with the transition from S0 to S1, and is an S-phase dependent transcriptional switch. It takes place in early S phase of the last CFU-e generation . The S0/S1 transition is characterized by a short cycle lasting only 6 hours, and by a short S phase, lasting only 4 hours. The preceding cycles in S0 have an average duration of 15 hours and an S phase that is 7 hours long .

Figure 5. Spring plots of mouse bone marrow (BM) and fetal liver (FL)

Force-directed graph projections of Kit⁺ BM and FL hematopoietic single-cell transcriptomes, each represented by one dot. Proximity of dots indicates transcriptome similarity. Cells that are in color (non grey) are either multipotential progenitors (MPP) or express marker genes of mature blood lineages, as follows: E, erythroid; Ba, basophil or mast cell; Meg, megakaryocytic; Ly, lymphocytic; D, dendritic; M, monocytic; G, granulocytic.

Figure 6. Predictions of Population Balance Analysis

The left panel represents the dominant cell fate probabilities of each cell, colored as indicated (key as in Figure 5). Note that cells in branches are largely committed to one specific lineage, whereas the central area of the graph contains cells with substatial probabilites for more than one lineage. The central panel illustrates predicted differentiation ordering of cells. The right panel illustrates the predicted differentiation hierarchy, based on co-occurrence of specific cell fates in individual transcriptomes (above a threshold probabilitiy). The resuting topology is similar to the classical hierarchy, differing principally in showing differentiation to be a continuous process.

Figure 7. scRNAseq analysis of the erythroid trajectory

A The erythroid trajectory includes all cells whose erythroid fate probability increases with increasing distance from MPP. It is divided into the following segments, based on transcriptomic and cell fate analysis: MPP, multi-potential progenitors; EBMP, erythroid/basophil-mast cell/megakaryocytic progenitors; EEP, early erythroid progenitors, functionallly BFU-e; CEP, committed erythroid progenitors, functionally CFU-e; ETD, erythroid terminal differentiation. B The CFU-e to ETD transition is a rapid transcriptional switch. Upper panel: erythroid trajectory cells are arranged along a linear axis. Rows denote differentially expressed genes along the trajectory, in order of their peak expression. Lower panel: the number of genes that turn on or off at each point of the trajectory. These data show that the CFU-e/CEP stage expresses a dedicated transcriptional program, which changes sharply at the switch to ETD (dashed red line).

Figure 8. The transcriptional context of the CFU-e/ETD switch

Expression profiles of transcription factors relevant to erythroid differentiation (upper panel) and signaling molecules (middle panel) show that expression of these regulators does not explain the timing of the CFU-e/ETD switch (dashed vertical line). By contast, cell cycle phase, as determined by chatacteristic expression of cell cycle genes, shows synchronization between G1/S and S phase of the cycle and the CFU-e/ETD switch (dashed line).

Figure 9. The CFU-e/ETD switch is preceded by gradual shortening in G1 phase of the cycle

A Cell cycle genes are among the genes whose expression is most highly correlated with progression along the erythroid trajectory. Shown are examples of genes expressed in S phase, whose expression ramps up gradually, peaking at the CFU-e/ETD switch. B Flow cytometric analysis of cell cycle phase and S phase speed along the erythroid trajectory shows a gradual increase in the fraction of cells in S phase, a complementary decrease of cells in G1, and a relatively constant S phase speed, as cells progress through the trajectory. These findings suggest gradual shortening of G1 phase preceding the CFU-e/ETD switch. The speed of S phase increases abruptly at the switch to ETD (dashed red line).

Figure 10. Regulatory role of the cycle in erythroid developmental decisions

Changes in cell cycle duration throughout the erythroid trajectory. The cycle is represented vertically. S phase speed influences whether CFU-e undergo a transcriptional swtich to ETD: a long S phase inhibits, and S phase shortening promotes, the switch. Glucocorticoids and erythropoietin exert their positive regulation of erythropoietic rate in response to stress through regulation of S phase and cell cycle speeds.