Cell–cell interaction networks regulate blood stem and progenitor cell fate

Daniel C Kirouac¹, Gerard J Madlambayan², Mei Yu¹, Edward A Sykes¹, Caryn Ito³ & Peter W Zandstra^1,4,5,6

Article highlights

Synopsis

Communication networks among cells, tissues, and organ systems are necessary for homeostasis in multicellular organisms. Intercellular communication networks are particularly relevant in stem cell biology, as stem cell fate decisions (self-renewal, proliferation, lineage specification) are tightly regulated based on physiological demand and responsive to external perturbations. Hematopoiesis, the process by which blood cells develop, serves as a prototype for other stem cell systems. Motivated by the fact that hematopoietic stem cell transplantation is a curative therapy for a number of hematopoietic and immunological diseases, herein we explore the behaviour of intercellular regulatory networks as tools to regulate cell fate during in vitro human blood stem cell propagation.

We have integrated a diverse range of experimental and theoretical literature to develop for predictive purposes a novel mathematical model of hematopoiesis. Our resulting model links functional cellular assays to specific model outputs, defines cell-level kinetic parameters such as cell cycle rates and self-renewal probabilities as functions of culture variables, and simulates feedback regulation using cell–cell interaction networks (Schematically depicted in Figure 1A). Computational analyses of the system dynamics indicate that non-cell autonomous parameters (cell–cell interactions) are dominant factors controlling stem cell growth. As a result of negative feedback interactions, our analysis suggest an antagonistic relationship between mature and primitive cell compartments, a finding empirically supported by numerous in vitro and in vivo observations.

Figure 1

Schematic depiction of blood stem cell development model incorporating functional assays and positive and negative feedback. (A) Stem cells (X₁) at the apex of the hematopoietic hierarchy may self-renew with probability f₁ to regenerate the stem cell compartment, or differentiate (1-f₁), giving rise to a series of increasing differentiated, and developmentally restricted progenitor populations (X_i). Transit between cell compartments is associated with mitosis (u_i). In vivo (SRC) and in vitro (LTC-IC, CFC) functional assays, and cell surface phenotype (Lin^- versus Lin⁺) can be used to quantify different cellular compartments within the hierarchy. Differentiated cell populations secrete factors that inhibit or enhance progenitor proliferation (SF1 and SF3, respectively), and undifferentiated (Lin^-) cells secrete factors that inhibit stem cell self-renewal (SF2). The inhibitory effects of SF2 on self-renewal are balanced by secretion of the self-renewal stimulator SF4 by differentiated cells. Differentiated populations are functionally lumped into those that secreted inhibitory factors (red), those that secrete stimulatory factors (green), and those that secrete factors with no effect on stem and progenitor cell growth (yellow). Phase portraits below the diagram display normalised proliferation rates (u_i/u_MAX) and self-renewal probabilities (f_i/f_MAX) as functions of model parameters varied between the constraints given in Table II (low to high values represented by blue to red as indicated). (B) Proliferation and self-renewal versus differentiation status (compartment number) for parameters n_MAX, D_GR, and D_SR left to right, respectively. (C) Proliferation versus time for parameters _D (left) and kt (right). (D) Proliferation or self-renewal versus secreted regulatory factor (SF1–4) concentrations for Hill coefficients (k1–4). Refer to Table II for parameter definitions.

Full figure and legend (806K)Figures & Tables index

To investigate the practical utility of the model, we cultured human umbilical cord blood stem cells under a range of conditions and dynamic perturbations (media exchange and mature cell depletion), and compared experimentally observed cell population outputs to model simulations. Our model simulations quantitatively recapitulate experimental results; culturing enriched progenitor populations at low initial cell densities, with frequent media exchange, and with progenitor re-selection enhances stem and progenitor expansion as a consequence of reduced inhibitory feedback signalling. Using a protein array system, we identify a limited number of secreted molecules (TGF-1, CCL4, CCL5, and CXCL8) displaying dynamic profiles consistent with predicted intercellular regulators. Additionally, we show the variability in secretion rates observed for these putative regulators is sufficient to explain the experimentally observed distribution of cell population outputs.

Computational analyses of the model revealed that system dynamics are relatively robust to changes in kinetic parameters, but highly sensitive to topological alterations of the proposed cell–cell regulatory network. We extend this concept by demonstrating that deleting a single regulatory connection (negative feedback control of self-renewal) is necessary and sufficient to reproduce the characteristic features of in vitro leukemic transformation (Figure 8).

Figure 8

Loss of responsiveness to self-renewal inhibitor SF2 alone is capable of inducing leukemic transformation. Simulated culture dynamics for UCB Lin^- cells re-plated at 10⁵ cell/ml every 7 days. Cumulative fold expansion of total cells (TNC) (A) and stem cells (SRC) (B) are shown for different parameter modifications. Experimental data for total cell output from control (black) and a transformed leukemic cell line (red) as published in Warner et al (2005) is overlaid. Error bars represent s.d.

Full figure and legend (222K)Figures & Tables index

In summary, our intercellular feedback model of blood stem cell growth predictively simulates the dynamic characteristics of both normal and malignant hematopoiesis in vitro. This model may therefore serve as a platform for further experimental interrogation of the regulatory mechanisms controlling stem cell fate in vitro and in vivo, and as a tool for the rational design of stem cell-therapy bioprocesses.

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated.