PAPER Reconstructing blood stem cell regulatory network COLLOQUIUM models from single-cell molecular profiles Fiona K. Hameya,1, Sonia Nestorowaa,1, Sarah J. Kinstona, David G. Kenta, Nicola K. Wilsona,2, and Berthold Gottgens¨ a,2 aDepartment of Haematology, Wellcome Trust–Medical Research Council Cambridge Stem Cell Institute, University of Cambridge, Cambridge Institute for Medical Research, Cambridge CB2 0XY, United Kingdom Edited by Ellen V. Rothenberg, California Institute of Technology, Pasadena, CA, and accepted by Editorial Board Member Neil H. Shubin January 18, 2017 (received for review July 25, 2016) Adult blood contains a mixture of mature cell types, each with Cellular decision-making is heavily influenced by transcription specialized functions. Single hematopoietic stem cells (HSCs) have factors acting as components of transcriptional regulatory net- been functionally shown to generate all mature cell types for the works (5). Identifying true transcriptional interactions remains lifetime of the organism. Differentiation of HSCs toward alterna- an enormous challenge, at least in part because the experimental tive lineages must be balanced at the population level by the fate validation of functional relationships between regulator and tar- decisions made by individual cells. Transcription factors play a key get genes does not readily scale to a system-wide approach. Com- role in regulating these decisions and operate within organized putational network inference methods have therefore become regulatory programs that can be modeled as transcriptional reg- widely used to predict these functional relationships. However, ulatory networks. As dysregulation of single HSC fate decisions the application of most network reconstruction methods has is linked to fatal malignancies such as leukemia, it is important been restricted to expression data measured on whole popu- to understand how these decisions are controlled on a cell-by- lations of cells. The power of single-cell data has previously cell basis. Here we developed and applied a network inference been recognized in discovering simple regulatory relationships method, exploiting the ability to infer dynamic information from in blood systems (6, 7). More recently, approaches have emerged single-cell snapshot expression data based on expression pro- basing network reconstruction on single-cell data (8–11). files of 48 genes in 2,167 blood stem and progenitor cells. This As well as identifying regulatory relationships, network infer- BIOPHYSICS AND approach allowed us to infer transcriptional regulatory network ence methods can also allow in silico simulation of gene expres- COMPUTATIONAL BIOLOGY models that recapitulated differentiation of HSCs into progeni- sion. Computational modeling of gene regulatory networks has tor cell types, focusing on trajectories toward megakaryocyte– been applied to a variety of systems, in particular, develop- erythrocyte progenitors and lymphoid-primed multipotent pro- mental gene networks, providing new understanding about gene genitors. By comparing these two models, we identified and regulatory processes. Several studies have used a mathemati- subsequently experimentally validated a difference in the regu- cal approach to study the role of gap genes in patterning the lation of nuclear factor, erythroid 2 (Nfe2) and core-binding fac- Drosophila embryo (12) where constructing gene circuit models tor, runt domain, alpha subunit 2, translocated to, 3 homolog improved understanding of the interactions present in the gap (Cbfa2t3h) by the transcription factor Gata2. Our approach con- gene network (13). In the developing sea urchin embryo, Peter et firms known aspects of hematopoiesis, provides hypotheses al. (14) used extensive experimental evidence of transcriptional about regulation of HSC differentiation, and is widely applica- regulation to create a computational network model that reca- ble to other hierarchical biological systems to uncover regulatory pitulated known patterning behavior, and was capable of making relationships. predictions by simulating perturbations. To address the question of how HSPC fate decisions are con- gene regulatory networks j hematopoiesis j single cell j Boolean trolled, we have used single-cell gene expression profiling to infer network j stem progenitor cells transcription factor regulatory relationships. To provide a large pool of cells for this investigation, qRT-PCR data we previously published (2) were extended to obtain comprehensive coverage hroughout adult life, the mammalian blood system is main- Ttained by hematopoietic stem cells (HSCs). HSCs are able to differentiate into all mature blood cell types, as well as self-renew to maintain the blood stem cell pool. Although alternative fate This paper results from the Arthur M. Sackler Colloquium of the National Academy of choices can be made by individual cells, the output toward differ- Sciences, ”Gene Regulatory Networks and Network Models in Development and Evolu- tion,” held April 12–14, 2016, at the Arnold and Mabel Beckman Center of the National ent mature cell types is balanced and regulated at the population Academies of Sciences and Engineering in Irvine, CA. The complete program and video level. An imbalance of fate choices leads to biased production recordings of most presentations are available on the NAS website at www.nasonline.org/ of cell types, which can result in severe blood disorders such as Gene Regulatory Networks. acute myeloid leukemia. It is therefore important to understand Author contributions: N.K.W. and B.G. designed research; F.K.H., S.N., S.J.K., D.G.K., and how fate decisions are controlled during blood cell development. N.K.W. performed research; F.K.H. contributed new reagents/analytic tools; F.K.H., S.N., Hematopoiesis is an extensively studied and well-characterized and N.K.W. analyzed data; and F.K.H., S.N., D.G.K., N.K.W., and B.G. wrote the paper. system (1), and yet it is only with the recent development of The authors declare no conflict of interest. high-throughput single-cell technologies that we are understand- This article is a PNAS Direct Submission. E.V.R. is a guest editor invited by the Editorial ing how heterogeneity within hematopoietic stem and progen- Board. itor cell (HSPC) populations is related to fate choices in the Data deposition: The ChIP data reported in this paper have been deposited in the Gene Expression Omnibus (GEO) database (accession no. GSE84328). The MEP and blood (2, 3). Unlike bulk population studies, which measure LMPP network models were deposited in BioModels and assigned the identifiers average states of expression and assume homogeneity within a MODEL1610060000 and MODEL1610060001, respectively. population, single-cell assays can resolve the molecular basis of 1F.K.H. and S.N. contributed equally to this work. cell type heterogeneity. Methods such as quantitative real-time 2To whom correspondence may be addressed. Email: [email protected] or nkw22@ PCR (qRT-PCR) and RNA sequencing (RNA-Seq) can be per- cam.ac.uk. formed in individual cells to obtain single-cell gene expression This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. profiles (4). 1073/pnas.1610609114/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1610609114 PNAS Early Edition j 1 of 8 Downloaded by guest on October 1, 2021 of the murine bone marrow HSPC compartment. Using these can capture a variety of more complex structures. The diffusion data, differentiation trajectories from HSCs to progenitor cells map method has been specifically adapted for use with single- were constructed. These were used to infer and validate regu- cell expression data (17) and has proved to be a powerful tool for latory network models, thereby gaining greater insight into the representing spatial heterogeneity in single-cell data from mouse transcriptional programs governing HSC differentiation. embryos (18), and branching differentiation dynamics for both single-cell qRT-PCR data describing embryonic blood develop- Results ment (10) and single-cell RNA sequencing (scRNA-Seq) data for Single-Cell Snapshot Measurements Capture Progression Through adult HSPCs (19). HSPC Differentiation. To study the transcriptional control of When applied to our data, diffusion map analysis using all of HSPC differentiation, we previously collected single-cell qRT- the genes analyzed by single-cell qRT-PCR demonstrated that PCR data for HSCs and progenitor cells, in which we quan- the new and old data sets integrated well (SI Appendix, Fig. S1). tified the expression levels of 48 genes in 1,626 HSPCs The location of specific HSPC populations in the diffusion using the Fluidigm Biomark system (2). This study profiled map was consistent with known lineage relationships between megakaryocyte–erythroid progenitors (MEPs), granulocyte– mature cell types and their respective precursor populations. monocyte progenitors (GMPs), lymphoid-primed multipotent Fig. 1B highlights two progenitor cell populations, MEPs and progenitors (LMPPs), common myeloid progenitors (CMPs), LMPPs, along with the so-called molecular overlap, or “MolO” HSCs with finite self-renewal (FSR-HSCs), and long-term HSCs HSCs, as identified by Wilson et al. (2). MolO cells are HSCs (LT-HSCs). However, the primary focus was to resolve hetero- with a shared transcriptional profile and increased probability geneity within four different LT-HSC populations isolated by of long-term multilineage reconstitution upon single-cell trans- fluorescence-activated cell sorting.