Person:

Sherwood, Richard

Summary: We present a method, LineageProgram, that uses the developmental lineage relationship of observed gene expression measurements to improve the learning of developmentally relevant cellular states and expression programs. We find that incorporating lineage information allows us to significantly improve both the predictive power and interpretability of expression programs that are derived from expression measurements from in vitro differentiation experiments. The lineage tree of a differentiation experiment is a tree graph whose nodes describe all of the unique expression states in the input expression measurements, and edges describe the experimental perturbations applied to cells. Our method, LineageProgram, is based on a log-linear model with parameters that reflect changes along the lineage tree. Regularization with L1 that based methods controls the parameters in three distinct ways: the number of genes change between two cellular states, the number of unique cellular states, and the number of underlying factors responsible for changes in cell state. The model is estimated with proximal operators to quickly discover a small number of key cell states and gene sets. Comparisons with existing factorization, techniques, such as singular value decomposition and non-negative matrix factorization show that our method provides higher predictive power in held, out tests while inducing sparse and biologically relevant gene sets.

MicroRNAs play a pivotal role in cellular maintenance, proliferation, and differentiation. They have also been implicated to play a key role in disease pathogenesis, and more recently, cellular reprogramming. Certain microRNA clusters can enhance or even directly induce reprogramming, while repressing key proteins involved in microRNA processing decreases reprogramming efficiency. Although microRNAs clearly play important roles in cellular reprogramming, it remains unknown whether microRNAs are absolutely necessary. We endeavored to answer this fundamental question by attempting to reprogram Dicer-null mouse embryonic fibroblasts (MEFs) that lack almost all functional microRNAs using a defined set of transcription factors. Transduction of reprogramming factors using either lentiviral or piggyBac transposon vector into two, independently derived lines of Dicer-null MEFs failed to produce cells resembling embryonic stem cells (ESCs). However, expression of human Dicer in the Dicer-null MEFs restored their reprogramming potential. Our study demonstrates for the first time that microRNAs are indispensable for dedifferentiation reprogramming.

Here we describe Protein Interaction Quantitation (PIQ), a computational method that models the magnitude and shape of genome-wide DNase profiles to facilitate the identification of transcription factor (TF) binding sites. Through the use of machine learning techniques, PIQ identified binding sites for >700 TFs from one DNase-seq experiment with accuracy comparable to ChIP-seq for motif-associated TFs (median AUC=0.93 across 303 TFs). We applied PIQ to analyze DNase-seq data from mouse embryonic stem cells differentiating into pre-pancreatic and intestinal endoderm. We identified (n=120) and experimentally validated eight ‘pioneer’ TF families that dynamically open chromatin, enabling other TFs to bind to adjacent DNA. Four pioneer TF families only open chromatin in one direction from their motifs. Furthermore, we identified a class of ‘settler’ TFs whose genomic binding is principally governed by proximity to open chromatin. Our results support a model of hierarchical TF binding in which directional and non-directional pioneer activity shapes the chromatin landscape for population by settler TFs.

Wnt signaling has been implicated in many developmental processes, but its role in early endoderm development is not well understood. Wnt signaling is active in posterior endoderm as early as E7.5. Genetic and chemical activation show that the Wnt pathway acts directly on endoderm to induce the intestinal master regulator Cdx2, shifting global gene away from anterior endoderm and toward a posterior, intestinal program. In a mouse embryonic stem cell differentiation platform that yields pure populations of definitive endoderm, Wnt signaling induces intestinal gene expression in all cells. We have identified a set of genes specific to the anterior small intestine, posterior small intestine, and large intestine during early development, and show that Wnt, through Cdx2, activates large intestinal gene expression at high doses and small intestinal gene expression at lower doses. These findings shed light on the mechanism of embryonic intestinal induction and provide a method to manipulate intestinal development from embryonic stem cells.

In embryonic stem (ES) cells, a well-characterized transcriptional network promotes pluripotency and represses gene expression required for differentiation. In comparison, the transcriptional networks that promote differentiation of ES cells and the blastocyst inner cell mass are poorly understood. Here, we show that Sox17 is a transcriptional regulator of differentiation in these pluripotent cells. ES cells deficient in Sox17 fail to differentiate into extraembryonic cell types and maintain expression of pluripotency-associated transcription factors, including Oct4, Nanog, and Sox2. In contrast, forced expression of Sox17 down-regulates ES cell-associated gene expression and directly activates genes functioning in differentiation toward an extraembryonic endoderm cell fate. We show these effects of Sox17 on ES cell gene expression are mediated at least in part through a competition between Sox17 and Nanog for common DNA-binding sites. By elaborating the function of Sox17, our results provide insight into how the transcriptional network promoting ES cell self-renewal is interrupted, allowing cellular differentiation.

Regulatory proteins can bind to different sets of genomic targets in various cell types or conditions. To reliably characterize such condition-specific regulatory binding we introduce MultiGPS, an integrated machine learning approach for the analysis of multiple related ChIP-seq experiments. MultiGPS is based on a generalized Expectation Maximization framework that shares information across multiple experiments for binding event discovery. We demonstrate that our framework enables the simultaneous modeling of sparse condition-specific binding changes, sequence dependence, and replicate-specific noise sources. MultiGPS encourages consistency in reported binding event locations across multiple-condition ChIP-seq datasets and provides accurate estimation of ChIP enrichment levels at each event. MultiGPS's multi-experiment modeling approach thus provides a reliable platform for detecting differential binding enrichment across experimental conditions. We demonstrate the advantages of MultiGPS with an analysis of Cdx2 binding in three distinct developmental contexts. By accurately characterizing condition-specific Cdx2 binding, MultiGPS enables novel insight into the mechanistic basis of Cdx2 site selectivity. Specifically, the condition-specific Cdx2 sites characterized by MultiGPS are highly associated with pre-existing genomic context, suggesting that such sites are pre-determined by cell-specific regulatory architecture. However, MultiGPS-defined condition-independent sites are not predicted by pre-existing regulatory signals, suggesting that Cdx2 can bind to a subset of locations regardless of genomic environment. A summary of this paper appears in the proceedings of the RECOMB 2014 conference, April 2–5.

Deriving lung progenitors from patient-specific pluripotent cells is a key step in producing differentiated lung epithelium for disease modeling and transplantation. By mimicking the signaling events that occur during mouse lung development, we generated murine lung progenitors in a series of discrete steps. Definitive endoderm derived from mouse embryonic stem cells (ESCs) was converted into foregut endoderm, then into replicating Nkx2.1+ lung endoderm, and finally into multipotent embryonic lung progenitor and airway progenitor cells. We demonstrated that precisely-timed BMP, FGF, and WNT signaling are required for NKX2.1 induction. Mouse ESC-derived Nkx2.1+ progenitor cells formed respiratory epithelium (tracheospheres) when transplanted subcutaneously into mice. We then adapted this strategy to produce disease-specific lung progenitor cells from human Cystic Fibrosis induced pluripotent stem cells (iPSCs), creating a platform for dissecting human lung disease. These disease-specific human lung progenitors formed respiratory epithelium when subcutaneously engrafted into immunodeficient mice.

Embryonic stem (ES) cells hold great promise with respect to their potential to be differentiated into desired cell types. Of interest are organs derived from the definitive endoderm, such as the pancreas and liver, and animal studies have revealed an essential role for Nodal in development of the definitive endoderm. Activin A is a related TGFβ member that acts through many of the same downstream signaling effectors as Nodal and is thought to mimic Nodal activity. Detailed characterization of ES cell-derived endodermal cell types by gene expression analysis in vitro and functional analysis in vivo reveal that, despite their similarity in gene expression, Nodal and Activin-derived endodermal cells exhibit a distinct difference in functional competence following transplantation into the developing mouse embryo. Pdx1-expressing cells arising from the respective endoderm populations exhibit extended differences in their competence to mature into insulin/c-peptide-expressing cells in vivo. Our findings underscore the importance of functional cell-type evaluation during stepwise differentiation of stem cells.

CD47 is a cell surface molecule that inhibits phagocytosis of cells that express it by binding to its receptor, SIRPα, on macrophages and other immune cells. CD47 is expressed at different levels by neoplastic and normal cells. Here, to reveal mechanisms by which different neoplastic cells generate this dominant ‘don't eat me' signal, we analyse the CD47 regulatory genomic landscape. We identify two distinct super-enhancers (SEs) associated with CD47 in certain cancer cell types. We show that a set of active constituent enhancers, located within the two CD47 SEs, regulate CD47 expression in different cancer cell types and that disruption of CD47 SEs reduces CD47 gene expression. Finally we report that the TNF-NFKB1 signalling pathway directly regulates CD47 by interacting with a constituent enhancer located within a CD47-associated SE specific to breast cancer. These results suggest that cancers can evolve SE to drive CD47 overexpression to escape immune surveillance.

A full description of the ontogeny of the β cell would guide efforts to generate β cells from embryonic stem cells (ESCs). The first step requires an understanding of definitive endoderm: the genes and signals responsible for its specification, proliferation, and patterning. This report describes a global marker of definitive endoderm, Claudin-6 (Cldn6). We report its expression in early development with particular attention to definitive endoderm derivatives. To create a genetic system to drive gene expression throughout the definitive endoderm with both spatial and temporal control, we target the endogenous locus with an inducible Cre recombinase (Cre-ERT2) cassette. Cldn6 null mice are viable and fertile with no obvious phenotypic abnormalities. We also report a lineage analysis of the fate of Cldn6-expressing embryonic cells, which is relevant to the development of the pancreas, lung, and liver.