Archived Projects

Transcriptional Profiling of Synovial Macrophages using Minimally Invasive Ultrasound-Guided Synovial Biopsies in Rheumatoid Arthritis

Rheumatoid arthritis (RA) is a heterogeneous disease characterized by chronic inflammation and joint destruction that likely results from a mixture of genetic predisposition and environmental factors.  Th emergence of biologic treatments, such as anti-TNF, has provided a surge of hope for RA patients.  However, only about a third of RA patients will respond to a given course of treatment and there is currently no way to predict this.

Our goal is to use cutting-edge technology to transcriptionally profile macrophages from the joint of RA patients in order to better understand the heterogeneity in disease progression and treatment response.  Macrophages have been strongly implicated in the pathogenesis of RA and reduction in the number of macrophages in the synovial sublining of the joint is a key biomarker for better outcomes. 

In the current study (Mandelin, Homan, Shaffer, et al. Arthritis & Rheumatology, 2018), we take advantage of tissue from the joints of RA paitents obtained by minimally invasive ultrasound-guided synovial biopsy.  These samples provide intact tissue suitable for histology, whole tissue RNA-seq, and cell sorting via FACS with comparable quality to samples from surgical joint replacement in osteoarthritis (OA) patients. Importantly, we then isolate synovial macrophages from RA patients in order to assess their patterns of gene expression by RNA-seq anaysis.  We find that the global gene expression of macrophages clusters RA patients into 2 major groups associated with disease severity.  In addition, patients express different combinations of co-regulated gene modules which are associated with the clinical phenotypes of patients.  These results suggest the the transcriptional profile of synovial macrophages may be used to assess disease state and  treatment response.  In the long term, we will extend these studies to test whether we can predict the best course of treatment for RA patients.

Other Projects:

• Macrophage-specific BIM deletion leads to lupus-like symptoms

• A "How-To" analyze RNA-seq for lung biologists

• Review on Functional Genomics in Rheumatology

At Weizmann Institute

The Gene Regulatory Networks of Innate Lymphoid Cells (ILCs) are Shaped by the Microbiome

Gury-BenAri, Thaiss et. al. Cell 2016

ILCs represent a unique intersection of the adaptive and innate immune systems.  They are derived from the lymphoid lineage through the common lymphoid progenitor (CLP) and are functionally similar to T cells, even sharing transcription factors (TFs), but do not have antigen-specific receptors.  Nevertheless, they play a key role in homeostasis and host defence.  ILCs are defined by 3 distinct subsets based on surface markers but whether these subtypes cover the full range of functional diversity is unclear.

In this project, we first analyze the transcriptional and epigenomic profiles of the 3 defined intestinal ILC subsets to determine how their gene regulatory networks (GRNs) differ.  Each subtype is defined by distinct genes, regulatory elements, and key TFs (T-bet for ILC1, GATA3 for ILC2, and RORC for ILC2).  Next, we used single-cell RNA-seq to define ILC subpopulations in an unbiased manner that does not rely on surface markers.  We were able to recapitulate the 3 subtypes, which further divided into transcriptional subclasses, as well as 2 additional subpopulations.  When the signals from the microbiome were disrupted, using antibiotics or germ-free mice, we found, among other changes, that ILC1s and ILC2s more closely resembled ILC3s by bulk assays and the number of cells assigned to ILC3-like subpopulations increased by single-cell analysis.  This suggested that the microbiome played a role in shaping the the distinct GRNs of ILC subpopulations.

Microglia development proceeds in a step-wise manner to complement brain development

Matcovitch-Natan, Winter et al. 2016

Microglia are the tissue-resident immune cell of the central nervous system.  They originate as erythroid-myeloid progenitors (EMPs) in the embryonic yolk sac and migrate to the brain early in the organism's development.  Microglia play an important role in brain homeostasis throughout life and their activity during immune challenge and injury has been highly studied.  However, how microglia are programmed to adapt to the varying needs of the brain during development had not previously been addressed.

We analyzed gene expression data for microglia 13 time points from the early embryo through to adult and found that development progressed through 3 distinct stages: early (until day 13 or 14), pre (day 14 to a few weeks after birth), and adult (from a few weeks after birth).  These stages represent coordinated transitions that are controlled by distinct regulatory factors as evidenced by the chromatin landscape and single cell transcriptome data. In particular, the transcription factor MafB, serves an major role in brain homeostasis as immune modulation at the adult stage.  These findings could have important implications in understanding and treating neurodevelopmental disorders such as autism and schizophrenia.

Unprecedented Heterogeneity in Level and Direction of Commitment in Myeloid Progenitors

Paul, Arkin, Giladi, et al. 2015

The common myeloid progenitor (CMP) represents the decision of a hematopoietic stem cell (HSC; the progenitor of all immune cells) to commit to the myeloid, as opposed to lymphoid, branch of hematopoiesis (literally blood-making).  Thus, the CMP compartment is responsible for producing all cells of the myeloid and erythroid lineages from erythrocytes (red blood cells) and thrombocytes (platelets) to various white blood cells (neutrophils, monocytes, macrophages, dendritic cells, basophils, etc.).  Previously, researchers believed that any individual CMP cell, which is recognized by canonical surface markers, was capable of making any and all terminal cells.

In this work, the Amit and Amos labs used single cell transcriptome technology and analysis to profile bone marrow cells based on traditional definitions representing the entire myeloid progenitor compartment including CMP, granulocyte/macrophage progenitors (GMP), and megakaryocyte/erythrocyte progenitors (MEP).  They found great variability in the transcriptional profiles of these cells suggesting that many had already committed to distinct fates,  In some cases, such as in the erythrocyte lineage, they identified clusters of cells forming a continuum from precursor to mature cell.  My contribution was to analyze chromatin data from some of these subpopulations, which exhibited long-term remodeling in agreement with their suspected fate, and identify differential regulators.  Of particular interest, was a population that expressed factors for megakaryocyte differentiation that is thought to occur after the MEP stage.  However, we found that this population contained both erythrocyte and megakaryocyte regulatory elements, the latter of which were not found in the supposed precursor, MEPs.  This paper has strong implications for how we think of the hematopoietic tree and calls into question studies based solely on identification from surface markers alone.

The Tissue Microenvironment Defines the Identity of Resident Macrophages

Lavin, Winter, Blecher et al 2014

Macrophages are a type of myeloid cell in the immune system with phagocytic properties.  Almost every tissue in our bodies have a population of macrophages with unique phenotype and function.  We were interested in how a single cell type adapts to fill such distinct roles.

We analyzed the expression and chromatin state of 7 different tissue-resident macrophage populations (lung macrophages, peritonical macrophages, microglia in the brain, Kupffer cells in the liver, spleen macrophages, and intestinal (small and large) macrophages) in mice.  Compared to the closely related myeloid populations, monocytes and neutrophils, macrophages have many regulatory regions.  However, very few (<2%) of these regions are conserved across all macrophage populations, suggesting that each macrophage subtype is triggered by the local microenvironment to form unique regulatory networks that allow for tissue-specific function.  In support of this, we found that regulatory regions (enhancers) that showed tissue-specific activity were enriched for the binding sites (motifs) of distinct regulatory factors (TFs).  Thus, the tissue environment sends signals to infiltrating macrophages that initiate the unique regulatory networks required for tissue-specific function.

Cebpb and Irf8 form a negative feedback loop in Dendritic Cells

Bornstein et al 2014

Bone marrow derived cells can differentiate into populations that resemble monocyte-derived (mo) or plasmacytoid (p) dendritic cells (DCs) depending on the medium.  But, what factors regulate this fate decision?  By identifying enhancer regions that were differentially regulated between these fates, we found that Cebpb and Irf8 in combination with PU.1 were at the apex of the regulatory networks for moDCs and pDCs, respectively.  Moreover, through knockouts and overexpression followed by chromatin profiling, we showed that these transcription factors both auto-regulate themselves and provide negative feedback to each other.  Such regulatory circuits are likely to be common in controlling fate decisions throughout hematopoiesis and other developmental pathways.

At Duke University

DNase Annotated Regions of Nucleosome Stability (DARNS)

Winter et al 2013

DNase-seq is an assay commonly used to identify genome-wide regions of open chromatin by taking advantage of the preference for DNaseI to digest naked DNA between nucleosomes.  In this study, we take advantage of another proven feature of DNase digestion, its tendency to digest nuclesome-bound DNA at low levels every 10bp or so.  Using Fourier transforms, we show that the 10bp frequency can be observed even at individual loci.  Then, we design an HMM to predict the location of nucleosome regions with strong periodic positioning.  These nucleosomes are generally conserved across cell types but cell-type-specific regulatory elements, such as enhancers, often exhibit a displaceable nucleosome in cell types in which they are inactive.  The ubiquity of our predicted regions provides a view of the nucleosome-filled genome. 

Divergent Strategies of Transcription Initiation

Rach, Winter, et al 2011

Generally, we think of genes as having a defined transcription start site (TSS) -- in other words, each copy of a gene's mRNA will begin with an identical base.  Previous studies have shown that, in reality, some genes (the minority) have a narrow peak or 'focused' transcription initiation sites (see A above), others have are much more weak peak or 'dispersed' (see C above).  We find that these promoters represent divergent strategies of gene regulation.  The focused promoters, which tend to be tissue-specific, are regulated by TFs that bind specific sequence motifs and, thus, require less organized nucleosomes around their TSSs.  On the other hand, dispersed promoters, rely on the canonical nucleosome arrangement with a defined Nucleosome Free Region (NFR) to regulate transcription.

At University of Toronto

The eFP Browser for Visualizing Gene Expression 

Winter, Vinegar et al 2007

Are you interested in the expression pattern of a particular gene but can not find an intuitive way to access or visualize the data?  Then, the eFP browser if for you.  As on of the resources on the BAR website, the eFP browser can help you do exactly that.  It was originally designed for development and various other conditions on Arabidopsis genes.  It has been since adapted for mouse, human, and other plants.