The developmental origin of functional heterogeneity of HOXB8 microglia

The role the innate immune system plays in neuropsychiatric disorders is rapidly becoming illuminated. While functional heterogeneity of innate immune cells has been demonstrated, whether the developmental origin(s) and ontogeny contribute to these diverse functions is not known. A growing body of work centers around the origin(s) and ontogeny of 1 class of innate immune cells, called tissue-resident macrophages. These fully differentiated cells arise from hematopoiesis (development of blood cells) during embryonic development. A subpopulation of tissue-resident macrophages of the brain (i.e., microglia) and adult blood cells transiently express the protein Hoxb8 very early on in differentiation, suggesting that Hoxb8 plays a role in the hematopoietic stem or progenitor cells. Until recently, microglia were thought to be homogeneous in development and function. Recent RNA sequencing of microglia highlights a cell population diverse in function, distribution, and response to injury and disease, which may be a product of their lineage origin and development. What is the biological significance of Hoxb8 expression in a lineage that gives rise to microglia? A loss-of-function mutation in Hoxb8 results in a pathological grooming phenotype in adult mice like that seen in humans with Trichotillomania, an obsessive-compulsive disorder. Dysfunctional Hoxb8 microglia have been suggested to be causative for pathological grooming. This dissertation provides a broad overview of the development of tissue-resident macrophages of the Hoxb8 lineage; explores the ontogeny, functions, and characteristics of Hoxb8 microglia; and establishes a Hoxb8 microglia transplantation mouse model. A distinct ontogeny may provide opportunities for Hoxb8 microglia to develop a set of distinct functions. Functional heterogeneity among microglia offers a preliminary framework to create therapeutic strategies directed at microglia, such as microglial replenishment, to correct the abnormal function of specific subsets of microglia, like Hoxb8 microglia, that appear to be causative in pathological behavior.

THE DEVELOPMENTAL ORIGIN OF FUNCTIONAL HETEROGENEITY OF HOXB8 MICROGLIA by Donn A. Van Deren, Jr. A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Neuroscience Interdepartmental Program in Neuroscience The University of Utah December 2019 Copyright © Donn A. Van Deren, Jr. 2019 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL Donn A. Van Deren, Jr. The dissertation of has been approved by the following supervisory committee members: Mario R. Capecchi , Chair Monica L. Vetter , Member David Krizaj , Member Karen S. Wilcox , Member Christopher T. Gregg , Member and by the Department/College/School of David Krizaj 7/31/2019 Date Approved 7/31/2019 Date Approved 7/31/2019 Date Approved 7/31/2019 Date Approved 7/31/2019 Date Approved , Chair/Dean of Interdepartmental Program in Neuroscience and by David B. Kieda, Dean of The Graduate School. ABSTRACT The role the innate immune system plays in neuropsychiatric disorders is rapidly becoming illuminated. While functional heterogeneity of innate immune cells has been demonstrated, whether the developmental origin(s) and ontogeny contribute to these diverse functions is not known. A growing body of work centers around the origin(s) and ontogeny of 1 class of innate immune cells, called tissue-resident macrophages. These fully differentiated cells arise from hematopoiesis (development of blood cells) during embryonic development. A subpopulation of tissue-resident macrophages of the brain (i.e., microglia) and adult blood cells transiently express the protein Hoxb8 very early on in differentiation, suggesting that Hoxb8 plays a role in the hematopoietic stem or progenitor cells. Until recently, microglia were thought to be homogeneous in development and function. Recent RNA sequencing of microglia highlights a cell population diverse in function, distribution, and response to injury and disease, which may be a product of their lineage origin and development. What is the biological significance of Hoxb8 expression in a lineage that gives rise to microglia? A loss-offunction mutation in Hoxb8 results in a pathological grooming phenotype in adult mice like that seen in humans with Trichotillomania, an obsessive-compulsive disorder. Dysfunctional Hoxb8 microglia have been suggested to be causative for pathological grooming. This dissertation provides a broad overview of the development of tissueresident macrophages of the Hoxb8 lineage; explores the ontogeny, functions, and characteristics of Hoxb8 microglia; and establishes a Hoxb8 microglia transplantation mouse model. A distinct ontogeny may provide opportunities for Hoxb8 microglia to develop a set of distinct functions. Functional heterogeneity among microglia offers a preliminary framework to create therapeutic strategies directed at microglia, such as microglial replenishment, to correct the abnormal function of specific subsets of microglia, like Hoxb8 microglia, that appear to be causative in pathological behavior. iv “To the individuals who dare to be great while remaining humble.” -Donn A. Van Deren, Jr. TABLE OF CONTENTS ABSTRACT....................................................................................................................... iii LIST OF FIGURES ............................................................................................................ x ACKNOWLEDGMENTS ............................................................................................... xiii Chapters 1 INTRODUCTION TO THE DEVELOPMENTAL ORIGIN OF FUNCTIONAL HETEROGENEITY: A STUDY OF HOXB8 LINEAGE TISSUE-RESIDENT MACROPHAGES AND MICROGLIA ............................................................................. 1 1.1 Introduction ....................................................................................................... 2 1.1.1 Overview of hematopoiesis...................................................................... 2 1.1.2 Hematopoiesis begins in the yolk sac ...................................................... 2 1.1.3 Hematopoiesis in the yolk sac generates 2 waves of hematopoietic progenitor cells.................................................................................................. 3 1.1.4 Fetal HSCs emerge from the aorta-gonad-mesonephros (AGM) region ................................................................................................................ 4 1.1.5 Expansion and differentiation of EMPs and f-HSCs occur in the fetal liver ................................................................................................................... 6 1.1.6 Hox gene function and its implications for hematopoiesis ...................... 6 1.1.7 Hoxb8 in hematopoiesis ........................................................................... 8 1.1.8 Evidence for the developmental origin of functional heterogeneity ........ 9 1.1.9 Ontogeny of tissue-resident macrophages ............................................. 10 1.1.10 Ontogeny of microglia ......................................................................... 11 1.1.11 Functional heterogeneity of microglia ................................................. 14 1.1.12 Function and role of microglia in neuropsychiatric disorders ............. 15 1.1.13 Identification of Hoxb8 microglia ....................................................... 17 1.1.14 Dissertation outline .............................................................................. 18 1.2 References ....................................................................................................... 20 2 THE DEVELOPMENT OF TISSUE-RESIDENT MACROPHAGES DERIVED FROM THE HOXB8 LINEAGE ...................................................................................... 37 2.1 Introduction ..................................................................................................... 38 2.2 Experimental procedures ................................................................................ 41 2.2.1 Animals .................................................................................................. 41 2.2.2 Embryo isolation .................................................................................... 41 2.2.3 Embryonic and adult tissue processing for flow cytometry................... 42 2.2.4 Postnatal blood isolation and processing for flow cytometry ................ 43 2.2.5 Statistical analysis .................................................................................. 44 2.3 Results ............................................................................................................. 44 2.3.1 EMPs of the Hoxb8 lineage ................................................................... 44 2.3.2 Fetal monocytes and macrophages of the Hoxb8 lineage ...................... 45 2.3.3 Tissue-resident macrophages of the Hoxb8 lineage .............................. 47 2.3.4 Adult hematopoietic progenitors of the Hoxb8 lineage ......................... 48 2.3.5 Adult white blood cells of the Hoxb8 lineage ....................................... 48 2.4 Discussion ....................................................................................................... 49 2.5 References ....................................................................................................... 54 3 TWO DISTINCT ONTOGENIES CONFER HETEROGENEITY TO MOUSE BRAIN MICROGLIA .................................................................................................................... 66 3.1 Abstract ........................................................................................................... 67 3.2 Introduction ..................................................................................................... 67 3.3 Results ............................................................................................................. 68 3.3.1 Hoxb8 microglial progenitors are born during the second wave of yolk sac hematopoiesis and expanded in the AGM and fetal liver prior to colonizing the E12.5 brain .......................................................................... 68 3.3.2 Early postnatal dynamics of Hoxb8 microglia....................................... 70 3.3.3 Gene expression profile of Hoxb8 microglia ......................................... 71 3.3.4 Hoxb8 hematopoietic progenitor cells derived from E12.5 fetal liver give rise to mature parenchymal microglia when introduced into neonatal mouse brains ................................................................................................... 73 3.3.5 Physiological properties of Hoxb8 microglia ........................................ 73 3.3.6 Synaptic pruning .................................................................................... 74 3.3.7 Distribution of Hoxb8 microglia in the adult mouse brain .................... 75 3.4 Discussion ....................................................................................................... 76 3.5 Materials and methods .................................................................................... 77 3.5.1 Animals .................................................................................................. 77 3.5.2 Embryo isolation .................................................................................... 77 3.5.3 Immunohistochemistry .......................................................................... 77 3.5.4 Antibodies for immunohistochemistry .................................................. 77 3.5.5 EdU labeling .......................................................................................... 77 3.5.6 Embryonic tissue processing for FACS ................................................. 77 3.5.7 FACS gating strategy for embryonic and adult tissue populations........ 77 3.5.8 Postnatal brain injections ....................................................................... 78 3.5.9 qRT-PCR................................................................................................ 78 3.5.10 Microglia isolation ............................................................................... 78 3.5.11 RNA sequencing isolation and analyses ............................................... 78 3.5.12 Microglial activation using focused laser ablation .............................. 78 3.5.13 Microglial activation using intracranial injections .............................. 78 3.5.14 Synaptic pruning assay ........................................................................ 78 3.5.15 Confocal imaging parameters .............................................................. 79 3.5.16 Imaris image analysis ........................................................................... 79 3.5.17 Statistical analysis ................................................................................ 79 vii 3.6 References ....................................................................................................... 79 4 EFFECT OF A LOSS OF FUNCTION MUTATION OF C-MYB ON HOXB8 MICROGLIA AND HOXB8 HEMATOPOIETIC PROGENITORS .............................. 88 4.1 Introduction ..................................................................................................... 89 4.2 Experimental procedures ................................................................................ 91 4.2.1 Animals .................................................................................................. 91 4.2.2 Embryo isolation .................................................................................... 92 4.2.3 Cryosectioning and immunohistochemistry .......................................... 92 4.2.4 Antibodies for immunofluorescence detection ...................................... 93 4.2.5 Confocal imaging parameters ................................................................ 93 4.2.6 Imaris image analysis ............................................................................. 94 4.2.7 Embryonic tissue processing for flow cytometry .................................. 94 4.2.8 Flow cytometry gating strategy for embryonic tissue populations ........ 95 4.2.9 Generation of c-Myb chimeric mice ...................................................... 96 4.2.10 c-Myb mutant generation ..................................................................... 97 4.2.11 Statistical analysis ................................................................................ 98 4.3 Results ............................................................................................................. 98 4.3.1 Recapitulating the c-Myb phenotype using CRISPR/Cas9 technology . 98 4.3.2 The gross phenotype of CRISPR/Cas9-targeted c-Myb mutant embryos is comparable to that of c-Myb germline null embryos ................... 99 4.3.3 Molecular analyses revealed that CRISPR/Cas9-targeted c-Myb embryos have multiple mutated alleles ......................................................... 100 4.3.4 Fetal livers of CRISPR/Cas9-targeted embryos exhibit a partial elimination of the c-Myb-dependent Hoxb8 hematopoietic progenitor cell population ..................................................................................................... 102 4.3.5 Hoxb8 hematopoietic progenitors in the fetal liver require c-Myb ..... 102 4.3.6 CRISPR/Cas9-targeted c-Myb mutant embryos display a marked reduction of Hoxb8 microglia ....................................................................... 104 4.3.7 c-Myb null embryos display a modest reduction of Hoxb8 microglia 105 4.3.8 c-Myb may have a cell non-autonomous effect on postnatal Hoxb8 microglia ....................................................................................................... 106 4.4 Discussion ..................................................................................................... 107 4.5 References ..................................................................................................... 111 5 ESTABLISHMENT OF A HOXB8 MICROGLIA TRANSPLANTATION MOUSE MODEL .......................................................................................................................... 126 5.1 Introduction ................................................................................................... 127 5.2 Experimental procedures .............................................................................. 131 5.2.1 Animals ................................................................................................ 131 5.2.2 Embryo isolation .................................................................................. 132 5.2.3 Embryonic and postnatal tissue processing for flow cytometry .......... 132 5.2.4 Flow cytometry gating strategy for embryonic and postnatal tissue populations .................................................................................................... 133 viii 5.2.5 in utero injections into the fetal liver and lateral ventricles of the developing brain............................................................................................ 134 5.2.6 Neonatal injections into motor cortices of the developing mouse brain .............................................................................................................. 135 5.2.7 Cryosectioning and immunihistochemistry ......................................... 136 5.2.8 Antibodies for immunofluorescence detection .................................... 137 5.2.9 Confocal imaging parameters .............................................................. 137 5.2.10 Imaris image analysis ......................................................................... 138 5.2.11 Behavioral testing and analysis .......................................................... 138 5.2.12 Statistical analysis .............................................................................. 139 5.3 Results ........................................................................................................... 139 5.3.1 Hoxb8 hematopoietic progenitor survive engraftment following in utero fetal liver injections....................................................................................... 139 5.3.2 Postnatal injections of Hoxb8 hematopoietic progenitors ................... 141 5.4 Discussion ..................................................................................................... 143 5.5 References ..................................................................................................... 148 6 CONCLUSIONS ON THE DEVELOPMENTAL ORIGIN OF FUNCTIONAL HETEROGENEITY OF HOXB8 MICROGLIA ........................................................... 159 6.1 Discussion ..................................................................................................... 160 6.1.1 The ontogeny of immune cells derived from the Hoxb8 lineage......... 163 6.1.2 Two populations of microglia: Hoxb8 and non-Hoxb8 microglia....... 165 6.1.3 The ontogeny of Hoxb8 microglia ....................................................... 166 6.1.4 The development of Hoxb8 microglia does not require c-Myb function ......................................................................................................... 168 6.1.5 Determining the causal link between dysfunctional Hoxb8 microglia and pathological grooming in mice............................................................... 169 6.2 References ..................................................................................................... 173 ix LIST OF FIGURES Figures 1.1 Hematopoiesis of the developing mouse embryo. ...................................................... 30 1.2. Circulating blood cells are labeled with the Hoxb8-IRES-Cre driver. ...................... 31 1.3 Embryonic microglia are not affected by a loss-of-function of c-Myb. ..................... 32 1.4 The "yolk sac" origin of microglia. ............................................................................ 33 1.5 Pathological grooming is observed in Hoxb8 null mice. ............................................ 34 1.6 Pathological grooming behavior is rescued with bone marrow transplantation of normal bone marrow. ........................................................................................................ 35 1.7 Microglia label with the Hoxb8-IRES-Cre driver. ..................................................... 36 2.1. EMPs arise from the Hoxb8 lineage. ......................................................................... 57 2.2. Fetal liver-derived monocytes and fetal macrophages are derived from the Hoxb8 lineage. .............................................................................................................................. 58 2.3. Developmental dynamics of Hoxb8-lineage fetal monocytes and macrophages. ..... 59 2.4. Postnatal dynamics of Hoxb8 lineage tissue-resident macrophages.......................... 61 2.5. Tissue-resident macrophages, derived from the Hoxb8 lineage, are largely found in active sites of hematopoiesis in the adult mouse. ............................................................. 62 2.6. All immune cells in peripheral blood are derived from the Hoxb8 lineage............... 63 2.7. All Cx3cr1-expressing peripheral immune cells are derived from the Hoxb8 lineage. .............................................................................................................................. 65 3.1. Two-color microglia mouse model. ........................................................................... 69 3.2. Hoxb8 microglia enter the brain significantly later than non-Hoxb8-microglia. ...... 70 3.3. Hoxb8 microglia progenitors are selectively expanded during AGM and fetal liver hematopoiesis. ................................................................................................................... 71 3.4. Hoxb8 expression through embryonic development ................................................. 72 3.5. Few genes are differentially expressed between Hoxb8 and non-Hoxb8-microglia. 72 3.6. E12.5 Hoxb8 fetal liver-derived hematopoietic progenitors give rise to parenchymal microglia in vivo ............................................................................................................... 73 3.7. Hoxb8 microglia have similar response to damage as non-Hoxb8-microglia ........... 74 3.8. Synaptic pruning behavior of Hoxb8 and non-Hoxb8-microglia is similar, whereas the distributions of the two microglial populations in the adult brain are significantly different ............................................................................................................................. 75 3.S1. Hoxb8-lineage progenitors in fetal liver hematopoiesis .......................................... 81 3.S2. Hoxb8-lineage progenitors comprises of ~95 of the Runx1 cell population in the yolk sac ............................................................................................................................. 82 3.S3. Postnatal development of Hoxb8- and non-Hoxb8-microglia. ................................ 83 3.S4. Developmental dynamics of Hoxb8- and non-Hoxb8-microglia ............................ 84 3.S5. Tmem119 expression. .............................................................................................. 85 3.S6. E12.5 Hoxb8-lineage fetal liver hematopoietic progenitors differentiate and repopulate the brain parenchyma in Csf1r null mice. ....................................................... 86 3.S7. Resident brain microglia are predominantly present at the site of injury. ............... 87 4.1. Targeting strategy of c-Myb by CRISPR/Cas9.. ..................................................... 114 4.2. Bright-field images of E14.5-15.5 embryos following pronuclear injection with Cas9 and c-Myb-targeting gRNA. ........................................................................................... 115 4.3. Molecular analysis and summary of CRISPR/Cas9-induced mutations within exon 1 of c-Myb.......................................................................................................................... 116 4.4. Disruption of c-Myb, a transcription factor specific for fetal liver hematopoiesis, by a CRiSPR/Cas9-induced mutation, partially eliminates the generation of Hoxb8 hematopoietic progenitors............................................................................................... 117 4.5. Germline c-Myb null embryos show that fetal liver-derived Hoxb8 hematopoietic progenitors require c-Myb. ............................................................................................. 118 4.6. The gross phenotype of Hoxb8-tdTomato reporter embryos lacking c-Myb. ......... 119 4.7. Hoxb8 hematopoietic progenitors in AGM, but not yolk sac, are affected by disruption of c-Myb. ....................................................................................................... 120 xi 4.8. Embryonic Hoxb8 microglia are reduced in CRISPR/Cas9-targeted c-Myb mutant embryos. .......................................................................................................................... 121 4.9. Embryonic Hoxb8 microglia are slightly reduced in c-Myb null brains. ................ 122 4.10. Variation in Hoxb8 and non-Hoxb8 microglial densities in c-Myb null brains. ... 123 4.11. c-Myb chimerism in P5 pups. ................................................................................ 124 4.12. c-Myb may have a cell non-autonomous effect on postnatal Hoxb8 microglia. ... 125 5.1. Donor fetal liver-derived Hoxb8 hematopoetic progenitors engraft embryonic and postnatal tissues following the in utero injection procedure. .......................................... 151 5.2. in utero brain lateral ventricle and postnatal brain injections of donor fetal liverderived Hoxb8 hematopoetic progenitors engraft in the postnatal brain. ....................... 152 5.3. Postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors engraft in brains of Csf1rΔ/+ mice. ............................................................... 153 5.4. Postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors repopulate the brains of Csf1rΔ/+ mice. ........................................................ 154 5.5. Postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors engraft in brains of mice that have a conditional loss-of-function of Csf1r in Iba1+ or Cx3cr1+ microglia-expressing cells................................................................... 155 5.6. Nearly all Iba1+ microglia in the motor cortex are derived from donor fetal liverderived Hoxb8 hematopoietic progenitors. ..................................................................... 156 5.7. Few resident Tmem119+-only microglial cells are detected in recipient brains following engraftment. .................................................................................................... 157 5.8. Recipient mice with engrafted wild-type Hoxb8 microglia show no signs of overgrooming. ........................................................................................................................ 158 6.1 Innate immune cells derived from the Hoxb8 lineage are located predominantly in areas of hematopoiesis. ................................................................................................... 178 6.2 Origin and ontogeny of Hoxb8 tissue-resident macrophages. .................................. 179 6.3 Origin and ontogeny of non-Hoxb8 and Hoxb8 microglia. ...................................... 180 xii ACKNOWLEDGMENTS I want to extend much gratitude to my committee members for their support, Drs. Mario R. Capecchi, Monica Vetter, David Krizaj, Karen Wilcox, and Christopher Gregg. In particular, I thank Dr. Mario R. Capecchi, my mentor, for his knowledge, guidance, and care throughout my education as a Ph.D student at the University of Utah School of Medicine. Dr. Capecchi has provided me the essential techniques required for studies involving microglial development and microglial-mediated neuropsychiatric disorders. I dearly thank the entire Capecchi lab, past and current lab members: Shrutokirti De, Ph.D, Simon Titen, Ph.D, Anne Boulet, Ph.D, Petr Tvrdik, Ph.D, Ben Xu, Ph.D, Matt Hockins, Ph.D, Dimitri Tränkner, Ph.D, Rick Focht (microinjection specialist), Jim Hayes (microinjection specialist), Carol Lenz (stem cell culture specialist), Sheila Barnett (stem cell culture specialist), Kay Higgins (lab manager), Joan Zhang (lab technician), and the Capecchi lab vivarium personnel (Karl Lustig, June Wangerin, Kerry Prettyman, Rob Beglarian). Additionally, I especially thank Drs. Simon Titen, Anne Boulet, Shrutokirti De, and Joan for being my bench mentors, supporters, and helpers in which they have helped me progress and mature during my graduate studies at the University of Utah. Let it be known that my family provided me their unconditional support throughout my Ph.D. education. My Mother, Erika Van Deren, and Father, Donn A. Van Deren, Sr., voluntarily spent countless hours listening to and discussing my research progress. I honor their support and patience and share my Ph.D. in Neuroscience with them. This work was supported by the National Institutes of Health [R01 MH093595 to Mario R. Capecchi], the National Cancer Institute [5P30CA042014-14], and the National Center for Research Resources [1S10RR026802-01]. No conflicts of interest exist between the subject matter and the authors included in the dissertation. xiv CHAPTER 1 INTRODUCTION TO THE DEVELOPMENTAL ORIGIN OF FUNCTIONAL HETEROGENEITY: A STUDY OF HOXB8 LINEAGE TISSUE-RESIDENT MACROPHAGES AND MICROGLIA 2 1.1 Introduction 1.1.1 Overview of hematopoiesis Hematopoiesis is the production and development of blood cells in a series of spatiotemporally defined steps in the developing embryo. This process begins with pluripotent stem cells undergoing asymmetric division to either self-renew or generate lineage-restricted committed progenitor cells. Committed progenitor cells proliferate, differentiate, and mature to generate immature blood cells specific for the lymphoid, erythroid, and myeloid lineages (1, 2). The lymphoid lineage contains T-cells, B-cells, and natural killer cells, arising from a common lymphoid progenitor cell that resides in the thymus. The erythroid lineage contains erythrocytes (red blood cells) that transport oxygen to tissues via the circulatory system. The myeloid lineage contains monocytes, macrophages, granulocytes, and megakaryocytes (platelets). Both the erythroid and myeloid lineages arise from a common myeloid progenitor (3, 4). Progenitors fated to each hematopoietic lineage are generated at different developmental times, in different hematopoietic tissues, giving rise to the spatiotemporal nature of hematopoiesis. 1.1.2 Hematopoiesis begins in the yolk sac Hematopoiesis in the mouse embryo begins in an extraembryonic tissue called the yolk sac (5), a membranous sac encapsulating the embryo, between embryonic mouse day (E) 7.0 and E7.5 (Figure 1.1). Multipotent progenitor cells derived from mesoderm form a ring of hemangioblasts (blood islands) that surround the embryo. Following expression of the transcription factor Scl (stem cell leukemia), hemangioblasts differentiate into bipotent hematogenic endothelial cells that possess endothelial and 3 hematopoietic potential (6). The transcription factor Runx1 (Runt-related transcription factor 1) is expressed in these hematopoietic progenitors of the yolk sac, and Runx1 expression is restricted to the yolk sac during primitive hematopoiesis (7). Runx1 facilitates hematopoietic commitment and is required to generate immature hematopoietic progenitor cells that exhibit a CD41+ CD45- immunophenotype. CD41 and CD45 are cell surface markers frequently used to identify immature and mature hematopoietic progenitor cells, respectively (8, 9). Once mature, hematopoietic progenitor cells will have a CD41- CD45+ immunophenotype (9). For endothelial commitment, there is evidence that the transcription factor Hoxa3 (homeobox a3) represses Runx1 expression to generate structural endothelial cells and vascular smooth muscle cells (10). Both cell types can contribute to the development of the vasculature system. 1.1.3 Hematopoiesis in the yolk sac generates 2 waves of hematopoietic progenitor cells It is currently accepted that there are 2 waves of hematopoiesis that originate from the yolk sac (11, 12). The first wave, to enter the embryo proper from the yolk sac, consists of bipotent hematopoietic progenitor cells, or erythro-myeloid progenitors (EMPs). EMPs are first detected at E7.25, just after gastrulation, in the blood islands of the yolk sac and generate the erythroid lineage (13-15) and contribute to the myeloid cell lineage (i.e., yolk sac macrophages and microglia) (12, 16-20). In the second wave beginning at E8.25, EMPs give rise to erythroid, and myeloid, but not lymphoid cell lineages (21). Which cell type gives rise to tissue-resident macrophages remains an open debate (17, 20, 22). Hoeffel et al. showed that while E7.5 yolk sac-derived EMPs give 4 rise to local macrophages in the yolk sac and microglia in the developing brain, E8.5 yolk sac-derived EMPs give rise to most adult tissue-resident macrophages, via fetal monocyte intermediates. E8.5 EMPs begin seeding the primordial fetal liver at E9.5 (13, 22) and rapidly proliferate and differentiate into different types of cells, including fetal liver monocytes (21, 23, 24). 1.1.4 Fetal HSCs emerge from the aorta-gonad-mesonephros (AGM) region Hematopoiesis begins to transition from the yolk sac to internal hematopoietic sites in the embryo beginning at E8.5 (initiation of the developing vasculature and heartbeat) (Figure 1.1). Intra-embryonic hematopoiesis is thought to begin in the paraaortic splanchnopleura (P-Sp) region at E8.5 where multipotent hematopoietic cells are found (25), concomitant with the emergence of EMPs in the yolk sac (21). The P-Sp region originates from the hemogenic endothelium of the developing embryo and is the site where immature hematopoietic stem cells (HSCs) are born. Immature HSCs mature into fetal HSCs (f-HSCs) in the AGM beginning at E10.0 when they emerge from the ventral lining of the dorsal aorta (26-30). f-HSCs are multipotent cells capable of giving rise to the lymphoid and myeloid lineages (25, 31) and have long-term reconstitution (LTR) capabilities (32). f-HSCs then seed the fetal liver and rapidly proliferate to establish the fetal liver as one of the primary sites of hematopoiesis in the developing embryo. It is also a matter of continuing debate whether the P-Sp is a bone fida site of de novo hematopoietic progenitors. While Muller et al. demonstrated that LTR capability 5 was a unique attribute of fetal HSCs isolated from the AGM after E10.0, these data do not exclude the possibility that precursors from preceding hematopoietic tissues, such as the yolk sac, may colonize the primordial AGM. Further evidence suggests that the P-Sp is a second site of hematopoietic progenitor cells (31). The potency of intra-embryonic hematopoietic cells was tested by Cumano et al. with explant cultures of splanchnic mesoderm (tissue fated to become the P-Sp and AGM) and the yolk sac, isolated before initiation of blood circulation at E8.5. Multipotent hematopoietic cells were detected as early as E7.5 in the splanchnic mesoderm, suggesting that the P-Sp and AGM, both intraembryonic sites for hematopoiesis, generate multipotent hematopoietic progenitor cells that become f-HSCs that exhibit multipotency and LTR capability in situ. Hematopoietic cells found in the yolk sac explants displayed limited differentiation potential and were unable to give rise to lymphoid cells. Interestingly, lymphoid potential can be detected in the yolk sac shortly after the initiation of blood circulation (31, 33). These data provide evidence that progenitors generated from the P-Sp could potentially colonize the yolk sac as well as the well-established model of yolk sac-derived cells colonizing tissue of the embryo proper. Additionally, yolk sac-derived EMPs are functionally different from HSCs. Transplantation of EMPs into immune-compromised recipient mice showed that these progenitor cells, unlike HSCs, were not capable of successfully taking up residency following engraftment and have no lymphoid potential, long-term repopulating potential, or Sca-1 (Stem cells antigen-1) expression on their cell surface. Sca-1 is a cell surface marker that is commonly used to identify HSCs in the developing embryo (20, 34) and the adult mouse (35). 6 1.1.5 Expansion and differentiation of EMPs and f-HSCs occur in the fetal liver From E11.5-E16.5, the fetal liver becomes the primary tissue for embryonic hematopoiesis (Figure 1.1). Johnson et al. have postulated that the fetal liver itself does not display HSC properties (e.g., LTR and multipotency capability) and cannot generate HSCs in situ; instead, it is colonized by EMPs (21, 22) and f-HSCs (20, 36) from the yolk sac and AGM, respectively. The fetal liver is comprised of a complex array of cell types that create a microenvironment necessary for the expansion and differentiation of f-HSCs and EMPs. These cell types include liver epithelium, Kupffer macrophages, hepatic stellate cells, fibroblasts, myofibroblasts, myoid cells, vascular endothelium, and mesenchymal stromal cells. During gestation, the lymphoid, erythroid, and myeloid lineages give rise to cells that will colonize the thymus and bone marrow for the remainder of the life of the organism. 1.1.6 Hox gene function and its implications for hematopoiesis Homeobox (Hox) genes encode a family of transcription factors containing 2 exons that are characterized by a DNA binding domain, termed the homeodomain (37). Hox genes were first discovered in Drosophila and later found to be conserved in all bilaterian organisms. The Hox transcription factors specify tissue patterning and segment specification by defining regional fates along the axial domains of the developing embryo (38) and are crucial to hematopoiesis (39). In general, the more complex the organism, the larger the number of Hox genes. Mammals have 39 Hox genes clustered into 4 different loci (i.e., Hoxa, Hoxb, Hoxc, Hoxd). Hox genes exhibit a phenomenon termed 7 spatial colinearity in which genes that are found in the 3’ region of the cluster are expressed in the posterior region of the embryo and genes in the 5’ region of the cluster are expressed in the anterior region of the embryo. Hox genes are required for various developmental processes in mammals, and abnormal expression can result in detrimental effects. For example, Hox11 genes are required for kidney formation. Knockout of all 3 Hox11 genes (Hoxa11, Hoxc11, Hoxd11) in mice results in the absence of kidney formation (40). Hox13 genes are required for normal digit formation — abnormal Hox13 expression results in syndactyly (fusion of digits) (41). Hoxa9 is required to regulate lymphocyte numbers and the cellular response to granulocyte-colony stimulating factor. Loss of this gene product results in lower levels of lymphocytes (42). Hoxb8 is required to prevent abnormal grooming behavior in mice. Mice lacking this gene display a pathological grooming behavior (43, 44) similar to an obsessive-compulsive disorder seen in humans, called Trichotillomania. The Hox transcription factor family is involved in regulating hematopoietic cell differentiation via regulation of gene transcription in HSCs throughout development (45); expression of Hox genes from all 4 clusters can be found in myeloid differentiation. Genes from the HoxA cluster are typically expressed in HSCs. In the AGM, there is evidence that Hoxa3 can induce endothelial cell fate within the HSC population, by suppressing Runx1 expression (10). Hoxb cluster gene expression is found in cells undergoing granulocytic differentiation. Hoxb4 has been suggested to modulate selfrenewal in hematopoietic stem cells (46). Overexpression of HOXB4 can promote the differentiation of embryonic stem cells into hematopoietic cells in vitro (47). In the lymphocytic lineage (T-cells, B-cells, and NK cells), there is increased expression of the 8 Hoxb cluster (48), although expression of these genes is not required for normal hematopoiesis. Several studies have shown that the role of Hox genes can be masked in knockout mice by functional redundancies of HOX proteins (49-51). Knockout of the Hoxb cluster did not disrupt normal hematopoiesis (52), possibly due to the functional redundancy of other Hox genes. 1.1.7 Hoxb8 in hematopoiesis The role of Hoxb8 in hematopoiesis is not well understood. Hoxb8, whose expression was found in the myelomonocytic cell line WEHI-3B, was the first Hox gene implicated in murine acute myeloid leukemia formation (53, 54). Since then, the expression of Hoxb8 in hematopoiesis has been further characterized primarily in vitro. Hoxb8 is thought to be involved in the maintenance and differentiation of myeloid progenitor cells. It has been reported that Hoxb8 expression is not detected in mature hematopoietic cells (54, 55). Abnormal Hoxb8 gene expression can result in defects in cellular differentiation and can contribute to the formation of leukemia. For example, overexpression of Hoxb8 results in immortalization of myeloid progenitor cells in vitro (56) and can result in fatal murine acute myeloid leukemia (57). Lineage tracing is a crucial tool to identify and study progeny from a single cell or a population of cells that no longer express a gene of interest but may have been developmentally affected by the transient expression of that gene. In other words, cell lineage tracing experiments mark a cell permanently, and all its progeny, following expression of a specific preselected gene to trace developmental origin(s) and cell fate. Using the cell lineage, one can also study perturbations of cell-fate choices in 9 backgrounds where the lineage marker, or other developmental regulators, has been mutated. Experiments using Hoxb8 as the lineage marker, a transgenic mouse line in which the site-specific recombinase Cre is expressed under the control of the Hoxb8 promoter and Cre-mediated recombination causes the cell to fluoresce, have shown that various immune cell populations are labeled in the adult mouse (43, 58). Examination of circulating blood cells in adult transgenic Hoxb8IRES-Cre/+; Rosa26CAG-LSL-YFP/+ mice show that hematopoietic cell populations (i.e., platelets, granulocytes, monocytes, B-cells, Tcells) are derived from the Hoxb8 lineage (43) (Figure 1.2), which suggests that, since Hoxb8 expression is restricted to early progenitors, expression of Hoxb8 occurs in the common hematopoietic stem/progenitor of all of these lineages. Nearly all Sca-1+ c-Kit+expressing HSCs isolated from the bone marrow of Hoxb8IRES-Cre/+; Rosa26CAG-LSL-YFP/+ adult mice were Hoxb8-YFP+ (Figure 1.2), which demonstrated, at the time, that Hoxb8 is expressed in HSCs. This finding also supported the hypothesis that Hoxb8 is expressed during the early stages of hematopoiesis and is downregulated in mature hematopoietic cells. Whether other innate immune cells (e.g., hematopoietic progenitor cells, f-HSCs, fetal macrophages/monocytes, erythromyeloid progenitor cells, tissue-resident macrophages) are also derived from the Hoxb8 lineage has not been reported. 1.1.8 Evidence for the developmental origin of functional heterogeneity Functional heterogeneity exists across all cell lineages; however, whether ontogeny contributes to cell diversity is currently debated (59-63). In the brain, neurons have a diverse array of physiological properties, functions, and morphology. The 10 premitotic model states that diversity arises from neuronal progenitors before mitotic cell division. Thus, the distribution of neuronal progenitors or differentiation at different developmental times may lead to neuronal diversity. The postmitotic model argues that neuronal progenitors are homogenous, and that diversity arises in developing neurons through interactions with their surrounding microenvironment. A third model, the ‘mixed’ model, proposes that neuronal progenitors are relatively homogenous, but there are subtle differences in gene expression that generate diversity (65, 66). In astrocytes, the developmental origin of functional heterogeneity is beginning to be exposed (64-67). Astrocytes are diverse in function and morphology and have multiple origin sites (68-70). Heterogeneity exists within the hematopoietic lineages. Innate immune cells such as monocytes, dendritic cells, and lymphocytes contain subpopulations with an incredible diversity of expression profiles and functions (71, 72). 1.1.9 Ontogeny of tissue-resident macrophages Tissue-resident macrophages are phagocytic cells of the innate immune system that have critical roles in removing pathogens, clearing cellular debris from dead or dying neurons, as well as mediating inflammation, and tissue repair. These cells are found in all tissues and have tissue-specific functions, thereby exhibiting remarkable functional diversity. It was believed that tissue-resident macrophages arose from 1 source – circulating monocytes derived from the bone marrow (73). Later, the yolk sac was discovered to be a second source of tissue-resident macrophages. Macrophage progenitor cells born in the yolk sac (74) populate developing tissues during embryogenesis to become self-renewing tissue-resident macrophages (21, 22, 75), including microglia (16- 11 18, 22). The fetal liver was shown to be a third source of tissue-resident macrophages by cell transplantation experiments in which fetal liver-derived monocytes gave rise to tissue-resident macrophages of the skin (Langerhans cells). Langerhans cells of the skin have been reported to have 3 origins: yolk sac, fetal liver, and bone marrow (76). In the heart, cardiac macrophages can be divided into multiple subsets arising from multiple origins (77, 78). The diversity of the developmental origin of tissue-resident macrophages of the lungs (79) suggests that developmental origin has a dominant role in determining heterogeneity and distribution of tissue-resident macrophages within a tissue. Several models have been proposed to explain the diverse ontogeny of tissueresident macrophages of the embryo proper. The Hoeffel model states that E7.5 EMPs give rise to microglia, whereas E8.5 EMPs give rise to tissue-resident macrophages of all other organ systems (11, 22). The Gomez-Perdiguero model argues that E7.5 myeloid progenitors give rise to local yolk sac macrophages that do not exit the yolk sac and E8.5 EMPs give rise to microglia and tissue-resident macrophages of other organ systems (17, 80). The Sheng model proposes that E7.5 EMPs give rise to microglia while tissueresident macrophages are generated from E8.5 f-HSCs (20). Therefore, whether tissueresident macrophages arise from E7.5 EMPs, E8.5 EMPs, f-HSCs, or all 3 is unclear. 1.1.10 Ontogeny of microglia The origin of microglia, tissue-resident macrophages of the brain, has been the subject of a long-standing debate since the late nineteenth century and continues to be a focus of intense research. Microglia, the innate immune cells of the brain, were thought to be a homogenous cell population in the brain originating from a single hematopoietic 12 source – the yolk sac. It is widely accepted that microglia are exclusively the product of hematopoiesis. Direct evidence supporting this theory is based on lineage-tracing studies using Cre recombinase to demonstrate that microglia originate from EMPs during the first wave of yolk sac hematopoiesis at E7.5 (12, 16-20). Ginhoux et al. demonstrated that resident microglia are derived from the Runx1 lineage-labeled hematopoietic progenitor cells of the yolk sac at E7.5 using the tamoxifen-inducible Runx1MER-Cre-MER; Rosa26R26R-eYFP transgenic mouse (16). Tamoxifen injections before E7.5 showed the highest percentage of recombination-labeled microglia in 8-week-old brains. Tamoxifen injections after E7.5 resulted in significantly fewer labeled microglia but increasing labeled monocytes and lung macrophages, cell types known to arise from definitive hematopoiesis. These findings corroborated previous studies that yolk sac-derived macrophage cells marked with an F4/80 antibody were found in the mesenchyme surrounding the neuroepithelium and appeared to be entering the neuroepithelium at E10.0 (85). Using Runx1MER-Cre-MER; Rosa26R26R-LacZ embryos, Runx1 lineage-labeled hematopoietic progenitor cells could be seen colonizing the developing brain at E9.5 and appeared to be associated with blood vessels. The requirement of blood vessels for the migration of yolk sac-derived microglial progenitors to the E9.5 brain was shown using Ncx1-/- embryos (16). Ncx1 is a sodium-potassium exchanger that is necessary and sufficient for the initiation of blood circulation at E8.5 (81). Characterization of yolk sac-derived microglia showed that these cells are independent of c-Myb (18, 19) and f-HSC development in the fetal liver (19). c-Myb is a proto-oncogene transcription factor that belongs to the myeloblastosis oncogene family critical for tumorigenesis, neurogenesis, and hematopoiesis. c-Myb is required for the 13 establishment of fetal liver hematopoiesis (82) and the renewal of adult HSCs (83). cMyb dependence has been used to distinguish between EMPs that give rise to adult tissue-resident macrophages or microglia. c-Myb-independent EMPs, progenitors that do not express c-Myb, are detected in the E7.5 yolk sac and give rise to yolk sac macrophages and microglia (22). c-Myb-dependent EMPs are detected in the E8.5 yolk sac and give rise to most adult tissue-resident macrophages. Furthermore, microglial populations do not appear to be affected in the developing brains of c-Myb knockout mice (Figure 1.3) (18, 19). Microglial progenitor cells acquire the expression of 2 critical factors (Csf1r and Pu.1). Csf1r (colony stimulating factor 1 receptor) is essential for the regulation of myeloid cells and necessary for the viability of microglia (84). Pu.1 (Sfi1) is a critical transcription factor for myeloid cell development (85, 86) and survival of microglia and macrophages in the yolk sac (18), but not necessary for f-HSC development in the fetal liver (19). Kierdorf et al. refined the original characterization studies done by Schulz et al. by identifying the progenitors that give rise to yolk sac-derived microglia as EMPs that are dependent on the Pu.1 and Irf8 cellular differentiation pathways (18), which are downstream targets of Runx1 (87). Pu.1 mutant embryonic and newborn brains lack microglia (18, 19, 88), and Irf8 null adult mice had a significantly reduced number of microglia (18). Shortly following the development of the embryonic vasculature and initiation of blood circulation at E8.5, E7.5-derived EMPs directly colonize the neural folds of the neural tube at E9.0 and begin to proliferate within the neuroepithelium to become resident brain microglia (16, 18) (Figure 1.4A). Therefore, the present model states that resident adult microglia are immediate descendants of the extra-embryonic 14 (yolk sac), not intra-embryonic (AGM and fetal liver) phase of hematopoiesis (Figure 1.4B). Collectively, these studies were pivotal in demonstrating that microglia arise from hematopoietic progenitor cells present during yolk sac hematopoiesis. 1.1.11 Functional heterogeneity of microglia As mentioned above, heterogeneity exists across cell lineages, including tissueresident macrophages; however, microglia are still believed to be homogeneous regarding lineage origin and development – originating from 1 hematopoietic source, the yolk sac. It is widely accepted that all microglia originate exclusively during the first wave of hematopoiesis in the yolk sac (E7.5) and subsequently populate the embryonic brain by E9.5 (Figure 1.4). However, it remains unclear whether this population accounts for the total resident microglia population following gestation, and whether microglia have more than 1 origin and ontogeny during embryonic development. Limitations exist with the Cre-loxP system used to study the origin of microglia. Using this system to fate map hematopoietic progenitors, Ginhoux et al. reported that only ~30% of microglia were labeled in the adult mouse brain. A recent study revisited the ‘yolk sac origin of microglia’ hypothesis using the zebrafish model (89). Xu et al. reasoned that as previous studies relied on the Cre-loxP system to temporally control Cre recombination, spatial control was needed to eliminate the potential contribution from other sources of hematopoiesis. Their results showed that there are 2 distinct sources of microglia in zebrafish, the rostral blood island (equivalent to mouse yolk sac) and the ventral wall of the dorsal aorta (equivalent to mouse AGM). Current studies reveal that microglia have different functions in different brain 15 regions suggesting microglial heterogeneity. For example, a subpopulation of microglia expressing CD11c (complement component 3 receptor 4) is indispensable for the proper development of oligodendrocytes and in regulating myelination in the postnatal mouse brain (90, 91). CD11c was not previously known to be expressed on microglia, but CD11c+ microglia are derived from microglial progenitors (90). The nature of this functional diversity and whether functional heterogeneity of microglia is the result of heterogeneity of the microglial lineage is unclear. 1.1.12 Function and role of microglia in neuropsychiatric disorders Microglial cells comprise the innate immune system of the fetal and adult mouse brain. These cells possess ramified processes that project from the cell soma. Their processes are dynamic and continuously survey their microenvironment by making frequent transient contacts with neuronal synapses (92). Microglia are distributed uniformly in the gray matter through a process called cell tiling, whereby the cell somas do not overlap with each other (93). Microglia are essential in the shaping and fine-tuning of developing brain circuits, neural activity, and behavior (43, 94-99). Dysfunctional microglia have been implicated in many neurological disorders and diseases, including stroke (100), multiple sclerosis (101), Parkinson’s disease (102), Huntington’s disease (103), amyotrophic lateral sclerosis (104), traumatic brain injury (105), epilepsy (106), Alzheimer’s disease (107), and neuropsychiatric disorders (108, 109), including autism spectrum disorders (110, 111), schizophrenia (110) obsessive-compulsive disorders (43, 16 44). While current research reports the roles of microglia in neurological diseases, less is known about the potential use of microglia as a therapeutic avenue. Microglia can divide and become reactive around a site of injury (112, 113), which results in neuronal tissue damage, neuroinflammation, and oxidative stress (114-117). Microglial ablation strategies have been conducted to rectify neurological disorders in mice using pharmacological inhibition and gene targeting strategies (84, 118, 119). Although these studies have positive outcomes on reducing neuroinflammation and multiple reports show no apparent signs of behavioral alterations in microglial-depleted mice (84, 120-122), these transient-immunodeficient mice may be at risk for a brain infection or the loss of microglia may show long-term effects on homeostatic brain activity. Given that microglia can repopulate the brain in microglial-depleted mouse models in a relatively short amount of time (118, 121, 123), microglial replacement therapy using hematopoietic progenitors has been proposed to correct neurological disorders (124, 125). Transplantation studies using adult bone marrow-derived cells such as hematopoietic stem and progenitor cells (HSPCs) and macrophages have been shown to prevent diseases such as EAE (126, 127), spinal cord injury (127), and pathological grooming (43). Interestingly, the transcriptional identities of engrafted myeloid cells differ with respect to the tissue (63, 128-132). For example, monocyte replacement of lung alveolar macrophages (130) and liver Kupffer cells (129) results in similar transcriptional identities. In contrast, monocyte replacement of microglia results in distinct transcriptional identities (63, 131). As a result, a limiting factor in establishing hematopoietic progenitor-derived microglial replacement therapy for neurological disorders is whether the engrafted cells become true parenchymal microglia with the 17 same properties and functions of healthy, homeostatic microglia (63, 132-135). 1.1.13 Identification of Hoxb8 microglia Hox genes were originally only known to be required for patterning and tissue specification during embryonic development (38). However, 1 Hox gene, Hoxb8, appears to have an unexpected role in pathological grooming in adult mice (44). Hoxb8 null mice exhibited an excessive amount of self-grooming and hair removal (Figure 1.5A, B). In many instances, Hoxb8 null mice develop self-inflicted skin lesions due to the severe nature of their excessive grooming (Figure 1.5C, D). An interesting observation in this study was that the backs of wild-type cage littermates were also groomed excessively (Figure 1.5E, F), most likely from the Hoxb8 null mice, which prompted the notion that the central nervous system, rather than the peripheral nervous system, plays a role in this behavior. A Hoxb8 cell lineage reporter system was developed that consisted of a Hoxb8IRES-Cre driver allele and a Rosa26-YFP reporter allele (43). Microglia were identified as the only cell type in the brain to be labeled using this mouse model (Figure 1.6). At the time (before the publication of Ginhoux et al., 2010), this finding suggested that a subpopulation of resident adult microglia was of bone marrow origin because bone marrow transplants from Hoxb8 null mice into irradiated wild-type mice recapitulated the pathological grooming behavior (43). Adult mice that lack Hoxb8 display a pathological grooming behavior similar to that observed in humans suffering from an obsessive-compulsive disorder called Trichotillomania. Excessive removal of the fur on the ventral chest can be seen on adult Hoxb8 null mice (Figure 1.6A). Behavioral analysis of the time spent grooming over a 18 24-hour period showed that Hoxb8 null mice spent twice as long grooming themselves compared to their wild-type littermates (Figure 1.6D). Because the only cells in the brain that label with the Hoxb8-IRES-Cre driver are microglia (Figure 1.7) and because microglia are derived from the immune system, bone marrow transplantations were performed in which wild-type bone marrow cells were transplanted into lethally irradiated Hoxb8 null mice to see if their behavioral phenotype could be rescued. Indeed, when a behavioral analysis was performed on these rescued Hoxb8 null mice, their time spent grooming was equivalent to wild-type controls (Figure 1.6D). Furthermore, Hoxb8 null mice had a complete restoration of fur on their chest (Figure 1.6C). Conversely, mutant Hoxb8 bone marrow cells were transplanted into lethally irradiated wild-type mice to see if pathological grooming could be observed. Excessive removal of fur and lesions, like that in Hoxb8 null mice, was observed (Figure 1.6E) and the time spent grooming was elevated but not equivalent to that of Hoxb8 null mice (Figure 1.6F) 1.1.14 Dissertation outline The work presented here encompasses a detailed examination of the ontogeny of tissue-resident macrophages derived from the Hoxb8 lineage. Further exploration was conducted on the development and functions of a subset of tissue-resident macrophages of the brain, Hoxb8 microglia, including a possible requirement for c-Myb function and establishment of a Hoxb8 microglia transplantation mouse model. The hematopoietic, neuronal, and glial cell lineages are beginning to be understood as having remarkable functional diversity (59-62, 71, 72). Could functional diversity be a result of cells having different origins and ontogenies? Chapter 2 is a comprehensive examination of a broad 19 range of immune cells in various tissues throughout embryonic and postnatal development designed to determine to what extent immune cells are derived from the Hoxb8 lineage. This chapter also helps to reconcile the existing controversial models of the ontogeny of tissue-resident macrophages. Chapter 3 is a detailed study characterizing the origin, ontogeny, functions, and characteristics of Hoxb8 microglia. This chapter provides the first piece of evidence that lineage diversity exists within the microglial population and suggests a distinct ontogeny of Hoxb8 microglia compared to other microglia studied in mice. Chapter 4 is an in-depth genetic analysis of the requirement for c-Myb function in embryonic and postnatal Hoxb8 microglia and their progenitor cells using various genetic tools. Chapter 5 examines the engraftment dynamics of embryonic Hoxb8 hematopoietic progenitor cells to determine the capability of these cells to give rise to Hoxb8 microglia and to establish a transplantation mouse model to test pathological grooming. Chapter 6 is a discussion of the ontogenies of microglia and tissue-resident macrophages derived from the Hoxb8 lineage and the steps needed, using Hoxb8 microglia replacement therapy, to elucidate how aberrant development and function of Hoxb8 microglia is causative for pathological grooming. 20 1.2 References 1. D. D. Chaplin, 1. Overview of the immune response. J Allergy Clin Immunol 111, S442-S459 (2003). 2. C. V. Laiosa et al., Determinants of lymphoid-myeloid lineage diversification. Annu Rev Immunol 24, 705-738 (2006). 3. P. Laslo et al., Gene regulatory networks directing myeloid and lymphoid cell fates within the immune system. Semin Immunol 20, 228-235 (2008). 4. A. Alagha et al., Asymmetry in erythroid-myeloid differentiation switch and the role of timing in a binary cell-fate decision. Front Immunol 4, 426 (2013). 5. W. Jollie, Development, morphology, and function of the yolk-sac placenta of laboratory rodents. Teratol 41, 361-381 (1990). 6. C. Lancrin et al., The haemangioblast generates haematopoietic cells through a haemogenic endothelium stage. Nature 457, 892-895 (2009). 7. I. M. Samokhvalov et al., Cell tracing shows the contribution of the yolk sac to adult haematopoiesis. Nature 446, 1056-1061 (2007). 8. M. J. Ferkowicz, CD41 expression defines the onset of primitive and definitive hematopoiesis in the murine embryo. Development 130, 4393-4403 (2003). 9. H. K. Mikkola et al., Expression of CD41 marks the initiation of definitive hematopoiesis in the mouse embryo. Blood 101, 508-516 (2003). 10. M. Iacovino et al., HoxA3 is an apical regulator of haemogenic endothelium. Nat Cell Biol 13, 72-78 (2011). 11. F. Ginhoux et al., Tissue-resident macrophage ontogeny and homeostasis. Immunity 44, 439-449 (2016). 12. G. Hoeffel et al., Ontogeny of tissue-resident macrophages. Front Immunol 6, 486 (2015). 13. J. Palis et al., Development of erythroid and myeloid progenitors in the yolk sac and embryo proper of the mouse. Development 126, 5073-5084 (1999). 14. J. Y. Bertrand et al., Three pathways to mature macrophages in the early mouse yolk sac. Blood 106, 3004-3011 (2005). 15. J. Tober et al., The megakaryocyte lineage originates from hemangioblast precursors and is an integral component both of primitive and of definitive hematopoiesis. Blood 109, 1433-1441 (2007). 21 16. F. Ginhoux et al., Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841-845 (2010). 17. E. Gomez Perdiguero et al., Tissue-resident macrophages originate from yolk sacderived erythro-myeloid progenitors. Nature 518, 547-551 (2015). 18. K. Kierdorf et al., Microglia emerge from erythromyeloid precursors via Pu.1and Irf8-dependent pathways. Nat Neurosci 16, 273-280 (2013). 19. C. Schulz et al., A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336, 86-90 (2012). 20. J. Sheng et al., Most tissue-resident macrophages except microglia are derived from fetal hematopoietic stem cells. Immunity 43, 382-393 (2015). 21. K. E. McGrath et al., Distinct sources of hematopoietic progenitors emerge before HSCs and provide functional blood cells in the mammalian embryo. Cell Rep 11, 1892-1904 (2015). 22. G. Hoeffel et al., C-Myb(+) erythro-myeloid progenitor-derived fetal monocytes give rise to adult tissue-resident macrophages. Immunity 42, 665-678 (2015). 23. J. Palis et al., Yolk-sac hematopoiesis: The first blood cells of mouse and man. Exp Hematol 29, 927-936 (2001). 24. P. Kumaravelu et al., Quantitative developmental anatomy of definitive haematopoietic stem cells/long-term repopulating units (HSC/RUs): Role of the aorta-gonad-mesonephros (AGM) region and the yolk sac in colonisation of the mouse embryonic liver. Development 129, 4891-4899 (2002). 25. I. Godin et al., Emergence of multipotent hematopoietic cells in the yolk sac and paraaortic splanchnopleure in mouse embryos, beginning at 8.5 days postcoitus. Proc Natl Acad Sci USA 92, 773-777 (1995). 26. T. Yokomizo et al., Three-dimensional cartography of hematopoietic clusters in the vasculature of whole mouse embryos. Development 137, 3651-3661 (2010). 27. T. Jaffredo et al., Intraaortic hemopoietic cells are derived from endothelial cells during ontogeny. Development 125, 4575-4583 (1998). 28. A. L. Medvinsky et al., Definitive hematopoiesis is autonomously initiated by the AGM region. Cell 86, 897-906 (1996). 29. R. A. Smith et al., "Hemogenic endothelium" of the embryonic aorta: Does it exist? Dev Comp Immunol 6, 359-368 (1982). 30. M. Tavian et al., Aorta-associated CD34+ hematopoietic cells in the early human embryo. Blood 87, 67-72 (1996). 22 31. A. Cumano et al., Lymphoid potential, probed before circulation in mouse, is restricted to caudal intraembryonic splanchnopleura. Cell 86, 907-916 (1996). 32. A. M. Muller et al., Development of hematopoietic stem cell activity in the mouse embryo. Immunity 1, 291-301 (1994). 33. C.-P. Liu et al., In vitro development of murine T cells from prethymic and preliver embryonic yolk sac hematopoietic stem cells. Development 113, 13151323 (1991). 34. M. A. Inlay et al., Identification of multipotent progenitors that emerge prior to hematopoietic stem cells in embryonic development. Stem Cell Rep 2, 457-472 (2014). 35. A. W. Wognum et al., Identification and isolation of hematopoietic stem cells. Arch Med Res 34, 461-475 (2003). 36. G. R. Johnson et al., Role of stem cell migration in initiation of mouse foetal liver haemopoiesis. Nature 258, 726-728 (1975). 37. J. C. Pearson et al., Modulating Hox gene functions during animal body patterning. Nat Rev Genet 6, 893-904 (2005). 38. N. Shah et al., The Hox genes and their roles in oncogenesis. Nat Rev Cancer 10, 361-371 (2010). 39. R. A. Alharbi et al., The role of HOX genes in normal hematopoiesis and acute leukemia. Leukemia 27, 1000-1008 (2013). 40. D. M. Wellik et al., Hox11 paralogous genes are essential for metanephric kidney induction. Genes Dev 16, 1423-1432 (2002). 41. F. R. Goodman, Limb malformations and the human HOX genes. Am J Med Genet 112, 256-265 (2002). 42. H. J. Lawrence et al., Loss of expression of the Hoxa-9 homeobox gene impairs the proliferation and repopulating ability of hematopoietic stem cells. Blood 106, 3988-3994 (2005). 43. S. K. Chen et al., Hematopoietic origin of pathological grooming in Hoxb8 mutant mice. Cell 141, 775-785 (2010). 44. J. M. Greer et al., Hoxb8 is required for normal grooming behavior in mice. Neuron 33, 23-34 (2002). 45. A. Seifert et al., Role of Hox genes in stem cell differentiation. World J Stem Cells 7, 583-595 (2015). 23 46. B. Schiedlmeier et al., HOXB4’s road map to stem cell expansion. Proc Natl Acad Sci USA 104, 16952-16957 (2007). 47. C. D. Helgason et al., Overexpression of HOXB4 enhances the hematopoietic potential of embryonic stem cells differentiated in vitro. Blood 87, 2740-2749 (1996). 48. R. Morgan et al., Differential expression of HOX genes upon activation of leukocyte sub-populations. Int J Hematol 87, 246-249 (2008). 49. F. Chen et al., Targeted mutations in Hoxa-9 and Hoxb-9 reveal synergistic interactions. Dev Biol 181, 186-196 (1997). 50. G. S. B. Horan et al., Compound mutants for the paralogous Hoxa-4, Hoxb-4, and Hoxd-4 genes show more complete homeotic transformations and a dosedependent increase in the number of vertebrae transformed. Genes Dev 9, 16671677 (1995). 51. D. E. Rancourt et al., Genetic interaction between Hoxb-5 and Hoxb-6 is revealed by nonallelic noncomplementation. Genes Dev 9, 108-122 (1995). 52. J. Bijl et al., Analysis of HSC activity and compensatory Hox gene expression profile in Hoxb cluster mutant fetal liver cells. Blood 108, 116-122 (2006). 53. C. Blatt et al., DNA rearrangement of a homeobox gene in myeloid leukaemic cells. EMBO J 7, 4283-4290 (1988). 54. K. Kongsuwan et al., Expression of Hox-2.4 homeobox gene directed by proviral insertion in a myeloid leukemia. Nucleic Acids Res 17, 1881-1892 (1989). 55. M. Petrini et al., Expression of selected human HOX-2 genes in B/T acute lymphoid leukemia and interleukin-2/interleukin-l P-stimulated natural killer lymphocytes. Blood 80, 185-193 (1992). 56. A. C. Perkins et al., Conditional immortalization of mouse myelomonocytic, megakaryocytic and mast cell progenitors by the Hox-2.4 homeobox gene. EMBO J 12, 3835-3845 (1993). 57. C. Scholl et al., The homeobox gene CDX2 is aberrantly expressed in most cases of acute myeloid leukemia and promotes leukemogenesis. J Clin Invest 117, 1037-1048 (2007). 58. S. De et al., Two distinct ontogenies confer heterogeneity to mouse brain microglia. Development 145, 1-14 (2018). 59. D. Jabaudon, Fate and freedom in developing neocortical circuits. Nat Commun 8, 1-9 (2017). 24 60. P. Rakic, Specification of cerebral cortical areas. Science 241, 170-176 (1988). 61. D. D. M. O'Leary, Do cortical areas emerge from a protocortex? Trends Neurosci 12, 400-406 (1989). 62. L. Puelles et al., Forebrain gene expression domains and the evolving prosomeric model. Trends Neurosci 26, 469-476 (2003). 63. F. C. Bennett et al., A combination of ontogeny and CNS environment establishes microglial identity. Neuron 98, 1170-1183 e1178 (2018). 64. L. E. Clarke et al., Diversity reaches the stars. Nature 548, 396-397 (2017). 65. H. Chai et al., Neural circuit-specialized astrocytes: Transcriptomic, proteomic, morphological, and functional evidence. Neuron 95, 531-549 e539 (2017). 66. B. S. Khakh et al., Diversity of astrocyte functions and phenotypes in neural circuits. Nat Neurosci 18, 942-952 (2015). 67. L. Ben Haim et al., Functional diversity of astrocytes in neural circuit regulation. Nat Rev Neurosci 18, 31-41 (2017). 68. S. W. Levison et al., Both oligodendrocytes and astrocytes develop from progenitors in the subventricular zone of postnatal rat forebrain. Neuron 10, 201212 (1993). 69. J. G. Parnavelas, Glial cell lineages in the rat cerebral cortex. Exp Neurol 156, 418-429 (1999). 70. C. A. G. Marshall et al., Subpallial Dlx2-expressing cells give rise to astrocytes and oligodendrocytes in the cerebral cortex and white matter. Journal Neurosci 22, 9821-9830 (2002). 71. S. Yona et al., Monocytes: Subsets, origins, fates and functions. Curr Opin Hematol 17, 53-59 (2010). 72. A. T. Satpathy et al., Re(de)fining the dendritic cell lineage. Nat Immunol 13, 1145-1154 (2012). 73. R. van Furth, The origin and kinetics of mononuclear phagocytes. J Exp Med 128, 415-435 (1968). 74. K. Takahashi et al., Differentiation, maturation, and proliferation of macrophages in the mouse yolk sac: A light-microscopic, enzyme-cytochemical, immunohistochemical, and ultrastructural study. J Leukoc Biol 45, 87-96 (1989). 75. K. Higashi et al., Ontogenetic development, differentiation, and phenotypic expression of macrophages in fetal rat lungs. J Leukoc Biol 51, 444-454 (1992). 25 76. G. Hoeffel et al., Adult langerhans cells derive predominantly from embryonic fetal liver monocytes with a minor contribution of yolk sac-derived macrophages. J Exp Med 209, 1167-1181 (2012). 77. S. Epelman et al., Embryonic and adult-derived resident cardiac macrophages are maintained through distinct mechanisms at steady state and during inflammation. Immunity 40, 91-104 (2014). 78. J. M. den Haan et al., Innate immune functions of macrophage subpopulations in the spleen. J Innate Immun 4, 437-445 (2012). 79. S. Y. Tan et al.,Developmental origin of lung macrophage diversity. Development 143, 1318-1327 (2016). 80. E. G. Perdiguero et al., The development and maintenance of resident macrophages. Nat Immunol 17, 2-8 (2016). 81. S. V. Koushik et al., Targeted inactivation of the sodium-calcium exchanger (Ncx1) results in the lack of a heartbeat and abnormal myofibrillar organization. FASEB 15, 1209-1211 (2001). 82. M. L. Mucenski et al., A functional c-Myb gene is required for normal murine fetal hepatic hematopoiesis. Cell 65, 677-689 (1991). 83. R. Sumner et al., Initiation of adult myelopoiesis can occur in the absence of cMyb whereas subsequent development is strictly dependent on the transcription factor. Oncogene 19, 3335-3342 (2000). 84. M. R. Elmore et al., Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking a microglia progenitor cell in the adult brain. Neuron 82, 380-397 (2014). 85. R. P. DeKoter et al., PU.1 regulates both cytokine-dependent proliferation and differentiation of granulocyte/macrophage progenitors. EMBO J 17, 4456-4468 (1998). 86. S. R. McKercher et al., Targeted disruption of the PU.1 gene results in multiple hematopoietic abnormalities. EMBO J 15, 5647-5658 (1996). 87. G. Huang et al., PU.1 is a major downstream target of AML1 (RUNX1) in adult mouse hematopoiesis. Nat Genet 40, 51-60 (2008). 88. D. R. Beers et al., Wild-type microglia extend survival in PU.1 knockout mice with familial amyotrophic lateral sclerosis. Proc Natl Acad Sci USA 103, 1602116026 (2006). 89. J. Xu et al., Temporal-spatial resolution fate mapping reveals distinct origins for embryonic and adult microglia in zebrafish. Dev Cell 34, 632-641 (2015). 26 90. N. Hagemeyer et al., Microglia contribute to normal myelinogenesis and to oligodendrocyte progenitor maintenance during adulthood. Acta Neuropathol 134, 441-458 (2017). 91. A. Wlodarczyk et al., A novel microglial subset plays a key role in myelinogenesis in developing brain. EMBO J 36, 3292-3308 (2017). 92. H. Wake et al., Resting microglia directly monitor the functional state of synapses in vivo and determine the fate of ischemic terminals. J Neurosci 29, 3974-3980 (2009). 93. H. Kettenmann et al.,Microglia: New roles for the synaptic stripper. Neuron 77, 10-18 (2013). 94. Y. Li et al., Reciprocal regulation between resting microglial dynamics and neuronal activity in vivo. Dev Cell 23, 1189-1202 (2012). 95. K. M. Lenz et al., Microglia are essential to masculinization of brain and behavior. J Neurosci 33, 2761-2772 (2013). 96. R. C. Paolicelli et al., Synaptic pruning by microglia is necessary for normal brain development. Science 333, 1456-1458 (2011). 97. D. P. Schafer et al., Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron 74, 691-705 (2012). 98. M. Ueno et al., Layer V cortical neurons require microglial support for survival during postnatal development. Nat Neurosci 16, 543-551 (2013). 99. N. Nagarajan et al., Corticostriatal circuit defects in Hoxb8 mutant mice. Mol Psychiatry 23, 1-10 (2018). 100. R. Guruswamy et al., Complex roles of microglial cells in ischemic stroke pathobiology: New insights and future directions. Int J Mol Sci 18, (2017). 101. J. F. Bogie et al., Macrophage subsets and microglia in Multiple Sclerosis. Acta Neuropathol 128, 191-213 (2014). 102. S. R. Subramaniam et al., Targeting microglial activation states as a therapeutic avenue in Parkinson's Disease. Front Aging Neurosci 9, 176 (2017). 103. H. M. Yang et al., Microglial activation in the pathogenesis of Huntington's Disease. Front Aging Neurosci 9, 193 (2017). 104. M. C. Geloso et al., The dual role of microglia in ALS: Mechanisms and therapeutic approaches. Front Aging Neurosci 9, 242 (2017). 27 105. C. K. Donat et al., Microglial activation in traumatic brain injury. Front Aging Neurosci 9, 208 (2017). 106. U. B. Eyo et al., Microglia-neuron communication in epilepsy. Glia 65, 5-18 (2017). 107. D. V. Hansen et al., Microglia in Alzheimer's disease. J Cell Biol 217, 459-472 (2018). 108. V. Mondelli et al., Pariante, Brain microglia in psychiatric disorders. Lancet Psychiatry 4, 563-572 (2017). 109. M. Prinz et al., Microglia and brain macrophages in the molecular age: From origin to neuropsychiatric disease. Nat Rev Neurosci 15, 300-312 (2014). 110. M. W. Salter et al., Microglia emerge as central players in brain disease. Nat Med 23, 1018-1027 (2017). 111. C. A. Edmonson et al., A non-inflammatory role for microglia in Autism Spectrum Disorders. Front Neurol 7, 1-6 (2016). 112. C. Madore et al., Microglial confetti party. Nat Neurosci 20, 762-763 (2017). 113. T. L. Tay et al., A new fate mapping system reveals context-dependent random or clonal expansion of microglia. Nat Neurosci 20, 793-803 (2017). 114. R. Wada et al., Microglial activation precedes acute neurodegeneration in Sandhoff disease and is suppressed by bone marrow transplantation. Proc Natl Acad Sci USA 97, 10954-10959 (2000). 115. K. Ohmi et al., Activated microglia in cortex of mouse models of mucopolysaccharidoses I and IIIB. Proc Natl Acad Sci USA 100, 1902-1907 (2003). 116. S. Boillee et al., Onset and progression in inherited ALS determined by motor neurons and microglia. Science 312, 1389-1392 (2006). 117. C. S. Jack et al., TLR signaling tailors innate immune responses in human microglia and astrocytes. J Immunol 175, 4320-4330 (2005). 118. N. H. Varvel et al., Microglial repopulation model reveals a robust homeostatic process for replacing CNS myeloid cells. Proc Natl Acad Sci USA 109, 1815018155 (2012). 119. A. Waisman et al., Homeostasis of microglia in the adult brain: Review of novel microglia depletion systems. Trends Immunol 36, 625-636 (2015). 28 120. N. N. Dagher et al., Colony-stimulating factor 1 receptor inhibition prevents microglial plaque association and improves cognition in 3xTg-AD mice. J Neuroinflammation 12, 139 (2015). 121. M. R. Elmore et al., Characterizing newly repopulated microglia in the adult mouse: Impacts on animal behavior, cell morphology, and neuroinflammation. PLoS One 10, e0122912 (2015). 122. R. A. Rice et al., Microglial repopulation resolves inflammation and promotes brain recovery after injury. Glia 65, 931-944 (2017). 123. H. Lund et al., Competitive repopulation of an empty microglial niche yields functionally distinct subsets of microglia-like cells. Nat Commun 9, 4845 (2018). 124. Z. Shen et al., Microglia-targeted stem cell therapies for Alzheimer disease: A preclinical data review. J Neurosci Res 95, 2420-2429 (2017). 125. R. K. Srivastava et al., Migratory potential of transplanted glial progenitors as critical factor for successful translation of glia replacement therapy: The gap between mice and men. Glia 66, 907-919 (2018). 126. X. M. Zhang et al., Adoptive transfer of cytokine-induced immunomodulatory adult microglia attenuates experimental autoimmune encephalomyelitis in DBA/1 mice. Glia 62, 804-817 (2014). 127. S. F. Ma et al., Adoptive transfer of M2 macrophages promotes locomotor recovery in adult rats after spinal cord injury. Brain Behav Immun 45, 157-170 (2015). 128. M. Guilliams et al., Does niche competition determine the origin of tissueresident macrophages? Nat Rev Immunol 17, 451-460 (2017). 129. C. L. Scott et al., Bone marrow-derived monocytes give rise to self-renewing and fully differentiated Kupffer cells. Nat Commun 7, 10321 (2016). 130. L. van de Laar et al., Yolk sac macrophages, fetal liver, and adult monocytes can colonize an empty niche and develop into functional tissue-resident macrophages. Immunity 44, 755-768 (2016). 131. J. C. Cronk et al., Peripherally derived macrophages can engraft the brain independent of irradiation and maintain an identity distinct from microglia. J Exp Med 215, 1627-1647 (2018). 132. A. Shemer et al., Engrafted parenchymal brain macrophages differ from microglia in transcriptome, chromatin landscape and response to challenge. Nat Commun 9, 5206 (2018). 29 133. B. Ajami et al.,Local self-renewal can sustain CNS microglia maintenance and function throughout adult life. Nat Neurosci 10, 1538-1543 (2007). 134. A. Capotondo et al., Brain conditioning is instrumental for successful microglia reconstitution following hematopoietic stem cell transplantation. Proc Natl Acad Sci USA 109, 15018-15023 (2012). 135. M. L. Bennett et al., New tools for studying microglia in the mouse and human CNS. Proc Natl Acad Sci USA 113, E1738-1746 (2016). 30 Figure 1.1 Hematopoiesis of the developing mouse embryo. Illustrations represent the major sites of embryonic hematopoiesis (shown in red). The first hematopoietic progenitors are born in the blood islands of the yolk sac (YS) between E7.0-7.5. Following blood circulation at E8.5, the AGM region, within the embryo proper (EP), becomes the next active hematopoietic tissue generating hematopoietic stem cells de novo from E9.5-11.5. Circulating HPSCs mainly derived from the yolk sac or AGM region begin populating the primordial fetal liver as early as E10.0. Hematopoiesis is established in the fetal liver (FL) by E12.5, where yolk sac-derived and AGM-derived HPSCs rapidly expand and differentiate into lineage-restricted hematopoietic cells. 31 Figure 1.2. Circulating blood cells are labeled with the Hoxb8-IRES-Cre driver. A-C FACS profiles of adult circulating blood cells gated for myeloid cells (MAC1/GR1+ ), B cells (CD19+ ), and T cells (CD4/CD8+) on the y-axis and YFP on the x-axis isolated from ROSA26CAG-LSL-YFP/CAG-LSL-YFP or D-F Hoxb8IRES-Cre/+; ROSA26CAG-LSL-YFP/+ (D-F). G Representative fluorescence microscopy image of bone marrow cells expressing Hoxb8YFP isolated from an Hoxb8IRES-Cre/+; ROSA26CAG-LSL-YFP/+ adult mouse. H FACS profile of bone marrow cells gated for Sca-1 (x-axis) and c-Kit (y-axis) to examine HSCs. Sca-1+ c-Kit+ cells were further analyzed for Hoxb8-YFP detection. Black histogram: ROSA26CAG-LSL-YFP/CAG-LSL-YFP adult mouse, white histogram: Hoxb8IRES-Cre/+; ROSA26CAGLSL-YFP/+ adult mouse. (Adapted from Chen et al., 2010). 32 Figure 1.3 Embryonic microglia are not affected by a loss-of-function of c-Myb. A Fluorescence microscopy micrograph of cells labeled with DAPI (blue) a microglial cell labeled with Iba-1 (red) from the neuroectoderm of c-Myb+/+ and c-Myb-/- E16.5 embryos. Scale bar: 25μm. (Adapted from Schulz et al., 2012). B Bar graph showing the number of Iba-1+ cells/mm2 (mean±SEM) in the brains of c-Myb+/+ and c-Myb-/- E16.5 embryos. (Adapted from Schulz et al., 2012). C Bar graph showing the number of Iba-1+ cells/mm2 (mean±SEM) in the brains of c-Myb+/+ and c-Myb-/- E14.5 embryos. (Adapted from Kierdorf et al., 2013). 33 Figure 1.4 The "yolk sac" origin of microglia. A Representative sagittal brain sections of Cx3cr1GFP/+ embryos at different stages of embryonic development showing the neural tube (N) and mesenchyme (M) (red, anti-nestin antibody). Myeloid cells (green) are Cx3cr1-GFP+. By E9.0, Cx3cr1-GFP+ myeloid cells surround the border of the neural tube. At E9.5, Cx3cr1-GFP+ myeloid cells begin colonizing the neural tube and rapidly develop into microglial cells with ramified processes by E14.0. Arrows mark myeloid cells. Arrowheads mark microglia. Scale bars: 100μm (overviews) and 50μm (inserts) (Modified from Kierdorf et al., 2013). B Illustration of the origin and ontogeny of nonHoxb8-microglia: non-Hoxb8 microglial progenitors are born in the E7.5 yolk sac (YS), enter the embryo proper (EP), colonize the E9.5 brain, and become non-Hoxb8 microglia. 34 Figure 1.5 Pathological grooming is observed in Hoxb8 null mice. A-B Images showing fur removal on the lateral and ventral chest areas of Hoxb8-/- mice. B-C Some Hoxb8-/- mice exhibited deep open wounds on their lateral and ventral chest areas. E-F Hair removal on the backside of wild-type adult mice housed with Hoxb8-/- mice. (Adapted from Greer et al., 2002). 35 Figure 1.6 Pathological grooming behavior is rescued with transplantation of normal bone marrow. A Fur removal on the ventral chest of Hoxb8-/- mice. B-C Complete restoration of the fur of a Hoxb8-/- mouse 3 months following transplantation with normal bone marrow. D Behavioral analysis on time spent grooming over a 24-hour period in wild-type (white bar, n = 22), Hoxb8-/- (black bar, n = 25), and rescued Hoxb8-/(gray bar, n = 6) mice. All values are mean ± SEM. *p < 0.05 versus mutant. E Wild-type adult mouse transplanted with bone marrow from Hoxb8-/- mice showing hair removal and lesions. F Behavioral analysis on time spent grooming over a 24-hour period in wildtype (white bar, n = 22), Hoxb8-/- (black bar, n = 25), and wild-type mice transplanted with bone marrow from Hoxb8-/-mice (gray bar, n = 2) mice. Error bars represent SEM. *p < 0.05 versus wild-type. (Adapted from Chen et al., 2010). 36 Figure 1.7 Microglia label with the Hoxb8-IRES-Cre driver. Representative sagittal section of the cerebral cortex of an Hoxb8IRES-Cre/+; Rosa26CAG-LSL-YFP/+ adult mouse showing A fluorescence detection of Hoxb8-YFP+ cells B costained with an anti-CD11b antibody. C Colocalization of Hoxb8-YFP (yellow) and CD11b (red) signals. Scale bar: 40μm, (Modified from Chen et al., 2010). CHAPTER 2 THE DEVELOPMENT OF TISSUE-RESIDENT MACROPHAGES DERIVED FROM THE HOXB8 LINEAGE 38 2.1 Introduction There are at least 3 successive waves of hematopoiesis in the developing mouse embryo (1-3). These hematopoietic waves overlap with each other to timing and the sequential activity levels of hematopoietic tissues due to blood circulation. The first wave initiates in the blood islands of the yolk sac and generates erythromyeloid progenitors (EMPs) at embryonic day (E) E7.5. The second wave initiates in the hemogenic endothelium of the yolk sac and generates a second set of EMPs at E8.5. The third wave initiates in the hemogenic endothelium of the para-aortic splanchnopleura (P-Sp), which then becomes the aorta-gonad-mesonephros (AGM) region, and generates fetal hematopoietic stem cells (f-HSCs) at E10.5. The ontogeny of tissue-resident macrophages, including microglia, encapsulating their origin and hematopoietic progenitors, the pathway of development, and the genes required for differentiation are under debate. For example, while it is largely believed that yolk sac EMPs generate microglia through yolk sac macrophage intermediates, it is unclear whether yolk sac EMPs originate from the first wave or second wave of hematopoiesis, or both. Currently, the field is split between 2 models proposing different origins and ontogenies of tissueresident macrophages and microglia. The first model, proposed by Hoeffel et al., states that yolk sac EMPs are heterogeneous in that they are generated sequentially and can be defined functionally. The authors demonstrated that EMPs born in the E7.5 yolk sac (1st wave) differentiate into nucleated erythrocytes and yolk sac macrophages. Following blood circulation at E8.5, yolk sac macrophages migrate via the developing vasculature to populate all developing tissues to generate tissue-resident macrophages and in the brain, microglia. 39 E7.5 yolk sac EMPs require Colony stimulating factor 1 receptor (Csf1r) expression for their survival and differentiation but do not require c-Myb expression (4, 5). Csf1r is essential for the regulation of myeloid cells (6) and necessary for the viability of microglia (7). c-Myb is required for the establishment of hematopoiesis in the fetal liver and the renewal of adult HSCs (8-10). EMPs born in the E8.5 yolk sac (2nd wave) do not require Csf1r expression but do require c-Myb expression (4, 11). While a subset of E8.5 yolk sac EMPs gives rise to local macrophages that do not exit the yolk sac, most E8.5 yolk sac EMPs leave the yolk sac and seed the primordial fetal liver by E9.5, where they become fetal liver myeloid progenitor cells expressing Csf1r. Fetal liver myeloid progenitor cells differentiate into fetal liver-derived monocytes, via fetal liver common monocyte progenitor cells, that have an F4/80low CD11bhigh immunophenotype signature (3, 4). F4/80low CD11bhigh fetal liver-derived monocytes rapidly proliferate in the fetal liver from E12.5 to E14.5. Another hematopoietic site within the embryo proper generates immature f-HSCs at E8.5 in the P-Sp region (3rd wave). These immature fHSCs mature into f-HSCs and begin seeding the primordial fetal liver by E10.5. f-HSCs, like E8.5 yolk sac EMPs, also contribute to the generation of fetal liver-derived monocytes (3, 4, 12). At E14.5, fetal liver-derived monocytes exit the fetal liver to populate all tissues of the developing embryo except the brain. The second model, proposed by Gomez-Perdiguero et al., argues that all yolk sac EMPs are born in the yolk sac at E8.5 (2nd wave). The authors showed that these cells differentiate into yolk sac macrophages which, in turn, populate all tissues of the developing embryo, including the brain, to become tissue-resident macrophages and microglia, respectively (13-15). They determined that hematopoietic progenitors that are 40 born in the yolk sac at E7.5 (1st wave) only give rise to local yolk sac macrophages that do not exit the yolk sac. They also found that E8.5 yolk sac EMP-derived and AGM fHSC-derived fetal liver monocytes minimally contribute to tissue-resident macrophage populations, including microglia. c-Myb is expressed in E8.5 yolk sac-derived EMPs and is required for the differentiation of erythroid, but not myeloid, progenitor cells (15). In cMyb null embryos, there is an absence of F4/80low CD11bhigh fetal liver-derived monocytes, but F4/80high CD11blow fetal macrophages remain present (9), suggesting that fetal macrophages arise from a c-Myb-independent lineage originating in the yolk sac (i.e., E8.5 yolk sac EMPs). The models proposed by Hoeffel et al. and Gomez-Perdiguero et al. both agree that yolk sac EMPs generate microglia through yolk sac macrophage intermediates. The debate lies in whether these yolk sac EMPs are born during the first or second wave of hematopoiesis. The ontogeny of tissue-resident macrophages, whether these cells originate from yolk sac macrophages (Gomez-Perdiguero model) or fetal liver-derived monocytes (Hoeffel model), also remains unclear. Adding to this complexity, the precise origin of fetal liver-derived monocytes is uncertain because of the contribution from AGM-derived f-HSCs (12) and yolk sac-derived E8.5 EMPs (4). Furthermore, tamoxifen-induced Cre recombination is occurring during a time when developmental changes are happening at a rapid pace. There is an overlap of hematopoietic activity in different hematopoietic tissues due to blood circulation, adding to this challenge. Could there be a single Cre driver that helps to clarify the proposed models of the ontogeny of tissue-resident macrophages? The Hoxb8-IRES-Cre driver has been shown to label immune cell populations in adult blood such as the erythroid, myeloid, and 41 lymphoid cell populations and a subpopulation of microglia (16, 17). However, the extent to which this driver labels immune cell populations, including tissue-resident macrophages and their hematopoietic progenitors, remains to be identified. Here, we show the ontogeny of tissue-resident macrophages, their hemopoietic progenitors, and hematopoietic cell intermediates using the Hoxb8-tdTomato lineage reporter mouse as described by De et al. throughout embryonic development and adulthood. 2.2 Experimental procedures 2.2.1 Animals Gt(ROSA)26Sortm14(CAT-tdTomato)Hze (Ai14, #007908) mice were obtained from the Jackson Laboratory. Cx3cr1GFP mice (#005582, Jackson Laboratory) were a kind gift from Dr. Monica Vetter, Department of Neurobiology & Anatomy, University of Utah, Salt Lake City, UT. Hoxb8IRES-Cre mice were generated in our lab and reported in Chen et al. 2010. 2.2.2 Embryo isolation Timed matings were set up between male and female mice of the desired genotypes. Plug checks of the female were done daily. The day a female was successfully plugged was marked as Day 0.5 (E0.5) of gestation. On the days of dissection, females were euthanized, and laparotomy was performed to remove the uterine horns containing the embryos. The uterine horns were placed on ice-cold 1X HBSS, and embryos with their attached yolk sac were carefully transferred from the placenta. Yolk sac and fetal liver were dissected and processed for flow cytometry as described below. 42 2.2.3 Embryonic and adult tissue processing for flow cytometry Embryos were removed from the uterus and placed on ice in 5% fetal bovine serum (FBS, Atlanta Biologicals) in 1X Hanks’ balanced salt solution (HBSS, Gibco). Similarly, postnatal tissue was dissected and placed on ice in 5% fetal bovine serum (FBS, Atlanta Biologicals) in 1X HBSS. Dissections of the following embryonic tissues were performed at their respective time points: yolk sac (E8.0, 8.5, 9.5, 10.5, 11.5, 12.5) and fetal liver (E11.5, 12.5, 14.5, 15.5, 17.5, 18.5). Dissected tissues before E12.5 were pooled together to obtain enough cells at these time points for flow cytometry. Dissections of liver, lung, spleen, thymus, and heart were performed at 2 and 6 months of age. The tissue was broken up into single cells using a 26G needle for embryonic tissue and a 19G needle for adult tissue. The dissociated tissues were spun down at 1,500 x g for 5 minutes at 4˚C. Cells were washed 3 times with 1X HBSS and spun down at 1,500 x g for 5 minutes. The cells were passed through an 80 µm mesh to obtain a single-cell suspension. The suspension was incubated in 20µl of red blood cell lysis buffer (0.15M ammonium chloride, 1mM KHCO3) for 5 minutes at room temperature, then ~13 ml of 1X HBSS was added, and incubation continued for 5 minutes to lyse red blood cells, followed by another 5-minute spin at 200 x g. The anti-mouse antibodies used consisted of the following: TER-119 PerCP-Cy5.5 (1:50, BioLegend, #116228), CD11b BV510 (1:100, Biolegend, #101263), CD11b APC (1:100, BioLegend, #101212), F4/80 PE-Cy7 (1:100, BioLegend, #123114), c-Kit PE-Cy7 (1:100, Biolegend, #105814), CD45 APC (1:100, BioLegend, #103112), CD16/32 APC (1:100, BioLegend, #101326), Ly6C APCCy7 (1:100, BioLegend, #126026), CD41 APC-eFluor® 780 (1:100, eBioscience, #470411-80) and CD41 APC-Cy7 (1:100, BioLegend, #133928) in 1X HBSS. Cells were 43 incubated in the dark with this antibody cocktail for 30 minutes on ice. Cells were washed with 1X HBSS and spun down at 1,500 x g for 5 minutes, then resuspended in 1X HBSS with DAPI (1:1000). Flow cytometry data were obtained using the BD Bioscience FACSCanto II flow cytometry sorter. All FACS data were analyzed with FlowJo 10.0.7 (Celeza GmbH). 2.2.4 Postnatal blood isolation and processing for flow cytometry Blood was collected into microfuge tubes containing 10μL 0.5M EDTA-AC. Blood was isolated at 2, 3, and 6 months of age. Blood samples were inverted a few times gently and transferred to a 15mL conical tube containing 2mL 1x RBC lysis buffer. Samples were mixed gently, and removal of any blood clots present was done with a 1mL wide bore pipet tip. Amount of 1x RBC lysis buffer was adjusted to 2mL/0.1g of blood collected. Tubes were gently inverted a few times, followed by incubation for 10 minutes at room temperature. Lysed blood samples were then centrifuged for 5 minutes at 1,500 x g at room temperature. Blood cells were resuspended in 5mL fresh RBC lysis buffer and spun down for 5 minutes at 1,500 x g at room temperature. Blood cells were then resuspended in 2mL 1X HBSS and spun down again for 2 minutes at 1,500 x g. The antimouse antibodies used consisted of the following: TER-119 PerCP-Cy5.5 (1:50, BioLegend, #116228), CD11b APC (1:100, BioLegend, #101212), CD115 PE-Cy7 (1:100, Biolegend, #135524), and B220 APC-Cy7 (1:100, BioLegend, #103224) in 1X HBSS. Cells were incubated in the dark with this antibody cocktail for 30 minutes on ice. Cells were washed with 1X HBSS and spun down at 1,500 x g for 5 minutes, then resuspended in 1X HBSS with DAPI (1:1000). Flow cytometry data were obtained using 44 the BD Bioscience FACSCanto II flow cytometry sorter. All FACS data were analyzed with FlowJo 10.0.7 (Celeza GmbH). 2.2.5 Statistical analysis Unless otherwise stated, all results are reported as mean ± SEM. 2.3 Results 2.3.1 EMPs of the Hoxb8 lineage Previously, we identified the earliest developmental time point at which Hoxb8 is expressed during embryonic hematopoiesis. We were able to detect Hoxb8-tdTomato+ cells in hematopoietic tissues as early as E8.5 in the yolk sac (17), coincident with the emergence of E8.5 yolk sac EMPs (18). While several studies have classified EMPs differently, 2 common cell surface markers have been used, CD41 and c-Kit (4, 11, 18). We chose to study E8.5 yolk sacderived EMPs having the immunophenotype c-Kithigh CD41+ CD16/32+ proposed by McGrath et al., for 2 reasons. First, E8.5 EMPs express the cell surface marker CD16/32 while earlier E7.5 EMPs, or progenitor cells with lymphoid potential, do not. Second, cKithigh CD41+ CD16/32+ EMPs lack the expression of Sca-1, which makes them phenotypically distinct from f-HSCs. Combining these markers with our Hoxb8tdTomato reporter mouse, we were able to determine that a subset of EMPs were tdTomato+ in the yolk sac at E8.5 (9.17%) and peaked at E9.5 (13.5%) (Figure 2.1A, C). E8.5 EMPs have been shown to exit the yolk sac and seed the fetal liver by E9.5 (4, 18). Examining the fetal liver as early as E11.5 (Figure 2.1B), the percentage of 45 tdTomato+ EMPs (48.1%) was higher compared to all time points examined in the yolk sac (Figure 2.1C) further supporting the seeding of yolk sac EMPs during the early stages of fetal liver hematopoiesis. By E14.5 in the fetal liver, the percentage significantly increased (81.4%), and by E17.5, the percentage peaked to 91.1% (Figure 2.1C). Collectively, these data show that a subpopulation of E8.5 yolk sac-derived EMPs (~9%), derived from the Hoxb8 lineage, transits to the fetal liver and rapidly expands up to ~91%. Erythrocyte development begins in the yolk sac and continues in the fetal liver. We wondered whether erythroid lineage cells (Ter119+) are could also be derived from the Hoxb8 lineage. We found that the percentages of erythrocytes that were tdTomato+ were negligible in the yolk sac and early stages of fetal liver development but noticed a significant increase in the fetal liver between E15.5 (7.39%) to E17.5 (55.5%) and peaked at E18.5 (63.2%) (Figure 2.1D), showing that more than half of erythrocytes are derived from Hoxb8-expressing cells. 2.3.2 Fetal monocytes and macrophages of the Hoxb8 lineage We next looked at circulating and tissue-infiltrating fetal monocytes for the Hoxb8 lineage marker. Since several studies provide evidence that E8.5 EMPs give rise to tissue-resident macrophages via fetal monocytic intermediates found in the fetal liver (4, 18), flow cytometry analysis to determine whether the Hoxb8-tdTomato reporter lineage is found in these cell populations was performed. Fetal monocytes first appear in the fetal liver by E12.5 (4, 19, 20). E8.5 yolk sac EMPs and AGM-derived f-HSCs have been shown to differentiate into fetal monocytes 46 in the fetal liver by E12.5 and exit the tissue by E14.5 to other developing tissues to become tissue-resident macrophages (4, 12, 18). Since the Hoxb8-tdTomato lineage reporter labels a significant fraction of EMPs (Figure 2.1) and f-HSCs (17) in the fetal liver, we asked whether our reporter also labels fetal monocytes (CD45+ F4/80low CD11bhigh) in the fetal liver at different developmental time points. Cells were isolated from the fetal liver of E12.5 – E18.5 Hoxb8-tdTomato reporter embryos and analyzed by flow cytometry. tdTomato+ fetal monocytes substantially increased in percentage from E12.5 (37.3%) to E14.5 (91.6%) and increased up to 97.7% by E18.5 (Figure 2.2A, C). Fetal monocytes, which also express Cx3cr1 during differentiation (4), have been reported to rapidly expand in the fetal liver from E12.5 to E14.5 (4). At E12.5 and E14.5, most Cx3cr1-GFP+ cells were Hoxb8-tdTomato+ (Figure 2.3A), which suggests that these cells are differentiating into tissue-resident macrophages, as previously reported by Hoeffel et al. Fetal monocytes have been shown to exist in the fetal liver as 2 populations differentially expressing Ly6C (4, 21) In both subsets of fetal monocytes (Ly6C+ and Ly6C-), the percentage of tdTomato+ cells were equivalent (avg Ly6C+: 94.2%, avg Ly6C-: 90.8%) and >90% of each subset were tdTomato+ (Figure 2.3B, C). We also examined whether fetal macrophages (CD45+ F4/80high CD11blow) in the yolk sac and fetal liver are derived from the Hoxb8 lineage. Cells were isolated from yolk sac of E9.5 – E12.5 and fetal liver of E12.5 – E18.5 Hoxb8-tdTomato reporter embryos and analyzed by flow cytometry. We could detect tdTomato+ fetal macrophages in the E9.5 yolk sac (11.1%) but were virtually absent by E12.5 (0.30%) (Figure 2.2B, D). In the developing fetal liver, there was a marked increase in the percentage of tdTomato+ fetal macrophages from E12.5 (18.4%) to E14.5 (58.0%) (Figure 2.2B, D). Cx3cr1 is 47 expressed on some fetal macrophages of the yolk sac and fetal liver (4, 5, 9, 11). At E10.5 in the yolk sac, ~43% tdTomato+ fetal macrophages were Cx3cr1-GFP+ whereas, in the E12.5 fetal liver, ~24% were labeled and increased to ~53% by E14.5 (Figure 2.3D). 2.3.3 Tissue-resident macrophages of the Hoxb8 lineage Based on our data that a subpopulation of fetal macrophages in the yolk sac (~11%) and fetal liver (~58%) labeled with the Hoxb8-tdTomato reporter, we examined whether tissue-resident macrophages are derived from the Hoxb8 lineage in postnatal tissues. Cells were isolated from heart, liver, lungs, thymus, and spleen at 2 and 6 months of age and analyzed by flow cytometry. Cells were sorted with monocyte (CD45+ F4/80low CD11bhigh Hoxb8-tdTomato+ Cx3cr1-GFP+ or -) and macrophage (CD45+ F4/80high CD11blow Hoxb8-tdTomato+ Cx3cr1-GFP+ or -) markers. When we examined adult tissues not known to be sites for adult hematopoiesis (e.g., heart, lungs, liver), most macrophages were either GFP+ or tdTomato+ (Figure 2.4A-C). Conversely, nearly all GFP+ splenic macrophages and thymic macrophages were Hoxb8-tdTomato+ (Figure 2.4D-E). Any evidence of tdTomato+ cells within the TRM cell population was virtually nonexistent in the heart (avg: 0.15%), lungs (avg: 1.49%), and liver (avg: 0.18%) (Figure 2.5A, B). Analysis of cells isolated from thymus and spleen, known adult hematopoietic sites, however, showed that half of the thymic macrophages (avg: 54.3%) and virtually all splenic macrophages were Hoxb8-tdTomato+ (avg: 90.4%) (Figure 2.5A, B). 48 2.3.4 Adult hematopoietic progenitors of the Hoxb8 lineage Our lab has recently reported that hematopoietic progenitor cells (CD41+ c-Kithigh, CD41+ CD45+ c-Kithigh, or CD45+ c-Kithigh), f-HSCs, and adult bone marrow-derived HSCs are nearly all derived from the Hoxb8 lineage (16, 17), indicating that the Hoxb8 lineage is predominantly located in active hematopoietic sites during embryonic development and adulthood. In the adult bone marrow, an active site for adult hematopoiesis, nearly 100% of the hematopoietic progenitor cells (Lineage- Sca-1+ cKithigh [LSK+]) were tdTomato+. We then examined other adult hematopoietic sites, thymus and spleen, for Hoxb8 lineage hematopoietic progenitors (CD45+ c-Kithigh Hoxb8tdTomato+). Most hematopoietic cells were CD45+ (data not shown). While only a low number of CD45+ c-Kithigh cells exist in each tissue, the vast majority of these hematopoietic progenitor cells were of the Hoxb8-tdTomato lineage, spleen (avg: 78.4%) and thymus (avg: 86.5%) (Figure 2.4F, H). The remaining CD45+ hematopoietic cells were c-Kit- tdTomato+, suggesting that these cells are TRMs as described above. Erythrocytes are present in lymphoid organs such as the spleen and thymus of adult mice. We wondered whether these cells would label with the Hoxb8-tdTomato lineage reporter. The percentages of erythrocytes that were tdTomato+ were low in the adult thymus (avg: 3.28%) and spleen (avg: 1.85%) (Figure 2.4G, I). 2.3.5 Adult white blood cells of the Hoxb8 lineage The Hoxb8-tdTomato lineage reporter has been shown to mark the major immune cell populations in adult mouse blood (16), but the extent to which each of these cell populations is labeled has not been reported. Cells were isolated from adult blood at 2, 3, 49 and 6 months of age and analyzed by flow cytometry. Cells were marked with the CD11b antibody to mark the lymphocyte (CD11b-), monocyte (CD11blo), and granulocyte (CD11bhi) populations (Figure 2.6A, Supplemental Figure 2.7A). In each adult blood cell population examined (i.e., lymphocytes, granulocytes, monocytes), nearly 100% of cells are derived from the Hoxb8 lineage (Figure 2.6B, C). CX3CR1 is expressed on some blood cells, including circulating monocytes and T-cell subpopulations (22, 23). Nearly all Cx3cr1-GFP+ cells detected in the monocyte, granulocyte, and T cell populations were also tdTomato+ except for B cells, with the majority of cells in each cell compartment as GFP- tdTomato+ (Figure 2.7B). However, differences exist with the frequency of GFP+ tdTomato+ cells in each population (T cells: ~5%, monocytes: ~31%, granulocytes: ~53%) (Figure 2.7B). Similarly, a closer examination of T cells (CD11b- B220-) and B cells (CD11b- B220+) of the lymphocyte compartment showed that nearly all cells in both compartments were tdTomato+ (Figure 2.6E, F). Examination of the erythroid lineage in adult mouse blood showed that the percentages of erythrocytes that are tdTomato+ were nearly 100% in the adult (Figure 2.6G). 2.4 Discussion The Hoxb8-tdTomato lineage reporter labels a multitude of hematopoietic cell populations in the developing embryo and adult mouse: yolk sac EMPs, AGM-derived fHSCs, adult HSCs, hematopoietic progenitor cells, erythrocytes, white blood cells, fetal and adult monocytes, fetal macrophages, TRMs, and microglia (16, 17). The variety of immune cells labeled by the Hoxb8-tdTomato reporter indicates that most hematopoieticfated cells are derived from the Hoxb8 lineage, provides further evidence to the current 50 models of tissue-resident macrophage ontogeny, and demonstrates the usefulness of this reporter mouse system to study various immune cell populations, both in the developing embryo and adult mouse. In adult blood, nearly 100% of granulocytes, monocytes, and lymphocytes (i.e., T cells, B cells) are of the Hoxb8 lineage, which strongly indicates that a more in-depth analysis of the different subpopulations of blood cells such as neutrophils, basophils, eosinophils, natural killer cells, and dendritic cells would likely be labeled with the Hoxb8-tdTomato reporter mouse. A plausible explanation as to why nearly all adult immune blood cells are derived from the Hoxb8 lineage is that these cells are generated from bone marrow HSCs (LSK+ cells) that are almost all labeled with the Hoxb8tdTomato reporter (17). Our data show that ~9% of E8.5 yolk sac-derived EMPs are derived from the Hoxb8 lineage. According to current models of hematopoiesis, yolk sac-derived EMPs seed the fetal liver by E9.5, where they rapidly expand (3, 4, 18, 20). Coincidentally, we found a rapid expansion of Hoxb8 EMPs in the fetal liver, where ~91% of the total EMP population was also tdTomato+. These data suggest that either Hoxb8-tdTomato- EMPs become Hoxb8-tdTomato+ by activation of Hoxb8 expression in these cells during expansion in the fetal liver or Hoxb8-tdTomato- EMPs exit the yolk sac, transit to the fetal liver, and proliferate during the early stages of fetal liver hematopoiesis, thereby allowing Hoxb8-tdTomato+ EMPs to become the dominant EMP population in the fetal liver. Some aspects of the current ontological models for TRMs remain to be resolved. The Hoeffel model proposes that fetal monocytic intermediates become tissue-resident 51 macrophages while the Gomez-Perdiguero model argues that fetal monocytes only contribute minimally to the generation of tissue-resident macrophages (2, 4, 12, 15). Also, whether fetal monocytes are generated either from AGM-derived f-HSCs (12) or from E8.5 yolk sac-derived EMPs (4) is also an open question. We found that most fHSCs in the AGM (~88%) and EMPs in the yolk sac (~13%) are derived from the Hoxb8 lineage, which suggests fetal monocytes are derived from 2 sources. According to the Hoeffel model, E8.5 yolk sac EMPs require c-Myb expression once they seed the fetal liver by E9.5 (4, 11). Given that E8.5 Hoxb8 EMPs express a high level of c-Kit and, in c-Myb null embryos, c-Kithigh-expressing cells are eliminated in the fetal liver, it is likely that Hoxb8 EMPs also require c-Myb expression to differentiate into fetal liver-derived monocytes. At E11.5, 1 day prior to the expansion of fetal monocytes, the percentages of yolk sac-derived EMPs and AGM-derived f-HSCs in the fetal liver labeled with the Hoxb8tdTomato reporter are ~48% and ~51%, respectively, and significantly increase by E12.5 (Hoxb8 EMPs: ~73%; Hoxb8 f-HSCs: ~88%). At E12.5 in the fetal liver, ~39% of total monocytes are derived from the Hoxb8 lineage and by E14.5, ~92%. According to the Gomez-Perdiguero model, yolk sac macrophages predominantly become tissue-resident macrophages (13, 15). We show that ~11% of local macrophages in the E9.5 yolk sac are derived from the Hoxb8 lineage and this percentage decreases to <1% by E12.5, which is an 11-fold reduction in the number of Hoxb8 macrophages in the yolk sac from E10.5 to E12.5. These data indicate that either Hoxb8 macrophages are exiting the yolk sac and seeding other developing tissues, or Hoxb8 macrophages in the yolk sac are replaced by fetal liver-derived macrophages that are not of the Hoxb8 lineage. In the E14.5 fetal liver, 52 ~92% of fetal monocytes (F4/80low CD11bhigh) and ~58% of fetal macrophages (F4/80high CD11blow) are tdTomato+. Our data support both models of tissue-resident macrophage ontogeny during embryonic development: yolk sac-derived Hoxb8 EMPs and AGMderived Hoxb8-f-HSCs generate Hoxb8 fetal monocytes. However, it appears that Hoxb8 tissue-resident macrophages are concentrated at active sites of hematopoiesis, whereas non-Hoxb8 macrophages are distributed throughout all organ tissues. Combined with our previous observations regarding bone marrow-derived adult LSK+ HSCs (16, 17), it appears that the Hoxb8 lineage predominantly labels HSPCs and tissue-resident macrophages in active sites of adult hematopoiesis (e.g., bone marrow, spleen, thymus), and very few in non-hematopoietic sites (e.g., lungs, heart, liver). In bone marrow, nearly 100% of LSK+ HSCs are tdTomato+ (17). In the spleen, ~78% of CD45+ c-Kithigh hematopoietic progenitor cells and in the thymus, ~87% of CD45+ cKithigh hematopoietic progenitor cells are derived from the Hoxb8 lineage. These findings are congruent with our previous results that during embryonic development, hematopoietic progenitor cells and f-HSCs, as well as EMPs and fetal monocytes, are largely derived from the Hoxb8 lineage (17). During this time, the AGM and fetal liver are active sites of hematopoiesis in the developing embryo. Interestingly, tissue-resident macrophages of the spleen and thymus are also largely derived from the Hoxb8 lineage. In the thymus, ~54% of tissue-resident macrophages are tdTomato+, and in the spleen, ~90% are Hoxb8-tdTomato+, whereas <1.5% of tissue-resident macrophages are tdTomato+ in the heart, liver, and lungs. Using another Hoxb8 reporter line, Hoxb8Cre-induced lacZ expression, β-gal activity was reported in non-neural tissues (24). For example, lacZ activity was absent in liver and 53 heart, supporting our observations that <1% in liver and heart of tissue-resident macrophages are of the Hoxb8 lineage. The significance of Hoxb8 expression in the progenitors of these hematopoietic cells remains to be revealed. Functional heterogeneity exists across all cell lineages, especially the hematopoietic lineage. Innate immune cell populations such as monocytes, dendritic cells, and lymphocytes each contain subpopulations with an incredible diversity of expression profiles and functions (25, 26). Tissue-resident macrophages represent a diverse population of cells amongst tissues and within a single tissue and their developmental origins appear to contribute to their functional diversity and distribution in a tissue (19, 27-29), e.g., the tissue-resident macrophages of the lungs (29). Debate on the precise ontogeny of this diverse population of cells is ongoing (4, 12, 13). While functional heterogeneity of innate immune cells has been demonstrated, whether the developmental origin(s) and ontogeny contribute to the diverse functions are not known. We recently showed that a subpopulation of resident macrophages of the brain (Hoxb8lineage microglia) have a unique ontogeny and a protective function against obsessive compulsion and anxiety (17, 30). Therefore, it would be conceivable to postulate that Hoxb8 tissue-resident macrophages in tissues of active hematopoiesis could also have protective roles for maintaining a homeostatic balance. 54 2.5 References 1. I. Godin et al., Emergence of multipotent hematopoietic cells in the yolk sac and paraaortic splanchnopleure in mouse embryos, beginning at 8.5 days postcoitus. Proc Natl Acad Sci USA 92, 773-777 (1995). 2. G. Hoeffel et al., Ontogeny of tissue-resident macrophages. Front Immunol 6, 486 (2015). 3. F. Ginhoux et al., Tissue-resident macrophage ontogeny and homeostasis. Immunity 44, 439-449 (2016). 4. G. Hoeffel et al., C-Myb(+) erythro-myeloid progenitor-derived fetal monocytes give rise to adult tissue-resident macrophages. Immunity 42, 665-678 (2015). 5. F. Ginhoux et al., Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841-845 (2010). 6. E. R. Stanley et al., CSF-1 receptor signaling in myeloid cells. Cold Spring Harb Perspect Biol 6, (2014). 7. M. R. Elmore et al., Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking a microglia progenitor cell in the adult brain. Neuron 82, 380-397 (2014). 8. M. L. Mucenski et al., A functional c-Myb gene is required for normal murine fetal hepatic hematopoiesis. Cell 65, 677-689 (1991). 9. C. Schulz et al., A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336, 86-90 (2012). 10. R. Sumner et al., Initiation of adult myelopoiesis can occur in the absence of cMyb whereas subsequent development is strictly dependent on the transcription factor. Oncogene 19, 3335-3342 (2000). 11. K. Kierdorf et al., Microglia emerge from erythromyeloid precursors via Pu.1and Irf8-dependent pathways. Nat Neurosci 16, 273-280 (2013). 12. J. Sheng et al., Most tissue-resident macrophages except microglia are derived from fetal hematopoietic stem cells. Immunity 43, 382-393 (2015). 13. E. Gomez Perdiguero et al., Tissue-resident macrophages originate from yolk-sacderived erythro-myeloid progenitors. Nature 518, 547-551 (2015). 14. K. Kierdorf et al., Development and function of tissue resident macrophages in mice. Semin Immunol 27, 369-378 (2015). 55 15. E. G. Perdiguero et al.,The development and maintenance of resident macrophages. Nat Immunol 17, 2-8 (2016). 16. S. K. Chen et al., Hematopoietic origin of pathological grooming in Hoxb8 mutant mice. Cell 141, 775-785 (2010). 17. S. De et al., Two distinct ontogenies confer heterogeneity to mouse brain microglia. Development 145, 1-14 (2018). 18. K. E. McGrath et al., Distinct sources of hematopoietic progenitors emerge before HSCs and provide functional blood cells in the mammalian embryo. Cell Rep 11, 1892-1904 (2015). 19. G. Hoeffel et al., Adult langerhans cells derive predominantly from embryonic fetal liver monocytes with a minor contribution of yolk sac-derived macrophages. J Exp Med 209, 1167-1181 (2012). 20. G. Hoeffel et al., Fetal monocytes and the origins of tissue-resident macrophages. Cell Immunol 330, 5-15 (2018). 21. F. Geissmann et al., Blood monocytes consist of two principal subsets with distinct migratory properties. Immunity 19, 71-82 (2003). 22. L. Landsman et al., CX3CR1 is required for monocyte homeostasis and atherogenesis by promoting cell survival. Blood 113, 963-972 (2009). 23. M. Lee et al., Tissue-specific role of CX3CR1 expressing immune cells and their relationships with human disease. Immune Netw 18, e5 (2018). 24. R. Witschi et al., Hoxb8-Cre mice: A tool for brain-sparing conditional gene deletion. Genesis 48, 596-602 (2010). 25. S. Yona et al., Monocytes: Subsets, origins, fates and functions. Curr Opin Hematol 17, 53-59 (2010). 26. A. T. Satpathy et al., Re(de)fining the dendritic cell lineage. Nat Immunol 13, 1145-1154 (2012). 27. S. Epelman et al., Embryonic and adult-derived resident cardiac macrophages are maintained through distinct mechanisms at steady state and during inflammation. Immunity 40, 91-104 (2014). 28. J. M. den Haan et al., Innate immune functions of macrophage subpopulations in the spleen. J Innate Immun 4, 437-445 (2012). 29. S. Y. Tan et al., Developmental origin of lung macrophage diversity. Development 143, 1318-1327 (2016). 56 30. D. Trankner et al., A microglia sublineage protects from sex-linked anxiety symptoms and obsessive compulsion. Cell Rep 29, 791-799 e793 (2019). 57 Figure 2.1. EMPs arise from the Hoxb8 lineage. (A) FACS profiles of c-Kithigh CD41+ CD16/32+ cells gated for FSC-A (x-axis) and tdTomato (y-axis) from yolk sac of E8.5 and E9.5 Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. (B) FACS profiles of c-Kithigh CD41+ CD16/32+ cells gated for FSC-A (x-axis) and tdTomato (y-axis) isolated from fetal liver of E11.5 and E14.5 Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. (C) Graph is showing the percentage of Hoxb8 EMPs (c-Kithigh CD41+ CD16/32+ tdTomato+) through development in the yolk sac and fetal liver. (D) Graph is showing the percentage of tdTomato+ cells in the Ter119+ population during development in the yolk sac and fetal liver. n=3 biological replicates or pooled biological replicates per time point; bar represents the mean ± SEM. 58 Figure 2.2. Fetal liver-derived monocytes and fetal macrophages are derived from the Hoxb8 lineage. (A) FACS profiles of CD45+ F4/80low CD11bhigh cells gated for FSCA (x-axis) and tdTomato (y-axis) isolated from fetal liver (E12.5, E14.5, E18.5) of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. (B) FACS profiles of CD45+ F4/80high CD11blow cells gated for FSC-A (x-axis) and tdTomato (y-axis) isolated from yolk sac (E10.5) and fetal liver (E14.5, E18.5) of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. (C) Graph is showing the percentage of Hoxb8 fetal monocytes (CD45+ F4/80low CD11bhigh tdTomato+) through embryonic development in the fetal liver. (D) Graph is showing the percentage of Hoxb8 fetal macrophages (CD45+ F4/80high CD11blow tdTomato+) through embryonic development in the yolk sac and fetal liver. n=3 biological replicates per time point; bar represents the mean ± SEM. 59 Figure 2.3. Developmental dynamics of Hoxb8-lineage fetal monocytes and macrophages. (A) FACS profiles of fetal monocytes (CD45+ F4/80low CD11bhigh) gated for tdTomato (x-axis) and GFP (y-axis) isolated from fetal liver (E12.5, E14.5) of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. (B) FACS profiles of fetal monocytes (CD45+ F4/80low CD11bhigh Ly6C+/-) gated for FSC-A (x-axis) and tdTomato (y-axis) and isolated from fetal liver of E14.5 Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. (F) Graph is showing the percentage of tdTomato+ cells in the Ly6C+ and Ly6C- fractions through development in the fetal liver. (D) FACS profiles of fetal macrophages (CD45+ F4/80high CD11blow) gated for tdTomato (x-axis) and GFP (y-axis) isolated from yolk sac (E12.5) and fetal liver (E12.5, E14.5) of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. n=3 biological replicates per time point; bar represents the mean ± SEM. 60 61 Figure 2.4. Postnatal dynamics of Hoxb8-lineage tissue-resident macrophages. (A-E) FACS profiles of CD45+ F4/80high CD11blow cells gated for tdTomato (x-axis) and GFP (y-axis) isolated from heart, lungs, liver, thymus, and spleen of 2-month- and 6-monthold Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ mice. (F, H) Graphs showing percentages of tdTomato+ cells in the hematopoietic progenitor cell population (CD45+ cKithigh) of adult thymus and spleen. (G, I) Graphs showing the percentages of tdTomato+ cells in the Ter119+ population. n=3 mice per time point, bars represent the mean ± SEM. 62 Figure 2.5. Tissue-resident macrophages, derived from the Hoxb8 lineage, are largely found in active sites of hematopoiesis in the adult mouse. A FACS profiles of CD45+ F4/80high CD64+ Ly6C- cells gated for Cx3cr1-GFP (y-axis) and Hoxb8-tdTomato (x-axis) isolated from heart, lung, liver, thymus, and spleen of 2-month- and 6-month-old Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ mice. B Graph showing percentage of Hoxb8 macrophages (CD45+ F4/80high CD64+ Ly6C- tdTomato+) through postnatal development in adult tissues. n=4-5 mice per time point, bars represent the mean ± SEM. 63 Figure 2.6. All immune cells in peripheral blood are derived from the Hoxb8 lineage. (A) Representative FACS profile of white blood cells (lymphocytes, monocytes, granulocytes) isolated from an adult mouse. (B) FACS profiles of lymphocytes (CD11b-), monocytes (CD11blo), and granulocytes (CD11bhi) gated for FSC-A (x-axis) and tdTomato (y-axis) isolated from blood of 2-, 3-, and 6-month-old Cx3cr1GFP/+; Hoxb8IRESCre/+ ; ROSA26CAG-LSL-tdTomato/+ mice. (C) Graphs showing the percentage of Hoxb8 lymphocytes, monocytes, and granulocytes through postnatal development in adult blood. (D) Representative FACS profile of lymphocytes isolated from an adult mouse and gated for FSC-A (x-axis) and B220 (y-axis). (E) FACS profiles of B cells (CD11b- B220+) and T cells (CD11b- B220-) gated for FSC-A (y-axis) and tdTomato (y-axis). (F) Graphs showing the percentage of tdTomato+ cells in the B cell and T cell populations in Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ adult mice. (G) Graph is showing the percentage of tdTomato+ cells in the Ter119+ population. n=3 mice per time point, bars represent the mean ± SEM. 64 65 Figure 2.7. All Cx3cr1-expressing peripheral immune cells are derived from the Hoxb8 lineage. (B) Graphs showing the relative frequencies of lymphocytes, monocytes, and granulocytes of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ adult mice. (B) FACS profiles of T cells (CD11b- B220-), B cells (CD11b- B220+), monocytes (CD11blow), and granulocytes (CD11bhigh) gated for tdTomato (x-axis) and GFP (y-axis). n=3 mice per time point, bars represent the mean ± SEM. CHAPTER 3 TWO DISTINCT ONTOGENIES CONFER HETEROGENEITY TO MOUSE BRAIN MICROGLIA Reprinted with permission from De*, S., Van Deren*, D., Peden, E., Hockin, M., Boulet, A., Titen, S., Capecchi, M.R. (2018) Two distinct ontogenies confer heterogeneity to brain microglia. Development 145, 1-14. *These authors contributed equally to this work. 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 CHAPTER 4 EFFECT OF A LOSS OF FUNCTION MUTATION OF C-MYB ON HOXB8 MICROGLIA AND HOXB8 HEMATOPOIETIC PROGENITORS Data contribution by Shrutokirti De 89 4.1 Introduction Microglia, the resident macrophages of the brain, are derived from the hematopoietic system during development. As detailed in Chapter 1, the current model for the ontogeny of microglia states that all microglia arise from E7.5 yolk sac EMPs that become yolk sac macrophages, and directly populate the embryonic brain by E9.5 via the developing blood vasculature (1-6). Previous work from our lab suggested the existence of a second population of microglia in the adult mouse brain, identified by using a Hoxb8 lineage reporter mouse (7), potentially having a unique ontogeny. As discussed in Chapter 3, ramified Hoxb8 microglia were first detected in the developing brain parenchyma at ~E12.5 (8), rather than E9.5 for non-Hoxb8-microglia, suggesting that Hoxb8 microglia take a different developmental route that requires transit through other hematopoietic tissues such as the AGM and fetal liver. Flow cytometry analysis using Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos showed that the Hoxb8tdTomato lineage reporter labeled ~5% of immature hematopoietic progenitor cells (CD41+ tdTomato+ c-Kithigh) in the yolk sac from E8.5 to E10.0, whereas >82% of mature hematopoietic progenitor cells (CD45+ tdTomato+ c-Kithigh) in the fetal liver were labeled from E12.5 onward. These data suggest that the Hoxb8-tdTomato lineage reporter marks the tdTomato+ c-Kithigh hematopoietic progenitor cell population responsible for the active hematopoiesis in the fetal liver. Hematopoiesis in the developing mouse embryo is regulated differently in different tissues. c-Myb is a critical transcription factor for HSC development in the fetal liver (9-11) but is dispensable for hematopoiesis in the yolk sac (9). Mice homozygous for a loss-of-function mutation at the c-Myb locus die by E15.5 due to a lack of 90 erythrocyte production; c-Myb null embryos phenotypically exhibit a pale color compared to c-Myb heterozygous and c-Myb wild-type embryos. Schulz et al. utilized cMyb null mouse embryos to demonstrate that c-Myb is not required for the development of the microglial lineage (2), further supporting the yolk sac hypothesis of microglial origin. Flow cytometry analysis was performed on cells isolated from the E10.5 yolk sac to identify mature hematopoietic cells (CD45+) and macrophages (CD45+ CD11b+ F4/80+) as defined by their respective immunophenotype. The percentages of both cell populations were not affected in mice lacking c-Myb. The same method was performed on cells isolated from different tissues of the mouse embryo to identify CD45+ F4/80low CD11bhigh monocytes and CD45+ F4/80high CD11blow macrophages. c-Myb null embryos exhibited a lack of CD45+ F4/80low CD11bhigh monocytes whereas the CD45+ F4/80high CD11blow macrophages were not affected in each of the embryonic tissues examined, indicating that the F4/80low CD11bhigh monocyte population is c-Myb-dependent. Also, the brains of c-Myb null embryos displayed normal numbers of microglia as marked with a mature microglial marker Iba1 antibody (2, 3). Therefore, the absence of c-Myb does not appear to affect the number of yolk sac-derived microglial cells present in the embryonic brain. Conversely, flow cytometry analysis performed on cells isolated from fetal livers (CD11b+ c-Kit+) in c-Myb null embryos revealed an absence of c-Kit+ cells while the number of CD11b+ cells remained normal compared to c-Myb wild-type embryos (2). These data suggest that the absence of c-Myb affects the population of CD45+ F4/80low CD11bhigh monocytes and not the CD45+ F4/80high CD11blow macrophages/microglia. We hypothesized that these hematopoietic progenitor cells in the fetal liver, as well as Hoxb8 microglia, could be affected by a deletion of both alleles at the c-Myb 91 locus for 2 reasons. First, c-Myb is believed to be required for hematopoiesis in the fetal liver (9, 10). Second, the majority of the c-Kithigh hematopoietic progenitor cell population labels with the Hoxb8-tdTomato lineage reporter in the fetal liver by E12.5 (8). Two complimentary genetic methods were used to determine whether there is an effect on the number of embryonic Hoxb8 microglia and Hoxb8 hematopoietic progenitors in the fetal liver. The first used a germline deletion of c-Myb to analyze the Hoxb8 lineage in Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ mouse embryos. The second method was to generate a deletion of c-Myb, with CRISPR/Cas9, early in the developing embryo as a rapid diagnostic tool to examine embryos immediately compared to the traditional breeding experiment. Studying the effect of c-Myb function on Hoxb8 microglia in the developing embryo can be difficult for several reasons. Because c-Myb homozygous null embryos die by E15.5 (9) and the number of Hoxb8 microglia compared to non-Hoxb8 microglia in the E15.5 brain is approximately 4.1% (8), we generated c-Myb chimeric mice on a Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ background to better examine whether c-Myb is required cell autonomously in Hoxb8 microglia in the postnatal brain. 4.2 Experimental procedures 4.2.1 Animals Gt(ROSA)26Sortm14(CAG-tdTomato)Hze (Ai14, #007908) mice were obtained from the Jackson Laboratory. Cx3cr1GFP mice (#005582, Jackson Laboratory) were a kind gift from Dr. Monica Vetter, Department of Neurobiology & Anatomy, University of Utah, Salt Lake City, UT. Hoxb8IRES-Cre mice were generated in our lab and reported in Chen et 92 al. 2010. c-Myb+/- mice (9) were a kind gift from Dr. James Palis, Department of Pediatrics, University of Rochester Medical Center, Rochester, NY. 4.2.2 Embryo isolation Timed matings were set up between male and female mice of the desired genotypes. Plug checks of the female were done daily. The day a female was successfully plugged was marked as Day 0.5 (E0.5) of gestation. On the days of dissection, females were euthanized, and laparotomy was performed to remove the uterine horns containing the embryos. The uterine horns were placed on ice-cold 1X HBSS and embryos with their attached yolk sac were carefully removed from the placenta. Embryonic brains were dissected and processed for immunohistochemistry, and yolk sac, AGM, and fetal liver were dissected and processed for flow cytometry as described below. 4.2.3 Cryosectioning and immunohistochemistry Embryos and adult tissue were incubated in 4% (wt/vol) paraformaldehyde (PFA, EMS 15713) in 1X PBS overnight. Embryonic and adult samples were then kept in 10% sucrose overnight at 4˚C on a shaker, followed by another overnight incubation at 4˚C in 30% sucrose until brain tissues sink to the bottom. Samples were then embedded in Tissue-TekR O.C.T.TM (Sakura 4583), frozen in dry ice and stored at -80˚C. For immunohistochemistry, brain samples were sectioned at 20-25 μm using a Leica CM1900 cryostat and mounted on positively charged microscope slides (FisherbrandTM Tissue Path SuperFrostTM Plus Gold slides, 22-035813 Fisher Scientific). Sections were washed with 1X PBS and permeabilized with 0.2% Triton X-100, 1% Sodium deoxycholate 93 solution, then incubated overnight with primary antibody mixture at 4˚C. The following day the sections were washed and incubated with secondary antibodies for 2 hours at room temperature. Finally, the sections were stained with DAPI (D1306, Molecular Probes) and mounted with ProLong® Gold antifade reagent (P36934 Molecular Probes) and microscope cover glass (FisherbrandTM 22-266882). Images were acquired on the Leica TCS SP5 confocal microscope and processed and analyzed using Imaris 7.7 (Bitplane), as described below. 4.2.4 Antibodies for immunofluorescence detection Primary antibodies used: chicken anti-GFP (1:500, GFP-1020, Aves Labs) and rabbit anti-RFP (1:500, 600-401-379, Rockland). Secondary antibodies used: goat antichicken Alexa Fluor 488 (1:500, A-11039, Thermo Fisher Scientific) and goat anti-rabbit Alexa Fluor 555 (1:500, A-21428, Thermo Fisher Scientific). Both the primary and secondary antibodies that we have used have been tested for specificity and crossreactivity. 4.2.5 Confocal imaging parameters Images were acquired on a Leica TCS SP5 confocal system. For image acquisition, the brain sections were imaged with a 20X objective (0.4 NA, Leica) and 1.5X digital zoom, and images were acquired at 1024 x 1024 resolution and 400-600 Hz scan speed, using a 5.0 µm z-depth through the tissue. For microglial counting, the brain sections were imaged with a 10X objective (0.4 NA, Leica) and 1.5X digital zoom, and images were acquired at 512 x 512 resolution and 400-600 Hz scan speed, using a 5.0 µm 94 z-depth through the tissue. 4.2.6 Imaris image analysis The images acquired with confocal microscopy were processed using Imaris Image Analysis Software x64 (v 7.7.2, Bitplane). The “Spots” function was used to count the number of microglia per unit area (mm2). To further identify the subsets of cells that co-label with another marker (i.e., Cx3cr1-GFP), the spots were filtered using “mean intensities” of the fluorescence of the marker. To quantify the area of the region of analysis, “Surface” function was used. 4.2.7 Embryonic tissue processing for flow cytometry Embryos were removed from the uterus and placed on ice in 5% fetal bovine serum (FBS, Atlanta Biologicals) in 1X Hanks’ balanced salt solution (HBSS, Gibco). Dissections of the following tissues were performed at the time points indicated: yolk sac (E10.0), AGM region (E10.0, E11.5), and fetal liver (E10.0, E11.5, E12.5, E14.5, E15.5). Dissected tissues before E12.5 were pooled together to obtain enough cells for flow cytometry. The tissue was broken up into single cells using a 26G needle. The dissociated tissues were spun down at 1,500 x g for 5 minutes at 4˚C. Cells were washed 3 times with 1X HBSS and spun down at 1,500 x g for 5 minutes. The cells were passed through an 80 µm mesh to obtain a single-cell suspension. 20µl of red blood cell lysis buffer (0.15M ammonium chloride, 1mM KHCO3) was added to the suspension and incubated for 5 minutes at room temperature. Samples were transferred to a 15 mL conical tube and ~13 mL of 1X HBSS was added, and incubation continued for 5 minutes 95 to lyse red blood cells, followed by another 5-minute spin at 200 x g. The anti-mouse antibodies used were a combination of the following: TER-119 PerCP-Cy5.5 (1:50, BioLegend, #116228), CD41 APC-eFluor 780 (1:100, eBioscience, #47-0411-82), CD45 APC (1:100, BioLegend, #103112), and c-Kit PE-Cy7 (1:100, BioLegend, #105814) in 1X HBSS. Cells were incubated in the dark with this antibody cocktail for 30 minutes on ice. Cells were washed with 1X HBSS and spun down at 1,500 x g for 5 minutes, then resuspended in 1X HBSS with DAPI (1:500). Flow cytometry data were obtained using the BD Bioscience FACSCanto II flow cytometry analyzer. All FACS data were analyzed with FlowJo 10.0.7 (Celeza GmbH). 4.2.8 Flow cytometry gating strategy for embryonic tissue populations For all embryonic tissue populations, live cells were gated with side scatter (SSCA) and forward scatter (FSC-A) parameters. Viability and singlets were gated with FSC and DAPI parameters. For yolk sac, AGM, and fetal liver, the TER119- gating parameter excluded platelets, red blood cells, and erythrocytes, and was followed by a selection of either CD41+ only, CD45+ only, or CD41+ CD45+ hematopoietic cells. These cells were further analyzed for tdTomato and c-Kit to examine the percentage of tdTomato+ cells in the hematopoietic progenitor cell population. 96 4.2.9 Generation of c-Myb chimeric mice Timed matings were set up between c-Myb+/-; Hoxb8IRES-Cre/IRES-Cre male and cMyb+/-; Cx3cr1GFP/GFP; ROSA26LSL-CAG-tdTomato/LSL-CAG-tdTomato female mice to generate progeny having the genotypes: c-Myb+/+, c-Myb+/-, or c-Myb-/- on a Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26LSL-CAG-tdTomato/+ background. Carol Lenz (Lab technician specialist, Capecchi Lab, University of Utah) and Sheila Barnett (Lab technician specialist, Capecchi Lab, University of Utah) performed ES cell derivation from mouse blastocysts. Briefly, E3.5 blastocysts were flushed from the uterine horns of females and placed individually into a multi-dish 4-well plate containing an MEF feeder layer and 0.5mL blastocyst medium (high glucose DMEM, 20% FBS, 100X non-essential amino acids, 100X nucleosides, 2000X inhibitor PD0325901, 2000X inhibitor CHIR99021, 1000n/mL LIF, 7.5% NaHCO3). After 1-2 days of culture, embryos hatched from the zona pellucida and attached to the surface of the feeder layer by the migration of the trophoblast cells. Once attached, the inner cell mass (ICM) became evident and proliferated. Approximately 4.5 days after placing blastocysts in culture, well was rinsed with 1X PBS to dislodge the ICM clump using a Pasteur pipette and the clump was transferred to a drop of 0.25% Trypsin and incubated for 5 minutes at 37°C. The ICM clump was disaggregated by gentle pipetting into small clusters of cells (~4-5 cells per cluster). Cells were pipetted in a fresh MEF feeder well containing 1 mL blastocyst medium. At this point, cells were genotyped for c-Myb, then either propagated or frozen down. 15 – 20 ES cells from each clone (2 c-Myb+/+ and 2 c-Myb-/- clones) were microinjected by Rick Focht (Microinjection specialist, Capecchi Lab, University of Utah) and Jim Hayes (Microinjection specialist, Capecchi Lab, University of Utah) into 97 recipient blastocysts derived from C57BL/6J mice to generate chimeras. Approximately ten blastocysts were implanted per uterine horn into CBA/BL6 F1 hybrid pseudopregnant female mice. The spinal cord, kidney, and brain from c-Myb chimeric pups were harvested at P5 for immunohistochemical analysis as described above. The percentage of chimerism was estimated from analysis of the Hoxb8 lineage in the skin, spinal cord, and kidney, regions where Hoxb8 is known to be expressed in embryos and postnatal mice (12-16). c-Myb chimeric pups at P5 were scored by visual inspection of tdTomato expression and patterning on the skin. c-Myb wild-type (c-Myb+/+; Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+) pups at P5 were used as controls and given an arbitrary chimerism score of 5 out of 5 (maximum tdTomato expression in the Hoxb8 lineage). The number of Hoxb8 lineage microglia and non-Hoxb8 microglia was determined by confocal microscopy as described above. 4.2.10 c-Myb mutant generation In order to generate the equivalent of a c-Myb null embryo by CRISPR/Casmediated mutagenesis, 2 guide RNA sequences targeting exon 1 of the c-Myb locus (GTGTCGGGGTCTCCGGGCCA and TGGCCCGGAGACCCCGACAC) were cloned into the pX330-U6-Chimeric_BB-CBh-hSpCas9 plasmid, a gift from Feng Zhang (The Broad Institute, MIT) (Addgene plasmid # 42230). The 2 guide RNA plasmids were mixed at 2.5ng/µL each and injected into the pronucleus of early CX3CR1GFP/+; Hoxb8IRES-Cre/+; Rosa26LSL-CAG-tdTomato/+ or Rosa26CAG-LSL-tdTomato/+; Hoxb8IRES-Cre/+ fertilized eggs. Embryos were allowed to develop to E15.5 for dissection following implantation into a pseudopregnant female recipient. Recovered embryos were imaged 98 with a Leica M205 FA. Brain and fetal liver tissues were processed for immunofluorescence detection flow cytometry, respectively, as described above. DNA was isolated from the tail of each embryo and primers were used to amplify exon 1 (cmyb ex1 fwd - GGCGGCAACCGGTGCGGTCC and c-myb St +211u GAGCCTCAGGGCATGCTTCCCA), and the PCR product was cloned into a TOPO TA cloning vector (Cat# 45-0641, Invitrogen). Sanger sequencing was used to analyze the cMyb locus using the T7 primer. 4.2.11 Statistical analysis Unless otherwise stated, all results are reported as mean ± SEM, and statistical tests were considered significant when p<0.05. Statistical calculations (Unpaired student’s t-test with Welch’s correction, One-way ANOVA with Turkeys multiple comparisons test) were performed with GraphPad Prism 6.0 (GraphPad Software). 4.3 Results 4.3.1 Recapitulating the c-Myb phenotype using CRISPR/Cas9 technology Genetically manipulating each embryo by targeting the c-Myb locus for CRISPR/Cas9 gene editing was used as a rapid diagnostic method to test whether the production of Hoxb8 microglia is c-Myb-dependent. Using CRISPR/Cas9, we can directly modify the genome of the zygote by a Cas9-induced double-stranded-break at the c-Myb locus of both homologs. Cas9-mediated double-strand breaks are often repaired by the imperfect non-homologous end-joining pathway that generates insertions or deletions 99 at the site of cleavage (17). Cas9-injected zygotes are implanted into pseudopregnant female mice and embryos from these females can be analyzed at E14.5 in embryos carrying a Cas9-induced mutation at both alleles. This method dramatically reduces the amount of time needed to generate mutant embryos compared to the conventional crossing method. This Cas9 enzyme was targeted to exon 1 of the c-Myb locus using guide RNAs (sgRNA) (Figure 4.1) to induce a Cas9 double-stranded break on a Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ background. CRISPR/Cas9mediated disruption of c-Myb would, therefore, allow us to determine whether Hoxb8 hematopoietic progenitor cells and Hoxb8 microglia are present in the fetal liver and developing brain, respectively, before E15.5. For this 1-step generation of c-Myb mutants, pronuclear microinjection was performed in zygotes with Cas9 mRNA and guide RNA targeting c-Myb. We anticipated one of two possible outcomes in embryos where both alleles of c-Myb were successfully disrupted. One, if Hoxb8 microglia are not detected in the embryonic brains of CRISPR/Cas9-generated c-Myb mutants, this would suggest that the ontogeny of Hoxb8 microglia is c-Myb-dependent. Two, if Hoxb8 microglia are detected in the embryonic brain of c-Myb mutants, then Hoxb8 microglia development is not dependent on c-Myb. 4.3.2 The gross phenotype of CRISPR/Cas9-targeted c-Myb mutant embryos is comparable to that of c-Myb germline null embryos We compared injected zygotes carrying de novo CRISPR/Cas9-generated c-Myb mutants at E14.5-E15.5 to age-matched embryos homozygous for the previously described c-Myb null allele (9). Figure 4.2 shows the embryos generated from 2 100 independent CRISPR/Cas9 experiments. The c-Myb germline null phenotype was readily detectable in CRISPR/Cas9-generated c-Myb mutants CC6 and CC8 compared to littermate embryos with a wild-type-like phenotype, CC5 and CC7. The CC5 and CC7 embryos were considered wild-type-like because they appeared normal concerning all major arteries and blood vessels, showed no visible skin discoloration, and the size of the embryo was equivalent to age-matched c-Myb wild-type embryos. In agreement with other studies (9) and our observations with c-Myb germline null embryos (described below), c-Myb- embryos were pale in color and appeared anemic due to the failure to generate red blood cells (Figure 4.2). This result suggested that the CRISPR/Cas9 gene editing system could generate embryos with the c-Myb mutant phenotype. 4.3.3 Molecular analyses revealed that CRISPR/Cas9-targeted c-Myb embryos have multiple mutated alleles We used Sanger sequencing to analyze genomic DNA isolated from the tails of CC5, CC6, CC7, and CC8 embryos and examined the molecular nature of the c-Myb alleles in each embryo (Figure 4.3). Following PCR amplification of the target region in each embryo, the amplicon was subcloned into E. coli. The total numbers of DNA clones sequenced for each embryo were as follows: CC5: 27 clones, CC6: 25 clones, CC7: 24 clones, and CC8: 18 clones. Figure 4.3A provides a summary describing the CRISPR/Cas9-induced mutations within exon 1 of c-Myb. Embryo CC5 carried 3 alleles, Allele 1 (frequency of recovery: 40.7%) had a 1 nucleotide insertion resulting in a frameshift mutation, Allele 2 (frequency of recovery: 40.7%) had an 18 nucleotide deletion and a splice mutation, and Allele 3 (frequency of recovery: 18.5%) was wild- 101 type (Figure 4.3B). We also detected 3 alleles in embryo CC7: Allele 1 (frequency of recovery: 50.0%) contained a 51 nucleotide deletion that included the initiation start codon ATG presumably resulting in no protein production, Allele 2 (frequency of recovery: 41.7%) had a 5 nucleotide deletion resulting in a frameshift mutation, and Allele 3 (frequency of recovery: 8.30%) had 3 nucleotides, or 1 amino acid, deletion (Figure 4.3B). CC6 contained at least 4 alleles: Allele 1 (frequency of recovery: 52.0%) had a 1 nucleotide insertion resulting in a frameshift mutation, Allele 2 (frequency of recovery: 32.0%) had a 9 nucleotide deletion that included the ATG start codon resulting in no protein production, Allele 3 (frequency of recovery: 12.0%) was wild-type, and Allele 4 (frequency of recovery: 4.00%) had a 5 nucleotide deletion resulting in a frameshift mutation (Figure 4.3B). Embryo CC8 had at least 5 alleles: Allele 1 (frequency of recovery: 44.4%) had a 1 nucleotide substitution and a 15 nucleotide deletion resulting in a frameshift mutation, Allele 2 (frequency of recovery: 33.3%) had a 46 nucleotide deletion and no ATG start codon resulting in no protein production, Allele 3 (frequency of recovery: 11.1%) had a 1 nucleotide substation and a 5 nucleotide insertion resulting in a frameshift mutation, Allele 4 (frequency of recovery: 5.60%) had a 1 nucleotide substitution, a 5 nucleotide insertion, and a 3 nucleotide or 1 amino acid deletion resulting in a frameshift mutation, and Allele 5 (frequency of recovery: 5.60%) had a 46 nucleotide deletion and no ATG codon resulting in no protein production (Figure 4.3B). Based on the molecular analysis of these embryos, CRISPR/Cas9 generates mosaic organisms with multiple alleles of the targeted locus, c-Myb, with as many as 5 alleles in 1 embryo 102 4.3.4 Fetal livers of CRISPR/Cas9-targeted embryos exhibit a partial elimination of the c-Myb-dependent Hoxb8 hematopoietic progenitor cell population We examined fetal livers of the CRISPR/Cas9 c-Myb--targeted embryos for the loss of tdTomato+ c-Kithigh hematopoietic progenitor cells, similar to the phenotype of cMyb null embryos (described below). Flow cytometry of fetal livers from the CRISPR/Cas9 c-Myb--targeted embryos with a c-Myb-null-like phenotype (CC6, CC8) exhibited a marked reduction in a subset of tdtomato+ CD41+ c-Kithigh, tdtomato+ CD41+ CD45+ c-Kithigh, and tdtomato+ CD45+ c-Kithigh hematopoietic progenitor cell populations relative to the c-Myb-targeted embryos with a wild-type phenotype, CC5 and CC7 (Figure 4.4). This result indicated that CRISPR/Cas9-generated c-Myb- embryos partially recapitulated the mutant c-Myb phenotype previously described in hematopoiesis of the fetal liver (Figure 4.5) (2, 10). 4.3.5 Hoxb8 hematopoietic progenitors in the fetal liver require c-Myb c-Myb null embryos die of severe anemia by E15.5 and appear pale in color due to a lack of red blood cell production (9). We observed similar results when we examined the gross appearance of c-Myb null embryos containing the Hoxb8-IRES-Cre and Rosa26-tdTomato alleles (Figure 4.6). Several studies have shown that there is a lack of hematopoietic progenitor cells in fetal livers of c-Myb null embryos (2, 10) but an examination of these hematopoietic progenitor cells at different stages of maturation during fetal liver development has not been reported. Since Hoxb8-labeled hematopoietic 103 progenitor cells are in abundance in the fetal liver before E15.5, and Hoxb8 microglia are detected in the brain before E15.5 (8) in a wild-type embryo, we were able to determine whether the generation of these cells requires functional c-Myb. Fetal livers of E11.5 to E14.5 c-Myb-/-; Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAGLSL-tdTomato/+ embryos showed a paucity of tdTomato+ c-Kithigh cells in the CD41+ and CD45+ hematopoietic cell populations (Figure 4.5A, B), suggesting that tdTomato+ cKithigh cells are dependent on c-Myb for cellular differentiation and maturation. The lack of hematopoietic cell populations in the fetal livers of E11.5 embryos suggests that either these hematopoietic progenitor cells are unable to seed the fetal liver or are unable to expand and differentiate, possibly due to a failure to upregulate the expression of c-Kit. In c-Myb mutant fetal livers, there is a relatively sizeable residual tdTomato+ cell population, which may have low or transient levels of c-Kit expression (Figure 4.5A, B). Several studies have identified a c-Kitlow cell population in the fetal livers of c-Myb null embryos (2, 10). The presence of this tdTomato+ c-Kitlow cell population suggests that cells migrate and seed the fetal liver, but these hematopoietic progenitor cells lose c-Kit expression and are unable to support hematopoiesis. Previous studies suggest c-Myb is not necessary for the establishment of hematopoiesis in the yolk sac(9, 10) but appear to have a role in the generation of hematopoietic cells (11) and HSPC migration (18) in the P-Sp/AGM region. Cells isolated from c-Myb null embryos from E10.0 yolk sac showed no change in tdTomato+ c-Kithigh hematopoietic progenitors compared to c-Myb wild-type embryos (Figure 4.7), confirming that c-Myb does not appear to be required for hematopoiesis in the yolk sac. Interestingly, in the E11.5 AGM region of c-Myb null embryos, there appeared to be an increase in the tdTomato+ c-Kithigh hematopoietic 104 progenitor population (Figure 4.7). In contrast, tdTomato- c-Kithigh cells are reduced in the E10.0 yolk sac and E11.5 AGM (Figure 4.7), suggesting a subpopulation does require c-Myb. 4.3.6 CRISPR/Cas9-targeted c-Myb mutant embryos display a marked reduction of Hoxb8 microglia Since c-Myb null embryos exhibited a noticeable loss of tdTomato+ c-Kithigh hematopoietic progenitor cells in the fetal liver; we wondered whether the loss of these hematopoietic progenitor cells would be concomitant with the elimination of Hoxb8 microglia in the brain. The number and percentage of Hoxb8 microglia (GFP+ tdTomato+) was reduced in the brains of E14.5-E15.5 CRISPR/Cas9-targeted c-Myb-null-like phenotype embryos (CC6: 0.19±0.05 cells/mm2, 0.56±0.13% ; CC8: 0.65±0.14 cells/mm2, 2.99±0.62%) compared to CRISPR/Cas9-targeted wild-type-looking embryos (CC5: 1.02±0.08 cells/mm2, 3.60±0.35% ; CC7: 2.49±0.22 cells/mm2, 9.09±0.79%) (Figure 4.8A-D). The number and percentage of non-Hoxb8-microglia (GFP+ only) was not affected by CRISPR/Cas9 editing of the c-Myb locus (CC6: 29.29±1.16 cells/mm2, 99.44±0.13% ; CC8: 23.49±1.46 cells/mm2, 97.01±0.62; CC5: 28.38±1.14 cells/mm2, 96.39±0.35% ; CC7: 25.92±1.40 cells/mm2, 90.90±0.79%) (Figure 4.8 A-B, E-F), which was in line with previous reports (2, 3). 105 4.3.7 c-Myb null embryos display a modest reduction of Hoxb8 microglia Prior studies have reported that microglial development is independent of c-Myb function and is therefore independent of fetal liver hematopoiesis (2, 3). The number of Iba1+ microglia in brains of c-Myb null embryos was statistically equivalent in the c-Myb wild-type cohort. In agreement with these studies, we also found no difference in the number of non-Hoxb8 microglia in wild-type and c-Myb null brains at E14.5 (c-Myb wild-type: 22.3±0.89 cells/mm2; c-Myb null: 21.4±0.93 cells/mm2, p=0.637) (Figure 4.9A, C). Although Hoxb8 microglia were detected (Figure 4.9A, B), there is a noticeable and consistent reduction, although not statistically significant, in the number and percentage of Hoxb8 microglia in c-Myb null brains at E14.5 (c-Myb wild-type: 1.81±0.18 cells/mm2; c-Myb null: 0.90±0.08 cells/mm2, p=0.226) (Figure 4.9B). While the non-Hoxb8 microglia population showed no reduction (Figure 4.9C), the number and percentage of Hoxb8 microglia expected to be present in a wild-type brain are relatively low at E14.5, compared to non-Hoxb8 microglia, ~1.5 vs. ~22 microglia/mm2, respectively (Figure 4.9B). Because slight variability associated with the natural development from embryo to embryo within the same uterine horn would have a greater statistical effect on the Hoxb8 microglial population, E14.5, the latest developmental time point that can be measured in c-Myb null embryos, may be too early to reliably measure microglial density (Figure 4.10). To determine whether the slight reduction in the number of Hoxb8 microglia is due to a c-Myb cell-autonomous function, we generated c-Myb chimeric mice. 106 4.3.8 c-Myb may have a cell non-autonomous effect on postnatal Hoxb8 microglia While the generation of yolk sac-derived embryonic microglia does not require cMyb (2, 3), it is not known whether c-Myb is required to generate Hoxb8 microglia or to maintain the population of all postnatal microglia. Because of the embryonic lethality associated with c-Myb loss-of-function, we chose to address this question by producing chimeric mice. Chimeric mice were generated by injecting c-Myb null or c-Myb wild-type embryonic stem (ES) that also carried alleles of CX3CR1GFP, Hoxb8IRES-Cre, and ROSA26CAG-LSL-tdTomato to label all microglia (Cx3cr1-GFP) and Hoxb8 microglia (Cx3cr1GFP, Hoxb8-tdTomato). Cells were injected into wild-type C57BL/6J blastocysts. c-Myb chimeric pups at P5 were scored for tdTomato expression and patterning in skin cells, Hoxb8-tdTomato lineage cells in the kidney, and Hoxb8-tdTomato lineage interneurons in the spinal cord (12-16) (Figure 4.11A, B). P5 c-Myb wild-type and c-Myb null ES cells contributed equally to the generation of kidney cells (Figure 4.11C). When we examined the brains of P5 c-Myb chimeric pups, we observed variable chimerism in the number and percentage of both non-Hoxb8 microglia (~5-15 and ~3-16 microglia/mm2 in mice injected with c-Myb+/+ or c-Myb-/- ES cells, respectively) and Hoxb8 microglia (~0.2-2.3 and ~0.3-1 microglia/mm2 in mice injected with c-Myb+/+ or c-Myb-/- ES cells, respectively)(Figure 4.12A, B). Since c-Myb wild-type and c-Myb null ES cells both contributed to the generation of Hoxb8 microglia, c-Myb may function in a cell non-autonomous manner or may not be required for Hoxb8 microglia development. If c-Myb has a cell autonomous function in Hoxb8 microglia development, then we would expect to find only Cx3Cr1-GFP+ microglia present in chimeric mice, as data 107 suggest these microglia develop independently of c-Myb. Alternatively, if c-Myb has a cell non-autonomous role, or is not required, for Hoxb8 microglia development, then we would expect to find Cx3cr1-GFP+ and Hoxb8 microglia (Cx3cr1-GFP, Hoxb8tdTomato) in the chimeric brain. 4.4 Discussion Based on research from several labs, the hematopoietic progenitors that give rise to microglia are believed to originate from E7.5 EMPs, born in the yolk sac, develop in a c-Myb-independent manner, migrate directly to the brain after exiting the yolk sac, and differentiate, in situ, into tissue-resident macrophages of the brain, microglia (1-3, 5, 6). Two findings support this hypothesis: 1) following the development of the vasculature and initiation of blood circulation, yolk sac-derived progenitors are seen to populate the developing brain of the mouse embryo at E9.5 (1) and 2) fetal liver hematopoiesis, which may be partially c-Myb-dependent, is characterized by the expansion and differentiation of EMPs and HSCs (5, 6, 19), originating from the yolk sac and AGM, respectively, to generate nearly all immune cell types except, it is argued, adult microglia. Thus, the current dogma states that adult microglia are immediate descendants of yolk sac hematopoiesis at E7.5 and are independent of intra-embryonic hematopoiesis, in the AGM and fetal liver. Here we show, in c-Myb null embryos, that the propagation of Hoxb8 hematopoietic progenitor cells in the fetal liver is dependent on c-Myb function. In fact, tdTomato-labeled progenitor cells (tdTomato+ c-Kithigh) in the CD41+, CD41+ CD45+, and CD45+ hematopoietic cell populations are almost eliminated in the absence of c-Myb. 108 Despite the loss of this Hoxb8 progenitor pool in the fetal liver, early colonization of the brain by non-Hoxb8 microglial precursors is unperturbed. These data support the hypothesis that a subpopulation of microglia develops independently of intraembryonic hematopoiesis and c-Myb function. We also saw a modest, but nonsignificant, reduction in Hoxb8 microglia in c-Myb null embryos and CRISPR/Cas9-targeted embryos with a cMyb null-like phenotype. Since our data indicate that the (tdTomato+ c-Kithigh) Hoxb8 progenitor population is lost in c-Myb null embryos and that HoxB8 microglia are derived from the second wave of hematopoiesis through the AGM and fetal liver, we wondered why c-Myb null mutants retain Hoxb8 microglia. We think the most likely explanation is that there are 2 developmental pathways, c-Myb-dependent and c-Myb-independent, for fetal monocytes and tissue-resident macrophages. Additionally, Schulz et al. showed that skin, spleen, pancreas, kidney, and lungs all have macrophages derived from both pathways. There is a relatively sizeable residual tdTomato+ cell population that expresses a low level of c-Kit in the fetal liver of c-Myb null embryos. When we examined fetal livers as early as E10.0, tdTomato+ c-Kithigh cells, in both the CD41+ and CD45+ cell populations, were lost in cMyb null embryos. However, we did detect a tdTomato+ c-Kitlow population in the CD41+, CD41+ CD45+, and CD45+ hematopoietic cell compartments in the fetal liver at all time points analyzed. It has been reported that c-Kitlow HSCs proliferate significantly less than c-Kithigh HSCs but do have increased self-renewal capacity and long-term reconstituting potential (20). Therefore, the Hoxb8 microglial progenitors may be the remaining tdTomato+ c-Kitlow population in c-Myb null embryos. Additionally, a tdTomato+ c-Kithigh population persists in the AGM in c-Myb null 109 mice that could also represent the Hoxb8 microglial progenitor population. Interestingly, in E8.5 EMPs that express c-Myb (6) and transit to the fetal liver (6, 19), the myeloid fate is not affected, whereas the erythroid fate is affected in c-Myb null embryos (2, 4, 10, 21). Thus, EMPs fated to become myeloid cells and potentially microglia are not targeted by a lack of c-Myb function. Also, there may be functional redundancy provided by b-Myb in the absence of c-Myb function. In the CRISPR/Cas9 experiments, some c-Myb-targeted mice have readily visible vasculature and erythrocytes, while some had a pale appearance due to a lack of erythrocytes. CRISPR/Cas9 c-Myb-targeted mice of both phenotypes had multiple alleles generated but non-homologous end-joining repair of the Cas9-induced DNA doublestrand break. Elimination of tdTomato+ c-Kithigh cells in immature (CD41+) and mature (CD45+) hematopoietic cell populations from in c-Myb null embryos is nearly completely recapitulated in targeted c-Myb embryos. Conversely, a more significant reduction of Hoxb8 microglia was observed in the brains of CRISPR/Cas9-targeted disruption of cMyb when compared to the c-Myb wild-type brains. This difference could be a result of off-target mutations caused by the CRISPR/Cas9 gene editing system (22-26). Detecting off-target mutations can be challenging (27), but these mutations could be minimized, for example, either by truncating the sgRNA sequence at the 3’ end to increase target specificity (24, 28, 29) or by altering the amounts of Cas9 and sgRNA that are injected (23-25, 30). A common product of CRISPR/Cas9-mediated gene editing in developing embryos is multiple alleles of the target locus, resulting in a genetically mosaic organism which might also affect the variability of Hoxb8 microglia in the developing brain. Despite the challenges that CRISPR/Cas9 gene editing technology must 110 overcome, this technology is widely used in genome engineering across taxa. Using CRISPR/Cas9, we can directly modify the genome of the zygote by a Cas9-induced double-stranded break at any gene of interest. Injected zygotes are then implanted into pseudo-pregnant females. Analysis can be completed as embryos, in early postnatal stages, as adults, in the soma or germline, or propagate a germline mutation. 1-step generation of mice using the CRISPR/Cas9 system can be used to edit the genome with insertions or deletions at single or multiple genes, and by generating conditional alleles or reporter alleles (31, 32). This 1-step generation of mice has the advantage over conventional gene targeting by directly modifying the genome of a zygote which, in turn, saves money and labor required to generate genetically modified mice. This method can also be used to manipulate ES cells (31). Manipulated ES cells could then be injected, and the chimeric progeny could be studied. 111 4.5 References 1. F. Ginhoux et al., Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841-845 (2010). 2. C. Schulz et al., A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336, 86-90 (2012). 3. K. Kierdorf et al., Microglia emerge from erythromyeloid precursors via Pu.1and Irf8-dependent pathways. Nat Neurosci 16, 273-280 (2013). 4. E. Gomez Perdiguero et al., Tissue-resident macrophages originate from yolk-sacderived erythro-myeloid progenitors. Nature 518, 547-551 (2015). 5. J. Sheng et al., Most tissue-resident macrophages except microglia are derived from fetal hematopoietic stem cells. Immunity 43, 382-393 (2015). 6. G. Hoeffel et al., C-Myb(+) erythro-myeloid progenitor-derived fetal monocytes give rise to adult tissue-resident macrophages. Immunity 42, 665-678 (2015). 7. S. K. Chen et al., Hematopoietic origin of pathological grooming in Hoxb8 mutant mice. Cell 141, 775-785 (2010). 8. S. De et al., Two distinct ontogenies confer heterogeneity to mouse brain microglia. Development 145, 1-14 (2018). 9. M. L. Mucenski et al., A functional c-Myb gene is required for normal murine fetal hepatic hematopoiesis. Cell 65, 677-689 (1991). 10. R. Sumner et al., Initiation of adult myelopoiesis can occur in the absence of cMyb whereas subsequent development is strictly dependent on the transcription factor. Oncogene 19, 3335-3342 (2000). 11. Y. Mukouyama et al., Hematopoietic cells in cultures of the murine embryonic aorta–gonad–mesonephros region are induced by c-Myb. Curr Biol 9, 833-836 (1999). 12. A. Graham et al., The murine Hox-2 genes display dynamic dorsoventral patterns of expression during central nervous system development. Development 112, 255264 (1991). 13. J. C. Holstege et al., Loss of Hoxb8 alters spinal dorsal laminae and sensory responses in mice. Proc Natl Acad Sci USA 105, 6338-6343 (2008). 14. E. van den Akker et al., Targeted inactivation of Hoxb8 affects survival of a spinal ganglion and causes aberrant limb reflexes. Mech Dev 89, 103-114 (1999). 112 15. R. Witschi et al., Hoxb8-Cre mice: A tool for brain-sparing conditional gene deletion. Genesis 48, 596-602 (2010). 16. O. El-Mounayri et al., Regulation of smooth muscle-specific gene expression by homeodomain proteins, Hoxa10 and Hoxb8. J Biol Chem 280, 25854-25863 (2005). 17. P. D. Hsu et al., Development and applications of CRISPR-Cas9 for genome engineering. Cell 157, 1262-1278 (2014). 18. Y. Zhang et al., c-Myb regulates hematopoietic stem/progenitor cell mobilization during zebrafish hematopoiesis. Blood 118, 4093-4101 (2011). 19. K. E. McGrath et al., Distinct sources of hematopoietic progenitors emerge before HSCs and provide functional blood cells in the mammalian embryo. Cell Rep 11, 1892-1904 (2015). 20. J. Y. Shin et al., High c-Kit expression identifies hematopoietic stem cells with impaired self-renewal and megakaryocytic bias. J Exp Med 211, 217-231 (2014). 21. P. Bartunek et al., GATA-1 and c-Myb crosstalk during red blood cell differentiation through GATA-1 binding sites in the c-Myb promoter. Oncogene 22, 1927-1935 (2003). 22. P. Mali et al., CAS9 transcriptional activators for target specificity screening and paired nickases for cooperative genome engineering. Nat Biotechnol 31, 833-838 (2013). 23. P. D. Hsu et al., DNA targeting specificity of RNA-guided Cas9 nucleases. Nat Biotechnol 31, 827-832 (2013). 24. V. Pattanayak et al., High-throughput profiling of off-target DNA cleavage reveals RNA-programmed Cas9 nuclease specificity. Nat Biotechnol 31, 839-843 (2013). 25. Y. Fu et al., High-frequency off-target mutagenesis induced by CRISPR-Cas nucleases in human cells. Nat Biotechnol 31, 822-826 (2013). 26. X. H. Zhang et al., Off-target effects in CRISPR/Cas9-mediated genome engineering. Mol Ther Nucleic Acids 4, e264 (2015). 27. R. Gabriel et al., Mapping the precision of genome editing. Nat Biotechnol 33, 150-152 (2015). 28. S. W. Cho et al., Analysis of off-target effects of CRISPR/Cas-derived RNAguided endonucleases and nickases. Genome Res 24, 132-141 (2014). 113 29. Y. Fu et al., Improving CRISPR-Cas nuclease specificity using truncated guide RNAs. Nat Biotechnol 32, 279-284 (2014). 30. C. Kuscu et al., Genome-wide analysis reveals characteristics of off-target sites bound by the Cas9 endonuclease. Nat Biotechnol 32, 677-683 (2014). 31. H. Wang et al., One-step generation of mice carrying mutations in multiple genes by CRISPR/Cas-mediated genome engineering. Cell 153, 910-918 (2013). 32. H. Yang et al., One-step generation of mice carrying reporter and conditional alleles by CRISPR/Cas-mediated genome engineering. Cell 154, 1370-1379 (2013). 114 Figure 4.1. Targeting strategy of c-Myb by CRISPR/Cas9. Exon 1 of the c-Myb locus was targeted by 2 gRNAs to enhance target specificity. The size of c-Myb is 36 kB with 16 exons. gRNAs are drawn as green arrows. A portion of the sequence of exon 1 is expanded to detail the initiation start codon ATG (highlighted in yellow), c-Myb gRNA 1 (green text), c-Myb gRNA 2 (green text), and the PAM site AGG (blue text). 115 Figure 4.2. Bright-field images of E14.5-15.5 embryos following pronuclear injection with Cas9 and c-Myb-targeting gRNA. CC5 and CC7 embryos (c-Myb+) phenotypically display blood circulation. CC6 and CC8 embryos (c-Myb-) exhibit a lack of circulating erythrocytes resulting in a pale appearance. 116 Figure 4.3. Molecular analysis and summary of CRISPR/Cas9-induced mutations within exon 1 of c-Myb. A Molecular analysis of CC5 – CC8 embryos is showing the number of detected alleles and the frequency of recovery of each allele. Red text denotes exon 1 of c-Myb. The sequences for each allele are compared to the wild-type c-Myb sequence with the frequency of recovery of each allele in parentheses. Exon 1 is highlighted in red and bold. Deletions are denoted by an underscore, insertions are denoted in bold and underlined, and substitutions are denoted in bold. B Description of different alleles isolated from CC5 – CC8 embryos. The table shows a description of each allele present in embryos CC5 – CC8, the likely molecular nature of each allele, the genotype of the embryo, and the observed phenotype based on the micrographs in Figure 4.2. CC5 and CC7 were determined to be wild-type-like (c-Myb+). CC6 was determined to be a hypomorph (c-Myb-) and CC8 as a c-Myb null (c-Myb-). 117 Figure 4.4. Disruption of c-Myb, a transcription factor specific for fetal liver hematopoiesis, by a CRISPR/Cas9-induced mutation, partially eliminates the generation of Hoxb8 hematopoietic progenitors. Flow cytometry profiles of hematopoietic cells (CD41+, CD41+CD45+, CD45+) gated for c-Kit (y-axis) and tdTomato (x-axis) isolated from fetal livers of CC5 - CC8 embryos to identify hematopoietic progenitor cells that are descendants of the Hoxb8 lineage. Red arrows indicate the absent tdTomato+ c-Kithigh cell population in CC6 and CC8 embryos. 118 Figure 4.5. Germline c-Myb null embryos show that fetal liver-derived Hoxb8 hematopoietic progenitors require c-Myb. A FACS profiles of CD41+ cells gated for cKit (y-axis) and tdTomato (x-axis) from fetal liver of E11.5, E12.5, and E14.5 cMyb+/+or c-Myb-/-; Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. B FACS profiles of CD45+ cells gated for c-Kit (y-axis) and tdTomato (x-axis) isolated from fetal liver of E11.5, E12.5, and E14.5 c-Myb+/+or c-Myb-/-; Cx3cr1GFP/+; Hoxb8IRESCre/+ ; ROSA26CAG-LSL-tdTomato/+ embryos. n=2-4 replicates per time point. Red arrows indicate the absent tdTomato+ c-Kithigh cell populations. 119 Figure 4.6. The gross phenotype of Hoxb8-tdTomato reporter embryos lacking cMyb. Bright-field image of E14.5 c-Myb+/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryo displaying blood vasculature and erythrocyte production. Bright-field image of cMyb-/-; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryo appear smaller in size and lack blood vasculature and erythrocyte production resulting in a pale appearance. Hoxb8tdTomato reporter expression on the skin does not appear to be affected in c-Myb null embryos. 120 Figure 4.7. Hoxb8 hematopoietic progenitors in AGM, but not yolk sac, are affected by the disruption of c-Myb. A FACS profiles of CD41+ and CD45+ cells gated for c-Kit (y-axis) and tdTomato (x-axis) from yolk sac of E10.0 c-Myb+/+or c-Myb-/-; Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. B FACS profiles of CD41+ and CD45+ cells gated for c-Kit (y-axis) and tdTomato (x-axis) isolated from AGM of E11.5 cMyb+/+or c-Myb-/-; Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos. n=2-4 replicates per time point. 121 Figure 4.8. Embryonic Hoxb8 microglia are reduced in CRISPR/Cas9-targeted cMyb mutant embryos. Representative micrographs of embryonic brain sections of striatum from A CC5 and B CC6 embryos showing the reduction in the number of Hoxb8 microglia. An in-depth analysis is showing the number and percentage of C-D Hoxb8 and E-F non-Hoxb8-microglial populations in CC5 – CC8 embryonic brains. A-B White asterisks indicate Hoxb8 microglia. Scale bar: 50µm, n=4 slides per embryo (27 – 47 total brain sections imaged per embryo), 2 independent experiments, bars represent the mean ± SEM, ns, non-significant, ****P<0.0001. (Data contribution by Shrutokirti De). 122 Figure 4.9. Embryonic Hoxb8 microglia are slightly reduced in c-Myb null brains. A Representative images of brain sections from E14.5 c-Myb+/+; Cx3cr1GFP/+; Hoxb8IRESCre/+ ; Rosa26CAG-LSL-tdTomato/+ and c-Myb-/-; Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSLtdTomato/+ embryos. Graphs show the number and percentage of B Hoxb8 and C nonHoxb8-microglial populations. Scale bar: 50µm, n=4-5 embryos, bars represent the mean ± SEM, ns, non-significant. (Data contribution by Shrutokirti De). 123 Figure 4.10. Variation in Hoxb8 and non-Hoxb8 microglial densities in c-Myb null brains. Graphs show the number of A Hoxb8 and B non-Hoxb8 microglial populations in c-Myb null embryos (cMg 7, cMg 28, cMg 48, and cMg 53) compared to c-Myb wildtype embryos. n=4-5 embryos, bars represent the mean ± SEM, One-way ANOVA, Turkeys multiple comparisons test, ns, non-significant, *P<0.05, **P<0.005, ****P<0.0001. (Data contribution by Shrutokirti De). 124 Figure 4.11. c-Myb chimerism in P5 pups. A Representative micrographs of P5 pups displaying Hoxb8-tdTomato fluorescence. B Representative images of sections from kidney and spinal cord from P5 c-Myb wild-type chimera pups exhibiting Hoxb8tdTomato fluorescence. C Graph is showing the percentage of chimerism in the kidney from P5 pups. n=3 mice, bars represent the mean ± SEM, ns=not significant. (Data contribution by Shrutokirti De). 125 Figure 4.12. c-Myb may have a cell non-autonomous effect on postnatal Hoxb8 microglia. A Hoxb8 and B non-Hoxb8 microglia in c-Myb wild-type and c-Myb KO chimeric mice at P5. n=5 mice, bars represent the mean ± SEM, ns, non-significant. (Data contribution by Shrutokirti De). CHAPTER 5 ESTABLISHMENT OF A HOXB8 MICROGLIA TRANSPLANTATION MOUSE MODEL Data contribution by Shrutokirti De 127 5.1 Introduction Replacement of microglia through transplantation of myeloid cells, including hematopoietic stem and progenitor cells (HSPCs), can serve as a valuable tool to rectify neurological diseases and neuropsychiatric disorders. Several studies have revealed that this type of myeloid cell-based therapy can have beneficial outcomes in mouse models of neurodegenerative disorders such as amyotrophic lateral sclerosis (1) and Alzheimer’s disease (2), as well as spinal cord injury (3) and pathological grooming (4). These studies, along with others (5, 6), show that myeloid-engrafted cells generate resident myeloid cells, including microglia-like cells (7-13). Microglia have been shown to play critical roles in the development of neurological diseases (14) that involve neuronal tissue damage, neuroinflammation, and oxidative stress (15-18). It is vital to have a rapid and reliable engraftment system with robust delivery to study the viability of replacement therapy in a multitude of neurological diseases. In mice, it has been shown that engraftment efficiency dramatically increases in recipient mice that have been conditioned by ablation of resident immune cells (10, 19); ablation reduces the rejection of donor cells. A recent study has shown that injections of adult bone marrow-derived HSPCs into the lateral ventricles of the brain, of conditioned adult mice, result in the engraftment of microglia-like cells (20). The engrafted cells are morphologically, immunologically, and transcriptionally similar to endogenous microglia, similar to earlier findings that intracerebroventricular delivery of AGMderived HSCs generate microglia-like cells (6). Capotondo et al. were able to achieve engraftment of donor-derived HSPCs to ~40% on day 45 and ~60% on day 180 of total myeloid cells in the brain post-transplantation. 128 Engraftment dynamics of bone marrow-derived Hoxb8 lineage cells suggest that Hoxb8 function is necessary and sufficient to prevent pathological grooming in adult mice (4). Chen et al. showed that transplantation of adult wild-type bone marrow cells isolated from CAG-GFP mice into irradiated adult wild-type recipient mice showed few, ameboid-shaped GFP+ cells in the brain 4 weeks after transplantation. At 12 weeks posttransplantation, a higher number of GFP+ cells with microglial morphology was observed in the adult mouse brain. Interestingly, when mutant Hoxb8 bone marrow was transplanted into wild-type recipient mice, engraftment efficiency was low compared to engraftment of wild-type Hoxb8 bone marrow into mutant Hoxb8 recipient mice. Examination of circulating blood of irradiated wild-type recipient mice 2 months after competitive engraftments of wild-type Hoxb8 and mutant Hoxb8 bone marrow cells, at equal concentrations, revealed that mutant Hoxb8 bone marrow had a significantly lower engraftment efficiency, myeloid cells (~30% compared to ~70% wild-type myeloid cells), and T cells (~20% compared to ~80% wild-type T cells). In the B cell lineage, mutant Hoxb8 bone marrow-derived cells had a significantly higher engraftment level compared to wild-type B cells (~20% compared to ~80% mutant B cells). While the bone marrow transplantation experiments of wild-type bone marrow cells corrected the Hoxb8 mutant phenotype, experimental procedures make it challenging to conclude rescue definitively. There are several limiting factors associated with each of these studies. First, irradiation causes cellular apoptosis resulting from a faulty repair of double-stranded DNA breaks (21, 22) and has been shown to induce microglial proliferation in situ and upregulate cytokines that may permit peripheral immune cells to cross an X-ray-damaged 129 blood-brain barrier and perform functions similar to that of microglia (7, 23-26). As a result, it is possible that peripheral hematopoietic cells cross the blood-brain-barrier, enter the brain, and adopt microglial functions that may impact grooming behavior in adult mice. Second, transplantation of HSPCs takes time: replenishment of myeloid cells derived from HSPC-engrafted cells is slow. Third, which is controversial, is whether engrafted cells become true parenchymal microglia with the same transcriptional identity, properties, and functions of normal, homeostatic microglia (7, 10, 12, 13, 27). As hematopoietic transplantation mouse models have primarily used adult bone marrow as the source for donor-derived myeloid cells and HSPCs, could hematopoietic progenitors, derived from embryonic hematopoietic sources, be used to correct defects in neurological diseases? Vitry et al. demonstrated that postnatal injections of cultured fetalHematopoietic Stem Cells (f-HSCs) into newborn brains resulted in microglia-like cells expressing the macrophage marker F4/80 (6). f-HSCs were isolated from the E10.5 AGM region of transgenic actin-EGFP embryos. These AGM-derived f-HSCs were expanded in culture for 1 week and then injected into 1 of the lateral ventricles of P1-P3 wild-type pups. Brain analysis at 17 days post-injection detected engrafted donor cells with microglia morphology, e.g., ramified processes projecting from the cell soma, near the lateral ventricles. These cells were positive for the macrophage marker F4/80. These results provided strong evidence that AGM-derived f-HSCs, when cultured, can become microglia in the postnatal brain. Another study examined the differentiation potential of yolk sac-derived hematopoietic progenitor cells at different time points in situ (5). Adoptive transfer of these cells on to organotypic hippocampal slice cultures, where endogenous microglia were ablated, revealed that more mature progenitors gave rise to 130 microglia with thin ramified processes versus immature progenitors giving rise to immature microglia with fewer ramified processes. A more recent study from Bennet et al. demonstrated that macrophage populations in the brain that share the same developmental origin with microglia express only microglia-specific genes. Conversely, engrafted HSC-derived cells, although expressing the microglia-specific marker Tmem119, attain a microglia-like identity, and express HSC genes such as Ms4a7, Clec12a, and Apoe (12). Hoxb8 microglia have recently been shown to be bona fide parenchymal microglia (28). RNA sequencing of Hoxb8 microglia shows that the number of transcripts of the microglial-specific marker Tmem119 is equivalent to that of non-Hoxb8 microglia. Immunofluorescence microscopy of Hoxb8 microglia also shows that ~98% of these cells are Tmem119+, compared to ~99% of non-Hoxb8-microglia. Hoxb8 lineage analysis, and Hoxb8 expression analysis, in hematopoietic tissues, suggests that the Hoxb8 microglial progenitor cells are born in the yolk sac, and possibly the AGM, and transit through the fetal liver before colonizing the E12.5 brain. Further, Hoxb8 hematopoietic progenitor cells derived from the ~E11.5-12.5 fetal liver give rise to Hoxb8 microglia. Are Hoxb8 hematopoietic progenitors of the yolk sac, AGM, and primordial fetal liver also capable of becoming Hoxb8 microglia? Furthermore, because a loss of function of Hoxb8 results in overgrooming (29) and the only cells in the brain that are derived from the Hoxb8 lineage are a subset of microglia (4, 28), could there be a therapeutic benefit to replacing dysfunctional Hoxb8 microglia with functional Hoxb8 microglia with respect to pathological grooming? Establishment of a Hoxb8 hematopoietic progenitor transplantation mouse model 131 would be pivotal in the study, and potential rescue, of the Hoxb8 pathological grooming phenotype. This chapter examines engraftment dynamics of Hoxb8 hematopoietic progenitor cells isolated from different hematopoietic tissues and explores alternative, optimized, methods of engraftment of Hoxb8 hematopoietic progenitor cells, which include in utero fetal liver and brain lateral ventricle injections as well as postnatal motor cortex injections. In utero fetal liver engraftment experiments were designed to determine whether E12.5 fetal liver-derived Hoxb8 hematopoietic progenitor cells are a viable cell population for long-term engraftment in various embryonic and postnatal tissues. In utero brain lateral ventricle engraftment experiments examined whether Hoxb8 hematopoietic progenitor cells can take up residency in the postnatal brain. Hoxb8 hematopoietic progenitor cells were injected into the postnatal motor cortices of neonatal recipient mice to determine if these cells are competent to give rise to parenchymal Hoxb8 microglia in recipient mice possessing a germline deletion of the Csf1r or a myeloid-specific Csf1r deletion. 5.2 Experimental procedures 5.2.1 Animals Gt(ROSA)26Sortm14(CAGtdTomato)Hze (Ai14, #007908), Cx3cr1tm1.1(cre)Jung (#025524) , and Csf1rtm1.2Jwp (#021212) mice were obtained from the Jackson Laboratory. Cx3cr1GFP mice (#005582, Jackson Laboratory) were a kind gift from Dr. Monica Vetter, Department of Neurobiology & Anatomy, University of Utah, Salt Lake City, UT. Hoxb8IRES-Cre mice were generated in our lab and reported in Chen et al. 2010. Hoxb8XIRES-Cre and Iba1IRES-Cre mice were generated in our lab by Dr. Eric Peden and Dr. Ben Xu, 132 respectively. 5.2.2 Embryo isolation Timed matings were set up between male and female mice of the desired genotypes. Plug checks of the female were done daily. The day a female was successfully plugged was marked as Day 0.5 (E0.5) of gestation. On the days of dissection, females were euthanized, and laparotomy was performed to remove the uterine horns containing the embryos. The uterine horns were placed on ice-cold 1X HBSS, and embryos with their attached yolk sac were carefully transferred from the placenta. The yolk sac, AGM, and fetal liver were dissected and processed for flow cytometry as described below. 5.2.3 Embryonic and postnatal tissue processing for flow cytometry Embryos were removed from the uterus and placed on ice in 5% fetal bovine serum (FBS, Atlanta Biologicals) in 1X Hanks’ balanced salt solution (HBSS, Gibco). Similarly, postnatal tissue was dissected and placed on ice in 5% fetal bovine serum (FBS, Atlanta Biologicals) in 1X HBSS. Dissections of the following embryonic tissues were performed at their respective time points: yolk sac (E10.5), AGM (E11.5) and fetal liver (E11.5, 12.5, 13.5, 18.5), and brain (E18.5). Dissected tissues before E12.5 were pooled together to obtain enough cells at these time points for flow cytometry. Dissections of bone marrow, liver, lung, and spleen were performed at P14. The tissue was broken up into single cells using a 26G needle for embryonic tissue and a 19G needle for postnatal tissue. The dissociated tissues were spun down at 1,500 x g for 5 minutes at 4˚C. Cells were washed 3 times with 1X HBSS and spun down at 1,500 x g for 5 minutes. 133 The cells were passed through an 80 µm mesh to obtain a single-cell suspension. The suspension was incubated in 20µl of red blood cell lysis buffer (0.15M ammonium chloride, 1mM KHCO3) for 5 minutes at room temperature, then ~13 ml of 1X HBSS was added, and incubation continued for 5 minutes to lyse red blood cells, followed by another 5-minute spin at 200 x g. For embryonic tissues used for cell sorting experiments, the anti-mouse antibodies used consisted of the following: TER-119 PerCP-Cy5.5 (1:50, BioLegend, #116228), and c-Kit PE-Cy7 (1:100, Biolegend, #105814) in 1X HBSS. For embryonic and postnatal tissues used for flow cytometry analysis, TER-119 PerCP-Cy5.5 (1:50, BioLegend, #116228) in 1X HBSS was used. Cells were incubated in the dark with this antibody cocktail for 30 minutes on ice. Cells were washed with 1X HBSS and spun down at 1,500 x g for 5 minutes, then resuspended in 1X HBSS with DAPI (1:1000). Flow cytometry data were obtained using the BD Bioscience FACSCanto II flow cytometry analyzer and BD Bioscience FACS ARIA flow cytometry sorter. All FACS data were analyzed with FlowJo 10.0.7 (Celeza GmbH). 5.2.4 Flow cytometry gating strategy for embryonic and postnatal tissue populations For all embryonic and postnatal tissue populations, live cells were gated with side scatter (SSC-A) and forward scatter (FSC-A) parameters. Viability and singlets were gated with FSC and DAPI parameters. For embryonic tissues used for cell sorting (i.e., yolk sac, AGM, and fetal liver), the TER119- gating parameter excluded platelets, red blood cells, and erythrocytes, and was followed by a selection of tdTomato+ c-Kithigh cells for in utero and postnatal injection experiments. For embryonic and postnatal tissues used 134 for flow cytometry analysis, DAPI- cells were gated with the FSC-A and tdTomato parameters in order to detect tdTomato signal in the tissues examined. To apply the appropriate gates for tdTomato- and tdTomato+ cells, embryonic and postnatal tissues from C57Bl6/J, Rosa26CAG-LSL-tdTomato/CAG-LSL-tdTomato, and Hoxb8IRES-Cre/+; Rosa26CAG-LSLtdTomato/+ were isolated and examined for tdTomato detection. No tdTomato signal was detected in tissues isolated from C57Bl6/J and Rosa26CAG-LSL-tdTomato/CAG-LSL-tdTomato mice. tdTomato signal was detected in tissues isolated from Hoxb8IRES-Cre/+; Rosa26CAG-LSLtdTomato/+ mice. 5.2.5 in utero injections into the fetal liver and lateral ventricles of the developing brain Pregnant female mice were deeply anesthetized with 4% isoflurane within a gas chamber for approximately 5 minutes. Once anesthetized, the animals were transferred to a sterile field table for surgery. Glass capillaries (1.5 mm diameter) were prepared using a micropipette puller (Sutter P-1000 micropipette puller). The micropipette taper was cut to 1-1.5 cm length. Hematopoietic progenitor cells were isolated (described above) from E11.5 and E12.5 fetal livers of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryos and sorted by flow cytometry using the BD Bioscience FACS ARIA flow cytometry sorter. Sorted Hoxb8 hematopoietic progenitor cells consisted of the following immunophenotype: DAPI- TER119- c-Kithigh tdTomato+. 10 –15µL of freshly sorted cells were loaded into the pulled pipettes. A ½ inch incision was made along the midline in the lower abdomen of the pregnant female. The peritoneal cavity is visible through the uterine wall in E12.5 embryos. The micropipette was gently pushed through the uterine 135 wall and peritoneal space, making sure to avoid any visible blood vessels. Once the needle was positioned, ~2K Hoxb8 hematopoietic progenitor cells were injected into the fetal livers of C57Bl6/J embryos (E12.5). Following injections, the embryos were carefully placed back within the body cavity of the female. Fetal livers and brains were harvested from E18.5 embryos and processed for flow cytometry to examine tdTomato+ cells (described above) to examine early engraftment efficiency. For late engraftment efficiency, bone marrow, liver, spleen, and lungs were harvested from P14 mice and processed for flow cytometry to examine tdTomato+ cells (described above). For a separate set of experiments, lateral ventricle brain injections were performed. The lateral ventricles of the developing mouse brain are visible through the uterine wall at E13.5. Freshly sorted Hoxb8 hematopoietic progenitor cells were injected directly into the lateral ventricles (~1K/lateral ventricle) of the brains of C57Bl6/J embryos (E13.5). Brains were harvested, sectioned, and stained with antibodies against tdTomato, GFP, and Tmem119 from P14 mice. 5.2.6 Neonatal injections into motor cortices of the developing mouse brain Neonatal mice (P0-P4) were anesthetized on a sterile field table with 2.5% Isoflurane using an anesthetic vaporizer attached to a tube with a nose piece. Glass capillaries were prepared as described above. Hematopoietic progenitor cells were isolated (described above) from yolk sac (E10.5), AGM (E11.5), and fetal liver (E11.5, 12.5, 13.5) of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryos and sorted by flow cytometry using the BD Bioscience FACS ARIA flow cytometry sorter. Sorted 136 Hoxb8 hematopoietic progenitor cells consisted of the following immunophenotype: DAPI- TER119- c-Kithigh tdTomato+. 10 –15µL of freshly sorted cells were loaded into the pulled pipettes. The micropipette was gently pushed through the skin and developing skull and into the motor cortices, making sure to avoid any visible blood vessels. Once the needle was positioned, freshly sorted Hoxb8 hematopoietic progenitor cells were injected into Csf1r+/+, Csf1rΔ/+, or Csf1rΔ/Δ newborn mice (P0-P4). For yolk sac injections, 300 – 600 cells were injected per hemisphere. For AGM injections, ~12.K cells were injected per hemisphere. For fetal liver injections, 12.5-25K cells were injected per hemisphere. The number of DAPI- TER119- c-Kithigh tdTomato+ cells obtained from yolk sac is significantly smaller compared to the number of cells obtained from AGM and fetal liver. Injections were repeated on each brain hemisphere for each neonatal mouse. Brains from P14-17 were harvested, sectioned, and stained with antibodies against tdTomato, GFP, and Tmem119. 5.2.7 Cryosectioning and immunohistochemistry Adult brain tissue was harvested and incubated in 4% (wt/vol) paraformaldehyde (PFA, EMS 15713) in 1X PBS overnight. Brain tissues were then kept in 10% sucrose overnight at 4˚C on a shaker, followed by another overnight incubation at 4˚C in 30% sucrose until brain tissues sink to the bottom. Samples were then embedded in TissueTekR O.C.T.TM (Sakura 4583), frozen in dry ice and stored at -80˚C. For immunohistochemistry, brain samples were sectioned at 20-25 μm using a Leica CM1900 cryostat and mounted on positively charged microscope slides (FisherbrandTM Tissue Path SuperFrostTM Plus Gold slides, 22-035813 Fisher Scientific). Sections were washed 137 with 1X PBS and permeabilized with 0.2% Triton X-100, 1% Sodium deoxycholate solution, then incubated overnight with primary antibody mixture at 4˚C. The following day, the sections were washed and incubated with secondary antibodies for 2 hours at room temperature. Finally, the sections were stained with DAPI (D1306, Molecular Probes) and mounted with ProLong® Gold antifade reagent (P36934 Molecular Probes) and microscope cover glass (FisherbrandTM 22-266882). Images were acquired on the Leica TCS SP5 confocal microscope and processed using Imaris 7.7 (Bitplane), as described below. 5.2.8 Antibodies for immunofluorescence detection Primary antibodies used: chicken anti-GFP (1:500, GFP-1020, Aves Labs), guinea pig anti-tdTomato (1:500, tdTomato-GP-Af430, Frontier), rabbit anti-Tmem119 (28-3) (1:500, 209064, Abcam), and rabbit anti-Iba1 (1:500, 019-19741, Wako). Secondary antibodies used: goat anti-chicken Alexa Fluor 488 (1:500, A-11039, Thermo Fisher Scientific), goat anti-rabbit Alexa Fluor 555 (1:500, A-21428, Thermo Fisher Scientific), and goat anti-rabbit Alexa Fluor 647 (1:500, 111-605-144, Jackson ImmunoResearch). Both the primary and secondary antibodies that we have used have been tested for specificity and cross-reactivity. 5.2.9 Confocal imaging parameters Images were acquired on a Leica TCS SP5 confocal system. For image acquisition, the brain sections were imaged with either a 20X objective (0.4 NA, Leica) and 1.5X digital zoom, and images were acquired at 1024 x 1024 resolution and 400 Hz 138 scan speed, using a 5.0 µm z-depth through the tissue. For microglial counting, the brain sections were imaged with a 10X objective (0.4 NA, Leica) and 1.5X digital zoom, and images were acquired at 1024 x 1024 resolution and 400 Hz scan speed, using a 5.0 µm z-depth through the tissue. 5.2.10 Imaris image analysis The images acquired from confocal microscopy were processed using Imaris Image Analysis Software ×64 (v 7.7.2, Bitplane). The ‘Spots’ function was used to count the number of microglia per unit area (mm2). To further identify the subsets of cells that co-label with another marker (i.e., tdTomato, GFP), the spots were filtered using ‘mean intensities’ of the fluorescence of the marker. To quantify the area of the region of analysis, ‘Surface’ function was used. 5.2.11 Behavioral testing and analysis The LABORAS behavioral assay (Metris B.V.) is a fully automated system that monitors specific mouse behaviors based on vibration (e.g., grooming, itching, locomotion, eating, drinking, and rest) and has been used in our laboratory (4). On test day, mice were placed in cages on the LABORAS platforms, and their behavior was recorded over a 2-hour testing period. Data were processed and categorized in different behaviors by the LABORAS software. All animal experiments carried out in this study were approved by the Institutional Animal Care and Use Committee of the University of Utah. 139 5.2.12 Statistical analysis Unless otherwise stated, all results are reported as mean ± SEM, and statistical tests were considered significant when p<0.05. Statistical calculations (Student’s t-test) were performed with GraphPad Prism 6.0 (GraphPad Software). 5.3 Results 5.3.1 Hoxb8 hematopoietic progenitors survive engraftment following in utero fetal liver injections Hoxb8 hematopoietic progenitors (c-Kithigh Hoxb8-tdTomato+) were isolated from fetal livers of E12.5 Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryos by flow cytometry. Hoxb8 hematopoietic progenitor cells were injected into E11.5 and E12.5 fetal livers of C57Bl6/J embryos. Low levels of Hoxb8-tdTomato+ cells were detected in the E18.5 fetal liver of both E11.5 and E12.5 recipient embryos (Figure 5.1A). At E18.5, a few Hoxb8-tdTomato+ cells were detected in the brains of the E11.5injected recipients but not in E12.5-injected recipients. (Figure 5.1A). Following engraftment, fetal liver-derived Hoxb8 hematopoietic progenitor cells survived in the recipient fetal liver and brain for at least 6 days. Are these progenitors competent to give rise to resident cells in postnatal tissues such as the lungs, spleen, liver, and bone marrow? Since bone marrow is known to be seeded by cells from the fetal liver and nearly 100% of HSCs (Lin- Sca-1+ c-Kithigh Hoxb8-tdtomato+) in the adult bone marrow are derived from the Hoxb8 lineage (28)(Chapter 3), bone marrow was chosen to determine engraftment efficiency in the postnatal stage. E12.5 fetal liver-derived Hoxb8 hematopoietic progenitor cells were transplanted into E11.5 fetal livers of C57BL/6J 140 embryos and engraftment of Hoxb8-tdTomato+ cells was analyzed at P14. Hoxb8tdTomato+ cells were detected in each of the postnatal tissues examined (Figure 5.1B), except for the postnatal brain (data not shown). Given that the donor cells survived up to 22 days but failed to colonize the postnatal brain after in utero fetal liver injection, we surmised this might be due to the destruction of donor cells by the endogenous peripheral immune system of the recipient. To bypass this issue, in utero brain lateral ventricle injections were done to examine whether c-Kithigh Hoxb8-tdTomato+ cells, when directly placed in the microenvironment of the developing brain, could give rise to Hoxb8 microglia, in vivo. Freshly sorted Hoxb8 hematopoietic progenitor cells were directly injected bilaterally into the lateral ventricles of E13.5 C57Bl6/J embryos. The lateral ventricles of E13.5 embryos are easily identified, and ventricular injections would allow Hoxb8 hematopoietic progenitor cells to populate the entire brain by circulation. Approximately 20 days after injections (P14), Hoxb8-tdTomato+ cells were indeed found in the brain but morphologically appeared to be part of the vascular endothelium (Figure 5.2A). Collectively, these in utero experiments demonstrated that fetal liver-derived Hoxb8 hematopoietic progenitor cells could survive in the endogenous immune system of the recipient, are capable of taking up residence in various postnatal tissues, including the brain, and can survive at least 22 days post engraftment. 141 5.3.2 Postnatal injections of Hoxb8 hematopoietic progenitors Although Hoxb8 hematopoietic progenitor cells were able to engraft in the postnatal brain, they did not become bone fida microglia. A previous study demonstrated that postnatal injections of cultured f-HSCs into newborn brains resulted in microglia-like cells expressing the macrophage marker F4/80 (6). This transplantation method was adapted to perform injections, with E12.5 fetal liver-derived Hoxb8 hematopoietic progenitor cells, bilaterally, directly into the motor cortices of C57Bl6/J newborn pups (P0 – P4, ~1,000 cells/motor cortex). Hoxb8-tdTomato+ cells were detected in the brains of P21 recipient mice. These cells were also Cx3cr1-GFP+ and Tmem119+, suggesting that they were Hoxb8 microglia, despite a difference in morphological appearance (Figure 5.2B). Additionally, engraftment efficiency was low: when ~2,000 cells were injected into newborn brains, only 1 or a few cells were engrafted. Colony-stimulating factor 1 receptor (Csf1r) mutant mice were used as recipients to eliminate the possible destruction of donor cells by eliminating the resident microglia population. Csf1r is critical for microglial survival, and deletion of Csf1r results in the elimination of microglia (30, 31). Hoxb8 hematopoietic progenitor cells were isolated from a range of time points (E10.5 – E13.5) and various hematopoietic tissues (i.e., yolk sac, AGM, and fetal liver). Sorted cells were directly injected bilaterally into the motor cortices of Csf1rΔ/+ and Csf1rΔ/Δ neonatal mice from P0 – P4. Brains were harvested from P14-17 recipient mice. No Hoxb8-tdTomato+ or Cx3cr1-GFP+ signal was detected in the brains of Csf1rΔ/+ mice when injected with E10.5 yolk sac-derived or E11.5 AGMderived Hoxb8 hematopoietic progenitor cells (data not shown). The brains of Csf1rΔ/+ recipient mice injected with E11.5 fetal liver-derived Hoxb8 hematopoietic progenitor 142 cells showed that Hoxb8-tdTomato+ Cx3cr1-GFP+ cells were detected, albeit at small numbers (Figure 5.3). When brains of Csf1rΔ/+ mice were injected with E12.5 or E13.5 fetal liver-derived Hoxb8 hematopoietic progenitor cells, Hoxb8-tdTomato+ Cx3cr1GFP+ microglia were detected at higher numbers with ramified processes (Figure 5.3). These results confirmed that fetal liver-derived Hoxb8 hematopoietic progenitors could become Hoxb8 microglia when isolated from E11.5 - E13.5 fetal livers. In the brains of Csf1rΔ/Δ recipient mice, where resident microglia are not present, a significant number of Hoxb8 microglia were detected following engraftment of Hoxb8 hematopoietic progenitor cells isolated from E11.5, E12.5 (28) (Chapter 3), or E13.5 fetal livers of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryos (Figure 5.4). After successful, robust engraftment of Hoxb8 microglia in Csf1r null mice, we decided to try engraftment following myeloid-specific Csf1r deletion using specific Cre mouse lines (e.g., Cx3cr1-Cre, Iba1-IRES-Cre). These mice would potentially provide suitable “microglia-depleted” recipients to examine the level of long-term engraftment and engraftment efficiency of wild-type Hoxb8 microglia while minimizing any other effects associated with a null mutant of Csf1r. Recipient mice (Iba1IRES-Cre/+; Csf1rfl/Δ or Cx3cr1Cre/+; Csf1rfl/Δ) were injected between P0-P4 with ~25,000 Hoxb8 hematopoietic progenitor cells per motor cortex isolated from E12.5 fetal livers of Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ embryos. Five-week-old recipient mice were engrafted with Hoxb8 microglia (Cx3cr1-GFP+ Hoxb8-tdTomato+) that co-labeled with the microglial marker Tmem119 at the site of injection (Figure 5.5) and, remarkably, throughout the entire brain. The level of engraftment was determined by using the microglial marker Iba1 or Tmem119. The percentage of Iba-1+ cells labeled with Hoxb8- 143 tdTomato and Cx3cr1-GFP signals was nearly 100% in the motor cortex (Figure 5.6). Notably, Tmem119+-only cells were detected in small clusters in the anterior visual cortex (Figure 5.7). These results undoubtedly demonstrate that resident microglia are effectively ablated in 2 myeloid-specific Cre mouse lines, Iba1IRES-Cre and Cx3cr1Cre, and the brain is efficiently engrafted with Hoxb8 microglia from an injection of Hoxb8 progenitor cells from the E12.5 fetal liver. To examine whether there are any associated behavioral consequences, particularly overgrooming from 8- to 12-week-old transplanted recipients, these mice were tested for the amount of time spent grooming over a 2-hour period. The number of seconds spent grooming in Iba1IRES-Cre/IRES-Cre; Csf1rfl/+ and Iba1IRES-Cre/IRES-Cre; Csf1rfl/Δ mice was equivalent to age-matched Iba1+/+; Csf1r+/+ siblings and less than Hoxb8X-IRESCre homozygous mutant mice (Figure 5.8). Hoxb8X-IRES-Cre homozygous mutant mice contain an X-IRES-Cre allele at the Hoxb8 locus. The X-IRES-Cre allele was generated by replacing 6 amino acids in the homeobox DNA-binding domain with 5 alanines and a glutamic acid. These mice exhibit the typical pathological grooming behavior previously reported (29). Overall, the behavioral results show that motor cortex injections of fetal liver-derived Hoxb8 hematopoietic progenitor cells into recipients that lack Csf1r in microglia have no noticeable effect on grooming behavior in adult mice. 5.4 Discussion Hoxb8 hematopoietic progenitor cells isolated from E12.5 fetal livers can be successfully engrafted in the brains of neonatal mice possessing either a germline deletion of Csf1r or conditional deletion of Csf1r in myeloid-specific lineages. Engrafted 144 cells give rise to parenchymal Hoxb8 microglia expressing the microglial markers Tmem119 and Iba1. These results establish a robust and reliable transplantation mouse model that could be used to study mouse behavior, such as pathological grooming, and unlock a potential mechanism between dysfunctional Hoxb8 microglia and the overgrooming phenotype. The in utero and postnatal injection experiments confirmed that Hoxb8 hematopoietic progenitor cells indeed survive and engraft in embryonic (E18.5) and postnatal tissues (P14, P21) of wild-type recipients, although at low numbers. The low level of engraftment is likely due to activation of the recipient’s immune response, thereby destroying engrafted donor cells. To bypass this issue, in utero brain lateral ventricle injections were performed because the ventricles would allow donor-derived cells to circulate throughout the brain. Hoxb8-tdTomato+ cells did engraft throughout the brain, but they did not resemble microglia, lacking microglial morphology and Cx3cr1GFP and Tmem119 expression. Interestingly, engrafted cells appeared to resemble vasculature epithelium. The postnatal injections of fetal liver-derived Hoxb8 hematopoietic progenitor cells led to successful engraftment, as was the case with previous studies using various cell types from different sources (6, 12, 13, 20). Because the low engraftment of Hoxb8 hematopoietic progenitors in C57Bl6/J recipients, likely a consequence of endogenous resident microglia function, high concentrations, ~50,000 Hoxb8 hematopoietic progenitor cells (c-Kithigh Hoxb8-tdTomato+) (~25,000 cells/motor cortex), were injected into the brains of Csf1r heterozygous mutant mice, which led to the successful engraftment of Hoxb8 microglia (28). 145 The use of Csf1r germline deletion mice to eliminate resident microglia has been pivotal in this study. Fetal liver-derived Hoxb8 hematopoietic progenitor cells give rise to Hoxb8 microglia using these mice as recipients (28). Interestingly, engrafted Hoxb8 microglia, in a Csf1r null brain, have fewer ramified processes compared to engrafted Hoxb8 microglia in a Csf1r heterozygote. This result suggests that Hoxb8 microglia, in the Csf1r null brain, have not fully matured due to the lack of extrinsic signaling inputs from the microenvironment of the brain, or these cells are activated because these brains are not at a normal homeostatic level. The brains of Csf1r heterozygous mutant adult mice have an increased density of microglia (32) and, therefore, provide a fertile environment to examine competition of recipient-derived and donor-derived microglial cells to occupy the brain niche. In this environment, the capability for Hoxb8 hematopoietic progenitors to give rise to Hoxb8 microglia appears to occur in the fetal liver beginning at E11.5. An apparent increase in the density of engrafted Hoxb8 microglia is seen around the site of injection when injecting E12.5 and E13.5 fetal liver-derived Hoxb8 hematopoietic progenitors. These data suggest that E11.5 – E13.5 fetal liver-derived Hoxb8 microglial progenitor cells acquire the capability to become Hoxb8 microglia. Our current data suggest that yolk sacderived and AGM-derived Hoxb8 hematopoietic progenitor cells are not competent, as Hoxb8 microglia were not detected in any of the Csf1r heterozygous brains. The number of yolk sac-derived cells injected (~3,000/motor cortex) was limited due to the low number of c-Kithigh Hoxb8-tdTomato+ cells present in the E10.5 yolk sac (~8% of the total c-Kithigh population). However, when equal concentrations of Hoxb8 hematopoietic progenitor cells (~12,500 cells/motor cortex) derived from E11.5 AGM or E12.5 fetal 146 liver were injected into Csf1rΔ/+ mice, Hoxb8 microglia were only detected in recipient brains introduced with E12.5 fetal liver-derived progenitor cells. If the competitive environment is removed in the brain, as in the case with Csf1rΔ/Δ recipients, the capability of Hoxb8 hematopoietic progenitors derived from E11.5 – E13.5 fetal livers is increased, resulting in a remarkably higher number of Hoxb8 microglia distributed throughout the brain. Csf1r homozygous mutants could not be used to examine long-term engraftment efficiency because these mice die by 3 weeks of age due to several developmental issues (e.g., lack of incisor development, dome-shaped skull, kinked tail, truncated limbs, and osteopetrosis) (33). However, the postnatal engraftments of fetal liver-derived Hoxb8 hematopoietic progenitor cells in the brains of Csf1r null mice prompted us to ask whether this transplantation technique could be used to examine long-term engraftment of donor cells in mice that have a conditional loss-of-function of Csf1r in either Iba1+ or Cx3cr1+ myeloid-expressing cells (e.g., microglia). Indeed, Hoxb8 microglia with ramified processes, derived from E12.5 fetal liver progenitor cells, took up residence throughout the entire brains of both Iba1IRES-Cre/+; Csf1rfl/Δ and Cx3cr1Cre/+; Csf1rfl/Δ adult mice. Hoxb8 microglia are also Tmem119+, which confirms they are bona fide parenchymal microglia. Interestingly, recent studies show that intracerebral engraftment of various myeloid populations can result in transcriptomic identities that are different from microglia (12, 13). For example, Bennett et al. demonstrated that fetal liver-derived fetal monocytes could populate the Csf1r null brain and acquire a microglia-like phenotype, including expression of Tmem119. However, at the transcriptome level, these engrafted cells, including HSC-derived microglia-like cells, did not adopt a complete 147 microglial signature. Notably, the expression levels of Sal1 and P2ry12 (microgliaspecific markers) are low whereas the expression levels of Clec12a and Apoe (HSCspecific markers) are high in HSC-derived microglia-like cells, which is in line with other studies (13, 34). Therefore, a complete assessment of gene expression of engrafted Hoxb8 microglia will be of great interest in future studies. Engraftment of Hoxb8 microglia expressing the microglial marker Iba1 in the motor cortex, the injection site, was nearly 100% of total microglia. Notably, Tmem119+only microglia (resident microglia) were detected in the anterior visual cortex (a brain region distant from the site of injection), which shows that some resident microglial cells escaped conditional ablation of Csf1r. Furthermore, engraftment efficiency per injected brain is 100% of brains analyzed to date. Practically, this means we have generated a mouse with ~100% Hoxb8 microglia population rather than ~25% Hoxb8 microglia in a wild-type brain. Because of the high level of engraftment per recipient, we tested 8-weekold Iba1IRES-Cre/IRES-Cre; Csf1rfl/∆ transplanted recipient mice and saw no behavioral effects, particularly the overgrooming phenotype. These data appear to demonstrate that these mice have suffered no lasting neurological damage that affects their behavior as result of transient microglial loss, injection, cell engraftment, or a novel distribution of microglia in the adult mouse brain. We will continue to test these mice for long-term behavioral effects associated with a Hoxb8 microglia brain and then analyze the extent of engraftment afterward. Overall, these experiments unequivocally show that nearly all of the resident brain immune system is capable of being replaced with donor-derived Hoxb8 microglial cells. 148 5.5 References 1. S. Corti et al., Wild-type bone marrow cells ameliorate the phenotype of SOD1G93A ALS mice and contribute to CNS, heart and skeletal muscle tissues. Brain 127, 2518-2532 (2004). 2. A. R. Simard et al., Bone marrow-derived microglia play a critical role in restricting senile plaque formation in Alzheimer's disease. Neuron 49, 489-502 (2006). 3. S. F. Ma et al., Adoptive transfer of M2 macrophages promotes locomotor recovery in adult rats after spinal cord injury. Brain Behav Immun 45, 157-170 (2015). 4. S. K. Chen et al., Hematopoietic origin of pathological grooming in Hoxb8 mutant mice. Cell 141, 775-785 (2010). 5. K. Kierdorf et al., Microglia emerge from erythromyeloid precursors via Pu.1and Irf8-dependent pathways. Nat Neurosci 16, 273-280 (2013). 6. S. Vitry et al., Primordial hematopoietic stem cells generate microglia but not myelin-forming cells in a neural environment. J Neurosci 23, 10724-10731 (2002). 7. B. Ajami et al., Local self-renewal can sustain CNS microglia maintenance and function throughout adult life. Nat Neurosci 10, 1538-1543 (2007). 8. B. Ajami et al., Infiltrating monocytes trigger EAE progression, but do not contribute to the resident microglia pool. Nat Neurosci 14, 1142-1149 (2011). 9. A. Biffi et al., Gene therapy of metachromatic leukodystrophy reverses neurological damage and deficits in mice. J Clin Invest 116, 3070-3082 (2006). 10. A. Capotondo et al., Brain conditioning is instrumental for successful microglia reconstitution following hematopoietic stem cell transplantation. Proc Natl Acad Sci USA 109, 15018-15023 (2012). 11. A. Mildner et al., Microglia in the adult brain arise from Ly-6Chi CCR2+ monocytes only under defined host conditions. Nat Neurosci 10, 1544-1553 (2007). 12. F. C. Bennett et al., A combination of ontogeny and CNS environment establishes microglial identity. Neuron 98, 1170-1183 e1178 (2018). 13. A. Shemer et al., Engrafted parenchymal brain macrophages differ from microglia in transcriptome, chromatin landscape and response to challenge. Nat Commun 9, 5206 (2018). 149 14. J. Han et al., Enforced microglial depletion and repopulation as a promising strategy for the treatment of neurological disorders. Glia 67, 217-231 (2019). 15. R. Wada et al., Microglial activation precedes acute neurodegeneration in Sandhoff disease and is suppressed by bone marrow transplantation. Proc Natl Acad Sci USA 97, 10954-10959 (2000). 16. K. Ohmi et al., Activated microglia in cortex of mouse models of mucopolysaccharidoses I and IIIB. Proc Natl Acad Sci USA 100, 1902-1907 (2003). 17. S. Boillee et al., Onset and progression in inherited ALS determined by motor neurons and microglia. Science 312, 1389-1392 (2006). 18. C. S. Jack et al., TLR signaling tailors innate immune responses in human microglia and astrocytes. J Immunol 175, 4320-4330 (2005). 19. F. L. Wilkinson et al., Busulfan conditioning enhances engraftment of hematopoietic donor-derived cells in the brain compared with irradiation. Mol Ther 21, 868-876 (2013). 20. A. Capotondo et al., Intracerebroventricular delivery of hematopoietic progenitors results in rapid and robust engraftment of microglia-like cells. Sci Adv 3, 1-12 (2017). 21. V. Probin et al., Busulfan selectively induces cellular senescence but not apoptosis in WI38 fibroblasts via a p53-independent but extracellular signalregulated kinase-p38 mitogen-activated protein kinase-dependent mechanism. J Pharmacol Exp Ther 319, 551-560 (2006). 22. A. Meng et al., Ionizing radiation and busulfan inhibit murine bone marrow cell hematopoietic function via apoptosis-dependent and -independent mechanisms. Exp Hematol 31, 1348-1356 (2003). 23. M. Disero et al., Blood-brain barrier permeability after gamma whole-body irradiation: An in vivo microdialysis study. Can J Physiol Pharmacol 80, 670-678 (2002). 24. Y.-Q. Li et al., Early radiation-induced endothelial cell loss and blood–spinal cord barrier breakdown in the rat spinal cord. Radiat Res 161, 143-152 (2004). 25. S. Kyrkanides et al., Cyclooxygenase-2 modulates brain inflammation-related gene expression in central nervous system radiation injury. Mol Brain Res 104, 159-169 (2002). 26. C. Linard et al., Acute induction of inflammatory cytokine expression after γirradiation in the rat: Effect of an NF-κB inhibitor. Int J Radiat Oncol Biol Phys 58, 427-434 (2004). 150 27. M. L. Bennett et al., New tools for studying microglia in the mouse and human CNS. Proc Natl Acad Sci USA 113, E1738-1746 (2016). 28. S. De et al., Two distinct ontogenies confer heterogeneity to mouse brain microglia. Development 145, 1-14 (2018). 29. J. M. Greer et al., Hoxb8 is required for normal grooming behavior in mice. Neuron 33, 23-34 (2002). 30. F. Ginhoux et al., Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841-845 (2010). 31. M. R. Elmore et al., Colony-stimulating factor 1 receptor signaling is necessary for microglia viability, unmasking a microglia progenitor cell in the adult brain. Neuron 82, 380-397 (2014). 32. V. Chitu et al., Phenotypic characterization of a Csf1r haploinsufficient mouse model of adult-onset leukodystrophy with axonal spheroids and pigmented glia (ALSP). Neurobiol Dis 74, 219-228 (2015). 33. X.-M. Dai et al., Targeted disruption of the mouse colony-stimulating factor 1 receptor gene results in osteopetrosis, mononuclear phagocyte deficiency, increased primitive progenitor cell frequencies, and reproductive defects. Blood 99, 111-120 (2002). 34. H. Lund et al., Competitive repopulation of an empty microglial niche yields functionally distinct subsets of microglia-like cells. Nat Commun 9, 4845 (2018). 151 Figure 5.1. Donor fetal liver-derived Hoxb8 hematopoietic progenitors engraft embryonic and postnatal tissues following the in utero injection procedure. A Hoxb8 hematopoietic progenitor cells survive the harvesting, injection, and engraftment steps in embryonic tissues. FACS profiles of DAPI- cells gated for tdTomato (y-axis) from fetal liver and brain of E18.5 C57BL/6J embryos. n=4 embryos B Postnatal tissues show a limited degree of engraftment of fetal liver-derived Hoxb8 hematopoietic progenitor cells. FACS profiles of DAPI- cells gated for tdTomato (y-axis) from bone marrow, liver, spleen, and lungs of P14 C57BL/6J mice. n=4 mice, bars represent the mean ± SEM. 152 Figure 5.2. in utero brain lateral ventricle and postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors engraft in the postnatal brain. A Representative micrograph of a brain section of the substantia nigra of P14 C57BL/6J mice showing an immature microglia-like cell (white asterisk). Scale bar: 100µm, n=4 mice. B Representative micrographs of a brain section of the cortex of P21 C57BL/6J mice showing Hoxb8-tdTomato, Cx3cr1-GFP, and TMEM119 positive fluorescence. LV: lateral ventricle. Scale bar: 50µm, n=5 mice. (Data contribution by Shrutokirti De). 153 Figure 5.3. Postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors engraft in the brains of Csf1rΔ/+ mice. Representative micrographs of brain sections of the motor cortex of P17 Csf1rΔ/+ mice showing parenchymal Hoxb8 microglia (Hoxb8-tdTomato+ Cx3cr1-GFP+) following injections of sorted E11.5, E12.5, and E13.5 fetal liver donor-derived Hoxb8 hematopoietic progenitor cells (c-kithigh Hoxb8-tdTomato+). Scale bar: 60µm, n=5-7 mice. 154 Figure 5.4. Postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors repopulate the brains of Csf1rΔ/Δ mice. Representative micrographs of brain sections of the motor cortex of P17 Csf1rΔ/Δ mice showing parenchymal Hoxb8 microglia (Hoxb8-tdTomato+ Cx3cr1-GFP+) following injections of sorted E11.5 and E13.5 fetal liver donor-derived Hoxb8 hematopoietic progenitor cells (c-kithigh Hoxb8-tdTomato+). Scale bar: 60µm, n=2 mice. 155 Figure 5.5. Postnatal brain injections of donor fetal liver-derived Hoxb8 hematopoietic progenitors engraft in brains of mice that have a conditional loss-offunction of Csf1r in Iba1+ or Cx3cr1+ microglia-expressing cells. Representative micrographs of brain sections of motor cortex of P39 Iba1IRES-Cre/+; Csf1rfl/Δ and Cx3cr1Cre/+; Csf1rfl/Δ mice showing parenchymal Hoxb8 microglia (Hoxb8-tdTomato+ Cx3cr1-GFP+) colocalized with Tmem119 following injections of sorted E12.5fetal liver donor-derived Hoxb8 hematopoietic progenitor cells (c-kithigh Hoxb8-tdTomato+). Scale bar: 60µm, n=4 mice. 156 Figure 5.6. Nearly all Iba1+ microglia in the motor cortex are derived from donor fetal liver-derived Hoxb8 hematopoietic progenitors. Representative micrographs of brain sections of motor cortex of P39 Iba1IRES-Cre/+; Csf1rfl/Δ and Cx3cr1Cre/+; Csf1rfl/Δ mice showing parenchymal Hoxb8 microglia (Hoxb8-tdTomato+ Cx3cr1-GFP+) colocalized with Iba1 following injections of sorted E12.5fetal liver donor-derived Hoxb8 hematopoietic progenitor cells (c-kithigh Hoxb8-tdTomato+). Scale bar: 100µm, n=4 mice. 157 Figure 5.7. Few resident Tmem119+-only microglial cells are detected in recipient brains following engraftment. Representative micrographs of brain sections of the anterior visual cortex of P39 Iba1IRES-Cre/+; Csf1rfl/Δ mice showing parenchymal Hoxb8 microglia (Hoxb8-tdTomato+ Cx3cr1-GFP+) colocalized with Tmem119 and Tmem119+only microglial cells (purple asterisks) following injections of sorted E12.5 fetal liver donor-derived Hoxb8 hematopoietic progenitor cells (c-kithigh Hoxb8-tdTomato+). Scale bar: 100µm, n=4 mice. 158 Figure 5.8. Recipient mice with engrafted wild-type Hoxb8 microglia show no signs of over-grooming. The amount of time spent grooming for 8-week-old mice was assessed over a 2-hour testing period. Iba1-IRES-Cre; Csf1r-flox mice were transplanted with E12.5 fetal liver donor-derived Hoxb8 hematopoietic progenitor cells (c-kithigh Hoxb8-tdTomato+). Iba1IRES-Cre/IRES-Cre; Csf1rfl/+ and Iba1IRES-Cre/IRES-Cre; Csf1rfl/Δ mice groomed an equivalent amount of time to Iba1+/+; Csf1r+/+ but not to Hoxb8X-IRES-Cre/XIRES-Cre mice. n=2-5 mice. CHAPTER 6 CONCLUSIONS ON THE DEVELOPMENTAL ORIGIN OF FUNCTIONAL HETEROGENEITY OF HOXB8 MICROGLIA 160 6.1 Discussion Functional heterogeneity exists across all cell lineages, especially the hematopoietic lineage. Innate immune cell populations such as monocytes, dendritic cells, and lymphocytes each contain subpopulations with an incredible diversity of expression profiles and functions (1, 2). Tissue-resident macrophages represent a diverse population of cells amongst tissues and within a single tissue and their developmental origins appear to contribute to their functional diversity and distribution in a tissue (3-6), e.g., the tissue-resident macrophages of the lungs (6). Debate on the precise ontogeny of this diverse population of cells is ongoing (7-9). While functional heterogeneity of innate immune cells has been demonstrated, whether the developmental origin(s) and ontogeny contribute to the diverse functions is not known. Microglia are the innate immune cells of the brain and have several homeostatic functions. Microglia are involved with phagocytosis of dead/dying neurons (10-12), synaptic pruning of superfluous or unnecessary synapses (13-16), acute and chronic responses to injury by removing cellular debris and promoting wound repair (10, 17), regulation of neural progenitor cells during neurogenesis (18-21), regulation of wiring in the brain (22, 23), and defending the brain against pathogens. The current model in the field posits that there is a single uniform microglia population that originates from a single hematopoietic tissue, the yolk sac, during embryonic development. Since microglia are part of the immune system, these cells arise from hematopoiesis. Briefly, hematopoiesis begins in an extraembryonic tissue called the yolk sac (24). This tissue is the primary source of hematopoietic progenitors from E7.0 – E10.0 (25, 26). Following vascularization and initiation of blood flow at E8.5 via the 161 beating heart, intraembryonic tissues such as the P-Sp/AGM region and fetal liver initiate hematopoiesis. These tissues eventually become the primary sources of hematopoiesis from E10 – E11.5 (AGM) and E11 – E16.5 (fetal liver) (27) in the developing embryo. While it is unclear if cells from the yolk sac seed the P-Sp/AGM, evidence suggests that c-Myb expressing cells do transit to, and colonize, the fetal liver. The AGM is thought to be the hematopoietic source of HSCs (28-33), and the fetal liver is the site of hematopoietic stem and progenitor cells expansion, differentiation, and a primary source for monocytes and macrophages that home to target tissues (34). Although a distinct population of microglia is already present in the developing brain at E9.5, there are other hematopoietic tissues (e.g., AGM region and fetal liver) that represent potential sources for diverse microglia progenitors. A second population of microglia exists in the brain. Hoxb8 null mice are characterized by the behavioral phenotype of pathological grooming (35). Wild-type bone transplant experiments suggested that mutant Hoxb8 microglia may be responsible for the pathological grooming phenotype (36). Hoxb8 lineage analysis uniquely identifies microglia in the brain (36, 37), suggesting that Hoxb8 microglia, when defective, are causative in the disruption of normal adult grooming behavior. The Hoxb8 subpopulation represents ~25% of adult microglia, and the remaining ~75% are non-Hoxb8-microglia (37). De et al., showed that Hoxb8 microglia appear in the developing brain 3 days after non-Hoxb8 microglia are first detected, suggesting the possibility of an alternate ontogeny. Finally, the fact that Hoxb8 dysfunction, in a subpopulation of ~25% of microglia in Hoxb8 KO mice, cannot be corrected by the presence of non-Hoxb8 microglia prompted us to ask: “What is the biological significance of Hoxb8 expression 162 in the lineage that gives rise to Hoxb8 microglia?” Adult mice with a loss-of-function mutation of Hoxb8 display a pathological grooming behavior, similar to that seen in humans with the disease Trichotillomania, exhibiting excessive removal of fur on the ventral side near the forelimbs (35, 36). Behavioral analysis of time spent grooming over a 24-hour period shows that Hoxb8 null mice spend twice as long grooming compared to their wild-type littermates. Since the only cells in the brain that label with the Hoxb8IRES-Cre driver (indicating these cells expressed the Hoxb8 protein any time during their developmental program) are microglia, and microglia are derived from the immune system, bone marrow transplants of wildtype bone marrow cells into lethally irradiated Hoxb8 null mice successfully rescued the behavioral pathology 4 months after transplant. The time spent grooming in these transplanted mice was now equivalent to wild-type controls, and complete restoration of fur was seen on their chests. These remarkable behavioral outcomes suggest that Hoxb8 microglia are different functionally from non-Hoxb8 microglia, suggesting diversity within the microglia population. The body of work presented in this study provides an in-depth examination of the ontogeny of hematopoietic cells derived from the Hoxb8 lineage (i.e., EMPs, f-HSCs, macrophages, monocytes, granulocytes, lymphocytes, adult HSCs, and microglia) and elaborates on the role c-Myb plays in fetal liver hematopoiesis and possible function in Hoxb8 hematopoietic progenitors and Hoxb8 microglia. Cells across the hematopoietic, neuronal, and astrocyte populations are diverse in form and function (1, 2, 6, 38, 39). This diverse array of functions may be conferred upon the cell by signaling from the local microenvironment, the interactions of long-range signaling and cell-specific expression 163 patterns, as well as the tissues though which the cells transit as it matures, differentiation from precursor cell to its definitive state, and the ontogeny of the cell (6). Therefore, in order to understand how a minor subpopulation of microglia is responsible for aberrant organismal behavior, we need to understand the developmental history of microglial cells, as well as other hematopoietic immune cells derived from the Hoxb8 lineage. Understanding how the different developmental programs confer different functions on the 2 populations will shed light on the abnormal function of mutant Hoxb8-microglia that could be causative for pathological grooming. 6.1.1 The ontogeny of immune cells derived from the Hoxb8 lineage Hox genes are known to be involved in hematopoietic processes, mainly regulating self-renewal and differentiation of hematopoietic stem cells at different developmental stages (40-49). Our lab has discovered that Hoxb8 is expressed during the development of a subset of brain microglia and that mice lacking Hoxb8 show pathological grooming in adult mice (35, 36, 50). Other hematopoietic immune cells, e.g., lymphocytes, erythrocytes, and myeloid cells in adult blood, also express Hoxb8 during development, in their lineage, which we can mark with a Hoxb8 reporter (36, 37). Further characterization using this Hoxb8 lineage reporter in the developing mouse embryo and adult mouse, we have found that the majority, if not all, hematopoietic stem and progenitor cells examined (i.e., EMPs, f-HSCs (37), adult LSK+ HSCs (36, 37), CD41+, CD41+ CD45+, and CD45+ hematopoietic progenitor cells (37)) arise from the Hoxb8expressing cells (Chapters 2, 3). These stem and progenitor cell populations are found in 164 active hematopoietic tissues such as the yolk sac, AGM, fetal liver, bone marrow, thymus, and spleen (Chapter 2). Examination of embryonic and adult circulating monocytes show that nearly all these cells are also derived from the Hoxb8 lineage while, at most, ~41% of fetal macrophages in the fetal liver were labeled with the Hoxb8tdTomato lineage reporter (Figure 6.1)(Chapter 2). Tissue-resident macrophages were labeled with this reporter at significant percentages in adult spleen and thymus, both active sites of hematopoiesis, whereas macrophages in liver, lungs, and heart (nonhematopoietic sites in the adult mouse) were labeled at negligible levels (Figure 6.1)(Chapter 2). Collectively, these results show a tendency for hematopoietic stem and progenitor cells, as well as differentiated immune cells (i.e., monocytes, macrophages), located in active tissues of hematopoiesis to express or be descendants of Hoxb8expressing cells. The significance of Hoxb8 expression in the progenitors of these hematopoietic cells remains to be revealed. Based on fate-mapping experiments, using S100a4Cre/eYFP, it has been suggested that most tissue-resident macrophages are derived from fetal monocytes (8). S100a4 (FSP1) is expressed in fetal monocytes (51). Approximately 50-65% of tissue-resident macrophages in the liver, skin, kidneys, and lungs labeled with the S100a4Cre/eYFP fate mapping reporter, whereas only ~20% of microglia were marked. This minimal labeling of microglia may suggest that a separate source or developmental route, dependent on S100a4, could exist for microglia. According to the same model, fetal monocytes are descendants of E8.5 yolk sac-derived EMPs (8), and our data show that a subpopulation of these EMPs are Hoxb8-tdTomato+. Since it has been suggested that fetal monocytes, which differentiate into microglia, are derived from E8.5 EMPs (8, 52) or f-HSCs (7), 165 and there are equal percentages of Hoxb8 lineage cells in both populations (Chapters 2,3), it remains unclear whether E8.5 EMPs or f-HSCs or both are Hoxb8 microglial precursors. A second model suggests that microglia are derived solely from E8.5 EMPs (9, 53). Based on our data, we hypothesize that Hoxb8 hematopoietic progenitors (i.e., E8.5 yolk sac-derived Hoxb8 EMPs or Hoxb8 f-HSCs) develop into fetal monocytes in the fetal liver before becoming tissue-resident macrophages and seeding the embryonic brain at ~E12.5 to become Hoxb8 microglia (Figure 6.2). 6.1.2 Two populations of microglia: Hoxb8 and non-Hoxb8 microglia Our lab generated a triple transgenic mouse model to distinguish and image Hoxb8 and non-Hoxb8 microglial populations in real-time using 3 alleles: Cx3cr1GFP, Hoxb8IRES-Cre, and ROSA26CAG-LSL-tdTomato (37)(Chapter 3). We first tracked entry of both microglial populations (Hoxb8 and non-Hoxb8 microglia) into the brain. At E9.5, the only microglial cells detected in the brain were Cx3cr1-GFP+ only (non-Hoxb8 microglial cells), which agrees with the current dogma that microglia are detected in the brain by this time point. Three days later, at E12.5, Hoxb8 microglia were first detected in the brain. This delay suggests that Hoxb8 microglia may arise from a different source and follow a different developmental route that may require transit through other hematopoietic tissues, such as the AGM and fetal liver. 166 6.1.3 The ontogeny of Hoxb8 microglia As we traced the developmental route of Hoxb8-expressing cells, we found that a subset of hematopoietic progenitor cells expresses Hoxb8 (37)(Chapter 3). Furthermore, the earliest detection of Hoxb8-tdTomato expression in the yolk sac was at E8.5, suggesting that the earliest the Hoxb8 gene could be expressed would be approximately 24 hours earlier. The appearance of Hoxb8 microglia in the developing brain 3 days later than the arrival of non-Hoxb8 microglia suggests that if Hoxb8 hematopoietic progenitors are born in the yolk sac, they are likely transiting through other hematopoietic tissues, such as the AGM and fetal liver, before colonizing the developing brain (Figure 6.3). In the AGM and fetal liver, we observed a rapid amplification of Hoxb8-tdTomato+ hematopoietic progenitor cells, and nearly all hematopoietic progenitor cells are derived from the Hoxb8 lineage. Since Hoxb8 is likely to be expressed in the E7.5 yolk sac and the Hoxb8 lineage is predominantly detected in the AGM and fetal liver (37)(Chapter 3), Hoxb8 microglia may transit through all 3 tissues. Hoxb8 is predominantly expressed in the E8.5 yolk sac, moderately expressed in the AGM, and negligibly expressed in the fetal liver, indicating that Hoxb8 microglia most likely originate in the yolk sac, with a possible contribution from the AGM. In order to determine if non-Hoxb8 microglia express Hoxb8 after colonizing the brain, we examined whether Hoxb8 transcripts could be detected in the developing brain. Hoxb8 gene activity could not be detected at any time point examined, in the head or brain. We still do not definitively know the hematopoietic tissue source (e.g., yolk sac or AGM) and precise developmental route these cells take to enter the brain. 167 Our Hoxb8 gene expression analysis and Hoxb8 lineage mapping of Hoxb8 hematopoietic progenitor cells suggest that the yolk sac and AGM are likely candidate hematopoietic sources of Hoxb8 microglia. While recent studies in mice support the yolk sac as the origin of microglia (7-9, 54-56), the fate-mapping experiments presented in these studies are not able to account for a significant percentage of tissue-resident microglia (as mentioned above). The AGM is a potential source of microglia in vertebrates (57) and is the preceding hematopoietic tissue to the fetal liver (28, 58). In zebrafish, the AGM is a separate hematopoietic source that generates adult microglia. Xu et al. showed that there are 2 distinct sources of microglia, the rostral blood island (equivalent to mouse yolk sac), which generated embryonic microglia, and the ventral wall of the dorsal aorta (equivalent to mouse AGM), which generates adult microglia. Since the hematopoietic system in zebrafish is like that of the mouse (59, 60), the origin of microglia should be revisited in mice. Studies have shown that EMPs from yolk sac seed the fetal liver (8, 26, 61). It is not known whether yolk sac-derived Hoxb8 hematopoietic progenitor cells seed the AGM or are generated de novo in the AGM; however, we do show that these progenitor cells seed the fetal liver. The cis-element enhancer +9.5 is an E box-GATA composite element site in an intron located +9.5 kb downstream of the Gata2 promoter and is critical for HSC emergence from the AGM (62, 63). To test if the AGM is a source of, or is required for, the development of Hoxb8 microglia, embryos carrying a deletion of the cis-element enhancer +9.5 at the Gata2 locus could be used to test whether there is an effect on the number of Hoxb8 hematopoietic progenitor cells and Hoxb8 embryonic microglia present in the developing embryo. If Hoxb8 microglia are not detected in the embryonic brain in the absence of 168 Gata2 +9.5, this would suggest that Hoxb8 microglia depend on the emergence of HSCs in the AGM and that the AGM is the source of Hoxb8 microglia. On the other hand, if Hoxb8 microglia are detected in the embryonic brain in the absence of Gata2 +9.5, this would suggest that these cells do not depend on the emergence of HSCs in the AGM. One more approach to dissect the developmental route Hoxb8 microglial progenitor cells take and determine whether the Hoxb8 hematopoietic progenitors we have identified in the AGM and fetal liver can become Hoxb8 microglia in vivo is to perform cell transplantation experiments. Recently, we have demonstrated that neonatal brain injections of freshly sorted fetal liver-derived Hoxb8 hematopoietic progenitors can give rise to Hoxb8 microglia in vivo (37)(Chapters 3, 5), further supporting the fetal liver as a hematopoietic tissue essential for the development of Hoxb8 microglia. 6.1.4 The development of Hoxb8 microglia does not require c-Myb function We also made multiple attempts to dissect the developmental route of Hoxb8 microglial progenitors by removal of c-Myb function in Cx3cr1GFP/+; Hoxb8IRES-Cre/+; ROSA26CAG-LSL-tdTomato/+ embryos in order to uncover an effect on the number Hoxb8 microglia in the developing brain (Chapter 4). c-Myb is a transcription factor critical for the proper functioning of hematopoiesis in the fetal liver (64-66) but not in the yolk sac (64). The rationale to perform this experiment was that >99% of hematopoietic progenitors were derived from the Hoxb8 lineage (37)(Chapter 3). We predicted that by eliminating fetal liver-derived hematopoietic progenitors expressing c-Myb, Hoxb8 microglia would also be either eliminated or significantly reduced in number. Our results 169 from the 2 complementary genetic experiments (i.e., germline deletion of c-Myb or CRISPR/Cas9-targeted mutation of c-Myb in Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAGLSL-tdTomato/+ mouse embryos) showed that loss-of-function of c-Myb did not significantly influence Hoxb8 microglia development in the embryo, although there was a modest reduction in the number of Hoxb8 microglia (Chapter 4). We also generated c-Myb chimeric mice on a Cx3cr1GFP/+; Hoxb8IRES-Cre/+; Rosa26CAG-LSL-tdTomato/+ mouse background to examine whether c-Myb was required cell autonomously in postnatal Hoxb8 microglia. Again, postnatal Hoxb8 microglial numbers were not affected, which indicated that c-Myb does not have an autonomous cell effect (Chapter 4). Interestingly, in Drosophila, a study has shown that there may be a compensatory mechanism between c-Myb and b-Myb (67), suggesting that similar compensation by b-Myb in the absence of c-Myb in the mouse could be an alternative explanation as to why Hoxb8 microglia, lacking c-Myb, still appear in the brain. 6.1.5 Determining the causal link between dysfunctional Hoxb8 microglia and pathological grooming in mice The mechanism connecting the loss of Hoxb8 function to pathological grooming has not been identified, but the evidence that Hoxb8 microglia play a role in this behavior is beginning to be revealed. The successful engraftment of fetal liver-derived Hoxb8 hematopoietic progenitors into the brains of neonatal mice (i.e., Csf1rΔ/Δ, Iba1IRES-Cre/+; Csf1rfl/Δ, Cx3cr1Cre/+; Csf1rfl/Δ) has primarily established a reliable and robust Hoxb8 microglial transplantation mouse model (Chapter 5). As engraftment efficiency of wildtype Hoxb8 microglia was robust, with nearly all Iba1+ microglia or Tmem119+ microglia 170 donor-derived in the brains of 5-week-old mice, and no apparent behavior abnormalities detected in recipient mice (Chapter 5), we are in a position to address the question: Are dysfunctional Hoxb8 microglia directly causative for pathological grooming in adult mice? Grooming behavior would be tested in 8-week-old recipient mice transplanted with mutant Hoxb8 hematopoietic progenitors, followed by an examination of recipient brains to detect whether there are any abnormalities in microglial morphology, function, and expression. Two possible outcomes of this experiment can be predicted. In the first, pathological grooming is observed in either Cx3cr1Cre/+; Csf1rfl/Δ or Iba1IRES-Cre/+; Csf1rfl/Δ recipient mice when injected with mutant Hoxb8 hematopoietic progenitors, suggesting that mutant Hoxb8 microglia directly cause pathological grooming. In the second, pathological grooming would not be observed in either recipient mouse, suggesting that mutant Hoxb8 microglia do not cause pathological grooming. If mutant Hoxb8 microglia do cause the behavioral phenotype, we could begin to tease out a potential mechanism, such as an inability of mutant Hoxb8 microglia to maintain homeostasis or the lack of a critical function or factor that results in a dysregulation of grooming behavior, by examining the dominance of mutant Hoxb8 microglia or wild-type Hoxb8 microglia in the same mouse. Regardless of the outcome, this experimental avenue would help us determine whether defective Hoxb8 microglia are causative for pathological grooming in adult mice. Another approach to investigate the relationship between loss of Hoxb8 function and pathological grooming and whether microglia play a role in this behavior would be to acquire RNA profiling of mutant Hoxb8 hematopoietic progenitors from mutant Hoxb8 mice during yolk sac hematopoiesis at E8.5. We have shown that Hoxb8 mRNA is 171 detected in the E8.5 yolk sac during yolk sac hematopoiesis at much higher levels than in other hematopoietic tissues examined, such as the AGM and fetal liver (37)(Chapter 3). In vitro studies have proposed that Hoxb8 function during hematopoiesis is no longer required when immature hematopoietic cells differentiate into mature hematopoietic cells (68, 69). Our lab has recently generated a mutant Hoxb8 mouse model to disrupt Hoxb8 function, coined Hoxb8XIRES-Cre/null, that appears to recapitulate the grooming phenotype as previously described (35, 36, 50). These mutant mice could be used to enable RNA profiling from Hoxb8 hematopoietic progenitors isolated from the Hoxb8 mutant E8.5 yolk sac. We would hypothesize that the abnormal function of Hoxb8 microglia, resulting in pathological grooming, may be due to the absence of Hoxb8 protein function in Hoxb8 hematopoietic progenitors generated in the yolk sac. If differentially expressed genes are detected in Hoxb8 hematopoietic progenitor populations in yolk sacs from Hoxb8XIRESCre/null ; ROSA26LSL-CAG-tdTomato/+ embryos compared to age-matched Hoxb8IRES-Cre/+; ROSA26LSL-CAG-tdTomato/+ embryos, genetic mouse models lacking individual differentially expressed candidate genes could be used to test recapitulation or rescue of the pathological overgrooming phenotype. These genetic mouse models would undergo behavioral testing to assess for the excessive grooming phenotype. Microglia are the immune cells of the brain and are critical players in the maintenance of homeostasis in the brain. Dysfunction in microglia has been associated with many neuropsychiatric disorders and diseases, but causation has not been demonstrated. The Hoxb8 mouse models have provided key findings to show that Hoxb8 microglia may have a role in pathological grooming in adult mice (35-37, 50). This study provided a detailed investigation of the origin, ontogeny, and function of Hoxb8 172 microglia (Chapter 3), a possible requirement of c-Myb function (Chapter 4), engraftment dynamics of Hoxb8 hematopoietic progenitors (Chapter 5), as well as the ontogeny of other hematopoietic cells derived from the Hoxb8 lineage (Chapter 2). This body of work provides avenues to explore the inherent characteristics that make mutant Hoxb8 microglia defective, and the steps needed to elucidate how aberrant development and function of Hoxb8 microglia may be causative for pathological grooming. Understanding the role and behavior of Hoxb8 microglia in healthy and diseased brains might shed light onto the mechanisms that drive such neuropsychiatric disorders, and possibly provide answers concerning the pathology of other neurological conditions. 173 6.2 References 1. S. Yona et al., Monocytes: Subsets, origins, fates and functions. Curr Opin Hematol 17, 53-59 (2010). 2. A. T. Satpathy et al., Re(de)fining the dendritic cell lineage. Nat Immunol 13, 1145-1154 (2012). 3. G. Hoeffel et al., Adult langerhans cells derive predominantly from embryonic fetal liver monocytes with a minor contribution of yolk sac-derived macrophages. J Exp Med 209, 1167-1181 (2012). 4. S. Epelman et al., Embryonic and adult-derived resident cardiac macrophages are maintained through distinct mechanisms at steady state and during inflammation. Immunity 40, 91-104 (2014). 5. J. M. den Haan et al., Innate immune functions of macrophage subpopulations in the spleen. J Innate Immun 4, 437-445 (2012). 6. S. Y. Tan et al., Developmental origin of lung macrophage diversity. Development 143, 1318-1327 (2016). 7. J. Sheng et al., Most tissue-resident macrophages except microglia are derived from fetal hematopoietic stem cells. Immunity 43, 382-393 (2015). 8. G. Hoeffel et al., C-Myb(+) erythro-myeloid progenitor-derived fetal monocytes give rise to adult tissue-resident macrophages. Immunity 42, 665-678 (2015). 9. E. Gomez Perdiguero et al., Tissue-resident macrophages originate from yolk sacderived erythro-myeloid progenitors. Nature 518, 547-551 (2015). 10. A. Sierra et al., Microglia shape adult hippocampal neurogenesis through apoptosis-coupled phagocytosis. Cell Stem Cell 7, 483-495 (2010). 11. J. Caldero et al., Development of microglia in the chick embryo spinal cord: Implications in the regulation of motoneuronal survival and death. J Neurosci Res 87, 2447-2466 (2009). 12. K. Ashwell, Microglia and cell death in the developing mouse cerebellum. Brain Res Rev 55, 219-230 (1990). 13. A. Holtmaat et al., Experience-dependent structural synaptic plasticity in the mammalian brain. Nat Rev Neurosci 10, 647-658 (2009). 14. M. E. Tremblay et al., Microglial interactions with synapses are modulated by visual experience. PLoS Biol 8, e1000527 (2010). 174 15. R. C. Paolicelli et al., Synaptic pruning by microglia is necessary for normal brain development. Science 333, 1456-1458 (2011). 16. D. P. Schafer et al., Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron 74, 691-705 (2012). 17. H. Neumann et al.,Debris clearance by microglia: An essential link between degeneration and regeneration. Brain 132, 288-295 (2009). 18. M. Ueno et al., Layer V cortical neurons require microglial support for survival during postnatal development. Nat Neurosci 16, 543-551 (2013). 19. D. Battista et al., Neurogenic niche modulation by activated microglia: Transforming growth factor beta increases neurogenesis in the adult dentate gyrus. Eur J Neurosci 23, 83-93 (2006). 20. A. D. Bachstetter et al., Fractalkine and CX3CR1 regulate hippocampal neurogenesis in adult and aged rats. Neurobiol Aging 32, 2030-2044 (2011). 21. C. L. Cunningham et al., Microglia regulate the number of neural precursor cells in the developing cerebral cortex. J Neurosci 33, 4216-4233 (2013). 22. P. Squarzoni et al., Microglia modulate wiring of the embryonic forebrain. Cell Rep 8, 1271-1279 (2014). 23. Y. Shigemoto-Mogami et al., Microglia enhance neurogenesis and oligodendrogenesis in the early postnatal subventricular zone. J Neurosci 34, 2231-2243 (2014). 24. W. Jollie, Development, morphology, and function of the yolk sac placenta of laboratory rodents. Teratol 41, 361-381 (1990). 25. A. Cumano et al., Ontogeny of the hematopoietic system. Annu Rev Immunol 25, 745-785 (2007). 26. K. E. McGrath et al., Distinct sources of hematopoietic progenitors emerge before HSCs and provide functional blood cells in the mammalian embryo. Cell Rep 11, 1892-1904 (2015). 27. I. Godin et al., Emergence of multipotent hematopoietic cells in the yolk sac and paraaortic splanchnopleure in mouse embryos, beginning at 8.5 days postcoitus. Proc Natl Acad Sci USA 92, 773-777 (1995). 28. M. F. T. R. de Bruijn, N. A. Speck, M. C. E. Peeters, E. Dzierzak, Definitive hematopoietic stem cells first develop within the major arterial regions of the mouse embryo. EMBO J 19, 2465-2474 (2000). 175 29. T. Jaffredo et al., Intraaortic hemopoietic cells are derived from endothelial cells during ontogeny. Development 125, 4575-4583 (1998). 30. A. L. Medvinsky et al., Definitive hematopoiesis is autonomously initiated by the AGM region. Cell 86, 897-906 (1996). 31. R. A. Smith et al., "Hemogenic endothelium" of the embryonic aorta: Does it exist? Dev Comp Immunol 6, 359-368 (1982). 32. M. Tavian et al., Aorta-associated CD34+ hematopoietic cells in the early human embryo. Blood 87, 67-72 (1996). 33. T. Yokomizo et al., Three-dimensional cartography of hematopoietic clusters in the vasculature of whole mouse embryos. Development 137, 3651-3661 (2010). 34. G. R. Johnson et al., Role of stem cell migration in initiation of mouse foetal liver haemopoiesis. Nature 258, 726-728 (1975). 35. J. M. Greer et al., Hoxb8 is required for normal grooming behavior in mice. Neuron 33, 23-34 (2002). 36. S. K. Chen et al., Hematopoietic origin of pathological grooming in Hoxb8 mutant mice. Cell 141, 775-785 (2010). 37. S. De et al., Two distinct ontogenies confer heterogeneity to mouse brain microglia. Development 145, 1-14 (2018). 38. H. Zeng et al., Neuronal cell-type classification: Challenges, opportunities and the path forward. Nat Rev Neurosci 18, 530-546 (2017). 39. L. Ben Haim et al., Functional diversity of astrocytes in neural circuit regulation. Nat Rev Neurosci 18, 31-41 (2017). 40. S. McKinney-Freeman et al., The transcriptional landscape of hematopoietic stem cell ontogeny. Cell Stem Cell 11, 701-714 (2012). 41. C. Blatt et al., DNA rearrangement of a homeobox gene in myeloid leukaemic cells. EMBO J 7, 4283-4290 (1988). 42. K. Kongsuwan et al., Expression of Hox-2.4 homeobox gene directed by proviral insertion in a myeloid leukemia. Nucleic Acids Res 17, 1881-1892 (1989). 43. M. Iacovino et al., HoxA3 is an apical regulator of haemogenic endothelium. Nat Cell Biol 13, 72-78 (2011). 44. B. Schiedlmeier et al., HOXB4’s road map to stem cell expansion. Proc Natl Acad Sci USA 104, 16952-16957 (2007). 176 45. L. M. Forrester et al., Mechanism of action of HOXB4 on the hematopoietic differentiation of embryonic stem cells. Stem Cells 30, 379-385 (2012). 46. R. Morgan et al., Differential expression of HOX genes upon activation of leukocyte sub-populations. Int J Hematol 87, 246-249 (2008). 47. J. Bijl et al., Analysis of HSC activity and compensatory Hox gene expression profile in Hoxb cluster mutant fetal liver cells. Blood 108, 116-122 (2006). 48. A. C. Perkins et al., Conditional immortalization of mouse myelomonocytic, megakaryocytic and mast cell progenitors by the Hox-2.4 homeobox gene. EMBO J 12, 3835-3845 (1993). 49. C. Scholl et al., The homeobox gene CDX2 is aberrantly expressed in most cases of acute myeloid leukemia and promotes leukemogenesis. J Clin Invest 117, 1037-1048 (2007). 50. N. Nagarajan et al., Corticostriatal circuit defects in Hoxb8 mutant mice. Mol Psychiatry 23, 1-10 (2018). 51. D. Hashimoto et al., Tissue-resident macrophages self-maintain locally throughout adult life with minimal contribution from circulating monocytes. Immunity 38, 792-804 (2013). 52. G. Hoeffel et al., Ontogeny of tissue-resident macrophages. Front Immunol 6, 486 (2015). 53. E. G. Perdiguero et al., The development and maintenance of resident macrophages. Nat Immunol 17, 2-8 (2016). 54. F. Ginhoux et al., Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841-845 (2010). 55. K. Kierdorf et al., Microglia emerge from erythromyeloid precursors via Pu.1and Irf8-dependent pathways. Nat Neurosci 16, 273-280 (2013). 56. C. Schulz et al., A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336, 86-90 (2012). 57. J. Xu et al., Temporal-spatial resolution fate mapping reveals distinct origins for embryonic and adult microglia in zebrafish. Dev Cell 34, 632-641 (2015). 58. M.-J. Sanchez et al., Characterization of the first definitive hematopoietic stem cells in the AGM and liver of the mouse. Immunity 5, 513-525 (1996). 59. M. Jagannathan-Bogdan et al., Hematopoiesis. Development 140, 2463-2467 (2013). 177 60. L. Jing et al., Zebrafish as a model for normal and malignant hematopoiesis. Dis Model Mech 4, 433-438 (2011). 61. J. Palis et al., Development of erythroid and myeloid progenitors in the yolk sac and embryo proper of the mouse. Development 126, 5073-5084 (1999). 62. K. D. Johnson et al., Cis-element mutated in GATA2-dependent immunodeficiency governs hematopoiesis and vascular integrity. J Clin Invest 122, 3692-3704 (2012). 63. R. Sanalkumar et al., Mechanism governing a stem cell-generating cis-regulatory element. Proc Natl Acad Sci USA 111, E1091-1100 (2014). 64. M. L. Mucenski et al., A functional c-Myb gene is required for normal murine fetal hepatic hematopoiesis. Cell 65, 677-689 (1991). 65. Y. Mukouyama et al., Hematopoietic cells in cultures of the murine embryonic aorta–gonad–mesonephros region are induced by c-Myb. Curr Biol 9, 833-836 (1999). 66. R. Sumner et al., Initiation of adult myelopoiesis can occur in the absence of cMyb whereas subsequent development is strictly dependent on the transcription factor. Oncogene 19, 3335-3342 (2000). 67. C. J. Davidson et al., Functional evolution of the vertebrate Myb gene family: bMyb, but neither a-Myb nor c-Myb, complements Drosophila Myb in hemocytes. Genetics 169, 215-229 (2005). 68. P. S. Knoepfler et al., HoxB8 requires its Pbx-interaction motif to block differentiation of primary myeloid progenitors and of most cell line models of myeloid differentiation. Oncogene 20, 5440-5448 (2001). 69. M. Rosas et al., Hoxb8 conditionally immortalised macrophage lines model inflammatory monocytic cells with important similarity to dendritic cells. Eur J Immunol 41, 356-365 (2011). 178 Figure 6.1 Innate immune cells derived from the Hoxb8 lineage are located predominantly in areas of hematopoiesis. Illustration showing the tissue locations of Hoxb8 innate immune cells in the adult mouse. Nearly all bone marrow-derived HSCs, circulating blood cells (monocytes, granulocytes, lymphocytes) and splenic macrophages, as well as ~50% of thymic macrophages, are derived from the Hoxb8 lineage. In the brain, ~25% of microglia are of the Hoxb8 lineage. The Hoxb8 lineage represents <12% of macrophages in the liver, lungs, and heart. 179 Figure 6.2 Origin and ontogeny of Hoxb8 tissue-resident macrophages. Illustration showing the possible developmental route of Hoxb8 macrophages during embryonic development. Yolk sac (YS)-derived Hoxb8 EMPs and AGM-derived Hoxb8 f-HSCs transit through the fetal liver (FL) of the embryo proper (EP) where they rapidly proliferate and differentiate into Hoxb8 fetal monocytes. Hoxb8 fetal monocytes exit the fetal liver and further differentiate into Hoxb8 tissue-resident macrophages. Percentages show the proportion of macrophages derived from the Hoxb8 lineage in each tissue examined. 180 Figure 6.3 Origin and ontogeny of non-Hoxb8 and Hoxb8 microglia. Illustrations are showing the developmental routes of non-Hoxb8 and Hoxb8 microglia during embryonic development. Non-Hoxb8 microglial progenitors are born in the E7.5 yolk sac (YS) and directly colonize the E9.5 brain of the embryo proper (EP). Hoxb8 microglial progenitors are born in the E8.5 yolk sac or possibly the AGM, and transit through the fetal liver (FL) before colonizing the E12.5 brain.