Oxygenated Fatty Acids Enhance Hematopoiesis Via the Receptor GPR132

Home , CCRL2, Class C GPCR, Eicosanoid receptor, ELTD1, EMR1, EMR3

Oxygenated Fatty Acids Enhance Hematopoiesis via the Receptor GPR132

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation Lahvic, Jamie L. 2017. Oxygenated Fatty Acids Enhance Hematopoiesis via the Receptor GPR132. Doctoral dissertation, Harvard University, Graduate School of Arts & Sciences.

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:42061504

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA

Oxygenated Fatty Acids Enhance Hematopoiesis via the Receptor GPR132

A dissertation presented

Jamie L. Lahvic

The Division of Medical Sciences

in partial fulfillment of the requirements

for the degree of

Doctor of Philosophy

in the subject of

Developmental and Regenerative Biology

Harvard University

Cambridge, Massachusetts

May 2017

Oxygenated Fatty Acids Enhance Hematopoiesis via the Receptor GPR132

Abstract

After their specification in early development, hematopoietic stem cells (HSCs) maintain the entire blood system throughout adulthood as well as upon transplantation. The processes of HSC specification, renewal, and homing to the niche are regulated by protein, as well as lipid signaling molecules. A screen for chemical enhancers of marrow transplant in the zebrafish identified the endogenous lipid signaling molecule 11,12-epoxyeicosatrienoic acid (11,12-EET). EET has vasodilatory properties, but had no previously described function on HSCs. EET treatment enhanced mouse marrow transplant, and time-lapse imaging showed that this lipid increased HSC specification in zebrafish embryos. These phenotypes were dependent on PI3Kγ signaling as well as AP-1 signaling, which in the zebrafish was specifically required in endothelial cells of the AGM or CHT. EET is known to signal via a G-protein coupled receptor (GPCR), but the identity of this receptor remains unknown, impeding the progress of EET to the clinic and preventing genetic studies of EET signaling. A novel bioinformatic approach identified 10 candidate EET receptors that are expressed in three EET-binding human cell lines, but missing from an EET-non-binding line. Of these, only GPR132 showed EET-responsiveness in vitro. Knockdown of zebrafish gpr132b prevented EET- induced hematopoiesis in the embryo, and marrow from GPR132 KO mice showed a decreased ability to transplant long-term. Others have shown that GPR132 has affinity for a variety of oxygenated fatty acids in vitro. Treatment of zebrafish embryos with these putative GPR132 ligands produced EET-like phenotypes in vivo. Structure-activity-relationship analyses using both in vitro and in vivo assays revealed that a carboxylic acid moiety is required for activity, and oxygenated, unsaturated fatty acids are stronger activators. I have identified GPR132 as an oxygenated fatty acid receptor that mediates both embryonic and adult hematopoiesis. This receptor is a promising target for therapeutic modulation of hematopoiesis and for genetic interrogation of fatty acid signaling. Together these studies reveal the biological and potentially therapeutic importance of lipid signaling molecules and their target GPCRs in regulating stem cell behavior.

iii

Contents

Abstract...... iii

Acknowledgements ...... v

Chapter 1. Introduction ...... 1

Chapter 2. Epoxyeicosatrienoic acids enhance embryonic haematopoiesis and adult marrow

engraftment...... 20

Chapter 3. Oxygenated Fatty Acids Enhance Hematopoiesis via the Receptor GPR132 ...... 54

Chapter 4. GATA Factor-G-Protein-Coupled-Receptor Circuit Suppresses Hematopoiesis ...81

Chapter 5. Discussion ...... 109

References ...... 134

Appendix 1: Zebrafish Microarray...... 165

Appendix 2: GPCR expression levels ...... 167

Appendix 3: Chapter 4 Supplemental Experimental Procedures ...... 172

Appendix 4: Digital Supplemental Files ...... 176

Acknowledgements

This thesis has been many years in the making and all along I’ve been supported by so many. At

Harvard, the BBS office smoothes the many logistical and administrative hurdles, always with a smile. My

DAC was full of helpful members- Stu Orkin for gave insightful comments on hematopoiesis, Patricia

D’Amore was always ready with tough questions for me and for Len. Charles Serhan turned out to be essential for his expert perspective on lipid biochemistry, and I’m grateful to have learned many new techniques in his lab as well. Most of all, Caroline Burns encouraged me at every step of my graduate career; her advice on my science and professional development has always been spot-on.

In the Zon lab, Len has recruited a talented group of positive, collaborative scientists. Ilya

Shestapolov provided an upbeat welcome and mentored my rotation project. Pulin Li gave me a whirlwind introduction to her project, and dazzled me with her scientific precision and sheer efficiency. Vera Binder helped me through my first real experience of publishing a paper. Ellen Durand and Alison Taylor served as wise older grad students. Michelle Lin taught me how to do Western blots and Jon Henninger taught me how to properly CRISPR a zebrafish gene, and they, along with John Gansner, served as cheerful baymates and nearby sources of advice. Julie Perlin and Eva Fast became good friends in the lab who I could turn to for discussions about science or life in general. I’ve surely learned from every single Zon lab member, and I’ve been motivated and inspired by the quality work they present.

The data I present here relies directly on the work of many people. Kara Maloney and Christian

Lawrence kept my fish healthy and thriving. Ron Bernier supported my FACS analysis. Bruce Barut,

Dorothy Giarla, and Hannah DiCicco made sure that reagents arrived and meetings got booked. Two summer students, Michael Chase and Olivia Weiss, produced outstanding work. And I’ve had the honor to work with talented technicians who’ve arrived early, stayed late, and followed my project wherever it leads:

Michelle Ammerman, and Emma Stillman. I’m so grateful to them, and to Megan Blair, especially for performing the mouse procedures that I couldn’t handle.

Lastly, Dr. Len Zon has been an outstanding mentor throughout these years. His advice has been invaluable in designing and executing these projects, and especially in bringing them to the scientific world in the form of papers, presentations, and grants. Len has a deep understanding of stem cell biology that gives him an impressive intuition for explaining results and linking ideas. Len also cares deeply about the

v success of his mentees. He spent dozens of hours with me discussing my future, my goals, and how best to reach them. His encouragement, his confidence in my abilities, and his optimistic, can-do attitude about science helped me through the slow months and the dead ends that are inevitable in research.

Of course, a lot of life happens during grad school, as well, and many people become dear. Elaine biked the whole Cape with me, through sunshine and rain and flat tires. Marc drove halfway down the coast and back with me in a long weekend, just so we could each see our families for Thanksgiving. Deepali and

Evan never hesitated to share their joys or their sorrows. So many other members of my BBS class have been great company- Huixin, Charles, Jennifer, John, Hao, Linda, and many more. I’m grateful too to my fellow DRBers, Adrianna, Olivia, and Tyler. Jamie and Devereux made fabulous roommates and always reminded me of the good old times. My extended circle of friends in the Boston dance and music community, in turn, made sure to remind me that there is life beyond science. I’m grateful to Sharon and Ari for baking, cooking, and plotting the resistance with me. And I’m grateful to Will, who brought me so much light and love, so many hugs and so many new perspectives in this past year.

But Harvard, BBS, and science, none of this would have happened without my family, a constant source of support and encouragement. Mom and Dad raised me with a confidence that I could do anything, and with an outgoing, collaborative spirit that has served me well. They sustained me during my graduate career with phone calls, visits, and shipments of baked goods- they once even flew to Boston with frozen fruit cobbler in their suitcase. Grandpa Segler’s Skype calls were always brief, but frequent- he would light up my evening with smiles and showers of praise. My sister Kelly and I have only grown closer over these years, as our lives and careers have grown farther apart. I’m so lucky to have her as a listening ear. And I was lucky enough to grow up receiving just as much love from my “second family,” the Davies.

Thanks to all of you, and to the many friends farther afield. You’ve made the hard work bearable, you’ve made this thesis possible.

Dedicated to Herbert Steinbeisser

vii

Chapter 1. Introduction

1.1. Hematopoietic Stem Cells

1.1.1. The Prototypical Stem Cell

Blood has fascinated humanity for thousands of years. Diverse cultures have understood it as being vital, even the essence of health and life. But it was only in the late 19th century that scientists came to understand the link between bone marrow and blood. Ernst Neumann described bone marrow as “an important organ for blood formation which has been unrecognized. It operates continually in the de novo formation of red blood cells.” [1] This led to the understanding that blood is a regenerative organ, constantly being replenished even in adulthood. In the 20th century, scientists wondered whether that capacity for replenishment could be transferred from one individual to another. The first bone marrow transplant in human patients was performed by E Donnall Thomas in 1956 [2]. Further studies showed that specific cell populations within the bone marrow- hematopoieic stem cells (HSCs) carry this potential to re-establish the blood system upon transplant. HSC transplants remain our most successful stem cell based treatment today.

In addition to being a breakthrough medical advancement, bone marrow transplant became the seminal tool for the functional evaluation of HSCs. Using this technique, scientists described the defining characteristics of stem cells- a capability for self-renewal and for differentiation into multiple cell types [3].

Later, Richard Schofield proposed a niche hypothesis, whereby neighboring cells or environment would support stem cell function [4]. These concepts have since informed our understanding of stem cells in many solid tissues of the body. HSCs have become the prototypical stem cell; no other stem cell’s identity, behavior, or characteristics have been described in such great detail.

Like most fields of biology, studies of HSCs and their niche have focused on the role of genes and the proteins they encode, neglecting the lipids that make up cell membranes and fat deposits, and also function as signaling molecules. However, recent data has revealed an important role for lipid signaling molecules in regulating stem cell behavior. The discovery of new classes of endogenous lipids such as resolvins and maresins indicates that there exists a complex network of lipids regulating signaling and metabolism in many contexts [5, 6]. In this thesis, I have investigated the role of fatty acid signaling molecules in regulating hematopoietic stem cell specification and engraftment. I have revealed one node

2 of this signaling network, whereby multiple fatty acids can activate one G-protein coupled receptor to influence hematopoiesis.

1.1.2. Hematopoietic Stem Cell Transplant

Allogeneic hematopoietic stem cell transplant is a potentially curative and life-saving treatment performed on thousands of patients yearly suffering from malignant leukemias and lymphomas, as well as non-malignant blood disorders such as major anemias [7]. This treatment is a final line of defense however, because it requires irradiation and chemotherapy to clear the patient’s bone marrow of their own blood stem cells and open the niche for donor stem cells. This exposes the patient to a serious risk of infection, in addition to the many unpleasant side effects of chemotherapy and irradiation. Additionally, transplant success is not guaranteed. Primary risks include rejection of the transplant, or rejection of the host by the grafted immune system, a life threatening complication known as graft versus host disease (GVHD). To avoid these, an HLA-matched adult sibling donor is best. However, less than 30% of patients have such a sibling, so for more than 25 years physicians have turned to unrelated donors that are HLA matched or slightly mismatched. Adult unrelated donors can sometimes be found, but umbilical cord blood (UCB) provides a promising alternative (Figure 1.1). UCB contains minimal T cells, requiring less stringent HLA matching and leading to reduced incidence of GVHD. However, compared to adult donor sources, single cord blood units have about 10-fold fewer HSCs, leading to graft failure, delayed engraftment and extended vulnerability to infection. Patients receiving UCB transplants thus experience higher transplant related mortality, although they have lower rates of relapse, leading to similar rates of overall survival [8-10].

A current major goal in regenerative medicine is to reduce transplant related mortality in UCB transplant by expanding cord blood HSCs ex vivo or enhancing their ability to home to and engraft the bone marrow. These methods have been summarized in recent review articles [10-14]. It is now common practice to perform double UCB transplant, with two separate cords, but this is logistically and financially challenging, and leads to a higher rate of acute GVHD [10, 15]. Many groups have sought to expand UCB

HSPCs ex vivo in order to achieve a sufficient cell dose, typically using cytokines such as TPO, SCF, and

Flt3L [16]. More recently, groups have achieved additional expansion by using small molecules to target signaling via Delta/Notch [17], SIRT1 [18], or the aryl hydrocarbon receptor [19-21]. Co-culture of cord blood with mesenchymal stem cells causes expansion and showed safety in patients [22]. Our group found

3 that dimethyl-prostaglandin E2 (dmPGE2) expands HSCs ex vivo [23-25]. A recently completed Phase I clinical trial, where one out of two cords was treated with dmPGE2 before both were transplanted, demonstrated the safety of dmPGE2 in humans, and suggested this will be an efficacious treatment [26].

A systematic review of completed trials using expanded UCBs revealed that overall, these protocols are safe for patients and they reduce the time to neutrophil engraftment. Ongoing randomized controlled trials should reveal whether they improve patient survival [16]. Other groups have focused on improving the homing and engraftment of UCB cells, with improvements seen with hyperbaric oxygen treatment to decrease EPOR activity [27], or forced fucosylation to improve P and E-selectin binding [28, 29].

Figure 1.1 Hematopoietic stem cell transplant. A) Recipients suffering from anemias or hematological malignancies receive a transplant containing hematopoietic stem cells (HSC) as well as other progenitor cell types. These are sourced either from an adult, HLA-matched sibling donor or from umbilical cord blood. Many groups have attempted expansion protocols for UCB, and others are attempting to direct the differentiation of an HSC from iPSCs. B) Within the recipient, HSCs home to the bone marrow niche and then proliferate and differentiate to give rise to the many downstream cell types. (CMP, common myeloid progenitor; CLP common lymphoid progenitor; MEP myelo-erythro-progenitor; GMP, granulocyte monocyte progenitor; RBC, red blood cells.)

1.1.3. Hematopoietic Stem Cells in Development

The zebrafish Danio rerio is a useful model for hematopoietic studies, especially in the embryo, which is optically transparent and develops external to the mother, allowing easy visualization of developmental processes. The earliest true HSCs are born in the aorta-gonad-mesonephros (AGM) region of a developing vertebrate embryo (Figure 1.2a,b). Endothelial cells of the dorsal aorta undergo an endothelial-to-hematopoietic transition, causing them to round up and bud off of the ventral wall of the dorsal aorta [30-32]. This budding begins in the zebrafish around 24 hours post fertilization (hpf) and in the mouse at day E10.5, and requires the expression of runx1, a marker that continues to identify HSPCs even into adulthood [33]. Newly born HSPCs enter circulation and travel to a secondary developmental niche, the fetal liver of mammals and the caudal hematopoietic tissue (CHT) of fish [34]. The CHT consists of a network of vessels that emerges from the posterior caudal vein beginning around 25 hpf (Figure 1.2c) [35], as well as other cell types including stromal cells, macrophages, and nerves [36]. As in the fetal liver, the

CHT supports HSC proliferation and differentiation, but not de novo specification. HSCs colonize the zebrafish CHT from day two to day five post fertilization, and beginning at day four, egress into circulation to populate the kidney marrow niche, a functional equivalent of mammalian bone marrow [33].

Studies of developmental hematopoiesis have contributed to our understanding of adult HSC behavior. Prostaglandin E2 enhances the specification of embryonic HSCs in the zebrafish, as well as the proliferation and engraftment of adult HSCs treated ex vivo [25]. Many signaling pathways are also shared between the two contexts. Dynamic regulation of Notch signaling is necessary for HSC specification [37], while Notch ligands produced by niche cells regulate proliferation of adult HSCs [38, 39]. Zebrafish cxcl8/cxcr4 signaling regulates the endothelial CHT niche [40], while the related CXCL2 and CXCR2 may regulate the adult endothelial niche [41], especially in response to stresses such as radiation [42]. The transcription factor tfec is expressed in fish CHT endothelial cells, where it drives production of cytokines such as SCF and TPO and enhances HSPC proliferation in that niche [43]. SCF and TPO are essential to maintain mammalial HSCs [44], and SCF produced specifically by sinusoidal endothelial cells is required for this maintenance [45].

Figure 1.2 Developmental hematopoiesis in zebrafish and mice. A) In both mammals and fish, HSCs travel to multiple developmental niches. They are born in the aorta-gonad-mesonephros (AGM), then travel to a secondary niche in the caudal hematopoietic tissue (CHT) or fetal liver. After expansion in this niche, they re-mobilize and travel to adult niches in the thymus and kidney or bone marrow. Adapted from Baron et al 2012 [46] B) HSC birth in the zebrafish. HSCs are specified by the transcription factor runx1 from endothelial cells of the aorta, which can be labelled with flk1 expression. HSCs undergo an endothelial-to- hematopoietic transition, rounding up and budding off. C) Schematic depicting the distinct steps of HSC (green) engraftment in the CHT: 1) luminal attachment, 2) transendothelial migration, 3) interaction with macrophages (blue) in the extravascular space, and 4) abluminal endothelial remodeling and stromal cell (orange) attachment. From Elliott Hagedorn.

1.1.4. Hematopoietic Stem Cell Migration and Engraftment

Adult HSCs periodically mobilize from the bone or kidney marrow, enter circulation, and then home back to and re-engraft the marrow. Cord blood and other HSC transplants rely on this inherent ability of donor HSCs to home to and engraft the recipient niche [47]. Investigating the mechanisms of HSC mobilization and homing is therefore vital to both our understanding of HSC biology and to efforts to improve transplant success.

Lodgement of HSCs in the adult niche involves several steps. HSCs in circulation must first be tethered to endothelial cells to slow their movement and initiate a rolling behavior. This process is mediated

6 by P and E selectins as well as vascular cell adhesion molecule 1 (VCAM1) expressed by endothelial cells of the bone marrow microvessels. These interact with ligands such as P-selectin glycoprotein ligand

(PSGL)-1 and α1β4 integrin on the HSCs [48]. This step is followed by firm adhesion of HSCs to endothelial niche cells, which is also mediated by α1β4 integrin [49]. Next, matrix metalloproteinases (MMPs) degrade

ECM proteins of the endothelial basement membrane, allowing HSCs to extravasate into the niche [50].

Throughout all of these steps HSCs migrate towards the chemokine stromal cell derived factor-1 (SDF-1), secreted by niche cells. HSCs express high levels of the SDF-1 receptor, CXCR4, and also migrate in culture in response to SDF-1 gradients, a Rac1/Rac2 dependent process [51]. The SDF-1/CXCR4 signaling axis is then essential for maintaining HSCs in the niche, as proteolysis of these factors leads to mobilization of HSCs into the peripheral blood [52].

Less is known about the migration of HSCs to developmental niches, which are inaccessible in vivo in mammalian systems. Surprisingly, early HSCs are not responsive to SDF-1, and CXCR4 signaling appears to be important only as HSCs begin to colonize the bone marrow [53]. Goossens et al found a correlation between high CXCR4 expression caused by loss of the transcription factor Zeb2 and fetal liver hematopoietic defects. In conditional Zeb2 knockout mice, HSCs are born in the AGM and migrate successfully to the fetal liver, but then adhere tightly to endothelial cells and do not spread normally throughout the fetal liver. These cells then have impaired proliferation and differentiation in the fetal liver

[54]. Murayama et al observed zebrafish hematopoietic progenitors rolling within endothelial cells of the

CHT, a behavior reminiscent of adult mammalian HSCs [34]. Data from our lab also indicates that the steps to colonization of the CHT resemble those of adult HSC homing. Owen Tamplin developed a

Runx1+23:EGFP transgenic line that labels HSPCs. In vivo time-lapse microscopy of this line with dsRed marking Flk1+ endothelial cells showed HSCs entering the CHT via circulation, adhering to endothelial cells, and extravasating to the abluminal side of the vessel. Subsequently, endothelial cells remodeled around the HSC, cuddling it to form an individual niche (Figure 1.2c) [36].

1.2. Epoxyeicosatrienoic Acids and Bioactive Lipids

1.2.1. Oxygenated Free Fatty Acids

Free fatty acids play diverse metabolic and signaling roles in the body. In this thesis I will focus on the 18-20 carbon chain polyunsaturated free fatty acids (PUFAs), which are well described in their specific signaling and metabolism. The PUFAs are largely divided into the ω-6 and ω-3 series, which are defined by the position of their final double bond at the n=6 or n=3 position in the carbon chain, respectively. Linoleic acid and arachidonic acid are the major ω-6 fatty acids (Figure 1.3), while eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) are the major ω-3 fatty acids (Figure 1.4). We consume both ω-6 and ω-

3 fatty acids in the form of oils and fats in our diet. Much research has focused describing on the ideal balance of ω-6 and ω-3 PUFAs for human health, but these data are controversial and they are on not the focus of this thesis. All four PUFAs can be esterified into the plasma membrane of cells, and freed from the membrane by the enzyme phospholipase A2. Within each series, PUFAs can be interconverted by elongation, saturation or desaturation reactions. Arachidonic acid, EPA, DHA, and their oxygenated derivatives are known as eicosanoids, which refers simply to their 20-carbon backbone. [55].

Oxygenation, or the enzymatic addition of oxygen in the form of epoxide or hydroxyl groups, occurs for these PUFAs by three major pathways- the lipoxygenase (LOX), cyclooxygenase (COX), and

Cytochrome P450 expoxygenase (CYP450) pathways (Figure 1.3, Figure 1.4). There are three major LOX enzymes, 5-LOX, 12-LOX, and 15-LOX which have differing expression patterns as well as substrate specificity. They add hydroxyl groups to produce diverse molecules such as HETEs, leukotrienes, resolvins, and maresins. COX-1 and COX-2 produce ring structures within fatty acids to create the pro-inflammatory prostaglandin and thromboxane molecules. These enzymes are famously inhibited by NSAIDs such as aspirin and ibuprofen. The CYP450s are heme-containing proteins involved in the metabolism of diverse molecules including steroids, drugs, and PUFAs [56, 57]. Of the 18 CYP450 families in humans, the CYP2 and CYP4 families are most often associated with PUFA metabolism. CYP2C8, 2C9, and 2J2 in particular show preferential production of EETs in vitro, but their in vivo activity is likely dependent on tissue-specific expression patterns and substrate availability [58]. In the case of PUFA metabolism, CYP450s add an epoxide ring, which can be further oxygenated to a di-hydroxyl by the enzyme soluble epoxide hydrolase

(sEH) [59]. Many of the eicosanoids, especially those within the arachidonic acid cascade, are known to

8 specifically bind G-protein coupled receptors, including the leukotrienes, prostaglandins, thromboxane A2, and resolvin E1 [60-62].

There may be cross-talk occurring between different lipid pathways; for instance, one EET isomer seems to inhibit an enzyme involved in the synthesis of prostaglandin E2 [63], while another decreased

LPS-stimulated production of PGE2 by monocytes [64]. EPA treatment of endothelial cells can stimulate them to produce EETs [65], and recent data shows that EETs can be further metabolized by COX enzymes to form highly angiogenic products [66].

Figure 1.3 Omega-6 series PUFAs and their oxygenation. PUFAs are oxygenated by three major families of enzymes, the LOX, COX, and CYP450 enzymes. One example lipid is shown for each major category. Linoleic acid can be converted to arachidonic acid by a series of elongation and desaturation reactions.

Figure 1.4 Omega-3 series PUFAs and their oxygenation. PUFAs are oxygenated by three major families of enzymes, the LOX, COX, and CYP450 enzymes. One example lipid is shown for each major category. EPA can be converted to DHA by a series of desaturation reactions.

1.2.2. Biochemistry and physiology of EETs

As illustrated in Figure 1.3, the CYP450 family of epoxygenases synthesizes EETs from arachidonic acid, a membrane phospholipid derivative. Four regioisomers are possible based on epoxygenation of arachidonic acid’s four double bonds, namely (5,6), (8,9), (11,12), or (14,15)-EET. The

11,12 and 14,15 isomers are more abundant and more physiologically active [67]. Intracellularly, EETs are quickly metabolized by the enzyme soluble epoxide hydrolase (sEH) to their corresponding dihydroxyeicosatrienoic acids (DHETs), which are generally thought to be less biologically active [68].

EETs are synthesized by endothelial cells [69], and they can be released into the circulation [70, 71]. They are found as constituents of membrane phospholipids [72], and they are present in diverse human tissues.

The EETs were initially identified for their role in the relaxation of vascular smooth muscle [73].

Their vasodilator function has since been verified by many studies [70, 74-79]. Since then, EETs have been shown to have anti-inflammatory [80, 81] and pro-angiogenic effects [82-86]. On a cellular level, EETs contribute to the proliferation and migration of both endothelial and cancer cells [83, 87-90], although in one case EETs inhibited smooth muscle cell migration [91]. Systemic EET treatment of mice who had received a human tumor xenograft caused tumor dormacy escape, though it is unclear whether in this case EET

10 worked directly on cancer cells or perhaps on host tissues [68, 92]. In human breast cancer, higher EET levels are associated with tumor aggressiveness [93].

EETs have several described roles in the blood, though a direct role for these molecules in regulating hematopoietic stem cells has not been investigated. EETs are produced and stored in red blood cells, and can be released to induce vasodilation [94-96]. Frömel et al found that inhibition of sEH in zebrafish, resulting in a buildup of EET and a lack of DHET, caused defects in the CHT endothelium as well as decreased numbers of hematopoietic progenitors in this region. EET and DHET treatment, on the other hand, both caused an increase in hematopoietic progenitors in the CHT [97]. 8,9-EET, but not the other isomers, inhibits antibody production by B cells [98]. EETs decreased the expression of P-selectin and inhibited adhesion to endothelial cells in platelets [99], monocytes [100], and polymorphonuclear cells [101].

1.2.3. Receptors and Signaling Pathways

EETs can have both paracrine and autocrine effects and have been shown to modulate a wide variety of downstream signaling pathways, including NFkB [102], EGFR [103], MAPK [104, 105], PI3K [106-

108], and NOS [109] A wealth of data indicates that EETs signal via a specific high affinity membrane receptor. A bead-tethered EET analog that was unable to enter the cell nonetheless showed efficient activation of downstream signaling pathways [110]. DHETs can, however, activate PPARα [111, 112]. EET signaling requires GTP and can be inhibited by the G-protein inhibitors GTPγS and GDPβS, suggesting that EET signals through a G-protein coupled receptor (GPCR) [73]. Chen et al recently constructed a radio- and photoaffinity-labeled EET analog which can be photo-crosslinked to the putative receptor. They identified a specific 47 kDa band present in U937 monocytes, a known EET-responsive line, but absent in

HEK293T cells, which show no response to EET. Overexpression of 79 orphan GPCRs in HEK293T cells failed to produce this band upon treatment with the EET analog [113].

Several eicosanoid and free fatty acid-responsive GPCRs are reported to also respond to EET, but only at micromolar concentrations, so they represent low affinity receptors. Screening in Xenopus oocytes revealed low affinity responses of several prostaglandin receptors [114]. The PTGER2 receptor also showed EET-responsiveness in rat mesenteric arteries [115]. Overexpression of the free fatty acid receptor

GPR40 caused EET-responsiveness in HEK293 cells [105]. One report showed EET antagonizing the thromboxane receptor [74]. Non-GPCR low affinity EET receptors have also been reported, primarily

TRPV1, TRPV4, and TRPC receptors [116-118]; although one group failed to find activity of EET at TRPV4

[119].

1.3. Lipid Signaling via G-Protein Coupled Receptors

1.3.1. GPCR Cycling

G-protein coupled receptors (GPCRs) are plasma membrane bound proteins with seven membrane-spanning domains. They function primarily to sense the extracellular environment; ligand binding at the extracellular surface of GPCRs induces conformational changes in the protein that activate intracellular signaling. GPCRs are the largest gene family in the human genome, including over 800 genes, and they are found in all eukaryotes. This diversity means that GPCRs are major regulators of physiology and health, in addition being important drug targets [120, 121].

GPCRS are so called because they couple to heterotrimeric GTPases known as G-proteins which have Gα, Gβ, and Gγ subunits (Figure 1.5). In the inactive state, all three G-protein subunits are complexed with each other and associated with the cytoplasmic face of the plasma membrane. The Gα subunit has

GDP/GTP binding ability and is bound to GDP in the inactive state. Ligand binding of the GPCR causes a conformational change which allows the GPCR to function as a guanine nucleotide exchange factor (GEF) for the G-protein. Gα loses GDP, which is quickly replaced by cytosolic GTP. This change causes dissociation of the Gα subunit from the Gβγ subunits. Each of these can induce downstream signaling, although Gα signaling is best studied. The Gα subunit eventually hydrolyzes its GTP to GDP, allowing re- establishment of the heterotrimeric GTPase, and the system is ready to signal again [122]. However, ligand activation also causes the recruitment of GPCR-inactivating proteins. GPCR receptor kinases (GRKs) can phosphorylate the intracellular tail of the GPCR, preventing its further interaction with G-proteins. β-arrestin molecules also bind to phosphorylated GPCRs, further blocking G-protein interaction and even inducing the internalization of the GPCR by endocytosis. More recent research has taught us that β-arrestin, too, can activate signaling downstream of GPCRs [122].

1.3.2. Lipid-GPCR Binding

GPCRs respond to a wide variety of ligands including proteins, small molecules, lipids, hormones, ions, and even photons. The best studied receptors are the photon- and retinal-responsive rhodopsin

12 receptor as well as the epinephrine-responsive β2-adrenergic receptor (β2AR). For both of these we have extensively documented crystal structures, as well as detailed understanding of their signaling. For instance, agonist binding of β2AR induces conformational changes in the 6th and 7th membrane spanning domains, as well as the cytoplasmic surface of the protein. However, pharmacological β-arrestin biased ligands, which activate β-arrestin signaling with minimal triggering of Gα signaling, cause changes in the

7th membrane spanning domain only [123].

Figure 1.5 GPCR cycling. Adapted from Sanchez-Fernandez et al, 2014 [122], re-printed according to license #4091520473504. GPCRs (green) are activated by diverse ligands. Activation results in loss of GDP by the Gα protein. Separate Gα and Gβγ subunits can then initiate signaling. Recruitment of β-arrestin causes arrestin-dependent signaling, as well as internalization of the GPCR. Internalized GPCRs can be degraded by the lysosome or recycled to the cell membrane.

A wide variety of lipids are known to activate GPCRs, including eicosanoids, other free fatty acids, and phospholipids. Two lipid-binding GPCRs have solved crystal structures, the sphingosine-1-phosphate receptor (S1PR), and free fatty acid receptor 1 (FFAR1). Both show occluded extracellular access to the central binding pocket relative to most GPCRs [124, 125]. S1PR additionally has a more open structure between the 7th and 1st membrane-spanning domains, suggesting that S1P may access the interior binding pocket from within the membrane [124], this binding mechanism was recently successfully simulated for a pharmacological ligand in silico [126]. Membrane access of lipid ligands has also been hypothesized for the cannabinoid receptors [127].

Figure 1.6 Theoretical mechanisms of lipid activation of GPCRs. Lipids (red) could access the binding pocket of a GPCR (yellow) by a traditional, extracellular route (A), or by first associating with the membrane (B). C) Some GPCRs require homo- or heterodimerization to signal, and membrane lipids could affect their dimerization. D) Lipid composition of plasma membranes or endosomal vesicles could affect fusion of those vesicles with the membrane and localization of a GPCR to the cell surface. E) Some GPCRs need to be localized within membrane microdomains in order to signal, and these microdomains have specialized lipid compositions.

It is important to note that some GPCRs are reported to have constitutive, ligand-independent activity [128-132]. In other cases, GPCRs are known to signal upon the formation of homo- and heterodimers [133-137], including receptors for the eicosanoids thromboxane [138, 139] and leukotriene

B4 [140]. Finally, GPCR signaling can be caused by incorporation of GPCRs into membrane microdomains

[141-144], or by relocalization from endosomal compartments to the plasma membrane [145-148]. GPCR dimerization, in particular, has been shown to be regulated by membrane lipids, including fatty acids [149].

FRET studies with reconstituted lipid bilayers showed that lipid composition affected the formation of GPCR

14 dimers [150]. Cholesterol can mediate dimer formation for β2AR [151] as well as mGlu1 [152]. Rhodopsin signaling is enhanced by high docosahexaenoic acid content in cell membranes, perhaps through enhanced dimerization or by partitioning rhodopsin molecules into certain membrane microdomains [153].

1.3.3. Gα Signaling

Vertebrate Gα proteins are categorized into four subfamilies, Gαs, Gαi/o, Gαq/11, and Gα12/13.

Gαs and Gαi were the first families characterized, and were described as having opposite effects either stimulating (Gαs) or inhibiting (Gαi) adenylyl cyclase. Gαs causes increased production of cyclic AMP, resulting in PKA signaling. Many GPCRs activate this pathway, most notably the β-adrenergic receptors.

In addition to inhibiting the adenylyl cyclase/cAMP/PKA pathway, Gαi/o has since been shown to inhibit calcium and potassium currents, and to activate MAPK and other pathways [154]. Gαq proteins activate phospholipase C β and γ signaling, resulting in the formation of IP3/DAG. Gα12/13 is primarily known for activating Rho/ROCK signaling, which can affect cytoskeletal dynamics and trigger diverse downstream signaling [155]. While many GPCRs couple sterotypically to particular Gα proteins, others can activate multiple signaling pathways, sometimes in a cell type- or ligand specific- fashion [156-158].

As I will describe in Chapter 2, we found that EET signaling activates Gα12/13 signaling, as well as PI3K signaling, and the transcription of AP-1 family members and other genes associated with migration and metastasis. Gα12/13 signaling increases the metastasis and invasiveness of diverse cancer types

[159-162]. Gα12/13 has also been shown to activate the migration of a wide variety of non-malignant cell types, including fibroblasts and endothelial cells [161], neural progenitor cells [163], splenic B cells [164], and gastrulating zebrafish mesodermal cells [165]. Additionally, the loss of Gα13 in endothelial cells results in a loss of VEGFR2 expression and impaired angiogenesis [166]. It is plausible that Gα12/13 activation in our system would result in PI3K activation, as this family has been linked to both activation [167-169] and inhibition [170] of PI3K in different contexts.

Gα12/13 proteins are thought to mediate many of their phenotypes by activating RhoA signaling, which was first shown for LPA receptors [171]. RhoA is a member of the Rho family of small monomeric G- proteins, which bear some homology to Gα proteins and undergo a similar cycle of activation and deactivation based on their bound state to GTP or GDP [172]. GDP-bound RhoA must be activated by a

RhoGEF. Gα12/13 can activate the RH-RhoGEF subfamily which has three family members. Gα13 is

15 known to bind and activate p115-RhoGEF and PDZ-RhoGEF [155, 173, 174], while Gα12 can activate leukemia associated RhoGEF (LARG) [175]. These three RhoGEFs have roles regulating neutrophil and smooth muscle physiology [172]. Rho signaling has many reported roles in hematopoiesis, including regulating HSC self-renewal, migration, and differentiation [176]. In the zebrafish, one study showed

Rho/ROCK signaling to be a negative regulator of the endothelial to hematopoietic transition [177].

1.3.4. GPR132- Physiology and Ligand Binding

I will discuss in Chapter 3 the identification and validation of GPR132 as a receptor for EET and other fatty acids. GPR132 was originally named G2A for its ability to cause G2 accumulation in the cell cycle. The Witte lab saw induction of GPR132 expression upon transformation of murine pre-B cells with

BCR-ABL. They found that GPR132 is expressed in lymphocytes, upregulated in response to various stresses, and that it can block transformation by BCR-ABL [178]. GPR132 expression was later reported in human and mouse macrophages [179], as well as human atrial myocytes [180]. Zebrafish have two

GPR132 homologs, which are broadly expressed in adult tissues [181]. Gpr132a is additionally expressed in zebrafish embryos [182]. The homology between the zebrafish and human orthologs of GPR132 at the amino acid level is 39% (gpr132a) and 44% (gpr132b) [181], while human and mouse GPR132 share 69% amino acid identity [183]. Interestingly, the mouse ortholog lacks a histidine residue that is conserved between the human and both fish orthologs, which may be responsible for GPR132’s proton sensitivity.

These initial studies relied only on GPR132 expression, rather than ligand activation, but attention soon turned to a search for a GPR132 ligand. GPR132 was described as a receptor for the bioactive lipids lysophosphatidylcholine (LPC) and sphingosylphosphatidylcholine (SPC). Experiments showing direct, high affinity binding were later retracted [184, 185], but extensive studies have demonstrated the responsiveness of GPR132 to LPC and SPC, so the protein is perhaps indirectly activated by these molecules, or it is a low affinity receptor. Indeed, one group showed that LPC treatment caused relocalization of GPR132 from endosomal vesicles to the cell surface, activating the receptor without typical ligand-protein binding [186]. While some groups have shown an anti-inflammatory role for LPC-GPR132 signaling, including in suppressing cytokine release in sepsis [187] and enhancing the activity of suppressor

T cells [188], most reports describe a pro-inflammatory role. The LPC-GPR132 signaling axis activates chemotaxis in T cells, macrophages, natural killer cells, and zebrafish microglia [182, 189-191]. In the

16 central nervous system, LPC requires GPR132 to cause inflammation [192], and multiple groups have shown that this signaling produces neurotoxic effects in vitro [193, 194]. Beyond the immune system, LPC stimulation of GPR132-expressing HeLa cells induced cAMP signaling as well as apoptosis [195], and LPC activated GPR132 in human atrial myocytes to induce potassium currents [180],

In contrast, the Shimizu lab described low pH activating GPR132 to cause inositol phosphate accumulation, which could be inhibited by LPC [196]. Another group found that zebrafish GPR132 orthologs expressed in cell lines were activated in acidic environments [181]. GPR132 is a member of the OGR1 family of GPCRs, which also includes GPR68 (OGR1), GPR65, and GPR4. The entire family is reported to be pH sensitive [197, 198], although one study found GPR132 to be the weakest pH sensor, perhaps because of five histidine residues shown to be important for proton-sensing by GPR68, only one is conserved in GPR132. [183]. As for GPR132, the other OGR1 family members have also shown LPC and

SPC dependent signaling [199].

GPR132 is a reported receptor for bioactive lipids such as 9-HODE and 11-HETE [191, 200, 201].

Like 11,12-EET, these molecules are oxygenated, unsaturated free fatty acids. While initial studies were done entirely in vitro using the PathHunter β-arrestin system, which can measure GPCR activation without relying on downstream signaling, a pair of studies found in vivo interactions of 9-HODE and GPR132. 9-

HODE treatment of human keratinocytes inhibited their proliferation, and this could be blocked by siRNA of

GPR132 [202]. Treatment with the chemotherapeutic oxaliplatin causes a hypersensitivity to pain in both mice and human patients. One group recently found that oxaliplatin treatment increases 9-HODE concentrations in the mouse, and that GPR132 knockout mice do not show oxaliplatin-induced pain hypersensitivity [203]. Whether GPR132 is primarily a fatty acid receptor, an LPC receptor, or a pH sensitive receptor remains debated, and these roles may be context-dependent.

There are reports of diverse signaling pathways downstream of GPR132, which again may be context dependent. GPR132 was shown in multiple case to couple to Gα13 [195, 204, 205], and is known to cause Rho kinase signaling, which is often activated downstream of Gα13 [180, 192, 194]. I will describe in Chapter 2 our own evidence that the Gα12/13 pathway is activated by EET. One group also found that activation of GPR132 by 9-HODE caused an opening of TRPV1 channels [203]. TRPV channels are reported targets of EETs, as well [84, 116, 206-210].

1.4. Summary of Thesis

Investigations into stem cell function have the potential to enhance transplantation and regenerative therapies, while increasing our understanding of basic cell biology. The hematopoietic stem cell is the most studied and certainly the most transplanted stem cell, but even there wide avenues remain unexplored. In particular, biologists have focused on the central dogma- seeking to understand the role of genes and proteins in regulating HSC self-renewal, migration, and engraftment. However, significant data from our lab and others shows that lipid signaling molecules, especially fatty acids, are common and important regulators of HSC function. These lipids have complex interactions with receptors and metabolic enzymes, constituting a network of bioactive molecules.

In the chapters of this thesis I will describe first the discovery that one such lipid, 11,12- epoxyeicosatrienoic acid, enhances the specification and transplantation of hematopoietic stem cells. This signaling relied on a Gα12/13-coupled GPCR, which in turn activated PI3Kγ signaling and AP-1 transcription factors. AP-1 expression was enhanced in both adult human HSPCs and the hemogenic endothelium of zebrafish, where it was required for EET phenotypes. AP-1 signaling was similarly required in endothelial cells in order to mediate EET-induced migration of HSPCs to the zebrafish caudal hematopoietic tissue, so these signaling pathways may have cell autonomous and cell non-autonomous effects depending on context.

I will then describe how I pinpointed GPR132 as the EET receptor that mediates these hematopoietic phenotypes. Bioinformatic profiling of EET-binding and non-binding cell lines revealed only

10 GPCRs that were expressed in common among only the binding cell lines. In vitro screening revealed that one of these, GPR132, responded dose-dependently to EET. I found that GPR132 was required for

EET phenotypes in the zebrafish embryo and for normal hematopoietic stem cell transplant in the mouse.

Additionally I found that GPR132 is responsive in vitro and in vivo to a wide variety of medium chain, unsaturated, free fatty acids. This promiscuity in a GPCR is unusual, and may provide hints about how to best target GPR132 therapeutically.

Finally, GPR132 is a member of the OGR1 family of GPCRs, and I will describe how one other member, GPR65, also has an important role regulating hematopoiesis. GPR65 is upregulated during the

18 endothelial to hematopoietic transition by GATA-2. However, loss of GPR65 in either the mouse or zebrafish causes an increase in HSPCs, suggesting that GPR65 is a negative regulator of HSPCs, and forms a negative feedback loop with the pro-hematopoietic GATA-2. GPR65 is not responsive to EET or other fatty acids, and may signal constitutively or in response to acidic conditions. Investigation into the other OGR1 family members, GPR4, and GPR68, remains an important question.

In conclusion, the study of one particular bioactive lipid, 11,12-EET, has led to a thorough elucidation of the signaling downstream of EET in hematopoiesis, and eventually to the identification of

GPR132 as its receptor and an important regulator of hematopoietic stem cells. At least one GPCR closely related to GPR132 can also influence developmental hematopoiesis, and GPR132 itself can respond to many different lipid signaling molecules. GPR132 is therefore a promising therapeutic target for the manipulation of HSCs and HSC transplant, and may represent a node in the network of bioactive fatty acids regulating hematopoietic stem cell function. These data reinforce the importance of lipid signaling molecules and their GPCR receptors in controlling stem cell behaviors. HSCs, the prototypical stem cell, have been thoroughly studied for their genetic and cell biological regulators. With the support of this vast body of knowledge, the hematopoietic system is an excellent context for new studies of systems-level lipidomics.

Chapter 2. Epoxyeicosatrienoic acids enhance embryonic haematopoiesis and adult marrow engraftment.

Pulin Li,* Jamie L. Lahvic,* Vera Binder,* Emily K. Pugach, Elizabeth B. Riley, Owen J. Tamplin, Dipak Panigrahy, Teresa V. Bowman, Francesca G. Barrett, Garrett C. Heffner, Shannon McKinney- Freeman, Thorsten M. Schlaeger, George Q. Daley, Darryl C. Zeldin, and Leonard I. Zon

*These authors contributed equally to this work.

2.1. Attributions

The following is a paper published in the journal Nature on July 23, 2015 describing the role of epoxyeicosatrienoic acid in enhancing HSPC specification and engraftment in zebrafish and mammals

[211]. I am a co-first author on the paper along with Pulin Li and Vera Binder. Dr. Li designed and performed the chemical screen leading to the identification of 11,12-EET as an enhancer of hematopoietic stem cell transplant in the zebrafish. She confirmed a similar phenotype in mouse transplants, and found that EET treatment of zebrafish embryos increases expression of HSPC markers by in situ. I confirmed this in situ phenotype and validated by performing in vivo imaging of transgenic zebrafish embryos treated with DMSO or EET. I performed cell-tracking to analyze HSPC behaviors (Figure 2.3, Figure 2.4). This data demonstrated that EET enhanced HSPC specification, without affecting cell division, or residency time in the aorta.

Dr. Li determined that EET signaling activated PI3K to lead to increased expression of AP-1 transcription factors, both of which were required for EET phenotypes in the zebrafish. I used genetic and chemical genetic assays to block PI3K signaling via the different catalytic subunits, and found that only

PI3Kγ was required for EET phenotypes in the AGM and CHT (Figure 2.9, Figure 2.11). I also created transgenic zebrafish lines to block AP-1 signaling in Runx1+23 positive HSPCs or in Flk1 positive endothelial cells. Loss of AP-1 signaling in endothelial cells was sufficient to block EET phenotypes, indicating that AP-1 signaling is required cell autonomously in the hemogenic endothelium (Figure 2.6). Dr.

Binder conducted gene expression analysis in human primary and transformed blood cells to show that the upregulation of AP-1 genes was conserved in human cells, and that EET treatment also led to increased expression of factors involved in cell-cell signaling and cell movement. Along with Drs. Li and Binder, I contributed extensively to the writing and editing of the text. Elizabeth Riley and Emily Pugach offered technical assistance with zebrafish injections and in situ hybridizations.

2.2. Abstract

Haematopoietic stem and progenitor cell (HSPC) transplant is a widely used treatment for life- threatening conditions such as leukaemia; however, the molecular mechanisms regulating HSPC engraftment of the recipient niche remain incompletely understood. Here we develop a competitive HSPC transplant method in adult zebrafish, using in vivo imaging as a non-invasive readout. We use this system to conduct a chemical screen, and identify epoxyeicosatrienoic acids (EETs) as a family of lipids that enhance HSPC engraftment. The pro-haematopoietic effects of EETs were conserved in the developing zebrafish embryo, where 11,12-EET promoted HSPC specification by activating a unique activator protein

1 (AP-1) and runx1 transcription program autonomous to the haemogenic endothelium. This effect required the activation of the phosphatidylinositol-3-OH kinase (PI(3)K) pathway, specifically PI(3)Kγ. In adult

HSPCs, 11,12-EET induced transcriptional programs, including AP-1 activation, which modulate several cellular processes, such as migration, to promote engraftment. Furthermore, we demonstrate that the EET effects on enhancing HSPC homing and engraftment are conserved in mammals. Our study establishes a new method to explore the molecular mechanisms of HSPC engraftment, and discovers a previously unrecognized, evolutionarily conserved pathway regulating multiple haematopoietic generation and regeneration processes. EETs may have clinical application in marrow or cord blood transplantation.

2.3. Results

2.3.1. Zebrafish whole kidney marrow (WKM) competitive transplantation-based chemical screen

identifies EETs as enhancers of marrow engraftment

Transplanting zebrafish whole kidney marrow (WKM), the equivalent of mammalian whole bone marrow (WBM), can rescue lethally irradiated zebrafish as well as mutants with haematopoietic defects

[212-214]. In an effort to quantify HSPC activity in zebrafish, we developed a competitive transplantation system in a transparent mutant zebrafish, casper, which allows visualization of engraftment in vivo [215].

We co-injected WKM cells from two ubiquitous green and red transgenic donors, Tg(β-actin:GFP) [213] and

RedGlo® fish [216], into sub-lethally irradiated casper (a). We could directly visualize donor-derived green and red fluorescence within the same kidney region, and calculate relative engraftment as the ratio of the

22 green-to-red fluorescence intensity (G/R) (Figure 2.1b). This assay enables quick, quantitative, and non- invasive evaluation of marrow engraftment repeatedly over time, even up to 3 months post transplant. We validated the quantitative potential of this imaging-based approach by comparing with flow cytometry-based analysis of WKM from the same recipient. The two results were linearly correlated (Figure 2.1c). The assay was also sensitive to changes in the relative number of green-tored donor cells. We observed an increase of recipient G/R readouts accompanying the increasing green-to-red ratio of transplanted donor cells

(Figure 2.1d). Additionally, our system successfully detected the effects of two known chemical modulators of HSPC engraftment: dmPGE2 (16, 16-dimethyl-prostaglandin E2), a stabilized derivative of PGE2 [25], and BIO (6-bromoindirubin-3′-oxime), a GSK-3β inhibitor [24]. Green WKM transiently exposed for 4 hrs to either of these chemicals and transplanted together with untreated red WKM showed significantly enhanced engraftment capability compared to vehicle-only controls (Figure 2.1e). Therefore, within the cell dose and ratio range tested above, our zebrafish competitive transplantation system can successfully detect changes of relative engraftment of HSPCs.

To our knowledge, a screen-based forward-genetic approach to understand transplantation biology has never been attempted. We used our assay to screen 480 compounds with known bioactivities, which had been selected to cover diverse signaling pathways (Fig. 2.2a). 10 compounds significantly increased the G/R ratio reproducibly, including PGE2 and Ro 20-1724, which activates the cAMP pathway downstream of PGE2 [25]. The other hits target pathways that previously have not been linked to HSPC engraftment, including 11,12-epoxyeicosatrienoic acid (EET) and 14,15-EET (Figure 2.1e). These are arachidonic acid-derived eicosanoids synthesized through the cytochrome P450 epoxygenase pathway

(Fig. 2.2b) [59, 80]. A gene expression study previously reported mouse Cyp2j6, a cytochrome P450 epoxygenase, as one of the 93 genes enriched in long-term HSC [217].

Figure 2.1 Zebrafish whole kidney marrow (WKM) competitive transplantation-based chemical screen identifies EETs as enhancers of marrow engraftment. a, Schematic of zebrafish WKM competitive transplantation. b, Calculation of relative engraftment capability (G/R). White dashed line: kidney; Gkid/Rkid, kidney fluorescence intensity; Gbkg/Rbkg, background fluorescence intensity. c, The G/R ratios from imaging linearly correlated with flow cytometry analysis of the same recipients (linear regression). wpt, weeks post transplant. d, Serial dilution competitive transplantation with varying donor GFP:DsRed2 ratios. e, 4 hr transient chemical treatment increased WKM engraftment. 11,12- and 14,15-EET, 0.5 μM. Unpaired two- tailed t-test, mean with s.e.m. (d–e).

Fig. 2.2 Whole kidney marrow screen and EET metabolism. a, WKM from Tg(β-actin:GFP) donors were dissected, dissociated as single cell suspension, and incubated with chemicals at room temperature for 4 hrs in a round-bottom 96-well plate. Meanwhile, WKM were dissected from RedGlo® zebrafish, counted and kept on ice. After the drug treatment, chemicals were washed off and cells were resuspended in 0.9 X DPBS + 5% FBS. 20,000 treated green WKM and 80,000 untreated red WKM were co-injected retro- orbitally into sublethally irradiated capser zebrafish (n=10 per chemical). For every independent screening day, negative control (DMSO) and positive control (10 μM dmPGE2) were used for normalization and quality assurance. The engraftment was measured at 4 wpt (weeks post transplant) by fluorescence imaging and ImageJ quantification as described in Figure 2.1b. b, EET metabolic pathway: arachidonic acid is released by PLA2 (phospholipase A2) from the membrane lipid bilayer. EETs (epoxyeicosatrienoic acids) are synthesized directly from arachidonic acid by the cytochrome P450 family of epoxygenases, especially 2C and 2J in human [77], and get degraded by soluble epoxide hydrolase (sEH), generating DiHET (dihydroxyeicosatrienoic acids). Four isomers of EET exist in vivo: 5,6-, 8,9-, 11,12- and 14,15-EET.

2.3.2. 11,12-EET enhances HSPC specification in the zebrafish embryo AGM

Despite years of research on the potent effects of EETs in numerous physiological processes [68,

83, 218], knowledge about their direct target(s) and downstream pathway(s) is still very limited. To tackle this problem, a robust system allowing easy genetic perturbation is crucial. Since adult regeneration often reactivates pathways important for development, we decided to probe EETs’ effects on haematopoiesis during embryo development. Analogous to mammalian development, zebrafish HSPCs form from a flk1+ population, named haemogenic endothelium, at 24 hpf (hours post fertilization) and become runx1+ at 36 hpf in the evolutionarily conserved aorta-gonad-mesonephros (AGM) region [30, 31,

219]. HSPCs enter circulation after they emerge from the AGM [30-32], and seed the caudal haematopoietic tissue (CHT), a secondary haematopoietic site equivalent to the mammalian fetal liver (Figure 2.3a) [34,

36]. 11,12-EET treatment between 24–36 hpf strongly increased the HSPC marker runx1 in the AGM, and surprisingly induced runx1 in a non-haematopoietic region of the tail mesenchyme, where runx1 is not

25 normally expressed (Figure 2.3b). This indicates 11,12-EET might be inducing a conserved transcriptional program. We confirmed this AGM phenotype with in vivo time-lapse imaging of HSPC birth from the haemogenic endothelium. Tg(CD41:GFP; flk1:DsRed2) embryos treated with 11,12-EET starting at 24 hpf showed a significant increase in the number of double-positive HSPCs in the AGM from 30–46 hpf (Figure

2.3c-d). Single-cell analysis showed this change is mainly due to a significant increase in the frequency of

HSPCs directly specified from the haemogenic endothelium, while no increase in the rate of cell division or

AGM retention was observed (Figure 2.4). The additional HSPCs produced upon 11,12-EET treatment successfully homed to their next niche, resulting in increased numbers of HSPCs in the CHT, which was verified by in situ hybridization for the HSPC marker cmyb (Figure 2.3e, Figure 2.5). Time-lapse imaging of

Tg(Runx1+23:GFP) zebrafish showed that 11,12-EET treatment between 24–48 hpf increased the rate of arrival of GFP+ HSPCs to the CHT (Figure 2.3f, Supplementary Video 1,2), presumably due to enhanced

HSPC specification in the AGM. 11,12-EET induces a PI3K-dependent AP-1/runx1 transcriptional program to increase HSPC specification

To further dissect the molecular mechanism leading to runx1 induction, we performed microarray analysis on 11,12-EET-treated 36 hpf embryos (Appendix 1). The upregulation of multiple AP-1 (Activator

Protein 1) family transcription factors, including fosl2, and duplicated orthologs of human

JUNB, junb and junbl, was among the most prominent changes. Whole-mount in situ hybridization confirmed the induction both in the AGM and the non-haematopoietic region of the tail mesenchyme (Figure

2.6d, top two rows). AP-1 mRNA transcripts were detectable within 1 hr of 11,12-EET treatment and insensitive to the protein translation inhibitor cycloheximide (Figure 2.7a-b), indicating AP-1 members are immediate targets of EET signaling. In contrast, runx1 induction required at least 4 hrs of 11,12-EET treatment and was completely blocked by cycloheximide (Figure 2.7c). Therefore, we hypothesized that

EET-induced AP-1 expression is necessary for increasing runx1 transcription.

Figure 2.3 11,12-EET enhances HSPC specification in the zebrafish embryo AGM. a, Schematic of HSPC development in zebrafish embryos. hpf, hours post fertilization; AGM, aorta-gonad-mesonephros; CHT, caudal haematopoietic tissue. b, Representative images of whole-mount in situ hybridization showing 11,12-EET (24–36 hpf treatment) induced HSPC marker runx1 in the AGM and a tail non-haematopoietic tissue (>8 independent experiments, n>100). c–d, 11,12-EET (24–46 hpf) enhanced CD41:GFP/flk1:NRAS-mCherry double positive HSPCs (white arrowheads) emerging in the AGM. Arrows indicate blood flow. e–f, Same treatment increased the number of HSPCs in the CHT. mCherry+ HSPCs quantified in the Tg(Runx1+23:mCherry) CHT (e). Representative montage images of Runx1+23:GFP HSPCs (white arrowheads) engrafting CHT. flk1:DsRed2, endothelial cells (f). Unpaired two-tailed t-test, mean with s.e.m. (d–e).

Figure 2.4 11,12-EET enhances HSPC specification in the AGM in zebrafish embryos. Tg(CD41:GFP/flk1:NRAS-mCherry) embryos were treated with DMSO or 5 μM 11,12-EET starting at 24 hpf (hours post fertilization), then mounted for spinning disc confocal timelapse imaging from 30–46 hpf in the presence of the chemicals. Bars show mean and s.e.m., unpaired two-tailed t-tests, n=10 for DMSO, n=7 for EET. a, More HSPCs are directly specified in EET-treated AGM. Graph shows HSPCs born by direct specification/budding only, excluding cells born by division of an already-budding cell. b–c, 11,12- EET does not influence the rate of HSPC division in the AGM, shown by per movie, percentage of budding HSPCs that divide at least once (b) and divide twice or more (c) before leaving the AGM or before the end of timelapse recording. n.s., not statistically significant.

Figure 2.5 11,12-EET treatment between 24–48 hpf increases the number of HSPCs in the CHT. a, Embryos were treated between 24–48 hpf with either DMSO or 5 μM 11,12-EET. Chemicals were washed off at 48 hpf, and embryos grew in drug-free environment for another 24 hrs. b, 11,12-EET treatment increased the number of mCherry+ HSPCs in the CHT in Tg(Runx1+23:mCherry) embryos (see also Figure 2.3e). Representative images of the CHT from the two groups. c, The same chemical treatment increased the staining of cmyb, a HSPC marker, by whole-mount RNA in situ hybridization. Representative images from each group (a total of n>60 from 3 independent experiments).

To genetically test this hypothesis, we globally knocked down AP-1 with anti-sense morpholinos targeting junb/junbl, which blocked runx1 expression without affecting endothelial cells of the AGM (Figure

2.8), suggesting that AP-1 might be required for HSPC specification from haemogenic endothelium. To test if AP-1 function is autonomous to the haemogenic endothelium, we delivered a dominant-negative form of

JunB protein (dnJUNB) specifically to the flk1+ endothelial cells, before the induction of runx1, to functionally inhibit all AP-1 activity. Although flk1:dnJUNB did not significantly reduce the expression of runx1 in DMSO-treated embryos, it suppressed the EET-induced increase of runx1 in the AGM (Figure

2.6a-b). Combined with the gene expression data, these genetic analyses showed that 11,12-EET activates an AP-1/runx1 transcriptional cascade of cell fate specification autonomous to the haemogenic endothelium.

In an effort to define downstream signaling events for 11,12-EET, we performed a chemical suppressor screen in zebrafish embryos by examining the capability of various chemicals to suppress the 11,12-EET- induced AP-1/runx1 gene signature (Figure 2.6c). Multiple PI3K inhibitors completely blocked the signature without detrimental effects to overall embryonic development (Figure 2.6d-e, Figure 2.9a). To interrogate specific PI3K catalytic subunits, we assayed subunit-specific chemical inhibitors and morpholinos targeting individual Class I PI3K subunits. Among α-, β-, γ-, and δ-subunits of PI3K, only PI3Kγ loss of function specifically abrogated the runx1 induction in the AGM and tail non-haematopoietic tissue (Figure 2.9b-c).

Additionally, 11,12-EET enhanced PI3K activity in immortalized human umbilical vein endothelial cells, assayed by Akt phosphorylation (data not shown). No such increase was seen in human umbilical cord blood CD34+ HSPCs, although EET-induced gene expression changes could be partially blocked in these cells by co-treatment with PI3K inhibitors. This indicates PI3K functions either directly downstream of 11,12-

EET or as a parallel pathway, depending on the cellular context. In either case, PI3K activity is required for inducing the AP-1/runx1 transcription cascade in the AGM.

Figure 2.6 11,12-EET induces a PI3K-dependent AP-1/runx1 transcriptional program to increase HSPC specification. a–b, Stable flk1:dnJUNB-2A-GFP expression blocking AP-1 function suppressed 11,12-EET- enhanced HSPCs in the AGM. Representative images of runx1/cmyb in situ hybridization (a) and quantification (b) after 11,12-EET treatment (24–36 hpf). Embryos scored as high, medium, or low runx1/cmyb, summed across 4 experiments. *, p=0.01; ***, p<0.0001 by Chi-square. WT, wildtype. c, Schematic of chemical screen for EET signaling pathway suppressors. d–e,11,12-EET induced AP- 1 family transcription factors (fosl2, junb/junbl) (d) and runx1 (e), suppressed by cotreatment with LY294002, a PI3K inhibitor, in the AGM and tail (d–e) (3 independent experiments, n>40). Same images from Figure 2.3b as staining controls (e).

Figure 2.7 EET signaling pathway activates AP-1 family members as primary transcriptional targets, and runx1 as secondary transcriptional target. a, Wild-type embryos were incubated with 300 μM cycloheximide, a translation blocker, for 30 min before the addition of 5 μM 11,12-EET at 24 hpf. Embryos were fixed for in situ hybridization at 25 hpf or 28 hpf. b. AP-1 transcription was induced upon 1 hr treatment with 11,12-EET, insensitive to cycloheximide inhibition. This means AP-1 induction does not depend on de novo protein synthesis, indicating AP-1 members are primary transcriptional targets of the EET signaling pathway. c. runx1 transcription was induced upon 4 hr treatment with EET (left two columns) and cycloheximide completely blocked EET-induced runx1 expression (right two columns). This suggests runx1 transcription depends on de novo protein synthesis of an upstream factor(s) upon EET stimulation, indicating that runx1 is a secondary transcriptional target of the EET signaling pathway. Representative images from each group (a total of n>30 from 2 independent experiments).

Figure 2.8 Knocking down junb/junbl inhibits HSPC specification in the AGM a, Wild-type (WT) embryos were injected with antisense morpholinos at 1-cell stage, and treated with DMSO or 5 μM 11,12-EET starting from 24 hpf. Embryos were fixed at 36 hpf for in situ hybridization of runx1. b, Knocking down junb completely blocked runx1 expression at 36 hpf both in the AGM and the tail non-haematopoietic tissue (middle row). In contrast, knocking down c-jun did not block the increase of runx1 (bottom row), consistent with the lack of c-jun upregulation in EET-treated embryos (data not shown). c, junb morphants still developed normal vascular structure in the AGM at 28 hpf, as shown by endothelial marker flk1 (c). Representative images from each group (a total of n>40 from 3 independent experiments).

Figure 2.9 PI3Kγ activation is specifically required for EET-induced gene expression signature. a, Similar to LY294002 (Figure 2.6e), another pan-PI3K/AKT inhibitor, wortmannin (1 μM) blocked EET- induced runx1 expression both in the AGM and tail. Representative images from each group (a total of n>60 from 3 independent experiments). b, Morpholinos specific to PI3Kγ, but not α, β, and δ subunits (data not shown), prevented EET-induced runx1 in the AGM and tail. Embryos injected at 1–2 cell stage with indicated amount of morpholino and treated with DMSO or 5 μM 11,12-EET from 24–36 hpf. In situ hybridization for runx1 performed at 36 hpf and percentages of embryos having high, medium, or low expression in the AGM and present or absent expression in the tail are shown. Graph summarizes 3 experiments, n≥10 embryos for each condition (0, 1, and 2 ng, bars show mean and s.e.m.) or one experiment n≥9 for all conditions (4 and 6 ng). c, The PI3Kγ specific inhibitor AS605240 (AS6) recapitulates the morpholino phenotype. Embryos treated from 24–36 hpf with DMSO or 5 μM 11,12-EET, with or without 0.3–1.0 μM AS6, then fixed and stained for runx1 at 36 hpf. DMSO, n=23; EET, n=33; EET+0.3 μM AS6, n=35; EET+1.0 μM AS6, n= 38. * p<0.05, *** p<0.001, two-tailed Fisher’s Exact Test.

To understand how 11,12-EET treatment leads to increased engraftment in already specified

HSPCs, we performed RNA-sequencing in human umbilical cord blood CD34+ HSPCs and a human myeloid cell line (U937), and utilized Ingenuity Pathway Analysis to decipher the biological pathways regulated by 11,12-EET in both cell types (Figure 2.10, digital supplement). Cell-to-cell signaling and cellular movement networks topped the list of activated biological pathways, including the AP-1 members, which have been shown to modulate cell migration in many cell types [220, 221]. AP-1 thus appears to be a common target of EET signaling, which leads to the induction of runx1 in the haemogenic endothelium

(Figure 2.6), and likely supports cell migration and cell-cell signaling of already-specified haematopoietic cells. In contrast, RUNX1 is not upregulated in already-specified HSPCs, which is consistent with previous studies showing that Runx1 is dispensable for HSPCs to engraft later haematopoeitic sites [33]. Several cytokines, such as CXCL8, OSM and CCL2, were also upregulated and involved in the cell migration network. These data show that besides promoting HSPC specification from the haemogenic endothelium,

11,12-EET can also directly induce gene expression programs beneficial for engraftment in already- specified HSPCs. Similarly, 11,12-EET treatment of zebrafish embryos after 48 hpf, when AGM HSPC production has already completed, leads to increased HSPCs in the CHT in a PI3Kγ-dependent manner, without affecting cell apoptosis or proliferation (Figure 2.11). Our data strongly suggest that 11,12-EET modulates cell migration and cell-cell interaction during HSPC engraftment.

2.3.3. 11,12-EET enhances HSPC engraftment and homing in mammals

To test the evolutionary conservation of EET-induced haematopoietic phenotypes, we examined the effect of 11,12-EET on HSPC engraftment in mammalian bone marrow competitive transplantation.

Consistently, 11,12-EET promoted greater short-term chimerism by 4 weeks post transplant (wpt) compared to control-treated cells (Figure 2.12a-b). Even up to 24 weeks, EET-treated marrow maintained greater multi-lineage contribution (Figure 2.12c). Enhanced short- and long-term engraftment suggests that

11,12-EET may impact both stem and progenitor cells, perhaps by establishing a competitive advantage at the early stage of engraftment. In a WBM homing assay, we found 11,12-EET promoted the initial seeding of progenitor cells in the bone marrow (Figure 2.12d-e). The early effect could be due to an enhanced cell migration and cell-cell signaling program, since assaying cell proliferation or apoptosis in whole marrow immediately after 11,12-EET treatment did not show significant changes (Figure 2.13). However, this does

33 not exclude the possibility of a later onset of anti-apoptotic effects upon transplantation. Finally we found transient inhibition of PI3K partially blocked EET-induced enhancement of long-term, multi-lineage engraftment following mouse bone marrow transplant (Figure 2.12f). Thus, the EET effect on enhancing

HSPC engraftment is evolutionarily conserved in fish and mammals.

Figure 2.10 11,12-EET up-regulates genes involved in cell-to-cell signaling and cellular movement in haematopoietic progenitors. a, Venn diagram showing a common set of 54 genes up-regulated (log2fc>0.5) after 2 hrs of 11,12-EET treatment (5 μM), both in human myeloid U937 cells and human umbilical cord CD34+ HSPCs (see also digital supplement for lists of up- and down-regulated genes). b–c, Ingenuity Pathway Analysis (IPA) of the overlapping gene set between the two cell types for enrichment of bio- functions. b, Biological processes, such as cell-to-cell signaling and cellular movement, were highly enriched, supporting EETs’ capability of enhancing engraftment (see also digital supplement for a comprehensive list of all biological functions predicted to be activated or suppressed based on the same gene set). c, Activation of recruitment of blood cells is caused by up-regulation of chemokines and cytokines such as CXCL8 and OSM after EET treatment, as well as by up-regulation of transcription factors, such as AP-1 genes (FOS). Orange dashed arrows depict activation. Shades of red represent the level of activation. Numbers underneath factors show RNAseq FPKM values in U937 cells.

Figure 2.11 11,12-EET treatment after HSPC specification still enhances the number of HSPCs in the CHT. a, Embryos were treated with DMSO or 5 μM 11,12-EET between 48–72 hpf to bypass the HSPC specification process in the AGM. 72 hpf embryos were fixed and tested on the following assays. b, In situ hybridization for cmyb, a marker for HSPCs. EET treatment significantly increased the staining, while LY294002, a pan-PI3K inhibitor, suppressed the effect. Representative images from each group (a total of n>60 from 4 independent experiments). c, A PI3Kγ-specific inhibitor AS605240 (AS6) also blocked the EET-induced increase of cmyb staining. Percentage of embryos having high, medium, or low expression in the CHT is shown. n≥11 for all conditions. Chi-square analysis. d, The increase of HSPCs in the CHT is not due to effects on proliferation. Immunofluorescence staining for phospho-Histone H3 (pH3) as a marker for proliferating cells. The number of pH3 positive cells was manually counted. Two-tailed t-test showed no significant difference between DMSO vs EET treated embryos. n=9 for DMSO, n=10 for EET. e, TUNEL staining as an assay for apoptotic cells. Apoptosis was minimal in the CHT at 72 hpf. As a staining control, obvious apoptosis was detected in the same embryos in the brain region, and was comparable between DMSO and EET treated embryos (data not shown).

Our unbiased chemical genetic studies elucidate a new eicosanoid pathway for haematopoiesis, which increases HSPC specification in the AGM by inducing AP-1/runx1, and also enhances HSPC engraftment by modulating several biological pathways, such as migration and cell-cell signaling. Previous work in our lab discovered a different eicosanoid, PGE2, could also enhance marrow engraftment [24, 25]. Both PGE2 and EETs are arachidonic acid-derived eicosanoids that are locally produced near wounds, and may facilitate progenitor recruitment, engraftment, and proliferation. Despite their common origin, the underlying molecular signaling mechanisms and activities of PGE2 and EETs are different (Table 2.1). Although the direct receptor for EETs is unknown, several studies have provided biochemical evidence that EETs bind to a G-protein coupled receptor (GPCR) [113, 222]. GPCRs signal through various Gα subunits [121].

Previously we showed that PGE2 signals through the cAMP-dependent Gαs-coupled PGE2 receptor for its pro-haematopoietic effects [25]. Using chemical inhibition and genetic loss-of-function approaches, we screened all families of zebrafish Gα subunits. Interestingly, we found that gna12/13 are specifically required for EET-induced AP-1 and runx1 expression (Figure 2.1). Inhibiting Gαs did not suppress the EET phenotypes, indicating EETs and PGE2 have different signaling mechanisms.

During marrow transplantation, the achieved chimerism over time is critical, and the time to adequate neutrophil engraftment is an important milestone for treatment success. In addition to improving long-term repopulation, EETs seem to have a prominent effect on progenitor engraftment, as shown by increased chimerism early after transplantation. Our studies highlight the importance of lipid mediators in regulating

HSPC engraftment, and the manipulation of these pathways could have clinical impact for patients undergoing transplantation.

Figure 2.12 11,12-EET enhances HSPC engraftment and homing in mammals. a, Schematic of mouse WBM (whole-bone marrow) competitive transplantation. RT, room temperature. b–c, 4 hrs of 11,12-EET treatment promoted short-term WBM engraftment at 4 wpt (b) and long-term multilineage engraftment at 24 wpt (c). WBC, white blood cells; M, myeloid cells; B, B cells; T, T cells. 2 independent experiments combined, n=20 total. d, Schematic of WBM competitive homing assay. hpt, hours post transplant. e,11,12- EET increased homing efficiency of Lin- cells and Lin-Kit+ HSPCs (n=5). f, PI3K activation is required for EET-enhanced mouse WBM engraftment (n=10). LY, 10 μM LY294002. Recipients characterized as engrafted or non-engrafted based on peripheral blood WBC chimerism, two-tailed Fisher’s Exact test (b,f); unpaired two-tailed t-test (c,e), mean with s.e.m..

Figure 2.13 11,12-EET treatment of mouse whole bone marrow (WBM) does not lead to immediate changes in cell proliferation or apoptosis. a, in vitro apoptosis assay on WBM treated with DMSO or 2 μM 11,12- EET for 4 hrs. The 7-AAD negative and AnnexinV positive population are the cells undergoing apoptosis. No significant differences between the two groups were observed either in Lin-Sca-Kit+ or Lin-Sca+Kit+ progenitor populations (n=4 each), mean with s.e.m.. b–c, in vitro proliferation assay on WBM treated with DMSO or 2 μM 11,12-EET for 4 hrs, in the presence of 10 μM BrdU. No significant differences between the two groups were observed either in Lin-Sca-Kit+ (b) or Lin-Sca+Kit+ populations (c) for any cell-cycle stage. Unpaired two-tailed t-test, n=4 each, bar showing mean. D, DMSO; E, EET; n.s., not significant.

Table 2.1 Comparison of major hematopoietic phenotypes and signaling pathways regulated by PGE2 and EETs.

Figure 2.14 Gα12/13 is specifically required for EET-induced phenotypes in zebrafish embryos. All embryos were treated with DMSO or 5 μM 11,12-EET between 24–36 hpf. Chemical inhibitors were added 30 min before EET. mRNA or morpholinos (MO) were injected at 1-cell stage. a–b, Inhibiting Gαs or Gαi had no effect on EET-induced runx1 expression. Embryos were categorized into two groups with either normal or increased runx1 expression level (n>20 each). PtxA, pertussis toxin A, 3 pg, inhibiting Gαi [223]; H89, 5 μM, PKA inhibitor downstream of Gαs[24]; SQ, SQ22536, 50 μM, adenylate cyclase inhibitor downstream of Gαs[24]. Representative images from each group (b) (a total of n>40 from 2 independent experiments). c–e, Synergistic effects of gna12/13a/13b knockdown on suppressing runx1 expression. Knocking down gna13a/b or gna12 alone partially inhibited EET-induced runx1 expression in the AGM and tail (c). gna12 MO: 2 ng; gna13a/13b MOs: 1 ng each. Triple morpholinos against gna12, gna13a and gna13b (0.67 ng each) completely blocked EET-induced multiple gene expression, including runx1, genes in regeneration (fosl2) and cholesterol metabolism (hmgcs1) (d), while other major tissue development processes were not significantly affected, such as notochord (shh), muscle (myoD), and blood vessels (flk, ephrinB2) (e). The results were quantified in (f). Embryos were categorized as having decreased, normal or increased runx1 expression. The bar graph represents the percentage of embryos in each group (n>30).

2.4. Methods

2.4.1. Zebrafish strains

Zebrafish were maintained in accordance with Animal Research Guidelines at Boston Children’s

Hospital (BCH). The following transgenic zebrafish were used in this study: Tg(β-actin:GFP) [214], casper

[215], Tg(flk1:DsRed2) [224], Tg(CD41:GFP) [225], Tg(Runx1+23:mCherry) [36] and Tg(Runx1+23:GFP)

[36]. The +23 enhancer region of mouse Runx1 was used to drive HSPC-specific expression

[226]. Tg(flk1:dnJUNB-2A-GFP) was constructed by cloning a human JUNBΔN into a tol2 transgenesis vector [227].

2.4.2. Chemical treatment

The ICCB Known Bioactive Library was purchased from BIOMOL (Enzo Life Sciences) and used for the adult zebrafish transplantation-based chemical screen. Chemicals were diluted at a 1:200 ratio. Chemicals used for the secondary round of screening for confirmation were from a different aliquot of the library, independent of the primary screen plate. 11,12-EET (Cayman Chemical, Cat. 50511) was resuspended in DMSO with original organic solvent evaporated. AS605240 (Sigma-Aldrich Cat. A0233) was resuspended in DMSO. The following chemicals were used for zebrafish marrow treatment: dmPGE2 (Cayman, Cat. 14750), 10 μM; BIO (EMD), 0.5 μM. 0.5 μM 11,12-EET and 14,15-EET were used for zebrafish WKM treatment (Figure 2.1e); 2 μM 11,12-EET for all mouse WBM treatment (Figure 2.12); 5 μM 11,12-EET for all zebrafish embryo treatment (Figure 2.3, Figure 2.6). The concentrations were chosen based on dose titration pilot experiments with doses spanning 0.1–50 μM. For the chemical suppressor screen, the suppressors were added 30 min prior to 11,12-EET. Zebrafish embryos were incubated with inhibitors at three different concentrations. The highest effective concentrations tested without causing general toxicity are listed in

Table 2.2.

Table 2.2 Chemicals used in the zebrafish embryo suppressor screen.

2.4.3 Adult zebrafish kidney marrow transplantation and chemical screen

Adult zebrafish transplantation-based chemical screen was done at the hESC core at BCH. 3- month old casper recipients (both male and female) received split-dose irradiation of 15 Gy each two days and one day prior to transplantation. Adult zebrafish kidney marrow cells from multiple donors were dissected, pooled together, processed into single-cell suspension and injected retro-orbitally as described previously [228]. Tg(β-actin:GFP) WKM cells were incubated with DMSO control or chemicals in 0.9xDPBS plus 5% heat-inactivated FBS for 4 hrs at room temperature, at a density of 1000 cells/μl. Chemicals were washed off before 20,000 treated Tg(β-actin:GFP) WKM and 80,000 untreated RedGlo® WKM were mixed together and co-injected into irradiated capser recipients. The number of recipients per treatment condition in the chemical screen (n=10) was estimated based on preliminary experiments comparing the WKM treated with DMSO or the positive control chemical, dmPGE2. In each experiment, recipients were randomly assigned to each treatment group. All primary hits were cherry-picked and tested in a secondary round of screening (n=10 each). Recipients that died before 4 wpt, mostly due to infection, were excluded from the analysis. No statistically significant association was observed between recipients’ survival rate and a particular drug treatment.

2.4.4 Adult zebrafish fluorescence imaging and quantification

All zebrafish WKM transplantation results shown were obtained at 4 wpt (weeks post transplant).

Transplanted adult casper recipients were anesthetized with 0.2% Tricaine and imaged using a Zeiss

Discovery V8 fluorescence stereomicroscope with GFP/RFP filters. To quantify the relative engraftment level in adult zebrafish, the kidney region was manually annotated for each fish, and the average fluorescence intensity of GFP and DsRed2 within the same region was measured (Gkid and Rkid) using

ImageJ. The average background fluorescence intensity (Gbkg and Rbkg) was measured in a region outside the fish and a mean from multiple images within an experiment was used for all the background subtraction.

The relative engraftment level was calculated as G/R= (Gkid - Gbkg) / (Rkid - Rbkg). The investigator analyzing the data was blind to the chemical treatment conditions. For the chemical treatment and screen results

(Figure 2.1e), the mean G/R in the DMSO group was normalized to 1, and all other groups were normalized to the mean G/R of DMSO. Normalized results from 2–3 independent experiments were pooled for the same chemical.

2.4.5 Zebrafish embryo live imaging

For live imaging, zebrafish embryos were embedded in agarose as described before [30, 36].

Single-frame image or time-lapse movies were taken on a spinning disk confocal microscope with an incubation chamber. Images of HSPC birth in the AGM were taken every 10 minutes. Images of the CHT engraftment process were taken every 2 minutes. Image post-processing and the creation of the supplementary videos were done with Fluorender, ImageJ, and Imaris.

2.4.6 Zebrafish embryo whole-mount in situ hybridization, anti-sense morpholino knockdown, and

mRNA overexpression

Whole-mount mRNA in situ hybridization experiments were performed based on the standard protocol with some modifications (http://zfin.org/zf_info/zfbook/chapt9/9.8.html). Embryos were scored blindly. All of the morpholinos were initially tested at 2, 4, and 6 ng to decide the effective dosage. If the morpholino did not produce a phenotype at 6 ng, additional higher doses were tested (8 ng, 12 ng), until the morpholino caused toxicity. See Table 2.3 for morpholino sequences. PtxA (Pertussis toxin A, Gαi inhibitor) mRNA (Addgene, Plasmid 16678) [223] was in vitro transcribed with SP6 RNA polymerase

(Ambion, mMESSAGE mMACHINE SP6, AM1340) and injected into 1-cell stage zebrafish embryos at 3 pg / embryo, causing morphological defects but no general toxicity.

2.4.7 Zebrafish embryo proliferation and apoptosis assays

Zebrafish embryos were chemically treated between 48–72 hpf and fixed at 72 hpf. For proliferation analysis, embryos were permeabilized and stained with primary antibody against phospho-Histone H3, and

FITC-conjugated secondary antibody. Embryos were imaged and phospho-H3 positive cells in the CHT were manually counted. Secondary antibody-only control showed no non-specific staining. For apoptosis analysis, embryos were stained using the colorimetric TUNEL staining kit (Promega).

Table 2.3 Sequences of morpholinos (MO) used in zebrafish embryo studies. * The junb_ATG morpholino is a published sequence [229]. But due to the high similarity of the first 23bp coding sequences between junb and junbl (only differ by 4 bp, including a consecutive 3 bp), this morpholino might potentially knock down both alleles at the same time. ** Published morpholino sequences [165]. *** Published morpholino sequence [230].

2.4.8 Cell culture

Human CD34+ cells were isolated from fresh umbilical cord blood by Ficoll separation of mononuclear cells and subsequent positive selection of CD34+ cells using magnetic beads (Miltenyi). Cells were treated in serum-free IMDM media (Sigma-Aldrich) with either DMSO or 5 μM 11,12-EET for 2 hrs at

37°C. U937 cells [231] were cultured in RPMI-1640 Medium (Sigma-Aldrich) and 10% FBS at 5% CO2 in air atmosphere according to the protocol (ATCC). For in vitro treatment, cells were serum-starved for one hour and then treated with either DMSO or 5 μM 11,12-EET for 2 hrs at 37°C. The conditions for use of human umbilical cord blood CD34+ cells are governed by the associated institution’s Internal Review Board

(IRB) on behalf of the DF/HCC in accordance with Department of Health and Human Services regulations at 45 CFR Part 46. Informed consent was obtained from all subjects.

2.4.9 Mouse bone marrow transplant

All mice were maintained according to IACUC approved protocols in accordance with BCH animal research guidelines. 9-week-old CD45.1 and CD45.2 (C57/BL6) male mice were purchased from Jackson

Laboratories and housed for 2–3 weeks before the experiments. All CD45.2 recipients received an 11 Gy split dose of γ-irradiation prior to transplantation, and were randomly assigned to each treatment group.

20,000 CD45.1 WBM cells from age- and gender- matched BL6 donors were treated in DMEM + 2% FBS

44 at room temperature for 4 hrs with 2 μM 11,12-EET. For the suppressor experiment (Figure 2.12f), 10 μM

LY294002 was added to the cells a half-hour prior to the addition of 11,12-EET. Chemicals were washed off before cells were resuspended in 1x PBS and mixed with 200,000 fresh CD45.2 mouse WBM cells.

Donor cells were retro-orbitally injected into CD45.2 recipients. Each treatment condition included 10 recipients per experiment. The 12-week survival rate in each experiment was 90–95%, and recipients that died before 12 wpt were excluded from the analysis.

2.4.10 Mouse peripheral blood chimerism analysis

Peripheral blood was stained with lineage-specific antibodies and analyzed on LSRII (BD

Biosciences) to assess engraftment. The following antibodies were used: Gr1 (RB6-8C5), Mac1 (M1/70),

B220 (RA3-B2), CD3 (145-2C11), and Ter119 from eBioscience; CD45.1 and CD45.2 from BD Biosciences.

The CD45.1 chimerisms in non-irradiated, untransplanted CD45.2 mice were used as a negative staining control. Recipients with multi-lineage chimerism above the average negative-control chimerism plus 3 standard deviations were considered to have multi-lineage engraftment (Figure 2.12f).

2.4.11 Mouse competitive homing assay

The mouse competitive homing experiment was performed as described, with modifications [232].

In brief, CD45.1 mouse WBM were treated with either DMSO or 2 μM 11,12-EET at room temperature for

3.5 hrs at a density of 2×106 cells/ml. DiO dye was added to the cell suspension (1:200) and incubated at

37°C for 30 min. At the same time, WBM from CD45.2 mice were incubated at RT for 3.5 hrs without chemical treatment, then labeled with DiD dye (1:200) at 37°C for 30 min. After the incubation and labeling, the chemicals and dyes were washed off. The DiO-labeled CD45.1 bone marrow and DiD-labeled CD45.2

WBM were mixed at a 1:1 ratio and competitively transplanted into CD45.2 recipients (2.5×106 from each donor). Recipients received total body irradiation of 11 Gy one day before transplantation. 16 hours after transplant, the recipients were sacrificed and bone marrow was analyzed by flow cytometry for both

DiO/DiD and surface lineage markers (Gr1, Mac1, B220, CD3, Ter119, from Ebioscience) and c-Kit (2B8,

BD Biosciences). The ratio between the percentages of DiO+ (Donor) and DiD+ (Competitor) cells within different cell populations was quantified. DiO and DiD are from VybrantTM Multicolor Cell-Labeling Kit

(Molecular Probes, V-22889).

2.4.12 Mouse bone marrow apoptosis and proliferation assays

For apoptosis analysis, mouse WBM cells were treated with DMSO or 2 μM 11,12-EET for 4 hrs in vitro and stained using the AnnexinV apoptosis kit (BD Biosciences), together with antibodies against lineage markers, Sca-1 (E13-161.7) and c-Kit (2B8). The 7-AAD-/AnnexinV+ cells are the apoptotic population. For proliferation analysis, mouse WBM were treated with DMSO or 2 μM 11,12-EET for 4 hrs in vitro, in the presence of 10 μM BrdU, then fixed, permeabilized and stained with anti-BrdU antibody (BD

Pharmingen BrdU Flow Kits) [233], together with antibodies against lineage markers, Sca-1 and c-Kit.

2.4.13 Gene expression profiling and IPA analysis

Gene expression profiling data have been deposited with GEO (Accession

No. GSE39707, GSE66767). For the zebrafish embryo gene expression study, total RNA was extracted from 36 hpf zebrafish embryos treated with DMSO or 5 μM 11,12-EET between 24–36 hpf, with 3 biological replicates each and n=25 in each group. Microarray hybridization was performed with the Affymetrix

GeneChip Zebrafish Genome Array. Hybridized microarray was background-corrected, normalized and multiple-tested using Goldenspike (http://www2.ccr.buffalo.edu/halfon/spike/) in R/Bioconductor [234].

Genes with q<0.1 by SNR test were considered differentially expressed (Appendix 1). For RNAseq analysis on human cells, total RNA was extracted from treated CD34+ and U937 cells with the RNeasy mini plus kit from Qiagen. After quality control on the Bioanalyzer (Agilent), total RNA was depleted of ribosomal RNA with the RiboZero gold kit (Epicentre). Enriched mRNA was applied to library preparation according to manufacturer’s protocol (NEBNext Ultra). After repeated quality control for average DNA input size of 300 bp, samples were sequenced on a HiSeq Illumina sequencer with 2×100 bp paired-end reads. Quality control of RNA-Seq datasets was performed by FastQC

(http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and Cutadapt [235] to remove adaptor sequences and low quality regions. The high-quality reads were aligned to UCSC build hg19 of the human genome using Tophat 2.0.11 without novel splicing form calls [236]. Transcript abundance and differential expression were calculated with Cufflinks 2.2.1 [237]. FPKM values were used to normalize and quantify each transcript. Log2fc (log2 fold change), p- and q- values were calculated. As the experiment was not performed in biological replicates, the p- and q- values were not taken into consideration for further analysis of the data. Results are listed with a cutoff of log2fc>0.5 for upregulated genes and log2fc<-0.5 for

46 downregulated genes in digital supplement. Analysis of overlapping up-regulated genes in both cell types after EET treatment was done using Venny (http://bioinfogp.cnb.csic.es/tools/venny/index.html). The list of overlapping genes was analyzed using Ingenuity® Pathway Analysis (IPA®, QIAGEN) to map enriched bio- functions.

2.4.14 Statistics

The comparison of multi-lineage engraftment in Figure 2.12d and f were done by two-tailed Fisher’s

Exact Test by comparing the number of engrafted vs non-engrafted recipients. Using the mean chimerism plus 2 × s.e.m. (standard error of mean) in the DMSO control group as the cutoff, recipients with a chimerism higher than the cutoff were considered engrafted (Figure 2.12d). Embryos in the in situ hybridization experiments were scored blindly and analyzed by Chi-square tests or two-tailed Fisher’s Exact Test in the case of small sample sizes. The rest of the statistics were done with unpaired two-tailed t-test. Graphs show mean with s.e.m..

2.5. Supplementary Discussion

The compatibility of the donor and recipient immune systems plays an essential role in the success of haematopoietic stem and progenitor cell (HSPC) transplantation. Among all of the factors influencing the immune systems, genes from the human major histocompatibility complex (MHC) locus have been shown to be the most important genes determining tissue histocompatibility. When the adult zebrafish whole kidney marrow (WKM) competitive transplantation assay was developed and the chemical screening was performed, knowledge of zebrafish MHC was very preliminary. In addition, although wild-type zebrafish have been inbred for generations, unlike the BL6 mice, they still harbor diverse sets of MHC genotypes.

This fact might have a substantial impact on the average survival and chimerism of transplanted zebrafish recipients in the adult zebrafish transplantation experiments performed in this study. In fact, on average we had about 50-70% of recipients surviving up to 4 wpt (weeks post transplant).

Even though the donors and recipients were not completely immune-matched, this does not compromise the reliability of evaluating the chemicals’ effects on enhancing WKM engraftment for the following reasons: (1) The donor WKM cells were dissected from multiple donors and pooled together before being split into different chemical treatment groups. Therefore, each treatment group should receive

47 cells with the same mixture of MHC genotypes; (2) In each chemical screen, more than 150 adult casper recipients were mixed, 10 of which were randomly assigned to each group. Therefore, the recipients’ MHC genotypes were also randomly distributed among all the treatment groups; (3) All screen hits were repeated in at least two more experiments with different groups of donors and recipients, the MHC genotypes of which should be independent of the previous experiments.

The next step in improving the overall efficiency and reproducibility of the zebrafish WKM transplantation assay is to more comprehensively understand the zebrafish MHC system and try to match the donor and recipient immune systems. We have already started this initiative recently in our lab. We think this will improve the survival rate and make the assay more suitable for evaluating long-term engraftment.

2.6. Acknowledgements

2.6.1. Acknowledgments

We thank C. R. Lee, M. L. Edin, and N. Gray for providing reagents; Y. Zhou, A. Dibiase, S. Yang,

S. Datta, P. Manos, R. Mathieu, and M. Ammerman for technical assistance; H. Huang for providing graphic illustration; R. M. White, T. E. North and C. Mosimann for discussion. Microarray studies were performed by the Molecular Genetics Core Facility at Boston Children’s Hospital, supported by NIH-P50-NS40828 and

NIH-P30-HD18655. S. Li in Y. Zhang’s lab at the Longwood HHMI joint core facility helped with RNA-seq.

This work was supported by HHMI and NIH grants R01 HL04880, P015PO1HL32262-32, 5P30

DK49216, 5R01 DK53298, 5U01 HL10001-05, R24 DK092760, and 1R01HL097794-04 (to L.I.Z.). This work was also funded, in part, by the Intramural Research Program of the NIH, National Institute of

Environmental Health Sciences (Z01 ES025034 to D.C.Z.), the National Cancer Institute grant

ROCA148633-01A5 (D.P.), and DFG and Care-for-Rare Foundation (V.B.). L.I.Z. is a founder and stockholder of Fate, Inc. and a scientific advisor for Stemgent. G.Q.D. is a member of the Scientific Advisory

Boards of MPM Capital, Inc., Epizyme, Inc., and iPierian, Inc.

2.6.2. Author Contributions

P.L. and L.I.Z. designed the study, analyzed data and wrote the manuscript, with help from J.L.L. and V.B.. P.L. developed the zebrafish competitive transplantation and performed the chemical screen with

48 technical help from E.K.P.. P.L. performed the mouse experiments with technical help from T.V.B., S.M. and G.C.H.. P.L. performed the zebrafish microarray and embryo chemical/genetic suppressor screens with technical help from E.B.R.. J.L.L. performed zebrafish embryo genetic studies and AGM timelapse imaging.

V.B. performed RNA-seq and analysis on human cells with technical help from F.G.B.. O.J.T. performed

CHT time-lapse imaging. T.M.S. provided the chemical library. D.P. and D.C.Z. offered reagents and information related to the EET study. All authors discussed the results and commented on the manuscript.

2.7. Addendum: Unpublished Cell Autonomy Data

We found that EET treatment of zebrafish embryos beginning at around 40hpf, after the bulk of

HSPC specification has already occurred in the AGM, increased the number of HSPCs in the CHT, likely by enhancing the migration of HSPCs to this niche (Figure 2.11). As HSPC colonization of the CHT may serve as a model of bone marrow colonization following transplant, I sought to understand whether EET enhances colonization autonomously via the HSPC, or non-autonomously through niche cells. A primary niche cell type of the CHT is the endothelial cell [36]. Sinusoidal endothelial cells are also an essential niche component of mammalian bone marrow [238]. I used Gateway cloning to create Tol2 transgenesis constructs featuring either the HSPC specific Runx1+23 promoter, or the endothelial specific flk1 promoter driving dominant negative forms of Gα13, PI3Kγ, and JunB, all pathways which we found to be downstream of EET signaling. These constructs were tagged with GFP in order to visualize their expression in the zebrafish. dnGα13 and dnPI3Kγ both appeared to be highly toxic to HSPCs and endothelial cells. I saw minimal GFP expression in the AGM or CHT of flk1: or Runx1+23:dnGα13 or dnPI3Kγ-GFP animals compared to control flk1:GFP or Runx1+23:GFP injections (Figure 2.15). None of these four constructs was transmitted through the germ line to F1 animals, suggesting that these signaling pathways are necessary for the survival of endothelial cells and HSPCs.

However, we did see expression of dnJUNB-GFP driven under both promoters, similar to what was seen with control GFP injections (Figure 2.15). Both the flk1 and the Runx1+23 constructs showed significant off-target expression in muscle, and this expression was maintained into the F1 generation in several independent lines, suggesting the dnJUNB construct may contain a cryptic zebrafish muscle promoter or enhancer. As described in our paper, flk1:dnJUNB-GFP blocked the enhancement of HSPC

49 specification in the AGM caused by EET treatment (Figure 2.6). We also investigated EET-driven colonization of the CHT in both lines. We found that only flk1:dnJUNB-GFP blocked EET enhanced HSPC staining in the CHT, while Runx1+23:dnJUNB-GFP had no effect (Figure 2.16). These data suggest that

EET-induced AP-1 signaling is required primarily in endothelial cells. In the AGM, this results in increased specification of HSPCs, and in the CHT there may be secondary signaling of endothelial cells to recruit or retain HSPCs. Alternatively, EET treatment may cause changes in CHT endothelial cell properties to create a better niche.

An endothelial site of action for EET is consistent with the literature [218, 239, 240]. EETs are known to be released by endothelial cells as well as smooth muscle [69, 241], to cause vasodilation [79,

242], to enhance endothelial wound closure in scratch assays[85], and to enhance angiogenesis in vitro

[82, 83, 85]. Additionally, in our zebrafish and mouse transplants, EET treatment was performed on whole marrow, which contains endothelial cells as well as HSPCs and many other blood cell types. EET treatment is performed for four hours, leaving ample time for secondary, non-autonomous cell signaling to occur.

Figure 2.15 Endothelial and HSPC driven expression of dominant negative factors. Embryos were injected at the one cell stage with DNA constructs. a) Runx1+23 driving expression mosaically in HSPCs. Left, Runx1+23:GFP control injections; right, Runx1+23: dominant negative factors as indicated. a) Flk1 driving expression mosaically in endothelial cells. Left, flk1:GFP control injections, right Flk:dominant negative factors as indicated. All images are of the CHT region of 60hpf zebrafish, taken with a Nikon Eclipse Ti spinning disc confocal microscope using a 20X. Scale bars are 50μM.

Figure 2.16 AP-1 signaling is required in endothelial cells for EET-induced engraftment of the CHT. All embryos were treated with DMSO or 5 μM 11,12-EET between 24–36 hpf. a) ISH for runx1 and cmyb expression in the CHT of 60hpf, F1 embryos of dnJUNB lines or GFP negative WT siblings. Scale bar is 100μM. b,c) Embryos were scored as having low, medium, or high expression in the CHT. p values are Fisher’s Exact test, two tailed.

2.8. Summary

In summary, in addition to its known roles in regulating cardiovascular tone and inflammation, we have identified EET to be an enhancer of hematopoietic stem cell engraftment. This role is conserved from development to adulthood in zebrafish, as well as in adult mouse transplant. EET’s pro-hematopoietic effects work through a Gα12/13 coupled receptor, which I will further investigate in the next chapter. This causes activation of PI3Kγ signaling and AP-1 signaling, the latter of which occurs autonomously in hemogenic endothelium cells, and is even conserved in human hematopoietic cells. Exogenous EETs, or a synthetic analog, hold therapeutic promise to improve outcomes for human transplant patients, especially in cases such as cord blood transplantation where HSPC numbers are limiting and an extended time to engraftment poses a danger to patients [15]. In the context of earlier results in our lab showing that a related eicosanoid, prostaglandin E2, also enhances HSPC specification and engraftment, though by different signaling pathways from EET, it seems that endogenous lipid signaling molecules are important regulators of HSC quiescence, migration, and differentiation. These lipid signalers are likely important as well in other stem cell contexts.

Chapter 3. Oxygenated Fatty Acids Enhance Hematopoiesis via the Receptor GPR132

Jamie L Lahvic, Michelle Ammerman, Pulin Li, Emma Stillman, Song Yang, Megan Blair, Constantina Christodoulou, Nan Chiang, Michael Chase, Olivia Weis, Yi Zhou, Charles N. Serhan, and Leonard I. Zon.

3.1. Attributions

In Chapter 2 I described the important role of EET signaling in regulating adult and embryonic hematopoietic phenotypes. In this chapter, I will describe experiments that led to the identification of

GPR132 as an EET receptor, a validation of this receptor’s requirement in zebrafish and mouse phenotypes, and an exploration of GPR132 ligands. RNAseq libraries from EET-binding and non-binding cell lines were prepared by myself and by Pulin Li. Song Yang and Yi Zhou mapped the raw reads to the genome. I compiled gene-level expression data for all GPCRs and used this to derive candidate EET receptors. I performed PathHunter β-arrestin assays to test GPCR candidates and GPR132 family members in vitro. I adapted the GPR132 β-arrestin assay to function in a 384 well plate, and used this to test a large number of fatty acids and other potential GPR132 activators.

I injected the gpr132b morpholino into zebrafish embryos. I designed zebrafish drug treatments for a large variety of fatty acids and other potential GPR132 activators. Drug treatments were performed by myself, Michelle Ammerman, or Emma Stillman. The three of us also performed in situs on treated embryos.

I scored AGM and tail expression, and imaged these embryos with assistance from Michelle and Emma.

Olivia Weis joined the lab as a summer student and also performed zebrafish drug treatment and in situ experiments.

I established the GPR132 knockout mouse colony in the Zon lab, and Michelle Ammerman and

Emma Stillman managed the breeding and genotyping of the colony. I designed and performed primary and secondary mouse transplant experiments and irradiated mice. I received technical assistance from

Constantina Christodoulou for retro-orbital injections, and from Michelle Ammerman, Emma Stillman and

Megan Blair for retro-orbital bleeds, peripheral blood stains, and general mouse care. I performed flow cytometry analysis on all peripheral blood samples. I depleted bone marrow samples by MACS, optimized the 8-color sort for LT/ST-HSCs and MPPs with assistance from Ron Bernier of the Children’s FACS core, and analyzed all bone marrow samples at the LSRII.

I performed all data analysis, compiled figures, and wrote the manuscript in preparation below.

3.2. Abstract

Epoxyeicosatrienoic acids (EETs) are endogenous lipid signaling molecules with cardioprotective and vasodilatory actions. We recently showed that exogenous addition of 11,12-EET causes pro- hematopoietic phenotypes in mouse and zebrafish marrow, as well as in zebrafish embryos. EET is known to signal via a G-protein coupled receptor, but the specific identity of this receptor remains unknown, impeding the progress of EET to the clinic and preventing genetic studies of the EET signaling pathway.

We developed a novel bioinformatic approach to identify the EET receptor. We found 10 candidate EET receptors that are expressed in three EET-binding human cell lines, but missing from an EET-non-binding line. Of these candidates, only GPR132 showed EET-responsiveness in vitro. Knockdown of zebrafish gpr132b prevented EET-induced hematopoiesis, and marrow from GPR132 KO mice showed a decreased ability to transplant long-term. Others have shown that GPR132 has affinity for a variety of fatty acids in vitro, and we found that these same fatty acids enhance hematopoietic stem cell specification in the zebrafish. We further conducted structure-activity-relationship analyses using both in vitro and in vivo assays on a variety of medium chain fatty acids. Oxygenated, unsaturated free fatty acids show high activation of GPR132, while unoxygenated or saturated fatty acids show a lower activity. 11,12-EET-methyl ester shows no activation of GPR132, suggesting that the carboxylic acid moiety is required for binding.

We have identified GPR132 as an EET receptor that mediates both embryonic and adult hematopoiesis, and we have further characterized the breadth of ligands able to activate GPR132.

3.3. Introduction

Eicosanoids are endogenous free fatty acids derived from arachidonic acid, responsible for a variety of physiological phenotypes. 11,12-epoxyeicosatrienoic acid (11,12-EET) is formed by Cytochrome

P450 enzymes, and is known to enhance endothelial migration, monocyte adhesion, and vasodilation, among other cell-type specific effects [59, 241]. Additionally, we recently showed that 11,12-EET enhances the specification of hematopoietic stem and progenitor cells (HSCs) in developing zebrafish embryos, as well as the transplant of HSPCs in both fish and mice [211]. We described the downstream signaling mediating these phenotypes, including the activation of PI3Kγ signaling and upregulation of AP-1

56 transcription factors, but the upstream receptor of 11,12-EET responsible for its hematopoietic and other phenotypes remains unknown.

Many EET phenotypes can be elicited with nanomolar or even picomolar concentrations of 11,12 or 14,15-EET [239, 240], suggesting that at least one specific, high-affinity receptor for EET exists. Although

EET has been proposed to have an intracellular site of action for certain phenotypes [112, 243], a bead- tethered EET without the ability to cross the plasma membrane maintained an aromatase inhibiting activity, suggesting that EET activates a membrane-bound receptor [110]. Several groups have demonstrated a requirement of G-protein signaling for EET phenotypes [73, 85, 222], and Chen et al demonstrated that

U937 and other EET-responsive cell lines express a single high affinity EET receptor of about 47 kDa in size [113],. We demonstrated that EET’s hematopoietic phenotypes were similarly dependent on G-protein signaling via Ga12/13 [211].

As traditional biochemical methods have so far failed to identify the EET receptor, here we used bioinformatic techniques to identify candidate receptors and assayed those candidates for EET- responsiveness in vitro. Only GPR132, a previously described fatty acid receptor, showed responsiveness to EET. We then demonstrated that GPR132 is required for EET-induced hematopoietic stem and progenitor cell specification in the zebrafish, and for normal hematopoietic stem cell transplant in the mouse.

Previously described fatty acid activators of GPR132 induced hematopoietic phenotypes in the zebrafish identical to those seen with EET, further confirming that these molecules activate the same pathway. As it is unusual for a GPCR to show such promiscuous binding to diverse molecules, we performed structure- activity relationship analyses to determine the full range of GPR132 activators. “EET-like” fatty acids with double bonds and oxygenated groups in the middle of the carbon chain all showed in vitro and in vivo activity, while deviation from this structure correlated with a loss in activity.

3.4. Results

3.4.1. Bioinformatic identification of EET receptor candidates

A single EET receptor has been described in multiple cell types, but the identity of this receptor remains unknown [113, 222, 244, 245]. As traditional biochemical methods have failed to identify the EET receptor, we employed a novel approach which relied on expression analysis. We performed RNAseq

57 profiling in duplicate on four human cell lines, three that show binding to a radiolabelled EET analog, and one that shows no such binding ([113], Campbell lab personal correspondence). We hypothesized that this non-binding cell line (HEK293) would fail to express the EET receptor at the RNA level. We filtered the expression data for non-olfactory GPCRs and averaged the FPKM (fragments per kilobase per million reads) data for the duplicates. While we detected reads for hundreds of GPCRs, each cell line expressed only a few dozen GPCRs above an FPKM of 0.3 (Figure 3.1, see Appendix 2 for full expression data).

Among the three EET-binding cell lines (U937 monocytes, EaHy endothelial cells, and PC3M-LN4 prostate cancer cells), 37 GPCRs were expressed in common above the conservative threshold of 0.3

FPKM (Figure 3.2a). Of these, 27 were also expressed at high levels (FPKM>0.9) in our EET non-binding cell line (Figure 3.2b). This left 10 candidate EET receptors that were expressed only in EET binding cell lines and were missing from the non-binding cell line. All candidates had predicted molecular weights within

20% of the predicted 47 kDa size of the EET receptor [113], and candidates included both well-studied

GPCRs such as the β-adrenergic receptor and prostaglandin receptors, as well as orphan GPCRs such as

GPR35 and GPR132 (Table 3.1).

3.4.2. GPR132 is only candidate receptor to respond to EET in vitro

To test for receptor activation by 11,12-EET, we assayed recruitment of β-arrestin by each of our receptor candidates using the PathHunter assay ([200, 201], Figure 3.3). While known positive control ligands robustly and dose-dependently activated their respective receptors for several candidates (Figure

3.4c, f-h, ADRB2, HRH1, PTGER2, PTGER4), EET showed no activity in these assays. EET also showed no activity in two assays that lacked positive control ligands, CCRL2 and GPR35 (Figure 3.4d,e). For three candidates (GPR68, LPAR6, and PTGER1), no β-arrestin assay was available. In contrast, 11,12-EET dose-dependently recruited β-arrestin to the GPR132 receptor, indicating binding and activation of that receptor (Figure 3.4b). GPR132 is a member of the GPR4 family of GPCRs, which includes GPR4, GPR65, and GPR68. No β-arrestin assay is available for GPR68, but 11,12-EET showed no activation of GPR4 or

GPR65 (Figure 3.5).

Figure 3.1 Histograms depicting FPKMs of GPCRs in EET-binding (A) and non-binding (B) cell lines. RNAseq was performed on duplicate biological samples, and FPKMs per gene were averaged. Gray region indicates GPCRs considered as EET receptor candidates, elaborated in Figure 3.2 and Table 3.1.

Figure 3.2 Bioinformatic identification of candidate EET receptors. A) RNAseq profiling was performed in duplicate on the EET-binding cell lines U937, PC3M-LN4, and EaHy. Expression levels of GPCRs in the duplicates were averaged, and GPCR expressed at greater than 0.3 fragments per kilobase per million were compared. B) GPCR expressed in common in all three EET-binding cell lines were compared to those found in HEK293 cells, an EET non-binding cell line.

Table 3.1 Candidate EET receptors. Average FPKM values from RNAseq of 2 biological replicates for each cell type is shown. Candates needed to be expressed at greater than 0.3 FPKM in the EET binding cell lines U937, EaHy, and PC3M, and expressed at less than 1.0 FPKM in the non-binding cell line HEK293. Predicted molecular weight according to SwissProt for each candidate is shown.

average FPKM MW Candidate U937 EaHy PC3M HEK293 (kDa) ADRB2 8.4 13.7 1.0 0.7 46.5 CCRL2 2.5 2.5 1.5 0.1 39.5 GPR132 2.8 0.4 3.2 0.2 42.5 GPR135 1.2 1.4 2.0 0.7 51.7 GPR68 0.9 0.6 0.4 0.0 39.5 HRH1 1.4 9.8 1.8 0.2 55.8 LPAR6 0.5 2.6 2.4 0.4 39.4 PTGER1 20.9 0.4 2.0 0.3 41.8 PTGER2 12.1 0.4 1.1 0.0 39.8 PTGER4 10.8 11.3 0.6 0.1 53.1

Figure 3.3 PathHunter β-arrestin recruitment assay. PathHunter cells express a tagged GPCR. Upon ligand activation, recruitment of the tagged β-arrestin causes complementation of a β-galactosidase enzyme, which can catalyze a luminescent reaction.

Figure 3.4 11,12-EET activates GPR132 in vitro. A) Diagram depicting PathHunter β-arrestin assays. B-H) β-arrestin assays for candidate GPCRs treated with their known ligands or 11,12 or 14,15-EET. At each dose, cells were treated in triplicate and luminescence values were averaged. As basal luminescence varies significantly across assays for different proteins, luminescence values were normalized to the no treatment control. Cells were treated in 384 well plates according to manufacturer’s protocol. Each graph is representative of at least two similar experiments. Bars represent standard error. Note varying y-axes.

Figure 3.5 EETs and 11-HETE fail to activate GPR132 family members GPR4 and GPR65. A,B) β-arrestin assays for GPR4 and GPR65. At each dose, cells were treated in triplicate and luminescence values were averaged. As basal luminescence varies significantly across assays for different proteins, luminescence values were normalized to the no treatment control. Cells were treated in 384 well plates according to manufacturer’s protocol. Each graph is representative of at least two similar experiments. BTB09089 is a reported agonist of GPR65. Bars represent standard error. Note varying y-axes.

3.4.3. Gpr132b is required for EET phenotypes in zebrafish embryo

We next investigated the requirement of GPR132 for EET’s hematopoietic phenotypes in zebrafish embryos. Zebrafish have two GPR132 homologs, gpr132a and gpr132b. At the amino acid level, they share

39% and 44% sequence similarity with human GPR132, respectively, and both conserve a histidine residue in the fourth transmembrane domain hypothesized to be important for acid sensing [181]. Previous RT-

PCR studies showed gpr132b but not gpr132a expression in whole zebrafish embryos [182].

We designed a splice-blocking morpholino against gpr132b and injected 4-6ng into 1-cell zebrafish embryos. Embryos were then treated from 24-36 hours post fertilization (hpf) with 5uM 11,12-EET. Embryos were fixed at 36 hours and stained for runx1. This marker is normally expressed in hemogenic endothelial cells of the aorta-gonad-mesonephros (AGM). We showed previously that 11,12-EET increases the specification of hematopoietic stem and progenitor cells in the zebrafish AGM, and causes ectopic expression of runx1 in the tail mesoderm [211]. As described previously, in wildtype, uninjected embryos,

EET caused an increase in runx1 staining in the AGM and tail (Figure 3.6). However, in gpr132b morpholino-injected embryos treated with EET, no increase in HSPC specification was seen in the AGM, and the percentage of embryos with tail staining was significantly decreased compared to uninjected controls. No difference was seen in the AGMs of wildtype versus gpr132b morpholino-injected embryos.

GPR132 is thus required for EET’s enhancement of HSPC specification in zebrafish embryos.

Figure 3.6 gpr132b is required for EET’s enhancement of zebrafish developmental hematopoiesis. A) in situ hybridization for gpr132b expression in the embryo. Single-cell zebrafish embryos were injected with 4-6ng of gpr132b MO, then treated with DMSO or 5uM 11,12-EET beginning at 24hpf. Embryos were fixed at 36hpf and stained for runx1 expression. Scale bar is 100μN. B) Embryos were scored as having high, medium, or low runx1 expression in the AGM, and high (present) or low (absent) runx1 expression in the tail mesoderm. Graph presents summary of two experiments. Fisher’s exact test two-tailed..

3.4.4. GPR132 KO mouse shows impaired long-term marrow transplant

GPR132 KO mice have normal overall physiology, although they are known to show an age-related autoimmune disease featuring an increase in B cells and T cells, and lymphocytic infiltration of tissues

[246]. We found that 3.5-5 month old GPR132 KO mice have normal distributions of B cells, T cells, and granulocytes, as well as comparable numbers of long-term HSCs, short-term HSCs and multipotent progenitors compared to their wildtype or heterozygous siblings (Figure 3.7). To stringently test HSPC function, we performed limiting dilution whole marrow transplants from GPR132 knockout and heterozygous mice. We combined 10,000, 50,000, or 200,000 donor cells from CD45.1 or CD45.1/2, GPR132 -/- or +/- mice with 100,000 wildtype CD45.2 competitor cells. We transplanted these into irradiated CD45.2 recipients and observed chimerism over time (Figure 3.8a). At a limiting, 10,000 cell dose, GPR132 knockout marrow shows a functional defect in competitive long-term engraftment compared to sibling heterozygous marrow (Figure 3.8b), indicating that GPR132 is required for normal HSC function. No difference was seen between heterozygous and knockout marrow at the higher cell doses. At the low cell dose, the defect was consistent across hematopoietic lineages and in Lin- and LSK+ fractions of the marrow

(Figure 3.9). There was no difference in lineage contribution between heterozygous or knockout marrow, and the proportion of mice showing multilineage chimerism decreased over time in the knockout case

(Figure 3.9c,e). These data suggest an HSC defect, rather than a progenitor defect in GPR132 -/- mice.

Figure 3.7 GPR132 KO mice have normal basal hematopoiesis. A) Peripheral blood lineage distribution of 3.5-4.5 month old WT (n=4), GPR132 +/- (n=5) and GPR132 -/- (n=4) mice. B cells were marked with B220, T cells with CD3, and myeloid cells with Gr-1. B) Percentage of LSK, MPP, ST-HSC, and LT-HSC in 3.5-5 month old WT (n=3), GPR132 +/- (n=2), and GPR132 -/- (n=3) marrow.

We chose representative recipient mice from the 10,000 and 50,000 cell doses and dissected marrow from these to observe CD45.1/2 chimerism and perform secondary transplants (Figure 3.10).

Generally, peripheral blood and marrow chimerism correlated within each mouse, and matched peripheral blood chimerism from secondary recipients at 4 weeks post-transplant. At the limiting 10,000 cell dose, knockout marrow continued to show impaired engraftment compared to heterozygous marrow. We will continue to follow these mice until 16 weeks post secondary transplant and perform final peripheral blood and marrow readouts.

Figure 3.8 GPR132 is required for normal marrow transplant in the mouse. A) Diagram of primary and secondary marrow transplant using GPR132 KO mice or heterozygous siblings as donors. Donor mice were GPR132 +/- or GPR132 -/- and CD45.1 or CD45.1/2. 10,000-200,000 donor cells were combined with 100,000 CD45.2 wildtype competitor cells and transplanted into lethally irradiated CD45.2 recipients. Mice were bled at indicated timepoints and cells were stained for donor chimerism. At 6 months post transplant, representative primary recipients that had received 10,000 or 50,000 cells were sacrificed and 2 million whole marrow cells were transplanted into secondary recipients. The remainder marrow was stained for HSC-SLAM markers and CD45.1/2. B) Peripheral blood chimerism in transplant recipients. Graphs summarize data from two separate experiments with a total of 13 (10,000 and 50,000 cell dose) or 9 (200,000 cell dose) primary recipients per condition. 6 primary recipients were used as donors for 3 secondary recipients each. Mean with SEM is plotted. *, p<0.01, Fisher’s exact test two-tailed setting, considering mice with >0.5% chimerism to be engrafted. C) Peripheral blood lineage contribution of donor cells for each genotype and cell dose at 6 months post primary transplant. B cells were marked with B220, T cells with CD3, and myeloid cells with Gr-1.

Figure 3.9 GPR132 is required for normal marrow transplant in the mouse. A,B) Peripheral blood donor chimerism over time for individual mice transplanted with 10,000 GPR132 +/- cells (A) or 10,000 GPR132 -/- cells (B) and 100,000 competitor cells. Chimerism in each lineage is shown. Some mice have not yet reached 6 months post transplant. Secondary data is average of 3 secondary recipients transplanted from a single secondary donor. Secondary data is further broken down in Figure 3.10, below. For myeloid and B lymphoid engraftment, some mice had 0% chimerism, these were plotted as 0.01% in order to represent them on a log scale. C) Lineage contributions of donor derived cells at 6 months post transplant. There were no significant differences among groups. D) Marrow chimerism of individual mice transplanted with 10,000 cells at 6 months post transplant. Lineage- and Lineage-,Sca-1+,cKIT+ populations are shown.One mouse had 0% LSK chimerism, this was plotted as 0.001% in order to represent it on a log scale. E) Mice receiving 10,000 cells with peripheral blood chimerism above 0.5% in myeloid, B cell, and T cell compartments at indicated time-points. Some mice have not yet reached 6 months post transplant.

Figure 3.10 Bone marrow chimerism of primary recipients correlates with peripheral blood chimerism of primary as well as secondary recipients. 6 mice which had received 10,000 or 50,000 GPR132 heterozygous or knockout cells were chosen based on their 6 month peripheral blood readout to be secondary donors. Shown is donor chimerism in peripheral blood and marrow LSK cells of primary recipients 6 months post transplant, as well as peripheral blood of secondary recipients 4 weeks post transplant. 2 million cells from each primary recipient was transplanted into 3 lethally irradiated secondary recipients. Some recipients died due to infection before the 4 week time-point. In particular, secondary donors 2 and 3 from the 50,000 cell GPR132 +/- condition had no surviving recipients. PB, peripheral blood; LSK, lineage-, Sca-1+, cKit+.

We dissected marrow from GPR132 +/- and GPR132 -/- mouse donors and treated with DMSO,

11,12-EET, or 9-HODE, then transplanted 20,000 cells into irradiated WT recipients along with 200,000 untreated WT competitor cells. We saw no enhancement of transplant by fatty acid treatments in either genotype (Figure 3.11b). This contrasted with results seen in similar experiments treated GPR132 +/+ marrow, where 11,12-EET treatment robustly enhances transplant (Figure 3.11b, Figure 2.12). These results could indicate that GPR132 must be present at full wildtype levels in order to respond to fatty acid treatment, and that even heterozygous loss of GPR132 prevents fatty acid responsiveness in the mouse marrow. We are eager to repeat these experiments comparing GPR132 +/+ marrow head to head with heterozygous or knockout marrow to confirm that the difference between genotypes has a biological basis and is not a technical difference between the two experiments. In these experiments we saw no differences in engraftment of GPR132 +/- and -/- marrow, suggesting that the impairment in -/- marrow is only apparent at limiting cell doses (10,000 cells).

Figure 3.11 GPR132 may be required at full wildtype levels for responsiveness to fatty acids in marrow transplant. A) Wildtype donor marrow was treated with DMSO or 4uM 11,12-EET for 4hrs at 37C. B) In a separate experiment, GPR132 heterozygous or knockout marrow was treated in identical conditions with DMSO, 11,12-EET, or 9-HODE. In each case 20,000 treated marrow cells were mixed with 200,000 untreated competitor cells and transplanted into irradiated wildtype recipients Overall peripheral blood chimerism at 12 weeks post transplant is shown.

3.4.5. Diverse oxygenated fatty acids activate GPR132 in vitro and cause EET-like phenotypes in

vivo

GPR132 is a reported receptor for bioactive lipids such as 9-HODE and 11-HETE [191, 200, 201,

203]. Like 11,12-EET, these molecules are oxygenated, unsaturated free fatty acids (See Table 3.2 for chemical structures). We confirmed this previous data and showed that both of these molecules recruited

β-arrestin downstream of GPR132 in vitro dose-dependently (Figure 3.12a). Additionally, we treated developing zebrafish embryos with 9-HODE and 11-HETE from 24-36hpf, then fixed the embryos and performed an in situ hybridization to examine expression of the hematopoietic markers runx1 and c-myb.

9-HODE and 11-HETE caused identical phenotypes in zebrafish embryos, increasing runx1/c-myb expression in both the AGM and tail of the fish (Figure 3.13a,b,Table 3.2). The tail staining of runx1/c-myb is a phenotype we have uniquely seen with EET treatment, despite conducting a wide variety of chemical treatments of embryos in our lab. The high similarity between EET, 9-HODE, and 11-HETE phenotypes strongly suggests these molecules are activating the same receptor, namely GPR132.

11,12-EET or related activators of GPR132 represent a potential therapeutic for enhancing hematopoietic stem cell transplant. As GPR132 showed comparable reactivity to 11,12-EET, 9-HODE, and

11-HETE, we sought to further explore potential GPR132 ligands by conducting structure activity relationship analyses. We assayed a variety of bioactive lipids for their ability to recruit β-arrestin to GPR132 in vitro and to enhance runx1/c-myb expression in zebrafish embryos in vivo. We found that simple chain, oxygenated, unsaturated free fatty acids including 11,12-EET, 11-HETE, 9-HODE, and 11,12-DHET robustly enhanced both of these phenotypes (Figure 3.12a, Figure 3.13). More saturated lipids such as

9,10 and 12,13-EpOME showed lower affinity in β-arrestin assays and showed little activity in the zebrafish in vivo. Fatty acids such as 13-HODE and 20-HETE, with their oxygenated groups displaced farther from the carboxylic acid, similarly showed decreased activity in vivo and in vitro. A recent paper also identified

GPR75 as the 20-HETE receptor [247]. Oxygenation seems to contribute to GPR132 activation, as the unoxygenated free fatty acids linoleic acid, arachidonic acid, eicosapentaenoic acid, and docosahexaenoic acid also showed reduced activity (Figure 3.12a, Figure 3.13,Table 3.2).

Figure 3.12 Diverse oxygenated free fatty acids activate GPR132 in vitro. A-F) GPR132 β-arrestin recruitment assays. Drugs were treated in triplicate and each graph represents at least two similar experiments. Error bars represent SEM. Note differing y-axis scale.

Figure 3.13 Fatty acid GPR132 activators enhance expression of HSPC markers in zebrafish embryos. A) in situ hybridization of zebrafish embryos treated with DMSO, 11,12-EET, or 9-HODE showing AGM and tail expression of runx1/c-myb. For full quantification, see Table 1. B) in situ hybridization of zebrafish embryos treated with diverse free fatty acids, EET-methyl ester, and lysophosphatidylcholine (LPC) showing tail expression of runx1/c-myb. For full quantification, see Table 1. Embryos were dechorionated at 24hpf and incubated with the indicated quantities of drug from 24-36hpf. Embryos were fixed at 36hpf and stained using a probe combination of runx1 and c-myb. Scale bar is 100μm.

Table 3.2 Diverse free fatty acids enhance HSPC specification in zebrafish embryos. Embryos were treated with the indicated concentrations of drug from 24-36hpf, then fixed and stained for runx1/c-myb expression. AGMs were scored as having high, medium, or low expression, and tails were scored as having high (present) or low (absent) expression. Doses which caused morphological defects or circulation problems were disregarded. Data here is a summary of many experiments. Percentages of embryos having high expression in the AGM and tail are shown, as well as the total number of embryos assayed for each molecule. Each drug was tested in at least two separate experiments. Each experiment included clutch- matched negative (DMSO or ethanol) and positive (11,12-EET) control treatment conditions

AGM Tail Drug Structure Dose (μM) % High N % High N DMSO 12.6 % 388 0.5 % 393 Ethanol 12.2 % 278 0.0 % 278 11,12-EET 5 41.2 % 585 63.8 % 585 3 24.2 % 33 3.0 % 33 11,12-DHET 5 51.6 % 31 16.1 % 31 10 61.0 % 77 48.1 % 77 7 17.2 % 29 48.4 % 31 9-HODE 10 57.4 % 47 10.6 % 47 5 23.1 % 26 0.0 % 27 8 6.3 % 32 3.3 % 30 13-HODE 12 14.9 % 67 0.0 % 66 20 17.2 % 29 0.0 % 29 3 14.3 % 28 7.1 % 28 11-HETE 5 15.2 % 33 12.1 % 33 10 68.1 % 69 13.0 % 69 5 19.0 % 21 50.0 % 22 20-HETE 8 13.3 % 30 26.7 % 30 12 13.0 % 23 72.7 % 22 5 14.0 % 57 0.0 % 57 9,10-EpOME 10 37.0 % 54 13.0 % 54 20 32.8 % 58 81.0 % 58 5 16.7 % 42 0.0 % 42 12,13-EpOME 10 31.1 % 61 1.7 % 60 20 20.4 % 49 0.0 % 49 10 4.8 % 42 0.0 % 42 Arachidonic Acid 15 17.4 % 23 0.0 % 23 5 12.0 % 25 0.0 % 26 8 8.5 % 59 0.0 % 60 Linoleic Acid 12 12.5 % 56 0.0 % 54 20 29.2 % 24 4.2 % 24 7 8.8 % 34 0.0 % 34 Eicosapentaenoic 15 34.2 % 38 0.0 % 38 Acid 30 20.5 % 44 10.0 % 40 7 3.4 % 59 0.0 % 59 Docosahexaenoic 15 13.7 % 51 0.0 % 51 Acid 30 11.1 % 54 0.0 % 54 5 22.2 % 27 0.0 % 27 11,12-EET-methyl 7 23.2 % 82 0.0 % 82 ester 10 17.5 % 57 0.0 % 57 15 15.1 % 53 1.9 % 53 10 35.3 % 34 0.0 % 34 LPC 20 35.7 % 28 0.0 % 28 10 24.3 % 37 0.0 % 37 SPC 20 17.6 % 17 0.0 % 17

11,12-EET-methyl ester, which replaces the carboxylic acid of 11,12-EET with a methyl group, had no activity either in β-arrestin assays or in zebrafish embryos, suggesting the carboxylic acid is absolutely required for binding to GPR132 (Figure 3.12a, Figure 3.13, Table 3.2). This may substantiate other reports that have identified GPR132 and closely related proteins as pH-sensitive receptors that activate in acidic conditions [181, 183, 196-198]. However, organic acids such as isethionic acid, gluconic acid, and boric acid did not cause recruitment of β-arrestin by GPR132 (Figure 3.12f). Another group recently showed that lactic acid can activate GPR132 in mouse and human macrophages [248]. However, in our hands, lactic acid did not activate β-arrestin recruitment by GPR132, and did not induce HSPC marker staining in zebrafish (Figure 3.14). Many groups have reported lysophosphatidylcholine (LPC) and sphingosylphosphorylcholine (SPC), to be activators of GPR132 ([179, 180, 186-196], reviewed in [249,

250]), although direct binding experiments linking LPC to GPR132 were later retracted [184, 185], and some of these studies have been disputed [251].

Figure 3.14 Lactic acid fails to activate GPR132 in vitro or enhance HSPC marker expression in vivo. A) PathHunter β-arrestin assay for GPR132 activation. Drugs were tested in triplicate at indicated concentrations. Data is representative of two experiments. B) Summary in situ data for AGM and tail expression of runx1/c-myb in drug treated embryos. Embryos were treated with ethanol, 5uM 11,12-EET, or the indicated concentrations of lactic acid beginning at 24hpf. Embryos were fixed at 36hpf and stained for runx1/c-myb expression. Embryos were scored as having high, medium, or low expression in the AGM, and high (present) or low (absent) runx1 expression in the tail mesoderm. Graph presents summary of two experiments, total number of embryos per group is indicated at the base of columns. P-values are Fisher’s Exact Test, two-tailed.

3.5. Discussion

Our data show that 11,12-EET and related lipids such as 9-HODE and 11-HETE are able to activate

GPR132 to enhance developmental hematopoiesis. Our mouse studies additionally show that GPR132 is an endogenous regulator of HSC function. Further studies will elucidate whether GPR132 is also the receptor that mediates classical EET phenotypes in other physiological contexts, such as the hyperpolarization of smooth muscle cells or endothelial cells [70, 73-79, 116]. These phenotypes sometimes show greater stereospecificity than we saw in our assays. For instance, EET stimulation of

BK(Ca) channels in the mouse lung was enhanced by inhibition of she, the enzyme that converts EET to

DHET [252], suggesting only the epoxide is active. However, in our β-arrestin assays and zebrafish embryo treatments, EET and DHET were both active. EET activation of BK(Ca) channels has also been reported to be a high affinity interaction [240]. The affinity of GPR132 for oxygenated fatty acids remains unclear, as accurate affinity measures cannot be made from either our in vivo or in vitro methods. In zebrafish embryos, limited tissue accessibility of fatty acids may artificially depress the observed affinity. Serum in the media of β-arrestin assays can bind fatty acids and decrease their availability to cells [253]. Direct biochemical interrogation of EET-GPR132 binding is therefore still needed. Another assay of interest would be macrophage activation. Mouse M1 macrophages exposed to lipopolysaccharide (LPS) undergo macrophage polarization and produce inflammatory cytokines, but EET treatment at a 1uM concentration can effectively prevent this effect [254]. GPR132 is expressed in macrophages, and was recently reported to induce macrophage polarization in response to lactic acid [248].

It is possible that EET or other fatty acids activate GPR132 indirectly by affecting membrane dynamics and receptor localization, conformation, or dimerization. One study found that LPC activated

GPR132 by inducing its localization from endosomes to the plasma membrane, where it signaled constitutively [186]. DHA concentrations within the plasma membranes of rod cells can affect the likelihood of rhodopsin to enter the active conformation [153]. However, lipid ligands are also known to directly bind

GPCRs. While traditionally GPCR ligands interact with the extracellular loops and N-terminal domain of the protein, an intra-membrane site of entry has been hypothesized for sphingosine-1-phosphate and its receptor [126], as well as for cannabinoid receptors [127].

As an oxygenated free fatty acid receptor regulating hematopoiesis, GPR132 is a promising target for further biological and therapeutic studies. Ex vivo EET treatment of human cord blood may activate

GPR132 and improve transplantation, helping to overcome limiting cell numbers. Alternatively, since the structure-activity relationship analyses described here have defined the major chemical properties of

GPR132 agonists (simple chain, oxygenated, unsaturated, free fatty acids), medicinal chemistry could be used to synthesize even stronger GPR132 agonists. For basic biological studies, GPR132 provides a genetic handle by which we can study the EET signaling axis. GPR132 is expressed in B cells, T cells, and macrophages, and its potential expression in hematopoietic stem cells warrants further study. Thus, EET-

GPR132 signaling could regulate HSPC engraftment autonomously or non-autonomously. Conditional and tissue-specific knockout of GPR132 could be used to interrogate the endogenous role of EET signaling and determine which cell types are directly EET-responsive.

Other groups have reported that EET activates diverse receptors including TRPV/TRPC receptors

[116-118], prostaglandin receptors [114] [115], and in one case the G-protein coupled receptor FFAR1

[105]. We do not see expression of FFARs in our zebrafish embryos (data not shown), and voltage gated channels or nuclear hormone receptors do not explain the G-protein dependence of many EET phenotypes, including the hematopoietic phenotypes described here. It remains to be seen whether these receptors may be downstream of GPR132, or whether EET is able to activate different receptors depending on cellular context and concentration.

GPR132, also known as G2A, has been described variously as a receptor for free fatty acids [191,

200, 201], a receptor for lysophosphatidylcholine and sphingosylphosphatidylcholine [179, 180, 186-196], or a proton and pH sensitive receptor [181, 183, 196-198]. Both our β-arrestin recruitment assays and our in vivo experiments in the fish and mouse suggest that GPR132 is responsive only to free fatty acids.

Neither LPC, SPC, nor a variety of organic acids showed any effect on GPR132 activity. GPR132 is a member of the OGR1 family of GPCRs, which includes GPR4, GPR65 (TDAG8), as well as GPR68

(OGR1), which was an additional candidate EET receptor in this study. We saw no activity of EET in β- arrestin recruitment via GPR4 or GPR65 (Figure 3.5), and the Campbell lab has seen no activity in GPR68 assays (Bill Campbell, personal correspondence). One overexpression study suggested that GPR68 and

GPR132 heterodimerization could increase their sensitivity to low pH [255]. It’s possible that GPR132 alone

76 is fatty acid responsive, while the heterodimer is pH-sensing. Further experiments are needed to test these proteins in combination.

This work further confirms the importance of bioactive lipids in regulating hematopoiesis. Our group has now shown regulation of developmental hematopoiesis and transplant by EETs as well as prostaglandins. The roles of fatty acids such as HETEs, HODEs, and leukotrienes in regulating inflammation [62, 256-258], as well as B cell production, adhesion, and Ig production [259-262], and T cell function [263-265] are well studied. These reports hint at a complex network of lipid signaling molecules regulating diverse biological phenomena. Systematic understanding of this network is sorely needed.

3.6. Methods

3.6.1. Materials

All fatty acids, EET-methyl ester, and isoproterenol were purchased from Cayman Chemicals at 5-

10mM stock concentrations in ethanol or DMSO. Histamine dihydrochloride (H7250), LPC (L1881), SPC

(S4257) and lactic acid (L6661) were purchased from Sigma-Aldrich and diluted in ethanol. Recombinant human chemerin and MIP-3β were purchased from PeproTech, Inc.

3.6.2. Cell culture and RNAseq

U937 and EaHy cells were cultured in RPMI media with Pen/Strep and heat inactivated FBS.

PC3M-LN4 cells were cultured in RPMI media with Pen/Strep, heat inactivated FBS, and L-glutamine.

HEK293 cells were cultured in DMEM with Pen/Strep and heat inactivated FBS. For each cell type, 5 million cells were collected in duplicate and RNA was isolated by Qiagen RNeasy Plus Kit. Ribosomal RNA was depleted using the RiboGone H/M/R kit from Illumina, and libraries were prepared using the NebNext platform. Fragments were sequenced at a depth of approximately 30 million reads. Quality control of RNA-

Seq datasets was performed by FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and

Cutadapt [235] to remove adaptor sequences and low quality regions. The high-quality reads were aligned to

UCSC build hg19 of human genome using Tophat 2.0.11 without novel splicing form calls [236]. Transcript abundance was calculated with Cufflinks 2.2.1 [237]. FPKM values were used to normalize and quantify each transcript. FPKM for GPCRs (according to the gpcr.org database) were compiled. FPKM values from two

77 replicates for each cell type were averaged. EET receptor candidates had FPKM>0.3 in U937, EaHy, and

PC3M-LN4 cells, and FPKM<0.9 in HEK293 cells.

3.6.3. PathHunter β-arrestin assays

PathHunter β-arrestin assays were purchased from DiscoverX and performed in 384 well plates according to manufacturer instructions. In each experiment, drugs were plated in triplicate across a range of concentrations. Luminescence values were normalized to baseline within each experiment. Each drug was tested in a minimum of two separate experiments. Drug stocks were prepared in DMSO or ethanol.

EC50 values and non-linear regressions were performed in GraphPad Prism using the log(agonist) vs response function. In some cases, drug concentrations above 100μM could cause precipitation, resulting in inhibition of the luminescence reaction and highly variable readouts across technical replicates, in this case, max concentrations were disregarded for analysis.

3.6.4. Zebrafish embryo experiments

Zebrafish were housed and cared for according to IACUC protocol 14-10-2789. Casper or AB zebrafish were used for all experiments. For morpholino injections, embryos were injected at the 1 cell stage with 4-6ng MO, mixed with phenol red as an indicator of injection success. The Gpr132b_sp MO was produced by GeneTools, with a sequence of TAAAATGGCGTTGCTCTTACCTCTA. For drug treatments, embryos were dechorionated using pronase at 24hpf, and drug was added to E3 media and embryos in

12-well plates. Embryos were incubated at 32°C. 11,12-EET was always included as a positive control, and live embryos were observed 1-4 hours after treatment for evidence of tail “bubbling”- a precursor to runx1/c- myb expression in the tail and an indicator that the 11,12-EET is functional. Any experiment without bubbling was deemed unsuccessful and discarded. Embryos were also monitored for normal morphology and circulation. Embryos were fixed at 36 hpf in 4% PFA, and in situs were performed as described [266].

3.6.5. Mouse blood and marrow analysis and transplant

Mice were housed and cared for according to IACUC protocol 15-06-2964. GPR132 knockout mice

(Jackson Labs #008576) were kindly provided by Donna Bratton. Mice were genotyped for GPR132 status by PCR as previously described [246]. For peripheral blood analysis, mice were bled retro-orbitally at 3.5-

4.5 months old. Cells were incubated with fluorescently labeled antibodies as described below to stain B cell, T cell, and myeloid lineages, and for transplanted mice CD45.1 and CD45.2. For marrow analysis,

78 mice were sacrified at 3.5-5 months old. Femur, tibia, and hip bones were dissected and crushed to get whole marrow cells. Cells were stained with lineage and HSC-SLAM antibodies as described below.

For limiting dilutions transplants, two male donor mice between 10-12 weeks of age were sacrificed.

Donors were siblings, one KO and one heterozygous for GPR132. In the first transplant, the GPR132 HET donor was CD45.1/CD45.2 heterozygous, while the GPR132 KO donor was CD45.1 homozygous. In the second transplant, both donors were CD45.1 homozygous. For fatty acid treatment transplants, female

CD45.1+ donors of the indicated genotype between 10-12 weeks of age were sacrificed. GPR132 +/- and

GPR132 -/- donors were siblings, while the GPR132 +/+ donor was an unrelated CD45.1+ mouse.

Donor and competitor marrow from femur, tibia, and hip was harvested by crushing. A sample of the donor was red cell lysed by incubation with ACK lysing buffer (Thermo Fisher) and remaining cells were stained with trypan blue and counted. For drug treatments, cells were suspended in StemSpan SFEM media at a density of 1 million cells per mL. 5μM 11,12-EET or 9-HODE or an equivalent volume of 9-HODE was added and cells were incubated in a 37°C water bath for 4 hours. Drug was washed off before transplant. In limiting dilution transplants, the indicated number of donor cells were mixed with 100,000

CD45.2 wildtype (Jackson Labs #000664), sex- and age-matched competitor marrow cells. In fatty acid treatment transplants, 20,000 treated cells were mixed with 200,000 untreated sex- and age-matched

CD45.1/2 double positive competitor cells. Cells were transplanted retro-orbitally into sex- and age- matched CD45.2 wildtype recipients which had received 10 gy irradiation in a split dose. Recipients were maintained on sulfatrim antibiotics for the first 4 weeks after transplant, and were bled to observe chimerism at 4 weeks, 12 weeks and 6 months post-transplant. In the limiting dilution transplants, at 6 months, recipient mice were sacrificed and their marrow was analyzed for CD45.1 chimerism as described below and used in secondary transplants. In secondary transplants, 2 million cells from a total of 24 primary recipients were transplanted into 3 CD45.2 secondary recipients each. Secondary recipients were bled at 4 weeks, 12 weeks, and 16 weeks post-transplant to observe CD45.1 chimerism. Recipients were sacrificed at 16 weeks for final marrow analysis.

3.6.6. Antibody staining and flow cytometry analysis

Mouse peripheral blood was incubated with ACK lysing buffer (Thermo Fisher) two times for 5 minutes at room temperature in order to lyse and deplete red blood cells. Lysing buffer was washed off,

79 and the remaining cells were incubated for 1 hour at 4°C with 1:100 dilutions of the following antibodies:

Ter119-PE-Cy5 (red blood cells, eBioscience), CD3-APC (T cells, eBioscience), B220-Pacific Blue (B cells, eBioscience), Gr1-PE-Cy7 (clone RB6-8C5, granulocytes, eBioscience), CD45.1-PE (BD Pharmingen), and CD45.2-FITC (BD Pharmingen).

To quantify numbers of long-term HSCs, short-term HSCs, MPPs, and LSK cells in GPR132 KO mice, mouse bone marrow was incubated with the following antibodies: Ter119-PE-Cy5 (eBioscience

1:100), CD11b-PE-Cy5 (eBioscience 1:100), CD3e-PE-Cy5 (eBioscience 1:100), Gr-1-PE-Cy5

(eBioscience 1:100), B220-PE-Cy5 (eBioscience 1:100), Ly6A/E-PE (eBioscience 1:100), CD117-APC

(eBioscience, 1:100), CD48-eFluor450 (eBioscience 1:100), CD34 FITC (eBioscience, 1:33), and CD150

(SLAM)-PeCy7 (Biolegend, 1:100).

To determine marrow engraftment at 6 months post transplant, marrow cells were lineage-depleted by MACS (Miltenyi Biotec), and then stained with CD117-APC (eBioscience, 1:100) Ly-6A/E-PE-eFluor®

610 (eBioscience, 1:100), CD48-Alexa Fluor 700 (Biolegend, 1:50), CD34 FITC (eBioscience, 1:33), CD150

(SLAM)-PeCy7 (Biolegend, 1:100), CD45.1-PE (BD Pharmingen, 1:100), and CD45.2-APC/780

(eBioscience, 1:100) as well as with Streptavidin-eFluor450 lineage antibodies from eBioscience (Ter119,

CD11b, CD3e, Gr-1, B220). Cells were gated for LSK, CD48-, and MPP: CD34+, CD150-; ST-HSC: CD34+,

CD150+; and LT-HSC: CD34-, CD150+ as well as for CD45.1/2 positivity.

All stainings were performed in mouse buffer (PBS + 1% Pen/strep + 2%FBS) at 4°C for 1 hour.

Antibodies were then washed off and cells were analyzed on an LSRII Flow Cytometer. FlowJo software was used to process the data.

Chapter 4. GATA Factor-G-Protein-Coupled-Receptor Circuit Suppresses Hematopoiesis

Xin Gao, Tongyu Wu, Kirby D. Johnson, Jamie L. Lahvic, Erik A. Ranheim, Leonard I. Zon, and Emery H. Bresnick

4.1. Attributions

What follows is a paper published in Stem Cell Reports in March of 2016 on which I am a co-author

[267]. I performed all of the zebrafish experiments for this paper, including designing and injecting morpholinos against GPR65, and observing the effects on HSPC markers as well as more broad hematopoietic markers, and analyzing this data. Michelle Ammerman provided technical support with some of these in situ hybridizations. Xin Gao and members of the Bresnick lab performed the mouse and bioinformatics experiments. Xin Gao and Emery Bresnick wrote the manuscript, and I contributed significantly to the editing.

4.2. The OGR1 family of GPCRs

In the previous chapter I described how GPR132 is the likely receptor for a variety of fatty acids that can enhance hematopoiesis. GPR132 is one member of the small OGR1 family of GPCRs, which also includes OGR1 (GPR68), TDAG8 (GPR65), and GPR4. All four of these molecules have been reported to have proton-sensing activity, although GPR132’s activity is weaker [183]. One report recently found that co-expression of GPR68 and GPR132 can lead to increased proton sensing activity, suggesting the two can heterodimerize and respond to pH [255]. Indeed there is also one report of heterodimerization of

GPR68 and GPR4 [268]. In PathHunter β-arrestin assays, we did not see EET-sensitivity of GPR4 or

GPR65 (Figure 3.5), and the Campbell lab did not see EET-sensitivity of GPR68 in their β-arrestin assay

(W. Campbell, personal correspondence). In our hands, GPR132 also did not show proton-sensitivity

(Figure 3.12, Figure 3.14).

Until recently, the OGR1 family had few described roles in hematopoiesis. However, the Bresnick lab found that Gata2 is a major regulator of the endothelial-to-hematopoietic transition during HSPC development, and they set out to look for GPCRs regulated by this activity. They found that mouse Gpr65 was specifically regulated by GATA-2 and GATA-1, while Gpr132 is regulated only by GATA-2. We realized then, that multiple members of the OGR1 family may be able to regulate hematopoiesis. So, to understand the importance of this link, I performed experiments in the zebrafish to manipulate gpr65 and gata2a expression. We were surprised to observe that knockdown of gpr65 led to an increase in HSPC staining in

82 the AGM, suggesting that GPR65 is a negative regulator of hematopoiesis. This contrasts with the role of

GPR132 as well as with the role of GATA-2, the upstream regulator of GPR65. This data was the first hint that GPR65 might be part of a negative feedback loop. The Bresnick lab was then able to confirm with experiments in mice and cell culture that GPR65 is a negative regulator of GATA-2 and hematopoiesis.

The format of this article has been adjusted for publication in this thesis. The article can be found in its original format at: http://dx.doi.org/10.1016/j.stemcr.2016.01.008. The Creative Commons license for this article can be found at: http://dx.doi.org/10.1016/j.stemcr.2016.01.008

4.3. Abstract

Hematopoietic stem cells (HSCs) originate from hemogenic endothelium within the aorta-gonad- mesonephros (AGM) region of the mammalian embryo. The relationship between genetic circuits controlling stem cell genesis and multi-potency is not understood. A Gata2 cis element (+9.5) enhances Gata2 expression in the AGM and induces the endothelial to HSC transition. We demonstrated that GATA-2 rescued hematopoiesis in +9.5−/− AGMs. As G-protein-coupled receptors (GPCRs) are the most common targets for FDA-approved drugs, we analyzed the GPCR gene ensemble to identify GATA-

2-regulated GPCRs. Of the 20 GATA-2-activated GPCR genes, four were GATA-1-activated, and only Gpr65 expression resembled Gata2. Contrasting with the paradigm in which GATA-2-activated genes promote hematopoietic stem and progenitor cell genesis/function, our mouse and zebrafish studies indicated that GPR65 suppressed hematopoiesis. GPR65 established repressive chromatin at the +9.5 site, restricted occupancy by the activator Scl/TAL1, and repressed Gata2 transcription. Thus, a Gata2 cis element creates a GATA-2-GPCR circuit that limits positive regulators that promote hematopoiesis.

4.4. Introduction

Establishment and maintenance of the adult hematopoietic system requires the generation of hematopoietic stem cells (HSCs) from a unique endothelial cell (hemogenic) in the aorta-gonad- mesonephros (AGM) region of the mammalian embryo [269]. HSCs develop in clusters that bud off from hemogenic endothelium [30, 32], a process termed endothelial to hematopoietic transition (EHT). HSCs migrate to and colonize the fetal liver and, subsequently, the bone marrow [3]. Whereas major efforts have focused on defining regulatory proteins/networks governing EHT, many questions remain unanswered regarding the molecular constituents and mechanisms.

Master regulatory transcription factors co-localize at cis elements of target genes in hematopoietic stem and progenitor cells (HSPCs) to establish genetic networks that control hematopoiesis [270-276]. The combinatorial mechanisms operating in hemogenic endothelium, and the relationship between mechanisms governing EHT and HSC multi-potency, are unclear. A shared component of the mechanisms involves the

84 transcription factor GATA-2, which is required for definitive hematopoiesis [277]. GATA-2 functions in hemogenic endothelium to induce EHT and regulates HSC function [278-282].

Since GATA-2 induces EHT, it is instructive to consider factors/signals upstream of GATA-2.

Deletion of a cis element 9.5 kb downstream of the Gata2 promoter (+9.5) in mice decreased Gata2 expression in AGM hemogenic endothelium, deregulated genes encoding positive regulators of hematopoiesis, and abrogated EHT [279]. Deletion of a cis element 77 kb upstream of the promoter (−77) reduced Gata2 expression in myelo-erythroid progenitors and impaired progenitor function without affecting EHT [283]. The defective HSC generator of +9.5−/− embryos depleted HSPCs in the fetal liver and caused lethality at embryonic day 13–14 (E13–14) [280]. Since the +9.5 controls Gata2 expression and EHT [279, 284], and GATA-2 occupies the +9.5 [271, 285], one aspect of the +9.5 mechanism involves GATA-2-mediated positive autoregulation. Factors implicated upstream of

GATA-2 include bone morphogenetic protein 4 [286, 287], Notch signaling [288, 289], the Ets factor Etv2

[290], and the methylcytosine dioxygenases Tet2/Tet3 [291].

The Gata2 +9.5 site regulates a large gene cohort in hemogenic endothelium, and these genes do not parse into a single pathway [279]. The genes include Runx1, Lyl1, and Mpl, positive mediators of HSC generation and/or function. The identification of vital constituents of the +9.5-dependent genetic network will reveal how a cis element triggers EHT. As GATA-2 lacks features that can be leveraged for drug binding, identifying +9.5 network components will reveal strategies to promote HSC generation/function for transplantation and inhibit leukemia cell proliferation and survival.

The most common targets for drugs approved by the Food and Drug Administration are GPCRs

[292]. The GPCR family consists of 341 non-olfactory receptors classified as rhodopsin, secretin, glutamate, adhesion, and frizzled/taste2, based on sequence homology [120]. CXCR4, a rhodopsin-like GPCR, recognizes SDF-1/CXCL-12 and controls HSPC survival, proliferation, migration, and engraftment [293].

CXCR4 antagonists are used as mobilizing agents for stem cell transplantation. The prostaglandin PGE2, which functions through a group of related GPCRs [294], expands HSPCs [23, 25, 295]. We described +9.5- mediated upregulation of Gpr56 expression [279]. Loss-of-function analyses indicated that GPR56 promotes EHT [296] and contributes to HSC maintenance [297]. Galpha(s)-mediated signaling controls

HSPC engraftment of bone marrow [298].

By profiling expression of the GPCR cohort in the AGM, we discovered a subset of GATA-2- regulated GPCRs and a GATA-2- and GATA-1-regulated cohort, including GPR65, which suppresses AGM hematopoiesis. The +9.5 site, and its regulated protein GATA-2, which positively regulate hematopoiesis, upregulated Gpr65 encoding a negative regulator of hematopoiesis. These results provide evidence for a

GATA factor-GPCR type I incoherent feedforward loop as a vital component of the genetic network that controls HSPC generation and function.

4.5. Results

4.5.1. GATA-2 Expression in 9.5-/- AGM Rescues Hematopoiesis

Deletion of the +9.5 site reduced Gata2 expression in AGM hemogenic endothelium and abrogated

EHT [279] (Figure 4.1a). These results suggest that factors/signals conferring +9.5 activity control Gata2 expression, which promotes EHT. In principle, the +9.5 might regulate other genes in cis or in trans that control EHT. To determine whether the HSC generation defect of +9.5−/− AGM results from insufficient GATA-2 production, we tested whether GATA-2 expression rescues the defect. E11.5 mouse embryo AGMs were infected with GATA-2-expressing retrovirus (Figure 4.1a). After 96 hr of explant culture, we quantitated endothelial and hematopoietic cell populations in infected (GFP+) cells by flow cytometry using CD31 and c-KIT surface markers. The infection efficiency was similar among the three conditions

(+9.5+/+-empty vector, +9.5−/−-empty vector, +9.5−/−-GATA-2), as indicated by the indistinguishable percentage of GFP+ live cells (Figure 4.1b). GATA-2 expression restored Gata2 mRNA to the wild-type level in CD31+c-KIT− endothelial and CD31+c-KIT+ hematopoietic cells (Figure 4.1c), and rescued both populations (Figure 4.1d,e). Since the infected cells express GFP, the percentage of GFP+cells in each population increased in +9.5−/− AGM infected with GATA-2-expressing retrovirus (Figure 4.1f). However, retroviral-mediated GATA-2 expression did not rescue CD31+c-KIT− endothelial and CD31+c-

KIT+ hematopoietic cell populations in the GFP− cells (Figure 4.2). Thus, the +9.5−/− AGM hematopoiesis defect resulted from insufficient GATA-2 production, is associated with reduced CD31+c-KIT−endothelial cells, and can be rectified by restoring GATA-2.

To assess whether the GATA-2-rescued CD31+c-KIT+ hematopoietic cells are functional, we measured their capacity to generate myelo-erythroid colonies in a colony-forming unit (CFU) assay. After

86 culturing retroviral-infected AGMs for 96 hr, GFP+ CD31+c-KIT+ hematopoietic cells were isolated by fluorescence-activated cell sorting (FACS) and assayed for their capacity to generate BFU-E (erythroid burst-forming units), CFU-GM (granulocyte, macrophage), and CFU-GEMM (granulocyte, erythrocyte, monocyte/macrophage, megakaryocyte) colonies. Whereas +9.5−/−AGM failed to generate CFUs, GATA-2 expression in the +9.5−/− AGM induced CFUs comparable with +9.5+/+ AGM (Figure 4.3a). Quantification of colony types revealed comparable numbers of rescued CFU-GM colonies in comparison with wild-type

AGM (Figure 4.3b). GATA-2 expression induced BFU-E and CFU-GEMM colonies (Figure 4.3b). Colonies derived from wild-type and rescued samples were morphologically indistinguishable (Figure 4.3c). Wright-

Giemsa staining of cells from colonies revealed normal myeloid and erythroid cell generation from the

GATA-2-expressing +9.5−/− AGM (Figure 4.3d). The GATA-2-induced CD31+c-KIT+hematopoietic cells exhibited qualitatively and quantitatively normal activity.

4.5.2. Global GPCR Analysis in the AGM: Discovery of a GATA Factor-Regulated GPCR Cohort.

The +9.5 site confers Gata2 expression and establishes a genetic network involving known HSC regulators and genes not implicated in hematopoiesis [279]. To discover vital constituents of the network, especially those with potential for modulation by small molecules/drugs, we systematically analyzed the expression pattern of non-olfactory GPCRs (Figure 4.4a). The human genome encodes greater than 800

GPCRs, which are categorized into rhodopsin, secretin, glutamate, adhesion and frizzled/taste2 families based on sequence homology; 341 are distinct from the olfactory and taste GPCRs [120, 292]. Using our

AGM RNA-sequencing (RNA-seq) data [279], we parsed AGM-expressed GPCRs into the canonical categories: secretin, 5% (15); adhesion, 7% (22); glutamate, 5% (15); frizzled/taste2, 6% (20); and rhodopsin, 70% (221) (Figure 4.4b).

Figure 4.1 GATA-2 Expression in +9.5−/− AGM Rescues CD31+c-KIT+ Hematopoietic and CD31+c- KIT− Endothelial Cells. (A) AGM ex vivo retroviral infection and culture. (B) Flow cytometric analysis of GFP+ cells within total live cells (6 litters: +9.5+/+-Empty [n = 8 embryos]; +9.5−/−-Empty [n = 4 embryos]; +9.5−/−-GATA-2 [n = 6 embryos]). (C) RT-PCR analysis of Gata2 mRNA levels in FACS-sorted CD31+c-KIT− and CD31+c-KIT+ cells (six litters: +9.5+/+-Empty [n = 8 embryos]; +9.5−/−-Empty [n = 4 embryos]; +9.5−/−-GATA-2 [n = 6 embryos]). (D) Representative flow cytometric plots of CD31+c-KIT+ and CD31+c-KIT− cell populations in infected AGMs after 96 hr of ex vivo culture. (E and F) Quantitation of flow cytometry data expressed as percentage of CD31+c-KIT− and CD31+c-KIT+cells in GFP+ cells (E) and the percentage of GFP+ cells in CD31+c-KIT− and CD31+c-KIT+ cells (F) (6 litters: +9.5+/+-Empty [n = 8 embryos]; +9.5−/−-Empty [n = 4 embryos]; +9.5−/−-GATA-2 [n = 6 embryos]). Error bars represent SEM. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 (two-tailed unpaired Student's t test).

Figure 4.2 Retroviral-mediated GATA-2 expression does not rescue CD31+c-KIT+ hematopoietic and CD31+c-KIT- endothelial cells in the GFP- cell population. (A) Representative plots from flow cytometric analysis of CD31+c-KIT+ and CD31+c-KIT- cell populations in GFP- cells after 96h of ex vivo culture. (B) Quantitation of flow cytometry data expressed as the percentage of CD31+c-KIT- and CD31+c-KIT+ cells in GFP- cells (6 litters: +9.5+/+- Empty [n=8 embryos]; +9.5-/--Empty [n=4 embryos]; +9.5-/--GATA-2 [n=6 embryos]). Error bars represent SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (two-tailed unpaired Student’s t-test).

Figure 4.3 Rescued CD31+c-KIT+ Cells Exhibit Normal Colony-Forming Unit Activity. (A and B) Quantitative analysis of colony-forming activity of FACS-sorted CD31+c-KIT+ cells (6 litters: +9.5+/+-Empty [n = 9 embryos]; +9.5−/−-Empty [n = 7 embryos]; +9.5−/−-GATA-2 [n = 10 embryos]). (C) Representative BFU-E, CFU-GM, and CFU-GEMM colonies from FACS-sorted CD31+c-KIT+ cells. Scale bar, 2 mm. (D) Representative images of Wright-Giemsa-stained cells from colonies. Ery, erythroblast; Mac, macrophage; My, myeloid precursor; Neu, neutrophil. Scale bar, 40 μm. Error bars represent SEM. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 (two-tailed unpaired Student's t test).

Figure 4.4 Global GPCR Analysis in the AGM. (A) Global GPCR analysis.(B) Non-olfactory GPCRs (N = 314) were categorized into Secretin, Adhesion, Glutamate, Frizzled/Taste2, and Rhodopsin families based on sequence homology. (C) Classification of 85 GPCRs expressed in the AGM (>5 transcripts per million) into five families. (D) Bar graph depicting GATA-2-regulated genes from RNA-seq analysis of +9.5+/+ and +9.5−/− AGMs [279]. Black bars, genes co-regulated by GATA-1 according to our prior microarray analysis of G1E-ER-GATA with or without β-estradiol treatment [299]. (E) Gata2, Adora3, Gpr65, Ltb4r1, and P2ry1 expression during erythropoiesis. B, basophilic erythroblast; O, polyorthochromatic erythroblast; P, proerythroblast; R, reticulocyte (http://www.cbil.upenn.edu/ErythronDB/). (F) Time course of Gata2 and Gpr65 expression following β-estradiol treatment in G1E-ER-GATA cells (n = 3 independent experiments). (G) RT-PCR analysis of Gata2 and Gpr65 in FACS-sorted R1, R2, R3, and R4/5 populations from fetal liver (n = 3 independent experiments). (H and I) ChIP signal map for Gpr65 in human CD34 cells (H) [270], mouse HPC7 cells [274], Lin− bone marrow cells [300], and G1E cells [301] (I). (J and K) RT- PCR analysis of Gata2 and Gpr65 mRNA in +9.5+/+and +9.5−/− AGM (5 litters: +9.5+/+ [n = 8 embryos]; +9.5−/− [n = 6 embryos]) and yolk sac (three litters: +9.5+/+ [n = 7 embryos]; +9.5−/−[n = 5 embryos]) (J), and MAE cells expressing GATA-2 (K) (n = 3 independent experiments). (L) RNA-seq of Gata2 and Gpr65 mRNA in FACS-sorted endothelial cells (EC), hemogenic endothelial cells (HEC), hematopoietic cells (HC), and hematopoietic stem cells (HSC) from the AGM [296]. Error bars represent SEM. ∗p < 0.05; ∗∗∗p < 0.001 (two-tailed unpaired Student's t test).

Figure 4.4 Continued.

To discover GPCRs that control HSC generation and/or activity, we evaluated GPCR expression in the AGM. Of the 314 GPCRs annotated by RNA-seq, 85 were expressed at >5 transcripts per million

(Figure 4.4c), a level that can be validated with high frequency by real-time RT-PCR. Of the 85 GPCRs, 20 were downregulated in the +9.5−/− AGM versus +9.5+/+ AGM (Figure 4.4d), indicating GATA-2-regulation.

Using our previous microarray dataset (untreated or β-estradiol-treated G1E-ER-GATA-1 erythroid precursor cells) [299], we found that four of the 20 GATA-2-regulated GPCRs were GATA-1-regulated

(Figure 4.4d). Our strategy refined the 314 non-olfactory, AGM-expressed GPCRs to yield Adora3, Gpr65, Ltb4r1, and P2ry1, which are GATA-2- and GATA-1-regulated.

4.5.3. A GATA Factor-GPCR Incoherent Feedforward Loop Suppresses AGM Hematopoiesis

As shared gene expression patterns can infer functional interconnectivity, we compared Adora3, Gpr65, Ltb4r1, P2ry1, and Gata2 expression patterns in mouse cells and tissues.

Only Gpr65 resembled Gata2, both being expressed in HSPCs and mast cells (Figure 4.5)

(http://biogps.org). As GATA-1 represses Gata2 during erythroid maturation via a GATA switch, Gata2 expression declines upon erythroid maturation. Mining the Erythron Database, which provides transcriptomics data during erythroid precursor cell maturation into erythrocytes [302], revealed Gata2 and Gpr65 repression upon erythroid differentiation. Gpr65 was the only one of the four

GPCR genes to have a Gata2-like expression pattern (Figure 4.4e). To further compare expression patterns, we quantitated Gata2 and Gpr65 mRNA in G1E-ER-GATA-1 cells treated with β-estradiol to induce erythroid maturation and in FACS-sorted R1, R2, R3, and R4/5 fetal liver cell populations. Gpr65 and Gata2 were repressed during erythroid maturation (Figure 4.4f,g). These correlations are consistent with GATA-2 upregulating Gpr65 expression in the AGM and may point to a functional link between GATA-2 and GPR65. Chromatin immunoprecipitation (ChIP)-sequencing analysis in human (Figure 4.4h) and mouse (Figure 4.4i) cells revealed GATA-2 occupancy at Gpr65, suggesting that GATA-2 directly regulates Gpr65 transcription. Comparison of +9.5+/+ and +9.5−/− AGM revealed that reduced Gata2 expression in the +9.5−/− AGM decreased Gpr65 expression, and Gpr65 expression was undetectable in the yolk sac (Figure 4.4j). Previously, we demonstrated that GATA-2 expression in mouse aortic endothelial cells increases transcription of certain GATA-2 target genes [303]. In this system, GATA-

2 increased Gpr65 expression (Figure 4.4k), indicating that GATA-2 regulates Gpr65 expression in multiple

93 contexts. To analyze the Gpr65 expression pattern in distinct cell types within the AGM, we mined RNA- seq data obtained with FACS-sorted endothelial cells, hemogenic endothelial cells, hematopoietic cells,

94 and HSCs from the AGM [296]. This analysis revealed that Gpr65 is detectable in all cell types, and the levels are lower in HSCs (Figure 4.4l).

Figure 4.5 Expression profiles from normal mouse tissues, organs, and cell lines (biogps.org).

To test whether GPR65 controls hematopoiesis in the AGM, we conducted a loss-of-function analysis using a Gpr65 short hairpin RNA (shRNA) retrovirus. E11.5 AGMs were infected, and after culturing for 96 hr hematopoietic cell populations were quantitated by flow cytometry. Quantitation of

GFP+ live cells indicated that control shRNA (shLuc) and shGpr65 retroviruses had an indistinguishable infection efficiency (Figure 4.6a,b). Gpr65 knockdown reduced Gpr65 mRNA by 60%–70% (Figure 4.6c).

While downregulating Gpr65 did not alter CD31+c-KIT+hematopoietic cells (Figure 4.6a,d), it increased

CD31+c-KIT+SCA1+ HSC-containing, multipotent hematopoietic cells (Figure 4.6a,e). Retroviral- mediated Gpr65 shRNA expression did not alter CD31+c-KIT+ hematopoietic cells and CD31+c-

KIT+SCA1+ HSC-containing cells in GFP− cells (Figure 4.7). These results indicate that GPR65 suppresses hematopoiesis in the AGM.

To assess whether Gpr65 activity to suppress hematopoiesis operates in other systems, we used a morpholino (MO) targeting the Gpr65 translation start site (Gpr65_ATG MO) to reduce Gpr65 expression in zebrafish embryos. The Gpr65_ATG MO was injected into 1-cell stage embryos, which were analyzed for expression of the HSPC markers Runx1/cmyb using in situ hybridization (ISH) at 36 hr post-fertilization.

The Gpr65_ATG MO dose-dependently increased expression of the HSPC markers Runx1/cmyb in the embryos (Figure 4.6f,g). A second MO, which blocks Gpr65 splicing (Gpr65_SP MO), yielded an identical result; Gpr65 downregulation induced expression of the HSPC markers Runx1/cmyb (Figure 4.6g). Thus,

GPR65 suppresses hematopoiesis in mouse and zebrafish embryos.

shRNA and MO-based loss-of-function strategies may be confounded by off-target effects. As an alternative strategy we used a GPR65 antagonist, the lysosphingolipid galactosylsphingosine (psychosine).

Psychosine was initially proposed to be a GPR65 agonist, based on the GPR65 requirement for psychosine-induced multi-nuclear cell formation [304]. Subsequently, it was demonstrated that GPR65 is a proton-sensing receptor and that psychosine antagonizes GPR65 [183, 198, 305]. We treated E11.5 AGMs with vehicle or psychosine, and after 96 hr, hematopoietic cells were quantitated by flow cytometry. While psychosine did not alter CD31+c-KIT+ cells (Figure 4.8a,b), it increased the CD31+c-KIT+SCA1+ cell population, which is known to contain multipotent hematopoietic precursors (Figure 4.8a,c). In aggregate, the mouse and zebrafish studies indicate that GPR65 is an endogenous suppressor of hematopoiesis.

Figure 4.6 GPR65 Suppresses Hematopoiesis in the Mouse and Zebrafish AGM (A) Representative flow cytometric plots of CD31+c-KIT+ and CD31+c-KIT+SCA1+ cell populations in control or Gpr65 shRNA- treated AGMs after 96 hr of culture. (B) Quantitation of GFP+ cells with total live cells (9 litters: shLuc [n = 22 embryos]; shGpr65 [n = 26 embryos]). (C) RT-PCR analysis of Gpr65 mRNA levels in FACS-sorted GFP+ cells (6 litters: shLuc [n = 15 embryos]; shGpr65 [n = 15 embryos]). (D and E) Analysis of flow cytometry data expressed as percentage of CD31+c-KIT+ (D) and CD31+c-KIT+Scal1+ (E) cells in GFP+ cells (D: 9 litters: shLuc [n = 22 embryos]; shGpr65 [n = 26 embryos]; E: 7 litters: shLuc [n = 18]; shGpr65 [n = 22]). (F) Representative images of ISH with the HSPC markers Runx1/cMyb at 36 hr post- fertilization. (G) Quantitation of ISH data expressed as percentage of embryos with high, medium, and low Runx1/cMyb staining in total embryos (ATG MO 0 ng [124 embryos]; ATG MO 4 ng [75 embryos]; ATG MO 6 ng [66 embryos]; SP MO 0 ng [97 embryos]; SP MO 4 ng [49 embryos]; SP MO 6 ng [58 embryos]). Gpr65_ATG MO: morpholino targeting the translation start site of Gpr65; Gpr65_SP MO: morpholino blocking the splicing of Gpr65. Error bars represent SEM. ∗p < 0.05; ∗∗∗p < 0.001 (two-tailed unpaired Student's t test).

Figure 4.7 Retroviral-mediated shGPR65 expression does not alter the CD31+c-KIT+ cell populations in the GFP- cell population. A) Representative plots from flow cytometric analysis of CD31+c-KIT+ and CD31+c-KIT+SCA1+ cell populations in GFP- cells after 96 h of culture. B) Quantitative analysis of flow Cytometry data expressed as the percentage of CD31+c-KIT+ and CD31+c-KIT+SCA1+ cells in GFP- cells (Top: 9 litters: shLuc [n=22 embryos], shGpr65 [n=26 embryos]; Bottom: 7 litters: shLuc [n=18 embryos]; shGpr65 [n=22 embryos]).

Figure 4.8 Psychosine Promotes Hematopoiesis in the AGM and GPR65 Suppresses Hematopoiesis by Repressing Gata2 Expression. (A) Representative flow cytometric plots of CD31+c-KIT+ and CD31+c- KIT+SCA1+ cell populations in the AGM after 4 days of culture with 20 μM psychosine. (B and C) The average percentage of CD31+c-KIT+ (B) and CD31+c-KIT+SCA1+ (C) cell populations with vehicle or psychosine treatment (6 litters: control [n = 19 embryos]; psychosine [n = 20 embryos]). (D and E) RT-PCR analysis of Gpr65, Gata2, and Runx1 mRNA (D) and Gata2 primary transcript (E) in FACS-sorted CD31+c- KIT− cells (n = 3 independent experiments). Error bars represent SEM. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 (two-tailed unpaired Student's t test).

4.5.4. GPR65 Establishes Repressive Chromatin and Disrupts and Activating Complex on a cis

Element Required for Gata2 Transcription

To elucidate the mechanism underlying GPR65 suppression of hematopoiesis, we considered whether GPR65 might downregulate key regulators of HSC generation/activity. After knocking down

GPR65, we isolated infected CD31+c-KIT−endothelial cells that give rise to HSCs.

Downregulating Gpr65 mRNA by 60%–70% increased Gata2 mRNA 2.9-fold (p = 0.03) and its downstream target Runx1 mRNA 2.9-fold (p = 0.04) (Figure 4.8d). The knockdown elevated Gata2 primary transcripts

3.9-fold (p = 0.04) (Figure 4.8e), indicating that GPR65 suppresses Gata2 transcription.

To determine whether GPR65 regulates Gata2 expression in zebrafish, we analyzed Gata2 expression using ISH at 36 hr post-fertilization. Zebrafish have two Gata2 homologs; Gata2b is enriched in hemogenic endothelium and regulates HSPC emergence [306].

Whereas the majority of uninjected embryos exhibited broad staining in the AGM, there was no clear linear zone enriched in hemogenic endothelium. However, Gpr65_ATG MO-injected embryos exhibited the linear zone (Figure 4.9), suggesting that GPR65 suppresses Gata2 expression in zebrafish embryos. Given that

GATA-2 promotes HSC emergence in the AGM and regulates HSC activity [278, 279], we propose that

GPR65 suppresses hematopoiesis by limiting Gata2 expression and GATA-2 levels.

As Gpr65 is expressed in AGM endothelium and HSPCs, we asked whether GPR65 represses Gata2 expression in other contexts. Fetal liver hematopoietic precursors express Gata2, and as

GATA-1 rises upon erythroid maturation, Gata2 is repressed [307]. We isolated Lin− hematopoietic precursors from E14.5 fetal livers [307] and tested whether reducing Gpr65 expression with the Gpr65 shRNA retrovirus alters Gata2 expression. Cells were expanded for 72 hr to increase HSPCs, while suppressing differentiation. Downregulating Gpr65 increased the proerythroblast-enriched R2 population 1.9-fold (p = 0.0002) and reduced the R3 population (early and late basophilic erythroblasts)

1.6-fold (p = 0.0003) (Figure 4.10a). Increased R2 cells, concomitant with reduced R3 cells, suggests that

GPR65 promotes erythroid maturation. Downregulating Gpr65 mRNA by 70%–80%, which lowered GPR65 protein by 50%, increased Gata2 mRNA and primary transcripts 2.4 (p = 3.65 × 10−6) and 2.8-fold (p =

0.014), respectively (Figure 4.10b,c). Western blot analysis of fetal liver cells revealed that reducing GPR65 expression upregulated GATA-2 (Figure 4.10d). Increased GATA-2 selectively elevated GATA-2 target

100 gene expression (Figure 4.10e). To test whether increased Gata2 expression reflected a change in cellularity, we used FACS to isolate the GATA-2-expressing R2 population and compared gene expression in control and knockdown R2 cells. Gpr65 knockdown upregulated Gata2 mRNA 2.3-fold (p = 0.02), primary transcript 4.8-fold (p = 0.01), GATA-2 protein, and GATA-2 target genes in FACS-sorted R2 cells (Figure

4.10f-i). These results indicate that GPR65 represses Gata2 in the AGM and fetal liver. Since GATA-1 needs to repress Gata2 transcription early in erythroid maturation, upregulated Gata2 expression caused by the Gpr65 knockdown would be expected to increase immature R2 cells as observed (Figure 4.10a).

Figure 4.9 GPR65 suppresses Gata2 expression in the zebrafish embryo. A) Representative images of in situ hybridization for Gata2 at 36 hours post fertilization. The red rectangle illustrates the region enriched in hemogenic endothelium. B) Quantitative analysis of in situ hybridization data expressed as a percentage of embryos with high, medium, and low Gata2 staining in total embryos (ATG_MO 0ng [n=104 embryos]; ATG_MO 4nf [n=22 embryos]; ATG_MO 6ng [n=40 embryos]) Error bars represent SEM. *, P< 0.05 (two- tailed unpaired Student’s t-test).

101

Figure 4.10 GPR65 Enhances H4K20me1 and Limits Scl/TAL1 Occupancy at the +9.5 Enhancer. (A) Representative flow cytometric plots of erythroid maturation based on CD71 and Ter119 expression 3 days after HSPC expansion. The average percentage of cells in R1 through R5 populations after infection with control or Gpr65 shRNA is depicted on the right (n = 3 independent experiments). (B) RT-PCR analysis of Gpr65 mRNA in fetal liver cells (n = 5 independent experiments) (left). Western blot of GPR65 in fetal liver cells (top right), and quantification of GPR65 and α-tubulin intensities by densitometry (n = 8 independent experiments) (bottom right). (C) RT-PCR analysis of Gata2 mRNA and primary transcript in total fetal liver cells (n = 5 independent experiments). (D) Western blot of GATA-2 in fetal liver cells (top) and quantification of GATA-2/α-tubulin ratio by densitometry (n = 6 independent experiments) (bottom). (E) RT-PCR analysis of GATA-2 target gene expression in fetal liver cells (n = 5 independent experiments). (F–I) Real-time RT-PCR analysis of Gpr65 mRNA (F), Gata2 mRNA (G), Gata2 primary transcript (G), and GATA-2 target genes in FACS-sorted R2 cells (n = 3 independent experiments) (I). Western blot of GATA- 2 in FACS-sorted R2 cells and quantification of intensities of GATA-2 to α-tubulin band by densitometric analysis (n = 4 independent experiments) (H). (J) Fetal liver HSPCs were isolated from E14.5 embryos heterozygous for the +9.5 site at Gata2. Cells were infected with control or shGpr65 and expanded for 3 days (left). Allele-specific real-time RT-PCR analysis of Gata2 transcripts from the wild-type and mutant +9.5 alleles in total fetal liver cells (n = 3 independent experiments) (middle) and FACS-sorted R2 cells (n = 3 independent experiments) (right). (K) RT-PCR analysis of Setd8 mRNA in fetal liver cells treated with control or Gpr65 shRNA (n = 5 independent experiments) (top left). H4K20me1, GATA-1, and Scl/TAL1 chromatin occupancy measured by quantitative ChIP at the +9.5 site in expanding fetal liver HSPCs treated with control or Gpr65 shRNA (H4K20me1: n = 8 independent experiments; GATA-1: n = 4 independent experiments; Scl/TAL1: n = 6 independent experiments) (right). H4K20me1 levels at the repressed MyoD promoter and the active Eif3k promoter (middle left). Scl/TAL1 chromatin occupancy 114 kb upstream from the c-Kit promoter and Lyl1 exon1 as controls (bottom left). The dashed line illustrates the highest value obtained with PI antibody. (L) GATA-2, a positive regulator of hematopoiesis, upregulates Gpr65, which encodes a negative regulator of hematopoiesis, to control HSC emergence. GPR65 represses Gata2 expression by increasing H4K20me1, which restricts +9.5 occupancy by the activator Scl/TAL1. Error bars represent SEM. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 (two-tailed unpaired Student's t test).

102

Figure 4.10 Continued

103

The +9.5 enhances Gata2 expression in the AGM and fetal liver [279, 280, 308]. Deleting the +9.5 abrogated HSC generation in the AGM [279] and disrupted establishment of the HSPC compartment in the fetal liver [280]. To determine whether increased Gata2 transcription upon Gpr65 knockdown requires the +9.5, we isolated Lin− HSPCs from E14.5 + 9.5+/− fetal livers and infected cells with Gpr65 shRNA retrovirus. After 72 hr of expansion culture, allele-specific primers were used to quantitate primary transcripts from wild-type and mutant 9.5 alleles in fetal liver cells and FACS-purified R2 cells (Figure

4.10j). Gpr65 knockdown upregulated Gata2 primary transcripts from the wild-type, but not the +9.5 mutant, allele (Figure 4.10j), demonstrating importance of the +9.5 for Gata2 transcription.

Previously, we demonstrated that SetD8, the enzyme that monomethylates H4K20 [309], represses Gata2 expression via the +9.5 site [310]. We tested the relevance of this mechanism to GPR65- mediated Gata2 repression. We conducted quantitative ChIP analysis for H4K20me1 in fetal liver cells

(control or Gpr65 knockdown) after culturing for 72 hr. Downregulating Gpr65 did not alter the SetD8 mRNA level (Figure 4.10k). Gpr65 downregulation reduced H4K20me1 at the +9.5 site (p = 0.002) and at sites

480 bp upstream (p = 0.002), 466 bp downstream (p = 0.006), and 880 bp downstream (p = 0.02) of the +9.5 site (Figure 4.10k). H4K20me1 was unaltered at the repressed muscle-specific MyoD promoter and the constitutively active Eif3k promoter. These results support a mechanism in which GPR65 represses Gata2 by increasing a repressive chromatin mark at the +9.5 site. The basic helix-loop-helix transcription factor Scl/TAL1 activates Gata2 transcription through the +9.5 site, and Scl1/TAL1 chromatin occupancy is reduced at the +9.5 site during GATA-1-mediated Gata2 repression [308].

As Gpr65 downregulation decreased H4K20me1 at the +9.5 site, this alteration may generate chromatin that is more accessible to cognate binding factors. To determine whether GPR65 alters GATA-1 and

Scl/TAL1 occupancy, we quantitated GATA-1 and Scl/TAL1 occupancy at the +9.5 site in fetal liver cells infected with control shRNA or Gpr65 shRNA-expressing retrovirus. While knocking down Gpr65 did not alter GATA-1 occupancy at the +9.5, the knockdown increased Scl/TAL1 occupancy at the +9.5 1.6-fold

(p = 0.02) but not at other target loci (Lyl1 and Kit) (Figure 4.10k and Figure 4.11). These results suggest that GPR65 represses Gata2 by establishing repressive chromatin that limits Scl/TAL1 occupancy at the +9.5 site (Figure 4.10l).

104

Figure 4.11 ChIP-seq analysis of GATA-2 and TAL1 occupancy at Lyl1 (A) and KIT (B) loci in HPC7 cells. The red box depicts the amplified sequence from the real-time RT-PCR analysis of Figure 4.10k.

4.6. Discussion

Genetic networks orchestrating stem and progenitor cell transitions can be deconstructed into regulatory modules termed network motifs [311, 312]. Whereas major progress has been made to identify individual genes controlling hematopoiesis [279, 296, 313-316], many questions remain unanswered regarding how the genes form network motifs, how the network motifs amalgamate into circuitry, and whether the circuitry has an inherent plasticity and undergoes remodeling in states of altered hematopoiesis such as stress, aging, and malignancy.

Since the GATA-2-dependent genetic network promotes hematopoiesis [279, 296], it is reasonable to infer that GATA-2-induced factors are positive mediators of key steps, including EHT, HSPC self-renewal, and differentiation. In erythroid cells, GATA-1 activates heme biosynthetic genes, globin subunits, and constituents of the red cell cytoskeleton [271, 276, 317, 318], all required for erythroid maturation. Moreover,

GATA-1 represses Gata2, Lyl1, and Kit [319-321], to enable erythroid maturation. GATA-2 activates Kit [322], an essential regulator of HSPCs [323, 324], consistent with GATA-2 promoting HSPC genesis and function.

105

Using a coupled bioinformatics-experimental strategy, we analyzed the large GPCR gene family to discover GATA-2-induced GPCRs that promote HSPC transitions. This analysis led to the surprising finding that GATA-2 activates Gpr65 expression, which suppresses hematopoiesis via negative feedback on Gata2. This negative feedback mechanism conforms to a type I incoherent feedforward loop [325], which is defined as an input (GATA-2) that elicits an output (EHT or hematopoiesis) through positive

(increased Gpr65 expression) and negative (GPR65 suppression of hematopoiesis) paths. Type I incoherent feedforward loops shape the dynamics of a mechanism, with the feedforward loop conferring a pulse of activity to accelerate the reaction [325]. Since the constituents and network motifs governing HSPC transitions are still being identified, the dynamics of steps in these transitions are largely unexplored. The mechanism of GPR65-mediated Gata2 repression involved elevation of a repressive histone mark,

H4K20me1, linked to reduced occupancy by the activator Scl/TAL1. GPR65 might directly target chromatin to restrict Scl/TAL1 occupancy or might reduce Scl/TAL1 occupancy prior to chromatin modification.

Regardless of the order of events, our analysis established a GATA factor-GPCR-dependent type I incoherent feedforward loop that suppresses, rather than promotes hematopoiesis.

Gpr65, or T cell death-associated gene 8 (Tdag8), is a proton/acid-activated GPCR [198, 326].

Whereas GPR65 had not been linked to HSPC transitions, it was reported to be a pro-apoptotic factor in glucocorticoid-induced lymphocyte apoptosis [327] and to regulate cytokine production from macrophages

[328]. Targeted deletion of Gpr65 revealed it was dispensable for glucocorticoid-induced thymocyte apoptosis [329]. These results were suggested to reflect redundancy with related pH-sensing GPCRs, which would create an obstacle to dissect Gpr65 function in vivo. Although the impact of the pH-sensing mechanism to GATA-2 function in the AGM and fetal liver is unclear, the endogenous GPR65 antagonist psychosine resembled Gpr65 shRNA in upregulating Gata2 expression and hematopoiesis.

As GPR65 suppressed HSPC generation from mouse AGM explants, it is instructive to consider the consequences of ablating this mechanism in vivo. Abrogation of the mechanism would promote HSPC genesis and function. Since ablating Gpr65 would be expected to increase HSPCs, evaluating the role of the GATA-2-GPR65 circuit in vivo will require a careful quantitative analysis. Considerations of type I incoherent feedforward loops [311, 325] would infer an alteration of the dynamics of specific steps of hematopoiesis, which might not be evident from steady-state analyses. It is attractive to propose that this

106 suppressive mechanism is vital when the demand for hematopoiesis increases, e.g. during stress, since an unopposed increase could yield deleterious blood cell elevations.

In summary, we discovered that GATA-2 activity to promote hematopoiesis involves induction of both positive and negative mediators, the balance of which establishes the physiological output. Based on the GPR65 pH-sensing function, it will be crucial to analyze the pH microenvironment of the AGM and other sites of hematopoiesis and to establish the role of GPR65 as a mediator of pH-dependent HSPC transitions.

4.7. Experimental Procedures

4.7.1. Mice and Embryo Generation

Gata2 +9.5-mutant mice were described by Johnson et al. (2012, [280]). Pregnant females were euthanized with CO2, and freshly isolated embryos were transferred into cold PBS for dissection. Animal experiments were carried out with ethical approval of the Association for the Assessment and Accreditation of Laboratory Animal Care at the University of Wisconsin-Madison.

4.7.2. AGM Explant Culture

AGMs were dissected from E11.5 embryos. AGMs were infected (see below) and cultured as described by Gao et al. (2013 [279]). Intact AGMs were cultured for 4 days on Durapore filters (Millipore) at the air-liquid interface in IMDM+ (Iscove's modified Dulbecco's medium) (Gibco) supplemented with 20% fetal bovine serum (FBS; Gemini), 4 mM L-glutamine (Gibco), 1% penicillin/streptomycin (Cellgro), 0.1 mM mercaptoethanol, 100 ng/ml interleukin-3 (R&D Systems), 100 ng/ml Flt3L (R&D), and 1.5% conditioned medium from a Kit ligand-producing Chinese hamster ovary cell line.

4.7.3. Retroviral Infection

Detailed methodology is presented in Appendix 3. Gpr65 shRNA were cloned into the MSCV-PIG

(IRES-GFP) vector (kindly provided by Dr. Mitchell Weiss) [307] using BglII and XhoI restriction sites.

Retroviral expression vectors for Gata2 and shLuc were described by Katsumura et al (2014, [303]).

Retroviruses were packaged by co-transfecting 293T cells with pCL-Eco packaging vector. Retroviral supernatant was collected 24 and 48 hr post transfection and centrifuged to remove cells and debris. The

Retro-X Concentrator (Clontech #631455) was used to concentrate retrovirus. E11.5 AGMs were infected

107 with 20 μl concentrated retrovirus and 8 μg/ml polybrene in 500 μl explant culture media at 1,200 × g for

90 min at 30°C. AGMs were subjected to explant culture. Primary erythroid precursors were spinfected with

100 μl retrovirus supernatant and 8 μg/ml polybrene in 400 μl fetal liver expansion medium at 1,200 × g for

90 min at 30°C. After centrifugation, 500 μl pre-warmed fetal liver expansion medium was added, and cells were incubated at 37°C for 72 hr. Details for concentrating retrovirus and Gpr65 shRNA sequences are available in Appendix 3.

4.7.4. Statistical Analysis

Student's t tests were conducted using GraphPad Software or Microsoft Excel. p Values in all figures are denoted by asterisks: ∗p < 0.05, ∗∗p < .01, and ∗∗∗p < .001.

Supplemental experimental procedures can be found in Appendix 3.

4.7.5. Author Contributions and Acknowledgments

All authors participated in designing experiments and/or analyzing data. J.L.L. conducted the zebrafish experiments and analyzed the data. X.G. and E.H.B. wrote the manuscript in consultation with other authors. The work was supported by NIH DK68634 and DK50107 (to E.H.B.), Cancer Center Support

Grant P30CA014520 , and NIH HL04880 , HL032262 , HL10001-05 , and DK092760 (to L.I.Z). We thank

Michelle Ammerman and Prathibha Sanalkumar for assistance with zebrafish experiments and Western blotting, respectively.

108

Chapter 5. Discussion

Hematopoietic stem cell transplant is a lifesaving treatment for thousands of patients yearly, but limiting cell numbers remain a challenge, especially in the case of cord blood transplants. Fatty acids are important regulators of cell behavior in many contexts, but more study is needed about their roles in hematopoiesis and hematopoietic stem cell transplant. In particular, knowledge of their downstream signaling pathways and receptors will identify targets for therapeutic or experimental manipulation.

A chemical screen in adult zebrafish identified the endogenous lipid signaling molecule 11,12-EET as an enhancer of HSC transplant. EET also enhances specification of HSPCs during zebrafish development. Using chemical and genetic studies, we found that these phenotypes are mediated by

Gα12/13, PI3Kγ, and AP-1 signaling. Comparative RNAseq in EET-binding human cell lines identified candidate EET receptors. Only one of these candidates, GPR132, was activated by EET in vitro, and this gene was required in zebrafish embryos for EET to enhance hematopoietic signaling. Marrow from GPR132 knockout mice shows a defect in long-term transplant at limiting dilution, suggesting that these mice have decreased HSC numbers or decreased HSC functionality. Structure-activity-relationship analyses revealed that medium chain, polyunsaturated, oxygenated free fatty acids consistently activated GPR132 in vitro, and these molecules also cause EET-like phenotypes in zebrafish embryos. GPR132 is likely a receptor for EETs and other oxygenated fatty acids that can regulate hematopoietic phenotypes. GPR132 is a member of the OGR1 family of GPCRs, which also includes GPR65. GPR65 is upregulated by GATA2 during the specification of HSCs in mouse AGM. Knockdown studies in zebrafish embryos showed that

GPR65 is a negative regulator of GATA2 and hematopoiesis- loss of GPR65 caused increased expression of GATA2 as well as the hematopoietic markers runx1 and c-myb in the zebrafish AGM. GATA2 and GPR65 thus form a negative feedback loop to regulate HSC specification in the early embryo, a role that is conserved in mammals. Thus, two OGR1 family members play opposite roles in regulating HSPC specification.

5.1. Mechanisms of EET’s enhancement of hematopoiesis

11,12-EET treatment enhances the expression of hematopoietic markers during development and the transplant of HSCs during adulthood. Time-lapse imaging using the Runx1+23:GFP reporter line, which marks HSPCs in zebrafish, confirmed the increase of HSPCs in the AGM, and showed that it was due to

110 increased specification, without affecting cell division or AGM residence time. Our studies have shown that

EET induces HSC specification by activating the receptor GPR132, which results in Gα12/13, PI3Kγ, and

AP-1 signaling, increasing the transcription of the hematopoietic transcription factor runx1 autonomously in hemogenic endothelial cells (Figure 5.1). Our data indicates that this GPR132-PI3K-AP-1 signaling axis which we characterized in developmental hematopoiesis also functions downstream of EET in HSPC transplant. EET enhances short-term homing of HSPCs to the marrow (Figure 2.12), so its enhancement of engraftment likely results from improving the migratory abilities of HSPCs. Our working hypothesis is that

EET signaling via GPR132 enhances HSPC migration by activating Gα12/13, PI3K, and AP-1 signaling autonomously in HSPCs, although non-autonomous effects are also possible. I will discuss our evidence for each of these molecular mechanisms in developmental and adult hematopoiesis, as well as potential links between these signaling pathways and HSPC migration.

5.1.1. Identifying a candidate EET receptor

EET was first described as a biologically active signaling molecule causing vasodilation nearly three decades ago, and in 1997 one group found that 14,15-EET has a specific high affinity binding site on U937 cells [244]. However, the identity of an EET receptor to mediate these phenotypes has remained unknown.

Bill Campbell’s lab and others showed that EET likely activates a G-protein coupled receptor [73]. In an attempt to identify this receptor, they developed 14,15-EET-APSA-I25, a photo-crosslinkable, radiolabelled

EET analog. Binding with this molecule showed a single interacting protein of 47kDa in size, and this band was consistent across many cell types, but missing from HEK293 cells. However, overexpression of 90 candidate GPCRs, including GPR132, in HEK293 failed to restore the band [113]. This relatively high throughput overexpression assay might be prone to false negative results, especially because receptors were FLAG-tagged N-terminally. This allowed for easy validation of expression by FACS, but could interfere with ligand binding, as GPCR ligands are known to interact with the extracellular N-terminus of the protein.

Others have also used overexpression analyses. The Alkayed lab overexpressed 105 candidates in

Xenopus oocytes and assayed for calcium flux. They saw no high affinity responses to EET, but found that

EET interacted with prostaglandin receptors in a low affinity manner [114]. Another group overexpressed

FFAR1 (GPR40) in HEK293 cells and saw responsiveness to EET, again only with low affinity [105].

111

Figure 5.1 Mechanisms of oxygenated fatty acid signaling in hematopoiesis. The investigated steps of the EET signaling pathway in enhancing embryonic hematopoiesis and adult marrow transplant are shown. The systems used to validate each signaling factor are also shown. A link between Gα12/13 and PI3K has not yet been shown, but may rely on Rho signaling. The mechanistic relationship of each of these pathways to adult marrow transplant phenotypes also needs to be investigated.

112

The difficulty of isolating an EET receptor biochemically by pull-down methods stems from the difficulty of tagging or biotinylating a small molecule like EET without interfering with its binding. Difficulties also arise due to the likely transient and weak interactions between EET and its receptor. Additionally,

GPCRs are known to robustly signal even when they are present at low RNA or protein levels within a cell.

So, even a successful pull-down might not yield enough material to identify the receptor. The fact that diverse overexpression screens have failed to identify an EET receptor may imply something biological about the interaction of EET and its receptor. Screens of a similar nature have successfully identified high affinity receptors for similar fatty acids including resolvins [330], and 20-HETE [247]. Perhaps a high affinity interaction between EET and its receptor requires homo or heterodimerization of the receptor, or the interaction of cofactors such as receptor activity modifying proteins (RAMPS), which are required for the signaling of several B-type GPCRs [331]. In this case, the overexpression of individual GPCRs will not yield signaling.

Our approach for defining a candidate list avoids many of these limitations by asking what GPCRs are present in cells that bind EET and respond to signaling. Bill Campbell found that U937 cells, EaHy cells, and PC3M-LN4 cells all show binding to 14,15-EET-APSA-I25. These cell types also show physiological responses to EET treatment. U937 cells increase their adhesion to endothelial cells in vitro [332], and

PC3M-LN4 cells explanted in limiting cell numbers to immune deficient mice show tumor dormancy escape when the mice are treated with EET [92]. In developing a candidate list based on our RNAseq data, we needed to choose a threshold cut-off above which GPCRs would be considered as “present” within these cells. While other groups doing screening of this type have used FPKMs of 1 or 5 as presence cut-offs, we wanted to be extremely conservative in order to not miss a candidate. It was also known that GPCR expression can be quite low, especially at an RNA level, so it was important in our case to choose a low cut-off [333]. Hundreds of GPCRs were expressed in each cell type at 0 or 0.1 FPKMs, and this could represent a very low raw number of reads, so this seemed too low to be a reasonable cut-off. We chose

0.3FPKM, in part because there was a drop-off in the number of GPCRs expressed at this amount. Similarly, when looking at GPCRs missing from HEK293 cells, to be conservative we drew a maximum FPKM cut-off of 1.

113

Once a list of receptor candidates is in hand, the next step is to test each one in an in vitro or in vivo assay. Possible in vivo assays would include knockdown of GPCR expression in zebrafish embryos by morpholino or knockout by CRISPR/cas9. Homozygous knockout with CRISPR/Cas9 requires validation of guide RNAs followed by multiple generations of breeding, which can in total take several months.

Morpholino knockdown requires titration of the correct dose, as well as validation to ensure that knockdown is successful and that any phenotypes occur from on target effects. In vitro methods can be categorized into loss of function methods by knockdown or knockout in cell lines, or gain of function methods by overexpression of receptor candidates in cell lines. For knockdown, we would begin with cells that express the receptor and have known phenotypic responses to EET. The literature on EET phenotypes focuses primarily on in vivo phenotypes, or complex in vitro phenotypes requiring multiple cell types interacting, such as U937 cells adhering to endothelial cells [332]. Loss of function approaches also have similar caveats as in vivo methods requiring validation of siRNAs for knockdown or gRNAs for knockout. For overexpression, we would add candidates to a cell line where they are not expressed (ie HEK293 cells) and then look for a new phenotype with EET. What phenotype to look for is a question, as EET produces different readouts in different cell types and for any given readout we would need to ensure that HEK293 cells have all of the necessary downstream signaling components.

PathHunter β-arrestin assays avoid many of these issues, as they do not rely on downstream signaling for a readout. Additionally, β-galactosidase signaling can only come from the tagged, overexpressed receptor, so we can be confident that any luciferase signaling is coming from a true interaction of EET with the receptor and not from other proteins expressed in the cells. These assays are easy, fast, and quantitative. However, PathHunter assays are not available for all GPCRs, indeed 3 of our

10 candidates lacked PathHunter assays. Additionally, for orphan GPCRs without a known ligand, this assay would lack a positive control, leaving a vulnerability for false negative results. In our case, of the 7 candidates we tested, only GPR132 responded to EET treatment. This response was specific, as the related eicosanoid prostaglandin did not activate GPR132.

5.1.2. GPR132 as a mediator of zebrafish HSPC specification

In a process similar to mammalian development, zebrafish HSPCs are first specified from hemogenic endothelium in the aorta-gonad-mesonephros by expression of the master transcription factor

114 runx1. 11,12-EET treatment enhances this expression and causes ectopic expression in the tail mesoderm.

Zebrafish have two GPR132 homologs, gpr132a and gpr132b, but only gpr132b shows expression in zebrafish embryos [182]. We injected a morpholino to knockdown expression of gpr132b and found that this prevented EET’s enhancement of runx1/c-myb staining in the AGM and the tail, indicating that gpr132b is required for EET phenotypes in the zebrafish embryo. To ensure that this phenotype is due to loss of gpr132b and is not an off-target effect, we are performing RNA rescue experiments using a human

GPR132, and we are quantifying the amount of gpr132b knockdown by qPCR. Injections of human GPR132 proved to be toxic at doses greater than 75pg/embryo, although this was somewhat variable across clutches of embryos (data not shown). A first attempt at co-injecting the gpr132b morpholino with GPR132 RNA showed promising results, but needs to be repeated with larger numbers of embryos (data not shown).

CRISPR knockout zebrafish could be used to verify the morpholino result, and to conduct transplant experiments in the adult. We received 2 gRNA sequences from the Wen lab in Beijing targeting zebrafish gpr132b, which is expressed in early embryonic development. We also designed three gRNAs targeting gpr132a, which is not expressed in early development. However, gpr132a could be relevant in adult transplant assays, and upon genetic loss of gpr132b we could see compensation and upregulation of gpr132a [334]. Both gpr132b gRNAs and 2 of 3 gpr132a gRNAs target successfully according to T7E1 assays in zebrafish embryos. Injected together, the gpr132b gRNAs have the potential to generate a deletion of around 573bp in size, out of a total of 1123 bp of coding sequence. This would knockout 4 of the 7 transmembrane domains. Together, the two gpr132a guides have the potential to generate a deletion of around 210bp, out of a total of 969 bp of coding sequence. This would knockout the first two transmembrane domains. Injection of these gRNAs in any combination failed to block EET signaling in F0 embryos (data not shown), so we are breeding these fish to create homozygous lines. We intend to create lines with small insertions or deletions from the injection of single guides, as well as lines with large deletions. We have co-injected guides for both genes and intend to grow up both single and double mutant fish. Additionally, a zebrafish strain with an early nonsense mutation in gpr132a is available from Sanger, we have incrossed heterozygotes of this line and are growing up those offspring.

GPR132 is reported in in vitro studies to be a receptor for lipids such as 9-HODE and 11-HETE

[191, 200, 201, 203]. Like EET, these molecules enhanced runx1/c-myb expression in the zebrafish AGM

115 and tail. We performed extensive in vitro and in vivo assays on a variety of fatty acids and showed that overall, molecules which activated GPR132 in vitro also caused increased AGM and tail staining of hematopoietic markers in zebrafish embryos in vivo. These strong activators were always oxygenated, unsaturated free fatty acids such as 9,10-EpOME and 11,12-DHET. Unoxygenated fatty acids showed little to no activity.

However, the correlation between the in vivo and in vitro assays is not perfect. For instance, the unoxygenated free fatty acids arachidonic acid, linoleic acid, EPA, and DHA all showed weak activation of

GPR132 in vitro, but showed either no (arachidonic acid and DHA) or very poor (linoleic acid and EPA) activity in vivo. Possibly these have low affinity for the receptor and would be required at very high, toxic concentrations to see activity in the fish. Among the strong activators, activity in vitro did not perfectly match activity in vivo. For instance, 9-HODE is by far the strongest in vitro activator, with the lowest Kd and highest maximal activation, but higher concentrations of 9-HODE are required in vivo to give phenotypes equivalent to 5μM 11,12-EET. Both the in vitro and in vivo assays have caveats to their exact measurement of affinity and activity. In the zebrafish, the bioavailability of these molecules at the AGM or tail mesoderm is unknown.

In the PathHunter assays, serum within the media is known to bind lipids and may restrict their availabilility at the cell surface. For instance, our PTGER2 β-arrestin assay suggests that prostaglandin has a Kd of around 100nM at this receptor, while radioligand binding assays have shown the true affinity to be around

12nM [335]. In some cases the tail staining phenotype also did not perfectly correlate with the AGM staining.

An “EET-style” pattern of staining would be an approximate doubling of the percentage of embryos with high AGM staining compared to DMSO treatment, as well as greater than half of the embryos showing tail staining. EET created this effect at 5μM, 11,12-DHET gives a similar phenotype at 10μM, and 9,10 EpOME at 20μM. However, 11-HETE, despite greatly increasing AGM staining at 10μM, resulted in relatively few embryos with tail staining. 20-HETE on the other hand caused tail staining with a potency similar to that of

EET, without increasing AGM staining. These could result from differences of penetration to these two sites of the embryo. The AGM lies deeper within the embryo while the tail mesoderm is more exposed.

Alternatively, runx1 expression in the two locations could rely on slightly different downstream signaling, or perhaps co-receptors or receptor modifying enzymes. It’s possible that different fatty acids could differentially activate these receptor conformations or downstream pathways. It will be important to confirm

116 that the other fatty acids require GPR132 for their embryonic phenotypes, and to test for their ability to enhance HSC transplant in adult zebrafish or mice.

GPR132 and its related family members have been described as receptors for protons, LPC and

SPC, and in the case of GPR132 for fatty acids. The zebrafish embryo assays and PathHunter assays used here provide evidence only for responsiveness of GPR132 to fatty acids. LPC, SPC, and a variety of organic acids were ineffective. However, 11,12-EET methyl ester was also unable to activate GPR132, suggesting that the acid moiety is indeed important for binding. Intriguingly, LPCs come in variable forms featuring diverse fatty acid chains, although they do not have a free acid. I tested only one form of LPC, where the fatty acid chain featured one double bond and no oxygenated group. It’s possible that polyunsaturated, oxygenated LPCs would be able to activate GPR132 in our assays. Other explanations for the difference between our results and what other groups have seen with protons and LPC would include a differential presence of GPR132 homo or heterodimer in the contexts studied, or the presence or absence of other

GPCR-modifying proteins such as RAMPs. In particular, one group showed that GPR68 and GPR132 are able to heterodimerize, and that this increases sensitivity to low pH [255].

5.1.3. Role of GPR132 in mouse marrow transplant

GPR132 knockout mice have been studied for their autoimmune and inflammatory phenotypes

[246]. We saw no difference in overall blood lineages or in HSPC number in WT vs GPR132 heterozygous or knockout mice, but these should be studied in greater numbers to rule out the presence of subtle differences between the genotypes. Transplantation of a limiting number of knockout whole marrow cells resulted in decreased engraftment compared to heterozygous marrow. This defect was consistent in all blood lineages and in the marrow, and resulted in a smaller proportion of knockout mice displaying multi- lineage chimerism. This suggests that GPR132 knockout mice have a decreased number of functional LT-

HSCs. This should be verified with single cell transplantation of sorted LT-HSCs. These results raise the question of whether GPR132 functions constitutively in WT mice, or whether EET or other fatty acids activate this receptor throughout the lifetime of the mouse to improve HSC specification or maintenance.

We attempted EET and 9-HODE treatments of GPR132 knockout mouse marrow and that of heterozygous siblings followed by transplant into irradiated recipients. Although we achieved good chimerism, we saw no improvement of engraftment by EET or 9-HODE treatment in either case (Figure

117

3.11). This is in contrast to the enhanced engraftment that we have seen with EET treatments of WT unrelated marrow. This may indicate that full wildtype levels of GPR132 are needed for signaling, and both

GPR132 +/- and -/- marrow are unable to respond to EET. A mechanistic explanation for this could be that

GPR132 functions as a homodimer. In that case, loss of only 50% of the GPR132 protein level could result in a much more drastic loss of functional dimers, explaining a total lack of response to EET. We are eager to perform direct comparisons between EET treatment of wildtype and heterozygous marrow, as well as wildtype and knockout marrow to rule out any possible technical explanation for the lack of response seen in our heterozygous and knockout transplant experiments. For instance, there could be a problem with the background of the GPR132 KO mouse colony which prevents EET-responsiveness in transplant. We are also attempting experiments with more short-term readouts, such as looking for gene expression changes in marrow cells upon fatty acid treatment. If the difference in fatty acid responsiveness between GPR132

+/+ and +/- marrow is confirmed, it will be important to also compare transplantation of untreated GPR132

+/+ and +/- marrow at limiting cell doses, as there may already be a defect in the heterozygous case.

5.1.4. Gα12/13 signaling in EET phenotypes

GPCRs couple to Gα proteins belonging to the Gαs, Gαi, Gαq, or Gα12/13 families. GPR132 has been shown to specifically activate Gα13 signaling [204, 205]. We used morpholino knockdown of Gα12/13 factors to show the requirement of these proteins in EET-enhanced HSPC specification. Interestingly, we saw a separation of the roles of Gα12 and Gα13 in controlling runx1 expression in the AGM and tail mesoderm. In the tail, knockdown of each individually failed to block EET signaling, but a combined knockdown prevented the EET-induced increase in hematopoietic markers. This likely means that either factor is able to mediate EET signaling in the tail. In the AGM, knockdown of Gα12 prevented an EET- induced increase in hematopoietic markers, without decreasing the basal level of runx1/c-myb expression.

Gα13 knockdown instead completely ablated runx1/c-myb expression. Gα13 is likely required early for basal runx1 expression. It may be functioning downstream of GPR132 and endogenous fatty acid ligands, or it may be functioning in tandem with another GPCR. Gα12 instead seems to be specifically required for

EET-induced functions, and unlike in the tail, its loss in the AGM cannot be replaced by endogenous Gα13.

We have not directly interrogated a role for Gα12/13 signaling in EET-enhanced HSPC transplant, but this family is known to regulate cell migration and metastasis. Cell types where this has been shown

118 include endothelial cells [161], B cells [164], zebrafish mesodermal cells [165], and a variety of cancers

[159-162]. Gα12/13 may induce this migration by activating RhoA signaling and thereby affecting cytoskeletal dynamics [171]. Indeed, GPR132 activation of Gα13 is known to induce RhoA signaling [204,

205]. Rho family members may also be the mechanistic link between Gα12/13 and PI3K signaling. The

Rho family members RAC1 and CDC42 demonstrated direct binding to PI3K, with RhoA showing much weaker binding [336]. In another study, multiple Rho family members were shown to work cooperatively to activate PI3K [337].

5.1.5. PI3K signaling in EET phenotypes

PI3K signaling is a major cellular signaling pathway regulating growth and proliferation by activating

Akt. Overactive PI3K or Akt signaling is a hallmark of many cancers. There are 4 major PI3K catalytic subunits, termed PI3Kα, β, γ, and δ. A chemical suppressor screen found that multiple pan-PI3K inhibitors blocked EET’s enhancement of hematopoiesis in zebrafish embryos. Chemical treatments of zebrafish embryos using specific inhibitors found that EET phenotypes were sensitive specifically to PI3Kγ inhibition and not to perturbation of other PI3K catalytic subunits. Morpholino knockdown of the zebrafish PI3Kγ subunit confirmed this phenotype. However, PI3Kα and β both appeared to have important early developmental roles, as these morpholinos caused major morphological defects, precluding genetic analysis of their role in EET-enhanced AGM specification.

Pan-PI3K inhibition similarly blocked EET-enhanced engraftment of mouse bone marrow. Further experiments are needed to determine whether the γ subunit is also involved in transplant. Mechanistically,

PI3K likely enhances HSPC transplant by influencing migration, either through upregulation of AP-1 factors, as I will discuss below, or through other mechanisms. CXCR4 signaling activates PI3K, stimulating the migration of several different cell types [338-340]. However, most experiments examining HSCs have shown that activation of either PI3K or Akt inhibits migration and decreases homing and engraftment efficiency [341, 342]. Akt can mediate signaling through Rho and Rac1 which in turn affect cytoskeletal dynamics [343]. In EET enhanced hematopoiesis, Gα12/13 and PI3K could converge to activate RhoA.

Additionally, the Akt target GSK-3β influences adhesion by modulating integrin expression [344]. Integrin mediated adhesion is essential for the interaction of HSPCs with niche cells as they enter the marrow [48].

119

5.1.6. AP-1 signaling in EET phenotypes

We found that PI3K signaling was required for EET’s upregulation of AP-1 family members in zebrafish embryos. AP-1 proteins are basic leucine zipper factors that bind to TRE sites in the genome to activate their targets. They exist typically as heterodimers between Jun (c-jun, junB, and junD) and Fos (c- fos, fosl1, fosl2) family members [345]. In the AGM, EET treatment increased expression of zebrafish JUNB homologs and induced expression of fosl2, which is not normally present. Neither of these factors are expressed in tail mesoderm, but they were induced by EET treatment. Morpholino knockdown of junb/junbl impaired HSPC specification and blocked any enhancement by EET. This reflects data from Xenopus, where the JunD/c-fos complex induces early hematopoietic markers [346]. Runx1 has a single TRE site in its promoter, so AP-1 factors could directly increase runx1 expression and therefore increase HSPC specification in development.

EET treatment also increased expression of AP-1 genes in human U937 and CD34 cells, suggesting that AP-1 could be involved in EET’s enhancement of adult transplant. However, runx1 is dispensable for adult HSPCs, so AP-1 would require other mechanisms for enhancing transplant. In both normal and cancer cells AP-1 activation leads to increased cell migration [347, 348]. In many cases, this migration is mediated by the upregulation of matrix metalloproteinases, an important target of AP-1 [349-

351], which have recently been implicated in regulating HSC birth and CHT colonization in the zebrafish

[352]. However, no role has yet been identified for AP-1 factors in the migration of hematopoietic cells, and transplanted junb-/- mouse cells home successfully to new niches [353]. In adulthood, AP-1 appears to be a negative regulator of hematopoiesis, as mouse knock-outs of AP-1 factors develop myeloproliferative disorders and leukemia [354]. Similar AP-1 deficiencies have been identified in human leukemic patients.

5.1.7. Cell autonomy of EET signaling

Studies using transgenic zebrafish showed that AP-1 activity is required autonomously in the hemogenic endothelium for EET-enhanced specification. We have not directly assayed the autonomy of

PI3K or GPR132/Gα12/13 signaling in this context. However, blocking translation with cycloheximide failed to prevent AP-1 upregulation, so any cell non-autonomous signaling would have to occur without translation of new proteins. The zebrafish caudal hematopoietic tissue (CHT) presents a tempting model for studying

HSPC engraftment. As in the case of adult transplant, HSPCs migrate to the CHT via the circulation and

120 enter that niche after interacting with endothelial cells. Work in our lab has shown that endothelial cells undergo complex changes to provide a niche within the CHT [36], and that stromal cells and macrophages also show close interaction with HSPCs (Elliott Hagedorn, unpublished data). We are exploring methods to specifically mark CHT endothelial cells, which would allow us to manipulate these cells genetically, and determine if there are CHT-specific expression profiles in these endothelial cells compared to others.

When we treated zebrafish embryos with EET starting at 40hpf, after the major wave of specification in the CHT, we saw increased runx1/c-myb staining in the CHT. This was not seen with the

Runx1+23:GFP line, which marks a more specific population of hematopoietic stem cells, so it may indicate a progenitor, rather than a stem cell phenotype. Nevertheless, these EET-treated progenitors show enhanced engraftment of the CHT, and we found this phenotype to be reliant on AP-1 function specifically in endothelial cells. Blocking AP-1 in HSPCs did not prevent this phenotype. This suggests a cell non- autonomous role in preparing the niche for HSPC arrival, although it does not preclude an additional, cell- autonomous role for EET.

In our mouse and zebrafish transplant experiments, only donor marrow is treated with EET, not the recipient niche. However, diverse cell types are present in whole marrow, including endothelial cells and a variety of blood cells. Based on GPR132 expression patterns in mouse and human hematopoietic cell types,

HSPCs could be responding directly to EET, or B cells, T cells, macrophages or endothelial cells could provide indirect cues. In vitro, we found that two adult human hematopoietic cell types increase expression of AP-1 factors when treated with EET. Autonomous EET-GPR132 signaling in hematopoietic stem cells is therefore an exciting avenue for further investigation.

5.2. Future directions- further study of EET-GPR132 signaling

We have shown that EET and other oxygenated fatty acids can activate GPR132 to induce signaling via Gα12/13, PI3K, and AP-1 autonomously in hemogenic endothelial cells in order to increase the specification of HSPCs during development. We expect the same pathways to be important in EET’s enhancement of HSPC transplant, and hypothesize that these signals also occur autonomously in adult

HSPCs. Further studies on this signaling axis should validate a direct binding interaction between EET and

GPR132, explore RhoA signaling as a possible mechanism by which EET and GPR132 would regulate the

121 migration of HSPCs during transplant, and test the hypothesis of cell autonomous EET-GPR132 signaling in HSPCs.

5.2.1. Direct binding of fatty acids to GPR132

To definitively show that GPR132 is a direct free fatty acid receptor, radioligand binding assays must be performed. In these assays, GPR132 would be overexpressed in cell lines, and these cells would be incubated with radiolabeled fatty acids such as tritiated EET (H3-EET) or H3-9-HODE. Cells would be co-incubated with or without non-radiolabelled fatty acids as competitors; in the competitive condition these should be present at a minimum of 3 log fold excess. Several technical challenges exist for this approach.

GPR132 was initially named G2A for G2 arrest, because its overexpression causes cells to arrest the cell cycle. Accordingly, we found that HEK293 and CHO cells transfected with GPR132 were unhealthy, they had more rounded cell bodies and decreased proliferation. We were unable to establish stably expressing

GPR132 cell lines, but continued binding efforts with transiently transfected cells.

Western blots for GPCRs are known to be challenging, as receptors are stable only in the cell membrane. However, we found that the G-5 monoclonal GPR132 antibody from Santa Cruz successfully bound HEK293 cells transfected with GPR132, and FACS analysis showed that transfections were highly efficient. If EET binding to GPR132 requires cofactors or another GPCR as a heterodimer, individual overexpression of GPR132 will not be permissive for binding. Another option is to use GPR132-β-arrestin cells from DiscoverX. Since EET and other fatty acids activate GPR132 in these cells, any necessary cofactors must also be present.

Tritiated arachidonic acid is commercially available from American Radiolabelled Chemicals, and arachidonic acid showed slight activity in GPR132 β-arrestin assays, although it had no activity in zebrafish embryos. We first attempted to optimize binding using this compound, but we saw no specific binding in several different binding buffers, looking at attached cells or cells in solution. All binding assays were performed at 4°C, which should prevent the esterification of fatty acids into cell membranes, and decrease the amount of non-specific radioactivity. We contracted with ARC to produce H3-EET methyl ester, which must be saponified to H3-EET, as well as H3-9-HODE. Experiments with these molecules are ongoing.

Technical challenges that could prevent specific binding include insufficient expression of the receptor on the cell membrane; the copy number of GPR132 in β-arrestin cells is unknown. Binding can be

122 performed in very small volumes of cells in solution, however the detachment of cells may cause internalization of the receptor, preventing binding. Attached cells would likely not have this problem, but these would require large volumes of radioligand.

There are also biological explanations for a lack of specific binding. If GPR132 is a low affinity, rather than high affinity receptor, the quantities of radioligand required to see binding would be limiting, as would the necessary quantities of competitor ligand. If EET or 9-HODE activation of GPR132 is indirect,

GPR132 expressing cells would show no greater incorporation of radioligand than other cells. If EET has a membrane path of entry to GPR132, as has been described for lipid-sensing GPCRs such as S1PR and

FFAR1 [124, 125], rather than a more typical extracellular entry point, we might see very little binding at

4°C, and we might see high levels of background. Additionally, it is possible that fatty acids activate

GPR132 not by direct binding, but by changing the localization, conformation, or dimerization state of

GPR132. It has been proposed that DHA incorporation into the membrane can affect the conformation state of the rhodopsin receptor, for instance, possibly by regulating dimerization or localization [153]. Cholesterol concentrations can contribute to dimer formation for proteins such as β2AR [151] and mGlu1 [152]. Again, in these cases we would see no additional accumulation of radioligand in GPR132 expressing cells compared to GPR132 negative cells. These indirect activation hypotheses would likely require high concentrations of fatty acid, which may be consistent with the micromolar quantities necessary in our assays. To test these alternative hypotheses, GFP-tagged GPR132 should be imaged at a subcellular level under conditions of fatty acid treatment. Additionally, FRET probes have been developed for certain GPCRs to assay their oligomeric state [355, 356], and similar methods could be adapted for studying GPR132.

5.2.2. Rho signaling as a mediator of fatty acid/GPR132 phenotypes

GPR132 was shown to activate RhoA via Gα13 in NIH3T3 fibroblasts [204, 205]. LPC-GPR132 signaling was also shown to activate RhoA in several contexts [180, 192, 194]. Overexpression of GPR132 in vitro caused activation of a RhoA reporter, which could be enhanced by low pH [196]. Rho signaling is widely known to be involved in cell migration through the regulation of actin cytoskeletal dynamics [357].

Canonically, Rho activates Rho-associated protein kinase (ROCK), which in turn phosphorylates and inactivates myosin light chain phosphatase, leading to increased cell contractility by activation of myosin II

[358], and causing the formation of stress fibers and focal adhesions [359]. Focal adhesions anchor cells

123 to the ECM, providing mechanical force; they must be regularly turned over for cell movement. Actomyosin contractility is also important for contraction of the cell body and retraction of the trailing cell edge [359].

Gα13 and Rho are activated by CXCR4 in Jurkat T cells, and are required for directional migration of those cells to SDF-1 [360]. In zebrafish, RhoA is required for germ cell migration [361]. Increased expression or activity of Rho proteins also adds to the motility of cancer cells [362]. Rho therefore may have a role downstream of fatty acid/GPR132 signaling in enhancing the migration of HSPCs upon transplant. EET treated HSPCs do show early colonization of the recipient niche, and this could explain their enhanced competitiveness relative to DMSO treated cells.

Many tools exist to study Rho signaling in zebrafish and mouse cells. RhoA is ubiquitously expressed in early zebrafish embryos [361]. RhoA activity can be monitored in live zebrafish and other cell types either directly by expressing a RhoA YPET FRET biosensor [361, 363], or indirectly by monitoring actin dynamics with LifeAct constructs [364]. Small molecule inhibitors targeting RhoGEFs, RhoA itself, and

ROCK are all available [365-368], including the clinically approved ROCK inhibitor fasudil [369], and could be used for co-treatment with EET in zebrafish embryos, or fish or mouse marrow.

Two zebrafish RhoA morpholinos have been validated, but these cause early gastrulation defects, as Rho seems to be important for convergent extension and other gastrulation movements. Morpholino experiments are therefore unlikely to be useful for studying post-gastrulation phenotypes such as HSC specification [370]. Tissue-specific methods may therefore be more useful. Kardash et al [361] expressed a Rho-inhibiting protein as well as a constitutively active RhoA in zebrafish germ cells. Similar experiments could be performed using the flk1 promoter to drive expression in hemogenic endothelium, or the Runx1+23 promoter to drive expression in zebrafish HSPCs. The Gross Lab has developed stable zebrafish lines expressing similar Rho modifiers under a UAS promoter; this allows crossing to tissue specific GAL4 lines to control expression [371]. If Rho is indeed responsible for activating PI3K and AP-1 downstream of fatty acid signaling, then Rho inhibition in flk1+ cells should prevent enhanced expression of hematopoietic markers, and Rho inhibition in specified HSPCs should prevent their enhanced colonization of the CHT.

Similarly, constitutively active Rho signaling should mimic the effects of fatty acid/GPR132 signaling.

124

5.2.3. Cell autonomy of EET signaling

As discussed previously, further experiments are needed to determine whether EET/GPR132 signaling functions autonomously in HSPC transplant. A first functional validation of this hypothesis would be to isolate mouse HSCs, perform EET treatment on this isolated population, followed by transplant with helper marrow into irradiated recipients. If EET is still able to enhance engraftment, then it very likely functions cell autonomously on HSCs, although EET may have combined cell autonomous and non- autonomous effects. The identification of GPR132 as the EET receptor also allows genetic investigation of the autonomous requirement of GPR132. Expression patterns suggest that GPR132 is present in hematopoietic stem cells. We saw expression in sorted mouse LT-HSCs (Eva Fast and Ellen Durand, unpublished data). In mouse tissues, Gpr132 is highest in the spleen and thymus [372], while human

GPR132 is highly expressed in spleen, whole blood and especially transformed lymphocytes [373]. Among mouse blood progenitor populations, ENCODE project data shows Gpr132 expression in hematopoietic stem cells, and to a lesser extent megakaryocyte progenitors. Human GPR132 was also positive for expression in a human CD34+ dataset [374]. B cells, T cells, and macrophages are known to express

GPR132 [184] and to regulate HSPCs [375]; these cells represent the most likely candidates for non-cell autonomous regulation of HSPCs in response to EET/GPR132 signaling.

While RNA expression is indicative of where GPR132 might be functional, protein expression on the cell surface is of course more definitive. We have validated a GPR132 antibody which recognizes the human protein. This could be used for sorting of human marrow or cord blood cells to see what lineages

GPR132 + cells fall into, or what morphology they have in cytospins. The mouse hematopoietic system is even more characterized than the human system, especially for isolation of very specific LT-HSCs. While this antibody did not recognize mouse Gpr132, other antibodies could be tested in order to conduct similar experiments in the mouse. We could also monitor which cell types are functionally responsive to EET using antibodies to pAKT.

A requirement for GPR132 in a particular cell type would also need to be shown functionally. This could be done through tissue-specific knockout of GPR132. Our lab has developed a protocol for tissue- specific expression of gRNAs and Cas9 in the zebrafish, and has validated this method in the blood system

[376]. We could drive Cas9 under the Runx1+23 promoter to knock out gpr132b specifically in HSPCs. We

125 would then assay whether the adult fish has normal basal hematopoiesis, and whether marrow cells still show enhanced transplant upon EET treatment. We can additionally create a floxed Gpr132 allele in the mouse, this can be crossed into Cre lines for a variety of blood cell types to determine the requirement of

Gpr132 in each cell type, in experiments similar to those that have found that SCF must be produced by endothelial and perivascular stromal cells for normal HSC maintenance [377]. Induction of a tamoxifen- inducible Cre line at different stages of development could determine when the HSC defect in GPR132 knockout mice occurs. This defect could be explained by insufficient production of HSCs during development, which would be in line with our zebrafish studies, or it could be the result of impaired maintenance of HSCs in adulthood.

The transplant assay also allows the mixing of cells of different genotypes ex vivo, so that complex mouse genetics might not be needed. We can sort HSCs or other potentially EET-responsive cell types from wildtype or knockout marrow and EET treat these cell types separately. For instance, if B cells are indeed required for EET-enhanced transplant, incubation of wildtype HSPCs with EET-treated wildtype, but not GPR132 knockout, B cells should still enhance transplant. Similarly, if there is no autonomous role of

GPR132 in HSPCs, then GPR132 knockout HSPCs incubated with wildtype, EET-treated marrow should show enhanced transplant.

5.3. Future directions- beyond EET and GPR132 in hematopoiesis

In addition to describing a GPR132-dependent signaling pathway downstream of fatty acid treatment in hematopoiesis, the experiments in this thesis have also highlighted several broader fields of inquiry which merit further investigation. These include the role of GPR132 family members in hematopoiesis generally, a possible role for GPR132 in mediating non-hematopoietic EET phenotypes, and the importance of systems-level lipidomics to understand lipid signaling networks in hematopoiesis and elsewhere.

5.3.1. GPR132 family members as regulators of hematopoiesis

GPR132 has three close family members: GPR4, GPR65, and GPR68. These proteins show similar expression patterns in human cell types, and many studies have shown them having similar responsiveness to pH or LPC and SPC. While we saw no pH or LPC/SPC responsiveness of GPR132 in our assays, it was

126 clear that these molecules remain interesting candidate regulators of hematopoiesis. GPR4 and GPR65 showed no responsiveness to EET in vitro, although one study has suggested that GPR4 can respond to the eicosanoid 11-HETE. While a pre-made assay for GPR68 is not available, the Campbell lab saw no activation of this protein by EET (personal correspondence). However, these assays test one receptor at a time, and it is possible that heterodimerization of the receptors could affect their ligand binding.

RNA-seq revealed that GPR65 and GPR132 are 2 of 85 GPCRs expressed above 5 fragments per million in mouse AGM, and expression assays suggest that both show regulation by GATA2. GPR65 is further regulated by GATA1. Interestingly, another eicosanoid receptor, the leukotriene b4 receptor is also expressed in the AGM and regulated by both GATA1 and GATA2. Knockdown of GPR65 in the zebrafish causes an increase in the expression of hematopoietic markers, while knockdown of GPR132 blocks the

EET-induced increase in hematopoietic markers, and at high doses can decrease basal levels of these markers. These closely related family members therefore have opposite effects on developmental hematopoiesis. The GPR65 effects occur through inhibition of GATA2 in a negative feedback loop. We have not determined whether GPR132 or EET treatment regulates GATA2, or whether GPR65 can regulate pathways such as PI3K and AP-1 that we showed to be downstream of EET. Mouse studies confirmed functionally that GPR65 is conserved as a negative regulator of HSC specification, these studies remain to be done for GPR132.

While the other two family members, GPR4 and GPR68, did not show high expression in mouse

AGM, it will be important to investigate potential hematopoietic phenotypes of these genes as well. GPR68 in particular was one of our receptor candidates expressed specifically in EET-binding cell lines, and may be important in sensing fatty acids, either on its own or as a heterodimer with GPR132 [255]. GPR4 and

GPR68 each have a single homolog in zebrafish, which could be targeted by CRISPR or morpholino. In situ hybridization for these family members could determine whether they are expressed within hematopoietic regions of the embryo. As these four closely related family members may share some function, they may also be related by gene regulatory networks. qPCR should be performed in the context of single receptor knockout or knockdown to determine whether other family members undergo compensatory upregulation. These receptors may also undergo expression changes upon fatty acid treatment.

127

5.3.2. GPR132 in non-hematopoietic EET phenotypes.

The data presented in this thesis strongly supports a role for GPR132 regulating hematopoiesis downstream of free fatty acids. The question remains of whether GPR132 is the high affinity EET receptor thought to mediate phenotypes such as vasodilation, cardioprotection, and anti-inflammation, or whether

GPR132 is a low affinity EET receptor functioning only in the hematopoietic phenotypes described here.

Radioligand binding assays can help answer the question of GPR132’s affinity for EET and other fatty acids.

Additionally, the role of GPR132 in other EET phenotypes should be assayed. The highest potency activity of EET is reported as the activation of calcium gated potassium channels (BK(Ca) channels) in smooth muscle cells, leading to the dilation of the associated vessels, and this function is known to be G-protein dependent [73]. One group reported an EC50 for EET as low as the picomolar range [240], although most reports place this in the low nanomolar range. These assays should be repeated using cells from the

GPR132 KO mouse, although technical challenges exist as these experiments have typically been conducted on bovine or porcine coronary microvessels. Another important assay is macrophage activation.

Mouse M1 macrophages exposed to lipopolysaccharide (LPS) undergo macrophage polarization and produce inflammatory cytokines, but EET treatment at a 1uM concentration can effectively prevent this effect [254]. GPR132 is expressed in macrophages, and was recently reported to induce macrophage polarization in response to lactic acid [248]. We are working with collaborators to determine whether

GPR132 knockout macrophages fail to respond to EET treatment in this assay.

Our GPR132 β-arrestin assays and zebrafish embryo treatments suggest a lack of stereospecificity of GPR132, as it responds robustly to a variety of oxygenated free fatty acids. In contrast, other high potency

EET phenotypes have demonstrated extreme stereospecificity. For instance, 11(R),12(S)-EET causes G- protein dependent translocation of TRPC channels as well as angiogenesis in endothelial cells, while the

11(S),12(R)-EET enantiomer does not [85]. Additionally, while we saw similar activity of 11,12-EET and its derivative 11,12-DHET, many studies using inhibitors of soluble epoxide hydrolase (sEH), which converts

EET to DHET, have found differing roles for these molecules. EET stimulation of BK(Ca) channels in the mouse lung was enhanced by sEH inhibition [252], although another group found that DHETs stimulate rat smooth muscle BK(Ca) channels more potently than EETs [378]. Loss of sEH decreased hematopoietic progenitor proliferation in vivo, suggesting a role for DHET but not EET [97]. We found that sEH inhibitors

128 interfered non-specifically with PathHunter β-arrestin assays (data not shown), but a preliminary experiment conducting sEH inhibition in zebrafish embryos did not attenuate EET treatment phenotypes, suggesting that both EET and DHET are active (data not shown). This small molecule inhibition should be validated with genetic knockdown or knockout; two successful morpholinos for zebrafish sEH have already been described [97].

5.3.3. Lipidomics to investigate fatty acid signaling

Multiple studies from our group and others have now shown the potential for diverse fatty acids to regulate hematopoiesis as well as many other physiological functions. It is time for biology as a whole to look beyond the central dogma of genes, RNAs, and proteins, and to consider the important role of lipid signalers. The hematopoietic system is highly characterized- specific cell types can be isolated, transplants can assay for cellular function, and gene function can be manipulated in a cell-type specific manner. This prototypical stem cell field can now lead the way into systems level lipidomics studies. The bone marrow in particular has a high lipid content that increases with age, and these lipids are able to regulate hematopoiesis. Several groups found that human marrow cells are capable of producing leukotrienes, lipoxins, resolvins, and HETE family members in response to external stimuli [379-381]. One group found altered ratios of leukotrienes and HETEs in chronic myeloid leukemia patients compared to controls [382].

Arachidonic acid, the precursor molecule of EET, is found in rat bone marrow, making up about 2% of the fatty acid composition [383].

Diverse fatty acids and other lipid signaling molecules can be interconverted by metabolic enzymes, and their signaling can result in the regulation of other lipid modifying proteins. The relationship between prostaglandin and EET signaling is especially interesting- PGE2 does not activate GPR132, but PGE2 treatment of mouse marrow causes increased expression of GPR132 (Eva Fast, Ellen Durand, data not shown). PGE2 treatment also increased expression of COX-2, the molecule which synthesizes PGE2 from arachidonic acid. PGE2 treatment could thus create a positive feedback loop and overall increase the responsiveness of a tissue to eicosanoid signaling. In contrast, EET’s anti-inflammatory effects may be mediated in part by decreasing the production of the pro-inflammatory PGE2 [64], perhaps by directly inhibiting COX enzymes [63]. Another group recently showed that EETs are not COX inhibitors but

129 alternative substrates, and that the resulting ct-epoxy-hydroxyeicosatrienoic acids can promote angiogenesis [66].

Given such crosstalk, which may of course vary across different biological contexts, it is important to consider the entire network of lipid signaling molecules as well as their interacting proteins at a systematic level. Systems-level quantification of the lipids present in a given biological sample is known as lipidomics

[384]. In this subfield of metabolomics, most experiments are heavily reliant on mass spectrometry. Two primary methods of lipidomics exist: “shotgun lipidomics,” where the entire complexity of a given class of lipids within a sample can be queried and quantified, and LC-MS based lipidomics, where a complex lipid sample is separated by liquid chromatography, followed by quantification by mass spec [385]. Lipidomics on human tissues is providing potential biomarkers for disease states [386], and is expanding our understanding of lipid signaling as a whole. Currently, lipid extraction procedures as well as mass spec parameters determine the scope of lipids that can be quantified.

There are many challenges to direct manipulation of the endogenous lipidome. Most lipids cannot be pulled down using an antibody, and fluorescent or other tags interfere greatly with their chemistry and signaling. Lipids show enormous diversity in vivo and it is unlikely that we currently understand the full diversity of fatty acid signaling molecules. Fatty acid receptors and modifying enzymes can be inhibited or genetically targeted, but these often function on an entire class of molecules. The primary role of sEH, for instance, was long thought to be hydrolysis of the four EETs isomers to DHETs. However, sEH also produces DiHOMEs from EpOMEs [97, 387], and has more recently been shown to function on epoxide derivatives of EPA and DHA [388]. The studies in this thesis have revealed GPR132 to be an important node in the lipidome, integrating the signaling of several different fatty acids. This means that genetic or pharmacological manipulation of GPR132 would perturb not just EET signaling, but also DHET, 9-HODE,

11-HETE, and 9,10-EpOME signaling, and likely signaling downstream of other oxygenated fatty acids not tested here.

The Serhan lab at Brigham and Women’s hospital has designed an HPLC-MS/MS method for precise quantification of arachidonic acid, EPA, DHA, and important oxygenated products of these molecules, constituting a small lipidomics screen of 22 lipids. In a preliminary experiment, we treated zebrafish embryos with EET for 1 or 8 hours and then performed HPLC-MS/MS. At 1 hour post treatment,

130 increased AP-1 gene expression is already seen in embryos, although increased staining of hematopoietic markers takes longer. We saw no changes in the 22 lipids at 1 hour of treatment. However, at 8 hours we saw a loss of resolvin D5 and increased production of PGE2 and PGF2α. One study found that Gα13 can activate Rho to cause increased transcription of COX-2, which could explain the production of prostaglandins [389]. While these changes are likely too late to be regulating the specification of HSCs in zebrafish embryos, they may be relevant in EET’s HSC transplant phenotypes. These assays should be repeated and expanded to a wider list of lipid mediators, and they should be performed in zebrafish and mouse marrow. We are currently in consultation with James Cox, the head of the metabolomics core at the

University of Utah, to attempt to establish lipid quantification assays that could be performed on very small samples including zebrafish embryos or marrow.

Lipidomic studies looking at the changes in the lipid network upon treatment with a given fatty acid do not distinguish between direct conversion of the added fatty acid to another species and indirect stimulation of production of that other species. To follow the treated lipid, radiolabeled or deuterated fatty acids can be employed. The change in mass can distinguish the treatment molecule from endogenous lipids. We treated zebrafish embryos with deuterated 11,12-EET for 1 or 8 hours and found that at only 1 hour post treatment, over 90% of the d-11,12-EET remained present as 11,12-EET, but by 8 hours post treatment over 90% had been converted into other lipid species (data not shown). With the appropriate lipid extraction and mass spec parameters, shotgun lipidomics approaches would enable us to follow that d-

11,12-EET as it undergoes metabolic conversions.

Basic questions that could be answered with lipidomics analysis of marrow include, how does lipid content of the marrow change with age or other stress such as bleeding or irradiation? Are there sex- specific differences in marrow lipid content? How does lipid content differ across cell types within the marrow? How does marrow content change with fatty acid treatment? Do transplanted or diseased organisms show altered lipidomic profiles? Data from these experiments could point to protein targets that are regulating the different states. The mouse is currently the best situated model organism for these experiments, as hematopoietic cell types are highly defined, and experiments such as transplants and repeated bleeds can be performed in a highly controlled manner. However, as zebrafish can be maintained in large numbers and are not inbred, they could be used for studies of lipidomic diversity within a

131 heterogeneous population. Regulators of the lipid signaling network might be identified by correlating differences in the lipidome to genetic or epigenetic differences across zebrafish strains.

Lipidomics of the marrow could also shed light on the responsiveness of GPR132 to endogenous ligands. Our studies, applying micromolar concentrations of EET and other fatty acids, have shown GPR132 to be a pharmacologically relevant target of these molecules. Additionally, our mouse transplant experiments have shown that GPR132 has an endogenous role in regulating HSCs. However, we have not determined whether this endogenous role is fatty acid dependent, or whether GPR132 signals constitutively or in response to another ligand. We should investigate the endogenous presence of fatty acids such as

11,12-EET and 9-HODE in the marrow. With knowledge of the affinity of GPR132 for these molecules from direct binding studies, we can determine whether they are present in the bone marrow or peripheral blood at sufficient concentrations to activate GPR132 endogenously. It will be important to consider that as

GPR132 is responsive to a variety of lipids, if different fatty acids are present at insufficient concentrations individually, they may converge to activate GPR132. Additionally, local concentrations of the ligand would be very important. Membrane deposition of fatty acids provides an intriguing possibility for achieving high concentrations within in a single cell or even sub-cellular domain.

132

5.4. Conclusion

As we use lipidomics to expand our knowledge of the diversity and quantity of fatty acids present within tissues under different experimental perturbations, we must also continue to explore the proteins which regulate lipid signaling by processing or responding to lipids. GPR132 may represent a node within the lipid signaling network, able to integrate the signaling of diverse oxygenated fatty acids. We have shown that oxygenated free fatty acids activate GPR132 to enhance hematopoietic signaling pathways in zebrafish embryos, and that GPR132 is required for normal HSC function in adult mice. The closely related GPR65 is instead a negative regulator of HSC specification, and the family members GPR4 and GPR68 remain to be investigated for their roles in hematopoiesis. Specific binding assays will reveal the affinity of GPR132 for free fatty acids and bolster its identity as a direct lipid receptor. Expression studies and conditional or tissue specific knockout of GPR132 will determine whether this protein is responsible for canonical EET phenotypes such as vasodilation or macrophage activation, and can illuminate the cell autonomy of fatty acid/GPR132 signaling in regulating HSPCs. EET and GPR132 activate their hematopoietic phenotypes by increasing Gα12/13, PI3Kγ, and AP-1 signaling. Further experiments using transgenic zebrafish or mouse marrow can reveal how these factors relate to one another and to the transplant of HSPCs. Rho signaling is an especially promising target of both Gα12/13 and PI3K with known roles in cell migration. The oxygenated fatty acid-GPR132 signaling axis is an important regulator of hematopoiesis and a potential therapeutic target for enhancement of HSC transplant. The discovery of this lipid-GPCR interaction underscores the need for systems level study of lipid mediators and their protein partners.

133

References

1. Cooper, B., The origins of bone marrow as the seedbed of our blood: from antiquity to the time of Osler. Proc (Bayl Univ Med Cent), 2011. 24(2): p. 115-8.

2. Thomas, E.D., et al., Supralethal whole body irradiation and isologous marrow transplantation in man. J Clin Invest, 1959. 38: p. 1709-16.

3. Orkin, S.H. and L.I. Zon, Hematopoiesis: an evolving paradigm for stem cell biology. Cell, 2008. 132(4): p. 631-644.

4. Schofield, R., The relationship between the spleen colony-forming cell and the haemopoietic stem cell. Blood Cells, 1978. 4(1-2): p. 7-25.

5. Serhan, C.N., et al., Resolvins: a family of bioactive products of omega-3 fatty acid transformation circuits initiated by aspirin treatment that counter proinflammation signals. J Exp Med, 2002. 196(8): p. 1025-37.

6. Serhan, C.N., et al., Maresins: novel macrophage mediators with potent antiinflammatory and proresolving actions. J Exp Med, 2009. 206(1): p. 15-23.

7. Ballen, K.K., et al., The national marrow donor program 20 years of unrelated donor hematopoietic cell transplantation. Biology of blood and marrow transplantation : journal of the American Society for Blood and Marrow Transplantation, 2008. 14(9 Suppl): p. 2-7.

8. Kekre, N. and J.H. Antin, Cord blood versus haploidentical stem cell transplantation for hematological malignancies. Semin Hematol, 2016. 53(2): p. 98-102.

9. Kekre, N. and J.H. Antin, Hematopoietic stem cell transplantation donor sources in the 21st century: choosing the ideal donor when a perfect match does not exist. Blood, 2014. 124(3): p. 334-343.

10. Berglund, S., et al., Advances in umbilical cord blood cell therapy: the present and the future. Expert Opin Biol Ther, 2017.

11. Horwitz, M.E., Ex Vivo Expansion or Manipulation of Stem Cells to Improve Outcome of Umbilical Cord Blood Transplantation. Curr Hematol Malig Rep, 2016. 11(1): p. 12-8.

12. Baron, F., A. Ruggeri, and A. Nagler, Methods of ex vivo expansion of human cord blood cells: challenges, successes and clinical implications. Expert Rev Hematol, 2016. 9(3): p. 297-314.

134

13. Baron, F. and A. Nagler, Novel strategies for improving hematopoietic reconstruction after allogeneic hematopoietic stem cell transplantation or intensive chemotherapy. Expert Opin Biol Ther, 2017. 17(2): p. 163-174.

14. Hagedorn, E.J., et al., Getting more for your marrow: Boosting hematopoietic stem cell numbers with PGE2. Experimental cell research, 2014.

15. Cutler, C. and K.K. Ballen, Improving outcomes in umbilical cord blood transplantation: state of the art. Blood reviews, 2012. 26(6): p. 241-246.

16. Kiernan, J., et al., Clinical Studies of Ex Vivo Expansion to Accelerate Engraftment After Umbilical Cord Blood Transplantation: A Systematic Review. Transfus Med Rev, 2016.

17. Delaney, C., et al., Notch-mediated expansion of human cord blood progenitor cells capable of rapid myeloid reconstitution. Nature Medicine, 2010.

18. Horwitz, M.E., et al., Umbilical cord blood expansion with nicotinamide provides long-term multilineage engraftment. J Clin Invest, 2014. 124(7): p. 3121-8.

19. Wagner, J.E., Jr., et al., Phase I/II Trial of StemRegenin-1 Expanded Umbilical Cord Blood Hematopoietic Stem Cells Supports Testing as a Stand-Alone Graft. Cell Stem Cell, 2016. 18(1): p. 144-55.

20. Gori, J.L., et al., Efficient generation, purification, and expansion of CD34(+) hematopoietic progenitor cells from nonhuman primate-induced pluripotent stem cells. Blood, 2012. 120(13): p. e35-44.

21. Boitano, A.E., et al., Aryl hydrocarbon receptor antagonists promote the expansion of human hematopoietic stem cells. Science, 2010. 329(5997): p. 1345-8.

22. de Lima, M., et al., Cord-blood engraftment with ex vivo mesenchymal-cell coculture. N Engl J Med, 2012. 367(24): p. 2305-15.

23. Goessling, W., et al., Prostaglandin E2 enhances human cord blood stem cell xenotransplants and shows long-term safety in preclinical nonhuman primate transplant models. Cell Stem Cell, 2011. 8(4): p. 445-58.

24. Goessling, W., et al., Genetic interaction of PGE2 and Wnt signaling regulates developmental specification of stem cells and regeneration. Cell, 2009. 136(6): p. 1136-47.

25. North, T.E., et al., Prostaglandin E2 regulates vertebrate haematopoietic stem cell homeostasis. Nature, 2007. 447(7147): p. 1007-11.

135

26. Cutler, C., et al., Prostaglandin-modulated umbilical cord blood hematopoietic stem cell transplantation. Blood, 2013. 122(17): p. 3074-81.

27. Aljitawi, O.S., et al., Erythropoietin modulation is associated with improved homing and engraftment after umbilical cord blood transplantation. Blood, 2016. 128(25): p. 3000-3010.

28. Popat, U., et al., Enforced fucosylation of cord blood hematopoietic cells accelerates neutrophil and platelet engraftment after transplantation. Blood, 2015. 125(19): p. 2885-92.

29. Robinson, S.N., et al., Ex vivo fucosylation improves human cord blood engraftment in NOD-SCID IL-2Rgamma(null) mice. Exp Hematol, 2012. 40(6): p. 445-56.

30. Bertrand, J.Y., et al., Haematopoietic stem cells derive directly from aortic endothelium during development. Nature, 2010. 464(7285): p. 108-11.

31. Kissa, K. and P. Herbomel, Blood stem cells emerge from aortic endothelium by a novel type of cell transition. Nature, 2010. 464(7285): p. 112-5.

32. Boisset, J.C., et al., In vivo imaging of haematopoietic cells emerging from the mouse aortic endothelium. Nature, 2010. 464(7285): p. 116-20.

33. Chen, M.J., et al., Runx1 is required for the endothelial to haematopoietic cell transition but not thereafter. Nature, 2009. 457(7231): p. 887-91.

34. Murayama, E., et al., Tracing hematopoietic precursor migration to successive hematopoietic organs during zebrafish development. Immunity, 2006. 25(6): p. 963-75.

35. Wiley, D.M., et al., Distinct signalling pathways regulate sprouting angiogenesis from the dorsal aorta and the axial vein. Nat Cell Biol, 2011. 13(6): p. 686-92.

36. Tamplin, O.J., et al., Hematopoietic stem cell arrival triggers dynamic remodeling of the perivascular niche. Cell, 2015. 160(1-2): p. 241-52.

37. Butko, E., C. Pouget, and D. Traver, Complex regulation of HSC emergence by the Notch signaling pathway. Dev Biol, 2016. 409(1): p. 129-38.

38. Butler, J.M., et al., Endothelial cells are essential for the self-renewal and repopulation of Notch- dependent hematopoietic stem cells. Cell Stem Cell, 2010. 6(3): p. 251-64.

39. Calvi, L.M., et al., Osteoblastic cells regulate the haematopoietic stem cell niche. Nature, 2003. 425(6960): p. 841-6.

136

40. Blaser, B.W., et al., CXCR1 remodels the vascular niche to promote hematopoietic stem and progenitor cell engraftment. J Exp Med, 2017. 214(4): p. 1011-1027.

41. Pelus, L.M., et al., Peripheral blood stem cell mobilization. A role for CXC chemokines. Crit Rev Oncol Hematol, 2002. 43(3): p. 257-75.

42. Hale, S.J., et al., CXCR2 modulates bone marrow vascular repair and haematopoietic recovery post-transplant. Br J Haematol, 2015. 169(4): p. 552-64.

43. Mahony, C.B., et al., tfec controls the hematopoietic stem cell vascular niche during zebrafish embryogenesis. Blood, 2016. 128(10): p. 1336-45.

44. Kaushansky, K., Thrombopoietin and its receptor in normal and neoplastic hematopoiesis. Thromb J, 2016. 14(Suppl 1): p. 40.

45. Asada, N., et al., Differential cytokine contributions of perivascular haematopoietic stem cell niches. Nat Cell Biol, 2017. 19(3): p. 214-223.

46. Baron, M.H., J. Isern, and S.T. Fraser, The embryonic origins of erythropoiesis in mammals. Blood, 2012. 119(21): p. 4828-37.

47. Lensch, M.W., An evolving model of hematopoietic stem cell functional identity. Stem Cell Rev, 2012. 8(2): p. 551-60.

48. Mazo, I.B., et al., Hematopoietic progenitor cell rolling in bone marrow microvessels: parallel contributions by endothelial selectins and vascular cell adhesion molecule 1. J Exp Med, 1998. 188(3): p. 465-74.

49. Peled, A., et al., The chemokine SDF-1 stimulates integrin-mediated arrest of CD34(+) cells on vascular endothelium under shear flow. J Clin Invest, 1999. 104(9): p. 1199-211.

50. Shirvaikar, N., L.A. Marquez-Curtis, and A. Janowska-Wieczorek, Hematopoietic Stem Cell Mobilization and Homing after Transplantation: The Role of MMP-2, MMP-9, and MT1-MMP. Biochem Res Int, 2012. 2012: p. 685267.

51. Gu, Y., et al., Hematopoietic cell regulation by Rac1 and Rac2 guanosine triphosphatases. Science, 2003. 302(5644): p. 445-9.

52. Levesque, J.P., et al., Disruption of the CXCR4/CXCL12 chemotactic interaction during hematopoietic stem cell mobilization induced by GCSF or cyclophosphamide. J Clin Invest, 2003. 111(2): p. 187-96.

137

53. Mazo, I.B., et al., Total body irradiation causes profound changes in endothelial traffic molecules for hematopoietic progenitor cell recruitment to bone marrow. Blood, 2002. 99(11): p. 4182-91.

54. Goossens, S., et al., The EMT regulator Zeb2/Sip1 is essential for murine embryonic hematopoietic stem/progenitor cell differentiation and mobilization. Blood, 2011. 117(21): p. 5620-5630.

55. Jeppesen, C., K. Schiller, and M.B. Schulze, Omega-3 and omega-6 fatty acids and type 2 diabetes. Curr Diab Rep, 2013. 13(2): p. 279-88.

56. Buczynski, M.W., D.S. Dumlao, and E.A. Dennis, Thematic Review Series: Proteomics. An integrated omics analysis of eicosanoid biology. J Lipid Res, 2009. 50(6): p. 1015-38.

57. Wang, J.F. and K.C. Chou, Molecular modeling of cytochrome P450 and drug metabolism. Curr Drug Metab, 2010. 11(4): p. 342-6.

58. Xu, X., X.A. Zhang, and D.W. Wang, The roles of CYP450 epoxygenases and metabolites, epoxyeicosatrienoic acids, in cardiovascular and malignant diseases. Adv Drug Deliv Rev, 2011. 63(8): p. 597-609.

59. Spector, A.A. and H.Y. Kim, Cytochrome P450 epoxygenase pathway of polyunsaturated fatty acid metabolism. Biochim Biophys Acta, 2015. 1851(4): p. 356-65.

60. Im, D.S., New intercellular lipid mediators and their GPCRs: an update. Prostaglandins Other Lipid Mediat, 2009. 89(3-4): p. 53-6.

61. Singh, R.K., et al., A review on leukotrienes and their receptors with reference to asthma. J Asthma, 2013. 50(9): p. 922-31.

62. Powell, W.S. and J. Rokach, Biosynthesis, biological effects, and receptors of hydroxyeicosatetraenoic acids (HETEs) and oxoeicosatetraenoic acids (oxo-ETEs) derived from arachidonic acid. Biochim Biophys Acta, 2015. 1851(4): p. 340-55.

63. Fang, X., et al., 14,15-Epoxyeicosatrienoic acid inhibits prostaglandin E2 production in vascular smooth muscle cells. Am J Physiol, 1998. 275(6 Pt 2): p. H2113-21.

64. Kozak, W., et al., 11,12-epoxyeicosatrienoic acid attenuates synthesis of prostaglandin E2 in rat monocytes stimulated with lipopolysaccharide. Exp Biol Med (Maywood), 2003. 228(7): p. 786-94.

65. Wang, D., et al., Eicosapentaenoic acid increases cytochrome P-450 2J2 gene expression and epoxyeicosatrienoic acid production via peroxisome proliferator-activated receptor gamma in endothelial cells. J Cardiol, 2009. 54(3): p. 368-74.

138

66. Rand, A.A., et al., Cyclooxygenase-derived proangiogenic metabolites of epoxyeicosatrienoic acids. Proc Natl Acad Sci U S A, 2017.

67. Spector, A., et al., Epoxyeicosatrienoic acids (EETs): metabolism and biochemical function. Progress in lipid research, 2004. 43(1): p. 55-90.

68. Panigrahy, D., et al., EET signaling in cancer. Cancer metastasis reviews, 2011. 30(3-4): p. 525- 540.

69. Rosolowsky, M. and W.B. Campbell, Synthesis of hydroxyeicosatetraenoic (HETEs) and epoxyeicosatrienoic acids (EETs) by cultured bovine coronary artery endothelial cells. Biochim Biophys Acta, 1996. 1299(2): p. 267-77.

70. Fisslthaler, B., et al., Cytochrome P450 2C is an EDHF synthase in coronary arteries. Nature, 1999. 401(6752): p. 493-7.

71. Karara, A., et al., Arachidonic acid epoxygenase: structural characterization and quantification of epoxyeicosatrienoates in plasma. Biochem Biophys Res Commun, 1992. 182(3): p. 1320-5.

72. Karara, A., et al., Endogenous epoxyeicosatrienoyl-phospholipids. A novel class of cellular glycerolipids containing epoxidized arachidonate moieties. J Biol Chem, 1991. 266(12): p. 7561-9.

73. Li, P.L. and W.B. Campbell, Epoxyeicosatrienoic acids activate K+ channels in coronary smooth muscle through a guanine nucleotide binding protein. Circulation research, 1997. 80(6): p. 877- 884.

74. Behm, D.J., et al., Epoxyeicosatrienoic acids function as selective, endogenous antagonists of native thromboxane receptors: identification of a novel mechanism of vasodilation. J Pharmacol Exp Ther, 2009. 328(1): p. 231-9.

75. Carroll, M.A., et al., Adenosine2A receptor vasodilation of rat preglomerular microvessels is mediated by EETs that activate the cAMP/PKA pathway. American journal of physiology. Renal physiology, 2006. 291(1): p. 61.

76. Hye Khan, M.A., et al., Epoxyeicosatrienoic acid analogue lowers blood pressure through vasodilation and sodium channel inhibition. Clinical science (London, England : 1979), 2014. 127(7): p. 463-474.

77. Lee, C.R., et al., Endothelial expression of human cytochrome P450 epoxygenases lowers blood pressure and attenuates hypertension-induced renal injury in mice. Faseb j, 2010. 24(10): p. 3770- 81.

139

78. Yang, C., et al., Vasodilatory effect of 14,15-epoxyeicosatrienoic acid on mesenteric arteries in hypertensive and aged rats. Prostaglandins & other lipid mediators, 2014. 112: p. 1-8.

79. Zhang, P., et al., Nitric oxide and protein kinase G act on TRPC1 to inhibit 11,12-EET-induced vascular relaxation. Cardiovascular research, 2014. 104(1): p. 138-146.

80. Node, K., et al., Anti-inflammatory properties of cytochrome P450 epoxygenase-derived eicosanoids. Science (New York, N.Y.), 1999. 285(5431): p. 1276-1279.

81. Chen, W., et al., CYP2J2 and EETs protect against lung ischemia/reperfusion injury via anti- inflammatory effects in vivo and in vitro. Cellular physiology and biochemistry : international journal of experimental cellular physiology, biochemistry, and pharmacology, 2015. 35(5): p. 2043-2054.

82. Munzenmaier, D. and D. Harder, Cerebral microvascular endothelial cell tube formation: role of astrocytic epoxyeicosatrienoic acid release. American journal of physiology. Heart and circulatory physiology, 2000. 278(4): p. 7.

83. Wang, Y., et al., Arachidonic acid epoxygenase metabolites stimulate endothelial cell growth and angiogenesis via mitogen-activated protein kinase and phosphatidylinositol 3-kinase/Akt signaling pathways. The Journal of pharmacology and experimental therapeutics, 2005. 314(2): p. 522-532.

84. Su, K.H., et al., Implication of transient receptor potential vanilloid type 1 in 14,15- epoxyeicosatrienoic acid-induced angiogenesis. Int J Biol Sci, 2014. 10(9): p. 990-6.

85. Ding, Y., et al., The biological actions of 11,12-epoxyeicosatrienoic acid in endothelial cells are specific to the R/S-enantiomer and require the G(s) protein. The Journal of pharmacology and experimental therapeutics, 2014. 350(1): p. 14-21.

86. Pozzi, A., et al., Characterization of 5,6- and 8,9-epoxyeicosatrienoic acids (5,6- and 8,9-EET) as potent in vivo angiogenic lipids. J Biol Chem, 2005. 280(29): p. 27138-46.

87. Jiang, J.G., et al., Cytochrome P450 2J2 promotes the neoplastic phenotype of carcinoma cells and is up-regulated in human tumors. Cancer Res, 2005. 65(11): p. 4707-15.

88. Jiang, J.G., et al., Cytochrome p450 epoxygenase promotes human cancer metastasis. Cancer Res, 2007. 67(14): p. 6665-74.

89. Nithipatikom, K., et al., Inhibition of carcinoma cell motility by epoxyeicosatrienoic acid (EET) antagonists. Cancer Sci, 2010. 101(12): p. 2629-36.

90. Mitra, R., et al., CYP3A4 mediates growth of estrogen receptor-positive breast cancer cells in part by inducing nuclear translocation of phospho-Stat3 through biosynthesis of (+/-)-14,15- epoxyeicosatrienoic acid (EET). J Biol Chem, 2011. 286(20): p. 17543-59.

140

91. Sun, J., et al., Inhibition of vascular smooth muscle cell migration by cytochrome p450 epoxygenase-derived eicosanoids. Circ Res, 2002. 90(9): p. 1020-7.

92. Panigrahy, D., et al., Epoxyeicosanoids stimulate multiorgan metastasis and tumor dormancy escape in mice. The Journal of clinical investigation, 2012. 122(1): p. 178-191.

93. Wei, X., et al., Elevated 14,15- epoxyeicosatrienoic acid by increasing of cytochrome P450 2C8, 2C9 and 2J2 and decreasing of soluble epoxide hydrolase associated with aggressiveness of human breast cancer. BMC Cancer, 2014. 14: p. 841.

94. Jiang, H., G.D. Anderson, and J.C. McGiff, Red blood cells (RBCs), epoxyeicosatrienoic acids (EETs) and adenosine triphosphate (ATP). Pharmacol Rep, 2010. 62(3): p. 468-74.

95. Jiang, H., G.D. Anderson, and J.C. McGiff, The red blood cell participates in regulation of the circulation by producing and releasing epoxyeicosatrienoic acids. Prostaglandins Other Lipid Mediat, 2012. 98(3-4): p. 91-3.

96. Jiang, H., et al., Red blood cells: reservoirs of cis- and trans-epoxyeicosatrienoic acids. Prostaglandins Other Lipid Mediat, 2005. 75(1-4): p. 65-78.

97. Fromel, T., et al., Soluble epoxide hydrolase regulates hematopoietic progenitor cell function via generation of fatty acid diols. Proc Natl Acad Sci U S A, 2012. 109(25): p. 9995-10000.

98. Gao, Y., et al., 8,9-Epoxyeicosatrienoic acid inhibits antibody production of B lymphocytes in mice. PLoS One, 2012. 7(7): p. e40258.

99. Krotz, F., et al., Membrane-potential-dependent inhibition of platelet adhesion to endothelial cells by epoxyeicosatrienoic acids. Arterioscler Thromb Vasc Biol, 2004. 24(3): p. 595-600.

100. Li, D., et al., Inhibition of soluble epoxide hydrolase alleviated atherosclerosis by reducing monocyte infiltration in Ldlr(-/-) mice. J Mol Cell Cardiol, 2016. 98: p. 128-37.

101. Pratt, P.F., M. Rosolowsky, and W.B. Campbell, Effects of epoxyeicosatrienoic acids on polymorphonuclear leukocyte function. Life Sci, 2002. 70(21): p. 2521-33.

102. Rompe, F., et al., Direct angiotensin II type 2 receptor stimulation acts anti-inflammatory through epoxyeicosatrienoic acid and inhibition of nuclear factor kappaB. Hypertension, 2010. 55(4): p. 924- 31.

103. Cheng, L.M., et al., The epoxyeicosatrienoic acid-stimulated phosphorylation of EGF-R involves the activation of metalloproteinases and the release of HB-EGF in cancer cells. Acta Pharmacol Sin, 2010. 31(2): p. 211-8.

141

104. Pidkovka, N., et al., Epoxyeicosatrienoic acids (EETs) regulate epithelial sodium channel activity by extracellular signal-regulated kinase 1/2 (ERK1/2)-mediated phosphorylation. J Biol Chem, 2013. 288(7): p. 5223-31.

105. Ma, S.K., et al., Overexpression of G-protein-coupled receptor 40 enhances the mitogenic response to epoxyeicosatrienoic acids. PloS one, 2015. 10(2).

106. Qu, Y.Y., et al., The protective effect of epoxyeicosatrienoic acids on cerebral ischemia/reperfusion injury is associated with PI3K/Akt pathway and ATP-sensitive potassium channels. Neurochem Res, 2015. 40(1): p. 1-14.

107. Batchu, S.N., et al., Role of PI3Kalpha and sarcolemmal ATP-sensitive potassium channels in epoxyeicosatrienoic acid mediated cardioprotection. J Mol Cell Cardiol, 2012. 53(1): p. 43-52.

108. Bodiga, S., et al., Protective actions of epoxyeicosatrienoic acid: dual targeting of cardiovascular PI3K and KATP channels. J Mol Cell Cardiol, 2009. 46(6): p. 978-88.

109. Zhang, L., et al., 11,12-Epoxyeicosatrienoic acid activates the L-arginine/nitric oxide pathway in human platelets. Mol Cell Biochem, 2008. 308(1-2): p. 51-6.

110. Snyder, G., et al., Evidence for a membrane site of action for 14,15-EET on expression of aromatase in vascular smooth muscle. American journal of physiology. Heart and circulatory physiology, 2002. 283(5): p. 42.

111. Fang, X., et al., 14,15-Dihydroxyeicosatrienoic acid activates peroxisome proliferator-activated receptor-alpha. American journal of physiology. Heart and circulatory physiology, 2006. 290(1): p. 63.

112. Ng, V., et al., Cytochrome P450 eicosanoids are activators of peroxisome proliferator-activated receptor alpha. Drug metabolism and disposition: the biological fate of chemicals, 2007. 35(7): p. 1126-1134.

113. Chen, Y., et al., 20-Iodo-14,15-epoxyeicosa-8(Z)-enoyl-3-azidophenylsulfonamide: photoaffinity labeling of a 14,15-epoxyeicosatrienoic acid receptor. Biochemistry, 2011. 50(18): p. 3840-8.

114. Liu, X., et al., Functional screening for G protein-coupled receptor targets of 14,15- epoxyeicosatrienoic acid. Prostaglandins Other Lipid Mediat, 2016.

115. Yang, C., et al., 14,15-Epoxyeicosatrienoic acid induces vasorelaxation through the prostaglandin EP(2) receptors in rat mesenteric artery. Prostaglandins Other Lipid Mediat, 2010. 93(1-2): p. 44- 51.

142

116. Ma, Y., et al., Epoxyeicosatrienoic acids act through TRPV4-TRPC1-KCa1.1 complex to induce smooth muscle membrane hyperpolarization and relaxation in human internal mammary arteries. Biochimica et biophysica acta, 2015. 1852(3): p. 552-559.

117. Fleming, I., et al., Epoxyeicosatrienoic acids regulate Trp channel dependent Ca2+ signaling and hyperpolarization in endothelial cells. Arterioscler Thromb Vasc Biol, 2007. 27(12): p. 2612-8.

118. Watanabe, H., et al., Anandamide and arachidonic acid use epoxyeicosatrienoic acids to activate TRPV4 channels. Nature, 2003. 424(6947): p. 434-8.

119. Zheng, X., et al., Arachidonic acid-induced dilation in human coronary arterioles: convergence of signaling mechanisms on endothelial TRPV4-mediated Ca2+ entry. J Am Heart Assoc, 2013. 2(3): p. e000080.

120. Lagerstrom, M.C. and H.B. Schioth, Structural diversity of G protein-coupled receptors and significance for drug discovery. Nat Rev Drug Discov, 2008. 7(4): p. 339-57.

121. Lappano, R. and M. Maggiolini, G protein-coupled receptors: novel targets for drug discovery in cancer. Nat Rev Drug Discov, 2011. 10(1): p. 47-60.

122. Sanchez-Fernandez, G., et al., Galphaq signalling: the new and the old. Cell Signal, 2014. 26(5): p. 833-48.

123. Liu, J.J., et al., Biased signaling pathways in beta2-adrenergic receptor characterized by 19F-NMR. Science, 2012. 335(6072): p. 1106-10.

124. Hanson, M.A., et al., Crystal structure of a lipid G protein-coupled receptor. Science, 2012. 335(6070): p. 851-5.

125. Srivastava, A., et al., High-resolution structure of the human GPR40 receptor bound to allosteric agonist TAK-875. Nature, 2014. 513(7516): p. 124-7.

126. Stanley, N., L. Pardo, and G.D. Fabritiis, The pathway of ligand entry from the membrane bilayer to a lipid G protein-coupled receptor. Sci Rep, 2016. 6: p. 22639.

127. Hurst, D.P., et al., A lipid pathway for ligand binding is necessary for a cannabinoid G protein- coupled receptor. J Biol Chem, 2010. 285(23): p. 17954-64.

128. Agosti, F., et al., Melanocortin 4 receptor constitutive activity inhibits L-type voltage-gated calcium channels in neurons. Neuroscience, 2017. 346: p. 102-112.

143

129. Bond, R.A. and A.P. Ijzerman, Recent developments in constitutive receptor activity and inverse agonism, and their potential for GPCR drug discovery. Trends Pharmacol Sci, 2006. 27(2): p. 92- 6.

130. Costa, T. and A. Herz, Antagonists with negative intrinsic activity at delta opioid receptors coupled to GTP-binding proteins. Proc Natl Acad Sci U S A, 1989. 86(19): p. 7321-5.

131. Damian, M., et al., High constitutive activity is an intrinsic feature of ghrelin receptor protein: a study with a functional monomeric GHS-R1a receptor reconstituted in lipid discs. J Biol Chem, 2012. 287(6): p. 3630-41.

132. Meye, F.J., G.M. Ramakers, and R.A. Adan, The vital role of constitutive GPCR activity in the mesolimbic dopamine system. Transl Psychiatry, 2014. 4: p. e361.

133. Birdsall, N.J., Class A GPCR heterodimers: evidence from binding studies. Trends Pharmacol Sci, 2010. 31(11): p. 499-508.

134. Borroto-Escuela, D.O., et al., The G protein-coupled receptor heterodimer network (GPCR-HetNet) and its hub components. Int J Mol Sci, 2014. 15(5): p. 8570-90.

135. Ferre, S., et al., G protein-coupled receptor oligomerization revisited: functional and pharmacological perspectives. Pharmacol Rev, 2014. 66(2): p. 413-34.

136. Herrick-Davis, K., et al., Serotonin 5-HT2C receptor homodimer biogenesis in the endoplasmic reticulum: real-time visualization with confocal fluorescence resonance energy transfer. J Biol Chem, 2006. 281(37): p. 27109-16.

137. Teitler, M. and M.T. Klein, A new approach for studying GPCR dimers: drug-induced inactivation and reactivation to reveal GPCR dimer function in vitro, in primary culture, and in vivo. Pharmacol Ther, 2012. 133(2): p. 205-17.

138. Sasaki, M., et al., Physiological significance of thromboxane A(2) receptor dimerization. J Pharmacol Sci, 2006. 100(4): p. 263-70.

139. Laroche, G., et al., Oligomerization of the alpha and beta isoforms of the thromboxane A2 receptor: relevance to receptor signaling and endocytosis. Cell Signal, 2005. 17(11): p. 1373-83.

140. Baneres, J.L. and J. Parello, Structure-based analysis of GPCR function: evidence for a novel pentameric assembly between the dimeric leukotriene B4 receptor BLT1 and the G-protein. J Mol Biol, 2003. 329(4): p. 815-29.

141. Gasperi, V., et al., GPR55 and its interaction with membrane lipids: comparison with other endocannabinoid-binding receptors. Curr Med Chem, 2013. 20(1): p. 64-78.

144

142. Huang, Y.S., et al., Activation of myeloid cell-specific adhesion class G protein-coupled receptor EMR2 via ligation-induced translocation and interaction of receptor subunits in lipid raft microdomains. Mol Cell Biol, 2012. 32(8): p. 1408-20.

143. Insel, P.A., et al., Caveolae and lipid rafts: G protein-coupled receptor signaling microdomains in cardiac myocytes. Ann N Y Acad Sci, 2005. 1047: p. 166-72.

144. Villar, V.A., et al., Localization and signaling of GPCRs in lipid rafts. Methods Cell Biol, 2016. 132: p. 3-23.

145. Bowman, S.L., A.L. Soohoo, and M.A. Puthenveedu, Visualizing and quantitating sequence- dependent GPCR recycling. Methods Cell Biol, 2015. 130: p. 333-45.

146. Magalhaes, A.C., H. Dunn, and S.S. Ferguson, Regulation of GPCR activity, trafficking and localization by GPCR-interacting proteins. Br J Pharmacol, 2012. 165(6): p. 1717-36.

147. Ray, K., Calcium-Sensing Receptor: Trafficking, Endocytosis, Recycling, and Importance of Interacting Proteins. Prog Mol Biol Transl Sci, 2015. 132: p. 127-50.

148. Tower-Gilchrist, C., E. Lee, and E. Sztul, Endosomal trafficking of the G protein-coupled receptor somatostatin receptor 3. Biochem Biophys Res Commun, 2011. 413(4): p. 555-60.

149. Gahbauer, S. and R.A. Bockmann, Membrane-Mediated Oligomerization of G Protein Coupled Receptors and Its Implications for GPCR Function. Front Physiol, 2016. 7: p. 494.

150. Goddard, A.D., et al., Lipid-dependent GPCR dimerization. Methods Cell Biol, 2013. 117: p. 341- 57.

151. Prasanna, X., A. Chattopadhyay, and D. Sengupta, Cholesterol modulates the dimer interface of the beta(2)-adrenergic receptor via cholesterol occupancy sites. Biophys J, 2014. 106(6): p. 1290- 300.

152. Wu, H., et al., Structure of a class C GPCR metabotropic glutamate receptor 1 bound to an allosteric modulator. Science, 2014. 344(6179): p. 58-64.

153. Mitchell, D.C., S.L. Niu, and B.J. Litman, Enhancement of G protein-coupled signaling by DHA phospholipids. Lipids, 2003. 38(4): p. 437-43.

154. Jiang, M. and N.S. Bajpayee, Molecular mechanisms of go signaling. Neurosignals, 2009. 17(1): p. 23-41.

145

155. Siehler, S., Regulation of RhoGEF proteins by G12/13-coupled receptors. Br J Pharmacol, 2009. 158(1): p. 41-9.

156. Lee, M., et al., P2Y5 is a G(alpha)i, G(alpha)12/13 G protein-coupled receptor activated by lysophosphatidic acid that reduces intestinal cell adhesion. Am J Physiol Gastrointest Liver Physiol, 2009. 297(4): p. G641-54.

157. Lee, C.W., et al., LPA(4)/GPR23 is a lysophosphatidic acid (LPA) receptor utilizing G(s)-, G(q)/G(i)- mediated calcium signaling and G(12/13)-mediated Rho activation. J Biol Chem, 2007. 282(7): p. 4310-7.

158. Lee, C.W., et al., GPR92 as a new G12/13- and Gq-coupled lysophosphatidic acid receptor that increases cAMP, LPA5. J Biol Chem, 2006. 281(33): p. 23589-97.

159. Grzelinski, M., et al., Critical role of G(alpha)12 and G(alpha)13 for human small cell lung cancer cell proliferation in vitro and tumor growth in vivo. Clinical cancer research : an official journal of the American Association for Cancer Research, 2010. 16(5): p. 1402-1415.

160. Liu, S.-C., et al., G(alpha)12-mediated pathway promotes invasiveness of nasopharyngeal carcinoma by modulating actin cytoskeleton reorganization. Cancer research, 2009. 69(15): p. 6122-6130.

161. Shan, D., et al., The G protein G alpha(13) is required for growth factor-induced cell migration. Developmental cell, 2006. 10(6): p. 707-718.

162. Yagi, H., et al., A synthetic biology approach reveals a CXCR4-G13-Rho signaling axis driving transendothelial migration of metastatic breast cancer cells. Science signaling, 2011. 4(191).

163. Iguchi, T., et al., Orphan G protein-coupled receptor GPR56 regulates neural progenitor cell migration via a G alpha 12/13 and Rho pathway. J Biol Chem, 2008. 283(21): p. 14469-78.

164. Rieken, S., et al., Lysophospholipids control integrin-dependent adhesion in splenic B cells through G(i) and G(12)/G(13) family G-proteins but not through G(q)/G(11). J Biol Chem, 2006. 281(48): p. 36985-92.

165. Lin, F., et al., Essential roles of G{alpha}12/13 signaling in distinct cell behaviors driving zebrafish convergence and extension gastrulation movements. The Journal of cell biology, 2005. 169(5): p. 777-787.

166. Kishor Kumar, S., et al., G13 Controls Angiogenesis through Regulation of VEGFR-2 Expression. Developmental Cell, 2013.

146

167. Kveberg, L., et al., Sphingosine 1 phosphate induces the chemotaxis of human natural killer cells. Role for heterotrimeric G proteins and phosphoinositide 3 kinases. Eur J Immunol, 2002. 32(7): p. 1856-64.

168. Liu, J., et al., G-protein alpha-s and -12 subunits are involved in androgen-stimulated PI3K activation and androgen receptor transactivation in prostate cancer cells. Prostate, 2011. 71(12): p. 1276-86.

169. Shi, C.S. and J.H. Kehrl, PYK2 links G(q)alpha and G(13)alpha signaling to NF-kappa B activation. J Biol Chem, 2001. 276(34): p. 31845-50.

170. Wu, E.H., B.H. Tam, and Y.H. Wong, Constitutively active alpha subunits of G(q/11) and G(12/13) families inhibit activation of the pro-survival Akt signaling cascade. FEBS J, 2006. 273(11): p. 2388- 98.

171. Ridley, A.J. and A. Hall, The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell, 1992. 70(3): p. 389-99.

172. Aittaleb, M., C.A. Boguth, and J.J. Tesmer, Structure and function of heterotrimeric G protein- regulated Rho guanine nucleotide exchange factors. Mol Pharmacol, 2010. 77(2): p. 111-25.

173. Hart, M.J., et al., Direct stimulation of the guanine nucleotide exchange activity of p115 RhoGEF by Galpha13. Science, 1998. 280(5372): p. 2112-4.

174. Mao, J., et al., Guanine nucleotide exchange factor GEF115 specifically mediates activation of Rho and serum response factor by the G protein alpha subunit Galpha13. Proc Natl Acad Sci U S A, 1998. 95(22): p. 12973-6.

175. Suzuki, N., et al., Galpha 12 activates Rho GTPase through tyrosine-phosphorylated leukemia- associated RhoGEF. Proc Natl Acad Sci U S A, 2003. 100(2): p. 733-8.

176. Cancelas, J.A. and D.A. Williams, Rho GTPases in hematopoietic stem cell functions. Curr Opin Hematol, 2009. 16(4): p. 249-54.

177. Yue, R., et al., Thrombin receptor regulates hematopoiesis and endothelial-to-hematopoietic transition. Dev Cell, 2012. 22(5): p. 1092-100.

178. Weng, Z., et al., A DNA damage and stress inducible G protein-coupled receptor blocks cells in G2/M. Proceedings of the National Academy of Sciences of the United States of America, 1998. 95(21): p. 12334-12339.

147

179. Rikitake, Y., et al., Expression of G2A, a receptor for lysophosphatidylcholine, by macrophages in murine, rabbit, and human atherosclerotic plaques. Arterioscler Thromb Vasc Biol, 2002. 22(12): p. 2049-53.

180. Ding, W.G., et al., Lysophosphatidylcholine enhances I(Ks) currents in cardiac myocytes through activation of G protein, PKC and Rho signaling pathways. J Mol Cell Cardiol, 2011. 50(1): p. 58- 65.

181. Ichijo, Y., et al., Two zebrafish G2A homologs activate multiple intracellular signaling pathways in acidic environment. Biochemical and biophysical research communications, 2016. 469(1): p. 81- 86.

182. Xu, J., et al., Microglia Colonization of Developing Zebrafish Midbrain Is Promoted by Apoptotic Neuron and Lysophosphatidylcholine. Developmental cell, 2016. 38(2): p. 214-222.

183. Radu, C.G., et al., Differential proton sensitivity of related G protein-coupled receptors T cell death- associated gene 8 and G2A expressed in immune cells. Proc Natl Acad Sci U S A, 2005. 102(5): p. 1632-7.

184. Kabarowski, J.H., et al., Lysophosphatidylcholine as a ligand for the immunoregulatory receptor G2A. Science (New York, N.Y.), 2001. 293(5530): p. 702-705.

185. Witte, O.N., et al., Retraction. Science (New York, N.Y.), 2005. 307(5707): p. 206.

186. Wang, L., et al., Lysophosphatidylcholine-induced surface redistribution regulates signaling of the murine G protein-coupled receptor G2A. Molecular biology of the cell, 2005. 16(5): p. 2234-2247.

187. Chen, G., et al., Suppression of HMGB1 release by stearoyl lysophosphatidylcholine:an additional mechanism for its therapeutic effects in experimental sepsis. J Lipid Res, 2005. 46(4): p. 623-7.

188. Hasegawa, H., et al., Lysophosphatidylcholine enhances the suppressive function of human naturally occurring regulatory T cells through TGF-beta production. Biochem Biophys Res Commun, 2011. 415(3): p. 526-31.

189. Radu, C.G., et al., T cell chemotaxis to lysophosphatidylcholine through the G2A receptor. Proc Natl Acad Sci U S A, 2004. 101(1): p. 245-50.

190. Yang, L.V., et al., Gi-independent macrophage chemotaxis to lysophosphatidylcholine via the immunoregulatory GPCR G2A. Blood, 2005. 105(3): p. 1127-34.

191. Rolin, J., Z. Al-Jaderi, and A.A. Maghazachi, Oxidized lipids and lysophosphatidylcholine induce the chemotaxis and intracellular calcium influx in natural killer cells. Immunobiology, 2013. 218(6): p. 875-83.

148

192. Sheikh, A.M., et al., Lysophosphatidylcholine induces glial cell activation: role of rho kinase. Glia, 2009. 57(8): p. 898-907.

193. Qin, Z.X., H.Y. Zhu, and Y.H. Hu, Effects of lysophosphatidylcholine on beta-amyloid-induced neuronal apoptosis. Acta Pharmacol Sin, 2009. 30(4): p. 388-95.

194. Inose, Y., et al., Activated microglia in ischemic stroke penumbra upregulate MCP-1 and CCR2 expression in response to lysophosphatidylcholine derived from adjacent neurons and astrocytes. Neuropathology, 2015. 35(3): p. 209-23.

195. Lin, P. and R.D. Ye, The lysophospholipid receptor G2A activates a specific combination of G proteins and promotes apoptosis. J Biol Chem, 2003. 278(16): p. 14379-86.

196. Murakami, N., et al., G2A is a proton-sensing G-protein-coupled receptor antagonized by lysophosphatidylcholine. J Biol Chem, 2004. 279(41): p. 42484-91.

197. Ludwig, M.G., et al., Proton-sensing G-protein-coupled receptors. Nature, 2003. 425(6953): p. 93- 8.

198. Wang, J.-Q.Q., et al., TDAG8 is a proton-sensing and psychosine-sensitive G-protein-coupled receptor. The Journal of biological chemistry, 2004. 279(44): p. 45626-45633.

199. Tomura, H., et al., Proton-sensing and lysolipid-sensitive G-protein-coupled receptors: a novel type of multi-functional receptors. Cell Signal, 2005. 17(12): p. 1466-76.

200. Yin, H., et al., Lipid G protein-coupled receptor ligand identification using beta-arrestin PathHunter assay. The Journal of biological chemistry, 2009. 284(18): p. 12328-12338.

201. Obinata, H., et al., Identification of 9-hydroxyoctadecadienoic acid and other oxidized free fatty acids as ligands of the G protein-coupled receptor G2A. The Journal of biological chemistry, 2005. 280(49): p. 40676-40683.

202. Hattori, T., et al., G2A plays proinflammatory roles in human keratinocytes under oxidative stress as a receptor for 9-hydroxyoctadecadienoic acid. J Invest Dermatol, 2008. 128(5): p. 1123-33.

203. Hohmann, S.W., et al., The G2A receptor (GPR132) contributes to oxaliplatin-induced mechanical pain hypersensitivity. Sci Rep, 2017. 7(1): p. 446.

204. Zohn, I.E., et al., G2A is an oncogenic G protein-coupled receptor. Oncogene, 2000. 19(34): p. 3866-77.

149

205. Kabarowski, J., et al., Direct genetic demonstration of G alpha 13 coupling to the orphan G protein- coupled receptor G2A leading to RhoA-dependent actin rearrangement. Proceedings of the National Academy of Sciences of the United States of America, 2000. 97(22): p. 12109-12114.

206. Earley, S., et al., TRPV4 forms a novel Ca2+ signaling complex with ryanodine receptors and BKCa channels. Circ Res, 2005. 97(12): p. 1270-9.

207. Fernandes, J., et al., IP3 sensitizes TRPV4 channel to the mechano- and osmotransducing messenger 5'-6'-epoxyeicosatrienoic acid. J Cell Biol, 2008. 181(1): p. 143-55.

208. Loot, A.E., et al., Role of cytochrome P450-dependent transient receptor potential V4 activation in flow-induced vasodilatation. Cardiovasc Res, 2008. 80(3): p. 445-52.

209. Sakuta, H., et al., Nax signaling evoked by an increase in [Na+] in CSF induces water intake via EET-mediated TRPV4 activation. Am J Physiol Regul Integr Comp Physiol, 2016. 311(2): p. R299- 306.

210. Vriens, J., et al., Modulation of the Ca2 permeable cation channel TRPV4 by cytochrome P450 epoxygenases in vascular endothelium. Circ Res, 2005. 97(9): p. 908-15.

211. Li, P., et al., Epoxyeicosatrienoic acids enhance embryonic haematopoiesis and adult marrow engraftment. Nature, 2015. 523(7561): p. 468-471.

212. Stachura, D.L., et al., Clonal analysis of hematopoietic progenitor cells in the zebrafish. Blood, 2011. 118(5): p. 1274-82.

213. Traver, D., et al., Transplantation and in vivo imaging of multilineage engraftment in zebrafish bloodless mutants. Nat Immunol, 2003. 4(12): p. 1238-46.

214. Traver, D., et al., Effects of lethal irradiation in zebrafish and rescue by hematopoietic cell transplantation. Blood, 2004. 104(5): p. 1298-305.

215. White, R.M., et al., Transparent adult zebrafish as a tool for in vivo transplantation analysis. Cell Stem Cell, 2008. 2(2): p. 183-9.

216. Blake, A., et al., Recombinant constructs and transgenic fluorescent ornamental fish therefrom. . US Patent, 2010. US7,700,825 B2.

217. Forsberg, E.C., et al., Molecular signatures of quiescent, mobilized and leukemia-initiating hematopoietic stem cells. PLoS One, 2010. 5(1): p. e8785.

150

218. Pfister, S.L., K.M. Gauthier, and W.B. Campbell, Vascular pharmacology of epoxyeicosatrienoic acids. Adv Pharmacol, 2010. 60: p. 27-59.

219. Lam, E.Y., et al., Live imaging of Runx1 expression in the dorsal aorta tracks the emergence of blood progenitors from endothelial cells. Blood, 2010. 116(6): p. 909-914.

220. Renaud, S.J., et al., The FOS transcription factor family differentially controls trophoblast migration and invasion. J Biol Chem, 2014. 289(8): p. 5025-39.

221. Gilan, O., et al., PR55alpha-containing protein phosphatase 2A complexes promote cancer cell migration and invasion through regulation of AP-1 transcriptional activity. Oncogene, 2015. 34(10): p. 1333-9.

222. Yang, W., et al., Characterization of epoxyeicosatrienoic acid binding site in U937 membranes using a novel radiolabeled agonist, 20-125i-14,15-epoxyeicosa-8(Z)-enoic acid. J Pharmacol Exp Ther, 2008. 324(3): p. 1019-27.

223. Slusarski, D.C., V.G. Corces, and R.T. Moon, Interaction of Wnt and a Frizzled homologue triggers G-protein-linked phosphatidylinositol signalling. Nature, 1997. 390(6658): p. 410-3.

224. Kikuchi, K., et al., Retinoic acid production by endocardium and epicardium is an injury response essential for zebrafish heart regeneration. Dev Cell, 2011. 20(3): p. 397-404.

225. Ma, D., et al., The identification and characterization of zebrafish hematopoietic stem cells. Blood, 2011. 118(2): p. 289-97.

226. Bee, T., et al., The mouse Runx1 +23 hematopoietic stem cell enhancer confers hematopoietic specificity to both Runx1 promoters. Blood, 2009. 113(21): p. 5121-4.

227. Ikebe, D., et al., Suppression of keratinocyte stratification by a dominant negative JunB mutant without blocking cell proliferation. Genes to cells : devoted to molecular & cellular mechanisms, 2007. 12(2): p. 197-207.

228. Pugach, E.K., et al., Retro-orbital injection in adult zebrafish. J Vis Exp, 2009(34).

229. Meder, B., et al., JunB-CBFbeta signaling is essential to maintain sarcomeric Z-disc structure and when defective leads to heart failure. J Cell Sci, 2010. 123(Pt 15): p. 2613-20.

230. Yoo, S., et al., Differential regulation of protrusion and polarity by PI3K during neutrophil motility in live zebrafish. Developmental cell, 2010. 18(2): p. 226-236.

151

231. Sundstrom, C. and K. Nilsson, Establishment and characterization of a human histiocytic lymphoma cell line (U-937). Int J Cancer, 1976. 17(5): p. 565-77.

232. Lam, B.S., C. Cunningham, and G.B. Adams, Pharmacologic modulation of the calcium-sensing receptor enhances hematopoietic stem cell lodgment in the adult bone marrow. Blood, 2011. 117(4): p. 1167-75.

233. Challen, G.A., et al., Mouse hematopoietic stem cell identification and analysis. Cytometry A, 2009. 75(1): p. 14-24.

234. Choe, S.E., et al., Preferred analysis methods for Affymetrix GeneChips revealed by a wholly defined control dataset. Genome Biol, 2005. 6(2): p. R16.

235. Martin, M., Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.Journal, 2011. 17(1): p. 10-12.

236. Trapnell, C., L. Pachter, and S.L. Salzberg, TopHat: discovering splice junctions with RNA-Seq. Bioinformatics, 2009. 25(9): p. 1105-11.

237. Trapnell, C., et al., Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol, 2010. 28(5): p. 511-5.

238. Kopp, H.G., et al., Functional heterogeneity of the bone marrow vascular niche. Ann N Y Acad Sci, 2009. 1176: p. 47-54.

239. Baron, A., M. Frieden, and J.L. Bény, Epoxyeicosatrienoic acids activate a high-conductance, Ca(2+)-dependent K + channel on pig coronary artery endothelial cells. The Journal of physiology, 1997. 504 ( Pt 3): p. 537-543.

240. Zhang, Y., et al., EET homologs potently dilate coronary microvessels and activate BK(Ca) channels. American journal of physiology. Heart and circulatory physiology, 2001. 280(6): p. 40.

241. Bellien, J. and R. Joannides, Epoxyeicosatrienoic acid pathway in human health and diseases. Journal of cardiovascular pharmacology, 2013. 61(3): p. 188-196.

242. Sudhahar, V., S. Shaw, and J.D. Imig, Epoxyeicosatrienoic acid analogs and vascular function. Curr Med Chem, 2010. 17(12): p. 1181-90.

243. Wolf, G., Fatty acids bind directly to and activate peroxisome proliferator-activated receptors alpha and gamma. Nutr Rev, 1998. 56(2 Pt 1): p. 61-3.

152

244. Wong, P., et al., Post-receptor signal transduction and regulation of 14(R),15(S)- epoxyeicosatrienoic acid (14,15-EET) binding in U-937 cells. Journal of lipid mediators and cell signalling, 1997. 16(3): p. 155-169.

245. Wong, P.Y., P.S. Lai, and J.R. Falck, Mechanism and signal transduction of 14 (R), 15 (S)- epoxyeicosatrienoic acid (14,15-EET) binding in guinea pig monocytes. Prostaglandins & other lipid mediators, 2000. 62(4): p. 321-333.

246. Le, L.Q., et al., Mice lacking the orphan G protein-coupled receptor G2A develop a late-onset autoimmune syndrome. Immunity, 2001. 14(5): p. 561-571.

247. Garcia, V., et al., 20-HETE Signals Through G Protein-Coupled Receptor GPR75 (Gq) to Affect Vascular Function and Trigger Hypertension. Circ Res, 2017.

248. Chen, P., et al., Gpr132 sensing of lactate mediates tumor-macrophage interplay to promote breast cancer metastasis. Proc Natl Acad Sci U S A, 2017. 114(3): p. 580-585.

249. Kabarowski, J.H., G2A and LPC: regulatory functions in immunity. Prostaglandins & other lipid mediators, 2009. 89(3-4): p. 73-81.

250. Meyer zu Heringdorf, D. and K.H. Jakobs, Lysophospholipid receptors: signalling, pharmacology and regulation by lysophospholipid metabolism. Biochim Biophys Acta, 2007. 1768(4): p. 923-40.

251. Seuwen, K., M.G. Ludwig, and R.M. Wolf, Receptors for protons or lipid messengers or both? J Recept Signal Transduct Res, 2006. 26(5-6): p. 599-610.

252. Loot, A.E., et al., 11,12-EET stimulates the association of BK channel alpha and beta(1) subunits in mitochondria to induce pulmonary vasoconstriction. PLoS One, 2012. 7(9): p. e46065.

253. Spector, A.A., K. John, and J.E. Fletcher, Binding of long-chain fatty acids to bovine serum albumin. J Lipid Res, 1969. 10(1): p. 56-67.

254. Dai, M., et al., Epoxyeicosatrienoic acids regulate macrophage polarization and prevent LPS- induced cardiac dysfunction. J Cell Physiol, 2015. 230(9): p. 2108-19.

255. Huang, Y.H., et al., Heteromerization of G2A and OGR1 enhances proton sensitivity and proton- induced calcium signals. J Recept Signal Transduct Res, 2016. 36(6): p. 633-644.

256. Shapiro, H., P. Singer, and A. Ariel, Beyond the classic eicosanoids: Peripherally-acting oxygenated metabolites of polyunsaturated fatty acids mediate pain associated with tissue injury and inflammation. Prostaglandins Leukot Essent Fatty Acids, 2016. 111: p. 45-61.

153

257. Back, M., Leukotriene receptors: crucial components in vascular inflammation. ScientificWorldJournal, 2007. 7: p. 1422-39.

258. Massoumi, R. and A. Sjolander, The role of leukotriene receptor signaling in inflammation and cancer. ScientificWorldJournal, 2007. 7: p. 1413-21.

259. Lauder, S.N., et al., Myeloid 12/15-LOX regulates B cell numbers and innate immune antibody levels in vivo. Wellcome Open Res, 2017. 2: p. 1.

260. Liu, X., P. Zhu, and B.D. Freedman, Multiple eicosanoid-activated nonselective cation channels regulate B-lymphocyte adhesion to integrin ligands. Am J Physiol Cell Physiol, 2006. 290(3): p. C873-82.

261. Lamoureux, J., J. Stankova, and M. Rola-Pleszczynski, Leukotriene D4 enhances immunoglobulin production in CD40-activated human B lymphocytes. J Allergy Clin Immunol, 2006. 117(4): p. 924- 30.

262. Yamaoka, K.A., H.E. Claesson, and A. Rosen, Leukotriene B4 enhances activation, proliferation, and differentiation of human B lymphocytes. J Immunol, 1989. 143(6): p. 1996-2000.

263. Yang, X.Y., et al., Interleukin (IL)-4 indirectly suppresses IL-2 production by human T lymphocytes via peroxisome proliferator-activated receptor gamma activated by macrophage-derived 12/15- lipoxygenase ligands. J Biol Chem, 2002. 277(6): p. 3973-8.

264. Luster, A.D. and A.M. Tager, T-cell trafficking in asthma: lipid mediators grease the way. Nat Rev Immunol, 2004. 4(9): p. 711-24.

265. Goodwin, J.S., Regulation of T cell activation by leukotriene B4. Immunol Res, 1986. 5(3): p. 233- 48.

266. Thisse, C. and B. Thisse, High-resolution in situ hybridization to whole-mount zebrafish embryos. Nat Protoc, 2008. 3(1): p. 59-69.

267. Gao, X., et al., GATA Factor-G-Protein-Coupled Receptor Circuit Suppresses Hematopoiesis. Stem Cell Reports, 2016. 6(3): p. 368-82.

268. Zaslavsky, A., et al., Homo- and hetero-dimerization of LPA/S1P receptors, OGR1 and GPR4. Biochim Biophys Acta, 2006. 1761(10): p. 1200-12.

269. Dzierzak, E. and N.A. Speck, Of lineage and legacy: the development of mammalian hematopoietic stem cells. Nat Immunol, 2008. 9(2): p. 129-36.

154

270. Beck, D., et al., Genome-wide analysis of transcriptional regulators in human HSPCs reveals a densely interconnected network of coding and noncoding genes. Blood, 2013. 122(14): p. e12-22.

271. Fujiwara, T., et al., Discovering hematopoietic mechanisms through genome-wide analysis of GATA factor chromatin occupancy. Mol Cell, 2009. 36(4): p. 667-81.

272. May, G., et al., Dynamic analysis of gene expression and genome-wide transcription factor binding during lineage specification of multipotent progenitors. Cell Stem Cell, 2013. 13(6): p. 754-68.

273. Tripic, T., et al., SCL and associated proteins distinguish active from repressive GATA transcription factor complexes. Blood, 2009. 113(10): p. 2191-201.

274. Wilson, N.K., et al., Combinatorial transcriptional control in blood stem/progenitor cells: genome- wide analysis of ten major transcriptional regulators. Cell Stem Cell, 2010. 7(4): p. 532-44.

275. Wozniak, R.J., et al., Molecular hallmarks of endogenous chromatin complexes containing master regulators of hematopoiesis. Mol Cell Biol, 2008. 28(21): p. 6681-94.

276. Yu, M., et al., Insights into GATA-1-mediated gene activation versus repression via genome-wide chromatin occupancy analysis. Mol Cell, 2009. 36(4): p. 682-95.

277. Tsai, F.Y., et al., An early haematopoietic defect in mice lacking the transcription factor GATA-2. Nature, 1994. 371(6494): p. 221-6.

278. de Pater, E., et al., Gata2 is required for HSC generation and survival. J Exp Med, 2013. 210(13): p. 2843-50.

279. Gao, X., et al., Gata2 cis-element is required for hematopoietic stem cell generation in the mammalian embryo. J Exp Med, 2013. 210(13): p. 2833-42.

280. Johnson, K.D., et al., Cis-element mutated in GATA2-dependent immunodeficiency governs hematopoiesis and vascular integrity. J Clin Invest, 2012. 122(10): p. 3692-704.

281. Ling, K.W., et al., GATA-2 plays two functionally distinct roles during the ontogeny of hematopoietic stem cells. J Exp Med, 2004. 200(7): p. 871-82.

282. Rodrigues, N.P., et al., Haploinsufficiency of GATA-2 perturbs adult hematopoietic stem-cell homeostasis. Blood, 2005. 106(2): p. 477-84.

283. Johnson, K.D., et al., Cis-regulatory mechanisms governing stem and progenitor cell transitions. Sci Adv, 2015. 1(8): p. e1500503.

155

284. Hsu, A.P., et al., GATA2 haploinsufficiency caused by mutations in a conserved intronic element leads to MonoMAC syndrome. Blood, 2013. 121(19): p. 3830-7, s1-7.

285. Grass, J.A., et al., Distinct functions of dispersed GATA factor complexes at an endogenous gene locus. Mol Cell Biol, 2006. 26(19): p. 7056-67.

286. Lugus, J.J., et al., GATA2 functions at multiple steps in hemangioblast development and differentiation. Development, 2007. 134(2): p. 393-405.

287. Maeno, M., et al., The role of BMP-4 and GATA-2 in the induction and differentiation of hematopoietic mesoderm in Xenopus laevis. Blood, 1996. 88(6): p. 1965-72.

288. Guiu, J., et al., Hes repressors are essential regulators of hematopoietic stem cell development downstream of Notch signaling. J Exp Med, 2013. 210(1): p. 71-84.

289. Robert-Moreno, A., et al., Impaired embryonic haematopoiesis yet normal arterial development in the absence of the Notch ligand Jagged1. Embo j, 2008. 27(13): p. 1886-95.

290. Liu, F., et al., Induction of hematopoietic and endothelial cell program orchestrated by ETS transcription factor ER71/ETV2. EMBO Rep, 2015. 16(5): p. 654-69.

291. Li, C., et al., Overlapping Requirements for Tet2 and Tet3 in Normal Development and Hematopoietic Stem Cell Emergence. Cell Rep, 2015. 12(7): p. 1133-43.

292. Roth, B.L. and W.K. Kroeze, Integrated Approaches for Genome-wide Interrogation of the Druggable Non-olfactory G Protein-coupled Receptor Superfamily. J Biol Chem, 2015. 290(32): p. 19471-7.

293. Broxmeyer, H.E., et al., Rapid mobilization of murine and human hematopoietic stem and progenitor cells with AMD3100, a CXCR4 antagonist. J Exp Med, 2005. 201(8): p. 1307-18.

294. Sugimoto, Y. and S. Narumiya, Prostaglandin E receptors. J Biol Chem, 2007. 282(16): p. 11613- 7.

295. Hoggatt, J., et al., Differential stem- and progenitor-cell trafficking by prostaglandin E2. Nature, 2013. 495(7441): p. 365-9.

296. Solaimani Kartalaei, P., et al., Whole-transcriptome analysis of endothelial to hematopoietic stem cell transition reveals a requirement for Gpr56 in HSC generation. J Exp Med, 2015. 212(1): p. 93- 106.

156

297. Saito, Y., et al., Maintenance of the hematopoietic stem cell pool in bone marrow niches by EVI1- regulated GPR56. Leukemia, 2013. 27(8): p. 1637-49.

298. Adams, G.B., et al., Haematopoietic stem cells depend on Galpha(s)-mediated signalling to engraft bone marrow. Nature, 2009. 459(7243): p. 103-7.

299. DeVilbiss, A.W., M.E. Boyer, and E.H. Bresnick, Establishing a hematopoietic genetic network through locus-specific integration of chromatin regulators. Proc Natl Acad Sci U S A, 2013. 110(36): p. E3398-407.

300. Li, L., et al., Nuclear adaptor Ldb1 regulates a transcriptional program essential for the maintenance of hematopoietic stem cells. Nat Immunol, 2011. 12(2): p. 129-36.

301. Trompouki, E., et al., Lineage regulators direct BMP and Wnt pathways to cell-specific programs during differentiation and regeneration. Cell, 2011. 147(3): p. 577-89.

302. Kingsley, P.D., et al., Ontogeny of erythroid gene expression. Blood, 2013. 121(6): p. e5-e13.

303. Katsumura, K.R., et al., Molecular basis of crosstalk between oncogenic Ras and the master regulator of hematopoiesis GATA-2. EMBO Rep, 2014. 15(9): p. 938-47.

304. Im, D.S., et al., Identification of a molecular target of psychosine and its role in globoid cell formation. J Cell Biol, 2001. 153(2): p. 429-34.

305. Tobo, A., et al., Characterization of Imidazopyridine Compounds as Negative Allosteric Modulators of Proton-Sensing GPR4 in Extracellular Acidification-Induced Responses. PLoS One, 2015. 10(6): p. e0129334.

306. Butko, E., et al., Gata2b is a restricted early regulator of hemogenic endothelium in the zebrafish embryo. Development, 2015. 142(6): p. 1050-61.

307. McIver, S.C., et al., The exosome complex establishes a barricade to erythroid maturation. Blood, 2014. 124(14): p. 2285-97.

308. Sanalkumar, R., et al., Mechanism governing a stem cell-generating cis-regulatory element. Proc Natl Acad Sci U S A, 2014. 111(12): p. E1091-100.

309. Beck, D.B., et al., PR-Set7 and H4K20me1: at the crossroads of genome integrity, cell cycle, chromosome condensation, and transcription. Genes Dev, 2012. 26(4): p. 325-37.

157

310. DeVilbiss, A.W., et al., Epigenetic Determinants of Erythropoiesis: Role of the Histone Methyltransferase SetD8 in Promoting Erythroid Cell Maturation and Survival. Mol Cell Biol, 2015. 35(12): p. 2073-87.

311. Alon, U., Network motifs: theory and experimental approaches. Nat Rev Genet, 2007. 8(6): p. 450- 61.

312. Shoval, O. and U. Alon, SnapShot: network motifs. Cell, 2010. 143(2): p. 326-e1.

313. Guo, G., et al., Mapping cellular hierarchy by single-cell analysis of the cell surface repertoire. Cell Stem Cell, 2013. 13(4): p. 492-505.

314. Lara-Astiaso, D., et al., Immunogenetics. Chromatin state dynamics during blood formation. Science, 2014. 345(6199): p. 943-9.

315. Li, Y., et al., Inflammatory signaling regulates embryonic hematopoietic stem and progenitor cell production. Genes Dev, 2014. 28(23): p. 2597-612.

316. Moignard, V., et al., Characterization of transcriptional networks in blood stem and progenitor cells using high-throughput single-cell gene expression analysis. Nat Cell Biol, 2013. 15(4): p. 363-72.

317. Cheng, Y., et al., Erythroid GATA1 function revealed by genome-wide analysis of transcription factor occupancy, histone modifications, and mRNA expression. Genome Res, 2009. 19(12): p. 2172-84.

318. Welch, J.J., et al., Global regulation of erythroid gene expression by transcription factor GATA-1. Blood, 2004. 104(10): p. 3136-47.

319. Grass, J.A., et al., GATA-1-dependent transcriptional repression of GATA-2 via disruption of positive autoregulation and domain-wide chromatin remodeling. Proc Natl Acad Sci U S A, 2003. 100(15): p. 8811-6.

320. Johnson, K.D., et al., Friend of GATA-1-independent transcriptional repression: a novel mode of GATA-1 function. Blood, 2007. 109(12): p. 5230-3.

321. Munugalavadla, V., et al., Repression of c-kit and its downstream substrates by GATA-1 inhibits cell proliferation during erythroid maturation. Mol Cell Biol, 2005. 25(15): p. 6747-59.

322. Jing, H., et al., Exchange of GATA factors mediates transitions in looped chromatin organization at a developmentally regulated gene locus. Mol Cell, 2008. 29(2): p. 232-42.

158

323. Huang, E., et al., The hematopoietic growth factor KL is encoded by the Sl locus and is the ligand of the c-kit receptor, the gene product of the W locus. Cell, 1990. 63(1): p. 225-33.

324. Heissig, B., et al., Recruitment of stem and progenitor cells from the bone marrow niche requires MMP-9 mediated release of kit-ligand. Cell, 2002. 109(5): p. 625-37.

325. Mangan, S. and U. Alon, Structure and function of the feed-forward loop network motif. Proc Natl Acad Sci U S A, 2003. 100(21): p. 11980-5.

326. Choi, J.W., S.Y. Lee, and Y. Choi, Identification of a putative G protein-coupled receptor induced during activation-induced apoptosis of T cells. Cell Immunol, 1996. 168(1): p. 78-84.

327. Malone, M.H., Z. Wang, and C.W. Distelhorst, The glucocorticoid-induced gene tdag8 encodes a pro-apoptotic G protein-coupled receptor whose activation promotes glucocorticoid-induced apoptosis. J Biol Chem, 2004. 279(51): p. 52850-9.

328. Mogi, C., et al., Involvement of proton-sensing TDAG8 in extracellular acidification-induced inhibition of proinflammatory cytokine production in peritoneal macrophages. J Immunol, 2009. 182(5): p. 3243-51.

329. Radu, C.G., et al., Normal immune development and glucocorticoid-induced thymocyte apoptosis in mice deficient for the T-cell death-associated gene 8 receptor. Mol Cell Biol, 2006. 26(2): p. 668- 77.

330. Chiang, N., et al., Identification of resolvin D2 receptor mediating resolution of infections and organ protection. The Journal of experimental medicine, 2015. 212(8): p. 1203-1217.

331. Sexton, P.M., et al., Complexing receptor pharmacology: modulation of family B G protein-coupled receptor function by RAMPs. Ann N Y Acad Sci, 2006. 1070: p. 90-104.

332. Pritchard, K.A., Jr., et al., 14, 15-Epoxyeicosatrienoic acid promotes endothelial cell dependent adhesion of human monocytic tumor U937 cells. Biochem Biophys Res Commun, 1990. 167(1): p. 137-42.

333. Regard, J.B., I.T. Sato, and S.R. Coughlin, Anatomical profiling of G protein-coupled receptor expression. Cell, 2008. 135(3): p. 561-71.

334. Rossi, A., et al., Genetic compensation induced by deleterious mutations but not gene knockdowns. Nature, 2015. 524(7564): p. 230-3.

335. Kiriyama, M., et al., Ligand binding specificities of the eight types and subtypes of the mouse prostanoid receptors expressed in Chinese hamster ovary cells. Br J Pharmacol, 1997. 122(2): p. 217-24.

159

336. Fritsch, R., et al., RAS and RHO families of GTPases directly regulate distinct phosphoinositide 3- kinase isoforms. Cell, 2013. 153(5): p. 1050-63.

337. Yang, H.W., et al., Cooperative activation of PI3K by Ras and Rho family small GTPases. Mol Cell, 2012. 47(2): p. 281-90.

338. Wang, K., et al., Overexpression of SDF-1alpha enhanced migration and engraftment of cardiac stem cells and reduced infarcted size via CXCR4/PI3K pathway. PLoS One, 2012. 7(9): p. e43922.

339. Yu, J., et al., SDF-1/CXCR4-mediated migration of transplanted bone marrow stromal cells toward areas of heart myocardial infarction through activation of PI3K/Akt. J Cardiovasc Pharmacol, 2010. 55(5): p. 496-505.

340. Zheng, H., et al., Migration of endothelial progenitor cells mediated by stromal cell-derived factor- 1alpha/CXCR4 via PI3K/Akt/eNOS signal transduction pathway. J Cardiovasc Pharmacol, 2007. 50(3): p. 274-80.

341. Buitenhuis, M., et al., Protein kinase B (PKB/c-akt) regulates homing of hematopoietic progenitors through modulation of their adhesive and migratory properties. Blood, 2010. 116(13): p. 2373-84.

342. Desponts, C., et al., SHIP deficiency enhances HSC proliferation and survival but compromises homing and repopulation. Blood, 2006. 107(11): p. 4338-45.

343. Buitenhuis, M., The role of PI3K/protein kinase B (PKB/c-akt) in migration and homing of hematopoietic stem and progenitor cells. Curr Opin Hematol, 2011. 18(4): p. 226-30.

344. Roberts, M.S., et al., Protein kinase B/Akt acts via glycogen synthase kinase 3 to regulate recycling of alpha v beta 3 and alpha 5 beta 1 integrins. Mol Cell Biol, 2004. 24(4): p. 1505-15.

345. Vesely, P.W., et al., Translational regulation mechanisms of AP-1 proteins. Mutat Res, 2009. 682(1): p. 7-12.

346. Lee, S.Y., et al., The role of heterodimeric AP-1 protein comprised of JunD and c-Fos proteins in hematopoiesis. J Biol Chem, 2012. 287(37): p. 31342-8.

347. Ramachandran, A., et al., An Akt- and Fra-1-dependent pathway mediates platelet-derived growth factor-induced expression of thrombomodulin, a novel regulator of smooth muscle cell migration. Am J Pathol, 2010. 177(1): p. 119-31.

348. Wang, J., et al., Sustained c-Jun-NH2-kinase activity promotes epithelial-mesenchymal transition, invasion, and survival of breast cancer cells by regulating extracellular signal-regulated kinase activation. Mol Cancer Res, 2010. 8(2): p. 266-77.

160

349. Shi, Z.D., et al., Interstitial flow induces MMP-1 expression and vascular SMC migration in collagen I gels via an ERK1/2-dependent and c-Jun-mediated mechanism. Am J Physiol Heart Circ Physiol, 2010. 298(1): p. H127-35.

350. Adhikary, S., et al., Signaling through cannabinoid receptor 2 suppresses murine dendritic cell migration by inhibiting matrix metalloproteinase 9 expression. Blood, 2012. 120(18): p. 3741-9.

351. Singh, N.K., et al., AP-1 (Fra-1/c-Jun)-mediated induction of expression of matrix metalloproteinase-2 is required for 15S-hydroxyeicosatetraenoic acid-induced angiogenesis. J Biol Chem, 2010. 285(22): p. 16830-43.

352. Theodore, L.N., et al., Distinct Roles for Matrix Metalloproteinases 2 and 9 in Embryonic Hematopoietic Stem Cell Emergence, Migration, and Niche Colonization. Stem Cell Reports, 2017.

353. Santaguida, M., et al., JunB protects against myeloid malignancies by limiting hematopoietic stem cell proliferation and differentiation without affecting self-renewal. Cancer Cell, 2009. 15(4): p. 341- 52.

354. Passegue, E., E.F. Wagner, and I.L. Weissman, JunB deficiency leads to a myeloproliferative disorder arising from hematopoietic stem cells. Cell, 2004. 119(3): p. 431-43.

355. Boyer, S.B. and P.A. Slesinger, Probing novel GPCR interactions using a combination of FRET and TIRF. Commun Integr Biol, 2010. 3(4): p. 343-6.

356. Cottet, M., et al., Fluorescent ligands to investigate GPCR binding properties and oligomerization. Biochem Soc Trans, 2013. 41(1): p. 148-53.

357. Sadok, A. and C.J. Marshall, Rho GTPases: masters of cell migration. Small GTPases, 2014. 5: p. e29710.

358. Kimura, K., et al., Regulation of myosin phosphatase by Rho and Rho-associated kinase (Rho- kinase). Science, 1996. 273(5272): p. 245-8.

359. Hanna, S. and M. El-Sibai, Signaling networks of Rho GTPases in cell motility. Cell Signal, 2013. 25(10): p. 1955-61.

360. Tan, W., D. Martin, and J.S. Gutkind, The Galpha13-Rho signaling axis is required for SDF-1- induced migration through CXCR4. J Biol Chem, 2006. 281(51): p. 39542-9.

361. Kardash, E., et al., A role for Rho GTPases and cell-cell adhesion in single-cell motility in vivo. Nat Cell Biol, 2010. 12(1): p. 47-53; sup pp 1-11.

161

362. Sahai, E. and C.J. Marshall, RHO-GTPases and cancer. Nat Rev Cancer, 2002. 2(2): p. 133-42.

363. Yoshizaki, H., et al., Activity of Rho-family GTPases during cell division as visualized with FRET- based probes. J Cell Biol, 2003. 162(2): p. 223-32.

364. Mizoguchi, T., K. Kawakami, and M. Itoh, Zebrafish lines expressing UAS-driven red probes for monitoring cytoskeletal dynamics. Genesis, 2016. 54(9): p. 483-9.

365. Ma, S., et al., Development of second-generation small-molecule RhoA inhibitors with enhanced water solubility, tissue potency, and significant in vivo efficacy. ChemMedChem, 2015. 10(1): p. 193-206.

366. Morckel, A.R., et al., A photoactivatable small-molecule inhibitor for light-controlled spatiotemporal regulation of Rho kinase in live embryos. Development, 2012. 139(2): p. 437-42.

367. Shang, X., et al., Small-molecule inhibitors targeting G-protein-coupled Rho guanine nucleotide exchange factors. Proc Natl Acad Sci U S A, 2013. 110(8): p. 3155-60.

368. Shang, X., et al., Rational design of small molecule inhibitors targeting RhoA subfamily Rho GTPases. Chem Biol, 2012. 19(6): p. 699-710.

369. Nagumo, H., et al., Rho kinase inhibitor HA-1077 prevents Rho-mediated myosin phosphatase inhibition in smooth muscle cells. Am J Physiol Cell Physiol, 2000. 278(1): p. C57-65.

370. Zhu, S. and B.C. Low, Using zebrafish for studying Rho GTPases signaling in vivo. Methods Mol Biol, 2012. 827: p. 321-37.

371. Hanovice, N.J., E. McMains, and J.M. Gross, A GAL4-inducible transgenic tool kit for the in vivo modulation of Rho GTPase activity in zebrafish. Dev Dyn, 2016. 245(8): p. 844-53.

372. Yue, F., et al., A comparative encyclopedia of DNA elements in the mouse genome. Nature, 2014. 515(7527): p. 355-64.

373. The Genotype-Tissue Expression (GTEx) project. Nat Genet, 2013. 45(6): p. 580-5.

374. An integrated encyclopedia of DNA elements in the human genome. Nature, 2012. 489(7414): p. 57-74.

375. Riether, C., C.M. Schurch, and A.F. Ochsenbein, Regulation of hematopoietic and leukemic stem cells by the immune system. Cell Death Differ, 2015. 22(2): p. 187-98.

162

376. Ablain, J., et al., A CRISPR/Cas9 vector system for tissue-specific gene disruption in zebrafish. Developmental cell, 2015. 32(6): p. 756-764.

377. Ding, L., et al., Endothelial and perivascular cells maintain haematopoietic stem cells. Nature, 2012. 481(7382): p. 457-62.

378. Lu, T., et al., Dihydroxyeicosatrienoic acids are potent activators of Ca(2+)-activated K(+) channels in isolated rat coronary arterial myocytes. J Physiol, 2001. 534(Pt 3): p. 651-67.

379. Gijon, M.A., S. Zarini, and R.C. Murphy, Biosynthesis of eicosanoids and transcellular metabolism of leukotrienes in murine bone marrow cells. J Lipid Res, 2007. 48(3): p. 716-25.

380. Lindgren, J.A., et al., Formation and effects of leukotrienes and lipoxins in human bone marrow. J Lipid Mediat, 1993. 6(1-3): p. 313-20.

381. Stenke, L., et al., Formation and proliferative effects of lipoxins in human bone marrow. Biochem Biophys Res Commun, 1991. 180(1): p. 255-61.

382. Stenke, L., et al., Leukotriene production by fresh human bone marrow cells: evidence of altered lipoxygenase activity in chronic myelocytic leukemia. Exp Hematol, 1987. 15(2): p. 203-7.

383. Poulsen, R.C., et al., Identification of inflammatory and proresolving lipid mediators in bone marrow and their lipidomic profiles with ovariectomy and omega-3 intake. Am J Hematol, 2008. 83(6): p. 437-45.

384. Han, X. and R.W. Gross, Global analyses of cellular lipidomes directly from crude extracts of biological samples by ESI mass spectrometry: a bridge to lipidomics. J Lipid Res, 2003. 44(6): p. 1071-9.

385. Wang, M., et al., Novel advances in shotgun lipidomics for biology and medicine. Prog Lipid Res, 2016. 61: p. 83-108.

386. Zhao, Y.Y., X.L. Cheng, and R.C. Lin, Lipidomics applications for discovering biomarkers of diseases in clinical chemistry. Int Rev Cell Mol Biol, 2014. 313: p. 1-26.

387. Hanif, A., et al., Deletion of soluble epoxide hydrolase enhances coronary reactive hyperemia in isolated mouse heart: role of oxylipins and PPARgamma. Am J Physiol Regul Integr Comp Physiol, 2016. 311(4): p. R676-r688.

388. Harris, T.R. and B.D. Hammock, Soluble epoxide hydrolase: gene structure, expression and deletion. Gene, 2013. 526(2): p. 61-74.

163

389. Slice, L.W., J.H. Walsh, and E. Rozengurt, Galpha(13) stimulates Rho-dependent activation of the cyclooxygenase-2 promoter. J Biol Chem, 1999. 274(39): p. 27562-6.

164

Appendix 1: Zebrafish Microarray

From Chapter 2. Total RNA was extracted from 36 hpf zebrafish embryos treated with DMSO or 5 μM 11,12-EET between 24–36 hpf, with 3 biological replicates each and n=25 in each group. Microarray hybridization was performed with the Affymetrix GeneChip Zebrafish Genome Array. Hybridized microarray was background-corrected, normalized and multiple-tested using Goldenspike (http://www2.ccr.buffalo.edu/halfon/spike/) in R/Bioconductor [234]. Genes with q<0.1 by SNR test were considered differentially expressed

Genes with increased expression probe mean.fc q gene hchr zchr Dr.14719.1.A1_at 4.359432 0 ARF1 chr1 chr11 Dr.24471.1.A1_at 2.896713 0.0796 PALLADIN chr10 chr13 Dr.19888.1.S1_at 2.52457 0 VED chr8 chr10 DrAffx.2.81.A1_at 2.114362 0.015873 CYP51A1 chr7 chr19 Dr.12110.1.S1_at 2.097974 0 SC4MOL chr4 chr1 Dr.4938.1.S1_at 1.998925 0.015873 FADS2 chr11 chr25 Dr.22219.1.S1_at 1.997586 0.0796 ACTN2 chr14 chr17 Dr.16634.1.A1_at 1.749673 0.01587 AK128797 chr14 chr5 Dr.20131.2.A1_at 1.516192 0 CLU chr8 chr20 Dr.13651.1.A1_at 1.47763 0.028986 FBP1 chr9 chr5 Dr.16392.1.A1_at 1.413491 0 C6 chr5 chr21 Dr.11572.1.A1_at 1.30334 0.015873 WDR34 chr9 chr8 Dr.2051.1.S1_at 1.158936 0 HMGCS1 chr5 chr12 Dr.25093.1.A1_s_at 1.157838 0.03774 TAR3 chr6 chr20 Dr.24233.1.S1_at 1.148094 0 FN1 chr2 chr1 Dr.9617.1.A1_at 1.094222 0 SOCS3 chr17 chr12 Dr.23811.1.A1_at 1.072796 0.079602 MOGAT1 chr11 chr10 Dr.10130.2.A1_at 0.971631 0 FOSL2 chr2 chr15 Dr.967.1.S1_at 0.967481 0 MMP9 chr20 chr8 Dr.9642.1.A1_at 0.844911 0.015873 ASPN chr9 chr22 Dr.19560.1.S1_at 0.781171 0.028986 INSIG1 chr7 chr7 MS4A4A / Dr.22334.1.S1_at 0.747641 0 chr11 chr4 ms4a17a.11 Dr.19560.1.S2_at 0.720476 0.015873 INSIG1 chr7 chr7 Dr.2487.1.S1_at 0.710377 0.037736 UPB1 chr22 chr8 Dr.10326.1.S1_at 0.704015 0.0796 JUNB chr19 chr1 Dr.1089.1.S1_at 0.701168 0.015873 THBS3 chr19 chr16 Dr.25191.1.S1_at 0.635516 0.01587 IDH1 chr2 chr9 Dr.737.1.A1_at 0.543973 0.037736 JUNB chr1 chr3 Dr.26534.1.A1_at 0.502621 0.079602 INSIG1 chr7 chr7 Dr.7010.1.S1_at 0.475427 0.074713 ELOVL7 chr1 chr8 Dr.24562.1.S1_a_at 0.387242 0.066667 MVP chr16 chr3

165

Genes with decreased expression probe mean.fc q gene hchr zchr Dr.20008.1.S1_at -0.234813 0.074713 SLC25A5 chrX chr10|chr14 AFFX-Dr-acta1-5_at -0.290608 0.037736 ACTA1 chr1 --- Dr.6826.1.S1_at -0.381291 0.08333 COL1A1b chr17 chr12 Dr.20272.1.A1_at -0.395476 0.034722 SNRP70 chr19 chr3 Dr.5531.3.S1_at -0.405781 0.015873 KRT15 chr17 chr19 Dr.12425.1.S1_at -0.415358 0.034722 KRT17 chr17 chr19 Dr.1377.1.A1_at -0.440683 0.015873 COL1A2 chr7 chr19 Dr.24292.5.S1_at -0.467295 0 KRT15 chr17 chr19 Dr.15266.1.A1_at -0.51983 0.015873 CHRNG chr2 chr5 Dr.24487.1.A1_at -0.526189 0.079602 KRT8 chr12 chr10|chr2 Dr.3484.1.S1_at -0.529253 0.07471 sap30bp chr17 chr8 Dr.5439.2.S1_at -0.566989 0.079602 NSFL1C chr20 chr11 Dr.2539.1.A1_at -0.577588 0.01587 wu:fy25c05 chr4 chr7 Dr.24399.1.A1_at -0.579821 0 ARGLU1B chr13 --- Dr.4681.1.A1_at -0.61717 0 TTN chr2 chr9 Dr.23502.1.A1_at -0.781435 0 APOE chr19 chr19 Dr.14514.1.S1_a_at -0.999159 0 RPL13 chr16 chr10 Dr.18173.1.A1_at -1.131191 0.015873 MPP2 chr17 chr3 Dr.17747.1.S1_at -1.366207 0.015873 SCEL chr13 chr9 Dr.9478.4.S1_at -3.077714 0.0796 MPX chr17 chr10 Dr.12749.1.A1_at -3.319542 0.023256 RHCG chr1 chr16

166

Appendix 2: GPCR expression levels

From Chapter 3. Average FPKM expression levels from two biological duplicates for each cell line. For genes with multiple transcripts, gene level expression data was compiled by summing all transcripts.

Gene U937 EaHy PC3M HEK293 Gene U937 EaHy PC3M HEK293 ADCYAP1R1 0.12 0.23 0.00 2.40 GPR68 0.89 0.62 0.41 0.04 ADORA1 0.13 2.55 0.03 0.65 GPR75 0.89 0.24 0.21 1.11 ADORA2A 0.33 1.16 0.03 2.40 GPR78 0.42 0.24 0.00 0.31 ADORA2B 0.38 3.57 13.75 8.59 GPR82 2.03 1.10 0.01 1.89 ADORA3 1.05 0.03 0.01 0.00 GPR83 0.38 0.14 0.03 0.18 ADRA1A 1.23 0.54 0.06 1.00 GPR84 0.42 0.08 0.00 0.00 ADRA1B 0.28 0.16 0.04 0.30 GPR85 0.56 0.33 0.11 0.36 ADRA1D 0.09 0.16 0.03 0.00 GPR87 0.04 0.12 2.67 0.02 ADRA2A 0.05 0.06 0.02 0.01 GPR88 0.06 0.05 0.03 0.08 ADRA2B 0.09 0.05 0.00 0.06 GPR97 0.09 0.07 0.00 0.02 ADRA2C 0.29 0.18 0.04 8.41 GPR98 0.04 0.06 1.19 1.23 ADRB1 0.64 0.92 0.30 0.33 GPRC5A 0.13 16.13 5.78 2.02 ADRB2 8.36 13.69 1.02 0.73 GPRC5B 0.04 4.40 0.92 3.87 ADRB3 0.09 0.05 0.04 0.03 GPRC5C 1.45 0.09 15.34 4.56 AGTR1 0.05 0.08 0.17 0.69 GPRC5D 0.01 0.66 0.00 0.02 AGTR2 0.04 0.03 0.00 0.00 GPRC6A 0.05 0.01 0.01 0.00 APLNR 0.02 0.03 0.03 0.00 GRM1 0.07 0.07 0.04 0.02 AVPR1A 0.26 0.17 0.07 0.22 GRM2 0.08 0.04 0.02 0.08 AVPR1B 0.03 0.05 0.27 0.01 GRM3 0.19 0.04 0.03 0.06 AVPR2 1.35 0.05 0.00 0.00 GRM4 0.29 0.19 1.34 0.23 BAI1 8.05 0.08 0.03 0.72 GRM5 0.02 0.05 0.01 0.00 BAI2 0.10 1.03 1.20 9.36 GRM6 0.62 0.33 0.04 0.51 BAI3 0.05 1.39 0.37 1.44 GRM7 0.05 0.04 0.00 0.00 BDKRB1 0.04 0.03 0.19 0.06 GRM8 0.03 0.05 0.15 0.03 BDKRB2 0.07 0.06 0.61 0.03 GRPR 0.11 0.07 0.88 0.01 BRS3 0.04 0.01 0.00 0.12 HCAR1 1.04 0.34 0.15 1.09 C3AR1 1.70 0.06 0.02 0.02 HCAR2 0.01 0.02 0.04 0.00 C5AR1 2.67 0.30 0.03 0.36 HCAR3 0.01 0.02 0.00 0.00 C5AR2 2.46 0.19 0.00 0.34 HCRTR1 0.35 0.03 0.00 0.19 CALCR 0.02 0.03 0.00 0.02 HCRTR2 0.02 0.03 0.00 0.01 CALCRL 0.04 7.01 0.08 0.36 HRH1 1.38 9.84 1.84 0.23 CASR 0.04 0.05 0.00 0.00 HRH2 4.66 0.22 0.00 0.09 CCKAR 0.05 0.04 0.04 0.00 HRH3 0.10 0.06 0.00 0.00 CCKBR 0.11 0.08 0.03 0.10 HRH4 0.87 0.41 0.23 0.61 CCR1 1.01 0.02 0.00 0.00 HTR1A 0.06 0.04 0.00 0.00 CCR10 0.24 0.70 0.46 2.07 HTR1B 0.03 0.21 0.06 0.01

167

Gene U937 EaHy PC3M HEK293 Gene U937 EaHy PC3M HEK293 CCR2 0.08 0.02 0.00 0.00 HTR1D 0.36 5.54 0.04 0.53 CCR3 0.98 0.03 0.00 0.00 HTR1E 0.04 0.02 0.16 0.00 CCR4 0.06 0.08 0.00 0.00 HTR1F 0.03 0.02 0.02 0.04 CCR5 0.04 0.02 0.00 0.00 HTR2A 0.06 0.03 0.00 0.04 CCR6 0.88 0.35 0.04 0.58 HTR2B 0.28 0.08 0.04 0.18 CCR7 0.11 0.02 0.00 0.02 HTR2C 0.02 0.03 0.00 0.05 CCR8 0.01 0.03 0.00 0.02 HTR4 0.33 0.08 0.00 0.00 CCR9 0.08 0.05 0.00 0.06 HTR5A 0.02 0.03 0.00 0.00 CCRL2 2.48 2.46 1.47 0.05 HTR6 0.22 0.34 0.02 1.47 CD97 11.64 0.51 5.87 9.89 HTR7 0.09 0.49 1.04 0.11 CELSR1 0.16 4.36 0.87 4.92 KISS1R 1.67 0.98 0.31 0.24 CELSR2 0.78 0.20 0.39 6.74 LGR4 0.04 6.17 4.64 8.67 CELSR3 1.06 3.96 0.35 5.94 LGR5 0.01 0.02 0.46 1.54 CHRM1 0.01 0.03 0.00 0.00 LGR6 0.04 0.13 0.00 0.18 CHRM2 0.06 0.09 0.00 0.05 LHCGR 0.05 0.04 0.00 0.00 CHRM3 0.76 1.87 4.26 2.40 LPAR1 0.07 11.55 21.90 1.79 CHRM4 0.09 0.07 0.06 0.08 LPAR2 8.03 2.12 1.87 1.98 CHRM5 0.05 0.02 0.01 0.02 LPAR3 0.06 0.05 16.45 7.23 CMKLR1 0.04 0.30 0.02 0.00 LPAR4 0.57 0.11 0.00 0.16 CNR1 0.04 0.04 0.02 0.04 LPAR5 0.33 0.52 0.12 0.09 CNR2 0.07 0.01 0.00 0.01 LPAR6 0.46 2.64 2.39 0.35 CRHR1 0.12 0.10 0.09 0.80 LPHN1 6.71 2.40 2.28 7.21 CRHR2 0.05 0.04 0.01 0.02 LPHN2 0.03 16.12 12.38 14.82 CX3CR1 1.36 0.05 0.00 0.00 LPHN3 0.03 0.09 0.01 3.70 CXCR1 0.37 0.01 0.05 0.22 LTB4R 14.32 4.00 1.32 4.57 CXCR2 4.17 0.26 0.10 0.09 LTB4R2 9.18 5.62 2.70 4.49 CXCR3 0.97 0.04 0.00 0.00 MAS1 0.61 0.01 0.00 0.00 CXCR4 64.40 0.42 0.86 6.41 MAS1L 0.00 0.01 0.02 0.00 CXCR5 0.06 0.03 0.04 0.04 MC1R 0.99 2.53 0.50 1.70 CXCR6 1.82 0.09 0.05 0.81 MC2R 0.06 0.12 0.01 0.19 CYSLTR1 6.64 0.01 0.00 0.01 MC3R 0.02 0.01 0.00 0.00 CYSLTR2 0.06 0.68 0.00 0.00 MC4R 0.01 0.05 0.03 0.00 DARC 0.05 0.01 0.00 0.00 MC5R 0.16 0.22 0.00 0.00 DRD1 0.02 0.03 0.00 0.03 MCHR1 0.04 0.04 0.00 0.00 DRD2 0.08 0.07 0.03 0.11 MCHR2 0.20 0.13 0.00 0.14 DRD3 0.03 0.02 0.01 0.00 MLNR 0.23 0.15 0.00 0.16 DRD4 0.35 0.28 0.01 1.73 MRGPRD 0.06 0.07 0.00 0.01 DRD5 0.06 0.05 0.06 0.05 MRGPRE 0.01 0.04 0.00 0.00 EDNRA 0.26 0.22 0.66 1.30 MRGPRF 0.21 0.21 0.00 0.00 EDNRB 0.06 2.29 0.00 1.24 MRGPRG 0.03 0.03 0.00 0.00 ELTD1 0.04 0.06 0.00 0.00 MRGPRX1 0.08 0.03 0.00 0.00 EMR1 0.31 5.91 0.06 0.02 MRGPRX2 0.10 0.06 0.00 0.00

168

Gene U937 EaHy PC3M HEK293 Gene U937 EaHy PC3M HEK293 EMR2 5.85 0.12 0.20 0.39 MRGPRX3 0.92 0.72 0.24 0.49 EMR3 0.10 0.03 0.00 0.00 MRGPRX4 1.70 0.02 0.00 1.63 EMR4P 1.44 1.74 0.01 1.65 MTNR1A 0.01 0.06 0.00 0.38 F2R 2.17 38.64 4.97 8.14 MTNR1B 0.01 0.02 0.00 0.00 F2RL1 0.32 61.34 22.02 8.01 NMBR 0.04 0.04 0.03 0.02 F2RL2 0.40 2.53 0.02 0.25 NMUR1 0.56 0.27 0.07 0.67 F2RL3 2.44 0.56 0.04 0.70 NMUR2 0.06 0.06 0.00 0.00 FFAR1 0.06 0.14 0.00 0.00 NPBWR1 0.21 0.06 0.00 0.02 FFAR2 1.09 0.43 0.04 0.38 NPBWR2 0.01 0.06 0.00 0.00 FFAR3 0.02 0.07 0.01 0.00 NPFFR1 0.33 0.21 0.00 0.00 FFAR4 0.62 0.31 0.02 0.47 NPFFR2 0.70 0.21 0.05 0.65 FPR1 0.38 0.08 0.03 0.00 NPSR1 0.01 0.02 0.00 0.00 FPR2 0.32 0.02 0.00 0.00 NPY1R 0.03 0.02 0.19 0.65 FPR3 0.04 0.17 0.00 0.00 NPY2R 0.02 0.04 0.01 0.00 FSHR 0.05 0.06 0.02 0.00 NPY5R 0.13 0.13 0.04 0.84 FZD1 1.90 1.13 0.88 3.24 NPY6R 0.03 0.03 0.01 0.12 FZD10 0.06 0.07 0.02 1.37 NTSR1 0.08 1.48 0.77 0.01 FZD2 5.29 5.14 1.40 5.03 NTSR2 0.06 0.06 0.00 0.00 FZD3 0.57 0.28 0.95 2.55 OPN3 7.46 3.42 12.16 2.56 FZD4 0.57 31.93 2.15 3.12 OPN4 0.05 0.04 0.00 0.00 FZD5 1.22 0.48 0.59 3.03 OPN5 0.04 0.13 0.00 0.00 FZD6 0.04 15.19 8.65 4.99 OPRD1 0.13 0.10 0.04 0.21 FZD7 1.25 0.18 1.68 6.51 OPRK1 0.18 0.15 0.00 0.27 FZD8 0.22 2.75 15.99 3.85 OPRL1 2.30 0.76 0.05 1.05 FZD9 0.28 0.07 0.05 2.70 OPRM1 0.09 0.11 0.02 0.01 GABBR1 2.41 1.77 0.33 8.27 OR51E1 0.02 0.02 0.00 0.00 GABBR2 0.19 0.23 0.09 0.59 OXER1 1.36 0.12 0.00 0.09 GALR1 0.11 0.07 0.08 0.00 OXGR1 0.04 0.01 0.00 0.54 GALR2 0.08 0.64 0.16 0.16 OXTR 1.01 0.61 0.34 0.37 GALR3 2.11 1.32 0.24 0.07 P2RY1 0.07 0.06 6.97 2.96 GCGR 0.28 0.29 0.00 0.02 P2RY10 6.83 0.02 0.00 0.02 GHRHR 0.11 0.15 0.00 0.00 P2RY11 7.00 0.38 0.36 2.72 GHSR 0.09 0.07 0.01 0.00 P2RY12 0.04 0.01 0.00 2.24 GIPR 0.52 0.24 0.10 1.13 P2RY13 0.06 0.03 0.03 0.01 GLP1R 0.14 0.08 0.03 0.03 P2RY14 0.04 0.02 0.02 0.11 GLP2R 0.04 0.04 0.00 0.00 P2RY2 6.99 3.48 0.29 0.37 GNRHR 0.05 0.08 0.00 0.03 P2RY4 0.33 0.04 0.00 0.08 GNRHR2 10.36 5.71 3.67 13.26 P2RY6 6.52 0.26 0.07 0.00 GPBAR1 0.24 0.17 0.19 0.03 P2RY8 35.67 0.23 0.00 0.17 GPR1 0.85 0.49 0.18 0.97 PRLHR 0.02 0.02 0.01 0.00 GPR101 0.01 0.05 0.00 0.00 PROKR1 0.20 0.23 0.00 0.03 GPR107 7.99 11.58 3.33 14.59 PROKR2 0.02 0.03 0.00 0.00

169

Gene U937 EaHy PC3M HEK293 Gene U937 EaHy PC3M HEK293 GPR110 0.75 0.29 12.72 0.52 PTAFR 4.21 0.51 0.14 0.98 GPR111 0.12 0.09 0.00 0.13 PTGDR 0.03 0.03 0.01 0.00 GPR112 0.03 0.04 0.00 0.08 PTGDR2 0.71 1.79 0.06 1.58 GPR113 0.25 0.08 0.11 0.04 PTGER1 20.90 0.36 1.99 0.29 GPR114 0.40 0.27 0.00 0.00 PTGER2 12.10 0.44 1.14 0.28 GPR115 0.12 0.02 0.05 0.01 PTGER3 0.14 0.11 0.01 0.01 GPR116 0.03 0.10 0.20 0.01 PTGER4 10.78 11.27 0.60 0.14 GPR119 0.04 0.04 0.00 0.00 PTGFR 0.02 0.35 0.68 0.76 GPR12 0.34 0.24 0.00 0.32 PTGIR 0.09 1.72 0.00 0.01 GPR123 0.04 0.04 0.02 0.00 PTH1R 0.52 2.69 0.10 0.52 GPR124 70.82 0.39 0.21 4.92 PTH2R 0.02 0.02 0.04 0.00 GPR125 6.32 4.19 3.39 6.78 QRFPR 0.72 0.32 0.04 1.84 GPR126 0.04 5.13 0.63 1.27 RXFP1 0.08 0.08 0.02 0.00 GPR128 0.05 0.03 0.00 0.00 RXFP2 0.03 0.03 0.00 0.00 GPR132 2.81 0.36 3.21 0.16 RXFP3 0.11 0.08 0.01 0.00 GPR133 0.19 0.05 0.00 0.01 RXFP4 0.25 0.50 0.12 0.00 GPR135 1.24 1.35 2.00 0.66 S1PR1 0.06 40.23 0.32 0.65 GPR137 4.95 6.05 2.10 13.32 S1PR2 4.75 2.57 0.87 5.69 GPR139 0.06 0.10 0.26 0.00 S1PR3 13.75 8.62 66.07 10.19 GPR141 1.76 0.06 0.15 0.00 S1PR4 16.78 0.12 0.01 0.03 GPR142 0.12 0.10 0.00 0.00 S1PR5 0.12 0.17 0.09 1.20 GPR143 0.33 2.05 0.03 0.02 SCTR 0.05 0.03 0.00 0.00 GPR144 0.09 0.07 0.00 0.16 SMO 0.13 0.09 0.49 17.11 GPR146 0.85 0.10 0.28 0.36 SSTR1 0.05 0.10 0.09 0.07 GPR148 0.06 0.03 0.00 0.00 SSTR2 0.83 0.29 0.04 1.39 GPR149 0.02 0.02 0.04 0.01 SSTR3 0.15 0.16 0.00 0.01 GPR15 0.01 0.01 0.19 0.00 SSTR4 0.04 0.05 0.00 0.00 GPR150 0.17 0.21 0.01 0.00 SSTR5 0.08 3.80 20.16 0.04 GPR151 0.02 0.01 0.09 0.00 SUCNR1 16.94 13.20 0.00 0.00 GPR152 0.10 0.04 0.14 0.00 TAAR1 0.03 0.02 0.00 0.00 GPR153 0.49 1.00 2.08 4.13 TAAR2 0.04 0.05 0.00 0.01 GPR156 0.05 0.22 0.07 0.33 TAAR3 0.23 0.21 0.03 0.00 GPR157 2.46 1.91 6.05 2.78 TAAR5 0.01 0.01 0.00 0.00 GPR158 0.04 0.05 0.58 0.01 TAAR6 0.01 0.02 0.00 0.00 GPR160 11.78 0.83 1.10 11.92 TAAR8 0.02 0.01 0.00 0.00 GPR161 0.79 5.94 0.90 3.37 TAAR9 0.05 0.03 0.00 0.00 GPR162 0.90 6.74 0.32 0.74 TACR1 0.04 0.08 0.05 0.00 GPR17 0.03 0.65 0.06 0.01 TACR2 0.32 0.18 0.03 0.02 GPR171 0.05 0.03 0.07 0.10 TACR3 0.04 0.03 0.00 0.00 GPR173 0.09 0.17 0.13 0.05 TAS1R1 0.02 0.18 0.04 0.01 GPR174 0.01 0.02 0.00 0.00 TAS1R2 0.02 0.04 0.03 0.00 GPR176 0.07 11.18 1.87 2.90 TAS1R3 0.37 0.34 3.10 0.07

170

Gene U937 EaHy PC3M HEK293 Gene U937 EaHy PC3M HEK293 GPR179 0.05 0.05 0.02 0.03 TAS2R1 0.01 0.03 0.03 0.00 GPR18 2.68 0.07 0.00 0.01 TAS2R10 1.07 0.04 0.05 0.02 GPR182 0.04 0.05 0.08 0.19 TAS2R13 0.74 0.33 0.00 0.00 GPR183 13.12 0.12 0.05 0.13 TAS2R14 1.39 0.50 0.05 0.36 GPR19 0.12 0.09 0.59 0.52 TAS2R16 0.04 0.05 0.00 0.00 GPR20 0.01 0.10 0.06 0.12 TAS2R19 1.29 1.15 0.19 0.04 GPR21 0.06 0.27 0.04 0.16 TAS2R20 1.39 0.57 0.30 0.57 GPR22 0.06 0.30 0.00 0.00 TAS2R3 0.37 0.01 0.04 0.80 GPR25 0.08 0.09 0.00 0.00 TAS2R30 0.24 0.08 0.00 0.00 GPR26 0.05 0.04 0.00 0.01 TAS2R31 0.24 0.33 0.30 0.68 GPR27 0.42 0.42 0.13 11.82 TAS2R38 0.01 0.03 0.00 0.26 GPR3 0.07 0.58 1.69 0.68 TAS2R39 0.01 0.02 0.00 0.00 GPR31 0.30 0.11 0.00 0.00 TAS2R4 0.16 0.18 0.00 1.03 GPR32 0.05 0.14 0.00 0.00 TAS2R40 0.03 0.01 0.00 0.00 GPR33 0.04 0.00 0.00 0.00 TAS2R41 0.03 0.03 0.00 0.00 GPR34 0.36 0.03 0.00 0.02 TAS2R42 0.02 0.02 0.00 0.22 GPR35 1.82 0.06 0.06 0.21 TAS2R43 0.03 0.40 0.00 0.58 GPR37 0.05 0.12 1.20 0.32 TAS2R46 0.02 0.28 0.08 0.19 GPR37L1 0.86 0.40 0.23 0.80 TAS2R5 0.10 0.07 0.04 0.48 GPR39 0.09 0.09 0.12 0.01 TAS2R50 0.33 0.10 0.00 0.02 GPR4 0.18 0.99 0.02 0.18 TAS2R60 0.08 0.03 0.00 0.00 GPR45 0.07 0.04 0.00 0.15 TAS2R7 0.01 0.05 0.00 0.00 GPR50 0.01 0.01 0.00 2.42 TAS2R8 0.08 0.01 0.00 0.04 GPR52 0.61 0.05 0.00 0.02 TAS2R9 0.00 0.01 0.00 0.00 GPR55 0.13 0.08 0.20 0.00 TBXA2R 2.23 0.71 0.21 2.11 GPR56 0.44 28.50 0.27 0.48 TPRA1 22.78 8.65 7.09 14.97 GPR6 0.04 0.03 0.00 0.00 TRHR 0.00 0.02 0.03 0.01 GPR61 0.03 0.03 0.00 0.01 TSHR 0.30 0.07 0.03 0.04 GPR62 0.22 0.13 0.09 0.07 UTS2R 0.10 0.10 0.00 2.94 GPR63 1.65 0.33 0.82 2.92 VIPR1 0.32 0.22 0.16 1.97 GPR64 0.32 0.51 0.13 0.49 VIPR2 0.45 0.25 0.02 0.38 GPR65 0.80 0.28 0.28 0.43 XCR1 0.02 0.03 0.07 0.00

171

Appendix 3: Chapter 4 Supplemental Experimental

Procedures

Primary erythroid precursor cell isolation. Primary erythroid precursors were isolated from

E14.5 fetal livers using the EasySep negative selection Mouse Hematopoietic Progenitor Cell Enrichment kit (StemCell Technologies) (Hewitt et al., 2015; McIver et al., 2014). Fetal livers were dissociated by pipetting and resuspended at 5 x 107 cells/ml in PBS containing 2% FBS, 2.5 mM EDTA, and 10 mM glucose. EasySep Mouse Hematopoietic Progenitor Cell Enrichment Cocktail was added at 50 µg/ml supplemented with 2.5 µg/ml biotin-conjugated CD71 antibody (eBioscience). After 15 min incubation on ice, the cells were washed by centrifugation for at 1200 rpm at 4°C. Cells were resuspended at 5 X 107 cells/ml in PBS containing 2% FBS, 2.5 mM EDTA, and 10 mM glucose, and EasySep Biotin Selection

Cocktail was added at 100 µg/ml. After 15 min incubation at 4°C, EasySep Mouse Progenitor Magnetic

Microparticles were added at 50 µg/ml. After 10 min incubation at 4°C, cells were resuspended to 2.5 ml and incubated with a magnet for 3 min. Unbound cells were carefully transferred into 15 ml tube and used for subsequent experiments.

Cell culture. G1E-ER-GATA-1 cells were cultured with or without 1 µM estradiol as described

(DeVilbiss et al., 2013). Fetal liver erythroid precursor cells were cultured and maintained at a density of

2.5 x 105 – 1 x 106 cells/ml in StemPro-34 (Gibco) supplemented with 10% nutrient supplement (Gibco), 2 mM L-glutamine (Cellgro), 1% penicillin/streptomycin (Cellgro), 100 µM monothioglycerol (SigmaAldrich),

1 µM dexamethasone (Sigma-Aldrich), 0.5 U/ml of erythropoietin, and 1% conditioned medium from a KIT ligand-producing CHO cell line for expansion. Cells were cultured in a humidified incubator at 37°C and 5% carbon dioxide.

Retroviral infection: For concentrating retrovirus, the retroviral supernatant was mixed with the

Retro-X Concentrator and rotated at 4°C for 1h. The mixture was then centrifuged at 1,500 g for 45 minutes to obtain a high-titer virus-containing pellet that was resuspended at 1% of the original volume in IMDM

Media supplemented with 20% FBS [Gemini] and 1% penicillin/streptomycin [Cellgro].

Gpr65 shRNA sequence:

172

TGCTGTTGACAGTGAGCGAGCAGGTTAAGTTACATGGTATTAGTGAAGCCA

CAGATGTAATACCATGTAACTTAACCTGCCTGCCTACTGCCTCGGA

Quantitative real-time RT-PCR. Total RNA was purified from TRIzol (Invitrogen) according to manufacturers’ instructions. cDNA was synthesized from 1 µg purified total RNA by Moloney murine leukemia virus reverse transcriptase (MMLV RT). Real-time PCR was performed with SYBR green master mix (Applied Biosystems), and product accumulation was monitored by SYBR green fluorescence using either a StepOnePlus or Viia7 instrument (Applied Biosystems). Relative expression was determined from a standard curve of serial dilutions of cDNA samples, and values were normalized to 18S RNA expression.

The sequences of primers used for RT-PCR and genotyping are provided below. Colony assay. FACS- sorted CD31+ c-KIT+ cells from infected AGMs were plated in MethoCult M03434 complete media

(StemCell Technologies) in a 35mm dish. After incubation in a humidified incubator at 37°C with 5% carbon dioxide, colonies were visualized by microscopy and quantitated. Cells isolated from colonies were subjected to Wright-Giemsa staining.

Flow cytometry. Antibodies used for flow cytometry: APC-conjugated antibody CD31 (MEC13.3,

Biolegend), PE-conjugated antibody c-KIT (2B8, eBioscience), PerCP-Cy5.5-conjugated antibody SCA1

(D7, eBioscience), APC-conjugated antibody Ter119 (TER119, eBioscience) and PE-conjugated CD71

(R17217, eBioscience) were used for flow cytometry and Fluorescence Activated Cell Sorting (FACS).

Dissociated cells from cultured AGM explants were resuspended in PBS containing 2% FBS and passed through 25 µm cell strainers to obtain single-cell suspensions prior to antibody staining. Fetal liver cells were washed with PBS once prior to antibody staining. Samples were analyzed on a FACSAria™ II cell sorter (BD Biosciences). Cells were gated on GFP to ensure retroviral expression. DAPI (Sigma-Aldrich) exclusion was utilized for live/dead discrimination. FACS-sorted cells were immediately processed for RNA isolation or colony assay.

Zebrafish morpholino knockdown. Detailed methodology is presented in supplemental experimental procedures.Wild-type zebrafish embryos were injected at the single cell stage with 0, 4, or 6 ng of Gpr65_ATG MO (MO sequence: CATCTCAAGGGAGCATAAGTGCGTC) or Gpr65_SP MO (MO sequence: TAAATCGACAACTCACCATAAGTGC). Embryos were grown to 36 hours post-fertilization. No

173 defects in gross morphology or circulation were noted at the MO doses used. Embryos were fixed in formaldehyde and stained by in situ hybridization with either a Gata2b probe or a runx1/c-myb probe mixture. Embryos were blindly scored as having either low, medium, or high AGM expression of runx1/c- myb or Gata2b. The Gata2b probe was generously provided by David Traver.

ChIP assay. Detailed methodology is presented in supplemental experimental procedures.

Quantitative chromatin immunoprecipitation (ChIP) was conducted as described using antibodies specific to monomethylated H4K20 (Millipore), GATA-1, and Scl/TAL1 (DeVilbiss et al., 2015). Samples were analyzed by quantitative real-time PCR using either a StepOnePlus or Viia7 instrument (Applied

Biosystems). The amount of product was determined relative to a standard curve generated from a serial dilution of input chromatin.

Primers.

+9.5 flanking forward: 5′-ATGTCCTTTCGGATCTCCTGCC-3′ +

9.5 flanking reverse: 5′-GGTAAACAGAGCGCTACTCCTGTGTGTT-3′

18S rRNA forward: 5’-CGCCGCTAGAGGTGAAATTCT-3’

18S rRNA reverse: 5’-CGAACCTCCGACTTTCGTTCT-3’

Gata2 mRNA forward: 5′-GCAGAGAAGCAAGGCTCGC-3′

Gata2 mRNA reverse: 5′-CAGTTGACACACTCCCGGC-3′

Gpr65 mRNA forward: 5’-CAAGAGAAGCATCCCTCCAGAA-3’

Gpr65 mRNA reverse: 5’-TGTTTTTATTTTCACGCCGTTTG-3’

Gata2 primary transcript forward: 5’-GACATCTGCAGCCGGTAGATAAG-3’

Gata2 primary transcript reverse: 5’-CATTATTTGCAGAGTGGAGGGTATTAG-3’

MyoD promoter forward: 5′-GGGTAGAGGACAGCCGGTGT-3’

MyoD promoter reverse: 5′-GTACAATGACAAAGGTTCTGTGGGT-3’

Eif3k promoter forward: 5′-GTGATTTCCTTCCAGCAGTTGTAA-3’

Eif3k promoter reverse: 5′-CTCACGCTATTGGTCTCTTTTAAGTG-3’

Gata2 _9.5 Site _933 bp forward: 5′-CTTGCTGCTGGCTCTGAGAAC-3’

Gata2 _9.5 Site _933 bp reverse: 5′-AGTCCAGGGTCTTTTAAGGATAAATTC-3’

Gata2 _9.5 Site_480 bp forward: 5′-AACCTTCAAATGCAGACACTTCAC-3’

174

Gata2 _9.5 Site_480 bp reverse: 5′-GAATCCGCCAGAACGAAGAC-3’

Gata2_9.5 Site forward: 5′-GACATCTGCAGCCGGTAGATAAG-3’

Gata2_9.5 Site reverse: 5′-CATTATTTGCAGAGTGGAGGGTATTAG-3’

Gata2_9.5 Site_446 bp forward: 5′-GCCGAGGGAGTTCAGTGCTA-3’

Gata2_9.5 Site_446 bp reverse: 5′-AGCGCTACTCCTGTGTGTTCTTC-3’

Gata2_9.5 Site_880 bp forward: 5′-TCCTGGCGACTCCTAGATCCTA-3’

Gata2_9.5 Site_880 bp reverse: 5′-GAAAGCCCTGAGGAAGTTGGA-3’

Lyl1 Exon 1 forward: 5′-TCAGCATTGCTTCTTATCAGCC-3’

Lyl1 Exon 1 reverse: 5′-CGCAGAGGCCAGAGGATG-3’

Kit_114 kb forward: 5′-GCACACAGGACCTGACTCCA-3’

Kit_114 kb reverse: 5′-GTTCTGAGATGCGGTTGCTG-3’

Hdc mRNA forward: 5’-ACCTCCGACATGCCAACTCT-3’

Hdc mRNA reverse: 5’-CCGAATCACAAACCACAGCTT-3’ c-Kit mRNA forward: 5’-AGCAATGGCCTCACGAGTTCTA-3’ c-Kit mRNA reverse: 5’-CCAGGAAAAGTTTGGCAGGAT-3’

Lyl1 mRNA forward: 5’-AAGCGCAGACCAAGCCATAG-3’

Lyl1 mRNA reverse: 5’-AGCGCTCACGGCTGTTG-3’ c-Myb mRNA forward: 5’-CGAAGACCCTGAGAAGGAAA-3’ c-Myb mRNA reverse: 5’-GCTGCAAGTGTGGTTCTGTG-3’

Runx1 mRNA forward: 5’-TCACTGGCGCTGCAACAA-3’

Runx1 mRNA reverse: 5’-TCTGCCGAGTAGTTTTCATCGTT-3’

TAL1 mRNA forward: 5’-GAGGCCCTCCCCATATGAGA-3’

TAL1 mRNA reverse: 5’-GCGCCGCACTACTTTGGT-3’

Sfpi1 mRNA forward: 5’-GGCAGCGATGGAGAAAGC-3’ fpi1 mRNA reverse: 5’-GGACATGGTGTGCGGAGAA-3

175

Appendix 4: Digital Supplemental Files

Video S1 & S2. Confocal time-lapse imaging of HSPC trafficking to and engrafting CHT (caudal hematopoietic tissue) in zebrafish embryos. Tg(Runx1+23:GFP;flk:DsRed2) embryos were treated with chemicals between 24-46 hpf. Chemicals were washed off at 46 hpg. Embryos were immediately mounted in 1% low-melting-poing agarose and imaged between 54-63 hpf. Time-lapse mocies were taken on a spinning disk confocal microscope with a 28°C incubation chamber. Images were taken every 2 minutes and focused on the CHT region.

Video S1: DMSO-treated embryo

Video S2: 5μM 11,12-EET-treated embryo

Supplementary Table: EET treatment causes gene expression changes in two human hematopoietic cell types, CD34+ cells and U937 cells. Cells were treated with 5μM 11,12-EET for 2hrs at 37°C. Cells were then harvested and RNA was extracted and prepped for sequencing.

Tab 1-CD34_OK_log2fc>0.5: genes which passed quality control in CD34+ cells and increased upon EET treatment with log2 of fold change of greater than 0.5.

Tab 2-CD34_OK_log2fc<-0.5: genes which passed quality control in CD34+ cells and decreased upon EET treatment with log2 of fold change of less than -0.5.

Tab 3- U937_OK_log2fc>0.5: genes which passed quality control in U937 cells and increased upon

EET treatment with log2 of fold change of greater than 0.5.

Tab 4- U937_OK_log2fc<-0.5: genes which passed quality control in U937 cells and decreased upon EET treatment with log2 of fold change of less than -0.5.

Tab 5- overlap up-regulated: genes upregulated in both CD34+ cells and U937 cells by log2fc>0.5

Tab 6- IPA_overap_up-regulated: Ingenuity Pathway Analysis of the overlapping upregulated genes.

176