|Home | About | Journals | Submit | Contact Us | Français|
During primitive hematopoiesis in Xenopus, cebpa and spib expressing myeloid cells emerge from the anterior ventral blood island. Primitive myeloid cells migrate throughout the embryo and are critical for immunity, healing, and development. Although definitive hematopoiesis has been studied extensively, molecular mechanisms leading to the migration of primitive myelocytes remain poorly understood. We hypothesized these cells have specific extracellular matrix modifying and cell motility gene expression.
In situ hybridization screens of transcripts expressed in Xenopus foregut mesendoderm at stage 23 identified seven genes with restricted expression in primitive myeloid cells: destrin; coronin actin binding protein, 1a; formin-like 1; ADAM metallopeptidase domain 28; cathepsin S; tissue inhibitor of metalloproteinase-1; and protein tyrosine phosphatase nonreceptor 6. A detailed in situ hybridization analysis revealed these genes are initially expressed in the aVBI but become dispersed throughout the embryo as the primitive myeloid cells become migratory, similar to known myeloid markers. Morpholino-mediated loss-of-function and mRNA-mediated gain-of-function studies revealed the identified genes are downstream of Spib.a and Cebpa, key transcriptional regulators of the myeloid lineage.
We have identified genes specifically expressed in migratory primitive myeloid progenitors, providing tools to study how different gene networks operate in these primitive myelocytes during development and immunity.
The innate immune system is evolutionarily ancient and is specified before the adaptive immune system (Nagata, 1977; Tochinai, 1980; Kau and Turpen, 1983; Maeno et al., 1985). Myeloid cells are capable of protecting embryos from 98% of all pathogens encountered (Jones, 2000) and develop in 2 distinct temporal waves. The first wave called primitive hematopoiesis often occurs in blood islands and results in a transient population of erythroid and myeloid cells (Costa et al., 2008; Chen et al., 2009b). The second wave, known as definitive hematopoiesis, produces hematopoietic progenitors that ultimately provide all adult blood lineages. Though much is known about the latter wave in higher vertebrates, the former remains poorly understood in all vertebrates.
Research in the lower vertebrates, zebrafish and Xenopus, has revealed that primitive hematopoiesis occurs in 2 separate embryonic compartments (Palis et al., 1999; McGrath and Palis, 2005; Lux et al., 2008). Myelopoiesis occurs in the Xenopus anterior blood island (rostral blood island derived from the anterior lateral plate mesoderm in zebrafish) while erythropoiesis occurs in the posterior ventral blood island in Xenopus (posterior lateral plate mesoderm in zebrafish) (Warga et al., 2009; Ciau-Uitz et al., 2010; Ciau-Uitz et al., 2014).
Primitive myeloid cells are the first blood cells to differentiate and become functional in the Xenopus embryo and along with neural crest are some of the earliest migratory cells. A critical function of primitive myeloid cells is their ability to move within and between tissues where they are quickly and efficiently recruited to sites such as embryonic wounds even before a functional vasculature is established (Chen et al., 2009b). Myeloid cells have been implicated in diverse contexts of organ repair and regeneration among higher vertebrates: skin (Mirza et al., 2009; Goren et al., 2009) where their depletion results in delayed re-epithelialization, reduced collagen deposition, impaired angiogenesis, and decreased cell proliferation in healing wounds; muscle where two populations of monocytes sequentially phagocytose then accumulate myofibroblasts, promote angiogenesis, and deposit collagen (Nahrendorf et al., 2007; Arnold et al., 2007); kidney where wnt7b is produced by macrophages which invade the injured tissues and reestablish a developmental program beneficial for repair and regeneration (Lin et al., 2010), liver where macrophages play critical roles in both the injury and recovery phases of inflammatory scarring (Takeishi et al., 1999; Meijer et al., 2000; Duffield et al., 2005), and colon where macrophages migrate to a wound and promote epithelial proliferation at the injury site (Pull et al., 2005; Seno et al., 2009). Genes conferring myeloid cell motility, repair, and regeneration functions remain to be identified in all vertebrates.
More recent research suggests that myeloid cells are also likely to have important functions during normal embryogenesis (Rae et al., 2007; Stefater et al., 2011). In Drosophila, haemocytic macrophages phagocytose apoptotic cells from the developing embryo (Tepass et al., 1994; Franc et al., 1999). Also, macrophages actively induce apoptosis of endothelial cells in the pupillary membrane of the developing mammalian eye (Diez-Roux et al., 1999). Of interest, homozygous null mutation of Colony stimulating factor 1 (Csf1), effectively ablating macrophages from the mouse embryo, results in a major insulin mass deficit in fetuses and adults, abnormal postnatal islet morphogenesis, and impaired pancreatic cell proliferation at weaning and late pregnancy (Banaei-Bouchareb et al., 2004). Most recently migratory primitive myeloid cells have been reported to be essential for normal Xenopus cardiac morphogenesis (Smith and Mohun, 2011). These findings suggest that myeloid cells play important roles during normal embryonic development (Savill and Fadok, 2000). Exactly what functions within the myeloid cells confer such developmental roles has proved difficult to examine in mouse and higher vertebrates because few molecular markers are available either to identify embryonic primitive myeloid cells or to trace their ontogeny.
The earliest known markers of the primitive myeloid lineage in Xenopus include cebpa and spib.a transcripts. Cebpa is a basic helix-loop-helix transcription factor critical for the differentiation of murine myeloid progenitors into granulocytemonocyte progenitors (Zhang et al., 2004). cebpa mutations are often found in human patients with myeloid leukemias (Nerlov, 2004; Mueller and Pabst, 2006). Gain- and loss-of-function studies reveal that cebp a is necessary and sufficient for myeloid differentiation in Xenopus embryos (Chen et al., 2009b). spib.a encodes an ETS domain transcription factors that marks the primitive myeloid cell lineage in Xenopus and is required for its development where it acts upstream of spi1 (also known as pu.1) in the molecular hierarchy of primitive myeloid development (Costa et al., 2008).
Although some of the target genes regulated by these key myeloid transcription factors have been identified (Smith and Mohun, 2011), in general the downstream genetic program activated in primitive myeloid cells that confer their ability to migrate and mediate immunity are still poorly understood (Smith and Mohun, 2011). This study identified a group of genes expressed in the early primitive myeloid lineage of Xenopus embryos. The temporal and spatial expression patterns suggest they emerge after myeloid specification and with the onset of migratory activity. We show that the expression of these genes is regulated by Spib.a and Cebpa. These genes encode proteins that are implicated in mediating different aspects of myeloid cell migration and should facilitate elucidating the cell biology underlying the essential developmental and immunologic functions of the migrating primitive myeloid lineage in the Xenopus embryo.
In a previously published microarray experiment we identified several hundred genes expressed in early foregut (Stage 23) of Xenopus embryos (Kenny et al., 2012), GSE38654. In the course of validation by in situ hybridization we identified seven genes with punctate expression within the mesodermal layer of the anterior ventral blood island, reminiscent of primitive myeloid cells (Fig. 1). These genes included destrin (dstn); coronin, actin binding protein, 1a (coro1a); formin-like 1 (fmnl1); ADAM metallopeptidase domain 28 (adam28); cathepsin S (ctss); tissue inhibitor of metalloproteinase-1 (timp-1); and protein tyrosine phosphatase nonreceptor 6 (ptpn6). Of interest, research in mammals has suggested a role for many of these genes in adult myeloid cells (Yan et al., 2005; Shimoda et al., 2007; Mersich et al., 2010; Seto et al., 2012; Arora et al., 2012; Rochelle et al., 2013; Zajac et al., 2013; McComb et al., 2014; Masmas et al., 2014). However whether or not these genes are expressed in embryonic primitive hematopoiesis-derived myeloid cells has not previously been documented in any vertebrate species.
To test whether these genes are expressed in the primitive myeloid lineage, we compared their developmental expression with spib.a, and its downstream target mmp7, which are known to be specifically expressed in the primitive myeloid cells during stages 15 through 28 (Costa et al., 2008). Spib.a preceded the expression of these genes in the anterior ventral blood island (aVBI) at stage 15 (presomitic stage) with the exception of the emerging expression of timp-1 (Nieuwkoop and Faber, 1994) (Fig. 1A–I). All seven genes exhibited an expression profile very similar to spib.a, although they were less abundant compared with spib.a before stage 24. At stage 32, all genes tested were dispersed throughout the embryo excluding the area of the branchial arches (data not shown), a phenomenon observed in previous primitive myeloid gene expression (Ciau-Uitz et al., 2010). The in situ analysis suggested that the seven genes were expressed in subsets of the spib.a+ cells and that expression between the genes appears nonoverlapping, possibly representing different cell subpopulations within the spib.a-expressing myeloid lineage (Fig. 1 and data not shown). Thus, overall the seven genes that we identified were expressed in the anterior ventral blood island in a pattern very similar to other known primitive myeloid marker, but in general they were expressed later than spib.a and cebpa (data not shown) which is consistent with them being myeloid-specific genes temporally downstream of spib.a.
We further examined the myeloid-specific co-expression of spib.a and one of the identified genes, destrin (Fig. 2). Double whole-mount in situ hybridization (WMISH) at stage 22 revealed significant overlap of expression for these two genes (Fig. 2A–C). Later stage 25 WMISH revealed spib.a(−) destrin(+) cells migrating posteriorly and laterally out of the blood island (Fig. 2D). This is consistent with gene expression coinciding with increasing myeloid cell motility and migration away from the anterior ventral blood island. An alternative explanation of this data is the possibility of independent sequential transient waves of expression.
Using gene ontology analysis we assessed the predicted molecular and cellular functions of the genes we isolated based on information from their human and mouse orthologs (Table 1) (on Web site ToppGene.cchmc.org; Table 1) and by literature review. Molecular function analysis revealed DSTN, CORO1A, FMNL1 are involved in actin binding and that ADAM28 and CTSS have endopeptidase activity. The biological process analysis revealed that actin filament severing activity was mediated through DSTN and FMNL1; also extracellular matrix disassembly activity was mediated through CTSS and TIMP-1. Furthermore, ToppGene coexpression analysis has revealed these genes to be expressed in blood development and cancer (Table 1).
DSTN is an actin-depolymerizing factor present in actin micro-filaments and can break down actin filaments (Hatanaka et al., 1996). It functions in actin skeleton dynamics affecting cell shape and motility downstream of LIM-kinase (Suetsugu and Takenawa, 2003) and in mammalian myeloid cells (Ichikawa et al., 1982). CORO1A is an actin binding protein that interacts with microtubules and in some cell types associated with microtubules. CORONIN1 has been found to exert an inhibitory effect on cellular steady-state F-actin formation by means of an ARP2/3-dependent mechanism. It is required in humans for chemokine-mediated migration. It plays important roles in cell signaling, cell migration, phagocytosis, and vesicle trafficking (Castro-Castro et al., 2011) in human myeloid and other cells (Federzoni et al., 2013). FMNL1 homologues are implicated in BMP-mediated sprouting in endothelial cells (Wakayama et al., 2015) and ame-boid invasive cell motility and actin filament binding downstream of RHOC in human tumor cells (Kitzing et al., 2010). FMNL1 has been found in the actin-rich cores of primary human macrophage podosomes (Mersich et al., 2010).
ADAM homolog overexpression in nonsmall cell lung cancer in humans correlates with cell migration and invasion through NOTCH1 signaling in humans (Guo et al., 2012) and has been implicated in lymphocyte development (Gibb et al., 2011). ADAM10 has been found to cleave various cell surface molecules, including adhesion molecules, chemokines, and growth factor receptors. ADAM10-deficient macrophages showed reduced migration and extracellular matrix degradation as well as altered IL-10, IL-12, NO, and TNF signaling in mice (van der Vorst et al., 2015). CTSS is a lysosomal cysteine protease that has been implicated as essential in macrophage migration and development by cleavage of Rip1 kinase in mice (Verollet et al., 2011; McComb et al., 2014). TIMP-1 homologues inhibit metalloproteases which regulate immune cell development and migration by signaling downstream of TNF receptor, IL-6 receptor, EGF, and NOTCH (Khokha et al., 2013) and are found in the monocyte/macrophage population in mice (Guedez et al., 2012). SH2 domain-containing PTP homologues have intracellular protein tyrosine phosphatase activity implicated in cell migration during normal cardiac development (Lauriol et al., 2014) and when knocked out in murine myeloid cells result in attenuated renal fibrosis after unilateral ureter obstruction (Teng et al., 2015).
All these observations and findings derived from work in higher vertebrates are consistent with the identified genes having putative roles in either actin cytoskeleton or extracellular matrix functional pathways in differentiating Xenopus primitive myeloid cells (Bonnans et al., 2014). These functions could mediate the essential developmental and immunologic roles of this myeloid population. Future work will investigate which cellular functions fulfill such roles.
To confirm that the genes we have identified are expressed in the myeloid lineage downstream of the key transcription factor Spib.a, we examined their expression in Spib.a-depleted embryos that are known to lack primitive myeloid cells. (Fig. 3) (Costa et al., 2008). Using a well-established morpholino WMISH analysis of the known Spib.a-target genes, mpo and mmp7 confirmed that injection of a well-characterized Spib.a antisense morpholino oligo resulted in a loss of primitive myeloid cells (Smith et al., 2002; Harrison et al., 2004; Costa et al., 2008) (Fig. 3A–C′). As expected, Spib.a depletion resulted in a substantial reduction of all of the candidate genes (Fig. 3D–J′), confirming that they are expressed in the myeloid lineage downstream of Spib.a.
Ectopic expression of cebpa mRNA in the ectoderm (“animal cap”) of the early Xenopus embryo is sufficient to induce ectopic myeloid gene expression and generate migrating primitive myeloid cells that can be in transplanted into host embryos (Chen et al., 2009b). We therefore examined whether injection of Xenopus tropicalis cebpa into the animal blastomeres was sufficient to induce the ectopic expression of our seven candidate myeloid genes. We found increased or ectopic expression in all of the tested genes except fmnl1 (which showed normal or decreased expression; Fig. 4 and data not shown). Thus, we conclude that six of our seven genes are downstream of the essential primitive myeloid transcriptional regulator, Cebpa.
Finally we combined the spib.a-MO-mediated loss-of-function and cebpa mRNA-mediated gain-of-function approaches. We transplanted Cebp-expressing animal cap cells into the blastocoel of Spib.a-depleted embryos (a procedure known as an Einsteck experiment) (Sive et al., 2000), which rescued dstn and coro1a expression in the Spib.a morphant host embryos anterior ventral blood island (Fig. 5A–C′). We conclude that these genes are specific to the primitive myeloid lineage.
In this “Patterns and Phenotypes” work, we report seven genes that are expressed specifically in the mesoderm-derived primitive myeloid cells in Xenopus embryos. This is significant because molecular markers specific to primitive myeloid cells in higher vertebrates remain unknown, and the expression of genes in myeloid cells produced from definitive hematopoiesis are not expressed in the primitive myeloid lineage (Palis et al., 1999; James Palis, personal communication). CD34 has been identified in murine primitive myeloid cells, although its expression is not specific to the myeloid lineage (Gottgens et al., 1997; Sidney et al., 2014). We have used a combination of loss-of-function and gain-of-function experiments supporting that these primitive myeloid genes are epistatically downstream of key primitive myeloid transcriptional regulators, spib.a and cebpa.
Given the highly conserved roles of the primitive myeloid lineage, characterization of the genetic signature of this cell type is critical to better understand their function in developmental mechanisms including repair, regeneration, patterning (blood vessels), morphogenesis (bone and early heart), survival and/or apoptosis (neuronal, adipose), and cell fate decisions (pancreatic β cell) (Banaei-Bouchareb et al., 2004, 2006; Stefater et al., 2011; Smith and Mohun, 2011). Future work will determine if these genes function in a coordinated pathway for myeloid cell migration and function. For instance, loss of lurp1 gene function in Xenopus perturbed myeloid gene migration but not specification, perturbing cardiac morphogenesis (Smith and Mohun, 2011). Future work will determine the role each of these primitive myeloid progenitor genes plays in primitive myeloid cell migration, intracellular signaling, and extracellular matrix remodeling functions and determine exactly which of these functions impart the known important developmental mechanisms of this cell type (Stefater et al., 2011).
Xenopus embryo manipulation and culture were performed as described previously (McLin et al., 2007).
Gene ontology analysis software (available through Web site ToppGene.cchmc.org) was used to assess the predicted molecular and cellular functions of the isolated genes based on information available from their human and mouse orthologs (Chen et al., 2009a).
Embryos with clear dorsal–ventral pigmentation differences were selected for four- to eight-cell stage injections targeting Xenopus laevis spib.a morpholino to the dorsal blastomeres that contribute to the embryo’s dorsoanterior. Eight-cell animal blastomeres were targeted for misexpression of Xenopus tropicalis cebpa mRNA.
The MO used in this study has been previously published to generate specific loss-of-function phenotypes: Xenopus laevis spib.a-e1i1-MO (40 ng) (Smith and Mohun, 2011). For animal cap Einsteck rescue experiments, we injected Xenopus tropicalis cebpa mRNA synthesized from pFTX11.HA.xtCEBPa expression plasmid (150–300 pg/embryo; kind gift from Dr. Enrique Amaya) (Chen et al., 2009b).
Donor fragments were generated as follows. Stage 10–10.5 gfp-and cebpa-mRNA coinjected embryos had their vitelline membranes removed. Fluorescing, mesoderm-free donor animal caps were cut using hair knives and hair loops in 0.5 × MBS with 1:500 gentamicin. Stage 10 recipient spib.a morpholino-injected embryos had a slit generated by a hair knife and the animal cap donor tissue was inserted into the blastocoele. The Einsteck embryos were cultured to stage 22 and harvested for in situ analysis (Sive et al., 2000).
In situ hybridization was performed as previously described (Kenny et al., 2012). A sequential procedure for double whole-mount in situ hybridization was adopted, where embryos were photographed successively after the chromogenic reactions for each probe were developed. Digoxigenin- or fluorescein-labeled Xenopus laevis probes were synthesized from: pExpress-1 ceb-paaaaaaa-SPORT6-spib.a, pCMV-SPORT6-mmp7, pCMV-SPORT6-mpo2, pCMV-SPORT6-adam28, pCMV-SPORT6-coro1a, pCMV-SPORT6-ctss, pCMV-SPORT6-dstn, pCMV-SPORT6-fmnl1, pCMV-SPORT6-ptpn6, and pExpress-1-timp-1.
We thank Dr. Enrique Amaya for the pFTX11.ha.xtcebpa expression plasmid. This project was supported in part by a CCHMC Procter Scholarship and K08 HL105661 to A.P.K., a T32 HD07463 award to E.T.S., and PHS Grants P30 DK078392 and R01 DK070858 to A.M.Z.
Grant sponsor: CCHMC Procter Scholarship; Grant numbers: K08 HL105661 and T32 HD07463; Grant sponsor: PHS; Grant numbers: P30 DK078392 and R01 HD73179.