Hemogenic endothelial (HE) cells in the dorsal aorta undergo an endothelial-to-hematopoietic transition (EHT) to form multipotent progenitors, lympho-myeloid biased progenitors (LMPs), pre-hematopoietic stem cells (pre-HSCs) and adult-repopulating HSCs. These briefly accumulate in intra-arterial hematopoietic clusters (IAHCs) before being released into the circulation. It is generally assumed that the number of IAHC cells correlates with the number of HSCs. Here, we show that changes in the number of IAHC cells, LMPs and HSCs can be uncoupled. Mutations impairing MyD88-dependent toll-like receptor (TLR) signaling decreased the number of IAHC cells and LMPs, but increased the number of HSCs in the aorta-gonad-mesonephros region of mouse embryos. TLR4-deficient embryos generated normal numbers of HE cells, but IAHC cell proliferation decreased. Loss of MyD88-dependent TLR signaling in innate immune myeloid cells had no effect on IAHC cell numbers. Instead, TLR4 deletion in endothelial cells (ECs) recapitulated the phenotype observed with germline deletion, demonstrating that MyD88-dependent TLR signaling in ECs and/or in IAHCs regulates the numbers of LMPs and HSCs.
Hematopoietic ontogeny occurs in multiple waves in which hematopoietic stem and progenitor cells (HSPCs) with different potentials differentiate from hemogenic endothelial (HE) cells via an endothelial-to-hematopoietic transition (EHT). HE cells in the extra-embryonic yolk sac undergo EHT to become committed erythro-myeloid progenitors (EMPs) and lymphoid progenitors, while HE cells in the caudal arteries (dorsal aorta in zebrafish; dorsal aorta, vitelline and umbilical in mouse) differentiate into multipotent, lympho-myeloid biased progenitors (LMPs), HSCs and HSC precursors called pre-HSCs (Auerbach et al., 1996; Hadland and Yoshimoto, 2018). Following EHT, the newly formed HSPCs in the major arteries briefly accumulate in intra-arterial hematopoietic clusters (IAHCs) consisting of cells that uniformly express endothelial markers, hematopoietic transcription factors (such as RUNX1) and high levels of KIT (Bollerot et al., 2005). At embryonic day (E) 10.5, when IAHC numbers peak, the IAHC cells contain several hundred LMPs, ∼25 pre-HSCs and ∼0.03 adult-repopulating HSCs (Kumaravelu et al., 2002; Li et al., 2014; Rybtsov et al., 2016). At E11.5, IAHCs contain ∼65 pre-HSCs and one adult repopulating HSC (Kumaravelu et al., 2002; Rybtsov et al., 2016). Hence, the IAHCs are heterogeneous, containing HSPCs with unique differentiation and engraftment potentials, and their composition changes during development.
Innate immune and inflammatory signaling pathways have been shown to regulate the number of HSPCs generated in the arteries of zebrafish and mouse embryos. Mutations in or morpholino knockdown of genes encoding type I and II interferons (Li et al., 2014; Sawamiphak et al., 2014), tumor necrosis factor (TNF) (Espín-Palazón et al., 2014), inflammasomes (Frame et al., 2020), interleukin 1β (IL1β) (Frame et al., 2020), granulocyte colony stimulating factor (G-CSF) (He et al., 2015), toll-like receptor 4 (TLR4) (He et al., 2015), nuclear factor κB (NF-kB) (Espín-Palazón et al., 2014; He et al., 2015) and RIG1-like receptors (Lefkopoulos et al., 2020) reduced the number of myb:GFP+ kdrl:mCherry+ HSPCs in transgenic reporter zebrafish embryos, and decreased the levels of runx1 and myb transcripts in the dorsal aorta (DA). Innate immune signaling is initiated by binding of ligands to pattern recognition receptors (PRRs). PRRs recognize pathogen-associated molecular patterns (PAMPs), which are small molecules common to a specific class of infectious agents, and host molecules referred to as danger-associated molecular patterns (DAMPs), which are endogenous proteins or small molecules that are released from the cell or the extracellular matrix in response to cell damage, stress or death. Although, in theory, PAMPs from infectious agents might be responsible for activating innate immune signaling in mammalian or pre-hatched zebrafish embryos, it is more likely that DAMPs are the PRR ligands in the relatively sterile environment of the mammalian embryo.
Activation of toll-like receptors (TLRs) drives the production of type I interferons, TNF, IL-1β and G-CSF, and therefore directly or indirectly stimulates multiple inflammatory pathways (Monlish et al., 2016). TLRs signal through two intracellular pathways that involve either the MyD88 (myeloid differentiation primary response protein) and TIRAP (toll-interleukin 1 receptor domain containing adaptor protein), or the TRIF [toll IL-1 receptor (TIR)-domain-containing adaptor inducing interferon beta] and TRAM (translocating chain-associated membrane protein) adaptor proteins. Signaling from TLR4 on the plasma membrane through MyD88/TIRAP results in the nuclear localization of the CREB, AP1 and NF-κB transcription factors, and expression of inflammatory cytokine genes such as TNF, IL1β and G-CSF. Signaling through endosomal TLR4, TRIF/TRAM and the interferon regulatory factors (IRFs) primarily activates the expression of interferon genes. Morpholino knockdown of the genes encoding TLR4 (tlr4bb) and MyD88 (myd88) in zebrafish embryos reduced the expression of runx1, the number of myb:GFP+ kdrl:mCherry+ HSPCs in the DA, and HSPCs in secondary sites of colonization, including the caudal hematopoietic tissue and thymus (He et al., 2015). However, markers that discriminate between committed progenitors (e.g. LMPs) and pre-HSCs/HSCs are not available for zebrafish embryos; therefore, determining how TLR4 signaling differentially affects them is difficult. Understanding how LMP and pre-HSC/HSC generation are regulated is important for selectively optimizing production of the desired cells from arterial HE cells ex vivo.
Here, we examine TLR signaling to determine in which cells it is required, through which intracellular adaptors it signals to regulate arterial HSPC ontogeny and the types of HSPCs it regulates. We show that signaling through the MyD88 arm of the TLR pathway increases the number of IAHC cells and LMPs, while restricting the number of multi-lineage adult-repopulating HSCs.
RESULTS AND DISCUSSION
TLR4 signaling regulates RUNX1 levels and the number of IAHC cells
It was reported that loss of TLR4 reduced expression of the transcription factor runx1 in the DA of zebrafish embryos and the number of RUNX1+ cells in the DA of mouse embryos (He et al., 2015). RUNX1 is a specific marker of HE and IAHC cells, and is required for EHT (North et al., 1999). To examine whether the reduction in runx1 expression could be due to fewer RUNX1+ HE cells, to lower RUNX1 protein levels in HE cells or to a decrease in IAHC cells, we analyzed Tlr4−/− mouse embryos and their wild-type littermates by whole-mount confocal microscopy (Fig. 1A-I). IAHC cells were identified by their rounded appearance and expression of RUNX1 and CD31, and high levels of KIT, while HE cells were flatter, located within the endothelial layer, and expressed RUNX1, CD31 and very low or undetectable levels of KIT. We found that the relative fluorescence intensity of RUNX1 protein was lower in HE cells in the DA of E10.5 Tlr4−/− embryos compared with Tlr4+/+ littermates, but the numbers of RUNX1+ HE cells were equivalent (Fig. 1A-C). As the number of HE cells at a single snapshot in time is a function of the number that were originally specified minus the number that have undergone EHT, we enumerated HE cells at E9.5, prior to their major egress from the endothelial cell layer via EHT. The numbers of phenotypic RUNX1+ HE cells were equivalent in E9.5 Tlr4+/+ and Tlr4−/− embryos (Fig. 1D), and the levels of RUNX1 in Tlr4−/− HE cells were slightly increased (Fig. 1E). We compared the frequencies of functional HE cells by sorting limiting numbers of ECs (CD31hiCD144+ESAM+CD44+CD45−Ter119−Mac1−CD41−) into individual wells containing OP9 stromal cells (≤5 ECs/well) and counting the number of wells containing hematopoietic cells 10 days later. The frequencies of functional HE cells were equivalent in E9.5 Tlr4+/− and Tlr4−/− embryos (Fig. 1F, Fig. S1A). Therefore, loss of TLR4 decreases RUNX1 levels in E10.5 HE cells but has no effect on the number of HE cells.
Loss of TLR4 reduced the number of IAHC cells by 42% in E10.5 embryos, and IAHC cells had lower levels of RUNX1 (Fig. 1G,H). However, the numbers of IAHC cells in Tlr4+/+ and Tlr4−/− embryos were equivalent at E11.5, indicating that the TLR4 requirement for normal numbers of IAHC cells was transient (Fig. 1I). The decreased number of IAHC cells at E10.5 could be caused by inefficient EHT, decreased proliferation or increased apoptosis. We observed no decrease in specified HE cells at E9.5 or accumulation of HE cells at E10.5, suggesting that the decrease in IAHC cells at E10.5 is not due to inefficient EHT, although we cannot rule this out as inefficient EHT might cause HE cells to lose RUNX1 expression, to assume an alternative cell fate or to undergo apoptosis. Instead, our data suggest that TLR4 deficiency decreases the number of IAHC cells by reducing their proliferation, as reflected by a smaller fraction of IAHC cells that were Ki-67 positive (Fig. 1J,K). EdU labeling and flow cytometric analysis were also consistent with an effect of TLR4 loss on proliferation, specifically of CD45+ IAHC cells, which lack pre-HSCs but are enriched for LMPs at E10.5 (Rybtsov et al., 2011; Zhu et al., 2020) (Fig. 1M,N and Fig. S1B). TLR4 deficiency did not affect the proliferation of ECs, HE cells or CD45− IAHC cells, which include pre-HSCs at E10.5 (Fig. 1L-N). In summary, loss of TLR4 decreased RUNX1 protein levels in HE and IAHC cells, and reduced the proliferation of CD45+ IAHC cells (LMPs) at E10.5, but did not affect the number IAHC cells at E11.5 or the number of HE cells specified at E9.5.
Both TLR4 and TLR2 are expressed in and regulate inflammatory signaling in ECs (Salvador et al., 2016); therefore, we determined whether TLR2 also regulates the number of IAHC cells. Loss of TLR2 reduced the number of IAHC cells in the DA to a similar extent to TLR4 deficiency (Fig. 2B). TLR2 signals exclusively through the MyD88 adaptor, while TLR4 signals through both the MyD88 and TRIF adaptors (Fig. 2A). To determine which signaling arm of TLR4 was important for regulating the number of IAHC cells, we enumerated IAHC cells in both Myd88−/− and TRIF-deficient (Ticam1−/−) embryos. Loss of MyD88 decreased, whereas loss of TRIF increased, the number of IAHC cells (Fig. 2B). Therefore, signaling via cell-surface TLR2 and TLR4 through the MyD88 adaptor increases the number of IAHC cells at E10.5, likely by promoting their proliferation, whereas endosomal TLR4 signaling through the TRIF adaptor constrains IAHC cell numbers.
MyD88-dependent TLR signaling positively regulates the number of LMPs but restrains the number of adult repopulating HSCs
To determine whether loss of TLR signaling also decreased the number of adult-repopulating HSCs capable of multilineage engraftment, we transplanted one embryo equivalent of aorta-gonad-mesonephros (AGM) regions from E11.5 wild type, Tlr4−/−, Tlr2−/−, Myd88−/−, and Ticam1−/− embryos into irradiated adult recipient mice. One embryo equivalent (ee) of an E11.5 wild-type AGM region is a limiting dose of adult repopulating HSCs (Kumaravelu et al., 2002). We measured contribution to peripheral blood and bone marrow 16 weeks post-transplant (Fig. S2A). Despite the decrease in the number of IAHC cells, loss of TLR4, TLR2 and MyD88 increased the proportion of mice with multi-lineage engraftment (Fig. 2C), indicating that adult-repopulating HSCs were not depleted in the AGM regions of TLR4, TLR2 or MyD88-deficient mouse embryos and, if anything, appeared to be increased. Loss of Ticam1, on the other hand, did not affect the proportion of engrafted recipients (Fig. 2C).
To quantify the numbers of multi-lineage adult repopulating HSCs in embryos with mutations affecting MyD88-dependent TLR signaling, we transplanted a cohort of mice with limiting doses (0.3 ee and 1.0 ee) of dissociated cells from the AGM regions of E11.5 wild-type, Tlr4−/− and Myd88−/− embryos. There was a trend towards increased numbers of multi-lineage repopulating HSCs in the AGM regions of Tlr4−/− embryos and a significant 4.2-fold increase in Myd88−/− embryos (Fig. 2D), indicating that MyD88-dependent TLR signaling decreases the number of HSCs.
On the other hand, there were no differences in the frequencies of type I (CD31+KIT+CD41loCD45−Ter119−) or type II (CD31+KIT+CD41loCD45+Ter119−) pre-HSCs in the arteries (AGM region, vitelline and umbilical) of E11.5 Tlr4+/+ and Tlr4−/− embryos (Fig. 2E, Fig. S2B). We also observed no changes in the frequencies or absolute numbers of phenotypic long-term repopulating HSCs (LT-HSCs) in E12.5 or E13.5 fetal livers (Fig. 2F, Fig. S2C). Therefore, loss of TLR4 signaling did not decrease the number of phenotypic pre-HSCs or functional HSCs in the AGM region, or the number of phenotypic HSCs in the fetal liver, whereas loss of MyD88-dependent TLR signaling significantly increased the number of adult-repopulating HSCs in the AGM region.
LMPs with the potential to produce B and T cells in culture are >10-fold more abundant in AGM regions of E10.5 embryos compared with HSCs and pre-HSCs, and are enriched in the CD45+ population (Li et al., 2014; Rybtsov et al., 2016; Zhu et al., 2020), which we observed was less proliferative in Tlr4−/− embryos (Fig. 1M,N). To determine whether the number of LMPs was altered by mutations affecting MyD88-dependent signaling, we performed a limiting dilution analysis of cells from the AGM regions of wild-type and Myd88−/− embryos by plating them into wells containing OP9 stromal cells to enumerate progenitors with B cell potential, and on OP9 cells expressing the Notch ligand delta-like 1 (OP9-DLL1) to measure the number of T cell progenitors (Li et al., 2014; Schmitt and Zúñiga-Pflücker, 2002). Myd88−/− embryos contained significantly fewer progenitors capable of producing B or T cells compared with wild-type embryos (Fig. 2G). Therefore, MyD88-dependent TLR signaling positively regulates the number of LMPs while restraining the number of HSCs.
TLR4 signaling in endothelial cells and/or IAHCs, but not in primitive innate immune myeloid cells, regulates IAHC cell numbers
Professional innate immune myeloid cells are the primary producers of inflammatory cytokines and chemokines, and are proposed to be an important cellular source of the inflammatory molecules promoting the generation of HSPCs in the DA. Depletion of macrophages and granulocytes in zebrafish embryos using morpholinos targeting spi1a and spi1b, of macrophages alone with morpholinos against irf8, or of neutrophils using morpholinos to cebp1 decreased runx1 expression in the DA and the number of myb:GFP+ kdrl:mCherry+ HSPCs (Espín-Palazón et al., 2014; He et al., 2015; Li et al., 2014). Depletion of macrophages in explanted AGM regions from mouse embryos using clodronate also decreased HSPCs (Mariani et al., 2019). Vital imaging documented macrophages intimately contacting IAHC cells, and macrophages were occasionally observed to extravasate through ECs to reach the IAHCs (Mariani et al., 2019; Travnickova et al., 2015).
We determined whether TLR4 signaling in yolk sac-derived myeloid cells regulates the number of IAHC cells in the DA by deleting Tlr4 floxed alleles (Tlr4f/f) in EMPs using Csf1r-Cre (Gomez Perdiguero et al., 2015; Schulz et al., 2012). Csf1r-Cre should result in TLR4 loss in both yolk sac-derived macrophages and granulocytes, although as granulocytes are a minor population in the mouse embryo, macrophages are the more likely source of inflammatory cytokines (Mariani et al., 2019). We confirmed that cell surface TLR4 was depleted on macrophages following Csf1r-Cre deletion (Fig. 3A,B and Fig. S3A), and that macrophages had an impaired response to LPS stimulation ex vivo, as evidenced by a lower percentage of TNF+ cells (Fig. 3C,D and Fig. S3B). Deletion of TLR4 by Csf1r-Cre did not significantly impact the number of IAHC cells (Fig. 3E). RUNX1 levels were modestly decreased (12.1%) in IAHC cells (Fig. 3F), but whether this was due to Csf1r-Cre mediated deletion in EMPs, or in differentiating IAHC cells that expressed Csf1r-Cre could not be determined. The data indicate that the cytokines produced by primitive myeloid cells in response to TLR4 signaling do not regulate the number of IAHC cells. However, these results do not exclude the possibility of a role for primitive myeloid cells in regulating IAHC cell numbers, as other signaling pathways that regulate inflammatory cytokine production or other processes mediated by primitive myeloid cells may be responsible for their effect on hematopoietic ontogeny in the DA.
We next assessed whether TLR4 signaling in endothelial or IAHC cells was necessary for generating normal numbers of IAHC cells by deleting Tlr4 with an endothelial cell-specific, tamoxifen-regulated Cre (Cdh5- CreERT) (Sørensen et al., 2009a). We timed the tamoxifen injections in pregnant dams to preferentially delete TLR4 in arterial ECs and IAHCs, but not in EMPs. EMPs differentiate from HE cells in the yolk sac vascular plexus between E8.5 and E9.5 (Frame et al., 2015), immediately prior to the major wave of EHT in arterial endothelium that produces IAHC cells (Tober et al., 2013; Yokomizo and Dzierzak, 2010). We previously performed a temporal analysis of Cdh5-CreERT activity following tamoxifen injection and found that deletion in EMPs could be mostly avoided by injecting tamoxifen at E9.0 or later in development (Tober et al., 2013). Therefore, to delete TLR4 in arterial ECs while sparing EMPs, we injected tamoxifen into pregnant dams at E9.5 and analyzed HE and IAHC cells 1 day later in Tlr4f/f and Tlr4f/f; Cdh5-CreERT embryos. We determined that the percentage of TNF+ macrophages was equivalent in E10.5 Tlr4f/f and Tlr4f/f; Cdh5-CreERT AGM regions stimulated ex vivo with LPS, confirming that deletion of Tlr4 using a carefully timed activation of Cdh5-CreERT has minimal impact on EMPs or their progeny (Fig. 3G,H). The number of IAHC cells at E10.5 was significantly decreased by 27% following deletion of Tlr4f/f alleles at E9.5 by Cdh5-CreERT. RUNX1 levels in IAHC cells were reduced by ∼14.0% , similar to the decrease caused by deletion with Csf1r-Cre (Fig. 3I,J). Deletion of Tlr4 in endothelial cells at E9.5 did not result in a phenotype as robust as that in Tlr4−/− embryos, suggesting that TLR4 function in either non-endothelial cells or endothelial cells prior to E9.5 contributes to the IAHC cell phenotype. Overall, however, the data indicate that TLR4 is required in arterial ECs and/or IAHC cells to generate normal numbers of IAHC cells.
To confirm that TLR4 signaling can be directly activated in embryonic ECs we purified CD31hiCD144+ESAM+CD44−CD45−Ter119−Mac1−CD41− ECs, stimulated them ex vivo with LPS and analyzed expression of the integrin CD62E (E-selectin, encoded by Sele), which is a direct NF-κB target gene (Collins et al., 1995), by flow cytometry. The percentage of CD62E+ ECs increased following stimulation with LPS, indicating that embryonic ECs respond to TLR4 signals (Fig. 3K and Fig. S3C). We were unable to reliably detect TLR4 by flow cytometry on the surface of ECs as its levels were low. However, as signaling through TLR4 on the cell surface requires MyD88, and the genetic data indicate that TLR signaling through MyD88 but not TRIF adaptors regulates IAHC numbers, we conclude that the relevant signaling is occurring through extracellular TLR4 signaling.
In summary, the number of IAHC cells and LMPs at E10.5 are positively regulated by MyD88-dependent TLR signaling (Fig. 4). The number of specified HE cells at E9.5, on the other hand, is not. The selective decrease in LMPs in the absence of MyD88-dependent signaling inversely correlates with an increase in adult-repopulating HSCs, indicating that MyD88-dependent TLR signaling oppositely regulates the numbers of stem cells and progenitors with different potentials (Fig. 4). MyD88-dependent TLR signaling is required in ECs or in the IAHC cells that differentiate from HE cells to regulate IAHC cell numbers. Recent work suggests multipotent progenitors and HSCs arise from distinct subsets of HE cells (Dignum et al., 2021). It is unclear whether loss of TLR signaling impacts the specification of these distinct subsets of HE cells or the formation of committed progenitors and HSCs from them. Alternatively (or additionally) MyD88-dependent TLR signaling may differentially affect the number of LMPs and HSCs by regulating their proliferation. TLR4 signaling in adult HSCs increases proliferation but reduces fitness (Takizawa et al., 2017), and the same may be true for embryonic HSCs.
Haploinsufficiency for RUNX1 decreases the number of IAHC cells and results in the precocious emergence of HSCs in the embryo (Cai et al., 2000). However, RUNX1 haploinsufficiency also decreases the number of HE cells (Zhu et al., 2020), whereas loss of MyD88-dependent TLR signaling does not. HE cells differentiate from a precursor population called pre-HE, and the efficiency at which pre-HE differentiates into HE is regulated by RUNX1 dose (Zhu et al., 2020). We believe the reason loss of MyD88-dependent TLR signaling does not reduce the number of HE cells is that RUNX1 levels are higher in mutant embryos at E9.5 when HE specification occurs. The subsequent decrease in RUNX1 levels in HE and IAHC cells at E10.5 may contribute to the lower number and proliferation of IAHC cells, and reflect a later requirement for RUNX1. RUNX1 promotes the G1- to S-phase transition in several contexts (Cai et al., 2011; Nimmo et al., 2005; Osorio et al., 2008), and loss of RUNX1 decreases proliferation, consistent with these observations.
Our data are consistent with those from zebrafish embryos that demonstrated lower levels of runx1 expression and decreased numbers of HSPCs upon loss of MyD88-dependent TLR signaling (He et al., 2015), but add an extra layer of complexity. They also provide a cautionary note for investigators studying HSC ontogeny – the level of RUNX1 expression, and the numbers of IAHC cells and progenitors do not always correlate with the number of HSCs.
MATERIALS AND METHODS
Mouse lines and embryo production
Tlr4+/− (B6(Cg)-Tlr4tm1.2Karp/J), Tlr4f/f (B6(Cg)-Tlr4tm1.1Karp/J), Csf1rCre (C57BL/6-Tg(Csf1r-cre)1Mnz/J), Ticam1+/− (C57BL/6J-Ticam1Lps2/J) and B6.SJL-Ptprca Pepcb/BoyJ (CD45.1) mice were purchased from Jackson Laboratory. Tg(Cdh5-cre/ERT2)1Rha mice (Sörensen et al., 2009b) were a gift from Ralf Adamson (Max Planck Institute for Molecular Biomedicine, Münster, Germany). Myd88+/− mice were a gift from Edward Behrens (Children's Hospital of Philadelphia, PA, USA).
The morning post-mating is considered embryonic day (E) 0.5. E9.5-E11.5 embryos were staged at the time of harvest by counting somites. Embryos that showed abnormal development were discarded. Mice were handled according to protocols approved by the University of Pennsylvania's Institutional Animal Care and Use Committee and housed in a specific-pathogen-free facility.
Cell sorting and flow cytometry
Antibodies used can be found in Table S1. All antibodies were used at a concentration of 1:200. DAPI, LIVE/DEAD Fixable Aqua (ThermoFisher L34957) or Fixable Viability Stain 780 (BD Biosciences 565388) were used to exclude dead cells. We used the Cytofix/Cytoperm Kit (BD Biosciences) for intracellular flow cytometry to detect TNF.
Yolk sacs were removed from embryos, and vitelline vessels were retained with the embryonic tissue. The head, cardiac and pulmonary regions, liver, digestive tube, tail and limb buds were removed. The remaining tissue containing the aorta-gonad-mesonephros (AGM) region, parts of somite, umbilical and vitelline vessels was collected. Tissues were dissected in phosphate-buffered saline (PBS)/10% fetal bovine serum (FBS) and penicillin/streptomycin (Sigma), followed by dissociation in 0.125% collagenase (Sigma) for 1 h. Tissues were washed and filtered through a 40 μm filter and resuspended in antibody solution. Cells were sorted on a BD Influx equipped with a 100 μm nozzle, run at a pressure of 17 psi with flow rates of less than 4000 events/second, and collected in PBS/20% FBS/25 mM HEPES. Flow cytometry was performed on a LSR II (BD Biosciences) and data were analyzed using FlowJo software (v10.7, BD Biosciences).
Ex vivo LPS stimulation
TNF production by embryonic macrophages was analyzed by incubating dissected AGM regions from E10 embryos ex vivo with 100 ng/ml LPS for 4 h in RPMI medium in the presence of a Golgi inhibitor (BD Biosciences 555029). Collagenase (0.125%) was added for the last 30 min of incubation. After 4 h, dissociated cells were stained with extracellular antibodies, fixed and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences), stained for intracellular TNF, and analyzed by flow cytometry.
T- and B-cell progenitor assays
OP9 and OP9-DLL1 stromal cells were maintained in αMEM containing 20% FBS and antibiotics, and plated into 96-well plates at a concentration of 5500 cells/well 1 day before dilutions of the dissociated embryonic tissues were added. Medium was supplemented with purified recombinant 5 ng/ml Flt3L and 10 ng/ml IL7 (OP9) or 5 ng/ml Flt3L and 1 ng/ml IL7 (OP9-DLL1) (Peprotech). Cells were harvested and analyzed by flow cytometry 12 days after culture. B cells were identified as CD19+ B220+, T cells as CD25+ Thy1.1+. The progenitor frequency was calculated using the method of maximum likelihood applied to Poisson distribution with L-Calc software (StemCell Technologies). OP9s were purchased from ATCC, and OP9-DLL1s were obtained from Juan Carlos Zún̋iga-Pflücker (Sunnybrook Research Institute, Toronto, Canada). Both OP9s and OP-Dll1 cells were tested for mycoplasma and found to be negative.
Limiting dose HE cell assay
OP9 stromal cells were cultured in αMEM with 20% FBS and antibiotics. One day before sorting, OP9s were plated at a concentration of 5000 cells/well in a 96-well plate. Tlr4+/− and Tlr4−/− embryos generated from a cross of Tlr4+/− and Tlr4−/− mice were used in this experiment in order to have a sufficient number of endothelial cells for the analysis. Endothelial cells were sorted on a BD Influx into individual wells with a maximum of five cells sorted into each well. Medium was supplemented with 10 ng/ml Flt3L, SCF, IL3 and IL7 (Peprotech). Cells were cultured for 10 days before being scored by eye for hematopoietic cells, and positive wells were harvested and analyzed by flow cytometry. Positive wells contained CD45+, CD41+, Mac1+, B220+ and/or Ter119+ cells.
Yolk sacs were removed and intact embryos placed in αMEM containing 20% FBS and antibiotics. Embryos were incubated in a 24-well plate at 37°C for 2 h in EdU (5-ethynyl-2′-deoxyuridine, 100 µM). Immediately following incubation with EdU, embryos were dissociated using 0.125% collagenase (Sigma) for 30 min and stained for flow cytometry. EdU detection was performed using the Click-iT Plus EdU Flow Cytometry Assay kit (Thermo Fisher Scientific, C10632).
B6.SJL-PtprcaPepcb/BoyJ (CD45.1) mice were subjected to a split dose of 900 cGy, 3 h apart. Each recipient received 1.0 or 0.3 embryo equivalents (ee) of E11.5 AGM regions (CD45.2), vitelline and umbilical arteries with 2.5×105 carrier spleen cells by retro-orbital injection. We assessed donor (CD45.2) engraftment in peripheral blood at weeks 4, 8, 12 and 16, and in bone marrow at week 16 post-transplantation. HSC frequencies were determined by extreme limiting dilution analysis (ELDA) (Hu and Smyth, 2009).
Embryos were prepared as previously described (Yokomizo et al., 2012). A Zeiss LSM 880 AxioObserver inverted confocal microscope with ZEN 2.3 SP1 software was used to acquire Z- and single optical projections. Images were processed using Fiji software (Schindelin et al., 2012). All primary antibodies were used at a concentration of 1:500, except for KIT, which was used at 1:250. All secondaries were used at a concentration of 1:1000. Corrected total cell fluorescence (CTCF) was calculated to measure RUNX1 intensity in HE and IAHC cells by measuring the integrated density value for each cell using FIJI software. CTCF was calculated as [integrated density of cell−(area of selected cell×mean intensity of background)]. Power analysis determined that detection of a 10% change in the intensity of RUNX1 levels at P≤0.01 required 220 cells per genotype.
Significance for multiple comparisons was determined using ordinary one-way ANOVA with Tukey's test or two-way ANOVA with Sidak's correction for multiple comparisons. A two-tailed Student's t-test was used for pairwise comparisons. All statistics were performed using Prism software (v. 9.0.0). For all tests, *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001.
We thank Andrea Stout and Jasmine Zhao at the Cell and Developmental Biology Microscopy Core for confocal microscopy assistance. We thank Andrew Morschauser and Charles Pletcher at the Penn Cytometrics and Cell Sorting Resource Laboratory for their assistance.
Conceptualization: N.A.S.; Formal analysis: L.F.B., M.D.M., Y.L.; Investigation: L.F.B., M.D.M., Y.L.; Writing - original draft: N.A.S.; Writing - review & editing: L.F.B., M.D.M., Y.L.; Project administration: N.A.S.; Funding acquisition: N.A.S.
This work was supported by the National Institutes of Health (R01HL091724, T32HL007439 to L.F.B and T32DK007780 to M.D.M.). Deposited in PMC for release after 12 months.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200025.
The authors declare no competing or financial interests.