Biomechanical forces promote blood development through prostaglandin E₂ and the cAMP–PKA signaling axis

Previously, we described specific characteristics of WSS that induce hematopoietic progenitor activity in two-dimensional adherent cultures of dissociated AGM (Adamo et al., 2009). In the present study, we hypothesized that WSS may play a role in specification of HSCs that support life-long adult hematopoiesis and so examined developmental time points that precede definitive HSC emergence in cells derived from the anlage of the AGM known as the PSp at embryonic day (E) 9.5 and from the AGM at E10.5. Runx1 and Myb, two master regulators of hematopoiesis, are transcriptionally up-regulated by shear stress typical of embryonic murine blood flow (5 dyn/cm²), as are other genes required for definitive hematopoiesis and lymphopoiesis, including Bcl11a, Etv6, Gata2, Gata3, Tcf, Rag1, and Rag2 (Fig. 1 A). Analysis of cell surface phenotype after WSS confirmed increases in two markers of hemogenic endothelium, CD144/VE-Cadherin and c-kit, in the live (DAPI⁻) population (Fig. 1 B). We observed a 5.2 ± 1.2–fold increase in the percentage of CD144⁺ ckit⁺ cells, a surface phenotype thought to distinguish a subset of endothelial cells with definitive HSC potential (Fig. 1 C; Eilken et al., 2009; Swiers et al., 2013).

E9.5 PSp has previously been found to produce primarily B1a lymphocyte progenitors when delivered into immunocompromised neonatal recipients (Yoshimoto et al., 2011) or intact under the adult kidney capsule (Godin et al., 1993) and requires whole organ culture for several days to acquire long-term multilineage repopulating activity in vivo (Cumano et al., 2001). To evaluate whether changes induced by WSS were capable of producing greater functional competence in stem and progenitor cells, we transplanted dissociated PSp from E9.5 embryos (21–29 somite pairs) cultured with or without WSS for 36 h into irradiated adult Rag2^−/− Il2rγ^−/− recipients. Based upon the expectation that dissociated E9.5 PSp would provide absent to scant engraftment of adult recipients long term (Godin et al., 1993; Cumano et al., 2001), we transplanted each recipient with 16 embryo equivalents (e.e.). We detected low-level engraftment in the peripheral blood of 4 of 17 animals injected with E9.5 PSp cultured under static conditions, which dropped further by 10 wk in the surviving 10 animals, suggesting that the majority of engrafting cells were exhaustible short-term progenitors (Fig. 2, A and B). At 20 wk, static cultures contributed to <1.4% of total peripheral blood leukocytes in all but one recipient (2.8% CD45.2⁺). In contrast, WSS produced higher initial engraftment levels in 6 of 15 recipients, which persisted long term in 5 animals (Fig. 2 B). PSp contributed to major adult blood lineages, including B lymphoid (B220, CD19, and IgM), T lymphoid (CD3, CD4, and CD8), and myeloid (Mac1 and Gr1) populations (Fig. 2 C). WSS exposure promoted earlier emergence of B lymphocytes and sustained B lymphopoiesis in the periphery beyond 20 wk (Fig. 2, D and E). Greater B cell maturity was also apparent in the bone marrow (Fig. 2 F). Specifically, greater numbers of cells were in stages of pre-pro- and pro-B commitment (B220⁺ CD43⁺), with modest increases in late-stage pre-B and B development (B220⁺ CD43⁻) typical of adult-type bone marrow progenitors (Fig. 2 G; Hardy and Hayakawa, 1991). CD45.2⁺ bone marrow cells were subsequently transferred to secondary recipients and found to reconstitute four of seven WSS and two of five static recipients (Fig. 3, A and B). Importantly, peak chimerism was substantially higher for three of the WSS donors (4.2%, 12.2%, and 61.3%) as compared with static donors (0.9% and 2.1%). B and T cells were detectable as discrete populations and persisted long term in WSS recipients (Fig. 3, C and D). Collectively, these data suggest that WSS serves as a critical developmental signal that promotes commitment of precursors to the blood lineage and endows nascent HSCs and progenitors with the functional competence necessary to engraft the adult hematopoietic niche.

Knockout of the Ncx1 gene that encodes the cardiac-specific sodium/calcium exchanger results in mutant embryos with no heartbeat and little to no hematopoietic progenitor activity (Lux et al., 2008). Development of mutant embryos proceeds up to 10 d of gestation, at which time cardiac defects result in embryonic lethality. Previously, we showed that hematopoietic activity could be rescued ex vivo after 36 h of WSS in cells derived from E9.5 Ncx1^−/− PSp (Adamo et al., 2009). We therefore hypothesized that hemogenic endothelium was present in Ncx1^−/− PSp and measured the frequency of cell surface markers typical of hematopoietic precursors and progeny. We found that Ncx1^−/− embryos harbor phenotypic hemogenic endothelium and hematopoietic cells (Fig. 4 A). Indeed, the frequency of CD144⁺ ckit⁺ CD45⁻ hemogenic endothelium was elevated in mutant PSp by ∼4.1 ± 1.1–fold (Fig. 4 B). As this contrasts with our observations of enhanced CD144⁺ ckit⁺ CD45⁻ cell production in ex vivo WSS cultures, we speculate that accumulation of hemogenic endothelial cells in the mutant PSp could be a consequence of defective circulation, a phenomenon demonstrated previously to lead to pooling of primitive erythrocytes during development of zebrafish embryos (Iida et al., 2010). We thus analyzed gene expression in Ncx1^−/− PSp and found decreases in Runx1, Gata2, Tal1, Etv2, and effectors of Notch (Hes1 and Hey1) and Wnt (Wnt3a and Lef1), supporting a role for WSS in determination of signaling events in the PSp (Fig. 4 C).

We and others previously identified NO as a critical regulator of hematopoietic progenitor expansion in response to WSS and blood flow (Adamo et al., 2009; North et al., 2009; Wang et al., 2011). To more precisely define the signaling mechanisms triggered by WSS, we conducted global gene expression profiling of ex vivo cultures of AGM. Notch, Wnt, and eicosanoid signaling have previously been found in proteomic analyses to respond rapidly to fluid shear stress in aortic endothelial cells (Wang et al., 2007). Informed by these studies and our own serial measurements of gene subsets, we selected an early time point (6 h) and one demonstrated previously to enhance hematopoietic activity (36 h) to perform global gene expression analysis (Adamo et al., 2009). In brief, WSS was applied to dissociated E10.5 AGM cultured within microfluidic channels as described previously (Li et al., 2014), replicates were lysed, and RNA was processed for analysis by Illumina Mouse WG-6 v2.0 Expression BeadChips (45,200 transcripts). Unsupervised hierarchical clustering of unfiltered genes positioned 6- and 36-h cultures in two distinct groups, followed by segregation of static and WSS cultures (not depicted). Differential gene expression analysis (P < 0.01, twofold threshold) revealed significant change in 1,435 unique transcripts at 6 h, 347 at 36 h, and 109 common to both time points (Fig. 5 A and Dataset S1). By Ingenuity pathway analysis, these genes were found to encode enzymes and transcription regulators required for signaling through several overlapping pathways, including G protein–coupled receptors (GPCRs), calcium transport, nuclear factor of activated T cells (NFAT), phospholipase C, and PI3K (Btk, Chp1, Nfatc3, and Plcb3; Fig. 5 B). Components of the NFκB signaling pathway were also enriched, including the TRAF family member associated NFκB activator (Tank) and IKBε regulatory subunit (Nfkbie). Several other kinases, enzymes, and transcription factors were found to contribute to signaling through CREB, such as the α regulatory subunit of cAMP-dependent protein kinase (Prkar1a), G protein–binding protein (Gnb1l), Creb5, and CREB-binding protein (Crebbp). Wnt/β-catenin regulators and ligands were also enriched (Akt3, Dkk1, Wnt1, and Wnt7b). Functional enrichment analysis of unfiltered genes by GSEA (Mootha et al., 2003; Subramanian et al., 2005) was used to construct a global network wherein overlapping gene sets clustered together (Merico et al., 2010) and connections between pathways could be visualized at 6 and 36 h in parallel (Fig. 5 C). Wnt signaling functionally clustered with the adaptive immune system and cell cycle, whereas Notch, TGF-β, and EGFR/CREB/MAPK clustered as distinct groups. Notably, several functional groups that emerged from network analysis are known regulators or effectors of prostaglandin production and signaling, including calcium, MAPK, CREB, Wnt, NFκB, and biological oxidation involving prostaglandin-endoperoxide synthase function of COX1 and COX2 (Ptgs1 and Ptgs2; Fig. 5 C and Dataset S2; Tsatsanis et al., 2006; Goessling et al., 2011).

PGE₂ has recently been documented to promote hematopoietic engraftment in mice and in human clinical trials (North et al., 2007; Hoggatt et al., 2009, 2013a,b; Cutler et al., 2013). Up-regulation of prostaglandin synthases Ptgs1, Ptgs2, Ptges, and Ptges3, as well as Wnt and Notch were verified by quantitative RT-PCR (qRT-PCR) in WSS-exposed populations of the E10.5 AGM, including the CD144⁺ CD45⁻ population identified as the hemogenic endothelium (Fig. 6, A and B). Importantly, we found that WSS promoted production and secretion of significantly higher levels of PGE₂ (Fig. 6 C). This enhanced PGE₂ production could be blocked by the COX1/2 antagonist indomethacin and the COX2-specific inhibitors NS-398 and CAY10404, suggesting that Ptgs2 up-regulation may be responsible for elevated PGE₂ production in WSS-exposed AGM (Fig. 6 D). In osteoblasts, calcium-dependent activity at focal adhesions is believed to regulate COX2 and PGE₂ synthesis (Ponik and Pavalko, 2004). We therefore hypothesized that WSS may also stimulate calcium flux upstream of PGE₂ production in the AGM. Live cell imaging with Fluo-4 AM confirmed sparks of intracellular calcium signaling in WSS-stimulated cells, as determined by increased intensity in cytoplasmic fluorescence (Fig. 6, E and F; and Videos 1 and 2). Although calcium flux was evident in single cells within static cultures, exposure to WSS induced more intense signaling across a greater number of cells. Sequestration of calcium by BAPTA-AM reduced WSS-dependent induction of Ptgs2 transcript (Fig. 6 G). In static cultures, pharmacological elevation of intracellular calcium concentration by pulse treatment of the calcium ionophore A23187 also stimulated accumulation of COX2 protein (Fig. 6 H). Observations in ex vivo cultures of AGM were further corroborated by reduction in COX2 protein levels in uncultured Ncx1^−/− PSp (Fig. 6 I).

To evaluate the contribution of PGE₂ to intracellular signaling, we subjected cells to WSS with or without indomethacin and profiled global gene expression as described above. Analysis of differential gene expression revealed that PGE₂ contributed to transcription of CREB targets, as well as core components of Wnt/calcium signaling, including Creb1, Ctbp1, Shc1, Calm2, and Camk1 (Fig. 7 A). In chondrocytes of the bone, WSS was recently shown to promote TOP-luciferase (LEF/TCF reporter) and CRE-luciferase (CREB reporter) activity (Ogawa et al., 2014). Whereas addition of exogenous PGE₂ was insufficient to drive activation of TOP-luciferase, PGE₂ was capable of promoting activity of the CRE-luciferase reporter. Consistent with these findings, we observed that WSS increased phosphorylation of CREB at serine 133, a posttranslational modification required for CREB-mediated transcription (Fig. 7 B). CREB phosphorylation can be mediated by PKA; thus, the increase in PKA transcript Prkaca by WSS at 6 and 36 h (unpaired Student’s t test, P = 0.016 [6 h] and P = 0.004 [36 h]) suggested that modulation of PKA downstream of WSS could contribute to enhanced CREB activity. Furthermore, concomitant reduction in Prkaca transcript, the prohematopoietic CREB target Etv2, and P-CREB levels in uncultured Ncx1^−/− PSp independently suggested a role for blood flow in coordinating PKA and CREB activity (Figs. 4 C and 7 C). Of six Ncx1 mutants analyzed in the study, five expressed lower levels of P-CREB relative to wild-type littermates. Total CREB protein was also reduced in four of six mutants, likely indicative of complex signaling inputs known to regulate CREB expression and activity (Shaywitz and Greenberg, 1999). Binding of the EP₂ and EP₄ receptors by PGE₂ stimulates the G_s-adenylyl cyclase pathway, which leads to cAMP production and subsequent PKA activation (Zhang and Daaka, 2011; Yamamizu et al., 2012). As expected, AGM exposed to WSS for 36 h produced higher levels of intracellular and circulating cAMP compared with static cultures (Fig. 7 A). Importantly, antagonists of COX2 or PKA abolished enhancement in hematopoietic colony forming activity with ex vivo exposure to WSS, as did a blocking antibody to PGE₂ (Fig. 7 E). The adenylyl cyclase agonist forskolin amplified colony-forming potential in WSS-exposed cultures but not static cultures, suggesting that PGE₂ cooperates with additional WSS-responsive signaling in initiation of the hematopoietic program. Consistent with this notion, inhibition of COX2 by indomethacin blocks WSS induction not only of CREB and MAPK signaling but also calcium, calmodulin, calcineurin, and NFAT, primary components of noncanonical Wnt/calcium signaling that function to control cell adhesion, migration, and ventral patterning (Fig. 7 F and Dataset S3; Kohn and Moon, 2005). Based on analyses of BrdU incorporation and Annexin V staining, WSS contributes significantly to survival of the CD144⁺ CD45⁻ population (Fig. 8). However, PGE₂ does not appear to influence cell proliferation or survival, but instead may serve to modulate fate selection through CREB. More detailed analyses of signal transduction will be required to understand the precise nature of how PGE₂–PKA interfaces with other developmental signaling pathways to modulate hematopoietic-specific gene expression downstream of WSS in the AGM.

The HSC enhancing ability of dmPGE₂ was first identified in pharmacological screens of zebrafish embryos and has since been used for expansion of umbilical cord blood in phase I clinical trials (North et al., 2007; Cutler et al., 2013). To date, the effects of dmPGE₂ in mouse and human have only been tested on committed hematopoietic stem and progenitor cells. We therefore evaluated the ability of transient dmPGE₂ treatment to promote short- and long-term hematopoietic activity from E10.5 and E11.5 murine AGM. Colony formation assays showed an asymptotic dose-dependent increase in progenitor activity with up to 10 µM dmPGE₂ (Fig. 9 A). 2-h dmPGE₂ treatment also increased chimerism in the peripheral blood from donor AGM above mean levels in three of five recipients of E10.5 AGM and in six of eight recipients of E11.5 AGM (Fig. 9, B and C). Notably, dmPGE₂ improved B and T lymphoid potential of hematopoietic progenitors/stem cells at both embryonic stages, although E11.5 produced the only HSC-like long-term engraftment of all lineages (Fig. 9, D and E). Collectively, our findings suggest that WSS and PGE₂ function together to regulate developmental signaling that determines lineage potential of nascent hematopoietic cells in the earliest stages of definitive hematopoiesis.

Here, we have shown that WSS activates developmental pathways that promote hematopoietic fate and potentiates long-term engraftment of embryonic hematopoietic stem and progenitor cells from the PSp and AGM. Prostaglandin, Wnt, and Notch signaling were all up-regulated in response to WSS. PGE₂ production in particular was found to be required for WSS-enhanced progenitor activity, as its stimulatory effects could be inhibited by blocking antibodies or pharmacological inhibition of COX enzyme function. PGE₂ was directly responsible for increased cAMP production and contributed to regulation of CREB through PKA. Mimicry of WSS by addition of dmPGE₂ induced greater peripheral blood chimerism and enhanced lymphoid potential from E10.5 and E11.5 AGM, supporting a possible mechanism for transduction of mechanical cues into chemical signaling and downstream gene activation within HSCs of the embryonic vasculature.

Prostaglandin signaling has emerged as an attractive target for enhancing hematopoietic function in cell therapy and as a medical countermeasure against radiological and nuclear threats (Goessling et al., 2011; Cutler et al., 2013; Hoggatt et al., 2013c). In animal models and humans, PGE₂ has been found to regulate survival, self-renewal, and trafficking of HSCs and progenitors (Hoggatt et al., 2009, 2013a). Ablation of Ptgs2 or the terminal prostanoid synthase Ptges results in a variety of hematopoietic defects, impacting macrophages and platelets, and widespread impairment in inflammatory response (Uematsu et al., 2002; Cheng et al., 2006; Yu et al., 2007; Seta et al., 2009). Furthermore, PGE₂ has very recently been found to possess morphogen-like effects on bipotential endodermal precursors that determine fate selection between hepatic and pancreatic commitment (Nissim et al., 2014), raising the possibility that PGE₂ and downstream PKA activation could direct specification of hematopoietic lineages from precursors of the hemogenic endothelium, perhaps in part through modulation of Wnt signaling.

Our data support a model in which multiple pathways downstream of WSS must converge to balance developmental signaling essential for hematopoiesis. In addition to Wnt and Notch, other pathways altered by WSS corroborate findings from Goessling et al. (2011), in which pulse treatment of human and macaque CD34⁺ cells with dmPGE₂ induced pathways such as PKA, NFAT, and ERK/MAPK. Mechanosensation in hemogenic endothelium and other cells of the AGM involves release of second messengers into the cytosol, including intracellular calcium, that directly stimulate enzyme functions required for prostaglandin synthesis. In osteocytes, calcium flux triggers rapid PGE₂ synthesis and secretion through a biochemical cascade that includes PLC, DAG, PKC, and PLA₂ (Ajubi et al., 1999). Prostaglandin production has also been found in osteoblasts to be dependent on focal adhesions and transcriptional transactivation of Ptgs2 by transcription factors, including C/EBP β, AP-1, and CREB (Ogasawara et al., 2001; Ponik and Pavalko, 2004; Rangaswami et al., 2012). Prostaglandin synthesis is elevated in CD144⁺ CD45⁻ hemogenic endothelial cells, but PGE₂ signaling could also originate from other cell types in the AGM, such as committed CD144^+/− ckit⁺ hematopoietic progenitors (unpublished data). As PGE₂ is released into the aortic lumen, activation of EP₂/EP₄ receptors expressed on hemogenic endothelium or committed hematopoietic stem and progenitor cells stimulates G protein–dependent adenylyl cyclase activity and elevates intracellular cAMP levels. PKA, in the presence of cAMP, can then modulate CREB-dependent gene expression. Interestingly, genetic ablation of CREB phenocopies Wnt1 and Wnt3a knockout in differentiating myoblasts, suggesting that activation of CREB by PKA is a prerequisite for transduction of some noncanonical Wnt signals (Chen et al., 2005). We find that WSS stimulates CREB phosphorylation and up-regulates core components of noncanonical Wnt signaling, including calmodulin, calcineurin, NFAT, and calmodulin-dependent kinase, which are known to play an important role in self-renewal and repopulating activity of adult HSCs (Nemeth et al., 2007; Sugimura et al., 2012). Wnt receptor and ligand transcript levels were statistically unchanged by indomethacin or the PKA antagonist H89, suggesting that Wnt machinery at the cell membrane is regulated independently of PGE₂ but that target gene activation may be regulated by WSS-induced phosphorylation of CREB (unpublished data). The nature of the relationship between PGE₂ and NFAT will require further investigation, though it appears that PGE₂ may contribute to signaling activity through calcium, calmodulin, calmodulin-dependent kinase, and NFAT. CREB and NFAT are essential for transduction of noncanonical Wnt signals; thus, regulation of the activity of these transcription factors may allow PGE₂ to “fine tune” the dosage of canonical Wnt signaling required for expansion and maintenance of HSCs (Luis et al., 2011).

The identification of biomechanical cues that support hematopoiesis has begun to redefine our understanding of the types of signals that comprise the hematopoietic niche. Data from the current study demonstrate that PGE₂ contributes essential signaling downstream of WSS to govern the expansion of hematopoietic populations with long-term reconstitution potential in the developing embryo. Moreover, we find that WSS-induced PGE₂ acts through calcium and cAMP–PKA to regulate CREB for induction of prohematopoietic developmental programs. Further studies will be required to precisely address how biomechanical forces coordinate these and other signaling pathways in the PSp and AGM to define hematopoietic potential.

Timed pregnancies were bred in house to E9.5–11.5 (C57BL/6J or Ncx1) or were purchased as pregnant females from Taconic at E9.5 for large-scale shearing experiments (Swiss Webster). Gestational age of embryos was determined by observation of a copulation plug on E0.5 and number of somite pairs. Rag2^−/− Il2rγ^−/− mice were back-bred to C57BL/6 SJL (Pep Boy) to produce CD45.1⁺ Rag2^−/− Il2rγ^−/− mice and were maintained by inbreeding to CD45.2⁺ Rag2^−/− Il2rγ^−/−. Ncx1 knockout mice were maintained on a C57BL/6J background (minimum ninth generation). All animal experiments were performed according to the University of Texas Medical School at Houston and Children’s Hospital Boston guidelines for laboratory animals.

Embryos from C57BL/6J or Swiss Webster timed pregnancies were microdissected for isolation of E9.5 PSp or E10.5 AGM regions. Tissues were dissociated either by treatment with Accutase (STEMCELL Technologies) at room temperature with gentle agitation for 20 min or, for WSS transplantation experiments, by 0.1% dispase (Gibco) at 37°C for 30 min. Cells were then resuspended in M5300 MyeloCult medium (STEMCELL Technologies) enriched with Hepes (Invitrogen; 12.5 ml of 1 M Hepes per 500 ml MyeloCult medium), nonessential amino acids (Gibco), sodium pyruvate (Gibco), 10 U/ml penicillin (Lonza), and 10 µg/ml streptomycin (Lonza) without addition of hydrocortisone. Cells were plated on tissue culture plastic coated with a 1:20 Matrigel-PBS solution (BD) at a concentration of 8 e.e. per 10 cm² in preparation for culture on a dynamic flow system (Adamo et al., 2009) or in microfluidics (Li et al., 2014), as described previously. Microfluidics consisted of IBIDI VI^0.4 channel slides in line with a recirculating medium system driven by a single Harvard Apparatus PHD ULTRA 4400 with remote syringe pump, 0.27 PSI crack pressure check valves (Qosina), and female luer lock three-way stop cocks (Qosina). After 6.5 h of incubation, nonadherent cells were washed away and cells were cultured in enriched MyeloCult medium with 10 ng/ml murine SCF (PeproTech) in static/low-flow conditions (<0.0001 dyn/cm²) or in the presence of WSS (5 dyn/cm²). The shear stress pattern used to induce hematopoiesis is described in our previous work (Adamo et al., 2009). At the end of the culture period, adherent and nonadherent cells were collected by a 5-min incubation in Accutase, washed, resuspended in MyeloCult, and transferred to functional or phenotypic assays. Cells analyzed immediately by flow cytometry were immunostained with antibodies available through BD, purified rat CD144 and APC c-kit (CD117); eBioscience, APC-Cy7 CD45.2; and Invitrogen, Alexa Fluor 488 anti–rat IgG. Colony formation assays were established in M3434 MethoCult medium (STEMCELL Technologies) at 100,000 cells per 1.5 ml media and counted after 10–14 d.

Cells were cultured with pharmacological compounds or antibodies to block PGE₂ production and/or signaling by incubation with 10 µM indomethacin, 10 µM NS-398, 1 µM CAY10404 (Cayman Chemical), or 1:1,000 anti-PGE₂ antibodies (Abcam) for the duration of culture with or without WSS. H89 (Cayman Chemical) was applied at 10 µM to inhibit PKA activity. Forskolin (Sigma-Aldrich) was administered at 10 µM as an adenylyl cyclase activator to stimulate cAMP production. The stabilized synthetic analogue dmPGE₂ (Cayman Chemical) was applied to freshly dissociated AGM cells for 2 h at 37°C from 1 to 100 µM for colony formation assays or at 10 µM for transplantation into adult recipient mice. Intracellular calcium was sequestered by 10 µM BAPTA-AM (Tocris Bioscience) and elevated by 2 µM of calcium ionophore A23187 (Tocris Bioscience).

Total RNA was extracted with QIAGEN RNeasy kits for analysis of gene expression by Illumina Mouse WG-6 v2.0 Expression BeadChips or by TaqMan Assays. RT of RNA for qRT-PCR was performed using Multiscribe DNA polymerase (Applied Biosystems), and PCR was performed in 10-µl reactions on a 7500 Real-Time PCR System (Applied Biosystems). All procedures were conducted according to the manufacturer’s instructions. Illumina data were checked for quality, background corrected, and quantile normalized with GenomeStudio (Illumina). Analysis of differential expression was conducted with GenomeStudio for probe sets filtered by spot detection (P < 0.01) and difference in mean signal intensity between treatment groups (dif > 20). Significantly changed genes were subjected to analysis of canonical pathways curated by Ingenuity (IPA, Ingenuity Systems). Gene set expression analysis (GSEA) was conducted with unfiltered genes using curated gene lists downloaded from MSigDB. Ranked gene list data were imported into Cytoscape for visualization of functional enrichments with Enrichment Map using edge-weighted force-directed layout. Heat maps were generated using genes selected from differential gene expression analysis in GenePattern. Expression data have been deposited for public access in the NCBI Gene Expression Omnibus (GEO) under accession no. GSE62463.

E10.5 AGM were dissociated and exposed to WSS or static culture conditions for up to 110 h. Cell lysates and/or medium was processed for measurement with the Prostaglandin E₂ Express EIA kit (Cayman Chemical) or the Direct cAMP ELISA kit (Enzo Life Sciences). Standard curves were established in parallel for quantification of sample concentrations.

Microdissected PSp from 64 E9.5 embryos for WSS experiments were pooled and dissociated in 0.1% dispase-PBS with calcium and magnesium. PSp cells were cultured as described above and prepared for transplantation by gentle dissociation with 0.1% dispase at 37°C for 20 min. CD45.1⁺ or CD45.1⁺ CD45.2⁺ Rag2^−/− Il2rγ^−/− recipient mice were lethally irradiated with a split dose of 9.25 Gy (separated by 3 h) just before transplantation. Each primary recipient received 16 e.e. of E9.5 PSp-derived cells plus 500,000 CD45.1⁺ CD45.2⁺ Rag2^−/− Il2rγ^−/− whole bone marrow competitor cells. Secondary recipients received 20,000 CD45.2⁺ cells after sorting from primary recipient bone marrow and 200,000 adult competitor marrow cells.

For transplantation of dmPGE₂-treated AGM cells, embryos were microdissected, dissociated at room temperature with Accutase, resuspended in 10 µM dmPGE₂-containing MyeloCult media, and incubated at 37°C for 2 h. Cells were then washed in PBS without calcium and magnesium, filtered through a cell strainer, and injected retroorbitally into adult CD45.1⁺ or CD45.1⁺ CD45.2⁺ Rag2^−/− Il2rγ^−/− recipients. Each recipient of dmPGE₂-treated cells received a split dose of 8.25 Gy irradiation (separated by 3 h), no competitor, and 4 e.e. of E10.5 AGM or 2 e.e. of E11.5 AGM.

Peripheral blood was collected at 4, 6, 10, and 20 wk after transplantation. Red blood cells were removed by 1% dextran sulfate–PBS-EDTA separation and treatment with RBC lysing buffer (Sigma-Aldrich). Leukocytes were immunostained for discrimination of CD45 allelic variants and detection of lineage markers, followed by analysis on a five-laser BD LSR II flow cytometer. Antibodies used from BD included FITC CD45.1, biotin CD8a, biotin CD4, biotin CD3e, APC CD45.2, PE-Cy7 IgM, APC-Cy7 CD45R/B220, and PE Ly-6G/Ly-6C (Gr-1). We also used Pacific Blue CD11b (BioLegend), PE-Cy5 CD19 (eBioscience), Pacific Orange streptavidin (Invitrogen), and DAPI for live-dead discrimination (Sigma-Aldrich). Engraftment was measured as the percentage of CD45.2 cells within the CD45⁺ population. Peripheral blood from untransplanted CD45.1⁺ CD45.2⁺ Rag2^−/− Il2rγ^−/− mice was used to evaluate background in the CD45.2 gate.

Bone marrow was collected at 20-wk after transplantation for analysis of chimerism and B lineage maturation. In brief, bones were crushed in PBS with a mortar and pestle and filtered through a 70-µm cell strainer. The resulting cell suspension was subjected to RBC lysis and incubation with antibodies from BD: APC-Cy7 CD45.1, APC CD45.2, PE-Cy5 CD45R/B220, FITC CD24, PE CD25, PE-Cy7 IgM, biotin CD43, and Pacific Blue streptavidin. Cells were sorted and analyzed on a 5-laser BD Aria II special sorter. Limiting numbers of CD45.2⁺ cells were subsequently transferred to lethally irradiated secondary recipients for further evaluation of long-term reconstitution potential, as described in the transplantation section above.

Cells were harvested in RIPA buffer (150 mM sodium chloride, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 7.5, and 2 mM EDTA) with 1% protease and phosphatase inhibitor cocktails (Sigma-Aldrich). Equal amounts of protein were separated by SDS-PAGE and analyzed by immunoblotting with standard procedures. Antibodies used for immunodetection included COX2 (Abcam), phospho-CREB (Ser133; Cell Signaling Technology), CREB (Cell Signaling Technology), and β-actin (Santa Cruz Biotechnology, Inc.). Gel images were scanned and the density of the protein bands was quantified as a ratio to the actin loading control by MCID Analysis 7.1 software (Imaging Research).

Dissociated E11.5 AGM cells were plated at lower density (120,000 cells/cm²) for discrimination of calcium sparks in individual cells. In brief, adherent cells were washed with isotonic Tyrode’s solution (139 mM NaCl, 3 mM KCl, 17 mM NaHCO₃, 12 mM d-glucose, 3 mM CaCl₂, and 1 mM MgCl₂) and incubated with the fluorescent calcium indicator 5 µM Fluo-4 AM (F14201; Invitrogen) in isotonic Tyrode’s solution for 5 min at 37°C. Fresh MyeloCult medium was applied, and cells were placed in an environmental chamber maintained at 37°C, 5% CO₂ for imaging. Successive images were collected at a 300-ms time interval for 60 s (201 images total) using MetaMorph v7.7.9.0 on an Olympus IX81 fluorescent microscope equipped with an Andor iXon X3 885 EMCCD camera. WSS was applied 3 s after initiation of image acquisition and thus represents static culture at start time. Images were subsequently analyzed for integrated intensity using manual selection of individual cell boundaries with MetaMorph software. Videos were compiled using ImageJ 1.46r software (National Institutes of Health) using 10 frames per second for visualization.

E10.5 AGM were dissociated and exposed to WSS or control culture conditions for 36 h. After the culture period, adherent and nonadherent cells were collected with Accutase and washed in MyeloCult media. Cells were immunostained on ice with each antibody in 2% FBS-PBS for 20 min, first with rat anti–mouse VE-Cadherin (BD) antibody and then with Alexa Fluor 488 anti–rat IgG (Invitrogen). Cells were washed and additional cell surface markers were detected with APC c-kit (CD117; BD) and APC-Cy7 CD45.2 (eBioscience). Annexin V binding buffer was used for subsequent washes and for incubation with PE-Cy7 Annexin V (eBioscience) at room temperature for 15 min. Cells were resuspended in binding buffer containing 1 µg/ml DAPI before analysis on a three-laser LSR II flow cytometer.

BrdU was added to a final concentration of 10 µM after culture for 34 h under WSS or static conditions. WSS or static culture was maintained for an additional 2 h, and then cells were washed in PBS and recovered by Accutase treatment. Cells were incubated sequentially in 2% FBS-PBS for 20 min with rat anti–mouse VE-Cadherin, Alexa Fluor 488 anti–rat IgG, and APC-Cy7 CD45.2. Cells were then fixed and stained according to the manufacturer’s instructions for the BD BrdU Flow kit. In brief, cells were fixed on ice for 15 min in Cytofix/Cytoperm buffer, washed in Perm/Wash buffer, and incubated for 10 min on ice in Cytoperm Permeabilization buffer. Cells were fixed again in Cytofix/Cytoperm buffer and subsequently treated with 300 µg/ml DNase for 1 h at 37°C. Cells were then stained in Perm/Wash buffer containing APC anti-BrdU antibody at room temperature for 20 min and resuspended in 10 µg/ml DAPI solution for analysis on an LSR II flow cytometer.

All data were analyzed with SigmaPlot 12.5 for statistical significance and are reported as mean ± SEM. Differences in gene expression, cell surface labeling, and colony formation were analyzed with the unpaired Student’s t test or with the Mann–Whitney Rank Sum test where assumptions of normality and homoscedasticity were not met. Two-way ANOVA and the Holm-Šídák method for multiple comparisons were used to evaluate differences in peripheral blood chimerism over time, drug effects on colony formation, and changes in levels of PGE₂ and cAMP. Significance levels of P < 0.05 and P < 0.01 are denoted in graphs by a single asterisk (*) or double asterisks (**), respectively. Representative results from at least three independent biological replicates are shown unless stated otherwise.

Videos 1 and 2 show calcium imaging in live cell cultures. Dataset S1, included as a separate Excel file, contains differential gene expression analysis. Datasets S2 and S3, included as separate Excel files, include lists of gene sets within functional clusters of enrichment analysis between static and WSS or WSS vehicle– and WSS indomethacin–treated AGM cells, respectively.

We thank Drs. S. McKinney-Freeman, C. Dessauer, G. Heffner, A.-L. Tsai, P. Kim, and D. Shah for critical discussions and S. Lazo-Kallanian, A. Hazen, and R. Mathieu for flow cytometry.

This work was funded by grants from the American Society of Hematology (to P.L. Wenzel), State of Texas Emerging Technology Fund (to P.L. Wenzel), and National Institutes of Health (to P.L. Wenzel, G.Q. Daley, and G. García-Cardeña). L. Adamo was partially funded by the Giovanni Armenise–Harvard Foundation. G.Q. Daley is an Investigator of the Howard Hughes Medical Institute.

The authors declare no competing financial interests.

Author contributions: M.F. Diaz, N. Li, H.J. Lee, L. Adamo, S.M. Evans, H.E. Willey, N. Arora, Y.-s. Torisawa, D.A. Vickers, O. Naveiras, and S.A. Murthy designed and performed experiments and analyzed the data. D.E. Ingber and S.A. Morris provided guidance on microfluidic WSS systems. G.Q. Daley and G. García-Cardeña conceived the study, directed the research, and revised the manuscript. P.L. Wenzel designed the study, performed experiments, analyzed the data, wrote the manuscript, and directed the research.