Skip to main content
NCBI home page
Stem Cell Reports logoLink to Stem Cell Reports
. 2018 Sep 11;11(3):784–794. doi: 10.1016/j.stemcr.2018.08.004

Analysis of Runx1 Using Induced Gene Ablation Reveals Its Essential Role in Pre-liver HSC Development and Limitations of an In Vivo Approach

Jordi Senserrich 1,3, Antoniana Batsivari 1, Stanislav Rybtsov 1, Sabrina Gordon-Keylock 1, Celine Souilhol 1,4, Frank Buchholz 2, David Hills 1, Suling Zhao 1, Alexander Medvinsky 1,∗
PMCID: PMC6135942  PMID: 30208304

Summary

Hematopoietic stem cells (HSCs) develop in the embryonic aorta-gonad-mesonephros (AGM) region and subsequently relocate to fetal liver. Runx1 transcription factor is essential for HSC development, but is largely dispensable for adult HSCs. Here, we studied tamoxifen-inducible Runx1 inactivation in vivo. Induction at pre-liver stages (up to embryonic day 10.5) reduced erythromyeloid progenitor numbers, but surprisingly did not block the appearance of Runx1-null HSCs in liver. By contrast, ex vivo analysis showed an absolute Runx1 dependency of HSC development in the AGM region. We found that, contrary to current beliefs, significant Cre-inducing tamoxifen activity persists in mouse blood for at least 72 hr after injection. This deferred recombination can hit healthy HSCs, which escaped early Runx1 ablation and result in appearance of Runx1-null HSCs in liver. Such extended recombination activity in vivo is a potential source of misinterpretation, particularly in analysis of dynamic developmental processes during embryogenesis.

Keywords: Runx1, hematopoietic stem cells, AGM, development, hematopoiesis, conditional knockout, tamoxifen

Highlights

  • •

    Runx1 ablation induced in vivo at the AGM stage yields null HSCs in fetal liver

  • •

    Controlled Runx1 ablation in cultured AGM region blocks HSC development

  • •

    Discrepancy is explained by persistence of Cre activity in vivo for at least 3 days

  • •

    Runx1 is essential at pre-liver stage of HSC development

The authors found that Cre-mediated Runx1 ablation induced in vivo at pre-liver stages resulted in appearance of Runx1-null HSCs in the fetal liver. By contrast, deletion of Runx1 in cultured AGM region fully blocked HSC development. Appearance of Runx1-null HSCs in the liver is explained by presence of uncontrolled long-lasting (at least 3 days) Cre-inducing tamoxifen activity in vivo.

Introduction

In the mouse, the first transient hematopoietic waves (primitive erythroid and erythromyeloid progenitors [EMPs]) emerge extra-embryonically in the early yolk (McGrath et al., 2015, Palis et al., 1999). HSCs, which give rise to the adult hematopoietic system, appear only by mid-gestation intra-embryonically within the aorta-gonad-mesonephros (AGM) region (Ciau-Uitz et al., 2016, Medvinsky and Dzierzak, 1996, Medvinsky et al., 1993, Muller et al., 1994). HSCs emerge predominantly from the endothelial floor of the dorsal aorta (de Bruijn et al., 2002, North et al., 2002, Taoudi and Medvinsky, 2007) through a maturation process defined by inductive interactions within the AGM region (Fitch et al., 2012, Souilhol et al., 2016a) and signals from somites (Clements et al., 2011, Lee et al., 2014, Pouget et al., 2014). The origin of HSCs from aortic endothelium is supported by strong evidence in non-mammalian vertebrates and in humans (Bertrand et al., 2010, Ivanovs et al., 2014, Kissa and Herbomel, 2010).

Various signaling pathways are involved in early hematopoietic development (reviewed in [Ciau-Uitz et al., 2016]). HSC maturation in the mouse is a multistep process driven by SCF and IL3 (Robin et al., 2006, Rybtsov et al., 2011, Rybtsov et al., 2014), accompanied by temporal changing signaling requirements for Notch, Shh, and Bmp (Gama-Norton et al., 2015, Lizama et al., 2015, Peeters et al., 2009, Richard et al., 2013, Souilhol et al., 2016a, Souilhol et al., 2016b). It is reasonable to assume that exact transcription factor machinery also attunes. Here we aimed to explore temporal requirements for Runx1, a transcription factor essential for HSC development, but which is far less significant in adult bone marrow HSCs (Cai et al., 2011, North et al., 1999, Okuda et al., 1996, Putz et al., 2006). EMPs and HSCs originate from different endothelial subsets through a Runx1-dependent process (Chen et al., 2011, Eliades et al., 2016, Lie et al., 2018). Conditional Runx1 ablation in vivo blocks the endothelial to EMP/HSC transition prior to liver colonization (Chen et al., 2009). More detailed analysis using a robust AGM explant system recapitulating HSC development, showed that, up to embryonic day 11.5 (E11.5) (immediately prior to fetal liver colonization), HSC development absolutely depends on Runx1, but EMPs become Runx1 independent at earlier stages (Tober et al., 2013).

Here we followed the consequences of induced Runx1 ablation in the AGM region on HSC development in the fetal liver. We used a tamoxifen (4-OHT)-inducible Cre-ERT2 system to delete Runx1 and assessed the presence of EMPs and definitive HSCs in the E13.5 fetal liver. We observed a significant reduction of EMP numbers, but, contrary to our expectations and in apparent discrepancy with a previous report (Tober et al., 2013), we detected significant numbers of Runx1-null HSCs in the liver, implying that HSC development in the AGM region does not require Runx1. We then performed in vivo injections of 4-OHT, but in this case isolated E11.5 AGM regions and after culture transplanted them into irradiated recipients. We found no Runx1-null HSCs in repopulated irradiated recipients, which is in conflict with results of in vivo analysis.

In search of an explanation, we tested how long 4-OHT or/and its active derivatives (cumulatively termed here 4-OHT/Der) remain in the circulation of adult mouse and found that, at least 3 days after injection, concentration of 4-OHT/Der in the peripheral blood of the animal remains sufficient to cause significant (up to 20%) Cre-mediated recombination in cells. Therefore, those HSCs, which have not recombined during the AGM stage could proceed with development, but could be hit by recombination later when Runx1 is no longer essential. This would inevitably lead to emergence of functionally active Runx1-null HSCs in the fetal liver. Such a problem does not exist when the AGM region is removed from the mouse and explanted in fresh culture medium free of 4-OHT/Der.

Thus, despite the advantages of in vivo analysis in principle, poor control over in utero development may bring about a significant uncertainty and cause misinterpretation of results. By contrast, the functionally validated and robustly controlled ex vivo system provides strong evidence for the essential role of Runx1 in HSC development, in keeping with previously published results.

Results

4-OHT-Induced Runx1 Ablation at E8.5–E9.5, but Not Thereafter, Reduces EMP Formation

Here, we studied effects of in-vivo-induced stage-specific Runx1 ablation on the appearance of EMPs in fetal liver. To this end, pregnant [Rosa26CreERT2:sGFP:Runx1fl/fl] females crossed with [Rosa26CreERT2::sGFP::Runx1wt/fl] males were treated with a single dose of 4-OHT between 8.5 and 10.5 days of gestation at a concentration of 2 mg/mouse. E13.5 litters containing both homozygous Runx1fl/fl and heterozygous Runx1wt/fl (control) embryos were harvested, and cell suspensions prepared from individual livers were analyzed in methylcellulose colony-forming unit (CFU) assays (Figure 1A). In homozygous Runx1fl/fl embryos ([Rosa26CreERT2::sGFP::Runx1fl/fl]), 4-OHT injections at E8.5 or E9.5 led to a significant reduction of fetal liver EMPs by ∼50% compared with Runx1wt/fl heterozygous [Rosa26CreERT2::sGFP::Runx1wt/fl] controls (Figure 1B, blue and red bars). However, 4-OHT injection at later stage (E10.5) had no effect on EMPs in fetal livers (Figure 1B, green bars), indicating the essential role of Runx1 in EMP development at early stages but not after E10.5. This result is in agreement with temporal Runx1 dependency in embryonic EMPs reported previously (Tober et al., 2013).

Figure 1.

Figure 1

Open in a new tab

Tamoxifen (4-OHT)-Induced Runx1 Ablation at E8.5–E9.5, but Not Thereafter, Reduces EMP Formation

(A) Experimental design: studying the effect of conditional Runx1 deletion on EMPs. A single dose of 4-OHT was administered intraperitoneally (i.p.) to [Rosa26CreERT2::sGFP::Runx1fl/fl] pregnant females crossed with [Rosa26CreERT2::sGFP::Runx1wt/fl] males at E8.5–E10.5. E13.5 embryos were collected and EMPs from individual FLs quantified by clonogenic methylcellulose assays. A representative 4-OHT-induced E13.5 GFP+ embryo is shown.

(B) Left: numbers of E13.5 FL EMPs per embryo equivalent (e.e.). Time of i.p. injection indicated for, e.g., as ipE8.5. Mean values and standard errors are shown (three independent experiments, at least three embryos in each experiment). Unpaired Student t statistical tests were performed (∗∗p = 0.001; ∗∗∗p = 0.0002). Right: Runx1 recombination in [Rosa26CreERT2::sGFP::Runx1fl/fl] E13.5 FL erythromyeloid progenitors after 4-OHT injection at pre-liver stages, indicating the percentage of each genotype (at least two independent experiments; mean values and standard errors shown).

(C) Representative Runx1 genotyping PCR gels of EMP colonies, showing wt, fl, and Δ bands.

To confirm this conclusion, we tested the status of Runx1 loci in individual fetal liver EMPs using PCR. We picked only liver-derived methylcellulose colonies expressing GFP, indicative of prior activity of Cre-recombinase. Due to their clonal origin, PCR analysis of individual CFU colonies reveals the Runx1 status of individual EMPs. We found that a fraction of tested colonies (20%–36%) escaped any Runx1 recombination despite activation of GFP (Figures 1B right, 1C, and S1A; Table 1). The recombined fraction (64%–80%) contained those colonies that retained one intact Runx1 allele (Runx1fl/Δ) or had both loci recombined (Runx1Δ/Δ). Of note, 4-OHT injection at E8.5, E9.5, and E10.5 resulted in progressive growth of the proportion of Runx1Δ/Δ EMPs in the FL (11%, 25%, and 47%, respectively) (Figures 1B right and 1C; Table 1). A particularly high fraction of Runx1Δ/Δ EMPs was observed after 4-OHT injection at E10.5, when no reduction of EMP numbers was observed. All together this implies progressive acquisition of Runx1 independency by EMPs during development, in line with a previous report (Tober et al., 2013).

Table 1.

Runx1 Status of E13.5 Fetal Liver GFP+ Erythromyeloid Progenitors (CFU-Cs) Derived from Individual Embryos after 4-OHT Injection at Pre-liver Stages

Experiment Embryo Genotype Embryos with Runx1 Recombined Alleles/Total Embryos (Recombination Range)
 Total Colonies Analyzed
CFU-Cs in E13.5 FL
wt/Δ fl/Δ Δ/Δ
iE8.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 9/9 (11%–42%) 111
R26CreERT2::Runx1fl/fl 12/13 (17–86%) 9/13 (6%–33%) 159
iE9.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 7/7 (8%–58%) 85
R26CreERT2::Runx1fl/fl 13/13 (14%–80%) 11/13 (8%–67%) 134
iE10.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 5/6 (33%–100%) 36
R26CreERT2::Runx1fl/fl 10/12 (20%–80%) 11/12 (20%–100%) 60
Open in a new tab

Recombination range refers to proportion of recombined colonies.

Induced Runx1 Ablation from E9.5 onward Does Not Eliminate HSC Development in the Fetal Liver

We tested whether a single 4-OHT injection between E8.5 and E10.5 would block HSC development in vivo. After injection, embryos continued development in utero until E13.5. Cells from individual E13.5 fetal livers were then isolated and transplanted separately into irradiated recipients, as depicted in (Figure 2A). To our surprise, robust multilineage donor-derived long-term engraftment was observed in all 37 recipient animals, regardless of injection time (Figure 2B, left, and data not shown). Cre activity in these experiments was very efficient as seen by the presence of a large GFP+ cell fraction in most recipients (ranging from 85% to 97%, Figure S2).

Figure 2.

Figure 2

Open in a new tab

Induced Runx1 Ablation from E9.5 onward Does not Eliminate HSC Development in the Fetal Liver

(A) Experimental design analogous to that described in Figure 1A, with cell suspensions from individual E13.5 FL embryos transplanted into irradiated adult mice.

(B) Left: donor-derived contribution in recipient mice blood 16 weeks after transplantation is shown. Each point represents an individual recipient mouse, representing data from at least two independent experiments. Numbers of analyzed mice for Runx1wt/fl+ 4-OH-T series (ipE8.5, ipE9.5, and ipE10.5) were 6, 10, and 10, respectively; for Runx1fl/fl+ 4-OH-T series (ipE8.5, ipE9.5, and ipE10.5) were 18, 13, and 6, respectively. Right: Runx1 status of [Rosa26CreERT2::sGFP::Runx1fl/fl] E13.5 FL HSCs, assessed by analysis of donor-derived erythromyeloid progenitors from long-term (16 weeks) engrafted recipients. Percentages of each genotype (at least two independent experiments; mean values and standard errors) are shown.

(C) Representative Runx1 genotyping PCR gels of HSC-derived progenitors, indicating wt, fl, and Δ bands, are shown.

Since Cre-mediated GFP activation does not guarantee Runx1 recombination in the same cells, the recombination status of Runx1 was verified using PCR. To this end, individual donor-derived GFP+ colonies obtained from blood and bone marrow of long-term (16 weeks) reconstituted animals were analyzed by PCR (Figures 2B right, 2C, and S1B; Table 2). When 4-OHT was injected at E8.5, all but one GFP+ donor-derived methylcellulose colonies obtained cumulatively from 18 recipients contained at least one functional Runx1 allele, indicating that Runx1Δ/Δ HSCs did not develop. However, injection at E9.5 gave a significant proportion of donor-derived Runx1Δ/Δ colonies (17%) in 13 repopulated recipients. Injection at E10.5 resulted in further increase (33%) of Runx1Δ/Δ donor-derived colonies obtained from long-term transplanted animals (a total of six recipients were analyzed) (Figures 2B right and 2C; Table 2).

Table 2.

Runx1 Status of E13.5 Fetal Liver HSCs Assessed by PCR Analysis of Donor-Derived GFP+ Erythromyeloid Progenitors from Long-Term Engrafted Recipients

Experiment Embryo Genotype Embryos with Runx1 Recombined Alleles/Total Embryos (Recombination Range)
 Total Colonies Analyzed
HSC-Derived CFU-Cs (BM + Blood)
wt/Δ fl/Δ Δ/Δ
iE8.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 3/7 (17%–50%) 33
R26CreERT2::Runx1fl/fl 8/11 (33%–100%) 1/11 (8%) 89
iE9.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 5/5 (6%–55%) 63
R26CreERT2::Runx1fl/fl 8/8 (18%–80%) 5/8 (10%–40%) 143
iE10.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 2/3 (71%–100%) 9
R26CreERT2::Runx1fl/fl 3/3 (31%–54%) 3/3 (23%–44%) 52
Open in a new tab

Recombination range refers to proportion of recombined colonies.

Taken alone, these results suggest that Runx1 is not required for HSC development from as early as E9.5, which contradicts the previous report showing Runx1 dependency of HSC development up to E11.5 (Tober et al., 2013). We investigated reasons for this discrepancy as described below.

Active Concentrations of 4-OHT or/and Its Derivatives Persist in the Mouse for No Less than 72 hr

We reasoned that if, after 4-OHT injection, recombination does not occur in all cells at once and 4-OHT activity continues further, HSCs (known to become Runx1 independent when fully mature) may undergo late Runx1 ablation without adverse consequences. Upon transplantation, such Runx1-null HSCs could be generating Runx1Δ/Δ progeny detectable in the methylcellulose assay.

To investigate this possibility, we tested how long 4-OHT/Der persists in the mouse and causes recombination. To this end, wild-type C57BL/6 adult females were intraperitoneally injected with 4-OHT, and serum obtained at different time points after injection was added to cultures of primary AGM cells isolated from E10.5 and E11.5 [RosaCreERT2::sGFP] embryos (Gilchrist et al., 2003) (Figure 3A). We observed maximum recombination efficiency (100% GFP activation) using serum obtained 5 hr after 4-OHT administration. Nevertheless, serum collected after 48 and 72 hr could also cause 20%–40% recombination (Figure 3B). Specifically, 72 hr serum caused 52% ± 8.7% recombination in CD45+ cells, 16% ± 7.6% in VE-Cad+ cells, and 26% ± 13.2% in stromal cells (CD45−VE-Cad–) (data not shown). (Note that recombination in non-hematopoietic cells is underestimated due to known mosaicism of GFP expression in non-hematopoietic cells in these mice [Gilchrist et al., 2003]).

Figure 3.

Figure 3

Open in a new tab

Active Concentrations of 4-OHT or/and Its Derivatives Remain in the Mouse for No Less than 72 hr

(A) Experimental schematics: wild-type C57BL/6 adult females were intraperitoneally injected with 4-OHT and serum obtained at different time points after injection. Then, it was added to cultures of primary E10.5 and E11.5 [RosaCreERT2::sGFP] AGM cells, and recombination was assessed by monitoring GFP expression by fluorescence-activated cell sorting after 24 hr.

(B) GFP recombination in AGM cells upon incubation with serum isolated at different time points (0, 5, 20, 48, and 72 hr) after 4-OH-tamoxifen intraperitoneal injection (at least two independent experiments; three mice/experiment; mean values and standard errors are shown).

These experiments show that, after injection of 4-OHT, cells in the mouse containing loxP sites can recombine over a period of at least 3 days. Thus, those cells in the embryo which have not undergone rapid Cre-mediated recombination of the Runx1flox allele continue to be exposed to 4-OHT/Der and can recombine, even 3 days after 4-OHT injection.

The extended circulation of 4-OHT/Der in vivo creates uncertainty as to when exactly during development Runx1 becomes dispensable for HSC development. The next set of experiments was designed to control and reduce the duration of HSC exposure to 4-OHT/Der to obtain more interpretable results.

Runx1 Deletion at E9.5 Abrogates HSC Activity in the E11.5 AGM Region

To avoid prolonged exposure to 4-OHT/Der, we switched from the entirely in vivo design to experiments incorporating a highly controllable in vitro HSC maturation system (Rybtsov et al., 2011, Rybtsov et al., 2014, Taoudi et al., 2008). As in previous setups, embryos for these experiments contained either both or one conditional Runx1 allele (Runx1fl/fl and Runx1wt/fl, respectively). 4-OHT was administered in vivo at E9.5 as described previously; but, instead of allowing the embryos develop in utero until E13.5, E11.5 AGM regions were dissected and individually explanted (Medvinsky and Dzierzak, 1996). Placement of the AGM region in fresh culture medium ends the exposure of developing HSCs to 4-OHT/Der. After 4 days in culture, AGM cells were transplanted into irradiated recipients (Figure 4A).

Figure 4.

Figure 4

Open in a new tab

Runx1 Deletion at E9.5 Abrogates HSC Activity in the E11.5 AGM Region

(A) 4-OHT-induced Runx1 deletion was performed at E9.5 as described in Figure 1A, but embryos were harvested at E11.5. AGM regions were dissected and cultured as explants for 4 days, and cell suspensions from individual explants were transplanted into irradiated adult mice.

(B) Donor-derived engraftment in the blood from recipient mice 16 weeks after transplantation is shown. Each point represents an individual recipient mouse (data from three independent experiments). Numbers of analyzed mice for Runx1wt/fl and Runx1fl/fl+ 4-OH-T conditions were 12 and 16, respectively.

(C) Top: recipient mice were sacrificed 16 weeks after transplantation and the Runx1 genotype was examined by PCR in individual donor-derived GFP+ CFU-Cs from bone marrow and blood. Runx1 genotype frequencies within donor-engrafted mice are plotted. Average and SE from embryos from three independent experiments are given (no production of Runx1Δ/Δ colonies observed). Bottom: representative Runx1 genotyping PCR gels of colonies, with Runx1 wt, fl, and Δ bands shown.

Twelve recipients transplanted with control [Rosa26CreERT2::sGFP::Runx1wt/fl] AGM cells showed high-level, long-term multilineage donor-derived engraftment. Fourteen of 16 recipients transplanted with [Rosa26CreERT2:: sGFP::Runx1fl/fl] AGM cells were also repopulated at high level (only one showed low-level donor-derived chimerism and one lacked engraftment) (Figure 4B). We observed highly efficient GFP recombination in [Rosa26CreERT2::sGFP::Runx1wt/fl] and [Rosa26CreERT2::sGFP::Runx1fl/fl] repopulated cells (98% and 89%, respectively) (Figure S3). However, analysis of methylcellulose colonies obtained from these 14 long-term repopulated mice showed that, by contrast to liver, [Rosa26CreERT2:: sGFP::Runx1fl/fl] AGM region transplants contained no Runx1Δ/Δ HSCs (repopulated mice gave no Runx1Δ/Δ myeloid colonies) (Figure 4C; Table 3), indicating an absolute requirement of Runx1 for HSC development prior to liver colonization. Of note, variance in efficiency of recombination suggests a considerable difference in accessibility of loxP sites for Cre-mediated recombination between GFP and Runx1 loci.

Table 3.

Runx1 Genotype of E11.5 AGM Explant HSCs Assessed by PCR Analysis of Donor-Derived GFP+ Erythromyeloid Progenitors from Long-Term Engrafted Recipients

Experiment Embryo Genotype Embryos with Runx1 Recombined Alleles/Total Embryos (Recombination Range)
 Total Colonies Analyzed
HSC-Derived CFU-Cs (BM + Blood)
wt/Δ fl/Δ Δ/Δ
iE9.5 -> E13.5 FL R26CreERT2::Runx1wt/fl 4/6 (11%–43%) 71
R26CreERT2::Runx1fl/fl 10/12 (14%–64%) 0/12 (NA) 127
Open in a new tab

Recombination range refers to proportion of recombined colonies.

NA, not applicable.

Discussion

During maturation, pre-HSCs undergo massive expansion in the AGM region, sequentially upregulating hematopoietic CD41, CD43, and CD45, and some other markers (Boisset et al., 2015, Ferkowicz et al., 2003, Rybtsov et al., 2011, Rybtsov et al., 2014, Rybtsov et al., 2016, Taoudi et al., 2008, Tober et al., 2018, Zhou et al., 2016). This stepwise developmental process is driven by dynamic gene actions. Temporal requirements for genes driving HSC development can vary, and accurate staging of gene involvement is essential for understanding this process. For example, acquisition of adult phenotype is accompanied by gradual loss of Notch, Shh, and Bmp dependency by HSCs in the AGM region (Gama-Norton et al., 2015, Lizama et al., 2015, Peeters et al., 2009, Richard et al., 2013, Souilhol et al., 2016a, Souilhol et al., 2016b).

Runx1 is a key transcription factor involved in EMP and HSC development (Cai et al., 2000, Chen et al., 2009, North et al., 2002, Okuda et al., 1996). However, exact action points of Runx1 during this multistep process require clarification. Initial hematopoietic specification associated with upregulation of CD41 does not require Runx1 (Liakhovitskaia et al., 2014). Runx1 sensitivity of the embryonic endothelium is limited to the period around E7.5–E8.5 (Tanaka et al., 2012, Yzaguirre et al., 2018). Induced stage-specific inactivation using explant cultures followed by long-term transplantations showed that, up to E11.5, Runx1 is absolutely required for HSC development but becomes dispensable after that (Tober et al., 2013). Here we explored whether the conclusions of this study can be corroborated by in vivo tracking.

To this end, we injected pregnant females with 4-OHT during active EMP/HSC formation. Instead of analysis of E11.5 AGM explants (Tober et al., 2013), we followed HSC development in utero up to E13.5 in the fetal liver. Injections at E8.5/E9.5, but not at E10.5, reduced EMP numbers in the E13.5 liver, indicating that by this stage EMPs become Runx1 independent, which was corroborated by increased emergence of Runx1Δ/Δ, in agreement with previously reported results (Tober et al., 2013).

We then focused on the HSC development. After a single 4-OHT injection at E8.5, E9.5, or E10.5, the embryos were allowed to develop in utero up to E13.5, after which fetal liver cells were transplanted into irradiated recipients. Unexpectedly, all long-term repopulated animals (transplanted with E9.5 cells onward) contained Runx1-null HSCs, in apparent contradiction with previous data obtained in AGM cultures (Tober et al., 2013).

We then reasoned that in vivo results might depend on the persistence of 4-OHT/Der activity in the mouse. Indeed, 4-OHT injections at E8.5/E9.5 yielded non-recombined (Runx1fl/fl) and partly recombined (Runx1fl/Δ) HSCs. If 4-OHT/Der persisted in the mouse beyond E11.5 stage (when HSCs become Runx1 independent), delayed recombination could transform non-recombined (Runx1fl/fl) and partly recombined (Runx1fl/Δ) HSCs into Runx1Δ/Δ HSCs without adverse effect. Indeed, we found that substantial 4-OHT/Der activity, capable of inducing CRE-ERT2, persists in the mouse for at least 3 days, introducing major uncertainty as to when exactly recombination took place and making in vivo experiments poorly interpretable. Crucially, under highly controlled experimental conditions, where after a single 4-OHT injection at E9.5, E11.5 AGM regions were isolated and cultured in fresh medium free of 4-OHT/Der, subsequent transplantations showed complete ablation of Runx1-null HSCs. This result points to a critical role for Runx1 in HSC development in the AGM region before colonization of the fetal liver, in agreement with a previous report (Tober et al., 2013).

Interestingly, although we observed near 100% recombination of the GFP transgene, the efficiency of Runx1 recombination was significantly lower, emphasizing that accessibility of loxP sites in various genetic loci further confounds interpretation of in vivo results. Our recent study exploring the role of Notch in HSC development (Souilhol et al., 2016b) failed to obtain any useful data using in-utero-induced gene ablation due to the highly resistant nature of the floxed RbpJk allele (our unpublished data).

In summary, our current study shows a crucial role for Runx1 in HSC development in the AGM region prior to liver colonization, and emphasizes the value of in vitro models recapitulating developmental processes. Although studying biological processes in vivo is undeniably important, experiments involving induced gene modifications to analyze dynamic developmental processes in utero can incur major uncertainties. We believe that recapitulation of developmental processes in vitro using highly controlled conditions, supported by functional validation, will continue to play an important role in various areas of developmental biology.

Experimental Procedures

Mice

All transgenic mice used to generate embryos and adult mice were on the C57BL/6 (CD45.2/2) background. Rosa26CreERT2::Runx1fl/fl and Rosa26CreERT2::sGFP::Runx1fl/fl mice were used to obtain Rosa26CreERT2::sGFP::Runx1fl/fl conditional knockout and Rosa26CreERT2::sGFP::Runx1wt/fl control embryos. The morning of discovery of the vaginal plug was designated as day 0.5. E11.5 embryos used had 41–45 somite pairs. C57BL/6 (CD45.1/CD45.2) animals were used as recipients for transplantations. Rosa26CreERT2 were a gift from Lars Grotewold and Austin Smith, previously used in (Souilhol et al., 2016b); Runx1fl/fl mice and silent GFP (sGFP) mice were generated previously as described in (Putz et al., 2006) and (Gilchrist et al., 2003), respectively. All experiments with animals were performed under a Project License granted by the Home Office (UK), University of Edinburgh Ethical Review Committee, and conducted in accordance with local guidelines.

E11.5 AGM Explant Culture

AGM regions were dissected from E11.5 embryos and incubated for 2 hr at 37°C and 5% CO2 in IMDM medium (Invitrogen) supplemented with 20% fetal calf serum (FCS) (HyClone, Fisher Scientific), L-glutamine (L-Gln), and penicillin/streptomycin (P/S). AGM regions were transferred individually onto 0.8 μm mixed cellulose MF membranes (AAWP02500, Millipore) for 5–7 days (37°C, 5% CO2) and cultured at the liquid-gas interface with IMDM+ medium consisting of 20% FCS, L-Gln, P/S IMDM, and growth factors (interleukin-3, SCF, and FLT3L, 100 ng/mL each; all from PeproTech). After 4 days of culture, explants were dissociated by enzymatic digestion (collagenase/dispase in PBS containing CaCl2 and MgCl2). Methylcellulose CFU-C and long-term repopulation assays were subsequently performed.

Hematopoietic Clonogenic Progenitor (CFU-C) and Colony Genotyping

Cell suspensions from either fresh or cultured embryonic tissues were plated into semi-solid Methocult M3434 medium (StemCell Technologies). After 7–12 days of culture, CFU-C colonies were scored and recombination in the Runx1 locus analyzed by PCR. In brief, colonies were individually transferred into 50 μL of lysis buffer, incubated for 1 hr at 56°C followed by 5 min at 95°C. Runx1wt (251 bp), Runx1fl (325 bp), and Runx1Δ (382 bp) products were amplified by simultaneously using the primers P5 (5′-TAGGGAGTGCTGCTTGCTCT-3′), P6 (5′- GCCGGGTGCAA.ATTAAGTC-3′), and P7 (5′-CTCTGGGAAACCAGGGAGTG-3′), under the following PCR conditions: 94°C 5 min; [94°C 30 s; 60°C 30 s; 72°C 1 min] × 35 cycles; 72°C 10 min.

Long-Term Hematopoietic Repopulation Assays

Cell suspensions from embryos at different stages were injected into irradiated adult recipients (CD45.1/2) either directly (E13.5 FL, 0.2 embryo equivalents per recipient mouse) or after explant culture (E11.5 AGM, 0.1 embryo equivalents per recipient mouse), along with 80,000 CD45.1/1 bone marrow carrier cells. Recipients were irradiated by split dose (472 + 472 rad with 3-hr interval) of γ irradiation. Donor-derived chimerism was monitored in blood at 6 and 16 weeks after transplantation using FACScalibur (BD). Peripheral blood was collected by bleeding the lateral tail vein into 500 μL 5 mM EDTA/PBS, and erythrocytes were depleted using PharM Lyse (BD). Cells were stained with anti-CD16/32 (Fc-block, cat. no. 16-0161-86), CD45.1-APC (clone A20, cat. no. 17-0453-82), and anti-CD45.2-PE (clone 104, cat. no. 12-0454-83) monoclonal antibodies (eBioscience). Appropriate isotype controls were used. Dead cells were excluded using 7AAD (eBioscience). Mice demonstrating ≥5% donor-derived multilineage chimerism after 8 weeks were considered to be reconstituted.

Intraperitoneal Injections

Rosa26CreERT2::sGFP::Runx1fl/fl and Rosa26CreERT2::sGFP::Runx1wt/fl embryos were induced in vivo by intraperitoneal injection of 2 mg 4-OHT (Sigma) into Rosa26CreERT2::sGFP::Runx1wt/fl x Rosa26CreERT2::sGFP::Runx1fl/fl pregnant females. 4-OHT was prepared in 100% ethanol, diluted 1:10 in corn oil. Embryos were allowed to develop in utero up to E11.5 or E13.5 before harvesting tissues for analysis by CFU-C, flow cytometry, and repopulation assays.

Testing Persistence of 4-OHT in Peripheral Blood of Adult Mice

C57BL/6 females were injected with 4-OHT as described above. Peripheral blood was collected by submandibular bleeding before 4-OHT injection (0 hr), 5, 20, 48, and 72 hr following injection. After centrifugation (400 × g/5 min), serum was collected and additionally centrifuged (400 × g/5 min) to remove remaining blood cells. Serum was frozen at −80°C and used at the same time in two independent sets of experiments. Total E10-11 AGM cells cultured for 24 hr in IMDM + 50% mouse serum were collected and analyzed by flow cytometry for GFP expression (dead cells were excluded by 7AAD). In some cases the cells were stained for CD45-PE (Pharmingen, cat. no. 553081) and VE-Cadherin-AlexaFluor-647 (eBioscience, cat. no. 51-1441-82; 138006).

Author Contributions

J.S., A.B., S.R., S.G.-K., C.S., D.H., and S.Z. performed experiments. J.S. and A.M. designed the research. J.S. made figures and analyzed data. J.S. and A.M. wrote the paper. F.B. provided essential materials. A.M. provided overall supervision.

Acknowledgments

The authors thank J. Verth and C. Manson for assistance with mouse maintenance and breeding; A. Dyer for irradiations. We thank Drs. Lars Grotewold and Austin Smith for the RosaCreERT2 mice. The authors thank Dr. A. Binagui-Casas for helpful comments. This work was supported by the Medical Research Council.

Published: August 30, 2018

Footnotes

Supplemental Information includes three figures and can be found with this article online at https://doi.org/10.1016/j.stemcr.2018.08.004.

Supplemental Information

Document S1. Figures S1–S3
mmc1.pdf (413.5KB, pdf)
Document S2. Article plus Supplemental Information
mmc2.pdf (2.1MB, pdf)

References

  1. Bertrand J.Y., Chi N.C., Santoso B., Teng S., Stainier D.Y., Traver D. Haematopoietic stem cells derive directly from aortic endothelium during development. Nature. 2010;464:108–111. doi: 10.1038/nature08738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Boisset J.C., Clapes T., Klaus A., Papazian N., Onderwater J., Mommaas-Kienhuis M., Cupedo T., Robin C. Progressive maturation toward hematopoietic stem cells in the mouse embryo aorta. Blood. 2015;125:465–469. doi: 10.1182/blood-2014-07-588954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Cai Z., de Bruijn M., Ma X., Dortland B., Luteijn T., Downing R.J., Dzierzak E. Haploinsufficiency of AML1 affects the temporal and spatial generation of hematopoietic stem cells in the mouse embryo. Immunity. 2000;13:423–431. doi: 10.1016/s1074-7613(00)00042-x. [DOI] [PubMed] [Google Scholar]
  4. Cai X., Gaudet J.J., Mangan J.K., Chen M.J., De Obaldia M.E., Oo Z., Ernst P., Speck N.A. Runx1 loss minimally impacts long-term hematopoietic stem cells. PLoS One. 2011;6:e28430. doi: 10.1371/journal.pone.0028430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chen M.J., Yokomizo T., Zeigler B.M., Dzierzak E., Speck N.A. Runx1 is required for the endothelial to haematopoietic cell transition but not thereafter. Nature. 2009;457:887–891. doi: 10.1038/nature07619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen M.J., Li Y., De Obaldia M.E., Yang Q., Yzaguirre A.D., Yamada-Inagawa T., Vink C.S., Bhandoola A., Dzierzak E., Speck N.A. Erythroid/myeloid progenitors and hematopoietic stem cells originate from distinct populations of endothelial cells. Cell Stem Cell. 2011;9:541–552. doi: 10.1016/j.stem.2011.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Ciau-Uitz A., Patient R., Medvinsky A. Ontogeny of the haematopoietic system. In: Ratcliffe M.J.H., editor. vol. 1. Academic Press; 2016. pp. 1–14. (Encyclopedia of Immunobiology). [Google Scholar]
  8. Clements W.K., Kim A.D., Ong K.G., Moore J.C., Lawson N.D., Traver D. A somitic Wnt16/Notch pathway specifies haematopoietic stem cells. Nature. 2011;474:220–224. doi: 10.1038/nature10107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. de Bruijn M.F., Ma X., Robin C., Ottersbach K., Sanchez M.J., Dzierzak E. Hematopoietic stem cells localize to the endothelial cell layer in the midgestation mouse aorta. Immunity. 2002;16:673–683. doi: 10.1016/s1074-7613(02)00313-8. [DOI] [PubMed] [Google Scholar]
  10. Eliades A., Wareing S., Marinopoulou E., Fadlullah M.Z.H., Patel R., Grabarek J.B., Plusa B., Lacaud G., Kouskoff V. The hemogenic competence of endothelial progenitors is restricted by Runx1 silencing during embryonic development. Cell Rep. 2016;15:2185–2199. doi: 10.1016/j.celrep.2016.05.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ferkowicz M.J., Starr M., Xie X., Li W., Johnson S.A., Shelley W.C., Morrison P.R., Yoder M.C. CD41 expression defines the onset of primitive and definitive hematopoiesis in the murine embryo. Development. 2003;130:4393–4403. doi: 10.1242/dev.00632. [DOI] [PubMed] [Google Scholar]
  12. Fitch S.R., Kimber G.M., Wilson N.K., Parker A., Mirshekar-Syahkal B., Gottgens B., Medvinsky A., Dzierzak E., Ottersbach K. Signaling from the sympathetic nervous system regulates hematopoietic stem cell emergence during embryogenesis. Cell Stem Cell. 2012;11:554–566. doi: 10.1016/j.stem.2012.07.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gama-Norton L., Ferrando E., Ruiz-Herguido C., Liu Z., Guiu J., Islam A.B., Lee S.U., Yan M., Guidos C.J., López-Bigas N. Notch signal strength controls cell fate in the haemogenic endothelium. Nat. Commun. 2015;6:8510. doi: 10.1038/ncomms9510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gilchrist D.S., Ure J., Hook L., Medvinsky A. Labeling of hematopoietic stem and progenitor cells in novel activatable EGFP reporter mice. Genesis. 2003;36:168–176. doi: 10.1002/gene.10209. [DOI] [PubMed] [Google Scholar]
  15. Ivanovs A., Rybtsov S., Anderson R.A., Turner M.L., Medvinsky A. Identification of the niche and phenotype of the first human hematopoietic stem cells. Stem Cell Rep. 2014;2:449–456. doi: 10.1016/j.stemcr.2014.02.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kissa K., Herbomel P. Blood stem cells emerge from aortic endothelium by a novel type of cell transition. Nature. 2010;464:112–115. doi: 10.1038/nature08761. [DOI] [PubMed] [Google Scholar]
  17. Lee Y., Manegold J.E., Kim A.D., Pouget C., Stachura D.L., Clements W.K., Traver D. FGF signalling specifies haematopoietic stem cells through its regulation of somitic Notch signalling. Nat. Commun. 2014;5:5583. doi: 10.1038/ncomms6583. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Liakhovitskaia A., Rybtsov S., Smith T., Batsivari A., Rybtsova N., Rode C., de Bruijn M., Buchholz F., Gordon-Keylock S., Zhao S. Runx1 is required for progression of CD41+ embryonic precursors into HSCs but not prior to this. Development. 2014;141:3319–3323. doi: 10.1242/dev.110841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lie-A-Ling M., Marinopoulou E., Lilly A.J., Challinor M., Patel R., Lancrin C., Kouskoff V., Lacaud G. Regulation of RUNX1 dosage is crucial for efficient blood formation from hemogenic endothelium. Development. 2018;145 doi: 10.1242/dev.149419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lizama C.O., Hawkins J.S., Schmitt C.E., Bos F.L., Zape J.P., Cautivo K.M., Borges Pinto H., Rhyner A.M., Yu H., Donohoe M.E. Repression of arterial genes in hemogenic endothelium is sufficient for haematopoietic fate acquisition. Nat. Commun. 2015;6:7739. doi: 10.1038/ncomms8739. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. McGrath K.E., Frame J.M., Fegan K.H., Bowen J.R., Conway S.J., Catherman S.C., Kingsley P.D., Koniski A.D., Palis J. Distinct sources of hematopoietic progenitors emerge before HSCs and provide functional blood cells in the mammalian embryo. Cell Rep. 2015;11:1892–1904. doi: 10.1016/j.celrep.2015.05.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Medvinsky A., Dzierzak E. Definitive hematopoiesis is autonomously initiated by the AGM region. Cell. 1996;86:897–906. doi: 10.1016/s0092-8674(00)80165-8. [DOI] [PubMed] [Google Scholar]
  23. Medvinsky A.L., Samoylina N.L., Muller A.M., Dzierzak E.A. An early pre-liver intraembryonic source of CFU-S in the developing mouse. Nature. 1993;364:64–67. doi: 10.1038/364064a0. [DOI] [PubMed] [Google Scholar]
  24. Muller A.M., Medvinsky A., Strouboulis J., Grosveld F., Dzierzak E. Development of hematopoietic stem cell activity in the mouse embryo. Immunity. 1994;1:291–301. doi: 10.1016/1074-7613(94)90081-7. [DOI] [PubMed] [Google Scholar]
  25. North T., Gu T.L., Stacy T., Wang Q., Howard L., Binder M., Marin-Padilla M., Speck N.A. Cbfa2 is required for the formation of intra-aortic hematopoietic clusters. Development. 1999;126:2563–2575. doi: 10.1242/dev.126.11.2563. [DOI] [PubMed] [Google Scholar]
  26. North T.E., de Bruijn M.F., Stacy T., Talebian L., Lind E., Robin C., Binder M., Dzierzak E., Speck N.A. Runx1 expression marks long-term repopulating hematopoietic stem cells in the midgestation mouse embryo. Immunity. 2002;16:661–672. doi: 10.1016/s1074-7613(02)00296-0. [DOI] [PubMed] [Google Scholar]
  27. Okuda T., van Deursen J., Hiebert S.W., Grosveld G., Downing J.R. AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell. 1996;84:321–330. doi: 10.1016/s0092-8674(00)80986-1. [DOI] [PubMed] [Google Scholar]
  28. Palis J., Robertson S., Kennedy M., Wall C., Keller G. Development of erythroid and myeloid progenitors in the yolk sac and embryo proper of the mouse. Development. 1999;126:5073–5084. doi: 10.1242/dev.126.22.5073. [DOI] [PubMed] [Google Scholar]
  29. Peeters M., Ottersbach K., Bollerot K., Orelio C., de Bruijn M., Wijgerde M., Dzierzak E. Ventral embryonic tissues and Hedgehog proteins induce early AGM hematopoietic stem cell development. Development. 2009;136:2613–2621. doi: 10.1242/dev.034728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Pouget C., Peterkin T., Simoes F.C., Lee Y., Traver D., Patient R. FGF signalling restricts haematopoietic stem cell specification via modulation of the BMP pathway. Nat. Commun. 2014;5:5588. doi: 10.1038/ncomms6588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Putz G., Rosner A., Nuesslein I., Schmitz N., Buchholz F. AML1 deletion in adult mice causes splenomegaly and lymphomas. Oncogene. 2006;25:929–939. doi: 10.1038/sj.onc.1209136. [DOI] [PubMed] [Google Scholar]
  32. Richard C., Drevon C., Canto P.Y., Villain G., Bollérot K., Lempereur A., Teillet M.A., Vincent C., Rosselló Castillo C., Torres M. Endothelio-mesenchymal interaction controls runx1 expression and modulates the notch pathway to initiate aortic hematopoiesis. Dev. Cell. 2013;24:600–611. doi: 10.1016/j.devcel.2013.02.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Robin C., Ottersbach K., Durand C., Peeters M., Vanes L., Tybulewicz V., Dzierzak E. An unexpected role for IL-3 in the embryonic development of hematopoietic stem cells. Dev. Cell. 2006;11:171–180. doi: 10.1016/j.devcel.2006.07.002. [DOI] [PubMed] [Google Scholar]
  34. Rybtsov S., Sobiesiak M., Taoudi S., Souilhol C., Senserrich J., Liakhovitskaia A., Ivanovs A., Frampton J., Zhao S., Medvinsky A. Hierarchical organization and early hematopoietic specification of the developing HSC lineage in the AGM region. J. Exp. Med. 2011;208:1305–1315. doi: 10.1084/jem.20102419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rybtsov S., Batsivari A., Bilotkach K., Paruzina D., Senserrich J., Nerushev O., Medvinsky A. Tracing the origin of the HSC hierarchy reveals an SCF-dependent, IL-3-independent CD43(–) embryonic precursor. Stem Cell Rep. 2014;3:489–501. doi: 10.1016/j.stemcr.2014.07.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rybtsov S., Ivanovs A., Zhao S., Medvinsky A. Concealed expansion of immature precursors underpins acute burst of adult HSC activity in foetal liver. Development. 2016;143:1284–1289. doi: 10.1242/dev.131193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Souilhol C., Gonneau C., Lendinez J.G., Batsivari A., Rybtsov S., Wilson H., Morgado-Palacin L., Hills D., Taoudi S., Antonchuk J. Inductive interactions mediated by interplay of asymmetric signalling underlie development of adult haematopoietic stem cells. Nat. Commun. 2016;7:10784. doi: 10.1038/ncomms10784. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Souilhol C., Lendinez J.G., Rybtsov S., Murphy F., Wilson H., Hills D., Batsivari A., Binagui-Casas A., McGarvey A.C., MacDonald H.R. Developing HSCs become notch independent by the end of maturation in the AGM region. Blood. 2016;128:1567–1577. doi: 10.1182/blood-2016-03-708164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Tanaka Y., Hayashi M., Kubota Y., Nagai H., Sheng G., Nishikawa S., Samokhvalov I.M. Early ontogenic origin of the hematopoietic stem cell lineage. Proc. Natl. Acad. Sci. USA. 2012;109:4515–4520. doi: 10.1073/pnas.1115828109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Taoudi S., Medvinsky A. Functional identification of the hematopoietic stem cell niche in the ventral domain of the embryonic dorsal aorta. Proc. Natl. Acad. Sci. USA. 2007;104:9399–9403. doi: 10.1073/pnas.0700984104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Taoudi S., Gonneau C., Moore K., Sheridan J.M., Blackburn C.C., Taylor E., Medvinsky A. Extensive hematopoietic stem cell generation in the AGM region via maturation of VE-cadherin+CD45+ pre-definitive HSCs. Cell Stem Cell. 2008;3:99–108. doi: 10.1016/j.stem.2008.06.004. [DOI] [PubMed] [Google Scholar]
  42. Tober J., Yzaguirre A.D., Piwarzyk E., Speck N.A. Distinct temporal requirements for Runx1 in hematopoietic progenitors and stem cells. Development. 2013;140:3765–3776. doi: 10.1242/dev.094961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Tober J., Maijenburg M.M.W., Li Y., Gao L., Hadland B.K., Gao P., Minoura K., Bernstein I.D., Tan K., Speck N.A. Maturation of hematopoietic stem cells from prehematopoietic stem cells is accompanied by up-regulation of PD-L1. J. Exp. Med. 2018;215:645–659. doi: 10.1084/jem.20161594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Yzaguirre A.D., Howell E.D., Li Y., Liu Z., Speck N.A. Runx1 is sufficient for blood cell formation from non-hemogenic endothelial cells in vivo only during early embryogenesis. Development. 2018;145 doi: 10.1242/dev.158162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Zhou F., Li X., Wang W., Zhu P., Zhou J., He W., Ding M., Xiong F., Zheng X., Li Z. Tracing haematopoietic stem cell formation at single-cell resolution. Nature. 2016;533:487–492. doi: 10.1038/nature17997. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S3
mmc1.pdf (413.5KB, pdf)
Document S2. Article plus Supplemental Information
mmc2.pdf (2.1MB, pdf)

Articles from Stem Cell Reports are provided here courtesy of Elsevier

RESOURCES