CFU-S₁₁ activity does not localize solely with the aorta in the aorta-gonad-mesonephros region

During mouse embryonic development, different hematopoietic cell types are successively generated at several locations within the conceptus. At embryonic day 7.5 (E7.5), primitive erythrocytes and their direct progenitors are found in the yolk sac (YS).1 Shortly thereafter, the first definitive, in vitro clonable hematopoietic progenitors are formed in the YS and embryo body.2-4 Cells capable of in vivo hematopoietic activity are found slightly later in both the YS and the embryo: at E9 colony-forming units–spleen day 8 (CFU-S₈)5and neonatal repopulating cells6 and at E10 CFU-S₁₁5 and adult repopulating cells.7 We previously showed that long-term adult repopulating hematopoietic stem cells (HSCs) autonomously emerge first in the intrabody aorta-gonad-mesonephros (AGM) region (more than 34 somite pair [sp]).8 More recently we subdissected the AGM region into separate pieces containing either the aorta and surrounding mesenchyme or urogenital ridges (UGRs) and demonstrated that adult repopulating HSCs are first found exclusively in the area of the dorsal aorta and not in the flanking UGRs.9 B-cell progenitors in E10 AGM were also found predominantly in the aorta/mesentery subregion.3 However, sublocalization has not been performed on CFU-S₁₁, a myeloid-restricted stem cell subset that is slightly downstream from HSCs in the adult hematopoietic differentiation hierarchy.

We previously showed that CFU-S₁₁ can be detected in the AGM region slightly earlier (32-33 sp) than adult repopulating HSCs (at least 34 sp).7,8 Spatial and temporal mapping of CFU-S₁₁ activity in the AGM should help delineate the relationship between the functionally distinct CFU-S₁₁ and HSCs and contribute to our insight into the complex process of the onset of definitive hematopoiesis. Here we show that unlike HSCs, which are initially localized to the area of the dorsal aorta, CFU-S₁₁ are found in both the aorta and UGRs of E10 to E12 embryos. The number of CFU-S₁₁ in the aorta and in UGRs increases over developmental time. Furthermore, our data demonstrate that the aorta and UGRs are capable of autonomously increasing CFU-S₁₁ activity in explant cultures, suggesting that CFU-S₁₁ are generated in situ either from direct precursors or from pre-existing HSCs and expand during the in vitro culture period.

C57BL/10 × CBA F1 female mice were mated with C57BL/10 × CBA F1 or ln7210 male mice to generate embryos of appropriate ages. The day of vaginal plug discovery was considered E0. At day 10, 11, or 12 of pregnancy, the females were killed by cervical dislocation. The mice were housed according to institutional guidelines, and animal procedures were carried out in compliance with the standards for Humane Care and Use of Laboratory Animals.

The embryos were taken out of the uterus, and AGMs were dissected.11 Subsequently, AGMs were subdissected into the aorta with its surrounding mesenchyme and the UGRs using 27-gauge insulin syringes (Becton Dickinson, San Jose, CA). The tissues were either cultured in 3-day explant cultures8 or directly processed for transplantation into irradiated adult mice. Cell suspensions of freshly isolated or cultured tissues were obtained by collagenase digestion as previously described.11 Viable cells were counted using trypan blue (Sigma Chemical, St Louis, MO).

Cell suspensions of AGM, aorta, or UGRs were injected intravenously into the tail vein of adult (8-14 weeks old) C57BL/10 × CBA F1 mice (male or female). The recipients were conditioned by 1000 rads of γ-irradiation, which was administered in a split dose, with 3 hours between doses. The mice were killed 11 days after transplantation by cervical dislocation. The spleens were removed and fixed in Tellesniczky's fixative, and the colonies were counted by macroscopic observation.

Isolated AGMs were fixed in 2.5% glutaraldehyde (Sigma), post-fixed in 1% osmium tetroxide (OsO₄), dehydrated in ethanol, and critically point-dried. The tissues were mounted on stubs, coated with gold/paladium, and examined with a JEOL (Tokyo, Japan) JSM-25 electron microscope.

The spatial and temporal distribution of CFU-S₁₁ in the E10 to E12 AGM region was examined after subdissecting this tissue into parts containing the aorta with its surrounding mesenchyme and the UGRs. As shown in scanning electron micrographs, the AGM doubles in size from E10 to E12 (Figure 1A). During this time the distinctive features of the maturing excretory and reproductive systems allow for the precise separation of the UGRs from the dorsal aorta and mesenchyme. Cell counts were performed on the whole AGM, subdissected aorta, and UGRs. All tissues showed a 2-fold increase in cell number from E10 to E12 (Figure 1B). After E12, following further organogenesis, the AGM region as such recedes, and hematopoiesis takes place in the fetal liver.12 

CFU-S₁₁ were present in both the area of the aorta and UGRs at E10, E11, and E12 (Figure 2A). This was in agreement with the sublocalization of CFU-S₈ that was found at E10.13 A significant increase in CFU-S₁₁ was observed between E10 and E12, ie, 4.8-fold for the aorta and 11.2-fold for the UGRs (P ≤ .003, as determined by linear regression analysis). At all time points there were more CFU-S₁₁ in the aorta than in the UGRs (75%, 70%, and 56% of the recovered CFU-S₁₁ for E10, E11, and E12, respectively). Also, the frequency of CFU-S₁₁ was highest in the aorta at E10 and E11 (Figure 2B). Although the aorta appears to harbor more CFU-S₁₁ than the UGRs, the precise site of CFU-S₁₁ origin remains uncertain. Similar to adult repopulating HSCs,9 CFU-S₁₁ may find their origin in association with the intra-embryonic vasculature. To examine this possibility, we also transplanted cells of the vitelline and umbilical vessels. At E10.5 and E11, we found 0.3 and 1.6 CFU-S₁₁, respectively, per embryo equivalent of transplanted vitelline and umbilical cells, suggesting that CFU-S₁₁ may be generated and/or expanded in association with vascular tissue.

The temporal increase in CFU-S₁₁ activity in the aorta and also in the UGRs may reflect either the autonomous generation and/or expansion of these cells in the aorta and UGRs or the influx of CFU-S₁₁ into these tissues. To examine these possibilities, organ explants of the E10 to E12 aorta and UGRs were cultured prior to assaying CFU-S₁₁ activity. At E10, CFU-S₁₁activity was maintained during culture (Figure 2C) and at best showed an average 2.5-fold increase compared to directly transplanted cells. In contrast, CFU-S₁₁ activity significantly increased in cultures of the E11 aorta (5-fold) and UGRs (16.6-fold) (P < .05 and P < .005, respectively, as determined by the Mann-Whitney U test). At E12, CFU-S activity was maintained in these cultured tissues. These data indicate that both the E11 aorta and UGRs are capable of significantly increasing CFU-S₁₁ in a manner that is independent of influx from other embryonic sites. This suggests that the increased CFU-S₁₁ activity between E11 and E12 (P < .02, as determined by the t test) reflects in situ generation and/or expansion of CFU-S₁₁. Interestingly, the culture of E11 UGR explants results in the greatest increase in CFU-S₁₁ numbers. This result is reminiscent of the dramatic increase in HSCs in the E11 UGR explant cultures we recently observed9 and leads us to speculate that this increase in CFU-S₁₁ results from their differentiation from newly formed HSCs in the UGR cultures. Indeed, CFU-S₁₁ were found in all the same sites where the first HSCs were found. However, the observation that CFU-S₁₁ were detected in the UGRs before HSCs emerged in this site (Figure 2D) leaves open the question of how these 2 hematopoietic cell types are related to each other during early embryogenesis. A possibility is that CFU-S₁₁and HSCs are derived separately from different precursors, such as putative hemangioblasts. This would be in agreement with data showing multiple sites of hematopoietic birth in avian embryos.14Thus, the CFU-S we detected may be derived by 2 different means.

In the present study we report the temporal and spatial emergence of CFU-S₁₁ activity within the mouse AGM. We show that from mid-E10 onward, CFU-S₁₁ are increasingly found both in the area of the aorta and in the UGRs, and that CFU-S₁₁ can be increased autonomously in both sites. It is of great interest to further examine the cells and factors involved in the increase of CFU-S₁₁ activity as well as the possible lineage relationships between putative hemangioblasts, CFU-S₁₁ and HSCs. Such studies are likely to provide additional insight in the origin and ex vivo manipulation of the clinically important HSCs.

We would like to thank Alexander Medvinsky and all the members of the laboratory for their helpful discussions. We also thank Rob Ploemacher and Sjaak Philipsen for critical comments on this manuscript and Jeroen Essers for help with the figures.