The analysis of the first steps of hematopoietic development historically performed though the study of the yolk sac, is somehow limited by the poor level of available precursors, as well as their extremely fast differentiation. Great progress in the understanding of the development of primitive hematopoiesis was achieved, owing to the ES model, once the similarity of developmental sequence with that which occurs in vivo was firmly established . The availability of large number of ES cells, amenable to genetic manipulations, further refined our knowledge of the mechanisms involved (e.g. through invalidated ES cells rescue experiments). Unfortunately, the gathered information cannot be easily placed back to back to similar experiments conducted in the "normal" YS developmental context.
Here we provide two gene transfer protocols, whose combination may allow a thorough study of both early and late events of hematopoietic development in the YS. Thanks to two markers which make it possible to discriminate between mesodermal/pre-hematopoietic (lmo2+/β-H1-)  and their erythroid progeny (lmo2+/β-H1+), we precisely determined the stage when the YS blood islands are enriched in immature precursors. This expression analysis also allowed to precisely locate the prospective blood islands in intact 7 dpc YS. We next used this delineation to perform in situ electroporation, which keeps intact the mesoderm-endoderm interactions shown to be crucial for proper hematopoietic development at these stages [28,29]. This approach was proven here to be effective to target mesodermal/pre-hematopoietic precursors and to be completely compatible with a normal viability and hematopoieisis of the transfected mesodermal/pre-hematopoietic precursors upon organ culture and subsequent in vitro culture on OP9 stromal cells. Thus, our work provides for the first time a detailed procedure allowing the analysis of the very early steps (determination) of primitive hematopoiesis. However, although our protocol ensures a significant enrichment of hematopoietic cells in the transfected population (three folds), the overall yield remains low (about 1 transfected mesodermal/pre-hematopoietic precursors per YS) for two reasons: 1- the intrinsic paucity of this precursor type in the 7.5 YS (about 10, after the 3 days in organ culture); 2- the relatively low, albeit usual in related experiments [9,23], level of gene transfer through in situ electroporation. Yet, this limitation does not disqualify the procedure since qualitative analyses following transfer of candidate genes can be performed.
A complementary protocol was then provided to transfer gene in « late » (8 dpc) YS which contains numerous determined cells. At this stage, hematopoietic precursors are much less precisely localized than their mesodermal ancestors, but rather dispersed and diluted among an abundant differentiated (essentially erythroid) progeny. Moreover, the differentiation of these cells is not anymore dependent upon interaction with endodermal cells. We thus reasoned that a « blind » (i.e. non selective) but efficient method would be the most adapted and therefore performed ecotropic retrovirus-mediated gene transfer onto isolated YS cells. A similar strategy has been previously used to express HoxB4 into YS cells at about the same stage . However, neither indication about the efficiency of the method, nor quantitative data were provided. This is of particular importance given that the nature of the transduced gene in this study may confer such a high competitive growth/survival advantage that even a weak efficiency of transduction may have been sufficient. We show here that even for a «neutral» gene (eGFP), the transfer through ecotropic retrovirus in 8 dpc YS hematopoietic precursors is very potent, therefore overcoming the problem of precursor dilution. Indeed, over 60% of the YS cells were transduced using a MOI as low as 1. In fact, YS cells appear highly susceptible to retroviral transduction as a very short contact (4–6 hours) with the viral supernatant is sufficient to produce a maximal transduction which therefore minimizes the potential side effects of concentrated viral supernatant or of polybrene, and allows a precise temporal targeting of the beginning of gene transfer. Moreover, we show that including a 24 h organ culture step before exposure to the viral supernatant markedly improves the targeting of immature precursors. This may result from an increased precursor frequency, as evidenced by the higher recovery of clonogenic precursors obtained when GFP+ cells are sorted from OrgD1-YS. As a consequence, even though the protocol that includes an organ culture step leads to a lower percentage of transduced cells, GFP+ cells steadily increase from day 4 post-infection, while they rapidly vanish when recovered from freshly transduced YS. Further analysis of the hematopoietic progeny of transduced cells is therefore easier under these conditions. Confirming the high permissiveness of YS cells to retroviral transduction, we found that a MPI vector carrying a 2 kb cDNA is almost as efficient as the empty one at transducing OrgD1-YS hematopoietic precursors.
As for in situ electroporation of 7 dpc YS, the recovery, survival and differentiation potentials of retrovirally transduced YS cells appear undistinguishable from that of non-transduced cells. Thus, retroviral transduction appears as a very convenient way to genetically modify hematopoietic precursors from 8 dpc YS.
Two potential problems, however, limit the use of retroviral transduction:
1- in our conditions, 7 dpc YS cells (i.e. containing mesodermal/pre-hematopoietic precursors cells), which are amenable to gene transfer by in situ electroporation, appear almost totally refractory to retroviral transduction. The reason for this pitfall is unknown, but we suspect that, at this early development stage, mesodermal cells have not yet reached some degree of commitment/differentiation that allows retroviral transduction. Indeed, 7 dpc YS mesodermal/pre-hematopoietic precursors become efficiently transduced after a two days of organ culture, which essentially recapitulates their in vivo progression toward committed precursors.
2- Retroviral vector are stably integrated, and therefore transduced cells are indelibly genetically modified. This is confirmed by the expression of GFP after at least two weeks of culture (latest time point analysed) in retrovirally transduced cells from late YS. In contrast while GFP is still detectable in electroporated cells from 7 dpc YS after 4 days, it completely vanished a few days later. In view of this limitation, we show that our in situ electroporation protocol transduce hematopoietic precursors from 8 dpc YS and thus deserves to be used notably if a transient expression is needed. Note that a 4 days expression may appear somewhat long with respect to the accelerated dynamic of embryonic events, especially in the case of early hematopoiesis. However, given the intense proliferation of early YS cells, we suspect that the plasmid delivered by in situ electroporation is in fact rapidly diluted so that the expression of labile proteins will be expected to be more really « transient » than that of GFP, whose half life exceeds 24 hours.