The use of hematopoietic stem cells for in utero transplantation to create permanent hematochimerism represents a new concept in fetal therapy, although this approach has provided quite heterogeneous results. Flake and Zanjani have provided an excellent updated review of the current knowledge of in utero stem cell transplantation and have formulated diverse possible reasons for its poor clinical success.1 In fact, the only clear success, or claims of success, has been obtained in fetuses affected by immunodeficiency syndromes such as severe combined immunodeficiency (SCID) or bare lymphocyte syndrome. Flake and Zanjani strongly support the possibility that in such immunodeficiency disorders there is a selective advantage for donor cells, which then overcome a biological barrier to engraftment in fetuses. To date, the nature of this biological barrier remains unknown.
Alternatively, or perhaps additionally, we have raised the possibility that the fetus may develop an allogeneic immune response that ultimately accounts for graft failure. Accordingly, it has been reported that fetal liver stem cells are capable of engraftment in a syngeneic, nonallogeneic host,2 and that successful allogeneic in utero transplantation of fetal stem cells in sheeps and monkeys occurs if both the donor and the recipient fetuses are in a preimmune condition.3 Finally, we have reported a case in which in utero fetus-to-fetus transplantation performed at the 19th week of gestation resulted in neither engraftment nor tolerance induction; rather, 2 years after birth, the recipient had developed a highly increased cytotoxic T-lymphocyte precursor (CTLp) frequency against donor cells.4
Toward this possibility, we have recently undertaken molecular, phenotypic, and functional studies aimed at identifying the presence of fully competent T lymphocytes in samples of fetal livers and cord blood. We have found the presence of mature VDJ TCRβ chain transcripts in fetal liver and cord blood cells taken from 7 to 16 weeks of gestation. T-cell clones obtained from fetal liver cells showed a mature TCR αβ+, CD8+ phenotype and displayed strong alloreactivity against allogeneic HLA class I molecules.5 The very low yield of such clones from fetal liver–derived T lymphocytes strongly supports the view that their frequency is rather low and we have estimated it at around 0.2/106 cells. Similar low, yet significant alloreactive response has been found within CD8+ T lymphocytes taken from cord blood.6
Based on our results, as well as experimental data from the literature, we favor the possibility that the presence of alloreactive T lymphocytes may explain the failure to engraft in fetuses older than 13 to 16 weeks. Overall, our results may provide useful informations on the stages of fetal T-cell development and can help in devising new strategies and planning further clinical trials in intrauterine transplantation.
Role of the alloimmune response after in utero hematopoietic stem cell transplantation
We appreciate Renda et al's comments regarding their important investigations into the presence of alloreactivity in the early gestational human fetus. We agree that the immune response likely plays an important role in the barrier to engraftment following in utero transplantation, particularly after the appearance of phenotypically mature T lymphocytes in the peripheral circulation. The question is not whether the immune system presents a barrier to engraftment but, rather, when the human immune response becomes an important consideration and what components of the immune response play an important role. It is clear from animal studies in sheep1-1,1-2 and mice1-3 that engraftment can occur across full allogeneic or xenogeneic1-4 barriers if transplantation is performed early enough in gestation. In studies that we have presented (manuscripts in preparation or submission),1-5,1-6 it is also clear that tolerance in the mouse model of in utero hematopoietic stem cell (HSC) transplantation can be achieved across full major histocompatibility complex (MHC) barriers and involves deletion of host-against-donor-reactive T cells, as well as donor-against-host-reactive T cells, and can be mediated by either direct or indirect antigen presentation, by donor- or host-derived antigen-presenting cells, respectively. Thus if the transplant is performed early enough in gestation and there is adequate antigen presentation in the thymus, it appears that T-cell mediated alloreactivity is not prohibitive to engraftment.
The natural immune system may also play a role in fetal immune response. NK cells are present early in gestation, but their function is poorly understood. We are in the process of investigating the effect of prenatal transplantation on the inhibitory receptor profile of host- and donor-derived NK populations.1-7 Until better studies are performed in defined animal models, the importance of the immune response and the parameters required for tolerance or immune sensitization will remain conjectural. There is a strong need for further direct studies, as Renda et al are pursuing, to define the immune response of the early gestational human fetus. These studies will allow informed extrapolation of findings in animal models to the human fetus for optimization of clinical in utero HSC transplantation.