Stem cells take a shortcut to the bone marrow

One of the major problems limiting stem cell–based therapies, beyond simple transplantation of mononuclear cells or positive selection of CD34⁺ cells, is the absence of a clear understanding of the composition of the stem cell (HSC) pool in humans. The right cell must be targeted for the right therapy. For gene therapy, HSCs capable of permanent repopulation must be transduced, whereas for cancer therapy an HSC capable of rapidly generating granulocytes, platelets, and erythroid cells is also required. Xenotransplantation of human cells into preimmune sheep or, more commonly, immune-deficient mice provides powerful assay systems to characterize the HSC compartment. Clonal tracking of retrovirally transduced cord blood cells has identified individual HSCs (termed SCID repopulating cells, or SRCs) with short-term (ST-SRCs) and long-term (LT-SRCs) repopulation capacity. Cell purification studies indicate that LT-SRCs are highly enriched in the Lin⁻CD34⁺CD38⁻ cell fraction, while the Lin⁻CD34⁺CD38⁺ fraction contains ST-SRCs that lack self-renewal potential. Thus a picture of the human HSC compartment is emerging in terms of functional repopulation properties of different classes of HSCs, their frequency, cell surface markers, cell cycle status, and response to in vitro cytokine stimulation. Importantly, the mechanism of migration and homing/adhesion of HSCs within the xenoenvironment is being defined. The chemokine SDF-1, expressed by both human and murine bone marrow (BM) endothelium and stroma, and its cognate receptor, CXCR4, expressed on human progenitors play a key role. The SRCs possess the capacity for migration to SDF-1 and the in vitro motility of human CD34⁺cells correlates with their repopulation potential in patients receiving transplants.

All HSC repopulation assays rely on intravenous (IV) injection; a complex process involving circulation through blood, recognition and extravasation through BM vascular endothelium, and migration to a supportive microenvironment. It is possible that cells with intrinsic HSC function might exist but either do not survive in the circulation or do not possess the machinery for homing and retention, rendering them poorly detectable by traditional IV-based HSC assays. Along this line, a rare class of human CD34⁻-SRCs, found within the Lin⁻CD34⁻CD38⁻fraction, was tentatively identified. These CD34⁻-SRCs have low CXCR4 expression, poor responsiveness to SDF-1–mediated migration, and very low in vivo homing, leading to very limited repopulation. But upon in vitro culture with cytokines and/or stroma these Lin⁻CD34⁻CD38⁻ cells generate high numbers of Lin⁻CD34⁺CD38⁻ cells that now possess high levels of SDF-1–mediated migration and repopulation capacity, suggesting that these HSCs possess significant intrinsic HSC potential. Complicating this story is the realization that expression of CD34 on both murine and human HSCs is subject to modulation either developmentally (murine fetal HSC are CD34⁺; adult HSC, mostly CD34⁻), after culture, or following transplantation. Thus we currently do not know whether these rare human CD34⁻-SRC exist as a distinct class of HSC and whether they have clinical significance.

One approach to overcome these inherent limitations of IV-based repopulation assays is to deliver cells directly into the BM. In this issue, 2 manuscripts elegantly describe a more sensitive assay for human HSCs involving intra–bone marrow transplantation (IBMT) into immune-deficient mice. Yahata and colleagues (page 2905) used limiting dilution analysis and retroviral-mediated clonal analysis to show that the IBMT method detected 15-fold more Lin⁻CD34⁺CD38⁻-derived LT-SRCs than did IV injection (1 SRC per 44 cells vs 1 SRC per 660 cells, respectively). In addition, treatment with neutralizing anti-CXCR4, anti–VLA-4, or anti–VLA-5 antibodies alone or in combination indicated that SRC retention was still dependent on SDF-1/CXCR4 and VLA-4/VLA-5 interactions with their respective ligands. Wang and colleagues (page 2924) demonstrate that IBMT of Lin⁻CD34⁻ cells resulted in repopulation of all mice receiving transplants, in contrast to no engraftment following IV injection. The CD34⁻-SRCs were engrafted with somewhat slower kinetics were 10-fold less frequent than CD34⁺-SRCs, suggesting that these 2 HSC classes are distinct. Interestingly, the IBMT method also detected low levels of CD34⁺-SRCs within a fraction of Lin⁻CD34⁺ cells that were unable to migrate in vitro to SDF-1 and consequently were unable to engraft NOD/SCID mice when injected intravenously. These nonmigrating CD34⁺-SRCs are likely those shown to contain intracellular CXCR4 that oscillates to the cell surface.

Thus, direct delivery by IBMT reveals cells with intrinsic stem cell function that were previously poorly detectable by the traditional IV assay. Collectively, these studies indicate that the human HSC compartment might be more complex than previously thought and that novel classes of HSCs might still await discovery using this new assay. With the flurry of recent excitement concerning putative transdifferentiation of organ-specific stem cells and the induction of embryonic cells along various lineages, one can imagine that the IBMT method may provide a generalized method to detect cells with intrinsic HSC function. IBMT may also potentiate engraftment of other cell types such as mesenchymal progenitors to cure inherited bone diseases. Finally, although human experience with intraosseous transplantations goes back to the 1930s, there is little strong underlying experimental basis. These studies provide such a framework and some specific stem cell–mediated therapeutic strategies might benefit from the clinical use of IBMT. For example, if HSCs are delivered more efficiently into humans via IBMT as the xenotransplantation studies suggest, it may be possible to extend the use of cord blood transplantations to adults.