Stem Cell Barcoding: Is it Time to Change Our Definition of a Hematopoietic Stem Cell?

Mature blood cells have a limited life span, necessitating the need for constant tissue regeneration. The standard model to explain the ability to maintain lifelong production of blood is based on the “hematopoietic tree,” in which a few numbers of hematopoietic stem cells (HSCs) go through a series of progressively lineage-restricted divisions to produce mature blood cells. For more than 50 years, HSCs and their downstream progeny have been studied to explore this remarkable ability to regenerate a highly dynamic tissue system without exhaustion and only relatively rare cases of neoplasia. 

The ability to experimentally evaluate HSCs is based largely on animal models of bone marrow transplantation. As part of their educational materials,¹  the National Institutes of Health (NIH) defines an HSC as “a cell isolated from the blood or bone marrow that can renew itself, can differentiate to a variety of specialized cells, can mobilize out of the bone marrow into circulating blood, and can undergo programmed cell death…” The NIH further states that “the ‘gold standard’ for proving that a cell derived from mouse bone marrow is indeed an HSC is… the cells are injected into a mouse that has received a dose of irradiation sufficient to kill its own blood-producing cells. If the mouse recovers and all types of blood cells reappear, the transplanted cells are deemed to have included stem cells.”

While experimentally this definition of an HSC has been useful, a fundamental problem with it is that most of us walking around today have not had a bone marrow transplant that requires an HSC to traffic through the bloodstream and home to the bone marrow niche, and rapidly divide to replenish the lost blood cells as a result of myeloablative conditioning. Much of what we know today about HSCs may therefore be largely influenced by transplantation and not reflective of normal homeostatic blood production.

To begin to define the regulation of native hematopoiesis, new work published by the Camargo laboratory describes clonal dynamics of hematopoiesis in mice using a novel cellular barcoding system.²  Mice were engineered with Sleeping Beauty Transposase, which is an enzyme that mediates the mobilization of a transposon, a DNA element that can jump to different positions within the genome. The Sleeping Beauty Transposase was under the control of doxycycline, allowing the initiation of the DNA transposition at a precise time in an adult mouse. The transposition results in a readable DNA sequence being randomly inserted into a unique position within each individual cell, and their subsequent progeny will have the same insertional site. This stable genomic tag allowed the researchers to then track blood production from individual clones, without the need for transplantation like prior lentiviral barcoding experiments.^3-6 

While tracking the mice approximately every six weeks for a period of more than 40 weeks after in situ genetic tagging, they found that the vast majority of granulocytes were produced by clones that were only present at a single time point. This suggests that native granulopoiesis is supported by a large number of successive clones that “awaken” at different time points throughout the life span of the organism, supporting the “clonal succession” model⁷  of hematopoiesis. When determining the potential of lymphoid clones, about half of the clones 10 months after genetic tagging were also present in myeloid cells, demonstrating that there is active production of blood from a pool of multipotent stem or progenitor cells. Remarkably, when evaluating genetic tags in granulocytes, very few of the genetic tags found in granulocytes were also found in lymphoid cells up to 45 weeks post-tagging. These data suggest that the bulk of clones producing granulocytes are myeloid-lineage–restricted and call into question the role of multipotent HSCs in the production of blood in the absence of transplantation. 

Since prior work has relied on transplantation to evaluate HSCs, the authors used their mouse model in a transplantation assay to determine the clonal repertoire in a mouse 64 weeks after genetic tagging, presumably when transient progenitors are exhausted and blood should be produced from HSCs. Surprisingly, only about 5 to 8 percent of the clones that were contributing to blood production in the donor mouse were found in the recipient mice. Furthermore, when genetic tags were evaluated in HSCs, less than 5 percent of the HSC tags were found in mature cells. Like Sleeping Beauty, the data here suggest that the classical “transplantable HSC” may be completely dormant in steady state.

Mice are considerably short-lived compared with humans, so there is potential that the results seen in this system may not be reflective of human HSCs. A plenary presentation at the 2014 ASH Annual Meeting highlighted the results of HSC clonal tracking for five years after gene therapy trials for adenosine deaminase–severe combined immunodeficiency and Wiskott-Aldrich Syndrome.⁸  Early post-transplantation, blood production in these patients was supported by waves of distinct clones. However, these waves of clonal output stabilized after 12 months, and consistent multilineage output of tagged HSCs was seen throughout a period of five years, without the manifestation of clonal quiescence phases. These data support more of a “clonal stability”⁷  model of hematopoiesis, where many stem cells contribute simultaneously and indefinitely to blood production. However, this “experiment” in humans relied on the same paradigm as those before it in mice; it required transplantation of HSCs.

The clonal output from an individual HSC (or progenitor cell) was recently monitored in humans without the need for transplantation.⁹  The authors tracked the output of individual clones due to a rare mutation of the phosphatidylinositol N-acetylglucosaminyl-transferase gene (PIG-A) that results in the loss of specific glycoproteins on the surface of HSCs and their progeny. The authors saw that the vast majority of blood output from an individual clone, particularly those contributing to granulopoiesis, was lineage restricted. These data, like that of the Camargo study in mice, suggest that human blood production at steady-state may come predominately from lineage-restricted progenitors rather than the classical multilineage, transplantable HSC.

Expanding the Sleeping Beauty model to settings of hematopoietic challenge such as major blood loss, sublethal myeloablation, aging, or infection is likely to provide further insights into the clonal contribution of HSC to hematopoiesis. In the future, we may have to change our definition of an HSC to distinguish the transplantable cell from that which contributes to normal hematopoiesis.