Cell expansion and maintenance of stemness

Manipulation of hematopoietic cells to expand output while maintaining stem-cell potential has been an elusive goal of experimental hematology. The development of a system using a chemically induced dimerizer and modified thrombopoietin receptor has now allowed the expansion of primitive hematopoiesis without sacrificing stem cells.

Hematopoietic-cell expansion represents a much-sought-after therapeutic goal of the biomedical sciences. With the cloning and characterization of a large and growing number of hematopoietic growth factors, a mechanism for hematopoietic expansion seemed to be at hand. However, ex vivo expansion strategies using cocktails of cytokines have failed to expand transplantable hematopoietic stem cells (HSCs). In contrast, most such approaches lead to the differentiation and extinction of the most primitive cells in the cultures. The explanation for these results is the requisite coupling of cell proliferation and differentiation that results when hematopoietic growth factors bind their cognate receptors.

The work of Abdel-Azim and colleagues in this issue of Blood has used a previously described cell-expansion strategy in a new target-cell population to massively expand hematopoietic cells of multiple lineages, including, apparently, the HSC. The approach involves chemically inducing dimerization of the cytoplasmic domain of the thrombopoietin receptor (c-Mpl) in highly purified, primitive human marrow cells. The rationale for this approach began with the discovery that c-Mpl and its ligand, thrombopoietin, provide important and nonredundant support for HSC survival and proliferation.¹ 

Hematopoietic growth factors act by binding to their cognate receptors, altering the conformation of the latter, resulting in cross-phosphorylation of 2 tethered Jak signaling kinases. Once phosphorylated, Jak kinases phosphorylate the receptors themselves as well as several secondary survival and proliferation signals, including signal transduction and activator of transcription 3 (STAT3) and STAT5, phosphoinositol-3-kinase (PI3K), and mitogen-activated protein kinases (MAPKs). Ultimately, some of these same signals lead to signal extinction, by inducing receptor internalization and STAT-induced expression of suppressors of cytokine signaling (SOCS) molecules, which block further Jak signaling.² 

Identification of the FK506 binding protein (FKBP), the target of the commonly used immunosuppressant drug FK506, and the demonstration by Spencer et al that a chemically synthesized dimeric form of FK506, FK1012, could artificially dimerize 2 molecules of FK506,³  led to the first chemical inducer of dimerization (CID) strategy. Following a minor modification in FKBP (F36V) to render it responsive to the nonimmunosuppressive AP20187compound, the stage was set to use this CID to mimic cytokine-induced cellular signaling. By transducing marrow cells with an FKBP (F36V)–c-Mpl fusion protein, Jin et al first established the ability of the CID approach to influence hematopoietic-cell proliferation.⁴  These efforts expanded mature blood cell production both in vitro and in vivo; however, HSC expansion was not demonstrable. By using highly purified CD34⁺/CD38⁻/lineage⁻/CD7⁻ human cells, Abdel-Azim and colleagues have moved this tech-nology closer to clinical utility.

Like all studies with clinical implications, the work of Abdel-Azim et al must be repeated by others to verify that the capacity to expand and maintain HSCs can be generalized to transplantation in humans. If so, this work opens a number of new avenues for the manipulation of human HSCs for therapeutic benefit, including the engineering of HSCs with therapeutic proteins, the expansion of HSCs when only small numbers are clinically available, and the ex vivo generation of multiple therapeutic products. But there is an equally important aspect to the mechanism described by Abdel-Azim and colleagues: understanding why it works.

As noted, dimerization of c-Mpl by thrombopoietin results in HSC survival, cellular proliferation and maturation into multiple types of committed hematopoietic progenitor cells, and ultimately, HSC extinction. In contrast, dimerization of a non–cell-membrane-bound form of the cytoplasmic domain of c-Mpl by a CID expands hematopoietic cells and maintains their “stemness.” In their discussion, Abdel-Azim and colleagues suggest that the difference between their study and prior work exists in the constant signaling induced by their engineered construct, one that cannot be down-modulated. Another possible explanation could be the failure of the CID-activated cytoplasmic c-Mpl domain to phosphorylate STAT5, the prerequisite for SOCS activation, an important mechanism for turning off growth signals. Or could it be the cytoplasmic site of origin of the c-Mpl signals? Numerous studies suggest that the subcellular site from which a signal emanates plays a major role in the physiological effect of that signal. Or is it, perhaps, the geometry of the induced receptor dimerization, which differs between thrombopoietin and the CID? Wilson et al, in their work with erythropoietin mimetic peptides and the erythropoietin receptor,⁵  revealed that as little as a 20° difference in homodimer conformational rotation turns a peptide agonist into an antagonist.

Thus, like all excellent studies, the work of Abdel-Azim and colleagues answers some questions, raises many others, and prods us to move a new technology forward; hopefully, such work will help us fulfill the goal of expanding hematopoietic cells for therapeutic benefit.