AKT induces erythroid-cell maturation of JAK2-deficient fetal liver progenitor cells and is required for Epo regulation of erythroid-cell differentiation

Protein kinase B (PKB) or AKT is phosphorylated and activated in response to the erythropoietin (Epo) engagement of its receptor, downstream of the PI3 kinase. In addition to the regulation of survival, AKT is involved in many fundamental biologic processes, including the regulation of cell cycle, differentiation, and intermediary metabolism, and is often activated in human cancers.^1-3  Activation of the PI3 kinase signaling pathway,^4-12  alone in the absence of other signaling pathways, may be required and sufficient, under certain conditions, to sustain erythroid-cell development.^4,13,14  In addition, the PI3 kinase signaling pathway, and AKT specifically, are associated with the pathogenesis of polycythemia vera in which they may play a role.^15-17 

Here the potential role of AKT kinase in Epo-induced maturation of primary fetal liver erythroid progenitor cells using an in vitro system we previously established¹⁴  was investigated.

Isolation of E12.5 JAK2^-/- and E14 wild-type fetal liver cells, fluorescence-activated cell sorting (FACS) of TER 119^- (glycophorin A–negative) cells¹⁸  enriched for hematopoietic progenitors, and erythroid colony-forming unit (CFU-E) assay were previously described.^14,19 

TER 119^- wild-type (E14) and JAK2^-/- (E12.5) fetal liver cells (2 × 10⁵/mL) were resuspended in viral supernatants and plated on 60 mm RetroNectin– (chimeric fibronectin peptide; Takara Biomedicals, Osaka, Japan) coated dishes as previously described^14,19  in the presence of 100 ng/mL each of interleukin-6 (IL-6) and Steel factor (SF) (PeproTech, Rocky Hill, NJ). GFP-positive cells were FACS sorted the next day and cultured under the same conditions with or without Epo (2 U/mL) for another 24 hours.

Duplices of oligonucleotides were cloned into BglII/HindIII sites of MSCV-U3-H1 plasmid and subcloned subsequently into NotI/ScaI sites of pMSCV-puromycin-IRES-EGFP-U3-H1 plasmid (provided by B.L.).

Real-time polymerase chain reaction (PCR) was performed in duplicates on LightCycler 2.0 (Roche, Indianapolis, IN) using SYBR Green Taq Ready-Mix (Sigma, St Louis, MO) (as described by Ghaffari et al²⁰ ). Gene-specific primers were designed to span intron-exon boundary by Primer Express 2.0 (ABI; Applied Biosystems, Foster City, CA). Relative quantification of gene expression between multiple samples was achieved by normalization against 2 of endogenous Hprt, β-actin, and/or Rpl32 using the LightCycler Relative Quantification Software (Roche).

JAK2-deficient and TER 119^- wild-type fetal liver cells are deprived of mature erythroid cells. As expected, erythroid maturation of CFU-E progenitors contained in TER 119^- wild-type (Figure 1A) and JAK2^-/- fetal liver cells (Figure 1C), as determined by diaminobenzidine staining of hemoglobin, required Epo and a functional JAK2 tyrosine kinase. Interestingly, overexpression of a constitutively active form of AKT (AKT^*; Flag-tagged myristylated AKT) (Figure 1A,C, lane 4) but not wild-type AKT (Figure 1A,C, lane 3) overrides the need for Epo and JAK2. Activation and not merely overexpression of AKT is thus required for induction of erythroid differentiation of fetal liver progenitor cells. The size of the clusters and the degree of hemoglobinization was comparable between AKT^*- and control-transduced cells of either JAK2^-/- or wild-type origin. Overexpression of a dominant negative (DN) form of AKT in erythroid progenitors resulted in a significant decrease in the number of mature CFU-E–derived colonies in the presence of Epo (Figure 1A, lane 5) without a major effect on the total cell numbers, suggesting that the inhibitory effect of AKT DN is not due to apoptosis. Equal viral titers used in these experiments generated routinely similar levels of expression of AKT wild-type and mutants in heterologous cells (Figure 1B). Activated AKT and Epo induced similar levels of expression of β-globin in differentiating wild-type progenitor cells (Figure 1D). Expression of the β-globin gene was up-regulated by 2-fold after 14 hours and increased with time in activated AKT-transduced progenitor cells in culture (Figure 1E). The degree of maturation of JAK2^-/- fetal liver cells transduced with either an active AKT in the absence of Epo or with JAK2 in the presence of Epo was comparable as judged by the up-regulation of TER 119 (glycophorin A),¹⁸  a marker of post–CFU-E erythroid cells (Figure 1F), and by their morphology and Wright Giemsa staining (Figure 1G). These results demonstrate that AKT supports Epo-mediated erythroid-cell maturation without the need for any additional signaling protein that requires JAK2 for its activation.

To further determine the stage of erythroid differentiation, the expression of TER 119 and CD71 (transferrin receptor) in transduced cells was analyzed. In the absence of Epo, expression of activated AKT (Figure 2A, bottom right panel, and Figure 2B, lane 3), in contrast to wild-type AKT (Figure 2A, bottom left panel, and Figure 2B, lane 2), induced maturation of erythroid cells to the CD71^-/low TER 119⁺ stage of differentiation. Overexpression of active AKT^* had some mild inhibitory effect in the later stages of fetal liver cell maturation (Figure 2A; compare top and bottom right panels), although the effect did not seem to be statistically significant in the limit of the number of experiments performed (Figure 2B; compare populations of CD71^- TER 119⁺ in lanes 3 and 4).

We next investigated whether expression of activated AKT resulted in a survival advantage that facilitated erythroid-cell maturation. To our surprise, the survival rate of activated AKT-transduced cells (in the absence of Epo) was significantly lower (Figure 2C, right plot) than that of control cells cultured in Epo (Figure 2C, left plot) as measured by annexin V (an early measure of apoptosis) binding. These results suggest that AKT may not substitute for the apoptosis-suppressing function of the Epo-STAT5-Bcl-_XL pathway.

The enhanced erythroid maturation of AKT^*-transduced cells was not accompanied by an increase in total cell number. AKT^*-transduced TER 119^- WT fetal liver cells cultured in the absence of Epo were found less in the S phase (S, 38.5%; G₀-G₁, 46.8%) as compared with cells cultured in Epo (S, 47.5%; G₀-G₁, 42%), suggesting that AKT induction of erythroid differentiation was not only due to its enhancement of cell-cycle progression. Thus, the ability of activated AKT to induce fetal liver erythroid-cell differentiation does not solely stem from its potential to promote cell survival or proliferation.

AKT involvement was further evaluated by using RNA interference to inhibit the expression of AKT during Epo-induced differentiation of primary fetal liver cells. Retrovirally transduced fetal liver progenitor-derived cells expressing short hairpin RNA (shRNA) targeting AKT1 and AKT2 exhibited a significant decrease in the expression of red cell–specific genes β-globin and Alas2 (Figure 2D, lane 2) as compared with controls. Inhibition of expression of AKT also induced a significant suppression of the number but not the size of CFU-E–derived colonies consistent with an inhibition of erythroid differentiation of fetal liver cells (data not shown). The presence of inhibitors of caspase 3, a general inhibitor of apoptosis in cultured fetal liver cells, did not affect these results. Collectively, these results further confirmed that the effect of AKT on erythropoiesis is not limited to its regulation of apoptosis.

The combined data from both JAK2^-/- and wild-type fetal liver cells demonstrate that AKT mediates a signal downstream of EpoR,²³  distinct from its survival signal, that supports differentiation of fetal liver erythroid progenitor cells. These data are consistent with an in vivo role for AKT in normal²⁴  and in polycythemia vera^15-17  hematopoiesis.

We thank Dr Zhou Songyang (Baylor College of Medicine) for AKT constructs,²⁵  Dr Klaus Pfeffer (Technical University of Munich, Germany) for JAK2^+/- mice, and Glenn Paradis (Massachusetts Institute of Technology [MIT], Cancer Center) for flow cytometry. We are grateful to Drs Gordon Keller, Bill Schiemann, and Nai-Wen Chi for critical reviews of the manuscript.