Erythropoietin Regulates GATA-1 Function and Erythroid Differentiation by a PKD-HDAC5 Signal Transduction Pathway

Abstract 85

How erythropoietin (Epo) signaling promotes erythroid differentiation remains unclear. Epo is known to regulate the function of the master regulator of erythroid gene transcription, GATA-1. Using murine proerythroblasts engineered with a conditional GATA-1-ER fusion, G1ER cells, Gregory et al. showed that GATA-1 induction of erythroid differentiation required signaling by Epo; in cultures lacking Epo, activation of GATA-1 caused cell death without differentiation (Blood 94:87-96, 1999). Although several models have been proposed for Epo regulation of GATA-1, no mechanism has been established. While analyzing PKC isozymes regulated by Epo and iron in primary human erythroblasts, our lab identified PKCμ , aka PKD, as activated in a dosage-dependent manner by Epo. Subsequent studies in Epo-starved G1ER cells, and in primary human progenitors, confirmed direct Epo induction of PKD phosphorylation on Ser744/748. A major PKD pathway involves its phosphorylation of class II HDACs (4, 5, 7 and 9), leading to release of HDAC-bound transcription factors (e.g. MEF2) from tonic repression. Watamoto et al. previously identified GATA-1 as a class II HDAC regulated factor, displaying physical and functional interaction with HDAC5 (Oncogene 22:9176-9184, 2003). Accordingly, we employed G1ER cells to determine whether Epo signaling to GATA-1 involved the PKD-HDAC pathway. In G1ER cells cultured in stem cell factor (SCF) minus Epo, endogenous HDAC5, but not HDAC4, co-immunoprecipitated with GATA-1. Epo stimulation induced dissociation of the HDAC5-GATA-1 complex, without affecting levels of either factor. The function of erythroid PKD-HDAC signaling was addressed by pharmacologic and shRNA approaches. Initial experiments tested whether inhibition of HDAC activity could substitute for Epo signaling in G1ER cells undergoing estradiol activation of GATA-1-ER. As described, GATA-1 activation in cells grown in SCF without Epo caused >90% cell death with minimal hemoglobinization after 48 hours. Addition of the HDAC inhibitor SAHA completely rescued viability and partially restored hemoglobinization of cells undergoing GATA-1 activation in the absence of Epo. Along similar lines, shRNA knockdown of HDAC5, but not HDAC4, significantly enhanced viability and hemoglobinization of G1ER cells undergoing GATA-1 activation under limiting Epo concentrations (0.05 U/ml). Conversely, shRNA knockdown of PKD3 impaired hemoglobinization of cells undergoing GATA-1 activation in the presence of Epo. Further implicating PKD in erythropoiesis, differentiation of G1ER cells was blocked by the kinase inhibitor Gö6976, which targets PKC and PKD, but not by the related compound Gö6983, which targets only PKC. Similar results were obtained with primary human progenitors, in which the PKD inhibitor Gö6976 but not the classical PKC inhibitor Gö6983 inhibited erythroid differentiation. In vivo roles of PKD-HDAC signaling in erythropoiesis were addressed by studying HDAC5-/- mice. In contrast to wild type counterparts, adult HDAC5-/- mice showed elevated steady state hematocrits (56.2 ± 0.8 vs 39.3 ± 1.4, P < 0.0001), resulting from increased MCVs (68.8 ± 0.7 vs 48.2 ± 0.1, P < 0.0001). In response to PHZ-induced hemolytic anemia, HDAC5-/- mice showed higher nadir RBC counts (5.0 ± 0.3 vs 3.2 ± 0.2, P < 0.001) and displayed an unexpected increase in Hb levels at the nadir point (15.9 ± 0.8 vs 8.6 ± 0.5, P < 0.0001). Spleens and livers obtained post recovery showed increased iron deposition in HDAC5-/- mice, consistent with increased net red cell turnover. To determine whether the observed abnormalities were cell-intrinsic, we sorted CD71++ Ter119- erythroid progenitors from the marrows of adult HDAC5-/- and wild type mice. The HDAC5-/- progenitors, but not wild type counterparts, displayed erythroid differentiation, manifested by Ter119 upregulation, in medium with no or low Epo (0.01-0.05 U/ml). Our data thus implicate PKD-HDAC signaling in Epo regulation of GATA-1 function and thereby provide a mechanistic basis for an instructional function of the Epo receptor. In addition, activation of PKD in non-erythroid cells could potentially explain some of the poorly-understood clinical complications of Epo therapy.