Abstract 1622

Despite considerable concerns and efforts, the mechanism of action of hydroxyurea (HU) for the induction of fetal hemoglobin (HbF) remains elusive. For example, clinical studies with HU suggest that bone marrow reserve is critical for HbF response to HU, but the underlying mechanism remains unknown. We and others have demonstrated that HU activates the cGMP signaling pathway in erythroid cells, which plays a role in HbF induction. However, the mechanisms by which intracellular signals are transduced to downstream cascades of cyclic nucleotide-dependent pathways in erythroid cells treated with HU remain to be established. Here we present evidence that HU induces HbF expression by activating the cAMP signaling pathway through two independent mechanisms: cAMP and cGMP. To study signal transduction by HU in cyclic nucleotide-dependent pathways, we initially focused on identifying substrates for cAMP-dependent protein kinase (PKA) and cGMP-dependent protein kinase (PKG) that are expressed in erythroid-lineage cells. We found that vasodilator-stimulated phosphoprotein (VASP), which is a 46/50-kDa phosphoprotein expressed in platelets at high levels, is also expressed in erythroid-lineage cells. VASP can be phosphorylated by cyclic nucleotide-elevating agents such as forskolin (activator of adenylate cyclase) and nitric oxide donors (activator of soluble guanylate cyclase). Interestingly, cAMP and cGMP phosphorylate distinct serine residues of VASP; Ser157 is phosphorylated by cAMP-elevating agents, while cGMP-elevating agents phosphorylated Ser239. Although HU increased both intracellular cAMP and cGMP levels in CD34+-derived erythroblasts, we found that Ser157, but not Ser239, is phosphorylated in adult erythroid cells treated with HU, suggesting activation of the cAMP signaling pathway. However, HU-induced HbF expression was down-regulated by inhibiting the activity of adenylate cyclase or soluble guanylate cyclase, suggesting that both enzymes are involved in HU-induced HbF expression. Our studies found that HU decreased the expression of cGMP-inhibitable phosphodiesterase 3B in a manner dependent on soluble guanylate cyclase, resulting in activation of the cAMP signaling pathway. Although a recent study showed that HU directly activates soluble guanylate cyclase, our studies showed that HU is unable to directly stimulate the enzyme activity of adenylate cyclase. Furthermore, HU induced the expression of cyclooxygenase-1 (COX-1) and increased the production of prostaglandin E2 (PGE2) that activates adenylate cyclase through G protein-coupled E-prostanoid receptors. Plasma PGE2 levels were also elevated in sickle cell patients upon HU therapy. These results demonstrate that HU induces HbF expression by activating the cAMP pathway by cAMP- and cGMP-dependent mechanisms, producing redundancy in the response of HbF to HU. Both cAMP and cGMP may represent major molecules that transduce signals from HU to the fetal globin gene. It is known that non-erythroid cells such as leukocytes and monocytes produce a large amount of PGE2. Thus, the involvement of PGE2 in HU-induced HbF expression may suggest an important role of non-erythroid cells as well as bone marrow reserve in the induction of HbF expression. More interestingly, several single nucleotide polymorphisms with amino acid changes have been demonstrated for COX-1; some genetic variants exhibit reduced COX activities. If SCD patients have some mutations in the COX-1 gene, such patients might be resistant to HU therapy.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution