Research Article: CREB: A Key Regulator of Normal and Neoplastic Hematopoiesis

Date Published: August 27, 2009

Publisher: Hindawi Publishing Corporation

Author(s): Salemiz Sandoval, Martina Pigazzi, Kathleen M. Sakamoto.


The cAMP response element-binding protein (CREB) is a nuclear transcription factor downstream of cell surface receptors and mitogens that is critical for normal and neoplastic hematopoiesis. Previous work from our laboratory demonstrated that a majority of patients with acute myeloid leukemia (AML) and acute lymphoid leukemia (ALL) overexpress CREB in the bone marrow. To understand the role of CREB in leukemogenesis, we examined the biological effect of CREB overexpression on primary leukemia cells, leukemia cell lines, and CREB overexpressing transgenic mice. Our results demonstrated that CREB overexpression leads to an increase in cellular proliferation and survival. Furthermore, CREB transgenic mice develop a myeloproliferative disorder with aberrant myelopoiesis in both the bone marrow and spleen. Additional research from other groups has shown that the expression of the cAMP early inducible repressor (ICER), a CREB repressor, is also deregulated in leukemias. And, miR-34b, a microRNA that negative regulates CREB expression, is expressed at lower levels in myeloid leukemia cell lines compared to that of healthy bone marrow. Taken together, these data suggest that CREB plays a role in cellular transformation. The data also suggest that CREB-specific signaling pathways could possibly serve as potential targets for therapeutic intervention.

Partial Text

CREB is a 43-kDa protein, memberof the CREB/ATF-1 family of transcription factors, conserved from Drosophila to humans [1]. In mammals, CREB family members include CREB, cAMP-responsive element modulator (CREM), and activating transcription factor 1 (ATF-1). CREB and ATF-1 are ubiquitously expressed in all tissues. The expression of CREM, however, is tissue specific and developmentally regulated. This family of transcription factors contains a 60 amino acid kinase-inducible domain (KID) with several phosphorylation sites, two hydrophobic glutamine-rich transactivation domains, Q1 and Q2, that function as constitutive activators in vitro, and a basic leucine zipper (bZip) dimerization domain. CREB is activated through phosphorylation at serine 133 in response to a variety of cellular and mitogen stress signals. These include peptide hormones, neurotransmitters, calcium influx, and growth factors [1–3]. Upon activation, CREB binds as a dimer to the cAMP response element (CRE), TGACGTCA, or CRE half sites CGTCA/TGACG, where it promotes the recruitment of the transcriptional coactivators CREB binding protein (CBP) and p300. These coactivators then promote the recruitment of components of the basal transcriptional machinery to initiate transcription of CREB target genes [3–5].

There are several serine/threonine kinases that have been reported to activate CREB. Stimuli such as cAMP, calcium, and growth factors and cellular stress activate kinases such as ribosomal protein S6 kinase (pp90rsk), protein kinase A (PKA), protein kinase C (PKC), protein kinase B/AKT, and (mitogen- and stress-activated protein kinase) MSK-1, subsequently activate CREB [2]. Mitogenic and ultraviolet stress, for example, lead to the activation of mitogen/stresses activated kinase-1 (MSK-1), a pp90rsk family member that in turn phosphorylates CREB [8]. CREB phosphorylation is abrogated in fibroblasts from MSK-1/MSK-2 double knockout mice when stimulated with mitogens or when it is under cellular stress [9]. CREB is also phosphorylated in response to Granulocyte-Macrophage Colony Simulating Factor (GM-CSF) by pp90rsk in myeloid cells (Figure 1), leading to the activation of immediate early genes, c-fos, activator protein 1 (AP-1/junB), and early growth response protein 1 (egr-1) [10]. And CREB activation by MAPK and AKT/B enhances the survival of cultured cells [11]. In 2002, Raes and others characterized a new signaling pathway leading to CREB activation. Using L929 murine fibrosarcoma and ρ0 143B human osteosarcoma cell lines, this group demonstrated that CREB was phosphorylated by Calmodulin kinase IV (CAMKIV) as a consequence of mitochondrial dysfunction. In these mitochondrially impaired cells, CREB is constitutively activated as a result of high intracellular calcium levels that disrupt the association between CaMKIV and the protein phosphatase 2A (PP2A), resulting in constitutively activated CAMKIV [12]. These CREB kinases are among a long list of kinases that activate CREB in response of many extracellular stimuli.

CREB has been implicated in a great number of cellular functions, including metabolism, proliferation, apoptosis, and differentiation. Studies have demonstrated that CREB is phosphorylated in response to up to 300 different stimuli [3]. And upon activation, CREB enhances the expression of up to 5000 putative genes [5]. More stringent genome wide analysis for CREB binding motifs identified 1349 mouse and 1663 human CREB binding sites [13]. A quarter of the CRE-containing sequences function in cellular metabolism. In the liver, for example, CREB regulates gluconeogenesis, through phosphoenol pyruvate carboxykinase [14, 15]. Other metabolic enzymes, such as pyruvate carboxylase, ornithine decarboxylase, and lactate dehydrogenase contain CRE consensus sites in their promoters [2]. It has also been well characterized that CREB plays a critical role in survival. In sympathetic and cerebral neurons, nerve growth factor (NGF) and brain-derived neurotrophic factor (BDNF) stimulate survival by activating the expression of the antiapoptotic protein B-cell lymphoma 2 (bcl-2) [16, 17]. This was further shown by overexpression of a dominant negative form of CREB in these cells resulting in increased cell death. Moreover, this effect was reverted by overexpression of bcl-2. CREB has also been shown to regulate proliferation through cyclins A1 and D1 [2, 18, 19]. For instance, knockdown of CREB in both TF-1 and K562 cells leads to a decrease in the expression of cell cycle regulators, Cyclin A1 (Figure 1) and Cyclin D1.

Many of the CREB target genes identified function in cell growth, survival, and cell-cycle regulation, and their aberrant expression has been associated with cancer. The genes identified include genes that enhance proliferation, cell cycle progression, and survival, such as c-fos, cyclins A1 and D1, and survival, such as the bcl-2 protein. A role for CREB family members in cancer was first identified in clear-cell sarcomas of the soft tissues (CSSTs). Most CSSTs contain the chromosomal translocation t(12;22)(q13;q12) that fuses the DNA-binding and bZip domain of ATF1 to the Ewing’s sarcoma gene, EWS. The EWS-ATF1 fusion protein promotes proliferation and inhibits apoptosis [21–23]. Recently, an EWS-CREB fusion protein was also identified in a subset of CSST patients [24]. EWS-ATF1 and EWS-CREB fusion onco-proteins are constitutively active and enhance the expression of CREB target genes, independent of growth or signal stimulus.

Hematopoiesis involves the expansion and differentiation of a small population of pluripotent stem cells into progenitor cells. Maturation of these stem cells requires the exposure of these cells to a combination of differentiation and growth signals. CREB is downstream of cell surface receptors and mitogens that control normal hematopoiesis and leukemogenesis. Recent studies have demonstrated that CREB modulates the expression of genes that regulate hematopoiesis [20, 34]. In the human erythroid cell line, TF-1 for example, CREB was shown to regulate the expression of immediate early gene, egr-1 [10, 35]. In these cells, interleukin-3 (IL-3) and GM-CSF signaling results in CREB activation of egr-1 in response to pp90RSK activation through a Mitogen-activated protein/extracellular regulated kinase- (MEK-) dependant signaling pathway. Other studies have shown that IL-3 and GM-CSF, but not IL-4, potentiate CREB activation through PKC-ε in TF-1 cells [36]. Egr-1 is critical for transcription of myeloid-specific proteins that function as determinants of myeloid cell proliferation and differentiation. CREB also appears to play a role in megakaryocytic differentiation [36]. Studies performed in the biphenotypic cell line, HEL (erythroid/megakaryocytic), and CD34+ cells from normal patients show that thrombopoietin (TPO), and forskolin (FK), and phorbol myristate acetate (PMA) leads to increased activation of CREB through a Mitogen-activated protein kinase- (MAPK-) dependent mechanism [37].

The mechanism of CREB overexpression in leukemia remains unknown. MicroRNAs (miRNAs) are small noncoding RNAs that can either positively or negatively regulate gene expression principally through translational repression and targeting mRNA for degradation. There is recent evidence that microRNAs also regulate CREB expression. Approximately 30% of human genes possess conserved miRNA binding sites and are presumed to be regulated by microRNAs [50]. Generally, miRNAs bind mRNA sequences located at the 3’-untranslated region (UTR) with imperfect complementarity. MicroRNAs avoid target mRNA on polysomes, that would lead to a block in translation or mRNA degradation. The mRNA partners of microRNAs, microRNA function, and their tissue specificity are currently being investigated in normal and diseased patient samples to increase our understanding of tumorigenesis.

In summary, we conclude that CREB is a proto-oncogene whose overexpression can potentiate transformation of hematopoietic cells. However, CREB’s biologic effect on proliferation and survival is not sufficient to induce leukemias in vivo. Similar observations have been described with other transgenic mouse models of leukemia. AML1-ETO, K-Ras and FLT3-internal tandem duplication (ITD) transgenic mice, for example, develop myeloproliferative disorders but not leukemias. Similarly, CREB in vivo contributes to the leukemia phenotype, but additional mutations or ‘hits’ are required for disease development.