The cAMP Response Element Binding Protein (CREB) Overexpression Induces Myeloid Transformation in Zebrafish

Transcription factors has been widely confirmed to play a central role in acute leukemia, and among them the cyclic-adenosine monophosphate response element-binding protein (CREB) was identified to be involved in triggering acute myeloid leukemia. CREB protein overexpression has been found in the bone marrow of most pediatric patients with acute leukemia, and it has been shown to induce myeloid leukemia progression in vitro and in vivo by driving the upregulation of a series of target genes. CREB is evolutionarily conserved from invertebrates to human, and although the main focus of zebrafish research has traditionally been developmental biology, this model is currently used for cancer research. In fact, tumors induced in ZF have similar morphology and activated signaling pathways of human cancers.

We aimed to develop a ZF harboring CREB overexpression in myeloid precursors, and by monitoring the induced myeloid malignancy, we will characterize CREB signaling and its involvement in the myeloid transformation process.

Using a Multisite Gateway System we constructed a vector containing human-CREB gene fused to EGFP, driven by an early myeloid promoter, zPu.1, to induce CREB overexpression specifically in the myeloid lineage of ZF. We injected the EGFP-CREB plasmid into one-cell stage zebrafish embryos, and monitored its expression during early development. Results showed that CREB was expressed in ZF zones typical of myelopoiesis, such as in the intermediate cell mass and anterior lateral mesoderm migrating through the yolk from 12 to 48 hours post-fertilization (hpf), and few cells circulating throughout the embryo from 24 hpf. CREB transcriptional activity on cAMP response elements (CREs) was measured and confirmed by luciferase assay. The co-localization of EGFP-CREB with the CRE-mCherry reporter was seen by fluorescence microscopy analysis. To test CREB target gene expression, embryos injected with CREB or Empty vector were subjected to RNA extraction and RQ-PCR. CREB over-expression was documented (up to 103 fold), and c-myb, stat3, rb, runx1, cyclins A, B, D1 and E2 were found upregulate at 24 and 48 hpf. By RNA whole mount in situ hybridization, we revealed pu.1, mpo, gata1 and fli.1 increased signals, suggesting a general disruption of the main hematopoietic factors.

Ten CREB-injected zebrafish have been grown to adulthood and all of them displayed an abnormal/sick phenotype with abdominal enlargement and swelling from 10 to 12 months. Histochemical H&E staining performed on paraffin sections revealed an abdominal tumor and metastatic infiltration in kidney, skin, gills, muscles, adipose tissue. PAS staining identified the myeloid character of the tumor mass and kidney marrow. Wright-Giemsa and ANAE staining showed the predominance of clonal monocytes on appositions of tumor mass and kidney marrow of sick ZF. Tumor mass cell sorting displayed a clonal feature of the tumor with an enrichment of the myeloid-monocitic compartment. RNA extraction from the mass shows CREB overexpression as well as its target genes (c-myb, runx1, cyclin A1, cyclin B1, cyclin B2, cyclin E1), confirming CREB involvement in this tumor formation.

We demonstrated that CREB overexpression recapitulates myeloproliferative disorder in ZF, supporting the ZF as a suitable model for studying CREB-induced leukemia. The discovery through which targets CREB would preferentially mediate the myeloid transformation will help to unravel leukemogenesis. ZF model might be useful to test CREB directed drugs.

No relevant conflicts of interest to declare.