Systems Analysis Reveals Regulators of Apoptosis, Cell Cycle, Signal Transduction and Transcription as Novel Direct Targets of the Acute Promyelocytic Leukemia Fusion Protein NuMA-RAR α

We used the hCG-NuMA-RARα transgenic model of acute promyelocytic leukemia (APL) (Sukhai et al, 2004) to determine the downstream genetic targets of NuMA-RARα, one of five APL-associated fusion proteins (X-RARα). X-RARα retain the C-terminal domains of retinoic acid receptor (RAR) α, and are thought to interfere with retinoid signaling pathways and thus inhibit neutrophil differentiation. However, X-RARα have a wider range of DNA binding specificities compared to RARα, and may deregulate novel downstream targets that play critical roles in APL. Our group previously identified PPARγ signaling as being one such pathway deregulated by X-RARα (Kamel-Reid et al, 2003). We sought to identify additional, nonretinoid signaling, target genes of NuMA-RARα by using a combined experimental and in silico approach. Affymetrix oligonucleotide array analysis (mouse 430A arrays) was performed on RNA harvested from bone marrow cultures established from wild-type and transgenic mice. Genes deemed to be significantly deregulated and of biological relevance were validated by quantitative real-time RT-PCR. In our analysis, we identified 260 significantly over-expressed and 278 significantly under-expressed genes in transgenic bone marrow cultures, compared to wild-type. As NuMA-RARα is an aberrant transcriptional repressor, we focused on under-expressed genes: 251/278 were putative direct targets of NuMA-RARα. 82/251 were retinoid response genes; 169/251 were novel targets, external to retinoid signaling pathways. 150/251 had known function. Genes involved in regulation of apoptosis, cell cycle, signal transduction and transcription were over-represented within this dataset, in comparison to their frequency within the mouse genome. NuMA-RARα deregulated a number of myeloid transcription factors, including PU.1 (11.5-fold under-expression compared to wild-type), members of the C/EBP (3.5–5.0 fold under-expression) and GATA families (8.5–11.0-fold over-expression), as well as c-Myb and AML1. Binding sites for these transcription factors were overrepresented within the promoters of NuMA-RARα-deregulated genes. NuMA-RARα also deregulated a set of genes involved in cell cycle regulation and DNA damage response, including Gadd45α (1.9-fold under-expressed) and Dusp1 (3.8-fold under-expressed), and components of signal transduction pathways, including Jak2 (20.0-fold under-expressed). By mapping the chromosomal addresses of deregulated genes onto the murine genome, we determined that changes in chromosome copy number may not be responsible for deregulation. Finally, our comparison with other, previously published, microarray analyses indicated that NuMA-RARα shared some common target genes with other leukemia fusion genes, and that gene deregulation caused by NuMA-RARα could give rise to a molecular phenotype reminiscent of hematopoietic stem cells. Several genes identified as candidate downstream targets in these analyses were also previously identified as putative secondary events in APL mouse models. NuMA-RARα therefore has specific effects on the transcriptome, distinct from retinoid signaling, and from other leukemia fusion genes. Future functional studies are required to determine the cooperative relationship between these novel target genes and NuMA-RARα.