POU4F1 Is Associated with t(8;21) AML and Contributes Directly to Its Unique Transcriptional Signature.

Abstract 2623

Poster Board II-599

The t(8;21)(q22;q22) translocation, present in 10–15% of adult acute myeloid leukemia (AML) cases produces the AML1/ETO (AE) fusion protein. The molecular mechanism of AE action is not well understood. Using Affymetrix RNA expression array analysis on RNA isolated from whole bone marrow (BM) aspirates from patients with de novo AML (n=113), we identified POU4F1 as an expression outlier uniquely associated with t(8;21). All t(8;21) positive patients in this cohort (n=4) had POU4F1 expression levels exceeding 2 SD above the mean for this cohort. None of the t(8;21) negative AML samples or normal CD34+ BM samples (n=5) had POU4F1 levels above this threshold. POU4F1 is a POU domain containing transcription factor that has been studied most extensively in neurons of the caudal CNS and peripheral sensory nervous system. Pou4f1 null mice die peripartum from motor neuron failure. Given the strong correlation between POU4F1 expression and t(8;21), we hypothesized that POU4F1 is a direct transcriptional target of AE and a mediator of its activity. We found that transient or stable over-expression of AE in primary mouse BM or human cell lines did not increase the levels of POU4F1 when analyzed by qRT-PCR, implying that POU4F1 over-expression in t(8;21) AML is AE independent. AE is known to cause increased self renewal of hematopoietic progenitor cells in vitro. To determine whether Pou4f1 may contribute to this phenotype we retrovirally transduced murine BM cells with AE, Pou4f1, or empty vector. We then performed weekly clonal serial replating of myeloid progenitors in cytokine enriched methocellulose media. Progenitors transduced with AE had robust replating activity (15–20/24 CFUs replated for at least five generations in three independent experiments), whereas cells transduced with Pou4f1 did not (1-4/24 CFUs replated; P <0.05). We next used Pou4f1 null mice to determine whether Pou4f1 is required for AE induced progenitor self renewal. We repeated the serial replating assay using Pou4f1 +/+, +/−, or −/− E14.5-16.5 fetal liver cells (3 independent experiments, n=3 mice per genotype). Regardless of Pou4f1 genotype, cells transduced with AE retained replating capacity for five generations. Taken together these results indicate Pou4f1 is neither necessary nor sufficient for enhanced self renewal of hematopoietic progenitors in vitro. To further elucidate the role of POU4F1 in t(8;21) AML, we investigated what effect Pou4f1 had on the transcriptional profile of hematopoietic cells. Pou4f1 +/+ or −/− fetal liver cells were retrovirally transduced to produce cells with high (MSCV-IRES-Pou4f1-GFP in +/+), normal (MSCV-IRES-GFP in +/+), or no (MSCV-IRES-GFP in −/−) Pou4f1 expression. GFP positive cells were sorted and RNA was prepared. Affymetrix MOE430v2.0 GeneChip microarrays were run and comparisons were made between the gene expression profiles of the three experimental groups (each group n=3). As expected, since Pou4f1 is not normally expressed in hematopoietic cells, the gene expression profiles between Pou4f1 wildtype and null cells showed no significant differences. Comparisons of the cells expressing high levels of Pou4f1 to Pou4f1 wildtype or null, however, revealed 114 differentially expressed genes (FDR<0.25). Using Gene Set Enrichment Analysis on our AML data, we found that the human homologs of these 114 genes were significantly enriched in t(8;21) positive compared to t(8;21) negative patients (FDR q value = 0.018), implying that the Pou4f1 gene expression signature is conserved from mice to humans and is retained throughout leukemic transformation. Remarkably, unsupervised clustering performed on the expression profiles of the 114 Pou4f1 signature genes alone was sufficient to separately group the t(8;21) patients from all the other AML patients. Of these 114 genes, the expression of eight (PPAPDC1B, PELI2, FAM69B, ROBO1, SOX4, CAV1, H2AFV, PLXDC2) was significantly increased (P<0.01 by t-test) up to 33-fold (CAV1) in t(8;21) positive vs. negative patients. In this cross species analysis we identified a POU4F1 gene expression signature that distinguishes between t(8;21) positive and negative AML samples. These surprising results directly refute the assumption that the gene expression profile of t(8;21) AML is solely attributable to AE and suggest that POU4F1 is an important independent contributor to the pathophysiology of these leukemias.