CREB Is Amplified in AML Blasts and Is Associated with an Increased Risk of Relapse and Decreased Event-Free Survival.

The cAMP Responsive Element Binding Protein (CREB) is a basic leucine zipper transcription factor that regulates cell proliferation and survival. Greater than 60% of AML patients have elevated CREB protein levels in the bone marrow. Transgenic mice that overexpress CREB in myeloid cells have monocytosis and splenomegaly. Bone marrow progenitor cells from CREB transgenic mice exhibit properties of transformed cells including immortalization and growth factor-independence in colony assays. Our results suggest that CREB acts as a proto-oncogene in hematopoietic cells. We first sought to understand the mechanism of CREB overexpression in AML blasts. Since both CREB protein and mRNA levels were increased in primary AML cells, we hypothesized that CREB could be amplified. We performed fluorescence in situ hybridization (FISH) on blast cells from 4 CREB overexpressing patients using a CREB specific BAC clone. The CREB gene has been previously localized to chromosome 2q32.3–q34. In blast cells from 3 out of 4 AML patients that overexpressed CREB, we detected 3 to 4 signals from one homolog and 1 signal from the other in over 100 interphase nuclei analyzed. However, in normal control lymphocytes in metaphase and CREB nonamplified bone marrow cells in interphase, the CREB BAC showed two clear and unambiguous signals. Three out of the 4 AML patients whose bone marrow overexpressed CREB had a normal karyotype with no evidence of aneuploidy or hyperdiploidy. These results indicate that in certain AML patients there are more than two copies of the CREB gene. To evaluate whether increased CREB expression predicts outcome in leukemia patients, Kaplan-Meier analysis was performed on data from 45 AML patients (14 age 18 or under and 31 over age 18). Time to relapse or persistent disease was markedly shorter in the 31 CREB overexpressing, i.e. CREB positive, patients in comparison to the 14 not expressing CREB (proportion with persistent disease or relapsing by 1 year 0.68 ± 0.10 vs. 0.23 ± 0.14, p<0.001 both before and after stratification on cytogenetics). Likewise, event-free survival was noticeably worse in the CREB positive group (p<0.003). The seven patients who showed persistent disease after initial induction therapy were all CREB positive, but excluding them from the analysis still yielded noticeably shorter relapse times for the CREB positive group (p=0.01). No variable other than CREB overexpression had a significant association with adverse time to relapse or event-free survival. Initial treatment, cytogenetics, AML subtype, age, gender, initial white blood count, absolute blast count and follow-up times were similar between the two CREB groups. Percent blasts at diagnosis (mean ± SD) were higher in the CREB overexpressing group (69 ± 28 vs 51 ± 24, p=0.02). However, neither stratification on blast percent nor on any other variable altered the significance of CREB as a predictor of initial adverse outcome. In contrast, overall survival tended to differ in the expected way among the cytogenetic prognostic groups (p=0.03), but without strong evidence for an association with CREB expression (p=0.35). Our results suggest that CREB expression may be an important independent predictor of remission duration and event-free survival in AML patients.