Abstract
Childhood acute lymphoblastic leukemia (ALL) is the most common childhood cancer and, despite a cure rate approaching 90%, relapse is a significant cause of death in young people. Recently it has been shown that inactivating mutations in the histone acetyltransferase, CREB binding protein (CREBBP or CBP) are frequently seen at relapse in childhood ALL, with enrichment in high hyperdiploid and hypodiploid cases. Mutations are usually heterozygous, suggesting haploinsufficiency, and are often acquired at relapse, implying a role in drug resistance. Since glucocorticoid (GC) response genes are known targets of CREBBP and, given the pivotal role of GCs in ALL therapy, it has been postulated that CREBBP mutations confer GC chemoresistance.
CREBBP is a multifunctional protein, playing a role in cAMP dependent signalling, acetylation mediated activation of p53 and inactivation of BCL6 and a range of DNA damage repair pathways including base excision repair (BER) and direct DNA damage repair. To assess the role of CREBBP haploinsufficiency in ALL, RNAi techniques were used to create isogenic CREBBP knockdown models of ALL. CREBBP knockdown was carried out using small hairpin RNA (shRNA) transduction (termed shCBP cells) or small interfering RNA (siRNA) transfection (termed siCBP cells) in the PreB 697 B-cell precursor cell line (t(1;19)) and the hypodiploid MHH-CALL-2 cell line, as well as high hyperdiploid primagraft ALL cells. Knockdown of at least 50% of control was confirmed at both mRNA and protein level. The functional impact of CREBBP knockdown in cells was determined by analysis of known CREBBP target residues; acetyl H3K18 and H3K27, and transcription of cAMP dependent genes (CXCR4, MKNK2, DUSP5, DUSP10 and RGS16). To assess the impact of CREBBP knockdown on response to GCs, cells were treated with dexamethasone and expression of the classic glucocorticoid receptor (GR) targets; GILZ and FKBP51, was assessed by quantitative reverse transcriptase PCR (QRT-PCR). Alamar blue cell viability assays were used to determine the sensitivity of each CREBBPknockdown model to dexamethasone compared to isogenic controls.
Three out of four cell models displayed a reduction in H3K18 or H3K27 acetylation compared to isogenic control, indicating a relevant functional impact of CREBBP knockdown. Cell lines showed a trend towards reduced induction of some of the selected cAMP dependent targets but statistical significance was not achieved (p values >0.2). Gene expression profiling and Ingenuity Pathway Analysis of PreB 697 shCBP cells compared to isogenic control predicted that upstream transcription of NR3C1, the gene encoding the GR, would be affected in CREBBP knockdown cells. However, while induction of GILZ and FKBP51 in PreB 697 shCBP cells in response to GC was significantly impaired in knockdown compared to control cells (GILZ p=0.009, FKBP51 p=0.03), they were no more resistant to dexamethasone (p=0.9). This was mirrored in siCBP cell lines and primagraft cells, where a significant impairment in basal expression of GILZ and/or FKBP51 was seen in some lines (GILZ reduction; p=0.03 PreB 697 shCBP, p=0.02 PreB 697 siCBP, FKBP51 reduction; p=0.01 primagraft siCBP cells) but no significant impairment in the transcriptional induction of these genes in response to GC compared to isogenic control was observed (p values >0.5). Importantly, no decreased sensitivity to dexamethasone was seen in any model after CREBBP knockdown (p values >0.1).
CREBBP knockdown in ALL cells had no significant effect on the induction of cAMP dependent genes, had a variable effect on GR target expression, but consistently showed no impact on GC sensitivity, regardless of cytogenetic context. These data show that the acquisition of CREBBP mutations at relapse in childhood ALL is not mediated through GC resistance and suggest that other CREBBP associated mechanisms, such as DNA damage repair, may influence drug response. Understanding the role of CREBBP in carcinogenesis and drug resistance is crucial as it is implicated as a tumour suppressor in a growing number of cancers, making it a potential multi-tumour target for novel therapies.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.