Histone deacetylases (HDACs) are a family of enzymes that remove acetyl groups from histones and non-histone proteins, thereby modifying the structure of chromatin and regulating gene expression. HDACs have frequently been found to be upregulated in solid tumors and in hematological malignancies, leading to clinical trials of HDAC inhibitors for the treatment of various cancers. Oncogenic effects of HDACs are thought to be mediated by the epigenetic silencing of tumor suppressors. However, understanding of the specific gene targets of individual HDACs is lacking. In order to elucidate the function of HDAC1 in a cancer model, we have performed experiments comparing T cell leukemias in hdac1 haploinsufficient zebrafish with leukemias in their wild type siblings.

For our initial studies, we tested whether loss of one allele of the hdac1 gene (haploinsufficiency) would reduce the oncogenic effects of c-Myc, a potent oncogene, and protect against leukemia. We generated T cell lymphoblastic leukemia (T-ALL) in both hdac1+/- haploinsufficient and wild type zebrafish siblings by overexpressing murine c-Myc in T cells using a rag2 promoter. We then monitored the hdac1+/- and wild type fish for tumor incidence, latency, growth and overall survival.

The tumor incidence rates and mean latencies were not significantly different for hdac1+/- and wild type tumors. However, we found significant differences in primary tumor growth between these genotypes. Tumor progression was significantly faster for hdac1+/- fish (p=0.001), with mean time to stage 3 tumor (>50% of the animal showing evidence of tumor dissemination) of 43.3 ± 10.1 days (mean ± S.D.) compared with 76.3 ± 9.2 days for wild type siblings. Furthermore, survival of hdac1+/- fish was significantly shorter (p<0.01) at 70.3 ± 36.2 days compared with wild type siblings at 129.2 ± 57.8 days.

In contrast, progression of transplanted hdac1+/- tumors was slower compared with transplanted wild type tumors. Only 8.3% of transplanted hdac1+/- tumors had extended into the thymus and other organs by 21 days following intra-peritoneal injection into recipients. In contrast, 92% of the wild type tumors had evidence of widespread tumor dissemination by 21 days post-transplant.

Several mechanisms were considered as potential contributing factors to these differences between the hdac +/- and wild type leukemias, including cell cycle differences, the effects of different tumor micro-environments, and differences in leukemia-initiating cell (LIC) frequencies. We did not find obvious differences in cell cycle parameters between genotypes. Also, we found no effect from different microenvironments of tumors following transplantation into either hdac1 +/- or wild type recipients. In contrast, we did find significant differences in leukemia-initiating cell frequencies between the genotypes (p<0.0001). The LIC frequency for hdac1 +/- tumors was measured at 1 in 410 compared with a higher LIC frequency of 1 in 30 for wild type tumors.

In summary, our data provide support for dual roles of hdac1 in leukemogenesis. In this model, hdac1 has an initial protective role acting as an oncosuppressor, leading to more rapid tumor progression and decreased survival of animals with hdac1 +/- haploinsufficient tumors compared with wild type tumors. Interestingly, hdac1 also behaved as an oncogene, with decreased tumor progression following transplantation of hdac1 haploinsufficient tumors compared with wild type tumors. Our data indicate that caution is warranted regarding the use of HDAC inhibitors in treating hematologic malignancy, as this treatment could have unintended consequences, particularly during early stages of tumor development. Explaining these apparently contradictory roles of hdac1 in leukemogenesis is the focus of ongoing investigation, including gene expression studies and validation of targets.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.