Abstract
Inhibitors of histone deacetylases (HDACIs) like valproic acid (VPA) display activity in murine leukemia models, and induce tumor-selective cytoxicity against blasts from patients with acute myeloid leukemia (AML). However, despite of the existing knowledge of the potential function of HDACIs, there remain many unsolved questions especially regarding the factors that determine whether a cancer cell undergoes cell cycle arrest, differentiation, or death in response to HDACIs. Furthermore, there is still limited data on HDACIs effects in vivo, as well as HDACIs function in combination with standard induction chemotherapy, as most studies evaluated HDACIs as single agent in vitro. Thus, our first goal was to determine a VPA response signature in different myeloid leukemia cell lines in vitro, followed by an in vivo analysis of VPA effects in blasts from adult de novo AML patients entered within two randomized multicenter treatment trials of the German-Austrian AML Study Group. To define an VPA in vitro “response signature” we profiled gene expression in myeloid leukemia cell lines (HL-60, NB-4, HEL-1, CMK and K-562) following 48 hours of VPA treatment by using DNA Microarray technology. In accordance with previous studies in vitro VPA treatment of myeloid cell lines induced the expression of the cyclin-dependent kinase inhibitors CDKN1A and CDKN2D coding for p21 and p19, respectively. Supervised analyses revealed many genes known to be associated with a G1 arrest. In all cell lines except for CMK we examined an up-regulation of TNFSF10 coding for TRAIL, as well as differential regulation of other genes involved in apoptosis. Furthermore, gene set enrichment analyses showed a significant down-regulation of genes involved in DNA metabolism and DNA repair. Next, we evaluated the VPA effects on gene expression in AML samples collected within the AMLSG 07-04 trial for younger (age<60yrs) and within the AMLSG 06-04 trial for older adults (age>60yrs), in which patients are randomized to receive standard induction chemotherapy (idarubicine, cytarabine, and etoposide = ICE) with or without concomitant VPA. We profiled gene expression in diagnostic AML blasts and following 48 hours of treatment with ICE or ICE/VPA. First results from our ongoing analysis of in vivo VPA treated samples are in accordance with our cell line experiments as e.g. we also see an induction of CDKN1A expression. However, the picture observed is less homogenous as concomitant administration of ICE, as well as other factors, like e.g. VPA serum levels, might substantially influence the in vivo VPA response. Nevertheless, our data are likely to provide new insights into the VPA effect in vivo, and this study may proof to be useful to predict AML patients likely to benefit from VPA treatment. To achieve this goal, we are currently analyzing additional samples, and we are planning to correlate gene expression findings with histone acetylation status, VPA serum levels, cytogenetic, and molecular genetic data.
