1. To identify key differentiation genes and characterize their molecular mechanisms of actionA. Identification, Cloning and Characterization of CASZ1, the human homolog of the drosophila neural fate determination gene castor (dCas). We previously identified that there are 2 predominant isoforms of CASZ1;hCas11 (now referred to as CASZ1a), which has 11 zinc fingers, and hCas5 (CASZ1b) which has 5 zinc fingers. Total CASZ1 mRNA levels dramatically increased upon induction of differentiation in neuron and muscle cell models. The study of CASZ1 is warranted because it;1) maps to chromosome 1p36.22 a region lost in almost 98% of NB tumors with 1p36 LOH;2) was a highly evolutionarily conserved neural fate determination gene;3) has not been functionally studied in mammalian model systems;4) is regulated during NB cell differentiation and 5) is expressed at high levels in primary tumors of NB patients with good overall survival (p=0.0009). The classic involvement of a suppressor gene in tumorigenesis requires biallelic gene inactivation. For CHD5 the NB SRD 1p36 tumor suppressor, 1 allele is lost via 1pLOH while the remaining allele is silenced via DNA methylation. Previously we had postulated a role for MYCN in suppressing CASZ1 expression, since MYCN bound to an E-box in the CASZ1 promoter and MYCN over-expression inhibited CASZ1 expression in single-copy MYCN expressing NB cells. However, silencing MYCN alone did not relieve suppression of CASZ1 expression, suggesting the existence of additional suppressive mechanisms. The finding that in both 1pLOH and intact1p NB cells, CASZ1 expression is induced after HDACi treatment suggests chromatin modifications play a role in suppressing CASZ1 expression. A bioinformatic analysis of the CASZ1 locus indicated that in Embryonic stem cells (ES) the CASZ1 transcriptional start site (TSS) contains a bivalent chromatin mark- an activation mark characterized by acetylation of histone 3 at lysine 4 (H3K4me3) as well a suppressive mark characterized by methylation of histone 3 at lysine 27(H3K27me3). Bivalent marks in ES cells are typically found at genes poised to be dynamically regulated, such as during develoment. Since the enzymatically active EZH2 component of the polycomb repressor complex (PRC2) mediates the H3K27me3 silencing mark, we performed a series of experiments that directly demonstrated that CASZ1 is a target of PRC2 complex mediated suppression. ChIP-PCR analyses indicated that PRC2 complex proteins and an H3K27me3 suppression mark were enriched over the CASZ1 TSS in NB cells under steady-state conditions. After treatment with an HDACi (Depsipeptide), binding of these proteins to CASZ1 TSS decreased, the CASZ1 TSS lost the chromatin silencing H3K27me3 mark and was enriched in a chromatin activation mark H3K4me3. Pharmacologic and genetic inhibition of EZH2 in NB tumor cells increased CASZ1 expression. Moreover, now we have found that the PRC2 complex proteins suppress not only CASZ1, but repress CLU(clusterin), NGFR, p73, Runx3 and TrkA(NTRK1) expression. This identifies PRC2 as mediating suppression of a number of genes with tumor suppressor activity in NB. A perplexing aspect of the biology of NB tumors has been their genetic heterogeneity and their biologic plasticity. This study identifies EZH2 as a potential regulatory mechanism that may link these two aspects and is therapeutically tractable. Despite prior studies implicating CASZ1 in neural (Drosophila) and heart (Xenopus) development, the key structural domains mediating its transcriptional programs have not been elucidated in any species. Since the CASZ1b isoform is the most evolutionarily conserved isoform, we utilized site-directed mutagenesis and deletion mapping to identify key domains in CASZ1b required to 1) activate transcription of a tyrosine hydroxylase promoter-luciferase (TH-Luc) reporter construct and 2) regulate endogenous transcription of several CASZ1 neural (TH, TrkA, NGFR) or muscle differentiation-associated (MYO7A) target genes. Mutation of the C2H2 motif-defining cysteine to an alanine in any one of ZF1, 2, 3, or 4 caused a 60-80% loss in CASZ1b transactivation activity. Mutation of ZF5 did not. Next, a series of N-terminal truncations revealed a critical activation domain at AA31-185, which despite retaining nuclear localization completely lost transactivation function. Deletion of other predicted domains as well as the region C-terminus up to ZF4 (38% of the protein) had no appreciable effect on CASZ1b activity. Additional deletions of a proline-rich region, ATP binding site and a Peroxisome targeting sequence had modest but variable effects on CASZ1b transcriptional activity. Finally, we engineered the transcriptionally inactive mutant (ZF4deletedCASZ1b) into NB cells and found that it has a 50% decrease in its ability to inhibit NB cell growth compared to the CASZ1WT. We also found that mutations in CASZ1b blocking its transcriptional activity also resulted a a decreased ability of CASZ1b to suppress NB tumorigenicity. These data defined important domains mediating CASZ1b's transcriptional activity and demonstrated that the transcriptome regulated by CASZ1 is important in mediating its growth and tumor suppressor properties. By identifying and analyzing the critical domains within this gene, we are beginning to unravel the molecular mechanisms of regulation of the CASZ1 differentiation program. Not only does this move us one step closer to understanding the potential impact of this gene on neuroblastoma pathogenesis, it provides the first molecular analysis of a gene that has already been shown to be important in neural and cardiomycte differentiation programs.
Showing the most recent 10 out of 22 publications
