Abstract 4802: High-resolution analysis of chromatin association to nuclear lamina in normal and cancer cells

Abstract

Listen

Translate article

Abstract Chromatin organization depends on attachment to nuclear lamina. In coordination with histone modifications and insulators, the nuclear lamina defines active and repressed chromatin domains, thus it is an important regulator of gene expression. Also, it has been shown that both histone methylase and deacetylase enzymes, as well methyl-binding domain proteins, bind directly to lamina proteins, indicating a role of the nuclear lamina in epigenetic marking. Whole-genome mapping of lamina-associated domains (LADs) was recently done by DamID (Guelen et al., Nature 453, 2008). This method, although efficient, is not suitable to study primary, uncultured cells. For this reason, we developed a chromatin immunoprecipitation (ChIP) protocol to capture chromatin bound to the nuclear lamina. We performed lamin B1 ChIP in IMR-90 and PC-3 cells, which are cultured cells from normal lung and prostate cancer, respectively. In order to validate our protocol, we tested IMR-90 ChIPed DNA for binding to 12 genomic regions (5 identified as LADs in normal lung fibroblast cells in the original report by Guelen et al., and 7 regions without evidence of association to nuclear lamina). All but one LAD could be validated, thus showing that ChIP can be used to study binding of chromatin to nuclear lamina. Next, massively parallel sequencing of ChIPed DNA from PC-3 was performed using Illumina Solexa technology, as well ChIPed DNA from this same cell line using antibodies to histone H3 marks K9me2 and K27me3. Promoter DNA methylation in PC-3 was determined by MCAM, a reduced-representation methylation microarray technique. For each one of the tested modifications by ChIP-Seq, we obtained between 8 and 10 million uniquely mapped tags, and identification of genome target regions was done using SICER software (Zang et al., Bioinformatics 25, 2009), in order to identify spatial clusters of large domains. In PC-3, LADs are found in higher number than in normal lung (2597 LADs, compared to 1344 in normal lung fibroblasts), however they are significantly smaller (30 kb in average, compared to 1 Mb in normal lung fibroblasts). Thus, in PC-3, LADs cover 2.7% of the human genome, and 1.9% of all RefSeq genes are inside LADs. These genes, as expected, have lower mRNA expression than genes outside LADs, as measured by mRNA microarrays. Looking only at gene promoter marking, LADs overlap little with DNA methylation (7.5%) or H3K9me2 (16%), and extensively co-localize with H3K27me3 (57%). Still, nearly 40% of the LADs overlap with neither DNA methylation, histone H3 K9me2 nor K27me3, but still show significant transcriptional repression. In conclusion, ChIP-Seq is a valid method for whole genome, high resolution study of chromatin association to nuclear lamina, and it is compatible with primary, uncultured cells. LADs in cancer are smaller in size than those found in normal tissues, and it is possible that association to the lamina can be a stand-alone repression mark. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4802.