- Histone modification profiling in breast cancer cell lines highlights commonalities and differences among subtypes. - Results: To define alterations in epigenetic landscapes in breast cancers, we profiled the distributions of 8 key histone modifications by ChIP-Seq, as well as primary (GRO-seq) and steady state (RNA-Seq) transcriptomes, across 13 distinct cell lines that represent 5 molecular subtypes of breast cancer and immortalized human mammary epithelial cells.. - This approach identified AFAP1-AS1 as a triple negative breast cancer-specific gene associated with cell proliferation and epithelial-mesenchymal-transition. - Conclusions: Together, these datasets provide a comprehensive resource for histone modification profiles that define epigenetic landscapes and reveal key chromatin signatures in breast cancer cell line subtypes with potential to identify novel and actionable targets for treatment.. - Keywords: Breast cancer subtypes, Epigenetics, Histone modifications, Chromatin states. - The Lonestar Oncology Network for EpigeneticS Therapy And Research (LONESTAR) consor- tium was created to define epigenetic factors associated with molecular changes in specific subtypes of breast cancer. - (GRO-seq) and steady state (RNA-Seq) transcriptomes across 13 distinct cell lines that represent immortalized human mammary epithelial cells (hereafter referred to as normal immortalized or immortalized cells) and 5 mo- lecular subtypes of breast cancer (Fig. - These datasets provide a unique resource for breast cancer researchers, and reveal new insights into breast cancer biology. - Definition of common and unique features of chromatin landscapes of all 5 breast cancer subtypes Demonstration that histone modifications associated. - We systematically profiled histone modification patterns and gene expression programs in a set of well character- ized cell lines that represent 5 major breast cancer sub- types, including two ER positive subtypes, Luminal-A and Luminal-B, the HER2 positive subtype, and two triple- negative subtypes, TNBC-Claudin Low, and TNBC-Basal, as well as two normal immortalized breast cell lines as controls (Fig. - Chromatin states. - marks to define a 13-chromatin-state model from all 13 cell lines (Fig. - This model was selected by examin- ing the reproducibility of these 13 chromatin states on models trained on individual cell lines (Additional file 4:. - Together these 13 chro- matin states represent the combinatorial histone modifica- tion patterns and define the whole genome chromatin state landscapes across the breast cancer cell lines. - Com- bined with individual histone modification occupancy pro- files, these chromatin state landscapes provide an integrated view of key epigenetic marks across all breast cancer cell lines and subtypes, as illustrated for a region on chromosome 19 (Fig. - Therefore, most chromatin states identified by the model built from analysis of 5 core histone modifications (Fig. - Genes associated with different chromatin states show distinct transcriptional activities. - For example, signifi- cantly higher expression levels were observed for loci enriched for chromatin states associated with active pro- moters (PrAct) and gene bodies (TxAct), as expected (Additional file 4: Figure S2B). - The variability of each chromatin state was evaluated by their consistency at particular loci across all cell lines (Additional file 4:. - We next determined whether specific chromatin state patterns were unique to one or several breast cancer. - The two normal immortalized cell lines, 76NF2V and MCF10A, displayed significantly dis- tinct chromatin state distributions from all the breast cancer cell lines. - The clustering structure further re- vealed hierarchical relationships among breast cancer subtypes. - These subtype spe- cific chromatin states were further confirmed by exam- ining individual histone modification signals in different genomic regions, including promoters, enhancers and gene bodies (Additional file 6: Figure S4), and these data again showed clear subtype specificity in terms of characteristic histone modification patterns, with good reproducibility between replicates.. - Active transcription-associated states, including both TxAct and TxFlk, made up the majority (47%) of chromatin states that were common between cell lines within the same subtype. - The dynamic nature of enhancer associated chromatin states in breast cancer subtypes are further characterized by the enhancer RNA (eRNA) signals de- fined in the GRO-seq profiles (Additional file 7: Table S5), as will be presented in a separate paper from the LONESTAR consortium.. - 2c) to more quantitatively characterize differential chromatin states associated with subtype specific patterns. - We analyzed specific active and repressive transcription chromatin states for individual subtypes as well between subtype groups. - Choline has been identified as a breast cancer biomarker, and a recent study suggested distinct choline regulation in xenograft models of different breast cancer subtypes [6, 7]. - Consistent with this finding, we observed several AR pathway regulators are expressed at significantly lower levels in TCGA basal subtype patients compared with other subtype breast cancer patients, including AR,. - a Summary of cell lines and histone marks. - b Thirteen chromatin states were defined using 5 key histone modifications, the left panel describes the chromatin state annotations and color scheme, the central panel describes the emission coefficients in ChromHMM model, the right panel describes the relative enrichment of coverage in whole genome and in different genomic regions. - c Integrated view of whole genome chromatin state landscapes in breast cancer cells and corresponding individual histone modification profiles in MCF7 cells for a region on chromosome 19. - Repressive chromatin states. - 2 Chromatin states signatures for breast cancer subtypes. - b Pie chart of major chromatin states categories in subtype specific patterns across all cell lines. - Similar to our analyses of active transcription states, we identified breast cancer subtype specific patterns for repressive chromatin states (Additional file 14: Table S6). - Distinct H3K27me3 signal patterns were observed, revealing subtype specificity similar to that in active chromatin states (Fig. - The non-malignant, immortal cells had the greatest number of enriched/depleted repressive chromatin states relative to the breast cancer cell lines, which suggests that repressive states might serve as a pan breast cancer signature.. - Further functional analysis on associated gene sets did not reveal significantly enriched regulatory pathways, in contrast to the results for the active chromatin states.. - However, we observed one striking enrichment in the cluster of NOD-like signaling receptor genes, which display significantly increased H3K27me3 occupancy in all breast cancer cell lines but not in the normal immortal- ized cells (Additional file 15: Figure S7). - Our results are consistent with decreased expression of the NOD-like signaling receptor, NLRP3, in all subtypes of breast cancer relative to normal controls in the TCGA datasets (Additional file 13: Figure S6). - The enrichment of repressive chromatin states on NOD-like gene family members in all breast cancer cells illustrates the regulatory potential for epigenetic silencing of cancer suppressive genes and also provides a potential marker for a pan breast cancer signature.. - Close inspection of these chromatin states signatures also identified individual genes that displayed highly specific patterns across the breast cancer subtypes. - Exclusive expression of AFAP-AS1 is in triple negative breast cancer cells was further confirmed by RT-qPCR (Fig. - regulatory target for AFAP1-AS1 [9]. - Limited knock down of AFAP1-AS1. - The functions of AFAP1-AS1 are not yet clear, but the identification AFAP1-ASI as a TNBC specific gene through analysis of subtype specific chromatin states illustrates the power of our approach in identifying novel molecular tar- gets for future development of TNBC therapies.. - ZNF597 is further validated to have lower expression in TCGA basal patients compared with other breast cancer subtype patients (Additional file 13: Figure S6). - These results highlight the power of chromatin states signatures to not only predict tran- scriptional status but also to infer other potential epigenetic patterns, such as DNA methylation states.. - The LONESTAR consortium data provide a comprehensive resource for histone modification profiles and transcription states across a novel collection of breast cancer cell lines that represent the 5 molecular subtypes of breast cancer.. - a Screen shot of TNBC specific chromatin states and H3K4me3 signals on AFAP1-AS1 promoter and genebody. - TNBC specific AFAP1-AS1 expression measured by (b) RNA-seq and (c) q-PCR. - Depletion of AFAP1-AS1 in MB231 (d) and HCC1937 cells (e) inhibits cell proliferation (f, g), and colony formation (h, i) respectively. - landscapes across human breast cancer cell lines and identi- fied subtype specific epigenetic signatures for major breast cancer subtypes. - These epigenetic signatures revealed functional gene sets for each of the five breast cancer subtypes. - Our results are consistent with previous breast cancer profiling studies, but also provide unique insights, such as discovery of AFAP1-AS1 (Fig. - The chromatin state landscapes defined here in breast cancer cells also demonstrate the complexity of interactions between covalent histone modifications.. - Ultimately, these data may provide new clues to the etiology of the different breast cancer subtypes and new avenues for therapy development. - We hope these data resources as well as our analyses will be broadly used in the breast cancer research community for mechanistic studies, biomarker discovery and precision therapy.. - When cells reached 60% confluence, two siRNAs targeting AFAP1-AS1 (ThermoFisher Scientific, Cat. - MDA-MB-231, HCC1937 cells were transfected with control siRNA, or siRNA targeting AFAP1-AS1. - Chromatin state model. - We also trained models using all 8 histone marks using same proced- ure, and the extended model (Additional file 5: Figure S3) identified 15 chromatin states that largely overlap with the chromatin states defined by the 5 core histone modification model, with additional states representing the 5′ or 3′ end of active transcription units or broad flanking regions of ac- tive promoters. - We followed the method used by NIH roadmap epigenetics consortium to evaluate the robustness of the 13 chromatin-states model jointly-trained on all cell lines by comparing it with models independently trained on individual cell lines [4]. - We first identify all genomic regions that have same within- subtype chromatin states and different between-subtype chromatin states. - This is done by detecting have same chro- matin states in cell lines of the same subtype, and filter out the regions that have same chromatin states in all subtypes.. - The subtype specific chromatin state signature is defined by the chromatin states that are uniquely present or absent in the subtype. - The pan breast cancer signature is defined by comparing normal like cells with breast cancer cells, and the triple negative signature is defined as comparing TNBC-. - Additional file 1: Figure S1. - Additional file 2: Table S1. - Additional file 3: Table S2. - Additional file 4: Figure S2. - (A) Clustering of chromatin states model learned on individual cells showing same enrichment pattern that can recover the chromatin state jointly learned using all 13 cells. - (B) RNA-seq expression levels for genes associated with different chromatin states. - Larger area under curve indicates more variability across breast cancer cells. - Additional file 5: Figure S3. - Additional file 6: Figure S4. - Additional file 7: Table S5. - Additional file 8: Figure S5. - (PDF 489 kb) Additional file 9: Table S4. - Additional file 10: Table S7. - (XLSX 23 kb) Additional file 11: Table S8. - Additional file 13: Figure S6. - Subtype expression patterns of TCGA breast cancer samples. - Additional file 14: Table S6. - Additional file 15: Figure S7. - Additional file 16: Figure S8. - Additional file 17: Table S10. - Additional file 18: Table S3. - Emission probability and transition probability of Chromatin States defined by ChromHMM. - AFAP1-AS1: Actin filament associated protein antisense RNA 1. - TNBC: Triple-negative breast cancer;. - provided critical cell culture support and expertise on breast cancer cells. - Interplay of choline metabolites and genes in patient- derived breast cancer xenografts. - Breast Cancer Res. - High expression of AFAP1-AS1 is associated with poor survival and short-term recurrence in pancreatic ductal adenocarcinoma. - The up-regulation of long non-coding RNA AFAP1-AS1 is associated with the poor prognosis of NSCLC patients. - Upregulated long non-coding RNA AFAP1-AS1 expression is associated with progression and poor prognosis of nasopharyngeal carcinoma. - Long noncoding RNA AFAP1-AS1 indicates a poor prognosis of hepatocellular carcinoma and promotes cell proliferation and invasion via upregulation of the RhoA/Rac2 signaling. - Critical role for the long non-coding RNA AFAP1-AS1 in the proliferation and metastasis of hepatocellular carcinoma
Xem thử không khả dụng, vui lòng xem tại trang nguồn hoặc xem
Tóm tắt