Limits...
Distinct gene regulatory programs define the inhibitory effects of liver X receptors and PPARG on cancer cell proliferation.

Savic D, Ramaker RC, Roberts BS, Dean EC, Burwell TC, Meadows SK, Cooper SJ, Garabedian MJ, Gertz J, Myers RM - Genome Med (2016)

Bottom Line: PPARG generated a rapid and short-term response while maintaining a gene activator role.By contrast, LXR signaling was prolonged, with initial, predominantly activating functions that transitioned to repressive gene regulatory activities at late time points.Through the use of a multi-tiered strategy that integrated various genomic datasets, our data illustrate that distinct gene regulatory programs elicit common phenotypic effects, highlighting the complexity of the genome.

View Article: PubMed Central - PubMed

Affiliation: HudsonAlpha Institute for Biotechnology, Huntsville, AL, 35806, USA.

ABSTRACT

Background: The liver X receptors (LXRs, NR1H2 and NR1H3) and peroxisome proliferator-activated receptor gamma (PPARG, NR1C3) nuclear receptor transcription factors (TFs) are master regulators of energy homeostasis. Intriguingly, recent studies suggest that these metabolic regulators also impact tumor cell proliferation. However, a comprehensive temporal molecular characterization of the LXR and PPARG gene regulatory responses in tumor cells is still lacking.

Methods: To better define the underlying molecular processes governing the genetic control of cellular growth in response to extracellular metabolic signals, we performed a comprehensive, genome-wide characterization of the temporal regulatory cascades mediated by LXR and PPARG signaling in HT29 colorectal cancer cells. For this analysis, we applied a multi-tiered approach that incorporated cellular phenotypic assays, gene expression profiles, chromatin state dynamics, and nuclear receptor binding patterns.

Results: Our results illustrate that the activation of both nuclear receptors inhibited cell proliferation and further decreased glutathione levels, consistent with increased cellular oxidative stress. Despite a common metabolic reprogramming, the gene regulatory network programs initiated by these nuclear receptors were widely distinct. PPARG generated a rapid and short-term response while maintaining a gene activator role. By contrast, LXR signaling was prolonged, with initial, predominantly activating functions that transitioned to repressive gene regulatory activities at late time points.

Conclusions: Through the use of a multi-tiered strategy that integrated various genomic datasets, our data illustrate that distinct gene regulatory programs elicit common phenotypic effects, highlighting the complexity of the genome. These results further provide a detailed molecular map of metabolic reprogramming in cancer cells through LXR and PPARG activation. As ligand-inducible TFs, these nuclear receptors can potentially serve as attractive therapeutic targets for the treatment of various cancers.

No MeSH data available.


Related in: MedlinePlus

LXRs and PPARG maintain temporally distinct gene regulatory functions. a Cumulative distribution functions display the fraction of target gene promoters (y-axis) at different distance cutoffs to the nearest LXRA, LXRB, and PPARG binding event (x-axis). The nearest binding event to each gene promoter was used. Data for promoters of responsive up-regulated genes (green) and repressed (red) genes (adjusted p < 0.01, fold change cutoff of ±2) are shown. The background distribution using all gene promoters in the genome is also displayed (Bkgd, black). The top panel represents 2-h ChIP-seq data compared with 24-h RNA-seq data while the bottom panel compares 48-h ChIP-seq and RNA-seq data. For LXR data, GW3965 + T0901317 responsive genes are shown. Down-regulated curves are absent for 24-h GW3965 + T0901317 datasets as only one target gene was identified. b Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRA binding event (x-axis). LXRA binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRA sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRA sites coincident with RNAP2 (green). c Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRB binding event (x-axis). LXRB binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRB sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRB sites coincident with RNAP2 (green). d Read depth ratios (x-axis) of RNAP2 enrichment at LXRA binding sites (LXRA + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment. e Read depth ratios (x-axis) of RNAP2 enrichment at LXRB binding sites (LXRB + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4940857&req=5

Fig5: LXRs and PPARG maintain temporally distinct gene regulatory functions. a Cumulative distribution functions display the fraction of target gene promoters (y-axis) at different distance cutoffs to the nearest LXRA, LXRB, and PPARG binding event (x-axis). The nearest binding event to each gene promoter was used. Data for promoters of responsive up-regulated genes (green) and repressed (red) genes (adjusted p < 0.01, fold change cutoff of ±2) are shown. The background distribution using all gene promoters in the genome is also displayed (Bkgd, black). The top panel represents 2-h ChIP-seq data compared with 24-h RNA-seq data while the bottom panel compares 48-h ChIP-seq and RNA-seq data. For LXR data, GW3965 + T0901317 responsive genes are shown. Down-regulated curves are absent for 24-h GW3965 + T0901317 datasets as only one target gene was identified. b Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRA binding event (x-axis). LXRA binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRA sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRA sites coincident with RNAP2 (green). c Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRB binding event (x-axis). LXRB binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRB sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRB sites coincident with RNAP2 (green). d Read depth ratios (x-axis) of RNAP2 enrichment at LXRA binding sites (LXRA + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment. e Read depth ratios (x-axis) of RNAP2 enrichment at LXRB binding sites (LXRB + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment

Mentions: We integrated our drug treatment gene expression datasets with our nuclear receptor ChIP-seq results to investigate whether there was direct activation and repression through PPARG, LXRA, and LXRB. To identify putative direct target genes of the assayed TFs, we determined the distance of the nearest binding site to promoters of genes and subsequently generated cumulative distribution functions comparing the fraction of genes with binding events at different distance cutoffs (Fig. 5a). For assessing distance enrichments, we used all the genes in the genome to obtain a background distribution. We further split drug-responsive genes (adjusted p < 0.01, fold change cutoff ±2) into activated and repressed targets to assign putative regulatory functions. To infer early regulatory functions, we compared 24-h RNA-seq data with 2-h ChIP-seq data and used the 48-h datasets to assess late activities (Fig. 5a). Notably, both LXRA and LXRB generated concordant profiles that were distinct from those of PPARG. During the early drug response, LXRA and LXRB binding events are situated near up-regulated GW3965 + T0901317 target genes. However, during the late response, LXR occupancy occurs near repressed genes. This pattern is in contrast to the PPARG profile, where strong enrichments are observed only for up-regulated target genes. Again, the same LXR enrichment transition was observed when we used a lower fold change cutoff (±1.5), and we further ruled out the possibility that this enrichment was driven by a negative feedback loop on genes significantly activated at 24 h by removing these genes from the analysis (Additional file 1: Figures S13 and S14). We also obtained concordant enrichment transitions for LXRs using only GW3965 gene expression data (Additional file 1: Figures S15–S17).Fig. 5


Distinct gene regulatory programs define the inhibitory effects of liver X receptors and PPARG on cancer cell proliferation.

Savic D, Ramaker RC, Roberts BS, Dean EC, Burwell TC, Meadows SK, Cooper SJ, Garabedian MJ, Gertz J, Myers RM - Genome Med (2016)

LXRs and PPARG maintain temporally distinct gene regulatory functions. a Cumulative distribution functions display the fraction of target gene promoters (y-axis) at different distance cutoffs to the nearest LXRA, LXRB, and PPARG binding event (x-axis). The nearest binding event to each gene promoter was used. Data for promoters of responsive up-regulated genes (green) and repressed (red) genes (adjusted p < 0.01, fold change cutoff of ±2) are shown. The background distribution using all gene promoters in the genome is also displayed (Bkgd, black). The top panel represents 2-h ChIP-seq data compared with 24-h RNA-seq data while the bottom panel compares 48-h ChIP-seq and RNA-seq data. For LXR data, GW3965 + T0901317 responsive genes are shown. Down-regulated curves are absent for 24-h GW3965 + T0901317 datasets as only one target gene was identified. b Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRA binding event (x-axis). LXRA binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRA sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRA sites coincident with RNAP2 (green). c Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRB binding event (x-axis). LXRB binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRB sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRB sites coincident with RNAP2 (green). d Read depth ratios (x-axis) of RNAP2 enrichment at LXRA binding sites (LXRA + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment. e Read depth ratios (x-axis) of RNAP2 enrichment at LXRB binding sites (LXRB + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4940857&req=5

Fig5: LXRs and PPARG maintain temporally distinct gene regulatory functions. a Cumulative distribution functions display the fraction of target gene promoters (y-axis) at different distance cutoffs to the nearest LXRA, LXRB, and PPARG binding event (x-axis). The nearest binding event to each gene promoter was used. Data for promoters of responsive up-regulated genes (green) and repressed (red) genes (adjusted p < 0.01, fold change cutoff of ±2) are shown. The background distribution using all gene promoters in the genome is also displayed (Bkgd, black). The top panel represents 2-h ChIP-seq data compared with 24-h RNA-seq data while the bottom panel compares 48-h ChIP-seq and RNA-seq data. For LXR data, GW3965 + T0901317 responsive genes are shown. Down-regulated curves are absent for 24-h GW3965 + T0901317 datasets as only one target gene was identified. b Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRA binding event (x-axis). LXRA binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRA sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRA sites coincident with RNAP2 (green). c Cumulative distribution functions display the fraction of 48-h GW3965 + T0901317 repressed gene promoters (y-axis, adjusted p < 0.01, fold change cutoff of ±2) at different distance cutoffs to the nearest LXRB binding event (x-axis). LXRB binding events coincident with RNAP2 are in dark purple (+RNAP2), while LXRB sites devoid of RNAP2 occupancy are denoted in pink (-RNAP2). The background distribution using all gene promoters in the genome for sites overlapping with RNAP2 (+RNAP2 Bkgd, black) or for sites devoid of RNAP2 (-RNAP2 Bkgd, gray) is also graphed. The pie chart depicts the fraction of LXRB sites coincident with RNAP2 (green). d Read depth ratios (x-axis) of RNAP2 enrichment at LXRA binding sites (LXRA + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment. e Read depth ratios (x-axis) of RNAP2 enrichment at LXRB binding sites (LXRB + RNAP2). Negative values highlight stronger enrichment under control culture conditions (DMSO; 0 h) while positive values denote stronger enrichment after 48 h of GW3965 treatment
Mentions: We integrated our drug treatment gene expression datasets with our nuclear receptor ChIP-seq results to investigate whether there was direct activation and repression through PPARG, LXRA, and LXRB. To identify putative direct target genes of the assayed TFs, we determined the distance of the nearest binding site to promoters of genes and subsequently generated cumulative distribution functions comparing the fraction of genes with binding events at different distance cutoffs (Fig. 5a). For assessing distance enrichments, we used all the genes in the genome to obtain a background distribution. We further split drug-responsive genes (adjusted p < 0.01, fold change cutoff ±2) into activated and repressed targets to assign putative regulatory functions. To infer early regulatory functions, we compared 24-h RNA-seq data with 2-h ChIP-seq data and used the 48-h datasets to assess late activities (Fig. 5a). Notably, both LXRA and LXRB generated concordant profiles that were distinct from those of PPARG. During the early drug response, LXRA and LXRB binding events are situated near up-regulated GW3965 + T0901317 target genes. However, during the late response, LXR occupancy occurs near repressed genes. This pattern is in contrast to the PPARG profile, where strong enrichments are observed only for up-regulated target genes. Again, the same LXR enrichment transition was observed when we used a lower fold change cutoff (±1.5), and we further ruled out the possibility that this enrichment was driven by a negative feedback loop on genes significantly activated at 24 h by removing these genes from the analysis (Additional file 1: Figures S13 and S14). We also obtained concordant enrichment transitions for LXRs using only GW3965 gene expression data (Additional file 1: Figures S15–S17).Fig. 5

Bottom Line: PPARG generated a rapid and short-term response while maintaining a gene activator role.By contrast, LXR signaling was prolonged, with initial, predominantly activating functions that transitioned to repressive gene regulatory activities at late time points.Through the use of a multi-tiered strategy that integrated various genomic datasets, our data illustrate that distinct gene regulatory programs elicit common phenotypic effects, highlighting the complexity of the genome.

View Article: PubMed Central - PubMed

Affiliation: HudsonAlpha Institute for Biotechnology, Huntsville, AL, 35806, USA.

ABSTRACT

Background: The liver X receptors (LXRs, NR1H2 and NR1H3) and peroxisome proliferator-activated receptor gamma (PPARG, NR1C3) nuclear receptor transcription factors (TFs) are master regulators of energy homeostasis. Intriguingly, recent studies suggest that these metabolic regulators also impact tumor cell proliferation. However, a comprehensive temporal molecular characterization of the LXR and PPARG gene regulatory responses in tumor cells is still lacking.

Methods: To better define the underlying molecular processes governing the genetic control of cellular growth in response to extracellular metabolic signals, we performed a comprehensive, genome-wide characterization of the temporal regulatory cascades mediated by LXR and PPARG signaling in HT29 colorectal cancer cells. For this analysis, we applied a multi-tiered approach that incorporated cellular phenotypic assays, gene expression profiles, chromatin state dynamics, and nuclear receptor binding patterns.

Results: Our results illustrate that the activation of both nuclear receptors inhibited cell proliferation and further decreased glutathione levels, consistent with increased cellular oxidative stress. Despite a common metabolic reprogramming, the gene regulatory network programs initiated by these nuclear receptors were widely distinct. PPARG generated a rapid and short-term response while maintaining a gene activator role. By contrast, LXR signaling was prolonged, with initial, predominantly activating functions that transitioned to repressive gene regulatory activities at late time points.

Conclusions: Through the use of a multi-tiered strategy that integrated various genomic datasets, our data illustrate that distinct gene regulatory programs elicit common phenotypic effects, highlighting the complexity of the genome. These results further provide a detailed molecular map of metabolic reprogramming in cancer cells through LXR and PPARG activation. As ligand-inducible TFs, these nuclear receptors can potentially serve as attractive therapeutic targets for the treatment of various cancers.

No MeSH data available.


Related in: MedlinePlus