Selective regulation of biological processes by vitamin D based on the spatio-temporal cistrome of its receptor

Abstract

The transcription factor vitamin D receptor (VDR) is the exclusive nuclear target of the biologically active form of vitamin D (1,25(OH)D). In THP-1 human monocytes we obtained a highly accurate VDR cistrome after 2 and 24h ligand stimulation comprising >11,600 genomic loci, 78% of which were detected exclusively after 24h. In contrast, a group of 510 persistent VDR sites occurred at all conditions and some 2100 VDR loci were only transiently occupied. Machine learning and statistical analysis as well as a comparison with the re-analyzed B cell VDR cistrome indicated a subgroup of 339 highly conserved persistent VDR sites that were suited best for describing vitamin D-triggered gene regulatory scenarios. The 1,25(OH)D-dependent transcriptome of THP-1 cells comprised 587 genes, 311 of which were primary targets with main functions in the immune system. More than 97% of the latter genes were located within 1,25(OH)D-modulated topologically associated domains (TADs). The number of persistent and transient VDR sites was found to be the main discriminator for sorting these TADs into five classes carrying vitamin D target genes involved in distinct biological processes. In conclusion, specific regulation of biological processes by vitamin D depends on differences in time-dependent VDR binding.

Key Vitamin D Target Genes with Functions in the Immune System

“…Inspection of vitamin D target gene lists: three different RNA-seq datasets were used [29][30][31], which had been analyzed by the algorithm DESeq2 [32] for differentially expressed genes. Moreover, the THP-1 dataset had been filtered by the machine learning method, self-organizing map (SOM) [28] or for the 100 genes showing either the highest basal activity, inducibility (fold change (FC)) or significance (lowest p-value) [33]. The in vitro treated PBMCs [31] were used either unfiltered or filtered for the top 100 genes with regard to basal activity, FC or p-value.…”

“…Epigenomic characterization of enhancers and TSS regions: published epigenome-wide data for THP-1 cells were used, which had been treated for 24 h with 10 nM 1,25(OH) 2 D 3 or solvent (0.1% EtOH): ChIP-seq data of VDR [28], H3K27ac [35] and H3K4me3 [35] as well as FAIRE-seq data [29]. Using the IGV-browser [36] all VDR-bound enhancers 1 Mb up-and downstream of the genes' TSSs as well as the TSS regions themselves were first visually inspected for VDR binding, markers of active chromatin and accessible chromatin.…”

“…Using the IGV-browser [36] all VDR-bound enhancers 1 Mb up-and downstream of the genes' TSSs as well as the TSS regions themselves were first visually inspected for VDR binding, markers of active chromatin and accessible chromatin. Then, VDR classification [28] and statistically significant (p < 0.05) responsiveness to 1,25(OH) 2 D 3 treatment of enhancers and TSS regions [28,29,35] was checked based on data provided by the respective publications. When the criteria for a super-enhancer were not fulfilled, i.e., three VDR binding sites within 20 kb, one of which needs to be a strong persistent or transient site and a continuous H3K27ac mark throughout its whole enhancer [33], the closest single enhancer to the TSS carrying a strong persistent or transient VDR site was selected.…”

“…These were from THP-1 cells, which had been treated in three biological repeats for 24 h with 10 nM 1,25(OH) 2 D 3 [29]; PBMCs of 12 healthy individuals, which were stimulated in vitro in a single repeat for 24 h with 10 nM 1,25(OH) 2 D 3 [31]; and PBMCs of five healthy individuals, which had been supplemented once for 24 h with a vitamin D 3 bolus of 2000 µg [30]. The THP-1 dataset was filtered by the machine learning method self-organizing map (SOM) [28] or for the top 100 genes with respect to basal activity, FC or p-value [33], which provided 587 and 264 target genes, respectively ( Figure 1). Similarly, in vitro treated, unfiltered PBMCs showed 933 target genes or 248 genes after filtering for the top 100 genes with respect to basal activity, FC or p-value.…”

“…A VDR ChIP-seq meta-analysis indicated that there are on average more than 10,000 VDR binding loci per cell type [27]. However, only a few hundred of these sites are persistently occupied by VDR and some 2000 are bound transiently, while the majority of them are only found at later points in time [28]. Based on the chromatin model of vitamin D signaling [4], the looping of these VDR-bound enhancers to TSS regions modulates the activity of hundreds of vitamin D target genes in VDR expressing tissues.…”