For a general understanding of the mechanisms of 1α,25(OH)2
signaling it is essential to monitor the genome-wide location of VDR in relation to primary 1α,25(OH)2
target genes. In this study, we investigated the sites of VDR chromatin occupancy in THP-1 human monocytes after 40
min treatment with 1α,25(OH)2
and in the unstimulated state using ChIP-seq, and related these to the VDR target genes identified using microarrays from the same cellular model 4
h after ligand treatment. As expected, GO term analysis of the 638 primary 1α,25(OH)2
target genes that we identified in THP-1 cells showed a clear enrichment for immune-related genes. In particular, the positive effect of 1α,25(OH)2
on IL1 production and secretion was highlighted, which fits well with previous reports (44
). However, we did not find IL1A
within our 638 1α,25(OH)2
target genes, but only the gene of the IL1 receptor antagonist (IL1RN
) downregulated. The latter suggests that the regulation of IL1 activity by 1α,25(OH)2
is indirect via the block of the IL1 receptor. In addition, Kaler et al.
) suggested that in THP-1 cells 1α,25(OH)2
downregulates the transcription factor STAT1, which is central in IL1 regulation.
Since the only nuclear target of 1α,25(OH)2
is VDR (47
) and the 4
h treatment is a reasonably short time, most of the genes that we found in our microarray analysis are assumed to be primary VDR target genes, i.e. VDR should bind a regulatory chromatin region that loops to the target gene’s TSS region. The majority (64%) of the 1α,25(OH)2
target genes were upregulated, which is typical for most VDR target tissues (48
Through the choice of early time points both in microarray and ChIP-seq experiments we aimed to detect mainly primary VDR effects. This is in contrast to the recent study of Ramagopalan et al.
), who reported VDR ChIP-seq data in HapMap lymphoblastoid cells after a long (36
treatment, focusing more on the association of VDR binding sites with disease SNPs and aspects of human evolution.
In this study, we identified 2340 genomic sites of VDR binding, 520 of which are exclusive to the unstimulated THP-1 cells. By size, our set of 1820 1α,25(OH)2
-treated genomic VDR binding sites in THP-1 cells is comparable with the 2776 VDR binding sites in 1α,25(OH)2
-treated lymphoblastoids (23
), but the overlap is only 18.2%. On the level of 1α,25(OH)2
target genes identified by microarray analysis the overlap between the two cellular models is even lower at 5.6%. Although the poor overlap may be partly explained by differences in cutoffs when considering high-confidence peaks and target genes as well as the rather different duration of 1α,25(OH)2
min or 4
h versus 36
h), it also suggests that VDR utilizes considerably variant binding sites in different cells. In silico
screening for VDR binding showed an approximately 100-fold higher number of putative REs than actually used (50
). The main regulator for the access to these sites is the organization of chromatin, which is very cell specific. Moreover, in cells of primary origin there is a significant inter-individual variation in chromatin organization and gene expression. Therefore, it is not surprising that monocytes and lymphoblasts have a rather small overlap of VDR binding sites. This is fully in line with the diversity of VDR target gene sets between different tissues (19–22
In the absence of ligand, VDR has been shown to actively repress its target genes, likely via a mechanism involving an interaction with co-repressor proteins (17
). Interestingly, only 14% of the 520 genomic VDR binding locations that occur uniquely in the absence of ligand contain a DR3-type VDRE. In contrast, up to 90% of the best 1α,25(OH)2
-dependent VDR binding regions contain at least one DR3-type RE. This means that upon ligand treatment the VDR occupancy is strongly shifted from non-DR3 locations to DR3-type RE locations. It is possible that the non-DR3 locations may serve as a nuclear store of VDR to be utilized rapidly upon the introduction of the ligand, partly substituting for the need to transport VDR into the nucleus from outside. This mechanism may also apply to retinoic acid and thyroid hormone receptors, but not for the classical steroid receptors, since they do not bind to DNA in the absence of ligand.
Interestingly, for 300 (73.5%) of the 408 upregulated 1α,25(OH)2
target genes we found VDR binding within 400
kb of their TSS, while this applied only for 104 (45.2%) of the 230 downregulated genes. This suggests that the mechanisms of downregulation of VDR target genes are rather complex and/or varied, and may require gene-specific investigations as demonstrated for the CYP27B1
). In this case the repressive function of VDR is known to result from indirect interaction with the chromatin, via transcription factor 3, also known as VDR interacting repressor (53
). An additional question is whether for 1α,25(OH)2
target genes there is a ligand-induced derepression mechanism as described for few other members of the nuclear receptor superfamily (54
). From the 408 1α,25(OH)2
target genes that we can explain via a VDR peak within 400
kb of their TSS only 11 (2.7%) meet the derepression criteria that they have a VDR peak in the unstimulated sample and no peak in the 1α,25(OH)2
-treated sample. Additional 32 genes (7.9%) can be called dominantly derepressed, since their main peak shows only in the unstimulated sample. This indicates that for some 10% of all 1α,25(OH)2
target genes a derepression mechanism may apply.
The fact that DR3-type REs are detected by our de novo
motif search as the most dominating sequence in the genomic VDR peak regions confirms earlier reports (9
) and the general belief in the field that DR3-type REs are the main type of VDREs. Nevertheless, only 31.7% of the identified VDR binding regions contain at least one DR3-type RE, i.e. the majority of genomic VDR binding sites do not use this type of VDRE. Since the most dominant ligand-induced VDR binding sites are highly enriched for DR3-type REs that, moreover, preferentially locate to the neighborhoods of upregulated 1α,25(OH)2
target genes, it is likely that DR3-type REs are primarily used in gene activation. However, this leaves the question about the nature of VDR binding at those 68.3% of locations that lack a DR3-type VDRE. Although we could not identify any non-DR3-type VDRE that would explain any significant proportion of the VDR binding, it does not exclude the possibility that a low number of such sites exist. Indeed, a case of repressive gene regulation via an ER9-type VDRE has been demonstrated by Polly et al.
). Other possible explanations include sites where VDR binds indirectly to genomic DNA via other transcription factors, such as SP1, although again our de novo
motif analysis did not identify specific enrichment of other transcription factor binding sites in VDR binding locations lacking a DR3-type VDRE compared to those that had it. A probably more likely explanation for both the VDR-peak-less target genes as well as the target gene-less VDR peaks comes from the recent demonstration that gene regulation by VDR is a very dynamic process with rapid changes of VDR binding site occupancy and, consequently, on the mRNA accumulation of the 1α,25(OH)2
target genes (13
). Therefore, the time points chosen in this study represent only snap-shots of the early actions of 1α,25(OH)2
and its receptor VDR. Due to the fast dynamics of VDR binding site occupancy and resulting fluctuations in target gene mRNA amounts, it is likely that without time-course data a considerable proportion of all transient VDR binding sites remains unknown.
Although most of the upregulated genes may be explained by a more unified mechanism of one or multiple VDREs, as already suggested for various VDR target genes, such as CDKN1A
) and CCNC
), we demonstrate here with three exemplary genomic regions that there is quite a variation in the gene regulatory scenarios of upregulated genes. Genome-wide there are only about 20 genes that show, as in the case of the SP100
gene, a single VDR location close to the target gene TSS. More common are situations where either one target gene has multiple VDR binding sites in its regulatory region or, as shown for ELL
, a pair of closely located VDR target genes share one or more VDR binding sites. The first constellation we have already demonstrated for the above mentioned VDR target genes, while the second scenario applies for the members of the IGFBP
gene family (15
). However, the most complex regulatory situations we found in the clusters around the VDR target genes THBD
, which each contains five or six VDR binding sites of different characteristics.
In conclusion, our genome-wide characterization of early-responding VDR binding sites close to primary 1α,25(OH)2D3 target genes identified a clear shift of VDR binding from non-DR3 sites to DR3-type REs upon ligand treatment. Furthermore, our data suggest regulatory explanations for a large majority of especially upregulated 1α,25(OH)2D3 target genes in THP-1 cells.