The recent advent of new technologies analyzing genome-wide chromatin changes has greatly improved our understanding of gene regulation and control. These global approaches provide a large-scale picture to the extent of changes and can be used for further dissection of gene-specific modifications. Our genome-wide mRNA expression profile and map of three major histone modifications associated with gene expression in monocytes, macrophages and DCs are an addition to previous studies in immune cells [23–27].
We found a high concordance rate between two active promoter marks, H3K4me3 and AcH3, in all three populations. Furthermore, the overall density of H3K4me3 and AcH3 marks on promoter regions displayed a similar pattern, with two elevated peaks surrounding transcription start site and a significant decline at the point of transcription, representing reduced levels of H3K4me3 and AcH3 in a region covered by RNA polymerase II and other regulatory proteins [28, 29]. As the respective peaks of H3K4me3 and AcH3 marks by and large coincide, it is likely that they also often co-occur in the same nucleosomes. Thus, our data confirm a tight association of H3K4me3 and AcH3 modifications with permissive chromatin for transcription in antigen-presenting cells. In contrast, H3K27me3 had an overall very low concordance with active marks and was equally distributed over the promoter regions with only slight decline at transcription start site.
The H3K4me3 mark was clearly a predominant modification on gene promoters, and most commonly was coupled with AcH3. This is in agreement with the earlier reported genome-wide studies from embryonic stem cells [30, 31], CD4+ and CD8+ T cells [24–26], all reporting H3K4me3 as the most characteristic modification in promoter regions. The H3K4me3 mark was also strongly correlated with the transcriptional activity of a gene. Interestingly, we noticed a gradual decrease of the H3K4me3 mark among active genes during the differentiation of monocytes to DCs, and accordingly, the number of genes without any modification increased. It should be noted that the global number of genes with a detectable expression signal between the three cell populations did not change significantly. This suggests that during differentiation, a subset of genes lose their H3K4me3 mark independently of their gene expression and also hints at a highly dynamic nature of histone modifications during differentiation. When studying the dynamics of histone modifications, we found that both the H3K4me3 and AcH3 marks are unstable when they occur alone and are prone to either lose the modification (i.e. become a promoter without mark) or to gain another permissive mark. In contrast, having the combination of both the H3K4me3 and AcH3 marks in a gene promoter was highly stable, and the majority of genes having this combination of histone marks in monocytes also had it in macrophages and DC. The overall decrease of the H3K4me3 mark and the reciprocal increase of genes with or without the H3K27me3 mark in the DC population suggests that the ex vivo differentiation in the presence of IL-4 induces a global shift towards less active chromatin.
At the single gene level we found many differences in histone modification profiles, in particular between monocytes and in vitro cultivated cells, which correlate with their transcriptional changes. As expected, the expression levels of many marker genes showed a good correlation with the presence of active histone modifications. When we analyzed the gain or loss of marks during differentiation, we found a remarkable decline in the H3K4me3 mark on several monocyte-specific marker genes and a corresponding increase in promoters of DC-specific genes. The prominent exceptions were the CD83 and CD40 genes, which were not associated with any histone mark. Similar to marker genes, the genes associated with antigen presentation activity and phagocytosis were expressed at high levels and had either the H3K4me3 or AcH3 mark. Notably, genes that are associated with phagocytosis had more permissive H3K4me3/AcH3 mark in their promoters in the monocyte population, whereas the genes associated with antigen presentation had more H3K4me3/AcH3 mark in macrophages and DCs. Although there was no dramatic difference between the cell populations in general, this correlation of histone modifications argues in favor of the role of histone modifications in these two key functions of antigen-presenting cells. In case of transcription factors, the situation was more complex, and in many cases the pattern of transcriptional activity of the gene did not match with expected histone modifications. However, the silencing H3K27me3 modification marked a vast majority of transcription factors involved in development or lineage decision in other cell types.
In our analysis, we identified several genomic clusters with changes in their chromatin status suggesting that histone modifications in genomic loci are coordinately regulated. In particular, we found that in macrophages and DCs, H3K4me3 levels increased in a large genomic cluster of proinflammatory and chemotactic CC chemokines on chromosomes 17q11.2 and 16q13 . Typically, CC chemokines attract mononuclear cells and/or are induced in an inflammatory environment, thus our findings suggest the requirement of histone modifications on chemokine gene clusters for their responses to inflammatory situations. These changes in H3K4me3 modifications were not limited to chemokine gene clusters as we found similar changes in other gene clusters involved in inflammatory processes (CD1, C1Q, SLAM, CD300 and IFITM families). Interestingly, we noted that even when the gene expression levels in macrophages and DCs differed, the histone marks in these two cell populations were often similar. This indicates that the histone modification status might be established before the actual expression of the gene and serve as a prerequisite for rapid activation by gene-specific transcription factors upon inflammatory stimuli. However, changes in chromatin modifications may also occur during subsequent activation steps. For example, studies in murine bone marrow derived macrophages showed that genes induced by TLR4 stimulation can be divided into two subclasses based on their gene-specific activation and histone H4 acetylation . Our work here provides a basis for the further studies to address the role of histone modifications in chemokine and other gene clusters in human macrophages and DCs after the activation with inflammatory stimuli.