UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)
Macrophage regulatory mechanisms in atherosclerosis
The interplay of lipids and inflammation
Neele, A.E.
Publication date
2018
Document Version
Other version
License
Other
Link to publication
Citation for published version (APA):
Neele, A. E. (2018). Macrophage regulatory mechanisms in atherosclerosis: The interplay of
lipids and inflammation.
General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).
Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 143PDF page: 143PDF page: 143PDF page: 143
Chapter 8
Fine tuning of DNA methylation during the
differentiation and activation of human
macrophages
Koen F. Dekkers*, Annette E. Neele*, P. Eline Slagboom, J. Wouter Jukema,
Bastiaan T. Heijmans#, Menno P.J. de Winther#
*, # Contributed equally
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 144PDF page: 144PDF page: 144PDF page: 144
144
Abstract
Introduction: Macrophages and their precursors monocytes play a key role in
inflammation and chronic inflammatory disorders like atherosclerosis. Monocyte-to-macrophage differentiation and activation is accompanied by significant epigenetic remodeling of which DNA methylation is one of the primary components. Here, we studied fine tuning of DNA methylation during differentiation and activation of macrophages, and identified transcription factors linked to differentially methylated single CpGs.
Material and methods: Human monocytes were isolated and differentiated into
macrophages in vitro and subsequently activated with LPS/IFNγ, IL-4 or modified lipids, to obtain different activation states. We assessed DNA methylation in these different subsets using the Illumina 450k array and performed an epigenome-wide analysis.
Results: We found that DNA methylation changes occur predominantly during
monocyte-to-macrophage differentiation and substantially less upon subsequent activation. We observed both gain and loss of methylation during differentiation. Differential DNA methylation was particularly found at single CpGs or small regions, which were enriched for enhancers. Furthermore, differentially methylated CpGs were located at binding sites for transcription factors known to be involved in monocyte-to-macrophage differentiation and were distinct for gain (C/EBP and ETS) and loss (AP-1) of methylation.
Conclusion: Here, we show that both gain and loss of DNA methylation occurs during
monocyte-to-macrophage differentiation. Differential DNA methylation is particularly found at single CpGs or small regions enriched for enhancers at positions for known macrophage transcription factor binding sides, which are important regulatory regions for macrophage identity.
145
Introduction
Inflammatory sites are characterized by the recruitment of monocytes that upon migration from the blood differentiate into macrophages to perform different regulatory functions (1). Macrophages are involved in many biological processes such as host defense, tissue remodeling and wound healing (2). Due to their broad range of functional capacities, macrophages are important regulators of disease outcome. Foam cells (i.e. macrophage that have engulfed modified lipids in the arterial wall) play an important role in the development of atherosclerosis, the primary cause of cardiovascular disease, which can result in a myocardial infarction or stroke (3). Local environmental triggers induce different activation states of the macrophages. In vitro, cells treated with lipopolysaccharide (LPS) plus interferon gamma (IFNγ) or interleukin-4 (IL-interleukin-4) are at the extreme ends of the macrophage activation spectrum, with the first having pro-inflammatory characteristics and the latter being considered anti-inflammatory (4). Understanding the molecular mechanisms controlling macrophage differentiation and activation will aid in understanding their functioning in health and disease.
The differentiation and activation processes of macrophages are accompanied by changes in their transcriptional and epigenetic make-up (5). DNA methylation is an essential component of the epigenome and defines cell development and identity. It occurs predominantly at CpG sites and previous studies identified loss of methylation as being dominant during monocyte-to-macrophage differentiation (6, 7). Here, we investigated DNA methylation changes at specific genomic positions in monocyte-to-macrophage differentiation and subsequent activation (LPS/IFNγ, IL-4, modified lipids). We found that in addition to loss of methylation also gain of methylation occurs substantially during macrophage differentiation. Moreover, we demonstrated that DNA methylation changes occur at single CpGs or small regions, primarily at enhancers. We identified that these local changes associate with distinct transcription factor binding sites for gain (ETS and C/EBP) and loss (bZIP mtoifs, AP-1 factors) of methylation necessary for macrophage identity.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 145PDF page: 145PDF page: 145PDF page: 145
144
Abstract
Introduction: Macrophages and their precursors monocytes play a key role in
inflammation and chronic inflammatory disorders like atherosclerosis. Monocyte-to-macrophage differentiation and activation is accompanied by significant epigenetic remodeling of which DNA methylation is one of the primary components. Here, we studied fine tuning of DNA methylation during differentiation and activation of macrophages, and identified transcription factors linked to differentially methylated single CpGs.
Material and methods: Human monocytes were isolated and differentiated into
macrophages in vitro and subsequently activated with LPS/IFNγ, IL-4 or modified lipids, to obtain different activation states. We assessed DNA methylation in these different subsets using the Illumina 450k array and performed an epigenome-wide analysis.
Results: We found that DNA methylation changes occur predominantly during
monocyte-to-macrophage differentiation and substantially less upon subsequent activation. We observed both gain and loss of methylation during differentiation. Differential DNA methylation was particularly found at single CpGs or small regions, which were enriched for enhancers. Furthermore, differentially methylated CpGs were located at binding sites for transcription factors known to be involved in monocyte-to-macrophage differentiation and were distinct for gain (C/EBP and ETS) and loss (AP-1) of methylation.
Conclusion: Here, we show that both gain and loss of DNA methylation occurs during
monocyte-to-macrophage differentiation. Differential DNA methylation is particularly found at single CpGs or small regions enriched for enhancers at positions for known macrophage transcription factor binding sides, which are important regulatory regions for macrophage identity.
145
Introduction
Inflammatory sites are characterized by the recruitment of monocytes that upon migration from the blood differentiate into macrophages to perform different regulatory functions (1). Macrophages are involved in many biological processes such as host defense, tissue remodeling and wound healing (2). Due to their broad range of functional capacities, macrophages are important regulators of disease outcome. Foam cells (i.e. macrophage that have engulfed modified lipids in the arterial wall) play an important role in the development of atherosclerosis, the primary cause of cardiovascular disease, which can result in a myocardial infarction or stroke (3). Local environmental triggers induce different activation states of the macrophages. In vitro, cells treated with lipopolysaccharide (LPS) plus interferon gamma (IFNγ) or interleukin-4 (IL-interleukin-4) are at the extreme ends of the macrophage activation spectrum, with the first having pro-inflammatory characteristics and the latter being considered anti-inflammatory (4). Understanding the molecular mechanisms controlling macrophage differentiation and activation will aid in understanding their functioning in health and disease.
The differentiation and activation processes of macrophages are accompanied by changes in their transcriptional and epigenetic make-up (5). DNA methylation is an essential component of the epigenome and defines cell development and identity. It occurs predominantly at CpG sites and previous studies identified loss of methylation as being dominant during monocyte-to-macrophage differentiation (6, 7). Here, we investigated DNA methylation changes at specific genomic positions in monocyte-to-macrophage differentiation and subsequent activation (LPS/IFNγ, IL-4, modified lipids). We found that in addition to loss of methylation also gain of methylation occurs substantially during macrophage differentiation. Moreover, we demonstrated that DNA methylation changes occur at single CpGs or small regions, primarily at enhancers. We identified that these local changes associate with distinct transcription factor binding sites for gain (ETS and C/EBP) and loss (bZIP mtoifs, AP-1 factors) of methylation necessary for macrophage identity.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 146PDF page: 146PDF page: 146PDF page: 146
146
Material and Methods
Monocyte isolation and macrophage culture
Peripheral blood mononuclear cells were isolated from 4 healthy donors (3 males; mean age 31.5, (SD, 4.7) from buffycoats (Sanquin blood supply, Amsterdam, the Netherlands) through density centrifugation using Lymphoprep™ (Axis-Shield, Dundee, Scotland). Monocytes were then purified using human CD14 magnetic beads and MACS® cell separation columns (Miltenyi Biotec, Bergisch Gladbach, Germany).
Monocytes were plated in 6-well tissue culture plates at a density of 1*106 cells/mL for
45 min allowing monocyte adherence in Iscove’s Modified Dulbecco’s Medium (IMDM, Sigma-Aldrich, Zwijndrecht, The Netherlands) supplemented with 2mM l-glutamine, penicillin (100 U/mL), Streptomycin (100 µg/mL) and 1 % fetal calf serum (FCS; All Gibco, Waltham, MA). Hereafter, monocytes were used for experiments or differentiated to macrophages by replacing the medium with IMDM plus 10 % FCS and 50 ng/mL MCSF (Miltenyi Biotec, Bergisch Gladbach, Germany) for 6 days. On day 3, half the medium was removed and substituted by fresh IMDM with 10 % FCS and 50 ng/mL MCSF. On day 6, all medium was removed and replaced by IMDM with 10 % FCS without MCSF and cells were activated by various stimuli for 24 hour to gain different macrophage activation states: LPS/IFNγ (10 ng/mL, Sigma-Aldrich, Zwijndrecht, The Netherlands; 50 ng/mL R&D Systems, Minneapolis, MN), IL-4 (50 ng/mL, PreProTech, Rocky Hill, NJ), oxLDL (50 µg/mL Sanbio B.V., Uden, The Netherlands) and acLDL (50 µg/mL Sanbio B.V., Uden, The Netherlands) (Figure 1).
Flow cytometry
As a control for monocyte/macrophage purity and activation, we performed flow cytometry on all subsets. 0.2*106 cells were blocked and stained with the following
antibodies for purity and differentiation: CD14, CD16, HLA-DR, CCR5, CD68; LPS/IFNγ activation: CCR7 and CD64 or IL-4 activation: CD200R and CD206 (Supplemental Table S1). CD68 was stained intracellular after fixation and permeabilization following manufactures instruction (eBioscience, San Diego, CA). Fluorescence was measured with BD Canto II and analyzed with FlowJo software version 7.6.5. (FlowJo, LLC, Ashland, OR). Monocyte purity was based on CD14+ or CD16+ gating and the expression of surface markers is presented as median fluorescence intensity (MFI).
Oil red O staining
To visualize lipid uptake, cells were plated on coverslips and lipids were stained with Oil Red O staining (0.3 % in 60 % isopropanol, Sigma). Pictures were made with a Leica DM3000 microscope.
147
DNA methylation
Genomic DNA was purified using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany), bisulfite-converted (500 ng) with the Zymo EZ DNA methylation kit (Zymo Research, Irvine, CA, USA) and hybridized (4 μl) on the Illumina 450k array using the manufacturer’s protocol (Illumina, San Diego, CA, USA). Data was generated by the Human Genotyping facility (HugeF) of ErasmusMC, The Netherlands.
Quality control and normalization
For all samples, Illumina 450k array data passed quality control using MethylAid (8). Probes with detection P-value > 0.01, bead number < 3, or zero intensity in at least one sample were removed (46718 probes were removed, resulting in a final data set of 440292 CpGs). Data were normalized using minfi’s (9) functional normalization (10) (5 principal components). A workflow for the quality control and normalization pipeline is available at: https://molepi.github.io/DNAmArray_workflow/index.html
Statistical analyses
All statistical analyses were performed using R 3.4.1 (11). The epigenome-wide analysis was performed on methylation beta values using a linear mixed model with donor as random effect for each CpG using the lmer and aov functions in lme4 (12) with P-values calculated using Satterthwaite's approximation (13). Differentially methylated CpGs (DMCs) were obtained after adjusting for multiple testing using the Benjamini-Hochberg method and deciding on a mean-squares cut-off of 0.0025; a threshold that would imply a 5 % difference in effect size if only 2 conditions were considered. Principal components were obtained using the prcomp function in stats to visualize the characteristics of the DNA methylation data.
Nearest genes were found based on distance to the nearest transcription start or end site. Gene ontology enrichment was performed using Metascape (14) (only GO Biological Processes) and upstream regulators were found using Ingenuity Pathway Analysis (15) (standard settings).
Blueprint ChIP-seq peak files were downloaded for histon marks H3K4me1, H3K4me3, H3K27ac and H3K27me3 for 5 donors (C005VG, S001S7, S0022I, S00390 and S01F8K) with both monocyte and macrophage data (16). Peak files were converted to a binary format (0 = no peak, 1 = peak) and ChromHMM (17) was used on these converted data to learn 9 chromatin states (standard settings), which were labeled according to Roadmap reference nomenclature (18). States including the H3K27ac mark, not covered in Roadmap, were designated using “active” (e.g. active enhancer, active transcription start site). For each cell type, chromatin states at the genomic position of
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 147PDF page: 147PDF page: 147PDF page: 147
146
Material and Methods
Monocyte isolation and macrophage culture
Peripheral blood mononuclear cells were isolated from 4 healthy donors (3 males; mean age 31.5, (SD, 4.7) from buffycoats (Sanquin blood supply, Amsterdam, the Netherlands) through density centrifugation using Lymphoprep™ (Axis-Shield, Dundee, Scotland). Monocytes were then purified using human CD14 magnetic beads and MACS® cell separation columns (Miltenyi Biotec, Bergisch Gladbach, Germany).
Monocytes were plated in 6-well tissue culture plates at a density of 1*106 cells/mL for
45 min allowing monocyte adherence in Iscove’s Modified Dulbecco’s Medium (IMDM, Sigma-Aldrich, Zwijndrecht, The Netherlands) supplemented with 2mM l-glutamine, penicillin (100 U/mL), Streptomycin (100 µg/mL) and 1 % fetal calf serum (FCS; All Gibco, Waltham, MA). Hereafter, monocytes were used for experiments or differentiated to macrophages by replacing the medium with IMDM plus 10 % FCS and 50 ng/mL MCSF (Miltenyi Biotec, Bergisch Gladbach, Germany) for 6 days. On day 3, half the medium was removed and substituted by fresh IMDM with 10 % FCS and 50 ng/mL MCSF. On day 6, all medium was removed and replaced by IMDM with 10 % FCS without MCSF and cells were activated by various stimuli for 24 hour to gain different macrophage activation states: LPS/IFNγ (10 ng/mL, Sigma-Aldrich, Zwijndrecht, The Netherlands; 50 ng/mL R&D Systems, Minneapolis, MN), IL-4 (50 ng/mL, PreProTech, Rocky Hill, NJ), oxLDL (50 µg/mL Sanbio B.V., Uden, The Netherlands) and acLDL (50 µg/mL Sanbio B.V., Uden, The Netherlands) (Figure 1).
Flow cytometry
As a control for monocyte/macrophage purity and activation, we performed flow cytometry on all subsets. 0.2*106 cells were blocked and stained with the following
antibodies for purity and differentiation: CD14, CD16, HLA-DR, CCR5, CD68; LPS/IFNγ activation: CCR7 and CD64 or IL-4 activation: CD200R and CD206 (Supplemental Table S1). CD68 was stained intracellular after fixation and permeabilization following manufactures instruction (eBioscience, San Diego, CA). Fluorescence was measured with BD Canto II and analyzed with FlowJo software version 7.6.5. (FlowJo, LLC, Ashland, OR). Monocyte purity was based on CD14+ or CD16+ gating and the expression of surface markers is presented as median fluorescence intensity (MFI).
Oil red O staining
To visualize lipid uptake, cells were plated on coverslips and lipids were stained with Oil Red O staining (0.3 % in 60 % isopropanol, Sigma). Pictures were made with a Leica DM3000 microscope.
147
DNA methylation
Genomic DNA was purified using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany), bisulfite-converted (500 ng) with the Zymo EZ DNA methylation kit (Zymo Research, Irvine, CA, USA) and hybridized (4 μl) on the Illumina 450k array using the manufacturer’s protocol (Illumina, San Diego, CA, USA). Data was generated by the Human Genotyping facility (HugeF) of ErasmusMC, The Netherlands.
Quality control and normalization
For all samples, Illumina 450k array data passed quality control using MethylAid (8). Probes with detection P-value > 0.01, bead number < 3, or zero intensity in at least one sample were removed (46718 probes were removed, resulting in a final data set of 440292 CpGs). Data were normalized using minfi’s (9) functional normalization (10) (5 principal components). A workflow for the quality control and normalization pipeline is available at: https://molepi.github.io/DNAmArray_workflow/index.html
Statistical analyses
All statistical analyses were performed using R 3.4.1 (11). The epigenome-wide analysis was performed on methylation beta values using a linear mixed model with donor as random effect for each CpG using the lmer and aov functions in lme4 (12) with P-values calculated using Satterthwaite's approximation (13). Differentially methylated CpGs (DMCs) were obtained after adjusting for multiple testing using the Benjamini-Hochberg method and deciding on a mean-squares cut-off of 0.0025; a threshold that would imply a 5 % difference in effect size if only 2 conditions were considered. Principal components were obtained using the prcomp function in stats to visualize the characteristics of the DNA methylation data.
Nearest genes were found based on distance to the nearest transcription start or end site. Gene ontology enrichment was performed using Metascape (14) (only GO Biological Processes) and upstream regulators were found using Ingenuity Pathway Analysis (15) (standard settings).
Blueprint ChIP-seq peak files were downloaded for histon marks H3K4me1, H3K4me3, H3K27ac and H3K27me3 for 5 donors (C005VG, S001S7, S0022I, S00390 and S01F8K) with both monocyte and macrophage data (16). Peak files were converted to a binary format (0 = no peak, 1 = peak) and ChromHMM (17) was used on these converted data to learn 9 chromatin states (standard settings), which were labeled according to Roadmap reference nomenclature (18). States including the H3K27ac mark, not covered in Roadmap, were designated using “active” (e.g. active enhancer, active transcription start site). For each cell type, chromatin states at the genomic position of
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 148PDF page: 148PDF page: 148PDF page: 148
148
measured CpGs were based on a majority call (same chromatin state in at least 3/5 donors); 5855 states were called in monocytes of the 5870 monocyte-to-macrophage differentiation DMCs, 5865 in macrophages. Enrichments for DMCs in chromatin states were calculated using Fisher’s exact test.
Blueprint whole genome bisulfite sequencing (WGBS) data were downloaded for monocytes (donors: C000S5, C0010K, C001UY, S007G7) and macrophages (donors: C005VG, S001S7, S0022I and S00390) (16) and methylation beta values were averaged for each cell type. DMRs were obtained using the following approach: the WGBS CpG overlapping with the DMC discovered in 450k data and each subsequent WGBS CpG both upstream or downstream in had a ≥ 5 % difference in methylation in the same direction.
Blueprint DNAseI hypersensitive sites sequencing peak files were downloaded and converted to a binary format (0/1) for monocytes (donors: C0010K, C0011I, C001UY, C00408, S00T4H, S00T5F, S00T6D, S00TA5, S00TT4, S00UKI, S00UME, S00YK2, S00YRP, S00YVH, S0100M, S010B0, S010MF, S010P9, S010VY, S01238, S01246, S0130A, S01342, S0137X, S013CN and S013DL) and macrophages (donors: C005VG, C006UE, S001S7 and S0022I)(16). For each cell type a DNAseI hypersensitive site was called at a genomic position using majority call (≥ 13 in monocytes, ≥ 2 in macrophages). Enrichments were calculated using Fisher’s exact test.
Motif analysis for transcription factor binding sites was performed using HOMER (19) with a 50 bp window around the DMCs. A random set of 50000 non-DMCs was used as a background.
Results
Marked DNA methylation changes occur during monocyte-to-macrophage differentiation but not during subsequent macrophage activation
Monocytes were isolated from 4 healthy donors and differentiated to macrophages in
vitro. Macrophages were subsequently activated with LPS/IFNγ, IL-4, oxidized
low-density lipoprotein (oxLDL) or acetylated LDL (acLDL) (foam cells) (Supplemental Figure S1). An overview of the study design and a summary of the primary results are shown in Figure 1.
149
We evaluated dynamic DNA methylation during differentiation and further activation at 440292 CpGs. In total, we identified 5870 differentially methylated CpGs (DMCs) (PFDR < 0.05, mean squares > 0.025) (Supplemental Table S2). The large majority of
DMCs was attributed to monocyte-to-macrophage differentiation (98%, 5780 DMCs). In contrast to previous data (6), DMCs included both gains of DNA methylation (n=4283) and loss (n=1497) of DNA methylation). Subsequent activation of macrophages resulted in DNA methylation changes at <10 CpGs, except activation with LPS/IFNγ (69 DMCs; 65 gain-DMC, 4 loss-DMC) (Figure 1, Supplemental Figure S2 and Supplemental Table S2).
We validated our results by re-analyzing a public 450k data set on monocytes, macrophages and LPS-activated macrophages (7). Despite differences in culture (G-MCSF vs M-CSF) and activation conditions (LPS vs LPS+IFNγ), we observed a high correlation of effect sizes for the 5870 DMCs (RMonocyte = 0.77, RMacrophage = 0.63 and
RM(LPS/IFNγ)= 0.72, Supplemental Figure S3).
Figure 1: Study design. Monocytes were isolated from 4 healthy donors and differentiated to macrophages in
the presence of M-CSF. Macrophages were subsequently activated to obtain different macrophage subsets. DNA methylation in all subsets was measured using the Illumina 450k array and a epigenome-wide analysis was performed using a linear mixed model with donor as random effect. Shown are the amount of specific differentially methylated CpGs (DMCs) for each of the subsets compared to all other subsets based on partial t-statistic. (PFDR < 0.05, mean squares > 0.025).
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 149PDF page: 149PDF page: 149PDF page: 149
148
measured CpGs were based on a majority call (same chromatin state in at least 3/5 donors); 5855 states were called in monocytes of the 5870 monocyte-to-macrophage differentiation DMCs, 5865 in macrophages. Enrichments for DMCs in chromatin states were calculated using Fisher’s exact test.
Blueprint whole genome bisulfite sequencing (WGBS) data were downloaded for monocytes (donors: C000S5, C0010K, C001UY, S007G7) and macrophages (donors: C005VG, S001S7, S0022I and S00390) (16) and methylation beta values were averaged for each cell type. DMRs were obtained using the following approach: the WGBS CpG overlapping with the DMC discovered in 450k data and each subsequent WGBS CpG both upstream or downstream in had a ≥ 5 % difference in methylation in the same direction.
Blueprint DNAseI hypersensitive sites sequencing peak files were downloaded and converted to a binary format (0/1) for monocytes (donors: C0010K, C0011I, C001UY, C00408, S00T4H, S00T5F, S00T6D, S00TA5, S00TT4, S00UKI, S00UME, S00YK2, S00YRP, S00YVH, S0100M, S010B0, S010MF, S010P9, S010VY, S01238, S01246, S0130A, S01342, S0137X, S013CN and S013DL) and macrophages (donors: C005VG, C006UE, S001S7 and S0022I)(16). For each cell type a DNAseI hypersensitive site was called at a genomic position using majority call (≥ 13 in monocytes, ≥ 2 in macrophages). Enrichments were calculated using Fisher’s exact test.
Motif analysis for transcription factor binding sites was performed using HOMER (19) with a 50 bp window around the DMCs. A random set of 50000 non-DMCs was used as a background.
Results
Marked DNA methylation changes occur during monocyte-to-macrophage differentiation but not during subsequent macrophage activation
Monocytes were isolated from 4 healthy donors and differentiated to macrophages in
vitro. Macrophages were subsequently activated with LPS/IFNγ, IL-4, oxidized
low-density lipoprotein (oxLDL) or acetylated LDL (acLDL) (foam cells) (Supplemental Figure S1). An overview of the study design and a summary of the primary results are shown in Figure 1.
149
We evaluated dynamic DNA methylation during differentiation and further activation at 440292 CpGs. In total, we identified 5870 differentially methylated CpGs (DMCs) (PFDR < 0.05, mean squares > 0.025) (Supplemental Table S2). The large majority of
DMCs was attributed to monocyte-to-macrophage differentiation (98%, 5780 DMCs). In contrast to previous data (6), DMCs included both gains of DNA methylation (n=4283) and loss (n=1497) of DNA methylation). Subsequent activation of macrophages resulted in DNA methylation changes at <10 CpGs, except activation with LPS/IFNγ (69 DMCs; 65 gain-DMC, 4 loss-DMC) (Figure 1, Supplemental Figure S2 and Supplemental Table S2).
We validated our results by re-analyzing a public 450k data set on monocytes, macrophages and LPS-activated macrophages (7). Despite differences in culture (G-MCSF vs M-CSF) and activation conditions (LPS vs LPS+IFNγ), we observed a high correlation of effect sizes for the 5870 DMCs (RMonocyte = 0.77, RMacrophage = 0.63 and
RM(LPS/IFNγ)= 0.72, Supplemental Figure S3).
Figure 1: Study design. Monocytes were isolated from 4 healthy donors and differentiated to macrophages in
the presence of M-CSF. Macrophages were subsequently activated to obtain different macrophage subsets. DNA methylation in all subsets was measured using the Illumina 450k array and a epigenome-wide analysis was performed using a linear mixed model with donor as random effect. Shown are the amount of specific differentially methylated CpGs (DMCs) for each of the subsets compared to all other subsets based on partial t-statistic. (PFDR < 0.05, mean squares > 0.025).
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 150PDF page: 150PDF page: 150PDF page: 150
150
Genes linked to DMCs are enriched for processes involved in monocyte-to-macrophage differentiation and monocyte-to-macrophage activation for both gain and loss of methylation
We mapped the DMCs to their nearest genes (Supplemental Table S2) and observed that many of these genes were hallmark examples of genes involved in monocyte-to-macrophage differentiation for both gain (e.g. IRF8, CEBPB) and loss (e.g. PPARG) of methylation (20, 21). Additionally we also found genes for LPS/IFNγ macrophage-specific activation (e.g. CCL5) (Figure 2).
Pathway analysis of the nearest genes showed enrichment for processes involved in monocyte and macrophage differentiation for genes that were either linked to DMCs with gain of methylation (2152 unique genes) or with genes linked to DMCs with loss of methylation (771 unique genes) (Figure 3A). Of interest, the top 20 enriched processes of gain and loss of methylation included 7 overlapping pathways between gain and loss of methylation such as ‘myeloid leukocyte activation’, ‘single organism cell adhesion’ and ‘inflammatory response’ (Figure 3A). Similarly, gain and loss shared tumor necrosis factor (TNF), a key inflammatory cytokine and known regulator of monocytes and macrophages, as upstream regulator (Figure 3B). These data imply that both gain and loss are functionally important in monocyte-to-macrophage differentiation.
Figure 2: DMCs mapped to their nearest gene, show hallmark examples of macrophage-related genes. DNA
methylation beta values for DMCs near genes involved in monocyte to macrophage differentiation (IRF8, CEPB, PPARG) and for LPS/IFNγ macrophage activation (CCL5) for the 4 donors.
151
Figure 3: Genes linked to DMCs are enriched for processes involved in monocyte-to-macrophage differentiation for both gain and loss of methylation. Enrichment analysis for differential DMCs mapped to
their nearest gene for gain and loss of DNA methylation. (A) Pathway analysis for GO-terms biological processes (B) Upstream regulator analysis by IPA. Shown is the Top 20.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 151PDF page: 151PDF page: 151PDF page: 151
150
Genes linked to DMCs are enriched for processes involved in monocyte-to-macrophage differentiation and monocyte-to-macrophage activation for both gain and loss of methylation
We mapped the DMCs to their nearest genes (Supplemental Table S2) and observed that many of these genes were hallmark examples of genes involved in monocyte-to-macrophage differentiation for both gain (e.g. IRF8, CEBPB) and loss (e.g. PPARG) of methylation (20, 21). Additionally we also found genes for LPS/IFNγ macrophage-specific activation (e.g. CCL5) (Figure 2).
Pathway analysis of the nearest genes showed enrichment for processes involved in monocyte and macrophage differentiation for genes that were either linked to DMCs with gain of methylation (2152 unique genes) or with genes linked to DMCs with loss of methylation (771 unique genes) (Figure 3A). Of interest, the top 20 enriched processes of gain and loss of methylation included 7 overlapping pathways between gain and loss of methylation such as ‘myeloid leukocyte activation’, ‘single organism cell adhesion’ and ‘inflammatory response’ (Figure 3A). Similarly, gain and loss shared tumor necrosis factor (TNF), a key inflammatory cytokine and known regulator of monocytes and macrophages, as upstream regulator (Figure 3B). These data imply that both gain and loss are functionally important in monocyte-to-macrophage differentiation.
Figure 2: DMCs mapped to their nearest gene, show hallmark examples of macrophage-related genes. DNA
methylation beta values for DMCs near genes involved in monocyte to macrophage differentiation (IRF8, CEPB, PPARG) and for LPS/IFNγ macrophage activation (CCL5) for the 4 donors.
151
Figure 3: Genes linked to DMCs are enriched for processes involved in monocyte-to-macrophage differentiation for both gain and loss of methylation. Enrichment analysis for differential DMCs mapped to
their nearest gene for gain and loss of DNA methylation. (A) Pathway analysis for GO-terms biological processes (B) Upstream regulator analysis by IPA. Shown is the Top 20.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 152PDF page: 152PDF page: 152PDF page: 152
152
Differential methylation during monocyte-to-macrophage differentiation preferentially occurs at enhancers
To characterize the regulatory landscape at the DMCs we identified 9 chromatin states in both monocytes and macrophages using a hidden Markov model based on H3K4me1, H3K4me3, H3K27ac and H3K27me3 histone marks using public BLUEPRINT data (Figure 4A). DMCs associated with monocyte-to-macrophage differentiation were enriched for enhancers (H3K4me1) and active enhancers (H3K4me1 + H3K27ac) in both monocytes and macrophages (PFDR < 0.05) (Figure 4B). Strikingly, this was not
only the case for loss-DMCs (OR > 5.6), but also gain-DMCs (OR > 3.8). Transcription start sites (H3K4me3) and repressed states (quiescent (none), polycomb (H3K27me3) and bivalent states (H3K27me3 + H3K4me1 + H3K4me3) were underrepresented in both monocytes and macrophages (OR < 1).
A direct comparison of loss- versus gain-DMCs revealed that, as expected, gain-DMCs were relatively enriched in regions that remain repressed during monocyte-to-macrophage differentiation (i.e. quiescent, polycomb or bivalent; OR > 3.11 & PFDR <
0.05) or that acquired a repressed state in macrophages while being an enhancer or flanking active TSS in monocytes (OR = 5.9 & 11.5, Figure 4C). Conversely, loss of methylation DMCs were generally enriched for regions that either lose a repressive or acquire a more active chromatin state during monocyte-to-macrophage differentiation (e.g. quiescent regions becoming enhancers (OR = 3.4) and activation of enhancers (OR = 4.9).
DNA methylation changes are located at single CpGs or small regions that are enriched for changes DNAseI hypersensitive sites and specific transcription factor binding sites
To assess whether or not DMCs identified using the sparse 450k array represented differentially methylated regions (DMRs), we overlaid DMCs with whole-genome bisulfite sequencing data of monocytes and macrophages (16). Of the 5870 DMCs, 4600 CpGs were sufficiently covered in the WGBS data and the direction of DNA methylation differences generally was concordant (Figure S4). 2213 WGBS-CpGs showed a DNA methylation difference ≥5% in the same direction during monocyte-to-macrophage differentiation, which was an enrichment when compared to non-DMCs (OR = 10.7, P-value = 0). Surprisingly, 26% of DMCs did not extend into a DMR encompassing multiple CpGs and the median number of CpGs was 3 (Figure 5A). Consequently, the length of DMRs was generally small (median 112 bp). We hypothesized that these very local differences were associated with specific transcription factor binding sites. Indeed, we found that the DMCs were enriched for
153
changes in DNAseI hypersensitive sites for both gain and loss of methylation (Figure 5B), a highly localized mark (~150bp) of open chromatin that associates with transcription factor (TF) binding (22). Further analysis revealed that gain of methylation was enriched for binding sites of ETS and C/EBP TFs, while loss of methylation was enriched for bZIP motifs and contained mainly AP-1 TFs (AP-1, ATF3, JUNB) (Figure 5C), all TFs known to be tightly involved in monocyte-to-macrophage differentiation.
Figure 4: Chromatin states at DMCs are enriched for enhancers and loss of methylation is enriched for regions that become more active during differentiation. (A) Emission parameters of the 9 chromatin states
learned using ChromHMM. (B) Enrichment analysis of DMCs based on chromatin states for monocytes (left) and
macrophages (right) for gain and loss of methylation compared to genome-wide chromatin states. (C)
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 153PDF page: 153PDF page: 153PDF page: 153
152
Differential methylation during monocyte-to-macrophage differentiation preferentially occurs at enhancers
To characterize the regulatory landscape at the DMCs we identified 9 chromatin states in both monocytes and macrophages using a hidden Markov model based on H3K4me1, H3K4me3, H3K27ac and H3K27me3 histone marks using public BLUEPRINT data (Figure 4A). DMCs associated with monocyte-to-macrophage differentiation were enriched for enhancers (H3K4me1) and active enhancers (H3K4me1 + H3K27ac) in both monocytes and macrophages (PFDR < 0.05) (Figure 4B). Strikingly, this was not
only the case for loss-DMCs (OR > 5.6), but also gain-DMCs (OR > 3.8). Transcription start sites (H3K4me3) and repressed states (quiescent (none), polycomb (H3K27me3) and bivalent states (H3K27me3 + H3K4me1 + H3K4me3) were underrepresented in both monocytes and macrophages (OR < 1).
A direct comparison of loss- versus gain-DMCs revealed that, as expected, gain-DMCs were relatively enriched in regions that remain repressed during monocyte-to-macrophage differentiation (i.e. quiescent, polycomb or bivalent; OR > 3.11 & PFDR <
0.05) or that acquired a repressed state in macrophages while being an enhancer or flanking active TSS in monocytes (OR = 5.9 & 11.5, Figure 4C). Conversely, loss of methylation DMCs were generally enriched for regions that either lose a repressive or acquire a more active chromatin state during monocyte-to-macrophage differentiation (e.g. quiescent regions becoming enhancers (OR = 3.4) and activation of enhancers (OR = 4.9).
DNA methylation changes are located at single CpGs or small regions that are enriched for changes DNAseI hypersensitive sites and specific transcription factor binding sites
To assess whether or not DMCs identified using the sparse 450k array represented differentially methylated regions (DMRs), we overlaid DMCs with whole-genome bisulfite sequencing data of monocytes and macrophages (16). Of the 5870 DMCs, 4600 CpGs were sufficiently covered in the WGBS data and the direction of DNA methylation differences generally was concordant (Figure S4). 2213 WGBS-CpGs showed a DNA methylation difference ≥5% in the same direction during monocyte-to-macrophage differentiation, which was an enrichment when compared to non-DMCs (OR = 10.7, P-value = 0). Surprisingly, 26% of DMCs did not extend into a DMR encompassing multiple CpGs and the median number of CpGs was 3 (Figure 5A). Consequently, the length of DMRs was generally small (median 112 bp). We hypothesized that these very local differences were associated with specific transcription factor binding sites. Indeed, we found that the DMCs were enriched for
153
changes in DNAseI hypersensitive sites for both gain and loss of methylation (Figure 5B), a highly localized mark (~150bp) of open chromatin that associates with transcription factor (TF) binding (22). Further analysis revealed that gain of methylation was enriched for binding sites of ETS and C/EBP TFs, while loss of methylation was enriched for bZIP motifs and contained mainly AP-1 TFs (AP-1, ATF3, JUNB) (Figure 5C), all TFs known to be tightly involved in monocyte-to-macrophage differentiation.
Figure 4: Chromatin states at DMCs are enriched for enhancers and loss of methylation is enriched for regions that become more active during differentiation. (A) Emission parameters of the 9 chromatin states
learned using ChromHMM. (B) Enrichment analysis of DMCs based on chromatin states for monocytes (left) and
macrophages (right) for gain and loss of methylation compared to genome-wide chromatin states. (C)
Enrichment of gain over loss for each of the transitions from monocyte (x-axis) to macrophage (y-axis).
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 154PDF page: 154PDF page: 154PDF page: 154
154
Figure 5: DNA methylation changes are located in small regions which are enriched for changes in DNAseI hypersensitive sites and TF-binding sites. (A) Histogram depicting the number of DMRs harbouring a specific
amount of CpGs. (B) Enrichment analysis for DNAseI hypersensitive sites for gain (red) and loss (blue) of DNA
methylation compared to genome wide DNAseI hypersensitive sites. (C) Motif analysis with HOMER for
transcription factor binding sites on differential DMCs for gain (left) and loss (right) of DNA methylation.
Discussion
We here report on dynamic DNA methylation during monocyte-to-macrophage differentiation and subsequent activation. We observed that DNA methylation occurs primarily during monocyte-to-macrophage differentiation and substantially less during macrophage activation. We identified that in addition to loss of methylation also gain of DNA methylation occurs markedly during differentiation, primarily at enhancers.
155
Our data also show that remodeling of the epigenome during differentiation predominantly happens by very local tweaking of the methylation status of the chromatin and does not involve changes in large regions. We identified that these local changes associate with distinct binding sites of transcription factors necessary for macrophage identity for gain (ETS and C/EBP) and loss (bZIP motifs, AP-1 factors) of methylation. This implicates that both gain and loss of methylation have a distinct yet important biological function in the differentiation process.
While previous studies observed primarily loss of methylation during monocyte-to-macrophage differentiation (6, 7) we find predominant gain of DNA methylation. We identified that both gain and loss of methylation are located near genes and upstream regulators involved in monocyte-to-macrophage differentiation and macrophage activation indicating that both gain and loss of methylation are biologically relevant. We found pathways involved in cell morphogenesis, leukocyte migration, inflammatory response, myeloid leukocyte activation, response to lipid, response to wounding and single organism cell adhesion to be overlapping for gain and loss of methylation. TNF, a pro-inflammatory cytokine secreted in response to activation of monocytes and macrophages and important regulator of macrophage function (4), was predicted as an primary upstream regulator for both gain and loss in DNA methylation. The fact that TNF is associated with gain and loss of methylation suggests that there is a rearrangement of the TNF response upon monocyte-to-macrophage differentiation.
While loss of methylation has previously been observed to be enriched at enhancer regions, we identified that this is also true for gain of methylation. Nevertheless, gain of methylation occurs at regions that are enriched to become more repressed during monocyte-to-macrophage differentiation when compared to loss of methylation and, conversely, loss of methylation occurs at regions that are enriched to become more active, consistent with previous studies where similar effects were observed in B-cell differentiation (23) and in fetal development (24).
We observed that these methylation changes occur at single CpG sites or at small regions. These small regions are enriched for changes in DNAseI hypersensitive sites, locations associated with TF-binding. As a consequence, studies that set out to find differentially methylated regions (DMRs), for example the study that reported differential methylation between monocytes and macrophages at DMRs (>4 CpGs) (6), are expected to miss half of the DNA methylation changes since the median of CpGs we found in a region was three.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 155PDF page: 155PDF page: 155PDF page: 155
154
Figure 5: DNA methylation changes are located in small regions which are enriched for changes in DNAseI hypersensitive sites and TF-binding sites. (A) Histogram depicting the number of DMRs harbouring a specific
amount of CpGs. (B) Enrichment analysis for DNAseI hypersensitive sites for gain (red) and loss (blue) of DNA
methylation compared to genome wide DNAseI hypersensitive sites. (C) Motif analysis with HOMER for
transcription factor binding sites on differential DMCs for gain (left) and loss (right) of DNA methylation.
Discussion
We here report on dynamic DNA methylation during monocyte-to-macrophage differentiation and subsequent activation. We observed that DNA methylation occurs primarily during monocyte-to-macrophage differentiation and substantially less during macrophage activation. We identified that in addition to loss of methylation also gain of DNA methylation occurs markedly during differentiation, primarily at enhancers.
155
Our data also show that remodeling of the epigenome during differentiation predominantly happens by very local tweaking of the methylation status of the chromatin and does not involve changes in large regions. We identified that these local changes associate with distinct binding sites of transcription factors necessary for macrophage identity for gain (ETS and C/EBP) and loss (bZIP motifs, AP-1 factors) of methylation. This implicates that both gain and loss of methylation have a distinct yet important biological function in the differentiation process.
While previous studies observed primarily loss of methylation during monocyte-to-macrophage differentiation (6, 7) we find predominant gain of DNA methylation. We identified that both gain and loss of methylation are located near genes and upstream regulators involved in monocyte-to-macrophage differentiation and macrophage activation indicating that both gain and loss of methylation are biologically relevant. We found pathways involved in cell morphogenesis, leukocyte migration, inflammatory response, myeloid leukocyte activation, response to lipid, response to wounding and single organism cell adhesion to be overlapping for gain and loss of methylation. TNF, a pro-inflammatory cytokine secreted in response to activation of monocytes and macrophages and important regulator of macrophage function (4), was predicted as an primary upstream regulator for both gain and loss in DNA methylation. The fact that TNF is associated with gain and loss of methylation suggests that there is a rearrangement of the TNF response upon monocyte-to-macrophage differentiation.
While loss of methylation has previously been observed to be enriched at enhancer regions, we identified that this is also true for gain of methylation. Nevertheless, gain of methylation occurs at regions that are enriched to become more repressed during monocyte-to-macrophage differentiation when compared to loss of methylation and, conversely, loss of methylation occurs at regions that are enriched to become more active, consistent with previous studies where similar effects were observed in B-cell differentiation (23) and in fetal development (24).
We observed that these methylation changes occur at single CpG sites or at small regions. These small regions are enriched for changes in DNAseI hypersensitive sites, locations associated with TF-binding. As a consequence, studies that set out to find differentially methylated regions (DMRs), for example the study that reported differential methylation between monocytes and macrophages at DMRs (>4 CpGs) (6), are expected to miss half of the DNA methylation changes since the median of CpGs we found in a region was three.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 156PDF page: 156PDF page: 156PDF page: 156
156
We identified that DMCs for both gain and loss of methylation are enriched for known macrophage TF-binding sites. Gain of methylation is enriched for binding sites of C/EBP and ETS TFs, while loss of methylation is enriched for binding sites with bZIP motifs that bind AP-1 factors (AP-1, ATF3, JUNB). Enhancers are in general enriched for motifs that bind lineage determing transcription factors (LDTFs) needed for cell identity (18). Moreover, it has been shown in mouse macrophages that enhancers are enriched for motifs that bind PU.1 (ETS factor) and CEBP, which are required for the differentiation and function of macrophages (19, 25). Lara-astiaso et al. performed motif analysis of the enhancers which overlapped with open chromatin and identified the LDTFs PU.1 and C/EBPβ and C/EBPα as central regulators of myeloid enhancers (26). Consistently, also in human monocytes and macrophages it was shown that C/EBP and PU.1 binding motifs were especially found in enhancers (27). Interestingly, we identified these LDTFs, which are important factors for macrophage lineage commitment, to be associated with gain of methylation. Contrastingly, we found AP-1-like TFs to be strongly associated with loss of methylation. AP-1 AP-1-like factors are TFs involved in differentiation but also particularly in regulating macrophage activation and production of inflammatory factors such as cytokines and chemokines (20, 28). That both gain and loss of DNA methylation are associated with TF-binding is consistent with previous studies that show that CpG methylation can influence TF-binding, both negatively and positively, depending on the methylated motif (29). They show that binding of bZIP & ETS TFs was inhibited by CpG methylation. These are also the motifs affected by gain and loss of methylation. Therefore, we speculate that loci that bind LDTF are switched off during differentiation (gain of methylation), while positions with loss of methylation can be bound by AP-1 factors in macrophages. AP-1 factors are stimulus dependent transcription factors (SDTFs) and loss of methylation may imply that chromatin remodeling during differentiation of macrophages prepares the cells to allow proper AP-1 mediated responses to viral and bacterial triggers. In order to draw any conclusions about exact TF binding, chromatin immunoprecipitation (ChIP) experiments for the predicted TFs should be performed in monocytes and macrophages.
In conclusion, we reveal that both gain and loss of DNA methylation at very local genomic positions contribute to monocyte-to-macrophage differentiation. We identified DNA methylation as regulator of macrophage differentiation, rather than macrophage activation. In chronic inflammatory disorders like atherosclerosis, circulating monocytes are exposed to risk factors for disease like elevated levels of lipids. A key question that remains is how disease risk factors fine-tune DNA methylation during cellular differentiation and thereby contribute to disease outcome.
157
Acknowledgements
This work was mainly supported by The Netherlands Heart Foundation (CVON 2011/ B019: Generating the best evidence-based pharmaceutical targets for atherosclerosis [GENIUS]). Menno PJ de Winther is an established investigator of the Netherlands Heart Foundation (2007T067), is supported by a Netherlands Heart Foundation grant (2010B022), Spark-Holding BV (2015B002), the European Union (ITN-grant EPIMAC), REPROGRAM (EU Horizon 2020) and holds an AMC-fellowship.
References
1. Murray PJ, Wynn TA. Protective and pathogenic functions of macrophage subsets. Nature reviews Immunology. 2011;11(11):723-37.
2. Martinez FO, Gordon S. The M1 and M2 paradigm of macrophage activation: time for reassessment. F1000prime reports. 2014;6:13.
3. Moore KJ, Sheedy FJ, Fisher EA. Macrophages in atherosclerosis: a dynamic balance. Nature reviews Immunology. 2013;13(10):709-21.
4. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activation. Nature reviews Immunology. 2008;8(12):958-69.
5. Saeed S, Quintin J, Kerstens HH, Rao NA, Aghajanirefah A, Matarese F, et al. Epigenetic programming of monocyte-to-macrophage differentiation and trained innate immunity. Science. 2014;345(6204):1251086. 6. Wallner S, Schroder C, Leitao E, Berulava T, Haak C, Beisser D, et al. Epigenetic dynamics of monocyte-to-macrophage differentiation. Epigenetics & chromatin. 2016;9:33.
7. Vento-Tormo R, Company C, Rodriguez-Ubreva J, de la Rica L, Urquiza JM, Javierre BM, et al. IL-4 orchestrates STAT6-mediated DNA demethylation leading to dendritic cell differentiation. Genome biology. 2016;17:4.
8. van Iterson M, Tobi EW, Slieker RC, den Hollander W, Luijk R, Slagboom PE, et al. MethylAid: visual and interactive quality control of large Illumina 450k datasets. Bioinformatics. 2014;30(23):3435-7.
9. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30(10):1363-9.
10. Fortin JP, Labbe A, Lemire M, Zanke BW, Hudson TJ, Fertig EJ, et al. Functional normalization of 450k methylation array data improves replication in large cancer studies. Genome Biol. 2014;15(12):503.
11. R Core Team. R: A language and environment for statistical computing. 2017.
12. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear Mixed-Effects Models Usinglme4. Journal of Statistical Software. 2015;67(1).
13. Satterthwaite FE. An Approximate Distribution of Estimates of Variance Components. Biometrics Bulletin. 1946;2(6):110.
14. Tripathi S, Pohl MO, Zhou Y, Rodriguez-Frandsen A, Wang G, Stein DA, et al. Meta- and Orthogonal Integration of Influenza "OMICs" Data Defines a Role for UBR4 in Virus Budding. Cell Host Microbe. 2015;18(6):723-35.
15. Kramer A, Green J, Pollard J, Jr., Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30(4):523-30.
16. Stunnenberg HG, International Human Epigenome C, Hirst M. The International Human Epigenome Consortium: A Blueprint for Scientific Collaboration and Discovery. Cell. 2016;167(5):1145-9.
17. Ernst J, Kellis M. ChromHMM: automating chromatin-state discovery and characterization. Nat Methods. 2012;9(3):215-6.
18. Roadmap Epigenomics C, Kundaje A, Meuleman W, Ernst J, Bilenky M, Yen A, et al. Integrative analysis of 111 reference human epigenomes. Nature. 2015;518(7539):317-30.
19. Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Molecular cell. 2010;38(4):576-89.
20. Lawrence T, Natoli G. Transcriptional regulation of macrophage polarization: enabling diversity with identity. Nature reviews Immunology. 2011;11(11):750-61.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 157PDF page: 157PDF page: 157PDF page: 157
156
We identified that DMCs for both gain and loss of methylation are enriched for known macrophage TF-binding sites. Gain of methylation is enriched for binding sites of C/EBP and ETS TFs, while loss of methylation is enriched for binding sites with bZIP motifs that bind AP-1 factors (AP-1, ATF3, JUNB). Enhancers are in general enriched for motifs that bind lineage determing transcription factors (LDTFs) needed for cell identity (18). Moreover, it has been shown in mouse macrophages that enhancers are enriched for motifs that bind PU.1 (ETS factor) and CEBP, which are required for the differentiation and function of macrophages (19, 25). Lara-astiaso et al. performed motif analysis of the enhancers which overlapped with open chromatin and identified the LDTFs PU.1 and C/EBPβ and C/EBPα as central regulators of myeloid enhancers (26). Consistently, also in human monocytes and macrophages it was shown that C/EBP and PU.1 binding motifs were especially found in enhancers (27). Interestingly, we identified these LDTFs, which are important factors for macrophage lineage commitment, to be associated with gain of methylation. Contrastingly, we found AP-1-like TFs to be strongly associated with loss of methylation. AP-1 AP-1-like factors are TFs involved in differentiation but also particularly in regulating macrophage activation and production of inflammatory factors such as cytokines and chemokines (20, 28). That both gain and loss of DNA methylation are associated with TF-binding is consistent with previous studies that show that CpG methylation can influence TF-binding, both negatively and positively, depending on the methylated motif (29). They show that binding of bZIP & ETS TFs was inhibited by CpG methylation. These are also the motifs affected by gain and loss of methylation. Therefore, we speculate that loci that bind LDTF are switched off during differentiation (gain of methylation), while positions with loss of methylation can be bound by AP-1 factors in macrophages. AP-1 factors are stimulus dependent transcription factors (SDTFs) and loss of methylation may imply that chromatin remodeling during differentiation of macrophages prepares the cells to allow proper AP-1 mediated responses to viral and bacterial triggers. In order to draw any conclusions about exact TF binding, chromatin immunoprecipitation (ChIP) experiments for the predicted TFs should be performed in monocytes and macrophages.
In conclusion, we reveal that both gain and loss of DNA methylation at very local genomic positions contribute to monocyte-to-macrophage differentiation. We identified DNA methylation as regulator of macrophage differentiation, rather than macrophage activation. In chronic inflammatory disorders like atherosclerosis, circulating monocytes are exposed to risk factors for disease like elevated levels of lipids. A key question that remains is how disease risk factors fine-tune DNA methylation during cellular differentiation and thereby contribute to disease outcome.
157
Acknowledgements
This work was mainly supported by The Netherlands Heart Foundation (CVON 2011/ B019: Generating the best evidence-based pharmaceutical targets for atherosclerosis [GENIUS]). Menno PJ de Winther is an established investigator of the Netherlands Heart Foundation (2007T067), is supported by a Netherlands Heart Foundation grant (2010B022), Spark-Holding BV (2015B002), the European Union (ITN-grant EPIMAC), REPROGRAM (EU Horizon 2020) and holds an AMC-fellowship.
References
1. Murray PJ, Wynn TA. Protective and pathogenic functions of macrophage subsets. Nature reviews Immunology. 2011;11(11):723-37.
2. Martinez FO, Gordon S. The M1 and M2 paradigm of macrophage activation: time for reassessment. F1000prime reports. 2014;6:13.
3. Moore KJ, Sheedy FJ, Fisher EA. Macrophages in atherosclerosis: a dynamic balance. Nature reviews Immunology. 2013;13(10):709-21.
4. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activation. Nature reviews Immunology. 2008;8(12):958-69.
5. Saeed S, Quintin J, Kerstens HH, Rao NA, Aghajanirefah A, Matarese F, et al. Epigenetic programming of monocyte-to-macrophage differentiation and trained innate immunity. Science. 2014;345(6204):1251086. 6. Wallner S, Schroder C, Leitao E, Berulava T, Haak C, Beisser D, et al. Epigenetic dynamics of monocyte-to-macrophage differentiation. Epigenetics & chromatin. 2016;9:33.
7. Vento-Tormo R, Company C, Rodriguez-Ubreva J, de la Rica L, Urquiza JM, Javierre BM, et al. IL-4 orchestrates STAT6-mediated DNA demethylation leading to dendritic cell differentiation. Genome biology. 2016;17:4.
8. van Iterson M, Tobi EW, Slieker RC, den Hollander W, Luijk R, Slagboom PE, et al. MethylAid: visual and interactive quality control of large Illumina 450k datasets. Bioinformatics. 2014;30(23):3435-7.
9. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30(10):1363-9.
10. Fortin JP, Labbe A, Lemire M, Zanke BW, Hudson TJ, Fertig EJ, et al. Functional normalization of 450k methylation array data improves replication in large cancer studies. Genome Biol. 2014;15(12):503.
11. R Core Team. R: A language and environment for statistical computing. 2017.
12. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear Mixed-Effects Models Usinglme4. Journal of Statistical Software. 2015;67(1).
13. Satterthwaite FE. An Approximate Distribution of Estimates of Variance Components. Biometrics Bulletin. 1946;2(6):110.
14. Tripathi S, Pohl MO, Zhou Y, Rodriguez-Frandsen A, Wang G, Stein DA, et al. Meta- and Orthogonal Integration of Influenza "OMICs" Data Defines a Role for UBR4 in Virus Budding. Cell Host Microbe. 2015;18(6):723-35.
15. Kramer A, Green J, Pollard J, Jr., Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics. 2014;30(4):523-30.
16. Stunnenberg HG, International Human Epigenome C, Hirst M. The International Human Epigenome Consortium: A Blueprint for Scientific Collaboration and Discovery. Cell. 2016;167(5):1145-9.
17. Ernst J, Kellis M. ChromHMM: automating chromatin-state discovery and characterization. Nat Methods. 2012;9(3):215-6.
18. Roadmap Epigenomics C, Kundaje A, Meuleman W, Ernst J, Bilenky M, Yen A, et al. Integrative analysis of 111 reference human epigenomes. Nature. 2015;518(7539):317-30.
19. Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Molecular cell. 2010;38(4):576-89.
20. Lawrence T, Natoli G. Transcriptional regulation of macrophage polarization: enabling diversity with identity. Nature reviews Immunology. 2011;11(11):750-61.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 158PDF page: 158PDF page: 158PDF page: 158
158
21. Zhu YP, Thomas GD, Hedrick CC. 2014 Jeffrey M. Hoeg Award Lecture: Transcriptional Control of Monocyte Development. Arterioscler Thromb Vasc Biol. 2016;36(9):1722-33.
22. Thurman RE, Rynes E, Humbert R, Vierstra J, Maurano MT, Haugen E, et al. The accessible chromatin landscape of the human genome. Nature. 2012;489(7414):75-82.
23. Kulis M, Merkel A, Heath S, Queiros AC, Schuyler RP, Castellano G, et al. Whole-genome fingerprint of the DNA methylome during human B cell differentiation. Nat Genet. 2015;47(7):746-56.
24. Slieker RC, Roost MS, van Iperen L, Suchiman HE, Tobi EW, Carlotti F, et al. DNA Methylation Landscapes of Human Fetal Development. PLoS Genet. 2015;11(10):e1005583.
25. Ghisletti S, Barozzi I, Mietton F, Polletti S, De Santa F, Venturini E, et al. Identification and characterization of enhancers controlling the inflammatory gene expression program in macrophages. Immunity. 2010;32(3):317-28.
26. Lara-Astiaso D, Weiner A, Lorenzo-Vivas E, Zaretsky I, Jaitin DA, David E, et al. Immunogenetics. Chromatin state dynamics during blood formation. Science. 2014;345(6199):943-9.
27. Pham TH, Benner C, Lichtinger M, Schwarzfischer L, Hu YH, Andreesen R, et al. Dynamic epigenetic enhancer signatures reveal key transcription factors associated with monocytic differentiation states. Blood. 2012;119(24):E161-E71.
28. Glass CK, Natoli G. Molecular control of activation and priming in macrophages. Nature immunology. 2016;17(1):26-33.
29. Yin Y, Morgunova E, Jolma A, Kaasinen E, Sahu B, Khund-Sayeed S, et al. Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science. 2017;356(6337).
159
Supplemental Tables
Supplemental Table S1: Antibodies used for flow cytometry analysis
Surface marker Flurochrome Dilution Supplier
CD14 PE-Cy7 1/50 BD CD16 APC-Cy7 1/50 BD HLA-DR PerCpCy5.5 1/50 BD CCR5 CD68 CCR7 CD64 CD200R CD206 FITC PE PE APC PE FITC 1/10 1/50 1/25 1/50 1/25 1/25 BD BD Biolegend Biolegend Biolegend Sony
Supplemental Table S2: Differentially methylated CpGs (DMCs). The table includes all 5870
DMCs with the characteristics for each position. The supplemental table can be found online at:
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 159PDF page: 159PDF page: 159PDF page: 159
158
21. Zhu YP, Thomas GD, Hedrick CC. 2014 Jeffrey M. Hoeg Award Lecture: Transcriptional Control of Monocyte Development. Arterioscler Thromb Vasc Biol. 2016;36(9):1722-33.
22. Thurman RE, Rynes E, Humbert R, Vierstra J, Maurano MT, Haugen E, et al. The accessible chromatin landscape of the human genome. Nature. 2012;489(7414):75-82.
23. Kulis M, Merkel A, Heath S, Queiros AC, Schuyler RP, Castellano G, et al. Whole-genome fingerprint of the DNA methylome during human B cell differentiation. Nat Genet. 2015;47(7):746-56.
24. Slieker RC, Roost MS, van Iperen L, Suchiman HE, Tobi EW, Carlotti F, et al. DNA Methylation Landscapes of Human Fetal Development. PLoS Genet. 2015;11(10):e1005583.
25. Ghisletti S, Barozzi I, Mietton F, Polletti S, De Santa F, Venturini E, et al. Identification and characterization of enhancers controlling the inflammatory gene expression program in macrophages. Immunity. 2010;32(3):317-28.
26. Lara-Astiaso D, Weiner A, Lorenzo-Vivas E, Zaretsky I, Jaitin DA, David E, et al. Immunogenetics. Chromatin state dynamics during blood formation. Science. 2014;345(6199):943-9.
27. Pham TH, Benner C, Lichtinger M, Schwarzfischer L, Hu YH, Andreesen R, et al. Dynamic epigenetic enhancer signatures reveal key transcription factors associated with monocytic differentiation states. Blood. 2012;119(24):E161-E71.
28. Glass CK, Natoli G. Molecular control of activation and priming in macrophages. Nature immunology. 2016;17(1):26-33.
29. Yin Y, Morgunova E, Jolma A, Kaasinen E, Sahu B, Khund-Sayeed S, et al. Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science. 2017;356(6337).
159
Supplemental Tables
Supplemental Table S1: Antibodies used for flow cytometry analysis
Surface marker Flurochrome Dilution Supplier
CD14 PE-Cy7 1/50 BD CD16 APC-Cy7 1/50 BD HLA-DR PerCpCy5.5 1/50 BD CCR5 CD68 CCR7 CD64 CD200R CD206 FITC PE PE APC PE FITC 1/10 1/50 1/25 1/50 1/25 1/25 BD BD Biolegend Biolegend Biolegend Sony
Supplemental Table S2: Differentially methylated CpGs (DMCs). The table includes all 5870
DMCs with the characteristics for each position. The supplemental table can be found online at:
http://www.macrophages.eu/downloads/
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 160PDF page: 160PDF page: 160PDF page: 160
160
Supplemental Figures
Supplemental figure S1: Monocytes were successfully differentiated to macrophages. (A) monocyte and
macrophage purity based on CD14 and CD16 expression as assessed by flow cytometry. (B) Monocyte and
macrophage purity (% CD14+, CD16+ cells) for the 4 donors and macrophage subsets. (C) surface expression of
CD14, CD16, CD68, HLA-DR and CCR5 for monocytes and macrophage subsets. (D) Surface expression of the
marker genes CCR7 and CD64 after stimulation with LPS (10ng/ml) + IFNy (50ng/ml) for all 4 donors. (E) Surface
expression of the marker genes CD200R and CD206 after stimulation with IL-4 (50ng/ml) for all 4 donors. (F) Oil
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 161PDF page: 161PDF page: 161PDF page: 161
161
Supplemental figure S2: DNA methylation clusters on donor and monocyte versus macrophage. Principal
components calculated on the methylation beta values for each sample. PC1-PC3 cluster on donor, PC4 clusters on monocyte versus macrophage.
Supplemental figure S3: Differentially methylated CpGs were validated using public data. Cell type specific
regression estimates obtained using a linear mixed model were compared for the 5870 differentially methylated CpGs with public data re-analyzed using the same method for monocytes, macrophages and LPS/IFNγ macrophages.
515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele 515552-L-bw-neele Processed on: 1-12-2017 Processed on: 1-12-2017 Processed on: 1-12-2017
Processed on: 1-12-2017 PDF page: 162PDF page: 162PDF page: 162PDF page: 162
162
Supplemental figure S4: Methylation differences for the differentially methylated CpGs were generally concordant with public WGBS data. Monocyte specific regression estimates obtained using a linear mixed
model were compared for the 4648 differentially methylated CpGs with differences in monocyte and macrophage methylation in public WGBS data.
