The authors have declared that no competing interests exist.
Conceived and designed the experiments: SW MG MP EO GS. Performed the experiments: SW MG AS MP GL. Analyzed the data: SW T. Konovalova AS MP GL GS. Contributed reagents/materials/analysis tools: T. Kopf. Wrote the paper: SW GS.
Dysregulation of monocyte-macrophage differentiation is a hallmark of vascular and metabolic diseases and associated with persistent low grade inflammation. Plasmalogens represent ether lipids that play a role in diabesity and previous data show diminished plasmalogen levels in obese subjects. We therefore analyzed transcriptomic and lipidomic changes during monocyte-macrophage differentiation
Elutriated monocytes from 13 healthy donors were differentiated
Gene expression analysis showed strong regulation of lipidome-related transcripts. Enzymes involved in fatty acid desaturation and elongation were increasingly expressed, peroxisomal and ER stress related genes were induced. Total plasmalogen levels remained unchanged, while the PE plasmalogen species pattern became more similar to circulating granulocytes, showing decreases in PUFA and increases in MUFA. A partial least squares discriminant analysis (PLS/DA) revealed that PE plasmalogens discriminate the stage of monocyte-derived macrophage differentiation. Partial correlation analysis could predict novel potential key nodes including DOCK1, PDK4, GNPTAB and FAM126A that might be involved in regulating lipid and especially plasmalogen homeostasis during differentiation. An
Monocyte to macrophage differentiation goes along with profound changes in the lipid-related transcriptome. This leads to an induction of fatty-acid desaturation and elongation. In their PE-plasmalogen profile macrophages become more similar to granulocytes than monocytes, indicating terminal phagocytic differentiation. Therefore PE plasmalogens may represent potential biomarkers for cell activation. For the underlying transcriptional network we were able to predict a range of novel central key nodes and underlying transcription factors using a bioinformatic approach.
Macrophages are key players in innate immunity and play an important role in the development of atherosclerosis and insulin resistance in diabesity
Lipids regulate biological processes either locally as membrane components or remotely as signaling molecules. The lipid composition of the plasma membrane determines membrane fluidity but direct lipid-protein interactions also play a role in cellular signaling
In a previous study we could demonstrate that a SREBP1 dependent induction of monocyte fatty acid synthesis is vital for monocyte-macrophage differentiation
If not otherwise stated all materials including 1α,25-dihydroxyvitamin D3 (VitD3), all−trans−retinoic acid (ATRA) and ethanol were obtained from Sigma (Germany). Carrier-free macrophage colony-stimulating factor (M-CSF) and granulocyte macrophage colony-stimulating factor (GM-CSF) were both from R&D systems. HPLC grade methanol and chloroform were from Merck (Germany), plasmalogen standards from Avanti Polar Lipids (Alabaster, AL, USA).
Blood samples were obtained from thirteen healthy normolipidemic volunteers recruited from blood donors with apoE3/E3 genotype. Informed consent and approval of the Hospital Ethics Committee were obtained (Universitätsklinikum Regensburg, Ethikkommission der medizinischen Fakultät, proposal 08/119). Donors were fully informed of the possible complications and gave their written consent for the procedure. Blood cells were collected by leukapheresis in a Spectra cell separator (Gambro BCT, CO, USA) followed by subsequent counterflow elutriation as described elsewhere
HL-60 cells (ATCC; CCL-240) were grown in an incubator at 37°C (95% humidity, 5% CO2) in Iscove's Modified Dulbecco's Medium (IMDM) (PAN-Biotech) containing 20% FBS (Biochrom) in polystyrene tissue culture flasks (Sarstedt). Cells were seeded at a density of 0.3×106 cells per ml and 50% of the medium was replaced once per week. Confluent cultures were split 1∶3 to 1∶4 once per week.
Differentiation experiments were performed in Cell+ tissue culture flasks (Sarstedt). 1 μM working solutions of Vitamin D3 (VitD3)and all-trans retinoic acid (ATRA) were prepared in ethanol. Cells were seeded at 0.3×106 cells per ml. Monocytic differentiation was achieved in macrophage serum-free medium (SFM) supplemented with M-CSF (50 ng/ml) and 0.5 μM VitD3. Granulocytic differentiation was carried out in macrophage SFM with GM-CSF (10 ng/ml) and 0.5 μM ATRA. Cells grown in the respective medium containing 0,05% ethanol were used as controls. Cells were harvested by scraping with a rubber policeman and centrifugation for 5 min at 500 g at room temperature. They were finally washed once with PBS. Cell number was determined by counting in a Casy automated cell counter (Roche). Differentiation phenotype and viability were monitored by flow cytometry and Pappenheim staining. Experiments were repeated in two biological replicates.
Cells from nine donors were harvested, washed in PBS, resuspended in buffer RLT and RNA was isolated using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Purity and integrity of the RNA were determined on the Agilent 2100 Bioanalyzer with the RNA 6000 Nano LabChip reagent set (Agilent Technologies). RNA was quantified using a Nanodrop ND-1000 - UV/Vis spectrophotometer (PeqLab).
For gene array analysis we used a modified Agilent 4×44K microarray (014850) containing 205 free positions (Agilent Technologies). To these positions we added 201 probes, corresponding to 119 genes of the lipidome (LipidomicNet). 300 ng of total RNA were labeled with Cy3 using the Agilent Quick-Amp Labeling Kit - 1 color according to the manufacturer's instructions. cRNA was purified with the RNeasy Mini Kit (Qiagen). The amount and labeling efficiency of cRNA was measured with the NanoDrop ND-1000. For hybridization the Agilent Gene Expression Hybridisation Kit was used and arrays were incubated in Agilent SureHyb chambers for 17 hours in a hybridization oven at 65°C while rotating. Washing was performed according to the manufacturer's instructions. Scanning was done with the Agilent G2565CA Microarray Scanner System. The resulting TIFF files were processed with Agilent Feature Extraction software (10.7). Microarray data are available in the ArrayExpress database (
Cell pellets were dissolved in 0.2% SDS solution. An aliquot corresponding to 100 μg was used for mass spectrometric lipid analysis. Lipid extraction was performed according to the method of Bligh and Dyer
Partial least square discrimination analysis was performed and VIP values were calculated as described in
Correlations between lipid species were calculated in SPSS 20 (IBM) using Pearson's product moment correlational analysis.
Microarray experiments were analyzed according to the method described by Kondrakhin et al.
Significantly expressed and regulated lipid related genes were subjected to transcription factor and virtual promoter analysis, predicting potential transcriptional regulators.
q-order partial correlation graphs (qpgraphs) were used to reverse engineer the molecular regulatory network from microarray and lipid data. All data was Z-transformed prior to analysis and average non-rejection rates (NRR) were calculated in R 2.15.3
Changes in the transcriptional profile during monocyte to macrophage differentiation were analyzed using microarrays for a comprehensive analysis of the cellular transcriptome at days 1, 4 and 5 of MCSF dependent
Normalized enrichment scores (NES) indicate the distribution of Gene Ontology categories across a list of genes ranked by hypergeometrical score (HGS). Higher enrichment scores indicate a shift of genes belonging to certain GO categories towards either end of the ranked list, representing up or down regulation (positive or negative values, respectively). Lipid related catergories are generally found to be shifted upwards.
Fatty acids are carboxylic acids with a long aliphatic tail. They are synthesized from acetyl-CoA and malonyl CoA precursors via six recurrent reactions that are catalyzed by the enzyme fatty acid synthase (FAS) until the 16-carbon product palmitic acid (C16:0) is produced. The main function of FAS is therefore to catalyze the synthesis of palmitate from acetyl-CoA and malonyl-CoA, in the presence of NADPH. The resulting saturated fatty acids are consecutively desaturated by desaturases or elongated by elongases at the endoplasmic-reticulum (ER). An overview over this process is given in
Fatty acid synthesis in the differentiation of primary human monocytes shows an induction of enzymes involved in the synthesis of palmitate (FAS), desaturation (SCD, D5D and D6D) and elongation (ELOVL5 & 6). Blue indicates significantly downregulated transcript levels, upregulated transcripts are shown in red and unchanged levels or unknown regulation in shown in grey.
Desaturases introduce double bonds between defined carbon atoms in fatty acyl chains. In humans four types of fatty acid desaturases could be identified, according to the position of the double bond introduced. In our experiments all analyzed fatty acid desaturases were upregulated with the notable exception of FADS3, which is interesting since the gene encoding FADS3 is clustered with FADS1 and FADS2
The mitochondrial very long-chain specific acyl-CoA dehydrogenase (ACADVL) was strongly up-regulated. Its function is the degradation of very long fatty acids by catalyzing the first step of the mitochondrial fatty acid beta-oxidation pathway. It can accommodate substrates with chain lengths as long as 24 C-atoms
Lipoproteins are aggregates of apo-lipoprotein (apo) and lipid that allow the transport of lipids in the aqueous blood stream. Especially plasmalogens are known to correlate positively to serum HDL levels and to decrease with aging
Accumulation of excess free cholesterol in macrophages contributes to the development of atherosclerosis by stimulating tumor necrosis factor-α (TNF-α) and interleukin-6 production
Plasmalogens are ether glycerol-phospholipids that are integral parts of the plasma membrane as well as storage lipids for acyl residues and cellular anti-oxidants. Their synthesis takes place in a non-redundant pathway in peroxisomes and at the ER-membrane. Interestingly, although the macrophage is prepared for an immune reaction, transcripts for enzymes in plasmalogen synthesis remained mostly unchanged. The rate limiting enzyme for this process, glyceronephosphate O-acyltransferase (GNPAT, also called DHAP-AT), was not regulated on the transcriptomic level. The acyl residues in phosphatidyl-ethanolamine (PE) plasmalogens are taken from the cellular free fatty acid pool. For this purpose the enzyme fatty acyl-CoA reductase 1 (FAR1) provides the long chain fatty alcohol by reduction of saturated fatty acyl-CoAs in the peroxisomal membrane. During monocyte-macrophage differentiation its mRNA was downregulated. The resulting alcohols are then further processed by alkyldihydroxyacetonephosphate synthase (AGPS/ADHAP-S) in the peroxisome, an enzyme whose transcript was downregulated on day 4 and increased on day 5. These findings are in agreement with constant levels of total plasmalogens that were detected in lipid mass spectrometry (
(A) Levels of total plasmalogens over thre course of 5 day primary monocyte macrophage differentiation. Individual PE plasmalogen species show specific changes during differentiation (B and C). Saturation specificity relative to day 0 is shown in panel (D).
Eicosanoids are inflammatory mediators that are synthesized in macrophages from eicosapentaenoic acid (EPA, n-3), arachidonic acid (AA, n-6), and dihomo-gamma-linolenic acid (DGLA, n-6)
Eicosanoid production is subsequently mediated by the PG-endoperoxide synthases PTGS1 and PTGS2, also known as cyclooxigenases COX1 and COX2
In summary these changes in the gene expression profile reflect the major alterations in lipid metabolism: increases in fatty acid synthesis, elongation and desaturation as well as increased cholesterol production, utilization and export from phagocytic macrophages in comparison to blood monocytes. Since the modulated release of inflammatory mediators is a key contribution of macrophages in the development and sustenance of atherosclerotic plaques we next focused on an in depth examination of plasmalogen lipidomics.
As already mentioned before plasmalogens regulate cell membrane fluidity and represent cellular stores for the precursor molecules of eicosanoid biosynthesis, mainly arachidonic acid
Clinically atherogenesis starts with increased deposition and modification of LDL in the subendothelium. Subsequently inflammatory cells such as monocytes are chemotactically attracted and migrate into the vessel wall. There they secrete reactive oxygen species (ROS) producing an environment with high levels of oxidative stress, that also affects the deposited LDL (reviewed in
The PE plasmalogen pattern in human blood cells has been reported by our group before
PE plasmalogen species in an
The correlation of HL-60 monocytic and granulocytic differentiation was strongly positive (Pearson r = 0.960, p<0.001). Furthermore we found a significant correlation between the changes in monocyte to macrophage differentiation and the regulation of plasmalogen species during monocytic differentiation of HL-60 cells (Pearson r = −0.522, p = 0.005). In contrast the changes found in HL-60 derived granulocytic cells were statistically different from monocyte-macrophage differentiation (Pearson r = −0.359, p = 0.066).
Since the plasmalogen pattern undergoes very specific changes during phagocytic differentiation we hypothesized that it could serve as a biomarker for predicting the monocyte to macrophage differentiation state. To this end a PLS/DA analysis was performed on the data obtained from primary human monocytes and macrophages. PLS is a mathematical method that allows to analyze the influence of a matrix of predictors on an outcome variable.
In this approach 4 different latent variables were computed and plotted graphically to evaluate their power to differentiate the differentiation stage. As shown in
While latent variable LV1 is able to discriminate between days 0/1 and days 4/5 of differentiation, LV4 can differentiate between days 4 and 5.
We also determined the most influential lipid species by calculating so called Variable Importance for the Projection (VIP) values as shown in
Chronic, systemic low-grade inflammatory diseases such as diabesity and the metabolic syndrome are characterized by a strong activation of tissue macrophages that chemotactically also attract additional monocytes from the circulating blood stream
In order to identify novel players influencing the regulation of cellular plasmalogens we performed partial correlation analysis of plasmalogen species with gene array data (
Plasmalogens are shown in the center (rectangles) and correlating transcripts are grouped around them (circles). Circle size represents the number of connected nodes, line width the strength of the correlation and color the tyoe of regulation (upregulated in red, downregulated in blue). Central key nodes can be identified that have a multitude of connections to other nodes.
Furthermore a range of proteins that have not yet been implicated in monocyte-macrophage differentiation were identified. DOCK1 is a Rac activator that regulates myoblast fusion
Further transcripts that were identified include ARHGAP26, a Rho GTPase involved in the formation of endocytotic vesicles, cell spreding and adhesion
Further studies are needed to clarify the role of these potential lipid related modulators of the differentiation process.
Major regulatory changes in the cellular transcriptional profile are mediated by transcription factors. We used a bioinformatics approach therefore affecting all target genes that contain a specific binding motif in their promoter region at the same time. This search was based on transcripts listed in the Gene Ontology category lipid metabolism. Transcription factors that are predicted to bind to promoters of these genes are shown in
Up-regulated genes | Down-regulated genes |
GZF1 | CHCH |
LXR | PPARG |
TATA | HIF2A |
IRF | ZBTB |
ZNF148 | ELF ETS |
ATF CREB | AP1 |
TFE | ZFX |
GKLF | FKLF |
USF | CNOT3 |
FOX | RNF96 |
MYCMAX | |
AHRARNTHIF | |
E2F |
Up- and down-regulated genes |
AP2 |
NFY |
P53 |
SP1SP4 |
EGR |
Motives that were enriched in the 1000 bp promoter region upstream of up-regulated lipid related genes are shown on the left, binding motives that were enriched in promoters of down regulated genes on the right and motifs that were associated with up- and down-regulated genes at the bottom. Abbreviations (as used in the Transfac database): AHRARNTHIF: aryl hydrocarbon receptor/aryl hydrocarbon receptor nuclear translocator/hypoxia inducible factor, AP1: activator protein 1, AP2: activating protein 2, ATF CREB: activating transcription factor/cAMP response element binding, CHCH: Churchill, CNOT3: CCR4-NOT transcription complex subunit 3, E2F: E2F transcription factor, EGR: early growth response, ELF ETS: E74-like factor/E26 transformation-specific, FKLF: fetal β-like globin gene-activating Krüppel-like factor, FOX: forkhead box, GKLF: gut-enriched Krüppel-like factor, GZF1: GDNF-inducible zinc finger protein 1, HIF2A: hypoxia-inducible factor 2a, IRF: interferon regulatory factor, LXR: liver X receptor, MYCMAX: Myc/Max transcriptional complex, NFY: nuclear factor Y, p53: phosphoprotein p53, PPARG: peroxisome proliferator-activated receptor-gamma, RNF96: RING finger protein 96, SP1SP4: specificity protein 1 and 4, TATA: TATA box, TFE: Transcription Factor E (Family), USF: upstream transcription factor, ZBTB: zinc finger and BTB (broad complex, tramtrack, and bric-à-brac) domain-containing, ZFX: zinc finger X-chromosomal protein, ZNF148: zinc finger protein 148.
During the differentiation process to mature macrophages monocytes undergo profound changes in their lipidomic and transcriptomic profile to prepare them for their phagocytic and inflammatory function.
Fatty acid synthesis pathways show a strong induction of transcripts involved in fatty acid desaturation and elongation. Lipid transport and lipoprotein related transcripts are differentially modulated partly in a pro-inflammatory way partly anti-inflammatory. On the one hand cholesterol biosynthesis is downregulated and export is strongly promoted. At the same time the cell prepares to facilitate the hydrolysis of triglycerides. In depth analysis was perfomed for plasmalogens. Plasmalogen biosynthesis pathways remained mostly unchanged on the mRNA level. This was reflected in lipidomic measurements showing constant total plasmalogen levels. Plasmalogens represent storage lipids for precursor acyl-residues in eicosanoid synthesis. Transcripts for the iso-enzymes involved in eicosanoid production were tightly regulated.
Lipidomic analysis of PE plasmalogens showed decreases in polyunsaturated species (PUFA) and increases in monounsaturated species (MUFA). These adaptions rendered monocyte derived macrophages more similar to mature granulocytes in their plasmalogen profile. This observation could be confirmed in an HL-60 model system for granulocytic and monocytic
A bioinformatic approach using partial correlation and molecular network reconstruction was able to identify novel potential hubs in the plasmalogen related transcriptional network, such as 70 kDa heatshock proteins, DOCK1, PDK4, GNPTAB, FAM126A, MNS1 and C12orf48. Further studies are currently underway to clarify the potential modulatory role of these proteins during differentiation. Transcription factor motif analysis revealed the involvement of several binding motifs with the remarkable involvement of motifs for NFY, P53, SP1 and EGR in up- and down- regulated genes. Therefore probably the interplay of multiple additional factors such as SREBP, determines the amount of transcription of lipid regulatory transcripts.
In summary we could for the first time show the specificity of the changes in the plasmalogen pattern during monocyte to macrophage differentiation. Furthermore using a bioinformatics approach we were able to predict novel regulatory hubs and transcription factors that might play central roles in the complex interplay of transcription-factors, transcripts and lipids during monocyte-macrophage differentiation.
(TIF)
(XLSX)
(DOCX)