Mapping macrophage polarization over the myocardial infarction time continuum

Mapping macrophage polarization over the myocardial infarction time continuum In response to myocardial infarction (MI), cardiac macrophages regulate inflammation and scar formation. We hypothesized that macrophages undergo polarization state changes over the MI time course and assessed macrophage polarization transcrip- tomic signatures over the first week of MI. C57BL/6 J male mice (3–6 months old) were subjected to permanent coronary artery ligation to induce MI, and macrophages were isolated from the infarct region at days 1, 3, and 7 post-MI. Day 0, no MI resident cardiac macrophages served as the negative MI control. Whole transcriptome analysis was performed using RNA- sequencing on n = 4 pooled sets for each time. Day 1 macrophages displayed a unique pro-inflammatory, extracellular matrix high low (ECM)-degrading signature. By flow cytometry, day 0 macrophages were largely F4/80 Ly6C resident macrophages, low high whereas day 1 macrophages were largely F4/80 Ly6C infiltrating monocytes. Day 3 macrophages exhibited increased proliferation and phagocytosis, and expression of genes related to mitochondrial function and oxidative phosphorylation, indicative of metabolic reprogramming. Day 7 macrophages displayed a pro-reparative signature enriched for genes involved in ECM remodeling and scar formation. By triple in situ hybridization, day 7 infarct macrophages in vivo expressed collagen I and periostin mRNA. Our results indicate macrophages show distinct gene expression profiles over the first week of MI, with metabolic reprogramming important for polarization. In addition to serving as indirect mediators of ECM remodeling, macrophages are a direct source of ECM components. Our study is the first to report the detailed changes in the macrophage transcriptome over the first week of MI. Keywords Myocardial infarction · Macrophage · Transcriptome · RNA-Seq · LV remodeling Introduction Myocardial infarction (MI) invokes a cardiac wound healing Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s0039 5-018-0686-x) contains response that involves early initiation of inflammation, fol - supplementary material, which is available to authorized users. lowed by robust scar formation in the infarct area. The mac- rophage is a key regulator of cardiac remodeling, providing * Merry L. Lindsey both strong pro-inflammatory signals early and reparative mllindsey@umc.edu cues later [13, 24, 25, 41, 44]. Macrophages naturally reside Department of Physiology and Biophysics, Mississippi in the healthy heart, largely derived from the embryonic yolk Center for Heart Research, University of Mississippi Medical sac, proliferating locally and overseeing normal tissue main- Center, 2500 North State St., Jackson, MS 39216-4505, USA tenance [22, 34]. Research Service, G.V. (Sonny) Montgomery Veterans high In response to MI, circulating pro-ina fl mmatory Ly6C Affairs Medical Center, Jackson, MS 39216, USA monocytes are rapidly recruited from bone marrow and The Roslin Institute, University of Edinburgh, Easter Bush, splenic reservoirs to the infarcted area by the CC chemokine Midlothian, Scotland, UK CCL2/MCP-1, and require CCR2 (the endogenous receptor Department of Biomedical Engineering, University for CCL2) for extravasation [14, 80]. Early post-MI, mono- of Virginia, Charlottesville, VA, USA cyte-derived macrophages promote inflammation through Department of Pharmacology and Toxicology, University of Mississippi Medical Center, Jackson, MS 39216, USA Vol.:(0123456789) 1 3 26 Page 2 of 18 Basic Research in Cardiology (2018) 113:26 release of pro-inflammatory cytokines such as interleukin Coronary artery ligation (IL)-1β, a key regulator of post-MI inflammation [70]. In the healthy heart and following MI, monocyte/mac- To produce permanent MI, mice underwent coronary artery rophage subpopulations are identified by a limited number of ligation surgery as described previously and according to cell surface markers. Resident cardiac macrophages express the Guidelines for Experimental Models of Ischemia and high levels of the myeloid marker CD11b, as well as canoni- Infarction [12, 28, 39, 47, 94]. Mice were anesthetized with cal macrophage markers including CD14, CD86, CX3CR1, 2% isoflurane, intubated, and ventilated. The left coronary F4/80, and MHC-II, and display an anti-inflammatory M2 artery was ligated with 8–0 suture, and MI was confirmed by phenotype [63]. Resident macrophages are characterized left ventricle (LV) blanching and ST-segment elevation on by high expression of F4/80, while infiltrating monocytes the EKG. Mice were administered buprenorphine (0.05 mg/ and monocyte-derived macrophages express F4/80 at lower kg body weight) immediately before surgery. high low levels [22, 62]. Following MI, F4/80 Ly6C resident macrophages are rapidly replaced by pro-inflammatory infil- low high high trating F4/80 Ly6C monocytes that are CCR2 [22]; Echocardiography and necropsy by post-MI day 7, the macrophage phenotype shifts to a pre- dominantly anti-inflammatory/pro-reparative M2 phenotype. LV physiology was determined by transthoracic echocar- diography (Vevo 2100, VisualSonics; Toronto, CA) as Specifically targeting macrophages to improve MI out- comes has proven both promising and challenging, as described before and according to the Guidelines for Meas- uring Cardiac Physiology in Mice [12, 28, 40, 47]. Mice therapeutic approaches successfully manipulating the mac- rophage depend on both temporal and spatial factors [13, were anesthetized under 1–2% isoflurane, and both long and short-axis images were obtained. Measurements were taken 18]. For example, the early inflammatory response is critical for wound healing, as too little or excess inflammation can on the terminal day and were averaged from three cardiac cycles for each mouse. Following imaging, the hearts were adversely affect remodeling; the same paradigm is true for the later reparative and fibrotic response [18]. While mac- removed and the left ventricle (LV) divided into remote and infarct (which included border zone) regions. Each region rophages have been extensively studied in steady state and aging hearts [10, 49, 61], as well as following pressure over- was separately weighed for infarct area estimation. The infarct sizes over the 3 MI time points had a coefficient of load [88], the full picture of the macrophage evolution over the first week after MI has not been developed. variation of 18%, indicating gene variation was not likely due to differences in infarct sizes. Accordingly, the objective of this study was to combine transcriptomics, flow cytometry, and cell physiology to provide a map of macrophage phenotypes in response to Isolation of LV infarct macrophages MI. We analyzed transcriptomic changes at days 1, 3, and 7 post-MI to reflect the early inflammatory, proliferative, LV macrophages were isolated from the infarct region by and maturation phases. We hypothesized that macrophages immunomagnetic separation as described previously [12, would undergo changes over the MI time course that range 28]. Excised LV tissue was rinsed and immediately minced from pro-inflammatory to reparative polarization. To our and digested by collagenase II (Worthington; Lakewood, NJ) knowledge, this is the first study to report in detail the full and DNase solution in Hanks buffered saline solution. After transcriptome changes that occur in cardiac macrophages digestion, a single-cell suspension was generated and filtered that mediate post-MI wound healing and remodeling. through a 30 µm pre-separation column. Cell suspensions were incubated at 4 °C with an anti-Ly6G-biotin antibody (Miltenyi Biotech, Bergisch Gladbach, Germany, 130-092- Methods 332) to remove neutrophils, followed by an anti-CD11b-bio- tin antibody (Miltenyi 130-049-601) for 15 min, followed by Animal use anti-biotin microbeads (Miltenyi 130-092-332) for 10 min. Cells conjugated to the antibody microbeads were separated All procedures involving animals were approved by the Insti- by magnetic columns (Miltenyi 130-042-201). tutional Animal Care and Use Committee at the University The average numbers of isolated macrophages ± SEM of Mississippi Medical Center. A total of 122 C57BL/6J and the coefficient of variation (CV) from each of adult (3–6 month old) male mice were used for this study. the individual pooled animals for each time point Mice within this age range are at similar physiological matu- 5 5 were as follows: day 0—1.62 × 10 ± 0 .1 8 × 10 , ration [26]. Groups were randomly assigned prior to the sur- 5 5 CV = 31%; day 1—2.61 × 10 ± 0.33 × 10 , CV = 43%; geries by one investigator (KYDP) and another investigator 6 6 d ay 3 — 1 . 7 5 × 1 0 ± 0 . 0 8 × 1 0 , CV = 34%; and day (YM) performed the majority of surgeries. 1 3 Basic Research in Cardiology (2018) 113:26 Page 3 of 18 26 5 5 7—7.64 × 10 ± 0.46 × 10 , CV = 17%. To evaluate whether Bioinformatic analyses pooling increased variability, we compared the CV of the day 7 MI macrophage pools with the CV of a previously Analyses tools available in the online resource Metabo- published day 7 MI macrophage RNA-seq experiment that analyst 3.0 (http://www.metab oanal yst.ca/) and GraphPad used macrophages isolated from individual mice [28]. The Prism were used for graphical and statistical analyses [92, CV for the day 7 pooled set was 17%, while the CV for the 93]. FPKM values were uploaded into Metaboanalyst, and individual mice was 48% (n = 5). There was no indication, one-way ANOVA with Tukey’s post hoc test was performed therefore, that pooling increased variability. to determine differentially expressed genes (defined as false An initial assessment of macrophages freshly iso- discovery rate (FDR) adjusted p < 0.05). For individual post- lated from the myocardium yielded RNA that was not MI days, differential expression was characterized by a fold of sufficient quality for RNA-sequencing (RNA-Seq). change threshold of > 2.0 or < 0.5 compared to day 0 no MI To determine an optimal culturing time, day 3 post-MI values and a p value of < 0.05 by unpaired two-tailed t test. + − 6 CD11b Ly6G macrophages (1.5 × 10 cells/well) were Markov clustering analysis was performed independently by plated in 6-well culture dishes for 2 or 20 h in RPMI 1640 two investigators (AJM and TCF) using Graphia Pro soft- medium supplemented with 0.1% FBS and 1% antibiotics. ware (Kajeka, Edinburgh, UK) using genes with a pairwise After incubation, non-adherent cells were washed off and Pearson correlation threshold of r > 0.95. Both investiga- the remaining adherent cells were used for transcriptomics tors obtained similar results. Enrichment analysis for dif- analysis by RNA-Seq. The 20 h culturing changed mac- ferentially expressed genes was performed using Enrichr rophage phenotype; Supplemental Fig. 1), and based on (http://amp.pharm .msm.edu/Enric hr/) gene ontology (GO) this result, 2 h was selected as the incubation period for biological processes and Ingenuity Pathway Analysis (Qia- the time course evaluation. gen) canonical pathways. For GO terms, the combined score (calculated from Z-score and p value) was reported. RNA‑Seq RT‑PCR validation To obtain high-quality RNA for sequencing (from day 0 hearts in particular), macrophages were pooled from A total of five genes (Arg1, Ifng, Il1b, Lgals3, and Tnf) n = 81 hearts to obtain four biological replicates of were evaluated by quantitative RT-PCR on the same mac- 1.5  ×  10 cells for each day post-MI. Whole transcrip- rophage RNA samples used for RNA-Seq and assessed for tome analysis was performed as described previously [28, correlation. RNA was reverse transcribed to cDNA using 47]. RNA was extracted using the Pure Link RNA Mini the High Capacity RNA-to-cDNA kit (Applied Biosystems Kit (Ambion, Foster City, CA) according to manufacturer 4387406). Gene expression was quantified using the Taqman instructions and assessed for quality control parameters Gene Expression Assay and primers for Arg1, Ifng, Il1b, of minimum concentration and size range. cDNA librar- Lgals3, and Tnf (Applied Biosystems). Values for the arrays ies were developed using the TruSeq Total Stranded RNA were normalized to the housekeeping gene Hprt1. with RiboZero Kit (Ambion), set-A, quantified with the Qubit System (Invitrogen, Carlsbad, CA), and assessed for quality and size with the Experion DNA 1K Chip (Bio- Flow cytometry Rad, Hercules, CA). The libraries (n = 12 pooled samples per library) were sequenced using the NextSeq 500 High LV tissue excised from day 0 and day 1 post-MI mice was Output Kit (300 cycles, paired end 100 bp) on the Illu- minced and digested with 600 U/ml collagenase II (Wor- mina NextSeq 500 platform (Illumina, San Diego, CA). thington, LS004177, Lot 47E17554B) and 60 U/ml DNase Sequenced reads (n = 30–50; Cloud Computing Platform), I in Hanks buffered saline solution and filtered through a and Fastq sequence files were used to align reads to the 30-µm separation filter to generate single-cell suspensions. reference genome USCS-GRCm38/mm10) using RNA-Seq Red blood cells were lysed (Red Blood Cell Lysis Solution, Alignment Application with STAR aligner. Fragments per Miltenyi 130-094-183) and non-specific interactions were kilobase of transcript per million mapped reads (FPKM) blocked with FcR Blocking Reagent (Miltenyi 130-092-575). values of reference genes and transcripts were generated Cells were stained with the following fluorophore-conju- using Cufflinks 2. Variability (coefficient of variation) gated antibody panels: CD45-FITC (Miltenyi 130-110-658), of pooled sets was compared to variability of a previous CD11b-APC-Vio770 (Miltenyi 130-109-288), F4/80- RNA-Seq experiment on day 7 post-MI macrophages from PerCP-Vio700 (Miltenyi 130-102-161), Ly6C-VioBlue individual mice [28]. (Miltenyi 130-111-921), and Ly6G-APC (Miltenyi 130- 107-914). Samples were quantified using the MACSQuant 1 3 26 Page 4 of 18 Basic Research in Cardiology (2018) 113:26 Analyzer 10 (Miltenyi). Cell populations were gated on live 1:1000), and Postn (418581, 1:1500); nuclei were stained singlets, with cells from monocyte-derived/macrophage line- with 4′,6-diamidino-2-phenylindole (DAPI). Probes were + + − age classified as CD45 CD11b Ly6G cells. conjugated to the following fluorophores (Perkin Elmer, Waltham, MA, TSA Plus): fluorescein (NEL741E001KT), In vivo phagocytosis assay Cy3 (NEL744E001KT), or Cy5 (NEL745E001KT). Images were acquired at 40× using the Mantra Quantitative Pathol- To evaluate macrophage phagocytosis, day 0 or day 3 post- ogy Imaging System (Perkin Elmer), and analyses were MI mice were injected with 100 µg of fluorescein-labeled performed using the cell phenotyping feature in the inForm Escherichia coli K-12 BioParticles (Molecular Probes, software (Perkin Elmer). Eugene, OR, V-6694) through the jugular vein [22]. After 2 h, cardiac macrophages were isolated, cultured for 2 h Statistics to remove unattached cells, and fixed with 100% ethanol. Nuclei were stained with DAPI. Images were acquired using All experiments were performed and analyzed in a blinded an Olympus IX81 microscope. Phagocytic macrophages design. Data are presented as mean ± SEM. Survival rate was (green fluorescence) were counted as a percentage of the analyzed by Kaplan–Meier survival analysis and compared total cells per field. by the log rank test. For echocardiography, comparisons were made using one-way ANOVA followed by Tukey’s In vivo proliferation assay post hoc test. Statistics and bioinformatics for the RNA- sequencing are described above. RNA-sequencing compari- To evaluate macrophage proliferation, day 0 or day 3 post- sons to quantitative RT-PCR were made by Pearson’s linear MI mice were injected with 1 mg BrdU (Sigma, St. Louis, regression analysis. Two group comparisons were analyzed MO, 11647229001) intraperitoneally 2 h before being killed by unpaired two-tailed t test. A value of p < 0.05 was con- [22]. Isolated infarct macrophages were adhered to slides, sidered statistically significant. fixed with 100% ethanol, permeabilized with Triton-X 100, and stained with anti-BrdU-FITC antibody (eBioscience, Waltham, MA, 11-5071-42, 1:20). Nuclei were stained with Results DAPI. Images were acquired using an Olympus IX81 micro- scope. Proliferating cells (green fluorescence) were counted Macrophages continually polarize over the post‑MI as a percentage of total cells per field. time continuum In vivo macrophage turnover Proof of successful MI To evaluate macrophage turnover, mice were injected with a Day 7 post-MI survival was 56% (Supplemental Fig. 2a), FITC-F4/80 antibody (Biolegend, San Diego, CA, 123107, consistent with others and our own past reports [11, 28, 43, 200 µg/kg) through the jugular vein at 24 h post-MI (day 75, 87, 96]. Of the day 7 mice that did not survive, 58% 1) and killed at day 3 post-MI. MI mice without injection (7/12) died from cardiac rupture, as assessed at autopsy. served as negative controls. Infarct macrophages were iso- Infarct areas (% LVI mass by total LV mass) were similar lated, and FITC + cells were quantified using a MACSQuant among days 1, 3, and 7 MI groups (Supplemental Fig. 2b). Analyzer 10. The data were analyzed using the MACSQuan- As expected, cardiac physiology was impaired after MI, with tify software and were presented as the percentage of FITC LV infarct wall thinning and dilatation evident beginning at cells to total macrophages. day 1 after MI (Supplemental Fig. 2c–e). Fractional short- ening decreased similarly across all post-MI groups (Sup- Triple in situ hybridization plemental Fig. 2f). Day 7 post-MI LV sections (n = 3) collected from the mid- Proof of cell isolation purity papillary region were fixed in 10% zinc-buffered formalin for 24 h at room temperature paraffin-embedded, and sectioned Supplemental Fig. 3 shows FPKM values plotted for mac- at 5 µm. In situ hybridization was performed with three probe rophage specific markers (Cd14, Cd68, Emr1, Fcgr3, Itgam, sets per section using the RNAscope Multiplex Fluorescent and Lgals3) and compared to non-macrophage cell-specific Reagent Kit v2 (Advanced Cell Diagnostics, Newark, CA). markers for endothelial, fibroblast, lymphocyte, myocyte, Samples were hybridized using probes (all from Advanced and neutrophil cell types. FPKM values for macrophage Cell Diagnostics) specific for Acta2 (319531, 1:1000), Ccr2 specific markers averaged 455, while for non-macrophage (433271, 1:1000), Col1a1 (319371, 1:1500), Emr1 (317961, specific markers, the average FPKM was ~ 3. These results 1 3 Basic Research in Cardiology (2018) 113:26 Page 5 of 18 26 indicate that our macrophage isolations were ultra-pure and (1707) upregulated and 9% (1547) downregulated, and day free of contamination from other cell types. 7 had 5% (899) upregulated and 7% (1189) downregulated compared to day 0 macrophages (Fig. 1e). Fold change anal- Differentially expressed genes ysis of commonly used M1 and M2 markers is displayed in Supplemental Fig. 4. From the 23,847 genes in the dataset, 4064 were removed RT-PCR validation of FPKM values was performed for from analysis as FPKM values were 0 for all biological rep- Il1b, Arg1, Lgals3, Tnf, and Ifng on the same samples used licates in all four groups and 2938 were removed as there for RNA-Seq (Supplemental Fig. 5). These genes were cho- was < 3 replicates with values > 0 for any one group. Of sen based on their strong association with macrophages. the remaining 16,845 genes, 150 had duplicate or tripli- Positive correlations for Il1b (r = 0.99, p < 0.0001), Arg1 cate measurements that were removed. This left a total of (r = 0.95, p < 0.0001), and Lgals3 (r = 0.51, p = 0.02) were 16,695 unique transcripts for analysis (Fig. 1a; Supplemental observed. Tnf and Ifng showed low agreement due to low Table 1). By principal component analysis, each of the days detection by RT-PCR, either due to sensitivity issues or separated out into its own pattern, indicating that each day primer design. had a unique gene expression profile (Fig.  1b). By one-way ANOVA with Tukey’s post hoc test, 8109 genes differed Pattern clustering among groups (with FDR adjusted p < 0.05; Fig. 1c, heat map Fig. 1d). By fold change analysis using cut-off points Gene expression patterns across the post-MI remodeling of two for fold change and p < 0.05 for p value by unpaired spectrum were also analyzed by Markov pattern cluster- two-tailed t test, day 1 had 6% of genes (1019) significantly ing. A total of three major clusters distinguished the time upregulated and 12% (2043) downregulated; day 3 had 10% points (Fig. 2a), with representative genes from each cluster Days Post-MI Fold Change (a) (d) (b) Principal Component Analysis (e () 01 37 Analysi Analysis s 23,847 2043 1019 transcripts Day 7 1 evaluated 1547 1707 Day 0 1189 899 7,152 Day 3 -2000 02000 transcripts Upregulated Upregulated removed -3 Day 1 Day 1 16,695 transcripts -100 0100 analyzed PC1 (27.4%) 879 293 (c) One-way ANOVA Day 7 Day 3 Cfh 12 Ephx1 8,109 Cd209f Cd9 Fn1 transcripts 10 Pgk1 Downregulated DE by Day 1 ANOVA (FDR<0.05) 2 403 Pathway 0 5000 10000 15000 Analysis Day 7 Day 3 Genes Fig. 1 Distinct time-dependent gene expression profiles in post-MI were distinct from day 0 and day 1 and showed overlap. c One-way macrophages. a Of 23,847 genes sequenced, 7152 did not meet qual- ANOVA plot showing significant genes in red and d heat map of all ity control standards and were removed. Of the remaining 16,695 differentially expressed genes. e Fold change analysis of differentially genes, 8109 were differentially expressed (DE) by one-way ANOVA expressed genes at each day post-MI (fold change threshold of 2, (FDR adjusted p value < 0.05). b Macrophages from different post-MI FDR adjusted p value < 0.05) and Venn diagrams of upregulated and days analyzed by principal component analysis. Day 1 macrophages downregulated genes showing distinction and overlap in gene expres- were most distinct from the other times. Day 3 and 7 macrophages sion among the times 1 3 -log10(p) PC2 (19.1%) -100 0 100 Days post-MI 26 Page 6 of 18 Basic Research in Cardiology (2018) 113:26 Cluster 1Cluster 2Cluster 3 Translation elongation Neutrophil degranulation ECM organization (a) (b) Il1a Il1b Nlrp3 Cytoplasmic translation Collagen catabolic process Inflammatory response Translation fidelity Cell migration Canonical glycolysis Ptgs2 Plastid translation Regulation of small GTPase signaling Positive regulation of NF-κB Nos2 Cd80 Mitochondrial translation Positive regulation of PI3K signaling Cellular response to hypoxia Gapd p h Hif1 Hif1a a 05 0 50 01 10 00 0 01 0 10 00 0 20 2000 0 30 300 0 0 50 50 10 100 0 Combined Score Combined Score Combined Score Ikbkb Fold Change Fold Change Fold Change 4 #$ 15 * * #$ * 3 3 3 5 # 01 37 01 37 01 37 Days Days post post-M MII Days Days post post-MI MI Days Days pos post- t MI MI Sparc Eef2 Col1a1 Cl Col3 31 a1 Rpl3, 5, 6, 7 Eln Rps2, 7, 8, 18 Lox Postn Fig. 2 Gene expression pattern clustering. a Markov clustering analy- 1 (yellow) contained 565 genes increased at day 1 and was enriched sis generated three distinct major clusters representing different post- for pro-inflammatory processes. Cluster 2 genes (blue, 1965 genes) MI gene expression patterns. Each node represents a single gene, and were increased at day 3 and 7 and were enriched for translation pro- genes within the same cluster (color) show similar gene expression cesses. Cluster 3 (green, 1222 genes) were increased at day 7 and patterns. Genes representing each cluster are highlighted in red. b enriched for ECM processes. *p < 0.05 versus day 0, p < 0.05 versus Enrichment analysis and expression patterns for each cluster. Cluster day 1, p < 0.05 versus day 3 highlighted. Cluster 1 (yellow, 565 genes) showed an overall upregulation of some, but not all ECM genes and indirect expression pattern of mRNAs being increased only at day 1 regulation of ECM processes. after MI compared to day 0 (Fig. 2b). GO enrichment for this cluster revealed that pro-inflammation (neutrophil degranu- Day 1 MI macrophages display a pro‑inflammatory lation, p = 3.2E−11; inflammator y response, p = 7.5E−7; monocyte‑derived signature canonical glycolysis, p = 5.3E−7; positive regulation of NF-κB activity, p = 0.0018; cellular response to hypoxia, Genes differentially expressed at day 1 post-MI are displayed p = 0.00082) was the major process represented by these by volcano plot (Fig. 3a), with representative genes from genes, suggesting that day 1 macrophages rapidly upregu- each GO process highlighted. The representative genes were late inflammatory and glycolytic processes which are then the top five ranked up- and down-regulated genes based on rapidly turned off. The neutrophil degranulation term is term fold change (≥ 2) and p value at each time point. Notable overlap and does not imply neutrophil contamination, as we upregulated GO processes included inflammatory response saw essentially no expression of neutrophil-specific mark - (p = 0.007), cytokine-mediated signaling (p = 0.002), canoni- ers, and genes in our analysis represented by this term are cal glycolysis (p = 0.002), ECM disassembly (p = 0.002), and known to be expressed in macrophages (e.g., Mmp8/9, Mif, cellular response to hypoxia (p = 0.02; Fig. 3a). Downregu- and Lgals3). Cluster 2 (blue, 1965 genes) showed a pattern lated biological processes at day 1 included ECM organiza- of increasing at day 3 and day 7, with a peak at day 3. GO tion (p = 9.3E−7) and cell–matrix adhesion (p = 0.007). processes related to mRNA translation were prominent in To determine macrophage subpopulation heterogene- this cluster. Cluster 3 (green, 1222 genes) showed a pattern ity within the 1 day post-MI time, we performed multi- of being decreased at days 1 and 3, and either increased or marker flow cytometry. Myeloid cells (neutrophils and decreased at day 7. The top two GO processes were extra- macrophages) were elevated (27 ± 4% of total cells vs. cellular matrix (ECM) organization (p = 1.06E−16) and 4 ± 1% at day 0, a 6.2-fold increase; Fig.  3b). We fur- collagen catabolic process (p = 3.3E−7), reflecting direct ther gated on the myeloid cell population to differentiate 1 3 Basic Research in Cardiology (2018) 113:26 Page 7 of 18 26 Cd9 Upregulated Do Downregu nreg late lated d (a) Neutrophil degranulation Extracellular matrix organization Gapdh Actr3 Hif1a GTPasesignaling pathway Inflammatory response Itga11 Eln Ccr2 Cytokine-mediated signaling Mitotic chromosome condensation Il1b Col1a2 Canonical glycolysis Col3a1 Cardiac muscle contraction 2 Mmp8 Extracellular matrix disassembly Cell-matrix adhesion -10-50 5102 0 040600 10 20 30 4 log2FC Combined Score Combined Score (b) Myeloid Cells Neutrophils Neutrophils Day 0 Resident 80 Macrophage Monocytes/ Macrophage 50 40 Myeloid Monocyte cells 01 01 Resident Day 1 Macrophages Monocytes 01 01 CD11b CD11b Ly6C Days post-MI Days post-MI (c) Top 5 Genes Ranked by fold change then p-value Top 5 Genes Ranked by p value then fold change Expression in Expression in Gene p FC Macrophage Macrophage Role Role othe otherc r cardia ardiacc ccells ells Gene FC p Macrophage Role other cardia cardiacc ccells ells Pcdha4 1.2E7 0.002 Cell adhesion unknown Tetraspanin; inhibits M1 Cd9 2E-8 4.5 phenotypeubiquitous Pro-inflammatory Inhibition of Orm1 834 0.026 cytokine release ubiquitous Inhibition of Tnip1 5E-7 3.3 TNFα/LPS/EGF signalingubiquitous Pappa 22 0.016 cholesterol efflux myocytes Uptake of TG-rich Apobr 6E-7 35 3.5 lipoprotein lipoproteins( s (ApoB) endothelial Ih Inhibiti ibitif on of Slfn4 19 0.005 monocytopoiesis unknown Bcl2l11 1E-63.4 apoptosis myocytes M1 marker; induced pH-sensing; M2 Inhba 18 0.001 by glycolysis myocytes Gpr132 1E-62.3 polarization lymphocytes Fig. 3 Day 1 post-MI macrophages showed a pro-inflammatory pro- myeloid cells, neutrophils, and monocytes significantly increased in file. a Volcano plot with all values normalized to day 0 no MI con- the infarct region, while resident macrophages decreased. c Top five trols and representative genes highlighted (left). Enrichment analysis upregulated day 1 post-MI genes ranked by fold change (left) and p of upregulated and downregulated genes (right). b Flow cytometry value (right). *p < 0.05 versus day 0 analysis of macrophage cell surface markers. At day 1 post-MI, total monocytes and macrophages from neutrophils based on shown indicating that infiltrating monocytes replace resi- Ly6G expression. Ly6G + neutrophils were significantly dent macrophages early after MI [13, 14, 22, 54, 62, 63]. increased in the infarct region (54 ± 4% of the total To define the day 1 post-MI macrophage, we ranked the myeloid cells, vs. 5 ± 1% at day 0, 10.8-fold increase). top five uniquely upregulated genes by fold change followed Gating on CD45 + CD11b + Ly6G- cells, F4/80 and by p value, and the top five ranked by p value followed by Ly6C further divided this population into resident mac- fold change (Fig. 3c). All of top ten ranked genes had previ- high low rophages (F4/80 Ly6C ) and infiltrating monocytes ously been associated with macrophages: Pcdha4 [15], Orm1 low high (F4/80 Ly6C ). Resident macrophages were signifi- [38], Pappa [78], Slfn4 [82], Inhba [37], Cd9 [74], Tnip1 cantly decreased in the infarct (14 ± 1% of at day 1 vs. [95], Apobr [79], Bcl2l11 (Bim) [30], and Gpr132 [4]. Day 78 ± 3% at day 0, 5.7-fold decrease) while monocytes were 1 macrophages displayed a unique signaling profile by prin- significantly increased (74 ± 1% at day 1 vs. 12 ± 2% at cipal component analysis compared to the other three time day 0). These results are consistent with what others have points (Fig. 4a), with expression of genes associated with 1 3 -log10(p) CD45 Ly6G F4/80 % of mono/mac % of total cells % of myeloid cells 26 Page 8 of 18 Basic Research in Cardiology (2018) 113:26 p<0.05 genes Principal Component Analysis (a) Day 0Day 1Day 3Day 7 Day 0 Mapk7 Day 1 Ilr2 D3 Day 3 Il1rap Mapk6 Day 7 Mapk8 Il1rn Stat5a Stat5b Stat5b Il1rl2 Nfkb2 Tnfrsf1b Nfkb1 -400 -200 0 200 400 Socs2 PC PC1 1( (68.4 68 4% %) ) Il1rapl1 Non-Significant (p>0.05) Stat4 Mapk9 Day 0Day 1Day 3Day 7 (b) Socs6 Stat1 Stat2 Mapk3 Il1r Il1r1 1 Tf Tgfbr1 b1 Socs1 Il4ra Socs3 Mapk14 Stat3 Mapk11 Tgfbr2 Akt3 Stat6 Tnfrsf1a Map pk4 Mapk12 Mapk12 Socs5 Mapk13 Socs7 Akt2 Socs4 Mapk10 Il1rl1 Akt1 Akt1s1 Tgfbr3 Mapk1 Fig. 4 Post-MI macrophage signaling profiles. a Principal component to days 0, 3, and 7. b Heat maps grouped by significance (left, non- analysis of genes involved in ubiquitous signaling pathways indicates significant genes (p > 0.05); right, significant genes (p < 0.05) by that day 1 macrophages display a unique signaling profile compared unpaired two-tailed t test) IL-1, TNF, NF-κB, MAPK, STAT5, and SOCS2 signaling (Fig. 5b). In line with increased phagocytic capacity, IPA pathways (Fig. 4b). analysis indicated that phagosome maturation was an upreg- ulated pathway at day 3 (20/148 genes; p = 0.003). Likewise, Day 3 MI macrophages are phagocytic, proliferative, the phagosome maturation GO pathway was increased at and display a metabolic reprogramming signature day 7 post-MI (11/148 genes; p = 0.02), albeit to a lesser extent than day 3. Phagosome formation, but not matura- Genes differentially expressed at day 3 are displayed by tion, was an upregulated pathway at day 1 (15/130 genes; volcano plot, with representative genes highlighted in yel- p = 0.0003). While day 3 proliferation was not significantly low (Fig. 5a). The major upregulated GO processes were different from day 0 (Fig.  5b), resident (day 0) macrophages related to mitochondrial function, including mitochon- have been shown to be proliferative [22]. Further, GO pro- drial translation termination (p = 8.4E−10) and elongation cesses such as DNA replication initiation (p = 0.002), DNA (p = 8.4E−10), electron transport (p = 4.4E−11), respiratory replication (p = 0.01), and mitotic cell cycle (p = 0.01) were chain complex I assembly (p = 8.4E−10), cristae formation upregulated, indicating a proliferative signature at day 3. (p = 2.9E−5), and mitochondrial ATP synthesis-coupled pro- Genes associated with phagocytosis and proliferation are ton transport (p = 7.1E−5). Downregulated GO processes displayed in Supplemental Fig. 7. Macrophage turnover was included ECM organization (p = 2.4E−7), and cell migration assessed by F4/80 antibody injection at day 1 post-MI; only (p = 0.0002). By IPA analysis, oxidative phosphorylation 5.2 ± 0.4% of macrophages at day 3 post-MI stained positive (42/109 genes; p = 2.5E−21) and mitochondrial dysfunc- for the F4/80 antibody (Fig. 5c), indicating rapid turnover as tion (51/171 genes; p = 8.0E−20) were the top upregulated reported by others [36]. canonical pathways, while leukocyte extravasation (38/211 To define the day 3 post-MI macrophage, we ranked the genes; p = 2.9E−9) and agranulocyte adhesion and diapede- top five uniquely upregulated genes by fold change followed sis (35/191 genes; p = 7.7E−9) were the major downregu- by p value, and the top five ranked by p value followed by lated pathways (Supplemental Fig. 6). fold change (Fig. 5d). All of the top ten ranked genes had We assessed in vivo macrophage phagocytosis, prolifera- previously been associated with macrophages: Klra22 [59], tion, and turnover. Macrophages from post-MI day 3 showed Ffar4 [91], Tdgf1 [72], Sarm1 [19], Il24 [73], Dek [50], significantly increased phagocytosis compared to day 0 Rab18 [21], Vdac1 [7], Dnajc15 [55], and Tmsb4x [83]. 1 3 PC2 (21 %) -400 -200 02 0 00 0 400 Basic Research in Cardiology (2018) 113:26 Page 9 of 18 26 UpregulatedDownregulated (a) Mitochondrial translation termination Extracellular matrix organization Mrpl34 Ccl7 Mitochondrial translation elongation Muscle contraction Ccl11 Sdhb Mitochondrial electron transport Cell migration Col4a1 Col4a2 Cell-cell signaling Atp5l Mitochondrial complex I assembly Sparc Mrps10 Neutrophil degranulation Muscle filament sliding -1 10 0-50 50 5 5 10 10 0 0 20 20 40 40 60 60 80 80 0 0 10 10 20 20 30 30 40 40 50 50 log2FC Combined ScoreCombined Score (b) No injection Day 3 Phagocytosis (c) Day 0 0.28±0.11 100 μm F4/80 F4/80Ab F4/80Ab injectio injection n Proliferatio Proliferation n 20 20 Day 0 Day 3 p=0.19 5.23±0.43 0 0 100 μm F4/80 Days post-MI (d) Top 5 Genes Ranked by fold change then p-value Top 5 Genes Ranked by p value then fold change Expression in Expression in Gene FC pRole other cardiac cells Gene pFCRole other cardiac cells MHC-binding lectin-like Secreted Klra22 7.2E+400.0002 receptor Natural killer cells Dek 6E-7 2.1 chemotacticfactor Ubiquitous Omega-3 FA receptor; Fibroblasts, Rab18 6E-6 2.3phagocytosisUbiquitous Ffar430366 0.035 anti-inflammatory myocytes Calcium signaling; Tdgf1 1770 0.027TGF-betasignaling unknown Vdac11E-52.5 apoptosisUbiquitous Regulation of CCL5 Transcriptional Sarm1458 0.047 production Ubiquitous Dnajc15 3E-5 2.6 response to IFNγUbiquitous IL-10family; Inflammation Il24 2010.036 inflammatory response Th2 cells Tmsb4x 4E-5 2.1 resolution Ubiquitous Fig. 5 Day 3 post-MI macrophages showed a phagocytic, prolifera- Phagocytic capacity significantly increased at day 3 post-MI, whereas tive, and metabolic reprogramming profile. a Volcano plot with rep- proliferation was similar to day 0. c In vivo turnover. At day 3 post- resentative genes highlighted and enrichment analysis of upregulated MI, ~ 5% of macrophages were remaining from day 1. d Top five and downregulated genes. All values are normalized to day 0 no MI upregulated day1 genes ranked by fold change (left) and p value controls and representative genes are highlighted. b Representative (right). *p < 0.05 versus day 0 images of phagocytic (top) and proliferative (bottom) macrophages. of ECM genes downregulated at days 1 and 3 and either Day 7 MI macrophages display a pro‑reparative upregulated or downregulated at day 7 compared to day 0 signature (Fig. 6b; Supplemental Table 2). To define the day 7 post-MI macrophage, we ranked the Genes differentially expressed at day 7 post-MI are displayed top five uniquely upregulated genes by fold change followed by volcano plot, with representative genes highlighted in by p value, and the top five ranked by p value followed by yellow (Fig. 6a). The major upregulated GO processes were fold change (Fig. 6c). All of the top ten ranked genes had related to ECM remodeling (Supplemental Table 2), includ- previously been associated with macrophages: Pcdha7 [15], ing ECM disassembly (p = 0.0004) and collagen fibril organ- Klra23 [59], Tac4 [3], Phgk1 [16], Slc3a1 [29], Odc1 [20], ization (p = 0.01), and inflammatory response (p = 0.004) Spsb1 [56], Crem [46], Sptlc2 [5], and Abca1 [77]. was the major downregulated GO process. By IPA analysis, Of note, Col1a1, Col3a1, and Postn, which contrib- inhibition of matrix metalloproteinases (MMPs) was a major ute to post-MI scar formation and ECM stiffness, were upregulated canonical pathway (7/39 genes, p = 0.0004), all elevated in day 7 macrophages, while Col4a4 and while agranulocyte adhesion and diapedesis was a major Lama2, which contribute to basement membrane forma- downregulated pathway (34/192 genes, p = 4.6E−12; Sup- tion, and Mmp15, were decreased. Using triple in  situ plemental Fig. 6). Clustering analysis identified a number 1 3 -log10(p) 0 Phagocytosis Proliferation o % % of total cells s % of total cells Count Count 26 Page 10 of 18 Basic Research in Cardiology (2018) 113:26 Upregulated Downregulated Ccl2 (a) Inflammatory Inflammatory respons response e Et Extracell llulM lar Mati trix Di Disassembl bly Col1a1 Muscle filament sliding Hyaluronan catabolic process Tnf Postn Cell surface receptor signaling Ccl3 Platelet degranulation Lox Cxcr2 Cardiac muscle contraction Synapse Assembly Positive regulation of ERK 1 and 2 Collagen Fibril Organization -10-50 5100 20 40 60 05 10 15 20 25 log2FC Combined ScoreCombined Score (b) Col1a1 Col3a1 Postn Upregulated # # $ $ * $ $ 6 6 60 60 # # * * 150 * * 200 * * 4 40 20 50 0 0 1 1 3 3 7 7 0 0 1 1 3 3 7 7 0 0 1 1 3 3 7 7 0 0 1 1 3 3 7 7 Downregulated Col4a4 Lama2 Mmp15 1.5 1.0 10 1.0 ** 0 05 .5 2 2 5 0.5 * * * * * * * * * * * * * * * * * * 0 0 0 13 71 0 37 00 1 3 7 1 3 7 Days post-MI Days post-MIDays post-MI Days post-MI (c) To Top5 p5 Gene GenesR sRanked ankedb by y Fol Fold dC Change hange T5 Top 5 GR Genes Rankd kedb by p vallue Expression in Expression in Gene pFCMacrophageRole other cardiac cells Gene FC pMacrophage Role other cardiac cells cell adhesion, cell- Ornithine metabolism; Pcdha7 42888 0.038 cell recognition unknown Odc1 1E-5 3.1 M2 polarization ubiquitous Negative regulation of MHC-binding Spsb13E-5 3.4 iNOS unknown Klra23 18750.024 receptor Natural killer cells cAMPsignaling; M2 Inhibition of cytokine B cells, dendritic Crem 7E-5 2.3 polarizationmyocytes Tac4 79 0.012 signaling cells inflammasome Glycolytic Sptlc2 Sptlc2 9E 9E-5 5 21 2.1 activation activation myocytes myocytes Ph Phgk1 k1 32 32 0 0.043 043 metb taboli lism myocyttes Cholesterol efflux, Mitochondrial Abca11E-4 2.9 anti-inflammatory unknown Slc3a1 25 0.005 respiration unknown Fig. 6 Day 7 post-MI macrophages showed a pro-reparative profile. genes upregulated at day 7 compared to day 1 and 3 post-MI (top) a Volcano plot with representative genes highlighted and enrichment or remained downregulated at day 7 (bottom). c Top five upregulated analysis of upregulated and downregulated genes. All values are nor- day 7 genes ranked by fold change (left) and p value (right). *p < 0.05 # $ malized to day 0 no MI controls and representative genes are high- versus day 0, p < 0.05 versus day 1, p < 0.05 versus day 3 lighted. b ECM genes that clustered together. Fold change values for hybridization, we assessed the number of Emr1 + cells Defining macrophage polarization phenotypes (i.e., macrophages), Acta2 + cells (i.e., fibroblasts), and Em r1 + C cr 2 + c el ls (in f i ltrat in g mo no cy tes a nd m ac- In addition to the prototypical M1 and M2 markers, we rophages) expressing Col1a1 and Postn mRNA in the investigated the most prominent genes uniquely upregulated infarct region of LV tissue from day 7 post-MI mice at each post-MI day (Figs. 3c, 5d, 6c). Genes expressed only (Fig. 7a–c). Macrophages in the infarct region heterogene- at day 0 (i.e., no or very low expression at all post-MI days) ously expressed both Col1a1 and Postn mRNA (Fig. 7d). were ranked to assess day 0 (resident cardiac macrophage) Of the Emr1 + macrophages, about half expressed Col1a1 markers. Day 0 genes were ranked by FPKM then ANOVA and about a quarter expressed Postn. ECM gene expres- p value or by ANOVA p value then FPKM (Supplemental sion was found only in Ccr2 + cells, which have been Table 3). The top ranked day 0 genes by FPKM included shown to drive post-MI inflammation and impair ECM Atf3 [23], Cbr2, Folr2 [42, 58], Actr3 [90], and Cd81 [76]. formation [45]. The majority of Acta2 + cells expressed The top ranked day 0 genes by p value included Cfh [2], Col1a1 (70%), while 25% also expressed Postn. Lilra5 [48], Cd209f [69], Cmah [60], and Tln2 [66]. 1 3 Fold d Change Fold Change e -log10(p) FP PKM FPKM Basic Research in Cardiology (2018) 113:26 Page 11 of 18 26 (a) (b) PostnCol1a1Emr1DAPI PostnCol1a1Acta2DAPI (c) (d) Other PostnCcr2Emr1DAPI 16% Emr1+ 47 47± ±7% 7% Acta2+ 36±1% Ccr2+ Ccr2- Col1a1 Col1a1- Col1a1+ Col1a1+ Cl Col1 11 a1- Col1a1 Col1a1- Postn- Postn+ Postn- Postn- 30±8% 25±12% 25±10% 28±4% Ccr2- Ccr2- Col1a1+ Col1a1+ Col1a1+ Col1a1+ Col1a1+ Col1a1+ Postn- Postn+ Postn- 45±10% 26±6% 21±12% Fig. 7 Macrophage expression of Col1a1 and Postn in LV infarct Col1a1 (yellow, Cy3) and Postn (magenta, FITC). Emr1 and Ccr2 region. Numbers of cells expressing Postn and Col1a1 mRNA in LV positive cells (c; yellow, Cy3) were assessed for expression of Postn infarct tissue at day 7 post-MI were determined by in  situ hybridi- (magenta, FITC). d Pie charts displaying proportion of LV infarct zation. Both Emr1 (a; red, Cy5, macrophages) and Acta2 (b; green, Emr1 + (macrophages) and Acta2 + (fibroblasts) cells expressing Cy5, fibroblasts) positive cells were assessed for expression of Col1a1 and Postn. Nuclei are stained with DAPI (cyan) As cell specificity is a good criterion for a marker, we (Cbr2, Folr2, Cfh, Lilra5, Cd209, Pcdha4, Slfn4, Tdgf1, compared the ranked genes to the literature to determine Pcdha7, Slc3a1, Spsb1, and Abca1). Based on these criteria, whether genes were distinctly expressed in macrophages Table 1 lists candidate markers that may uniquely identify relative to other cardiac MI-relevant cell types (myocytes, cardiac macrophages from each time after MI, including endothelial cells, fibroblasts, neutrophils, and lymphocytes). directed analysis of known M1, M2, and macrophage mark- While several genes are ubiquitously expressed across cell ers and unbiased evaluation of the top ranked genes for each types (Atf3, Cd81, Cmah, Tln2, Orm1, Cd9, Tnip1, Sarm1, time examined. Dek, Dnajc15, Tmsb4x, Rab18, Vdac1, and Odc1), there We assessed expression of genes involved in circadian were genes with potential macrophage restricted expression rhythm regulation, as time of surgery and cell isolation all Table 1 Candidate post-MI Day 0 Day 1 Day 3 Day 7 macrophage markers a a Il1b, Nos2, Arg1, Chi3l3, Informed Ccl17, Ifng, Il12a, Lgals3, Vegfa Cd14, Sparc, a a a a Cd14 Retnla, , Lgals3, VegfaLox, Lgals3 , Cd163 Vegfa Unbiased Cbr2, Folr2, Pcdha4, Slfn4, Apobr, Gpr132 Klra22, Tdgf1, Il24 Spsb1, Abca1, Cfh, Pcdha7, Lilra5, Klra23, Cd209f Tac4, Slc3a1 Unbiased candidates are genes based on our ranking system; informed candidates are genes previously used as macrophage markers Marker for more than one time point 1 3 26 Page 12 of 18 Basic Research in Cardiology (2018) 113:26 were performed between 8 a.m. and 12 noon (Supplemen- Macrophage metabolism is a reflection of and also a tal Fig. 8). No differences in Clock, Per2, or Nr1d1 were contributor to polarization status, as pro-inflammatory M1 observed, while Nr1d2, Nfil3, Arnt1, and Cry1 were dif- macrophages rely on glycolysis, while reparative M2 mac- ferentially expressed after MI. These results indicate that rophages use oxidative phosphorylation [33, 57]. Indeed, MI may compromise circadian rhythm regulation in mac- our results show that by day 3 post-MI, genes related to rophages, which has been implicated in release of cytokines mitochondrial ATP generation and oxidative phosphoryla- and other macrophage functions [84]. tion were upregulated, indicating metabolic reprogramming over the post-MI remodeling continuum. While studies have demonstrated that metabolism influences macrophage polarization, the role of metabolism in post-MI macrophage Discussion polarization has not been extensively evaluated [16, 33]. Our results indicate the metabolic shift of the day 3 macrophage The goal of this study was to map the continuum of changes may be an indicator of wound repair status. that occur in cardiac macrophages over the first week of MI. At day 3 post-MI, macrophages downregulate many of The key findings were: (1) cardiac macrophages undergo the inflammatory genes elevated at day 1, including Il1b, continual and distinct transcriptomic changes at post-MI while upregulating others, including Il12a, Pf4, and Il24. days 1, 3, and 7; and (2) day 1 macrophages have a pro- In addition, day 3 macrophages showed increased phago- inflammatory profile, day 3 macrophages have a phagocytic, cytic capacity and a return in proliferation. Phagocytosis proliferative, and metabolic reprogramming profile, and day of necrotic and apoptotic cells is a critical role of post-MI 7 macrophages have a reparative signature that includes macrophages and is required for the transition from a pro- to expression of extracellular matrix remodeling genes that anti-inflammatory environment and polarization towards a contribute to scar formation (e.g., collagen 1a1 and peri- reparative phenotype [18, 32, 65]. Resident cardiac mac- ostin). While macrophages in the post-MI heart have been rophages have a basal phagocytic rate, while day 1 MI mac- characterized in terms of cell surface markers, phenotypes, rophages have increased phagocytic capacity [22, 54]. Our and origins, the full transcriptome has not been mapped in results indicate that phagocytic capacity is elevated at day 3, detail [9, 12–14, 18, 22]. which coincides with downregulation of many of the inflam- Macrophages underwent distinct gene expression changes matory genes that were elevated at day 1. Following MI, reflecting shifts in phenotype over the first week of post- macrophage phagocytosis occurs in two sequential steps: ini- MI LV remodeling. The fact that day 7 normalized values tial phagocytosis of necrotic myocytes, followed by efferocy - showed distinction in expression from day 3 normalized val- tosis of apoptotic neutrophils [68]. The increased phagocytic ues (e.g., Col1a1, Col3a1, and Postn genes) indicates there capacity that we observed at day 3 post-MI may represent is still a continuum of changes occurring, at a slower kinetic efferocytosis of apoptotic neutrophils, whose numbers in the rate than day 0 through days 1 and 3. Principal component infarct concomitantly begin to decline after day 3. As resi- analysis indicates that day 1 macrophages are markedly dif- dent cardiac macrophages are proliferative, the similarity in ferent in phenotype than all other time points, whereas day proliferation rates between day 0 and day 3 macrophages 3 and 7 macrophages are closer to a day 0 phenotype, indi- indicates that by day 3 proliferation of macrophages is re- cating the transition to a new homeostatic-like phenotype. initiated, consistent with a previous report [22]. Consistent with the literature, our results indicate that At day 7 post-MI, macrophages exhibited a reparative the early day 1 post-MI macrophage shows a strong pro- phenotype, indicated by upregulation of ECM organization inflammatory, matrix-degrading phenotype [22]. In addition genes. Interestingly, genes typically considered fibroblast- to pro-inflammation, enrichment analysis indicated that gly - specific were upregulated in day 7 macrophages, including colysis and cellular response to hypoxia were upregulated Col1a1 and Postn. Recent studies on aging hearts have char- in the day 1 macrophage, suggesting an adaptation to the acterized a myeloid-derived cardiac fibroblast population, hypoxic environment of the early infarct. Hypoxia induces which is derived from M2a macrophages [81]. Our study activation of the hypoxia-inducible factor (HIF)-1α pathway, suggests that post-MI macrophages may assume a fibroblast- which turns on pro-inflammatory gene expression, as well as like phenotype, which remains to be fully investigated. metabolic reprogramming towards glycolysis [1, 9]. Hif1a While macrophages indirectly contribute to ECM forma- was high in day 1 and 3 post-MI macrophages and returned tion by releasing paracrine factors that stimulate cardiac towards day 0 values by day 7. While day 1 macrophages fibroblasts, macrophages can directly secrete ECM proteins upregulate pro-inflammatory genes early in response to MI, [6]. Indeed, fibronectin expression was highly upregulated several genes that inhibit these pathways are also highly in day 1 macrophages and returned towards baseline by induced (e.g., Slfn4, Cd9, Tnip1, and Gpr132), most likely day 7. Collagens type VI and VIII were also elevated at to exert negative feedback and limit excessive inflammation. day 7 post-MI and have been shown to increase in response 1 3 Basic Research in Cardiology (2018) 113:26 Page 13 of 18 26 to anti-inflammatory stimuli (e.g., TGF-β1 and IL-4) and reported to be undetectable in the infarcted LV, was decrease in response to pro-inflammatory stimuli (e.g., LPS decreased in macrophages at day 1 post-MI, increased at day and IFN-γ) [71, 89]. Our results reveal a previously unde- 3, and returned to day 0 values by day 7 [17]. Expressions fined role for macrophages in directly contributing ECM of the anti-inflammatory cytokines Tgfb1 and Il10 were proteins to the infarct scar. unchanged in macrophages following MI. One explanation is Macrophage polarization towards M1 and M2 subtypes that other cell types are the major source of these cytokines; has been well-defined in vitro for some, but not all stimuli. a second is that regulation occurs at the post-translational Post-MI in vivo macrophage polarization is less clear, as stage. Tgfb1 is regulated at the post-translational stage, as there are a number of factors over a broad spectrum within pre-formed TGFβ1 protein is produced and sequestered in the infarct environment that co-stimulate to generate an the ECM by binding to latent TGFβ binding protein. Bio- overall macrophage phenotype. After MI, the inflamma- availability is regulated by proteases (including MMPs) that tory M1 phenotype predominates early (day 1), whereas cleave the binding protein. Ym1 (Chi3l3) and Fizz1 (Retnla) the M2 reparative phenotype later becomes the major have been implicated as M2 macrophage markers [27, 67]. subtype (day 7). While our enrichment analyses indicated Ym1/Chi3l3, which was highly upregulated at days 1 and 3 a pro-inflammatory day 1 post-MI phenotype and an anti- post-MI and downregulated at day 7, is a lectin that binds inflammatory/reparative phenotype at day 7 post-MI, gene heparan-type glycosaminoglycans to influence inflammation expression for typical M1 and M2 markers did not reflect the [69]. Fizz1/Retnla was highly decreased in macrophages at in vitro scenario. For example, while Nos2 (an M1 marker) all days post-MI, and its role in inflammation is less clear, as was increased 5-fold at day 1, Arg1 (an M2 marker) was it possesses both pro- and anti-inflammatory roles depending increased over 900-fold. Nos2 returned to baseline at days 3 on context [69]. Our results support the growing consensus and 7, while Arg1 remained elevated. The concept of argin- to abandon the M1 and M2 nomenclature, as pro-inflam- ase-1 as an M2 marker has been challenged, as a variety of matory day 1 macrophages do not fully display typical M1 M1 stimuli induce arginase-1 expression [27]. It is possible features nor do pro-reparative day 7 macrophages display that ratios or combinations of profiles dictate net phenotype, typical M2 features [53]. with fine tuning occurring by fluxes in both pro- and anti- The post-MI macrophage polarization time contin- inflammatory constituents. uum map is shown in Fig.  8. Day 1 macrophages exhibit Our results indicate that IL-1β and IFN-γ are major a NF-κB and MAPK signaling repertoire, which are acti- macrophage-derived cytokines that promote the pro-inflam- vated in response to toll-like receptor signaling pathways matory environment early post-MI, as Il6 was unchanged and induce downstream pro-inflammatory molecules [27]. and Tnf was decreased. Il12a, which has previously been At day 3, the sirtuin and AMPK axis, which promote the Fig. 8 Map of macrophage Resident cM(D1) cM(D3) cM(D7) polarization over the MI time macrophage Monocyte continuum. Post-MI mac- rophage profiles in terms of inputs, signaling, outputs, and cell physiology over the three major post-MI remodeling Day 0 13 7 phases (inflammation, prolif- Days post-MI eration, and scar maturation). TIMP-1, ADAMs, DAMPs damage-associated Oxidative stress IL-1α/β collagen, Inputs* molecular patterns, ECM extra- Changes in metabolism DAMPs matricellularproteins cellular matrix TNFR1B, Sirtuins, AMPK, Integrins, TGFBR1, IL1R, TLRs metabolic/redox Signaling IL4Ra NF-kB, MAPK signals STAT3 IL-1α/β, IFN-γ, pro- IL-12, PF4, IL24, Collagen, ECM Outputs inflammatory cytokines reparative factors cytokines/chemokines metabolic outputs Cell Inflammation ECM synthesis and Phagocytosis, Proliferation Physiology ECM Breakdown Metabolic Reprogramming deposition Inflammation Remodeling Proliferation Phenotypes Maturation 1 3 26 Page 14 of 18 Basic Research in Cardiology (2018) 113:26 shift from glycolysis to mitochondrial oxidative phospho- Although fresh cardiac tissue is difficult to obtain from mice rylation and polarization towards a reparative phenotype, with cardiac rupture due to its spontaneous nature, serum may be important regulators of macrophage signaling [51, factors (i.e., factor XIII) or circulating monocyte markers 86]. In day 7 macrophages, STAT3/4, TGFBR1, and IL4Ra could be used to predict survival versus rupture [52]. were important pathways regulating the reparative pheno- In conclusion, macrophages show distinct transcriptomic type. Both IL-6 and IL-10 can promote M2 polarization profiles at different time points over the first week of post-MI through STAT3 activation [85]. Signaling through TGFBR1 wound healing. Metabolic shifts in post-MI macrophages influences macrophage polarization by inhibiting TNF-α may regulate polarization and cell physiology status and and iNOS expression and strongly induces collagen I and remain to be fully investigated. Of note, macrophages may III expression in fibroblasts [8 , 35, 72]. IL-4Ra, which was play a more direct role in post-MI ECM synthesis and increased at day 7 post-MI, drives M2 polarization and is remodeling than previously thought. Our findings provide critical for macrophage regulation of collagen remodeling novel insights into the potential mechanisms and pathways in skin wound healing [31]. Itga11, which was elevated at that regulate macrophage physiology during the inflamma- day 7 post-MI, encodes for a collagen-binding integrin that tory, granulation, and maturation phases of post-MI cardiac regulates intracellular collagen production [64]. remodeling. Overall, our work suggests that attention to tem- Resident (day 0) cardiac macrophages are derived from poral profiles should be given when considering therapeutic a different source than infiltrating monocyte-derived mac- strategies that alter macrophages. rophages. While the different origins of these two popula- Acknowledgements We thank Dr. Rugmani P. Iyer, Dr. Mira Jung, tions of macrophages could account for transcriptomic and and Presley C. Cannon for their technical support. We acknowledge phenotypic differences, the purpose of our study was to funding from the American Heart Association under Award Num- determine differences in macrophages within the cardiac ber 15SDG22930009, from the National Institutes of Health under Award Numbers GM103328, GM103476, GM104357, GM114833, environment that reflect cell physiological roles in mainte- GM115428, HL051971, HL075360, HL105324, HL129823, and nance and wound healing, rather than differences in circu- HL136438, and from the Biomedical Laboratory Research and Devel- lating monocytes before MI compared to infiltrating tissue opment Service of the Veterans Affairs Office of Research and Devel- monocytes after MI. Further, differentially expressed genes opment under Award Numbers 5I01BX000505 and IK2BX003922. T.C.F. is funded by an Institute Strategic Programme Grant funding were largely distinct across days 1, 3, and 7, indicating that from the Biotechnology and Biological Sciences Research Council the patterns of gene expression were not primarily reflect- (BB/J004227/1).The content is solely the responsibility of the authors ing cell source. Genes that were similarly differentially and does not necessarily represent the official views of any of the fund- expressed across all three of these days may be related to ing agencies. All authors have reviewed and approved the article. All authors have read the journal authorship agreement and policy on dis- either cell source or sustained expression. At last, the pro- closure of potential conflicts of interest and have nothing to disclose. cedure of isolating cardiac macrophages is different from isolating circulating monocytes, which could complicate Open Access This article is distributed under the terms of the Crea- interpretation of gene expression data. tive Commons Attribution 4.0 International License (http://creat iveco Transcriptomic analysis was performed on all cardiac mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- macrophages, rather than focusing on distinct macrophage tion, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the subpopulations. As this study is the first to report distinct Creative Commons license, and indicate if changes were made. changes in the macrophage transcriptome along the MI time course, our study provides the necessary framework for future studies to analyze distinct macrophage subpopula- tions within each time point. There was no indication that References pooling increased variation, as the coefficient of variation (CV) for individual cell counts was the smallest at day 7 1. Aarup A, Pedersen TX, Junker N, Christoffersen C, Bartels ED, Madsen M, Nielsen CH, Nielsen LB (2016) Hypoxia-inducible post-MI, whereas the CV for genes in Cluster 3 (Col1a1 and factor-1alpha expression in macrophages promotes development Postn) was highest at the time, indicating biological vari- of atherosclerosis. Arterioscler Thromb Vasc Biol 36:1782–1790. ability in response. Previous studies in our lab have observed https ://doi.org/10.1161/ATVBA HA.116.30783 0 biological variation in expression of these genes (Col1a1, 2. Aredo B, Li T, Chen X, Zhang K, Wang CX, Gou D, Zhao B, He Y, Ufret-Vincenty RL (2015) A chimeric Cfh transgene leads to CV = 140% and Postn, CV = 148%) in day 7 post-MI mac- increased retinal oxidative stress, inflammation, and accumulation rophages isolated from individual mice [28]. In the present of activated subretinal microglia in mice. Invest Ophthalmol Vis study, CV was 41% for Col1a1 and 46% for Postn, indicating Sci 56:3427–3440. https ://doi.org/10.1167/iovs.14-16089 there was no evidence of increased variation due to pooling. 3. Berger A, Tran AH, Paige CJ (2007) Co-regulated decrease of Neurokinin-1 receptor and Hemokinin-1 gene expression in mono- Our study focused only on surviving mice, which raises cytes and macrophages after activation with pro-inflammatory the question of changes in the macrophage transcriptome in surviving mice versus mice that undergo cardiac rupture. 1 3 Basic Research in Cardiology (2018) 113:26 Page 15 of 18 26 cytokines. J Neuroimmunol 187:83–93. https://doi.or g/10.1016/j. 16. Fleetwood AJ, Lee MKS, Singleton W, Achuthan A, Lee MC, jneur oim.2007.04.019 O’Brien-Simpson NM, Cook AD, Murphy AJ, Dashper SG, Reyn- 4. Bolick DT, Skaflen MD, Johnson LE, Kwon SC, Howatt D, olds EC, Hamilton JA (2017) metabolic remodeling, inflamma- Daugherty A, Ravichandran KS, Hedrick CC (2009) G2A defi- some activation, and pyroptosis in macrophages stimulated by ciency in mice promotes macrophage activation and atheroscle- Porphyromonas gingivalis and its outer membrane vesicles. rosis. Circ Res 104:318–327. https ://doi.or g/10.1161/CIRCR Front Cell Infect Microbiol 7:351. https ://doi.org/10.3389/fcimb ESAHA .108.18113 1 .2017.00351 5. Camell CD, Nguyen KY, Jurczak MJ, Christian BE, Shulman 17. Frangogiannis NG, Mendoza LH, Lindsey ML, Ballantyne CM, GI, Shadel GS, Dixit VD (2015) Macrophage-specific de novo Michael LH, Smith CW, Entman ML (2000) IL-10 is induced synthesis of ceramide is dispensable for inflammasome-driven in the reperfused myocardium and may modulate the reaction to inflammation and insulin resistance in obesity. J Biol Chem injury. J Immunol 165:2798–2808. https://doi.or g/10.4049/jimmu 290:29402–29413. https ://doi.org/10.1074/jbc.M115.68019 9 nol.165.5.2798 6. Chang MY, Chan CK, Braun KR, Green PS, O’Brien KD, Chait 18. Gombozhapova A, Rogovskaya Y, Shurupov V, Rebenkova M, A, Day AJ, Wight TN (2012) Monocyte-to-macrophage dif- Kzhyshkowska J, Popov SV, Karpov RS, Ryabov V (2017) Mac- ferentiation: synthesis and secretion of a complex extracellular rophage activation and polarization in post-infarction cardiac matrix. J Biol Chem 287:14122–14135. https ://doi.org/10.1074/ remodeling. J Biomed Sci 24:13. https ://doi.org/10.1186/s1292 jbc.M111.32498 8 9-017-0322-3 7. Chen H, Gao W, Yang Y, Guo S, Wang H, Wang W, Zhang S, 19. Gurtler C, Carty M, Kearney J, Schattgen SA, Ding A, Fitzger- Zhou Q, Xu H, Yao J, Tian Z, Li B, Cao W, Zhang Z, Tian Y ald KA, Bowie AG (2014) SARM regulates CCL5 production in (2014) Inhibition of VDAC1 prevents Ca(2)(+)-mediated oxida- macrophages by promoting the recruitment of transcription fac- tive stress and apoptosis induced by 5-aminolevulinic acid medi- tors and RNA polymerase II to the Ccl5 promoter. J Immunol ated sonodynamic therapy in THP-1 macrophages. Apoptosis 192:4821–4832. https ://doi.org/10.4049/jimmu nol.13029 80 19:1712–1726. https ://doi.org/10.1007/s1049 5-014-1045-5 20. Hardbower DM, Asim M, Luis PB, Singh K, Barry DP, Yang C, 8. Cheng R, Dang R, Zhou Y, Ding M, Hua H (2017) MicroRNA-98 Steeves MA, Cleveland JL, Schneider C, Piazuelo MB, Gobert inhibits TGF-beta1-induced differentiation and collagen pro- AP, Wilson KT (2017) Ornithine decarboxylase regulates M1 duction of cardiac fibroblasts by targeting TGFBR1. Hum Cell macrophage activation and mucosal inflammation via histone 30:192–200. https ://doi.org/10.1007/s1357 7-017-0163-0 modifications. Proc Natl Acad Sci USA 114:E751–E760. https :// 9. Cheng Y, Feng Y, Xia Z, Li X, Rong J (2017) omega-Alkynyl doi.org/10.1073/pnas.16149 58114 arachidonic acid promotes anti-inflammatory macrophage M2 21. Hashim S, Mukherjee K, Raje M, Basu SK, Mukhopadhyay A polarization against acute myocardial infarction via regulating (2000) Live Salmonella modulate expression of Rab proteins to the cross-talk between PKM2, HIF-1alpha and iNOS. Biochim persist in a specialized compartment and escape transport to lys- Biophys Acta 1862:1595–1605. https ://doi.org/10.1016/j.bbali osomes. J Biol Chem 275:16281–16288. https://doi.or g/10.1074/ p.2017.09.009 jbc.275.21.16281 10. Chiao YA, Dai Q, Zhang J, Lin J, Lopez EF, Ahuja SS, Chou 22. Heidt T, Courties G, Dutta P, Sager HB, Sebas M, Iwamoto Y, YM, Lindsey ML, Jin YF (2011) Multi-analyte profiling reveals Sun Y, Da Silva N, Panizzi P, van der Laan AM, Swirski FK, matrix metalloproteinase-9 and monocyte chemotactic protein-1 Weissleder R, Nahrendorf M (2014) Differential contribution of as plasma biomarkers of cardiac aging. Circ Cardiovasc Genet monocytes to heart macrophages in steady-state and after myo- 4:455–462. https://doi.or g/10.1161/CIRCGENETI CS.111.95998 1 cardial infarction. Circ Res 115:284–295. https: //doi.org/10.1161/ 11. DeLeon-Pennell KY, de Castro Bras LE, Iyer RP, Bratton DR, Jin CIRCR ESAHA .115.30356 7 YF, Ripplinger CM, Lindsey ML (2014) P. gingivalis lipopoly- 23. Hellmann J, Tang Y, Zhang MJ, Hai T, Bhatnagar A, Srivastava saccharide intensifies inflammation post-myocardial infarction S, Spite M (2015) Atf3 negatively regulates Ptgs2/Cox2 expres- through matrix metalloproteinase-9. J Mol Cell Cardiol 76:218– sion during acute inflammation. Prostaglandins Other Lipid 226. https ://doi.org/10.1016/j.yjmcc .2014.09.007 Mediat 116–117:49–56. https ://doi.or g/10.1016/j.pr os t ag lan 12. DeLeon-Pennell KY, Iyer RP, Ero OK, Cates CA, Flynn ER, Can- dins.2015.01.001 non PL, Jung M, Shannon D, Garrett MR, Buchanan W, Hall 24. Honold L, Nahrendorf M (2018) Resident and monocyte-derived ME, Ma Y, Lindsey ML (2017) Periodontal-induced chronic macrophages in cardiovascular disease. Circ Res 122:113–127. inflammation triggers macrophage secretion of Ccl12 to inhibit https ://doi.org/10.1161/circr esaha .117.31107 1 fibroblast-mediated cardiac wound healing. JCI Insight. https :// 25. Hulsmans M, Sam F, Nahrendorf M (2016) Monocyte and mac- doi.org/10.1172/jci.insig ht.94207 rophage contributions to cardiac remodeling. J Mol Cell Cardiol 13. Dutta P, Nahrendorf M (2015) Monocytes in myocardial infarc- 93:149–155. https ://doi.org/10.1016/j.yjmcc .2015.11.015 tion. Arterioscler Thromb Vasc Biol 35:1066–1070. https ://doi. 26. Iyer RP, Patterson NL, Zouein FA, Ma Y, Dive V, de Castro Bras org/10.1161/ATVBA HA.114.30465 2 LE, Lindsey ML (2015) Early matrix metalloproteinase-12 inhi- 14. Epelman S, Lavine KJ, Beaudin AE, Sojka DK, Carrero JA, Cal- bition worsens post-myocardial infarction cardiac dysfunction deron B, Brija T, Gautier EL, Ivanov S, Satpathy AT, Schilling by delaying inflammation resolution. Int J Cardiol 185:198–208. JD, Schwendener R, Sergin I, Razani B, Forsberg EC, Yokoyama https ://doi.org/10.1016/j.ijcar d.2015.03.054 WM, Unanue ER, Colonna M, Randolph GJ, Mann DL (2014) 27. Jablonski KA, Amici SA, Webb LM, Ruiz-Rosado Jde D, Popo- Embryonic and adult-derived resident cardiac macrophages are vich PG, Partida-Sanchez S, Guerau-de-Arellano M (2015) Novel maintained through distinct mechanisms at steady state and dur- markers to delineate murine M1 and M2 macrophages. PLoS ONE ing inflammation. Immunity 40:91–104. https://doi.or g/10.1016/j. 10:e0145342. https ://doi.org/10.1371/journ al.pone.01453 42 immun i.2013.11.019 28. Jung M, Ma Y, Iyer RP, DeLeon-Pennell KY, Yabluchanskiy A, 15. Feig JE, Vengrenyuk Y, Reiser V, Wu C, Statnikov A, Aliferis Garrett MR, Lindsey ML (2017) IL-10 improves cardiac remod- CF, Garabedian MJ, Fisher EA, Puig O (2012) Regression of eling after myocardial infarction by stimulating M2 macrophage atherosclerosis is characterized by broad changes in the plaque polarization and fibroblast activation. Basic Res Cardiol 112:33. macrophage transcriptome. PLoS ONE 7:e39790. h tt ps : // do i.https ://doi.org/10.1007/s0039 5-017-0622-5 org/10.1371/journ al.pone.00397 90 29. Karunakaran D, Thrush AB, Nguyen MA, Richards L, Geoffrion M, Singaravelu R, Ramphos E, Shangari P, Ouimet M, Pezacki 1 3 26 Page 16 of 18 Basic Research in Cardiology (2018) 113:26 JP, Moore KJ, Perisic L, Maegdefessel L, Hedin U, Harper ME, 42. London A, Cohen M, Schwartz M (2013) Microglia and mono- Rayner KJ (2015) Macrophage mitochondrial energy status regu- cyte-derived macrophages: functionally distinct populations that lates cholesterol efflux and is enhanced by anti-miR33 in athero- act in concert in CNS plasticity and repair. Front Cell Neurosci sclerosis. Circ Res 117:266–278. https ://doi.org/10.1161/CIRCR 7:34. https ://doi.org/10.3389/fncel .2013.00034 ESAHA .117.30562 4 43. Ma Y, Halade GV, Zhang J, Ramirez TA, Levin D, Voorhees 30. Kirschnek S, Ying S, Fischer SF, Hacker H, Villunger A, Hochrein A, Jin YF, Han HC, Manicone AM, Lindsey ML (2013) Matrix H, Hacker G (2005) Phagocytosis-induced apoptosis in mac- metalloproteinase-28 deletion exacerbates cardiac dysfunction rophages is mediated by up-regulation and activation of the Bcl-2 and rupture after myocardial infarction in mice by inhibiting homology domain 3-only protein Bim. J Immunol 174:671–679. M2 macrophage activation. Circ Res 112:675–688. https ://doi. https ://doi.org/10.4049/jimmu nol.174.2.671org/10.1161/CIRCR ESAHA .111.30050 2 31. Knipper JA, Willenborg S, Brinckmann J, Bloch W, Maass T, 44. Ma Y, Mouton AJ, Lindsey ML (2018) Cardiac macrophage biol- Wagener R, Krieg T, Sutherland T, Munitz A, Rothenberg ME, ogy in the steady-state heart, the aging heart, and following myo- Niehoff A, Richardson R, Hammerschmidt M, Allen JE, Eming cardial infarction. Transl Res 191:15–28. https:/ /doi.org/10.1016/j. SA (2015) Interleukin-4 receptor alpha signaling in myeloid trsl.2017.10.001 cells controls collagen fibril assembly in skin repair. Immunity 45. Majmudar MD, Keliher EJ, Heidt T, Leuschner F, Truelove 43:803–816. https ://doi.org/10.1016/j.immun i.2015.09.005 J, Sena BF, Gorbatov R, Iwamoto Y, Dutta P, Wojtkiewicz G, 32. Lambert JM, Lopez EF, Lindsey ML (2008) Macrophage roles Courties G, Sebas M, Borodovsky A, Fitzgerald K, Nolte MW, following myocardial infarction. Int J Cardiol 130:147–158. Dickneite G, Chen JW, Anderson DG, Swirski FK, Weissleder R, https ://doi.org/10.1016/j.ijcar d.2008.04.059 Nahrendorf M (2013) Monocyte-directed RNAi targeting CCR2 33. Langston PK, Shibata M, Horng T (2017) Metabolism sup- improves infarct healing in atherosclerosis-prone mice. Circu- ports macrophage activation. Front Immunol 8:61. https ://doi. lation 127:2038–2046. https ://doi.or g/10.1161/CIRCU L A TIO org/10.3389/fimmu .2017.00061 NAHA.112.00011 6 34. Lavine KJ, Epelman S, Uchida K, Weber KJ, Nichols CG, Schil- 46. Mead JR, Hughes TR, Irvine SA, Singh NN, Ramji DP (2003) ling JD, Ornitz DM, Randolph GJ, Mann DL (2014) Distinct Interferon-gamma stimulates the expression of the inducible macrophage lineages contribute to disparate patterns of cardiac cAMP early repressor in macrophages through the activation recovery and remodeling in the neonatal and adult heart. Proc of casein kinase 2. A potentially novel pathway for interferon- Natl Acad Sci USA 111:16029–16034. https://doi.or g/10.1073/ gamma-mediated inhibition of gene transcription. J Biol Chem pnas.14065 08111 278:17741–17751. https ://doi.org/10.1074/jbc.M3016 02200 35. Lee YU, de Dios Ruiz-Rosado J, Mahler N, Best CA, Tara S, 47. Meschiari CA, Jung M, Iyer RP, Yabluchanskiy A, Toba H, Yi T, Shoji T, Sugiura T, Lee AY, Robledo-Avila F, Hibino Garrett MR, Lindsey ML (2018) Macrophage overexpression N, Pober JS, Shinoka T, Partida-Sanchez S, Breuer CK (2016) of matrix metalloproteinase-9 in aged mice improves diastolic TGF-beta receptor 1 inhibition prevents stenosis of tissue-engi- physiology and cardiac wound healing after myocardial infarc- neered vascular grafts by reducing host mononuclear phagocyte tion. Am J Physiol Heart Circ Physiol 314:H224–H235. https :// activation. FASEB J 30:2627–2636. h t t p s : / / d o i . o r g /1 0 . 1 0 9 6 /doi.org/10.1152/ajphe art.00453 .2017 fj.20150 0179R 48. Mitchell A, Rentero C, Endoh Y, Hsu K, Gaus K, Geczy C, 36. Leuschner F, Rauch PJ, Ueno T, Gorbatov R, Marinelli B, Lee McNeil HP, Borges L, Tedla N (2008) LILRA5 is expressed by WW, Dutta P, Wei Y, Robbins C, Iwamoto Y, Sena B, Chudnovs- synovial tissue macrophages in rheumatoid arthritis, selectively kiy A, Panizzi P, Keliher E, Higgins JM, Libby P, Moskowitz MA, induces pro-inflammatory cytokines and IL-10 and is regulated Pittet MJ, Swirski FK, Weissleder R, Nahrendorf M (2012) Rapid by TNF-alpha, IL-10 and IFN-gamma. Eur J Immunol 38:3459– monocyte kinetics in acute myocardial infarction are sustained by 3473. https ://doi.org/10.1002/eji.20083 8415 extramedullary monocytopoiesis. J Exp Med 209:123–137. https 49. Molawi K, Wolf Y, Kandalla PK, Favret J, Hagemeyer N, Frenzel ://doi.org/10.1084/jem.20111 009 K, Pinto AR, Klapproth K, Henri S, Malissen B, Rodewald HR, 37. Li D, Duan M, Feng Y, Geng L, Li X, Zhang W (2016) MiR- Rosenthal NA, Bajeno ff M, Prinz M, Jung S, Sieweke MH (2014) 146a modulates macrophage polarization in systemic juvenile idi- Progressive replacement of embryo-derived cardiac macrophages opathic arthritis by targeting INHBA. Mol Immunol 77:205–212. with age. J Exp Med 211:2151–2158. h t t p s : / / d o i . o rg / 1 0 . 1 0 8 4 / https ://doi.org/10.1016/j.molim m.2016.08.007jem.20140 639 38. Ligresti G, Aplin AC, Dunn BE, Morishita A, Nicosia RF (2012) 50. Mor-Vaknin N, Punturieri A, Sitwala K, Faulkner N, Legendre M, The acute phase reactant orosomucoid-1 is a bimodal regulator Khodadoust MS, Kappes F, Ruth JH, Koch A, Glass D, Petruzzelli of angiogenesis with time- and context-dependent inhibitory L, Adams BS, Markovitz DM (2006) The DEK nuclear autoanti- and stimulatory properties. PLoS ONE 7:e41387. https ://doi. gen is a secreted chemotactic factor. Mol Cell Biol 26:9484–9496. org/10.1371/journ al.pone.00413 87https ://doi.org/10.1128/MCB.01030 -06 39. Lindsey ML, Bolli R, Canty JM, Du XJ, Frangogiannis NG, Frantz 51. Moreira D, Rodrigues V, Abengozar M, Rivas L, Rial E, Laforge S, Gourdie RG, Holmes JW, Jones SP, Kloner R, Lefer DJ, Liao R, M, Li X, Foretz M, Viollet B, Estaquier J, Cordeiro da Silva A, Murphy E, Ping P, Przyklenk K, Recchia FA, Schwartz Longacre Silvestre R (2015) Leishmania infantum modulates host mac- L, Ripplinger CM, Van Eyk JE, Heusch G (2018) Guidelines for rophage mitochondrial metabolism by hijacking the SIRT1- experimental models of myocardial ischemia and infarction. Am J AMPK axis. PLoS Pathog 11:e1004684. https://doi.or g/10.1371/ Physiol Heart Circ Physiol. https ://doi.or g/10.1152/a jphe ar t.00335 journ al.ppat.10046 84 .2017 52. Nahrendorf M, Aikawa E, Figueiredo JL, Stangenberg L, van den 40. Lindsey ML, Kassiri Z, Virag JAI, de Castro Bras LE, Scherrer- Borne SW, Blankesteijn WM, Sosnovik DE, Jaffer FA, Tung CH, Crosbie M (2018) Guidelines for measuring cardiac physiology Weissleder R (2008) Transglutaminase activity in acute infarcts in mice. Am J Physiol Heart Circ Physiol. https://doi.or g/10.1152/ predicts healing outcome and left ventricular remodelling: impli- ajphe art.00339 .2017 cations for FXIII therapy and antithrombin use in myocardial 41. Lindsey ML, Saucerman JJ, DeLeon-Pennell KY (2016) Knowl- infarction. Eur Heart J 29:445–454. https://doi.or g/10.1093/eurhe edge gaps to understanding cardiac macrophage polarization fol-artj/ehm55 8 lowing myocardial infarction. Biochim Biophys Acta 1862:2288– 2292. https ://doi.org/10.1016/j.bbadi s.2016.05.013 1 3 Basic Research in Cardiology (2018) 113:26 Page 17 of 18 26 53. Nahrendorf M, Swirski FK (2016) Abandoning M1/M2 for a net- 67. Raes G, De Baetselier P, Noel W, Beschin A, Brombacher F, Has- work model of macrophage function. Circ Res 119:414–417. https sanzadeh GhG (2002) Differential expression of FIZZ1 and Ym1 ://doi.org/10.1161/CIRCR ESAHA .116.30919 4 in alternatively versus classically activated macrophages. J Leukoc 54. Nahrendorf M, Swirski FK, Aikawa E, Stangenberg L, Wurdinger Biol 71:597–602. https ://doi.org/10.1189/jlb.71.4.597 T, Figueiredo JL, Libby P, Weissleder R, Pittet MJ (2007) The 68. Rienks M, Carai P, Bitsch N, Schellings M, Vanhaverbeke M, healing myocardium sequentially mobilizes two monocyte sub- Verjans J, Cuijpers I, Heymans S, Papageorgiou A (2017) Sema3A sets with divergent and complementary functions. J Exp Med promotes the resolution of cardiac inflammation after myocardial 204:3037–3047. https ://doi.org/10.1084/jem.20070 885 infarction. Basic Res Cardiol 112:42. https ://doi.org/10.1007/ 55. Navasa N, Martin-Ruiz I, Atondo E, Sutherland JD, Angel Pas-s0039 5-017-0630-5 cual-Itoiz M, Carreras-Gonzalez A, Izadi H, Tomas-Cortazar J, 69. Roszer T (2015) Understanding the mysterious M2 macrophage Ayaz F, Martin-Martin N, Torres IM, Barrio R, Carracedo A, through activation markers and effector mechanisms. Mediat Olivera ER, Rincon M, Anguita J (2015) Ikaros mediates the Inflamm 2015:816460. https ://doi.org/10.1155/2015/81646 0 DNA methylation-independent silencing of MCJ/DNAJC15 70. Sager HB, Heidt T, Hulsmans M, Dutta P, Courties G, Sebas M, gene expression in macrophages. Sci Rep 5:14692. https ://doi. Wojtkiewicz GR, Tricot B, Iwamoto Y, Sun Y, Weissleder R, org/10.1038/srep1 4692 Libby P, Swirski FK, Nahrendorf M (2015) Targeting interleu- 56. Nishiya T, Matsumoto K, Maekawa S, Kajita E, Horinouchi T, kin-1beta reduces leukocyte production after acute myocardial Fujimuro M, Ogasawara K, Uehara T, Miwa S (2011) Regula- infarction. Circulation 132:1880–1890. https: //doi.org/10.1161/ tion of inducible nitric-oxide synthase by the SPRY domain- and CIRCU LATIO NAHA.115.01616 0 SOCS box-containing proteins. J Biol Chem 286:9009–9019. 71. Schnoor M, Cullen P, Lorkowski J, Stolle K, Robenek H, Troyer https ://doi.org/10.1074/jbc.M110.19067 8 D, Rauterberg J, Lorkowski S (2008) Production of type VI col- 57. O’Neill LA, Pearce EJ (2016) Immunometabolism governs den- lagen by human macrophages: a new dimension in macrophage dritic cell and macrophage function. J Exp Med 213:15–23. https functional heterogeneity. J Immunol 180:5707–5719. https :// ://doi.org/10.1084/jem.20151 570doi.org/10.4049/jimmu nol.180.8.5707 58. O’Shannessy DJ, Somers EB, Wang LC, Wang H, Hsu R (2015) 72. Schridde A, Bain CC, Mayer JU, Montgomery J, Pollet E, Expression of folate receptors alpha and beta in normal and can- Denecke B, Milling SWF, Jenkins SJ, Dalod M, Henri S, Mal- cerous gynecologic tissues: correlation of expression of the beta issen B, Pabst O, McL Mowat A (2017) Tissue-specific differ - isoform with macrophage markers. J Ovarian Res 8:29. https :// entiation of colonic macrophages requires TGFbeta receptor- doi.org/10.1186/s1304 8-015-0156-0 mediated signaling. Mucosal Immunol 10:1387–1399. https :// 59. Oakes JL, O’Connor BP, Warg LA, Burton R, Hock A, Loader J, doi.org/10.1038/mi.2016.142 Laflamme D, Jing J, Hui L, Schwartz DA, Yang IV (2013) Ozone 73. Shao J, Zhang B, Yu JJ, Wei CY, Zhou WJ, Chang KK, Yang enhances pulmonary innate immune response to a Toll-like recep- HL, Jin LP, Zhu XY, Li MQ (2016) Macrophages promote the tor-2 agonist. Am J Respir Cell Mol Biol 48:27–34. https ://doi. growth and invasion of endometrial stromal cells by downregu- org/10.1165/rcmb.2012-0187O C lating IL-24 in endometriosis. Reproduction 152:673–682. https 60. Okerblom JJ, Schwarz F, Olson J, Fletes W, Ali SR, Martin PT, ://doi.org/10.1530/REP-16-0278 Glass CK, Nizet V, Varki A (2017) Loss of CMAH during human 74. Suzuki M, Tachibana I, Takeda Y, He P, Minami S, Iwasaki T, evolution primed the monocyte-macrophage lineage toward a Kida H, Goya S, Kijima T, Yoshida M, Kumagai T, Osaki T, more inflammatory and phagocytic state. J Immunol 198:2366– Kawase I (2009) Tetraspanin CD9 negatively regulates lipopol- 2373. https ://doi.org/10.4049/jimmu nol.16014 71 ysaccharide-induced macrophage activation and lung inflamma - 61. Pinto AR, Godwin JW, Chandran A, Hersey L, Ilinykh A, Debu- tion. J Immunol 182:6485–6493. https://doi.or g/10.4049/jimmu que R, Wang L, Rosenthal NA (2014) Age-related changes in nol.08027 97 tissue macrophages precede cardiac functional impairment. Aging 75. Takahashi T, Tang T, Lai NC, Roth DM, Rebolledo B, Saito (Albany NY) 6:399–413. https ://doi.org/10.18632 /aging .10066 9 M, Lew WY, Clopton P, Hammond HK (2006) Increased car- 62. Pinto AR, Godwin JW, Rosenthal NA (2014) Macrophages in diac adenylyl cyclase expression is associated with increased cardiac homeostasis, injury responses and progenitor cell mobi- survival after myocardial infarction. Circulation 114:388–396. lisation. Stem Cell Res 13:705–714. https ://doi.org/10.1016/j.https ://doi.org/10.1161/CIRCU LATIO NAHA.106.63251 3 scr.2014.06.004 76. Takeda Y, He P, Tachibana I, Zhou B, Miyado K, Kaneko H, 63. Pinto AR, Paolicelli R, Salimova E, Gospocic J, Slonimsky E, Suzuki M, Minami S, Iwasaki T, Goya S, Kijima T, Kumagai Bilbao-Cortes D, Godwin JW, Rosenthal NA (2012) An abundant T, Yoshida M, Osaki T, Komori T, Mekada E, Kawase I (2008) tissue macrophage population in the adult murine heart with a Double deficiency of tetraspanins CD9 and CD81 alters cell distinct alternatively-activated macrophage profile. PLoS ONE motility and protease production of macrophages and causes 7:e36814. https ://doi.org/10.1371/journ al.pone.00368 14 chronic obstructive pulmonary disease-like phenotype in mice. 64. Popova SN, Barczyk M, Tiger CF, Beertsen W, Zigrino P, Aszodi J Biol Chem 283:26089–26097. https ://doi.or g/10.1074/jbc. A, Miosge N, Forsberg E, Gullberg D (2007) Alpha11 beta1 inte-M8019 02200 grin-dependent regulation of periodontal ligament function in the 77. Tang C, Liu Y, Kessler PS, Vaughan AM, Oram JF (2009) The erupting mouse incisor. Mol Cell Biol 27:4306–4316. https: //doi. macrophage cholesterol exporter ABCA1 functions as an anti- org/10.1128/MCB.00041 -07 inflammatory receptor. J Biol Chem 284:32336–32343. https:// 65. Prabhu SD, Frangogiannis NG (2016) The biological basis for doi.org/10.1074/jbc.M109.04747 2 cardiac repair after myocardial infarction: from inflammation to 78. Tang SL, Chen WJ, Yin K, Zhao GJ, Mo ZC, Lv YC, Ouy- fibrosis. Circ Res 119:91–112. https ://doi.org/10.1161/CIRCR ang XP, Yu XH, Kuang HJ, Jiang ZS, Fu YC, Tang CK (2012) ESAHA .116.30357 7 PAPP-A negatively regulates ABCA1, ABCG1 and SR-B1 66. Praekelt U, Kopp PM, Rehm K, Linder S, Bate N, Patel B, expression by inhibiting LXRalpha through the IGF-I-mediated Debrand E, Manso AM, Ross RS, Conti F, Zhang MZ, Harris RC, signaling pathway. Atherosclerosis 222:344–354. https ://doi. Zent R, Critchley DR, Monkley SJ (2012) New isoform-specific org/10.1016/j.ather oscle rosis .2012.03.005 monoclonal antibodies reveal different sub-cellular localisations 79. Tetens J, Heuer C, Heyer I, Klein MS, Gronwald W, Junge W, Oefner PJ, Thaller G, Krattenmacher N (2015) Polymorphisms for talin1 and talin2. Eur J Cell Biol 91:180–191. https ://doi. within the APOBR gene are highly associated with milk levels org/10.1016/j.ejcb.2011.12.003 1 3 26 Page 18 of 18 Basic Research in Cardiology (2018) 113:26 of prognostic ketosis biomarkers in dairy cows. Physiol Genom- (2014) Ly6C(low) and not Ly6C(high) macrophages accumulate ics 47:129–137. https://doi.or g/10.1152/physiolg enomics .00126 first in the heart in a model of murine pressure-overload. PLoS .2014 ONE 9:e112710. https ://doi.org/10.1371/journ al.pone.01127 10 80. Tomczyk M, Kraszewska I, Szade K, Bukowska-Strakova K, Mel- 89. Weitkamp B, Cullen P, Plenz G, Robenek H, Rauterberg J (1999) oni M, Jozkowicz A, Dulak J, Jazwa A (2017) Splenic Ly6C(hi) Human macrophages synthesize type VIII collagen in vitro and monocytes contribute to adverse late post-ischemic left ventricular in the atherosclerotic plaque. FASEB J 13:1445–1457. https://doi. remodeling in heme oxygenase-1 deficient mice. Basic Res Car -org/10.1096/faseb j.13.11.1445 diol 112:39. https ://doi.org/10.1007/s0039 5-017-0629-y 90. Wenzel J, Ouderkirk JL, Krendel M, Lang R (2015) Class I myo- 81. Trial J, Heredia CP, Taffet GE, Entman ML, Cieslik KA (2017) sin Myo1e regulates TLR4-triggered macrophage spreading, Dissecting the role of myeloid and mesenchymal fibroblasts in chemokine release, and antigen presentation via MHC class II. age-dependent cardiac fibrosis. Basic Res Cardiol 112:34. https Eur J Immunol 45:225–237. https ://doi.org/10.1002/eji.20144 ://doi.org/10.1007/s0039 5-017-0623-4 4698 82. van Zuylen WJ, Garceau V, Idris A, Schroder K, Irvine KM, Lattin 91. Williams-Bey Y, Boularan C, Vural A, Huang NN, Hwang IY, JE, Ovchinnikov DA, Perkins AC, Cook AD, Hamilton JA, Hert- Shan-Shi C, Kehrl JH (2014) Omega-3 free fatty acids suppress zog PJ, Stacey KJ, Kellie S, Hume DA, Sweet MJ (2011) Mac- macrophage inflammasome activation by inhibiting NF-kappaB rophage activation and differentiation signals regulate schlafen-4 activation and enhancing autophagy. PLoS ONE 9:e97957. https gene expression: evidence for Schlafen-4 as a modulator of mye-://doi.org/10.1371/journ al.pone.00979 57 lopoiesis. PLoS ONE 6:e15723. https ://d oi.org/10.1371 /j our n 92. Xia J, Sinelnikov IV, Han B, Wishart DS (2015) MetaboAnalyst al.pone.00157 23 3.0–making metabolomics more meaningful. Nucleic Acids Res 83. Vasilopoulou E, Kolatsi-Joannou M, Lindenmeyer MT, White 43:W251–W257. https ://doi.org/10.1093/nar/gkv38 0 KE, Robson MG, Cohen CD, Sebire NJ, Riley PR, Winyard PJ, 93. Xia J, Wishart DS (2016) Using MetaboAnalyst 3.0 for compre- Long DA (2016) Loss of endogenous thymosin beta4 acceler- hensive metabolomics data analysis. Curr Protoc Bioinform 55:14 ates glomerular disease. Kidney Int 90:1056–1070. https ://doi. 10 11–14 10 91. https ://doi.org/10.1002/cpbi.11 org/10.1016/j.kint.2016.06.032 94. Zamilpa R, Zhang J, Chiao YA, de Castro Bras LE, Halade GV, 84. Virag JA, Lust RM (2014) Circadian influences on myocardial Ma Y, Hacker SO, Lindsey ML (2013) Cardiac wound healing infarction. Front Physiol 5:422. https ://doi.or g/10.3389/fphy s post-myocardial infarction: a novel method to target extracellu- .2014.00422 lar matrix remodeling in the left ventricle. Methods Mol Biol 85. Wang N, Liang H, Zen K (2014) Molecular mechanisms that influ- 1037:313–324. https ://doi.org/10.1007/978-1-62703 -505-7_18 ence the macrophage m1–m2 polarization balance. Front Immunol 95. Zhou J, Wu R, High AA, Slaughter CA, Finkelstein D, Rehg JE, 5:614. https ://doi.org/10.3389/fimmu .2014.00614 Redecke V, Hacker H (2011) A20-binding inhibitor of NF-kappaB 86. Weinberg SE, Sena LA, Chandel NS (2015) Mitochondria in the (ABIN1) controls Toll-like receptor-mediated CCAAT/enhancer- regulation of innate and adaptive immunity. Immunity 42:406– binding protein beta activation and protects from inflammatory 417. https ://doi.org/10.1016/j.immun i.2015.02.002 disease. Proc Natl Acad Sci USA 108:E998–E1006. https ://doi. 87. Weirather J, Hofmann UD, Beyersdorf N, Ramos GC, Vogel B, org/10.1073/pnas.11062 32108 Frey A, Ertl G, Kerkau T, Frantz S (2014) Foxp3+ CD4+ T cells 96. Zhu M, Goetsch SC, Wang Z, Luo R, Hill JA, Schneider J, improve healing after myocardial infarction by modulating mono- Morris SM Jr, Liu ZP (2015) FoxO4 promotes early inflamma- cyte/macrophage differentiation. Circ Res 115:55–67. https://doi. tory response upon myocardial infarction via endothelial Arg1. org/10.1161/CIRCR ESAHA .115.30389 5 Circ Res 117:967–977. https ://doi.org/10.1161/CIRCR ESAHA 88. Weisheit C, Zhang Y, Faron A, Kopke O, Weisheit G, Steinstrasser .115.30691 9 A, Frede S, Meyer R, Boehm O, Hoeft A, Kurts C, Baumgarten G 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Basic Research in Cardiology Springer Journals

Loading next page...
 
/lp/springer_journal/mapping-macrophage-polarization-over-the-myocardial-infarction-time-dzyJn4GDDF
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Cardiology
ISSN
0300-8428
eISSN
1435-1803
D.O.I.
10.1007/s00395-018-0686-x
Publisher site
See Article on Publisher Site

Abstract

In response to myocardial infarction (MI), cardiac macrophages regulate inflammation and scar formation. We hypothesized that macrophages undergo polarization state changes over the MI time course and assessed macrophage polarization transcrip- tomic signatures over the first week of MI. C57BL/6 J male mice (3–6 months old) were subjected to permanent coronary artery ligation to induce MI, and macrophages were isolated from the infarct region at days 1, 3, and 7 post-MI. Day 0, no MI resident cardiac macrophages served as the negative MI control. Whole transcriptome analysis was performed using RNA- sequencing on n = 4 pooled sets for each time. Day 1 macrophages displayed a unique pro-inflammatory, extracellular matrix high low (ECM)-degrading signature. By flow cytometry, day 0 macrophages were largely F4/80 Ly6C resident macrophages, low high whereas day 1 macrophages were largely F4/80 Ly6C infiltrating monocytes. Day 3 macrophages exhibited increased proliferation and phagocytosis, and expression of genes related to mitochondrial function and oxidative phosphorylation, indicative of metabolic reprogramming. Day 7 macrophages displayed a pro-reparative signature enriched for genes involved in ECM remodeling and scar formation. By triple in situ hybridization, day 7 infarct macrophages in vivo expressed collagen I and periostin mRNA. Our results indicate macrophages show distinct gene expression profiles over the first week of MI, with metabolic reprogramming important for polarization. In addition to serving as indirect mediators of ECM remodeling, macrophages are a direct source of ECM components. Our study is the first to report the detailed changes in the macrophage transcriptome over the first week of MI. Keywords Myocardial infarction · Macrophage · Transcriptome · RNA-Seq · LV remodeling Introduction Myocardial infarction (MI) invokes a cardiac wound healing Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s0039 5-018-0686-x) contains response that involves early initiation of inflammation, fol - supplementary material, which is available to authorized users. lowed by robust scar formation in the infarct area. The mac- rophage is a key regulator of cardiac remodeling, providing * Merry L. Lindsey both strong pro-inflammatory signals early and reparative mllindsey@umc.edu cues later [13, 24, 25, 41, 44]. Macrophages naturally reside Department of Physiology and Biophysics, Mississippi in the healthy heart, largely derived from the embryonic yolk Center for Heart Research, University of Mississippi Medical sac, proliferating locally and overseeing normal tissue main- Center, 2500 North State St., Jackson, MS 39216-4505, USA tenance [22, 34]. Research Service, G.V. (Sonny) Montgomery Veterans high In response to MI, circulating pro-ina fl mmatory Ly6C Affairs Medical Center, Jackson, MS 39216, USA monocytes are rapidly recruited from bone marrow and The Roslin Institute, University of Edinburgh, Easter Bush, splenic reservoirs to the infarcted area by the CC chemokine Midlothian, Scotland, UK CCL2/MCP-1, and require CCR2 (the endogenous receptor Department of Biomedical Engineering, University for CCL2) for extravasation [14, 80]. Early post-MI, mono- of Virginia, Charlottesville, VA, USA cyte-derived macrophages promote inflammation through Department of Pharmacology and Toxicology, University of Mississippi Medical Center, Jackson, MS 39216, USA Vol.:(0123456789) 1 3 26 Page 2 of 18 Basic Research in Cardiology (2018) 113:26 release of pro-inflammatory cytokines such as interleukin Coronary artery ligation (IL)-1β, a key regulator of post-MI inflammation [70]. In the healthy heart and following MI, monocyte/mac- To produce permanent MI, mice underwent coronary artery rophage subpopulations are identified by a limited number of ligation surgery as described previously and according to cell surface markers. Resident cardiac macrophages express the Guidelines for Experimental Models of Ischemia and high levels of the myeloid marker CD11b, as well as canoni- Infarction [12, 28, 39, 47, 94]. Mice were anesthetized with cal macrophage markers including CD14, CD86, CX3CR1, 2% isoflurane, intubated, and ventilated. The left coronary F4/80, and MHC-II, and display an anti-inflammatory M2 artery was ligated with 8–0 suture, and MI was confirmed by phenotype [63]. Resident macrophages are characterized left ventricle (LV) blanching and ST-segment elevation on by high expression of F4/80, while infiltrating monocytes the EKG. Mice were administered buprenorphine (0.05 mg/ and monocyte-derived macrophages express F4/80 at lower kg body weight) immediately before surgery. high low levels [22, 62]. Following MI, F4/80 Ly6C resident macrophages are rapidly replaced by pro-inflammatory infil- low high high trating F4/80 Ly6C monocytes that are CCR2 [22]; Echocardiography and necropsy by post-MI day 7, the macrophage phenotype shifts to a pre- dominantly anti-inflammatory/pro-reparative M2 phenotype. LV physiology was determined by transthoracic echocar- diography (Vevo 2100, VisualSonics; Toronto, CA) as Specifically targeting macrophages to improve MI out- comes has proven both promising and challenging, as described before and according to the Guidelines for Meas- uring Cardiac Physiology in Mice [12, 28, 40, 47]. Mice therapeutic approaches successfully manipulating the mac- rophage depend on both temporal and spatial factors [13, were anesthetized under 1–2% isoflurane, and both long and short-axis images were obtained. Measurements were taken 18]. For example, the early inflammatory response is critical for wound healing, as too little or excess inflammation can on the terminal day and were averaged from three cardiac cycles for each mouse. Following imaging, the hearts were adversely affect remodeling; the same paradigm is true for the later reparative and fibrotic response [18]. While mac- removed and the left ventricle (LV) divided into remote and infarct (which included border zone) regions. Each region rophages have been extensively studied in steady state and aging hearts [10, 49, 61], as well as following pressure over- was separately weighed for infarct area estimation. The infarct sizes over the 3 MI time points had a coefficient of load [88], the full picture of the macrophage evolution over the first week after MI has not been developed. variation of 18%, indicating gene variation was not likely due to differences in infarct sizes. Accordingly, the objective of this study was to combine transcriptomics, flow cytometry, and cell physiology to provide a map of macrophage phenotypes in response to Isolation of LV infarct macrophages MI. We analyzed transcriptomic changes at days 1, 3, and 7 post-MI to reflect the early inflammatory, proliferative, LV macrophages were isolated from the infarct region by and maturation phases. We hypothesized that macrophages immunomagnetic separation as described previously [12, would undergo changes over the MI time course that range 28]. Excised LV tissue was rinsed and immediately minced from pro-inflammatory to reparative polarization. To our and digested by collagenase II (Worthington; Lakewood, NJ) knowledge, this is the first study to report in detail the full and DNase solution in Hanks buffered saline solution. After transcriptome changes that occur in cardiac macrophages digestion, a single-cell suspension was generated and filtered that mediate post-MI wound healing and remodeling. through a 30 µm pre-separation column. Cell suspensions were incubated at 4 °C with an anti-Ly6G-biotin antibody (Miltenyi Biotech, Bergisch Gladbach, Germany, 130-092- Methods 332) to remove neutrophils, followed by an anti-CD11b-bio- tin antibody (Miltenyi 130-049-601) for 15 min, followed by Animal use anti-biotin microbeads (Miltenyi 130-092-332) for 10 min. Cells conjugated to the antibody microbeads were separated All procedures involving animals were approved by the Insti- by magnetic columns (Miltenyi 130-042-201). tutional Animal Care and Use Committee at the University The average numbers of isolated macrophages ± SEM of Mississippi Medical Center. A total of 122 C57BL/6J and the coefficient of variation (CV) from each of adult (3–6 month old) male mice were used for this study. the individual pooled animals for each time point Mice within this age range are at similar physiological matu- 5 5 were as follows: day 0—1.62 × 10 ± 0 .1 8 × 10 , ration [26]. Groups were randomly assigned prior to the sur- 5 5 CV = 31%; day 1—2.61 × 10 ± 0.33 × 10 , CV = 43%; geries by one investigator (KYDP) and another investigator 6 6 d ay 3 — 1 . 7 5 × 1 0 ± 0 . 0 8 × 1 0 , CV = 34%; and day (YM) performed the majority of surgeries. 1 3 Basic Research in Cardiology (2018) 113:26 Page 3 of 18 26 5 5 7—7.64 × 10 ± 0.46 × 10 , CV = 17%. To evaluate whether Bioinformatic analyses pooling increased variability, we compared the CV of the day 7 MI macrophage pools with the CV of a previously Analyses tools available in the online resource Metabo- published day 7 MI macrophage RNA-seq experiment that analyst 3.0 (http://www.metab oanal yst.ca/) and GraphPad used macrophages isolated from individual mice [28]. The Prism were used for graphical and statistical analyses [92, CV for the day 7 pooled set was 17%, while the CV for the 93]. FPKM values were uploaded into Metaboanalyst, and individual mice was 48% (n = 5). There was no indication, one-way ANOVA with Tukey’s post hoc test was performed therefore, that pooling increased variability. to determine differentially expressed genes (defined as false An initial assessment of macrophages freshly iso- discovery rate (FDR) adjusted p < 0.05). For individual post- lated from the myocardium yielded RNA that was not MI days, differential expression was characterized by a fold of sufficient quality for RNA-sequencing (RNA-Seq). change threshold of > 2.0 or < 0.5 compared to day 0 no MI To determine an optimal culturing time, day 3 post-MI values and a p value of < 0.05 by unpaired two-tailed t test. + − 6 CD11b Ly6G macrophages (1.5 × 10 cells/well) were Markov clustering analysis was performed independently by plated in 6-well culture dishes for 2 or 20 h in RPMI 1640 two investigators (AJM and TCF) using Graphia Pro soft- medium supplemented with 0.1% FBS and 1% antibiotics. ware (Kajeka, Edinburgh, UK) using genes with a pairwise After incubation, non-adherent cells were washed off and Pearson correlation threshold of r > 0.95. Both investiga- the remaining adherent cells were used for transcriptomics tors obtained similar results. Enrichment analysis for dif- analysis by RNA-Seq. The 20 h culturing changed mac- ferentially expressed genes was performed using Enrichr rophage phenotype; Supplemental Fig. 1), and based on (http://amp.pharm .msm.edu/Enric hr/) gene ontology (GO) this result, 2 h was selected as the incubation period for biological processes and Ingenuity Pathway Analysis (Qia- the time course evaluation. gen) canonical pathways. For GO terms, the combined score (calculated from Z-score and p value) was reported. RNA‑Seq RT‑PCR validation To obtain high-quality RNA for sequencing (from day 0 hearts in particular), macrophages were pooled from A total of five genes (Arg1, Ifng, Il1b, Lgals3, and Tnf) n = 81 hearts to obtain four biological replicates of were evaluated by quantitative RT-PCR on the same mac- 1.5  ×  10 cells for each day post-MI. Whole transcrip- rophage RNA samples used for RNA-Seq and assessed for tome analysis was performed as described previously [28, correlation. RNA was reverse transcribed to cDNA using 47]. RNA was extracted using the Pure Link RNA Mini the High Capacity RNA-to-cDNA kit (Applied Biosystems Kit (Ambion, Foster City, CA) according to manufacturer 4387406). Gene expression was quantified using the Taqman instructions and assessed for quality control parameters Gene Expression Assay and primers for Arg1, Ifng, Il1b, of minimum concentration and size range. cDNA librar- Lgals3, and Tnf (Applied Biosystems). Values for the arrays ies were developed using the TruSeq Total Stranded RNA were normalized to the housekeeping gene Hprt1. with RiboZero Kit (Ambion), set-A, quantified with the Qubit System (Invitrogen, Carlsbad, CA), and assessed for quality and size with the Experion DNA 1K Chip (Bio- Flow cytometry Rad, Hercules, CA). The libraries (n = 12 pooled samples per library) were sequenced using the NextSeq 500 High LV tissue excised from day 0 and day 1 post-MI mice was Output Kit (300 cycles, paired end 100 bp) on the Illu- minced and digested with 600 U/ml collagenase II (Wor- mina NextSeq 500 platform (Illumina, San Diego, CA). thington, LS004177, Lot 47E17554B) and 60 U/ml DNase Sequenced reads (n = 30–50; Cloud Computing Platform), I in Hanks buffered saline solution and filtered through a and Fastq sequence files were used to align reads to the 30-µm separation filter to generate single-cell suspensions. reference genome USCS-GRCm38/mm10) using RNA-Seq Red blood cells were lysed (Red Blood Cell Lysis Solution, Alignment Application with STAR aligner. Fragments per Miltenyi 130-094-183) and non-specific interactions were kilobase of transcript per million mapped reads (FPKM) blocked with FcR Blocking Reagent (Miltenyi 130-092-575). values of reference genes and transcripts were generated Cells were stained with the following fluorophore-conju- using Cufflinks 2. Variability (coefficient of variation) gated antibody panels: CD45-FITC (Miltenyi 130-110-658), of pooled sets was compared to variability of a previous CD11b-APC-Vio770 (Miltenyi 130-109-288), F4/80- RNA-Seq experiment on day 7 post-MI macrophages from PerCP-Vio700 (Miltenyi 130-102-161), Ly6C-VioBlue individual mice [28]. (Miltenyi 130-111-921), and Ly6G-APC (Miltenyi 130- 107-914). Samples were quantified using the MACSQuant 1 3 26 Page 4 of 18 Basic Research in Cardiology (2018) 113:26 Analyzer 10 (Miltenyi). Cell populations were gated on live 1:1000), and Postn (418581, 1:1500); nuclei were stained singlets, with cells from monocyte-derived/macrophage line- with 4′,6-diamidino-2-phenylindole (DAPI). Probes were + + − age classified as CD45 CD11b Ly6G cells. conjugated to the following fluorophores (Perkin Elmer, Waltham, MA, TSA Plus): fluorescein (NEL741E001KT), In vivo phagocytosis assay Cy3 (NEL744E001KT), or Cy5 (NEL745E001KT). Images were acquired at 40× using the Mantra Quantitative Pathol- To evaluate macrophage phagocytosis, day 0 or day 3 post- ogy Imaging System (Perkin Elmer), and analyses were MI mice were injected with 100 µg of fluorescein-labeled performed using the cell phenotyping feature in the inForm Escherichia coli K-12 BioParticles (Molecular Probes, software (Perkin Elmer). Eugene, OR, V-6694) through the jugular vein [22]. After 2 h, cardiac macrophages were isolated, cultured for 2 h Statistics to remove unattached cells, and fixed with 100% ethanol. Nuclei were stained with DAPI. Images were acquired using All experiments were performed and analyzed in a blinded an Olympus IX81 microscope. Phagocytic macrophages design. Data are presented as mean ± SEM. Survival rate was (green fluorescence) were counted as a percentage of the analyzed by Kaplan–Meier survival analysis and compared total cells per field. by the log rank test. For echocardiography, comparisons were made using one-way ANOVA followed by Tukey’s In vivo proliferation assay post hoc test. Statistics and bioinformatics for the RNA- sequencing are described above. RNA-sequencing compari- To evaluate macrophage proliferation, day 0 or day 3 post- sons to quantitative RT-PCR were made by Pearson’s linear MI mice were injected with 1 mg BrdU (Sigma, St. Louis, regression analysis. Two group comparisons were analyzed MO, 11647229001) intraperitoneally 2 h before being killed by unpaired two-tailed t test. A value of p < 0.05 was con- [22]. Isolated infarct macrophages were adhered to slides, sidered statistically significant. fixed with 100% ethanol, permeabilized with Triton-X 100, and stained with anti-BrdU-FITC antibody (eBioscience, Waltham, MA, 11-5071-42, 1:20). Nuclei were stained with Results DAPI. Images were acquired using an Olympus IX81 micro- scope. Proliferating cells (green fluorescence) were counted Macrophages continually polarize over the post‑MI as a percentage of total cells per field. time continuum In vivo macrophage turnover Proof of successful MI To evaluate macrophage turnover, mice were injected with a Day 7 post-MI survival was 56% (Supplemental Fig. 2a), FITC-F4/80 antibody (Biolegend, San Diego, CA, 123107, consistent with others and our own past reports [11, 28, 43, 200 µg/kg) through the jugular vein at 24 h post-MI (day 75, 87, 96]. Of the day 7 mice that did not survive, 58% 1) and killed at day 3 post-MI. MI mice without injection (7/12) died from cardiac rupture, as assessed at autopsy. served as negative controls. Infarct macrophages were iso- Infarct areas (% LVI mass by total LV mass) were similar lated, and FITC + cells were quantified using a MACSQuant among days 1, 3, and 7 MI groups (Supplemental Fig. 2b). Analyzer 10. The data were analyzed using the MACSQuan- As expected, cardiac physiology was impaired after MI, with tify software and were presented as the percentage of FITC LV infarct wall thinning and dilatation evident beginning at cells to total macrophages. day 1 after MI (Supplemental Fig. 2c–e). Fractional short- ening decreased similarly across all post-MI groups (Sup- Triple in situ hybridization plemental Fig. 2f). Day 7 post-MI LV sections (n = 3) collected from the mid- Proof of cell isolation purity papillary region were fixed in 10% zinc-buffered formalin for 24 h at room temperature paraffin-embedded, and sectioned Supplemental Fig. 3 shows FPKM values plotted for mac- at 5 µm. In situ hybridization was performed with three probe rophage specific markers (Cd14, Cd68, Emr1, Fcgr3, Itgam, sets per section using the RNAscope Multiplex Fluorescent and Lgals3) and compared to non-macrophage cell-specific Reagent Kit v2 (Advanced Cell Diagnostics, Newark, CA). markers for endothelial, fibroblast, lymphocyte, myocyte, Samples were hybridized using probes (all from Advanced and neutrophil cell types. FPKM values for macrophage Cell Diagnostics) specific for Acta2 (319531, 1:1000), Ccr2 specific markers averaged 455, while for non-macrophage (433271, 1:1000), Col1a1 (319371, 1:1500), Emr1 (317961, specific markers, the average FPKM was ~ 3. These results 1 3 Basic Research in Cardiology (2018) 113:26 Page 5 of 18 26 indicate that our macrophage isolations were ultra-pure and (1707) upregulated and 9% (1547) downregulated, and day free of contamination from other cell types. 7 had 5% (899) upregulated and 7% (1189) downregulated compared to day 0 macrophages (Fig. 1e). Fold change anal- Differentially expressed genes ysis of commonly used M1 and M2 markers is displayed in Supplemental Fig. 4. From the 23,847 genes in the dataset, 4064 were removed RT-PCR validation of FPKM values was performed for from analysis as FPKM values were 0 for all biological rep- Il1b, Arg1, Lgals3, Tnf, and Ifng on the same samples used licates in all four groups and 2938 were removed as there for RNA-Seq (Supplemental Fig. 5). These genes were cho- was < 3 replicates with values > 0 for any one group. Of sen based on their strong association with macrophages. the remaining 16,845 genes, 150 had duplicate or tripli- Positive correlations for Il1b (r = 0.99, p < 0.0001), Arg1 cate measurements that were removed. This left a total of (r = 0.95, p < 0.0001), and Lgals3 (r = 0.51, p = 0.02) were 16,695 unique transcripts for analysis (Fig. 1a; Supplemental observed. Tnf and Ifng showed low agreement due to low Table 1). By principal component analysis, each of the days detection by RT-PCR, either due to sensitivity issues or separated out into its own pattern, indicating that each day primer design. had a unique gene expression profile (Fig.  1b). By one-way ANOVA with Tukey’s post hoc test, 8109 genes differed Pattern clustering among groups (with FDR adjusted p < 0.05; Fig. 1c, heat map Fig. 1d). By fold change analysis using cut-off points Gene expression patterns across the post-MI remodeling of two for fold change and p < 0.05 for p value by unpaired spectrum were also analyzed by Markov pattern cluster- two-tailed t test, day 1 had 6% of genes (1019) significantly ing. A total of three major clusters distinguished the time upregulated and 12% (2043) downregulated; day 3 had 10% points (Fig. 2a), with representative genes from each cluster Days Post-MI Fold Change (a) (d) (b) Principal Component Analysis (e () 01 37 Analysi Analysis s 23,847 2043 1019 transcripts Day 7 1 evaluated 1547 1707 Day 0 1189 899 7,152 Day 3 -2000 02000 transcripts Upregulated Upregulated removed -3 Day 1 Day 1 16,695 transcripts -100 0100 analyzed PC1 (27.4%) 879 293 (c) One-way ANOVA Day 7 Day 3 Cfh 12 Ephx1 8,109 Cd209f Cd9 Fn1 transcripts 10 Pgk1 Downregulated DE by Day 1 ANOVA (FDR<0.05) 2 403 Pathway 0 5000 10000 15000 Analysis Day 7 Day 3 Genes Fig. 1 Distinct time-dependent gene expression profiles in post-MI were distinct from day 0 and day 1 and showed overlap. c One-way macrophages. a Of 23,847 genes sequenced, 7152 did not meet qual- ANOVA plot showing significant genes in red and d heat map of all ity control standards and were removed. Of the remaining 16,695 differentially expressed genes. e Fold change analysis of differentially genes, 8109 were differentially expressed (DE) by one-way ANOVA expressed genes at each day post-MI (fold change threshold of 2, (FDR adjusted p value < 0.05). b Macrophages from different post-MI FDR adjusted p value < 0.05) and Venn diagrams of upregulated and days analyzed by principal component analysis. Day 1 macrophages downregulated genes showing distinction and overlap in gene expres- were most distinct from the other times. Day 3 and 7 macrophages sion among the times 1 3 -log10(p) PC2 (19.1%) -100 0 100 Days post-MI 26 Page 6 of 18 Basic Research in Cardiology (2018) 113:26 Cluster 1Cluster 2Cluster 3 Translation elongation Neutrophil degranulation ECM organization (a) (b) Il1a Il1b Nlrp3 Cytoplasmic translation Collagen catabolic process Inflammatory response Translation fidelity Cell migration Canonical glycolysis Ptgs2 Plastid translation Regulation of small GTPase signaling Positive regulation of NF-κB Nos2 Cd80 Mitochondrial translation Positive regulation of PI3K signaling Cellular response to hypoxia Gapd p h Hif1 Hif1a a 05 0 50 01 10 00 0 01 0 10 00 0 20 2000 0 30 300 0 0 50 50 10 100 0 Combined Score Combined Score Combined Score Ikbkb Fold Change Fold Change Fold Change 4 #$ 15 * * #$ * 3 3 3 5 # 01 37 01 37 01 37 Days Days post post-M MII Days Days post post-MI MI Days Days pos post- t MI MI Sparc Eef2 Col1a1 Cl Col3 31 a1 Rpl3, 5, 6, 7 Eln Rps2, 7, 8, 18 Lox Postn Fig. 2 Gene expression pattern clustering. a Markov clustering analy- 1 (yellow) contained 565 genes increased at day 1 and was enriched sis generated three distinct major clusters representing different post- for pro-inflammatory processes. Cluster 2 genes (blue, 1965 genes) MI gene expression patterns. Each node represents a single gene, and were increased at day 3 and 7 and were enriched for translation pro- genes within the same cluster (color) show similar gene expression cesses. Cluster 3 (green, 1222 genes) were increased at day 7 and patterns. Genes representing each cluster are highlighted in red. b enriched for ECM processes. *p < 0.05 versus day 0, p < 0.05 versus Enrichment analysis and expression patterns for each cluster. Cluster day 1, p < 0.05 versus day 3 highlighted. Cluster 1 (yellow, 565 genes) showed an overall upregulation of some, but not all ECM genes and indirect expression pattern of mRNAs being increased only at day 1 regulation of ECM processes. after MI compared to day 0 (Fig. 2b). GO enrichment for this cluster revealed that pro-inflammation (neutrophil degranu- Day 1 MI macrophages display a pro‑inflammatory lation, p = 3.2E−11; inflammator y response, p = 7.5E−7; monocyte‑derived signature canonical glycolysis, p = 5.3E−7; positive regulation of NF-κB activity, p = 0.0018; cellular response to hypoxia, Genes differentially expressed at day 1 post-MI are displayed p = 0.00082) was the major process represented by these by volcano plot (Fig. 3a), with representative genes from genes, suggesting that day 1 macrophages rapidly upregu- each GO process highlighted. The representative genes were late inflammatory and glycolytic processes which are then the top five ranked up- and down-regulated genes based on rapidly turned off. The neutrophil degranulation term is term fold change (≥ 2) and p value at each time point. Notable overlap and does not imply neutrophil contamination, as we upregulated GO processes included inflammatory response saw essentially no expression of neutrophil-specific mark - (p = 0.007), cytokine-mediated signaling (p = 0.002), canoni- ers, and genes in our analysis represented by this term are cal glycolysis (p = 0.002), ECM disassembly (p = 0.002), and known to be expressed in macrophages (e.g., Mmp8/9, Mif, cellular response to hypoxia (p = 0.02; Fig. 3a). Downregu- and Lgals3). Cluster 2 (blue, 1965 genes) showed a pattern lated biological processes at day 1 included ECM organiza- of increasing at day 3 and day 7, with a peak at day 3. GO tion (p = 9.3E−7) and cell–matrix adhesion (p = 0.007). processes related to mRNA translation were prominent in To determine macrophage subpopulation heterogene- this cluster. Cluster 3 (green, 1222 genes) showed a pattern ity within the 1 day post-MI time, we performed multi- of being decreased at days 1 and 3, and either increased or marker flow cytometry. Myeloid cells (neutrophils and decreased at day 7. The top two GO processes were extra- macrophages) were elevated (27 ± 4% of total cells vs. cellular matrix (ECM) organization (p = 1.06E−16) and 4 ± 1% at day 0, a 6.2-fold increase; Fig.  3b). We fur- collagen catabolic process (p = 3.3E−7), reflecting direct ther gated on the myeloid cell population to differentiate 1 3 Basic Research in Cardiology (2018) 113:26 Page 7 of 18 26 Cd9 Upregulated Do Downregu nreg late lated d (a) Neutrophil degranulation Extracellular matrix organization Gapdh Actr3 Hif1a GTPasesignaling pathway Inflammatory response Itga11 Eln Ccr2 Cytokine-mediated signaling Mitotic chromosome condensation Il1b Col1a2 Canonical glycolysis Col3a1 Cardiac muscle contraction 2 Mmp8 Extracellular matrix disassembly Cell-matrix adhesion -10-50 5102 0 040600 10 20 30 4 log2FC Combined Score Combined Score (b) Myeloid Cells Neutrophils Neutrophils Day 0 Resident 80 Macrophage Monocytes/ Macrophage 50 40 Myeloid Monocyte cells 01 01 Resident Day 1 Macrophages Monocytes 01 01 CD11b CD11b Ly6C Days post-MI Days post-MI (c) Top 5 Genes Ranked by fold change then p-value Top 5 Genes Ranked by p value then fold change Expression in Expression in Gene p FC Macrophage Macrophage Role Role othe otherc r cardia ardiacc ccells ells Gene FC p Macrophage Role other cardia cardiacc ccells ells Pcdha4 1.2E7 0.002 Cell adhesion unknown Tetraspanin; inhibits M1 Cd9 2E-8 4.5 phenotypeubiquitous Pro-inflammatory Inhibition of Orm1 834 0.026 cytokine release ubiquitous Inhibition of Tnip1 5E-7 3.3 TNFα/LPS/EGF signalingubiquitous Pappa 22 0.016 cholesterol efflux myocytes Uptake of TG-rich Apobr 6E-7 35 3.5 lipoprotein lipoproteins( s (ApoB) endothelial Ih Inhibiti ibitif on of Slfn4 19 0.005 monocytopoiesis unknown Bcl2l11 1E-63.4 apoptosis myocytes M1 marker; induced pH-sensing; M2 Inhba 18 0.001 by glycolysis myocytes Gpr132 1E-62.3 polarization lymphocytes Fig. 3 Day 1 post-MI macrophages showed a pro-inflammatory pro- myeloid cells, neutrophils, and monocytes significantly increased in file. a Volcano plot with all values normalized to day 0 no MI con- the infarct region, while resident macrophages decreased. c Top five trols and representative genes highlighted (left). Enrichment analysis upregulated day 1 post-MI genes ranked by fold change (left) and p of upregulated and downregulated genes (right). b Flow cytometry value (right). *p < 0.05 versus day 0 analysis of macrophage cell surface markers. At day 1 post-MI, total monocytes and macrophages from neutrophils based on shown indicating that infiltrating monocytes replace resi- Ly6G expression. Ly6G + neutrophils were significantly dent macrophages early after MI [13, 14, 22, 54, 62, 63]. increased in the infarct region (54 ± 4% of the total To define the day 1 post-MI macrophage, we ranked the myeloid cells, vs. 5 ± 1% at day 0, 10.8-fold increase). top five uniquely upregulated genes by fold change followed Gating on CD45 + CD11b + Ly6G- cells, F4/80 and by p value, and the top five ranked by p value followed by Ly6C further divided this population into resident mac- fold change (Fig. 3c). All of top ten ranked genes had previ- high low rophages (F4/80 Ly6C ) and infiltrating monocytes ously been associated with macrophages: Pcdha4 [15], Orm1 low high (F4/80 Ly6C ). Resident macrophages were signifi- [38], Pappa [78], Slfn4 [82], Inhba [37], Cd9 [74], Tnip1 cantly decreased in the infarct (14 ± 1% of at day 1 vs. [95], Apobr [79], Bcl2l11 (Bim) [30], and Gpr132 [4]. Day 78 ± 3% at day 0, 5.7-fold decrease) while monocytes were 1 macrophages displayed a unique signaling profile by prin- significantly increased (74 ± 1% at day 1 vs. 12 ± 2% at cipal component analysis compared to the other three time day 0). These results are consistent with what others have points (Fig. 4a), with expression of genes associated with 1 3 -log10(p) CD45 Ly6G F4/80 % of mono/mac % of total cells % of myeloid cells 26 Page 8 of 18 Basic Research in Cardiology (2018) 113:26 p<0.05 genes Principal Component Analysis (a) Day 0Day 1Day 3Day 7 Day 0 Mapk7 Day 1 Ilr2 D3 Day 3 Il1rap Mapk6 Day 7 Mapk8 Il1rn Stat5a Stat5b Stat5b Il1rl2 Nfkb2 Tnfrsf1b Nfkb1 -400 -200 0 200 400 Socs2 PC PC1 1( (68.4 68 4% %) ) Il1rapl1 Non-Significant (p>0.05) Stat4 Mapk9 Day 0Day 1Day 3Day 7 (b) Socs6 Stat1 Stat2 Mapk3 Il1r Il1r1 1 Tf Tgfbr1 b1 Socs1 Il4ra Socs3 Mapk14 Stat3 Mapk11 Tgfbr2 Akt3 Stat6 Tnfrsf1a Map pk4 Mapk12 Mapk12 Socs5 Mapk13 Socs7 Akt2 Socs4 Mapk10 Il1rl1 Akt1 Akt1s1 Tgfbr3 Mapk1 Fig. 4 Post-MI macrophage signaling profiles. a Principal component to days 0, 3, and 7. b Heat maps grouped by significance (left, non- analysis of genes involved in ubiquitous signaling pathways indicates significant genes (p > 0.05); right, significant genes (p < 0.05) by that day 1 macrophages display a unique signaling profile compared unpaired two-tailed t test) IL-1, TNF, NF-κB, MAPK, STAT5, and SOCS2 signaling (Fig. 5b). In line with increased phagocytic capacity, IPA pathways (Fig. 4b). analysis indicated that phagosome maturation was an upreg- ulated pathway at day 3 (20/148 genes; p = 0.003). Likewise, Day 3 MI macrophages are phagocytic, proliferative, the phagosome maturation GO pathway was increased at and display a metabolic reprogramming signature day 7 post-MI (11/148 genes; p = 0.02), albeit to a lesser extent than day 3. Phagosome formation, but not matura- Genes differentially expressed at day 3 are displayed by tion, was an upregulated pathway at day 1 (15/130 genes; volcano plot, with representative genes highlighted in yel- p = 0.0003). While day 3 proliferation was not significantly low (Fig. 5a). The major upregulated GO processes were different from day 0 (Fig.  5b), resident (day 0) macrophages related to mitochondrial function, including mitochon- have been shown to be proliferative [22]. Further, GO pro- drial translation termination (p = 8.4E−10) and elongation cesses such as DNA replication initiation (p = 0.002), DNA (p = 8.4E−10), electron transport (p = 4.4E−11), respiratory replication (p = 0.01), and mitotic cell cycle (p = 0.01) were chain complex I assembly (p = 8.4E−10), cristae formation upregulated, indicating a proliferative signature at day 3. (p = 2.9E−5), and mitochondrial ATP synthesis-coupled pro- Genes associated with phagocytosis and proliferation are ton transport (p = 7.1E−5). Downregulated GO processes displayed in Supplemental Fig. 7. Macrophage turnover was included ECM organization (p = 2.4E−7), and cell migration assessed by F4/80 antibody injection at day 1 post-MI; only (p = 0.0002). By IPA analysis, oxidative phosphorylation 5.2 ± 0.4% of macrophages at day 3 post-MI stained positive (42/109 genes; p = 2.5E−21) and mitochondrial dysfunc- for the F4/80 antibody (Fig. 5c), indicating rapid turnover as tion (51/171 genes; p = 8.0E−20) were the top upregulated reported by others [36]. canonical pathways, while leukocyte extravasation (38/211 To define the day 3 post-MI macrophage, we ranked the genes; p = 2.9E−9) and agranulocyte adhesion and diapede- top five uniquely upregulated genes by fold change followed sis (35/191 genes; p = 7.7E−9) were the major downregu- by p value, and the top five ranked by p value followed by lated pathways (Supplemental Fig. 6). fold change (Fig. 5d). All of the top ten ranked genes had We assessed in vivo macrophage phagocytosis, prolifera- previously been associated with macrophages: Klra22 [59], tion, and turnover. Macrophages from post-MI day 3 showed Ffar4 [91], Tdgf1 [72], Sarm1 [19], Il24 [73], Dek [50], significantly increased phagocytosis compared to day 0 Rab18 [21], Vdac1 [7], Dnajc15 [55], and Tmsb4x [83]. 1 3 PC2 (21 %) -400 -200 02 0 00 0 400 Basic Research in Cardiology (2018) 113:26 Page 9 of 18 26 UpregulatedDownregulated (a) Mitochondrial translation termination Extracellular matrix organization Mrpl34 Ccl7 Mitochondrial translation elongation Muscle contraction Ccl11 Sdhb Mitochondrial electron transport Cell migration Col4a1 Col4a2 Cell-cell signaling Atp5l Mitochondrial complex I assembly Sparc Mrps10 Neutrophil degranulation Muscle filament sliding -1 10 0-50 50 5 5 10 10 0 0 20 20 40 40 60 60 80 80 0 0 10 10 20 20 30 30 40 40 50 50 log2FC Combined ScoreCombined Score (b) No injection Day 3 Phagocytosis (c) Day 0 0.28±0.11 100 μm F4/80 F4/80Ab F4/80Ab injectio injection n Proliferatio Proliferation n 20 20 Day 0 Day 3 p=0.19 5.23±0.43 0 0 100 μm F4/80 Days post-MI (d) Top 5 Genes Ranked by fold change then p-value Top 5 Genes Ranked by p value then fold change Expression in Expression in Gene FC pRole other cardiac cells Gene pFCRole other cardiac cells MHC-binding lectin-like Secreted Klra22 7.2E+400.0002 receptor Natural killer cells Dek 6E-7 2.1 chemotacticfactor Ubiquitous Omega-3 FA receptor; Fibroblasts, Rab18 6E-6 2.3phagocytosisUbiquitous Ffar430366 0.035 anti-inflammatory myocytes Calcium signaling; Tdgf1 1770 0.027TGF-betasignaling unknown Vdac11E-52.5 apoptosisUbiquitous Regulation of CCL5 Transcriptional Sarm1458 0.047 production Ubiquitous Dnajc15 3E-5 2.6 response to IFNγUbiquitous IL-10family; Inflammation Il24 2010.036 inflammatory response Th2 cells Tmsb4x 4E-5 2.1 resolution Ubiquitous Fig. 5 Day 3 post-MI macrophages showed a phagocytic, prolifera- Phagocytic capacity significantly increased at day 3 post-MI, whereas tive, and metabolic reprogramming profile. a Volcano plot with rep- proliferation was similar to day 0. c In vivo turnover. At day 3 post- resentative genes highlighted and enrichment analysis of upregulated MI, ~ 5% of macrophages were remaining from day 1. d Top five and downregulated genes. All values are normalized to day 0 no MI upregulated day1 genes ranked by fold change (left) and p value controls and representative genes are highlighted. b Representative (right). *p < 0.05 versus day 0 images of phagocytic (top) and proliferative (bottom) macrophages. of ECM genes downregulated at days 1 and 3 and either Day 7 MI macrophages display a pro‑reparative upregulated or downregulated at day 7 compared to day 0 signature (Fig. 6b; Supplemental Table 2). To define the day 7 post-MI macrophage, we ranked the Genes differentially expressed at day 7 post-MI are displayed top five uniquely upregulated genes by fold change followed by volcano plot, with representative genes highlighted in by p value, and the top five ranked by p value followed by yellow (Fig. 6a). The major upregulated GO processes were fold change (Fig. 6c). All of the top ten ranked genes had related to ECM remodeling (Supplemental Table 2), includ- previously been associated with macrophages: Pcdha7 [15], ing ECM disassembly (p = 0.0004) and collagen fibril organ- Klra23 [59], Tac4 [3], Phgk1 [16], Slc3a1 [29], Odc1 [20], ization (p = 0.01), and inflammatory response (p = 0.004) Spsb1 [56], Crem [46], Sptlc2 [5], and Abca1 [77]. was the major downregulated GO process. By IPA analysis, Of note, Col1a1, Col3a1, and Postn, which contrib- inhibition of matrix metalloproteinases (MMPs) was a major ute to post-MI scar formation and ECM stiffness, were upregulated canonical pathway (7/39 genes, p = 0.0004), all elevated in day 7 macrophages, while Col4a4 and while agranulocyte adhesion and diapedesis was a major Lama2, which contribute to basement membrane forma- downregulated pathway (34/192 genes, p = 4.6E−12; Sup- tion, and Mmp15, were decreased. Using triple in  situ plemental Fig. 6). Clustering analysis identified a number 1 3 -log10(p) 0 Phagocytosis Proliferation o % % of total cells s % of total cells Count Count 26 Page 10 of 18 Basic Research in Cardiology (2018) 113:26 Upregulated Downregulated Ccl2 (a) Inflammatory Inflammatory respons response e Et Extracell llulM lar Mati trix Di Disassembl bly Col1a1 Muscle filament sliding Hyaluronan catabolic process Tnf Postn Cell surface receptor signaling Ccl3 Platelet degranulation Lox Cxcr2 Cardiac muscle contraction Synapse Assembly Positive regulation of ERK 1 and 2 Collagen Fibril Organization -10-50 5100 20 40 60 05 10 15 20 25 log2FC Combined ScoreCombined Score (b) Col1a1 Col3a1 Postn Upregulated # # $ $ * $ $ 6 6 60 60 # # * * 150 * * 200 * * 4 40 20 50 0 0 1 1 3 3 7 7 0 0 1 1 3 3 7 7 0 0 1 1 3 3 7 7 0 0 1 1 3 3 7 7 Downregulated Col4a4 Lama2 Mmp15 1.5 1.0 10 1.0 ** 0 05 .5 2 2 5 0.5 * * * * * * * * * * * * * * * * * * 0 0 0 13 71 0 37 00 1 3 7 1 3 7 Days post-MI Days post-MIDays post-MI Days post-MI (c) To Top5 p5 Gene GenesR sRanked ankedb by y Fol Fold dC Change hange T5 Top 5 GR Genes Rankd kedb by p vallue Expression in Expression in Gene pFCMacrophageRole other cardiac cells Gene FC pMacrophage Role other cardiac cells cell adhesion, cell- Ornithine metabolism; Pcdha7 42888 0.038 cell recognition unknown Odc1 1E-5 3.1 M2 polarization ubiquitous Negative regulation of MHC-binding Spsb13E-5 3.4 iNOS unknown Klra23 18750.024 receptor Natural killer cells cAMPsignaling; M2 Inhibition of cytokine B cells, dendritic Crem 7E-5 2.3 polarizationmyocytes Tac4 79 0.012 signaling cells inflammasome Glycolytic Sptlc2 Sptlc2 9E 9E-5 5 21 2.1 activation activation myocytes myocytes Ph Phgk1 k1 32 32 0 0.043 043 metb taboli lism myocyttes Cholesterol efflux, Mitochondrial Abca11E-4 2.9 anti-inflammatory unknown Slc3a1 25 0.005 respiration unknown Fig. 6 Day 7 post-MI macrophages showed a pro-reparative profile. genes upregulated at day 7 compared to day 1 and 3 post-MI (top) a Volcano plot with representative genes highlighted and enrichment or remained downregulated at day 7 (bottom). c Top five upregulated analysis of upregulated and downregulated genes. All values are nor- day 7 genes ranked by fold change (left) and p value (right). *p < 0.05 # $ malized to day 0 no MI controls and representative genes are high- versus day 0, p < 0.05 versus day 1, p < 0.05 versus day 3 lighted. b ECM genes that clustered together. Fold change values for hybridization, we assessed the number of Emr1 + cells Defining macrophage polarization phenotypes (i.e., macrophages), Acta2 + cells (i.e., fibroblasts), and Em r1 + C cr 2 + c el ls (in f i ltrat in g mo no cy tes a nd m ac- In addition to the prototypical M1 and M2 markers, we rophages) expressing Col1a1 and Postn mRNA in the investigated the most prominent genes uniquely upregulated infarct region of LV tissue from day 7 post-MI mice at each post-MI day (Figs. 3c, 5d, 6c). Genes expressed only (Fig. 7a–c). Macrophages in the infarct region heterogene- at day 0 (i.e., no or very low expression at all post-MI days) ously expressed both Col1a1 and Postn mRNA (Fig. 7d). were ranked to assess day 0 (resident cardiac macrophage) Of the Emr1 + macrophages, about half expressed Col1a1 markers. Day 0 genes were ranked by FPKM then ANOVA and about a quarter expressed Postn. ECM gene expres- p value or by ANOVA p value then FPKM (Supplemental sion was found only in Ccr2 + cells, which have been Table 3). The top ranked day 0 genes by FPKM included shown to drive post-MI inflammation and impair ECM Atf3 [23], Cbr2, Folr2 [42, 58], Actr3 [90], and Cd81 [76]. formation [45]. The majority of Acta2 + cells expressed The top ranked day 0 genes by p value included Cfh [2], Col1a1 (70%), while 25% also expressed Postn. Lilra5 [48], Cd209f [69], Cmah [60], and Tln2 [66]. 1 3 Fold d Change Fold Change e -log10(p) FP PKM FPKM Basic Research in Cardiology (2018) 113:26 Page 11 of 18 26 (a) (b) PostnCol1a1Emr1DAPI PostnCol1a1Acta2DAPI (c) (d) Other PostnCcr2Emr1DAPI 16% Emr1+ 47 47± ±7% 7% Acta2+ 36±1% Ccr2+ Ccr2- Col1a1 Col1a1- Col1a1+ Col1a1+ Cl Col1 11 a1- Col1a1 Col1a1- Postn- Postn+ Postn- Postn- 30±8% 25±12% 25±10% 28±4% Ccr2- Ccr2- Col1a1+ Col1a1+ Col1a1+ Col1a1+ Col1a1+ Col1a1+ Postn- Postn+ Postn- 45±10% 26±6% 21±12% Fig. 7 Macrophage expression of Col1a1 and Postn in LV infarct Col1a1 (yellow, Cy3) and Postn (magenta, FITC). Emr1 and Ccr2 region. Numbers of cells expressing Postn and Col1a1 mRNA in LV positive cells (c; yellow, Cy3) were assessed for expression of Postn infarct tissue at day 7 post-MI were determined by in  situ hybridi- (magenta, FITC). d Pie charts displaying proportion of LV infarct zation. Both Emr1 (a; red, Cy5, macrophages) and Acta2 (b; green, Emr1 + (macrophages) and Acta2 + (fibroblasts) cells expressing Cy5, fibroblasts) positive cells were assessed for expression of Col1a1 and Postn. Nuclei are stained with DAPI (cyan) As cell specificity is a good criterion for a marker, we (Cbr2, Folr2, Cfh, Lilra5, Cd209, Pcdha4, Slfn4, Tdgf1, compared the ranked genes to the literature to determine Pcdha7, Slc3a1, Spsb1, and Abca1). Based on these criteria, whether genes were distinctly expressed in macrophages Table 1 lists candidate markers that may uniquely identify relative to other cardiac MI-relevant cell types (myocytes, cardiac macrophages from each time after MI, including endothelial cells, fibroblasts, neutrophils, and lymphocytes). directed analysis of known M1, M2, and macrophage mark- While several genes are ubiquitously expressed across cell ers and unbiased evaluation of the top ranked genes for each types (Atf3, Cd81, Cmah, Tln2, Orm1, Cd9, Tnip1, Sarm1, time examined. Dek, Dnajc15, Tmsb4x, Rab18, Vdac1, and Odc1), there We assessed expression of genes involved in circadian were genes with potential macrophage restricted expression rhythm regulation, as time of surgery and cell isolation all Table 1 Candidate post-MI Day 0 Day 1 Day 3 Day 7 macrophage markers a a Il1b, Nos2, Arg1, Chi3l3, Informed Ccl17, Ifng, Il12a, Lgals3, Vegfa Cd14, Sparc, a a a a Cd14 Retnla, , Lgals3, VegfaLox, Lgals3 , Cd163 Vegfa Unbiased Cbr2, Folr2, Pcdha4, Slfn4, Apobr, Gpr132 Klra22, Tdgf1, Il24 Spsb1, Abca1, Cfh, Pcdha7, Lilra5, Klra23, Cd209f Tac4, Slc3a1 Unbiased candidates are genes based on our ranking system; informed candidates are genes previously used as macrophage markers Marker for more than one time point 1 3 26 Page 12 of 18 Basic Research in Cardiology (2018) 113:26 were performed between 8 a.m. and 12 noon (Supplemen- Macrophage metabolism is a reflection of and also a tal Fig. 8). No differences in Clock, Per2, or Nr1d1 were contributor to polarization status, as pro-inflammatory M1 observed, while Nr1d2, Nfil3, Arnt1, and Cry1 were dif- macrophages rely on glycolysis, while reparative M2 mac- ferentially expressed after MI. These results indicate that rophages use oxidative phosphorylation [33, 57]. Indeed, MI may compromise circadian rhythm regulation in mac- our results show that by day 3 post-MI, genes related to rophages, which has been implicated in release of cytokines mitochondrial ATP generation and oxidative phosphoryla- and other macrophage functions [84]. tion were upregulated, indicating metabolic reprogramming over the post-MI remodeling continuum. While studies have demonstrated that metabolism influences macrophage polarization, the role of metabolism in post-MI macrophage Discussion polarization has not been extensively evaluated [16, 33]. Our results indicate the metabolic shift of the day 3 macrophage The goal of this study was to map the continuum of changes may be an indicator of wound repair status. that occur in cardiac macrophages over the first week of MI. At day 3 post-MI, macrophages downregulate many of The key findings were: (1) cardiac macrophages undergo the inflammatory genes elevated at day 1, including Il1b, continual and distinct transcriptomic changes at post-MI while upregulating others, including Il12a, Pf4, and Il24. days 1, 3, and 7; and (2) day 1 macrophages have a pro- In addition, day 3 macrophages showed increased phago- inflammatory profile, day 3 macrophages have a phagocytic, cytic capacity and a return in proliferation. Phagocytosis proliferative, and metabolic reprogramming profile, and day of necrotic and apoptotic cells is a critical role of post-MI 7 macrophages have a reparative signature that includes macrophages and is required for the transition from a pro- to expression of extracellular matrix remodeling genes that anti-inflammatory environment and polarization towards a contribute to scar formation (e.g., collagen 1a1 and peri- reparative phenotype [18, 32, 65]. Resident cardiac mac- ostin). While macrophages in the post-MI heart have been rophages have a basal phagocytic rate, while day 1 MI mac- characterized in terms of cell surface markers, phenotypes, rophages have increased phagocytic capacity [22, 54]. Our and origins, the full transcriptome has not been mapped in results indicate that phagocytic capacity is elevated at day 3, detail [9, 12–14, 18, 22]. which coincides with downregulation of many of the inflam- Macrophages underwent distinct gene expression changes matory genes that were elevated at day 1. Following MI, reflecting shifts in phenotype over the first week of post- macrophage phagocytosis occurs in two sequential steps: ini- MI LV remodeling. The fact that day 7 normalized values tial phagocytosis of necrotic myocytes, followed by efferocy - showed distinction in expression from day 3 normalized val- tosis of apoptotic neutrophils [68]. The increased phagocytic ues (e.g., Col1a1, Col3a1, and Postn genes) indicates there capacity that we observed at day 3 post-MI may represent is still a continuum of changes occurring, at a slower kinetic efferocytosis of apoptotic neutrophils, whose numbers in the rate than day 0 through days 1 and 3. Principal component infarct concomitantly begin to decline after day 3. As resi- analysis indicates that day 1 macrophages are markedly dif- dent cardiac macrophages are proliferative, the similarity in ferent in phenotype than all other time points, whereas day proliferation rates between day 0 and day 3 macrophages 3 and 7 macrophages are closer to a day 0 phenotype, indi- indicates that by day 3 proliferation of macrophages is re- cating the transition to a new homeostatic-like phenotype. initiated, consistent with a previous report [22]. Consistent with the literature, our results indicate that At day 7 post-MI, macrophages exhibited a reparative the early day 1 post-MI macrophage shows a strong pro- phenotype, indicated by upregulation of ECM organization inflammatory, matrix-degrading phenotype [22]. In addition genes. Interestingly, genes typically considered fibroblast- to pro-inflammation, enrichment analysis indicated that gly - specific were upregulated in day 7 macrophages, including colysis and cellular response to hypoxia were upregulated Col1a1 and Postn. Recent studies on aging hearts have char- in the day 1 macrophage, suggesting an adaptation to the acterized a myeloid-derived cardiac fibroblast population, hypoxic environment of the early infarct. Hypoxia induces which is derived from M2a macrophages [81]. Our study activation of the hypoxia-inducible factor (HIF)-1α pathway, suggests that post-MI macrophages may assume a fibroblast- which turns on pro-inflammatory gene expression, as well as like phenotype, which remains to be fully investigated. metabolic reprogramming towards glycolysis [1, 9]. Hif1a While macrophages indirectly contribute to ECM forma- was high in day 1 and 3 post-MI macrophages and returned tion by releasing paracrine factors that stimulate cardiac towards day 0 values by day 7. While day 1 macrophages fibroblasts, macrophages can directly secrete ECM proteins upregulate pro-inflammatory genes early in response to MI, [6]. Indeed, fibronectin expression was highly upregulated several genes that inhibit these pathways are also highly in day 1 macrophages and returned towards baseline by induced (e.g., Slfn4, Cd9, Tnip1, and Gpr132), most likely day 7. Collagens type VI and VIII were also elevated at to exert negative feedback and limit excessive inflammation. day 7 post-MI and have been shown to increase in response 1 3 Basic Research in Cardiology (2018) 113:26 Page 13 of 18 26 to anti-inflammatory stimuli (e.g., TGF-β1 and IL-4) and reported to be undetectable in the infarcted LV, was decrease in response to pro-inflammatory stimuli (e.g., LPS decreased in macrophages at day 1 post-MI, increased at day and IFN-γ) [71, 89]. Our results reveal a previously unde- 3, and returned to day 0 values by day 7 [17]. Expressions fined role for macrophages in directly contributing ECM of the anti-inflammatory cytokines Tgfb1 and Il10 were proteins to the infarct scar. unchanged in macrophages following MI. One explanation is Macrophage polarization towards M1 and M2 subtypes that other cell types are the major source of these cytokines; has been well-defined in vitro for some, but not all stimuli. a second is that regulation occurs at the post-translational Post-MI in vivo macrophage polarization is less clear, as stage. Tgfb1 is regulated at the post-translational stage, as there are a number of factors over a broad spectrum within pre-formed TGFβ1 protein is produced and sequestered in the infarct environment that co-stimulate to generate an the ECM by binding to latent TGFβ binding protein. Bio- overall macrophage phenotype. After MI, the inflamma- availability is regulated by proteases (including MMPs) that tory M1 phenotype predominates early (day 1), whereas cleave the binding protein. Ym1 (Chi3l3) and Fizz1 (Retnla) the M2 reparative phenotype later becomes the major have been implicated as M2 macrophage markers [27, 67]. subtype (day 7). While our enrichment analyses indicated Ym1/Chi3l3, which was highly upregulated at days 1 and 3 a pro-inflammatory day 1 post-MI phenotype and an anti- post-MI and downregulated at day 7, is a lectin that binds inflammatory/reparative phenotype at day 7 post-MI, gene heparan-type glycosaminoglycans to influence inflammation expression for typical M1 and M2 markers did not reflect the [69]. Fizz1/Retnla was highly decreased in macrophages at in vitro scenario. For example, while Nos2 (an M1 marker) all days post-MI, and its role in inflammation is less clear, as was increased 5-fold at day 1, Arg1 (an M2 marker) was it possesses both pro- and anti-inflammatory roles depending increased over 900-fold. Nos2 returned to baseline at days 3 on context [69]. Our results support the growing consensus and 7, while Arg1 remained elevated. The concept of argin- to abandon the M1 and M2 nomenclature, as pro-inflam- ase-1 as an M2 marker has been challenged, as a variety of matory day 1 macrophages do not fully display typical M1 M1 stimuli induce arginase-1 expression [27]. It is possible features nor do pro-reparative day 7 macrophages display that ratios or combinations of profiles dictate net phenotype, typical M2 features [53]. with fine tuning occurring by fluxes in both pro- and anti- The post-MI macrophage polarization time contin- inflammatory constituents. uum map is shown in Fig.  8. Day 1 macrophages exhibit Our results indicate that IL-1β and IFN-γ are major a NF-κB and MAPK signaling repertoire, which are acti- macrophage-derived cytokines that promote the pro-inflam- vated in response to toll-like receptor signaling pathways matory environment early post-MI, as Il6 was unchanged and induce downstream pro-inflammatory molecules [27]. and Tnf was decreased. Il12a, which has previously been At day 3, the sirtuin and AMPK axis, which promote the Fig. 8 Map of macrophage Resident cM(D1) cM(D3) cM(D7) polarization over the MI time macrophage Monocyte continuum. Post-MI mac- rophage profiles in terms of inputs, signaling, outputs, and cell physiology over the three major post-MI remodeling Day 0 13 7 phases (inflammation, prolif- Days post-MI eration, and scar maturation). TIMP-1, ADAMs, DAMPs damage-associated Oxidative stress IL-1α/β collagen, Inputs* molecular patterns, ECM extra- Changes in metabolism DAMPs matricellularproteins cellular matrix TNFR1B, Sirtuins, AMPK, Integrins, TGFBR1, IL1R, TLRs metabolic/redox Signaling IL4Ra NF-kB, MAPK signals STAT3 IL-1α/β, IFN-γ, pro- IL-12, PF4, IL24, Collagen, ECM Outputs inflammatory cytokines reparative factors cytokines/chemokines metabolic outputs Cell Inflammation ECM synthesis and Phagocytosis, Proliferation Physiology ECM Breakdown Metabolic Reprogramming deposition Inflammation Remodeling Proliferation Phenotypes Maturation 1 3 26 Page 14 of 18 Basic Research in Cardiology (2018) 113:26 shift from glycolysis to mitochondrial oxidative phospho- Although fresh cardiac tissue is difficult to obtain from mice rylation and polarization towards a reparative phenotype, with cardiac rupture due to its spontaneous nature, serum may be important regulators of macrophage signaling [51, factors (i.e., factor XIII) or circulating monocyte markers 86]. In day 7 macrophages, STAT3/4, TGFBR1, and IL4Ra could be used to predict survival versus rupture [52]. were important pathways regulating the reparative pheno- In conclusion, macrophages show distinct transcriptomic type. Both IL-6 and IL-10 can promote M2 polarization profiles at different time points over the first week of post-MI through STAT3 activation [85]. Signaling through TGFBR1 wound healing. Metabolic shifts in post-MI macrophages influences macrophage polarization by inhibiting TNF-α may regulate polarization and cell physiology status and and iNOS expression and strongly induces collagen I and remain to be fully investigated. Of note, macrophages may III expression in fibroblasts [8 , 35, 72]. IL-4Ra, which was play a more direct role in post-MI ECM synthesis and increased at day 7 post-MI, drives M2 polarization and is remodeling than previously thought. Our findings provide critical for macrophage regulation of collagen remodeling novel insights into the potential mechanisms and pathways in skin wound healing [31]. Itga11, which was elevated at that regulate macrophage physiology during the inflamma- day 7 post-MI, encodes for a collagen-binding integrin that tory, granulation, and maturation phases of post-MI cardiac regulates intracellular collagen production [64]. remodeling. Overall, our work suggests that attention to tem- Resident (day 0) cardiac macrophages are derived from poral profiles should be given when considering therapeutic a different source than infiltrating monocyte-derived mac- strategies that alter macrophages. rophages. While the different origins of these two popula- Acknowledgements We thank Dr. Rugmani P. Iyer, Dr. Mira Jung, tions of macrophages could account for transcriptomic and and Presley C. Cannon for their technical support. We acknowledge phenotypic differences, the purpose of our study was to funding from the American Heart Association under Award Num- determine differences in macrophages within the cardiac ber 15SDG22930009, from the National Institutes of Health under Award Numbers GM103328, GM103476, GM104357, GM114833, environment that reflect cell physiological roles in mainte- GM115428, HL051971, HL075360, HL105324, HL129823, and nance and wound healing, rather than differences in circu- HL136438, and from the Biomedical Laboratory Research and Devel- lating monocytes before MI compared to infiltrating tissue opment Service of the Veterans Affairs Office of Research and Devel- monocytes after MI. Further, differentially expressed genes opment under Award Numbers 5I01BX000505 and IK2BX003922. T.C.F. is funded by an Institute Strategic Programme Grant funding were largely distinct across days 1, 3, and 7, indicating that from the Biotechnology and Biological Sciences Research Council the patterns of gene expression were not primarily reflect- (BB/J004227/1).The content is solely the responsibility of the authors ing cell source. Genes that were similarly differentially and does not necessarily represent the official views of any of the fund- expressed across all three of these days may be related to ing agencies. All authors have reviewed and approved the article. All authors have read the journal authorship agreement and policy on dis- either cell source or sustained expression. At last, the pro- closure of potential conflicts of interest and have nothing to disclose. cedure of isolating cardiac macrophages is different from isolating circulating monocytes, which could complicate Open Access This article is distributed under the terms of the Crea- interpretation of gene expression data. tive Commons Attribution 4.0 International License (http://creat iveco Transcriptomic analysis was performed on all cardiac mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- macrophages, rather than focusing on distinct macrophage tion, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the subpopulations. As this study is the first to report distinct Creative Commons license, and indicate if changes were made. changes in the macrophage transcriptome along the MI time course, our study provides the necessary framework for future studies to analyze distinct macrophage subpopula- tions within each time point. There was no indication that References pooling increased variation, as the coefficient of variation (CV) for individual cell counts was the smallest at day 7 1. Aarup A, Pedersen TX, Junker N, Christoffersen C, Bartels ED, Madsen M, Nielsen CH, Nielsen LB (2016) Hypoxia-inducible post-MI, whereas the CV for genes in Cluster 3 (Col1a1 and factor-1alpha expression in macrophages promotes development Postn) was highest at the time, indicating biological vari- of atherosclerosis. Arterioscler Thromb Vasc Biol 36:1782–1790. ability in response. Previous studies in our lab have observed https ://doi.org/10.1161/ATVBA HA.116.30783 0 biological variation in expression of these genes (Col1a1, 2. Aredo B, Li T, Chen X, Zhang K, Wang CX, Gou D, Zhao B, He Y, Ufret-Vincenty RL (2015) A chimeric Cfh transgene leads to CV = 140% and Postn, CV = 148%) in day 7 post-MI mac- increased retinal oxidative stress, inflammation, and accumulation rophages isolated from individual mice [28]. In the present of activated subretinal microglia in mice. Invest Ophthalmol Vis study, CV was 41% for Col1a1 and 46% for Postn, indicating Sci 56:3427–3440. https ://doi.org/10.1167/iovs.14-16089 there was no evidence of increased variation due to pooling. 3. Berger A, Tran AH, Paige CJ (2007) Co-regulated decrease of Neurokinin-1 receptor and Hemokinin-1 gene expression in mono- Our study focused only on surviving mice, which raises cytes and macrophages after activation with pro-inflammatory the question of changes in the macrophage transcriptome in surviving mice versus mice that undergo cardiac rupture. 1 3 Basic Research in Cardiology (2018) 113:26 Page 15 of 18 26 cytokines. J Neuroimmunol 187:83–93. https://doi.or g/10.1016/j. 16. Fleetwood AJ, Lee MKS, Singleton W, Achuthan A, Lee MC, jneur oim.2007.04.019 O’Brien-Simpson NM, Cook AD, Murphy AJ, Dashper SG, Reyn- 4. Bolick DT, Skaflen MD, Johnson LE, Kwon SC, Howatt D, olds EC, Hamilton JA (2017) metabolic remodeling, inflamma- Daugherty A, Ravichandran KS, Hedrick CC (2009) G2A defi- some activation, and pyroptosis in macrophages stimulated by ciency in mice promotes macrophage activation and atheroscle- Porphyromonas gingivalis and its outer membrane vesicles. rosis. Circ Res 104:318–327. https ://doi.or g/10.1161/CIRCR Front Cell Infect Microbiol 7:351. https ://doi.org/10.3389/fcimb ESAHA .108.18113 1 .2017.00351 5. Camell CD, Nguyen KY, Jurczak MJ, Christian BE, Shulman 17. Frangogiannis NG, Mendoza LH, Lindsey ML, Ballantyne CM, GI, Shadel GS, Dixit VD (2015) Macrophage-specific de novo Michael LH, Smith CW, Entman ML (2000) IL-10 is induced synthesis of ceramide is dispensable for inflammasome-driven in the reperfused myocardium and may modulate the reaction to inflammation and insulin resistance in obesity. J Biol Chem injury. J Immunol 165:2798–2808. https://doi.or g/10.4049/jimmu 290:29402–29413. https ://doi.org/10.1074/jbc.M115.68019 9 nol.165.5.2798 6. Chang MY, Chan CK, Braun KR, Green PS, O’Brien KD, Chait 18. Gombozhapova A, Rogovskaya Y, Shurupov V, Rebenkova M, A, Day AJ, Wight TN (2012) Monocyte-to-macrophage dif- Kzhyshkowska J, Popov SV, Karpov RS, Ryabov V (2017) Mac- ferentiation: synthesis and secretion of a complex extracellular rophage activation and polarization in post-infarction cardiac matrix. J Biol Chem 287:14122–14135. https ://doi.org/10.1074/ remodeling. J Biomed Sci 24:13. https ://doi.org/10.1186/s1292 jbc.M111.32498 8 9-017-0322-3 7. Chen H, Gao W, Yang Y, Guo S, Wang H, Wang W, Zhang S, 19. Gurtler C, Carty M, Kearney J, Schattgen SA, Ding A, Fitzger- Zhou Q, Xu H, Yao J, Tian Z, Li B, Cao W, Zhang Z, Tian Y ald KA, Bowie AG (2014) SARM regulates CCL5 production in (2014) Inhibition of VDAC1 prevents Ca(2)(+)-mediated oxida- macrophages by promoting the recruitment of transcription fac- tive stress and apoptosis induced by 5-aminolevulinic acid medi- tors and RNA polymerase II to the Ccl5 promoter. J Immunol ated sonodynamic therapy in THP-1 macrophages. Apoptosis 192:4821–4832. https ://doi.org/10.4049/jimmu nol.13029 80 19:1712–1726. https ://doi.org/10.1007/s1049 5-014-1045-5 20. Hardbower DM, Asim M, Luis PB, Singh K, Barry DP, Yang C, 8. Cheng R, Dang R, Zhou Y, Ding M, Hua H (2017) MicroRNA-98 Steeves MA, Cleveland JL, Schneider C, Piazuelo MB, Gobert inhibits TGF-beta1-induced differentiation and collagen pro- AP, Wilson KT (2017) Ornithine decarboxylase regulates M1 duction of cardiac fibroblasts by targeting TGFBR1. Hum Cell macrophage activation and mucosal inflammation via histone 30:192–200. https ://doi.org/10.1007/s1357 7-017-0163-0 modifications. Proc Natl Acad Sci USA 114:E751–E760. https :// 9. Cheng Y, Feng Y, Xia Z, Li X, Rong J (2017) omega-Alkynyl doi.org/10.1073/pnas.16149 58114 arachidonic acid promotes anti-inflammatory macrophage M2 21. Hashim S, Mukherjee K, Raje M, Basu SK, Mukhopadhyay A polarization against acute myocardial infarction via regulating (2000) Live Salmonella modulate expression of Rab proteins to the cross-talk between PKM2, HIF-1alpha and iNOS. Biochim persist in a specialized compartment and escape transport to lys- Biophys Acta 1862:1595–1605. https ://doi.org/10.1016/j.bbali osomes. J Biol Chem 275:16281–16288. https://doi.or g/10.1074/ p.2017.09.009 jbc.275.21.16281 10. Chiao YA, Dai Q, Zhang J, Lin J, Lopez EF, Ahuja SS, Chou 22. Heidt T, Courties G, Dutta P, Sager HB, Sebas M, Iwamoto Y, YM, Lindsey ML, Jin YF (2011) Multi-analyte profiling reveals Sun Y, Da Silva N, Panizzi P, van der Laan AM, Swirski FK, matrix metalloproteinase-9 and monocyte chemotactic protein-1 Weissleder R, Nahrendorf M (2014) Differential contribution of as plasma biomarkers of cardiac aging. Circ Cardiovasc Genet monocytes to heart macrophages in steady-state and after myo- 4:455–462. https://doi.or g/10.1161/CIRCGENETI CS.111.95998 1 cardial infarction. Circ Res 115:284–295. https: //doi.org/10.1161/ 11. DeLeon-Pennell KY, de Castro Bras LE, Iyer RP, Bratton DR, Jin CIRCR ESAHA .115.30356 7 YF, Ripplinger CM, Lindsey ML (2014) P. gingivalis lipopoly- 23. Hellmann J, Tang Y, Zhang MJ, Hai T, Bhatnagar A, Srivastava saccharide intensifies inflammation post-myocardial infarction S, Spite M (2015) Atf3 negatively regulates Ptgs2/Cox2 expres- through matrix metalloproteinase-9. J Mol Cell Cardiol 76:218– sion during acute inflammation. Prostaglandins Other Lipid 226. https ://doi.org/10.1016/j.yjmcc .2014.09.007 Mediat 116–117:49–56. https ://doi.or g/10.1016/j.pr os t ag lan 12. DeLeon-Pennell KY, Iyer RP, Ero OK, Cates CA, Flynn ER, Can- dins.2015.01.001 non PL, Jung M, Shannon D, Garrett MR, Buchanan W, Hall 24. Honold L, Nahrendorf M (2018) Resident and monocyte-derived ME, Ma Y, Lindsey ML (2017) Periodontal-induced chronic macrophages in cardiovascular disease. Circ Res 122:113–127. inflammation triggers macrophage secretion of Ccl12 to inhibit https ://doi.org/10.1161/circr esaha .117.31107 1 fibroblast-mediated cardiac wound healing. JCI Insight. https :// 25. Hulsmans M, Sam F, Nahrendorf M (2016) Monocyte and mac- doi.org/10.1172/jci.insig ht.94207 rophage contributions to cardiac remodeling. J Mol Cell Cardiol 13. Dutta P, Nahrendorf M (2015) Monocytes in myocardial infarc- 93:149–155. https ://doi.org/10.1016/j.yjmcc .2015.11.015 tion. Arterioscler Thromb Vasc Biol 35:1066–1070. https ://doi. 26. Iyer RP, Patterson NL, Zouein FA, Ma Y, Dive V, de Castro Bras org/10.1161/ATVBA HA.114.30465 2 LE, Lindsey ML (2015) Early matrix metalloproteinase-12 inhi- 14. Epelman S, Lavine KJ, Beaudin AE, Sojka DK, Carrero JA, Cal- bition worsens post-myocardial infarction cardiac dysfunction deron B, Brija T, Gautier EL, Ivanov S, Satpathy AT, Schilling by delaying inflammation resolution. Int J Cardiol 185:198–208. JD, Schwendener R, Sergin I, Razani B, Forsberg EC, Yokoyama https ://doi.org/10.1016/j.ijcar d.2015.03.054 WM, Unanue ER, Colonna M, Randolph GJ, Mann DL (2014) 27. Jablonski KA, Amici SA, Webb LM, Ruiz-Rosado Jde D, Popo- Embryonic and adult-derived resident cardiac macrophages are vich PG, Partida-Sanchez S, Guerau-de-Arellano M (2015) Novel maintained through distinct mechanisms at steady state and dur- markers to delineate murine M1 and M2 macrophages. PLoS ONE ing inflammation. Immunity 40:91–104. https://doi.or g/10.1016/j. 10:e0145342. https ://doi.org/10.1371/journ al.pone.01453 42 immun i.2013.11.019 28. Jung M, Ma Y, Iyer RP, DeLeon-Pennell KY, Yabluchanskiy A, 15. Feig JE, Vengrenyuk Y, Reiser V, Wu C, Statnikov A, Aliferis Garrett MR, Lindsey ML (2017) IL-10 improves cardiac remod- CF, Garabedian MJ, Fisher EA, Puig O (2012) Regression of eling after myocardial infarction by stimulating M2 macrophage atherosclerosis is characterized by broad changes in the plaque polarization and fibroblast activation. Basic Res Cardiol 112:33. macrophage transcriptome. PLoS ONE 7:e39790. h tt ps : // do i.https ://doi.org/10.1007/s0039 5-017-0622-5 org/10.1371/journ al.pone.00397 90 29. Karunakaran D, Thrush AB, Nguyen MA, Richards L, Geoffrion M, Singaravelu R, Ramphos E, Shangari P, Ouimet M, Pezacki 1 3 26 Page 16 of 18 Basic Research in Cardiology (2018) 113:26 JP, Moore KJ, Perisic L, Maegdefessel L, Hedin U, Harper ME, 42. London A, Cohen M, Schwartz M (2013) Microglia and mono- Rayner KJ (2015) Macrophage mitochondrial energy status regu- cyte-derived macrophages: functionally distinct populations that lates cholesterol efflux and is enhanced by anti-miR33 in athero- act in concert in CNS plasticity and repair. Front Cell Neurosci sclerosis. Circ Res 117:266–278. https ://doi.org/10.1161/CIRCR 7:34. https ://doi.org/10.3389/fncel .2013.00034 ESAHA .117.30562 4 43. Ma Y, Halade GV, Zhang J, Ramirez TA, Levin D, Voorhees 30. Kirschnek S, Ying S, Fischer SF, Hacker H, Villunger A, Hochrein A, Jin YF, Han HC, Manicone AM, Lindsey ML (2013) Matrix H, Hacker G (2005) Phagocytosis-induced apoptosis in mac- metalloproteinase-28 deletion exacerbates cardiac dysfunction rophages is mediated by up-regulation and activation of the Bcl-2 and rupture after myocardial infarction in mice by inhibiting homology domain 3-only protein Bim. J Immunol 174:671–679. M2 macrophage activation. Circ Res 112:675–688. https ://doi. https ://doi.org/10.4049/jimmu nol.174.2.671org/10.1161/CIRCR ESAHA .111.30050 2 31. Knipper JA, Willenborg S, Brinckmann J, Bloch W, Maass T, 44. Ma Y, Mouton AJ, Lindsey ML (2018) Cardiac macrophage biol- Wagener R, Krieg T, Sutherland T, Munitz A, Rothenberg ME, ogy in the steady-state heart, the aging heart, and following myo- Niehoff A, Richardson R, Hammerschmidt M, Allen JE, Eming cardial infarction. Transl Res 191:15–28. https:/ /doi.org/10.1016/j. SA (2015) Interleukin-4 receptor alpha signaling in myeloid trsl.2017.10.001 cells controls collagen fibril assembly in skin repair. Immunity 45. Majmudar MD, Keliher EJ, Heidt T, Leuschner F, Truelove 43:803–816. https ://doi.org/10.1016/j.immun i.2015.09.005 J, Sena BF, Gorbatov R, Iwamoto Y, Dutta P, Wojtkiewicz G, 32. Lambert JM, Lopez EF, Lindsey ML (2008) Macrophage roles Courties G, Sebas M, Borodovsky A, Fitzgerald K, Nolte MW, following myocardial infarction. Int J Cardiol 130:147–158. Dickneite G, Chen JW, Anderson DG, Swirski FK, Weissleder R, https ://doi.org/10.1016/j.ijcar d.2008.04.059 Nahrendorf M (2013) Monocyte-directed RNAi targeting CCR2 33. Langston PK, Shibata M, Horng T (2017) Metabolism sup- improves infarct healing in atherosclerosis-prone mice. Circu- ports macrophage activation. Front Immunol 8:61. https ://doi. lation 127:2038–2046. https ://doi.or g/10.1161/CIRCU L A TIO org/10.3389/fimmu .2017.00061 NAHA.112.00011 6 34. Lavine KJ, Epelman S, Uchida K, Weber KJ, Nichols CG, Schil- 46. Mead JR, Hughes TR, Irvine SA, Singh NN, Ramji DP (2003) ling JD, Ornitz DM, Randolph GJ, Mann DL (2014) Distinct Interferon-gamma stimulates the expression of the inducible macrophage lineages contribute to disparate patterns of cardiac cAMP early repressor in macrophages through the activation recovery and remodeling in the neonatal and adult heart. Proc of casein kinase 2. A potentially novel pathway for interferon- Natl Acad Sci USA 111:16029–16034. https://doi.or g/10.1073/ gamma-mediated inhibition of gene transcription. J Biol Chem pnas.14065 08111 278:17741–17751. https ://doi.org/10.1074/jbc.M3016 02200 35. Lee YU, de Dios Ruiz-Rosado J, Mahler N, Best CA, Tara S, 47. Meschiari CA, Jung M, Iyer RP, Yabluchanskiy A, Toba H, Yi T, Shoji T, Sugiura T, Lee AY, Robledo-Avila F, Hibino Garrett MR, Lindsey ML (2018) Macrophage overexpression N, Pober JS, Shinoka T, Partida-Sanchez S, Breuer CK (2016) of matrix metalloproteinase-9 in aged mice improves diastolic TGF-beta receptor 1 inhibition prevents stenosis of tissue-engi- physiology and cardiac wound healing after myocardial infarc- neered vascular grafts by reducing host mononuclear phagocyte tion. Am J Physiol Heart Circ Physiol 314:H224–H235. https :// activation. FASEB J 30:2627–2636. h t t p s : / / d o i . o r g /1 0 . 1 0 9 6 /doi.org/10.1152/ajphe art.00453 .2017 fj.20150 0179R 48. Mitchell A, Rentero C, Endoh Y, Hsu K, Gaus K, Geczy C, 36. Leuschner F, Rauch PJ, Ueno T, Gorbatov R, Marinelli B, Lee McNeil HP, Borges L, Tedla N (2008) LILRA5 is expressed by WW, Dutta P, Wei Y, Robbins C, Iwamoto Y, Sena B, Chudnovs- synovial tissue macrophages in rheumatoid arthritis, selectively kiy A, Panizzi P, Keliher E, Higgins JM, Libby P, Moskowitz MA, induces pro-inflammatory cytokines and IL-10 and is regulated Pittet MJ, Swirski FK, Weissleder R, Nahrendorf M (2012) Rapid by TNF-alpha, IL-10 and IFN-gamma. Eur J Immunol 38:3459– monocyte kinetics in acute myocardial infarction are sustained by 3473. https ://doi.org/10.1002/eji.20083 8415 extramedullary monocytopoiesis. J Exp Med 209:123–137. https 49. Molawi K, Wolf Y, Kandalla PK, Favret J, Hagemeyer N, Frenzel ://doi.org/10.1084/jem.20111 009 K, Pinto AR, Klapproth K, Henri S, Malissen B, Rodewald HR, 37. Li D, Duan M, Feng Y, Geng L, Li X, Zhang W (2016) MiR- Rosenthal NA, Bajeno ff M, Prinz M, Jung S, Sieweke MH (2014) 146a modulates macrophage polarization in systemic juvenile idi- Progressive replacement of embryo-derived cardiac macrophages opathic arthritis by targeting INHBA. Mol Immunol 77:205–212. with age. J Exp Med 211:2151–2158. h t t p s : / / d o i . o rg / 1 0 . 1 0 8 4 / https ://doi.org/10.1016/j.molim m.2016.08.007jem.20140 639 38. Ligresti G, Aplin AC, Dunn BE, Morishita A, Nicosia RF (2012) 50. Mor-Vaknin N, Punturieri A, Sitwala K, Faulkner N, Legendre M, The acute phase reactant orosomucoid-1 is a bimodal regulator Khodadoust MS, Kappes F, Ruth JH, Koch A, Glass D, Petruzzelli of angiogenesis with time- and context-dependent inhibitory L, Adams BS, Markovitz DM (2006) The DEK nuclear autoanti- and stimulatory properties. PLoS ONE 7:e41387. https ://doi. gen is a secreted chemotactic factor. Mol Cell Biol 26:9484–9496. org/10.1371/journ al.pone.00413 87https ://doi.org/10.1128/MCB.01030 -06 39. Lindsey ML, Bolli R, Canty JM, Du XJ, Frangogiannis NG, Frantz 51. Moreira D, Rodrigues V, Abengozar M, Rivas L, Rial E, Laforge S, Gourdie RG, Holmes JW, Jones SP, Kloner R, Lefer DJ, Liao R, M, Li X, Foretz M, Viollet B, Estaquier J, Cordeiro da Silva A, Murphy E, Ping P, Przyklenk K, Recchia FA, Schwartz Longacre Silvestre R (2015) Leishmania infantum modulates host mac- L, Ripplinger CM, Van Eyk JE, Heusch G (2018) Guidelines for rophage mitochondrial metabolism by hijacking the SIRT1- experimental models of myocardial ischemia and infarction. Am J AMPK axis. PLoS Pathog 11:e1004684. https://doi.or g/10.1371/ Physiol Heart Circ Physiol. https ://doi.or g/10.1152/a jphe ar t.00335 journ al.ppat.10046 84 .2017 52. Nahrendorf M, Aikawa E, Figueiredo JL, Stangenberg L, van den 40. Lindsey ML, Kassiri Z, Virag JAI, de Castro Bras LE, Scherrer- Borne SW, Blankesteijn WM, Sosnovik DE, Jaffer FA, Tung CH, Crosbie M (2018) Guidelines for measuring cardiac physiology Weissleder R (2008) Transglutaminase activity in acute infarcts in mice. Am J Physiol Heart Circ Physiol. https://doi.or g/10.1152/ predicts healing outcome and left ventricular remodelling: impli- ajphe art.00339 .2017 cations for FXIII therapy and antithrombin use in myocardial 41. Lindsey ML, Saucerman JJ, DeLeon-Pennell KY (2016) Knowl- infarction. Eur Heart J 29:445–454. https://doi.or g/10.1093/eurhe edge gaps to understanding cardiac macrophage polarization fol-artj/ehm55 8 lowing myocardial infarction. Biochim Biophys Acta 1862:2288– 2292. https ://doi.org/10.1016/j.bbadi s.2016.05.013 1 3 Basic Research in Cardiology (2018) 113:26 Page 17 of 18 26 53. Nahrendorf M, Swirski FK (2016) Abandoning M1/M2 for a net- 67. Raes G, De Baetselier P, Noel W, Beschin A, Brombacher F, Has- work model of macrophage function. Circ Res 119:414–417. https sanzadeh GhG (2002) Differential expression of FIZZ1 and Ym1 ://doi.org/10.1161/CIRCR ESAHA .116.30919 4 in alternatively versus classically activated macrophages. J Leukoc 54. Nahrendorf M, Swirski FK, Aikawa E, Stangenberg L, Wurdinger Biol 71:597–602. https ://doi.org/10.1189/jlb.71.4.597 T, Figueiredo JL, Libby P, Weissleder R, Pittet MJ (2007) The 68. Rienks M, Carai P, Bitsch N, Schellings M, Vanhaverbeke M, healing myocardium sequentially mobilizes two monocyte sub- Verjans J, Cuijpers I, Heymans S, Papageorgiou A (2017) Sema3A sets with divergent and complementary functions. J Exp Med promotes the resolution of cardiac inflammation after myocardial 204:3037–3047. https ://doi.org/10.1084/jem.20070 885 infarction. Basic Res Cardiol 112:42. https ://doi.org/10.1007/ 55. Navasa N, Martin-Ruiz I, Atondo E, Sutherland JD, Angel Pas-s0039 5-017-0630-5 cual-Itoiz M, Carreras-Gonzalez A, Izadi H, Tomas-Cortazar J, 69. Roszer T (2015) Understanding the mysterious M2 macrophage Ayaz F, Martin-Martin N, Torres IM, Barrio R, Carracedo A, through activation markers and effector mechanisms. Mediat Olivera ER, Rincon M, Anguita J (2015) Ikaros mediates the Inflamm 2015:816460. https ://doi.org/10.1155/2015/81646 0 DNA methylation-independent silencing of MCJ/DNAJC15 70. Sager HB, Heidt T, Hulsmans M, Dutta P, Courties G, Sebas M, gene expression in macrophages. Sci Rep 5:14692. https ://doi. Wojtkiewicz GR, Tricot B, Iwamoto Y, Sun Y, Weissleder R, org/10.1038/srep1 4692 Libby P, Swirski FK, Nahrendorf M (2015) Targeting interleu- 56. Nishiya T, Matsumoto K, Maekawa S, Kajita E, Horinouchi T, kin-1beta reduces leukocyte production after acute myocardial Fujimuro M, Ogasawara K, Uehara T, Miwa S (2011) Regula- infarction. Circulation 132:1880–1890. https: //doi.org/10.1161/ tion of inducible nitric-oxide synthase by the SPRY domain- and CIRCU LATIO NAHA.115.01616 0 SOCS box-containing proteins. J Biol Chem 286:9009–9019. 71. Schnoor M, Cullen P, Lorkowski J, Stolle K, Robenek H, Troyer https ://doi.org/10.1074/jbc.M110.19067 8 D, Rauterberg J, Lorkowski S (2008) Production of type VI col- 57. O’Neill LA, Pearce EJ (2016) Immunometabolism governs den- lagen by human macrophages: a new dimension in macrophage dritic cell and macrophage function. J Exp Med 213:15–23. https functional heterogeneity. J Immunol 180:5707–5719. https :// ://doi.org/10.1084/jem.20151 570doi.org/10.4049/jimmu nol.180.8.5707 58. O’Shannessy DJ, Somers EB, Wang LC, Wang H, Hsu R (2015) 72. Schridde A, Bain CC, Mayer JU, Montgomery J, Pollet E, Expression of folate receptors alpha and beta in normal and can- Denecke B, Milling SWF, Jenkins SJ, Dalod M, Henri S, Mal- cerous gynecologic tissues: correlation of expression of the beta issen B, Pabst O, McL Mowat A (2017) Tissue-specific differ - isoform with macrophage markers. J Ovarian Res 8:29. https :// entiation of colonic macrophages requires TGFbeta receptor- doi.org/10.1186/s1304 8-015-0156-0 mediated signaling. Mucosal Immunol 10:1387–1399. https :// 59. Oakes JL, O’Connor BP, Warg LA, Burton R, Hock A, Loader J, doi.org/10.1038/mi.2016.142 Laflamme D, Jing J, Hui L, Schwartz DA, Yang IV (2013) Ozone 73. Shao J, Zhang B, Yu JJ, Wei CY, Zhou WJ, Chang KK, Yang enhances pulmonary innate immune response to a Toll-like recep- HL, Jin LP, Zhu XY, Li MQ (2016) Macrophages promote the tor-2 agonist. Am J Respir Cell Mol Biol 48:27–34. https ://doi. growth and invasion of endometrial stromal cells by downregu- org/10.1165/rcmb.2012-0187O C lating IL-24 in endometriosis. Reproduction 152:673–682. https 60. Okerblom JJ, Schwarz F, Olson J, Fletes W, Ali SR, Martin PT, ://doi.org/10.1530/REP-16-0278 Glass CK, Nizet V, Varki A (2017) Loss of CMAH during human 74. Suzuki M, Tachibana I, Takeda Y, He P, Minami S, Iwasaki T, evolution primed the monocyte-macrophage lineage toward a Kida H, Goya S, Kijima T, Yoshida M, Kumagai T, Osaki T, more inflammatory and phagocytic state. J Immunol 198:2366– Kawase I (2009) Tetraspanin CD9 negatively regulates lipopol- 2373. https ://doi.org/10.4049/jimmu nol.16014 71 ysaccharide-induced macrophage activation and lung inflamma - 61. Pinto AR, Godwin JW, Chandran A, Hersey L, Ilinykh A, Debu- tion. J Immunol 182:6485–6493. https://doi.or g/10.4049/jimmu que R, Wang L, Rosenthal NA (2014) Age-related changes in nol.08027 97 tissue macrophages precede cardiac functional impairment. Aging 75. Takahashi T, Tang T, Lai NC, Roth DM, Rebolledo B, Saito (Albany NY) 6:399–413. https ://doi.org/10.18632 /aging .10066 9 M, Lew WY, Clopton P, Hammond HK (2006) Increased car- 62. Pinto AR, Godwin JW, Rosenthal NA (2014) Macrophages in diac adenylyl cyclase expression is associated with increased cardiac homeostasis, injury responses and progenitor cell mobi- survival after myocardial infarction. Circulation 114:388–396. lisation. Stem Cell Res 13:705–714. https ://doi.org/10.1016/j.https ://doi.org/10.1161/CIRCU LATIO NAHA.106.63251 3 scr.2014.06.004 76. Takeda Y, He P, Tachibana I, Zhou B, Miyado K, Kaneko H, 63. Pinto AR, Paolicelli R, Salimova E, Gospocic J, Slonimsky E, Suzuki M, Minami S, Iwasaki T, Goya S, Kijima T, Kumagai Bilbao-Cortes D, Godwin JW, Rosenthal NA (2012) An abundant T, Yoshida M, Osaki T, Komori T, Mekada E, Kawase I (2008) tissue macrophage population in the adult murine heart with a Double deficiency of tetraspanins CD9 and CD81 alters cell distinct alternatively-activated macrophage profile. PLoS ONE motility and protease production of macrophages and causes 7:e36814. https ://doi.org/10.1371/journ al.pone.00368 14 chronic obstructive pulmonary disease-like phenotype in mice. 64. Popova SN, Barczyk M, Tiger CF, Beertsen W, Zigrino P, Aszodi J Biol Chem 283:26089–26097. https ://doi.or g/10.1074/jbc. A, Miosge N, Forsberg E, Gullberg D (2007) Alpha11 beta1 inte-M8019 02200 grin-dependent regulation of periodontal ligament function in the 77. Tang C, Liu Y, Kessler PS, Vaughan AM, Oram JF (2009) The erupting mouse incisor. Mol Cell Biol 27:4306–4316. https: //doi. macrophage cholesterol exporter ABCA1 functions as an anti- org/10.1128/MCB.00041 -07 inflammatory receptor. J Biol Chem 284:32336–32343. https:// 65. Prabhu SD, Frangogiannis NG (2016) The biological basis for doi.org/10.1074/jbc.M109.04747 2 cardiac repair after myocardial infarction: from inflammation to 78. Tang SL, Chen WJ, Yin K, Zhao GJ, Mo ZC, Lv YC, Ouy- fibrosis. Circ Res 119:91–112. https ://doi.org/10.1161/CIRCR ang XP, Yu XH, Kuang HJ, Jiang ZS, Fu YC, Tang CK (2012) ESAHA .116.30357 7 PAPP-A negatively regulates ABCA1, ABCG1 and SR-B1 66. Praekelt U, Kopp PM, Rehm K, Linder S, Bate N, Patel B, expression by inhibiting LXRalpha through the IGF-I-mediated Debrand E, Manso AM, Ross RS, Conti F, Zhang MZ, Harris RC, signaling pathway. Atherosclerosis 222:344–354. https ://doi. Zent R, Critchley DR, Monkley SJ (2012) New isoform-specific org/10.1016/j.ather oscle rosis .2012.03.005 monoclonal antibodies reveal different sub-cellular localisations 79. Tetens J, Heuer C, Heyer I, Klein MS, Gronwald W, Junge W, Oefner PJ, Thaller G, Krattenmacher N (2015) Polymorphisms for talin1 and talin2. Eur J Cell Biol 91:180–191. https ://doi. within the APOBR gene are highly associated with milk levels org/10.1016/j.ejcb.2011.12.003 1 3 26 Page 18 of 18 Basic Research in Cardiology (2018) 113:26 of prognostic ketosis biomarkers in dairy cows. Physiol Genom- (2014) Ly6C(low) and not Ly6C(high) macrophages accumulate ics 47:129–137. https://doi.or g/10.1152/physiolg enomics .00126 first in the heart in a model of murine pressure-overload. PLoS .2014 ONE 9:e112710. https ://doi.org/10.1371/journ al.pone.01127 10 80. Tomczyk M, Kraszewska I, Szade K, Bukowska-Strakova K, Mel- 89. Weitkamp B, Cullen P, Plenz G, Robenek H, Rauterberg J (1999) oni M, Jozkowicz A, Dulak J, Jazwa A (2017) Splenic Ly6C(hi) Human macrophages synthesize type VIII collagen in vitro and monocytes contribute to adverse late post-ischemic left ventricular in the atherosclerotic plaque. FASEB J 13:1445–1457. https://doi. remodeling in heme oxygenase-1 deficient mice. Basic Res Car -org/10.1096/faseb j.13.11.1445 diol 112:39. https ://doi.org/10.1007/s0039 5-017-0629-y 90. Wenzel J, Ouderkirk JL, Krendel M, Lang R (2015) Class I myo- 81. Trial J, Heredia CP, Taffet GE, Entman ML, Cieslik KA (2017) sin Myo1e regulates TLR4-triggered macrophage spreading, Dissecting the role of myeloid and mesenchymal fibroblasts in chemokine release, and antigen presentation via MHC class II. age-dependent cardiac fibrosis. Basic Res Cardiol 112:34. https Eur J Immunol 45:225–237. https ://doi.org/10.1002/eji.20144 ://doi.org/10.1007/s0039 5-017-0623-4 4698 82. van Zuylen WJ, Garceau V, Idris A, Schroder K, Irvine KM, Lattin 91. Williams-Bey Y, Boularan C, Vural A, Huang NN, Hwang IY, JE, Ovchinnikov DA, Perkins AC, Cook AD, Hamilton JA, Hert- Shan-Shi C, Kehrl JH (2014) Omega-3 free fatty acids suppress zog PJ, Stacey KJ, Kellie S, Hume DA, Sweet MJ (2011) Mac- macrophage inflammasome activation by inhibiting NF-kappaB rophage activation and differentiation signals regulate schlafen-4 activation and enhancing autophagy. PLoS ONE 9:e97957. https gene expression: evidence for Schlafen-4 as a modulator of mye-://doi.org/10.1371/journ al.pone.00979 57 lopoiesis. PLoS ONE 6:e15723. https ://d oi.org/10.1371 /j our n 92. Xia J, Sinelnikov IV, Han B, Wishart DS (2015) MetaboAnalyst al.pone.00157 23 3.0–making metabolomics more meaningful. Nucleic Acids Res 83. Vasilopoulou E, Kolatsi-Joannou M, Lindenmeyer MT, White 43:W251–W257. https ://doi.org/10.1093/nar/gkv38 0 KE, Robson MG, Cohen CD, Sebire NJ, Riley PR, Winyard PJ, 93. Xia J, Wishart DS (2016) Using MetaboAnalyst 3.0 for compre- Long DA (2016) Loss of endogenous thymosin beta4 acceler- hensive metabolomics data analysis. Curr Protoc Bioinform 55:14 ates glomerular disease. Kidney Int 90:1056–1070. https ://doi. 10 11–14 10 91. https ://doi.org/10.1002/cpbi.11 org/10.1016/j.kint.2016.06.032 94. Zamilpa R, Zhang J, Chiao YA, de Castro Bras LE, Halade GV, 84. Virag JA, Lust RM (2014) Circadian influences on myocardial Ma Y, Hacker SO, Lindsey ML (2013) Cardiac wound healing infarction. Front Physiol 5:422. https ://doi.or g/10.3389/fphy s post-myocardial infarction: a novel method to target extracellu- .2014.00422 lar matrix remodeling in the left ventricle. Methods Mol Biol 85. Wang N, Liang H, Zen K (2014) Molecular mechanisms that influ- 1037:313–324. https ://doi.org/10.1007/978-1-62703 -505-7_18 ence the macrophage m1–m2 polarization balance. Front Immunol 95. Zhou J, Wu R, High AA, Slaughter CA, Finkelstein D, Rehg JE, 5:614. https ://doi.org/10.3389/fimmu .2014.00614 Redecke V, Hacker H (2011) A20-binding inhibitor of NF-kappaB 86. Weinberg SE, Sena LA, Chandel NS (2015) Mitochondria in the (ABIN1) controls Toll-like receptor-mediated CCAAT/enhancer- regulation of innate and adaptive immunity. Immunity 42:406– binding protein beta activation and protects from inflammatory 417. https ://doi.org/10.1016/j.immun i.2015.02.002 disease. Proc Natl Acad Sci USA 108:E998–E1006. https ://doi. 87. Weirather J, Hofmann UD, Beyersdorf N, Ramos GC, Vogel B, org/10.1073/pnas.11062 32108 Frey A, Ertl G, Kerkau T, Frantz S (2014) Foxp3+ CD4+ T cells 96. Zhu M, Goetsch SC, Wang Z, Luo R, Hill JA, Schneider J, improve healing after myocardial infarction by modulating mono- Morris SM Jr, Liu ZP (2015) FoxO4 promotes early inflamma- cyte/macrophage differentiation. Circ Res 115:55–67. https://doi. tory response upon myocardial infarction via endothelial Arg1. org/10.1161/CIRCR ESAHA .115.30389 5 Circ Res 117:967–977. https ://doi.org/10.1161/CIRCR ESAHA 88. Weisheit C, Zhang Y, Faron A, Kopke O, Weisheit G, Steinstrasser .115.30691 9 A, Frede S, Meyer R, Boehm O, Hoeft A, Kurts C, Baumgarten G 1 3

Journal

Basic Research in CardiologySpringer Journals

Published: Jun 4, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off