Cardiac macrophages are heterogenous in phenotype and functions, which has been associated with differences in their ontogeny. Despite extensive research, our understanding of the precise role of different subsets of macrophages in ischemia/reperfusion injury remains incomplete. We here investigated macrophage lineages and ablated tissue macrophages in homeostasis and after I/R injury in a CSF1R-dependent manner. Genomic deletion of a fms-intronic regulatory element (FIRE) in the Csf1r locus resulted in specific absence of resident homeostatic and antigen-presenting macrophages, without affecting the recruitment of monocyte-derived macrophages to the infarcted heart. Specific absence of homeostatic, monocyte-independent macrophages altered the immune cell crosstalk in response to injury and induced proinflammatory neutrophil polarization, resulting in impaired cardiac remodelling without influencing infarct size. In contrast, continuous CSF1R inhibition led to depletion of both resident and recruited macrophage populations. This augmented adverse remodelling after I/R and led to an increased infarct size and deterioration of cardiac function. In summary, resident macrophages orchestrate inflammatory responses improving cardiac remodelling, while recruited macrophages determine infarct size after I/R injury. These findings attribute distinct beneficial effects to different macrophage populations in the context of myocardial infarction.
This study provides important findings on the distinct functions of resident and recruited macrophages during cardiac healing after myocardial ischemia. Using state-of-the-art fate-mapping models and genetic and pharmacological targeting approaches, the authors provide solid evidence that the absence of resident macrophages do not influence infarct size but instead alter the immune cell crosstalk in response to injury. However, the functional evaluation of resident macrophages is limited by potential off target effects in ∆FIRE mice. This study should be of interest to the fields of Development, Immunology and Cardiology.
Macrophages are important effectors of innate immunity. They are essential for host defense against infections but are also involved in different cardiovascular diseases. They represent the most abundant immune cell population in healthy cardiovascular tissues (Heidt et al, 2014; Weinberger et al, 2020), where they contribute to organ functions (Hulsmans et al, 2017) and maintenance of tissue homeostasis (Nicolas-Avila et al, 2020). In cardiovascular diseases such as atherosclerosis and its main sequelae, ischemic stroke and acute myocardial infarction (AMI), macrophage functions are central to both disease development and healing. AMI has remained a leading cause of mortality and morbidity worldwide (Ahmad & Anderson, 2021; Lozano et al, 2012). Although acute survival in this condition has improved through the broad availability of percutaneous coronary intervention, adverse myocardial remodeling and fibrosis frequently result in heart failure (Gerber et al, 2016). Pathophysiologically, the diminished blood supply to myocardial tissue during AMI leads to acute tissue necrosis, which induces a profound sterile inflammation and triggers complex cascade of immune processes and tissue remodeling (Hilgendorf et al, 2014; Honold & Nahrendorf, 2018; Nahrendorf et al, 2007). Consequently, uncontrolled immune reactions in the course of AMI are associated with impaired wound healing and adverse remodeling and can result in worsened cardiac outcome (Panizzi et al, 2010).
Macrophages play an essential role in cardiac injury, and thus represent a potential therapeutic target (Hilgendorf et al., 2014; Nahrendorf et al., 2007). However, they are an heterogeneous population (Kubota et al, 2019; Weinberger & Schulz, 2015; Zaman & Epelman, 2022), and a large body of work has shown that they can have both pro- and anti-inflammatory functions. The differential roles of macrophage populations in AMI have remained incompletely understood. Cardiac macrophages can derive from embryonic and adult hematopoietic progenitors (Epelman et al, 2014). Fate mapping analyses have identified yolk sac (YS) erythro-myeloid progenitors (EMPs) as a principal source of cardiac macrophages in adult life (Ginhoux et al, 2010). However, limited labeling in inducible cre reporter systems have not allowed for precisely differentiating and quantifying developmental origins of cardiac macrophages. Further, targeting of these macrophages has been challenging (Frieler et al, 2015; Ruedl & Jung, 2018).
In this study, we investigated the cellular identity of cardiac macrophages in association with their developmental paths and their immune responses to ischemia/reperfusion (I/R) injury. By combining lineage tracing with single cell RNA sequencing, we provide an in-depth analysis of the differential functions of resident and recruited cardiac macrophages. We then harnessed mice with genomic deletion of the fms-intronic regulatory element (FIRE) (Rojo et al, 2019), that allowed us to specifically address populations of resident macrophages in the infarcted heart and compared them to mice, in which both recruited as well as resident macrophages are depleted by pharmacological inhibtion of the CSF1R-signalling pathway. Using these approaches of selective macrophage depletion, we could attribute different beneficial functions to resident and also to recruited macrophages which impact differently on cardiac remodeling, infarct size and cardiac outcome.
Absence of resident cardiac macrophages in Csf1rΔFIRE mice
To quantify the contribution of YS EMPs to cardiac resident macrophages, we harnessed constitutive labeling in RankCre;Rosa26eYFP mice (Jacome-Galarza et al, 2019; Mass et al, 2016; Percin et al, 2018). ∼80% of cardiac macrophages expressed YFP in hearts of 12 week-old animals, whereas blood monocytes were not labeled (Fig 1A; Fig S1). EMP-derived microglia served as control confirming high efficiency of Cre-mediated recombination. In a comparative analysis, we traced fetal and adult definitive hematopoiesis in Flt3CreRosa26eYFP mice (Gomez Perdiguero et al, 2015; Schulz et al, 2012), indicating that ∼20% of cardiac macrophages derived from definitive HSC (Fig 1B). Thus, the majority of resident macrophages in the healthy heart is of early embryonic origin.
To determine the role of resident macrophages in the mouse heart, we generated RankCreRosa26DTR mice that express the avian diphtheria toxin (DT) receptor in the EMP lineage, rendering them susceptible to DT-mediated ablation. Administration of a single DT dose, however, resulted in premature death of all mice (n=6) within 24 hours (Fig S2). The injected mice presented with a systemic inflammatory response syndrome (SIRS)-like phenotype, indicating that acute depletion of EMP-derived macrophages is not viable.
In Csf1rΔFIRE mice, genetic deletion of the fms-intronic regulatory element (FIRE) in the first intron of the Csf1r gene leads to a selective absence of tissue macrophages without the developmental defects observed in Csf1r-/- mice (Munro et al, 2020; Rojo et al., 2019). We first confirmed that cardiac macrophages in Csf1rΔFIRE/ΔFIRE mice (further termed ΔFIRE) mice were reduced by ∼90% and the number of monocytes and neutrophils was not altered (Fig 1C-E). We thus crossed the ΔFIRE animals with the YS EMP fate mapping line (RankCreRosa26eYFP mice) and showed that the macrophages absent in ΔFIRE animals were EMP-derived macrophages (Fig 1F), thus offering a new model to investigate the role of resident macrophages in cardiac injury.
Changes in the cardiac immune phenotype in Csf1rΔFIRE mice in baseline conditions
To further characterize the macrophage populations absent in healthy hearts of adult ΔFIRE mice, we carried out single cell RNA-sequencing (scRNA-seq) of CD45+ immune cells in wildtype, ΔFIRE and RankCreRosa26eYFP mice (Fig 2A, Fig S3). We identified the presence of six macrophage clusters in control mice (Fig 2B, Fig S3). Two clusters of macrophages expressing homeostatic (e.g. Lyve-1, Gas6, Stab1) and antigen presentation-related genes (e.g. CD74, H2-Ab1) were specifically ablated in ΔFIRE mice (Fig 2C), which largely represent YFP-expressing macrophages in RankCreRosa26eYFPmice (Fig 2D). Specifically, we mapped ∼90% of yfp transcripts to the two clusters of homeostatic and antigen presenting macrophages that were ablated in ΔFIRE. Other clusters were not reduced in ΔFIRE, and also were not quantitatively represented in the YS macrophage lineage tracing model, i.e. Cx3cr1hi(2% yfp), Ccr2lowLy6clo (5% yfp) and Ccr2hiLy6chi (1% yfp) macrophages within respective clusters (Fig 2D). The specific impact of ΔFIRE on resident macrophages is further supported by the observation, that no changes in gene expression nor cell numbers were observed in Ly6chi monocytes and inflammatory Ccr2hiLy6chi macrophages (Fig 2C Fig S4). Together, ΔFIRE allows for investigating the role of resident cardiac macrophages that are largely of YS origin.
Absence of resident macrophages in ΔFIRE was associated with changes in gene expression in non-macrophage clusters such as T- and B-cells and natural killer (NK) cells (Fig S5). Further, overall phagocytic capacity, as inferred by expression of phagocytosis related genes (Amorim et al, 2022), was reduced (Fig 2E). This suggests that absence of resident macrophages is accompanied by distinct changes in immune functions in the homeostatic heart.
Adverse cardiac remodeling in Csf1rΔFIRE mice after I/R injury
To assess the impact of the absence of resident macrophages in cardiac injury, we subjected ΔFIRE mice to I/R injury and investigated remodeling and functional outcome by sequential positron emission tomography (PET) imaging after 6 and 30 days (Fig 3A). Function of the cardiac left ventricle (LV), as determined by ejection fraction (LVEF) and stroke volume (SV), improved in controls in the course of post-I/R remodeling (Fig 3B-C). In contrast, LVEF and SV remained unchanged or worsened in ΔFIRE mice, and longitudinal observations of mice indicated a negative net effect on ejection fraction from day 6 to day 30 post I/R. Thus, absence of resident macrophages had a negative effect on cardiac remodeling (Fig 3D-G). Nonetheless, infarct size as determined by viability defect (PET) and fibrotic area (histology) was not different after 30 days (Fig 3E-H).
Recruitment of BM-derived macrophages into infarct zone of Csf1rΔFIRE mice
To gain a deeper understanding of the inflammatory processes taking place in the infarcted heart, we quantified macrophage distribution by immunofluorescence and flow cytometry analysis of ischemic and remote areas after I/R. In ΔFIRE mice, macrophages were largely absent in the remote zone of infarcted hearts (Fig 4A), indicating sustained depletion of resident macrophages. However, macrophages strongly increased in the infarct area and their numbers were not different in both infarct and border zones between ΔFIRE and control mice (Fig 4A). This indicated recruitment and differentiation of ΔFIRE-independent macrophages from the circulation into these regions. Indeed, complementing lineage tracing of BM HSC in Flt3Cre mice (Fig 4B) and YS EMP in RankCre mice (Fig 4C) proved the recruitment of macrophages from BM HSC (Flt3 GFP+ and Rank RFP-, respectively) in the early phase of I/R injury. 30 days after I/R, BM-derived macrophages remained overrepresented in the infarct zone (∼75% HSC contribution), and differential contribution of BM HSC declined from the border zone (∼50%) to the remote zone (∼35%) (Fig 4D). Taken together, recruited BM-derived macrophages represent the main population in the infarct area and their recruitment was unaltered in ΔFIRE mice. This supports the notion that resident macrophages influence cardiac remodeling but recruited macrophages drive infarct size after I/R.
Transcriptional landscape of resident versus recruited macrophages in I/R injury
To address the differential responses of resident and recruited macrophages to I/R injury, we generated BM chimeric mice. We applied an irradiation-independent model using conditional deletion of c-myb to deplete BM hematopoietic cells in CD45.2 mice and replace them with CD45.1 donor HSC. 2 days after I/R injury, we FACS-sorted recruited (CD45.1+) and resident (CD45.2+) macrophages and carried out bulk RNA sequencing (Fig 5A). The two macrophage populations exhibited profound transcriptional differences after I/R injury. Expression of homeostasis-related genes like Timd4, Lyve-1, CD163 and Retnla was increased in resident CD45.2+ macrophages (Fig 5B, C). Substantiating the differential regulation of macrophage programs, CD45.1+ macrophages increased inflammatory- and host defense-related gene ontology (GO) terms (e.g. myeloid leukocyte related immunity, killing of cells), whereas CD45.2+ macrophages upregulated development- and homeostasis-related GO terms (e.g. extracellular matrix organization, vasculature development, heart muscle development) (Fig 5D). GO enrichment analysis further predicted that processes involved in innate (e.g. leukocyte migration, regulation of defense response, response to bacterium) as well as the adaptive immunity (e.g. T cell selection, T cell activation, B cell proliferation) were increased in CD45.1+ macrophages after infarction (Fig 5D). Thus, resident and recruited macrophages provide distinct transcriptional changes and inferred functions in response to I/R.
Altered inflammatory patterns and immune cell communication in Csf1rΔFIRE mice
To evaluate the immune response of resident and recruited macrophages to I/R injury in ΔFIRE mice, we interrogated the transcription profile of CD45+ cells from the infarct area (Fig 6A). In contrast to the absence of homeostatic and antigen-presenting macrophage clusters in healthy hearts of ΔFIRE mice (Fig 2B-C), there were less differences in immune cell clusters between ΔFIRE and control mice 2 days after I/R (Fig 6B, C). Abundance of homeostatic macrophages was also reduced at this time point, however, other clusters including Ccr2hiLy6chi inflammatory macrophages were not altered, which is in line with our histological findings.
ΔFIRE was associated with some changes in gene expression in cardiac non-macrophage immune cells. Across different clusters, including lymphocyte and neutrophil clusters, expression of anti-inflammatory genes like Chil3 (Ym1) and Lcn2 was reduced. Vice versa, expression of Bcl-family genes like Blc2a1a and Bcl2a1d, which are associated with apoptosis and inflammatory pathways, was higher in ΔFIRE mice (Fig S6-8). Other upregulated genes were related to antigen presentation (e.g.CD74, H2-Ab1), as identified in the Ccr2loLy6clo, homeostatic and Ccr2hiLy6chi macrophage clusters (Fig S6 and 7). Further, we interrogated inflammatory gene expression in neutrophils, which are abundant first responders to myocardial infarction. We found that a computed score summing up inflammasome activation (Amorim et al., 2022) was increased in all neutrophil clusters in ΔFIRE mice (Fig 6D). Thus, the absence of cardiac macrophages was associated with altered inflammatory properties of non-macrophage immune cells in the infarcted heart.
Altered intercellular crosstalk of macrophages is a hallmark of cardiac inflammation. We therefore assessed ligand-receptor (LR) interactions between immune cell populations after I/R injury. Indeed, the number of LR interactions with neutrophils and lymphocytes, as well as the strength of the macrophage-emitted communication signals was markedly reduced in homeostatic, antigen-presenting and Ccr2loLy6clo Mø macrophage clusters (Fig 6E, F). In contrast, immune cell communication in BM-derived macrophage clusters (e.g. Ccr2hiLy6chi Mø cluster, Fig S9) was not different to control mice. Taken together, deficiency in resident macrophages in I/R injury altered the intercellular immune crosstalk and induced a pro-inflammatory signature in e.g. cardiac neutrophils. However, transcriptional profiles of BM-derived inflammatory macrophages were largely unaltered. Together with our histological analysis showing the dominance of recruited BM-derived macrophages in the early phase of I/R, this potentially explained the limited impact of ΔFIRE on functional outcome after I/R injury.
Ablation of resident and recruited macrophages severely impacts on cardiac healing after I/R injury
To test this hypothesis, we determined the effect of combined ablation of resident and recruited macrophages. We therefore exposed mice to continuous treatment with the CSF1R-inhibitor PLX5622 (Fig 7A). In healthy hearts, inhibitor treatment resulted in the absence of cardiac macrophages within 3 days, which had no impact on cardiac function (Fig S10). We then subjected mice to I/R injury and investigated outcome by sequential PET imaging and histology (Fig 7A). Treatment with PLX5622 diminished macrophage numbers in both remote and infarct areas in the early phase (day 2) after injury. Recruitment of other myeloid cells e.g. neutrophils was not altered in this context (Fig 7B). This effect was pronounced in the chronic phase (day 30) after I/R injury, in which macrophages were largely absent in remote, border and infarct zones (Fig 7C,D). Absence of resident and recruited macrophages was associated increased infarct size, as determined by fibrosis area (WGA histology) as well as viability defect (PET), and resulted in deterioration of cardiac function (Fig 7E-I). Specifically, LVEF was reduced 6 days after I/R, and remained strongly impaired at 30 days (Fig 7E).
Taken together, sole absence of resident macrophages had limited negative impact on cardiac remodeling. Absence of both resident and recruited macrophages resulted in a significant increase in infarct size and deterioration of left ventricular function after I/R injury, highlighting a beneficial effect of recruited macrophages in cardiac healing.
Macrophages are key players in cardiac homeostasis and disease (Bajpai et al, 2019; Bajpai et al, 2018; Dick et al, 2019; Epelman et al., 2014; Hulsmans et al., 2017; Nahrendorf et al., 2007; Nicolas-Avila et al., 2020; Panizzi et al., 2010; Sager et al, 2016). The precise understanding of their developmental origin, their functions and their regulation could enable the identification of macrophage-targeted strategies to modify inflammation in the heart. This study sheds light on the differential role of resident and recruited macrophages in cardiac remodeling and outcome after AMI.
BM-derived recruited macrophages are a scarce population in the healthy heart but are recruited in vast numbers to the injured myocardium after I/R injury. These recruited cells exhibit substantially different transcriptional profiles in comparison to their resident counterpart, and show pro-inflammatory properties. Resident macrophages remain present in the remote and border zone and display a reparative gene expression profile after I/R injury. In comparison to recruited macrophages, resident macrophages expressed higher levels of genes related to homeostatic functions (e.g. Lyve-1, Timd4, CD163, Stab1). These markers have recently been associated with self-renewing tissue macrophages that are maintained independently of BM contribution (Dick et al, 2022). Biological processes associated with cardiac healing are upregulated in resident macrophages (e.g. regulation of vascular development, regulation of cardiomyocyte development, extracellular matrix organization). Transcriptional profiles of recruited macrophages show enhanced inflammatory biological processes also in macrophages harvested from the remote area, underlining their potentially detrimental influence on remote cardiac injury.
Two recent studies addressed the role of resident macrophages using DT-mediated macrophage depletion and reported impaired cardiac remodeling in chronic myocardial infarction (Dick et al., 2019) and after I/R injury (Bajpai et al., 2019). However, DT-mediated cell ablation is known to induce neutrophil recruitment and tissue inflammation (Frieler et al., 2015; Oh et al, 2017; Ruedl & Jung, 2018; Sivakumaran et al, 2016). This inflammatory preconditioning of cardiac tissue after DT-depletion is likely to impact on cardiac remodeling and influence assessment of tissue macrophage functions. Genomic deletion of FIRE in the Csf1r gene results in the near-absence of resident cardiac macrophages but circumvents the inflammatory stimulus of DT-induced depletion. ΔFIRE specifically diminished homeostatic and antigen-presenting macrophages. Substantiating the specificity to resident cardiac macrophages in this genetic model, the amount of macrophages recruited to the infarct area after I/R injury was not affected in ΔFIRE mice. Further, scRNAseq analysis revealed that gene expression of monocytes as well as Ccr2hiLy6chi and Cx3cr1hi Mф, representing mainly recruited immune cells, was mostly unchanged in ΔFIRE compared to control mice. The specific targeting of the resident macrophage population introduces an interesting in vivo model for studying their role in MI without the need for inflammation-prone conditional deletion of macrophages (Frieler et al., 2015).
Ablation of resident macrophages altered macrophage crosstalk to non-macrophage immune cells, especially lymphocytes and neutrophils. This was paralleled by a proinflammatory gene signature in neutrophils, with increased expression of inflammasome-related genes and a reduction in anti-inflammatory genes like Chil3 and Lcn2 (Gordon & Martinez, 2010; Guo et al, 2014; Parmar et al, 2018). Interestingly, inflammatory polarization of neutrophils is associated with poor outcome also after ischemic brain injury (Cuartero et al, 2013). Consequently, absence of resident macrophages worsened cardiac remodeling in the late postinfarction phase in ΔFIRE mice indicating their biological role in myocardial healing. However, absence of resident macrophages did not change infarct size or impact on LV function.
Depletion of macrophages by pharmacological inhibition of CSF1R induced the absence of both resident and recruited macrophages, allowing us to determine cardiac outcome in juxtaposition to the ΔFIRE mice. Continuous CSF1R inhibition induced the absence of macrophages also in the infarcted area and had deleterious effect on infarct size and LV function. In line with our findings, depletion of macrophages by anti-CSF1R treatment was associated with worsened cardiac function in a model of pressure overload induced heart failure (Revelo et al, 2021). Controversially, a recent study in which monocytes were depleted using DT-injections in Ccr2DTR mice reported beneficial effects on cardiac outcome after MI in mice. However, other immune cells including neutrophils express CCR2 and may therefore be affected directly (Talbot et al, 2015; Xu et al, 2017). In line with this, Ccr2-deficient mice exhibit reduced acute recruitment of neutrophils to the brain after I/R injury, which was associated with reduced infarct size and brain edema (Dimitrijevic et al, 2007).
Taken together our study underlines the heterogeneity of cardiac macrophages and the importance of ontogeny therein. Resident macrophages, which mainly derive from YS EMPs, govern cardiac homeostasis in the healthy heart, and contribute positively to cardiac healing after I/R injury by orchestrating anti-inflammatory programming of other cardiac immune cells. However, recruited macrophages contribute to the healing phase after AMI and their absence defines infarct size and cardiac outcome.
Materials and methods
RankCre (Jacome-Galarza et al., 2019; Percin et al., 2018), Flt3Cre (Benz et al, 2008), Mybfl/fl (Emambokus et al, 2003), Mx1Cre (Kuhn et al, 1995) (from The Jackson Laboratory (JAX), Stock No: 003556), Csf1rΔFIRE (Rojo et al., 2019), Rosa26DTR(Buch et al, 2005) (JAX Stock No: 007900), Rosa26mT/mG (Muzumdar et al, 2007) (JAX Stock No: 007676), Rosa26eYFP (Srinivas et al, 2001) (JAX Stock No: 006148), and Rosa26RFP (JAX Stock No: 034720) mice have been previously described. PCR genotyping was performed according to protocols described previously. Animals were aged between 10-16 weeks. We have complied with all relevant ethical regulations. Animal studies were approved by the local regulatory agency (Regierung von Oberbayern, Munich, Germany, record numbers ROB-55.2-2532.Vet_02-19-17 and ROB-55.2-2532.Vet_02-19-1; Care and Use Committee of the Institut Pasteur (CETEA), dap190119).
For fate-mapping analysis of Flt3+ precursors, Flt3Cremales (the transgene is located on the Y chromosome) were crossed to Rosa26eYFP or Rosa26mT/mG female reporter mice and only the male progeny was used. For all other experiments, the offspring of both sexes was used for experiments.
BM transplantation was enabled by conditional deletion of the transcription factor myb as previously described (Stremmel et al, 2018). In brief, we induced BM ablation by 4 injections of polyI:C into CD45.2; Mx1CreMybflox/flox every other day, and then transplanted BM from congenic C57BL/6 CD45.1 (Ly5.1; JAX Stock No: 006584). We confirmed the success (chimerism of above 90%) of the BM-transplantation after 4 weeks.
To deplete macrophages, we used the selective CSF1R-inhibitor PLX5622. Control and PLX5622 (300 ppm formulated in AIN-76A standard chow, Research Diets, Inc.) chows were kindly provided by Plexxikon Inc (Berkeley, CA). For depletion of RANK-lineage macrophages, RankCreRosa26DTR mice were injected intraperitoneally with 0.02 mg/kg body weight diphtheria toxin (DT; Sigma-Aldrich) at the timepoints mentioned in Fig ure and corresponding legend. Mice were closely monitored by veterinarians and according to score sheets that were approved by the regulatory agency. Single DT application in RankCreRosa26DTRmice let to rapid deterioration of their health (without infarction or other interventions), and consequently experiments needed to be aborted according to regulations; experimental outcomes were then documented as premature death as indicated.
Ischemia-reperfusion (I/R) injury
I/R injury was carried out as previously described (Nicolas-Avila et al., 2020). In brief, mice were anesthetized using 2% isoflurane and intraperitoneal injection of fentanyl (0.05 mg/kg), midazolam (5.0 mg/kg) and medetomidine (0.5 mg/kg), and then intubated orally (MiniVent Ventilator model nr. 845, Harvard Apparatus®) and ventilated (volume of 150 µL at 200/min). After lateral thoracotomy, the left anterior descending artery (LAD) was ligated with an 8-0 prolene suture producing an ischemic area in the apical left ventricle (LV). To induce the reperfusion injury, the suture was removed after 60 min and reperfusion was confirmed by observing the recoloring of the LV. For experiments with CD169-DTR mice, animals were anesthetized with xylazine and ketamine, and duration of ischemia was 45 min in the I/R model as described earlier (Nicolas-Avila et al., 2020). Postoperative analgesia was performed by injection of Buprenorphin (0.1 mg/kg) twice per day for 3 days. After 2, 6 or 30 days after I/R injury organs were harvested after cervical dislocation.
Mice were anaesthetized using 2% isoflurane and organ harvest was performed after cervical dislocation. Blood was harvested with a heparinized syringe (2mL) by cardiac puncture. After perfusion with 20 mL of ice-cold PBS hearts were excised and kept in PBS on ice until further tissue processing. For flow cytometry, hearts were divided into the ischemia (tissue distal of the LAD ligation) and remote area (tissue proximal of the LAD ligation). For histological examinations of tissues, hearts were incubated in 4% PFA for 30 min followed by an incubation in 30% sucrose-solution (Sigma Aldrich®) for 24 h. Afterwards, hearts were mounted onto a heart slicing device and cut transversally into 3 equal parts (termed level 1, 2 and 3) and stored in Tissue-Tek® (Sakura Finetek Germany GmbH) at -80°C.
Cryosections (10-12 µm) of heart tissue were fixed with 4% paraformaldehyde for 10 min. Blocking and permeabilization were performed with 0.5% Saponin and 10% goat serum for 1h. Primary antibodies were added and incubated for 2-18 h (see Supplemental Table 1). Slides were washed with PBS and secondary antibodies were added and incubated for 1h (see Supplemental Table 1). WGA-staining (Wheat Germ Agglutinin Alexa Fluor™ 647 conjugated antibody, ThermoFisher Scientific®) was used to locate and measure the infarct area and nuclei were stained with DAPI. Finally, slides were washed one more time and Fluorescence Mounting Medium was used to cover the stained sample.
Heart samples were evaluated using an Axio Imager M2 (Carl Zeiss®) and blinded picture analysis was performed using ZEN Imaging and Axiovision SE64 Rel. 4.9.1 (Carl Zeiss®). For the evaluation of cell numbers, 6 individual high-resolution images from each respective anatomical region (infarct area, border zone and remote zone) were analysed for each animal. To measure infarct size the heart was cut into 3 parts and the infarcted area was measured as WGA+ area in sections from each part.
100 µL of heparinized blood was used for FACS analysis. Erythrocytes were lysed with 1% ammonium chloride. After washing with PBS, the cell suspension was resuspended in purified rat anti-mouse CD16/CD32 (BD Pharmingen®) and incubated for 15 min at 4°C. Following this, cells were incubated with FACS-antibodies for 15 min at 4°C (table S1).
Heart tissue was dissected into remote and ischemic tissue as described above and minced into small pieces using forceps and a scalpel. When comparing baseline and I/R injury in FACS analysis, basal heart tissue was used for comparison with remote tissue and apical heart tissue for comparison with ischemia tissue. After enzymatic digestion (Collagenase XI 1200 U/mg, Collagenase I 125 U/mg, Hyaluronidase 500 U/mg, DNase I 1836 U/mg; Sigma Aldrich®) for 30 min at 37°C cells were washed and incubated with purified anti-CD16/32 (FcγRIII/II; dilution 1/50) for 10 min. Thereafter, cells were incubated wit FACS-antibodies (table S1) for 30 min at 4°C. FACS-analysis was performed on a BD Fortessa® or a BeckmanCoulter Cytoflex® flow cytometer and gating strategies are shown in Supplemental Fig ures 1 and 3. Data were analysed using FlowJo® (version 10.0.8r1).
Cell sorting was performed on a MoFlo Astrios (Beckman Coulter) to obtain cardiac macrophages from CD45.2; Mx1CreMybflox/flox after BM-transplantation of CD45.1 BM (n=3 for 2 days after I/R injury) for bulk sequencing, or all cardiac immune cells (CD45+ cells) of Csf1rΔFIRE and RankCreRosa26eYFP mice (n=3 for Csf1rΔFIRE/+ and Csf1rΔFIRE/ΔFIRE in baseline conditions, n=1 for RankCreRosa26eYFP in baseline conditions; n=2 for Csf1rΔFIRE/+ and Csf1rΔFIRE/ΔFIRE 2 days after I/R injury) for single cell analysis. CD45+ cells were enriched by using magnetic beads and MS columns (CD45 MicroBeads; Miltenyi Biotec). Cells were then sorted as single/live/CD45+ cells for single cell sequencing from baseline hearts and from the ischemic myocardium 2 days after I/R injury. Bulk sequencing was performed on single/live/CD45+/lin-/CD11b+/F4/80+/CD64+/CD45.1+ or CD45.2+ cells from the remote and the ischemic myocardium 2 days after I/R injury. Dead cells were identified with SYTOX Orange Dead Cell Stain.
Bulk sequencing and analysis
For each sample, ∼1000 macrophages were sorted into 75µL of RLT buffer (Qiagen, containing 1% beta-mercaptoethanol), vortexed for 1 min and immediately frozen (-80°C). RNA extraction (RNeasy® Plus Micro Kit, Qiagen), cDNA generation (SMART®-Seq v4 Ultra Low Input RNA Kit, Takara Bio) and library preparation (Nextera® XT DNA Library Prep Kit, Illumina) were performed according to the manufacturer’s specifications. Sequencing was performed on a HiSeq4000 system (Ilumina).
The obtained reads were trimmed using bbduk from the BBMap (https://sourceforge.net/projects/bbmap/) v38.87 collection using parameters “ktrim=r k=23 mink=11 hdist=1 tpe tbo”. The trimmed reads were aligned with Hisat2 2.2.1 against the Ensembl release 102 reference mouse genome (Yates et al, 2020).
Gene expression was quantified using the featureCounts (Liao et al, 2014) application from the subread package (v2.0.1) and with parameters “--primary -O -C -B -p -T 8 --minOverlap 5”. Differential expression analysis was performed using DESeq2 (v1.30.0) (Love et al, 2014). A Gene Ontology Set Enrichment analysis was performed using the ClusterProfiler R package (v3.18.1). The visualization of the respective clustermaps was layouted using the ForceAtlas2 implementation of the python fa2 (https://pypi.org/project/fa2/) package for networkx (v2.5) (https://networkx.org/).
On the library-size normalized count data, the pymRMR ((Peng et al, 2005), https://pypi.org/project/pymrmr/) package was used to derive the top 100 discriminatory genes (Mutual Information Quotient method) for subsequently calculating the UMAP 2D-embedding ((McInnes et al, 2020), (https://pypi.org/project/umap-learn/)) for all samples (umap-learn package v0.5.0rc1, 3 neighbors) ((McInnes et al, 2018), https://pypi.org/project/pymrmr/).
Single cell RNA sequencing and analysis
After sorting, cells were proceeded for single cell capture, barcoding and library preparation using Chromium Next GEM single cell 3’ (v3.1, 10x Genomics) according to manufacturer’s specifications. Pooled libraries were sequenced on an Illumina HiSeq1500 sequencer (Illumina, San Diego, USA) in paired-end mode with asymmetric read length of 28+91 bp and a single indexing read of 8bp.
The reads of heart1 sample (RankCreRosa26eYFP) were demultiplexed using Je-demultiplex-illu (Girardot et al, 2016) and mapped against a customized mouse reference genome (GRCm38.p6, Gencode annotation M24) including eYfp sequence using CellRanger (v3.1.0, 10x Genomics).
The six mouse samples 20133-0001 to 20133-0006 (baseline condition of Csf1rΔFIRE/+and Csf1rΔFIRE/ΔFIRE), were processed using Cellranger 4.0.0 using the 2020A mm10 reference.
The 4 mouse samples (MUC13956-13959, infarct condition of Csf1rΔFIRE/+ and Csf1rΔFIRE/ΔFIRE), were sequenced with 4 technical sequencing replicates and pooled using the cellranger (v4.0.0, 10x Genomics) pipeline. Cellranger 4.0.0 was called with default parameters and the 2020A mm10 reference for gene expression.
Finally, all 11 samples were integrated using Seurat 4.0.0 (on R 4.0.1) (Stuart & Satija, 2019). The samples were processed, and cells were filtered to contain between 200 and 6000 features, have at least 1000 molecules detected (nCount_RNA > 1000), have below 15% mtRNA content (^MT) and below 40% ribosomal RNA content (^Rps|^Rpl). After this filtering a total of 35759 cells remained.
After performing SCTransform (Hafemeister & Satija, 2019) on the samples, the SCTransform vignette for integrating the datasets was followed (with 2000 integration features). For dimensionality reduction, PCA was performed using default parameters, and UMAP and Neighbour-Finding was run on 50 PCs. Clustering was performed at a resolution of 0.8. A total of 18 clusters was identified using this approach. Cluster markers were calculated using the t-test in the FindMarkers function. Subsequently cell types were initially predicted using the cPred cell type prediction (https://github.com/mjoppich/scrnaseq_celltype_prediction). Upon manual curation, further fine-grained cell type annotations were made. Differential comparisons were performed against several subgroups of the data set. These comparisons were performed using the t-test in the FindMarkers function of Seurat. Differential results are visualized using the EnhancedVolcano library (Blighe K, 2021).
For the analysis of cell-cell interactions we downloaded the ligand-receptor pairs from Jin et al. (Jin et al, 2021) from the Lewis Lab GitHub repository (https://github.com/LewisLabUCSD/Ligand-Receptor-Pairs). For each interaction (ligand-receptor pair for a cluster-pair) the communication score is calculated as the expression product (Armingol et al, 2021) of the mean normalized expressions exported from the Seurat object. This ensures that little expression of either ligand or receptor in only one cluster results in a relatively low communication score, and only good expression of ligand and receptor will result in a high communication score. The direction of an interaction is fixed from ligand to receptor. The single ligand-receptor-communication-scores were then aggregated (sum) such that only interactions with a score greater six were taken into account.
Gene module scores for inflammasome, ROS and phagocytosis gene sets were calculated using Seurat’s AddModuleScore function. All scripts, including the ones for creating the visualizations of bulk and scRNA-seq data, are available online through https://github.com/mjoppich/myocardial_infarction.
In Vivo PET Imaging
ECG-gated positron emission tomography (PET) images were performed on day 6 and 30 after I/R injury of the LAD using a dedicated small-animal PET scanner (Inveon Dedicated PET, Preclinical Solutions, Siemens Healthcare Molecular Imaging, Knoxville, TN, USA), as previously described (Brunner et al, 2012). Anaesthesia was induced with isoflurane (2.5%), delivered via a face mask in pure oxygen at a rate of 1.2 L/min and maintained with isoflurane (1.5%). Approximately 15 MBq 2-deoxy-2-[18F]fluoro-D-glucose ([18F]-FDG) (∼100 μL) were administered through a teil vein catheter and slowly flushed immediately afterwards with 50 μL saline solution. Body temperature was monitored using a rectal thermometer and maintained within the normal range using a heating pad. After placing animals within the aperture of the PET scanner ECG electrodes (3M, St. Paul, MN, USA) were placed on both forepaws and the left hind paw and ECG was recorded using a dedicated physiological monitoring system (BioVet; Spin Systems Pty Ltd., Brisbane, Australia) (Todica et al, 2018). The PET emission acquisition (list-mode) was initiated 30 min after [18F]-FDG injection and lasted 15 min (Brunner et al., 2012; Gross et al, 2016). For scatter and attenuation correction and additional 7-min long transmission scan was performed using a Co-57 source.
The accuracy of the ECG trigger signal was verified retrospectively using in-house software programmed in MATLAB (The Mathworks, Natick, USA) and in C programming language and erroneous trigger events were removed when needed, as previously described (Boning et al, 2013). Further processing of the data was performed using the Inveon Acquisition Workplace (Siemens Medical Solutions, Knoxville, TN). As previously described, data was reconstructed as a static image or as a cardiac gated image with 16 bins in a 128 × 128 matrix with a zoom of 211% using an OSEM 3D algorithm with 4 and a MAP 3D algorithm with 32 iterations (Brunner et al., 2012). The reconstructed data was normalized, corrected for randoms, dead time and decay as well as attenuation and scatter.
PET images were analysed using the Inveon Research Workplace in a blinded manner (Siemens Medical Solutions, Knoxville, TN). Infarct sizes were determined from static reconstructed images using QPS® (Cedars-Sinai, Los Angeles, CA, USA). Hereby, datasets were compared to a normative database and the viability defect was calculated as percentage of the left ventricular volume, as described previously (Lehner et al, 2014; Todica et al, 2014). Left ventricular function volumes (EDV, ESV, SV), as well as the LVEF, were determined from ECG-gated images using QGS® (Cedars-Sinai, Los Angeles, CA, USA), as described previously (Brunner et al., 2012; Croteau et al, 2003).
Student’s t-test or Fisher’s LSD test was used (Prism GraphPad®). Welch’s correction for unequal variances was used when applicable. A p-value of p < 0.05 was considered significant. Data are presented as mean +/- standard deviation (SD).
We thank Nicole Blount, Beate Jantz, Michael Lorenz, Sebastian Helmer for excellent technical assistance.
DZHK (German Centre for Cardiovascular Research) and the BMBF (German Ministry of Education and Research) (grants 81Z0600204 to CS, 81X2600252 to TW and 81X2600256 to MF). The Deutsche Forschungsgemeinschaft (SCHU 2297/1-1 and collaborative research centers 1123 project Z02 (MJ, RZ) and A07 (CS), and CRC TRR332 project A6 (CS). LMUexcellent (TW). ESC Research Grant 2021 (TW). Friedrich-Baur Stiftung (TW). Deutsche Stiftung für Herzforschung (M.F.). Chinese Scholarship Council (CSC) to JF and LL.
Conceptualization: TW, CSchu
Methodology: TW, MJ, MF, JL, GP, AT, DE, CSchu
Investigation: TW, DM, MJ, MF, CG, KK, VW, SR, SA, JF, LL, WHL, JW, LT, DE, GP, SMS
Funding acquisition: TW, MF, RZ, EGP, CSchu
Supervision: AH, CW, SE, AT, RZ, CP, EGP, CSchu
Writing – original draft: TW, MJ
Writing – review & editing: all authors
Conflict of interest
The authors declare no conflict of interest.
- The Leading Causes of Death in the US for 2020JAMA
- IFNgamma and GM-CSF control complementary differentiation programs in the monocyte-to-phagocyte transition during neuroinflammationNat Immunol 23:217–228
- Deciphering cell-cell interactions and communication from gene expressionNat Rev Genet 22:71–88
- Tissue Resident CCR2- and CCR2+ Cardiac Macrophages Differentially Orchestrate Monocyte Recruitment and Fate Specification Following Myocardial InjuryCirc Res 124:263–278
- The human heart contains distinct macrophage subsets with divergent origins and functionsNat Med 24:1234–1245
- The stream of precursors that colonizes the thymus proceeds selectively through the early T lineage precursor stage of T cell developmentJ Exp Med 205:1187–1199
- Blighe K RS, Lewis M, 2021. EnhancedVolcano: Publication-ready volcano plots with enhanced colouring and labeling. R package version 1.10.0, https://githubcom/kevinblighe/EnhancedVolcano
- Erroneous cardiac ECG-gated PET list-mode trigger events can be retrospectively identified and replaced by an offline reprocessing approach: first results in rodentsPhys Med Biol 58
- Left ventricular functional assessment in murine models of ischemic and dilated cardiomyopathy using [18 F]FDG-PET: comparison with cardiac MRI and monitoring erythropoietin therapyEJNMMI Res 2
- A Cre-inducible diphtheria toxin receptor mediates cell lineage ablation after toxin administrationNat Methods 2:419–426
- Quantitative gated PET for the assessment of left ventricular function in small animalsJ Nucl Med 44:1655–1661
- N2 neutrophils, novel players in brain inflammation after stroke: modulation by the PPARgamma agonist rosiglitazoneStroke 44:3498–3508
- Self-renewing resident cardiac macrophages limit adverse remodeling following myocardial infarctionNat Immunol 20:29–39
- Three tissue resident macrophage subsets coexist across organs with conserved origins and life cyclesSci Immunol 7
- Absence of the chemokine receptor CCR2 protects against cerebral ischemia/reperfusion injury in miceStroke 38:1345–1353
- Progression through key stages of haemopoiesis is dependent on distinct threshold levels of c-MybEMBO J 22:4478–4488
- Embryonic and adult-derived resident cardiac macrophages are maintained through distinct mechanisms at steady state and during inflammationImmunity 40:91–104
- Depletion of macrophages in CD11b diphtheria toxin receptor mice induces brain inflammation and enhances inflammatory signaling during traumatic brain injuryBrain Res 1624:103–112
- Mortality Associated With Heart Failure After Myocardial Infarction: A Contemporary Community PerspectiveCirc Heart Fail 9
- Fate mapping analysis reveals that adult microglia derive from primitive macrophagesScience 330:841–845
- Je, a versatile suite to handle multiplexed NGS libraries with unique molecular identifiersBMC Bioinformatics 17
- Tissue-resident macrophages originate from yolk-sac-derived erythro-myeloid progenitorsNature 518:547–551
- Alternative activation of macrophages: mechanism and functionsImmunity 32:593–604
- FDG-PET reveals improved cardiac regeneration and attenuated adverse remodelling following Sitagliptin + G-CSF therapy after acute myocardial infarctionEur Heart J Cardiovasc Imaging 17:136–145
- Lipocalin 2 is a regulator of macrophage polarization and NF-kappaB/STAT3 pathway activationMol Endocrinol 28:1616–1628
- Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regressionGenome Biol 20
- Differential contribution of monocytes to heart macrophages in steady-state and after myocardial infarctionCirc Res 115:284–295
- Ly-6Chigh monocytes depend on Nr4a1 to balance both inflammatory and reparative phases in the infarcted myocardiumCirc Res 114:1611–1622
- Resident and Monocyte-Derived Macrophages in Cardiovascular DiseaseCirc Res 122:113–127
- Macrophages Facilitate Electrical Conduction in the HeartCell 169:510–522
- Developmental origin, functional maintenance and genetic rescue of osteoclastsNature 568:541–545
- Inference and analysis of cell-cell communication using CellChatNat Commun 12
- Matrix metalloproteinase-12 produced by Ly6C(low) macrophages prolongs the survival after myocardial infarction by preventing neutrophil influxJ Mol Cell Cardiol 131:41–52
- Inducible gene targeting in miceScience 269:1427–1429
- Lehner S, Todica A, Vanchev Y, Uebleis C, Wang H, Herrler T, Wangler C, Cumming P, Boning G, Franz WM et al (2014) In vivo monitoring of parathyroid hormone treatment after myocardial infarction in mice with [68Ga]annexin A5 and [18F]fluorodeoxyglucose positron emission tomography. Mol Imaging 13: 1-8In vivo monitoring of parathyroid hormone treatment after myocardial infarction in mice with [68Ga]annexin A5 and [18F]fluorodeoxyglucose positron emission tomography. Mol Imaging 13
- featureCounts: an efficient general purpose program for assigning sequence reads to genomic featuresBioinformatics 30:923–930
- Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2Genome Biol 15
- Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010Lancet 380:2095–2128
- Specification of tissue-resident macrophages during organogenesisScience
- UMAP: Uniform Manifold Approximation and Projection for Dimension ReductionarXiv
- UMAP: Uniform Manifold Approximation and ProjectionJ Open Source Softw 3
- CNS macrophages differentially rely on an intronic Csf1r enhancer for their developmentDevelopment
- A global double-fluorescent Cre reporter mouseGenesis 45:593–605
- The healing myocardium sequentially mobilizes two monocyte subsets with divergent and complementary functionsJ Exp Med 204:3037–3047
- A Network of Macrophages Supports Mitochondrial Homeostasis in the HeartCell 183:94–109
- Transient Depletion of CD169(+) Cells Contributes to Impaired Early Protection and Effector CD8(+) T Cell Recruitment against Mucosal Respiratory Syncytial Virus InfectionFront Immunol 8
- Impaired infarct healing in atherosclerotic mice with Ly-6C(hi) monocytosisJ Am Coll Cardiol 55:1629–1638
- Lipocalin 2 Plays an Important Role in Regulating Inflammation in Retinal DegenerationJ Immunol 200:3128–3141
- Feature selection based on mutual information: criteria of max-dependency, max-relevance, and min-redundancyIEEE Trans Pattern Anal Mach Intell 27:1226–1238
- CSF1R regulates the dendritic cell pool size in adult mice via embryo-derived tissue-resident macrophagesNat Commun 9
- Cardiac Resident Macrophages Prevent Fibrosis and Stimulate AngiogenesisCirc Res
- Deletion of a Csf1r enhancer selectively impacts CSF1R expression and development of tissue macrophage populationsNat Commun 10
- DTR-mediated conditional cell ablation-Progress and challengesEur J Immunol 48:1114–1119
- Proliferation and Recruitment Contribute to Myocardial Macrophage Expansion in Chronic Heart FailureCirc Res 119:853–864
- A lineage of myeloid cells independent of Myb and hematopoietic stem cellsScience 336:86–90
- Depletion of CD11c(+) cells in the CD11c.DTR model drives expansion of unique CD64(+) Ly6C(+) monocytes that are poised to release TNF-alphaEur J Immunol 46
- Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locusBMC Dev Biol 1
- Inducible disruption of the c-myb gene allows allogeneic bone marrow transplantation without irradiationJ Immunol Methods 457:66–72
- Integrative single-cell analysisNat Rev Genet 20:257–272
- CCR2 Expression in Neutrophils Plays a Critical Role in Their Migration Into the Joints in Rheumatoid ArthritisArthritis Rheumatol 67:1751–1759
- Monitoring of Cardiac Remodeling in a Mouse Model of Pressure-Overload Left Ventricular Hypertrophy with [(18)F]FDG MicroPETMol Imaging Biol 20:268–274
- In-vivo monitoring of erythropoietin treatment after myocardial infarction in mice with [(6)(8)Ga]Annexin A5 and [(1)(8)F]FDG PETJ Nucl Cardiol 21:1191–1199
- Ontogeny of arterial macrophages defines their functions in homeostasis and inflammationNat Commun 11
- Myocardial infarction: a critical role of macrophages in cardiac remodelingFront Physiol 6
- CCR2 dependent neutrophil activation and mobilization rely on TLR4-p38 axis during liver ischemia-reperfusion injuryAm J Transl Res 9:2878–2890
- Ensembl 2020Nucleic Acids Res 48:D682–D688
- Resident cardiac macrophages: Heterogeneity and function in health and diseaseImmunity 55:1549–1563