Mechanical regulation of macrophage metabolism by allograft inflammatory factor 1 leads to adverse remodeling after cardiac injury

Abstract

Myocardial infarction (MI) mobilizes macrophages, the central protagonists of tissue repair in the infarcted heart. Although necessary for repair, macrophages also contribute to adverse remodeling and progression to heart failure. In this context, specific targeting of inflammatory macrophage activation may attenuate maladaptive responses and enhance cardiac repair. Allograft inflammatory factor 1 (AIF1) is a macrophage-specific protein expressed in a variety of inflammatory settings, but its function after MI is unknown. Here we identify a maladaptive role for macrophage AIF1 after MI in mice. Mechanistic studies show that AIF1 increases actin remodeling in macrophages to promote reactive oxygen species-dependent activation of hypoxia-inducible factor (HIF)-1α. This directs a switch to glycolytic metabolism to fuel macrophage-mediated inflammation, adverse ventricular remodeling and progression to heart failure. Targeted knockdown of Aif1 using antisense oligonucleotides improved cardiac repair, supporting further exploration of macrophage AIF1 as a therapeutic target after MI.

Extended Data Fig. 1. Allograft inflammatory factor 1 (Aif1) is enriched within recruited inflammatory macrophage clusters, but not lymphocytes, after myocardial infarction (MI) in mice.

A Using the Infection Atlas (Rizzo 2022) to visualize publicly available single cell RNA sequencing data (GSE135310, GSE197441, GSE197853), Aif1 expression was analyzed among monocytes and macrophages isolated from the hearts of mice at various time points after MI. Inflammatory myeloid subsets include Ly6C^hi monocytes (Ly6Chi mo) and recruited macrophage clusters (Res MHCII, MHCII+IL1B+, and Isg15). B Aif1 gene expression in cardiac T and B cells 5 days after MI as visualized using the Infection Atlas. C AIF1 expression in peripheral blood lymphocytes 5 days after MI. Dashed lines represent fluorescence minus one (FMO) staining controls (n = 4 naïve and n = 4 day 5 MI, two-tailed unpaired t test, pooled from two independent experiments). Data are presented as mean ± SEM.

Extended Data Fig. 2. Allograft inflammatory factor 1 (Aif1) worsens cardiac repair after myocardial infarction (MI).

A Mice with whole body deletion of Aif1 (Aif1−/−) or controls (Aif1+/+) were subjected to permanent occlusion MI. Percent infarct (INF)/left ventricle (LV), percent area-at-risk (AAR)/LV, and percent INF/AAR were measured one week after MI (n = 5 Aif1+/+ and n = 6 Aif1−/−, two-tailed unpaired t test, pooled from two independent experiments). B Gene expression in whole cardiac extracts from Aif1+/+ or Aif1−/− bone marrow chimera mice (n = 3 naïve and n = 5 day 3 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). Data are presented as mean ± SEM.

Extended Data Fig. 3. Allograft inflammatory factor 1 (AIF1) in macrophages promotes glycolytic metabolism and inflammatory activation in response to necrotic myocardial cells but is dispensable for macrophage efferocytosis and efferocytic production of IL-10.

A Gene expression in bone marrow-derived macrophages (BMDM) after treatment with necrotic myocardial cell (NMC) extracts or lipopolysaccharide (LPS) (n = 4 untreated, n = 3 NMC or LPS, two-way ANOVA with Tukey’s test, two independent experiments). B Extracellular acidification rate (ECAR) in BMDMs after treatment with NMC extracts (n = 6 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). C Percent efferocytosis of apoptotic Jurkat cells co-cultured with BMDMs. Scale bar, 10μm. D IL-10 production by efferocytic BMDMs (for C and D, n = 4 for all groups, two-tailed unpaired t test, two independent experiments). Data are presented as mean ± SEM.

Extended Data Fig. 4. Allograft inflammatory factor 1 (AIF1) promotes Hypoxia Inducible Factor (HIF)-1α-dependent glycolysis after myocardial infarction (MI).

A HIF-1α activation in cardiac macrophages 3 days after MI (n = 3 day 0 MI and n = 6 day 3 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). B Isolation of cardiac macrophages 3 days after MI with representative enrichment assessed by flow cytometry. C Extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) by cardiac macrophages isolated from the infarct 3 days after MI (n = 4 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). D Publicly available microarray data (GSE112630) of cytokine and glycolytic gene expression among CCR2⁺ and CCR2⁻ macrophages isolated from the hearts of patients during ischemic (ICM) or dilated cardiomyopathy (DCM). Data are presented as mean ± SEM.

Extended Data Fig. 5. Calcium influx is required for inflammatory glycolytic reprogramming of macrophages after TLR4 stimulation.

A Calcium influx in bone marrow-derived macrophages (BMDM) after TLR4 or ionomycin (IM) stimulation. Images are representative of two independent experiments. Scale bar, 20 μm. B Cytokine production by BMDMs treated with a cell-permeable, calcium chelator (BAPTA) (n = 4 for IL-6 and TNF-α and n = 3 for IL-10, two-way ANOVA with Tukey’s test, three independent experiments). Extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) in BMDMs after treatment with calcium chelators, C BAPTA or D EGTA. Vehicle treated groups are the same in C and D (n = 6 vehicle and n = 8 for all other groups, two-way ANOVA with Tukey’s test, two independent experiments). E ECAR and OCR in BMDMs treated with ionomycin (n = 5 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). Data are presented as mean ± SEM.

Extended Data Fig. 6. Allograft inflammatory factor 1 (AIF1) is dispensable for calcium influx in macrophages after TLR4 stimulation.

A Mean fluorescent intensity (MFI) of the calcium indicator dye, Fluo-4, in bone marrow-derived macrophages (BMDM) 30 minutes after TLR4 stimulation as measured by fluorescent microscopy. Scale bar, 10μm (n = 10 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). B Kinetics of Fluo-4 MFI in BMDMs after TLR4 stimulation as measured by a fluorescent plate reader (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). Data are presented as mean ± SEM.

Extended Data Fig. 7. Actin polymerization is required for inflammatory glycolytic reprogramming of macrophages after TLR4 stimulation.

A Rac1 knockdown efficiency in bone marrow-derived macrophages (BMDM) (n = 4 for all groups, two-tailed unpaired t test, two independent experiments. B Cytokine production by BMDMs treated with a cell-permeable, actin polymerization inhibitor, Cytochalasin D (CytoD) (n = 4 for all groups, two-way ANOVA with Tukey’s test, three independent experiments). C Extracellular acidification rate (ECAR) in BMDMs after treatment with CytoD (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). D Actin polymerization in cardiac macrophages 3 days after MI (n = 3 day 0 MI and n = 6 day 3 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). Data presented as mean ± SEM.

Extended Data Fig. 8. Role of Hypoxia inducible factor 1 (HIF)-1α in Rac1 activation, membrane ruffling, glycolysis, and transcription after TLR4 stimulation.

A Rac1-GTP levels in bone marrow derived macrophages (BMDMs) after 1 hour of TLR4 stimulation with lipopolysaccharide (LPS). Dashed line represents background absorbance level (n = 4 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). B F-actin immunostaining of BMDMs after 1 hour of TLR4 stimulation with quantification of membrane ruffling area (μm²) (n = 30 for all groups, whiskers 10–90 percentile, two-way ANOVA with Tukey’s test, two independent experiments). C Extracellular acidification rate (ECAR) in BMDMs after TLR4 stimulation with LPS (n = 12 vehicle and n = 16 LPS, two-way ANOVA with Tukey’s test, two independent experiments). D ECAR of Hif1+/+ and Hif1−/− BMDMs after TLR4 stimulation (n = 14 unstimulated and n = 16 LPS, two-way ANOVA with Tukey’s test, two independent experiments). E Gene expression in BMDMs after TLR4 stimulation (n = 4 for all groups, one-way ANOVA with Dunnett’s test, two independent experiments). F Gene expression in Hif1+/+ and Hif1−/− BMDMs after 3 hours of TLR4 stimulation (n = 3 unstimulated and n = 4 LPS, two-way ANOVA with Tukey’s test, two independent experiments). Data presented as mean ± SEM.

Extended Data Fig. 9. Allograft inflammatory factor 1 (Aif1) antisense oligonucleotide (ASO) dose response.

ASOs were formulated in PBS, administered at doses of 1.6, 8, 40, and 75 mg/kg, and injected subcutaneously on days 0, 6, and 10. Aif1 tissue expression and general ASO tolerability was evaluated on day 14. Gene expression levels of Aif1 in A heart and B liver. Gene expression levels of Ccl2 in C heart and D liver. E Liver or F spleen weight as a percent of total body weight. Plasma levels of G alanine transaminase or H creatinine (n = 6 PBS, n = 3 for 1.6, 8, and 40 mg/kg, and n = 2 for 75 mg/kg, one-way ANOVA with Dunnett’s test, pooled from two independent experiments). Data presented as mean ± SEM.

Extended Data Fig. 10. Mechanical regulation of macrophage metabolism by Allograft Inflammatory Factor 1 (AIF1) leads to adverse remodeling after cardiac injury.

Myocardial infarction leads to necrotic death of cardiomyocytes and release of damage associated molecular patterns (DAMPs). Recognition of DAMPs on macrophages by Toll-like Receptors (TLR), including TLR4, leads to calcium influx and activation of Allograft Inflammatory Factor 1 (AIF1). AIF1 interacts with RAC1 to promote actin polymerization and mechanical stiffness. This activates NADPH oxidase (NOX) and increases reactive oxygen species (ROS) production. Elevated ROS levels stabilize Hypoxia Inducible Factor (HIF)-1α, leading to its nuclear translocation and transcription of genes involved in glycolysis and inflammation. This switch to glycolytic metabolism fuels increased production of proinflammatory cytokines, including interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α, which contribute to adverse left ventricular remodeling and progression to heart failure (HF).

Figure 1.. Allograft Inflammatory Factor 1 is increased in cardiac macrophages after acute MI (AMI) in humans.

A Identification of 9 distinct clusters by uniform manifold approximation and projection (UMAP) from the single nuclei transcriptomics dataset of Kuppe et al. B Feature plots representing expression of macrophage marker genes. C Violin plot of AIF1 expression among the distinct clusters. D Ridge plot of AIF1 expression in macrophages from ischemic (IZ) or remote zone (RZ) of patients after AMI compared to non-MI controls (n = 3 control, n = 4 RZ AMI, n = 4 IZ AMI, one-way ANOVA with Tukey’s test). E Pathway enrichment of differentially expressed genes in macrophage cluster. F Histological analyses of AIF1 in CD68⁺ macrophages (arrows) in cardiac autopsy samples collected at different timepoints after AMI. Patients with non-MI deaths were used as controls. Scale bar, 50μm (n = 3 for all groups, one-way ANOVA with Dunnett’s test). Data are presented as mean ± SEM.

Figure 2.. Allograft Inflammatory Factor 1 is increased in cardiac macrophages after MI in mice.

A Histological analysis of AIF1 in CX3CR1+ macrophages (arrows) in the hearts of mice before or after MI. Scale bar, 20μm. Expression of AIF1 in infarct-associated B macrophages, C Ly6C^hi monocytes, and D neutrophils after MI in mice. Dashed lines represent staining controls using Aif1−/− mice (n = 3 day 0 MI, n = 5 day 3 MI, and n = 3 day 7 MI, one-way ANOVA with Tukey’s test, pooled from two independent experiments). E Gating strategy for CCR2⁺ and CCR2⁻ cardiac macrophages. F Expression of AIF1 in infarct-associated CCR2⁺ and CCR2⁻ cardiac macrophages after MI in mice. Dashed line represents fluorescent minus one (FMO) staining control (n = 4 for all groups, two-way ANOVA with Tukey’s test, pooled from two independent experiments). Data are presented as mean ± SEM.

Figure 3.. Bone marrow-derived Allograft inflammatory factor 1 (AIF1) worsens cardiac repair after MI.

A Experimental design of Aif1+/+ and Aif1−/− bone marrow chimeras (BMC). B Gene expression of Aif1 in whole cardiac extracts after MI (n = 3 day 0 MI, n = 5 day 3 MI, n = 3 day 7 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). C Percent infarct (INF)/left ventricle (LV), percent area-at-risk (AAR)/LV, and percent INF/AAR measured one week after MI (n = 6 Aif1+/+ and n = 7 Aif1−/−, two-tailed unpaired t test, pooled from three independent experiments). D M-mode echocardiography measurements four weeks after sham or MI surgery with quantification of percent ejection fraction (EF), percent fractional shortening (FS), LV systolic and diastolic volume (microliter), LV mass (milligrams), interventricular septal diameter (IVSD, millimeters), and LV wall thickness (millimeters) (n = 5 Aif1+/+ sham and MI, n = 5 Aif1−/− sham, and n = 6 Aif1−/− MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). Data are presented as mean ± SEM.

Figure 4.. Bone marrow-derived Allograft inflammatory factor 1 (AIF1) promotes inflammatory macrophage responses after MI.

A Infarct-associated neutrophil, Ly6C^hi monocyte, and cardiac macrophage responses in Aif1+/+ or Aif1−/− bone marrow chimeric mice after MI. Neutrophil and Ly6C^hi monocyte abundance in B peripheral blood, C spleen, and D bone marrow after MI. For A-D, n = 3 day 0 MI, n = 4 day 1 MI, n = 5 day 3 MI, and n = 3 day 7 MI, two-way ANOVA with Tukey’s test, pooled from three independent experiments. E CD80 and F CD206 expression on cardiac macrophages after MI (n = 3 day 0 MI, n = 5 day 3 MI, and n = 3 day 7 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). G Efferocytosis of apoptotic mCherry-expressing cardiomyocytes by cardiac macrophages 3 days after MI (n = 5 for all groups, two-tailed unpaired t test, pooled from two independent experiments). H Gene expression of inflammatory mediators in whole cardiac extracts after MI n = 3 day 0 MI, n = 5 day 3 MI, and n = 3 day 7 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). Data are presented as mean ± SEM.

Figure 5.. Allograft Inflammatory Factor 1 (AIF1) promotes HIF-1⍺-dependent proinflammatory glycolytic reprogramming of macrophages after TLR4 stimulation.

A Cytokine gene expression in bone marrow-derived macrophages (BMDM) after 6 hours of TLR4 stimulation (n = 4 for all groups, two-way ANOVA with Tukey’s test, three independent experiments). B Cytokine production by BMDMs after 6 hours of TLR4 stimulation (n = 3 for all groups, two-way ANOVA with Tukey’s test, three independent experiments). C Gene or D protein expression of hypoxia inducible factor (HIF)-1⍺ in BMDMs after 1 hour of TLR4 stimulation (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). E Glycolytic gene expression in BMDMs after 3 hours of TLR4 stimulation (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). F Extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) in BMDMs after 3 hours of TLR4 stimulation (n = 6 for all groups, two-way ANOVA with Tukey’s test, three independent experiments). G ECAR in BMDMs treated with cobalt chloride (CoCl₂) during TLR4 stimulation n = 5 for all groups, one-way ANOVA with Tukey’s test, three independent experiments). H Glycolytic or I cytokine gene expression in BMDMs treated with CoCl₂ during TLR4 stimulation (n = 3 for all groups, one-way ANOVA with Tukey’s test, two independent experiments). Data are presented as mean ± SEM.

Figure 6.. Allograft Inflammatory Factor 1 (AIF1) interacts with RAC1 to promote proinflammatory glycolytic reprogramming of macrophages after TLR4 stimulation.

A F-actin immunostaining of bone marrow-derived macrophages (BMDMs) after 1 hour of TLR4 stimulation with lipopolysaccharide (LPS) and treatment with cytochalasin D (CytoD) or gene silencing of Aif1 or Rac1 with quantification of membrane ruffling area (μm²) (Scale bar, 5 μm, n = 30 for all groups, whiskers 10–90 percentile, two-way ANOVA with Tukey’s test, three independent experiments). B Atomic Force microscopy of BMDMs after 1 hour of TLR4 stimulation with quantification of Young’s modulus (Scale bar, 5 μm, n = 40 Aif1+/+ and Aif1−/− Ø, n = 32 Aif1+/+ LPS, n = 38 Aif1−/− LPS, whiskers 10–90 percentile, two-way ANOVA with Tukey’s test, two independent experiments). C Rac1-GTP levels in BMDMs after 1 hour of TLR4 stimulation. Dashed line represents background absorbance level (n = 4 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). D AIF1 and RAC1 immunostaining of BMDMs after 1 hour of TLR4 stimulation. Images are representative of three independent experiments (Scale bar, 5 μm). E Proximity ligation assay between AIF1 and RAC1 in BMDMs after 1 hour of TLR4 stimulation and treatment with the calcium chelator, BAPTA, with quantification of mean fluorescence intensity (n = 3 for all groups, two-way ANOVA with Tukey’s test, three independent experiments). F Immunoblot of hypoxia inducible factor (HIF)-1⍺ in BMDMs after 1 hour of TLR4 stimulation (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). G. Extracellular acidification rate (ECAR) in BMDMs after 3 hours of TLR4 stimulation (n = 6 siCtrl Ø and n = 8 for siCtrl LPS, siRac1 Ø, and siRac1 LPS, two-way ANOVA with Tukey’s test, three independent experiments). H Cytokine production by BMDMs after 6 hours of TLR4 stimulation (n = 3 for all groups, two-way ANOVA followed by Tukey’s test, three independent experiments). Data are presented as mean ± SEM.

Figure 7.. Allograft Inflammatory Factor 1 (AIF1)-dependent proinflammatory glycolytic reprogramming of macrophages after TLR4 stimulation requires NADPH Oxidase activation.

Immunoblot of hypoxia inducible factor (HIF)-1⍺ in bone marrow-derived macrophages (BMDM) treated with NADPH oxidase (NOX) inhibitors, A apocynin or B GSK2795039 (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). C Extracellular acidification rate (ECAR) in BMDMs treated with NOX inhibitors (n = 6 vehicle Ø and n = 8 for all other groups, two-way ANOVA with Tukey’s test, three independent experiments). D Cytokine production by TLR4-stimulated BMDMs treated with NOX inhibitors (n = 4 for all groups, one-way ANOVA with Tukey’s test, three independent experiments). E Reactive oxygen species (ROS) production in BMDMs after 1 hour of TLR4 stimulation as measured by the ROS indicator, 2’,7’-dichlorofluorescein (DCF) (n = 4 for all groups, two-way ANOVA with Tukey’s test, three independent experiments). F Immunoblot of phospho-p40phox in BMDMs after 1 hour of TLR4 stimulation (n = 3 for all groups, two-way ANOVA with Tukey’s test, two independent experiments). Data are presented as mean ± SEM.

Figure 8.. Therapeutic targeting of Allograft inflammatory factor 1 (AIF1) with antisense oligonucleotides (ASO) dampens inflammation and improves cardiac repair after MI.

A Experimental design using control (Ctrl) or Aif1 ASOs in C57BL/6J mice after MI. B Gene expression of Aif1 in whole cardiac extracts after MI (n = 3 day 0 MI and n = 5 day 3 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). C AIF1 expression in infarct-associated cardiac macrophages 3 days after MI. Dashed line represent fluorescence minus one (FMO) staining control (n = 5 for all groups, two-tailed unpaired t test, pooled from two independent experiments). D Infarct-associated cellular responses after MI (n = 3 day 0 MI and n = 5 day 3 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). E Absolute number of TIM4+ cardiac resident macrophages 3 days after MI (n = 5 for all groups, two-tailed unpaired t test, pooled from two independent experiments). F Gene expression of inflammatory mediators in whole cardiac extracts after MI (n = 3 day 0 MI and n = 5 day 3 MI, two-way ANOVA with Tukey’s test, pooled from two independent experiments). G Percent infarct (INF)/left ventricle (LV), percent area-at-risk (AAR)/LV, and percent INF/AAR measured 7 days after MI (n = 5 for all groups, two-tailed unpaired t test, pooled from two independent experiments). H M-mode echocardiography measurements 28 days after MI with quantification of percent ejection fraction (EF), percent fractional shortening (FS), LV systolic and diastolic volume (microliter), LV mass (milligrams), internal diameter (millimeters), and LV wall thickness (millimeters) (n = 6 for all groups, two-tailed unpaired t test, pooled from two independent experiments). Data are presented as mean ± SEM.