Loss of developmentally derived Irf8+ macrophages promotes hyperinnervation and arrhythmia in the adult zebrafish heart

Abstract

Recent developments in cardiac macrophage biology have broadened our understanding of the critical functions of macrophages in the heart. As a result, there is further interest in understanding the independent contributions of distinct subsets of macrophage to cardiac development and function. Here, we demonstrate that genetic loss of interferon regulatory factor 8 (Irf8)-positive embryonic-derived macrophages significantly disrupts cardiac conduction, chamber function, and innervation in adult zebrafish. At 4 months post-fertilization (mpf), homozygous irf8^st96/st96 mutants have significantly shortened atrial action potential duration and significant differential expression of genes involved in cardiac contraction. Functional in vivo assessments via electro- and echocardiograms at 12 mpf reveal that irf8 mutants are arrhythmogenic and exhibit diastolic dysfunction and ventricular stiffening. To identify the molecular drivers of the functional disturbances in irf8 null zebrafish, we perform single cell RNA sequencing and immunohistochemistry, which reveal increased leukocyte infiltration, epicardial activation, mesenchymal gene expression, and fibrosis. Irf8 null hearts are also hyperinnervated and have aberrant axonal patterning, a phenotype not previously assessed in the context of cardiac macrophage loss. Gene ontology analysis supports a novel role for activated epicardial-derived cells (EPDCs) in promoting neurogenesis and neuronal remodeling in vivo. Together, these data uncover significant cardiac abnormalities following embryonic macrophage loss and expand our knowledge of critical macrophage functions in heart physiology and governing homeostatic heart health.

Keywords: Up to 10; arrhythmia; development; epicardial; epicardium; heart; irf8; macrophage; single cell transcriptomics; zebrafish.

Figure 1.. Loss of embryonic macrophages produces electrical and transcriptional signatures of cardiac dysfunction.

(A) Cartoon depiction of optogenetic stimulation of channelrhodopsin (ChR2)-positive macrophages with blue light. (B-C) Heart rate of 7 dpf control larvae and larvae with ChR2+ macrophages (B) or neurons (C). Heart rate was determined at three different timepoints using the following stimulation protocol: 30 seconds before stimulation (“No stim.”), 30 seconds during stimulation with 405 nm light (“Stim.”), and 10 minutes post-stimulation (“Recovery”). #p < 0.05, “No Stim.” ChR2+ versus “Stim.” ChR2+; **p < 0.01, “Stim.” Control versus “Stim.” ChR2+. Welch’s t-test. [*]p < 0.05, “No Stim.” Control versus “Stim.” Control. (D-E’) Confocal micrographs of 12 mpf wild type (D-D’) and irf8 mutant (E-E’) hearts with transgenic expression of GFP-labeled macrophages (mpeg1+). Scale = 200 um (D,E) and 50 um (D’,E’). Images at 10x (D,E) and 40x (D’,E’). (F) Representative optical voltage mapping traces of isolated 4 mpf irf8 wild type and mutant (null) hearts bathed in 10 μm di-4-ANNEPS. Atrial activation (black); ventricular activation (red). Gray box indicates atrial action potential duration (APD). (G-I) Quantification of AV delay (G), atrial action potential duration (APD) (H), and ventricular APD (I). n = 9 per genotype. Welch’s t-test.

Figure 2.. Functional swim tunnel challenge reveals irf8 mutants are susceptible to arrhythmia.

(A) Cartoon depiction of the rack-compatible swim tunnel with metered flow. Following a 5 -minute acclimation period within the tunnel, 12 mpf zebrafish were subjected to increasing flow rate intervals for a total of 1 hour. (B) Cartoon depiction of an ECG recording and a representative ECG trace from an anesthetized fish placed ventral-side up in a sponge surrounded by anesthesia solution. (C) Representative ECG traces for wild type and irf8 mutant zebrafish following the swim assay. Black arrows indicate unique electrical disturbances occurring above any baseline noise. (D) Percentage of fish with an abnormal electrical event during the ECG recording. Male n = 7 per genotype, ****p < 0.0001. Female n = 4-8 per genotype, ****p < 0.0001. Fisher's exact test. (E) Categorization of ECG abnormalities observed throughout all wild type and irf8 mutant fish.

Figure 3.. Echocardiographic measurements further support arrhythmicity in irf8 mutants and identify a pumping deficiency.

(A) Cartoon depiction of an adult zebrafish undergoing an echocardiogram with the transducer in the long-axis (LAX) position. Echocardiograms were performed on unchallenged fish. (B) Cartoon of an adult zebrafish heart from the perspective of the echocardiography recording, denoting the atrium (A), atrioventricular (AV) valve, ventriculobulbar (VB) valve, and bulbus arteriosus (BA). The boxes around the AV valve and VB valve depict the gating strategies to determine inward flow and outward flow of blood from the ventricle. (C) Still image from an adult zebrafish echocardiography recording with the ventricle and bulbus arteriosus outlined in red. (D) Representative echocardiogram trace gating on the AV valve with the early filling (E) wave, late filling (A) wave, diastasis time, isovolumic contraction time (IVCT), aortic ejection time (AET), and isovolumic relaxation time (IVRT) labeled. (E-G) Quantifications of IVRT (E), AET (F), and pressure within the VB valve (G) in male and female wild type and irf8 mutant fish at 6 and 12 mpf. Male n = 10-21 per genotype and timepoint. Female n = 9-14 per genotype and timepoint. Each point represents individual beats per fish. 8-10 beats were analyzed per fish. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. (H) Quantification of heart in male and female wild type and irf8 mutants. (I) Representative echocardiogram traces from a wild type and irf8 mutant recording. (J-M) Quantification of diastasis time in wild type males (J) and females (K) at 6 mpf and 12 mpf, as well as diastasis time in irf8 mutant males (L) and females (M) at 6 mpf and 12 mpf. A Gaussian distribution was created by analyzing 8 beats per sample and grouping diastasis duration in defined increments. Male n = 10-21 per genotype and timepoint. Female n = 9-14 per genotype and timepoint. Horizontal bars represent the range of diastasis duration. (N-O) 95% confidence interval of the average diastasis time durations in wild type (gray) and irf8 mutant (blue) males (N) and females (O).

Figure 4.. irf8 mutant hearts are fibrotic and have epicardial abnormalities.

(A) scRNA clustering by Uniform Manifold Approximation and Projection (UMAP) from wild type and irf8 mutant hearts at 12 mpf with cluster 1 demarcated. (B) Enlarged images of cluster 1 showing expression maps of endothelial (flt1, tal1, kdrl) and mesenchymal and pro-fibrotic (cdh11, pdgfrα, fgf10a) transcripts. (C) Cartoon of the epicardium in an adult zebrafish heart. The epicardium is the outermost layer of the heart composed of squamous epithelial cells. Upon cardiac stress or injury, epicardial cells become activated and undergo epithelial-to-mesenchymal transition (EMT) to give rise to smooth muscle cells (SMCs), fibroblasts (FBs), and pericytes (PCs). Image made with Biorender. (D-E) Representative H&E sections of 12 mpf female (top) and male (bottom) wild type (D) and irf8 null (E) hearts. (F-G) Quantification of cardiomyocyte muscle density in females (F) and males (G) at 12 mpf. Female n = 5-8 per genotype. Male n = 6-10 per genotype. Welch’s t-test. *p < 0.05. **p < 0.01. (H) Confocal micrographs of sectioned 12 mpf wild type and irf8 mutant hearts stained for the mesenchymal marker, vimentin (Vim., magenta; gray), cardiac muscle (phalloidin, green), and nuclei (Hoechst, blue). Scale = 200 um. Images at 10x. (I) Epifluorescent images of sectioned 12 mpf wild type and irf8 mutant hearts with transgenic expression of GFP-labeled Tcf21+ nuclei, Tg(tcf21:nucEGFP). The epicardial and subepicardial space (compact myocardium) is marked in blue to demonstrate thickness. (J-L) Quantification of epicardial thickness (J), total epicardial cell number (K), and epicardial cell number relative to their positioning within the epicardial and subepicardial space (L). n = 3-6 per genotype. Welch’s t-test. *p < 0.05.

Figure 5.. Activated EPDCs in irf8 mutant hearts are neurogenic and hearts are hyperinnervated.

(A) Enriched gene ontology analysis of cluster 1 in irf8 mutant scRNA sequencing data at 12 mpf listing the top significantly differentially expressed genes (compared to all other clusters). (B) Co-localization analysis of activated EPDCs (fgf10a+) with neuronal-specific genes (lrtm2a, trim46b, adgrb3). (C-F”) Confocal micrographs of 12 mpf wild type (C-D”) and irf8 mutant (E-F”) hearts stained against anti-acetylated α-tubulin. Micrographs were skeletonized (C’, D’, E’, F’) and analyzed for axonal coverage and compartmental (patches >0.1% of ventricular area; red areas in 4x magnifications in C”, D”, E”, F”). (G-K) Quantification of 12 mpf anti-acetylated α-tubulin stained wild and irf8 mutant hearts for ventricular innervation (G), bulbus arteriosus innervation (H), axonal patch coverage over the ventricular surface (I), number of axonal patches >0.1% of the ventricular area (J), and average axonal patch size as a percent of ventricular area (K). Scale = 200 um. Images at 10x. n = 7-8 per genotype. Welch’s t-test. *p < 0.05; **p < 0.01.