Natural killer cells limit cardiac inflammation and fibrosis by halting eosinophil infiltration

Abstract

Myocarditis is a leading cause of sudden cardiac failure in young adults. Natural killer (NK) cells, a subset of the innate lymphoid cell compartment, are protective in viral myocarditis. Herein, we demonstrated that these protective qualities extend to suppressing autoimmune inflammation. Experimental autoimmune myocarditis (EAM) was initiated in BALB/c mice by immunization with myocarditogenic peptide. During EAM, activated cardiac NK cells secreted interferon γ, perforin, and granzyme B, and expressed CD69, tumor necrosis factor-related apoptosis-inducing ligand treatment, and CD27 on their cell surfaces. The depletion of NK cells during EAM with anti-asialo GM1 antibody significantly increased myocarditis severity, and was accompanied by elevated fibrosis and a 10-fold increase in the percentage of cardiac-infiltrating eosinophils. The resultant influx of eosinophils to the heart was directly responsible for the increased disease severity in the absence of NK cells, because treatment with polyclonal antibody asialogangloside GM-1 did not augment myocarditis severity in eosinophil-deficient ΔdoubleGATA1 mice. We demonstrate that NK cells limit eosinophilic infiltration both indirectly, through altering eosinophil-related chemokine production by cardiac fibroblasts, and directly, by inducing eosinophil apoptosis in vitro. Altogether, we define a new pathway of eosinophilic regulation through interactions with NK cells.

Figure 1

Depletion of natural killer (NK) cells increases cardiac inflammation and the severity of experimental autoimmune myocarditis (EAM). A: Schematic of phosphate-buffered saline (PBS), rabbit IgG, and asialogangloside GM-1 (ASGM-1) antibody treatment schedule throughout EAM. Representative histology from rabbit IgG (RaIgG) and ASGM-1–treated animals (B) and scores of hematoxylin and eosin–stained cardiac sections from PBS, rabbit IgG, and ASGM-1 antibody treated animals at day 21 of EAM (analysis of variance P < 0.01; C). D: Total CD45⁺ cells infiltrating the hearts of isotype control and ASGM-1–treated animals at day 21 of EAM, as assessed by flow cytometry (P < 0.001). E: Levels of anti-myocarditogenic peptide of cardiac myosin heavy chain (MyHC)_614-629 total IgG antibodies in the serum of PBS, isotype control, and ASGM-1–treated animals at day 21 of EAM (analysis of variance P < 0.001). Percentage of CD3⁻DX5⁺NKp46⁺ NK cells in the heart (P < 0.001) and spleen (P < 0.002) (F) and CD3⁺DX5⁺ NK T cells out of CD45⁺ cells (G) of rabbit IgG and ASGM-1 monoclonal antibody–treated wild-type (WT) mice at day 21 of EAM, as assessed by flow cytometry. ^∗∗P < 0.05, ^∗∗∗P < 0.001.

Figure 2

Depletion of natural killer cells increases collagen deposition and fibrosis during experimental autoimmune myocarditis (EAM). A: Representative histology of Masson's trichrome–stained cardiac sections from rabbit IgG (RaIgG) and asialogangloside GM-1 (ASGM-1)–treated animals at day 21 of EAM. B: Enumeration of fibrosis by calculation of red versus blue pixels on ImageJ software (P = 0.012). Statistics calculated by unpaired t-test. C: Ejection fraction (P = 0.026) calculated from fractional shortening and left ventricular (LV) diastole and systole. LV end diastolic dimension (LVEDD; D) and LV end systolic dimension (LVESD; P = 0.03; E). F: Intraventricular septum diameter (IVS), as measured by echocardiography. Statistics by ordinary one-way analysis of variance with post testing by Tukey's multiple-comparisons test. Collagen production as measured using real-time quantitative PCR for collagen 1a1 (Col1a1; P = 0.04; G), collagen 1a2 (Col1a2; P = 0.03; H), and collagen 3a1 (Col3a1P = 0.02; I). Values calculated as a function of hypoxanthine-guanine phosphoribosyltransferase (HPRT) and compared against rabbit IgG. Statistics calculated by unpaired t-test. ^∗P < 0.05.

Figure 3

Activated natural killer (NK) cells accumulate in the heart during experimental autoimmune myocarditis (EAM). A: Gating strategy for cardiac CD3⁻DX5⁺NKp46⁺ NK cells. Absolute numbers (analysis of variance P = 0.002; B) and percentages (C) of CD3⁻DX5⁺NKp46⁺ NK cells on days 0, 14, and 21 of EAM by flow cytometry. Significance by ordinary one-way analysis of variance with post testing by Tukey's multiple-comparisons test. To determine NK physiology, cardiac and splenic CD3⁻DX5⁺NKp46⁺ NK cells were stained at day 21 of EAM. Percentage of interferon γ (IFNγ; P < 0.001; D), Il-13 (E), and perforin (P < 0.001; F) positive cells were based on intracellular antibody staining after 4 to 6 hours of 4β-phorbol 12-myristate 13-acetate/ionomycin and Golgistop in vitro. G: Percentage of lysosomal-associated membrane protein 1 (LAMP-1⁺), a marker for granzyme B release, staining (P < 0.001). Percentage of NK cells positive for activation-associated markers CD27 (P < 0.001; H), CD69 (P < 0.001; I), tumor necrosis factor–related apoptosis-inducing ligand treatment (TRAIL; P < 0.001; J), and NKG2D (P < 0.001; K). D–F: Significance calculated by unpaired t-test. ^∗∗P < 0.01, ^∗∗∗P < 0.001. FSC, forward scatter; SSC, side scatter.

Figure 4

Depletion of natural killer (NK) cells increases eosinophilic infiltration in the heart. A: Representative side-scatter histograms of isotype and asialogangloside GM-1 (ASGM-1) antibody treated animals at day 21. Percentage of Ly6G^loSiglecF⁺ eosinophils from cardiac CD45⁺ cells on day 14 (P = 0.043) and day 21 (P < 0.001) (B) and SiglecF⁻Ly6G^hi neutrophils at days 14 and 21 (C). Significance calculated by unpaired t-test. ^∗P < 0.05, ^∗∗∗P < 0.001. SSC, side scatter.

Figure 5

Activated eosinophils infiltrate the heart in the absence of natural killer cells. A: Immunofluorescence staining of paraffin-embedded cardiac sections from asialogangloside GM-1 (ASGM-1)–treated mice on day 21 of experimental autoimmune myocarditis (EAM). Sections stained with 1:500 rat anti-major basic protein (MBP) primary or rat IgG antibody, 1:200 anti-rat donkey phosphatidylethanolamine (PE)–Texas Red antibody, and DAPI. Relative mean fluorescence intensity of SiglecF (P < 0.001; B) and CD11b (P < 0.001; C) on Ly6GloSiglecF⁺ eosinophils from the heart and spleen of ASGM-1–treated animals at day 21 of EAM by flow cytometry. Levels of eosinophil peroxidase (Epx; P = 0.05; D), major basic protein 2 (Prg2; P = 0.05; E), Il1b (P = 0.04; F), chemokine ligand (Ccl) 11 (P = 0.03; G), Il6 (P = 0.02; H), Il4 (I), and Il13 (J) mRNA from fluorescence-activated cell sorted Ly6G^loSiglecF^hi eosinophils from ASGM-1–treated mice at day 21 of EAM. Values shown as fold induction compared to spleen and controlled against hypoxanthine-guanine phosphoribosyltransferase (HPRT). Significance calculated by Student's t-test. ^∗P < 0.05, ^∗∗∗P < 0.001. MFI, mean fluorescence intensity.

Figure 6

Eosinophil-deficient mice are phenotypically unresponsive to natural killer depletion. Representative histology from rabbit IgG (RaIgG) and asialogangloside GM-1 (ASGM-1) antibody-treated (A) and scores of rabbit phosphate-buffered saline, IgG, and ASGM-1 antibody-treated ΔdblGATA1 animals (analysis of variance; B). Representative histology (C) and percentage of fibrosis (D) by Masson's trichrome staining of cardiac sections at day 21 of experimental autoimmune myocarditis (EAM) from rabbit IgG and ASGM-1–treated mice. Ejection fraction (E), left ventricular end diastolic dimension (LVEDD; F), and left ventricular end systolic dimension (LVESD; G) by echocardiography at day 21. Collagen 1 and 3 production as measured using real-time quantitative PCR for collagen 1a1 (Col1a1; H), collagen 1a2 (Col1a2; I), and collagen 3a1 (Col3a1; J) mRNA in hearts of rabbit IgG and ASGM-1–treated animals at day 21 of EAM. Values calculated as a function of hypoxanthine-guanine phosphoribosyltransferase (HPRT) levels and compared against rabbit IgG. Statistics calculated by unpaired t-test.

Figure 7

Chemokine ligand (Ccl) 11 is involved, but not required, for the suppression of eosinophils by natural killer (NK) cells. Ccl11 mRNA at day 21 of experimental autoimmune myocarditis (EAM) in whole heart (P = 0.004; A) and isolated cardiac fibroblasts of rabbit IgG and asialogangloside GM-1 (ASGM-1)–treated animals (B). C: CCL24 mRNA in isolated cardiac fibroblasts of rabbit IgG (RaIgG) and ASGM-1–treated animals at day 21 of EAM. D: Ccl11 in the supernatant after a 96-hour culture of wild-type cardiac fibroblasts with or without NK cells (P < 0.001). E: Percentage of Ly6G^loSiglecF⁺ eosinophils at day 21 of EAM in rabbit IgG and ASGM-1–treated Ccr3^−/− animals. Significance by Student's t-test. ^∗P < 0.05, ^∗∗∗P < 0.001.

Figure 8

Natural killer (NK) cells induce apoptosis in eosinophils (Eos) in vitro. A: Representative side scatter versus SiglecF and LIVE/DEAD viability dye versus annexin V plots of Ly6G^loSiglecF⁺ eosinophils after a 3-hours co-culture with primary naïve NK cells. Percentage of annexin V–positive (analysis of variance P < 0.001; B) and LIVE/DEAD stain–positive Ly6G^loSiglecF⁺ eosinophils (C). Average mean fluorescence intensity of activation markers SiglecF (D) and CD11b (E) of Ly6G^loSiglecF^hi eosinophils. Statistics calculated by one-way analysis of variance with post testing by Tukey's test. ^∗P < 0.05, ^∗∗∗P < 0.001. SSC, side scatter.