IL-1β enhances susceptibility to atrial fibrillation in mice by acting through resident macrophages and promoting caspase-1 expression

Abstract

Atrial fibrillation (AF) is more prevalent in patients with elevated interleukin (IL)-1β levels. Here we show that daily administration of IL-1β for 15 days sensitizes mice to AF, leading to fibrosis, accumulation of β-pleated sheet proteins in the left atrium, and systemic inflammation, resembling the pathophysiological changes observed in patients with AF. IL-1β administration creates a positive feedback loop, dependent on the IL-1 receptor (IL-1R) activity in cardiac resident macrophages. This results in increased caspase-1 maturation in the left atrium and elevated Il1b and Casp1 transcription in atrial macrophages. IL-1β treatment accelerated action potential and Ca²⁺ restitution in the left atrium, leading to action-potential shortening. This, along with increased caspase-1 maturation and IL-1R signaling, was essential for inducing AF. Lack of IL-1R in macrophages, but not cardiomyocytes, prevented IL-1β-induced AF sensitivity. By depleting recruited macrophages or deleting IL-1R specifically in cardiac resident macrophages, we further demonstrate that IL-1β/IL-1R signaling in these resident macrophages is responsible for increased AF susceptibility. These findings offer insights into the therapeutic potential of targeting IL-1β/IL-1R signaling in patients with AF and emphasize the importance of recognizing different underlying causes in this patient group.

Extended Data Fig. 1 |. Comparative mRNA expression in atrial samples of mice injected with saline or IL1β for 15 consecutives days.

a. Fold change was calculated dividing mean TPM values of IL1β -treated by the control group. Genes from the list of interest (n = 72) with absolute log2-fold change higher than 0.5 were displayed in red. The cumulative expression level of the genes is depicted in the x-axis as the sum of all the transcript per million of the four samples. b. Gene-set enrichment analysis interrogating KEGG data base. c. Gene-set enrichment analysis interrogating Reactome data base. d. Gene-set enrichment analysis interrogating Wikipathways data base. The top 10 up-regulated pathways are shown in blue and the top 10 down-regulated in orange (b-d).

Extended Data Fig. 2 |. IL-1β injection increases circulating neutrophils and macrophages.

a, Mice circulating biomarkers. b, Scheme of experimental design for leukocyte subsets in LA at 15 days post-injections. c, Mouse atria immune cells profile. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated with Mann-Whitney’s test (two-tailed) (a), and unpaired Student’s t-test (two-tailed) (c).

Extended Data Fig. 3 |. Human circulating biomarkers.

Patients in sinus rhythm (SR) or in atrial fibrillation (AF) circulating biomarkers. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated with Mann-Whitney’s test (two-tailed).

Extended Data Fig. 4 |. IL-1β injection does not produces maturation of IL-1β in left atria of Casp1^−/−.

a, Serum levels of IL-6 in saline- and IL-1β–injected Casp1^−/− mice. b, Western blot of IL-1β/pro-IL-1β in Casp1^−/− mice. Quantitation levels were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) loading control. Each sample represents one mouse heart; 5 saline-injected and 6 IL-1β-injected, samples were analyzed. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated with Mann-Whitney’s test (two-tailed) (a) and Student’s t-test (two-tailed) (b).

Extended Data Fig. 5 |. PKA and CaMKII expressions and the effect of Nifedipine on LA.

a, Western blot of Calcium-Calmodulin Kinase II δ (CaMKII), phospho-CaMKII Thr287 (p-CaMKII) and oxidized-CaMKII Met 281/282 (oxi-CAMKII). Quantitation levels were normalized to vinculin, used as loading control. Each sample represents one mouse heart; 5 saline-injected and 3 IL-1β-injected samples were analyzed. b Western blot of catalytic subunit of Protein kinase A (PKA). Quantitation levels were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) loading control. Each sample represents one mouse heart; 8 saline-injected and 8 IL-1β-injected samples were analyzed (see the second membrane in uncropped Material). c, APDs before and after Nifedipine perfusion (at 30%, 50%, 70%, and 90% of repolarization). Data are presented as individual values before and after condition. Data in a and b are presented as the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. Data in c is presented as individual values before and after condition. P values were calculated with unpaired Student’s t-test (two-tailed) (a,b), and paired samples Student’s t-test (two-tailed) (c).

Extended Data Fig. 6 |. Action Potential after 4-aminopyridine (4-AP) and atropine (Atrop) perfusion.

a, Graphical scheme of action potential recording after 15 min of 4-AP 2.5 mM perfusion. b, Representative traces of AP from saline and IL-1β groups perfused with Tyrode solution or 4-AP. c, Action potential duration (APD) at 30%, 50%, 70% and 90% of repolarization, from saline and IL-1β groups during perfusion with Tyrode (Tyr) or 4-AP. Data are presented as individual values before and after condition. d, Graphical scheme of action potential recording after 15 min of Atropine 10 μM perfusion. e, Representative traces of AP from saline and IL-1β groups perfused with Tyrode solution or atropine. f, APD at 30%, 50%, 70% and 90% of repolarization, from saline and IL-1β groups during perfusion with Tyrode (Tyr) or atropine (Atrop). Data are presented as individual values before and after condition. P values in c and f were calculated with paired samples Student’s t-test (two-tailed).

Extended Data Fig. 7 |. IL-1β injection increases arrhythmia susceptibility in a mechanism dependent of Casp1 expression and not IL-6.

a, graphical scheme of experiments. Casp1^−/− mice were treated for 15 days with saline or IL-1β subcutaneously (SC). b, representative EKG traces from saline and IL-1β treated mice. c, EKG parameter analyzed: RR interval, p wave duration, and PR interval. d, representative action potential traces from saline and IL-1β treated Casp1^−/− mice. e, action potential duration (APD) at 30%, 50%, 70% and 90% of repolarization, from saline and IL-1β treated mice. Lower inset panel show adjusted mean ± 95% C.I.f, triggered activities in Casp1^−/− mice. g, graphical scheme of IL-6 restitution in Casp1^−/− mice injected with IL-1β. h, representative EKG trace in sinus rhythm after burst-pacing in Casp1^−/− mice injected with IL-1β and IL-6. i, AF inducibility after burst-pacing in Casp1^−/− mice injected with IL-1β and IL-6. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. Categorical data are presented as bar plots (f,i). P values were calculated with Student’s t-test (two-tailed) (c), hierarchical analysis with mixed effect model (e), and Fisher’s exact test (f,i) (two-tailed).

Extended Data Fig. 8 |. IL-1β injection increases arrhythmia susceptibility through activation of IL-1R.

a, graphical scheme of experiments. IL-1r^−/− mice were treated for 15 days with saline or IL-1β subcutaneously (SC). b, representative EKG traces from saline and IL-1β treated mice. c, EKG parameters analyzed: RR interval, p wave duration, and PR interval. d, representative action potential traces from saline and IL-1β treated mice. e, action potential duration (APD) at 30%, 50%, 70% and 90% of repolarization, from saline and IL-1β treated mice. Lower inset panel show adjusted mean ± 95% C.I.All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. Categorical data are presented as bar plots (f). P values were calculated with unpaired Student’s t-test (two-tailed) (c), hierarchical analysis with mixed effect model (e), and Fisher’s exact test (two-tailed) (f).

Extended Data Fig. 9 |. IL-1β injection does not produces maturation of IL-1β in left atria of Csf1r^Cre IL-1r^fl/fl.

a, graphical scheme of experiments. Csf1r^Cre IL-1r^fl/fl mice were treated for 15 days with saline or IL-1β subcutaneously (SC). b, Western blot of IL-1β/pro-IL-1β in Csf1r^Cre IL-1r^fl/fl mice. Quantitation levels were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) loading control. Each sample represents one mouse heart; 5 saline-injected and 6 IL-1β-injected samples were analyzed. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated with Student’s t-test.

Fig 1.. IL-1β sensitizes mice to atrial fibrillation.

a, Schematic illustration of the experimental design for transoesophageal pacing. b, Representative EKG traces obtained during and after transoesophageal burst pacing. c, number of AF episodes/number of burst-pacing stimulations. d, Cumulative bar plots of AF inducibility during the whole burst-pacing protocol. e, Representative RR interval series after burst pacing. f, Average RR interval duration after burst pacing. g, Average standard deviation (SD) of RR intervals. h, Percentage of RR intervals 5 ms above or below RR mean duration. i, Heart rate variability (HRV) triangular index. j, Multiscale sample entropy of saline- and IL-1β-injected mice. All central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values for quantitative data. Categorical data are presented as relative frequencies and percentages in cumulative bar plots (d). In Figure 1j the data are present as mean ± s.d. P values were calculated using Mann-Whitney test (two-tailed) (c), two-tailed Fisher’s exact test (d), and unpaired Student’s t-test (two-tailed) (f-j).

Fig. 2.. IL-1β injection increased collagen deposition in a disarranged pattern.

a, b, Representative three-dimensional reconstruction of second harmonic generation (SHG) microscopy and two-dimensional reconstruction images of left atria from saline-injected (left) and IL-1β–injected (right) mice. Collagen fibres appear in green. c, Integrated optical density of SHG signals for collagen. d, Representative fast-Fourier-transform of SHG microscopy images of left atria from saline-injected (left) and IL-1β–injected (right) mice. e, Collagen orientation index (COI). All central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated using unpaired Student’s t test (two-tailed) (c,e).

Fig. 3.. IL-1β injections increase mature IL-1β and caspase-1 in mouse left atria.

a, Schematic illustration of the experimental design. C57Bl/6 mice were injected with IL-1β (4 ng) or saline daily for 15 days. b, Circulating biomarkers in mouse. c, Number, gender, and age of patients in sinus rhythm (SR) and with AF. d, Circulating biomarkers in patients in SR and with AF. e, Schematic illustration of LA whole-tissue lysate western blotting. f. Western blots for IL-1β/pro–IL-1β at 1, 4, 8, and 15 days, in LA whole-tissue lysates from saline- and IL-1β–injected mouse. Quantitation was normalized to GAPDH. Each sample represents one mouse heart; 6 saline-injected and 6 IL-1β-injected, 4 saline-injected and 5 IL-1β-injected, 4 saline-injected and 5 IL-1β-injected and 7 saline-injected and 7 IL-1β-injected samples were analyzed. g. The graph shows the differences on mature IL-1β concentration. The figure on the right panel depicts the temporal fold of change (IL-1β treated/saline). h. Caspase-1 (Casp1) /pro-Casp1 expression in LA whole-tissue lysates from saline- and IL-1β–injected mouse. Quantitation was normalized to GAPDH. Each sample represents one mouse heart; 24h: 5 saline-injected and 7 IL-1β-injected, 15: 7 saline-injected and 7 IL-1β-injected, samples were analyzed. i. Schematic illustration of the experimental design. Bone marrow derived-macrophages (BMDM) culture exposed or not to IL-1β. j. The heat maps show the different gene expression data of BMDM from WT mice (left map) after 1 or 3 days of IL-1β or not. The right map shows the different gene expression data of BMDM from WT, Casp1^−/− mice or IL-1r^−/− mice (left map) after 3 days exposure to IL-1β or not. Each column represents data from one experiment (n = 3 independent samples for both groups). Each row represents a single gene. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. The data in c are presented as % (n) and mean ± s.d. (Age). P values were calculated using the Mann–Whitney (two-tailed) (b, d) and unpaired Student’s t tests (two-tailed) (g–h).

Fig. 4.. IL-1β injections shorten the action potential (AP) duration and shape Ca²⁺-transient (Ca²⁺-T) kinetics.

a, Schematic illustration of the experimental design for electrophysiological studies using AP recording and local field fluorescence microscopy (LFFM). b, Representative LA AP traces. c, LA APDs at 30%, 50%, 70% and 90% repolarization. Detailed hierarchical analysis results are provided in Supplementary Table 3. Lower inset panel show adjusted mean ± 95% C.I.. d, Representative AP traces from the restitution experiments. e, f. APD restitution at 90% repolarization and slope thereof. Pooled group data were fitted with an exponential decay equation to calculate time constant (τ) values (n = number of hearts/mice). g, Representative normalized traces of Ca²⁺-T, recorded with local field fluorescence microscopy. h, Kinetic Ca²⁺-T parameters: rise time, time to peak, half duration and fall time. i, Western blots of Ryanodine (Ry) receptor and sarcoplasmic reticulum Ca²⁺ATPase 2a (SERCA2a). Quantitation was normalized to vinculin. Each sample represents one mouse heart; RyR: 7 saline-injected and 7 IL-1β-injected and SERCA2a: 5 saline-injected and 5 IL-1β-injected, samples were analyzed. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated using hierarchical analysis with a mixed-effects model (c) and using unpaired Student’s t test (two-tailed) (f,h,i).

Fig. 5.. IL-1β injections increase triggered activities

a, Representative heart traces of Ca²⁺-T amplitude restitution (left) and restitution curves as a function of the S2-S1 interval (right – n = number of hearts/mice). b, Time constant (τ) values of fitted exponential curves are also presented. c, Schematic illustration of the experimental design for Ca²⁺-T amplitude studies. Ry (10 μM) and Tg (2μM) perfusion. d, Representative traces of Ca²⁺-T amplitudes, normalised by those recorded with Tyrode perfusion. e, Ca²⁺-T amplitude measurements showing fractional inhibition after Ry + Tg perfusion. f, Schematic illustration of the experimental design for AP recording. Nifedipine (5 μM) perfusion. g, Representative traces of LA APs before and after 15 min of nifedipine perfusion (For APD measurements, see Extended Data Fig. 5c). h, Western blots of Cav1.2. Quantitation was normalized to GAPDH. Each sample represents one mouse heart; 6 saline-injected and 6 IL-1β-injected samples were analyzed. i, Schematic illustration of the experimental design for AP recording. j, Representative LA AP traces from WT IL-1β–injected mice recorded under an extra-systolic stimulation (S1-S2) protocol. I, regular pacing (S1); II, 1:1 S1-S2 stimulation response; III, onset of S1-S2–induced atrial triggered activities (TA). k, Cumulative bar plot of the TA incidence in saline- and IL-1β–injected WT mice perfused with Tyrode (Control solution), 4-aminopyridine (2.5 mM) or atropine (10 μM). In b, the data are presented as mean ± s.d.. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values. P values were calculated using Mann-Whitney test (two-tailed) (e), unpaired Student’s t test (two-tailed) (h), and Fisher’s exact test (two-tailed) (k).

Fig 6.. IL-1β injections require Casp1 expression to sensitise to atrial fibrillation.

a, Schematic illustration of the experimental design for transoesophageal pacing. b, Representative EKG traces obtained during and after (see “Burst pacing end”) transesophageal burst pacing. c, number of AF events/number of burst-pacing stimulations. d, Cumulative bar plots of AF inducibility during the whole burst-pacing protocol of WT and Casp1^−/−mice. e, Schematic illustration of IL-6 neutralization with Anti-IL-6 mAb in WT mice injected simultaneously with IL-1β. f, Serum IL-6 levels of WT mice injected with saline, IL-1β, and IL-1β + Anti-IL6 mAB. g, Representative EKG trace obtained during and after transesophageal burst pacing. h, number of AF events/number of burst-pacing stimulations. i, Cumulative bar plots of AF inducibility during the whole burst-pacing protocol. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values for quantitative data. Categorical data are presented as relative frequencies and percentages in cumulative bar plots (d,i). P values were calculated using Mann-Whitney test (two-tailed) (c), Fisher’s exact test (two-tailed) (d,i) and the Kruskal–Wallis test with the Dunn post-test (two-tailed) (f,h).

Fig 7.. IL-1β injections require IL-1R expression macrophages, but not in cardiomyocyte to sensitise to atrial fibrillation.

a, Schematic illustration of the experimental design for burst pacing in WT and IL-1r^−/− mice. b, Representative EKG traces during and after burst pacing in WT and IL-1r^−/− mice. c, number of AF events/number of burst-pacing stimulations. d, Cumulative bar plots of AF inducibility during the whole burst-pacing protocol of WT and IL-1r^−/−mice. e, Schematic illustration of experimental design in CreLox mice: top panel shows WT mice, expressing IL-1R in cardiomyocytes, monocytes and resident macrophages; second row panel, mice lacking of IL-1R only in cardiomyocyes (Myh6^Cre IL-1r^fl/fl ); third row panel, mice lacking of IL-1R expression in monocytes and resident macrophages (Csf1r^Cre IL-1r^fl/fl), under treatment with three cycles of Tamoxifen, (Tx); f, representative EKG traces during and after burst pacing in WT, Myh6^Cre IL-1r^fl/fl and Csf1r^Cre IL-1r^fl/fl (three cycles of tamoxifen). g, Number of atrial fibrillation episodes / number of stimuli after burst pacing. h, Atrial fibrillation inducibility after burst-pacing protocol. All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values for quantitative data. Categorical data are presented as relative frequencies and percentages in cumulative bar plots (d,h). P values were calculated using Fisher’s exact test (two-tailed) (d,h) and Mann-Whitney’s test (two-tailed) (c,g). P values shown in g,h were calculated for each comparison between two groups, comparing with WT + IL-1β AF inducibility.

Figure 8.. IL-1β injections require IL-1R expression in resident macrophages to sensitise to atrial fibrillation.

a, Schematic illustration of experimental design in KO mice: top panel depicts KO mice for Ccr2 (Ccr2^−/−); bottom panel, mice with specific ablation of IL-1R in resident macrophages with a single cycle of tamoxifen. b, representative EKG traces during and after burst pacing in Ccr2^−/− (Upper panel) and Csf1r^Cre IL-1r^fl/fl (Bottom panel; single cycle of tamoxifen). c, Number of atrial fibrillation episodes / number of stimuli after burst pacing (upper panel) and atrial fibrillation inducibility after burst-pacing protocol (bottom panel). d, Schematic illustration of experimental design for cardiac resident macrophage sorting. e, Representative gates of resident macrophages sorting. f, Quantitative PCR with reverse transcription (RT–qPCR) of FACS-purified resident cardiac macrophages (CD11b⁺CD64⁺CCR2⁻) from Wild-type (WT) or one tamoxifen cycle Csf1r^CreIL-1r^fl/fl treated with saline or IL-1β for 15 days. g, representative LA AP traces. h, LA APDs at 30% and 90% repolarization. Detailed hierarchical analysis results are provided in Supplementary Table 3. Lower inset panel show adjusted mean ± 95% C.I..; i, Cumulative bar plot of arrhythmia incidence in Csf1r^Cre IL-1r^fl/fl (single cycle of tamoxifen). All the central lines in the box plots represent the median, the box limits represent the first and third quartiles and the whiskers denote the minimum and maximum values for quantitative data. Categorical data are presented as relative frequencies and percentages in cumulative bar plots (c bottom panel). P values were calculated using ANOVA with post-hoc Bonferroni test (two-tailed) (c upper panel, f), Fisher’s exact test (two-tailed) (c bottom panel) and hierarchical analysis with a mixed-effects model (h).