Single-Cell Analyses Reveal IKKα Regulates the Interaction Between Macrophages and T Cells in the Doxorubicin-Induced Cardiomyopathy

Abstract

Doxorubicin (DOX) is a chemotherapeutic agent used to treat solid tumors and hematologic malignancies, though DOX-induced cardiomyopathy poses a risk of severe cardiac impairment and poor prognosis. Immune cells have been increasingly implicated in cardiovascular inflammation, with overproduction of inflammatory cytokines and macrophage accumulation. However, its molecular mechanism remains unclear and needs to be further investigated. This study investigated the involvement of IKKα in regulating cardiac function in response to early-stage DOX stimulation. Results indicated that IKKα^Lyz2−Cre mice were more susceptible to DOX-induced cardiac injury than IKKα^flox/flox mice, showing reduced heart function, extensive cardiac fibrosis, and elevated inflammatory markers. Single-cell transcriptomic analysis revealed cellular heterogeneity in DOX-induced cardiomyopathy tissue between IKKα^flox/flox and IKKα^Lyz2−Cre mice, identifying 11 cell types, 8 of which were immune cells. Bar plots and cell density analysis showed a higher proportion of T cells in IKKα^flox/floxmice, while IKKα^Lyz2−Cre mice had increased monocytes and macrophages. Notably, IKKα deletion promoted a shift in macrophage polarization from Fcna⁺ M2 to Jaml⁺ M1 and impaired T cell activation and differentiation. Additionally, IKKα played a critical role in mediating macrophage-T cell interactions. Loss of macrophage IKKα activated T cells through Jaml⁺ M1 macrophage, and activated T cells subsequently enhanced M1 macrophage activation via IFN-γ and IL-6. These findings highlight the potential of targeting immune cell interactions as a therapeutic strategy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10753-025-02365-7.

Introduction

Doxorubicin (DOX) is a chemotherapeutic drug widely used against solid tumors and hematologic malignancies [1]. However, its dose-dependent cardiotoxicity has limited broader clinical application [2]. DOX-induced cardiomyopathy is associated with a poorer prognosis than other forms of cardiomyopathy [3]. Consequently, understanding the pathological mechanisms underlying DOX-induced cardiomyopathy and developing effective therapeutic strategies are urgently required.

Extensive research has demonstrated that oxidative stress, chronic inflammation, and mitochondrial dysfunction contribute to DOX-induced cardiomyopathy [4–6]. Growing evidence also highlights the essential role of immune cells in maintaining cardiac homeostasis and in the progression of cardiac diseases [7]. Studies have shown that DOX-induced cardiomyopathy is marked by inflammatory cytokine overproduction and accumulation of heterogeneous macrophage populations in the heart, with macrophages undergoing dynamic changes throughout disease progression. Within DOX-affected heart tissue, Treg numbers and functionality are compromised, resulting in elevated IFN-γ production [8]. Additionally, activation of invariant natural killer T cells has been shown to mitigate DOX-induced cardiomyocyte death and cardiomyopathy [9]. However, the precise molecular mechanisms remain unclear, necessitating further investigation into the specific roles and interactions between macrophages and T cells in this context.

The NF-κB signaling pathway is critical in DOX-induced cardiomyopathy [10, 11]. NF-κB activation relies on the degradation of IκBα, a process initiated by phosphorylation through the IκB kinase (IKK) complex, which includes two catalytic subunits, IKKα and IKKβ, along with a regulatory subunit, IKKγ [12]. In macrophages, IKKα limits inflammatory gene activation by providing negative feedback within NF-κB signaling [13]. Our previous studies have demonstrated that macrophage-specific IKKα contributes to ischemic cardiomyopathy development [14]. Moreover, IKKα is essential not only for maintaining Treg cell homeostasis but also for promoting the expansion of Treg and effector CD4⁺ T cells [15]. Additionally, IKKα regulates MAPK phosphorylation to activate T cells [16]. However, the influence of IKKα on macrophage-T cell interactions in DOX-induced cardiomyopathy remains unknown.

In this study, we identified a role for IKKα in the early stages of cardiac injury under DOX stimulation. Single-cell analyses revealed that IKKα knockout shifted macrophage polarization from Fcna⁺ M2 to Jaml⁺ M1. Meanwhile, macrophage IKKα knockout impaired T cell activation and differentiation.

Materials and Methods

Animal Model Establishment

IKKα^flox/flox mice and IKKα^Lyz2−Cre (macrophage IKKα knockout) mice were acquired from the Model Animal Research Center of Nanjing University. At the age of 8 weeks, with weights between 23 and 25 g, mice were maintained in a pathogen-free environment with controlled conditions: temperature set at 22–24 °C, humidity at 50–60%, and a 12-hour light/dark cycle. Following a two-week acclimatization period, the mice were randomly assigned to groups for DOX-induced cardiomyopathy induction. Doxorubicin (DOX, Sigma-Aldrich, D1515) or saline was administered intraperitoneally seven times over two weeks at a 3 mg/kg dose. Daily monitoring was conducted for survival, and cardiac function was evaluated weekly via echocardiography. Isoflurane was used to induce (3%) and maintain (0.5%) anaesthesia for the mice included in this studyconstructing mice model. Euthanasia was achieved by subjecting the mice to accomplished after placing the mice under 100% CO2 inhalation and performing cervical dislocation.

Evaluation of Cardiac Function in Mice

For four weeks following DOX or saline administration, mice were anesthetized with 1% isoflurane weekly to assess cardiac function utilizing the Vevo 2100 Ultrasonic system (VisualSonics, Canada). Trained personnel from the Nanjing Medical University animal center measured left ventricular ejection fraction (EF%) and fractional shortening (FS%) to monitor cardiac performance.

Histological Analysis of Cardiac Tissues

Heart samples were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned into 5 μm slices. Histological staining was performed, including hematoxylin-eosin (HE) and Masson staining. Fibrosis analysis was performed on six non-overlapping fields per heart section, calculating the fibrotic area as the ratio of connective tissue area to total tissue area.

Single-Cell Solution Preparation

Mouse heart tissue was sectioned into 1–2 mm³ pieces and subjected to enzymatic digestion at 37 °C for 30–60 min. Digestion was stopped by adding RPMI-1640 medium, and the resulting cell suspension was filtered through a 4 μm cell strainer. The prepared single-cell suspension was maintained on ice until it was loaded into the Chromium Controller (10x Genomics) for single-cell transcriptome analysis.

Single-Cell Transcriptome Capture, Library Construction and Sequencing

Cells were loaded onto Chromium microfluidic chips and barcoded using the 10x Chromium Controller (10x Genomics). Single-cell transcriptomes were reverse-transcribed into cDNA libraries, incorporating 10x cell barcodes and unique molecular identifiers (UMIs). All steps were conducted with the Chromium Next GEM Single Cell 3′ Kit v3.1 (10x Genomics, PN-1000268) and the Chromium Next GEM Chip G Single Cell Kit (10x Genomics, PN-1000120). The resulting libraries were sequenced in paired-end 150 bp (PE150) mode on the Illumina NovaSeq platform.

Sequencing Data Processing

The raw reads from the cDNA library were aligned to the reference genome using Cell Ranger (v3.1.0) by 10x Genomics, with the Genome Reference Consortium Mouse Build 38 (GRCm38) serving as the reference genome for analysis.

Data integration, Dimensionality Reduction, Clustering and Visualization

Seurat v4.0.0 was utilized for clustering analysis and visualization. The gene expression matrix from each sample was imported and converted into Seurat objects. Cells with over 6% mitochondrial UMIs, fewer than 500 UMIs, or fewer than 200 genes were excluded from further analysis. Log-normalization based on total cellular UMI counts was applied, followed by canonical correlation analysis (CCA) to identify 2,000 anchor genes across samples, correcting for batch effects while preserving biological variation. Data were then scaled according to UMI counts, and PCA was conducted on the top 2,000 highly variable anchor genes. Clustering was performed at a resolution of 0.6, and Uniform Manifold Approximation and Projection (UMAP) was utilized for visualization.

Cell Type Annotation

SingleR (v1.4.1), a computational framework utilizing reference bulk transcriptomes, was used to unbiasedly annotate cell types for each cluster in the combined dataset. The Seurat function FindMarkers was then applied to identify specific markers for each cluster, cross-referencing these markers with the PanglaoDB cell type marker database (https://panglaodb.se/). Final adjustments to cell type annotations were informed by biological knowledge.

Differential Expressed Gene Analysis

We utilized the FindMarkers function in Seurat to identify specific markers across identities, applying the Wilcoxon test with a minimum expression threshold of min.pct > 0.25. Genes with a log2 fold change > 0.25 and p_val_adj < 0.05 were classified as upregulated, while those with a log2 fold change < −0.25 and p_val_adj < 0.05 were considered downregulated. Expression patterns of selected genes in each cluster were visualized using feature plots, violin plots, and heat maps.

Gene Ontology and KEGG Pathway Enrichment Analysis

GO and KEGG pathway enrichment analyses were conducted on differentially expressed genes (DEGs) utilizing the clusterProfiler R package with default parameters. GO enrichment analysis provided insights into the biological functions of DEGs, encompassing molecular functions, cellular components, and biological processes. KEGG enrichment analysis was utilized to identify the associated signaling pathways relevant to DEGs.

Scoring of Biological Processes

The AddModuleScore function in Seurat was utilized to score individual cells based on gene signatures representing specific biological processes. Gene sets for S-phase and G2/M-phase were sourced from Seurat’s embedded cell cycle list, while apoptosis (mcc04210), ferroptosis (mcc04216), necroptosis (mcc04217), and cellular senescence (mcc04218) gene sets were obtained from the KEGG database. Stemness and differentiation gene signatures were referenced from relevant literature (ref). Average comparison p-values between specified groups were calculated using the stat_compare_means function in ggpubr with default settings.

Pseudo-Time Trajectory Analysis

Pseudo-time trajectory analysis for selected cell types was carried out utilizing Monocle2 with the DDR-Tree reduction method and default settings. A Monocle object was created using expression matrices and metadata from the Seurat object, with feature selection based on top marker genes identified within Seurat clusters. Dimensionality reduction accounted for batch effects. Trajectory plots and heatmaps were generated to illustrate pseudo-time progression.

Cell Communication Analysis

CellphoneDB, a database of receptor-ligand interactions, was employed to infer cell-cell communication among various cell types. Cell communication analysis was performed with default parameters, with a minimum expression threshold of 10% for any ligand or receptor in a given cell type, and 1,000 permutations. A heatmap visualized the intensity of communication between cell types, while a dot plot displayed interaction pairs among different cell types.

Molecular Analysis of Gene Expression Using qRT-PCR

RNA was extracted from heart tissues using TRIzol (Takara). cDNA synthesis was performed using the PrimeScript RT Reagent Kit (RR047A, Takara). Quantitative real-time PCR analysis was conducted via the Applied Biosystems 7500 Real-Time PCR System. Gene-specific primers were obtained from Thermo Fisher, and the sequences can be found in Table 1. The mRNA expression levels of the target genes were normalized to those of the housekeeping gene Actb.

Table 1.

RT-PCR primers used in this study

	Forward	Reverse
TNF-α	5′-ACGGCATGGATCTCAAAGAC-3′	5′-AGATAGCAAATCGGCTGACG-3′
IL-6	5′-ACAACGATGATGCACTTGCAGA-3′	5′-GATGAATTGGATGGTCTTGGTC-3′
IL-1β	5′-CCAGCTTCAAATCTCACAGCAG-3′	5′-CTTTGGGTATTGCTTGGGATC-3′
Jaml	5′-ACTGATGAGGGGAGGGAGAC-3′	5′-CCAGATCCAAACCAGTCCCC-3′
Fcna	5′-TAAGGTCGTAGGTCTGGGGG-3′	5′-GTTGAAAAACGGTCCAGCCC-3′
IFN-γ	5′-AGGATCTGATGCCCCCTTCT-3′	5′-ACTCCCAGGGACTATGCCAT-3′
GAPDH	5′-AATGGATTTGGACGCATTGGT-3′	5′-TTTGCACTGGTACGTGTTGAT-3′

Results

IKKα Deficiency Aggravated DOX-Induced Cardiac Dysfunction at an Early Stage

Our previous study proved that IKKα played a key role in the DOX-induced cardiomyopathy at the late stage. To investigate whether IKKα influences DOX-induced cardiomyopathy in its early stages, we monitored changes in cardiac function in mice during the first week. Results indicated a marked decline in heart function in IKKα^Lyz2−Cre mice versus IKKα^flox/flox mice within the first week (Fig. 1A and B; Table 2). Meanwhile, IKKα^Lyz2−Cre mice showed increased susceptibility to doxorubicin-induced cardiac injury, with more pronounced cardiac fibrosis (Fig. 1C and Figure S1 A). Early secretion of inflammatory cytokines was also elevated in IKKα^Lyz2−Cre mice versus IKKα^flox/flox mice, indicating a significant increase in cardiac inflammatory factors (Fig. 1D and and Figure S1 B).

Fig. 1.

Macrophage-specific IKKα deficiency exacerbates doxorubicin (Dox)-induced cardiac injury in mice in the early stage. A. Representative echocardiography images of murine hearts and ejection fraction (EF) and fractional shortening (FS) determined by echocardiography 4 weeks after Dox inoculation. n = 5 per group. B. Representative morphological changes of hematoxylin and eosin (H&E) and Masson staining of myocardial Sect. 1 week after Dox inoculation. C. Serum levels of inflammatory cytokines in IKKα^flox/flox and IKKα^Lyz2−Cre mice 1 week after Dox inoculation. n = 6 per group

Table 2.

Echocardiography measurements for cardiac functions of DOX-induced cardiomyopathy mice

Genotypes	EF(%)	FS(%)	LVID; d(mm)	LVID; s(mm)	LV Vol; d(ul)	LV Vol; s(ul)
IKKαflox/flox + Saline	64.29 ± 2.03	34.53 ± 1.41	3.71 ± 0.10	2.51 ± 0.12	58.95 ± 3.87	24.01 ± 2.36
IKKαLyz−2−cre + Saline	61.91 ± 1.03	32.57 ± 0.95	3.75 ± 0.10	2.53 ± 0.09	62.76 ± 3.97	25.37 ± 2.79
IKKαflox/flox + DOX	50.22 ± 1.44 #, p < 0.0001	25.27 ± 0.67 #, p < 0.0001	4.19 ± 0.13 #, p < 0.0001	3.05 ± 0.13 #, p < 0.0001	69.39 ± 2.95 #, p = 0.0004	35.84 ± 2.33 #, p < 0.0001
IKKαLyz−2−cre + DOX	44.97 ± 1.36 †, p < 0.0001	20.96 ± 1.19 †, p < 0.0001	4.35 ± 0.15 †, p = 0.0842	3.33 ± 0.14 †, p = 0.0062	74.74 ± 4.48 †, p = 0.0347	46.70 ± 2.58 †, p < 0.0001

N = 6 each group (mean ± SEM)

#: IKKα^flox/flox + DOX vs. IKKα^flox/flox + Saline;

†: IKKα^{Lyz−2−cre} + DOX vs. IKKα^flox/flox + DOX.

Cell Composition Heterogeneity in DOX-induced Cardiomyopathy Tissue Between the IKKα^flox/flox and IKKα^Lyz2−Cre Mice

To elucidate the molecular mechanisms underlying the exacerbated tissue damage observed in early-stage DOX-induced cardiomyopathy following macrophage-specific IKKα knockout, we conducted single-cell transcriptome analysis on heart tissues from IKKα^flox/flox and IKKα^Lyz2−Cre mice under DOX stimulation. Following stringent quality control and filtering, we obtained 20,052 high-quality single-cell datasets (Fig. 2A). Using automated annotation and manual validation based on classical marker genes, we identified 11 cell types (Fig. 2B), including T cells, B cells, NK cells, monocytes, macrophages, neutrophils, vascular endothelial cells, cardiomyocytes, fibroblasts, smooth muscle cells, and Schwann cells (Figure S2A and B).

Fig. 2.

Comprehensive single-cell sequencing atlas of mouse heart tissue in a DOX-induced cardiomyopathy model. A. Schematic of the experimental workflow for this study. B. Uniform Manifold Approximation and Projection (UMAP) clustering plot of all cells from IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, colored by cell type. C. Heatmap of cell-cell communication frequency between 11 major cell types in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, analyzed using CellphoneDB. D. UMAP clustering plot of immune cells from IKKα^flox/flox and IKKα^Lyz2−Cre hearts, with cells colored by their immune cell types. E. FeaturePlot indicating marker genes for various immune cell types in the UMAP plot. F. DotPlot of marker gene expression for immune cell types in the UMAP plot. G. Bar plot comparing immune cell type proportions in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue. H. Density plots of spatial distribution for different immune cell types in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue

Analysis of intercellular communication networks between IKKα^flox/flox and IKKα^Lyz2−Cre mice revealed significant differences in intercellular communication frequency, particularly among immune cells such as monocytes, macrophages, and T cells Fig. 2C). Subsequently, we isolated all immune cells, applied additional quality filtering, and retained 4,547 high-quality immune cells for refined clustering. The clustering analysis identified eight immune cell types (Fig. 2D), including T cells expressing Cd3d and Trac, B cells expressing Cd19 and Ms4a1, NK cells expressing Ncr1 and Gzma, monocytes expressing Ccr2, macrophages expressing Cd14 and Fcgr3, dendritic cells expressing Cd209a and Ccl22, neutrophils expressing S100a9 and Il1b, and proliferative cells expressing Top2a and Mki67 (Fig. 2E and F). Bar plots and cell density analyses of cell composition between the IKKα^flox/flox and IKKα^Lyz2−Cre groups revealed a higher proportion of T cells in the IKKα^flox/flox group, while the IKKα^Lyz2−Cre group exhibited an increased proportion of monocytes and macrophages (Fig. 2G and H).

IKKα Knockout Shift Macrophage Polarization from Fcna⁺ M2 to Jaml⁺ M1

Building on the cell-cell communication analysis, which identified monocytes and macrophages as central components in the cardiac tissue microenvironment’s communication network, these cells likely play a critical role in the inflammatory response associated with DOX-induced cardiomyopathy. Therefore, we isolated monocytes and macrophages for detailed clustering analysis. Through cluster-specific marker genes and manual annotation, we identified nine distinct subtypes (Fig. 3A): monocyte subtype Mono_Ccr2, Mono_Ace and the macrophage subtype Mac_Mki67, Mac_Fcna, Mac_Jaml, Mac_Mmp14, Mac_Cd207, Mac_Flt1, Mac_Ifit3 (Fig. 3B and C). We then analyzed the differences in cell composition between the IKKα^flox/flox and IKKα^Lyz2−Cre groups, visualizing these distinctions through bar and density plots. The analysis showed significant enrichment of Mono_Ccr2, Mono_Ace, and Mac_Fcna in the IKKα^flox/flox group, whereas Mac_Jaml and Mac_Flt1 were notably enriched in the IKKα^Lyz2−Cre group (Fig. 3D and E).

Fig. 3.

Transcription characteristics and heterogeneity among the nine Monocyte/Macrophage subclusters. A. UMAP sub-clustering plot of Monocytes and Macrophages from IKKα^flox/flox and IKKα^Lyz2−Cre hearts, colored by subtype. B. FeaturePlot indicating marker genes for nine monocyte and macrophage subtypes in the UMAP plot. C. DotPlot of marker gene expression for nine monocyte and macrophage subtypes in the UMAP plot. D. Bar plot comparing the proportions of nine Monocyte and Macrophage subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue. E. Density plots of spatial distribution for nine monocyte and macrophage subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue. F. Violin plot showing the differences in NF-κB signaling pathway activity across nine Monocyte and Macrophage subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre mouse heart tissue, based on gene set scoring. G. Pseudotime trajectory plot of nine Monocyte and Macrophage subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, generated with Monocle. H. Distribution of IKKα^flox/flox and IKKα^Lyz2−Cre-derived cells within the pseudotime trajectory plot, as analyzed by Monocle. I. Heatmap of temporal changes of marker genes across nine Monocyte and Macrophage subtypes along the pseudotime trajectory. J. Heatmap showing the M1 and M2 polarization tendencies of nine Monocyte and Macrophage subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, based on gene set scoring

Previous studies indicate that IKKα can suppress NF-κB pathway activation in macrophages. Our gene set scoring analysis confirmed that the NF-κB pathway was significantly upregulated in the Mac_Fcna, Mac_Jaml, and Mac_Mmp14 macrophage subtypes following IKKα knockout (Fig. 3F). Given that Mac_Fcna and Mac_Jaml are predominantly enriched in the IKKα^flox/flox and IKKα^Lyz2−Cre groups, respectively, we employed Monocle to perform pseudotime trajectory inference on all monocyte and macrophage subtypes to map the differentiation landscape of these cells in DOX-induced cardiomyopathy cardiac tissue. The pseudotime analysis revealed that Mono_Ccr2 and Mono_Ace occupy the initial stages of the trajectory, which then progresses through Mac_Mki67 and Mac_Mmp14, eventually diverging into two distinct branches. At this branching point, the IKKα^flox/flox group-specific Mac_Fcna and the IKKα^Lyz2−Cre group-specific Mac_Jaml occupy separate branches. In summary, monocytes in the IKKα^flox/flox group primarily differentiate into the tissue-resident macrophage subtype Mac_Fcna, which expresses genes Fcna, Lyve1, and Folr2. Conversely, IKKα knockout directs monocytes toward the inflammatory Mac_Jaml subtype, marked by Jaml expression (Fig. 3G and H). This shift in differentiation induced by IKKα loss promotes an inflammatory macrophage phenotype over the tissue-repair phenotype (Fig. 3I). To validate the phenotypes of these macrophage subtypes, we referenced datasets from two studies for M1/M2 classification. Both analyses consistently indicated that Mac_Fcna is inclined toward the M2 phenotype, while Mac_Jaml aligns with the M1 phenotype (Fig. 3J and Figure S3A).

T Cell Activation and Differentiation Were Impaired in IKKα^Lyz2−Cre Mice

As another critical component of the cellular communication network within the cardiac tissue microenvironment, we proceeded to isolate T cells for further sub-clustering. This analysis identified seven distinct T cell subtypes (Fig. 4A), including the T_Ccr7 subtype expressing Ccr7 and Sell, the T_Gzma subtype expressing Gzma and Cx3cr1, the proliferative T_Mki67 subtype expressing Mki67 and Top2a, the helper T_Cd4 subtype expressing Cd4, the T_Il17 subtype expressing Il17a and Il23r, and the T_Calca subtype expressing Calca and Il1rl1 (Fig. 4B and C). Bar plots and cell density maps revealed notable differences in cell composition between the groups, with effector T cell subtypes T_Gzma and T_Cd4 significantly enriched in the WT group, whereas the naive T cell subtype T_Ccr7 was predominantly enriched in the IKKα^Lyz2−Cre group (Fig. 4D and E). Pseudotime trajectory inference showed that naive T cells (T_Ccr7) form the starting point of development, diverging into two primary differentiation pathways: one leading to effector T cells (T_Gzma and T_Cd4) and the other to proliferative T cells (T_Mki67) (Fig. 4F). Comparative analysis indicated that in IKKα conditional knockout tissues, T cells remained in a naive state, lacking the activation and differentiation processes observed in IKKα^flox/flox tissues (Fig. 4G and H).

Fig. 4.

Transcription characteristics and heterogeneity among the seven T cell subclusters. A. UMAP sub-clustering plot of T cells from IKKα^flox/flox and IKKα^Lyz2−Cre hearts, colored by subtype. B. FeaturePlot indicating marker genes for seven T cell subtypes in the UMAP plot. C. DotPlot of marker gene expression for seven T cell subtypes in the UMAP plot. D. Bar plot comparing the proportions of seven T cells subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue. E. Density plots of spatial distribution for seven T cell subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue. F. Pseudotime trajectory plot of seven T cells subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, generated with Monocle. G. Distribution of IKKα^flox/flox and IKKα^Lyz2−Cre-derived cells within the pseudotime trajectory plot, as analyzed by Monocle. H. Heatmap of temporal changes of marker genes across seven T cells subtypes along the pseudotime trajectory

Jaml + Macrophage is Resposible for T Cell Activation

To further investigate cell communication between the nine monocyte-macrophage subtypes and seven T cell subtypes in DOX-induced cardiomyopathy heart tissue, we used CellphoneDB and CellChat software to analyze intercellular communication frequency and specific ligand-receptor interactions. Heatmaps of communication frequency showed a notable reduction in interactions between monocyte-macrophage subtypes and T cell subtypes, as well as within monocyte-macrophage subtypes themselves, following IKKα knockout (Fig. 5A). Scatter plots of communication frequency revealed a marked decrease in both signals received by T cell subtypes and signals emitted by monocyte-macrophage subtypes in IKKα knockout tissues (Fig. 5B).

Fig. 5.

Cell-cell communications among 9 monocyte/macrophage subtypes and 7 T cell subtypes. A. Heatmap of cell-cell communication frequency among nine monocyte/macrophage subtypes and seven T cell subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, analyzed with CellphoneDB. B. Scatter plot generated by CellChat, illustrating cell-cell communication frequency between monocyte/macrophage and T cell subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissues. In the plot, the X-axis represents the signaling strength emitted by the cells, while the Y-axis indicates the signal reception strength. C. The bar plot illustrates the differential communication signals between IKKα^flox/flox and IKKα^Lyz2−Cre heart tissues. D. Heatmap displays the cell-cell communication intensity and roles of the CXCL, ICOS and MHC-I signaling pathways in IKKα^flox/flox heart tissues. E. Heatmap displays the cell-cell communication intensity and roles of the CXCL, ICOS and MHC-I signaling pathways in IKKα^Lyz2−Cre heart tissues. F. Heatmap displays the cell-cell communication intensity and roles of the IFN-II signaling pathways in IKKα^flox/flox heart tissues. G. Heatmap displays the cell-cell communication intensity and roles of the IFN-II and IL-6 signaling pathways in IKKα^Lyz2−Cre heart tissues. H. Violin plot showing the differences in chemokine signaling pathway/T cell receptor signaling pathway activity across seven T cells subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre heart tissue, based on gene set scoring. I. Violin plot showing the differences in IFN-γ signaling pathway/IL-6 signaling pathway activity across nine Monocyte and Macrophage subtypes in IKKα^flox/flox and IKKα^Lyz2−Cre mouse heart tissue, based on gene set scoring

Differential cell communication analysis indicated that most communication intensities were stronger in the IKKα^flox/flox group than in the IKKα^Lyz2−Cre group (Fig. 5C). Specifically, signals related to T cell chemotaxis and migration via the CXCL pathway, as well as T cell activation through ICOS and MHC-I pathways, were significantly elevated in the IKKα^flox/flox group (Fig. 5D and E). In contrast, communication signals linked to macrophage activation and M1 polarization, such as IL-6 and IFN-γ pathways, were considerably lower in the IKKα^flox/flox group compared to the IKKα^Lyz2−Cre group (Fig. 5F and G).

These differences in communication patterns aligned with functional gene set analysis of T cell subtypes, which showed that chemotaxis and T cell receptor signaling pathway activity was significantly higher in the IKKα^flox/flox group (Fig. 5H). Meanwhile, functional gene set analysis of macrophage subtypes demonstrated that IFN-γ and IL-6 signaling pathways were significantly more active in macrophages from the IKKα^Lyz2−Cre group (Fig. 5I and and Figure S3A).

Discussion

Although effective against various cancers, the clinical utility of DOX is limited by its cardiotoxic effects, highlighting the need to elucidate the underlying mechanisms. In this study, we aimed to characterize macrophage subpopulations and explore macrophages-T cells crosstalk in hearts of IKKα^Lyz2−Cre mice with DOX-induced cardiomyopathy by single-cell sequencing. Our results showed that IKKα knockout shifted macrophage polarization from Fcna⁺ M2-like towards Jaml⁺ M1-like phenotypes. Additionally, IKKα was identified as a key regulator in mediating macrophage-T cell interactions. Specifically, IKKα deficiency in macrophages activated T cells through Jaml⁺ M1 macrophages, which in turn enhanced M1 polarization through IFN-γ and IL-6 signaling pathways.

Single-cell transcriptomic analysis of DOX-induced cardiomyopathy tissues revealed distinct heterogeneity in immune cell composition between IKKα^flox/flox and IKKα^Lyz2−Cre mice. The IKKα^flox/flox group exhibited a higher proportion of T cells, whereas IKKα^Lyz2−Cre mice showed increased abundance of monocytes and macrophages. These findings align with prior observations that shifts in immune cell subsets, such as variations in T cell and monocyte distribution, can significantly influence organ-specific immune responses and disease progression [18]. Specifically, the elevated levels of monocytes and macrophages in IKKα^Lyz2−Cre mice likely exacerbate pro-inflammatory conditions, as higher monocyte abundance is associated with enhanced immune activation and inflammatory signaling pathways [19]. Moreover, the altered T cell dynamics in this model resemble observations in fibrotic disorders such as idiopathic pulmonary fibrosis, where increased CD8 + T cell fractions correlate with disease progression and dysregulated immune homeostasis [20]. These consistencies underscore the role of immune cell heterogeneity in disease pathogenesis, potentially mediated through disrupted intercellular communication networks, as further supported by significant differences in immune crosstalk identified in our analysis.

Heart tissue in pressure-induced heart failure, observed in both patients and animal models, exhibits significant infiltration and activation of inflammatory cells [21–23]. Immune cells have been shown to play a critical role in DCM as well [24]. Among cardiac immune cells, macrophages represent the earliest and most abundant responders in heart failure, with their abundance and phenotypic states being crucial in the response to cardiac injury or disease [25]. Macrophages have traditionally been classified into M1 and M2 subgroups based on functional and phenotypic characteristics. However, emerging research has highlighted the extensive heterogeneity and plasticity of macrophages in both homeostatic and injury states within the heart, expanding the traditional understanding of macrophage polarization. Signal transducers and activators of transcription (STATs), particularly STAT3, serve as central regulators of macrophage polarization [26], integrating multiple signaling pathways to influence phenotypic fate [27]. Previous study demonstrated that IKKα knockout promoted STAT3-S727 phosphorylation, further impairing cardiac function [28]. In the current study, we IKKα as a key regulator of macrophage polarization in DOX-induced cardiomyopathy. Fcna⁺ macrophage and Jaml⁺ macrophage were enriched in the hearts of IKKα^flox/flox and IKKα^Lyz2−Cre mice, respectively. IKKα knockout shifted macrophage polarization from Fcna⁺ to Jaml⁺, correlating with worsened cardiac function. Additionally, the NF-κB pathway was significantly upregulated in both Mac_Fcna and Mac_Jaml macrophage subtypes following IKKα knockout, driving macrophages toward an M1-like pro-inflammatory phenotype. These insights highlight the critical role of IKKα in regulating macrophage phenotypes and imply that its knockout exacerbates pro-inflammatory states in cardiac tissue, potentially influencing disease progression in cardiomyopathy contexts.

The observed impairment in T cell activation and differentiation in IKKαLyz2-Cre mice, characterized by the enrichment of naïve T cells (T_Ccr7) and reduction in effector T cell subtypes (T_Gzma and T_Cd4), aligns with established roles of IKKα in T cell biology. Specifically, the deficiency in IKKα likely disrupts the differentiation of naïve T cells into effector populations, as supported by studies showing that IKK1 (IKKα) promotes the differentiation of effector T cells, such as Th1 and Th17 subsets, in contexts like ischemia-reperfusion injury [29]. Furthermore, signaling through the IKK complex, which includes IKKα, is essential for the survival and maturation of T cells, and the persistence of naïve T cells in IKKαLyz2-Cre tissues may reflect a failure in NF-κB activation, a key pathway regulated by IKK that supports T cell survival and differentiation [30]. The enrichment of naïve T cells might also be exacerbated by impaired apoptosis repression, as IKK promotes naïve T cell survival by inhibiting RIPK1-dependent pathways [30]. Overall, these findings highlight the critical role of IKKα in facilitating T cell effector differentiation and activation, potentially through NF-κB-dependent mechanisms, and underscore the importance of IKK signaling in maintaining immune homeostasis within tissue microenvironments.

Intercellular crosstalk is essential for maintaining organ homeostasis and can drive dysfunction by modulating cell survival, proliferation, and death [31, 32]. Cell communication typically occurs through paracrine signaling, such as ligand-receptor interactions, or through direct physical connections, like gap junctions [33]. The primary cell types involved in cardiac inflammatory diseases are M1 and M2 macrophages, along with effector T (Teff) and regulatory T (Treg) cells [34]. Under different cellular metabolic conditions and intercellular interactions, macrophages and T cells can exhibit either pro-inflammatory or anti-inflammatory abilitie. For instance, inflammation in atherosclerosis is triggered by macrophage-mediated innate immune responses and sustained by Th1, Th17, and Th2 cells [35]. Conversely, a recent study demonstrated that therapeutic activation of Treg cells following myocardial infarction improved healing and survival by promoting macrophage differentiation toward the M2 phenotype [36]. Our study identified a novel role of IKKα in modulating T cell activation through IFN-γ and IL-6 secretion by Jaml⁺ macrophage in DOX-induced cardiomyopathy. The interaction between macrophages and T cells creates a fluid balance between anti-inflammatory and pro-inflammatory states in DOX-induced cardiomyopathy. These findings suggest that targeting macrophage-T cell crosstalk may represent a promising therapeutic approach for DOX-induced cardiomyopathy.

In conclusion, this study elucidates the pivotal role of IKKα in modulating macrophage polarization and T cell activation within the context of DOX-induced cardiomyopathy, employing single-cell transcriptomics to unveil intricate cellular interactions and heterogeneity. The findings highlight a significant shift in macrophage polarization from the reparative M2 phenotype (Mac_Fcna) to the inflammatory M1 phenotype (Mac_Jaml) following IKKα knockout, alongside impaired T cell activation and disrupted intercellular communication networks. These insights not only enhance our understanding of the cardiac immune microenvironment but also underscore the potential for targeted immunotherapeutic strategies aimed at modulating macrophage and T cell responses. However, the limitations of this study, including the small sample size and the absence of clinical validation, necessitate further investigation to translate these findings into effective therapeutic interventions. Future research integrating multi-omics approaches and clinical data will be essential to advance personalized medicine in the management of cardiac diseases, ultimately improving patient outcomes and quality of life.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

YFZ, SCL and ZHT designed and supervised the study. GYC and YWY collected and analyzed the data, and wrote the manuscript; YFJ and HYQ collected and analyzed the data; YFJ, CW, YML and JL interpreted the results. XC, LY and YDC edited and revised the manuscript. All the authors contributed to the final version of the paper.

Funding

This study was supported by grants from Nanjing Health Technology Development Project (YKK22125); the young Program of National Natural Science Foundation of China (No. 82200529 and No. 82200286).

Data Availability

All data are available from the corresponding author upon reasonable request.

Declarations

Ethics Approval and Consent to Participate

Not applicable.

Consent for Publication

Not applicable.

Competing Interests

The authors declare no competing interests.