The presence of 11 immune cell types in SLE environment

Autoimmune disorder SLE displays involvement across various bodily systems and the extensive generation of autoantibodies (Ameer et al. 2022). The emphasis in recent investigations has been on the contribution of Tfh cells to the etiology of SLE (Liu et al. 2022a, b). Occupying the follicular regions of lymph nodes and spleen, Tfh cells, a distinctive faction within CD4+ T cells, are essential for the stimulation, evolution, and antibody synthesis by B cells (Anang et al. 2022; Xin et al. 2023).

Aberrant activation of Tfh cells in SLE might trigger excessive B cell activation and production of autoantibodies, further worsening the disease (Wu et al. 2016). Additionally, SLE's pathogenic advancement is notably influenced by the communication linking Tfh cells with various immune entities, affecting the production and release of inflammatory mediators (Nakayamada and Tanaka 2021). Therefore, a thorough investigation of the role and mechanisms of Tfh cells in SLE is crucial for understanding the disease pathogenesis and providing new therapeutic strategies.

For a better grasp on the function and mechanisms of Tfh cells in SLE, we commenced by opting for a solitary single-cell transcriptomic dataset (GSE135779) linked to SLE from the GEO database. This dataset comprised peripheral blood mononuclear cell (PBMC) samples from 12 adult individuals, including 5 healthy controls and 7 SLE patients.

Employing the Seurat software package, we amalgamated the information and initially assessed the gene numbers (nFeature_RNA), mRNA molecule counts (nCount_RNA), and percentage of mitochondrial genes (percent. mt) in all cells of the scRNA-seq data. The results showed that a majority of cells exhibited nFeature_RNA < 6000, nCount_RNA < 40000, and percent.mt < 25% (Figure S1A).

Cells of inadequate quality were eliminated by adhering to the specifications of nFeature_RNA > 300, 500 < nCount_RNA < 30000, and percent.mt < 20%. Upon filtration, the sequencing depth correlation examination presented a correlation coefficient of r = 0.17 for nCount_RNA and percent.mt, as well as a correlation coefficient of r = 0.84 for nCount_RNA and nFeature_RNA (Figure S1B), suggesting the filtered cellular data's satisfactory quality suitable for subsequent analytical purposes.

To delve deeper into the examination of the filtered cell samples, we standardized the experimental data employing the "LogNormalize" function and opted for the top 1000 genes with significant variability for subsequent scrutiny (Figure S1C). Later on, PCA was applied to lower the dimensions of the 1000 genes, enabling the TSNE clustering of cellular entities.

After conducting PCA, a total of 40 PCs were acquired, among which the top 20 PCs were chosen for the ensuing tasks of TSNE clustering and annotating the cells (Figure S1D). Figure S1E displays the PCA outcomes regarding cells in individual samples. Figure S1F showcases the predominant genetic elements linked to the first set of PCs, whereas Figure S1G presents the gene expression heatmap for the leading four PCs.

The outcomes of the PCA analysis point to the dependability of the results and their applicability for the clustering of cells.

Subsequent to the TSNE cluster analysis, all cells were sorted into 15 cellular clusters (Fig. 1A), with the arrangement of each cluster depicted in Fig. 1B. We then annotated these cells by utilizing established lineage-specific marker genes gathered from academic sources and the internet-based tool CellMarker, leading to the detection of 11 different cellular categories (Fig. 1C).

Fig. 1

TSNE clustering and cell annotation of 12 PBMC samples after quality control. Note: (A) PBMC samples were isolated from 5 healthy controls and 7 SLE patients after rigorous quality control. The cells were clustered into 15 clusters using the TSNE clustering method. B Distribution of cell clusters in the normal control (NC) and SLE groups. C The cell clusters were annotated using known cell-specific marker genes obtained from literature and online resources like CellMarker, identifying a total of 11 cell types. D Heatmap of the marker genes for these 11 cell types. E Bubble plot showing the expression levels of these 11 cell marker genes. The abbreviations: NC: normal control group; SLE: systemic lupus erythematosus; nCD4 + T: naive CD4+ T cells; aCD4+ T: activated CD4+ T cells; CD1C+ B DC: CD1C+ B dendritic cells; Tem: effector CD8 + memory T cells; CD1C-CD141- DC: CD1C-CD141- dendritic cells; Treg: regulatory T cells, induced regulatory T cells; RBC: red blood cells; MPC: megakaryocyte progenitor cells; pDC: plasmacytoid dendritic cells; AXL+SIGLEC6+ DC: AXL+SIGLEC6+ dendritic cells

Cluster 0 was recognized as naïve CD4 + T cells (nCD4 + T) with marker genes CCR7 and NOSIP (Fig. 1E; Figure S2A-B). The investigation into GO enrichment unveiled the intimate correlation between the marker genes of this group and the stimulation and differentiation of T cells (Figure S3A).

Cluster 1 was denoted as activated CD4 + T cells (aCD4 + T) with the marker gene CD69 (Fig. 1D-E; Figure S2C). GO enrichment analysis showed that the marker genes of this cluster are associated with lymphocyte differentiation and T-cell activation (Figure S3B).

Clusters 2 and 9 were identified as CD1C + B dendritic cells (CD1C + B DC) with marker genes S100A9 and LYZ (Fig. 1D-E; Figure S2D-E). GO enrichment analysis revealed that the marker genes of this cluster are involved in antigen processing and presentation (Figure S3C).

Clusters 3 and 6 were annotated as CD8 + effector memory T cells (Tem) with marker genes GZMH and NKG7 (Fig. 1D-E; Figure S2F-G). GO enrichment analysis showed that the marker genes of this cluster are associated with T cell activation and regulation (Figure S3D).

Cluster 4 was designated as B cells with marker genes CD79A and MS4A1 (Fig. 1D-E; Figure S2H-I). GO enrichment analysis revealed that this cluster's marker genes are involved in B cell-mediated immune processes (Figure S3E).

Clusters 5 and 8 were identified as CD1C-CD141- dendritic cells (CD1C-CD141- DC) with marker genes IFITM3 and TYMP (Fig. 1D-E; Figure S2J-K). GO enrichment analysis showed that the marker genes of this cluster are associated with the MHC II protein complex and related functions (Figure S3F).

Cluster 7 was annotated as regulatory T cells (Treg) with the marker gene IL7R (Fig. 1D-E; Figure S2L). GO enrichment analysis revealed that the marker genes of this cluster are implicated in T cell differentiation and the Wnt signaling pathway (Figure S3G).

Cluster 10 was annotated as red blood cells (RBC) with marker genes HBA1 and HBB (Fig. 1D-E; Figure S2M-N). GO enrichment analysis showed that the marker genes of this cluster are associated with oxygen transport and hydrogen peroxide metabolism (Figure S3H).

Cluster 11 was characterized as megakaryocyte progenitor cells (MPC) with marker genes PPBP and PF4 (Fig. 1D-E; Figure S2O-P). GO enrichment analysis revealed that the marker genes of this cluster are involved in platelet activation and aggregation (Figure S3I).

Clusters 12 and 13 were annotated as plasmacytoid dendritic cells (pDC) with marker genes ITM2C and LILRA4 (Fig. 1D-E; Figure S2Q-R). GO enrichment analysis showed that the marker genes of this cluster are associated with antigen processing and presentation (Figure S3J).

Cluster 14 was annotated as AXL + SIGLEC6 + dendritic cells (AXL + SIGLEC6 + DC) with the marker gene SOX4 (Fig. 1D-E; Figure S2S). GO enrichment analysis revealed that the marker genes of this cluster are associated with T cell differentiation and regulation of immune response (Figure S3K).

In summary, we successfully annotated 15 cellular clusters and identified 11 distinct cell types, including 3 CD4 + T cell subtypes (nCD4 + T, aCD4 + T, and Treg), 4 dendritic cell subsets (CD1C + B DC, CD1C-CD141- DC, pDC, and AXL + SIGLEC6 + DC), 1 red blood cell (RBC), 1 megakaryocyte progenitor cell (MPC), 1 B cell (B cell), and 1 CD8 + T cell (Tem). Among these, we particularly focused on the two CD4 + T cell subtypes, namely, nCD4 + T cells and aCD4 + T cells. nCD4 + T cells serve as precursors of Tfh cells, while aCD4 + T cells can differentiate into Tfh cells under specific conditions.

Signal regulation of dendritic cells in SLE and their impact on T cell function

To further elucidate the role and mechanisms of Tfh cells in SLE, we decided to perform a cellular communication analysis. The analysis provides insights into the interplay among nCD4 + T cells, aCD4 + T cells, and various cell types, thereby driving their differentiation into Tfh cells. Therefore, we conducted cellular communication analysis on these eleven cell types. The display in Figure S4A demonstrates the aggregate quantity and degree of interplays among various cellular groups. Figure S4B-L illustrates the quantity and strength of connections among each cell type and other cells.

By examining the cellular interaction networks of each cell type with other cells, we identified four dendritic cell subtypes, namely CD1C + B DC (Figure S4D), CD1C-CD141- DC (Figure S4G), pDC (Figure S4K), and AXL + SIGLEC6 + DC (Figure S4L), primarily acting as signal senders, with nCD4 + T cells and aCD4 + T cells as their targets. Dendritic cells are highlighted as central regulators in the process of immune communication. Their impact on SLE development could be due to their influence on T-cell activation and functionality.

Next, the examination conducted centered on the signaling pathways and their associated ligand-receptor pairs essential for cellular communication networking. Through analysis, we identified eight signaling pathways that are pivotal in this network, including MIF, GALECTIN, ANNEXIN, RESISTIN, BAFF, IL16, CXCL, and MK (Figure S5A-B). Furthermore, Figure S5C provides a detailed overview of the ligand-receptor pairs involved in each signaling pathway.

Within the studied signaling pathways, nCD4 + T cells and aCD4 + T cells primarily act as receptors for the MIF and GALECTIN signaling pathways ((Figure S6). On the signaling sender side, plasmacytoid dendritic cells (pDC) and AXL + SIGLEC6 + dendritic cells are key senders of the MIF signaling pathway (Figure S6A); meanwhile, CD1C-CD141- dendritic cells and AXL + SIGLEC6 + dendritic cells serve as leading senders in the GALECTIN signaling pathway (Figure S6B).

In summary, our study identified four dendritic cell subtypes that are pivotal in the MIF and GALECTIN signaling pathways, particularly in activating nCD4 + T cells and aCD4 + T cells. This intercellular communication may be a crucial mechanism driving their differentiation into Tfh cells.

Subclustering and pseudotime analysis of aCD4 + T cells in SLE

Within our investigation, we have elucidated the central regulatory role of dendritic cells in the MIF and GALECTIN signaling pathways on both nCD4 + T cells and aCD4 + T cells. This intercellular communication may be a crucial mechanism driving their differentiation into Tfh cells. To further explore the diversity and specific functionality of aCD4 + T cells in SLE, particularly their relationship with Tfh cells, we conducted a sub-clustering analysis and pseudotime analysis of the aCD4 + T cells.

We initially filtered the dataset to identify the aCD4 + T cells and subsequently subjected them to preprocessing and gene expression-based clustering analysis, resulting in the identification of nine distinct subclusters (Fig. 2A). Utilizing existing literature and the online resource CellMarker, we annotated these cells, ultimately yielding nine cell types (Fig. 2B).

Fig. 2

Subclustering and pseudotime analysis of activated CD4 + T cells. Note: (A) TSNE clustering analysis divided all activated CD4 + T cells into 9 subclusters. B Annotation of the 9 cell subtypes. C Bubble plot showing the expression levels of marker genes for each sub-group. D Pseudotime analysis results of activated CD4 + T cells (Cluster 3, 5, and 8) and Tfh cells (Cluster 7)

Cluster 0 was designated as nCD4 + T cells type 1, distinguished by the marker gene CCR7 (Fig. 2C; Figure S7A). Gene enrichment analysis revealed significant activity in processes such as lymphocyte immune activation and growth factor response within this cell subcluster (Figure S8A).

Cluster 1 stood out as nCD4 + T cells type 2, identified by the marker gene EEF1B21 (Fig. 2C; Figure S7B). Gene enrichment analysis indicated involvement in processes such as cellular response to cytokines and cytokine signaling in the immune system (Figure S8B).

Cluster 2 was identified as CD8 + T cells, marked by the gene CD8B (Fig. 2C; Figure S7C). Gene enrichment analysis revealed participation in cellular responses to cytokine stimulation and leukocyte activation (Figure S8C).

Cluster 3 was categorized as aCD4 + T cells type 1, recognized by the marker gene ITGB1 (Fig. 2C; Figure S7D). Gene enrichment analysis showed involvement in processes such as cellular antioxidant detoxification and low oxygen response (Figure S8D).

Cluster 4 was identified as T helper 17 cells, identified by the marker gene KLRB1 (Fig. 2C; Figure S7E). Gene enrichment analysis showed participation in processes such as cytokine signaling in the immune system and antigen processing and presentation (Figure S8E).

Cluster 5 was identified as aCD4 + T cells type 2, marked by the gene CD69 (Fig. 2C; Figure S7F). Gene enrichment analysis indicated involvement in processes such as antigen processing and presentation and regulation of cell activation (Figure S8F).

Cluster 6 was identified as CD1C-CD141- dendritic cells, characterized by the marker gene ISG15 (Fig. 2C; Figure S7G). Gene enrichment analysis showed participation in processes such as interferon-gamma signaling and antiviral mechanisms induced by IFN stimulation (Figure S8G).

Cluster 7 was identified as Tfh T cells, marked by the gene ICOS and CXCR5 (Fig. 2C; Figure S7H and FigureS7I). Gene enrichment analysis revealed involvement in the management of T cell activation and adaptive immune system processes (Figure S8H).

Cluster 8 was labeled as aCD4 + T cells type 3, marked by the marker gene CD14 (Fig. 2C; Figure S7J). Gene enrichment analysis showed participation in inflammatory responses and cell activation processes (Figure S8I).

In summary, we conducted a comprehensive sub-clustering analysis of aCD4 + T cells, resulting in the identification of nine distinct subclusters. These subclusters comprised two types of nCD4 + T cells, three types of aCD4 + T cells, CD8 + T cells, Th 17 cells, Tfh cells, and CD1C-CD141- dendritic cells.

For a more profound insight into the developmental pathways of aCD4 + T cells (Clusters 3, 5, and 8) and Tfh cells (Cluster 7) in SLE, we performed pseudo-time analysis. The results revealed that Clusters 3 and 8 primarily occupy the initial phase of the developmental trajectory, hinting that they represent earlier cell subsets. On the other hand, Cluster 5 is located in the middle of the trajectory, while Cluster 7 is concentrated at the terminal end, indicating that Tfh cells may be formed in a later stage (Fig. 2D).

These findings imply that within the framework of SLE, aCD4 + T cells go through an early activation stage initially, followed by gradual differentiation into Tfh cells with specific functions. This differentiation and developmental trajectory may be closely associated with the pathophysiological mechanisms of SLE.

Observation of ferroptosis inhibition in Tfh cells in SLE model mice

Examining the involvement of Tfh cells in the pathogenesis of SLE involved the use of MRL/lpr mice as an SLE model and MRL/Mpj mice as the healthy controls. The results demonstrated that, compared to MRL/Mpj mice, MRL/lpr mice exhibited typical SLE pathological features, including significantly increased serum anti-double-stranded DNA antibodies and slight elevation in 24-h proteinuria (P < 0.05), as well as markedly reduced complement levels (P < 0.05). Enlargement was observed in both the lymph nodes and spleens (Fig. 3A-E). Elevated levels of CD4 + ICOS + CXCR5 + Tfh cells were detected in the peripheral blood, spleen, and lymph nodes of MRL/lpr mice in relation to MRL/Mpj mice, as indicated by FCM analysis (P < 0.05) (Fig. 3F-H, Figure S9A).

Fig. 3

Changes in Tfh cell quantity in the peripheral blood of MRL/lpr mice. Note: (A) Left panel: Assessment of anti-dsDNA titers in mouse serum using ELISA; right panel: Evaluation of proteinuria levels in 8-week-old mice. B Measurement of complement 3 levels in serum. C Representative images and length quantification of mouse axillary lymph nodes. D-E Representative images and length quantification of mouse spleens. F Detection of Tfh cell counts in mouse blood using FCM. G Detection of Tfh cell counts in mouse spleens using FCM. H Detection of Tfh cell counts in mouse lymph nodes using FCM. I Detection of iron content in Tfh cells from each group using FerroOrange staining. J Spectrophotometric measurement of lipid peroxide levels in Tfh cells. K Determination of MDA content in each group. L Measurement of GSH/GSSG ratio in each group. M–N Detection of mRNA levels of ferroptosis inhibitors STAT3 and GPX4 in Tfh cells using PCR. O Observation of mitochondrial morphology in cells using transmission electron microscopy, with a scale bar of 1 μm. P Detection of MMP in cells from each group. *** indicates P < 0.001 between groups, and the cell experiments were repeated three times; in the animal experiments, each group included 6 mice

Tfh cell abnormalities have been extensively reported in various autoimmune diseases. In SLE, disease severity is directly connected to the extent of pathogenic antibody generation facilitated by the interaction of Tfh cells and B cells (Zhao et al. 2023; Yu et al. 2022; Lee et al. 2019). Recent studies suggest that inhibiting ferroptosis can enhance Tfh cell survival and function (Yao et al. 2021).

We further examined ferroptosis indicators in Tfh cells, such as iron (Fe2+) content, LPO, MDA, and the GSH/GSSG ratio. The FCM technique was utilized for the sorting of CD4 + ICOS + CXCR5 + Tfh cells from MRL/lpr and MRL/Mpj mice. The study outcomes indicated that Tfh cells in MRL/lpr mice exhibited diminished Fe2+, LPO, and MDA levels contrasted with MRL/Mpj mice, whereas the GSH/GSSG balance was substantially enhanced (P < 0.05), indicating that ferroptosis was inhibited in MRL/lpr Tfh cells (Fig. 3I-L).

The occurrence of lipid peroxidation sets apart ferroptosis as an iron-reliant cell death process (Bock and Tait 2020; Gao et al. 2019). GPX4 functions prominently in transforming LPO to harmless alcohols, alleviating peroxidation (Zhang et al. 2022; Huang et al. 2023a, b). Activation of STAT3 upregulates GPX4 levels, enhancing cell resistance to peroxidation and influencing ferroptosis (Zhang et al. 2022).

Our experiments indicated that MRL/lpr Tfh cells exhibited markedly elevated STAT3 and GPX4 expression levels (P < 0.05) (Fig. 3M-N). Mitochondrial damage, a hallmark of ferroptosis (Bock and Tait 2020; Gao et al. 2019), was observed, showing mitochondrial recovery in MRL/lpr Tfh cells and damage in MRL/Mpj Tfh cells (Fig. 3O). Analysis of MMP indicated hyperpolarization in MRL/lpr Tfh cells (P < 0.05) (Fig. 3P).

In conclusion, using MRL/lpr mice as an SLE model, we observed typical SLE features and a significant increase in Tfh cell counts. Our findings further indicated that ferroptosis was inhibited in these cells, supporting their potential involvement in SLE pathogenesis.

We isolated CD4+ICOS+CXCR5+ Tfh cells from the MRL/lpr and MRL/Mpj mouse models using FCM (Figure S9B). In this study, CD4 + ICOS + CXCR5 + Tfh cells were isolated using FCM from MRL/lpr and MRL/Mpj mice. The RNA-seq analysis was executed on RNA extracted from sorted Tfh cells. Post quality validation of data, all samples were merged, and batch effects were removed. The R programming language's "limma" package was employed for the differential expression analysis. A total of 617 DEGs were identified as differentially expressed, meeting the criteria of |logFC|> 1.5 and P < 0.05. Out of these, 330 genes showed upregulation, whereas 287 genes demonstrated downregulation (Fig. 4A). The expression of these DEGs was visualized through a heatmap in Fig. 4B, highlighting notable variances in gene expression among the distinct Tfh cell groups.

Fig. 4

Tfh cell RNA-Seq data analysis. Note: (A) The volcano plot presents the expression of DEGs based on RNA-Seq sequencing analysis of three Tfh cell samples (CD4 + ICOS + CXCR5 +) from MRL/lpr and MRL/Mpj mice. Blue dots represent downregulated genes, red dots represent upregulated genes, and black dots represent genes without significant differences. B The hierarchical clustering heatmap displays the top 20 DEGs. C The bar chart shows the results of GO functional analysis and KEGG pathway enrichment analysis of DEGs, including biological processes (BP), cellular components (CC), and molecular functions (MF), as well as the associated network diagram. D The Venn diagram illustrates the intersection of DEGs and the marker genes of Tfh cells (Cluster 7). E The box plot represents the expression of HMGA1 in RNA-Seq data. F Western blot analysis was conducted to examine the protein level of HMGA1 in Tfh cells, with each experiment involving 6 mice. Cell experiments were repeated three times, with three replicates each time. *** indicates P < 0.001 compared to MRL/Mpj mice

In-depth scrutiny of the potential functions of the DEGs involved conducting a thorough analysis on GO enrichment. The examination uncovered that these genetic elements were predominantly linked to functions such as fatty acid metabolism, T cell differentiation, type 1 immune response, immune complex, cation channel complex, AMPA-type glutamate receptor complex, immunoglobulin receptor binding, glutathione peroxidase activity, and peroxidase activity (Fig. 4C).

Of note, the lipoxygenase pathway is implicated in ferroptosis due to its involvement in lipid peroxidation, a key feature of iron-dependent cell death (Chen et al. 2021a, b; Shintoku et al. 2017). Additionally, both glutathione peroxidase activity and peroxidase activity are related to lipid peroxidation and iron-dependent cell death (Yang and Stockwell 2016). These enrichment outcomes indicate a potential involvement of these DEGs in the progression of SLE by influencing the iron-dependent cell death pathway in Tfh cells.

To identify key genes involved in disease progression, an intersection analysis was performed between DEGs and the marker genes of Tfh cells (belonging to Cluster 7), resulting in the identification of HMGA1 as a potential key gene implicated in SLE (Fig. 4D). HMGA1 is a chromatin-associated protein belonging to the AT-hook subfamily of the High Mobility Group (HMG) protein family, primarily exerting its function through the regulation of gene transcription (Wu et al. 2021; Zhao et al. 2015; Cai et al. 2021).

Based on our transcriptomic sequencing results, HMGA1 exhibited a prominent rise in Tfh cells from MRL/lpr mice during our observation (P < 0.05) (Fig. 4E), which was further validated by Western blot experiments (P < 0.05) (Fig. 4F). Taken together, these observations indicate that the heightened HMGA1 expression could be vital in the pathological advancement of SLE by inhibiting iron-dependent cell death in Tfh cells.

HMGA1 regulates Tfh cell proliferation and ferroptosis

In exploring the function of HMGA1 in Tfh cells, we purchased human-derived Tfh cells. Subsequently, we successfully constructed HMGA1 knockdown and overexpression Tfh cell models using adenoviral infection and sorted CD4+CXCR5+ICOS+ cells by FCM (Figure S9C), which displayed normal cell viability (Figure S9D). The knockdown and overexpression of HMGA1 were accurately validated through Western blot and immunofluorescence experiments (P < 0.05) (Fig. 5A and Figure S9E).

Fig. 5

Impact of HMGA1 knockdown or overexpression on Tfh cells. Note: (A) Western blot analysis confirms the knockdown and overexpression effect of HMGA1. B The CCK8 experiment measures the influence of HMGA1 on the proliferation capability of Tfh cells. C Statistical results of iron content in Tfh cells in various groups using a detection kit. D Measurement of lipid peroxidation levels in Tfh cells using spectrophotometry. E Levels of MDA in each group were detected. F GSH/GSSG ratio in Tfh cells after different treatments. G Electron microscopic images of Tfh cells (Scale bar = 2 μm/50 nm). The scale bar in (G) represents 1 μm. Cell experiments were repeated three times, with three replicates each time. *** indicates a P < 0.001

The influence of HMGA1 on the proliferation of Tfh cells was investigated via CCK8 experiments. The data revealed that HMGA1 knockdown significantly suppressed the proliferative activity of Tfh cells in contrast with the control group (P < 0.05), while HMGA1 overexpression significantly enhanced their proliferative capacity (P < 0.05) (Fig. 5B).

Furthermore, we investigated the influence of HMGA1 on ferroptosis in Tfh cells. We measured intracellular iron levels (Fe2+), LPO, MDA levels, and the GSH/GSSG ratio. The results showed that HMGA1 knockdown led to an increase in intracellular Fe2+ levels, lipid peroxidation and MDA levels (P < 0.05), and a significant decrease in the GSH/GSSG ratio (Fig. 5C-F). Conversely, overexpression of HMGA1 resulted in opposite