Dynamic landscape of peripheral blood lymphocyte subsets in dengue patients: a multimodal single-cell and flow cytometry analysis

Abstract

Background

Dengue fever is a mosquito - borne infectious disease caused by the dengue virus (DENV). Understanding its immune response mechanisms is crucial for developing effective treatments and vaccines, especially regarding the dynamic changes of peripheral blood lymphocyte subsets. Traditional immunological methods have limitations in accurately analyzing lymphocyte subsets and their functional dynamics. However, the combination of single - cell RNA sequencing and flow cytometry can explore the immune response of dengue patients more deeply.

Methods

We integrated four public datasets from the Gene Expression Omnibus (GEO), comprising samples from acute dengue patients, healthy individuals, and convalescent patients, for single-cell RNA sequencing analysis. The analysis included data preprocessing, cell annotation, functional enrichment, T cell state scoring, intercellular communication analysis, and pseudotime analysis. Additionally, we collected peripheral blood samples from 81 DENV-1-infected individuals and 30 healthy controls to analyze lymphocyte subsets, activation status, and apoptosis using flow cytometry.

Results

Single-cell RNA sequencing identified 11 cell clusters (including T cells, B cells, and plasma cells). Functional analyses demonstrated diverse roles of lymphocyte subpopulations in immune responses and metabolic regulation, as well as frequent intercellular communication. Pseudotime analysis revealed the differentiation trajectory of CD4⁺ naïve T cells and related signaling changes. Flow cytometry showed significantly decreased absolute T cell counts, altered subpopulation distribution and activation, and increased apoptosis; immunological marker variations existed among T cell subpopulations, with correlations between absolute and relative counts.In conclusion, dengue virus infection significantly impairs lymphocyte composition and function, causing immune dysregulation. This study provides insights into the dengue immune regulatory network, with potential implications for targeted immunotherapy and vaccine development, though its small sample size limits the findings and requires further research.

Conclusions

Dengue virus infection significantly affects the composition and function of lymphocyte subsets, leading to immune dysregulation. This study provides new insights into the immune regulatory network of dengue fever and holds potential implications for the development of targeted immunotherapies and vaccines. However, the study has limitations, such as small sample size, and further research is needed in the future.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-13511-3.

Background

Dengue fever is a mosquito-borne infectious disease caused by the dengue virus (DENV), widely distributed in tropical and subtropical regions [1, 2]. This disease has a high incidence rate and may trigger a variety of clinical manifestations ranging from mild to severe due to a cytokine storm, including dengue hemorrhagic fever and dengue shock syndrome. These severe cases are often accompanied by a vigorous immune system response [3–5]. The immune response plays a crucial role in the pathological process of dengue fever. Among them, lymphocytes, especially T cells and B cells, are involved in the clearance of the virus and the regulation of immune tolerance. However, the specific dynamic changes of immune cells during dengue fever are still not fully understood, particularly the functional states and transitions of different lymphocyte subsets during the infection and recovery stages. The compositional and functional changes of peripheral blood lymphocyte subsets are regarded as important markers for evaluating the immune response to dengue fever. Although traditional immunological analysis methods can provide a preliminary understanding of immune cell subsets, there are still certain limitations in precisely typing cell subsets and revealing their functional dynamics. In recent years, advanced technologies such as single-cell RNA sequencing (scRNA-seq) and flow cytometry have provided a higher resolution and more comprehensive perspective for immunological research. Single-cell technology can conduct in-depth analysis of the gene expression of individual cells, thereby revealing the heterogeneity among cells. Flow cytometry, on the other hand, can provide real-time quantitative information on lymphocyte subsets through the detection of surface markers and the evaluation of cell functions. The combination of these two techniques can precisely depict the changes in peripheral blood lymphocyte subsets of dengue fever patients and their roles in the immune response.In this study, through the comprehensive analysis of single-cell and flow cytometry, we systematically evaluated the dynamic changes of peripheral blood lymphocyte subsets of dengue fever patients. Our aim is to reveal how these immune cell subsets participate in the immune response during dengue virus infection, especially their roles in virus clearance, immune tolerance, and possible immune escape. Through an in-depth analysis of immune function and subset remodeling, this study hopes to provide a new perspective for understanding the immune mechanism of dengue fever and offer potential theoretical basis for the development of future immunotherapeutic strategies and vaccines.

Methods

Data Sources In this study, we integrated three public datasets, which included single-cell RNA sequencing data (downloaded from the Gene Expression Omnibus (GEO) repository: https://www.ncbi.nlm.nih.gov/geo). Inthisstudy, weintegrated four publicly available datasets, including single-cell RNA sequencingdata retrieved from the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo). The datasets, “GSE280258”,“GSE147104”, “GSE167363” and “GSE220969” consist of 8 peripheral blood lymphocyte samples from acute dengue patients, 4 peripheral blood mononuclear cell (PBMC) samples from healthy individuals, and 1 PBMC sample froma patient 180 days post-DENV infection during the recovery phase.

Single-cell RNA sequencing analysis

We used the Seurat (V4.4.0) R software package to process and analyze the gene expression matrix. Samples with less than 200 transcripts per cell and more than 2500 transcripts per cell, as well as samples with a mitochondrial gene percentage exceeding 10%, were filtered out. We used the “NormalizeData” function in the Seurat software package to perform logarithmic transformation and normalization of the data using the LogNormalize method. The “FindVariableFeatures” function was used to identify highly variable genes. The variance stabilizing transformation (vst) method was applied to stabilize the variance, and the top 1000 genes with the highest variability were retained. Finally, the “ScaleData” function was used to remove the batch effects between samples. Principal component analysis (PCA) was used for dimensionality reduction, and the top 50 principal components were retained. The “FindCluster” function was used to partition the cells into multiple clusters. Harmony was used to remove the batch effects between samplings. The Uniform Manifold Approximation and Projection (UMAP) algorithm was used to map the cells onto a two-dimensional space for visualization.

Annotation and enrichment analysis of cell data

The “FindAllMarker” function was used to identify the differentially expressed genes (DEGs) in the cell populations in scRNA-seq. Genes that were positively expressed in more than 25% of the cells in each cluster were selected. Classic cell markers and the top 50 DEGs of each cluster were combined for cell annotation. To analyze the functions of each cluster, the R package clusterProfiler was used for GO and KEGG enrichment analysis [6].

Scoring of T cell states T lymphocytes are highly diverse

Their origins, differentiation trajectories, and functions vary. The remarkable diversity observed in T cells is reflected in the multiple unique states they exhibit. Exploring the complexity of T cell states can help us better understand how the immune system interacts with diseases and provide important clues for enhancing the efficacy of immunotherapies. We used the TCSS_Calculate function of the TCellSI package to score the states of different types of T cells [7].

Analysis of intercellular communication

There are complex interaction relationships among T cells. To understand the intercellular interactions between different T cell types, we used the CellChat package to construct an interaction network of peripheral blood T cells from dengue fever patients [8]. This R package is based on the consistent interactions between signaling ligands, receptors, and their cofactors. We set the ligand-receptor database as “CellChatDB.human”. We applied the “subsetCommunication” function to infer the CellChat network, used the “computeCommunProb” function to calculate the probabilities of intercellular communication, and the “computeCommunProbPathway” function to infer the probabilities of intercellular communication at the signaling pathway level. We filtered out the communication data observed in less than 10 cells.

Pseudotime analysis

To simulate the developmental trajectories of T lymphocyte subsets, we performed pseudotime analysis using the Monocle3 package [9]. After extracting the RNA expression data, cellular metadata, and gene metadata from the previously constructed single-cell Seurat object, we used the newcelldata_set function to construct a Monocle3 object. After removing the batch effects, we performed dimensionality reduction using the Umap method. Subsequently, we identified the starting point through unbiased clustering of the data and visualized the results using the Umap method.

Peripheral blood sample collection

We selected patients infected with DENV-1 and untreated from the Second Affiliated Hospital of Shantou University Medical College between July 2019 and December 2023.All enrolled patients were in the febrile phase of dengue infection (defined as within 0–7 days after onset of fever, with confirmed DENV-1 infection by nucleic acid detection). Cases with major diseases significantly affecting T cells were excluded. Finally, 81 eligible patients infected with DENV were screened as the experimental group (DENV group). Thirty healthy volunteers (healthy donor, HD) who underwent physical examinations at the Second Affiliated Hospital of Shantou University Medical College were selected as the control group (HD group). Five milliliters of fasting peripheral blood from the research subjects was drawn into an EDTA-K2 blood collection tube and thoroughly mixed immediately after blood collection.

Flow cytometry detection

The corresponding fluorescent antibodies (see Supplementary Materials Table S1) were added to the BD Trucount counting tube, 50 µL of anticoagulant-mixed peripheral blood was added, and the mixture was incubated in the dark at room temperature for 15 min. 450 µL of 1× lysin (BD, USA) was added for lysis for 10 min, 1 mL of PBS (Sigma, USA) was added and mixed, centrifuged at 500 g for 5 min, and after pouring off the supernatant, 1 mL of PBS was added to resuspend the cells. After adjusting the voltage and compensation of the flow cytometer (CATTO Ⅱ, BD, USA), the samples were detected on the machine. After removing debris and cell clumps, 50,000 cells were collected from each tube. The gating analysis was performed on the FCSA Diva software to obtain the number and percentage of gated cells for absolute counting. The calculation method is as follows: Absolute count of lymphocytes (cells/µL) = D * A / C / Vsample (Note: D: the number of sample events collected; A: the total number of standard microspheres in the tube; C: the number of standard microspheres collected; Vsample: the volume of the sample stock solution to be counted.)

Statistics

The SPSS.26 software was used for statistical analysis of the data. The Kruskal-Wallis test was used for significance analysis, the Spearman test was used for correlation analysis, and GraphpadPrism8.0 was used for graphing. A P value ≤ 0.05 was considered statistically significant.

Results

Lymphocyte subsets in the peripheral blood of acute dengue fever patients revealed by ScRNA-seq

In this study, peripheral blood lymphocyte samples from multiple sources were used for analysis. Specifically, it included 8 peripheral blood lymphocyte samples from acute dengue fever patients, which were processed through the 10×Genomics platform;4 peripheral blood PBMC sample from a healthy individual; and 1 peripheral blood PBMC sample from a convalescent patient 180 days after DENV virus infection.

Single-cell RNA sequencing analysis

We used the Seurat (V4.4.0) R software package to process and analyze the gene expression matrix. Samples with less than 200 transcripts per cell and more than 2500 transcripts per cell, as well as samples with a mitochondrial gene percentage exceeding 10%, were filtered out. We used the “NormalizeData” function in the Seurat software package to perform logarithmic transformation and normalization of the data using the LogNormalize method. The “FindVariableFeatures” function was used to identify highly variable genes. The variance stabilizing transformation (vst) method was applied to stabilize the variance, and the top 1000 genes with the highest variability were retained. Finally, the “ScaleData” function was used to remove the batch effects between samples. Principal component analysis (PCA) was used for dimensionality reduction, and the top 50 principal components were retained. The “FindCluster” function was used to partition the cells into multiple clusters. Harmony was used to remove the batch effects between samplings. The Uniform Manifold Approximation and Projection (UMAP) algorithm was used to map the cells onto a two-dimensional space for visualization.

Annotation and enrichment analysis of cell data

The “FindAllMarker” function was used to identify the differentially expressed genes (DEGs) in the cell populations in scRNA-seq. Genes that were positively expressed in more than 25% of the cells in each cluster were selected. Classic cell markers and the top 50 DEGs of each cluster were combined for cell annotation. To analyze the functions of each cluster, the R package clusterProfiler was used for GO and KEGG enrichment analysis [6].

Scoring of T cell states T lymphocytes are highly diverse

Their origins, differentiation trajectories, and functions vary. The remarkable diversity observed in T cells is reflected in the multiple unique states they exhibit. Exploring the complexity of T cell states can help us better understand how the immune system interacts with diseases and provide important clues for enhancing the efficacy of immunotherapies. We used the TCSS_Calculate function of the TCellSI package to score the states of different types of T cells [7].

Analysis of intercellular communication

There are complex interaction relationships among T cells. To understand the intercellular interactions between different T cell types, we used the CellChat package to construct an interaction network of peripheral blood T cells from dengue fever patients [8]. This R package is based on the consistent interactions between signaling ligands, receptors, and their cofactors. We set the ligand-receptor database as “CellChatDB.human”. We applied the “subsetCommunication” function to infer the CellChat network, used the “computeCommunProb” function to calculate the probabilities of intercellular communication, and the “computeCommunProbPathway” function to infer the probabilities of intercellular communication at the signaling pathway level. We filtered out the communication data observed in less than 10 cells.

Pseudotime analysis

To simulate the developmental trajectories of T lymphocyte subsets, we performed pseudotime analysis using the Monocle3 package [9]. After extracting the RNA expression data, cellular metadata, and gene metadata from the previously constructed single-cell Seurat object, we used the newcelldata_set function to construct a Monocle3 object. After removing the batch effects, we performed dimensionality reduction using the Umap method. Subsequently, we identified the starting point through unbiased clustering of the data and visualized the results using the Umap method.

Peripheral blood sample collection

We selected patients infected with DENV-1 and untreated from the Second Affiliated Hospital of Shantou University Medical College between July 2019 and December 2023.All enrolled patients were in the febrile phase of dengue infection (defined as within 0–7 days after onset of fever, with confirmed DENV-1 infection by nucleic acid detection). Cases with major diseases significantly affecting T cells were excluded. Finally, 81 eligible patients infected with DENV were screened as the experimental group (DENV group). Thirty healthy volunteers (healthy donor, HD) who underwent physical examinations at the Second Affiliated Hospital of Shantou University Medical College were selected as the control group (HD group). Five milliliters of fasting peripheral blood from the research subjects was drawn into an EDTA-K2 blood collection tube and thoroughly mixed immediately after blood collection.

Flow cytometry detection

The corresponding fluorescent antibodies (see Supplementary Materials Table S1) were added to the BD Trucount counting tube, 50 µL of anticoagulant-mixed peripheral blood was added, and the mixture was incubated in the dark at room temperature for 15 min. 450 µL of 1× lysin (BD, USA) was added for lysis for 10 min, 1 mL of PBS (Sigma, USA) was added and mixed, centrifuged at 500 g for 5 min, and after pouring off the supernatant, 1 mL of PBS was added to resuspend the cells. After adjusting the voltage and compensation of the flow cytometer (CATTO Ⅱ, BD, USA), the samples were detected on the machine. After removing debris and cell clumps, 50,000 cells were collected from each tube. The gating analysis was performed on the FCSA Diva software to obtain the number and percentage of gated cells for absolute counting. The calculation method is as follows: Absolute count of lymphocytes (cells/µL) = D * A / C / Vsample (Note: D: the number of sample events collected; A: the total number of standard microspheres in the tube; C: the number of standard microspheres collected; Vsample: the volume of the sample stock solution to be counted.)

Statistics

The SPSS.26 software was used for statistical analysis of the data. The Kruskal-Wallis test was used for significance analysis, the Spearman test was used for correlation analysis, and GraphpadPrism8.0 was used for graphing. A P value ≤ 0.05 was considered statistically significant.

Results

Lymphocyte subsets in the peripheral blood of acute dengue fever patients revealed by ScRNA-seq

In this study, peripheral blood lymphocyte samples from multiple sources were used for analysis. Specifically, it included 8 peripheral blood lymphocyte samples from acute dengue fever patients, which were processed through the 10×Genomics platform;4 peripheral blood PBMC sample from a healthy individual; and 1 peripheral blood PBMC sample from a convalescent patient 180 days after DENV virus infection. After quality control, a total of 42,574 high-quality cells were obtained from 13 samples. Through standardization, normalization, an unsupervised cluster detection algorithm (Seurat), and dimensionality reduction clustering, 13 different cell clusters were identified and projected(Fig. 1A, B]. Most of these clusters could be distinguished by classical hematopoietic lineage-defining genes, such as CD3, CD4, and CD8. To further verify the identities of these cell clusters, their transcriptomes were also compared with the published transcriptomes of human peripheral blood mononuclear cells [10–12]. Finally, we identified 8 T cell subsets, 1 B cell subset, 1 plasma cell subset, 1 NK cell subset, 1 neutrophil subset, and 1 DNT cell subset.Comparing cell proportions in healthy individuals, dengue-infected patients, and those in recovery, we found that dengue infection significantly increases total CD4⁺ T cell proportions, with a decrease in Naive CD4⁺ T cells and an increase in Th1 and Th17 subsets, indicating a disruption of CD4⁺ T cell homeostasis. In recovery, memory subsets (Tcm and Tem) rise, suggesting immune memory formation. For CD8⁺ T cells, Naive CD8⁺ T cell proportions decrease, while Tcm and Trm subsets expand, reflecting a strong antiviral response. In recovery, CD8⁺ T cells return to near-healthy levels, with sustained higher memory subset proportions, signaling a shift to memory-dominant immunity (Fig. 1C).For gene expression analysis of these subsets, a heatmap (Fig. 1D) was generated using the top 5 genes with the highest expression levels in each subset.In addition, conventional multicolor flow cytometry with a panel of 10 markers was applied to the peripheral blood cells of dengue fever patients, and a recognized gating strategy was applied (Fig. S1) [13].

Fig. 1.

Single-Cell Immune Landscape of Peripheral Blood Mononuclear Cells (PBMC) Across Healthy Controls, Acute Dengue Infection, and 180-Day Convalescence. A) UMAP visualization of immune cell clusters in PBMC, colored by cell type identity. B) UMAP visualization of immune cell clusters, colored by clinical status (Healthy, Acute Dengue, 180-day Convalescence). C) Bar plot showing the proportion of each immune cell type across healthy, acute dengue, and convalescent conditions. D) Heatmap of key marker gene expression across distinct immune cell clusters, with color-coded cell type identity. E) Feature plots showing the expression of canonical marker genes for immune cell types and functional states overlaid on the UMAP embedding

Functional analysis of lymphocyte subsets and intercellular regulatory network in dengue fever patients

We characterized the functional dynamics of T cells in healthy controls, acute dengue patients, and convalescent individuals. In healthy subjects, CD4 + Th1/Th17 and other T-cell subsets showed specific functional activation, mediating helper, regulatory, and cytotoxic functions. These processes rely on tightly regulated activation and quiescence controlled by transcription factors, cytokines, and immune microenvironmental homeostasis.During acute dengue infection, CD4+ naïve T cells, Tcm/Tem, and CD8 + Tcm were strongly activated. DNT cells and CD4 + Th1/Th17 cells displayed activation in both proliferative and quiescent states with partial senescence, but other functional programs remained inactive. This pattern indicates an early T-cell response in which proliferation was initiated but downstream effector functions were not fully engaged.After 180 days of recovery, viral clearance and restored immune homeostasis repaired T-cell signaling. All subsets regained quiescence and regulatory functions, with gradual recovery of proliferation and effector activities (Fig. 2A) [14–16].

Fig. 2.

Functional and enrichment analysis of immune cell subsets in dengue virus infection and recovery. A) Heatmap of T cell functional state enrichment scores. B) KEGG pathway enrichment analysis of immune cell clusters from dengue infection. C) Cell–cell communication analysis (CellChat) of dengue-derived immune cells: left, interaction strength network; right, ligand–receptor pair count network, with node size and edge thickness indicating interaction magnitude. D) Gene Ontology (GO) biological process enrichment analysis of immune cell subsets from dengue infection, with dot size and color representing gene count and adjusted P-value, respectively

Functional enrichment analysis (Fig. 2B, D) showed that CD4⁺ naïve T, Th1/Th17, and CD8⁺ Tcm cells were enriched for T-cell activation pathways, indicating early activation but incomplete differentiation. DNT cells were enriched for immune regulatory pathways with reduced function. Total immune clusters showed enrichment in viral infection pathways and upregulated metabolism for activation. B-cell immunity was not significantly enriched, indicating an ineffective response. The overall profile showed T-cell activation without full effector function, impaired regulation, co-activation of viral responses, and silenced humoral immunity. A number of studies have shown that there is a wide range of mutual regulatory relationships among lymphocyte subsets [17–19]. We built an intercellular communication network in dengue patients to analyze lymphocyte interactions. T-cell subsets frequently communicated, highlighting their central role in antiviral immunity (Fig. 2C). These findings reflect the host’s response to dengue and reveal the crucial roles of these cells in immunity, linked to metabolic regulation.

Detection of regulatory patterns in dengue fever lymphocytes based on the cell regulatory network

Lymphocyte subsets communicate and regulate functions through incoming and outgoing signals. Network analysis revealed that lymphocytes receive signals via various molecules.

The heatmap (Fig. 3A) shows CD4 + Tcm and CD4 + Tem cells as leaders in communication, with over 200 interactions, followed by CD8 + Tcm and CD8 + Trm cells. These memory T cells are “core nodes” in intercellular interactions, more active than naïve T cells and other immune cells (e.g., B and NK cells). The communication chord diagram (Fig. 3B) highlights their dominant role in signaling to B and NK cells, confirming their central regulatory function in the immune network.Signal integration (Fig. 3C) showed CD4 + Tcm and CD8 + Tcm with the highest signal output. MHC signaling revealed that memory T cells are key MHC signal producers, supporting MHC molecules’ role in antigen recognition and immune response (e.g., MHC-I for CD8 + T cell cytotoxicity, MHC-II for CD4 + T cell helper functions) [20–23]. Memory T cells enhance interactions with CD4 + T cells and immune memory by acquiring MHC molecules (e.g., CD8 + T cells via phagocytosis) [24].

Fig. 3.

Cell–Cell Communication Network of Immune Cell Subsets in Dengue Virus Infection A) Heatmap of intercellular communication frequency between immune cell subsets, with color intensity representing the number of ligand–receptor interactions. B) CellChat chord diagrams showing the strength (left) and count (right) of ligand–receptor interactions between major immune cell types. C) Scatter plot of efferent interactions intensity of immune cell subsets: Top: Global communication network; Bottom: MHC signaling pathway-specific network. The size of the dots represents the number of interactions. D) Dot plot of differentially expressed ligand–receptor pairs mediating communication between CD4 + Tcm and other immune cell subsets, with color representing communication probability and size indicating significance (p-value)

Figure 3D shows high communication probabilities for MHC-related signals (e.g., HLA, CNTF), with CD4 + Tcm as the signal source. CD4 + Tcm cells use MHC-II and other ligands to transmit activation/co-stimulatory signals to CD8 + T cells, B cells, and other immune cells, driving proliferation and activation, making them central to immune network coordination [25].

In conclusion, memory T cell subsets (especially CD4 + Tcm and CD8 + Tcm) form the core immune regulatory network through frequent interactions and signaling. This confirms their pivotal role in immune responses and provides targets for studying immune mechanisms in diseases like dengue. Lymphocyte pathways coordinate to precisely regulate immune responses, crucial in immune-related diseases such as dengue.

Pseudotime analysis of T cells in lymphocyte subsets

The pseudotime analysis (Fig. 4A, B) revealed a continuous differentiation trajectory of T cells from the naive to the effector/memory subsets during dengue virus infection.

Fig. 4.

Dynamic expression of key ligand-receptor genes during T cell differentiation trajectories. A) UMAP visualization of immune cell clusters, colored by cell type, showing the spatial distribution of distinct immune cell populations. B) UMAP visualization of pseudotime trajectory, colored by pseudotime, depicting the progression of cellular differentiation. C) Gene expression dynamics of CCL5, CD55, CLEC2D, HLA-B, HLA-C, and ITGB2 across pseudotime, colored by cell type. D) Gene expression dynamics of CCL5, CD55, CLEC2D, HLA-B, HLA-C, and ITGB2 across pseudotime, colored by pseudotime, revealing expression changes during differentiation

CD4⁺ T cells, DENV antigens are continuously presented via MHC-II, and together with γc cytokines like IL-2 and IL-7, drive naive T cell proliferation and differentiation into central memory T cells (Tcm) [26]. With ongoing antigen stimulation and inflammatory cytokines like IL-12 and IFN-γ, Tcm further differentiate into effector memory T cells (Tem) [27]. CCL5 and ITGB2 expression increases over time, supporting Tem cell migration to infected tissues. IL-12 and IFN-γ promote Th1 differentiation via the STAT4/T-bet pathway, while IL-6 and TGF-β induce Th17 differentiation via the STAT3/RORγt pathway, contributing to mucosal immunity and inflammation [28, 29].

In CD8⁺ T cells, naive cells differentiate into Tcm under IL-7 and IL-15, maintaining memory potential and expanding rapidly upon secondary infection. Tcm further differentiate into Tem or Trm [30, 31]. During acute infection, Tem cells, driven by antigen stimulation and IL-2, produce granzyme and perforin, enhancing killing of DENV-infected cells [32]. Trm formation relies on tissue signals (e.g., TGF-β) and adhesion molecules like ITGB2, allowing them to reside in peripheral tissues like skin for local protection [33]. HLA-B/C supports MHC-I-mediated antigen recognition and memory maintenance, while ITGB2 is key for Trm retention [38]. This differentiation is driven by DENV antigen presentation, cytokine signaling, and tissue-specific cues [34–37]. Imbalance in this process can lead to immune dysregulation, as seen in dengue infection.

Gene expression (Fig. 4C, D) shows CCL5, CLEC2D, and ITGB2 upregulation, reflecting enhanced effector function and migration, while CD55 downregulation may modulate complement sensitivity [38]. HLA-B/C shows biphasic expression, indicating dynamic MHC-I signaling during T cell activation and memory [28]. We suggest CD4⁺ Tcm coordinate CD8⁺ T and B cell responses via MHC-II interactions, aligning with previous studies [39–41], and that dengue drives T cell differentiation into protective memory subsets [42–44].

Flow cytometry analysis of peripheral blood T cells and subsets in dengue - infected patients

To validate the results of bioinformatics analysis and explore the functions of peripheral blood T cells and the immune response in dengue fever patients, this study included 81 individuals infected with DENV − 1 and 30 healthy controls, and analyzed lymphocyte subsets using flow cytometry. Lymphocytes were distinguished from bead - labeled particles by gating with CD45 and SSC, and T cell subsets were classified based on CD3, CD4, and CD8 [45–47]. Additionally, CD25, CD28, CD45RA, CD45RO, CD38, HLA - DR, and CD95 were further labeled to evaluate the activation and apoptosis status of T cells [48–53]. The study found that the absolute counts of peripheral blood T cells and their subsets in patients infected with DENV significantly decreased, including total lymphocytes, CD3⁺T, CD4⁺T, CD8⁺T cells, etc. The subsets of T cells labeled with CD25, CD28, CD45RA, and CD45RO also showed a downward trend (Fig. 5A).

Fig. 5.

Comparison of absolute counts of immune subsets between dengue fever patients and healthy individuals. (B) Comparison of the percentage change amplitudes of CD25, CD28, CD45RA, CD45RO, CD38, HLA - DR, and CD95 in CD4 + T cell and CD8 + T cell subsets in patients infected with dengue virus. (C) Correlation analysis of relative counts (left) and absolute counts (right) among T cell subsets, activation phenotypes, and apoptosis phenotypes in patients infected with dengue virus. (* p < 0.05, ** p < 0.01, *** p < 0.001)

In terms of the proportions and activation status of cell subsets, DENV infection significantly changed the proportions of T cell subsets and enhanced the activation response. The proportion of CD4⁺T cells decreased, while the proportions of CD8⁺T, CD4⁺CD8⁺T, and CD4⁻CD8⁻T cells increased, especially the proportion of CD4⁺CD8⁺T cells increased significantly. The proportions of T cells labeled with CD25⁺and CD45RO⁺increased, the proportion of CD45RA⁺T cells decreased, the proportion of CD8⁺CD28⁻T cells increased, and the proportion of CD8⁺CD28⁺T cells decreased (Fig. S2). Analysis of activation markers showed that the numbers of CD38⁺ and HLA - DR⁺ positive T cells increased significantly. The absolute counts of CD3⁺CD38⁺T and CD8⁺CD38⁺T cells increased significantly, while the count of CD4⁺CD38⁺T cells decreased. The absolute counts and percentages of CD38 and HLA - DR double - positive T cells increased significantly. Analysis of the apoptosis status of T cells showed that after DENV infection, the absolute counts of CD3⁺CD95⁺, CD4⁺CD95⁺, and CD8⁺CD95⁺T cells decreased, but their percentages increased significantly, indicating enhanced apoptosis of T cells (Fig. 5A, Figure S2, 3).

Further analysis of the changes in immune markers of CD4⁺T and CD8⁺T cell subsets revealed that among CD8⁺T cell subsets, the decrease in CD8⁺CD25⁺T cells was relatively small, the number of CD8⁺CD38⁺T cells increased significantly, and the number of CD8⁺HLA - DR⁺T cells decreased significantly. In terms of relative counts, the number of CD4⁺T cells decreased while the number of CD8⁺T cells increased in the DENV group. The increases in the expressions of CD25, CD38, and CD95 in CD8⁺T cell subsets were significantly higher than those in CD4⁺T cells, and the expression of CD28 decreased in CD8⁺T cells but increased in CD4⁺T cells. There were no significant differences in the double - expression changes of CD45RA, CD45RO, and CD38⁺HLA - DR⁺ between the two subsets. It is worth noting that for some T cell subsets, their relative counts increased while their absolute counts decreased. In patients infected with DENV, the expression of HLA - DR increased significantly in CD4⁺T cell subsets, the expressions of CD25, CD38, and CD95 increased significantly in CD8⁺T cell subsets, and the decrease in the expression of CD28 was significant in CD8⁺T cell subsets (Fig. 5B).

The study also found that DENV infection led to an increase in activated T cells in peripheral blood, which was consistent with our previous analysis that CD4⁺Treg and CD4⁺naive cells had higher scores, and CD4⁺Tcm cells had a higher score in the proliferative state (Fig. 2A). The systemic inflammatory response triggered by dengue virus infection allows the virus to invade multiple organs [54]. After the immune response was activated by the virus, cytotoxic T cells such as CD8⁺Tem were recruited to the site of infection [55]. At the same time, DENV infection induced apoptosis of T cells [56], resulting in a decrease in the absolute count of peripheral blood T cells, which was consistent with the higher scores of CD8⁺Tem, CD4⁺Th1, and CD4⁺Trm cells in the exhausted state obtained from the TCellSI analysis (Fig. 2A).

In addition, the impaired function and proliferative inhibition of CD4⁺T cells led to a significant decrease in their numbers, resulting in an increase in the relative proportions of other subsets (Fig. 5C) [57]. When the absolute counts of some T cell subsets, such as CD4⁺T cells, decreased significantly, the proportional balance was disrupted, and the relative counts of CD8⁺T cells increased. To cope with the immune imbalance, compensatory differentiation and regulation of T cell subsets may occur in the body [54]. In the study of T cell subsets in dengue fever patients, the relative counts of subsets such as CD4⁺CD25⁺ and CD3⁺CD25⁺ were positively correlated with the absolute counts of subsets such as CD4⁻CD8⁻T and CD8⁺T, possibly due to their joint participation in the immune response or functional collaboration [58, 59]. The absolute counts of subsets such as CD3⁺T and CD8⁺HLA - DR⁺, and the relative counts of subsets such as CD4⁺CD95⁺ and CD8⁺CD45RA⁺ were negatively correlated, possibly due to immune feedback regulation and functional antagonism [60].

Discussion

Our study combines scRNA-seq and flow cytometry to track lymphocyte subset dynamics during dengue virus (DENV) infection, revealing immune mechanisms that drive the shift from acute infection to recovery. In line with Zhao et al. [61], we find DENV induces specific T cell activation, differentiation, and apoptosis, along with B cell and NK cell activation, confirming findings in prior research [36, 62]. This emphasizes the importance of dengue-specific immunology.

Our study reveals that memory T cell subsets (CD4⁺Tcm and CD8⁺Tcm) are key regulatory nodes in the dengue patient communication network, consistent with Ioannidis et al.‘s work on “memory-driven immune responses” in chronic infections [44, 63, 64]. scRNA-seq analysis shows MHC-II and MHC-I pathways as central communication axes for these T cells. We propose memory T cells coordinate antiviral responses through co-stimulatory signals, not just initial T cell activation [65]. Functional enrichment supports this, with upregulated viral and inflammatory metabolic pathways during acute infection.

We found incomplete T cell effector functions during acute DENV infection, supporting the “immune hyporesponsiveness” concept by Waickman et al. [66, 67]. Despite early activation, CD4⁺ and CD8⁺ T cells fail to fully engage effector functions, which may be an adaptive mechanism to prevent cytokine storms [43, 67].TCellSI analysis shows higher exhaustion scores for CD8⁺Tem, CD4⁺Th1, and CD4⁺Trm cells during acute infection, with flow cytometry confirming increased apoptosis (CD95⁺). This exhaustion-apoptosis axis reduces T cell numbers, affecting viral clearance and increasing secondary infection risk [14]. CD8⁺T cells exhibit more activation and apoptosis features, while CD4⁺ T cells only show HLA-DR upregulation. These findings align with de Almeida et al. [37], and correlation analysis shows a negative correlation between CD8⁺HLA-DR⁺ and CD4⁺CD95⁺, suggesting these markers as potential non-invasive biomarkers for early dengue diagnosis and intervention.

Our study suggests potential directions for dengue immunotherapy and vaccine development, especially for seronegative individuals. The role of CD4⁺Tcm and CD8⁺Tcm indicates that vaccines inducing memory T cell responses, rather than just neutralizing antibodies, could improve protection [27, 68]. Vaccines targeting MHC-II/MHC-I epitopes from DENV non-structural proteins (e.g., NS3, NS5) may generate functional memory T cells [69, 70]. Additionally, T cell subset-specific markers support therapies like PD-1/PD-L1 inhibitors to prevent T cell exhaustion or caspase inhibitors to regulate apoptosis, restoring effector function and reducing immune pathology [71, 72]. MHC signaling (HLA, CNTF) offers potential small molecule targets to enhance CD4⁺Tcm co-stimulatory signals or inhibit excessive MHC signaling, improving immune responses in severe cases [37, 44]. Overall, our findings provide new insights for immune intervention, but further validation is needed.

This study has limitations. The small sample size, focused on mild dengue patients’ peripheral blood, may not fully reflect the immune response in all cases.Immune cells from other organs were not included, limiting our overall understanding.Although preliminary lymphocyte interactions were observed, the molecular mechanisms and key signaling pathways remain unclear and need further investigation.Additionally, dengue serotypes were not identified, and viral load data were unavailable, preventing correlation analysis between viral burden and immune parameters.The limited convalescent samples and incomplete data on dengue severity and infection stages also hindered stratified and longitudinal analyses.We hope future research, with larger and more comprehensive sample collections, will address these gaps and explore the long-term impact of dengue virus on immune memory and the role of different strains.

Conclusion

In conclusion, our study has put forward some key findings. Firstly, we have observed the dynamic change patterns of immune cell subsets in dengue fever patients, providing preliminary data support for a better understanding of the changes in immune responses during disease progression. By combining single - cell and flow cytometry analyses, we have explored the mechanisms of interactions and functional regulation among lymphocyte subsets, thus deepening our understanding of the immune regulatory network in dengue fever. In addition, our discussion of the roles of different lymphocyte subsets in the immune response has broadened our understanding of lymphocyte crosstalk and provided potential ideas for the treatment of dengue fever in the future. We hope that these findings can contribute to the development of more targeted immunotherapeutic strategies and vaccines, although there is still much room for further exploration.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

DENV

Dengue virus

scRNA-seq

Single-cell RNA sequencing

GEO

Gene Expression Omnibus

PBMC

Peripheral blood mononuclear cell

PCA

Principal Component Analysis

UMAP

Uniform Manifold Approximation and Projection

DEGs

Differentially expressed genes

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

TCSS_Calculate

A function used to calculate the state scores of T cells (from the TCellSI package)

CellChat

R package for constructing intercellular communication networks

Tregs

Regulatory T cells

Th1

T helper 1 cell

Trm

Tissue-resident memory T cell

Tem

Effector memory T cell

Tcm

Central memory T cell

MHC

Major Histocompatibility Complex

TCR

T cell receptor

APCs

Antigen-presenting cells

IL

Interleukin

IFN-γ

Interferon-gamma

TNF-α

Tumor Necrosis Factor-alpha

Author contributions

J.C.W.designed the study, conducted scRNA-seq bioinformatics analysis, drafted the manuscript, and coordinated the research. D.H. performed flow cytometry experiments, processed samples, and analyzed related data. J.X.Y. assisted with scRNA-seq data processing, conducted pseudotime analysis, and prepared key figures. Q.X.G. and Y.L. supported intercellular communication analysis and validated pathway interactions. Z.W.Z. performed T cell state scoring and analyzed subset dynamics. J.Y. conceived the study, supervised the work, revised the manuscript critically, and secured funding. All authors reviewed and approved the final version.

Funding

This work was supported by the Basic and Applied Basic Research Foundation of Guangdong Province, No. 2024A1515220129, and the Guangdong Provincial Basic and Applied Basic Research Fund, Natural Science Foundation Project (2024A1515012663).

Data availability

The single - cell RNA sequencing data used in this study were downloaded from the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo), including datasets GSE280258, GSE147104,GSE167363, and GSE220969. The datasets during the current study related to flow cytometry are available from the corresponding author upon reasonable request.

Declarations

Ethical approval

As a retrospective study, it did not involve clinical trials. This retrospective study was granted an exemption from ethical review by the Ethics Committee of the Second Affiliated Hospital of Shantou University Medical College (Exemption Approval No. 2025-25). The study was conducted in accordance with the Declaration of Helsinki. Informed consent was waived due to the retrospective nature of the study and the ethical review exemption obtained.

Competing interests

The authors declare no competing interests.