This study was approved by the Ethics Committee of The First Affiliated Hospital of Guangxi Medical University (IRB Protocol Number: 2022-KT-Guike-127), and all donors signed written informed consent. Eight AIGAs-positive patients (P1–P8) and three healthy controls (H1–H3) were enrolled for this study; all patients were from the Department of Respiratory and Critical Care Medicine, The First Affiliated Hospital of Guangxi Medical University, and were diagnosed with AIGAs immunodeficiency syndrome. Among these patients, P2, P3, P4, and P5 were in the stable phase with immune damage, while P1, P6, P7, and P8 were in the infective phase.

In the study, serum AIGAs were assessed using an indirect enzyme-linked immunosorbent assay (ELISA), while the neutralizing capacity of AIGAs was determined via Western blotting. All patients tested positive for AIGAs through ELISA, which were subsequently confirmed to possess neutralizing abilities via Western blotting.

6 mL of peripheral venous blood was collected from each sample, and PBMCs were isolated using density gradient centrifugation with Ficoll-Hypaque.

Infective phase was defined as patients exhibiting symptoms of infection, such as chills, fever, cough, sputum, skin, and soft tissue infections, along with elevated infective indices. The stable phase with an immune damage was defined as the initial clinical infective manifestation being controlled, but patients still having high levels of AIGAs and increased immune indices such as GLB, IgG, IgE, IgG4, eosinophils, and ESR; or high levels of AIGAs as well as clinical symptoms, such as nausea, anorexia, rash, and immune damage to the eyes, which are not caused by infection.

Raw sequencing data were processed with Illumina’s bcl2fastq and the 10x Genomics Cell Ranger pipeline (v3.1.0) for demultiplexing, alignment to the GRCh38/GRCm38 genome, and gene counting. This yielded 136,782 single cells from 3 healthy donors and 8 patients sequenced on the 10x Genomics Chromium platform.

Downstream analysis was performed in Seurat (v3.1.1). We filtered the data to retain 108,709 high-quality cells by removing cells with <500 or >5,000 genes, <500 UMIs, or >25% mitochondrial content, and genes detected in <3 cells. Data were log-normalized, and the top 10 principal components from PCA were used for SNN-based clustering and UMAP visualization. Cluster markers were identified using Seurat’s FindAllMarkers (Wilcoxon test), requiring expression in >10% of cells and an average log₂FC > 0.25.

Flow cytometry was performed to validate scRNA-seq findings. Lymphocyte subsets (T, NK, B, CD4⁺, CD8⁺) were quantified using the clinically approved BD Multitest™ 6-Color TBNK Reagent and analyzed in Kaluza software after FSC/SSC and CD45⁺ gating (Supplementary Figure S1).

T-helper subsets (Th1, Th2, Th17, Tc1, Tc2) were analyzed using a cytometric bead array (CBA) for IFN-γ⁺ (Th1/Tc1), IL-4⁺ (Th2/Tc2), and IL-17A⁺ (Th17) cells. Data were acquired on a calibrated flow cytometer and analyzed with FCAP Array v3.0 software, following quality control criteria of >1,800 total beads and ≥150 events per cytokine population.

B-cell subpopulations were characterized by multi-step FACS gating. After doublet exclusion and lymphocyte (FSC/SSC, CD45⁺) gating, total B cells were identified as CD19⁺. Subsets were defined using antibodies against CD27, IgD, IgM, CD38, CD24, and CD21 as: naïve (CD27⁻IgD⁺), marginal zone-like (CD27⁺IgD⁺), memory (CD27⁺CD38^dim) including class-switched (IgD⁻IgM⁻) and unswitched (IgM⁺) variants, plasmablasts (CD27^hiCD38^hi), transitional (CD27⁻IgM⁺CD38^hiCD24⁺), and CD21^low B cells (CD38^lowCD21^low).

Serum levels of the following cytokines were measured in 20 AIGAs-positive patients and 10 healthy controls: TNF-α (Cat. #m1064303V), GM-CSF (Cat. #m1025281V), IL-6 (Cat. #m1058097V), IL-5 (Cat. #m1058095V), IL-17 (Cat. #m1058051V), IL-4 (Cat. #m1058093V), TGF-β (Cat. #m1064258V), IL-12 (Cat. #m1058044), IL-2 (Cat. #m1058063V), and IFN-γ (Cat. #m1057856V). An indirect ELISA was performed according to the manufacturer’s instructions (Shanghai Enzyme-linked Biotechnology).

The continuous data were presented as mean ± standard deviation (x ± s) for normally distributed data, and comparisons were made using independent samples t-tests. For non-normally distributed continuous data, the median and interquartile range were used, and comparisons were made using independent samples nonparametric rank sum tests. A significant difference was defined as a two-tailed P value < 0.05. Statistical analysis and graphing were performed using SPSS (version 27.0) and GraphPad Prism (version 10).

This study enrolled 8 patients positive for AIGAs (case group), designated as P1 to P8, and 3 healthy controls (control group), designated as H1 to H3. The baseline characteristics of all participants are presented in Supplementary Table 1. There were no statistically significant differences in gender or age distribution between the case and control groups (P > 0.05).

After rigorous quality filtering, a total of 75,631 and 33,078 PBMCs were extracted from AIGAs-positive patients and healthy individuals, respectively. PBMCs immunocytes in AIGAs immunodeficiency syndrome mainly include T cells, B cells, monocytes, and natural killer T (NKT) cells. The immunological landscape of immunocytes was shown by uniform manifold approximation and projection (UMAP, Figure 1A). T cells specifically expressed CD3D and CD3E, which distinguished them from other clusters. B cells were characterized by high expression of MS4A1 and CD79A, and NKT cells were identified based on the high expression of NKG7. Two subsets of monocytes were defined; that is, CD14⁺ monocytes had enriched CD14, S100A8, and S100A9, and CD16⁺ monocytes had enriched FCGR3A (CD16, Figure 1B).

The proportion of immunocytes differed significantly between AIGAs-positive patients and healthy controls. Specifically, patients exhibited a pronounced increase in CD14⁺ monocytes (16.35% vs. 7.42% in controls) and a moderate increase in B cells (11.6% vs. 10.24%). Conversely, we observed a decrease in T cells (56.78% vs. 62.56%) and NKT cells (14.43% vs. 18.86%) (Figure 1C).

The immunocyte proportion also varied among patients and across different disease phases in AIGAs-positive patients (Figures 1D, E). Compared with healthy individuals (Health), patients in the stable stage exhibited a marked increase in CD14⁺ monocytes (11.35% vs. 7.42%) alongside decreases in T cells (63.97% vs. 62.56%), NKT cells (13.98% vs. 18.86%), and B cells (10.04% vs. 10.24%). During the infective stage, this immune landscape shifted further: CD14⁺monocytes expanded dramatically to 22.52%, while T cells dropped substantially to 47.9%. Conversely, B cells increased to 13.51% compared to healthy (10.24%) group. NKT cells in the infective stage (14.97%) remained lower than healthy controls (18.86%).

T cells were categorized into two groups, namely, CD8⁺ T and CD4⁺ T cells, based on the expression of CD4, CD8A, and CD8B. The immunological landscape of T cell subsets was visualized using uniform manifold approximation and projection (UMAP) in Figure 2A. Figure 2B presents a heatmap of the expression levels of specific gene markers in each T cell subset. An increase in the number of CD4⁺ T cells (51.78%) and a decrease in the level of CD8⁺ T cells (44.53%) were observed in the case group (Figure 2C). The proportion of T cell subsets varied between individual patients and patients under different conditions (Figures 2D, E).

Naive CD8⁺ T cells (CD8_Tn; LEF1, SELL, TCF7), GNLY effector memory CD8⁺ T cells (GNLY_CD8_Tem; GNLY, GZMH, GZMB), and GZMK effector memory CD8⁺ T cells (GZMK_CD8_Tem; GZMK, GZMA) were identified in the CD8⁺ T cell population (17). CD4⁺ T cells mainly consist of effector memory CD4⁺ T cells (CD4_Tem; LTB, IL7R), naive CD4⁺ T cells (CD4_Tn; LEF1, SELL, TCF7), and regulatory cells (Treg; FOXP3).

The level of CD4_Tem increased in the case group, and that of CD4_Tn decreased. Upon further analysis, CD4_Tem showed a higher expression of the IFN-γ gene (IFNG) in the case group (16.62%) than in the control group (6.94%) (Figure 3).

The B cells subgroup consists of naive B cells expressing IGHD, IGHM, and TCL1A (naive_B), memory B cells expressing AIM2 and TNFRSF13B (memory_B), plasma cells expressing IGHA1and IGHG1 (plasma), and an atypical B cell subpopulation expressing IFN-stimulated genes (ISG) (ISG_B) (Figures 4A, B). The proportions of these B cell subpopulations varied among individuals (Figure 4C). B cell subpopulation analysis revealed a profound shift in AIGAs-positive patients, characterized by an expansion of naive B cells (78.21% vs. 54.32% in health) and ISG⁺ B cells (2.95% vs. 0.53%), alongside a contraction of memory B cells (9.54% vs. 27.35%) and plasma cells (9.30% vs. 17.79%). This distribution was maintained across disease stages, with ISG⁺ B cells persistently elevated (stable: 4.03%; infective: 2.70%) and memory B cells remaining suppressed (stable: 14.35%; infective: 8.41%). Notably, plasma cells were decreased in both the stable and infective stages (21.57% and 6.44%, respectively) compared to health (17.79%) (Figures 4D, E).

Differentially expressed genes for each cell subsets were analyzed using the likelihood-ratio test. Upregulated genes were selected based on both a log2 FC ≥ 0.36 and a P value ≤ 0.01 in AIGAs immunodeficiency syndrome patients. Analysis of T cell subsets revealed a significant upregulation of interferon response genes, including IFNG, IFRD1, ISG15, ISG20, IFITM2, IFI6, and IFI44L, in most T cell subtypes (Figure 5). HLA-DRB1, HLA-DQB1, and HLA-DQA1, were also upregulated in most T cell subtypes (Figure 5). Genes belonging to the Fos gene family, such as FOS and FOSB, along with those from the JUN family, including JUN, JUNB, and JUND, were broadly expressed in all T cell subtypes (Figure 5).

Subsequently, the DEGs of B cell subsets were analyzed. IFI44L, IFITM1, HLA-DRB1, HLA-DQB1, HLA-DQA1, JUND, JUNB, JUN, FOSB, FOS, and CD69 were significantly upregulated in most B cell subtypes (Figure 5). IGHE was significantly upregulated in plasma cells in patients (Supplementary Figure S2).

Monocyte upregulated differential gene analysis also revealed a series of interferon response genes, including IFITM3, IFI6, IFI44L, IFI44, IFI16, IFNGR1, IFITM2, IFRD1, and ISG15 genes. Genes such as HLA-DRB1, HLA-DQB1, HLA-DQA1, JUND, JUNB, JUN, FOSB, and FOS, which are upregulated in most subtypes of T and B cells, were also highly expressed in monocytes. Additionally, S100A9, S100A12, S100A4, S100A11, and S100A8 were highly upregulated in monocytes (Figure 5).

Moreover, the upregulated differential genes of NKT cells were simultaneously analyzed, and a series of interferon-stimulated genes were found, including IFNG, IFI44L, IFI6, ISG15, ISG20. The expression of HLA-DRB1, HLA-DQB1, and HLA-DQA1 was also upregulated (Figure 5).

The analysis of the up-regulated genes in the above cell subsets revealed that a number of interferon genes and human leukocyte antigen genes, as well as the FOS and JUN gene families, were up-regulated in most cell subsets of the patients (Figure 5).

Finally, we analyzed downregulated differentially expressed genes for each cell subset and chose downregulated genes based on log2 FC ≤ 0.36 and P value ≤ 0.01 in AIGAs immunodeficiency syndrome patients. IRF1 genes were downregulated in all cell subgroups, while HLA-DQA2 and HLA-DRB6 were downregulated in most cell subgroups (Figure 5). IGHA1 and IGHA2 genes was significantly downregulated in plasma cells (Supplementary Figure S2). The complete datasets corresponding to all these analyses are provided in Supplementary Table 2.

Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were utilized to investigate cell subset-specific pathways in patients with differentially expressed genes (DEGs). GO analysis revealed that the majority of the gene products were located in the plasma membrane, suggesting a focus on cell-surface communication and environmental sensing. The analysis also indicated that the metabolic pathways of multiple cell subsets were associated with immune processes or immune responses. Notably, interferon signaling, including type I IFN (mainly IFN-α and IFN-β) and type II IFN, was activated in B cells, CD14+ monocytes, and NKT cells (Figure 6). This broad activation of IFN pathways across distinct immune lineages implies a robust, systemic antiviral state or a heightened inflammatory environment in the patients. Furthermore, enrichment was observed in pathways related to defense responses to viruses and fungi, antigen processing and presentation, as well as B cell activation (Figure 6). The co-enrichment of antiviral defense and antigen presentation pathways highlights a coordinated innate and adaptive immune mobilization, while enhanced B cell activation signals suggest an ongoing humoral immune response.

The KEGG pathway analysis was simultaneously conducted, revealing enrichment of several immune-associated pathways across various cell subsets. These pathways include Antigen Processing and Presentation, Th17 Cell Differentiation, Th1 and Th2 Cell Differentiation, and the IL-17 Signaling Pathway. The enrichment of these T cell polarization and IL-17 pathways points to a complex T-cell-mediated immune landscape, potentially involving both pro-inflammatory (Th1/Th17) and other adaptive mechanisms. Additionally, pathways related to infectious diseases such as Coronavirus Disease (COVID-19), Epstein-Barr Virus Infection, and Salmonella Infection were identified (Figure 7). This pattern suggests that the patients’ immune gene expression profile shares similarities with responses to specific viral and bacterial pathogens, which could provide clues about the disease etiology or triggering factors. Moreover, signal transduction pathways such as the TNF Signaling Pathway and the JAK-STAT Signaling Pathway were enriched as well (Figure 7). The enrichment of these key pro-inflammatory (TNF) and cytokine signaling (JAK-STAT) hubs further underscores the presence of a strong inflammatory milieu and offers potential mechanistic insights and therapeutic targets.

Given the marked increase of CD4_Tem and CD14⁺ monocytes in AIGAs immunodeficiency syndrome patients, we specifically investigated their intercellular communication networks. CellChat analysis disclosed a bidirectional augmentation of both pro-inflammatory and antigen-presenting pathways in AIGAs peripheral blood (Figures 8, 9).

Key upregulated ligand-receptor axes included: the adhesion/recruitment pair SELPLG-SELL, the inflammatory amplifier MIF-(CD74+CD44), MHC-II antigen presentation (e.g., HLA-DRB5/DRB1/DRA-CD4), and the potential negative feedback mediator HLA-F-LILRB1. Notably, the anti-inflammatory resolution axis ANXA1-FPR1 was also prominently active, indicating an insufficient compensatory response against the overall inflammatory milieu. These results collectively demonstrate, at single-cell resolution, a communication imbalance in AIGAs peripheral blood defined by heightened antigen presentation and leukocyte recruitment alongside overwhelmed inhibitory signaling, thereby elucidating potential mechanisms underlying sustained Th1 inflammation and monocyte hyperactivation.

Pseudotemporal ordering, seeded by CD4_Tn and CD8_Tn and sequentially resolved to CD4_Tem and GNLY⁺CD8_Tem, corroborated the fidelity of T-cell sub-cluster annotation (Supplementary Figure S3).

Compared with healthy controls, AIGAs-derived trajectories were globally shifted toward terminal effector fates; notably, the second bifurcation was significantly skewed toward CD4_Tem commitment, indicating that peripheral T cells in patients persist in a “high-effector, low-quiescent” state and providing further evidence of disease-associated hyper-immunity. Guided by this observation, we focused on CD4⁺ T-cell subsets. Lineage-restricted pseudotime (Supplementary Figure S4) bifurcated from CD4_Tn into CD4_Tem and Treg branches, recapitulating a “naive→ effector/regulatory” dichotomy. Cell density at pseudotime node 4 was markedly increased in patients and preferentially routed to CD4_Tem. Differential gene analysis of patient-versus control-derived CD4_Tem at this node identified significant up-regulation of LYZ, S100A8 and S100A9—perfectly mirroring the global differential signature(Supplementary Figure S5).

B-cell pseudotime reconstruction revealed that, although both groups displayed the canonical naive→memory→plasma-cell axis, patients exhibited a substantially more intricate differentiation topology (Supplementary Figure S6). Three disease-specific features were evident: (i) an ISG_B-cell cluster emerged exclusively in patients, implying aberrant interferon-pathway activation within the patient-specific trajectory (Figure 10); (ii) branch-point analysis showed an extra bifurcation (node 1) in patients that generated plasma-cell and memory-B-cell subsets completely absent in controls, whereas both groups shared these populations at node 2 (Figure 10); and (iii) BEAM analysis of node 2 demonstrated that the earlier-emerging plasma cells were enriched for IGHG1, IGHA2, IGHG3 and IGHG4 (Figure 11). Consequently, we performed differential gene analysis on node 2 and node 1 from the patient group. This analysis revealed that at node 2, the later-emerging plasma cell fraction displayed selective up-regulation of IGLC3 (Figure 12).

A total of 25 individuals were enrolled, including 15 AIGAs-positive patients and 10 healthy controls. As shown in and Figure 13A, the proportion of Th1(IFN-γ⁺CD4⁺) cells(21.84[14.87-27.57] vs.11.96[7.19-15.74] in control), Th17(IL-17⁺CD4⁺) cells(2.56[1.58-4.98] vs. 1.13[0.86-1.74]), Treg (CD25⁺FOXP3⁺)(6.22[3.51-7.86] vs.2.11[1.54-3.53]) in CD4⁺ T cells increased in AIGAs-positive patients (P < 0.05). Conversely, no statistically significant variances were observed in the distribution of Th2(IL-4⁺CD4⁺), Tc1(IFN-γ⁺CD8⁺), and Tc2(IL-4⁺CD8⁺) between the two groups (P>0.05).

Subsequently, B cell subsets were evaluated. Compared to healthy controls, AIGAs-positive patients exhibited a significant reduction in the frequencies of marginal zone B cells (CD19⁺CD27⁺IgD⁺; 2.87% [IQR 1.71-4.45] vs. 8.60% [IQR 6.77-15.65]), memory B cells (CD19⁺CD27⁺CD38^dim; 13.85% [5.72-20.23] vs. 22.96% [16.39-33.83]), and class-switched memory B cells (CD19⁺CD27⁺IgD⁻IgM⁻; 6.11% [2.39-9.10] vs. 10.18% [5.35-15.77]) (Figure 13B).

In our study, we assessed the serum levels of TNF-α, GM-CSF, IL-6, IL-5, IL-17, IL-4, TGF-β, IL-12, IL-2, and IFN-γ in 20 AIGAs-positive patients and 10 healthy controls. As shown in Table 1, there were no significant differences in IFN-γ, IL-2, TNF-α, IL-12, IL-4, IL-6, and TGF-β levels between the two groups (P = 0.008). However, AIGAs-positive patients had significantly lower levels of IL-5 (P = 0.009) and IL-17 (P = 0.044), while GM-CSF levels were higher (P = 0.002) compared to healthy controls.