Multi-omics characterises early immune divergence associated with infection-related deterioration after lung transplantation

Abstract

Background

Early adverse outcomes after lung transplantation often manifest as infection-associated clinical deterioration. Such events are closely intertwined with dysregulated perioperative immune reconstitution and are influenced by immunosuppressive therapy. A systematic characterisation of the perioperative immune landscape and potentially actionable pathways may facilitate understanding of early trajectory divergence and inform perioperative monitoring and management.

Methods

We conducted longitudinal sampling and immune profiling at predefined perioperative time points in lung transplant recipients. A deeply profiled discovery cohort was constructed using multi-omics approaches including single-cell transcriptomics, B-cell receptor sequencing, plasma proteomics, and flow cytometry, and key findings were externally validated in an independent validation cohort. Healthy donors provided reference baselines for peripheral immune states.

Results

Outcome-associated immune divergence was detectable early and became more apparent by post-operative Day 7 (T2). At baseline (T0), recipients who later died tended to show a higher burden of CD177⁺ neutrophils, which may reflect a subclinical inflammatory state and a propensity for NETs-related activity. By T2, the Death group more frequently exhibited persistent inflammatory and infection-related deterioration with features of maladaptive humoral reconstitution, including plasmablast/plasma cell expansion, naïve B-cell depletion, skewed BCR clonality with restricted somatic hypermutation, and heightened BAFF/NETs activity alongside IgA-biased responses. The Live group generally showed a more resolving inflammatory course with relative preservation of B-cell homeostasis.

Conclusions

These data suggest that CD177⁺ neutrophil–linked NETs–BAFF activity may contribute to early adverse trajectories after lung transplantation and may be informative for perioperative risk stratification. A composite assessment at T0 and T2 integrating CD177⁺ neutrophils, BAFF/NETs readouts, B-cell dynamics and immunoglobulin profiles may help identify recipients at risk of infection-related deterioration and nominate the first post-operative week as a potential window for carefully timed immune modulation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12931-026-03654-x.

Introduction

Lung transplantation (LTx) remains the only effective treatment for end-stage lung disease, yet early mortality is still largely driven by infections linked to dysregulated immune reconstitution [1, 2]. Surgical stress, alloimmune activation and immunosuppression can amplify innate inflammation while delaying adaptive recovery, yet perioperative immune trajectories across outcomes remain incompletely defined [3].

Neutrophils are rapidly responding innate effector cells and display marked heterogeneity and contribute to tissue injury through degranulation and the formation of neutrophil extracellular traps (NETs) [4]. Recent studies have identified CD177⁺ neutrophils as enriched in lung transplant recipients and associated with poor clinical outcomes [5]. CD177 promotes neutrophil migration, reactive oxygen species production, and NETs formation [6]. NETs-associated signals have also been implicated in promoting BAFF (B cell activating factor) release from myeloid cells and in shaping dysregulated humoral responses in inflammatory settings such as systemic lupus erythematosus [7]. These findings suggest a potential regulatory axis between neutrophils and B cells via NETs and BAFF, especially in the immunologically vulnerable setting of LTx.

While transplant immunology has traditionally emphasised T-cell–centred alloimmunity, increasing evidence supports a complementary contribution of B cells and antibody-mediated processes to allograft outcomes [8–10], particularly through effector subsets such as plasmablasts (PB) and plasma cells (PC), which can expand rapidly under inflammatory or infectious stress and, through antigen presentation and cytokine production, further modulate T-cell programmes [8]. IgA responses have been linked to inflammatory and infectious stress states, but their perioperative context and cellular origin after LTx remain unclear [11].

To dissect mechanisms linking immune dysregulation, infection, and mortality, we performed longitudinal multi-omics profiling of postoperative peripheral blood using single-cell transcriptomics, B cell receptor (BCR) sequencing, and proteomics. We longitudinally profiled recipients transplanted for end-stage pulmonary fibrosis and focused on post-operative Day 7 (T2), when inflammatory peaks and clinical deterioration often emerge. We identified outcome-associated expansion and hyperactivation of CD177⁺ neutrophils with heightened NETosis and BAFF programmes, and functional assays supported a CD177⁺ neutrophil–NETs/BAFF axis linked to plasmablast differentiation and IgA-skewed responses.

Collectively, our study identifies a neutrophil-NETs-BAFF axis that is associated with B cell hyperactivation and aberrant IgA responses following LTx. By linking these features to early adverse outcomes and peripheral readouts, our findings may inform perioperative immune monitoring and early identification of patients at risk of clinical deterioration, and nominate potentially tractable myeloid–humoral targets for future evaluation.

Methods

The complete materials and methods are included in the Supplementary Material.

Study design and participants

This prospective, multicentre observational study enrolled adults with end-stage pulmonary fibrosis undergoing LTx at the Eighth Medical Center of the Chinese PLA General Hospital and Wuxi People’s Hospital. Two independent cohorts of transplant recipients, together with healthy controls, were recruited and sampled longitudinally at perioperative time points: T0 (preoperative baseline), T1 (post-operative day 1), T2 (post-operative day 7), T3 (post-operative day 14), and T4 (approximately 2 months after transplantation). A total of 55 lung transplant recipients were included (44 in the Live group and 11 in the Death group). Baseline donor and recipient characteristics are summarized in Table 1. The study was conducted in accordance with the Declaration of Helsinki and its later amendments; ethics approvals and informed consent procedures are detailed in the Declarations. Detailed inclusion and exclusion criteria, clinical definitions, immunosuppression regimens, and sample processing procedures are provided in the Supplementary Methods. Sample-level assay availability is provided in Supplementary Tables 1–4.

Donor procurement and allocation

Donor lungs were obtained from voluntary deceased donors and allocated through the China Organ Transplant Response System (COTRS), in accordance with national regulations and the Declaration of Istanbul, with appropriate consent.

Single-cell RNA sequencing and B cell receptor analysis

Whole blood from lung transplant recipients and healthy controls was processed for single-cell profiling using the 10x Genomics Chromium 5′ platform and sequenced on an Illumina NovaSeq 6000. Gene-expression libraries were processed with Cell Ranger and Seurat using standard quality control and batch correction; cell types were annotated by canonical markers and differential expression. In parallel, single-cell V(D)J libraries were generated to obtain paired BCR sequences, which were processed with Cell Ranger vdj and analyzed using scRepertoire to define clonotypes, clonal expansion, diversity, somatic hypermutation, and isotype usage.

Plasma proteomics and immunoassays

Plasma samples were subjected to quantitative proteomics by data-independent acquisition (DIA) mass spectrometry. Briefly, high-abundance proteins were depleted prior to tryptic digestion, and peptides were analyzed by nanoLC–MS/MS on a high-resolution Orbitrap platform. DIA raw files were searched against the human UniProt reference proteome for protein identification and quantification, followed by across-sample normalization to enable longitudinal and between-group comparisons. Differential protein abundance, pathway enrichment, and module-based analyses were conducted as detailed in the Supplementary Methods. Plasma levels of BAFF, APRIL, NETs (MPO–DNA complexes), and immunoglobulins (IgA, IgG, and IgM) were quantified by ELISA according to the manufacturers’ instructions.

Flow cytometry

Multiparameter flow cytometry was performed on thawed peripheral blood mononuclear cells (PBMCs). B cell subsets were defined within the CD19⁺ lymphocyte gate as follows: naïve B cells (CD27⁻IgD⁺), unswitched memory B cells (CD27⁺IgD⁺), class-switched memory B cells (CD27⁺IgD⁻), double-negative (DN) B cells (CD27⁻IgD⁻), and plasmablasts (CD27⁺CD38^hi). Freshly isolated neutrophils were stained for CD45, CD66b, and CD177 to quantify CD177⁺ and CD177⁻ neutrophil subsets.

Neutrophil and B cell isolation and co-culture

Peripheral blood B cells from healthy donors were purified from PBMCs by negative selection, and neutrophils were isolated from whole blood using density-based separation. In vitro co-culture experiments were performed with purified B cells and CD177⁺ or CD177⁻ neutrophils, recombinant BAFF, or isolated NETs under defined conditions, followed by flow cytometric quantification of PB and other B cell subsets.

Immunofluorescence staining of NETs and lung tissue

NETs were visualized by immunofluorescence on purified neutrophils cultured on coverslips and stained for citrullinated histone H3, myeloperoxidase (MPO), and DNA. In explanted recipient lungs and normal donor lungs, formalin-fixed paraffin-embedded sections were subjected to multiplex tyramide signal amplification (TSA), based staining for CD177, CitH3, and CD19, followed by DAPI counterstaining and fluorescence imaging.

Statistical analysis

Statistical analyses were performed using R software, SPSS, and GraphPad Prism. Continuous variables were compared using parametric or nonparametric tests, as appropriate, and categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. Longitudinal continuous variables measured repeatedly from T0 to T4 were analyzed using linear mixed-effects models to account for within-subject correlation. A two-sided p value < 0.05 was considered statistically significant.

Results

Perioperative immune profiling highlights T2 as a major window of immune divergence after lung transplantation

To delineate temporal dynamics, we established a multi-omics longitudinal cohort of 55 lung transplant recipients and 14 healthy controls, integrating clinical parameters, plasma proteomics, single-cell transcriptomics, and BCR sequencing (Fig. 1A). Baseline donor and recipient characteristics were broadly comparable between groups and are summarised in Table 1. Clinical trajectories began to diverge early between groups (Fig. 1B). Importantly, multi-omics profiling was designed to capture immune perturbations that may precede or accompany early clinical deterioration. Compared with the Live group, the Death group showed an earlier rise in serum IL-6 beginning at T1, whereas differences in serum procalcitonin (PCT), neutrophil counts (NEU), C-reactive protein (CRP), became apparent at T2.

Fig. 1.

Perioperative immune profiling identifies T2 as a key transition point during immune reconstitution after LTx. A Study design. Fifty-five patients with advanced pulmonary fibrosis undergoing LTx (Death, n = 11; Live, n = 44) and 14 healthy controls were enrolled. A multi-omics cohort (n = 12; Death, n = 4; Live, n = 8) underwent longitudinal single-cell RNA sequencing (scRNA-seq), B cell receptor (BCR) sequencing, and plasma proteomics. The remaining patients (n = 43; Death, n = 7; Live, n = 36) formed a validation cohort for clinical and proteomic analyses. Peripheral blood was collected at four perioperative time points (T0-T3), with an additional late time point (T4) for proteomic profiling. B Line plots show temporal changes in clinical and inflammatory parameters in Death (yellow) and Live (blue) groups, including procalcitonin (PCT), neutrophil count (NEU), C-reactive protein (CRP), platelet count (PLT), D-dimer, absolute lymphocyte count (ALC), interleukin-6 (IL-6), and tacrolimus (FK506) concentrations. Asterisks indicate between-group differences (Death vs. Live) at the corresponding time point. C Line plots show dynamic proportions of CD3⁺ T cells, CD4⁺ T cells, CD8⁺ T cells, NK cells, B cells, and CD4/CD8 ratio across time points. Asterisks indicate between-group differences (Death vs. Live) at the corresponding time point. D Box plots compare serum IgA, IgG, and IgM levels between Death and Live groups at T2. Statistical analyses used one-way ANOVA with Tukey’ s multiple-comparison test unless indicated otherwise.*p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

Table 1.

Baseline characteristics and clinical outcomes of lung transplant recipients stratified by survival status

Variable	Total (n=55)	Live (n=44)	Death (n=11)	Statistic (χ2 / T /H value)	p value
Demographic of donors
Male, n (%)	44.0 (80.0)	35.0 (79.5)	9.0 (81.8)	0.028	0.866
Type, n (%)				<0.001	1.000
Donation after brain death	50.0 (90.9)	40.0 (90.9)	10.0 (90.9)
Donation after brain and cardiac death	5.0 (9.1)	4.0 (9.1)	1.0 (9.1)
Age (year), mean (SD)	37.8 (7.3)	38.6 (7.5)	34.5 (5.4)	1.730	0.090
Ventilator time (day), mean (SD)	4.7 (1.3)	4.6 (1.3)	4.8 (1.3)	-0.465	0.644
Oxygenation index (mmHg), mean (SD)	439.3 (27.0)	438.1 (25.7)	444.2 (32.9)	-0.660	0.512
BMI (kg/m²), mean (SD)	22.8 (0.7)	22.8 (0.7)	22.9 (0.5)	-0.458	0.649
Infection, n (%)	7.0 (12.7)	5.0 (11.4)	2.0 (18.2)	3.973	0.204#
Pseudomonas aeruginosa, n (%)	2.0 (3.6)	2.0 (4.5)	0.0 (0.0)
Acinetobacter baumannii, n (%)	3.0 (5.5)	1.0 (2.3)	2.0 (18.2)
Staphylococcus aureus, n (%)	2.0 (3.6)	2.0 (4.5)	0.0 (0.0)
ABO blood type/n (%)				1.033	0.855#
A	12 (21.8)	10 (22.7)	2 (18.2)
B	15 (27.3)	13 (29.5)	2 (18.2)
AB	5 (9.1)	4 (9.1)	1 (9.1)
O	23 (41.8)	17 (38.6)	6 (54.5)
Demographic of recipients
Male, n (%)	44.0 (80.0)	34.0 (77.3)	10.0 (90.9)	0.348	0.555
Age (year), mean (SD)	56.7 (10.8)	57.0 (10.8)	55.6 (11.4)	0.364	0.717
Primary lung disease, n (%)				1.341	0.511
IPF	8.0 (14.5)	7.0 (15.9)	1.0 (9.1)
CTD_ILD	13.0 (23.6)	9.0 (20.5)	4.0 (36.4)
Other ILD	34.0 (61.8)	28.0 (63.6)	6.0 (54.5)
History of smoking, n (%)	29.0 (52.7)	23.0 (52.3)	6.0 (54.5)	0.018	0.893
BMI (kg/m2), mean (SD)	22.9 (0.7)	22.9 (0.8)	22.6 (0.4)	1.994	0.056*
ABO blood type/n (%)				0.507	0.888#
A	13 (23.6)	10 (22.7)	3 (27.3)
B	18 (32.7)	14 (31.8)	4 (36.4)
AB	4 (7.3)	3 (6.8)	1 (9.1)
O	20 (36.4)	17 (38.6)	3 (27.3)
Recipient preoperative characteristics
ECMO bridging pretransplant, n (%)	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)		NA
Pretransplant Infection, n (%)				2.589	0.762#
Pseudomonas aeruginosa	1.0 (1.8)	1.0 (2.3)	0.0 (0.0)
Acinetobacter baumannii	5.0 (9.1)	4.0 (9.1)	1.0 (9.1)
Stenotrophomonas maltophilia	2.0 (3.6)	1.0 (2.3)	1.0 (9.1)
Escherichia coli	4.0 (7.3)	3.0 (6.8)	1.0 (9.1)
Procedure type, n (%)				1.289	0.256
BLT	40.0 (72.7)	30.0 (68.2)	10.0 (90.9)
SLT	15.0 (27.3)	14.0 (31.8)	1.0 (9.1)
Routine blood tests
WBC (10⁹/L), mean (SD)	8.9 (3.2)	8.8 (2.9)	9.5 (4.5)	-0.472	0.645*
ALC (10⁹/L), mean (SD)	1.4 (0.9)	1.5 (0.9)	0.9 (0.6)	1.909	0.062
AMC (10⁹/L), mean (SD)	0.6 (0.3)	0.6 (0.3)	0.5 (0.3)	1.310	0.196
NEU (10⁹/L), mean (SD)	7.1 (3.8)	6.9 (3.7)	7.8 (4.5)	-0.655	0.516
CRP (mg/L), mean (SD)	13.4 (24.9)	12.2 (27.0)	18.3 (13.0)	-0.724	0.472
PCT (U/L), mean (SD)	0.05 (0.3)	0.06 (0.3)	0.01 (0.0)	0.521	0.604
Comorbidities
Hypertension, n (%)	8.0 (14.5)	8.0 (18.2)	0.0 (0.0)	1.106	0.293
Cardiovascular disease, n (%)	5.0 (9.1)	3.0 (6.8)	2.0 (18.2)	0.344	0.558
Diabetes, n (%)	5.0 (9.1)	5.0 (11.4)	0.0 (0.0)	0.344	0.558
Drugs before LT
Antibiotic, n (%)	9.0 (16.4)	8.0 (18.2)	1.0 (9.1)	0.075	0.785
Glucocorticoid, n (%)	53.0 (96.4)	42.0 (95.5)	11.0 (100)		1.000#
Antifibrotic drug, n (%)	51.0 (92.7)	42.0 (95.5)	9.0 (81.8)		0.175#
Postoperative characteristics and outcomes
Duration of mechanical ventilation, mean (SD)	2.9 (1.0)	2.8 (1.0)	3.3 (1.2)	-1.268	0.210
ICU days, mean (SD)	8.0 (7.3)	6.5 (3.6)	14.3 (13.3)	-1.932	0.081*
Time to event, mean (SD)	50.3 (19.4)	58.8 (8.1)	16.3 (12.3)	13.881	<0.001
Early postoperative infection, n (%)	17.0 (30.9)	7.0 (15.9)	10.0 (90.9)		<0.001#
Stenotrophomonas maltophilia	4.0 (7.3)	2.0 (4.5)	2.0 (18.2)
Acinetobacter baumannii	10.0 (18.2)	3.0 (6.8)	7.0 (63.6)
Filamentous fungi	1.0 (1.8)	1.0 (2.3)	0.0 (0.0)
Candida albicans	1.0 (1.8)	1.0 (2.3)	0.0 (0.0)
Staphylococcus aureus	1.0 (1.8)	0.0 (0.0)	1.0 (9.1)
Time to first infection(day), mean (SD)	5.6 (1.7)	6.4 (2.0)	5.0 (1.3)	1.782	0.095
FK506  concentration (ng/mL), mean (SD)	9.5 (3.1)	9.6 (3.1)	9.2 (3.0)	0.419	0.677

Abbreviations: ALC absolute lymphocyte count, AMC absolute monocyte count, BLT bilateral lung transplantation, CRP C-reactive protein, CTD-ILD connective tissue disease–associated interstitial lung disease, ECMO extracorporeal membrane oxygenation, ILD interstitial lung disease, IPF idiopathic pulmonary fibrosis, LTx lung transplantation, NEU neutrophil count, PCT procalcitonin, SLT single lung transplantation, WBC white blood cell count

*χ², T, H: test statistics for chi-square, independent t-test, Kruskal-Wallis tests, respectively. #: p from Fisher’s exact test; *: p from Kruskal-Wallis test. Other p values from chi-square or t-test as appropriate

Peripheral immune cell composition further illuminated the immunological basis of these differences (Fig. 1C). The Death group displayed an early and persistent reduction in CD3⁺ total T cells and CD8⁺ T cells, together with a declining CD4/CD8 ratio, without compensatory expansion of natural killer (NK) cells. In contrast, the proportion of B cells remained consistently higher in the Death group throughout the perioperative period and peaked at T2, indicating progressive B cell expansion against a background of impaired T cell recovery.

Serum immunoglobulin profiling showed globally elevated IgA levels in the Death group compared with the Live group, whereas IgG and IgM levels were largely comparable between groups (Fig. 1D), indicating an IgA-dominant humoral response in patients with poor outcomes. Thus, T2 corresponds to a clinically relevant window in which sustained inflammation, impaired T cell reconstitution, and B cell/IgA expansion coexist, providing a key observation window for identifying immune dysregulation and an increased risk of early complications.

Divergent clinical outcomes are shaped by distinct trajectories of immune reconstitution

To define the cellular basis of the T2 transition, we performed single-cell RNA sequencing on 45 peripheral blood samples from 12 lung transplant recipients (T0, T1, T2 and T3) and 5 healthy controls. Using canonical marker genes, we identified six major immune lineages (T cells, B cells, monocytes, neutrophils, erythroid cells and megakaryocytes) together with their principal subclusters (Fig. 2A, B; Fig. S1A-C).

Fig. 2.

Longitudinal comparison of clinical parameters and immune cell composition between death and live groups across perioperative time points of LTx. A UMAP visualization of major immune populations identified by scRNA-seq, including T, B, monocyte, neutrophil, erythroid, and megakaryocyte lineages. B UMAP visualization of canonical marker gene expression across all single-cell transcriptomes. C Stacked bar plot illustrating the relative abundance of major immune compartments across time points. D Principal component analysis (PCA) of scRNA-seq data showing sample distribution based on global transcriptomic profiles across groups and time points. Each point represents one sample, and ellipses indicate 95% confidence intervals. E Heatmap of Z-scored immune subset frequencies across all groups and time points. F Correlation heatmap integrating immune cell frequencies and clinical markers. Red-labeled variables represent immune cell frequencies derived from scRNA-seq data, while others indicate clinical or inflammatory parameters from laboratory tests. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

Over time, the proportion of B cells rose sharply at T2, whereas T cells had not yet recovered (Fig. 2C; Fig. S1D). Dimensionality reduction and proportion-based clustering demonstrated persistent separation of the Death group from the Live group/controls, with early enrichment of B cells/PC and immature neutrophils and sustained depletion of T/NK compartments in the Death group (Fig. 2D, E). Supplementary time-cluster analysis showed that distinct gene clusters exhibited different temporal expression patterns, with T2 mainly enriched for pathways related to leukocyte activation and immune regulation (Fig. S1E, F). Immune composition correlated with clinical inflammation and immunosuppression indices, linking B-cell/neutrophil predominance to inflammatory burden and impaired T cell recovery (Fig. 2F).

Proteomic principal component analysis and pathway enrichment were in line with these findings. At T2, samples were most enriched for pathways related to inflammation, granulocyte activation and acute-phase responses. Thereafter, the Live group shifted toward complement activation, humoral immunity, lymphocyte activation, and tissue repair, whereas the Death group retained strong inflammatory and pro-apoptotic signaling (Fig. S2A, B). Consistently, the Death group displayed an IgA-biased humoral profile, with upregulation of IgA-related proteins enriched in complement activation and B cell–mediated immunity (Fig. S2C, D).

Collectively, our multi-omics data identify T2 as a key transition point from innate immune hyperactivation to adaptive immune reconstitution.

T cell regulatory impairment is associated with failure of adaptive immune reconstitution

Given the central role of T cells in post-transplant immune reconstitution and their marked reduction in the Death group, we comprehensively analyzed T-cell subsets and functional states.

UMAP clustering resolved ten T/NK subsets (Fig. 3A, B; S3A). Compared with the Live group, the Death group showed a longitudinal shift toward CD4⁺ effector memory/Tregs and CD8⁺ naïve/cytotoxic cells, whereas the Live group were enriched for CD4⁺ naïve and CD8⁺ effector (GZMK⁺/GZMB⁺) populations, consistent with impaired effector differentiation in the Death group (Fig. 3C, D).

Fig. 3.

Single-cell characterization and longitudinal profiling of T cell subsets across perioperative time points of LTx. A UMAP visualization of T and NK cell subsets identified from scRNA-seq data. B Dot plot showing representative marker gene expression for each subset. Dot size represents the percentage of expressing cells, and color intensity indicates the average expression level. C UMAPs colored by time points (top) and clinical groups (bottom), with stacked bar plots on the right showing the relative proportions of each subset across time points and groups. D Line plots showing temporal changes in the frequencies of individual T/NK cell subsets across the control, death, and live groups. E Violin plots showing the distribution of Tfh cell–related scores across groups and time points. F Violin plots illustrating longitudinal changes in plasma protein levels of CXCL13 and ICOSLG measured by proteomic profiling across groups and time points. G Violin plots showing normalized expression levels of representative Tfh-related genes (CD40LG, BCL6, IL21R, and CXCR5) across groups and time points. Statistical significance was determined using Wilcoxon tests: *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

Given that CD4⁺ naïve T cells are the origin of both effector and Tfh lineages, we focused on their transcriptional and functional dynamics. Module scoring based on canonical Tfh-associated genes revealed that the control group had the highest scores overall, which declined post-transplantation; notably, the Live group exhibited a partial recovery at T2, whereas the Death group remained persistently low, with the most significant intergroup difference observed at T2 (Fig. 3E). These findings suggest that the recovery capacity of the CD4⁺ naïve pool is closely associated with the immune outcome after transplantation.

At the molecular level, differential gene analysis demonstrated upregulation of S100A8/A9/A12, ACTB, and BANF1—genes linked to inflammation and cellular stress—in the Death group, whereas the live group showed enrichment of TNFAIP3, RASA3, and SSH2, which are involved in immune tolerance and homeostatic regulation (Fig. S3B). Pathway analysis further revealed functional divergence between groups, with the Live group showing enrichment of T-cell activation and immune response pathways, whereas the Death group was characterized by stress-, apoptosis-, and inflammatory response-related programs over time (Fig. S3C, D). In the Tfh-related signaling axis, proteomic profiling showed that CXCL13 was elevated at T2 only in the Live group (Fig. 3F), while single-cell transcriptomic analysis demonstrated significantly reduced expression of CXCR5, BCL6, IL21R, and CD40LG in the Death group (Fig. 3G). Together, these findings indicate compromised Tfh differentiation and B cell helper function in the Death group.

CellChat analysis further indicated reduced interaction number and strength between T and B cells in the Death group, whereas the Live group retained key regulatory signals, linking CD4⁺ naïve T-cell dysfunction with weakened Tfh differentiation and impaired T–B communication (Fig. S3E–F). Collectively, these data suggest that quantitative loss and functional perturbation of CD4⁺ naïve T cells are associated with compromised Tfh differentiation and reduced T–B communication, representing a potential regulatory gap during early immune reconstitution.

B cell subset remodeling and functional imbalance define divergent post-transplant immune outcomes

To further delineate the cellular basis of impaired immune reconstitution, we focused on the dynamic evolution of B cell subsets. Single-cell transcriptomic analysis identified 10 B cell populations, including TCL1A⁺, FYB1⁺ and FAM129C⁺ naïve B cells (Bn), unswitched and class-switched memory B cells (Bm), transitional PB, MKI67⁺ proliferating PB, and IGHM⁺, IGHA⁺ and IGHG⁺ PC (Fig. 4A, B; Fig. S4A). Longitudinal analysis showed that the B cell compartment was markedly remodeled after transplantation. On T2, the Death group displayed a pronounced expansion of B cells, with a striking increase in PB and PC, whereas the Live group showed a gradual normalization and a more balanced distribution of B cell subsets (Fig. 4C, D; Fig. S4B, C). These patterns suggest that B cells play a central regulatory role in perioperative immune reconstitution.

Fig. 4.

Single-cell characterization and longitudinal profiling of B cell subsets after LTx. A UMAP visualization of B cell subsets identified from scRNA-seq data. B Dot plot displaying representative marker gene expression of B cell subsets. Dot size indicates the percentage of expressing cells, and color intensity represents average expression levels. C UMAP plots colored by time points (top) and clinical groups (bottom), with stacked bar plots on the right showing the relative proportions of each subset across time points and groups. D Line plots illustrating temporal changes in the proportions of B cell subsets across control, death, and live groups. E Heatmaps displaying Z-scored module scores of B cell functional signatures in the Death and Live groups. F Representative gating strategy in peripheral blood mononuclear cells (PBMCs) to identify CD19⁺ B cells and the following B-cell subsets: plasmablasts (PB; CD27⁺CD38hi), naïve B cells (Naive B; CD27⁻IgD⁺), unswitched memory B cells (CD27⁺IgD⁺), class-switched memory B cells (CD27⁺IgD⁻), and double-negative B cells (DN B; CD27⁻IgD⁻). G Longitudinal changes in the proportions of CD19⁺ B cells (percentage of PBMCs), naïve B, PB, unswitched mem, class-switched mem, and DN B (all as percentages of CD19⁺ B cells) from T0 to T3 in the Death group (yellow) and the Live group (blue). Each point represents an individual patient, with lines connecting repeated measurements from the same patient. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

To connect compositional shifts with functional states, module scoring revealed distinct activation states of B cell subsets in the Live group and the Death group (Fig. 4E). Module scoring indicated that, in the Death group, naïve/memory B cells were skewed towards interferon/stress-associated programmes, while PB/PC showed sustained BCR/antibody-secretion activity at T2. The Live group showed a more resolving trajectory with progressive engagement of activation/differentiation modules and restoration of homeostasis.

Building on this, we next examined the molecular features of each B cell subset at T2 (Fig. S4B; Fig. S5). Among naïve B cells, memory B cells, PB and PC, the largest number of differentially expressed genes (DEGs) and the most pronounced transcriptional changes were observed in PB and PC (Fig. S4B). DEG and enrichment analyses suggested reduced activation/antigen-presentation programmes in naïve/memory B cells and a stress-associated PB/PC state in the Death group (Fig. S4D; Fig. S5A-F). These data indicate that B cells in the Death group adopt a dysregulated configuration characterized by weakened precursor programs and overactivated effector cells under high stress, failing to complete the transition from an inflammatory state to stable adaptive immune reconstitution.

To validate the abnormal B cell remodeling observed at single-cell resolution, we performed multicolor flow cytometry on perioperative peripheral blood from all recipients (Fig. 4F). Consistent with the transcriptomic data, the Death group showed a higher overall proportion of CD19⁺ B cells early after transplantation, with a marked increase in PB and double-negative (DN) B cells (CD27⁻IgD⁻) and a relative reduction in naïve B cells at T2. In contrast, the Live group displayed a gradual restoration of B cell subset distribution over time (Fig. 4G). Together, these findings support B cell lineage remodeling as a key feature of perioperative immune imbalance.

Distinct BCR clonal-expansion and somatic hypermutation patterns define divergent immune reconstitution trajectories

To characterize B cell immune reconstitution at the molecular level, we analyzed BCR repertoires across time points and outcome groups. UMAP projections revealed marked heterogeneity in the clonal distribution of immunoglobulin isotypes, with IGHM, IGHA and IGHG each enriched in distinct B cell clusters (Fig. 5A). Over time, the overall proportion of IGHM gradually decreased, whereas IGHA and IGHG increased specifically in the Live group, indicating more complete class-switch recombination in this group (Fig. 5B).

Fig. 5.

B cell receptor (BCR) repertoire analysis reveals distinct clonal expansion and mutation profiles between clinical outcomes. A UMAP visualization of BCR clonotypes colored by immunoglobulin isotypes. B Stacked bar plots showing the proportional distribution of immunoglobulin isotypes across time points (left), across clinical groups including Control, Death, and Live (middle), and between the Death and Live groups at T2 (right).C UMAP visualization of BCR clonotypes classified by clone size: hyperexpanded (30 < X ≤ 100), large (10 < X ≤ 30), medium (3 < X ≤ 10), small (1 < X ≤ 3), and single (0 < X ≤ 1). D Stacked bar plots displaying the distribution of clonal expansion categories across time points (left), across clinical groups including Control, Death, and Live (middle), and between the Death and Live groups at T2 (right). E Box plot comparing the overall mutation frequency among the Control, Death, and Live groups. F Sankey diagram illustrating the relationship between clonal expansion (left), B cell subtypes (middle), and clinical groups (right). G–H Box plots showing mutation frequencies grouped by B cell subtype (G) and by immunoglobulin isotype (H). I Comparison of mutation frequencies for each immunoglobulin isotype between the Death and Live groups at T2. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

Analysis of clonal structure and somatic hypermutation (SHM) further highlighted these differences. The Death group harbored a higher fraction of large and hyperexpanded clones, while the Live group showed a more even clone-size distribution (Fig. 5C, D). Overall SHM frequencies in the Death group were lower than in healthy controls and the Live group (Fig. 5E). Sankey plots demonstrated that hyperexpanded clones in the Death group were largely concentrated within PC, whereas the Live group retained clonotypes derived from naïve B cells, memory B cells and PB (Fig. 5F). Stratified analyses indicated persistently low SHM in the Death group across precursor subsets, with expanded PB/PC exhibiting limited mutation burden and diversity. At the isotype level, IGHA and IGHG clones in the Live group carried higher SHM, whereas the Death group were dominated by low-mutated IGHM and IGHD clones (Fig. 5G-I), consistent with incomplete class switching and affinity maturation.

Extended repertoire analyses (Fig. S6) were concordant with these findings. Across additional repertoire metrics, the Death group showed reduced diversity and limited accumulation of SHM over time, consistent with impaired affinity maturation and a bias towards rapid extrafollicular activation (Fig. S6A-F). Taken together, patients with poor outcomes display restricted BCR diversity, insufficient SHM and aberrant PC-dominated clonal expansion, whereas the Live group achieve a more complete humoral reconstitution through ongoing clonal diversification and accumulation of somatic mutations.

Neutrophil-driven BAFF secretion and NETosis activation contribute to a hyperinflammatory milieu associated with poor outcomes

The above findings suggest that aberrant B-cell remodeling may be driven, at least in part, by upstream innate immune signals. On this basis, we next shifted our analytical focus to myeloid cells, particularly neutrophils, to explore their potential regulatory role. Neutrophils exhibited the greatest transcriptional heterogeneity and were clustered into six subpopulations, including CD177⁺, CXCR4⁺, IFN⁺, NFKBIZ⁺, CD74⁺ and immature neutrophils (Fig. 6A,B ; Fig. S7A). In the Death group, CD177⁺ neutrophils peaked at T2 and represented the most divergent subset between outcome groups, whereas NFKBIZ⁺ and immature neutrophils were also enriched overall. Notably, the CD177⁺ subset was already shifted at enrolment (T0), preceding its marked T2 expansion and functional hyperactivation. By contrast, CXCR4⁺, IFN⁺ and CD74⁺ subsets remained relatively stable in the Live group (Fig. 6C, D).

Fig. 6.

Single-cell characterization and longitudinal profiling of neutrophil subsets across perioperative time points of LTx. A UMAP visualization of neutrophil subsets identified from scRNA-seq data. B Dot plot displaying representative marker gene expression of neutrophil subsets. Dot size indicates the percentage of expressing cells, and color intensity represents average expression levels. C UMAP plots colored by time points (top) and clinical groups (bottom), with stacked bar plots on the right showing the relative proportions of each subset across time points and groups. D Line plots illustrating temporal changes in the proportions of neutrophil subsets across groups. E Violin plots showing UCell scores of NETosis-related gene sets among the six neutrophil subsets. F Violin plots displaying NETosis scores across different time points and groups. G Violin plots showing normalized expression levels of BAFF-axis ligands TNFSF13B and TNFSF13 in neutrophils across groups and time points. H Violin plots showing the expression levels of TNFSF13B and TNFSF13 among individual neutrophil subsets. I CellChat chord diagrams showing the inferred BAFF signalling network at T2 in the Death and Live group. Chord thickness reflects inferred communication strength (communication probability) between sender and receiver cell populations, highlighting strengthened neutrophil-to-B-cell connectivity in the Death group. J CellChat signalling-role heatmaps for the BAFF pathway at T2, illustrating the relative importance of each cell population as sender, receiver, mediator, or influencer within the BAFF signalling network in the Death and Live group. Colour intensity denotes relative importance (higher values indicate greater contribution). *p< 0.05, **p< 0.01, ***p < 0.001; ns, not significant

Functional profiling showed that CD177⁺ neutrophils had the highest NETosis scores and were enriched for oxidative and degranulation genes such as MPO, ELANE and PRTN3 (Fig. 6E,F), indicating a highly activated inflammatory state. TNFSF13B, encoding BAFF, was markedly upregulated in this subset (Fig. 6G, H), and its expression correlated positively with NETosis scores within CD177⁺ cells (Fig. S7B). Differential expression and pathway analyses indicated that CD177⁺ neutrophils from the Death group upregulated stress- and metabolism-related genes and pathways linked to oxidative phosphorylation, antigen processing, and cytokine-mediated inflammation, whereas those from the Live group showed relative enrichment of molecules associated with immune regulation and resolution (Fig. S7C-E). To map the interaction landscape linking neutrophil activation and B-cell effector skewing at T2, we conducted CellChat analysis on T2 samples. BAFF pathway chord plots indicated stronger neutrophil–B-cell communication in the Death group, with CD177⁺ neutrophils prioritised as major BAFF signal senders and B-cell subsets as key receivers (Fig. 6I, J). Ligand–receptor analysis further highlighted increased TNFSF13B (BAFF) signalling to its canonical receptors (BCMA, TACI and BAFF-R) within neutrophil–B-cell interactions in the Death group (Fig. S7F, G). These observations support an immune landscape in which BAFF-related neutrophil–B-cell communication is accentuated during infection-associated deterioration. Multi-omics integration supported a neutrophil-BAFF-NETosis axis at the systems level. Proteomic profiling showed that, at T2, the Death group had significantly higher levels of BAFF-related proteins (TNFSF13 and TNFSF13B) and NETosis-associated factors (ELANE, MPO, MMP9, LTF, S100A8/A9 and PRTN3), whereas the Live group gradually returned toward baseline (Fig. S8A). In parallel, plasma IgA1 and IgA2 levels remained elevated in the Death group, consistent with PB/PC overexpansion and linking excessive BAFF signaling to amplified humoral responses (Fig. S8B).

Because BAFF and APRIL are largely produced by myeloid cells, we further examined monocyte dynamics. UMAP re-clustering separated monocytes into intermediate, CD14⁺ classical, and CD16⁺ non-classical subsets (Fig. S9A–C). Across myeloid lineages, neutrophils almost exclusively expressed TNFSF13B with minimal TNFSF13, indicating neutrophils are the dominant source of BAFF, whereas APRIL was minimal (Fig. S9D). Within monocytes, TNFSF13B expression was generally higher in the Live group, whereas at the critical T2 time point the TNFSF13B peak localized predominantly to neutrophils (Fig. S9E, F), consistent with an abnormally amplified neutrophil–BAFF–IgA cascade in patients with poor outcomes.

Together, these data support a model in which CD177⁺ neutrophils constitute the core of an innate-humoral imbalance. In the Death group, CD177⁺ neutrophils at T2 adopt a high-NETosis, high-BAFF inflammatory phenotype that is associated with excessive B cell activation and IgA overproduction through combined BAFF signaling and NETs release. This myeloid-driven inflammatory cascade may represent an important mechanism underlying failed immune reconstitution and early adverse outcomes after LTx.

CD177⁺ neutrophils contribute to pathological B-cell activation through NETs and BAFF

Based on the marked enrichment of CD177⁺ neutrophils in the Death group, we first compared their frequencies across outcome groups. Flow cytometry showed that the proportion of CD177⁺ neutrophils in peripheral blood was significantly higher in lung transplant recipients than in healthy controls, with the greatest increase observed in the Death group (Fig. 7A, B). To assess NETs-forming capacity, we sorted CD177⁺ and CD177⁻ neutrophils for immunofluorescence and functional assays. Compared with unstimulated controls, PMA robustly induced NETs release, and CD177⁺ neutrophils generated markedly more extensive spontaneous CitH3⁺/MPO⁺ extracellular web-like structures than CD177⁻ cells (Fig. 7C, D). Immunofluorescence further showed that MPO signals closely co-localized with extracellular DNA, a hallmark of NETs formation. Together with CitH3 positivity, these findings support active NETs release and an intrinsically higher NETosis propensity in CD177⁺ neutrophils under identical conditions.

Fig. 7.

Phenotypic and functional characterization of CD177⁺ neutrophils. A Representative flow-cytometry gating strategy for identifying CD177⁻ and CD177⁺ neutrophil (NEU) subsets from peripheral blood. Cells were sequentially gated as total leukocytes by FSC/SSC, live cells, singlets, and neutrophils defined as CD45⁺CD66b⁺ events, followed by stratification into CD177⁻ and CD177⁺ subsets based on CD177 expression. The same gating strategy was applied to groups. B Proportion of CD177⁺ NEU among total NEU in the three groups. C Quantification of neutrophil extracellular trap formation (NETs area %) in unstimulated neutrophils (Control), PMA-stimulated bulk neutrophils (PMA), sorted CD177⁻ NEU (CD177⁻), and sorted CD177⁺ NEU (CD177⁺). D Corresponding immunofluorescence images with DAPI staining for nuclei, MPO (myeloperoxidase), and CitH3 (citrullinated histone H3) highlighting NETs structures. E Levels of MPO–DNA complexes (MPO–DNA OD450) in culture supernatants under the indicated conditions. F–G Concentrations of BAFF (B cell activating factor, TNFSF13B) (F) and APRIL (a proliferation-inducing ligand, TNFSF13) (G) in neutrophil supernatants. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

Consistently, culture supernatants from CD177⁺ neutrophils contained significantly higher levels of MPO-DNA complexes and BAFF, whereas APRIL levels were similar across groups (Fig. 7E-G), indicating that CD177⁺ neutrophils are a major perioperative source of NETs and BAFF.

To obtain in situ evidence supporting the CD177⁺ neutrophil–NETs–B-cell axis, we performed multiplex TSA-based immunofluorescence on explanted lungs from recipients with severe early infection or death and on normal donor lungs (n = 3 per group). In donor lungs, CitH3 and CD19 signals were minimal with only sparse CD177⁺ neutrophils. In contrast, lungs from high-risk recipients exhibited prominent accumulation of CD177⁺ neutrophils accompanied by CitH3 positivity, with CD19⁺ B cells enriched in close proximity and partially overlapping with these CD177⁺CitH3⁺ regions (Fig. 8A), consistent with a pathological NETs-associated neutrophil–B-cell niche within the transplanted lung.

Fig. 8.

CD177⁺ neutrophil–derived NETs and BAFF drive PB differentiation and IgA production. A Immunofluorescence (IF) validation of CitH3⁺CD177⁺ neutrophils in close proximity to CD19⁺ B cells in explanted lung tissue from lung transplant recipients with severe early infection/death versus normal donor lungs. Scale bars = 100 μm (magnified image). B Schematic overview of the in vitro coculture system. B cells from healthy donors were cultured alone (B only), with recombinant BAFF, with purified NETs, or cocultured with sorted CD177⁻ or CD177⁺ neutrophils. In some conditions, DNase or TACI-Fc was added to degrade NETs or block BAFF signaling, respectively. C IgA concentrations in B-cell culture supernatants under the indicated conditions. D Representative flow-cytometry gating strategy for CD19⁺ B cells, showing identification of naïve B cells, unswitched memory B cells, class-switched memory B cells , double-negative B cells, and PB. E Frequencies of PB (percentage of CD19⁺ B cells) across the different stimulation conditions. F BAFF concentrations in neutrophil supernatants under control (Control), PMA-stimulated (PMA), and NETs-treated (NETs) conditions. G–H Plasma levels of BAFF (G) and MPO–DNA complexes (H) at T2 in groups. I–J Scatter plots showing the correlation between plasma BAFF and MPO–DNA complex levels at T2 in the Death group (I, n = 11) and the Live group (J, n = 14). Each dot represents one individual. r and p values indicate Spearman’s correlation coefficient and corresponding significance level. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant

These observations prompted us to test whether CD177⁺ neutrophils contribute to PB differentiation and an IgA-skewed response via NETs and BAFF. In an in vitro coculture system, exogenous BAFF, purified NETs, and CD177⁺ neutrophils each increased IgA secretion, with CD177⁺ neutrophils exerting the most pronounced effect; CD177⁻ neutrophils were less potent, and DNase or the BAFF decoy receptor TACI-Fc partially attenuated IgA production (Fig. 8B, C). Flow cytometry confirmed that these stimuli promoted PB differentiation, particularly under CD177⁺ neutrophil or NETs conditions, whereas blockade of NETs or BAFF signaling mitigated this effect (Fig. 8D, E). At the neutrophil level, PMA and NETs stimulation enhanced BAFF release, and NETs further amplified BAFF secretion, consistent with a self-reinforcing NETs–BAFF loop (Fig. 8F). In patient plasma at T2, BAFF and MPO–DNA levels were higher in the Death group than in the Live group and healthy controls (Fig. 8G, H), and BAFF correlated positively with MPO–DNA only in the Death group (Fig. 8I, J). Together, these data support a model in which CD177⁺ neutrophils, through coordinated NETs and BAFF production, are associated with pathogenic B-cell activation tightly linked to poor clinical outcome.

Discussion

LTx triggers a complex process of immune reconstitution, in which early infection-related mortality remains a major challenge. Although the mechanisms of graft tolerance and rejection have been extensively investigated [12–15], the longitudinal immune dynamics under standard immunosuppression remain poorly defined, particularly in B cells [16]. In this study, we integrated single-cell transcriptomics, BCR sequencing, proteomics and flow cytometry to construct a perioperative immune atlas of LTx. Collectively, our data highlight T2 as the time window when outcome-related immune differences are most pronounced, and suggest that myeloid activity may contribute to maladaptive humoral reconstitution. In parallel, we observed incomplete adaptive recovery characterised by sustained T/NK depletion and impaired CD4⁺ naïve/Tfh-associated programmes. Together, these findings suggest that around T2, delayed cellular recovery coincides with exaggerated humoral activation.

To guide interpretation of the B-cell results, we focused on an outcome-linked pattern of effector expansion coupled with limited repertoire maturation. We observed that the Death group tended to show aberrant B cell activation, dysregulated humoral responses and cytokine disturbance. Consistent across single-cell and flow cytometric analyses, these patients showed an expansion of PB and PC. In parallel, BCR repertoire profiling revealed clonal skewing, restricted SHM and reduced repertoire diversity, features consistent with rapid but low-quality extrafollicular antibody responses. Similar dysregulated B-cell programmes have been described across organ transplantation [17, 18], as well as in other severe inflammatory conditions. In severe COVID-19 and active SLE, extrafollicular B-cell responses are associated with plasmablast expansion and reduced maturation, supporting the interpretation that the lower observed in the Death group may reflect not only a shorter clinical course, but also a rapid, low-quality humoral response skewed away from full germinal-center maturation [19–21].

Accumulating evidence indicates that part of this B cell imbalance may arise from strong regulation by myeloid cells, especially neutrophils. This axis may be particularly prominent after lung transplantation because the graft is uniquely exposed to an open airway mucosal interface and early ischaemia–reperfusion/epithelial injury, together providing strong innate cues that favour neutrophil recruitment and NETosis [22]. Neutrophil-derived BAFF can influence B-cell survival and differentiation. Xu et al. reported that activated neutrophils release BAFF while forming NETs, thereby augmenting B cell responses [23]. Our in vitro data suggest that NETs may feed back on neutrophils to enhance BAFF release and contribute to an amplified inflammation–humoral loop; DNase partially attenuated downstream B-cell activation, supporting a role for NETs structural integrity, although the relative contributions of the DNA scaffold versus NETs-associated granular proteins remain unclear. However, the lung allograft microenvironment is more complex than our co-culture system, with epithelial injury–repair programmes and vascular niche cues shaping immune responses in vivo [24, 25]. Therefore, these experiments should be interpreted as supportive mechanistic evidence rather than proof that neutrophils unidirectionally drive B-cell activation in vivo. In this broader perioperative context, tissue injury can create an oxidative stress milieu that amplifies inflammatory circuits, providing an additional context in which neutrophil activation and downstream humoral skewing may be reinforced [26]. Consistent with this feed-forward model, NETosis has been linked to BAFF upregulation and pathogenic humoral activation in other inflammatory settings [27, 28]. Together, these observations are consistent with a framework in which NETs-associated inflammation can amplify BAFF-linked B-cell activation and differentiation [29, 30] and are compatible with reported correlations between IgA and NETosis biomarkers [31].

We collectively propose that perioperative imbalance in neutrophil–B-cell crosstalk represents a key immunological node associated with early immune reconstitution after lung transplantation, with the NETs–BAFF–IgA axis potentially operating in a feed-forward manner. CD177⁺ neutrophils represent an activated subset linked to NETs formation and lung ischaemia–reperfusion injury and have been proposed as an early biomarker for severe PGD after lung transplantation [5]. Consistent with prior work, we observed baseline and perioperative expansion of CD177⁺ neutrophils in the Death group, linking heightened NETosis/BAFF programmes to maladaptive humoral reconstitution. Notably, elevated CD177⁺ neutrophil abundance was already evident at T0 in the Death group, suggesting that pre-transplant clinical heterogeneity may contribute to baseline immune variation. One plausible contributor is occult pre-transplant infection or airway microbial colonisation that primes neutrophil activation before surgery. On this basis, we speculate that perioperative joint monitoring of CD177⁺ neutrophil burden, NET/BAFF-associated readouts, and peripheral B-cell/plasmablast dynamics together with antibody levels (including IgA-skewed responses) may help identify recipients at heightened risk of infection-related deterioration [32]. Future studies should test whether selectively reducing NET generation and/or limiting BAFF induction can rebalance extrafollicular B-cell activation without broadly suppressing protective humoral immunity. Given infection vulnerability in transplant recipients, any immunomodulation targeting this axis should be carefully evaluated and timed, potentially around the T2 window.

This study has several limitations. First, we restricted enrolment to recipients with advanced pulmonary fibrosis undergoing lung transplantation; therefore, these findings may not fully generalise to other indications (e.g., cystic fibrosis or COPD). Second, the non-survivor group was relatively small (n = 11), which may limit statistical power and generalisability. Third, the Death group more often required intensive post-operative organ support, which may confound immune phenotypes and reflects severe physiological stress that could contribute to the observed trajectory. Finally, perioperative infection status, antimicrobial exposure, and immunosuppressive dosing may also act as confounders despite standardised clinical management. Moreover, perioperative DSA monitoring (including IgG/IgA isotypes) and biopsy-based AMR assessment (e.g., C4d staining) were not performed systematically in this cohort, which precludes definitive exclusion of subclinical or early AMR as a contributor to the observed plasmablast/plasma-cell expansion. Larger, prospective studies with in vivo mechanistic testing will be required to validate the prognostic utility of the proposed markers and to assess the safety and efficacy of interventions targeting the NETs–BAFF axis.

Abbreviations

ALC

Absolute lymphocyte count

AMR

Antibody-mediated rejection

ACR

Acute cellular rejection

APRIL

A proliferation-inducing ligand (TNFSF13)

BAFF

B cell activating factor (TNFSF13B)

BCMA

B cell maturation antigen (TNFRSF17)

BCR

B cell receptor

Bm

Memory B cells

Bn

Naïve B cells

CDR3

Complementarity determining region 3

CitH3

Citrullinated histone H3

COTRS

China Organ Transplant Response System

CRP

C-reactive protein

CSR

Class-switch recombination

D-dimer

D-dimer

DIA

Data-independent acquisition

DN B

Double-negative B cells

DNase

Deoxyribonuclease

DSA

Donor-specific antibodies

ELISA

Enzyme-linked immunosorbent assay

FBS

Fetal bovine serum

FDR

False discovery rate

FK506

Tacrolimus

GO

Gene Ontology

GSA-Human

Genome Sequence Archive for Human

HBSS

Hank’s balanced salt solution

HCD

Higher-energy collisional dissociation

HLA

Human leukocyte antigen

IF

Immunofluorescence

IgA

Immunoglobulin A

IgG

Immunoglobulin G

IgM

Immunoglobulin M

IL-6

Interleukin-6

IL-8

Interleukin-8

IL-21

Interleukin-21

IL-21R

Interleukin-21 receptor

IRB

Institutional Review Board

LC–MS/MS

Liquid chromatography–tandem mass spectrometry

LTx

Lung transplantation

MPO

Myeloperoxidase

NEU

Neutrophil count

NETs

Neutrophil extracellular traps

NGDC

National Genomics Data Center

NK

Natural killer cell

PB

Plasmablasts

PBMCs

Peripheral blood mononuclear cells

PC

Plasma cells

PCA

Principal component analysis

PCT

Procalcitonin

PBS

Phosphate-buffered saline

PLT

Platelet count

RPMI

Roswell Park Memorial Institute medium

scRNA-seq

Single-cell RNA sequencing

SHM

Somatic hypermutation

SLE

Systemic lupus erythematosus

TACI

Transmembrane activator and CAML interactor (TNFRSF13B)

TACI-Fc

TACI–Fc fusion protein

Tfh

T follicular helper cell

TSA

Tyramide signal amplification

UMAP

Uniform Manifold Approximation and Projection

V(D)J

Variable, diversity, and joining

WGCNA

Weighted gene co-expression network analysis

T0–T4

Perioperative time points (T0, preoperative baseline; T1, postoperative day 1; T2, postoperative day 7; T3, postoperative day 14; T4,2 months after transplantation)

Authors’ contributions

Y.L., L.S., Y.Y., J.L., and W.G. contributed equally to this work. L.X. and B.W. conceived and supervised the study. J.C. provided critical clinical guidance and resources. Y.L., L.S., and Y.Y. performed data analysis and visualization. J.L., W.G., P.S. and K.W. collected clinical samples and performed laboratory experiments. R.S., D.S., J.D. and J.L. contributed to data interpretation and manuscript revision. Y.L. drafted the manuscript with input from all authors. All authors reviewed and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant No. 82172109 to L.X.) and by the Wuxi Health and Family Planning Commission (Grant No. Z202215 to B.W.).

Data availability

The datasets generated and analyzed during the current study have been deposited in the Genome Sequence Archive for Human (GSA-Human) at the National Genomics Data Center (NGDC, China) under accession PRJCA031535. Processed data supporting the main findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This multicentre study was approved by the Medical Ethics Committee of the Chinese PLA General Hospital (Approval No. 2022052701006) and the Research Ethics Committee of Wuxi People’s Hospital (Approval No. (2025) KY25010). The study was conducted in accordance with the Declaration of Helsinki and its later amendments. Written informed consent was obtained from all lung transplant recipients and healthy donors prior to participation.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.