Inhibition of mitochondrial ROS by TACI sustains bone marrow plasma cells

Abstract

Vaccines establish humoral protection via neutralizing antibodies, which are sustained by bone marrow long-lived plasma cells (LLPC). The lifespan of LLPCs determines the duration of protection, however, the mechanisms underlying LLPC survival remain poorly understood. Here, we employ plasma cell–specific conditional knockout mice to systematically dissect the roles of receptors for candidate niche factors. Unexpectedly, we find that the cytokine receptor TACI is essential for LLPC survival. Loss of TACI reduces polyclonal plasma cell numbers and abrogated LLPCs induced by both T cell-dependent and T cell-independent antigens. Importantly, TACI deficiency severely compromises protection elicited by both SARS-CoV-2 and influenza vaccines. Mechanistically, loss of TACI causes accumulation of mitochondrial reactive oxygen species and subsequent plasma cell death. Importantly, pharmacologic antioxidant treatment with the FDA-approved drug, N‑acetylcysteine, mitigates ROS accumulation, rescues LLPC numbers, and enhances influenza vaccine efficacy in vivo. Together, these results establish TACI as a non-redundant regulator of LLPC longevity and vaccine-induced protection by limiting oxidative stress, providing a potential target for enhancing vaccine efficacy and durability.

Long-lived plasma cells (LLPC) in the bone marrow are indispensable for longer-term protection against infections following vaccination. Here the authors show that the cytokine receptor TACI, but not BCMA or BAFFR, protects LLPCs from oxidative stress and thus facilitates their survival, and that genomic deletion of TACI specifically in plasma cells results in reduced numbers of LLCs and reduced vaccine efficacy in mice, which could be mitigated a pharmacologic antioxidant.

Introduction

The induction of durable neutralizing antibodies is a cornerstone of successful vaccination¹. Yet, the longevity of these crucial antibodies varies considerably across different vaccines, ranging from months to decades and resulting in substantial heterogeneity in the duration of protective immunity^2–7. These neutralizing antibodies are produced by terminally differentiated, non-dividing long-lived plasma cells (LLPC) that reside in the bone marrow^8–12. Consequently, the survival of these LLPCs is a critical determinant of the duration of antibody-mediated immunity^13–17. LLPCs persist in an antigen-independent manner^18–22, providing rapid and potent protection by neutralizing invading pathogens through secreted antibodies before infection can fully establish—a process known as sterilizing immunity^{3,4,8–12,23}. Therefore, the efficient generation and long-term survival of pathogen-specific LLPCs are essential not only for preventing reinfection but also for limiting viral shedding and transmission within populations, thereby contributing to herd immunity^12,23.

Plasma cells (PC) are generated in secondary lymphoid organs and only a small percentage of PCs successfully engraft into the bone marrow or mucosal tissues and become LLPCs^10,16. Despite their critical role, the cellular and molecular pathways that enable durable LLPC survival remain incompletely defined. Among the various potential mediators of PC survival, the TNF superfamily members B cell activating factor (BAFF) and a proliferation inducing ligand (APRIL) have been considered to be key niche factors^14,24. The receptors for these niche factors include BCMA (B cell maturation antigen), TACI (transmembrane activator and CAML interactor) and BAFF receptor (BAFFR)¹⁴. Previous studies employing global knockout mouse models or blocking antibodies have yielded conflicting results regarding the in vivo functions of BCMA and TACI^25–32. Although BCMA has been generally considered to be essential for the survival of LLPCs^29,31, other investigations suggest that it is dispensable for humoral immune responses^25,32. Importantly, Menzel et. al. developed two different genomic knockout mouse lines for BCMA and showed that BCMA deletion did not alter PC numbers, their maintenance/longevity, or antibody responses, suggesting that BCMA is dispensable LLPCs³². Similarly, while some reports suggested that TACI deficiency impairs LLPC survival induced by T cell-dependent antigens^26,27, others indicate TACI is dispensable for T cell-dependent IgG responses^28,29. The role of BAFFR in PC maintenance has remained particularly challenging to resolve, despite previous predictions regarding its dispensable function in PCs.

Here, using plasma cell–specific conditional knockout mouse models, we directly interrogate candidate survival regulatory receptors in LLPCs. We find that the TACI is required for the maintenance of bone marrow LLPCs and for the long-term persistence of vaccine‑elicited protective antibodies. Mechanistically, TACI sustains LLPC survival by suppressing mitochondrial reactive oxygen species (ROS); loss of TACI results in elevated mitochondrial ROS and progressive loss of antigen‑specific LLPCs. Importantly, pharmacologic antioxidant treatment with the FDA-approved drug, N‑acetylcysteine, mitigates ROS accumulation, rescues LLPC numbers, and enhances influenza vaccine efficacy in vivo. These results reveal a previously underappreciated TACI–mitochondrial ROS axis that governs LLPC longevity and identify a potential approach to improve the efficacy and durability of vaccine‑induced humoral immunity.

Results

TACI is essential for the maintenance of bone marrow polyclonal PCs

To systematically analyze the roles of TACI, BCMA, and BAFFR in the maintenance of PCs in vivo, we generated Taci ^flox/flox, Bcma ^flox/flox, and Baffr ^flox/flox mice (Supplementary Fig. 1) and crossed each line with Jchain-CreERT2 to delete these genes specifically in PCs33. We chose to knock out different exons of each gene based on factors including the efficiency of homologous recombination and Cre-loxP mediated recombination, and the potential impact of loxP insertion on the expression of neighboring genes. First, we analyzed the role of TACI in PCs using Jchain-Cre ^ERT2;Taci ^flox/flox. To investigate the mechanism by which TACI sustains LLPCs, we first analyzed the effect of TACI in the maintenance of bone marrow polyclonal PCs. To control for any effect caused by different genotypes, we included both Jchain-Cre ^ERT2 mice and Taci ^flox/flox mice as controls. We induced the deletion of Taci by giving these mice tamoxifen for 7 days and analyzed polyclonal PCs in bones including femur, tibia, pelvis, humerus, sternum, skull and thoracic vertebrate using flow cytometry and ELISpot assay (Fig. 1a). We found that Taci was efficiently deleted after 7-day tamoxifen treatment in PCs from Jchain-Cre ^ERT2; Taci ^flox/flox mice compared to Jchain-Cre ^ERT2 and Taci ^flox/flox control mice (Fig. 1b). FACS analysis showed that deletion of Taci in PCs reduced both the percentage and number of PCs in all the bones we examined (Fig. 1cand Supplementary Figs. 2a and 3a). Furthermore, ELISpot assay also revealed a significant reduction of IgA-, IgM-, and IgG-secreting cells caused by the loss of Taci in all the bones we examined (Fig. 1d, e). To investigate whether loss of Taci had a sustained depletion effect on bone marrow PCs, we administered tamoxifen for 3 weeks (Fig. 1f) and found that long-term deletion efficiently ablated Taci in PCs (Supplementary Fig. 3b), significantly reduced the number of PCs in the bone marrow (Fig. 1g) and the serum concentrations of IgM, IgA, and IgG1 (Fig. 1h). Strikingly, just 3 days of tamoxifen treatment (Fig. 1i), efficiently deleted Taci in PCs (Supplementary Fig. 3c) and significantly reduced both the percentage and number of PCs in the bone marrow as shown by both flow cytometry (Fig. 1j) and ELISpot assay (Fig. 1k), suggesting that Taci deletion caused a quick reduction of PCs in the bone marrow. Notably, Jchain-Cre ^ERT2 and Taci ^flox/flox control mice did not show significant difference between them. While these data clearly highlight the crucial role of TACI in PC survival, it is important to note that other factors may contribute to the survival of the remaining TACI¯ PCs. Additionally, the number of GC B cells, PBs, and PCs was not affected by the deletion of Taci in PCs when the Jchain-Cre ^ERT2;Taci ^flox/flox and Taci ^flox/flox mice were treated with tamoxifen for 3 days (Supplementary Figs. 2b and 3d), excluding the effect of TACI deletion on the generation of PCs in the spleen. Collectively, these data show that TACI promotes the survival of polyclonal PCs under steady state.

Fig. 1. TACI promotes the maintenance of polyclonal PCs in the bone marrow.

a Experimental scheme for evaluating the effect of 7-day Taci depletion on PCs in Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice. b Representative FACS plots from 3 independent experiments, the percentage of TACI⁺ PCs in Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice after 7-day Tamoxifen treatment (n = 6 mice/group). c The percentage and number of PCs in the indicated bones (humerus, pelvis, tibia, skull, sternum, and thoracic vertebrate) from Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice (n = 6 mice/group). d Representative ELISpot images of IgA, IgM and IgG-secreting PCs in the indicated bones (humerus, pelvis, tibia, skull, sternum, and thoracic vertebrate) from Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice from 2 independent experiments. e ELISpot quantification of IgA, IgM and IgG-secreting PCs in the indicated bones (humerus, pelvis, tibia, skull, sternum, and thoracic vertebrate) from Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice (n = 6 mice/group). f Experimental scheme for evaluating the effect of 3-week Taci depletion on PCs in Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice, or Jchain-Cre ^ERT2;Taci ^flox/flox mice. g Representative FACS plots from 3 independent experiments, the percentage and number of PCs in the femur of Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice, or Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen administration for 3 weeks (n = 5 mice/group). h Concentration of total IgA, IgM, and IgG1 in serum from Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice (n = 5 mice/group) after tamoxifen treatment for 3 weeks. i Experimental scheme for evaluating the effect of 3-day Taci depletion on PCs. j Representative FACS plots from 3 independent experiments, the percentage and number of PCs in the femur of Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen treatment for 3 days (n = 4 mice/group). k Representative ELISpot images and quantification of IgA, IgM, and IgG-secreting PCs in the femur of Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice, or Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen treatment for 3 days (n = 4 mice/group). Data are mean ± s.e.m. P-values were determined by One-way ANOVA. Source data are provided as a Source Data file.

TACI is essential for the maintenance of immunization-induced LLPCs in the bone marrow

LLPCs survive for an extended time without proliferating and sustain the serum antibodies after vaccination. We next investigate whether TACI regulates antigen-specific PCs induced by T cell-dependent (TD) and T cell-independent (TI) antigens. To test whether Taci regulates the survival of LLPCs induced by TD antigen, we immunized Taci ^flox/flox, Jchain-Cre ^ERT2 and Jchain-Cre ^ERT2;Taci ^flox/flox mice with NP-KLH and induced the Taci deletion 8 weeks post immunization (Fig. 2a). This experimental design allowed us to bypass the NP-specific PC generation phase and directly analyze the effect of Taci deletion on the survival of existing NP-specific LLPCs without the interference of new NP-specific PC input. We found that deletion of Taci significantly reduced the concentration of anti-NP IgG1 at 4 weeks and 8 weeks after tamoxifen treatment compared to Jchain-Cre ^ERT2 and Taci ^flox/flox controls (Fig. 2b). Consistent with these results, the number of anti-NP IgG-secreting cells was also reduced by the loss of Taci (Fig. 2c). These data suggest that TACI promotes the survival of LLPCs induced by TD antigens.

Fig. 2. TACI promotes the maintenance of immunization-induced LLPCs the bone marrow.

a Experimental scheme for evaluating the effect of Taci depletion on TD antigen-induced LLPCs in Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with NP-KLH. b Concentration of anti-NP IgG1 in serum from Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with NP-KLH (n = 8 mice/group). c Representative ELISpot images from 2 independent experiments, and quantification of NP-specific IgG-secreting LLPCs in the femur of Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with NP-KLH 8 weeks after tamoxifen treatment (n = 8 mice/group). d Experimental scheme for evaluating the effect of Taci depletion on TI antigen-induced LLPCs in Taci ^flox/flox mice, Jchain-Cre ^ERT2 mice, or Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with NP-Ficoll. e Concentration of anti-NP IgM in serum from Taci ^flox/flox mice or Jchain-Cre ^ERT2x mice, Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with NP-Ficoll (n = 8 mice/group). f Representative ELISpot images from 2 independent experiments and quantification of NP-specific IgM-secreting LLPCs in the femur of Taci ^flox/flox mice or Jchain-Cre ^ERT2 mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with NP-Ficoll 8 weeks after tamoxifen treatment (n = 8 mice/group). Data are mean ± s.e.m. P-values were determined by One-way ANOVA. Source data are provided as a Source Data file.

To analyze whether Taci regulates the survival of LLPCs induced by TI antigen, we immunized Taci ^flox/flox and Jchain-Cre ^ERT2; Taci ^flox/flox mice with NP-ficoll and induced the Taci deletion 8 weeks post immunization (Fig. 2d). Our results revealed that deletion of Taci significantly reduced the concentration of anti-NP IgM at 4 weeks and 8 weeks after tamoxifen treatment (Fig. 2e). Consistent with these results, the number of anti-NP IgM-secreting cells was also reduced by the loss of Taci (Fig. 2f). Collectively, these data suggest that TACI promotes the survival of LLPCs induced by both TD and TI antigens.

To exclude the potential impact of tamoxifen treatment and the Jchain-Cre ^ERT2 genotype on mouse health and immune homeostasis, we systematically analyzed the effect of tamoxifen treatment on body weight, food intake, water intake, the weight and histology of different organs, different immune cell number, and antibody titer of Jchain-Cre ^ERT2 mice and WT mice that were treated with tamoxifen for 21 days. Although the treatment of tamoxifen diet caused reduced food intake and weight loss in the initial 7 days, they quickly recovered afterwards and there was no difference between Jchain-Cre ^ERT2 mice and WT mice (Supplementary Fig. 4a–c). Furthermore, we did not observe obvious differences between these two control groups in general health parameters, histology, immune cell populations, or antibody concentration after treatment with tamoxifen for 21 days (Supplementary Figs. 2, 4 and 5). Therefore, the observed PC phenotype in Jchain-Cre ^ERT2;Taci ^flox/flox mice was caused by the loss of Taci, but not the impact of tamoxifen side effects or the Jchain-Cre ^ERT2 genotype.

TACI is essential for the vaccine-induced protection against viruses by maintaining LLPCs

To investigate whether TACI affects the durability of vaccine-induced protection, we first immunized Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice with inactivated influenza vaccine and induced Taci deletion 6 weeks after vaccination to investigate the role of TACI in the survival of already-generated influenza vaccine-induced LLPCs (Fig. 3a). Our results showed that deletion of Taci significantly reduced the concentration of anti-influenza virus haemagglutinin (HA) IgG1 in the serum (Fig. 3b). The number of anti-HA IgG-secreting LLPCs in the bone marrow was also significantly reduced by Taci ablation (Fig. 3c). When we challenged the immunized mice 8 weeks after tamoxifen treatment, we found that Jchain-Cre ^ERT2;Taci ^flox/flox mice exhibited greater weight loss and succumbed to the infection more rapidly than control Taci ^flox/flox mice (Fig. 3d and e). Consistently, the acute lung injury was more severe in the absence of Taci (Fig. 3f). In addition, the lung viral load was significantly higher in the challenged Jchain-Cre ^ERT2;Taci ^flox/flox mice compared to Taci ^flox/flox controls (Fig. 3g), suggesting impaired virus control due to lower neutralizing antibody concentration in the Jchain-Cre ^ERT2;Taci ^flox/flox mice. Our data suggest that TACI sustains the influenza vaccine-induced protection by maintaining LLPCs in the bone marrow.

Fig. 3. TACI sustains the vaccine-induced protection against viruses by maintaining LLPCs.

a Experimental scheme for evaluating the effect of Taci depletion on influenza vaccination-induced protection in Taci ^flox/flox mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice. b Remaining serum anti-HA IgG1 at different time points relative to that before tamoxifen treatment in Taci ^flox/flox mice and Jchain-Cre ^ERT2;Taci ^flox/flox mice (Taci ^flox/flox: n = 7 mice; Jchain-Cre ^ERT2;Taci ^flox/flox: n = 11 mice). c Representative ELISpot images from 3 independent experiments and quantification of HA-specific IgG-secreting LLPCs in the femur of Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice immunized with influenza vaccine 8 weeks after tamoxifen treatment (Taci ^flox/flox: n = 7 mice; Jchain-Cre ^ERT2;Taci ^flox/flox: n = 8 mice). d Body weight of influenza vaccine-immunized Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice after influenza virus challenge (n = 7 mice/group). e Survival of influenza vaccine-immunized Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice after influenza virus challenge (n = 7 mice/group). f Representative lung histology pictures from 3 independent experiments and acute lung injury (ALI) score in influenza vaccine-immunized Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice after influenza virus challenge (n = 4 mice/group). Scale bar, 2.5 mm or 100 μm. g Viral load in the lung of influenza vaccine-immunized Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice after influenza virus challenge (n = 7 mice/group). h Experimental scheme for evaluating the effect of Taci depletion on Ad5-nCo vaccine-induce protection in Taci ^flox/flox mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice. i Remaining serum anti-RBD IgG1 at different time points relative to that before tamoxifen treatment in Taci ^flox/flox or Jchain-Cre ^ERT2;Taci ^flox/flox mice (n = 10 mice/group). j Neutralizing activity of the serum from Ad5-nCo-vaccinated Taci ^flox/flox or Jchain-Cre ^ERT2;Taci ^flox/flox mice against SARS-CoV-2 spike protein-pseudotyped virus as determined by SARS-CoV-2 pseudovirus neutralization assay (n = 10 mice/group). k Representative ELISpot images from 2 independent experiments and quantification of RBD-specific IgG-secreting LLPCs in the femur of Ad5-nCo-vaccinated Taci ^flox/flox or Jchain-Cre ^ERT2;Taci ^flox/flox mice 8 weeks after tamoxifen treatment (n = 10 mice/group). Data are mean ± s.e.m. P-values were determined by unpaired two-tailed Student’s t-test except (e), which was analyzed using Log-rank (Mantel-Cox) test for the survival analysis. Source data are provided as a Source Data file.

To further confirm this finding, we immunized Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice with SARS-CoV-2 vaccine Ad5-nCoV and followed the antibody response over time after tamoxifen injection 6 weeks after vaccination (Fig. 3h). The concentration of IgG1 against the spike protein receptor binding domain (RBD) of SARS-CoV-2 declined more rapidly in Jchain-Cre ^ERT2;Taci ^flox/flox mice than in Taci ^flox/flox mice (Fig. 3i). Consistently, deletion of Taci significantly reduced the neutralizing activity of serum from Ad5-nCoV-vaccinate mice against the SARS-CoV-2 spike protein-pseudotyped virus (Fig. 3j). Additionally, ELISpot assay revealed a significant reduction of anti-RBD IgG-secreting LLPCs in the bone marrow of Jchain-Cre ^ERT2;Taci ^flox/flox mice compared to Taci ^flox/flox controls (Fig. 3k). Collectively, these data suggest that TACI sustains the different vaccines-induced protection against viral infection by maintaining LLPCs.

BCMA and BAFFR are dispensable for the maintenance of bone marrow PCs

Next, we analyzed the role of the other two receptors, BCMA and BAFFR, in PCs. BCMA was initially shown to be essential, but recently reported to be dispensable, for bone marrow LLPCs^31,32. To analyze the role of BCMA in the maintenance of PCs, we deleted Bcma specifically in PCs using Jchain-Cre ^ERT2;Bcma ^flox/flox mice (Supplementary Fig. 1b). Bcma was efficiently deleted by tamoxifen treatment for 3 days (Supplementary Fig. 6a and b). However, inducible deletion of Bcma in PCs did not affect the percentage or the absolute number of PCs in the bone marrow (Supplementary Fig. 6c and 6d). Importantly, an alternative gating strategy based on TACI, CD138, CD19, and B220 yielded the same conclusion, showing that Bcma deletion had no significant effect on bone marrow PC populations (Supplementary Fig. 6e and f). We then increased the tamoxifen treatment to 3 weeks and again found that loss of Bcma for 3 weeks did not affect the percentage or number of PCs in the bone marrow either (Supplementary Fig. 6g and h). The same conclusion was reached using another gating strategy, which again showed no significant effect of prolonged Bcma deletion on bone marrow PC populations (Supplementary Fig. 6i and j). Consistent with these results, serum concentrations of IgA, IgG1, and IgM were not altered by the loss of Bcma (Supplementary Fig. 6k). Our data support the conclusion that BCMA is not required for PCs, as previously reported by Menzel et al.³². The number of germinal center (GC) B cells, plasmablasts (PBs) and PCs in the spleen were not significantly changed after Bcma deletion, indicating that the generation of PCs was not affected (Supplementary Fig. 6l). Collectively, these data suggest that BCMA is dispensable for the survival of polyclonal PCs under steady state.

To investigate the role of BCMA in the maintenance of LLPCs induced by TD or TI antigen, we immunized Bcma ^flox/flox and Jchain-Cre ^ERT2;Bcma ^flox/flox mice with NP-KLH or NP-Ficoll and induced Bcma deletion 8 weeks after immunization (Supplementary Fig. 7a and d). Our results showed that deletion of Bcma did not affect the concentration of TD antigen-induced anti-NP IgG1 or TI antigen-induced anti-NP IgM (Supplementary Fig. 7b and e). Consistent with these results, the numbers of anti-NP IgG-secreting cells and anti-NP IgM-secreting cells were not affected by the loss of Bcma (Supplementary Fig. 7c and f). These data suggest that BCMA is dispensable for the survival of LLPCs induced by TD and TI antigens. Collectively, these data suggest that BCMA is dispensable for the survival of bone marrow PCs.

The role of BAFFR in the survival of PCs has not been investigated, as global Baffr knockout caused B cell deficiency³⁴. To overcome this problem, we specifically deleted Baffr in PCs using Jchain-Cre ^ERT2;Baffr ^flox/flox mice (Supplementary Fig. 1c). These mice were analysed the same as the conditional Bcma conditional knockout mice. Our results showed that deletion of Baffr had no significant effects on PCs or the antibody levels, suggesting that BAFFR is dispensable for the function of PCs (Supplementary Figs. 8 and 9).

TACI protects PCs from mitochondrial dysfunction and ROS-mediated cell death

To investigate the mechanism by which TACI promotes PC survival, we performed RNA-seq analysis on bone marrow PCs isolated from Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen treatment. Differential gene expression analysis identified 469 genes altered by Taci loss (Supplementary Fig. 10a). Gene set enrichment analysis (GSEA) analysis revealed, among other pathways, a significant enrichment of mitochondrion-related ones following the loss of Taci (Fig. 4a and Supplementary Fig. 10b, c), suggesting mitochondrial function was dysregulated. To test this, we measured mitochondrial mass with mitotracker Green (MTG) and mitochondrial membrane potential with tetramethylrhodamine-ethyl-ester (TMRE) by flow cytometry. Indeed, deletion of Taci reduced mitochondrial mass (Fig. 4b), increased mitochondrial membrane potential (Fig. 4c), and raised the TMRE/MTG ratio (Fig. 4d), suggesting fewer or smaller mitochondria that are individually more polarized and potentially more metabolically active. Electron microscopy confirmed a reduced number of mitochondria in bone marrow PCs from Jchain-Cre ^ERT2;Taci ^flox/flox mice (Fig. 4e, f). To assess mitochondrial respiratory function directly, we performed Seahorse metabolic flux analyses on isolated bone marrow PCs. Taci-deficient PCs exhibited higher basal and maximum oxygen consumption rates (OCR) (Fig. 4g, h), indicating that PCs switch from a normally low oxidative phosphorylation (OXPHOS) state to a high OXPHOS state after TACI deletion. Because elevated mitochondrial activity can increase reactive oxygen species (ROS), we measured intracellular ROS and mitochondrial superoxide by flow cytometry. We found that both intracellular ROS and mitochondrial superoxide were elevated in Jchain-Cre ^ERT2;Taci ^flox/flox mice compared to Taci ^flox/flox controls (Fig. 4i, j), suggesting that loss of TACI promotes mitochondrial ROS accumulation.

Fig. 4. TACI protects PCs from mitochondrial dysfunction and ROS-mediated cell death.

a Gene set enrichment analysis (GSEA) of PCs from Jchain-Cre ^ERT2;Taci ^flox/flox mice and control mice. Mitochondrion-related gene sets are significantly enriched in PCs after Taci deletion. b–d Representative FACS plots from 3 independent experiments, and mitochondrial mass (b), mitochondrial membrane potential (c) and TMRE/MTG ratio (d) of PCs from Taci ^flox/flox or Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen treatment for 7 days (n = 8 mice/group). e Representative electron micrographs of PCs isolated from Jchain-Cre ^ERT2;Taci ^flox/flox and Taci ^flox/flox mice from 3 independent experiments. Scale bar, 1 μm. f Quantification of mitochondria in PCs isolated from Jchain-Cre ^ERT2;Taci ^flox/flox and Taci ^flox/flox mice (Taci ^flox/flox: n = 45 cells; Jchain-Cre ^ERT2;Taci ^flox/flox: n = 46 cells). g OXPHOS levels measured by oxygen consumption rates (OCR) in freshly isolated PCs from Jchain-Cre ^ERT2;Taci ^flox/flox and Taci ^flox/flox mice at baseline and after treatment with the indicated drugs (n = 6 mice/group). h Basal and maximal OCR of PCs (n = 6 mice/group). i-j, Representative FACS plots from 3 independent experiments, and intracellular ROS levels (i) and mitochondrial superoxide levels (j) in PCs of Taci ^flox/flox or Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen treatment for 7 days (n = 8 mice/group). k Experimental scheme for evaluating whether ROS scavenger, NAC, can rescue Taci deficiency-induced PC depletion. l Representative FACS plots, the percentage and number of PCs in the femur of Taci ^flox/flox mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice with or without NAC treatment (Taci ^flox/flox+vehicle, Jchain-Cre ^ERT2;Taci ^flox/flox + NAC: n = 8 mice/group; Taci ^flox/flox + NAC, Jchain-Cre ^ERT2;Taci ^flox/flox +vehicle: n = 9 mice/group). m Experimental scheme for evaluating whether NAC can rescue the depletion of immunization-induced antibody and LLPCs induced by Taci deletion. n Concentration of anti-NP IgG1 in serum from NP-KLH-immunized Taci ^flox/flox mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice with or without NAC treatment (Taci ^flox/flox+vehicle: n = 8 mice; Taci ^flox/flox + NAC, Jchain-Cre ^ERT2;Taci ^flox/flox+vehicle, Jchain-Cre ^ERT2;Taci ^flox/flox + NAC: n = 9 mice/group). o Representative ELISpot images from 3 independent experiments and quantification of NP-specific IgG-secreting LLPCs in the femur of NP-KLH-immunized Taci ^flox/flox mice or Jchain-Cre ^ERT2;Taci ^flox/flox mice with or without NAC treatment 8 weeks after tamoxifen treatment (Taci ^flox/flox+vehicle: n = 8 mice; Taci ^flox/flox + NAC, Jchain-Cre ^ERT2;Taci ^flox/flox+vehicle, Jchain-Cre ^ERT2;Taci ^flox/flox + NAC: n = 9 mice/group). Data are mean ± s.e.m. P-values were determined by unpaired two-tailed Student’s t-test. Source data are provided as a Source Data file.

To test whether elevated ROS contributes to PC loss after Taci deletion, we treated Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice with N-acetylcysteine (NAC), a ROS scavenger, in drinking water together with Taci deletion with tamoxifen (Fig. 4k). Our results showed that NAC treatment rescued the reduction in PC number in the Jchain-Cre ^ERT2;Taci ^flox/flox mice (Fig. 4l). To test whether NAC can rescue the loss of antigen specific LLPC after Taci deletion, we immunized Taci ^flox/flox and Jchain-Cre ^ERT2; Taci ^flox/flox mice with NP-KLH, and treated them 6 weeks later with tamoxifen together with NAC or control water (Fig. 4m). As expected, loss of Taci dramatically decreased the serum anti-NP IgG1 concentration (Fig. 4n) and NP-specific bone marrow LLPCs in the absence of NAC (Fig. 4o). Strikingly, NAC treatment rescued both the serum anti-NP IgG1 concentration (Fig. 4n) and the number of NP-specific bone marrow LLPCs in Jchain-Cre ^ERT2;Taci ^flox/flox mice (Fig. 4o). Together, these data indicate that TACI supports LLPCs survival by limiting mitochondrial ROS–driven cell death.

ROS causes rapid death of PCs in vitro

To further elucidate the mechanism by which PC numbers decrease after TACI deletion, we first examined the expression levels of apoptosis-related genes and found no differences between Taci ^flox/flox and Jchain-Cre ^ERT2;Taci ^flox/flox mice (Fig. 5a). This is consistent with the absence of a significant change in the percentage of cleaved caspase 3⁺ PCs (Fig. 5b), arguing against caspase-dependent apoptosis as the primary mechanism. To probe alternative mechanisms, we modeled mitochondrial dysfunction and ROS accumulation in vitro by treating isolated bone marrow PCs with CCCP (carbonyl cyanide m‑chlorophenyl hydrazone), an uncoupler of oxidative phosphorylation (Fig. 5c). As expected, CCCP induced a rapid rise in intracellular ROS within 30 minutes (Fig. 5d). Interestingly, CCCP treatment depleted PCs in just 4 hours (Fig. 5e). We next tested inhibitors of major regulated death pathways—apoptosis (Z‑VAD), necroptosis (Necrostatin‑1, Nec‑1), ferroptosis (Ferrostatin‑1, Fer‑1), and autophagy (chloroquine), as well as NAC, in CCCP-induced PC death. Unexpectedly, none of the canonical death‑pathway inhibitors rescued CCCP-induced PC loss, whereas NAC completely prevented it (Fig. 5f). To further investigate how PCs died after CCCP treatment, we imaged the PCs from Prdm1-EYFP mice³⁵, a PC reporter murine line in which PCs are identified as Prdm1-EYFP ⁺ cells, during CCCP treatment using live cell imaging in vitro (Fig. 5g). We found that CCCP exposure caused a rapid loss of GFP signal and cell-body swelling within 2 minutes (Fig. 5h and Supplementary Movie 1), suggesting that CCCP induced rapid cell rupture and release of intracellular contents. These observations suggest that loss of TACI may cause ROS‑driven necrotic cell death rather than classical apoptotic, necroptotic, or ferroptotic cell death. In summary, these data indicate that TACI supports LLPCs survival by limiting mitochondrial ROS–driven cell death.

Fig. 5. Mitochondria dysfunction induced quick death of PCs in vitro.

a Expression level of genes in cell survival, apoptosis, and PC signature pathways (n = 4 mice/group). b Representative FACS plots from 3 independent experiments and the percentage of apoptotic PCs determined by cleaved Caspase 3 assay in the bone marrow of Taci ^flox/flox mice and Jchain-Cre ^ERT2;Taci ^flox/flox mice after tamoxifen treatment (n = 4 mice/group). c Experimental scheme for evaluating the effect of mitochondria dysfunction induced by CCCP on ROS level and the survival of BM PCs in vitro. d Intracellular ROS levels in PCs with or without CCCP treatment in vitro for 30 minutes (n = 3 wells/group). e Representative FACS plots from 3 independent experiments and the number of Prdm1-eYFP⁺ PCs after treatment with CCCP or vehicle for 4 hours in vitro (n = 4 wells/group). f Representative FACS plots from 3 independent experiments and the number of PCs cultured in vitro with CCCP in the presence of different small molecule inhibitors (apoptosis: Z-VAD; necroptosis: Nec-1; ferroptosis: Fer-1; autophagy: chloroquine; ROS: NAC) (n = 6 wells/group). g Experimental scheme of in vitro live cell imaging of Prdm1-eYFP⁺ PCs cultured with CCCP. h Sequential images of Prdm1-eYFP⁺ PCs after treatment with CCCP from 3 independent experiments. Scale bar, 20 μm. Arrowhead: the cell lost Prdm1-eYFP signal within 2 minutes and became swollen. Data are mean ± s.e.m. P-values were determined by unpaired two-tailed Student’s t-test. Source data are provided as a Source Data file.

Pharmacologic antioxidant treatment with NAC enhances influenza vaccine efficacy in mice

To investigate whether targeting the TACI-mitochondria ROS pathway can enhance the immunogenicity of NP-KLH, we immunized wild-type mice with NP-KLH, administered NAC or vehicle control immediately after immunization, and analyzed NP-specific PC responses in the spleen and bone marrow at days 7, 10, and 21 after immunization (Fig. 6a). We found that NAC treatment significantly increased NP-specific GC B cells, plasmablasts, and PCs in the spleen, as well as NP-specific PCs in the bone marrow (Fig. 6b–g). These results suggest that NAC enhances vaccination efficacy by boosting the initial magnitude of the B cell response and promoting the successful seeding of the bone marrow PC niche, rather than solely acting on the maintenance of established LLPCs.

Fig. 6. Pharmacologic antioxidant treatment with NAC enhances influenza vaccine efficacy in mice.

a Experimental scheme for evaluating the effect of NAC on PC generation. b–d The number of NP-specific germinal center (GC) B cells (b), plasmablasts (c), and PCs (d) in the spleen of vehicle and NAC-treated mice by flow cytometry (n = 6 mice/group). e The number of NP-specific PCs in the bone marrow of Vehicle and NAC-treated mice by flow cytometry (n = 6 mice/group). f, g ELISpot images quantification of NP-specific IgG antibody-secreting cells in the spleen (f) and bone marrow (g) of vehicle and NAC-treated mice (n = 6 mice/group). h Experimental scheme for evaluating the effect of NAC on immunization-induced antibody and LLPCs in WT mice. i Concentration of anti-NP IgG1 in serum from NP-KLH-immunized mice with or without NAC treatment (vehicle: n = 9 mice; NAC: n = 8 mice). j Representative ELISpot images from 3 independent experiments and quantification of NP-specific IgG-secreting LLPCs in the femur of vehicle and NAC-treated mice (vehicle: n = 9 mice; NAC: n = 8 mice). k Experimental scheme for evaluating whether NAC can enhance influenza vaccination-induced protection in WT mice. l Concentration of anti-HA IgG1 in serum from influenza-vaccinated mice with or without NAC treatment (vehicle: n = 11 mice; NAC: n = 10 mice). m Titer of neutralizing antibodies measured by hemagglutination inhibition assay in serum vehicle and NAC treated of mice (vehicle: n = 11 mice; NAC: n = 10 mice). n Representative ELISpot images from 3 independent experiments and quantification of HA-specific IgG-secreting LLPCs in the femur of vehicle and NAC-treated mice (n = 7 mice/group). o, p Body weight (o) and survival (p) of vehicle and NAC-treated mice post challenge (n = 8 mice/group). q Representative lung histology pictures from 3 independent experiments and acute lung injury (ALI) score in mice with or without NAC treatment post challenge (n = 5 mice/group). Scale bar, 2.5 mm or 100 μm. r Viral load in the lung of vehicle and NAC-treated mice post challenge (n = 8 mice/group). Data are mean ± s.e.m. P-values were determined by unpaired two-tailed Student’s t-test except (i), which was analyzed using Log-rank (Mantel-Cox) test for the survival analysis. Source data are provided as a Source Data file.

To further assess the effect of NAC on the generation of antigen-specific LLPCs, wild-type mice were immunized with NP-KLH, treated with NAC or vehicle for 4 weeks, and followed by analysis of serum anti-NP IgG1 and bone marrow LLPCs for 8 weeks (Fig. 6h). Consistent with the early response data, NAC treatment significantly increased the concentration of anti-NP IgG1 in the serum compared with control mice at 2, 4, and 8 weeks after immunization (Fig. 6i). The number of anti-NP IgG-secreting LLPCs in the bone marrow was also significantly increased by NAC treatment (Fig. 6j), suggesting that the NAC treatment during early phase of immunization boost the LLPCs generation and serological memory

To investigate whether targeting the TACI-mitochondria ROS pathway can enhance vaccine efficacy, we vaccinated wild-type mice with influenza vaccine and treated them with antioxidant NAC or vehicle control immediately after vaccination (Fig. 6k). Neutralizing antibodies against the influenza viral surface glycoprotein HA blocks viral entry into cells and represents the major component of vaccine-mediated protection³⁶. NAC treatment significantly increased the vaccine-induced anti‑HA IgG1 titers compared with control (Fig. 6l). We next assessed serum neutralization using the hemagglutination inhibition (HI) assay and found that sera from NAC‑treated mice exhibited significantly higher HI activity (Fig. 6m). Consistently, the HA-specific LLPCs in the bone marrow was also increased by NAC treatment (Fig. 6n).

To determine whether NAC improved functional protection, we challenged the influenza vaccine-immunized mice with a lethal dose of the live parental influenza virus (Fig. 6k). NAC‑treated animals showed less body‑weight loss (Fig. 6o), increased survival (Fig. 6p), reduced lung pathology (Fig. 6q), and lower viral genome levels in the lung (Fig. 6r) compared with vehicle controls. Together, these findings show that NAC-mediated reduction of mitochondrial ROS enhances humoral responses and confers improved protection against lethal influenza challenge, supporting antioxidant modulation as a strategy to boost vaccine efficacy.

Discussion

Vaccines represent a milestone in medical history, providing long-term protection through the generation of durable, antibody-secreting long-lived plasma cells (LLPCs)^1,37. The long-term protection of successful vaccines depends on the durable protective antibodies constantly secreted by the vaccine-induced LLPCs, sometimes over decades post-vaccination⁴. Thus, the long-term survival of vaccine-induced LLPCs is the key to the efficacy of vaccines. However, the mechanism that regulates the lifespan of LLPCs still remain a puzzle. Unravelling these mechanisms is critical for the rational design of more effective vaccines. Recently, we discovered that fasting impaired the maintenance of LLPCs through β-hydroxybutyrate-mediated downregulation of CXCR4¹⁷. In this study, using PC-specific conditional genetic deletion models, we demonstrate that the TNFRSF member TACI, but not BAFFR or BCMA, promotes the maintenance of PCs in the bone marrow by inhibiting ROS-induced cell death.

The mechanisms governing LLPC survival have been extensively investigated, implicating both intrinsic and extrinsic regulatory factors^31,38–55. However, due to the inherent limitations of in vitro co-culture systems and studies employing conventional global knockout mouse models used in previous studies, the roles of individual molecule and cell population remain poorly defined. Previous studies employing global knockout mouse models or blocking antibodies have yielded conflicting results regarding the in vivo functions of BCMA and TACI^25–31. Our findings provide independent validation of the paradigm-shifting study by Menzel et al., which utilized genomic knockout lines to show that BCMA is dispensable for PC longevity³². However, our results do not exclude the possibility that the APRIL-BCMA axis may become important in the absence of TACI. BCMA is redundant in the presence of TACI, while it may promote the survival of PCs in the absence of TACI.

TACI appears to exert a stage-specific dual role in humoral immunity. Mutations in Taci or Tnfrsf13b are associated with common variable immunodeficiency (CVID) and certain Taci mutations predispose CVID patients to autoimmunity and lymphoproliferation⁵⁶. In mice, TACI deficiency causes lymphoproliferation and autoimmunity, suggesting that TACI is a negative regulator of humoral responses^30,57. Other reports showed that TACI regulates the class switching and T-independent humoral responses^28,58,59. Moreover, loss of TACI in mice was reported to enhance antibody avidity and the clearance of intestinal pathogenes²⁹. These findings suggest that TACI regulates B cells at different activation stages, likely exerting an inhibitory effect on B cell activation and proliferation during early stages. This dual role complicates the analysis of the effect of TACI on PCs using global knockout models. Thus, TACI appears to restrict B cell activation at early stages while promoting PC survival at later stages. Consistent with previous studies showing that metabolism regulates the survival of PCs44^60–63, we found that TACI sustains the survival of PCs by restraining the mitochondrial ROS, and treatment of mice with the approved antioxidant drug⁶⁴, NAC, enhances the efficacy of the influenza vaccine.

The mechanistic link between TACI signaling and mitochondrial integrity is a critical finding of this study. TACI is known to activate the non-canonical NF-κB and PI3K-AKT pathways⁵⁶, both of which are central regulators of metabolic fitness. TACI-mediated signaling may sustains the expression of antioxidant enzymes that stabilize the mitochondrial membrane. In the absence of TACI, the loss of these survival signals likely leads to mitochondrial depolarization and increased superoxide leakage from the electron transport chain, culminating in the ROS-mediated apoptosis we observed. This aligns with emerging evidence that PCs are uniquely sensitive to proteotoxic and oxidative stress due to their massive secretory burden⁴⁷. To mitigate this ROS-induced death, we utilized N-acetylcysteine (NAC), a clinically approved thiol-containing compound⁶⁵. NAC serves as a potent antioxidant both by directly scavenging reactive oxygen species and, more importantly, by acting as a rate-limiting precursor for reduced glutathione (GSH) synthesis⁶⁶. By replenishing the intracellular pool of GSH, NAC enhances the buffering capacity of the PC against mitochondrial oxidative stress. Our finding that NAC promotes the initial seeding of the PC pool and rescues TACI-deficient LLPCs highlights the metabolic vulnerability of the plasma cell lineage during its transition from the lymphoid organs to the bone marrow niche. Consequently, manipulating TACI signaling or utilizing antioxidant interventions like NAC represents a promising strategy to enhance vaccine immunogenicity and may offer a therapeutic vulnerability to target in PC-related malignancies such as multiple myeloma.

Methods

Mice

All mice were maintained on a C57BL/6 J background and housed on a 12-hour light-dark cycle at 22–25 °C with 40–70% humidity under specific pathogen-free conditions in the Laboratory Animal Centre of Sun Yat-sen University and fed with autoclaved food. Experimental procedures were approved by the Animal Care and Ethics Committee of Zhongshan School of Medicine, Sun Yat-sen University (SYSU-IACUC). B6.Cg-Tg(Prdm1-EYFP)1Mnz/J mice³⁵ and JchainCre ^ERT2 mice³³ were purchased from the Jackson Laboratory. Bcma ^flox/flox, Taci ^flox/flox, and Baffr ^flox/flox mice were generated by Cyagen (Guangzhou, Guangdong, China). C57BL/6 J mice were purchased from GemPharmatech (Nanjing, Jiangsu, China). All mice used in the experiment were 8-12 weeks of age. Both male and female mice were used in the experiment, and no significant differences in immunological response were observed. Mice were euthanized by CO₂ inhalation followed by confirmatory cervical dislocation.

Immunization

For NP-KLH immunization, 50 µg NP-keyhole limpet hemocyanin (NP-KLH, Biosearch Technologies) precipitated onto Alum (InvivoGen) was administered to mice via intraperitoneal injection. For NP-Ficoll immunization, 25 µg NP-Ficoll (Biosearch Technologies) precipitated onto Alum (InvivoGen) was administered to mice via intraperitoneal injection. For SARS-CoV-2 vaccination using Ad5-nCoV, mice were immunized intramuscularly with 1 × 10¹⁰ of Ad5-nCoV particles. For Influenza vaccine immunization, mice were immunized with 6 µg inactivated A/Puerto Rico/8/34 (PR8, H1N1) vaccine adjuvanted with AddaVax (InvivoGen). Over six weeks post-immunization, tamoxifen was injected intraperitoneally to induce the deletion of specific genes.

Tamoxifen treatment

Tamoxifen was administered as previously described⁶⁷. Briefly, Jchain-cre ^ERT2 mice and control mice were injected once daily with 100 mg/kg tamoxifen in corn oil for 3-7 consecutive days. For long-term treatment, mice were fed a diet containing 400 mg tamoxifen/kg.

In vivo NAC treatment

Mice were given 0.1 % NAC in drinking water without restriction. The water was changed twice a week.

Flow cytometry

Cell isolation and analysis were performed as previously described⁶⁸. Briefly, femurs were gently flushed with 1 mL ice-cold PEB buffer (PBS supplemented with 0.5% BSA and 2 mM EDTA) through a 1 mL syringe (KDL) with a 21 G needle into 1.5 mL Eppendorf tubes. Other bones—including humerus, pelvis, tibia, skull, sternum, and thoracic vertebrae —were crushed with mortar and pestle in PEB to release bone marrow cells. Spleens were mashed through a 40-µm filter mesh into 6-well plates containing 2 mL ice-cold PEB. Red blood cells were lysed by ammonium chloride. Cells were first treated with Fc blocker (Biolegend), then stained with antibody against surface antigens in PEB buffer for 30 min at 4 °C, and analyzed on either LSR Fortessa (BD Biosciences) or Attune (Life Technologies) flow cytometer. The data were analysed with FlowJo software (Tree Star). Dead cells were excluded by FSC, SSC, and DAPI (Sigma) or propidium iodide staining (PI, Thermo Fisher). Plasma cells were gated as CD138^high B220^¯ in WT mice or Prdm1-eYFP⁺CD138^high in Prdm1-eYFP mice. Fluorophore-conjugated antibodies against CD138 (281-2), CD19 (6D5), CD3 (145-2C11), Gr1 (RB6-8C5), F4/80 (BM8), CD115 (AFS98), CD38 (S21016F), IgD (11-26 c.2a) and GL7 (GL7) were from Biolegend. Antibodies against B220 (RA3-6B2) and CD267 (8F10-3) were from eBioscience. Antibodies were used at 1:100 dilution for cell staining. For cell sorting, samples were processed under sterile conditions and sorted on a BD FACSAria or Influx cell sorter.

ELISA

96-well plates (Thermo Scientific or Jet) were precoated with NP-BSA (10 µg/mL, Biosearch Technologies), SARS-CoV-2 RBD recombinant protein (2 µg/mL, lab-made), Influenza A HA protein (1 µg/mL, Sino Biological) in 100 mM bicarbonate/carbonate buffer or PBS overnight at 4°C. Sera were diluted and incubated on the plates for 2 hours together with antigen-specific antibodies of known concentration as standards (anti-NP IgG1 from Novus Biologicals and anti-RBD IgG1 from Sino Biological) when available. After incubation with HRP-conjugated goat anti-mouse IgA (Southern Biotech), HRP-conjugated goat anti-mouse IgG1 (Southern Biotech), or HRP-conjugated goat anti-mouse IgM (Southern Biotech) for 2hours, TMB (3,3’,5,5’-Tetramethylbenzidine) substrate reagents were added for color development. The absorbance was read at 450 nm with a reference read at 630 nm on a microplate reader (Allsheng).

ELISpot

Multiscreen-HA plates (Millipore) were coated with 5 µg/mL anti-mouse Ig (Southern Biotech), 10 µg/mL NP-BSA (Biosearch Technologies), 10 µg/mL RBD protein (lab-made), or 3 µg/mL influenza A H1N1 (PR8) hemagglutinin (HA) protein overnight at 4°C. Bone marrow cells or splenocytes were cultured on the plates for 8 h at 37 °C in a 5% CO₂ incubator. After culture, cells were removed, and capture antibodies HRP-conjugated goat anti-mouse IgA (Southern Biotech), HRP-conjugated goat anti-mouse IgM (Southern Biotech), or HRP-conjugated goat anti-mouse IgG (Southern Biotech) were incubated in the plate overnight at 4°C, and spots were developed using the substrate 3-amino-9-ethyl-carbazole (AEC) buffer.

RNA extraction and Real-time PCR analysis

RNA was extracted using TRIzol reagent (Invitrogen) following the product manual. cDNA was generated from 500 ng RNA using a Reverse Transcription kit from TaKaRa. Universal SYBR qPCR master mix (Vazyme Biotech) and a CFX 96 touch Real-Time PCR System were used for Real-time PCR analysis.

Population RNA-seq analysis

Total RNA from 10,000 sorted PCs was extracted using the EZ-press RNA Purification Kit PLUS (EZBioscience). Reverse transcription, PCR preamplification, and purification of PCR products were conducted using the Discover-sc WTA Kit V2 (Vazyme Biotech). The RNA-seq library was generated by TruePrep DNA Library Prep Kit V2 for Illumina (Vazyme Biotech) using 10 ng amplified cDNAs as input. RNA sequencing data generated from the Novaseq 6000 platform (Illumina) were processed using the following pipeline. In brief, the FastQC(v0.11.9) package was used to check the quality of raw data, and the Trim-Galore(v0.6.7) was used for trimming the adapter sequence and filtering reads with low quality. Then, paired-end sequencing reads were aligned to the mouse genome (GRCm38.primary_assembly.genome) using Spliced Transcripts Alignment to a Reference (STAR) 2.7.10a. Package featureCounts v.2.0.1 was used to count the read numbers mapped to each gene, and then the fragments per kilobase of exon model per million mapped reads (FPKM) of each gene was calculated based on the length of the gene and the read count mapped to this gene. Differential expression analysis between two groups was performed using the DESeq2 R package, which provides statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution.

Influenza vaccine experiments

An inactivated influenza vaccine was generated and used to vaccinate mice as described before⁶⁹. Briefly, A/Puerto Rico/8/34 (PR8, H1N1) influenza virus expanded in MDCK cells was inactivated with BPL (Millipore Sigma), concentrated by ultracentrifugation, and purified using discontinuous sucrose density gradient ultracentrifugation. The total protein content of the purified influenza vaccine was determined using a BCA assay kit (Thermo Fisher Scientific). For intramuscular immunization in mice, 6 µg of vaccine in 25 µl PBS was mixed with 25 µl adjuvant AddaVax (InvivoGen) and administered to the mice twice at a 2-week interval. For the Jchain-Cre ^ERT2; Taci ^flox/flox and control mice challenge experiment, 5× LD₅₀ of H1N1 (A/PR8) virus in 20 μL solution was given nasally to the mice. For the NAC-treated mice challenge experiment, 20× LD₅₀ of H1N1 (A/PR8) virus in 20 μL solution was given nasally to the mice. Body weight was monitored and mice were humanely euthanized if more than 25% of initial body weight was lost.

Hemagglutination inhibition assay

Collected serums were treated with 56 °C heat to remove nonspecific inhibitors of hemagglutination. Influenza virus strain (PR8, H1N1) was mixed with the first diluted forty-fold, followed by serial two-fold dilutions in PBS in U-bottomed 96-well plates. After 30 min of incubation, chicken erythrocytes were added to the mixtures. Plates were incubated at room temperature for 1 h. The HI titer was defined as the highest dilution of the serum able to inhibit hemagglutination.

Assessment and scoring of lung injuries

To investigate lung injury in mice challenged with influenza virus, a standardized lung injury scoring system was employed and scored by two persons independently, as described in the American Thoracic Society Workshop Report⁷⁰. A minimum of 20 random fields (400× magnification) from each sample were analyzed. Five histological parameters were evaluated: (A) alveolar neutrophils (0: none; 1: 1-5; 2: 5), (B) interstitial neutrophils (0: none; 1: 1-5; 2: >5), (C) hyaline membranes (0: none; 1: 1; 2: > 1), (D) proteinaceous debris (0: none; 1: 1; 2: > 1), and (E) alveolar septal thickening (0: <2 x; 1: 2 × - 4 ×; 2: > 4 ×). Score =[(20×A) + (14×B) + (7×C) + (7×D) + (2×E)]/(number of fields×100).

Analysis of PC generation

Mice were immunized with NP-KLH as described above and immediately assigned to NAC or control treatment after immunization. Splenocytes were collected for flow cytometric analysis of NP-specific GC B cells, plasmablasts, and plasma cells, whereas bone marrow cells were collected for analysis of NP-specific plasma cells. NP-specific IgG-secreting PCs in the spleen and bone marrow were further quantified by ELISpot.

Analysis of PCs in vitro

Bone marrow mononuclear cells of Prdm1-eYFP mice were cultured in the presence or absence of 50 μM carbonyl cyanide m-chlorophenyl hydrazine (CCCP, Sigma) at 37 °C in the cell culture incubator. To analyse how CCCP induces PC death, inhibitors including Z-VAD (20 µM, MCE), Necrostatin-1 (Nec-1, 20 µM, MCE), Ferrostatin-1 (Fer-1, 20 µM, MCE), chloroquine (300 µM, MCE), and N-acetyl cysteine (NAC, 5 mM, MCE) were added to the culture. The number of PCs identified as Prdm1-eYFP⁺CD138^high was determined by flow cytometry.

Live cell imaging in vitro

PCs were purified from the bone marrow mononuclear cells of Prdm1-eYFP mice by a mouse CD138⁺ Plasma Cell Isolation Kit (Miltenyi Biotec) and cultured in the presence or absence of 50 μM CCCP at 37 °C. For live cell imaging, PCs were continuously imaged at 20X magnification on a Lionheart FX automated microscope (BioTek). Images were processed using ImageJ.

Mitochondrial analysis

For mitochondrial mass and membrane potential analysis, cells were stained with 0.1 μM MitoTracker Green (MTG, Thermo Fisher) and 0.1 μM Tetramethylrhodamine, ethyl ester (TMRE, MCE) following the manufacturer's protocols after staining with antibodies against surface molecules.

ROS detection

Diacetyldichlorofluorescein (DCFH-DA, MCE) and MitoSOX (MCE) were used to detect cellular ROS and mitochondrial superoxide respectively, following the protocols recommended by the manufacturer after the cells were stained with antibodies against surface molecules.

Mitochondrial function analysis

Cellular bioenergetics were assessed using a Seahorse XFe96 Extracellular Flux Analyzer. One day before the assay, the sensor cartridge was hydrated in calibration buffer at 37 °C in a non-CO₂ incubator. XFe96 microplate wells were coated with Poly-D-lysine (Corning) at a concentration of 100 μg/ml at 4°C and rinsed with sterile water. Cells were resuspended in XF base minimal 1640 media containing 10 mM glucose, 2 mM glutamine, and 1 mM sodium pyruvate. Approximately 1 × 10⁵ FACS sort-purified PCs in 180 μL of medium were added per well in the microplate. The plates were centrifuged at 200 g for 1 min to promote cell adhesion and incubated for 1 hour at 37 °C in a non-CO₂ incubator. OCR were measured at baseline and following sequential addition of oligomycin (1 μM), FCCP (0.75 μM), and antimycin A and rotenone (0.5 μM each) using the XFe96 Analyzer.

Electron microscopy

FACS-sorted PCs were fixed with 2.5% glutaraldehyde in PBS for 2 h, followed by rinsing 3 times with PBS. Post-fixation staining was performed with 1% osmium tetroxide/K₃Fe(CN)₆ for 2 hours. After washing with distilled water, cells were dehydrated in a graded series of 50%, 70%, 80%, 90%, 95%, 100%, and 100% ethanol for 15 minutes each. After the final incubation, cells were incubated in a 1:1 mix of ethanol and 812 embedding agent (SPI) overnight and then in 100% 812 embedding agent overnight. Cells were then infiltrated and embedded in Pon 812 resin (Sigma). The blocks were trimmed, and 60-80 nm sections were contrast-stained with uranyl acetate and citrate. Cells were then examined with TEM (FEI, TECNAI G2 20 TWIN). Numbers of mitochondria per cell section and fraction of mitochondrion-occupied area in the cell were quantified with ImageJ.

SARS-CoV-2 pseudovirus neutralization assay

Neutralizing activity of the serum from mice vaccinated with SARS-CoV-2 vaccine Ad5-nCoV after different treatments against SARS-CoV-2 pseudovirus expressing the full-length spike protein was determined as previously reported⁷¹. Briefly, SARS-CoV-2 pseudovirus bearing the full-length spike protein of SARS-CoV-2 was produced in an Env-defective, luciferase-expressing HIV-1 backbone using HEK293T cells. Twenty thousand HEK293T-ACE2 cells/well were seeded in 96-well plates overnight. Diluted serum by a factor of 90 was mixed with the pseudovirus and incubated at 37 °C for 1 hour. Subsequently, the mixture of serum and pseudovirus was added to the HEK293T-ACE2 cells and centrifuged at 800 × g for 30 minutes. After 48 hours, the luminescence unit (RLU) value was measured using the Luciferase Assay System (Promega, USA) and the inhibition rate was used to evaluate the neutralization of test serum against pseudovirus.

Ad5-nCoV adenovirus vaccine production

The recombinant adenovirus vaccine against SARS-CoV-2 (Wuhan-Hu-1) Ad5-nCoV was amplified by serial passage on HEK293A cells and purified chromatographically using Adeno-X Maxi Purification Kit (Clontech)⁷¹.

Statistics and reproducibility

Statistical analyses were performed with Prism 9 (GraphPad Software). Results are presented as means ± SEM. All experiments were repeated at least once, except the RNAseq experiment. Unpaired two-tailed Student’s t-tests were used to compare two groups. One-way ANOVA analysis was used for multiple group comparisons. The log-rank (Mantel-Cox) test was used for the survival analysis.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

C.X. designed and supervised the study, analyzed results, wrote the manuscript, and obtained funding; YM.Z. designed the experiments, performed experiments, analyzed results, and wrote the manuscript; J.C. and H.L. performed experiments and analyzed results; Z.T, X.H., Z.J., H.J., H.M. and C.Y analyzed results; JM.C. performed experiments.

Peer review

Peer review information

Nature Communications thanks Andrew Getahun, Pascal Schneider and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This study was supported by grants from the National Key R&D Program of China (2022YFA1105000 and 2022YFA1104100), the National Natural Science Foundation of China (92168102 and 82271783), the Guangdong Basic and Applied Basic Research Foundation (2023A1515012294), and Guangzhou Basic and Applied Basic Research Foundation (2024A04J6379) to C. Xu.

Data availability

RNA-seq data of PCs isolated from Taci ^flox/flox mice and Jchain-Cre ^ERT2; Taci ^flox/flox mice (accession#: GSE306258) are available through Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE306258). Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-75430-w.