The immune system has evolved to discriminate between “self” and “nonself,” thereby controlling alternative reactions: self-tolerance in the case of the former and aggressive responses in that of the latter. This principle appears to be consistently upheld in allogeneic transplantation, where highly polymorphic proteins derived from allografts, such as human leukocyte antigens (HLAs) and minor histocompatibility proteins, are perceived as “nonself” molecules by the recipient. When these donor-specific antigens are recognized and presented by dendritic cells and B cells, antigen-cognate T cell clones are activated and differentiate into effector cells, including follicular helper T (Tfh) cells, proinflammatory Th17 cells, and cytotoxic T lymphocytes. Some activated B cells give rise to plasma cells (PCs) that secrete potentially pathogenic alloantibodies, in a process that may be dependent on or independent of Tfh cells. In kidney transplantation, approximately 7% of adult recipients reportedly develop donor-specific antibodies (DSAs; primarily donor HLA-specific antibodies) within 5 years of transplantation, and this proportion rises to 20% after 10 years [1–4]. Posttransplant DSAs have also been identified in significant numbers of recipients of other solid organ transplants [5–8].

In addition to being produced de novo during the posttransplantation period, alloantibodies preexist in many individuals on transplant waiting lists [9–11]. This preexistence can result from prior exposure to allogeneic tissues via pregnancy, blood transfusion, or earlier transplantation [12]. In addition, natural antibodies, which are established independently of such exposures, contribute to presensitization to alloantigens. This is likely because they are polyreactive and can cross-react with diverse epitopes from microbes, alloantigens, and autoantigens [13,14]. Indeed, transplant recipients with preformed DSAs are known to face a higher risk of antibody-mediated rejection (AMR) and subsequent allograft loss across all solid organ transplants. This evidence supports the potential critical role of preformed DSAs in the promotion of AMR [6,8,15–19].

Both preexisting and posttransplant DSAs can trigger pathological events associated with AMR. These events are characterized by swelling of endothelial cells, microvascular inflammation, and the formation of intravascular immune cell infiltrates, which may occur with or without complement deposition. Accumulating evidence indicates that AMR is a major contributor to decreased long-term graft survival. For instance, 40% of patients lose their renal allografts within 5 years following the emergence of de novo DSAs—a rate double that observed in patients without DSAs [4]. Moreover, over half of late allograft failures can be attributed to HLA DSA-mediated chronic rejection [2,20–24]. Consequently, to improve transplant outcomes, it is crucial to understand the immunopathological mechanisms underpinning DSA production and to develop strategies targeting DSAs.

Despite decades of attempts to understand the biology of antibody-secreting cells and to apply this knowledge in developing treatments, AMR remains a formidable obstacle to successful transplantation. Current desensitization strategies, such as B cell depletion protocols, are often ineffective. One potential explanation is the inadequate depletion of alloreactive memory B cells and long-lived PCs (LLPCs), which together form the basis of humoral memory. Although the characteristics of LLPCs have not been fully elucidated, they possess unique features that set them apart from other B cell subsets, including conventional PCs. Furthermore, they have been implicated in AMR, especially chronic AMR.

Here, we present a contemporary perspective on the biology of antibody-secreting B cells, focusing on LLPCs. We explore their developmental pathways, survival mechanisms, and functional characteristics. Given the increasingly recognized importance of LLPCs in AMR, we also discuss potential strategies for targeting these cells.

AMR is initiated by the binding of alloantibodies to alloantigens, predominantly HLA molecules, on the surface of vascular endothelial cells. This interaction induces inflammatory responses mediated by the complement system and/or Fc receptors [25–27]. Local complement activation results in cell lysis, the recruitment and activation of neutrophils, and thrombus formation. Concurrently, the engagement of Fc receptors by alloantibodies authorizes neutrophils and natural killer cells to attack and destroy endothelial cells. These combined mechanisms contribute to endothelial injury and the development of intravascular thromboses, leading to graft dysfunction.

The processes that underlie chronic AMR are more complex than those associated with acute AMR. They develop insidiously over months to years and may occur with or without clinically evident episodes of acute rejection. In the majority of vascularized grafts, chronic inflammatory responses induce the proliferation of intimal smooth muscle cells within blood vessels and fibroblasts in the interstitial tissue, leading to vascular occlusion and interstitial fibrosis [28,29]. Ultimately, the grafts fail, predominantly due to the resulting ischemic damage [30].

The adaptive immune response to an allograft begins in the peripheral lymphoid organs that drain the graft (Fig. 1). Donor dendritic cells within the allograft migrate to the lymphoid tissue, where they can directly present allogeneic HLA molecules to host T cells. Alternatively, in the indirect pathway, recipient dendritic cells that have infiltrated the allograft capture and present HLA peptides to alloreactive host T cells. Upon priming by both direct and indirect allorecognition pathways, the alloantigen-cognate T cells become activated and differentiate into effector cells. A subset of activated CD4⁺ T cells, known as extrafollicular helper T cells, migrate toward the B cell follicles [31]. Concurrently, antigen-engaged B cells move to the interface of T and B cell zones to interact with the extrafollicular helper T cells. This interaction triggers the initial T cell-dependent B cell activation, leading to the formation of extrafollicular foci. Within these foci, B cells proliferate and differentiate into PCs, which are predominantly short-lived [32,33].

Figure 1. Humoral alloimmune responses in the context of solid organ transplantation. Adaptive immune responses to alloantigens are initiated by the activation of alloreactive dendritic cells. Within secondary lymphoid organs, B cells that have engaged with alloantigens interact with cognate CD4⁺ T cells at the interface of the T cell zone and the follicle. This interaction prompts the B cells to proliferate and differentiate into either extrafollicular PCs or GC B cells. The latter undergo a mutation-prone proliferation process and eventually give rise to memory B cells and affinity-matured PCs. Whether generated by the extrafollicular or GC responses, most PCs are short-lived. However, a subset migrates to the bone marrow, which provides survival niches. In the bone marrow, they further differentiate into LLPCs. LLPCs can also be found in intragraft tissues, where they may develop in situ within tertiary lymphoid clusters that form in the graft, or they may migrate from secondary lymphoid organs. DC, dendritic cell; SLPC, short-lived plasma cell; eT_FH, extrafollicular helper T cell; T_FH, follicular helper T cell; GC, germinal center; FDC, follicular dendritic cell; SHM, somatic hypermutation; PC, plasma cell; TLC, tertiary lymphoid cluster; LLPC, long-lived PC.

Of the T and B cells activated in the extrafollicular foci, some return to the follicles to initiate germinal center (GC) reactions [34–36]. Within the GCs, B cells undergo a repetitive cycle, moving between the light and dark zones. They experience mutation-prone proliferation in the dark zone before shuttling to the light zone, where high-affinity GC B cells are positively selected under the control of Tfh cells. Ultimately, affinity-matured B cells exit the GC reaction by differentiating into either memory B cells or PCs.

The stepwise differentiation of naive B cells into PCs is governed by a genetic program. The transcription factors PAX5, BCL6, and BLIMP1 are relatively well-established master regulators that delineate the identities of naive B cells, GC B cells, and PCs, respectively [37–39]. These molecules, along with other transcription factors including IRF4, BACH2, and XBP1, form a complex genetic regulatory network that orchestrates the elaborate differentiation program [40–42].

Most PCs generated via the GC reaction are proliferating cells that undergo apoptotic cell death in situ within a week, as these cells appear to possess few survival niches. To extend their lifespan, PCs migrate out of the organs in which they originated, seeking survival niches that are primarily found in the bone marrow (BM) [43,44]. The homing of PCs to the BM is mediated by the CXCL12-CXCR4 axis; BM stromal cells release CXCL12, and PCs expressing the CXCR4 receptor follow the chemokine gradient to their destination. Upon arrival in the BM, they differentiate further into mature antibody-secreting cells and can survive for years or even decades. Consequently, LLPCs, which secrete high-affinity isotype-switched antibodies, are predominantly generated during the GC reaction and are mainly located in the BM. However, recent studies have revealed that a substantial proportion of LLPCs are found in various anatomical locations that provide their own survival niches. These tissue-resident PCs display a transcriptional signature specific to their tissue of origin [45]. Intriguingly, analysis of the transcriptional signatures indicates that the BM contains a diverse collection of PCs from all tissues, suggesting that tissue-resident PCs may not be sessile but instead migrate to the BM.

Antibody-secreting cells can be categorized into at least three subsets: plasmablasts, short-lived PCs (SLPCs), and LLPCs. These subsets appear to represent stages in a maturation continuum, as evidenced by their shared PC molecular markers, which include CD138 (syndecan-1), BLIMP1, XBP1, and IRF4. The initial PC precursors, which arise from either extrafollicular or GC responses, are rapidly proliferating plasmablasts. As they differentiate into mature PCs, they lose B cell markers such as CD19, CD20, B220, major histocompatibility complex class II, and SLAMF6 [46]. These mature PCs continue to proliferate and are turned over within a few days. While most of these cells end their lifecycle as SLPCs, a subset acquires the necessary competence to become LLPCs. These LLPCs can secrete antibodies for decades without requiring antigenic stimulation.

LLPCs are a rare and diverse group of antibody-secreting cells that lack unique markers [47]. While most LLPCs are found in the BM, they do not represent all BM PCs. In naive mice and healthy humans, LLPCs make up less than 25% of the BM PC population, which itself is a minor fraction, comprising only 0.1%–5% of the total immune cell population [46,48,49]. Following deliberate immunization, the proportion of LLPCs in the BM can increase markedly to about 65%, but it never approaches 100% [50]. A previous study has shown that the BM-resident CD38^hi antibody-secreting cell population can be divided into four subsets based on the presence or absence of CD19 and CD138 on the cell surface [48]. While LLPCs have been found to be enriched among CD38^hiCD19⁻CD138^hi cells, this subset includes both LLPCs and SLPCs. In mice, LLPCs were identified within the B220⁻BLIMP1⁺CD138^hi cell fraction, which similarly included SLPCs. Transcriptome profiling and biochemical analyses have revealed several genetic markers that are differentially expressed in LLPCs compared to SLPCs, including BLIMP1, XBP1, CD138, B cell maturation antigen (BCMA), CD28, and MCL-1 (Fig. 2) [44,51–57]. However, these markers are not exclusive to LLPCs and are also present in SLPCs at varying levels. Consequently, the physiology of LLPCs remains incompletely understood.

Figure 2. Characteristics of LLPCs. LLPCs exhibit numerous characteristics that distinguish them from SLPCs. Molecules associated with plasma cell survival and function are upregulated in LLPCs, whereas B cell markers are downregulated. These molecular changes not only contribute to the resistance of LLPCs to traditional cytostatic treatments and B cell depletion therapies but also inform the mechanisms of action for ongoing therapeutic trials. SLPC, short-lived plasma cell; LLPC, long-lived plasma cell; ATG, autophagy gene; BCMA, B cell maturation antigen.

What determines the difference in cell fate between SLPCs and LLPCs, as well as the longevity of PCs? It appears to hinge on a variation in responsiveness to prosurvival signals from the microenvironment, rather than on distinct genetic programs. For extended survival, PCs exit the differentiation site and migrate toward their survival niches, predominantly located in the BM. These niches are structured by BM stromal cells that secrete CXCL12 and attract CXCR4⁺ PCs [58]. Additional niche components include megakaryocytes, eosinophils, and osteoclasts, which produce prosurvival factors such as BAFF, APRIL, tumor necrosis factor, and interleukin-6 (IL-6) [59–61]. These factors are also concentrated in inflamed tissues, drawing CXCR3-expressing PCs that persist until the inflammation subsides [62–64]. Our research, along with that of others, has demonstrated that in certain severe autoimmune diseases in both mice and humans, a substantial proportion (over 30%) of splenic PCs are LLPCs [65–69]. These ectopic LLPC populations are sustained by myeloid-derived suppressor cells (MDSCs) originating from the spleen, which secrete BAFF [70]. Notably, these cells have recently been found to promote Tfh cell development [71]. This cellular network, termed the Tfh-LLPC-MDSC axis, may be pivotal in the in situ perpetuation of autoimmune responses.

Only PCs that respond effectively to these survival factors can persist as LLPCs. The PC surface molecules BCMA and CD138 are well-established factors responsible for this survival. Among the three receptors for BAFF and APRIL, only BCMA, not BAFF-R or TACI, is elevated in BM PCs [56]. The survival of BM PCs is notably compromised by the deletion of BCMA, indicating its critical role in LLPC survival. Supporting this, signals associated with BCMA engagement have been shown to induce the expression of MCL-1, a crucial prosurvival factor for PCs [57]. CD138 is a canonical surface marker of PCs that is upregulated upon maturation [54]. It increases heparan sulfate levels on the PC surface, which in turn promotes IL-6 and APRIL signaling. Consequently, mature CD138^hi PCs gain a selective survival advantage over their less mature CD138^lo PCs counterparts due to this enhanced prosurvival cytokine signaling. Certain adhesion molecules expressed on BM PCs, including VLA4, LFA-1, CD44, and PSGL-1, also contribute to PC survival by interacting with molecules in their niche [72,73].

The acquisition of the molecular competence mentioned above appears to largely rely on the increased activity of transcription factors that are essential for PC development. The transcription repressor BLIMP1 serves as a master regulator of PC differentiation, as its ectopic expression is sufficient to drive this process [74,75]. Utilizing a BLIMP1-GFP reporter system, Kallies et al. [52] demonstrated that LLPCs express higher levels of BLIMP1 compared to SLPCs. The induced conditional deletion of Prdm1, the gene encoding BLIMP1, results in the loss of previously formed PCs, likely due to endoplasmic reticulum (ER) stress-induced apoptosis. This indicates that BLIMP1 is not only necessary for PC formation but also for the maintenance of LLPC populations [53].

The influence of BLIMP1 on PC maintenance appears to be partially mediated by XBP1, which is itself an essential factor for PC development [39,76,77]. BLIMP1 upregulates XBP1 expression by directly suppressing the transcription of Pax5, which in turn derepresses genes, including Xbp1, that are normally repressed by PAX5. The spliced, active form of XBP1 can initiate the unfolded protein response (UPR) in PCs and thereby prevent ER stress-associated cell death. This process is crucial for PC survival due to the high levels of ER stress these cells experience from their extensive antibody production and secretion. ER stress activates the UPR to maintain cell viability and reestablish protein homeostasis. The XBP1 pathway of the UPR primarily serves to facilitate ER expansion, enabling PCs to secrete large quantities of immunoglobulins (Ig). XBP1 induces the transcription of numerous target genes, including those associated with molecular chaperones; this improves protein folding capacity and ER-associated degradation, which ultimately delivers ubiquitinated substrates to proteasomes. Based on our research, one target of XBP1 is the molecular chaperone FKBP13, which plays a pivotal role in the quality control of Ig within the ER [78]. FKBP13 levels are higher in LLPCs than in SLPCs, and inhibiting FKBP13 with rapamycin treatment reduces the PC lifespan. These findings indicate that proteostatic mechanisms involving the XBP1-FKBP13 axis are key to the prolonged survival of LLPCs. Additionally, this work highlights a previously unappreciated effect of rapamycin in the depletion of LLPCs.

A second mechanism implicated in the adaptation of cells to protein stress is autophagy. Autophagy is a natural, conserved process that eliminates unnecessary or dysfunctional components via a lysosome-dependent regulated pathway [79]. Pengo et al. [80] have reported that autophagy is critical for LLPC survival. They showed that the number of LLPCs was significantly reduced in mice immunized with a T cell-dependent antigen after deletion of the autophagy gene Atg5.

PCs have a higher metabolic demand than naive B cells because they require ample energy and nutrient resources to secrete large amounts of glycosylated Ig [81]. As a cell’s metabolic program is closely associated with its differentiation status, it is reasonable to anticipate that LLPCs would exhibit metabolic profiles distinct from those of SLPCs. The results of one study support this notion [82], demonstrating that LLPCs import more glucose than SLPCs, and when faced with metabolic stress, LLPCs—but not SLPCs—redirect glucose metabolism towards the production of pyruvate. Consistent with these observations, the genetic ablation of Mpc2, a gene encoding a component of the mitochondrial pyruvate carrier, resulted in a reduced lifespan for LLPCs. Thus, mitochondrial pyruvate import may represent a critical pathway for the survival of LLPCs.

However, there are some important exceptions to the traditional paradigm describing the identity of LLPCs. First, while this paradigm limits their origin to GC B cells, recent studies have shown that early-wave extrafollicular plasmablasts and B-1 cell-derived PCs can also persist for extended periods in the BM [47,83–85]. Second, not all BM-residing PCs exhibit long-term survival [51]. Third, the longstanding belief that most PCs in the spleen are short-lived is now being questioned, following the identification of substantial numbers of splenic LLPCs in both mouse models and patients with autoimmune diseases [65–69].

Despite these caveats, resistance to traditional therapies remains a defining characteristic of LLPCs. Numerous studies have demonstrated that LLPCs are not eliminated by treatments that target B lineage cells, including irradiation, prednisone, cyclophosphamide, and anti-CD20 antibodies [65,67,86–88]. This resistance, coupled with their sustained antibody production, contributes to the challenge of treating LLPC-mediated pathologies.

LLPCs contribute to AMR by producing a diverse array of alloantibodies. They appear to be the primary source of panel-reactive antibodies, including preformed DSAs, which can result in a positive serum cross-match prior to kidney transplantation. Furthermore, alloreactive LLPCs may emerge after transplantation and continue to produce de novo DSAs over extended periods.

As in autoimmune spleens, alloreactive LLPCs can accumulate in grafted organs. The microenvironment of the transplant organ may induce PCs to migrate to the graft, where they subsequently differentiate into LLPCs. Alternatively, the inflammatory milieu within the graft could enable circulating or tissue-resident immune cells to initiate lymphoid organogenesis within the graft niche. Subsequently, some memory B cells and GC B cells may transform into PCs and ultimately LLPCs. This possibility is supported by a study showing that chronically rejected human kidneys contain tertiary lymphoid foci where GC B cells, memory B cells, and PCs are concentrated [89–92]. Regardless of the initial site of PC formation, the presence of LLPCs in a graft can contribute to organ rejection through the continuous local production of alloantibodies.

Current therapies commonly used to treat AMR include plasmapheresis and intravenous Ig (IVIg). Plasmapheresis works by physically removing antibodies from the circulation. In turn, IVIg has multiple mechanisms of action: it neutralizes the effects of pathogenic antibodies, inhibits complement activation, blocks neonatal Fc receptor to reduce antibody recycling, and interacts with FcγRIIb to send inhibitory signals to immune cells [93]. Although these protocols are considered first-line treatments for AMR, their efficacy is limited [93–97]. Consequently, it is necessary to establish more aggressive protocols that target antibody-secreting cells, rather than simply removing or inhibiting the antibodies produced.

Anti-CD20 therapy has been developed to deplete B cells. Rituximab, a type 1 anti-CD20 monoclonal antibody, is utilized for desensitization in HLA-incompatible and ABO-incompatible transplantation, typically in conjunction with plasmapheresis and IVIg administration [98]. However, in multiple clinical trials, its efficacy has not been demonstrated [99,100]. Other mechanistic studies have revealed that while rituximab depletes circulating B cells, primarily through complement-dependent cytotoxicity, it does not effectively reduce tissue-resident B cells and mature PCs [101,102]. The mechanism of memory B cell resistance is unclear, but a gradual loss of CD20 expression on the surface of PCs during their maturation may underlie the resistance observed in mature PCs, including LLPCs. Notably, rituximab treatment has been found to exacerbate primary immune thrombocytopenia by increasing the number of LLPCs [68]. This suggests that the milieu generated by B cell depletion may actually encourage the development and local tissue settlement of LLPCs, offering insights from an autoimmune disease into the challenges of transplantation. Obinutuzumab, a glycoengineered type 2 anti-CD20 monoclonal antibody, has shown greater efficacy in depleting various B cell subsets, including memory B cells, plasmablasts, and PCs, via antibody-dependent cellular cytotoxicity [103]. Despite its beneficial effects in this context and its efficacy in the treatment of systemic lupus erythematosus, the combination of obinutuzumab with IVIg has not led to a reduction in HLA antibody levels [104]. The presence of CD20⁻ LLPCs likely contributed to the lack of effectiveness of this treatment approach.

Inhibiting proteasomes could be a promising strategy for eliminating LLPCs, as proteasome activity is essential for the survival of all PCs, suggesting that proteasome inhibitors would likely not discriminate between LLPCs and SLPCs. However, the proteasome inhibitor bortezomib did not significantly reduce DSA levels or improve the histological or molecular signs of AMR in certain clinical settings [105,106]. Furthermore, bortezomib therapy has been associated with significant toxicity [107]. Another proteasome inhibitor, carfilzomib, has been shown to be more effective than bortezomib in lung transplant recipients; those who responded to carfilzomib were less prone to experiencing chronic lung allograft dysfunction or progression compared to nonresponders [108]. Nevertheless, concerns have arisen regarding rebound DSA and AMR-driven allograft failure following the discontinuation of treatment. Consequently, despite their theoretical potential, the current evidence seems insufficient to endorse the widespread use of proteasome inhibitors for the treatment of AMR.

Since BCMA is essential for the survival of LLPCs, it is considered an attractive therapeutic target. Strategies aimed at BCMA were initially developed for B cell malignancies and have since been expanded to include desensitization to DSAs in allograft transplants. A monoclonal antibody specific for both BCMA and CD3 has been developed to target BCMA on myeloma cells and PCs, as well as CD3 on T cells [109,110]. This bispecific antibody functions by recruiting and activating cytotoxic T cells to kill BCMA⁺ cells. Although anti-BCMA-CD3 bispecific antibodies have shown promising efficacy in the treatment of multiple myeloma, these antibodies—like other T cell-engaging therapies—may cause adverse effects, including cytokine release syndrome and T cell depletion [111]. To mitigate such adverse effects, a novel monoclonal antibody with low affinity for CD3 (REGN 5458) has been developed and is currently under investigation as a means to decrease HLA antibody levels and improve transplant rates in patients who are highly HLA-sensitized and awaiting kidney transplantation [110].

Chimeric antigen receptor (CAR) T cell therapy has been employed to deplete BCMA⁺ PCs. CAR T cells are autologous T cells that have been genetically modified to express a synthetic receptor specific to a target antigen. Upon encountering their designated antigen, CAR T cells are activated and initiate cytotoxic attacks on cells expressing that antigen. Currently, CAR T cells that react to the pan-B cell antigen CD19 and to BCMA are being employed as secondary treatments for various B cell malignancies, including multiple myeloma [112]. The potential for CAR T cells to deplete LLPCs has also been explored in the field of transplantation [113]. Zhang et al. [114] demonstrated that CAR T cells targeting CD19 and BCMA efficiently eliminated DSAs, memory B cells, and LLPCs and prevented AMR in allosensitized mice. They also observed a clinically meaningful reduction in alloantibodies in a human trial (ClinicalTrials.gov: NCT03549442). These findings provide evidence that LLPCs are a key cellular source of alloantibodies and suggest that immunotherapy targeting LLPCs holds promise for desensitization and/or treatment of allograft rejection.

Because preexisting and posttransplant alloantibodies are major contributors to AMR, it is essential to understand the biology of antibody-secreting cells to inform the development of immunotherapies with curative potential. LLPCs are a mature subset of antibody-secreting cells that can survive for decades without dividing. Emerging through diverse pathways of B cell differentiation, they are found not only in the BM but also in transplanted tissues that provide the necessary survival factors. The capacity of LLPCs to respond to external survival signals and to manage intracellular protein stress allows them to have extended lifespans accompanied by the constitutive production of antibodies. Chronic and refractory AMR may be explained by the resistance of LLPCs to traditional immunosuppressive and B cell depletion therapies. Therefore, developing a method to eliminate alloreactive LLPCs is a critical element in any strategy aimed at preventing AMR.

Conflict of Interest

No potential conflict of interest relevant to this article was reported.

Author Contributions

All the work was done by Eunkyeong Jang and Jeehee Youn. All authors read and approved the final manuscript.