Age-associated B cells and double-negative B cells: two sides of the same coin? The answer depends on the context

B cells are central players in adaptive immunity, orchestrating humoral responses through antibody production, antigen presentation, and cytokine secretion. During aging, cumulative immunological changes reshape the Bcell compartment in a process termed immunosenescence, which has profound consequences for infection control, vaccine efficacy, and susceptibility to autoimmunity (1,2). Among the most striking features of B-cell remodeling is the expansion of atypical B cell subsets, such as ageassociated B cells (ABCs) and double-negative (DN) B cells. Both subsets have become focal points of intense research, because of their enrichment in elderly individuals, their expansion in autoimmune diseases, and their presence in chronic infections (3)(4)(5)(6). Yet, their precise lineage relationships remain controversial. Some evidence suggests that DN2 cells, which refer to an extrafollicular DN sub-population (7), may overlap functionally and transcriptionally with ABCs, raising the central question: are ABCs and DN B cells two sides of the same coin, or distinct entities? Clarifying their identity is complicated further by nomenclature inconsistencies across species, tissues, and disease contexts (7)(8)(9). This manuscript synthesizes current knowledge on ABCs and DN subsets, examining their biology in aging, autoimmunity, and infection, with a focus on their transcriptomic signatures and therapeutic relevance. Taking into account these data, we provide an opinion on whether ABCs and DN B cells should be considered identical populations or two distinct B cell subsets.ABCs were first identified in aged mice, as CD19⁺CD21⁻CD23⁻ B cells expressing the transcription factor T-bet and the integrin CD11c (5). Their percentages in young healthy individuals are low but steadily increase with age (5). In healthy aging, a subset of ABCs exhibits reduced BCR-mediated antibody production, consistent with aspects of immunosenescence, though these cells retain other immune functions (2). However, their exact role is not yet completely understood.ABC populations were also reported to expand in autoimmune diseases, as well as in chronic-infections (9,10). In the context of autoimmunity, T-bet expression in B cells is elevated, leading to increased antibody production, enhanced antigen presentation to T cells and also formation of germinal centers (GCs), all drivers of immunological reaction (11). In the context of infectious diseases, on the other hand, ABCs' role is less clear. It seems that these cells contribute to pathogen clearance, thus display a protective role (12).Up to this day, the origin of ABCs remains a mystery. It has been shown, though, that these cells express a diverse Ig repertoire , portrayed by somatic hypermutations and antigen-driven activation (2). Based on these features, a GC-experienced origin for at least a subset of ABCs has been proposed, although alternative hypotheses cannot be excluded. For instance, formation through homeostatic proliferation is also a potential route leading to the generation of ABCs (2). Moreover, it's worth mentioning that B cell activation involves, not only GC reactions, but also extrafollicular and other GC-independent pathways (6).ABCs exhibit heightened responsiveness to endosomal Toll-like receptor (TLR) signals, particularly TLR7 and TLR9, and can be driven to differentiate by TLR stimulation in combination with IFN-γ and/or IL-21, depending on the experimental context. (2,9,(13)(14)(15). IL-21 robustly induces CD11c expression, contributing to the ABC phenotype, whereas IFN-γ primarily promotes T-bet expression, reflecting context-dependent effects reported in primary studies (2,14,15). BCR signaling can also contribute to ABC activation, in conjunction with TLR, IFN-γ, IL-21, and/or CD40 signals, although it is insufficient to drive differentiation on its own (11,16).Of note, TLR7 is an X linked transmembrane receptor, closely related to ABC activation, clarifying the higher rate of autoimmunity onset in female patients (4).As far as transcriptomic profiling of ABCs is concerned, data derived from mice has shown that ABCs are a unique sub-population of B cells, discrete from B1 and FO B cells. In detail, ABCs highly express CD11c and Tbet, along with transcripts of immunoglobulin heavy chain and CD138 (13). Intermediate expression of transcription factors involved in plasma cell differentiation has also been reported, thus indicating that ABCs are probably plasma cell precursors (13). In the context of autoimmunity, transcriptomic analyses of ABCs have revealed that IL-21 inducible genes, as well as genes associated with cell adhesion, are strongly up-regulated (9).In systemic autoimmune diseases, such as systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), Sjögren's syndrome (SS) and systemic sclerosis (SSc), ABCs expand prematurely and -in some cases -correlate with disease activity, autoantibody titers, and organ involvement (4,17,18). More specifically, T-bet+ B cells (which are considered as ABCs) are expanded in SLE patients and correlate with disease activity index and lupus nephritis, indicating that these cells can be used as potential biomarkers for the disease (19). In SSc, ABC-like cells were found to be expanded and contribute to vascular complications of the disease (18). In addition, in the case of multiple sclerosis (MS), ABCs expand and exhibit a significantly upregulated inflammatory cytokine profile, in terms of mRNA expression (20). Beyond autoimmunity, chronic infections such as HIV, hepatitis C and malaria, induce similar ABC expansions, linking persistent immune activation to their emergence (10,21).Regarding ABCs, it is important to mention that specific targeting of this population of B cells can alleviate symptoms in various autoimmune diseases, including SLE, RA and MS (22). Currently therapeutic interventions targeting BAFF (e.g., belimumab) and/or CD20 (e.g., rituximab) reduce ABC frequencies in the human blood of SLE patients, indicating their implementation in autoimmunity pathogenesis (4,23). Additional important ABC reducing therapies include an extended portfolio of IRF5 inhibitors (genetic and/or chemical inhibition), JAK inhibitors (such as baricitinib), triggering of adenosine receptor 2a (A2a) via A2a agonists, TNF inhibitors, administration of tocilizumab (antagonist of IL-6 receptor) and ROCK-kinase inhibitors (such as fasudil) (19,22).Double negative (DN) B cells are a heterogeneous population, defined by the absence of IgD and CD27 markers (6). These cells are expanded in the elderly and, moreover, seem to showcase different immune functions in different pathological conditions (refers to infectious diseases and autoimmunity) (6). For instance, in case of HIV infection and/or malaria, a proportion of DN B cells displays an exhausted phenotype, while in SLE cases these cells are considered as the main source of autoantibody secretion (6,7,24). DN B cells are comprised of at least four subsets and are categorized into subgroups, based on the expression of CXCR5 chemokine, CD11c integrin and transcription factor T-bet: DN1 (CXCR5 + CD11c -T-bet -), DN2(CXCR5 -CD11c + T-bet + ), DN3 (CXCR5 -CD11c -T-bet low ), and DN4 (CXCR5 + CD11c -T-bet -) (6,25,26). DN1 cells are enriched in healthy elderly individuals and are primarily associated with immunosenescence (8). Of note, DN1 appear relatively non-pathogenic and lack a strong T-bet-driven phenotype (6,8,25,26). DN2 cells, on the other hand, are highly responsive to TLR7 and have been strongly associated with extrafollicular plasmablast differentiation, particularly in inflammatory settings, and are strongly linked to active SLE, especially in African American women with lupus nephritis (7). Interestingly, DN2 is considered to be the only DN subset highly expressing Tbet (6,7,25,26). Initially, it was presumed that these cells were lupus specific (7). Subsequently, however, their presence was also confirmed in other rheumatic conditions, including rheumatoid arthritis (27). As far as DN3 cells are concerned, this subset constitutes a newly recognized DN subset that has been linked to extrafollicular immune activation and has been associated with severe COVID-19 and hypoxic conditions (6,26). In addition, the population has also recently been implicated in autoimmune fibrosis in the context of IgG4-related disease, where it infiltrates inflamed tissues (28), as well as in SLE, where it shows a significant correlation with disease activity (29). Lastly, DN4 cells remain a poorly defined population. According to some studies, though, the aforementioned DN cells are closely related to allergic reactions (6,25,26). Overall, although DN subsets -particularly DN2 and DN3 -are strongly associated with extrafollicular immune responses, their developmental trajectories are likely context-dependent and may vary across disease settings (6,7,25,26).Transcriptomic analyses confirm that DN2 cells share common features with ABCs, including T-bet and CD11c expression (7), although these similarities may vary depending on the inflammatory or pathological context. In contrast, DN3 cells display a distinctive signature, characterized by a strong signature of proliferation and unfolded protein response, along with lowestamong DN subsets -levels of CD22, CD72, CD69 and BAFFR expressions (26,28). DN1 are found in healthy elderly individuals, while DN4 seem to be strongly associated with allergies (8,25,26). The former have been shown to transcriptomically resemble memory B cells, while the latter express genes linked to the Notch signaling pathway and protein ubiquitination, and thus are distinguished from DN1 (28). However, both two sub-types express CXCR5⁺ chemokine, in contrast to DN2 and DN3 cells (6). Interestingly, DN2 and DN3 cells are those that are mostly associated with autoimmune diseases (7,28,29). In total, these differences in function and transcriptomic characterization suggest that every DN sub-population is a complete discrete entity (30).Regarding disease involvement, DN B cells, similar to ABCs, expand in systemic autoimmune diseases and infections (6, 7, 10, 20, 24, ). For instance, in MS, both ABCs and DN have been reported to expand and exhibit some pro-inflammatory characteristics (31). In SLE, furthermore, both populations increase in numbers and seem to drive disease pathogenesis via differentiating into plasma cells and producing autoantibodies (7,9). In infectious diseases, such as COVID-19 and malaria, DN expansion has also been reported and linked to poor clinical outcomes (32,33). Importantly, these observations suggest that the functional relationship between ABCs and DN B cells may depend on the specific inflammatory or pathological context.ABCs and DN B cells share many similarities, and are therefore often considered related populations (4). Both subsets have been reported to expand in elderly individuals (5, 8), during infections (10,24), and in autoimmune diseases (7,9). Regarding immunophenotype, they exhibit overlapping features, including expression of CD11c and/or T-bet (6,7,9). Moreover, both populations display similar activation requirements, which involve IFN-γ, IL-21, and TLR7 or TLR9 signaling (7,9,13,15). Supporting this parallel, pharmacological agents that reduce ABC frequencies in human blood also appear to decrease DN B cell percentages (22). Notably, in both healthy individuals and patients with lupus, ABC frequencies correlate with DN B cell frequencies (19,29). Collectively, these findings indicate that ABCs and DN B cells are highly similar at first glance. However, the extent of these similarities may vary depending on the inflammatory or pathological context, highlighting that their functional and transcriptomic relationship is not absolute.Beyond antibody secretion, ABCs have been shown to produce proinflammatory cytokines and/or release chemokines, serve as antigenpresenting cells with strong phagocytic capacity, and also contribute to T-cell activation (34). Similarly, DN B cells -particularly the DN2 subset -exhibit comparable effector functions: upon stimulation, they produce proinflammatory cytokines, express antigen-presentation markers such as HLA-DR and CD86, and can activate T cells (35,36). These findings indicate that both ABCs and DN B cells are not merely antibody precursors, but can perform broader immune functions, with their activity often depending on the inflammatory or pathological context.Despite their similarities, ABCs differ from DN B cells in several key aspects. Some ABCs highly express the memory marker CD27 (13), which is absent from DN B cells (7,8). Although the majority of ABCs are classswitched B cells, predominantly expressing IgG or IgA, single-cell RNA-seq analyses reveal that this population also contains unswitched IgD+ cells (37). Furthermore, DN1, DN3 and DN4 subsets (which are less well-characterized than DN2) lack expression of CD11c and T-bet (25,26,30), two defining features of ABCs (9). With respect to autoantibody production, a hallmark of ABC function (2,13), only DN2 cells among DN B cell subsets efficiently differentiate into plasma cells (7). Regarding their origins, ABCs are commonly thought to include GC-experienced cells, although alternative developmental pathways have been proposed (2,5,38), whereas DN B cells have been strongly associated with extrafollicular differentiation, particularly under inflammatory or autoimmune settings (7,29,32), suggesting that their relationship with ABCs may be context-dependent. Lastly, according to a comparative transcriptomic analysis, ABCs are distinct from other CD11c + B cell populations, such as DN2, as they display an elevated expression in multiple cytokines and chemokines, which are not detected as increased in the other CD11c + subsets (39).In recent literature, the term "ABCs" refers to murine B cells, whereas in humans the "DN B cells" -particularly DN2 cells -are regarded as their corresponding counterparts (30,40,41). This scenario is plausible, although ABC cells (or at least ABC-like cells) have also been reported in humans (9,10,13,18,19,31). ABCs represent a heterogeneous B-cell population (42), and several related circulating and/or splenic subsets have been described in both mice and humans (43). Similarly, DN B cells are heterogeneous, with four discrete subsets identified to date (25,26). Considering the resemblances and differences between ABCs and DN B cells (Table 1), we propose that one ABC subset -lacking IgD and CD27 expression -closely corresponds to DN2 cells, though this relationship may be influenced by the specific inflammatory or pathological environment. Memory (CD27+) and phenotypically naïve-like (IgD+) ABCs cannot be classified as DN cells (7,8,13,37), and conversely, DN B cells lacking CD11c and T-bet cannot be classified as ABCs (9,25,26,30). Notably, IgD+ ABCs are not considered naïve cells, as their population carries somatic hypermutations and is clonally related to IgD-cells, indicating prior antigen experience (39). As consensus has not yet been reached regarding the optimal immunophenotypic markers defining ABCs, we suggest that this specific B-cell population should be regarded not only as murine, but also as human, being closely related to DN2 cells (7). Interestingly, while Knox JJ et al. (2025) represent an important refinement of murine ABC definitions (44), further observations across studies highlight ongoing heterogeneity of ABCs. In more detail, Knox et al. report relative resistance of ABCs to anti-BLyS and anti-CD20 depletion in murine lupus models (44), whereas Ramsköld D et al. (2018) and Faustini F et al. (2022) found sensitivity of ABCs to these same interventions, in SLE patients (23,45). Clearly, such observations further highlight the continuing need for ABC characterization in both human and murine systems. Similar to the case of ABCs, the optimal characterization of DN B cells is of utmost importance. This need is reflected, for example, in the observation that a DN B cell population with cytoplasmic FOXO1 has been identified in patients with SLE (46). However, it is still unknown whether this population corresponds to any known DN subset observed in SLE patientssuch as DN2 or DN3 -or whether it represents another distinct population (6,7,29,46). Of note, FOXO1 is a transcription factor that plays a key role in B cell development (47).ABCs constitute a heterogeneous population of B cells, comprising CD27+ B cells, IgD+ B cells and -predominantly -IgD-CD27-(DN) B cells (48). Compelling evidence from immunophenotypic, functional, and transcriptomic analyses indicates that ABCs with an IgD⁻CD27⁻ phenotype are closely related to the DN2 B cell subset, which has been well-characterized in lupus (7,49). Notably, the extent of this relationship may vary depending on the inflammatory or pathological context.