Regulatory B-cell states in NSCLC immunotherapy resistance: mechanisms, spatial context and translational implications Authors: Xin Wang, Xi Li, Jing Zhou, Kai Yang, Rong Hu, Bin Pan Published: 2026 ABSTRACT AbstractNon-small cell lung cancer (NSCLC) remains a leading cause of cancer-related mortality worldwide. Although immune checkpoint inhibitors (ICIs) have improved clinical outcomes, durable benefit is limited by primary and acquired resistance arising from dynamic interactions between tumour cells and the tumour microenvironment. While current models remain largely T cell-centred, accumulating evidence indicates that B-cell states also shape therapeutic response. Here, we propose that regulatory B cells (Bregs) in NSCLC are best understood not as a fixed lineage, but as inducible Breg-associated states that may amplify immunotherapy resistance. We discuss how these states arise under tumour-driven conditions and are further shaped by spatial context, particularly within tertiary lymphoid structures (TLSs) and related immune niches. We further examine how Breg-associated states restrain antitumour immunity and interact with other immunosuppressive populations. This perspective reframes resistance beyond T cell-centric models and supports the development of biomarker-guided, functionally selective, and spatially informed therapeutic strategies.Keywords: Non-small cell lung cancer, Immune checkpoint inhibitors, Immunotherapy resistance, Regulatory B cells, Tumour microenvironment, Tertiary lymphoid structures Introduction Lung cancer remains the leading cause of cancer-related mortality worldwide, with non-small cell lung cancer (NSCLC) accounting for approximately 85% of lung cancer cases [1]. Immune checkpoint inhibitors (ICIs) have improved outcomes in selected patients, yet durable benefit remains limited by both primary and acquired resistance [2, 3]. Although resistance has often been interpreted mainly through tumour-intrinsic immune evasion and T cell-centred dysfunction, growing evidence indicates that broader immune cell states within the tumour microenvironment (TME) also shape therapeutic outcome [4, 5]. Among these, B-cell biology has attracted increasing interest because B cells in the TME can exert both antitumour and immunosuppressive functions, depending on their phenotype, functional state and tissue context [6, 7]. However, the role of B-cell programmes in NSCLC remains incompletely defined. Much of the current evidence is discussed within broader frameworks of B-cell heterogeneity, tertiary lymphoid structures (TLSs) and tumour immune contexture, often drawing on pan-cancer observations rather than lung cancer-specific data [8–10]. This has limited the development of coherent frameworks linking B-cell-centred immune regulation to spatial immune organisation and immunotherapy resistance in NSCLC [11, 12]. In this Review, we argue that regulatory B cells in NSCLC are better conceptualised as Breg-associated functional states than as a discrete fixed lineage [6, 9]. We examine how these states emerge within tumour-associated microenvironments, how they are shaped by the spatial architecture of B cells and tertiary lymphoid structures, and how they might promote immunotherapy resistance through crosstalk with broader immunosuppressive cellular networks [6, 9]. We further discuss the translational relevance of this framework, focussing on pathway-oriented therapeutic opportunities and the development of spatially resolved biomarkers anchored in B-cell and tertiary lymphoid structure biology [6, 9]. By adopting this perspective, this Review aims to connect mechanistic discussion more directly to clinically actionable questions while avoiding overstatement of the degree of validation currently available in NSCLC [9, 11]. Biological characteristics of Breg-associated states Induction of Breg-associated states in the TME In NSCLC, Breg-associated states are likely to arise from tumour-driven and microenvironmental cues that shape infiltrating B-cell states and functions [6, 11]. This process is unlikely to follow a single developmental route. Rather, it arises from the combined influence of soluble mediators, sustained cellular interactions, and local tissue constraints within the TME [13, 14]. Multiple classes of signals are likely to contribute [7]. Cytokines such as IL-10 and TGF-β, produced by tumour cells and surrounding immune populations, may bias B cells towards regulatory function [6, 15]. Persistent antigen exposure, chronic inflammatory stimulation, and checkpoint-rich conditions may reinforce this shift by maintaining B cells in a prolonged activation state that favours suppressive over effector differentiation [7, 16]. In parallel, hypoxia, nutrient limitation, and metabolic stress within the tumour microenvironment may stabilise these states by enforcing persistent immunosuppressive metabolic programmes [17, 18]. Tumour-derived extracellular vesicles, including exosomal cargo enriched in immunoregulatory molecules, may further promote such immune reprogramming both locally and in tumour‑draining lymphoid tissues, including sentinel lymph nodes [19, 20]. Importantly, the induction and functional polarisation of Bregs are embedded within broader suppressive cellular networks, in which myeloid populations may act as active instructors rather than passive neighbours [8, 21]. MDSC-associated indoleamine 2,3-dioxygenase (IDO) promotes tryptophan catabolism and kynurenine accumulation, which can activate aryl hydrocarbon receptor (AhR)-dependent transcriptional programmes in B cells and drive differentiation towards IL-10-producing regulatory states [22, 23]. In addition, myeloid-derived suppressor cells (MDSCs) may enhance the suppressive activity of PD-1+/PD-L1+ Breg-associated states through PD-L1/PI3K/AKT/NF-κB-linked signalling [21]. Tumour-associated macrophages (TAMs) may provide complementary immunoregulatory signals through IL-6, IL-10, prostaglandin E2, and metabolically restrictive tumour-microenvironmental conditions [24, 25]. Regulatory T cells (Tregs) may further reinforce suppressive immune circuits through IL-10, IL-35, and TGF-β, in reciprocal crosstalk with regulatory B cells [8, 26]. Taken together, these observations support a model in which Breg-associated states in NSCLC arise through tumour-educated functional reprogramming [27, 28]. The central biological question is therefore which signals, cellular interactions, and spatial contexts most strongly drive suppressive B-cell induction during tumour progression and under therapeutic pressure [29, 30]. Phenotypic heterogeneity and limitations of marker-based definition Having outlined the mechanisms of induction, the next question concerns the phenotypic consequences of these states. Breg-associated states in cancer display substantial heterogeneity, and NSCLC is unlikely to be an exception [29, 31]. Reported Breg-like populations across tumour settings include transitional-like CD19+CD24^hi^CD38^hi^ cells, memory-like subsets, and populations expressing PD-1, PD-L1, TIM-1, CD5, or CD1d, depending on tissue context, activation status, and analytical platform used [8, 32]. This diversity creates a major interpretive challenge. No single marker or marker combination consistently defines regulatory B cells across tumour types [8, 33]. Moreover, several putative Breg markers are not uniquely specific to Bregs and may also be detected in activated B-cell states, underscoring the context dependence of Breg phenotyping [8, 34]. In inflamed tumours such as NSCLC, where B cells may contribute both to productive immunity and to local immune restraint, phenotype alone is therefore insufficient to define clinically meaningful Breg populations [34, 35]. For this reason, a more useful framework in NSCLC is to define Breg-associated states through the integration of phenotype, suppressive function, and spatial context [34, 35]. Phenotypic profiling can identify candidate populations, but functional evidence is required to establish regulatory relevance [8, 34]. Spatial localisation is equally important, including whether these states arise within TLS-associated regions, stromal immune aggregates, tumour–stroma interfaces, or tumour-draining lymphoid tissues [36, 37]. Such an integrated approach is likely to be more informative than marker-based classification alone for biologically grounded interpretation and clinically relevant biomarker development in NSCLC [34, 37]. Functional mechanisms through which Breg-associated states may restrain antitumour immunity in NSCLC To provide a conceptual overview of Breg-associated immunosuppressive mechanisms in NSCLC, Fig. 1 summarises key pathways—including cytokine-mediated suppression, contact-dependent checkpoint regulation, metabolic constraints, and integration within suppressive cellular networks—highlighting their collective impact on antitumour immunity [8, 35]. Functional mechanisms of Breg-associated states in NSCLC. Schematic representation of the principal pathways through which Breg-associated states may restrain antitumour immunity in the NSCLC tumour microenvironment (TME). A Cytokine-mediated suppression via IL-10, TGF-β, and IL-35, modulating effector T cell responses and antigen-presenting cell function. B Contact-dependent regulation, including PD-1/PD-L1 and additional co-inhibitory interactions. C Metabolic and enzymatic constraints, such as CD39/CD73-mediated adenosine production and nutrient competition, limiting effector immune cell function. D Integration within suppressive cellular networks involving regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs), and tumour-associated macrophages (TAMs), which reinforce the immunosuppressive milieu. TME Tumour microenvironment; Tregs Regulatory T cells; MDSCs Myeloid-derived suppressor cells; TAMs Tumour-associated macrophages. Cytokine-mediated immunosuppression A central mechanism through which Breg-associated states may restrain antitumour immunity is the production of immunosuppressive cytokines [8]. Among these, interleukin-10 (IL-10) is the most widely implicated and remains the best-established functional mediator of regulatory B-cell activity across tumour settings [8]. IL-10 can suppress effector T cell activation, impair antigen-presenting cell function, and reinforce broader immunoregulatory circuits within the TME [8]. Mechanistically, IL-10 may act, at least in part, through STAT3-dependent modulation of antigen-presenting cells, thereby limiting their maturation and co-stimulatory function and weakening support for effective T cell priming [38, 39]. In this way, Breg-associated IL-10 may contribute not only to downstream T cell restraint but also to impaired upstream immune activation [40, 41]. TGF-β represents another major cytokine axis potentially linked to Breg-associated suppression [42, 43]. TGF-β is well known for its ability to inhibit cytotoxic lymphocyte activity, support immune exclusion, and promote regulatory immune differentiation [40, 44]. Within Breg-associated states, this cytokine may extend immunosuppression beyond the B-cell compartment by reinforcing a local milieu that progressively constrains effector immune activation [40, 42]. Although direct attribution of TGF-β production to regulatory B-cell activity in NSCLC is less consistently documented than IL-10, this axis remains mechanistically plausible given the established involvement of TGF-β signalling in lung cancer immune suppression and resistance to immune checkpoint therapy [40, 42]. Other cytokines, including IL-35, may also participate in Breg-associated immune restraint [33, 42]. IL-35 has been linked to the reinforcement of regulatory immune programmes and suppression of Th1-oriented responses across several inflammatory and tumour settings [45, 46]. In NSCLC, direct functional evidence remains limited. Nonetheless, available data and related spatial observations support the possibility that IL-35-associated activity contributes to local suppressive immune niches, particularly in B-cell-rich or TLS-adjacent regions [47, 48]. In NSCLC, however, direct functional validation of these cytokine-producing B-cell states remains limited, and most mechanistic inferences should therefore be interpreted with appropriate caution [35, 49]. Overall, cytokine-mediated immunosuppression appears to represent a major functional dimension of Breg-associated states in NSCLC [8, 35]. However, the strength of evidence supporting individual cytokine mediators in NSCLC varies substantially, and mechanistic interpretations should be aligned with the extent of direct functional validation currently available [8, 35]. Contact-dependent suppression and checkpoint-associated regulation In addition to soluble mediators, Breg-associated states may suppress antitumour immunity through direct cell–cell interactions [31]. This form of regulation is likely to be most relevant in spatially restricted immune niches where B cells localise near activated or dysfunctional T cells and antigen-presenting cells [47, 50]. Under these conditions, regulatory B-cell activity may influence immune responses not only by secreting suppressive cytokines, but also by engaging receptor–ligand pathways that directly constrain neighbouring immune cells [8]. Among these contact-dependent pathways, PD-1/PD-L1 signalling is of particular interest [8, 51]. PD‑L1‑expressing B cells have been described in multiple tumour contexts, and Breg‑associated states with checkpoint‑related phenotypes may contribute to local T‑cell inhibition by reinforcing exhaustion‑like or functionally restrained states [8, 49]. Mechanistically, PD-L1 engagement of PD-1 can recruit SHP2 phosphatase and attenuate proximal T cell receptor signalling, thereby weakening cytotoxic activation and effector function [51, 52]. In NSCLC, where PD-1/PD-L1 signalling is a central determinant of immunotherapeutic response, this mechanism is of particular interest [53, 54]. At the same time, PD-L1 expression on B cells should not automatically be interpreted as evidence of a regulatory state, because it may also reflect activation or inflammatory conditioning [55, 56]. Functional and spatial context therefore remain essential. This distinction is particularly important in NSCLC, where checkpoint-related phenotypes may reflect either suppressive function or ongoing inflammatory activation [35, 53]. Other contact-dependent pathways, including Fas ligand, CD80/CD86-associated signalling, and additional co-inhibitory or co-stimulatory interactions, may also shape the suppressive potential of Breg-associated states [8, 57]. Although the relative importance of these pathways remains incompletely defined in human NSCLC, their possible contribution underscores the importance of interpreting regulatory B-cell activity within local cellular neighbourhoods rather than on the basis of marker expression alone [35, 47]. Taken together, these observations support the view that contact-dependent regulation constitutes a distinct mechanism through which Breg-associated states may restrain antitumour immunity in NSCLC, particularly in immune niches where proximity-dependent checkpoint signalling is likely to be functionally important [35, 49]. Metabolic and enzymatic pathways of immune restraint Breg-associated states may also contribute to immune suppression through metabolic and enzymatic pathways that alter the biochemical conditions of the local TME [58, 59]. This dimension of regulatory activity is important because it extends beyond classical cytokine-based suppression and suggests that B-cell-associated states may help shape the molecular setting in which antitumour immunity either persists or fails [7, 60]. One pathway of particular interest involves ectonucleotidase-mediated purinergic signalling, especially the CD39/CD73 axis [61]. By promoting the extracellular conversion of ATP to adenosine, this pathway can shift immune environments towards suppression, limiting effector T cell and NK-cell activity while favouring regulatory immune function [61, 62]. Mechanistically, adenosine can signal through A2A receptor-dependent cAMP pathways in cytotoxic lymphocytes, thereby suppressing activation, cytokine production, and killing capacity [61]. In tumour settings where B cells express CD39, CD73, or related ectonucleotidase programmes, Breg-associated states may therefore contribute to the formation or maintenance of adenosine-rich suppressive microenvironments [61, 63]. In NSCLC, where adenosine-mediated immune restraint has increasingly been implicated in checkpoint resistance, this pathway is biologically plausible, although the specific contribution of regulatory B-cell subsets remains incompletely resolved [64, 65]. Other metabolic regulators may also intersect with Breg-associated suppression [63]. Enzymatic pathways linked to tryptophan metabolism, redox balance, and nutrient competition may be especially relevant in niches characterised by chronic inflammation, hypoxia, or myeloid accumulation [17, 66]. These processes are not necessarily B-cell-specific, but they provide a useful framework for understanding how Breg-associated states may participate in broader suppressive immune conditions [63, 67]. At present, however, the extent to which these metabolic pathways are directly executed by Breg-associated states rather than by surrounding myeloid or stromal compartments remains to be clarified in human NSCLC [27, 53]. Overall, these metabolic and enzymatic pathways suggest that Breg-associated states may restrain antitumour immunity not only through secreted cytokines or direct cell–cell interactions, but also through local biochemical environments that progressively constrain effector immune function [8, 67]. Integration within suppressive cellular networks Breg-associated states should not be viewed as isolated immunosuppressive units [63]. Rather, they are better understood as components of a broader regulatory network in which myeloid, lymphoid, and stromal signals converge to induce, stabilise, and spatially organise suppressive immune function [63, 68]. In NSCLC, this perspective is particularly important because regulatory B-cell activity is unlikely to arise independently of the surrounding TME [27]. Instead, Breg-associated states are more plausibly sustained through reciprocal interactions with other suppressive populations, including MDSCs, Tregs, and TAMs [27, 69]. Among these interactions, the relationship between MDSCs and Breg-associated states is likely to be inductive rather than merely associative [69]. MDSCs may actively instruct suppressive B-cell differentiation through IDO-dependent tryptophan catabolism, kynurenine accumulation, and downstream AhR activation in B cells [22]. In parallel, MDSC-derived checkpoint-related signals may enhance the suppressive activity of PD-L1+ Breg-associated states through PI3K/AKT/NF-κB-linked pathways [21]. Much of this evidence remains preclinical, and direct validation in human NSCLC is still limited [27, 35]. Even so, these findings support a model in which suppressive myeloid populations act upstream of Breg-associated states and help initiate their emergence within the TME, including in lung cancer-relevant settings [22, 69]. The interaction between Tregs and Breg-associated states is better understood as reciprocal stabilisation than as a one-directional suppressive effect [8, 70]. Breg-associated production of IL-10, TGF-β, and potentially IL-35 may favour Treg induction, maintenance, or functional reinforcement, thereby extending suppression beyond the B-cell compartment [8]. Conversely, Treg-derived suppressive cytokines and checkpoint-linked signals may contribute to conditions that favour the persistence of regulatory B-cell phenotypes, especially in chronically inflamed immune niches in which suppressive feedback loops are already established [70, 71]. TAMs may contribute to this network not only through soluble mediators such as cytokines, chemokines, and prostaglandins, but also by shaping the metabolic architecture of suppressive immune niches [24, 72]. In addition to providing signals that favour regulatory polarisation, TAMs can help establish hypoxic, nutrient-restricted, or adenosine-enriched microenvironments that constrain effector lymphocyte activity while supporting immunoregulatory persistence [73, 74]. Within such niches, Breg-associated states may therefore be maintained not only by direct molecular instruction but also by the broader metabolic conditions that TAM-rich regions help generate [8, 72]. Taken together, these observations support a network model in which Breg-associated states are induced by suppressive myeloid instruction, stabilised through reciprocal interaction with Tregs, and maintained within TAM-shaped metabolic niches [8, 24, 30]. This integrated framework provides a stronger mechanistic basis for understanding how Breg-associated functions may contribute to immune dysfunction in NSCLC [30, 75]. Spatial context and immune architecture of Breg-associated states in lung cancer Spatial distribution across tumour compartments and immune niches The significance of Breg-associated states in NSCLC depends not only on phenotype and function, but also on where they arise within the tumour ecosystem [11, 27]. Lung tumours contain spatially distinct immune compartments, including tumour parenchyma, stromal regions, TLSs, tumour-draining lymph nodes, and circulating immune pools [47, 76]. Breg-associated functions are unlikely to be uniformly distributed across these compartments [27, 76]. Instead, they appear to emerge within specific anatomical niches shaped by local cytokine gradients, cellular density, vascular organisation, and metabolic stress [27, 77]. Within tumour tissue, Breg-associated states may localise to stromal immune aggregates, tumour–stroma interfaces, and B-cell-rich regions surrounding malignant nests [47, 78]. These locations are well positioned to influence antigen presentation, T cell activation, and local checkpoint signalling [79]. Their biological role may therefore differ from that of circulating B-cell populations, which are less constrained by local tissue architecture and may not fully capture tumour-specific suppressive states [27, 80]. This distinction may partly explain why blood-based Breg measurements have shown inconsistent associations with clinical outcomes across studies [8, 80]. TLSs are particularly important in this context. Mature, well-organised TLSs are generally associated with improved antitumour immunity and better response to checkpoint blockade [47, 81]. However, not all B-cell-rich structures are functionally equivalent. Immature TLSs, dysregulated lymphoid aggregates, or stromal B-cell clusters lacking effective immune organisation may instead support regulatory activity [82, 83]. The presence of B cells in a lymphoid niche should therefore not be interpreted as uniformly favourable without functional and spatial resolution [47, 81]. As illustrated in Fig. 2, several molecular mechanisms may help explain how Breg-associated states operate within these spatially restricted immune niches [47, 84]. IL-35-producing Breg-associated states in or near TLS-related compartments may promote Treg expansion and suppress Th1-oriented immunity, thereby shifting local immune balance towards tolerance rather than activation [6, 47]. Likewise, PD-L1-expressing Breg-associated states in TLSs or adjacent stromal regions may inhibit nearby PD-1+ T cells in a proximity-dependent manner, limiting the functional benefit of lymphocyte infiltration [6, 37]. CD39/CD73-mediated adenosine production may further reinforce suppression in hypoxic or metabolically stressed niches, generating local conditions that restrain T cell and NK-cell function [17, 85]. Spatial organisation of Breg-associated states in NSCLC. Schematic representation of the spatial organisation of Breg-associated states in NSCLC. A Mature tertiary lymphoid structure (TLS) in NSCLC, showing the germinal centre, T cell zone, follicular dendritic cells (FDCs), and T follicular helper (Tfh) cells. B Tumour-associated suppressive niche in which Breg-associated B cells may contribute to local immune suppression through PD-L1/PD-1 signalling, CD39/CD73-mediated adenosine production, and interactions with Treg cells and suppressive myeloid populations. TLS Tertiary lymphoid structure; NSCLC Non-small cell lung cancer; FDCs Follicular dendritic cells; Tfh T follicular helper; Treg Regulatory T cell. Tumour-draining lymph nodes may represent an additional site of regulatory B-cell priming or maintenance [8, 86]. Because these structures are central to antigen presentation and immune education, Breg-associated states arising there may influence the quality of subsequent intratumoural immunity [8, 86]. By contrast, peripheral blood may reflect only a partial and diluted view of these functions [47, 87]. Thus, Fig. 2 should be read not as an isolated mechanistic schematic, but as a spatially contextualised model of how Breg-associated states may operate within distinct immune niches in NSCLC. Overall, current evidence supports a spatial model in which Breg-associated states in NSCLC are shaped by, and act within, discrete immune niches rather than as uniformly distributed suppressive cells [47, 88]. Immune-context differences across lung cancer subtypes The role of Breg-associated states is also likely to depend on the broader immune architecture of different lung cancer subtypes [27, 89]. In NSCLC, inflamed tumours with substantial lymphocyte infiltration and organised TLSs are generally more likely to support productive antitumour immunity and to respond to immune checkpoint blockade [11, 90]. However, these same immune-rich settings may also permit the emergence of Breg-associated suppressive activity, particularly when the tumour microenvironment reshapes B-cell function through chronic antigenic stimulation, checkpoint signalling, and local suppressive circuits [27, 91]. Immune infiltration alone therefore does not reliably distinguish effective from suppressive B-cell activity, which depends on B-cell functional state, spatial organisation, and TLS maturity [91, 92]. Direct head-to-head comparisons between lung adenocarcinoma and lung squamous cell carcinoma, particularly in the immune microenvironment, remain relatively limited [93, 94]. Differences in Breg abundance or phenotype reported across studies may reflect variation in stromal organisation, TLS composition, sampling site, or analytical definitions rather than true subtype-specific biology [47, 95]. In the absence of standardised criteria and lineage-specific markers for defining Breg-associated states, it remains difficult to determine whether apparent subtype differences reflect biologically robust variation or instead arise from spatial context, phenotypic plasticity, and methodological differences [29, 96]. The situation is even less clear in small cell lung cancer (SCLC) [97]. Compared with NSCLC, SCLC generally exhibits a less developed adaptive immune milieu and less well-characterised TLS organisation, although immune-inflamed subsets have been described [97–99]. Direct evidence for Breg-associated states in SCLC is currently limited, and it remains unclear whether regulatory B-cell activity in this setting parallels that seen in NSCLC or is instead shaped by a distinct immune contexture [29, 96]. These uncertainties suggest that Breg-associated biology should not be interpreted through histological subtype alone [29, 96]. A more informative framework is to consider subtype together with immune architecture, spatial organisation, and treatment context, which may better clarify when Breg-associated states are biologically relevant and when they could influence therapeutic response [29, 96]. Breg-associated mechanisms of immunotherapy resistance in lung cancer Resistance as a tumour–immune ecological process: where Breg-associated states fit Resistance to immune checkpoint blockade in NSCLC is increasingly understood as an ecological process shaped by dynamic interactions between tumour cells and the surrounding immune microenvironment [100, 102]. Within this framework, Breg-associated states are unlikely to act as universal initiators of resistance; instead, they may function as context-dependent amplifiers that reinforce defective immune priming, sustain inhibitory signalling, and stabilise suppressive cellular networks [29, 96]. This perspective is particularly relevant in tumours that display immune infiltration or PD-L1 expression without durable clinical benefit, indicating that immune cell presence alone does not guarantee productive antitumour immunity [56, 102]. Tumour-driven induction and functional reprogramming during resistance In NSCLC, Breg-associated states may be better viewed as tumour microenvironment-shaped functional adaptations that emerge under suppressive or therapy-resistant conditions [27, 96]. Persistent antigenic and microenvironmental stressors, including chronic inflammation, checkpoint signalling, metabolic stress, and regulatory cytokines, may reprogram infiltrating B cells away from antigen-supportive and TLS-associated functions towards suppressive phenotypes [29, 96]. Tumour-derived extracellular vesicles may further reinforce this process both locally and in tumour-draining lymphoid tissues [20, 103]. The TLS context further illustrates the importance of local immune architecture in shaping adaptive immune responses within the TME [47, 104]. Mature TLSs are generally associated with improved antitumour immunity and better responses to checkpoint blockade. However, B-cell-rich niches are not uniformly protective [95, 96]. Immature or dysregulated lymphoid aggregates may instead sustain regulatory B-cell activity and reduce the functional benefit of immune infiltration [95, 105]. Tumour-driven reprogramming may therefore shift the balance between effector and suppressive B-cell states within spatially defined immune niches during the evolution of resistance [96, 104]. Breg-mediated regulation of antigen presentation and T cell dysfunction Breg-associated states may contribute to immunotherapy resistance by suppressing both upstream immune activation and downstream effector function [8, 106]. Through IL-10-dependent mechanisms, these states may impair the antigen-presenting and co-stimulatory capacity of dendritic cells and other antigen-presenting cells, potentially through STAT3-linked reprogramming, thereby limiting the priming and amplification of antitumour T cell responses [8]. In this way, Breg-associated activity may reinforce tumour-associated defects in immune priming without needing to act as the primary initiating lesion [8, 107]. Breg-associated cytokines and inhibitory ligands may also directly restrain cytotoxic lymphocytes [8]. TGF-β suppresses CD8+ T cell proliferation and effector differentiation while promoting immune tolerance and Treg stabilisation [108]. IL-35 may further weaken Th1-oriented immunity and favour regulatory polarisation [46]. PD-L1 expression on Breg-associated states may sustain PD-1 signalling independently of tumour-cell PD-L1, helping maintain exhaustion-associated T cell states even in immune-infiltrated lesions [49, 106]. Together, these mechanisms may help explain why some NSCLC tumours remain functionally immunosuppressed despite apparent immune engagement and checkpoint blockade [10, 109]. Persistence of suppressive networks under therapeutic pressure Under therapeutic pressure, Breg-associated states may become increasingly important by preserving immune dysfunction despite partial restoration of antitumour immunity [6, 8]. Checkpoint blockade can reactivate cytotoxic T cell responses in some tumour regions, yet this reactivation may remain incomplete where pre-existing suppressive circuits are spatially organised and biologically reinforced [10, 110]. In this setting, Breg-associated states likely matter less as isolated suppressor cells than as components of a network that stabilises resistant immune conditions over time [4, 49]. This network-based interpretation is particularly relevant in NSCLC, where treatment response may depend not only on immune cell abundance but also on whether infiltrating immune populations are embedded within productive or suppressive local architectures [56, 100]. Breg-associated states positioned within suppressive local niches may therefore contribute to persistent immune dysfunction even when PD-1/PD-L1 blockade is pharmacologically active [5, 6]. Their role under treatment pressure is thus best understood in terms of maintenance and amplification rather than initiation [3]. Implications for interpreting resistance in NSCLC These observations suggest that Breg-associated states form part of the broader ecological organisation of resistance in NSCLC rather than a single dominant mechanism [100]. Their relevance likely depends on timing, spatial immune architecture, and the extent of coupling to other suppressive populations [56]. This framework also cautions against interpreting B-cell abundance alone as favourable or unfavourable [56]. Instead, B-cell activity must be understood in context [5]. These considerations provide the rationale for the translational strategies outlined below [5]. Therapeutic strategies and translational challenges Why Breg-associated programmes are therapeutically relevant Breg-associated states are therapeutically relevant because they identify suppressive immune functions that sit at the intersection of lymphoid regulation, myeloid instruction, checkpoint signalling, and metabolic restraint [49, 67]. Their translational importance lies less in absolute abundance than in their ability to mark spatially organised immune niches in which antitumour immunity is functionally weakened [7, 56]. In NSCLC, this shifts the clinical question from whether B cells are present to whether B-cell-rich compartments remain immunologically productive or instead become suppressive [35, 111]. Viewed in this way, Breg-associated biology becomes relevant not only to mechanism, but also to patient stratification and rational therapy design [7, 8]. The principal translational strategies currently relevant to Breg-associated programmes in lung cancer, together with their mechanistic rationale, biomarker implications, and present level of evidence, are summarised in Table 1 [8, 35]. Table 1Translational approaches relevant to Breg-associated programmes in lung cancerStrategyRationaleRelevance to immunotherapyCandidate biomarkers/readoutsClinical exampleEvidence levelReferencesBroad B-cell depletionReduces tumour-associated B cells, but also depletes protective B-cell and TLS-related compartmentsMay attenuate suppressive B-cell activity, but risks loss of beneficial antitumour immunityBulk CD20+ B-cell abundance; limited functional specificityNo clearly established biomarker-guided strategy in NSCLCConceptual / preclinical[6, 35]Surface-marker targetingSeeks to enrich for suppressive B-cell states using putative regulatory markersMore selective than pan-B-cell depletion, but constrained by phenotypic instabilityTIM-1-related phenotypes and other candidate Breg markersNo established clinical programme in NSCLCPreclinical / translational[8, 35]Cytokine modulationTargets suppressive mediators associated with Breg-related functionMay reduce immune suppression without depleting the broader B-cell compartmentIL10, TGFB1, and EBI3/IL12A-related programmesBiomarker-guided combination approaches remain investigationalTranslational[8, 27]PD-L1–TGF-β co-targetingTargets convergent suppressive pathways linked to immune exclusion and regulatory nichesMay be relevant to primary or adaptive resistance to immune checkpoint blockadePD-L1 expression, TGF-β-high immune-excluded niches, and TLS maturityBintrafusp alfa in advanced NSCLCTranslational / clinical[8, 35]Adenosine-axis targetingDisrupts CD39/CD73-mediated metabolic suppression within the TMEMay complement immune checkpoint blockade in suppressive immune nichesENTPD1/CD39, NT5E/CD73, and adenosine-rich spatial nichesPACIFIC-9Translational / clinical[8, 35]IDO / AhR-linked modulationInterferes with myeloid–B-cell regulatory circuits that may sustain suppressive immune statesMay be relevant in adaptive resistance settingsIDO1, kynurenine/AhR-related signatures, and myeloid–B-cell couplingIO102-IO103 plus pembrolizumabEarly clinical / translational[8, 27, 35]Spatial biomarker stratificationDistinguishes immune-active TLSs from suppressive B-cell-rich nichesMay improve patient selection for pathway-directed combinationsTLS maturity, GC/FDC features, and suppressive spatial neighbourhoodsVisium/Visium HD, Xenium/CosMx, GeoMx DSP, mIF/mIHCPlatform-enabling[27, 35, 112]Breg Regulatory B cell; TLS Tertiary lymphoid structure; NSCLC Non-small cell lung cancer; GC Germinal centre; FDC Follicular dendritic cell; DSP Digital spatial profiling; mIF Multiplex immunofluorescence; mIHC Multiplex immunohistochemistry; Ref. representative reference(s) supporting each translational strategy Translational approaches relevant to Breg-associated programmes in lung cancer Breg Regulatory B cell; TLS Tertiary lymphoid structure; NSCLC Non-small cell lung cancer; GC Germinal centre; FDC Follicular dendritic cell; DSP Digital spatial profiling; mIF Multiplex immunofluorescence; mIHC Multiplex immunohistochemistry; Ref. representative reference(s) supporting each translational strategy Why indiscriminate B-cell depletion is unlikely to be the optimal strategy Despite the therapeutic relevance of regulatory B cells, broad B-cell depletion is unlikely to represent an optimal therapeutic strategy [113, 114]. B cells can support antitumour immunity through antibody production, antigen presentation, cytokine secretion and participation in TLSs, which are frequently associated with favourable responses to immunotherapy [6, 9]. Non-selective B-cell depletion could therefore eliminate both immunosuppressive and beneficial B-cell subsets, with uncertain overall consequences for tumour control [7, 9]. This concern is particularly relevant in NSCLC, where organised (mature) TLSs have been repeatedly associated with enhanced immune responsiveness and favourable responses to immunotherapy [81, 115]. A lineage-level B-cell depletion strategy might attenuate regulatory B-cell activity in some tumour contexts, but could also compromise beneficial B-cell-mediated and TLS-associated lymphoid immunity in others [7, 36]. Accordingly, translational strategies should prioritise selective targeting of Breg-associated suppressive programmes, signalling pathways, or tumour immune niches rather than indiscriminate B-cell depletion [8, 116]. This view is also consistent with the notion that regulatory B-cell activity in NSCLC is context dependent and may reflect inducible functional states rather than a rigid lineage-defined entity [8, 35]. Pathway-directed therapeutic options A more promising translational approach may be to selectively target pathways that mediate Breg-associated immunosuppression, rather than relying on indiscriminate B-cell depletion [8, 116]. From this perspective, the therapeutic relevance of Breg-associated biology lies in identifying selectively targetable immunosuppressive pathways or circuits that may be exploited in combination with existing immunotherapies [8, 116]. One major therapeutic option is to target cytokine- and checkpoint-associated pathways that sustain Breg-linked immunosuppressive activity [8, 116]. In tumours where Breg-associated immunosuppressive states are linked to persistent T cell dysfunction, combinatorial strategies involving PD-1/PD-L1 blockade combined with the targeting of TGF-β-associated immune exclusion or other suppressive cytokine pathways may be mechanistically justified [8, 117]. A second avenue is metabolic targeting, particularly in tumours with CD39/CD73-high adenosine-rich microenvironments in which Breg-associated states are likely to be embedded within broader suppressive immune niches [64, 116]. A third strategy is to target upstream pathways that may sustain suppressive B-cell reprogramming through myeloid instruction [22, 106]. In this context, interventions linked to IDO, kynurenine, and AhR signalling may be relevant in tumours showing strong myeloid–B-cell coupling [22, 118]. Likewise, settings marked by macrophage-dominant suppressive architecture may justify consideration of approaches that disrupt TAM-supported regulatory pathways or the metabolic conditions that help maintain suppressive niches [24, 119]. These strategies are best viewed as mechanism-guided rather than Breg-exclusive interventions. Even so, they offer a more selective and biologically coherent framework than pan-B-cell depletion [116, 120]. Where available, early-phase combination strategies targeting TGF-β-, adenosine-, or related suppressive axes may provide clinically informative settings in which to test this framework in NSCLC [121, 122]. Biomarker strategy: from marker panels to composite spatial readouts A major translational barrier is the absence of a universally accepted definition of Breg-associated states in lung cancer [120, 123]. Marker panels alone may not fully discriminate regulatory B cells, as current Breg-specific markers remain under investigation and Bregs display marked phenotypic heterogeneity [124, 125]. Clinically useful biomarker strategies will therefore need to move beyond simple phenotyping towards composite readouts that integrate phenotype, suppressive function, spatial location, and treatment timing [102, 125]. In practical terms, Breg-relevant biomarker models in NSCLC should combine several layers of information [102, 126]. These may include TLS maturity, B-cell functional state, local expression of suppressive cytokine or checkpoint activity, and spatial proximity to exhausted T cells, Tregs, TAMs, or MDSC-rich regions [9, 78]. Candidate readouts may therefore include IL10-, TGFB1-, EBI3/IL12A-, CD274-, ENTPD1-, and NT5E-related features, interpreted in the context of local immune architecture rather than in isolation [9, 124]. Such composite models are more likely to capture suppressive immune niches than single-marker approaches [9, 78]. This shift in emphasis is important because it reframes the translational question. Rather than asking whether a tumour contains “Bregs”, the more useful question is whether it contains a spatially organised suppressive B-cell state likely to influence response to immunotherapy [9, 29]. Spatial and platform strategy for translational development Spatially resolved approaches are likely to be essential for translating Breg-associated biology in NSCLC [78, 127]. The biological significance of a suppressive B-cell state depends not only on phenotype, but also on whether it is located within a mature TLS, an immature lymphoid aggregate, a stromal suppressive niche, or the tumour–stroma interface [9, 128]. Discovery and validation pipelines should therefore be built around spatial immune architecture rather than bulk immune cell abundance alone [78, 129]. A staged platform strategy may be especially useful [130, 131]. Discovery-oriented studies can use spatial transcriptomic approaches such as Visium or Visium HD to map candidate B-cell-rich suppressive niches and identify pathway-level features associated with resistance [132, 133]. Imaging-based single-cell platforms such as Xenium or CosMx can then define cellular neighbourhoods and proximity-dependent interactions among B cells, PD-1+ T cells, Tregs, macrophages, and other myeloid populations [78, 134]. These findings can subsequently be translated into larger, clinically annotated cohorts using more scalable spatial assays, including multiplex immunofluorescence, multiplex immunohistochemistry, and digital spatial profiling [135, 136]. Such a workflow can also support trial design [78, 137]. Tumours with TGF-β-high immune-excluded regions may be more suitable for cytokine- or stroma-directed combinations, whereas CD39/CD73-high adenosine-rich microenvironments may be more relevant to adenosine-axis targeting [138, 139]. Likewise, tumours characterised by strong IDO/AhR-linked myeloid–B-cell coupling may represent settings in which broader immunoregulatory combinations deserve evaluation [22, 140]. Although these approaches remain exploratory, they offer a practical route for translating Breg-associated biology into patient stratification and more rational combination immunotherapy development in NSCLC [6, 141]. Conclusion Regulatory B-cell biology in NSCLC is increasingly understood not as a lineage-defined entity, but as a set of inducible, context-dependent suppressive states shaped by the TME [6, 142]. This perspective helps reconcile the marked phenotypic heterogeneity of B-cell populations across lung tumours and shifts attention from static marker-based definitions towards integrated interpretation based on function, spatial organisation, and immune context [9, 76]. Within this framework, Breg-associated states are unlikely to act as isolated drivers of immune dysfunction [6, 143]. Rather, they appear to participate in broader suppressive cellular networks involving MDSCs, Tregs, TAMs, and metabolically restrictive immune niches [143, 144]. Their relevance in NSCLC therefore lies not simply in their presence, but in the extent to which they are embedded within local architectures that impair antigen presentation, restrain cytotoxic lymphocytes, and sustain resistance to immune checkpoint blockade [145, 146]. At the translational level, these observations argue against indiscriminate B-cell depletion and instead support mechanism-guided strategies directed at cytokine, checkpoint, metabolic, or myeloid-linked suppressive circuits [147, 148]. They also highlight the need for composite biomarker models that combine phenotype, function, and spatial context rather than relying on single-marker definitions of regulatory B cells [35, 112]. In this regard, spatially resolved transcriptomic and imaging-based platforms are likely to be especially important for identifying clinically meaningful B-cell states and their suppressive niches in NSCLC [88, 112]. Nevertheless, mechanistic validation in human NSCLC remains limited for many proposed B-cell–associated pathways [35]. Future progress will depend on integrating mechanistic studies, spatial profiling, and clinically annotated cohorts to define when Breg-associated states are most relevant, how they interact with other suppressive populations, and which patient subsets may benefit from pathway-directed therapeutic strategies [35, 112]. A more precise understanding of these context-dependent suppressive B-cell states may ultimately refine biomarker development and support more rational combination immunotherapy in lung cancer [8, 149]. Author contributions XW: Writing – original draft, Investigation, Visualization. XL: Writing – original draft, Investigation, Data curation. JZ: Writing – original draft, Methodology, Validation. KY: Conceptualization, Supervision, Writing – review & editing, Project administration. RH: Conceptualization, Supervision, Writing – review & editing. BP: Conceptualization, Supervision, Writing – review & editing. All authors contributed to the article and approved the submitted version. Funding This work was supported by the Sichuan Medical Association Research Project (No. S2024024), the Scientific Research Project of Sichuan Administration of Traditional Chinese Medicine (No. 26MSZX573), and the Scientific Research Project of Sichuan Medical and Health Promotion Association (No. KY2025SJ0052). Data availability No datasets were generated or analysed during the current study. Declarations Conflict of interest The authors declare that they have no conflict of interest. Generative AI statement During the revision of this manuscript, the authors used a generative AI-based language tool to assist with English-language editing, including grammar, spelling, punctuation, wording consistency, and readability. The scientific framing, literature selection and interpretation, figures, tables, and conclusions were developed and verified by the authors. All AI-assisted edits were critically reviewed and revised where necessary by the authors, who take full responsibility for the integrity, accuracy, and final content of the manuscript. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xin Wang, Xi Li and Jing Zhou have contributed equally to this work and should be considered co-first authors. Contributor Information Kai Yang, Email: yangkai@cmc.edu.cn. Rong Hu, Email: 15882193236@163.com. Bin Pan, Email: 13348886545@163.com. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.