Systemic lupus erythematosus (SLE) is an autoimmune disorder characterized by a deleterious immune response that affects several tissues and organs, including kidneys. Lupus nephritis (LN) is one of the most common manifestations of this chronic disease, constituting a poor prognosis feature that increases morbidity and mortality (1–3). Over the years, different factors that contribute to LN pathogenesis have been identified being one of the most relevant, the deleterious function of B cells, as it generates hyper-reactive plasma cells that produce autoantibodies that can cause permanent damage within the glomerular, vascular, and tubulo-interstitial compartments of the kidneys, leading to acute or chronic renal failure (4). Nevertheless, the role of B cells in SLE beyond their differentiation to plasma cells and subsequent autoantibodies secretion is still unclear (5).

Analyzing specific circulating B cell subsets to identify alterations in their frequency or phenotype is one of the main approaches that would allow identifying possible roles of this lymphocyte lineage and their specific subpopulations in autoimmune diseases (6). Accordingly, the discovery and characterization of the Aged-associated B Cells (ABCs) subset has been relevant by its association with pro-inflammatory contexts. ABCs have been found numerically altered in conditions of chronic inflammation that encompass autoimmune diseases (SLE, rheumatoid arthritis, systemic sclerosis, etc.) or chronic infection diseases (HIV infection, malaria, hepatitis C virus infection, etc.) (7–10). This rare B cell subset has been defined by the expression of the α-integrin subunit CD11c and the transcription factor T-bet, as well as a low-level expression of the complement receptor type 2 (CR2)/CD21 (11). Regarding their functional properties, ABCs are highly responsive to innate stimuli such as TLR-7 ligands; they may produce inflammatory cytokines (including TNF-α and IFNγ) and block the generation of conventional B cells progenitors by inducing apoptosis through the expression of high levels of surrogate light chain (11–13). Furthermore, upon activation, these cells can rapidly expand, differentiate and produce antibodies (mainly IgG); in fact, higher frequencies of circulating ABCs have been correlated with elevated titers of autoantibodies in patients with SLE (14).

Although ABCs features are particularly well described in mouse models, there are many unsolved questions about their implication in human autoimmune diseases, including SLE, particularly how they are linked to the evolution of the disease, their impact on patients’ treatments, or if alterations in its phenotype could be used as prognostic markers. Consequently, we analyzed the ABCs subset and similar B cell phenotypes in cohorts of SLE patients exhibiting or not LN. We found that patients with LN present low or null frequencies of the atypical “ABC-like” CD11c⁺T-bet⁺CD21^hi B cell subset, in contrast with healthy individuals and SLE patients without LN. This subpopulation rendered significant correlations with different features of SLE. Interestingly, both ABCs and CD21^hi subsets presented alterations in kidney disease patients after induction therapy, making them attractive candidates to constitute markers that could define patients’ status or prognosis in LN onset.

One of the main factors contributing to SLE development and its clinical manifestations are B cells. Several reports concerning these cells have shown their functional, phenotypic, or activation-related alterations. Therefore, the analyses of different B cell subsets have been useful for understanding the pathogenesis of SLE; among them, the T-bet-expressing ABCs represent one of the most intriguing examples. Previous studies in infectious or immune-mediated diseases have evaluated the T-bet transcription factor, integrin CD11c and complement receptor CD21 in these B cells, as molecules associated with aberrant cell activity. However, the relevance of these cells in the pathological, clinical, and therapeutic contexts during autoimmune disorders such as SLE is still not fully clarified.

According to our analysis of ABCs subset, the presence of these cells was detected in our three study groups. In this regard, it has been previously reported that ABCs are increased in lupus patients compared to healthy controls, with a significant difference even more marked when they develop LN (14, 20, 21). As mentioned, our cohorts do not to display that trend, as we could not detect any difference among healthy subjects and lupus patients without active lupus nephritis. Additionally, the LN group of patients only showed a moderate significant increase compared to healthy donors but no difference with non-LN lupus patients. Of course, this trend could be the result of the limited number of recruited patients in our study but also, it could suggest that ABCs are not necessarily as robust for segregating SLE and LN outcomes. Another possible explanation could be that non active disease is associated with absence of a proinflammatory state that could be the main trigger for the increase in ABCs subpopulation.

Prior reports aimed to understand the abnormal activity of B lymphocytes in lupus have detected alterations of ABCs in peripheral blood. The first report to describe a disruption in a similar B cell compartment described an incremented frequency of a CD19⁺ CD21^lo/- B cell subset in patients with common variable immunodeficiency (CVID) and SLE (22, 23). Upon this observation, several reports emerged describing and expanding this cell phenotype (including its T-bet expression), besides its association with different chronic infectious diseases and autoimmune disorders (7–10). Despite that, conflicting data about B cells’ CD21 expression in autoimmunity contexts have gone unnoticed. One of the few related approaches was reported by Dash R. et al., which detected an increase in CR2 (CD21) transcript levels in PBMCs of rheumatoid arthritis patients that negatively correlates with disease activity (24). Regarding SLE, a decrease in the frequency of total CD19⁺ CD21⁺ B cells was previously reported (25). In our present report, the absence of the CD21^hi subset in LN patients contrasts with healthy individuals and non-LN patients. Accordingly, the importance of the analysis of CD21 expression can be highlighted. Considering that one of the most important features to development of SLE/LN is the abnormal function of B cells, the integration of these atypical CD21^hi B cells to the whole disease landscape, could be relevant to understand phenotypic and functional changes of this cell lineage, thus maybe associated with LN progression.

ABCs have also been proposed as a good marker when following SLE patients since their numbers correlate with some clinical and laboratory features, but mainly with auto-antibody titers (14). These findings also showed that the higher frequency of ABCs implies, in most cases, an increase in the activity of the disease (14, 21). Nevertheless, these correlations are weak or not as significant as those detected in the present study for the CD21^hi subpopulation that seems to be more associated to an immunological state of homeostasis.

Beyond the mentioned, there are few descriptions about the effects of induction therapy in LN over B cells; some findings have shown decreased frequencies of plasmablasts/plasma cells associated with mycophenolate mofetil (MMF) administration or the selective depletion of naïve B cells in patients treated with cyclophosphamide (CYC), with little or no effect on class-switched memory B cells in both cases (5). More recently, the effects of the immunosuppressive treatment over a IgD^- CD27^- B cell subpopulation (double negative; DN) were documented in SLE patients with kidney damage, demonstrating a decreased frequency of these cells in patients who responded to therapy (26). To our knowledge, this is the first study about the effect of induction therapy on ABCs or similar B cell phenotypes. Trying to explain the mechanisms involved in ABCs decrease after the immunosuppressive treatment, we could mention the possible role of IL-21: a cytokine that supports the proliferation of this subset (14, 17). Since MMF therapy has been associated with the inhibition of STAT3 phosphorylation (27) and this transcription factor is linked to the stimulation pathway mediated by IL-21, it would not be surprising that MMF could be directly affecting the functionality of ABCs. However, when we analyzed the potential effects of the different immunosuppressants (MMF or CYC) or dosages employed for induction therapy of followed up LN patients in ABCs or CD21^hi subsets, we could not detect any significant shift in their absolute counts regarding the administered drugs (Supplementary Table 3) nor any significant correlation between these cell numbers and dosage of the same treatments (Supplementary Table 4). Therefore, we hypothesize that additional yet unknown factors, independent of the influence of therapeutics, could contribute to numeric changes in these specific B cell subsets in circulation.

On the other hand, the recovery of CD21^hi subset, regardless of a direct effect of induction therapy, could be due to a decrease of different CD21 ligands in circulation after treatment, that includes molecules such as IFN-α, DNA, and C3d, which are classically associated with the physiopathology of SLE (28–30). At this point, we cannot discard that CD21^hi cells could not represent an independent B cell subset, thus maybe constituting a transitory stage derived from ABCs. However, as we still do not understand the biological function of this subset, many questions about their identity or origins remain.

To gain insight about functional roles of these cells and considering that the expression of chemoreceptors plays an important role in the activation and maturation of B cells, we evaluated the expression of CXCR3, CCR7 and CXCR5 associated with extra-follicular/follicular pathway, to characterize the putative activation/effector sites of the CD21^hi population. An increase in B cells’ CXCR5 expression has been demonstrated when high levels of serum CXCL13 are detected in patients with SLE and nephropathy (31–33). Conflictingly, it has been stated that higher frequencies of CXCR5^- CXCR3⁺ B cells correlate with active SLE, besides the presence of CXCR3⁺ B cells in human kidney tissue (18, 34, 35). Since all these reports do not perform extensive phenotyping of these cells, it is possible that our subsets of interest can be overlapped. As ABCs and the recently defined DN subset DN2 (ABCs-like phenotype) have been characterized by a high expression of CXCR3 in contrast to CXCR5 (10, 14, 21), our results regarding these cells correspond to those defining a CXCR3⁺ CXCR5^-/lo CCR7^- phenotype for ABCs. On the other hand, the neglected CD21^hi subset that possesses a higher density of CXCR5 and increases in frequency in post-treatment LN patients, would represent those cells mentioned in the CXCL13/CXCR5 axis reports (31–33). Perhaps by responding to CXCL13, these lymphocytes could migrate towards tertiary lymphoid aggregates to exert putative tissue-associated functions; or maybe, this phenotype could be associated with a follicular pathway that these cells need to follow to be adequately immunologically trained as memory precursors. Accordingly, we recognize that our study is currently restricted by the limited phenotyping of these B cell subsets, that will be further deeply characterized by our group to identify a possible effector role that would promote or cease inflammatory responses.

In summary, we confirmed that ABCs increment their frequency in the circulation of LN patients, and although they cannot be discarded as a factor promoting the pathogenesis of this SLE feature, their numerical alterations could not be as robustly related to LN as those from the non-previously described CD11c⁺ T-bet⁺ CD21^hi cells that are almost absent when renal manifestations arise. Importantly, this CD21^hi subset could be used as a prognostic factor considering that their numbers strongly correlate with less activity of SLE in our cohort. Furthermore, we found that either ABCs or CD21^hi B cell subsets could be considered to assess response to induction therapy in LN. Finally, the identity and functional roles that CD21^hi cells perform during the LN onset/course must be studied considering that they may represent a divergent subset different from the already studied ABCs.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics and Research Committees of the Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán (Ref. 2555). The patients/participants provided their written informed consent to participate in this study.

VS-H contributed to the design and performance of experiments, analysis, and interpretation of data. SR-R and RC-D performed experiments and analyzed data. DC-V, IN-H, LW-G, JT-R, RR-H, and DM-S assisted in obtaining, processing and preservation of patient samples. LW-G, AA-A, and AV-C recruited and collected patient data, generated, and organized our clinical database. GJ-V supports on flow cytometry analyses. VS-H and NM-D performed bioinformatics and statistical analyses. VS-H, VO-N, DG-M, and JM-M designed experiments, supervised general work, wrote, and edited the manuscript. All authors contributed to the article and approved the submitted version.

VS-H was supported by CONACyT fellowship 756882/CVU 854621. This work was supported by CONACyT [FOSISS A3-S-36875] and UNAM-DGAPA-PAPIIT Program [IN213020 and IN212122] granted to JM-M, and CONACyT [652260] granted to GJ-V.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.