C57BL/6 (B6), BALB/c and IL-4/IL-13fl/fl mice (BALB/c) were purchased from the Jackson Laboratories. hCD20-ERT2.Cre^+/− (BALB/c; kindly provided by Mark Shlomchik, University of Pittsburgh) were used as Cre-controls (24). We crossed hCD20-ERT2.Cre and IL-4/IL-13fl/fl mice to generate hCD20-ERT2.Cre^+/− IL-4/IL-13fl/fl mice (B-IL-4/13 iKO). B IL-4/13 iKO and hCD20.Cre-control mice were placed on a 1:1 mixture of normal and tamoxifen-containing chow (Envigo RMS LLC; 250 mg/kg) for 7 days to acclimate, then placed on tamoxifen chow alone (~40 mg intake/d/mouse) for 7 days prior to initiating experiments. Mice remained on tamoxifen chow for the duration of each experiment. Mice were used between 8 and 12 weeks of age. The University of Pittsburgh Animal Care and Use Committee approved all animal studies.

Mice were immunized with 30 × 10⁶ mitomycin C-treated allogeneic (C57BL/6) splenocytes intraperitoneally (i.p.) and animals were then sacrificed at the indicated times (3, 25, 26). To determine cytokine expression, splenocytes were cultured in RPMI complete media (10% FBS, 2 mM L-glutamine, 100 U/ml Penicillin/Streptomycin, and 50 μM 2-mercaptoethanol) and stimulated for 5 h, as follows: B cell IL-4: 10 μg/mg LPS, 0.75 μg/ml ionomycin, 50 ng/ml PMA (Sigma Aldrich Co.) and monensin (eBioscience Inc.). B cell IL-10: splenocytes were stimulated as above except ionomycin was used at 50 ng/ml. T cells cytokines: 0.50 μg/ml ionomycin and 50 ng/ml PMA. Cells were stained with Aqua Live-Dead Fixable dye (ThermoFisher Scientific Inc.), washed and incubated with Fc block (anti-CD16/CD32), then stained with anti- CD3, CD4, CD8, Siglec-1 and CD19 fluorochrome conjugates (Fisher Scientific Inc. and BioLegend Inc.). For intracellular staining, cells were fixed and permeabilized using BD Fixation/Permeabilization kit (BD Biosciences Inc.). Cytokines were detected using fluorochrome conjugated anti-IL-4, IL-5, IL-10, and IFN-γ (Fisher Scientific Inc. and BioLegend Inc.). Intracellular staining controls were incubated with relevant isotype/fluorochrome controls. BD LSRII or Fortessa flow cytometers were used to collect the data and FlowJo software was used for data analysis (BD Biosciences Inc.).

B6 pancreata were digested with collagenase V (Sigma Aldrich) for 20 min at 37°C, as we described (3). Digested tissue was layered onto a Ficoll gradient and islets collected from the 11 and 20.5% interface after centrifugation at 1,800 rpm for 10 min at 4°C. Islets were washed and incubated in RPMI complete medium overnight at 37°C in 5% CO₂. Islets were collected and 250 islets/mouse were placed under the kidney capsule of sex-matched allogeneic recipients with streptozotocin-induced diabetes (blood glucose levels >250 mg/dl) (3). Recipients whose glycemia normalized to <150 mg/dl within 2 days were considered successful. Grafts were considered rejected if blood glucose was >250 mg/dl for two consecutive days. Recipients with normal glycemia beyond 100 days were nephrectomized to demonstrate that euglycemia was due to transplanted islets.

Mice were immunized with 40 μg of Ova precipitated in Alum by intraperitoneal injection on days -14 and -7. On day 0, mice received 15 μg of intranasal Ova for 3 consecutive days. Bronchoalveolar lavage (BAL) was carried out by injecting 1ml of RPMI into the lungs through the trachea and recovering the injected media. This procedure was repeated 9 more times. The lungs were washed with total of 10 ml per mouse. Total lung tissue infiltrate was isolated by injecting mice with 5 μg of anti-CD45-FITC mAb (i.v.) to stain intravascular lymphocytes for exclusion (27). After 5 min, mice were perfused with 20 ml of PBS containing 2% heparin through the left ventricle while the right ventricle was cut in order to drain the blood. Lungs were removed, diced and transferred to gentle-MACS tubes containing RPMI medium, 0.25 mg/ml of Liberase TL and 0.20 mg/ml of DNAase I. The tissue was then dissociated with a gentle-MACS tissue disrupter and further digested for 30–40 min at 37°C in 5% CO₂ according to instructions of the manufacturer. The digested tissue was passed through a 40 μm filter and cell suspension washed with PBS. Cells were prepared for flow cytometry as described above, with the addition of anti-CD45-APC mAb to identify leukocytes within lung tissue.

Statistical analyses and graphing were carried out using GraphPad Prism 9. Data were analyzed using unpaired 2-tailed Student’s t-test, ANOVA, or log-rank (Mantel–Cox) test, as indicated. Differences were considered to be significant at p-values <0.05.

To examine the role of B cell IL-4 on Th2 differentiation and Th2-mediated immune responses, we crossed hCD20-ERT2.Cre and IL-4/IL-13fl/fl BALB/c mice to generate mice with a tamoxifen (Tam)-inducible B cell-specific deletion of IL-4 and IL-13 (B-IL-4/13 iKO). Mice were placed on tamoxifen chow for two weeks prior to initiating experiments. At baseline, naïve Tam-treated B-IL-4/13 iKO and hCD20.Cre-control mice did not have any significant differences in splenic CD3⁺, CD4⁺, or CD8⁺ T cells or B cell frequency or number (Supplemental Figures 1A, B). Further, there were no differences in cytokine expression by CD4⁺ T cells or B cells (Supplemental Figures 1C, D). Consistent with previous findings (28), IL-4 expression is essentially undetectable in B cells from naive hCD20.Cre-control or B-IL-4/13 iKO mice (Supplemental Figure 1D). To evoke an immune response, Tam-treated B-IL-4/13 iKO and hCD20.Cre-control mice were alloimmunized and spleens harvested on day 14. After alloimmunization, a small but reproducible fraction of B cells expresses IL-4 (Figure 1A). B cells from B-IL-4/13 iKO mice had a 60% decrease in frequency of IL-4 expression compared to those from hCD20.Cre-control B cells (Figure 1A). Of note, we did not detect IL-13 protein expression by B cells from alloimmunized B-IL-4/13 iKO or hCD20.Cre-control mice (Supplemental Figure 2A).

We previously demonstrated that TIM-1⁺ B cells, a population with potent Breg activity, were highly enriched for both IL-4 and IL-10 expression (3). Moreover, B cells from constitutive IL-4^−/− and IL-4Rα^−/− mice exhibited markedly decreased expression of both TIM-1 and IL-10, suggesting that IL-4 signaling was important for Breg development. In alloantigen immunized B-IL-4/13 iKO mice, acute reduction in B cell IL-4 expression resulted in an ~4-fold decrease in the frequency of IL-10⁺ B cells (Bregs) compared to hCD20.Cre-controls (Figure 1B and Supplemental Figure 2A). Because the number of B and T cells in spleens of control and iKO mice are comparable (Supplemental Figure 1A), these changes in frequency of B and T cell subsets reflect changes in overall number. Thus, acutely reducing endogenous B cell-derived IL-4 markedly reduces Breg IL-10 expression. Examination of CD4⁺ T cells from the same mice revealed that reduced B cell IL-4 resulted in a 7.5-fold decrease in frequency of IL-4 expression, while the expression of IL-10, IFNγ, and Foxp3 were unchanged (Figures 1C–F and Supplemental Figure 2). Moreover, there was no difference in IFNγ expression by CD8 T cells (Figure 1G). Thus, reduction in B cell IL-4 results in a specific decrease in IL-10⁺ Bregs and Th2 cells. This indicates that B cell IL-4 normally promotes Th2-polarization in response to alloimmunization.

Studies performed several decades ago suggested that allograft rejection is exacerbated by Th1 immune responses while Th2 responses promote allograft acceptance (29–34). Given that B cell IL-4 supports Th2 responses after alloimmunization, we examined its role in allograft survival. To accomplish this, diabetic Tam-treated B-IL-4/13 iKO versus hCD20.Cre-control mice were transplanted with fully MHC-mismatched islet allografts. B-IL-4/13 iKO transplant recipients exhibited accelerated rejection of allografts compared to hCD20.Cre-control mice (MST = 21 days vs. 31 days, respectively; p = 0.04, Figure 2A). This suggests that acute rejection is exacerbated by reduced B cell IL-4 and resultant decreased Th2 responses in B-IL-4/13 iKO mice, at least when combined with Breg deficiency. In contrast, we previously showed deficiency of either Bregs or B cell IL-10 alone, do not accelerate rejection in untreated recipients (9, 35).

Treatment of islet allograft recipients with tolerogenic anti-TIM-1 mAb (RMT1-10) was shown to increase IL-10 expression by B cells ~3–5 fold and prolongation of allograft survival was dependent on B cell IL-10 expression (3, 9, 36, 37). To determine whether anti-TIM-1 remains effective in the face of reduced B cell IL-4, diabetic Tam-treated B-IL-4/13 iKO and hCD20.Cre-control mice received islet transplants as above but were treated with anti-TIM-1. Anti-TIM-1 markedly prolonged allograft survival in hCD20.Cre-control mice, resulting in long-term survival in 66% of recipients (MST 17days vs. >100 days, respectively; p = 0.04, Figure 2A). In contrast, anti-TIM-1 did not prolong allograft survival of B-IL-4/13 iKO mice. After anti-TIM-1 treatment, alloimmunized B-IL-4/13 iKO mice maintained an ~60% reduction in splenic IL-4⁺ B cells compared to control hCD20.Cre mice (Figure 2B). However, in this setting, anti-TIM-1 elicited a similar increase in splenic IL-10⁺ Bregs in both strains (Figures 2B, C). This contrasts with the decrease in B cell IL-10 expression in alloimmunized B-IL-4/13 iKO mice in the absence of anti-TIM-1 (Figure 1B). Regardless, the frequency of splenic IL-4⁺ CD4 T cells was still reduced in anti-TIM-1 treated B-IL-4/13 iKO recipients compared to hCD20.Cre-control mice, while the frequency of IFNγ, IL-17 and Foxp3 expressing CD4 T cells and IFNγ expressing CD8 T cell remained unchanged (Figures 2D–H). These data support the notion that B cell IL-4 supports Th2 differentiation and that B cell IL-4 and Th2 polarization play an important role in anti-TIM-1 mediated allograft survival that is independent from the role of IL-10⁺ Bregs, Tregs or Th1 responses.

IL-4 expression by Th2 cells is a major contributing factor to pathophysiology of AAD including airway inflammation (38, 39). To directly address the role of B cell IL-4 expression on Th2 polarization and airway inflammation, we subjected Tam-treated B-IL-4/13 iKO and hCD20.Cre-control mice to Ova-induced AAD. In naïve mice, ~120,000 CD45⁺ leukocytes were recovered in BAL from either hCD20.Cre-control or B-IL-4/13 iKO mice (Supplemental Figure 3A). After induction of AAD, there was virtually no increase in number of CD45⁺ leukocytes in the bronchoalveolar space of B-IL-4/13 iKO mice, whereas those from hCD20.Cre-control mice increased ~5-fold (Figure 3A). BAL fluid had 3.5× more CD3 T cell and 12.5× more eosinophils in hCD20.Cre-controls (Figures 3B, C). Very few, if any, B cells were found in the BAL of either mouse strain (data not shown).

Lung histopathology in hCD20.Cre-control mice revealed marked thickening and infiltration of inflammatory cells into the alveolar septae compared to B-IL-4/13KO mice, which had preserved lung architecture (Figures 4A, B). Moreover, in hCD20.Cre-control mice, an inflammatory cell infiltrate and proteinaceous debris was observed in the alveolar space and an inflammatory infiltrate was observed around a proportion of small airways (Figure 4B). In contrast, B-IL-4/13 iKO mice had clear alveolar spaces and no infiltrate around small airways (Figures 4A, B). No differences were observed in epithelial thickening between the two groups. To quantitate and characterize the parenchymal infiltrate, lung tissue was digested and intravascular lymphocytes were gated out (see Materials and Methods; Supplemental Figure 4). As seen in alloimmunized mice, after induction of AAD, no B cell IL-13 could be detected in lung tissue from either B-IL-4/13 iKO or hCD20.Cre control mice (Supplemental Figure 5). Naïve Control and B-IL-4/13 iKO mice, had a similarly low number of resident CD45⁺ cells in lung tissue, of which only a small fraction was comprised of T or B lymphocytes or eosinophils (Figures 4C–E, H, I). After induction of AAD, lung tissue from hCD20.Cre-control mice exhibited a >5-fold increase in infiltration by (CD45⁺) leukocytes (from 4.9×10⁶ to 21×10⁶), ~75% of which was comprised by a marked expansion of T cells. B cell infiltration increased ~5-fold but still comprised only about 5% of the total infiltrate. Eosinophil numbers increased markedly, but their contribution to the total infiltrate was minor. Despite the preservation of lung architecture and normalization of BAL seen in B-IL-4/13 iKO mice after AAD induction, infiltration of leukocytes and lymphocytes into lung parenchyma was increased compared to baseline levels seen in naive mice (Figures 4B–F). However, in comparison to hCD20.Cre control mice, after AAD induction, B-IL-4/13 iKO mice exhibited a 2-fold decrease in the number of infiltrating (CD45⁺) leukocytes, including an ~2-fold decrease in total (CD3⁺) and CD4⁺ T cells (Figures 4C–E). Moreover, amongst CD4 T cells in lung tissue, there was an ~2-fold decrease in IL-4 and IL-5 expressing cells in B-IL-4/13 iKO mice (Figures 4F, G). Despite the marked differences in BAL, there was no significant difference between B-IL-4/13 iKO and hCD20.Cre control mice in eosinophil infiltration into lung tissue, nor were there differences in B cell infiltration (Figures 4H, I). Thus, acute reduction in B cell IL-4 significantly reduced Th2 polarization, lung T cell infiltration, and airway inflammation in AAD.