Male and female transgenic mice expressing HLA-II alleles, DRB1^*03:01 (DR3), DRB1^*15:01 (DR15), or HLA-DQA1^*0301 and DQB1^*0302 (DQ8), were bred in-house in a specific pathogen-free room at the University of North Dakota. These transgenic mice have been shown to express functional HLA (32, 33), and the presence of the appropriate HLA-II transgenes was confirmed by PCR-based genotyping using allele-specific oligonucleotide primers as described previously (33, 34). Animals were housed under a 12-h light/12-h dark cycle and had ad libitum access to water and a whey-protein-free diet (Teklad 2018, Envigo Corporation, Indianapolis, IN, USA). All procedures were carried out in strict accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the University of North Dakota Institutional Animal Care and Use Committee.

BLG sensitization was carried out as previously described (18, 35). Briefly, 4-week-old genotyped transgenic mice were randomly assigned to sham or BLG-sensitization groups (n = 6–11). Once a week for 5 weeks, the BLG-sensitization group was intragastrically gavaged with 1 mg of purified BLG (Millipore Sigma, Burlington, MA, USA) in 200 μL sodium bicarbonate buffer (pH 9.0) with 10 μg cholera toxin (CT; List Biologicals, Campbell, CA, USA) as an adjuvant, while the sham group received only the CT-containing vehicle (Figure 1A). All mice were fasted for 2 h before and after sensitization. At the beginning of Week 6, both the sham and BLG-sensitized mice were orally challenged with 50 mg BLG (without CT) in 200 μL sodium bicarbonate buffer. At 30 min post-challenge, their body temperature was measured using a MicroTherma 2 T Handheld Thermometer with a RET-3 probe (Braintree Scientific, Inc., Braintree, MA, USA), and their physical symptoms were scored based on the scale by Li et al. (36).

Mice were euthanized by CO₂-asphyxiation 2 days after the BLG challenge. Atrial blood was collected into microfuge tubes or EDTA-coated Microvette tubes (Sarstedt, Inc., Newton, NC, USA). The blood collected in microfuge tubes was centrifuged at 2,000 × g for 15 min at 4°C to collect the serum and stored at −80°C to determine BLG-specific IgG and IgE levels. Ten microliter of whole blood from the EDTA-coated tubes was thinly smeared onto a glass slide for later microscopic analysis. The remaining blood in the circulation was cleared by transcardial perfusion with phosphate-buffered saline (PBS; pH 7.4). The ileum was dissected from the rest of the small intestine, fixed in 4% paraformaldehyde (pH 7.4) for 24 h at 4°C, and equilibrated in 30% sucrose in PBS prior to sectioning.

Isotype-specific detection of immunoglobulins (Igs) was carried out as previously described (18, 39). Briefly, 8-well RIA strips (Corning, Inc., Corning, NY, USA) were coated with 2 μg/mL of sterile BLG solution in sodium carbonate/bicarbonate buffer (pH 9.5) overnight at 4°C. Wells were washed and blocked in an assay buffer containing 0.5% bovine serum albumin (BSA) in PBS. Serum samples were diluted to 1:40 in assay buffer and incubated with magnetic protein-G Dynabeads™ (Thermo Fisher, Waltham, MA, USA) to adsorb total IgG. The supernatant was collected for IgE ELISA, while bead-bound IgG was eluted with 50 mM glycine buffer (pH 2.8) and neutralized with 1 M tris buffer (pH 7.5) for IgG1 ELISA. Both sample preparations were incubated in the BLG-coated plates overnight at 4°C. Wells were washed thoroughly with the assay buffer and incubated with a secondary anti-mouse IgE (Thermo Fisher Scientific, Cat# 13-5992-82) or IgG1 (ThermoFisher Scientific, Cat# 13-4015-82) at 1:1,000 dilution. The amounts of IgE and IgG1 were quantified by the colorimetric detection of 3,3′,5,5′-tetramethylbenzidine (TMB) as the substrate after incubating with avidin-HRP (eBioscience, San Diego, CA, USA). The substrate reactions were terminated with 2 N H₂SO₄, and the plates were immediately read at 450 nm with a reference wavelength at 550 nm (BioTek Instruments, Winooski, VT, USA).

Mast cell activities were determined by the amounts of circulating MCPT-1 as described previously (40). The serum samples collected at the time of sacrifice were diluted 1:10 with the assay buffer, and 100 μL of the diluted samples were used to quantify MCPT-1 using the Mouse MCPT-1 Uncoated ELISA Kit (Thermo Fisher Scientific) according to the manufacturer's instructions.

The blood smears from the anticoagulated whole blood samples were fixed in methanol for 10 min and allowed to dry. The slides were stained with Wright-Giemsa stain (Volu-Sol, Salt Lake City, UT, USA) according to the manufacturer's instructions and coverslipped with Permount mounting medium (Fisher Scientific, Hampton, NH, USA). Two blinded observers, a board-certified medical laboratory scientist (DG) and a pathology resident (YW), performed the 5-part differential blood counting and calculated the relative percentage of lymphocytes, neutrophils, monocytes, eosinophils, and basophils of 200 leukocytes.

Sucrose-equilibrated ileum samples were embedded in Tissue-Tek® O.C.T. Compound (Sakura Finetek USA, Inc, Torrance, CA, USA) and sectioned at 10 μm on a Leica CM1850 cryostat (Leica Biosystems, Buffalo Grove, IL, USA). Ileum sections were then stained with H&E and coverslipped at the University of North Dakota Histology Core Facility. Whole-slide images were captured using the Hamamatsu NanoZoomer 2.0HT Brightfield + Fluorescence Slide Scanning System (Hamamatsu Photonics, Bridgewater, NJ, USA). Using the measure tool on the NDP.view2 software (Hamamatsu Photonics), 5 villi per section were measured from the intestinal crypts to the apical tip of the villi. The villi measurements were averaged to calculate the villi length of each animal.

The spleens were excised from 3 to 4 months-old naïve male and female HLA-DR3, HLA-DR15, and HLA-DQ8 mice after euthanasia with CO₂ asphyxiation and individually collected in ice-cold PBS (pH 7.4). Each spleen was mashed on a sterile 70-μm cell strainer placed over a 50-mL conical tube and rinsed through the strainer with PBS. Splenocytes were pelleted by centrifuging the resulting cell suspension at 400 × g for 5 min at 4°C and resuspended in red blood cell lysis buffer (BioLegend, San Diego, CA, USA) for 2 min at room temperature. Lysis was terminated by adding 30 mL PBS and removing the lysis buffer by centrifugation. The resulting cell pellets were resuspended in RPMI-1640 medium supplemented with 25 mM glucose, 25 mM HEPES, 25 mM sodium bicarbonate, heat-inactivated 10% fetal bovine serum (FBS; Peak Serum, Inc., Wellington, CO, USA), pen-strep cocktail (100 IU penicillin, 100 μg/mL streptomycin; Corning, Inc.), and 50 μM β-mercaptoethanol (MilliporeSigma, Burlington, MA, USA). The number of viable cells was counted by trypan blue exclusion, and 1 × 10⁶ viable cells per well were plated in 96-well round-bottom plates. The cells were cultured in the supplemented RPMI-1640 medium alone or with BLG at the concentration indicated (0, 1.0, or 10 mg/mL) for 72 h at 37°C with 5% CO₂. The responsiveness of the cells to an immunological stimulus was validated using concanavalin A (ConA; data not shown).

The splenocytes culture plates were centrifuged at 400 × g for 5 min at 4°C, and the media were removed from the wells and stored at −80°C. The remaining cells were washed twice with PBS and incubated with Ghost Dye™ Violet 510 viability dye (1:1,000 in PBS; Tonbo Biosciences, San Diego, CA, USA) for 30 min on ice in the dark. After washing in PBS and blocking with anti-CD16/32 antibody (Biolegend, Cat# 101302) at a 1:50 dilution in PBS with 2% FBS, the cells were incubated in one of 3 designated antibody cocktails, each containing a panel of specific primary antibodies for immunophenotyping (Table 1). The cells were stained for 1 h at 4°C in the dark, washed in PBS with 2% FBS, and fixed for 1 h at 4°C in Fixation Buffer (Biolegend). For the antibody panel 3, which included an intracellular target (CD206), the fixed cells were washed in Perm/Wash Buffer (BioLegend), blocked with 3% rat serum in Perm/Wash Buffer for 10 min at 4°C in the dark, and incubated with anti-CD206 antibody diluted in Perm/Wash Buffer for 1 h at 4°C in the dark. Labeled cells were washed and resuspended in PBS containing 2% FBS and 1 mM EDTA and analyzed using a BD FACSymphony A3 flow cytometer (BD Biosciences, San Jose, CA, USA) at the North Dakota Flow Cytometry and Cell Sorting Core Facility. Unstained, single-stained, and fluorescence-minus-one (FMO) controls for every fluorochrome were used to set gating parameters to exclude dead cells, determine specificity, and establish the compensation panel to avoid spectral overlaps. At least 140,000 total cells were acquired.

The results were further analyzed using FlowJo v10.8.1 software (BD Life Sciences). For each panel, individual .fcs files from all samples were run on the same workspace. Using built-in FloJo plugins, .fcs files were inspected and cleaned for any anomalous signals and spurious events that might confound downstream data analysis. After gating on singlets from good events, the number of events from each sample was downsampled to the fewest number of live CD45⁺ events in each antibody panel. The minimum number of live/CD45⁺ for downsampling was determined based on the sample with the fewest numbers of CD45⁺ events. Downsampled CD45⁺ events from all samples were then selected for concatenation to define immune subpopulations (Supplementary Figure 1). The t-distributed stochastic neighbor embedding (t-SNE) was performed on the concatenated file using the default parameters in FloJo software (Supplementary Figure 2). To identify immune subsets and changes within genotypes and stimulation conditions, heatmaps with hierarchical clustering were generated with the heatmap tool, Morpheus (Supplementary Figure 3; software.broadinstitute.org/morpheus/), using fold change values and applying 1-Pearson correlation metric with average linkage in both columns and rows.

All statistical analyses were performed using GraphPad Prism v9.0 software (GraphPad Software, Inc., La Jolla, CA, USA), with males and females independently analyzed. In vivo intra-HLA strain differences were compared using multiple Mann-Whitney tests or multiple uncorrected t-tests as indicated in the figure legends. In vivo inter-HLA strain differences were compared using two-way ANOVA with Fisher's least significant difference (LSD) test. In vitro splenocyte experiments comparing unstimulated and stimulated splenocytes used a one-way ANOVA with a Dunnett's post-hoc test. The positive percentage of respective markers within each sample was first calculated, and the frequency of BLG-specific cells is presented as fold change by dividing the frequency of respective markers in the unstimulated condition. The ROUT method (Q = 1%) was used to identify outliers in a group when appropriate, and the values were removed from the final results. A p < 0.05 (p < 0.05) was considered statistically significant.

As we have previously demonstrated with C57BL/6J and BALB/c mouse strains (18, 37), the growth of mice was unaffected by the BLG sensitization procedure in sham and BLG mice expressing HLA-II (Figures 1B,C). However, when challenged with a large dose of the allergen, the physical reactions of BLG-sensitized mice ranged widely by HLA-II strain and sex. On a symptom-based scale (36), BLG-sensitized DR3 males scored the highest by exhibiting the most severe reactions (BLG DR3: 1.9 ± 0.5; p = 0.001), with one mouse dying after the allergen challenge (Figure 2A). In contrast, BLG-sensitized DR15 and DQ8 males were largely asymptomatic, with their scores not significantly above their sham counterparts (Figure 2A; BLG DR15: 0.4 ± 0.3; BLG DQ8 0.4 ± 0.3). For all HLA-II strains, BLG-sensitized females showed only mild or no physical response after the allergen challenge (Figure 2B). However, the difference in the scores between sham and BLG-sensitized mice was significant for DQ8 females (BLG DQ8: 0.6 ± 0.2; p = 0.04) and near statistical significance for DR3 females (BLG DR3: 0.7 ± 0.4; p = 0.07). In addition to the symptom scoring, the hypothermic response was recorded 30 min post-challenge as an additional indicator of the immediate hypersensitivity response. While not statistically significant, BLG-sensitized DR3 males showed a clear trend for hypothermia (sham DR3: 36.5 ± 0.2°C; BLG DR3: 34.5 ± 1.0°C; p = 0.7) (Figure 2C). This decrease was also observed with BLG-sensitized DR3 females (sham DR3: 36.8 ± 0.2°C; BLG DR3: 36 ± 0.7°C; p = 0.6) and DQ8 females (sham DQ8: 37.7 ± 0.2°C; BLG DQ8: 35.4 ± 1.2°C; p = 0.2) (Figure 2D). These results suggested that the severity of the immediate physical reactions to the BLG challenge was both HLA-II and sex-dependent.

One day after the BLG-challenge, the general mobility of mice was assessed with the OFT, including the total distance traveled, time immobile, and the number of immobile episodes. Furthermore, anxiety-like behavior was also investigated by the avoidance and the time spent in the center zone of the OFT arena. We have previously reported that mice become acclimated to the OFT apparatus over the 10 min testing period (18). Thus, after quantifying their activities at each 1-min interval (not shown), the first 4 min of the testing period prior to the apparent acclimation between the treatment groups was analyzed. Of the BLG-sensitized male HLA-II mice, only the DR15 mice showed significant behavioral changes. Upon assessing the overall locomotion with the OFT (Figures 3A–F), the sensitized DR15 males traveled significantly less distance (Figure 3A; sham DR15: 11.1 ± 1.4 m; BLG DR15: 7.2 ± 0.7 m; p = 0.03) and spent nearly twice as long immobile (Figure 3B; sham DR15: 57.7 ± 14.0 s; BLG DR15: 103.5 ± 0.7 m; p = 0.04) with a greater number of immobile episodes than sham mice (Figure 3C; sham DR15: 10 ± 2; BLG DR15: 16 ± 1; p = 0.005). In addition, the sensitized DR15 females showed a trend for lower mobility than their sham counterparts (Figures 3D–F). When anxiety-like behavior was evaluated with the OFT (Figure 4), male DR15 mice showed a tendency toward entering the center zone less frequently (Figure 4A; sham DR15: 9 ± 2; BLG DR15: 5 ± 1; p = 0.08) and spent a shorter amount of time in the center than sham mice (Figure 4B; sham DR15: 13.3 ± 2.4 s; BLG DR15: 7.0 ± 1.2 s; p = 0.04), although no significant differences were observed in the female groups (Figures 4C,D). When spatial memory was evaluated by the CMT (Figures 5A–D), BLG-sensitized male DR15 mice exhibited decreased mobility as in the OFT, with fewer numbers of arm entries than the strain-matched sham mice (Figure 5A; sham DR15: 29 ± 3; BLG DR15: 19 ± 1; p = 0.02). However, the percent alternations by these mice were comparable between the treatment groups (Figure 5B), suggesting that their sensitization status did not influence their spatial memory after the allergen challenge. In contrast, while no differences in the number of arm entries were observed between the strain-matched female groups (Figure 5C), sensitized DR15 female mice showed significantly fewer successful alterations (Figure 5D; sham DR15: 36.9 ± 2.8%; BLG DR15: 29.8 ± 1.2%; p = 0.03) while their total number of arm entries was similar to the sham mice. Inter-HLA strain comparisons of the sensitized mice are shown in Supplementary Figures 4–6. These results indicated that the behavioral manifestations of CMA were observed only in BLG-sensitized DR15 mice when mice were challenged acutely, and changes in their behavioral phenotypes were sex-specific.

Circulating levels of BLG-specific IgE and IgG1 from terminal blood samples were quantified to assess hypersensitivity immune responses by the sham and sensitized mice. Allergen-specific IgE was significantly elevated by 2-fold in the sensitized DR3 male mice (sham DR3: 0.4 ± 0.03; BLG DR3: 0.8 ± 0.1; p = 0.03) but not in DR15 and DQ8 male mice (Figure 6A). However, when the serum samples from female mice were assayed, the levels of BLG-specific IgE were 1.7-fold greater in sensitized DR3 mice (sham DR3: 0.4 ± 0.05; BLG DR3: 0.7 ± 0.07; p = 0.006) and 2-fold in DR15 (sham DR15: 0.3 ± 0.03; BLG DR15: 0.6 ± 0.08; p = 0.006) compared to the strain- and sex-matched sham mice (Figure 6B). The allergen-specific IgE levels in sensitized DQ8 mice were comparable to the sham counterparts as observed in males (Figure 6B). On the other hand, BLG-specific IgG1 levels showed small but significant elevation in sensitized male DR3 mice compared to sham (Figure 6C; DR3 sham 0.6 ± 0.03; DR3 BLG 0.9 ± 0.08; p = 0.03) and in sensitized female DR15 mice (Figure 6D; DR15 sham 0.2 ± 0.05; DR15 BLG 0.4 ± 0.06; p = 0.03). Interestingly, the allergen-specific IgG1 levels were robustly elevated in both male and female DQ8 sensitized mice, with approximately 8.5- and 6-times greater levels of IgG1, respectively, than their sham mice (Figures 6C,D; male sham DQ8: 0.4 ± 0.04; male BLG DQ8: 3.4 ± 0.03; p < 0.0001; female sham DQ8: 0.4 ± 0.01; female BLG DQ8: 2.5 ± 0.5; p = 0.0006). Inter-HLA strain comparisons are shown in Supplementary Figure 7. Together, these results indicated that the humoral response to the BLG allergen was affected both by the HLA-II allele expressed and sex.

Mast cells are the effector cells of the immediate allergic response and release their granular contents, such as histamine and proteases, upon allergen binding to their cell-surface IgE-FcεRI complex (41). To compare the differences in mast cell activation in the three HLA-II transgenic strains, we measured MCPT-1 from the serum samples of sham and BLG-sensitized mice. MCPT-1 was significantly elevated in sensitized DR3 male mice by 2-fold (Figure 7A; sham DR3: 4.0 ± 0.7 pg/mL; BLG DR3: 8.8 ± 0.8 pg/mL; p = 0.002). Sensitized female mice exhibited notable variabilities in the MCPT-1 levels within each strain. In particular, one of the sensitized DR3 and DR15 mice showed a much greater level of MCPT-1 than others in the respective group (Figure 7B). However, only the MCPT-1 levels in sensitized HLA-DQ8 mice were significantly different from the sham mice (Figure 7B; sham DQ8: 6.3 ± 1.1 pg/mL; BLG DQ8: 12.8 ± 5.5 pg/mL; p = 0.009). Inter-HLA strain comparisons are shown in Supplementary Figure 8. These observations suggested that mast cells were differentially activated among the three strains of HLA-II mice, resulting in the varying levels of mast cell-derived proteases in their circulation at the time of euthanasia.

Blood smears were prepared at the time of sacrifice to investigate changes in circulating leukocyte composition (Figures 8A–H). HLA-DR15 mice had the most dynamic change among the BLG-sensitized male mice, with a 3% decrease in circulating lymphocytes (Figure 8A; sham DR15: 84.1 ± 1.5%; BLG DR15: 81.0 ± 1.1%; p = 0.0001). In addition, a 2% increase in neutrophils (Figure 8B; sham DR15: 3.3 ± 0.4%; BLG DR15: 5.6 ± 0.8%; p = 0.01) and approximately 1.5% more monocytes were observed in these mice than their sham counterparts (Figure 8C; sham DR15: 1.5 ± 0.2%; BLG DR15: 3.1 ± 0.6%; p = 0.01). BLG-sensitized male DQ8 mice had 4% greater number of neutrophils (Figure 8B; sham DQ8: 5.8 ± 0.5%; BLG DQ8: 9.9 ± 1.5%; p = 0.02) and an almost 2% increase in monocytes (Figure 8C; sham DQ8: 2.6 ± 0.3%; BLG DQ8: 4.4 ± 0.6%; p = 0.008). Among the BLG-sensitized female groups, only HLA-DR3 mice showed significant changes in their blood leukocytes, with a 3% increase in circulating monocytes (Figure 8G; sham DR3: 3.1 ± 0.7%; BLG DR3: 6.4 ± 1.4%; p = 0.03). Although the immediate reaction to the challenge had occurred over 2 days prior to sacrifice, small but significant differences in blood leukocyte composition were detectable, particularly in the male sensitized transgenic mice. These findings suggest that these persisting changes occurred during the sensitization or were triggered by the allergen challenge.

Ileum sections were stained with H&E to determine intestinal inflammatory changes due to the BLG sensitization and challenge among the transgenic mouse strains. While inflammation was evident in all the ileums of BLG-sensitized male (Figures 9A–F) and female mice (Figures 9G–L), the extent of the structural changes depended on HLA-II expression and sex. BLG-sensitized DR3 males were the most severely affected with highly disorganized mucosal/submucosal architecture (Figures 9A,D) and significant blunting of the villi (Figure 10A; sham DR3: 249 ± 16 μm; BLG DR3: 187 ± 13 μm; p = 0.02). Sensitized DR15 males also showed a significant decrease in villi length (Figure 10A; sham DR15: 322 ± 5 μm; BLG DR15: 289 ± 12 μm; p = 0.04), though the changes in their overall mucosa and submucosa morphology were less extensive (Figures 9B,E). Disorganized villi and crypt structure were also noticeable in DQ8 male mice (Figures 9C,F), but the trend of decrease in their average villi length did not reach statistical significance (Figure 10A; sham DQ8: 292 ± 20 μm; BLG DQ8: 251 ± 23 μm). In female ileums, BLG-sensitized DR3 mice again had the most drastic structural changes in the mucosal and submucosal organization (Figures 9G,J), while overall morphology of the sensitized DR15 and DQ8 ileums were similar to their respective sham controls (Figures 9H,I,K,L). There was no significant shortening of the villi in any of the female HLA-II strains, though BLG-sensitized mice tended to have shorter villi than their sham counterparts (Figure 10B). While the BLG challenge increased inflammatory changes in sensitized mice of all the transgenic strains, the severity of the changes was greatly influenced by both the HLA-II strain and the sex of the sensitized mice.

To better understand the influence of BLG on the immune responses of the transgenic mouse strains as an antigen, we directly stimulated splenocytes from naïve mice with either 1 mg/mL (BLG_low) or 10 mg/mL (BLG_high) of BLG in vitro. The phenotypic changes in BLG-stimulated cells were expressed as fold changes from the strain- and sex-matched unstimulated controls (Figures 11–13), and the samples were classified according to the similarities in their expression profiles of several markers using Morpheus. The algorithm-assisted hierarchical clustering grouped the phenotypic markers based on distinct concentration- and genotype-specific immune profiles and revealed three major profiles of immune subsets (Supplementary Figure 3).

The first cluster grouped phenotypic markers for T cells (CD3⁺, CD4⁺, CD8⁺) and myeloid cells (Ly6C⁺), such as dendritic cells, monocytes, and/or macrophages (Figures 11A–F). As anticipated, the splenocytes from the three HLA-II variant strains of male and female mice altered their immunophenotypes in the presence of the allergen. While BLG did not elicit notable changes in the T cell populations in the splenocytes from DR3, particularly from female cells (Figures 11A,D), the allergen altered the T-cell-associated profiles in DR15 and DQ8 of both sexes. The effects of BLG on the CD3⁺ population in DR15 and DQ8 were concentration-dependent, with the expansion of this population promoted by BLG_low and suppressed by BLG_high. The Ly6C⁺ populations in the cells from these genotypes showed similar profile changes. However, female DR3 splenocytes were stimulated to expand this population at the higher concentration of BLG.

Cluster 2 largely grouped innate markers, including CD11b⁺, F4/80⁺, CD14⁺, CD206⁺, Ly6G⁺, CD80⁺, and the innate-to-adaptive bridging phenotype, CD8⁺CD49b⁺, and the effector phenotype, CD8⁺CD25⁺. Among this group, the most prominent finding was that BLG_high induced profound expansion of CD45⁺CD206⁺ cells in the stimulated splenocytes from all three genotypes of both sexes (Figure 12). In addition, concentration-dependent increases were observed with the CD45⁺CD11b⁺ cell populations in all but male DQ8 splenocyte preparations after stimulation with BLG_high. In contrast, the expansion of leukocyte populations labeled with the neutrophil marker, Ly6G, increased with BLG_high in all samples except in splenocytes from male DR15. Furthermore, the cells from male and female DR15 failed to expand the CD14⁺ population, indicating that the allergen did not stimulate monocytes in this genotype even at the higher concentration of BLG. DQ8 was uniquely stimulated by BLG_high, showing increased F4/80⁺ macrophage and CD80⁺ leukocytes populations, regardless of sex.

The phenotypes identified by the third cluster included CD45⁺, CD11c⁺, HLA-II⁺, CD4⁺CD49b⁺, CD4⁺CD25⁺, CD4⁺TCRβ⁺ (TCRβ: the variable beta chain of the T cell receptor), CD8⁺TCRβ⁺, CD86⁺, B220⁺, CD19⁺, and IgM⁺. Among this group, CD4⁺TCRβ⁺, CD8⁺TCRβ⁺, CD86⁺, and HLA-II⁺ populations were uniquely and significantly increased in the splenocytes from DR15 and DQ8, but not DR3, for both sexes (Figure 13). Additionally, the population of B cells expressing B220, CD19, and IgM were also upregulated by BLG_high in the splenocytes from male and female DR15 mice but not in any other samples. These results suggested that the expansion of B cell populations likely facilitated T cell activation in a manner specific to the HLA-II variant, particularly highlighting the classic HLA-II/TCR interactions and cognate B cell/CD4⁺ T cell interactions in DR15. Innate responses were more pronounced in DQ8.

Together our in vitro results indicated that BLG, especially at a higher concentration, stimulated innate and adaptive immune responses differentially among the three genotypes, suggesting that the HLA-II allele dictates the downstream cellular responses.