Marginal Zone B Cells are Necessary for the Formation of Anti-Donor IgG After Allogeneic Sensitization

Abstract

Background:

The formation of anti-major histocompatibility (MHC) antibody is a significant barrier for many patients awaiting organ transplantation. Patients with pre-formed anti-MHC antibodies have limited options for suitable donors, and the formation of donor-specific anti-MHC antibodies after transplantation is a harbinger of graft rejection. Despite the recognized importance of anti-MHC antibodies, the mechanisms responsible for the differentiation of B cells after exposure to allogeneic antigen are poorly understood.

Methods:

In order to evaluate the differentiation of B cells in response to allogeneic antigen, we used a model of H-2^b C57Bl/6 sensitization with H-2^d antigen. We used a Class I MHC tetramer-based approach to identify allogeneic B cells and flow cytometric crossmatch to identify allogeneic IgM and IgG.

Results:

We found that although the formation of anti-H-2^d IgG was robust, few class switched B cells and germinal center B cells were formed. Antigen-specific B cells did not express classical memory B cell markers after sensitization but had an IgM⁺CD21⁺ marginal zone B cell phenotype. The frequency of marginal zone B cells increased after sensitization. Depletion of marginal zone B cells prior to sensitization or skin grafting resulted in a significant diminution of anti-H-2^d IgG and fewer germinal center B cells. Adoptive transfer experiments revealed that marginal zone B cells more efficiently differentiated into germinal center B cells and anti-donor IgG producing cells than follicular B cells.

Conclusions:

These results demonstrate an important role for marginal zone B cells as a reservoir of alloreactive B cells that are activated by allogeneic antigen.

INTRODUCTION

Solid organ transplantation is curative therapy for end-stage organ failure due to a wide range of diseases. However, the successful impact of transplantation is limited by the formation of anti-MHC antibodies, termed allogeneic sensitization.^1,2 For patients awaiting transplantation, allogeneic sensitization can dramatically limit the number of compatible donors, and the formation of anti-donor HLA antibodies after transplantation leads to graft rejection and loss.³ Further mechanistic insight into the cellular and molecular pathways that lead to allogeneic sensitization could lead to improved risk stratification of patients prior to transplant and pathway-specific treatments for desensitization and antibody-mediated graft rejection ^4–7.

Although plasma cells are major source of anti-MHC IgG, the pathways by which B cells differentiate into plasma cells after encountering allogeneic antigen are poorly understood ^4,5. Naïve mature B cells differentiate into two distinct subsets, follicular and marginal zone populations. Relative to follicular B cells, which are IgD^hi and reside within the follicle of secondary lymphoid organs, marginal zone B cells are IgM^hi and reside along the follicle periphery in the marginal zone ^8,9. Marginal zone B cells are described as innate-like, owing to their capacity to rapidly respond to bloodborne pathogens and pattern recognition receptors (PRR) such as the TLR4 agonist lipopolysaccharide (LPS).^10,11 Relative to follicular B cells, marginal zone B cells are poised to rapidly form IgG secreting plasma cells.^11,12 Despite these differences, there is significant functional overlap between follicular and marginal zone B cells, as each can acquire the opposite phenotype in certain contexts, form germinal center (GC) B cells, and form memory B cells.^11,13,14

In transplantation, there is growing evidence that innate-like and MZ B cells play a role in HLA sensitization and graft rejection. Marginal zone precursors play a critical role in CD40 blockade tolerance to heart allografts.¹⁵ Depletion of marginal zone B cells prolongs heart graft survival after T cell depletion, and marginal zone B cells are required for antibody responses against a red blood cell alloantigen.^16–18 Recent studies found that in rejecting hearts and kidneys, infiltrating B cells populations are responsive to innate stimuli and have self-reactivity.^6,7,19,20 In this study, we sought to understand the differentiation of B cells after exposure to allogeneic antigen by tracking antigen-specific B cells after sensitization. We found that marginal zone B cells are critical for the formation of germinal center B cells and anti-donor IgG and thus represent an enriched reservoir of alloreactive B cells.

MATERIALS AND METHODS

Mice.

C57Bl/6 (H-2^b), Balb/c (H-2^d), RAG1 knockout, and MD4 (IgHEL) mice were obtained from Jackson Laboratories or bred in-house. The study was conducted in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals. The protocol was approved by the Animal Care and Use Committee at Johns Hopkins University. Animals were housed in specific pathogen-free animal facilities. ^21–25

Allogeneic Sensitization and Skin Grafting.

For allogeneic sensitization, Balb/c H-2^d splenocytes were harvested and processed to a single-cell suspension.^21–25 Approximately 20×10⁶ cells were administered via intraperitoneal (IP) injection into C57Bl/6 H-2^b mice for 3 consecutive days. Full-thickness tail and ear skins were transplanted onto the dorsal thorax of recipient mice and secured with adhesive bandages as previously described.²⁶

MHC Tetramer Enrichment and Flow Cytometry.

Biotinylated MHC monomers with human β2-microglobulin specific for H-2L^d β-galactosidase (TPHPARIGL) were obtained from the NIH Tetramer Core Facility and tetramerized with streptavidin-PE (Prozyme). Decoy reagent was made according to published protocols.²⁷ Briefly, streptavidin-PE (Prozyme) was conjugated to AlexaFluor647 using the Antibody Labeling Kit (ThermoFisher). The concentration decoy reagent was quantified using a NanoDrop. For flow cytometry and magnetic enrichment were performed using FACS buffer comprised of 1X PBS (without Ca²⁺/Mg²⁺, pH 7.2), 0.25% BSA, 2 mM EDTA, and 0.09% azide. Single cell suspensions of splenocytes were incubated with decoy reagent, followed by MHC tetramer, and enriched using anti-PE beads and paramagnetic LS columns (Miltenyi) according to published protocols. The column bound fraction was analyzed as the MHC tetramer enriched fraction, and the flow-through was collected as the unbound or bulk fraction.

Flow Cytometry Analysis.

B cell populations were identified as live (Zombie Near IR⁻; Biolegend), lineage (Thy1.2, NK1.1, F4/80)⁻, and CD19⁺B220⁺. Both fractions were stained with surface antibodies for 30 min at room temperature and intracellular antigens were assessed using the Transcription Factor Staining Kit (eBiosciences). Populations were identified using antibodies against CD19, CD21/35, CD23, CD38, CD45R (B220), CD73, CD80, F4/80, GL7, IgD, IgM, IRF4, NK1.1, PD-L2, and/or Thy1.2 (Biolegend, ThermoFisher, BD Biosciences). All experiments were acquired using a 4L Cytek Aurora, and gating was performed using FlowJo (BD Biosciences). For dimensionality reduction analysis, manually gated bound H-2L^d tetramer⁺ IgD^lo B cell events were imported into R (4.1.1) through CytoML (2.40), flowWorkspace (4.4.0), and flowCore (2.4.0). The data were further analyzed using CATALYST (1.16.2) with FlowSOM (2.0.0) clustering and ConsensusClusterPlus (1.56.0) meta clustering. UMAP dimensionality reductions were generated with scater (1.20.1). Visualizations were generated in ggplot2 (3.3.5) and ComplexHeatmap (2.8.0).

Cellular H-2^d Crossmatch.

Serum was collected from C57Bl/6 mice before or at indicated times after allogeneic sensitization. Single cell suspensions of Balb/c splenocytes were prepared, and 1×10⁶ splenocytes were incubated with 1 μL of serum for 1 h at room temperature. Cells were washed 2X with FACS buffer, followed by anti-mouse IgM, anti-mouse IgG, and anti-B220.

Depletion of Marginal Zone Subsets.

Depletion of the Marginal Zone population in vivo was performed via intraperitoneal injection with 100 μg each anti-CD11a/anti-CD49d or IgG isotype control (BioXCell) on days −4 and −2 prior to sensitization, according to established protocols.^{16,17,28–31}

Adoptive Transfer of B Cell Populations.

CD45.1 B cells were isolated using the MZ and FO B Cell Isolation kit (Miltenyi Biotec). Cells were resuspended in PBS and 1–2×10⁶ MZ or 1–2×10⁷ FO B cells were adoptively transferred via lateral tail vein injection.

Immunofluorescence Microscopy.

Triple immunolabeling was performed on formalin‐fixed, paraffin embedded sections on a Ventana Discovery Ultra autostainer (Roche Diagnostics). Briefly, following dewaxing and rehydration on board, epitope retrieval was performed using Ventana Ultra CC1 buffer (Roche Diagnostics) at 96^oC for 64 minutes. Primary antibodies for CD4 (Abcam), CD19 (Roche Diagnostics), and Ki67 (Abcam) were successively applied at 36^oC for 40 minutes and detected using anti-rabbit HQ detection (Roche) with OPAL 520, OPAL 570 (FP1488001KT), and OPAL 690 (FP1497001KT), respectively, diluted 1:150 in 1X Plus Amplification Diluent (Akoya Biosciences). Following each antigen detection, primary and secondary antibodies were stripped on board using Ventana Ultra CC1 buffer at 95^oC for 12 minutes followed by neutralization using Discovery Inhibitor (Roche Diagnostics). Tissue was then counterstained with spectral DAPI (FP1490, Akoya Biosciences) and mounted with Prolong Gold (P36930, ThermoFisher Scientific). Whole slide images were generated using a Zeiss Axioscan 7. Representative images were captured with QuPath.

Statistical Analysis.

All data points represent individual animals. Expression levels were compared using Student’s t-test (two-tailed) or ANOVA where appropriate. Statistics were performed using GraphPad Prism 9. Significance was determined as *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

RESULTS

Allogeneic sensitization of H-2^d cells induces anti-donor antibodies and modest frequencies in isotype-switched B cells.

In order to investigate the differentiation of B cells induced by sensitization with allogeneic antigen, we utilized an established method of repeatedly exposed C57Bl/6 H-2(b) MHC haplotype hosts with allogeneic H-2(d) MHC haplotype Balb/c antigen (Figure 1A).²⁵ This model allows the assessment of anti-H-2(d) IgM and IgG via cellular crossmatch and antigen-specific B cells using MHC Class I H-2L(d) tetramers. We assessed the formation of anti-donor IgM and IgG after sensitization and found that anti-donor IgM was significantly increased at day 7 and diminished at day 14 and 21 (Figure 1B). Anti-donor IgG increased at day 7 and remained elevated at day 14 and 21 (Figure 1C). Thus, exposure to allogeneic antigen in this manner effectively induces anti-donor IgM and IgG responses.

Figure 1. Allogeneic sensitization induces IgG and IgM antibodies and modest germinal center populations.

(A) C57Bl/6 mice were sensitized by intraperitoneal administration of Balb/c (H-2^d) splenocytes (days 1, 2, and 3). (B) Anti-donor IgM and (C) anti-donor IgG were assessed in the serum on day 0–21 post-transplant. (D) Representative staining and absolute cell counts for H-2L(d) tetramers in naïve and MD4 RAG1 KO mice (E) The expression of IgM and IgD on bulk CD19⁺ and H-2L^d tetramer⁺ B cells on day 0 and 14 post-sensitization. Class switched Ig (SwIg) are defined as IgM^loIgD^lo B cells. (F) The frequency of CD38^loGL7⁺ germinal center B cells on day 0 and 14 post-sensitization. The frequency of IgM^lo cells among germinal center B cells and non-germinal center (IgD^loCD38⁺GL7⁻) (G) bulk CD19⁺ B cells and (H) H-2L(d) tetramer⁺ B cells. Each data point represents an individual mouse, and all summary data represents compiled data from 2–3 independent experiments including (B-C) n=10–20/group, (D) 6–12/group, (E-F) n=5–9/group, (G)_ n=6/group. ***p<0.001, ****p<0.0001.

In order to assess the antigen-specific B cell compartment, we used an MHC Class I tetramer-based approach to sensitively identify H-2L(d)-specific B cells and the bulk B cells (Figure S1A–B). This approach enabled the specific identification of B cells in pre-sensitized naïve C57Bl/6 mice (Figure 1D), as no tetramer binding B cells were found in mice transgenic for the hen egg lysozyme (MD4 mice; Figure 1D). Consistent with prior reports, approximately 7400 ± 999 L(d) tetramer⁺ B cells were found in naïve C57Bl/6 mice prior to sensitization.²¹

We next assessed the B cell receptor isotype among bulk and H-2L^d-specific B cells after sensitization. Despite the formation of anti-donor IgG in the serum by day 14, we found no significant changes in the isotype profile of B cells on day 14 relative to day 0 (Figure 1E). We assessed the formation of germinal center CD38^loGL7⁺ B cells, and we found that within the IgD^lo compartment, approximately 20% of L(d) tetramer⁺ B cells had differentiated into germinal center B cells (Figure 1F). Although the overall frequency of isotype class switching was not different among bulk or antigen-specific B cells (Figure 1E), the germinal center B cells were universally IgM^lo (Figure 1G–H). Thus, allogeneic sensitization induces a durable donor-specific IgG formation but relatively modest number of class-switched B cells.

Allogeneic sensitization induces few classical memory B cell populations

Antigen-experienced B cells are typically defined by the expression of three markers, CD73, PD-L2, and CD80.^14,32 We assessed the expression of these receptors in bulk and antigen-specific H-2L^d B cells on day 21 after allogeneic sensitization by comparing naive IgD^hi and activated IgD^lo populations. Although CD73 and PD-L2 were modestly increased in bulk IgD^lo B cells relative to IgD^hi B cells (Figure 2A and Figure S1C), there was no difference in the expression of these receptors in antigen-specific B cells (Figure 2B and Figure S1D). Thus, although these receptors can be used to distinguish antigen-experienced and memory B cells induced after infection or hapten immunization, they do not appear to be specifically upregulated after allogeneic sensitization.

Figure 2. Analysis of the phenotype of antigen-experienced B cells after allogeneic sensitization.

The expression (MFI and % positive) of anti-experienced B cell markers CD73, CD80, and PD-L2 were assessed on (A) bulk CD19⁺ B cells and (B) H-2L(d) tetramer⁺ B cells on day 21 after allogeneic sensitization. (C) The phenotypic profiles of naïve and day 21 H-2L(d) tetramer⁺ B cells were clustered and visualized using t-SNE. (D) Frequencies of each cluster in both B cell populations. (E) Histograms of phenotypic markers in each individual cluster on day 21 post-sensitization. Each data point represents an individual mouse, and all summary data represents compiled data from 2–3 independent experiments including (A-B) n=6–8/group, (C-E) 1–3/group. ***p<0.001.

In order to evaluate the global phenotypic changes that occur after allogeneic sensitization, we evaluated the phenotype of antigen-specific IgD^lo B cells on day 0 and day 21 after sensitization using hierarchical clustering and dimensionality reduction visualization with t-SNE (Figure 2C–D). Only one cluster, Cluster 3, appeared to be IgM^lo, consistent with a class switched memory phenotype B population, but this cluster was not high for the memory markers PD-L2 or CD73, consistent with our manual gating (Figure 2A–B). Cluster 2 was consistent with a germinal center population that was GL7⁺Ki67⁺ and intermediate for CD73 and CD38 expression (Figure 2E). Clusters 4, 5, 6, and 7 were high or intermediate for the complement receptors CD21/35. Three of these clusters appeared to be enriched in day 21 versus day 0 and expressed low levels of CD23 and Ki67. These clusters expressed high levels of IRF4 and variable levels of PD-L2. Thus, clustering and dimensionality reduction analysis of IgD^lo B cells revealed that post-sensitization B cells appear to be enriched for an IgM^hi CD21/35⁺ marginal zone B cell phenotype.

Allogeneic sensitization induces an increased frequency of IgM^hi marginal zone phenotype B cells that have high IRF4 expression.

We next evaluated whether the phenotype of enriched clusters corresponded to populations changes detected by manual gating from day 0–42 post-sensitization. We evaluated both the frequency of IgD^hi follicular B cells and the frequency of CD21⁺ B cells among IgD^lo cells in both antigen-specific and bulk B cells. While the frequency of IgD^hi follicular B cells was unchanged after sensitization, we found that the frequency of marginal zone B cells increased at day 21 and 42 post-sensitization in both subsets relative to naïve mice (Figure 3A). Among IgD^lo B cells at day 21 post-sensitization, both bulk and tetramer⁺ CD21⁺ B cells expressed higher levels of IgM versus CD21^lo populations (Figure 3B), consistent with our dimensionality reduction analysis results (Figure 2). We also evaluated the expression of the transcription factor IRF4, expression levels of which correspond to the strength of prior antigen signaling in lymphocytes.³³ We found that IRF4 expression was significantly enriched in both bulk and tetramer+ CD21^hi B cell populations relative to CD21^lo populations (Figure 3C). Taken together, these data demonstrate that allogeneic sensitization is associated with an increase in marginal zone B cells with an antigen-experienced IRF4⁺ phenotype.

Figure 3. The frequency of CD21^hiIgM^hi marginal zone B cells increases after allogeneic sensitization.

(A) The frequency of IgD^hi B cells and IgD^lo CD21⁺ B cells among bulk and H-2L(d) tetramer⁺ B cells on day 0, 14, 21, and 42 post-transplant. (B-C) Gated on IgD^lo B cells. (B) The frequency of IgM^hi bulk and H-2L(d) tetramer+ B cells on day 21 among CD21^hi and CD21^lo populations. (C) The frequency of IRF4⁺ B cells among CD21^hi and CD21^lo bulk and H-2L(d) tetramer⁺ B cells. Each data point represents an individual mouse, and all summary data represents compiled data from 2–3 independent experiments including (A) n=6–9/group (B-C) n=6/group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Depletion of the marginal zone B cells after allogeneic sensitization significantly restrains the formation of anti-donor IgG and germinal center B cells

In order to evaluate whether the marginal zone B cells are a precursor population of allogeneic B cells, we depleted marginal zone B cells using an antibody-based approach^{16,17,28–31} We examined the specificity of this approach and found that while the depletion of marginal zone B cells was nearly complete (Figure S2A), the total B cell and T cell compartments were unperturbed (Figure S2B–D). In addition, consistent with published work,¹⁶ marginal zone depleted mice had in tact follicular structures (Figure S2E). We first assessed the formation of anti-donor antibody in marginal zone depleted or isotype control treated mice (Figure 4A–B). Using a cellular crossmatch, we found that at day 7 post-sensitization the anti-donor IgM was diminished in marginal zone-depleted hosts (Figure 4C). We found that anti-donor IgG was diminished in marginal zone-depleted hosts at both day 7 and Day 14 (Figure 4D).

Figure 4. Depletion of marginal zone B cells diminishes anti-donor IgG and the formation of germinal center B cells.

(A) Mice were treated with marginal zone depletion antibodies or isotype control prior to sensitization. (B) Representative flow plots of CD21^hi marginal zone B cells after depletion or control IgG treatment. (C) Anti-donor IgM in the serum on day 7 and 14 in mice marginal zone depleted (MZ Depl.) or isotype control (Control) mice. (D) Anti-donor IgG in the serum on day 7 and 14 in mice marginal zone depleted (MZ Depl.) or isotype control (Control) mice. (E) The frequency of germinal center B cells among bulk and H-2L(d) tetramer⁺ B cells on day 14 post-sensitization. Each data point represents an individual mouse, and all summary data represents compiled data from 2–3 independent experiments including 6–9/group. **p<0.01, ***p<0.001.

We next assessed the phenotype of bulk and antigen-specific B cells in wild-type and marginal zone depleted hosts. We found that the frequency of germinal-center B cells was slightly diminished in bulk B cell populations, but this did not reach statistical significance (Figure 4E). The formation of antigen-specific germinal center B cells, however, was significantly diminished (Figure 4E). Taken together, these data demonstrate that marginal zone B cells are required for the full formation of anti-donor IgG and the majority of germinal center B cells after sensitization.

Marginal Zone B cells are required for the formation of anti-donor IgG and GC B cells after allogeneic skin grafting

While our data demonstrates the importance of marginal zone B cells in response to cell-based allogeneic sensitization, we next sought to determine whether the same was true for a model of transplantation. We performed fully allogeneic Balb/c skin grafts onto control or marginal zone depleted C57Bl/6 mice (Figure 5A). We assessed the formation of anti-donor IgM and IgG using a cellular crossmatch. We found similar levels of anti-donor IgM and IgG in both groups on day 7 post-graft (Figure 5B). Although marginal zone B cells have been shown to rapidly differentiate into antibody-secreting plasmablast populations after antigen-encounter, we did not observe an increase in CD138+ or IRF4+ B cells among anti-L(d) tetramer+ or bulk B cells in the spleen (Figure S3).

Figure 5. Depletion of marginal zone B cells diminishes anti-donor IgG and the formation of germinal center B cells.

(A) Mice were treated with marginal zone depletion antibodies or isotype control prior to Balb/c skin grafting. Anti-donor IgM and anti-donor IgG in the serum on (B) day 7 post-graft or (C) day 14 post-graft in marginal zone depleted (MZ Depl.) or isotype control (Control) mice. Frequency of germinal center B cells among bulk and L(d) tetramer+ B cells in the (D) spleen and (E) graft draining lymph nodes. Numbers in (E) depict the frequencies of IgD^lo germinal center B cells. (F) Fluorescence immunohistochemistry of spleen and draining lymph nodes on day 14 post-graft from MZ depleted or Control mice. (G) Anti-donor IgG on day 42 or later post-graft in MZ depleted or Control mice. Each data point represents an individual mouse, and all summary data represents compiled data from 2–3 independent experiments including (B-C, G) n=6–8/group, (D) n=8–10/group. **p<0.01, ***p<0.001.

On day 14 post-graft, marginal zone depleted mice had significantly reduced levels of anti-donor IgG (Figure 5C). In the spleen at this timepoint, there was a trend towards fewer germinal center B cells in both the antigen-specific and bulk B cells (Figure 5D). In the draining lymph nodes, however, marginal zone B cell depleted mice had significantly lower frequency of germinal center B cells relative to control treated mice (Figure 5E). We also sought to corroborate this finding in situ using immunofluorescence microscopy. We found that the marginal zone depleted mice had noticeably reduced formation of germinal centers relative to control mice in both the draining lymph nodes and the spleen (Figure 5F). The impact of marginal zone depletion during antigen priming was prolonged, as less anti-donor IgG was produced in depleted mice at memory timepoints after transplantation (Figure 5G).

Marginal zone B cells preferentially form germinal center B cells after skin grafting

While our results demonstrate that marginal zone B cells are required for the optimal formation of germinal centers after skin grafting, it is not clear whether this is due to direct differentiation of marginal zone B cells into germinal center B cells or an indirect effect on follicular B cells. To assess the fate of follicular and marginal zone B cells after skin grafting, we conducted adoptive transfer experiments (Figure 6A). We isolated naïve follicular and marginal zone B cells from CD45.1 mice, and adoptively transferred them into naïve CD45.2 MD4 hosts, which are B cell transgenic for hen egg lysozyme. In this experimental setup, hosts have an intact T cell compartment but the host B cells are unable to respond to allogeneic antigen. Follicular and marginal zone B cells are present at an approximate 10:1 ratio in vivo, and accordingly we transferred 1–2×10⁷ and 1–2×10⁶ of each respective population.

Figure 6. Marginal zone B cells efficiently differentiate into germinal center B cells and produce anti-donor IgG.

(A) Naïve B cells from CD45.1 mice were isolated, and 1–2×10⁷ follicular or 1–2×10⁶ marginal zone B cells were adoptively transferred into CD45.2 MD4 hosts. Hosts were grafted with Balb/c skin and assessed on day 7–14 post-transplant. (B) Anti-donor IgM and anti-donor IgG production on day 14 post-graft. (C) Absolute number of transferred CD45.1 B cells in the spleen, draining lymph nodes, and blood. (D) Frequency of germinal center B cells among transferred follicular or marginal zone B cells on day 14 in the draining lymph nodes and spleen. Each data point represents an individual mouse, and all summary data represents compiled data from 2–3 independent experiments including n=5–8/group. *p<0.05, **p<0.01, ***p<0.001.

We assessed the formation of anti-donor IgM and IgG after Balb/c skin grafting. We found that mice with marginal zone B cells produced similar levels of anti-donor IgG versus follicular B cell mice on day 14 post-graft (Figure 6B). We next evaluated the fate of each population on day 14 in the spleen and draining lymph nodes. We found that marginal zone originating B cells persisted in primarily in the spleen, lymph nodes, on day 14 post-graft, and few were found circulating in the blood (Figure 6C). Finally, we found that the frequency of germinal center B cells was higher among marginal zone originating B cells relative to follicular originating B cells (Figure 6D). Together, this experiment demonstrates that marginal zone B cells efficiently form germinal center phenotype B cells and form anti-donor IgG producing populations.

DISCUSSION

We used an MHC tetramer-based approach to study the differentiation of naïve B cells in a model of fully allogeneic sensitization and transplantation.²⁷ Using a cellular sensitization approach,^21–24,34 we found that although soluble anti-donor IgM and IgG were readily formed, very few class switched B cells were generated. We also found that allogeneic sensitization did not result in antigen-specific B cells differentiating into subsets of memory B cells that express high levels of the surface receptors CD73, CD80, and PD-L2.^14,35 These receptors have been used for identifying memory B cells after hapten immunization and viral infection in mice.³² It is not clear whether the relative absence of these markers is due to differences in the antigen properties (e.g. affinity, duration, concentration) or inflammatory signals between these different models.

Using clustering analysis, we found that antigen-specific B cells primarily expressed high levels of IgM and CD21 after allogeneic sensitization, consistent with a marginal zone B cell phenotype. Using manual gating, we found that the frequency of CD21⁺ marginal zone B cells were increased after day 21–42 following sensitization. Functionally, we found that marginal zone B cell depleted mice had diminished formation of anti-donor IgG and germinal center B cells relative to controls. We extended these findings to a model of skin graft rejection, in which we found that marginal zone B cell depleted mice had reduced anti-donor IgG and germinal center B cells. Fate tracking experiments showed that relative to follicular B cells, marginal zone B cells preferentially formed germinal center B cells. These data suggest a model in which the marginal zone B cell population is a precursor population for differentiating allogeneic B cells, and that class switching via both early plasmablasts and germinal center reactions draws heavily from the marginal zone B cell pool.

Marginal zone B cells are defined by their localization in the marginal zone surrounding the B cell follicle of the spleen. The fate decision of transitional B cells to become marginal zone B cells is based on antigen signal strength and dependent on continuous Notch2 signaling.⁸ Relative to follicular B cells, marginal zone B cells are more sensitive to innate signaling via PRR, (such as LPS/TLR4), participate in T-independent B cell responses, and readily differentiate into plasma cells.^11–13 However, in addition to being sensitive to innate PRR signaling, marginal zone B cells are also recognized to participate in T cell-dependent responses and have several features of differentiated or antigen-experienced B cells^11,14 In humans, marginal zone B cells are considered a memory B cell subset.^9,32,36 Thus, despite the innate-like features that distinguish them from follicular B cells, there is considerable evidence that marginal zone B cells can participate in the response to T dependent responses to antigen such as foreign MHC. Recent reports have implicated marginal zone B cells in the formation of DSA and graft rejection after T cell depletion and have shown that Notch2 deletion can prolong cardiac allograft survival ^18,37.

Multiple studies have found that humoral immunity to foreign MHC involves germinal center reaction and the formation of follicular T helper cells (T_FH).² Circulating T_FH have been identified as a marker of ongoing alloimmunity and graft rejection ^2,38,39. Our findings are not incongruent with this work, but rather demonstrate that the marginal zone B cells, not follicular B cells, are the major reservoir from which the germinal center B cells are derived. It is not clear whether the selective activation of MZ B cells versus follicular B cells is due to differences in the threshold for activation (e.g. signaling through inflammatory mediators) or the BCR specificity for allogeneic antigen. Future studies are needed to investigate the interaction between transplant-relevant inflammatory mediators (e.g. the ischemia reperfusion injury, and cell death), and the activation of marginal zone B cells ⁴⁰. Furthermore, this work did not extensively evaluate the direct impact of marginal zone B cells on T cell alloreactivity, which could be altered due to the cell extrinsic functions (e.g. antigen presentation, cytokine production, costimulation) of the marginal zone B cell population.

The use of MHC tetramers to identify antigen-specific B cells has been previously established in multiple models,^21,41–43 and the specificity of allogeneic tetramer+ B cells was recently demonstrated using biophysical methods.⁴⁴ However, given the abundance of naïve B cells identified using tetramers by our group and others, one limitation to this approach is the potential for staining non-specific or non-functional B cells.²¹ In addition, B cells have known antigen-presentation functions in the setting of allogeneic antigen.⁴⁵ Additional work is needed to dissect these facets of allogeneic B cell function in transplantation.

Allogeneic sensitization remains a significant barrier to improved outcomes for transplant patients prior to and after transplantation.⁴⁶ Deeper understanding of the mechanisms by which HLA sensitization occurs could lead to better risk stratification of transplant patients and original approaches to prevent rejection.^4,5 This work has implications for both desensitization approaches and prevention of antibody-mediated rejection. For example, the anti-CD20 monoclonal antibody rituximab, which is often used in desensitization protocols to deplete circulating B cells, has variable efficacy at lymphoid tissue resident B cell populations such as marginal zone B cells.⁴⁷ Because of the importance of the germinal center reaction in the formation of anti-MHC IgG from either follicular or marginal zone B cell precursors, marginal zone B cell-based anti-donor IgG production is likely susceptible to the major classes of immunosuppressive drugs (e.g. calcineurin inhibitors and CD28 blockaders). In particular, as blockade of the CD28 pathway has been shown to potently inhibit germinal center B cells through the inhibition of follicular helper CD4⁺ T cells,^48–51 CD28 blockade is likely to effectively inhibit the differentiation of allogeneic marginal zone B cells. However, future studies will be needed to evaluate whether marginal zone and follicular B cells are differentially impacted by immunosuppressive agents.

ABBREVIATIONS:

GC

germinal center

MHC

Major Histocompatibility Complex

MZ

Marginal Zone

FO

Follicular