B lymphocytes contribute to indirect pathway T cell sensitization via acquisition of extracellular vesicles

B cells have been implicated in transplant rejection via antibody‐mediated mechanisms and more recently by presenting donor antigens to T cells. We have shown in patients with chronic antibody‐mediated rejection that B cells control the indirect T cell alloresponses. To understand more about the role of B cells as antigen‐presenting cells for CD4+ T cell with indirect allospecificity, B cells were depleted in C57BL/6 mice, using an anti‐CD20 antibody, prior to receiving MHC class I–mismatched (Kd) skin. The absence of B cells at the time of transplantation prolonged skin graft survival. To study the mechanisms behind this observation, T cells with indirect allospecificity were transferred in mice receiving a Kd skin transplant. T cell proliferation was markedly inhibited in the absence of recipient B cells, suggesting that B cells contribute to indirect pathway sensitization. Furthermore, we have shown that a possible way in which B cells present alloantigens is via acquisition of MHC‐peptide complexes. Finally, we demonstrate that the addition of B cell depletion to the transfer of regulatory T cells (Tregs) with indirect alloresponse further prolonged skin graft survival. This study supports an important role for B cells in indirect T cell priming and further emphasizes the advantage of combination therapies in prolonging transplant survival.


| INTRODUC TI ON
Solid organ transplantation is the best treatment available for organ failure; however, many transplants are lost within 12 years.
The biggest single cause of transplant failure is immune-related injury.
T cell-mediated rejection is initiated when T cells are stimulated in response to graft alloantigens (alloAgs) via three major pathways of allorecognition. (1) The "direct pathway" where recipient T cells recognize intact donor major histocompatibility complex (MHC) molecules from donor antigen-presenting cells (APCs), and (2) the "indirect pathway" where presentation of donor alloAgs occurs after processing and presentation by host recipient APCs. 1 In the third pathway of allorecognition, the "semi-direct," recipient APCs acquire intact donor MHC molecules and present to recipient T cells with direct allospecificity. 2  The presence of antibodies (Abs) against HLA is in most cases a predictive marker of transplant failure, and the level of circulating Abs against donor antigens correlates with chronic Ab-mediated rejection. 6,7 B cells are best known for their unique role as Ab-producing cells and Abs binding to the graft lead to the deterioration of the organ by Ab-mediated rejection mechanisms. Recent data have shown an additional role for recipient B cells, demonstrating that these cells can contribute to graft survival/rejection by presenting alloAgs to recipient T cells. [8][9][10] As discussed later, we have shown that B cells specific for alloAgs are particularly efficient in amplifying T cell responses with indirect allospecificity. 9 It is important to note that there is compelling data from murine models suggesting that the antigen-presenting functions of B cells, rather than antibody formation, are the primary reason why B cells are required for chronic rejection. [11][12][13][14] Depletion of B cells has been applied in different murine models to improve transplant survival of MHC mismatched allografts.
However, a diverse range of outcomes has been observed. In one study, B cell depletion did not alter the survival of fully mismatched kidney or heart allografts, 8 whereas in other studies accelerated rejection was observed. 15 The latter was hypothesized to be a consequence of depletion of IL-10-producing regulatory B cells (Bregs), which serve to suppress antigen-specific responses. [16][17][18][19][20] Furthermore, although some of the early experiments using depleting Abs suggested that B cells play an important role during T cell priming, [21][22][23] further research using genetically modified mice deficient in the B cell compartment produced conflicting results. 14,[24][25][26][27][28][29] In the clinical setting, several desensitization therapies aiming to target donor-specific antibody (DSA) levels in highly sensitized patients prior to transplantation are being tested, many of which involve B cell depletion. However, the differences in protocols, particularly the time of administration of the B cell-depleting drug, such as an anti-CD20 Ab rituximab, make the interpretation of the results difficult. One study has reported that rituximab induction therapy may promote rather than inhibit acute cellular rejection in some patients, 30 with one of the mechanisms suggested being depletion of B cells with regulatory function. In agreement with this idea, we have recently shown that operationally tolerant kidney-transplanted patients have "transitional" regulatory B cells that maintain high levels of IL-10 expression, further supporting the concept that some subpopulations of B cells might enhance graft survival. 20 Furthermore, we have recently published that B cells make complex and dynamic contributions to the patterns of alloresponses seen in patients with chronic Ab-mediated rejection, including patterns consistent with suppression of alloresponses in some. The data suggest that B cells, in addition to allo-Ab production, are able to induce indirect T cell alloresponses by acting as APCs in some, but suppress indirect alloreactive T cell responses, at least in vitro, in others. 9 In this study, the role of B cells as APCs for indirect CD4 + T cell alloresponses was examined in vivo. B cells were depleted (using an anti-CD20 Ab) in recipient mice receiving MHC class I-mismatched (K d ) skin grafts. We demonstrate that B cells contribute to the activation of CD4 + T cells with indirect alloresponses via acquisition of donor MHC molecules through the uptake of extracellular vesicles.

| Skin transplantation
Donor tail skin grafts were performed and monitored as previously described. 31 In some experiments, anti-CD8 antibody (clone YTS169, 250µg/injection/mouse) was injected intraperitoneally at day 1 before and day 1 after skin graft, and weekly thereafter. When described, B cells were depleted with 200µg of anti-CD20 (clone 5D2, kindly provided by Genentech) by intravenous injection 7 days before skin transplantation.

| Generation of CD4 + CD25 + Treg line
Tregs with indirect specificity were created as outlined by Tsang et al. 32 (For more details see SI.)

| CFSE labeling of DCs and isolation of EVs
Bone marrow (BM)-derived dendritic cells (BM-DCs) were generated as previously described. 33 EVs were prepared as described in the SI. 34 To measure the concentration and size of the particles iso-

| Statistical analysis
Graft survival data were analyzed using the Kaplan-Meier method, with the Wilcoxon rank test and the log-rank test used to verify the significance of difference between the groups (GraphPad Prism).
Statistical analysis of other data was performed using the two-tailed Student's t test for unpaired samples with unequal variance.

| B cell subpopulations in the spleen express high levels of CD20
To further understand the contribution of B cells as APCs in transplant rejection, an anti-CD20 Ab was used to deplete B cells. CD20 is first expressed in human pre-B cells in the BM, and its expression continues until B cells differentiate into plasma cells. 35,36 In mouse, several reports have shown a similar CD20 expression pattern. 37 In this study, we used a multi-parametric analysis to investigate the expression of CD20 by B cells during their development. 38,39 From pre-pro-B cell (B220 + CD43 + CD24 low BP-1 neg ) through pro-B cell (B220 + CD43 + CD24 + BP-1 neg ) until pre-B-I cell (B220 + CD43 + CD24 + BP-1 + ), stage CD20 was not expressed (Figure 1, Figure S2). However, CD20 was expressed at low densities at pre-B-II cell stage (B220 + CD43 + CD24 high BP-1 + ) and increased further during B cell maturation ( Figure S3). This is in line with previous reports, where it was shown that CD20 expression is parallel to the expression of immunoglobulin heavy chain. 37 B cells egress from BM and migrate to secondary lymphoid organs.
We analyzed the expression of CD20 on B cell subpopulations from spleen. Although all subpopulations analyzed expressed CD20, transitional 1 (T1) and T2 B cells (B220 + CD21 high sIgM high CD93 + CD23 neg and CD23 + , respectively), as well as marginal zone (MZ) and MZ precursors (B220 + CD21 low sIgM high CD23 neg and CD23 + , respectively) expressed CD20 at very high levels ( Figure 1, Figure S3). Of note, MZ precursors are also described by some authors as T2 CD21 low . 38 Although CD20 expression increases with B cell maturation, it has been reported that once B cells enter the B220 high B cell pool, CD20 expression decreases. This suggests that transitional B cells that express high levels of CD20 are recent BM emigrants, whereas mature B cells (eg, follicular B cells) express lower densities of CD20 compared to T1, T2, and MZ B cells ( Figure S3).

| Anti-CD20 antibody efficiently depletes mature B cells
After showing the distribution of CD20 in the different B cell subpopulations in the BM and in the periphery, we injected B6 naïve mice intravenously with the anti-CD20 Abs, and the B cells populations were monitored. In our hands, the 5D2 clone administered in a single dose efficiently depleted B cells rapidly from blood, and by day 14, late pre-B cells to mature stages ( Figure 2) could not be detected in BM, whereas early precursors (ie, pre-pro-B cells to early pre-B cells) remained intact ( Figure 2). These results are in line with those of previous reports using anti-CD20 Abs and consistent with CD20 expression patterns by BM B cells ( Figure S2).
In the spleen, CD20 is highly expressed by recent BM emigrants and transitional B cells and MZ precursor cells ( Figure S3).
Concurrently, these IgM high CD24 high B cell populations with the highest CD20 expression levels were the first cells to disappear from the spleen after treatment with anti-CD20 Abs ( Figure S4, day 4).
Mature B cells, such as follicular B cells, which express CD20 at a lower density than recent emigrants were not completely depleted until day 7 in spleen ( Figure 3) and lymph nodes (LNs) ( Figure S5).
We confirmed that the depletion of B cells was long-lasting in the absence of inflammation, as after 2months B cells were still depleted ( Figure S4, day 60).

| The absence of B cells at the time of skin transplant is necessary to induce prolongation of skin graft survival
After we confirmed that anti-CD20 Ab treatment depletes B cells, B6 mice transplanted with skin from a B6 mouse strain transgenic for K d molecules (B6.K d ) received anti-CD20 Ab either 7days before (−d7) or 2days after (+d2) transplantation ( Figure 4A). Furthermore, to focus on the antigen-presenting function of B cells for T cells with indirect allospecificity, recipient mice were injected with an anti-CD8 Ab at day −1 and +1, and then every week until mice completely rejected the skin graft, as published previously. 31 Of note, anti-CD8 treatment depletes both CD8 + T cells and CD8 + DCs ( Figure S6). Skin graft survival was prolonged for 8 days but only when anti-CD20 Ab was injected 7 days before transplantation but not 2days after (phosphate-buffered saline [PBS]; +d2 and -d7 mean graft survival time were 11, 11.5, and 19 days, respectively) ( Figure 4B).
Although anti-CD8 Ab alone did not prolong graft survival, anti-CD20 prolonged skin graft survival only when combined with anti-CD8 Abs ( Figure S7 and Figure 4B). These results indicate that B cell depletion prolongs graft survival only after depletion of both CD8 + T cells and CD8 + DCs, strongly suggesting that B cells are presenting the K d alloantigen to CD4 + T cells, most likely because graft rejection in this experimental setup is dependent on indirect allorecognition by CD4 + T cells.
A regulatory IL-10-producing B cell population has been re-   Figure S1. Splenocytes were gated on lymphocyte live CD11b neg CD19 + cells. Following Allman and Pillai gating strategy, cells were gated based on their CD21 and IgM surface (sIgM) expression. sIgM high CD21 high cells were further divided into transitional 1 and transitional 2; sIgM + CD21 + were divided into follicular I and follicular II; sIgM high CD21 + were divided into marginal zone (MZ) and MZ precursor cells, as detailed in Figure S2. Overimposed black dot plots showed CD20-expressing cells. Data are representative of 3 independent experiments

| B cell depletion led to a delay in priming of CD4 + T cells with indirect allospecificity
To understand the mechanisms behind the increased skin survival following B cell depletion, and in the absence of CD8 + T cells and

| B cells are efficient in donor antigen capture
We have shown previously that graft rejection was attributed to   Figure 7B). This suggests that intact class II (I-A b ) molecules presenting K d peptide were transferred onto CBA B cells.
Altogether these results demonstrate that recipient B cells can acquire both intact MHC class I molecules and MHC class II molecules containing class I-derived peptides via EV transfer, and that both can activate T cells with indirect specificity, the first requiring re-cycling of the newly acquired, EV-transferred membrane MHC class I via the endogenous MHC class II antigen processing pathway for presentation onto MHC class II, but the second not. Therefore these data support our hypothesis that B cells can acquire intact MHC via EV transfer to efficiently stimulate CD4 + T cells recognizing alloantigen in a self-MHC class II-restricted manner.

| D ISCUSS I ON
In this study we have investigated the contribution of B cells in skin graft transplant survival/rejection. Uchida J et al observed that the effectiveness of anti-CD20 Abs to deplete B cells in vivo closely correlated with the subtype. The most efficient depletion was achieved by using immunoglobulin G isotype 2a (IgG2a), through a mechanism involving both Fc-γ receptor and macrophages. 46 In this study we used a murine version of rituximab to deplete B cells, the monoclonal IgG2a anti-CD20 Ab, clone 5D2, which mediates cellular phagocytosis of circulating B cells in the liver. 47 The anti-CD8-depleting Ab was administered to prevent the CD8 + T cell direct alloresponse to donor K d , but by also depleting the major population of lymphoid resident DC, it allowed us to assess the importance of B cells as APCs to prime CD4 + T cell indirect alloresponse in the absence of CD8 + DCs. Of F I G U R E 5 B cell depletion delays CD4 + T cells indirect alloresponses. B6 mice (CD90.2) were injected intravenously with anti-CD20 Ab (200µg) 7days before receiving a dorsal B6.K d skin transplant (day 0). Anti-CD8 depleting Ab was administered at day −1 and day +1 to mice that have received or not anti-CD20 Ab. Congenic (CD90.1) CFSE-labeled TCR75 cells were adoptively transferred at early and late time points as indicated in the experimental design (A). Proliferation by CFSE dilution of CD90.1 + cells of 2 independent experiment from pooled draining lymph node and spleens (n = 4-5/ group for each experiment) was analyzed by flow cytometry (B). *P < .05, **P < .01   In the clinic, rituximab is currently being added to immunosuppressive regimens, mostly when transplanting across blood-group antibody barriers, however, with no real understanding of whether it gives benefit above other depleting agents. Anti-CD20 Abs have no effect on plasma cells, since they do not express CD20;  Care.

D I SCLOS U R E
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author, G.L., upon reasonable request.