Reversing pregnancy-induced B cell sensitization to facilitate the induction of transplant tolerance in postpartum recipients

Abstract

Pregnancy can result in the development of antibodies to fetal human leukocyte antigens, and this humoral sensitization contributes to reduced access to transplantation. Although desensitization protocols have enabled more sensitized patients to undergo transplantation, their efficacy is variable due in part to the persistence of donor-specific memory B cells. Better constraint of memory B cells through the induction of a B cell–intrinsic tolerant state may result in improved humoral desensitization. This study uses mouse models of semiallogeneic pregnancy to show that, in the absence of fetus-specific antibodies, pregnancy-sensitized B cells can acquire a tolerant state that manifests as the inability to differentiate into germinal center B cells. Adoptive transfer of serum containing fetus-specific antibodies prevented tolerance induction and reversed established B cell tolerance. Strikingly, B cells sensitized through skin rejection did not acquire this tolerant state even when donor-specific antibodies were absent. Furthermore, antepartum CTLA-4Ig treatment prevented humoral sensitization, induced B cell tolerance, and preserved the ability of postpartum dams to accept F1 grafts with transient anti-CD154 treatment. Together, these findings reveal the distinct plasticity of pregnancy-induced memory B cells that may be leveraged toward new strategies to improve access to transplantation and outcomes in postpartum women.

1. Introduction

Highly sensitized patients with pretransplant panel-reactive antibodies >98% experience longer wait times, higher pretransplant mortality, and increased risk of antibody-mediated rejection.^1–4 Approximately 50% of postpartum (PP) women develop anti-HLA IgG, and the risk of humoral sensitization to paternal human leukocyte antigens (HLAs) increases with each subsequent pregnancy.^5–11 As a result, PP transplant candidates are overrepresented as highly sensitized and have reduced access to organ transplantation.^6,12 Furthermore, long-term follow-up studies indicate an increased likelihood of graft loss in parous women.^3,13 Although desensitization protocols have enabled more sensitized patients to undergo transplantation, their efficacy remains variable due in part to the persistence of pregnancy-induced memory B cells.^14–16 Memory B cells can rapidly differentiate into antibody-secreting cells with reduced requirement for T cell help, thus making their control more challenging with calcineurin inhibitor–based immunosuppression that primarily targets T cells.^17–19 In a subset (~40%) of PP recipients, fetal HLA-specific IgG gradually becomes undetectable over time, but memory B cells can persist to mount anamnestic donor-specific antibody (DSA) responses posttransplant.^20–23 Importantly, donor-specific memory B cells are not routinely quantified in the clinic due to the absence of tractable assays capable of quantifying these low frequency cells in the blood with sensitivity and specificity.²⁴

In prior work, we reported that semiallogeneic pregnancy induces T cell tolerance and expansion of fetus-specific regulatory T cell (Tregs), while simultaneously sensitizing B cells to produce fetus-specific antibodies (FSAs).²⁵ This humoral sensitization prevented anti-CD154–induced acceptance of transplanted fetus-matched (F1) hearts whereas pregnancy-sensitized sIgm^−/−aicda^−/− (sIgKO) mice that cannot secrete antibodies accepted F1 heart grafts.²⁵ Those observations raise the possibility that preventing pregnancy-induced allosensitization as well as tolerizing PP memory B cells would enhance the armamentarium for preventing and reversing humoral desensitization in the clinic. One population that may benefit from preventing sensitization during pregnancy would be women with preexisting chronic kidney disease (CKD), because pregnancy can trigger CKD progression leading to increased likelihood of requiring renal transplantation after pregnancy.^26,27

Our investigation shows that pregnancy-sensitized sIgKO B cells can acquire a cell-intrinsic tolerant state that manifests as the inability to differentiate into germinal center (GC) B cells. However, circulating FSA prevents and reverses this state of tolerance. In contrast, sIgKO B cells sensitized through skin rejection do not acquire this tolerant state even when the FSAs are absent. Finally, we demonstrate that pregnancy-induced humoral sensitization can be mitigated with transient CTLA-4Ig treatment administered during and immediately after pregnancy, leading to a B cell–intrinsic tolerant state and preserved susceptibility of PP mice to anti-CD154 + donor spleen cells (TolRx)-induced long-term acceptance of F1 grafts.

2. Results

2.1. Anti-CD154 induces acceptance of fetus-matched heart grafts in pregnancy but not in skin-sensitized sIgKO mice

We previously reported that semiallogeneic pregnancy elicits humoral sensitization despite the induction of fetus-specific T cell tolerance.²⁵ Using the same experimental models, we confirmed that PP C57Bl/6 females mated with 2W-OVA.BALB/c males and then treated with TolRx, rejected fetus-matched 2W-OVA.F1 (BALB/c × C57Bl/6) heart grafts between postoperative day (POD) 28 to 45 (Fig. 1A–C). In contrast, virgin mice treated with TolRx uniformly accepted 2W-OVA.F1 heart grafts for >45 days (Fig. 1A–C). FSAs generated during pregnancy, and not memory B cells, mediated resistance to anti--CD154–mediated allograft acceptance, as sIgKO mice lacking secreted immunoglobulins uniformly accepted 2W-OVA.F1 hearts under TolRx (Fig. 1D–F). Rejected allografts recovered on ~POD 45 to 60 from these PP recipients showed CD4⁺ and CD8⁺ T cell infiltration that was absent in PP sIgKO recipients. These observations support the conclusion that FSAs generated during pregnancy prevented TolRx-induced T cell tolerance (Fig. 1G). Sensitization by pregnancy was distinct from sensitization by skin graft rejection, as skin graft–sensitized sIgKO mice were resistant to TolRx and acutely rejected 2W-OVA.F1 heart grafts by POD 14; these results also confirm the ability of sIgKO mice to develop memory B cell responses (Fig. 1D–F). Thus, pregnancy sensitization poises the maternal immune response distinctly from F1 skin graft rejection, with pregnancy-sensitized B cells preserving their susceptibility to TolRx-induced acceptance of 2W-OVA.F1 hearts in the absence of circulating FSAs/DSAs.

Figure 1.

Anti-CD154 induces the acceptance of 2W-OVA.F1 heart allografts in pregnancy- but not skin-sensitized sIgm^−/−aicda^−/− (sIgKO) recipients. (A) Experimental model. Virgin C57Bl/6 received fetus-matched 2W-OVA.F1 (F1) heart transplant with (blue) or without (red) -anti-CD154 + donor spleen cells [TolRx]treatment. 2W-OVA.BALB/c males were mated with C57Bl/6 females. Postpartum (PP) C57Bl/6 received an F1 heart transplant and was treated with aTolRx to induce graft acceptance. (B, C) Percent survival and graft palpation scores of F1 hearts transplanted into virgin, PP recipients treated with TolRx, or virgin recipients without TolRx; n = 4–7/group. Log-rank (Mantel-Cox) test for statistical significance. (D) Female sIgKO mice were mated with 2W-OVA.BALB/c males, or received 2W-OVA.F1 skin transplant (Sk), and both groups received F1 heart grafts + TolRx. Virgin sIgKO mice received an F1 heart transplant without TolRx. (E, F) Percent survival and graft palpation scores of F1 heart grafts, with log-rank (Mantel-Cox) test for statistical significance. (G) Histology of F1 grafts from PP C57Bl/6 and PP sIgKO recipients on day ~50 posttransplant. Graft scores were determined based on abnormalities, decellularization, and infiltration for H&E stain and IHC for infiltrating CD4⁺ and CD8⁺ T cells. Each symbol represents 1 mouse. Data are presented as mean ± SEM, and statistical significance was assessed by the Mann–Whitney test. *P < .05; **P < .01; ***P < .001. AR, acute rejection of F1 hearts by virgin mice.

2.2. Pregnancy-sensitized but not skin-sensitized sIgKO B cells are susceptible to anti-CD154–induced heart allograft acceptance

Consistent with previous findings,²⁵ the anti-F1 IgG response in semiallogeneic pregnancy (2W-OVA.BALB/c males mated with C57BL/6 females) was reduced compared to fetus-matched F1 skin graft rejection, whereas the frequency of fetus-specific (I-E^d) B cells was not significantly different, when analyzed at day 21 to day 30 PP or skin transplantation (Supplementary Fig. S1A, B). We next tested whether B cells sensitized by pregnancy were intrinsically different from B cells sensitized by skin rejection. To minimize the differential effects of skin sensitization vs pregnancy on donor-specific T cells and anti-F1 IgG, we enriched for B cells from day ~30 PP or skin transplanted sIgKO mice and adoptively transferred 35 × 10⁶ B cells/mouse into MD4 hosts that then received F1 heart transplantation (HTx) and TolRx (Fig. 2A). MD4 mice were selected as hosts because they have normal B cell numbers; however ~95% of their B cells express a B cell receptor specific for hen egg lysozyme and consequently they have a substantially reduced anti-BALB/c B cell repertoire and mount a minimal anti-BALB/c responses.^28,29 Additionally, 5 × 10⁶ T cells from naïve C57Bl/6 mice were cotransferred with B cells hosts to ensure sufficient donor-specific T cell help and minimize the possibility that the restricted B cell repertoire may, in turn, limit the endogenous T cell repertoire in MD4 mice.^30–32

Figure 2.

Pregnancy-sensitized but not skin-sensitized sIgm^−/−aicda^−/− (sIgKO) B cells are susceptible to anti-CD154–induced F1 heart allograft acceptance. (A) Experimental design. Female sIgKO mice were sensitized by mating with 2W-OVA.BALB/c males or 2W-OVA.F1 skin transplant (Sk). After 30 days, pregnancy (postpartum [PP]) or Sk sensitized sIgKO B cells (35 × 10⁶/mouse) were adoptively transferred (AdTr) with 5 × 10⁶ naïve C57Bl/6 T cells into MD4 hosts. These mice then received 2W-OVA.F1 (F1) heart transplantation (HTx) and anti-CD154 + donor spleen cells (TolRx). (B, C) Percentage of graft survival and palpation scores. Log-rank (Mantel-Cox) test for statistical significance. (D) Histology of allografts on day ≥45 posttransplant from MD4 hosts that received PP or Sk sIgKO B cells. Graft scores were determined based on abnormalities, decellularization, and infiltration for H&E stain and IHC for infiltrating CD4⁺, CD8⁺ T cells and B220⁺ B cells. (E) Total donor (I-E^d)-specific B cells recovered. (F) Representative histogram and mean fluorescence intensity (MFI) of donor endothelial MHC-II expression. Each symbol represents 1 mouse. Data are presented as mean ± SEM, and statistical significance was assessed by the Mann–Whitney test. *P < .05; **P < .01; ***P < .001. EC, Endothelial cell.

MD4 mice that received sIgKO B cells sensitized by pregnancy, as well as F1 HTx + TolRx, exhibited significantly prolonged heart allograft survival, with superior palpation scores, compared to mice receiving skin-sensitized sIgKO B cells (Fig. 2B, C). Histology and immunohistochemistry performed on allografts recovered on day ~50 posttransplant confirmed that allografts in MD4 recipients of PP sIgKO B cells had minimal B or T cell infiltrates while recipients of skin-sensitized sIgKO B cells had infiltration characteristic of rejection (Fig. 2D). The frequency of donor MHC II–specific B cells recovered from lymph nodes and identified based on binding to I-E^d tetramers was ~2-fold lower in recipients of pregnancy-sensitized compared to skin-sensitized sIgKO B cells (Fig. 2E). Finally, MHC-II expression on the donor graft endothelial cells was significantly lower, consistent with reduced inflammation (Fig. 2F). Together, these data support the hypothesis that skin-sensitized sIgKO B cells play an antigen-presenting role in resisting CD154 blockade and driving T cell–mediated rejection. In contrast, pregnancy-sensitized sIgKO B cells retained susceptibility to TolRx and permitted the induction of long-term acceptance of F1 HTx. These observations raise the possibility that PP sIgKO B cells, like naïve B cells, could develop B cell–intrinsic tolerance.²⁹

2.3. sIgKO B cells from PP recipients with accepted heart allografts exhibit reduced ability to differentiate into GC B cells

We previously reported that TolRx-induced acceptance of allogeneic hearts in naïve mice drove donor-specific B cells into a cell-intrinsic tolerant state marked by reduced differentiation into GC B cells and antibody-secreting cells.²⁹ Because TolRx was able to induce F1 heart acceptance in pregnancy-sensitized sIgKO dams, we hypothesized that PP memory B cells post HTx + TolRx would also acquire a similar tolerant state. To test this, 2 × 10⁷ B cells from PP sIgKO dams with accepted F1 hearts were adoptively transferred into MD4 hosts, followed by sub-cutaneous immunization (IMZ) with F1 donor spleen cells (DSCs) (Fig. 3A). Controls were MD4 hosts with B cells from PP sIgKO mice without F1 heart transplants. DSAs were not quantifiable due to sIgKO B cells inability to secrete immunoglobulins. However, on day 14 post immunization, the total number of B cells binding to donor I-Ed tetramers recovered from PP and PP + HTx + TolRx groups were comparable (Fig. 3B, C). Furthermore, donor-specific B cells with a GC phenotype (Fas⁺GL7⁺) in lymph nodes were quantified by flow cytometry (Fig. 3D); B cells from PP sIgKO dams that received HTx + TolRx demonstrated a significantly reduced ability to differentiate into GC B cells upon DSC immunization compared to PP sIgKO B cells (Fig. 3D).

Figure 3.

sIgm^−/−aicda^−/− (sIgKO) B cells from tolerant postpartum (PP) recipients exhibit reduced ability to differentiate into germinal center (GC) B cells. (A) Experimental design. Female sIgKO were mated with 2W-OVA.BALB/c male, and 30 days later, received 2W-OVA.F1 (F1) heart transplantation (HTx) and anti-CD154 + donor spleen cells (TolRx). B cells were enriched from PP sIgKO + F1 HTx + TolRx or PP sIgKO and adoptively transferred (AdTr) into MD4 hosts (20 × 10⁶/mouse) with 5 × 10⁶ naïve C57Bl/6 T cells, followed by subcutaneous donor spleen cell immunization (IMZ). (B) MHC-II–specific B cells were quantified from lymph nodes on day 13 of IMZ. Representative flow plot to detect MHC-II (I-E^d) specific B cells using I-E^d-PE tetramer together with decoy K^b (recipient MHC) tetramer conjugated to PE and AF-647. (C) Total I-E^d specific sIgKO B cells from PP or PP + F1HTx + TolRx mice. (D) Representative flow plots to quantify I-E^d specific GC B cells (Fas⁺GL7⁺), and total GC B cells recovered from MD4 hosts on day 14 post-IMZ. Each symbol represents 1 mouse. (E) Experimental design for B cell adoptive transfer of PP B6 + F1 HTx + TolRx or PP B6 into MD4 hosts as in Figure 2A. (F) Representative flow plots to quantify C57BL/6 B cells that are I-E^d specific and have a GC phenotype (Fas⁺GL7⁺), and total GC B cells recovered from MD4 hosts on day 14 post-IMZ. Data are presented as mean ± SEM, and statistical significance was assessed by the Mann–Whitney test. ***P < .001. SAC, Sacrifice.

We also investigated the fate of fetus-specific B cells from PP C57Bl/6 (Fig. 3E). The total donor I-E^d tetramer–binding B cells recovered from PP and PP + HTx + TolRx C57Bl/6 dams was comparable (Fig. 3C). However, in contrast to sIgKO B cells, wild-type B cells from PP and PP + HTx + TolRx C57Bl/6 showed comparable differentiation into Fas⁺GL7⁺ GC B cells (Fig. 3F). Together, these findings demonstrate that in the absence of circulating FSAs/DSAs, PP B cells can acquire a cell-intrinsic inability to differentiate into GC B cells when they reencounter F1 antigens in the setting of HTx + TolRx. This reduced ability to differentiate into GC B cells persisted when transferred with T cells into MD4 hosts that were immunized with F1 DSCs. The ability of PP sIgKO B cells to acquire a persistent state of B cell–intrinsic tolerance is reminiscent of that observed with naïve B cells,²⁹ but distinct from skin rejection–sensitized sIgKO B cells.

2.4. FSAs prevent acquisition and maintenance of PP sIgKO B cell tolerance

B cells from sIgKO mice may functionally differ from wild-type B cells beyond the absence of secreted immunoglobulins; potential differences can include distinct activation thresholds and the absence of antibody-mediated feedback regulation and AID-driven GC dynamics. To address these concerns, we tested whether PP FSAs (250 μL/mouse) adoptively transferred at the time of HTx + TolRx could prevent the acquisition of B cell tolerance (Fig. 4A). PP serum had significantly increased total (Fig. 4B) and fetus-specific (Fig. 4C) IgM, IgG, and IgA. Transfer of PP serum into sIgKO HTx recipients prevented HTx acceptance induced by TolRx, whereas recipients that did not receive PP serum accepted transplanted hearts (Fig. 4D). To test whether PP serum prevented the induction of B cell–intrinsic hypofunction, B cells isolated from HTx recipients were transferred into MD4 recipients that were then immunized with F1 splenocytes (Fig. 4A). B cells from HTx recipients with PP serum differentiated into GC B cells in MD4 hosts immunized with F1 splenocytes, contrasting with B cells from HTx recipients without PP serum that did not differentiate into GC B cells (Fig. 4E, G).

Figure 4.

Fetal-specific antibodies (FSAs) prevent the acquisition and maintenance of B cell–intrinsic hypofunction in postpartum (PP) sIgm^−/−aicda^−/− (sIgKO) B cells. (A) Experimental design. Female sIgKO were mated with 2W-OVA.BALB/c males and received 2W-OVA.F1 (F1) heart transplantation (HTx) at PP day ≥30. Recipients were treated with anti-CD154 + donor spleen cells (TolRx) to induce graft acceptance. Sera from C57BL/6 dams that had experienced semiallogeneic or syngeneic pregnancy were recovered on PP day 0 to 14 (FSA). For some groups, serum was administered as a single dose (250 μL, intravenous) on the day of HTx. Enriched B cells from PP sIgKO HTx recipients were adoptively transferred (AdTr) into MD4 hosts, which were then challenged with donor spleen cell immunization (IMZ; subcutaneous). Some MD4 hosts (bottom 2 rows) received FSAs and donor spleen cells. MD4 hosts were euthanized on day 14 post-IMZ, and MHC-II–specific B cells from lymph nodes were quantified. (B, C) Total (B) and FSAs (C) in virgin and PP serum. (D) Percent graft survival in PP sIgKO recipients with or without FSAs at heart transplant. (E, F) Representative flow plots to quantify I-E^d specific germinal center (Fas⁺GL7⁺) B cells for the indicated experimental groups. (G) Total I-E^d specific germinal center B cells recovered at day 14 post-IMZ. Each symbol represents 1 mouse. Data are presented as mean ± SEM, and statistical significance was assessed by the Mann–Whitney test. *P < 0.05; **P < .01; ***P < .001. Ab, antibody; MFI, mean fluorescence intensity. SAC, Sacrifice.

We next asked whether PP serum, when adoptively transferred into MD4 hosts, could reverse established B cell tolerance (Fig. 4A). B cells from tolerant PP sIgKO HTx recipients differentiated into GC B cells in MD4 hosts that received PP serum at the time of F1 immunization (Fig. 4F, G). In these MD4 hosts, the GC B cell differentiation was comparable to PP sIgKO B cells that were nontolerized by HTx + TolRx (Fig. 3D). Controls were MD4 hosts that received PP serum from syngeneic pregnancy or did not receive any PP serum, wherein adoptively transferred PP sIgKO B cells did not differentiate into GC B cells (Fig. 4E, G). Together, these findings reveal the potent ability of FSAs to not only prevent but also reverse TolRx-induced tolerance in PP B cells.

2.5. Antepartum CTLA4-Ig prevents humoral sensitization and preserves susceptibility to anti-CD154–induced allograft acceptance

We previously reported that antepartum treatment with CTLA-4Ig, which blocks CD28:CD80/86 interactions necessary for activating T cells to differentiate into cells providing help to B cells, was able to prevent FSA production.²⁵ Here, we tested whether antepartum CTLA-4Ig could induce a cell-intrinsic tolerant state in fetus-specific B cells and preserve susceptibility to anti-CD154–mediated F1 heart acceptance in PP dams. CTLA4-Ig administered during pregnancy (embryonic days E7, E11, E15) and immediately postpartum (PP3, PP5) (Fig. 5A) did not impact litter sizes, which were comparable to untreated controls (Fig. 5B). However, CTLA-4Ig significantly reduced fetus-specific (anti-K^d, anti-I-E^d) IgG responses, as quantified using a multiplex bead assay (Fig. 5C–E).

Figure 5.

Antepartum CTLA4-Ig prevents humoral sensitization and preserves susceptibility to anti-CD154/donor spleen cell (DSC)-induced allograft acceptance. (A) Experimental design. Female C57Bl/6 mated with male 2W-OVA.BALB/c and treated with CTLA4-Ig during gestation and the immediate postpartum (PP). (B) Litter sizes in PP mice ± CTLA-4Ig. (C) Representative flow plots of a multiplex bead assay to measure donor MHC-I/II–specific IgG. (D, E) Quantification of fetal MHC-I–specific and MHC-II–specific IgG of PP mice ± CTLA4-Ig. (F) Experimental design. Female C57Bl/6 mated with male 2W-OVA. BALB/c and treated with antepartum CTLA-4Ig. B cells were isolated and adoptively transferred (AdTr) into MD4 hosts that were then immunized with 2W-OVA.F1 (F1) DSC. (G, H) On day 14 postimmunization, MHC-I and MHC-II antibodies were quantified. (I) Pregnant C57Bl/6 dams treated with antepartum CTLA-4Ig and 30 days after delivery, received 2W-OVA.F1 (F1) heart transplantation (HTx) + TolRx. (J, K) Graft survival and palpation scores. Log-rank (Mantel-Cox) test for statistical significance. Each symbol represents 1 mouse. Data are presented as mean ± SEM, and statistical significance was assessed by the Mann–Whitney test. *P < .05; **P < .01. DSA, donor-specific antibody; E, embryonic day; TolRx, anti-CD154 + donor spleen cells.

To track fetus-specific T follicular helper (Tfh) cell responses and assess the effects of antepartum CTLA-4Ig, TCR75 CD4⁺ T cells recognizing fetus K^d-derived peptide presented on maternal I-A^b were adoptively transferred (5 × 10⁴/mouse) into pregnant dams on embryonic day E3 (Supplementary Fig. S2A).³³ TCR75 cells expressed the congenic marker CD45.1, allowing identification in CD45.2 C57Bl/6 recipients (Supplementary Fig. S2B, C). UMAP analysis of PP TCR75 cells showed activation during pregnancy, with increased expression of Bcl6, PD-1, Icos, and cMaf (Supplementary Fig. S2D, E), indicating modest Tfh-like cell differentiation. Notably, antepartum CTLA-4Ig inhibited both TCR75 activation and differentiation into Tfh-like cells (Supplementary Fig. S2). Similar modest Tfh-like differentiation was similarly observed in endogenous polyclonal fetus-specific CD4⁺ T cells identified using 2W:I-A^b tetramers from PP C57Bl/6 mated with 2W-OVA.-BALB/c males (Supplementary Fig. S3).

We next tested whether antepartum CTLA-4Ig treatment not only prevented humoral sensitization but also induced fetus-specific B cells to acquire a cell-intrinsic tolerant state. B cells from CTLA-4Ig–treated dams were isolated on PP day 30 and transferred into naïve MD4 hosts that were then immunized with F1 DSCs (Fig. 5F). B cells from CTLA-4Ig–treated dams produced significantly less donor MHC class I–specific and MHC class II–specific IgG than B cells from untreated dams (Fig. 5G, H), consistent with antepartum CTLA-4Ig–induced B cell tolerance.

Finally, we tested whether CTLA-4Ig–treated dams resembled naïve mice in their ability to accept F1 hearts when treated with anti-CD154/DST. F1 hearts were transplanted on PP day 30 (Fig. 5I), when circulating CTLA-4Ig is estimated to be reduced by ~95% from day 14 (T_1/2 ~90 hours³⁴). We observed that 67% (4 of 6) of dams treated with antepartum CTLA-4Ig accepted their grafts long-term (>100 days) with strong palpation scores. In contrast, only 14% (1 of 7) of recipients without antepartum CTLA-4Ig treatment accepted their grafts when treated with TolRx (Fig. 5J, K). These results demonstrate that antepartum CTLA-4Ig prevents both humoral sensitization and preserved susceptibility to anti-CD154–mediated F1 allograft acceptance in PP dams.

3. Discussion

In this study, we demonstrate that pregnancy-sensitized sIgKO mice can be tolerized to F1 grafts with TolRx, but only in the absence of circulating FSAs, as PP C57Bl/6 were resistant to TolRx. Importantly, skin-sensitized sIgKO mice were resistant to anti-CD154, revealing the ability of pregnancy to imprint a unique immune-experienced state that is permissive to anti-CD154–induced allograft acceptance. We further demonstrated that PP sIgKO recipients of F1 hearts + anti-CD154 acquire a B cell–intrinsic tolerant state, evidenced by their failure to differentiate into GC B cells after adoptive transfer into secondary MD4 hosts and F1 DSC challenge. This tolerant state was observed only with fetus-specific B cells from PP sIgKO but not from PP C57Bl/6. Furthermore, the adoptive transfer of FSAs at the time of F1 HTx prevented the acquisition of B cell–intrinsic tolerance, whereas the adoptive transfer of FSAs into MD4 hosts reversed the established tolerant state. Collectively, these findings underscore the potent antitolerogenic properties of FSAs and the importance of humoral desensitization to reduce pretransplant DSAs as well as prevent posttransplant DSA production by memory B cell recall responses.

The mechanisms by which FSAs prevent or reverse B cell tolerance were not elucidated in the current study. Rejection in the presence of FSAs was characterized by increased MHC-II expression by graft endothelial cells, as well as T and B cell infiltration, in the allograft, consistent with FSAs overriding the ability of anti-CD154 to induce donor-specific T and B cell tolerance. We had previously reported that pretransplant memory B cells or DSAs prevent anti-CD154–induced graft acceptance through the ability of DSAs to bind to donor antigens, facilitate antigen presentation, and enhance CD40-CD154–independent activation of alloreactive T cells.^35,36 We therefore hypothesize that FSAs will similarly bind to DSCs to generate opsonins/immune complexes; these complexes then bind to the BCR and activate complement receptors, CD21 (CR2) and CD35 (CR1), to enhance B cell differentiation.^37–39 Immune complexes can also bind to CR1/CR2 expressed on follicular dendritic cells that play critical roles in the formation of the proper microenvironment for GC responses.^40,41 Future experiments that allow the conditional elimination of CD21/CD35 only in PP B cells will be necessary to test this hypothesis, but to our knowledge, such mice are currently not available.

Several mechanisms constraining T cell responses to the semiallogeneic fetus and ensuring successful pregnancy have been identified at the maternal-fetal interface and in the SLO; these include the systemic induction of anergic fetus-specific CD4⁺ T cells and expansion of fetus-specific Tregs.^42–51 More recently, Rizzuto et al⁵² and Rizzuto⁵³ described glycan-mediated immune regulation in which trophoblast-derived sialylated glycoproteins engage inhibitory Siglec receptors, particularly CD22, that signal through Lyn kinase to suppress antigen-specific B cell responses. This CD22-Lyn axis constrains the accumulation of fetus-specific and GC B cells, which in turn modulate fetus-specific CD4⁺ T cell responses. Nevertheless, pregnancy induces fetus-specific IgG responses in both humans and mice,^54,55 which we previously reported in a mouse model of semiallogeneic pregnancy to occur without evidence of GC B cell responses.²⁵ This is in contrast to allograft rejection, where strong GC B cell responses are observed.^17,18,29,56 We speculate that the susceptibility of pregnancy-sensitized but not skin rejection–sensitized sIgKO B cells to acquire a tolerant state upon exposure to F1 hearts with TolRx may be explained by the difference in how these sIgKO memory B cells were generated. Consistent with this speculation are the recent observations that GC B cells undergo extensive epigenetic reprogramming to emerge as long-lived plasma cells and memory B cells, whereas extrafollicular B cells experience a transient activation and acquire an epigenetic landscape that favors the generation of short-lived plasma cells and memory B cells.^57–59 It is also possible that the magnitude of the B cell response in pregnancy vs rejection contributed to the observed differential susceptibility to TolRx.

A substantial subset of primiparous and multiparous women who develop FSAs no longer have detectable antibodies at 3 years PP.^20,22 The discordance between declining anti-HLA antibodies and persistent memory B cells highlights the need to quantify donor-specific memory B cells in the clinic,^16,60–63 and to develop therapies that restrain their differentiation into antibody-secreting cells posttransplant.¹⁸ In support of this concept, recent nonhuman primate studies demonstrated that, proteasome inhibition combined with maintenance belatacept in pregnancy-sensitized monkeys significantly improved kidney transplant survival, suppressed posttransplant DSAs, and lowered frequencies of circulating Tfh-like and memory B cells.⁶⁴ Notably, although belatacept is superior to calcineurin inhibition at constraining DSA responses,^65–67 only ≥10% of kidney Tx recipients currently receive belatacept due to concerns of higher incidence of acute rejection and increased risk of posttransplant lymphoproliferative disorder and infection.^68–70 Our study supports the hypothesis that patients at the highest risk for calcineurin inhibition–resistant recall antibody responses, would most benefit from belatacept-based immunosuppression, thus allowing for a more individualized approach to posttransplant immunosuppression in multiparous women.

Pregnancy alloimmunization is one of the most common modes of sensitization in women, and there is a major unmet need to mitigate pregnancy-induced humoral sensitization. We here demonstrate that antenatal CTLA-4Ig treatment prevents humoral sensitization during pregnancy by inducing a tolerant state in fetus-specific B cells, thereby preserving susceptibility to TolRx-induced allograft acceptance in PP dams. These proof-of-principle studies suggest a potential strategy for preventing humoral sensitization during pregnancy in women with late-stage CKD, especially because pregnancy can accelerate renal failure and increase transplant need.^26,27 Importantly, abatacept use during pregnancy has not shown teratogenic effects to the developing fetus, and spontaneous abortion and live birth complication rates were within the normal range.^71,72 Whether anti-CD154 would be superior to CTLA-4Ig, due to preservation of Tregs, is an important question, especially if anti-CD154 is successful in clinical trials and becomes a therapeutic option.

In summary, our current study shows that pregnancy induces distinct states of humoral sensitization from F1 skin rejection, wherein PP recipients preserved their susceptibility to anti-CD154–induced acceptance of F1 hearts when FSAs are absent, and B cell–intrinsic tolerance can be induced. We further demonstrate that antepartum CTLA-4Ig treatment prevents pregnancy-induced humoral sensitization, leading to the induction of B cell tolerance and preserved susceptibility of PP C57Bl/6 mice to anti-CD154–induced acceptance of F1 grafts. Limitations of the study include the need for additional in-depth mechanistic studies on the potent effects of FSAs on B cell tolerance and to define the transcriptional and epigenetic differences in PP vs skin-sensitized B cells that permit the acquisition of this state of tolerance. Taken together, this study illustrates the unique immunobiology of pregnancy sensitized B cells and a strategy for how they may be controlled to facilitate access to transplantation of postpartum women with end-stage renal failure and promote long-term survival of the allografts.

4. Material and methods

4.1. Mice

Eight- to 12-week-old female C57Bl/6 (H-2^b) and HEL-specific BCR-transgenic (Tg) MD4 mice on a C57Bl/6 background were purchased from The Jackson or Harlan Laboratories. Geetha Chalasani (University of Pittsburgh, Pittsburgh, PA) provided the sIgm^−/−aicda^−/− mice (H-2^b, sIgKO), and TCR75 TCR-Tg mice were from R. Pat Bucy (University of Alabama, Birmingham, Alabama, USA). Act-2W-OVA Tg mice (2W-OVA.C57Bl/6) from James Moon (Massachusetts General Hospital, Charlestown, MA) were bred for >50 generations with BALB/c mice to generate 2W-OVA.-BALB/c. PP mice were generated by mating male 2W-OVA.-BALB/c mice with virgin C57Bl/6 or sIgKO females, and the detection of the copulation plug was designated embryonic day 0.5. Antepartum CTLA4-Ig (250 μg/mouse) was administered intraperitoneal (i.p.) on embryonic days, E5, E11, and E15, and PP days, PP3 and PP6. All animal experiments were approved by the Institutional Animal Care and Use Committee at the University of Chicago and adhered to the standards of the NIH Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011).

4.2. Heart and skin transplantation and TolRx treatment

Heterotopic heart transplants were performed as previously described, by grafting 2W-OVA.F1 (BALB/c × C57Bl/6) donor hearts into the peritoneal cavity of recipient mice.⁷³ Flank skin from 2W-OVA.F1 was transplanted onto the flank of recipient mice.³⁵ For B cell adoptive transfer, MD4 mice received 3.5 × 10⁷ B cells/mouse on day −1 of HTx or subcutaneous immunization with F1 splenocytes. Graft acceptance was induced with anti-CD154 (MR1, BioXCell) at a dose of 500 μg/mouse on POD 0 (intravenous), and 250 μg/mouse on POD 7 and 14 (i.p.), in combination with 2 × 10⁷ DSCs (i.p.) on POD 0. Allograft survival was monitored twice weekly using transabdominal heart palpation and scored on a scale of 0 to 4. Allograft rejection was defined as the day when the heartbeat was no longer palpable.

4.3. Tetramers for antigen-specific T and B cells

B cells and CD4 T cells were enriched by negative selection using a Pan-B Cell Isolation Kit (catalog 130–095-813; Miltenyi Biotec) and a CD4 T Cells Isolation Kit (catalog 130–095-130; Miltenyi Biotec). Following enrichment, cells were stained with Fixable Aqua Live/Dead stain (Thermo Fisher Scientific or Invitrogen) and blocked with anti-CD16/32 (clone 2.4G2). To identify donor MHC-II-reactive B cells, pCons-CDR1 (FIEWNKLRFRQGLEW):I-E^d tetramers (PE-conjugated) and H-2K^b (SIINFEKL) decoy tetramers (AF647-PE–conjugated) were used⁷⁴ (NIH Tetramer Core Facility). Lymphocytes were first incubated with supersaturating concentrations of decoy tetramer (10 nM) at 4 °C for 10 minutes, followed by saturating concentrations of PE-conjugated I-E^d tetramer (5 nM/10⁷) and incubation on ice for 30 minutes. Adoptively transferred congenic TCR75-Tg cells were enriched with APC-Cy7–conjugated anti-CD45.1 (clone A20). 2W:I-A^b specific CD4⁺ T cells staining was performed with PE and APC-conjugated 2W (EAWGALANWAVDSA):I-A^b tetramers⁷³ (NIH Tetramer Core Facility).

4.4. Statistical analysis

Statistical significance analyses were conducted using GraphPad Prism. Kaplan–Meier/Mantel–Cox log-rank tests were used to assess graft-survival significance. Statistical differences between experimental groups were determined using ANOVA or the Mann–Whitney unpaired t test. P values ≤.05 were considered statistically significant.

Abbreviations:

CKD

chronic kidney disease

DSA

donor-specific antibody

DSC

donor spleen cell

FSA

fetus-specific antibody

GC

germinal center

HLA

human leukocyte antigen

HTx

heart transplantation

i.p.

intraperitoneal

POD

postoperative day

PP

postpartum

sIgKO

sIgm ^−/− aicda ^−/−

Tfh

T follicular helper

Tg

transgenic

TolRx

anti-CD154 + donor spleen cell

Treg

regulatory T cell

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ajt.2026.06.008.

Data availability

The data that support the findings of this study are available on request from the corresponding author.