Abstract
Acute allograft rejection requires the activation of alloreactive CD4 T cells. Despite the capacity of B cells to act as potent APCs capable of activating CD4 T cells in vivo, their role in the progression of acute allograft rejection was unclear. To determine the contribution of B cell APC function in alloimmunity, we engineered mice with a targeted deficiency of MHC class II-mediated Ag presentation confined to the B cell compartment. Cardiac allograft survival was markedly prolonged in these mice as compared to control counterparts (median survival time, >70 vs 9.5 days). Mechanistically, deficient B cell-mediated Ag presentation disrupted both alloantibody production and the progression of CD4 T cell activation following heart transplantation. These findings demonstrate that indirect alloantigen presentation by recipients’ B cells plays an important role in the efficient progression of acute vascularized allograft rejection.
Activation of alloreactive CD4 T cells occurs via two distinct pathways of alloantigen presentation: 1) the direct pathway, driven by graft-derived APCs; and 2) the indirect pathway, mediated by recipient APCs (1, 2, 3, 4). Host dendritic cells (DCs)4 and macrophages have been the dominant focus of studies aimed at elucidating mechanisms of indirect alloantigen presentation (4, 5). Notwithstanding, B cells also constitutively express MHC class II and costimulatory molecules, which enable them to act as potent APCs for alloreactive CD4 T cell activation (6, 7, 8). Activated alloreactive B cells, by virtue of their Ag-specific BCR, have the capacity to selectively concentrate specific Ags making them a repository of alloantigen-presenting APCs. Despite this unique ability to sequester alloantigens and subsequently activate alloreactive CD4 T cells via MHC class II, the role that B cells play in indirect presentation of alloantigens to alloreactive CD4 T cells in acute graft rejection remains undefined. Furthermore, the fact that alloreactive B cells become activated and participate in alloimmune responses is clearly demonstrated by the production of alloantigen-specific Abs by the recipient’s B cell compartment (9, 10, 11). To date, the question of whether induction of B cell tolerance is required for transplantation tolerance remains an open one (12). Indeed, the participation of B cells in allograft rejection has been assumed to be limited to hyperacute or chronic allograft rejection driven by alloantibodies (13). The fact that B lymphocyte-deficient mice (μMT0/0) maintain their capacity to mediate allograft rejection has been taken as definitive evidence that neither B cells nor their alloantibody products make a significant contribution to acute allograft rejection (14, 15). Nonetheless, the immune system of μMT0/0 mice, which lacks a B cell compartment during its ontogeny, may impact the differentiation of T cell responses—as suggested by the potentiated capacity of DCs from μMT0/0 mice to deviate Th responses toward a dominant Th1 phenotype (16).
Thus, to assess the relative importance of B cell-mediated Ag presentation in vascularized cardiac allograft rejection, we sought to disrupt B cell-mediated Ag presentation via MHC class II in recipient mice. To this end, we used two complementary bone marrow (BM) chimera strategies to engineer recipient mice whose B cells exhibited a selective deficiency in MHC class II Ag presentation. One set of chimeras was engineered such that their B cells expressed no MHC class II. A second set of chimeras expressed MHC class II on their B cells, but lacked the ability to load exogenous peptides onto the MHC class II molecule. In this study, the fate of allografts in the absence of B cell-mediated MHC class II Ag presentation in the above-mentioned chimeras was determined. Moreover, we assessed the impact of impaired B cell-mediated MHC class II Ag presentation on two key in vivo effector processes following transplantation: 1) alloreactive CD4 T cells activation and 2) alloantibody production.
Materials and Methods
Mice
C57BL/6, C57BL/6 Scid, C57BL/6 I-Aβ0/0, C57BL/6 DM0/0, C57BL/6 μMT0/0, BALB/c, and C3H mice were purchased from The Jackson Laboratory. TS1 Tg BALB/c mice are transgenic for an αβ TCR clonotype specific for the major I-Ed-restricted peptide determinant of influenza virus PR8 hemagglutinin (HA) (17). HACII transgenic heart donors were also described previously (18). All mice were housed in a specific pathogen-free barrier facility at the Children’s Hospital of Philadelphia in accordance with protocols by the Institutional Animal Care and Use Committee.
Generation of chimeric mice
Chimeric BCIID, intermediate BCIID (BCIIDint), BDM, or control mice were generated by i.v. inoculating a total of 1 × 107 T and B cell-depleted BM cells into lethally irradiated (1000 rad) C57BL/6 wild-type or 100-rad-irradiated C57BL/6 Scid recipients. In multiple experiments, no difference was observed between reconstituted C57BL/6 wild-type or Scid recipients, and therefore they were used interchangeably throughout these experiments. All BM preparations, regardless of origin, were depleted of T and B cells using anti-CD45R (B220) and anti-CD90 (Thy1.2) microbeads in concert with the VarioMACS system (Miltenyi Biotec). The BCIID and BDM BM inoculates were comprised of 5 × 106 μMT0/0 and 5 × 106 I-Aβ0/0 or DM0/0 T and B cell-depleted BM cells per mouse, respectively. The BCIIDint BM inoculum was a mixture of 5 × 106 μMT0/0, 2.5 × 106 I-Aβ0/0, and 2.5 × 106 wild-type cells per mouse. The BALB/c BCIID recipients were reconstituted identically, albeit using BM derived from μMT0/0 congenic BALB/c mice and I-Aβ0/0 mice. Control chimeric mice were reconstituted with 5 × 106 μMT0/0 and 5 × 106 wild-type BM cells per mouse. Following reconstitution, all mice were housed in a pathogen-free animal facility at the University of Pennsylvania Medical Center for up to 10 wk before analysis or transplantation.
Flow cytometry and immunohistology
A total of 1 × 106 splenocytes or lymph node cells were surface stained in 96-well microtiter plates with various combinations of the following Abs obtained from BD Pharmingen: RA3-6B2-PerCP (anti-B220), AF6-78-FITC (anti-IgMb), RM4-5-allophycocyanin (anti-CD4), 53-6.7-PerCP (anti-CD8), AF6-120.1-FITC (anti-I-Ab), M1/70-PE (anti-CD11b), and HL3-PE (anti-CD11c). For immunohistochemical staining, spleens from representative mice were snap frozen and sectioned. Serial splenic tissue sections were then single or double stained with AF6-78-FITC (anti-IgMb; BD Pharmingen) and AF6-120.1-PE (anti-I-Ab; BD Pharmingen). Slides were then examined by fluorescent microscopy and photographed using a digital camera system.
HA-specific ELISA
Virus-specific ELISAs were done as described previously (19). Briefly, 96-well flat-bottom microtiter Immunlon 1B plates (Dynex Technologies) were coated with 50 μl of purified influenza PR8 and J1 viruses (1000 HAU/ml) overnight. Plates were washed and blocked for 1 h with 100 μl of PBS plus 1%BSA, then incubated with serum samples (diluted 1/100 in PBS plus 1%BSA) for 90 min at room temperature. Plates were washed, and bound Ab was detected using alkaline phosphatase-conjugated goat anti-mouse IgG (Southern Biotechnology Associates). Plates were developed using p-nitrophenylphosphate, and optical densities were read at 405 nm using a microplate reader. PR8 influenza virus contains the same HA as the HACII mice, whereas J1 virus contains a disparate HA. Therefore, the HA-specific component of the Ab response is the difference of OD taken on PR8 less that measured on J1.
CFSE labeling and in vivo CD4 T cell activation
Purified lymph node cells from C57BL/6 mice or TS1 Tg BALB/c mice were B cell depleted using anti-CD45R microbeads and VarioMACS (Miltenyi Biotec). Following depletion, 0.5–1 × 107 lymph node T cells were CFSE labeled and i.v. injected into BCIID or control mice. Labeling with CFSE (Molecular Probes) was conducted as described previously (20). After an overnight rest period, C57BL/6 strain mice were injected with 15 μg of staphylococcal enterotoxin A (SEA) i.v. Forty-eight hours later, mice were euthanized and spleen and lymph nodes cells were purified and analyzed by flow cytometry as described above. BALB/c BCIID mice were transplanted with cardiac allografts from HA-CII transgenic donors as described below and previously (18). Four days following transplantation, these mice were euthanized and spleen and lymph nodes cells were purified and analyzed by flow cytometry as described above and previously (18).
Skin and cardiac transplantation
All surgeries were performed under general anesthesia as prescribed by the Institutional Animal Care and Use Committee guidelines. Donor hearts were harvested for heterotopic cardiac transplantation, and an end-to-site anastomosis of donor and recipient aortas was performed. A venous anastomosis was also performed in an end-to-site fashion between the donor pulmonary artery and the inferior vena cava of the recipient. All anastomoses were done using 7-0 prolene sutures. The fate of all allografts was monitored by daily manual abdominal palpation and, following euthanasia, by gross and histological examination. For skin grafting a round, full-thickness graft was harvested and stripped of the panniculus carnosus. The grafts were then transplanted onto recipient mice and their fate was assessed every 2–3 days following transplantation.
Results
Strategy for creating mice with a MHC class II deficiency restricted to the B cell compartment
To generate mice with a B cell-specific abrogation of MHC class II-mediated Ag presentation, we used BM chimeras. Immunodeficient mice were reconstituted with a mixture of BM cells from B cell-deficient (μMT0/0) and MHC class II presentation-deficient (I-Aβ0/0 or DM0/0) mice. I-Aβ0/0 mice are homozygous null for the I-Aβ-chain of MHC class II. Thus, on the C57BL/6 background, which does not express I-E, I-Aβ0/0 mice do not express any MHC class II on the surface of their APCs. DM0/0 mice are deficient in the H2-M molecule responsible for loading peptides onto MHC class II, resulting in a crippled capacity of their APCs to present endogenous (other than CLIP) or exogenously derived peptides (21). The APC compartments of these mixed chimeric mice should, in theory, be derived partially from the μMT0/0 BM, and thus give rise to MHC class II-sufficient non-B cell APCs (i.e., DC and macrophages) and partially from the I-Aβ0/0 or DM0/0 BM, giving rise to B cells (in addition to DCs and macrophages) that are deficient in MHC class II Ag presentation (Table I⇓). These mice are referred to as BCIID and BDM mice, respectively. A third category of mice reconstituted with μMT0/0 BM cells and a mixture of I-Aβ0/0 and I-Aβ+/+ BM cells, are termed BCIIDint mice. BCIIDint mice were constructed to generate mice whose B cell compartment contains both MHC class II-deficient and MHC class II-sufficient B cells. Control mice were generated using mixed BM cells from μMT0/0 and I-Aβ+/+, thus reconstituting with a MHC class II-sufficient B cell compartment.
Origin and MHC class II status of B cell and non-B cell APCs in chimeric BCIID, BCIIDint, and BDM mice
Phenotypic analysis of the peripheral lymphoid compartment in BCIID mice
Chimeric mice were analyzed 8 wk postreconstitution for phenotypic and functional characterization of the immune systems. Absolute numbers of B220+/IgM+ B cells were similar in BCIID mice and control chimeric mice (48.2 × 106 ± 11.3 × 106 in BCIID mice and 43.3 × 106 ± 8.7 × 106 in control chimeric mice). As expected, splenic B220+/IgM+ B cells from control chimeras, but not BCIID chimeras, expressed MHC class II (Fig. 1⇓). In BCIIDint mice, ∼50% of the splenic B cells were MHC class II-deficient and 50% MHC class II-sufficient (data not shown). The MHC class II expression of the non-B cell APC compartment was examined by assessing macrophage and DCs from BCIID and control mice. As shown in Fig. 2⇓, e and f, splenic CD11c+ DCs and CD11b+ macrophages were present in similar proportions/absolute numbers and express MHC class II in BCIID and control mice. Interestingly, we found that MHC class II-sufficient non-B cell APCs preferentially populated the periphery of the BCIID mice, presumably due to a survival advantage over their MHC class II-deficient counterparts. Therefore, the non-B cell APC compartment of BCIID mice was primarily made of MHC class II-sufficient cells, despite the expectation that it would contain a proportion of MHC class II-deficient cells derived from I-Aβ0/0 progenitors. As a second measure of the integrity of reconstitution of the chimeric mice, histological analyses of the spleens were performed (Fig. 2⇓, a–d). B cell follicles from BCIID and control chimeric mice were similar in their distribution of B220 staining (Figs. 2⇓, a and c); however, MHC class II expression in the BCIID mice was predominantly localized to the T cell zone of the follicle (Fig. 2⇓b), whereas in control mice, B220+ cells in the B cell zone of the follicle also stained positive for MHC class II (Fig. 2⇓d). Splenic and lymph node CD4 and CD8 T cells were comparable, both numerically and proportionally in BCIID, BDM, and control chimeric mice (Fig. 1⇓b). In addition, the thymus of BCIID mice contained a normal proportion of CD4 single-positive (SP), CD8 SP, and CD4/CD8 double-positive thymocytes, indicating that T cell development proceeds normally in these mice (Fig. 1⇓b). The existence of a normal sized CD4 T cell compartment indicated that the amount of MHC class II expressed in these chimeric mice is sufficient to permit the development and survival of a phenotypically and numerically normal CD4 T cell compartment in BCIID mice (22, 23).
The T and B cell compartments in BCIID mice. a, Representative flow cytometric analysis of splenic B cells of BCIID (bottom) and control (top) chimeric mice. MHC class II expression by BCIID or control (red line) chimeric vs wild-type (blue line) IgM+/B220+ splenocytes is illustrated. Absolute numbers and proportions of splenic B cells are comparable in BCIID and control chimeric mice (48.2 × 106 ± 11.3 × 106 in BCIID vs 43.3 × 106 ± 8.7 × 106 in control mice; n = 4). b, Representative flow cytometric analysis of CD4 and CD8 expression in thymus and lymph nodes of BCIID and control chimeric mice. The percentages of CD4 and CD8 lymph node T cells are indicated in the upper left and lower right quadrants, respectively, and are comparable in BCIID and control chimeras. Pooled peripheral lymph node cells were isolated from the cervical, axillary, and inguinal regions. The absolute numbers of lymph node T cells were 14.8 × 106 ± 6 × 106 vs 12.6 × 106 ± 3 × 106 for CD4 T cells and 9.3 × 106 ± 2.1 × 106 vs 10 × 106 ± 3 × 106 for CD8 T cells in BCIID vs control mice, respectively. The distribution of CD4 and CD8 SP and double-positive thymocytes was comparable in BCIID and control mice. c, Frequency of CD4+/CD25+ Treg cells in the lymph nodes of BCIID and control chimeric mice (range, 3–6% in peripheral lymph nodes). There is no notable expansion of CD4 T cells with the Treg phenotype in BCIID mice.
Macrophage and DC compartments are intact in BCIID chimeric mice. a and c, IgM (green) single staining of spleen from a representative BCIID (a) and control (c) chimeric mouse. b and d, IgM (green) and MHC class II (red) double staining of spleen from a representative BCIID (b) and control (d) chimeric mouse. The B cell follicle in the BCIID spleen is devoid of MHC class II staining, in contrast to the control spleen. e and f, MHC class II+ DCs (CD11c+) and macrophages (CD11b+) are present at a comparable frequency in both BCIID and control chimeric mice. Note that the population of cells expressing MHC class II but not CD11b or CD11c in the control chimeric mice is absent from the BCIID chimeras. This population represents B cells, which, in the BCIID chimeras, do not express MHC class II; CD4/CD8 T cells were gated out. The indicated percentages are of the circled populations as a function of total splenocytes and are representative of four mice analyzed. g, In vivo division of CFSE-labeled CD4 T cells in the spleen of BCIID and control chimeric mice following 48 h of stimulation with SEA. The dotted line is drawn after division 3 to provide a point of reference for comparison of the two histograms.
CD4 T cell activation in BCIID mice
Before assessing the kinetics of allograft rejection in BCIID mice, it was important to determine whether the MHC class II expressing non-B cell APCs of BCIID mice were sufficient to support CD4 T cell activation in vivo. A prior study in a chimeric system similar to the BCIID chimeras demonstrated intact CD4 T cell priming in response to protein immunogens in such mice (24). BCIID and control mice were injected with SEA, which primarily stimulates Vβ3+ and Vβ11+ T cells through binding to MHC class II and the constant region of the TCR β-chain (25, 26). In response to SEA administration, a marker population of CFSE-labeled CD4 T cells underwent nearly identical numbers of cell divisions in BCIID and control mice (Fig. 2⇑g). This result indicated that MHC class II-mediated T cell activation occurred in BCIID mice, despite the fact that the B cell compartment is deficient in MHC class II expression and was functional confirmation of the integrity of the MHC class II-sufficient non-B cell APC compartment in BCIID mice.
Allograft rejection in BCIID mice
To characterize the participation of B cell MHC class II-mediated Ag presentation in the pathogenesis of allograft rejection, we used BCIID and BDM mice as transplant recipients. C57BL/6 (H-2b) BCIID and BDM mice were used as recipients of skin or vascularized abdominal cardiac allografts from allogeneic BALB/c (H-2d) and C3H (H-2k) strain donors. Allogeneic skin grafts were promptly rejected by BCIID and BDM recipients with a kinetics similar to that seen in control C57BL/6 recipients (median rejection times of 11 days for control, 11 days for BCIID, and 9 days for BDM recipients of BALB/c skin grafts (see Table II⇓). In contrast, vascularized heterotopic cardiac allografts displayed markedly prolonged survival in BCIID and BDM recipients relative to control recipients (median rejection times of 9.5 days and 7 days for rejection of BALB/c (H-2d) and C3H (H-2k), respectively, vs >70 days (p < 0.001) and >30 days (p < 0.05) in BCIID recipients and >197.5 days (p < 0.005) and >66 days (p < 0.001) for BDM recipients (Table III⇓). As exemplified in Fig. 3⇓, histological examination of cardiac allografts transplanted into BCIID (Fig. 3⇓a) and BDM (data not shown) mice revealed patchy areas of myocyte destruction with perivascular and parenchymal infiltration by mononuclear cells, as well as arterial intimal thickening; processes reminiscent of a chronic rejection process. This result is in contrast to cardiac allografts transplanted into control recipients, which demonstrated widespread and almost complete myocyte destruction by day 15 following transplantation, consistent with an acute rejection process (Fig. 3⇓b). We next sought to determine whether the presence of both MHC class II-sufficient and -deficient B cells in recipient mice led to efficient acute cardiac allograft rejection. Indeed, BCIIDint mice rejected vascularized cardiac allografts with the same kinetics seen in control counterparts (Table III⇓). Therefore, it was concluded that MHC class II-deficient B cells do not exert a dominant inhibitory effect upon acute vascular allograft rejection in the BCIID model.
Cardiac allograft histology in BCIID and control chimeric mice. Representative H&E staining of BALB/c cardiac allografts in B6 BCIID (a) and control (b) chimeric mice at 55 and 20 days posttransplantation, respectively. Cardiac allograft myocytes in BCIID recipients are markedly spared from destruction, despite frequent foci of perivascular mononuclear cell infiltration (arrow), as compared to counterparts in control chimeras.
Skin allograft survival
Cardiac allograft survival
IgG alloantibody production in BCIID mice following transplantation
It is generally accepted that alloantibody production by B cells requires CD4 T cell help via MHC class II-mediated Ag presentation (27). Therefore, the absence of cognate T-B interaction in BCIID recipients could abrogate the production of alloantibodies, which may be an important effector pathway in the pathogenesis of acute vascularized allograft rejection. Thus, to directly assess alloantibody production in BCIID mice, BCIID and control C57BL/6 mice (H-2b) were transplanted with cardiac allografts from BALB/c mice carrying influenza HA under control of the MHC class II promoter (termed BALB/c HA-CII; Ref. 18). These recipients were then serially bled and examined for the presence of Abs specific for the surrogate alloantigen, HA. Tracking the Ab response to HA as a surrogate alloantigen afforded the ability to quantify a highly specific alloantibody response to a single allograft-derived Ag (i.e., HA) without the high background and temporal/quantitative mouse-to-mouse variability inherent to the polyclonal nature of a total alloantibody response (data not shown). Serum was examined every 3 days following transplantation for ∼4 wk. In control B6 mice that rejected the HACII transgenic allografts acutely (median survival time 10 days; n = 5), alloantibody production was detectable by 2 wk following heart transplantation and peaked at day 18 (Fig. 4⇓a). In B6 BCIID mice, however, no detectable levels of IgG alloantibody were present in the serum up to 4 wk following transplantation (Fig. 4⇓a). In addition, when BCIID mice were immunized with an influenza virus carrying the same HA epitope (i.e., PR8) as the one used as our surrogate alloantigen, an IgG response to HA was not mounted up to 1 mo following inoculation (Fig. 4⇓b). Overall, although MHC class II-deficient B cells are not intrinsically defective in IgG Ab production (28), deficient MHC class II-mediated cognate T/B cell collaboration abrogates the Ab responses to T-dependent Ags, including the IgG alloantibody production following vascularized solid-organ transplantation.
a, Measurement of alloantibody titers in control and BCIID chimeric mice 18 days following cardiac allografts. Mice were bled the day before transplant and then 18 days after transplant. ELISA was done on PR8 virus (P) (which shares the HA molecule present in HACII mice) and J1 virus (J) (an influenza virus with an irrelevant HA as a specificity control). The absolute OD was calculated by subtracting the OD value obtained on each indicated virus with serum at day 0 from that obtained with serum from day 18 posttransplant. Circles indicate individual mice, and bars indicate the mean OD for each designated group. The dotted line indicates threshold for nonspecific binding. Excess of binding on PR8 virus over J1 indicates HA-specific Ab. b, Fold increase in serum anti-HA IgG titer 22 days following immunization of control and BCIID chimeric mice with PR8 virus. The fold increase is calculated from the preimmunization background in the same mice.
Alloreactive CD4 T cell activation in BCIID mice following transplantation
In addition to the impact of a B cell-specific MHC class II deficiency on IgG alloantibody production, we sought to determine whether in vivo alloreactive CD4 T cell activation was also impacted by defective cognate T/B collaboration. Thus, we used a TCR transgenic system we previously developed for the study of CD4 T cell alloimmunity in vivo (18). Briefly, this system uses mice carrying the TS1 transgene, which encodes an αβ TCR specific for the major I-Ed peptide determinant of the influenza virus PR8 HA. Purified TS1 CD4 T cells are CFSE labeled and transferred into recipients, which are then transplanted with allografts from mice expressing HA on MHC class II+ cells (i.e., HA-CII mice). In this fashion, the in vivo CD4 T cell alloimmune response to our surrogate alloantigen can be tracked. We previously characterized this response in detail (18). In this model of CD4 T cell alloimmunity, vascularized heterotopic cardiac allografts elicited a vigorous systemic response, which was detectable 4 days following transplantation (18). In this study, 5 × 106 purified CFSE-labeled TS1 lymph node T cells from Thy1.1 congenic (Thy1.1+) BALB/c donors were adoptively transferred into the following: 1) wild-type control, or 2) BCIID BALB/c mice. The recipients’ indigenous T cell compartment expressed Thy1.2, thereby making it distinguishable from the adoptively transferred cells. These mice subsequently received a HA-CII cardiac allograft. Fig. 5⇓a demonstrates the division profiles of TS1 CD4 T cells in the spleens and pooled lymph nodes of the recipient mice on day 4 following transplantation. Overall, TS1 CD4 T cells clearly underwent fewer divisions and gave rise to a smaller number of divided daughter cells in BCIID recipients, as compared to wild-type controls. As shown in Fig. 5⇓b, using the ratio of divided to undivided CFSE+/CD4+/Thy1.1+ CD4 T cells as a measure of the degree of activation of the alloreactive CD4 T cell population, an ∼2- to 3-fold lower degree of activation is achieved in BCIID mice, as compared to wild-type controls.
a, In vivo division profile of alloreactive CD4 T cells 4 days following heterotopic cardiac transplantation. Representative flow cytometric analysis of the division profile of Thy1.1 congenic, TS1 transgenic, CD4+ T cells from the spleen and pooled lymph nodes of control and BCIID mice 4 days following transplantation with an HACII cardiac allograft (n = 3). The presented dot-plots are gated on Thy1.1+ cells to eliminate the recipients’ endogenous Thy1.2+ T cells from the analysis. To allow for an accurate visual comparison, each dot-plot contains ∼5–10 × 103 Thy1.1+ events. b, Ratio of divided to undivided Thy1.1+, TS1 transgenic, CD4+ T cells in spleen and pooled lymph nodes, as a measure of the activation state of alloreactive CD4 T cells in control (solid bars) and BCIID (striped bars) recipients.
Discussion
It is well established that CD4 T cells are critical regulators of the alloimmune response (29) and that B-cells alone cannot cause acute allograft rejection. Moreover, the fact that B cell-deficient, μMT0/0, mice reject skin and cardiac allografts, has led to a general consensus that B cells or their secreted alloantibody products have no significant role in the progression of acute allograft rejection (14, 15). In agreement with these studies, the present work also demonstrates that μMT0/0 mice reject both skin and cardiac allografts, albeit with a somewhat protracted tempo (Table III⇑). Nevertheless, we have been ambivalent in definitively accepting the conclusions derived from the μMT0/0 system regarding the participation of B cells in acute allograft rejection in wild-type mice. Specifically, the developmental absence of B lymphocytes in μMT0/0 mice causes functional changes in the immune system, which are likely to be confounding. An important study demonstrated that DC development in μMT0/0 mice markedly potentiates the capacity of these non-B cell APCs to polarize CD4 T cell activation toward a Th1 phenotype (16). In fact, several experimental models of transplantation have demonstrated that favoring a Th1 CD4 T cell response potentiates allograft rejection (30). Thus, we chose to re-examine the role of B cell APC function in the pathogenesis of acute allograft rejection by generating the BCIID and BDM chimeric systems in which a selective deficiency in cognate T-B interactions via MHC class II was engineered. In these chimeric recipients cardiac, but not skin, allograft survival was markedly prolonged. This finding indicated that, in the absence of B cell-mediated MHC class II Ag presentation, the normal progression of acute vascularized allograft rejection is disrupted. The fact that skin allografts were rejected with the normal tempo of acute rejection indicated that BCIID and BDM recipients are not generally immunodeficient. Furthermore, the non-B cell APCs in BCIID mice are capable of stimulating CD4 T cells in a comparable fashion to that seen in wild-type controls (Fig. 2⇑g), using a MHC class II-dependent superantigen; this finding was in line with a previous study demonstrating efficient CD4 T cell priming in mice similar to the BCIID chimeras used in this study (24). Thus, the prolonged survival of cardiac allografts in BCIID mice is not a result of a generalized state of immunodeficiency.
The results of this study have led to the conclusion that B cell APC function plays an important role in the efficient progression of acute vascularized cardiac allograft rejection. Mechanistically, the absence of MHC class II-mediated cognate T-B cell collaboration abrogated the production of IgG alloantibodies following transplantation in BCIID recipients, in agreement with previous studies (27). It is possible that by abrogating the allospecific IgG response, cardiac allografts in BCIID and BDM mice enjoy a degree of protection from alloantibody-mediated effector processes leading to their prolonged survival. However, whether this feature alone is responsible for prolongation of cardiac allograft survival in our studies remains unclear. Indeed, we noted a marked overall diminution in the level of alloreactive CD4 T cell activation in BCIID mice following transplantation (Fig. 5⇑). This novel finding provides a likely mechanistic explanation for the diminished vigor of acute rejection in these mice, because it has been clearly established that Th differentiation and effector function are closely linked with the division state of the responding CD4 T cell population (30, 31). Notwithstanding, our study does not prove that the 2- to 3-fold-diminished CD4 T cell division causes the prolonged survival of cardiac allografts in BCIID mice. It is possible that this diminished level of CD4 T cell division in the absence of B cell-mediated MHC class II Ag presentation reduces the probability of alloreactive CD4 T cells fully differentiating into effector Th1 cells (30, 31). In fact, differentiation of Th1 responses are known to promote cardiac allograft rejection (32). Our future studies will be focused on determining whether B cell-mediated Ag presentation drives complete differentiation of an alloreactive Th1 response to vascularized allografts. Finally, we determined that the prolonged survival of cardiac allografts in the BCIID recipients does not result from expansion of CD4+/CD25+ T regulatory (Treg) cells in these mice, because they do not harbor an increased frequency of such cells in their lymph nodes (Fig. 1⇑c) or spleen (data not shown) as compared to control counterparts.
It is important to note that studies by Yamada et al. (33, 34) suggest that acute allograft rejection proceeds normally in the complete absence of indirect alloantigen presentation. In these studies recipient mice with a MHC class II deficiency confined to peripheral APCs, but not thymic epithelial cells were used as recipients of cardiac allografts. These mice are reported to have normal CD4 and CD8 T cell compartments and reject cardiac allografts with normal or accelerated kinetics, leading the authors to the conclusion that indirect pathway presentation is not necessary for the pathogenesis of acute allograft rejection. These results pose a clear contradiction to the results of our studies, which indicate that indirect alloantigen presentation by B cells plays an important role in the efficient progression of acute cardiac allograft rejection. However, a study by Bhandoola et al. (35) sheds some light on this seeming conundrum. This latter study makes two important points regarding CD4 T cell behavior in a MHC class II-deficient environment: 1) CD4 T cells acquire an activated/memory phenotype, and 2) they acquire autoreactivity against syngeneic MHC class II and reject MHC class II+ syngeneic grafts. Bhandoola et al. (35) demonstrated that the very definition of CD4 T cell tolerance changes when these T cells are confronted with MHC class II+ grafts in a MHC class II-deficient environment. Therefore, relying on alloreactive CD4 T cell activation in MHC class II-deficient animals to delineate the role of indirect pathway participation in acute cardiac allograft rejection involves a major confounding factor that should temper definitive conclusions.
Overall, this study demonstrates a necessity for B cell MHC class II-mediated alloantigen presentation for progression of acute vascularized allograft rejection. Mechanistically, we have defined two downstream effector functions important in allograft rejection that are impaired in the absence of B cell-mediated Ag presentation; namely, IgG alloantibody production and CD4 T cell activation. These results establish an important role of B cells in the pathogenesis of acute allograft rejection and support the contention that achievement of transplantation tolerance may require induction of, both, B and T cell tolerance to alloantigens (12, 36, 37, 38). Thus, at a translational level, it is essential to determine whether B cell-specific immunomodulatory agents are effective adjuncts to T cell-specific agents in achieving immunological tolerance to vascularized solid organ allografts.
Acknowledgments
We thank Jessica Dias, Siri A. Greeley, Arjun Jeganathan, Negin Noorchashm, Lauren Noto, Alison Perate, Robert Roses, Alexander Schlachterman, and Daniel J. Trainer for technical and scientific input at various stages of this study.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by National Institutes of Health Grants DK064603 and DK049814.
↵2 H.N. and A.J.R. contributed equally to the design, execution, and preparation of this manuscript.
↵3 Address correspondence and reprint requests to Dr. Hooman Noorchashm, University of Pennsylvania Medical Center, Department of Surgery, 4 Silverstein Pavilion, 3400 Spruce Street, Philadelphia, PA 19104; E-mail address: Hooman.Noorchashm{at}uphs.upenn.edu or Dr. Ali Naji, University of Pennsylvania Medical Center, Department of Surgery, 4 Silverstein Pavilion, 3400 Spruce Street, Philadelphia, PA 19104; E-mail address: Ali.Naji{at}uphs.upenn.edu
↵4 Abbreviations used in this paper: DC, dendritic cell; BM, bone marrow; HA, hemagglutinin; int, intermediate; SEA, staphylococcal enterotoxin A; SP, single positive; Treg, T regulatory.
- Received July 10, 2006.
- Accepted September 5, 2006.
- Copyright © 2006 by The American Association of Immunologists
References
In this issue
