The role of secondary lymphoid organs in adaptive immune responses following transplantation is controversial. To examine the requirement for peripheral lymphoid organs in mounting immune responses to transplantation antigens, lymphotoxin α-deficient (LTα–/–) and LTβ-receptor-deficient (LTβR–/–) mice that lack lymph nodes and Peyer's patches were used as recipients of fully allogeneic heart and skin grafts. Splenectomized LTα–/– and LTβR–/– mice effectively rejected skin and cardiac allografts, although with delayed kinetics when compared with wild-type controls. In addition, initial skin allograft challenge in splenectomized LTβR–/– mice resulted in accelerated rejection of subsequent donor cardiac allografts when compared with heart rejection in nonsensitized controls. Thus, although peripheral lymphoid organs play an important role in allowing allograft responses to occur, they do not appear to be absolutely required for either acute allograft rejection, or T-cell priming. These results suggest that immunologic events capable of leading to allograft rejection can successfully occur at sites other than classical secondary lymphoid organs.
The initial stimulation of T cells by antigen presented in the context of self-antigen-presenting cells (APCs) is thought to occur in secondary lymphoid organs. Recognition of alloantigen in transplantation settings may be more complex, in that antigen can be presented both by donor cells, including graft stromal cells and passenger APCs (direct presentation) and, following processing, by host-antigen-presenting cells (indirect presentation). Direct presentation has been shown to play an important role in cardiac transplantation settings for stimulation of both CD4+ (1) and CD8+ T cells (2,3), and is sufficient to promote acute allograft rejection. However, whether direct presentation occurs at the graft site or following migration of donor cells to host secondary lymphoid organs has not been fully elucidated.
Studies conducted using aly/aly mice suggested that acute allograft rejection depends on the presence of secondary lymphoid organs (4). Mutant aly/aly mice lack lymph nodes as a result of a mutation in NF-κB-inducing kinase (NIK), an enzyme upstream of NF-κB activation (5). Unmanipulated aly/aly mice were unable to reject skin allografts, whereas splenectomized aly/aly mice failed to reject cardiac allografts (6), suggesting that lymph nodes were required for acute rejection of skin, whereas either lymph nodes or spleen were sufficient for rejection of allogeneic hearts. NIK is activated downstream of LTβR in stromal cells, an event necessary for secondary lymphoid structure formation. However, NIK activation has also been described downstream of TCR/CD28 engagement and may play a role in T-cell function (7) and T-cell migration (8), such that aly/aly mice might have impaired T-cell responses. In addition, NIK has also been implicated in B-cell (9,10) and macrophage function (11), which might affect indirect presentation of alloantigen in aly/aly mice.
Other mouse models displaying altered lymphoid architecture have failed to reveal such an absolute requirement for secondary lymphoid organs during T-cell responses to antigen. Thus, paucity of lymph node T cells (plt) mice that lack CCL21 and CCL19 and display consequent defects in T-cell migration to lymph node T-cell zones, were shown to have delayed but ultimately enhanced T-cell responses to antigen (12). In fact, when plt mice were splenectomized and immunized with antigen, T cells from draining lymph nodes displayed similar proliferation to antigen as those from splenectomized wild-type mice, suggesting that T-cell activation can occur outside of conventional T-cell zones in lymph nodes and spleen. LTα–/– mice have also been used as another mouse model devoid of lymph nodes. In an airway inflammation model, these mice generate even stronger responses than wild-type mice. However, airway inflammation is diminished following splenectomy (13). Other studies have shown than LTα–/– mice can clear a productive viral infection but fail to develop splenomegaly or lymphocytosis (14). However, one study shows clear defects in the generation of antiviral CTL responses and IFN-γ-producing cells (15), and another report indicates that LTα–/– mice cannot mount contact sensitivity responses (16). We have previously shown that LTα–/– mice effectively rejected allogeneic small bowel transplants (17) and skin allografts (18), suggesting competent T-cell responses in these animals, and indicating that the dependence on intact lymph nodes and/or spleen for immune res-ponses may vary with the stimulus and type of response.
The formation of secondary lymphoid organs is a complex process that requires lymphotoxin (LT) signaling. The predominant form of membrane LT (LTα1β2) binds to LTβR expressed on stromal cells, whereas its minor form, LTα2β1, and soluble LT (α3) bind to tumor necrosis factor receptor (TNFR). LTα-deficient mice are devoid of morphologically detectable lymph nodes and Peyer's patches (19). In addition, the white pulp of the spleen is abnormal, lacking the usual segregation between T and B cells. In rare mice, some mesenteric lymph nodes are found (20). In contrast, LTβ–/– mice lack most lymph nodes and Peyer's patches but retain cervical and mesenteric lymph nodes, suggesting that signaling mediated by soluble LTα3 may be responsible for the formation of specific lymph nodes (21). Finally, LTβR–/– mice have been generated, that have a similar phenotype to LTα–/– mice in terms of lymph node formation, as they lack all lymph nodes and Peyer's patches (22).
We have used LTα–/– and LTβR–/– mice to evaluate the role of secondary lymphoid organs in acute rejection of allogeneic skin and heart grafts. We found that allograft rejection and T-cell priming occurred in splenectomized LTα–/– and LTβR–/– mice, albeit in an impaired fashion when compared with intact animals. Therefore, we conclude that secondary lymphoid organs play an important role but are not absolutely required for immune responses to occur in transplantation settings.
Materials and Methods
Eight-to-twelve-week-old BALB/c (H-2d), C57BL/6 (B6, H-2b), C3H/HEN (H-2 k), and LTα–/– mice were purchased from The Jackson Laboratories (Bar Harbor, Maine, USA). LTβR–/– mice were obtained from Klaus Pfeffer (University of Düsseldorf, Düsseldorf, Germany). LTα–/– and LTβR–/– were back-crossed onto the B6 background for 10 and 6 generations, respectively. Animals were housed in individually ventilated cages in a specific pathogen-free animal facility. Experiments were performed in agreement with our Institutional Animal Care and Use Committee and according to the NIH guidelines for animal use.
Heart and skin transplantation
Abdominal heterotopic cardiac transplantation was performed using a technique adapted from that originally described by Corry et al. (23). Cardiac allografts were transplanted in the abdominal cavity by anastomosing the aorta and pulmonary artery of the graft end-to-side to the recipient's aorta and vena cava, respectively. The day of rejection was defined as the last day of a detectable heartbeat in the graft. Graft rejection was verified in selected cases by necropsy and pathological examination of hematoxylin/eosin-stained graft sections.
Skin transplantation was performed as previously described (24). Briefly, full-thickness donor tail-skin pieces (0.5 cm2) were positioned on a flank graft bed. Bandages were removed on the seventh postoperative day. The time-point of rejection was defined as the complete necrosis of the graft.
Splenocytes or peripheral blood cells from B6 and LTβR–/– mice left untreated, or transplanted with B6 or BALB/c hearts were stained with FITC-coupled anti-CD4, anti-CD8, or anti-Thy1.2 mAbs and PE-coupled isotype control, CD44, or CD62L mAbs (PharMingen, San Diego, CA, USA). Cells were analyzed by flow cytometry.
Mixed lymphocyte reactions
B6 and LTα-deficient mice were left untreated, or were transplanted with BALB/c heart allografts. All animals were sacrificed a few days after rejection of heart transplants by B6 mice. Splenocytes used as responders were prepared into single cell suspensions, depleted of red blood cells in ammonium/potassium/chloride (ACK) lysis buffer, and plated at 2 × 105 cells/well in round-bottom, 96-well plates. BALB/c and C3H/HEN stimulator splenocytes were depleted of erythrocytes in ACK lysis buffer and γ-irradiated (2000 rads). Stimulators were plated at 4 × 105 cells/well. Plates were incubated for 24 h at 37 °C in the presence of 7% CO2. Supernatants were assayed for IL-2 content by ELISA using anti-IL-2 antibody pairs (PharMingen), as directed by the manufacturer.
Histopathologic assessment of cardiac grafts
Grafts were removed at different time-points, embedded with OCT (Tissue-Tek Miles Inc, Elkhart, IN, USA), and immediately frozen in liquid nitrogen. The samples were sliced into 6-µm-thick sections at − 20 °C and stained with anti-CD8 rat IgG supernatant (neat) or anti-CD4 purified rat IgG antibody as previously described (25). Slides were evaluated under light microscopy by a pathologist blinded to the clinical rejection status of the heart. The number of CD4+ and CD8+ cells was counted in 3–4 randomly chosen high-powered visual fields per section ( ×400 magnification, approximately 254 mm2).
Heart graft mean survival time (MST), standard error, and p-values were calculated using Kaplan–Meier/log rank test methods. The t-test was used to compare numbers of infiltrating cells.
Effective skin graft rejection by LTβR –/– mice
To examine the role of secondary lymphoid organs in the rejection of unvascularized tissue allografts, fully allogeneic BALB/c skins were transplanted into intact or splenectomized B6 and LTβR–/– mice. Acute rejection of skin grafts occurred with similar kinetics in intact wild-type and LTβR–/– recipients (MST 8 ± 1 days, Figure 1). Surprisingly, splenectomized LTβR–/– mice effectively rejected skin allografts, albeit with slower kinetics than intact mice (MST 17 days, p < 0.05). Rejection was not restricted to this particular strain combination, as C3H/HEN (H-2 k) skin grafts were also effectively rejected in splenectomized LTβR–/– mice (MST 11 ± 2, n = 5, data not shown). These data suggest that secondary lymphoid structures are not absolutely required for acute rejection of skin allografts.
Cardiac allograft rejection occurs in splenectomized LTβR –/– and LTα–/– mice
To address the role of secondary lymphoid organs in rejection of a vascularized allograft, BALB/c hearts were transplanted into intact and splenectomized wild-type and LTβR–/– mice. Acute rejection of cardiac allografts occurred rapidly in intact B6 and LTβR–/– recipients (MST 8, and 11 days, respectively, Figure 2A). The surgical removal of the spleen prolonged survival of cardiac allografts in B6 mice (MST 16 days, p < 0.001) and markedly delayed rejection in LTβR–/– recipients (p < 0.001). However, between 60 and 87 days post-transplant, 5 out of 6 splenectomized LTβR–/– mice eventually rejected their grafts.
All LTβR–/– transplanted animals were carefully examined at necropsy and we were unable to detect any lymphoid structures. However, as an alternative model, experiments were repeated using LTα–/– mice as recipients. Similarly to what was observed in LTβR–/– mice, rejection of cardiac allografts effectively occurred in both intact (MST 26 days) and splenectomized (MST 50 days) LTα–/– recipients (Figure 2B). Taken together, these results suggest that secondary lymphoid tissues are not absolutely required for rejection of either skin or heart allografts.
Increased numbers of CD4 + cells in cardiac allografts rejected by splenectomized LTβR–/– mice
To compare the cell types infiltrating rejected cardiac allografts, histology and immunohistochemistry were performed on heart transplants harvested from splenectomized B6 and LTβR–/– recipients at different time-points. Histological features in all mice were consistent with cellular-mediated acute rejection (scores 2, data not shown). Greater numbers of CD8+ than CD4+ cells were observed in grafts from intact (data not shown) and splenectomized (Figure 3A) B6 recipients, confirming our previous findings in this strain combination (25). Surprisingly, a significant increase in the number of CD4+ cells was observed in grafts from splenectomized LTβR–/– when compared with grafts from splenectomized B6 mice at the same time-point (day 10, p < 0.01). Numbers of CD8+ cells were not significantly different. Numbers of infiltrating CD4+ and CD8+ cells both gradually increased over time in splenectomized LTβR–/– mice (Figure 3A,B). To determine if the increase in infiltrating CD4+ cells was a feature of LTβR–/– mice or was secondary to the removal of the spleen, the number of intragraft CD4+ cells was compared at the time of rejection in intact and splenectomized LTβR–/– mice. Interestingly, significantly increased numbers of CD4+ cells were detected within allografts from splenectomized LTβR–/– mice when compared with intact LTβR–/– (p < 0.001), suggesting that acute allograft rejection in the absence of secondary lymphoid organs may require increased influx of CD4+ cells.
Effective acute allograft rejection in LTβR –/– and LTα–/– mice is probably not due to greater numbers of memory T cells than in wild-type mice
It has been reported that previously activated T cells or memory cells do not require trafficking to secondary lymphoid organs for mediating acute allograft rejection (26). Thus, it is conceivable that effective allograft rejection by splenectomized LTβR–/– mice in our model is due to the presence of memory T cells. Therefore, we examined the expression of CD44 and CD62L on splenic CD4+ and CD8+ T cells from wild-type and LTβR–/– mice. As shown in Figure 4(A), similar percentages of splenic T cells from wild-type and LTβR–/– displayed the CD44high and CD62Llow phenotypes. Transplantation of syngeneic and especially of allogeneic grafts resulted in an increased proportion of T cells expressing activation markers when compared with untreated mice. No difference between the percentage of T cells expressing activation markers (Figure 4A) or the expression levels (mean fluorescence intensity, data not shown) of surface markers was detected between T cells from wild-type and LTβR–/– mice.
To exclude the possibility of an accumulation over time of an activated population of T cells in splenectomized animals, we examined the expression of activation markers on peripheral blood T cells from splenectomized wild-type and LTα–/– mice. Figure 4(B) indicates similar percentages of activated T cells over the course of the experiment in both strains of mice. Together, these results suggest that effective allograft rejection observed in splenectomized LT-deficient mice is unlikely to be due to a higher percentage of allospecific memory T cells.
Finally, we compared the degree of alloreactivity in wild-type and LTα-deficient mice, as a readout of the number of alloantigen-specific T cells in both strains of mice. To this end, splenocytes from B6, LTα-deficient, and B6 mice that had rejected BALB/c hearts were stimulated with irradiated BALB/c or C3H/HEN splenocytes. As shown in Figure 4(C), similar levels of IL-2 production were found in the supernatants of B6 and LTα-deficient splenocytes in response to both stimulator strains, whereas significantly more IL-2 was found in the cultures of splenocytes from previously transplanted B6 mice restimulated with donor but not third-party cells. This result indicates that the degree of alloreactivity in LTα-deficient mice is comparable to unmanipulated rather than to sensitized animals. Taken together, these data suggest that LTα-deficient mice do not have more activated or antigen-experienced T cells than wild-type mice.
Secondary lymphoid organs are not required for generation of memory responses
We next investigated whether priming to alloantigens could occur in the absence of secondary lymphoid organs. To address this question, LTβR–/– mice were first splenectomized and transplanted with BALB/c skins. Following skin graft rejection, the animals were challenged with donor cardiac allografts. As shown in Figure 5, splenectomized LTβR–/– mice previously sensitized with skin allografts rejected cardiac allografts significantly faster (MST 41 ± 35 days) than nonsensitized splenectomized LTβR–/– mice (MST 84 ± 17 days, p < 0.01), suggesting that T-cell priming by skin alloantigens did occur in the absence of secondary lymphoid structures. As expected, when skin grafts were first placed onto intact LTβR–/– recipients that were later splenectomized at the time of cardiac allograft challenge, heart rejection occurred even faster (MST 11 ± 3 days, p < 0.001), indicating that the presence of the spleen allowed for more effective T-cell priming. Together, these results indicate that the presence of a spleen facilitates, but is not required for, generation or expansion of memory responses to alloantigens.
Using splenectomized LTα–/– and LTβR–/– mice that are devoid of lymph nodes and Peyer's patches, we have evaluated the requirement for secondary lymphoid organs for mounting immune responses to alloantigens in vivo. Our results in this model suggest that lymph nodes and spleen are important but not absolutely necessary for either rejection of skin and heart allografts, or generation of memory responses.
The implications of these results are in conflict with those of a recent report using aly/aly mice that indicated the presence of lymph nodes was required for the effective rejection of skin allografts and the combined deficit in lymph nodes and spleen prevented cardiac allograft rejection (6). Both models utilize mice genetically devoid of lymph nodes and Peyer's patches in which the spleen was surgically removed. Using chimeric mice obtained by fusing morulae from wild-type and aly/aly or LTα–/– mice, it was reported that LTα and the aly gene product control lymphoid organogenesis in a similar manner, LT acting as a membrane-bound and/or soluble ligand and the aly gene product acting as a LTβR-signaling molecule expressed by the stroma of lymphoid organs (27). However, T cell, B cell, and macrophage function may be impaired in aly/aly mice (7,9), affecting responses to alloantigens. In addition, it was also reported that aly/aly T cells have reduced signaling responses to the chemokine secondary lymphoid-tissue chemokine (SLC) and decreased consequent migratory properties (8), which might also affect migration into transplanted organs. Alternatively, it is possible that the formation of tertiary lymphoid structures (i.e. antigen-driven de novo formation of lymph-node-like structures) occurs more readily in LT-deficient than in aly/aly mice. This is not likely, however, as LT signaling is thought to be necessary for tertiary lymph node formation. Indeed, the formation of these structures in the pancreas of diabetic NOD mice is abrogated when these mice are treated with LTβR-Ig (28,29). Finally, although our data do not support an increase in the global activation status of T cells from LTα–/– or LTβR–/–, we cannot rule out that pre-existing allospecific memory cells are responsible for rejection in the absence of secondary lymphoid organs in this model. Thus, disparities between these two very similar models of mice devoid of spleen and lymph nodes must caution result interpretations, as other features than lack of secondary lymphoid organs could be altered in both models. Comparisons between the different models may allow further dissection and understanding of immune responses in vivo.
Immunohistochemical analysis of cardiac allografts at the time of rejection revealed increased numbers of CD4+ cells infiltrating the grafts of splenectomized LTβR–/– mice when compared to those from intact LTβR–/– and from wild-type animals. This increase is compatible with CD4+ T cells being stimulated at the graft site. Indeed, direct presentation by MHC class II-expressing donor cells has been shown to be essential for CD4+ T-cell-dependent rejection of cardiac allografts (1,2). In splenectomized LTβR–/– mice, more CD4+ T cells may therefore be activated at the graft site, as conventional secondary lymphoid structures are not available. Recently, it has been reported that presentation of alloantigen to CD8+ T cells can be achieved by nonhematopoietic cells from the graft, perhaps directly by donor endothelial cells (3). As mouse endothelial cells can express MHC class II (30), it is possible that this pathway also exists for CD4+ T cells. Alternatively, an increased number of CD4+ cells at the graft site may be the result of impaired CD8+ T-cell function in this model and greater consequent requirement for CD4+ T cells. Indeed, the complex lymphotoxin pathway is known to be important for CD8+ T-cell responses, either for providing costimulation (31,32), or for facilitating T-cell migration (17).
Cardiac allografts in splenectomized LTβR–/– mice all survived for at least 60 days. Following this lag time, most grafts were rejected within a short interval. This lag time is not due to the acute reduction in lymphocyte numbers that results from splenectomizing these animals, as the same delay in rejection was observed when the splenocytes were reinfused intravenously at the time of cardiac transplantation (data not shown). The delay may reflect the fact that LT signaling promotes allogeneic responses. Blockade of LT signaling using LTβR-Ig treatment resulted in a 3-day prolongation of cardiac allograft survival, whether transplanted hearts were obtained from wild-type or LTβR–/– donors (p < 0.05, data not shown). This is consistent with a previous report showing similar 3-day prolongation of cardiac allograft survival in mice genetically deficient in LIGHT, a ligand expressed on dendritic cells and activated T cells that is blocked by LTβR-Ig (33).
Surprisingly, intact LTα–/– mice were found to reject cardiac allografts more slowly than intact LTβR–/–animals. LTα–/– mice lack both soluble and membrane-bound LT signaling, whereas LTβR-deficient mice lack signaling via membrane-bound LT and LIGHT. It is possible that soluble LT normally accelerates rejection, perhaps via binding to TNFR.
Generation of memory responses occurred in splenectomized LTβR–/– mice, as sensitizing these mice with skin grafts resulted in faster rejection of subsequent donor cardiac allografts (41 days) when compared with the rejection time in splenectomized, nonsensitized LTβR–/– mice (84 days). Thus, secondary lymphoid organs are not absolutely required for memory responses to occur, although their presence clearly facilitates priming. CCR7– memory cells that express low levels of the lymph node homing receptor CD62L and may not require trafficking to lymph nodes to acquire effector function have been described (34). In addition, memory cells have been shown to promote cardiac allograft rejection in the absence of lymph nodes and spleen (26). Our data support the notion that the generation of effector memory cells or the expansion of pre-existing memory cells also does not require the presence of lymph nodes and spleen.
Thus, our data using splenectomized LTα–/– and LTβR–/– mice suggest that initiation of immune responses and generation of memory responses can occur in the absence of spleen and lymph nodes, albeit much less efficiently than when secondary lymphoid structures are present.
We would like to thank Drs Anita Chong and Anne Sperling for helpful discussions and J. Richard Thistlethwaite for critical reading of this manuscript.This study was supported by the Roche Organ Transplantation Research Foundation Award #193412071 and by an American Society of Transplantation Faculty Grant.