Lymphotoxin β receptor and tertiary lymphoid organs shape acute and chronic allograft rejection

Solid organ transplantation remains the life-saving treatment for end-stage organ failure, but chronic rejection remains a major obstacle to long-term allograft outcomes and has not improved substantially. Tertiary lymphoid organs (TLOs) are ectopic lymphoid structures that form under conditions of chronic inflammation, and evidence from human transplantation suggests that TLOs regularly form in allografts undergoing chronic rejection. In this study, we utilized a mouse renal transplantation model and manipulation of the lymphotoxin αβ/lymphotoxin β receptor (LTαβ/LTβR) pathway, which is essential for TLO formation, to define the role of TLOs in transplantation. We showed that intragraft TLOs are sufficient to activate the alloimmune response and mediate graft rejection in a model where the only lymphoid organs are TLOs in the allograft. When transplanted to recipients with a normal set of secondary lymphoid organs, the presence of graft TLOs or LTα overexpression accelerated rejection. If the LTβR pathway was disrupted in the donor graft, TLO formation was abrogated, and graft survival was prolonged. Intravital microscopy of renal TLOs demonstrated that local T and B cell activation in TLOs is similar to that observed in secondary lymphoid organs. In summary, we demonstrated that immune activation in TLOs contributes to local immune responses, leading to earlier allograft failure. TLOs and the LTαβ/LTβR pathway are therefore prime targets to limit local immune responses and prevent allograft rejection. These findings are applicable to other diseases, such as autoimmune diseases or tumors, where either limiting or boosting local immune responses is beneficial and improves disease outcomes.


Introduction
In solid organ transplantation, immunosuppressive therapy has substantially improved short-term organ allograft survival by reducing acute rejection rates.However, chronic rejection -mediated by T cells, antibodies (Abs), or both -has not markedly declined in incidence and remains an important obstacle to long-term allograft survival (1,2).Further understanding of the pathophysiology of chronic rejection is therefore necessary.
A likely important contributor to the pathogenesis of chronic rejection is the formation of tertiary lymphoid organs (TLO) within the graft.TLO are ectopic lymphoid structures resembling lymph nodes (LN) that arise in chronically inflamed tissues by a process called lymphoid neogenesis (3).Pathognomonic features of TLO include distinct T cell zones, B cell zones, and high endothelial venules (HEV), normally not found outside LNs and Peyer's patches.In the non-transplant setting, TLO have been described in autoimmunity, chronic infection, atherosclerosis, and cancer (4).They correlate with disease severity except in cancer, where they portend better prognosis (5,6).In transplantation, they have been extensively documented in heart, kidney, and lung allografts in both laboratory animals and humans and are associated with chronic rejection and shorter allograft survival (7)(8)(9)(10).For example, 78% of mouse heart allografts undergoing chronic rejection, and up to 95% of human renal allograft explants due to chronic rejection have features of lymphoid neogenesis (7,8).Some reports have also demonstrated roles in tolerance maintenance in mouse models of lung transplantation, where Treg seem to exert their regulatory function in TLO in recipients treated with costimulatory blockade (11).Recent work by Rosales et al. have also described the presence of Treg-rich organized lymphoid structures (TOLS) in a kidney transplantation model in mice using a specific donor -recipient stain combination (12).These structures, contrary to TLO, do not contain HEVs, reflected by lack of PNAd expression.
TOLS have been shown to be important for long-term renal allograft survival, which is dependent on Treg, and can develop in the absence of secondary lymphoid tissue.Although these studies outline specific functions of TLO in different disease models, and associations with specific disease outcomes, cause-effect experiments delineating the contribution of TLO to allograft rejection are sparse.
The lymphotoxin alpha (LTab) -lymphotoxin beta receptor (LTbR) pathway is important for lymphoid neogenesis.The ligands for LTbR are the heterotrimer LTa1b2 and LIGHT, while the homotrimer LTa3 can bind to other members of the TNF receptor superfamily (TNFR1, TNFR2 and HVEM).Signaling through LTbR activates the NFkb as well as the JNK pathway.The alternative pathway of NFkb activation involves activation p100, which is dependent on IKKa and NIK and is the major LTbR pathway responsible for lymph node development.This is evident by the absence of secondary lymphoid tissue in LTbR, IKKa and NIK deficient mice.
Studies utilizing skin transplantation in recipients that do not have secondary lymphoid organs have demonstrated that skin containing TLO can mediate allograft rejection at the same site or of skin transplanted elsewhere (13).
Despite these different roles of TLO in immunity and allograft rejection, several questions remain: Are TLO contributing to allograft rejection?What immune functions do TLO support in vivo?
In this manuscript, we are utilizing renal allograft transplantation model in mice and manipulation of the LTbR -LTab pathway to elucidate the role of TLO in allograft rejection.
Moreover, we developed an intravital microscopy model to visualize immune cell interactions in renal TLO to investigate if TLO support activation of T and B cells.We found that TLO are sufficient for renal allograft rejection, that they contribute to rejection even in the presence of lymph nodes and that disrupting the LTbR pathway prolongs allograft survival.Intravital microscopy showed that TLO support T and B cell activation.

TLO are sufficient for renal allograft rejection
To investigate if renal TLO are sufficient to initiate an alloimmune response and cause graft rejection, we utilized splenectomized LTbR-deficient (LTbR-ko) mice as recipients of F1 (B6 x Balb/c, CB6F1) or F1-RIP-LTa kidneys (Figure 1A).LTbR-ko mice do not have lymph nodes or peyer's patches and after splenectomy are devoid of all secondary lymphoid organs.
Secondary lymphoid organs are necessary to mount an alloimmune response and reject an allograft (14).F1-RIP-LTa mice express lymphotoxin alpha under control of the rat insulin promoter and develop spontaneous TLO in the pancreas, skin, and kidney at 4 to 6 months of age (15).F1-RIP-LTa donor kidneys therefore contain preformed TLO at the time of transplantation, while F1 WT kidneys do not.In this model, the only lymphoid tissue present is the TLO in the donor graft.To rule out that the presence of inflammatory TLO or LTa overexpression in the donor graft has a functional consequence independent of an alloimmune response, we performed syngeneic F1-RIP-LTa kidney transplants to F1 recipients as controls.
As shown in Figure 1B, F1 allografts survive beyond 200 days, while F1-RIP-LTa grafts containing TLOs are rejected with a mean survival time (MST) of 23 days.Syngeneic F1-RIP-LTa grafts were maintained beyond day 90.Donor specific antibody measurements in the serum on day 50 show lack of DSA in recipients of WT allografts while IgG DSA was present in recipients that received F1-RIP-LTa allografts, suggesting that TLO provide a place for B cell activation and antibody formation (Figure 1C).Histopathology (Figure 1D) demonstrates the presence of TLO pre-transplantation as well as at time of rejection in F1 RIPLTa allografts.F1 allografts show less infiltration and lower proportions of severe rejection (Banff >1B) (Figure 1, E and F) on day 200 but are characterized by the presence of lymphoid aggregates around small arteries.Syngeneic F1-RIP-LTa grafts display presence of TLO before and after transplantation, but no other immune infiltrate (Figure 1D) and surpass rejection time of the F1-RIPLTa allografts transplanted to LTbR -/-recipients.Treg-rich organized lymphoid structures (TOLS) have been previously documented, notably even in the absence of secondary lymphoid tissue (12).To further characterize the lymphoid aggregates present in allografts, we performed IF staining for T, B, FoxP3 and PNAd in both F1 and F1-RIP-LTa grafts.As shown in Supplemental Figure 1, lymphoid aggregates in F1 allografts are characterized by T and B cell areas, the presence of FoxP3 + Treg and the absence of PNAd, fulfilling the main criteria for TOLS.Lymphoid aggregates in F1-RIPLTa grafts also contain T and B cell areas, but lack FoxP3 + Treg cells.In addition, PNAd staining is present in these structures, a hallmark of inflammatory tertiary lymphoid organs.The long-term life sustaining function of the syngeneic F1-RIP-LTa kidney grafts suggests that graft failure in allogeneic kidney transplants is a consequence of rejection rather than the mere presence of TLO or LTa overexpression in the graft.These data demonstrate that preformed TLO are sufficient for allograft rejection and support a full alloimmune response with T and B cell activation / DSA production.

TLO accelerate renal allograft rejection
We next investigated if TLO contribute to renal allograft rejection in the presence of a normal set of secondary lymphoid tissue.We transplanted either F1 or F1-RIPLTa kidneys to WT B6 recipients and monitored allograft survival (Figure 2A).As LTa can bind as a heterotrimer LTa1b2 to LTbR and as a homotrimer LTa3 to TNFR family members, mediating inflammatory signals, we also transplanted F1-RIPLTa kidneys from young donors (8 weeks old), in which no TLO have formed at the time of transplantation, but LTa is overexpressed.Allograft survival of F1-RIPLTa kidneys was significantly shorter (MST=63 days) than survival of F1 allografts (MST=225 days), indicating that preformed TLO in the graft accelerate allograft rejection (Figure 2B).Renal allografts from young F1-RIPLTa donors were also rejected significantly earlier (MST=72.5 days) than F1 allografts.No significant differences were detected in DSA formation (Figure 2C).Histopathology at the time of rejection demonstrated the presence of TLO in all allografts.F1-RIPLTa grafts displayed prominent TLO pre-and post-transplantation.In young F1-RIPLTa donor grafts, only occasional lymphoid aggregates were present pretransplantation, but TLO developed quickly post-transplantation (Figure 2D), which makes it impossible to separate the inflammatory effects of LTa from TLO functions.F1 allografts demonstrated development of TLO with HEV (PNAd expression) (Supplemental Figure 2) at the time of rejection, suggesting that de novo TLO formation in WT grafts requires a longer time to occur (Figure 2D).Histological quantitation of the cellular infiltrate (Figure 2E) and Banff rejection scores (Figure 2F) confirmed that the presence of preformed or rapidly forming TLO in the allograft leads to a larger immune infiltrate and higher Banff scores, reflecting the differences in median survival time.These data support that TLO and LTa-LTbR signaling contribute to chronic allograft rejection in WT recipients.

Blocking donor LTbR signaling prolongs allograft survival
To further elucidate the role of TLO in allograft rejection, we performed transplantation survival experiments where donor LTbR signaling is disrupted.LTbR is critically important for secondary lymphoid organ and TLO formation and binds two different ligands, the heterotrimeric LTa1b2 and LIGHT.While the heterotrimer LTa1b2 only binds to LTbR, the LTa3 homotrimer has inflammatory properties similar to TNFa, binds to TNFRI, II and HVEM, but not to LTbR.LTa3 has been associated with autoimmunity and inflammatory diseases.In this model, we are therefore able to separate the proinflammatory effects of LTa3 signaling from the effects of blocking LTbR signaling.We used B6 WT or B6 LTbR-ko donor grafts transplanted to Balb/c recipients (Figure 3A).The B6 to Balb/c kidney transplantation model results in acute rejection of renal allografts.As LTbR expression on stromal cells is essential for lymphoid neogenesis, the absence of LTbR on donor graft tissue prevents intragraft TLO formation and inflammatory signals mediated through LTbR.As shown in Figure 3B, B6 WT allografts are quickly rejected (MST=11 days), while B6 LTbR-ko allografts survive significantly longer (MST=24 days).No statistically significant difference in IgG DSA production in the B6 WT allograft recipients compared to B6 LTbR-ko graft recipients were observed (Figure 3C).Histopathology of a subset of grafts procured on day 9 after transplantation showed more severe immune cell infiltration in B6 WT allografts compared to B6 LTbR-ko allografts (Figure 3D).Immunofluorescence for PNAd showed presence of HEV in B6 allografts but not in B6 LTbR-ko allografts, indicating that B6 WT renal allografts promote lymphoid neogenesis very early after transplantation (Figure 3D).Quantitation of the immune infiltrate at the time of rejection revealed a significant higher infiltration in B6 WT compared to B6 LTbRko allografts (Figure 3E), and Banff rejection scores at the time of rejection were significantly higher (>IB compared to <IA)) in B6 than in B6 LTbR-ko allografts (Figure 3F).

Renal TLO support naïve immune cell activation
To further investigate the function of TLO, we developed an intravital microscopy model to study cell-cell interactions in vivo in TLO under the kidney capsule.This allowed us to compare immune cell interactions in TLO to that observed in lymph nodes.We imaged kidneys of bone marrow chimeric CD11c-YFP B6 RIP-LTa mice, where we could identify TLO by (a) the lack of normal kidney structure (capillaries), and (b) the accumulation of CD11c-YFP+ dendritic cells.To clearly define the time point of antigen introduction, we utilized transgenic OT-I (dsRed) and OT-II (CFP) T and NP-specific (CellTracker Red) B cells, imaged at time 0, and after 1, 3, 6, 24 and 72 hours after immunization with either anti-DEC-205 (OT-I T cells) or NP-Ovalbumin (OT-II and B cells) (Figure 4A).As shown in Figure 4B, TLO can be identified by MAdCAM-1 expression (HEV), accumulation of dendritic cells, presence of naïve T or B cells, which are absent in normal surrounding kidney tissue, and lack of normal kidney tissue structure.
Naïve T and B cells accumulate in distinct zones (Figure 4B).We first evaluated motility parameters of OT-I T cells before and after immunization with anti-DEC205-OVA and FGK4.5 (16).OT-I T cells displayed a reduction in mean speed and displacement after immunization and an increased arrest coefficient over time (Figure 4C and Supplemental Movie 1), which is reflected in total track length shown in the bottom panel of Figure 4C.To investigate B cell activation, we transferred NP-specific B cells (labeled with CellTracker Red) and OT-II T cells (CFP) and performed imaging at day 0, and days 1 and day 3 after immunization with NPovalbumin and adjuvant.For B cells, we observed an increase in motility and displacement after immunization, which is similar to observations made in lymph nodes and consistent with B cell activation (Figure 4D and Supplemental Movie 2) (17,18).CD4 OT-II T cells were imaged at the same time points and displayed lower mean speed and increased arrest coefficient on day 1 after immunization with increased motility parameters on day 3 (Figure 4E and Supplemental Movie 3) and associated changes in total track length (Figure 4E, bottom row).The observations in T cells are indicative of stable cell-cell interactions needed for activation and similar to motility changes that have been observed during T cell activation in lymph nodes (19,20).

Discussion
The formation of TLO through the process of lymphoid neogenesis has long been associated with chronic inflammatory conditions where antigen persists (21).This is the case during chronic infections, autoimmune diseases, and organ transplantation (22)(23)(24)(25).The discovery that many cancers also promote TLO formation and that the presence of TLO often is a predictor of better outcomes, has led to a resurgence of interest in TLO (26).TLO have immunomodulatory effects, they can either promote immunity or can be associated with immune regulation (11,27,28).In transplantation, TLO have been associated with both, chronic rejection outcomes and graft acceptance, at least in animal models (7,11,13,27).In this study, we have performed causeeffect experiments to further define the role of TLO in acute and chronic rejection in a mouse model of kidney transplantation.We are not only defining the role of TLO by manipulating the LTa -LTbR pathway in survival experiments, but also developed a model of intravital microcopy to capture for the first time the cellular events and interactions in TLO, similar to what has been studied in lymph nodes.
We demonstrate that TLO are sufficient to mediate allograft rejection in recipients that do not have secondary lymphoid tissue and do not reject allografts in the absence of TLO.This highlights that TLO are fully functional lymphoid organs that are capable of providing the environment needed for activation of an adaptive immune response in a model of vascularized, solid organ transplantation.This includes the activation of B cells and production of donorspecific antibodies (DSA), which do not develop if WT, non-TLO containing, allografts are transplanted.The chronic kidney transplantation model used in our studies is not dependent on DSA, of which we are only able to detect low levels in the serum, independent of the presence of preformed graft TLO.The significance of B cell activation and DSA production in graft TLO needs to be further investigated.
We observed that F1 allografts, although maintained long-term and not undergoing rejection, contained lymphoid aggregates that resemble the Treg-rich organized lymphoid structures (TOLS) that have been previously reported (12).Our data confirm that TOLS can form in a donor -recipient strain combination where the recipient lacks secondary lymphoid tissue as first described by Rosales et al (12).These aggregates were characterized by the presence of Treg, the absence of HEV expressing PNAd and a location around a central blood vessel.These aggregates were not present when F1 allografts were transplanted to B6 WT recipients, where we observed formation of inflammatory TLO over time.The formation of TOLS in renal allografts transplanted to LTbR -/-recipients offers an opportunity for further research to elucidate what conditions and mechanisms govern their formation and function.
This study demonstrates that TLO accelerate allograft rejection in the presence of a normal set of secondary lymphoid tissue.While WT F1 allografts also demonstrated TLO at the time of rejection, there was a marked difference in rejection tempo if TLO were present at the time of transplantation, suggesting a local contribution of TLO in the alloimmune response.A caveat of TLO studies is that the function of TLO is linked to local inflammation, as a chronic inflammatory environment is essential to provide the conditions necessary for the development of TLO.This is also applicable to the model utilized in this study.The RIP-LTa model causes local overexpression of the inflammatory mediator LTa, which then provides the signal for TLO formation.We attempted to address this by transplanting donor kidneys from young F1 RIP-LTa mice which did not harbor TLO at the time of transplantation, but the histological presence of TLO after accelerated allograft rejection confirmed that the function of TLO and inflammatory signals could not be separated in this experiment.A separate TLO-independent function of LTa can therefore not be ruled out.

In an acute kidney rejection model (B6 to
Balb/c), disrupting the LTbR pathway in donor allografts led to prolonged allograft survival compared to WT B6 donors.Prolongation of allograft survival took place with intact LTa3 and TNFa signaling pathways that mediate inflammatory signals, which highlights the importance of the LTbR pathway in allograft rejection.The outcome of prolonged allograft survival cannot be attributed solely to lymphoid neogenesis as LTbR signaling involves activation of both, the NFkB and JNK pathways, which play roles not only in lymphoid neogenesis but also mediate inflammatory signals.It is therefore possible that not only the prevention of TLO formation is causative for better allograft survival, but that the absence of inflammatory signals mediated though LTbR also contribute to this outcome.However, physiologically the presence of TLO is intrinsically linked to inflammatory signals, making TLO and the LTbR pathway relevant targets to improve allograft outcomes.The intravital microscopy data of TLO in this model suggests that productive cell-cell interactions, leading to activation of naïve T cells, are taking place in TLO.Together with the survival data presented, TLO are therefore likely to contribute to local immune activation and maintenance in this model.As the cell-cell interactions and motility parameters observed in TLO are similar to what has been described in lymph nodes, it is likely that TLO support similar immunological functions, including promoting regulatory functions under the appropriate conditions.It is promising that interfering with the LTbR pathway in the donor organ can delay allograft rejection significantly in an acute kidney graft rejection model in mice.This work used transplantation as a model to investigate the function of TLO, but the results are equally relevant to autoimmunity, cancer, and other chronic inflammatory conditions.

Figure 4 :
Figure 4: Intravital microscopy of TLO and immune cell interactions