Role of TLRs and DAMPs in allograft inflammation and transplant outcomes

HMGB1.

High mobility group box 1 (HMGB1), a nuclear protein that regulates gene transcription, is one of the best-characterized of the intracellular DAMPs that induce inflammation in transplanted organs and affect the rate of development of acute allograft rejection. In non-transplanted murine models of hepatic and renal IRI, HMGB1 levels are increased in the damaged organs and pharmacological blockade of HMGB1 attenuates inflammation^24–27. These results, together with findings from other studies which indicate that HMGB1 induces inflammation either by activating TLR4 (REF. 28) or the receptor for advanced glycation end-products (RAGE)²⁹, have led researchers to examine the function of HMGB1 in experimental organ transplantation. A 2014 study found that inhibition of necroptosis, a form of regulated cell necrosis, reduced Hmgb1 expression in cardiac isografts and administration of recombinant HMGB1 increased the expression of the proinflammatory cytokine IL-17 in the grafts³⁰. Consistent with these findings, other studies have shown that hypothermic preservation of cardiac allografts increases Hmgb1 expression after graft implantation, and IRI leads to release of HMGB1 from cardiac allografts³¹. Importantly, pharmacological blockade of HMGB1 activity extended cardiac allograft survival by 1 week, with reduced graft inflammation, in a full MHC-mismatched murine cardiac-transplant model³². In agreement with this latter study, another study of cardiac transplantation in mice found that inhibition of RAGE extended the median survival of cardiac allografts to almost 3 weeks, compared to just 1 week for allografts transplanted under control conditions³³. Of note, RAGE is not critical for the induction of renal injury in non-transplant models of IRI in mice³⁴ and a definitive role for RAGE in the inflammatory response to renal transplantation is yet to be determined. Additionally, in a rat-to-mouse cardiac xenograft model, expression of HMGB1 was upregulated in the xenograft and inhibition of HMGB1 function via an anti-HMGB1 monoclonal antibody increased xenograft survival by about 1 week, with evidence of reduced graft inflammation³⁵.

The above findings indicate that HMGB1 participates in early allograft rejection and accelerates acute allograft rejection. Findings from a 2014 study suggest that HMGB1 also has a role in chronic allograft rejection, the leading cause of graft loss for which no effective new therapies are available. In an MHC class II mismatched murine model of chronic cardiac allograft rejection, levels of HMGB1 were elevated in the graft 2 months after transplantation³⁶. Furthermore, administration of an anti-HMGB1 monoclonal antibody every 4 days during the first month after transplantation reduced the degree of graft vasculopathy by 50% and was accompanied with reduced graft levels of IL-17A, IFN-γ and inflammatory macrophages³⁶.

Repetitive minor ischaemic insults, either to the organ of interest or to remote tissues, including muscle, might elicit a protective response within the organ through a process called ischaemic preconditioning³⁷. In a mouse model of kidney IRI, prior administration of recombinant HMGB1 provided significant protection against kidney dysfunction, histological damage and inflammation³⁸. Further mechanistic analyses revealed that HMGB1 preconditioning elicited upregulation of Siglec-G, a negative regulator of TLR4 signalling¹⁴. Findings from studies in experimental infection models suggest that innate immune cells can exhibit a form of memory — they can enhance immune responses after repeat stimulation through a phenomenon known as ‘trained immunity’ (REF. 39). Whether this phenomenon occurs in organ transplantation has not yet been determined but will need to be investigated if repeated administration of DAMPs is to be considered as a ischaemic-preconditioning strategy in patients undergoing organ transplantation.

Overall, these studies indicate that HMGB1 is released after organ implantation, contributes to graft inflammation and accelerates graft rejection. However, transplant survival in the various models was typically only extended from 1 week to 3 weeks after transplantation with HMGB1 inhibition³², indicating that HMGB1 is not the only contributor to graft inflammation.

Hyaluronan.

Hyaluronan (also called hyaluronic acid) is a glycosaminoglycan produced by mesenchymal cells that has been implicated in a variety of experimental models of sterile inflammation, most notably in bleomycin-induced lung injury^40–42 and in murine non-transplant IRI models^43,44. In quiescent, noninflammatory states, hyaluronan exists in a large molecular weight form, which undergoes degradation upon inflammation. Hyaluronan fragments can activate TLR2 and TLR4 to induce inflammation⁴⁵, and the large form of hyaluronan is thought to be immunoregulatory⁴⁶. However, a 2014 study found that overexpression of the enzyme that fragments hyaluronan in murine skin does not lead to skin inflammation, but rather promotes the migration of dendritic cells out of the skin, which reduces subsequent delayed-type hypersensitivity responses in the skin⁴⁷. This study indicates that hyaluronan fragments could be important to control immune-cell emigration before immune activation. Whether this is also the case in internal organs requires further study.

As well as activating TLR2, TLR4 and their downstream cytoplasmic signal adaptor MYD88⁴⁰, hyaluronan can activate CD44 in T cells in the lung⁴¹. Levels of hyaluronan are elevated in human kidney allografts undergoing acute rejection⁴⁸, in minor mismatched skin grafts undergoing rejection and in human lung transplants with evidence of chronic rejection⁴⁹, compared to levels in nonrejecting grafts. Fragmented, but not full length hyaluronan, acts via TLR2, TLR4 and MYD88 to activate dendritic cells to prime alloreactive T cells in vitro ⁴⁹. These findings are compatible with the reported presence of hyaluronan in fibrotic lung tissue from human lung transplant recipients⁵⁰. Interestingly, in an orthotopic lung allograft model in which transplant tolerance was induced by combined anti-CD154 and CTLA4 immunoglobulin treatment, fragmented, but not full length hyaluronan, abrogated transplant tolerance via a TLR2, TLR4 and MYD88-dependent pathway⁵⁰. Fragmented hyaluronan is therefore likely to contribute to inflammation and acute graft rejection and might also impair the tolerance of some organ allografts. Investigation of the role of fragmented hyaluronan in chronic allograft rejection requires tools to efficiently deplete hyaluronan or hyaluronan fragments in murine models.

Other DAMPs.

Other factors that contribute to sterile inflammation, such as uric acid⁵¹ and mitochondrial DNA (which is involved in inflammation after trauma⁵²), are not known to be causally involved in inflammation after solid organ transplantation. Some intracellular proteins, such as members of the heat shock protein-70 family, are thought to trigger inflammation; yet, when examined in experimental murine skin transplant models, Hsp1b deletion in the donor or Hsp1a deletion in both the donor and the recipient, did not affect the timing of allograft rejection^53,54. Using proteomic approaches, the protein haptoglobin, which is known for its haeme binding and antioxidant properties⁵⁵, was found to be upregulated in both syngeneic and allogeneic skin grafts in mice⁵⁶. Purified haptoglobin activates dendritic cells via MYD88 to enhance the priming of alloreactive T cells⁵⁶. In a minor histocompatibility mismatched skin transplant model in which MYD88 is required for graft rejection⁵⁷, donor-derived haptoglobin accelerated acute graft rejection⁵⁶. In a full MHC mismatched cardiac transplant model, recipient-derived haptoglobin, but not donor-derived haptoglobin, impaired the ability of CTLA4 immunoglobulin to induce indefinite cardiac allograft survival⁵⁸. Haptoglobin amplified intragraft inflammation by stimulating the accumulation of dendritic cells in the graft, by increasing the expression of the proinflammatory cytokine IL-6 and of the neutrophil attracting chemokine CXCL2 (also known as MIP2), and by reducing the levels of the immunosuppressive cytokine IL-10 within the first 2 weeks of transplantation. Finally, haptoglobin was present in biopsy samples from human cardiac transplants that exhibited moderate cellular rejection, providing initial clinical translation of the experimental findings⁵⁸. These studies indicate that discovery approaches such as proteomics can be powerful tools to identify novel DAMPs after organ transplantation.

Haptoglobin is probably not involved in the initiation of IRI-induced inflammation in the graft after transplantation. Therefore, we speculate that known (and possibly unknown) DAMPs, such as HMGB1, are released upon IRI and activate innate immune signalling pathways within the graft, which heighten the local inflammatory milieu and subsequently lead to the local release of proinflammatory cytokines such as IL-6 or IL-1. These inflammatory mediators then enter the circulation to induce the production of inflammatory amplifiers such as haptoglobin by the recipient (FIG. 1). These amplifiers tip the balance from graft acceptance to graft rejection by increasing inflammation, which leads to an increase in subsequent anti-donor T-cell responses.

DAMP-mediated signalling.

HMGB1 is upregulated in human cadaveric kidney transplants⁷⁴. Moreover, a follow-up analysis in human kidney transplant recipients showed that urinary HMGB1 levels were elevated as early as 3 h after transplantation⁸¹. In addition, the presence of HMGB1 in the peripheral blood of kidney transplant recipients was associated with an upregulation of TLR2 and TLR4 in peripheral blood leukocytes, which suggests that HMGB1 induces systemic TLR activation⁸¹. Similar observations have been made in liver transplant recipients⁸². Exposure of human tubular cells to HMGB1 in vitro induces the release of inflammatory cytokines (including IL-6, IL-1β, TNF and CCL2), suggesting that HMGB1 promotes inflammation in human organ transplantation⁷⁴ (TABLE 1).

Transcriptomic analysis of bronchoalveolar lavage fluid and lung tissue biopsy samples demonstrated that activation of innate immune pathways occurs in lung transplantation^83,84. Specifically, gene expression analysis of bronchoalveolar lavage fluid before and after reperfusion in lung transplant recipients showed upregulation of the inflammasome (including NLRP3, IL-1β and caspase-1), inflammatory chemokines and cytokines (CXCL1, CXCL2, CCL7 and IL-6) and TLRs (TLR2, TLR4, TLR6 and TRL9) — key mediators of early graft dysfunction⁸³. Of note, the gene expression analysis conducted in these studies does not detect post-translational modifications of proteins such as caspase-1 and pro-caspase-1, which are important inflammatory events. These studies also demonstrated that longer ventilation times were associated with a stronger inflammatory profile, including a higher expression of TLRs and inflammatory cytokines^83,84, indicating that mechanical ventilation directly causes lung injury and inflammation, and impairs early graft function.

TLR-mediated signalling.

Interestingly, in lung transplant recipients expression of TLRs strongly correlated with that of inflammatory cytokines, providing evidence that TLRs are the main mediators of graft inflammation after human lung transplantation⁸³ (TABLE 1). The importance of TLR4 signalling in IRI after organ transplantation has been further illustrated by genetic studies of specific single nucleotide polymorphisms (SNPs) in the Tlr4 allele. Presence of the Tlr4 loss-of-function alleles Asp299Gly or Thr399Ile was associated with lower rates of acute allograft rejection after kidney, lung and liver transplantation, and with lower intragraft levels of proinflammatory cytokines such as TNF and CCL2 (REFS 74,85,86). In addition, TLR gene-expression profiling of biopsy samples from kidney transplant recipients showed increased expression of TLR1, TLR2, TLR4, TLR7 and TLR8 in patients undergoing acute rejection, and TLR4 expression within the graft correlated with cellular infiltration and graft damage, according to the 2007 Banff classification⁸⁷, indicating that TLRs, and TLR4 in particular, are important for graft inflammation after kidney transplantation⁷⁴. However, in a 2015 study, TLR SNPs did not correlate with outcomes after kidney transplantation⁸⁸. These discrepancies could stem from differences in cohort sizes, SNPs analysed and statistical methods. Nonetheless, given that loss-of-function of Tlr4 was affected by the expression of inflammatory cytokines in the graft⁷⁴, we suggest that the weight of current evidence indicates a functional role of TLR4 in graft inflammation.

The mechanisms by which TLR4 is upregulated during IRI after organ transplantation remain unclear, although one can speculate that hypoxic and oxidative stress mediated by IRI or subsequent inflammation induced by DAMPs promote TLR4 upregulation in human epithelial and immune cells^89–91. In support of a role for DAMPs in triggering TLR4 upregulation, release of RAGE after graft implantation is predictive of a poor short-term outcome after lung transplantation⁹². Specifically, plasma levels of RAGE 4 h after reperfusion correlate with a longer duration of mechanical ventilation and length of stay in the intensive care unit, independently of graft ischaemia time⁹².

TLRs.

Genetic analyses in 787 liver allografts identified three donor SNPs in the Tlr4 allele that were associated with graft failure⁹³. However, in a separate study, when liver transplant recipients were stratified on the basis of their hepatitis C virus infection status, the association between Tlr4 polymorphisms and late graft failure was only found in patients with hepatitis C, suggesting that Tlr4 polymorphisms synergize with hepatitis C virus infection to enhance graft rejection⁸⁶. In lung transplantation, recipient Tlr4 SNPs influence graft rejection. An analysis of 110 lung transplant recipients, including 20 patients with bronchiolitis obliterans syndrome (BOS), identified an association between polymorphisms in Tlr4, Tlr2 and Tlr9 and the development of BOS⁹⁴, which indicates that these TLR polymorphisms enhance the development of chronic rejection after lung transplantation.

In kidney transplantation, loss-of-function polymorphisms in donor TLR4 are associated with a reduction in the risk of acute rejection⁹⁵; however, these outcomes do not translate into improvements in long-term kidney allograft function. Indeed, allograft function 1–5 years after transplantation was reported as similar, regardless of donor polymorphism status⁹⁶. Similarly, in a case-controlled study of kidney transplant recipients, graft SNPs in TLRs 1–8 were not associated with biopsy-proven graft rejection or mortality by 4 years after transplantation, although a comparison of kidney transplant recipients and donors identified an association between TLR SNPs and end-stage renal disease⁸⁸.

Although genetic association studies have yielded inconsistent results, transcriptional analysis of tissue biopsy samples from starndard criteria living-donor kidney grafts showed an ongoing injury response and TLR-mediated inflammation after transplantation^97,98. Specifically, development of inflammation during the first year after kidney transplantation is associated with upregulation of TLR4 and TLR2 (REFS 97,98). Higher TLR expression was only observed in patients with both graft interstitial fibrosis and cellular infiltrates and not in patients with normal graft histology or fibrosis alone⁹⁷. TLR4 expression is also increased in human kidney grafts undergoing chronic rejection compared to stable controls⁹⁹. Consistent with these findings, the expression pattern of MYD88 and TLR4 in the peripheral blood leukocytes of human kidney transplant recipients enables distinction between patients with chronic rejection and operationally tolerant recipients⁹⁹. In heart transplant recipients, expression of TLR4 on circulating monocytes associates with endothelial dysfunction within the allograft up to 3 years after transplantation, although evidence of chronic rejection was not reported in this study¹⁰⁰. Although upregulation of the TLR pathway has been reported in these studies, it is often associated (and sometimes masked) by the overexpression of genes involved in the adaptive immune response and other components of innate immunity, which illustrates the molecular complexity of the immune response during graft rejection^97,101,102.

DAMPs.

DAMPs are present in chronically rejected human allografts^50,103. Notably, lung transplant recipients with BOS have an accumulation of hyaluronan in the fibrous tissue of the airway lumen. These patients have a higher expression of hyaluronan synthase 1 in their lung tissue and higher levels of hyaluronan in their bronchoalveolar lavage fluid and plasma, as compared to grafted patients without BOS⁵⁰. Extensive profiling of DAMPs in the bronchoalveolar lavage fluid of lung transplant recipients identified a specific biological pattern for BOS and for restrictive allograft syndrome. Specifically, levels of HMGB1 and S100 family proteins (including S100A8, S100A9, S100A8/A9, S100A12 and S100P), which can activate Tlr4 and induce innate immune-cell recruitment and endothelial cell activation¹⁰⁴, were elevated in patients with restrictive allograft syndrome and BOS compared to control patients with stable graft function¹⁰³. The production of these DAMPs was slightly increased in the restrictive allograft syndrome group compared to levels in the BOS group¹⁰³. By contrast, expression of S100A8 in human kidney allografts was associated with a reduced incidence of chronic allograft vasculopathy¹⁰⁵. This finding is consistent with experimental evidence showing that S100A8 contributes to resolution of inflammation after renal injury in non-transplant models of IRI¹⁰¹ and reduces inflammation after allotransplantation^106,107. Differences between lung and kidney transplants (for example lung transplants are exposed to the environment whereas kidney grafts are not), might explain the divergent results in regard to S100 proteins.

Inhibition of TLR signalling.

Several antibodies and compounds that target TLR4 are in development or in clinical trials. Given that TLR4 is ubiquitously expressed and crucial for providing protection against infection, strategies that aim to nonspecifically block the TLR4 signalling pathway might incur a considerable risk of infectious complications.

Genetic deletion of pivotal molecules in the innate immune response (individual TLRs or molecules essential for their downstream signalling, such as MYD88) is achievable in mice, particularly when performed in specific pathogen-free environments, and has provided mechanistic insights into the pathogenic roles of these factors, as discussed above. Such major interruptions of innate immune signalling pathways are also theoretically achievable in humans through the use of blocking antibodies or small molecular antagonists, but are likely to carry high risks of infection and cancer. Inhibiting the ligands that activate TLRs directly in the allograft before implantation is a conceptually attractive strategy to avoid compromising host defence in transplant recipients. This strategy has yet to be tested experimentally or clinically. Despite the above-mentioned concerns, several clinical studies that inhibit TLRs are in progress.

A phase II clinical trial is currently testing the therapeutic efficacy of OPN-305, a monoclonal antibody that blocks TLR2, to dampen inflammation and prevent early dysfunction of extended-criteria-donor organs (donor age >60 years)¹⁰⁸. Another clinical trial will assess the safety and tolerability of an anti-TLR4 antibody in healthy individuals¹⁰⁹. The results of these studies will undoubtedly provide interesting insights as to the therapeutic potential of blocking specific TLRs to reduce inflammation after organ transplantation.

Stimulation of resolution pathways.

Tissues express a large array of evolutionary conserved cellular pathways that control tissue damage and counterbalance the deleterious effects of IRI¹¹⁰. Acute inflammation triggers inflammation resolution pathways that are required for tissue repair, including those regulated by haemoxygenase 1, heat shock proteins and nuclear factor erythroid 2-related factor 2 (NFE2-related factor 2, also known as NRF2)^111,112. An experimental study found that heat shock preconditioning reduces damage after renal IRI through the expansion of regulatory T cells¹¹³. Hypoxia-inducible factor 1α (Hif-1α) is an oxygen-sensitive transcription factor that regulates tissue metabolism and angiogenesis¹¹⁴ and its induction reduces infarct size after myocardial infarction in mice¹¹⁵. Hif-1α is also critical for ischaemia preconditioning in murine models of myocardial IRI¹¹⁶. These results have given rise to strategies to precondition allografts using drugs that enhance the activity of Hif-1α or by inducing minor and remote ischaemic insults¹¹⁷. Such graft preconditioning should induce protective mechanisms that enable a greater degree of tolerance to tissue damage¹¹⁸. To date, the majority of the approaches being investigated in clinical trials to reduce inflammation have employed remote ischaemic preconditioning. In organ transplantation, this procedure consists of three 5-min cycles of left lower limb ischaemia, induced by an automated cuff inflator placed on the limb and inflated to 300 mm Hg, with an intervening 5 min of reperfusion during which the cuff is deflated¹¹⁹. However, this approach failed to demonstrate beneficial effects on early graft function or rates of acute rejection in human kidney transplantation¹¹⁹. Other experimental approaches that directly target stress responses in the damaged organ by enhancing Hif-1α activity via small molecules such as l-mimosine¹²⁰, or promoting HO1 activity by exposing the graft to haeme arginate¹²¹ have shown beneficial effects and have provided an impetus for several ongoing clinical studies^122–124. Patient safety data are already available for haeme arginate and this agent is now under investigation in a randomized controlled clinical trial in recipients of deceased donor kidneys¹²².