Innate immune memory responses in organ transplantation

Abstract

Innate immunity is a critical contributor to graft rejection and cardiovascular complications following organ transplantation, with increasing evidence indicating that innate immune memory significantly influences graft outcomes. The most extensively studied form, trained immunity, involves epigenetic and metabolic reprogramming of innate immune cells, altering their inflammatory responsiveness. Numerous transplantation-related factors, including metabolic disturbances such as hypercholesterolemia and hyperglycemia, can induce trained immunity, and both experimental and clinical studies have linked it to graft survival. While trained immunity reflects responses to nonspecific stimuli, innate allogeneic memory shows that innate immune cells can recognize non-self and mount donor-specific memory responses. This review covers the current knowledge on both forms of innate immune memory in the context of solid organ transplantation.

Graphical Abstract

1. Introduction

Innate immune cells play a central role in the response to solid organ transplantation. Beyond host defense, they orchestrate inflammatory responses to alloimmune injury and can directly contribute to graft rejection. Early studies demonstrated that innate immune cells can mediate robust allograft responses, with macrophages dominating acutely rejecting kidney allografts [1], and graft rejection occurring even in the absence of adaptive immune cells [2]. Besides acute rejection, innate immune mechanisms contribute to cardiac allograft vasculopathy, a major cause of late graft failure in heart transplant recipients [3]. Beyond graft outcomes, innate immune pathways are implicated in the heightened cardiovascular risk after transplantation. Cardiovascular disease remains a leading cause of morbidity and mortality after transplantation, partly driven by post-transplant factors such as immunosuppressant-induced hyperlipidemia and hyperglycemia [4]. These metabolic disturbances are potent activators of innate immune pathways [5, 6], which can regulate vascular homeostasis but can also promote chronic inflammation, fibrosis, and atherosclerosis [7]. Consequently, dysregulated innate immune activation is a key mechanism linking transplantation to graft vasculopathy and cardiovascular disease.

Innate immune cell functions are dynamically shaped by environmental cues. Upon activation, they can undergo durable functional reprogramming that enhances their responsiveness to subsequent, unrelated stimuli. This phenomenon, termed trained immunity, challenges the paradigm that immunological memory is exclusive to the adaptive immune system [8]. In contrast to adaptive immune memory, which depends on antigen receptor gene rearrangements, trained immunity is mediated by epigenetic and metabolic remodeling. While trained immunity can enhance host defense, it may also exacerbate chronic inflammation and promote vascular pathology [9].

Although trained immunity represents memory to nonspecific stimuli, recent studies have shown that innate immune cells can also develop antigen-specific memory responses to non-self antigens [10–12]. Whether these reflect distinct mechanisms or shared molecular pathways remains unresolved.

Interest in innate immune memory has grown rapidly in transplantation immunology, a field traditionally focused on adaptive immune responses. Increasing evidence links innate immune memory to graft injury and long-term outcomes [13, 14]. Although its relevance has been recognized for several years [15], recent advances have deepened our understanding of how innate immune memory shapes interactions between immune and non-immune cells, including T cells, endothelial cells, and the microbiome [16, 17]. Furthermore, the first clinical studies have supported its role in influencing long-term graft survival [13, 18–20].

In this review, we summarize current insights into innate immune memory in solid organ transplantation.

2. Trained immunity responses in organ transplantation

Trained immunity refers to durable functional reprogramming of innate immune cells, in which an initial stimulus induces long-lasting changes that enhance inflammatory responses to subsequent, non-specific challenges [8]. In contrast, innate immune tolerance is a hyporesponsive state in which prior exposure to stimuli such as lipopolysaccharide (LPS) dampens subsequent responses. Both training and innate immune tolerance occur not only in peripheral cells, but also in myeloid progenitor cells in hematopoietic organs, resulting in systemic, long-lasting reprogramming of the innate immune system [21].

Epigenetic remodeling is central to trained immunity. In resting myeloid cells, pro-inflammatory genes are repressed by condensed chromatin, with accessibility regulated by DNA methylation and histone modifications [21, 22]. DNA methylation is generally repressive and more stable than histone changes [22]. Key histone marks include: histone 3 lysine 27 acetylation (H3K27ac) at enhancers, typically accompanied by histone 3 lysine 4 monomethylation (H3K4me1); histone 3 lysine 4 trimethylation (H3K4me3) at promoters; and histone 3 lysine 18 lactylation (H3K18la) at distal regulatory sites of immune genes. Together, these modifications increase chromatin accessibility and promote RNA polymerase II recruitment, enabling rapid transcription of inflammatory genes [21, 23, 24]. Notably, LPS exposure can delay or prevent H3K27ac deposition, consistent with innate immune tolerance [25]. The mechanisms governing these changes are increasingly defined. For example, nuclear factor of activated T cells (NFAT)–dependent long non-coding RNAs (lncRNAs) recruit WD repeat-containing protein 5/mixed lineage leukemia 1 (WDR5/MLL1) methyltransferase complexes to promoters within the same topologically associating domains, enabling locus-specific H3K4me3 [26].

Metabolic rewiring further supports trained immunity. To meet the energetic demands of inflammation, cells shift from oxidative phosphorylation to aerobic glycolysis, glutaminolysis, and cholesterol metabolism, largely through mTOR activation [Figure 1] [27, 28]. Beyond energy production, metabolic intermediates directly regulate epigenetic remodeling. For instance, acetyl-CoA and lactate serve as substrates for the enzymes driving H3K27ac and H3K18la deposition respectively [23, 24, 28]. Epigenetic changes in turn enhance transcription of metabolic enzymes, creating a positive feedback loop that reinforces metabolic activation [23, 28].

Figure 1: Metabolic and epigenetic interplay in trained immunity.

Dectin-1–mediated calcium influx activates calmodulin–calcineurin–dependent transcription of long non-coding RNAs (lncRNAs) via Nuclear factor of activated T-cells (NFAT). These lncRNAs promote histone 3 lysine 4 trimethylation (H3K4me3) at immune gene promoters by recruiting WD Repeat-Containing Protein 5/Mixed Lineage Leukemia-1 (WDR5/MLL1) complexes. In parallel, Dectin-1 signaling induces mechanistic target of rapamycin(mTOR)–mediated metabolic reprogramming, including increased glycolysis, leading to accumulation of metabolites that drive histone lactylation and (de)acetylation. Together, these epigenetic modifications enhance pro-inflammatory gene transcription.

TCA cycle: Tricarboxylic acid cycle

Since trained immunity involves an increased responsiveness of innate immune cells, it is conceivable that it could have negative consequences in the context of organ transplantation. In recent years, the first steps have been taken to investigate this possibility, with the central questions being: How is trained immunity induced in organ transplantation? What is the effect of trained immunity on allograft outcomes? And how can we therapeutically target trained immunity?

2.1. Endogenous and microbial drivers of trained immunity in organ transplantation

Microbial triggers of trained immunity

Various pathogen associated molecular patterns (PAMPs) induce trained immunity [Figure 2A] [8], which poses challenges in organ transplantation, where recipients face increased infection risk before transplantation due to comorbidities and after transplantation due to lifelong immunosuppression [29]. Infections have been associated with graft rejection in multiple human studies [29–31]. While several mechanisms may contribute to this association, experimental evidence suggests involvement of innate immune responses. For example, Simon et al. showed that infection with Staphylococcus aureus, a potent inducer of trained immunity [32], prevented skin graft acceptance in mice receiving costimulatory blockade [33]. This immune reaction strictly required innate immune recognition and interleukine-6 (IL-6) signaling, as Myeloid differentiation primary response 88 (MyD88)-, Recombinase-Activating Gene (RAG)-, or IL-6-deficient recipients developed graft acceptance despite infection. Similarly, Listeria monocytogenes infection prevented tolerance induction in a mouse skin and cardiac allograft model and reversed graft acceptance in a mouse cardiac allograft model [34, 35].

Figure 2: Trained immunity and innate allogeneic memory in organ transplantation.

In organ transplantation, pathogen associated molecular patterns (PAMPs), danger associated molecular patterns (DAMPs) and T cell interactions via interferon gamma (IFNy) or CD40-CD40L signaling induce trained immunity (1A) through pathways involving mammalian target of rapamycine (mTOR), nuclear factor of activated T cells (NFAT) and nuclear factor kappa-B (NF-κB). This drives metabolic and epigenetic reprogramming, enhancing pro-inflammatory cytokine responses and T cell activation and differentiation.

Innate allogeneic memory (1B) is triggered when donor signal regulatory protein α (SIRPa) variants bind CD47 with higher affinity than the host’s SIRPa variant, or when paired immunoglobulin-like receptor A (PIR-A, human ortholog: leukocyte immunoglobulin-like receptors A (LILR-A)) recognizes non-self major histocompatibility complex-I (MHC-I, human ortholog: human leukocyte antigen A (HLA-A)) inducing CD40-CD40L-dependent PIR-A transcription via NF-κB. This promotes intragraft monocyte-derived dendritic cell (Mono-DC) accumulation, T cell activation and differentiation, and increased macrophage-mediated alloreactive cytotoxicity.

PRRs: pattern recognition receptors, IL-6: interleukin 6, TNF: tumor necrose factor, IL1B: interleukin 1 beta

The relation between PAMP-induced innate immune responses and rejection raises the question whether tissue-resident macrophages with frequent PAMP exposure are more prone to trigger rejection. For instance, does training of lung-resident macrophages by repeated PAMP exposure [36, 37] contribute to the high incidence of lung allograft rejection [38]? By contrast, acute rejection is less frequent in liver transplantation and rarely leads to chronic graft damage, even though studies show liver-resident Kupffer cells can be trained and are continuously exposed to microbial stimuli via the portal circulation [39, 40]. This suggests that tolerogenic mechanisms within the hepatic microenvironment may override pro-inflammatory training effects.

The gut microbiome represents another association between microbes and graft outcomes. In chronic kidney disease (CKD), gut dysbiosis is common, with reduced microbial diversity, lower short-chain fatty acid (SCFA) production, and increased uremic toxin levels. Interestingly, Johannes et al. [19] showed that the microbiome partially normalizes after kidney transplantation, but shifts back to a dysbiotic, CKD-like composition with reduced SCFA production preceding rejection, before recovering again. The transient decline in SCFAs may favor trained immunity, as SCFAs, particularly butyrate, can induce innate immune tolerance through metabolic and epigenetic reprogramming, resulting in reduced cytokine production upon stimulation [41], while simultaneously enhancing antimicrobial activity [42]. Microbial metabolites may also exert systemic effects. For example, microbiota-derived lactate has been reported to promote proliferation and differentiation of hematopoietic stem cells [43], although its impact on innate immune memory in transplantation remains unclear. Importantly, this interaction appears bidirectional, as trained immunity responses after Bacille Calmette-Guérin (BCG) vaccination correlate with microbiota composition [20].

These studies highlight how the microbe-innate immune cell interactions may play a role in developing trained immunity responses and affecting graft outcomes. How these processes influence each other during transplantation, and what the net effect on graft outcomes is, remains to be clarified.

Sterile triggers of trained immunity

In addition to pathogens, innate immune cells respond to sterile danger signals released during tissue damage or stress. These danger-associated molecular patterns (DAMPs) can induce trained immunity [Figure 2A] [13–15] and are abundant throughout the transplantation process.

Already prior to transplantation, organ failure and organ replacement therapies markedly increase the presence of DAMPs in the circulation. For example, uremic toxins have been shown to induce trained immunity in healthy human monocytes in vitro and in mice in vivo [44] and accumulate during end-stage renal disease. The uremic state of chronic kidney failure patients and the mechanic procedure of dialysis were linked to an increased production of pro-inflammatory monocytes compared to healthy controls [44–46]. Furthermore, recent work by Nakayama et al. demonstrated that heart failure induces epigenetic reprogramming of hematopoietic stem cells, enhancing their capacity to generate pro-inflammatory macrophages in mice. Interestingly, bone marrow transplantation of mice experiencing cardiac distress was sufficient to induce spontaneous heart failure and related comorbidities in recipient animals [47].

Moreover, metabolic DAMPs and cardiovascular risk factors such as hyperlipidemia and hyperglycemia are highly prevalent in patients with end-stage organ failure and can induce trained immunity [4–6]. These metabolic perturbations may persist or even worsen after transplantation as a consequence of immunosuppressive therapy [4].

During and directly after transplantation, large amounts of DAMPs, like vimentin, High Mobility Group Box 1 (HMGB1), adenosine triphosphate (ATP) and histones, are released from dying cells during ischemia-reperfusion injury (IRI) [13–15]. This has led to the hypothesis that induces a trained innate immune phenotype, supported by experimental evidence. Dong et al. demonstrated that blood monocytes and bone marrow-derived macrophages from mice after myocardial infarction displayed epigenetic modifications at the spleen tyrosine kinase (SYK) locus, a critical regulator of proinflammatory genes including Interleukin 1 beta (Il1b), Tumor necrosis factor (Tnf), and Intercellular Adhesion Molecule-1 (ICAM-1). These cells showed enhanced SYK expression upon stimulation with oxidized low-density lipoprotein. Notably, myocardial infarction-induced trained immunity could be transferred via bone marrow transplantation, increasing pro-inflammatory cytokine production by blood monocytes and atherosclerosis in recipient mice [48]. Similarly, liver IRI in mice upregulated metabolic enzymes involved in trained immunity, an effect further amplified by ischemic pre- and post-conditioning [49]. Pre-operative exercise could inhibit this training in Kupffer cells, promoting an anti-inflammatory metabolic program and reducing liver injury [50], highlighting promising therapeutic potential.

However, the relation between trained immunity and IRI remains poorly understood, with conflicting findings across studies. For example, Jonkman et al. found that renal ischemia failed to induce trained immunity of murine bone marrow cells and splenocytes [13]. Such discrepancies may reflect differences in experimental design, including the organ, immune cell populations and trained immunity readouts that are being studied. This complexity is further compounded by immunosuppressive therapies that may override IRI-induced training effects in a clinical setting.

2.2. Cellular interactions in trained immunity in organ transplantation

T cells

In addition to PAMPs and DAMPs, recent research has identified T cells as inducers of trained immunity [Figure 2A]. This is particularly relevant in transplantation, where T cell activation plays a central role. Jacobs et al. demonstrated that T cells modulate monocyte training via CD40–CD40L interactions [16]. In addition to direct cell-cell contact, interferon gamma (IFN-γ) produced by T or natural killer (NK) cells also plays a vital role in the induction of trained immunity [36]. However, the extent to which T cells contribute to the induction of trained immunity appears to be both model- and context-dependent. In vivo studies employing T cell-deficient or -depleted mouse models have demonstrated that trained immunity can, under certain conditions, be induced in the absence of an adaptive immune system [51], whereas in other settings the presence of T cells is indispensable [36]. Similarly, some studies investigating primary human monocytes in vitro have reported that trained immunity is impaired in the absence of T cells [16], while in other contexts T cells are not strictly required but exert an amplifying effect [51].

On the other hand, trained immunity also shapes the three classical T cell activation signals. Antigen presentation and costimulation (signal I and II) are enhanced by trained innate immune cells: BCG-trained murine macrophages upregulate major histocompatibility complex (MHC)-II, CD80, and CD40, while β-glucan–trained human monocyte-derived dendritic cells (mono-DCs) increase CD11c, Human Leukocyte Antigens (HLA-)DR, CD86, and CD40 expression [52, 53]. In contrast, LPS-induced tolerance reduces MHC-II and CD86 on monocytes and macrophages [52]. Signal III, cytokine signaling, is similarly affected: trained cells produce more TNF, IL-6, and Interleukin-1 beta (IL-1β), favoring T helper 17 differentiation [54], which is implicated in allograft rejection [55]. In contrast, tolerized macrophages produce more Interleukin-10 (IL-10), supporting regulatory T cell responses [56, 57]. These changes can alter alloreactivity: BCG-trained murine macrophages enhance allogeneic CD4⁺ T cell proliferation, while LPS-tolerized macrophages suppress it [52]. Similarly, human β-glucan–trained dendritic cells (DCs) augment allogeneic CD8⁺ T cell proliferation and IFN-γ production [53].

Endothelial cells

Vascular pathology critically determines allograft outcome, as vascular inflammation underlies acute rejection and allograft vasculopathy is a major cause of long-term graft failure [58, 59]. Endothelial cells are the first to encounter the host’s immune defense and thereby play a central role in orchestrating graft inflammation. They can be activated by non-immune triggers, such as DAMPs [60], but also by immunological activators such as anti-HLA antibodies, IFN-γ, and TNF [60, 61]. Activated endothelial cells upregulate HLA, costimulatory molecules, and adhesion molecules such as ICAM-1, while secreting chemokines and cytokines (e.g., Monocyte Chemoattractant Protein-1 (MCP-1), IL-1β, IL-6) that can recruit and functionally reprogram innate immune cells [17, 60, 62–64].

Notably, endothelial cells themselves exhibit features of innate immune memory: prior exposure to TNF or IFN-γ induces persistent epigenetic and transcriptional changes that can enhance monocyte recruitment and antigen presentation upon re-stimulation [17]. Moreover, transient hyperglycemia induces persistent H3K4me1 enrichment at the nuclear factor kappa-B (NF-κB) p65 promoter through Set7 methyltransferase activity, consistent with trained immunity [65, 66]. This increases NF-κB–dependent inflammatory gene expression which drive monocyte recruitment [65]. Furthermore, lipoprotein(a) and its component lysophosphatidylcholine promote glycolytic and epigenetic reprogramming, reinforcing a persistent pro-inflammatory state marked by increased IL-6 production and enhanced monocyte recruitment [64, 67].

Together, these findings support a model in which endothelial cells can be trained by immune and non-immune stimuli, thereby amplifying innate immune recruitment and activation. Further research is needed to determine whether trained endothelial cells can drive training and recruitment of innate immune cells within the graft, and whether this contributes to vascular allograft injury.

2.3. Trained immunity and allograft outcomes

The association between trained immunity and graft survival has been explored in only one observational human study. Jonkman et al. analyzed sera from 96 kidney transplant recipients before and one week after transplantation, with long-term follow-up [13]. Healthy human peripheral blood mononuclear cells (PBMCs) were incubated with these sera for 24 hours, rested for 5 days, and subsequently restimulated with LPS. Trained immunity was assessed by measuring IL-6 and TNF production. While pre-transplant serum induced variable cytokine responses upon restimulation, post-transplant serum consistently suppressed inflammatory responsiveness. This challenges the prevailing assumption that trained immunity-inducing factors dominate after transplantation and instead underscores the tolerizing influence of clinical immunosuppression in this model. Supporting this, the cytokine response after training with post-transplant serum did not correlate with ischemia-related factors, including DAMP levels, living versus deceased donor transplantation, delayed graft function, or ischemia duration. Only whole blood tacrolimus levels one week post-transplantation correlated with IL-6 responses, suggesting a key role for immunosuppression in modulating innate immune memory.

To investigate the association between graft survival and trained immunity, Jonkman et al. stratified patients into tertiles based on the LPS-induced IL-6/TNF response after training with their post-transplant serum. Strikingly, those in the lowest response group showed a 96.9% death-censored graft survival, compared to just 68.8% in the highest response group, after a mean follow-up of 6.8 years. Acute rejection rates did not differ between groups, suggesting that chronic immunological mechanisms may underlie the differences in long-term graft survival. These findings motivated a prospective clinical trial and support innate immune memory as a potential biomarker and therapeutic target in transplantation.

2.4. Therapeutic targeting of trained immunity to improve graft outcomes

Efforts to target innate immune memory in transplantation has driven the development of nanobiologics for myeloid-specific delivery of tolerizing agents, offering mechanistic insight into graft survival and a strategy to reduce off-target toxicity. Myeloid-specific targeting is achieved by incorporating drugs in high density lipoprotein (HDL)-based particles with phospholipids and apolipoprotein A-I (apoA-I) [68]. Braza et al. were the first to test this approach in a mouse heart transplant model using mammalian target of rapamycin (mTOR) inhibitor-loaded HDL nanobiologics (mTORi-HDL) [14]. mTORi-HDL induced metabolic and epigenetic innate immune tolerance in vivo. Macrophages from treated mice could prevent CD8⁺ T cell proliferation and expanded regulatory T (Treg) cells in vitro. In line with this, short term mTORi-HDL treatment increased Treg cell infiltration in the allograft in vivo and significantly increased graft survival. Similarly, graft survival was extended when blocking T cell-driven monocyte training using CD40-Tumor Necrosis Factor Receptor-Associated Factor 6 (TRAF6) inhibitor-loaded HDL nanobiologics (CD40-TRAF6i-NBs) [14, 16]. Long-term graft acceptance was achieved by combining CD40-TRAF6i-NBs with mTORi-HDL inducing innate immune tolerance through two complementary mechanisms [14]. Similar results were seen when trained immunity responses were suppressed with CD40-TRAF6i-NBs, and combined with Cytotoxic T-lymphocyte–associated protein 4–immunoglobulin fusion protein (CTLA4-Ig) treatment, a myeloid–T cell co-stimulation blocker that inhibits T cell activation [16]. These findings demonstrate the potential of targeting innate immunity to promote a tolerogenic adaptive immune response and long-term graft acceptance with short term targeting. This strategy could reduce reliance on continuous, daily immunosuppression and its chronic off-target side effects (e.g., hyperglycemia, hypertension, renal dysfunction), but the stability of innate immune tolerance is still unclear. It may be reversible with new stimuli, whereas irreversible tolerance could limit the ability to reduce immunosuppression when clinically necessary, such as during severe infections.

To date, no human studies have tested targeted therapy for trained immunity in organ transplant recipients. However, current antirejection therapies can already modulate trained immune responses. mTOR inhibitors are well recognized suppressors of trained immunity [27], and similarly corticosteroids and the calcineurin inhibitor tacrolimus have been shown to potently suppress trained immunity responses [13]. Fanucchi et al. demonstrated that tacrolimus could prevent training by blocking NFAT-mediated transcription of lncRNAs required to guide epigenetic reprogramming [26].

3. Innate immune memory responses triggered by non-self recognition

While trained immunity describes nonspecific innate immune memory triggered by danger or pathogen signals, growing evidence indicates that innate immune cells can also mount allogeneic memory, a specific response to non-self, that contributes to graft rejection [10–12, 69].

Oberbarnscheidt et al. provided strong support for this concept. In B6-Rag^−/−Il2rg^−/− mice, which lack B cells, T cells, and innate lymphoid cells, they observed greater infiltration of DCs, including mono-DCs, in allogeneic grafts compared to syngeneic ones. Importantly, mono-DCs from allogeneic grafts retained a mature phenotype and could induce both T cell proliferation and IFN-γ production ex vivo. In contrast, mono-DCs from syngeneic grafts showed only transient maturation, supporting T cell proliferation but failed to induce IFN-γ production [69].

Supporting these findings, Dai et al. demonstrated that pre-immunization with allogeneic splenocytes in the same mouse model led to increased mono-DC infiltration in grafts from the same donor. This response was donor-specific and not observed with syngeneic or unrelated allogeneic grafts [12]. These findings reveal that monocytes can exhibit donor-specific memory, a functional hallmark that was long thought to be exclusive to the adaptive immune system.

Further evidence comes from observations that macrophages developed cytotoxic activity against allogeneic cells only after prior immunization with splenocytes from the same donor. Unlike mono-DCs, macrophages required both immunization and CD40–CD40L signaling during the priming phase to acquire this cytotoxic function [12, 70].

3.1. Signaling pathways underlying innate immune allorecognition

Although early evidence suggested that innate immune cells could participate in allorecognition, the mechanisms by which they recognize non-self remained elusive. More recent studies have identified specific receptors that mediate both innate allorecognition and antigen-specific memory.

SIRPα/CD47 signaling

The first such mechanism discovered was MHC-independent and involved polymorphisms in Signal regulatory protein α (SIRPα) [11]. SIRPα, a member of the immunoglobulin superfamily expressed on myeloid and parenchymal cells, interacts with the widely expressed, non-polymorphic ligand CD47 [71]. CD47 typically delivers activating signals to monocytes, but this is normally counterbalanced by inhibitory SIRPα signaling, preventing autoimmunity [Figure 2B]. However, in B6-Rag^−/−Il2rg^−/− mice, transplantation of donor cells expressing a SIRPα variant with higher affinity for host CD47 than the host’s own SIRPα has for donor CD47 results in net monocyte activation and enhanced mono-DC infiltration into the graft. This response is abolished by CD47 blockade or in CD47-deficient hosts, and is significantly reduced when using donor splenocytes from non-obese diabetic-resistant (NOR) mice, which, like B6 mice, express a low-affinity SIRPα variant [11].

More recently, Zhao et al. showed that introducing a SIRPα mismatch into a MHC-mismatched kidney transplant model worsens chronic rejection–like histological changes [18]. This reflects a delicate balance between activating and inhibitory signals in distinguishing self from non-self, resembling NK cell allorecognition. However, unlike NK cells which rely on different receptors, the SIRPα–CD47 pathway uses the same receptor-ligand pair to transmit either activating or inhibitory signals, depending on the affinity of the interaction. In addition to initiating primary monocyte responses upon allorecognition, SIRPα polymorphism is also required for the development of monocyte alloreactive memory [12].

Interest in the role of SIRPα mismatches is growing in the human transplantation field. In a study of 350 patients with acute myeloid leukemia or myelodysplastic syndrome undergoing HLA-matched related hematopoietic stem cell transplantation, Saliba et al. found that SIRPα mismatches occurred in 39% of donor-recipient pairs and were associated with higher rates of graft-versus-host disease and lower relapse rates [72]. Building on this, Zhao et al. performed a large study involving 455 kidney transplant recipients and found that SIRPα mismatches correlated with increased acute rejection, greater fibrosis at one year, and reduced allograft survival, independently of HLA mismatch status [18]. Consistent with findings in mice, adverse outcomes were observed only when the donor expressed a SIRPα variant with higher affinity for CD47 than that of the host. These studies indicate that SIRPα compatibility may serve as a novel biomarker for risk stratification and matching in transplantation.

PIR-A-MHC signaling

The second class of receptors implicated in innate allorecognition are the paired immunoglobulin-like receptors A (PIR-As) [12]. These polymorphic members of the immunoglobulin superfamily, orthologous to human leukocyte Ig-like receptors A (LILR-As), are expressed in a variegated manner on monocytes/macrophages, mast cells, B cells, and neutrophils. PIR-As recognize MHC class I molecules and deliver activating signals upon engagement, with different PIR-A isoforms binding to distinct MHC-I alleles [12, 73].

Upon allostimulation combined with CD40 signaling, PIR-A expression is upregulated in a clonal manner, driven by activation of the transcription factor Nuclear factor kappa B subunit 2 (NF-κB2) [Figure 2B], which promotes PIR-A transcription [73]. In lymphocyte-deficient mice, genetic deletion or functional blockade of PIR-A isoforms impaired monocyte and macrophage memory responses to grafts expressing the corresponding MHC-I alleles, while leaving their primary effector responses intact [12]. These findings suggest that monocyte and macrophage allospecific memory arises from CD40-assisted clonal expansion of PIR-A isoforms following recognition of specific non-self MHC-I molecules.

In a mouse heart transplant model of chronic rejection, conditional deletion of Pira in recipient macrophages or CD40–CD40L blockade reduced PIR-A expression, attenuated chronic rejection, and prolonged graft survival [73]. These findings identify PIR-A and CD40–CD40L signaling as key mediators of innate allorecognition in chronic rejection and highlight them as potential therapeutic targets in transplantation.

3.2. NK cell allorecognition induces innate immune memory

In addition to monocytes, natural killer (NK) cells have been shown to contribute to allorecognition and graft rejection, even in the absence of adaptive lymphocytes. Notably, NK cells can mediate allograft rejection in lymphocyte-deficient mice [70]. The foundational concept of NK cell allorecognition is based on the “missing-self” hypothesis, first proposed in 1986 in the context of tumor immunology [74]. NK cell activity is regulated by a balance between inhibitory and activating signals. Inhibitory receptors recognize self MHC class I molecules. When these are absent or altered, as in allogeneic tissues, the balance shifts toward activation, triggering an immune response.

Both experimental and clinical studies have linked NK cell activation due to missing self-MHC I with graft rejection and failure [75–77]. However, the role of NK cells in alloantigen-specific memory remains less well understood. A study by Nabekura and Lanier provided an interesting insight by demonstrating that NK cells can develop memory responses specific to the MHC I molecule H-2Dᵈ, mediated by the activating Lymphocyte antigen 49D (Ly49D) receptor, but only in the context of inflammation [78]. This response was suppressed in NK cells co-expressing the inhibitory Lymphocyte antigen 49A (Ly49A) receptor, which also binds H-2D^d. These findings suggest that alloreactive NK cell memory is shaped by the net signaling balance between activating and inhibitory Ly49 receptors, reinforcing the importance of receptor co-expression patterns in modulating NK cell responses to alloantigens.

4. Conclusion and future perspective

Experimental and clinical studies indicate that innate immune memory responses can be activated during the transplantation process and can affect graft outcomes. Both the trained immunity profile induced by a patient’s plasma and donor-recipient SIRPα compatibility are indicative for graft survival and may therefore serve as a biomarker for risk prediction or matching [13, 18]. In addition, targeting innate immune memory offers compelling therapeutic potential. Current treatments are aimed at inhibiting T cell proliferation but thereby also block Treg cell formation. In contrast, modulation of innate immune memory as induction therapy can steer the immune response towards acceptance, with an increase in Treg cells and improved long-term graft survival [14].

Despite advances in our understanding of innate immune memory in organ transplantation, key questions remain to fully harness its predictive and therapeutic potential. Although the triggers and molecular mechanisms of trained immunity are well characterized [21], its impact on key transplant-related functions, such as alloreactive T cell activation and cytotoxicity, remains less clear. In contrast, innate allogeneic memory was identified through enhanced effector functions following innate allorecognition [69], while its triggers, SIRPα and PIR-A polymorphism, have only recently been defined [11, 12]. This poses the question whether these molecules are only part of a broader set of factors driving innate allogeneic memory and potential targets for future matching. Furthermore, the transcriptomic, epigenetic and metabolic machinery behind innate allogeneic memory is still largely unknown. Although both trained immunity and innate allogeneic memory involve NF-κB signaling, trained immunity enhances responses to heterologous stimuli, whereas innate allogeneic memory selectively augments responses to alloantigen re-exposure. Clarifying the underlying molecular mechanisms will help determine whether these represent distinct programs or context-specific variants of shared pathways, and may inform multifaceted or selective therapeutic strategies.

Another important topic requiring further exploration is the relationship between innate immune memory and chronic allograft rejection and fibrosis. These are detrimental processes that affect long-term graft survival and for which treatment options are limited. Both transient triggers that reprogram the innate immune system and sustained activation through innate allorecognition may drive chronic inflammation within the graft. Clinical studies are needed to gain further insight into the role of innate immune memory in the development of chronic rejection and allograft fibrosis.

Overall, the concept of innate immune memory offers a novel perspective for understanding the immunological processes that determine graft survival and opens new avenues for the development of therapeutic strategies. Future experimental and clinical studies will be required to determine how these insights can be leveraged to improve transplant outcomes for patients.

Highlights.

• Experimental and clinical studies indicate that innate immune memory responses are related to graft outcomes.

• Many factors in organ transplantation are known to induce trained immunity, whereas only a limited number of molecular triggers of innate allogeneic memory are currently known.

• Trained immunity and innate allogeneic memory may alter interactions between innate and adaptive immune cells in organ transplantation.

• Further studies are needed to translate these insights into strategies to improve transplant outcomes.

Non-standard Abbreviations and Acronyms:

LPS

Lipopolysaccharide

H3K27ac

Histone 3 lysine 27 acetylation

H3K4me1

Histone 3 lysine 4 monomethylation

H3K4me3

Histone 3 lysine 4 trimethylation

H3K18la

Histone 3 lysine 18 lactylation

lncRNAs

Long non-coding RNAs

WDR5/MLL1

WD Repeat-Containing Protein 5/Mixed Lineage Leukemia-1

PAMPs

Pathogen associated molecular patterns

MyD88

Myeloid differentiation primary response 88

RAG

Recombinase-Activating Gene

SCFA

Short-chain fatty acid

DAMPs

Danger-associated molecular patterns

HMGB1

High Mobility Group Box 1

IRI

Ischemia-reperfusion injury

SYK

Spleen tyrosine kinase

ICAM-1

Intercellular Adhesion Molecule-1

NF-κB

Nuclear factor kappa-B

PBMCs

Peripheral blood mononuclear cells

mTOR

Mammalian target of rapamycin

mTORi-HDL

inhibitor-loaded HDL nanobiologics

Treg

T Regulatory

TRAF6

Tumor Necrosis Factor Receptor-Associated Factor 6

CTLA4-Ig

Cytotoxic T-lymphocyte–associated protein 4–immunoglobulin fusion protein

NFAT

Nuclear factor of activated T cells

DCs

Dendritic cells

mono-DCs

Monocyte-derived dendritic cells

SIRPα

Signal regulatory protein α

NOR

Non-obese diabetic-resistant

PIR-As

Paired immunoglobulin-like receptors A

LILR-As

Leukocyte Ig-like receptors A

NF-κB2

Nuclear factor kappa B subunit 2

Ly49D

Lymphocyte antigen 49D

Ly49A

Lymphocyte antigen 49A

NIAID

National Institute of Allergy and Infectious Diseases