Although there were sporadic reports of tissue transplants in ancient times, skin grafting did not become an accepted practice until the late 1800s. Even then, however, many workers did not distinguish between autografts (where the donor and recipient are the same individual) and allografts (where the donor and recipient are of the same species), or even xenografts (where donor and recipient are of different species). The last of these formed the basis for an extensive practice known as zoografting in which patients were subjected to grafts from animals ranging from pigs to frogs. According to Billingham, no one apparently cared whether the grafted skin “took” or merely promoted healing of the wound.¹ The results of these efforts led to a period of confusion in transplantation. Without any clear understanding of the processes involved, surgeons embarked on all sorts of transplants, and a series of operations were reported that we know, from our present understanding of the laws of transplantation, could not possibly have been successful. The transplantation of internal organs awaited the development of techniques for vascular anastomosis. In 1908, Carrel, one of the pioneers of vascular surgery, reported the results of en bloc allotransplantation of both kidneys in a series of nine cats.² He was able to obtain up to 25 days of urine output in some cats, but ultimately all of them died. The first successful clinical renal transplant was not performed until 1952 in Boston, using the kidney of an identical twin.³

During this same period, the closely related field of tumor transplantation gained momentum. Although many of the experiments with tumor transplants provided information essential to an understanding of transplantation immunology, this was not clear at the time. Many workers were committed to the idea that they were studying an effect peculiar to tumor tissues. In his Harvey lecture, Medawar summed up the confusion neatly by the statement, “Nearly everyone who supposed that he was using transplantation to study tumors was in fact using tumors to study transplantation.” Medawar was largely responsible for establishing the immune basis of transplant reactions. Following a series of grafting experiments in rabbits and mice, he concluded in 1945 that rejection of skin “belongs to the general category of actively acquired immune reactions.”⁴

As discussed previously, the genetic analysis of graft rejection indicated that the antigens encoded within the MHC are of particular importance in graft rejection. Table 46.4 summarizes important aspects of the MHC antigens that are especially relevant for transplantation, while a much more detailed description of their structure and function can be found in Chapter 21.

While initially defined by their ability to cause rapid graft rejection, MHC antigens are defined in part by the location of the genes encoding them and in part by the well-characterized structure of both class I and class II antigens (see Chapter 21). MiHAs, on the other hand, are those capable of eliciting a T-cell immune response, but which lack the structural characteristics of MHC products.⁶⁵ Rather than being allelic cell surface proteins, MiHAs are donor-specific peptides presented by MHC molecules that are shared by donor and recipient.^{66,67,68,69,70,71,72,73} As individuals are tolerant to the peptides derived from their own proteins, they only respond to the peptides of another individual’s proteins that have allelic variation. Unlike MHC antigens, MiHAs do not readily stimulate primary in vitro cell-mediated responses in mixed lymphocyte response and cell-mediated lympholysis assays, reflecting the low frequency of T cells recognizing them in the unprimed T-cell repertoire.

It has been estimated that there may be as many as 720 minor histocompatibility loci in mice,⁷⁴ some of which are autosomal and others of which are encoded on the Y chromosome. MiHAs can be expressed ubiquitously or in a tissue-selective manner.⁷⁵ Many proteins producing MiHAs have been identified,^{71,72,73,76,77,78,79,80,81,82,83,84,85,86,87,88,89} some of which are intracellular proteins such as nuclear transcription factors and myosin, while others, like CD31 and CD19, are polymorphic cell surface glycoproteins.^90,91,92

Some MiHAs are diallelic peptides, both of which can be represented by a particular MHC molecule,^76,78 resulting in bidirectional recognition (eg, the murine H13 locus⁷⁸). Alternatively, allelic variation in MHC-binding capacity of a peptide can result in one allele being presented and the other not (eg, the human HA-1 minor antigen, in which only one of two allelic peptides binds effectively to HLA-A2⁷⁶). Minor antigenic determinants can also result from the failure of one allele to be processed to a peptide. An example is HA-8, for which only one allele is effectively transported by the TAP complex, resulting in a null allele despite the presence of the MHC-binding peptide sequence in the molecule.⁷⁹

Both helper determinants, recognized by CD4+ cells, and cytotoxic determinants, recognized by CD8+ cells, are required for effective cytotoxic T-cell responses to MiHAs.^93,94 When multiple MiHA disparities exist, a phenomenon known as “immunodominance” may occur.^{95,96,97,98,99,100} Removal of the immunodominant recognition can reveal strong responses to antigens that evoked weak or no responses before. This phenomenon may be due to competition between peptides for presentation by MHC molecules,⁷⁴ as well as differing durations of antigen presentation and TCR avidities.¹⁰¹ An exceptional peptide, H60, is derived from an NKG2D-binding protein and produces responses that are comparable in potency to those elicited by MHC alloantigens, apparently due to the existence of a very high frequency of TCR in the naïve repertoire recognizing this peptide,¹⁰² which may be a useful target for graft-versus-leukemia (GVL) effects in HCT.¹⁰³ However, immunodominance of CTL responses measured in vitro does not necessarily reflect the immunodominance of the same antigens in vivo.^74,104 possibly reflecting the importance of tissue distribution of minor antigen expression or of helper T-cell responses.⁸⁴

The H-Y antigens are encoded on the Y chromosome and therefore expressed only by males.^{81,82,105,106,107,108} They are of biologic significance, as they cause rejection of male skin grafts in syngeneic female mice and, in humans, female-to-male HCT is associated with increased GVHD rates compared to other combinations in the HLA-identical-related donor setting.

Human MiHAs have been identified as determinants recognized by CTLs, mainly in the setting of HLA-identical sibling donor HCT, in association with GVHD and marrow graft rejection.^{77,79,84,85,109,110} Certain MiHA incompatibilities (eg, HA-1) may predispose to GVHD.^77,111 Immunodominance of CTL responses to particular H-Y and HA determinants as well as expansion of minor antigen-specific CTLs detected with tetrameric complexes of HLA molecules and minor antigenic peptides^85,110,112 have been associated with GVHD and marrow graft rejection.¹¹³

Most MiHAs identified to date are determinants recognized by CTL, reflecting the relative ease with which CTL assays can be used to measure peptide-specific responses. Newer techniques have allowed the recent identification of minor antigenic peptide epitopes recognized by CD4+ T cells.^114,115,116

It is difficult to detect humoral responses to MiHAs, presumably because individual peptide/MHC complexes are too low in abundance on the cell surface to either stimulate an antibody response or be detected on the cell surface with antibodies. An exception is the recent discovery of antibodies against H-Y (Y chromosome-encoded) antigens in association with chronic GVHD (cGVHD).

HAR occurs within minutes to hours after blood flow is established to a transplanted vascularized organ.^143,144,145 The phenomenon is visible and dramatic: the organ turns blue and its function ceases. Microscopically, there is extensive evidence of vascular thrombosis and hemorrhage. The important components involved in the mechanism of HAR include 1) donor endothelial antigens, 2) preformed antibodies that can bind these antigens, 3) complement, and 4) coagulation cascades, which are activated by the binding of preformed antibodies to the vascular endothelium. The interaction of these components leading to hyperacute rejection is diagrammed in Figure 46.8.

The role of complement in HAR is inferred both from the accumulation of various complement components in the grafts and from the fact that complement depletion leads to prolonged survival of xenografts.¹⁴⁶ Complement activation leads to production of active protein fragments and complexes of complement components, which cause tissue injury either directly or by recruiting effector cells that mediate destruction of the graft. In allogeneic combinations, this is initiated by antibody-mediated activation of complement through the classical pathway, whereas in xenogeneic combinations, the alternative pathway may also be involved.¹⁴⁷ In both cases, the membrane attack complex, produced by the ordered interaction of several complement components, initiates the destructive pathway.

Complement activation is controlled by several regulatory molecules, including complement receptor 1, decay accelerating factor (DAF; CD55), membrane cofactor protein (CD46), and CD59, which act at different stages along the cascade (see Chapter 36). Many of these molecules are produced by the vascular endothelial cells. Because these regulatory proteins prevent unwanted complement activation in the face of low levels of perturbation to the system, the titer and avidity of the preformed antibodies must be high enough to activate despite these downregulating molecules. Thus, preformed antibodies directed at MHC antigens almost always accomplish this activation, whereas the lower-affinity blood group antibodies lead to hyperacute rejection in only about 25% of kidneys. One of the reasons that hyperacute rejection is such an important feature in xenogeneic transplantation is that the complement regulatory proteins produced by the donor vascular endothelium of one species often do not function effectively with complement molecules derived from a different species. Because of this incompatibility, lower levels of an initial triggering signal lead to explosive complement activation.

Although the membrane attack complex is often thought of as a lytic molecule, its effect on the donor vascular endothelium, even before cell lysis, is to cause endothelial activation. This occurs rapidly, before there is time for new gene transcription or protein synthesis, and has been referred to as type I endothelial activation. The two principal manifestations of this activation are cell retraction, leading to gaps between endothelial cells, and initiation of coagulation pathways due to the loss of antithrombotic molecules from the endothelium.¹⁴⁸ Thus, type I endothelial activation is responsible for the two principal pathologic findings in hyperacute rejection: extravascular hemorrhage and edema, and intravascular thrombosis.

There are no known treatments that can stop the process of hyperacute rejection once it has started and, thus, it is essential to avoid the circumstances that initiate it. Experimentally, this can be accomplished for relatively short periods of time by administration of complement inhibitors, such as cobra venom factor, which depletes complement. In clinical practice, this is accomplished by avoiding transplantation in the face of preformed antibodies, both by avoiding blood-group antigen disparities and by testing recipients before transplantation (ie, a “cross-match”) to determine whether they have performed antidonor antibodies.

Not all organs and tissues are equally susceptible to hyperacute rejection. Most primarily vascularized organs, such as kidneys and hearts, are very susceptible, but the liver can often survive without hyperacute rejection despite preexisting antidonor antibodies.¹⁴⁹ It is not clear whether this unusual feature of the liver reflects the large surface area of its vascular endothelium or an intrinsic property of liver endothelial cells. Nonetheless, hyperacute rejection of the liver has occurred in some cases, especially involving xenogeneic transplantation, indicating that its resistance to hyperacute rejection is not absolute. Skin grafts are relatively resistant to hyperacute rejection but high levels of antibody can cause a “white graft” (ie, a failure of blood vessels to communicate with those of the recipient)¹⁵⁰ Pancreatic islets are likewise resistant to this form of rejection.¹⁵¹ Free cellular transplants, such as bone marrow cells or hepatocytes, that express some of the antigens recognized by preformed antibodies, are cleared quickly from the circulation by the reticular endothelial system, leading to resistance to engraftment.¹⁵² In the case of HCT, this resistance can be overcome by transplanting larger numbers of cells.^152,153,154 Additionally, antibody-independent complement activation has been shown to be a significant factor diminishing the engraftment of porcine bone marrow in mice.¹⁵⁵

A second mechanism of rejection, also caused by antibodies, occurs as a result of antibodies that are induced very rapidly after a transplant is performed. This type of rejection has been called “acute humoral” or “accelerated” rejection because it typically occurs within the first 5 days after transplant. The process is characterized by fibrinoid necrosis of donor arterioles with intravascular thrombosis.¹⁵⁶

Accelerated rejection is rare in allogeneic combinations because it requires that an antibody response occur before the T-cell response that is typically responsible for acute rejection episodes (see following discussion). Indeed, most allogeneic B-cell responses are T-cell-dependent. The best examples of accelerated rejection are probably those observed in vascularized organ transplants between closely related, concordant xenogeneic species and between discordant species following adsorption of anti-Gal antibodies. In these cases, the levels of preformed antibodies are not sufficient to cause hyperacute rejection, but antidonor antibodies appear rapidly (within 3 to 4 days) in association with the onset of rejection. The pathology of acute humoral rejection reveals a paucity of lymphocytes infiltrating the donor graft, antibody binding to donor vascular endothelium, and fibrinoid necrosis of the donor vessels. Vigorous anti-T-cell immunosuppression has little effect on acute humoral rejection, whereas immunosuppression with reagents that affect B-cell responses, such as cyclophosphamide, delays its onset until more typical T-cell-mediated rejection occurs.¹⁵⁷

As in hyperacute rejection, the process of acute humoral rejection is usually initiated by antibody binding to antigens on the donor vascular endothelium. In this case, however, the subsequent endothelial changes occur more slowly, allowing time for gene transcription and new protein synthesis. This later form of activation has been called type II endothelial activation.^147,158 Many of its features appear to be mediated by the transcription factor NF-κB, which generates many of the responses associated with inflammation, including the secretion of inflammatory cytokines such as interleukin (IL)-1 and IL-8, and the expression of adhesion molecules such as E-selectin and intercellular adhesion molecule (ICAM)-1.¹⁵⁹ In addition, type II endothelial activation causes the loss of thrombomodulin and other prothrombotic changes.¹⁶⁰ Thus, the events following type II endothelial activation are associated with the pathologic changes that occur with “accelerated” rejection, including the tendency toward intravascular thrombosis and the inflammatory destruction of donor vessels that occurs in the absence of infiltrating lymphocytes.

Just as there are regulatory processes for complement activation, there are regulatory molecules that counter the tendency toward intravascular coagulation and the process of type II endothelial activation (eg, tissue factor protein inhibitor [expressed by vascular endothelium, which inhibits factor Xa of the clotting cascade] and a number of other protective molecules, including Bcl-x_L, Bcl-2, and A20).^147,161 Although these are often thought of as antiapoptotic molecules, they also tend to inhibit activation mediated by NF-κB. Like the regulatory molecules of complement, some of these regulators may not function across species differences, leading to disorder regulation of the coagulation system.¹⁶²

Although vigorous early antibody responses generate type II endothelial activation and accelerated rejection, later antibody responses often fail to do so. The process that enables transplanted organs to survive in the face of circulating antibodies that can bind endothelial antigens has been called “accommodation.”^147,163 This phenomenon has been observed in some allogeneic and xenogeneic combinations with preformed antibodies, but has so far been disappointingly ineffective in discordant xenotransplants. Resistance to type II endothelial activation has been achieved in vitro by pretreatment with low levels of antiendothelial antibodies that are insufficient to trigger activation.¹⁶⁴ The achievement of accommodation is associated with increased expression of the antiapoptotic genes described previously and with changes in the isotype of the recipient’s antibody responses.^158,165

An important difference between HAR and acute humoral rejection is that there is no known therapy to stop graft destruction by HAR, whereas acute humoral rejection can sometimes be reversed by desensitization.

“Acute cellular rejection,” which is characterized by a mononuclear cell infiltrate in the graft, is the most common type of organ allograft rejection. Acute rejection is most common during the first 3 months after transplant, but may occur at any time, especially if immunosuppressive medication is withdrawn. Acute rejection is T cell-dependent, and its treatment, which is usually successful, includes increased doses of standard immunosuppressive drugs or antilymphocyte antibodies.

The use of newer immunosuppressive drugs and anti-T-cell antibodies has markedly reduced acute rejection rates. For example, the vast majority of kidney transplant recipients never experience an episode of acute rejection. It is now quite rare to lose a transplanted organ to cellmediated rejection during the first year after transplantation. However, the use of these highly effective immunosuppressive treatments is associated with significant morbidity.

Experimental models for acute rejection include nonprimarily vascularized skin grafts, heart graft fragments, artificial “sponge” allografts, or islet transplants in rodents, which may not accurately reflect the processes of rejection for primarily vascularized organs. While there are models of heart, kidney, liver, and other types of primarily vascularized organ transplants in rodents, these types of transplants are more tolerogenic and hence more easily accepted than similar transplants in large animals and humans. Studies of primarily vascularized organ transplants in large animals, such as monkeys or pigs, have obvious clinical relevance, but are expensive and require many special resources.

Acute GVHD is the counterpart of cellular rejection that involves graft-versus-host alloreactivity, usually in the context of HCT, but also sometimes with organ transplants that carry significant amounts of donor lymphoid tissue (eg, liver). Like acute rejection, acute GVHD is T cell-dependent. T-cell depletion of the donor hematopoietic cell graft prevents GVHD but is associated with increased rates of graft rejection and relapse of malignant diseases.

The concepts of “direct” and “indirect” allorecognition introduced previously must be considered at both the sensitization and effector phases of an immune response. The definition of “indirect” recognition used in this chapter is based on which set of APCs (donor versus recipient) is presenting donor antigen. “Cross-priming” is a term specifically denoting sensitization of CD8 T cells through the indirect pathway. As shown in Figure 46.9, there are three major T-cell pathways to consider in relation to graft rejection. These include 1) direct recognition of donor alloantigens by CD4+ T cells, which generate effector CD4 cells and provide help for the generation of effector CD8 cells; 2) direct activation of CD8 T cells by donor APCs; and 3) CD4+ T-cell activation by recipient APCs presenting re-processed donor antigens (the indirect pathway of sensitization). This pathway is important in providing help for immunoglobulin production by B cells. A role in rejection for cross-primed CD8+ T cells is not included in Figure 46.9, because this pathway has not been shown to play a role in allograft rejection except under certain circumstances, as described in the following. The multiplicity of T-cell sensitization and effector pathways involved in graft rejection and GVHD is demonstrated by the frequent observation of rejection or GVHD when only CD4+ or CD8+ T cells are depleted.^166,167,168 As a result of the high precursor frequency of T cells that respond to allogeneic MHC antigens directly, populations of T cells that ordinarily have minimal significance become functionally important.

Most experimental studies of rejection are performed without immunosuppression and, therefore, graft destruction usually occurs within the first several days or weeks by one of the mechanisms described previously. In clinical practice, however, the use of immunosuppression usually allows graft survival for much longer periods of time. Nonetheless, clinical survival statistics reveal that even when 1-year graft survival has been achieved, the loss of transplanted organs continues to occur at a rate of about 3% to 5% per year, and a significant proportion of this delayed or late graft failure appears to be due to immunologic mechanisms.

The term “chronic rejection” is commonly used to describe this later process of delayed graft destruction, although in kidney transplantation the Banff classification schema has proposed to replace this term with interstitial fibrosis and tubular atrophy.⁴⁸⁴ As immunosuppressive reagents have become more effective at controlling acute rejection, chronic rejection has emerged as one of the most important problems in clinical practice. Indeed, while there has been ongoing improvement over the past 30 years in the 1-year graft survival rates for kidney transplants, the halflife for organs that have survived for 1 year has not changed significantly over that entire period of time; as a result of this ongoing loss, only about 50% of transplants are still functioning 10 years later.

Although almost every type of organ transplant suffers from deterioration in function over time, the pathologic manifestations are different in each case. Kidney biopsies tend to show interstitial fibrosis along with arterial narrowing from hyalinization of the vessels—hence the terminology interstitial fibrosis and tubular atrophy. In the heart, the process is manifested principally as a diffuse myointimal hyperplasia, proceeding to fibrosis of the coronary arteries that has often been referred to as “accelerated atherosclerosis” or “transplant arteriosclerosis.” Chronic rejection in lung transplants primarily affects the bronchioles with progressive narrowing of these structures and is referred to as “bronchiolitis obliterans.” The liver may be the one type of organ transplant that is relatively resistant to chronic rejection, but the progressive destruction of bile ducts referred to as the “vanishing bile duct syndrome” may be another manifestation of this process.

Some of the causes of chronic graft destruction may not be immunologic in origin.^485,486 Analysis of sequential kidney transplant biopsies suggests that chronic rejection represents cumulative and incremental damage to the graft from time-dependent nonimmunologic and immunologic causes.^487,488 Potential nonimmunologic factors that have been considered to contribute to the development of chronic rejection include the initial ischemic insult, the reduced mass of transplanted tissue (especially in the case of kidney transplants leading to hyperfiltration injury), the denervation of the transplanted organ, the hyperlipidemia and hypertension associated with immunosuppressive drugs, the immunosuppressive drugs themselves, and chronic viral injury, amongst others. Nonetheless, while these factors undoubtedly contribute to the process, there is a marked difference in survival between syngeneic and allogeneic transplants in experimental models. Thus, there is almost certainly an important immunologic component in most cases of chronic rejection.

Several important observations regarding the pathogenesis of chronic rejection have emerged from clinical practice, particularly the analysis of biopsy samples. In kidney transplants, two distinctive phases of injury of chronic allograft nephropathy have been described.⁴⁸⁷ Previous studies have suggested that there is a high correlation between the onset of chronic rejection and a history of early acute and subclincial rejection episodes.^489,490 Analysis of protocol biopsies has revealed that the onset of mild chronic rejection by 1 year after kidney transplantation is associated with an initial phase of early tubulointerstitial damage from ischemic injury that occurs before severe rejection is detected. Beyond 1 year, a later phase of chronic allograft nephropathy was characterized by microvascular and glomerular injury.⁴⁸⁷ Importantly for long-term outcomes, the clinical data show that the process of chronic rejection is usually refractory to increases in immunosuppressive therapy, in contrast to acute rejection episodes that almost always respond to treatment. The development of chronic rejection has also frequently been associated with the presence of antidonor antibodies,^491,492 and the deposition of complement component C4d in the allograft.

Taken together, these clinical observations have suggested to some that chronic rejection is the result of donorspecific alloantibody production.⁴⁹³ Moreover, there are also now data emerging to suggest that components of the innate immune system such as NK cells can also contribute,⁴⁹⁴ in some cases in conjunction with alloantibody.⁴⁹⁵ All of these suggestions may be correct, but neither the logic nor the evidence fully supports the exclusive involvement of one mechanism. The relationship between alloantibody, C4d deposition, and neointimal fibrosis is complex.⁴⁹⁶ In the first place, alloantibody production often reflects activation of indirect pathway T cells (see previous discussion), and hence it might equally well be a marker for other rejection mechanisms as opposed to a cause of chronic rejection. Moreover, the presence of alloantibody and C4d may be transient, and there are clear examples of chronic changes in the graft in their absence.⁴⁹⁶ In addition, early rejection episodes probably reflect primarily the degree of antidonor immunoreactivity, and there is no proven direct link with later deterioration of graft function. Therefore, even if sufficient
immunosuppression were given to prevent acute rejection, chronic rejection might still occur when the levels of immunosuppression are reduced over the long term, even if acute rejection had never occurred. Finally, experimental studies have suggested that the mechanisms of chronic rejection are not absolutely dependent on either antibody formation or on the occurrence of acute rejection episodes.

The uncertainties that arise from the interpretation of the clinical data make it important to develop experimental models for studying the mechanisms of chronic rejection. It is difficult in the laboratory, however, to mimic a process that may take 5 or 10 years to develop in patients treated with immunosuppressive drugs. Thus, the effort to study chronic rejection experimentally has depended on surrogate short-term pathologic markers that are thought to predict the long-term changes of chronic rejection. In particular, these studies have concentrated on the development of the myointimal proliferation that is thought to be the precursor of the chronic vascular changes typically observed in patients. In rodents, pigs, and primates, this has often been done with grafts after an initial period of immunosuppression that prevents acute rejection.^497,498,499 All of these experimental studies are subject to the caveat that the surrogate pathologic lesion occurs much earlier than the typical changes of chronic rejection in patients. Thus, the process being studied experimentally may not be the same as the clinical process.

Interactions between Fas and Fas ligand (FasL), which is upregulated during rejection responses, can mitigate GVHD and graft rejection by killing activated T cells and APCs.³⁰⁶ FasL-deficient recipients are more susceptible than normal mice to the development of GVHD,⁵²³ and FasL can promote resistance to rejection of tissues transplanted to some “privileged sites,” such as the testis or the anterior chamber of the eye.⁵²⁴ However, forced overexpression of FasL causes a nonspecific inflammatory syndrome associated with prominent neutrophil infiltration⁵²⁵ and can promote graft destruction.^526,527 Overexpression of FasL has, however, been reported to promote survival of heart allografts in recipients of FasL-expressing donor-specific transfusions (DSTs),⁵²⁸ and of bone marrow⁵²⁹ and islet⁵³⁰ allografts.

CTL antigen (CTLA)4 helps maintain self-tolerance, as evidenced by the T-cell lymphoproliferative autoimmune syndrome that develops in CTLA4 knockout mice.^531,532 CTLA4 has been shown to play a role in T-cell tolerance in many systems.^{384,533,534,535,536,537,538,539,540,541,542,543,544,545,546} Blockade of CTLA4 accelerates cardiac allograft rejection⁵⁴⁷ and increases GVHD.⁵⁴⁸ PD1, an additional downregulatory molecule expressed by activated T cells that recognizes the B7 family members PDL-1 and PDL-2, also mitigates graft rejection^{549,550,551,552,553,554} and GVHD.^555,556,557 Interaction of one of the PD1 ligands, PD-L1, with its alternative receptor B7.1, also mitigates graft rejection.⁵⁵⁸ While PD1 plays an important modulatory role in the presence of extensive antigenic disparities between murine heart graft donors and recipients, the B- and T-lymphocyte attenuator (BTLA)-herpesvirus entry mediator (HVEM) inhibitory pathway predominates in the presence of more restricted antigenic disparities.⁵⁵⁹ BTLA-HVEM interactions may also control GVHD,⁵⁶⁰ but the opposing effects of BTLA and HVEM expressed by T cells on T-cell activation complicate interpretation of experiments, which have produced conflicting results with respect to GVHD^560,561 and islet allograft survival.^562,563

There is considerable evidence of roles for IL-10 and TGF-β in downmodulating graft rejection^564,565,566 and GVHD,^{372,567,568,569,570} and these cytokines may have therapeutic utility.^{571,572,573,574,575} These
activities may reflect the important role of these cytokines in regulatory T (T_reg)-cell and B-cell generation and function. On the other hand, IL-10 can enhance cytolytic mechanisms of islet graft rejection,⁵⁷⁶ and high doses can accelerate GVHD.³⁷² Moreover, TGF-β is an important mediator of fibrotic pathologies in both chronic rejection and GVHD.^577,578 TGF-β, in concert with proinflammatory cytokines such as IL-1 and IL-6, also promotes the development of proinflammatory Th-17 cells.⁵⁷⁹

The inflammatory condition that develops in IL-2 knockout mice is clear evidence of the important immunomodulatory role of this cytokine, and studies in several graft rejection⁵⁸⁰ and GVHD⁵⁸¹ models have confirmed such a role in transplantation. Much of this effect is due to the important role of IL-2 in survival and expansion of T_reg cells,^582,583 as is discussed elsewhere in this chapter. IL-2 can also promote activation-induced cell death of alloreactive CD8 cells.^582,584

While IFNγ can also promote graft rejection, rejection is rapid or even accelerated in IFNγ knockout mice.^{383,564,585,586,587} IFNγ has also been shown to play a downregulatory role in GVHD.^371,385,588 IFNγ has antiproliferative effects on T cells,^384,587 increases activation-induced cell death via the Fas/FasL pathway,^589,590,591 upregulates nitric oxide production,^{592,593,594,595} and is necessary for T_reg-cell function in certain conditions.^596,597,598 Consistently, the major cytokines that induce IFNγ, IL-12 and IL-18, can inhibit graft rejection^382,384,599 and GVHD.^385,600,601 IL-12 and IL-18 act in an IFNγ and Fas-FasL-dependent manner,^385,601,602 which preserves or enhances GVL effects.^{588,603,604,605,606} IFNγ promotes the GVL effect of CD8 T cells,^385,607,608 apparently by promoting lymphohematopoietic grat-versus-host responses while inhibiting tissue GVHD,⁶⁰⁹ perhaps due in part to increased PDL1 expression by APCs,⁶¹⁰ promotion of donor T_reg-cell expansion,⁶¹⁰ and reduced Th17 differentiation.³³⁷

Despite interest in the notion that Th2 cytokines are antiinflammatory and may suppress rejection and GVHD,^{359,360,361,362,363,364,365,366,367} IL-4 deficiency does not accelerate graft rejection^564,611 and can actually downmodulate GVHD.³⁷¹ However, there clearly are situations in which an immunomodulatory effect is achieved by Th2 cytokines. For example, the use of total lymphoid irradiation can alter the balance between NKT cells and conventional T cells in BMT recipients, and IL-4 production by enriched recipient NKT cells downmodulates GVHD.⁶¹² This approach, which also apparently involves T_reg cell enrichment when combined with antithymocyte serum,⁶¹³ has recently been extended to clinical trials of HLA-identical HCT for treatment of hematologic malignancies and renal allograft tolerance induction.^{614,615,616,617}

Vascularized organ allografts may be accepted spontaneously^{618,619,620,621} or with a short course of immunosuppression in rodents or pigs.^{621,622,623,624,625,626,627,628,629} The long survival of these transplanted organs can prevent rejection of other allografts from the same donor.^618,630,631 In clinical transplantation, long survival of a transplanted organ may diminish the rejection response, as much less immunosuppression is required late after transplantation than in the early period. Studies in transiently chimeric monkeys and patients achieving tolerance with HLA-mismatched combined kidney and BMT strongly implicate a role for the kidney itself in promoting tolerance.^{632,633,634,635,636} T_reg cells have been implicated in many of these models (see following discussion). The capacity of transplanted organs to regulate their own survival is often confused with the capacity of a treatment to induce tolerance. For example, many transient immunosuppressive regimens achieve vascularized allograft survival in rodents, but the role of the immunosuppression is to allow a strong T_reg cell response induced by the graft to predominate under the controlled experimental conditions. These conditions do not usually apply to human transplantation, thus explaining the failure to translate the many tolerance regimens that succeed in rodents into clinical practice.

Graft injury, such as that associated with ischemia-reperfusion and host tissue injury induced by conditioning therapy in the case of HCT, play an important role in promoting graft rejection and GVHD, respectively. Local inflammatory processes activate innate and consequently adaptive immunity and T-cell activation. Inflammation plays an important role in promoting APC migration from tissues to lymph nodes^{637,638,639,640,641} and also promotes trafficking of activated T cells into tissues, as is illustrated dramatically in HCT models. Administration of large numbers of nontolerant donor lymphocytes to established mixed bone marrow chimeras (ie, animals not recently treated with conditioning therapy) leads to a graft-versus-host response that attacks only lymphohematopoietic tissues and does not cause GVHD, a disease of epithelial tissues such as skin, intestines, and liver.^642,643 In contrast, similar numbers of T cells cause rapidly lethal, severe GVHD in freshly irradiated hosts.^642,644 Conditioning rapidly induces production of chemokines in the GVHD target tissues, promoting immigration of T cells that then elicit a further cascade of chemokines that amplifies the response.⁶⁴⁵ Upregulated adhesion molecules also promote leukocyte infiltration through the microvasculature of these tissues. Lethal total body irradiation (TBI) and cyclophosphamide, for example, upregulate the proinflammatory cytokines IL-1, IL-6 and TNF-α,^646,647,648 which can upregulate endothelial cell E-selectin, P-selectin, ICAM-1, and vascular cell adhesion molecule-1.^649,650 In the absence of such host target tissue inflammation, mature, activated graft-versus-host-reactive effector T cells are unable to traffic into skin and induce injury.⁶⁴⁴ Provision of a local tolllike receptor (TLR) stimulus promotes the entry of such cells into the skin and induces localized GVHD,⁶⁴⁴ indicating that tissue inflammation provides a critical checkpoint for T-cell recruitment to GVHD target tissues. A systemic TLR stimulus in this setting promotes severe, multiorgan GVHD.⁶⁴⁴ GVHD can also occur when very large numbers of nontolerant parental T cells are administered to genetically tolerant F1 hosts,⁶⁵¹ indicating that, in the presence of sufficiently powerful graft-versus-host responses, the need for tissue inflammation to induce GVHD can be bypassed, possibly due to inflammation induced by high systemic cytokine levels.

All forms of organ and tissue transplantation involve ischemic and traumatic injury to the donor tissue, which may be one of the reasons that rejection episodes occur most frequently early after transplantation. The surgical trauma associated with transplantation is associated with very early production of chemokines in the graft,^433,652 promoting infiltration of NK cells⁴³³ and neutrophils,⁶⁵³ which in turn perpetuate inflammation that promotes subsequent T-cell infiltration.⁶⁵⁴ At least partly due to the influence of these cells of the innate immune system,⁴³³ chemokines are produced before T-cell infiltration is seen. IFNγ, whose early production may require CD8 T-cell activation,⁶⁵⁵ also activates macrophages to become effective APCs and release chemokines.^655,656,657 These phenomena, along with microbial exposures that drive innate immunity, contribute to “danger” signals that promote graft rejection.⁶⁵⁸ Nevertheless, skin and cardiac allografts that are allowed to heal before being exposed to alloreactive T cells are rapidly rejected if there is sufficient antigenic disparity between donor and host.^659,660 Similarly, patients with long-standing allografts are rarely able to terminate immunosuppressive therapy without rejection. Therefore, “danger” signals are not a critical requirement for graft rejection, and it is better to picture the antigenic disparity and the recipient’s immune responsiveness as the dominant features controlling graft rejection, while danger signals may influence the timing, intensity, or character of the rejection response.

The innate immune system comprises a group of cells and molecules (Table 46.5) that provide a first line of defense against pathogens and which also play an important role in allograft rejection. Primary adaptive immune system responses that rely on the activation and expansion of antigenspecific T and B cells take several days to reach maturity. In contrast, the innate immune system can be considered as a “preformed” defense mechanism that is immediately available to defend the host until either the dangerous stimulus is cleared or the adaptive immune system is able to mount an antigen-specific response.⁶⁶¹ Clearly, this is a somewhat simplistic view; while many components of the innate immune system can be recruited very quickly after transplantation, their activity can be amplified after activation.

The physical process of graft retrieval and implantation generates signals within the graft and the recipient that trigger rejection. The concept of “danger” triggering an immune response has evolved as an idea over many years.^170,661 Pattern recognition receptors exist to detect the unwanted presence of bacterial or viral pathogen-associated molecular patterns, but after transplantation the TLRs that form part of the pattern recognition receptor family can also be used to detect the molecules produced as a result of implantation of the graft, so called damage-associated molecular patterns (DAMPS). These signals include heat shock proteins, reactive oxygen species, high mobility group protein B1, complement breakdown products, nucleic acids (deoxyribonucleic acid and RNA), mitochondrial components, and molecules associated with tissue fibrosis that activate cells of the innate immune system via TLR ligation.

The immune system monitors the health of cells and responds to ones that have been injured and killed. Cell death is an inevitable consequence of the ischemia and reperfusion injury that is caused by organ and tissue retrieval. Dying cells expose intracellular DAMPs that can be recognized by components of the innate immune system.⁶⁶² The role of DAMPs may vary depending on the type of transplant performed and the degree of injury resulting from cell isolation or organ retrieval. Some DAMPs may be expressed in a tissue-restricted manner (eg, one class of DAMPs called alarmins include molecules such as β-defensins that are expressed primarily by leukocytes and therefore may be more relevant after hematopoietic stem cell transplantation or BMT). In addition, the location or type of injury may contribute to the outcome of dying cells triggering a response. Thus, the contribution of different DAMPs to triggering the innate immune system may vary with different types of transplant.

Macrophages and other phagocytic cells can ingest dying cells and necrotic tissue; when activated, they release cytokines such as TNF-α, IL-1, and IL-6, which all contribute to the local inflammatory environment. Production of active IL-1β requires proteolytic cleavage of the inactive form, pro-IL-1, a process linked to a multiprotein complex called an inflammasome that acts as a platform for caspase 1, the cysteine protease involved in the proteolytic maturation and secretion of IL-1β.⁶⁶³ A number of different inflammasomes exist, but the nucleotide oligomerization domain-like receptor protein 3 inflammasome is the best studied to date. IL-1β can be highly deleterious to the tissue in which is produced; therefore, inflammasome activity is tightly controlled. As a consequence of these events, the early infiltration of macrophages into a graft at the onset of rejection has been suggested to be a poor prognostic sign for transplant survival. Macrophage colony-stimulating factor, produced by tubular and mesangial cells, promotes macrophage infiltration and proliferation, and may play a pathogenic role in acute rejection.

Damaged tissue can also trigger complement activation in the absence as well as the presence of antibody; complement has been demonstrated to contribute to ischemia reperfusion
injury.⁶⁶⁴ Activated complement components constitute a proteolytic cascade that generates a range of effector molecules.¹⁷⁰ The anaphylatoxins C5a and C3a are chemoattractant molecules that assist leukocytes to home to the graft while other soluble mediators are able to opsonise cells, targeting them for destruction by phagocytes.⁶⁶⁵ Recognition of C3b, C4b, or their fragments covalently bound to target cells by complement receptors on the surface of leucocytes facilitates antigen presentation and T-and B-cell activation.^666,667 Generation of the terminal components of the complement cascade (C5b-9) results in formation of the membrane attack complex within the target cell membrane and initiation of target cell lysis. This has been demonstrated to play an important role in ischemia reperfusion injury.⁶⁶⁴ In addition to the potential of the damaged tissue itself to activate complement, there is also evidence that natural immunoglobulin M antibody can trigger complement activation via both the classical and mannose-binding lectin pathway.⁶⁶⁸ Studies in muscle reperfusion models initially identified natural immunoglobulin M as a major initiator of pathology through the activation of the complement system and recruitment of inflammatory cells. When the repertoire of natural immunoglobulin M antibodies was altered, significant protection of the myocardial tissue was observed with only limited apoptosis of cardiomyocytes and decreased neutrophil infiltration compared to when natural antibody was present.⁶⁶⁹ As mentioned previously, there is also increasing evidence that complement can influence graft outcomes, contributing to the development of acute and/or chronic rejection, either directly or through antibodydependent mechanisms.^511,670

NK cells are innate immune mediators that express cell surface receptors, including activating receptors that bind to widely expressed carbohydrate residues on self-cells and inhibitory receptors that bind self-MHC class I molecules. Activating NK cell receptors including NKG2D recognize natural stress signal ligands, whereas the inhibitory receptors include the CD94-NKG2A complex, KIR family in humans, and Ly49 family in mice. The possible role of NK cells in graft rejection is discussed previously. Absence of an appropriate MHC class I ligand on an allogeneic cell informs the NK cell that the allogeneic cell should be killed. Some malignant or virally infected cells downregulate MHC class I expression or express altered class I molecules as a strategy to evade CD8+ T-cell cytotoxicity. As a result, they are unable to stimulate inhibitory receptors and are vulnerable to NK cell killing. Thus, NK cells could contribute to tissue damage following cell and solid organ transplantation. While the role of NK cells in rejecting bone marrow (at least in mice) is clearer than for solid organ transplantation,⁶⁷¹ NK cells can have a marked and lasting impact in this setting,⁶⁷² particularly with a form of chronic cardiac graft vasculopathy in mice.⁴⁹⁴ NK cells likely contribute to acute rejection in certain donor-recipient combinations where, even if they are not the major drivers of the responses, they have a significant impact by secreting IFNγ. Nevertheless the precise role of NK cells requires further elucidation as NK cells have also been shown to promote tolerance induction (see the following discussion) by killing donor APCs.⁶⁷³