Immunobiology of Heart and Heart-Lung Transplantation

	INTRODUCTION

	THE MAJOR HISTOCOMPATIBILITY COMPLEX

 
An allogeneic organ is one that is transferred from one individual to another of the same species but with a different genetic repertoire. A donor heart or lung is immunologically incompatible with the host tissues, and an immunologic reaction or alloresponse is directed against donor proteins or antigens located on the surface of the endothelial, mesenchymal, and epithelial cells of the allograft.

The major histocompatibility locus (MHC) is a complex of polymorphic genes whose glycoprotein MHC molecule products are expressed on the surface of cells. The protein products are the principal determinants of whether an organ is deemed self or nonself, and are the primary targets of the immune response to allografts. The MHC, also known as human leukocyte antigens (HLA), guide the development of T lymphocytes to have a low affinity to self and to use the reaction to self as a way in which foreign peptides are recognized (MHC restriction). The genes that express HLA are among the most variable (or have the largest number of polymorphisms) in the human genome.

Immune responses to organs with different HLA gene types define the alloresponse in humans. The HLA complex encodes class I HLA molecules A, B, and C, which present intracellular antigen to stimulate cytotoxic T lymphocytes expressing the cell surface receptor CD8 (Fig. 59-1). In addition, the HLA complex encodes HLA class II molecules DP, DQ, and DR, which are expressed on antigen-presenting cells that bind extracellular, foreign antigen that is recognized by the proinflammatory CD4-positive T lymphocytes.

View larger version (16K): [in this window] [in a new window]   FIGURE 59-1 Class I and II HLA molecules are made up of polypeptide chains with intrachain disulfide bonds. The 1 and 2 distal domains of class I and 1 and ß1 domains of class II make up the peptide binding site for alloantigen. (Adapted with permission from Parham M, in Haber E (ed): Immunobiology of Transplantation Molecular Cardiovascular Medicines. New York, Scientific American Press, 1995.)

In heart and lung transplantation, several single-institution studies examining the effect of HLA matching on outcome have found an association between the degree of serologic HLA-DR matching and actuarial graft survival at 1, 5, and 10 years. In general, an association was not present for HLA-A and HLA-B matching. In fact, in a report from the Texas Heart Institute, Kerman et al reviewed 448 heart transplants⁴ and found an inverse relationship between HLA-A and HLA-B mismatches and death from cardiac allograft vasculopathy.

Most studies draw from a system of random allocation of donor organs, resulting in less than 8% of closely matched donor-recipient pairs (i.e., 0, 1, or 2 mismatches). Given this low frequency, an adequate pool of closely matched pairs for comparison to recipients that are mismatched at multiple loci with the donor is made possible only by a multi-institutional study of significant size. Two cardiac transplantation registries have fulfilled this size requirement and verified the relationship between HLA matching and acute graft survival: the Collaborative Transplant Study,⁵ with 8331 recipients, and the United Network for Organ Sharing/International Society for Heart-Lung Transplantation (UNOS/ISHLT) Registry⁶ with 10,752. In the Collaborative Transplant Study, 128 patients (1.5%) with either 0 or 1 combined HLA-A, HLA-B, or HLA-DR mismatches were compared to those with 2 mismatches and 3 to 6 mismatches. Mean rates of survival at three years were a striking 83%, 76%, and 71%, respectively. Multifactorial regression analysis further established that HLA matching had a strong independent effect on graft survival, with the most pronounced effect at 6 months. While the timing might suggest a predominant role for acute rejection, only graft survival and not rejection rates were reported.

The UNOS/ISHLT Registry investigators also found a progressive reduction in risk for greater donor-recipient HLA matching. As opposed to the Collaborative Transplant Study, data obtained from this registry is derived from a database whose use is compulsory for all transplant centers in the United States and subject to auditing and verification. Follow-up in this patient population was found to be virtually complete. The primary benefit of matching appeared to be at the A and DR loci with no independent effect of matching of the B loci. However, these retrospective data, as with the previous study, were based on serologic methods of tissue typing that are less accurate than current recombinant techniques. Again, the effect of HLA matching was greatest at 6 months with the survival curves between matched and mismatched patients becoming parallel at later time points. Considering the results of these two registry studies in light of prior work,⁷ HLA matching is unlikely to influence chronic graft rejection. Larger numbers of well-matched transplants studied at a prolonged follow-up are needed to investigate the influence of HLA matching on chronic rejection.

Data demonstrating the effect of HLA matching on outcomes in heart-lung and lung transplantation are sparse. In one study from the University of Pittsburgh, 74 single- and double-lung transplant recipients were analyzed, and a strong effect of HLA-DR matching on 6-month graft survival was evident (100% vs. 75% vs. 56% for 0, 1, and 2 DR mismatches, respectively).⁸ Combining A, B, and DR mismatches showed 100% survival for 0 to 2 mismatches, 78% for 3 or 4, and 58% for 5 or 6. The Collaborative Transplant Study showed a trend toward improved survival for well-matched grafts in both heart-lung and lung recipients, but this did not reach statistical significance (1176 patients enrolled in the lung transplant group and 640 in the heart-lung group).⁵ The UNOS/ISHLT registry data also showed a less impressive effect of matching on lung allograft outcome. A significant reduction in risk with any degree of HLA matching was seen but no progressive improvement with increasing levels of matching.

These data support the conclusion that HLA matching confers an important benefit after heart transplantation, and probably after heart-lung and lung transplantation. The conventional wisdom that deems prospective HLA matching to be logistically unfeasible in thoracic organ transplantation may be changing. The former requirement of HLA typing using serologic methods for donor splenic tissue retrieved during procurement did not provide sufficient time for prospective typing given that heart and lung allografts tolerate only 4 to 6 hours of ischemic time. This formidable restriction has been overcome by the use of PCR-based HLA typing on peripheral blood lymphocytes prior to the procurement procedure. Future advances in our current preservation methods such as the use of continuous, warm, sanguinous perfusion of the ex vivo cardiac allograft⁹ will permit procurements from longer distances, a certain requirement of an organ allocation system that takes into consideration HLA matching.

Some groups have, in fact, reported impressive advances in achieving prospective HLA matching. One is the Harefield group, which recently reported that within their donor allocation zone, HLA typing was available before organ retrieval in 69% of cases performed in 1994. Based on outcome data from their institution, this group has focused on HLA-DR matching only and has seen an increase in prospective matching from 5% to 25% of transplants in a recent 1-year time period and a reduction in acute rejection in those matched. However, widespread adoption of cardiac HLA matching has been hindered by continued limitations. A benefit on early cardiac allograft survival was seen mainly for those with more than 3 antigen matches, an infrequent event (less than 8% of cases) in the current system of random allocation of donor organs. Increasing this frequency of close matches via prospective matching would require significantly prolonged ischemic and waiting list times for the donor organ and transplant candidate, respectively. Given current methods of organ preservation and the high recipient waiting list mortality, such a requirement would seem unacceptable in light of the modest impact on acute organ survival and lack of evidence supporting an effect on long-term graft outcome and chronic rejection.

	ANTIGEN PROCESSING/AFFERENT RESPONSE

 
HLA class I and class II molecules are not uniformly expressed on the same cells. HLA expression is low at baseline in human donor hearts and lungs but can be found to stain prominently for both classes in response to inflammatory stimuli (Fig. 59-2). After transplantation, class I molecules present protein products produced from the endogenous breakdown of their own MHC protein to previously activated CD8⁺ cytotoxic lymphocytes (CTL) (Fig. 59-3). Class I molecules are found on the surfaces of all cells except the erythrocyte, which incidentally protects a malarial red cell infection from CD8⁺ cell surveillance. Class II expression is constitutively found on the professional antigen-presenting cells (APCs), dendritic cells, B lymphocytes, macrophages, and thymic epithelium, and induced on other cell types by cytokines like interferon (INF), also produced as part of the alloimmune response. Originating either from donor organ (i.e., passenger cells) or from the host, these APC internalize, process, and present shed fragments of donor MHC protein on class II molecules (see Fig. 59-3). The allogenic class II molecule and bound peptide exclusively react with a large number of CD4⁺ T cells (perhaps 1%) in a process called the direct or allorestricted pathway.^10,11

View larger version (26K): [in this window] [in a new window]   FIGURE 59-2 Donor antigen derives from either the endogenous pathway or exogenous path.

View larger version (24K): [in this window] [in a new window]   FIGURE 59-3 Generally HLA class I molecules bind protein fragments of donor MHC protein produced from endogenous cellular processes. Allopeptides from MHC donor cell membrane fragments are brought into special APCs that process and bind them to HLA class II molecules for presentation to host T cells.

View larger version (21K): [in this window] [in a new window]   FIGURE 59-4 Host T lymphocytes can respond to alloantigens by the direct or indirect pathway. On direct presentation, donor cells bind endogenous MHC protein to donor MHC molecules (allorestricted). In the direct path, host cells respond to processed donor MHC peptide bound to host MHC (self-restricted). The direct pathway can stimulate many T cells and is responsible for most acute rejections, whereas fewer T cells respond to the small foreign peptide presented in host MHC molecule. The indirect pathway has been implicated as an important part of the late chronic rejection process when host APCs replace those of the donor within the allograft.

Only mature DC provide the appropriate costimulatory signal for the conversion of naïve CD8 T cells to activated cytotoxic lymphocytes (CTL). In addition, this process typically requires exogenous IL-2 from CD4 "helper" T cells. A more stringent requirement for stimulation of CD8⁺ versus CD4⁺ T cells assures that CTL are formed only when evidence of their need is unambiguous.¹

	EFFERENT ALLORESPONSE

 
After recognizing a heart or lung allograft as foreign, the immune system unleashes cellular and humoral (antibody) attacks (Fig. 59-5). The efferent response usually begins when activated CD4 cells secrete various cytokines that drive the inflammatory response.¹⁷ IL-2 increases the expression of IL-2R on CD4 cells, driving proliferation and further differentiation of CD4 cells. The activated CD4 cells secrete additional lymphokines, including INF, which with IL-2 stimulate the activated CTL cells to bind to the allograft cells presenting donor MHC protein molecules. The CTL proliferate and specifically kill the allotarget by at least two mechanisms.¹⁸ In the presence of Ca²⁺, the protein perforin polymerizes onto the target cell and causes 16 to 20 nm pores to open in the cell membrane, resulting in osmotic collapse. The other likely method is by stimulation of apoptosis or programmed cell death by interaction of the lymphocyte Fas ligand with the APO-1/Fas receptor of the target cell. Second messengers are elicited that activate endonucleases and proteases to cause fragmentation of DNA and T-cell dissolution. Tissues that appear to have an immune privilege, such as the testis, eye, brain, and some tumors, utilize this Fas/Fas-L apoptotic pathway to destroy autoreactive lymphocytes. This pathway is being exploited for the development of tolerance to allogenic tissues in experimental models.¹⁹

View larger version (32K): [in this window] [in a new window]   FIGURE 59-5 The alloresponse is a complicated cellular and humoral process that generally begins when a CD4 cell recognizes a class II donor HLA molecule peptide complex presented on a donor heart or lung APC (direct path) and a precursor CTL cell of CD8 lineage binds to a class I donor molecule. The CD4 proliferates and produces IL-2 that drives the process further. The activated CTL cells seek class I donor specific targets and are stimulated by IL-2 and INF to kill targets. These CTL CD8 cells are primarily responsible for destruction of the allograft. Antibodies are selectively produced by B cells and draw inflammatory cells to the targets by antibody-dependent cytotoxicity. The complement is also activated by the humoral arm and initiates lytic changes and thrombosis in the allograft. Other inflammatory cytokines attract polymorphonuclear cells into the response and TNF and INF mix to activate macrophages.

On the other hand, recent data suggest that host reparative responses may mitigate the immunogenic effects of RI. By analyzing female cardiac allografts transplanted into male recipients, Quaini et al documented the migration of host stem cells that matured into myocytes, endothelium, and capillaries in the donor hearts as soon as 4 days posttransplantation.²⁴ With the discovery of this capacity for rapid formation of an organ chimera with up to 20% mature host-derived tissue, it is hypothesized that RI may actually result in a reduction, rather than enhancement, of allograft immunogenicity. The overall effect of RI on the allograft likely depends on which pathway plays a greater role in any given donor-recipient pair, illustrating an important future area for investigation.

The humoral response begins as host and B cells are drawn into the alloresponse by the lymphokines and by their own class I and II cell receptor engagement with the donor cells. The activated B cells evolve into plasma cells that produce allospecific antibodies against the donor class I and II HLA molecules and engage the complement cascade.

	T-CELL LYMPHOCYTE MATURATION AND ALLOACTIVATION

 
T cells form receptors (TCR) in the thymus that, similar to antibody, recognize specific peptide sequences. However, unlike antibody, the TCR cannot be released from the cell membrane and requires an association with five invariant polyproteins collectively called the CD3 complex. The TCR cannot recognize free antigen but is restricted by the HLA molecule with which it interacts. Genes responsible for the TCR randomly rearrange within the thymus to provide an astonishing array (10¹⁶) of potential binding sites necessary for diversity. Immature T cells are selected to survive in the thymus based upon whether and how strongly their TCR binds the HLA class I and II molecules expressed on the thymic epithelium (Fig. 59-6). The importance of the thymus is illustrated when it fails to develop in DiGeorge's syndrome. This disease results in an increased risk for a wide range of opportunistic infections that is reversed by thymic transplantation.²⁵ It is believed that when T cells react too strongly or too weakly to HLA molecules in the thymus, they are negatively selected and die of apotosis or DNA fragmentation to prevent the establishment of autoreactive clones and impotent cells. In fact, 95% of the thymocytes do not survive this selection.

View larger version (31K): [in this window] [in a new window]   FIGURE 59-6 T cells enter the thymus from the bone marrow where they differentiate and assemble diverse TCRs that cause the cells to be pl or mi selected, based on their usefulness. Diversity is based on rearrangement of genes responsible for variable portions of the chains that form the TCR heterodimer. When the TCR is pl selected on a class I HLA molecule, CD8 will form part of the receptor complex, and when a class II molecule is involved, then CD4 will form.

View larger version (24K): [in this window] [in a new window]   FIGURE 59-7 CD4 cells TCR has engaged its specific MHC class II molecule and bound allopeptide. The TCR complex initiates cytotoxic signal transduction that enables a Ca2+ influx activation of calcineurin nuclear factor for activating T cells, loses phosphate, and enters the nucleus to begin the promotor sequence to activate the IL-2 gene. (Adapted with permission from Parham M, in Haber E (ed): Immunobiology of Transplantation Molecular Cardiovascular Medicines. New York, Scientific American Press, 1995.)

	REJECTION OF HEART AND LUNGS

Hyperacute rejection (HAR) is said to occur when edema, hemorrhage, and thrombosis are noted shortly following revascularization. This process involves preformed antibodies that immediately bind to and activate the endothelium, initiating the complement and coagulation cascades. These antibodies bind to oligosaccharide antigens of the ABO blood group and xenoreactive antigens that are similar to those found on numerous endemic bacteria, protozoa, and viruses. The cross-reactivity of antibodies directed against these endemic microbes is likely to be responsible for the preexisting natural antibodies that cause HAR after transplantation with either ABO-incompatible or xenogenic organs. Because the titer and avidity of preformed antibodies against the blood group antigens in newborn infants is low, ABO-incompatible cardiac allografts have shown greater success in these patients.³² HAR also occurs from antibodies directed against nonself HLA antigens, especially in patients with a prior history of exposure to allogenic HLA through blood transfusions and pregnancies. Mechanical support with a ventricular assist device is also a strong risk factor for development of anti-HLA antibodies, which can be alleviated in part by the use of leukocyte-depleted, CMV-negative blood transfusions.³³

Although anti–class I Ab are more destructive of the graft endothelial cells, class II HLA is induced on the graft vasculature during periods of inflammation and can also provoke HAR when bound by anti–class II Ab following allotransplantation. Although HAR can be treated by cobra venom factor to deplete complement,³⁴ it is best prevented during allotransplantation by avoiding blood group disparities and identifying preexisting antibodies to HLA antigens. This is accomplished by exposing candidate serum to panels of donor cells with most HLA types. If a patient is determined to react to more than 10% of the panel, specific pretransplant (i.e., prospective) crossmatching is recommended between lymphocytes from the proposed donor and candidate's serum.

Despite their high endothelial avidity, neither the pre- nor posttransplant presence of anti-HLA antibodies guarantees HAR. By using an aggressive perioperative regimen of plasmapheresis, IVIG, and cytoxan, patients have successfully avoided HAR after transplantation despite a positive prospective crossmatch.³⁵ Anti–donor HLA antibodies have developed in some after transplant despite a negative prospective crossmatch. Titers may rise as early as 3 to 4 days after transplant, which implies a secondary antibody response with undetectable levels of preformed anti-HLA antibodies despite prior exposure. Although a process known as accelerated, acellular rejection occurs in a few, the induction of a protective phenotype (e.g. bcl-x_L, bcl-2, and A20) inhibits endothelial activation and prevents vascular injury in the vast majority.³⁶

Acute Rejection

Acute rejection involves both cellular and humoral immunity and is most common within weeks to months after transplantation. Although late acute episodes can occur, they often do so in the setting of a change in the balance of immunosuppression versus host immunity. A decrease in the blood level of immunosuppressant either by prescription drug interaction or by an upregulation in alloreactivity owing to viral infection can cause a late allorejection. Myocardial and pulmonary cytolysis on endomyocardial or pulmonary biopsy is the finding that supports a higher rejection grade (Table 59-1). Nonimmunological modalities, including measurement of hemodynamic parameters, radionuclide scanning, and magnetic resonance imaging have shown good correlation with established high-grade rejections but have not demonstrated sufficient predictive value to be included in routine clinical management.

Chronic Rejection of the Heart

Although chronic, persistent cell-mediated rejection causes progressive myocardial fibrosis and dysfunction, the term chronic allograft vasculopathy (AV) takes into consideration the role of multiple nonimmune factors in the etiology of this process. AV has a prevalence of at least 60% within 5 years of transplantation.³⁹ This obstructive process can progress to near-complete occlusion of the epicardial coronary arteries causing micro- and macroinfarction (Figs. 59-8 and 59-9) and is the leading cause of death after the first year following cardiac transplantation. The histologic findings differ from those seen in typical atherosclerosis with a uniform pattern of near-luminal occlusion by neointimal proliferation, and fewer early accumulations of extracellular lipid. Infiltrates of T cells that encircle the entire vessel are characteristic.⁴⁰ The concentric nature of the lesion has led to emergence of intravascular ultrasound (IVUS) as the optimal method for clinical detection of AV.⁴¹ Endothelial cells generally remain intact but are known to be dysfunctional based on a paradoxical constrictive response to acetylcholine.⁴²

View larger version (146K): [in this window] [in a new window]   FIGURE 59-8 Obliterative arteriopathy, or chronic allograft vasculopathy (CAV), results in concentric narrowing of the epicardial coronary arteries and their large intramyocardial branches. This section was taken from an explanted cardiac allograft resected at the time of retransplantation for chronic rejection. Note the fibrointimal hyperplasia, and adventitial and mural inflammation.

View larger version (72K): [in this window] [in a new window]   FIGURE 59-9 (A) Predisposition to thrombosis is a complication of chronic allograft vasculopathy (CAV). In this photomicrograph, an artery already narrowed by CAV shows a complicating thrombus (arrow). (B) A higher magnification of the artery shown in A illustrates the adventitial (a), medial (m), and intimal (i) mononuclear inflammation, which is more prevalent and severe in CAV than in atherosclerosis seen in the general population.

Our current limited pathophysiologic understanding of this relentless process is based largely on small animal models. By systematically isolating possible etiologic factors, these models have provided significant insight into the basic science of the vasculopathy process in cardiac allografts. However, out of logistical necessity, the surrogate pathologic lesion occurs much earlier than the typical changes of chronic rejection in clinical patients. Thus, the pathogenesis of the process being studied experimentally is almost certainly not the same as that occurring clinically. Indeed, many of the commonly used rodent models demonstrate suppression of AV lesion formation with standard immunosuppression such as cyclosporin,⁵⁸ a finding that significantly limits clinical relevance (Fig. 59-10).

View larger version (71K): [in this window] [in a new window]   FIGURE 59-10 (A) Experimental animal models of chronic allograft vasculopathy suggest that tolerance induction via the introduction of hematolymphoid chimerism can prevent chronic class II antigens on the microvasculature and large intramyocardial coronary arteries, which show early changes of chronic allograft vasculopathy (arrow). (B) In contrast, staining for donor MHC class II antigens in a cardiac allograft that is resistant to chronic rejection shows staining only in the interstitial hematolymphoid cells (small arrows), whereas the arteries are normal appearing (large arrows).

Treatment strategies will remain elusive unless more complete control of the alloresponse can be maintained by the newer xenobiotics and monoclonal antibodies or induction of tolerance. Clinicians are anxious to explore the potential for new xenobiotics that have demonstrated striking reduction in experimental AV based on their suppression of the smooth muscle response to the growth factors.⁵⁸

Chronic Rejection of Lung Allografts

The lung allograft, too, appears to be affected by a chronic process that limits the long-term usefulness of the organ. This chronic attrition can affect 30% to 50% of recipients within 3 years of pulmonary transplantation and 60% to 70% of patients who survive for 5 years.⁶¹ In the majority of patients, the problem is difficult to resolve once it develops, and the mortality rate at 3 years after diagnosis is 40% or higher. The term used to describe this chronic loss of function, obliterative bronchiolitis (OB), comes from the histologic findings of obliteration and fibrotic scarring of the terminal bronchioles (Fig. 59-11).⁶² However, the clinical diagnosis is rarely based on histology given the low sensitivity of transbronchial biopsy.⁶³ The lesions of OB involve the lung in a nonuniform manner and biopsy is performed mostly to rule out other causes of graft dysfunction such as acute rejection, infection, and airway complications. Because symptoms are nonspecific, the most sensitive test for early detection of OB is a fall in forced expiratory flow between 25% and 75% of the FVC (FEF 25-75).⁶⁴ The term "bronchiolitis obliterans syndrome" (BOS) was formulated to describe chronic allograft dysfunction in the absence of confirming histology by a progressive decline in FEV1, deemed a more reliable and reproducible pulmonary function test.⁶⁵ BOS grades 0 to 3 are assigned according to the percentage of FEV1 to best postoperative baseline value obtained. Bronchial wall thickening, distention of distal airways with air trapping, and frequent association of secondary acute infection have been detected on high-resolution chest CT scanning and are proposed as helpful in making the diagnosis.⁶⁶

View larger version (148K): [in this window] [in a new window]   FIGURE 59-11 Bronchiolitis obliterans. A small bronchiole has its lumen completely obliterated by dense scar tissue and mononuclear inflammatory cells.

The transplanted lung is unique amongst solid organ transplants in that it is exposed to the outside world. As a result, these lungs are particularly susceptible to the immunomodulary effects of respiratory viruses. By serving as an adjuvant for the cellular immune response or by a direct cytopathic effect on the airway, respiratory viruses such as CMV, respiratory syncytial virus, adenovirus, influenza, and parainfluenza infection have all been implicated as risk factors for BOS.⁶⁹ CMV infection, in particular, has a potent effect on donor-specific and nonspecific immune responses. An increase in INF and MHC class II antigen expression has been noted in bronchoalveolar lavage (BAL) cells during infections with CMV.⁷⁰ Most transplant centers believe in an association between CMV infection and BOS, although a precise relationship is far from uniformly accepted.

Although other facilitating factors certainly exist, several lines of clinical evidence support the alloresponse as a more important force behind the development of OB than cardiac AV (Figs. 59-12 and 59-13). First, acute rejection has consistently been found to be the leading risk factor for the eventual development of OB.⁷¹ In particular, OB develops in the setting of indirect alloimmune⁷² and alloantibody⁷³ responses to HLA-A mismatches and is occasionally stabilized by augmented immunosuppression (Fig. 59-13).^74,75 Second, an identical form of OB can occur in bone marrow transplant recipients with graft-versus-host disease following the recognition of the host lungs by the grafted alloreactive T cells.⁷⁶ Third, cells from bronchial lavages of OB patients have demonstrated TH1 cytokine mRNA profiles (IL-1, IL-2, IL-6 and IFN).⁷⁷ Finally, in the Pittsburgh study of microchimerism, it appeared that those patients with OB had less evidence of microchimerism in blood, lymph nodes, and skin, which follows the general concept of less immune reactivity for those patients with a generalized chimeric state.⁷⁸

View larger version (127K): [in this window] [in a new window]   FIGURE 59-12 Bronchiolitis obliterans. Stains for S100 protein decorate antigen-processing dendritic cells (arrows) present in increased numbers in the airways of lung allografts experiencing rejection.

View larger version (115K): [in this window] [in a new window]   FIGURE 59-13 Bronchiolitis obliterans. Chronic airway rejection is characterized by increased expression of HLA class II antigens, especially HLA-DR (shown here) on respiratory epithelial cells.

	NEW IMMUNOSUPPRESSIVE DRUGS

 
The improved outlook for transplant recipients has followed the introduction of xenobiotic immunosuppressants, that is, those drugs produced by organic synthesis or microorganisms that suppress the immune system. Between 1960 and 1985 only steroids, azathioprine (AZA), and cyclosporin (CsA) had been adopted for use in clinical transplantation. These agents have been more recently joined by polyclonal and monoclonal anti–T-cell antibodies. In the last few years, progress in the molecular understanding of the alloresponse has made new discoveries possible and allowed agents to be classified by mechanism of molecular action (Fig. 59-14 and Table 59-2).

View larger version (39K): [in this window] [in a new window]   FIGURE 59-14 Immunosuppressants available to clinicians are directed toward inhibiting T-cell activation at various steps and by varied mechanisms, including interference with TCR complex (OKT3 Mab) and other surface ligands (anti–ICAM-1, anti-CD2, others); Ca2+-dependent (CsA, FK) signal transduction; inhibition of cytokine IL-2 action in promoting cellular proliferation (RPM, LFM); and inhibition of purine (AZA, MZR, MMP) or pyrimidine synthesis BQR.

Transplant physicians have recognized the benefits of corticosteroids from the very early days of clinical transplantation. These molecules have protean effects that, like any steroid, are mediated through intracellular receptors that alter gene transcription. The predominant anti-inflammatory effects of glucocorticoids such as the blockade of NFKB-induced transcription of inflammatory cytokines and adhesion molecules derive from the inhibition of gene transcription. On the other hand, the metabolic side effects such as muscle wasting and diabetes derive from positive transcriptional effects.⁸¹ This concept of differing mechanisms of action has prompted investigations to develop corticosteriod analogues that bind to intracellular receptors to promote the inflammatory effects without the metabolic effects.

Glucocorticoids have been found to induce apoptosis in malignant T cells⁸² and are therefore an especially appropriate choice of sole immunosuppression in the setting of posttransplant lymphoproliferative disorder. Outside of this subgroup, however, recent advances in the development of tolerance protocols have suggested that steroid use blocks certain immune signaling pathways necessary to induce donor-specific anergy or suppressor cells.³⁰ In addition, steroid weaning reduces the tendency towards diabetes and dyslipidemia, which may decrease the incidence of AV.⁸³

Cytokine Synthesis Inhibitors

Cyclosporin (CsA) inhibits the gene activation necessary for IL-2 production. To accomplish this, CsA inhibits the function of a Ca^{2 +} activated calcineurin phosphatase when bound to its cytoplasmic receptor.⁸⁴ This prevents the activation and nuclear translocation of the nuclear factor for activation of T cells (NFAT), precluding its engagement with the promoter sequence of the IL-2 gene. Blockade of the Ca²⁺ calcineurin phosphatase complex also inhibits the production of nitric oxide synthetase, a potential mechanism by which CsA seems to promote AV in animal models.⁸⁵

Originally oil-based, CsA has been replaced by a novel microemulsion formula, Neoral, which has significantly improved its bioavailability and reduced pharmacokinetic variability between patients. Approximately 30% of heart transplant recipients develop nephrotoxicity, the primary toxicity of CsA, which appears to be mediated by the inhibition of prostaglandin metabolites. However, the prostaglandin analogue misoprostol has afforded little clinical benefit.⁸⁶ In two recent series of heart transplant patients, calcineurin inhibition was the sole indication for metachronous kidney transplantation.^87,88 CsA-induced alterations in cell phenotype explain other side effects such as hypertension and dyslipidemia.^89,90

CsA was widely embraced as the central component for effective multidrug immunosuppression until FK506 (tacrolimus) was introduced to patients in Pittsburgh in 1988. Tacrolimus combines with a different cytosolic protein than CsA (FK binding protein) but complexes with the same Ca²⁺ activated calcineurin to prevent the activation of NFAT.⁷⁶ Tacrolimus has proven to be at least as effective in heart transplant patients⁹¹ and possibly better in lung transplant patients.⁹² It has found particular success following a switch from CsA-based immunosuppression when faced with a refractory acute rejection of the heart or lung⁹³ or bronchiolitis obliterans.⁷⁴ Given a mechanism of action similar to that of CsA, the reason for the improved effectiveness of tacrolimus in refractory rejection likely relates to more predictable pharmacokinetics.⁹⁴ Therefore, ongoing clinical trials comparing tacrolimus with Neoral are of great interest with regard to efficacy but are not likely to change the improved side effect profile already demonstrated with tacrolimus in multicenter trials.⁹¹ Compared to recipients receiving CsA, tacrolimus was found to be associated with less facial disfigurement, hirsutism, hypertension, and hyperlipidemia but equal nephrotoxicity, and was perhaps associated with greater neurotoxic and diabetogenic effects.

Inhibitors of DNA Synthesis

Antimetabolites are immunosuppressive because they inhibit the synthesis of nucleotides necessary for DNA's rapidly dividing cells. The classic antimetabolite has been azathioprine (AZA), which inhibits purine synthesis and therefore DNA and RNA synthesis throughout all dividing cells. Mycophenolate mofetil (MMF) appears to be more selective for T and B cells than AZA⁹⁵ based on its ability to block the activity of enzyme inosine monophosphate dehydrogenase, and therefore the synthesis of purines in the de novo pathway. Unlike other parenchymal and peripheral blood cells, T cells and B cells cannot use the salvage pathway and depend solely on the de novo pathway for purine synthesis. Compared to AZA, randomized clinical trials have shown a reduction in acute rejection events and antibody production with MMF in both heart⁹⁶ and lung⁹⁷ transplant patients. The reduction in chronic graft loss that has been demonstrated with the use of MMF versus AZA in renal transplantation has not yet been confirmed in cardiothoracic transplants. Neutropenia has not been a limiting factor as it has been with AZA.

Brequinar sodium (BQR) is a new addition to the antimetabolite group.⁹⁸ Unlike MMF, its action appears to be directed against dihydroorotate dehydrogenase (DHODH), an enzyme in the pathway leading to synthesis of pyrimidines. The rationale for use of BQR is similar to that for MMF, given that activated immune cells are relatively more dependent on de novo synthesis of pyrimidine's effects than nonimmune cells, although unlike for purines, a salvage pathway does exist. As a result, BQR appears to be less selective for immune cells. Another DHODH inhibitor, leflunomide, has demonstrated a much more favorable therapeutic window although a profound weight loss has been seen in animal and human trials.⁹⁹ Leflunomide depletes ATP-dependent enzymes, which inhibits the glycosylation of adhesion molecules, providing another possible mechanism of immunosuppression.¹⁰⁰ Clinical utility of both BQR and leflunomide has been limited by myelotoxicity and GI effects; planned clinical trials have been stopped.

IL-2 Signal Transduction Inhibitor

Rapamycin (sirolimus; RPM) is structurally similar to tacrolimus and binds with FK binding protein (FKBP) but surprisingly does not inhibit the calcium-activated calcineurin.⁹⁴ Instead, RPM acts at a point downstream from the cytokine inhibitors and upstream from the antiproliferative agents. In alloreactive T cells, stimulation of the IL-2 receptor leads to clonal proliferation following initiation of the cell cycle and conversion from the resting (G₀) to proliferative (G₁/S) state. The RPM/FKBP complex binds the so-called "target of rapamycin," a lipid kinase,¹⁰¹ and prevents the signaling between IL-2 receptor activation and cell-cycle initiation. Because of theoretical concerns of competition for FK binding protein, sirolimus was combined initially with CsA¹⁰² but recent clinical trials in renal transplantation have actually demonstrated greater success when combined with FK506.¹⁰³ In vitro studies have demonstrated that RPM also induces cell cycle arrest in B cells and smooth muscle cells.¹⁰⁴ This smooth muscle cell antiproliferative effect is thought responsible for the arrest of AV in both small⁶⁰ and large¹⁰⁵ animal experimental models, and for the clinical prevention of restenosis after using RPM-coated intracoronary stents.¹⁰⁶ In preliminary randomized studies, the use of RPM instead of AZA following heart transplantation has resulted in reduced AV by IVUS evaluation at 6 months¹⁰⁷ and 1 year.¹⁰⁸ Sirolimus is not nephroxic but it may enhance the renal toxic effects of calcineurin inhibitors¹⁰³; its main toxicity is hyperlipidemia.

Combinations of these drugs that act at the level of cytokine production, the proliferative response to cytokines, and/or the signaling between the two have demonstrated additive immunosuppressive effects.¹⁰⁹ This will not only effectively reduce the alloresponse but also potentially do so with lower doses of each.

Inhibition of T- and B-Cell Maturation

Deoxyspergualin (DSG) does not inhibit the synthesis or actions of cytokine but has been shown to inhibit the maturation of T and B cells and APC.¹¹⁰ Clinical trials that were conducted in high-risk renal allograft recipients were stopped due to a high incidence of leukopenia.

Receptor Antagonists and Monoclonal Antibodies

Polyclonal anti–T-cell preparations (ATG) have been developed that recognize T-cell surface structures and kill these targets by inducing F_C-receptor–mediated cell lysis or by complement-dependent cell lysis. In the mid-1980s, the murine anti–human CD3 monoclonal antibody (OKT3) was developed that recognizes the epsilon protein of the CD3 complex on all T cells. When used as induction agents in thoracic transplantation, ATG and OKT3 have been found only to delay the onset of acute rejection at the expense of a profound, uncontrolled immunosuppression that increases the risk for opportunistic infections and malignancy.¹¹¹ Furthermore, their main toxicity, the cytokine release syndrome, is tolerated particularly poorly in heart and lung transplant recipients. As a result, their current use is limited in most centers for the treatment of refractory acute rejection and as a calcineurin-inhibitor sparing agent in those with prolonged postoperative renal dysfunction.

The development of a humanized monoclonal antibody (mAb) against the IL-2 receptor provided the opportunity for a more selective targeting of activated T cells, the only cells that express this receptor. In a small (55 heart transplant recipients), randomized, clinical trial, induction therapy using this mAb, dacluzimab, reduced the frequency and severity of acute rejection events over the study period. In addition, there were essentially no side effects and no increased risk of infections or malignancy.¹¹² Pilot studies using mAb against the cell adhesion molecules LFA-1¹¹³ or ICAM-1¹¹⁴ showed promise in preventing reperfusion injury but variable success against acute rejection. The combination of the two, which was synergistic against acute rejection in rodent models, has not been tried clinically. Also awaiting clinical trial is a strategy which inhibits T-cell costimulation such as the anti–CD154 mAb or CTLA-4 Ig which have produced tolerance in the nonhuman primate model.³⁰ Other monoclonal antibodies that are in various stages of clinical development may have a specific role in therapy, but hopes for a magic bullet likely will not be realized, as the immune response is far from simple and is based on redundancy by way of alternative pathways. A combination of various mAb would likely be the best protocol to address this redundancy. Unfortunately, no preclinical or clinical trials using a combination strategy have been performed in large part due to financial conflicts between the different pharmaceutical companies that own the rights to these agents.

	TOLERANCE

 
While immunosuppressive agents have permitted the replacement of heart and lungs to become realities, their toxic side effects and inability to prevent more chronic forms of rejection have driven the search for alternative strategies. Experimental models have suggested that the induction of tolerance is the most effective way to prevent chronic rejection.¹¹⁵ Indeed, the absence of AV or OB has been considered to be the best clinical end point for the evaluation of successful tolerance induction in future trials by the National Heart, Lung and Blood Institute Heart and Lung Tolerance Working Group.¹¹⁶ While a myriad of protocols have induced tolerance in small animals, only a few have been reproduced in swine or nonhuman primate transplant models. The list of studies relevant to thoracic surgery shrinks further when considering the elusiveness of translating tolerance protocols from one organ to the other. In addition, there is a general belief that induction of tolerance in thoracic organs is more difficult than for other organs such as the liver or kidney.¹¹⁷ To date, only three protocols have induced prolonged survival of cardiac allografts in large animals without immunosuppression (none in lung allografts): (1) the induction of mixed chimerism which is thought to work through central tolerance; (2) the use of costimulatory blockade to induce peripheral tolerance; and (3) the cotransplantation of heart and kidney allografts, which works by unknown mechanisms.¹¹⁷

By introducing allogenic bone marrow cells into newborn mice, Billingham et al induced a mixed chimeric state that was the first demonstration of allograft tolerance in 1953.¹¹⁸ A recent analysis of renal transplant recipients from sibling donors has shown this type of tolerance to occur towards noninherited maternal HLA due to prior in utero exposure to this foreign antigen.¹¹⁹ In these mixed chimeras, bone marrow–derived elements of both host and donor appear in the thymus and present ligands for negative selection of newly developing T cells that are either donor or host reactive.¹²⁰ This induces clonal deletion, one of the most reliable approaches to achieving long-term donor-specific tolerance. In an adult organism, creation of a mixed chimera requires the infusion of donor hematopoietic cells along with a conditioning regimen to enhance engraftment. Conditioning using antilymphocyte serum has been attempted but has produced little¹²¹ to no¹²² evidence of durable hematopoietic cell engraftment and minimal impact on allograft survival. The most common method used in preclinical and clinical trials has been the use of a toxic dose of total lymphoid irradiation (TLI), similar to that used to treat Hodgkin disease.¹²³ The use of a nonmyeloablative conditioning regimen of CD3 monoclonal Ab bound to the diphtheria immunotoxic instead of TLI has enabled the induction of stable mixed chimerism using donor stem cells with significantly less toxicity. Subsequent transplantation resulted in long-term tolerance in the swine model despite a minor-antigen mismatched histocompatibility barrier. Recently, the use of pleuripotent embryonic stem cells has allowed the development of mixed chimerism and tolerance without conditioning in rodents.¹²⁴ These preliminary results have not yet been reproduced in large animal models.

It has been observed that in heart¹²⁵ and lung¹²⁶ transplant recipients enjoying long-term survival, donor-type lymphoid and dendritic cells migrate from the graft and establish themselves in the unconditioned recipient's periphery.¹²⁷ In this case, donor cells exist at levels less than cytometric detection in the periphery of solid organ transplant recipients (1 in 10⁵ cells as detected by polymerase chain reaction techniques) with a kinetics and patchy distribution that resemble a spreading infection. In lung transplant recipients, this microchimerism has been associated with donor-specific hyporeactivity and lower incidence of OB.¹²⁸

The observation that microchimerism is a common event after heart and lung transplantation and associated with long-term graft acceptance has stimulated attempts to augment microchimerism with perioperative bone marrow transfusion. Pham et al have provided proof of the principle that induced microchimerism influences acute and chronic rejection in clinical heart¹²⁹ and lung¹³⁰ transplantation without producing GVHD. However, a causal relationship between microchimerism and a decrease in target events such as rejection or graft survival was not established. The association between microchimerism and graft survival is inconsistent with the identification of both long-term survivors who do not have it and patients with multiple episodes of rejection who do.¹³¹ Furthermore, only mixed chimerism, and not microchimerism, has been shown in animal models to induce systemic, stable allograft-specific tolerance.¹³²

The second mechanism for the induction and maintenance of tolerance that occurs following costimulatory blockade is anergy, also known as peripheral tolerance.¹³³ The two-signal model of T-cell activation holds that an alloresponse to the interaction of the T-cell receptor with its antigen requires a second signal prompted by the other molecular participants of the immunologic synapse. Engagement of the T-cell receptor without these signals induces anergy or a lack of T-cell proliferation on antigen stimulation. The advantage of this mechanism is achieved by exposing T cells to alloantigen under the umbrella of mAb blocking costimulatory molecules such as CD28,³¹ LFA-1, or CD2 ligand. The period of immunosuppression lasts only as long as the mAb are at therapeutic levels in the host. As the mAb clear, the immune competence returns to all antigens other than those to which tolerance had been induced.

It has long been recognized that passenger leukocytes, that is, cells derived from the organ transplant most commonly dendritic in nature, can escape the organ and accumulate in peripheral host lymphoid tissues. It is possible that donor-derived immature passenger dendritic cells mediate a form of anergy by indirect and direct presentation of alloantigen with limited secondary signals necessary for T-cell proliferation. Kidney transplants are thought to have enhanced numbers of these cells as the proposed mechanism of their immunosuppressive effect on cardiac allograft rejection.

	XENOTRANSPLANTATION

 
Animal-to-human transplantation, known as xenotransplantation, has been proposed to alleviate the critical shortage of human donor organs. Approximately 30% of the patients waiting for hearts and lungs will die without receiving a transplant. In 1964, 4 years before the first human allotransplant, Hardy attempted to replace a 68-year-old man's failing heart with one obtained from a chimpanzee.¹³⁴ Since then, there have been only six additional attempts of xenotransplantation reported using pig, sheep, baboon, or chimpanzee donors.^135–138 These cases, although unsuccessful, have provided insight and promise to the field. In 1984, Bailey et al placed an ABO-incompatible baboon heart into a child with a hypoplastic left heart syndrome. Baby Faye, as the child was known, made remarkable progress for 20 days when rather suddenly the xenograft stopped functioning. Examination of the heart gave evidence consistent with a humoral rejection, perhaps related to the blood group incompatibility.

Although nonhuman primates provide concordant organs for cross-species transplantation, phylogenetically disparate or discordant donor organs from pigs are favored for several reasons. First, broad use of the relatively rare and sentient primates is unlikely to gain societal acceptance. Second, retroviruses from pigs are much less likely than those from nonhuman primates to transmit disease to humans.¹³⁹ Finally, the short gestation, time to maturation, and large litters of pigs relative to primates simplifies their breeding and improves their candidacy for germ-line gene therapy. Accordingly, the porcine heart has been the focus of most of the experimental work in heart and lung xenotransplantation and was the most recent clinical xenograft reported for use in cardiac transplantation in 1992.¹³⁸

Xenograft Hyperacute Rejection

The first major obstacle to discordant cross-species transplantation is generally believed to be the process described as hyperacute rejection (HAR). Hyperacute rejection is mediated largely by preformed xenoreactive antibodies and the relative incompatibility of discordant xenograft complement regulatory proteins (e.g., decay accelerating factor) with the human complement system. The primary human xenoreactive antibodies that initiate HAR against discordant pig hearts are specific for the blood group carbohydrate Gal (1–3 Gal), an antigen not present in concordant nonhuman primates.¹⁴⁰ Utilizing the classical pathway, the binding of xenoreactive antibodies to the pig endothelium leads to the unregulated activation of complement due to inadequate function of swine counter-regulatory proteins for human complement. The resulting uncontrolled deposition of the terminal complement complexes (C5b67) on the swine endothelial cells disrupts the endothelial cell barrier function as they retract and generate intracellular gaps (type 1 activation). Platelets then are attracted to exposed extracellular matrix and release vasoactive substances, including thromboxane A2, that stimulate vasoconstriction. The procoagulant state is intensified because of the loss of heparin sulfate proteoglycans from their surface.

These findings have led White, Pedor, and Platt to develop swine that are transgenic for human DAF and CD-59. By overexpressing human DAF and CD-59, these transgenic organs successfully avert hyperacute rejection following pig to primate renal, heart and lung transplantation.^141–143 The temporary, pretransplant depletion of complement using cobra venom³⁴ and anti-Gal antibodies using any of several different methods have provided further success against HAR.¹⁴⁴ Evidence exists that if HAR and AVR can be prevented initially, then the xenograft may "accommodate" in a manner in which it becomes resistant to future exposure to human antibody and complement.^36,145 This is thought to be mediated by increased expression of antiapoptotic genes and inhibition of NFKB transcriptional activation in the xenograft endothelium. The enhancement of this pathway would serve obvious benefits in xenotransplantation. However, the greatest potential for a significant advance has been achieved by the recent creation of Gal-1,3 galactosyl transferase knockout pigs.

Acute Vascular Xenograft Reaction

Despite prevention of HAR by either disrupting antibody binding or by depletion or inhibition of complement, xenografts are subjected to a process named acute vascular rejection (AVR).¹⁴⁵ Although the histologic picture of AVR with its hemorrhage and thrombosis is very characteristic of HAR, it appears to be a distinct process not dependent on complement nor appearing in concordant transplant combinations. The pathophysiology begins with naturally occurring anti–pig antibodies binding to the endothelial surface. This leads to levels of complement activation through the membrane attack complex (MAC) that are below lytic levels but that lead to the induction of IL-1, which mediates other changes on the surface of the endothelial cell that, by and large, create a strongly procoagulant state (type 2 activation). These changes include the induction of procoagulant tissue factor, release of plasminogen activator inhibitor, decrease in tissue plasminogen activator, and a loss in thrombomodulin activity. Thrombomodulin is expressed on the surface of vascular endothelial cells and reduces thrombotic process by thrombin-dependent activation of protein C, which in turn degrades the procoagulant cofactor's factors Va and VIIIa. It has been noted also that E-selectin, responsible for leukocyte rolling on the endothelium, is also upregulated during AVR.

Cell-Mediated Xenograft Rejection

Although cell-mediated rejection has not been studied extensively in the xenograft model because of difficulties in overcoming HAR and AVR, recent investigations have suggested that xenografts have increased susceptibility to cell-mediated injury and also to attack by NK cells.¹⁴⁶ NK cells normally are inhibited by class I MHC receptors, yet when added to xenograft tissue culture, NK cells have been demonstrated to cause cytotoxicity and phenotypic changes in a disruptive endothelial cell monolayer consistent with retraction and gap formation typical of the activated endothelium described in HAR.^147,148 It was hoped that thymic selection, which permits T lymphocytes to recognize allogeneic cells directly, might be less effective in producing T lymphocytes that might recognize the porcine xenogenic cells. However, it has been determined that human T cells can recognize porcine cells directly through MHC class II antigen.¹⁴⁹

Significant progress has been made with the discordant porcine-to-primate model. However, an impact on human transplantation awaits additional studies that promote a further understanding of the effects of the inhibition of early HAR, methods to persistently reduce AVR, and finally a way of dealing with an enhanced cell-mediated rejection. It is likely that a combination of immunosuppressants, transgenic animals, and even tolerance-induction protocols may provide a suitable therapeutic cocktail. Chief clinical investigators in the field of xenotransplantation have stressed the need to look for intermediate end points as means of understanding processes, since the long-term goal of routinely successful discordant xenografting will require solving multiple complex processes. It is likely then that well-prepared surgical groups will soon initiate bridge trials in which short-term survival of the xenograft might be predicted, and information gained will be invaluable to the science.