Springer Semin Immunopathol (2003) 25:141Ð165 DOI 10.1007/s00281-003-0134-2 © Springer-Verlag 2003

Immunopathology of organ transplantation

Lorraine C. Racusen Department of Pathology, Carnegie 4, The Johns Hopkins University School of Medicine, 600 North Wolfe Street, Baltimore, MD 21287, USA

Abstract.&p.1: Immunopathology techniques are useful and sometimes indispensable tools to diagnose and/or define pathogenesis of disease processes in an organ allo- graft. Using immunostaining and/or in situ hybridization, cells and molecules can be identified and localized in organ grafts. Considered in this chapter are the “priming” of tissues by ischemia and infection, cell-mediated and antibody-mediated acute and chronic rejection in xenografts and allografts, graft infection, and recurrent or de novo immune disease in organ grafts. The chapter is designed to provide an over- view, with illustrative examples of the contribution of immunopathology to analysis of solid organ allografts.&bdy:

Introduction

Immunopathology has contributed and continues to contribute greatly to diagnosis and to definition of pathogenesis of important pathological processes in solid organ allografts. These processes include ischemia/reperfusion injury with tissue “prim- ing”, allograft rejection, infection, and recurrent disease in the allograft. Immunopa- thology is useful in confirming the presence of acute and chronic cell-mediated re- jection, and is particularly crucial for tissue diagnosis of antibody-mediated rejec- tion, each of which will be considered in this chapter. The immunopathology of both xenografts and allografts will be discussed. Technical aspects of immunostaining will be outlined briefly as well.

Techniques

Immunopathology is a broad term encompassing immunohistochemistry, immuno- fluorescence and in situ hybridization. Broadly, antibodies or probes specific for tar- get proteins or nucleic acids are reacted with tissue sections and attach to the target antigen. The type of tissue that can be utilized for such studies depends on the tech- nique, the stability of the antigen, and the avidity of the probe. For detection of some antigens, staining must be performed on frozen tissue. Other protein antigens and nu- 142 L.C. Racusen cleic acid sequences can generally be detected in formalin-fixed paraffin-embedded tissue. Some preparation of the tissue, such as protease digestion or microwaving, may be required to “unmask” the antigen; these techniques must be carefully con- trolled and standardized, to avoid over-interpretation of staining. Intracellular mole- cules such as cytokines cannot be detected reliably in formalin-fixed tissue, but can be detected by immunostaining of tissue fixed in acetone. In immunohistochemistry, the antibody or probe carries an enzyme (e.g., peroxi- dase) which can trigger a chemical reaction, resulting in localized precipitation of a colored reaction product at the site of antigen that is detectable by light microscopy. For direct immunofluorescence, the antibody carries a dye that can be activated by UV light (e.g., fluorescein, rhodamine). In indirect immunofluorescence, the primary antibody specific for the antigen is not labeled, but tissue sections are reacted with a second labeled antibody to the primary antibody. With appropriate positive and nega- tive controls run concurrently, very precise detection and localization can be achieved. Intensity and extent of staining can be quantitatively assessed. In in situ hybridization, a labeled riboprobe is utilized that is complementary to a unique nucleic acid sequence in target cells or organisms. With a radio-labeled probe, following reaction of the probe with the tissue, sections are dipped in nuclear emulsion and exposed in the dark; silver grains from the emulsion precipitate at the site of the radio-label. Probes can also be labeled with digoxigenin; the probe can be detected using enzyme-labeled anti-digoxigenin antibody, with amplification to in- tensify precipitated reaction product. This technique also allows precise localization and can be quantitatively assessed. Relative localization of several different target antigens in a specimen can be as- sessed by staining of immediately sequential sections. However, the availability of a variety of colorimetric histochemical reactions and UV-activated dyes now allows detection of several different target antigens simultaneously on a single tissue sec- tion, which can provide very precise information about spatial relationships, with overlapping signals suggesting intimate interactions between antigens. A variety of different techniques for such multiple labeling are available (e.g., [147]). Recent ad- vances in immunostaining including new reporter molecules and methods to increase sensitivity and improve multiple immunostaining have been recently reviewed [146]. Recent advances in microscopy have further advanced the field. Confocal micro- scopy can be used to refine three-dimensional (3D) localization of labeled molecules. Quantitative digital fluorescence microscopy using laser scanning (thick sections) or wide-field microscopy (thin sections) allows collection of large data sets for molecu- lar localization in cells and tissues [6]. Different wavelengths for activation have en- abled deeper tissue penetration for 3D imaging in very localized volumes [29]. De- velopment of fluorescent tagging with green fluorescent protein can be used to study protein mobility in living cells [119]. These technologies have yet to be applied on any scale to studies of allograft biology, but doubtless will be applied to the field in the near future. While the focus of this review is primarily on immunopathology as applied to tis- sue sections, the same techniques can be and have been used to study material ob- tained via less invasive methods. Specimens obtained via fine needle aspiration of organs, bronchoalveolar lavage or collection of urine sediment, for example, can be examined for expression of phenotypic markers, markers signaling activation and proliferation, or “stress response”. Indeed, immunohistology of allograft aspirates may contribute significantly to specificity of diagnosis [120]. Suspended cells can be Immunopathology 143 exposed to one or more labeled antibodies and examined by light microscopy. With adequate numbers of cells, flow cytometry can be employed and cells analyzed by cell sorting. With appropriate purification and positive and negative controls, excel- lent characterization and quantitation of cells by phenotype is possible. Large num- bers of available dyes and technological advances have made this an increasingly useful technology (see [59] for recent review).

Tissue priming/“Innate” immunity

It is almost impossible to avoid injury to an organ allograft. Ischemia/reperfusion in- jury, supervening infection and other processes can trigger nonspecific innate im- mune and inflammatory responses in the graft. These injury responses in turn can trigger and amplify antigen-specific adaptive immune responses. This tissue “prim- ing” may lead not just to acute alloimmune reaction to the graft [84], but may precip- itate indolent chronic processes, eventuating in fibrosing changes in the allograft [50]. These processes have particular relevance in organs from cadaveric donors [8]. Immunohistology is being utilized as a tool to understand this “priming” process, and in kidney allografts, at least, may be useful in identifying recipients at particular risk for early graft dysfunction [130]. A number of investigators have used immuno- histology to demonstrate increased expression of adhesion molecules, HLA antigens (class II), and complement in both native and allograft organs which have undergone ischemia/reperfusion [42, 68, 136]. Infection, and especially cytomegalovirus infec- tion, may also produce analogous tissue changes [151]. Adhesion molecules contributing to homing and targeting of inflammatory infil- trates include intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1) and selectins. In experimental models, adhesion molecule ex- pression is altered in ischemia/reperfusion injury (reviews in [77, 142]). Analogous changes have been demonstrated in clinical renal biopsy specimens studied by im- munohistochemistry. For example, expression of ICAM-1 and VCAM-1 have been shown to be elevated in pre-transplant donor biopsy samples from cadaveric com- pared to live donor kidney allografts; expression of ICAM-1 on tubular cells appears to be a predictor for delayed graft function, although not rejection, in cadaveric kid- neys [130]. Adhesion molecule expression, including ICAM, VCAM and E-selectin, have also been shown to be increased in cardiac allograft biopsy material post-trans- plant [19, 57]; in one study these changes were coincident with elevated serum tro- ponin T levels, indicative of ischemia [57]. New or enhanced expression of HLA antigens has been demonstrated by immu- nohistochemistry in organs with ischemic injury. In particular, HLA class II antigens, which are constitutively expressed on a limited number of cell types, may be ex- pressed by parenchymal cells following tissue injury. Class II DR expression has also been shown to be predictive of later allograft function [38]. Using immunohis- tology in sequential studies of unilateral kidney ischemia, Ibrahim et al. [68] were able to demonstrate increased expression of class II antigens and localize expression to tubular epithelial cells, peritubular capillaries and interstitial cells; this was associ- ated with influx of monocytes, macrophages, and T cells and increased cytokine pro- duction . The response was strain dependent, and progressive after the ischemic in- sult. Up-regulation of HLA molecule expression may be asynchronous. For example, in a mouse model of unilateral renal ischemia, class I antigen expression increased 144 L.C. Racusen

Fig. 1. Immunofluorescence staining for C3d in a post-perfusion heart biopsy specimen with ischemic in- jury (courtesy of Dr. W. Baldwin) by day 3, class II antigens by day 7, primarily in interstitial cells, but also in tubular cells. Increased expression was associated with infiltrate of Thy 1.2+ cells [136]. Immunohistochemical studies have shown not only that dysregulation of class I and class II antigens may have differing kinetics, but increases in expression may vary between different organ allografts in the same strain [94]. Fuggle et al. [39] used locus-specific or polymorphic antibodies to examine specificity of induced tu- bular class II antigens in renal allografts. While HLA-DR was expressed on proximal tubules in pre-transplant biopsy specimens, HLA-DR, DQ and DP antigens were all expressed on tubular cell membranes and in cytoplasm in serial post-transplant biop- sy material; the class II antigens expressed were of donor origin, even at 1 year post transplant [39]. Similarly, class II expression can be demonstrated in lung allografts within 24 h of transplant, preceding rejection, and correlating with influx of neutro- phils, lymphocytes and increase in cytotoxicity in bronchoalveolar lavage fluid [2]. Persistent class II antigen expression on endothelium may provide antigen-specific signaling to leukocytes [28]. Recent studies using monoclonal antibodies and indi- rect immunohistochemistry have demonstrated increased expression of MHC class I chain-related antigens A and B (MIC A and B) in renal and pancreatic allografts in ischemic injury, as well as in acute and chronic rejection [56] Ischemia/reperfusion and extracorporeal circulation devices, relevant in organ har- vesting, can also lead to activation of complement, potentially via “natural antibodies” and/or acute-phase proteins such as C-reactive protein (reviewed in [8]). Immunohistol- ogy has proven to be an invaluable tool in analyzing the role of complement in injury models and clinical tissue samples. In human cardiac allografts, ischemic injury has been shown to be associated with widespread deposition of complement, and especially C3d and C4d, on the endothelium of small vessels [9], (Fig. 1). This has not been de- monstrable in cadaveric renal transplants, at least in the post-perfusion biopsy material [47], but this has not been extensively investigated. Recent studies have demonstrated that macrophages in graft infiltrates can be a source of complement production [110]. The nonspecific innate immunity response triggered by ischemia, infection and other insults involves some of the same effector mechanisms as adaptive reactions, which are antigen specific and involve immunologic memory [111]. There is an in- Immunopathology 145

Fig. 2. Immunoperoxidase staining for monocytes/macrophages following ischemic injury in the kidney. Numerous positive cells are seen in glomerular and peritubular capillaries creasing body of data supporting overlapping functions of components of these sys- tems. The T cell has been proposed as the bridge between innate and adaptive immu- nity [111]. Cellular components of the innate immune system are detectable in in- jured organ grafts. These cells include professional antigen-presenting dendritic cells, which have been demonstrated to migrate into the kidney following transplant surgery [105]. Natural killer (NK) cells can be cytolytic and produce proinflammato- ry cytokines, and have also been detected in ischemia/reperfusion injury in experi- mental models and clinical transplant biopsy specimens. Indeed, NK and NK-like cells have been detected by flow cytometry in liver grafts before transplantation (do- nor origin) as well as after revascularization (recipient origin) [98]. T cells are early responders to micro-organisms and stressed cells, rapidly releasing interferon, poten- tially influencing allo-immune responses; these cells have also been demonstrated in allografts by immunostaining (e.g., [98]). Finally, monocytes/macrophages are part of the innate immune response, demonstrable in tissue following ischemic injury [9, 95, 143] (Fig. 2). These cells in turn can activate complement, activate the coagula- tion cascade, present antigen, and serve as effectors of tissue injury. Heat shock proteins (HSPs) are cytoprotective proteins induced by organ harvesting and preservation, ischemia/reperfusion injury, and surgery [72, 101, 145]. Inflammatory responses, including rejection, correlate with increased HSP expression [101]. Immuno- histology has been used to demonstrate HSPs in allografts; there may be differential ex- pression of different HSPs in the various tissue components of organ grafts [31, 101]. Of note, these proteins may be induced in other (native) organs in the recipient as well as in the allograft [101]. These approaches have contributed to theories that, on the one hand, immunity to HSP promotes development of acute and chronic rejection [31, 104], per- haps via formation of antibodies [78]. On the other hand, these proteins may protect the graft from injury and/or enhance recovery [62, 138]. Recent studies suggest that HSPs indeed have the potential to modulate allograft rejection [14]. Flohe et al. [35] have found that expression of HSP 70 is a potential prognostic marker for acute rejection in human liver transplants. This complex area remains one of very active investigation to further define what is likely a delicate balance between injury and protective mecha- nisms [109], and immunohistology should continue to play a role in this evolving area. 146 L.C. Racusen

Fig. 3. Immunofluorescence stain for IgM in rejecting xenograft (courtesy of Dr. W. Baldwin)

Immunopathology of rejection

Xenogeneic organ grafts

The earliest attempted human organ transplants were xenografts of animal limbs and organs. These grafts quickly thrombosed and became necrotic following engraft- ment, in a “hyperacute” rejection response. While the vascular nature of the reaction to xenografts was recognized early, immunohistology became a tool which greatly advanced understanding of the mechanisms of loss of these grafts. Deposition of im- munoglobulins, primarily IgM, in the vasculature has been demonstrated in a variety of animal models of hyperacute xenograft rejection [41, 76, 107] (Fig. 3). Engage- ment of antibody with target antigen triggers fixation of complement, and recruit- ment of the coagulation cascade and platelet aggregation, responses preceding neu- trophil adherence and infiltration. Recent studies have shown that the classical com- plement pathway is activated in xenograft rejection [125]. The major targets of the xeno-reaction on the endothelium have been defined by immunohistochemistry. Gal1alpha-1gal3 is a ubiquitous endothelial antigen ex- pressed in several xeno-donor species; humans and a few other species lack this anti- gen and produce high titers of antibody against it. In addition to endothelial expres- sion, immunohistochemistry has also revealed high expression of this antigen on epi- thelial cells in kidney and lung. Absence of endothelial expression of this antigen can be demonstrated in animals genetically engineered to lack the gene for expres- sion of the antigen (reviewed in [125]). These animals could, in theory, serve as or- gan donor animals for xenografting into humans, avoiding the engagement of pre- formed antibody with the endothelium and the cascade of injury that follows. Unfortunately, while hyperacute rejection with near-immediate thrombosis due to preformed antibody can be avoided by several different strategies, an aggressive re- jection ensues in these models. This response is characterized by progressive cellular infiltrates and platelet and fibrin deposition in the microvasculature, and has proven relatively refractory to immunosuppressive therapy (reviewed in [53, 108]). Im- munohistology of experimental liver allografts with delayed “cellular” xenograft re- Immunopathology 147 jection has revealed graft infiltration by NK cells and CD4 lymphocytes [76]. If im- mediate xenograft rejection is prevented by immunosuppresssion, a delayed rejection occurs on withdrawal of immunosuppression. This rejection response may be charac- terized by different patterns in different organs, with an antibody-mediated process involving IgM, IgG and C3 in heart xenotransplants, but marked portal and peripor- tal cellular infiltrate, including lymphocytes, especially CD4+ cells, and monocytes, and with large numbers of Kupffer cells in liver [41]. Prevention of hyperacute rejec- tion by transplantation into complement-depleted animals has resulted in vascular IgM (but no complement) deposition, and infiltrates of granulocytes, monocytes, lymphocytes and small numbers of NK cells [23, 82]. Studies to further define and potentially avert or successfully treat this delayed form of xenograft rejection are ongoing, with the hope of removing this barrier to clinical use of xenografts. Immunopathology will doubtless remain an important in- vestigative tool in this area.

Allogeneic organ grafts

Inflammatory responses targeted against allo-antigens remain important causes of al- lograft injury and loss. These processes may be acute and clinically overt, acute and subclinical, or more chronic and indolent, ultimately leading to graft fibrosis. Once again, immunhistology can be used as a tool to define the nature of inflammatory re- sponses to the allograft, enhancing diagnosis of rejection in allografts, and informing and enabling development and tailoring of immunosuppressive therapy. Expression of target antigens by the tissues can also be elucidated, potentially identifying those at increased risk of rejection and chronic sequelae. In addition, the cytokine and che- mokine milieu within which these reactions take place can be defined, providing an understanding of relevant molecules that can be targeted by novel therapies.

Acute rejection

Acute rejection (AR) may be primarily antibody-driven, cell-mediated, or a combi- nation of both. Using appropriate antibodies and immunhistology, as outlined below, these complex reactions can be defined and appropriate therapy initiated to prevent graft damage and loss of function. Each of these types of AR is considered below.

Antibody-mediated rejection

As in the xenograft, antibody-mediated AR (AbAR) occurs in response to antigens expressed on endothelium of allografts. In large part due to improved definition and diagnosis, AbAR is now recognized in up to 20Ð30% of AR cases [18, 87]. AbAR may occur due to pre-formed anti-HLA antibody often present in individuals sensi- tized by previous antigen exposure via previous transplant, pregnancy or transfu- sions. Humans also have natural antibodies to major (and some minor) blood group antigens that their own cells do not express; these antigens are also expressed on en- dothelium. A solid organ allograft expressing these antigens will incite an immediate antibody-mediated rejection in recipients with pre-existing antibodies not depleted of 148 L.C. Racusen

Fig. 4. Immunoperoxidase stain for blood group A antigen, expressed uniformly on all endothelial cells antibodies prior to transplantation. Immunohistology can be used to identify target antigens along the endothelium (Fig. 4). AbAR reactions involve engagement of antibody with antigen, and fixation of complement with triggering of the complement cascade. Antibodies directed against HLA antigens are generally of the IgG class, while anti-ABO antibodies are of the IgM class. It is possible to demonstrate deposition of these antibodies along the en- dothelium in the acute phase of AbAR using immunohistology (e.g., [24, 52]). How- ever, deposition of identifiable antibody and of the early complement components is apparently evanescent, due in part to injury and loss of endothelium and presumably to clearance mechanisms, and these molecules are generally not detectable in clinical biopsy samples from organs undergoing AbAR (e.g., [24]). Later complement split products, and especially C3d and C4d, have proven to be much more reliable tissue immuno-markers of AbAR, probably because these factors bind covalently to the vessel wall (Fig. 5). Feucht et al. [34] were the first to report that immunostaining of peritubular capillaries for C4d in biopsy specimens from sen- sitized recipients with severe rejection correlated with increased graft loss . Collins et al. [24] subsequently demonstrated that C4d deposition in peritubular capillaries was invariably present in biopsies from patients with circulating de novo donor-specific antibodies (DSA) and either neutrophils in peritubular capillaries or arterial fibrinoid necrosis, identifying this immunostaining pattern as a useful marker for AbAR. A fol- low-up study of a cohort of 67 patients with AbAR confirmed this immunopathologic finding as a sensitive and specific marker for AbAR [88]. C4d immunostaining can appear and disappear within days on serial biopsies [100], and is variable over time [117]. Several series have shown that positive immunostaining for C4d correlates with poor graft survival [17, 58, 80, 118]. While typically detected by immunofluorescence on frozen tissue, an antibody is now available that can be used to detect C4d in paraf- fin sections. [118], expanding ability to study archival tissue specimens. This body of evidence has led to incorporation of immunostaining for C4d into the Banff schema for kidney allograft rejection as a defining feature of AbAR [112]. In heart allografts, there is increasing interest in refining criteria for diagnosing AbAR, and immunopathology potentially could play a major role in diagnosis [124]. Hammond et al. [52] evaluated morphological manifestations of AbAR in cardiac al- Immunopathology 149

Fig. 5. Immunofluorescence stain for C4d in a renal allograft with documented anti-donor antibody

Fig. 6. Immunoperoxidase stain for monocytes/macrophages in a case of antibody-mediated rejection lograft biopsy specimens; they found immunoglobulin and complement deposition in small vessels and HLA DR expression by immunostaining. A series of 44 patients have been described who had biopsy samples showing evidence of AbAR; criteria included positive immunostaining for IgG and C3 [92]. In another study, positive im- munostaining for C4d in surveillance biopsies performed in the first 3 months fol- lowing transplantion was shown to correlate with patient death [10], indirectly sug- gesting an unusual potentially refractory rejection component. Monocytes/macrophages have been proposed as a marker of rejection that may correlate with a poorer long-term outcome [26]. One potential explanation for this is the relatively recent recognition of an association of monocytes/macrophages and AbAR. In kidney allografts, macrophages have been shown by immunostaining to accumulate in peritubular capillaries and glomeruli in AbAR (Fig. 6), a finding that correlated strongly with C4d immunostaining [85]. Swollen macrophages have also been detected in capillaries in cardiac allografts with AbAR [83]. 150 L.C. Racusen

Cell-mediated rejection

Cell-mediated AR (CMAR) is characterized by infiltration of cells into interstitial ar- eas of organ grafts, and especially infiltration and/or engagement of effector cells with structural components within the tissues. Immunohistology can be used to de- fine expression of potential target antigens, as well as to define the nature of the in- filtrating cells. Indeed, immunostaining enables detection of very low levels of inter- stitial infiltrates. Activated and proliferating cells can also be demonstrated by im- munostaining, providing a means of identifying functionally active cells in the infil- trate, as well as reactive changes in parenchymal cells of the graft. While not a for- mal component of schemas defining criteria for diagnosing and grading rejection, immunostaining can also be very useful in equivocal cases to confirm the diagnosis of cell-mediated rejection. Since inflammatory cells can be detected in ischemic inju- ry and in stable allografts, defining the localization and nature of the infiltrate may be critical in evaluating allografts for CMAR [37], and immunostaining can make a valuable contribution to this process. Using immunohistochemistry, it has been shown that the nature of cell infiltration into allografts in CMAR is time dependent (reviewed in [37]). In early/reversible al- lograft rejection, CD4+ cells may be numerous, while CD8+ cells may predominate in established/irreversible CMAR. Based on immunostaining studies, Sanfilippo et al. [129] suggested that the intensity of CD8 staining in the cortex of kidney allo- grafts may be predictive of response to therapy; CD8+ cells have also been correlated with rejection grade in heart allografts [60]. In addition to basic cell phenotypes, pro- liferation and activation states of the cells can be defined using immunostaining for proliferation-associated antigens such as proliferating cell nuclear antigen (PCNA) and MIB-1 (Ki67) (e.g., [132]). For example, T cells with a CD45RO phenotype, in- dicating activated/memory cells, are more numerous than naïve cells in established AR in kidney [66], heart [67] and liver [65] allografts. Activated T cells can also be identified by immunostaining cells for expression of surface antigens such as the IL- 2 receptor (CD25) [135]. Detection of small numbers of such cells can be enhanced by multilayered immunostaining techniques (e.g., [55]) or use of gold enhancement of immunostaining [132]. Inducible lymphocyte costimulatory molecules can also be detected in allografts [55]. For example, in human renal allografts during acute rejec- tion, focal intense infiltrates of B7Ð1 and B7Ð2 cells (mainly CD20-Ð CD14+) and of CTLA-4 T lymphocytes (mainly CD8+) have been demonstrated. Scattered CD40 li- gand-positive (CD40L+) cells were detected as well, which exhibited CD4+ pheno- type [13]. A particularly strong (and specific) correlate of clinically overt AR is the pres- ence of cells expressing perforin and granzyme in the allograft [51, 70, 75]; these are lytic enzymes expressed by cytotoxic T cells which are presumed effectors of tissue injury in CMAR (Fig. 7). These findings are consistent with studies of molecular markers of AR, in which perforin and granzyme, as well as Fas ligand, were strongly associated with CMAR [140]. Cytotoxic cells can also be detected by immunostain for GMP-1, a cytotoxic granule protein (e.g., [91]). NK cells have also been shown to express granzymes A and B during AR episodes. B cells, NK cells, and monocytes/macrophages are present as minority popula- tions in rejection infiltrates (reviewed in [37]). While not proven to be major effec- tors in clinical allograft rejection, in some experimental models, NK cells of the TCRÐ, CD3Ð, NKRÐ phenotype have been shown to be alloreactive, and represent a Immunopathology 151

Fig. 7. Immunoperoxidase stain for perforin showing positive granules (arrow) in a cytotoxic cell infiltrat- ing a rejecting cardiac allograft. (courtesy of Dr. R. Hruban) larger proportion of rejection infiltrates [15, 106]. Rejection infiltrates characterized by immunohistology have also been shown to contain variable numbers of mono- cytes/macrophages. Indeed, monocytes/macrophages may play a pivotal role in CMAR [139]. They are capable of antigen presentation, and, upon activation, can produce tissue injury via release of free radicals and other toxic mediators, are the ef- fector cells in antibody-mediated cell cytotoxicity, and can activate the coagulation cascade. Cells harvested via vascular perfusion of native, isograft and allograft kidneys were examined by flow cytometry analysis; almost 75% of cells flushed from the vasculature of these experimental grafts were monocytes, most expressing an activated phenotype [45]. Croker et al. [26] have demonstrated in semiquantita- tive analysis of immunostained kidney allografts that macrophages predict graft sur- vival post transplant, and in another study, the presence of HLA-DR+ macrophages has been associated with poor prognosis in allografts [114]. Immunohistochemistry can be used to detect other monocyte activation markers, including NKR-P1, the ex- pression of which is increased in experimental AR [131]. Induction of HLA class I and class II antigens can be demonstrated in rejecting allografts, potentially serving to trigger/amplify the rejection response, and target re- jection infiltrates. In the kidney allograft, DR expression on tubular epithelium de- tected by immunostaining has been shown to correlate strongly with AR, and has been proposed as one of the most reliable markers for CMAR in renal allografts [100]. There is also evidence that expression of class II antigens and adhesion mole- cules on endothelium may serve as a surrogate marker of AR, correlating with the degree of CD3+ T cells in infiltrates [19, 28]. In addition to these markers, immuno- detection of altered coagulation factors and pro-inflammatory cytokines on the endo- thelium has been demonstrated [127]. Cells infiltrating the interstitium in allografts are less specific for rejection than cells that are actually targeting and infiltrating structural elements in the grafts (Fig. 8). Immunohistology has been very useful in characterizing these important effector cell populations. Semiquantitative assessment of CD57+ cells infiltrating tubules in kidney allografts has been shown to correlate with CMAR [12]. Double staining for CD3+ cells and Ki67 (MIB-1 proliferation marker) has been used to detect a popula- 152 L.C. Racusen

Fig. 8. Immunoperoxidase stain for T cells (CD3) in a rejecting renal allograft. Arrows mark cells infiltrat- ing into the tubular epithelium tion of proliferating intratubular T cells in AR. Robertson et al. [122] have localized CD8+ cells expressing perforin to the allograft tubules. Recent studies have identi- fied CD8+ TCR αβ+ cytotoxic cells expressing high levels of CD103 in allografts during rejection. Immunohistochemical analysis has demonstrated that these cells specifically home to the tubular epithelium; CD8 cells also exhibiting cytotoxic phe- notype but not expressing CD103 were predominantly restricted to the graft intersti- tium. These studies provide evidence that CD103, a beta 7 integrin, serves as a hom- ing receptor targeting CD8+ cytotoxic T lymphocytes to graft epithelium [49]. Anal- ogous mechanisms are doubtless present in other solid organ allografts. Targeting of structural elements via vectorial recruitment of immune cell popula- tions is achieved in part by proteoglycan-immobilized chemokines [54, 121]. Chemo- kines are a large family of molecules involved in leukocyte recruitment and the adap- tive immune response, and play important roles in rejection and organ allograft out- comes (see [25, 99] for recent reviews). These molecules and their receptors may be demonstrated and localized by immunohistochemistry in experimental and human or- gan transplants. For example, Eitner et al. [32] used in situ hybridization to detect large numbers of CXCR4-expressing cells in interstitial infiltrates in CMAR; double labeling for CD3 identified a large percentage of these cells as T lymphocytes. La- beled cells were also identified in neo-intimal T lymphocytes in vascular rejection. RANTES/CCL5 protein has been localized by immunostaining in 17 of 20 human kidney allografts with CMAR. The protein was localized to endothelium of peritubu- lar capillaries, infiltrating cells and renal tubular cells; RANTES/CCL5 mRNA was detected by in situ hybridization on infiltrating mononuclear cells and tubular epithe- lial cells. By elegantly combining indirect immunofluorescence staining and laser confocal microscopy, Robertson and Kirby [121] localized RANTES, MIP 1 α and MIP 1 β to graft tubular epithelium, with highest concentrations at the basolateral sur- face and within the basement membrane. Increased chemokine expression has also been demonstrated in experimental heart, lung and liver allografts [123, 150]; re- viewed in [99]. Immunostaining to assess expression of RANTES+ cells in AR in car- diac allografts may be useful in differentiating rejection from Quilty infiltrations [93]. Immunopathology 153

Cytokines within the allograft influence and reflect the nature of the infiltrates in immune reactions such as rejection (see [27, 69] for reviews). The cytokine milieu can be very complex, reflecting the multiple pathways to allograft rejection [81], and potentially reflecting other processes and even tolerizing infiltrates (see below). While often analyzed using molecular approaches, data on localization of expressed cytokines are critical for understanding the role of cytokines in the allograft. Immu- nostaining for cytokines can be performed on tissue fixed in acetone (although not in paraffin-embedded tissue), enabling documentation of expression and localization of cytokine expression to particular tissue components of the allograft, and to specific cell types with the use of double-labeling techniques (e.g., [23, 55]. Adhesion molecules, including LFA1-ICAM, VLA4-VCAM and the selectins play a potentially critical role in AR, facilitating interaction of inflammatory cells with endothelium and other target structures in allografts (for review see [137]). ICAM-1 is constitutively expressed on endothelium in normal kidney. In renal allo- grafts, expression on tubular epithelial cells, especially proximal tubular cells, is in- creased in rejection [5, 40]. This finding is somewhat nonspecific [5], but the obser- vation has been used to differentiate AR from drug toxicity in kidney allografts [33]. VCAM-1 expression is markedly increased in AR in capillaries, arteries and tubular epithelium in kidney [20]. In one study, intense staining was detected for both ICAM-1 and VCAM-1 in AR in heart allografts, while staining for ELAM-1 was rare; in liver allograft biopsy specimens, induction of ICAM-1, VCAM-1 and ELAM-1 was demonstrated in AR in vascular and sinusoidal endothelium [79]. While ICAM-1 and VCAM-1 were also induced in cytomegalovirus (CMV) infec- tion, induction of ELAM-1 on vascular endothelium was only seen in rejection. Immunohistochemistry may also be useful in defining differential diagnoses in grafts with infiltrates of inflammatory cells. With bacterial infection, B cells will be predominant. Immunohistochemistry can also be used to define post-transplant lym- phoproliferative disorders, which are usually a B cell proliferation; immunostaining may be used to define oligoclonal or monoclonal infiltrates, and in situ hybridization can be used to identify the Epstein-Barr Virus genome (EBV) (Fig. 9), which drives the disorder in the majority of cases. Some infiltrates may be protective/immuno- modulatory (see tolerance/accommodation below). As noted above, some infiltrates may also signal triggering of innate immunity/injury response [111], providing an additional rationale to identify cytotoxic cells and invasion of structural components by effector cells, features not seen in the injury response.

Chronic rejection

Fibrosing changes in allografts remain a common cause of graft loss in all vascular- ized organ allografts. There are many potential causes of fibrosis in allografts, in- cluding “true” chronic rejection (CR), i.e., fibrosing changes occurring due to previ- ous/ongoing alloimmune injury to the graft. The often mulifactorial etiology of the fibrosing changes may be difficult to define completely. Cellular infiltrates in fibros- ing allografts have been identified as predominantly T lymphocytes and mono- cytes/macrophages. Persistent CD3+ infiltrates have been identified in experimental models of CR [135], and in clinical allograft vasculopathy [128], presumably a marker of ongoing allo-immune injury. Both CD4+ and CD8+ cells have been dem- onstrated infiltrating grafts in experimental chronic allograft vasculopathy. In one 154 L.C. Racusen

Fig. 9. In situ hybridization for Epstein-Barr virus (EBER) in a kidney with post-transplant lymphoprolif- erative disorder. Arrows mark positive cells study, anti-chemokine therapy abrogated development of allograft vasculopathy, and in those animals a marked decline in CD4+ cells was detected by flow cytometry [153]. T cells and macrophages in CR in kidney allografts can be localized to glo- meruli as well as tubules using immunostaining [71]. Macrophages are central to initiation of fibrosis, producing a variety of fibrogenic factors and matrix proteins [97]. Macrophages are demonstrable localized in perivas- cular areas during early phases of clinical and experimental cardiac allograft vascu- lopathy (e.g., [128, 153]. Up-regulation of cytokines associated with macrophage ac- tivation (MCP-1, IFN-γ, IL-6) has been demonstrated in experimental cardiac CR; immunostaining localized the gene products to infiltrating mononuclear cells in the interstitium and vasculature [126]. Similar findings have been reported in renal allo- graft CR (e.g., [95]). Macrophages also precede fibrosis in CsA toxicity [152]. More recent studies have attempted to define allo-responses that may be actively promoting fibrosis in allografts by examining phenotypes of infiltrating cells in CR. For example, recent immunohistochemical studies using antibodies to co-stimulation factors have detected CD40L+CD4+ cells in infiltrates in CR, with intense expression of CD40 on endothelium; only a few B7-1 and B7-2 cells were detectable [13]. T cells in intima of affected arteries may express cytotoxic enzymes [36, 91]. There is now evidence that defining a component of ongoing antibody-mediated immune in- jury may enable specific therapy that can stabilize graft function. Mauiyyedi et al. [89] used immunofluorescence staining for C4d to identify a subgroup of patients with CR, defined by pathological and clinical criteria, in whom C4d staining was correlated with presence of anti-donor antibody. In a large cohort study, Regele et al. [117] have recently demonstrated C4d in about a third of late renal allograft biopsies. C4d staining was associated with basement membrane injury and multi-layering in glomerular and peritubular capillaries, changes pathognomonic for CR. Adhesion molecules can also be demonstrated in chronic rejection, potentially initiating and/or reflecting tissue injury. Immunostaining for VCAM-1 has been pro- posed as the strongest immunohistological correlate of CR [28, 61, 137]. VCAM ex- pression has been proposed as a surrogate marker for allograft vasculopathy [28]. Immunopathology 155

Expression on peritubular capillaries may signal ongoing inflammatory cell recruit- ment and injury [137], potentially culminating in thickening of capillary walls and eventual disappearance of capillaries. Increased expression of ICAM as well as VCAM may be demonstrated in chronic allograft vasculopathy [44]. Mechanisms of fibrogenesis in organ allografts have not been as well studied as, but presumably are analogous to, those producing fibrogenesis in native organs. Pathogenesis involves a variety of fibrogenic factors in the milieu, including trans- forming growth factor β (TGF-β), angiotensin II, and plasminogen activator inhibitor (PAI-1) ([135]; reviewed in [148]). Some studies have utilized immunostaining to confirm and localize expression of these factors. For example, TGF-β isoforms have been shown to be increased in CR as well as AR, either in tubulointerstitium or espe- cially in glomeruli and vessels [63, 133] in kidney, and in experimental cardiac allo- grafts [74, 149]. Fibroblast growth factor-1 (FGF-1) and receptor have been localized to glomerular lesions in kidney allografts with chronic rejection [71]. Platelet-de- rived growth factor (PDGF) A and B have been demonstrated by immunohistochem- istry and in situ hybridization in smooth muscle cells and macrophages in vessels with allograft vasculopathy [3]. More recent studies have demonstrated similar local- ized expression of the matricellular protein SPARC (secreted protein acidic and rich in cysteine) in vessels and interstitium, suggesting that SPARC could function as an accessory molecule in sclerosing changes in these vessels [4]. Chemokines also play a role in CR as well as acute inflammation and rejection in allografts [99]. Chemokines may contribute to interaction and activation of cells such as monocytes and endothelial cells, leading to the vasculopathic changes of CR. Ele- gant experimental studies have localized RANTES/CCL5 in infiltrating lymphocytes and macrophages in arterioles and venules adjacent to coronary arteries and in myo- fibroblasts within the neo-intima in graft atherosclerosis following cardiac transplan- tation [153]. Yun et al. [153] used multicolor flow cytometry to define the types of cells expressing RANTES in a murine model of allograft vasculopathy; NK, NKT and γδ cells, as well as lymphocytes and macrophages produced RANTES. As these examples illustrate, immunopathology is contributing significantly to this active area of transplant research. The development of fibrosing changes in allografts involves interstitial cells in the grafted organs, and especially fibroblasts, which differentiate during fibrogenesis into myofibroblasts. Identification of myofibroblasts is possible using immunostain- ing for α-actin. Proliferating myofibroblasts have been identified and localized by these techniques in experimental kidney allografts undergoing rejection. Shimizu et al. [135] have shown that myofibroblasts were markedly increased with progressive fibrosis and dysfunction; these cells were demonstrated to be proliferating using dou- ble-labeling for PCNA. Some of the actin-positive cells were in tubules, perhaps rep- resenting transdifferentiation of tubular cells. In another study, expression of fibro- blast-specific protein-1 has been used in vivo to demonstrate transdifferentiation of tubular epithelial cells into fibroblasts [141]. In human renal allografts, myofibro- blasts have been associated with progressive interstitial fibrosis [103]. Increased myofibroblasts are also seen in glomerular mesangium [73] and in the intima of ves- sels with allograft vasculopathy. MCP-1 expression demonstrated by immunohisto- chemistry on paraffin sections has been shown to parallel mononuclear cell recruit- ment as characterized by flow cytometry in a model of bronchiolitis obliterans [11]. Matrix components accumulating during fibrosis have not been extensively stud- ied in allografts, but a new focus on potential stabilization and reversal of fibrogene- 156 L.C. Racusen sis in the allograft will doubtless stimulate increased investigation in this area [113]. Collagen type III staining has been shown to progressively increase with time in hu- man renal allografts [73, 103]. De novo expression of laminin β2 and collagen 4α3 has been demonstrated at the proximal tubular cell basement membrane in CR of re- nal allografts [1]. Tenascin and the EDA isoform of fibronectin are expressed in glo- meruli in CR [43]; tenascin expression is also an early feature of the development of transplant renal arteriopathy in humans [144]. While still not completely defined, the nature of matrix components may ultimately be helpful in differentiating major caus- es of allograft fibrosis. In a study by Abrass et al. [1], for example, renal biopsy specimens with chronic cyclosporine A toxicity were found to have increased colla- gen I and III in the interstitium, in contrast to the pattern seen in CR. Based on these immunostaining studies, the authors suggest that the two processes may have a fun- damentally different pathogenesis. Finally, viral infections can lead to fibrosis in allografts, and viral infections have been implicated in pathogenesis of chronic allograft vasculopathy in a number of sol- id organ allografts (see below and [21]).

Tissue acceptance/accommodation

The development of tolerance involves antigen processing and transient interstitial infiltrates in animal models, and presumably in human organ allografts. Immunohis- tochemistry has defined features of those infiltrates in experimental models, which are similar to infiltrates in CMAR. In a miniature swine model following tolerizing strategies, some animals have progressive graft injury, some do not. In those that do not (i.e., appear tolerant), CD3+ T cells and macrophages are seen, and T cell subsets are present in proportions similar to those in rejection [134]. Immunostains for IL-2+ receptor have shown few activated T cells, and staining for PCNA is reduced, along with less apoptosis in parenchymal cells and persistent apoptosis in infiltrating cells when compared to grafts that were lost to rejection. Phenotypic analysis of infiltrat- ing cells by immunohistochemistry has revealed a dominance of CD8+ cells, with 78% single positive, and 19% CD4/CD8 double positive [16]. Only mild tubulitis and no vasculitis are seen in these reactions. Suppressor T cells, identifiable as CD4+/CD25+ cells, can be localized in organ allografts [90], and may be providing an important immunomodulatory function. These and other studies emphasize the need to carefully characterize and localize cell infiltrates in transplanted organs to avoid misinterpretation and suppression of potentially advantageous inflammation. Indeed, many allografts have cell infiltrates that appear not to be harming the allograft [37]. Immunohistochemistry may ulti- mately play a role in differentiating benign potentially “tolerizing” from injurious re- jection-related infiltrates. Cytokine profiles may also be useful in defining and un- derstanding tolerance [27].

Infection

Immunohistology plays a critical role in diagnosis of the infections that are so com- mon in immunosuppressed organ transplant recipients. This is a very broad topic, and the focus of this brief discussion will be on use of immunostaining techniques to Immunopathology 157 diagnose infections which directly involve the solid organ allografts. Infections may affect allografts indirectly via triggering an often systemic immune response which elicits a reaction in the organ allograft; this reaction in turn may prime the graft for allo-immune injury. Secondly, the allograft may be directly infected. Immunostain- ing can be very valuable in identifying organisms, and in particular viruses, infecting allografts. These viral infections may trigger not only acute inflammatory responses which are primarily T cell driven in the graft, but may lead to fibrosing changes in vessels and/or parenchyma of the organ graft. CMV is a common cause of allograft infection, varying among types of allografts. In experimental models, CMV has been shown to increase cytotoxic T cells, activa- tion of macrophages and expression of cell adhesion molecules [151], and CMV has been implicated in tissue priming for allograft rejection reactions. Histological find- ings in CMV infection in the allograft may be subtle, and immunostaining can be critical in identifying and defining the extent of infection. Rapid and accurate diag- nosis is particularly critical in CMV pneumonitis in the lung allograft, since effective anti-viral therapy can markedly reduce mortality rates [102]. For more chronic and indolent infection, in situ hybridization using probes for viral genomes may provide important insights into pathogenesis of processes such as chronic allograft vasculop- athy. Indeed, there is evidence potentially linking not only CMV, but adenovirus, en- terovirus, parvovirus and herpes simplex virus infections to cardiac allograft vascu- lopathy (see [21] for recent review). CMV may affect the severity of accelerated graft arteriosclerosis by augmenting vascular growth factors, directly altering the al- loimmune response, or altering cytokines and adhesion molecules [64]. EBV-associated post-transplant lymphoproliferative disorder (PTLD) can involve any solid organ allograft, and especially lung, intestine and pancreas allografts. Rec- ognition of this process and differentiation from allograft rejection is critical for proper management of the graft recipient (for review see [96]. The phenotype of in- filtrating cells can be defined using immunostaining; most cases of PTLD are B cell proliferations, while active acute cell-mediated rejection involves predominantly T cells. Clonality of the infiltrate can be assessed by immunostaining for kappa and lambda light chains. Finally, since most of these proliferations are EBV related, in situ hybridization on tissue sections for EBV antigens (commonly EBER) remains the gold standard for establishing the diagnosis [46] (Fig. 9). Polyoma virus (PPV) is a recently emerging pathogen in the renal allograft popula- tion. Known to commonly infect the urothelium, usually without serious sequelae, with modern immunosuppressive regimens, PPV-associated nephritis has emerged. PPVs, which include BK virus, JC virus, and SV40 virus, are members of the retrovi- rus family. Antibodies to the SV40 large T antigen are typically used to confirm pres- ence of the virus and determine extent of the infection in the kidney allograft (Fig. 10). When more specific probes are used, the exact viral pathogen can be defined; most cases are due to BK virus, some with co-infection with JC virus [7]. Immunostaining for PPV is used to confirm the presence of the virus and determine extent [30, 116]; the technique may not increase the sensitivity of diagnosis beyond suspicious findings on light microscopy for the experienced pathologist, but it may be useful in ruling out infection in cases with very reactive tubular cell changes [116]. Hepatitis C infection, and especially as recurrent disease, is a major problem in liver allografts. Unfortunately, there are currently no reliable immunoprobes to iden- tify hepatitis C in tissues. Hepatitis B infection can be diagnosed by immunostaining for any of several virus-associated antigens. 158 L.C. Racusen

Fig. 10. Immunoperoxidase stain for polyoma virus (large T antigen). Nuclei in some tubular cells stain strongly for the viral antigen

Fig. 11. Finely granular immunofluorescence staining for IgG along glomerular capillary walls in a patient with membranous glomerulopathy in a transplant kidney

Recurrent/de novo immune diseases The kidney allograft is particularly susceptible to development of recurrent or de novo immune complex disease, especially glomerulonephritis. Some diseases pro- ducing end-stage renal disease requiring renal replacement recur with high frequency in kidney allografts, and can lead to graft loss (see [22, 115] for reviews). IgA neph- ropathy, membranoproliferative glomerulonephritis (MPGN), membranous nephrop- athy, fibrillary glomerulopathy, dense deposit disease, anti-glomerular basement membrane disease, and lupus nephritis can all recur, and have characteristic findings on immunohistology. Immunohistology may be particularly critical in the differential diagnosis of MPGN and chronic allograft glomerulopathy, which is presumed to be caused by allo-injury to endothelium and has many histological and ultrastructural features in common with MPGN. Characteristic peripheral capillary loop staining for Immunopathology 159

C3 may be the only feature enabling accurate diagnosis [48]. The most common de novo immune complex disease in the kidney allograft is membranous nephropathy (Fig. 11). Some have suggested that slow release of antigens from the allograft may lead to membranous nephropathy in some patients. Other de novo immune complex glomerulonephritides may develop, including post-infectious glomerulonephritis, with immunohistology once again playing an important role in diagnosis. Finally, pauci-immune necrotizing glomerulonephritis may occur in the allograft; in these cases, the absence of staining on immunofluorescence helps establish the diagnosis. While serological testing is primarily used to diagnose recurrent hepatitis in the liv- er allograft, it has been proposed that immunoperoxidase staining of native liver for hepatitis B surface and core antigen may be useful in predicting recurrence of hepatitis B in liver allografts [86]. Recurrence of amyloidosis in organ allografts may be charac- terized by immunostaining for light chains or other amyloid-associated proteins.

Conclusion Immunopathology, as applied to tissue sections, aspirates, and other fluids from the allograft such as urine and bronchoalveolar lavage, continues to be an important di- agnostic modality, and to provide critical mechanistic insights regarding diseases aff- tecting organ allografts. These techniques are a critical complement to molecular studies, enabling localization at the tissue and cellular level. Technological advances will likely make immunopathology even more useful in the future.

References

1. Abrass CL, Berfield AK, Stehman-Breen C, et al (1999) Unique changes in interstitial extracellular matrix composition are associated with rejection and cyclosporine toxicity in human renal allograft bi- opsies. Am J Kidney Dis 33:11 2. Adoumie R, Serrick C, Giaid A, et al(1992) Early cellular events in the lung allograft. Ann Thorac Surg 54:1071 3. Alpers CE, Davis CL, Barr D, et al (1996) Identification of platelet-derived growth factor A and B chains in human renal vascular rejection. Am J Pathol 48:439 4. Alpers CE, Hudkins KL, Segerer S, et al (2002) Localization of SPARC in developing, mature, and chronically injured human renal algrafts. Kidney Int 62:2073 5. Andersen CB, Ladefoged SD, Larsen S (1994) Acute kidney graft rejection: a morphologic and im- munohistologic study of “zero hour” and follow-up biopsies with special emphasis on cellular infil- trates and adhesion molecules. APMIS 102:23 6. Andrews PD, Harper IS, Swedlow JR (2002) To 5D and beyond: quantitative fluorescence microscopy in the postgenomic era. Traffic 3:29 7. Baksh FK, Finkelstein SD, Swalsky PA, et al (2001) Molecular genotyping of BK and JC viruses in human polyoma virus-associated interstitial nephritis after transplantation. Am J Kidney Dis 38:354 8. Baldwin WM, Larsen CP, Fairchild RL (2001) Innate immune response to transplants. A significant variable with cadaver donors. Immunity 14:369 9. Baldwin WM, Samaniego-Picota M, Kasper EK, et al (1999) Complement deposition in early cardiac transplant biopsies is associated with ischemic injury and subsequent rejection episodes. Transplanta- tion 68:894 10. Behr TM, Richter K, Fischer P, et al (1998) Incidence of humoral rejection in heart transplant recipi- ents treated with tacrolimus or cyclosporin A. Transplant Proc 30:1920 11. Belperio JA, Keane MP, Burdick MD, et al (2001) Critical role for the chemokine MCP-1/CCR2 in the pathogenesis of bronchiolitis obliterans. J Clin Invest 108:547 12. Beschorner WE, Burdick JF, Williams GM, et al (1985) The presence of Leu 7 reactive lymphocytes in renal allografts undergoing acute rejection. Transplant Proc 17:618 160 L.C. Racusen

13. Biancone L, Segoloni G, Turello E, et al (1998) Expression of inducible lymphocyte costimulatory molecules in human renal allograft. Nephrol Dial Transplant 13:716 14. Birk OS, Gur SL, Elias D, et al (1999) The 60-dKA heat shock protein modulates allograft rejection. Proc Natl Acad Sci 96:5159 15. Blancho G, Buzelin F, Dantal J, et al (1992) Evidence that early acute renal failure may be mediated by CD3-CD16+ cells in kidney graft recipient with large granular lymphocyte proliferation. Transplan- tation 53:1242 16. Blancho G, Gianello PR, Lorf T, et al (1997) Molecular and cellular events implicated in local toler- ance to kidney allografts in miniature swine. Transplantation 63:26 17. Bohmig GA, Exner M, Habicht A, et al (2002) Capillary C4d deposition in kidney allografts: a specif- ic marker of alloantibody-dependent graft injury. J Am Soc Nephrol 13:1091 18. Bohmig GA, Exner M, Watschinger B, et al (2002) Acute humoral renal allograft rejection. Curr Opin Urol 12:95 19. Briscoe DM, Yeung AC, Schoen FJ, et al (1995) Predictive value of inducible endothelial cell adhe- sion molecule expression for acute rejection of human cardiac allografts. Transplantation 59:204 20. Brockmeyer C, Ulbrecht M, Schenckel DJ, et al (1993) Distribution of cell adhesion molecules (ICAM-1, VCAM-1, ELAM-1) in renal tissue during allograft rejection. Transplantation 55:610 21. Cainelli F, Vento S (2002) Infections and solid organ transplant rejection: a cause and effect relation- ship? Lancet Infect Dis 2:539 22. Cameron JS (1994) Recurrent renal disease after renal transplantation. Curr Opin Nephrol Hypertens 3:602 23. Candinas D, Belliveau S, Koyamada N, Miyatake et al (1996) T cell independence of macrophage and (NK) cell infiltration, cytokine production, and endothelial activation during delayed xenograft rejec- tion. Transplantation 62:1920 24. Collins AB, Schneeberger EE, Pascual MA, et al (1999) Complement activation in acute humoral re- nal allograft rejection: diagnostic significance of C4d deposits in peritubular capillaries. J Am Soc Nephrol 10:2208 25. Colvin BL, Thomson AW (2002) Chemokines, their receptors and transplant outcome. Transplantation 74:149 26. Croker BP, Clapp WL, Abu Shamat ARF, et al (1996) Macrophages and chronic renal allograft neph- ropathy. Kidney Int 50 (Suppl 57):S42 27. Dallman MJ (1993) Cytokines as mediators of organ graft rejection and tolerance. Curr Opin Immunol 5:788 28. Denton MD, Davis SF, Baum MA, et al (2000) The role of the graft endothelium in transplant rejec- tion: evidence that endothelial activation may serve as a clinical marker for the development of chron- ic rejection. Pediatr Transplant 4:252 29. Diaspro A, Robello M (2000) Two-photon excitation of fluorescence for three-dimensional optical im- aging of biological structures. J Photochem Photobiol 55:1 30. Drachenberg C, Beskow CO, Cangro CB, et al (1999) Human polyoma virus in renal allograft biop- sies; morphological findings and correlation with urine cytology. Hum Pathol 30:970 31. Duquesnoy RJ, Liu K, Fu XF, et al (1999) Evidence for heat shock protein immunity in a rat cardiac allograft model of chronic rejection. Transplantation 67:156 32. Eitner F, Cui Y, Hudkins KL, et al (1998) Chemokine receptor (CXCR4) m-RNA expressing leuko- cytes are increased in human renal allograft rejection. Transplantation 66:1551 33. Faull RJ, Russ GR (1990) Tubular expression of intercellular adhesion molecule-1 during renal allo- graft rejection. Transplnatation 48:226 34. Feucht HE, Schneeberger H, Hillebrand G, et al (1993) Capillary deposition of C4d complement frag- ment and early renal graft loss. Kidney Int 43:1333 35. Flohe S, Speidel N, Flach R, et al (1998) Expression of HSP70 as a potential prognostic marker for acute rejection in human liver transplantation. Transplant Int 11:89 36. Fox WMD, Hameed A, Hutchins GM, et al (1993) Perforin expression localizing cytotoxic lympho- cytes in the intima of coronary arteries with transplant related accelerated arteriosclerosis. Hum Pathol 24:477 37. Fuggle SV (1991) Immunphenotypic analysis of leukocyte infiltration in the renal transplant. Immunol Lett 29:143 38. Fuggle SV, McWhinnie DL, Chapman JR, et al (1986) Sequential analysis of HLA class II antigen ex- pression in human renal allografts. Induction of tubular class II antigens and correlation with clinical parameters. Transplantation 42:144 Immunopathology 161

39. Fuggle SV, McWhinnie DL, Morris PJ (1987) Precise specificity of induced tubular HLA class II anti- gens in renal allografts. Transplantation 44:214 40. Fuggle SV, Sanderson JB, Gray DW, et al (1993) Variation in expression of endothelial adhesion mol- ecules in pretransplant and transplanted kidneys Ð correlation of intragraft events. Transplantation 55:117 41. Ginesta MM, Mollevi DG, Ribas Y, et al (2002) Histology and immunopathology of heart and lung xenografts under low-dose tacrolimus. Transplant Proc 34:317 42. Goes N, Hobart M, Ramassar V, et al (1997) Many forms of renal injury induce a stereotyped re- sponse with increased expression of MHC, IFN-gamma, and adhesion molecules. Transplant Proc 29: 1085 43. Gould VE, Martinez-Lacabe V, Virtanen I, et al (1992) Differential distribution of tenascin and cellu- lar fibronectins in acute and chronic renal allograft rejection. Lab Invest 67:71 44. Gouldesbrough RA, Axelsen RA (1994) Arterial endothelialitis in chronic renal allograft rejection: a histopathological and immunocytochemical study. Nephrol Dial Transplant 9:35 45. Grau V, Stehling O, Garn H, et al (2001) Accumulating monocytes in the vasculature of rat renal allo- grafts; phenotype, cytokine inducible NO synthase, and tissue factor mRNA expression. Transplanta- tion 71:37 46. Gulley MJ (2001) Molecular diagnosis of Epstein Barr virus-related diseases. J Mol Diagn 3:1 47. Haas M, Ratner LE, Mongomery RA (2002) C4d staining of perioperative renal transplant biopsies. Transplantation 74:711 48. Habib R, Broyer M (1993) Clinical significance of allograft glomerulopathy. Kidney Int Suppl 43:S95 49. Hadley GA, Charandee C, Weir MR, et al (2001) CD103+ CTL accumulate within the graft epithelium during clinical renal allograft rejection. Transplantation 72:1548 50. Halloran PF, Homick J, Goes N, et al (1997) The “injury response”: a concept linking non-specific in- jury, acute rejection, and long-term transplant outcomes. Transplant Proc 29:79 51. Hameed A, Truong LD, Price V, et al (1991) Immunohistochemical localization of granzyme B anti- gen in cytotoxic cells in human tissues. Am J Pathol 138:1069 52. Hammond EH, Yowell RL, Nunoda S, et al (1989) Vascular (humoral) rejection in heart transplanta- tion: pathologic observations and clinical implications. J Heart Transplant 8:430 53. Hancock WW (1997) Delayed xenograft rejection. World J Surg 21:917 54. Hancock WW (2002) Chemokines and transplant immunobiology. J Am Soc Nephrol 13:821 55. Hancock WW, Sayegh MH, Zheng X-G, et al (1996) Costimulatory function and expression of CD40L, CD80 and CD8 in vascularized murine cardiac allograft rejection. Proc Natl Acad Sci USA 93:13967 56. Hankey KG, Drachenberg CB, Papadimitrious JC, et al (2002) MIC expression in renal and pancreatic allografts. Transplantation 73:30 57. Herskowitz A, Mayne AE, Willoughby SB, et al (1994) Patterns of myocardial cell adhesion molecu- lar expression in human endomyocardial biopsies after cardiac transplantation: induced ICAM-1 and VCAM-1 related to implantation and rejection. Am J Pathol 145:1082 58. Herzenberg AM, Gill JS, Djurdjev O, et al (2002) C4d deposition in acute rejection: an independent long-term prognostic factor. J Am Soc Nephrol 13:234 59. Herzenberg LA, Parks D, Sahaf B, et al (2002) The history and future of the fluorescence activated cell sorter and flow cytometry. Clin Chem 48:1819 60. Higuchi ML, Assiss RV de, Sambiase NV, et al (1991) Usefulness of T-cell phenotype characteriza- tion in endomyocardial biopsy fragments from human cardiac allografts. J Heart Lung Transplant 10:235 61. Hill PA, Main IW, Akins RC (1995) I-CAM-1 and V-CAM-1 in human renal allograft rejection. Kid- ney Int 47:1383 62. Hiratsuka M, Yano M, Mora BN, et al (1998) Heat shock pretreatment protects pulmonary isografts from subsequent ischemia-reperfusion injury. J Heart Lung Transplant 17:123 63. Horvath LZ, Freis H, Schilling M, et al (1996) Altered expression of transforming growth factor-betas in chronic renal rejection. Kidney Int 50:489 64. Hosenpud JD (1999) Coronary artery disease after heart transplantation and its relation to cytomegalo- virus. Am Heart J 138:S469 65. Ibrahim S, Dawson DV, Killenberg PG, et al (1993) The pattern and phenotype of TÐcell infiltration associated with human liver allograft rejection. Hum Pathol 24:1365 66. Ibrahim S, Dawson DV, Sanfilippo F (1995) Predominant infiltration of rejecting human renal allo- grafts with T cells expressing CD8 and CD45RO. Transplantation 59:724 162 L.C. Racusen

67. Ibrahim S, Dawson DV, Van Tryt P, et al (1993) Differential infiltration by CD45RO and CD45RA subsets of T cells associated with human heart allograft rejection. Am J Pathol 142:1794 68. Ibrahim S, Jacobs F, Zukin Y, et al (1996) Immunohistochemical manifestations of unilateral kidney ischemia. Clin Transplant 10: 646 69. Kamoun M (2001) Cellular and molecular parameters in human renal allograft rejection. Clin Bio- chem 34:29 70. Kataoka K, Naomoto Y, Shiozaki S, et al (1992) Infiltration of perforin-positive mononuclear cells in- to the rejected renal allograft. Transplantation 53:240 71. Kerby JD, Verran DJ, Luo KL, et al (1996) Immunolocalization of FGF-1 and receptors in human re- nal allograft vasculopathy associated with chronic rejection. Transplantation 62:467 72. Knowlton AA, Brecher P, Apstein CS (1991) Rapid expression of heat shock in rabbit after brief car- diac ischemia. J Clin Invest 87:139 73. Ko YJ, Sugar L, Zaltzman J, et al (1997) Alpha-smooth muscle actin and collagen deposition in chronic allograft rejection. Transplantation 63:156 74. Korom S, Hancock WW, Coito AJ, et al (1998) Blockade of very late antigen-4 integrin binding to fi- bronectin in allograft recipeints II. Treatment with connecting segment-1 peptides prevents chronic re- jection by attenuating arteriosclerotic development and suppressing intragraft T cell and macrophage activation. Transplantation 65:854 75. Kummer JA, Wever PC, Kamp AM, et al (1995) Expression of granzyme A and B proteins by cyto- toxic lymphocytes involved in acute renal allograft rejection. Kidney Int 47:70 76. Langer A, Valdivia LA, Murase N, et al (1993) Humoral and cellular immunopathology of hepatic and cardiac hamster-into-rat xenograft rejection. Am J Pathol 143:85 77. Laskowski I, Pratschke J, Wilhelm MJ, et al (2000) Molecular and cellular events associated with ischemia/reperfusion injury. Ann Trasnplant 5:29 78. Latif N, Yacoub MH, Dunn MJ (1997) Association of pre-transplant anti-heart antibodies against heat shock protein 60 with clinical course following cardiac transplant. Transplant Proc 29:1039 79. Lauterschlager I, Hockerstedt K, Taskinen E, et al (1996) Expression of adhesion molecules and their ligands in liver allografts during cytomegalovirus (CMV) infection and acute rejection. Transplant Int 9 (Suppl 1):S213 80. Lederer SR, Kluth-Pepper B, Schneeberger H, et al (2001) Impact of humoral alloreactivity early after transplantation on long-term survival of renal allografts. Kidney Int 59:334 81. LeMoine A, Goldman M, Abramowicz D (2002) Multiple pathways to allograft rejection. Transplan- tation 73:1373 82. Leventhal JR, Matas AJ, Sun LH, et al (1993) The immunopathology of cardiac xenograft rejection in the guinea pig-to-rat model. Transplantation 56:1 83. Lones MA, Czer LS, Trento A, et al (1995) Clinical-pathologic features of humoral rejection in cardi- ac allografts: a study in 81 consecutive patients. J Heart Lung Transplant 14:151 84. Lu CY, Penfield JG, Kielar ML, et al (1999) Hypothesis: is renal allograft rejection initiated by the re- sponse to injury sustained during the transplant process? Kidney Int 55:2157 85. Magil AB, Tinckam K (2003) Monocytes and peritubular capillary C4d deposition in acute renal allo- graft rejection. Kidney Int 63:1888 86. Mann SP, Painter DM, McCaughan GW (1994) Histopathological prediction of hepatitis B recurrence in liver allografts. Pathology 26:251 87. Mauiyyedi S, Colvin RB (2002) Humoral rejection in kidney transplantation: new concepts in diagno- sis and treatment. Curr Opin Nephrol Hypertens 11:609 88. Mauiyyedi S, Crespo M, Collins AB, et al (2002) Acute humoral rejection in kidney transplantation. II. Morphology, immunopathology and pathological classification. J Am Soc Nephrol 13:779 89. Mauiyyedi S, Pelle PD, Saidman S, et al (2001) Chronic humoral rejection: identification of antibody- mediated chronic renal allograft rejection by C4d deposits in peritubular capillaries. J Am Soc Nephr- ol 12:574 90. McHugh RS, Shevach EM (2002) The role of suppressor T cells in regulation of immune responses. J Allergy Clin Immunol 110:693 91. Meehan SM, McCluskey RT, Pascual M, et al (1997) Cytotoxicity and apoptosis in human renal allo- grafts: identification, distribution and quantitation of cells with a cytotoxic granule protein GMP-17 (TIA1) and cells with fragmented nuclear DNA. Lab Invest 76:639 92. Michaels PJ, Esperjo BA, Kobashigawa J, et al (2002) Humoral rejection in cardiac transplants; risk factors, hemodynamic consequences and relationship to transplant coronary artery disease. J Heart Lung Transplant 22:58 Immunopathology 163

93. Michaels PJ, Kobashigawa J, Laks H, et al (2001) Differential expression of RANTES chemokine, TGF-beta and leukocyte phenotype in acute cellular rejection and Quilty B lesions. J Heart Lung Transplant 20:407 94. Milton AD, Spencer SC, Fabre JW(1986) Detailed analysis and demonstration of differences in ki- netics of induction of class I and II major histocompatibility antigens in rejecting cardiac and kidney allografts. Transplantation 41:499 95. Nadeau KC, Azuma H, Tilney NL (1995) Sequential cytokine dynamics in chronic rejection of rat renal allografts: roles for RANTES and MCP-1. Proc Natl Acad Sci USA 92:8729 96. Nalesnik MA (2002) Clinicopathologic characteristics of post-transplant lymphoproliferative disor- ders. Recent Results Cancer Res 159:9 97. Nathan CF (1987) Secretory products of macrophages. J Clin Immunol 79:319 98. Navarro F, Portales P, Candon S, et al (2000) Natural killer cells and alphabeta and gammadelta lymphocytes traffic into the liver graft immediately after liver transplantation. Transplantation 69:633 99. Nelson PJ, Krensky AM (2001) Chemokines, chemokine receptors and allograft rejection. Immunity 14:377 100. Nickeleit V, Zeiler M, Gudat F, et al (2002) Detection of the complement degradation product C4d in renal allografts: diagnostic and therapeutic implications. J Am Soc Nephrol 13:242 101. Ogita K, Hopkinson K, Nakao M, et al (2000) Stress responses in graft and native intestine after rat heterotopic small bowel transplantation. Transplantation 69:2273 102. Paradis IL, Williams P (1993) Infection after lung transplantation. Semin Respir Infect 8:207 103. Pedagogos E, Hewitson TD, Walker RG, et al (1997) Myofibroblast involvement in chronic trans- plant rejection. Transplantation 64:1192 104. Perdrizet GA (1996) The heat shock response and organ transplantation. Transplant Rev 10:78 105. Penfield JG, Wang Y, Li S, et al (1999) Transplant surgery injury recruits recipient MHC class II- positive leukocytes into the kidney. Kidney Int 56:1759 106. Petersson E, Ostraat O, Ekberg H, et al (1997) Allogeneic heart transplantation activates alloreactive NK cells. Cell Immunol 175:25 107. Platt JL, Fischel RJ, Matas AJ, et al (1991) Immunopatholgy of hyperacture xenograft rejection in a swine-to-primate model. Transplantation 52:214 108. Platt JL, Lin SS, McGregor CGA (2001) Acute vascular rejection. Xenotransplantation 5:169 109. Pockley AG (2001) Heat shock proteins, anti-heat shock protein reactivity and allograft rejection. Transplantation 71:1503 110. Qian Z, Wasawska BA, Behrens E, et al (1999) C6 produced by macrophages contributes to cardiac allograft rejection. Am J Pathol 155:1293 111. Rabb H (2002) The T cell as a bridge between innate and adaptive immune systems: implications for the kidney. Kidney Int 61:1935 112. Racusen L, Colvin RB, Solez K, et al (2003) Antibody-mediated rejection criteria Ð An addition to the Banff classification of renal allograft rejection. Am J Transplant 3:708Ð714 113. Racusen L, Solez K, Colvin R (2003) Fibrosis and atrophy in the renal allograft Ð Interim report and new directions. Am J Transplant 2:203 114. Raferty M, Seron D, Koffman G, et al(1989) The relevance of induced class II HLA antigens and macrophage infiltration in early renal allograft biopsies. Transplantation 48:238 115. Ramos E, Tisher CC (1994) Recurrent diseases in the kidney transplant. Am J Kidney Dis 24:142 116. Randhawa PS, Demetris AJ (2000) Nephropathy due to polyomavirus type BK. N Engl J Med 342:1361 117. Regele H, Bohmig GA, Habicht A, et al (2002) Capillary deposition of complement split product C4d in renal allografts is associated with basement membrane injury in peritubular and glomerular capillaries: A contribution of humoral immunity to chronic allograft rejection. J Am Soc Nephrol 13:2371 118. Regele H, Exner M, Watschinger B, et al (2001) Endothelial C4d deposition is associated with in- ferior kidney allograft outcome independently of cellular rejection. Nephrol Dial Transplant 16:2058 119. Reits EA, Neefjes JJ (2001) From living cells to PRAP: measuring protein mobility and activity in living cells. Nat Cell Biol 3:E145 120. Ribeiro-David DS, David-Neto E, Castro MC, et al (1998) Contribution of the expression of ICAM- 1, HLA-DR and IL-2R to the diagnosis of acute rejection in renal allograft aspiration cytology. Transplant Int 11(Suppl 1):S19 164 L.C. Racusen

121. Robertson H, Kirby JA (2003) Post-transplant tubulitis: the recruitment, differentiation and persis- tence of intra-epithelial T cells. Am J Transplant 3:3 122. Robertson H, Wheeler J, Kirby JA, et al (1996) Renal allograft rejection: in situ demonstration of cy- totoxic intratubular cells. Transplantation 61:1546 123. Robinson LA, Nataraj C, Thomas DW, et al (2000) A role for fractaline and its receptor in cardiac al- lograft rejection. J Immunol 165:6067 124. Rodriguez ER (2003) The pathology of heart transplantation. Revisiting the 1990 ISHLT working formulation. J Heart Lung Transplant 21:3 125. Roos A, Daha MR (2002) Antibody-mediated activation of the classical complement pathway in xe- nograft rejection. Transplant Immunol 9:257 126. Russell ME, Wallace AF, Hancock WW, et al (1995) Upregulation of cytokines associated with mac- rophage activation in the Lewis-to-F344 rat transplantation model of chronic cardiac rejection. Transplantation 59:572 127. Salom RN, Maguire JA, Hancock WW (1998) Endothelial activation and cytokine expression in hu- man acute cardiac allograft rejection. Pathology 30:24 128. Salomon RN, Hughes CC, Schoen FJ, et al (1991) Human coronary transplantation-associated arter- iosclerosis. Evidence for a chronic immune reaction to activated graft endothelial cells. Am J Pathol 138:791 129. Sanfilippo F, Kolbeck PC, Vaughn WK, et al (1985) Renal allograft cell infiltrates associated with ir- reversible rejection. Transplantation 40:679 130. Schwarz C, Regele H, Steininger R, et al (2001) The contribution of adhesion molecular expression in donor kidney biopsies to early allograft dysfunction. Transplantation 71:1666 131. Scriba A, Grau V, Steiniger B (1998) Phenotype of rat monocytes during acute kidney allograft rejec- tion: increased expression of NKR-P1 and reduction of CD43. Scand J Immunol 47:332 132. Seron D, Alexopoulos E, Raftery MJ, et al (1989) Diagnosis of rejection in renal allograft biopsies using the presence of activated and proliferating cells. Transplantation 47:811 133. Shihab FS, Yamamoto T, Nast CC, et al(1995) Transforming growth factor-beta and matrix protein expression in acute and chronic rejection of human renal allografts. J Am Soc Nephrol 6:286 134. Shimizu A, Yamada K, Meehan SM, et al (2002) Acceptance reaction: intragraft events associated with tolerance to renal allografts in miniature swine. J Am Soc Nephrol 11:2371 135. Shimizu A, Yamada K, Sachs DH, et al (2002) Persistent rejection of peritubular capillaries and tu- bules is associated with progressive interstitial fibrosis. Kidney Int 61:1867 136. Shoskes D, Parfery NA, Halloran PF (1990) Increased major histocompatibility complex antigen ex- pression in unilateral ischemic acute tubular necrosis in the mouse. Transplantation 49:201 137. Solez K, Racusen LC, Abdulkareem F, et al (1997) Adhesion molecules and rejection of renal allo- grafts. Kidney Int 51:1476 138. Squiers EC, Bruch D, Buelow R, et al (1999) Pretreatment of small bowel isografts donors with co- balt-protoporphyrin decreases preservation injury. Transplant Proc 31:585 139. Steiniger B, Stehling O, Scriba A, et al V (2001) Monocytes in the rat: phenotype and function dur- ing acute allograft rejection. Immunol Rev 184:38 140. Strehlau J, Pavlakis M, Lipman M, et al (1997) Quantitative detection of immune activation transcripts as a diagnostic tool in kidney transplantation. Proc Natl Acad Sci USA 94:695 141. Strutz F, Muller GA, Neilson EG (1996) Transdifferentiation: a new angle on renal fibrosis. Exp Nephrol 4:67 142. Tailor A, Granger DN (2000) Role of adhesion molecules in vascular regulation and damage. Curr Hypertens Rep 2:78 143. Takada M, Nadeau KC, Shaw GD, et al (1997) The cytokine-adhesion molecule cascade in isch- emia/reperfusion injury of the rat kidney: inhibition by a soluble P-selectin ligand. J Clin Invest 99:2682 144. Tanabe S, Ueda M, Han YS, et al (1996) Increased tenascin expression is an early feature of the de- velopment of transplant renal arteriopathy in humans. Transplant Int 9 (Suppl 1):S45 145. Udelsman R, Blake MJ, Holbrook NJ (1991) Molecular response to surgical stress; specific and si- multaneous heat shock protein induction in the adrenal cortex, aorta and vena cava. Surgery 110:1125 146. Van Noorden S (2002) Advances in immunocytochemistry. Folia Histochem Cytobiol 40:121 147. Waiser J, Schwaar S, Bohler T, et al (2002) Immunohistochemical double-staining of renal allograft tissue: critical assessment of three different protocols. Virchows Arch 440:648 148. Waller JR, Nicholson L (2001) Molecular mechanisms of renal allograft fibrosis. Br J Surg 88:1429 Immunopathology 165

149. Waltenberger J, Wanders A, Fellstrom B, et al (1993) Induction of transforming growth factor-beta during cardiac allograft rejection. J Immunol 151:1147 150. Yamaguchi Y, Ichiguchi O, Matsamura F, et al (1997) Enhanced expression of cytokine-induced neu- trophil chemoattractant in rat hepatic allografts during acute rejection. Hepatology 26:1546 151. Yilmaz S, Koskinen PK, Kallio E, et al (1996) Cytomegalovirus-enhanced chronic kidney allograft rejection is linked with intercellular adhesion molecule-1 expression. Kidney Int 50:526 152. Young BA, Burdmann EA, Johnson RJ, et al (1995) Cellular proliferation and macrophage influx precede interstitial fibrosis in cyclosporine toxicity. Kidney Int 48:439 153. Yun JJ, Fischbein MP, Whiting D, et al (2002) The role of MIG/CXCL9 in cardiac allograft vascu- lopathy. Am J Pathol 161:1307