SEARCH

SEARCH BY CITATION

Keywords:

  • Kidney allograft;
  • kidney transplantation;
  • rejection;
  • renal allograft;
  • renal allograft rejection;
  • T cell;
  • transplantation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

In kidney allografts, T cell mediated rejection (TCMR) is characterized by infiltration of the interstitium by T cells and macrophages, intense IFNG and TGFB effects, and epithelial deterioration. Recent experimental and clinical studies provide the basis for a provisional model for TCMR. The model proposes that the major unit of cognate recognition in TCMR is effector T cells engaging donor antigen on macrophages. This event creates the inflammatory compartment that recruits effector and effector memory CD4 and CD8 T cells, both cognate and noncognate, and macrophage precursors. Cognate T cells cross the donor microcirculation to enter the interstitium but spare the microcirculation. Local inflammation triggers dedifferentiation of the adjacent epithelium (e.g. loss of transporters and expression of embryonic genes) rather than cell death, via mechanisms that do not require known T-cell cytotoxic mechanisms or direct contact of T cells with the epithelium. Local epithelial changes trigger a response of the entire nephron and a second wave of dedifferentiation. The dedifferentiated epithelium is unable to exclude T cells, which enter to produce tubulitis lesions. Thus TCMR is a cognate recognition-based process that creates local inflammation and epithelial dedifferentiation, stereotyped nephron responses, and tubulitis, and if untreated causes irreversible nephron loss.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

T-cell-mediated rejection (TCMR) is an important event in organ transplantation and a model for T-cell-mediated inflammatory diseases. With contemporary immunosuppression, TCMR is less frequent but remains the dominant early rejection phenotype and the end point for clinical trials. The key to TCMR is inflammation. Carrel described kidney transplant rejection (1) and noted in his Nobel lecture: ‘the transplanted kidneys … presented some lesions … of diffuse nephritis ….’ Da Fano in 1912 observed that the infiltrate in rejecting mouse tumor allografts consisted of lymphocytes and macrophages, and the lymphocytes were not proliferating, anticipating the concept that the inflammatory compartment is not the site of clonal expansion and effector generation (2). Mitchison (3) showed that primed lymphocytes, subsequently shown to be T cells, could adoptively transfer rejection, as had previously been shown for delayed-type hypersensitivity. Thus the key elements in TCMR of organ allografts are infiltration of the donor tissue interstitium by host T cells and macrophages, followed by deterioration of the donor tissue. (Note that the term ‘macrophages’ here includes a range of cell types including monocytes and dendritic cells in various stages of maturation and activation.)

We have recently had the opportunity to study thousands of mouse and human organ transplants by conventional assessments and microarrays, creating large amounts of new data and suggesting new concepts (http://transplants.med.ualberta.ca). The present article summarizes the lessons from these studies and proposes a provisional framework, incorporating the new data from mouse and human kidney allografts (Table 1). The model aims to explain how molecular mechanisms create the diagnostic histopathology lesions – interstitial infiltrate and tubulitis (4) – incorporating earlier basic and clinical observations with our recent observations. Critical elements that shape the model are (a) the distinction between TCMR and antibody-mediated rejection (ABMR), giving a clearer picture of true TCMR; (b) observations in mouse kidney allografts that manifest TCMR lesions; and (c) the emerging knowledge from studies of gene expression (5–8). The key novel points are outlined in Table 2 and expanded in the text. I have not focused on endarteritis because it is an ambiguous lesion, reflecting either TCMR or ABMR. I will not discuss the generation of effector T cells.

Table 1.  The model for T-cell-mediated rejection of kidney allografts
1. The cognate recognition unit in TCMR in the allograft: the effector T cells encountering donor antigens on interstitial host or donor antigen-presenting cells, creating a local zone of inflammation
2. Individual units explain the microspecificity of TCMR
3. Injury is not essential to initiate TCMR: injury effects are additive
4. TCMR triggers loss of function by parenchymal dedifferentiation, not cell death
5. Parallel human and mouse TCMR molecular picture in microarrays:
   a. T-cell burden and IFNG effects
   b. Macrophage burden and alternative macrophage activation
   c. Stereotyped injury response
       i. Increased embryonic and developmental transcripts
       ii. Loss of transcripts defining differentiated epithelium, e.g. solute carriers
       iii. Altered endothelial transcripts due to remodeling of the microcirculation
6. Mouse model with lesions such as human TCMR permits mechanistic insights:
   a. Early perivascular T-cell infiltrate by day 1 with IFNG effects
   b. Rising T-cell burden and IFNG effects plateau by day 7, then sustained
   c. Progressive time-dependent epithelial deterioration and loss of integrity lead to tubulitis because epithelium can no longer exclude T cells
   d. TCMR does not require host B cells/antibody, granzymes, perforin, CD103 or donor Fas
7. Examples of unsolved problems
   a. What T cells engage antigen in the graft at day 1?
   b. How do cognate T cells cross the donor microcirculation without destroying it?
   c. How can we identify cognate T cells engaging donor antigens?
   d. How many infiltrating T cells are noncognate and what is their role?
   e. Can specialized antigen-presenting cells be defined?
   f. Is there a role for regulatory T cells in the graft (as opposed to the lymphoid organs)?
Table 2.  New concepts of T-cell-mediated rejection
 1. TCMR in kidney is defined by lesions (infiltrate (i), tubulitis (t), endothelialitis (v)) and molecules
 2. The interstitial infiltrate is stereotyped: effector and effector memory CD4 and CD8 T cells; macrophages and dendritic cells with features of classical and alternative activation
 3. Immunosuppression and antigen differences affect probability and intensity but not quality
 4. TCMR does not lead to ‘chronic rejection’. Separation of TCMR from ABMR has clarified that the phenomenon of ‘chronic rejection’ is due to ABMR
 5. Early phase of TCMR (24 hours in experimental kidney and heart transplants) may reflect specialized peripheral surveillance memory T cells that patrol normal tissues
 6. IFNG and donor class I MHC molecules stabilize early TCMR. Experimental organ allografts lacking host IFNG, donor IFNGR, donor IRF1, or donor class I develop early microcirculation failure with ischemic necrosis, as well as increased AMA
 7. Effector and effector memory T cells have similar transcriptomes whether CD4 or CD8, and whether antigen triggered or not
 8. Non-cognate T cells attracted to the inflammatory compartment by cognate T cells contribute to TCMR through innate immunity functions such as IFNG production, triggered by local cytokines
 9. The response of renal epithelium to TCMR is a preprogrammed stereotyped injury response shared with other types of injury
10. Nephrons respond to local changes in their epithelium. The snake-like nephron can shut down pending repair, or permanently with atrophy and fibrosis if repair fails
11. Promptly treated TCMR is relatively benign even with v lesions or occurring late, if ABMR is absent
12. Treated TCMR may leave some permanent nephron loss (atrophy and fibrosis), but this is not progressive

A key point in our mechanistic understanding is that the lesions and molecular changes of TCMR that develop in human renal transplants despite immunosuppression are reproduced in nonimmunosuppressed mice receiving kidney allografts incompatible for major histocompatibility complex (MHC) plus non-MHC antigens (9–15). Immunosuppressive drugs act mainly on effector generation. Our experience has been that immunosuppressive drugs such as cyclosporine reduce the incidence and severity of TCMR in mouse models but do not qualitatively alter the histologic and molecular features of TCMR when it develops. TCMR is stereotyped: when it arises despite immunosuppression, it is qualitatively the same as on no immunosuppression (an exception may be TCMR developing in humans after intense T-cell-depleting therapy – see below).

Mouse kidney allografts do not manifest lesions of ABMR during the first 2–3 weeks, permitting us to explore TCMR mechanisms without ABMR. TCMR in this model is very destructive: most B6 allografts into nephrectomised CBA hosts fail due to rejection by about 3 weeks, with a minority (about 20%) spontaneously surviving, usually with continuing histologic abnormalities (unpublished observations). For pathology studies, the mouse model is performed with one host kidney left in place to stabilize the host and avoid premature death, which would preclude histopathology studies.

The challenge of studying the mechanisms operating within the rejecting kidney

The separation of effector T-cell generation from the inflammatory compartment poses challenges for studies that focus on the mechanisms that operate within the inflammatory compartment, distal to effector generation. Many molecules participate in rejection at multiple levels (e.g. effector generation, effector homing and cognate triggering in the graft), creating a problem in defining a distinct role at any one level. Moreover, it is easy to disrupt effector generation (i.e. to create a degree of immunodeficiency), but it is difficult to disrupt the TCMR mechanisms in the graft while sparing effector generation. Adoptive transfer of primed T cells can ensure a supply of effector cells, but whether it truly recapitulates the pathogenesis of TCMR is not known. New experimental systems are required to target the specific mechanisms operating in the allograft without compromising effector generation.

Creation of the Inflammatory Compartment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

We postulate that the fundamental unit of cognate recognition in the graft is probably the infiltrating effector T cell encountering host or donor cells with antigen-presenting abilities, creating a local inflammatory zone or compartment (Figure 1A). The antigen-specific effector T cells include naive T cells primed in the secondary lymphoid organs and memory T cells responding by heterologous memory, after being initially primed by infections (16). The site of memory cell reactivation and clonal expansion is not clear, but memory T cells are less reliant on secondary lymphoid organs (17).

imageimage

Figure 1. Evolving model of mechanisms of epithelial deterioration. TCMR is mediated by a local process analogous to delayed-type hypersensitivity, independent of contact between effector T cells and epithelial cells. (A) Cognate interactions between antigen-specific effector T cells and host (or donor) antigen-presenting cells could trigger release of soluble factors, either by effector T cells, by macrophages and dendritic cells (DC) activated by effector T cells, and possibly by noncognate effector memory T cells, which induce local epithelial changes either through direct effects on the renal parenchyma or effects on the matrix and microcirculation. The key change is epithelial dedifferentiation. (B) Dedifferentiation of the epithelium with loss of cadherins, intercellular junctions, polarity and epithelial integrity eventually permits entry of lymphocytes, causing tubulitis lesions.

The TCMR mechanisms operating in the graft produce stereotyped lesions and molecular changes despite differences in the mechanisms for antigen presentation or antigenic disparities. While MHC products activate naïve T cells by direct, indirect or semi-direct presentation and recognition (18), these details of antigen presentation do not seem to influence the lesions of TCMR in the graft. The Banff TCMR lesions are the same whether TCMR occurs five days or 30 years posttransplant, arguing that priming by distinct presentation mechanisms leads to the same effector phase in the graft. Moreover, in clinical renal transplantation, TCMR lesions have not been shown to be influenced by the nature of the mismatched antigens (i.e. class I, class II or non-HLA).

Interactions between injury and allorecognition

Although the unavoidable injury of donation and implantation is a serious issue in transplantation, the belief that injury promotes T-cell allorecognition in fact has not been proved. Injury increases MHC expression and IFNG production (19,20), but has not been shown to increase rejection across MHC disparities. Moreover, MHC class I expression and IFNG actually have stabilizing roles as outlined below.

In population studies, injury effects on the graft are additive with rejection, not synergistic, and the injury phenotypes such as delayed graft function have complex relationships with covariates such as donor age (21). Injury complicates transplant assessment and management. For example, TCMR is diagnosed more often in kidneys with delayed graft function, but this observation is hard to interpret because (a) damaged kidneys are biopsied more frequently than well-functioning kidneys; (b) early ABMR can mimic ischemic injury; and (c) acute renal injury can induce tubulitis and mimic TCMR. Thus one should not assume that interventions directed against organ injury would automatically have immunologic benefits.

Cell entry: crossing the microcirculation

Transendothelial migration follows the general multistep process of rolling via selectin–selectin ligand interactions, triggering of chemokine receptors, tight adhesion through integrins and immunoglobulin superfamily adhesion molecules, and transmigration (22). Chemokine and adhesion systems of inflamed sites are specialized and differ from those of the secondary lymphoid organs, the effector generator. The allograft adhesion molecules and chemokines are induced by the stress of the donation – implantation process, by the inflammation triggered by T cells, and by the injury induced by TCMR. T cells are guided to extravasate through the endothelium and within the graft by chemokine gradients on glycosaminoglycans. Prominent adhesion molecule transcripts in mouse and human transplants include L selectin and P selectin ligands, LFA-1 (ITGAL/ITGB2) and its ligands ICAM1 and ICAM2, and VLA4 (ITGA4/ITGB1) and its ligand VCAM1. However, preventing rejection by blocking individual adhesion molecules or chemokines has been unsuccessful, probably reflecting the robustness of the strategies used by inflammatory mechanisms, which avoid excessive reliance on single molecules.

An intriguing observation is that alloimmune T cells and IFNG effects appear within 24 hours posttransplant in mouse kidney and heart allografts (13,23). In mouse kidney allografts, T cells are first observed in periarterial areas rather than interstitial areas (13). This may reflect heterologous memory, although some T cells found at day 1 in mouse kidney allografts lack memory markers (13). Early T-cell entry and activation may represent a scouting mechanism that anticipates later events. The recently described peripheral surveillance memory T cells may be relevant in this regard. Such cells play a ‘first line of defense’ role, due to their capacity to patrol normal tissues (22,24). The ability of peripheral surveillance memory T cells to enter allogeneic tissue without previous injury has implications for our understanding of whether organ injury is essential for T-cell entry to organs.

There are two models for how cognate effector T cells home to the graft: by recognizing antigen on endothelium, or by simply entering the graft in their role of patrolling peripheral tissues and injured sites and encountering donor antigens on interstitial antigen-presenting cells (22,24). In either case, the encounter of cognate T cells with donor antigen activates a local inflammatory compartment that recruits cognate and noncognate T cells and macrophages. Once inflammation is triggered, most T cells that enter the graft from the blood engage the allograft endothelium through a mechanism independent of antigen recognition on the endothelium, in response to the endothelial changes induced by inflammation. As in other inflamed sites, the microcirculation becomes permeable, creating edema.

In TCMR, effector T cells cross the donor capillary endothelium without killing the endothelial cells, at least initially, indicating the existence of mechanisms that protect the endothelium from destruction (25). The need to protect the microcirculation in inflamed sites is empirically obvious: in host defense, T cells must purge organs of virus infections without compromising the microcirculation, which would destroy the organ and kill the host. Endothelium expresses MHC antigens and has some antigen-presenting abilities (26), but it is relatively resistant to killing by effector T cells, and its representation of MHC is specialized to protect it from injury by effector T cells entering the tissue (25). TCMR does not typically trigger ischemic changes in the parenchyma, and can smolder for days or weeks as a tubulointerstitial process that remains treatable with steroids.

Leukocyte recruitment is selective, involving both positive (attraction) and negative (exclusion) strategies. Chemokines can act both as agonists and antagonists, attracting some leukocytes (e.g. T cells and macrophages) but excluding others (e.g. neutrophils and eosinophils). The paucity of eosinophils in TCMR may reflect inhibitory processes mediated by IFNG, which suppresses eotaxins and antagonizes eotaxin receptors via CXCL9, CXCL10 and CXCL11 (27–29). Many other examples will emerge of regulation of cell entry; for example, macrophage products such as MMP12 inhibit neutrophil entry by inactivating the chemokines they require (30).

Features of the Inflammatory Compartment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

The infiltrating T cells

The T cells in the infiltrate are CD4 and CD8 effector and effector memory cells. The transcriptomes of primed CD4 versus CD8 effector T cells, and of effector memory T cells versus recently primed effector T cells, are surprisingly similar (31–33). One cannot identify which T cells in infiltrates are cognate, i.e. antigen-specific. This may explain why the T-cell transcripts expressed in TCMR are relatively stereotyped and do not qualitatively separate TCMR from the inflammation associated with nonspecific injury (8,34).

While the number of T cells in the graft that are specific for donor antigens is unknown, some observations suggest that they are a minority. Only a few thousand alloreactive T cells can mediate TCMR (35), and limiting dilution analysis suggests that cognate T cells are a small percentage of infiltrating cells (36). In the clinic, TCMR episodes can occur in severely T-cell-depleted patients (37), with many macrophages and IFNG effects but very few T cells.

Noncognate T cells, recruited and activated as a consequence of the activities of cognate T cells, may contribute to the inflammation in TCMR. Memory T cells in inflamed sites can be triggered to express inflammatory (innate immune) functions such as IFNG production through cytokine signals without antigen engagement (38,39).

Macrophages

The macrophage population is heterogeneous and constantly responding to the changing conditions. Macrophages in TCMR have complex cause–effect relationships to the parenchymal deterioration, including antigen presentation, effector functions and regulatory functions, but also response to injury and tissue remodeling and repair (40). In mouse kidneys with TCMR, the transcripts include macrophage markers such as EMR1 (F4/80), CD68 and ITGAM (‘Mac1’), potential dendritic cell markers such as ITGAX (CD11c), and transcripts reflecting alternative macrophage activation. Studies of the impact of macrophages on the phenotype of rejection are limited by the complexity and dynamism of the macrophage population and by the lack of animal models with macrophage deficiency without undesirable features such as impaired effector T-cell generation.

Other cell types

Small numbers of B cells and plasma cells are present in the TCMR infiltrate, probably recruited nonspecifically to the inflamed site. However, B-cell knockouts have little effect on the early course of TCMR (9), and the role of B cells and plasma cells in the graft is undefined. Extensive B-cell and plasma cell infiltrates are late developments, usually in damaged grafts beyond 6 months posttransplant, and have no clear relationship to ABMR or TCMR (21).

NK cells are rare in human TCMR, and their role is likely to be minor. The rarity of NK cells compared to T cells may reflect differential homing strategies: NK cells lack CXCR6, whereas most effector T cells express CXCR6, which has been implicated in T-cell recruitment to inflamed sites (32,41,42).

The role of IFNG and the IFNG-MHC axis

TCMR manifests intense IFNG production and effects in the local cells. IFNG, presumably from T cells, plays a major role in establishing the typical phenotype of TCMR, and IFNG effects correlate strongly with the T-cell burden (8). IFNG induces expression of many genes in the donor tissue and in host infiltrating cells, inducing chemokines such as CCL5, CXCL9, CXCL10, CXCL11 and MHC class I and II (43).

Host IFNG has many nonredundant effects, which are reflected in the powerful phenotypes of IFNG deficiency in knockout mice (44,45). Contrary to expectations, the absence of host IFNG is not beneficial but evokes a deviant rejection phenotype: as TCMR develops around days 5–7, kidney allografts (not isografts) develop ischemic necrosis and congestion due to failure of the microcirculation, plus alternative macrophage activation. These features are shared by kidney allografts in wild-type hosts if the donor lacks IFNGRs, IFNG-regulated factor 1 (Irf1) (46) or expression of MHC class I (29). Thus IFNG stabilizes the graft during TCMR by an action on donor cells, preventing injury to the microcirculation and alternative macrophage activation. IFNG-induced MHC class I molecules may influence the inflammatory phenotype by acting as ligands for inhibitory receptors on T cells and other inflammatory cells (47).

Tubulitis reflects dedifferentiation of the epithelium

As TCMR begins, the epithelium dedifferentiates before tubulitis develops (14), losing transcripts and proteins associated with function, such as solute carriers, and expressing embryonic genes, injury genes and cell cycle genes (15,48). Mononuclear cells first accumulate in the interstitium (i score) before T cells cross the basement membrane to enter the epithelium as tubulitis (t score), a major diagnostic lesion in TCMR (4). In clinical transplantation, tubulitis correlates strongly with interstitial inflammation, suggesting that tubulitis reflects epithelial dedifferentiation in response to the interstitial inflammation. Tubulitis occurs in other situations involving epithelial injury and dedifferentiation, including atrophy, acute tubular necrosis and interstitial nephritis. Thus tubulitis does not imply cognate engagement of epithelial cells by T cells. The i score and t score correlate with the accumulation of transcripts, reflecting the T cell and macrophage burden, IFNG effects, alternative macrophage activation and parenchymal injury (8,15). Intimal arteritis in small arteries is less frequent and less specific, being also seen in ABMR.

Known cytotoxic mechanisms are not required for TCMR

While T-cell cytotoxic molecules (perforin, granzymes, Fas ligand) are prominent in the infiltrating T cells in TCMR, tubulitis and epithelial deterioration in TCMR do not require T-cell cytotoxic mechanisms or for that matter epithelial cell death. In mouse allograft rejection, cell death by apoptosis is demonstrable in the infiltrating T cells, but the main feature of the epithelium is dedifferentiation, not necrosis or apoptosis (44). TCMR lesions are dependent on host T cells but independent of host B cells and alloantibody (9). Hosts lacking integrin CD103 reject kidney allografts with typical tubulitis, indicating that epithelial damage is independent of CD103-E cadherin interactions (12). TCMR lesions are unchanged in hosts deficient in perforin or granzymes A and B (10,12) and when donor tissue lacks Fas (11). One report suggested that donor Fas deficiency transiently reduces TCMR (49), but we were unable to confirm this result. However, there remain effector functions of T cells that are not well characterized, perhaps including other cytotoxic mechanisms such as granzyme K (50,51).

Despite the diffuse inflammation, it is likely that destructive effects of TCMR in an allograft are microspecific, i.e. confined to the area around the cognate T cell, damaging only the adjacent allogeneic tissue and sparing syngeneic tissue in the same graft (52). This could be explained by soluble factors produced by antigen-engaged T-cell macrophage unit acting only at short range, or a contact-dependent cytotoxic mechanism that has not been defined. However, the fact that deterioration of the epithelium begins long before T cells have demonstrable contact with the epithelium argues against but do not exclude contact-dependent mechanisms (14).

Mediators of epithelial dedifferentiation: potential role for TGFB1

One mediator of the epithelial changes is probably TGFB1, which can induce loss of transporters, cadherins, and other epithelial features, expression of mesenchymal markers (vimentin), and potential progression to epithelial mesenchymal transition. TGFB1 is regulated not only by production but by many mechanisms of activation, including integrin ITGB6 expressed by injured renal epithelium, and other integrins expressed by inflammatory cells (53). TGFB1 also has powerful anti-inflammatory effects, reminding us that each mediator is a player in an inflammatory orchestra, not a solo act. Many soluble products of T cells and macrophages (e.g. cytokines and growth factors, enzymes, reactive oxygen species, nitric oxide, CO and eicosanoids) could contribute to epithelial dedifferentiation, either acting directly on the epithelium or indirectly through changes in the extracellular matrix or microcirculation.

The regulated nephron response to injury

Local effects of cognate T-cell-induced inflammation on adjacent epithelial cells in a nephron segment could trigger a nephron regulatory response. We often conceptualize individual donor epithelial cells as being affected individually by TCMR, ignoring the macroorganization of epithelial cells into regulated snake-like nephrons, the functional units of the kidney. Nephrons can respond to injury in ways that determine the fate of all of their member cells. Many changes in rejection may reflect such nephron responses triggered by segmental injury. A candidate sensor for epithelial injury is the juxtaglomerular apparatus, which monitors the performance of the epithelium via the electrolyte contents in the distal convoluted tubule fluid, inducing vasoconstriction in the afferent arteriole and shutting down filtration (54). This would impact every cell in the nephron, suspending function and triggering dedifferentiation until the regulatory mechanism senses that the integrity of the damaged epithelium is restored and filtration can once again be permitted. If injury persists and repair fails, shutdown becomes irreversible, and nephrons that have been permanently shut down will resolve by atrophy, fibrosis, global glomerulosclerosis and mechanisms such as epithelial mesenchymal transition. This concept of nephron loss is crucial: prevention of fibrosis will not be beneficial if the nephrons cannot be saved.

Natural History of TCMR and Response to Therapy

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

The natural history of untreated TCMR can be modeled in mouse kidney allografts, which by day 21 manifest universal tubulitis and profound epithelial deterioration but typically not global infarction. Such kidneys often show hemorrhage and patchy ischemic necrosis, reflecting failure of the microcirculation or possibly effects of severe arteritis, and resemble nephrectomies of end-stage failed human kidney transplants. Alternative macrophage activation and TGFB1 effects are prominent in advanced TCMR despite persistent IFNG effects (15,34).

Both in mice and humans, it is difficult to exclude alloantibody-induced damage in such end-stage cases (9). Mouse kidney allografts in B-cell knockout hosts show attenuation of the late stages of TCMR, either because of the absence of alloantibody or because the absence of the antigen presenting function of B cells is needed to sustain effector T-cell generation (9). End-stage human kidney transplants with severe TCMR usually have concomitant ABMR. Analyses of failed human kidney allografts with TCMR but no ABMR would be instructive.

The natural history of untreated TCMR in the absence of ABMR and in the absence of treatment may not be relentlessly destructive. T-cell responses faced with persistence of antigen activate complex regulatory mechanisms of clonal exhaustion and anergy, as well as regulation. Ultimately, effector generation is probably reduced. The role of regulatory T cells in the lymphoid organs is well established, although their antigen specificity remains unclear. Cells with features of regulatory T cells such as FOXP3 expression also appear over time in inflamed compartments (55), but their role in controlling TCMR and their antigen specificity are not clear.

TCMR in the absence of ABMR is very responsive to steroid therapy and if necessary T-cell-depleting antibodies, and does not lead to progressive late parenchymal deterioration following treatment, even if endothelialitis is present (34). Our experience with late TCMR in the absence of ABMR is that it is responsive to therapy (34). Thus the bad reputation of late rejection episodes is probably due to unrecognized ABMR. Some nephrons may shut down irreversibly after a TCMR episode, as manifested by an increase in atrophy and fibrosis and permanent loss of some renal function. However, in the absence of ABMR, the prognosis after treatment of TCMR is good in compliant patients (56). The well-known association between ‘rejection’ and reduced late survival (57) is probably due to unrecognized ABMR (58).

Advantages of the Model

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

This provisional framework is proposed to promote critical re-examination of how molecular mechanisms produce the features observed in clinical biopsies. This is the ultimate question in bringing the richness of molecular biology into the clinic. This will allow us to see molecular, histologic and clinical features as all representing the same biological events. The model permits deconstruction of the processes of TCMR into their components on a biological basis, assigning molecular changes to biologic processes, as exemplified by our pathogenesis-based transcript sets (http://transplants.med.ualberta.ca). In addressing these changes, investigators must decide on the comparator: normal tissue (a ‘sick vs. well’ analysis) or another disease state such as antibody-mediated rejection (a ‘sick vs. sick’ analysis). Both are useful, but the comparator limits the interpretation.

The model makes best guesses at the mechanisms operating based on the data available because such hypotheses once formulated can potentially be tested. A weakness of the model is our inability to model the fundamental cognate recognition unit (effector T-cell macrophage) in vitro, or to visualize the unit in vivo. The challenge now is to accomplish this and to frame new questions and strategies to address them: for example, why do cognate effector T cells in TCMR not use their cytotoxic mechanisms? How is TCMR microspecific? Such questions can be addressed productively within this overall framework, and the answers will probably more interesting than the model. The mechanisms proposed are applicable to interpreting TCMR in other organs, and other disease states in which the pathogenesis is driven by cognate T-cell recognition of antigens in the tissue.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References

This viewpoint relies on the work of many members of our research team, including (alphabetic order) Gunilla Einecke, Konrad Famulski, Luis Hidalgo, Michael Mengel, Jeff Reeve and Banu Sis. Thanks also to Allan Kirk for helpful suggestions on the manuscript; Dellice Berezan and Danielle Stewart for preparing the manuscript; and Ken North of North Design Group for artwork. This work was supported by funding and/or resources from Genome Canada, Genome Alberta, the University of Alberta, the University of Alberta Hospital Foundation, Roche Molecular Systems, Hoffmann-La Roche Canada Ltd., Alberta Ministry of Advanced Education and Technology, the Roche Organ Transplant Research Foundation, the Kidney Foundation of Canada, and Astellas Canada. Dr Halloran also holds the Muttart Chair in Clinical Immunology.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Creation of the Inflammatory Compartment
  5. Features of the Inflammatory Compartment
  6. Natural History of TCMR and Response to Therapy
  7. Advantages of the Model
  8. Acknowledgments
  9. References