Chronic allograft nephropathy

Purpose of review Chronic allograft nephropathy is the major cause of late renal allograft loss. This disease is heterogeneous and the diagnosis is nonspecific, with both immune and nonimmune causes. Increasingly, we are able to recognize specific contributors to the disease. Recent findings Further understanding of chronic allograft nephropathy comes from a large study detailing the natural history of the disease, from protocol biopsies revealing subclinical cellular rejection, and from studies using C4d staining to distinguish antibody-mediated chronic rejection from nonspecific causes. Also made more clear are nonimmune mechanisms of chronic allograft nephropathy, such as the effect of decreased dosing of calcineurin inhibitors, and the concept of senescence as a mechanism of the disease. Summary Chronic allograft nephropathy is a heterogeneous disease with immune and nonimmune causes. Some features recognizable by histology and detected by other laboratory tests can help to categorize specific causes of the disease in particular cases. In addition, recent studies have contributed to our knowledge of the pathogenesis of the disease. In order to advance our understanding, we must be able to distinguish the various recognizable causes of chronic allograft dysfunction. Further research is warranted on the subset of the disease with indeterminate cause.


Introduction
As acute rejection has become a less common cause of renal allograft loss, the major group of pathologic processes that affects renal allograft function is referred to as chronic allograft nephropathy (CAN) [1]. CAN is the main cause of graft loss after 1 year, with a 60-70% prevalence in protocol biopsies [2,3]. The disease is characterized clinically as chronically deteriorating renal allograft function, often with proteinuria. 'CAN' was introduced in the Banff 91 schema as a term that encompassed several pathologic processes that could not always be distinguished on biopsy [4]. The diagnostic features were mild to severe interstitial fibrosis and corresponding tubular atrophy. Since then, at times the term has been used loosely to designate virtually any pathologic process affecting the renal allograft. Further complicating the picture is the increased attention CAN has received in the scientific literature in the past few years (Fig. 1). A major challenge in the future of renal transplantation is to dissect out the distinct and identifiable causes of CAN and to develop treatments for the specific contributing entities.
Causes of CAN are both immunologic and nonimmunologic. Rejection, both cellular and humoral, is a major contributor to CAN. Here, the term chronic rejection is defined as active but slowly progressive injury caused by alloreactivity to the graft. Nonimmunologic causes are also important, and include chronic calcineurin inhibitor toxicity, donor disease, ischemia, infection (e.g. polyoma virus), obstruction, hypertension, and recurrence of the original kidney disease [5,6].
The heterogeneity of processes leading to CAN raises several questions. What are the underlying causes of CAN? To what degree do these factors contribute to CAN? Can we identify features suggestive of a specific diagnosis? In order to develop treatments for CAN, we must be able to identify the distinct causes of renal failure in individual patients. Here we review recent literature that helps to elucidate CAN.

General histologic features of chronic allograft nephropathy
Chronic changes in the renal allograft can be seen in the glomeruli, vessels, tubules, and interstitium. Classically, in chronic rejection, the glomerular basement membrane is duplicated as seen by periodic acid-Schiff or silver stains (Fig. 2). This feature, which when widespread is known as chronic allograft glomerulopathy, is thought to be the most specific glomerular lesion in chronic rejection [7,8]. Glomerular basement membrane duplication, however, may be seen in other conditions present in the graft, such as recurrent or de-novo glomerular disease or chronic thrombotic microangiopathy. In addition to glomerular basement membrane duplication, the glomeruli in CAN may also show an increase in mesangial cells and matrix [9].
Chronic allograft arteriopathy is characterized by concentric thickening of the intima, sometimes with inflammatory cells under the endothelium [7] (Fig. 3). This thickening is accompanied by narrowing of the arterial lumen. These arterial features, suggestive of chronic immune-mediated damage, should be distinguished from arterial changes due to hypertension, which include thickening of the media and duplication of the internal elastic lamina and do not show an inflammatory cell infiltrate [10]. Also suggestive of immunologic injury is duplication of the lamina densa in peritubular capillaries as seen by electron microscopy [11,12].
Specific features of chronic calcineurin inhibitor toxicity, another contributor to CAN, can be seen in renal arterioles. Chronic calcineurin inhibitor toxicity is characterized by nodular hyaline deposits on the outer part of the arteriole [13,14] (Fig. 4a). This type of hyaline deposit must be distinguished from hyaline arteriolosclerosis seen in hypertension and the aging kidney, where hyaline material appears on the intimal surface and is generally not nodular (Fig. 4b). Hyaline arteriolosclerosis may also be more prominent in diabetes mellitus [15].
The interstitium in CAN shows an increase in fibrosis with corresponding tubular atrophy [8]. Although this feature is necessary to diagnose CAN, on its own, it does not point towards any specific etiology.
Nankivell et al. [16 ] recently published a paper detailing the natural history of CAN. This study included 961 biopsies from 120 patients, all but one of whom received kidney-pancreas transplants for type 1 diabetes. The authors observed two phases of renal allograft injury. An early phase consisted of acute tubulointerstitial damage and rapidly increasing interstitial fibrosis and tubular atrophy. This early damage and subsequent CAN correlated with acute tubular necrosis at the time of transplantation and with episodes of severe acute rejection. Acute and subacute rejection was found in 60.8% of patients at 1 month following transplantation; this percentage fell to 45.7% and 25.8% at 3 months and 1 year, respectively. Subclinical rejection and acute rejection correlated with more severe CAN later. The later phase of renal allograft injury, occurring after 1 year, was characterized by findings of worsening CAN, with severe CAN present in 58.4% of patients at 10 years. More severe disease was associated with calcineurin inhibitor use.
One of the difficulties in determining specific causes of CAN is that many of these pathologic processes have not been recognized on renal biopsy. More recent advances, such as the ability to detect C4d deposition in cases of chronic rejection, have contributed to our knowledge of CAN and our ability to make a more specific diagnosis [17,18 ].

Chronic allograft nephropathy and alloreactivity
CAN is often seen as a result of rejection, both acute and chronic. Patients with a history of acute rejection 230 Renal immunology and pathology  The duplicated glomerular basement membrane is seen on this silver stain.
are at increased risk for developing CAN, and some studies have shown HLA mismatch to be a risk factor for CAN. Acute and chronic rejection may be cellular, humoral, or both. Because immune-mediated injury may respond to specific treatment, and current or future treatments may differ for cellular versus humoral rejection, it will be increasingly important for the pathologist to recognize the components of immune-mediated injury [19][20][21][22][23][24].

Cellular rejection
CAN is related to subclinical rejection as has been observed in protocol biopsy studies. In a study by Shishido et al. [25 ] looking at protocol biopsies over 5 years in 95 grafts, subclinical acute cellular rejection was often found in association with CAN, especially earlier in the disease. Subclinical acute cellular rejection, defined histologically according to the Banff 97 schema in the absence of a rise in creatinine, was seen in 50% of CAN cases at 1 year, 32% at 2 years, 19% at 3 years, and 16% at 5 years following transplantation. The patients with subclinical acute cellular rejection showed worse renal function and graft survival. These findings in this respect are similar to those reported by Nankivell et al. [16 ,26 ]. Of course, the frequent association of subclinical rejection with CAN does not prove that rejection causes CAN, but these findings certainly are suggestive of acute cellular rejection playing a role in CAN.

Humoral rejection
The importance of antidonor antibody in renal allograft rejection has become widely recognized in the past several years [27]. Antigraft antibody causes both acute and chronic rejection, recognized by the presence of widespread C4d deposition in peritubular capillaries as seen by immunofluorescence microscopy [17,28,29].
Recently, further studies have contributed to our knowledge of the histopathology and pathogenesis of antibodies in CAN and chronic humoral rejection in particular.
A recent study [18 ] surveyed a subset of chronically injured renal allografts, those with chronic allograft glomerulopathy (CAG). Among 1111 kidney transplants that functioned for at least 6 months, 18 (1.6%) showed this glomerular lesion, which was present at a median of 8.3 years following transplantation. Deposition of C4d Chronic allograft nephropathy Cornell and Colvin 231  was found in 10/11 CAG cases, while in only 2/13 controls, indicating ongoing antibody deposition with complement activation in the graft. Several of these CAG patients also had demonstrable donor-specific anti-HLA antibodies. Risk factors for CAG included higher panel-reactive antibody percentage and acute rejection 3 months following transplantation.
Poggio et al. [30 ] evaluated patients with CAN for the presence of antibodies against donor HLA molecules, a target of antigraft antibodies in humoral rejection. The authors included 20 patients with CAN, diagnosed by biopsy or by clinical criteria, and 25 control patients, and looked for donor-reactive immune cells in peripheral blood from these patients using an IFN-g enzyme-linked immunosorbent spot assay. Their results suggested that patients with CAN harbored cells that reacted specifically against donor antigens. This study provides evidence for specific antibody-mediated rejection as a cause of CAN.

Rejection or accommodation?
Some stable patients demonstrate the presence of antidonor antibody, or at least focal C4d deposition, without full-scale antibody-mediated rejection [31]. Graft survival in the presence of antigraft antibodies has been termed 'accommodation' [32]. This concept supposes that the graft, in the presence of specific antibodies, may be protecting itself from antibody and complement-mediated damage, perhaps by means of increased expression of complement regulatory or antiapoptotic factors on the endothelium. Human renal allografts show increased expression of bcl-xL in patients with antidonor antibody, and upregulation of Duffy antigen receptor for chemokines has been shown in episodes of acute rejection [33,34]. We have found increased expression of the complement regulatory factor protectin (CD59) in peritubular capillaries in acute and chronic renal allograft rejection and decay-accelerating factor (CD55) in a subset of chronic humoral rejection cases [35]. Increased expression of factors such as these, in the setting of antigraft antibodies, may contribute to accommodation and to the attenuation of rejection.

Nonimmunologic contributors to chronic allograft nephropathy
Nonimmunologic factors are also important in the pathogenesis of CAN. Several donor factors seemingly unrelated to alloreactivity correlate with an increased risk of CAN, including increased donor age, brain death as the cause of donor death, preservation injury, infection, and various recipient factors [36][37][38]. Chronic calcineurin inhibitor toxicity is a major contributor to CAN. The renin-angiotensin system (RAS) may also play a role in CAN. Finally, the cumulative insults may push the graft into a state of senescence, such that the graft can no longer respond normally to injury.

Drug toxicity
Advances in immunosuppressive therapy have made an improvement in preventing acute rejection and in short-term graft survival. Currently, a 'triple therapy' consisting of prednisone, a calcineurin inhibitor, and mycophenolate mofetil is the most commonly used drug combination for maintenance in transplant patients [39 ].
In the long term, however, calcineurin inhibitors are associated with increased interstitial fibrosis and tubular atrophy. Because of the considerable nephrotoxicity caused by calcineurin inhibitors later in the life of the graft, Nankivell et al. [16 ] suggested that a two-stage treatment, with less reliance on calcineurin inhibitors in the second phase, may provide for better renal outcomes than current strategies.
One potential approach to avoiding chronic calcineurin inhibitor toxicity is through the use of sirolimus, which has only become more widely used in the past few years [39 ]. Sirolimus inhibits cell proliferation by blocking signal transduction from the activated IL-2 receptor [40]. A recent study by Ruiz et al. [41 ] looked at early withdrawal of calcineurin inhibitors in renal transplant patients. Sixty-four patients received steroids, cyclosporine, and sirolimus for the first 3 months after transplantation. After 3 months, patients were randomly assigned to two treatment groups. The control group received continued therapy with steroids, cyclosporine, and sirolimus. The cyclosporine-withdrawal group received steroids and decreased doses of cyclosporine with withdrawal of cyclosporine after several weeks and a corresponding increased dose of sirolimus. At biopsy after 1 year, the patients in the cyclosporine-withdrawal group overall showed significantly less progression in CAN indices, namely interstitial fibrosis and tubular atrophy, when compared with the baseline (pretransplant) biopsy. Another recent study [42] also looked at reduced dosing or withdrawal of tacrolimus, again with the addition of sirolimus, with similar results, although there was a higher incidence of subclinical rejection in patients not taking tacrolimus.

Renin-angiotensin system
Blockage of the RAS is beneficial in a variety of native kidney diseases, presumably by decreasing glomerular capillary pressure [43]. Angiotensin-converting enzyme inhibitors (ACEIs) are thought to decrease fibrosis and improve renal survival in native kidney disease [44,45]. One mechanism of the RAS in renal fibrosis probably involves the regulation of plasminogen activator inhibitor-1 (PAI-1) by angiotensin II. Angiotensin II increases levels of PAI-1, which normally inhibits plasminogen activators in the blood [46]. When PAI-1 levels are increased, the generation of plasmin is reduced; plasmin, then, is not sufficiently present to degrade excessive matrix proteins [47].
Besides reducing the dose of calcineurin inhibitors, another strategy for avoiding CAN may be through the use of ACEIs. In a recent retrospective study in 72 patients with biopsy-proved CAN, Artz et al. [48 ] found that patients who were taking ACEIs had overall less severe CAN and longer graft survival, a finding not related to decreased blood pressure in this patient group. Also, a particular angiotensin-converting enzyme genotype, DD, is associated with worse graft survival in patients with CAN, a finding further supporting the importance of the RAS in graft survival [49 ].

Senescence
One intriguing hypothesis of CAN has to do with the concept of senescence in the allograft. 'Senescence' refers to a phenomenon observed in cultured cells, where cells stop dividing after a number of passages and show different characteristics from normal somatic cells, such as resistance to growth factors [50,51]. Several recent studies have looked for the presence of senescent cells in cases of CAN. Ferlicot et al. [52 ] used a specific marker, senescence-associated b-galactosidase (SA-B-gal), in 67 renal biopsies with CAN, and correlated staining to telomere length in the cells [52 ,53]. In particular, tubular epithelial cells were found to express SA-B-gal in cases that showed staining. Senescent cells were present in 67% of CAN cases, and the amount of staining correlated with the severity of CAN. In addition, the presence of senescent cells was related to the age of the donor, but not to the duration of cold ischemia time, the time of biopsy after transplantation, or the age of the recipient.
Chkhotua et al. [54 ] looked at expression of the cyclindependent kinase inhibitor genes p16 and p27, thought to play a role in the development of cell senescence, in cases of CAN and in normal aging kidneys. The authors found increased tubular and interstitial expression of p16 and p27 in CAN as compared with controls. Moreover, they observed a greater degree of staining in CAN cases than would be predicted for the age of the kidney, such that the 'biological age' of the chronically rejected kidney is higher than the 'chronological age'. These observations support the hypothesis that the various stresses a graft encounters -ischemia, reperfusion injury, infection, rejection, and so on -lead to 'accelerated senescence' [55]. These senescent cells in the graft do not respond normally to stimuli and may not repair normally, but instead may function to produce fibrogenic factors, leading to increased interstitial fibrosis.

Conclusion
CAN is currently the major cause of late renal allograft loss, but this disease process is heterogeneous, with both immune and nonimmune mediators, and the diagnosis of CAN is nonspecific. Recent studies have elucidated various specific mechanisms contributing to CAN. In order to make progress in the prevention and treatment of CAN, we must be able to make diagnoses as specific as possible, with the remainder of cases falling into the idiopathic category of CAN. Further work must then be done to uncover the causes of idiopathic CAN.