Overlapping pathways to transplant glomerulopathy: chronic humoral rejection, hepatitis C infection, and thrombotic microangiopathy

Transcription

1 & 2011 International Society of Nephrology see commentary on page 801 Overlapping pathways to transplant glomerulopathy: chronic humoral rejection, hepatitis C infection, and thrombotic microangiopathy Seema Baid-Agrawal 1, Alton B. Farris III 2, Manuel Pascual 3, Shamila Mauiyyedi 4, Mary Lin Farrell 5, Nina Tolkoff-Rubin 5, A. Bernard Collins 6, Ulrich Frei 1 and Robert B. Colvin 6 1 Department of Nephrology and Medical Intensive Care, Campus Virchow Clinic, Charite Universitatsmedizin Berlin, Berlin, Germany; 2 Department of Pathology, Emory University, Atlanta, Georgia, USA; 3 Transplantation Center, Departments of Medicine and Surgery, University Hospital of Lausanne (CHUV), Lausanne, Switzerland; 4 Department of Pathology and Laboratory Medicine, University of Texas, Health Science Center at Houston, Houston, Texas, USA; 5 Renal and Transplantation Units, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA and 6 Department of Pathology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA Transplant glomerulopathy (TG) has received much attention in recent years as a symptom of chronic humoral rejection; however, many cases lack C4d deposition and/or circulating donor-specific antibodies (DSAs). To determine the contribution of other causes, we studied 209 consecutive renal allograft indication biopsies for chronic allograft dysfunction, of which 25 met the pathological criteria of TG. Three partially overlapping etiologies accounted for 21 (84%) cases: C4d-positive (48%), hepatitis C-positive (36%), and thrombotic microangiopathy (TMA)-positive (32%) TG. The majority of patients with confirmed TMA were also hepatitis C positive, and the majority of hepatitis C-positive patients had TMA. DSAs were significantly associated with C4dpositive but not with hepatitis C-positive TG. The prevalence of hepatitis C was significantly higher in the TG group than in 29 control patients. Within the TG cohort, those who were hepatitis C-positive developed allograft failure significantly earlier than hepatitis C-negative patients. Thus, TG is not a specific diagnosis but a pattern of pathological injury involving three major overlapping pathways. It is important to distinguish these mechanisms, as they may have different prognostic and therapeutic implications. Kidney International (2011) 80, ; doi: /ki ; published online 22 June 2011 KEYWORDS: antibody-mediated rejection; hepatitis; kidney transplantation Correspondence: Seema Baid-Agrawal, Department of Nephrology and Medical Intensive Care, Campus Virchow Clinic, Charite Universitatsmedizin Berlin, Augustenburger Platz 1, Berlin, Germany. seema.baid-agrawal@charite.de Received 14 October 2010; revised 13 March 2011; accepted 15 April 2011; published online 22 June 2011 Transplant glomerulopathy (TG) is a glomerular lesion common in long-standing kidney allografts, first described by Porter et al. 1 in 1967 and is characterized by duplication of glomerular basement membrane. In recent years, it has received much attention as a manifestation of chronic humoral rejection (CHR). 2,3 The prevalence of TG has been shown to be approximately 5 10% in all allograft biopsies performed for clinical indications. 4 9 A few contemporary studies of surveillance and clinical biopsies in kidney allografts have shown that the incidence of TG is higher than expected, affecting 4% of conventional transplants at 1 year post-transplant, progressively increasing up to 20% at 5 years. 10,11 TG is clinically manifested by low-grade to nephrotic-range proteinuria with progressive allograft dysfunction and has an extremely poor prognosis, with eventual graft loss in 40 70% of the affected patients. 4 9 The precise pathogenesis of TG remains unclear; theories include alloreactivity to the donor or chronic infection. Recent studies increasingly support the association of TG with circulating donor-specific human leukocyte antigen antibody (DSA) and/or C4d deposits in peritubular capillaries (PTCs), suggesting that humoral immune-mediated mechanisms of tissue injury have a role in the development of TG. 9,11 16 In a series from Vienna, C4d deposition preceded and predicted development of TG. 13 However, a substantial number of cases of TG have no associated C4d deposits or circulating DSA, indicating that yet unknown, other non-alloantibodymediated processes may be involved in the development of TG in a significant subset of patients Interestingly, the histopathological features of TG, which include glomerular basement membrane duplication and increase in mesangial matrix, are reminiscent of membranoproliferative glomerulonephritis (MPGN). 20 Both TG and MPGN induce proteinuria and appear similar by light microscopy. The mesangial and glomerular endocapillary hypercellularity of MPGN have also been recently described Kidney International (2011) 80,

2 S Baid-Agrawal et al.: Overlapping pathways to transplant glomerulopathy in cases of TG. 21 However, by immunofluorescence (IF) and electron microscopy (EM), immune complexes and electrondense deposits (subendothelial and/or mesangial) are typically seen in MPGN but not usually in TG. 1,2,4,20,22,23 In the past few years, MPGN has been clearly associated with hepatitis C virus (HCV) infection, both in non-transplant and transplant settings The pathogenesis of HCVassociated glomerular disease involves the formation of immune complexes within the glomeruli However, immune complexes may be scant or absent in a subgroup of recipients with chronic glomerular lesions after transplantation because of the immunosuppression, and may render a clear distinction between TG and MPGN difficult in most cases. Owing to the similarities between the histopathological features of TG and HCV-associated MPGN, a causal association between TG and HCV infection has been suggested in a few reports. In 1995, Gallay et al. 33 first suggested a possible association between HCV infection and TG in two kidney allograft recipients at 3 and 7 years after transplantation. These authors speculated that some patients diagnosed with TG may actually have HCVassociated MPGN. Cosio et al. 34 also found an increased prevalence (33%) of HCV infection in 27 kidney allograft recipients with TG. In their study, IF and EM were available in only a minority of cases, and therefore pathological analysis was limited. Corroborating these findings, in a recent study of protocol biopsies in kidney allografts, pretransplant HCV antibody positivity was found to be an independent risk factor for the development of TG. 11 However, the diagnosis of TG in their study was based only on light microscopy findings. In contrast to the above findings, Cruzado et al. 35 did not find any significant difference in the prevalence of TG in HCV-positive and HCV-negative kidney recipients when they included EM examination in all biopsies. Another disorder that can produce many of the histological findings analogous to TG is thrombotic microangiopathy (TMA), also initiated by endothelial injury and/or dysfunction. TMA has usually been associated with calcineurin inhibitor toxicity (CNIT) or with CHR in kidney transplant recipients. 36,37 In addition, TMA has also been identified in cases of HCV However, systemic clinicopathological studies looking at the association between TG, TMA, and HCV infection and contribution of these additional potential causes of TG apart from CHR in kidney transplant recipients are lacking. There is an unmet need to better characterize precise etiology and pathogenesis of TG, as it could contribute to prevention or interventions to preserve optimal graft function and survival in patients with this lesion. In this clinicopathological study with a long-term follow-up, we aimed to delineate the pathogenesis of TG after kidney transplantation and sought clinical and pathological evidence for additional potential causes of TG apart from CHR. RESULTS Clinical features and course of TG All patients had received ABO-compatible grafts and had a negative cross-match before transplantation. The maintenance immunosuppressive protocol consisted of cyclosporine, prednisone, and azathioprine (n ¼ 21), or prednisone and azathioprine (n ¼ 4). None of these patients received antibody induction therapy. The prevalence of TG was 12% (25/209) in all biopsies performed for evaluation of chronic allograft dysfunction and/or proteinuria. TG was diagnosed at a median of 66 months (range months) after transplantation, with a mean serum creatinine at biopsy of 4.4±2.1 mg/dl and lowgrade or nephrotic-range proteinuria (43.5 g/24 h). A significantly higher prevalence of HCV infection was observed in patients with TG as compared with the 29 control patients with chronic CNIT (36% versus 7%; P ¼ 0.02). All patients were hepatitis B surface antigen negative. None of the HCV þ patients had received previous interferon therapy. Pathology and C4d staining Three overlapping categories of TG were identified in 21/25 (84%) cases: C4d þ TG (12/25; 48%), HCV þ TG (9/25; 36%), and TMA þ TG (8/25; 32%). These overlapping features are illustrated through two-way hierarchical clustering (Figure 1). A total of 15 cases (60%) had only one of these features, probably representing an independent cause of TG (9 C4d þ, 3 HCV þ, and 3 TMA þ ). Six cases (24%) had combinations of C4d þ and HCV þ or HCV þ and TMA. Four cases (16%) had no evidence of HCV, CHR, or TMA. Of these, three had severe hyaline arteriolopathy attributed to CNIT, and therefore may have had an element of TMA. The fourth was a graft that had been in place for 34 years. The majority of the TMA cases were HCV þ (5/8; 62%) and the majority of HCV þ cases had TMA (5/9; 56%), suggesting that these are interrelated. The pathological features comparing HCV TG with HCV þ TG and also according to the C4d status are summarized in Table 1. Only 2/9 (22%) HCV þ TG cases had evidence of mesangial hypercellularity as compared with 11/16 (66%) HCV TG cases (P ¼ 0.04). We did not find any significant differences in presence of mononuclear cells in the glomeruli or PTC and PTC basement membrane multilayering when analyzed according to either the HCV or C4d status. These features were also present in some HCV þ TG cases even in the absence of C4d and/or DSA. No significant differences were found in PTC C4d deposition between HCV TG and HCV þ TG patients (9/16, 56% versus 3/9, 33%; P ¼ nonsignificant). No case of MPGN in native kidneys had C4d deposits in PTC. Immune complexes and subendothelial electron-dense deposits were present in all cases of HCV-associated MPGN in native kidneys as expected and in none of the TG cases (by definition, as described in the Materials and Methods section). 880 Kidney International (2011) 80,

3 S Baid-Agrawal et al.: Overlapping pathways to transplant glomerulopathy Clinical characteristics of HCV þ TG versus HCV TG patients The two groups of patients were similar with respect to demographics and baseline characteristics, such as age, race, gender, donor status (deceased or living donor), number of retransplants, panel reactive antibodies peak and current at TMA HCV C4d DSA TMA HCV C4d DSA NA Figure 1 The heterogeneity of TG cases is displayed through two-way hierarchical clustering with regard to the presence of C4d, DSA, TMA, and HCV. Each line displays a TG case. The dendrograms on the right show how closely related the cases are, and the dendrogram on the bottom shows how closely related the features (C4d, DSA, TMA, and HCV) are. DSA, donor-specific antibody; HCV, hepatitis C virus;, positive; NA, not available;, negative; TG, transplant glomerulopathy; TMA, thrombotic microangiopathy. the time of transplantation, and human leukocyte antigen matching. All patients in both groups were followed until progression to graft failure or death with a functioning graft. The clinical features of these two groups are compared in Table 2. A striking observation was that the duration from transplantation to total graft loss was significantly shorter in HCV þ TG than the HCV TG patients (71.1±52.7 versus 153.7±120.5 months; P ¼ 0.03), as also shown by Kaplan - Meier survival analyses (Figure 2, P ¼ 0.02, log-rank test). More HCV þ TG patients (4/9; 44%) had persistently abnormal liver function tests (defined by elevation of transaminases, aspartate aminotransferase, and/or alanine aminotransferase, 450 U/l on at least two occasions and maintained for more than 6 months) in contrast to none (0%) in the HCV TG group, suggesting the presence of chronic liver disease in HCV þ patients (P ¼ 0.01). Complement levels and cryoglobulin measurements were available in three HCV TG patients and one HCV þ TG patient. C3 and C4 values were normal, except in one HCV TG patient who had low C3 with normal C4; cryoglobulins were undetectable in all the four patients. Multivariate analysis showed that the presence of HCV had the highest likelihood ratio for more rapid progression to graft failure using both proportional hazards (Cox) and parametric survival fit methods (likelihood ratio ¼ 4.56 and 5.06, and P ¼ 0.03 and 0.02, respectively), with the other factors (C4d and TMA) having likelihood ratios that were not statistically significant (Table 3). Donor-specific antibody Of the 12 TG patients tested in the peri-biopsy period for circulating DSA, 7/8 C4d þ TG patients had positive de novo DSA (that is, negative pretransplant and positive at the time of biopsy) compared with 1/4 C4d TG patients (88% versus 25%; Po0.02). Furthermore, eight of nine HCV TG patients had detectable de novo DSA as compared with none in the three HCV þ TG patients (89% versus 0%, P ¼ 0.04). The information on DSA is displayed in Figure 1. DISCUSSION The findings of this study provide further insight into natural history, causes, and pathogenesis of TG in kidney transplant Table 1 Pathological features of patients with TG according to HCV and C4d status HCV TG (n=16; %) (n=9; %) C4d TG (n=13) C4d + TG (n=12) GBM duplication (both LM+EM) 16 (100%) 9 (100%) 13 (100%) 12 (100%) Mesangial hypercellularity 11 (69%) a 2 (22%) a 8 (62%) 5 (42%) Mononuclear cells in glomeruli b (X10 cells/glomerulus) 10/14 (71%) 5/7 (71%) 10/13 (77%) 5/8 (63%) Mononuclear cells in PTC c (in 410% PTC) 3/13 (23%) 2/7 (29%) 1/12 (8%) 4/8 (50%) PTCBMML (X3 layers) d 4/8 (50%) 4/4 (100%) 1/3 (33%) 7/9 (78%) Additional features of TMA (n=8) 3 (19%) 5 (56%) 6 (46%) 2 (17%) Complement C4d deposits in PTC (n=12) 9 (56%) 3 (33%) 0 12 HCV positivity 0 9 6/13 (46%) 3/12 (25%) Abbreviations: EM, electron microscopy; GBM, glomerular basement membrane; HCV, hepatitis C virus; LM, light microscopy; PTC, peritubular capillary; PTCBMML, peritubular capillary basement membrane multilayering; TG, transplant glomerulopathy; TMA, thrombotic microangiopathy. a P=0.04. b Data could be analyzed in 21/25 cases. c Data could be analyzed in 20/25 cases. d Data could be analyzed in 12/25 cases. Kidney International (2011) 80,

4 S Baid-Agrawal et al.: Overlapping pathways to transplant glomerulopathy Table 2 Clinical features and outcome of patients with TG HCV TG (n=16) (n=9) P-value Time from Tx to Bx (months) 124.3± ± Creatinine at Bx (mg/dl) 4.3±2.3 (median 3.4) 4.6±1.9 (median 3.8) 0.74 Proteinuria at Bx (g/24 h) a 2.6± ± Elevated liver transaminases (%) 0 4 (44%) 0.01 Total graft loss (%) 16 (100%) 9 (100%) 1.00 Death-censored graft loss (%) 15 (94%) 9 (100%) 1.00 Time from Bx to total graft loss (months) 29.4± ± Time from Tx to total graft loss (months) 153.7± ± Death (%) 1 (6%) 1 (11%) 1.00 Abbreviations: Bx, biopsy; HCV, hepatitis C virus; TG, transplant glomerulopathy; Tx, transplantation. a Only 10 HCV TG and 7 had 24 h urine collections available for analysis. Proportion surviving P= Time from Tx to total graft loss (months) Figure 2 Kaplan Meier curves showing time to graft loss after transplantation in HCV TG and HCV þ TG patients. HCV þ TG patients had a significantly faster progression to graft failure after Tx as compared with HCV TG patients. HCV, hepatitis C virus; TG, transplant glomerulopathy; Tx, transplantation. recipients. We found that TG was associated with a dismal graft outcome, with a median survival of only 9 months following the biopsy diagnosis. Pathogenetically, three overlapping categories of TG could be identified in a majority (84%) of patients: C4d þ TG (48%), HCV þ TG (36%), and TMA þ TG (32%). In a small proportion of patients, none of these causes could be identified, suggesting that there may be still unknown factors that have a role in the development of TG in some patients. Of particular interest was the striking prevalence of HCV infection in recipients with TG (36%), which was significantly higher than the prevalence seen in the chronic CNIT control group without TG (7%) or in our overall kidney transplant population (4.8%). 38 These results extend and confirm previous observations 11,33,34 and suggest a strong causal association between HCV infection and TG. Furthermore, HCV þ TG recipients differed from HCV TG patients clinically by a significantly more rapid progression to graft failure after transplantation and increased frequency of Table 3 Multivariate analysis of clinicopathological factors and time to graft loss Likelihood ratio (P-value) based on proportional hazards (Cox) fit Likelihood ratio (P-value) based on parametric survival fit C4d 0.82 (P=0.36) 1.86 (P=0.17) TMA 0.14 (P=0.71) 0.29 (P=0.59) HCV 4.56 (P=0.03) 5.06 (P=0.02) Abbreviations: HCV, hepatitis C virus; TMA, thrombotic microangiopathy. abnormal liver function tests. By multivariate analysis analyzing the impact of HCV, C4d, and TMA on graft failure, HCV alone was found to have a highest likelihood ratio for more rapid progression to graft failure, suggesting that HCV per se was associated with more rapid graft loss. We found C4d deposition in PTC in approximately half of the patients with TG, supporting an antibody-mediated (humoral) alloreactive mechanism of tissue injury in TG in a subset of patients, as suggested by previous studies. However, despite the growing body of evidence in support of an antibody-mediated mechanism, there are also some conflicting data that not all cases of TG are antibody-mediated. Vongwiwatana et al. 14 showed C4d deposition only in 25% of biopsies with TG. Furthermore, in protocol biopsies 10 years after transplantation, 6 of 11 of the C4d-negative patients had TG in absence of DSA. 41 In another comprehensive study, C4d was not correlated with TG among all patients. 18 Corroborating these results, C4d deposition was absent in half of the patients in our series as well. One possible explanation of these conflicting data may be the presence of C4d-negative antibody-positive TG associated with elevated intragraft endothelial transcripts as recently described by Sis et al. 42,43 Another explanation may be that, in addition to antibody-mediated TG, there is also alloantibody-independent TG where chronic viral infections such as HCV may have a role, as strongly suggested by our observations. Interestingly, in the 12 TG patients where DSA testing was available, 89% of HCV TG patients had circulating DSA as compared with none in the HCV þ TG patients. Various pathogenetic mechanisms may be postulated for this link between HCV and TG. First, immune complexes containing HCV, with or without cryoglobulins, may deposit 882 Kidney International (2011) 80,

5 S Baid-Agrawal et al.: Overlapping pathways to transplant glomerulopathy within the glomeruli similar to native kidneys. 26,27,29 After transplantation, owing to immunosuppression, enhanced viral replication in combination with modulation of lymphocyte activity may produce an antigen antibody imbalance, which may predispose to the development of de novo glomerular lesions in the allograft. Cryoglobulin detection can be difficult in transplant patients, possibly because of the immunosuppressive therapy that may reduce the cryoglobulin level and/or alter the antigen antibody ratio. 30,31 Even undetectable levels of cryoglobulins may be of pathogenic significance after kidney transplantation and may lead to absence of their extrarenal manifestations and attenuated histological expression on the allograft in a subgroup of HCV þ patients. 30 Therefore, it may be difficult to distinguish clearly between TG and MPGN in these cases. Second, HCV infection may promote TG in some patients by increasing humoral immune alloreactivity against the donor tissue, suggesting an antibody-mediated process. This is supported by our finding of C4d positivity in about onethird of HCV þ cases and the presence of other features of CHR such as PTC basement membrane multilayering and presence of mononuclear cells in PTC and glomeruli in some HCV þ patients even in the absence of C4d deposition. Last but not the least, approximately one-third of patients with TG had evidence of TMA in their biopsies, which may be postulated as another potential overlapping pathway for development of TG. Of note, the majority of the TMA cases were HCV þ and majority of HCV þ cases had TMA, suggesting that these two entities are interrelated. HCV infection may trigger TMA by virtue of mechanisms as postulated for cytomegalovirus. These include increased cytokine production (for example, interleukin-6 and interferon-a and g) and vascular endothelial cell activation that promotes alloreactivity. 44,45 From these results, it is tempting to postulate that a humoral mechanism is associated with TG in HCV patients and an alloantibody-independent process such as TMA in HCV þ patients. However, because of the limited DSA data in our study, and presence of C4d deposits in one-third of the HCV þ TG patients, definite conclusions regarding this possibility will require larger prospective studies. Whatever the exact mechanism of association between HCV infection and lesions of TG may be, the recognition of this association is potentially important, as the high prevalence of HCV infection among patients on dialysis and consequently in kidney transplant recipients puts this population at an increased risk for developing HCV-associated TG. 46 Furthermore, we have shown HCV positivity to be associated with a more rapid progression to graft failure compared with HCV recipients. Evaluation for HCV infection is therefore warranted in all patients who develop lesions of TG after kidney transplantation. Overall, it appears that a number of above described pathogenetic mechanisms (that is, chronic antibodymediated rejection or CHR, HCV-associated de novo glomerular lesions, and TMA) are involved in the pathogenesis of TG and may operate simultaneously producing an overlap, making it difficult to recognize the relative role of a single factor. In future, more work will be needed to find specific pathological markers that may differentiate these mechanisms. Detection of HCV antigens in the glomeruli using specific monoclonal antibodies would seem to be a useful solution, but the results have been difficult to reproduce. 28 Recently, glomerular C4d staining using a polyclonal anti- C4d antibody has been suggested as a potential marker of humoral immune responses in TG. 47 However, those authors did not look at the possible association of TG with HCV infection in their series. It would be interesting to study whether glomerular C4d deposition or endothelial gene expression differ significantly in HCV þ and HCV patients with lesions of TG. The strengths of our study are a comprehensive investigation of multiple clinical and histopathological factors including C4d testing and EM, exclusion of patients with confounding diagnosis of immune complex glomerulonephritis, and long-term follow-up. But there are also some limitations that deserve specific consideration. First, being a retrospective analysis, it suffers from some of the inherent limitations of this kind of study design. Second, the sample size is small, limiting the power of the study. However, still a statistically significant difference in the prevalence of HCV between the TG and the chronic CNIT groups was obtained, strongly suggesting a link between the two. Some caution is warranted when interpreting the negative results of the study. Third, DSA testing during the peri-biopsy period was not performed in all patients, and the stored sera were not available for a retrospective analysis in the remaining patients. Furthermore, the DSA numbers may be biased, as testing for DSA was done more frequently in C4d þ cases and may have been possibly as a result of the positive C4d. Last but not the least, according to the local practice of the institution, the biopsies were performed only by indication and not on a prospective protocol basis. However, the indication for biopsies was not biased by the HCV status of the patients, as many of the biopsies were performed before the introduction of routine screening for HCV serology at our institution and we retrospectively tested all the pretransplant sera for the presence of HCV RNA. In summary, TG is not a specific diagnosis, but rather a pattern of pathological injury with at least three overlapping causes including C4d-associated antibody-mediated rejection, TMA, and chronic HCV infection-associated glomerular lesions. It is important to distinguish between these mechanisms, as they may have different prognostic and therapeutic implications. Further larger, prospective studies carefully looking at the association among TG, TMA, and HCV infection utilizing protocol biopsies evaluating C4d staining, EM, endothelial gene expression, and circulating DSA analysis are required to confirm our findings. Early diagnosis of TG and better understanding of its pathogenesis may improve management and outcome of this deleterious post-transplant complication. Kidney International (2011) 80,

6 S Baid-Agrawal et al.: Overlapping pathways to transplant glomerulopathy MATERIALS AND METHODS The study was approved by the Institutional review board of the Massachusetts General Hospital. Kidney biopsy selection We searched the database of the Department of Pathology for kidney allograft biopsies performed between January 1990 and December 2000 for evaluation of chronic allograft dysfunction (typically a gradual and progressive rise in creatinine over a 6-month to 1-year period and/or proteinuria 4500 mg/24 h). Of a total of 209 such indication biopsies, we identified 25 biopsies performed in 25 patients with lesions of TG, as defined by light microscopy findings of 410% duplication of the glomerular basement membranes of non-sclerotic glomeruli and absence of immune complexes by IF and/or EM. 2,20 For comparing the prevalence of HCV infection, all patients with isolated biopsy-proven chronic CNIT (n ¼ 29) from the same time period were used as controls. Chronic CNIT was defined by the presence of arteriolar peripheral nodular medial hyalinosis typically accompanied by interstitial fibrosis and tubular atrophy on biopsy in patients receiving either cyclosporine or tacrolimus. For pathological comparison, the biopsies from 14 cases of MPGN in native kidneys (7 HCV þ and 7 HCV )from thesameperiodwerealsoidentified.allbiopsieswiththediagnosis of de novo or recurrent MPGN or other glomerular diseases including diabetic nephropathy were excluded from the study. Clinical data and donor-specific human leukocyte antigen antibody Clinical and laboratory data including peri-biopsy DSA analysis were reviewed for all patients with TG (n ¼ 25). Allograft loss was defined as development of end-stage renal disease requiring dialysis or death with a functioning graft. The pretransplant sera of all patients were checked for presence of HCV RNA by reverse transcriptase PCR using the Roche Amplicor 2.0 (Roche Molecular Diagnostics, Branchburg, NJ). The sera had been stored at 80 1Cin the Tissue Typing Laboratory. Testing for circulating DSA in peribiopsy period was available in 12/25 TG patients (9 HCV and 3 HCV þ ). Cytotoxic cross-matches were performed using patient sera against donor mononuclear cells that had been stored at 80 1C. T and B cells were isolated at the time of cross-matching using immunomagnetic beads. Both antihuman globulin-enhanced T-cell and standard complement-dependent cytotoxicity B-cell assays were used. 12 Pathology and tissue processing Samples were routinely divided and processed in a standard fashion to perform light microscopy, direct IF, and EM evaluation as described earlier Humoral antidonor reactivity was evaluated retrospectively using a three-step IF technique for C4d deposition in PTC in all available frozen biopsy samples. 51 It could be performed in all patients with TG, 23/29 controls with CNIT, and all 14 additional native MPGN controls. EM study of glomeruli was available in 23/25 patients with TG (7/9 HCV þ and 16/16 HCV patients), 6/23 controls with chronic CNIT, and all 14 native MPGN cases. EM study for PTC basement membrane multilayering was available only in 12/25 TG cases (3 C4d þ HCV þ, 6 C4d þ HCV, 1 C4d HCV þ, and 2 C4d HCV ) and in 6/29 chronic CNIT cases; as this being a retrospective analysis, the EM sections were not specifically prepared for evaluation of PTC in the earlier biopsies. All biopsies of 25 TG cases and other biopsies that these patients had before the diagnosis of TG, 29 CNIT and 14 native kidney MPGN cases were reviewed again. The glomerular lesions of TG were analyzed systematically and scored by two observers independently without the knowledge of clinical or serological data. In addition to the required features of TG to be included in this study, the biopsies were specifically analyzed for mesangial hypercellularity, the presence/margination of inflammatory cells (mononuclear cells) in glomeruli (21/25 cases), and PTC (20/25 cases). The remaining cases had no sufficient material available for analysis. On the basis of the evaluation of 3 5 PTC per biopsy, PTC basement membrane multilayering was categorized as absent, if no or focal splitting of the BM was noted, or present, if three or more layers of BM were identified. We also looked for evidence of TMA in all biopsies. TMA was recognized by the presence of thrombi or mucoid intimal thickening in arteries/arterioles or red blood cell fragments in the walls of arteries/arterioles. Statistical analysis Statistical analysis and hierarchical clustering was conducted in SAS JMP version 8.0 (SAS Institute, Cary, NC). Continuous variables are expressed as mean±standard deviation, and categorical variables are presented as proportions. Comparisons between groups were done using Fisher s exact test for categorical variables, unpaired t-test for normally distributed, and Wilcoxon sum-rank test for nonnormally distributed continuous variables. Two-sided P-values o0.05 were considered significant. The survival curves were calculated by Kaplan Meier survival analyses and compared by the log-rank and Wilcoxon tests. Multivariate analysis was conducted to analyze the influence of the presence of C4d, TMA, and HCV on censored survival data for progression to graft failure using two different methods: a Cox (semiparametric) proportional hazards method and parametric (linear) iterative maximum likelihood survival method using a Weibull distribution. 52 DISCLOSURE All the authors declared no competing interests. ACKNOWLEDGMENTS We are thankful to Dr Susanne Saidman and Donna Fitzpatrick from the Tissue Typing Laboratory, Massachusetts General Hospital, for their kind help in providing the panel reactive antibodies and the DSA data. REFERENCES 1. Porter KA, Dossetor JB, Marchioro TL et al. Human renal transplants. I. Glomerular changes. Lab Invest 1967; 16: Colvin RB. Antibody-mediated renal allograft rejection: diagnosis and pathogenesis. J Am Soc Nephrol 2007; 18: Cosio FG, Gloor JM, Sethi S et al. Transplant glomerulopathy. Am J Transplant 2008; 8: Maryniak RK, First MR, Weiss MA. Transplant glomerulopathy: evolution of morphologically distinct changes. Kidney Int 1985; 27: Habib R, Broyer M. Clinical significance of allograft glomerulopathy. Kidney Int Suppl 1993; 43: S95 S Shu KH, Lu YS, Chang CH et al. Transplant glomerulopathy a clinicopathological study. Transplant Proc 1996; 28: Suri DL, Tomlanovich SJ, Olson JL et al. Transplant glomerulopathy as a cause of late graft loss. Am J Kidney Dis 2000; 35: Banfi G, Villa M, Cresseri D et al. The clinical impact of chronic transplant glomerulopathy in cyclosporine era. Transplantation 2005; 80: Sis B, Campbell PM, Mueller T et al. Transplant glomerulopathy, late antibody-mediated rejection and the ABCD tetrad in kidney allograft biopsies for cause. Am J Transplant 2007; 7: Cosio FG, Grande JP, Wadei H et al. Predicting subsequent decline in kidney allograft function from early surveillance biopsies. Am J Transplant 2005; 5: Kidney International (2011) 80,

7 S Baid-Agrawal et al.: Overlapping pathways to transplant glomerulopathy 11. Gloor JM, Sethi S, Stegall MD et al. Transplant glomerulopathy: subclinical incidence and association with alloantibody. Am J Transplant 2007; 7: Mauiyyedi S, Pelle PD, Saidman S et al. Chronic humoral rejection: identification of antibody-mediated chronic renal allograft rejection by C4d deposits in peritubular capillaries. J Am Soc Nephrol 2001; 12: Regele H, Bohmig GA, Habicht A et al. 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 2002; 13: Vongwiwatana A, Gourishankar S, Campbell PM et al. Peritubular capillary changes and C4d deposits are associated with transplant glomerulopathy but not IgA nephropathy. Am J Transplant 2004; 4: Kieran N, Wang X, Perkins J et al. Combination of peritubular c4d and transplant glomerulopathy predicts late renal allograft failure. J Am Soc Nephrol 2009; 20: Hidalgo LG, Campbell PM, Sis B et al. De novo donor-specific antibody at the time of kidney transplant biopsy associates with microvascular pathology and late graft failure. Am J Transplant 2009; 9: Nickeleit V, Zeiler M, Gudat F et al. Detection of the complement degradation product C4d in renal allografts: diagnostic and therapeutic implications. J Am Soc Nephrol 2002; 13: Akalin E, Dinavahi R, Dikman S et al. Transplant glomerulopathy may occur in the absence of donor-specific antibody and C4d staining. Clin J Am Soc Nephrol 2007; 2: John R, Konvalinka A, Tobar A et al. Determinants of long-term graft outcome in transplant glomerulopathy. Transplantation 2010; 90: Racusen LC, Solez K, Colvin RB et al. The Banff 97 working classification of renal allograft pathology. Kidney Int 1999; 55: Papadimitriou JC, Drachenberg CB, Munivenkatappa R et al. Glomerular inflammation in renal allografts biopsies after the first year: cell types and relationship with antibody-mediated rejection and graft outcome. Transplantation 2010; 90: Ponticelli C, Banfi G. Transplant glomerulopathy: new clues in the puzzle of chronic allograft nephropathy? Am J Transplant 2003; 3: Andresdottir MB, Assmann KJ, Koene RA et al. Immunohistological and ultrastructural differences between recurrent type I membranoproliferative glomerulonephritis and chronic transplant glomerulopathy. Am J Kidney Dis 1998; 32: Pascual M, Perrin L, Giostra E et al. Hepatitis C virus in patients with cryoglobulinemia type II. J Infect Dis 1990; 162: Agnello V, Chung RT, Kaplan LM. A role for hepatitis C virus infection in type II cryoglobulinemia. N Engl J Med 1992; 327: Johnson RJ, Willson R, Yamabe H et al. Renal manifestations of hepatitis C virus infection. Kidney Int 1994; 46: Pascual M, Schiferli J. Hepatitis C virus infection and glomerulonephritis. In: Andreucci V, Fine L (eds). International Yearbook of Nephrology. Oxford University Press: New York, 1996, pp Sansonno D, Gesualdo L, Manno C et al. Hepatitis C virus-related proteins in kidney tissue from hepatitis C virus-infected patients with cryoglobulinemic membranoproliferative glomerulonephritis. Hepatology 1997; 25: D Amico G. Renal involvement in hepatitis C infection: cryoglobulinemic glomerulonephritis. Kidney Int 1998; 54: Roth D, Cirocco R, Zucker K et al. De novo membranoproliferative glomerulonephritis in hepatitis C virus-infected renal allograft recipients. Transplantation 1995; 59: Cruzado JM, Gil-Vernet S, Ercilla G et al. Hepatitis C virus-associated membranoproliferative glomerulonephritis in renal allografts. JAmSoc Nephrol 1996; 7: Baid S, Cosimi AB, Tolkoff-Rubin N et al. Renal disease associated with hepatitis C infection after kidney and liver transplantation. Transplantation 2000; 70: Gallay BJ, Alpers CE, Davis CL et al. Glomerulonephritis in renal allografts associated with hepatitis C infection: a possible relationship with transplant glomerulopathy in two cases. Am J Kidney Dis 1995; 26: Cosio FG, Roche Z, Agarwal A et al. Prevalence of hepatitis C in patients with idiopathic glomerulopathies in native and transplant kidneys. Am J Kidney Dis 1996; 28: Cruzado JM, Carrera M, Torras J et al. Hepatitis C virus infection and de novo glomerular lesions in renal allografts. Am J Transplant 2001; 1: Naesens M, Kuypers DRJ, Sarwal M. Calcineurin inhibitor nephrotoxicity. Clin J Am Soc Nephrol 2009; 4: Satoskar AA, Pelletier R, Adams P et al. De novo thrombotic microangiopathy in renal allograft biopsies-role of antibody-mediated rejection. Am J Transplant 2010; 10: Baid S, Pascual M, Williams Jr WW et al. Renal thrombotic microangiopathy associated with anticardiolipin antibodies in hepatitis C-positive renal allograft recipients. J Am Soc Nephrol 1999; 10: Baid S, Pascual M, Cosimi AB et al. Viruses and thrombotic microangiopathy. Transplantation 1999; 68: Wong W, Denton M, Rennke HG et al. Hepatitis C, proteinuria, and renal insufficiency. Am J Kidney Dis 2004; 44: Namba Y, Oka K, Moriyama T et al. Incidence of positive C4d deposition in long-term survival cases over 10 yr after renal transplantation. Clin Transplant 2006; 20(Suppl 15): Sis B, Jhangri GS, Bunnag S et al. Endothelial gene expression in kidney transplants with alloantibody indicates antibody-mediated damage despite lack of C4d staining. Am J Transplant 2009; 9: Sis B, Halloran PF. Endothelial transcripts uncover a previously unknown phenotype: C4d-negative antibody-mediated rejection. Curr Opin Organ Transplant 2010; 15: Rubin RH. Impact of Cytomegalovirus Infection on Renal Transplantation, 3rd edn. Marcel Dekker: New York, Cacoub P, Ghillani P, Revelen R et al. Anti-endothelial cell auto-antibodies in hepatitis C virus mixed cryoglobulinemia. J Hepatol 1999; 31: Baid-Agrawal S, Pascual M, Moradpour D et al. Hepatitis C virus infection in haemodialysis and kidney transplant patients. Rev Med Virol 2008; 18: Sijpkens YW, Joosten SA, Wong MC et al. Immunologic risk factors and glomerular C4d deposits in chronic transplant glomerulopathy. Kidney Int 2004; 65: Guesdon JL, Ternynck T, Avrameas S. The use of avidin-biotin interaction in immunoenzymatic techniques. J Histochem Cytochem 1979; 27: Hsu SM, Raine L. The use of avid-biotin-peroxidase complex (ABC) in diagnostic and research pathology. In: DeLellis RA (ed) Advances in Immunohistochemistry. Masson: New York, 1984, pp Schneeberger EE, Collins AB, Latta H et al. Diminished glomerular accumulation of colloidal carbon in autologous immune complex nephritis. Lab Invest 1977; 37: Collins AB, Schneeberger EE, Pascual MA et al. Complement activation in acute humoral renal allograft rejection: diagnostic significance of C4d deposits in peritubular capillaries. J Am Soc Nephrol 1999; 10: Cox DR. Regression models and life-tables. J R Stat Soc Series B Stat Methodol 1972; 34: Kidney International (2011) 80,