J Am Soc Nephrol 13: , 2002

Transcription

1 J Am Soc Nephrol 13: , 2002 Capillary Deposition of Complement Split Product C4d in Renal Allografts is Associated with Basement Membrane Injury in Peritubular and Glomerular Capillaries: A Contribution of Humoral Immunity to Chronic Allograft Rejection HEINZ REGELE,* GEORG A. BÖHMIG, ANTJE HABICHT, DANIELA GOLLOWITZER,* MARTIN SCHILLINGER, SUSANNE ROCKENSCHAUB, BRUNO WATSCHINGER, DONTSCHO KERJASCHKI,* and MARKUS EXNER *Clinical Institute of Pathology, Departments of Internal Medicine III, Internal Medicine II, Transplant Surgery, and Laboratory Medicine, University of Vienna, Vienna, Austria. Abstract. Endothelial deposition of the complement split product C4d is an established marker of antibody-mediated acute renal allograft rejection. A contribution of alloantibody-dependent immune reactions to chronic rejection is under discussion. In this study, the association of immunohistochemically detected endothelial C4d deposition in peritubular capillaries (PTC) with morphologic features of chronic renal allograft injury was investigated in a large study cohort. C4d deposits in PTC were detected in 73 (34%) of 213 late allograft biopsies performed in 213 patients more than 12 mo after transplantation (median, 4.9 yr) because of chronic allograft dysfunction. Endothelial C4d deposition was found to be associated with chronic transplant glomerulopathy (CG) (P ), with basement membrane multilayering in PTC (P 0.01), and with an accumulation of mononuclear inflammatory cells in PTC (P 0,001). Furthermore, C4d deposits in PTC (in biopsies with normal glomerular morphology) were associated with development of CG in follow-up biopsies. Other morphologic features of chronic allograft nephropathy (with exception of tubular atrophy) were not associated with C4d deposits in PTC. Analyses of previous and follow-up biopsies revealed that C4d deposits may occur de novo and may also disappear at any time after transplantation. In conclusion, the data suggest that complement activation in renal microvasculature, indicating humoral alloreactivity, contributes to chronic rejection characterized by chronic transplant glomerulopathy and basement membrane multilayering in PTC. Acute rejection is the most important threat to transplanted kidneys in the early phase after transplantation. Advances in understanding immunologic mechanisms underlying acute allograft rejection enabled the development of efficient diagnostic tools and therapeutic strategies against early immune-mediated graft loss. This led to an increase of 1-yr graft survival rates to more than 90%. Long-term graft survival did not, however, improve to a similar degree (1). A steady decline of renal function over years is still the rule in the majority of cases. Chronic allograft nephropathy (CAN) is a common morphologic finding in late allograft dysfunction. The term CAN (as defined by the Banff classification [2]) does not, Received March 20, Accepted May 20, Correspondence to Dr. Heinz Regele, Clinical Institute of Pathology, University of Vienna, Währinger Gürtel 18-20, A-1090 Vienna, Austria. Phone: ; Fax: ; heinz.regele@akh-wien.ac.at / Journal of the American Society of Nephrology Copyright 2002 by the American Society of Nephrology DOI: /01.ASN F however, describe a specific pathologic process; rather, it comprises various types of chronic injuries that may result from a variety of underlying pathogenic mechanisms. Both alloantigen-dependent and alloantigen-independent (e.g., drug toxicity, hypertension, recurrence of disease) mechanisms may be involved (1,3,4). Use of the term CAN reflects the limited ability to discriminate alloantigen-dependent damage from other causes of tissue injury (3). This uncertainty about pathogenic mechanisms of chronic allograft failure is a major obstacle to the development of specific therapeutic interventions. Chronic transplant glomerulopathy (CG) and distinct morphologic changes in arteries and capillaries have been reported as signs of chronic immunologic damage, i.e., true chronic rejection (2,5). The precise nature of the underlying immunologic mechanisms is, however, still unclear (6,7). Capillary deposition of C4d, reflecting complement activation via the classical pathway (8), is an established marker of antibody-mediated allograft rejection early after transplantation (9 16). In acute rejection, this complement split-product has been shown to be closely associated with the presence of alloantibodies (9 16). C4d was also identified in renal allo-

2 2372 Journal of the American Society of Nephrology J Am Soc Nephrol 13: , 2002 graft biopsies with morphologic signs of chronic rejection (5,17). Previous studies on C4d in chronically dysfunctioning allografts have suggested a pathogenic role of humoral immunity in a subset of patients (5,18). This finding has been challenged by others (19). Controversial results might at least partially be due to differences in the design of the cited studies, which were performed in highly selected patient groups (5), included small study populations (5,17), did not analyze allograft morphology (20 22), or made no clear distinction between early and late allograft biopsies (19). In this study, a large series of unselected late renal allograft biopsies was examined immunohistochemically for endothelial deposition of C4d as surrogate marker antibody-mediated alloreactivity in renal allografts. The aim of the study was to systematically analyze the incidence of C4d deposits in late allograft biopsies and to determine the pattern of histomorphologic changes in chronic humoral allograft injury. Materials and Methods Study Population From 1992 to 1998, a total of 1139 individuals were seen with a functioning graft in the outpatient department of our transplant unit. According to local practice, biopsies were not performed on a prospective protocol basis but were based on the clinical course of the patients (e.g., deterioration of kidney function, proteinuria, etc.). Of 1139 recipients, 213 were enrolled in this retrospective cohort study on the basis of the following inclusion criteria: (1) availability of at least one allograft biopsy (index biopsy) performed more than 12 mo after transplantation and during the observation period between July 1992 and September 1998; (2) availability of adequate tissue for histologic diagnosis and immunohistochemistry; and (3) patient follow-up (including all previous and subsequent allograft biopsies) at our institution. The study population consisted of patients transplanted between 1977 and In this period, 2339 patients were transplanted at our institution (University of Vienna, Austria). For patients with two or more biopsies performed more than 12 mo after transplantation, the first of late biopsies was enrolled as index-biopsy. Patients were not selected for indication of biopsy, clinical history, or specific diagnoses. Indications for biopsy were a significant increase of serum creatinine levels and/or de novo occurrence of proteinuria ( 1.5 g/24 h). In 96 patients, at least one biopsy preceding the index biopsy was available ( previous biopsy ). In 58 patients, follow-up biopsies were studied. In cases with more than one previous or follow-up biopsy, the first biopsy preceding or following the indexbiopsy was evaluated. In 43 cases, renal cortical tissue samples were available for morphologic analysis of peritubular capillaries (PTC) on electron microscopy (EM). Patient demographics and basic clinical data are listed in Table 1. Histopathology A total of 367 kidney allograft biopsies were evaluated by two independent observers. Pathohistologic evaluation was performed on formalin-fixed paraffin sections stained with hematoxylin and eosin, periodic acid-schiff, Methenamin-Silver, and with a trichrome stain (acidic-fuchsine-orange-g). Histologic lesions were classified and scored according to the Banff 97 classification (2) without previous disclosure of C4d staining results. Additional morphologic changes potentially related to acute humoral rejection (granulocytes or mononuclear inflammatory cells in PTC, thrombotic microangiopathy) were also scored semiquantitatively. The diagnosis of chronic cyclosporine toxicity was made only in cases with de novo arteriolopathy in absence of severe hypertension or diabetes mellitus. Immunohistochemistry and Immunofluorescence Generation, characterization, and staining properties of the polyclonal anti-c4d antibody, C4dpAb (Biomedica, Vienna, Austria), which is suitable for immunohistochemistry on paraffin sections, have been reported previously (9). On frozen sections, staining patterns obtained with C4dpAb are identical to those obtained with a commercially available monoclonal antibody (9). For immunohistochemical detection of C4d, 2- m sections were deparaffinized and endogenous peroxidase activity was blocked with hydrogen peroxide/methanol. Antigen retrieval was carried out by pressure-cooking for 10 min at 1 bar in citrate-buffer (ph 6.0) as described previously (9). Endogenous biotin was blocked using a Biotin blocking kit (Vector Laboratories, Burlingame, CA). After 30-min incubation with C4dpAb (1:50), bound IgG was visualized using the Supersensitive Kit (BioGenex, San Ramon, CA) according to the manufacturer s protocol. PTC staining was categorized semiquantitatively according to the extent and intensity of staining. Scoring was performed without knowledge of histologic diagnoses or clinical history of recipients. Biopsies with circumferential linear endothelial C4d staining in at least 25% of PTC were designated as C4d PTC -positive, all other biopsies as C4d PTC -negative. This definition and the relevance of a cut-off point of 25% for positive staining were validated in a previous study by flow-cytometric detection of circulating antibodies (12). Table 1. Patient demographics and clinical parameters a Clinical Parameter All Patients (n 213) C4d in PTC (Index Biopsy) Yes (n 73) No (n 140) Male patients, n (%) 123/213 (58) 43/73 (59) 81/140 (57) Recipient age (yr), median (IQR) 45 (34 to 61) 44 (32 to 54) 49 (35 to 61) Year of TX, median (IQR) 1990 (1988 to 1993) 1990 (1987 to 1993) 1990 (1987 to 1993) Biopsy interval after TX b (yr), median (IQR) 4.9 (2.4 to 7.5) 4.8 (2.4 to 7.0) 4.9 (2.4 to 7.2) First transplant, n (%) 171/213 (80) 54/73 (74) 117/140 (84) Previous episodes of acute rejection, c n (%) 50/213 (23) 13/73 (18) 37/140 (26) a IQR, interquartile range; PTC, peritubular capillaries; TX, transplantation. b The first biopsy taken later than 1 yr was included into the study. c Biopsy-proven episodes of acute rejection.

3 J Am Soc Nephrol 13: , 2002 Complement Activation in Chronic Rejection 2373 Immunofluorescent double staining was carried out on paraffin sections of ten C4d PTC -positive and 5 C4d PTC -negative cases. Sections were pretreated as described above and were incubated with C4dpAb (1:10) together with an anti-collagen type IV (1:50) (DAKO, Glostrup, Denmark) or an anti-cd31 (1:20; DAKO) monoclonal antibody for 30' at room temperature. Bound primary antibody was detected with an Alexa-486 conjugated goat anti-rabbit IgG antibody (1:1000; Molecular Probes, Eugene, OR) for C4dpAb or an Alexa- 594-conjugated goat anti-mouse IgG antibody (1:1000; Molecular Probes) for anti-collagen IV and anti-cd31. Electron Microscopy Electron microscopy could be performed in 43 of 213 index biopsies. Biopsies were analyzed for the presence of basement membrane multilayering in PTC (MLPTC). Biopsy specimens were fixed in 4% paraformaldehyde in 0.1 M sodium cacodylate buffer with 0.1% glutaraldehyde, post-fixed with 1% OsO 4 in 0.1 M sodium cacodylate buffer, and stained with 1% aqueous uranyl acetate. Specimens were embedded in epoxy-resin (Serva electrophoresis, Heidelberg, Germany) (polymerization for 36 h at 60 C). Ultrathin sections were prepared on an Ultracut-E Ultramicrotome (Reichert Jung, Vienna, Austria). After staining with uranyl acetate in methanol and with lead citrate, sections were examined with a JEOL JEM 1010 transmission electron microscope (JEOL, Tokyo, Japan). On the basis of the examination of at least six PTC per biopsy,we categorized MLPTC as being absent if no or only focal splitting of basement membranes (BM) was present, as low-grade MLPTC if three to four BM layers could be observed (with more severe lesions only focally distributed). High-grade MLPTC was defined by five and more BM layers as predominant lesion. Immune Electron Microscopy Biopsy specimens obtained from four patients with C4d deposits detected by conventional immunohistochemistry were fixed in freshly prepared 4% formaldehyde in PBS for 8 h and then embedded in Lowicryl K4M (Chemische Werke LOWI, Waldkraiburg, Germany) (23). Ultrathin sections were blocked with 1% bovine serum albumin in PBS for 1 h followed by incubation with C4dpAb (diluted 1:1 in 1% BSA, 0.5% Tween-20, 0.1% Triton-X100 in PBS) for 20 h at 4 C. After washing in PBS, slides were incubated for 1 h at room temperature with 10 nm of gold-labeled goat-anti-rabbit IgG (Amersham Pharmacia Biotech, Little Chalfont, UK). Sections were stained with 2% aqueous uranyl acetate and lead citrate (24) and examined with a JEOL JEM 1010 transmission electron microscope. Statistical Analyses Data are presented as the median and interquartile range (range from the 25 th to the 75 th percentile). Percentages were calculated for dichotomous variables. The 2 test or, if appropriate, Fisher s exact test were applied for comparison of proportions. The cumulative proportion of CG-free patients in C4d-postitive and C4d-negative subjects during the follow-up time interval was calculated by the Kaplan Meier method and compared by means of the log rank test. A P value 0.05 was accepted as statistically significant; all P values were two sided. Calculations were performed with SPSS for Windows (Version 10.0; SPSS Inc, Chicago, IL). Results Linear C4d Staining in Peritubular Capillaries of Late Allograft Biopsies Seventy-three of the 213 index biopsies (34%) showed linear C4d deposition along endothelial cells of PTC (Figure 1A). C4d PTC -positive cases did not differ significantly from those without C4d deposits regarding gender, age, date of transplantation, time from transplantation to index-biopsy, number of previous transplants, or incidence of acute rejection within the first year after transplantation (Table1). Linear C4d Staining in Glomerular Capillaries is Uncommon In 13 (6%) of 213 index biopsies, linear C4d deposits were also observed along glomerular endothelial cells. In all cases with glomerular endothelial staining, C4d deposits were also present in PTC. In addition, C4dpAb detected glomerular immune deposits in 30 cases of immune complex-mediated glomerulonephritis (GN). Our evaluations were based on C4d staining in PTC (C4d PTC ) C4d Staining in PTC of Allograft Biopsies Reflects Endothelial C4d Deposits To determine the exact localization of C4d deposits in PTC, immune electron microscopy and immunofluorescent double staining were performed in selected C4d PTC -positive biopsies. By immune electron microscopy, antibody-conjugated gold particles were located at the luminal surface of peritubular capillary endothelial cells, between endothelial cells and the basement membrane of PTC and also within some intracellular vesicles (Figure 2). A perfect co-localization of C4d with the endothelial cell marker CD31, demonstrated by immunofluorescent double staining, confirms that C4d is bound to endothelial cells (Figure 1, C through E), whereas double labeling for C4d and collagen IV (as marker for basement membranes) only reveals focal overlap but no true co-localization of staining signals (Figure 1, F through H). C4d Staining in PTC is Associated with Chronic Transplant Glomerulopathy (CG) Next, the 213 index biopsies were analyzed for associations of C4d deposition in PTC with morphologic signs of rejection. C4d staining was neither associated with the diagnosis of acute rejection (as defined by the Banff classification) nor with the presence of CAN (as defined by the Banff classification). The former was a rare finding (7% of index-biopsies), whereas the latter was found in the majority of both C4d PTC -positive (92%) and C4d PTC -negative (87%) biopsies (Tables 2 and 3). Analysis of individual criteria defining Banff chronic allograft nephropathy revealed a highly significant association of C4d PTC staining with CG (Figure 3A) (53% of C4d PTC -positive versus 14% of C4d PTC -negative biopsies; P ). Glomerular endothelial C4d staining, a rare finding in general (6%), was observed in 12% of the biopsies showing CG. Tubular atrophy was present in nearly all C4d PTC -positive biopsies (99%) and was less common in biopsies without C4d PTC deposits (89%; P 0.01). Interstitial fibrosis, intimal fibrosis, and mesangial sclerosis of glomeruli were not linked to the presence or absence of C4d PTC staining (Table 2). Analyzing individual diagnostic lesions of acute rejection, we found a higher percentage of cases with glomerulitis (20% of C4d PTC -positive versus 9% of C4d PTC -negative

4 2374 Journal of the American Society of Nephrology J Am Soc Nephrol 13: , 2002 Figure 1. Immunohistochemical detection of endothelial C4d deposits in peritubular capillaries (PTC) on paraffin sections of renal allograft biopsies. (A) Linear C4d deposits along the walls of PTC. (B) C4d-positive PTC congested with mononuclear inflammatory cells (triangles) (detail from A). (E) Yellow staining along the inner aspect of PTC indicating perfect co-localization of green staining for C4d (C) and red staining for CD31 (labeling endothelial cells) (D). (H) PTC showing only focal overlap (yellow staining) but no widespread co-localization of green staining signals for C4d (F) and red staining for collagen type IV (G). Both tubular and capillary (triangles) basement membranes are labeled by an anti-collagen type IV antibody (F and G). (A and B) Indirect immunoperoxidase staining with C4dpAb on paraffin sections. (C through G) Indirect immunofluorescent double-staining of paraffin sections with C4dpAb, monoclonal anti-cd31, and monoclonal anticollagen type IV; C, peritubular capillary; T, tubule. Magnifications: 275 in A; 550 in B; 230 in C, D, F and G; 460 in E and F. biopsies; P 0.02) or (predominantly low-grade) tubulitis (37% of C4d PTC -positive versus 24% of C4d PTC -negative biopsies; P 0.04) in C4d PTC -positive biopsies. Interstitial inflammation and arterial intimal arteritis were not associated with C4d deposition in PTC. Neither an accumulation of granulocytes in PTC nor thrombotic microangiopathy (both reported as being typical for acute humoral rejection) were found to be more common in biopsies with C4d PTC deposits. In our study population, however, these lesions were very rare findings (Table 3). C4d Staining in PTC is Associated with Multilayering of Basement Membranes in PTC Lamellar splitting and thickening of the BM (Figure 3, C and D) represent the ultrastructural lesions characterizing basement membrane multilayering in PTC (MLPTC). Suitable material

5 J Am Soc Nephrol 13: , 2002 Complement Activation in Chronic Rejection 2375 Figure 2. Immune electron microscopy with C4dpAb on a representative C4d PTC -positive renal allograft specimen. Antibody-conjugated gold particles are located at the luminal surface of PTC (A, B, C) and between endothelial cells and the basement membrane (A, C and D as indicated by arrows). Gold particles were also found within some intracellular vesicles (D). CL, capillary lumen; N, nucleus; v, vesicle. for EM was available for 43 cases of our study population. MLPTC was identified in 28 specimens (Table 4). Twenty-one of 43 biopsies were C4d PTC -positive (Table 4). Analyzing at least six PTC per biopsy, we quantified the lesion on the basis of the number of newly formed lamellae. MLPTC was more common on C4d PTC -positive biopsies (Table 4). In particular, high-grade MLPTC demonstrated a strong association with C4d PTC deposits (15 of 21 in C4d PTC -positive cases versus 3of 22 in cases without C4d PTC ; P 0.001). MLPTC was also strongly associated with CG (16 of 18 cases with MLPTC also showed CG; P 0.009) Mononuclear Inflammatory Cells Accumulate in C4d- Positive PTC A common finding in C4d PTC -positive biopsies was a conspicuous accumulation of mononuclear cells in (frequently dilated) lumina of PTC (Figure 1B) (51% of C4d PTC -positive versus 25% of C4d PTC -negative biopsies; P 0.001). Fortythree percent (31 of 72) of biopsies with accumulation of mononuclear cells in PTC also showed tubulitis. Acute interstitial rejection (Banff I) was, however, diagnosed in only four cases because tubulitis and interstitial inflammation were mild in most biopsies (t1 and i1 according to Banff). In biopsies without increased numbers of inflammatory cells in PTC, tubulitis occurred in only 22% (31 of 141; P 0.02) of cases. Chronic cyclosporine toxicity showed no association with C4d deposits in PTC (Table 2). Glomerulonephritis (membranous GN, n 16; IgA nephritis, n 11; mesangial proliferative GN, n 5; membranoproliferative GN, n 1; crescentic GN, n 1) was more commonly observed in C4d PTC -positive biopsies (25% of C4d PTC -positive versus 11% of C4d PTC - negative biopsies; P 0.02) (Table 3). C4d Staining is Variable in Sequential Biopsies To evaluate the dynamics of endothelial C4d deposition over time, we examined allograft biopsies preceding index biopsies (96 cases) and first follow-up biopsies after the index biopsy (58 cases) for the presence of C4d PTC deposits. In 31 cases, all three biopsy types (previous, index, and follow-up) were available. Analysis of all 367 biopsies revealed a time interval from transplantation to the first biopsy with C4d PTC deposits ranging from 1 wk to 17.4 yr. We found C4d PTC to be highly variable over time. In many cases C4d PTC deposits developed de novo several years after transplantation. In some cases, disappearance of C4d PTC deposits was observed in follow-up biopsies (Table 5). Thirty-four of 96 index biopsies for which previous biopsies were available, were C4d PTC -positive. In only eight (24%) of these cases, C4d PTC deposits were present already in a previous biopsy (Table 5). On the other hand, in 5 of 13 patients who had C4d PTC detectable in their previous biopsy (performed within the first year after transplantation), no more C4d PTC was present in the subsequent index biopsy (Table 5). A similar variability of C4d PTC staining was observed between index biopsy and follow-up biopsies (Table 5). Eight (38%) out of 21 patients with a C4d PTC -positive index biopsy were C4d PTC -negative in follow-up biopsies. After index biopsy, immunosuppressive therapy was modified in only one of the eight recipients (conversion to MMF). Furthermore, dosage of immunosuppressive drugs was not increased after index biopsy. The median time interval between C4d PTC -positive and subsequent C4d PTC -negative biopsies was 16 mo (range, 6 to 61). Twelve (48%) out of 25 patients with a C4d PTC -positive follow-up biopsy were C4d PTC -negative in their corresponding index biopsy (Table 5). In only three patients (out of a group of 31 with 3 sequential biopsies) C4d PTC was constantly present from the first (previous) to the last (follow-up) biopsy. C4d Staining Is Associated with Progression to CG in Subsequent Biopsies C4d PTC deposition in biopsies with unaffected glomeruli showed a significant association with the development of CG in follow-up biopsies (Figure 4). In nine (82%) of eleven allografts (with follow-up biopsy available) showing C4d PTC deposits but no detectable CG in the index biopsy, progression to CG could be observed in the follow-up biopsy. Of 33 C4d PTC -negative cases (in index biopsy), only 9 (27%) developed CG in a follow-up biopsy (log rank test, P 0.04) (Table 6). Similarly, 6 (46%) of 13 patients with C4d PTC -positive previous biopsies (performed during first year after transplantation) developed CG in a subsequent C4d PTC -negative index biopsy. In only 5 (6%) of 82 C4d PTC -negative previous biopsies, CG was detectable in the following (still C4d PTC -negative) index biopsy (log rank test, P ). Discussion Chronic allograft dysfunction represents a major clinical problem after transplantation. It is widely accepted that pathogenic mechanisms underlying CAN are heterogeneous in nature (1 4). Both alloantigen-dependent (true chronic rejection) and alloantigen-independent mechanisms may result in very similar patterns of injury. The term chronic allograft nephropathy (CAN) has been coined for this common but nonspecific morphologic appearance of chronically dysfunctioning allo-

6 2376 Journal of the American Society of Nephrology J Am Soc Nephrol 13: , 2002 Table 2. C4d PTC staining and morphologic signs of chronic allograft injury a Pathohistology Index Biopsies (n 213) C4d in PTC (Index Biopsy) Yes (n 73) No (n 140) P b Chronic allograft nephropathy, n (%) 189/213 (87) 67/73 (92) 122/140 (87) 0.4 Chronic cyclosporine toxicity, n (%) 26/213 (12) 8/73 (11) 18/140 (13) 0.8 Chronic glomerulopathy, c n (%) 58/213 (27) 39/73 (53) 19/140 (14) Interstitial fibrosis, c n (%) 207/213 (97) 73/73 (100) 134/140 (96) 0.1 Tubular atrophy, c n (%) 197/213 (93) 72/73 (99) 125/140 (89) 0.01 Intimal fibrosis, c n (%) 169/213 (79) 56/73 (77) 113/140 (81) 0.6 Mesangial sclerosis, c n (%) 101/213 (48) 36/73 (49) 65/140 (46) 0.7 Arteriolar hyalinosis, c n (%) 139/213 (65) 42/73 (56) 97/140 (69) 0.07 a PTC, peritubular capillaries. b 2 test, C4d PTC -positive versus C4d PTC -negative subgroup. c Lesions are defined and scored according to the Banff 97 classification. Table 3. C4d PTC staining and histomorphology a Pathohistology Index Biopsies (n 213) C4d in PTC (Index Biopsy) Yes (n 73) No (n 140) P b Banff I, II, or III, n (%) 13/213 (6) 6/73 (8) 7/140 (5) 0.4 Glomerulitis, c n (%) 27/213 (13) 15/73 (20) 12/140 (9) 0.02 Interstitial inflammation, c n (%) 135/213 (63) 49/73 (67) 86/140 (61) 0.5 Tubulitis, c n (%) 62/213 (29) 28/73 (37) 34/140 (24) 0.04 Intimal arteritis, c n (%) 5/213 (2) 3/73 (4) 2/140 (1) 0.3 Granulocytes in PTC, n (%) 8/213 (4) 5/73 (7) 3/140 (2) 0.1 Mononuclear cells in PTC, n (%) 72/213 (34) 37/73 (51) 35/140 (25) Thrombotic microangiopathy, n (%) 5/213 (2) 3/73 (3) 2/140 (1) 0.3 Glomerulonephritis, n (%) 34/213 (14) 18/73 (25) 16/140 (11) 0.02 a PTC, peritubular capillaries. b 2 test, C4d-positive versus C4d-negative subgroup. c Lesions are defined and scored according to the Banff 97 classification. grafts (2,3). Reliable identification of the various conditions that lead to morphologic and functional damage is a prerequisite for specific therapeutic interventions. In most patients with chronic allograft dysfunction, however, estimation of the relative contribution of individual damaging processes to a given injury (3) is a very difficult task. It is generally accepted that alloimmune mechanisms are among the major causes of chronic allograft injury (1,7). The exact nature of chronic alloimmunity and its contribution to graft injury are, however, far from being defined, and reliable diagnostic criteria are lacking (2,3). Several studies emphasize an important role of alloantibody-mediated processes in chronic rejection (5,7,17,20 22,25). Some of these studies employed C4d staining as a marker of humoral immunity (5,17). A recent report, however, failed to confirm a link between C4d and chronic allograft injury (19). A novel anti-c4d antibody (9) suitable for detection of C4d on paraffin sections enabled this retrospective study on archival biopsies. Protocol biopsies of kidneys with stable function are not available at our center; the study was therefore by definition limited to patients with functional impairment of the renal allograft. In this study population, C4d deposits were detected in 34% of late allograft biopsies. Previously, Mauiyeddi et al. (5) reported C4d deposits in 61% of cases with histomorphologic evidence of chronic rejection (i.e., transplant glomerulopathy or arteriopathy), which is quite similar to our observation of 53% C4d PTC -positivity in biopsies with transplant glomerulopathy. Only 13% of C4d PTC -positive biopsies were observed by same group in a separate analysis of unselected cases of chronic allograft dysfunction (26). The true incidence of C4d deposits cannot be deduced from our data. This issue has to be addressed in a study employing protocol biopsies. However, the results of this study clearly demonstrate a close link between complement activation on capillary endothelial cells of renal allografts and morphologic signs of chronic rejection. C4d deposits on endothelial cells of renal microvessels can be observed in a fraction of allograft biopsies but not in native kidneys specimens (9,27). Co-localization of C4d with CD31 in immunofluorescent double staining confirms the endothelial

7 J Am Soc Nephrol 13: , 2002 Complement Activation in Chronic Rejection 2377 Figure 3. (A) Chronic transplant glomerulopathy with widespread basement membrane reduplication (Methenamin Silver staining on paraffin section). (B, C, and D) PTC of late allograft biopsies in electron microscopy. (B) Normal PTC with single-layered delicate basement membrane (triangles). (C) Low-grade multilayering in PTC (MLPTC) with slight basement membrane splitting (triangles). (D) High-grade MLPTC with severe thickening and multi-lamellation (triangles) of basement membrane. Transmission electron microscopy, CL, capillary lumen of peritubular capillaries; Magnifications: 600 in A; 12,400 in B; 7300 in C; 6100 in D. localization of C4d deposits. As shown by immune electron microscopy, C4d deposits are restricted to the surface (i.e., the luminal and also the abluminal cell membrane), and sometimes to intracellular vesicles of endothelial cells (possibly indicating a transcytotic process). This suggests targeted activation of complement rather than passive precipitation of serum components. Several lines of evidence suggest a close relation of immunohistochemically detectable C4d and the presence of donor-specific antibodies; C4d deposits in renal allograft biopsies are associated with pretransplant (9,12,14) and posttransplant (12 14,16) detection of circulating anti-hla antibodies. In addition, removal of circulating anti-donor antibodies by immunoadsorption leads to restoration of renal function in otherwise therapy-resistant C4d-positive acute rejection (10,11). Endothelial C4d deposits seem, however, to be unevenly distributed throughout the renal microvasculature. The infrequent detection of C4d deposits on glomerular endothelial cells (even in cases with CG) might be due to biologic mechanisms that counteract the persistence of complement deposits on glomerular endothelial cells, rendering endothelial C4d deposits undetectable by immunohistochemistry. A possible explanation could be an enhanced glomerular endothelial regeneration, resulting in a higher rate of endothelial cell turnover, which leads to a more efficient elimination of complementloaded endothelial cells in glomeruli than in PTC (28). This study established a close link of endothelial C4d deposition with chronic transplant glomerulopathy (CG) and multilayering of basement membranes in PTC (MLPTC). MLPTC has previously been described in association with antibodymediated immune mechanisms (18,29). Presumably, the association of C4d PTC with CG was not detected in previous studies either because CG was one of the inclusion criteria (5,17) or late allograft biopsies were analyzed together with early allograft biopsies (with a very low incidence of CG), thus obscuring the association between C4d PTC and CG (19). CG and MLPTC are commonly regarded as signs of chronic allograft rejection in humans as well as in experimental models (5,29 33). These lesions tend to occur simultaneously (29,30,33,34) and show similar patterns of basement membrane thickening and reduplication, which are believed to result from endothelial injury (29,30,33). Complement activation (reflected by endothelial C4d deposition) is likely to play a major causative role in endothelial injury of allografts. It might thus represent the pathogenic link between CG and MLPTC, thereby explaining the frequently observed simultaneous occurrence of these two morphologic lesions. In view of the close relation of endothelial C4d deposition with serologically detectable alloantibodies (5,10,12 14) it is very likely that endothelial-bound alloantibodies (inducing complementmediated development of CG and MLPTC) represent an important driving force in chronic rejection. Other morphologic signs of CAN (with the exception of tubular atrophy being more common in C4d PTC -positive cases) were not associated with C4d PTC. This is also true for arterial intimal fibrosis, a lesion being potentially related to chronic rejection (2,5). Several non alloantigen-related conditions (e.g., hypertension, donor-derived arteriosclerosis) may, however, also contribute to intimal fibrosis. There also was no association of C4d PTC staining with acute rejection (defined according to the Banff classification), which is of course rare in late allograft biopsies. Analysis of individual morphologic criteria for acute cellular and/or humoral rejection revealed only tubulitis (mostly as low-grade lesion and therefore not sufficient for diagnosis of acute rejection Banff type I) and glomerulitis to be more common in C4d PTC - positive biopsies. A remarkable finding was a frequently observed accumulation of mononuclear cells (but not of granulocytes as observed in acute humoral rejection) in PTC of C4d PTC -positive biopsies. A study on ultrastructural alterations of PTC in antibodymediated rejection (mainly dealing with the acute phase of rejection) reported a similar accumulation of not further specified mononuclear inflammatory cells in PTC in a few biopsies performed 8 mo after transplantation (29). We failed, however, to establish an association between either inflammatory cells or C4d PTC deposition and interstitial rejection. The pathogenic or

8 2378 Journal of the American Society of Nephrology J Am Soc Nephrol 13: , 2002 Table 4. C4d PTC staining and peritubular capillary basement membrane lesions in electron microscopy a Morphological Lesion All Biopsies (n 43) C4d in PTC Yes (n 21) No (n 22) P b MLPTC (low- and high-grade), n (%) 28/43 (65) 18/21 (86) 10/22 (45) 0.01 MLPTC (only high-grade), n (%) 18/43 (42) 15/21 (71) 3/22 (14) a MLPTC, basement membrane multilayering in PTC; PTC, peritubular capillaries. b 2 test, C4d PTC -positive versus C4d PTC -negative subgroup. Table 5. C4d PTC staining in previous and follow-up biopsies a C4d PTC -Positive Index Biopsy, n (%) C4d PTC -Negative Time to/from Index Biopsy (mo), median (IQR) Previous biopsy b C4d PTC -positive, n (%) 13/96 (14) 8/13 (62) 5/13 (38) 20 (13 to 58) C4d PTC -negative, n (%) 83/96 (86) 26/83 (31) 57/83 (69) 44 (19 to 67) Follow-up biopsy b C4d PTC -positive, n (%) 25/58 (43) 13/25 (52) 12/25 (48) 26 (10 to 45) C4d PTC -negative, n (%) 33/58 (57) 8/33 (24) 25/33 (76) 15 (8 to 29) a IQR, interquartile range; PTC, peritubular capillaries. b Latest previous biopsy (taken within the first year) and earliest follow-up biopsy were investigated, patients with previous biopsies available are not necessarily identical with patients with follow-up biopsies. Figure 4. Kaplan Meier curve displaying the cumulative frequency of continuously chronic transplant glomerulopathy (CG) free patients after the early biopsy (until the index biopsy) and after the index biopsy (until the late follow-up biopsy) according to the absence or presence of C4d. diagnostic relevance of inflammatory cells accumulating in PTC is therefore currently not clear. In acute humoral rejection, endothelial C4d deposition mostly develops within few days or weeks after transplantation (9,12,14,15). There is scarce information, however, about the dynamics of C4d deposits in later phases after transplantation. C4d deposits were found to occur de novo at virtually any time after transplantation. In certain cases, C4d PTC deposition was detectable in sequential biopsies over years. On the other hand, there were a considerable number of C4d PTC -positive cases without C4d PTC depositions in subsequent biopsies. The trigger for the occurrence of alloantibodies late after transplantation is yet unclear; however, the indirect pathway of allorecognition could be one attractive candidate. In this pathway recipient CD4( ) T cells are activated by alloantigen when presented by self antigen-presenting-cells (35). Its relevance in renal transplant recipients with chronic allograft dysfunction has been shown earlier (36), and it was suggested that indirect allorecognition is crucial for effector mechanisms of allograft destruction to occur (i.e., DTH, alloantibodies, cytotoxicity) (37). Recent studies in animal models for cardiac Table 6. Chronic transplant glomerulopathy in index biopsies and follow-up biopsies a Index Biopsy n (%) a Follow-up Biopsy, b n (%) CG-Positive CG-Negative Time from Index Biopsy (mo), median (IQR) C4d PTC -pos. CG-pos. 10/21 (48) 10/10 (100) 0/10 (0) 16 (12 to 32) C4d PTC -pos. CG-neg. 11/21 (52) 9/11 (82) 2/11 (18) 23 (13 to 47) C4d PTC -neg. CG-pos. 4/37 (11) 2/4 (50) 2/4 (50) 8,9,15,45 a C4d PTC -neg. CG-neg. 33/37 (89) 9/33 (27) 24/33 (73) 19 (16 to 39) a CG, chronic allograft nephropathy; IQR, interquartile range; PTC, peritubular capillaries. b Follow-up biopsies were performed 8, 9, 15, and 45 mo after index biopsy.

9 J Am Soc Nephrol 13: , 2002 Complement Activation in Chronic Rejection 2379 allograft vasculopathy elegantly demonstrate that indirect allorecognition and CD4( ) T cells are crucial for the development of transplant arteriosclerosis and that alloantibody production is dependent on CD4( ) T cell help after indirect allorecognition (38 40). Higher frequencies of indirectly activated CD4( ) T cells were reported in patients with chronic allograft nephropathy, whereas significant titers of alloantibodies could only be detected in a minority of patients (41). It is thus tempting to speculate that deposition of C4d in biopsies of patients with chronic allograft rejection may represent a remaining trace of alloantibodies, originally produced via the help of indirectly primed allo-specific CD4( ) T cells. Given the strong association of C4d PTC and CG, we were particularly interested in the follow-up of cases showing normal glomerular morphology despite the presence of C4d PTC deposits. Interestingly, progression to CG in a subsequent biopsy was significantly more common in C4d PTC -positive than in C4d PTC - negative cases. This further underscores a close association of endothelial complement activation and the development of CG. CG (42) and severe proteinuria (43 45) (being usually associated with CG) are predictors of an unfavorable outcome. Detection of C4d in allograft biopsies could therefore be of diagnostic and therapeutic relevance, because complement activation (presumably reflecting antibody-mediated alloreactivity) precedes the evolution of structural damage in glomeruli and PTC. In recent studies, plasmapheresis together with tacrolimus-mycophenolate or treatment with immunoadsorption were proposed as anti-humoral therapy in acute rejection (46). The feasibility of a therapy for chronic C4d-positive kidney allograft rejection is not yet established. Recently, Theruvath et al. (17) demonstrated that rescue therapy with tacrolimus and mycophenolate mofetil effectively reduces posttransplant DSA levels in chronic humoral rejection. The clinical efficacy of this approach will have to be evaluated in a larger prospective study. In conclusion, this study reveals a close association of endothelial C4d deposits with CG and MLPTC (both being regarded as signs of chronic allograft rejection) and with an accumulation of mononuclear inflammatory cells in PTC. These findings support the previously suggested contribution of humoral immune mechanisms to chronic rejection. In most cases, C4d PTC deposition in late allograft biopsies does not reflect persistence of early onset humoral injury but rather may occur de novo at any time after transplantation. C4d PTC deposition seems to be a reliable marker of humoral alloreactivity; its detection might therefore be a simple and inexpensive way to identify patients who are likely to benefit from anti-humoral therapy. Acknowledgments The authors wish to thank Katalin Nagy-Bojarszky and Ingrid Raab for developing and applying immunohistocemical staining methods and Brigitte Langer, Anny Howorka, and Helga Poczewski for excellent technical assistance in electron microscopy. References 1. Kreis HA, Ponticelli C: Causes of late renal allograft loss: Chronic allograft dysfunction, death, and other factors. Transplantation 71: S5 S9, Racusen LC, Solez K, Colvin RB, Bonsib SM, Castro MC, Cavallo T, Croker BP, Demetris AJ, Drachenberg CB, Fogo AB, Furness P, Gaber LW, Gibson IW, Glotz D, Goldberg JC, Grande J, Halloran PF, Hansen HE, Hartley B, Hayry PJ, Hill CM, Hoffman EO, Hunsicker LG, Lindblad AS, Yamaguchi Y: The Banff 97 working classification of renal allograft pathology. Kidney Int 55: , Halloran PF, Melk A, Barth C: Rethinking chronic allograft nephropathy: The concept of accelerated senescence. JAmSoc Nephrol 10: , Paul LC: Chronic allograft nephropathy: An update. Kidney Int 56: , Mauiyyedi S, Pelle PD, Saidman S, Collins AB, Pascual M, Tolkoff-Rubin NE, Williams WW, Cosimi AA, Schneeberger EE, Colvin RB: Chronic humoral rejection: Identification of antibodymediated chronic renal allograft rejection by C4d deposits in peritubular capillaries. J Am Soc Nephrol 12: , Hayry P, Aavik E, Savolainen H: Mechanisms of chronic rejection. Transplant Proc 31: 5S 8S, Paul LC: Immunologic risk factors for chronic renal allograft dysfunction. Transplantation 71: S17 S23, Zwirner J, Felber E, Herzog V, Riethmüller G, Feucht HE: Classical pathway of complement activation in normal and diseased human glomeruli. Kidney Int 36: , Regele H, Exner M, Watschinger B, Wenter C, Wahrmann M, Österreicher C, Säemann MD, Mersich N, Hörl WH, Zlabinger GJ, Böhmig GA: Endothelial C4d deposition is associated with inferior kidney allograft outcome independently of cellular rejection. Nephrol Dial Transplant 16: , Böhmig GA, Regele H, Säemann MD, Exner M, Druml W, Kovarik J, Hörl WH, Zlabinger GJ, Watschinger B: Role of humoral immune reactions as target for antirejection therapy in recipients of a spousal-donor kidney graft. Am J Kidney Dis 35: , Böhmig GA, Regele H, Exner M, Derhartunian V, Kletzmayr J, Säemann MD, Hörl WH, Druml W, Watschinger B: C4d-positive acute humoral renal allograft rejection: Effective treatment by immunoadsorption. J Am Soc Nephrol 12: , Böhmig GA, Exner M, Habicht A, Schillinger M, Lang U, Kletzmayr J, Säemann MD, Hörl WH, Watschinger B, Regele H: Capillary C4d deposition in kidney allografts: A specific marker of alloantibodydependent graft injury. J Am Soc Nephrol 13: , Collins AB, Schneeberger EE, Pascual MA, Saidman SL, Williams WW, Tolkoff-Rubin N, Cosimi AB, Colvin RB: Complement activation in acute humoral renal allograft rejection: diagnostic significance of C4d deposits in peritubular capillaries. J Am Soc Nephrol 10: , Crespo M, Pascual M, Tolkoff-Rubin N, Mauiyyedi S, Collins AB, Fitzpatrick D, Farrell ML, Williams WW, Delmonico FL, Cosimi AB, Colvin RB, Saidman SL: Acute humoral rejection in renal allograft recipients: I. Incidence, serology and clinical characteristics. Transplantation 71: , Feucht HE, Schneeberger H, Hillebrand G, Burkhardt K, Weiss M, Riethmüller G, Land W, Albert E: Capillary deposition of C4d complement fragment and early renal graft loss. Kidney Int 43: , Lederer SR, Kluth-Pepper B, Schneeberger H, Albert E, Land W, Feucht HE: Impact of humoral alloreactivity early after transplantation on the long-term survival of renal allografts. Kidney Int 59: , Theruvath TP, Saidman SL, Mauiyyedi S, Delmonico FL, Williams WW, Tolkoff-Rubin N, Collins AB, Colvin RB, Cosimi AB, Pascual M: Control of antidonor antibody production with

10 2380 Journal of the American Society of Nephrology J Am Soc Nephrol 13: , 2002 tacrolimus and mycophenolate mofetil in renal allograft recipients with chronic rejection. Transplantation 72: 77 83, Takeuchi O, Oikawa T, Koyama K, Usami T, Shimano Y, Ito A, Katoh M, Otsuka Y, Takeda A, Haba T, Uchida K, Morozumi K: Multilayering of peritubular capillary is a specific diagnostic criterion for immunologic chronic rejection: Does a humoral factor contribute to the pathogenesis of peritubular capillary lesions in chronic rejection? Transplant Proc 32: , Nickeleit V, Zeiler M, Gudat F, Thiel G, Mihatsch MJ: Detection of the complement degradation product c4d in renal allografts: Diagnostic and therapeutic implications. J Am Soc Nephrol 13: , Piazza A, Poggi E, Borrelli L, Servetti S, Monaco PI, Buonomo O, Valeri M, Torlone N, Adorno D, Casciani CU: Impact of donorspecific antibodies on chronic rejection occurrence and graft loss in renal transplantation: Posttransplant analysis using flow cytometric techniques. Transplantation 71: , Kerman RH, Susskind B, Kerman DH, Lam M, Gerolami K, Williams J, Kalish R, Campbell M, Katz SM, Van Buren CT, Kahan BD: Anti-HLA antibodies detected in posttransplant renal allograft recipient sera correlate with chronic rejection. Transplant Proc 29: , Kimikawa M, Tojimbara T, Nakajima I, Fuchinoue S, Teraoka S, Toma H, Agishi T: Posttransplant antidonor antibodies and chronic rejection in renal transplantation. Transplant Proc 31: , Armbruster BL, Carlemalm E, Chiovetti R, Garavito RM, Hobot JA, Kellenberger E, Villiger W: Specimen preparation for electron microscopy using low temperature embedding resins. J Microsc 126: 77 85, Horvat R, Hovorka A, Dekan G, Poczewski H, Kerjaschki D: Endothelial cell membranes contain podocalyxin The major sialoprotein of visceral glomerular epithelial cells. J Cell Biol 102: , Abe M, Kawai T, Futatsuyama K, Tanabe K, Fuchinoue S, Teraoka S, Toma H, Ota K: Postoperative production of antidonor antibody and chronic rejection in renal transplantation. Transplantation 63: , Baid S, Saidman SL, Tolkoff-Rubin N, Williams WW, Delmonico FL, Cosimi AB, Pascual M: Managing the highly sensitized transplant recipient and B cell tolerance. Curr Opin Immunol 13: , Collins CM, Mendoza SA, Griswold WR, Tanney D, Lieberman E, Reznik VM: Pediatric renal transplantation in Laurence- Moon-Biedl syndrome. Pediatr Nephrol 8: , Nangaku M, Alpers CE, Pippin J, Shankland SJ, Adler S, Kurokawa K, Couser WG, Johnson RJ: A new model of renal microvascular endothelial injury. Kidney Int 52: 182, 194, Lajoie G: Antibody-mediated rejection of human renal allografts: An electron microscopic study of peritubular capillaries. Ultrastruct Pathol 21: , Monga G, Mazzucco G, Messina M, Motta M, Quaranta S, Novara R: Intertubular capillary changes in kidney allografts: A morphologic investigation on 61 renal specimens. Mod Pathol 5: , Ivanyi B, Fahmy H, Brown H, Szenohradszky P, Halloran PF, Solez K: Peritubular capillaries in chronic renal allograft rejection: A quantitative ultrastructural study. Hum Pathol 31: , Yilmaz A, Yilmaz S, Kallio E, Rapola J, Hayry P: Evolution of glomerular basement membrane changes in chronic rejection. Transplantation 60: , Gough J, Yilmaz A, Miskulin D, Gedeon I, Burama A, Yilmaz S, Supanj F, Muruve D, McKenna R, Benediktsson H: Peritubular capillary basement membrane reduplication in allografts and native kidney disease: A clinicopathologic study of 278 consecutive renal specimens. Transplantation 71: , Drachenberg CB, Steinberger E, Hoehn-Saric E, Heffes A, Klassen DK, Bartlett ST, Papadimitriou JC: Specificity of intertubular capillary changes: Comparative ultrastructural studies in renal allografts and native kidneys. Ultrastruct Pathol 21: , Watschinger B: Indirect recognition of allo MHC peptides Potential role in human transplantation. Nephrol Dial Transplant 14: 8 11, Vella JP, Spadafora-Ferreira M, Murphy B, Alexander SI, Harmon W, Carpenter CB, Sayegh MH: Indirect allorecognition of major histocompatibility complex allopeptides in human renal transplant recipients with chronic graft dysfunction. Transplantation 64: , Sayegh MH: Why do we reject a graft? Role of indirect allorecognition in graft rejection. Kidney Int 56: , Ensminger SM, Spriewald BM, Witzke O, Pajaro OE, Yacoub MH, Morris PJ, Rose ML, Wood KJ: Indirect allorecognition can play an important role in the development of transplant arteriosclerosis. Transplantation 73: , Szeto WY, Krasinskas AM, Kreisel D, Krupnick AS, Popma SH, Rosengard BR: Depletion of recipient CD4 but not CD8 T lymphocytes prevents the development of cardiac allograft vasculopathy. Transplantation 73: , Womer KL, Stone JR, Murphy B, Chandraker A, Sayegh MH: Indirect allorecognition of donor class I and II major histocompatibility complex peptides promotes the development of transplant vasculopathy. J Am Soc Nephrol 12: , Baker RJ, Hernandez-Fuentes MP, Brookes PA, Chaudhry AN, Cook HT, Lechler RI: Loss of direct and maintenance of indirect alloresponses in renal allograft recipients: implications for the pathogenesis of chronic allograft nephropathy. J Immunol 167: , Suri DL, Tomlanovich SJ, Olson JL, Meyer TW: Transplant glomerulopathy as a cause of late graft loss. Am J Kidney Dis 35: , Roodnat JI, Mulder PG, Rischen-Vos J, van Riemsdijk IC, van Gelder T, Zietse R, JN IJ, Weimar W: Proteinuria after renal transplantation affects not only graft survival but also patient survival. Transplantation 72: , Yildiz A, Erkoc R, Sever MS, Turkmen A, Ecder ST, Turk S, Kilicarslan I, Ark E: The prognostic importance of severity and type of post-transplant proteinuria. Clin Transplant 13: , Hohage H, Kleyer U, Bruckner D, August C, Zidek W, Spieker C: Influence of proteinuria on long-term transplant survival in kidney transplant recipients. Nephron 75: , Böhmig GA, Exner M, Watschinger B, Regele H: Acute humoral renal allograft rejection. Curr Opin Urol 12: 95 99, 2002 See related editorials, C4d and the Fate of Organ Allografts (pp ) and Capillary C4d Deposition as a Marker of Humoral Immunity in Renal Allograft Rejection (pp ).