Abstract
Abstract. There is increasing evidence for an important pathogenetic role of alloantibodies in acute renal allograft rejection. Acute humoral rejection (AHR) has been reported to be associated with a poor transplant survival. Although treatment modalities for cellular rejection are fairly well established, the optimal treatment for AHR remains undefined. Ten of 352 kidney allograft recipients transplanted at the authors' institution between November 1998 and September 2000 were diagnosed as having AHR, supported by severe graft dysfunction, C4d deposits in peritubular capillaries (PTC), and accumulation of granulocytes in PTC. AHR was diagnosed 18.9 ± 17.5 d posttransplantation. All patients were subjected to immunoadsorption (IA) with protein A (median number of treatment sessions, 9; range, 3 to 17). Seven recipients with additional signs of cellular rejection (according to the Banff classification) received also antithymocyte globulin. In nine of ten patients, AHR was associated with an increase in panel reactive antibody reactivity. A pathogenetic role of alloantibodies was further supported by a positive posttransplant cytotoxic crossmatch in all tested recipients (n = 4). In nine of ten recipients, renal function recovered after initiation of anti-humoral therapy. One patient lost his graft shortly after initiation of specific therapy. Another recipient with partial reversal of AHR returned to dialysis 8 mo after transplantation. Mean serum creatinine in functioning grafts was 2.2 ± 1.2 mg/dl after the last IA session (n = 9) and 1.5 ± 0.5 mg/dl after a follow-up of 14.2 ± 7.1 mo (n = 8). In conclusion, this study suggests that AHR, characterized by severe graft dysfunction, C4d staining, and peritubular granulocytes, can be effectively treated by timely IA. In the majority of patients, IA treatment can restore excellent graft function over a prolonged time period.
Over the last decades, graft survival after kidney transplantation has dramatically improved. This is partly due to improvements in immunosuppression that have led to a significant reduction of acute kidney allograft rejections and a higher success rate of antirejection treatments. Nevertheless, a considerable number of allografts is still lost because of refractory early rejection. There is increasing evidence that, besides cellular immunity, antibody-mediated immune mechanisms also may cause acute rejection associated with a high rate of graft loss (1,2,3,4).
Currently, the diagnosis of acute humoral rejection (AHR) is predominantly based on the detection of donor-specific anti-bodies by posttransplant crossmatching (1,2,3,4). The presence of donor-specific antibodies has been shown to be associated with typical histopathologic findings, such as accumulation of neutrophils in peritubular capillaries (PTC) (5,6,7). Recent studies have suggested that endothelial deposition of the stable complement-split product C4d may also be indicative of antibody-mediated rejection (8,9,10,11,12). Endothelial C4d deposits have been demonstrated to be clearly associated with antibody-positive AHR (12). Furthermore, C4d deposition has been shown to correlate with poor graft survival (9,10,11,13). In a recent retrospective analysis, we found that this association of endothelial C4d deposition with inferior graft outcome might be independent of the occurrence of acute cellular rejection (13).
Recently, plasmapheresis, combined with tacrolimus-mycophenolate mofetil rescue or with intravenous immune globulin, has been demonstrated to effectively reverse AHR (14,15). Some reports suggest efficacy of immunoadsorption (IA) with protein A in refractory antibody-positive kidney transplant rejection (16,17,18,19).
A few months ago, we reported our first case of C4d-positive AHR that was resistant to conventional therapy but responded well to IA (19). Prompted by this observation, we instituted routine staining for C4d at our unit >2 yr ago. Patients with positive peritubular C4d staining, severe early graft dysfunction, and accumulation of granulocytes in PTC, findings that are highly suggestive for AHR, were subjected to specific antihumoral therapy, i.e., IA therapy. We herein provide a summary of the first ten kidney transplant recipients who fulfilled these criteria that led to the treatment with IA. This initial trial suggests that, in the majority of patients with AHR, IA treatment can restore excellent graft function over a prolonged time period.
Materials and Methods
Patients
Between November 1998 and September 2000, 352 patients received a kidney allograft at our unit. Sixty recipients (17%) had delayed graft function; 304 transplant biopsies were performed in 178 (51%) of the 352 patients (≥2 biopsies in 82 patients) during follow-up until February 2001. Eighty recipients (22.7%) had one or more biopsy-proven acute rejection episodes (rejection defined by the Banff classification). Ten kidney allograft recipients (2.8%) fulfilled the following criteria: (1) C4d deposition in PTC, (2) granulocytes in PTC, and (3) severe allograft dysfunction. These patients were subjected to IA. Patient demographics and clinical data are listed in Tables 1 and 2. In these patients, biopsies were performed at a mean of 18.9 ± 17.5 (range, 5 to 65) d after transplantation. Immunohistochemical and histopathologic results are listed in Table 3. All patients showed positive staining for C4d in PTC and granulocyte accumulation in PTC. An example of strong C4d staining along PTC is depicted in Figure 1. Other histopathologic features that pointed to AHR were found at lesser frequencies (Table 3). The features of AHR occurred either alone (“pure” rejection, n = 3) or in combination with established signs of cellular rejection, according to the criteria of the Banff 97 Working Classification of Renal Allograft Pathology (20) (“mixed” rejection, n = 7).
Patient demographics and immunologic risk factorsa
Clinical course in patients with AHRa
Immunohistochemical and pathohistological evaluation at the time of rejectiona
Histomorphology in a case of mixed (cellular and humoral) acute kidney allograft rejection (patient 10). First biopsy performed 15 d after renal transplantation with (A) granulocytes in peritubular capillaries (PTC; indicated by arrows) (periodic acid—Schiff [PAS]), (B) signs of vascular (indicated by arrows) and interstitial rejection (PAS), and (C) endothelial C4d deposition in PTC (immunohistochemistry on a paraffin section). Follow-up biopsy performed 32 d after transplantation without signs of cellular rejection and disappearance of peritubular granulocytes (D) (PAS), but (E) persistence of C4d staining (immunohistochemistry on a paraffin section). Magnifications: ×650 in A; ×250 in B and D; ×400 in C and E.
Immunosuppression
Primary immunosuppression consisted of cyclosporin A (n = 10; dose adjusted to trough levels), mycophenolate mofetil (n = 8; initial dose, 2 g/d orally; in four patients, temporary reduction or transient withdrawal was necessary because of leukopenia or diarrhea) or FTY720 (patient 9, 2.5 mg/d and patient 10, 0.25 mg/d; both patients were included in a multicenter randomized open label exploratory dose finding study; Novartis, Basel, Switzerland), and steroids (n = 10; dose adjusted according to a local regimen). One patient with early immunologic loss of a first renal allograft (patient 6) received prophylactic therapy with horse antilymphocyte serum (Lymphoglobuline, IMTIX SangStat, France). Anticellular rejection therapy consisted of rabbit antithymocyte globulin (ATG; Thymoglobuline, IMTIX SangStat), in some patients combined with high dose steroids. To minimize the possibility of a reduction in ATG levels by IA due to antibody binding to protein A, ATG was infused shortly after termination of IA treatment sessions. During antibody therapy, cytomegalovirus prophylaxis with gancyclovir was given. Antihumoral therapy consisted of IA therapy. Two patients were converted to tacrolimus in addition. IA therapy was performed with the Citem 10 Immunoadsorption System (Fresenius HemoCare, Inc., Redmond, WA) and a staphylococcal protein A column (Fresenius). Plasma was separated with a traditional plasma separator, passed through the column for adsorption, and then reinfused. During each session, two to three plasma volumes were processed. Initially, recipients received daily treatment sessions (2 to 5 d) and two to three sessions per week afterward.
Renal Allograft Pathology
Histopathologic evaluation was done on formalin-fixed paraffin sections stained with hematoxilin and eosin, periodic acid-Schiff, Methenamin-Silver, and the trichrome stain Acidic-Fuchsin-Orange-G. For C4d staining, we used either frozen (patient 1) or, if frozen material was not available, paraffin sections (patients 2 to 10). For immunohistochemical detection of C4d on paraffin-embedded tissue sections, we used a polyclonal anti-C4d antibody (C4dpAb) generated by immunization of a rabbit with a 15-mer peptide that corresponds to amino acid 1242 to 1256 of C4, as described elsewhere (13). The identical peritubular staining pattern of C4d, when our polyclonal and a commercially available monoclonal anti-C4d antibody (Quidel, Alkmaar, Netherlands) were used, was earlier proved in a series of kidney biopsies (n = 37; 12 allografts and 25 normal native kidneys), of which both paraffin-embedded and frozen material was available. Immunostaining of paraffin-embedded sections with C4dpAb revealed the same peritubular staining patterns as those observed on corresponding frozen sections stained with anti-C4d mAb.
For immunohistochemistry, sections (2 μm) 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 elsewhere (21). Endogenous biotin was blocked by use of a Biotin blocking kit (Vector Laboratories, Burlingame, CA). After 30-min incubation with C4dpAb (5 μg/ml), bound IgG was visualized by use of the Supersensitive Kit (BioGenex, San Ramon, CA), according to the manufacturer's protocol. For detection of C4d on frozen sections, we used anti-human C4d mAb (Quidel). Untreated 4-μm sections were incubated with anti-C4d mAb at 2 μg/ml for 30 min, and bound IgG was visualized as described above. Intensity of endothelial C4d staining was staged as weak (+), moderate (++), or strong (+++). Some of the biopsies with weak C4d staining showed an inhomogeneous distribution of C4d (only few C4d positive peritubular capillaries). In biopsies with moderate or strong C4d staining, all capillaries stained positive.
Immunologic Methods
Cytotoxic crossmatch testing was performed according to the protocol of the Eurotransplant Organization and used the standard microcytotoxicity technique described by Terasaki and McClelland (22). For inactivation of IgM, sera were pretreated with dithiotreitol. Cytotoxic panel reactive antibody (PRA) reactivity was assessed by use of a panel of cells from 50 phenotyped donors. For flow cytometry crossmatch (FCXM) testing, 2 × 105 donor cells (cryopreserved splenocytes or, in the case of living-donor transplantation, freshly isolated peripheral blood mononuclear cell) were incubated with undiluted serum for 1 hr at room temperature. The cells were washed and then incubated with pretitered FITC—conjugated goat anti-human IgG (Dako, Carpinteria, CA) and with phycoerythrin-conjugated CD3 mAb (T cell FCXM) or CD19 mAb (B cell FCXM; Immunotech S. A., Marseille, France) at 4°C for 30 min. Fluorescence intensity was measured with a FACscan flow cytometer (Becton Dickinson, Sunnyvale, CA). When the mean fluorescence intensity (MFI) was >2 SD of normal controls (sera from at least four different nonsensitized healthy volunteers of blood type AB), the FCXM was considered positive. Positivity is graded + (<2-fold mean control MFI) or ++ (>2-fold mean control MFI).
Results
Effect of Specific Therapy on Clinical Outcome
All 10 patients were subjected to IA (median number of IA treatment sessions, 9; range, 3 to 17; mean duration of therapy, 18.1 ± 8.6 d). Laboratory evaluation revealed a substantial decrease of total serum IgG levels by IA (mean level before initiation of IA, 639 ± 194 mg/dl; after the last IA session, 23 ± 75.3 mg/dl). Patients with pure AHR received IA treatment without concomitant ATG (one recipient was converted to tacrolimus in addition). In one case, graft dysfunction had been resistant to prior ATG (14-d course) and tacrolimus rescue (1 wk posttransplantation). Recipients with mixed rejection received IA plus ATG (7- to 13-d course of ATG, plus conversion to tacrolimus in one and steroid bolus therapy in four patients). In three cases of mixed rejection (patients 3, 9, and 10), ATG was initiated ≥4 (4 to 10) d before IA. In those patients, ATG alone failed to improve graft function. In four cases, IA and ATG were initiated simultaneously.
Graft function dramatically improved in eight patients after implementation of specific therapy (Table 3). In one case of pure humoral rejection (patient 6), however, reversal of severe graft dysfunction was incomplete. Because of progressive deterioration of graft function, this patient returned to dialysis 8 mo later. In patient 5 (mixed rejection), immunologic graft loss lead to transplant nephrectomy 7 d after initiation of antirejection therapy. After the last IA session, serum creatinine in functioning grafts (n = 9) was 2.2 ± 1.2 mg/dl. After a follow-up of 14.2 ± 7.1 mo, serum creatinine was 1.5 ± 0.5 mg/dl (functioning grafts, n = 8).
Effect of Specific Therapy on Immunologic Results and Histopathology
In five patients, a follow-up biopsy was performed 11.4 ± 4.3 (range, 7 to 17) d after the first biopsy (IA only, n = 1; IA plus ATG, n = 4). All follow-up biopsies showed prominent endothelial C4d staining, although with a reduced staining intensity in two patients. In recipients who responded to specific therapy, peritubular accumulation of granulocytes, glomerulitis, or intimal arteritis disappeared or were less prominent in follow-up biopsies. Furthermore, in those patients, control biopsies showed no signs of rejection according to the Banff classification. A representative case is depicted in Figure 1. In the patient with immunologic graft loss, however, subsequent evaluation revealed clear histopathologic features of both humoral and cellular rejection.
Serologic results are given in Table 4. At the time of biopsy a clear increase of PRA reactivity was observed in nine of ten patients. In one case, however, no complement-fixing anti-HLA antibodies were detected. In eight patients with increased posttransplant PRA levels, specific treatment significantly decreased PRA reactivity (Table 4). In the patient with refractory rejection, PRA reactivity was not reduced.
Serologic findings in patients with C4d-positive AHRa
In four patients, donor lymphocytes were available for post-transplant crossmatch testing (Table 4). All four patients had a positive cytotoxic crossmatch at the time of biopsy. After the last IA session, three of the four tested recipients (recovery of renal function in all four patients) had a negative cytotoxic crossmatch.
Further evaluation by flow cytometry crossmatching before IA revealed a positive T cell FCXM in all four and a positive B cell FCXM in two tested patients. After IA, the T cell FCXM became either negative or MFI levels were clearly reduced. In the patient in whom complement-fixing anti-donor antibodies were still detectable after IA, T cell FCXM testing revealed a clear reduction in anti-HLA class I antibody levels. In one patient, a positive B cell FCXM became negative with IA treatment; in the other recipient, it remained positive, whereas the post-IA cytotoxic cross-match turned negative and graft function was completely restored (Table 4).
A retrospective evaluation of pretransplant serum samples revealed a positive pretransplant T cell FCXM in three of four cases. Notably, the history of one of these patients revealed no apparent risk factor for presensitization. In contrast, the pretransplant B cell FCXM was negative in all tested patients.
Infectious Complications
Seven patients received antibiotic treatment for urinary tract infection. One patient developed bacterial pneumonia after 2 mo, and one recipient developed osteomyelitis 17 mo after transplantation. With the exception of patient 5, who developed severe bacterial pneumonia after transplant nephrectomy, no serious infectious complications occurred. Laboratory evaluation pointed to subclinical cytomegalovirus infection in four patients (de novo appearance of cytomegalovirus-specific IgM antibodies and/or pp65 antigenaemia), which was reversible by therapy with gancyclovir.
Discussion
In this study, we have reported on ten renal allograft recipients with severe acute humoral rejection. In nine of these patients, therapeutic IA successfully reversed rejection. Regarding the reported poor graft outcome after AHR not subjected to specific treatment (5,6,7,23,24), our observation of an 80% graft survival after a mean follow-up of 14 mo indicates a high efficacy of this treatment modality.
AHR remains a diagnostic challenge. The diagnosis of AHR is primarily based on posttransplant crossmatching with donor lymphocytes (1,2,3,4). In cadaveric transplantation, donor cells are often not available because storage of donor material is usually not performed. Therefore, reliable diagnostic criteria for AHR other than posttransplant crossmatching are highly desirable.
In our study, we used the concurrent presence of three criteria, i.e., severe allograft dysfunction (nine of ten patients were dialysis-dependent at the time of rejection) and two distinct biopsy-based findings (accumulation of granulocytes in PTC and C4d deposition in PTC) as indicator of AHR. The presence of these criteria was recently proposed to be highly suggestive for AHR (3,12).
Endothelial deposition of the stable complement split product C4d was earlier suggested to be a specific marker for classical complement activation reflecting antibody-mediated graft injury (8,9,10,11,12). Endothelial C4d deposition has also been reported to be associated with poor kidney graft outcome (9,10,11,12,13). More recently, Collins et al. (12) demonstrated a strong association of endothelial C4d deposition in PTC with AHR defined by the occurrence of donor-specific anti-HLA antibodies and characteristic histopathologic changes, i.e., accumulation of granulocytes in PTC in all biopsies. Earlier, this accumulation of granulocytes in PTC as well as other histopathologic features, such as fibrin thrombi and arterial wall necrosis, were shown to correlate well with antibody-positive acute renal allograft rejection (5,6,7). On the basis of these elegant studies, we used C4d staining in combination with granulocyte accumulation in PTC for the diagnosis of AHR.
There is increasing evidence that antibody-mediated kidney allograft rejection can occur in the absence of cellular rejection defined by the Banff classification (2,5,6,12,13). Humoral graft injury without features of cellular rejection was considered to represent pure antibody-mediated rejection (2,3). These data suggest that, with the exception of type III rejection, which can be a reflection of AHR, pathohistologic evaluation according to the Banff classification might not be useful in discriminating acute humoral graft injury. In our study patients, we used the classification of biopsies according to the Banff classification to discriminate between “pure” and “mixed” rejection.
The applicability of the chosen criteria of AHR is strongly supported by serologic results obtained in our study population. In nine of ten recipients, AHR was associated with a substantial increase of PRA reactivity. In one patient, however, no complement-binding HLA antibodies were detectable at the time of biopsy. Because no donor lymphocytes were available for posttransplant crossmatch testing in this case, the presence of non-complement-binding anti-HLA alloantibodies cannot be excluded. Alternatively, one could speculate that positive C4d staining in this patient reflects a pathogenetic role of donor-specific non-HLA antibodies. It has previously been reported that alloantibodies directed against polymorphic non-HLA antigens can cause graft rejection (25,26,27). The reliability of the inclusion criteria is further corroborated by the finding of a positive cytotoxic crossmatch and a positive T cell FCXM in those four patients for whom donor lymphocytes were available. A positive B cell FCXM in two of these patients might point to a contribution of anti-HLA class II antibodies to AHR. Humoral immunity to HLA class II antigens was earlier suggested to contribute to a subset of antibody-mediated allograft rejections (1). However, our observation that, in one patient, IA did not affect the B cell FCXM but effectively reversed rejection might argue against a crucial pathogenetic role of donor-specific anti-HLA class II antibodies in this particular case.
All ten patients with AHR were subjected to specific anti-humoral therapy with IA. Therapy with ATG, with or without high-dose steroids, was instituted in all seven patients who showed signs of cell-mediated rejection in addition to features of humoral rejection. Two patients were converted to tacrolimus. Recently, conversion to tacrolimus as an adjunct to plasmapheresis has been suggested to be effective in the treatment of AHR (14).
In previous reports, AHR not subjected to specific treatment was associated with a high graft failure rate and poor long-term graft survival (5,6,7,23,24). In our analysis, nine of ten patients showed an initial response to specific antirejection therapy that was sustained over an extended period, resulting in an 80% graft survival with excellent graft function after 14 mo of follow-up (mean serum creatinine, 1.5 ± 0.5 mg/dl). In six patients, there was a clear relationship between initiation of IA and improvement of graft function. IA monotherapy restored graft function without prior or concomitant antibody therapy in two recipients with pure rejection. In four patients (pure rejection, n = 1; mixed rejection, n = 3) who were resistant to prior antibody therapy, subsequent IA effectively restored graft function. In addition, three patients treated with IA and ATG simultaneously responded well to therapy. In these cases, an additional effect of ATG for the restoration of renal function cannot be excluded. Support for the efficacy of IA also comes from histopathologic evaluation of follow-up biopsies. In patients who responded to IA, a complete disappearance of granulocytes in PTC was observed. Other histologic signs of AHR were reduced or disappeared upon therapy as well. These results are in accordance with the observation of reduced PRA reactivity in successfully treated patients. In contrast, in a patient with refractory AHR, specific therapy failed to lower increased PRA reactivity and to eliminate histologic features of humoral rejection. Other than histopathologic signs of AHR, deposition of C4d along peritubular or glomerular endothelium was not or was only slightly altered by IA. This might be explained by covalent binding of C4d to target structures being stable over a longer period of time. Similar to other observational reports that have tested antihumoral therapies, our study is limited by a generally low incidence of AHR (2.8% in our series; 5% to 10% in other reports [5,6,7,24]). For definitive elaboration of the efficacy of therapeutic IA in an adequate number of cases, we thus suggest the performance of a prospective controlled multicenter trial.
Selected cases of antibody-positive kidney allograft rejections refractory to prior antilymphocyte antibody have earlier been shown to be reversible by IA (16,17,18). The higher graft failure rate in these studies, compared with our analysis, might in part be explained by a comparatively late implementation of IA, after antilymphocyte antibody therapy had failed. Our approach of prospective C4d staining in all patients allowed an early diagnosis of AHR and thus enabled a timely start of antihumoral therapy. It can be speculated that early removal of pathogenetic alloantibodies might prevent severe, probably irreversible, graft injury.
In three out of four tested recipients, retrospective pretransplant FCXM testing revealed the presence of preformed anti-donor HLA class I antibodies that escaped detection by the standard cytotoxic crossmatch test. With FCXM testing, a complement-independent technique, donor antigen-specific antibodies can be detected with substantially higher sensitivity than the cytotoxic crossmatch, which requires complement-fixation and lysis of target lymphocytes (28). In many studies, a positive pretransplant FCXM in the face of a negative cytotoxic crossmatch has been shown to be associated with poor kidney transplant outcomes, i.e., decreased graft survival and higher rejection rates (29,30,31,32,33). The retrospective crossmatch results obtained in our small series indicate that a considerable proportion of patients with AHR might possess preformed non-complement-fixing donor-specific antibodies not detectable by conventional crossmatching. It can be speculated that in the T cell FCXM-positive patients, preemptive IA started at the time of transplantation could have influenced the development of AHR, as was recently suggested by Montgomery et al. (15), who tested an alternative preemptive antihumoral therapy (plasmapheresis plus intravenous Ig) in FCXM-positive renal allograft recipients. In summary, the present initial trial suggests high efficacy of IA in the therapy of C4d-positive AHR.
Acknowledgments
We thank Claus Wenhardt and Bettina Haidbauer for excellent technical assistance.
- © 2001 American Society of Nephrology
References
Jump to section
More in this TOC Section
Cited By...
- Antibody-Mediated Rejection of the Kidney after Simultaneous Pancreas-Kidney Transplantation
- Antibody-Mediated Rejection in Renal Allografts: Lessons from Pathology
- Sensitization after Kidney Transplantation
- 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
- Capillary C4d Deposition as a Marker of Humoral Immunity in Renal Allograft Rejection
- Capillary C4d Deposition in Kidney Allografts: A Specific Marker of Alloantibody-Dependent Graft Injury