The ability of alloantibody to mediate renal graft injury is well established (1). In hyperacute rejection the injury is immediate and massive, and the classical pathological features are infiltrating polymorphs, fibrinoid necrosis, and platelet thrombi(1-3). Fortunately, current cross-matching technology has largely eliminated this entity. However, alloantibody can also be associated with syndromes distinct from hyperacute rejection. We have previously described the role of anti-class I antibody in one form of acute accelerated rejection (4, 5). The presence of anti-class I antibody at the time of rejection is associated with a worse prognosis than rejection in the absence of antibody (5, 6). In acute rejection associated with anti-class I antibodies, recovery of renal function can be achieved with aggressive early treatment, particularly anti-CD3. A favorable outcome is usually associated with loss of the antibody response (5).
In antibody-mediated acute rejection, antibody may be undetectable at the time of transplantation. After transplantation, re-exposure to the antigen to which the recipient has been previously sensitized triggers the production of large quantities of antibody. Clinically, the recipient may have a normal initial postoperative course until there is sudden deterioration in graft function, which probably reflects the release of antibody that binds to the graft endothelium, activates complement and other effector mechanisms, and initiates cell injury. The incidence of delayed graft function is increased in sensitized patients, suggesting that one type of antibody-mediated injury may present as poor initial graft function or acute tubular necrosis. It is also probable that vascular injury has a worse prognosis than tubulointerstitial lesions (7, 8) and that antibody plays a role in such injury.
Since the mechanisms initiating the graft injury in antibody-associated rejection are different from those in cell-mediated injury, the pathology is also likely to be different. However, no antibody-positive rejection episodes have been analyzed using the Banff criteria. Therefore, we sought to determine whether the presence of antibody influences the pathological findings in renal biopsy specimens showing features of acute rejection. We identified a population of patients who had biopsy-proven acute renal rejection according to Banff criteria (9) and who also had antibody against class I antigens of their donor. We compared their histopathological findings to those of patients with biopsy-proven rejection who tested negative for anti-class I antibodies. We found that in biopsy-proven acute rejection, the presence of anti-class I antibody is associated with more severe vasculitis, whereas the absence of antibody is associated with more severe tubulitis.
PATIENTS AND METHODS
Patients. We prospectively identified patients whose sera could be tested against donor cells, and who also had instructive clinical courses and satisfactory biopsy material. We reviewed the records of all patients who had received renal transplants (from both cadavers and living donors) at the University of Alberta Hospital between January 1987 and December 1994. During this period, 360 cadaver and 99 living-related kidneys were transplanted. The patients who were tested for the presence of antidonor reactivity (anti-class I antibody) during an episode of biopsy-proven acute rejection were selected for study. Patients who (1) did not have a biopsy during the episodes of renal functional impairment, (2) had biopsy findings incompatible with acute rejection, or (3) had an insufficient number of serum samples for antibody testing were excluded from the study. Information on long-term graft function was obtained from the transplant patient database. A total of 44 patients met these criteria: 24 patients tested antibody positive (Ab+R*) and 20 patients tested antibody negative (Ab-R). Thirteen of the Ab+R patients had been included in our previous studies, at a time when the Banff classification(9) was unavailable. Clinical records were reviewed for patient demographics, previous sensitization by transfusion, pregnancy and transplantation, number of days to rejection, rejection therapy, and peak panel reactive antibody (PRA) before, at the time of transplantation and, if available, after the onset of rejection.
Detection of donor-specific antibody. Donor-specific antibody detection was performed as described previously (4, 5). In brief, we performed pretransplant cross-matches using the NIH complement-dependent cytotoxicity assay with extended incubation time; cross-matching was carried out on two to three of the most recent serum samples against donor peripheral blood lymphocytes (PBL). In highly presensitized patients (PRA >50%) or in retransplants, the final cross-match was done using a fresh serum sample tested against the donor's T and B cells (isolated either from peripheral blood or from spleen), and for autoreactivity against recipient PBLs. In retransplant cases, historical(remote) serum samples were tested as well for the presence of antidonor reactivity. A positive T cross-match, either final or historical, was regarded as a contraindication for transplantation. A positive B cell cross-match was regarded as admissible when autoreactivity was also present. In living-related donors (LRD) a positive B cell cross-match in the absence of autoantibodies required pooled platelet absorbtion to distinguish between anti-class I and anti-class II antibodies.
All patients experiencing acute deterioration in graft function, clinically determined as a rise in serum creatinine and a decrease urine output, had serum samples tested for the presence of antidonor reactivity, using frozen spleen cells of PBL.
Immunosuppression. Before surgery, all recipients received 1.5 mg/kg methylprednisolone and 3 mg/kg azathioprine. Patients with high immunological risk (high PRA, previous graft loss due to aggressive rejection) and those with poor initial graft function were treated postoperatively with 10 mg/kg Minnesota antilymphocyte globulin (MALG). Since June 1992 we have used OKT3 induction at a dose of 2.5 mg/day in high-risk patients and those with poor initial function. The remaining recipients received 3 mg/kg i.v. cyclosporine (CsA); oral CsA was introduced on day 1 after transplant and intravenous CsA was discontinued when adequate blood levels were achieved with oral therapy. Maintenance immunosuppression included CsA, azathioprine, and prednisone.
Patients with mild rejection received methylprednisolone (250 mg) daily for 3 consecutive days. More severe cases were treated with either OKT3 (5 mg/day) for 10 to 14 days or MALG (20 mg/kg) for 14 days. We discontinued the maintenance immunosuppression with CsA during the courses of antilymphocytic therapy; it was reintroduced 2-3 days before the end of the treatment to provide adequate blood levels.
Pathology. Renal transplant biopsies were performed on all patients included in the study, resulting in 144 biopsies from 44 patients (69 from 24 Ab+R patients and 75 from 20 Ab-R patients). The transplant was localized under ultrasound guidance, and the biopsy was performed using an 18-gauge needle with a Biopty gun. Specimens for light microscopy were fixed in 10% formalin and embedded in paraffin. Sections (3 μm thick) were stained with hematoxylin and eosin, periodic acid-Schiff, periodic acid-methenamine-silver, Masson's trichrome, Martius scarlet blue, and Congo red. Usually we examined nine slides per biopsy. Immunofluorescence was performed on 33/44 patients (21/24 Ab+R patients and 12/20 Ab-R patients). The usual panel of antibodies included anti-IgG, IgM, IgA, κ, λ, C1q, C3, C4, albumin, and fibrinogen. Electron microscopy was carried out on 18/24 specimens from Ab+R patients and 14/20 specimens from Ab-R patients.
Kidney biopsy specimens were originally read by one of the pathologists(K.S.) and subsequently reviewed by both pathologists (K.T., K.S.). We applied the Banff classification of kidney transplant pathology for scoring the presence and degree of rejection on kidney biopsy specimens(9). Banff scores were done on the biopsy closest to the time of confirmed serological status (antibody positive or negative). Mild acute rejection indicated the presence of foci of moderate tubulitis (5-10 mononuclear cells per tubular cross section/10 tubular epithelial cells) in cases with significant interstitial infiltration (>25% of parenchyma affected). Moderate acute rejection was defined as the presence of significant interstitial infiltration with foci of severe tubulitis (>10 mononuclear cells per tubular cross section/10 tubular epithelial cells) and/or mild or moderate intimal arteritis. Severe acute rejection included the cases with severe intimal arteritis in many arterial cross-sections and/or“transmural” arteritis with fibrinoid change and necrosis in medial smooth muscle cells. Recent focal infarctions (and interstitial hemorrhage without other obvious cause) were also regarded as evidence for severe acute rejection. Glomerulitis was defined as presence of mononuclear cells in peripheral loops of glomerular capillaries with focal or diffuse endothelial swelling. We also assessed the following morphologic changes: presence of polymorphonuclear leukocytes (PMNs) in glomeruli, tubules, peritubular capillaries (PTC) and interstitium, dilatation of PTC, fibrin thrombi, fibrinoid necrosis in vascular walls, and infarctions.
Statistical analysis was performed using Fisher's exact test; the vasculitis score and tubulitis score were compared using the Mann-WhitneyU test. True Epistat software was used for both statistical analyses.
RESULTS
The patients in the antibody-negative rejection (Ab-R) and antibody-positive rejection (Ab+R) groups were matched with respect to age, average peak PRA, number of B,DR mismatches, and the incidence of acute tubular necrosis (Table 1). Men comprised 67% of the Ab+R group, compared with 45% of the Ab-R group (NS). The Ab-R group had 4/20 LRDs compared with the Ab+R group, in which all the patients had received cadaveric renal transplants (P=0.036). Nevertheless, the degree of B,DR mismatching was not different between the groups.
The Ab+R patients had a biopsy-proven rejection episode earlier than the Ab-R patients (11 days vs. 18 days, NS; Table 2). Treatment with antilymphocytic therapy (either OKT3 or MALG) was similar in both groups: 80% of Ab-R patients and 92% of Ab+R patients. The incidence of early graft loss was striking in Ab+R patients: 50% of Ab+R patients versus only 15% of Ab-R patients lost their grafts within 3 months. One Ab+R patient died with a functioning graft, but re-analysis of the data excluding this patient did not change the significance of the results.
Nevertheless, the presence of antidonor activity during rejection did not preclude long-term success. Many of the grafts that survived beyond 3 months had long-term survival without chronic rejection. One Ab+R patient did not recover renal function and eventually had a nephrectomy at day 107, and was counted as a loss before 3 months. The 11 grafts that regained renal function survived during the study period, except for one graft from an Ab+R patient that was lost 68 months after transplantation following episodes of late acute rejection, possibly due to noncompliance. No cases of chronic rejection have yet been diagnosed in the Ab+R group. In the Ab-R group, two grafts were lost in the follow-up period because of chronic rejection.
The evidence of presensitization is presented in Table 3. The average peak PRA was similar in both groups. In the Ab+R group, 50% had previously lost kidney transplants, compared with 10% in the Ab-R group. The pregnancy and transfusion history did not differ significantly between the two groups, nor were previous pregnancies or transfusions associated with the development of antidonor activity after transplantation. Although the peak PRA did not differ significantly between the Ab+R and Ab-R groups, the presensitized Ab+R patients (PRA >10%) did worse than the Ab+R patients without PRA (Table 4). Indeed, the patients with PRA>10% who developed Ab+R had almost 70% graft loss by 3 months, indicating that this group carries a high risk of early graft loss. In Ab-R patients, PRA>10% did not predict the risk of graft loss. Thus, in the presence of a negative pretransplant cross-match, history of PRA >10% affected only those patients who developed Ab+R, compatible with the view that antibody-positive rejection was more serious if it was a secondary antibody response.
Pathology. Table 5 summarizes the histologic findings in Ab+R and Ab-R biopsy specimens. The frequency of glomerulitis in the Ab+R patients was significantly higher (P=0.01)(Table 5; Fig. 1A). It was usually accompanied by severe vasculitis (6/11, 55%) and five of these patients lost their grafts. Another patient who had glomerulitis lesions (G2) also lost the graft, and showed evidence of moderate rejection. Five of the six patients who lost their grafts presented with moderate glomerulitis (G2).
Ab-R biopsies consistently demonstrated evidence of tubulitis (19/20 of the Ab-R patients); 95% of those (18/19) showed either moderate or severe tubulitis (Table 5 and Fig. 2A). Only 50% of the Ab+R patients (12/24) showed moderate or severe tubulitis, and absence of tubulitis lesion was noted in five (21%) of the Ab+R patients. Thus, tubulitis correlated negatively with antibody in this population, which suggests that tubulitis reflects T-cell-mediated injury.
The predominant feature of the Ab+R biopsies was vascular injury(Fig. 3A; Tables 5 and 6). Vasculitis in Ab-R patients was either absent or mild, often affecting only one arterial cross section. Moderate or severe arteritis was present in 14/24 (58%) Ab+R patients, compared with only 2/20 (10%) Ab-R patients. Some patients without arteritis were seen in both Ab+R and Ab-R groups, suggesting that antibody can sometimes be present without prominent pathologic changes. In patients with no signs of arteritis in the Ab+R group (6/24), no graft was lost.
The mean vasculitis score was higher in the Ab+R biopsy specimens, while the mean tubulitis score was higher in the Ab-R biopsy specimens(Table 6). Thus, the Ab+R and Ab-R specimens were not merely different in severity but were qualitatively different, perhaps reflecting different mechanisms of immunologic injury.
Early endothelial arterial proliferation was noted in Ab+R patients with severe rejection, and resulted in focal, near complete vascular occlusion. Glomerular (Fig. 1B) and vascular thrombi, occasionally accompanied by fibrinoid necrosis of the vascular walls (Fig. 3A), resembling hemolytic uremic syndrome, also indicated extensive endothelial injury in the microcirculation (Table 7). Infarction (Fig. 3B) was noted during the acute rejection episodes in 9/24 (38%) of the Ab+R group, versus none of the Ab-R group. Infarction portended dismal prognosis for the allograft and was present in 6 of 12 patients who lost their grafts early.
When PMN were assessed in various parenchymal compartments, only the presence of PMN in PTC (>2 PTC affected) was significantly different between the groups (Fig. 2B). Marked dilatation of PTC was more common in the Ab+R group (33% vs. 10%).
Immunofluorescence and electron microscopy findings were similar in both groups (not shown). Diffuse staining for IgG or IgM was never seen in the Ab+R or Ab-R groups, consistent with our previous findings(4, 5). Mild vascular and mesangial staining for IgM was present in nine Ab+R patients and two Ab-R patients, and mainly glomerular staining for IgG was present in three Ab+R patients and one Ab-R patient. The incidence of IgM was higher in the Ab+R patients (43% vs. 17%, NS). Staining for the C3 complement component and fibrinogen was not different between the groups (71% vs. 66% and 81% vs. 75%), reflecting the relative nonspecificity of these findings in biopsy specimens. Electron microscopy showed widening of the endothelial space and swelling of the endothelial cells, which were slightly more frequent in the Ab+R group (61% vs. 57% and 61% vs. 43%). Presence of platelets and fibrin in the capillary lumina was noted in 28%(5/18) of the Ab+R patients and in 14% (2/14) of the Ab-R patients.
DISCUSSION
Acute rejection in the early posttransplant period may be caused by T-cell-mediated mechanisms, by alloantibody, or both. In most rejection episodes in renal transplants, the biopsy shows the characteristic features of acute cellular rejection, such as tubulitis and intimal arteritis, as defined by the Banff schema (9). However, the Banff schema did not address the issue of whether some lesions were mediated by alloantibody. The present study was confined to biopsy-proven rejection episodes fulfilling the Banff criteria, comparing the Ab+R and Ab-R groups. The presence of anticlass I antibody associated with acute rejection predicted a poorer outcome, as in previous reports (5, 10), and a different pathological picture. Thus a positive donor-specific cross-match after transplantation in the presence of biopsy-proven rejection indicates increased risk of early graft loss. The prognosis if the lesion is reversed is good, however, arguing that an aggressive approach may be justified. The immunoglobulin class of the anti-class I antibodies detected in the present study was not assessed, but in other studies specific antibodies associated with increased incidence of rejection and graft loss have been shown to be of the IgG class (10-12).
The main pretransplant differences between the Ab+R and Ab-R groups were in fewer previous transplants and more living donors in the Ab-R group. The tendency toward antibody-mediated rejection may explain some of the increased risk of early graft loss in retransplants. Conversely, lower probability of antibody response during a rejection episode may be a benefit of LRD transplantation. Transfusion is a complex variable(13, 14) and did not differ between the antibody-positive and antibody-negative rejection episodes. Peak PRA as a measure of sensitization did not differ between the Ab+R and Ab-R groups overall; however, within the Ab+R group, previous PRA was associated with a poor outlook, probably reflecting the difficulty of suppressing secondary antibody responses. The morphologic picture associated with the presence of anti-class I antibody during a rejection episode is of prominent vascular damage, and probably explains why these patients are at increased risk of early graft loss.
Both the Ab+R and the Ab-R groups had clinically severe rejections: mild rejections in our center are not biopsied and were excluded from this study, and over 80% of the patients in both groups received antilymphocyte antibodies for rejection, reflecting clinical assessment as severe or resistant. Although both Ab+R and Ab-R episodes were severe, the pathology was different: Ab+R patients showed arteritis. glomerulitis, and PMNs in the capillaries, whereas Ab-R patients had more severe tubulitis, reflecting different effector mechanisms. The Ab-R group exhibited T-cell-mediated injury, which in its“pure” form causes tubulitis and mild intimal arteritis. The antibody-positive rejection episodes reflect the effects of the antibody: more severe vasculitis and glomerulitis, and accumulation of PMNs in PTCs. Some rejection episodes, particularly in the Ab+R group, probably reflect the operation of both mechanisms, but the proof of this awaits a reliable test for T cell effector activity.
The present findings show that presensitization, particularly with previous transplants, increases the risk of graft loss in patients with antibody-mediated rejection. The second set of acute allograft rejection episodes in animal models is accelerated and augmented, at least in part through the vigor of the alloantibody response (15). Antibody against membrane antigens may exhibit a threshold, where the amount of injury is critically dependent on a burst of antibody reaching the target cell at a sufficient concentration within a critical time, before the cell has a chance to accommodate or modulate its antigen expression. If the critical concentration is reached, the effector mechanisms are activated and injury occurs. Postulated effector mechanisms include: (1) complement activation by the classical pathway, followed by recruitment of PMNs and mononuclear cells by virtue of their Fc and C3 receptors; (2) antibody-dependent cell-mediated cytotoxicity (ADCC), via recruitment of large granular lymphocytes. Immunofluorescence findings did not show staining for IgG, C3, or other complement components (C1q, C4) in Ab+R patients, consistent with our previous studies (4, 5). Perhaps the high turnover of activated endothelium and limited sensitivity of immunohistology for small amounts of immunoglobulin preclude demonstration of diffuse staining. Nevertheless, the complement-PMN pathway is probably the major mechanism of antibody injury.
Glomerulitis (or acute allograft glomerulopathy) has a dismal prognosis, with increased vascular injury and early graft loss(16-23). Some investigators have used the term “glomerular rejection”(17, 18, 20, 21). However, the strong association of glomerulitis with vascular rejection, as in our Ab+R group, probably accounts for the increased severity and graft loss when glomerulitis is present (24). Early descriptions of combined glomerular and vascular lesions of this type establish a connection with circulating anti-class I antibodies detected in the posttransplant period(12, 25). Positive B lymphocyte cross-match(anti-class II) can also be associated with glomerulitis and graft injury(20, 21). Endothelial damage and/or necrosis may be prominent in glomerulitis (26), and glomeruli with glomerulitis lesion stain more intensely for HLA class I antigens than tubules, in contrast to nonglomerulitis rejection cases(16) or glomeruli in nonrejecting cases(27). This up-regulation of HLA class I antigens, combined with the phenotype of the mononuclear cells within glomeruli (CD3, CD8, CD57), suggests that ADCC may be operating in glomerulitis. Charpentier et al. (28) found ADCC and anti-HLA complement-dependent antibodies in the eluates from the rejecting kidneys.
The present pathologic findings indicate that the endothelium of the arteries, glomeruli, and the microcirculation is the main target of anti-class-I-mediated rejection, confirming our previous findings(5). Severe vascular lesions, resulting in frequent infarctions, fibrin thrombi, and fibrinoid necrosis in vessel walls and accompanied by frequent glomerular lesions, represent the morphologic evidence of extensive endothelial injury. Accumulation of PMN in PTC observed on postvascularization biopsy specimens is associated with increased incidence of acute rejection episodes, and subsequent graft loss (29). Our findings suggest that the endothelium of PTCs, glomeruli, and large and small arteries may all be targets of antibody-mediated rejection.
Although this study was not designed or powered to address the role of antibody in chronic rejection, our results do not support the concept that the antibody response strongly increases the risk of chronic graft loss, despite the increased acute graft loss. A role for alloantibody in the pathogenesis of obliterative arterial lesions in chronic rejection has been suggested(25) and Barr et al. (30) reported an association between development of anti-HLA antibodies (class I and class II) and chronic allograft rejection in renal and heart allografts: the 5-year renal allograft survival rate was 70% in recipients without antibodies and 53% in recipients who developed anti-HLA alloantibodies during the first year after transplantation. However, alloantibody is strongly associated with acute rejection in the early posttransplant period, and acute rejection with or without antibody is known to increase the risk of late graft loss. Whether antibody-positive rejection episodes have a greater propensity to chronic rejection than other rejection episodes, and the role of anti-class I antibodies in mediating chronic rejection, should be regarded as open questions for future studies.
Footnotes
Parts of this study were presented at the 14th Annual American Society of Transplant Physicians Meeting (Chicago, IL, May 14-17, 1995), the 13th International Congress of Nephrology (Madrid, Spain, July 2-6, 1995), and the 3rd Banff Conference on Allograft Pathology (Banff, Alberta, Canada, July 20-24, 1995).
REFERENCES
1. Kissmeyer-Nielsen F, Olsen S, Peterson VP, Fjeldborg O. Hyperacute rejection of kidney allografts associated with pre-existing humoral antibodies against donor cells. Lancet 1966; 2: 662.
2. Andres GA, Accinni L, Hsu KC, et al. Human renal transplants: III. Immunopathologic studies. Lab Invest 1970; 22: 588.
3. Porter KA. Morphological aspects of renal homograft rejection. Br Med Bull 1965; 21: 171.
4. Halloran PF, Wadgymar A, Ritchie S, Falk J, Solez K, Srinivasa NS. The significance of the anti-class I antibody response: I. Clinical and pathologic features of anti-class I mediated rejection. Transplantation 1990; 49: 85.
5. Halloran PF, Schlaut J, Solez K, Srinivasa NS. The significance of the anti-class I response: II. Clinical and pathologic features of renal transplants with anti-class I-like antibody. Transplantation 1992; 53: 550.
6. Martin S, Dyer PA, Mallick NP, Gokal R, Johnson RW. Posttransplant anti-donor lymphocytotoxic antibody production in relation to graft outcome. Transplantation 1987; 44: 50.
7. van Saase JLCM, van der woude FJ, Thorogood J, et al. The relation between acute vascular and interstitial renal allograft rejection and subsequent chronic rejection. Transplantation 1995; 59: 1280.
8. Herbertson BM, Evans DB, Calne RY, Banerjee AK. Percutaneous needle biopsies of renal allografts: the relationship between morphological changes present in biopsies and subsequent allograft function. Histopathology 1977; 1: 161.
9. Solez K, Axelsen RA, Benediktsson H, et al. International standardization of criteria for the histologic diagnosis of renal allograft rejection: the Banff working classification of kidney transplant pathology. Kidney Int 1993; 44: 411.
10. Lobo PI, Spencer CE, Stevenson WC, Pruett TL. Evidence demonstrating poor kidney graft survival when acute rejections are associated with IgG donor-specific lymphocytotoxin. Transplantation 1995; 59: 357.
11. Scornic JC, Salomon DR, Lim PB, Howard RJ, Pfaff WW. Posttransplant antidonor antibodies and graft rejection. Transplantation 1989; 47: 287.
12. Lucas ZJ, Coplon N, Kempson R, Cohn R. Early renal transplant failure associated with subliminal sensitization. Transplantation 1970; 10: 522.
13. Opelz G, Sengar DP, Mickey MR, Terasaki PI. Effect of blood transfusions on subsequent kidney transplants. Transplant Proc 1973; 5: 253.
14. Iwaki Y, Cecka JM, Terasaki PI. The transfusion effect. In: Terasaki P, ed. Clinical transplants. Los Angeles, CA: UCLA Tissue Typing Laboratory, 1988: 283.
15. Gritsch HA, Grimm PC, Ettenger RB, Rosenthal JT. Sensitized B lymphocytes contribute to acute allograft rejection. J Surg Res 1991; 51: 204.
16. Tuazon TV, Schneeberger EE, Bhan AK, et al. Mononuclear cells in acute allograft glomerulopathy. Am J Pathol 1987; 129: 119.
17. Hiki Y, Leong AS-Y, Mathew TH, Seymour AE, Pascoe V, Woodroffe AJ. Typing of intraglomerular mononuclear cells associated with transplant glomerular rejection. Clin Nephrol 1986; 26: 244.
18. Axelsen RA, Seymour AE, Mathew TH, Canny A, Pascoe V. Glomerular transplant rejection: a distinctive pattern of early graft injury. Clin Nephrol 1985; 25: 1.
19. Maryniak RK, First MR, Weiss MA. Transplant glomerulpathy: evolution of morphologically distinct changes. Kidney Int 1985; 27: 799.
20. Russ GR, Nicholls C, Sheldon A, Hay J. Positive B lymphocyte crossmatch and glomerular rejection in renal transplant recipients. Transplant Proc 1987; 19: 785.
21. Faull RJ, Russ GR, Nicholls C, Mathew TH. Early glomerular rejection in sensitized patients: treatment with plasma exchange and antithymocyte globulin. Transplant Proc 1988; 20: 1476.
22. Habib R, Antignac C, Hanglais N, Gagnadoux MF, Broyer M. Glomerular lesions in the transplanted kidney in children. Am J Kidney Dis 1987; 10: 198.
23. Boyce NW, Hayer K, Gee D. Cytomegalovirus infection complicating renal transplantation and its relationship to acute transplant glomerulopathy. Transplantation 1988; 45: 706.
24. Herrera GA, Alexander RW, Cooley CF, et al. Cytomegalovirus glomerulopathy: a controversial lesion. Kidney Int 1986; 29: 725.
25. Jeannet M, Pinn VW, Flax MH, Winn HJ, Russell PS. Humoral antibodies in renal allotransplantation in man. N Engl J Med 1970; 282: 111.
26. Mihatsch MJ, Ryffel B, Gudat F. Morphological criteria of chronic rejection: differential diagnosis, including cyclosporine nephropathy. Transplant Proc 1993; 25: 2031.
27. Hancock WW, Kraft N, Atkins RC. The immunohistochemical demonstration of major histocompatibility antigens in the human kidney using monoclonal antibodies. Pathology 1982; 14: 409.
28. Charpentier BM, Bach MA, Lang P, Martin B, Fries D. Phenotypic composition and in vitro functional capacities of unmodified fresh cells infiltrating acutely rejected human kidney allografts. Transplantation 1987; 44: 38.
29. Gaber LW, Gaber AO, Tolley EA, Hathaway DK. Prediction by postrevascularization biopsies of cadaveric kidney allografts of rejection, graft loss, and preservation nephropathy. Transplantation 1992; 53: 1219.
30. Barr ML, Cohen DJ, Benvenisty AI, et al. Effect of anti-HLA antibodies on the long-term survival of heart and kidney allografts. Transplant Proc 1993; 25: 262.
