• Agrin;
  • chronic allograft nephropathy;
  • glomerular basement membrane;
  • humoral immunity;
  • kidney transplantation;
  • transplant glomerulopathy


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Chronic allograft nephropathy (CAN) of renal allografts is still the most important cause of graft loss. A subset of these patients have transplant glomerulopathy (TGP), characterized by glomerular basement membrane (GBM) duplications, but of unknown etiology. Recently, a role for the immune system in the pathogenesis of TGP has been suggested. In 11 of 16 patients with TGP and in 3 of 16 controls with CAN in the absence of TGP we demonstrate circulating antibodies reactive with GBM isolates. The presence of anti-GBM antibodies was associated with the number of rejection episodes prior to diagnosis of TGP. Sera from the TGP patients also reacted with highly purified GBM heparan sulphate proteoglycans (HSPG). Indirect immunofluorescence with patient IgG showed a GBM-like staining pattern and colocalization with the HSPGs perlecan and especially agrin. Using patient IgG, we affinity purified the antigen and identified it as agrin. Reactivity with agrin was found in 7 of 16 (44%) of patients with TGP and in 7 of 11 (64%) patients with anti-GBM reactivity. In conclusion, we have identified a humoral response against the GBM-HSPG agrin in patients with TGP, which may play a role in the pathogenesis of TGP.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Chronic allograft nephropathy (CAN) is one of the most important causes of graft failure after the first 6 months post-transplantation (Tx) and may affect over half of the renal allograft recipients (1,2). CAN is characterized by a slow but variable decline in renal function after the initial 3 post-Tx months, often in combination with proteinuria and hypertension (3,4). Histologically, biopsies with CAN show fibrous intimal thickening of arteries, glomerulosclerosis, interstitial fibrosis and tubular atrophy (5). Clinically, the lesions of CAN are still difficult to treat. Risk-factor analysis for the development of CAN has revealed that late acute rejection episodes (after the first 3 months post-Tx), recipient age and race, pre-transplant sensitization and injury, HLA-matching and inadequate immunosuppression are important in the pathogenesis (6).

In 1964, Hamburger et al. described transplant glomerulopathy (TGP) (7), a lesion characterized by enlargement of the glomeruli with swelling of the endothelial and mesangial cells, mesangiolysis, infiltration of the glomeruli with mononuclear cells, mesangial matrix expansion and show splitting of the glomerular basement membrane (GBM) with electron lucent subendothelial material (8–10). TGP can be discriminated from recurrent or de novo membranoproliferative glomerulonephritis using electron and immunofluorescence microscopy. In TGP the GBM is electron-lucent whereas in membrane proliferative glomerulonephritis (MPGN) electron-dense deposits are present (11). Furthermore, patients with TGP show no immunoglobulin deposits or scanned IgM with a greater intensity than C3, whereas MPGN patients showed a greater intensity of C3 (11). The presence of TGP in a renal biopsy is associated with accelerated graft loss (12). About 5–15% of patients with CAN have TGP, which is about 1.5–3% of all renal allograft recipients (9,13). The etiology of this entity is not yet known although it is thought that immunological mechanisms are involved (14). Recently, we characterized a group of 18 patients with TGP and confirmed the diagnosis by reexamination of the original biopsies (13). Clinically, TGP was diagnosed in 14% of patients with CAN (i.e. 1.6% of renal allograft recipients) and was diagnosed at a median of 8.3 years post-Tx. Although TGP was diagnosed later as compared to CAN, graft survival after diagnosis was similar. Clinical data from these patients, also used in this study, were used to perform risk-factor analysis. We identified late acute rejection episodes beyond 3 months post-Tx and pre-transplant sensitization as independent risk factors for the development of TGP (13), supporting an immunological pathogenesis. Moreover, in the majority of patients with TGP, the presence of C4d has been demonstrated in a GBM-like pattern on paraffin sections (13,15) and in peritubular capillaries (14,16). C4d is considered a ‘stable’ marker of previous humoral immunity (17–19). The presence of antibodies reactive with human leukocyte antigens (HLA) has been associated with CAN (20–22). Although antibodies to HLA antigens seem important, not all patients that suffer from CAN have antibodies reactive with donor-type HLA-antigens (22). Therefore, antibodies reactive with other (organ-specific) antigens might contribute to the pathogenesis of TGP.

The most characteristic lesion of TGP is duplication of the GBM. The GBM is a basement membrane specialized in ultrafiltration and consists of various matrix molecules, including fibronectin, collagens and HSPGs. In the rat, injection of antibodies reactive with GBM antigens, in particular HSPGs, can result in structural alterations of the GBM and proteinuria (23,24). Furthermore, it has been demonstrated that the amount of HSPG in the GBM is decreased in several glomerular diseases (25,26). Agrin is the most abundantly expressed HSPG in the GBM, while perlecan expression is more prominent in the mesangial matrix and Bowman's capsule (26,27).

Recently, we identified antibodies reactive with GBM antigens in the F344 to LEW renal transplant model for chronic rejection, that has glomerular lesions characteristic for TGP including GBM duplications (28). The antibodies developed in the first few weeks after transplantation and increased in time. The response was donor-GBM specific and we identified the GBM heparan sulphate proteoglycan (HSPG) perlecan as the major antigen (28).

In the present study we investigated whether patients with TGP also develop antibodies reactive with GBM proteins that may be involved in the development of specific lesions. We found that a large proportion of sera of patients with TGP indeed showed reactivity with total GBM preparations as well as with highly purified GBM-HSPG. Using immuno-affinity purification we identified agrin as the antigen recognized by 44% of TGP patients.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References


Sixteen patients with TGP were selected from all 1111 renal transplant recipients with at least 6 months of graft function, transplanted in Leiden in the period from January 1983 to January 2001 as described in a previous study (13). Sera (n = 16) were available from 14 of 22 patients described in (13), and from two additional patients that were recently diagnosed. Histopathologically, TGP was defined as double contours of the GBM or enlargement of the glomeruli with swelling of endothelial and mesangial cells, mesangiolysis. Infiltration of the glomeruli with mononuclear cells was present in the four acute TGP cases. In addition, all patients had a negative immunofluorescence staining for IgA, IgM and C3 or scant IgM deposits with greater intensity than C3 to exclude de novo membrane proliferative glomerulonephritis (MPGN). Furthermore, patients that had MPGN as original disease or patients that were hepatitis C seropositive at the time of transplantation were excluded. Patients received an immunosuppressive regimen consisting of prednisone and cyclosporin A and/or azathioprine. In 1996, Sandimmune was changed to Neoral.

Sera were collected at the time of biopsy and stored at −80°C. Patient characteristics are summarized in Table 1. Patient numbers are similar as in our previous study (13). The presence of circulating antibodies reactive with the HLA class I and/or II mismatches between donor and recipient were determined at the time of biopsy. Sera were tested using standard flow-cytometric cross-matching with splenocytes, the specificity of the anti-HLA antibodies was determined by ELISA (Lambda Antigen Tray, LAT 1240, One Lambda, Inc. Canoga Park, CA).

Table 1.  Characteristics of TGP patients
Patient*Original diseaseTime (years)C4dglomAntibodies reactive with
HLAGBM p-valueGoodpasture#HSPG
  1. *Patient numbers 1–20 are similar as previously reported (13), patients 23 and 24 were recently diagnosed.

  2. Time in years after transplantation until the diagnosis of TGP.

  3. The glomerular deposition of C4d is indicated in this table as previously described (13).

  4. Anti-GBM antibodies were measured in at least six independent ELISAs for all individual patients, data were expressed as ratios to the normal human serum included in all plates. A Wilcoxon signed rank test with a theoretical median of 1 was performed and p-values < 0.05 were considered significant (in bold), that is, patients have anti-GBM antibodies. Patient 23 reached statistical significance because all ratios were below 1, therefore this patient does not have anti-GBM antibodies.

  5. #Reactivity with Goodpasture epitope.

  6. Reactivity with highly purified GBM-HSPG.

  7. §Original disease of patients is unknown, patients were referred to our center for transplantation after they had started dialysis.

  8. APKD: autosomal polycystic kidney disease; GN: glomerulonephritis; DM: diabetes mellitus; CIN: chronic interstitial nephritis; IgAN: IgA nephropathy; HT: hypertension; HSPG: heparan sulphate proteoglycan; NA = not available.

6DM I5.20.2274
16DM I11.3NA+<0.0001+
19DM II0.6++0.0001+

Patient 19 was transplanted because of type II diabetes, and received a cadaveric renal allograft with two HLA class I and one class II mismatches. After 3 months she developed proteinuria (>1 g/24 h) and a biopsy revealed acute TGP. In a second biopsy, performed after 6 months, GBM duplications were present. Serologically, she produced anti-HLA DR 1 antibodies (anti-donor HLA antibodies) and anti-GBM antibodies. She was treated with plasmapheresis to remove these antibodies. Plasma obtained during plasmapheresis was collected and used for some of the experiments described below.

A control group consisted of 16 randomly selected patients with CAN in the absence of TGP-specific lesions. All patients in the control group were from the same transplant era as the TGP patients, were at least 3 years after transplantation and were treated with an immunosuppressive regimen similar to the TGP patients.

GBM isolation and ELISA

Human GBMs were isolated from human donor kidneys not suitable for transplantation due to anatomical reasons. Isolations were performed as previously described for rat GBM (28).

Collagenase digested GBM (0.16 μg/well) was coated to 96 well ELISA plates (Greiner, Alphen aan den Rijn, the Netherlands). All patient sera (of all 16 patients) have been tested on two independently isolated GBM preparations derived from two different donor kidneys. After blocking with phosphate-buffered saline (PBS)/1% bovine serum albumin (BSA), plates were incubated with serial dilutions (starting at 1 to 25) of patient or control sera. IgG binding was detected using digoxigenin (DIG)-conjugated mouse anti-human IgG (HB43, hybridoma obtained from the American Type Culture Collection, Manassas, VA), followed by incubation with horseradish peroxidase (HRP) -conjugated sheep F(ab′) fragments anti-DIG (Roche Diagnostics, Almere, the Netherlands) and staining with the peroxidase substrate ABTS (2,2′-amino-bis-3-ethylbenzthiazoline-6-sulfonic acid, Sigma Chemical Co., St. Louis, MO). The optical density was measured at 415 nm using a Titertek multiscan plate reader (Flow Laboratories, Zwanenburg, the Netherlands). To determine the isotype of the response, binding of antibodies to the GBM coated plates was detected with antibodies specific for human immunoglobulin isotypes. IgG antibodies were detected as described above, IgA and IgM were detected using mouse anti-human monoclonal antibodies 4E8 (our own laboratory (29)) and HB57 (hybridoma obtained from the American Type Culture Collection). Purified immunoglobulins were used as positive controls.

The presence of antibodies reactive with the Goodpasture epitope were measured using a commercial ELISA kit that contained wells coated with recombinant α3 collagen type IV (anti-GBM antibodies, Wielisa kit, MP products, Amersfoort, the Netherlands).

HSPG isolation and ELISA

Human GBM-HSPGs were purified by a combination of anion exchange and gel filtration chromatography, as previously described (30). Purified HSPGs, derived from pooled human kidneys, were coated in 96 well ELISA plates (200 ng/well; Nunc Maxisorp F96, Life Technologies, Breda, the Netherlands). Plates were blocked with PBS/1% BSA and incubated with 1–30 diluted sera of all 16 patients. The ELISA was developed using the same reagents as described above.

Indirect immunofluorescence

To perform indirect immunofluorescence stainings, patient and control IgG were purified. Since only limited amounts of serum were available, a specific IgG isolation method was used (31). Serum (400 μL, all four patients that showed HSPG reactivity and had sufficient amounts of serum left) was diluted 1–5 in H2O and applied on mini-columns containing SPC 25 (Pharmacia, Uppsala, Sweden) and DE 52 (diethylaminoethyl cellulose; Whatman International Ltd., Maidstone, UK). After 1 h columns were centrifuged and IgG was present in the fall through (31). Protein content was measured using OD 280.

IgG preparations were conjugated with DIG using DIG-NHS (=digoxigenin-3-0-methylcarbonyl-ɛ-amino caprioc acid-N-hydroxysuccinimide ester, Roche Diagnostics) and free probe was removed by dialyzation against PBS.

Frozen sections (3 μm) from normal human kidneys (HLA-DR1 negative) were acetone fixed and endogenous peroxidase was blocked using hydrogen peroxide and sodium azide. Aspecific protein binding was blocked with PBS/1% BSA. Subsequently the sections were incubated with 0.3 mg/mL DIG-conjugated patient or control IgG overnight at 4°C. Finally, the sections were incubated with HRP conjugated sheep F(ab′) anti DIG followed by staining with tyramid-FITC and embedding in DABCO-glycerol (1,4-diazabicyclo-(2,2,2)-octane; Sigma).

Double stainings of DIG-conjugated patient IgG with a monoclonal antibody against the core protein of agrin (JM72; 1 to 200) (32) or a cocktail of monoclonal antibodies against perlecan (1 to 500; kind gift of Dr. G. David, Leuven, Belgium) (33) were performed. Monoclonal antibodies were stained with Alexa 548 conjugated goat anti-mouse IgG antibodies (Molecular Probes, Leiden, the Netherlands). Slides were analyzed by confocal microscopy on a Zeiss LSM 510 (Zeiss b.v., Sliedrecht, the Netherlands).

Immuno-affinity column

IgG was isolated from plasma of patient 19 derived from the first two plasmapheresis treatments. One and a half liters of plasma were precipitated with 40% (w/v) ammoniumsulphate for 1 h at room temperature and 1 h at 4°C followed by centrifugation at 6310 g for 30 min. The pellet was dissolved in 400 mL H2O and dialyzed against PBS/2 mM EDTA. Thirty milliliters of this sample was applied on a 100 mL DEAE-sephacel column (Pharmacia). IgG containing samples were pooled, concentrated and tested for anti-GBM reactivity.

Purified IgG (68 mg) from patient 19 was coupled to 10 mL Biogel A5 (Bio-Rad Laboratories, Richmond, CA) using activated cyanogen-bromide. The IgG-Biogel column was equilibrated with 10 column volumes of PBS/2 mM EDTA before applying the sample. Pooled, collagenase-digested GBM (1 mg) from two different donor kidneys (negative for HLA-DR1) was applied on the immuno-affinity column. Bound antigens were eluted using 0.1 M glycin/0.3 M NaCl pH 2.8, and 2 mL fractions were collected, which were immediately neutralized with a Tris-buffer. Fractions were tested for their total protein content using the bicinchoninic acid (BCA) protein assay (Pierce Chemical Co., Rockford, IL). Subsequently the fractions were coated to ELISA plates (1/20 in PBS) and detected using patient sera (1/25) and anti-IgG antibodies (HB43). Based on the protein and ELISA data peak-fractions were pooled and further designated as fall through and eluate.

Identification of the antigen

Total GBM, fall-through and eluate were coated to 96-well ELISA plates and incubated with antibodies against HLA-DR and known matrix proteins. HLA-DR was detected using a pan-DR antibody (clone B.11,2). Matrix proteins were detected with the following antibodies: goat anti-fibronectin (Sigma), rabbit anti-laminin (E-Y laboratories, San Mateo, CA), goat anti-collagen type IV (Immunologicals Direct, Oxfordshire, UK), mouse anti-perlecan (33) and mouse anti-agrin (JM72) (32,34).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Sera from TGP patients contain antibodies reactive with GBM isolates

Serum from two independent patients with TGP and a normal human serum (NHS) were tested in a GBM-ELISA. Serum from both TGP patients bound dose-dependently to GBM-coated plates, whereas the normal human serum did not bind the plates (Figure 1A). Since maximal binding was achieved using 4% of patient serum, with minimal signal of the NHS, we decided to use 4% of serum for the rest of the experiments. Initially sera from patients were tested on GBM preparations derived from two independent donor kidneys (A and B). Some patients exclusively recognized GBM derived from one kidney (patient 1) but not from the other, whereas other patients (patient 19) recognized GBM derived from both kidneys (Figure 1B). Therefore, subsequent experiments were performed with a mix of both preparations.


Figure 1. Anti-GBM antibodies are present in sera from TGP patients. (A) Serum from two different TGP patients and a NHS was tested in dose-response on GBM coated plates and stained for IgG binding. (B) Sera (1/25) were tested on GBM preparations derived from two different kidneys (A/B) and stained for measured IgG binding.

Download figure to PowerPoint

Sera of all 16 patients with TGP were tested for the presence of anti-GBM antibodies in at least three independent ELISAs. To compare the independent experiments, data were normalized to NHS and the mean of all samples was calculated. Patients were called positive if the ratios of patient serum to NHS were significantly (p < 0.05) higher than 1, tested with the Wilcoxon signed rank test. Anti-GBM antibodies were detected in serum of 11 of 16 patients with TGP and confirmed in at least 3 independent ELISA's (Figure 2A) (Table 1). All reactivity with the GBM was of the IgG isotype, whereas IgM and IgA could not be detected (data not shown). A similar analysis was performed using sera of 16 patients with CAN in the absence of TGP. In this control group anti-GBM antibodies were present in 3 of 16 patients (Figure 2B). Significantly more patients had anti-GBM antibodies in the TGP group compared to the control (CAN) group (p = 0.0044, chi-square test).


Figure 2. Sera from TGP patients contain antibodies reactive with GBM isolates. (A) Sera (1/25) from 16 patients (numbers identical to previous publication (13)) with TGP were incubated on total, collagenase digested GBM, coated to ELISA plates and IgG binding was measured. Sera were measured in at least three independent ELISAs using GBM isolated from two different kidneys. Ratios were calculated using the optical density reached by patient serum on the NHS included in the same plate. Data are expressed as the mean plus the standard error of the mean (SEM) of these ratios. Patients have anti-GBM antibodies if the Wilcoxon signed rank test was significantly different from 1. (*)A p-value <0.05 was considered significant. The line indicates a ratio of 1, that is, equal to NHS. (B) Sera from 16 patients with CAN in the absence of TGP were tested in a similar fashion as (A). Data were tested using the Wilcoxon signed rank test and a p-value of <0.05 was considered significant. (C) Sera of 11 of the TGP patients were tested in a ELISA specific for the Goodpasture epitope (NC1 α3 collagen type IV). The cut-off (line) is based on standard samples included in the assay and is used to determine which patients have Goodpasture antibodies.

Download figure to PowerPoint

The anti-GBM antibodies found in TGP patients were not reactive with the well-characterized Goodpasture antigen, NC1 α3 chain of collagen type IV (Figure 2C), indicating that other basement membrane antigens are involved.

TGP patients with antibodies have experienced more rejection episodes

Patients with and without anti-GBM antibodies (Table 1) were compared for several clinical parameters. Patients with or without anti-GBM antibodies did not show statistically significant differences in serum creatinin levels (Figure 3A) (p = 0.3957 (T-test)) or creatinin clearance (data not shown; p = 0.3927) measured at the time of biopsy. Similarly, proteinuria at the time of biopsy is equal between TGP patients with or without anti-GBM antibodies (Figure 3B) (p = 0.2841 (T-test)). If the number of rejection episodes is compared between patients with and without antibodies a significant difference is found. Patients with anti-GBM antibodies have experienced significantly more rejection episodes as TGP patients without anti-GBM antibodies (Figure 3C) (p = 0.0008 (T-test)).


Figure 3. TGP patients with anti-GBM antibodies experienced more rejection episodes. (A) Serum creatinin values (mmol/L) at the time of biopsy in patients with and without anti-GBM antibodies (p = 0.3957, T-test). (B) Proteinuria (g/24 h) at the time of biopsy in patients with and without antibodies (p = 0.2841, T-test). (C) Number of rejection episodes prior to diagnosis of TGP in patients with and without anti-GBM antibodies (p = 0.0008, T-test).

Download figure to PowerPoint

Antibodies specific for donor-HLA antigens were detected in 8 of 16 patients with TGP (Table 1). Six patients had anti-GBM antibodies in the absence of anti-HLA antibodies. Six TGP patients had both anti-HLA and anti-GBM antibodies, whereas two patients had only anti-donor and two patients had no antibodies at all. The number of rejection episodes did not correlate with the presence of anti-HLA antibodies (p = 0.5709 (T-test)) (data not shown).

TGP patients recognize highly purified HSPGs

In the F344 to LEW rat model for CR, we have previously identified antibodies reactive with the GBM-HSPG perlecan (28). Glomeruli of these rat kidney allografts showed GBM duplications from day 30 onwards. Banff scores for TGP were on average 1.5–2 and the majority of glomeruli was affected. Since the lesions observed in the rat model and in patients with TGP are very similar, HSPG molecules might also be involved in patients with TGP. Therefore, highly purified GBM-HSPGs containing at least perlecan and agrin (Figure 4A) were coated in an ELISA and tested as antigen. Sera from 7 of 16 patients (44%) tested reacted with purified GBM-HSPGs (Figure 4B, Table 1). All seven patients also recognized total GBM, whereas none of the anti-GBM negative patients recognized the purified HSPGs. These seven patients were used for more detailed analysis of the antigen(s) involved.


Figure 4. TGP patients recognize highly purified HSPGs. (A) Highly purified GBM-HSPGs were coated and incubated with monoclonal antibodies directed against the most important glomerular HSPGs perlecan and agrin. (B) Sera (1/30) from TGP patients were incubated on the highly purified GBM-HSPGs coated to ELISA plates. IgG binding was measured and sera were considered positive when binding was higher than the mean + 2 times the standard deviation of pooled NHS (line).

Download figure to PowerPoint

Purified IgG from TGP patients colocalizes with agrin in the GBM

Patient IgG was purified from all four patients that recognized the purified GBM-HSPGs and of whom sufficient serum was available. Indirect immunofluorescence with these sera on normal human kidney sections revealed glomerular as well as interstitial staining. The glomeruli showed a linear GBM staining pattern (Figure 5A). In addition, some tubular basement membranes and Bowman's capsule were stained. IgG purified from normal human serum showed no reactivity (Figure 5B). Glomerular staining was observed using IgG from all four patients and in 0/4 control sera. Glomerular staining was inhibited by pre-incubation of patient IgG with soluble GBM (data not shown).


Figure 5. Purified IgG from TGP patients recognizes agrin in the GBM. Patient and control IgG was purified and used to stain (0.3 mg/mL) normal human kidney sections. (A, C, F) Purified, serum IgG from TGP patient 16 results in linear GBM staining. (B) Purified, normal IgG does not show glomerular staining. (D) Perlecan (1/500) is expressed in the GBM, mesangial matrix, Bowman's capsule and on the tubular basement membranes. (E) On the GBM there is minor colocalization between patient IgG and perlecan (merge picture). (G) Agrin (1/200) is expressed most prominently in the GBM and the a lesser extent in Bowman's capsule and on the tubular basement membranes. (H) Agrin and patient IgG colocalize on the GBM (merge picture). Original magnification × 250.

Download figure to PowerPoint

Double stainings were performed using patient IgG and monoclonal antibodies against the most abundantly expressed glomerular HSPGs, perlecan and agrin. Double staining of patient IgG with anti-perlecan antibodies showed a partial colocalization in the glomerulus, especially in the mesangial matrix and Bowman's capsule, and to a lesser extent along the GBM (Figures 5C–E). In addition, staining with the patient IgG was demonstrated inside as well as outside the glomeruli, including the TBM, which did not colocalize with the perlecan staining. Double staining of patient IgG with the anti-agrin monoclonal antibody showed a more abundant colocalization in the glomerulus, especially along the GBM (Figures 5F–H). However, the patient IgG also recognized antigens on the TBM that are not recognized by this monoclonal antibody against agrin.

The antigen is recognized by all patients that recognized GBM isolates

One patient (patient 19) was treated with plasmapheresis to remove anti-donor HLA and anti-GBM antibodies from circulation. IgG was purified from the plasma obtained at plasmapheresis, coupled to sepharose beads and used for immuno-affinity purification. The largest amount of GBM proteins was present in the fall through, whereas small amounts of protein could be detected in the eluted fractions (Figure 6A). Specific reactivity of patient serum was only observed with the eluted fractions of the column. This reactivity of TGP patient serum was similar using serum from the patient used to isolate the antigen (patient 19) and another patient (Figure 6B, patient 16). Reactivity of sera of the patients was tested on the pooled fall through and the eluate of the immuno-affinity column. All patients did recognize the eluate (Figure 7A). In addition, five patients also recognized to a variable extent additional proteins still present in the fall through (Figure 7B, Table 2). SDS-PAGE electrophoresis of the eluate with subsequent silver staining showed one band of over 200 kDa, however this can not exclude the presence of low amounts of other proteins. Unfortunately mass spectrometry did not help to identify the protein(s) involved.


Figure 6. Purification and reactivity of the antigen recognized by TGP patient sera. Immuno-affinity purification of the antigen recognized by patient 19, elution with glycin, NaCl at pH 2.8 started at fraction 40 (arrow). (A) The majority of proteins (BCA assay, OD 562) are present in the fall through fractions of the immuno-affinity column and not in the eluate. (B) Sera (1/25) from patient 19 and other patients (#16) recognize specifically the eluate (1/25) but not the fall through of the affinity purification.

Download figure to PowerPoint


Figure 7. TGP patient sera recognize the eluate and not the fall through of the immuno-affinity column. (A) All TGP patient sera (1/25) that were positive on purified GBM-HSPG proteins recognize the proteins eluted (1/25) from the immuno-affinity column. (B) Five TGP patients recognize proteins in the fall through of the immuno-affinity column in a variable extent in addition to the eluate. The line indicates the mean + 2 times the standard deviation of pooled NHS.

Download figure to PowerPoint

Table 2.  Summary of patients positive on HSPG
PatientHSPG*IFFall throughEluate
  1. *Reactivity with highly purified GBM-HSPG.

  2. Glomerular staining.

  3. Reactivity with fall through or eluate of GBM-loaded immuno-affinity column; ±: borderline positive.

  4. HSPG: heparan sulphate proteoglycans; IF: immunofluorescence; NA: not available.


Agrin is the antigen purified using patient IgG

Antibodies reactive with known GBM proteins were used to identify proteins present in the eluate of the column. The eluate did not contain detectable levels of HLA-DR. We found that major components of the GBM-like laminin, collagen type IV and fibronectin were all present in the fall through of the column but not in the eluate (data not shown). Next we tested two of the HSPGs that are expressed in the GBM, perlecan (Figure 8A) and agrin (Figure 8B). Perlecan was present in the fall through but not in the eluate of the column. In contrast, agrin was present in the eluate of the column. In summary, the eluate of the column contained agrin and not any of the other GBM molecules (fibronectin, laminin, collagen type IV, perlecan) that were tested (data not shown).


Figure 8. Agrin is the antigen recognized by TGP patients. Proteins purified by the immuno-affinity column were probed with antibodies against the major glomerular HSPG molecules. (A) Perlecan (1/50, 1/100, 1/500) is present in the fall through but not the eluate of the affinity column. (B) Agrin (1/200, 1/800, 1/3200) is present in the eluate of the affinity column. Data are expressed as mean OD 415 nm of two samples (+standard deviation), white bars represent the second antibody only.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Antibodies reactive with highly purified HSPGs were found in patients with TGP. Indirect immunofluorescence using patient sera on normal human kidneys showed linear GBM staining patterns that can be inhibited by addition of soluble GBM. The glomerular staining of patient IgG colocalized with the HSPG agrin along the GBM and to a lesser extent with perlecan. After immuno-affinity purification of GBM with TGP patient serum, the HSPG agrin was identified as the antigen recognized by 44% of the TGP patient sera.

In the F344 to LEW model of chronic renal allograft rejection antibodies reactive with GBM isolates were detected and the HSPG perlecan was identified as one of the GBM antigens (28). So both in an experimental model for TGP (28) and in patients with TGP antibodies against a GBM-HSPG were detected. This may imply that these antibodies play a role in the pathogenesis of TGP (14). The GBM-HSPGs are important for the structure of the GBM and specifically the glomerular permselectivity. Alterations in the structure of the HSPG molecules will lead to structural GBM changes and protein leakage (26,27,35,36). It is known that injection of antibodies against HSPGs in rats results in basement membrane duplications and proteinuria (23,24). Although we do not know whether the antibodies reactive with agrin in patients are involved in the induction of proteinuria, it is known that all patients with TGP have proteinuria. In patient 19 the antibodies preceded proteinuria at a moment that serum creatinin levels were stable, speculating that antibodies are involved (data not shown). In the total group of 16 patients we found no correlation between the presence of antibodies and the level of proteinuria, serum creatinin or creatinin clearance. However, the presence of antibodies was associated with an increased number of previous rejection episodes. This supports the hypothesis that the anti-GBM antibodies detected in the circulating are the result of graft damage or local inflammation and exposure of cryptic epitopes upon previous acute rejection episodes. One argument that the antibodies also recognize non-damaged basement membrane antigens is that isolated immunoglobulins stain normal kidney sections. However this does not exclude the possibility that a certain amount of damage is necessary for the initial induction of antibody production. Before these issues can be addressed, more detailed analyses will be necessary to obtain insights in the molecular variation and alterations upon damage in basement membrane molecules. In addition, longitudinal follow-up of serum antibody levels of a group of patients at risk for developing TGP will be helpful to answer this question.

Patients with TGP are generally diagnosed years after Tx and present with a gradual decline in function, suggesting that the pathogenesis develops slowly. The production of antibodies against agrin might be a consequence of previous damage. Late acute rejection episodes are an important risk factor present in the majority of TGP patients. Late acute rejection episodes or primary non-function might induce damage resulting in the presentation of otherwise hidden epitopes and thereby in the induction of humoral responses. This hypothesis is further supported by the finding of a relation between the presence of antibodies and the number of previous rejection episodes. Alternatively, antibodies reactive with donor HLA antigens might be a trigger for the production of antibodies reactive with matrix proteins-like agrin. The timing of TGP, generally years after transplantation, suggests that a previous event is necessary to induce the antibodies and the subsequent damage. Next to these autoimmune-based explanations, also some arguments for the involvement of alloimmune responses are present. It is generally observed that patients with TGP do not present with pulmonary problems despite the fact that agrin is expressed in the lung (37). In addition, agrin can be found in skin and muscle basement membranes and in neurological and immunological synapses. Furthermore, some of the patients did not recognize all batches of GBM isolated from different kidneys. Finally, the experiments in the F344 to LEW model showed a strict donor-specific response. TGP patients may recognize aberrant or alternatively spliced forms of agrin or agrin polymorphisms, which normally are not ubiquitously expressed in the GBM. Several alternative splice forms of (renal) agrin have been described (38–40). Detailed analysis of epitopes on agrin that are recognized by the TGP patients will be necessary to obtain more insights in the origin of the responses.

Circulating and glomerular deposits of anti-GBM antibodies are frequently found in patients with Goodpasture syndrome (41,42). These patients produce antibodies reactive with the non-collagenous (NC1) part of the α3 chain of collagen type IV and show IgG and C3 deposits in a renal biopsy (41–43). Although others have found occasional deposits of immunoglobulins and complement proteins in glomeruli of patients with TGP (44), routine direct immunofluorescence did not reveal clear immunoglobulin or complement depositions in the glomeruli of these biopsies. In addition, patient sera did not react with the Goodpasture epitope, excluding the de novo development of Goodpasture syndrome upon renal Tx. None of the patients in this group had Goodpasture syndrome as original disease, excluding the recurrence of Goodpasture syndrome.

In the group of TGP patients we have investigated, not all patients reacted with the GBM preparations used. This might be the consequence of antigenic differences in GBM, as described above, or relate to the limited number of serum samples available from the patients. In the present study we were able to follow one patient (#19) in time. The anti-HSPG antibody titers varied in time, first detectable a few weeks after renal Tx (data not shown). In addition to the glomerular reactivity, some of the patients recognized structures outside of the glomerulus suggesting that other antigens were recognized in addition to agrin. This is further supported by the finding of strong reactivity with the fall through of the immuno-affinity column in two of the patients. The nature of these antigens is at present unclear. However, they may include alternatively spliced agrin variants or perlecan, which are not retained by the coupled IgG from patient 19. The staining of mesangial matrix, Bowman's capsule and tubular basement membranes by some TGP patient sera may support the presence of antibodies specific for both agrin splice-variants and perlecan (26,27,40). In addition, the immunofluorescent stainings only showed a partial colocalization with agrin and perlecan, next to additional staining with the patient IgG. This suggests that other antigens are also recognized by the patient sera, or that the monoclonal anti-agrin antibody does not recognize all forms of agrin present in the kidney.

The presence of anti-GBM antibodies was not exclusive for patients with TGP, but also observed in 3 of 16 patients with CAN. Two out of these three patients also recognized both the eluate and fall through of the immunoabsorbance column, whereas the third CAN patients did recognize neither the fall trough nor the eluate (data not shown). If antibody production precedes the development of TGP, these patients might be in the process of developing TGP. Alternatively, if antibody production is the result of altered epitopes after induction of chronic damage and inflammation, the patients might not develop TGP. Unfortunately, no follow-up material of these three patients is available to investigate the development of TGP later in time.

In conclusion, we identified an antibody response against agrin, which is the most abundantly expressed HSPG in the GBM, in patients with TGP. This finding supports the concept that TGP is mediated by humoral immune responses against non-HLA antigens.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This work was supported by a grant from the Dutch Kidney Foundation (C98.1783). We thank Maria Borrias (Department of Nephrology, LUMC, Leiden, the Netherlands) for technical assistance, Frans Prins (Department of Pathology, LUMC, Leiden, the Netherlands) for help with confocal microscopy and Prof. Dr. F.H.J. Claas (Department of Immunohematology and Bloodbank, LUMC, Leiden, the Netherlands) for the critical reading of the manuscript.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Paul LC. Chronic renal transplant loss. Kidney Int 1995; 47: 14911499.
  • 2
    Hariharan S, Johnson CP, Bresnahan BA, Taranto SE, McIntosh MJ, Stablein D. Improved graft survival after renal transplantation in the United States, 1988 to 1996. N Engl J Med 2000; 342: 605612.DOI: 10.1056/NEJM200003023420901
  • 3
    Paul LC, Hayry P, Foegh M et al. Diagnostic criteria for chronic rejection/accelerated graft atherosclerosis in heart and kidney transplants: joint proposal from the Fourth Alexis Carrel Conference on chronic rejection and accelerated arteriosclerosis in transplanted organs. Transplant Proc 1993; 25: 20222023.
  • 4
    Kasiske BL, Andany MA, Danielson B. A thirty percent chronic decline in inverse serum creatinine is an excellent predictor of late renal allograft failure. Am J Kidney Dis 2002; 39: 762768.
  • 5
    Mauiyyedi S, Colvin RB. Pathology of Kidney Transplantation. In: MorrisPJ, ed. Kidney Transplantation. Oxford : W.B. Saunders Company; 2002: 343376.
  • 6
    Sijpkens YW, Joosten SA, Paul LC. Chronic rejection in renal transplantation. Transplant Rev 2003; 17: 118130.
  • 7
    Hamburger J, Crosnier J, Dormont J. Observations in patients with a well-tolerated homotransplanted kidney: possibility of a new secondary disease. Ann N Y Acad Sci 1964; 120: 558577.
  • 8
    Maryniak RK, First MR, Weiss MA. Transplant glomerulopathy: evolution of morphologically distinct changes. Kidney Int 1985; 27: 799806.
  • 9
    Habib R, Broyer M. Clinical significance of allograft glomerulopathy. Kidney Int Suppl 1993; 43: S9598.
  • 10
    Olsen TS. Pathology of allograft rejection. In: BurdickJF, RacusenLC, SolezK, WilliamsGM, eds. Kidney transplant rejection. Diagnosis and treatment. New York, Basel , Hong Kong : Marcel Dekker Inc; 1992: 333.
  • 11
    Andresdottir MB, Assmann KJ, Koene RA, Wetzels JF. Immunohistological and ultrastructural differences between recurrent type I membranoproliferative glomerulonephritis and chronic transplant glomerulopathy. Am J Kidney Dis 1998; 32: 582588.
  • 12
    Suri DL, Tomlanovich SJ, Olson JL, Meyer TW. Transplant glomerulopathy as a cause of late graft loss. Am J Kidney Dis 2000; 35: 674680.
  • 13
    Sijpkens YW, Joosten SA, Wong M-C et al. Immunological risk factors and glomerular C4d deposits in chronic transplant glomerulopathy. Kidney Int 2004; 65: 24092418.DOI: 10.1111/j.1523-1755.2004.00662.x
  • 14
    Vongwiwatana A, Gourishankar S, Campbell PM, Solez K, Halloran PF. Peritubular capillary changes and C4d deposits are associated with transplant glomerulopathy but not IgA nephropathy. Am J Transplant 2004; 4: 124129.DOI: 10.1046/j.1600-6143.2003.00294.x
  • 15
    Regele H, Bohmig GA, Habicht A et al. Capillary deposition of complement split product C4d in renal allografts is associated with basement membrane injury in peritubular and glomerular capillaries: a contribution of humoral immunity to chronic allograft rejection. J Am Soc Nephrol 2002; 13: 23712380.DOI: 10.1097/01.ASN.0000025780.03790.0F
  • 16
    Horita S, Nitta K, Kawashima M et al. C4d deposition in the glomeruli and peritubular capillaries associated with transplant glomerulopathy. Clin Transplant 2003; 17: 325330.DOI: 10.1034/j.1399-0012.2003.t01-1-00014.x
  • 17
    Bohmig GA, Exner M, Habicht A et al. Capillary C4d deposition in kidney allografts: a specific marker of alloantibody-dependent graft injury. J Am Soc Nephrol 2002; 13: 10911099.
  • 18
    Mauiyyedi S, Pelle PD, Saidman S et al. Chronic humoral rejection: identification of antibody-mediated chronic renal allograft rejection by C4d deposits in peritubular capillaries. J Am Soc Nephrol 2001; 12: 574582.
  • 19
    Collins AB, Schneeberger EE, Pascual MA et al. Complement activation in acute humoral renal allograft rejection: diagnostic significance of C4d deposits in peritubular capillaries. J Am Soc Nephrol 1999; 10: 22082214.
  • 20
    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 2001; 59: 334341.DOI: 10.1046/j.1523-1755.2001.00495.x
  • 21
    Sumitran-Holgersson S. HLA-specific alloantibodies and renal graft outcome. Nephrol Dial Transplant 2001; 16: 897904.DOI: 10.1093/ndt/16.5.897
  • 22
    McKenna RM, Takemoto SK, Terasaki PI. Anti-HLA antibodies after solid organ transplantation. Transplantation 2000; 69: 319326.DOI: 10.1097/00007890-200002150-00001
  • 23
    Miettinen A, Stow JL, Mentone S, Farquhar MG. Antibodies to basement membrane heparan sulfate proteoglycans bind to the laminae rarae of the glomerular basement membrane (GBM) and induce subepithelial GBM thickening. J Exp Med 1986; 163: 10641084.DOI: 10.1084/jem.163.5.1064
  • 24
    Van Den Born J, Van den Heuvel LP, Bakker MA, Veerkamp JH, Assmann KJ, Berden JH. A monoclonal antibody against GBM heparan sulfate induces an acute selective proteinuria in rats. Kidney Int 1992; 41: 115123.
  • 25
    Van Den Born J, Van den Heuvel LP, Bakker MA et al. Distribution of GBM heparan sulfate proteoglycan core protein and side chains in human glomerular diseases. Kidney Int 1993; 43: 454463.
  • 26
    Raats CJ, Van Den Born J, Berden JH. Glomerular heparan sulfate alterations: mechanisms and relevance for proteinuria. Kidney Int 2000; 57: 385400.
  • 27
    Groffen AJ, Ruegg MA, Dijkman H et al. Agrin is a major heparan sulfate proteoglycan in the human glomerular basement membrane. J Histochem Cytochem 1998; 46: 1927.
  • 28
    Joosten SA, Van Dixhoorn MGA, Borrias MC et al. Antibody response against perlecan and collagen types IV and VI in chronic renal allograft rejection in the rat. Am J Pathol 2002; 160: 13011310.
  • 29
    De Fijter JW, Van Den Wall Bake AW, Braam CA, Van Es LA, Daha MR. Immunoglobulin A subclass measurement in serum and saliva: sensitivity of detection of dimeric IgA2 in ELISA depends on the antibody used. J Immunol Methods 1995; 187: 221232.DOI: 10.1016/0022-1759(95)00188-8
  • 30
    Van den Heuvel LP, Van Den Born J, Van De Velden TJ et al. Isolation and partial characterization of heparan sulphate proteoglycan from the human glomerular basement membrane. Biochem J 1989; 264: 457465.
  • 31
    Seino J, Vd Wall Bake WL, Van Es LA, Daha MR. A novel ELISA assay for the detection of C3 nephritic factor. J Immunol Methods 1993; 159: 221227.DOI: 10.1016/0022-1759(93)90160-9
  • 32
    Van Den Born J, Van den Heuvel LP, Bakker MA, Veerkamp JH, Assmann KJ, Berden JH. Monoclonal antibodies against the protein core and glycosaminoglycan side chain of glomerular basement membrane heparan sulfate proteoglycan: characterization and immunohistological application in human tissues. J Histochem Cytochem 1994; 42: 89102.
  • 33
    Heremans A, Van Der Schueren B, De Cock B et al. Matrix-associated heparan sulfate proteoglycan: core protein-specific monoclonal antibodies decorate the pericellular matrix of connective tissue cells and the stromal side of basement membranes. J Cell Biol 1989; 109(6 Pt 1): 31993211.DOI: 10.1083/jcb.109.6.3199
  • 34
    Van Den Born J, Van den Heuvel LP, Bakker MA, Veerkamp JH, Assmann KJ, Berden JH. Production and characterization of a monoclonal antibody against human glomerular heparan sulfate. Lab Invest 1991; 65: 287297.
  • 35
    Groffen AJ, Veerkamp JH, Monnens LA, Van den Heuvel LP. Recent insights into the structure and functions of heparan sulfate proteoglycans in the human glomerular basement membrane. Nephrol Dial Transplant 1999; 14: 21192129.DOI: 10.1093/ndt/14.9.2119
  • 36
    Iozzo RV. Matrix proteoglycans: from molecular design to cellular function. Annu Rev Biochem 1998; 67: 609652.DOI: 10.1146/annurev.biochem.67.1.609
  • 37
    Groffen AJ, Buskens CA, Van Kuppevelt TH, Veerkamp JH, Monnens LA, Van den Heuvel LP. Primary structure and high expression of human agrin in basement membranes of adult lung and kidney. Eur J Biochem 1998; 254: 123128.DOI: 10.1046/j.1432-1327.1998.2540123.x
  • 38
    O'Toole JJ, Deyst KA, Bowe MA, Nastuk MA, McKechnie BA, Fallon JR. Alternative splicing of agrin regulates its binding to heparin alpha-dystroglycan, and the cell surface. Proc Natl Acad Sci USA 1996; 93: 73697374.DOI: 10.1073/pnas.93.14.7369
  • 39
    Burgess RW, Skarnes WC, Sanes JR. Agrin isoforms with distinct amino termini: differential expression, localization, and function. J Cell Biol 2000; 151: 4152.
  • 40
    Raats CJ, Bakker MA, Hoch W et al. Differential expression of agrin in renal basement membranes as revealed by domain-specific antibodies. J Biol Chem 1998; 273: 1783217838.DOI: 10.1074/jbc.273.28.17832
  • 41
    Hudson BG, Tryggvason K, Sundaramoorthy M, Neilson EG. Alport's syndrome, Goodpasture's syndrome, and type IV collagen. N Engl J Med 2003; 348: 25432556.DOI: 10.1056/NEJMra022296
  • 42
    Borza DB, Hudson BG. Molecular characterization of the target antigens of anti-glomerular basement membrane antibody disease. Springer Semin Immunopathol 2003; 24: 345361.DOI: 10.1007/s00281-002-0103-1
  • 43
    Leinonen A, Netzer KO, Boutaud A, Gunwar S, Hudson BG. Goodpasture antigen: expression of the full-length alpha3(IV) chain of collagen IV and localization of epitopes exclusively to the noncollagenous domain. Kidney Int 1999; 55: 926935.DOI: 10.1046/j.1523-1755.1999.055003926.x
  • 44
    Shu KH, Lu YS, Chang CH et al. Transplant glomerulopathy-a clinicopathological study. Transplant Proc 1996; 28: 15271528.