SEARCH

SEARCH BY CITATION

Keywords:

  • Accommodation;
  • allograft function;
  • antibody-mediated rejection;
  • complement activation;
  • complement component C3;
  • complement C4d;
  • complement regulation;
  • donor-specific antibodies;
  • heart allograft;
  • heart transplantation;
  • immunofluorescence;
  • pathology

Abstract

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

Antibody-mediated rejection (AMR) is an immunopathologic process in which activation of complement often results in allograft injury. This study correlates C4d and C3d with HLA serology and graft function as diagnostic criteria for AMR. Immunofluorescence staining for C4d and C3d was performed on 1511 biopsies from 330 patients as part of routine diagnostic work-up of rejection. Donor-specific antibodies were detected in 95% of those with C4d+C3d+ biopsies versus 35% in the C4d+C3d– group (p = 0.002). Allograft dysfunction was present in 84% in the C4d+ C3d+ group versus 5% in the C4d+C3d− group (p < 0.0001). Combined C4d and C3d positivity had a sensitivity of 100% and specificity of 99% for the pathologic diagnosis of AMR and a mortality of 37%. Since activation of complement does not always result in allograft dysfunction, we correlated the expression pattern of the complement regulators CD55 and CD59 in patients with and without complement deposition. The proportion of patients with CD55 and/or CD59 staining was highest in C4d+C3d− patients without allograft dysfunction (p = 0.03). We conclude that a panel of C4d and C3d is diagnostically more useful than C4d alone in the evaluation of AMR. CD55 and CD59 may play a protective role in patients with evidence of complement activation.


Introduction

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

The diagnosis of acute antibody-mediated rejection (AMR) in cardiac transplantation requires histopathologic, immunologic and clinical correlation. The current recommendation in the 2004 revised working formulation of the International Society for Heart and Lung Transplantation (ISHLT) is histologic screening in every biopsy for features of AMR that include endothelial swelling and intravascular macrophage accumulation. If these features are identified, then one should proceed with pathologic confirmation of AMR by performing immunofluorescence (IF) or immunoperoxidase staining (1). However, the reported sensitivities of endothelial cell activation and swelling at 63% and of intravascular macrophages at 30% in a large retrospective study were low (2).

The evaluation of AMR has traditionally been performed on frozen tissue using antibodies directed against immunoglobulins (IgG, IgM and IgA) and complement components (C3c, C1q). The diagnostic value of antibodies against Ig is questionable because deposition of these markers has been repeatedly shown to correlate poorly with clinical status, outcome or presence of circulating anti-human leukocyte antigens (HLA) antibodies in transplant patients (3–7). The subsequent availability of commercial monoclonal antibodies against complement split products C4d and C3d has proven the usefulness of these markers (6,7). Application of C4d staining alone gained wide acceptance when it was shown to work in formalin-fixed paraffin-embedded tissue (8). To date, consensus does not exist in the methodology of detection of complement split products and their interpretation by pathologists. Furthermore, there are only limited studies that correlate the presence of complement split products in the biopsy with circulating alloantibodies (7,9).

The purpose of this study was to evaluate the diagnostic utility of performing a panel of C4d and C3d in frozen sections of endomyocardial biopsies (EMBs) for the investigation of AMR in a large unbiased cohort of heart transplant recipients. IF patterns of C4d and C3d staining were correlated with hemodynamic status and presence of HLA class I and II donor-specific antibodies (DSA). In addition, expression of the membrane-bound regulators of complement activation (RCA) CD55 and CD59 and their possible role in AMR was studied.

Materials and Methods

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

Patient population

We studied 1511 consecutive EMBs performed between November 2006 and December 2007 at our institution. Protocol surveillance EMBs were obtained weekly for the first month, every 2 weeks for month 2, every 4–6 weeks for months 3–6, every 6–8 weeks for months 7–9, every 8 weeks for months 10–12, every 3 months for months 13–24, every 4 months for months 25–36 and every 6 months for months 37–60. Beyond 5 years, biopsies were not routinely scheduled. Nonprotocol indication EMBs were obtained when there was clinical suspicion of rejection. Follow-up biopsies were performed within 1–3 weeks after episodes of cellular rejection or AMR. Biopsies that were insufficient for evaluation of acute cellular rejection per ISHLT working formulation were excluded from this study. Pediatric transplant patients were not included in this evaluation.

There were 330 adult patients (266 males and 64 females) followed up during this period with age ranging from 20 to 73 years at the time of transplantation. Fifty patients were transplanted <1 year, while 241 patients were within 1 to 5 years posttransplant. The remaining 39 patients were >5 years after cardiac transplantation.

Specimen handling and processing

EMBs consisted of four or more pieces received in the pathology laboratory wrapped in saline-moistened filter paper. All tissue fragments were quickly blotted to remove excess moisture, mounted on a chuck covered with Optimal Cutting Temperature (OCT) compound (Tissue-Tek, Sakura Finetek, Torrance, CA) and submerged in liquid nitrogen until frozen. Cryostat sections were cut in three-step levels for hematoxylin-eosin (H&E) staining. Additional slides were obtained and air-dried for IF studies.

IF staining technique

Slides were fixed in acetone for 10 min, dried and rinsed in phosphate buffered saline (PBS) at pH 7.2 for 5 min. Tissue sections were then incubated for 30 min with mouse monoclonal antibodies to human complement components C4d (dilution 1:50, AbD Serotec, Raleigh, NC) and C3d (dilution 1:20, Quidel, San Diego, CA). Slides were rinsed in PBS twice for 5 min. Fluorescein isothiocyanate (FITC)-labeled affinity purified goat F(ab’)2 anti-mouse IgG (dilution 1:50, Protos Immunoresearch, Burlingame, CA) was used as the secondary antibody with 30-min incubation followed by two 5-min rinses in PBS. Negative controls using mouse primary antibodies (dilution 1:100, Dako, Carpinteria, CA) and appropriate positive control slides for each antibody were run in parallel.

Staining for CD55 and CD59 was done retrospectively from the remaining specimen. Biopsy tissue was not available for 6 of the 330 patients. For the detection of RCA, acetone-fixed slides were rehydrated in 1X PBS-0.5% Tween-20 for 10 min followed by a 30-min incubation with 10% rabbit serum. Sections were incubated overnight at 4°C with goat polyclonal antibodies to CD55 and CD59 (dilution 1:100, Santa Cruz Biotechnology, Santa Cruz, CA). Slides were washed three times in 1X PBS-0.5% Tween-20 for 3 min each and incubated in FITC-conjugated rabbit anti-goat IgG (dilution 1:300, Vector Laboratories, Burlingame, CA) for 30 min at room temperature. Slides were washed three times in 1X PBS-0.5% Tween-20 before coverslipping with Vectashield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA).

Pathologic evaluation

Acute cellular rejection was graded according to the revised ISHLT classification (10). IF intensity was scored semi-quantitatively from 0 to 3+. The localization (capillary or perimyocytic) of C4d and C3d staining was recorded. Diffuse pattern was interpreted as the presence of staining in ≥50% of the myocardium. Focal positivity was defined as staining present in <50% of the myocardium. In the assessment of CD55 and CD59 staining, biopsy was considered positive if there was diffuse granular staining in capillary endothelial cells. Positive biopsies were also categorized semi-quantitatively from 0 to 3+ based on the intensity of granular capillary staining. The staining intensity varied from 0 to 3+ with time and also with occurrence of AMR. Therefore, a mean of the IF score per patient was used to analyze the data. Staining in the connective tissue of the endocardium, arteries or arterioles by complement or by RCA was not included in the analysis. Clinical status of the recipients was unknown at the time of biopsy interpretation.

Immunohistochemical staining for CD68 was performed retrospectively on biopsy material in all cases that were positive for C4d and C3d.

Serologic testing

Panel-reactive antibodies (PRA) were screened by complement-dependent cytotoxicity assay and FlowPRA class I and class II Single Antigen beads (One Lambda, Canoga Park, CA). Sensitization was defined as a PRA value of >10% or the detection of specific HLA IgG antibody pretransplant. Patients with a C4d+ biopsy were tested for HLA class I and II DSA as follows: 27 of 62 were tested because of clinical suspicion and 26 of 62 were tested retrospectively because of C4d+ result. Sera matching the C4d+ biopsy were not available in 9 of 62 patients. DSA testing was performed in blood drawn at or around the time of biopsy by flow cytometry using donor cells and/or beads.

Clinical data

Electronic medical records with follow-up ending in December 2008 were reviewed. Cardiac allograft dysfunction was defined as a decline in left ventricular ejection fraction, decrease in cardiac index, elevation of right-sided cardiac pressures and a need for inotropic support. Institution and response to treatment for AMR were recorded.

Statistical analysis

Intergroup means were compared by analysis of variance. Statistical significance between groups was determined by Fisher's exact test to compare categorical variables. IF data for RCA were compared using Mann–Whitney-U test. A p-value of < 0.05 was considered significant. Sensitivity, specificity, positive and negative predictive values of a panel of C4d and C3d versus C4d alone were determined. Calculations were performed using SPSS 16.0 software (SPSS Inc., Chicago, IL).

Results

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

IF staining patterns of complement split products

An average of five EMBs per patient was examined. One hundred fifty-two (10%) biopsies from 62 patients showed positive staining for complement split products. All but five of these biopsies were graded 0R or 1R for cellular rejection. The 62 patients were subdivided into four groups based on the pattern and greatest extent of complement staining among multiple biopsies in a particular patient (Table 1). Group 1 patients (n = 19) showed diffuse linear capillary staining for both C4d and C3d (Figure 1). This staining pattern was typically observed in at least two consecutive EMBs. Group 2 (n = 19) showed diffuse capillary staining of C4d only. Eight of 19 patients had this result in more than one biopsy and 11 of 19 in only 1 EMB. Group 3 (n = 13) showed focal capillary staining with C4d occurring in only one occasion. Group 4 (n = 11) showed perimyocytic pattern that was almost always diffuse and seen in only 1 EMB in 10 patients and in 2 EMBs in 1 patient (Figure 1). Perimyocytic pattern was observed more commonly with staining for both C4d and C3d (seven patients) than with C4d alone (four patients). Staining for C3d alone was only seen focally in one EMB from a patient who had diffuse capillary staining for C4d on previous occasions. C3d staining was never observed alone in diffuse capillary pattern or in perimyocytic pattern.

Table 1.  Immunofluorescence staining pattern of C4d and C3d in 330 patients
 Diffuse capillaryFocal capillaryDiffuse perimyocytic
C4d and C3d positive19 (6%)07 (2%)
C4d positive only19 (6%)13 (4%)4 (1%)
C3d positive only000
image

Figure 1. Immunofluorescence staining patterns of C4d and C3d. Intense linear staining of capillary endothelium in a diffuse pattern is identical between C4d (A) and C3d (B) during episodes of AMR (×400, FITC anti-C4d and anti-C3d). (C) Diffuse linear capillary deposits of C4d are still visible 1 week after plasmapheresis. In addition, there is linear staining around myocytes in the interstitium (×400, FITC anti-C4d). (D) Perimyocytic pattern shows C4d deposits outlining individual myocytes without capillary staining (×400, FITC anti-C4d).

Download figure to PowerPoint

In sequential biopsies of patients with diffuse capillary C4d+C3d+ deposition who underwent plasmapheresis, there was a decrease in staining intensity after 1 week of therapy. This was commonly accompanied by a transition from capillary to interstitial staining with clinical resolution of AMR (Figure 1). Capillary staining of C3d was cleared within 2 weeks to 1 month. C4d deposits were cleared within 1 to 2 months (Figure 2).

image

Figure 2. AMR course in a patient at 8 months posttransplant. Cardiac index began to drop concurrent with the appearance of C4d and C3d in the biopsy. Testing for donor-specific antibodies (DSA) showed high titers for DR51 expressed as molecules of equivalent soluble fluorochrome. Clearance of alloantibodies was evident after plasmapheresis. C3d was cleared earlier than C4d in the biopsy.

Download figure to PowerPoint

Among the 19 patients in group 1 with diffuse capillary deposition of C4d and C3d, four patients were noted to present initially with C4d capillary staining only. The time interval for progression to C4d+C3d+ pattern ranged from 6 weeks to 6 months.

Of 37 biopsies that were positive for both C4d and C3d, capillary endothelial swelling and intravascular macrophages were identified in 54% and 43% of the biopsies, respectively (Table S1). Capillary endothelial swelling, when present, appeared as a diffuse change and was observed in nine of 16 patients with diffuse capillary C4d+C3d+ biopsies. Intravascular macrophages were scant and present only in a patchy distribution in six of 19 these patients. These changes appear as a function of progression of the AMR episode, being more common in biopsies with 3+ intensity of C4d and C3d.

DSA and allograft dysfunction

Eighteen of 19 (95%) patients in group 1 had detectable DSA. Of these 18 patients, 16 showed allograft dysfunction (Table 2). The only patient with positive complement deposition and no evidence of DSA and allograft dysfunction had received a total artificial heart.

Table 2.  Demographic and clinical data of patients with positive immunofluorescence staining
 Group 1Group 2Group 3Group 4p-Value
C4dDiffusely positive in capillary patternDiffusely positive in capillary patternFocally positive in capillary patternDiffusely positive in perimyocytic pattern 
C3dDiffusely positive in capillary patternNegativeNegativeDiffusely positive in perimyocytic pattern; negative 
No. of patients19191311 
M:F13 M:6 F17 M:2 F 8 M:5 F10 M:1 FNS
Age (mean)51 (27–67)49 (21–65)48 (24–68)52 (23–64)NS
Time of biopsy with complement staining (months–median, range)21 (0.25–157)15 (0.25–56)15 (0.25–117)41 (2–147)NS
Sensitized patients58% (11/19)37% (7/19)54% (7/13)36% (4/11)NS
LVAD15% (3/19)21% (4/19)30% (4/13)45% (5/11)NS
Donor-specific antibody95% (18/19)35% (6/17)56% (5/9) 0% (0/8)p = 0.002 
Allograft dysfunction84% (16/19) 5% (1/19) 0% (0/13) 0% (0/11)p < 0.0001
Mortality42% (8/19) 0% (0/19) 8% (1/13) 9% (1/11) 

DSA was present in 6 of 17 (35%) patients in group 2 and 5 of 9 (56%) of group 3. However, only one patient with diffuse C4d staining and negative DSA complained of fatigue and found to have elevated right-sided pressures.

None of group 4 patients with perimyocytic pattern of complement deposition had detectable DSA. This group of patients had no allograft dysfunction and remained asymptomatic during a mean follow-up period of 20 months after a positive biopsy.

Prevalence and course of AMR

Sixteen of 330 patients had evidence of concomitant complement deposition in the biopsy, allograft dysfunction and presence of DSA (Figure 3). Using these criteria, the prevalence of AMR in this cohort is 5%. A higher proportion of females developed AMR with a prevalence of 8% compared to 4% in males (not statistically significant). The median occurrence of AMR in the whole cohort was at 25 months (range: 1 week–157 months) posttransplant. In sensitized patients, the median occurrence was at 18 months while those who were not sensitized developed AMR at a median of 75 months. Treatment of AMR consisted of high-dose pulse steroid and plasmapheresis in a majority of the patients. A few patients also received intravenous immunoglobulin.

image

Figure 3. Correlation of C4d and C3d staining with donor-specific antibodies (DSA), allograft dysfunction and survival.*Sera were not available for DSA testing in two patients of this group. **This patient died of glioblastoma multiforme. ***This patient died of coronary thrombosis.

Download figure to PowerPoint

Prior to this study period, surveillance of AMR was performed only in the first month posttransplant and when clinically indicated. In the group of patients whose biopsies were negative for C4d and C3d during this study period (n = 268), four patients had one episode of ‘vascular rejection’ using the ISHLT 1990 criteria within the first 3 months and one patient at 115 months posttransplant. There was no recurrence of AMR after a single episode. A sensitized female patient in group 1 had her first episode of AMR at 11 months and a second episode occurred within the study period at 21 months. Last, a highly sensitized group 2 patient also had AMR within the first 3 months that did not recur.

Eight of 19 patients in group 1 died within the observation period (Figure 3). AMR was the immediate cause of death in three patients and was the underlying cause of death in another three patients in this group. Deaths unrelated to AMR were due to glioblastoma multiforme and coronary thrombosis. There were no deaths in the group 2 patients. One patient in group 3 died of acute cellular rejection. Another patient died of sepsis and multiorgan failure in group 4.

Comparisons between groups

No differences were found between the groups in mean age, sex distribution, indication for transplantation, time of posttransplant biopsy with positive complement deposition, frequency of sensitization and use of ventricular assist devices prior to transplantation (Table 2).

Since none of the patients with perimyocytic staining pattern were found to have DSA or cardiac allograft dysfunction, only patients with capillary pattern of staining were included in subsequent analyses. The frequency of detection of DSA or allograft dysfunction did not differ between patients with diffuse (group 2) or focal (group 3) C4d capillary staining. The association of DSA and allograft dysfunction with diffuse C4d+C3d+ capillary staining (group 1) was compared to those with C4d capillary staining only (groups 2 and 3) and found to be statistically significant (p = 0.002 and p ≤ 0.0001, respectively).

C4d and C3d as diagnostic tests

In this study, combined positivity for diffuse capillary C4d and C3d was more specific than diffuse C4d staining alone (99% vs. 93%) for the diagnosis of AMR as defined by the presence of both DSA and allograft dysfunction (Table 3). More important, the predictive value of a positive C4d and C3d staining was 84% compared to 41% if C4d staining alone was employed.

Table 3.  Diagnostic test analysis of a panel of C4d and C3d versus C4d alone
 C4d and C3dC4d alone
Sensitivity100%  100%  
Specificity99.00%92.99%
Positive predictive value84.21%42.11%
Negative predictive value99.68%100%  
Positive likelihood ratio104.33  14.27 
Negative likelihood ratio00

Regulators of complement activation

EMBs from 264 out of 330 patients without allograft dysfunction and without evidence of complement staining served as the control group. Within this group, 72% (191 of 264) had staining with CD55 (Figure 4) and/or CD59 and 28% (73 of 264) showed absent staining with both CD55 and CD59 (Table 4). The proportion of control patients with and without RCA expression was compared with those of patients in groups 1 to 4. A statistically significant difference was observed in group 2 in which 95% (18 of 19) of patients had RCA expression (p = 0.03). In serial biopsies, staining of CD55 and CD59 was noted to fluctuate from 0 to 3+ in relation to time from transplant and to the presence of AMR. Therefore, staining of CD55 and CD59 was further analyzed by calculating a mean IF intensity score from multiple biopsies per patient. Within the control group, a trend of increasing CD55 and CD59 expression over time as represented by the average IF intensity score was observed (Figure 5). There was no significant difference in average CD55 score between patients transplanted within 1 year, 1 to 5 years and >5 years. However, the difference in average CD59 scores between groups within the control patients was statistically significant and suggestive of sustained CD59 expression over time. The mean scores of CD55 and CD59 in groups 1 to 4 were not significantly different from each other. Interestingly, during episodes of AMR, it was noted that 65% of group 1 patients showed a transient decrease or loss of CD55 and CD59 staining.

image

Figure 4. Indirect immunofluorescence stain with antibody for regulator of complement activation CD55 or decay accelerating factor. (A–C) Negative biopsy. (A) Nucleic acid stain of myocyte and endothelial cell nuclei (×400, DAPI). (B) There is a complete lack of immunofluorescence of myocardial capillaries (×400, FITC anti-CD55). (C) Overlapping images of A and B (×400, DAPI and FITC anti-CD55). (D–F) Positive anti-CD55 stain in myocardial capillaries (1+ intensity). (D) Nucleic acid stain of myocyte and endothelial cell nuclei (×400, DAPI). (E) Positive granular staining pattern of myocardial capillary endothelial cells in longitudinal orientation. Scored 1+ (×400, FITC anti-CD55). (F) Overlapping images of D and E showing endothelial cell nuclei in the capillaries with granular staining for CD55 (×400, DAPI and FITC anti-CD55). (G–I) Positive anti-CD55 stain in myocardial capillaries (3+ intensity). (G) Nucleic acid stain of myocyte and endothelial cell nuclei (×400, DAPI). (H) Positive granular staining pattern of myocardial capillary endothelial cells in oblique and cross-section orientation. Scored 3+ (×400, FITC anti-CD55). (I) Overlapping images of G and H showing endothelial cell nuclei in the capillaries with granular staining for CD55 (×400, DAPI and FITC anti-CD55). (J–L) Positive anti-CD55 stain of the cytoplasmic membrane of lymphocytes in a Quilty (type A) endocardial infiltrate. (J) Nucleic acid stain of a cluster of lymphocytes (×1000, DAPI). (K) Linear immunofluorescent signal is seen that highlights the contour of the cytoplasm of the lymphocytes present in the Quilty infiltrate (×1000, FITC anti-CD55). (L) Overlapping images of J and K showing the lymphocyte nuclei in blue and the cytoplasmic membrane staining with anti-CD55 (×1000, DAPI and FITC anti-CD55).

Download figure to PowerPoint

Table 4.  Immunofluorescence staining result of CD55 and CD59
 CD55 stainingCD59 stainingCD55 and CD59 stainingAbsent CD55 and CD59 stainingAverage CD55 scoreAverage CD59 score
  1. 1The proportion of patients with CD55 and/or CD59 staining is highest in the group 2 (i.e. patients who are C4d+C3d– on biopsy and without allograft dysfunction) compared to control group (patients with no complement in their biopsies) (p = 0.0309). There was no statistically significant difference between the control group and groups 1, 3 and 4.

  2. 2There were statistically significant differences in CD59 expression in the control group as a function of time (<1 year vs. 1–5 years, p = 0.0318; <1 year vs. >5 years, p < 0.0001; 1–5 years vs. >5 years, p = 0.0001).

Control group (n = 264)107 (41%)9 (3%)75 (28%)73 (28%)0.85520.2439
<1-year posttransplant (n = 40) 24 (60%)010 (25%) 6 (15%)0.6345 0.05242
1 to 5 years posttransplant (n = 192) 78 (40%)1 (1%)53 (28%)60 (31%)0.8888 0.22032
> 5 years posttransplant (n = 32)  5 (15%) 8 (25%)12 (38%) 7 (22%)0.92970.6252
Group 1 (n = 17)  7 (41%)0 8 (47%) 2 (12%)0.87990.2376
Group 2 (n = 19)   8 (42%)10 10 (53%)11 (5%)1.176 0.1678
Group 3 (n = 13)  6 (46%)0 6 (46%)1 (8%)0.84850.2777
Group 4 (n = 11)  6 (55%)0 2 (18%) 3 (27%)1.23910.2727
image

Figure 5. Average immunofluorescence intensity scores for CD55 and CD59 in the control group (left) and patients with evidence of complement activation. Group 1 represents C4d+C3+ biopsies, group 2 are diffuse C4d+C3d− biopsies, group 3 are focal C4d+C3d− biopsies and group 4 are biopsies with perimyocytic staining pattern.

Download figure to PowerPoint

Discussion

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

The lack of uniformity in the approach to diagnostic testing for AMR impacts on the reporting of its incidence, timeliness of diagnosis and institution of appropriate treatment. The histologic features suggestive of AMR have low sensitivity and specificity to recognize early rejection in EMBs (2). Similarly, we find that capillary endothelial swelling and intravascular macrophages occurred in only about half of the patients with AMR (Table S1). However, there is no published temporal characterization of the appearance of endothelial changes and intravascular macrophages as markers for AMR. In our data, they seem to appear once AMR and injury to the graft are well established. The more severe cases with neutrophilic infiltration and interstitial hemorrhage and edema are now rarely seen in the clinical setting. More important, there are no specific histologic findings that are predictive of abnormal hemodynamic status (11,12). This study demonstrates a more reliable method of diagnostic screening and work-up for AMR in biopsies by the detection of complement split products in capillaries.

The markers used in the early 1990s have been replaced by the more abundant and stable C4d and C3d fragments. During the activation of complement via the classical pathway, there is an amplification factor of 28 and 241 for the number of bound C4b and C3b molecules per C1q molecule, respectively (13). C4b and C3b are rapidly degraded but the cleaved products, C4d and C3d, remain covalently bound to the membrane. For these reasons, it could be expected that detection of C4d and C3d deposition would be more sensitive than C1q.

C4d became well established as a marker of AMR in clinical transplantation as it can be reliably detected in formalin-fixed tissue, thus obviating the need for an extra piece of tissue to be frozen. Consequently, investigation of C3d in cardiac transplantation has received less attention because it has not been systematically evaluated in formalin-fixed tissue. In our institution, evaluation of acute cellular rejection and AMR is routinely carried out on frozen sections without compromise in the quality of the tissue for interpretation (14), thus allowing same-day interpretation. Therefore, we have tailored our screening for AMR by using a panel of C4d and C3d.

Our study differs from previously reported large series of biopsies because we comply with the requirement for the presence of both allograft dysfunction and DSA in the diagnosis of AMR (1). In this cohort of 330 patients studied at various time points posttransplant, the prevalence of AMR was 5%. This is slightly higher than the 3% prevalence in a population of unsensitized patients (6) and is comparable with earlier reported prevalence of AMR in studies that evaluated graft dysfunction (12,15,16). Our data further support previous observations that AMR more commonly occurs late in the course of heart transplantation and this should be considered in the formulation of institutional screening algorithms (5,6).

Our results show that detection of C3d along with C4d in cardiac allografts significantly enhances the diagnostic utility of testing for complement activation products. In 38 patients with diffuse linear capillary deposition of C4d, half of these patients had concomitant staining with C3d (Figure 3). Evidence of C3d deposition was associated with allograft dysfunction in 84% and presence of circulating antibodies to donor HLA in 95% of the patients. Sixteen patients fulfilled the clinicopathologic criteria for AMR. In contrast, none of the patients with C4d staining only had concurrent allograft dysfunction and circulating DSA. More important, the clinical outcome in the C4d+C3d+ group was significantly worse (58% alive at 2 years) than those with C4d+C3d− biopsies (100% alive at 2 years). Although C4d is considered a very sensitive marker for the diagnosis of AMR, a positive stain cannot be equated to AMR. As shown in Table 3, C4d+ staining alone has a positive predictive value of only 42%. The addition of C3d staining results in significant improvement of the positive predictive value to 84% when both markers are positive.

The artifacts of C3d IF staining are comparable to C4d in that staining of ischemic myocytes, endocardium and internal elastic lamina of arteries and arterioles was observed. C4d was cleared in the cardiac allograft within months similar to the experience in renal transplants (17). C3d was cleared from the biopsies 2 weeks to a month earlier than C4d for unknown mechanisms. Loss of bound complement split products on the endothelial cell surface may be accounted for by mechanisms analogous to shedding of antibodies and clearance of the membrane attack complex by exocytosis or endocytosis (18,19). One possible explanation for the apparent slower clearance of C4d is that regulators of complement activation effectively halt downstream activation of C3 but allow for low-level ongoing C4d deposition.

A shift from vascular to interstitial staining of complement components is usually observed with resolution of AMR. However, isolated interstitial perimyocytic pattern of staining was occasionally observed in 11 patients without prior episodes of AMR or identifiable predisposing events that could lead to complement activation. In a majority of patients, only a single biopsy specimen was affected. The significance of this staining pattern is unclear because it was not associated with circulating DSA or graft dysfunction at the time of biopsy or on follow-up.

In the course of routine transplant surveillance, 31 clinically stable patients were found to have C4d staining with diffuse (18) or focal (13) positivity. DSA testing showed that only 42% of these patients were positive for circulating anti-HLA antibodies; the frequency did not differ significantly between the diffuse and focal C4d groups. Non-HLA antibodies that were not evaluated in this study could possibly trigger complement activation in some cases (20). Failure to detect circulating DSA may also be due to absorption of antibodies by the graft, formation of immune complex of anti-HLA antibodies with soluble HLA antigens in circulation and development of antiidiotypic antibodies (21,22).

The correlation of C4d IF staining and alloantibodies has been investigated in an earlier study of 80 biopsies from 38 patients (9). C4d was positive in 84% of biopsies with alloantibodies and 12% without alloantibodies. Two of nine patients with C4d+ biopsies and alloantibodies were reportedly asymptomatic in that study. A similar phenomenon of C4d deposition without evidence of rejection in renal transplantation was acknowledged in the 2007 Banff Conference (23).

Since C4d deposition occurs without allograft dysfunction in some instances, this suggests that some protective mechanisms exist that prevent further activation of the complement cascade. Accommodation is a condition in which there are antibodies bound to target antigens in the graft but clinical dysfunction of the allograft is not evident (24). Accommodation has been extensively documented in experimental models and this acquired resistance to immune-mediated injury may, in part, be due to upregulation of membrane-bound RCAs (25,26). CD55 or decay-accelerating factor inhibits C3 convertase, thereby preventing cleavage of C3 to C3b and reducing or ablating the deposition of detectable C3d. In a pilot study, we found no detectable tissue expression of CD35 and CD46 in EMBs from patients with and without AMR, whereas CD55 and CD59 were more readily detectable by IF (27). In the current coordinated study of both complement and two of its regulators, the expression levels of CD55 and CD59 appeared to increase over time in 264 control patients. This observation suggests that there may be a protective effect of CD55 and CD59 even if there is no activation or low-level activation of complement. Furthermore, we also observed that patients with AMR, as defined in clinicopathologic terms (1), lacked expression of these markers during the acute episode. However, the correlation of intensity of detectable membrane-bound RCAs in myocardial capillaries by IF to molar amount of bound RCAs has not been reported in AMR. This type of quantitative validation of mRNA expression of RCAs in capillaries would be difficult to interpret, since RCAs are constitutively expressed in lymphocytes (in Quilty lesions and in cellular rejection) (Figure 4) (28). Further studies to elucidate the function of other RCAs including soluble factors in the plasma and how their expression in tissue or plasma is regulated are needed.

In summary, our data indicate that concomitant use of C4d and C3d have a strong positive predictive value for the presence of circulating DSA and allograft dysfunction in heart transplant patients.

Together, these markers are more helpful in establishing the diagnosis of AMR. Routine screening of surveillance biopsies revealed that C4d+C3d− biopsies, even with the presence of alloantibodies, were not associated with allograft dysfunction. Thus, it cannot be assumed that activation of C4 alone indicates clinically significant AMR. The long-term significance of C4d deposition without acute allograft dysfunction is uncertain and graft outcome in these patients need to be assessed.

Acknowledgment

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

The assistance of Lynne Klingman is gratefully acknowledged.

Funding source: PO1-HL-50691 NHLBI.

References

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

Supporting Information

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

Table S1: Correlation of histologic parameters with positive capillary staining for complement split products

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
AJT_2748_sm_TableS1.doc67KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.