Deposition of C4d and C3d in cardiac transplants: A factor in the development of coronary artery vasculopathy

Methods 

Patient population and study design 

Fifty-two patients evaluated from 1991 to 1997 were identified for this study. The CAV group consisted of 26 patients with irreversible CAV, as defined by angiography within 4 years of transplantation (age 49 ± 11 years, 4 females). A control group of patients (n = 26) were matched to cases with at least 8 years of CAV-free angiography (age 49 ± 8.5 years, 2 females), at an individual level, by transplant date to ensure that changes over time did not confound the comparison. In addition, at a group level, cases and controls were matched for pre-transplant diagnosis, age and cytomegalovirus (CMV) status. Despite this, in the CAV-positive group, more patients had a pre-transplant diagnosis of ischemic heart disease, an organ from an older donor, and a slightly longer ischemic time at transplantation, although the differences were not statistically significant (Table 1). However, case–control pairs did not differ with regard to gender or CMV status of both recipients and donors (Table 1).

Table 1. Summary of Case–Control Pairs

		No CAV (n = 26)	CAV (n = 26)	p-value
	Recipient gender (male)	22(85%)	24(92%)	0.668
	Donor gender (male)	19(73%)	22(85%)	0.499
	Mean (SD) recipient age (y)	49(11)	49(9)	1.000
	Mean (SD) donor age (y)	31(12)	49(9)	0.056
	Recipient CMV-positive	15(58%)	15(58%)	1.000
	Donor CMV-positive	9(35%)	13(50%)	0.400
	Pre-Tx diagnosis IHD	13(50%)	19(73%)	0.153
	Mean (SD) ischemia time (min)	185(64)	196(67)	0.520

Pairs were matched in respect to recipient gender and age, donor gender, CMV status of donor and recipient, pre-transplant diagnosis and ischemic times. CAV, coronary artery vasculopathy; CMV, cytomegalovirus; IHD, ischemic heart disease, SD, standard deviation; Tx, transplant.

All patients were on a triple immunosuppressive regimen consisting of cyclosporine, steroids and azathioprine.

Grading and statistical analysis 

Two independent observers used a semi-quantitative system to grade deposition of the complement markers. Intensity of the deposition on the endothelium and the distribution throughout the tissue was graded 0 (absent), 1 (weak staining in a few areas), 2 (moderate staining in several areas) or 3 (strong staining throughout the tissue), as shown in Figure 1 (and described by Chantrauwat et al20). Biopsies were scored as positive for staining on the endothelium of intramyocyte capillaries and small arterioles. Serum staining within vessels and diffuse background interstitial staining in the myocardium were not scored. To confirm that this staining was specific we stained serial sections with the endothelial-specific marker CD31 to ensure that the complement staining seen and graded was due to complement deposition on endothelial-lined channels.

View Large Image Download to PowerPoint

Figure 1. (A) Immunohistochemistry (IHC) staining of C3d on post-transplant myocardial biopsies. A biopsy with Grade 1 deposition. Weak staining, on a few of the larger vessels, not found ubiquitously throughout the tissue (original magnification ×100). (B) Grade 1 deposition at a higher magnification. Very few of the intramyocyte vessels have C3d deposition (original magnification ×400). (C) Grade 3 C3d deposition. Throughout the tissue there is strong staining on intramyocyte vessels. There is some serum staining in the tissue, but only deposition on the endothelium was scored on the grading system. Larger capillaries also have C3d deposits (original magnification ×100). (D) Grade 3 deposition at a higher magnification. All intramyocyte vessels have C3d deposition on the endothelium (original magnification ×400).

Exploratory analyses showed that there were very few negative or Grade 3 biopsies for the two complement markers. Thus, biopsy results were categorized as 0 to 1 or 2 to 3 for subsequent analysis. Initially, the CAV and no-CAV groups were compared for complement deposition across the time period, with the pooled data providing a group analysis. Similarly, the overall incidence of Grade ≥3 acute rejection in each group was compared. Finally, logistic regression models were fitted to assess the effect of C3d and C4d deposition on subsequent development of CAV, with time post-transplant and CAV group included as fixed effects.

Results 

In total, 312 biopsies should have been collated and studied, but 30 were either missing from the records or considered inadequate for grading, resulting in a total of 282 biopsies.

In keeping with previous studies, complement deposition was seen in small capillaries. Complement deposition was found, at varying levels, in almost all biopsies at all time-points in both groups, with few completely negative biopsies for complement deposition (C4d, 4% [12 of 282]; C3d, 5% [15 of 280]) (Table 2, Table 3).

Table 2. Summary of C4d Expression Results for CAV-positive (n = 134) and CAV-negative (n = 148) Groups Across 2 Years

	Time-point	No CAV (grade)				CAV (grade)
		0	1	2	3	0	1	2	3
	1 week	1	14	10	1	1	13	5	1
	4 weeks	1	14	11	0	1	12	9	2
	8 weeks	1	15	9	0	1	11	10	2
	6 months	2	11	11	0	0	13	10	2
	1 year	1	16	6	1	0	11	12	0
	2 years	2	16	5	0	1	9	7	1
	Total	8	86	52	2	4	69	53	8

Intensity was graded from 0 (absent) to 3 (intense and diffuse staining). Results are grouped by biopsy time-points, separated according to number at each grade of intensity. Biopsies of Grade ≥2 were considered for statistical analysis. No statistical difference was detected (p = 0.155). CAV, coronary artery vasculopathy.

Table 3. Summary of C3d Expression Results of CAV-positive (n = 143) and CAV-negative (n = 137) Groups Across 2 Years

	Time-point	No CAV (grade)				CAV (grade)
		0	1	2	3	0	1	2	3
	1 week	1	9	14	1	2	7	9	6
	4 weeks	1	11	10	3	0	13	10	0
	8 weeks	0	13	8	2	0	10	11	2
	6 months	2	11	10	0	3	3	17	2
	1 year	1	14	8	2	2	9	12	1
	2 years	2	12	6	2	1	7	8	2
	Total	7	70	56	10	8	49	67	13

Biopsies of Grade ≥2 were considered for statistical analysis. There is a statistical difference for C3d Grade ≥2 biopsies between the CAV-negative and CAV-positive groups (p = 0.043). CAV, coronary artery vasculopathy.

Over the first 2 years, the patients who were CAV-free had 54 of 148 (36%) biopsies with a C4d score of 2 or 3 and were compared with CAV patients who had 61 of 134 (46%) biopsies with a C4d score of 2 or 3 (Fisher's exact test, p = 0.155; Table 2). The CAV-free group had an average of 39% (49 of 125) of biopsies scored 2 or 3 for deposition of C4d across the first year, which dropped to 22% (5 of 23) at the 2-year time-point. From logistic regression models there was no increase in the chance of getting a C4d Grade ≥2 biopsy in the CAV group (odds ratio = 1.43, 95% confidence interval 0.87 to 2.36, p = 0.141).

With C3d deposition over the first 2 years, the patients who were CAV-free had 66 of 143 (46%) biopsies with C3d Grade ≥2, compared with CAV patients who had 80 of 137 (58%) with Grade ≥2 biopsies (Fisher's exact test, p = 0.043). Thus, there was some evidence of an increase in C3d Grade ≥2 biopsies among the CAV patients (Table 3). The number of C3d-positive (Grade ≥2) biopsies in the CAV-negative group gradually decreased across the 2-year sampling period from 60% (15 of 25) at 1 week to 36% (8 of 22) at 2-year data collection. Logistic regression did not show a significant change in C3d complement deposition over time, but there was a significant difference between the two groups at the 5% level in C3d Grade ≥2 biopsies (odds ratio = 1.63, 95% confidence interval 1.01 to 2.64, p = 0.044).

When looking at dual positivity (Grade ≥2 for both C4d and C3d) we found that 25 biopsies in the CAV-negative group were dual positives (biopsies from 13 patients). Dual-negative biopsies accounted for 43 biopsies in this group (Grade <2 for both C4d and C3d). In the CAV-positive group, 39 biopsies were dual positives (from 17 patients) and 33 were dual negatives. Among all of the dual-positive biopsies, only 3 had other positive features of antibody-mediated rejection (AMR), such as endothelial swelling.

Our investigation was a retrospective study of archived biopsies that were not previously assessed for features of AMR. We therefore analyzed all biopsies in both groups for histologic features of AMR, as defined by the ISHLT criteria,10 with edema, mild diffuse endothelial infiltrate, intravascular macrophages confirmed with macrophage marker CD68 (Figure 2), and swollen cells. On review, using these parameters, 14 of 280 (5%) biopsies were considered possibly positive for AMR. No parallel serum samples were available to test for the presence of specific antibodies. Eleven patients had one morphologic AMR event and 1 patient had three events in the first 2 years. Six of these patients were from the CAV-negative group and 6 were from the CAV-positive group, with the patient with multiple events from the CAV-negative group. Therefore, no difference in the early incidence of AMR and the later development of CAV was found in this study. Comparison of AMR and complement deposition demonstrated that, of the 14 biopsies with features of AMR, only 9 had a positive complement deposition score (Grade ≥2) and 67% (6 of 9) of these occurred at ≥8 weeks post-transplant. The patient who had three AMR-positive biopsies (at 4 weeks, 8 weeks and 1 year) was from the CAV-positive group. When compared with the complement deposition grades, only one of these three biopsies (at 1 year) had a positive C4d/C3d score, but these numbers were too low for formal statistical testing.

View Large Image Download to PowerPoint

Figure 2. IHC staining for the macrophage marker CD68 (original magnification ×400). This marker was used to demonstrate intravascular macrophages in biopsies that had been selected from routine hematoxylin–eosin staining as positive for AMR.

We also assessed each group for episodes of treatable (Grade ≥3A) acute rejection, as defined by the ISHLT's 1990 classification.19 Data are summarized in Table 4. Both CAV-negative and CAV-positive groups had 19 biopsies (14% and 15% of biopsies, respectively) with Grade ≥3A acute rejection across the 2-year time period. There was no relationship between high complement deposition in biopsies and features of acute rejection.

Table 4. Summary of Treatable Acute Rejection Grades (Grade ≥3a, ISHLT Classification 1990)

	Time of biopsy	No CAV		CAV
		Grade <3a	Grade ≥3a	Grade <3a	Grade ≥3a
	1 week	23	3(13%)	21	1(5%)
	4 weeks	19	6(32%)	23	3(13%)
	8 weeks	20	5(25%)	20	6(30%)
	6 months	24	2(8%)	23	3(13%)
	1 year	23	2(9%)	20	5(25%)
	2 years	23	1(4%)	19	1(5%)
	Total	132	19(14%)	126	19(15%)

This table compares the group with no coronary artery vasculopathy (CAV) at 8 years post-operation to the group that was CAV-positive at 4 years post-operation. CAV-negative biopsies: n = 151; CAV-positive biopsies: n = 145.

Discussion 

This study was designed to investigate the role of complement deposition in the first 2 years after cardiac transplantation and its relationship to the later development of CAV. By comparing a CAV-positive group and a CAV-negative group we aimed to ascertain any role for complement activation and whether it was a predictor of future development of graft vasculopathy. This study employed a relatively small sample group, which limits the power to assess even moderate effects. However, despite these limitations, we detected a significantly increased chance of a C3d-positive biopsy in the CAV-positive group. These data support a role for complement activation in the pathogenesis of CAV, which raises further questions about the diagnostic application of C3d as a marker for AMR.

The most noteworthy and surprising observation was the constant deposition of C4d and C3d throughout the 2-year period. Although expression levels varied from very low to high, both within an individual patient's biopsy set as well as across the groups, only a few biopsies (C4d, 4% [12 of 282]; C3d, 5% [15 of 280]) were completely free of complement deposition. Our findings coincide with those from a recent study by Fedson et al21 who also demonstrated that complement deposition occurred through the first 18 months post-transplant, with very few (13%) completely negative biopsies. Given our findings and those of Rodriguez et al,11 this led us to establish a level of significance for our observations based on a 0 to 3 grading system, where Grade ≥2 was considered significant. This decision was based in part on a previous study demonstrating that assessment of C4d, when expressed with Grade ≥2 intensity, is sensitive enough to support a diagnosis of humoral rejection in areas without acute cellular rejection or Quilty lesions.11

Using this grading system, we were able to determine a significant association between C3d deposition and the development of CAV, but not C4d, which is an accepted marker of AMR. The observation of continuous complement deposition, albeit at low levels in some biopsies, could suggest that persistent complement activation might result in a continued insult to or activation of the endothelial lining of the vessels within the heart. This finding, along with acute rejection episodes, suggests a challenging period of immune attack on the donor organ within the first 2 years, as studied in this investigation. Complement deposition is a feature of ischemic reperfusion injury (IRI), and is thought to play a key role in its pathogenesis.22 Interestingly, in this study, both groups had similar mean ischemic times, so this risk factor was not studied further in this experiment. Importantly, the presence of complement deposition in late biopsies, out to 2 years, would suggest that complement activation and deposition at these time-points would be independent of IRI.

Complement markers C4d and C3d are widely accepted and used as markers of AMR in transplantation. Many renal studies have utilized these markers to demonstrate a link between AMR and the development of chronic rejection.23, 24 The cardiac transplantation study by Poelzl et al18 further supports the contention that C4d deposition and AMR is involved in the development of CAV. However, given our findings and the results of others18, 21 there appears to be a state of consistent complement activation and deposition in the graft. This makes interpretation of biopsies with complement deposition problematic in terms of diagnosing AMR. Although C4d complement activation occurs in the absence of other features of AMR, it does not appear to relate directly to the development of CAV in the larger study presented herein, whereas C3d does have a correlation. The development of detectable CAV can occur years after transplant, so persistent levels of complement may not be a direct cause, but may be a factor, in the progression of the disease. Furthermore, investigation of complement activation within the capillaries of the intramyocardium may not best represent the processes occurring within the coronary arteries of the heart that are actually affected by the development of CAV. Therefore, antibody/antigenic differences in endothelial- and smooth-muscle-cell phenotypes, which constitute the majority of the cell types associated with CAV, may differ. Thus, the degree of complement activation and complement-mediated damage may differ between coronary and myocardial sites based on antibody affinity.

A limitation of our study is the lack of serum samples for investigating the presence of circulating antibodies and complement fragments. This study was performed retrospectively and utilized tissue collected over a 10-year period and, unfortunately, serum samples were unavailable for comparison. Several groups have reported the presence of anti-graft antibodies post-transplantation, with the incidence and titers varying extensively between different patient populations.25, 26 Hosenpud et al27 demonstrated that most patients develop IgM anti-graft antibodies, albeit at low titers. In their in vitro culture model utilizing patients' donor-specific endothelial cells, they were able to show that all of patients studied generated anti-graft IgM antibodies. Furthermore, they showed that, although these antibodies bound to donor-specific endothelium, and thus would result in complement deposition in vivo, they did not induce cytotoxicity. Given these data, it is perhaps unsurprising that complement activation was seen in almost all biopsies analyzed, presumably a result associated with anti-graft antibody deposition. Another study limitation was using archival routine acquired myocardial biopsies, where, on occasion, the remaining tissue was inadequate for the entire panel of staining and grading. This resulted in slightly different biopsy numbers for the comparisons, although this was taken in to account in the statistical analysis.

Comparison of AMR grading to complement deposition demonstrated that only 5% (14 of 280) of biopsies had features of AMR compared with 41% (115 of 282) with a positive complement deposition grade, indicating a need for caution when interpreting the deposition of complement within the first 2 years post-transplantation. Dual-positive biopsies were found, but these did not always reflect AMR grade, as only three of these incidences coincided with other features of AMR. The persistence of complement activation and the reported presence of anti-graft IgM antibodies intuitively suggest that complement activation occurs via the classical pathway. However, given the higher prevalence of C3d at levels associated with significant deposition (Grade ≥2), it is possible that complement activation ensues by activation of the alternative pathway independently of anti-graft antibodies. An alternative explanation may be that the increased deposition of C3d may simply reflect that C3 is the convergence point of all three pathways and that increased C3d deposition relative to C4d may be associated with the amplification loop of complement activation. Regardless of the route of activation, these data, together with those of other investigators,11, 15 support the need to include C3d as a marker of complement activation in the diagnosis of AMR. With several groups investigating the relationship with complement and AMR, it is increasingly difficult to compare our results with those of other studies due to the lack of standardized techniques and interpretation of results.

In conclusion, we have shown that complement deposition is common throughout the first 2 years post-transplantation and was even detected in the absence of histologic features of AMR. Our study demonstrates that early C3d deposition can be linked to later development of CAV. A role for low-level complement causing ongoing damage cannot be excluded, but it was not shown to be an identifiable risk factor in this study. A larger study may allow for more statistically powerful analyses. Due to the long time-frame of this retrospective study, with the CAV-negative group being angiographically free from disease at 8 years, we were unable to assess patients on any of the newer anti-proliferation immunosuppressive drugs. This may have included a slightly different development group, because, although the incidence of acute rejection has been shown to lower the rate of CAV, disease development remains at high levels.