Complement Independent Antibody-Mediated Endarteritis and Transplant Arteriopathy in Mice

Abstract

Complement fixation, as evidenced by C4d in the microvasculature, is a widely accepted criterion of antibody-mediated rejection. Complement fixation has been shown to be essential in acute antibody-mediated rejection, but its role in chronic rejection has not been addressed. Previous studies showed that passive transfer of complement fixing monoclonal IgG2a anti-H-2K^k into B6.RAG1^−/− KO recipients of B10.BR hearts led to progressive chronic transplant arteriopathy (CTA) over 14–28 days, accompanied by C4d deposition. The present studies were designed to test whether complement was required for these lesions. We report that a noncomplement fixing donor-specific alloantibody (DSA, monoclonal IgG1 anti-H-2K^k) injected into B6.RAG1^−/− KO recipients of B10.BR hearts also promotes CTA, without C4d deposition. Furthermore, a passive transfer of DSA (monoclonal IgG2a anti-H-2K^k) initiated endarteritis followed by CTA in B6.RAG1^−/− mice genetically deficient in the third component of complement (RAG1^−/−C3^−/−). These studies indicate that antibody to class I MHC antigens can trigger chronic arterial lesions in vivo without complement participation, in contrast to acute antibody-mediated rejection. This pathway may be relevant to C4d-negative chronic rejection sometimes observed in patients with DSA, and argues that lack of C4d deposition does not exclude antibody-mediated chronic rejection.

Introduction

Chronic transplant arteriopathy (CTA) is a major cause of late allograft loss after heart or kidney transplantation (1–3). Studies done in a small number of patients in our institution over 30 years ago showed that CTA occurred primarily in patients who developed donor-specific antibodies (DSAs) (4). Multicenter prospective trials have since shown that an adverse effect of circulating anti-MHC antibodies on long-term graft survival is clearly demonstrable (5,6). Furthermore, pathological studies of kidney transplants have shown that lesions of chronic rejection are associated with circulating antidonor HLA antibodies (7–11) and with the deposition of the C4d in peritubular capillaries in the transplants in about 50% of the recipients. Recent studies of heart transplants have also reported an association between C4d deposition in myocardial capillaries and graft loss (12,13), circulating DSA (14) and sometimes CTA (15). However, in all series a substantial fraction of the patients with chronic rejection had DSA without deposition of C4d. Clinical studies investigating the role of DSA on the development of CTA are necessarily observational, but have motivated experimental studies to determine the mechanisms.

Recent evidence from experimental studies in animals make it clear that DSA can mediate CTA (16–18). We showed that complement fixing DSA to class I MHC antigen is sufficient to cause CTA in cardiac allografts in immunodeficient mice (RAG1^−/−) and was associated with C4d deposition in the capillaries (18). Complement fixation is necessary for antibody-mediated acute rejection of allo- and xenografts (19–23), but its requirement in chronic lesions has not been tested. Because alloantibody can cause activation and proliferation of cultured endothelial cells without complement or leukocytes (24), we postulated that the arterial lesions might arise independent of complement activation and we have tested this hypothesis in mouse heart allografts using noncomplement fixing alloantibodies and in recipients deficient in C3.

Materials and Methods

Mice

C57BL/6 (H-2b), B10.BR (H-2^k) and B6.129S7-Rag1^tm1Mom (B6.RAG1^−/−, H-2^b) mice aged 5–7 weeks were purchased from The Jackson Laboratory (Bar Harbor, ME). C3 deficient B6.RAG1^−/−(B6.RAG1^−/−C3^−/−) double knockout (DKO) male mice were kindly provided by Dr. Michael Carroll, Harvard Medical School. B6. RAG1^−/−C3^−/− male mice were bred with C3 deficient RAG1^+/+ mice (Jackson Laboratories). Female F1 mice (C3^−/− RAG1^−/+) mice were backcrossed with C3^−/−RAG1^−/− males. DKOs for further breeding were selected for the absence of circulating T cells using flow cytometry. C3 deficiency was confirmed by immunohistochemistry of the spleens. All mice were maintained under pathogen-free conditions in filter-top cages throughout the experiments with an automatic water system and were cared for according to methods approved by the American Association for the Accreditation of Laboratory Animal Care.

Murine heterotopic heart transplantation and histological techniques

Heart grafts were transplanted heterotopically into the abdomen of the recipients using our previously described technique (25). Briefly after removing the donor heart, the donor aorta and pulmonary artery were anastomozted to the recipient infrarenal aorta and inferior venacava, respectively, in an end-to-side fashion. Cold ischemia time was less than 30 min throughout all experiments. The transplanted hearts were removed on day 28, and cut into three parts (base, middle and apex). The basal and middle parts of transplanted hearts were embedded and frozen in OCT compound (Sakura Finetek USA Inc., Torrance, CA), and stored at −20°C. The remaining apical blocks were fixed in 10% formalin and embedded in paraffin. Frozen sections including proximal coronary arteries were cut at 4–6 µm and stained using Weigert’s method for elastic fibers to evaluate the severity of coronary lesions of transplanted hearts. Although not optimal for histology, frozen sections were found to be more efficient to obtain suitable cross-sections of the proximal coronaries and are required for some immunohistochemical stains (e.g. Ly49G2).

Complement components

C4d was detected using a rat monoclonal antibody against mouse C4 that reacts with C4b/C4d (16D2; Abcam Inc., Cambridge, MA, as previously described [18]). Immunofluorescence stain of C3 was performed on frozen sections (ICN Biomedicals, Aurora, OH). Spleens from BL/6 wild-type mice showed deposition of C3 in the germinal centers and capillaries, however, no deposition of C3 was detected in the B6.RAG1^−/−C3^−/−DKO (data not shown). To identify infiltrating cells around coronary lesions and in the neointima, antibodies directed to mouse Mac1 (CD16/32, clone 93, eBioscience) and NK1.1 (Ly49G2, clone 4D11, BD Pharmingen) were utilized. Immunoperoxidase stained sections were then developed in a solution of 3-amino-9-ethyl carbazole (Aldrich Chemical Co., Inc., Milwaukee, WI) or alkaline phosphatase, postfixed in 4% formaldehyde, counterstained with hematoxylin and mounted in Gelvatol (Monsanto Co., Springfield, MA) as previously described (16).

Morphometric analysis

Morphometric analysis was performed on digital microscopic images of coronary arteries near the ostia on tissue sections stained with Weigert’s elastin stain. This is the preferential and earliest site of CTA in the mouse; these vessels are similar in size to intramyocardial vessels in the human (26). The best oriented cross-section of each proximal coronary was photographed by digital light microscopy at 100×–200× magnification. Each heart had one or two arterial images (left and/or right coronary arteries). Image processing and analysis with ImageJ software (NIH) was used to demarcate the borders of the lumen and the intima of the artery. One evaluator, who was blinded to the diagnosis and treatment of the hearts, demarcated the areas on all the sections. Tangential arterial sections were demarcated similar to arterial cross-sections, but in tangential sections measurement was made on the coronary artery only, at the junction of the coronary artery and the aorta. The software then quantitated the manually demarcated luminal and intimal areas. From these area values the ‘neointimal index’ or percent of luminal encroachment, defined as neointimal area divided by neointimal area plus luminal area 100×, was calculated, similar to what has been described previously (27). A higher value of the neointimal index indicates a more severe coronary lesion (100% = completely occluded).

Passive transfer of monoclonal antibodies

All injected monoclonal antibodies; anti H-2K^k IgG1(clone AF 3–12.1.3) and anti H-2K^k IgG2a (clone 36–7-5) were obtained from BioXCell, Lebanon, NH. B6.RAG1/ KO or B6.RAG1/C3/ DKO mice were given repeated injections of anti-H-2K^k mAb at a dose of 30 µg in 200 µl PBS i.p., beginning the day after transplantation and continuing twice a week until completion of the experiments.

Statistical analysis

Excel statistical software for Windows (version 2002, Social Survey Research Information Co., Ltd., Japan) was used for statistical analysis. Statistical differences between two groups were analyzed by either the Fisher’s exact test or two-tailed Student’s t-test. The Mann-Whitney test was performed in cases where the data were skewed. A p-value of < 0.05 was considered significant. Numerical data were expressed as a median (range).

Results

Transfer of IgG2a and IgG1 anticlass I DSA into B6.RAG1^−/− recipients

Hearts from B10.BR mice were transplanted into B6.RAG1^−/− recipients given anti-H-2K^k IgG2a mAb at a dose of 30 µg twice per week. Anti-H-2K^k IgG2a mAb was also utilized as nondonor-specific antibody in which hearts from B6.RAG1^−/−mice were transplanted into B6.RAG1^−/− recipients. In these isografts, the same amount of mAb was received. Cardiac allografts or isografts continued to beat vigorously until the recipients were euthanized 28 days after transplantation. Eight of nine cardiac allografts treated with complement fixing DSA (anti-H-2K^k IgG2a) developed coronary lesions, as did six of eight cardiac grafts treated with a noncomplement fixing DSA (anti-H-2K^k IgG1) (Table 1, Figure 1). In contrast, neither B6.RAG1^−/− isografts treated with non-DSA (anti-H-2K^k IgG2a) nor B10.BR allografts with no treatment developed coronary lesions. The frequency of coronary lesions in the IgG2a DSA group was significantly higher than in both the untreated control group and the anti-H-2K^k IgG2a treatment group with cardiac isografts (p < 0.005, p < 0.01, respectively). Similarly, the frequency of coronary lesions in the IgG1 DSA group was significantly higher than in the untreated control group (p < 0.05).

Table 1.

Frequency of CTA in B6.RAG1^−/−/ recipients of allografts or isografts

Group	Strain combination	Rx1	Hearts with CTA	Comparison of groups2
1		None	0 / 5
2	B10.BR to B6.RAG1−/−	Anti-H-2Kk IgG2a	8 / 9	Group 1–2 p < 0.005
3		Anti-H-2Kk IgG1	6 / 8	Group 1–3 p < 0.005
4	B6.RAG1−/− to B6.RAG1−/−	Anti-H-2Kk IgG2a	0 / 4	Group 2–4 p < 0.01 Group 3–4 p = 0.06

¹mAbs given at dose of 30 µg twice per week. Animals euthanized on day 28 posttransplant.

²p-values calculated using Fisher’s exact test or two-tailed Student’s t-test.

Figure 1. Proximal coronary arteries from B10.BR or B6.RAG1^−/− to B6.RAG1^−/− heart transplants at 28 days.

(A) Control B10.BR heart allograft from no treatment group, no evidence of CTA; (B) B10.BR heart allograft from group treated with anti-H-2Kk IgG2a mAb, beginning on postoperative day +1, thereafter twice per week. Severe coronary transplant arteriopathy developed with pronounced neointimal proliferation and infiltration of mononuclear cells. A portion of the lumen is indicated by an arrow. Even though most of the lumen is occluded, this is a cross-section as judged by the thickness of the media. (C) B10.BR heart allograft from group treated with anti-H-2Kk IgG1 mAb. The noncomplement fixing DSA caused the development of severe CTA. This lesion is at a later stage than that in Figure 1B. (D) B6.RAG1^−/− heart isograft from group treated with anti-H-2Kk IgG2a mAb (nondonor specific antibody). No CTA or infiltrate is evident. All sections were stained with Weigert’s elastin stain. Magnification: (A) 100×; (B,C,D) 200×.

The severity of coronary lesions in hearts as judged by the neointimal index, was greater in those treated with IgG2a DSA than in control groups (Figure 2). The median value of the neointimal index in the IgG1 DSA group was significantly higher than those in the no treatment or isografts treated with anti-H-2K^k IgG2a mAb groups. The cellularity of the infiltrate and the myointimal fibrosis were similar in the IgG1 and IgG2 DSA groups (data not shown).

Figure 2. Comparison of the severity of coronary lesions in the B10.BR to B6.RAG1^−/− combination using the neointimal index.

The allografts in the anti-H-2Kk IgG2a and IgG1 DSA treatment groups showed the most severe CTA. Both are significantly greater than that in either the no treatment group or the isograft group treated with anti-H-2Kk IgG2a mAb. The IgG1 DSA group is not significantly less than the allograft treated with IgG2a DSA. The results are displayed with the 25th and 75th percentile as the area of the box and the 10th and 90th percentile as the whiskers. The median is indicated beside the box. The comparison was performed using the Mann-Whitney test and the p-values were calculated using the same test. The numbers in the parenthesis on the X axis indicate the numbers of coronary arteries in each group.

Complement components in B10.BR hearts into B6.RAG1^−/− recipients

As expected from previous studies (18), C4d deposition was observed in myocardial allograft capillaries in animals treated with IgG2a DSA. In contrast, no C4d deposition was observed in cardiac allografts from IgG1 DSA treated animals, indicating that IgG1 DSA was effective in the absence of detectable activation of the classical complement pathway, (Figure 3). C3 was detected in the microvasculature of the donor hearts from recipients treated with IgG2a DSA.

Figure 3. C4d deposition in hearts in the B10.BR to B6.RAG1^−/− combination.

(A) anti-H-2Kk IgG2a group with allografts, (B) anti-H-2Kk IgG1 group. Anti-H-2Kk IgG2a treatment group had prominent C4d deposition (A). In the anti-H-2 Kk IgG1 treatment group (B), C4d deposition was not observed despite the development of coronary lesions. No C4d was detected in untreated allografts or treated isografts (not shown). Immunoperoxidase C4d stain. Magnification 400×.

Complement-Deficient Recipients

These results raise the possibility that the complement system might not be necessary for antibodies to induce CTA. However, we could not exclude an alternative pathway for complement activation by the IgG1 DSA. Therefore, we tested the ability of alloantibodies to mediate CTA in recipients deficient in C3 (B6.RAG1^−/− C3^−/− DKO).

Twelve B10.BR hearts were transplanted into B6.RAG1^−/−C3^−/− DKO recipients. A passive transfer of anti-H-2K^k IgG2a mAb was given according to the same dose schedule as described above. The transplanted hearts continued to beat vigorously until the animals were euthanized at 28 days after transplantation. Ten of 12 cardiac grafts treated with anti-H-2K^k IgG2a developed coronary lesions, compared with none in 10 cardiac grafts without treatment (p < 0.001, Table 2). The neointimal index was also significantly higher in the group treated with anti-H-2K^k IgG2a than the control group (p < 0.0001, Figure 4).

Table 2.

Frequency of CTA in B6.RAG1^−/−C3^−/− recipients of B10.BR allografts

Group	Strain combination	Rx1	n	Hearts with CTA	Comparison of groups2
1	B10.BR to	None	10	0 / 10
2	B6.RAG1−/−C3−/−DKO	Anti-H-2Kk IgG2a	9	10 / 12	Group 1–2 p < 0.001

¹mAb given at dose of 30 µg twice per week. Animals euthanized on day 28 post transplant.

²p-values calculated using Fisher’s exact test or two-tailed Student’s t-test.

Figure 4. Comparison of the severity of coronary lesions in the B10.BR to B6.RAG1^−/−C3^−/− DKO combination using the neointimal index.

The severity of coronary lesions in B10.BR hearts transplanted into B6.RAG1^−/−C3^−/− DKO recipients is significantly greater than that in the no treatment group. Median values of neointimal index in anti-H-2Kk IGg2a treatment and no treatment groups are 65.6 (8.3–100) and 7.2 (4.5–9.9), respectively. Results presented as in Figure 2.

As expected, since C3 is distal to C4 in the complement cascade, diffuse C4d deposition was observed in cardiac tissue from animals treated with anti-H-2 K^k IgG2a mAb. Allografts in B6.RAG1^−/−C3^−/− DKO treated with anti-H-2K^k IgG2a mAb showed no vascular C3 deposition (data not shown).

Coronary lesions in recipients treated with anti-H2K^k IgG2a mAb showed intense infiltration by inflammatory cells in the neointima and around the coronary arteries (Figure 5). Fcγ RIII+ cells with the appearance of macrophages were detected with CD16/32 antibody in the intima and adventitia of coronary arteries (Figure 6A, B). Furthermore, at early time points, for example day 14 or 17, when the mononuclear infiltrate was most intense, NK cells were observed in the intima and adventitia of coronary lesions with Ly49g2 staining (Figure 6C, D). Alpha smooth muscle actin+ myofibroblasts or smooth muscle cells accumulated in the intima, as in lesions in the human (Figure 7). These findings were similar to the B10.BR to B6.RAG1^−/− combination.

Figure 5. Coronary arteries in the B10.BR to B6.RAG1−/−C3−/− DKO combination.

(A) Coronary artery from no treatment group (arrow). (B) Coronary artery from heart treated with anti-H-2Kk IgG2a mAb. Florid coronary arterial vasculopathy apparent. There are many inflammatory cells in the intima and around the coronary artery. Magnification ×200.

Figure 6. Graft infiltrating cells of transplanted hearts from B6.RAG1^−/−C3^−/−DKO.

The cellular phase of CAV has abundant mononuclear cells in the intima and adventitia. (A) CD16⁺ cells (macrophages and/or NK cells were detected in the adventitia of coronary arteries (×400). (B) CD16⁺ cells with the appearance of macrophages at higher magnification (×800). (C) Ly49G2+ NK cells were observed in the intima and adventitia of the coronary lesions (×200). (D) Ly49G2 NK cells at higher magnification (×400).

Figure 7. Coronary artery from a B6.RAG1 recipient of a B10.BR heart and anti-H-2Kk IgG1 antibody.

The immunohistochemical stain shows extensive alpha-smooth muscle cells in the neointima (arrows). The media (asterisk) is also positive.

Discussion

We have demonstrated that donor-specific alloantibody can be sufficient to initiate arterial intimal inflammation (endarteritis) and neointimal proliferation in the absence of complement activation in murine heart allografts. We showed complement independence by two different approaches. The first used a noncomplement fixing isotype (IgG1), which showed no detectable complement fixation in vivo at the level of the graft endothelium (C4d). However, a low level of complement activation could not be excluded, nor could we rule out a role for the alternative pathway of complement activation. Therefore, a second series of experiments was performed in recipients genetically lacking C3. In this setting, the vascular lesions produced by the passive transfer of alloantibody were equally severe to those with an intact complement system. Pratt et al. reported that in kidney transplantation in mice, production of C3 by the graft was associated with acute rejection via stimulation of T-cell priming (28). Although we could not rule out a contribution of local synthesis of C3 by the graft, no C3 was detected at the level of the endothelium and we believe it unlikely that this is involved.

Complement fixation is essential for the pathogenesis of acute and hyperacute rejection (19–22). Wasowska et al. showed that using an MHC fully mismatched mouse heart transplant model, a passive transfer of complement fixing DSA into immunoglobulin knock out recipients caused acute rejection (22). Noncomplement fixing IgG1 DSA did not trigger rejection by itself, although it could augment the effects of complement fixing antibodies via the lectin pathway (20).

Similarly, studies with a passive transfer of monoclonal xenoantibodies into RAG^−/− mice with rat heart xenografts showed that acute rejection was complement dependent, in that it was prevented by either cobra venom factor or anti-C5 antibody (23).

Several previous clinical studies have raised the possibility that antibodies may mediate graft injury without obvious complement activation as measured by C4d deposition in capillaries. In heart allografts some find a strong association between C4d deposition in myocardial capillaries and the development of CTA (13). However, some studies do not detect this association (14,29). Even in those studies that show an association, 27% develop CTA without C4d deposition (13), implying either another pathway is involved or that the C4d stain is insensitive. In renal allografts, transplant glomerulopathy is strongly associated with DSA and C4d deposition, but many cases do not have C4d at the time of biopsy (30–32). For example, 55% of patients with transplant glomerulopathy and DSA were C4d negative (32). Recently Sis and colleagues reported that an endothelial cell associated gene expression signature of antibody-mediated rejection can be detected in grafts without C4d deposition in patients with circulating antibody (33). This could be explained either by a relative lack of sensitivity of C4d staining or alternatively, by a complement-independent pathway by which antibody may cause these gene expression effects.

Antibodies do affect endothelial cell function in vitro, without the participation of complement. In vitro, alloantibody to class I MHC molecules, without leukocytes or complement, can activate endothelial cells and promote endothelial proliferation via increased expression of FGF-1 receptors (24,31–36).

Antibody binding to class I molecules leads to tyrosine phosphorylation of Src family kinases and the binding to FAK and phosphorylation of downstream molecules including PI3K and Akt (37,38). Recently, Jindra et al. showed that siRNA knockdown of mTOR inhibited the ability of class I antibody to induce phosphorylation of Src and related proteins, indicating that mTOR is involved in antibody mediated endothelial activation (39). Fc receptors are not needed for this endothelial response, since F(ab’)₂ fragments of DSA trigger Akt phosphorylation (34,38). However, a passive transfer of F(ab’)₂ fragments of DSA in vivo can induce similar endothelial expression of phosphorylated proteins, no chronic lesions were detected, suggesting that Fc dependent mechanisms may be needed to promote neointimal proliferation.

Certain responses of endothelial cells in vitro require the participation of cells with Fc receptors. Lee and colleagues showed that IgG1 DSA had little effect on cultured mouse endothelial cells without the addition of peritoneal mononuclear cell populations containing macrophages. With macrophages, increased endothelial production of IL-6 and MCP-1 was detected (40). The effect was blocked by antibodies to FcgRIII and was not elicited with F(ab’)₂. Our immunohistochemical studies show that the cells that infiltrate around the affected coronary arteries and in the intima express Fcγ RIII (CD16), consistent with the possibility that Fc receptor interaction may be involved. NK cells, which express CD16, were also present in the CTA lesions. Further studies to be reported subsequently have supported a mechanistic role of NK/Fc receptors in the pathogenesis of antibody-mediated CTA in C3 deficient recipients (41).

The present study provides evidence that endarteritis can be caused by antibodies independent of T cells. The lesions begin with a cellular infiltrate in the intima and adventitia, and progress to myointimal fibrosis, similar to that in the human (42,43). In clinical practice, endarteritis is widely regarded as a T-cell-mediated lesion, defining type II cellular rejection in the Banff classification (44). In the human, the evidence is that T cells predominate in the intimal infiltrate, endarteritis is reversed by OKT3 and endarteritis is not associated with C4d deposition (42) Indeed, mice deficient in B cells develop florid endarteritis in heart allografts (17). Here, 2 weeks after a passive DSA in T-cell-deficient recipients, we find intense mononuclear cell infiltration of the intima, very similar to the usual endarteritis in humans. However, the infiltrating cells were not T cells, but rather cells with Fc receptors, including macrophages and NK cells. Perhaps analogous cellular markers can be used to detect a non-T-cell variant of endarteritis in humans.

We have shown that using a passive transfer of anti-class I MHC DSA, transplanted hearts develop chronic neointimal proliferation in coronary arteries without the participation of complement. The lack of dependence on complement distinguishes chronic from acute antibody-mediated rejection and may be a factor in the different time course and pathological features (lack of neutrophils, thrombi). New molecular markers other than C4d will be needed to detect the activity of this complement- independent antibody pathway, which may be relevant to C4d-negative chronic rejection not uncommonly observed in patients with donor reactive alloantibodies.