Cardiac allograft vasculopathy (CAV) is a major source of late posttransplant mortality. Although numerous cell types are implicated in the pathogenesis of CAV, it is unclear which cells actually induce the vascular damage that results in intimal proliferation. Because macrophages are abundant in CAV lesions and are capable of producing growth factors implicated in neointimal proliferation, they are leading end-effector candidates. Macrophages were depleted in a murine heterotopic cardiac transplant system known to develop fulminant CAV lesions. C57BL/6 hearts were transplanted into (C57BL/6 × BALB/c)F1 recipients, which then received anti-macrophage therapy with intraperitoneal carrageenan or i.v. gadolinium. Intraperitoneal carrageenan treatment depleted macrophages by 30–80% with minimal effects upon T, B or NK cells as confirmed by flow cytometry and NK cytotoxicity assays. Carrageenan treatment led to a 70% reduction in the development of CAV, as compared to mock-treated controls (p = 0.01), which correlated with the degree of macrophage depletion. Inhibition of macrophage phagocytosis alone with gadolinium failed to prevent CAV. Macrophages may represent the end-effector cells in a final common pathway towards CAV independent of T-cell or B-cell alloreactivity and exert their injurious effects through mechanisms related to cytokine/growth factor production rather than phagocytosis.
The long-term clinical success of cardiac transplantation is constrained by chronic rejection as manifested by cardiac allograft vasculopathy (CAV). This process is characterized by concentric proliferation of smooth muscle and endothelial cells in the allograft vascular intima, provoking a rapidly progressive and diffuse arteriosclerosis of allograft arteries that often culminates in luminal occlusion of arterial vessels, ischemia-induced replacement fibrosis of the allograft parenchyma and eventual allograft failure (1).
The development of more effective therapies and preventative strategies for CAV has been hindered by an insufficient understanding of many basic aspects of its pathophysiology. It is now clear that there are numerous immunologic and nonimmunologic factors that can promote CAV (1). However, emerging evidence suggests that multiple pathways contributing to vascular lesion formation may all converge on a final common pathway. If so, identifying the end-effector element in the final common pathway and targeting it for treatment may provide the most effective and comprehensive means of eliminating this insidious disease.
Although three pathways to CAV, involving T cells, antibody or NK cells, are well documented (1), it is unclear if they possess the ability to drive the neointimal proliferation and fibrosis observed in CAV lesions. Macrophages are particularly appealing end-effector candidates for several reasons. They are known to widely infiltrate both allografts and CAV lesions themselves (2). They are potently activated by cytokines such as IFN-γ that are produced at high levels by upstream triggers of CAV such as T and NK cells (3). They can produce both reactive oxygen species and potent degradative enzymes which might cause adventitious injury of allograft vascular endothelium (4). Finally, macrophages can produce growth factors such as transforming growth factor-beta (TGF-β), platelet-derived growth factor (PDGF) and insulin-like growth factor-1 (IGF-1) and chemokines such as MIG/CXCL9 and RANTES/CCL5, all of which might drive the proliferation of neointimal cells (2,5–9). In this study, we evaluate the contribution of macrophages to chronic lesion formation by depleting macrophages from recipients of parental to F1 cardiac transplants, a system that reliably produces CAV lesions, independent of T-cell or antibody alloreactivity (10).
Materials and Methods
C57BL/6 (H-2b), BALB/c (H-2d) and (C57BL/6 × BALB/c)F1 (H-2b×d) mice were all purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were kept in filter top cages and remained healthy throughout the experiments. All animals were cared for according to American Association for the Accreditation of Laboratory Animal Care approved methods.
Monoclonal antibodies against mouse CD4 (GK1.5, a rat IgG2b) and CD8 (2.43, a rat IgG2b) were produced from cell clones acquired from ATCC (Rockville, MD). Anti-mouse CD49b mAb (DX5, a rat IgM) and anti-mouse F4/80 mAb (CI:A3-1, a rat IgG2b) were purchased from Caltag Laboratories (Burlingame, CA). Anti-mouse CD45/B220 (RA3-6B2, a rat IgG2a) was purchased from BD PharMingen (San Diego, CA). Antibodies used in immunohistochemistry included anti-mouse F4/80 mAb (BM8, eBioscience, San Diego, CA), anti-CD3 (YCD3-1; C.A. Janeway, Yale University), and anti-Ly49G2 (4D11, ATCC).
Murine heart transplantation techniques
Mouse hearts were transplanted to a heterotopic abdominal location with appropriate microsurgical anastomoses according to our previously described technique (11). For parental to F1 transplant experiments, C57BL/6 [B6] donor hearts were transplanted into (C57BL/6 × BALB/c)F1 recipients [designated CB6F1] as described earlier (10). Transplanted hearts were monitored by direct transabdominal palpation at least twice a week. The vigor of contraction of the transplants were recorded on a scale of 0–3+ (10). All transplant recipients received aspirin (Sigma) ad libitum in their water supply (60 mg/L water), starting on the day of transplantation.
Macrophages were either depleted using carrageenan or defunctionalized using gadolinium which blocks phagocytosis. Depletion of macrophages was achieved in transplant recipients with intraperitoneal injections of 1 mg λ-carrageenan type IV (Sigma) dissolved in saline on postoperative day 1, and twice weekly thereafter for the duration of the experiment (12–14). Mice receiving carrageenan all received aspirin anticoagulation, as above, which prevented allograft thrombosis in >85% of transplant recipients. Transplant recipients generally tolerated this regimen well, although almost all developed sterile peritonitis, a known complication of intraperitoneal carrageenan (22,23). Inhibition of macrophage phagocytosis was achieved with i.v. injections of gadolinium(III) chloride (10 mg/kg body weight) on preoperative day -2 and weekly thereafter (12,15). Alternative methods to deplete macrophages using clodronate-encapsulated liposomes (gift of Nico van Rooijen, ref (20)) failed due to systemic thrombosis, which could not be prevented with aspirin, heparin or different liposome batches, but was not observed with saline encapsulated (control) liposomes (data not shown). Apparently, mice are more susceptible to this toxicity than rats (21).
Transplanted hearts were typically removed from recipients on postoperative day 31 and frozen sections were prepared to determine the presence and degree of CAV, as previously described (16). Sections were prepared with Weigert's elastin stain, and intimal proliferative changes were classified into three stages, as previously reported in detail (16). Immunohistological analysis of tissue macrophages was performed as in previous studies using mAbs directed to the mouse F4/80 pan-macrophage marker (17).
Morphometric analysis, similar to what has been described previously (18), was performed on images of coronary arteries near the ostia on tissue sections stained with Weigert's elastic stain. An image of the most representative section was captured digitally by light microscopy at 100×, 200×, or 400× magnification. IP lab imaging software (Scanalytics, BD Biosciences, Rockville, MD) was used to demarcate the borders of the lumen and the intima of the artery. One evaluator, who was blinded as to the diagnosis and treatment of the hearts, demarcated the areas on all the sections. Tangential arterial sections were demarcated similarly to arterial cross-sections, but in tangential sections demarcation 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’, defined as [intimal area/(luminal + intimal area)]× 100, was calculated.
Flow cytometric analysis
Peripheral blood was taken from direct puncture of the inferior vena cava at the time of sacrifice, and peripheral blood leukocytes (PBLs) were isolated following two rounds of water osmolysis of erythrocytes. Nonspecific FcR binding was blocked by anti-mouse FcR mAb 2.4G2 (BD PharMingen) and the PBLs were then washed and incubated with fluorophore-conjugated antibodies specific for F4/80, CD3, DX5 and B220. All incubations were performed for 30 min at 4°C, using 100 μL staining volumes and 500 000 cells per staining reaction. Lymphocytes were analyzed on a FACScan (Becton Dickinson, San Jose, CA).
NK cell cytotoxicity assay
A FACS-based cytotoxicity assay was used as previously described (19). Briefly, YAC target cells were labeled with DiOC18(3) membrane stain for 20 min at 37°C. Simultaneously, splenocytes were recovered from cardiac transplant recipients and prepared with ACK lysing buffer. Labeled YAC target cells were then coincubated for 3 h at 37°C with effector splenocytes at effector:target ratios ranging from 1:2.5 to 80:1, prepared in duplicate. Negative controls of target cells alone and positive controls of target cells treated with 0.1% saponin were also prepared. Following incubation, all samples were stained with membrane-impermeable propidium iodide dye, which only labeled the dead cells. Samples were analyzed on a FACScan (Becton Dickinson), comparing double-staining cells (dead targets) to all cells staining with DiOC18(3) and correcting for spontaneous and max lysis. Cytotoxicity curves were obtained by plotting %adjusted target cell lysis versus actual splenocyte:target ratios. Assay was repeated three times, with representative cytotoxicity curves depicted.
Significant differences between group means were determined using the Fisher's exact test. A p-value of <0.05 was considered significant.
To evaluate the participation of macrophages in CAV pathogenesis, we employed our previously characterized parental to F1 transplant system, involving the heterotopic transplantation of hearts from C57BL/6 donors into (C57BL/6 × BALB/c)F1 hybrid (CB6F1) recipients. This transplant combination consistently produces CAV lesions within 21 days through an alloresponse involving both recipient NK and T cells (10). Moreover, the parental to F1 transplant system does not require the use of potentially confounding immunosuppressive agents to prevent acute allograft rejection.
Carrageenan treatment depletes macrophages in transplant recipients
Flow cytometry was performed on transplant recipients at the time of sacrifice (POD 30) to verify the selectivity of carrageenan-mediated macrophage depletion. Compared to control transplant recipients treated with aspirin alone, carrageenan treatment resulted in a notable 30–80% depletion of macrophages in peripheral blood leukocytes (Figure 1A). The effect of carrageenan on other cell populations such as NK and B-cell populations was negligible, with only a mild (≈20%) depletion of T cells being observed (Figures 1B). The total number of leukocytes extracted from the spleen and peripheral blood of carrageenan-treated recipients was almost identical to those of controls indicating that changes in cell population proportions reflect changes in absolute cell counts (data not shown).
Due to our recent finding that NK cells are important triggers of CAV formation in this parental to F1 transplant system (10), it was necessary to exclude the possibility that carrageenan treatment affected NK cell number or function. Flow cytometry confirmed that carrageenan treatment did not affect the numbers of NK cells in the circulation and spleen (Figure 1A). Furthermore, a comparison of splenocyte effectors from either control transplant recipients (treated with aspirin alone) or carrageenan-treated recipients revealed no significant difference in NK cytotoxicity against YAC target cells (Figure 2).
Macrophage depletion suppresses CAV formation
The effects of macrophage depletion on CAV formation were assessed by comparing F1 recipients of parental hearts treated with aspirin and carrageenan to F1 controls treated with aspirin alone. All recipients were sacrificed on postoperative day 30, as earlier studies with this transplant system identified initial CAV development by postoperative day 21 (10). Transplanted hearts continued to beat vigorously up to the day of sacrifice and showed no signs of acute rejection (Table 1).
1Aspirin given in recipient water supply ad libitum (60 mg/L water), starting on day of transplant.
2Intraperitoneal injection of 1 mg carrageenan given on postoperative day +1 and twice weekly thereafter, along with above aspirin regimen.
3Intraperitoneal injection of 1 mg carrageenan started on postoperative day +30 and twice weekly thereafter, along with above aspirin regimen. Transplant recipients sacrificed on postoperative day 51.
4Gadolinium(III) chloride (10 mg/kg body weight) administered i.v. via dorsal penile vein injection on preoperative day -2 and weekly thereafter.
5All transplanted hearts continued to beat vigorously at time of recipient sacrifice.
6p < 0.01 compared to aspirin-only controls in group 3.
7Mean neointimal index ± standard error. Neointimal index =[intimal area/(luminal + intimal area)]× 100.
8p < 0.001 compared to aspirin-only controls in group 3.
10.7 ± 2.3
78.2 ± 6.9
Carrageenan + aspirin2
26.8 ± 10.18
All parental to F1 transplants treated with aspirin alone showed evidence of fulminant vasculopathy whereas no isografts developed lesions (Table 1, groups 1–3). Previous results showing that parental to F1 transplants receiving no treatment also develop CAV indicate that CAV in this system is not a consequence of aspirin administration (10). Each arterial lesion exhibited collagen and smooth muscle cell accumulation, often resulting in near complete luminal occlusion (Figure 3B). Cuffs of infiltrating cells were present around affected vessels, although there was little or no accompanying inflammation in the myocardium. These lesions reproduced with fidelity the vascular lesions observed in human heart transplant recipients undergoing chronic rejection.
In striking contrast, only 3 of 10 carrageenan-treated transplant recipients developed lesions of CAV (Table 1, group 4). Histological analysis showed that most of these macrophage-depleted recipients had pristine allograft coronary arteries with no evidence of vasculopathy (Figure 3A), a notable difference from the aspirin-only controls (Figure 3B). Quantitative assessment of the coronary arteries with computerized morphometry analysis verified that macrophage depletion with carrageenan substantially suppresses the formation of neointimal proliferation compared to controls treated with aspirin alone (Table 1). Importantly, the mild T-cell depletion observed with carrageenan treatment did not likely contribute to the observed abrogation of CAV because earlier studies demonstrated that treatment of these transplant recipients with anti-CD4 and anti-CD8 mAbs (achieving >90% T-cell depletion) could not prevent CAV (10). Immunohistochemistry revealed collections of macrophages surrounding coronary vessels in both carrageenan-treated and control transplant recipients, although the macrophage-specific F4/80 staining was far more intense in the controls (Figures 3C and 3D). The presence of infiltrating macrophages in carrageenan-treated recipients demonstrates the limitations of macrophage depletion achieved by this regimen. However, the degree of macrophage depletion in the three carrageenan-treated recipients that did develop vascular lesions was significantly less than in those that did not develop vasculopathy (p < 0.05), suggesting that abrogation of CAV correlated directly with the degree of macrophage depletion achieved (Figure 4). Infiltrating CD3+ T cells were also noted in both groups (data not shown).
Delayed treatment had no effect on CAV lesion formation
To determine whether delayed macrophage depletion could reverse established CAV lesions, parental to F1 cardiac transplant recipients were treated with a 3-week course of carrageenan starting on postoperative day 30, by which time CAV lesions should have already formed in the allografts (Table 1, group 5). Upon sacrifice, all four of the recipients receiving this delayed carrageenan regimen showed evidence of persistent vasculopathy (Figure 3E).
Inhibition of macrophage phagocytosis fails to abrogate CAV
In order to distinguish mechanisms by which macrophages may mediate the pathogenesis of CAV, we treated transplant recipients with i.v. gadolinium chloride. Unlike carrageenan, gadolinium specifically inhibits macrophage phagocytosis without causing depletion (12,15). Treatment of parental to F1 cardiac transplant recipients with gadolinium chloride failed to suppress CAV in six of six recipients (Table 1, group 6). The morphology of lesions in the gadolinium-treated recipients closely resembled those observed in untreated controls, showing an advanced degree of neointimal proliferation (Figure 3F). Of note, treatment with gadolinium did not deplete splenic macrophages (data not shown). This finding suggests that the role of macrophages in promoting CAV is independent of phagocytosis or antigen-presentation.
We have demonstrated that depleting host macrophages with carageenen can suppress the development of CAV in a parental to F1 heterotopic cardiac transplant system. The success of this strategy was correlated with the degree of macrophage depletion achieved. While apparently critical for the initial pathogenesis of CAV, macrophages did not appear to be required for the maintenance of established CAV lesions, as delayed treatment had no effect on lesion development. Finally, the ability of macrophages to promote CAV formation was likely related to cytokine or growth factor release rather than phagocytosis or antigen presentation, as selective inhibition of macrophage phagocytosis with gadolinium chloride could not prevent CAV. Together, these findings suggest a role for macrophages in the pathogenesis of CAV, possibly as end-effectors in a final common pathway toward CAV. To our knowledge, this present work marks the first demonstration that specific macrophage depletion can suppress CAV formation in an unmanipulated vascularized organ transplant system.
The precise contribution of macrophages to the pathogenesis of transplant vasculopathy is ill-defined in the literature. While macrophages are abundantly present in CAV lesions (2) and cytokines associated with macrophage activation are upregulated during CAV development (3), it remains unproven whether this is a cause or effect of CAV. Earlier reports have failed to establish a definitive causal link between macrophages and CAV, in part due to the difficulties of achieving selective macrophage inhibition or depletion. Although treatment of murine transplant recipients with purported macrophage inhibitors such as gamma-lactone and 15-deoxyspergualin managed to partially suppress transplant vasculopathy, the specificity and precise effects of these agents on macrophages are unknown (24,25). The severe reduction of vasculopathy observed with experimental transplants using the supposedly macrophage-deficient op/op osteopetrotic mice as recipients is also raised to implicate macrophages in CAV formation (26). The op/op mouse, however, also suffers from pleiotrophic immune defects such as aberrant granulocyte function, impaired Th1 differentiation and irregular B-cell lymphopoeisis (27–29). These confounding immune defects prevented the definitive interpretation of transplant experiments utilizing op/op mice, leaving the actual role of macrophages in CAV pathogenesis unresolved. The findings of this study help establish a firmer link between macrophages and CAV.
A more comprehensive examination of macrophage activation pathways and effector functions would be required to identify specific mechanisms by which macrophages may contribute to allograft vasculopathy. However, several distinct pathways of macrophage activation and function have been elucidated in recent years which have potential relevance to CAV. Classical activation of macrophages requires the dual stimulation of macrophages by IFN-γ and engagement of their toll-like receptors (TLRs) (30,31). Organ transplantation might induce the classical activation of macrophages, as IFN-γ is abundantly produced by the activated T cells and NK cells that infiltrate allografts in the early postoperative period, and macrophage TLRs recognize ligands such as heat shock proteins and altered lipoproteins that are liberated from allograft cells damaged by ischemia-reperfusion injury (32–34). An alternative activation pathway for macrophages was also recently described. This pathway develops in the setting of a local cytokine milieu enriched in IL-4 and/or IL-13, cytokines traditionally associated with a TH2-type immune response (31). Macrophages activated via the alternative pathway have a less robust respiratory burst response, but they instead promote angiogenesis and proliferation of vascular endothelium, characteristics which might facilitate macrophage involvement in CAV pathogenesis (31,35).
Once activated by organ transplantation, recipient macrophages possess three distinct sets of effector functions which may enable them to promote allograft arteriopathy. First, their phagocytic properties and constitutive expression of MHC class II products establish macrophages as excellent antigen-presenting cells to CD4+ T cells, although their failure to constitutively express costimulatory molecules prevents them from activating naïve T cells (32). Through the presentation of donor MHC peptides, macrophages may contribute to the indirect allorecognition pathway, culminating in smoldering sub-acute alloresponses that are implicated in vasculopathy development (36). Macrophage facilitation of the adaptive immune alloresponse is also promoted by production of chemokines such as RANTES/CCL5 and MIG/CXCL9. However, increased leukocyte recruitment and donor antigen presentation by macrophages cannot be invoked as the dominant mechanism by which macrophages contribute to CAV formation, as inhibition of macrophage phagocytosis and antigen presentation with gadolinium chloride administration failed to suppress allograft vasculopathy.
Next, activated macrophages may directly injure allograft vascular endothelium through the release of reactive oxygen species. Injury of this vascular endothelium is hypothesized to be the common insult that underpins all causes of CAV, perhaps through a pathway involving dysregulation of endothelial nitric oxide production (37). Finally, in addition to inducing direct vascular endothelial injury, macrophages may promote an aberrant healing response fostering allograft arteriopathy through the production of a large array of growth factors (such as PDGF, IGF-1 and TGF-β). Many of these growth factors are known to promote vascular smooth muscle proliferation and migration, essential steps in the evolution of the neointimal proliferation that characterizes CAV (2,5–7).
This study has several limitations. First, the peritonitis induced by carrageenan is associated with granulocytosis (38), so the possibility cannot be fully excluded that CAV suppression was caused by neutrophil-produced protective factors rather than macrophage depletion. Additionally, while carrageenan has been employed extensively in the literature to induce selective macrophage depletion, the possibility that carrageenan may deplete other immune cells (particularly other phagocytic cell lineages such as dendritic cells) was not specifically examined, and therefore the suppression of CAV may not actually result from macrophage depletion. However, the leukocyte subsets most relevant to transplant rejection (including T cells, B cells and NK cells) were examined, and minimal adventitious effects on these cell lineages were noted following carrageenan administration.
Next, while evidence that macrophages promote CAV through growth factor release is suggestive, these findings cannot definitively prove that macrophages function as end-effectors in CAV pathogenesis rather than as intermediaries in this process. Other cell populations implicated in CAV (such as T, B and NK cells) might instead serve as both triggers and end-effectors in the CAV cascade. Supporting this hypothesis, both NK and T cells can produce abundant IFN-γ, a cytokine which may be sufficient for CAV formation (39). However, other studies have cast doubt on the premise that IFN-γ can induce neointimal proliferation by itself (10). The final notable limitation of this study involves its use of a single specific system of CAV formation (the parental to F1 hybrid transplant system) which may not be fully relevant to CAV lesions attained in clinical transplantation, despite the verisimilitude of the CAV lesions developed in this experimental system. This well-characterized transplant system was selected because it offered several unique advantages for the purposes of this study, including the rapid development of almost universal allograft vasculopathy and freedom from acute alloresponses, eliminating a requirement to use a potentially confounding immunosuppressant regimen.
Multiple lines of evidence support a role for macrophages in chronic alloresponses against other solid organ grafts. For example, the intensity of macrophage infiltration in early biopsies of human renal allografts is predictive of subsequent development of fibrosis (40). In miniature swine, persistence of macrophage infiltration following the resolution of an episode of acute rejection also predicted the ultimate evolution of chronic renal allograft rejection (41). An anti-macrophage regimen (gadolinium) also was able to suppress the development of bronchiolitis obliterans, the manifestation of chronic rejection in lung transplants, in a rat model of heterotopic tracheal transplantation (42). Furthermore, our studies complement a growing body of evidence implicating macrophages in clinical vasculopathy arising outside the setting of organ transplantation. Macrophage depletion partially ameliorates the neointimal proliferation seen in vein grafts and in animal models of balloon angioplasty (43–45). Thus, the therapeutic implications of these findings may have broader applicability in the field of transplantation and vascular biology.
It is notable that CAV arises despite long-term immunosuppression that is effective in preventing acute rejection. Macrophages are not specifically targeted by calcineurin inhibitors (46), but rapamycin partially inhibits macrophage proliferation in response to macrophage colony-stimulating factor (M-CSF) (47). Rapamycin also suppresses monocyte chemotaxis into tissues, hindering their development into macrophages (48). Perhaps not coincidentally, rapamycin is able to reduce the incidence of CAV in human heart transplant recipients more than twofold compared to other conventional immunosuppressants (49). These results therefore offer the hope that specifically targeted macrophage inactivation or depletion in humans may improve the long-term outcome of transplanted organs.
We thank Dr. Bruce Rosengard for critical review of this manuscript. This work was funded in part by grants from the National Heart Blood and Lung Institute of the National Institutes of Health (RO1 HL071932) and a grant from the Roche Organ Transplant Research Foundation. W.H.K. is a Howard Hughes Medical Institute Research Training Fellow.