PROLONGATION OF PRIMATE CARDIAC ALLOGRAFT SURVIVAL BY... : Transplantation (original) (raw)
CD40 ligand (CD154) was initially identified in rodents (1) and man (2) as a T-cell surface molecule critical to the induction of T-helper cell-dependent antibody responses (3). In recent studies, perioperative CD154 blockade dramatically prolonged primate renal allograft survival (4-6). Based on these observations, we investigated whether monotherapy with a humanized anti-CD154 monoclonal antibody, hu5c8, prolongs survival of primate heterotopic cardiac allografts in a preclinical model.
Monoclonal humanized 5c8 (hu5c8*) incorporates the complementarity-determining region of the anti-human CD154 antibody, 5c8, in a human IgG1 framework (Biogen, Inc., Cambridge, MA) (2, 4-6). Two hu5c8 dosing strategies were evaluated. All treated animals received hu5c8 intravenously at the time of graft implantation (day 0) and on postoperative days 5 and 14. Six animals (group 1) received 15 mg/kg perioperatively (day 0) and 5 mg/kg subsequently (cumulative dose, 25 mg/kg). In the sustained therapy group (group 2), one animal received perioperative, day 5, and day 14 therapy as for group 1, followed by 10 mg/kg on day 28 and biweekly thereafter. The other two animals in this group received 20 mg/kg on day 0, 10 mg/kg on days 5 and 14, and then 20 mg/kg on day 28 and monthly until graft demise. Control animals received donor-specific transfusion and either no treatment (n=1) or polyclonal human IgG (15 mg/kg on the day of transplantation and 5 mg/kg on postoperative day 5; Gamimune, Bayer, Pittsburgh, PA; n=3). No hemodynamic compromise was observed during administration of hu5c8, nor was there obvious evidence of cutaneous, gastrointestinal, or renal toxicity. Treated animals were generally vigorous, and resilient after procedures. We did not observe any opportunistic or other infections.
Outbred, wild-caught cynomolgus (Macaca fascicularis) monkeys (Coulston Foundation, Holman AFB, NM) weighing 3-7 kg were paired based on compatible ABO blood type and class II MHC disparity, as determined by a stimulation index (SI) greater than 3 (usually >8) in unidirectional mixed lymphocyte reaction (MLR). Anesthesia was induced with intramuscular ketamine (10 mg/kg) and xylazine (0.1 mg/kg) and maintained with endotracheal isoflurane. After systemic heparinization and collection of 120-200 ml of donor whole blood into CPDA anticoagulant, diastolic arrest of the donor heart was induced with University of Wisconsin solution (15-20 ml/kg, 4°C; Viaspan, DuPont-Merck, Wilmington, DE) via the aortic root, and the heart explanted and prepared.
All group 1, one group 2, and three control animals underwent traditional nonworking heterotopic cardiac allograft transplantation wherein the donor aorta was anastomosed end-to-side to the infrarenal abdominal aorta of the recipient, and the donor pulmonary artery was anastomosed to the adjacent vena cava. In two group 2 animals and one control animal a working heterotopic technique was used, with a single aortic anastomosis, as described by Klima et al. (7). Each recipient received fresh donor blood (10-15 ml/kg) intravenously during the transplantation procedure. Postoperative analgesia consisted of intramuscular buprenorphine (0.01 mg/kg) and flunixin (1 mg/kg). All procedures were approved by the Vanderbilt University IACUC and were carried out in compliance with the Guide for the Care and Use of Laboratory Animals (HHS, NIH Publication 86-23, 1985).
Graft contractility was assessed daily by palpation. Confirmatory abdominal ultrasounds were performed at the time of protocol biopsies and whenever an examiner appreciated decreased contractility. Graft failure was defined as a loss of palpable graft activity, with ultrasound confirmation of weak or absent myocardial contractility. Failed grafts were explanted promptly and examined histologically. Kaplan-Meier graft survival was compared by log-rank analysis. Statistical calculations were performed using SPSS 8.0 (SPSS, Inc., Chicago, IL).
Median graft survival time was prolonged significantly to 49 days (range 14 to 56) in the perioperative treatment group (group 1) and 106 days (range 56 to 245) in the sustained dosing group (group 2), in comparison to 5 days (range 5 to 6) among control recipients (P<0.05 for all comparisons). Histologic characteristics of the graft biopsy specimens and explants for controls and treated animals are shown in Table 1. One animal in group 1 (M1158) rejected its graft at day 14. Another animal (M1150) was killed at day 19 due to iatrogenic small bowel obstruction; there was vigorous graft function at the time of death, and this animal is therefore not included in the statistical analysis. Histologic examination of the graft explanted at 35 days (M1344) demonstrated venous thrombosis and absence of interstitial cellular infiltrate in the edematous myocardium; perivenular infiltration with neutrophils was present. Graft failure in this animal may have been either due to an atypical inflammatory process primarily involving coronary veins or technical, due to venous anastomotic stricture. Since we cannot exclude a contribution of immune mechanisms to failure of this graft, it is included in the survival statistics.
Graft survival and histologic summary of serial allograft biopsy specimens
In the three remaining group 1 animals treated with perioperative low-dose hu5c8, graft function persisted beyond 6 weeks. In two of these animals (M1122 and M1128), declining graft function by palpation and ultrasound over several days heralded subsequent graft failure and correlated histologically with advanced rejection and zonal myocardial infarction. The third animal (M1514) exhibited similar signs of depressed graft function on day 51 and received a single additional "rescue" dose of hu5c8 (15 mg/kg). Biopsy results 2 days later demonstrated some areas of healing scar and other areas of myocardium containing a lymphocytic infiltrate consistent with mild "resolving" rejection (International Society for Heart and Lung Transplantation [ISHLT] grade 1B, Fig. 1A). The infiltrate was increased in intensity relative to that observed in a protocol biopsy specimen from day 42 (grade 0, Fig. 1B), obtained 9 days before the clinical rejection episode. Graft function stabilized until day 92, at which time graft function ceased despite a second "rescue" dose of hu5c8 on day 90. For statistical purposes, graft failure in this animal was taken as day 51. With this exception, rejection episodes were not treated.
Histology before (A) and after (B) treatment of a clinical episode of acute rejection. (A) Core biopsy specimen of M1514 on day 42 showing no evidence of acute cellular rejection (grade 0). This biopsy specimen was obtained 9 days before a clinical episode of rejection. (hematoxylin & eosin; ×400). (B) M1514 biopsy specimen on day 53, two days after a "rescue" dose of hu5c8 for clinical acute rejection, showing mild acute cellular rejection with a sparse interstitial lymphocytic infiltrate, but without distortion of the overall myocyte architecture or myocyte necrosis (grade 1B) (hematoxylin & eosin; ×400).Biopsy specimens of hu5c8-treated heart allografts during the first week exhibited various grades of acute rejection (D and E), whereas control hearts (C) uniformly had advanced rejection. (C) Control heart on day 5 (M1274), explanted with minimal residual graft contractility. Histologic examination shows severe acute cellular rejection, with a diffuse aggressive lymphocytic and eosinophilic interstitial infiltrate combined with marked myocyte necrosis, edema, and hemorrhage (grade 4) (hematoxylin & eosin; ×400). Vasculitis was not seen in this group, and coronary thrombosis with associated myocardial necrosis was observed in only one instance (M1374). (D) Biopsy specimen from M1128 on postoperative day 5 showed a mild acute cellular rejection with a scant interstitial lymphocytic infiltrate (grade 1B) (hematoxylin & eosin; ×400). (E) Moderate acute cellular rejection (grade 3A) in animal M1360 on day 5, with multiple aggressive lymphocytic infiltrates expanding the interstitial spaces resulting in distortion of overall myocyte fiber architecture. Myocyte necrosis was not prominent, and graft contractility was normal (hematoxylin & eosin; ×400). (F) Biopsy specimen of the same animal pictured in E (M1360) at day 90, demonstrating complete resolution of acute rejection (grade 0) (hematoxylin & eosin; ×400).Vascular changes in an explanted group 2 heart (M1360, day 245) (G-I). (G) A dense perivascular lymphocytic infiltrate, endothelial activation, and cellular intimal proliferation are illustrated, associated with grade 4 severe acute cellular rejection. No fibrosis is detected in this vessel on routine stains (hematoxylin & eosin; ×100). (H) Large artery from M1360 explant, showing intact internal elastic lamina (black), endothelial activation and proliferation, scattered intraintimal lymphocytes, and detectable early fibrosis (green staining indicates collagen deposition within the proliferating intima) (Movats Pentachrome stain; ×400). (I) Little fibrosis (blue) is demonstrated within vessel walls even in areas of prolific intimal proliferation, suggesting that these are active lesions and represent an early-to-intermediate stage in the spectrum of chronic allograft vasculopathy (M1360 at day 245; Masson's Trichrome stain; ×100).
One animal in the sustained dosing group (M1044) was treated using the group 1 regimen (25 mg/kg cumulative dose from day 0-14) followed by 10 mg/kg biweekly, starting on day 28; this graft continued to contract well until day 106, although persistent moderate-grade rejection (grade 3A/B) was present from day 42 onward. Two animals received intensified perioperative hu5c8 dosing (40 mg/kg cumulative dose from day 0 to 14). One graft failed at day 56 (M1132), just before a second scheduled monthly maintenance hu5c8 dose. The other graft functioned well until it was rejected at 245 days, over 2 months after the last monthly hu5c8 dose (20 mg/kg) on day 180 (M1360).
Pathologic analysis: Open allograft biopsy specimens were obtained surgically twice within the first 2 weeks postoperatively and biweekly or monthly thereafter, using a 14-gauge core biopsy needle device (Tru-Cut, Baxter Diagnostics, Stone Mountain, GA) to obtain multiple full-thickness samples of the ventricle. Formalin-fixed biopsy specimens and hearts explanted after cessation of graft function were stained with hematoxylin and eosin, and selectively with Masson trichrome or Movats Pentachrome. Rejection was graded (M.A.S.) according to standard ISHLT criteria (8). We interpreted the finding of plump, raised endothelial cells with abundant cytoplasm and rounded nuclei as indicative of endothelial activation, and defined vasculitis as the presence of intimal invasion by inflammatory cells. All available biopsy and explanted tissue was examined for evidence of cellular intimal proliferation or fibrous lesions in vessels, and the presence and prevalence of these lesions were separately noted.
The results of the histologic evaluation of serial graft biopsy specimens and explanted grafts are summarized in Table 1. Explanted control hearts (Table 1A) demonstrated acute cellular rejection (grade 3A-4), with multifocal aggressive lymphocytic infiltrates, diffuse interstitial hemorrhage, and myocyte necrosis (Fig. 1C). Endothelial activation was noted, but no intimal proliferation or fibrosis was seen.
Serial biopsy specimens from groups 1 and 2 grafts consistently revealed lymphocytic infiltrates during the first 3 weeks, which ranged from diffuse and mild (grade 1B, Fig. 1D) to focally aggressive in one (grade 2) or more areas (grade 3A/3B, Fig. 1E). In every graft that survived the first month at least one rejection score was obtained that was lower than the score assigned on the previous biopsy. In five instances, resolution of infiltrate (grade 0) was documented on at least one subsequent biopsy or at explant (Fig. 1F). Recrudescent infiltrates progressed slowly (M1514, 1132, 1044). Graft contraction (assessed by daily palpation, and confirmed intermittently by ultrasound or direct inspection at the time of biopsy) generally remained vigorous during this interval, often despite the presence of prominent, multifocal, "aggressive" lymphocytic infiltrates and associated vasculitis (Table 1B). Diminished graft function heralded cessation of contractility by several days and, except as noted above, histologic examination of explanted grafts confirmed severe acute cellular rejection (grade 4, Fig. 1G).
Vasculitis, with lymphocytes, eosinophils, and occasional neutrophils in and around vessel walls, was typically associated with acute rejection, both in biopsy specimens and in explants. Cellular intimal proliferation, often with luminal obliteration and surrounding focal, old myocardial infarction, was often observed in conjunction with graft failure in group 1 but not in control hearts, suggesting that vessel occlusion contributed to graft demise in this group. Group 2 protocol biopsy specimens obtained after day 14 and before explant differed from group 1 specimens obtained at similar times in that intimal proliferative lesions were milder, even in the setting of intense, persistent perivascular lymphocytic infiltrates and endothelial activation (grade 3A/3B, M 1044, days 42-84). Zonal myocardial necrosis was not seen in protocol biopsy specimens in group 2, suggesting that regional ischemic infarction due to vascular insufficiency had not occurred. Intimal collagen deposition was appreciated associated with prolific endothelial cellular proliferation and intact elastic lamina in the graft explanted at 245 days (M 1360, Fig. 1, H and I). We interpret this finding as consistent with an intermediate stage in the evolution of chronic allograft vasculopathy.
MLR assay: Within groups of wild-caught mature animals obtained from a single vendor (Coulston Foundation, Holman AFB, NM), heart donors were matched to ABO-compatible recipients to achieve maximal donor-specific MLR responses. Assays performed preoperatively or on the day of transplantation (day 0) were performed using fresh recipient peripheral blood mononuclear cells (PBMC) and donor PBMC or splenocytes. After transplantation, MLR was performed on days 5 and 14, and biweekly or monthly thereafter, using freshly collected recipient lymphocytes stimulated with cryopreserved splenocytes from the heart donor and a consistent third-party source. Splenocytes and PBMC were prepared and purified using standard techniques, and were plated in triplicate 4-day MLR microcultures (105 responders and stimulators/well) (9).
The median SI before transplantation was 7.3 (range 3.2 to 13.1). In group 1 animals, the SI tended to decrease over time during therapy (week 3, Fig. 2, B and C), and fell further around the time of graft rejection, 3-5 weeks after the last hu5c8 dose (week 6). This pattern was similar to that observed in controls around the time of graft loss (1 week, 3 weeks; Fig. 2A). However, even in animals (e.g., M1122, M1344; Fig. 2C) in which the SI fell, donor-specific proliferation remained detectable (SI>2). All group 2 animals demonstrated vigorous proliferative responses in donor and third-party MLR throughout the postoperative period (Fig. 2, D-F). Comparison of proliferation to donor and third-party antigens revealed no consistent evidence for donor-specific hyporesponsiveness in the circulating recipient lymphocyte pool in any animal. Nor did increased MLR activity consistently correlate with or predict graft loss.
Hu5c8 treatment does not induce donor-specific hyporesponsiveness in MLR. MLR was performed with fresh recipient peripheral blood lymphocytes and irradiated peripheral blood lymphocytes or splenocytes from the donor (•) or a consistent irrelevant animal (third-party, ○) at the indicated times before and after transplantation. (A) Control animals; (B and C), group 1 animals treated with perioperative hu5c8; (D-F), individual group 2 animals treated with sustained hu5c8 therapy. An MLR against the animal subsequently used as the third-party stimulator was not obtained preoperatively for M1360.
Our results demonstrate that regimens using perioperative hu5c8 monotherapy are associated with prolonged survival of primate cardiac allografts. Increasing the intensity of therapy further delays graft failure without obvious toxicity. Whether this effect is attributable to higher perioperative dosing or continuing treatment beyond day 14, and whether administration of donor-specific transfusion influenced outcome, are important questions not addressed by our study. That monotherapy targeting the CD154 pathway reliably prevents early graft failure corroborates for cynomolgus heart allografts the observations of Kirk et al. (4, 5) and Kenyon et al. (6) in rhesus renal and islet allograft models. Our findings differ importantly from theirs in that we have not observed prevalent long-term survival even during continued therapy, and vascular lesions are prominent. Whether these differences reflect differences between our models (species, organ, recipient maturity), or rather reflect the consequence of differences in hu5c8 dosing regimen, are also important questions.
Our MLR findings are similar to those reported using a short duration of hu5c8 (4) and different from those described more recently with prolonged therapy (5). The evolution of hyporesponsiveness with prolonged therapy is consistent with the modulation of donor-directed immunity. This finding supports the hypothesis that intense, long-term anti-CD154 therapy may facilitate the emergence of regulatory immune responses to donor antigens not induced with shorter or less intense CD154 blockade. The cumulative dose administered in our group 2 animals was roughly half of that used by Kirk et al. (4, 5) during the first month after grafting. The pharmacokinetic behavior of hu5c8 is consistent between nonhuman primate species, including cynomolgus and rhesus monkeys (10, and Biogen, unpublished observations), as confirmed by occasional measurements obtained during the course of these studies, suggesting substantially lower drug exposure in our experiments. Our data demonstrate that, at the doses we evaluated, inhibition of donor antigen recognition and antidonor proliferative responses do not occur in the periphery, and thus cannot account for the observed effect on graft survival.
We initiated these studies based on the hypothesis that the immunomodulatory effects of CD154 blockade involve interruption of the signal 2-driven, CD154-mediated costimulatory activation events classically associated with initial antigen recognition. Intact donor-specific responses in MLR, particularly in group 2, coupled with the presence of cellular infiltrates in the graft during the first weeks after transplantation lead us to speculate that anti-CD154 therapy does not block an active immune response to donor antigens, but rather modulates that response, or attenuates its consequences. Analysis of graft cytokine mRNA expression (11, and A.C.C., preliminary data, not shown) confirms that lymphocyte activation in the graft is not prevented at the doses of hu5c8 we used, substantially corroborating results reported by Larsen et al. (12) in rodents. The apparent discrepancy between the evidence of histologic "rejection" (using standard ISHLT criteria) and the persistence of vigorous graft function for weeks thereafter suggests that hu5c8 blocks expression of cytotoxic effector function. Further, hu5c8 apparently arrested progression to graft loss (M1514), confirming Kirk's (4) similar finding. We interpret these results to suggest that CD154 plays an important role in recruitment and amplification of the effector phase in acute allograft rejection.
An important issue raised by our studies and those recently reported in renal allografts (5) is the significance and interpretation of lymphocytic infiltrates during hu5c8 therapy. In contrast to clinical renal and liver transplantation, where evidence of graft dysfunction is used to guide immunosuppressive management, in cardiac transplantation the development of clinically detectable ventricular dysfunction with rejection often portends life-threatening graft failure. Therefore, protocol biopsies are performed routinely, and immunosuppression is based primarily upon histologic criteria. If lymphocytic infiltrates are a concomitant of modulation of effector function by anti-CD154, and particularly if addition of other "conventional" immunosuppressive agents block this effect (5), then novel means must be developed to distinguish pathogenic from adaptive or regulatory immune responses to apply anti-CD40 ligand therapy safely and effectively in the clinic.
Whether perioperative administration of donor peripheral blood is important to the effect of CD154 blockade on heart allograft survival in our system is an important question. In our estimation, the available evidence from many other models suggests that exposure to additional donor antigen perioperatively may be particularly important for cardiac (as opposed to renal or liver) allografts, if not for long-term graft acceptance with ongoing immunosuppression, then for induction of stable peripheral tolerance after perioperative induction therapy. This hypothesis, and its potential importance to anti-CD154-based regimens, remains to be tested in primates or in man. Finally, whether the intimal proliferative lesions we see are due to subtherapeutic hu5c8 dosing or represent the consequence of CD154-independent pathogenic mechanisms will also require further study.
In conclusion, CD154 blockade prolongs the survival of cardiac allografts in primates. It is possible that this occurs by promoting or allowing adaptive modulation of an active response to donor graft antigens while also inhibiting myocardial injury by those cells that do infiltrate the graft. Alternatively, rejection occurs that is mechanistically similar to that observed without immunosuppression or on conventional agents, but is delayed in its tempo. In either case, we speculate that this feature of anti-CD154 therapy may allow manipulation of donor-specific immune responses around the time of engraftment, and thus may provide the foundation for a clinically applicable strategy to induce stable peripheral tolerance in man based on improved understanding of the mechanisms involved.
REFERENCES
1. Armitage RJ, Fanslow WC, Strockbine L, et al. Molecular and biological characterization of a murine ligand for CD40. Nature 1992; 357(6373): 80.
2. Lederman S, Yellin MJ, Krichevsky A, Belko J, Lee JJ, Chess L. Identification of a novel surface protein on activated CD4+ T cells that induces contact-dependent B cell differentiation (help). J Exp Med 1992; 175(4): 1091.
3. Noelle RJ, Ledbetter JA, Aruffo A. CD40 and its ligand, an essential ligand-receptor pair for thymus-dependent B-cell activation. Immunol Today 1992; 13(11): 431.
4. Kirk AD, Harlan DM, Armstrong NN, et al. CTLA4-Ig and anti-CD40 ligand prevent renal allograft rejection in primates. Proc Natl Acad Sci USA 1997; 94(16): 8789.
5. Kirk AD, Burkly LC, Batty DS, et al. Treatment with humanized monoclonal antibody against CD154 prevents acute renal allograft rejection in nonhuman primates. Nature Med 1999; 5(6): 686.
6. Kenyon NS, Chatzipetrou M, Masetti M, et al. Long-term survival and function of intrahepatic islet allografts in rhesus monkeys treated with humanized anti-CD154. Proc Natl Acad Sci USA 1999; 96: 8132.
7. Klima U, Guerrero JL, Levine RA, Vlahakes GJ. A new, biventricular working heterotopic heart transplant model: anatomic and physiologic considerations. Transplantation 1997; 64(2): 215.
8. Billingham ME, Cary NR, Hammond ME, et al. A working formulation for the standardization of nomenclature in the diagnosis of heart and lung rejection: Heart Rejection Study Group. The International Society for Heart Transplantation. J Heart Transplant 1990; 9(6): 587.
9. Chapman A, Stewart SJ, Nepom GT, et al. CD11b+CD28−CD4+ human T cells. J Immunol 1996; 157: 4771.
10. Gobburu JV, Tenhoor C, Rogge MC, et al. Pharmacokinetics/dynamics of 5c8, a monoclonal antibody to CD154 (CD40 ligand) suppression of an immune response in monkeys. J Pharmacol Exp Ther 1998; 286(2): 925.
11. Chang AC, Blum MG, Blair KSA, et al. Modulation of interleukin-12 gene expression in primate cardiac allografts during anti-CD40 ligand monoclonal antibody treatment. Surg Forum 1998; 49: 382.
12. Larsen CP, Alexander DZ, Hollenbaugh D, et al. CD40-gp39 interactions play a critical role during allograft rejection: suppression of allograft rejection by blockade of the CD40-gp39 pathway. Transplantation 1996; 61(1): 4.
* Abbreviations: hu5c8, monoclonal humanized 5c8; ISHLT, International Society for Heart and Lung Transplantation; MLR, mixed lymphocyte reaction; PBMC, peripheral blood mononuclear cells; SI, stimulation index.
© 1999 Lippincott Williams & Wilkins, Inc.