| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2007;133:1354-1363
© 2007 The American Association for Thoracic Surgery
Cardiothoracic Transplantation |
a University of Maryland and Baltimore Veterans Administration Medical Center, Baltimore, Md
b Immerge Biotherapeutics Inc, Cambridge, Mass
c Transplantation Biology Research Center, Massachusetts General Hospital, Boston, Mass
d Revivicor, Blacksburg, Va.
Read at the Eighty-sixth Annual Meeting of The American Association for Thoracic Surgery, Philadelphia, Pa, April 29–May 3, 2006.
Received for publication April 28, 2006; revisions received November 10, 2006; accepted for publication November 20, 2006. * Address for reprints: Richard N. Pierson III, MD, Division of Cardiac Surgery, University of Maryland, 22 S Greene St N4W94, Baltimore, MD 21201. (Email: rpierson{at}smail.umaryland.edu).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2 hours) relative to wild-type (<15 minutes) and swine lungs expressing human decay accelerating factor (<1 hour). In this study, the in vivo behavior of galactosyl transferase knockout lungs was evaluated. Methods: Three galactosyl transferase knockout swine left lungs were transplanted into baboons in a life-supporting model. One baboon lung allograft and two swine lung xenografts transgenic for human membrane cofactor protein (CD46) served as controls.
Results: Whereas two membrane cofactor protein lungs exhibited high pulmonary vascular resistance (>500 mm Hg · min/L) and failed to support life within 21 minutes, two of three galactosyl transferase knockout lungs supported life, for 90 and 215 minutes, and displayed low peripheral vascular resistance (48 ± 12 mm Hg · min/L at 60 minutes), similar to the allogeneic control. Complement activation (delta C3a < 250 ng/mL through 60 minutes) and C5b-9 deposition were minimal in both galactosyl transferase knockout and membrane cofactor protein lungs. Neutrophils, monocytes, and platelets were rapidly sequestered in galactosyl transferase knockout and human membrane cofactor protein lung recipients, unlike the allogeneic control (<20%); and thrombin formation (delta plasma fraction 1+2 > 0.5 nmol/L) was seen in the galactosyl transferase knockout recipients. Platelet activation (ß-thromboglobulin rise > 200) and appearance of capillary congestion and vessel thrombosis confirmed coagulation activation associated with galactosyl transferase knockout lung failure.
Conclusions: Galactosyl transferase knockout swine lungs are significantly protected in vivo from the physiologic consequences (increased pulmonary vascular resistance, capillary leak) associated with hyperacute lung rejection. As during ex vivo perfusion, dysregulated coagulation—thrombin elaboration, platelet activation, and intravascular thrombosis—mediates galactosyl transferase knockout lung xenograft injury.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
although hyperacute rejection has remained particularly problematic. To this end, a variety of transgenic swine have been engineered in the hope of abrogating hyperacute rejection by inhibiting complement activation. Although this approach has been met with some success in human complement regulatory protein transgenic kidneys and hearts,6-8 protection of human decay accelerating factor-positive (hDAF) (+) lungs was disappointing.4,8-12 In our prior experience with three hDAF transgenic swine lungs transplanted into cynomolgus monkeys,13 life-supporting xenogenic lung function was rarely observed for more than a few minutes. The combination of an hDAF (+) lung, C1 esterase inhibitor (Behrinert-P; Behringer, Mannheim, Germany), thromboxane and histamine inhibition, methylprednisolone, and inhaled nitric oxide was also tried in three baboon lung xenograft recipients without dramatic success. In one case, additional use of 
Gal-column immunoadsorption daily for 3 days before transplant yielded 7 hours of life-supporting survival in a baboon lung transplant recipient without any inotrope requirement.14 The lung in this recipient ultimately had marked perivascular edema and lymphatic dilation in the lobular septa, with intravascular thrombi, and histologic characteristics in small arterioles evidencing endothelial activation. This approach was not successful when applied in two additional animals that had previously been sensitized during rejection of a previous swine heart graft.
Because the carbohydrate structure Gal1-3Gal is the antigen recognized by over 80% of anti-swine antibody naturally found in human beings, genetically modified galactosyl transferase knockout (GalT-KO) swine have been developed that do not express this moiety.15,16 Although heart and kidney transplants from GalT-KO swine appear to be reliably protected from hyperacute rejection,17,18 our own results in ex vivo lung perfusion studies have been disappointing.19 Although physiologic transpulmonary blood flow, low peripheral vascular resistance (PVR), and markedly prolonged survival compared with wild-type and hDAF lungs were observed, lung failure nevertheless occurred after an average of 2 hours resulting from loss of pulmonary vascular barrier function (pulmonary edema) and/or intravascular thrombosis.
Although our ex vivo lung perfusion model is widely established and simulates most physiologic in vivo conditions, exposure of blood to artificial surfaces and concomitant intensive anticoagulation introduce potentially confounding variables. Thus, this in vivo study, the first to describe GalT-KO swine lung function in nonhuman primates, was designed to determine whether the observations ex vivo predict in vivo performance, and specifically to determine whether a GalT-KO lung xenograft is able to support the life of a nonhuman primate.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Lung Transplant Model
GalT-KO lungs were harvested remotely either in Boston or Pittsburgh and transported with a cold ischemic time of approximately 6 hours. Two hMCP swine lung donors and one baboon lung donor were sedated with 20 mg/kg ketamine and 1 mg/kg xylazine (Phoenix Pharmaceutical Inc, St Joseph, Mo) and given diphenhydramine (Benadryl; Pfizer, Morris Plains, NJ) 10 to 25 mg and ranitidine 2 mg/kg. Anesthesia was induced and maintained with isoflurane 1% to 3%. After median sternotomy and vascular isolation, heparin was administered (500 IU/kg), and an aortic cannula was secured in the pulmonary artery. Next, 1-benzylimidazole (5 mg/kg, a thromboxane synthase inhibitor; Sigma No. 116416, St Louis, Mo) and prostacyclin analog (Flolan 0.03 mg/kg; GlaxoSmithKline, Research Triangle Park, NC) were injected by slow bolus infusion into the right heart. Cardiac inflows were occluded, and the lungs were flushed with cold (4°C) Perfadex (Vitrolife, Kungsbacka, Sweden) containing Flolan 0.5 mg/L and nitroglycerin 10 mg/L. Cold saline slush was applied to the lungs. After excision of the heart–lung block, the left lung was prepared for transplant into the baboon, and the right lung was prepared for ex vivo perfusion with heparinized fresh human blood as previously described.19
Baboon recipients were sedated with 10 mg/kg ketamine, intubated, and maintained under general anesthesia with isofluorane 1% to 3%. A femoral arterial line, a femoral vein sheath, and additional intravenous access were secured. Heparin was bolused at 70 IU/kg and then infused at 200 U/h to maintain activated clotting time greater than 200. The recipients were treated with diphenhydramine (0.50–1 mg/kg), ranitidine (2 mg/kg), and 1- benzylimidazole (7 mg/kg). Nitric oxide (INO Therapeutics, Clinton, NJ) was delivered to the inspiratory limb of the ventilator circuit and calibrated to deliver 40 ppm. One recipient of an hMCP lung was additionally treated with bivalirudin (Hirulog; The Medicines Company, Cambridge, Mass) 1 mg/kg bolus before graft revascularization and continuous infusion of 2.5 mg · kg–1 · h–1. A clamshell incision was made at the fourth intercostal space. A flow-directed catheter introduced from the femoral vein was positioned in the pulmonary artery with assistance of direct cardiac manipulation. The right pulmonary artery was isolated with a snared vessel loop, the left hilar structures dissected and occluded, and the left lung removed. The bronchial anastomosis was performed first, followed by arterial and venous anastomoses. Just before revascularization, methylprednisolone (SoluMedrol, 10 mg/kg; Pfizer-Pharmacia, Morris Plains, NJ), C1 esterase inhibitor (Behrinert, 50IU/kg; Behringer), and 1- benzylimidazole (2.5 mg/kg; Sigma) were given to the recipients by slow bolus infusion. Before and after graft reperfusion, blood and tissue biopsy specimens were collected serially at predetermined time points for in vitro analysis. Vascular flow probes (Transonic Systems Inc, Ithaca, NY) on the aorta and left pulmonary artery continuously measured cardiac output and flow to the transplanted lung, respectively. Thirty minutes after graft reperfusion, flow to the right (native) lung was progressively occluded by snaring the right pulmonary artery to assess the capacity of the transplanted lung to support life. If the animal was unstable with right pulmonary artery occlusion, the right pulmonary artery snare was released, and hemodynamic support was initiated with titrated norepinephrine, vasopressin, and dopamine. The right pulmonary artery was resnared at 15- to 30-minute intervals for 90 to 120 minutes, and hemodynamics was recorded; if improvement in lung function was not observed on serial assessment, the experiment was terminated. Graft survival was defined as duration of life-supporting lung function.
Histology and Immunochemistry
Serial lung biopsy specimens after reperfusion and multiple terminal samples were obtained, trisected, and processed as follows. Formalin-fixed, paraffin-embedded sections were stained with hematoxylin and eosin for light microscopy and analyzed qualitatively for intravascular thrombosis and interstitial, alveolar, or larger airway edema and hemorrhage. Two other pieces were snap-frozen in liquid nitrogen, one without further manipulation and the other after gentle syringe instillation into a visible bronchiole of OCT diluted with saline using a small-gauge angiocatheter. Frozen tissue sections from OCT-infused biopsy specimens were assessed by immunohistochemistry as previously described4 with monoclonal antibodies against human C5b-9 (Dako, Copenhagen, Denmark) at 1:50, C4d at 1:50 (Quidel Corporation, San Diego, Calif), CD41 (Immunotech, Marseilles, France) at 1:100, Cy3-labeled goat-anti-mouse immunoglobulin G (Jackson Laboratories, West Grove, Pa), rabbit anti-von Willebrand factor antibody (Dako), and Cy2-labeled goat-anti-rabbit immunoglobulin G antibody (Jackson Laboratories). The deposition of platelets in lung tissue was judged by the staining intensity (from – to +++), taking into account the extent and intensity of staining from a standard control section.
Beta-thromboglobulin, Thrombin, and Complement Enzyme-linked Immunosorbent Assays
Beta-thromboglobulin (ßTG) and prothrombin fragments 1+2 (F1+2) were measured by commercial enzyme-linked immunosorbent assay (ELISA; Asserachrome-ßTG; Diagnostica Stago, Parsippany, NJ; and Enzygnost micro F1+2; Dade Behring, Marburg, Germany) in plasma samples collected in CTAD tubes (Becton Dickinson, Franklin Lakes, NJ; BD No. 367947) containing 0.109 mol/L buffered sodium citrate, 15 mmol/L theophylline, 3.7 mmol/L adenosine, and 0.198 mmol/L dipyridamole. C3a was measured by ELISA (C3a ELISA; Quidel, San Diego, Calif) in plasma samples stored in ethylenediaminetetraacetic acid at –70°C.
Flow Cytometry Staining of Platelet Activation
Blood samples collected in CTAD tubes were immediately fixed with HEPES-buffered saline plus 1% paraformaldehyde to prevent in vitro activation. Platelet activation was detected by detecting the expression of CD62P by flow cytometry as previously described.5 Owing to technical difficulties with sample processing and platelet staining protocols, these data are available for only the last two GalT-KO lung xenograft experiments and for the allograft study.
Hematologic Analysis
White blood cells, neutrophils, monocytes, and platelets were enumerated by both standard automated (Cell-Dyn; Abbott, Abbott Park, Ill) and manual differential techniques in blood samples collected in ethylenediaminetetraacetic acid.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
The right pulmonary artery was reclamped periodically over the next 90 minutes, but flow to the transplanted lung remained low, and the animal was put to death.
The hMCP lung recipient treated with bivalirudin exhibited blood flow to the transplanted left lung (0.45 L/min) that was 40% of cardiac output (1.1 L/min). When the right pulmonary artery was snared, the left lung accommodated the full cardiac output of 1.0 L/min, and the transplanted lung supported the recipient life for a total of 21 minutes. Ultimately, hypoxemia (arterial oxygen tension 35 mm Hg) was associated with bradycardia and hypotension and led to the termination of the experiment.
One GalT-KO lung (No. 1) failed to support life owing to high PVR. The other two GalT-KO lungs (Nos. 2 and 3) were life-supporting for 225 and 90 minutes, respectively. The recipient of GalT-KO No. 2 lung exhibited tracheal edema at 2 hours, associated with decline in arterial oxygen tension from 390 to 84 mm Hg; arterial oxygen tension remained low for the remainder of the experiment. PVR remained low and cardiac output stable for approximately 4 hours, when pulmonary flow dropped from 0.5 to 0.2 L/min. During the last 5 minutes of the experiment, systemic blood pressure dropped and the animal died. GalT-KO No. 3 lung survived 90 minutes with physiology similar to the lung allograft (PVR remained low at 100 mm Hg/L/min), but progressive tracheal edema impeding ventilation and oxygenation led to recipient death. Of note, this recipient had some native lung pulmonary edema before xenograft revascularization. The cause of the acute injury to the native lung in this instance is unknown.
Complement Activation
Relative to prolific complement activation seen during perfusion of wild-type lungs ex vivo (>1000 ng/mL),4 in vivo perfusion of five genetically modified lungs in this experience was associated with minimal increase in C3a (Figure 1, B). Lung C4d and C5b-9 deposition was minimal at all time points (not shown), suggesting that in vivo perfusion of these lungs with baboon blood did not result in significant complement activation systemically or in the swine lung.
Platelet Sequestration and Activation Markers
Platelet counts were minimally perturbed after graft revascularization and institution of life-supporting lung function by the allograft (Figure 2, A), and there was no detectable platelet activation (CD62P expression in circulating platelets, Figure 2, B; ßTG elaboration, Figure 2, C). In contrast, platelet counts fell by 18% to 60% after xenograft reperfusion and/or conversion to life-supporting graft perfusion, accompanied by varying degrees of activation in residual circulating platelets. hMCP expression or additional thrombin inhibition did not prevent platelet sequestration or activation. GalT-KO No. 1, which had minimal graft perfusion, exhibited little decline in platelet number after reperfusion and partial recovery during the observation interval. GalT-KO lungs 2 and 3 exhibited a sharp initial fall in circulating platelet number after xenograft reperfusion and a further subsequent decline over 90 to 225 minutes of life-supporting lung xenograft function. Activation of those platelets remaining in the circulation was minimal in the recipient of GalT-KO lung No. 2, and marked (78% at 90 minutes) in the recipient of GalT-KO No. 3. In the three experiments in which both CD62P expression and ßTG were assayed, results were concordant.
|
Blood Cell Counts
Neutrophil counts fell with high PVR and low transpulmonary blood flow through the MCP lung without bivalirudin treatment, but not in the allograft that had physiologic transpulmonary blood flow (Figure 3). Neutrophil counts declined 35% to 45% after reperfusion of GalT-KO lungs 2 and 3, which subsequently supported life, but not in the life-supporting bivalirudin-treated MCP lung or in GalT-KO lung 1, which did not support life. No consistent trend was manifest in association with institution of life-supporting function or subsequent graft failure. Circulating monocyte counts fell 30% to 70% during initial reperfusion of all grafts with physiologic blood flow (bivalirudin-treated MCP, GalT-KO Nos. 2 and 3, as well as the allograft) and tended to fall in association with xenograft failure except in the instance in which thrombin was inhibited.
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Similarly, we find that direct thrombin inhibition significantly improves transpulmonary blood flow and permits transient life-supporting function of a baboon by a swine lung expressing a human complement regulatory protein, hMCP. However, by itself, hMCP expression was not sufficient to support life in one experiment. As part of our ongoing studies using a different model, the right lungs of the GalT-KO No. 2 and No. 3 experiments were perfused ex vivo with human blood (manuscript in preparation). Interestingly, in these ex vivo experiments, both the tempo of graft failure (>6 hours in GalT-KO No. 2 and 56 minutes for GalT-KO No. 3) and mode of graft failure (sudden loss of transpulmonary graft flow versus loss of pulmonary capillary barrier function, respectively) were similar for each pair of lungs studied contemporaneously in both in vivo and ex vivo experiments. The histologic features of ex vivo perfused lungs also showed intravascular thrombosis at the time of graft failure. These preliminary observations suggest that our ex vivo model is useful for predicting in vivo lung xenograft behavior, while illustrating the importance of confirming hypotheses developed with ex vivo perfusion in a life-supporting model.
In outbred swine, Yeatman,9,10 Kulick,11 Gonzalez-Stawinski,20 and their colleagues have previously obtained prolonged lung function by inhibiting complement, adsorbing anti-swine antibody, and supporting animals with high-dose pressors. However, these models only rendered the recipient intermittently dependent on function of the graft.9-11,20 Although constructs such as this can provide useful mechanistic information, in vivo models that do not mandate continuous life support and ex vivo perfusion systems using nonphysiologic perfusion criteria (low flow, high pressure, or perfusates lacking critical formed blood elements such as platelets or nucleated cells) may not accurately predict functional capacity of the graft in vivo.21
On the basis of C3a elaboration and immunohistologic analysis, both hMCP and GalT-KO lung xenografts are well protected with respect to antibody-driven classic pathway complement activation. Although these genetic modifications are effective in blocking their targeted pathway, hyperacute rejection still occurs, presumably driven primarily by other, complement-independent pathways.22-24 Our current and prior observations suggest that efficient regulation of complement alone is inadequate to permit clinically useful function of porcine lung in a human blood environment.4,22-24
Platelet and coagulation activation occur consistently in the context of discordant organ xenotransplantation.25-28 Whereas the lung allograft recipient had minimal platelet sequestration and only a modest degree of thrombin formation, in all three recipients of physiologically reperfused lung xenografts 30% to 60% of platelets were sequestered from the circulation within 30 minutes after revascularization, and thrombin formation was also high in all three GalT-KO lung recipients. Intravascular thrombosis and tissue platelet deposition at the time of lung failure further strongly support our working hypothesis that coagulation activation is the cause of ultimate graft demise. However, since we have not examined potential differences in levels of donor-reactive anti-swine antibodies among the primate recipients, it remains possible that anti-swine antibodies (anti-Gal for MCP lungs and anti–non Gal for GalT-KO lungs) might serve as the primary initiator of the events that ultimately lead to lung injury. Nonetheless, it appears that, to achieve long survival of GalT-KO lungs, the etiology of coagulation dysregulation needs to be addressed independently from complement activation.
Coagulation cascade activation can be prevented either directly or indirectly at the level of thrombin formation and/or platelet activation. In this study, bivalirudin, a direct thrombin inhibitor, controlled thrombin production, limiting the rise in PVR, and prolonged life-supporting time. This replicates our previous experience ex vivo where bivalirudin significantly improved mean lung survival, from 8 to 77 minutes, even in the context of intact antibody and complement-driven mechanisms.29 Importantly, bivalirudin also consistently delayed complement activation, similar to combined platelet receptor blockade.22 We infer that endothelial thrombin formation, which occurs in the context of xenotransplantation even with quiescent endothelial cells,27 may be a crucial primary stimulus for the initiation of platelet aggregation and activation, and thus for complement cascade amplification during hyperacute lung rejection, and may be expected to occur even if antibody and complement are removed from the equation. Although thrombin is one important platelet activator, there are many other agonists/platelet receptor pairs that likely will need to be considered, including von Willebrand factor and glycoprotein Ib.
Although bivalirudin effectively blocked thrombin formation and intravascular thrombosis in the hMCP graft, lung failure was associated with interstitial and alveolar hemorrhage, indicative of profound loss of endothelial barrier function. If caused by antibody-mediated injury, as seems probable, GalT-KO lungs should be less susceptible to this phenomenon.
Peterson and associates30 demonstrated that human monocytes adhere to swine endothelial cells by binding to the -Gal antigen, which not only causes monocyte sequestration in the tissue but also acts as a trigger to upregulation of endothelial cells’ adhesion molecule, cytokine, and tissue factor expression. It is interesting that, in this study, we found sequestration of monocytes from circulation in all four functional lungs, including the allogeneic lung and the two GalT-KO lungs. However, more data are needed to confirm these observations and discern their physiologic significance.
In summary, we have shown that whereas human complement regulatory protein transgenic lungs failed to support life without additional treatment to the recipient, GalT-KO lungs were able to do so for as long as 3.5 hours. Our associated findings support our working paradigm that dysregulated coagulation participates directly in GalT-KO lung injury. Our model postulates that primate monocytes and porcine intravascular macrophages trigger local coagulation pathway activation through release of tissue factor when activated by antibody, complement, or local thrombin formation. Tissue factor in turn initiates and amplifies platelet aggregation as well as activation of endothelial cells and a plethora of other cell types in lung and blood. Dysfunctional thromboregulatory processes in the setting of xenotransplantation fail to abort amplification of several prothrombotic cascades,25-28 leading to prolific intravascular coagulation and concomitant complement activation. In our estimation, improved mechanistic definition of the interaction between these various triggers—antibody, complement, adhesion molecules, platelets, and other coagulation pathway constituents—will be necessary to devise efficient, practical strategies for aborting the various pathogenic positive feedback loops outlined here, and thus to develop clinically suitable approaches for lung xenografting.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 David Ayares reports equity options in Revivicor, the producer of knockout swine. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
). The legend in this figure applies to all following figures. The long-dash line represents the allogenic lung. Two short-dash lines stand for membrane cofactor protein (MCP) lungs, and three solid lines for galactosyl transferase knockout (GalT-KO) lungs. Open symbols represent nonfunctional lungs and filled symbols represent lungs that were able to support life.
