If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Lung remains the least-utilized solid organ for transplantation. Efforts to recover donor lungs with reversible injuries using ex vivo perfusion systems are limited to <24 hours of support. Here, we demonstrate the feasibility of extending normothermic extracorporeal lung support to 4 days using cross-circulation with conscious swine.
A swine behavioral training program and custom enclosure were developed to enable multiday cross-circulation between extracorporeal lungs and recipient swine. Lungs were ventilated and perfused in a normothermic chamber for 4 days. Longitudinal analyses of extracorporeal lungs (ie, functional assessments, multiscale imaging, cytokine quantification, and cellular assays) and recipient swine (eg, vital signs and blood and tissue analyses) were performed.
Throughout 4 days of normothermic support, extracorporeal lung function was maintained (arterial oxygen tension/inspired oxygen fraction >400 mm Hg; compliance >20 mL/cm H2O), and recipient swine were hemodynamically stable (lactate <3 mmol/L; pH, 7.42 ± 0.05). Radiography revealed well-aerated lower lobes and consolidation in upper lobes of extracorporeal lungs, and bronchoscopy showed healthy airways without edema or secretions. In bronchoalveolar lavage fluid, granulocyte-macrophage colony-stimulating factor, interleukin (IL) 4, IL-6, and IL-10 levels increased less than 6-fold, whereas interferon gamma, IL-1α, IL-1β, IL-1ra, IL-2, IL-8, IL-12, IL-18, and tumor necrosis factor alpha levels decreased from baseline to day 4. Histologic evaluations confirmed an intact blood–gas barrier and outstanding preservation of airway and alveolar architecture. Cellular viability and metabolism in extracorporeal lungs were confirmed after 4 days.
We demonstrate feasibility of normothermic maintenance of extracorporeal lungs for 4 days by cross-circulation with conscious swine. Cross-circulation approaches could support the recovery of damaged lungs and enable organ bioengineering to improve transplant outcomes.
Conventional ex vivo lung perfusion systems offer limited time for recovery and therapeutic intervention in extracorporeal lungs. We demonstrate that normothermic preservation of extracorporeal lung tissue structure and respiratory function can be maintained for 4 days using cross-circulation. This system could serve as a platform for lung bioengineering and organ recovery and regeneration.
Lung transplantation, the only life-saving intervention for patients with end-stage lung disease, remains limited by the shortage of usable donor organs. Although the number of patients waiting to receive a lung transplant continues to rise, only 20% of donor lungs meet functional criteria for transplantation.
Many of the conditions that render donor lungs unacceptable for transplantation (eg, aspiration, infection, and pulmonary contusions) are potentially reversible, but conventional methods of donor lung preservation rely on nonphysiologic cold static ischemia and preclude endogenous repair and recovery.
Ex vivo lung perfusion (EVLP) aims to address these limitations by providing initially unacceptable donor lungs with physiologic conditions—normothermia, perfusion, ventilation—to recover function outside the body to a level acceptable for transplantation.
However, EVLP has been unable to recover the majority of unusable donor lungs, likely due to the inability of an isolated single-organ support system to provide an appropriate physiologic milieu that enables endogenous repair.
Despite efforts to address the shortage of transplantable lungs, physiologic constraints limit lung perfusion and preservation times, and thus restrict opportunities for donor organ recovery and bioengineering. To overcome these limitations, our group previously used a swine model of cross-circulation (XC) to establish a lung support system that extended normothermic extracorporeal support to 36 hours and enabled statistically significant ex vivo recovery of severely injured lungs.
The use of an XC system to achieve ex vivo lung recovery mimics the clinical setting where in situ recovery of marginal quality lungs is achieved in patients placed on extracorporeal membrane oxygenation support for several days after transplantation.
Motivated by this clinical practice, we hypothesize that extending the duration of extracorporeal support from hours to days could not only enable the recovery of damaged lungs not currently salvageable using EVLP systems but also enable investigation of bioengineering strategies to improve or personalize organs before transplantation.
In this proof-of-feasibility study, we developed an extracorporeal lung support system capable of maintaining lungs for 4 days using XC (Figure 1). A notable difference between this study and our previous studies, wherein swine recipients remained under anesthesia for the duration of extracorporeal support and recovery, is that we established a configuration in which swine recipients remained conscious throughout the procedure. Such an approach avoids the adverse effects of anesthetic agents and recipient immobility, provides access to nutrition ad libitum, and models a translational setup of XC between a patient and a donor organ. Throughout 4 days of normothermic support, all extracorporeal lungs and conscious swine recipients were subjected to longitudinal analyses to assess the safety and stability of the organ support system.
This study received approval from the Institutional Animal Care and Use Committee at Columbia University. In this proof-of-feasibility study, we investigated healthy swine lungs (n = 3) as a reproducible experimental input to assess the ability of the extracorporeal lung support system to maintain the structure, function, and integrity of extracorporeal lungs for 4 days, and to establish baseline values and methodologies. The mean total normothermic extracorporeal support time of all procedures was 100.7 ± 1.2 hours.
Six Yorkshire swine (3 donor–recipient pairs, aged 4-6 months) were used in this study. Swine lung donors had a mean weight of 43.3 kg (range, 35.7-57.0 kg), and swine recipients had a mean weight of 53.0 kg (range, 41.5-59.0 kg). No animals died in the course of this study.
Donor Lung Procurement
Swine lungs were procured in standard fashion as previously described.
The aortic arch, serving as an endothelialized biobridge between the lungs and the extracorporeal circuit, was secured to the left atrial cuff with a running 6–0 polypropylene suture. A 36F venous drainage cannula was secured to the biobridge with a 2–0 braided polyester tie. Lungs were placed in the organ preservation chamber in prone position in a sterile, double-lined organ basin containing warm normal saline.
Recipient Swine Cannulation
Recipient swine underwent general anesthesia after intramuscular induction with tiletamine (5 mg/kg). Cefazolin (30 mg/kg) and enrofloxacin (5 mg/kg) were administered before skin incision and re-dosed every 8 and 24 hours, respectively. Immunosuppression was administered intravenously: tacrolimus (5 mg/kg), mycophenolate mofetil (500 mg), each re-dosed every 12 hours, and methylprednisolone (125 mg), re-dosed every 8 hours. After exposing the right internal jugular vein, a heparin bolus (15,000 U) was administered, and the vein was cannulated with a 19F to 23F dual-lumen cannula (Avalon Elite; Maquet Cardiopulmonary, Rastatt, Germany) (Figure 2, A and B).
XC and Extracorporeal Lung Support
Calcium gluconate (1 g) was administered intravenously to recipient swine, and cross-circulation of blood between recipient swine and extracorporeal lungs was initiated, as previously described.
The extracorporeal circuit contained a pump console (Jostra HL-20; Maquet Cardiopulmonary), disposable pump (Rotaflow Centrifugal Pump; Maquet Cardiopulmonary), and continuous monitoring software (VIPER; Spectrum Medical, Cheltenham, England). Circuit flow rate was maintained within a protective regime between 5% and 10% of the estimated cardiac output of recipient swine, with pulmonary artery pressures <20 mm Hg, and pulmonary vein pressures between 3 and 5 mm Hg (Figure E1, A).
Extracorporeal lungs were ventilated (Oxylog 3000 plus; Dräger, Lübeck, Germany) (Video 1) with the following settings: respiratory rate, 6 to 8 bpm; tidal volume, 6 to 8 mL/kg; positive end-expiratory pressure, 5 cm H2O; inspired oxygen fraction, 40%; and maintained on XC for a mean duration of 1.5 ± 0.1 hours before initiating recovery of recipient swine from general anesthesia.
Management of Recipient Swine During Recovery From Anesthesia
Anesthetized recipient swine were transferred into a Panepinto sling suspended within a custom enclosure (Figure E2) and allowed to recover from anesthesia for a mean duration of 2.1 ± 1.3 hours (Figure 2, C and D). Following transfer to the Panepinto sling, recipient swine were weaned from general anesthesia and administered ketamine (0.5-3.0 mg/kg) and dexmedetomidine (0.1-0.4 mg/kg/h) as needed. When spontaneous breathing was achieved, recipient swine were extubated while suspended in the Panepinto sling, and subsequently lowered onto the floor of the custom enclosure, where they were allowed to fully recover from anesthesia (Video 2). Swine recipients were maintained in the custom enclosure for 4 days (Figure 2, E, and Video 3).
Analyses of Extracorporeal Lungs and Recipient Swine
Multiscale analyses of extracorporeal lungs were performed every 24 hours. Recipient swine were continuously monitored, and hemodynamic and biochemical parameters were recorded every 12 hours. Detailed methods are available in Appendix E1.
Extracorporeal lungs were connected to the normothermic support system and periodically evaluated over 4 days (Figure 1). Recipient swine were continuously monitored through behavioral observation, vital signs, and blood samples analyses.
Extracorporeal Circuit Stability
All extracorporeal circuit parameters were maintained within target lung-protective ranges throughout 4 days of normothermic support. Pulmonary artery pressures remained below 25 mm Hg (Figure 3, A); and the transpulmonary pressure gradient, the difference between pulmonary artery and vein pressures, was maintained within the target range of 5 to 15 mm Hg (Figure 3, C). Flows were maintained on average at 0.28 ± 0.03 L/min (8%-9% of estimated cardiac output
) (Figure 3, B), and the circuit had a mean temperature of 35.1°C ± 1.2°C throughout the 4 days (Figure 3, D). The pH of the perfusate stayed within the physiologic range of 7.42 ± 0.05 (Figure 3, E), and lactate remained below 2 mmol/L until day 4, when lactate increased slightly to 3.10 ± 0.14 mmol/L (Figure 3, F).
Recipient Swine Safety and Stability
All recipient swine tolerated venous neck cannulation (Figure 2, B), the experimental custom enclosure (Figure 2, E), and exhibited normal food and water consumption and excretion throughout all procedures. Safety and stability were assessed by monitoring of recipient swine hemodynamic parameters, which were maintained within normal ranges (mean heart rate, 112 ± 18 bpm and mean systolic pressure, 120 ± 29 mm Hg), and by hemogas analysis (pH day 0, 7.36 ± 0.02; pH day 4, 7.37 ± 0.08) (Table E1). Hematocrit decreased gradually over 4 days of support (day 0, 27.2% ± 16.6%; day 4, 10.7% ± 2.3%) due to repeated blood sampling from both recipient swine and the extracorporeal circuit, and from minor transient bleeding from repeated lung tissue sampling. The inflammatory response of recipient swine was evaluated by quantification of serum inflammatory cytokine levels. From baseline to day 4, mean serum concentrations of granulocyte-macrophage colony-stimulating factor, interferon gamma (IFNγ), interleukin (IL) 1β, Il-1ra, IL-6, IL-10, and tumor necrosis factor alpha (TNFα) variably increased, with the largest increases in IL-1β (10.4-fold), IFNγ (19.2-fold), and IL-1ra (69-fold). Mean serum concentrations of IL-1α, IL-2, IL-4, IL-8, IL-12, and IL-18 decreased, with the largest decrease in IL-8 (14.8-fold) (Table E2). All serum cytokine concentrations after 4 days of normothermic support were within or below ranges reported in swine EVLP studies that provided a maximum of 12 hours of support (Table E3).
Functional Maintenance of Extracorporeal Lungs
Respiratory function of extracorporeal lungs was preserved over 4 days of normothermic support. Robust gas exchange with mean arterial oxygen tension/inspired oxygen fraction values above 430 mm Hg (day 0, 439.4 ± 227.1 mm Hg; day 4, 548.5 ± 176.9 mm Hg) (Figure 4, A), and dynamic compliance with mean values above 20 mL/cm H2O (day 0, 22.2 ± 1.7 mL/cm H2O; day 4, 20.0 ± 1.0 mL/cm H2O) (Figure 4, B) were maintained consistently for the duration of all procedures. Mean peak inspiratory pressures increased slightly from day 0 to day 4 (day 0, 20.5 ± 0.7 cm H2O; day 4, 25.0 ± 4.2 cm H2O) (Figure 4, C), but always remained below 30 cm H2O for equivalent tidal volumes. All other functional parameters demonstrated minimal changes from baseline to day 4 (Table E4). Lung weight gradually increased over 4 days (day 0, 0.84 ± 0.17 kg; day 4, 1.18 ± 0.07 kg) (Figure 4, D), which was likely due to edema resulting from changes in hydrostatic pressure caused by variations in position of conscious swine recipients (eg, prone to standing) (Figure E1, B-D).
Multiscale Analyses of Extracorporeal Lungs
Gross photography of extracorporeal lungs showed normal appearance of the pleural surface with areas of localized consolidation periodically observed in upper lobes after day 2 (Figure 5 and Figure E3, A). Radiography confirmed that extracorporeal lungs remained aerated, with diffuse radiopacities in upper lobes after day 2 (Figure 5, C). Surface thermography revealed that lobes with lower surface temperatures were consistent with radiolucent areas (Figure 5, B), suggesting that consolidated regions in upper lobes had ventilation–perfusion mismatch with decreased ventilation leading to increased surface temperatures.
Bronchoscopy confirmed normal appearance of large airways with no evidence of airway edema, erythema, or secretions after 4 days (Figure 5, D). Although histologic evaluations revealed edema in upper lobes, structural preservation of lung parenchyma, pulmonary airways, and the vascular tree was confirmed throughout the lungs (Figure 5, E and F, and Figure E3, B-D). Transmission electron microscopy confirmed preservation of the blood–gas barrier with intact alveolar epithelial lining and abundant type II pneumocytes with normal cuboidal morphology (Figure 5, G).
Inflammation and Histopathologic Assessment of Extracorporeal Lungs
Airway inflammation was assessed by quantification of inflammatory cytokines in bronchoalveolar lavage (BAL) fluid. From baseline to day 4, mean concentrations of granulocyte-macrophage colony-stimulating factor, IL-4, IL-6, and IL-10 trended upward but did not increase drastically, and mean concentrations of IFNγ, IL-1α, IL-1β, IL-1ra, IL-2, IL-8, IL-12, IL-18, and TNFα decreased in BAL fluid (Figure 6, A-F, and Table E5 and E6). Notably, the largest increase of inflammatory cytokine concentrations in BAL fluid was IL-4 (5.5-fold), and the largest decrease was IFNγ (104.9-fold). To assess the degree of injury in extracorporeal lungs, tissue samples were subjected to blinded histopathologic review and assigned lung injury scores (Figure 6, G) according to an established injury scoring rubric (Figure E4, C). Polymorphonuclear cells, indicators of immune response to injury, remained low in airways but gradually increased in alveoli. Alveolar and interstitial edema increased slightly, which was consistent with the observed increase in lung weight. Nevertheless, the lack of significant increase in any lung injury category score from baseline to day 4 suggests that the extracorporeal lungs experienced minimal to no injury over 4 days of normothermic support.
Cellular Integrity and Function of Extracorporeal Lungs
Pentachrome staining confirmed preservation of bronchial structures, including airway mucosa, smooth muscle, cartilage (Figure 7, A), and bronchial epithelium with intact pseudostratified epithelium and airway cilia (Figure 7, B). In the respiratory zone, alveolar capillaries, and distal venules were well perfused (Figure 7, C), consistent with the outstanding respiratory performance at day 4. Immunohistochemical staining for vascular endothelial cadherin enabled visualization of the intact endothelial lining of pulmonary vessels (Figure 7, E). Following administration of nebulized methacholine on day 4, extracorporeal lungs demonstrated rapid bronchoresponsiveness (Figure 7, H) of airway smooth muscle (Figure 7, D). Cell viability in extracorporeal lungs was confirmed by uptake of viability marker carboxyfluorescein succinimidyl ester (Figure 7, F), and cellular metabolism in the parenchyma of extracorporeal lungs remained within range of the metabolic activity of lungs in vivo (Figure 7, G).
In this proof-of-feasibility study, we describe an extracorporeal organ support system capable of maintaining the structure, viability, and function of extracorporeal lungs for 4 days (Figure 8). This system extends the duration of normothermic extracorporeal lung support significantly beyond the capability of current ex vivo lung perfusion systems, from hours to multiple days. Our hypothesis is that homeostatic normothermic extracorporeal support for days to weeks could offer new opportunities for the assessment, recovery, and regeneration of donor lungs. Additionally, we established new methods to enable multiday XC of whole blood between extracorporeal lungs and a conscious large animal. Our results demonstrate feasibility that lungs can be maintained outside the body for 4 days with outstanding preservation of respiratory function, lung tissue structure, and cellular integrity and metabolism. Such a system offers a developmental platform for advanced therapeutic interventions such as gene or cell therapies for extracorporeal or intracorporeal organs.
Three new methodologies were developed in this study: a swine behavioral training program (Table E7 and Figure E5) implemented 2 weeks before the start of XC to acclimate recipient swine to the custom enclosure used during the procedure, a technique to manage full recovery of anesthetized cannulated recipient swine to consciousness using a Panepinto sling and custom enclosure, and a novel extracorporeal circuit configuration with a single-site dual-lumen cannula to enable maintenance of circuit parameters within acceptable ranges. Altogether, these methods may prove useful in future translational studies involving the connection of conscious large animals or humans to extracorporeal circuits for organ support.
The function of extracorporeal lungs in this study was robustly maintained throughout 4 days of normothermic support. In comparison to previous studies, wherein swine lungs were supported by EVLP systems for a total of 12 hours
lungs in this study at day 4 demonstrated superior mean arterial oxygen tension/inspired oxygen fraction (12 hours, ∼400 mm Hg; 24 hours, ∼400 mm Hg; XC 4 days, 548 mm Hg), equivalent mean compliance (12 hours, 20 mL/cm H2O; 24 hours, not reported; XC 4 days, 21 mL/cm H2O), and in-range mean peak inspiratory pressures (12 hours: 25 cm H2O; 24 hours, 15 cm H2O; XC 4 days, 25 cm H2O). IL-1β and IL-8 in BAL fluid decreased from baseline to day 4, and were only 1.4-fold and 2.2-fold higher, respectively, than in healthy swine lungs after 4 hours
of EVLP support. Furthermore, airway concentrations of IL-6, TNFα, and IFNγ after 4 days were, respectively, 2.2-fold, 2.8-fold, and 1260-fold lower. Altogether, these data suggest that the overall function and condition of swine lungs after 4 days were superior to the condition of lungs placed on 12 to 24 hours of EVLP support.
revealed that lungs after 4 days showed less injury in all categories except for the mild interstitial edema reflected by increased lung weight. Overall, the composite injury score at day 4 in this study (4.6) was markedly lower than composite injury scores reported for ischemia reperfusion injury (6.8) and gastric aspiration (9.3) studies. When compared with recipient swine lungs at the conclusion of the procedure, extracorporeal lungs showed only a minimal increase in mean injury score (day 4, 4.6; recipient lungs, 3.8) (Figure E6).
The maintenance of vascular pressures within physiological range is critical for the prevention of pulmonary edema and the preservation of extracorporeal lungs.
In this study, the pressure at the pulmonary veins was dependent on the hydrostatic pressure difference between the extracorporeal lungs and recipient swine, and was regulated by adjusting the height difference between the lungs and recipient swine (target, 10 cm). In our previous studies, this height difference was fixed, as recipient swine were anesthetized and therefore immobile for the duration of all procedures. In this study, recipient swine were conscious and free to stand upright or lay prone at will, which resulted in intermittent changes in the height difference between the extracorporeal lungs and recipient swine. Adjustments to the height of the lungs were therefore necessary to maintain the target height difference of 10 cm. Controlled adjustments of lung height were performed manually using a hydraulic lift, but were technically challenging to perform in real time. Consequently, variabilities in pulmonary vein pressures resulting from changes in swine position, in conjunction with the persistent prone position of extracorporeal lungs, likely contributed to the development of dependent interstitial edema, most notably observed in the upper lobes.
There are several limitations to the present study. This study involved a small number of procedures (n = 3). Although such a study size limits the opportunity for statistical analyses, the results demonstrated feasibility of normothermic support of extracorporeal lungs for 4 days. Future studies will investigate larger numbers of lungs, recipient swine, and a wider variety of experimental conditions including increased flow rates, recovery of damaged lungs, and therapeutic interventions. Further, because immunosuppression was used, immunological markers of injury may be different than in the absence of immunosuppression, which may account for differences observed between cytokine concentrations in BAL fluid and serum. Notably, several of the cytokines investigated in this study have been shown to have both pro- and anti-inflammatory roles,
so their functions in this system are difficult to interpret. Future studies using a multiday extracorporeal support system could help elucidate the roles of cytokines in extracorporeal lung support systems. Several technical challenges remain to be resolved in the current system: specification and regulation of a long-term extracorporeal organ environment, controlled variability of extracorporeal organ orientation, and appropriate ventilation and perfusion management strategies. The inability to strictly regulate dynamic hydrostatic pressure changes led to the development of edema, necessitating future development of feedback-regulated pump controls and an automated organ height adjustment system capable of responding precisely in real time to dynamic changes in swine position, transpulmonary pressure gradient, and lung weight. Although this study did not investigate deposition of recipient cells or platelets in extracorporeal lungs, which could result in platelet-induced injury and chimerism, future studies of hematologic and immunologic interactions will be critical to assess the safety of clinical translation.
Despite these limitations, this study demonstrated that cross-circulation enables a quality and duration of extracorporeal lung support not previously shown by EVLP systems. The use of XC to recover damaged lungs ex situ could be applicable in clinical settings where patients receiving extracorporeal membrane oxygenation fail to match suitable donor lungs. In such patients, XC of lungs with reversible injuries for several days could enable functional assessment, lung-protective strategies, and graft recovery while avoiding the physiologic insult associated with major surgical intervention and severe primary graft dysfunction. Lungs recovered by XC would then be transplanted into the patient, thereby potentially decreasing the morbidity and mortality associated with transplantation of injured lungs. Future investigations using multiday extracorporeal organ support could also enable advanced interventions through immunomodulation,
and ultimately serve as a platform to improve transplant outcomes.
In this study we demonstrate the longest duration of normothermic support of extracorporeal lungs reported to date (4 days), with outstanding maintenance of lung tissue and respiratory functions. We envision that this system could be applicable in clinical settings to recover and regenerate damaged donor organs, and in translational research settings as a platform to investigate new strategies for lung bioengineering.
Drs Guenthart, O'Neill, Vunjak-Novakovic, Bacchetta, and Mr Fung have a pending patent for a cross-circulation platform for recovery, regeneration, and maintenance of extracorporeal organs. All other authors have nothing to disclose with regard to commercial support.
The authors thank the Institute of Comparative Medicine veterinary staff, including A. Romanov, S. Robertson, R. Ober, A. McLuckie, G. Geist, N. Herndon, S. Hastings, D. Ordanes, and A. Rivas for supporting animal studies; Weill Cornell Microscopy and Image Analysis Core Facility staff, including L. Cohen-Gould and J. P. Jimenez for transmission electron microscopy imaging services. The authors also thank the Herbert Irving Comprehensive Cancer Center Molecular Pathology Shared Resources, including T. Wu, D. Sun, and R. Chen for histology services; E. Lopes, G. Pierre, and I. Fedoriv for support with technical analytics; and S. Pistilli and S. Halligan for administrative support.
Appendix E1. Details of preprocedure and procedure methods
Preprocedure Behavior Training of Recipient Swine
Before initiation of cross-circulation, a swine behavior training program (Table E7) was implemented to acclimate recipient swine to the custom enclosure (Figure E2). Target training (ie, conditioning swine to touch nose to target when the trainer activated an audible clicker) was initiated at least 14 days before the start of cross-circulation and was used to encourage recipient swine to enter the custom enclosure by day 4 of training (Figure E5, A). Initially, swine were maintained in the custom enclosure for up to 1 hour. Through days 5 to 14 of training, recipient swine were maintained in the custom enclosure for at least 2 hours per day to increase comfort and familiarity with the experimental environment. The width of the custom enclosure was incrementally reduced to limit rotational movements (Figure E5, B). Behavior training was essential to minimize stress and maximize comfort of recipient swine, and thus decrease risk of decannulation during cross-circulation procedures. Throughout the preprocedure behavior training program, recipient swine experienced enrichment with manipulata (Figure E5, C), social interactions with procedure personnel (Figure E5, D), and standard and high-quality edible treats.
Management and Monitoring of Recipient Swine
During cannulation of the right internal jugular vein, physiologic parameters of anesthetized recipient swine, including heart rate, electrocardiogram, blood pressure (cuff and arterial line), oxygen saturation, end-tidal carbon dioxide, temperature, and respiratory rate were continuously monitored and recorded. Following recovery from anesthesia and throughout the duration of procedures, recipient swine were continuously monitored. Physical examinations to monitor and record respiratory signs (ie, dyspnea, tachypnea, coughing, and hypoxia), change in appetite, abnormal attitude or mentation, and signs of pain or discomfort were conducted by a large-animal veterinarian and animal behavior specialist at least twice daily. Throughout the duration of cross-circulation, recipient swine were maintained on a continuous intravenous heparin infusion (initial rate, 25 U/kg/h). Activated clotting time was measured and recorded hourly, and heparin infusion rates were titrated accordingly to remain within the target activated clotting time range of 200 to 300 seconds. Recipient swine were allowed to eat, drink, sleep, and play freely in the custom enclosure. Twice daily, recipient swine received food (Laboratory Mini-Pig Grower Diet; LabDiet) and water ad libitum. To prevent interference with recipient swine activities, extracorporeal circuit tubing was secured at the top of the custom enclosure. One unit of whole blood was collected from each donor swine and was stored at 4°C. A transfusion was given to the recipient swine if hemoglobin decreased below 4 g/dL.
Before recovery of recipient swine from anesthesia, blood samples were collected from an auricular arterial line. Following recovery from anesthesia, blood samples were collected every 12 hours from a central venous line placed in the left external jugular vein. Hemogas analysis was performed using a point-of-care blood analysis system (epoc; Siemens Healthineers). Complete blood counts, basic metabolic panels, liver function tests, and coagulation panels were performed by a diagnostic laboratory service (Antech Diagnostics). Mycophenolate levels were not measured, but FK506 (tacrolimus) levels were measured (Architect System; Abbott). Inflammatory cytokines (granulocyte-macrophage colony-stimulating factor, interferon gamma, interleukin (IL) 1α, IL-1β, IL-1ra, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, IL-18, and tumor necrosis factor alpha) in recipient swine serum were quantified in triplicate by multiplex cytokine array (Discovery Assay Pig Cytokine Array; Eve Technologies).
Extracorporeal Lung Monitoring
Blood samples were collected from sample ports at the pulmonary artery and pulmonary veins every 12 hours throughout the duration of normothermic support. Arterial oxygen tension/inspired oxygen fraction and dynamic compliance (Cdyn = tidal volume/(peak inspiratory pressure − positive end-expiratory pressure)) were calculated at baseline and every 24 hours. Lung weight was obtained every 24 hours using a scale (Denver Instrument Company) housed inside the organ chamber. Gross photographs of extracorporeal lungs were acquired using a high-definition camera (Hero 5; GoPro), and radiographs were acquired using a portable unit (PXP-16HF; United Radiology Systems) with images captured at 2.2 mAs and 90 kVp. Thermographs of extracorporeal lungs were acquired using an infrared camera (T430sc; FLIR).
Bronchoscopy and Bronchoalveolar Lavage Fluid Analysis
Bronchoscopic assessment of the airways was performed, and bronchographs were obtained at baseline and every 24 hours. Bronchoalveolar (BAL) fluid samples were collected by introducing a 3.8 mm flexible bronchoscope (aScope 3; Ambu) into subsegmental bronchi of the left and right lower lobes of extracorporeal lungs and injecting sterile normal saline (5 mL) with subsequent aspiration and collection of BAL fluid in a sterile specimen trap. BAL fluid samples were centrifuged at 3500 rpm for 10 minutes at 4°C. Supernatants were snap frozen in liquid nitrogen and stored at –80°C until further processing. Inflammatory cytokines (granulocyte-macrophage colony-stimulating factor, interferon gamma, IL-1α, IL-1β, IL-1ra, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, IL-18, and tumor necrosis factor alpha) in BAL fluid of extracorporeal lungs were quantified in triplicate by multiplex cytokine array (Discovery Assay Pig Cytokine Array; Eve Technologies).
Extracorporeal Lung Tissue Sampling
A surgical stapler (GIA Auto Suture; Covidien) was used to collect tissue samples from extracorporeal lungs at baseline and every 24 hours. Tissue sampling locations were predetermined using randomization software (randomizer.org) (Figure E4, A and B). Tissue samples were immediately fixed in cold phosphate-buffered 4% paraformaldehyde for 24 to 48 hours, embedded in paraffin, and sectioned at 5-μm thickness. All sections were stained for hematoxylin and eosin and Movat's pentachrome by histology services in the Department of Molecular Pathology at Columbia University Medical Center. All sections were examined under light microscopy and imaged with a slide scanner (SCN400; Leica).
Recipient Lung Tissue Sampling
After each experiment the recipient lungs were procured via a median sternotomy, and a surgical stapler (GIA Auto Suture) was used to collect tissue samples from recipient lungs from all 5 lobes. Tissue samples were processed in a similar manner to extracorporeal lungs described above.
Transmission electron microscopy
Lung tissue samples were fixed with 2.5% glutaraldehyde, 4% paraformaldehyde, and 0.02% picric acid in 0.1M Na-cacodylate buffer (pH 7.2). Samples were then postfixed with 1% osmium tetroxide in Sorenson's buffer for 1 hour, dehydrated, and embedded in Lx-112 (Ladd Research Industries). Sections (thickness, 60 nm) were prepared using an ultramicrotome, stained with uranyl acetate and lead citrate, and examined with an electron microscope (JEM-1200 EXII; JEOL). Images were captured with a digital camera (ORCA-HR; Hamamatsu Photonics).
An experienced lung transplant pathologist blinded to the study protocol performed all histopathologic review that included evaluation of lung tissue samples from upper, middle, and lower lobes. A previously reported lung injury scoring rubric (Figure E4, C) was used to assess the degree of lung injury.
Following deparaffinization, lung sections were subjected to boiling citrate buffer (pH 6.0) for antigen retrieval and blocked with 10% normal donkey serum in phosphate-buffered saline for 2 hours at room temperature. Next, primary antibodies (VE-cadherin: Abcam, ab232880; alpha smooth muscle actin: Abcam, ab5694) were diluted 1:200 and incubated for 2 hours. Secondary antibody (Abcam, ab150077) was diluted 1:300 and incubated for 2 hours at room temperature. Sections were subsequently mounted, and images were obtained using a fluorescent microscope (FSX100; Olympus).
Airway smooth muscle tone in extracorporeal lungs was assessed after 4 days of normothermic support by intratracheal delivery of nebulized methacholine chloride (4 mg). Changes in peak inspiratory pressures were recorded for 1 minute before and 5 minutes after delivery of methacholine chloride.
Cell viability assay
To assess cell viability in the parenchyma of extracorporeal lungs, carboxyfluorescein succinimidyl ester (eBioscience) was reconstituted in dimethyl sulfoxide at a concentration of 1.06 M. After 4 days of normothermic support, carboxyfluorescein succinimidyl ester was delivered via flexible bronchoscope into distal regions of left and right lower lobes of extracorporeal lungs and incubated for 15 minutes. Lung tissue samples were then collected, washed 5 times with phosphate-buffered saline, fixed in cold phosphate-buffered 4% paraformaldehyde for 48 hours, embedded, deparaffinized, mounted, and imaged using a fluorescence microscope (FSX100).
Metabolic activity assay
To assess cellular metabolism in extracorporeal lungs, tissue samples were collected from the parenchyma at baseline, day 2, and day 4 of normothermic support. Tissue samples (volume: 250 μL; triplicates) were finely minced, gently homogenized, and placed in a 96-well plate. Cell metabolism assay reagent (Alamar Blue; ThermoFisher) was diluted 1:10 in DMEM supplemented with 10% fetal bovine serum, and 100 μL Alamar Blue reagent was added to wells containing lung sample homogenates.
Alamar Blue reagent alone (100 μL) was added to wells containing no lung homogenate (negative controls). The multiwell plate was protected from light and incubated at 37°C with gentle shaking for 2 hours. Following incubation, absorbance was measured at 570 nm and normalized to absorbance at 600 nm. To obtain a benchmark comparison of metabolic activity to healthy swine lungs in vivo, fresh lung tissues were collected from healthy swine immediately following thoracotomy, and metabolic activity assays were performed on both extracorporeal swine lung tissue samples and fresh healthy swine lung tissue samples.
Table E1Safety and stability of recipient swine during multiday extracorporeal lung support. Analysis of recipient swine vitals, hemogas, biochemistry, coagulation, and electrolytes throughout 4 days of cross-circulation
Oxygen tension values at day 0 were obtained while anesthetized swine were intubated and mechanically ventilated with inspired oxygen fraction of 100%. Oxygen tension values at days 1 to 4 were obtained from a central venous catheter while conscious swine were breathing room air.
274 ± 34
45 ± 23
39 ± 16
30 ± 2
42 ± 3
Carbon dioxide tension (mm Hg)
57 ± 3.4
46 ± 3.3
50 ± 8.5
47 ± 8.7
42 ± 5.8
32 ± 1.2
33 ± 1.7
39 ± 15.9
34 ± 3.8
30 ± 3.0
1.42 ± 0.49
1.58 ± 0.38
1.08 ± 0.27
0.91 ± 0.31
3.1 ± 0.14
112 ± 77
199 ± 49
127 ± 33
177 ± 55
159 ± 38
26.1 ± 19.6
18.3 ± 4.2
12.3 ± 6.4
14.3 ± 1.9
9.5 ± 5.2
47 ± 11
67 ± 5
53 ± 14
65 ± 19
68 ± 12
1 ± 1
1 ± 1
1 ± 1
1 ± 1
3 ± 1
423 ± 222.4
307 ± 47.7
203 ± 97.1
118 ± 53.8
94 ± 61.6
8.1 ± 4.9
7.6 ± 0.8
7.9 ± 3.3
4.9 ± 0.6
3.1 ± 0.7
27.2 ± 16.6
25.7 ± 3.8
20.8 ± 3.9
16.5 ± 2.5
10.7 ± 2.3
47 ± 16.9
54 ± 13.6
49 ± 13.4
28 ± 19.0
37 ± 3.5
67 ± 12.3
61 ± 23.2
51 ± 23.4
40 ± 1.2
29 ± 3.6
1.33 ± 0.26
1.15 ± 0.13
1.01 ± 0.14
1.05 ± 0.06
1.74 ± 0.58
0.93 ± 1.04
13.37 ± 9.06
20.23 ± 8.93
22.37 ± 8.89
15.21 ± 3.83
Activated clotting time (s)
296 ± 83
222 ± 35
225 ± 65
250 ± 66
239 ± 47
Electrolytes and other
136 ± 3.8
141 ± 2.4
136 ± 1.0
136 ± 1.0
140 ± 2.8
7.9 ± 2.9
6.3 ± 2.5
5.6 ± 1.1
5.6 ± 1.1
5.0 ± 0.5
7.8 ± 0.8
7.2 ± 1.1
8.1 ± 0.8
8.1 ± 0.8
6.4 ± 1.4
11.4 ± 2.0
9.5 ± 0.8
9.2 ± 0.6
9.2 ± 0.6
10.9 ± 0.6
78.6 ± 47.1
78.1 ± 40.3
87.2 ± 61.5
64.1 ± 56.0
75.5 ± 63.8
14.9 ± 1.2
14.5 ± 0.8
17.0 ± 5.1
10.5 ± 7.0
10.4 ± 5.4
Values are presented as mean ± standard deviation. BP, Blood pressure; WBC, white blood cells; Hgb, hemoglobin; Hct, hematocrit; AST, aspartate transaminase; ALT, alanine transaminase; PTT, partial thromboplastin time; PT, prothrombin time.
∗ Oxygen tension values at day 0 were obtained while anesthetized swine were intubated and mechanically ventilated with inspired oxygen fraction of 100%. Oxygen tension values at days 1 to 4 were obtained from a central venous catheter while conscious swine were breathing room air.
Table E3Comparison of reported values of serum inflammatory cytokines in swine lung and ex vivo lung perfusion (EVLP) studies
Reported value (pg/mL)
0 900 ± 1500 1300 ± 1900 300 to 625 Not detected Not detected 67.3 ± 105.6 2201.8 ± 1746.9 5941 ± 5868.4
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate Control swine not exposed to PRCV 36 h XC 36 h XC, after 18 h cold static ischemia Baseline, before gastric aspiration 6 h after gastric aspiration 36 h XC, after gastric aspiration; Interventional treatments
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 4 h EVLP, after HCl aspiration; control (no treatment) 4 h EVLP, after HCl aspiration; Surfactant treatment Control swine Swine with PRRSV infection 6 h EVLP, after gastric aspiration; control (no treatment) 6 h EVLP, after gastric aspiration; aurfactant treatment 6 h EVLP, after gastric aspiration; lavage treatment 6 h EVLP, after gastric aspiration; surfactant + lavage treatment 36 h XC 36 h XC, after 18 h cold static ischemia Baseline, before gastric aspiration 6 h after gastric aspiration 36 h XC, after gastric aspiration; Interventional treatments 12 h EVLP
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 4 h EVLP, after HCl aspiration; control (no treatment) 4 h EVLP, after HCl aspiration; surfactant treatment Baseline, before EVLP 6 h EVLP Control swine Swine with PRRSV infection Control swine Swine with ventilator-induced lung injury Control swine Swine with PRCV infection 6 h EVLP, after gastric aspiration; control (no treatment) 6 h EVLP, after gastric aspiration; Sufactant treatment 6 h EVLP, after gastric aspiration; lavage treatment 6 h EVLP, after gastric aspiration; surfactant + lavage treatment 36 h XC 36 h XC, after 18 h cold static ischemia Baseline, before gastric aspiration 6 h after gastric aspiration 36 h XC, after gastric aspiration; interventional treatments 12 h EVLP, cellular perfusate 12 h EVLP, acellular perfusate 12 h EVLP
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 6 h EVLP Control swine Swine with PRRSV infection Control swine Swine with ventilator-induced lung injury 6 h EVLP, after gastric aspiration; control (no treatment) 6 h EVLP, after gastric aspiration; surfactant treatment 6 h EVLP, after gastric aspiration; lavage treatment 6 h EVLP, after gastric aspiration; surfactant + lavage treatment 36 h XC 36 h XC, after 18 h cold static ischemia Baseline, before gastric aspiration 6 h after gastric aspiration 36 h XC, after gastric aspiration; interventional treatments 12 h EVLP, cellular perfusate 12 h EVLP, acellular perfusate 12 h EVLP
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate Baseline, before EVLP 6 h EVLP Control swine Swine with ventilator-induced lung injury Control swine Swine with PRCV infection 36 h XC 36 h XC, after 18 h cold static ischemia Baseline, before gastric aspiration 6 h after gastric aspiration 36 h XC, after gastric aspiration; Interventional treatments 12 h EVLP
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 6 h EVLP Control swine Swine with PRRSV infection Control swine Swine with ventilator-induced lung injury 36 h XC 36 h XC, after 18 h cold static ischemia Baseline, before gastric aspiration 6 h after gastric aspiration 36 h XC, after gastric aspiration; Interventional treatments 12 h EVLP, cellular perfusate 12 h EVLP, acellular perfusate 12 h EVLP
Values are presented as mean ± standard deviation of cytokine concentrations in BAL fluid collected from both left and right lungs. GM-CSF, Granulocyte-macrophage colony-stimulating factor; IFNγ, interferon-gamma; IL, interleukin; TNFα, tumor necrosis factor alpha.
∗ Fold change represents the change (– or +) in cytokine concentration from day 0 to day 4.
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate Thoracotomy only Transplantation, after 6 h cold static ischemia 36 h XC, control (no injury) 36 h XC, after gastric aspiration; Interventional treatments
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 4 h EVLP, after HCl aspiration; control (no treatment) 4 h EVLP, after HCl aspiration; surfactant treatment 4 h cold static ischemia Transplantation, after 4 h cold static ischemia and 4 h EVLP Thoracotomy only Transplantation, after 6 h cold static ischemia Control swine Swine with PRRSV infection 36 h XC 36 h XC, after 18 h cold static ischemia 24 h after gastric aspiration, control (no treatment) 4 h EVLP, control (no injury) 4 h EVLP, after gastric aspiration 36 h XC, control (no injury) 36 h XC, after gastric aspiration; Interventional treatments
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 4 h EVLP, after HCl aspiration; Control (no treatment) 4 h EVLP, after HCl aspiration; Surfactant treatment Control swine Swine with PRRSV infection Control swine Swine with ventilator-induced lung injury 36 h XC 36 h XC, after 18 h cold static ischemia 24 h after gastric aspiration, control (no treatment) 4 h EVLP, control (no injury) 4 h EVLP, after gastric aspiration 36 h XC, control (no injury) 36 h XC, after gastric aspiration; Interventional treatments
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 4 h EVLP, after HCl aspiration; Control (no treatment) 4 h EVLP, after HCl aspiration; Surfactant treatment 4 h cold static ischemia Transplantation, after 4 h cold static ischemia and 4 h EVLP Thoracotomy only Transplantation, after 6 h cold static ischemia Control swine Swine with PRRSV infection Control swine Swine with ventilator-induced lung injury 36 h XC 36 h XC, after 18 h cold static ischemia 24 h after gastric aspiration, control (no treatment) 4 h EVLP, control (no injury) 4 h EVLP, after gastric aspiration 36 h XC, control (no injury) 36 h XC, after gastric aspiration; Interventional treatments 2 h EVLP, after 24 h cold static ischemia (control)
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate Thoracotomy only Transplantation, after 6 h cold static ischemia Control swine Swine with ventilator-induced lung injury 36 h XC 36 h XC, after 18 h cold static ischemia 36 h XC, control (no injury) 36 h XC, after gastric aspiration; Interventional treatments
2.5 ± 0.6 25.9 ± 4.6 0.5 to 1.5 20.6 × baseline 202 ± 252 46 ± 71 0.8 ± 0.5 to 2.7 ± 0.8 53.8 ± 13.3 to 96.6 ± 10.5 5.1 ± 8.9 to 32.5 ± 16.2 5.1 ± 8.9 to 61.7 ± 20.8
4 h EVLP, STEEN perfusate 4 h EVLP, STEEN + blood perfusate 4 h EVLP, Papworth–blood perfusate 4 h cold static ischemia Transplantation, after 4 h cold static ischemia and 4 h EVLP Thoracotomy only Transplantation, after 6 h cold static ischemia Control swine Swine with PRRSV infection Control swine Swine with ventilator-induced lung injury 36 h XC 36 h XC, after 18 h cold static ischemia 36 h XC, control (no injury) 36 h XC, after gastric aspiration; Interventional treatments
Table E7Preprocedure training program for recipient swine. Before initiation of cross-circulation between extracorporeal lungs and recipient swine, preprocedure behavior training was conducted for 13 days. On day 14, cross-circulation was initiated and continued through day 18
Arrival at animal housing facility where recipient swine were placed into standard housing enclosure.
Initiation of target training in swine housing enclosure. Swine conditioned to touch nose to target when trainer used clicker. Behavior reinforced with standard quality edible treats. Duration: 1 h daily.
Continuation of target training in housing enclosure. Introduction of custom enclosure within the standard housing enclosure area. Duration: 1 h.
Progression of target training to encourage swine to enter custom enclosure. Duration: 1 h. Swine placed into custom enclosure and transported to operating room. Duration: 1 h.
Continuation of daily decreases in width of custom enclosure to inhibit swine from rotating within the cage while remaining comfortable and allowing forward, backward, upward, and downward movements. Swine encouraged to remain within the enclosure using high quality edible treats. Duration: 2 h daily.
Transportation to operating room where swine were anesthetized, cross-circulation was initiated, and swine were placed into custom enclosure.
Active enrichment during multiday extracorporeal organ support studies using manipulata (ie, toys), social interactions with procedure personnel, and standard and high quality edible treats.
Supported by the National Institutes of Health (grant Nos. HL120046 , HL134760 , HL007854 , HL143733 , and EB027062 ), the Richard Bartlett Foundation , the Blavatnik Foundation , and the Mikati Foundation .