Extracorporeal liver and lung perfusion was performed with organs from four genetically modified pigs. To enable the perfusion, pigs lacking the major xenoantigen αGal (αGal-KO), that were created by knocking out the GGTA1 gene were used for this study [12]. Animals deficient for GGTA1 were generated by breeding of parental lines carrying a homozygous or heterozygous knockout of the GGTA1 gene. Offspring was tested by PCR using saliva or tissue biopsy. The knockout was confirmed on porcine peripheral blood mononuclear cells (PBMCs) by flow cytometry analysis prior to organ procurement. Organ procurement was performed according to German animal welfare guidelines using a non-heart-beating donor model at the Friedrich-Loeffler Institute of Farm Animal Genetics in Mariensee (Neustadt). During the procedure, liver and lung were perfused with PERFADEX^® Plus solution containing 2,000 I.U. Heparin in preparation for subsequent extracorporeal machine perfusion. For perfusion, the organs were transported on ice to the surgical research lab, at Hannover Medical School (MHH). The extracorporeal machine perfusion of lung and liver was performed on XVIVO XPS System (Serial number: XPS0132), and the XVIVO Liver Assist™ for up to 6 h. Tissue and perfusate samples were taken to monitor machine perfusion and organ-specific functions on a regular basis. Additionally, hyperspectral images (HSI) were captured prior to and during perfusion in the respective organs. Perfusion success, defined as maintained organ perfusion over time, was evaluated in consideration of the respective tissue, perfusate, bile and HSI data.

Without premedication, animals were electrically stunned and killed according to standard procedure. After animals were exsanguinated, the abdominal cavity was opened immediately and the abdominal aorta and vena cava were identified.

Following cannulation of both vessels, perfusion of the organs was started with 2 L of ice-cold PERFADEX^® Plus solution containing 2,000 I.U. Heparin. Simultaneously, the abdominal aorta was ligated shortly behind passing through the diaphragm. During in situ perfusion, a sternotomy was performed, and lung and liver were explanted according to the German guidelines of organ procurement [13]. Following explantation, both organs were perfused again at the side of retrieval with 2 L of ice-cold PERFADEX^® Plus containing 2,000 I.U. Heparin (lungs) or with Histidine-tryptophan-ketoglutarate/Custadiol^® (HTK, plus 2,000 I.U. Heparin) (liver) in preparation for extracorporeal organ perfusion.

Following liver procurement procedure and 2 h of transport period to the MHH at +4°C, the explanted liver was prepared for extracorporeal organ perfusion. Back table, the portal vein and hepatic artery were cannulated using a 24 Fr portal vain and 3,33 Fr hepatic artery cannula. A representative liver tissue sample from the median liver lobe was taken for subsequent histological analysis. Furthermore, the common bile duct was drained by inserting a 10 CH suction tube (Asid Bonz, Herrenberg, Germany). Simultaneously, the XVIVO Liver Assist™ was primed with two bags of human erythrocytes of type 0 Rh⁺ and Gelafundin to reach a total perfusion volume of 2 L. Heparin (25,000 I.U.), Ilomedin (20 μg/mL (1:10 dilution); 2 mL/h), Insulin (1 mL/h minimum), 2 L/min of oxygen and 0.5–1 L/min of CO₂ were added continuously for maintaining physiological perfusion conditions. The graft was connected to the XVIVO Liver Assist™ to perfuse the liver under physiological conditions with a portal vain flow of 550 mL to 450 mL/h and an arterial flow of 0.1–0.2 L/min for up to six hours. Hourly blood gas analysis was performed throughout the perfusion period to monitor the composition of the perfusion solution.

After approximately two hours of cold ischemia during transport to the MHH, EVLP was performed according to the Toronto protocol [14]. The system was first primed with 1.5 L of STEEN Solution™ (XVIVO Perfusion AB, Moelndal, Sweden) and a total of 10,000 I.U. of heparin. The flow rate was set at approximately 40% of the equivalent cardiac output and the temperature was set at 34°C. Next to priming the device, the lung was prepared for the perfusion by inserting a endotracheal tube with 9 mm inner diameter (Ruesch, Rommelshausen, Germany), and usage of the XVIVO Lung Canulla Set™ for the pulmonary artery and right atrium. A standard retrograde flush with 1 L PERFADEX^® Plus was performed in the grafts prior to connecting the organ to the EVLP. Then, recruitment was performed at least one time per run, with positive end expiratory pressure (PEEP) continuously increased to approximately 10 and maintained at 100% FIO₂ for a total of 10 min. Within the first run, recruitment was conducted for longer times. EVLP runs were proceeded from 60 up to 160 min and a maximum of 1 L STEEN Solution™ was added during procedure.

In two EVLP runs (2nd & 4th), additional simulation of extracorporeal perfusion using human blood was achieved by using one bag of human erythrocytes of type 0 Rh+ added to the circulation at the end of the respective perfusion time. This modified perfusion setting was conducted for up to 15 min.

During extracorporeal perfusion, the physical perfusion parameters were monitored continuously for both organs. For liver graft perfusion, additional samples of the perfusate were taken prior to perfusion and in two-hour intervals to monitor the metabolic function of the grafts (urea, creatinine, AST, ALT, GLDH, alkaline phosphatase, gamma-GT, total bilirubin, and ammonia). The sampling was conducted in our clinical laboratory. Additionally, a complete blood count was conducted prior to perfusion and again at 3 h and 6 h after initiating the perfusion to monitor any hematological changes.

HSI, using TIVITA^® 2.0 (TIVITA^® Tissue System, Diaspective Vision GmbH, Am Salzhaff, Germany), was conducted hourly during the perfusion. In addition, HSI was used during one exemplary run of EVLP prior to the perfusion and after each recruitment during perfusion STEEN Solution™ as well as alongside with the hemoperfusion.

The parameters hemoglobin oxygen saturation (StO₂; 1 mm deep, 500–600 nm), tissue hemoglobin index (THI, 500–600 nm), near-infrared (NIR; 4–6 mm, 700–1,000 nm) perfusion index and tissue water index (TWI, 900–980 nm) were recorded.

PBMCs were isolated from the blood of αGal-KO pigs, which were used as donors for perfusion experiments. PBMCs from a wildtype (wt) pig served as control. αGal epitopes were detected by staining with FITC conjugated Griffonia simplicifolia isolectin B4 (IB4-FITC). Analysis was performed on a FACS Calibur flow cytometer (Becton Dickinson, San Jose, CA, United States) and data were processed using FCS Express 7 (De Novo software, Pasadena, CA, United States).

Flow cytometry was employed to analyze perfusate samples during liver NMP to detect and quantify the release of porcine immune cells into the perfusate. Specific porcine antibodies detecting CD45 (Clone K252.1E4, Acris), CD3 (Clone BB23-8E6-8C8, BDBiosciences) CD21 (Clone LT21, Origene), CD4 (Clone 74-12-4, Acris), CD8 (Clone 76-2-11, Acris), CD14 (Clone MIL2, Serotec) and CD56 (Clone MEM-188, Biolegend) were used, allowing for identification and characterization of cellular components in the perfusate. For flow cytometry analysis, we used the BD FACSCalibur with CellQuest™ Pro software and FCS Express 7.

Histological examination was performed in both organs, on different time points. During lung perfusion, tissue samples (2 cm × 2 cm) of each lobe were taken before and at the end of the perfusion interval. From liver grafts, samples (2 cm × 2 cm) were taken prior to connecting the organ to the extracorporeal system as well as at the end of the perfusion interval. All histological samples were fixed in 10% methanol buffered formalin for 24 h at room temperature prior to further processing.

Dehydration using increasing concentrations of ethanol and embedding in Paraffin was performed according to standard procedure. Sections of 2 µm thickness were cut using a microtome and stained applying a standard Hematoxylin and Eosin (H&E) protocol.

Data export from the XVIVO XPS and Liver Assist™ systems was performed after each run. Perfusion data were then manually organized and processed. Statistical analysis was performed using appropriate methods for data validation and analysis to ensure reproducibility and significance of results. EVLP-Data were analyzed using R version 4.2.3 (2023-03-15), utilizing the ggpubr package to examine pulmonary arterial systolic pressure and left atrium oxygen partial pressure over the time during EVLP. Continuous variables were summarized using median values with interquartile ranges [IQR].

In total, four female genetically modified pigs (chronological order of experiments #1529, #1544, #1421, #1396), and aged between 20 and 45 months, with a weight range of 250 kg–350 kg, were used for the extracorporeal organ perfusion experiments. Prior to organ procurement, the flow cytometry analysis of peripheral blood mononuclear cells (PBMCs) verified the absence of αGal epitopes, in three of four animals. Despite a confirmed bi-allelic knockout of the GGTA1-gene (deletions of one base (Δ1) and five bases (Δ5) within the coding from of GGTA1), PBMCs of pig #1396 stained positive for αGal (Figure 1). Further genetic analysis revealed a third, most likely functional, copy of the GGTA1-gene (not shown, probably reflecting a copy number variation for this locus). Due to the size of the animals, the liver weight was between 2,800 and 3,200 g, weight of the lungs was not determined. No evidence of intrahepatic clotting or pulmonary edema prior to the perfusion was observed in any of the organs.

Extracorporeal liver perfusion could be maintained for up to 252.5 ± 86.7 min. The extracorporeal liver perfusion was conducted under physiological conditions, with the following portal venous and hepatic artery flows: #1529, arterial flow 483.3 ± 169.4 mL/min, portal vein flow 0.518 ± 0.178 L/min, #1396, arterial flow 130.4 ± 70.4 mL/min, portal vein flow 0.503 ± 0.024 L/min, #1544, arterial flow 13.3 ± 1.3 mL/min, portal vein flow 0.492 ± 0.031 L/min (Figures 2A, B). The vascular resistance of liver #1544 in the portal venous system branches was already high at the beginning of perfusion (VR of 4.65 vs. 0.3 in liver #1529) and increased by more than 25% in the first hour of liver perfusion. In consideration of the decreasing hemoglobin and hematocrit values found in the blood gas analysis, perfusion of this particular organ was terminated after 3 h. The liver of #1421 was discarded, due to initial perfusion problems during graft procurement after DCD.

Analysis of ALT and AST levels in three pigs over time during perfusion was performed (Figures 3A, B). ALT and AST levels increased over time in all three pigs, with pig #1544 showing the highest ALT and AST levels. The slope for pig #1544 is the steepest, indicating a faster increase in ALT and AST levels compared to the other pigs. ALT and AST levels increases during perfusion suggest liver stress or damage over time in all three pigs.

Also, a continuous increase in lactate levels was observed by blood gas analysis in all three perfusions. This rise in lactate suggests a shift towards anaerobic metabolism. Possible causes include hypoxia, insufficient perfusion into capillary structures or cellular stress or damage (Figure 3C).

The decline in hemoglobin levels, which were below the measurable threshold at termination of all three perfusions, is indicative of a substantial loss of erythrocytes.

Accelerated hemolysis could be due to mechanical stress of erythrocytes passing the perfusion pump or through contact with non-biocompatible surfaces within the perfusion apparatus. Additionally, the presence of porcine immune cells or the lack of human plasma in the perfusate could also lead to increased hemolysis.

The accompanying figure (Figure 3D), clearly shows a steady decline, ultimately reaching unmeasurable hemoglobin levels during perfusion. All measured clinical parameters underscore the physiological challenges encountered during the perfusion process.

Extracorporeal lung perfusion was maintained with STEEN Solution™ in all organs as planned for up to two hours. A significant increase of oxygen saturation in the left atrium (pO₂(LA)) was observed in each run during recruitment maneuvers (100% FiO₂) (Figures 4A–D). However, the pO₂(LA) returned to its initial level following these maneuvers and did not improve further after repeated recruitments (Figures 4A–D).

According to the Toronto protocol, left atrium pressure (PLA) was intended to be between 3 and 5 mmHg. Although there were some brief periods in runs 1–3 where the pressure exceeded 5 mmHg, a median left atrium pressure of 4 mmHg was maintained across all runs (first run: 4 (4–5); second run: 4 (4–5), third run: 4 (3–4; fourth run: 4 (4–5); p < 0.001) (Figure 4E).

During regular perfusion with STEEN Solution™, pulmonary arterial pressure (PAP) should remain below a cut-off of 20 mmHg. With the exception of the initial phase of the first run and after erythrocyte administration in the fourth run, pressure was maintained within the targeted range (Figure 4F). The median PAP across all runs was 10 (9–14) mmHg (first run: 12 (11–12); second run: 16 (15–17), third run: 9 (8–10; fourth run: 9 (9–10); p < 0.001) (Figure 4F). The addition of human erythrocytes in run 4 (pig #1396) caused an increase in PAP to 30 mmHg, which subsequently declined (Figure 4F). In congruence with PAP, pulmonary vascular resistance (PVR) increased from 100 to 1,500 dyn/s/cm⁻⁵ within 5 min (Data not shown) immediately after addition of one erythrocyte concentrate at the end of run 4 (Figure 4F). Simultaneously, the Horowitz-index decreased from 400 to 160 mmHg within 3 min, and remained steady over 15 min and subsequently increased to 300 mmHg (not shown). Within 15 min after reaching peak values, PAP and PVR decreased to normal levels.

Flow cytometry analysis demonstrated the presence of specific porcine immune cell populations, including NK cells (CD56), monocytes (CD14), T cells (CD4/CD8), and B cells (CD21)) in the perfusate during perfusion. FSC, SSC dot blots show no immune cells in the perfusate before perfusion, but significant amounts at 5 min and 3 h into perfusion (Figure 5A). The analysis indicated an enrichment of B and T cells and a decrease in monocytes after 3 h NMP compared to the 5 min time point (Figure 5B). Only few NK cells were detectable at both time points with no significant difference in NK cell numbers between the 5 min and 3 h time point (Figure 5B). The persistence of immune cells in the perfusate until late into the NMP suggests active immune cell release from the liver graft during the perfusion period.

HSI provided additional insights into the tissue condition during perfusion (Figures 6A–Y). The NIR index started at a baseline value indicating undisturbed perfusion of the liver. During the first 2 h, there was a noticeable increase in NIR values (Figure 6H). From the third hour onwards, NIR values decreased and plateaued at a low level at the end of the 6 h perfusion (Figure 6M). Initial THI measurements suggested a normal distribution of hemoglobin within the liver tissue. Over the 6-h period, the THI values remained within a consistent range, showing no signs of significant hemoglobin depletion or concentration (Figures 6D, I, N). Perfusion heat maps displayed uniform perfusion across superficial liver tissue, with high perfusion areas correlating with high oxygenation levels. When comparing StO₂ values at the start of the perfusion (0 h) with measurements at 2 h, an increase can be observed (0 h: central 41%/periphery 0%; 2 h: 49%/42%), at the peripheral measurement point. After 6 h, we observed decreased StO₂ values (35%/29%). NIR values at 2 h (37/26) versus 6 h NMP (5/0) also show a significant decrease. The same decrease is seen for TWI (2 h: 44/38 vs. 6 h: 36/32), but not for THI (2 h: 98/99 vs. 6 h: 93/100).

During EVLP, substantial differences caused by the addition of human erythrocytes to the cell-free STEEN Solution have been observed applying HSI. After addition of the erythrocyte concentrate, oxygenation and THI increased compared to only STEEN Solution™ (Figures 6Q, S, V, X). In contrast, TWI decreased in most areas (Figures 6T, Y). Area of venous congestion (*) could be identified exhibiting low StO₂ and high THI in lungs perfused with erythrocytes (Figures 6V, X).

The histological examination of peripheral liver tissue samples taken prior to perfusion (Figures 7A–C) and at the end of perfusion (Figures 7D–F), revealed shrunken periportal fields and partly collapsed sinusoidal space at 6 h perfusion. No significant necrotic areas were detected. Nuclear staining remained nearly unchanged as most cell nuclei were well-defined (Figures 7B, C, E, F). The imperfect perfusion described above, did not seem to cause extensive cell death or tissue damage in the liver.

Lung tissue exhibited intact lung architecture before and after perfusion with cell-free STEEN Solution™ (Figures 7G–N). Compared to lungs directly after harvesting, vessels and capillaries were free of erythrocytes (Figures 7K–N). Edema formation could not be observed. After addition of human erythrocytes, extravasation of erythrocytes into the airspace, (alveoli as well as bronchi) could be observed, indicating a loss of barrier function (Figures 7O–R).

In this study, we utilized genetically modified pigs that do not express αGal, aiming to reduce immunogenicity, a key target in xenotransplantation [15]. This approach of genetic modification, particularly the knockout of xenoantigens such as αGal, Neu5Gc and Sda, alongside the introduction of immunoprotective human transgenes, is well documented in the literature. For instance, Detelich et al. noted maximum liver perfusion times of 5–7 h in wild-type pigs, 4–6 h in αGal-KO/hCD55 pigs, and 8–14 h in pigs with multiple knockouts and transgenes. Here, we achieved satisfactory perfusion parameters in 3 out of 4 normothermic liver perfusions. We stopped perfusion at 6 h at the latest, but we achieved comparable perfusion times as in the αGal-KO/hCD55 group reported by Detelich et al. [15, 16].

Histological analysis revealed intact tissue, lacking major necrotic areas and well-defined hepatocyte nuclei after 6 h of perfusion, indicating that the liver’s structural integrity had been preserved. These encouraging results suggests that the present NMP method is capable of maintaining liver structures over extended periods.

Hyperspectral imaging demonstrated consistent perfusion during the initial hours. Despite a constant THI at the end of the perfusion period, StO₂, NIR and TWI decreased, indicating hypoxia or ischemia. A drop in StO₂ values has also been observed in other studies, often attributed to reduced oxygen delivery or increased oxygen consumption [17, 18].

The question of how long a liver can be safely perfused prior to irreversible damage needs to be addressed in future experiments. Therefore, further developments should involve optimizing the perfusion parameters to ensure consistent oxygen delivery throughout the perfusion period. Additionally, exploring other genetically modified pig lines could yield better outcomes.

Another significant observation was the release of porcine immune cells into the perfusate, including NK cells (CD56), monocytes (CD14), T cells (CD4/CD8), and B cells (CD21). This phenomenon of passenger leukocytes is also observed in lung perfusion, leading to the implementation of leukocyte filters in clinically used perfusion circuits [19]. Understanding the nature and impact of these immune cells is crucial for developing strategies to mitigate any adverse effects [20, 21].

The lung is particularly challenging for xenotransplantation, because there is no preclinical or clinical model demonstrating long-term survival after xenotransplantation. Even human lung transplantation remains suboptimal compared to other organs, with a median survival of approximately 7 years according to ISHLT data [22].

Here, by applying Steen solution we have successfully demonstrated reproducible cell-free perfusion of genetically modified porcine donor lungs exhibiting excellent functionality. This method forms the basis for our future preclinical research to enable xenogeneic lung transplantation. In pig-to-non-human primate (NHP) lung transplantations most lungs failed to function and showed rapid loss of barrier function and extravasation of human erythrocytes [23]. Although, tested only as an example and without functional human immune cells, platelets, antibodies or complement, PAP, PVR and extravasation of human erythrocytes increased after addition of human erythrocyte concentrates. In our opinion, specific interactions between porcine lungs and human erythrocytes are responsible for the extravasation of human erythrocytes, even though some damage to the endothelium will occur in any organ perfusion. Some mechanisms have been reported, such as binding of human erythrocytes by porcine macrophages using sialoadhesin [24]. We speculate that there may be other, as of today unknown, mechanisms or molecular incompatibilities responsible for these observations.

In general, the use of human erythrocyte concentrates in the perfusate instead of human whole blood, is the major limitation of this study. The absence of functional human complement, antibodies and immune cells might explain the good perfusion results achieved with organ from pig #1396 that appeared to express αGal in certain amounts on PBMCs, possibly through chromosomal recombination.

Our results show that the use of porcine DCD organs without premedication of the animal is feasible in this context, especially for lungs confirming previous experiments using wildtype pigs [25, 26]. Tissue integrity and functional recovery of retrieved organs could improve by further reducing cold and warm ischemia times and donor animal anti-coagulation [27].