The Effect of Normothermic Machine Perfusion on the Immune Profile of Donor Liver

Abstract

Background: Normothermic machine perfusion (NMP) allows viability assessment and potential resuscitation of donor livers prior to transplantation. The immunological effect of NMP on liver allografts is undetermined, with potential implications on allograft function, rejection outcomes and overall survival. In this study we define the changes in immune profile of human livers during NMP.

Methods: Six human livers were placed on a NMP device. Tissue and perfusate samples were obtained during cold storage prior to perfusion and at 1, 3, and 6 hours of perfusion. Flow cytometry, immunohistochemistry, and bead-based immunoassays were used to measure leukocyte composition and cytokines in the perfusate and within the liver tissue. Mean values between baseline and time points were compared by Student's t-test.

Results: Within circulating perfusate, significantly increased frequencies of CD4 T cells, B cells and eosinophils were detectable by 1 hour of NMP and continued to increase at 6 hours of perfusion. On the other hand, NK cell frequency significantly decreased by 1 hour of NMP and remained decreased for the duration of perfusion. Within the liver tissue there was significantly increased B cell frequency but decreased neutrophils detectable at 6 hours of NMP. A transient decrease in intermediate monocyte frequency was detectable in liver tissue during the middle of the perfusion run. Overall, no significant differences were detectable in tissue resident T regulatory cells during NMP. Significantly increased levels of pro-inflammatory and anti-inflammatory cytokines were seen following initiation of NMP that continued to rise throughout duration of perfusion.

Conclusions: Time-dependent dynamic changes are seen in individual leukocyte cell-types within both perfusate and tissue compartments of donor livers during NMP. This suggests a potential role of NMP in altering the immunogenicity of donor livers prior to transplant. These data also provide insights for future work to recondition the intrinsic immune profile of donor livers during NMP prior to transplantation.

Keywords: flow cytometry; immune profile; immunohistochemistry; liver transplantation; normothermic machine perfusion.

Figure 1

Illustration of normothermic machine perfusion set up and experimental design. (A) Schematic illustration- Perfusion of the circuit is propelled by a centrifugal pump; the perfusate passes through the heater and oxygenator where the perfusate is oxygenated by sweep gas that contains high concentration of O₂ and low concentration of CO₂; effluent perfusate from hepatic vein drains directly into the cannula which routes the perfusate to venous reservoir. (B) Experimental Design- Timeline of perfusion and sample collection. (NMP, normothermic machine perfusion).

Figure 2

Normal physiologic parameters in all 6 livers were maintained during normothermic machine perfusion. (A) Portal vein flow rate (liter/minute) remained stable. (B) Hepatic artery flow rate (liter/minute) remained stable. (C) Portal vein pressure (mmHg) remained stable. (D) Hepatic artery pressure (mmHg) remained stable.

Figure 3

Metabolic parameters improved or remained stable in all 6 livers during normothermic machine perfusion. (A) Lactate level (mmol/liter) decreased during perfusion. (B) Hepatic artery pH normalized during perfusion. (C) Aspartate aminotransferase level (unit/Liter) remained stable. (D) Alanine aminotransferase level (unit/Liter) remained stable. (* P<0.05; ** P<0.01).

Figure 4

Representative H & E histology of liver tissue. Liver 6 exhibited cirrhotic change. Livers 9, 10 and 11 exhibited variable degree of microsteatotic change. At the end of perfusion, histology from livers 6, 9, 10, and 11 showed completely viable tissue with no structural change. Livers 7 and 8 exhibited patchy subcapsular hepatocyte necrosis (black arrows) with viable liver parenchyma underneath (white arrows).

Figure 5

Composition of leukocytes of perfusates and liver tissues prior to perfusion and at the end of perfusion (N=6). Neutrophils were the most abundant cell types in both perfusates and tissues both prior to and at 6^th hour of perfusion. Percentages were obtained by dividing the count of each immune cell type by CD45+ cell count for each tissue type at each time point.

Figure 6

Percentage change of frequency of cells from NK cell lineage during perfusion. A percentage change of 0% suggests there is no change in the innate immune cell frequency from time point 0^th hour. Average values of percentage changes (red line) as well as standard deviations were calculated (vertical bar). A two-sided, paired t-test was performed to compare percentages of each time point to those at time 0H. (** P<0.01). (A) Perfusate NK cell frequency decreased at the 1^st hour of perfusion (-63.4 ± 17.8%) and remained decreased at the 6^th hour of perfusion (-70.7 ± 9.8%), without changes to tissue NK cell frequency. (B) Perfusate NK T cell frequency decreased at the 6^th hour of perfusion (-56.8 ± 28.5%) without changes to tissue NK T cell frequency.

Figure 7

Percentage change of frequency of cells from lymphocyte lineage during perfusion. A percentage change of 0% suggests there is no change in the innate immune cell frequency from time point 0^th hour. Average values of percentage changes (red line) as well as standard deviations were calculated (vertical bar). A two-sided, paired t-test was performed to compare percentages of each time point to those at time 0H. (* P<0.05; ** P<0.01). (A) Perfusate CD4 T cell frequency increased at the 1^st hour of perfusion (+218.9 ± 104.6%) and continued to increase at the 6^th hour (+472.6 ± 172.5%) without changes to tissue CD4 T cell frequency. (B) Perfusate CD8 T cell frequency increased at the 3^rd hour of NMP (+147.4 ± 111.3%) and returned to baseline at the 6^th hour without changes to tissue CD8 T cell frequency. (C) Perfusate B cell frequency increased at the 1^st hour of perfusion (+147.4 ± 127.5%) and continued to increase at the 6^th hour of perfusion (+288.5 ± 189.2%), while tissue B cell frequency increased at 6^th hour of perfusion (+26.3 ± 16.0%).

Figure 8

Percentage change of frequency of cells from granulocyte lineage during perfusion. A percentage change of 0% suggests there is no change in the innate immune cell frequency from time point 0^th hour. Average values of percentage changes (red line) as well as standard deviations were calculated (vertical bar). A two-sided, paired t-test was performed to compare percentages of each time point to those at time 0H. (* P<0.05). (A) No changes were observed in perfusate neutrophil frequency but tissue neutrophil frequency decreased at the 6^th hour of perfusion (-24.5 ± 21.3%). (B) Perfusate eosinophil frequency increased at the 1^st hour of perfusion (+249.3 ± 175.3%) and continued to increase at the 6^th hour of perfusion (+763.1 ± 533.4%), without changes to tissue eosinophil frequency.

Figure 9

Percentage change of frequency of cells from monocyte lineage during perfusion. A percentage change of 0% suggests there is no change in the innate immune cell frequency from time point 0^th hour. Average values of percentage changes (red line) as well as standard deviations were calculated (vertical bar). A two-sided, paired t-test was performed to compare percentages of each time point to those at time 0H. (** P<0.01). (A) Perfusate classical monocyte frequency decreased at the 1^st hour of perfusion (-91.6 ± 5.8%) and remained decreased at the 6^th hour of perfusion (-95.6 ± 4.3%), with no changes observed in tissue classical monocyte frequency. (B) Perfusate intermediate monocyte frequency decreased at the 1^st hour of perfusion (-98.3 ± 1.5%) and remained decreased at the 6^th hour of perfusion (-98.5 ± 2.7%), while tissue intermediate monocyte frequency decreased at the 1^st hour of perfusion (-53.2 ± 17.6%) and returned to baseline at the 3^rd hour of perfusion. (C) No changes were observed in perfusate non-classical monocyte frequency and tissue non-classical monocyte frequency.

Figure 10

Identification of Treg using IHC (DAPI+CD3+CD4+FoxP3+). (A) Representative Slides (Liver 6) from before (0^th hour) and following perfusion (6^th hour). DAPI (Blue), CD3 (Cyan), CD4 (Yellow), FoxP3 (Red). (B) Treg cell frequency out of total CD3+ T cells both before and following perfusion.