Impact of Three Methods of Ischemic Preconditioning on Ischemia-Reperfusion Injury in a Pig Model of Liver Transplantation

Abstract Background Ischemic preconditioning (IPC), either direct (DIPC) or remote (RIPC), is a procedure aimed at reducing the harmful effects of ischemia-reperfusion (I/R) injury. Objectives To assess the local and systemic effects of DIPC, RIPC, and both combined, in the pig liver transplant model. Materials and methods Twenty-four pigs underwent orthotopic liver transplantation and were divided into 4 groups: control, direct donor preconditioning, indirect preconditioning at the recipient, and direct donor with indirect recipient preconditioning. The recorded parameters were: donor and recipient weight, graft-to-recipient weight ratio (GRWR), surgery time, warm and cold ischemia time, and intraoperative hemodynamic values. Blood samples were collected before native liver removal (BL) and at 0 h, 1 h, 3 h, 6 h, 12 h, 18 h, and 24 h post-reperfusion for the biochemical tests: aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), creatinine, BUN (blood urea nitrogen), lactate, total and direct bilirubin. Histopathological examination of liver, gut, kidney, and lung fragments were performed, as well as molecular analyses for expression of the apoptosis-related BAX (pro-apoptotic) and Bcl-XL (anti-apoptotic) genes, eNOS (endothelial nitric oxide synthase) gene, and IL-6 gene related to inflammatory ischemia-reperfusion injury, using real-time polymerase chain reaction (RT-PCR). Results There were no differences between the groups regarding biochemical and histopathological parameters. We found a reduced ratio between the expression of the BAX gene and Bcl-XL in the livers of animals with IPC versus the control group. Conclusions DIPC, RIPC or a combination of both, produce beneficial effects at the molecular level without biochemical or histological changes.

Liver ischemia-reperfusion injury (I/R) is inevitable during liver transplantation surgery and it is the main cause of primary graft dysfunction, one of the most serious complications of this surgery in children and adults [1][2][3]. It is known that an approach to reduce the harmful effects of liver I/R is an ischemic preconditioning (IPC) procedure. It consists of producing brief periods of ischemia followed by brief periods of reperfusion prior to the prolonged ischemic insult, which should induce greater organ tolerance to prolonged ischemia and subsequent reperfusion [4,5]. Its beneficial effects are thought to result from the release of adenosine by the ischemic tissue promoting vasodilation, inhibition of platelet aggregation and neutrophil adherence, inhibition of endothelin synthesis and reactive oxygen species, in addition to increased production of nitric oxide [5].
IPC can be direct at the target organ or indirect (remote). Direct ischemic preconditioning (DIPC) has the disadvantage of causing mechanical stress to the main vascular structures of the organ [6]. In remote ischemic preconditioning (RIPC), the procedure is applied to another organ, with the ORIGINAL RESEARCH protective effect on the target organ being exerted by biochemical mediators activated at a distance and carried by the blood stream, without direct stress or trauma to the organ [7]. The effect of RIPC was first demonstrated on the myocardium of rats submitted to renal IPC [6]. So far, there is no consensus about the best method of RIPC, i.e., the number of I/R cycles, the effective I/R time required to trigger the protective stimulus, and the choice of the I/R site to maximize the beneficial effects of IPC with the least possible damage to the body. However, notwithstanding these unanswered questions, the short-term occlusion of the mesenteric artery has been proven of great importance, with positive effects on several organs [6].
Several studies demonstrate the beneficial effects of DIPC and RIPC. A recent study also showed that the combination of both types of IPC enhanced the protective effects of this procedure in a mouse liver transplant model [8].
The pig model has clear anatomical, morphological, and physiological similarities with humans. It also allows for easy and extensive monitoring and ensures the feasibility of the orthotopic liver transplantation, thus being useful to investigate issues that have direct clinical relevance [8,9]. The potential benefits of RIPC in transplants performed on medium-sized animals, the association of direct and remote ischemic preconditioning, and the comparative effects of these two methods have been reported in very few studies. Thus, the objectives of the present study were to evaluate the local and systemic effects of three methods of ischemic preconditioning in a pig model of liver transplantation. The animals were evaluated by biochemical, histopathological and molecular analyses.

Materials and methods
The experimental study protocol was approved by the Institutional Animal Use Ethics Committee. The funding body had no participation in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript. All procedures performed in the animals were in accordance with the ethical standards of our Institution. Twenty-four hybrid pigs of both sexes weighing on average 28.7 ± 2.4 kg were commercially obtained from the Company Granja RG (Suzano -São Paulo, Brazil). The animals were randomly divided into the following 4 groups (n = 6 each) and submitted to orthotopic liver transplantation. The number of animals in each group was based on previous similar studies [10,11]. The animals (donors and recipients) were fasted for 12 hours, then at 7 a.m. injected with intramuscular xylazine (2.0 mg/Kg) and ketamine (10.0 mg/Kg) as pre-anesthetic 15 minutes before anesthesia, which was induced with propofol (5.0-10.0 mg/Kg) and maintained with endotracheal intubation, 40% oxygen supply, and isoflurane (1.3 to 2.0%) in inspired air, along with a continuous intravenous infusion of fentanyl (0.05 µg/Kg/min). Catheters were introduced into the jugular vein for fluid infusion and central venous pressure (CVP) measurement, and into the carotid artery for invasive mean arterial pressure (iMAP) measurement and blood sampling for biochemical analyses. The catheters were sutured to the skin with mononylon sutures to prevent its release during the postoperative period. Recipient animals were continuously monitored until the end of the surgery and post-surgery recovery with electrocardiogram, oximetry, end tidal carbon dioxide monitoring (EtCO2), respiratory rate, and pressure measurements -CVP and iMAP.
According to the groups where they were allocated, the animals underwent one of two types of ischemic preconditioning (groups D and R) or both combined (group D + R). Direct organ preconditioning was performed by clamping the donor whole hepatic pedicle (portal vein, hepatic artery, and bile duct); indirect preconditioning was applied to the recipient gut, by clamping the superior mesenteric artery. Both types of preconditioning consisted of three cycles with 5 minutes' ischemia followed by 5 minutes' reperfusion. These periods were standardized in a previous pilot procedure, in which we verified that they did not provoke animal arterial hypotension.
Surgical procedures on donor and recipient animals were performed according to previously published technique by our laboratory. After a large experience, we standardized the porcine liver transplantation without the use of veno-venous bypass [10,11].
The liver grafts were perfused with Euro Collins solution through the aorta and the portal vein. The grafts were kept in this solution cooled with ice cubes until the implant procedure in the recipient.
After surgery, recipient animals remained extubated and conscious in private stalls in our laboratory for 24 hours, with catheters in the jugular vein and carotid artery for medication infusion and blood sampling. Blood samples were collected at the following time points: before native liver removal (BL) and periods after graft reperfusion: 0 h, 1 h, 3 h, 6 h, 12 h, 18 h, and 24 h. At the end of this period, the animals were anesthetized, intubated and connected to the mechanical ventilator for gut, kidney, lung, and liver biopsies. After that, the animals were euthanized with an overdose of inhaled anesthetic 5% isoflurane and intravenous administration of 10 mL/kg of 19.1% potassium chloride.
The following parameters were recorded: donor and recipient weight, surgery time, warm and cold ischemia time, and intraoperative hemodynamic values. Additionally, the following biochemical tests were performed: aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), creatinine, BUN, lactate, total and direct bilirubin.
Histopathological analyses were performed on 4 liver fragments obtained at the following time points: before removal from the donor (BL), immediately after release of the hepatic artery flow of the recipient (0 h), after 1 hour of reperfusion (1 h) and after 24 hours of reperfusion (24 h). For the other organs (gut, kidney, and lung), only one biopsy was performed after 24 hours of observation. Samples were fixed in neutral formalin, dehydrated and embedded in paraffin blocks, then sectioned (4 µm-sections) and stained with hematoxylin-eosin.
The liver injury was assessed based on endpoints analyzed and quantified according to the criteria described by Scheuer et al [12]. This score was used for its easy reproducibility and for describing the inflammatory response. Lung histology assessment was based on the modified VILI score, with 4 parameters evaluated and quantified to measure the degree of tissue damage [13]. Kidney injury was assessed using the Banff score with 4 parameters evaluated and quantified [14]. To assess histological damage to the gut, the Chiu score was used [15].
All the tissue sections were examined under light microscopy by 3 different blind readers who assigned scores to the identified injuries. For the statistical analysis, each section was assigned the average of these 3 scores.
For the molecular analyses, a fragment from each biopsy specimen was examined for expression of the apoptosis-related genes, i.e., BAX (pro-apoptotic) and Bcl-XL (anti-apoptotic), the eNOS (endothelial nitric oxide synthase) gene, and the IL-6 gene related to the inflammatory ischemia-reperfusion injury, using RT-PCR. To study the balance between proand anti-apoptotic gene expression, the BAX/Bcl-XL ratio for all organs was calculated at different time points.

Statistical analysis
Data were recorded and stored on a spreadsheet of the Stata statistical package. For the qualitative variables, the absolute and relative frequencies were calculated. For quantitative variables, the mean, standard deviation, median, minimum and maximum values were calculated and displayed in graphic format with the values.
Data with normal distribution regarding quantitative variables and gene expression were assessed using analysis of variance (ANOVA); differences between groups were identified using the Bonferroni test; and differences between timepoints in the groups were identified using the t-test. Histomorphometric (qualitative) data were compared using the nonparametric Kruskall-Wallis method.
The null hypothesis of equality of means was rejected when p < 0.05.

Results
The recipient weights and ischemia times in different groups were similar (p > 0.05, Fig. 1). The duration of warm ischemia, cold ischemia, and recipient surgery periods are shown in Table 1. We can verify that there were no differences among the groups, regarding these parameters.

Biochemical analysis
The values of the serum biochemical markers in the different groups throughout the experimental period are shown in Fig. 2.
Regarding serum AST levels, all groups showed similarly disperse patterns, with values increasing over the experimental time points and stabilizing between 12 and 24 hours. There was no difference between groups.  For gamma-GT, some groups showed great variation within the group at the experimental time points, but no difference was observed between groups (not included in the figure). Also, alkaline phosphatase levels showed no differences between groups over the experimental time points.
All groups showed an increase in creatinine values compared to baseline. At 3 h and 6 h, values in the R group were lower than in controls (p = 0.003 and p = 0.042, respectively).
There was no difference in arterial lactate levels between groups during the experimental period. All groups showed a consistent increase in the immediate post-transplant period, followed by a return to baseline values at the end of the experiment.
There was wide variation in total bilirubin levels in the R and D + R groups at 6 h, 12 h, 18 h, and 24 h, but the only significant difference observed in this variable was an increase in the D + R group compared to the controls at 24 h (p = 0.046). The direct bilirubin levels showed the same behavior as total bilirubin, i.e., wide variation within the groups, but with higher levels in the D + R group compared to the controls at 6 h (p = 0.033) (not included in the figure).

Histopathological analysis
The histopathological analysis of the liver revealed typical ischemia-reperfusion lesions found in major surgeries: lobular and inflammatory cells infiltrate, hepatic sinusoid hyperemia and hepatocellular necrosis. In the small intestine, cell lysis, Grünhagen spaces formation, enlargement of the distance between villi, dilated capillaries, presence of inflammatory cells, and destruction of the villi with intraluminal hemorrhage were observed. In the kidney, the typical lesions were capillary tuft retraction plus interstitial, tubule and vessel congestion. In the lung, alveolar congestion was observed and in some cases hemorrhage and leukocyte infiltrates inside the alveoli and bronchi (Fig. 3). Histopathological scores in the different tissues and groups are summarized in Table 2. Despite individual observations, no statistically significant differences were found in tissue scores among the groups.

Molecular analysis
The gene expressions in the liver, gut, kidney, and lung of the study animals are shown in Fig. 4. In the liver, concerning the BAX gene expression, the D + R group showed lower values at 24 h when compared to the controls (p = 0.039). The expression of the IL-6 gene in D group animals was higher at 24 h when compared to the R group (p = 0.001) and the D + R group (p = 0.02). The Bcl-XL values did not increase during the experimental period in the controls, while in the groups that received any preconditioning there was an increase in the gene expression toward the end of the experimental period; the D group and the D + R group showed higher values at 24 h (p = 0.034 and   In the gut, no differences in the expression of the BAX and IL-6 genes were observed in the groups at 24 h. However, the Bcl-XL gene expression was higher in the R group than in the controls (p = 0.004). In addition, the eNOS gene expression was higher in the D group animals when compared to the controls (p = 0.001) and the R group (p = 0.001). In the kidney, it was observed that there was no difference in the expression of the BAX and Bcl-XL genes in the various groups at 24 h. IL-6 in the R group showed higher values than in the controls (p = 0.004). Also, no difference between groups in the expression of the BAX, IL-6 and Bcl-XL genes in the lung tissue, although the eNOS gene expression was higher in the D group than in the R group (p = 0.007). Finally, the ratio of pro-apoptotic BAX gene expression to anti-apoptotic Bcl-XL gene expression was lower in the liver by 24 h in all groups receiving IPC (D, R, and D + R groups) when compared to the controls (Fig. 5).

Discussion
After the advent of the experimental evidence of positive effects of IPC in I/R injury, several clinical studies started to show conflicting results about the real benefits of this procedure in medical practice. Despite the evidence of beneficial effects of direct and remote IPC in clinical situations involving myocardial [16,17], renal [18], and neuronal ischemia [19], it is known that different tissue and organ sensitivities to ischemia are diverse so that such benefits cannot be generalized, mainly in renal [20] and liver transplantation surgeries [21].
Presently, there is no clear evidence in the literature about the real effects of IPC in the liver transplantation surgery. Besides, different modalities of IPC are impractical procedures that may further increase the already high complexity of transplantation surgery. So, although several experimental investigations have shown promising results of IPC in liver IRI of small animals [22][23][24][25], there are few studies with models of cold and warm ischemia in liver transplantation performed on medium-sized animals [26,27] and this motivated us to evaluate such effects utilizing a model that simulates the human condition.
We have been utilizing the porcine model in our laboratory because it mimics the clinical situation of liver transplantation [10,11,28,29]. In a previous study, utilizing a model similar to the present one, we observed that IPC resulted in partial attenuation of the harmful effects of I/R injury [11]. In the current study, we also aimed at identifying if the positive effects of IPC remained after a longer observation time, thus providing a rationale for its use in pediatric recipients.
While studies of IPC are not new, still there is no consensus about technical issues such as ischemia time and number of I/R cycles needed to effectively achieve protective effects. Therefore, in our model, we used 3 alternating 5-minute cycles of ischemia followed by the same reperfusion time, to prevent severe hemodynamic repercussions in the animals, while avoiding a too short IPC period that might have less evident effects [6,30]. Another objective of this experimental study was to assess RIPC as a means to protect the target organ without causing the direct stress of ischemia-reperfusion. By using the two techniques in separate groups as well as combined, in a specific group that received the two types of IPC, our goal was to ascertain if there would be a difference between preconditioning the graft or the recipient, and to check potential cumulative effects when both procedures are used concomitantly.
For RIPC, the most widely used technique is clamping the artery to the limb of a patient [31] or animal [32]. However, for our project we chose a different target territory, i.e. the gut, by clamping the superior mesenteric artery. This organ was chosen because it has one of the highest levels of metabolic activity and is very susceptible to oxygen pressure variations, and therefore is quickly responsive to short periods of ischemia, which should maximize the protective effect of the RIPC. Finally, it is important to stress that the gut is a territory that markedly suffers from blood stasis during the anhepatic phase of the transplant procedure.
We performed the current experiments to clarify local and systemic effects of IPC in liver transplantation and assess the potential influence of our conditioning models on common problems caused by I/R injury in organ transplants, such as acute kidney injury [33][34][35]. Unlike previous publications, our biochemical results fail to show any benefit of DIPC or RIPC in liver transplantation. Serum AST and ALT levels are consistent with the degree of hepatocellular injury and are used as indicators of graft distress, bearing correlation with different levels of primary graft dysfunction. All groups in our study were comparable in terms of enzymatic profile. In the D + R group, AST showed high variability, with a trend toward higher median values at 12 h, 18 h, and 24 h, suggesting a possible harmful effect of the addition of the two IPC procedures, even though the difference was not significant.
Our results differ from those of studies of liver transplantation in small animals (rats), which showed lower transaminase values and improvement of histopathological aspects in animals submitted to RIPC compared to controls after 24 h of reperfusion [32]. On the other hand, our results are consistent with those from human studies in which direct and remote IPC promoted increased AST and ALT values 24 h after reperfusion [36].
We made a refinement in this current investigation, by adding some molecular analyses that could show some beneficial effects of IPC. In the liver tissue, the combined IPC approach resulted in marked positive changes in gene expression. The eNOS gene expression in the liver tissue was higher in the D + R group at 24 h, and such expression is usually related to improvement in ischemia. In addition, lower expression of pro-inflammatory genes (BAX) and higher expression of anti-inflammatory genes (Bcl-XL) were observed in this same group. DIPC also had positive effects, leading to increased expression of the IL-6 and Bcl-XL genes. Finally, these results are similar to those obtained in a human model of liver ischemia-reperfusion many years ago [37].
The BAX/Bcl-XL ratio showed lower values for all treated groups when compared with controls at 24 h. This finding may suggest an IPC-driven potential decrease in cell apoptosis secondary to I/R injury. Furthermore, in the gut, kidney, and lung tissues, some molecular changes were detected demonstrating beneficial effects of each IPC separately.
Although gene expression suggests positive effects, we cannot infer that in the complex situation of liver transplantation these results would be sufficient to indicate an IPC procedure. In addition, the results of biochemical and histopathological analyses, consistent with human studies, did not confirm any benefits from IPC, which raises questions about the feasibility of extrapolating results obtained in small animals to medium-sized animals and humans. The first hypothesis to explain this difference would be a higher susceptibility of certain species, e.g., rodents, to the IPC procedure, either in situ (direct) or remote, with this propensity diminishing as we advance phylogenetically toward humans, as in the case of pigs.
Also, worth mentioning are the physiological, anatomical, and surgical variations involved in organ transplantation in different species. Despite the highly ingenious technical solutions found to overcome difficulties and enable organ transplantation in rats and mice [38,39], the surgical procedure in these animals cannot match what happens in human surgery, including technical hurdles, hemodynamic instability, surgical time, postoperative follow-up, and mainly ischemic susceptibility at the intestinal level, as recently emphasized. The cellular and molecular interspecies differences should always be considered in the interpretation of findings when relating them to the clinical setting [40].
Finally, it remains an important question: although considered a good idea, why the ischemic preconditioning in the pig model did not work out? Probably the ischemia-reperfusion represents intense stress to the liver graft, that all benefits of ischemic preconditioning are covered by the real ischemic impact.
Considering the great complexity of transplant surgery even in the experimental field, we may acknowledge the limitations of the current study due to the small group size and healthy animals. Based on our results, we may conclude that the three methods of IPC herein utilized may not be utilized in clinical practice.

Author's contributions
ARB, ACAT, RRG, and JLF performed the animal experiments; AMS, SS, JOG, CSF, ASA, VRP performed all laboratory studies; DARM performed the statistical analyses and UT was the responsible for the manuscript. All authors have read and approved the manuscript.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The laboratory studies of the current investigation were supported by Fundação de Amparo a Pesquisa do Estado de São Paulo -(Research number 2014/25676-0).