Exercise training decreases hepatic injury via changes in immune response to liver ischemia/reperfusion in mice

Abstract

Background & Aims

Liver ischemia-reperfusion injury (IRI) induces local and systemic inflammation in which neutrophil extracellular traps (NETs) are major drivers. IRI markedly augments metastatic growth consistent with the notion that the liver IRI can serve as a premetastatic niche. Exercise training (ExT) confers a sustainable protection reducing IRI in some animal models and has been associated with improved survival in cancer patients; however, the impact of ExT on liver IRI or development of hepatic metastases is unknown.

Approach & Results

Mice were randomized into exercise (ExT) and sedentary groups prior to liver IRI and tumor injection. Computerized dynamic network analysis (DyNA) of 20 inflammatory mediators was utilized to dissect the sequence of post-I/R mediator interactions that induce injury. ExT mice showed a significant decrease in hepatic IRI and tissue necrosis. This coincided with disassembly of complex networks among inflammatory mediators seen in sedentary mice. Neutrophil infiltration and NETs formation were decreased in the ExT group, which suppressed the expression of liver endothelial cell adhesion molecules. Concurrently, ExT mice revealed a distinct population of infiltrating macrophages expressing M2 phenotypic genes. In a metastatic model, fewer metastases were present three weeks after I/R in the ExT mice, a finding that correlated with a marked increase in tumor-suppressing T cells within the tumor microenvironment.

Conclusions

ExT preconditioning mitigates the inflammatory response to liver IRI, protecting the liver from injury and metastases. In light of these findings, potential may exist for the reduction of liver pre-metastatic niches induced by liver IRI through the use of exercise training as a non-pharmacologic therapy prior to curative surgical approaches.

INTRODUCTION

Liver ischemia-reperfusion injury (IRI) is a common clinical consequence of all forms of liver surgery and transplantation (1). IRI results from direct cellular damage caused by blood flow interruption that is accentuated upon organ reperfusion (2). IRI during major hepatic surgery decreases oxygen supply and leads to impaired adenosine triphosphate (ATP) production, an excessive burst of reactive oxygen species (ROS), and the systemic activation of inflammatory pathways (3), leading to parenchymal damage, organ dysfunction, morbidity, and increased healthcare costs.

Besides jeopardizing patients’ outcomes in the early postoperative period, there are growing concerns that liver IRI can promote progression of metastatic disease and worsen long-term oncological outcomes for patients undergoing potential curative surgery for liver metastases (4,5). Surgical manipulation of the liver can result in the release of cancer cells into the circulation, which could in turn form new metastatic foci in the remnant liver. In addition, the pro-tumorigenic changes accompanying liver IRI can promote the growth of clinically undetectable micrometastatic disease (6–8). Despite its obvious clinical importance, the mechanisms that account for liver IRI are only partially understood and remain an understudied area. Although an array of agents has shown protective effects against liver IRI in animal models, only a few have been tested in randomized, placebo-controlled clinical trials, with varying degrees of success (9,10). Thus, new therapeutic strategies to minimize IRI are warranted to improve short- and long-term outcomes.

Exercise training confers a sustainable protection against ischemia-reperfusion injury in several animal models of ischemia, such as cardiac or cerebral ischemia, and has been associated with improved survival in patients with cancer (11,12). During IRI, hepatocytes and liver endothelial cells release mediators and damage-associated molecular pattern molecules (DAMPs) that trigger the immune system as part of a robust inflammatory cascade (13). Neutrophils, macrophages, dendritic cells, natural killer (NK) cells, and a variety of lymphocytes all play crucial roles in a complex of inflammatory responses (14–16) in addition to playing a potential role in future metastatic progression. Part of the protective effect of exercise (17) is postulated to be the result of profound and pleiotropic changes in the immune system, resulting in alterations in wound healing, the release of various inflammatory mediators, the polarization of macrophages, and the proliferation of various lymphocyte sub-populations (18).

Given the obvious benefits of exercise in inflammatory states and cancer, we sought to determine whether similar protective effects can be achieved in our metastatic liver ischemia-reperfusion model through modulating the inflammatory response. Indeed, studies described herein indicate that moderate-intensity aerobic exercise training attenuates liver inflammatory responses and decreases injury and the development of metastases in a mouse model of liver IRI. In light of these findings, there may be a rationale for the reduction of liver IRI-associated promotion of metastatic disease through the use of aerobic exercise training as a non-pharmacological therapy prior to curative surgical approaches.

METHODS

Animals I/R and metastases model

Male C57BL/6 mice of 8–12 weeks of age where subjected to 70% non-lethal hepatic ischemia and reperfusion as described previously (8,13,19,20). Colorectal cancer liver metastases were induced in mice by injecting (1×10⁶/100ul PBS) MC38 (murine colorectal adenocarcinoma) cells via portal vein right before inducing the liver IRI. Experimental groups were sacrificed 3 weeks after the I/R and tissue was harvested. All experiments were performed according to protocols approved by Institutional Animal Care and Use Committee of the University of Pittsburgh.

Exercise Training Model

ExT animals were subjected to 1 h treadmill running (Columbus Instruments, USA) at a horizontal position with opaque dividers separating the six running lanes preventing visualization. Mice were trained at the speed of 12.5 meter/minute for 5 consecutive days a week followed by a 48-hour recovery time. All experiments were performed 48 hours following the end of the 4-week training period.

Neutrophil isolation, NETs formation and quantification of NET Chromatin [NETchr]

Mouse neutrophils were isolated by flushing the bone marrow out of tibias and femurs as described previously (21). Cells were collected 4 hours after the phorbol-12-myristate-13-acetale (PMA) treatment (250 nM) and spun at 480 × g. The Pellet was discarded, and the supernatant containing NET Chromatin (NETchr) was spun at 18,000 × g. Quantification of NETchr was analyzed using Nano Drop. In vivo, quantification of NETchr was analyzed using myeloperoxidase (MPO) associated with DNA ELISA as described previously (19).

Network Analysis

Network Analysis was carried out to define the central inflammatory network nodes in both experimental groups (Sed+I/R vs. ExT+I/R) at 6 hrs using a modified version (22) of our previously published algorithm implemented in MATLAB® (MathWorks, Natick, MA) for Dynamic Network Analysis (DyNA) (19,23,24). Connections between inflammatory mediators were created if the Pearson correlation coefficient between any two nodes (inflammatory mediators) was greater or equal to a threshold of 0.7–0.95, as indicated. The “network complexity” for each experimental condition was calculated using the following formula: Sum (N1+ N2+…+ Nn)/n-1, where N represents the total number of connections for each mediator and n is the number of mediators analyzed.

Statistical Analysis

Results are expressed as the mean ± standard error of mean (SEM). Multiple comparisons were performed using Student’s t-test and ANOVA. The “P” value <0.05 was considered statistically significant. SigmaPlot™ 11 software was used to analyze cytokine data (Systat Software, Inc., San Jose, CA).

RESULTS

ExT decreases liver IRI when compared with sedentary mice

The impact of ExT on liver IRI is unknown. ExT Mice were subjected to four weeks (5x/week) of 1 hour of treadmill running at a pace of 12 meters per minute followed by 2 days of rest prior to liver ischemia-reperfusion. The IRI outcomes were compared with those seen in sedentary mice (Figure 1A). As expected, ExT inhibited weight gain (Figure 1B). ExT significantly reduced IRI as evidenced by reduced serum ALT, AST, and LDH levels 6-hours after reperfusion (Figure 1C). The histological changes after I/R were consistent with the serum markers of liver damage, with markedly decreased sinusoidal congestion, vacuolization, and hepatocellular necrosis in the exercise-trained group (Figure 1D and 1E).

Fig. 1. Pre-operative exercise training decreases liver ischemia reperfusion injury.

(A) Schematic illustration of murine pre-operative ExT model indicating duration and period of liver IRI. (B) Line graph analysis showing difference in the mean total body weight between sedentary and ExT group over 4 weeks of time period (n=10), *P<0.05. (C) Serum levels of ALT, AST and LDH were decreased in ExT group compared to sedentary at 6 h (n=10) *P<0.05 **P<0.01. (D-E) Liver H&E staining and graph bar indicating extent of necrotic area between the groups. Magnification 10x (n=5) **P<0.01.

Differential dynamic inflammatory networks identified 6-hours after ischemia-reperfusion injury

We next sought to determine potential mechanisms by which ExT ameliorates liver IRI. Following liver I/R, a complex chain reaction develops between inflammatory cells and their cytokine products. We have previously found that dynamic networks in the livers of mice subjected to IRI were characterized by interconnections among both innate immune and lymphoid mediators (19). By 6 hours, the networks evolve into a much larger, reverberating mediator complex, a phenomenon often labelled as a “cytokine storm”. This cytokine storm is responsible for maintaining and propagating an inflammatory response that exacerbates the injury after liver I/R. We hypothesized that ExT prior to the I/R stress will precondition the mouse to a lesser inflammatory state and ameliorate this cytokine storm. Dynamic Network Analysis (DyNA) was performed on the inflammatory mediators assessed in the I/R liver lobes to compare ExT mice to sedentary mice. After ExT, there was a marked decrease in the number of significant mediators and their interconnections, with a decreased network complexity as compared to sedentary controls (Figure 2A, 2B, 2C, Supplementary Table 1). DyNA showed that ExT preconditioning of the mice disassembled the cytokine networks by radically reducing the number of participating cytokines and their mutual interconnections after IRI.

Fig. 2. Differential dynamic inflammatory networks identified 6 hours after ischemia reperfusion injury.

(A) Network Analysis of liver I/R tissue (Sed+I/R vs. ExT+I/R) showing an overview of the networks and mediator connections. The red circles represent mediators with one or more connection, whereas yellow circles represent mediators with no connections to other mediators. The black connecting line indicate positive connections; the red connecting lines indicate negative connections. (B) Simplified representation of inflammatory network connections between innate immune mediators and lymphoid mediators in ExT as compared to Sed group 6 h after I/R. (C) Network complexity in Sed+I/R vs. ExT+I/R at different stringency levels as described in Materials and Methods. (n=5)

ExT mitigates inflammation by inhibiting neutrophil recruitment and NET formation

Neutrophils are known to play a crucial role in the response to sterile inflammation induced by liver IRI (25,26). The formation of NETs plays a major mechanistic role in the damaging effect of IRI (6,8,13,19). Figure 2B and 2C show that ExT significantly reduced the involvement of mediators related to neutrophil recruitment and activation in the ischemic lobe after liver IRI. We thus hypothesized that ExT may play a protective role by altering neutrophil trafficking and NET formation. Using flow cytometry, there was a significant decrease in neutrophil recruitment to the ischemic lobe 6 hours after reperfusion in the ExT cohort compared to the sedentary one (Figure 3A). This was paralleled by a significant decrease in the mRNA levels of neutrophil attractant chemokines, CXCL1/KC, CXCL2/MIP2, and CXCL5/ LIX, in the ischemic lobe of livers in ExT mice (Figure 3B). We found that ischemic lobes in sedentary mice exhibited significantly higher levels of citrullinated-histone H3 (citH3), a specific marker of NET formation. Cith3 was significantly reduced when mice were exercise-trained prior to I/R. (Figure 3C). Furthermore, the serum levels of MPO-DNA complexes, a marker of circulating nucleosomes that are derived from NET formation were significantly decreased after I/R in the ExT group compared to sedentary controls (Figure 3D). In addition, there was a small but significant increase in DNase 1 activity, an endogenous enzyme responsible for degrading NETs, after liver IRI in the ExT group (Figure 3E). In vitro, we isolated neutrophils from either the bone marrow or blood of both sedentary and ExT mice and treated them with PMA (250nM), a NET stimulator (27), for 4 hours. Neutrophils from the ExT group showed decreased ability of neutrophils from either source to form NETs as observed by the expression of citH3 via immunofluorescence imaging and western blot analysis (Figure 3F and 3G and Supplementary Figure 2A).

Fig 3. Exercise training mitigates inflammation by inhibiting neutrophil recruitment and NET formation into the I/R liver lobe.

(A) Representative flowcytometry analysis showing decrease in percent neutrophil (CD11b⁺ Ly6G⁺) infiltration into the ischemic lobe of ExT mice compared to sedentary, 6 h after reperfusion. (B) Real Time PCR analysis showing significant decrease in neutrophil attractant chemokine ligands in the reperfused tissue of ExT and sedentary [Sed] mice *P<0.05. (C-D) Quantification of serum MPO-DNA levels and lysate citH3 protein expression show decreased NET formation in ExT mice compared to sedentary mice. ***P<0.001 ****P<0.0001. (E) Serum DNase levels are significantly increased in the sera of ExT group 6 hrs after I/R. (n=10) *P<0.05. (F-G) Bone marrow derived neutrophils were cultured and stimulated with PMA (250nM) for 4 hours from both sedentary and ExT mice. Decreased expression of citH3 protein was observed in the ExT group as evident by Immunofluorescence imaging and western blot analysis. (n=5) Magnification 40x, scale bar 50 μm.

ExT polarizes macrophages into anti-inflammatory state after liver I/R

Thus far, we have shown that decrease IRI in ExT mice is associated with decrease inflammatory network complexity and decrease neutrophil infiltration and NET formation within the I/R liver lobe. ExT might also modulate the activity of other innate immune cells in specific tissues. It has been shown that in healthy mice, when exposed to chronic exercise, a shift in macrophage phenotype takes place from a pro-inflammatory M1 state to an anti-inflammatory M2 within the adipose tissue (28). We therefore quantified macrophage infiltration within the I/R liver lobes. Flow cytometry analysis shows that liver I/R promoted increased macrophage infiltration in the ExT group 6 hours after reperfusion compared to the sedentary group (Figure 4A). Furthermore, we found that among these infiltrating macrophages, the genes (Arg1 and CD206) associated with the anti-inflammatory subpopulation of the M2 phenotype macrophages were upregulated in the ExT group (Figure 4B). By contrast iNOS and CD86, markers of the M1 phenotype, were both markedly elevated in the infiltrating macrophages of the I/R lobes of sedentary mice but remained normal or even reduced in the I/R lobes of ExT mice. We next evaluated in vitro the cytokines produced by the subpopulations of macrophages harvested from the I/R lobes of ExT or sedentary mice. There was significant increase in gene expression of anti-inflammatory cytokines (IL-10/TGF-β) in the ExT groups; whereas, pro-inflammatory cytokines (IL-1β/TNFα) gene expression was significantly higher in the sedentary groups (Figure 4C). The change in macrophage polarization after ExT parallels the decreased NET formation within the tissues (Figure 3).

Fig 4. Exercise training polarizes macrophages into anti-inflammatory state.

(A) Representative flowcytometry analysis showing increase in percent macrophages (CD11b⁺F480⁺) infiltrating into the I/R lobe in ExT mice compared to sedentary, 6 h after reperfusion. (B) Quantitative PCR (qPCR) analysis showing gene expression associated with proinflammatory M1 (iNOS and CD86) and anti-inflammatory M2 (Arg1 and CD206) macrophages in the reperfused lobe of ExT vs Sed mice (n=10) *P<0.05 **P<0.01 ***P<0.001. (C) ExT I/R treated mice showed significantly increased expression of anti-inflammatory cytokines (IL-10 and TGF-β) in liver derived macrophages *P<0.05 **P<0.01 ***P<0.005 ****P<0.001 (D) Macrophage lineage RAW264.7 cells were cultured and treated with unstimulated neutrophils or NETchr for 12 hours and gene analysis was performed using qPCR. *P<0.05 ***P<0.001.

In vitro, RAW 264.7 cells of macrophage lineage were treated with NETchr. PCR analysis showed that treatment of macrophages with NETchr significantly increased the expression of a gene (CD86) associated with the M1 type population (Figure 4D). These results are consistent with the notion that the products of NETosis, which are increased in sedentary mice and which may shift the phenotype of macrophages toward M1 expression and thus may contribute to the dominance of M1 macrophages in the I/R lobes of sedentary mice. By contrast, the reduced level of NETosis in the I/R lobes of ExT mice may account in part for the predominance of M2 macrophages therein. ExT may thus promote its protective effect by decreasing NETosis in the I/R liver and shifting the macrophage population into the M2 state. One mechanism by which ExT promotes its protective effect against IRI is through decreased NET formation and thus promoting M2 polarization.

ExT and NET alter expression of LSEC adhesion molecules and immune trafficking

Liver sinusoidal endothelial cells normally play important protective roles controlling inflammation by acting as a gate to incoming immune cells. Endothelial adhesion molecules regulate the transmigration of activated neutrophils and monocytes towards the interstitial space of the I/R tissue (29). In our model, following liver IRI, qPCR analysis of isolated LSECs from ischemic lobes showed a significant increase in the expression of Intracellular Adhesion Molecule (ICAM)-1, (ICAM)-2 and Vascular cell Adhesion Molecule (VCAM)-1 in the sedentary mice group compared to the ExT mice (Figure 5A), consistent with the observed increase in invasive neutrophils in the ischemic lobe of sedentary mice (Fig 3A). Figure 5B, 5C and 5D further show a decrease in the protein expression of these activated endothelial markers in the ischemic lobes of the liver of the ExT mice by western blot analysis and immunofluorescent imaging. As neutrophils are among the first cells to interact with LSECs and infiltrate the ischemic lobes, we next sought to delineate the role of NETs in the activation of endothelial cells. LSECs were harvested from the liver tissue of healthy mice and were cultured with NETchr for 12 hours. Figure 5E shows that LSECs, when treated with NETchr in vitro, upregulate the expression of ICAM-1, ICAM-2, and VCAM-1. Thus, NETs can induce expression of integrins on LSEC to increase infiltration of further neutrophils and pro-inflammatory immune cells contributing to the IRI. Thus, by inhibiting NET formation, ExT may act to mitigate further damage within the ischemic lobe by reducing the transmigration of additional infiltrating neutrophils.

Fig 5. Comparative expression of integrins on liver sinusoidal endothelial cells [LSEC] in ExT and sedentary mice after I/R.

(A) Graph plot analysis showing gene expression of endothelial cell activation integrins among normal (control) liver lobes and sedentary and ExT mice after I/R (n=5)*P<0.05 **P<0.01. (B-D) ExT treated I/R group showed decreased expression of endothelial integrins as evident by western blot analysis and Immunofluorescence staining. Magnification 40x, scale bar 50 μm (E) LSECs from normal sedentary mice were isolated and cultured with unstimulated neutrophils or NETchr for 12 hours. Quantitative PCR (qPCR) analysis showed significant increase expression of ICAM-1, ICAM-2 and VCAM-1 in NETchr treated LSEC compared to neutrophil group ***P<0.001.

ExT halts the IRI-induced accelerated growth of liver metastases

Surgical manipulation of the liver induces the release of cancer cells in the blood stream, which could in turn engraft into the remnant liver and form new metastatic foci as a source of tumor recurrence (30). We have shown previously that NETs lead to a chain of inflammatory responses that foster both the number of liver metastases and the rate of tumor growth when I/R is immediately preceded by intraportal injection of MC38 tumor cells. We used our model (Fig 6A) of liver IR and metastases to study the effect of ExT on metastatic growth. Mice in the ExT group displayed significantly decreased tumor growth compared with sedentary mice, which was grossly appreciable as smaller and less numerous tumors three weeks after tumor inoculation followed by IRI (Figure 6B, 6C and 6D). In addition, ExT significantly decreased the tumor load after I/R, as evidenced by decreased liver-to-body weight ratio and tumor hepatic replacement area (Figure 6E and 6F).

Fig 6. Exercise training halts IRI-induced accelerated growth of liver metastases.

(A) Schematic representation of murine pre-operative ExT training model showing training duration prior to I/RI induced liver metastasis. (B-C) Representative images of liver metastases showing reduced number of tumor nodules in the ExT group compare to sedentary 3 weeks after I/R treatment. Mice without I/R (sedentary) treatment served as a control group. (D-F) Graph bar analysis showing decreased gross surface metastases, liver to body weight ratio and percent hepatic replacement in the ExT group. (n=10) *P<0.05 **P<0.01 ***P<0.001.

ExT activates the infiltration of lymphocytes into the ischemic lobe of the liver after I/R

In addition to the pro-tumorigenic acute inflammatory response accompanying liver IRI, we further explored the effect of ExT on the immune cell infiltration into the tumor. Tumor tissues from mice subjected to liver lobar I/R plus intraportal MC38 tumor cells were harvested and digested into a single-cell suspension for analysis 3 weeks after liver IRI. Control mice received only intraportal tumor cells, but I/R was not induced. Flow cytometry analysis showed a significant decrease in the percent of tumor infiltrating neutrophils (TIN) (Figure 7A) within the tumor microenvironment of the ExT group post IRI compared to the sedentary group. We observed a reciprocal significant increase in tumor infiltrating macrophages (TIM) and NK cells in the ExT treated tumor microenvironment (Figure 7B and 7C). In addition, using both immunohistochemical staining and flow cytometry, we found that the proportion of CD4⁺ and CD8⁺ T-cells and CD3⁻NK1.1⁺ NKT cells infiltrating the tumor microenvironment after IRI was significantly greater in the ExT mice than the sedentary mice (Figure 7D and 7E).

Fig 7. Exercise training alters relative neutrophil and mononuclear cell infiltration of metastases in IRI liver lobes.

(A-C) ExT in the I/R treated mice significantly reduced neutrophil (CD11b⁺Ly6G⁺) recruitment and increased macrophage (CD11b⁺F480⁺) and NK cells (CD3⁻F480⁺) infiltration in the metastatic tumor tissue compared to sedentary 3 weeks after cancer injection. Mice without I/R (sedentary) treatment served as a control group. (n=5) *P<0.05 **P<0.01 (D) Representative immunofluorescence images showing decreased staining of CD4+ and CD8+ T cells in tumors growing in ExT group 3 weeks after I/R and tumor cell infusion. Magnification 40x, scale bar 50 μm. (E) Representative bar graph showing increased infiltration of lymphoid cells (CD3⁺CD4⁺ T cells, CD3⁺CD8⁺ T cells and CD3⁺NK1.1⁺ T cells) in the tumor microenvironment of the ExT group compared to sedentary. (n=5) *P<0.05 **P<0.01 ****P<0.001.

DISCUSSION

Acute liver injury and metastatic progression are known detrimental effects of liver ischemia-reperfusion, a common complex dynamic process occurring during liver surgery. This holds significant clinical implications on patients’ short- and long-term outcomes, and thus a strategy to ameliorate this unavoidable injury is critical. In this study, using a murine model of liver IRI and metastases, we show that ExT prior to ischemia-reperfusion is associated with a significant decrease in both liver injury after reperfusion and IRI-associated metastatic progression. The protective effects of ExT were predominately mediated through decreased NET formation, which resulted in a modification of the inflammatory environment and downstream effects on liver sinusoidal endothelial cells, macrophages and lymphocytes (Figure 8). Taken together, our findings are the first to provide evidence that ExT can directly impact liver pathophysiology and tumor biology by refining the immune environment and thus supports this as a promising therapeutic strategy.

Fig 8. Schematic diagram illustrating the proposed effect of exercise training in decreasing hepatic tissue injury and metastases after liver IRI.

Exercise training alters the acute inflammatory response, decreases the recruitment of neutrophil and concurrent NET formation, increases macrophage recruitment and its phenotypic switch to M2 subtype, decreases activation of endothelial cell adhesion molecules and subsequently protects the liver from the IRI damage. Exercise further promotes favorable oncological outcome by inducing marked increase in the tumor suppressing T cells with in the TME resulting in decreased formation of new metastatic niche and establishment of micrometastases.

All forms of liver surgery are inevitably accompanied by periods of liver ischemia and subsequent IRI: not only is the liver manipulated and compressed, but the portal triad is occluded routinely for significant periods during major resections. In fact, evidence of residual ischemic tissue has been found to adversely affect prognosis (31). The critical pathways involved are difficult to dissect. One central node noted by a number of investigators is the dominance of neutrophils and NETs with their pro-inflammatory and pro-tumorigenic properties (32). Tumors with dominant neutrophil infiltration have generally poorer prognosis (33,34). Neutrophils bear receptors for a number of neutrophil chemokines including CXCL1 and CXCL2 shown here to be produced in IRI liver lobes and are also expressed on some tumor cells (32). These chemokines, in addition to a number of inflammatory mediators, can trigger NETosis. We have shown that NETs can directly foster tumor cell mitochondrial biogenesis and tumor proliferation in vitro even in the absence of other cells, an effect mediated by interaction of neutrophil elastase with tumor surface TLR4 (21). Cools-Lartigue et al were the first to demonstrate that NETs in the liver could capture and promote metastasizing tumor cells (35), and Demers et al showed that NETs could promote thrombosis (36). These findings are consistent with the finding of large areas of liver necrosis seen in sedentary mice after IRI, but significantly mitigated in ExT mice (37). NETs also can awaken dormant tumor cells in the lung (38) and even act to establish premetastatic niches in the omentum for ovarian cancer (39). We have previously presented data to support a central role of NETs in the growth of metastases after liver surgery (6,8,20). In that sense, liver IRI creates a similar pre- metastatic niche; massive tumor growth occurs in the liver lobes subjected to IRI, and not in the continuously perfused lobe. ExT seems to mitigate against the creation of such a niche.

The current study shows that if tumor cells are injected into the portal system during ischemia-reperfusion in sedentary mice, the number of metastases and their size 3 weeks later was found proportional to liver damage at 6 hours, findings supporting the notion that metastatic growth is augmented by the inflammatory environment initiated by NETs. NETosis is accompanied by early infiltration of M1-polarized macrophages and increased expression of activated sinusoidal endothelial cell adhesion molecules involved in the extravasation of leukocytes (40). By contrast, ExT mice showed much less liver necrosis, significantly diminished hepatic chemokine levels, fewer infiltrating neutrophils, less evidence of NET formation, a disassembled inflammatory cytokine network, reduced expression of endothelial adhesion molecules, and a dominance of M2 anti-inflammatory macrophages. When circulating tumor cells were added to the model, ExT suppressed tumor metastases. Importantly, bone marrow and peripheral blood neutrophils of normal ExT mice produced fewer NETs after stimulation in vitro. The finding of increased NK cells in the tumors of ExT mice is consistent with the common reports of increased numbers of circulating NK cells in ExT mice or mice treated with IL-6 in the literature (41). Thus, ExT seems to modulate many of the characteristics of the normal pre-metastatic niche in the IRI liver downstream of NET infiltration. However, it impossible to attribute the entire protective effect on the reduced NETs infiltration because the changes in endothelial cells and infiltrating macrophages are seen concurrently. The changes induced by ExT may occur independently of each other and may even precede neutrophil infiltration. Also, ExT could conceivably alter the responsiveness of the liver tissue itself to I/R, leading to reduced chemokine production as an initial modulating effect and thereby reduce NET formation. In addition, previous studies have shown that skeletal muscle cells during exercise release a vast array of polypeptide hormones and signals, of which the most well studied is IL-6; a few of these myokines have tumor inhibitory effects in vitro (42). Furthermore, catecholamines released during exercise also have anti-cancer properties in vivo (43).

Basic research using animal models has elucidated several dominant molecular pathways in the pathogenesis of liver IRI. Although an array of agents has shown protective effects against IRI in animal models, only a few have been tested in clinical trials investigating therapeutic strategies to reduce IRI. Each trial was based on strong experimental evidence implicating a specific pathway to the pathophysiology of IRI (44). Although each pathway explored in the above trials is important in liver IRI, the translation to the bedside has been rather disappointing when compared to the success in experimental models. Liver IRI represents a continuum of complex processes that involve multiple cellular and molecular pathways. Thus, some of the varying successes in trials could be attributed to adopting a reductionist approach focusing on a single pathway. In this study, ExT resulted in a global anti-inflammatory effect on the cytokine response, endothelial cells, and the immune cells and thus a promising therapeutic strategy against liver IRI and metastatic growth.

There is much epidemiologic and observational evidence that regular physical exercise reduces the risk of cancer, slows tumor progression, and improves outcomes when combined with other traditional oncologic therapies (45). In a number of therapeutic trials of regular exercise, cancer survival prognosis was improved. Unmasking the mechanisms for the oncologic benefits of ExT might reveal useful therapeutic targets (46,47). In recent years, more emphasis has been placed on preoperative rehabilitation or prehabilitation (prehab) to optimize functional status prior to surgery. Evidence indicates that individuals who have better preoperative fitness experience lower rates of morbidity and mortality during their hospital stay and have better physical functional status postoperatively (48,49). Furthermore, physical exercise is also increasingly being integrated into the care of cancer patients, and for good reason. Evidence is accumulating that exercise improves the wellbeing of these patients combating physical and mental struggles during their fight against cancer (50). Targeting potentially modifiable behaviors such as activity in cancer patients could reduce risk of complications and recurrences and also empower cancer patients to take a more active role in their recovery. We are currently enrolling patients with colorectal liver metastases in a randomized trial (the DASH Study) to monitor activity levels with the intervention group receiving an exercise regimen and active prompts to comply with it. We will be looking at short-term postoperative and long-term oncologic outcomes.

In conclusion, this study demonstrates the novel finding that ExT prior to surgery significantly protected the liver from I/R injury which is associated with disassembly of the complex network between the inflammatory mediators. In addition, the in vivo and in vitro experiments document a global effect of ExT on immune cells activated after IR including decreased neutrophil infiltration, NET formation, macrophage phenotypic changes, and changes in endothelial cells’ response. The cumulative effect was protective against acute liver IRI and the associated pro-metastatic TME. The clinical implications of our findings are significant, as ExT may offer a powerful therapeutic approach for the treatment of liver IRI and could improve both short postoperative and long oncologic outcomes. While much remains to be learned about how physical ExT influences formation and progression of metastatic disease after IRI, evidence shows that exercise is safe and feasible for patients undergoing major or minor surgery.

Abbreviations

IRI

ischemia reperfusion injury

NETs

neutrophil extracellular traps

ExT

exercise training

DyNA

dynamic network analysis

NK

natural killer

MPO

myeloperoxidase

citH3

citrullinated histone H3

DNAse

deoxyribonuclease

NETchr

neutrophil extracellular trap chromatin

LSECs

liver sinusoidal endothelial cells