PECAM‐1 modulates liver damage induced by synergistic effects of TNF‐α and irradiation

Abstract The mechanisms of radiation‐induced liver damage are poorly understood. We investigated if tumour necrosis factor (TNF)‐α acts synergistically with irradiation, and how its activity is influenced by platelet endothelial cell adhesion molecule‐1 (PECAM‐1). We studied murine models of selective single‐dose (25 Gy) liver irradiation with and without TNF‐α application (2 μg/mouse; i.p.). In serum of wild‐type (wt)‐mice, irradiation induced a mild increase in hepatic damage marker aspartate aminotransferase (AST) in comparison to sham‐irradiated controls. AST levels further increased in mice treated with both irradiation and TNF‐α. Accordingly, elevated numbers of leucocytes and increased expression of the macrophage marker CD68 were observed in the liver of these mice. In parallel to hepatic damage, a consecutive decrease in expression of hepatic PECAM‐1 was found in mice that received radiation or TNF‐α treatment alone. The combination of radiation and TNF‐α induced an additional significant decline of PECAM‐1. Furthermore, increased expression of hepatic lipocalin‐2 (LCN‐2), a hepatoprotective protein, was detected at mRNA and protein levels after irradiation or TNF‐α treatment alone and the combination of both. Signal transducer and activator of transcription‐3 (STAT‐3) seems to be involved in the signalling cascade. To study the involvement of PECAM‐1 in hepatic damage more deeply, the liver of both wt‐ and PECAM‐1‐knock‐out‐mice were selectively irradiated (25 Gy). Thereby, ko‐mice showed higher liver damage as revealed by elevated AST levels, but also increased hepatoprotective LCN‐2 expression. Our studies show that TNF‐α has a pivotal role in radiation‐induced hepatic damage. It acts in concert with irradiation and its activity is modulated by PECAM‐1, which mediates pro‐ and anti‐inflammatory signalling.


| INTRODUC TI ON
Irradiation is one of the current treatment options for cancer. With the enormous progress in this field, it became technically possible to selectively irradiate most types of tumours. However, one of the shortcomings of this procedure is the unwanted effect on healthy tissues adjacent to the tumour. In the liver, a limiting factor for irradiation is radiation-induced liver disease (RILD) also known as radiation hepatitis. RILD is regarded as a major limiting factor for radiotherapy of primary and secondary liver tumours. 1,2 Therefore, the identification of molecular interactions involved in RILD may help to improve radiotherapeutic options, for example by the protection of non-diseased tissue from unwanted side effects of irradiation. Furthermore, the irradiation of liver tumours is often performed in patients with additional local or general diseases, which may increase radiation toxicity. 3 Of note, our in vitro studies have recently shown that normal liver cells are highly resistant against radiation, 4 which appears to be in contrast to the high incidence of RILD. However, combinatory effects of tumour necrosis factor (TNF)-α application and radiation increase radiosensitivity in healthy liver cells, associated with the release of mediators of cellular injury. [5][6][7] Tissue damage outside or within the liver leads to the induction of acute-phase (AP)-response, which is an early defense mechanism required to maintain homeostasis and initiate repair. Clinically, APresponse is characterized by fever, anorexia, as well as change or dramatic increase in concentration of serum proteins (positive APproteins). Thereby, release of positive AP-proteins such as C-reactive protein (CRP) in humans, or lipocalin-2 (LCN-2) and serum amyloid A (SAA) in mice, subsequent to pro-inflammatory cytokines, is a hallmark of AP-reaction. TNF-α is one of the major AP-cytokines. 8 It is well known for its role in cell injury probably through the generation of reactive oxygen species (ROS) 7,9,10 or apoptosis. 11 In liver, an active involvement of TNF-α in both acute and chronic liver inflammation has extensively been investigated. 7,10,12 Kupffer cells are the main source for TNF-α, but it is also produced by other liver cells during stress conditions, such as hepatocytes. 13,14 TNF-α levels have been reported to be elevated in viral hepatitis, alcoholic and non-alcoholic fatty liver disease and liver injury in humans and rodents. 7,15-17 Accordingly, we observed elevated TNF-α levels in various animal models of selective liver irradiation 4,18 and acute toxin-induced liver damage. 17 Several reports have demonstrated a positive correlation between tissue damage and the immigration of inflammatory cells. 7,[19][20][21] Thereby, TNF-α, which is of major importance for the maturation of humoural immune response, also regulates leucocyte transmigration via adhesion molecules and inflammatory mediators. 22,23 PECAM-1/CD31 is an adhesion molecule constitutively expressed by endothelial cells, macrophages, neutrophils and lymphocytes. 23,24 Initially, the role of PECAM-1 in inflammation was considered to be a minor one and was controversially debated. Moreover, the down-regulation of PECAM-1, both at the endothelial and the leucocyte surface, was considered to be a consequence of transmigration of leucocytes through the vessel wall and not the pre-condition for their transmigration. In fact, PECAM-1 down-regulation can be induced by cytokines in vitro without the need for contact between endothelial cells and leucocytes. 23,25 However, in recent years, data supporting its possible role as 'anti-inflammatory' adhesion molecule are increasing. 19,24 Lack of PECAM-1 is associated with greater tissue damage in a model of acute inflammation induced by lipopolysaccharides (LPS). 19,26 In line with this, we have recently shown that PECAM-1ko mice are more sensitive to irradiation, which is associated with increased production of inflammatory mediators, especially TNF-α. 4 Therefore, the current study investigated the probability of synergistic effects of TNF-α and irradiation in liver damage and the potential role of PECAM-1.

| Materials
All chemicals and reagents were purchased from commercial sources Sigma-Aldrich (St. Louis, USA) or Merck (Darmstadt, Germany).

| Animal models
All mice used in this study were male and 8-12 weeks old with a body weight of 20-28 g. Mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Wild-type (wt) (C57BL/6J) mice

| Measurement of aspartate aminotransferase (AST) in murine serum
At regular intervals between 1 and 48 hours following selective liver irradiation, blood samples from the V. cava inferior were collected from irradiated and sham-irradiated mice and used for AST measurement using analysis kits (DiaSys Deutschland, Flacht, Germany) according to the suppliers instructions.

| Immunofluorescence double-staining of mice liver sections
Immunofluorescence staining was performed as described before. 4

| RNA isolation and real-time PCR analysis
Total RNA from the livers of irradiated and sham-irradiated mice was isolated after homogenization in Trizol (Invitrogen, Carlsbad, USA) as described previously. 4 q-RT-PCR was performed with cDNA as described previously with primers (Invitrogen) listed in Table 1.

| Protein extraction and Western blot analysis
Protein extraction and Western blot analysis were performed as described before. 23 Briefly, 50 μg of protein was loaded on polyacrylamide gels (NuPAGE 4%-12% Bis-Tris Gel, Invitrogen, Carlsbad, CA, USA) under reducing conditions. After electrophoresis, proteins were transferred to Hybond-ECL (enhanced chemiluminescence) nitrocellulose membranes. Immunodetection was carried out as described previously 23 with antibodies and concentrations listed in Table 2.

| Statistical analysis
The data were analysed using GraphPad Prism version 4 software (San Diego, USA). All experimental errors are shown as SEM.

| Elevated aspartate aminotransferase levels in serum after irradiation of wt-mice
To examine liver damage, serum levels of aspartate aminotransferase (AST) were analysed after irradiation with and w/o injection of TNF-α in wt-mice. Serum levels of AST were significantly increased in both groups in comparison to sham-treated controls. AST levels started to rise after 3 hours, and remained elevated until 24 hours in mice that received both irradiation and TNF-α. The maximum enzyme activity was detected at 24 hours (305 ± 55 U/L). Irradiation-only induced maximum AST levels at 6 hours (188 ± 16 U/L) compared to sham-irradiated mice (96 ± 16 U/L). Of note, there was a significant difference between mice that received TNF-α immediately before irradiation compared to treatment with irradiation-only ( Figure 1).

| Changes in CD68 expression and recruitment of leucocytes into liver of wt-mice
RNA expression in liver of wt-mice was analysed by qRT-PCR and revealed an increase in CD68 expression (indicating higher numbers of macrophages and granulocytes) after combined administration of TNF-α and irradiation as compared to irradiation-only. mRNA expression of CD68 was significantly increased 3 hours after the combined administration of TNF-α and irradiation with a further significant increase at 6 hours (2.68 ± 0.8-folds) and 12 hours (3.22 ± 0.4-folds).
The level of CD68 decreased thereafter (Figure 2A).
Our qRT-PCR data were further confirmed by immunofluorescence studies. F4/80 + cells (marker for macrophages/Kupffer cells) and CK-19 + cells (marker of biliary cells) were regularly detected in liver of sham-irradiated mice ( Figure 2B). Consistent with our qRT-PCR-data, an increase in the number of F4/80 + cells was detected after 6 hours in mice treated with TNF-α plus irradiation, and

| Influence of irradiation on PECAM-1 expression in liver of wt-mice
Total RNA extracted from the livers of wt-mice treated by irradiation, TNF-α or both was analysed by real-time PCR. After irradiation, PECAM-1-specific transcripts started to decrease early and levels remained significantly low until 48 hours (Figure 3).

| Kinetics of LCN2 in liver of wt-mice after irradiation and TNF-α
In contrast to PECAM-1, the expression of LCN-2 was increased in the liver of wt-mice after single-dose selective liver irradiation with or w/o TNF-α injection at RNA and protein level as compared to sham-irradiated controls. The level of LCN-2 started to increase immediately (1 hour) after irradiation (4 ± 2-fold), further increased after 3 hours, remained high after 12 hours (163 ± 74-fold) and 24 hours, and decreased thereafter ( Figure 4A). More efficiently, TNF-α administration alone induced RNA expression of LCN-2, with a dramatic increase after 1 hour (263 ± 1-fold) and reached a peak value after 3 hours (1190 ± 180-fold). The levels of LCN-2 remained elevated until 24 hours and decreased thereafter. A dramatic induction in LCN-2 RNA expression was observed in mice who received irradiation in combination with TNF-α. The combined treatment raised LCN-2 expression after 1 hour (203 ± 80-fold), followed by a massive induction at 3 hours (2149 ± 553-fold). Then, the expression of LCN-2 started to decrease but remained drastically elevated even at 24 hours (1237 ± 29-fold). Thereby, from 3 to 24 hours, LCN-2 expression after combination treatment was significantly higher than each single treatment ( Figure 4A). Importantly, qRT-PCR results could be confirmed at protein level by Western blotting ( Figure 4B).

| Phosphorylation of STAT-3 in the liver of wtmice after irradiation and TNF-α
Signal transducer and activator of transcription-3 (STAT-3) is an important transcription factor in inflammatory signalling pathways.
By Western blot analysis with specific antibodies against pSTAT-3, phosphorylation of STAT-3 was noticed at 3 and 6 hours in mice after irradiation ( Figure 4B). A similar result was observed in mice who received TNF-α only. There, phosphorylation already occurred after 1 hour. Of note, the most intense signals of pSTAT-3 were visible in mice who received radiation plus TNF-α. The strength of this band increased further at 6 hours, and rapidly decreased thereafter.
Of note, the magnitude of phosphorylation was the highest in mice, who received both irradiation and TNF-α ( Figure 4B).

| Detection of the liver damage in wt and PECAM-1-ko-mice after irradiation
Liver damage after irradiation was confirmed by measuring the serum levels of AST both in wt and PECAM-1-ko-mice. Both groups of mice showed an increase in AST as compared to sham-irradiated controls ( Figure 5). The levels of AST rose immediately after irradiation with a maximum at 6 hours (238 ± 22 U/L). However, there was a significant difference between wt (238 ± 22 U/L) and ko-mice (422 ± 33 U/L) at this time-point, the ko-mice showing significantly higher levels of the enzyme. AST levels then decreased ( Figure 5).

| Changes in level of hepatic LCN-2 after irradiation
To validate the role of PECAM-1 in radiation-induced liver stress, we

| TNF-α and radiation-induced liver disease
The use of radiation therapy may be extended to hepatic cancers reported that the intracellular defense system of hepatocytes is weakened when irradiation is administered in combination with inflammatory mediators such as TNF-α. 5 TNF-α, when released into the extracellular matrix together with other mediators of cellular noxae, initiates inflammatory processes followed by the recruitment of leucocytes. It is now well-accepted that immigration of inflammatory cells plays an important role in tissue injury, as has been shown in animal models of toxicity. 20,21 We previously showed that selective liver irradiation results in cytokine and chemokine expression, mainly TNF-α and MCP-1. 4,28 Their increased synthesis is accompanied by the immigration of distinct numbers of granulocytes, but not mononuclear phagocytes, into the liver and induces an immediately reversible mild hepatic damage. 28 In the current study, we demonstrate synergistic effects of irradiation and TNF-α application, which results in considerable numbers of mononuclear phagocytes in liver, in addition to granulocytes, which is in line with previous studies where an accumulation of inflammatory cells is shown to be associated with tissue damage. 28 This suggests an increased level of liver damage as compared to sole irradiation by increased numbers of newly recruited leucocytes.
Similarly, in a rat model of thioacetamide (TAA)-induced acute liver damage an amplified release of TNF-α was followed by immigration of both granulocytes and mononuclear phagocytes, which was associated with significant liver damage. 30

| Adhesion molecules and liver irradiation
In addition to cytokines and chemokines, adhesion molecules play a crucial role in the activation of infiltrating cells into liver, their attachment to endothelial cells, primarily those of the portal vessels, and their transmigration towards stressed hepatocytes. [31][32][33][34] Thereby, for the transmigration and activation of inflammatory cells. 22,23,31,34 In the current study, we examined the significance of TNF-α-induced signalling for (down)-regulation of PECAM-1. We and others could evidently show the functional relevance of TNF-α for the regulation of PECAM-1 in inflammatory processes, 24,34 which was considered to be of minor importance previously. Here we show that lack or reduction of PECAM-1 is associated with elevated and prolonged liver damage as shown by AST serum levels. This is in accordance with other reports, where lack or reduction of PECAM-1 expression corresponded with greater liver damage after administration of CCl 4 to animals. 34 In the same lines, an enhanced liver damage was reported in PECAM-1-ko-mice after LPS exposure, indicating clearly that lack of PECAM-1 can enhance acute liver damage. 19,26 It was shown that both IFN-γ and TNF-α can down-regulate PECAM-1 in various cell types, such as liver endothelial cells, sinusoidal macrophages, peripheral blood leucocytes (PBLs) and granulocytes. 22,23 Furthermore, the administration of anti-TNF-α antibodies (Infliximab) reverses the reduction caused by either IFN-γ or TNF-α in PBLs, macrophages and granulocytes. 23 Anti-TNF-α therapy exerts its anti-inflammatory effects by neutralizing soluble TNF-α and consequently blocking IFN-γ signalling. 27

| Hepatoprotective mechanisms and liver irradiation
Lipocalin-2 (LCN-2) is not only a marker for irradiation-induced damage but also a positive acute-phase protein. 36 It is up-regulated In fact, regulation of LCN-2 through STAT-3, and the liver protective functions of this pathway, has been well documented. [38][39][40] Taken together, our studies give insight into irradiation-induced liver injury through TNF-α-regulated PECAM-1, which may be important not only for leucocyte transmigration through the vascular wall, but also for the regulation of inflammation per se. Furthermore, we could show that after application of the pro-inflammatory agent TNF-α the liver becomes more sensitive to radiation. This may explain the high sensitivity of the liver to radiotherapy in patients, when the diseased/inflamed liver is irradiated and additionally treated with chemotherapeutics. Our results also indicate that antiinflammatory treatment may help to prevent RILD and its sequel.
Thereby, liver protection might be achieved by PECAM-1 up-regulation. Our results point to the possibility of developing therapeutics to reduce radiation-induced damage in normal tissue, as well as agents that may enhance the effects of radiation in tumours.

CO N FLI C T O F I NTE R E S T
The authors declare that no actual or potential conflict-of-interest in relation to this article exists.