CXC-chemokine regulation and neutrophil trafficking in hepatic ischemia-reperfusion injury in P-selectin/ICAM-1 deficient mice

Gene-targeted double mutant mice deficient in P-selectin and ICAM-1 (P/I double mutant), C57BL/6-Icam1^tm1BaySelp^tm1Bay, were used in this study. The breeding pairs of double-knockout mice were purchased from Jackson Laboratory (Bar Harbor, ME) and bred under the guidance of University Laboratory Animal Resources at Michigan State University. The wild-type mice were male C57BL/6. Before and after surgery, all the animals had unlimited access to food and water.

A murine model of lobar hepatic ischemia, as previously described by our laboratory, was used [13]. The experimental procedures were performed under aseptic conditions. Adult male mice (8–10 wk) weighing between 23–30 g were anesthetized with inhaled methoxyflurane (Baxter Caribe, Inc., Guayama, PR) followed by an intraperitoneal injection (35 mg/kg body wild-type) of sodium pentobarbital (Abbott Laboratories, North Chicago, IL). A midline laparotomy was performed. The ligamentous attachments of the left lateral and median lobes were carefully divided. The left lateral and median lobes were freed. The portal circulation to both of these lobes was carefully dissected and the portal vein and hepatic artery supplying the median and left lateral lobes were then interrupted with an atraumatic vascular clamp (Accurate Surgical and Scientific Instruments Corporation, Westbury, NY). The left lateral lobe was also rotated 180 degrees counter-clockwise on its vascular pedicle to eliminate any potential perfusion that might occur with an imperfect clamp occlusion. The caudate and right lateral lobes, as well as the papillary and quadrate processes, retained an intact portal and arterial inflow and venous outflow to prevent intestinal venous congestion. This procedure resulted in the induction of ischemia to approximately 65–70 percent of the liver. The mortality due to the surgical procedure was minimal (< 1–2%). After 90 minutes of partial hepatic ischemia the clamp was removed, the left lobe was rotated back 180 degrees clockwise, and reperfusion was initiated. The midline laparotomy was closed in a single layer fashion using 5-0 nylon suture. Sterile lactated Ringer's solution (0.8 ml) was administered subcutaneously to compensate for operative blood and fluid losses. Animals were divided into two groups; the test group underwent I/R and the sham group underwent the same anesthesia and midline laparotomy dissection of the portal vessels and liver, but without vascular occlusion. Mice were euthanized after 6 and 15 h of reperfusion and the blood and liver tissue were collected and processed, as described below. Additionally, a survival study was carried on in which the length of survival from the start of reperfusion was recorded up to three weeks at the time the mice were euthanized.

Blood samples were collected from the right ventricle via a left anterior thoracotomy in a sterile heparinized syringe containing 50 μl of heparin (100 USP Units/ml). The blood samples were centrifuged and plasma were collected and stored at -30°C until further use. Portions of the ischemic and non-ischemic liver lobes were fixed in buffered 10% formalin, embedded in paraffin, and used for hematoxylin and eosin (H&E) staining. Other portions of ischemic and non-ischemic liver lobes were snap frozen in liquid nitrogen and stored at -70°C, until use for immunohistochemistry staining, and MPO analysis.

The plasma ALT levels were determined spectraphotometrically, as previously described [13]. The ALT values are expressed in international units per liter (IU/L).

H&E staining was performed on tissue sections prepared at 5-μm intervals. A pathologist, blinded to the experimental procedure of the mice, examined the histopathology of the hepatic tissue sections.

ICAM-1 expression of the hepatic tissue was detected by an immunohistochemistry technique as previously published by our laboratory [13]. Briefly, cryosections (5-um thick) from ischemic and nonischemic hepatic lobes fixed in acetone were stained using an anti-mouse ICAM-1 antibody (3E2, IgG, Pharmingen, San Diego, CA) and a biotin-conjugated goat anti-hamster IgG secondary antibody (Pharmingen). ICAM-1 molecules were visualized using a Vectastain avidin-biotin complex reagent and 3,3'-diaminobenzidine chromogen kits (Vector Lab, Inc., Burlingame, CA). The tissue sections were examined using a Nikon light microscope interfaced with a Spot 24-Bit Digital Color Camera. Similarly, immunohistochemical staining for neutrophils was performed using a primary antibody (IgG_2a) specific to the mouse neutrophil (Cedarlane International Distributor, Ontario, Canada).

Plasma TNF-α, IL-6, KC, and MIP-2 levels were determined in a 96-well Nunc-Immuno microplate (VWR Scientific, Chicago, IL), using a sandwich e nzyme-l inked i mmunos orbent a ssay (ELISA) technique, as previously described [17]. The capture antibody was a polyclonal anti-mouse TNF-α, IL-6, KC, or MIP-2 specific goat IgG (R&D Systems, Minneapolis, MN) and the detection antibody was a biotinylated polyclonal anti-mouse TNF-α, IL-6, KC or MIP-2 specific goat IgG, (R&D Systems). All plasma samples were tested in duplicate. The minimal detectable protein concentration was 20 pg/ml.

Liver MPO content was measured according to the previously published method by our laboratory [17]. Briefly, the frozen liver tissues were homogenized using a Tissue Tearor, centrifuged and the pellets were resuspended in the buffer. The MPO activity was determined using a tetramethylbenzidine substrate kit (ImmunoPure, Pierce, Rockford, IL) and read at 450 nm using a human MPO as a standard. One unit of MPO activity was defined as the quantity of enzyme degrading 1 μmol peroxide/min at 25°C.

All data are expressed as means ± SEM. Comparison between two groups was performed using an unpaired Student t-test. Comparisons between multiple groups and various time points were performed using a Kruskal-Wallis One-Way Analysis of Variance (ANOVA) followed by a Bonferroni test. Survival data was assessed using the Kaplan-Meier log rank test. Analysis was performed using the Number Cruncher Statistical System (Number Cruncher Statistical Systems, Kaysville, UT). P ≤ 0.05 was considered significant.

The absence of ICAM-1, and P-selectin expression in the P/I null mice was confirmed in randomly selected litter mice by specific immunohistochemical staining and reverse transcriptase polymerase chain reaction (RT-PCR), as previously published by our laboratory [13]. The ICAM-1 expression was determined in all the animals used in this study. ICAM-1 was constitutively expressed in the wild-type control mice as indicated by brown staining along endothelial lining of the sinusoids, and hepatic vasculatures (Figure 1A). The ICAM-1 expression was markedly increased in wild-type mice following hepatic I/R (Figure 1C). In contrast, in P/I null liver tissue, ICAM-expression was absent (Figures 1B and 1D). The RT-PCR data are not shown.

Hepatic I/R caused significant hepatocellular damage as demonstrated by plasma ALT levels. The plasma ALT levels of both wild-type and P/I null mice after 90 minutes of ischemia followed by 6 and 15 h of reperfusion were significantly elevated when compared to their respective sham-operated mice (Figure 2). ALT levels were negligible in both groups by the survival time-point. Although there was no statistically significant difference in ALT levels between the wild-type and P/I null mice at either time point of reperfusion (i.e. 6 and 15 h), P/I null mice showed decreased ALT levels compared to the wild-type mice.

The histopathologic injury of the liver tissue was evaluated based on sinusoidal congestion, cytoplasmic vacuolization, hepatocellular necrosis, and neutrophil infiltration. The liver sections from the sham-operated mice displayed minimal/no necrosis, similar to that of the non-operated control mice (Figure 3A and 3B). Additionally, there was no apparent evidence of hepatic injury due to ischemia alone (i.e., at zero hour of reperfusion; image not shown). However, reperfusion of the ischemic liver induced an extensive hepatocellular necrosis, sinusoidal congestion, and neutrophil infiltration after 6 and 15 h of reperfusion in both wild-type and P/I null mice (Figures 3C, 3D, 3E, and 3F). There was sparing of the periportal areas with progressively increased injury approaching the central vein. In general, it appeared that the wild-type I/R mice exhibited larger areas of coagulative necrosis when compared to the P/I null mice. The injury was associated with a marked number of neutrophils infiltrated into the midzonal region of ischemic liver after 6 and 15 h of reperfusion in both wild-type and P/I null mice, which was confirmed with in situ immunohistochemical staining of the neutrophils (Figure 3, last row). There were a minimal number of neutrophils present in the livers of sham-operated mice and in the non-ischemic lobes of mice subjected to I/R, indicating that ischemia was a pre-requisite for reperfusion injury to occur. Further, neutrophil infiltration was quantitated by measuring the liver MPO content (Figure 4). The liver MPO levels of both wild-type and P/I null mice after 90 minutes of ischemia followed by 6 and 15 h of reperfusion were significantly elevated compared to the sham-operated mice at the corresponding time points. Further, there was no statistically significant difference in MPO levels between the wild-type and P/I null mice, at either time point.

In order to determine whether plasma cytokine/chemokine levels correlated with tissue injury, the plasma cytokine/chemokine levels were measured using an ELISA. As figure 5 displays, the plasma levels of all cytokines (i.e., TNF-α, IL-6, MIP-2, and KC) were significantly increased in response to I/R, in both the wild-type and P/I null groups, and reached their maximum at 6 h of reperfusion, which then declined to baseline by 15 h of reperfusion. The data related to the wild-type is consistent with our previous studies, where a similar pattern was observed in wild-type mice subjected to hepatic I/R [20].

As Figure 5A. shows, hepatic IR caused significant production of TNF-α, by 6 h of reperfusion in both wild-type and P/I mice, which declined by 15 h of reperfusion. However, the data indicates no significant difference in I/R-induced TNF-α production between the wild-type and P/I mice. This data is consistent with plasma ALT data, which showed a maximal increase in ALT levels at I/R 6 h, followed by a decrease in levels by I/R 15 h of reperfusion. A similar pattern was observed in plasma IL-6 levels (Figure 5B). In contrast, chemokine production showed a different pattern, in that the P/I null mice had significantly lower levels of plasma KC and MIP-2 at I/R 6 h than the wild-type mice (Figure 5C. and 5D).

Ten wild-type and ten P/I null mice were subjected to 90 minutes of partial hepatic ischemia and were observed post-operatively and their approximate times of death were recorded. At the end of three weeks, all surviving mice were euthanized. All the P/I null mice survived the full three weeks while only 7 out of 10 of the wild-type mice survived that length of time. The Kaplan-Meier log rank test showed no statistically significant difference between the two groups (p = 0.067), although a trend towards improved survival in the P/I null group was apparent (Figure 6).