Depressed phagocytic activity of Kupffer cells after warm ischemia-reperfusion of the liver

Journal of Hepatology 1994; 20:301-304 Printed in Denmark. All rights reserved Munksgaard. Copenhagen

Copyright© Journalof Hepatology1994 Journal of Hepatology ISSN 0168-8278

Rapid Publication

Depressed phagocytic activity of Kupffer cells after warm ischemiareperfusion of the liver Brigitte Vollmar ~, Julia Glasz 1, Stefan Post 2 and Michael D. Menger ~ tlnstitute for Surgical Research, University o f Munich and 2Department o f Surgery, University o f Heidelberg, Germany

Phagocytic activity of Kupffer cells following hepatic ischemia/reperfusion was studied in 39 livers of male SpragueDawley rats by in vivo fluorescence microscopy. Animals were subjected to either 20 min (group B, n=9) and 60 min left hepatic lobar ischemia (group C, n=9) or to 20 min of global hepatic ischemia (group D, n= 11). Sham-operated animals without ischemia served as controls (group A, n= 10). After 60 min postischemic reperfusion, fluorescent latex beads (3.108. kg body wt-~; diameter: 1.1 ~rn) were injected intra-arterially. The zonal distribution and kinetics of adherence of latex beads were quantified by off-line video analysis. After 20 min of left hepatic lobar ischemia, 50%, 38% and 12% of injected latex beads adhered in zones 1, 2 and 3, respectively, and did not significantly differ from control livers (group A: 57%, 32% and 11%). In contrast, after 60 min of left hepatic lobar ischemia (group C) as well as after 20 min of global hepatic ischemia (group D), a more homogeneous distribution of latex beads adherent in zones 1, 2 and 3 was observed (group C: 48%, 36% and 16%; group D: 48%, 36% and 16%). Kinetic analysis of phagocytosis (% adherence of visible latex beads 1 rain and 3 rain after injection) showed no significant difference between 20 min left hepatic lobar ischemia (group B: 84% and 95%) and control (group A: 81% and 95%). However, sinusoidal adherence of latex beads was significantly delayed after both 60 min of left hepatic lobar ischemia (group C: 76% and 91%) and 20 min of global ischemia (group D: 74% and 92%) (p<0.01 vs. group A and B). We conclude that prolonged warm ischemia/reperfusion of the liver, and additional splanchnic vascular congestion, depresses the phagocytic activity of Kupffer cells. This probably contributes to the development of hepatic failure following severe ischemic insults to the liver. Key words." Kupffer cell; Phagocytic activity; Reperfusion; Warm hepatic ischemia

The vascular clearance of particulate matter, endotoxin and other antigenic material is a physiological function of Kupffer cells, the resident macrophages of the reticuloendothelial system of the liver, which account for about 90% of intravascular phagocytosis (1). Functional alterations of the reticulo-endothelial system have been demonstrated in many pathophysiological conditions including ischemia/reperfusion (I/R) of the liver (1). The frequent development of gram-negative bacteremia and sepsis following hepatic ischemic insult has been postulated to be, in part, due to depression of the reticulo-endothelial system with subsequent impairment of the hostdefense mechanism (2). Recently, our laboratory developed a method to moni-

tor phagocytic activity of Kupffer cells in vivo based on the adherence of fluorescent latex beads (3). To confirm the hypothesis that ischemia/reperfusion deranges hepatic reticulo-endothelial function, the postischemic phagocytic activity of Kupffer ceils was assessed using in vivo fluor, escence microscopy. Materials and Methods Animals and preparation

Thirty-nine male Sprague-Dawley rats weighing 175-260 g (mean body wt 200___7 g) were used. After overnight fasting, but free access to tap water, the animals were anesthetized with chloraihydrate (36 mg- 100 g body

Correspondence to: B. Vollmar, M.D., Institute for Surgical Research, Universityof Munich, Marchioninistr. 15, 81366 Munich, Germany

302 wt -~ i.p.) and atropine (0.25 mg s.c.). Tracheotomy was performed to facilitate spontaneous respiration (room air, flow rate 2 1- min-~). The animals were placed in a supine position on a heating pad to maintain body temperature between 36°C and 37°C. Polyethylene catheters (PE-50, 0.58 mm ID, Fa. Portex, Hythe, UK) were placed in the carotid artery and the jugular vein to assess central hemodynamics (arterial blood pressure, MAP; heart rate, HR) as well as to inject fluorescent latex beads and sodiumfluorescein. Supplementary chloralhydrate (3.6 mg- 100 g body wt -~) was given intraperitoneally if required. After transverse laparotomy, ligamentous attachments from the liver to the diaphragm and abdominal wall were dissected. Left hepatic lobar ischemia was induced by clamping the left hepatic artery together with the portal branch, while global ischemia was induced by clamping the proper hepatic artery and portal vein. An ischemic period of either 20 min or 60 rain was followed by reperfusion for a total of 60 min. The animals were positioned on their left side and the left liver lobe was exteriorized on an adjustable stage for intravital fluorescence microscopy. Sham-operated animals underwent an identical operative procedure without induction of ischemia and served as controls.

Experimental groups and protocol After a 30-min stabilization period, the animals were divided into four groups: non-ischemic animals served as controls (group A, n= 10) and underwent an identical experimental protocol except for clamping of the hepatic vessels. The period of sham-ischemia was 60 min. Animals in group B and C underwent left hepatic lobar ischemia of 20 min (n=9) and 60 min (n=9), respectively. Animals in group D were subjected to 20 min of global ischemia (n=ll). After a reperfusion period of 60 min, the phagocytic activity of Kupffer cells was analyzed by m vivo microscopy. Intravital fluorescence microscopy After 60 min of postischemic reperfusion, & vivo microscopy was performed using a modified Leitz-Orthoplan microscope (Leitz, Wetzlar, FRG) with epi-illumination, a CCD video camera (FK 6990; Prospective Measurements Inc., San Diego, CA, USA) and a videotape recorder (VO-5800 PS; Sony GmbH, Munich, FRG). Using a water immersion objective (W 25x/0.60, Leitz, Wetzlar, FRG), magnification of x600 was achieved on the video screen (PVM-1442 QM, diagonal: 330 mm, Sony GmbH, Munich, FRG). After 60 min reperfusion, a single bolus injection of plain fluorescent latex beads was performed intra-arterially through the carotid catheter (3-108-kg body wt-~ in 1 ml isotonic saline; diameter 1.1/am; Poly-

B_ VOLLMAR et al. sciences Inc., Warrington, PA) (3). Background staining was achieved by i.v. injection of sodium-fluorescein (2 ~a'nol- kg body wt-~).

Video analysis Quantitative assessment of the zonal distribution and kinetics of adherence of the latex beads was performed off-line by frame-to-frame analysis of the videotaped images, as described in detail previously (3). Within 10-15 randomly selected acini of each animal, the number of adherent latex beads was counted. Zonal distribution of adherent latex beads was facilitated by dividing sinusoids of each acinus into three segments of equal length and expressed as a percentage of particles visible within the acinus. For assessment of kinetics of adherence, 10-15 randomly selected microscopic fields (405×540 /am) per experiment were analyzed successively within 5 rain after injection. The kinetics of adherence was quantified by the number of beads moving in sinusoids as a percentage of all beads visible in the acini during observation for 10 s. Since variations in absolute number of beads per acinus were found in association with alterations of sinusoidal perfusion, all data were normalized and expressed as the percentage of particles visible in sinusoids per microscopic field (kinetic of adherence) or per acinus (zonal distribution). Statistics All data are expressed as mean_SEM. Nested-design two-way ANOVA (group and individual animal within group) was used for data of zonal distribution of beads adherence in acini. Non-normally distributed data were rank-transformed; multiple comparisons of means were performed with Tukey's honestly significant difference test. Adherence kinetics were calculated by a general linear model with log-transformed percentages of moving beads as dependent variables. Independent variables, included in the model, were time, time'-, time 3, experimental group, time*group interactions and animals nested within groups (3). Differences were considered to be significant if p<0.05. Calculations were performed with SAS procedure GLM (SAS Institute, Cary, NC). Results

No significant differences in central hemodynamics (MAP) were observed between control animals (group A: 89---5 mmHg) and animals subjected to lobar ischemia (group B: 93___4 mmHg and group C: 92---3 mmHg) throughout the experiment. Animals with global ischemia (group D) experienced transient systemic hypotension during ischemia (MAP: 53+__2 mmHg) with only incom-

ACTIVITY OF KUPFFER CELLS AFTER LIVER ISCHEMIA

303

were found in animals subjected to 20 min left hepatic lobar ischemia: l-rain and 3-min values were 15.9% (13.7% to 18.6%) and 5,3% (4.2% to 5.9%). However, after both 60 min left lobar ischemia (Fig. 2) and 20 min global ischemia, bead adherence was significantly delayed, when compared with non-ischemic controls, with 23.8% (20.3% to 27.9%) (Fig. 2) and 26.5% (23.1 to 30.4%) still moving after 1 min and 8.7% (7.4 to 10.1%) (Fig. 2) and 7.8% (6.8 to 8.9%) moving 3 min after injection, respectively.

Discussion

Fig. 1. Epi-illumination in vivo microscopy of the rat liver. Regular distribution of latex beads (appearing as bright white spots) with predominant localization of particles in periportal segments of sinusoids in a control liver. Background staining with sodium-fluorescein. V=postsinusoidal venule. Bar represents 100 ~n.

plete recovery o f M A P (79+_4 m m H g ) after 60 min o f reperfusion when c o m p a r e d to baseline ( M A P 102+_3 m m H g , p<0.01). Analysis o f zonal distribution of adherent latex beads revealed p r e d o m i n a n t periportal adherence in control livers (group A; Fig. 1, Table 1). Hepatic I/R resulted in a m o r e homogeneous distribution of latex beads within zonal segments o f sinusoids. After both 60 rain o f left l o b a r ischemia and 20 min of global ischemia, a significantly smaller percentage o f latex beads was found located in zone I, whereas significantly more beads adhered in zone 3 (Table 1; p < 0.01 for multiple comparisons o f group C and D vs. group A). Analysis o f the kinetics o f bead adherence in control animals revealed that 18.9% o f the beads visible per screen and per 10 s o f observation were still moving 1 rain after injection (95% confidence interval: 16.8% to 21.1%; Fig. 2). Three minutes after injection, this number had decreased to 5.3% (4.7% to 6.1%). N o significant differences

By use o f in vivo fluorescence microscopy in the rat liver, adherence o f latex beads, as described here, offers the distinct advantage of direct measurement o f phagocytic activity o f Kupffer cells. We have demonstrated that warm hepatic I/R depresses phagocytic activity of Kupffer cells with a consequent change in the pattern of zonal distribution o f adherent latex beads within sinusoids. In control livers the zonal gradient o f adherent latex beads reflects neither zonal differences in phagocytic activity of Kupffer cells, nor their local distribution (4), but reflects the marked first-pass effect with rapid proximal adherence and reduced delivery o f particles to downstream regions of the liver sinusoids (3). Therefore, I/Rinduced depression of phagocytic activity o f Kupffer cells accounts for the shift from periportal to midzonal and pericentral adherence o f latex beads, resulting in a more homogeneous zonal distribution. The underlying p a t h o m e c h a n i s m for the depression of

~

100-

E

TABLE 1 Zonal distribution of adherent latex beads

Group Group Group Group

A B C D

n

Beads distribution [%] Zone 1

Zone 2

Zone 3

261 116 128 177

56.9 + 1.2 49.8+__1.4* 48.2±1.2" 48.1_+1.3"

32.2 + I. 1 38.0±1.4 36_1±1.1 35.9_+1_2

10.9_+_0.7 12_2±0.9 15.7_+0.8 * 15.9+0.9 *

Group A: non-ischemic controls (n-- 10); group B: 20 min left hepatic lobar ischemia (n=9); group C: 60 min left hepatic lobar ischemia (n=9); group D: 20 min global hepatic ischemia (n=l 1). Mean-+ SEM. n, number of acini analyzed. Two-way ANOVA with Tukey's honestly significant difference test: * p<0.01 vs group A. All other intergroup comparisons were non-significant.

minutes alter

injection

Fig. 2. Kinetics of bead adherence. Logarithmic scale representation of moving beads as percentage of visible beads in microscopic fields observed for I0 s. Individual measurements in control animals (solid line, group A, n= 10) and in animals subjected to 60 min left hepatic lobar ischemia (dashed line, group C, n=9). Regression curves and 95% confidence limits of mean were calculated with a general linear model. Note that beads in group C (dashed line) adhered less rapidly than controls (solid line). The kinetics of adherence of latex beads in animals subjected to 20 min of global ischemia (group D, data not shown) was comparable to animals in group C_

304 phagocytosis of Kupffer cells has not been clarified. The fact that 60 rain, but not 20 min, of left hepatic lobar ischemia reduced phagocytic activity implies the impact of ischemia on Kupffer cell activity. However, since 20 min of global hepatic ischemia followed by reperfusion caused depression of phagocytic activity of Kupffer ceils comparable to that with 60 min lobar ischemia, reperfusion-associated events due to splanchnic vascular congestion may cause additional insult to Kupffer cells. Moreover, pathomechanisms associated with warm I/R seem to differ from those after cold I/R in liver transplantation, since increased phagocytic activity of Kupffer cells was reported after cold storage, transplantation and reperfusion of rat livers (3). Depression of the phagocytic activity of Kupffer cells may seriously impair the ability of the liver to counteract the arrival of bacterial endotoxin/lipopolysaccharides from the gut and may result in exposure of other cell types, i.e. hepatocytes, to injurious agents which are norreally removed from the circulation by Kupffer cells_ Since the function of sinusoidal lining Kupffer cells is critical to hepatocytes, the failure of Kupffer cells to remove lipopolysaccharides may have direct effects on hepatocytes and markedly affect hepatocellular structure and function. Using the hepatic uptake of 51Cr-labeled erythrocytes as a measure of reticulo-endothelial function, Crafa et al, (5) have shown that depressed phagocytic activity after 90 min of warm hepatic ischemia in the rat is associated with a marked release of transaminases (AST, ALT), extensive liver necrosis, and a survival rate of only 40%_ Holper et al. (2) reported that hepatic parenchymal cell lesions in baboons induced by 90 min ischemia were compatible with survival when associated with adequate reticulo-endothelial function, but recovery did not occur when reticulo-endothelial function was significantly inhibited. In view of increased lipopolysaccharide-translocation

B. VOLLMAR et al. from the gut following splanchnic vascular congestion (group D), Kupffer cell depression may have major consequences for the course of postischemic hepatic dysfunction and subsequent multiple organ failure, since blockade of these cells during intra-abdominal sepsis in mice has been shown to predispose to high mortality because of septic complications (6). In addition, there is clinical evidence that compromised hepatic reticulo-endothelial function increases the prevalence of systemic endotoxemia (7). Therefore, we suggest (i) that depression of the phagocytic activity of Kupffer cells following I/R plays a role in subsequent organ failure and (ii) that novel therapeutic regimes should include preservation of Kupffer cell function, h7 vivo monitoring of phagocytic activity of Kupffer cells by adherence of latex beads seems to be an adequate tool to clarify the significance of depressed Kupffer cell phagocytosis in postischemic liver damage and remote organ failure. References

1. Nolan JP. Endotoxin, reticuloendothelial function, and liver injury. Hepatology 1981; 1: 458-65. 2. Holper K, Olcay I, Kitahama A, et al. Effect of ischemia on hepatic parenchymal and reticuloendothelial function in the baboon. Surgery 1974; 76: 423-32. 3. Post S, Gonzalez AP, Palma P, Rentsch M, Stiehl A, Menger M_ Assessment of hepatic phagocytic activity by h7 vivo microscopy after liver transplantation in the rat. Hepatology 1992; 16: 803-9. 4. Te Koppele JM, Thurman RG. Phagocytosis by Kupffer cells predominates in pericentral regions of the liver lobule. Am J Physiol 1990; 259: G814-21. 5. Crafa F, Gugenheim J, Saint-Paul M-C, et al. Protective effects of prostaglandin El on normothermic liver ischemia_ Eur Surg Res 1991; 23: 278-84. 6. Callery MP, Kamei T, Flye MW. Kupffer cell blockade increases mortality during intra-abdominal sepsis despite improving systemic immunity. Arch Surg 1990; 125: 36~11. 7. Rimola A, Soto R, Bory F, Arroyo V, Piera C, Rodes J. Reticuloendothelial system phagocytic activity in cirrhosis and its relation to bacterial infections and prognosis. Hepatology 1984; 4: 53-8.