- Email: [email protected]
Role of hepatocytes in direct clearance of lipopolysaccharide in rats
GASTROENTEROLOGY1995;109:1969-1976
Role of Hepatocytes in Direct Clearance of Lipopolysaccharide in Rats YOSHiHIRO MIMURA, SHOTARO SAKISAKA, MASARU HARADA, MICHIO SATA, and KYUICHI TANIKAWA Second Departmentof Medicine, Kurume UniversitySchool of Medicine, Kururne,Japan
Background & Aims: The liver is the clearance organ for lipopolysaccharide (LPS). The aim of this study was to investigate the biliary excretion of LPS using fluorescein isothiocyanate (FITC)-Iabeled LPS. Methods: After FITC-LPS was injected intravenously into rats, the cellular localization of fluorescence in the liver was examined and the biliary excretion of fluorescence was measured. The effects of gadolinium chloride, a blocker of Kupffer cells, and colchicine, an inhibitor of microtubules, on the biliary excretion of fluorescence was investigated, and bile was analyzed using high-performance liquid chromatography. Results: Laser scanning confocal microscopy showed that fluorescence was taken up by hepatocytes 5 minutes after injection of FITC-LPS into the portal vein. When FITC-LPS was injected into the portal vein, fluorescence was rapidly secreted into bile, peaking at 20 minutes, and 25.1% of the injected dose appeared in bile within 60 minutes. When the same dose of FITC-LPS was injected into the tail vein, 15.8% appeared in bile within 60 minutes. Chromatography showed that FITC-LPS was excreted into bile in an unchanged form over a period of 20 minutes after injection. Colchicine significantly reduced the biliary excretion of fluorescence, but gadolinium chloride had no effect. Conclusions: LPS was directly and effectively processed by hepatocytes and secreted into the bile canalicular system via a microtubule-dependent vesicular pathway.
t ndotoxin is the general term for a class of lipopolyt saccharide (LPS) molecules located in the outer cell membrane of gram-negative enteric bacteria. The presence of endotoxin in the systemic circulation can produce a variety of pathophysiological effects that are manifested during gram-negative infections, LPS is therefore considered a probable cause of septic shock. 1 The liver has been found to play a major role in clearing circulating LPS from the blood. 2-v In a previous study in which 5*Cr-labeled LPS was injected intravenously into normal mice, > 8 5 % of the dose was found to be trapped in the liver within 1 hour.: High concentrations of radioactivity were observed in the gallbladder bile of rabbits given an intravenous injection of
E
125I-labeled LPS. 8 When *4C-labeled endotoxin was injected into rats, radioactivity was excreted mainly in the feces. 4 The presence of radioactivity in gallbladder bile and feces after intravenous administration of radiolabeled LPS suggests that LPS and/or its metabolites may be processed by hepatocytes and secreted into the biliary system. Evidence suggests that LPS is internalized, modified, and released into the bloodstream by macrophages, including Kupffer cells. 2'5'<8-1. Furthermore, in the liver, a histochemical study using an anti-LPS antibody showed that LPS was initially associated with Kupffer cells and later with hepatocytes. 5 On the other hand, it has been reported that LPS binds to receptors on hepatocytes in vitro. 12'13 However, the physiological function of LPS receptor on hepatocytes has not been well defined. It has not been shown whether hepatocytes take up LPS without prior metabolism by Kupffer cells and secrete it into bile, although it is known that a peculiar structure of LPS was not metabolized and secreted into bile by hepatocytes. 14 To show the role of hepatocytes in LPS clearance, we investigated the cellular distribution and biliary excretion of fluorescein isothiocyanate (FITC)-labeled Escherichia coli O55:B5 LPS in rats. In addition, we studied the effects of gadolinium chloride (GdC13), a blocker of Kupffer cells, and colchicine, an inhibitor of microtubules, on the biliary excretion of LPS. Materials
and Methods
Chemicals FITC-labeled and -unlabeled LPS were purchased from List Biological Laboratories Inc. (Campbell, CA). LPS was prepared from E, coli (serotype O55:B5) using phenol extraction and purified chromatographically. FITC-LPS has 6.4 mol of FITC (isomer I) per mole of-LPS (technical data of FITC-LPS). Abbreviations used in this paper: FITC,fluoresceinisothiocyanate; GdCla, gadolinium chloride; HPLC, high-performanceliquid chromatography; LPS, lipopolysaccharide. © 1995 by the AmericanGastroenterologicalAssociation 0016-5085/95/$3.00
1970 MIMURAET AL.
The same lots were used for one set of experiments. The two LPS preparations were stored at 4°C and dissolved in isotonic, sterile saline before use. FITC-labeled bovine albumin (FITCalbumin), colchicine, and lumicolchicine were purchased from Sigma Chemical Co. (St. Louis, MO). FITC isomer and sodium pentobarbital were obtained from Polysciences Inc. (Warrington, PA) and Abbott Laboratories (North Chicago, IL), respectively. The acetonitrile used in the mobile phase was of high-performance liquid chromatography (HPLC) grade, and all other chemicals were of analytical-reagent grade.
GASTROENTEROLOGYVol. 109, No. 6
FITC-LPS to assess the participation of Kupffer ceils in clearance of LPS. Colchicine, an inhibitor of microtubules, and lumicolchiclue, an isomer of colchicine, were administered intraperitoneally (1 gg/g body wt) 4 hours before the administration of FITC-LPS to investigate the intercellular transport and processing kinetics of FITC-LPS in the liver. In addition, we examined the relationship between the injected dose of FITC-LPS at a concentration of 5 - 6 0 /.tg/300 g body wt and an amount of the FITC moiety secreted into bile.
Laser Scanning Confocal Microscopy HPLC of FITC-LPS The cellular distribution of FITC-LPS in the liver was examined in normal and GdC13-treated rats. FITC-LPS was injected into the portal vein at a dose of 50 btg/300 g body wt. Livers were then perfused in situ with 50 mL ofperiodatelysine-paraformaldehyde medium to remove gross blood and for fixation 5, 10, 20, or 60 minutes after injection of FITCLPS. Furthermore, pieces of the liver tissue were fixed in 4% phosphate-buffered formaldehyde for 12 hours and processed for paraffin embedding. Liver sections (5 btm) were examined using laser scanning confocal microscopy (LSM-GB200; Olympus, Tokyo, Japan) at an excitation wavelength of 488 nm and an emission wavelength of 515 nm.
Experimental Protocol Male Wistar rats weighing 2 5 0 - 3 5 0 g and bred under specific pathogen-free conditions were used in all experiments, which were performed between 10 AM and 4 PM. After rats were anesthetized with intraperitoneal injection of sodium pentobarbital (Nembutal, 39 mg/kg body wt), the common bile duct was exposed and cannulated with a polyethylene catheter, which had outer and inner diameters of 900 and 320 ~tm, respectively, and a length of 10 cm. Bile collection for experiments was started 3 0 - 4 5 minutes after bile duct cannulation when bile flow became stabilized. Using a 27-gauge needle, we injected 50 p.g/300 g body wt of FITC-LPS or FITC-albumin in 0.5-mL of saline into a branch of the portal vein or the tail vein. Control rats were given the same amount of unlabeled LPS in 0.5 mL of normal saline. Bile was collected every 10 minutes for 70 minutes after administration of FITCLPS while the abdomen was covered with wet gauze. The collected bile was centrifuged at 5000 rpm for 15 minutes at room temperature. The concentration of the FITC moiety in the supernatant was determined using an F3000 fluorescence spectrophotometer (Hitachi Ltd., Tokyo, Japan) at an excitation wavelength of 496 nm and an emission wavelength of 517 nm. In a preliminary study, the fluorescent intensity of FITC-LPS dissolved in bile was linearly related to FITC-LPS concentration ranging from 0.1 to 10 ~g/mL. Thus, the FITCLPS concentration in bile was determined by this standard curve. The FITC moiety represents the amount of unchanged LPS and/or a metabolic product carrying the FITC molecule. GdC13, a blocker of Kupffer cells, was administered intravenously (5 mg/kg body wt) 24 hours before administration of
To determine whether the LPS secreted in bile was unchanged or represented a metabolic product carrying the FITC molecule, we analyzed bile using HPLC according to the method of Martin and Ghabrial. 15 Bile obtained over a period of 20 minutes after administration of FITC-LPS was centrifuged at 5000 rpm for 15 minutes at room temperature. The supernatant was collected and then dissolved in 20 volumes of chloroform-methanol (1:2 by volume). The lower phase contents were dried, and the residue was redissolved in 200 mL of distilled water and centrifuged at 10,000 rpm for 30 minutes at 4°C. Chromatography was performed on 20 btL of the supernatant using a constant-flow high-pressure liquid chromatograph consisting of an LC-6A pump system equipped with an SCL-6A controller and an RF-530 spectrofluorometer (Shimadzu, Kyoto, Japan) at an excitation wavelength of 496 nm and an emission wavelength of 517 nm with excitation and emission slit widths of 10 nm. The column used was an Ultrahydrogel linear column (ID, 300 × 7.8 mm; Waters Associates, Milford, MA) packed with a gel of cross-linked hydroxylated polymethacrylate. The mobile phase was 100 mmol/L phosphate buffer (pH 7.75) containing 20% (vol/vol) acetonitrile at a flow rate of 0.4 mL/min.
Statistical Analysis Statistical analysis of the data was performed using Student's t test. P values of <0.05 were considered to indicate significance of differences between means.
Results Cellular Distribution of FITC-LPS in the Liver Laser scanning confocal microscopy revealed fluorescence in hepatic macrophages (Kupffer cells) and hepatocytes of liver specimens 10 minutes after 50 ~g/300 g body wt of FITC-LPS was injected into a branch of the portal vein (Figure 1A). Fluorescence intensity was stronger in Kupffer cells than in hepatocytes. Faint but significant fluorescence was detected in Kupffer cells and hepatocytes 5 minutes after injection (data not shown). W h e n FITC alone was injected, cellular fluorescence was much weaker, and its intensity was similar in Kupffer cells and hepatocytes
December 1995
HEPATOCYTIC CLEARANCE OF LIPOPOLYSACCHARIDE 1971
Biliary Excretion of FITC-LPS Neither 50 ~g/300 g body wt of FITC-LPS (Figure 2A) nor the same dose of unlabeled LPS (data not shown) affected bile flow during a 70-minute period after the portal vein administration. Excretion of the FITC moiety into bile peaked at 20 minutes, gradually decreasing during the next 50 minutes (Figure 2B). The cumulative amount of the FITC moiety excreted into bile was 25.1% +__5.7% (mean ___SD; n = l l ) of the injected dose of FITC-LPS at 60 minutes (Figure 2C).
Biliary Excretion of FITC-LPS at the Various Injected Doses When 5-60 ~g/300 g body wt of FITC-LPS was injected into the portal vein, the biliary excretion of the FITC moiety increased gradually according to the injected dose of FITC-LPS (Figure 3A). The peak concentration of the excreted FITC moiety grew in a linear fashion (Figure 3B). However, administration of >80 btg/300 g body wt of FITC-LPS reduced bile flow (data not shown).
Biliary Excretion of FITC-Albumin
Figure 1. Laser scanning confocal micrograph of fluorescence in the rat liver. (A) Ten minutes after portal vein injection of 50 pg/300 g body wt of FITC-LPS. Fluorescence was localized in hepatocytes and Kupffer cells (arrows) in the sinusoidal space. (B) Sixty minutes after portal vein injection of FITC alone. Fluorescence intensity was not different between hepatocytes and Kupffer cells when FIT(:; alone was injected. (C) Sixty minutes after portal vein injection of FITC-LPS. Fluorescence was detected in hepatocytes and nonparenchymal cells, and fluorescence intensity was reduced along the acinus of the liver. Kupffer cells (arrows) showed stronger fluorescence than hepatocytes. C, central vein; P, portal area.
(Figure 1B). Sixty minutes after injection of FITC-LPS, fluorescence was observed mainly in hepatocytes and nonparenchymal ceils close to the terminal portal venule, and fluorescence intensity was reduced in the pericentral area (Figure 1C). Fluorescence was observed within the hepatic acinus from 5 to 60 minutes after injection of FITC-LPS. Neither margination of neutrophils and mononuclear cells nor fibrinous deposits in the hepatic sinusoids and the spaces of Disse were observed within 60 minutes after injection of FITC-LPS (Figure 1C). Furthermore, the hepatic acinus was composed of normal-appearing hepatocytes (Figure 1C). The fluorescence for FITC-LPS was also distributed in splenic macrophages and pulmonary dust ceils (data not shown).
When 50 J,tg/300 g body wt of FITC-albumin was injected into the portal vein, excretion of the FITC moiety did not peak at 20 minutes (Figure 4A) and only 2% of the injected dose appeared in bile within 30 minutes (Figure 4B).
HPLC of FITC-LPS Chromatograms showed that fluorescent peaks of the preinjected FITC-LPS (Figure 5A) and the FITC moiety secreted into bile (Figure 5B) had the same retention time.
Biliary Excretion After Portal Vein and Systemic Injections of FITC-LPS Biliary excretion of the FITC moiety by the liver after intravenous injection via the tail vein was significantly lower than that after portal vein injection via the mesenteric vein (Figure 6). Sixty minutes after intravenous administration of FITC-LPS, 15.8% + 4.7% (mean ___ SD; n = 5) of the injected dose appeared in bile. Biliary excretion of the FITC moiety in rats given a systemic injection with FITC-LPS decreased to 62.9% of that found in rats given a portal vein injection (P < 0.001).
Effect of GdCla on Biliary Excretion of FITC-LPS When 50 btg/300 g body wt of FITC-LPS was injected intraportally into GdC13-treated rats, biliary ex-
1972
MIMURA ET AL.
GASTROENTEROLOGY Vol. 109, No. 6
cretion of the FITC moiety was the same as in control animals (Figure 7).
Uptake of FITC-LPS by Kupffer Cells Pretreated With GdCla The fluorescence intensity in Kupffer cells was reduced by GdCI 3 10 minutes after injection of FITC-LPS (Figure 8). GdC13 blocked endocytosis of FITC-LPS by Kupffer cells from 5 to 60 minutes after injection of FITC-LPS.
Effect of Colchicine on Biliary Excretion of FITC-LPS Sixty minutes after administration of FITC-LPS, 17.8% + 2.7% (mean + SD; n = 8) of the injected
A
250
p V10min
II[-i i I
200
lso
i
I i
O
m
I00
50
a
0
I
|
I
|
I
I
I
dose appeared in bile when 50 [.tg/300 g body wt of FITC-LPS was injected intraportally into colchicinetreated rats. Colchicine, an inhibitor of microtubules, significantly inhibited biliary excretion of the FITC moiety to 70.9% of control within 60 minutes after administration of FITC-LPS (Figure 9). On the other hand, lumicolchicine, an isomer of colchicine that does not affect microtubules, did not affect biliary excretion of the FITC moiety (data not shown). Colchicine and lumicolchicine did not affect bile flow.
Discussion The liver is the main clearance organ for LPS, and Kupffer cells have been reported to be responsible for hepatic clearance of LPS. 2'5'6's However, it is not well known whether the parenchymal liver cells (hepatocytes) are responsible for the direct clearance of LPS, whether hepatocytes take up LPS and secrete it into bile, or which transport system excretes LPS into bile from the liver. In the present study, we studied the route and rate of in vivo excretion of LPS in normal rats using FITC-LPS. The FITC molecule in FITC-LPS has been found to bind covalently to amine groups of LPS. 15 FITC-LPS was positive for the limulus amebocyte lysate assay (our unpublished observation). FITC-LPS is complexed to LPS-binding protein, and the resulting complex is recognized by CD14 receptor on monocytes in vitro) 6 FITC-LPS also
20' A
g.~ ""
B
,,g/~ 40-
t,~nl
10'
0
,~ 30
0
C 30'
% of injectedFITC-LPS
! •
•
°°, N
20,
/ °,
.//
,g (r=0.822107) °.,° • ,,,
//
/°'*
• ~
10'
/° //•
•
.,° •
1#•• /~*'°e •
0
10
20 30 40 50 60 Minutes after injection
70
80
Figure 2. Biliary excretion of FITC moiety after portal vein injection of 50 # g / 3 0 0 g body wt of FITC-LPS. (A) Effect of FITC-LPS on bile flow. (B) Excretion of FITC moiety into bile quantified as fluorescence intensity. (C) Cumulative amount of the FITC moiety excreted into bile. Results are expressed as mean _+ SD obtained from 11 rats.
0
20 4O 6O 80 Mitres afterinjection
J 10
Injected
,
,
20 dose
•
i
•
i
,
i
-
i
30 40 50 60 ( b~g/300 gm body wt)
•
)
70
Figure 3. Relationship between the injected dose of FITC-LPS and biliary excretion of the FITC moiety. (A) Representative data of biliary excretions of the FITC moiety after portal vein injection of 5 - 6 0 #g/ 300 g body wt of FITC-LPS. (B) The peak concentration of the secreted FITC moiety in A is plotted along the ordinate against the injected dose of FITC-LPS on the abscissa.
December 1995
HEPATOCYTIC CLEARANCE OF BPOPOLYSACCHARIDE
activates B cells in vivoJ 7'*s In our confocal microscopy study, fluorescence of FITC-LPS was predominantly distributed in splenic macrophages and pulmonary dust cells as well as in Kupffer cells after venous injection of FITCLPS. This cellular distribution was similar to that of LPS found in previous studies using autoradiography or immunohistochemical methods. ~'5'<8 These results suggested that FITC-LPS could be used as a tracer of LPS. Electron-microscopic autoradiography, cell isolation, and partial purification of cell types showed that 15Crlabeled LPS was taken up by Kupffer cells, endothelial cells, and hepatocytes in rats. .9 Van Bossuyt et al. showed that 3H-labeled LPS was localized in the mitochondria, cell membranes, and the periphery of hepatocytes 15 minutes after injection of the tracer. 2° In the present study, laser scanning confocal microscopy also showed hepatocytic uptake of FITC-LPS. Microscopy showed that FITC-LPS was processed by hepatocytes as well as by Kupffer cells 5 minutes after injection via the mesenteric vein. Uptake was greater in hepatocytes of acinar zones 1 and 2 than in those of acinar zone 3, probably because hepatocytes surrounding the terminal portal re-
~t
1973
nule (acinar zone 1) were exposed to a high concentration of FITC-LPS. In the present study, biliary excretion of the FITC moiety increased rapidly, peaking 20 minutes after injection of FITC-LPS, unlike FITC-albumin. Colchicine inhibited the biliary excretion of FITC-LPS. The biliary secretory curve of FITC-LPS was similar to those found in previous studies for immunoglobulin A , 21 horseradish peroxidase, 22'23 insulin, 24 inulin, 25 and copper, 26 which appeared to be transported directly to bile without passing through the lysosomal compartment of hepatocytes. Because colchicine was known to interfere with microtubules and related membrane functions, 22'25-27 these resuits suggested that FITC-LPS was transported to bile via a microtubule-dependent vesicular pathway that rapidly transverses hepatocytes from the sinusoidal membrane to the bile canaliculus. LPS has been found to be taken up in all organs by
A
i-r., CO
30 p g/ml
~ 20'
g ~
F1TC-LPS 10
F1TC-AIb 0
i
B
i
~
!
i
aj oJ
~. . . .
%of injecteddose 2O
~
FITC-LPS
!1o
B
FITC-Alb 0
0
10 20 Minutes after injection
30
.
40
Figure 4. Biliary excretion of the FtTC moiety after portal vein injection of 50 p g / 3 0 0 g body wt of FITC-albumin (Alb) (O) and FITC-LPS (0). (A) Excretion of FITC moiety into bile quantified as fluorescence intensity. (B) Cumulative amount of the FITC moiety excreted into bile. Results of FITC-albumin administration are expressed as mean ± SD obtained from 3 rats.
I 0
Retention time
I 40 rain
Figure 5, HPLC of (A) preinjected FITC-LPS and (B) FITC moiety in collected bile run on an Ultrahydrogel linear column at a flow rate of 0.4 mL/min.
1974
MIMURA ET AL.
GASTROENTEROLOGY Vol. 109, No. 6
granulocytes and macrophages, including Kupffer cells, and then LPS or its metabolites are later redistributed into hepatocytes. 28'29 However, it probably takes longer than 20 minutes for phagocytic cells to process LPS; then LPS and/or its metabolites are transferred from the blood stream to bile through hepatocytes. In the present study, we used 50 [.tg/300 g body wt of FITC-LPS. It is possible that the high dose of LPS saturates the Kupffer cell pathway, resulting in preferential processing by hepatocytes. However, fluorescence intensity in Kupffer cells after injection of 50 /.tg/300 g body wt of FITC-LPS was weaker than after an injection of 70 g/300 g body wt (our unpublished observation), suggesting that the Kupffer cell was unsaturated with LPS after injection of 50 btg/300 g body wt of FITCLPS. Furthermore, our dose-excretion data showed that the biliary efflux of FITC moiety increased with the high dose of injected FITC-LPS, representing a straight line passed through the nearby origin of the coordinates (Figure 3B). This suggested that hepatocytes may participate
in LPS clearance at the low concentration of LPS in the blood. These findings suggested that hepatocytes and Kupffer cells independently took up LPS from the blood stream. Administration of 50 /.tg/300 g body wt of FITCLPS had no effect on either the rate of bile flow or the histological alterations in various tissues. However, administration of > 8 0 ~g/300 g body wt of FITC-LPS reduced bile flow so that we could not examine the maximum capacity of hepatocytes in biliary secretion of LPS. After intravenous injection of radiolabeled LPS, the radioactivity is secreted into the biliary system. 4'8 After injection of unlabeled LPS, ~-hydroxymyristic acid, a peculiar structure of lipid A moiety in LPS, is also secreted into bile. 14 However, it is not known whether the molecules secreted into bile were intact LPS or its metabolites. Our HPLC suggested that FITC-LPS was processed and secreted into bile by hepatocytes predominantly in an unmetabolized form during a period of 20 minutes after injection. Intravenous injection of GdC13 strongly reduced liver uptake of colloidal carbon and radioactively labeled for-
0 o
A
. t* g l O m i n
i
5.0 4.0 3.0
g
2.0 1.0
0 B 30 ] %of injectedFITC-LPS
m
0.0
•
m
-
|
!
|
m
e
m
B ~8
"~
10'
i 0
10
20
30
40
50
60
70
3O
10' 0
80
m
10
|
20
i
u
i
30
40
50
!
60
m
70
80
Minutes after injection
Minutes after injection Figure 6. Biliary excretion of the FITC moiety after portal vein and systemic injections of FITC-LPS. Fifty micrograms per 300 g body wt of FITC-LPS was administered via the mesenteric vein (0) or the tail vein (t). (A) Biliary output of the FITC moiety. (B) Cumulative amount of the FITC moiety in bile. Results of systemic injections of FITC-LPS are expressed as mean -- SD obtained from 3 rats. * * P < 0.05; * * * P < 0.001.
0
Figure7.
Effect of GdCI3 on biliary excretion of the FITC moiety after portal vein injection of FITC-LPS, Fifty micrograms per 300 g body wt of FITC-LPS was administered into GdCl3-treated rats (©) and control rats (e), (A) Biliary output of the FITC moiety. (B) Cumulative amount of the FITC moiety in bile. Results in GdCls-treated rats are expressed as mean ± SD obtained from 6 rats, There was no statistical significance between the groups.
December 1995
eign erythrocytes)° It is likely that accumulation of these particles by the liver is mainly caused by phagocytosis by Kupffer cells, j° The present study also showed that GdC13 inhibited endocytosis of FITC-LPS by Kupffer cells. However, GdCI 3 did not reduce the biliary secretion of the FITC moiety by hepatocytes. These findings suggest that Kupffer cells may not participate in the early excretion of LPS into bile. If all of the FITC-LPS had bound to some plasma proteins or been processed by ceils other than hepatocytes, there would probably have been no difference in the biliary excretion of the FITC moiety between peripheral and portal vein administration. However, hepatic clearance of the FITC moiety in rats given systemic injections decreased to 62.9% of that in animals given portal vein injections. Portal vein administration of FITC-LPS would increase the concentration of LPS delivered to hepatocytes compared with systemic administration. Our results suggest that hepatocytes may take up FITC-LPS directly and excrete it into bile. Previous studies have shown that LPS administered via the portal vein is cleared more effectively than that administered by other routes, 31'32 which may explain why portal vein injection of LPS was associated with a lower mortality rate than systemic administration in rats. 33 Recently, Parent showed that rat hepatocytes had membrane receptors for LPS. 1~ However, the physiological function of hepatocyte LPS receptors is unclear. The receptors may be related to the direct clearance of LPS to bile. In this study, a significant amount of LPS was directly and effectively removed from the circulation and trans-
HEPATOCYTIC CLEARANCE OF LIPOPOLYSACCHARIDE
1975
O2
!'1 2"
8
1"
ol %of injectedFITC-LPS
T
l
30-
O
2°i
~t~~4~
~
10'
6 0' 0
10
20
30
40
50
60
70
80
Minutes after injection Figure 9. Effect of colchicine on (A) biliary output and (B) cumulation amount of the FITC moiety secreted into bile after portal vein injection of 50 btg/300 g body wt of FITC-LPS. e, Colchicine-pretreated rats; O, data from control rats as in Figure 2. Results in colchicine-pretreated rats are expressed as mean _+ SD obtained from 8 rats. * * P < 0.05; * * * P < 0.001.
ferred to bile by liver parenchymal cells through a microtubule-dependent vesicular pathway. The rest of the injected LPS might be cleared by phagocytic cells (granulocytes, monocytes, and macrophages), as described previously.2,5,6,8 Studies have shown that endotoxemia is associated with liver diseases such as obstructive jaundice,34 cholestasis, ~5 and liver cirrhosis) 6 Disturbed biliary excretion is also associated with these diseases, 3v suggesting that a disturbance in LPS excretion in addition to increased absorption of LPS from the intestine 38 may contribute to endotoxemia.
References
Figure 8. Laser scanning confocal micrograph of fluorescence in the liver pretreated with GdCI3. Ten minutes after portal vein injection of 50 btg/300 g body wt of FITC-LPS. Compare with Figure 1A. GdCl3 blocked endocytosis of FITC-LPS by Kupffer cells.
1. Nogare ARD. Southwestern Internal Medicine Conference: septic shock. Am J Med Sci 1 9 9 1 ; 3 0 2 : 5 0 - 6 5 . 2. Carey FJ, Braude AL, Zalesky M. Studies with radioactive endotoxin. III. The effect of tolerance on the distribution of radioactivity after intravenous injection of Escherichia coil endotoxin labelled with 51Cr. J Clin Invest 1 9 5 8 ; 3 7 : 4 4 1 - 4 5 7 . 3. Braude AI, Carey FJ. Zalesky M. Studies with radioactive endotoxin. II. Correlation of physiological effects with distribution of radioactivity in rabbits injected with lethal doses of E. coli endo-
1976
4.
5.
6.
7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
MIMURA ET AL.
toxin labelled with radioactive sodium chromate. J Clin Invest 1955; 3 4 : 8 5 8 - 8 6 6 . Klein B, Freudenberg MA, Galanos C. Excretion of radioactivity in faeces and urine of rats injected with 3H,14C-lipopolysaccharide. Br J Exp Pathol 1 9 8 5 ; 6 6 : 3 0 3 - 3 0 8 . Freudenberg MA, Freudenberg N, Galanos C. Time course of cellular distribution of endotoxin in liver, lungs and kidneys of rats. Br J Exp Pathol 1 9 8 2 ; 6 3 : 5 6 - 6 5 . Musson RA, Morrison DC, Uievitch RJ. Distribution of endotoxin (lipopolysaccharide) in the tissues of lipopoiysaccharide-responsive and unresponsive mice. Infect Immun 1 9 7 8 ; 2 1 : 4 4 8 - 4 5 7 . Zlydaszyk JC, Moon RJ. Fate of 5~Cr-labeled lipopolysaccharide in tissue culture cells and livers of normal mice. Infect Immun 1976; 1 4 : 1 0 0 - 1 0 5 . Mathison JC, Ulevitch RJ. The clearance, tissue distribution and cellular localization of intravenously injected lipopolysacchadde in rabbits. J Immunol 1979; 1 2 3 : 2 1 3 3 - 2 1 4 3 . Peterson AA, Munford RS. Dephosphorylation of the lipid A moiety of Escherichia coil lipopolysaccharide by mouse macrophages. Infect Immun 1 9 8 7 ; 5 5 : 9 7 4 - 9 7 8 . Munford RS, Hall CL. Uptake and deacylation of bacterial lipopolysaccharides by macrophages from normal and endotoxin-hyperresponsive mice. Infect Immun 1 9 8 5 ; 4 8 : 4 6 4 - 4 7 3 . Fox ES, Thomas P, Broitman SA. Clearance of gut-derived endotoxins by the liver. Gastroenterology 1 9 8 9 ; 9 6 : 4 5 6 - 4 6 1 . Pagani R, Portoles MT, Municio AM. The binding of Escherichia coil endotoxin to isolated rat hepatocytes. FEBS Lett 1981; 1 3 1 : 1 0 3 - 1 0 7 . Parent JB. Membrane receptors on rat hepatocytes for the inner core region of bacterial lipopolysaccharide. J Biol Chem 1990; 2 6 5 : 3 4 5 5 - 3 4 6 1 . Maitra SK, Rachmilewitz D, Eberle D, Kaplowitz N. The hepatocellular uptake and biliary excretion of endotoxin in the rat. Hepatoiogy 1981; 1 : 4 0 1 - 4 0 7 . Martin G, Ghabrial H. High-performance liquid chromatographic method to resolve and determine lipopolysaccharide sub-groups of Escherichia coil endotoxin in isolated perfused rat liver perfusate. J Chromatogr 1 9 9 2 ; 5 7 4 : 2 0 5 - 2 1 1 . Gallay P, Jongeneel CV, Barras C, Burnier M, Baumgartner JD, Glauser MP, Heumann D. Short time exposure to lipopolysacchadde is sufficient to activate human monocytes. J Immunol 1993; 1 5 0 : 5 0 8 6 - 5 0 9 3 . Moeller G, Gronowicz E, Persson U, Coutinho A, Moiler E, Hammarstrom L, Smith E. Spleen cells from animals tolerant to a thymus-dependent antigen can be activated by lipopolysaccharide to synthesize antibodies against the tolerogen. J Exp Med 1976; 1 4 3 : 1 4 2 9 - 1 4 3 8 . Skelly RR, Munkenbeck P, Morrison DC. Stimulation of T-independent antibody responses by hapten-lipopolysaccharides without repeating polymeric structure. Infect Immun 1979;23: 287-293. Ruiter DJ, Van Der Meulen J, Brouwer A, Hummel M JR, Mauw BJ, Van Der Ploeg JCM, Wisse E. Uptake by liver cells of endotoxin following its intravenous injection. Lab Invest 1 9 8 1 ; 4 5 : 3 8 - 4 5 . Van Bossuyt H, De Zanger RB, Wisse E. Cellular and subcellular distribution of injected lipopolysaccharide in rat liver and its inactivation by bile salts. J Hepatol 1 9 8 8 ; 7 : 3 2 5 - 3 3 7 . Renston RH, Jones AL, Christiansen WD, Hradek GT, Underdrown BJ. Evidence for a vesicular transport mechanism in hepatocytes for biliary secretion of immunoglobulin A. Science 1980; 208:1276-1278.
GASTROENTEROLOGY Vol. 109, No. 6
22. Sakisaka S, Ishii Y, Ueda H, Matsumoto H, Eguchi T, Tanikawa K. The mechanism of intracellular transport of horseradish peroxidase and the role of micro~ubules in the hepatocyte. J Clin Electron Microsc 1 9 8 2 ; 1 5 : 5 - 6 . 23. Gondo K, Sakisaka S, Abe H, Tanikawa K. Observation of 1251insulin in hepatocytes by light and electron microscope autoradiography. J Clin Electron Microsc 1 9 8 5 ; 3 4 : 6 6 - 6 7 . 24. Lorenzini 1, Sakisaka S, Meier PJ, Boyer JL. Demonstration of a transcellular vesicle pathway for biliary excretion of inulin in rat liver. Gastroenterology 1 9 8 6 ; 9 1 : 1 2 7 8 - 1 2 8 8 . 25. Harada M, Sakisaka S, Yoshitake M, Shakadoh S, Gondo K, Sata M, Tanikawa K. Biliary copper excretion in acutely and chronically copper-loaded rats. Hepatology 1 9 9 3 ; 1 7 : 1 1 1 - 1 1 7 . 26. Sakisaka S, Ng OC, Boyer JL. Tubulovesicular transcytotic pathway in isolated rat hepatocyte couplets in culture.Effect of colchicine and taurocholate. Gastroenterology 1 9 8 8 ; 9 5 : 7 9 3 - 8 0 4 . 27. Manfredi JJ, Horwitz SB. An antimitotic agent with a new mechanism of action. Pharmacol Ther 1 9 8 4 ; 2 5 : 8 3 - 1 2 5 . 28. Freudenberg MA, Kleine B, Galanos C. The fate of lipopolysaccharide in rats: evidence for chemical alteration in the molecule. Rev Infect Dis 1 9 8 4 ; 6 : 4 8 3 - 4 8 7 . 29. Freudenberg MA, Galanos C. Bacterial lipopoiysaccharides: structure, metabolism and mechanisms of action. Int Rev Immunol 1990;6:207-221. 30. Husztik E, Lazar G, Parducz A. Electron microscopic study of Kupffer-cell phagocytosis blockade by gadolinium chloride. Br J Exp Pathol 1 9 8 0 ; 6 1 : 6 2 4 - 6 3 0 . 31. Jacob AI, Goldberg PK, Bloom N, Degenshein GA, Kozinn PJ. Endotoxin and bacteria in portal blood. Gastroenterology 1977; 7 2 : 1 2 6 8 - 1 2 7 0 . 32. Wolter J, Liehr H, Grun M. Hepatic clearance of endotoxins: differences in arterial and portal venous infusion. J Reticuloendothel Soc 1 9 7 8 ; 2 3 : 1 4 5 - 1 5 2 . 33. Mori K, Matsumoto K, Gans H. On the in vivo clearance and detoxification of endotoxin by lung and liver. Ann Surg 1973;177:159-163. 34. Ddvas G, James O, Wardle N. Study of reticuloendothelial phagocytic capacity in patients with cholestasis. BMJ 1 9 7 6 ; 1 : 1 5 6 8 1569. 35. Greve JW, Gouma DJ, Buurman WA. Complications in obstructive jaundice: role of endotoxins. Scand J Gastroenterology Suppl 1992; 1 9 4 : 8 - 1 2 . 36. Tachiyama G, Sakou M, Kambayashi J, lijima S, Tsujinaka T, Mori T. Endogenous endotoxemia in patients with liver cirrhosis--a quantitative analysis of endotoxin in portal and peripheral blood. Jpn J Surg 1 9 8 8 ; 1 8 : 4 0 3 - 4 0 8 . 37. Klaassen CD, Watkins III JB. Mechanisms of bile fo;mation, hepatic uptake, and biliary excretion. Pharmacol Rev 1 9 8 4 ; 3 6 : 1 67. 38. Kocsar LT, Bertok L, Varteresz V. Effect of bile acids on the intestinal absorption of endotoxin in rats. J Bacteriol 1969;100:220-223. Received February 27, 1995. Accepted July 28, 1995. Address requests for reprints to: Yoshihiro Mimura, M.D., Second Department of Medicine, Kurume University School of Medicine, 67 Asahi-machi, Kurume 830, Japan. The authors thank T. Yasukouchi, T. Shinozaki and Y. Ogou for expert technical assistance and Drs. H. Yoshida, K. Noguchi, S. Shakado, and M. Yoshitake for discussion of concepts in this manuscript.
Role of Hepatocytes in Direct Clearance of Lipopolysaccharide in Rats YOSHiHIRO MIMURA, SHOTARO SAKISAKA, MASARU HARADA, MICHIO SATA, and KYUICHI TANIKAWA Second Departmentof Medicine, Kurume UniversitySchool of Medicine, Kururne,Japan
Background & Aims: The liver is the clearance organ for lipopolysaccharide (LPS). The aim of this study was to investigate the biliary excretion of LPS using fluorescein isothiocyanate (FITC)-Iabeled LPS. Methods: After FITC-LPS was injected intravenously into rats, the cellular localization of fluorescence in the liver was examined and the biliary excretion of fluorescence was measured. The effects of gadolinium chloride, a blocker of Kupffer cells, and colchicine, an inhibitor of microtubules, on the biliary excretion of fluorescence was investigated, and bile was analyzed using high-performance liquid chromatography. Results: Laser scanning confocal microscopy showed that fluorescence was taken up by hepatocytes 5 minutes after injection of FITC-LPS into the portal vein. When FITC-LPS was injected into the portal vein, fluorescence was rapidly secreted into bile, peaking at 20 minutes, and 25.1% of the injected dose appeared in bile within 60 minutes. When the same dose of FITC-LPS was injected into the tail vein, 15.8% appeared in bile within 60 minutes. Chromatography showed that FITC-LPS was excreted into bile in an unchanged form over a period of 20 minutes after injection. Colchicine significantly reduced the biliary excretion of fluorescence, but gadolinium chloride had no effect. Conclusions: LPS was directly and effectively processed by hepatocytes and secreted into the bile canalicular system via a microtubule-dependent vesicular pathway.
t ndotoxin is the general term for a class of lipopolyt saccharide (LPS) molecules located in the outer cell membrane of gram-negative enteric bacteria. The presence of endotoxin in the systemic circulation can produce a variety of pathophysiological effects that are manifested during gram-negative infections, LPS is therefore considered a probable cause of septic shock. 1 The liver has been found to play a major role in clearing circulating LPS from the blood. 2-v In a previous study in which 5*Cr-labeled LPS was injected intravenously into normal mice, > 8 5 % of the dose was found to be trapped in the liver within 1 hour.: High concentrations of radioactivity were observed in the gallbladder bile of rabbits given an intravenous injection of
E
125I-labeled LPS. 8 When *4C-labeled endotoxin was injected into rats, radioactivity was excreted mainly in the feces. 4 The presence of radioactivity in gallbladder bile and feces after intravenous administration of radiolabeled LPS suggests that LPS and/or its metabolites may be processed by hepatocytes and secreted into the biliary system. Evidence suggests that LPS is internalized, modified, and released into the bloodstream by macrophages, including Kupffer cells. 2'5'<8-1. Furthermore, in the liver, a histochemical study using an anti-LPS antibody showed that LPS was initially associated with Kupffer cells and later with hepatocytes. 5 On the other hand, it has been reported that LPS binds to receptors on hepatocytes in vitro. 12'13 However, the physiological function of LPS receptor on hepatocytes has not been well defined. It has not been shown whether hepatocytes take up LPS without prior metabolism by Kupffer cells and secrete it into bile, although it is known that a peculiar structure of LPS was not metabolized and secreted into bile by hepatocytes. 14 To show the role of hepatocytes in LPS clearance, we investigated the cellular distribution and biliary excretion of fluorescein isothiocyanate (FITC)-labeled Escherichia coli O55:B5 LPS in rats. In addition, we studied the effects of gadolinium chloride (GdC13), a blocker of Kupffer cells, and colchicine, an inhibitor of microtubules, on the biliary excretion of LPS. Materials
and Methods
Chemicals FITC-labeled and -unlabeled LPS were purchased from List Biological Laboratories Inc. (Campbell, CA). LPS was prepared from E, coli (serotype O55:B5) using phenol extraction and purified chromatographically. FITC-LPS has 6.4 mol of FITC (isomer I) per mole of-LPS (technical data of FITC-LPS). Abbreviations used in this paper: FITC,fluoresceinisothiocyanate; GdCla, gadolinium chloride; HPLC, high-performanceliquid chromatography; LPS, lipopolysaccharide. © 1995 by the AmericanGastroenterologicalAssociation 0016-5085/95/$3.00
1970 MIMURAET AL.
The same lots were used for one set of experiments. The two LPS preparations were stored at 4°C and dissolved in isotonic, sterile saline before use. FITC-labeled bovine albumin (FITCalbumin), colchicine, and lumicolchicine were purchased from Sigma Chemical Co. (St. Louis, MO). FITC isomer and sodium pentobarbital were obtained from Polysciences Inc. (Warrington, PA) and Abbott Laboratories (North Chicago, IL), respectively. The acetonitrile used in the mobile phase was of high-performance liquid chromatography (HPLC) grade, and all other chemicals were of analytical-reagent grade.
GASTROENTEROLOGYVol. 109, No. 6
FITC-LPS to assess the participation of Kupffer ceils in clearance of LPS. Colchicine, an inhibitor of microtubules, and lumicolchiclue, an isomer of colchicine, were administered intraperitoneally (1 gg/g body wt) 4 hours before the administration of FITC-LPS to investigate the intercellular transport and processing kinetics of FITC-LPS in the liver. In addition, we examined the relationship between the injected dose of FITC-LPS at a concentration of 5 - 6 0 /.tg/300 g body wt and an amount of the FITC moiety secreted into bile.
Laser Scanning Confocal Microscopy HPLC of FITC-LPS The cellular distribution of FITC-LPS in the liver was examined in normal and GdC13-treated rats. FITC-LPS was injected into the portal vein at a dose of 50 btg/300 g body wt. Livers were then perfused in situ with 50 mL ofperiodatelysine-paraformaldehyde medium to remove gross blood and for fixation 5, 10, 20, or 60 minutes after injection of FITCLPS. Furthermore, pieces of the liver tissue were fixed in 4% phosphate-buffered formaldehyde for 12 hours and processed for paraffin embedding. Liver sections (5 btm) were examined using laser scanning confocal microscopy (LSM-GB200; Olympus, Tokyo, Japan) at an excitation wavelength of 488 nm and an emission wavelength of 515 nm.
Experimental Protocol Male Wistar rats weighing 2 5 0 - 3 5 0 g and bred under specific pathogen-free conditions were used in all experiments, which were performed between 10 AM and 4 PM. After rats were anesthetized with intraperitoneal injection of sodium pentobarbital (Nembutal, 39 mg/kg body wt), the common bile duct was exposed and cannulated with a polyethylene catheter, which had outer and inner diameters of 900 and 320 ~tm, respectively, and a length of 10 cm. Bile collection for experiments was started 3 0 - 4 5 minutes after bile duct cannulation when bile flow became stabilized. Using a 27-gauge needle, we injected 50 p.g/300 g body wt of FITC-LPS or FITC-albumin in 0.5-mL of saline into a branch of the portal vein or the tail vein. Control rats were given the same amount of unlabeled LPS in 0.5 mL of normal saline. Bile was collected every 10 minutes for 70 minutes after administration of FITCLPS while the abdomen was covered with wet gauze. The collected bile was centrifuged at 5000 rpm for 15 minutes at room temperature. The concentration of the FITC moiety in the supernatant was determined using an F3000 fluorescence spectrophotometer (Hitachi Ltd., Tokyo, Japan) at an excitation wavelength of 496 nm and an emission wavelength of 517 nm. In a preliminary study, the fluorescent intensity of FITC-LPS dissolved in bile was linearly related to FITC-LPS concentration ranging from 0.1 to 10 ~g/mL. Thus, the FITCLPS concentration in bile was determined by this standard curve. The FITC moiety represents the amount of unchanged LPS and/or a metabolic product carrying the FITC molecule. GdC13, a blocker of Kupffer cells, was administered intravenously (5 mg/kg body wt) 24 hours before administration of
To determine whether the LPS secreted in bile was unchanged or represented a metabolic product carrying the FITC molecule, we analyzed bile using HPLC according to the method of Martin and Ghabrial. 15 Bile obtained over a period of 20 minutes after administration of FITC-LPS was centrifuged at 5000 rpm for 15 minutes at room temperature. The supernatant was collected and then dissolved in 20 volumes of chloroform-methanol (1:2 by volume). The lower phase contents were dried, and the residue was redissolved in 200 mL of distilled water and centrifuged at 10,000 rpm for 30 minutes at 4°C. Chromatography was performed on 20 btL of the supernatant using a constant-flow high-pressure liquid chromatograph consisting of an LC-6A pump system equipped with an SCL-6A controller and an RF-530 spectrofluorometer (Shimadzu, Kyoto, Japan) at an excitation wavelength of 496 nm and an emission wavelength of 517 nm with excitation and emission slit widths of 10 nm. The column used was an Ultrahydrogel linear column (ID, 300 × 7.8 mm; Waters Associates, Milford, MA) packed with a gel of cross-linked hydroxylated polymethacrylate. The mobile phase was 100 mmol/L phosphate buffer (pH 7.75) containing 20% (vol/vol) acetonitrile at a flow rate of 0.4 mL/min.
Statistical Analysis Statistical analysis of the data was performed using Student's t test. P values of <0.05 were considered to indicate significance of differences between means.
Results Cellular Distribution of FITC-LPS in the Liver Laser scanning confocal microscopy revealed fluorescence in hepatic macrophages (Kupffer cells) and hepatocytes of liver specimens 10 minutes after 50 ~g/300 g body wt of FITC-LPS was injected into a branch of the portal vein (Figure 1A). Fluorescence intensity was stronger in Kupffer cells than in hepatocytes. Faint but significant fluorescence was detected in Kupffer cells and hepatocytes 5 minutes after injection (data not shown). W h e n FITC alone was injected, cellular fluorescence was much weaker, and its intensity was similar in Kupffer cells and hepatocytes
December 1995
HEPATOCYTIC CLEARANCE OF LIPOPOLYSACCHARIDE 1971
Biliary Excretion of FITC-LPS Neither 50 ~g/300 g body wt of FITC-LPS (Figure 2A) nor the same dose of unlabeled LPS (data not shown) affected bile flow during a 70-minute period after the portal vein administration. Excretion of the FITC moiety into bile peaked at 20 minutes, gradually decreasing during the next 50 minutes (Figure 2B). The cumulative amount of the FITC moiety excreted into bile was 25.1% +__5.7% (mean ___SD; n = l l ) of the injected dose of FITC-LPS at 60 minutes (Figure 2C).
Biliary Excretion of FITC-LPS at the Various Injected Doses When 5-60 ~g/300 g body wt of FITC-LPS was injected into the portal vein, the biliary excretion of the FITC moiety increased gradually according to the injected dose of FITC-LPS (Figure 3A). The peak concentration of the excreted FITC moiety grew in a linear fashion (Figure 3B). However, administration of >80 btg/300 g body wt of FITC-LPS reduced bile flow (data not shown).
Biliary Excretion of FITC-Albumin
Figure 1. Laser scanning confocal micrograph of fluorescence in the rat liver. (A) Ten minutes after portal vein injection of 50 pg/300 g body wt of FITC-LPS. Fluorescence was localized in hepatocytes and Kupffer cells (arrows) in the sinusoidal space. (B) Sixty minutes after portal vein injection of FITC alone. Fluorescence intensity was not different between hepatocytes and Kupffer cells when FIT(:; alone was injected. (C) Sixty minutes after portal vein injection of FITC-LPS. Fluorescence was detected in hepatocytes and nonparenchymal cells, and fluorescence intensity was reduced along the acinus of the liver. Kupffer cells (arrows) showed stronger fluorescence than hepatocytes. C, central vein; P, portal area.
(Figure 1B). Sixty minutes after injection of FITC-LPS, fluorescence was observed mainly in hepatocytes and nonparenchymal ceils close to the terminal portal venule, and fluorescence intensity was reduced in the pericentral area (Figure 1C). Fluorescence was observed within the hepatic acinus from 5 to 60 minutes after injection of FITC-LPS. Neither margination of neutrophils and mononuclear cells nor fibrinous deposits in the hepatic sinusoids and the spaces of Disse were observed within 60 minutes after injection of FITC-LPS (Figure 1C). Furthermore, the hepatic acinus was composed of normal-appearing hepatocytes (Figure 1C). The fluorescence for FITC-LPS was also distributed in splenic macrophages and pulmonary dust ceils (data not shown).
When 50 J,tg/300 g body wt of FITC-albumin was injected into the portal vein, excretion of the FITC moiety did not peak at 20 minutes (Figure 4A) and only 2% of the injected dose appeared in bile within 30 minutes (Figure 4B).
HPLC of FITC-LPS Chromatograms showed that fluorescent peaks of the preinjected FITC-LPS (Figure 5A) and the FITC moiety secreted into bile (Figure 5B) had the same retention time.
Biliary Excretion After Portal Vein and Systemic Injections of FITC-LPS Biliary excretion of the FITC moiety by the liver after intravenous injection via the tail vein was significantly lower than that after portal vein injection via the mesenteric vein (Figure 6). Sixty minutes after intravenous administration of FITC-LPS, 15.8% + 4.7% (mean ___ SD; n = 5) of the injected dose appeared in bile. Biliary excretion of the FITC moiety in rats given a systemic injection with FITC-LPS decreased to 62.9% of that found in rats given a portal vein injection (P < 0.001).
Effect of GdCla on Biliary Excretion of FITC-LPS When 50 btg/300 g body wt of FITC-LPS was injected intraportally into GdC13-treated rats, biliary ex-
1972
MIMURA ET AL.
GASTROENTEROLOGY Vol. 109, No. 6
cretion of the FITC moiety was the same as in control animals (Figure 7).
Uptake of FITC-LPS by Kupffer Cells Pretreated With GdCla The fluorescence intensity in Kupffer cells was reduced by GdCI 3 10 minutes after injection of FITC-LPS (Figure 8). GdC13 blocked endocytosis of FITC-LPS by Kupffer cells from 5 to 60 minutes after injection of FITC-LPS.
Effect of Colchicine on Biliary Excretion of FITC-LPS Sixty minutes after administration of FITC-LPS, 17.8% + 2.7% (mean + SD; n = 8) of the injected
A
250
p V10min
II[-i i I
200
lso
i
I i
O
m
I00
50
a
0
I
|
I
|
I
I
I
dose appeared in bile when 50 [.tg/300 g body wt of FITC-LPS was injected intraportally into colchicinetreated rats. Colchicine, an inhibitor of microtubules, significantly inhibited biliary excretion of the FITC moiety to 70.9% of control within 60 minutes after administration of FITC-LPS (Figure 9). On the other hand, lumicolchicine, an isomer of colchicine that does not affect microtubules, did not affect biliary excretion of the FITC moiety (data not shown). Colchicine and lumicolchicine did not affect bile flow.
Discussion The liver is the main clearance organ for LPS, and Kupffer cells have been reported to be responsible for hepatic clearance of LPS. 2'5'6's However, it is not well known whether the parenchymal liver cells (hepatocytes) are responsible for the direct clearance of LPS, whether hepatocytes take up LPS and secrete it into bile, or which transport system excretes LPS into bile from the liver. In the present study, we studied the route and rate of in vivo excretion of LPS in normal rats using FITC-LPS. The FITC molecule in FITC-LPS has been found to bind covalently to amine groups of LPS. 15 FITC-LPS was positive for the limulus amebocyte lysate assay (our unpublished observation). FITC-LPS is complexed to LPS-binding protein, and the resulting complex is recognized by CD14 receptor on monocytes in vitro) 6 FITC-LPS also
20' A
g.~ ""
B
,,g/~ 40-
t,~nl
10'
0
,~ 30
0
C 30'
% of injectedFITC-LPS
! •
•
°°, N
20,
/ °,
.//
,g (r=0.822107) °.,° • ,,,
//
/°'*
• ~
10'
/° //•
•
.,° •
1#•• /~*'°e •
0
10
20 30 40 50 60 Minutes after injection
70
80
Figure 2. Biliary excretion of FITC moiety after portal vein injection of 50 # g / 3 0 0 g body wt of FITC-LPS. (A) Effect of FITC-LPS on bile flow. (B) Excretion of FITC moiety into bile quantified as fluorescence intensity. (C) Cumulative amount of the FITC moiety excreted into bile. Results are expressed as mean _+ SD obtained from 11 rats.
0
20 4O 6O 80 Mitres afterinjection
J 10
Injected
,
,
20 dose
•
i
•
i
,
i
-
i
30 40 50 60 ( b~g/300 gm body wt)
•
)
70
Figure 3. Relationship between the injected dose of FITC-LPS and biliary excretion of the FITC moiety. (A) Representative data of biliary excretions of the FITC moiety after portal vein injection of 5 - 6 0 #g/ 300 g body wt of FITC-LPS. (B) The peak concentration of the secreted FITC moiety in A is plotted along the ordinate against the injected dose of FITC-LPS on the abscissa.
December 1995
HEPATOCYTIC CLEARANCE OF BPOPOLYSACCHARIDE
activates B cells in vivoJ 7'*s In our confocal microscopy study, fluorescence of FITC-LPS was predominantly distributed in splenic macrophages and pulmonary dust cells as well as in Kupffer cells after venous injection of FITCLPS. This cellular distribution was similar to that of LPS found in previous studies using autoradiography or immunohistochemical methods. ~'5'<8 These results suggested that FITC-LPS could be used as a tracer of LPS. Electron-microscopic autoradiography, cell isolation, and partial purification of cell types showed that 15Crlabeled LPS was taken up by Kupffer cells, endothelial cells, and hepatocytes in rats. .9 Van Bossuyt et al. showed that 3H-labeled LPS was localized in the mitochondria, cell membranes, and the periphery of hepatocytes 15 minutes after injection of the tracer. 2° In the present study, laser scanning confocal microscopy also showed hepatocytic uptake of FITC-LPS. Microscopy showed that FITC-LPS was processed by hepatocytes as well as by Kupffer cells 5 minutes after injection via the mesenteric vein. Uptake was greater in hepatocytes of acinar zones 1 and 2 than in those of acinar zone 3, probably because hepatocytes surrounding the terminal portal re-
~t
1973
nule (acinar zone 1) were exposed to a high concentration of FITC-LPS. In the present study, biliary excretion of the FITC moiety increased rapidly, peaking 20 minutes after injection of FITC-LPS, unlike FITC-albumin. Colchicine inhibited the biliary excretion of FITC-LPS. The biliary secretory curve of FITC-LPS was similar to those found in previous studies for immunoglobulin A , 21 horseradish peroxidase, 22'23 insulin, 24 inulin, 25 and copper, 26 which appeared to be transported directly to bile without passing through the lysosomal compartment of hepatocytes. Because colchicine was known to interfere with microtubules and related membrane functions, 22'25-27 these resuits suggested that FITC-LPS was transported to bile via a microtubule-dependent vesicular pathway that rapidly transverses hepatocytes from the sinusoidal membrane to the bile canaliculus. LPS has been found to be taken up in all organs by
A
i-r., CO
30 p g/ml
~ 20'
g ~
F1TC-LPS 10
F1TC-AIb 0
i
B
i
~
!
i
aj oJ
~. . . .
%of injecteddose 2O
~
FITC-LPS
!1o
B
FITC-Alb 0
0
10 20 Minutes after injection
30
.
40
Figure 4. Biliary excretion of the FtTC moiety after portal vein injection of 50 p g / 3 0 0 g body wt of FITC-albumin (Alb) (O) and FITC-LPS (0). (A) Excretion of FITC moiety into bile quantified as fluorescence intensity. (B) Cumulative amount of the FITC moiety excreted into bile. Results of FITC-albumin administration are expressed as mean ± SD obtained from 3 rats.
I 0
Retention time
I 40 rain
Figure 5, HPLC of (A) preinjected FITC-LPS and (B) FITC moiety in collected bile run on an Ultrahydrogel linear column at a flow rate of 0.4 mL/min.
1974
MIMURA ET AL.
GASTROENTEROLOGY Vol. 109, No. 6
granulocytes and macrophages, including Kupffer cells, and then LPS or its metabolites are later redistributed into hepatocytes. 28'29 However, it probably takes longer than 20 minutes for phagocytic cells to process LPS; then LPS and/or its metabolites are transferred from the blood stream to bile through hepatocytes. In the present study, we used 50 [.tg/300 g body wt of FITC-LPS. It is possible that the high dose of LPS saturates the Kupffer cell pathway, resulting in preferential processing by hepatocytes. However, fluorescence intensity in Kupffer cells after injection of 50 /.tg/300 g body wt of FITC-LPS was weaker than after an injection of 70 g/300 g body wt (our unpublished observation), suggesting that the Kupffer cell was unsaturated with LPS after injection of 50 btg/300 g body wt of FITCLPS. Furthermore, our dose-excretion data showed that the biliary efflux of FITC moiety increased with the high dose of injected FITC-LPS, representing a straight line passed through the nearby origin of the coordinates (Figure 3B). This suggested that hepatocytes may participate
in LPS clearance at the low concentration of LPS in the blood. These findings suggested that hepatocytes and Kupffer cells independently took up LPS from the blood stream. Administration of 50 /.tg/300 g body wt of FITCLPS had no effect on either the rate of bile flow or the histological alterations in various tissues. However, administration of > 8 0 ~g/300 g body wt of FITC-LPS reduced bile flow so that we could not examine the maximum capacity of hepatocytes in biliary secretion of LPS. After intravenous injection of radiolabeled LPS, the radioactivity is secreted into the biliary system. 4'8 After injection of unlabeled LPS, ~-hydroxymyristic acid, a peculiar structure of lipid A moiety in LPS, is also secreted into bile. 14 However, it is not known whether the molecules secreted into bile were intact LPS or its metabolites. Our HPLC suggested that FITC-LPS was processed and secreted into bile by hepatocytes predominantly in an unmetabolized form during a period of 20 minutes after injection. Intravenous injection of GdC13 strongly reduced liver uptake of colloidal carbon and radioactively labeled for-
0 o
A
. t* g l O m i n
i
5.0 4.0 3.0
g
2.0 1.0
0 B 30 ] %of injectedFITC-LPS
m
0.0
•
m
-
|
!
|
m
e
m
B ~8
"~
10'
i 0
10
20
30
40
50
60
70
3O
10' 0
80
m
10
|
20
i
u
i
30
40
50
!
60
m
70
80
Minutes after injection
Minutes after injection Figure 6. Biliary excretion of the FITC moiety after portal vein and systemic injections of FITC-LPS. Fifty micrograms per 300 g body wt of FITC-LPS was administered via the mesenteric vein (0) or the tail vein (t). (A) Biliary output of the FITC moiety. (B) Cumulative amount of the FITC moiety in bile. Results of systemic injections of FITC-LPS are expressed as mean -- SD obtained from 3 rats. * * P < 0.05; * * * P < 0.001.
0
Figure7.
Effect of GdCI3 on biliary excretion of the FITC moiety after portal vein injection of FITC-LPS, Fifty micrograms per 300 g body wt of FITC-LPS was administered into GdCl3-treated rats (©) and control rats (e), (A) Biliary output of the FITC moiety. (B) Cumulative amount of the FITC moiety in bile. Results in GdCls-treated rats are expressed as mean ± SD obtained from 6 rats, There was no statistical significance between the groups.
December 1995
eign erythrocytes)° It is likely that accumulation of these particles by the liver is mainly caused by phagocytosis by Kupffer cells, j° The present study also showed that GdC13 inhibited endocytosis of FITC-LPS by Kupffer cells. However, GdCI 3 did not reduce the biliary secretion of the FITC moiety by hepatocytes. These findings suggest that Kupffer cells may not participate in the early excretion of LPS into bile. If all of the FITC-LPS had bound to some plasma proteins or been processed by ceils other than hepatocytes, there would probably have been no difference in the biliary excretion of the FITC moiety between peripheral and portal vein administration. However, hepatic clearance of the FITC moiety in rats given systemic injections decreased to 62.9% of that in animals given portal vein injections. Portal vein administration of FITC-LPS would increase the concentration of LPS delivered to hepatocytes compared with systemic administration. Our results suggest that hepatocytes may take up FITC-LPS directly and excrete it into bile. Previous studies have shown that LPS administered via the portal vein is cleared more effectively than that administered by other routes, 31'32 which may explain why portal vein injection of LPS was associated with a lower mortality rate than systemic administration in rats. 33 Recently, Parent showed that rat hepatocytes had membrane receptors for LPS. 1~ However, the physiological function of hepatocyte LPS receptors is unclear. The receptors may be related to the direct clearance of LPS to bile. In this study, a significant amount of LPS was directly and effectively removed from the circulation and trans-
HEPATOCYTIC CLEARANCE OF LIPOPOLYSACCHARIDE
1975
O2
!'1 2"
8
1"
ol %of injectedFITC-LPS
T
l
30-
O
2°i
~t~~4~
~
10'
6 0' 0
10
20
30
40
50
60
70
80
Minutes after injection Figure 9. Effect of colchicine on (A) biliary output and (B) cumulation amount of the FITC moiety secreted into bile after portal vein injection of 50 btg/300 g body wt of FITC-LPS. e, Colchicine-pretreated rats; O, data from control rats as in Figure 2. Results in colchicine-pretreated rats are expressed as mean _+ SD obtained from 8 rats. * * P < 0.05; * * * P < 0.001.
ferred to bile by liver parenchymal cells through a microtubule-dependent vesicular pathway. The rest of the injected LPS might be cleared by phagocytic cells (granulocytes, monocytes, and macrophages), as described previously.2,5,6,8 Studies have shown that endotoxemia is associated with liver diseases such as obstructive jaundice,34 cholestasis, ~5 and liver cirrhosis) 6 Disturbed biliary excretion is also associated with these diseases, 3v suggesting that a disturbance in LPS excretion in addition to increased absorption of LPS from the intestine 38 may contribute to endotoxemia.
References
Figure 8. Laser scanning confocal micrograph of fluorescence in the liver pretreated with GdCI3. Ten minutes after portal vein injection of 50 btg/300 g body wt of FITC-LPS. Compare with Figure 1A. GdCl3 blocked endocytosis of FITC-LPS by Kupffer cells.
1. Nogare ARD. Southwestern Internal Medicine Conference: septic shock. Am J Med Sci 1 9 9 1 ; 3 0 2 : 5 0 - 6 5 . 2. Carey FJ, Braude AL, Zalesky M. Studies with radioactive endotoxin. III. The effect of tolerance on the distribution of radioactivity after intravenous injection of Escherichia coil endotoxin labelled with 51Cr. J Clin Invest 1 9 5 8 ; 3 7 : 4 4 1 - 4 5 7 . 3. Braude AI, Carey FJ. Zalesky M. Studies with radioactive endotoxin. II. Correlation of physiological effects with distribution of radioactivity in rabbits injected with lethal doses of E. coli endo-
1976
4.
5.
6.
7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
MIMURA ET AL.
toxin labelled with radioactive sodium chromate. J Clin Invest 1955; 3 4 : 8 5 8 - 8 6 6 . Klein B, Freudenberg MA, Galanos C. Excretion of radioactivity in faeces and urine of rats injected with 3H,14C-lipopolysaccharide. Br J Exp Pathol 1 9 8 5 ; 6 6 : 3 0 3 - 3 0 8 . Freudenberg MA, Freudenberg N, Galanos C. Time course of cellular distribution of endotoxin in liver, lungs and kidneys of rats. Br J Exp Pathol 1 9 8 2 ; 6 3 : 5 6 - 6 5 . Musson RA, Morrison DC, Uievitch RJ. Distribution of endotoxin (lipopolysaccharide) in the tissues of lipopoiysaccharide-responsive and unresponsive mice. Infect Immun 1 9 7 8 ; 2 1 : 4 4 8 - 4 5 7 . Zlydaszyk JC, Moon RJ. Fate of 5~Cr-labeled lipopolysaccharide in tissue culture cells and livers of normal mice. Infect Immun 1976; 1 4 : 1 0 0 - 1 0 5 . Mathison JC, Ulevitch RJ. The clearance, tissue distribution and cellular localization of intravenously injected lipopolysacchadde in rabbits. J Immunol 1979; 1 2 3 : 2 1 3 3 - 2 1 4 3 . Peterson AA, Munford RS. Dephosphorylation of the lipid A moiety of Escherichia coil lipopolysaccharide by mouse macrophages. Infect Immun 1 9 8 7 ; 5 5 : 9 7 4 - 9 7 8 . Munford RS, Hall CL. Uptake and deacylation of bacterial lipopolysaccharides by macrophages from normal and endotoxin-hyperresponsive mice. Infect Immun 1 9 8 5 ; 4 8 : 4 6 4 - 4 7 3 . Fox ES, Thomas P, Broitman SA. Clearance of gut-derived endotoxins by the liver. Gastroenterology 1 9 8 9 ; 9 6 : 4 5 6 - 4 6 1 . Pagani R, Portoles MT, Municio AM. The binding of Escherichia coil endotoxin to isolated rat hepatocytes. FEBS Lett 1981; 1 3 1 : 1 0 3 - 1 0 7 . Parent JB. Membrane receptors on rat hepatocytes for the inner core region of bacterial lipopolysaccharide. J Biol Chem 1990; 2 6 5 : 3 4 5 5 - 3 4 6 1 . Maitra SK, Rachmilewitz D, Eberle D, Kaplowitz N. The hepatocellular uptake and biliary excretion of endotoxin in the rat. Hepatoiogy 1981; 1 : 4 0 1 - 4 0 7 . Martin G, Ghabrial H. High-performance liquid chromatographic method to resolve and determine lipopolysaccharide sub-groups of Escherichia coil endotoxin in isolated perfused rat liver perfusate. J Chromatogr 1 9 9 2 ; 5 7 4 : 2 0 5 - 2 1 1 . Gallay P, Jongeneel CV, Barras C, Burnier M, Baumgartner JD, Glauser MP, Heumann D. Short time exposure to lipopolysacchadde is sufficient to activate human monocytes. J Immunol 1993; 1 5 0 : 5 0 8 6 - 5 0 9 3 . Moeller G, Gronowicz E, Persson U, Coutinho A, Moiler E, Hammarstrom L, Smith E. Spleen cells from animals tolerant to a thymus-dependent antigen can be activated by lipopolysaccharide to synthesize antibodies against the tolerogen. J Exp Med 1976; 1 4 3 : 1 4 2 9 - 1 4 3 8 . Skelly RR, Munkenbeck P, Morrison DC. Stimulation of T-independent antibody responses by hapten-lipopolysaccharides without repeating polymeric structure. Infect Immun 1979;23: 287-293. Ruiter DJ, Van Der Meulen J, Brouwer A, Hummel M JR, Mauw BJ, Van Der Ploeg JCM, Wisse E. Uptake by liver cells of endotoxin following its intravenous injection. Lab Invest 1 9 8 1 ; 4 5 : 3 8 - 4 5 . Van Bossuyt H, De Zanger RB, Wisse E. Cellular and subcellular distribution of injected lipopolysaccharide in rat liver and its inactivation by bile salts. J Hepatol 1 9 8 8 ; 7 : 3 2 5 - 3 3 7 . Renston RH, Jones AL, Christiansen WD, Hradek GT, Underdrown BJ. Evidence for a vesicular transport mechanism in hepatocytes for biliary secretion of immunoglobulin A. Science 1980; 208:1276-1278.
GASTROENTEROLOGY Vol. 109, No. 6
22. Sakisaka S, Ishii Y, Ueda H, Matsumoto H, Eguchi T, Tanikawa K. The mechanism of intracellular transport of horseradish peroxidase and the role of micro~ubules in the hepatocyte. J Clin Electron Microsc 1 9 8 2 ; 1 5 : 5 - 6 . 23. Gondo K, Sakisaka S, Abe H, Tanikawa K. Observation of 1251insulin in hepatocytes by light and electron microscope autoradiography. J Clin Electron Microsc 1 9 8 5 ; 3 4 : 6 6 - 6 7 . 24. Lorenzini 1, Sakisaka S, Meier PJ, Boyer JL. Demonstration of a transcellular vesicle pathway for biliary excretion of inulin in rat liver. Gastroenterology 1 9 8 6 ; 9 1 : 1 2 7 8 - 1 2 8 8 . 25. Harada M, Sakisaka S, Yoshitake M, Shakadoh S, Gondo K, Sata M, Tanikawa K. Biliary copper excretion in acutely and chronically copper-loaded rats. Hepatology 1 9 9 3 ; 1 7 : 1 1 1 - 1 1 7 . 26. Sakisaka S, Ng OC, Boyer JL. Tubulovesicular transcytotic pathway in isolated rat hepatocyte couplets in culture.Effect of colchicine and taurocholate. Gastroenterology 1 9 8 8 ; 9 5 : 7 9 3 - 8 0 4 . 27. Manfredi JJ, Horwitz SB. An antimitotic agent with a new mechanism of action. Pharmacol Ther 1 9 8 4 ; 2 5 : 8 3 - 1 2 5 . 28. Freudenberg MA, Kleine B, Galanos C. The fate of lipopolysaccharide in rats: evidence for chemical alteration in the molecule. Rev Infect Dis 1 9 8 4 ; 6 : 4 8 3 - 4 8 7 . 29. Freudenberg MA, Galanos C. Bacterial lipopoiysaccharides: structure, metabolism and mechanisms of action. Int Rev Immunol 1990;6:207-221. 30. Husztik E, Lazar G, Parducz A. Electron microscopic study of Kupffer-cell phagocytosis blockade by gadolinium chloride. Br J Exp Pathol 1 9 8 0 ; 6 1 : 6 2 4 - 6 3 0 . 31. Jacob AI, Goldberg PK, Bloom N, Degenshein GA, Kozinn PJ. Endotoxin and bacteria in portal blood. Gastroenterology 1977; 7 2 : 1 2 6 8 - 1 2 7 0 . 32. Wolter J, Liehr H, Grun M. Hepatic clearance of endotoxins: differences in arterial and portal venous infusion. J Reticuloendothel Soc 1 9 7 8 ; 2 3 : 1 4 5 - 1 5 2 . 33. Mori K, Matsumoto K, Gans H. On the in vivo clearance and detoxification of endotoxin by lung and liver. Ann Surg 1973;177:159-163. 34. Ddvas G, James O, Wardle N. Study of reticuloendothelial phagocytic capacity in patients with cholestasis. BMJ 1 9 7 6 ; 1 : 1 5 6 8 1569. 35. Greve JW, Gouma DJ, Buurman WA. Complications in obstructive jaundice: role of endotoxins. Scand J Gastroenterology Suppl 1992; 1 9 4 : 8 - 1 2 . 36. Tachiyama G, Sakou M, Kambayashi J, lijima S, Tsujinaka T, Mori T. Endogenous endotoxemia in patients with liver cirrhosis--a quantitative analysis of endotoxin in portal and peripheral blood. Jpn J Surg 1 9 8 8 ; 1 8 : 4 0 3 - 4 0 8 . 37. Klaassen CD, Watkins III JB. Mechanisms of bile fo;mation, hepatic uptake, and biliary excretion. Pharmacol Rev 1 9 8 4 ; 3 6 : 1 67. 38. Kocsar LT, Bertok L, Varteresz V. Effect of bile acids on the intestinal absorption of endotoxin in rats. J Bacteriol 1969;100:220-223. Received February 27, 1995. Accepted July 28, 1995. Address requests for reprints to: Yoshihiro Mimura, M.D., Second Department of Medicine, Kurume University School of Medicine, 67 Asahi-machi, Kurume 830, Japan. The authors thank T. Yasukouchi, T. Shinozaki and Y. Ogou for expert technical assistance and Drs. H. Yoshida, K. Noguchi, S. Shakado, and M. Yoshitake for discussion of concepts in this manuscript.