Microvascular consequences of Kupffer cell modulation in rat liver fibrogenesis

J. Pathol. 189: 85–91 (1999)

MICROVASCULAR CONSEQUENCES OF KUPFFERCELL MODULATION IN RAT LIVER FIBROGENESIS

*, , .

Institute for Clinical and Experimental Surgery, University of Saarland, Homburg/Saar, Germany

SUMMARY

The process of Ito cell activation, which is thought to be the central pathogenic mechanism in liver fibrogenesis, may involve distinctinteractions with Kupffer cells (KCs) mediated by various cytokines and growth factors. The aim of this study was to determine whethertargeting KC function using gadolinium chloride (GdCl3) interferes with the manifestation of carbon tetrachloride (CCl4)-inducedhepatic fibrosis, placing special emphasis on the process of microvascular remodelling. Using in vivo fluorescence microscopy,characteristic microvascular features of CCl4-induced liver fibrosis, progressively observed within the 8-week period of toxin exposure,were the significant reduction in sinusoidal density; the increase of venular vascular space; the perivenular accumulation of Ito cells, withconcomitant collagen deposition; and the collapse of parenchymal tissue. GdCl3 effectively attenuated sinusoidal rarefaction and delayed,but did not prevent, the process of Ito cell activation-associated collagen deposition. Strikingly, the 8-week modulation of KC functionby GdCl3 exhibited sustained hepatocellular fatty vacuolation with organ weight increase, liver enzyme release, and bile flow reduction.Thus, GdCl3 treatment attenuates the hepatic microvascular response, but favours fatty change and only delays the development of liverfibrosis following CCl4-exposure. Copyright ? 1999 John Wiley & Sons, Ltd.

KEY WORDS—gadolinium chloride; intravital fluorescence microscopy; Kupffer cells; Ito cells; hepatic microcirculation; fibrosis; cirrhosis

INTRODUCTION

Recent integrative studies of the cellular and molecu-lar mechanisms of liver fibrogenesis have proposed athree-step cascade mechanism of Ito cell activation,including a sequential cross-talk between parenchymaland non-parenchymal cells.1 Within this puzzle of inter-acting cells, fibrogenic cytokines, regulatory loops, andextracellular matrix molecules,2 an initial inflammatoryresponse to toxin-induced hepatocyte necrosis is thoughtto include expansion3 and activation of resident Kupffercells (KCs) with the release of various peptide growthfactors.4 These mediators in turn stimulate proliferationwithin the Ito cell population and their subsequentactivation and modulation towards ‘myofibroblast-like’cells, with resultant increased synthesis of matrix pro-teins.5 In previous in vivo fluorescence microscopicstudies in chronically CCl4-exposed rats,6,7 we havedemonstrated specific progressive changes of microvas-cular morphology with disordered acinar architecture,non-parenchymal cell transformation with collagendeposition, and parenchymal cell loss, as well as rarefac-tion and capillarization of sinusoids. Since transforma-tion of Ito cells with deposition of collagen and otherextracellular matrix proteins is considered a key compo-nent of the pathogenesis of liver fibrosis,8 and KCactivation with the release of fibrogenic mediators isknown to precede Ito cell activation,9,10 it is reasonableto postulate that modulation of KCs could disrupt thesequence of events leading to organ injury by dampingdown the fibrogenic stimulus.

CCC 0022–3417/99/100085–07$17.50Copyright ? 1999 John Wiley & Sons, Ltd.

To define the significance of KCs in the developmentof liver fibrosis/cirrhosis, we therefore treated rats withGdCl3 and assessed the nature and magnitude of CCl4-induced fibrogenesis of those animals in comparisonwith saline-treated controls.

MATERIALS AND METHODS

*Correspondence to: Brigitte Vollmar, MD, Institute for Clinicaland Experimental Surgery, University of Saarland, 66421 Homburg/Saar, Germany. E-mail: [email protected]

Contract/grant sponsor: Wilhelm-Sander Stiftung; Contract/grantnumber: 93.019.2.

Animal model

In accordance with German legislation on the protec-tion of animals and the Guide for the Care and Use ofLaboratory Animals (NIH publication No. 86-23,revised 1985), male Sprague Dawley rats (Charles River,Fa. Wiga, Sulzfeld, Germany) with a starting weight of2170 g were injected with GdCl3 (Sigma, Deisenhofen,Germany) through the penile vein (10 mg/kg) on day"2 and "1.11 On day 0, animals received phenobarbi-tal sodium (35 mg/dl) with their drinking water. Threedays later, the rats were subcutaneously injected with0·15 ml of CCl4/100 g body weight (Merck, Darmstadt,Germany) in an equal volume of olive oil (Merck) twicea week. CCl4 treatment was administered for 1 (n=5)and 8 (n=5) weeks. These distinct time points werechosen, since we have demonstrated in a previous studythat CCl4 exposure for 1 week already caused notablemicrovascular and cellular changes, which fully devel-oped within 4–8 weeks of CCl4 treatment.7 During thetime period of CCl4 exposure, GdCl3 (10 mg/kg) wassimultaneously administered intravenously twice a week.A minimal delay of 3 days after the last dose of CCl4 wasallowed before microcirculation studies were under-taken. Additional animals received phenobarbital andCCl4 for either 1 (n=5) or 8 weeks (n=5), but instead ofGdCl3, equal volumes of saline intravenously. A totalof eight additional animals were given either olive oil or

Received 9 June 1998Revised 18 February 1999

Accepted 14 April 1999

86 B. VOLLMAR ET AL.

GdCl3 alone, for 1 and 8 weeks. Since these animals didnot demonstrate significant differences in any of theparameters analysed, they were combined as a non-cirrhotic control group. All animals were kept on astandard light cycle and were fed ad libitum with a stockpellet dish.

In vivo studies

Under pentobarbital anaesthesia (50 mg/kg bodyweight i.p.), the animals were tracheotomized and theright carotid artery and jugular vein were cannulated(PE-50; ID 0·58 mm; Portex, Hythe, U.K.) for haemo-dynamic monitoring and injection of fluorescent dyes.After transverse laparotomy and cannulation of the bileduct (PE-50) for continuous collection of bile (given inìl/min#g liver weight), the rats were positioned on theirleft side and the livers were prepared for intravitalfluorescence microscopy, as previously described indetail.6,7,11 After the microscopic procedure, arterialblood samples were taken for spectrophotometric deter-mination of serum activities of aspartate aminotrans-ferase (AST) and alanine aminotransferase (ALT). Totallivers were excised for determination of wet weight andliver tissue samples were obtained for routine histo-pathology and immunohistochemistry, with subsequentquantitative analysis of tissue specimens.

relative to the whole area of observation.
Intravital fluorescence microscopy and microcirculatoryanalysis
In vivo microscopy was performed using a modifiedZeiss Axio-Tech microscope (Zeiss, Jena, Germany) andepi-illumination technique with a 100 W mercury lamp,as described previously.6,7 Microscopic images wereregistered by a charge-coupled device video camera (FK6990; Prospective Measurements Inc., San Diego,CA, U.S.A.) and were transferred to a video system(VO-5800 PS; Sony, Munich, Germany). With the useof different objectives (10#/0·30, W 20#/0·5; Zeiss)magnifications of #350 and #700 were achievedon the video screen (PVM-1442 QM, diagonal:330 mm, Sony). By means of different Zeiss filter sets(excitation/emission wavelengths: 330–390/>430 nm,450–490 nm/>520 nm, and 530–560/>580 nm), micro-fluorographic images of autofluorescence of the hepaticsurface [10–15 observation areas (560#440 ìm) peranimal] were taken. Ultraviolet epi-illumination allowedfor observation of multiple patchy fluorescent activitiesderived from vitamin A in Ito cells and completelyeliminated vitamin A autofluorescence within 220 s ofexposure, due to its rapid photobleaching property.12

Emission maxima of collagen above 435 nm13 allowedfor the detection of these extracellular matrix com-posites by switching to blue and green fluorescenceepi-illumination, respectively. Subsequently, contrastenhancement for visualization of the hepatic microvas-culature was achieved by intravenous injection of 5 percent fluorescein isothiocyanate (FITC)-labelled dextran(MW 150 000; 0·1 ml/100 g body weight; Sigma) andepi-illumination with blue light. For staining of paren-chymal tissue, the fluorescent dye bisbenzamide H33342

Copyright ? 1999 John Wiley & Sons, Ltd.

(2 ìmol/100 g body weight; Sigma) and ultravioletepi-illumination were used.14

As described previously,7 quantitative analysis wasperformed off-line by a computer-assisted image analysissystem (CapImage; Zeintl, Heidelberg, Germany). Spa-tial distribution of sites of Ito cell-associated vitamin Aautofluorescence was assessed by densitometric record-ing of positive sites of fluorescence per single frame(75#75 ìm) within the individual zones 1, 2, and 3 ofliver acini. The area of positive vitamin A sites wascalculated automatically as a percentage of the wholearea of the single frame. For the assessment of collagenautofluorescence, grey levels were also determined den-sitometrically, with subsequent automatic calculation ofthe area of positive deposits as a percentage of the wholearea of observation (560#440 ìm). The post-sinusoidal/hepatic venular lumen was quantitated planimetrically(area measurement, given as a percentage of thecomplete area under investigation) in the ultravioletepi-illuminated microfluorographic images using alsoCapImage.

After contrast enhancement by FITC-dextran, thesinusoidal density was determined by counting thenumber of blood cell-perfused sinusoids crossing a200 ìm raster line.7 For the assessment of hepatocellularbisbenzamide fluorescence intensity, grey levels weredetermined densitometrically; the area of positivelystained cells was calculated automatically and is given

7,14

Histopathology and immunohistochemistry

Samples of liver tissue were fixed in 4 per centphosphate-buffered formalin for 2–3 days and embed-ded in paraffin. Sections (5 ìm) were cut and mountedon poly--lysine slides for immunohistochemical studiesand for haematoxylin and eosin for routine histologyand a trichrome stain for assessment of collagendeposition (Ladewig staining). Quantification of areaswith hepatocellular fatty vacuolation and of fibroussepta was performed planimetrically (given as a percent-age of the whole observation area) in 10–15 observationareas of interest per tissue section and animal usingCapImage.

For immunohistochemical detection of macrophages,deparaffinized sections were incubated with H2O2 toblock endogenous peroxidase and subsequently pre-treated with pronase (2·5 mg/ml) for 10 min at37)C. Mouse monoclonal anti-ED2 (Biotrend, Koln,Germany) diluted 1:200 was used as the primary anti-body, incubated for 90 min at 37)C. A biotinylated horseanti-mouse antibody (Vector Laboratories, Burlingame,CA, U.S.A.) diluted 1:200 was used as the secondaryantibody for streptavidin–biotin–complex peroxidasestaining (Vectastain ABC-peroxidase kits; Camona,Wiesbaden, Germany). The sections were treated with3,3*-diaminobenzidine and counterstained with haema-toxylin. For quantitative analysis, the numbers of ED2-positive KCs, which were identified by positive stainingand by their morphology, were counted in a total of 50high-power fields (HPFs) per tissue section under alight microscope (Model BX60F; Olympus Optical Co.,

J. Pathol. 189: 85–91 (1999)

87KUPFFER CELL MODULATION IN LIVER CIRRHOSIS

Tokyo, Japan) at 400# magnification and are given ascells/HPF.

Statistics

All values are expressed as means&SEM. After prov-ing the assumption of normality and homogeneity ofvariance across groups, differences between groups weretested by analysis of variance (ANOVA) followed by theappropriate post hoc comparison test (SigmaStat, JandelCorporation, San Rafael, CA, U.S.A.). The criterion forsignificance was taken to be p<0·05.


Table I—Effect of GdCl3 treatment on the area with hepato-cellular fatty vacuolation (fatty change) and fibrous septa (as apercentage of the observation area) as well as on the number ofED2-positive cells (n/HPF) in liver tissue of saline- and GdCl3-treated animals exposed for 1 and 8 weeks to CCl4

Fattychange

Fibroussepta

ED2-positiveKCs

Control 0 0 10·0&1·21 week CCl4/saline 26·9&8·1* 3·5&0·9* 10·0&0·41 week CCl4/GdCl3 32·2&6·9* 1·7&1·1* 8·1&1·58 weeks CCl4/saline 21·0&8·2* 10·0&0·4* 8·2&0·48 weeks CCl4/GdCl3 73·7&11·3*† 7·2&1·7* 8·6&1·4

Means&SEM; CCl4 exposure for either 1 (n=5) or 8 (n=5) weeksand CCl4 exposure for either 1 (n=5) or 8 (n=5) weeks with simul-taneous GdCl3 treatment each. Control=non-CCl4-treated controlanimals (n=8).

*p<0·05 vs. control;†p<0·05 vs. 8 weeks CCl4/saline.

RESULTS

Animals of the GdCl3- or saline-treated groupsexposed to CCl4 for 1 and 8 weeks did not differ in termsof body weight (1 week: 233&10 and 234&8 g; 8 weeks:345&24 and 359&18 g). The mean body weight ofthe control animals was 258&17 g. Calculation of liverto body weight ratios indicated comparable organhypertrophy after 1 week [4·4&0·1 and 4·6&0·2 percent, p<0·05 vs. controls: 3·1&0·2 per cent (olive oil:3·3&0·2 per cent; GdCl3: 3·1&0·5 per cent)]. After 8weeks of CCl4 exposure, GdCl3-treated animals exhib-ited significantly (p<0·05) higher values of liver to bodyweight ratios (5·7&0·3 per cent) than those ofKC-unmanipulated animals (4·5&0·5 per cent). Alongwith this pronounced organ hypertrophy, markedlyincreased serum AST (453&188 U/l) and ALT(492&201 U/l) activities as well as reduced bile flow(0·9&0·1 ìl/min#g) were found in GdCl3-treated ani-mals at 8 weeks of CCl4 exposure, whereas hepatocellu-lar disintegration (AST: 148&17 U/l; ALT: 81&20 U/l;p<0·05) and excretory dysfunction (1·3&0·1 ìl/min#g;p>0·05) were less pronounced in animals solely exposedto CCl4 for 8 weeks. Short-term exposure to CCl4 for1 week did not cause significantly different serum liverenzyme activities or bile flow values, either between bothCCl4-exposed groups (AST and ALT <60 U/l; bile flow2·2–2·4 ìl/min#g) or between those groups and thecontrol groups (olive oil: AST and ALT <50U/L; bileflow 2·5&0·2 ìl/min#g; GdCl3: AST and ALT <50 U/l;bile flow 2·3&0·3 ìl/min#g).

Quantitative analysis of liver tissue specimens in termsof hepatocellular fatty vacuolation, fibrous septa, andnumber of ED2-positive KCs is given in Table I. Fattychange of parenchymal tissue and the formation offibrotic septa were found in CCl4-exposed animals butnot in controls. While in 1 week-treated animals, fattyaccumulation with cytoplasmic vacuolation of mostlypericentrally located hepatocytes dominated the rarelyobserved delicate fibrous septa, rats treated for 8 weeksshowed more dense fibrous septa, dividing the hepaticparenchyma into multiple discrete nodules (Table I andFig. 1). Fibrotic septa were only slightly reduced byGdCl3 treatment, but fatty change was found to be evenmarkedly (p<0·05) increased (Table I and Fig. 1).Immunohistochemistry of either 1- or 8-week CCl4-exposed livers did not reveal significant changes in terms

of numbers of ED2-positive KCs, ranging between 8 and10 cells/HPF. Moreover, GdCl3 treatment, either alone(9·9&1·4 vs. olive oil: 10·0&1·4 cells/HPF) or in com-bination with CCl4 (Table I), was not associated with apronounced depletion in the number of KCs.

In contrast to the homogeneous distribution of sites ofvitamin A autofluorescence in control livers, CCl4-exposed livers exhibited a change of spatial distributionof vitamin A-associated autofluorescent sites, with areduction in zones 1 and 2 in favour of maximalaccumulation in zone 3 (Fig. 2A). In GdCl3-treatedanimals, local redistribution was not found to be sopronounced within the first week of CCl4 exposure(p=0·06 vs. saline-treated animals), but was fully evidentafter the 8-week period (Fig. 2A). Overall, sites ofvitamin A-associated autofluorescence decreased in bothgroups at 8 weeks of CCl4 exposure (Fig. 2B). Theassessment of yellow-green and red autofluorescence,both of which are consistent with collagenous deposits,revealed a moderate accumulation after CCl4 exposurefor 1 week, with a three- to four-fold increase at 8 weeksof CCl4 exposure (Fig. 3). In the GdCl3-treated animals,collagen deposition was slightly, but not significantlyreduced in comparison with the KC-unmanipulatedanimals, which is in line with the planimetricallyassessed histological data (Table I). In both groups,collagen deposition strongly coincided with sites ofaccumulation of vitamin A autofluorescence. Increaseof venular vascular space within 1–8 weeks of CCl4exposure did not differ between animals of the twoexperimental groups (Fig. 4).

Liver fibrogenesis was associated with a pronouncedreduction in sinusoidal density to 3·9&0·3 and 0·6&0·1sinusoids per 200 ìm (1 and 8 weeks of CCl4 exposure),compared with values in control livers (GdCl3: 5·3&0·1and olive oil: 5·2&0·1 sinusoids per 200 ìm; Fig. 5). Incontrast, nutritive perfusion was found to be less(p<0·05) impaired in the GdCl3-treated animals afterboth 1 and 8 weeks of CCl4 exposure, with sinusoidaldensities of 4·9&0·1 and 2·5&0·4 sinusoids per 200 ìm,respectively (Fig. 5). Parenchymal tissue staining using

J. Pathol. 189: 85–91 (1999)


Fig. 1—Histological sections of liver tissue (A and B, haematoxylin and eosin staining; C and D, Ladewig staining) following CCl4exposure for 8 weeks with either GdCl3 (A, C) or saline treatment (B, D). Note the markedly more pronounced fatty vacuolation ofhepatocytes (A), but fewer fibrotic septa (C) in the cirrhotic animals after GdCl3 treatment, compared with the saline-treatedcirrhotic controls (B and D). (A, B) #185; (C, D) #92·5—reduced to 65 per cent in printing

Fig. 2—Sites of vitamin A-associated autofluorescence in zones 1, 2, and 3 (A) as well as in total acinar regions (B) in livers of animals following CCl4exposure for 1 (n=5) and 8 (n=5) weeks with either GdCl3 (open bars) or saline (closed bars) treatment. Non-CCl4-treated animals served as controls(C, n=8). Note the non-homogeneous distribution of sites of vitamin A autofluorescence, with accumulation in zone 3 at the expense of zones 1 and2 in the CCl4-exposed animals, in contrast to their uniform distribution in controls. GdCl3 treatment (1 week) delayed the kinetics of pericentralaccumulation of sites of vitamin A autofluorescence. Means&SEM; *p<0·05 vs. control

Copyright ? 1999 John Wiley & Sons, Ltd. J. Pathol. 189: 85–91 (1999)


bisbenzamide revealed a comparable reduction betweenboth groups by approximately 15 and 44 per cent after 1and 8 weeks of CCl4 exposure, respectively (Fig. 6).

Fig. 3—Area of autofluorescent collagen (percentage) in livers ofanimals following CCl4 exposure for 1 (n=5) and 8 (n=5) weeks(closed bars) and following CCl4 exposure for 1 (n=5) and 8 (n=5)weeks with simultaneous GdCl3 treatment (open bars). C=non-cirrhotic control animals (n=8). CCl4-induced deposition of autofluo-rescent collagen did not significantly differ between groups, althoughthere was a tendency towards reduced values in GdCl3-treated cir-rhotic animals. No collagen deposition could be detected in non-cirrhotic control animals. Means&SEM; *p<0·05 vs. control

3


Fig. 4—Venular vascular space (mm2) in livers of animals followingCCl4 exposure for 1 (n=5) and 8 (n=5) weeks (closed bars) andfollowing CCl4 exposure for 1 (n=5) and 8 (n=5) weeks with simul-taneous GdCl3 treatment (open bars). C=non-cirrhotic control ani-mals (n=8). Increases of venular vascular space within 8 weeks of CCl4exposure, which reflect dilation of post-sinusoidal venules and theirtransformation into shunts, did not differ between the GdCl3-treatedand the non-GdCl3-treated group. Means&SEM; *p<0·05 vs. control

DISCUSSION

The effects of GdCl3 on KCs are reported to bemanifold in nature and more complex than merely apossible change in cell number, including the inductionof apoptosis;15 the reduction in nitric oxide synthaseexpression;16 the decrease of phagocytic activity,11

probably by interfering with calcium-dependent cellsurface interactions;17,18 the attenuation of endotoxin-induced TNF-alpha mRNA production and proteinrelease;11,19,20 the impairment of their capacity to gener-ate superoxide;21 the abolition of expression of certaincell-specific antigens;22 and the switch of phenotypicshape with loss of IL-10 expression.23 This broad spec-trum of action might be the reason for an observeddiscrepancy: GdCl3 treatment was demonstrated to leadboth to impairment of host defence, predisposing tohigher mortality,24 and to attenuation of post-ischaemicand endotoxaemic liver injury, with increased sur-vival.11,19,20,25,26 Given these results, KCs have thepotential to act as a double-edged sword in liver disease,making their modulation at least not an indisputabletherapeutic rationale. Accordingly, the major finding ofour study is that GdCl treatment does moderately

attenuate the CCl4-induced hepatic fibrogeneticresponse, but favours ongoing hepatocellular fattyvacuolation.

To our surprise, GdCl3-treated animals revealed a stillhigher density of perfused sinusoids after 8 weeks ofCCl4 exposure, despite the obvious organ injury anddysfunction. The evidence that GdCl3 pretreatmentdecreases portal resistance and increases perfusion aftercold ischaemia/reperfusion injury,27 as well as amelio-rates portal pressure of fat-loaded livers from ethanol-treated rats using a low-flow reflow hypoxia reperfusionmodel of liver injury,28 is consistent with the notion thatamongst many other factors, disturbance of the hepaticmicrocirculation can be attributed to swollen KCs,protruding into the sinusoidal lumina.29 Given theseresults, GdCl3 might preserve sinusoidal density in CCl4-exposed livers by the reduced liberation of vasoconstric-tive mediators upon KC activation,30 as postulated byothers.27 Moreover, prevention of sinusoidal endothelialcell damage with preservation of sinusoidal architecture,as reported for GdCl3 in experimental settings of ratliver cold ischaemia/reperfusion31 and endotoxaemia,32

may also account for the distinct maintenance ofsinusoidal perfusion in the present study. In contrast,hepatocellular disintegration and excretory dysfunctionseem to be mediated by KC-independent mechanisms;similar mechanisms are postulated for the failure ofGdCl3 treatment to prevent parenchymal cell membranedamage and dysfunction in post-ischaemic livers, but forthe effectiveness of GdCl3 with regard to the hepaticmicrocirculation.27

J. Pathol. 189: 85–91 (1999)


In line with our previously demonstrated results indeveloping cirrhosis in rats,7 CCl4 exposure is charac-terized by an initial change of the spatial distribution ofIto cells, with coincidence of autofluorescent fibroticmaterial around dilated post-sinusoidal venules. More-over, within the first weeks of CCl4 exposure, hepato-cellular fatty change is typically prominent, whilefibrogenetic features predominate in long-term CCl4-exposed livers.7 However, GdCl3-treated animals exhib-ited extensive fatty change after 8 weeks of CCl4exposure, with less pronounced fibrogenetic organ alter-ations. Primary maintenance of homogeneity of acinarIto cell distribution in the GdCl3-treated animals mightmirror the initial inhibition of the CCl4-induced KC-dependent Ito cell proliferation and migration towardszone 3, with collagenous matrix deposition, whichindeed was still found slightly reduced after 8 weekscompared with KC-unmanipulated animals. However,in general, we failed to demonstrate long-term protectionfrom CCl4-induced liver injury, using GdCl3. This seemslargely due to the fact that GdCl3 may transiently reducethe number of large KCs, but is immediately followed bya repopulation of monocytic cells, which are less vulner-able to GdCl3.22 In addition to the influx of bloodmonocytes, local proliferation of KCs might account forthe numbers of phagocytic cells remaining unchanged,as in the present study, or even increasing, as observedby others.3 Besides decreased sensitivity to GdCl3,these invading cells exhibit an incomplete phagocytic


capacity.22 This may aggravate endocytic dysfunction ofthe reticuloendothelial system, which itself is known toexist in experimental and human liver cirrhosis,33–36

thereby causing an increase in susceptibility to toxicagents. Thus, altered responsiveness of KCs can be thereason for the observation of sustained hepatocellularfatty change, with increased tissue weight, enhancedenzyme release, and excretory dysfunction in GdCl3-treated fibrotic rats. At the same time, unaltered releaseof tissue growth factor-beta 1 by KCs, despite GdCl3treatment,23 nevertheless creates the micro-enviromentprimarily necessary for the initiation and perpetuation ofIto cell activation, and, might thus account for the fibro-sis seen in the GdCl3-treated animals. In the isolatedperfused liver, GdCl3 was reported to be effective interms of inhibition of initial injury upon a single CCl4challenge.37 Despite this, we conclude from our presentresults that GdCl3 does not prove effective by its long-term modulation of KC functions in the experimentalsetting of developing cirrhosis in rats.

ACKNOWLEDGEMENTS

This study was supported by a grant from theWilhelm-Sander Stiftung (No. 93.019.2). BV is sup-ported by a Heisenberg-Stipendium of the DeutscheForschungsgemeinschaft (Vo 450/6-1).

Fig. 5—Sinusoidal density (perfused sinusoids per 200 ìm) in livers ofanimals following CCl4 exposure for 1 (n=5) and 8 (n=5) weeks(closed bars) and following CCl4 exposure for 1 (n=5) and 8 (n=5)weeks with simultaneous GdCl3 treatment (open bars). C=non-cirrhotic control animals (n=8). Note the pronounced sinusoidalrarefaction within 1–8 weeks of CCl4 exposure, whereas GdCl3 treat-ment significantly inhibited the reduction in sinusoidal density.Means&SEM; *p<0·05 vs. control; # p<0·05 vs. KC-unmanipulatedlivers (closed bars)

Fig. 6—Parenchymal (hepatocellular) tissue (percentage) in livers ofanimals following CCl4 exposure for 1 (n=5) and 8 (n=5) weeks(closed bars) and following CCl4 exposure for 1 (n=5) and 8 (n=5)weeks with simultaneous GdCl3 treatment (open bars). C=non-cirrhotic control animals (n=8). Cirrhosis-associated collapse ofparenchymal tissue within 1–8 weeks of CCl4 exposure was compar-able in extent in the two groups. Means&SEM; *p<0·05 vs. control

REFERENCES1. Gressner AM, Bachem MG. Molecular mechanisms of liver fibrogenesis—a

homage to the role of activated fat-storing cells. Digestion 1995; 56:335–346.

J. Pathol. 189: 85–91 (1999)


2. Gressner AM. Hepatic fibrogenesis: the puzzle of interacting cells, fibro-genic cytokines, regulatory loops, and extracellular matrix molecules.Z Gastroenterol 1992; 30: 5–16.

3. Geerts A, Schellinck P, Bouwens L, Wisse E. Cell population kinetics ofKupffer cells during the onset of fibrosis in rat liver by chronic carbontetrachloride administration. J Hepatol 1988; 6: 50–56.

4. Matsuoka M, Tsukamoto H. Stimulation of hepatic lipocyte collagenproduction by Kupffer cell-derived transforming growth factor â: implica-tion for a pathogenic role in alcoholic liver fibrogenesis. Hepatology 1990;11: 599–605.

5. Tanaka Y, Nouchi T, Yamane M, et al. Phenotypic modulation of lipocytesin experimental liver fibrosis. J Pathol 1991; 164: 273–278.

6. Vollmar B, Wolf B, Siegmund S, Katsen AD, Menger MD. Lymph vesselexpansion and function in the development of hepatic fibrosis and cirrhosis.Am J Pathol 1997; 151: 169–175.

7. Vollmar B, Siegmund S, Menger MD. An intravital fluorescence micro-scopic study of hepatic microvascular and cellular derangements in devel-oping cirrhosis in rats. Hepatology 1998; 27: 1544–1553.

8. Bachem MG, Meyer DM, Melchior R, Sell KM, Gressner AM. Activationof rat liver perisinusoidal lipocytes by transforming growth factor derivedfrom myofibroblast-like cells—a potential mechanism of self perpetuation inliver fibrogenesis. J Clin Invest 1992; 89: 19–27.

9. Johnson SJ, Hines JE, Burt AD. Macrophage and perisinusoidal cellkinetics in acute liver injury. J Pathol 1992; 166: 351–358.

10. Burt AD. Cellular and molecular aspects of hepatic fibrosis. J Pathol 1993;170: 105–114.

11. Vollmar B, Ruttinger D, Wanner GA, Leiderer R, Menger MD. Modu-lation of Kupffer cell activity by gadolinium chloride in endotoxemic rats.Shock 1996; 6: 434–441.

12. Suematsu M, Oda M, Suzuki H, et al. Intravital and electron microscopicobservation of Ito cells in rat hepatic microcirculation. Microvasc Res 1993;46: 28–42.

13. Swatland HJ. Fluorimetry of bovine myotendon junction by fibre optics andmicroscopy of intact and sectioned tissues. Histochem J 1987; 19: 276–280.

14. Vollmar B, Rucker M, Menger MD. A new method for the intravitalmicroscopic quantification of hepatic sinusoidal perfusion failure using thedye bisbenzamide H33342. Microvasc Res 1996; 51: 250–259.

15. Mizgerd JP, Molina RM, Stearns RC, Brain JD, Warner AE. Gadoliniuminduces macrophage apoptosis. J Leukoc Biol 1996; 59: 189–195.

16. Roland CR, Naziruddin B, Mohanakumar T, Flye MW. Gadoliniumchloride inhibits Kupffer cell nitric oxide synthase (iNOS) induction.J Leukoc Biol 1996; 60: 487–492.

17. Lazar G. The reticuloendothelial-blocking effect of rare earth metals in rats.J Reticuloendothel Soc 1973; 13: 231–237.

18. Husztik E, Lazar G, Parducz A. Electron microscopic study of Kupffer cellphagocytosis blockade induced by gadolinium chloride. Br J Exp Pathol1980; 61: 624–630.

19. Suzuki S, Nakamura S, Serizawa A, et al. Role of Kupffer cells andthe spleen in modulation of endotoxin-induced liver injury after partialhepatectomy. Hepatology 1996; 24: 219–225.

20. Fujita S, Arii S, Monden K, et al. Participation of hepatic macrophages andplasma factors in endotoxin-induced liver injury. J Surg Res 1995; 59:263–270.


21. Liu P, McGuire GM, Fisher MA, Farhood A, Smith CW, Jaeschke H.Activation of Kupffer cells and neutrophils for reactive oxygen formation isresponsible for endotoxin-enhanced liver injury after hepatic ischemia.Shock 1995; 3: 56–62.

22. Hardonk MJ, Dijkuis FW, Hulstaert CE, Koudstaal J. Heterogeneity of ratliver and spleen macrophages in gadolinium chloride-induced eliminationand repopulation. J Leukoc Biol 1992; 52: 296–302.

23. Rai RM, Loffreda S, Karp CL, Yang SQ, Lin HZ, Diehl AM. Kupffer celldepletion abolishes induction of interleukin-10 and permits sustained over-expression of tumor necrosis factor alpha messenger RNA in the regener-ating rat liver. Hepatology 1997; 25: 889–895.

24. Callery MP, Kamei T, Flye MW. Kupffer cell blockade increases mortalityduring intra-abdominal sepsis despite improving systemic immunity. ArchSurg 1990; 125: 36–41.

25. Suzuki S, Toledo-Pereya LH, Rodriguez F, Lopez F. Role of Kupffer cellsin neutrophil activation and infiltration following total hepatic ischemia andreperfusion. Circ Shock 1994; 42: 204–209.

26. Iimuro Y, Yamamoto M, Kohno H, Itakura J, Fujii H, Matsumoto Y.Blockade of liver macrophages by gadolinium chloride reduced lethality inendotoxemic rats—analysis of mechanisms of lethality in endotoxemia.J Leukoc Biol 1994; 55: 723–728.

27. Kukan M, Vajdova K, Horecky J, Nagyova A, Mehendale HM, Trnovec T.Effects of blockade of Kupffer cells by gadolinium chloride on hepatobiliaryfunction in cold ischemia–reperfusion injury of rat liver. Hepatology 1997;26: 1250–1257.

28. Zhong Z, Qu W, Connor HD, Thurman RG. Inactivation of Kupffer cellsminimizes reperfusion injury in fat-loaded livers from ethanol-treated rats.Transplant Proc 1995; 27: 528–530.

29. McCuskey RS, Urbaschek R, McCuskey PA, Urbaschek B. In vivo micro-scopic observations of the responses of Kupffer cells and the hepaticmicrocirculation to Mycobacterium bovis BCG alone and in combinationwith endotoxin. Infect Immunol 1983; 42: 362–367.

30. Decker K. Biologically active products of stimulated liver macrophages(Kupffer cells). Eur J Biochem 1990; 192: 245–261.

31. Niwano M, Arii S, Monden K, et al. Amelioration of sinusoidal endothelialcell damage by Kupffer cell blockade during cold preservation of rat liver.J Surg Res 1997; 72: 36–48.

32. Sarphie TG, D’Souza NB, Deaciuc IV. Kupffer cell inactivation preventslipopolysaccharide-induced structural changes in the rat liver sinusoid: anelectron-microscopic study. Hepatology 1996; 23: 788–796.

33. Rimola A, Soto R, Bory F, Arroyo V, Piera C, Rodes J. Reticuloendothelialsystem phagocytic activity in cirrhosis and its relation to bacterial infectionsand prognosis. Hepatology 1984; 4: 53–58.

34. Arii S, Monden K, Itai S, et al. Depressed function of Kupffer cells in ratswith CCl4-induced liver cirrhosis. Res Exp Med 1990; 190: 173–182.

35. Gomez F, Ruiz P, Schreiber AD. Impaired function of macrophageFc-gamma receptors and bacterial infection in alcoholic cirrhosis. N Engl JMed 1994; 331: 1122–1128.

36. Shiratori Y, Teraoka H, Matano S, Matsumoto K, Kamii K, Tanaka M.Kupffer cell function in chronic ethanol-fed rats. Liver 1989; 9: 351–359.

37. Edwards MJ, Keller BJ, Kauffman FC, Thurman RG. The involvement ofKupffer cells in carbon tetrachloride toxicity. Toxicol Appl Pharmacol 1993;119: 275–279.

J. Pathol. 189: 85–91 (1999)

Documents

Microvascular consequences of Kupffer cell modulation in rat liver fibrogenesis