Stimulation of interstitial collagenase in co-cultures of rat hepatocytes and sinusoidal cells

G,\STROIINTEROL,OGY

1986:90:829-36

Stimulation of Interstitial Collagenase in Co-cultures of Rat Hepatocytes and Sinusoidal Cells KAZUO

KASHIWAZAKI,

CARLO

L. MAINARDI,

Departments and Veterans

MARGARET

and ANDREW

S. HIBBS,

of Medicine and Biochemistry, University Administration Medical Center, Memphis.

Although the fibrosis observed during chronic liver injury is the result of a complex process, the striking accumulation of collagen in end stage liver disease has provoked interest in the mechanisms that regulate both collagen production and degradation in the diseased liver. The present studies have examined the cell interactions that may be important in the regulation of collagen degradation. Although minimal amounts of interstitial collagenase activity were noted in cultures of normal hepatocytes and sinusoidal cells, the co-cultures of these cells in the presence of lipopolysaccharide showed a substantial increase in collagenase activity. When the hepatocytes were obtained from rats that had been treated with carbon tetrachloride in vivo, the enhanced activity seen in the co-cultures did not require the addition of Jipopolysaccharide. Further characterization of this interaction suggested that the increase in collagenolytic activity was partially due to the elaboration of soluble factors by the hepatocyte, which stimulated collagenase production by the sinusoidal cell population. Elaboration of collagenase activity by the sinusoidal cells was inhibited by cycloheximide, suggesting that protein synthesis was required. The proteolytic activity was abrogated by inhibitors of metalloproteinases but not by serine or thiol proteinase inhibitors. The degradation products of type I collagen were typical Received February 5, 1985. Accepted October 8, 1985. Address requests for reprints to: Margaret S. Hibbs, M.D.. Research Ser\rice (151). Veterans Administration Medical Center, 1030 Jefferson Avenue, Memphis, Tennessee 38104. K. Kashiwazaki’s present address is: Keio University, School of Medicine, 35 Shinanomacho Shinjuku-ku, Tokyo 160, Japan. This work was supported by grants from the National Institutes of Health (AM 16506, HL 27614, AA 83732, AM 01138) and funds from the Veterans Administration. M. Hibbs is a recipient of a Clinical Investigator Award from the National Institute of Arthritis. Diabetes, Digestive and Kidney Diseases [AM 01138). 8 1986 bv the American Gastroenterological Association 0016.5085/861$3.50

JEROME

M. SEYER,

H. KANG of Tennessee Tennessee

Center

for Health

Sciences

of the expected products seen with vertebrate collagenuses. Thus, it appears that the increased collagenolytic activity detected in this co-culture system is attributable to the production of interstitial collagenase by the sinusoidal cell population. Such cell-cell interactions may play an important role in the maintenance of normal connective tissue structure of the liver during disease processes. Cirrhosis is the end result of many -forms of chronic liver injury. The irreversible stage of this fibrosing process is characterized by an absolute increase in collagen content as well as an alteration in the ratio of type I to type III collagen (1,2). Although several studies have suggested a simultaneous increase in collagen production and degradation during the early phases of this fibrosing process (3,4), the findings in established cirrhosis suggest that overall hepatic collagen accumulation exceeds its degradation. Associated with these alterations in the connective tissue scaffold is a striking abnormality in cellular architecture that may play an important role in liver dysfunction. The major alterations in the collagenous content of the cirrhotic liver have been observed in the interstitial collagens, types I and III (1,5). The degradation of these collagens (6) is primarily initiated by the proteinase designated as vertebrate collagenase (E.C. 3.4.23.7). Studies correlating the activity of this proteinase with the progressive fibrosis seen in chronic liver disease have suggested that the collagenolytic activity in the liver is increased during early fibrosis but is diminished once irreversible fibrosis is established (4,7,8). Although it is possible that this initial increase in degradative ability facilitates the fibrotic process by destroying the normal extracellular matrix, another alternative is that the Abbreviation

used

in this pnper:

LPS. lil):,polvsaccharidrt.

830

KASHIWAZAKI

degradative ability is a response to injury and a physiologic component of connective tissue remodeling (9). In such an instance, the pathologic imbalance between collagen synthesis and degradation may be dependent on the loss of normal cellular communication in the face of continuing injury. Thus, it appears that studies on the production and regulation of this proteinase during liver injury are important in the determination of the overall functional role of collagenase in hepatic fibrogenesis. While interstitial collagenase has been identified in homogenized liver specimens (10,ll) and in organ culture (12), little information is available on the cell types that are responsible for collagenase production. In vitro cultures of hepatocytes (13) and Kupff er cells (l&15) have been found to produce small amounts of collagenase. The studies of Maruyama et that when rabbit hepatocytes al. (16)demonstrated and hepatic fibroblasts were co-cultured at a permissive ratio the collagenase activity could be stimulated by the addition of phorbol diesters, suggesting that cell-cell interactions may be important in the regulation of collagenase production. However, the role of cellular interactions between the myriad of cell types found in the liver in the control of collagenase production has been relatively unexplored. In this study, two components of hepatic tissue were isolated from rat liver by using primary culture techniques. These cell types included hepatocytes and sinusodial cells, the latter representing a combination of Kupffer cells and endothelial cells as well as minor cell types such as Ito and pit cells. The innate ability of these cell populations to produce interstitial collagenase as well as their ability to elaborate collagenolytic activity in mixed culture was determined. The results indicate that cell-cell interactions may play an important role in the production of interstitial collagenase in the liver. Materials

GASTROENTEROLOGY

ET AL.

and Methods

Animals Male Wistar rats (weighing 80-150 g) were maintained under standard conditions. In some experiments the rats were injected subcutaneously with Ccl, (1 mgikg in an equal volume of mineral oil) twice a week for 4 wk or thioacetamide (50 mgikg) intraperitoneally daily for 6 wk.

Collagen Preparation Type I collagen was extracted from fetal calf skin and purified by the method of Glimcher et al. (17). The purified collagen was radiolabeled using [14C]acetic anhydride (New England Nuclear, Boston, Mass.) according to the method of Cawston and Barrett (18). The specific activity of the collagen preparation ranged from 80,000 to 120,000 cpmimg of collagen, depending on the preparation used.

Collagenuse

Vol. 90, No. 4

Assays

Interstitial collagenase activity was determined using a reconstituted fibril assay as described by Nagai et al. (19). One hundred micrograms of 14C-labeled collagen were incubated with 120 ~1 of concentrated media in a reaction mixture containing 50 mM Tris-HCl, 5 mM CaC12, and 0.02% NaN, in a final volume of 300 ~1 for 16 h at 37°C. To determine the total enzyme activity present, aminophenylmercuric acetate (final concentration, 1 mM) was added to the reaction mixture to activate any latent enzyme that might be present (20). After the intact collagen was precipitated by centrifugation at 10,000 g for 5 min, an aliquot of the supernatant was counted in a liquid scintillation counter. Under the experimental conditions, ~5% of the acetylated substrate was susceptible to nonspecific proteolysis by trypsin or purified neutrophil gelatinase. Background activity was subtracted from each sample before determination of the specific activity. One unit of enzyme activity is defined as the degradation of 1 pg of collagen per minute. The determinations were performed in duplicate or triplicate and the values shown represent mean values. Intraexperimental variation was ~5%. In assays where soluble collagen was used, the reaction mixture was identical to that used in the fibril assay, except that 50 mM arginine was added to prevent fibril formation and 35 ~1 of concentrated media was utilized. Assays with soluble collagen were done at 25°C for 40 h. After the termination of the assay by the addition of ethylenediaminetetraacetic acid (final concentration, 10 mM), the reaction products were analyzed on 7.5% polyacrylamide gels.

Polyacrylamide

Gel Electrophoresis

Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate was performed by the method of Laemmli (21). The gels were stained with 0.1% Coomassie Brilliant Blue in 50% methanol:lO% acetic acid (volivol) and destained in 10% methanol:lO% acetic acid (volivol).

Isolation

of Hepatocytes

Hepatocytes from normal rats and rats treated with Ccl4 or thioacetamide were isolated by a modification of the method of Jeejeebhoy et al. (22). Briefly, 0.05% bacterial collagenase (type I, Sigma Chemical Co., St. Louis, MO.) in Ca”, Mg’ ’ free Hank’s balanced salt solution [Grand Island Biological Co. (GIBCO), Grand Island, N.Y.], pH 7.4, saturated with 5% C02--95% 02, was perfused through the liver via the portal vein of heparinized, anesthetized rats at a flow rate of 10 mlimin. After 15 min, the liver was harvested and minced gently with scissors. The cells were filtered through surgical gauze and sedimented at 1 g for 10 min at 22°C. Using low speed centrifugation conditions (50 g for 90 s), the precipitated cells containing the hepatocytes were washed three times in cold Dulbecco’s phosphate buffered saline (GIBCO). Viability of the isolated cells was determined by trypan blue exclusion. Only hepatocyte preparations with a viability of 585% were used. The hepatocytes were cultured in 75-cm”, 100-mm plastic dishes (Falcon, Oxnard, Calif.) at a cell

April

1986

COLLAGENASE

PRODUCTION

BY K,4T LIVER CELLS

83

1

density of 4 x 10” viable cells in 10 ml of Williams’ medium E (GIBCO) containing 2 mM glutamine, 100 U/ml of penicillin, 100 &ml of streptomycin, 0.25 pg/ml of amphotericin B, 10 mM HEPES, and 0.2% lactalbumin hydrolysate. The medium was changed at 2 and 24 h. No contamination by peripheral blood cells, Kupffer cells, or fibroblasts was noted on microscopic inspection of the cell population. Isolation

of Sinusoidal

Cells

Sinusoidal cells from normal livers were isolated by the method of Knook and Sleyster (23). Briefly, perfused livers were minced and then stirred in Gey’s balanced salt solution (GIBCO), pH 7.4, containing 0.2% (wtivol) pronase E (type XIV, Sigma Chemical Co.) for 60 min at 37°C. After filtration through gauze, the dissociated cells were collected by centrifugation at 300 g for 5 min. The cells were resuspended in metrizamide (17.5% wtivol, final density = 1.089) and centrifuged at 1400 g for 45 min at 22°C. The sinusoidal cells were harvested from the upper layer of the tube and were washed three times in Gey’s balanced salt solution. Viable cells (6 x 10”) were incubated in lOO-mm dishes in 10 ml of Williams’ medium E as described for hepatocyte cultures. Nonadherent cells were removed after 24 h. No contamination by fibroblasts, hepatocytes, or peripheral blood cells was noted on microscopic inspection. The isolated sinusoidal cell population was estimated to contain 44%) Kupffer cells as determined by nonspecific esterase staining (24).

DAYS Figure

Cell Culture Both hepatocytes and sinusoidal cells were maintained in serum-free Williams’ medium E supplemented with 0.2% lactalbumin hydrolysate, and the medium was changed every l-3 days. Some cultures were stimulated by the addition of 30 pg/ml of lipopolysaccharide (LPS) or 50 ngiml of phorbol myristic acetate. Initially, co-culture experiments used the same number of cells as were used in

the individual cultures (e.g. 4 x 10” cells/ml for hepatocytes and 6 x 10” cells/ml for sinusoidal cells). While the cell ratio was always 2 parts hepatocyte to 3 parts sinusoidal cells, the absolute cell number was later decreased to 5 X 10” cells/ml. There were no significant differences in the results obtained with variation of the absolute cell number when the cell ratios were maintained. The harvested medium was concentrated 60-fold by lyophilization before determination of enzymatic activity. The results represent the cumulative experience from a minimum of three separate experiments. As is common in biologic systems, the amount of collagenase produced by the cells obtained from different groups of animals was variable such that the interexperimental levels of collagenase varied by 25+500/o. When the data were normalized on a percentage control basis, the interexperimental variation was 10%15”/0. In both cases, the stimulated cultures produced significantly greater amounts of collagenase (p 5 0.05) as compared to controls. Representative experiments are shown.

I. Collagenase activity in individual cell preparations and co-culture using cells obtained from normal livers Cells were cultured at 4 x 10’ cells/ml for hepatocytes for sinusoidal cells both in the and 6 x 10" cells/ml isolated cell cultures and in co-cultures in serum-free medium. Media were harvested every 48 h for 6 days. concentrated 60-fold, and collagenase activity was determined using the reconstituted fibril assay. Aminophenylmercuric acetate (1 mM final concentration) was included in the assays to activate latent collagenase. Solid line, controls: broken line, LPS-stimulated cultures; A-A. hepatocytes alone: 0-O. sinusoidal cella alone; m-m, co-culture of hepatocytes and sinusoidal cells. Results are expressed in terms of mU/ml of concentrated media.

Results Collagenase Production Normal Livers

by Cells

From

To determine the relative ability of normal hepatocytes and sinusoidal cells to produce collagenase, each of the cell types was cultured alone as described in Methods. No significant activity was noted with either cell population at 2,4, or 6 days of culture. Additionally, when the two cell populations were co-cultured at a cell ratio of 2: :I (hepatocyte to sinusoidal cell), no significant collagenolytic activity could be detected [Figure 1). As previous studies have suggested that agents such as phorbol diesters (16) and LPS (15~5) enhance

the

these

agents

production

of collagenase,

to stimulate

collagenase

the was

ability

The addition of LPS to hepatocyte cultures failed stimulate the production of significant amounts collagenase

activity.

tivity

was

noted

most

dramatic

Although

in the

some

sinusoidal

enhancement

in

increase cell

activity

of

examined. to of

in ac-

cultures,

the

was

seen

when hepatocytcs were co-cultured with sinusoidal cells in the presence of LPS. There was a fourfold greater stimulation of collagenase production in the co-cultures treated with LPS as compared to the

832

KASHIWAZAKI

GASTROENTEROLOGY

ET AL.

B.

Ls 3

D.

~

DAYS Figure

2. Collagenase activity in cells isolated from normal livers and livers of rats treated with Ccl,. Cell density in the for isolated cell cultures was 4 x lo5 cells/ml hepatocytes and 6 x lo5 cells/ml for sinusoidal cells. In the co-culture experiments, 2 x 10'hepatocytesiml and 3 x lo5 sinusoidal cells/ml were used. Media were harvested every 24 h, concentrated 60-fold, and then assayed in the reconstituted fibril assay for 16 h at 37°C in the presence of aminophenylmercuric acetate to assess the total amount of collagenase present. A. Normal liver cells. B. Normal liver cells in the presence of LPS. C. Hepatocytes from Ccl,-treated rat. D. Hepatocytes from CC&-treated rat in the presence of LPS. O-e, hepatocytes alone; A-A, sinusoidal cells alone; m-w, co-culture of hepatocytes and sinusoidal cells.

ment of collagenolytic activity seen in co-cultures using hepatocytes from CC14-treated animals represented a 1%fold increase in activity when compared with the activity seen in unstimulated co-cultures of normal cells. When hepatocytes from Ccl.+-treated livers were co-cultured with sinusoidal cells in the presence of LPS, a marked increase in collagenolytic activity was also noted; however, the increase in collagenase activity was greater during the first 2 days of culture in the LPS-treated group (Figure 2D). In the absence of LPS, the collagenase activity in the co-cultures using CC14-treated hepatocytes was predominately in latent form (>70%). In the presence of LPS, an increase in the amount of active enzyme present was detected (data not shown). Similar results were obtained when hepatocytes from thioacetamide-treated livers were used (data not shown). These data demonstrated that co-cultures of in vivo, damaged hepatocytes and normal sinusoidal cells did not require the addition of LPS to induce increased collagenase production.

Determination of the Role of Various Cell Types in the Enhanced Production of Collagenase Although it was clear that the co-culture of hepatocytes with sinusoidal cells produced in-

sinusoidal cell cultures treated with the same agent (Figure 1). The collagenase activity detected was predominantly in an active form (>90%). Similar data were obtained when the cultures were treated with phorbol myristic acetate (data not shown). These data suggested a synergism between normal hepatocytes and sinusoidal cells in the production of collagenase in response to stimulatory agents. Collagenase Hepatocytes

1

!

,100 f E

Production in Cultures Using Obtained From Injured Liver

As the initial studies were performed with cells obtained from normal livers, it was of interest to determine if a similar synergism occurred when hepatocytes obtained from livers injured in vivo with Ccl, or thioacetamide were studied. When hepatocytes from normal animals were used, only the LPS-stimulated co-cultures of hepatocytes and sinusoidal cells produced significant quantities of collagenase [Figure 2A and 2B), and the activity was present predominately in active form. When the hepatocytes from the injured livers were co-cultured with sinusoidal cells, the induction of collagenolytic activity in the co-cultures was markedly increased even in the absence of LPS (Figure 2C). The enhance-

Vol. 90, No. 4

Figure

n

l-l

3. Stimulation of interstitial collagenase production by sinusoidal cells by conditioned media from hepatocyte cultures. Hepatocyte or sinusoidal cell culture media conditioned in the presence or absence of LPS ware incubated with other cell types for 24 h, after which the media from the secondary cultures were harvested, concentrated, and assayed for collagenolytic activity using the rkconstituted fibril assay. Control media were obtained by incubating media in the absence of cells in the presence or absence of LPS under the same conditions. Open bars, conditioned media added to sinusoidal cell cultures; solid bars, conditioned media added to hepatocyte cultures. A, media conditioned in the absence of cells; B, media conditioned in the absence of cells with LPS; C, media conditioned in the presence of cells: D, media conditioned in the presence of cells stimulated with LPS.

April

1986

creased amounts of collagenase, it remained to be determined which cell was responsible for this enhanced activity. Thus, the culture medium from each isolated cell type was examined for its ability to stimulate collagenase production by the other cell type. In these studies, the separated cells were cultured in the presence or absence of LPS. After 24 h, the media were aspirated, cleared of cellular debris by centrifugation, and added to the other cell type. After an additional 24 h, the media were removed and assayed for collagenase activity. The media from hepatocytes cultured in the presence of LPS significantly stimulated the production of collagenase by the sinusoidal cell population (Figure 3). This increase in activity was 2.5fold greater than the activity that would be seen if sinusoidal cells were stimulated with media that had been conditioned with LPS in the absence of cells. No enhancement of collagenase activity was detected when media that had been initially incubated with sinusoidal cells were incubated with hepatocytes. These findings suggested that, after stimulation with LPS, hepatocytes produced a soluble factor that stimulated the production of collagenase by the sinusoidal cells. To exclude the possibility that the stimulation of sinusoidal cell collagenase production by the LPSstimulated hepatocyte culture media was attributable to LPS in the hepatocyte culture media, the hepatocytes were incubated in the presence of LPS for 2 h and then washed extensively before incubation in media without LPS for 6 h. The LPS-free, hepatocyte-conditioned media obtained were then incubated with sinusoidal cells for 24 h, after which time the media were concentrated and analyzed for collagenolytic activity. Sinusoidal cells incubated

COLLAGENASE

Figure

PRODUCTION

BY RAT LIVER CELLS

833

5. Collagenase production by fibroblasts stimulated with hepatocyte media. Conditioned media from cultures of hepatocytes that had been incubated in the presence or absence of LPS were added to cultures of rat dermal fibroblasts. After 6 days the media were collected. concentrated Z&fold, and assayed for collagenase activity in the presence of aminophenylmercuric acetate. Results are expressed in terms of milliunits per 10” fibroblasts. A, LPS-containing media conditioned in the absence of cells: I?, media from unstimulated hepatocyte cultures; C, media from LPS-stimulated hepatocyte cultures.

with hepatocyte-conditioned media obtained after the removal of LPS from the system showed enhanced collagenolytic activity (Figure 4). This increased collagenase production by the sinusoidal cells represented a more than twofold increase in activity in comparison to the activity seen when the sinusoidal cells were stimulated with the original concentration of LPS. These data indicated that the effect of LPS in this system could not be solely attributed to its direct effect on the sinusoidal cells, but rather, it appeared to act by inducing hepatocytes to elaborate soluble factors that stimulated the sinusoidal cells.

Effect of Hepatocyte-Conditioned Medium on Fibroblast Collagenase Production

Figure

4. Ability of lipopolysaccharide (LPS) to induce hepatocytes to produce soluble factors that stimulate collagenase production by the sinusoidal cells. Isolated hepatocytes (4 x 10’ cells/ml) were incubated with 30 &ml of LPS for 2 h, washed extensively, and then incubated for 6 h in the absence of LPS. The conditioned media were incubated with sinusoidal cells (6 x 10’ cells/ml) for 24 h with and without additional LPS (30 pgiml): the media from the sinusoidal cell cultures were concentrated and assayed for collagenase in the reconstituted fibril assay. A, media conditioned in the absence of cells: U, media conditioned in the absence of cells + LPS: C, hepatocyte-conditioned media.

During the cirrhotic process, an increase in the number of fibroblastlike cells has been noted. Thus, the ability of rat fibroblasts to produce increased amounts of collagenase in response to the hepatocyte-conditioned media was examined. In these experiments, dermal fibroblasts obtained from neonatal rats [American Type Culture Collection, Rockville, Md., ATCC-CRL 1439) were incubated with LPS-stimulated, hepatocyte-conditioned culture media. The enhanced collagenolytic production that was detected represented a more than sevenfold increase in activity as compared with fibroblasts cultured with media containing LPS or media from unstimulated hepatocytes (Figure 5). Thus, it appears that not only sinusoidal cells but also fibroblasts could be induced to produce collagenase in response to soluble factors from hepatocytes.

834

KASHIWAZAKI

ET AL.

GASTROENTEROLOGY

DAYS -

Figure

6. Effect of cycloheximide on interstitial collagenase production by sinusoidal cells. Sinusoidal cell cultures were incubated with conditioned media from LPSstimulated hepatocyte cultures in the presence or absence of cycloheximide. The sinusoidal cell culture media were harvested at 24 and 48 h, concentrated, and assayed for collagenase activity. Open bars, no cycloheximide; solid bars, 0.25 pg/ml of cycloheximide.

Characterization of the Interstitial Collagenase Produced by the Sinusoidal Cells To determine whether the collagenase activity in the sinusoidal cell cultures seen after stimulation with hepatocyte-conditioned media represented an enhanced release of the enzyme or an increase in synthesis of the proteinase, 0.25 kg/ml of cycloheximide was added to sinusoidal cell cultures at the time of stimulation with hepatocyte-conditioned media. As seen in Figure 6, the stimulation of colla-

Table

1. Inhibition Produced Stimulation Media”

Additive APMA p-Chloromecuribenzoate N-Ethylmaleimide Phenylmethanesulfonylfluoride (in 10% dimethylsulfoxide) Dimethylsulfoxide Dithiothreitol 2-Mercaptoethanol L-Cysteine EDTA Fetal calf serum Rat serum Hz0

of the Interstitial Collagenase by Sinusoidal Cells After With Hepatocyte-Conditioned

Concentration

genase activity expressed by the sinusoidal cells was reduced to basal levels by inclusion of cycloheximide in the culture media. These data suggested that the increase in activity seen in the sinusoidal cell cultures after stimulation with hepatocyte culture media reflected an increase in the synthesis of this enzyme by the sinusoidal cells. To further characterize the collagenolytic activity derived from the sinusoidal cells, the ability of various proteinase inhibitors to diminish the enzyme activity was examined. The proteolytic activity seen in the sinusoidal cell culture media was inhibited by ethylenediaminetetraacetic acid, dithiothreitol, and serum, but not by thiol or serine proteinase inhibitors, suggesting that the activity is due to a metalloendoproteinase (Table 1).When the culture media from the stimulated sinusoidal cells were incubated with type I collagen, the specific reaction products characteristic of interstitial collagenase were noted (Figure 7). These results indicated that the collagenolytic activity observed represented a specific degradation of collagen similar to the results obtained with classic mammalian collagenases. Additionally, they demonstrated that the collagenolytic activity observed could not be attributed to residual bacterial collagenase.

8

Activity (% Control)

1 mM 3mM 5mM 3mM 1mM

97.0 88.3 59.7 62.7 62.1

10% 10 mM 1 mM 20 mM 5mM 20 mM 10% 10%

58.4 4.5 69.2 30.2 48.1 5.5 14.6 16.7 100

APMA, aminophenylmercuric acetate; EDTA, ethylenediaminecells were incubated for 24 h with tetraacetic acid. n Sinusoidal media from hepatocyte cultures that had been stimulated with 30 pgiml lipopolysaccharide for 24 h. The media obtained from the sinusoidal cell cultures were concentrated 60-fold, and the ability of various agents to inhibit activity in the reconstituted fibril assay was assessed.

Vol. 90. No. 4

1 HI 2Hl

auli

a2rll

ENZYME B. F.

1 2 Figure

3

7. Degradation products of type I collagen. Concentrated media from sinusoidal cells stimulated with hepatocyte-conditioned medium were incubated with 100 pg of type I collagen for 40 h at 25°C. Reaction products were separated on a 7.5% polyacrylamide gel. Lane 1, collagen alone; lane 2, hepatocyte-stimulated sinusoidal cell culture media alone: lane 3, collagen + hepatocyte-stimulated sinusoidal cell culture media.

April 1986

Discussion The excessive accumulation of interstitial collagen during chronic liver injury suggests an abnormal balance in collagen synthesis and degradation In this study, we have examined the cellular interactions that may be important in the regulation of the production of interstitial collagenase, the proteinase required for the initiation of interstitial The results indicate that collagen degradation. hepatocytes stimulated the sinusoidal cell population to produce interstitial collagenase, indicating that cellular interaction may be an important determinant in the regulation of collagenase production. Further characterization of this cellular interaction revealed that the increased collagenase production by the sinusoidal cells was partially attributable to soluble factors elaborated by the hepatocyte. Other studies examining the effect of cellular interactions on collagenase production have indicated that cytokines may be important in this process. Epithelial-fibroblast interactions, which lead to enhanced collagenase production by the rabbit cornea1 stroma, are mediated by a 20,000-dalton cytokine secreted by the cornea1 epithelium (26,27). Interleufor the kin 1 has been shown to be responsible enhanced collagenase production by rabbit synovial cells (28) and human fibroblasts (29) after stimulation with monocyte-conditioned media. Studies are in progress to characterize the soluble factors generated by the hepatocyte that stimulate sinusoidal cells to produce collagenase. Although the data indicate that the induction of collagenase is due in part to the elaboration of soluble factors by the hepatocyte, the profound difference in activity observed in co-cultures as opposed to the activity seen in the sinusoidal cell cultures stimulated with hepatocyte-conditioned media indicates that cell-cell communication by mechanisms other than soluble mediators that stimulate collagenase production could also be an important determinant of collagenolytic potential. Similar findings were obtained when the stimulation of collagen production was examined in a co-culture system that used mononuclear cells and fibroblasts (30). It is unclear at present whether this is related to factors secreted by the sinusoidal cell population or direct cell-cell communication. Although the relative contribution of soluble factors and direct cell-cell interaction remains to be resolved, it is evident that the co-culture of hepatocytes with sinusoidal cells leads to increased collagenase production. Although the cells from normal livers required the addition of LPS for this stimulatory activity to be detected, co-cultures using hepatocytes from in vivo injured livers showed en-

COLLAGENASE

PRODUCTION

BY KAT LIVER CELLS

835

hanced collagenase production in the absence of LPS. These findings suggested that the in vitro stimulation of hepatocytes by substances such as LPS may be a relevant model for in vivo tissue injury. The sinusoidal cell population is a heterogeneous cell population including endothelial cells, Kupffer cells, and minor cell types such as Ito cells and pit cells; thus, it is not possible to determine from the present study which cell is responsible for the increased production of collagenase. As endothelial (32) from other cells (31) and tissue macrophages sources have been reported to produce interstitial collagenase, either cell type could be responsible for the increased production of collagenase seen after stimulation with the hepatocyte-conditioned medium. Perhaps more intriguing is the possibility that the Ito cell may be the responding cell. Although this cell is a minor component of the nonparenchymal cell population in the normal liver (33), it has been demonstrated that this cell increases in number (34) and adopts a fibroblastic phenotype in the injured liver (35). Our results, indicating that fibroblasts responded to hepatocyte-conditioned media with a sevenfold increase in collagenase activity, indicate a potential role for cells of the fibroblastic phenotype in the response to the hepatocyte factors. Thus, multiple cell types could be the source of collagenase in the injured liver. The collagenase identified in the stimulated sinusoidal cell cultures was similar to the traditional mammalian interstitial collagenases identified in other cell systems (36-38). In cultures stimulated with LPS, the proteinase was predominately detected in an active form (>90%). In the absence of LPS, the enzyme was secreted in a predominately latent form. These findings suggest that LPS may induce the release of an activator from the hepatocyte or the sinusoidal cells, or both. Such physiologic activators have been reported in other collagenolytic systems (39) and may play a major role in the expression of collagenase activity in vivo. Though previous studies have demonstrated the production of collagenase by liver cells, the present study has sought to better define the cell-cell interactions that may be important in the regulation of collagenase expression. Collagenase production is a complex process; the overall regulation of collagenolytic activity is dependent on the presence of physiologic activators and inhibitors of mterstitial collagenase. Although the mechanism of the increased collagenase activity seen in hepatocyte-sinusoidal cell co-cultures remains to be determined, it is possible that the loss of hepatocytes during ongoing injury could lead to a decrease in factors that stimulate collagenase production while collagen deposition continues. Thus, the ability of th’e hepatocyte to

836

GASTROENTEROLOGY Vol. 90, No. 4

KASHIWAZAKI ET AL.

stimulate the production of interstitial collagenase by the nonparenchymal cell populations could be critical in maintaining the balance between collagen deposition and degradation required to maintain normal connective tissue composition and cellular architecture.

References

19. Nagai Y, Lapierre C, Gross J. Tadpole collagenase: 20.

21.

22.

1. Seyer JM. Interstitial

collagen polymorphism in rat liver with CC&-induced cirrhosis. Biochim Biophys Acta 1980;629: 490-8. 2. Seyer JM, Hutchison ET, Kang AH. Collagen polymorphism in normal and cirrhotic human liver. J Clin Invest 1977; 59241-8. 3. Okasaki I, Maruyama K, Ono A, Kobayashi T, Suzuki H. Reversibility of hepatic fibrosis in toxic liver injury. J UOEH (Japan) 1982;4(Suppl):169-81. 4. Maruyama K, Feinman L, Fainsilber Z, Nakano M, Okazaki I, Leiber C. Mammalian collagenase increases in early alcoholic liver disease and decreases with cirrhosis. Life Sci 1982; 30:1379-84. 5. Rojkind M, Giambrone MA, Biempica L. Collagen types in normal and cirrhotic livers. Gastroenterology 1979;76:710-9. 6. Harris ED, Cartwright EC. Mammalian collagenase. In: Barrett AJ, ed. Proteinases in mammalian cells and tissues. New York: Academic, 1977:249-78. 7. Takahashi S, Dunn MA, Seifter S. Liver collagenase in murine schistosomiasis. Gastroenterology 1980;78:1425-31. R. Collagenase in experimental 8. Montford I, Perez-Tamayo carbon tetrachloride cirrhosis of the liver. Am J Path01 1978;92:411-20. 9. Grill0 HC, Gross J. Collagenolytic activity during mammalian wound repair. Dev Biol 1967;15:300-17. 10. Maruyama K, Feinman L, Okasaki I, Leiber CS. Direct measurement of neutral collagenase activity in homogenates from baboon and human liver. Biochim Biophys Acta 1981; 658:124-31. of 11. Fujiwara K, Sakai T, Oda T, Igarashi S. Demonstration collagenase activity in rat liver homogenates. Biochem Biophys Res Commun 1973;54:531-7. K, Kashiwazaki K, Kamegaya K, 12. Okasaki I, Maruyama Tsuchiya M. Hepatic collagenase activity in experimental hepatic fibrosis. Biochem Exp Biol 1976;12:124-38. 13. Nagai Y, Hori H, Hata R, Konomi H, Suriada H. Collagenase production by adult rat hepatocytes in primary culture. Biomed Res 1982:3:345-g. 14. Fujiwara K, Sakai T, Oda T, Igarashi S. Presence of collagenase in Kupffer cells of rat liver. Biochem Biophys Res Commun 1973;54:531-7. 15. Bhatanagar R, Schade U, Rietschel E, Decker K. The involvement of prostaglandin E and adenosine-3’,5’-monophosphate in lipopolysaccharide-stimulated collagenase release by rat Kupffer cells. Eur J Biochem 1984;125:125-30. 16. Maruyama K, Okazaki I, Kobayashi T, Suzuki H, Kashiwazaki K, Tsuchiya M. Collagenase production by rabbit liver cells in monolayer culture. J Lab Clin Med 1983;102:543-50. 17. Glimcher MJ, Francois CT, Richards K, Krane S. The presence of organic phosphorus in collagens and gelatins. Biochim Biophys Acta 1964;93:585-602. 18. Cawston TE, Barrett A. A rapid and reproducible assay for collagenase using [l-‘Xl-acetylated collagen. Anal Biochem 1979;99:340-5.

23. 24.

25.

26. 27.

28.

29.

30.

31.

32.

33.

preparation and purification. Biochemistry 1966;5:3123-30. Seller A, Cartwright E, Murphy G, Reynolds JJ. Evidence that latent collagenases are enzyme-inhibitor complexes. Biochem J 1977;163:303-7. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227:680-5. Jeejeebhoy KN, Ho J, Greenberg GR, Phillips J, BruceRobertson A, Sodtke U. Albumin, fibrinogen and transferrin synthesis in isolated rat hepatocyte suspensions. Biochem J 1975;146:141-55. Knook DL, Sleyster EC. Separation of Kupffer and endothelial cells by centrifugal elutriation. Exp Cell Res 1976;99:444-9. Yam LT, Li CY, Crosby WH. Cytochemical identification of monocytes and granulocytes. Am J Clin Path01 x971:55: 283-90. Wahl LM, Olsen CE, Sandberg AL, Mergenhagen SE. Prostaglandin regulation of macrophage collagenase production. Proc Nat1 Acad Sci USA 1977;74:4955-8. Johnson-Muller B, Gross J. Regulation of cornea1 collagenase production. Proc Nat1 Acad Sci USA 1978;75:4417-21. Johnson-Wint B, Bauer EA. Stimulation of collagenase synthesis by a 20,000 dalton epithelial cytokine. J Biol Chem 1985;260:2080-5. Mizel SB, Dayer JM, Krane SM, Mergenhagen SE. Stimulation of rheumatoid synovial cell collagenase and prostaglandin production by partially purified lymphocyte activating factor (interleukin 1). Proc Nat1 Acad Sci USA 1981;78:2474-7. Postlethwaite AE, Lachman CB, Mainardi CL, Kang AH. Interleukin 1 stimulation of collagenase production by cultured fibroblasts. J Exp Med 1983;157:801-6. Hibbs MS, Postlethwaite AE, Mainardi CL, Seyer JM, Kang AH. Alterations in collagen production in mixed mononuclear-hbroblast cultures. J Exp Med 1983;157:47-59. Gross JL, Moscatelli D, Jaffe EA, Rifkin DB. Plasminogen activator and collagenase production by cultured capillary endothelial cells. J Cell Biol 1982;95:974-81. Horwitz AL, Kelman JA, Crystal RG. Activation of alveolar macrophage collagenase by a neutral protease secreted by the same cell. Nature 1976;264:772-4. Voss B, Rauterberg J, Pott G, et al. Nonparenchymal cells cultivated from explants of fibrotic liver resemble endothelial and smooth muscle cells from blood vessel walls. Hepatology 1982;2:19-28.

34. Kent G, Gay S, Minick

35.

36.

37.

38.

39.

OT. Vitamin A-containing lipocytes and formation of type III collagen in liver injury. Proc Nat1 Acad Sci USA 1976;73:3719-22. DeLeeuw AM, McCarthy SP, Geerts A, Knook DL. Purified rat liver fat-storing cells in culture divide and contain collagen. Hepatology 1984;4:392-403. Stricklin GP, Eisen AZ, Bauer EA. Jeffrey JJ. Human skin fibroblast collagenase: chemical properties of precursor and active forms. Biochemistry 1978;17:2331-7. Evanson JM, Jeffrey JJ, Krane SM. Human collagenase: identification and characterization of an enzyme from rheumatoid synovium in culture. Science 1967;158:499-502. Cawston TE, Murphy G. Mammalian collagenases. In: Lorand L, ed. Methods in enzymology. Vol. 80. New York: Academic, 1981:711-22. Vater CA, Nagase H, Harris ED. Purification of an endogenous activator of procollagenase from rabbit synovial fibroblast culture medium. J Biol Chem 1983;256:9374-82.