0270-9139/84/0405-0843$02.00/0 HEPATOLOGY Copyright ici 1984 by the American Association for the Study of Liver Diseases

Vol. 4, No. 5, pp. 843-848, 1984 Printed in U. S.A.

Acetaldehyde Stimulates Collagen and Noncollagen Protein Production by Human

Fibroblasts KENNETH HOLT, MICHAEL BENNETT AND MARIO CHOJKIER

D iuis ion of Gastroenterology, Depart rnent of Med icine, Veterans Ad ministrat ion Medical Center and University of California at Sun Diego, Sun Diego, California 92161

The mechanigms responsible for the increased hepatic collagen deposition in alcoholic cirrhosis remain unknown. The question of whether ethanol or acetaldehyde has a direct effect on collagen and noncollagen protein production was investigated in human fibroblasts with no detectable activity of alcohol dehydrogenase to distinguish the effects of these metabolites. To eliminate environmental factors, protein production by confluent human skin, fetal and hepatic fibroblasts was studied after three passages. Cells were labeled with [5-3H]proline for 4 hr in the presence of 0.2 mM ascorbate alone or with addition of either ethanol (50 mM) or acetaldehyde (0 to 300 p M j . Rates of protein production were calculated from the radioactivities of collagenase-sensitive and collagenase-resistant proteins. Skin fibroblasts from alcoholic individual either with cirrhosis or without liver disease have comparable rates of collagen and noncollagen protein production. Acetaldehyde, in a concentration found in the liver during ethanol abuse, significantly increased collagen production by human skin fibroblasts (up to 140%), fetal fibroblasts (up to 240%) and hepatic fibroblasts (up to 70%) but the addition of ethanol had no significant effect on basal collagen production. The effect of acetaldehyde was dose-related and affected noncollagen protein produc- tion in a similar manner. Acetaldehyde did not cause changes in either proline transport or the specific activity of the proline precursor pool. This newly recognized stimulation of collagen production by acetaldehyde may be a possible mechanism of fibrogenesis in alcoholic individuals.

Epidemiological studies have shown that the likelihood of developing alcoholic liver disease is directly related to the amount of ethanol consumed over a lifetime (1, 2). Ethanol causes various forms of hepatic damage (3) but only some alcoholics develop cirrhosis (1). Therefore, factors other than quantity of alcohol consumed must be operative in development of cirrhosis. These factors, which may be genetic or acquired are still poorly under- stood. The concordance for liver cirrhosis was greater in monozygotic twins than in dyzygotic twins (4) and could not be explained by differences in the concordance for alcoholism. These studies suggest a genetically influ-

Received February 6, 1984; accepted April 23, 1984 This study was supported by the Research Service, Veterans Admin-

istration and by a grant from the Academic Senate, University of California, San Diego (R-F143).

M. Bennett and K. Holt were supported by a National Institutes of Health Training Program Grant (AM-07202).

This work was presented in part a t the 84th Annual Meeting of the American Gastroenterological Association and published as an abstract (Gastroenterology 1983; 84:1363).

Address reprint requests to: Mario Chojkier, M.D., Veterans Admin- istration Medical Center (111D), 3350 La Jolla Village Drive, San Diego, California 92161.

enced susceptibility to develop liver fibrosis among al- coholic individuals.

Alcohol-induced hepatic fibrogenesis occurs in the ab- sence of obvious hepatocellular necrosis in baboons (5 ) and in patients (6). These findings indicate that hepatic fibrosis may be stimulated by mechanisms other than necrosis or inflammation. If factors other than necroin- flammation influence hepatic fibrogenesis, then effects of ethanol or its metabolites on connective tissue metab- olism represent likely primary factors. Although there are data showing that ethanol increases in uitro incor- poration of radioactive proline into collagen in liver tissue from patients with cirrhosis (7), there is still no clear evidence that this increase represents a specific effect on collagen production.? The lack of a stimulatory effect of ethanol on collagen production in cultured liver fibroblasts (8) and lung fibroblasts (9) suggests that ethanol itself does not contribute directly to the alcohol- induced hepatic fibrogenesis.

t The radioactivity present in high molecular weight protein is considered to reflect net production resulting from incorporation during biosynthesis and loss due to degradation. It can be presented as:

Net Production = Biosynthesis - Degradation.

843

a44 HOLT, BENNETT AND CHOJKIER HEPATOLOGY

Although acetaldehyde has been shown to mediate many of the biological effects of ethanol (10, ll), its role on collagen metabolism is unknown (12-14). The ques- tion of whether ethanol or acetaldehyde has a direct effect on collagen production was investigated in human fibroblasts with no detectable activity of alcohol dehy- drogenase to distinguish the effects of these metabolites. We find that acetaldehyde at concentrations found in the liver during alcohol abuse (15, 16), but not ethanol, stimulates collagen and noncollagen protein production by human fibroblasts.

MATERIALS AND METHODS [5-3H]Proline (19 mCi per pmole), ~~-hydroxy-[2-’~C]

proline (15 mCi per mmole) and ACS scintillation fluid were obtained from Amersham Corporation (Ar- lington Heights, Ill.); [3H] toluene and [ 14C] toluene stan- dards were purchased from New England Nuclear (Bos- ton, Mass.). Chromatographically purified bacterial col- lagenase form I11 was from Advance Biofactures Com- pany (Lynbrook, NY). Sources of other chemicals were: bovine IgG standard, protein dye reagent and AG50W- X8 resin (100 to 200 mesh) from Bio-Rad Laboratories (Richmond, Calif.); L-ascorbic acid, 0, a’-dipyridyl(2,2’- bipyridine), L-proline, hydroxy-L-proline, cycloheximide and trichloroacetic acid from Sigma Chemical Company (St. Louis, Mo.); absolute ethanol from Aaper Alcohol and Chemical Company (Louisville, Ky.); and certified acetaldehyde from Fisher Scientific Company (Fair Lawn, NJ). Human fetal fibroblasts AF2 were obtained from the Core Cell Culture Facility (University of Cali- fornia, San Diego).

PATIENTS Studies were performed in 17 volunteers selected from

the Outpatient Clinic at the San Diego Veterans Admin- istration Medical Center; 9 with liver cirrhosis and 8 without liver disease. All individuals were Caucasian male alcoholics and their age was similar in both groups (59 f 1, mean f S.E. vs. 61 f 2 years). The daily mean ethanol intake was also similar in both groups (79 f 15 vs. 72 f 11 gm per kg X years*). Patients were excluded if they had a history of drug abuse, blood transfusions or evidence of hepatitis, hemochromatosis, Wilson’s or drug-induced liver disease. The individuals in the control group had no evidences of liver disease by history, phys- ical examination, laboratory tests, ultrasonography and radionuclide scan. The patients in the other group were medically stable with biopsy-proven cirrhosis. Informed consent for participation in the investigation was ob- tained from each individual. This investigation was ap- proved by the Human Subjects Committee, University of California a t San Diego.

FIBROBLAST CULTURES Primary cultures of fibroblasts were initiated from

normal skin biopsies in 17 individuals. In two individuals, primary cultures of hepatic fibroblasts were initiated

* This unit is the equivalent daily mean intake of 1 gm of absolute ethanol per kg of body weight per year.

from liver biopsies. All cells were cultured under an atmosphere of 5% C02/95% air in tissue culture dishes using Eagle’s minimal essential medium containing 10% fetal calf serum modified by reducing the NaHC03 con- centration to one-half (0.013M) and adding Tricine buffer (0.025 M , pH 7.4). For growth and radiolabeling studies, cells were detached using 0.25% trypsin (17). Human fetal AF2 fibroblasts were grown as described for skin fibroblasts and used at subcultivation 17 to 21. Cells were plated at a density of 9 x 105/P-100 dish, and radiolabeling studies were performed after 6 days, as described below.

RADIOLABELING OF CELLS In preliminary experiments, we found a linear rate of

collagen and noncollagen protein production by human fibroblasts up to 6 hr; we chose 4-hr incubations for the studies. Confluent cell cultures were studied after 3 to 5 passages to eliminate environmental factors. Duplicate cell cultures were incubated in 2.5 ml of Eagle’s minimal essential medium without serum but containing 0.1 mh4 L-proline, at 37°C in 5% c02/95% air. After a 15-min preincubation period in the presence of 0.2 mM ascorbic acid, cofactor for prolyl and lysyl hydroxylases, approx- imately 25 pCi of ~-[5-~H]proline were added, and the incubation was continued for 4 hr (18). Before use, ~ 4 5 - 3H]proline was purified on a 1 x 15 cm AG50W-X8 ion- exchange column by elution of contaminants with 1.0 N HCl followed by elution of proline with 6.0 N HCl (19). The effects of ethanol or acetaldehyde on collagen me- tabolism were examined by the addition of either ethanol (50 mM) or acetaldehyde (50 p M ) to duplicate plates at the time of preincubation. In other experiments, the effects of varying concentrations of acetaldehyde (0 to 300 p M ) upon collagen production was assessed in skin fibroblasts, hepatic fibroblasts and fetal fibroblasts.

MEASUREMENT OF COLLAGEN PRODUCTION After incubation of cells, the plastic dishes were chilled

to O”C, the media were removed and the cell layers were rinsed with cold 0.11 M NaC1/0.05 M Tris (pH 7.4). Media and washes were pooled and stored at -20°C. The cell layers were scraped off the dishes with 0.11 M NaCl/ 0.05 M Tris (pH 7.4) and sonicated at 0°C for 15 sec at 20% maximum voltage in a Biosonik sonicator (needle probe) to rupture the cells (18). Trichloroacetic acid was added to yield a final concentration of 10%. Samples were centrifuged at 1000 X g for 5 min. Precipitates were resuspended and recentrifuged in 5% trichloracetic acid. The combined trichloracetic acid-soluble fraction of cell layers were stored at -20°C to determine the specific activity of proline in the free cellular pool. The pooled media and washes were added to the precipitate from cell layers, and the combined suspension was precipitated with 10% trichloroacetic acid proteins were separated by centrifugation and washed three times with 5% tri- chloroacetic acid. Combined cell layer and medium pre- cipitates were dissolved in 0.2 N NaOH. The amount of radioactive collagen was determined by the collagenase digestion method (20) as previously described (21). The collagenase-resistant material was hydrolyzed in 6 N HC1

Vol. 4, No. 5, 1984 STIMULATION OF PROTEIN PRODUCTION BY ACETALDEHYDE 845

- s I

z 20 - c ” 8 : l o : 0

4 s t 0 - w z I ”

at 120°C for 3 hr (21). The 3H-radioactivity of collagen- ase-sensitive and -insensitive proteins was used to cal- culate the relative rate of collagen production (21). The absolute rate of collagen production was determined from the ‘H-radioactivity incorporated into collagenase-sen- sitive proteins. It was expressed as disintegrations per minute per milligram of protein per hour. Protein con- tent in the 0.2 N NaOH solutions was analyzed by the dye-binding method of Bradford (22), using IgG as a standard.

P I 001

-

ETHANOL

ACETALDEHYDE

NS- Not significant

SPECIFIC RADIOACTIVITY OF THE FREE PROLINE POOL

In one experiment, the specific activities of the free L3H]proline pools were determined in skin fibroblasts incubated in the presence of acetaldehyde. Separate por- tions of the trichloroacetic acid-soluble fraction of cell sonicates were used for quantitation of proline by nin- hydrin reaction on a Durrum D-500 high-pressure Ami- noacid Analyzer (DC-6A ion-exchange resin) and for measurement of radioactivity (22). The specific radioac- tivity was then calculated as disintegrations per minute per pmole.

PROLINE TRANSPORT In order to correct for possible differences in imino

acid transport, the total radioactive proline taken up into the cells (trichloroacetic acid-soluble plus trichloroacetic acid-precipitable) was determined (24). This included the radioactivity of the free cellular pool and the radioactiv- ity of cellular and media proteins. In other experiments, the transport of imino acids was determined in confluent skin fibroblasts in the absence or presence of 50 pM acetaldehyde. Cells were incubated for 1 hr in the pres- ence of 0.2 mM ascorbic acid and 0.3 mM cycloheximide, in order to inhibit protein synthesis, with 5 pCi of [5- ‘Hlproline and 1 pCi of [2-14C]hydroxyproline. The ra- dioactivities of the acid-soluble fraction of cell layers were measured to calculate the imino acid transport.

STATISTICAL ANALYSIS All the results are expressed as mean f S.E. The

Student’s t test (paired or unpaired) was used to evaluate the differences of the means between groups, accepting p < 0.05 as significant (25).

RESULTS BASAL COLLAGEN PRODUCTION

The absolute rates of collagen production are compa- rable in control and cirrhotic groups (5,157 f 941 vs. 5,281 k 2,104 dpm per mg protein per hr; NS). To correct for possible differences in proline transport, the produc- tion of collagen was calculated as the percentage of total [3H]proline taken up into the cells which had been in- corporated into collagenase-sensitive protein. These val- ues are similar in both groups (1.9 2 0.4 vs. 1.8 & 0.6%; NS). Also, the rates of collagen production relative to total protein production are similar in both groups (data not shown).

EFFECTS OF ETHANOL OR ACETALDEHYDE ON COLLAGEN AND NONCOLLAGEN PROTEIN PRODUCTION Addition of ethanol (50 mM) has no significant effect

upon basal collagen production in the entire group (-4 k 0.8%; 15 pairs; t: 0.74, NS) or in either subgroup (Figure 1). However, addition of acetaldehyde (50 pM) increases basal collagen production by human fibroblasts in the entire group (+ 25 & 4.5%; 16 pairs; t: 3.20, p < 0.01) and to a similar extent in control and cirrhotic groups (Figure 1). Acetaldehyde also stimulates collagen production by fetal and hepatic fibroblasts. This effect of acetaldehyde upon collagen production by fibroblasts is dose-related at concentrations found in the liver during ethanol abuse (15, 16) (Table 1). The stimulation of collagen production by acetaldehyde is observed in skin fibroblasts (up to 140%), fetal fibroblasts (up to 240%) and hepatic fibroblasts (up to 70%) (Table 1). However, the effect of acetaldehyde is similar for both collagen and noncollagen proteins (Table 1). Thus, the rate of collagen production relative to total protein production is essen- tially unchanged by the addition of acetaldehyde to skin fibroblasts (5.3 & 0.4 vs. 5.4 & 0.4%; 16 pairs, NS), fetal fibroblasts, or hepatic fibroblasts (data not shown).

EFFECT OF ETHANOL OR ACETALDEHYDE ON PROLINE TRANSPORT The uptake of proline upon addition of either ethanol

(50 mM) or acetaldehyde (50 pM) is similar to baseline values in both groups. Thus, the increased collagen ra- dioactivity induced by acetaldehyde (Figure 1 and Table 1) cannot be ascribed to differences in the cellular uptake

CONTROLGROUP CIRRHOTIC GROUP ENTIRE GROUP

30 r T T T

10 NS NS NS

FIG. 1. Effect of ethanol or acetaldehyde on collagen production by human fibroblasts. Primary cultures of fibroblasts were initiated from normal skin biopsies in 16 individuals, 8 healthy (control group) and 8 cirrhotics (cirrhotic group). Cells were grown as described in “Materials and Methods.” After 3 to 5 passages, duplicate plates of confluent cell were incubated with 25 pCi of [5-3H]proline (0.1 mM) in the presence of 0.2 mM ascorbic acid alone or with the addition of either 50 mM ethanol or 50 pM acetaldehyde. Collagen production (disintegrations per minute per milligram of protein per hour) was determined from the radioactivity in collagenase-sensitive protein as described in “Materials and Methods”. Values shown are percentage change from basal collagen production. Ethanol has no significant effect on collagen production either in the entire group (-4 f 0.8%; 15 pairs; t: 0.74, NS) or in the subgroups. Acetaldehyde stimulates collagen production in the entire group (+25 & 4.5%; 16 pairs; t: 3.20, p < 0.01) and to a similar extent in both subgroups.

846 HOLT, BENNETT AND CHOJKIER HEPATOLOGY

TABLE 1. DOSE-RESPONSE EFFECT OF ACETALDEHYDE ON COLLAGEN AND NONCOLLAGEN PROTEIN PRODUCTION BY

HUMAN FIBROBLASTS

had been reached (21). The specific activity of the cellular free proline pool is not modified by the addition of acetaldehyde (Table 3). Taken together with the lack of an acetaldehyde effect on proline transport, the data

concentrationb 10-3 radio- change productionc indicate that acetaldehyde causes an increase in collagen Collagen production‘ Noncollagen

protein Experimental conditions”

( a M ) activity (dpm) (%) change (%) production. Skin fibroblasts

Experiment 1

Experiment 2

Experiment 3

Experiment 4

Fetal fibroblasts Experiment 5

Hepatic fibroblasts Experiment 6

Experiment 7

0 50

100 150

0 50

100 0

100 200

0 100 200

0 50

100

0 100

0 50

100 200

19.8 30.4 46.3 47.7 93.8

119 148 64.4 78.5 94.2

416 463 523

252 620 853

19.0 23.7

297 507 387 414

0 0 +54 +77

+134 +86 +141 +121

0 0 +27 +30 +57 +43

0 0 +22 0 +46 +25

0 0 +12 +11 +26 +21

0 0 +146 +151 +238 +232

0 0 +25 +12

0 0 +71 +122 +31 +61 .^ -- +4u +53

a Duplicate plates of confluent cells were studied in the presence of 0.2 mM ascorbic acid, and proteins were labeled with [5-3H]proline (0.1 mM) for 4 hr as described in “Materials and Methods.”

Acetaldehyde was added to the incubations at the described con- centrations 15 min prior to the addition of the isotope. ‘ Collagen and noncollagen protein production per plate were deter-

mined from the radioactivity of collagenase-sensitive and -resistant protein as described in “Materials and Methods.” Percentage change from basal values are shown. Acetaldehyde increased protein produc- tion by skin, fetal and hepatic fibroblasts (p < 0.05). There was a linear dose-response to acetaldehyde (r = 0.86, p < 0.02).

of proline (Table 2). When protein synthesis by human fibroblasts is inhibited by 99% with cycloheximide (21), the transport of proline and hydroxyproline, both of which use the same amino acid transport system, remains essentially unchanged upon addition of 50 pA4 acetalde- hyde (Table 3).

EFFECT OF ACETALDEHYDE O N THE SPECIFIC RADIOACTIVITY OF T H E PROLINE P O O L

There is much evidence to suggest that the free amino acid pool is the immediate precursor for protein synthesis and that there is rapid equilibration between amino acids in the extracellular fluid and the intracellular pool (26, 27). To assure equilibration of the medium and internal pools, unlabeled proline was added with the radioisotope. The fact that a linear rate of incorporation was attained in the preliminary experiments suggests that equilibrium

DISCUSSION In the present study, we have shown that acetaldehyde,

but not ethanol, stimulates collagen and noncollagen protein production by human fibroblasts. Acetaldehyde, at concentrations found in the liver during ethanol abuse (15,16) stimulates protein production by skin fibroblasts (Figure 1 and Table l), fetal fibroblasts and hepatic fibroblasts (Table 1). There is a variable but linear dose- response curve up to about 150 to 200 pM of acetalde- hyde. The mechanisms responsible for the variable stim- ulation of protein production, induced by acetaldehyde in cultured fibroblasts, have not been yet identified.

TABLEZEFFECT OF ETHANOL OR ACETALDEHYDE ON PROLINE UPTAKE BY HUMAN FIBROBLASTS

Experimental conditions” Proline uptakeb

Control 1,440 * 150 Ethanol (50 mM) 1,210 f 245 Acetaldehyde (50 pM) 1,370 * 263

Cells and experimental conditions are the same as those described in Figure 1.

Values from dpm of [3H]proline (trichloroacetic acid-precipitable plus trichloroacetic acid-soluble) transported into the cells per micro- gram of cellular protein in 4 hr. The difference between groups was not statistically significant.

TABLE 3. EFFECTS OF ACETALDEHYDE ON IMINO ACID TRANSPORT AND SPECIFIC ACTIVITY OF THE FREE PROLINE

P O O L I N HUMAN FIBROBLASTS Imino acid transportb

Hydroxy- specific activityC Acetaldehyde Droline (dpm/pmole) concentration” ( p M ) [‘H]proline

0 345 * 24 33 f 3 304 50 295 f 23 3 1 k 3 287

100 NDd ND 321

Acetaldehyde was added to the incubation at the described concen- trations 15 min prior to addition of the isotope.

* Confluent skin fibroblasts were incubated in the presence of 0.2 mM ascorbic acid and 0.3 mM cycloheximide with 5 pCi ~-[~H]prol ine (0.1 mM) and 1 pCi DL-[“C]hydroxyproline for 1 hr as described in “Materials and Methods”. The transport of [3H]proline and [“C] hydroxyproline, in the presence of cycloheximide, were determined from the radioactivities of the acid-soluble fraction of cell layers. Imino acid transport was expressed as x radioactivity (dpm) per plate per 60 min. Values of triplicate plates are shown. ’ Confluent skin fibroblasts from Experiment 1 of Table 1 were used

as described in “Materials and Methods.” Proline was purified by ion- exchange chromatography of the acid-soluble fraction of cell sonicates. Separate portions were used for quantitation of proline on amino acid analyzer and for measurement of radioactivity. Results of combined duplicate samples are shown. The difference between the groups were not statistically significant.

ND, not done.

Vol. 4, No. 5, 1984 STIMULATION OF PROTEIN PRODUCTION BY ACETALDEHYDE 847

Differences in the susceptibility to acetaldehyde could be due to differences in cell growth and/or cell metabolism. Whether acetaldehyde, one of its metabolites or a by- product of its metabolism is responsible for the observed stimulation remains to be determined. In any case, the lack of response to ethanol (50 mM) (Figure l), can be explained by absence of any detectable activity of alcohol dehydrogenase in these human skin fibroblasts (T.-K. Li, unpublished observations). The effect of acetaldehyde on collagen and noncollagen protein radioactivity cannot be ascribed to an increased imino acid transport (Tables 2 and 3) and in addition, the specific activity of the cellular proline precursor pool is essentially unchanged by the addition of acetaldehyde to the incubations (Table 3). Thus, the increased incorporation of [3H]proline into protein induced by acetaldehyde represents a stimulation of protein production and is not a spurious increase secondary to changes in the cellular precursor pool of proline. Consistent with our data, in the isolated perfused rat liver the uptake of a-aminoisobutyric acid, a non- metabolized amino acid that uses the imino acid trans- port system, was also unchanged by the addition of acetaldehyde or acetate (28).

Acetaldehyde does not alter the rate of collagen pro- duction relative to total proteins. Thus, the effect of acetaldehyde is not preferential for collagenous proteins and presumably affects a common protein biosynthetic pathway through a mechanism not yet identified. Bar- aona and coworkers (29) found that the absolute synthesis of hepatic protein was increased in uivo in ethanol-fed rats after correcting for the specific activity of the [“C] leucine precursor pool. Furthermore, cell-free protein synthesis with liver membrane-bound polysomes of ethanol-fed rats was increased and was probably due to an augmented level of biologically active mRNA (30). The mechanism by which ethanol induced these changes in uiuo is unknown, but they may be related to our present in uitro finding of stimulation of collagen and noncolla- gen protein production by acetaldehyde.

Lieber and coworkers (3,5,6) have demonstrated that ethanol causes hepatic damage and that alcoholic liver fibrosis may develop without significant necroinflam- matory reaction (5, 6). Several mechanisms have been postulated to explain the hepatic collagen deposition of chronic alcohol abuse (31-34) but they remain contro- versial (11-14). Chen and Leevy (7) found that ethanol increased the incorporation of radioactive proline into collagen by liver tissue of patients with alcoholic hepa- titis and cirrhosis. However, contamination of collagen by radioactive noncollagen protein may have occurred (35) as indicated by a very low percentage of proline hydroxylation in cirrhosis (28%) and alcoholic hepatitis (15%) (as calculated from the data in Ref. 7). Collagen production by cultured fibroblasts (8, 9), no longer ex- posed to their in uivo environment, was either unchanged (8) or decreased (9) upon addition of ethanol. Although these findings suggest that ethanol itself does not stim- ulate collagen production (8, 9), a direct role of acetal- dehyde on collagen metabolism has not been previously considered (11-14). The newly recognized stimulation of collagen production by acetaldehyde may play an impor-

tant role in the hepatic fibrogenesis of alcoholic liver disease independent of a necroinflammatory reaction. A recent study (Savolainen, E.-R. et al., Hepatology 1983; 3:801, Abstract) found that acetaldehyde also increased collagen synthesis of cultured baboon liver myofibro- blasts without change in proline pools in agreement with our findings.

In this study, we have also shown that the rates of collagen production by skin fibroblasts of alcoholic pa- tients with cirrhosis or without liver disease are quite similar (see “Results” and Figure 1). These results differ somewhat from those reported by other investigators (7, 8, 36).

Production of collagen (7, 8, 36) and other connective tissue proteins (8) have been determined in hepatic tissue obtained from patients with alcoholic liver disease. An increased relative rate of collagen production was re- ported by Mann and coworkers (36) in liver biopsies of patients with cirrhosis. However, the use of [4-3H]proline for labeling proteins introduces another variable. Since 69% of the tritium is at the 4-trans position (18) and it is displaced upon proline hydroxylation, collagen radio- activity will be higher if peptidyl proline remains under- hydroxylated, independent of the rate of collagen pro- duction. Similarly, Chen and Leevy (7) found that biopsy specimens of patients with alcoholic hepatitis, and cir- rhosis had an increase in radioactive proline and hydrox- yproline, but contamination of collagen by radioactive noncollagen proteins may have occurred (35). Nonethe- less, the marked increase of salt-soluble radioactive hy- droxyproline in cirrhosis is a provocative finding. It should be noted, however, that contamination of this fraction is a possible source of error in the analysis of hydroxyproline (19). Lastly, these investigators (7, 36) did not attempt to eliminate environmental factors and therefore their studies may not be comparable to our study.

Galambos and coworkers (8) studied cultured fibro- blasts derived from liver biopsies, and no longer exposed to the direct effect of their in uiuo environment. These investigators found an increased production of collagen by fibroblasts of patients with alcoholic hepatitis and cirrhosis. These data suggest a genetic influence since environmental factors had probably been eliminated after several cell subcultivations. The discrepancy be- tween this study and ours could be explained in several ways. It is conceivable that the genetic defect in collagen production, if it exists, is not expressed in skin fibro- blasts. In our study, cirrhotic patients were matched to control patients by ethnic background, age and sex. All these variables which may influence collagen production by fibroblasts (37), were not considered by Galambos and coworkers (8). Furthermore, in the previous studies (7, 8, 36), neither the proline transport nor the specific activity of the free proline pool were determined. To the extent that fibroblasts are responsible for the hepatic fibrogenesis in uiuo, our results would suggest that any genetic influence on the individual susceptibility to de- velop alcoholic cirrhosis is not mediated through an increased collagen production per cell.

In conclusion, we have shown that acetaldehyde, at

848 HOLT, BENNETT AND CHOJKIER HEPATOLOGY

concentrations found in the liver during ethanol abuse, stimulates collagen and noncollagen protein production by human fibroblasts. These findings may provide new insights into the mechanisms of alcoholic liver fibrosis.

Acknowledgments: We are grateful to Dr. T.-K. Li (Department of Medicine, Indiana School of Medicine, Indianapolis, Ind.) for performing the alcohol dehydrog- enase assays, to Dr. K. Hostetler for his critical review of this manuscript and to Dr. M. Schuckit for allowing us to include two of his patients in the study. We thank Gary Deming for his skillful preparation of this manu- scrint.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

-x- -.

REFERENCES Lelbach WK. Quantitative aspects of drinking in alcoholic liver cirrhosis. In: Khanna JM, Israel Y, Kalant H, eds. Alcoholic liver pathology. Toronto: Addiction Research Foundation of Ontario,

Takeuchi J , Takada A. Alcohol and its hepatotoxic effect. In: Khanna JM, Israel Y, Kalant H, eds. Alcoholic liver pathology. Toronto: Addiction Research Foundation of Ontario, 1975 199- 224. Rubin E, Lieber CS. Fatty liver, alcoholic hepatitis and cirhosis produced by alcohol in primates. N Engl J Med 1974; 290128-135. Hrubec Z, Omenn GS. Evidence of genetic predisposition to alco- holic cirrhosis and psychosis: twin concordances for alcoholism and its biological end points by zygosity among male veterans. Alcoholism: Clinical and Experimental Research 1981; 5:207-214. Popper H, Lieber CS. Histogenesis of alcoholic fibrosis and cirrho- sis in the baboon. Am J Pathol 1980; 98695-716. Nakano M, Worner TM, Lieber CS. Perivenular fibrosis in alco- holic liver injury: ultrastructure and histologic progression. Gastro- enterology 1982; 83:777-785. Chen TSN, Leevy CM. Collagen biosynthesis in liver disease of the alcoholic. J Lab Clin Med 1975; 85:103-112. Galambos JT, Hollingsworth MA Jr., Falek A, et al. The rate of synthesis of glycosaminoglycans and collagen by fibroblasts cul- tured from adult human liver biopsies. J Clin Invest 1977; 60:107- 114. Thanassi NM, Rokowski RJ, Sheehy J , et al. Non-selective de- crease of collagen synthesis by cultured fetal lung fibroblasts after non-lethal doses of ethanol. Biochem Pharmacol 1980; 29:2417- 2424. Lindros KO, Eriksson CJP. The role of acetaldehyde in the actions of ethanol. Helsinki: The Finnish Foundation for Alcohol Studies, 1975. Lieher CS. Medical disorders of alcoholism. Pathogenesis and treatment. Philadelphia: W. B. Saunders Co, 1982. Orrego H, Israel Y, Blendis LM. Alcoholic liver disease: informa- tion in search of knowledge? Hepatology 1981; 1:267-283. Mezey E. Alcoholic liver disease. In: Popper H, Schaffner F, eds. Progress in liver diseases, Vol VII. New York Grune and Stratton, 1982: 555-572. Popper H, Martin GR. Fibrosis of the liver: the role of the ectoske- leton. In: Popper H, Schaffner F, eds. Progress in liver diseases, Vol VII. New York: Grune and Stratton, 1982: 133-156. Eriksson CJP, Sippel HW. The distribution and metabolism of acetaldehyde in rats during ethanol oxidation-I. Biochem Phar- macol 1977: 26:241-247.

1975: 1-18.

17. Peterkofsky B. The effect of ascorbic acid on collagen polypeptide synthesis and proline hydroxylation during the growth of cultured fibroblasts. Arch Biochem Biophys 1972; 152:318-328.

18. Chojkier M, Peterkofsky B, Bateman J. A new method for deter- mining the extent of proline hydroxylation by measuring changes in the ratio of [4-3H]:[14C]proline in collagenase digests. Anal Biochem 1980; 108385-393.

19. Chojkier M, Bateman J , Phang JM, et al. Formation of proline metabolites in chick embryo bone: interference with the measure- ment of free hydroxyproline by ion-exchange chromatography. Anal Biochem 1982; 120:330-338.

20. Peterkofsky B, Diegelmann R. Use of a mixture of proteinase-free collagenases for the specific assay of radioactive collagen in the presence of other proteins. Biochemistry 1971; 10:988-994.

21. Peterkofsky B, Chojkier M, Bateman J. Determination of collagen synthesis in tissue and cell culture systems. In: Furthmayr H, ed. Immunochemistry of the extracellular matrix, Vol 11. Boca Raton, Florida: CRC Press, Inc., 1982; 19-47.

22. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein- dye binding. Anal Biochem 1976; 72:248-254.

23. Chojkier M, Spanheimer R, Peterkofsky B. Specifically decreased collagen biosynthesis in scurvy dissociated from an effect on pro- line hydroxylation and correlated with body weight loss. J Clin Invest 1983; 72:826-835.

24. Eichhorn JH, Peterkofsky B. Local anesthetic-induced inhibition of collagen secretion in cultured cells under conditions where microtubules are not depolymerized by these agents. J Cell Biol

25. Colton T. Inference on means. In: Colton T, ed. Statistics in medicine. Boston: Little Brown and Co., 1974: 99-150.

26. Loftfield RB, Harris A. Participation of free amino acids in protein synthesis. J Biol Chem 1956; 219: 151-159.

27. Fern EB, Garlick PJ. The specific radioactivity of the tissue free amino acid pool as a basis for measuring the rate of protein synthesis in the rat in uiuo. Biochem J 1974; 142:413-419.

28. Piccirillo VJ, Chambers JW. Inhibition of hepatic uptake of alpha aminoisobutyric acid by ethanol: effects of pyrazole and metabo- lites of ethanol. Res Commun Chem Pathol Pharmacol 1976; 13:297-320.

29. Baraona E, Leo MA, Borowsky SA, et al. Pathogenesis of alcohol- induced accumulation of protein in the liver. J Clin Invest 1977;

30. Zern MA, Chakraborty PR, Ruiz-Opazo N, et al. Development and use of a rat albumin cDNA clone to evaluate the effect of chronic ethanol administration on hepatic protein synthesis. Hepatology

31. Kershenobich D, Fierro FJ, Rojkind M. The relationship between free pool of proline and collagen content in human liver cirrhosis. J Clin Invest 1970; 49:2246-2249.

32. McGee JO'D, O'Hare RP, Patrick RS. Stimulation of the collagen biosynthesis pathway by factors isolated from experimentally- injured liver. Nature (New Biology) 1973; 243:121-123.

33. Comstock JP, Udenfriend S. Effect of lactate on collagen proline hydroxylase activity in cultured L-929 fibroblasts. Proc Natl Acad Sci USA 1970; 66552-557.

34. Maruyama K, Feinman L, Fainsilber Z, et al. Mammalian colla- genase increases in early alcoholic liver disease and decreases with cirrhosis. Life Sci 1982; 301379-1384.

35. Perez-Tamayo R. Is cirrhosis of the liver experimentally produced by CC1, an adequate model of human cirrhosis? Hepatology 1983; 31 12-120.

36. Mann SW. Fuller GC. Rodil JV. et al. Heaatic arolvl hvdroxvlase

1979; 81:26-42.

60:546-554.

1983; 3:317-322.

16. Lindros KO. Regulatory factors in hepatic acetaldehyde metabo-

The role of acetaldehyde in the actions of ethanol. Helsinki: The Finnish Foundation for Alcohol Studies, 1975: 67-81.

and collagen synthesis in patients with alcbholi; liver dkease: Gut

37. Prockop DJ, Kivirikko KI, Tuderman L, et al. The biosynthesis of lism during ethanol oxidation. In: Lindros KO, Eriksson CJP, eds. 1979; 20825-832.

collagen and its disorders. N Engl J Med 1979; 301:77-85.