Stimulation of Collagen Gene Expression by Ascorbic Acid in Cultured ...

THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 266. No. 28, Issue of October 5, pp. 16957-16962,1989 Printed in U.S. A.

Stimulation of Collagen Gene Expression by Ascorbic Acid in Cultured Human Fibrcoblasts A ROLE FOR LIPID PEROXIDATION?* (Received for publication, April 27, 1989)

Mario ChojkierS, Karl Houglum, Jose Solis-Herruzog, and David A. Brennerl From the Llepartment of Medicine, Veterans Administration Medical Center and University of California, San Diego, California 92161

Ascorbic acid stimulates collagen gene expression in cultural fibroblasts (Lyons, B. L., and Schwartz, R. L. (1984) Nucleic Aci& Res. 12, 2569-2579), but the mechanisms responsible forthis effect are poorly understood. In the presence of the transitional metal iron, ascorbic acid could induce lipid peroxidation with the formation of reactive aldehydes. Since another aldehyde, acetaldehyde, the first metabolite of ethanol, also stimulates collagen transcription in cultured fibroblasts (Brenner, D. A., and Chojkier, M. (1987) J. Biol. Chem 262,17690-17696),we investigated whether ascorbic acid induces lipid peroxidation in cultured cells and if this is the mechanism by which ascorbic acid stimulates collagen gene expression. Ascorbic acid (0.2 mM) induced lipid peroxidation in cultured human fibroblasts judging by the production of thiobarbituric acid-reactive substances and carbonyl groups, and by the presence of malondialdehyde- and 4-hydroxynonenal-1~roteinadducts. Ascorbic acid stimulated (2-3-fold) the net production of collagen relative to total proteins,the levels of procollagen al(I) mRNA and the transcription of this gene. Inhibition of the ascorbic acid-induced lipid peroxidation in cultured human fibroblasts with a-tocopherol (50 PM) or ) the stimulation of methylene blue (10 i c ~ prevented collagen gene expression. The addition of malondialdehyde (200 PM), a product of lipid peroxidation, to cultured human fibroblasts also increased 2-3-fold collagen production and procollagen al(1) mRNA levels. Thus, ascorbic acid induces lipid peroxidation and reactive aldehydes andl this stepmay be necessary for the stimulation of collagen gene expression by ascorbic acid in cultured human fibroblasts.

Ascorbic acid has a well documented role in collagen metabolism as a direct requirement for prolyl (1)and lysyl (2) hydroxylases. It is required for hydroxylation of peptidylproline incell culture systelrns (3,4),unless a microsomal reducing cofactor is present ( 5 ) .Ascorbic acid also stimulates collagen production in cultured chick tendon (6, 7) and human skin (8, 9) fibroblasts. Ascorbic acid increases procollagen mRNA * This study was supported inpart by United States Public Health Service Grants DK39996,DK07202, and DK38652 and by grants from the Veterans Administration. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Recipient of a Research Career Development Award, Veterans Administration. § Supported in part by the Del Amo Foundation, Spain. li Pew Scholar in the Biomedical Sciences.

levels in cultured fibroblasts (9, IO), and this effect is the result of the stimulation of procollagen gene transcription and thedecreased degradation of the procollagen mRNA (10). The mechanism responsible for the stimulation of collagen gene expression by ascorbic acid in culturedcells is unknown, but it is not related to the role of ascorbic acid on collagen hydroxylation (8,9,11). It has been suggested that procollagen secretion is the rate-controling step in the ascorbic acid induction of collagen gene expression (12). In thisreport, we analyze the mechanisms by which ascorbic acid regulates collagen gene expression in cultured cells. Our work is derived from the reported increase in numerous aldehydes resulting from the peroxidation of polyunsaturated fatty acids that is induced by ascorbic acid in vitro in the presence of the transitional metal iron (13), and from recent observations linking another aldehyde, acetaldehyde, the first metabolite of ethanol, with selective stimulation of collagen production (14) and collagen gene transcription (15) in cultured human fibroblasts. Since cultured cells accumulate a sizeable amount of iron (16),we wondered whether the addition of ascorbic acid would induce lipid peroxidation with the formation of reactive aldehydes, and if this is the mechanism by which ascorbic acid stimulates collagen gene expression. In this study, we demonstrate that ascorbic acid induces lipid peroxidation and reactive aldehydes and that this step maybe necessary for the stimulation of TypeI collagen synthesis and collagen a(1) gene transcription by ascorbic acid in cultured human fibroblasts. EXPERIMENTALPROCEDURES

Muterials-~-[5-~H]Proline(22 mCi/mmol) and aqueous counting scintillant fluid were purchased from Amersham Corp. Chromatographically purified bacterial collagenase form I11 was obtained from Advance Biofactures Company (Lynbrook, NY). (a-32P]2’-Deoxycytidine 5”triphosphate was purchased from ICN. Deoxyribonucleotides, Klenow fragment of DNA polymerase I, bovine serum albumin, and dextran sulfate were purchased from Pharmacia LKB Biotechnology Inc. Agarose and restriction endonucleases were purchased from Bethesda Research Laboratories, AG-50W-X8 resin (100-200 mesh) and low melting agarose were from Bio-Rad. CsCl wasobtained from IBI (New Haven, CT). Byodine transfer membrane was purchased from Pall (Glen Cove, NY)and malondialdehyde from Aldrich. Guanidine isothiocyanate was obtained from Eastman. L-Ascorbic acid, a-tocopherol succinate, N-ethylmaleimide, thiobarbituric acid, 2,4-dinitrophenylhydrazine,and ethidium bromide were obtained from Sigma. Sephadex ‘2-50 prespun columns were obtained from Boehringer Mannheim. Agar, tryptone, and yeast extract were obtained from Difco. Vectastin kits and reagents were from Vector (Burlingame, CAI. Guinea pig antibodies against specific malondialdehyde- and 4-hydroxynonenal-lysine protein adducts were generously provided by J. Witztum (University of California, San Diego). Human fetal AF, fibroblasts and the media used for their in uitro

16957

16958

Ascorbic Acid, Lipid Peroxidation, and

Collagen Gene Expression

culturing were obtained from the Core Cell Culture Facility (Univer- was eluted from a C18 column with 11%methanol (1 ml/min) after sity of California, San Diego, CA). 5 min and measured at 546 nm. Fibroblast Cultures-Human fetal AF, fibroblasts wereused at Lipid peroxidation was also assessed in the cell layer by measuring subcultivations 5-21 as previously described (14, 17). Cells were the 2,4-dinitrophenylhydrazinederivates of aldehydes formed form cultured under an atmosphere of 5% CO,, 95% air in tissue culture the degradation of unsaturated fatty acids. Cells were harvested and dishes using Eagle's minimal essential medium (MEM)' containing washed with phosphate saline buffer a t 0 "C, and reacted with 2,410% fetal calf serum. Cells were plated at a density of 9 X 106/P-10 dinitrophenylhydrazine. The derivates were separated by TLC as dish, and radiolabeling studies were performed after 6 days as de- described by Benedetti and co-workers (35), but omitting the silicic acid preparative chromatography. Polar and moderately polar fracscribed below. Radiolabeling of Cells-Confluent cell cultures were incubated in tions were combined and analyzed separately from the nonpolar 10 ml of MEM with 10% fetalcalf serum containing 0.1 mM L-proline, carbonyl derivatives, by measuring their absorption at 365 nm. Values a t 37 "C in 5% CO,; 95% air. The effects of lipid peroxidation on were expressed as nanomoles of carbonyls compounds produced using collagen metabolism were examined by addition of ascorbic acid (0.2 the extinction coefficient of 25,500 (35). Immunohistochemistry-Cells were grown to 60-80% confluency mM) to plates at thetime of preincubation (0-72 h). In some experiments FeSO, (0.1 mM) was added to minimize variability in media on glass slides or in tissue culture slides (Lab-tek) and incubated for fetal calf serum iron. Also, the influence of antioxidants, a-tocopherol 20 b in the presence of ascorbic acid (0.2mM) alone or with the succinate (5-50 p ~ ) and , dimethyl sulfoxide (25 mM), as well as addition ofFeSO, (0.1 mM) and in the absence or presence of amethylene blue (10 p ~ )a, scavenger of reducing equivalents (18), on tocopherol (50 pM). Cells werefixed for less than 20 min in 4% the ascorbic acid-treated cells was assessed by the addition of these paraformaldehyde, 20 p~ BHT, 2 mM EDTA, and 5% (w/v) sucrose. compounds to the preincubation at the time of addition of ascorbic Guinea pig antibodies against specific malondialdehyde- and 4-hyacid. After the preincubation period, 0.2 mM ascorbic acid (a cofactor droxynonenal-lysine protein adducts were utilized as primary antifor prolyl and lysyl hydroxylases) and 740kBqof ~-[5-~H]prolinebodies to identify these modified proteins as evidence of in uiuo lipid were added, and the incubation was continued for up to 4 hr (14). peroxidation (36). The primary antibodies were used in conjunction The effects of exogenous malondialdehyde on collagen production with biotinylated goat anti-guinea pigIgG and the ABC-alkaline was assessed by adding malondialdehyde (200 p ~ in) MEM without phosphatase systems essentially as recommended by the manufacfetal calf serum to thecell cultures at thetime of the incubation with turer. Nonspecific antibody reactions were assessed by omitting the the radiolabeled proline. Labeling of cells was terminated by cooling first antibody. Statistical Analysis-All the results are expressed as mean f S.E. the plates to 0 "C. Measurement of Collagen Production-Collagen and noncollagen unless stated otherwise. The Student's t test or analysis of variance protein production in combined cell layer and medium were deter- were used to evaluate the differences of the means between groups, acceptingp < 0.05 as significant (37). mined by the collagenase method (19), as previously described (2022). The radioactivity of collagenase-sensitive and -insensitive proteins was used to calculate the relative rate of collagen production. RESULTS Eualuution of Cellular Tonicity-Loss of cellular lactic dehydrogenAscorbic Acid Induction of Lipid Peroxidation in Cultured ase was measured in the supernatant of cell incubations. The lactic dehydrogenase assay was carried out by a Cobas-Bio Analyzer. En- Cells-We assessed whether ascorbic acid induces lipid perzyme release was normalized from experiment to experiment by oxidation in cultured human fibroblasts. Incubation of cells expressing the data as percent of total cellular lactic dehydrogenase with ascorbic acid induced the production of TBARS (Table (23) as described previously (21). Intact cell counts were determined I), anindex of lipid peroxidation. This was determined by the using a Coulter counter. Preparation of cDNA Probes-The plasmid pHF677 (24) contain- thiobarbituric acid assay using either fluorometric detection ing the cDNA for the human al(1) collagen was provided by Dr. F. or HPLC identification of the malondialdehyde-thiobarbituRamirez. The plasmid pKaI containing the cDNA for human a- ric acid complex. The addition of iron, but not ascorbic acid, tubulin was provided by Dr. N. Cowan (25). The plasmid pcllase 1 at thetime of cell harvest led to a spurious increase in TBARS containing the cDNA for human fibroblast collagenase was provided (data not shown). To assess lipid peroxidation in the presence by Dr. P. Angel (26). The plasmid p k b 3 . u ~containing the human c- of additional iron in the preincubation period, the 2,4-dinitrofos cDNA (27) was provided by Dr. I. Verma. Using the random phenylhydrazone derivatives of aldehydes formed from the primer synthesis method (28), the cDNA fragments were radiolabeled with [cP~*P]~CTPa specific to activity of approximately 1X lo9cpm/ degradation of unsaturatedfatty acids were measured. A significant increase in nonpolar derivatives was found (6.4 f pg of DNA. Northern Blotting and Slot Blotting-AF, cells were incubated as 2 nmol/plate; p < 0.05) in cells preincubated with ascorbic described above with ascorbic acid or malondialdehyde (200 p ~ )with , acid/FeS04. Similarly, there was an increase in polar and or without a-tocopherol (5-50 p M ) or methylene blue (10 pM). Cells moderately polar carbonyl compounds. In addition, in cells were processed as previously described (15), and RNA was isolated incubated with ascorbic acid we detected a marked increase by the guanidine thiocyanate/phenol/chloroformextraction method (29). Total RNA was usedfor Northern and slot blotting as previously in malondialdehyde-protein adducts (Fig. 1B)and 4-hydroxdescribed (15, 30).The slot blot autoradiograms were quantitated by ynonenal-protein adducts (data not shown) by immunohisa scanning laser densitometer interfaced with an integrator. The level tochemistry utilizing specific antibodies against these aldeof collagen nl(1) mRNA in each sample was normalized to the level hyde-protein adducts. Sections incubated in the absence of of a-tubulin mRNA. the first antibody did not demonstrate any specific staining. Run-off Transcription-The nuclear run-off transcription assay Therefore, both ascorbic acid alone and ascorbic acid/FeS04 which was developed by Groudine et al. (31) and adapted by Wang and Calame (32) was used as previously described (15). The average TABLE I experiment produced 1 cpm of radiolabeled RNA per nucleus. Effects of ascorbic acid on lipid peroxidation Determination of Lipid Peroxidation-Lipid peroxidation was determined by measuring thiobarbituric acid-reacting substances Experimental condition" TBARS (TBARS) in the cell layer as described by Ohkawa and co-workers nmol (33). TBARS were measured fluorometrically (excitation 515 nm, Control 0.29 & 0.04 emission 553 nm) using malondialdehyde and tetramethoxypropane Ascorbic acid 1.01 f 0.08 standards. To some plates, FeS04 (0.1 mM) and/or ascorbic acid (0.2 Ascorbic acid/u-tocopherol 0.33 f 0.02 mM) wereadded before harvesting to assess their influence on TBARS formation in the assay conditions. The thiobarbituric acid-malondiCells wereincubated for 20 h inMEM, 10% fetal calf serum alone aldehyde complex was also determined by HPLC (34). This complex (control), or with the addition of ascorbic acid (0.2 mM) or ascorbic acid, a-tocopherol (50 pM). * Thiobarbituric acid-reactive substances were determined in cell 'The abbreviations used are: MEM, Eagle's minimal essential medium; TBARS, thiobarbituric acid-reacting substances; HPLC, layers as described under "Experimental Procedures." Values are means f S.E. of triplicate samples; p < 0.05. high performance liquid chromatography.

Ascorbic Acid, Lipid Peroxidation, and Collagen GeneExpression

16959

or absence of ascorbic acid for 20 h. Cells were then labeled for an additional 4-h period with[3H]proliie in thepresence of ascorbic acid, a cofactor for prolyland lysyl hydroxylases. The net production of collagen was increased by ascorbic acid, while noncollagen protein production remained unchanged (Fig. 2). In agreement with previous studies (7-9,11,12), the stimulatory effect of ascorbic acid on collagensynthesis was observed when cells wereincubated with ascorbic acid for24 h and persisted for up to 72 h. A similar induction of collagen production was achieved when ascorbic acid with FeSOr was added to the medium, but addition of FeS04without ascorbic acid had no effect on collagen production (Table 11). We found no evidence of cellular toxicity in AF, cells treated with ascorbic acid for 24 h; the release of lactic dehydrogenase, intact cell counts, and noncollagen protein production were similar to that of control cells. If the products of lipid peroxidation are responsible forthe stimulation of collagen production in these cultured cells, then antioxidants should prevent this effect. As expected, coincubation of cells with ascorbic acid and a-tocopherol, an inhibitor of lipid peroxidation (38), blocked the production of TBARS (Table I) and malondialdehyde-protein adducts (Fig. IC)and 4-hydroxynonenal-protein adducts (data not shown). Also, a-tocopherol blocked the ascorbic acid stimulation of collagen production(Table 11).Similarly, dimethyl sulfoxide, a scavenger of hydroxyl radicals (391, prevented the stimulation of collagen production induced by ascorbic acid(data not shown). We have previously shownthat theincrease in both collagen production and procollagen mRNA levels induced by acetaldehydecouldbeblocked by methylene blue (15), a scavenger of reducing equivalents suchas NADH (18).Those experiments suggested that reducing equivalents, which enhance the formation of aldehyde adducts as well as the proportion of stable adducts, may be required for the effect. Similarly, methylene blue prevented the stimulation of collagen gene expression induced by ascorbicacid (Table 11),and this effect was associated with inhibition of lipid peroxidation (TBARS, -40%; p < 0.05). Since malondialdehyde is a product of lipid peroxidation we studied its effects on collagen production.The addition of malondialdehyde (200 PM),to thecultured human fibroblasts

r

FIG. 1. Effects of ascorbic acid on malondlaldehyde-protein adducts. Cells were grown on glass slides and incubated for 20 h with MEM, 10% fetal calf serum alone (control, A ) or with the addition of ascorbic acid (0.2 mM) ( B ) or ascorbic acid and CYtocopherol (50 p ~ (C). ) Immunohistochemistry was performed as described under “Experimental Procedures.”

induced measurable lipid peroxidation in cultured A F 2 cells. Ascorbic Acid-inducedStimulation of Collagen GeneExpression. Role of Lipid Peroxidation-Cultured human fibroblasts were incubated inMEM, 10%fetal calf serum inthe presence

AA

CONTROL

FIG.2. Effects of ascorbic acid on collagen production in cultured human fibroblasts. Cell cultures were preincubated for 20 h with or without ascorbic acid (0.2 mM) and labeled, in the presence of 0.2 mM ascorbic acid, for 4 h a t 37 “C with 740 kBq of L[5-3H]proline.The collagen and noncollagen protein production was calculated from the 3H radioactivities in collagenase-sensitive and -resistant proteins. Values are mean S.E. of quadruplet samples. p < 0.05 for collagen.

*

Ascorbic Acid, Lipid Peroxidation, and

16960

Collagen Gene Expression TABLE I1

Effects of ascorbic acid on collagen production Relative rate of collagen production'

Noncollagen proteinh

Collagenb

Experimental conditions"

dpm

96

X

Experiment 1 Control 3.0 f 0.1 95.0 f 4.1 100 f 5 Ascorbic acid/Fe*+ 5.7 f 0.2 103.7 f 5.0 173 f 2 Fez+ 3.5 f 0.2 107.0 f 4.4 103 f 2 Experiment 2 122.4 f 1.8 100 f 2 Control 9.3 f 0.3 109.5 f 9.8 151 f 11 Ascorbic acid 12.9 f 0.2 90.5 f 3.4 103 f 2 Ascorbic acid/a-tocopherol 7.1 f 0.3 Experiment 3 222.1 f 13.1 100 f 13 Control 5.2 f 0.9 269.3 k 26.0 196 f 17 Ascorbic acid 11.9 f 0.7 Ascorbic acid/methylene blue 3.1 f 0.3 183.0 f 20.4 74 f 4 a Confluent cell cultures were incubated for 20 h in MEM, 10% fetal calf serum, and labeled in the presence of 0.2 mM ascorbic acid for 4 h as described in Fig. 2. Experiment 1 , no additions (control)or with the addition of 0.2 mM ascorbic acid/FeSO, (0.1 mM) or 0.1 mM FeS04;Experiment 2, no additions (control) or with the addition of ; 3, no additions (control)or with the ascorbic acid (0.2 mM) or ascorbic acid and a-tocopherol (50 p ~ )Experiment addition of ascorbic acid (0.2 mM), or ascorbic acid and methylene blue (10 pM). 'Determined from the 'H radioactivity incorporated into collagenase-sensitive and -resistant proteins after labeling cell cultures with 740 kBq of [5-'H]proline for 4 h. Values are means f S.E. of a t least triplicate samples. p < 0.05 for collagen (ascorbic acid/Fe*+ (Experiment 1) and ascorbic acid (Experiments 2 and 3)). Calculated from the formula ['H]collagen dpm/([3H]noncollagen dpm X 5.4 + ['H]collagen dpm) and expressed as percentage of control values. The S.E. reported is the S.E. of the individual samples in each group.

incubatedin MEM withoutfetal calf serum (to minimize malondialdehyde binding to media proteins), increased 2-fold the production of collagen without affectingnoncollagen protein production (Table 111). The concentration of malondialdehyde in the media was 80% of the initial concentration after a 3-h incubation. The step at which ascorbic acid (and lipid peroxidation) acts tostimulate collagen synthesis inthese cultured cells was TABLE 111 Malondialdehyde-inducedstimulation of collagen production Experimental conditions"

Collagenb

Noncollagen proteinb

&Iative rate of collagen production'

dpm x

96

Control 9.7 f 0.6 529.0 f 33.2 100 f 6.0 178 f 16.0 Malondialdehyde 18.3 f 1.7 547.6 f 15.2 Confluent cell cultures were incubated in MEM, with or without 200 p~ malondialdehyde, and labeled in the presence of 0.2 mM ascorbic acid and 740 kBq of [5-3H]proline for 3 h. Determined as described in Table 11. Values are mean f S.E. of triplicate samples; p < 0.05 for collagen. Calculated as described in Table 11.

studied. Total RNA was extracted from human fibroblasts incubated in the presence of 0.2 mM ascorbic acid for 24 h or malondialdehyde for 3 h. The fibroblast RNAs were analyzed in Northern blots and slot blots, usinghuman a tubulin, collagenase, and procollagen al(1) DNA probes. The latter was selected because collagen Type I is the major collagen expressed inthese cell cultures (30). The Northern blots revealed that theRNAs were intact and thatneither ascorbic acid nor malondialdehyde alter the size of the procollagen al(1) (Fig. 3), a-tubulin,or collagenase (data not shown) transcripts. The two transcripts of the procollagen al(1) (4.8 and 5.8 kilobases in length) have been described previously by others andrepresent alternativechoices of polyadenylation sites (40). We found a 2-3-fold increase in procollagen al(1) mRNA, without changes in a-tubulin mRNA, in cells incubated with either ascorbic acid (Fig. 3, lane 2) or malondialdehyde (data not shown). Co-incubation of cells with atocopherol or methylene blue prevented the increase in procollagen mRNA induced by ascorbic acid (Fig. 3, lanes 3 and 4 ) . Incubation of cells with ascorbic acid/Fe2+also increased the level of collagen mRNA (Fig. 3, lane 5), and this effect was blocked when cells were co-incubated with a-tocopherol (Fig. 3, lane 6 ) .

PROCOLLAGEN

5.5 kb

m

~i

1

Q1

2

3

4

5

6

FIG.3. Northern blot analysis of human fibroblast collagen mRNA. Confluent human fibroblasts were incubated for 24 h in MEM, 10% fetal calf serum alone ( l a n e 1 ) with the addition of ascorbic acid (0.2 mM) ( l a n e 2 ) ; ascorbic acid and a-tocopherol (50 p ~ ( l)a n e 3 ) ;ascorbic acid and methylene blue (10 p M ) ( l a n e 4 ) ; ascorbic acid (0.2 mM), FeS04 (0.1 mM) ( l a n e 5);or ascorbic acid, FeS04, anda-tocopherol (50 pM) ( l a n e 6).10 pg of total RNA wereelectrophoresed on a formaldehyde 1%agarose gel and transferred to a nylon filter by capillary blotting. The filters were hybridized to radiolabeled human collagen al(1)cDNA.

Ascorbic Acid, Lipid Peroxidation, and Collagen Gene Expression

16961

that the increased collagen synthesis and gene expression induced bv ascorbic acid are related to the stimulation of 4 / procollagen secretion. For example, a reduction of hydroxylation in procollagen of cultured chick embryo bone cells inhibited secretion of procollagen by 80% without affecting its synthesis (44). Also, we have shown previously that shortterm impairment of proline hydroxylation in vitro with (Y,cY'dipyridyl, an inhibitor of proline hydroxylation, did not affect collagen synthesis in guinea pig bone (49) or in AF2 human fibroblasts (21) cultures. Furthermore,the stimulation of procollagen mRNA induced by ascorbic acid in chick tendon fibroblasts was not prevented by a,d-dipyridyl (50). Analogously, inhibition of procollagen secretion without affecting procollagen hydroxylation in different culture systems did not result in a block in collagen synthesis (51,52). Our study provides an alternative explanation for the role FIG. 4. Run-off transcriptional assay. Nuclei were isolated of ascorbic acid in collagen gene expression in cultured cells. from control (open bars) and ascorbic acid (0.2 mM)/FeS04 (0.1 mM) We found that ascorbic acid induced lipid peroxidation in AF2 (closed bars)-treated human fibroblasts incubated as described for human fibroblasts judging by the production of TBARS and Fig. 2. Purified radiolabeled nascent RNAs (2 X 10' cpm) from each carbonyl compounds and by the presence of malondialdehydesample were hybridized t.0 1p g of each plasmid DNA immobilized on nitrocellulose filters. The autoradiograms were quantitated by scan- and 4-hydroxynonenal-proteinadducts. Since the previous (6-12) and present experiments used fetal calf serum in the ning laser densitometry. incubations, these cells would accumulate a sizeable amount In order to determ:ine the effects of ascorbic acid on the of iron (16), which upon the addition of ascorbic acid could relative rates of transcription of collagen and noncollagen induce lipid peroxidation (13). a-Tocopherol, a lipophilic antioxidant, and other antioxiprotein genes, we performed transcription run-off assays. Results of a transcription run-off assay using nuclei isolated dants prevented lipid peroxidation as well as the stimulation from control fibroblasts and ascorbic acid-treated fibroblasts of collagen gene expression and collagen production by ascorare shown in Fig. 4. Ascorbic acid increased procollagen al(1) bic acid in culturedfibroblasts. In addition, exogenous malontranscriptional activit.y about 3-fold.Ascorbic acid had no dialdehyde, a product of lipid peroxidation, induced similar effects on collagen gene expression inthesehuman fibroeffect on c-fos or collagenase gene transcription. blasts. Collectively, these data strongly suggest that products DISCUSSION of lipid peroxidation mediate the ascorbic acid-induced stimThis study analyzes the mechanisms by which ascorbic acid ulation of collagen gene expression. The mechanisms by which ascorbic acid and exogenous stimulates collagen gene expression in cultured cells. The addition of ascorbic acid induced an increase incollagen gene malondialdehyde stimulate collagen gene transcription are transcription, collagen mRNA levels, and collagen production unknown. We suspect that they are similar to those responsible for the acetaldehyde stimulation of collagen gene tranin cultured AF2 human fibroblasts. These findings are in agreement with those]previously reported by other investiga- scription (15). Perhaps, protein adducts formed with aldetors (6-12). Understanding how ascorbic acid stimulates col- hydes derived from peroxidation of unsaturated fatty acids, lagen synthesis may provide insights into the regulation of such as 4-hydroxynonenal and malondialdehyde, may play a collagen gene expressi(on.There are only three other factors role in the increased transcription of the collagen gene. For known to stimulate collagen gene transcription. These factors example, the formation of aldehyde adducts with a transactiare acetaldehyde ( E ) , a fibrogenic factor purified from rat vating protein may alter itsbinding to thecollagen gene. We liver (41), and TGF@under some conditions (42) but not found 2-%fold more adducts of malondialdehyde in nuclear others (43). proteins extracted from cells incubated with ascorbic acid Ascorbic acid stimulates the rate of procollagen secretion than incontrol cells.2Adducts of aldehydes and proteinshave in cell culturesystems (8, 44) because of its role inthe been shown to occur in vivo inatherosclerotic lesions of formation of hydroxyproline which provides stability to the hyperlipidemic rabbits (53), in the livers of rats fed ethanol procollagen triple helix: (45). It has been proposed that accu- (54) or treated with the hepatotoxin carbon tetra~hloride,~ mulation of nonhelical, underhydroxylated procollagen in the and in theliver and plasma proteins of iron-overloaded rats.' rough endoplasmic reticulum in cells incubated without as- However, the significance of these adducts in gene regulation, corbic acid (46) may inlhibitcollagen production and that this if any, remainsto be determined. Alternatively, a direct effect block may be relieved by ascorbic acid (12). Since the per- of products of lipid peroxidation on either the collagen gene centage of collagen degradation is essentially the same incells or cytosolic regulatory pathways of collagen synthesis is posincubated inthe presence or absence of ascorbic acid (11,471, sible. the ascorbic acid-induced stimulation of collagen production It is not clear whether this modulation of collagen synthesis is a consequence of increased synthesis. Although addition of by ascorbic acid in cultured cells occurs also in uioo. A differamino-procollagen peptides selectively inhibits thesynthesis ent hypothesis has been suggested to explain the specific of procollagen by fibroblasts in culture (48), it is not known decrease in collagen polypeptide synthesis in scorbutic guinea whether asimilar effect could be induced by these propeptides pigs (49). Ascorbic acid deficiency leads to decreased food within the endoplasmic reticulum while still part of the prointake and associated weight loss, which either directly or collagen molecule. A dissociation betweten the extent of procollagen hydroxM. Filip and M. Chojkier, unpublished observations. ylation and the ratesof collagen synthesis has been observed K. Houglum and M. Chojkier, unpublished observations. in various cell culture systems,and therefore it seems unlikely ' K. Houglum and M. Chojkier, unpublished observations.

16962

Ascorbic Acid, Lipid Peroxidation, and Collagen Gene Expression

through another regulatory mechanism inhibits collagen gene 24. Chu, M.-L., Myers, J. C., Bernard, M. P., Ding, J.-F., and Ramirez, F. (1982)Nucleic Acids Res. 10,5925-5933 expression (49,55). 25. Cowan, N. J., ,Dobner, P. R., Fuchs, E. V., and Cleveland, D.W. (1983) Mol. Cell. Bwl. 3. 1738-1745 Lipid peroxidation is associated with the tissue injury and P., Baumann, I., Stein, B., Delius, H., Rahmsdorf, H. J.,and fibrogenesis of several pathological disorders, including ath- 26. Angel, Herrlich, P. (1987)Mol. Cell. Bid. 7 , 2256-2266 27. Mitchell, R. L., Zokas, L., Schreiber, R. D., and Verma, I. M. (1985)Cell erosclerosis (53), iron overload (56), porphyria (57), and 40.2119-21 ,-.- - -7. ethanol (58)-,bleomycin (59)-,and carbon tetrachloride (60)- 28. Feinberg, A. P., and Vogelstein, B. (1983)Anal. Biochem. 132,6-13 induced toxicity. The mechanisms by which tissue injury is 29. Chomczynski, P., and Sacchi, N. (1987)Anal. Bkchem. 162,156-159 30. Solis-Herruzo, J. A,, Brenner, D. A., and Chojkier, M. (1988)J. BioL Chem. sometimes followedby fibrogenesis in uiuo are unknown. 263,584-5845 M., Peretz, M., and Weintraub,H. (1981)Mol. Cell. Biol. 3,281Whether they are related to the stimulation of collagen gene 31. Groudine, 9RR expression induced by products of lipid peroxidation remains 32. Wa;, X.-F., and Calame, K. (1985)Cell 43,659-665 33. Ohkawa, H Ohishi N. and Yagi, K. (1979)Anal. Biochem. 96,351-358 to be established. 34. Acknowledgments-We thank Michael Filip, Martina Buck, and Linda Veloz for their excellent technical assistance and Kris Beaver for her skillful preparation of this manuscript. REFERENCES

1. Peterkofsky, B., and Udenfriend, S. (1965)Proc. Natl. Acad. Sci. U. S. A. 63,335-342 2. Kivirikko, K. L., and Prockop, D. J. (1967)Proc. Natl. Acad. Sci. U. S. A. 67,782-789 3. Peterkofsky, B. (1972)Arch. Biochem. Biophys. 162,318-328 4. Bates, C. J., Prynne, C. J., and Levene, C. I. (1972)Bzochim Bkphys. Acta 263,397405 5. Mata, J. M., Assad, R., and Peterkofsky, B. (1981)Arch. Biochem. Biophys. 206,93-104 6. Schwarz, R. I., and Bissell, M. J. (1977)Proc. Natl. Acad. Sci. U. S. A. 74, 4453-4457 7. Schwarz, R. I., Farson, D. A., Soo, W.-J., and Bissell, M. J. (1978)J. Cell BioL 79,672-679 8. Murad, S., Grove, D., Lindberg, K. A., Reynolds, G., Sivarajah, A., and Pinnell, S. R. (1981)Proc. Natl. Acad. Sci. U. S. A. 78,2879-2882 9. Takehara, K., Grotendorst, G. R., Trojanowska, M., andLeroy, E. C. (1986) CoUagen Relat. Res. 6,455-466 10. Lyons, B. L., and Schwarz, R. I. (1984)Nucleic Acids Res. 12,2569-2579 11. Schwarz, R. I., Mandell, R. B., and Bissell, M. J. (1981)Mol. Cell. Bwl. 1, 843-853 12. Schwarz, R. I. (1985)J . Biol. Chem. 260,3045-3049 13. Minotti, G., and Aust, S. D. (1986)J. Biol. Chem. 262,1098-1104 14. Holt, K., Bennett, M., and Chojkier, M. (1984)Hepatology 4,843-848 15. Brenner, D. A., and Chojkier, M. (1987)J. Bwl. Chem. 262,17690-17695 16. Schraufstatter, I., Hyslop, P. A., Jackson, J. H., and Cochrane, C. G. (1988) J. Clin. Invest. 82, 1040-1050 17. Flaherty, M., and Chojkier, M. (1984)Anal. Biochem. 142,386-394 18. Ryle, R. P., Chakrahorty, J., and Thomson,A. D. (1985)Biochem. J. 232, 877-882 19. Peterkofsky, B., and Diegelmann, R. (1971)Biochemistry 10,988-994 20. Chojkler, M., Peterkofsky, B., and Bateman,J. (1980)Anal. Biochem. 108, 385-393 21. Flaherty, M., and Chojkier, M. (1986)J. Biol. Chem. 261,12060-12065 22. Peterkofsky, B., Chojkier, M., and Bateman,J. (1982)in Immunochemistry of the Extracellular Matrix (Furthymayr, H., ed) Vol. 2,pp. 19-47,CRC Press Inc., Boca Ratnn, FL 23. Quissell, D. O.,Barzen, K. A., and Lafferty, J. L. (1981)Am. J. Physiol. 241,C76-C85

Bird, R. P.,';nd Driper: H. H. (1984)Methods Entymol. 106,299-305 35. Benedetti, A., Pompella, A., Fulceri, R., Romani, A., and Comporti, M. (1986)Biochim. Bio hys Acta 876,658-666 36. Palinski, W., RosenfePd, M . E., Yla-Herttuala, S., Gurtner, G. C., Socher, S.S., Butler, S. W., Parthasarathy, S., Carew, T. E., and Steinberg, D. (1989)Proc. Natl. Acad. Sci. U. S. A. 86,1372-1376 37. Colton, T. (1974)Statistics in Medicine, pp. 99-150,Little, Brown and Co., Boston 38. Ames,B. N., Cathcart, R., Schwiers, E., and Hochstein, P. (1981)Proc. Natl. Acad. Sci. U. S . A. 78,6858-6862 39. Granger, N. D., Hernandez, L. A., and Grisham, M. B. (1986)Viewpoints Dig. Dis. 18,13-16 40. Chu, M.-L, de Wet, W., Bernard, M., and Ramirez, F. (1985)J. Biol. Chem. 260,2315-2320 41. Ra how,R., Gossage, D., Seyer, J. M., and Kang, A. H. (1984)J. Biol. &em. 269,12718-12723 42. Rossi P., Karsenty G., Roberts, A. B., Roche, N. S., Sporn, M. B., and de Crdmbrugghe, B.'(1988) Cell 62,405-414 43. Ra how, R., Postlethwaite, A. E., Keski-Oja, J., Moses, H. L., and Kang, H. (1987)J. Clin Invest. 79,1285-1288 44. Blanck, T. J. J., and Peterkofsky, B. (1975)Arch. Biochem. Biophys. 171, 259-267 45. Berg, R. A., and Prockop, D. J. (1973)Bbchem. Biophys. Res. Commun. 62, 115-120 46. Pacifici, M., and Iozzo, R. V. (1988)J. Biol. Chem. 263,2483-2492 47. Geesin, J., Murad, S., and Pinnell, S. R. (1986)Biochim Biophys. Acta 886,272-274 48. Wiestner, M., Krieg, T., Horlein, D., Glanville, R. W., Fietzek, P.,and Muller, P. K. (1979)J. Bkl. Chem. 254, 7016-7023 49. Chojkier, M., Spanheimer, R., and Peterkofsky, B. (1983)J. Clin. Invest. 72,826-835 50. Rowe, L. B., and Schwarz, R. I. (1983)Mol. Cell. Biol. 3,241-249 51. Diegelmann, R. F., and Peterkofsky, B. (1972)Proc. Natl. Acad. Sci. U. S. A. 69,892-896 52. Eichhorn, J. H., and Peterkofsky, B. (1979)J . Cell Biol. 81,26-42 53. Haberland, M. E., Fon D , and Cheng, L. (1988)Science 241,215-218 54. Lin, R. C.,Smith, R. and Lumeng, L. (1988)J. Clin. Invest. 81, 615619 55. Spanheimer, R. G., and Peterkofsky, B. (1985)J. Biol. Chem. 260,39553962 56. Bacon, B. R., Tavill, A. S., Brittenham, G. M., Park, C. H., and Recknagel, R. 0.(1983)J. Clin. Invest. 71,429-439 57. Poh-Fit atrick, M.B. (1982)Semin. Liver Dis. 2,164-176 58. Muller, and Sies, H.(1982)Biochem. J. 206,153-156 59. Clark, J. G., Kostal, K. M., and Marino, B. A. (1982)J. Biol. Chem. 267, 8098-8105 60. Burk, R. F., Lane, J. M., and Patel, K. (1984)J. Ctin. Inuest. 74, 19962001

1.

8:.

x,