JOURNAL OF CELLULAR PHYSlOLOGY 152:27&280 (1992)

Induction of Fibronectin Gene Expression by Transforming Growth Factor Beta-1 Is Attenuated in Bronchial Epithelial Cells by ADP-Ribosyltransferase Inhibitors JOE D. BECKMA”,* M A R Y ILLIG, DEBRA ROMBERGER, AND STEPHEN I. R E N N A R D Section of Pulmonary and Critical Care, Department of lnternal Medicine, Un~versityof Nebraska Medical Center, Omaha, Nehraska 68 798-2465 Transforming growth factor-beta (TGF-P) exerts several effects on cultured airway epithelial cells including inhibition of proliferation and stimulation of fibronectin gene expression. ADP-ribosylation is one potential regulatory mechanism of gene expression by TGF-P. We tested this possibility by exposing cultured bovine bronchial epithelial cells to the chemical inhibitor of ADP-ribosyl transferase enzymes, 3-aminobenramide (3-AB) and, for comparison, 3-aminobenzoic acid (3-ABA), which i s structurally similar to 3-A6 but which does not inhibit ADPribosyl transferases. Exponential cell growth rate ( I .2 doublings/day) or cellular morphology observed by phase contrast microscopy were not affected by 3 m M 3-A6 or 3-ABA. Neither compound antagonized inhibition of cell division or induction of squamous morphology by TGF-P1. In contrast, the sixfold stimulation of fibronectin production by exposure of cells to 30 pM TCF-PI for 48 h was reduced by 50% in the presence of 3 mM 3-AB, whereas 3 mM 3-ABA had no effect. The antagonistic effect was augmented by administration of 3-AB 24 h prior to induction by TGF-Pl . Northern blot hybridization analyses demonstrated that 3-AB, but not 3-ABA, attenuated the induction of fibronectin mRNA by TGF-P1 by up to 50%. The5e observations may implicate il role of cellular ADP-ribosylation in the regulation of some gene expression by TGF-P. o I Y Y ~Wiky ~ 1 5 5 ,Inc.

The family of transforming growth factors beta (TGF-P) have profound regulatory effects on cell growth and differentiation (Pelton and Moses, 1990). TGF-P inhibits airway epithelial cell growth (Masui et al., 1986), and we have shown (see below) that TGF-P1 also induces fibronectin production by bronchial epithelial cells in culture (Romberger et al., 1989). Similar inductive effects of TGF-P are observed on mesenchyma1 cells (Penttinen et al., 1988), which are growth stimulated by this factor (Roberts et al., 1985). The effects of TGF-P on epithelial vs. fibroblast cells are consistent with the histologic (Lazenby et al., 1990) and biochemical (Khalil et al., 1989) changes observed in post-inflammatory lung fibrosis. Notably, the production of extracellular matrix proteins increases dramatically in response to bleomycin (Khalil et al., 1989), and of these proteins, fibronectin is a major component that may be temporally expressed before types I and I11 collagens (Hoyt and Lazo, 1988; Kelley et al., 1985; Raghow et al., 1985). Elegant immunohistological evidence (Khalil et al., 1989) revealed the spatiotemporal expression of TGF-P in bleomycin-induced pulmonary fibrosis; the initial expression of this cytokine was localized to the bronchiolar epithelium, followed by macrophages. Therefore, it seems quite likely that the airway epithelium may be an initial target in the progression of fibrosis. Current evidence indicates that TGF-$ may mediate, at least in part, both the benefi0 1992 WILEY-LISS. INC.

cia1 effects of wound repair and the deleterious effects of fibrotic reactions (Khalil et al., 1989; Phan and Kunkel, 1992; Raghow et al., 1989) and overt squamous metaplasia (Masui et al., 1986).In this regard, controlling the effects of TGF-p on cells of the respiratory system, and perhaps other locations, appears potentially beneficial. There are many reports describing attempts to alter the phlogistic and fibrotic responses of the lung t o agents such as bleomycin (e.g., Haschek et al., 1989; Riley et al., 1982; Wang et al., 1989; Ward et al., 1988). Niacin and 3-aminobenzamide (3-AB) were recently used t o prevent bleomycin-induced pulmonary fibrosis in a hamster model (Giri et al., 1989; Wang et al., 1990a,b).The rationale for this effect rested in the diminution of NAD levels due to DNA damage and subsequent polyADP-ribosyl transferase (pADPRT) activation (Hussain et al., 1985). Niacin and 3-AB, a competitive pADPRT inhibitor, may antagonize this drop of NAD concentration, thereby allowing normal metabolic flux through NAD-dependent dehydrogenases to continue. The ability of pADPRT inhibitors to

Received December 23,1991; accepted March 6,1992.

*To whom reprint requests/correspondence should be addressed.

ADPRT INHIBITION ANTACXlNIZES TGF-BI

attenuate a cellular response to pro-fibrotic cytokines has not, however, been tested. We have examined this possibility in this investigation by exposing cultured bovine bronchial epithelial cells to nicotinamide analogues, which are either inhibitory to ADP-ribosyl transferase or not inhibitory (Purnell and Whish, 1980). We report that 3-aminobenzamide antagonizes the induction of fibronectin protein and mRNA levels by TGF-P1 in cultured bronchial epithelial cells. However, 3-aminobenzoic acid, which is not a n inhibitor of ADP-ribosyl transferases, does not have these effects.

MATERIALS AND METHODS Cell culture Procedures, media, reagents, and supplements for the serum-free culture of bovine bronchial epithelial cells were as previously described (Beckmann et al., 1992). For all experiments reported here, second or third passage cells were cultured in a 1 : l mixture of LHC-9 and RPMI 1640 media (Beckmann et al., 1992). LHC-9, developed for culture of human bronchial epithelial cells (Lechner and LaVeck, 19851, contains the following supplements: insulin (5 pgiml), hydrocortisone (0.2 pM), EGF (5 ngiml), transferrin (10 pg/ml), phosphoiethanolamines (0.5 pM each 1, dl-epinephrine (0.5 pg/ml), retinoic acid (0.33 nM), triiodothyronine (10 nM), trace elements, and bovine pituitary extract (50 pg proteiniml). The concentrations of these supplements in the final medium are reduced by 50% due to mixing with non-supplemented RPMI 1640. Media were changed every 1-3 days a s indicated. The nicotinamide analogues 3-aminobenzamide (3-AB) and 3-aminobenzoic acid (3-ABA) obtained from Sigma (St. Louis, MO) were dissolved in culture medium at 20 mM prior to sterile filtration and dilution with fresh medium to concentrations as shown. Porcine platelet transforming growth factor p l (TGF-pl) was purchased from R & D Systems (Minneapolis, MN). Bovine bronchial fibroblasts were obtained as outgrowths onto tissue culture plastic from explants (ca. 1 cm2)cultured in DMEMi10% fetal calf serum at 37°C in a humidified atmosphere of 95% air/5% CO,. These cells dispersed by 0.05% trypsin were seeded onto 35 mm wells, cultured for 3 days in DMEM/10% serum until about 50% confluent, incubated overnight in the serum-free LHC-9IRPMI 1640 medium (see above), and finally fed with fresh LHC-9iRPMI media containing & 30 pM TGF-B and +- 4 mM 3-aminobenzamide. After 48 h , supernatant media and cells were harvested for fibronectin quantitation (see below). Cell growth and morphology measurements Three approaches were used t o quantitate cell proliferation. First, cells were detached with 0.05% trypsin and manually counted with a hemacytometer. This approach was used for measurement of cell growth rates, which were determined by dividing the slopes of semilog plots by 0.301 (log 2) to give population doublings per day. Second, culture dishes were washed once with Hepes buffered saline, fixed 2 min with methanol, stained 2 min with eosin, stained 2 min with methylene blue, and then washed with water (Leukostat reagents; Fisher Scientific). The overall surface area occupied by cells on the entire dish was then quantitated using a n

275

Optomax image analyzer equipped with a video camera and 35 mm lens (Beckmann et al., 1992). Third, mitotic indices (% cells in mitosis) were determined by inspection of cultures by phase contrast microscopy. At least six random fields containing 50-250 cells each were manually counted, with cells containing distinct chromatin bands being counted as mitotes. Cell planar areas were determined as previously described (Beckmann et al., 1992) using a n Optomax image analyzer. Occupied surface areas in six random fields under 2 0 magnification ~ were divided by nuclei counts for each field.

Quantitation of fibronectin protein We have previously reported the use of a competitive inhibition enzyme linked immunosorbant assay for bovine fibronectin (Romberger e t al., 1989). The quantity of antigen is then divided by time elapsed since the previous change of medium and also divided by cell number as determined by hemacytometer. Results are therefore reported as ng fibronectinlcellih. All assays were performed in triplicate. Northern and dot-blot hybridization analyses Total RNA was extracted and purified from bronchial epithelial cells on culture dishes using a previously described method (Chomczynski and Sacchi, 1987). All samples had 260 nmi280 n m absorbance ratios of 1.9-2.0 and were quantitated using standard Warburg coefficients a s a component of the Beckman DU-62 spectrophotometer software. RNA samples (10-20 pg) were denatured with formaldehyde/ formamide (Maniatis et al., 19821, made to contain 30 pg/ml ethidium bromide (EtBr), and subjected to electrophoresis on 0.8% (wiv) agarose gels containing 6.6% formaldehyde in 40 mM MOPSilO mM Na-acetateil mM EDTAipH 7.0 buffer (Maniatis et al., 1982). After Polaroid photography of the gel using ultraviolet transillumination, the RNA was directly transferred to nitrocellulose (BA-85; Schleicher and Schuell, Keene, NH) by capillary flow using 20x SSPE as the reservoir. ( I x SSPE contains 0.18 M NaClilO mM NaPJ1 mM EDTAipH 7.6.) Alternatively, formaldehyde denatured RNA samples ( 2 . 5 4 pg) were dot-blotted to nitrocellulose using a S & S Minifold I1 apparatus. Blots were baked 2 h a t 80°C in vacuo and stored at room temperature until use. Purified restriction enzyme fragments of the mouse p5 tubulin (Sullivan and Cleveland, 1986) and human F H l l l fibronectin (Kornblihtt et al., 1983) cDNAs were 32P-labeled in low melting agarose by the random hexamer priming method (Feinberg and Vogelstein, 1984). After 70-90% incorporation of radiolabeled nucleotide was achieved, 10 volumes of water were added, the samples heated for 5 min at 95"C, mixed with hybridization fluid a t 50°C, and then added to nitrocellulose membranes which had been blocked. Blocking solution included 5 x SSPEi5 x Denhardt's (Denhardt, 1966111% SDS, and was used at 42°C. Hybridization fluid contained 5x SSPE/l x Denhardt's/5% dextran sulfateiO.l% SDS/50%formamide and was used at 42°C for 16-24 h with continuous agitation. Preliminary experiments established the specificities of the probes after washing conditions of 0 . 1 SSPEiO.l% ~ SDS at 42°C.

BECKMANN ET AL

276 7.0

I

I1

lo3mM3-AB 45

'

A

3mM3-ABA

; + ;

I

5

7

Days Fig. 1. Bronchial epithelial cell growth rates are not affected by 3-aminobenzamide (3-AB) and 3-aminobenzoic acid (3-ABA). On day 0,50,000 cells were plated per 35 mm dish. Media were changed on day 2 (arrow)to include no analogue (closed circles), 3 mM 3-aminobenzamide (open circles), or 3 mM 3-aminobenzoic acid (closed triangles). The initial growth rate under all conditions was 1.2 population doublings per day.

Washed blots were exposed to Kodak XAR-5 film a t -80°C using one DuPont Cronex Lightning Plus intensifying screen. Northern blot signals were quantitated using a GS300 scanning densitometer (Hoefer Scientific Instruments, San Francisco, CA) interfaced to a n IBM-PC. Relative dot-blot signals were measured using a n Image Analyzer with 256 grey shade resolution which was calibrated with a n autorad containing a series of known relative signals.

Statistical analyses Unpaired two-tailed student's t-test was used to make comparisons, which were taken as significant when P < 0.05. RESULTS Effects of analogues on cell growth and morphology Prior to testing the effects of nicotinamide analogues on bronchial epithelial cell responses to TGF-P1, it was first necessary to determine if the chemicals had any undesired effects on cellular growth and morphology. Concentrations of analogues of up to 5 mM were chosen based on published values found useful in other investigations (Exley et al.. 1987; Hunting et al., 1985; Milam and Cleaver, 1984; Nishio et al., 1983; Porteous et al., 1979). In initial experiments, two day exposures of up to 5 mM of either 3-aminobenzamide (3-AB) or 3-aminobenzoic acid (3-ABA) had no effects on cell proliferation as measured by a clonal growth assay (Beckmann et al., 1992) (not shown). Secondly, continuous exposures to either 3-AB or 3-ABA did not significantly impair exponential cell growth rates of 1.2 population doublings per day (Fig. 1).Thirdly, inspection of the cultures by phase contrast microscopy did not reveal any overt morphological changes of the normal cobblestone appearance in response to the analogues (Fig. 2). Although these results indicate a n absence of overt cy-

totoxic effects of 3-AB or 3-ABA, we cannot eliminate the possibility of low level toxic effects not detectable by these controls. Therefore, subsequent results should be interpreted with caution. TGF-p1 inhibits bronchial epithelial cell division and induces a squamous phenotype (Jetten et al., 1986; Masui et al., 1986; Beckmann et al., 1992). Therefore, 3-AB and 3-ABA have been tested for their abilities to alter these cellular responses to TGF-P1. Culture mitotic index decreased from 2.58 0.38% (control & SEM) to 0.56 2 0.30% in response t o 20 pM TGF-P1 for 48 h. Similarly, in the presence of 3 mM 3-AB or 3-ABA, the mitotic indices decreased from 2.74 +0.29% to 0.38 F 0.20%, o r 2.62 0.32% to 0.53 2 0.3976, respectively. Hence, the analogues did not affect the ability of TGF-p1 to inhibit cell growth. Morphological analysis (Beckmann et al., 1992) revealed that 48 h exposure to 20 pM TGF-p1 increased the average cell planar area from 498 k 66 pm2 to 1550 5 320 pm'. Inclusion o f 3 mM 3-AB or 3-ABA resulted in increases from 534 t 53 pm2 to 1020 2 110 pm', or from 700 k 125 pm2 to 1420 ? 310 pm2, respectively. Although this indicates a n approximate 50% attenuation of the induction of the squamous morphology by 3-AB, this response did not reach statistical significance (P = 0.14 by t-test).

*

*

Effects of analogues on fibronectin induction by TGF-(31 Exposure to 30 pM TGF-P1 for 48 h resulted in the expected (Romberger et al., 1989) increase of fibronectin released into the culture medium (Fig. 3). Although a n increase was also observed in the presence of 3 mM 3-AB, the observed induction was reduced by 60%. In contrast to 3-AB, 3-ABA did not antagonize this induction. Northern blot hybridization analyses (Fig. 4) have revealed comparable antagonism of cognate transcript levels (53% reduction of fibronectin RNA induction compared to control or 3-ABA exposure). These results closely parallel the observed decreased fibronectin protein release (Fig. 3), and thereby suggest t h a t 3-AB may be acting at a pre-translational level which affects the induction of fibronectin gene expression by TGF-p. Time-course experiments measuring the release of fibronectin after TGF-P1 induction were performed (not shown). Maximal rate of secretion was observed a t 48 h post-TGF-pl exposure in the absence or presence of 3-AB. Therefore, the analogue does not merely delay the response to TGF-p1. To test for possible cell specificity of the above effects, bovine bronchial fibroblasts were also examined (see Methods). Exposure to 30 pM TGF-P1 for 48 h increased the fibronectin released into the medium (1.28 to ngicellih, P < 0.05). Inclusion of 3-ami3.88 x nobenzamide (4 mM) reduced the fibronectin secretion by 44% (2.16 x lop3ngicellih, P < 0.05). Therefore, the ability of 3-AB to antagonize the induction of fibronectin production by TGF-P may not be limited to epithelial cells. Dose responses for 3-AB attenuation As shown in Figure 5, 2 mM or 4 mM 3-AB nearly equally antagonized the induction of fibronectin release at up to 30 pM TGF-PI. At 40 pM TGF-P1,4 mM

AUPRT INHIBITION ANTAGONIZES TGF-PI

277

Fig. 2. Nicotinamide analogues do not affect bronchial epithelial cell morphology as observed by phase microscopy. These representative photomicrographs were taken after two day exposures to 3 mM 3-aminobenzamide (middle)or 3 mM 3-aminobenzoic acid (right).The width of each photomicrograph is 240 Pm.

Control 15 1

10

-

3-AB

3-ABA

i

bronectin response on both the protein and mRNA levels (Fig. 6). There was significant decrease of protein levels in the cultures not exposed to exogenous TGF-P1. We speculate that this effect may be the result of autocrine TGF-@production by these cells, which 3-AB may also attenuate.

DISCUSSION These studies demonstrate that a reported inhibitor 5of nuclear poly(ADP)-ribosyl transferase, 3-AB, can antagonize the ability of TGF-Pl to induce bronchial epithelial cell fibronectin protein secretion and cognate steady state mRNA levels. In contrast to 3-aminobenzamide, 3-aminobenzoic acid is a poor inhibitor of PADPRT activity (Purnell and Whish, 1980). The lack of a n 4. + + effect of 3-aminobenzoic acid on fibronectin protein or 30 pM TGF-I31 RNA induction by TGF-f3 (Figs. 3, 4)lends credence to Fig. 3. The stimulation of bronchial epithelial cell fibronectin re- the possible role of ADP-ribosylation a t some point in lease by TGF-B1 is antagonized by 3-aminobenzamide (3-AB), but not the regulation of TGF-Pl signal transduction regulat3-aminobenzoic acid (3-ABAj. Cells were pre-exposed to the analogues ing fibronectin gene expression in bronchial epithelial (3 mM) for 24 h prior to 48 h co-exposure with 30 pM TGF-B1. Supercells. This does not hold, however, for the growth inhinatant media were then assayed by ELISA for fibronectin, and cell counts determined by trypsinization and hemacytometer. Symbol bition by the cytokine, and i t therefore seems possible sizes encompass the SEM of averaged triplicates unless shown other- that two signal pathways may exist. Evidence for both wise. G protein-dependent and -independent signal pathways for TGF-P binding to fibroblasts has been reported (Howe et al., 1989). Reports on the ability of niacin (nicotinic acid) and 3-AB was slightly more antagonistic than 2 mM analogue. The level of fibronectin induction shown in Fig- 3-aminobenzamide to antagonize bleomycin induced ure 5 was less than shown in Figure 3, which was prob- pulmonary fibrosis (Giri et al., 1989; Wang et al., ably due to variation of culture confluence. We have 1990a) are relevant to this investigation. Bleomycin observed that fibronectin induction by TGF-P decreases can induce DNA damage and subsequently cause large as cell density increases in our culture system (results increases in nuclear poly(ADPRT) activity. Adminisnot shown). Although coadministration of the analogue tration of niacin presumably helps prevent consequent with TGF-P1 reduced fibronectin release by about 40- large drops of cellular NAD levels, which would be del50%,pre-exposure of the cells for 24 h to 3-AB increased eterious due to decreased NAD-dependent dehydrogenase activities. In addition, however, it is now known the inhibition to 60-80%. Variation of the 3-AB concentration prior to TGF-P1 that bleomycin increases the production of TGF-P first (30 pM) exposure revealed that millimolar concentra- by bronchiolar epithelial cells, followed by pulmonary tions of analogue were required to attenuate the fi- macrophage production (Khalil et al., 1990). Because

/-

BECKMANN ET AL

278 EtBr

Autorads

1 -

I

Control

3-AB

rn

+

TGFB. -

3-ABA

n +

+--

-

1

Control

3-AB

+ --

-

3-ABA

n +

-r-+

- FN !8 S

-

-

ias-

cDNA (upperright panel) or mouse mp5 tubulin cDNA (lowerright panel) probes. The autoradiographs are aligned with the ethidium bromide (EtBr) stained gel (left panel) to allow direct size comparison. Equivalent RNA loads in each lane are indicated by the intensity of the ethidium bromide fluorescence and by the intensities of the tuhulin bands.

Fig. 4. The increase of bronchial epithelial cell fibronectin mRNA levels in response to TGF-Pl is antagonized by 3-aminobenzamide (3-AB), but not 3-aminobenzoic acid (3-ABA). At approximately 70% confluency, cell cultures were co-exposed for 48 h to 4 mM 3-AB or 3-ABA ? 30 pM TGF-pl. Total RNAs were extracted, 15 kg RNA electrophoresed (Materials and Methods), the gel capillary blotted to nitrocellulose, and the blot hybridized to 32P-labeled fibronectin

Co-exposure with 3-AB

.-

c

A

TUB

Pre-exposure to

3-AB

6

c)

5

0 -

aJ

=: E

g '0

4 Y

3

f

= 2

a a , a m 21 o c a m O c

l

0

'

~

10

'

20

TGF-I31

i

'

l

30

(PM)

.

1

40

I

0

10

20

30

40

TGF-I31 (pM)

Fig. 5. Inhibition of fibronectin induction at varied doses of TGF-pl by 3-aminobenzamide (3-AB). Cultures were either incubated simultaneously with TGF-p1 and 3-aminobenzamide (left panel) or preincuhated with analogue for 24 h before co-exposure with TGF-p1 (right

panel). Closed circles are 0 mM 3-AB, open circles contained 2 mM 3-AB, and closed triangles included 4 mM 3-AB. After 48 h, supernatant media were harvested for fibronectin quantitation by ELISA, and cell counts were determined.

this cytokine is probably a n important mediator for inducing the increased deposition of fibrotic matrix, niacin or nicotinamide therapy could also be protective by virtue of antagonizing cellular responses to TGF-6, as shown in this investigation. Although our data have been obtained from bronchial epithelial cells, which may play little role in post-inflammatory pulmonary

fibrosis, we have also observed that 3-AB antagonizes the induction of fibronectin release from bronchial fibroblasts by TGF-f31. Therefore, this effect may not be cell type specific. What is the mechanism by which 3-aminobenzamide attenuates the cellular response to TGF-P? Although TGF-P proteins are known t o bind to specific cell sur-

ADPRT INHIBITION ANTAGONIZES TGF-p1

w* T

-00

- 2

- 1

0

log (mM 3-Aminobenzamide) Fig. 6. Effect of 3-aminobenzamide concentration on induction of secreted fibronectin (top panel) and fibronectin mRNA level (lower panel) by TGF-p1. Cultures at about 50% conflucncc were pre-incubated with the indicated 3-AB concentrations prior to co-exposure with (closed circles) or without (open circles) 30 pM TGF-Pl for 48 h. Asterisks indicate significant changes (P< 0.05) compared to values at 0 mM 3-AB (-m).

face receptors (Massague and Like, 1985), a precise mechanism of signal transduction for TGF-P is currently unclear (Fafeur et al., 1991). However, recent investigations have demonstrated that the pro-a2(1) collagen (Rossi et al., 1988) and fibronectin (Dean, 1989) genes are responsive, a t least in part, to TGF-p stimulation through the action of a transcription activator, NF-1, which comprises a family of genetically related proteins (Gil et al., 1988; Santoro et al., 1988). Interestingly, the DNA sequence which binds NF-1 can also specifically bind histone H 1 (Ristiniemi and Oikarinen, 1989). which is generally considered to be a repressor molecule (Weintraub, 1985). Therefore, H 1 and NF-1 proteins will likely compete for these regulatory DNA sequences. Since H1 is known to be ADPribosylated (Levy-Wilson, 1983), and since this covalent modification decreases the activities of many nuclear proteins (Gaal and Pearson, 19851, i t follows that nicotinamide analogues could increase the binding of repressive H1 to NF-1 consensus DNA sequences. Such a n effect could result in a suppression of some cellular genetic responses to TGF-p. The above model assumes that the results of this investigation are due to nuclear effects. Indeed, there is evidence from several systems to suggest that nuclear ADP-ribosylation plays a role in cell differentiation (Caplan et al., 1979; Exley e t al., 1987; Nishio et al., 1983; Porteous e t al., 1979; Williams and Johnstone, 1983). However, our results did not reveal attenuation of growth inhibition or induction of a squamous phenotype by TGF-P. It is important to note that nicotinamide and chemical analogues may also inhibit the cytoplasmic (Moss et al., 1980) and membrane associated

279

(De Wolf et al., 1981) mono(ADPRT) enzymes. The roles of these latter enzymes are not entirely clear, although evidence indicates their abilities to regulate adenylate cyclase (Inageda et al., 1991). It has also been suggested that the cytosolic mono(ADPRT1 may covalently modify histones prior to migration into the nucleus (Gaal and Pearson, 1985). These additional possibilities could be consistent with our results. Distinguishing the relative importance of the nuclear vs. cytosolic ADPRT enzymes in mediating the effects in this investigation could possibly be accomplished through the use of more specific ADPRT inhibitors or competitive alternative substrates, such as m-iodobenzylguanidine (Loesberg et al., 1990). In conclusion, we have shown that a chemical ADPRT inhibitor can antagonize the inductive effect of TGF-pl on bronchial epithelial cell fibronectin gene expression. It will be interesting to determine if this holds for cell types which are key players in fibrotic diseases, and if the induction of other extracellular matrix genes by TGF-p is similarly affected. This effect may also become important as therapeutic metabolic modulation of oxidant-induced cytotoxicity is explored (Berger, 1991), or in the use of ADPRT inhibitors as potentiating chemotherapeutic agents (Gaal and Pearson, 1985). Finally, these results are consistent with a working model for the mechanism of TGF-P which provides direction for several further lines of investigation.

ACKNOWLEDGMENTS We thank Lorene Claassen for help in preparation of primary epithelial cells and quantitation of fibronectin by ELISA, and Mr. Gregory Prorok for preparation of bronchial fibroblasts. Dr. Stanley Cox generously allowed the use of his scanning densitometer. We are grateful to Drs. Austin B. Thompson and Joseph Sisson for reviewing the manuscript. Apology is offered for the omission of many important citations due to spatial constraints. This research was supported in part by the Larsen Endowment. LITERATURE CITED Beckmann, J.D., Takizawa, H., Romberger, D., Illig, M., Claassen, L., Rickard, K., and Rennard, S.I. (1992) Serum free culture offractionated bovine bronchial epithelial cells. In Vitro Cell. Dev. Biol., 28At39-46. Berger, N.A. (1991) Oxidant-induced cytotoxicity: A challenge for metabolic modulation. Am. J. Respir. Cell Mol. Biol., 4:l-3. Caplan, A.I., Niedergang, C.?Okazaki, H., and Mandel, P. (1979)Poly(ADPRibosei levels a s a function of chick limb mesenchymal cell development as studied in vitro and in vivo. Dev. Riol., 72:102-109. Chomczynski, P., and Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroformextraction. Anal. Biochcm., 162t156-159. De Wolf, M.J.S., Vitti, P., Ambesi-Impiombato, F.S., and Kohn, L.D. (1981) Thyroid membrane ADP ribosyltransferase activity. J. Biol. Chem., 256t12287-12296. Dean, 11.C. (1989) Expression of the fibronectin gene. Am. J. Respir. Cell Mol. Biol., lt5-10. Denhardt, D.T. (1966) A membrane-filter techniaue for the detection of complementary DNA. Biochem. Riophys. Res. Commun., 23541646. Exley, R., Gordon, J., and Clemens, M.J. (1987) Induction of B-cell differentiation antigens in interferon- or phobol ester-treated Daudi cells is impaired by inhibitors of ADP-ribosyltransferase. Proc. Natl. Acad. Sci. U S A . , 84,6467-6470. Fafeur, V., Jiang, Z.P., and Bohlen, P. (1991) Signal transduction by bFGF, but not TGFP1, involves arachidonic acid metabolism in endothelial cells. J . Cell. Physiol., 149r277-283.

280

BECKMANN

Feinberg, A.P., and Vogelstein, B. (1984) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem., 137.266-267. Gaal, J.C., and Pearson, C.K. (1985) Eukaryotic nuclear ADP-ribosylation reactions. Biochem. J., 230:l-18. Gil, G., Smith, J.R., Goldstein, J.L., Slaughter, C.A., Orth, K., Brown, M.S., and Osborne, T.F. (1988)Multiple genes encode nuclear factor 1-like proteins that bind to the promoter for 3-hydroxy-3-methylglutaryl-coenzyme A reductase. Proc. Natl. Acad. Sci. U.S.A., 85:8963-8967. Giri, S.N., Hyde, D.M., Nakashima, J.M., Wang, Q., Javadi, I., and Schiedt, M.J. (1989)Niacin and 3-aminobenzamide ameliorate bleomycin-induced lung fibrosis in hamsters. Am. Rev. Respir. Dis., 139tA206. Haschek. W.M.. Baer. K.E.. and Rutherford. J.E. (1989) Effects of dimethyl sulfoxide (DMSO) on pulmonary fibrosis in rats and mice. Toxicology, 54.197-205. Howe, P.H., Bascom, C.C.. Cunningham, M.R.. and Leof. E.B. (1989) Regulation of transforming growth factor p l action by multiple transducing pathways: Evidence for both G protein-dependent and -independent signaling. Cancer Res., 49t6024-6031. Hoyt, D.G., and Lazo, J.S. 11988) Alterations in pulmonary mRNA encoding procollagens, fibronectin and transforming growth factor+ precede bleomycin-induced pulmonary fibrosis in mice. J. Pharmacol. Exp. Ther., 246.765-771, Hunting, D.J., Gowans, B.J., and Henderson, J.F. (1985) Specificity of inhibitors of poly(ADP-ribose)synthesis. Mol. Pharmacol., 28200206. Hussain, M.Z., Giri, S.N., and Bhatnagar, R.S. (1985) Poly(ADP-rihose) synthetase activity during bleomycin-induced lung fibrosis in hamsters. Exp. Mol. Pathol., 43:162-176. Inageda, K., Nishina, H., and Tanuma, S. (1991) Mono-ADP-ribosylation of Gs by a n eukaryotic arginine-specific ADP-rihosyltransferase stimulates the adenylate cyclase system. Biochem. Biophys. Res. Commun., 176t1014-1019. Jetten, A.M., Shirley, J.E., and Stoner, G. (1986) Regulation of proliferation and differentiation of respiratory tract epithelial cells by TGFp. Exp. Cell Res., 167t539-549. Kelley, J.,Chrin, L., Shull, S.,Rowe, D.W., and Cutroneo, K.R. (1985) Bleomycin selectively elevates mRNA levels for procollagen and fihronectin following acute lung injury. Biochem. Biophys. Res. Commun., 131t836-843. Khalil, N., Bereznay, O., Sporn, M., and Greenherg, A.H. (1989) Macr0Dhat.e Droduction of transformine Prowth factor 6 and fibroblast collagen synthesis in chronic pulmonary inflammation. J. Exp. Med., 170r727-737. Khalil, N., Bereznay, O., and Greenherg, A.H. (29903 Alveolar macrophages from bleomycin-induced pulmonary inflammation synthesize and secrete transforming growth factor-beta (TGF-P).Am. Rev. Respir. Dis., 141:A138 (abstr.). Kornblihtt, A.R., Vibe-Pedersen, K., and Baralle, F.E. (1983)Isolation and characterization of cDNA clones for human and bovine fibronectins. Proc. Natl. Acad. Sci. U.S.A., 80t32183222. Lazenby, A.J., Crouch, E.C., McDonald, J.A., and Kuhn, C. (1990) Remodeling of the lung in bleomycin-induced pulmonary fibrosis in the rat. Am. Rev. Respir. Dis., 142:206-214. Lechner, d.F., and LaVeck, M.A. (1985) A serum-free method for culturing normal human bronchial epithelial cells at clonal density. J. Tiss. Cult. Meth., 9t43-48. Levy-Wilson, B. (1983) Glycosylation, ADP-rihosylation, and methylation of tetrahymena histones. Biochemistry, 22r48P489. Loesherg, C., van Rooij, H., and Smets, L.A. (1990) Meta-iodobenzylguanidine (MIBG),a novel high-affinity substrate for cholera toxin that interferes with cellular mono(ADP-rihosylation).Biochim. Biophys. Acta, 1037.92-99. Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982) In: Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, pp. 202-203. Massague, J., and Like, B. (19851 Cellular receptors for type p transforming growth factor. J . Biol. Chem., 260.2636-2645. Masui, T., Wakefield, L.M., Lechner, J.F., LaVeck, M.A., Sporn, M.B., and Harris, C.C. (1986) Type beta transforming growth factor is the primary differentiation-inducing serum factor for normal human bronchial epithelial cells. Proc. Natl. Acad. Sci. U S A . ; 83:243% 2442. Milam, K.M., and Cleaver, J.E. (1984) Inhibitors of poly(adenosine diphosphate-ribose) synthesis: Effect on other metabolic processes. Science, 223589-591. 1

Y

l

.I

.I

wr AL.

Moss, J.,Stanley, S.J., and Watkins, P.A. (1980) Isolation and properties of an NAD- and guanidine-dependent ADP-ribosyltransferase rrom turkey erythrocytes. J. Biol. Chem., 255:5838-5839. Nishio, A,, Nakanishi, S., Doull, J.,and Uyeki, E.M. (1983) Enhanced chondrocytic differentiation in chick limb bud cell cultures by inhibitors of poly(ADP-ribose) synthetase. Biochem. Biophys. Res. Commun., 111:750-759. Pelton, R.W., and Moses, H.L. (1990)The p type transforming growth factors: Mediators of cell regulation in the lung. Am. Rev. Respir. Dis., 142tS31435. Penttinen, R.P., Kobayashi, S., and Bornstein, P. (1988)Transforming growth factor p increases mRNA for matrix proteins both in the presence and in the absence of changes in mRNA stability. Proc. Natl. Acad. Sci. U.S.A., 85t110S1108. Phan, S.H., and Kunkel, S.L. (1992) Lung cytokine production in bleomycin-induced pulmonary fibrosis. Exp. Lung Res., 18t29-43. Porteous, J.W., Furneaux, H.M., Pearson, C.K., Lake, C.M., and Morrison, A. (1979) Poly(adenosine diphosphate ribose) synthetase activity in nuclei of dividing and of non-dividing but differentiating intestinal epithelial cells. Biochem. J., 180t455-463. Purnell, M.R., and Whish, W.J.D. (1980) Novel inhibitors of poly(ADP-ribose)synthetase. Biochem. J., 185:775-777. Raghow, R., Lurie, S., Seyer, J.M., and Kang, A.H. (1985) Profiles of steady state levels of messenger RNAs coding for type I procollagen, elastin, and fihronectin in hamster lungs undergoing bleomycin induced intersitial pulmonary fibrosis. J. Clin. Invest., 76r173P 1739. Raghow, R., Irish, P., and Kang, A.H. (1989) Coordinate regulation of transforming growth factor p gene expression and cell proliferation in hamster lungs undergoing hleomycin-induced pulmonary fibrosis. J. Clin. Invest., 84t18361842. Riley, D.J., Kerr, J.S., Berg, R.A., Ianni, B.D., Pietra, G.G., Edelman, N.H., and Prockop, D.J. (1982) Beta-aminopropionitrile prevents Bleomycin-induced pulmonary fibrosis in the hamster. Am. Rev. Respir. Dis., 125.67-73. Ristiniemi, J., and Oikarinen, J. (1989) Histone H1 binds to the putative nuclear factor I recognition sequence in the mouse alphaZ(1) collagen promoter. J. Biol. Chem., 264.2164-2174. Roberts, A.B., Anzano, M.A., Wakefield, L.M., Roche, N.S., Stern, D.F., and Sporn, M.B. (1985) Type P transforming growth factor: A bifunctional regulator of cellular growth. Proc. Natl. Acad. Sci. U.S.A., 82:119-123. Romberger, D., Beckmann, J., Claassen, L., Ertl, R., and Rennard, S.I. (1989) Transforming growth factor-beta 1 stimulates fibronectin release from bronchial epithelial cells. Am. Rev. Resp. Dis., 139:252 (abstr.). Rossi, P., Karsenty, G., Roberts, A.B., Roche, N.S., Sporn, M.B., andde Cromhrugghe, B. (1988)A nuclear factor 1binding site mediates the t,ranscriptional activation of a type 1 collagen promoter by transforming growth factor-p. Cell, 52.4055414. Santoro, C., Mermod, N., Andrews, P.C., and Tjian, R. (1988) A family of human CCAAT-box-binding proteins active in transcription and IINA replication: Cloning and expression of multiple cDNAs. Nature, 334;21%224. Sullivan, K.F., and Cleveland, D.W. (1986) Identification of conserved isotype-defining variable region sequences for four vertebrate p tubulin polypeptide classes. Proc. Natl. Acad. Sci. U.S.A., 83t43274331. Wang, Q., Giri, S.N., Hyde, D.M., andNakashima, J.M. (1989) Effects of taurine on bleomycin-induced lung fibrosis in hamsters. Proc. Soc. Exp. Biol. Med., 19Ot330-338. Wang, Q., Giri, S.N., Hyde, D.M., Li, C., and Schiedt, M.J. (199Oa) Prevention of bleomycin-induced pulmonary fibrosis in hamsters by combined treatment with taurine and niacin. Am. Rev. Respir. Dis., 141tA138 (abstr.1. Wang, Q . , Giri, S.N., Hyde, D.M., Nakashima, J.M., and Javadi, I. ( 1590b) Niacin attenuates bleomycin-induced lung fibrosis in the hamster. J. Biochem. Toxicol., 5:13-22. Ward, H.E., Hicks, M., Nicholson, A., and Berend, N. (1988) Deferoxamine infusion does not inhibit hleomycin-induced lung damage in the rat. Am. Rev. Respir. Dis., 137.1356-1359, Weintraub, H. (1985) Assembly and propagation of repressed and derepressed chromosomal states. Cell, 42t705-711. Williams, G.T., and Johnstone, A.P. (1983) ADP-ribosyl transferase, rearrangement of DNA, and cell differentiation. Biosci. Rep., 3r815830.