Inhibition of Myofibroblast Apoptosis by Transforming Growth Factor b1 Hong-Yu Zhang and Sem H. Phan Department of Pathology, University of Michigan Medical School, Ann Arbor, Michigan

Fibroblast differentiation to the myofibroblast phenotype is associated with a–smooth-muscle actin (a-SMA) expression and regulated by cytokines. Among these, transforming growth factor (TGF)-b1 and interleukin (IL)-1b can stimulate and inhibit myofibroblast differentiation, respectively. IL-1b inhibits a-SMA expression by inducing apoptosis selectively in myofibroblasts via induction of nitric oxide synthase (inducible nitric oxide synthase [iNOS]). Because TGF-b is known to inhibit iNOS expression, this study was undertaken to see if this cytokine can protect against IL-1b–induced myofibroblast apoptosis. Rat lung fibroblasts were treated with IL-1b and/or TGF-b1 and examined for expression of a-SMA, iNOS, and the apoptotic regulatory proteins bax and bcl-2. The results show that TGF-b1 caused a virtually complete suppression of IL-1b–induced iNOS expression while preventing the decline in a-SMA expression or the myofibroblast subpopulation. TGF-b1 treatment also completely suppressed the IL-1b–induced apoptosis in myofibroblasts. IL-1b–induced apoptosis was associated with a significant decline in expression of the antiapoptotic protein bcl-2, which was prevented by concomitant TGF-b1 treatment. The level of the proapoptotic protein bax, however, was not significantly altered by either cytokine. These data suggest that TGF-b1 inhibits IL-1b–induced apoptosis in myofibroblasts by at least two mechanisms, namely, the suppression of iNOS expression and the prevention of a decline in bcl-2 expression. Thus, TGF-b1 may be additionally important in fibrosis by virtue of this novel ability to promote myofibroblast survival by preventing the myofibroblast from undergoing apoptosis. Zhang, H.-Y., and S. H. Phan. 1999. Inhibition of myofibroblast apoptosis by transforming growth factor b1. Am. J. Respir. Cell Mol. Biol. 21:658–665.

The presence in active fibrotic lesions of a–smooth-muscle actin (a-SMA)–expressing fibroblasts, referred to as myofibroblasts, has been extensively documented (1–5). On the basis of the morphologic and biologic features of these cells, they may contribute to the increase in extracellular matrix deposition and contractility of lung parenchyma, which are associated with pulmonary fibrosis (3–5). Recent studies have additionally identified the myofibroblast as the primary source of the increased lung collagen gene expression in bleomycin-induced pulmonary fibrosis, as well as a major source of fibrogenic cytokines, such as transforming growth factor (TGF)-b, and chemokines (6–8). However, the mechanism responsible for the emergence of this cellular phenotype in pulmonary fibrosis is unclear. (Received in original form March 10, 1999 and in revised form June 1, 1999 ) Address correspondence to: Dr. Sem H. Phan, Dept. of Pathology, University of Michigan Medical School, Ann Arbor, MI 48109-0602. E-mail: [email protected] Abbreviations: a–smooth-muscle actin, a-SMA; bovine serum albumin, BSA; enzyme-linked immunosorbent assay, ELISA; immunoglobulin, Ig; interleukin, IL; inducible NO synthase, iNOS; monoclonal antibody, mAb; messenger RNA, mRNA; nitric oxide, NO; phosphate-buffered saline, PBS; polymerase chain reaction, PCR; reverse transcription, RT; standard error, SE; transforming growth factor, TGF; terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling, TUNEL. Am. J. Respir. Cell Mol. Biol. Vol. 21, pp. 658–665, 1999 Internet address: www.atsjournals.org

Alterations in the expression of a variety of cytokines, such as TGF-b, tumor necrosis factor-a, and interleukin (IL)-1b, are associated with alterations in a-SMA expression during the development of fibrosis (6–11). TGF-b is known to promote contraction of fibroblast-populated collagen gel and induce differentiation of myofibroblasts (12– 14), whereas IL-1b inhibits this process (15–17). Further, IL-1b is a potent inducer of high levels of nitric oxide (NO) production in rat lung fibroblasts (15, 18, 19), whereas induction of NO production is associated with a significant decline in a-SMA expression in vascular smooth-muscle cells (16, 17). More recently, IL-1b is reported to induce apoptosis and inhibit cell proliferation in rat lung fibroblasts, thyrocytes, and chondrocytes (15, 20, 21). Confirming the importance of NO in mediating the induction of apoptosis is the demonstration that exogenous NO or NO released by a variety of activated cells can induce DNA strand breaks and apoptosis (22–24). Further proof is provided by the observation that transfection of the inducible NO synthase (iNOS) gene into murine melanoma cell results in apoptosis in these cells (25). IL-1b–induced endogenous NO production is also capable of mediating apoptosis in pancreatic RINm 5F cells and lung fibroblasts (15, 26). In contrast to the effects of IL-1b, TGF-b is reported to be a potent inhibitor of iNOS in mouse macrophages and rat vascular smooth-muscle cells (27–30). TGF-b is capable of downregulating IL-1b–induced iNOS expression in these cells, as well as in microvascular endothelial cells and

Zhang and Phan: TGF-b1 Inhibits Myofibroblast Apoptosis

659

microglial cells (31, 32). However, it remains unclear whether suppression of iNOS expression by TGF-b leads to inhibition of apoptosis in these cells. This is an important issue because suppression of myofibroblast apoptosis may lead to the prolonged survival of this cell with expected deleterious consequences on the progression of pulmonary fibrosis. If TGF-b could inhibit apoptosis by suppression of iNOS, this would add another dimension to the fibrogenic properties of this cytokine; namely, by promoting the survival of the myofibroblast. The cellular mechanisms mediating myofibroblast apoptosis are as yet undefined. In other cell types or cell lines in which these pathways have been investigated, the process appears to be regulated by a complex system consisting of numerous proteins and cascading proteolytic and phosphorylation steps. Among the proteins that have been identified are members of the bcl-2 gene family, whose members could either suppress (bcl-2) or promote (bax) apoptosis (33). Hence, in many cell types the ratio of the level of expression of these two proteins with opposing activities vis-à-vis induction of apoptosis, may be decisive in determining whether a cell dies or lives (33). The expression of these proteins during TGF-b regulation of IL-1b– induced myofibroblast apoptosis is unknown. To investigate these issues, the effects of TGF-b on IL1–induced iNOS expression and apoptosis were examined in isolated rat lung fibroblasts. Also, the possible relationship of TGF-b modulation of apoptosis to expression of bcl-2 and bax was investigated, given their importance in inhibiting and promoting apoptosis, respectively.

Inc., Minneapolis, MN) at the indicated doses and times. After fixation and permeabilization with 4% paraformaldehyde and 0.5% Triton X-100 in PBS, respectively, the coverslips were stained as follows. To enumerate the percentage of apoptotic cells, a TUNEL fluorescein kit (APODIRECT; PharMingen, San Diego, CA) was used. After blocking with 10% BSA in PBS, fibroblasts on the coverslip were first subjected to TUNEL assay as described by the manufacturer but excluding the propidium iodide steps. To identify the cells (myofibroblast versus fibroblast) undergoing apoptosis, the slides were then stained with a mouse anti–a-SMA monoclonal antibody (mAb) (Boehringer Mannheim Corp., Indianapolis, IN) and a matched secondary antibody, Texas Red–conjugated goat antimouse immunoglobulin (Ig)G (Molecular Probes, Inc., Eugene, OR). To evaluate cellular localization of iNOS vis-à-vis myofibroblast phenotype, the possible coexpression of a-SMA with iNOS was analyzed by dual immunostaining for these antigens. Cell monolayers on coverslips were treated as described earlier for analysis of apoptosis. After fixation, they were first stained with a polyclonal rabbit anti-iNOS antibody (Cayman Chemical Corp., Ann Arbor, MI) and then incubated with Texas Red–conjugated goat antirabbit IgG (Molecular Probes). The cells were then stained with a fluorescein isothiocyanate–conjugated monoclonal anti–a-SMA antibody (Sigma, St. Louis, MO). After washing, the coverslides were mounted in Mowiol (Calbiochem, La Jolla, CA), examined, and photographed with a Zeiss Aristoplan fluorescence microscope. For each coverslip, the total number of cells and the number of cells positive for any of the above properties were counted in four or more randomly chosen, noncontiguous, high-power (340 objective) fields until a minimum of 100 total cells were counted. For each treatment group, at least three coverslips were examined. The results were expressed as the percentage of positive cells per high-power field.

Materials and Methods Cell Culture Rat lung fibroblasts were isolated from adult rat lungs as described previously (15). Briefly, 4- to 6-wk-old rats were killed and their lungs were perfused with phosphate-buffered saline (PBS). The lung tissue was digested in trypsin– ethylenediaminetetraacetic acid (EDTA) solution until the cells were released. The cells were cultured in complete medium composed of Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 mg/ml streptomycin, and 0.25 mg/ ml fungizone. The morphologic and synthetic characteristics of the cultured cells were consistent with those for fibroblasts, as previously described (34). All cells used in this study were between cell passage numbers 4 and 8 after primary culture. Confluent cell monolayers were treated with the indicated substances in serum-free media supplemented with 2 mg/ml bovine serum albumin (BSA) for the indicated times in the various experiments described later. In Situ Labeling of Apoptotic Cells and Immunofluorescence Analysis Cells undergoing apoptosis are identified using terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) as previously described (15). Monolayer cultures on coverslips were treated with or without human recombinant IL-1b (specific activity 5 108 U/mg protein; Genzyme Diagnostics, Cambridge, MA) and/or human recombinant TGF-b1 (R&D Systems,

Quantitative Immunoassay for iNOS Expression iNOS expression in cells was also analyzed using a modified enzyme-linked immunosorbent assay (ELISA) as previously described for quantitation of cellular a-SMA expression (15). Briefly, confluent cells in 96-well microtiter plates were treated with the indicated concentrations of IL-1b and/or TGF-b1 in serum-free DMEM containing BSA (2 mg/ml). At the indicated time points, the cells were immediately fixed by addition of methanol and then blocked with 10% FBS in TBST buffer (20 mM Tris-HCl [pH 7.5], 0.5 M NaCl, and 0.05% Tween-20) on a rotating platform. After incubation with anti-iNOS antibody, visualization of bound antibody was undertaken using biotinylated antirabbit IgG in conjunction with streptavidin-conjugated alkaline phosphate and chromogenic substrate (3 mM p-nitrophenyl phosphate, 0.05 M NaCO3, and 0.05 mM MgCl2). The absorbance was then read at 405 nm using a microplate reader (Titertek Multiskan, MCC/340; Flow Laboratories, McLean, VA). Nonimmune rabbit IgG (Sigma) was used as a negative control. Assays were performed in triplicate and repeated at least twice. Immunoblotting Analysis After undergoing treatments as described previously, cells were extracted directly into lysis buffer (10 mM Tris-HCl

660

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

[pH 7.6], 5 mM EDTA, 50 mM NaCl, 5 mg/ml aprotinin, 1 mg/ml pepstatin, 1 mM phenylmethylsulfonyl fluoride, and 1% Nonidet P-40 [Sigma]). The immunoblot analysis of these cell extracts was performed as described previously (15). Briefly, equal amounts of protein were electrophoresed through 7.5% (for iNOS), 12% (for a-SMA), or 15% (for bcl-2 and bax) polyacrylamide gels in the presence of sodium dodecyl sulfate. After being transferred onto polyvinylidene fluoride membranes (Millipore Corp., Bedford, MA), nonspecific binding was blocked with 10% nonfat milk. The blots were then incubated with the primary antibodies for detection of a-SMA (1:2,000), iNOS (1:2,000), bcl-2 (1:1,500; mouse mAb; Transduction Laboratories, Lexington, KY), or bax (1:1,000; mouse mAb; Trevigen, Gaithersburg, MD). Horseradish peroxidase–conjugated antimouse IgG or antirabbit IgG (1:10,000; Amersham, Arlington Heights, IL) was then used to detect the bound primary antibodies, and finally developed using enhanced chemiluminescence (ECL kit; Amersham). Photographs of the exposed and developed X-ray films were shown. For quantitative analysis, the films were subjected to densitom-

etry by scanning the images and analyzed using 1D Image Analysis Software (Kodak Digital Science, New Haven, CT). Results were expressed as random integration units based on summation of pixel intensity within the immunologically detected protein band of interest. Reverse Transcription/Polymerase Chain Reaction Analysis Rat lung fibroblasts were cultured to confluence in 100mm2 dishes and then treated with IL-1b and/or TGF-b1 at the indicated doses and time intervals as described previously. Total RNA was then extracted from the cells using Trizol reagent (GIBCO BRL, Gaithersburg, MD). Reverse transcription (RT) and polymerase chain reaction (PCR) was performed using the Superscript One-step RT-PCR kit (GIBCO BRL) according to the manufacturer’s instructions. Briefly, 0.5 mg of total RNA was used for each RT-PCR reaction and the following primers were used: iNOS, sense 59-AGG GAG TGT TGT TCC AGG TG-39, antisense 59-TCC TCA ACC TGC TCC TCC TCA CT-39; and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), sense 59-TCC AGT ATG ACT CTA CCC ACG-39, antisense 59-GTC TTC TGA GTG GCA GTG ATG-39. These primers were chosen on the basis of the published sequence for rat iNOS and GAPDH messenger RNA (mRNA) (35, 36). Forty cycles of amplification were performed using a thermal cycler (PCT 200 DNA Engine; MJ Research, Watertown, MA). PCR products were visualized after separation on 2% agarose gels by staining with ethidium bromide. The gels were photographed and the specific bands were quantified using Kodak Digital Science Image Analysis Software (Eastman Kodak Co., Rochester, NY). Statistical Analysis Results were expressed as means 6 standard error (SE), and the differences between means of various treatment control groups were analyzed using Student’s t test for paired data or analysis of variance followed by Scheffé’s test for comparisons of multiple ( . 2) group means. A value of P , 0.05 was considered significant.

Figure 1. Effects of IL-1b and TGF-b1 on apoptosis and a-SMA expression. Rat lung fibroblasts were treated with IL-1b, TGF-b1, or both as described in M ATERIALS AND METHODS for the indicated times. The cells were then subjected to TUNEL and immunohistochemical staining for detection of apoptotic and a-SMA– expressing cells, respectively. The percentage of apoptotic ( A) and a-SMA–expressing (B) cells was determined, as well as that of cells expressing both positive TUNEL (apoptotic cells) and actin staining (C). Because some apoptotic nuclei were devoid of visible cytoplasm, the number of apoptotic and a-SMA–positive cells was likely to be an underestimate of such cells. Numbers shown represent means 6 SE, with n 5 3. The effects of IL-1b (alone) on all three parameters were statistically significant relative to all other treatment groups at all time points examined. The effects of TGF-b1 treatment were also statistically significant in B relative to the no-treatment group (None) at all three time points. The IL-1b 1 TGF-b1 group was significantly different from the TGF-b1 group in A at all time points, and in B only at the 24-h time point.

Zhang and Phan: TGF-b1 Inhibits Myofibroblast Apoptosis

Results Effect of TGF-b1 on a-SMA Expression and Apoptosis Despite the fact that survival of fibroblasts depends upon the continued presence of growth factors (15, 23, 37), after total withdrawal of serum only a small percentage of rat lung fibroblasts (, 5%) was observed to undergo apoptosis in the first few days (Figure 1A). When these cells were treated with IL-1b (2 ng/ml), a noticeable increase in the percentage of apoptotic cells was observed by phase contrast microscopy, beginning at about 4 h and continuing to increase significantly with prolonged IL-1b treatment, reaching a maximum at just under 18% apoptotic cells. When treated with IL-1b for 4 h, some cells showed a reduction in size, cell contact was lost, and detachment of cells from the culture dish was observed. Beginning at this time point, highly condensed and contracted nuclei and typical apoptotic bodies were seen, and the presence of apoptosis was confirmed in situ using the TUNEL technique. When IL-1b–treated cells were concomitantly treated with TGF-b1, however, the number of apoptotic cells was significantly reduced to less than 9% (Figure 1A). TGF-b1 treatment alone did not significantly influence the number of apoptotic cells. IL-1b and TGF-b1 are known to regulate the expression of a-SMA in rat lung fibroblasts and other cells (12– 17). These studies show that IL-1b downregulates the expression of a-SMA via selective induction of apoptosis in myofibroblasts, whereas TGF-b1 upregulates this expression and, hence, myofibroblast differentiation. By implica-

Figure 2. Effects of IL-1b and TGF-b1 on a-SMA protein expression. Rat lung fibroblasts were treated with IL-1b, TGF-b1, or both as described in MATERIALS AND METHODS. The cell extracts were then subjected to immunoblotting for detection of a-SMA. The results of quantitation by densitometry were expressed as random integration units as described in M ATERIALS AND METHODS and shown as means 6 SE of triplicate samples. Statistical analysis showed significant differences between control (None) versus IL-1b–treated groups (P , 0.05) and between these two groups versus the TGF-b1–treated group (P , 0.05 and , 0.01, respectively).

661

tion, these findings suggest that TGF-b1 may prevent or reduce the IL-1b inhibitory influence on actin expression by promoting the survival of the myofibroblast. To evaluate this possibility, the effects of IL-1b and/or TGF-b1 on myofibroblast differentiation and apoptosis were examined simultaneously using TUNEL staining for apoptosis, combined with immunostaining for a-SMA expression. By counting cells with these properties on coverslip culture, the results show that after exposure to IL-1b the number of nonapoptotic (TUNEL-negative) actin-positive cells declined significantly (Figure 1B). In contrast, the number of apoptotic actin-positive cells increased with IL-1b treatment (Figure 1C). TGF-b1 treatment alone caused a significant increase in actin expression or myofibroblast differentiation, and prevented the IL-1b–induced decline in actin expression (Figure 1B). Additionally, TGF-b1 treatment had no significant effect on the number of apoptotic actin-positive cells, but prevented the IL-1b–induced increase in the proportion of apoptotic myofibroblasts (Figure 1C). The results of Western blotting confirmed the effect of TGF-b1 in preventing the reduction of a-SMA expression by IL-1b (Figure 2). Thus, IL-1b caused a significant reduction in a-SMA expression in control untreated cells,

Figure 3. Effects of IL-1b and TGF-b1 on iNOS expression by ELISA. Cells were treated with IL-1b, TGF-b1, or both as in Figure 2. The cells were then subjected to a modified ELISA for quantitation of iNOS protein expression as described in M ATERIALS AND METHODS. Readings from the ELISA reader in absorbance units are shown. The blank reading from an empty well (no cells plated) gave an absorbance reading of 0.39 6 0.03 (mean 6 SE), thus absorbances at or below this reading indicated undetectable iNOS expression. IL-1b was a potent and rapid inducer of iNOS (P , 0.01 versus None), which was significantly inhibited (P , 0.01) by TGF-b1 at time points . 2 h (IL-1b group versus IL-1b 1 TGF-b1 group). The effects of TGF-b1 alone were not statistically significant relative to the control (None) group, but were significantly different from the IL-1b 1 TGF-b1 group at 2 and 4 h. Data represent means 6 SE, with n 5 3.

662

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

Figure 4. Effects on iNOS expression by immunoblotting. Cells treated as in Figure 3 were also analyzed for iNOS expression by immunoblotting to confirm the results in Figure 3. A representative blot from two separate experiments is shown. Similar results were obtained in both experiments. The results show essentially complete abolition of IL-1b–induced iNOS expression by TGF-b1.

which was prevented by concomitant treatment with TGFb1. Further, the results confirmed the ability of TGF-b1 alone to stimulate a-SMA expression, consistent with its ability to promote myofibroblast differentiation. Inhibition of IL-1b–Induced iNOS Expression by TGF-b1 Previous studies have provided evidence that the production of endogenous NO is associated with apoptosis of normal cells (15, 21–26, 38, 39), and transfection of the iNOS gene into metastatic cells is associated with suppression of tumorigenicity by induction of apoptosis (25). Cytokines such as IL-1b can induce iNOS expression in macrophages and fibroblasts (15, 18, 19, 39) and thus induce apoptosis by secretion of endogenous NO from within the cells themselves or by exogenous NO from adjacent activated cells. Despite evidence that TGF-b can inhibit the activation of iNOS by IL-1 in a variety of cells (27–32), direct evidence that this cytokine could protect cells against apoptosis via inhibition of iNOS induction is lacking. To examine this issue, the effect of TGF-b1 on iNOS expression in fibroblasts was determined using a modified ELISA and double immunofluorescence to provide complementary information on the cellular distribution of iNOS vis-àvis a-SMA expression in certain cells. Although untreated

control cells expressed undetectable levels of iNOS, IL1b–treated cells rapidly expressed this enzyme, being detected as early as 2 h after treatment (Figure 3). Concomitant treatment with TGF-b1 inhibited this IL-1b–induced iNOS expression, whereas treatment solely with TGF-b1 did not significantly alter expression. The kinetics of inhibition revealed that by 8 h the level of iNOS expression in IL-1b 1 TGF-b1–treated cells was not significantly different from that in untreated control cells or in cells treated with TGF-b1 alone (Figure 3). Western blotting analysis confirmed that iNOS induction by IL-1b was rapid and persisted up to 72 h of treatment, but was completely inhibited after 12 h of concomitant treatment with TGF-b1 (Figure 4). RT-PCR analysis was then used to examine the effects of cytokine treatment on iNOS mRNA expression. Consistent with the protein expression studies, IL-1b induced the expression of iNOS mRNA after 2 h of exposure, increasing to a maximum after 12 h of treatment and persisting up to 72 h (Figure 5). Concomitant treatment with TGF-b1 essentially abolished IL-1b–induced iNOS mRNA expression. Together, these findings and the preceding results on selective susceptibility of myofibroblasts to undergo IL1b–induced apoptosis would suggest that iNOS expression may correlate with and/or perhaps mediate susceptibility to apoptosis induction in myofibroblasts. Because iNOS is considered in certain cells to be part of the apoptotic pathway, the expectation is that myofibroblasts would selectively express this enzyme in response to IL-1b–induced apoptosis. To examine this issue, cells were subjected to double immunofluorescence analysis for iNOS and a-SMA expression. Unexpectedly, although apoptosis occurred preferentially in a-SMA–positive cells or myofibroblasts (15), analysis by double immunofluorescence showed iNOS was detected exclusively in nonmyofibroblasts or fibroblasts that do not express a-SMA (Figure 6). Thus, iNOS expression in adjacent fibroblasts correlated with apoptosis in myofibroblasts. Effect of IL-1b and/or TGF-b1 on bcl-2 and bax Expression In view of the regulatory effects of IL-1b and TGF-b1 on myofibroblast apoptosis, the effects of these cytokines on certain apoptotic pathway protein expression were exam-

Figure 5. Effects on iNOS expression by RT-PCR. Cells were treated as in Figure 3 and cellular RNA was extracted for RT-PCR analysis of iNOS mRNA. A representative electropherogram is shown of RTPCR products from cells treated for the indicated times with the indicated cytokine or cytokines. The experiment was repeated once and yielded similar results. IL-1b was a potent and rapid inducer of iNOS at the mRNA level as well, and was also inhibitable by TGF-b1.

Zhang and Phan: TGF-b1 Inhibits Myofibroblast Apoptosis

Figure 6. Cellular localization of iNOS and a-SMA expression. Cells were treated with IL-1b for 6 h and then immunostained for both iNOS (red fluorescence) and a-SMA (green fluorescence) expression. The slides were then successively photographed for both green and red fluorescence. Two representative microscopic fields are shown. Expression of iNOS was seen only on cells not expressing a-SMA. Bar in bottom right corner represents 30 and 20 mm in A and B, respectively.

ined. Recent data demonstrated that a family of bcl-2– related genes is intimately involved in the regulation of apoptosis in several cell types (33, 40–42). Specifically, the overexpression of bcl-2 suppresses apoptosis, which would otherwise occur in response to a number of stimuli, including oxidative stress (33, 41, 42). In this study, inasmuch as TGF-b1 was found to inhibit IL-1b–induced apoptosis, the effects of both these cytokines on expression of bcl-2 and its proapoptotic counterpart, bax, were examined. Western blotting analysis of untreated control cell extracts revealed detectable bcl-2 and bax expression (Figures 7A and 7B, respectively). Upon treatment with IL-1b at doses that induce apoptosis in myofibroblasts (15), the level of bcl-2 expression was found to decline significantly after 24 h of treatment (Figure 7A). This reduction in bcl-2 expression was prevented by concomitant treatment with TGFb1, whereas treatment with TGF-b1 alone did not signifi-

663

Figure 7. Effects on bcl-2 and bax expression. Cells were treated as in Figure 4 with IL-1b, TGF-b1, or both for the indicated times. The cell extracts were then subjected to immunoblotting for detection of bcl-2 (A) and bax (B) expression. The results of densitometric quantitation of the immunoblots were expressed as random integration units as described in M ATERIALS AND METHODS. Data are shown as means 6 SE of triplicate samples. Statistical analysis showed that the inhibition of bcl-2 expression by IL-1b was significant (P , 0.05 versus the None or IL-1b 1 TGF-b1 groups) at 48 and 72 h (A). At these same time points, bcl-2 levels in cells concomitantly treated with TGF-b1 (the IL-1b 1 TGF-b1 group) were not significantly different from those of control untreated (None) cells or those treated with TGF-b1 alone. However, no significant effects on bax expression were detectable after any of the treatments examined (B).

cantly affect bcl-2 expression. In contrast, bax expression in control cells was not altered by any of these treatments, singly or in combination (Figure 7B). Thus, the level of bcl-2 but not of bax expression was associated with the protection afforded by TGF-b1 against IL-1b–induced apoptosis. The IL-1b–induced reduction in bcl-2 expression may represent a key mechanism or signal for apoptosis in myofibroblasts.

Discussion Previous studies have suggested that myofibroblasts disappear from healing wounds via apoptosis (43). Additionally, IL-1b is found to selectively induce apoptosis in myofibroblasts, which is mediated by the induction of iNOS (15). In contrast, TGF-b expression is associated with the

664

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

emergence of the myofibroblast, and treatment of cells in vitro with this cytokine promotes myofibroblast differentiation from the fibroblast and inhibits iNOS expression (13–15, 27–32, 44–46). However, the mechanism by which TGF-b promotes myofibroblast differentiation and protects it from apoptosis is undetermined. To examine this issue, the present study evaluated the relationship between iNOS expression, apoptosis, and the expression of certain apoptotic pathway proteins in the context of the functional responses to IL-1b and TGF-b1 treatments. Consistent with previous studies, IL-1b was found to induce apoptosis in myofibroblasts, whereas TGF-b1 promoted a-SMA expression or myofibroblast differentiation. The novel finding here was that TGF-b1 was able to prevent or inhibit the apoptosis induced by IL-1b. Thus, in addition to directly promoting a-SMA gene expression (46) or myofibroblast differentiation, TGF-b1 appears to favor the persistence of myofibroblasts by enhancing their survival via inhibition of IL-1b–induced apoptosis. A potential mechanism by which TGF-b1 protects myofibroblasts from IL-1b–induced apoptosis is suggested by the requirement for iNOS in the latter process (15). Indeed, the results show that TGF-b1 was a potent inhibitor of iNOS expression in rat lung fibroblasts, and that this inhibition correlated with the protection of the myofibroblast from IL-1b–induced apoptosis. Curiously, however, the iNOS expression induced by IL-1b was exclusively localized to fibroblasts and not detected in myofibroblasts, which were selectively targeted to undergo apoptosis by this treatment. This suggests that NO secreted by adjacent fibroblasts is responsible for mediating the induction of apoptosis in myofibroblasts. This paracrine mode of regulation implies that the fibroblasts themselves may be immune to the apoptotic inducing effects of NO, inasmuch as IL-1b selectively induces apoptosis in myofibroblasts (15). Although NO from adjacent fibroblasts represents an important signaling mechanism for inducing and/or mediating apoptosis in myofibroblasts, the actual mechanism involved in induction of apoptosis in myofibroblasts remains unclear. On the basis of studies of apoptosis in other cells, a large array of proteins and enzymes has been identified as important mediators or regulators of this process (33, 40). Among these are the antiapoptotic proteins of the bcl-2 family, whose activity is countered by their proapoptotic counterparts, such as bax (33). Examination of bcl-2 and bax expression revealed that only the level of expression of bcl-2 correlated with resistance to IL-1b–induced apoptosis. This conclusion was based on the observation that bcl-2 expression was reduced upon IL-1b–induced apoptosis, and that TGF-b1 could prevent this reduction in bcl-2 expression, which was in turn correlated with inhibition of iNOS expression and reduction in IL-1b–induced apoptosis. Because TGF-b1 treatment alone has no significant effects on bcl-2 expression, this would suggest that this cytokine protects against apoptosis not by increasing bcl-2 expression but by inhibiting the IL-1b–induced reduction in bcl-2 expression, possibly mediated by inhibition of NO production. These selective alterations in bcl-2 expression would result in alterations in the bcl-2–to–bax ratio, which is key in determination of whether a cell will undergo apoptosis or survive (33).

The findings in this study provide new clues as to how myofibroblasts appear de novo during the active/synthetic phase of wound healing and fibrosis; but more importantly, they suggest a mechanism by which these cells could persist or disappear as the fibrosis progresses or resolves, respectively. The importance of TGF-b in this process is 2-fold. First, its ability to directly upregulate a-SMA gene expression (46) represents an obvious mechanism by which it could induce the de novo appearance of myofibroblasts in healing tissue. Second, the results of this study additionally identified a novel mechanism by which myofibroblast numbers could be augmented; namely, by affording them protection from apoptosis. It is therefore logical to surmise that the persistent presence of TGF-b could result in the persistence of these cells, with deleterious consequences on normal repair by prolonging and/or intensifying the deposition of extracellular matrix by these cells. Hence, by implication, given the importance of iNOS in myofibroblast apoptosis, it may be inadvisable to use iNOS inhibitors in diseases with potential for progressive fibrosis. Further work is necessary to prove that these mechanisms are actually operative in vivo. Acknowledgments: This work was supported by National Institutes of Health grants HL28737, HL31963, and HL52285. The expert technical assistance of Bridget McGarry is also acknowledged.

References 1. Zhang, K., M. D. Rekhter, D. Gordon, and S. H. Phan. 1994. Co-expression of a-smooth muscle actin and type I collagen in fibroblast-like cells of rat lung with bleomycin-induced pulmonary fibrosis: a combined immuno-histochemical and in situ hybridization study. Am. J. Pathol. 145:114–121. 2. Doane, K., R. A. McReynolds, and F. J. Wilson. 1983. Immunofluorescence location of contractile proteins in the rat lung following bleomycin injury. Histochem. J. 15:82–88. 3. Low, R. B., J. Woodcock-Mitchell, J. N. Evans, and K. B. Adler. 1984. Actin content of normal and of bleomycin-fibrotic rat lung. Am. Rev. Respir. Dis. 129:311–316. 4. Mitchell, J., J. Woodcock-Mitchell, S. Reynolds, R. B. Low, K. O. Leslie, K. B. Alder, G. Gabbiani, and O. Skalli. 1989. a-Smooth muscle actin in parenchymal cells of bleomycin-injured rat lung. Lab. Invest. 60:643–650. 5. Kuhn, C., and J. A. McDonald. 1991. The role of the myofibroblast in pulmonary fibrosis. Am. J. Pathol. 138:1257–1265. 6. Phan, S. H., and S. L. Kunkel. 1992. Lung cytokine production in bleomycin-induced pulmonary fibrosis. Exp. Lung Res. 18:29–43. 7. Zhang, K., M. Gharaee-Kermani, M. L. Jones, J. S. Warren, and S. H. Phan. 1994. Lung monocyte chemoattractant protein-1 gene expression in bleomycin-induced pulmonary fibrosis. J. Immunol. 153:4733–4741. 8. Zhang, K., K. C. Flanders, and S. H. Phan. 1995. Cellular localization of transforming growth factor b expression in bleomycin-induced pulmonary fibrosis. Am. J. Pathol. 147:352–361. 9. Elias, J. A., B. Freundlich, J. A. Kern, and J. Rosenbloom. 1990. Clinical implications of basic research: cytokine networks in the regulation of inflammation and fibrosis in the lung. Chest 97:1439–1445. 10. Elias, J. A. 1988. Tumor necrosis factor interacts with interleukin-1 and interferons to inhibit fibroblast proliferation via fibroblast prostaglandin dependent and independent mechanisms. Am. Rev. Respir. Dis. 138:652–658. 11. Montesano, R., and L. Orci. 1988. Transforming growth factor-b stimulated collagen-matrix contraction by fibroblasts: implication for wound healing. Proc. Natl. Acad. Sci. USA 85:4894–4897. 12. Tingstrom, A., C. H. Heldin, and K. Rubin. 1992. Regulation of fibroblastmediated collagen gel contraction by platelet-derived growth factor, interleukin-1 alpha and transforming growth factor-beta 1. J. Cell Sci. 102:315– 322. 13. Zhang, H-Y., M. Gharaee-Kermani, K. Zhang, S. Karmiol, and S. H. Phan. 1996. Lung fibroblast a-smooth muscle actin expression and contractile phenotype in bleomycin-induced pulmonary fibrosis. Am. J. Pathol. 148:527– 537. 14. Desmouliere, A. 1995. Factors influencing myofibroblast differentiation during wound healing and fibrosis [Review]. Cell Biol. Int. 19:471-476. 15. Zhang, H.-Y., M. Gharaee-Kermani, and S. H. Phan. 1997. Regulation of lung fibroblast a-smooth muscle actin expression, contractile phenotype, and apoptosis by IL-1b. J. Immunol. 158:1392–1399.

Zhang and Phan: TGF-b1 Inhibits Myofibroblast Apoptosis

16. Trinkle, L. A., D. Beasley, and R. S. Moreland. 1990. Interleukin-1b alters actin expression and inhibits contraction of rat thoracic aorta. Am. J. Physiol. 262:C828–C833. 17. Beasley, D., R. A. Cohen, and N. G. Levinsky. 1989. Interleukin-1 inhibits contraction of vascular smooth muscle. J. Clin. Invest. 83:331–335. 18. Jorens, P. G., F. J. Van Overveld, P. A. Verneire, H. Bult, and A. G. Herman. 1992. Synergism between interleukin-1b and interferon-g, an inducer of oxide synthase, in rat lung fibroblasts. Eur. J. Pharmacol. 224:7–12. 19. Lavnikova, N., and D. L. Laskin. 1995. Unique patterns of regulation of nitric oxide production in fibroblasts. J. Leukoc. Biol. 58:451–458. 20. Giordano, C., G. Stassi, R. D. Maria, M. Todaro, P. Richiusa, G. Papoff, G. Ruberti, M. Bagnasco, R. Testi, and A. Galluzzu. 1997. Potential involvement of Fas and its ligand in the pathogenesis of Hashimoto’s thyroiditis. Science 275:960–963. 21. Blanco, F. J., R. L. Ochs, H. Schwarz, and M. Lotz. 1995. Chondrocyte apoptosis induced by nitric oxide. Am. J. Pathol. 146:75–85. 22. Estevéz, A. G., N. Spear, S. M. Machelle, R. Radi, C. E. Henderson, L. Barbeito, and J. S. Beckman. 1998. Nitric oxide and superoxide contribute to motor neuron apoptosis induced by trophic factor deprivation. J. Neurosci. 18:923–931. 23. Fehsel, K., K. Kroncke, K. L. Meyer, H. Huber, V. Wahn, and V. KolbBachofen. 1995. Nitric oxide induces apoptosis in mouse thymocytes. J. Immunol. 155:2858–2865. 24. Troy, C. M., D. Derossi, A. Prochiantz, L. A. Greene, and M. L. Shelanski. 1996. Downregulation of Cu/Zn superoxide dismutase leads to cell death via the nitric oxide-peroxynitrite pathway. J. Neurosci. 16:253–261. 25. Xie, K., S. Huang, Z. Dong, S.-H. Juang, M. Gutman, Q.-W. Xie, C. Nathan, and I. L. Fidler. 1995. Transfection of inducible nitric oxide synthase gene suppresses tumorigenicity and abrogates metastasis in K-1737 murine melanoma. J. Exp. Med. 181:1333–1343. 26. Ankarcrona, M., J. Dypdukt, M. B. Brune, and P. Nicotera. 1994. Interleukin-1b-induced nitric oxide production activates apoptosis in pancreatic RINm5F cells. Exp. Cell Res. 213:172–177. 27. Lin, J. Y., R. Sequin, K. Keller, and K. Chadee. 1995. Transforming growth factor-b primes macrophages for enhanced expression of the nitric oxide synthase gene for nitric oxide-dependent cytotoxicity against Entamoeba histolytica. Immunology 85:400–407. 28. Perrella, M. A., C. Patterson. L. Tan, S. F. Yet, C. M. Hsieh, M. Yoshizumi, and M. E. Lee. 1996. Suppression of interleukin-1b-induced nitric-oxide synthase promoter/enhancer activity by transforming growth factor-beta 1 in vascular smooth muscle cells. J. Biol. Chem. 271:13776–13780. 29. Finder, J., W. W. Stark, Jr., D. K. Nakayama, D. Geller, K. Wasserloos, B. R. Pitt, and P. Davies. 1995. TGF-b regulates production of NO in pulmonary artery smooth muscle cell by inhibiting expression of NOS. Am. J. Physiol. 268:L862–L867. 30. Perrella, M. A., M. Yoshizumi, Z. Fen, J. C. Tsai, C. M. Hsieh, S. Kourembanas, and M. E. Lee. 1994. Transforming growth factor-beta 1, but not dexamethasone, downregulates nitric oxide synthase mRNA after its induction by interleukin-1 beta in rat smooth muscle cells. J. Biol. Chem. 269:14595–14600. 31. Murata, J. S., B. Corradin, E. Felley-Bosco, and L. Juillerat-Jeanneret. 1995. Involvement of a transforming-growth-factor-beta-like molecule in tumor-

665

32.

33. 34. 35. 36.

37. 38.

39. 40. 41.

42. 43. 44.

45.

46.

cell-derived inhibition of nitric-oxide synthesis in cerebral endothelial cells. Int. J. Cancer 62:743–748. Vincent, V. A., F. J. Tilders, and A. M. Van Dam. 1997. Inhibition of endotoxin-induced nitric oxide synthase production in microglial cells by the presence of astroglial cells: a role for transforming growth factor beta. Glia 19:190–198. Adams, J. M., and S. Cory. 1998. The bcl-2 protein family: arbiters of cell survival. Science 281:1322–1326. Phan, S. H., J. Varani, and D. Smith. 1985. Rat lung fibroblast collagen metabolism in bleomycin-induced pulmonary fibrosis. J. Clin. Invest. 76:241– 247. Galea, E., D. J. Reis, and D. L. Feinstein. 1994. Cloning and expression of inducible nitric oxide synthase from rat astrocytes. J. Neurosci. Res. 37: 406–414. Fort, P., L. Marty, M. Piechaczyk, S. Sabrouty, C. Dani, P. Jeanteur, and J. M. Blanchard. 1984. Post-transcriptional regulation of glyceraldehyde-3phosphate-dehydrogenase gene expression in rat tissues. Nucleic Acids Res. 12:6951–6963. Guerne, P. A., A. Sublet, and M. Lotz. 1994. Growth factor responsiveness of human articular chondrocytes and fibroblasts. J. Cell. Physiol. 158:476– 484. Wong, H. R., J. D. Finder, K. Wasserloos, C. J. Lowenstein. D. A. Geller, T. R. Billiar, B. R. Pitt, and P. Davies. 1996. Transcriptional regulation of iNOS by IL-1b in cultured rat pulmonary artery smooth muscle cells. Am. J. Physiol. 271:L166–L171. Sveinbjornsson, B., R. Olsen, O. M. Seternes, and R. Seljielid. 1996. Macrophage cytotoxicity against murine meth A sarcoma involves nitric oxidemediated apoptosis. Biochem. Biophys. Res. Commun. 223:643–649. McConkey, D. J., and S. Orrenius. 1996. Signal transduction pathways in apoptosis. Stem Cells (Dayt.) 14:619–631. Herrmann, J. L., A. W. Beham, M. Sarkiss, P. J. Chiao, M. T. Rands, E. M. Bruckheimer, S. Brisbay, and T. J. McDonnell. 1997. Bcl-2 suppresses apoptosis resulting from disruption of the NF-kB survival pathway. Exp. Cell Res. 237:101–109. Melkova, Z., S. B. Lee, D. Rodriguez, and M. Esteban. 1997. Bcl-2 prevents nitric oxide-mediated apoptosis and polyADP-ribose polymerase cleavage. FEBS Lett. 403:273–278. Desmouliere, A., M. Redard, I. Darby, and G. Gabbiani. 1995. Apoptosis mediates the decrease in cellularity during the transition between granulation tissue and scar. Am. J. Pathol. 146:56–66. Desmouliere, A., A. Geinoz, F. Gabbiani, and G. Gabbiani. 1993. Transforming growth factor-beta 1 induces alpha-smooth muscle actin expression in granulation tissue myofibroblasts and in quiescent and growing cultured fibroblasts. J. Cell Biol. 122:103–111. Ronnov-Jessen, L., and O. W. Petersen. 1993. Induction of alpha-smooth muscle actin by transforming growth factor-beta 1 in quiescent human breast gland fibroblasts: implications for myofibroblast generation in breast neoplasia. Lab. Invest. 68:696–707. Hautmann, M. B., C. S. Madsen, and G. K. Owens. 1997. A transforming growth factor-b (TGFb) control element drives TGFb induced stimulation of smooth muscle a-actin gene expression in concert with two CArG elements. J. Biol. Chem. 272:10948–10956.