Inhibition of TGF-b-stimulated CTGF gene expression and anchorage-independent growth by cAMP identifies a CTGF-dependent restriction point in the cell cycle DEVASHISH KOTHAPALLI, NOBUKAZU HAYASHI, AND GARY R. GROTENDORST1 Department of Cell Biology and Anatomy, University of Miami School of Medicine, Miami, Florida 33136, USA
CTGF is a 38 kDa cysteine-rich peptide whose synthesis and secretion are selectively induced by transforming growth factor b (TGF-b) in connective tissue cells. We have investigated the signaling pathways controlling the TGF-b induction of connective tissue growth factor (CTGF) gene expression. Our studies indicate that inhibitors of tyrosine kinases and protein kinase C do not block the signaling pathway used by TGF-b to induce CTGF gene expression. In contrast, elevation of cAMP levels within the target cells by a variety of methods blocked the induction of CTGF by TGF-b. Furthermore, agents that elevate cAMP blocked the induction of anchorage-independent growth (AIG) by TGF-b. Inhibition of AIG could be overcome by the addition of CTGF, indicating that it was not a general inhibition of growth but a selective inhibition of CTGF synthesis that is responsible for the inhibition of TGF-b-induced AIG by cAMP. Kinetic studies of the induction of DNA synthesis by CTGF in cells arrested by cAMP indicate that the block occurs in very late G1. These and other studies in monolayer cultures suggest that the CTGF restriction point in the cell cycle is distinct from the adhesion-dependent arrest point.—Kothapalli, D., Hayashi, N., Grotendorst, G. R. Inhibition of TGF-b-stimulated CTGF gene expression and anchorage-independent growth by cAMP identifies a CTGF-dependent restriction point in the cell cycle. FASEB J. 12, 1151–1161 (1998) ABSTRACT
Key Words: signal transduction · promoter · wound repair · fibrosis · growth regulation
TRANSFORMING GROWTH factor beta (TGF-b)2 is a general regulator of cellular activities with a wide variety of actions on numerous cell types. TGF-b plays a central role in the regulation of development, differentiation, and wound repair, as well as in the initiation and progression of many fibrotic disorders. TGF-b’s unusual ability to stimulate the proliferation of mesenchymal cells (1–3) and inhibit the proliferation of
endothelial and epithelial cells (4–7) is not shared by other growth stimulatory or inhibitory factors. Secretion of TGF-b by platelets (3), macrophages, and neutrophils (8, 9) serves to induce the cascade of events leading to connective tissue formation that is essential for normal wound repair and leads to the overproduction of connective tissue in fibrotic disorders. CTGF is a Mr 38,000 cysteine-rich, heparin binding peptide first identified in media conditioned by human umbilical vein endothelial cells (10). The TGFb and CTGF genes are coordinately regulated during normal tissue regeneration (11) and CTGF is induced by injection of TGF-b in skin (12). Both TGF-b and CTGF are overexpressed in the lesions of numerous fibrotic disorders in skin (13) and in internal organs (12, 14). CTGF gene expression is selectively induced by TGF-b in fibroblastic cells but not in other cell types (12, 15). A brief exposure of fibroblasts to TGF-b (1 h) is sufficient to induce a prolonged highlevel expression of the CTGF transcript (24–36 h) (15). The regulation of CTGF appears to be controlled primarily at the level of transcription and is dependent on the action of a specific TGF-b control element (TbRE) present in human and murine CTGF promoters (15). This element is distinct from other reported TGF-b-responsive elements and is not present in other genes regulated by TGF-b (15–17). Collectively, these aspects of CTGF gene expression strongly suggest that CTGF is likely to function as a 1
Correspondence: Department of Cell Biology and Anatomy, University of Miami School of Medicine, 1600 N.W. 10th Ave., Miami, FL 33136, USA. E-mail:
[email protected]. miami.edu 2 Abbreviations: TGF, transforming growth factor; AIG, anchorage-independent growth; ATCC, American Type Culture Collection; DMEM, Dulbecco’s modified Eagle’s medium; FBS, fetal bovine serum; TPA, tetradecanoyl phorbol acetate; CTX, cholera toxin; SDS, sodium dodecyl sulfate; TCA, trichloroacetic acid; PMA, phorbol myristate acetate; PTX, pertussis toxin; PCR, polymerase chain reaction; BSA, bovine serum albumin; PDGF, platelet-derived growth factor; FGF, fibroblast growth factor; EGF, epidermal growth factor; FBS, fetal bovine serum.
0892-6638/98/0012-1151/$02.25 Q FASEB
/ 382f 0014 Mp 1151 Thursday Aug 06 12:11 PM LP–FASEB 0014
1151
downstream mediator of TGF-b action on fibroblastic cell types. This hypothesis is supported by cell culture studies where we have shown that inhibition of CTGF synthesis or action with antisense nucleotides or specific antibodies can block TGF-b-stimulated, anchorage-independent growth (AIG) (18). These results suggest that CTGF could serve as a selective therapeutic target for the inhibition of TGF-b actions on connective tissue cells. During the course of investigations to identify the signaling pathway components involved in the regulation of CTGF gene expression, we have identified an intracellular mechanism for inhibiting the TGF-b induction of CTGF gene expression. Here we report on those studies that demonstrate that elevation of intracellular cAMP levels by various toxins or pharmacologic agents or treatment of cells with membrane-permeable analogs of cAMP block the induction of the CTGF gene by TGF-b. These agents can also block TGF-b-induced cell proliferation, indicating that components of the cell that regulate cAMP levels or cAMP-regulated kinases could serve as a new class of therapeutic targets for anti-fibrotic therapies, and that cAMP analogs could serve as a new class of potential anti-fibrotic drugs.
night. TGF-b (10 ng/ml) (Gibco BRL) was then added and incubated for 24 h. To inhibit cell signaling, we used kinase inhibitors such as herbimycin A (Gibco BRL), tetradecanoyl phorbol acetate (TPA) (Gibco BRL), 8-Br-cAMP (Boehringer-Mannheim), 8Br-cGMP (Sigma Chemical Co., St. Louis, Mo.), and cholera toxin (CTX)(Gibco BRL). The cells were treated with the inhibitors for 2 h before adding TGF-b. To carry out long-term exposure of cells to TPA, we treated the cells with the compounds for 24 h after transfection. Each transfection experiment was carried out in duplicate. Luciferase activity was measured by using a luciferase assay system (Promega) and a scintillation counter (Beckman LS6000SC-Beckman Instruments Inc., Fullerton, Calif.) in the single photon monitor mode. The transfection efficiency was monitored by b-galactosidase activity, which was measured with a chemiluminescent assay using Galacto-Light (Tropix Inc., Bedford, Mass.).
MATERIALS AND METHODS
Anchorage-independent growth assay
Cell lines and culture conditions
The AIG assays were performed essentially as described by Guadagno and Assoian (19). Cells were grown initially as monolayer cultures. Cultures were trypsinized at 80% confluence and seeded (51104/well in a 12-well tissue culture plate) for use in AIG assays in NRK growth medium containing epidermal growth factor (EGF) (1 ng/ml). NRK fibroblasts were made quiescent as described previously by Zhu and Assoian (20). The growth response of the cells was determined 3 days after plating on the agarose layer. Cells were labeled with 1 mCi/ml of [methyl-3H]thymidine (NEN, Boston, Mass.) for 24 h. The cells were recovered from the wells, transferred to microcentrifuge tubes, and centrifuged at 8000 rpm for 10 min. The cell pellet was then treated with 1 ml of ice-cold 5% trichloroacetic acid (TCA), 100 mg BSA (as a carrier), and incubated at 47C for 1 h. The tubes were spun at 14,000 rpm for 10 min and the precipitate obtained was washed twice with cold 5% TCA. DNA was solubilized in 200 ml of 0.1% SDS, 0.1 N NaOH, and the incorporation of [3H]thymidine was determined using a liquid scintillation counter (21). Soft agar colony forming assays were performed as described by DeLarco and Todaro (1).
NIH/3T3 cells were obtained from the American Type Culture Collection (ATCC) (Rockville, Md.). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) at 377C in an atmosphere of 5% CO2 and 90% air. NRK 49F fibroblasts were also obtained from ATCC and cultured in DMEM containing 2.5% FBS and 2.5% NU-Serum (Collaborative Biomedical Products, Bedford, Mass.) in a 5% CO2 containing atmosphere. The antisense CTGF gene bearing fibroblasts (NRK-ASCTGF) was maintained in DMEM containing 2.5% FBS, 2.5% NUS, and 0.5 mg/ml Geneticin (Gibco BRL) (18). Luciferase reporter gene assay To study the secondary signaling pathway involved in the TGFb induction of CTGF gene expression, we looked at the effect of various signaling pathway inhibitors on the CTGF promoter luciferase reporter construct. A fragment of the CTGF promoter (0823 to /74) was cloned into the Sac I - Xho I cloning site of PGL2-basic vector (Promega Corp., Madison, Wis.). A part of promoter (0275 to 0106) was made using polymerase chain reaction (PCR), then cloned into the Sac I - Xho I cloning site of PGL2 promoter (Promega). Point mutants were made by using a mutated primer for PCR and then cloned into Bsm I deleted CTGF promoter driving the pGL2-basic firefly luciferase vector. These constructs were confirmed by sequencing. NIH/3T3 cells were transfected with 2 pg of luciferase plasmid and 0.5 pg of pSV-b-galactosidase vector (Promega) in 6-well plates using 4 ml of LIPOFECTIN reagent (Gibco BRL, Gaithersburg, Md.) for 6 h in serum-free DMEM with ITS (Collaborative Biomedical Products, Bedford, Mass.), followed by incubation with serum-free DMEM and ITS over1152
Vol. 12
September 1998
Preparation of cell extracts and immunoblotting NRK fibroblasts were grown to confluence in 100 mm dishes and starved overnight in DMEM containing ITS (Becton Dickinson) and bovine serum albumin (BSA) (0.1 mg/ml). The cells were then treated with 8-Br-cAMP and TGF-b for 24 h. The conditioned media was collected and affinity purified using anti-CTGF bound Affi-gel Sepharose beads. The CTGF was eluted and fractionated on a 12% sodium dodecyl sulfate (SDS) -polyacrylamide gel and electrophoretically transferred to a nitrocellulose membrane. Immunoblots were detected as described previously using a rabbit anti-CTGF specific antibody diluted 1:1000 in the blocking buffer (18).
RESULTS CTGF expression is blocked by cholera toxin and 8-Br-cAMP The CTGF gene is directly regulated by signaling events induced in fibroblasts by TGF-b, since inhibitors of protein synthesis have no effect on the induc-
The FASEB Journal
/ 382f 0014 Mp 1152 Thursday Aug 06 12:11 PM LP–FASEB 0014
KOTHAPALLI ET AL.
tion (11, 15). To evaluate what secondary signaling pathway might be involved in the TGF-b induction of CTGF gene expression, we screened a variety of compounds that act as inhibitors of various signaling pathways. These studies were performed using the CTGF promoter (0823 to /74) luciferase reporter construct that had been transiently transfected into NIH/3T3 fibroblasts. Inhibitors were added to the culture media 2 h before the addition of TGF-b and cells were maintained in the presence of the inhibitor for the duration of the experiment (24 h). None of the inhibitors exhibited toxic effects on the cells during this period. Herbimycin, a potent tyrosine kinase inhibitor, had no effect on the induction of the CTGF gene at concentrations of up to 300 ng/ml, but did completely inhibit platelet-derived growth factor (PDGF) -induced cell proliferation (Fig. 1A). The tumor promoter phorbol myristate acetate (PMA), which modulates protein kinase C activity,
had no effect on CTGF gene expression either as an inducer directly or as an inhibitor of TGF-b induction of the CTGF gene (Fig. 1B). These results suggest that neither tyrosine kinases or protein kinase C is involved in the signaling pathway from the TGF-b receptor that leads to increased CTGF gene transcription. We then evaluated the effects of agents that would modulate components of the cAMP signaling pathways, including cholera toxin, forskolin, and cell membrane-permeable analogs of cyclic nucleotides. Our initial studies demonstrated that agents that elevate intracellular levels of cAMP, such as CTX, acted as potent inhibitors of TGF-b induction of the CTGF gene. Cholera toxin exhibited a dose-dependent inhibition of TGF-b-induced CTGF promoter activity, with concentrations as low as 0.1 ng/ml completely blocking TGF-b-induced luciferase activity (Fig. 1C). No effect on basal reporter activity was detected at
Figure 1. Effect of phorbol ester, herbimycin, cholera toxin, and cyclic nucleotides on CTGF gene expression/promoter activity. NIH/3T3 cells were grown to 50% confluence in DMEM and 10% FBS. They were all transfected with the CTGF promoter driving the expression of the pGL2 basic firefly luciferase vector. After 24 h in DMEM/ITS media, either PMA (A), herbimycin (B), cholera toxin (C), or 8-Br-cAMP/8-Br-cGMP (D) were added 2 h prior to the addition of TGF-b (10 ng/ml). Cells were incubated for 24 h and the luciferase activity determined. The mitotic index represents the percent of cells that underwent mitosis. cAMP INHIBITION OF TGF-b STIMULATION OF ANCHORAGE-INDEPENDENT GROWTH
/ 382f 0014 Mp 1153 Thursday Aug 06 12:11 PM LP–FASEB 0014
1153
sion. To determine whether the endogenous CTGF gene expression was blocked by elevation of intracellular cAMP, Western blots were performed on media conditioned by NRK fibroblasts (Fig. 2) treated with TGF-b in the presence or absence of 8Br-cAMP (1 mM). The results of these studies demonstrate that elevation of intracellular cAMP levels can completely block CTGF production induced by TGF-b in these cells. Similar results were obtained with CTX and forskolin (data not shown). These data confirm those obtained with the reporter assay using the CTGF promoter controlling luciferase expression, and suggest that these agents could be used to study TGF-b signaling via the CTGF pathway. Figure 2. cAMP inhibits TGF-b-induced CTGF expression. NRK fibroblasts were grown to confluence and starved in DMEM containing ITS and BSA (0.1 mg/ml) for 24 h. The cells were then treated with or without 8-Br-cAMP (1 mM) and TGF-b for 24 h. The conditioned media was then collected, affinity purified using anti-CTGF-bound Affi-gel Sepharose beads, run on a 12% SDS polyacrylamide gel, and immunoblotted with an anti-CTGF specific antibody. CTGF standard was 10 ng.
concentrations of CTX up to 1 mg/ml (Fig. 1C). These data indicate that cAMP levels may modulate the TGF-b inducibility of the CTGF gene. To verify this, the reporter assays were carried out using 8-BrcAMP and 8-Br-cGMP (Fig. 1D). The 8-Br-cAMP inhibited CTGF promoter activity in a dose-dependent fashion, with a concentration of 1 mM giving complete inhibition. Identical concentrations of 8Br-cGMP had no effect on the TGF-b induction of the CTGF promoter. Neither 8-Br-cAMP nor 8-BrcGMP had any effect on the basal activity of the CTGF promoter. These observations were confirmed by using dibutyryl derivatives of the cyclic nucleotides and forskolin, which indicated that the elevation of intracellular cAMP levels but not cGMP levels completely blocked the TGF-b induction of CTGF gene expres-
Cholera toxin and 8-Br-cAMP do not block TGF-binduced changes in cell morphology We monitored the appearance of cells treated with the various inhibitors and noted some significant changes in the morphology of cells treated with 8-BrcAMP. Individual cells in cultures of NRK fibroblasts exposed to 8-Br-cAMP assumed a more rounded morphology, developed numerous neurite-like pseudopodia, and appeared to be poorly attached to the culture dish (Fig. 3). These changes were not due to toxic effects of the compounds, as trypan blue exclusion tests indicated greater than 90% viability of the cultures. Furthermore, the addition of TGF-b to the cAMP-treated cultures reversed the morphological changes. There were only subtle differences between the appearance of control TGF-b-treated and 8-BrcAMP-treated cultures in the presence of TGF-b. The cells in TGF-b-treated cultures were more flattened than controls, well spread, and attached to the culture dish. Cells in these cultures had a more organized appearance, with bipolar cells in tight swirls in the monolayer (Fig. 3). Because TGF-b was able to induce a nearly identical morphology in both control
Figure 3. Effect of cAMP on the morphology of NRK fibroblasts in monolayer cultures. NRK fibroblasts were grown to 80% confluence as in panel A and then treated with 8-Br-cAMP (1 mM) (B), TGF-b (10 ng/ml) (C), or both (D) in serum-free DMEM for 4 h and then photographed (101).
1154
Vol. 12
September 1998
The FASEB Journal
/ 382f 0014 Mp 1154 Thursday Aug 06 12:11 PM LP–FASEB 0014
KOTHAPALLI ET AL.
and 8-Br-cAMP-treated cultures, the effects of TGF-b on cell morphology must occur independent of CTGF action. Similar changes were seen in NRK cells exposed to CTX and in NIH/3T3 cells treated with either 8-Br-cAMP or CTX, indicating that these effects are not limited to 8-Br-cAMP or NRK fibroblasts. These data indicate that elevation of intracellular cAMP levels in fibroblasts selectively blocks TGF-b actions that require the synthesis of CTGF, suggesting that these agents could be useful to help dissect the actions of TGF-b, which require CTGF. Cholera toxin and 8-Br-cAMP block TGF-b-induced, anchorage-independent growth of NRK fibroblasts by inhibition of CTGF synthesis We wanted to determine whether cholera toxin or 8Br-cAMP would block the growth stimulatory action of TGF-b on cultured fibroblasts. The induction of anchorage-independent growth of nontransformed fibroblasts in culture is the hallmark assay for TGF-b and was used to follow TGF-b activity during its initial isolation. Because other growth factors cannot substitute for TGF-b in this assay, AIG serves as a specific assay for TGF-b. Earlier studies indicated that blockade of CTGF synthesis or action inhibited TGF-b-induced AIG (18). Accordingly, NRK cells were treated with cAMP analogs or agents such as cholera toxin or forskolin that would elevate intracellular cAMP. The presence of cholera toxin in the media resulted in a dose-dependent inhibition of TGF-b-induced, anchorage-independent growth of NRK fibroblasts with concentrations as low as 2.5 ng/ml, causing a measurable reduction in growth. TGF-b-induced AIG was nearly completely inhibited by cholera toxin at 20 ng/ml, with 5 ng/ml giving greater than 50% inhibition compared to control cultures (Fig. 4A). The 8Br-cAMP also inhibited TGF-b-induced AIG in a dose-dependent fashion, with complete inhibition at concentrations of 1 mM (Fig. 4B). In contrast, 8-BrcGMP exhibited no effect in the AIG assay at concentrations of up to 10 mM (Fig. 4B). Similarly, forskolin (an adenylate cyclase activator) was also active as an inhibitor of TGF-b-induced AIG with 20 ng/ml, causing a half-maximal inhibition and maximal inhibitory activity at concentrations of 50 ng/ml (Fig. 4C). These results indicate that agents that elevate intracellular cAMP or cell membrane-permeable analogs of cAMP can function as effective inhibitors of TGFb-induced AIG. Because cAMP inhibits TGF-b induction of CTGF, which is required for AIG (18), we wanted to determine whether the inhibition of TGF-b-induced AIG could be overcome by addition of pure recombinant CTGF to the cell cultures. This would demonstrate that the inhibition was due solely to the blockade of CTGF gene expression and not to a more general inhibition of cell proliferation. To accomplish this,
we performed experiments where rCTGF was added to TGF-b-treated NRK cells that were growth arrested by 8-Br-cAMP or CTX in suspension culture. The addition of pure rCTGF stimulated DNA synthesis in the TGF-b-treated cell cultures whose growth had been blocked with 8Br-cAMP (Fig. 4D). Neither PDGF nor fibroblast growth factor (FGF), at concentrations that are maximally active in growth arrested monolayer cultures, were effective in stimulating DNA synthesis in TGF-b-treated cells blocked with 8Br-cAMP (Fig. 4D). The CTGF stimulation of growth was dose dependent, with maximal DNA synthesis induced at CTGF concentrations of 20 ng/ml (0.6 nM) and half-maximal activity at 4 ng/ml (0.12 nM) (Fig. 4E). To determine whether CTGF could restore the growth of the NRK cells for a longer period of time, we performed colony assays to measure AIG. Cells were maintained for a period of 10 days in complete media containing either TGF-b (2 ng/ml), TGF-b and CTX (10 ng/ml), or TGF-b, CTX, and CTGF (10 ng/ml). Figure 5 contains photomicrographs of representative fields of the cultures that indicate the size and number of colonies that formed under the various culture conditions. In the positive control (cultures treated with TGF-b), large numbers of colonies greater than 50 mm in diameter are present (Fig. 5A). In contrast, the addition of CTX inhibits the formation of any large colonies and only individual cells or small colonies are present (Fig. 5B). Addition of CTGF to the media restored the growth of the NRK cells in the presence of the CTX (Fig. 5C), and these cultures displayed even larger colonies than the positive control TGF-b-treated cultures. These data demonstrate that the inhibition of cell growth caused by elevation of intracellular cAMP under these culture conditions is due solely to inhibition of CTGF synthesis, as no inhibition of cell growth was observed when cells were provided with CTGF in the presence of maximally inhibitory concentrations of cholera toxin or 8-Br-cAMP. The cAMP arrest of TGF-b-induced AIG identifies a CTGF-dependent G1 restriction point Previous studies have demonstrated that adhesion regulates progression of fibroblasts through a restriction point that is approximately in the middle of the G1 phase of the cell cycle (22). This restriction point can be overcome in suspension by the presence of TGF-b (23), and we wanted to determine whether elevation of intracellular cAMP prevented TGF-btreated cells from progressing from this restriction point or whether they were arrested at a different point in the cell cycle. To test this, we compared the kinetics of S-phase entry of the cells by measuring DNA synthesis based on cultures that were arrested by suspension (19) or by 1 mM 8-Br-cAMP in the presence of TGF-b. Quiescent NRK fibroblasts were main-
cAMP INHIBITION OF TGF-b STIMULATION OF ANCHORAGE-INDEPENDENT GROWTH
/ 382f 0014 Mp 1155 Thursday Aug 06 12:12 PM LP–FASEB 0014
1155
Figure 4. Inhibition of TGF-b-induced, anchorage-independent growth of fibroblasts by cholera toxin, cyclic nucleotides and forskolin and its reversal by CTGF. NRK fibroblasts were grown in AIG conditions in the presence of either cholera toxin (A), 8-BrcAMP/8-Br-cGMP (B), or forskolin (C) and TGF-b (5 ng/ml) and then processed for [3H]thymidine incorporation. D) Anchorageindependent growth of NRK fibroblasts was inhibited using 8-BrcAMP (1 mM) and TGF-b. In addition, FGF (20 ng/ml), PDGFBB (20 ng/ml), or rCTGF (20 ng/ml) were also present for the length of the assay. E) NRK cells were treated with TGF-b, 8-BrcAMP, and increasing concentrations of CTGF in suspension. [3H]Thymidine incorporation was then measured as described in the text. Assays were performed in duplicate; bars, SD.
tained in complete media [(DMEM/10% fetal bovine serum (FBS)] in suspension culture supplemented with either EGF (1 ng/ml) without any 8-Br-cAMP or TGF-b (2.5 ng/ml) and EGF (1 ng/ml) in the presence of 1 mM 8-Br-cAMP for 16 h. Cells from cultures arrested under both conditions were then returned to monolayer culture in fresh complete media (DMEM/10% FBS supplemented with EGF) and pulse-labeled for 1 h at various intervals after plating with [3H]thymidine (2 mCi/ml) to measure DNA synthesis. Cells that had been arrested in suspension cul1156
Vol. 12
September 1998
ture in the presence of EGF alone required 5–6 h to enter S-phase, a time comparable to that observed by Guadagno and Assoian (19) (Fig. 6A). In contrast, cells arrested in suspension in the presence of TGFb and 8-Br-cAMP began DNA synthesis 2–3 h after return to monolayer culture in the absence of 8-BrcAMP (Fig. 6A). A third set of cultures arrested in the presence of TGF-b and 8-Br-cAMP remained in suspension culture, but were stimulated with CTGF (Fig. 6A). Cells in these cultures also exhibited a shortened kinetic of S-phase entry similar to those cells that
The FASEB Journal
/ 382f 0014 Mp 1156 Thursday Aug 06 12:12 PM LP–FASEB 0014
KOTHAPALLI ET AL.
Figure 5. Effect of TGF-b, CTX, and CTGF on NRK fibroblasts in soft agar cultures/assays. NRK fibroblasts were grown in soft agar in the presence of TGF-b (2.5 ng/ml) (A), TGF-b and CTX (10 ng/ml) (B), or TGF-b, CTX, and rCTGF (20 ng/ml) (C) over a period of 10 days and then photographed (201).
were starved in the presence of cAMP but returned to monolayer in its absence. They began DNA synthesis within 2–3 h after addition of CTGF to the media. These data indicate the existence of a second adhesion-dependent restriction point in late G1, which can be overcome by the presence of CTGF in the media when the cells are in suspension or by providing the cells with an adhesive surface in the absence of 8-Br-cAMP. The CTGF dependence of this restriction point was confirmed by evaluating the kinetics of S-phase entry of a clone of NRK fibroblasts containing an antisense CTGF gene (NRK-ASCTGF). We have previously reported that these cells do not synthesize CTGF in response to TGF-b stimulation and are refractory to TGF-b as an inducer of AIG (18). These cells behave similarly to normal NRK fibroblasts in monolayer culture, but require the addition of both TGF-b and CTGF for growth in suspension. One set of cultures of quiescent NRK-ASCTGF fibroblasts was suspended in media (DMEM/10% FBS) containing EGF (1 ng/ ml), and two sets of cultures were arrested in suspension culture in the presence of EGF (1 ng/ml) and TGF-b (2.5 ng/ml) for 16 h. The suspension cultures treated with EGF, as well as one set of cultures that were arrested with EGF and TGF-b, were returned to monolayer in fresh DMEM/10% FBS supplemented with EGF (1 ng/ml) and pulse-labeled with [3H]thymidine (2 mCi/ml) for 1 h at the indicated
time points in order to measure DNA synthesis. The second set of suspension cultures treated with EGF and TGF-b were maintained in suspension culture, treated with CTGF, and pulse-labeled with [3H]thymidine at the same time points. Cells that had been suspension arrested in the presence of EGF alone required at least 5–6 h to enter S-phase, similar to the parental NRK cells cultured under the same conditions (Fig. 6B). However, NRK-ASCTGF cells arrested in suspension in the presence of EGF and TGF-b began DNA synthesis within 2–3 h after their return to monolayer culture in the absence of TGFb (Fig. 6B), similar to the parental NRK cells treated with EGF and TGF-b in the presence of 1 mM 8-BrcAMP. The cells that were arrested in the presence of EGF and TGF-b and stimulated with CTGF in suspension also had increased rates of [3H]thymidine incorporation 2–3 h after the addition of CTGF to the media (Fig. 6B), as seen with the 1 mM 8-BrcAMP-blocked NRK cells. The reason for the differences in the rates of DNA synthesis between these two sets of cultures is due to the continued synthesis of the antisense CTGF RNA in monolayer after the removal of TGF-b. These results support the observations made with the normal NRK cells blocked with cAMP in the presence of TGF-b and indicate the presence of a CTGF-dependent restriction point in the cell cycle. They also indicate that TGF-b induces some changes independently of CTGF that bring the cells closer to S-phase entry. We wanted to determine whether the CTGF-dependent restriction point could be overcome by adhesion alone. In the previous studies, we had removed the inhibitory blocks (8-Br-cAMP/CTGF antisense RNA) when cells were allowed to attach. This was essential as the attachment of the cells proceeded very slowly in the presence of 8-Br-cAMP or if the antisense CTGF gene was continually expressed by cells in suspension. To circumvent these problems, we set up monolayer cultures that remained attached throughout the course of treatment and assay. Cultures of parental NRK and NRKASCTGF were grown to confluence in complete media and shifted to DMEM/ITS for 3 days to ensure growth arrest. The cultures were treated with TGF-b and EGF overnight; in the case of the parental cells, 1 mM 8-Br-cAMP was added at the time of growth factor addition to ensure blockade of CTGF production. CTGF was then added to the culture media and DNA synthesis was determined by pulse labeling with [3H]thymidine as described in the above experiments. Cultures with no CTGF added served as negative controls for basal levels of S-phase entry under these conditions. Both the parental NRK cells in the presence of 1 mM 8-Br-cAMP and the NRK-ASCTGF cells exhibited very low levels of DNA synthesis 24 h after the addition of growth factors as compared to parental NRK cells activated
cAMP INHIBITION OF TGF-b STIMULATION OF ANCHORAGE-INDEPENDENT GROWTH
/ 382f 0014 Mp 1157 Thursday Aug 06 12:12 PM LP–FASEB 0014
1157
Figure 6. CTGF can reverse cAMP inhibition of TGF-b-induced, anchorage-independent growth of fibroblasts in a cell cycle-dependent manner. NRK (A) or NRK-ASCTGF (B) fibroblasts were grown to quiescence, trypsinized, and transferred (2.51106 cells/well) to suspension cultures in 12-well plates. They were treated with TGF-b and 8-Br-cAMP (1 mM) overnight and then with CTGF (25 ng/ml) in suspension or returned to monolayer for the indicated time period before being pulsed with [3H]thymidine (2 mCi/ml) for 2 h. Some cells were treated overnight with EGF (1 ng/ml) instead of TGF-b and 8-Br-cAMP and returned to monolayer. (C) NRK and NRK-ASCTGF fibroblasts were grown to confluence, then serum starved in DMEM/ITS for 3 days, and treated overnight with EGF and TGF-b (NRK-ASCTGF cells) or TGF-b (2.5 ng/ml) and 8-Br-cAMP (NRK cells). rCTGF was added to the cultures for the time period indicated before the cells were pulse-labeled as described above. Each experimental time point was assayed in duplicate and the experiments were repeated three times; bars, SD.
with growth factors in the absence of inhibitors, which represents maximal DNA synthesis as plotted in Fig. 6C. Addition of CTGF to the culture media of either growth-inhibited culture stimulated a significant and rapid induction of DNA synthesis, with kinetics similar to that observed in the suspension culture experiments. Cells from both sets of cultures began DNA synthesis within 2 h after the addition of CTGF to the media, which continued and increased to near maximal levels 6–8 h after growth factor addition. These data demonstrate that the TGF-b-stimulated DNA synthesis induced by 8-BrcAMP or antisense CTGF RNAs cannot be reversed by adhesion, but specifically requires the presence of CTGF or the removal of the inhibitory factor, allowing endogenous CTGF to be produced. Furthermore, these data indicate the existence of a CTGFdependent restriction point in the cell cycle that is distinct from growth factor or adhesion-dependent restriction points described previously. 1158
Vol. 12
September 1998
DISCUSSION Collectively, fibrotic disorders represent the largest segment of human disease that remain intractable to drug therapy. These disorders are characterized by an overproduction of connective tissue (primarily collagen), resulting in a scar that impedes the function of the normal tissues composing the affected organ or tissue. Numerous factors are involved in the initiation and progression of fibrotic disorders. Nonetheless, it is well recognized that overproduction of active TGF-b is a common observation in all of these disorders. Consequently, there has been an intensive effort to identify agents that could selectively block the biological actions of TGF-b, much of which has focused on TGF-b signaling pathways. Because CTGF is selectively induced by TGF-b in connective tissue cells, and blockade of CTGF expression or action can prevent TGF-b-induced connective tissue cell proliferation, we wanted to identify the com-
The FASEB Journal
/ 382f 0014 Mp 1158 Thursday Aug 06 12:12 PM LP–FASEB 0014
KOTHAPALLI ET AL.
ponents of the signaling pathway used by TGF-b to induce CTGF gene expression. Once identified, these components could serve as specific targets for anti-fibrotic drug development. To begin an investigation into this TGF-b signaling pathway, we have evaluated the effects of various compounds that modulate the action of identified signaling pathway components. Our initial assays to evaluate compounds took advantage of a CTGF promoter firefly luciferase reporter gene that behaves identically to the endogenous gene with respect to growth factor induction and cell specificity (15). This chimeric gene routinely exhibits a 20- to 30-fold increase in luciferase activity in fibroblastic cells treated with TGF-b compared to nontreated cells. Because of this large-fold induction and the highly reproducible effects of TGF-b in this system, this assay provides a convenient and simple method for analysis of inhibitors. Initially, we analyzed compounds that affected tyrosine kinases, protein kinase C, and protein kinase A. Since the TGF-b receptor acts via a serine-threonine kinase domain, herbimycin A, an inhibitor of tyrosine kinases, had no effect on the induction of the CTGF gene by TGF-b. Nonetheless, at identical concentrations, herbimycin was an effective inhibitor of PDGFstimulated DNA synthesis, demonstrating that it is biologically active in this system. Similarly, the tumor promoter, phorbol myristate acetate, a potent activator of protein kinase C, exhibited no effect as an inducer of the CTGF gene. Furthermore, long-term exposure of the cells to PMA, which causes a depletion of C-kinase, also had no effect on the induction of the CTGF gene by TGF-b. These results distinguish the regulation of the CTGF gene by TGF-b from the regulation of cef10/ cyr-61, which is modulated by phorbol ester (24, 25) but weakly induced by TGF-b (15). They also indicate that C-kinase is not a part of the signaling pathway used by TGF-b to induce CTGF nor does it modulate the activity of a component of this signaling pathway. This further distinguishes the TGF-b receptor signaling pathway that activates CTGF gene expression from the regulatory pathways that control other members of the CCN gene family (26–28). In contrast to the compounds mentioned above, agents that modulated the activity of protein kinase A had significant effects on the TGF-b induction of CTGF gene expression. Treatment of the cells with cholera toxin completely abolished induction of the CTGF gene by TGF-b. However, cholera toxin had no effect on the basal expression of the CTGF gene either as an inducer or suppressor. Other agents such as pertussis toxin, forskolin, and IBMX, which elevate cAMP levels in cells, were effective inhibitors of CTGF gene induction by TGF-b. Proof that these effects are mediated by cAMP comes from the results of experiments using a number of cell membrane-permeable analogs of cAMP including dibutyryl-cAMP, 8-Br-
cAMP, and benzyl cAMP. All of these cAMP analogs were effective inhibitors of CTGF gene induction, demonstrating that the actions of cholera toxin are via a cAMP-dependent mechanism. No effects were detected on the basal level of expression of the CTGF gene by cholera toxin or by cAMP analogs. Analogs of cGMP had no effect on the induction of CTGF by TGF-b, confirming the specificity of the cAMP analogs. The inhibition of CTGF gene induction is not unique to the transient expression of the transgene in target cells, as stable lines of NIH/3T3 cells expressing the gene show similar behavior (A. Leask, FibroGen, S. San Francisco, Calif.) and the induction of the endogenous CTGF gene is also blocked by these compounds. We previously demonstrated that inhibition of CTGF synthesis by antisense RNA can block TGF-binduced AIG of NRK fibroblasts (18). Here we find that agents that elevate cAMP in the NRK cells such as cholera toxin or forskolin are also effective inhibitors of TGF-b-induced AIG. This inhibition is mediated by cAMP since cell-permeable analogs of cAMP such as dibutyryl cAMP or 8-Br-cAMP were also effective inhibitors. Addition of recombinant CTGF to the cell cultures treated with either cholera toxin or 8Br-cAMP overcame the inhibition and allowed the cells to complete the cell cycle and replicate. Addition of other growth factors such as PDGF or FGF did not reverse the inhibition of AIG caused by cholera toxin or 8-Br-cAMP. These findings demonstrate that these agents are selectively interfering with the induction but not the action of CTGF and that the effects are not due to toxicity or a general inhibition of cell proliferation. Previous studies by several investigators have indicated that elevation of cAMP in target cells can antagonize the effects of TGF-b on gene expression. For example, Thalacker and Nilsen-Hamilton (29) reported that elevation of cAMP with cholera toxin or forskolin or the addition of dibutyryl-cAMP suppresses the absolute level of PAI-1 gene expression, but does not alter the fold induction by TGF-b. Howe et al. (30) reported that in AKR-2B fibroblasts, pertussis toxin reduced the TGF-b induction of growth regulatory genes such as c-sis, c-myc, and c-fos, but had no effect on the TGF-b induction of matrix protein genes such as fibronectin or collagen or on TGFb-induced morphologic changes. In AKR-2B cells, pertussis toxin (PTX) was found to be inhibitory to TGF-b-induced AIG in the soft agar assay but not in a monolayer assay, whereas cholera toxin was inhibitory in a monolayer assay of TGF-b-stimulated growth but not in the AIG assay (31). In our experiments with the NRK fibroblasts, we found that CTX was more effective than PTX on the inhibition of growth in both the soft agar assay and a monolayer growth assay. These effects are being mediated by elevated levels of cAMP, as both forskolin and 8-Br-cAMP are
cAMP INHIBITION OF TGF-b STIMULATION OF ANCHORAGE-INDEPENDENT GROWTH
/ 382f 0014 Mp 1159 Thursday Aug 06 12:12 PM LP–FASEB 0014
1159
potent inhibitors of TGF-b-induced proliferation of NRK fibroblasts. As seen in the AKR-2B cells, not all effects of TGF-b on the fibroblasts are affected by elevated levels of cAMP, which is most clearly demonstrated by our studies examining cell morphology where elevated levels of cAMP had no effect on the morphologic changes induced by TGF-b in monolayer cultures of NRK fibroblasts. Because elevation of cAMP blocks CTGF gene expression and the addition of pure recombinant CTGF can overcome the inhibition of cell proliferation, these results confirm and support the hypothesis for a requirement of TGF-b activation of both CTGF-dependent and CTGF-independent pathways to induce of AIG in NRK fibroblasts (18). Our results show that TGF-b-induced AIG is arrested by cAMP in late G1 and that this arrest can be reversed only in the presence of CTGF. This suggests the existence of a late G1 restriction point that is CTGF dependent (Fig. 7). Assoian et al. (22) have reported an anchorage-dependent restriction point 5–6 h from G1/S, which is overcome by the presence of TGF-b in the suspension culture media. This restriction point coincides with the suspension arrest point we see in our experiments in the absence of cAMP. Several lines of evidence indicate that the CTGF-dependent arrest point is distinct from arrest points previously reported (19, 32). First, we find that elevation of intracellular cAMP can effectively block the growth of TGF-b-stimulated NRK cells in either suspension or monolayer culture. This is distinct from the adhesion arrest point where cell adhesion to the tissue culture surface alone can overcome the growth arrest (23). Second, the kinetics of S-phase entry are shortened considerably for cells treated with TGF-b and 8-Br-cAMP and restimulated with CTGF in either suspension or monolayer culture, compared to cells arrested at the adhesion-dependent arrest point. Last, growth arrest is selectively overcome by the addition of CTGF, with other growth factors having no effect. Collectively, these data demonstrate the existence of a CTGF-dependent cell cycle restriction point in the late G1 phase. This restriction point seems to manifest itself only when cells have been activated with TGF-b. For example, the NRK-ASCTGF clones replicate in a manner similar to the parental NRK cells when maintained in normal culture media supplemented with serum, but are growth inhibited when stimulated with TGF-b. In contrast, the parental NRK cells exhibit DNA synthesis in response to TGF-b stimulation. In the parental NRK cells, we find that elevation of intracellular cAMP acts as a much more effective inhibitor of TGF-b-stimulated cell proliferation than of PDGF-stimulated cell proliferation, suggesting that the inhibition is selective for TGF-b-activated cells. TGF-b’s inhibitory actions are normally restricted to epithelial cells, which do not express the CTGF gene 1160
Vol. 12
September 1998
Figure 7. A schematic diagram of the relative positions of growth factor-dependent arrest points in the G1 phase of the fibroblast cell cycle.
in response to TGF-b and do not exhibit a response to CTGF as a mitogen (12). Potentially, the inhibitory actions of TGF-b in fibroblasts may be masked by the ability of the cells to produce and respond to CTGF. This raises the possibility that TGF-b’s inhibitory actions are induced in all cell types, with CTGF acting to override or bypass this inhibitory effect in cells that express functional CTGF receptors. Both cAMP and TGF-b have been reported to elevate kinase inhibitory proteins, which can inhibit cyclin-dependent kinases and arrest the cell cycle in mid to late G1 (33, 34). This work has been done with macrophages and epithelial cells and may not pertain to fibroblastic cells. Whether these same effects are occurring in the NRK cells is an area of active investigation. In summary, the data presented here demonstrate that elevation of cAMP levels in fibroblasts can block TGF-b-induced expression of CTGF and subsequent cell proliferation. This inhibition appears to be specific for TGF-b-induced growth. These characteristics of the inhibition of TGF-b-stimulated cell proliferation suggest that cAMP analogs or agents that could selectively elevate cAMP levels in connective tissue cells at sites of fibrosis could potentially serve as antifibrotic therapeutics. Because TGF-b is overexpressed and may be the dominant mitogen present, this could make the cells more sensitive to inhibitory effects of elevated levels of cAMP. This would allow for the selective inhibition of cells in the lesion. Thus, CTGF may provide a target for developing therapeutic agents, which would be potentially useful in the treatment of fibrotic conditions. The authors would like to thank Dr. Richard Assoian for constructive advice and criticism during the course of this work. The studies reported here were supported by grants from the National Institutes of Health (GM37223) and FibroGen, Inc., to G.R.G.
REFERENCES 1. 2.
DeLarco, J. E., and Todaro, G. J. (1978) Growth factors from murine sarcoma virus-transformed cells. Proc. Natl. Acad. Sci. USA 75, 4001–4005 Childs, C. B., Proper, J. A., Tucker, R. F., and Moses, H. L. (1982) Serum contains a platelet-derived transforming growth factor. Proc. Natl. Acad. Sci. USA 79, 5312–5316
The FASEB Journal
/ 382f 0014 Mp 1160 Thursday Aug 06 12:12 PM LP–FASEB 0014
KOTHAPALLI ET AL.
3.
4.
5. 6.
7. 8.
9. 10.
11.
12.
13.
14.
15.
16. 17. 18.
Assoian, R. K., Komoriya, A., Meyers, C. A., Miller, D. M., and Sporn, M. B. (1983) Transforming growth factor-beta in human platelets: identification of a major storage site, purification, and characterization. J. Biol. Chem. 258, 7155–7160 Coffey, R. J., Jr., Bascom, C. C., Sipes, N. J., Graves-Deal, R., Weissman, B. E., and Moses, H. L. (1988) Selective inhibition of growth-related gene expression in murine keratinocytes by transforming growth factor beta. Mol. Cell. Biol. 8, 3088–3093 Heimark, R. L., Twardzik, D. R., and Schwartz, S. M. (1986) Inhibition of endothelial regeneration by type-beta transforming growth factor from platelets. Science 233, 1078–1080 Takehara, K, LeRoy, E. C., and Grotendorst, G. R. (1987) TGFb inhibition of endothelial cell proliferation: alteration of EGF binding and EGF-induced growth-regulatory (competence) gene expression. Cell 49, 415–422 Tucker, R. F., Shipley, G. D., Moses, H. L., and Holley, R. W. (1984) Growth inhibitor from BSC-1 cells closely related to platelet type beta transforming growth factor. Science 226, 705–707 Assoian, R. K., Fleurdelys, B. E., Stevenson, H. C., Miller, P. J., Madtes, D. K., Raines, E. W., Ross, R., and Sporn, M. B. (1987) Expression and secretion of type beta transforming growth factor by activated human macrophages. Proc. Natl. Acad. Sci. USA 84, 6020–6024 Grotendorst, G. R., Smale, G., and Pencev, D. (1989) Production of transforming growth factor beta by human peripheral blood monocytes and neutrophils. J. Cell. Physiol. 140, 396–402 Bradham, D. M., Igarashi, A., Potter, R. L., and Grotendorst, G. R. (1991) Connective tissue growth factor: a cysteinerich mitogen secreted by human vascular endothelial cells is related to the SRCinduced immediate early gene product CEF10. J. Cell Biol. 114, 1285–1294 Igarashi, A., Okochi, H., Bradham, D. M., and Grotendorst, G. R. (1993) Regulation of connective tissue growth factor gene expression in human skin fibroblasts and during wound repair. Mol. Biol. Cell 4, 637–645 Frazier, K., Williams, S., Kothapalli, D., Klapper, H., and Grotendorst, G. R. (1996) Stimulation of fibroblast cell growth , matrix production and granulation tissue formation by connective tissue growth factor. J. Invest. Dermatol. 107, 404–411 Igarashi, A., Nashiro, K., Kikuchi, K., Sato, S., Ihn, H., Fujimoto, M., Grotendorst, G. R., and Takehara, K. (1996) Connective tissue growth factor gene expression in tissue sections from localized scleroderma, keloid and other fibrotic skin disorders. J. Invest. Dermatol. 106, 729–733 Hammes, M. S., Lieske, J. C., Pawar, S., Spargo, B. H., and Toback, F. G. (1995) Calcium oxalate monohydrate crystals stimulate gene expression in renal epithelial cells. Kidney Int. 48, 501– 509 Grotendorst, G. R., Okochi, H., and Hayashi, N. (1996) A novel transforming growth factor b response element controls the expression of the connective tissue growth factor gene. Cell Growth Differ. 7, 469–480 Edwards, D. R., and Heath, J. K. (1991) Regulation of transcription by transforming growth factor-b. In The Hormonal Control of Regulation of Gene Transcription, Vol. 16, Elsevier, Amsterdam Brunner, A. L., Chinn, J., Neubauer, M., and Purchio, A. F. (1991) Identification of a gene family regulated by transforming growth factor-beta. DNA Cell Biol. 10, 293–300 Kothapalli, D., Frazier, K. S., Welply, A., Segarini, P. R., and Grotendorst, G. R. (1997) TGF-b induces anchorage-independent
19. 20. 21.
22.
23.
24. 25. 26.
27.
28. 29.
30.
31.
32. 33.
34.
growth of NRK fibroblasts via a CTGF-dependent signaling pathway. Cell Growth Differ. 8, 61–68 Guadagno, T. M., and Assoian, R. K. (1991) G1/S control of anchorage-independent growth in the fibroblast cell cycle. J. Cell Biol. 115, 1419–1425 Zhu, X., and Assoian, R. K. (1995) Integrin-dependent activation of MAP kinase: a link to shape-dependent cell proliferation. Mol. Biol. Cell 6, 273–282 Assoian, R. K., Boardman, L. A., and Drosinos, S. (1989) A preparative suspension culture system permitting quantitation of anchorage-independent growth by direct radiolabeling of cellular DNA. Anal. Biochem. 177, 95–99 Assoian, R. K., Guadagno, T. M., and Dalton, S. L. (1993) Attachment control of fibroblast proliferation. In Cell Biology and Biotechnology: Novel Approaches to Increase Cellular Productivity, Springer-Verlag, New York Han, E. K., Guadagno, T. M., Dalton, S. L., and Assoian, R. K. (1993) A cell cycle and mutational analysis of anchorage-independent growth: cell adhesion and TGF-b1 control G1/S transit specifically. J. Cell Biol. 122, 461–471 Simmons, D. L., Levy, D. B., Yannoni, Y., and Erikson, R. L. (1989) Identification of a phorbol ester-repressible v-src-inducible gene. Proc. Natl. Acad. Sci. USA 86, 1178–1182 Latinkic, B. V., O’Brien, T. P., and Lau, L. F. (1991) Promoter function and structure of the growth factor inducible immediate early gene cyr6l. Nucl. Acids Res. 19, 3261–3267 Ryseck, R. P., Macdonald-Bravo, H., Mattei, M. G., and Bravo, R. (1991) Structure, mapping, and expression of fisp-12, a growth factor-inducible gene encoding a secreted cysteine-rich protein. Cell Growth Differ. 2, 225–233 Joliot, V., Martinerie, C., Dambrine, G., Plassiart, G., Brisal, B., Crochet, J., and Perbal, B. (1992) Proviral rearrangements and overexpression of a new cellular gene (nov) in myeloblastosisassociated virus type 1-induced nephroblastomas. Mol. Cell. Biol. 12, 10–21 Bork, P. (1993) The modular architecture of a new family of growth regulators related to connective tissue growth factor. FEBS Lett. 327, 125–130 Thalacker, F. W., and Nilsen-Hamilton, M. (1992) Opposite and independent actions of cyclic AMP and transforming growth factor b in the regulation of type 1 plasminogen activator inhibitor expression. Biochem. J. 287, 855–862 Howe, P. H., Cunningham, M. R., and Leof, E. B. (1990) Distinct pathways regulate transforming growth factor b1-stimulated proto-oncogene and extracellular matrix gene expression. J. Cell. Physiol. 142, 39–45 Howe, P. H., Bascom, C. C., Cunningham, M. R., and Leof, E. B. (1989) Regulation of transforming growth factor b1 action by multiple transducing pathways: evidence for both G protein-dependent and -independent signaling. Cancer Res. 49, 6024–6031 Pardee, A. (1989) G1 events and regulation of cell proliferation. Science 246, 603–608 Kato, J. Y., Matsuoka, M., Polyak, K., Massague, J., and Sherr, C. J. (1994) Cyclic AMP-induced G1 phase arrest mediated by an inhibitor (p27Kip1) of cyclin-dependent kinase 4 activation. Cell 79, 487–496 Reynisdottir, I., Polyak, K., Iavarone, A., and Massague, J. (1995) Kip/Cip and Ink4 Cdk inhibitors cooperate to induce cell cycle arrest in response to TGFb. Genes & Dev. 9, 1831–1845
cAMP INHIBITION OF TGF-b STIMULATION OF ANCHORAGE-INDEPENDENT GROWTH
/ 382f 0014 Mp 1161 Thursday Aug 06 12:12 PM LP–FASEB 0014
Received for publication January 27, 1998. Revised for publication April 16, 1998.
1161