Insulin-like growth factor I (IGF-I) and transforming growth factor-β1 (TGF-β1) are upregulated in myofibroblasts at sites of fibrosis in experimental enterocolitis and in Crohn's disease (CD). We compared the sites of expression of IGF-I and TGF-β1 in a rat peptidoglycan-polysaccharide (PG-PS) model of chronic granulomatous enterocolitis and fibrosis. We used the human colonic CCD-18Co fibroblast/myofibroblast cell line to test the hypothesis that TGF-β1 and IGF-I interact to regulate proliferation, collagen synthesis, and activated phenotype typified by expression of α-smooth muscle actin and organization into stress fibers. IGF-I potently stimulated while TGF-β1 inhibited basal DNA synthesis. TGF-β1 and IGF-I each had similar but not additive effects to induce type I collagen. TGF-β1 but not IGF-I potently stimulated expression of α-smooth muscle actin and stress fiber formation. IGF-I in combination with TGF-β1 attenuated stress fiber formation without reducing α-smooth muscle actin expression. Stress fibers were not a prerequisite for increased collagen synthesis. TGF-β1 upregulated IGF-I mRNA, which led us to examine the effects of IGF-I in cells previously activated by TGF-β1 pretreatment. IGF-I potently stimulated proliferation of TGF-β1-activated myofibroblasts without reversing activated fibrogenic phenotype. We conclude that TGF-β1 and IGF-I both stimulate type I collagen synthesis but have differential effects on activated phenotype and proliferation. We propose that during intestinal inflammation, regulation of activated phenotype and proliferation may require sequential actions of TGF-β1 and IGF-I, but they may act in concert to increase collagen deposition.
- Crohn's disease
- α-smooth muscle actin
fibrosis is a serious complication of Crohn's disease (CD) that involves excessive transmural deposition of collagen and disorganized overgrowth of the intestinal smooth muscle layers (7, 13, 14, 34).
Activated mesenchymal cells with myofibroblast phenotype, characterized by the presence of α-smooth muscle actin organized in prominent stress fibers (19, 20, 25, 26), play a major role in wound healing and fibrosis in a number of organs (19, 20). Transient appearance of activated myofibroblasts is a feature of normal wound healing (19, 20, 26), but the persistence of these cell types is associated with excessive collagen deposition and fibrosis. In some organs, like the liver, it is generally accepted that activation of α-smooth muscle actin-negative hepatic stellate cells to an α-smooth muscle actin-positive phenotype and expansion of this modified cell population lead to chronic liver fibrosis (5). Myofibroblasts and fibroblasts are increased in areas of fibrosis in the involved regions of the intestine in patients with CD (22) and in a rat model of peptidoglycan-polysaccharide (PG-PS)-induced granulomatous enterocolitis and fibrosis (31). This provides evidence that myofibroblasts mediate or contribute to fibrosis associated with chronic intestinal inflammation (20, 22). In the intestine, however, the origin of myofibroblasts present at sites of fibrosis is unknown. They may derive from resident collagen producing fibroblasts in the submucosa, resident subepithelial myofibroblasts, smooth muscle cells, or interstitial cells of Cajal (19, 20, 22, 23, 26).
A considerable amount of evidence suggests that transforming growth factor (TGF)-β1 and insulin-like growth factor (IGF)-I play a role in intestinal fibrosis (13, 20). Their expression is increased in regions of collagen deposition in CD (1, 22). In PG-PS-induced enterocolitis (31, 36, 37), both TGF-β1 (31) and IGF-I (36, 37) have been localized to myofibroblasts surrounding submucosal granulomas. IGF-I and TGF-β1 both induce collagen synthesis in multiple mesenchymal cell types (20), including rat intestinal myofibroblasts (6,31, 34). The role of TGF-β1 in activation of fibroblasts to α-smooth muscle actin-expressing myofibroblast phenotype in vivo and in vitro is well established (3, 6, 19, 20). TGF-β1, however, has known antiproliferative effects (19) while fibroblasts and myofibroblasts increase in number in intestinal fibrosis (20, 22, 31). A different mitogen, therefore, may act independently of TGF-β1 to stimulate growth of fibrogenic mesenchymal cells. Increased expression of IGF-I in intestinal fibrosis and its documented proliferative effects on intestinal mesenchymal cells (10, 27, 34) indicate that IGF-I could serve as such mitogen.
A family of IGF-binding proteins (IGFBPs) is known to modulate the actions of IGF-I. Depending on whether they are secreted or associated with cell surface or extracellular matrix (ECM), they can inhibit or potentiate IGF-I action (2, 15). IGFBPs, in particular IGFBP-3, can also have IGF-independent effects (11). Three IGFBPs, IGFBP-3, IGFBP-4, and IGFBP-5, are present in the normal intestine in vivo (29, 33-35). Increased expression of IGFBP-5 occurs in CD (35), suggesting that IGFBP-5 could modulate IGF-I action on fibrosis. IGF-I has been shown to stimulate expression of all three IGFBPs in intestinal smooth muscle cells in culture (10, 34), and TGF-β1 has been shown to induce an inhibitory IGFBP-3 in skin fibroblasts (16). The effects of IGF-I and TGF-β1 on IGFBPs synthesized in intestinal myofibroblasts are not defined.
The present study examined whether IGF-I and TGF-β1 show similar or overlapping sites of expression in the rat PG-PS model of intestinal inflammation and fibrosis. Partially overlapping sites of expression suggested that the two peptides may interact to regulate fibrosis in vivo. A well-characterized CCD-18Co cell line that can exhibit fibroblast (27) or myofibroblast (30) phenotype was therefore used to directly assess interactions between TGF-β1, IGF-I, and IGFBPs in regulating proliferation, collagen synthesis, and phenotype.
MATERIALS AND METHODS
TGF-β1 was purchased from Sigma Chemical (St. Louis, MO), human recombinant IGFBP-4 and -5 from Austral Biologicals (San Ramon, CA), and IGFBP-3 from Upstate Biotechnology (Lake Placid, NY). Rabbit polyclonal antibodies to IGFBP-3, -4, and -5 were purchased from Upstate Biotechnology, mouse monoclonal antibody specific for α-smooth muscle actin (Clone 1A4) from Dako (Carpinteria, CA), affinity-purified rabbit polyclonal antibody specific for procollagen α1(I) from Rockland Immunochemicals (Gilbertsville, PA), monoclonal antibody to β-tubulin (Clone Tub 2.1) from Sigma Chemicals, and goat polyclonal anti-human LAP TGF-β1 antibody from R&D Systems (Minneapolis, MN). Biotinylated horse anti-mouse antibody and Texas Red avidin D were purchased from Vector Laboratories (Burlingame, CA), peroxidase-conjugated donkey anti-rabbit, rabbit anti-goat, and rabbit anti-mouse immunoglobulins from Jackson ImmunoResearch Laboratories (West Grove, PA), enhanced chemiluminescence (ECL) reagent from New England Nuclear (Boston, MA), and [3H]thymidine and [125I]iodine from Amersham (Arlington Heights, IL). Polyclonal antibody to rat IGF-I precursor Ea-domain and human recombinant IGF-I (Genentech, San Francisco, CA) were kind gifts from Dr. L. E. Underwood (University of North Carolina, Chapel Hill, NC).
Localization of TGF-β1 and IGF-I in inflamed fibrotic intestine in vivo.
Immunoreactivity for TGF-β1 or IGF-I precursor was localized in adjacent sections of the cecum of rats with chronic PG-PS-induced enterocolitis (22, 37). Paraformaldehyde (4%)-fixed, paraffin-embedded tissue was sectioned at 6 μm. After quenching of endogenous peroxidase with methanol and hydrogen peroxide, sections were blocked with 4% normal rabbit serum (for IGF-I) or normal goat serum (for TGF-β1 antibody) for 30 min at room temperature (RT). Overnight incubation with anti-IGF-I antibody, diluted 1:300, and with anti-TGF-β1 antibody, diluted 1:100, in Triton X-100-containing phosphate-buffered saline (PBS, pH 7.4) was carried out overnight at 4°C. Sections were washed, then incubated with biotinylated secondary antibodies for 90 min at RT, followed by 90 min with avidin-biotin complex (Vectastain Kit, Vector Laboratories, Burlingame, CA). Peroxidase was visualized by diaminobenzidine treatment, and tissues were counterstained with Mayer's hematoxylin. Normal serum (4%) substituted for the primary antisera gave uniformly negative results. Adjacent sections were stained with hematoxylin and eosin for morphology and Sirius red for collagen.
CCD-18Co cells (CRL 1459) at passage 6 were obtained from American Type Culture Collection (Rockville, MD) and used betweenpassages 8 and 15. Cells were routinely split 1:3 when they reached 70–80% confluence and then were maintained in Dulbecco's modified Eagle's medium, supplemented with 10% fetal bovine serum (FBS), 50 U/ml penicillin, and 50 μg/ml streptomycin. For experiments, cells were plated at a density of 1 × 104 cells per well in 24-well tissue culture plates in medium containing FBS. Cells were grown overnight, then serum-deprived for 24 h, followed by treatment with serum-free medium with addition of indicated peptides. Concentrations of peptides required to induce maximal effects, IGF-I (10 ng/ml), TGF-β1 (1 ng/ml), and IGFBPs (5 nM), were established in preliminary experiments. In experiments where treatments were continued for longer than 24 h, medium was replaced every 24 h.
Assays of cell proliferation.
Incorporation of [3H]thymidine into DNA was used as a measure of mitogenic effects of peptide treatments. [3H]thymidine (2 μCi/ml) was added for the last 24 h of incubation in each experiment. To evaluate [3H]thymidine incorporation into DNA, medium was aspirated, and cells were washed twice with PBS and fixed with 10% trichloroacetic acid. Total cell extracts were harvested in 0.2 N NaOH and 0.1% SDS. Radioactivity incorporated into DNA was quantified by scintillation counting. All assays were performed in triplicate or quadruplicate and replicated in at least two separate experiments.
Western immunoblot or radioligand blot.
For immunoblots of whole cell lysates, cells were solubilized in SDS sample buffer. Equivalent amounts of protein were size fractionated on 8.5% SDS-polyacrylamide gels and transferred onto Immobilon-P polyvinylidene difluoride (PVDF) membranes (Millipore, Bedford, MA) using a Semi-Phor system (Hoefer Scientific Instruments, San Francisco, CA). After blocking in PBS containing 3% non-fat dry milk, blots were incubated with primary antisera against α-smooth muscle actin or procollagen α1(I) for 16 h at 4°C, washed in PBS containing 0.05% Tween, and then incubated with peroxidase conjugated secondary antibodies for 30 min at RT. Immunoreactive proteins were identified using the ECL detection system. Intensity of bands was quantified by densitometry of X-ray films using NIH Image software (version 1.61).
Radioligand blots were used to analyze IGFBPs in cell-conditioned media and ECM using methods previously described (9, 21). Medium was collected and analyzed directly. Cells were lysed in 0.5% Triton X-100 in PBS. Adherent nuclei and cytoskeleton were removed with 25 mM ammonium acetate in PBS (pH 9.0), and the remnant ECM was scraped from the plates and solubilized in 0.1% SDS and 0.1% sodium deoxycholate in PBS, pH 7.4 (9). Equal amounts of each sample were denatured in Laemmli buffer at 95°C and electrophoresed on 12.5% discontinuous SDS-polyacrylamide gels under nonreducing conditions. Size-separated proteins were transferred to PVDF membranes and incubated with 3 × 106 cpm/ml of recombinant human125I-IGF-I (specific activity 200–300 μCi/μg) for 16 h and washed, and then bound IGF-I was visualized by autoradiography (27). Immunoblot analyses were used to verify the identity of IGFBPs observed an radioligand blots.
Immunohistochemistry for α-smooth muscle actin.
Cells were grown in four-well Lab-Tech II plastic chambers (Nunc, Naperville, IL) and treated for different lengths of time (2–7 days) with various combinations of TGF-β and/or IGF-I. After treatment, cells were fixed in methanol (−20°C) for 10 min, blocked with 5% normal horse serum in PBS for 1 h at RT, and then incubated with the mouse monoclonal α-smooth muscle actin antibody (diluted 1:100 in PBS) for 2 h at RT. After washing in PBS, slides were incubated with biotinylated horse anti-mouse immunoglobulin, followed by Texas red-labeled avidin D (1:50 in PBS) for 30 min at RT. Fluorescent cells were visualized using a Zeiss Axiphot microscope with FITC (excitation = 488 nm, emission = 530/30BP) and Texas red (excitation = 568 nm, emission = 590LP) filters.
RNA extraction and analyses.
Total RNA was extracted from cells grown in 100-mm tissue culture dishes (Corning, Corning, NY) using the TRIzol Reagent (GIBCO BRL, Grand Island, NY), according to manufacturer's instructions. RT-PCR for mRNAs of human IGF-I, IGF-II, and constitutively expressed glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was performed on reverse-transcribed RNA using procedures described by Simmons et al. (27). Oligomers used for PCR were human IGF-I sense (5′-CCTCGCCTCTCTTCTACCTGGC-3′) and antisense (5′-CATGTCACTCTTCACTCCTCAGG-3′), human IGF-II sense (5′-CCTGGAGACGTACTGTGCTACC-3′) and antisense (5′-GCTCACTTCCGATTGCTGG-3′), and human GAPDH sense (5′-CTACTGGCGCTGCCAAGGCTGT-3′) and antisense (5′-GCCATGAGGTCCACCACCCTGTTG-3′). PCR conditions were 40 cycles with denaturation at 95°C, annealing at 55°C, and amplification at 72°C, each for 30 s.
Values for [3H]thymidine incorporation and abundance of procollagen α1(I) or α-smooth muscle actin were expressed as a ratio of levels in the serum-free control within the same experiment, and means ± SE were calculated across replicate experiments. Comparisons between treatments were analyzed by one-way ANOVA and post hoc pairwise comparisons using Tukey's test. AP value of <0.05 was considered statistically significant.
TGF-β1 and IGF-I show similar but not overlapping sites of expression during PG-PS-induced enterocolitis and fibrosis.
Figure 1 shows immunohistochemical localization of TGF-β1 and IGF-I precursor in areas of submucosal fibrosis in rats with chronic PG-PS-induced enterocolitis. Both peptides localize to cells surrounding granulomas, previously identified as myofibroblasts (31, 36, 37), but their sites of expression do not fully overlap. IGF-I precursor-positive cells are more numerous. In the granuloma, IGF-I precursor immunoreactivity localizes to mesenchymal cells that lie primarily at the periphery of the granuloma while TGF-β1 localizes more centrally, in closer proximity to immune cells. These findings suggest that some mesenchymal cells are exposed to TGF-β1 alone, others to TGF-β1 and IGF-I, and others primarily to IGF-I. This prompted us to examine the effects of TGF-β1 and IGF-I alone or in combination.
Effects of TGF-β1 and IGF-I on DNA synthesis.
IGF-I potently stimulated DNA synthesis in CCD-18Co cells, whereas TGF-β1 inhibited basal and IGF-I-stimulated DNA synthesis (Fig.2).
Effects of TGF-β1 and IGF-I on type I collagen and α-smooth muscle actin expression.
Figure 3 shows the time course of TGF-β1 action on type I collagen and α-smooth muscle actin. A prominent band of 175,000 kDa, corresponding in size to procollagen α1(I) (35), was detected in serum-deprived cells, which increased in abundance within 8 h of treatment with TGF-β1 followed by further increases to maximum abundance between 24 and 48 h of continuous exposure to TGF-β1. Low-level expression of α-smooth muscle actin was observed in serum-deprived cells and increased progressively over the course of 48 h of treatment with TGF-β1 (Fig. 3).
Next, we assessed interactions between IGF-I and TGF-β1 in regulating expression of type I collagen and α-smooth muscle actin. In cells treated for 48 h, IGF-I, TGF-β1, or IGF-I plus TGF-β1 induced similar increases in type I collagen (Fig.4). In contrast, IGF-I stimulated a much smaller but significant increase in α-smooth muscle actin expression than the major, 10-fold increase induced by TGF-β1 or TGF-β1 in combination with IGF-I (Fig. 4).
Because of their different effects on levels of expression of α-smooth muscle actin, we studied the ability of TGF-β1 and IGF-I alone or in combination to stimulate organization of α-smooth muscle actin into stress fibers, a typical feature of activated phenotype (5, 19, 20). Patterns of α-smooth muscle actin immunostaining are illustrated in Fig. 5, and the percentages of cells that showed α-smooth muscle actin-positive stress fibers after prolonged (4 or 7 days) treatment are summarized in Table 1. Most of the CCD-18Co cells grown on plastic in the absence of serum showed weak, diffuse α-smooth muscle actin immunostaining with only occasional cells showing stress fibers (Fig. 5). After 4 days of treatment with TGF-β1, stress fibers, most oriented into the same plane, were evident in virtually all cells. We note that, even though induction of α-smooth muscle actin by TGF-β1 was detected by Western immunoblot within 8 h, organization into stress fibers first became evident only in a subpopulation of cells after 2 days (data not shown), peaking at 4 days of treatment (Fig. 5). This prompted us to use days 4 and 7 for quantitative analyses of numbers of cells with α-smooth muscle actin organized into stress fibers. As shown in Table 1, TGF-β1 induced α-smooth muscle actin stress fibers in virtually every cell. IGF-I caused a slight but significant increase in the number of cells with stress fibers. When added in combination with TGF-β1, IGF-I caused a decrease in the number of cells with α-smooth muscle actin-positive stress fibers relative to TGF-β1 alone.
Together with Western immunoblotting data, these findings indicate that IGF-I, in contrast to TGF-β1, does not induce a typical activated phenotype in CCD-18Co cells despite similar effects of IGF-I and TGF-β1 on expression of type I collagen. Moreover, IGF-I combined with TGF-β1 attenuates organization of α-smooth muscle actin into stress fibers without affecting the ability of TGF-β1 to increase levels of α-smooth muscle actin or type I collagen.
TGF-β1 induces IGF-I in CCD-18Co cells.
We have previously reported that serum-deprived CCD-18Co cells express endogenous IGF-I and IGF-II that stimulate proliferation in an autocrine fashion (27). We used RT-PCR to test whether expression of IGFs is altered by TGF-β1. Indeed, treatment with TGF-β1 stimulated expression of endogenous IGF-I but not IGF-II (Fig.6). This effect has been observed in multiple intestinal myofibroblast cell lines (unpublished data). TGF-β1 induction of IGF-I may account for the lack of additive effect of TGF-β1 and IGF-I on collagen accumulation and indicates that TGF-β1 and IGF-I could act sequentially in regulating activated myofibroblast phenotype or proliferation. We therefore examined the effects of exogenous IGF-I on cells that had been pretreated with TGF-β1 followed by removal of conditioned medium before addition of IGF-I.
IGF-I stimulates proliferation of cells activated with TGF-β1 without reversing activated phenotype.
CCD-18Co cells were pretreated with TGF-β1 for 24 h or 7 days, times when we had established that few (<5% at 1 day) or many (96% at 7 days) of the TGF-β1-treated CCD-18Co cells show activated phenotype. TGF-β1 was then removed and IGF-I added. Cells, pretreated with TGF-β1, showed robust increases in DNA synthesis when subsequently exposed to IGF-I, even when TGF-β1 pretreatment was as long as 7 days (Fig. 7 A). Next, we asked if cells responding to IGF-I maintain the activated phenotype induced by TGF-β1. Cells pretreated with TGF-β1 for 7 days and then exposed to serum-free medium or IGF-I for 3 days were immunostained for α-smooth muscle actin (Fig. 7 B). A large percentage of the 96% of cells that exhibit stress fibers after 7 days of TGF-β1 treatment still maintained stress fibers after exposure for 3 days to serum-free medium (79 ± 3%) or IGF-I (82 ± 9.8%). Activated phenotype is therefore not reversed spontaneously or by IGF-I over this time frame.
We then asked whether proliferating myofibroblasts maintain their ability to produce collagen. Compared with cells continuously exposed to TGF-β1, production of collagen in TGF-β1-activated cells exposed to serum-free medium or IGF-I tended to drop over a 3-day period (Fig.8) but was still significantly higher than cells kept in serum-free medium throughout the entire study period. As cells pretreated with TGF-β1 followed by IGF-I proliferate (Fig. 7 A), this suggests that IGF-I can expand a population of cells converted to fibrogenic phenotype by prior exposure to TGF-β1 without reversing their fibrogenic phenotype.
IGF-I and TGF-β1 regulate IGFBPs expressed in CCD-18Co cells.
Radioligand and immunoblot data show that CCD-18Co cells express IGFBP-3, IGFBP-4, and IGFBP-5 (Fig. 9), the same IGFBPs that have been demonstrated in intestinal mesenchymal cells in vivo (13). IGF-I increased secretion of IGFBP-5 into cell-conditioned medium (Fig. 9, A and B), without affecting the association of any of the IGFBP with cells or ECM (Fig. 9 C). TGF-β1 induced IGFBP-3 secretion into medium as well as its association with ECM. TGF-β1 and IGF-I, in combination, resulted in increases in both IGFBP-3 and IGFBP-5 in medium and similar association of IGFBP-3 with ECM as observed with TGF-β1.
IGFBP-3 and -4 inhibit IGF-I-dependent proliferation of CCD-18Co cells without affecting phenotype.
Figure 10 summarizes the effects of IGFBPs on basal and IGF-I-stimulated DNA synthesis. Addition of IGFBP-3 to serum-deprived CCD-18Co modestly inhibited and IGFBP-5 modestly enhanced basal DNA synthesis. When coincubated with IGF-I, IGFBP-3 and IGFBP-4 significantly inhibited IGF-I-stimulated DNA synthesis while IGFBP-5 had no effect. IGFBPs did not affect the inhibitory effect of TGF-β1 on basal DNA synthesis and did not affect basal or IGF-I- or TGF-β1-stimulated expression of α-smooth muscle actin and type I collagen (data not shown).
TGF-β1 is a well-established mediator of wound healing and fibrosis in a number of organs, including skin, lungs, and the liver (3, 5, 12, 19). In the intestine, however, increased expression of TGF-β1 accompanies ulcerative colitis (UC), which generally is not associated with fibrosis, and CD, where fibrosis is a common complication (1). Our recent findings of IGF-I upregulation in the intestine of patients with CD, and not in UC (22), suggest a specific role of IGF-I in the fibrosis associated with CD. We have also shown that an expanded population of myofibroblasts and fibroblasts heavily populates areas of fibrosis in CD (22). We therefore used a human CCD-18Co cell line that can exhibit fibroblast (27) or myofibroblast (30) phenotype to directly test the potential interactions between TGF-β1 and IGF-I in regulating proliferation, phenotype, and collagen synthesis.
Our findings indicate that, like in other systems, TGF-β1 potently activates CCD-18Co cells to a myofibroblast phenotype and induces expression of type I collagen, a major component of ECM in fibrotic intestine in CD (17, 22, 28) and in animal models of chronic intestinal inflammation (31). TGF-β1, however, inhibits proliferation of CCD-18Co cells. These antiproliferative effects of TGF-β1 in CCD-18Co cells are shared by other mesenchymal cells (8, 19). Thus, while TGF-β1 may have a prominent role in the induction of ECM synthesis during intestinal inflammation, it is not likely to be a mediator of expansion of fibrogenic cells. In the liver, where TGF-β1 does stimulate growth of activated hepatic stellate cells (5), it also induces other peptide growth factors that further increase proliferation (5). Our novel observations that TGF-β1 induces expression of IGF-I in CCD-18Co cells and that IGF-I stimulates growth of CCD-18Co cells activated to a myofibroblast phenotype by TGF-β1 suggest that this model may be operative in the intestine. This is further supported by our prior (31, 36, 37) and current observations that TGF-β1 and IGF-I are both expressed in myofibroblasts surrounding submucosal granulomas in the PG-PS model of chronic enterocolitis. Close, but not overlapping, localization is consistent with a model whereby TGF-β1 could induce IGF-I in neighboring mesenchymal cells, which, in turn, could stimulate proliferation of mesenchymal cells in an autocrine or paracrine fashion.
Our study supports a role of IGF-I, along with TGF-β1, as a mediator of increased ECM synthesis during intestinal inflammation. Our study, however, provides no evidence for additive or synergistic effects of TGF-β1 and IGF-I on collagen synthesis in intestinal myofibroblasts as has been observed for type II collagen synthesis in cultured chondrocytes (32). This may reflect the fact that TGF-β1-induced IGF-I may, in part, mediate the actions of TGF-β1 on type I collagen synthesis.
The expression of IGFBP-3, IGFBP-4, and IGFBP-5 by CCD-18Co cells is consistent with observations that these are the primary IGFBPs expressed in the intestine in vivo (18, 29, 33, 35). As IGFBP-3 inhibits DNA synthesis and is induced by TGF-β1, it may contribute to TGF-β1-mediated inhibition of basal or IGF-I-dependent proliferation. It is noteworthy that the effect of TGF-β1 to induce both IGF-I and IGFBP-3 indicates that, if such an effect occurs in vivo, the spatial and temporal distribution of IGF-I and IGFBP-3 may have major effects on the degree to which fibrogenic populations of myofibroblasts are expanded. Other than modest effects to enhance basal DNA synthesis, our observations provide no evidence that IGFBP-5 potentiates IGF-I-mediated proliferation or collagen synthesis in intestinal myofibroblasts, which contrasts with its potentiating effects on IGF-I action in vascular smooth muscle (2).
The functional role of α-smooth muscle actin in myofibroblasts during wound healing and fibrosis is not well defined. Transient appearance of α-smooth muscle actin-positive myofibroblasts plays a role in wound contraction in the skin (3) and is accompanied by brief collagen gene induction (3, 4). In contrast, continuous activation of myofibroblasts results in fibrosis and excessive tissue scarring (3, 5). Our findings demonstrate that TGF-β1 stimulates increases in collagen synthesis, coincident with increases in α-smooth muscle actin expression, but at times that precede the organization of α-smooth muscle actin into stress fibers that typify activated phenotype. Furthermore, IGF-I strongly induces collagen but has very weak effects on α-smooth muscle actin expression or stress fiber formation. Thus, in intestinal myofibroblasts/fibroblasts, the appearance of α-smooth muscle actin stress fibers is not a prerequisite for increased collagen synthesis. Consistent with these findings, a recent report demonstrated that vanadate, an inhibitor of tyrosine phosphatases, reduced the appearance of α-smooth muscle actin stress fibers in skin myofibroblasts without impeding wound healing or ECM deposition (4). Inhibition of organization of α-smooth muscle actin into stress fibers has also been shown to enhance cell motility (24). This may suggest that factors that limit organization of α-smooth muscle actin into stress fibers, like IGF-I, may promote myofibroblast migration and could contribute to the disorganized mesenchymal overgrowth observed in fibrosis associated with chronic intestinal inflammation. This is speculative at present but worthy of further investigation.
In the intestine, the sequence of cellular events that underlie fibrosis is not well defined due, in part, to the complexity of mesenchymal cell subtypes. Subepithelial myofibroblasts and enteric smooth muscle cells (19) both are α-smooth muscle actin positive, making it difficult to trace activated myofibroblasts during intestinal inflammation (23). Nonetheless, our findings in a simple model, CCD-18Co cells, indicate that the relative abundance and spatial and temporal relationship of TGF-β1 and IGF-I could profoundly influence the phenotype and number of fibrogenic cells. These findings, in conjunction with localization of TGF-β1 and IGF-I to distinct mesenchymal cell populations in submucosal granulomas at regions of intestinal fibrosis in vivo, support a model put forward in Fig. 11. Initially, TGF-β1 expressed in immune cells recruits fibroblasts, converts them to activated phenotype, and induces IGF-I in these or neighboring fibroblasts/myofibroblasts. The combined actions of TGF-β1 and IGF-I would stimulate collagen synthesis and healing, but proliferation would be limited by TGF-β1 and IGFBP-3 induced by TGF-β1. The persistent expression of IGF-I in cells exposed to TGF-β1, but at sites distant from TGF-β1 expression, would permit expansion of collagen-producing myofibroblasts and fibrosis.
We thank Dr. Balfour Sartor for provision of fixed tissue from PG-PS-treated rats and K. McNaughton for assistance with histology.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-40247 and DK-47769 and Minority Opportunities in Research Division of the National Institutes of General Medical Sciences Grant GM-000678. The study was facilitated by the molecular histopathology core of the Center for Gastrointestinal Biology and Disease (NIH P30-DK-34987) and the tissue culture and DNA synthesis cores of the Lineberger Cancer Center (NIH CA-16086).
Address for reprint requests and other correspondence: J. G. Simmons, Dept. of Cell and Molecular Physiology, The Univ. of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7545 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
April 24, 2002;10.1152/ajpgi.00057.2002
- Copyright © 2002 the American Physiological Society