Collagen Metabolism Is a Novel Target of the Neuropeptide (cid:1) -Melanocyte-stimulating Hormone*

Suppression of collagen synthesis is a major therapeutic goal in the treatment of fibrotic disorders. We show here that (cid:1) -melanocyte-stimulating hormone ( (cid:1) -MSH), a neuropeptide well known for its pigment-inducing capacity, modulates collagen synthesis and deposition. (cid:1) -MSH in vitro suppresses the synthesis of collagen types I, III, and V and down-regulates the secretion of procollagen type I C-terminal peptide (PICP) in human dermal fibroblasts treated with the fibrogenic cytokine transforming growth factor- (cid:2) 1 (TGF- (cid:2) 1 ). (cid:1) -MSH did not interfere with TGF- (cid:2) 1 signaling, because TGF- (cid:2) 1 -in- duced expression of collagen mRNA was not affected, implying a posttranscriptional mechanism. Human dermal fibroblasts in vitro express a high affinity binding site for MSH, which was identified by reverse transcription PCR and immunofluorescence

Fibrotic and sclerotic diseases comprise a large and heterogeneous group of inflammatory, idiopathic, toxic, hereditary, and pharmacologically induced disorders such as hypertrophic scars, keloids, localized scleroderma, systemic sclerosis, sclerodermic graft versus host disease of the skin, cirrhosis of the liver, idiopathic and bleomycin-induced lung fibrosis, or cyclosporine-induced nephropathy. The therapeutic options are limited, and treatment of these disabling disorders is still a challenge.
A key feature of fibrotic disorders is excessive production of extracellular matrix, mainly type I collagen, followed by a gradual loss of organ function which, in some cases, can be fatal. In recent years it became apparent that transforming growth factor-␤ 1 (TGF-␤ 1 ), 1 a multifunctional cytokine, is crucially involved in the pathogenesis of fibrotic disorders (1)(2)(3)(4)(5). It induces fibrosis by various ways (reviewed in Ref. 6). It enhances the expression of several collagens including types I, III, and V. TGF-␤ 1 decreases the production of matrix-degrading proteases and enhances the synthesis of inhibitors of such proteases. TGF-␤ 1 also increases extracellular cross-linking of collagen by enhancing the expression and the activity of lysyl oxidase (7). These multiple activities explain the potent fibrotic effect of TGF-␤ 1 . Therefore, strategies aimed at antagonizing the strong profibrotic effect of TGF-␤ 1 are regarded as providing a promising approach to preventing excessive collagen accumulation in fibrotic disorders (8 -10).
␣-Melanocyte-stimulating hormone (␣-MSH) is a tridecapeptide generated from pro-opiomelanocortin (POMC) by proteolytic cleavage (reviewed in Ref. 11). It was originally isolated from the pituitary gland and characterized as a pigment-inducing factor regulating the coat color of many vertebrate species, but it turned out to regulate many other biological activities with regard to the skin (reviewed in Refs. 11 and 12). The biological activities of ␣-MSH are mediated by a family of structurally related receptors that are known as the melanocortin receptors (MC-Rs). They belong to the superfamily of G protein-coupled receptors with seven trans-membrane domains, and they activate adenylate cyclase after ligand binding. Five MC-R subtypes have been cloned that differ in their relative affinities to ␣-MSH and the other melanocortins (13,14).
Here we show that, in addition to its multiple biological effects, ␣-MSH suppresses TGF-␤ 1 -induced collagen synthesis by human dermal fibroblasts (HDF) in vitro. This effect is mediated via the MC-1R. ␣-MSH also exerts its anti-fibrogenic activity in vivo, because injection of ␣-MSH into mice reduces TGF-␤ 1 -induced fibrosis. Our data establish a role for melanocortins in fibroblast biology and point toward a therapeutic potential of ␣-MSH and its analogues in the treatment of fibrotic and sclerotic diseases.

MATERIALS AND METHODS
Cells and Culture Conditions-HDF from neonatal foreskin and adult skin as well as normal human melanocytes were purchased from Cell Systems, St. Katharinen, Germany. The human fibrosarcoma cell line HT-1080 was obtained from the American Type Culture Collection (ATCC). Fibroblasts were routinely cultured in RPMI 1640 (PAA, Cölbe, Germany), 1% glutamine, 1% penicillin/streptomycin (both from PAA), and 10% fetal calf serum (FCS) (Biochrom, Berlin, Germany) in a humidified atmosphere of 5% CO 2 at 37°C. Normal human melanocytes were cultured in MBM2 medium plus MGM-3 aliquots as indicated by the manufacturer (Clonetics, Walkersville, MD).
RNA Extraction, RT-PCR, and Sequencing-Total RNA was isolated from cells using a commercial purification kit (Promega, Madison, WI). After DNA digestion, 1 g of total RNA was reverse transcribed with 15 units of avian myeloblastosis virus reverse transcriptase (Promega). The resulting cDNA was amplified with 2.5 units of Taq polymerase (Promega) and MC-R primers under conditions identical to those described previously (15)(16)(17). Primer sequences and the sizes of their amplification products are given in Table I. Only RNA samples that did not yield amplification products were reverse transcribed. For some positive controls, genomic DNA from HDF was prepared by routine protocols. Amplicons were separated in 1.5% agarose gels. The resulting MC-1R-related band in HDF was purified using a gel extraction kit (Qiagen, Santa Clarita, CA), cloned into pGEM-T easy vectors (Promega), and sequenced (4base lab GmbH, Reutlingen, Germany).
Quantitative Real Time PCR-Quantification of mRNA levels of the various procollagen chains was carried out by real time fluorescence detection as described previously (18). cDNA was prepared amplified by PCR in the ABI Prism 7700 sequence detector (PE Biosystems, Foster City, CA). Primer and probe sequences were designed by the Primer Express software (PE Biosystems) or supplied by PE Biosystems (glyceraldehyde-3-phosphate dehydrogenase) (18). The primers used were as follows: COL(I)␣ 1 sense, 5Ј-CAGCCGCTTCACCTACAGC-3Ј, COL(I)␣ 1 antisense, 5Ј-AATCACTGTCTTGCCCCAGG-3Ј, and COL(I)␣ 1 probe, 5Ј-ACTGTCGATGGCTGCACGAGTCACAC-3Ј; COL(I)␣ 2 sense, 5Ј-GA-TTGAGACCCTTCTTACTCCTGAA-3Ј, COL(I)␣ 2 antisense, 5Ј-GGGT-GGCTGAGTCTCAAGTCA-3Ј, and COL(I)␣ 2 probe, 5Ј-TCTAGAAAGA-ACCCAGCTCGCACATGC-3Ј; and COL(III)␣ 1 sense, 5Ј-TCCAACTGC-TCCTACTCGCC-3Ј, COL(III)␣ 1 antisense, 5Ј-GAGGGCCTGGATCTC-CCTT-3Ј, and COL(III)␣ 1 probe 5Ј-CCTAATGGTCAAGGACCTCAAG-GCCC-3Ј. Probes were labeled at the 5Ј-end with the reporter dyes 6-carboxyfluorescein or VIC and at the 3Ј-end with the quencher dye 6-carboxy-tetramethyl-rhodamine. The 5Ј-nuclease activity of the Taq polymerase (Applied Biosystems) cleaved the probe and released the fluorescent dyes, which were detected by the laser detector of the sequence detector. After the detection threshold was reached, the fluorescence signal was proportional to the amount of PCR product generated. The initial template concentration could be calculated from the cycle number when the amount of PCR product passed a threshold set in the exponential phase of the PCR. Relative gene expression levels were calculated using standard curves generated by serial dilutions of cDNA from HT1080 cells. The relative amounts of gene expression were calculated by using the expression of glyceraldehyde-3-phosphate dehydrogenase as an internal standard. Expression of each gene was assessed by three independent PCR analyses and calculation of the mean Ϯ S.E. Data were analyzed by the Student's t test.
Binding Studies-[Nle 4 ,D-Phe 7 ]␣-MSH (NDP-MSH) (Bachem, Bubendorf, Switzerland) was radioiodinated by the chloramine T method and purified by high pressure liquid chromatography. Binding studies were performed as described previously (19). In short, cells were washed with binding buffer and distributed into 96-well plates. Cells were then incubated for 2 h at 37°C with 50 l of binding buffer in each well containing a constant concentration of 0.2 nM 125 I-NDP-MSH and appropriate concentrations of unlabeled ligand. After incubation, the cells were washed with 0.2 ml of ice-cold binding buffer and detached with 0.2 ml of 0.1 N NaOH, and the radioactivity was counted in each well. The binding assays were performed in duplicate wells. Radioactivity was determined by a gamma counter (Wallac, Wizard Automatic), and data were analyzed with a software package for radioligand binding analyses. Data were analyzed by fitting it to formulas derived from the law of mass action by the method generally referred to as computer modeling.
Immunofluorescence-HDF were seeded into chamber slides and fixed with methanol for 30 min at Ϫ20°C or, for surface staining, with 4% paraformaldehyde for 30 min at room temperature. Nonspecific binding was blocked with 5% goat/donkey serum for 1 h at room temperature. Cells were then incubated for 1 h with a rabbit polyclonal antibody against the human MC-1R (1 g/ml). Production and characterization of the anti-human MC-1R is described in detail elsewhere (20,21). In some experiments, double staining with a monoclonal antibody against protein disulfide isomerase (1:100; Dako, Hamburg, Germany), a cytoplasmic marker (22,23), was performed. Bound antibodies were visualized with a donkey anti-rabbit antibody coupled to Texas Red (1:100; Dianova, Hamburg, Germany) and a goat anti-mouse antibody coupled to fluorescein isothiocyanate (1:100; Dako). After mounting, specimens were examined with a confocal laser-scanning microscope (TCS E, Leica, Heidelberg, Germany).
Determination of cAMP-For intracellular cAMP measurements, 2 ϫ 10 4 HDF were seeded into 96-well tissue culture plates. The next day, the routine culture medium was changed to RPMI 1640 containing 1% FCS. Cells were cultured for additional 24 h followed by stimulation with ␣-MSH as indicated for 20 min in the presence of 0.1 mM isobutyl methylxanthine. 0.1-5 M forskolin was used as a positive control. After incubation, supernatants were removed, and cells were lysed. cAMP levels in the lysates were determined by a specific enzyme immunoassay according to the instructions of the manufacturer (Amersham Biosciences). Triplicate wells were used for each individual treatment, and statistical analysis was performed using the Student's t test.
Collagen Analysis-HDF were seeded into 6-well tissue culture plates (250,000 per well) and allowed to attach and grow for 16 h. Subconfluent cell monolayers were then incubated with minimal essential medium containing 0.5% FCS and 50 g/ml L-ascorbic acid for 24 h in the presence of TGF-␤ 1 (10 ng/ml), ␣-MSH (10 Ϫ6 M), or a combination of both substances. HDF were rinsed and depleted for 1 h in methionine/cysteine-free minimal essential medium (ICN, Costa Mesa, CA). Cells were labeled with [ 35 S]methionine/cysteine mix (50 Ci/ml) for 16 h in the presence of ascorbate and the above agents. Media were collected, cell cultures rinsed three times with ice-cold Tris-buffered saline, and cells were scraped into 1% Nonidet P-40 in Tris-buffered saline with a rubber policeman. Cell lysates were centrifuged, and aliquots of the supernatant were subjected to liquid scintillation counting to determine cell mass. Media and combined cell lysates/scrapings were treated with pepsin to destroy non-collagenous proteins (24). Proteins in either fraction were precipitated using methanol/chloroform and processed for SDS-PAGE (5% acrylamide; acryl/bisacryl, 37.5:1). All loaded aliquots were calibrated for cell mass so that all slots contained pepsin-treated collagen derived from the same amount of cells. Slab gels were fixed, and radiolabeled collagens were detected by autoradiography (25). Representative gels were subjected to densitometry using the Biostep Phoretix Grabber (Biostep, Jahnsdorf, Germany) Determination of Procollagen I C-terminal Peptide-The amounts of procollagen I C-terminal peptide used as a marker for procollagen I secretion were determined using a commercially available ELISA (TaKaRa, Shiga, Japan). HDF were seeded into 12-well tissue culture plates at a density of 250,000 cells per well. Confluent HDFs were then deprived of FCS for 2 days and subsequently stimulated with ␣-MSH (10 Ϫ6 -10 Ϫ10 M), TGF-␤ 1 (10 ng/ml), or both agents in the presence of 50 g/ml ascorbate. In some experiments, cells were coincubated with a synthetic peptide corresponding to the amino acids 87-132 of the human Agouti signaling peptide (Phoenix Pharmaceuticals, Belmont, CA) at a 10-fold molar excess. Culture supernatant were harvested after 48 h, centrifuged, and frozen at Ϫ70°C until use. Statistical evaluation from triplicate wells was performed using the Student's t test.
Mouse Model for Cutaneous Fibrosis-For in vivo evaluation of the anti-fibrogenic effect of ␣-MSH, a mouse model described previously by Shinozaki et al. (26) was used with slight modifications. Accordingly, cutaneous fibrosis was induced by intracutaneous injections of 800 ng of TGF-␤ 1 into the neck of newborn Balb/c mice on three consecutive days. Treatment groups (four groups of three mice each) consisted of mice injected with TGF-␤ 1 , ␣-MSH (25 g), TGF-␤ 1 plus ␣-MSH, and the solvent (0.1% BSA in PBS) in which TGF-␤ 1 had been solubilized. On day 4, mice were sacrificed, and 4-mm punch biopsies were taken from the sites of injection for immunohistochemical analysis.
Immunohistochemistry-After fixation in 4% paraformaldehyde and embedment in paraffin, biopsies from mouse skin were processed with the following stains: (i) hematoxylin and eosin; (ii) van Gieson stain, in which collagen appears red; and (iii) resorcin-fuchsin stain, according to Weigert, in which elastic tissue appears black. For collagen staining, the sections were treated with 1 mg/ml pepsin (Sigma) in 0.5 M acetic acid, washed, and incubated with a rabbit antibody against collagen type I (1:100; DPC Biermann, Bad Nauheim, Germany) for 1 h. For the staining of vimentin, sections were microwave-treated to unmask epitopes, followed by incubation with a polyclonal antibody from Abcam (Cambridge, UK) for 30 min at 37°C. For the staining of ␣-smooth muscle actin, a monoclonal antibody from Dunn Labortechnik (Asbach, Germany) was incubated for 1 h at 2 g/ml without prior unmasking. Immunohistochemistry for MC-1R in sections of normal adult human skin (n Ͼ 5) was performed exactly as outlined previously (17,21). Sections were developed by the indirect immunoperoxidase technique using 3-amino-9-ethylcarbazole (Sigma) as a chromogen. Negative controls included incubation with control IgG at the same protein concentration as the primary antibody, omission of the first antibody, or pre-incubation with the immunogenic peptide in 10-fold weight excess in the case of MC-1R immunostaining. Vimentin immunostaining and ␣-smooth muscle actin immunostaining in sections of mouse skin were quantitatively assessed by counting the number of immunoreactive interfollicular dermal cells in three high power fields (ϫ400). Means Ϯ S.D. from 3-4 independent experiments were analyzed by analysis of variance.

␣-MSH Modulates Collagen Expression by HDF in
Vitro-We addressed the question of whether ␣-MSH can modulate the key function of HDF, namely the expression and secretion of collagen. To this end, cultured, normal HDF from neonatal foreskin were treated with ␣-MSH, TGF-␤ 1 , or both substances. The amounts of collagen present in cell lysates and culture supernatants were separately determined after metabolic labeling, pepsin digestion, and SDS-PAGE.
TGF-␤ 1 , a well known inducer of collagen synthesis (6), increased the amount of secreted collagens in the culture medium (Fig. 1A). ␣-MSH alone appeared to reduce the extracellular amount of collagens I, III, and V by 50 -70% as determined by densitometry (Fig. 1A, and data not shown). This reduction was not due to intracellular retention, although there was some increase in ␣ 1 (I) chains that we attribute to a higher synthesis rate. The intracellular bands showed no delayed migration and, thus, excluded significant posttranslational overmodification due to abnormal intra-endoplasmatic retention (Fig. 1A). Most strikingly, ␣-MSH dramatically reversed the stimulatory effects of TGF-␤ 1 on the extracellular collagen presence with the strongest effects on collagens I and III and somewhat milder effects on collagen V (Fig. 1A). In the conspicuous absence of intracellular retention, these findings implicate either intracellular or extracellular proteolytic degradation or a combination of both.
To further substantiate the activity of ␣-MSH on collagen synthesis and/or secretion, we determined the amount of procollagen I C-terminal peptide (PICP) in the culture media of HDF stimulated with ␣-MSH, TGF-␤ 1 , or both agents. The addition of TGF-␤ 1 led to a dramatic increase of PICP in the culture medium by Ͼ500%. In accordance with the modulatory effect of ␣-MSH on the TGF-␤ 1 -induced collagen biosynthesis and subsequent secretion, we found significantly reduced secreted amounts of PICP by HDF (677.9 Ϯ 60.1 pg/ml versus 1313.9 Ϯ 136 pg/ml; p Ͻ 0.005) (Fig. 1B). ␣-MSH alone, in contrast, did not affect the basal amounts of secreted PICP. These findings suggested either an intracellular or an extracellular cause for the reduction of secreted procollagen I. ulatory activity of ␣-MSH on collagen synthesis is regulated at the transcriptional level. HDF from neonatal foreskin were stimulated with ␣-MSH, TGF-␤ 1 , or both agents for 12 h. The relative mRNA levels for the ␣ 1 (I) and ␣ 2 (I) chains of collagen I (alleles COL1A1 and COL1A2, respectively) and for the ␣ 1 (III) chains for collagen III (allele COL3A1) were subsequently determined by quantitative real-time PCR. TGF-␤ 1 significantly increased the mRNA levels of collagen type I ␣ 1 and ␣ 2 as well as that of collagen type III ␣ 1 as compared with non-treated cells (Table II). The observed rate of increase in the amount of these collagens by TGF-␤ 1 was in accordance with earlier reports (27). Despite some variation, neither ␣-MSH alone nor coincubation of ␣-MSH and TGF-␤ 1 caused a significant reduction in the relative levels of the collagen mRNAs (Table II). Similar results were obtained when HDFs were treated with TGF-␤ 1 and ␣-MSH for 24 h (data not shown). These findings show that ␣-MSH does not interfere with TGF-␤ 1 signaling and that ␣-MSH may affect collagen expression at the posttranscriptional level.
Detection of High Affinity Binding Sites for MSH on HDF-The identified effects of ␣-MSH on the amount of extracellular collagen suggested the presence of specific binding sites in HDF. Therefore, we examined HDF from neonatal foreskin for competitive radioligand binding using an iodinated synthetic ␣-MSH analogue, NDP-MSH. Displacement was performed with an unlabeled ligand at varying concentrations, and COS-1 cells were used as a negative control. HDF exhibited a specific and saturable binding kinetic with 125 I-NDP-MSH. The affinity of the radioligand was similar to COS-1 cells transfected with the human MC-1R (Fig. 2). The K i values were 0.058 Ϯ 0.012 nM for the HDF and 0.086 Ϯ 0.033 nM for COS-1 cells transfected with the human MC-1R, the latter value being similar to previous studies (19). HDF, therefore, exhibited similar affinity but slightly lower expression levels of high affinity MSH binding sites than did COS-1 cells transfected with MC-1R. These data strongly suggested that ␣-MSH binds to specific surface receptors on the surface of HDF, which appear to mediate its biological action.
Expression of MC-1R in HDF in Vitro and in Situ-To investigate in detail the expression of MC-Rs in HDF, we performed RT-PCR analysis using primers against all known MC-Rs (Table I). MC-1R was the only MC-R expressed in HDF derived from neonatal foreskin (Fig. 3A). Similarly, HDF derived from adult human skin expressed MC-1R at the RNA level (data not shown). The MC-1R amplification product of HDF comigrated exactly with that of normal human melanocytes used as a positive control (Fig. 3A). The identity of the amplification product in HDF (416 bp) was determined by DNA sequencing and found to be identical with the mRNA sequence of MC-1R as deposited in the National Center for Biotechnology Information (Table I, and data not shown). In contrast to MC-1R, no other MC-R was expressed in HDF as shown by RT-PCR (Fig. 3A).
The amplification products of the positive controls were all of the expected size ( Fig. 3A and Table I) (15)(16)(17).
To examine the expression of the MC-1R at the protein level in HDF in vitro, we next performed immunofluorescence studies. For melanoma cells in culture, it has been reported that binding sites for MSH are confined to certain areas on the cell surface (17,28), whereas in other cutaneous cell types MC-1R immunoreactivity appeared more randomly distributed (17). Immunofluorescence studies with HDF fixed either in paraformaldehyde (data no shown) or methanol gave similar images. MC-1R antigenicity was visible as a characteristic punctate staining randomly distributed with accentuation in the cell periphery (Fig. 3B). Control experiments with preimmune serum or neutralization with the immunogenic peptide used for generation of the anti-MC-1R antibody did not produce any staining (Fig. 3B).
To check if expression of MC-1R is maintained in HDF in situ, skin sections of normal adult human skin were processed for immunohistochemistry. MC-1R immunoreactivity was absent in interfollicular dermal fibroblasts at the light microscopic level. However, in distinct fibroblastic cells of the connective tissue sheath of hair follicles, MC-1R immunoreactivity was consistently detectable (Fig. 3C). MC-1R immunostaining in these cells had a punctate pattern and was localized mainly in the cytoplasm (Fig. 3C). Immunostaining with an antibody against vimentin confirmed the nature of these cells as fibroblasts (data not shown). Pre-incubation with the antigenic peptide or pre-immune serum, in contrast, did not produce any staining (Fig. 3C). As reported previously (21), MC-1R immunoreactivity was also detected in distinct epithelia of the skin appendages, for example of the outer root sheath hair follicle keratinocytes (Fig. 3C). Collectively, the data from these studies demonstrate that MC-1R expression is not restricted to HDF in culture but is also detectable in situ in distinct dermal fibroblast populations of normal human skin.
Functional Coupling of MC-1R Expressed in HDF-To investigate the functioning of the identified MC-1R on HDF, we  performed cAMP measurements of cells stimulated with varying doses of ␣-MSH. It has been shown previously that all members of the MC-R family are G protein-coupled receptors whose interaction with the ligand results in stimulation of adenylate cyclase. ␣-MSH increased the amount of intracellular cAMP in a dose-dependent manner as compared with nontreated cells (Fig. 4). This effect was maximal at 10 Ϫ6 M (p Ͻ 0.05) and similar to stimulation of HDF with 0.1 M forskolin. Concentrations of ␣-MSH higher than 10 Ϫ9 M did not lead to significant changes in the amount of intracellular cAMP as compared with non-stimulated HDF (Fig. 4).
To clarify if MC-1R mediates the inhibitory action of ␣-MSH on TGF-␤ 1 -induced collagen synthesis in HDF, we performed blocking experiments with a synthetic peptide corresponding to the amino acids 87-132 of human Agouti signaling peptide (ASIP), a natural and highly potent antagonist of MC-1R but also of MC-4R (29). Cysteine-rich C-terminal ASIP fragments were shown previously to be as potent as full-length ASIP (30). Because HDF express only MC-1R (Fig. 3A) we hypothesized that the synthetic ASIP peptide would block the antagonistic effect of ␣-MSH on collagen synthesis induced by TGF-␤ 1 . Coincubation of the synthetic ASIP fragment at 10 Ϫ7 M plus ␣-MSH at 10 Ϫ8 M and TGF-␤ 1 (10 ng/ml) completely abrogated the antagonistic effect of ␣-MSH on PICP secretion (Fig. 5). In contrast, the synthetic ASIP fragment alone did not affect secretion of PICP in a significant manner (Fig. 5). As outlined above, ␣-MSH alone or coincubation of TGF-␤ 1 plus the synthetic ASIP did not exert any modulatory effect on PICP secretion (data not shown). Taken together, these data strongly support the concept that ␣-MSH, via acting on MC-1R, modulates fibroblast activity or collagen synthesis, respectively.
␣-MSH has Anti-fibrogenic Activity in Vivo-We next wished to know if ␣-MSH can also modulate collagen synthesis and secretion in vivo. Therefore, we employed an animal model in which cutaneous fibrosis is elicited by repetitive intracutaneous injections of high doses of TGF-␤ 1 (26). We chose newborn mice because they contain significantly less collagen in their skin than adult mice, thus rendering the former suitable for evaluation of fibrogenic and anti-fibrogenic stimuli. Accordingly, newborn mice were injected into the neck for three consecutive days with TGF-␤ 1 (800 ng), ␣-MSH (25 g), ␣-MSH plus TGF-␤ 1 , or PBS. On day 4, punch biopsies were taken from the injection sites and subjected to biochemical and histological analysis. In contrast with samples from adult murine skin, the   FIG. 3. Expression of the MC-1R in HDF in vitro and in situ. A, RT-PCR analysis of total RNA derived from cultured HDF using specific primers against the five MC-Rs. Negative controls (NC) consisted of a genomic contamination control using total RNA without reverse transcription and a reaction mixture control using H 2 O instead of the template. Normal human melanocytes (NHM) and genomic DNA from HDF were used as positive controls (PC). B, MC-1R immunoreactivity in cultured HDF as shown by double immunofluorescence. Cells were stained with a polyclonal MC-1R antibody (red signal) and a monoclonal against protein disulfide isomerase (green signal), a cytoplasmic marker. Scale bar, 20 m. C, MC-1R immunohistochemistry of normal adult human skin using a polyclonal anti-MC-1R antibody and the immunoperoxidase technique. MC-1R immunoreactive fibroblasts were detected within the connective tissue sheath of the hair follicle (arrows) but not in the negative control (pre-incubation with the immunogenic peptide). Asterisk (*) indicates MC-1R immunoreactivity that was also detected in outer root sheath epithelia. Magnification in the top two sections is ϫ100. MC-1R immunoreactivity in perifollicular fibroblasts displayed a punctate staining pattern. Magnification is ϫ200 in the bottom section. hydroxy-proline content in the newborn mice samples was below the detection limit to allow collagen analysis by this approach (data not shown). Therefore, we used semiquantitative histochemical and immunohistochemical analysis to assess the effect of ␣-MSH on TGF-␤ 1 -induced skin fibrosis. When compared with PBS, injections with TGF-␤ 1 induced dermal thickening and fibrosis as well as increased numbers of collagen fibers as shown by hematoxylin and eosin and Van Gieson stains and also by immunohistochemistry using an anti-collagen type I antibody (Fig. 6, A, D, and G versus B, E, and H). Elastic fibers in the skin of untreated mice, in contrast, were sparse and were detected primarily in the dermal vasculature with no increase upon injection with TGF-␤ 1 (data not shown). Injections with ␣-MSH alone did not produce any changes as compared with mice treated with PBS/BSA (data not shown). On the other hand, coinjection of mice with TGF-␤ 1 plus ␣-MSH resulted in a significant reduction in the amount of extracellularly deposited collagen as compared with mice injected with TGF-␤ 1 alone (Fig. 6, C, F, and I versus B, E, and H). To further corroborate the anti-fibrogenic activity of ␣-MSH in vivo, we examined the in situ number of dermal cells immunoreactive for vimentin, an established fibroblast marker, as well as for ␣-smooth muscle actin (␣-SMA), a fibroblast activation and myofibroblast trans-differentiation marker. It has been shown previously that ␣-SMA is strongly induced in fibroblasts by TGF-␤ 1 in vitro (31). Enhanced in situ expression of ␣-SMA has also been associated with increased tissue injury and the progression of interstitial fibrosis (32)(33)(34). As compared with mice injected with PBS/BSA, TGF-␤ 1 significantly increased the number of dermal cells immunoreactive for both vimentin and SMA. This effect was strongly antagonized by coinjection of TGF-␤ 1 plus ␣-MSH (Table III; p Ͻ 0.001 for vimentin; p Ͻ 0.02 for ␣-SMA). In accordance with the above data, ␣-MSH alone did not have any modulatory activity on the number of vimentin or ␣-SMA positive cells in murine skin (data not shown). These findings demonstrate that the modulatory activity of ␣-MSH on TGF-␤ 1 -induced collagen synthesis and deposition is not confined to HDF in vitro but is also operational in vivo. DISCUSSION We have shown here that the neuropeptide ␣-MSH antagonizes the action of TGF-␤ 1 on collagen synthesis in HDF in vitro as well as in a mouse model of cutaneous fibrosis in vivo.
This represents an analogy to the antagonistic action of this neuropeptide on interleukin-1-mediated responses that appears to be affected by suppression of NF-B activation (35). It has been shown previously that the TGF-␤ 1 -mediated effect on type I procollagen ␣ 1 transcription in rat kidney fibroblasts depends on the expression of connective tissue growth factor (CTGF), a downstream target of TGF-␤ 1 that is negatively regulated by cAMP (36). Because ␣-MSH increases intracellular cAMP, we originally hypothesized that treatment with ␣-MSH would lead to reduced connective tissue growth factor expression and, consequently, to reduced mRNA expression of collagen types I and III. However, the lack of any effect of ␣-MSH on the mRNA levels of types I and III collagen precludes an interference of ␣-MSH with the signal transduction of TGF-␤ 1 . The latter conclusion is in accordance with findings showing that ␣-MSH in HDF neither inhibits TGF-␤ 1 -induced phosphorylation or blocks the nuclear translocation of Smad2/3 (data not shown). It is possible that the observed differences between the aforementioned findings and our data are due to cell-specific differences (i.e., HDF versus rat kidney fibroblasts) or differences in the individual experimental setting (e.g. the use of artificial cAMP versus natural cAMP inducers).
The fact that ␣-MSH does not interfere with TGF-␤ 1 -mediated transcription of collagen type I suggests a posttranscriptional mechanism for the suppressive effect on collagen synthesis, e.g. translational repression or enhanced extracellular proteolytic degradation. Regarding the latter, ␣-MSH may suppress the activity of the C-and N-terminal procollagen proteinases that remove the propeptides from secreted procollagen. This would lead to a reduced formation of collagen fibers. Alternatively, ␣-MSH might stimulate degradation of secreted procollagen by activating members of the matrix metalloproteinase (MMP) family, for example MMP-1 (37) and/or other members such as MMP-2 and MMP-9. Because C-and Nproteinases are only marginally active in fibroblast cell cultures (38), the observed reduction of collagens in the culture medium may point to extracellular degradation of single procollagen trimers. In addition to the above potential mechanisms, it is possible that ␣-MSH affects the intracellular free pool of selected amino acids required for collagen synthesis. Recently, it was demonstrated that selected environmental changes such as hypoxia inhibit proline uptake while leaving methionine uptake relatively unaffected (39). With regard to ␣-MSH, however, nothing is known about the potential influence on the uptake and transport of amino acids. Further studies are thus necessary to elucidate the molecular mechanism by which ␣-MSH modulates collagen synthesis.
The presence of MC-1R in HDF as shown by radioligand binding, RT-PCR, and immunofluorescence in this paper explains a number of previously reported activities of ␣-MSH on human fibroblasts. It was reported that ␣-MSH can block the interleukin-1-induced production of prostaglandin E in a lung fibroblast cell line (40). We have shown that ␣-MSH in vitro increases the secretion of interleukin-8 and modulates the activation of the transcription factors NF-B and AP-1 in HDF (41).
The inhibitory action of ␣-MSH on TGF-␤ 1 -induced collagen synthesis and/or secretion by HDF in vitro and its anti-fibrogenic activity in vivo adds another dimension to the broad spectrum of biological activities of this neuropeptide. The skin itself contains the full capacity to produce POMC peptides (42), and HDF in vitro have recently been shown to generate immunoreactive amounts of adrenocorticotropin and ␣-MSH (43). However, transgenic mice with a signaling-deficient MC-1R are not sclerodermic (44). In light of our findings, it is puzzling that peripheral blood levels of POMC peptides are elevated in pa- tients with systemic sclerosis (45), and increased in situ expression of POMC has been detected in inflammatory cells of keloids (46). It remains to be determined whether these findings are part of the cutaneous stress response (11) or are related to the involvement of ␣-MSH in collagen metabolism. Because the amounts of ␣-MSH used in our studies were higher than the plasma concentration of this neuropeptide in man, the exact role of ␣-MSH in fibroblasts under physiologic and pathophysiologic conditions requires further investigation.
Our findings on the modulating activity of ␣-MSH on collagen synthesis finally highlight a novel biological activity that may be exploited in the treatment of fibrotic disorders. ␣-MSH is a small molecule with a molecular mass of 1.66 kDa. Preliminary data have been shown that nickel-induced contact dermatitis in humans can be suppressed by a topical ␣-MSH (100 M) (11). It is known that systemic or intradermal injection of ␣-MSH or its analogue, NDP-MSH, into humans is well tolerated and has little toxicity (47,48). The latter ␣-MSH derivative was found to be 10 -1000-fold active than ␣-MSH, depending on the applied bioassay (49). In the past, a variety of ␣-MSH analogues with increased potency and prolonged activity have been synthesized (49). These bioactive peptides include minimal fragment analogues of ␣-MSH containing the core sequences 6 -9 and 7-9. The truncated ␣-MSH peptides are active at micromolar concentrations and are MC-R subtype-specific. The low molecular weight of such peptide fragments may render them suitable for transdermal delivery in vivo.