There is increasing evidence that serotonin (5-hydroxytryptamine [5-HT]) and distinct 5-HT receptors are involved in the pathogenesis of systemic sclerosis. The aim of this study was to test the hypothesis that tropisetron, a routinely used antiemetic agent previously characterized as a 5-HT3/4 receptor–modulating agent, can directly affect collagen synthesis in vitro and attenuate experimentally induced fibrosis in vivo.
Functional in vitro studies were performed using human dermal fibroblasts (HDFs). Signal transduction studies included immunofluorescence analysis, Western immunoblotting, promoter reporter assays, cAMP/Ca2+ measurements, and use of pharmacologic activators and inhibitors. Gene silencing was performed using small interfering RNA. Putative receptors of tropisetron were detected by semiquantitative reverse transcription–polymerase chain reaction (RT-PCR) and immunofluorescence. The murine model of bleomycin-induced scleroderma was used to assess the antifibrogenic and antifibrotic effects of tropisetron in vivo. Collagen expression in vitro, ex vivo, and in situ was determined by real-time RT-PCR analysis, Western immunoblotting, sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and immunohistochemical analysis.
Tropisetron suppressed collagen synthesis induced by transforming growth factor β1 (TGFβ1). This effect was independent of 5-HT3/4 receptor but was mediated via α7 nicotinic acetylcholine receptor (α7nAChR). Suppression of TGFβ1-induced collagen synthesis occurred via an unknown molecular mechanism not involving modulation of the Smad, cAMP, Akt, c-Jun, or MAPK pathway. In vivo, tropisetron not only prevented skin fibrosis but also reduced the collagen content in established dermal fibrosis induced by bleomycin.
Tropisetron directly reduces collagen synthesis in HDFs via an α7nAChR-dependent mechanism. The antifibrogenic and antifibrotic effects of this agent observed in a mouse model of bleomycin- induced scleroderma indicate the future potential of tropisetron in the treatment of fibrotic diseases such as scleroderma.
Systemic sclerosis (SSc) is a chronic inflammatory connective tissue disease that affects the skin and internal organs. Key pathogenetic features are vascular damage and excessive tissue fibrosis/sclerosis due to increased expression of profibrotic cytokines such as transforming growth factor β1 (TGFβ1) or constitutive activation of TGFβ1-mediated signaling pathways (1, 2). This results in aberrant production of extracellular matrix (ECM). The pathogenesis of SSc is incompletely understood (3, 4), and new effective antifibrotic therapies are needed.
Targeting neuroendocrine receptors that modulate collagen metabolism may be a novel approach for the treatment of SSc. Our group recently observed that α-melanocyte–stimulating hormone suppresses collagen synthesis in human dermal fibroblasts (HDFs) and has antifibrogenic activity in a mouse model of SSc (5, 6). Perhaps the earliest neuroendocrine mediator implicated in the pathogenesis of SSc was serotonin (5-hydroxytryptamine [5-HT]). Serotonin is a vasoactive substance that may mediate the initial phase of vasoconstriction in Raynaud's phenomenon in patients with SSc. Antiserotoninergic strategies have therefore been used in SSc (7). Interestingly, 5-HT stored in platelets induced ECM synthesis, and mice lacking the 5-HT2 receptor were protected against experimentally induced fibrosis (8). Further, cyproheptadine and other 5-HT receptor antagonists are used for the treatment of carcinoid syndrome, in which tumor cells secrete 5-HT and SSc-like skin changes occur (9–11).
Interestingly, 5-HT receptors are expressed in various nonneuronal cells, including those in skin (12). The biology of 5-HT receptors is complex. At least 17 different receptor subtypes have been described (13). Antagonists selective for 5-HT receptor have been developed, and some of them are used in daily medical practice. Tropisetron was characterized as a 5-HT3 receptor antagonist that binds with high affinity to 5-HT3 receptor and with lower affinity to 5-HT4 receptor (14). It has antiemetic activity by antagonizing 5-HT3 receptor in the gut and is approved for the treatment of nausea and vomiting during chemotherapy in patients with cancer.
In this study, we investigated whether tropisetron suppresses experimentally induced skin fibrosis in a mouse model of SSc. We also investigated the mechanism behind this action of tropisetron.
MATERIALS AND METHODS
Cell culture and reagents.
Normal neonatal and adult HDFs (Tebu-Bio) were cultured as previously described (5). Fibroblasts (3T3) (PromoCell) were cultured in RPMI 1640 medium, as previously described (6). Human colon tissue was provided by D. Bettenworth, MD (University of Münster, Germany), and synovial fibroblasts were provided by S. Grässel, MD (University Hospital Regensburg, Germany). Tropisetron was obtained from Novartis Pharma, and 5-HT and bleomycin were purchased from Sigma. TGFβ1 was obtained from PeproTech, forskolin (FSK) and a Phosphodiesterase Inhibitor Set I were from Calbiochem-Novabiochem, and α-bungarotoxin was from Tocris. The specific α7 nicotinic acetylcholine receptor (α7nAChR) agonist AR-R17779 was synthesized at AstraZeneca Pharmaceuticals.
Establishment of HDFs from patients with SSc.
HDFs from the lesional skin of 6 patients with SSc (mean age 54.2 years, range 31–72 years) were established as previously described (15). All patients fulfilled the American College of Rheumatology criteria for a diagnosis of SSc (16). The establishment of HDFs from patients with SSc was approved by the local ethics committees of the Universities of Münster and Cologne.
Cell viability test.
The cytotoxicity of tropisetron was assessed by XTT assay (Roche Diagnostics), as previously described (17).
Total RNA was prepared from cells using an RNeasy Mini Kit (Qiagen) or from skin using TRIzol (Invitrogen). Complementary DNA synthesis was performed with a RevertAid cDNA Kit (Fermentas). Established primers and conditions were used for semiquantitative and real-time RT-PCRs of 5-HT receptors (18), α7nAChR (19, 20), type I collagen (5, 6), plasminogen activator inhibitor 1 (PAI-1), and tissue inhibitor of metalloproteinases 1 (TIMP-1) (21). Primers for α-smooth muscle actin (α-SMA) were as follows: forward 5′-GACCGAATGCAGAAGGAGAT-3′ and reverse 5′-CCACCGATCCAGACAGAGTA-3′. Quantification of the gene expression of each sample was performed using the 2 method (22), using GAPDH as an internal standard with the ground condition set at 1.
Plasmid constructs and promoter reporter assays.
The luciferase plasmid construct (CAGA)9-Lux, which is specific for Smad3, was a gift from S. Dennler and J. M. Gauthier (Glaxo Wellcome) (23). The 3T3 fibroblasts were transfected with DOTAP (Roth), treated with TGFβ1 and tropisetron, followed by promoter reporter assay. Luciferase activity was measured with a kit from Promega.
Determination of C-terminal type I procollagen propeptide (CPI).
We used a commercially available enzyme-linked immunosorbent assay (ELISA) (Takara) to measure type I collagen secretion. HDFs were stimulated with TGFβ1 alone or with tropisetron. Cell culture supernatants were harvested after 24 hours.
Gene silencing of α7nAChR.
Small interfering RNA (siRNA) oligonucleotides with the target sequence GAUAACAGUCUUACUCUCU and the nontargeting control siRNA #2 were designed by Thermo Scientific Dharmacon. HDFs were transfected with siRNA for 24 hours according to the manufacturer's protocol.
Determination of cAMP levels.
HDFs were stimulated with tropisetron, alone or with TGFβ1, for 30 minutes in the presence of a Phosphodiesterase Inhibitor Set I. FSK served as a positive control. Intracellular cAMP levels were measured by a commercially available enzyme immunoassay (Amersham Pharmacia Biotech).
Determination of 5-HT by high-performance liquid chromatography (HPLC).
HDFs were stimulated with TGFβ1 (10 ng/ml) or were left untreated. Supernatants were harvested after 24 hours. Ten microliters of cold 0.4M perchloric acid was added to 50 μl of the cell supernatants or 10−7M 5-HT standards. After centrifugation, supernatants were subjected to isocractic electrochemical detection–HPLC separation using a 25-cm Supelcosil LC-18 column and 50 mM citrate/phosphate buffer solution (pH 3.9) containing 10% acetonitril as eluent. The detection limit was 10 pg 5-HT in 40 μl injection volume (0.25 ng/ml).
Cells were either lysed with boiling Laemmli sample buffer or digested with pepsin in 500 μl 0.5M acetic acid for 24 hours at 4°C for the detection of type I collagen protein. Lysates were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by immunoblotting, as previously described (6). The following antibodies were used: anti–type I collagen (DPC Biermann), anti-Smad2/3 (BD Biosciences), anti–phospho-Smad3 (Ser433/435) (Cell Signaling Technology), anti–5-HT3 receptor (Imgenex), anti–5-HT4 receptor (Millipore), anti-α7nAChR (R&D Systems), anti–phospho-p38, anti–phospho–ERK-1/2, anti-ERK-1/2, anti–phospho–c-Jun, and anti–phospho-Akt (all from Cell Signaling Technology). Equal protein loading was assured by membrane stripping and reprobing with an antibody against β-actin (Santa Cruz Biotechnology) or α-tubulin (Calbiochem-Novabiochem). Densitometric analysis was performed using ImageJ software. All Western immunoblots were repeated at least 3 times.
Immunofluorescence analysis in vitro.
Cells were seeded in 8-well chamber slides, stimulated as indicated, fixed with methanol at −20°C or with 4% paraformaldehyde. Slides were blocked with 5% rabbit or donkey serum followed by incubation for 1.5 hours with a mouse polyclonal antibody against Smad2/3 (1 μg/ml; BD Biosciences) or a rabbit α7nAChR antibody (R&D Systems). Bound primary antibodies were detected with rabbit anti-mouse or donkey anti-rabbit antibodies coupled to Texas Red (Dianova). DAPI (Fluka) was used for nuclear labeling. Cells were imaged by fluorescence microscopy (Carl Zeiss).
HDFs were seeded on round 20-mm–diameter glass slides at 25,000 cells per slide. After 24 hours of fetal calf serum (FCS) deprivation, cells were incubated with 4 μM Fluo-4 AM (Invitrogen) in Tyrode's solution (2 mM CaCl, 140 mM NaCl2, 4 mM KCl, 1 mM MgCl2, 5 mM HEPES, 10 mM glucose, pH 7.4) for 15 minutes, and then the dye was washed out for 10 minutes with Tyrode's solution. Agents were directly applied into the incubation chamber, and fluorescence was recorded using a Zeiss LSM 710 confocal fluorescence microscope (excitation 488 nm, Fluo-4 emission 493–622 nm).
Mouse model of scleroderma.
The bleomycin-induced skin fibrosis model initially described by Yamamoto et al (24) was used. All animal experiments were performed with permission from the local veterinary authorities of the University of Ulm, Germany. C3H mice (4 groups of 5 mice each) were treated with bleomycin (10 μg), tropisetron (10 μg), bleomycin plus tropisetron, or NaCl (negative control). The mice were killed after 21 days. To test whether tropisetron has effects on established skin fibrosis, a post-bleomycin approach was chosen (25). Accordingly, C3H mice were injected with bleomycin (10 μg) for 21 days followed by injections with NaCl (negative control) or tropisetron (10 μg) for a further 14 days. Skin biopsy specimens were obtained for RNA extraction, collagen protein determination, and immunohistochemical analysis.
Histochemistry and immunostaining.
Masson's trichrome staining of mouse skin cryosections was performed according to the manufacturer's protocol (Sigma). Type I collagen immunostaining was performed using rabbit anti-mouse type I collagen or rabbit IgG as an isotype control as primary antibodies (Chemicon). An Alexa Fluor 555–labeled goat anti-rabbit antibody (Caltag) was used as secondary antibody, followed by nuclear staining with DAPI (Fluka). Digital pictures were acquired with a Zeiss Axiophot microscope.
Assessment of dermal thickness.
Skin sections from mice treated with NaCl, bleomycin, tropisetron, or bleomycin plus tropisetron were stained with hematoxylin and eosin. Sections were viewed under a Zeiss Axiophot microscope. Three measurements were performed for each microscopic field, and 5 different fields were analyzed for each mouse (n = 3). Densitometric analysis was performed using ImageJ software.
Determination of collagen protein content ex vivo.
Mouse skin samples were weighed and incubated in 0.5M acetic acid in the presence of protease inhibitor cocktail (Roche Diagnostics). One milligram of pepsin (Sigma) per 500 μl 0.5M acetic acid was added to each sample, followed by incubation for 24 hours at 4°C. After centrifugation, supernatants were precipitated with methanol/chloroform, after which 2× Laemmli sample buffer was added to dissolve the protein pellets. After normalization for wet weight, samples were subjected to SDS-PAGE and stained with 0.5% Coomassie brilliant blue (Bio-Rad).
All experiments were performed at least 3 times. Expression and activity levels were calculated as the mean ± SD, and deviations from normality were assessed by the Shapiro-Wilk test. Differences between 2 mean values were assessed by Student's t-test. One-way analysis of variance and Tukey's test were used for the analysis of Ca2+ fluorescence intensity. P values less than 0.05 were considered significant.
Suppressive effect of tropisetron on TGFβ1-induced collagen synthesis in HDFs.
We first investigated the effect of tropisetron in neonatal HDFs stimulated with TGFβ1. TGFβ1 is a key profibrotic factor that up-regulates the synthesis of collagen and other ECM genes implicated in fibrosis (26, 27). First, the amount of secreted type I collagen was measured by CPI ELISA. In HDFs treated with TGFβ1, CPI secretion was significantly increased compared with control. Tropisetron dose-dependently suppressed TGFβ1-induced CPI secretion by 21–48% (Figure 1A, lower panel). To clarify whether this effect of tropisetron is limited only to deposited type I collagen or involves a reduction in intracellular collagen synthesis, Western immunoblotting was performed. As expected, treatment of HDFs with TGFβ1 (10 ng/ml) increased the expression of type I collagen protein. Tropisetron reduced the intracellular protein expression of type I collagen when costimulated with TGFβ1 compared with stimulation with TGFβ1 alone (Figure 1A, upper panel).
Next, real-time PCR analysis of Colα1(I) and Colα2(I) was performed to investigate whether tropisetron reduces collagen synthesis at the RNA level. TGFβ1 (10 ng/ml) increased the expression of type I collagen messenger RNA (mRNA). Coincubation with tropisetron at a dose of 10 μg/ml suppressed this effect of TGFβ1 by ∼40% for Colα1(I) and by ∼30% for Colα2(I) compared with incubation with TGFβ1 alone (Figure 1B). Tropisetron per se did not alter type I collagen gene expression (additional information is available from the corresponding author). Moreover, the tropisetron doses used in the above experiments were noncytotoxic (additional information is available from the corresponding author).
The suppressive effect of tropisetron on collagen synthesis was confirmed using neonatal HDFs from 2 different donors (Figure 1C). Furthermore, tropisetron attenuated TGFβ1-induced type I collagen mRNA expression in HDFs established from the normal skin of 3 adult donors (Figure 1D). Interestingly, tropisetron also reduced TGFβ1-mediated induction of α-smooth muscle actin (α-SMA), a myofibroblast marker (additional information is available from the corresponding author). In summary, tropisetron suppressed TGFβ1-induced type I collagen synthesis as well as α-SMA mRNA expression in HDFs.
No effect of tropisetron on several canonical signal transduction pathways activated by TGFβ1.
To gain insight into the mechanism by which tropisetron suppresses TGFβ1-mediated ECM synthesis, we analyzed the Smad pathway, which is activated by TGFβ (28, 29). Treatment of HDFs with TGFβ1 led to nuclear translocation of Smad2/3 in HDFs within 30 minutes, as shown by immunofluorescence analysis (Figure 2A). Simultaneous stimulation of cells with tropisetron and TGFβ1, however, did not alter TGFβ1-mediated Smad2/3 translocation (Figure 2A). Moreover, preincubation with tropisetron for 24 hours also did not affect TGFβ1-induced Smad2/3 nuclear accumulation (data not shown). In accordance with this finding, tropisetron did not reduce TGFβ1-induced Ser433/435 phosphorylation of Smad3 in HDFs (Figure 2B).
Next, we performed promoter reporter assays with (CAGA)9-Lux, a Smad3/4-specific luciferase construct (21, 30). In these experiments, 3T3 fibroblasts were used, because transfection of HDFs with this plasmid was not efficient. TGFβ1 treatment increased (CAGA)9-Lux promoter activity compared with no treatment. However, tropisetron failed to alter this TGFβ1 effect, and the tropisetron alone also did not change basal Smad promoter activity (Figure 2C). In support of the lack of an effect of tropisetron on Smad signaling, expression of 2 TGFβ1-regulated Smad target genes, PAI-1 and TIMP-1 (31), was unaffected by the drug (additional information is available from the corresponding author). These data showed that tropisetron does not abrogate TGFβ1-mediated collagen synthesis in HDFs via Smad interference.
We also questioned whether tropisetron affects TGFβ1-mediated activation of additional signal transduction pathways involved in collagen expression (32). TGFβ1 treatment of HDFs resulted in weak phosphorylation of p38, c-Jun, and Akt, but this response was not altered by treatment with tropisetron (additional information is available from the corresponding author). Moreover, neither TGFβ1 nor tropisetron, alone or in combination, altered the phosphorylated state of ERK-1/2 under the given experimental conditions (additional information is available from the corresponding author).
Mediation of the suppressive effect of tropisetron on collagen synthesis by HDFs.
Previously, we demonstrated that cAMP plays an important role in modulating the TGFβ-mediated responses of several ECM genes (21, 30). In order to investigate whether the effect of tropisetron is mediated by activation of the cAMP pathway, we used the adenylate cyclase inhibitor SQ22536. SQ22536 did not affect the suppressive effect of tropisetron on TGFβ1-induced collagen synthesis, as shown by real-time RT-PCR analysis (Figure 3A), suggesting that tropisetron acts in a cAMP-independent manner in HDFs. In accordance with this, tropisetron alone at several doses failed to increase intracellular cAMP levels over control, as shown by ELISA (Figure 3B). In contrast, FSK, an activator of adenylate cyclase, strongly elevated cAMP expression (Figure 3B). Furthermore, neither TGFβ1 alone nor TGFβ1 plus tropisetron at different doses affected intracellular cAMP levels (Figure 3C). Therefore, tropisetron did not elicit its suppressive effect on TGFβ1-induced collagen synthesis in HDFs via activation of the cAMP pathway.
Because both 5-HT3 receptor and 5-HT4 receptor were previously identified as binding sites for tropisetron (14), we next investigated whether HDFs express these receptors. Surprisingly, mRNA expression of both 5-HT3 receptor and 5-HT4 receptor was undetectable in HDFs, as shown by semiquantitative RT-PCR, whereas these receptors were present in the human colon tissue used as positive control (Figure 3D, left panel). In accordance with this, both receptors were not expressed in HDFs at the protein level, as demonstrated by Western blotting (Figure 3D, right panel). In contrast, single protein bands of the expected sizes of 5-HT3 receptor (60 kd) and 5-HT4 receptor (45 kd) were observed in cell lysates of human colon (33, 34) (Figure 3D, right panel). The data showed that 5HT3/4 receptors cannot mediate the suppressive effect of tropisetron on TGFβ1-induced collagen synthesis, because HDFs lack these receptors.
Because the suppressive effect of tropisetron on TGFβ1-mediated collagen synthesis occurred in the absence of exogenous 5-HT, we further sought to determine whether the drug antagonizes endogenously produced 5-HT by HDFs. Conditioned medium from cells was analyzed by HPLC (detection limit 0.25 ng/ml) for the presence of 5-HT, which was detectable in neither the experimental culture medium (0% FCS) nor in conditioned medium from HDFs. In addition, no detectable amounts of 5-HT were observed in supernatants of HDFs stimulated with TGFβ1 for 24 hours (data not shown). These results showed that tropisetron acts in the complete absence of 5-HT.
Effect of α7nAChR on the anti-TGFβ1 effect of tropisetron in HDFs.
Because both putative tropisetron receptors, 5-HT3 receptor and 5-HT4 receptor, are not present in HDFs, we investigated whether α7nAChR could mediate the effect of tropisetron on TGFβ1-induced collagen synthesis. It has been reported that α7nAChR binds tropisetron due to its high homology to 5-HT3 receptor (35, 36). However, the functional significance of this observation is unknown. Expression of α7nAChR in neonatal HDFs was detected via semiquantitative RT-PCR and immunofluorescence analysis (additional information is available from the corresponding author), in accordance with earlier studies (37). Indeed, Ca2+ signaling after treatment of HDFs with the specific α7nAChR agonist AR-R17779 was detected. This effect was blocked by α-bungarotoxin, an α7nAChR antagonist (38). Tropisetron also led to Ca2+ signaling, but it was less potent than that induced by AR-R17779 (see Supplementary videos 1–3, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37809/abstract; additional information is available from the corresponding author).
Next, we performed functional readouts with the above agents. The suppressive effect of tropisetron on TGFβ1-induced type I collagen gene expression in HDFs was completely neutralized by α-bungarotoxin (Figure 4A). Like tropisetron, AR-R17779 attenuated the inductive effect of TGFβ1 on type I collagen transcripts (Figure 4B).
Finally, we confirmed the critical role of α7nAChR in the context of TGFβ1-mediated collagen synthesis by gene silencing. Transfection of HDFs with α7nAChR siRNA markedly decreased mRNA expression of this receptor by ∼90%, whereas nontargeting siRNA had no effect (Figure 4C, inset). Cells were then treated with α7nAChR siRNA or nontargeting siRNA and stimulated with TGFβ1 or TGFβ1 plus tropisetron, respectively. Gene silencing of α7nAChR abolished the suppressive effect of tropisetron on TGFβ1-mediated Colα1(I) and Colα2(I) mRNA expression, as shown by real-time RT-PCR (Figure 4C). Moreover, α7nAChR siRNA but not nontargeting siRNA neutralized the suppressive effect of tropisetron on TGFβ1-mediated CPI secretion (Figure 4D). In summary, these findings identified α7nAChR as the essential and functional tropisetron receptor that mediates the suppressive effect on TGFβ1-induced collagen synthesis in HDFs.
Effect of tropisetron on experimentally induced fibrosis and established skin fibrosis induced by bleomycin.
To assess the in vivo significance of the observed in vitro effects of tropisetron in HDFs, we used an established mouse model of SSc in which dermal fibrosis is induced by subcutaneous injections of bleomycin (24). First, the possible preventive effect of tropisetron on bleomycin-induced skin fibrosis was assessed by determining the amount of collagen at both the RNA and protein levels in skin samples, using real-time RT-PCR. Subcutaneous injection of bleomycin led to significantly increased mRNA expression levels of Colα1(I) and Colα1(III) compared with control. Tropisetron strongly reduced the mRNA expression levels of Colα1(I) (mean ± SD −79 ± 21%) and Colα1(III) (−79 ± 11%) in skin samples from bleomycin-treated mice (Figure 5A). At the protein level, bleomycin treatment likewise resulted in an increase in the amount of extractable collagen in skin, as shown by SDS-PAGE after pepsin digestion. This effect of bleomycin was suppressed by tropisetron (Figure 5B).
The extent of dermal fibrosis was also assessed by histologic and immunohistochemical analysis. Masson's trichrome staining revealed intense dermal fibrosis in bleomycin-treated mice, as visualized by blue staining (Figure 5C). Cotreatment with tropisetron suppressed this effect of bleomycin, while tropisetron alone did not have any effect. The beneficial effect of tropisetron on bleomycin-induced skin fibrosis was confirmed by type I collagen immunostaining. Bleomycin injection increased the extent of deposited dermal collagen compared with the negative control as well as tropisetron alone, while cotreatment with bleomycin and tropisetron led to a strong decline in the extent of collagen staining (Figure 5C). In accordance with these readouts, treatment of mice with bleomycin resulted in an increase of dermal thickness, whereas tropisetron significantly reduced this effect (additional information is available from the corresponding author).
To evaluate whether tropisetron is also operative on established skin fibrosis, we used a post-bleomycin approach described initially by other investigators (25). Accordingly, cutaneous fibrosis was first induced by bleomycin injections for 21 days followed by daily injections with either NaCl (negative control) or tropisetron (10 μg) for a further 14 days. Tropisetron reduced bleomycin-induced collagen induction after the onset of skin fibrosis, as shown by mRNA (Figure 5D, left panel) and protein readouts (Figure 5D, right panel). These findings demonstrated that tropisetron had both preventive (antifibrogenic) and therapeutic (antifibrotic) activity in the given mouse model of bleomycin-induced SSc.
Effect of α7nAChR on dermal fibroblasts from patients with SSc.
In a final set of experiments, we determined the expression of α7nAChR in skin and fibroblasts from patients with SSc. As shown by semiquantitative RT-PCR, transcripts for α7nAChR were detectable in both normal and diseased adult human skin (Figure 6A, top panels). Moreover, α7nAChR transcripts were present in normal fibroblasts and fibroblasts from patients with SSc (n = 3 each) (Figure 6A, bottom panels). To check whether the detected α7nAChR is functionally active in the latter cells, we treated them with TGFβ1 and tropisetron. Tropisetron reduced TGFβ1-induced type I collagen mRNA expression as well as CPI secretion in these cells (Figures 6B and C). In addition, tropisetron reduced TGFβ1-mediated α-SMA induction, as shown by real-time RT-PCR (Figure 6D).
Here, we provide evidence for a modulating effect of tropisetron on collagen synthesis in vitro and in vivo. In vitro, tropisetron dose-dependently suppressed type I collagen synthesis, and this effect was mediated by α7nAChR. Validation of α7nAChR as the crucial mediator of the anticollagenic effects of tropisetron was shown by pharmacologic approaches using specific antagonists and agonists of α7nAChR as well as by gene silencing of this receptor. The receptor α7nAChR belongs to a family of neurotransmitter-gated ion channel proteins, mediating the exchange of ions such as Na+, Ca2+, and K+ (39). Notably, expression of these receptors is not confined to neuronal cells (40) but also occurs in non-neuronal tissue, including skin (41). The α7nAChR detected in HDFs was shown to be functionally active and to mediate the biologic effects of tropisetron in the context of TGFβ1-mediated expression of several ECM target genes.
In order to elucidate the molecular mechanism of the anti-TGFβ1 effect of tropisetron, we investigated several canonical signal transduction pathways of TGFβ that are known to regulate collagen metabolism (30, 32). Tropisetron did not affect the activation of canonical Smad, c-Jun, Akt, or MAPK signal transduction pathways in response to TGFβ1. Moreover, the drug did not alter intracellular levels of cAMP. Previous studies revealed that α7nAChR agonists can activate the JAK/STAT pathway (42). However, inhibition of STAT-3 suppressed collagen production in keloid fibroblasts (43). Moreover, STAT-4 was recently described to exert profibrotic effects by controlling cytokine release in a mouse model of SSc (44). Based on these findings, involvement of the STAT-3/4 pathways in the anticollagenic effect of tropisetron in HDFs is unlikely. Notably, the α7nAChR agonist nicotine was shown to inhibit the production of inflammatory mediators via modulation of the NF-κB pathway in macrophages (45). Further studies are needed to clarify the signaling pathways downstream of α7nAChR that lead to the anticollagenic effect of tropisetron in HDFs.
To assess the in vivo relevance of our in vitro findings, we tested tropisetron in a mouse model of established bleomycin-induced SSc. Most importantly, tropisetron not only prevented bleomycin-induced skin fibrosis but also had antifibrotic effects in established fibrosis; the latter finding suggests a therapeutic potential of this drug in SSc patients with advanced skin (and organ) fibrosis. Notably, it has been reported that mice with targeted disruption of α7nAChR have decreased amounts of type I collagen protein and elastin (46). The discrepancy between the suppressive effect of tropisetron acting via α7nAChR in our in vitro experiments and these in vivo observations in which no inflammatory stimulus was included needs to be further investigated. Indeed, the bleomycin-induced fibrosis model is regarded as an inflammation-driven model for SSc. Therefore, tropisetron, via α7nAChR, may also elicit antiinflammatory effects that could contribute to the attenuating effect on bleomycin-induced skin fibrosis. In animal models of inflammation, selective α7nAChR agonists have been effective in suppressing macrophage cytokine production and inflammatory responses (47–51).
Regarding the future use of tropisetron in SSc, we evaluated the expression of α7nAChRs in the skin and fibroblasts of patients with SSc. The presence of these receptors in the affected skin and fibroblasts of these patients in particular indicated that this target site for tropisetron is preserved. Most importantly, tropisetron was shown to be operative even in HDFs from patients with SSc. As in normal HDFs, tropisetron suppressed TGFβ1-mediated type I collagen synthesis and also attenuated expression of α-SMA, a myofibroblast marker.
In summary, we demonstrated a novel effect of tropisetron, namely the inhibition of collagen synthesis, and provided insight into the molecular mechanism of this effect in HDFs. We identified α7nAChR as the essential receptor that mediates the anti-TGFβ1 effect of tropisetron in HDFs and confirmed the relevance of these observations in vivo. Tropisetron not only prevented experimentally induced skin fibrosis but also had antifibrotic effects on established fibrosis induced by bleomycin. Our findings encourage not only a detailed investigation of the cutaneous cholinergic system in SSc but also the exploitation of tropisetron and other α7nAChR agonists as a future treatment option for fibrotic diseases.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Böhm had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Stegemann, Sindrilaru, Grando, Fiebich, Böhm.
Acquisition of data. Stegemann, Sindrilaru, Heinick, Schulte.
Analysis and interpretation of data. Stegemann, Eckes, del Rey, Müller, Scharffetter-Kochanek, Luger, Böhm.
We are grateful to Mara Apel, Nicole Gross, and Anne Erpenbeck-Leuer (Münster), and Gabi Scherr (Cologne) for expert technical assistance.