Enhanced asthma-related fibroblast to myofibroblast transition is the result of profibrotic TGF-β/Smad2/3 pathway intensification and antifibrotic TGF-β/Smad1/5/(8)9 pathway impairment

Airway remodelling with subepithelial fibrosis, which abolishes the physiological functions of the bronchial wall, is a major issue in bronchial asthma. Human bronchial fibroblasts (HBFs) derived from patients diagnosed with asthma display in vitro predestination towards TGF-β1-induced fibroblast-to-myofibroblast transition (FMT), a key event in subepithelial fibrosis. As commonly used anti-asthmatic drugs do not reverse the structural changes of the airways, and the molecular mechanism of enhanced asthma-related TGF-β1-induced FMT is poorly understood, we investigated the balance between the profibrotic TGF-β/Smad2/3 and the antifibrotic TGF-β/Smad1/5/9 signalling pathways and its role in the myofibroblast formation of HBF populations derived from asthmatic and non-asthmatic donors. Our findings showed for the first time that TGF-β-induced activation of the profibrotic Smad2/3 signalling pathway was enhanced, but the activation of the antifibrotic Smad1/5/(8)9 pathway by TGF-β1 was significantly diminished in fibroblasts from asthmatic donors compared to those from their healthy counterparts. The impairment of the antifibrotic TGF-β/Smad1/5/(8)9 pathway in HBFs derived from asthmatic donors was correlated with enhanced FMT. Furthermore, we showed that Smad1 silencing in HBFs from non-asthmatic donors increased the FMT potential in these cells. Additionally, we demonstrated that activation of antifibrotic Smad signalling via BMP7 or isoliquiritigenin [a small-molecule activator of the TGF-β/Smad1/5/(8)9 pathway] administration prevents FMT in HBFs from asthmatic donors through downregulation of profibrotic genes, e.g., α-SMA and fibronectin. Our data suggest that influencing the balance between the antifibrotic and profibrotic TGF-β/Smad signalling pathways using BMP7-mimetic compounds presents an unprecedented opportunity to inhibit subepithelial fibrosis during airway remodelling in asthma.

. HBFs derived from patients with asthma transdifferentiate into myofibroblasts efficiently and faster than those of their healthy counterparts. HBFs (AS = 10; NA = 6) were cultured in serum-free conditions in the absence or presence of TGF-β 1 (5 ng/mL) for 0-7 days. (A) Then, the cells were fixed with 3.7% formaldehyde, permeabilized, and immunostained for α-SMA (green) and counterstained for DNA (blue), as shown on representative images. Scale bar = 25 μm. (B) The fraction of cells with prominent α-SMA-positive stress fibres in HBF populations was determined using fluorescence microscopy, each in three independent experiments. (C,D) Main markers of myofibroblast formation: α-SMA and EDA-fibronectin levels were assessed using in-cell-ELISA tests, and the results are presented as the mean value of absorbance (450 nm) reflecting the protein content. Data represent the mean ± SEM carried out on HBFs (AS = 10; NA = 6), each in triplicate. (E,F) HBFs (AS = 7; NA = 7) were cultured in serum-free conditions in the absence or presence of TGF-β 1 for 24 h. RT-qPCR analyses of alpha smooth muscle actin (ACTA2) and TGF-β 1 expression were performed. Statistical significances of all experiments were tested using the non-parametric Mann-Whitney test; *p ≤ 0.05, **p ≤ 0.01. Scientific RepoRtS | (2020) 10:16492 | https://doi.org/10.1038/s41598-020-73473-7 www.nature.com/scientificreports/ expression, which was similar in the fibroblasts derived from both study groups. The effect of TGF-β 1 on Smad5 expression was undetectable (Fig. 3A). Parallel immunoblot analyses revealed differences between the Smads at the protein level in the HBFs from both study groups (approximately threefold increases in both the Smad1 and Smad5 levels in unstimulated HBFs from asthmatic donors) (Fig. 3B). However, TGF-β 1 stimulation strongly decreased Smad1 but not Smad5 contents in HBFs from patients with asthma. The level of the E3 ubiquitinprotein ligase Smurf2 was also increased in AS HBFs. A comparison of the Smad1/5/9 pathway activity between HBFs derived from asthmatic and the non-asthmatic donors was performed by immunoblots and immunofluorescence analysis. A representative immunoblot membrane showed a time-dependent increase in the phosphorylation of Smad1/5/9 after TGF-β 1 treatment, resulting in faster and stronger (three times) phosphorylation of Smad1/5/9 in the HBFs from non-asthmatic donors than those from asthmatic donors (Fig. 3C). Moreover, the nuclear localization of phosphorylated Smad1/5/9 complexes, shown on representative images and measured fluorometrically, was also enhanced in HBFs from non-asthmatic donors after TGF-β 1 treatment (Fig. 3D). Since it is well documented that phosphorylated R-Smads translocate into nucleus in complexes with Co-Smads we compared the expression of Smad 4 at the mRNA and protein levels in the HBFs from both study groups. Our results of RT-qPCR analyses showed significantly higher Smad 4 expression in HBF populations derived from asthmatics (Fig. S1D). Interestingly, immunoblot analyses revealed no differences between Smad 4 at the protein level in the HBF populations from AS and NA donors (Fig. S1E).
Silencing of Smad1 in the HBFs derived from non-asthmatic donors leads to Smad2/3 pathway stimulation and FMT activation. To illustrate the impact of the opposing Smad-dependent pathways on FMT, we silenced the Smad1 gene in HBFs. The immunofluorescence staining of p-Smad2/3 proteins after 1 h of TGF-β 1 administration showed that Smad1 silencing had no impact on Smad2/3 pathway activation in the HBFs from asthmatic donors, but it caused a twofold increase in p-Smad2/3 activity (enhanced nuclear translocation of p-Smad2/3 complexes) in the HBFs from non-asthmatic donors (Fig. 4A). This observation was confirmed by fluorometric analysis (Fig. 4B,C). This change resulted in the intensification of the FMT process (Fig. 4A). The immunofluorescence analyses showed that stress fibres consisting of α-SMA were increased after Smad1 silencing in the HBF populations from non-asthmatic donors (Fig. 4A, inserts with plots).
Activation of Smad1/5/(8)9 signalling prevents FMT in HBFs from asthmatics. We used BMP7, which is described as a main activator of the Smad1/5/(8)9 pathway [42][43][44] , to assess the impact of Smad1/5/9 activation on the FMT process in HBF populations. Immunofluorescence analyses showed that BMP7 stimulates the myofibroblast transition of HBFs from asthmatic donors, but this process was much weak than after TGF-β 1 administration (Fig. 5A,B,C). These microscopic observations were confirmed by fluorometric analyses, which revealed shorter, thinner and less enriched α-SMA actin stress fibres in BMP7-treated HBFs compared to TGFβ 1 -stimulated HBFs (Fig. 5A,B). The combination of BMP7 and TGF-β 1 decreased the FMT potential (measured as the number of myofibroblasts) in the HBFs derived from asthmatic patients but not in those from their nonasthmatic counterparts (Fig. 5C). However, α-SMA levels were not reduced after TGF-β 1 and BMP7 treatment in the HBFs derived from asthmatic donors (Fig. 5D,E). These results suggested that BMP exerts an inhibitory effect on the TGF-β 1 -induced FMT by affecting the actin cytoskeleton architecture. Further analyses were performed to estimate the effect of isoliquiritigenin (ISL), described as a small molecule activator of BMP signalling 45 , on the TGF-β 1 -induced FMT in the HBF populations derived from asthmatic patients. ISL (used at a concentration that did not affect cell viability and proliferation, Fig. S1F,G) attenuated the TGF-β 1 -induced phenotypic differentiation of HBFs, as shown by a reduction in the fraction of myofibroblasts in the HBF populations (Fig. 6A,B). Concomitantly, ISL decreased the α-SMA and fibronectin levels in TGFβ 1 -treated HBFs (Fig. 6C,D). Moreover, immunoblot analyses confirmed that ISL-induced FMT attenuation is dependent on Smad1/5/9 pathway overactivation (Fig. 6E).

Discussion
A major problem of bronchial remodelling in asthma is progressive subepithelial fibrosis, which causes irreversible changes in the bronchial wall 1,2,5,46 . The strong accumulation of contractible (overexpression of α-SMA) myofibroblasts secreting excessive amounts of extracellular matrix proteins (fibronectin, collagens, proteoglycans) abolishes the physiological functions of the bronchi. FMT, a crucial event during subepithelial fibrosis observed Figure 2. The TGF-β 1 /Smad2/3 pathway is intensified in HBFs derived from patients with asthma compared to their healthy counterparts. Cells (AS = 7, NA = 7) were cultured in serum-free medium with or without TGF-β 1 (5 ng/mL) for 24 h (A), 48 h (B), 1-60 min (C,D) or 0-240 min (E,F). (A) Smad expression was analysed at the mRNA level using RT-qPCR. (B-D) Smad2, Smad3, and their phosphorylated forms were detected using Western blots. Representative membranes are shown. Densitometric quantification of Smad proteins in relation to β-actin and p-Smad2 or 3 in relation to appropriate Smads (as control proteins) are presented on the graphs as values of relative optical densities (ROD) (n = 6). (E-F) HBFs were fixed, permeabilized, and immunostained for p-Smad2 or p-Smad3. Representative photos were selected. Scale bar = 25 μm. The percentage of cells with p-Smad + nuclei was determined using fluorescence microscopy. Alternatively, p-Smad levels in HBF nuclei were quantified with the fluorometric approach using ImageJ, and the results are presented as the mean fluorescence intensity in relation to the nuclei area. Data on photos (yellow) represent the mean ± SEM of AS = 5, NA = 5; each in min. 100 cells. Statistical significance (C-F) * between HBFs AS TGF vs HBFs NA TGF; (F) # HBFs AS CTRL vs HBFs NA CTRL. Statistical significance was tested using the non-parametric Mann-Whitney test; * p ≤ 0.05, **p ≤ 0.01, *** p ≤ 0.001. www.nature.com/scientificreports/ in asthma, depends on the combined action of inflammatory mechanisms and inherent features of airway wall fibroblasts facilitating their phenotypic transitions in response to humoural factors 4 . The induction of the FMT in HBF populations in vitro by the addition of exogenous TGF-β 1 mimics the in vivo asthmatic process triggered by local secretion of TGF-β 1 (the most important pro-fibrotic cytokine 22 ) within the inflammatory milieu. Therefore, the experimental model, based on the primary HBFs expanded directly from bronchial biopsies of different patients with or without diagnosed asthma, is suitable for the analyses of the FMT background in asthma, as well as for pharmacokinetic studies. Our model may have some limitations related to potential confounders such as a younger age and steroid use in the asthmatic patients in comparison to control subjects. However, due to propagation of HBFs for several passages in a culture, in our opinion, this model gives a possibility to describe the differences between cells isolated from individual patients, which are related to genetic and epigenetic background rather than to a chronic inflammation or to consequences of drugs taken by the patients. When it comes to the age differences between the both groups of patients used in our study they result Figure 3. The TGF-β 1 /Smad1/5/9 pathway is intensified in HBFs derived from patients without asthma compared to asthmatic patients. (A) Cells were cultured in serum-free medium supplemented with TGF-β 1 (5 ng/mL) for 24 h. Then, the mRNA was isolated, and transcripts were analysed using RT-qPCR. (B) Smad1, Smad5, and Smurf2 were detected using Western blots. Representative membranes are shown. Densitometric quantification is presented on the graphs as values of the relative optical densities (ROD) (n = 6) of the bands in relation to that of β-actin (control protein). (C) Phosphorylated forms of Smad1/5/9 (p-Smad1/5/9) were detected using Western blots. Representative membranes are shown. Densitometric quantification is presented on the graph as values of the relative optical densities (ROD) (n = 4) of p-Smads in relation to Smad1 (as the control protein) (D) HBFs cultured in serum-free medium were fixed, permeabilized, and immunostained for p-Smad1/5/9, and the percentage of cells with p-Smad1/5/9-positive nuclei was determined using fluorescence microscopy. Representative photos were selected. Scale bar = 25 μm. The mean fluorescence intensity in relation to the nuclei area (AS = 3, NA = 3 each in 200 cells) was measured using ImageJ. Statistical significance was tested using the non-parametric Mann-Whitney test; *p ≤ 0.05, ***p ≤ 0.001 or the T-test; # p ≤ 0.05.
Scientific RepoRtS | (2020) 10:16492 | https://doi.org/10.1038/s41598-020-73473-7 www.nature.com/scientificreports/ from difficulties in recruiting individuals to the control group, from which we could have obtained consent and take samples from the lower respiratory tract by bronchoscopy. Although aging is recognized as a major risk factor for fibrotic diseases 47 and FMT has been shown in in vitro studies to strongly increase with age 48 , in our studies a younger age of the asthmatic patients, in comparison to the control subjects, did not affect the enhanced TGF-β 1 -induced asthma-related FMT. The FMT process is inextricably linked to profibrotic signalling from TGF-β receptors through a TGF-β/ Smad2/3-dependent pathway. Our previous 12,15,16 and current observations indicate that FMT is more efficient in the fibroblasts from asthmatic donors after TGF-β 1 treatment than those from their non-asthmatic counterparts. Based on many reports describing the antifibrotic role of the Smad1/5/9 pathway in various diseases 35,49,50 , we compared profibrotic Smad2/3 and antifibrotic Smad1/5/9 signalling activity between the HBFs from asthmatic and non-asthmatic donors. The association of TGF-β/Smad2 signalling activity with airway remodelling in asthma has been poorly investigated 51 . We are the first to demonstrate that the intensification of the fibrogenic potential of AS HBFs was associated with increased activity of the profibrotic TGF-β/Smad2/3-dependent pathway in these cells. Both the level of phosphorylation of the Smad2 and Smad3 proteins and the nuclear translocation of their complexes were increased in AS HBFs, while NA HBFs showed much weaker activation of this pathway in vitro (Fig. 7). Such an increase in the phosphorylation of Smad2 and Smad3 was previously observed in mouse lung tissues after prolonged allergen challenge (ovalbumin) 52,53 and in humans in acute allergen-induced remodelling during asthma 54,55 . The observed intensification of the TGF-β/Smad2/3 pathway in AS HBFs may be related to the enhanced expression of TGF-β receptors, especially ALK5 ( Fig. S1B) 56 , because the TGF-βRII level is the same in NA and AS HBFs 36,56 . Dysregulation of the Smad2/3 pathway through blocking of TGF-β receptors (pharmacologically and using specific antibodies) was used to decrease the FMT potential and fibrosis, but this treatment may disturb the balance of normal processes in other cells of the organism 37,49,57 . Therefore, substances that act downstream in the TGF-β/Smad2/3 pathway, such as inhibitors of Smad2/3 phosphorylation or nuclear translocation inhibitors (SIS3, statins, fenofibrate, etc.), are continually being assessed 16,37,38 .
However, some studies have shown that TGF-β can induce not only Smad2/3 phosphorylation via TGF-βRII/ALK5 binding but also Smad1/5 phosphorylation dependent on different receptor complex activities (e.g., ALK1/2/3/6) in a cell type-specific manner 58,59 . Taken together, these observations suggest that further studies are needed to determine whether TGF-β 1 can differentially bind to heteromeric complexes comprised of varied mammalian TGF-β family type I and type II receptors and can thus activate TGF-β/Smad pathways in diverse mechanisms in HBFs from asthmatic and non-asthmatic donors.
Smad1/5/9 pathway activation has been shown previously to have potent antifibrotic effects in liver fibrosis 41 , renal fibrosis 39, 60 , pulmonary fibrosis [61][62][63] and myocardial fibrosis 64 in animals and humans. We reported that TGF-β/Smad1/5/9 pathway activity is increased in HBFs from non-asthmatic donors compared to HBFs obtained from patients with asthma. A co-Smad4 is necessary for the nuclear translocation of Smad complexes, such as p-Smad2/3/4 and p-Smad1/5/4 28 . The observed impairment of Smad1/5 phosphorylation in TGF-β-treated HBFs from asthmatic donors along with a lack of differences in Smad4 levels between NA and AS HBFs (Fig. S1D,E) www.nature.com/scientificreports/ indicated overactivation of the profibrotic Smad2/3 pathway in AS HBFs. Our data are consistent with the observations that the basal expression of R-BMPs and Smad1/5 pathway activation are reduced in bronchial cells from patients with mild asthma 65 . The imbalance between the Smad1/5/9 and Smad2/3 ratio in AS HBFs can intensify the production of TGF-β 1 , fibronectin, α-SMA and collagens, leading to enhanced FMT and subepithelial fibrosis. We showed that TGF-β 1 stimulation strongly decreased the Smad1 level in AS HBFs. This change may be caused by the increased Smurf2 level, which can degrade Smad1 29, 66 , resulting in additional impairment of the Smad1/5/9 pathway and inhibition of its protective antifibrotic role. Moreover, the silencing of Smad1 in NA HBFs by siRNA mediated the signal transfer from the antifibrotic TGF-β/Smad1/5/9 to the profibrotic TGF-β/ Smad2/3 pathway and stimulated FMT in vitro. BMP7 is known as a cytokine that can antagonize the TGF-β 1 -dependent fibrogenic activity of mouse pulmonary myofibroblastic cells 62 . This cytokine is also a main stimulator of the Smad1/5/9 pathway 43 . We showed that canonical BMP7 stimulation of the Smad1/5/9 pathway dysregulates α-SMA synthesis in both untreated and TGF-β 1 -treated HBFs from healthy donors. Similar results have been obtained by Pegorier et al. and Midgley et al. on normal human lung fibroblasts cultured in vitro in combination with TGF-β 1 67, 68 . In our model, BMP7 stimulation of the Smad1/5/9 pathway reduced the efficiency of TGF-β 1 -induced FMT in AS HBFs by decreasing the percentage of myofibroblasts in HBF populations. Surprisingly, cotreatment of AS HBFs with TGF-β 1 and BMP7 did not decrease α-SMA production; however, we observed disruption of α-SMA incorporation into actin stress fibres 16,38 and a consequent reduction in the formation of myofibroblasts. It is well documented that myofibroblast formation in a variety of tissues, including human bronchi, demands not only humoral factors (e.g. TGF-β) but also mechanical stimuli (e.g. appropriate cell stiffness and/or the interaction of cells with ECM proteins) 4 . Previously we have shown that HBFs from asthmatics form an increased number of thick stress fibres accompanied by enlarged focal adhesions which correlate with high elastic modulus of asthmatic HBFs and their increased efficiency to the TGF-β-induced FMT 14 . Taken together, our results demonstrate that BMP-7 (without affecting the level of α-SMA) induces changes in the actin cytoskeleton architecture in HBFs causing decreased cell stiffness, which correlate with the reduced FMT in BMP-7-treated HBF populations. These findings may be related to the previously shown impairment of the TGF-β/Smad1/5/9 pathway in AS HBFs. This impairment may be a reason for the dysregulation of the Smad1/5/9 pathway by upstream stimulation by BMP7 and TGF-β 1 (Fig. 7). Additionally, a recent report showed that in a mouse model of asthma, the balance between TGF-β 1 and BMP7 could be used to predict the intensity of lung fibrosis 61 . Moreover, the levels of the TGF-β 1 and TGF-β receptors, unlike the level of BMP7 and its receptors, were higher in the bronchi of asthmatics than the non-asthmatic controls 61,65 . The results from multiple fibrotic disease models demonstrated that BMP7 expression is downregulated during tissue fibrosis and that treatment with BMP7 prevented fibrotic changes 36,46,61 . Despite the promising reports from various animal studies, exogenous BMP7 has not yet been approved for use in human fibrotic diseases or even in preclinical studies, probably because BMP7 is costly and its clinical use is limited due to the supraphysiological doses required for therapeutic efficacy, which cause severe side effects. BMP7 is also likely rapidly cleared from the blood 49 . Given the above results, we decided to search for small molecule activators of BMP signalling. Based on the results of Vrijens et al., we chose ISL as a small activator of Smad1/5/9 signalling 45 . We demonstrated for the first time that in our in vitro model, ISL at non-cytotoxic and non-cytostatic concentrations notably attenuated the TGF-β 1 -induced phenotypic transition of AS HBFs into myofibroblasts through Smad1/5/9 pathway stimulation (Fig. 7). As we noted previously, the experimental model used in this study mimics the basic properties of HBFs during asthma and the chronic inflammation crucial for this disease. Thus, our results indicated that this agent may attenuate in vivo airway wall remodelling via a TGF-β desensitizing effect on HBFs. Earlier studies showed that ISL decreased the levels of IL-4 and IL-5 in allergic asthma, significantly improving the function of the lungs 69,70 , and it may relax guinea-pig trachea through multiple intracellular actions 71 . ISL supplementation reduced high fat diet (HFD)-induced adipose tissue fibrosis and the expression of fibrosis-related genes in mice 72 . A recent report showed that ISL can also attenuate monocrotaline-induced pulmonary hypertension via its anti-inflammatory and antiproliferative actions in rats 73 . Thus, our and other results suggest that ISL is a strong candidate for antifibrotic therapy in asthma.
Taken together, our data shed light on FMT signalling in bronchial asthma when the disordered balance between the Smad2/3 and Smad1/5/9 pathways is considered. Based on our results, we strongly suggest that enhanced asthma-related FMT (the key event of subepithelial fibrosis in the bronchial wall of asthmatic patients) Figure 5. Smad2/3 versus Smad1/5/9 activation balance plays a pivotal role in fibroblast to myofibroblast transition of HBFs. (A,B) HBFs were cultured in serum-free medium without (Ctrl) or with TGF-β 1 (5 ng/ ml) in the absence or presence of BMP7 (100 ng/ml) for 7 days. Then, the cells were fixed, permeabilized, and immunostained for α-SMA (green) and counterstained for F-actin (red) and DNA (blue), as shown on representative images. The intensity of actin fibre fluorescence in the sections is presented on plot profiles and quantified in graphs (under the photos). Scale bar = 200 μm. (C,E) The fraction of cells with prominent α-SMApositive stress fibres in HBF populations was determined using fluorescence microscopy in three independent experiments. (D,F) Analyses of α-SMA content were carried out using an in-cell ELISA test, and the results are presented as the mean value of absorbance (450 nm) reflecting the protein content. Data represent the mean ± SEM carried out on HBFs (AS = 10; NA = 6), each in triplicate. (G) α-SMA was detected using Western blots. Representative membranes are shown. Glyceraldehyde 3 phosphate dehydrogenase (GAPDH) was used as a loading control. (H) Densitometric quantification of membranes is presented on the graph as values of the relative optical densities (ROD) (n = 2) of α-SMA in relation to GAPDH (as control protein) (A,C,D) HBFs from asthmatics, (B,E,F) HBFs from non-asthmatics. Statistical significance was tested using were determined using Statistical significance was tested using one-way ANOVA with the Bonferroni multiple comparison post hoc test; ns -non statistically significant, # p ≤ 0.05, ## p ≤ 0.01, ### p ≤ 0.001. www.nature.com/scientificreports/ depends on TGF-β/Smad1/5/(8)9 pathway impairment (Fig. 7). In connection with the above findings, the usage of substances stimulating this antifibrotic pathway (such as ISL) may prove to be an effective way to suppress FMT and subepithelial fibrosis in asthma, but the exact molecular mechanism of the phenomenon reported here requires further studies.

Materials and methods
Patient characteristics. For the study, 10 patients with asthma severity 3-5 (according to the Global Initiative for Asthma (GINA) classification) and 7 control subjects were included. All patients were treated in the Department of Medicine of the Jagiellonian University Medical College and remained in stable clinical conditions. The control subjects were selected from the group of patients who were referred for bronchoscopy due to prolonged cough. These patients had previous thorough, extensive diagnostics, during which other diseases causing cough were excluded by means of anamnesis and additional tests (imaging, lung function tests, histopathological examination of bronchial samples, allergic tests, microbiological tests). They were finally diagnosed with post-infectious or idiopathic cough. All participants have never smoked. COPD was excluded by the history and the correct result of spirometry, including FEV 1 /FVC ratio after bronchodilator > 0.7. Therefore, it did not affect the results. www.nature.com/scientificreports/ Detailed characteristics of asthma patients and control subjects are presented in Table 1.
Although the studied groups of donors differ in the mean age, in our opinion, a younger age of asthma patients in comparison to control subjects has no effect on the study outcomes. Age differences are mainly due to the fact that it was difficult to recruit theoretically "healthy" subjects for the control group and to collect from them samples from the lower respiratory tract by bronchoscopy. Therefore, we recruited all suitable subjects regardless their age.
The study was approved by the Jagiellonian University Ethics Committee (Decision No. 122.6120.16.2016). All patients and control subjects gave written informed consent to participate in the study.
All methods were performed in accordance with the relevant guidelines and regulations.
Isolation and culture of primary HBFs. Primary HBFs were established from the explants of bronchial biopsies obtained from patients during bronchoscopy using a fiberoptic bronchoscope (Olympus). Then, the explants were washed with cold phosphate buffered saline (PBS, Corning) and transferred to fibroblast growth medium (FGM; Fibroblast Basal Medium containing 2% FBS and supplements from the FGM-2 Bullet Kit; Lonza) with collagenase IV (Worthington, USA; 1 mg/ml) at 37 °C. After 5-6 h of incubation with intensive mixing several times, the digested samples were centrifuged for 5 min at 300×g, transferred to Petri dishes and Immunofluorescence studies. HBFs were cultured for different periods (1 min-2 h for p-Smads; 7 days for α-SMA staining) in serum-free medium in the presence or absence of TGF-β 1 and then fixed with 3.7% formaldehyde/PBS. The following antibodies were used: anti-Smads/p-Smads from Smad1/5/9 (#12656) and Smad2/3 (#12747) Antibody Sampler Kits (Cell Signaling; all: 1:100). Stress fibres containing α-SMA were visualized using mouse monoclonal IgG against α-SMA (Sigma-Aldrich, St. Louis, MO, USA) with appropriate secondary antibodies (goat-anti mouse IgG conjugated with Alexa Fluor 488). Samples were counterstained by phalloidin conjugated with Alexa Fluor 546 (Life Technologies, Thermo Fisher Scientific). Some samples were counterstained with Hoechst 33258 (1 µg/ml, Sigma-Aldrich, St. Louis, MO, USA). The images were acquired using a Leica DMI6000B inverted microscope (Leica Microsystems, Wetzlar, Germany) equipped with LAS-X software for image processing. Smad signalling pathway activation is presented as a percentage of cells with strong signals of the phosphorylated forms of Smads from the nuclear area. In turn, the effectiveness of FMT was determined by the percentage of cells with α-SMA-positive microfilaments. Quantification of the obtained results was performed using plot profiles perpendicular to bundles of stress fibres with quantification of the signal for particular spikes or fluorescence intensity of the nuclear areas (for Smad activity). Fiji ImageJ 1.51 s was used to measure these signals. Finally, the averaged fluorescence signal of the analysed p-Smads (for at least 100 cell nuclei) in relation to the averaged DNA-specific fluorescence intensity www.nature.com/scientificreports/ from the surface of proper cell nuclei, was compared between the conditions. Cytofluorimetric analyses of p-Smad levels were performed using the same excitation/exposure settings 16, 38 . Immunoblots. Preparation of protein supernatant from cultured cells was described previously 16,38 . Samples containing 20 µg of proteins were electrophoresed on a 10% SDS-polyacrylamide gel and transferred to PVDF membranes (Bio-Rad) according to a previously described protocol 16,38 . Next, the membranes were incubated overnight at 4 °C with primary antibodies against Smad proteins from the Smad2/3 Antibody Sampler Kit and the Smad1/5/9 antibody Sampler Kit, mouse monoclonal IgG against α-SMA, mouse monoclonal IgG against β-tubulin, mouse monoclonal IgG against β-actin, mouse monoclonal IgG against vinculin and mouse monoclonal IgM against glyceraldehyde 3 phosphate dehydrogenase (GAPDH) (All: Sigma-Aldrich, St. Louis, MO, USA; 1:1000), diluted in 1% BSA/PBS. After three washes with Tris-buffered saline with Tween 20 (TBST), the membranes were exposed to horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG (all: 1:3000, Life Technologies). Band detection was performed using Luminata Crescendo Western HRP Substrate (Merck Millipore), and the chemiluminescence imaging system ChemiDoc XRS + (Bio-Rad) was used. Band intensities were quantified using Fiji ImageJ 1.51 s freeware.
Cell-based enzyme-linked immunosorbent assay (in-cell-ELISA). HBFs were grown in 96-well plates in serum-free medium alone or in combination with human recombinant TGF-β 1 (5 ng/ml) for 7 days. An in cell ELISA protocol 38 was performed using antibodies against α-SMA (the same as used in the immunofluorescence analyses and Western blot) or fibronectin (1:2000 in 1% BSA/PBS). The results are presented as an absorbance value (450 nm; Microplate Reader, Thermo Scientific, Multiskan FC) corresponding to the relative amount of protein levels.
Real-time PCR. Isolation of mRNA from HBFs was performed using the GeneMATRIX Universal RNA/ miRNA Purification Kit (EURx, Gdańsk, Poland) according to the manufacturer's protocol. The concentration of isolated mRNA was measured in a NanoDrop spectrophotometer (Implen) at OD260/280 nm. Two thousand nanograms of extracted mRNA, the NG dART RT-PCR Kit (EURx) and C1000 Touch Thermal Cycler (Bio-Rad) were used for cDNA synthesis. Gene expression was measured with SYBR Green PCR Master Mix (Applied Biosystems) and specific primer sets (described in Table 2; all from Genomed, Warszawa, Poland) using the 7500 Fast System (Applied Biosystems). Relative amounts of genes were estimated using the quantification threshold value recalculated against GAPDH transcripts by the ∆ CT method [ ∆ CT refers to CT (tested gene) -CT (GAPDH) ] and are presented as a 2 -ΔCt mean value. Statistical analysis. The statistical significances were determined using the non-parametric Mann-Whitney test (the comparison between groups HBFs NA and HBFs AS) or the t-test (the comparison in one group between two different experimental conditions) due to the small number of groups and a lack of normal distribution of data. For the data with normal distribution of data (number of measurements > 50), statistical significance was tested using one-way ANOVA with the Bonferroni multiple comparison post hoc test. The differences were considered to be statistically significant at probability levels of *p < 0.05%, **p < 0.01%, ***p < 0.001%. Each parameter was calculated as the mean (± SEM).  TTC CAG CCA TCC TTC AT  CCG TGA TCT CCT TCT GCA TT   Smad1  ACC TGC TTA CCT GCC TCC TG  CAT AAG CAA CCG CCT GAA CA   Smad2  CGT CCA TCT TGC CAT TCA CG  CTC AAG CTC ATC TAA TCG TCCTG   Smad3  GCG TGC GGC TCT ACT ACA TC  GCA CAT TCG GGT CAA CTG GTA   Smad5  CTG GGA TTA CAG GAC TTG ACC AAG TTC CAA TTA AAA AGG GAGGA   TGF-β 1  AGG GCT ACC ATG CCA ACT TCT CCG GGT TAT GCT GGT TGT ACA   GAPDH  GAA GGT GAA GGT CGG AGT  GAA GAT GGT GAT GGG ATT TC