Design and Synthesis of Matrine Derivatives as Novel Anti-Pulmonary Fibrotic Agents via Repression of the TGFβ/Smad Pathway

A total of 18 matrine derivatives were designed, synthesized, and evaluated for their inhibitory effect against TGF-β1-induced total collagen accumulation in human fetal lung fibroblast MRC-5 cell lines. Among them, compound 3f displayed the most potent anti-fibrotic activity (IC50 = 3.3 ± 0.3 μM) which was 266-fold more potent than matrine. 3f significantly inhibited the fibroblast-to-myofibroblast transition and extracellular matrix production of MRC-5 cells. The TGF-β/small mothers against decapentaplegic homologs (Smad) signaling was also inhibited by 3f, as evidenced by inhibition of cytoplasm-to-nuclear translocation of Smad2/3 and suppression of TGF-β1-induced upregulation of TGF-β receptor type I (TGFβRI). Additionally, 3f exhibited potent inhibitory effects against TGF-β1-induced fibroblasts migration. These data suggested that 3f might be a potential agent for the treatment of idiopathic pulmonary fibrosis via repression of the TGFβ/Smad signaling pathway.


Introduction
Idiopathic pulmonary fibrosis (IPF) is a progressive and irreversible lung disease characterized by abnormally excessive accumulation of extracellular matrix (ECM) [1]. It is associated with high morbidity and mortality, with a median survival ranging from only three to five years from diagnosis [2]. Unfortunately, treatments for IPF were quite limited, and to date, only two small molecule drugs (Pirfenidone and Nintedanib) were approved by FDA for IPF treatment; nonetheless their treatment effect was not satisfactory [3,4]. Therefore, novel anti-IPF agents were urgently needed.
Matrine, a pyrrolizidine alkaloid isolated from traditional Chinese medicine Sophora flavescens, attracted our interest for its proved anti-fibrotic activities. Yu et al. [5] demonstrated that matrine exhibited significant inhibitory effect against liver fibrosis by reducing the expression of transforming growth factor-β1 (TGF-β1) and enhancing the activity of hepatocyte growth factor. The anti-fibrotic activities of matrine on cardiac fibrosis via inhibition of TGF-β/Smad pathway [6] or affection of ATF6 signaling pathway [7] had also been suggested by previous studies. In addition, two independent clinical studies [8,9] revealed that 150 mg/d matrine in combination of prednisone could significantly demonstrated that oxymatrine, the N-oxide analogue of matrine, could significantly reduce serum matrix metalloproteinase-9 (MMP-9) and TGF-β1 level in IPF patients. Collectively, matrine is of great potency for IPF treatment and worthy of further investigation. MASM ( Figure 1) was a matrine derivative discovered by Xu et al. [11] which exhibited potent anti-fibrotic activities. Their research also suggested that the introduction of alkyl amine groups at the 13-position of matrine had a significant impact on anti-liver fibrosis [11] and anti-inflammatory activities [12]. Taking MASM as a lead, we changed the thioamide motif back to the original amide structure of matrine due to the water-instability of MASM [13], and performed structural optimization by varying the alkylamino substituent groups at the 13-position of matrine.
Indole is a privileged scaffold in anti-fibrotic drug discovery. For instance, Xia and coworkers [14] demonstrated that indole derivative DIM could attenuate renal fibrosis via inhibiting fibroblastto-myofibroblast transition and Smad2/3 phosphorylation. Additionally, in an anti-fibrotic screening of indole compound library, Shima et al. [15] observed that MA-35 displayed potent anti-fibrotic activities via inhibition of TGF-β1 signaling and suppression of ECM production. Moreover, indole scaffold could also be found in the chemical structure of anti-fibrotic drug Ruxolitinib [16]. Therefore, a series of 13-indolyl substituted derivatives was also designed and synthesized. Herein, we describe the synthesis of 18 matrine derivatives, SAR analysis, as well as the primary mechanism of action of the key compound.

Chemistry
The synthetic routes of all target compounds were outlined in Scheme 1. The aliphatic amine substituted derivatives 2a-d were synthesized by microwave-assisted water-mediated Aza-Michael addition. Reactions of aliphatic amines and sophocarpine (1), a natural dehydrogenation analogue of matrine, proceeded smoothly under microwave irradiation to give desired products in yields of 90%-95%. Then, a series of indole substituted derivatives 3a-n were obtained by Michael addition of sophocarpine and indole in the presence of KHMDS in yields of 35-77%. All of the desired products were purified with silica gel chromatography using dichloromethane/methanol as eluents. The Michael addition reaction introduced a new chiral center at the 13 position of matrine. Previous literature [12] determined the configuration of 13-position to be 13S, however, no experimental evidence supporting this determination was reported. In this study, we found that most of the synthesized compounds appeared in pairs of 13S/13R stereoisomers, of which the 13Sisomerswere the major stereoisomers. Nonetheless, four compounds (2a, 2b, 3f, 3j) appeared in pure 13S form, and we determined the configuration of 13 position based on NOE experiment of 3j (See Supplementary Materials). As illustrated in Figure 2, we observed NOE correlation between H-13 and H-7, indicating that these protons were in the same face. Thus, we confirmed that the major stereoisomers were 13S-isomers.

Inhibition of TGF-β1-Induced Collagen Accumulation by Target Compounds
Excessive accumulation of ECM in the alveolar parenchyma and progressive scarring of lung tissue are major characteristics of IPF [17]. TGF-β1 is the most potent inducer of ECM production as well as the most significant fibrotic cytokines characterized so far [18]. It promotes IPF disease progression by inducing differentiation of fibroblast to myofibroblast and upregulating the secretion of ECM proteins [18,19], especially collagen, which is the most abundant protein in ECM [20]. Therefore, we established an in vitro anti-fibrotic screening model based on TGF-β1-induced collagen accumulation in human fetal lung fibroblasts (MRC-5 cell lines). The inhibitory effect against collagen accumulation of the synthesized compounds were quantitively determined by Sircol collagen assay, a colorimetric bioassay which was also adopted by Intermune incorporation (the innovator of Pirfenidone) for anti-fibrotic screening of Pirfenidone and its derivatives [21]. Cytotoxicity of all The Michael addition reaction introduced a new chiral center at the 13 position of matrine. Previous literature [12] determined the configuration of 13-position to be 13S, however, no experimental evidence supporting this determination was reported. In this study, we found that most of the synthesized compounds appeared in pairs of 13S/13R stereoisomers, of which the 13S-isomerswere the major stereoisomers. Nonetheless, four compounds (2a, 2b, 3f, 3j) appeared in pure 13S form, and we determined the configuration of 13 position based on NOE experiment of 3j (See Supplementary Materials). As illustrated in Figure 2, we observed NOE correlation between H-13 and H-7, indicating that these protons were in the same face. Thus, we confirmed that the major stereoisomers were 13S-isomers. The Michael addition reaction introduced a new chiral center at the 13 position of matrine. Previous literature [12] determined the configuration of 13-position to be 13S, however, no experimental evidence supporting this determination was reported. In this study, we found that most of the synthesized compounds appeared in pairs of 13S/13R stereoisomers, of which the 13Sisomerswere the major stereoisomers. Nonetheless, four compounds (2a, 2b, 3f, 3j) appeared in pure 13S form, and we determined the configuration of 13 position based on NOE experiment of 3j (See Supplementary Materials). As illustrated in Figure 2, we observed NOE correlation between H-13 and H-7, indicating that these protons were in the same face. Thus, we confirmed that the major stereoisomers were 13S-isomers.

Inhibition of TGF-β1-Induced Collagen Accumulation by Target Compounds
Excessive accumulation of ECM in the alveolar parenchyma and progressive scarring of lung tissue are major characteristics of IPF [17]. TGF-β1 is the most potent inducer of ECM production as well as the most significant fibrotic cytokines characterized so far [18]. It promotes IPF disease progression by inducing differentiation of fibroblast to myofibroblast and upregulating the secretion of ECM proteins [18,19], especially collagen, which is the most abundant protein in ECM [20]. Therefore, we established an in vitro anti-fibrotic screening model based on TGF-β1-induced collagen accumulation in human fetal lung fibroblasts (MRC-5 cell lines). The inhibitory effect against collagen accumulation of the synthesized compounds were quantitively determined by Sircol collagen assay, a colorimetric bioassay which was also adopted by Intermune incorporation (the innovator of Pirfenidone) for anti-fibrotic screening of Pirfenidone and its derivatives [21]. Cytotoxicity of all

Inhibition of TGF-β1-Induced Collagen Accumulation by Target Compounds
Excessive accumulation of ECM in the alveolar parenchyma and progressive scarring of lung tissue are major characteristics of IPF [17]. TGF-β1 is the most potent inducer of ECM production as well as the most significant fibrotic cytokines characterized so far [18]. It promotes IPF disease progression by inducing differentiation of fibroblast to myofibroblast and upregulating the secretion of ECM proteins [18,19], especially collagen, which is the most abundant protein in ECM [20]. Therefore, we established an in vitro anti-fibrotic screening model based on TGF-β1-induced collagen accumulation in human fetal lung fibroblasts (MRC-5 cell lines). The inhibitory effect against collagen accumulation of the synthesized compounds were quantitively determined by Sircol collagen assay, a colorimetric bioassay which was also adopted by Intermune incorporation (the innovator of Pirfenidone) for anti-fibrotic screening of Pirfenidone and its derivatives [21]. Cytotoxicity of all compounds were also determined against MRC-5 cells, so as to calculate selectivity index (SI, CC 50 /IC 50 ). Higher SI values indicates better safety profile. The structures and activity data were summarized in Table 1. Pirfenidone was used as positive control, and its anti-fibrotic IC 50 value determined in this study was 1320 ± 98 µM, which was consistent with Intermune's results (23-69% inhibition of collagen accumulation at 1000 µM concentration for Pirfenidone derivatives [21]). compounds were also determined against MRC-5 cells, so as to calculate selectivity index (SI, CC50/IC50). Higher SI values indicates better safety profile. The structures and activity data were summarized in Table 1. Pirfenidone was used as positive control, and its anti-fibrotic IC50 value determined in this study was 1320 ± 98 μM, which was consistent with Intermune's results (23-69% inhibition of collagen accumulation at 1000 μM concentration for Pirfenidone derivatives [21]). The SAR investigation was initiated with variation of the alkylamino substituent groups at the 13 position of matrine, by which four derivatives were obtained (2a-d). As shown in Table 1, the IC50 values of 2a-d were 65.3 μM to 255.8 μM, which improved the potency of matrine by 3-to 13-folds, corroborating that 13 position substitution was beneficial for anti-pulmonary fibrosis activity. However, it seemed that the anti-fibrotic potency was largely influenced by steric effect of alkylamino substituents, and smaller substituent was favored. Introduction of bulky substituents such as cycloalkylamino group (2c-d) or dimethylamino group (2b) resulted in 1.4-to 3.9-fold decrease in anti-fibrotic potency compared to methylamino substituted derivative (2a). As a consequence, the steric influence in alkylamino substituted derivatives limited the scope for further improvement.  compounds were also determined against MRC-5 cells, so as to calculate selectivity index (SI, CC50/IC50). Higher SI values indicates better safety profile. The structures and activity data were summarized in Table 1. Pirfenidone was used as positive control, and its anti-fibrotic IC50 value determined in this study was 1320 ± 98 μM, which was consistent with Intermune's results (23-69% inhibition of collagen accumulation at 1000 μM concentration for Pirfenidone derivatives [21]). The SAR investigation was initiated with variation of the alkylamino substituent groups at the 13 position of matrine, by which four derivatives were obtained (2a-d). As shown in Table 1, the IC50 values of 2a-d were 65.3 μM to 255.8 μM, which improved the potency of matrine by 3-to 13-folds, corroborating that 13 position substitution was beneficial for anti-pulmonary fibrosis activity. However, it seemed that the anti-fibrotic potency was largely influenced by steric effect of alkylamino substituents, and smaller substituent was favored. Introduction of bulky substituents such as cycloalkylamino group (2c-d) or dimethylamino group (2b) resulted in 1.4-to 3.9-fold decrease in anti-fibrotic potency compared to methylamino substituted derivative (2a). As a consequence, the steric influence in alkylamino substituted derivatives limited the scope for further improvement. compounds were also determined against MRC-5 cells, so as to calculate selectivity index (SI, CC50/IC50). Higher SI values indicates better safety profile. The structures and activity data were summarized in Table 1. Pirfenidone was used as positive control, and its anti-fibrotic IC50 value determined in this study was 1320 ± 98 μM, which was consistent with Intermune's results (23-69% inhibition of collagen accumulation at 1000 μM concentration for Pirfenidone derivatives [21]). The SAR investigation was initiated with variation of the alkylamino substituent groups at the 13 position of matrine, by which four derivatives were obtained (2a-d). As shown in Table 1, the IC50 values of 2a-d were 65.3 μM to 255.8 μM, which improved the potency of matrine by 3-to 13-folds, corroborating that 13 position substitution was beneficial for anti-pulmonary fibrosis activity. However, it seemed that the anti-fibrotic potency was largely influenced by steric effect of alkylamino substituents, and smaller substituent was favored. Introduction of bulky substituents such as cycloalkylamino group (2c-d) or dimethylamino group (2b) resulted in 1.4-to 3.9-fold decrease in anti-fibrotic potency compared to methylamino substituted derivative (2a). As a consequence, the steric influence in alkylamino substituted derivatives limited the scope for further improvement. The SAR investigation was initiated with variation of the alkylamino substituent groups at the 13 position of matrine, by which four derivatives were obtained (2a-d). As shown in Table 1, the IC 50 values of 2a-d were 65.3 µM to 255.8 µM, which improved the potency of matrine by 3to 13-folds, corroborating that 13 position substitution was beneficial for anti-pulmonary fibrosis activity. However, it seemed that the anti-fibrotic potency was largely influenced by steric effect of alkylamino substituents, and smaller substituent was favored. Introduction of bulky substituents such as cycloalkylamino group (2c-d) or dimethylamino group (2b) resulted in 1.4-to 3.9-fold decrease in anti-fibrotic potency compared to methylamino substituted derivative (2a). As a consequence, the steric influence in alkylamino substituted derivatives limited the scope for further improvement.
Next, we sought to replace the alkylamino substituents of the 13 position to indolyl group, and 14 derivatives were obtained (3a-n). To our delight, most of the compounds displayed more significant improvement in anti-fibrotic potency than matrine. Among them, compounds 3d, 3f, and 3g gave inspiring IC 50 values of 4.3 µM, 3.3 µM, and 6.5 µM, respectively, which was 204-fold, 266-fold, and 135-fold more potent than matrine; or 307-fold, 400-fold, and 203-fold more potent than Pirfenidone. Compounds 3d and 3f also showed marked improvement in SI over matrine (from 2.8 to 7.6-8.0).
Further analysis of 3a-n suggested a remarkable impact of electronic properties of indolyl substitution on anti-fibrotic potency. Compounds with strong electron withdrawing groups such as halogen (3f-g, 3k-l) or nitro group (3d, 3i, 3m) gave more potent IC 50 values (3.3-27.0 µM), an exception being compound 3h with cyano group whose IC 50 was 72.1 µM. Contrarily, the introduction of electron donating groups resulted in significant loss of anti-fibrotic potency, and the IC 50 values of methyl (3b) or methoxy (3c, 3e) substituted compounds were around 39.0 to 68.1 µM.
Additionally, we found an interesting phenomenon that the anti-fibrotic potency was positively correlated with lipophilicity of the corresponding compounds, and a bubble chart based on calculated log partition coefficients (ClogP) and IC 50 values was established. As depicted in Figure 3, compounds exhibiting better anti-fibrotic activities tended to possess higher ClogP values. The introduction of indoles significantly increased the ClogP values of matrine (from 1.36 to 1.98-4.18), which might permit better membrane penetration and consequently more potent activities. Next, we sought to replace the alkylamino substituents of the 13 position to indolyl group, and 14 derivatives were obtained (3a-n). To our delight, most of the compounds displayed more significant improvement in anti-fibrotic potency than matrine. Among them, compounds 3d, 3f, and 3g gave inspiring IC50 values of 4.3 μM, 3.3 μM, and 6.5 μM, respectively, which was 204-fold, 266fold, and 135-fold more potent than matrine; or 307-fold, 400-fold, and 203-fold more potent than Pirfenidone. Compounds 3d and 3f also showed marked improvement in SI over matrine (from 2.8 to 7.6-8.0).
Further analysis of 3a-n suggested a remarkable impact of electronic properties of indolyl substitution on anti-fibrotic potency. Compounds with strong electron withdrawing groups such as halogen (3f-g, 3k-l) or nitro group (3d, 3i, 3m) gave more potent IC50 values (3.3-27.0 μM), an exception being compound 3h with cyano group whose IC50 was 72.1 μM. Contrarily, the introduction of electron donating groups resulted in significant loss of anti-fibrotic potency, and the IC50 values of methyl (3b) or methoxy (3c, 3e) substituted compounds were around 39.0 to 68.1 μM.
Additionally, we found an interesting phenomenon that the anti-fibrotic potency was positively correlated with lipophilicity of the corresponding compounds, and a bubble chart based on calculated log partition coefficients (ClogP) and IC50 values was established. As depicted in Figure 3, compounds exhibiting better anti-fibrotic activities tended to possess higher ClogP values. The introduction of indoles significantly increased the ClogP values of matrine (from 1.36 to 1.98-4.18), which might permit better membrane penetration and consequently more potent activities. Bubble size reflects SI value. Yellow, green, and grey bubbles represent derivatives 2a-d, 3a-n, and positive control, respectively. Mat, Sop, PF was abbreviation for Matrine, Sophocarpine, and Pirfenidone, respectively.

Inhibition Effects of Expression Levels of ECM Proteins by Key Compound 3f
Due to the significance of collagen in IPF pathogenesis, we first evaluated the inhibitory effect of key compound 3f against collagen accumulation. Immunofluorescence staining was employed to investigate collagen protein expression. As indicated in Figure 4A,B and 4D,E, collagen synthesis was significantly elevated upon TGF-β1 stimulation, and the expression of Collagen type I alpha-1 chain (COL1A1) and Collagen type III alpha-1 chain (COL3A1) was upregulated by 178.2% and 183.1%, respectively. 10 μM 3f treatment resulted in remarkable decrease in expression level of COL1A1 and

Inhibition Effects of Expression Levels of ECM Proteins by Key Compound 3f
Due to the significance of collagen in IPF pathogenesis, we first evaluated the inhibitory effect of key compound 3f against collagen accumulation. Immunofluorescence staining was employed to investigate collagen protein expression. As indicated in Figure 4A,B,D,E, collagen synthesis was significantly elevated upon TGF-β1 stimulation, and the expression of Collagen type I alpha-1 chain (COL1A1) and Collagen type III alpha-1 chain (COL3A1) was upregulated by 178.2% and 183.1%, respectively. 10 µM 3f treatment resulted in remarkable decrease in expression level of COL1A1 and COL3A1 relative to TGF-β1 stimulation group (46.2% and 31.1%, respectively), indicating a potent inhibitory effect of 3f against TGF-β1-induced collagen accumulation.
Molecules 2019, 24, x FOR PEER REVIEW 6 of 19 COL3A1 relative to TGF-β1 stimulation group (46.2% and 31.1%, respectively), indicating a potent inhibitory effect of 3f against TGF-β1-induced collagen accumulation. Next, we evaluated the inhibitory effect of 3f against fibronectin expression. Fibronectin is another major constituent of ECM and plays a critical role in fibrogenesis by promoting the migration [22], proliferation [23], and adhesion [24] of fibroblasts. We observed that 3f treatment reduced the expression level of fibronectin by 40.0% comparing to TGF-β1 stimulation group, which resulted in even lower expression level than vehicle group ( Figure 4C,F). These results suggested that 3f could revert the increase of ECM production induced by TGF-β1 stimulation.

Inhibition Effects against Fibroblast-to-Myfibroblast Transition by 3f
Myofibroblasts, a further differentiated subset of fibroblasts, are primary effector cells for collagen synthesis in fibrotic disorders. Inhibition of fibroblast differentiation is an important therapeutic target for IPF treatment. To examine whether 3f could inhibit TGF-β1-induced fibroblastto-myofibroblast transition, we investigated the expression level of alpha-smooth muscle actin (α-SMA) and S100A4, the key markers which distinguish myofibroblasts from their precursor fibroblasts [25]. We observed dramatic upregulation of myofibroblast markers in response to TGF-β1 stimulation (774.9% and 537.7% for α-SMA and S100A4, respectively), characterizing a strong expression of myofibroblast phenotype. 3f treatment resulted in disruption of fibroblast-to-myofibroblast transition, as evidenced by 45.8% and 17.7% downregulation in expression levels of α-SMA and S100A4, respectively, compared to TGF-β1 stimulation group ( Figure 5). Next, we evaluated the inhibitory effect of 3f against fibronectin expression. Fibronectin is another major constituent of ECM and plays a critical role in fibrogenesis by promoting the migration [22], proliferation [23], and adhesion [24] of fibroblasts. We observed that 3f treatment reduced the expression level of fibronectin by 40.0% comparing to TGF-β1 stimulation group, which resulted in even lower expression level than vehicle group ( Figure 4C,F). These results suggested that 3f could revert the increase of ECM production induced by TGF-β1 stimulation.

Inhibition Effects against Fibroblast-to-Myfibroblast Transition by 3f
Myofibroblasts, a further differentiated subset of fibroblasts, are primary effector cells for collagen synthesis in fibrotic disorders. Inhibition of fibroblast differentiation is an important therapeutic target for IPF treatment. To examine whether 3f could inhibit TGF-β1-induced fibroblast-to-myofibroblast transition, we investigated the expression level of alpha-smooth muscle actin (α-SMA) and S100A4, the key markers which distinguish myofibroblasts from their precursor fibroblasts [25]. We observed dramatic upregulation of myofibroblast markers in response to TGF-β1 stimulation (774.9% and 537.7% for α-SMA and S100A4, respectively), characterizing a strong expression of myofibroblast phenotype. 3f treatment resulted in disruption of fibroblast-to-myofibroblast transition, as evidenced by 45.8% and 17.7% downregulation in expression levels of α-SMA and S100A4, respectively, compared to TGF-β1 stimulation group ( Figure 5).

Inhibition Effects against Smad-Mediated Signaling by 3f
Small mothers against decapentaplegic homologs (Smad) are key mediators and regulators of TGFβ signaling. The binding of TGF-β1 to TGFβ receptors leads to the phosphorylation of Smad2/3 complex, which then translocate into nucleus to enhance gene transcription by cooperating with DNA transcription factors [26]. Inhibition of the nuclear translocation of Smad2/3 could attenuate the transcription of profibrotic genes such as collagens and α-SMA. Therefore, we investigated the cellular localization of Smad2/3 by immunofluorescence assay. As depicted in Figure 6A, Smad2/3 predominately localize in cytoplasm under normal conditions, but translocate into nucleus upon TGF-β1 stimulation. We observed that 3f could significantly inhibit the cytoplasm-to-nuclear translocation of Smad2/3.
Another emerging effect of Smad which acts as the amplifier of TGF-β signaling also contribute to fibrotic diseases. The nuclear translocation of Smad3 leads to the activation of the miR-433 promotor region. miR-433 promotes the decrease of antizyme inhibitor I (Azin1) and eventually results in depletion of polyamine, which causes the upregulation of expression level of TGF-β receptor type I (TGFβRI) and TGF-β1. This series of signaling creates a positive-feedback loop which amplify the TGFβ signaling [27]. TGF-β1-induced upregulation of expression level of TGFβRI was also observed in this study ( Figure 6B), which was reduced by 46.7% under 3f treatment ( Figure 6C), suggesting an inhibitory effect of 3f against the positive-feedback loop of TGFβ signaling.

Inhibition Effects against Smad-Mediated Signaling by 3f
Small mothers against decapentaplegic homologs (Smad) are key mediators and regulators of TGFβ signaling. The binding of TGF-β1 to TGFβ receptors leads to the phosphorylation of Smad2/3 complex, which then translocate into nucleus to enhance gene transcription by cooperating with DNA transcription factors [26]. Inhibition of the nuclear translocation of Smad2/3 could attenuate the transcription of profibrotic genes such as collagens and α-SMA. Therefore, we investigated the cellular localization of Smad2/3 by immunofluorescence assay. As depicted in Figure 6A, Smad2/3 predominately localize in cytoplasm under normal conditions, but translocate into nucleus upon TGF-β1 stimulation. We observed that 3f could significantly inhibit the cytoplasm-to-nuclear translocation of Smad2/3.

Inhibition Effects against TGF-β1 Induced Migration of MRC-5 Cells by 3f
Fibroblast migration is the key event in pulmonary fibrogenesis. Lung fibroblasts migrate across basement membranes and deposit excessive collagen which resulted in loss of function and eventually destruction of alveoli [28]. Inhibition of lung fibroblasts migration is a promising potential therapeutic target for IPF. We performed wound healing assay to investigate the inhibitory effect of 3f and matrine against fibroblast migration (Figure 7). The results indicated that MRC-5 fibroblasts displayed limited migration capacity under normal conditions, with 65.2% closure of wound area over 24 h period. However, wound closure increased to 85.0% over 24 h period upon TGF-β1 stimulation, indicating a significant enhancement in migration capacity, whilst treatment of 3f resulted in potent inhibition of TGF-β1-induced fibroblast migration, with only 37.7% closure of wound area over 24 h period. It is also worth mentioning that the inhibitory effect of 10 μM 3f was more potent than that of 1000 μM matrine (56.6% closure of wound area over 24 h period). Another emerging effect of Smad which acts as the amplifier of TGF-β signaling also contribute to fibrotic diseases. The nuclear translocation of Smad3 leads to the activation of the miR-433 promotor region. miR-433 promotes the decrease of antizyme inhibitor I (Azin1) and eventually results in depletion of polyamine, which causes the upregulation of expression level of TGF-β receptor type I (TGFβRI) and TGF-β1. This series of signaling creates a positive-feedback loop which amplify the TGFβ signaling [27]. TGF-β1-induced upregulation of expression level of TGFβRI was also observed in this study ( Figure 6B), which was reduced by 46.7% under 3f treatment ( Figure 6C), suggesting an inhibitory effect of 3f against the positive-feedback loop of TGFβ signaling.

Inhibition Effects against TGF-β1 Induced Migration of MRC-5 Cells by 3f
Fibroblast migration is the key event in pulmonary fibrogenesis. Lung fibroblasts migrate across basement membranes and deposit excessive collagen which resulted in loss of function and eventually destruction of alveoli [28]. Inhibition of lung fibroblasts migration is a promising potential therapeutic target for IPF. We performed wound healing assay to investigate the inhibitory effect of 3f and matrine against fibroblast migration (Figure 7). The results indicated that MRC-5 fibroblasts displayed limited migration capacity under normal conditions, with 65.2% closure of wound area over 24 h period. However, wound closure increased to 85.0% over 24 h period upon TGF-β1 stimulation, indicating a significant enhancement in migration capacity, whilst treatment of 3f resulted in potent inhibition of TGF-β1-induced fibroblast migration, with only 37.7% closure of wound area over 24 h period. It is Molecules 2019, 24, 1108 9 of 19 also worth mentioning that the inhibitory effect of 10 µM 3f was more potent than that of 1000 µM matrine (56.6% closure of wound area over 24 h period).

Action on TGFβ/Smad Pathway of 3f
As illustrated earlier, 3f could significantly inhibit fibroblast-to-myofibroblast transition induced by TGF-β1, which led to the downregulation of ECM production. 3f could also inhibit Smad signaling by inhibiting cytoplasm-to-nuclear translocation of Smad2/3. The inhibition of positivefeedback loop of TGFβ signaling was evidenced by suppression of TGF-β1-induced upregulation of TGFβRI by 3f as well. Therefore, it was speculated that 3f might exert anti-pulmonary fibrosis activity via repression of the TGF-β/Smad signaling pathway, as depicted in Figure 8.

Action on TGFβ/Smad Pathway of 3f
As illustrated earlier, 3f could significantly inhibit fibroblast-to-myofibroblast transition induced by TGF-β1, which led to the downregulation of ECM production. 3f could also inhibit Smad signaling by inhibiting cytoplasm-to-nuclear translocation of Smad2/3. The inhibition of positive-feedback loop of TGFβ signaling was evidenced by suppression of TGF-β1-induced upregulation of TGFβRI by 3f as well. Therefore, it was speculated that 3f might exert anti-pulmonary fibrosis activity via repression of the TGF-β/Smad signaling pathway, as depicted in Figure 8.

General Procedure for the Synthesis of Compounds 2a-d
To a 10 mL microwave vial was added sophocarpine (0.3 mmol) and aliphatic amine (1.5 mmol), then water (1 mL) was added as solvent. The solution was irradiated under the microwave power of 100 W for 30 min. After irradiation, the reaction solution was poured into saturated aqueous ammonium chloride (20 mL), and then extracted with dichloromethane (20 mL) twice. The organic layer was washed by saturated sodium chloride, dried over anhydrous sodium sulfate, and concentrated under reduced pressure. The residue was purified by flash chromatography on silica gel with dichloromethane/methanol as the eluent. Compounds 2a-d were mostly presented in pairs

General Procedure for the Synthesis of Compounds 2a-d
To a 10 mL microwave vial was added sophocarpine (0.3 mmol) and aliphatic amine (1.5 mmol), then water (1 mL) was added as solvent. The solution was irradiated under the microwave power of 100 W for 30 min. After irradiation, the reaction solution was poured into saturated aqueous ammonium chloride (20 mL), and then extracted with dichloromethane (20 mL) twice. The organic layer was washed by saturated sodium chloride, dried over anhydrous sodium sulfate, and concentrated under reduced pressure. The residue was purified by flash chromatography on silica gel with dichloromethane/methanol as the eluent. Compounds 2a-d were mostly presented in pairs of 13S/13R stereoisomers, of which 13S-isomers were the major stereoisomers. The NMR peaks recorded below were signals of the major stereoisomers.

General Procedure for the Synthesis of Compounds 3a-n
To a solution of the corresponding substituted indole (0.36 mmol) in THF (2 mL) was added 2 mol/L KHDMS (0.36 mmol) at −20 °C; the resulting solution was stirred at the same temperature for 20 min, and then sophocarpine (0.3 mmol) was added. The reaction temperature was allowed to raise to r.t. and stirred overnight. After the completion of the reaction, the reaction mixture was diluted with dichloromethane (50 mL). The organic layer was washed with water and brine, dried over anhydrous sodium sulfate, and concentrated under reduced pressure. The residue was purified by To a solution of the corresponding substituted indole (0.36 mmol) in THF (2 mL) was added 2 mol/L KHDMS (0.36 mmol) at −20 • C; the resulting solution was stirred at the same temperature for 20 min, and then sophocarpine (0.3 mmol) was added. The reaction temperature was allowed to raise to r.t. and stirred overnight. After the completion of the reaction, the reaction mixture was diluted with dichloromethane (50 mL). The organic layer was washed with water and brine, dried over anhydrous sodium sulfate, and concentrated under reduced pressure. The residue was purified by flash chromatography on silica gel with dichloromethane/methanol as the eluent. Compounds 3a-n were mostly presented in pairs of 13S/13R stereoisomers, of which 13S-isomers were the major stereoisomers. The NMR peaks recorded below were signals of the major stereoisomers.