Fenofibrate inhibits TGF‐β‐induced myofibroblast differentiation and activation in human lung fibroblasts in vitro

Fenofibrate (FF), a peroxisome proliferator‐activated receptor‐alpha (PPAR‐α) agonist and a lipid‐lowering agent, can decrease experimental pulmonary fibrosis. However, the mechanisms underlying the antifibrotic effect of FF remain unknown. Hence, this study was conducted to evaluate the effects of FF on transforming growth factor‐beta (TGF‐β)‐induced myofibroblast differentiation and activation in lung fibroblasts. The results showed that FF inhibited alpha‐smooth muscle actin (α‐SMA) and connective tissue growth factor expression, collagen production, cell motility, SMAD3 phosphorylation and nuclear translocation, and metabolic reprogramming in TGF‐β‐exposed cells. The inhibitory effect of FF did not decrease with the addition of a PPAR‐α antagonist. Moreover, the inhibitory effect given by FF could not be reproduced with the addition of an alternative PPAR‐α agonist. FF inhibited mitochondrial respiration. However, rotenone, a complex I inhibitor, did not suppress TGF‐β‐induced myofibroblast differentiation. Furthermore, the TGF‐β‐induced nuclear reduction of protein phosphatase, Mg2+/Mn2+‐dependent 1A (PPM1A), a SMAD phosphatase, was inhibited by FF. These results showed that FF suppressed TGF‐β‐induced myofibroblast differentiation and activation independent of PPAR‐α activation and impaired mitochondrial respiration. In conclusion, this study provides information on the effects of FF on anti‐TGF‐β mechanisms.

Fenofibrate (FF), a peroxisome proliferator-activated receptor-alpha (PPAR-a) agonist and a lipid-lowering agent, can decrease experimental pulmonary fibrosis. However, the mechanisms underlying the antifibrotic effect of FF remain unknown. Hence, this study was conducted to evaluate the effects of FF on transforming growth factor-beta (TGF-b)-induced myofibroblast differentiation and activation in lung fibroblasts. The results showed that FF inhibited alpha-smooth muscle actin (a-SMA) and connective tissue growth factor expression, collagen production, cell motility, SMAD3 phosphorylation and nuclear translocation, and metabolic reprogramming in TGF-b-exposed cells. The inhibitory effect of FF did not decrease with the addition of a PPAR-a antagonist. Moreover, the inhibitory effect given by FF could not be reproduced with the addition of an alternative PPAR-a agonist. FF inhibited mitochondrial respiration. However, rotenone, a complex I inhibitor, did not suppress TGF-b-induced myofibroblast differentiation. Furthermore, the TGF-b-induced nuclear reduction of protein phosphatase, Mg 2+ /Mn 2+ -dependent 1A (PPM1A), a SMAD phosphatase, was inhibited by FF. These results showed that FF suppressed TGF-b-induced myofibroblast differentiation and activation independent of PPAR-a activation and impaired mitochondrial respiration. In conclusion, this study provides information on the effects of FF on anti-TGF-b mechanisms.
Myofibroblast differentiation and activation are the key pathogenic events involved in many pulmonary diseases, such as asthma, chronic obstructive pulmonary disease, and idiopathic pulmonary fibrosis [1][2][3][4]. In these diseases, airway or parenchymal fibrosis is mechanistically mediated by common and/or disease-specific pathways, among which transforming growth factorbeta (TGF-b) signaling plays a central role. TGF-b induces the phosphorylation of SMAD proteins, stimulating signal transduction pathways that lead to myofibroblast differentiation and transcription factor activation and alpha-smooth muscle actin (a-SMA) and collagen production [5][6][7].
Fenofibrate (FF) is a widely used antihyperlipidemic agent exhibiting a lipid-lowering effect by activating peroxisome proliferator-activated receptor-alpha (PPARa) [8]. In addition, FF exerts pleiotropic actions on multiple pathways to reduce inflammation, oxidative stress, apoptosis, angiogenesis, and fibrosis [9,10]. Notably, recent clinical studies on patients with diabetes have revealed that FF has protective effects on retinopathy and nephropathy [11,12]. Moreover, several animal studies have shown that FF inhibits experimental fibrosis in different organs, including the retina, heart, liver, kidney, and lung [13][14][15][16][17][18][19][20]. For example, the oral administration of FF in rats has been shown to attenuate the severity of bleomycin-induced pulmonary fibrosis [20]. Since the overexpression of TGF-b is critical in fibrotic disease, we hypothesized that FF inhibits the TGF-b signaling pathway. The molecular mechanism of action of FF appears to be complex, which includes not only PPAR-a-dependent but also PPAR-a-independent mechanisms [21][22][23][24][25][26][27][28][29][30][31]. Thus, this study was conducted to assess the effects of FF on TGF-b-induced myofibroblast differentiation and activation in vitro and to determine whether the effects of FF depend on PPAR-a.

Immunofluorescence staining
Cells in the NuncÒ Lab-TekÒ eight-well chamber slides (Thermo Fisher Scientific K.K., Yokohama, Japan) were fixed with 10% formalin and permeabilized with 0.3% Tri-tonÒ X-100 in PBS for 5 min. After blocking the nonspecific binding sites with 3% BSA, the slides were incubated with mouse monoclonal anti-a-SMA antibody (Novus Biologicals USA, Centennial, CO, USA) or rabbit monoclonal anti-SMAD3 (Abcam Japan, Tokyo, Japan). Next, the primary antibody was allowed to react with a secondary anti-mouse immunoglobulin G antibody conjugated with Alexa Fluor 488 (Invitrogen, Carlsbad, CA, USA). Then, the cell nuclei were counterstained with 4 0 ,6-diamidino-2-phenylindole. Fluorescence images were obtained using a microscope (Olympus IX71; Olympus Optical Co., Ltd., Tokyo, Japan) equipped with a digital camera.

Collagen assay
The collagen content in the culture medium was assessed using the SircolÒ Soluble Collagen Assay Kit (Biocolor, Carrickfergus, UK).

In vitro wound closure assay
The cells were cultured to reach confluence in 24-well plates. After serum starvation for 24 h, they were scratched in a straight line using a sterile 200-lL tip and were washed twice with PBS. Thereafter, the cells were incubated in serum-free DMEM with or without TGF-b (5 ngÁmL À1 ) in the presence or absence of FF (25 lM). After 24 h, the cells were fixed with 10% formalin and were washed twice with PBS. Wound healing was assessed under a phase-contrast microscope (Olympus IX71; Olympus Optical Co., Ltd.) equipped with a digital camera. The wound area was measured using image analysis software (Lumina Vision; Mitani Corporation, Fukui, Japan) on a Microsoft XP computer. Five measurements were taken from five fields of each well obtained from six wells in each experiment.

Metabolic flux analysis
Cells were cultured in eight-well XF assay plates and then treated with or without TGF-b (5 ngÁmL À1 ) in the presence or absence of FF (25 lM) or WY14643 (50 lM) for 48 h. The oxygen concentration rate (OCR) and extracellular acidification rate (ECAR) were measured using the Seahorse XFp Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA).

Assessment of glucose consumption
Glucose concentrations in the culture and original medium (not cultured with cells) were assessed using an amperometric blood glucose sensor (Nipro Care Fast C; Nipro Co., Osaka, Japan).

Statistical analysis
Statistical analyses were performed using Microsoft Excel X with Statcel 3 (OMS; Tokyo, Japan) as an add-in software. Data were expressed as means AE standard error of the mean. Equality of variance was examined using Bartlett's test. Data with equality of variance rejected were logarithmically transformed and retested. When equality of variance was recognized, statistical significance was analyzed using ANOVA. If significant results were obtained, the Tukey-Kramer test was used as a post hoc test for multiple comparisons. P-values < 0.05 were used to denote statistically significant differences.

Results
First, we examined whether FF modulates the effects of TGF-b on IMR-90 cells, a human lung fibroblast cell line. The treatment of IMR-90 cells with TGF-b (5 ngÁmL À1 ) for 48 h enhanced the expression of a-SMA, indicating myofibroblast differentiation (Fig. 1A,B). However, cotreatment with FF (1-25 lM) reduced the enhanced expression of a-SMA in TGF-btreated cells in a dose-dependent manner. This finding showed that FF inhibits myofibroblast differentiation induced by TGF-b. Moreover, treatment of IMR-90 cells with TGF-b increased the expression of CTGF (Fig. 1C), an important mediator of myofibroblast activation and extracellular matrix synthesis [35], and the production of collagen (Fig. 1D). Similar to the inhibition of a-SMA expression by FF, cotreatment with FF also reduced the aforementioned enhanced CTGF expression and collagen production induced by exposure to TGF-b (Fig. 1C,D). Since the activation of myofibroblasts is associated with increased cell motility [36], the effects of FF on cell migration were evaluated. The in vitro wound closure assay showed that the mobility of TGF-b-treated IMR-90 cells was significantly decreased in the presence of FF (Fig. 1E). SMADs, including SMAD3, were found to play a critical role in TGF-b signaling pathways leading to myofibroblast differentiation and extracellular matrix protein production [5]. Western blotting and immunofluorescence revealed that pretreatment with FF inhibited TGF-b-induced phosphorylation and nuclear translocation of the SMAD3 (Fig. 1F,G).
Myofibroblast differentiation is associated with metabolic reprogramming, which was proposed as a molecular target of TGF-b action [37,38]. According to previous reports [37,38], treatment with TGF-b increased both OCR, as evidenced by augmented basal respiration (Fig. 1H) and mitochondrial ATP production (Fig. 1I), and glycolysis, as indicated by increased ECAR (Fig. 1J) and glucose consumption (Fig. 1K). Cotreatment with FF decreased TGF-b-induced mitochondrial respiration (Fig. 1H,I). However, interestingly, it enhanced TGF-b-mediated stimulation of glycolysis (Fig. 1J,K). Taken together, these results showed that FF suppressed SMAD activation, myofibroblast differentiation, collagen production, cell mobility, and mitochondrial respiration in human lung fibroblasts treated with TGF-b.
The lipid-lowering effect of FF is mediated by the activation of PPAR-a, a nuclear receptor to which FF binds as a ligand [39]. Thus, we next examined whether the inhibition of TGF-b-induced myofibroblast differentiation by FF was attributed to the activation of PPARa. The results showed that treatment with GW6471 (20 lM), a PPAR-a antagonist, did not change the inhibitory effect of FF on TGF-b-induced a-SMA expression ( Fig. 2A). Furthermore, treatment with WY14643 (50 lM), a PPAR-a agonist, did not inhibit TGF-binduced a-SMA expression (Fig. 2B,C). These findings showed that the inhibition of TGF-b-induced myofibroblast differentiation by FF is mediated by a mechanism independent of PPAR-a activation.
FF inhibits complex 1 of the mitochondrial electron transfer chain [40]. In this study, in contrast to FF, WY14643 did not suppress the high mitochondrial  (Fig. 3A,B). Hence, it was unlikely that the inhibition of mitochondrial respiration by FF was caused by PPAR-a activation. Then, we addressed the possibility that the PPAR-aindependent inhibition of mitochondrial complex I by FF is mechanistically involved in inhibiting TGF-binduced myofibroblast differentiation. However, treatment with rotenone, a complex I inhibitor, did not inhibit either a-SMA expression (Fig. 3C) or SMAD3 phosphorylation (Fig. 3D) in TGF-b-treated cells. This result indicated that the inhibition of TGF-b-induced myofibroblast differentiation by FF is not attributed to the inhibition of mitochondrial respiration.
A time-course analysis of FF inhibition of TGF-binduced phosphorylation of SMAD3 revealed that FF reduced the cellular levels of phosphorylated SMAD3 after 1 h or thereafter but not as early as 30 min after treatment with TGF-b. This result showed that FF might have reduced TGF-b-induced SMAD3 phosphorylation at a later time by dephosphorylating SMAD3  4). Data were expressed as means AE standard error of the mean. *P < 0.05 and **P < 0.01 using the Tukey-Kramer test. Co, control; p-SMAD3, phosphorylated SMAD3. IMR-90 cells were pretreated with WY14643 (50 lM) or vehicle alone for 1 h and were then treated with or without TGF-b (5 ngÁmL À1 ) for 48 h in the presence or absence of WY14643. The expression of a-SMA protein was evaluated using western blot analysis (B, n = 4) or immunofluorescence staining (C, n = 4). green, a-SMA; blue, nuclear staining. Insets are magnified images. Scale bar: 100 lm. Data were expressed as means AE standard error of the mean. *P < 0.05 and **P < 0.01 using the Tukey-Kramer test. n.s., not significant. Co, control; GW, GW6471; WY, WY14643. (Fig. 4A). Studies have shown that PPM1A dephosphorylates and promotes the nuclear export of SMAD3 to terminate TGF-b signaling [41]. Lastly, we found that treatment with TGF-b decreased the nuclear levels but not the total levels of PPM1A, which was partially recovered after cotreatment with FF (Fig. 4B). These findings suggest that FF partially inhibited TGF-binduced reduction of nuclear PPM1A.

Discussion
This study showed that treatment with FF inhibits TGF-b-induced myofibroblast differentiation and activation, as evidenced by the suppression of a-SMA and CTGF expression, collagen production, cell mobility, SMAD3 phosphorylation and nuclear translocation, and mitochondrial respiration in human lung fibroblasts treated with TGF-b. The lipid-lowering effect of FF is reportedly mediated by activating PPAR-a [39]. However, this study revealed that the inhibition of TGF-b-induced myofibroblast differentiation by FF is not caused by PPAR-a activation because the PPAR-a antagonist GW6471 did not suppress the inhibitory effect of FF on TGF-b-induced myofibroblast differentiation and treatment with the PPAR-a agonist WY14643 did not inhibit TGF-b-induced myofibroblast differentiation.
This study showed that FF inhibits mitochondrial respiration via a PPAR-a-independent mechanism. However, the inhibition of TGF-b-induced myofibroblast differentiation by FF was not attributed to the inhibition of mitochondrial respiration because rotenone did not suppress TGF-b-induced myofibroblast differentiation. This study also showed that FF partially inhibits TGF-b-induced reduction of nuclear Effects of rotenone on the expression of a-SMA (C) and phosphorylation of SMAD3 (D). IMR-90 cells were pretreated with rotenone (0.5 lM) or vehicle alone for 1 h and were then treated with or without TGF-b (5 ngÁmL À1 ) in the presence or absence of rotenone for 1 h (D) or 48 h (C). The a-SMA expression (C) and SMAD3 phosphorylation (D) levels were evaluated using western blot analysis (n = 4). Data were expressed as means AE standard error of the mean. *P < 0.05 and **P < 0.01 using the Tukey-Kramer test. n.s., not significant. Co, control; WY, WY14643; Rot, rotenone; p-SMAD3, phosphorylated SMAD3.
PPM1A, a SMAD phosphatase that terminates TGF-b signaling [41]. Thus, the FF-mediated regulation of PPM1A may mechanistically be accounted at least in part for the FF-mediated inhibition of TGF-b-induced myofibroblast differentiation. Interestingly, Paw et al. [32] have recently shown that FF inhibits TGF-binduced myofibroblast differentiation by disturbing the organization of actin cytoskeleton architecture. They have found that FF inhibits TGF-b-stimulated incorporation of a-SMA into stress fibers by reducing connexin 43, a protein that constitutes gap junctional channels. However, in contrast to this study, they have found that FF does not suppress TGF-b-induced a-SMA expression despite the inhibition of SMAD2/3 signaling. The discrepancy between their study and this study may be attributed to different sources of fibroblasts: We used fibroblasts derived from normal fetal lungs, whereas Paw et al. used bronchial fibroblasts derived from patients with asthma. Although the results of this study indicated PPAR-a-independent mechanisms by which FF suppresses TGF-b-induced myofibroblast differentiation, we still considered the possibility that PPAR-a activation has some antifibrotic effects because PPAR-a-knockout mice treated with bleomycin had more severe inflammation and injury than wild-type mice [43].  4. Effects of FF on SMAD3 phosphorylation (A) and the total and nuclear levels of PPM1A (B). IMR-90 cells were pretreated with FF (25 lM) or vehicle alone for 1 h and were then treated with or without TGF-b (5 ngÁmL À1 ) in the presence or absence of FF for 30 min, 1, 2, and 4 h (A, n = 4) or 1 h (B, n = 4). Total cell lysates (A; B, right) or nuclear lysates (B, left) were analyzed using western blot analysis. Data were expressed as means AE standard error of the mean. *P < 0.05 and **P < 0.01 using the Tukey-Kramer test. n.s., not significant. Co, control; p-SMAD3, phosphorylated SMAD3.
Recent animal studies have shown that treatment with FF decreases experimental fibrosis in the retina, lungs, liver, heart, and kidneys [13][14][15][16][17][18][19][20]. FF is a relatively safe and inexpensive agent widely used in clinical practice for treating hyperlipidemia. This study provided a better understanding of the antifibrotic effect of FF, which can be used as a basis for its clinical application for treating fibrotic disorders.