Chrysin Ameliorates Cyclosporine-A-Induced Renal Fibrosis by Inhibiting TGF-β1-Induced Epithelial–Mesenchymal Transition

Cyclosporine A (CsA) is a nephrotoxicant that causes fibrosis via induction of epithelial–mesenchymal transition (EMT). The flavonoid chrysin has been reported to have anti-fibrotic activity and inhibit signaling pathways that are activated during EMT. This study investigated the nephroprotective role of chrysin in the prevention of CsA-induced renal fibrosis and elucidated a mechanism of inhibition against CsA-induced EMT in proximal tubule cells. Treatment with chrysin prevented CsA-induced renal dysfunction in Sprague Dawley rats measured by blood urea nitrogen (BUN), serum creatinine and creatinine clearance. Chrysin inhibited CsA-induced tubulointerstitial fibrosis, characterized by reduced tubular damage and collagen deposition. In vitro, chrysin significantly inhibited EMT in LLC-PK1 cells, evidenced by inhibition of cell migration, decreased collagen expression, reduced presence of mesenchymal markers and elevated epithelial junction proteins. Furthermore, chrysin co-treatment diminished CsA-induced TGF-β1 signaling pathways, decreasing Smad 3 phosphorylation which lead to a subsequent reduction in Snail expression. Chrysin also inhibited activation of the Akt/ GSK-3β pathway. Inhibition of both pathways diminished the cytosolic accumulation of β-catenin, a known trigger for EMT. In conclusion, flavonoids such as chrysin offer protection against CsA-induced renal dysfunction and interstitial fibrosis. Chrysin was shown to inhibit CsA-induced TGF-β1-dependent EMT in proximal tubule cells by modulation of Smad-dependent and independent signaling pathways.


Introduction
Renal fibrosis is a common outcome of several forms of kidney injury. Irrespective of the etiology, chronic kidney disease (CKD) causes progressive decline in kidney function associated with loss of renal parenchyma and formation of scar tissue. Accumulation of scar tissue in the kidney leads to end-stage renal disease, requiring frequent dialysis or a transplant, both of which have been associated with complications such as infections or graft rejection. In spite of the treatment options available, patients suffering from CKD have a very high rate of mortality, making therapeutic intervention to inhibit renal fibrosis an urgent necessity [1].
Epithelial-mesenchymal transition (EMT) has been suggested as an important step in tissue repair following injury, through which the surviving epithelial cells transition into cells with a mesenchymal phenotype [2]. These cells migrate to the denuded basement membrane, proliferate to repair the injury [3,4] and finally re-differentiate to epithelial cells [5]. During EMT, renal proximal tubule cells lose epithelial characteristics, such as a cuboidal appearance and apical to basolateral polarity, and acquire mesenchymal traits including spindle shape, leading or trailing edge asymmetry and secretion of extracellular matrix proteins [6].
A key event in renal fibrosis is the initiation of EMT by TGF-β1, a cytokine produced by injured parenchymal cells and macrophages [5,7]. TGF-β1 elicits its effects on target genes, focusing primarily on Smad proteins and their role in the transcriptional regulation of extracellular matrix (ECM) genes (reviewed in [8]). Ligand activation of type 1 TGF-β1 receptor recognizes and phosphorylates the ligand-specific receptor-activated Smad, resulting in the formation of a heteromeric complex with co-Smad 4. Nuclear translocation of the complex induces transcription of various pro-ECM genes, including Col1a1, Col3a1, and Timp1, that eventually result in the development and progression of fibrosis. In addition, TGF-β1 can signal through Smad-independent signaling cascades (reviewed in [9,10]). The phosphoinositol 3-kinase (PI3K)-Akt-GSK3-β signaling axis is critical in the regulation of TGF-β1-induced EMT [11]. TGF-β1 triggers the PI3K/Akt/GSK-3β signaling cascade, resulting in the activation of Akt and subsequent inactivation of GSK-3β, which has been implicated as an early sequence in the induction of fibrosis [12][13][14]. Studies have also shown that phosphorylation of GSK-3β is a crucial step in triggering EMT [14,15]. Through these signaling cascades TGF-β1 regulates transcription and stabilization of DNA binding protein Snail [16][17][18]. The zinc finger transcription factor, Snail, functions as a potent repressor of E-cadherin expression that can, acting alone or in concert with the β-catenin, induce EMT [19][20][21].
Chrysin is a flavonoid ( Figure S1) that can be extracted from natural sources such as passionflower (Passiflora caerulea, Passiflora incarnata) and honeycombs. This naturally occurring plant flavonoid has many pharmacological activities, which include anti-inflammatory [22], anti-apoptotic [23], antioxidant [24] and anti-cancer properties [25]. Chrysin has also been documented to counter drug-induced hepatotoxicity and nephrotoxicity (reviewed in Pingili et al., 2019 [26]). Chrysin has been shown to alleviate tubulointerstitial fibrosis [27] and glomerulonephritis [28] associated with diabetic kidney disease. The chronic use of the immunosuppressant cyclosporine A (CsA) results in impaired renal function as a result of fibrosis, which can have devastating effects on transplant patients. In the present investigation we evaluated the potential efficacy of chrysin in preventing CsAinduced renal fibrosis and elucidated a mechanism by which chrysin inhibits CsA-induced TGF-β1-mediated EMT in vitro.

Effect of Chrysin on CsA-Induced Renal Dysfunction
To determine the effect of CsA and chrysin on kidney function, serum and urine creatinine as well as blood urea nitrogen (BUN) levels were measured in samples collected from male Sprague Dawley rats. A significant elevation in serum creatinine was observed with CsA alone ( Figure 1A); this was ameliorated by concomitant treatment with chrysin. Correspondingly, the creatinine clearance showed a significant decline after 28 days ( Figure 1B), which was not seen in any of the animals co-treated with CsA and chrysin. Further chrysin co-treatment attenuated increased BUN levels seen with CsA treatment alone ( Figure 1C).

Effect of Chrysin on Histopathological Damage Induced by CsA
Tubular damage induced by CsA treatment and the effects of co-treatment with chrysin were evaluated by H&E staining. Compared to control animals (Figure 2A), renal sections from animals treated with CsA alone ( Figure 2B) showed evidence of tubular dilation and edema. There was also evidence of hyaline casts and tubular vacuolization (magnified in Figure 2B) in animals treated with CsA. Co-treatment with chrysin prevented these changes ( Figure 2D,E).

Effect of Chrysin on CsA-Induced Collagen Deposition
The effect of chrysin on tubulointerstitial fibrosis induced by CsA was evaluated by Masson's trichrome staining. Compared to vehicle-treated rats ( Figure 3A), treatment with CsA ( Figure 3B) resulted in increased deposition of collagen in the interstitium of cortical nephrons as evidenced by intense blue staining. Tissue sections from the animals co-treated with both CsA and chrysin did not show any evidence of elevated collagen deposition ( Figure 3D,E). These histopathological observations were verified with quantification of collagen using a Sircol Sirius red assay ( Figure 3F). Treatment with CsA alone significantly increased the collagen content in the kidney, which was prevented by co-treatment with chrysin. (F) Collagen content quantification in samples of the renal cortex was carried out using the Sircol Sirius red assay. The data are represented as mean ± SEM (n = 6). Statistical significance was determined by one-way ANOVA followed by Tukey's test for multiple comparisons. * indicates p < 0.05 versus control; # indicates p < 0.05 vs. CsA alone.

Effect of Chrysin on TGF-β1-or CsA-Induced Changes in LLC-PK1 Cells
Treatment with either 5 ng/mL TGF-β1 or 4.2 µM CsA for 48 h induced alterations in LLC-PK1 cell morphology from an epithelioid-cobblestone-like appearance ( Figure 4A) towards elongated, spindle-shaped cells ( Figure 4B,D). Co-treatment with chrysin ( Figure 4C,E) inhibited the change in morphology, retaining the epithelioid cobblestone appearance. Treatment with chrysin alone did not appear to change cellular morphology ( Figure 4F). Immunofluorescence was used to evaluate the effect of chrysin on TGF-β1-or CsA-induced expression of mesenchymal markers. Compared to control cells ( Figure 4G) or cells treated with chrysin alone ( Figure 4L), treatment with TGF-β1 ( Figure 4H) or CsA ( Figure 4J) induced expression of α-SMA (green staining). Similarly, the expression of vimentin in cells treated with TGF-β1 ( Figure 4N) or CsA ( Figure 4P) was greater when compared against the basal level seen in controls ( Figure 4M) and with chrysin only ( Figure 4R). The induction in α-SMA ( Figure 4I,K) and vimentin ( Figure 4O,Q) by TGF-β1 or CsA was diminished on co-treatment with chrysin. . Fluorescence for α-SMA and vimentin was quantified and data are represented as percent intensity relative to the control group. Statistical significance was determined by two-way ANOVA followed by Tukey's multiple comparison test. * indicates p < 0.05 versus control; † indicates p < 0.05 versus chrysin co-treatment.
Protein expression analysis ( Figure 5A) further showed that CsA or TGF-β1 treatment reduced epithelial E-cadherin expression while increasing expression of the mesenchymal markers α-SMA and vimentin. Chrysin co-treatment attenuated these effects. Chrysin treatment was also shown to decrease collagen deposition induced by TGF-β1 or CsA treatments ( Figure 5B).
Furthermore, cell migration studies ( Figure 6) showed that CsA or TGF-β1 treatments significantly enhanced cellular movement and wound healing. Co-treatment with chrysin prevented these alterations showing a similar wound area as the controls.

Effect of Chrysin on CsA-Induced TGF-β1 Signaling Cascade
Western blotting was used to determine the influence of chrysin on CsA-induced TGF-β1 signaling ( Figure 7A). Treatment with 4.2 µM CsA increased the cellular expression of TGF-β1 whereas 5 ng/mL TGF-β1 treatment resulted in auto-induction of TGF-β1 expression. Co-treatment with chrysin prevented this observed increase. Downstream, CsA and TGF-β1 both induced phosphorylation of Smad 3 and increased expression of Snail and E-cadherin. Observations of TGF-β1-associated non-Smad pathways indicated that chrysin co-treatment decreased phosphorylation of Akt at the serine 473 position ( Figure 7B) and of GSK-3β at the serine 9 position ( Figure 7C), which were significantly increased after TGF-β1 or CsA treatments alone. The total Akt and GSK-3β levels were unchanged with the various treatments (data not shown). Additionally, intracellular localization of β-catenin was evaluated by immunofluorescence. The control cells had basal levels of β-catenin expression localized mostly to the cell-cell junctions ( Figure 7D). Treatment with TGF-β1 ( Figure 7E) or CsA ( Figure 7G) increased localization of β-catenin within the cytoplasm (white arrowheads), whereas co-treatment with chrysin inhibited increases in the release of membrane-bound β-catenin ( Figure 7F,H).

Discussion
Since the 1970s, cyclosporine A has revolutionized the field of organ transplants due to its immunosuppressant properties [29]. However, one of its major drawbacks has been chronic nephrotoxicity characterized by tubulointerstitial fibrosis [30,31]. Clinically, patients treated with CsA have noted loss of renal function accompanied by increased interstitial fibrosis and tubular atrophy [32][33][34]. Increased levels of TGF-β1 in the kidneys and blood have been documented in multiple types of renal fibrotic pathologies, including cyclosporine A nephropathy [35,36]. In vivo findings in rats treated with CsA [35,37] and in vitro studies using human proximal tubule cells [38,39] have demonstrated the role of EMT in CsA-induced renal fibrosis. Increased TGF-β1 secretion was observed followed by a decline in epithelial junction protein expression and increased mesenchymal characteristics [38]. Together, these studies suggest that targeting TGF-β1-mediated signaling pathways that cause EMT responses could be a potential therapeutic intervention against CsA-induced tubulointerstitial fibrosis. In the present investigation we show that co-treatment with the flavonoid chrysin offers nephroprotection against pharmacological induction of fibrosis by CsA treatment. We have further elucidated a potential mechanism by which chrysin prevents development of CsA-induced fibrosis by inhibiting TGF-β1-mediated EMT.
In this study, Sprague Dawley rats were used to investigate the potentially beneficial effects of chrysin on CsA-induced renal fibrosis. Treatment with CsA caused an increase in BUN and serum creatinine along with a decrease in creatinine clearance, markers for chronic CsA-induced nephropathy [31]. Similar to observations made by Kuruş et al. (2005) [40] and Ateşşahin et al. (2007) [41], kidneys from CsA-treated animals displayed classic signs of CsA nephrotoxicity, such as tubular dilatation, hyaline cast formation in the tubules and tubular vacuolization. Co-treatment with chrysin significantly improved renal function parameters and prevented CsA-induced tubular injury, demonstrating that chrysin offers protection against CsA-induced nephrotoxicity.
Fibrosis is a pathophysiological condition that occurs when chronic inflammatory stimuli that trigger wound healing lead to the excess deposition of extracellular matrix proteins [42]. Chrysin has been observed to attenuate fibrosis in numerous animal models of disease including asthma [43], carbon-tetrachloride-induced liver damage [44], adenineinduced chronic kidney disease [45] and diabetic nephropathy [27,28]. Studies on CsAinduced renal toxicity have shown increased collagen deposition indicative of tubulointerstitial fibrosis [46] as well as increased collagen content in cultured human renal epithelial (HK-2) cells [38], endothelial cells and fibroblasts [47]. In our investigation, chrysin cotreatment inhibited such CsA-induced increases in collagen deposition in the interstitium of rat renal cortices, as well as in cultured proximal tubule cells.
During organ fibrosis, type 2 EMT has been indicated as a mechanism responsible for reparative-associated processes in response to ongoing inflammation and oxidative stress [48]. A classic marker for EMT is change in cellular morphology [49] and cell motility [50], with increased expression of mesenchymal markers and decreased expression of epithelial proteins. Kang and coworkers (2015) [27] demonstrated the protective effects of chrysin against high-glucose-induced EMT in renal proximal tubule epithelial cells (RPTEC), characterized by reduced α-SMA and vimentin and increased E-cadherin expressions. Our observations indicate that CsA-induced morphological changes and expression of the mesenchymal markers, α-SMA and vimentin, are inhibited by co-treatment with chrysin. Chrysin co-treatment also prevented the corresponding decrease in E-cadherin expression. Chrysin has shown efficacy against cancer metastasis by decreasing in-cell motility via EMT inhibition [51]. Our findings reveal that CsA-induced cell migration is significantly decreased in cells co-exposed to chrysin. Both in vivo and in vitro observations point toward a novel role for chrysin in protection against CsA-induced fibrosis via inhibition of EMT.
Studies to understand the molecular basis for anti-fibrotic actions of chrysin suggest that inhibition of CsA-induced EMT occurs by two separate mechanisms: (1) inhibition of TGF-β1/Smad 3 signaling cascade and (2) inhibition of TGF-β1-mediated Akt/GSK-3β signaling, both leading to a decrease in cytoplasmic localization of β-catenin (Figure 8). TGF-β1 phosphorylation of Smad 2 and Smad 3 has previously been linked to the initiation of EMT in multiple cancer cells [52] as well as proximal tubule cells [53]. CsA treatment for up to 24 h increased Smad 3 phosphorylation in rat glomerular mesangial cells [54]. Induction of the Smad pathway by TGF-β1 increases Snail expression [55], a known regulator of EMT [56,57]. Studies by Cano and coworkers (2000) have shown that TGF-β1 can trigger EMT by decreasing expression of epithelial junction proteins through the activation of Snail [56,58]. Downregulation of these proteins releases β-catenin initially immobilized at the adherens junctions [19], resulting in its nuclear translocation and initiation of EMT [59]. This correlates with the findings of the present study, showing that CsA-induced activation of the TGF-β1-Smad cascade led to increased phosphorylation of Smad3 and subsequent expression of Snail. Triggering of this cascade resulted in a decrease in E-cadherin expression leading to increased localization of β-catenin in the cytoplasm. β-catenin accumulation in the cytoplasm is critical to TGF-β1-induced EMT, and blocking its translocation has been shown to inhibit EMT [60]. The present study demonstrated that co-treatment with chrysin prevents this increase in cytoplasmic β-catenin by inhibiting the TGF-β1-Smad-Snail-mediated signaling. TGF-β1 also acts via non-Smad pathways to mediate β-catenin release. Kang and coworkers (2007) [61]. have shown that TGF-β1 signaling activates the PI3K/Akt/GSK-3β signaling cascade, resulting in increased expression of the Ser-9-phosphorylated inactive form of GSK-3β. This prevents the ubiquitination and degradation of cytoplasmic β-catenin, thereby leading to its cytoplasmic accumulation and subsequent nuclear translocation [62,63]. We have shown that chrysin inhibits CsA-induced phosphorylation of Akt and the inactivation of GSK-3β preventing the cytoplasmic accumulation of β-catenin.
Agonists of PPAR-γ, such as telmisartan [64], troglitazone [18] and pioglitazone [65], have the ability to inhibit EMT. It is thus possible that chrysin may additionally act through a similar mechanism, as it is known to increase transcription and protein expression of PPAR-γ [18,66,67]. Increasing evidence of oxidative-stress-induced EMT has been documented in multiple pathologies including diabetic nephropathy [68], interstitial fibrosis of renal allograft [67], an in vitro model for chronic kidney disease [69] and carcinogenesis [70]. Antioxidant properties of chrysin [71][72][73] could be an additional avenue of protection against EMT. Chrysin has been shown to alleviate oxidant injury in lead-acetate-induced renal damage in rats [72]. Its antioxidant properties have also been useful in combating off-target effects of anticancer drugs such as mitomycin C [73]. Another molecule that has been implicated in the induction of EMT is HIF-1α [74,75]. The ability of chrysin to inhibit HIF-1α [76] and evidence of chrysin as a PPAR-γ agonist and antioxidant suggest that chrysin inhibition of CsA-induced EMT may involve additional mechanisms which require further investigation.
In summary, flavonoids such as chrysin offer protection against CsA-induced renal dysfunction and tubulointerstitial fibrosis. Chrysin prevents CsA-induced TGF-β1mediated EMT in proximal tubule cells by inhibiting activation of the Akt/GSK-3β signaling pathway and Smad phosphorylation, leading to a decrease in cytoplasmic β-catenin localization. In conjunction with the in vitro study, the in vivo findings support the potential use of chrysin and its analogs as inhibitors of EMT to prevent fibrosis associated with cyclosporine A therapy.

Animal Studies
Protocols involving the use of animals in these studies were submitted to and approved by the St. John's University Institutional Animal Care and Use Committee (IACUC, protocol #1824, 8 August 2014). Eight-week-old male Sprague Dawley rats (Taconic, Germantown, NY, USA) weighing between 260-280 g were used for the study. To induce renal injury, cyclosporine A was administered subcutaneously at 25 mg/kg [77]. The intraperitoneal doses of chrysin, 10 and 50 mg/kg, were based on published studies [27,45,78]. The animals had access to food and water ad libitum for the duration of the study. Body weights were measured initially and at 3-day intervals. Blood and urine samples were collected on days 0, 14 and 28. Blood samples were obtained from the lateral or ventral tail vein after placing the animals under isofluraneinduced anesthesia. Urine samples were collected using metabolism cages. The animals were acclimatized to the metabolism cages for 24 h and urine samples were collected over the next 24 h. The blood samples were analyzed for blood urea nitrogen (BUN) and creatinine, and urine samples were analyzed for creatinine content. Creatinine clearance was calculated by using the following formula: Creatinine clearence = Urinary creatinine (mg/mL) × Urine volume (mL) Serum creatinine (mg/mL) × Time (min)

Analysis of Urine and Serum Chemistry
Creatinine in urine and serum in addition to serum BUN were measured with clinical reagent kits (Stanbio Labs, Boerne, TX, USA). Serum BUN (Cat #2020-500) was measured by a modification of the Berthelot urease method and creatinine (Cat #0430-500) by the creatinine amidohydrolase/sarcosine oxidase reactions. The methods were adapted for 96-well plate and absorbance was recorded at 550 nm using Promega GloMax Multiplus plate reader (Promega, Madison, WI, USA).

Histology
On day 28 animals were euthanized and kidneys were removed, transected and fixed in 10% neutral buffered formalin. The tissues were processed and embedded in paraffin. For morphological examinations, tissue sections of 4 µm thickness were stained with Gill's hematoxylin and eosin-pholxine (H&E), whereas collagen localization was determined by staining tissue sections with Masson's trichrome.

Cell Culture
LLC-PK1 clone 5 cells, isolated by Dr. S. Ford in 1990 from the established cell line, obtained from ATCC, were grown on 35 or 100 mm BD Falcon™ polystyrene tissue culture dishes (Thermo Fisher Scientific, Somerset, NJ, USA; Cat #353003). The cells maintain the epithelial morphology of the renal proximal tubule as well as characteristic SGLT transporters [79] and oxidative metabolism [80]. They were maintained in the custom culture medium reconstituted to the standard SFFD composition with a final glucose concentration of 5 mM, and supplemented with 3% FBS. The osmolality of the media was 317 mOsm/kg. Cells used in experiments were between passages 218 and 270 and were maintained at 37 • C in an atmosphere of 95% air to 5% CO 2 . The cultures were periodically checked for mycoplasma (Bionique Labs, Saranac Lake, NY, USA; Cat #M100).

Treatment of Cells
The cells seeded at a density of 2 × 10 4 cells/mL were grown to about 60% confluence and then treated with fresh media containing CsA (4.2 µM), TGF-β1 (5 ng/mL), chrysin (25 µM) or combinations of chrysin with TGF-β1 or CsA. In vitro experiments were performed after 48 h of treatment. To determine phosphorylation of Smad 3, the cells were treated for 1 h before samples were collected.

Assessment of EMT in Vitro by Morphological Examination and Immunocytochemistry
LLC-PK1 cells were grown in the 4-well Nunc™ Lab-Tek II CC2™ chamber slide system (Thermo Fisher Scientific, Somerset, NJ, USA; Cat #154917) and treated for 48 h. After treatment, cell morphology was examined under the Nikon (Tokyo, Japan) Ts2R microscope and micrographs were captured. The cell monolayers were then fixed with 4% paraformaldehyde for 15 min at room temperature followed by permeabilization with 0.1% Triton X-100 for 10 min. Monolayers were blocked with 2% w/v bovine serum albumin (BSA) in PBS containing 0.3 M glycine for an hour and then incubated overnight at 4 • C with the primary mouse monoclonal antibodies against α-SMA (1:100), vimentin (1:500) or β-catenin (1:100). Monolayers were then washed with ice-cold PBS, incubated with FITC-labeled anti-mouse IgG antibody (1:1000) for 1 h, then washed thrice with PBS. After the final wash, the slides were mounted in Prolong™ diamond anti-fade reagent with DAPI (Thermo Fisher Scientific, Somerset, NJ, USA; Cat #P36962) and observed under a Nikon (Tokyo, Japan) Ts2R microscope.

Cell Migration Assay
Cell migration was quantified using an ImageJ macro for wound healing as described by Nunes and Dias (2017). In brief, LLC-PK1 cells were grown in 35 mm culture plates. After reaching 90% confluence, wounds~2 mm wide were made in the monolayer with a 1 mL sterile pipette tip. Detached cells were removed by washing with 1 mL media and cells were then treated for 48 h. Micrographs were captured using the Nikon (Tokyo, Japan) Ts2R microscope and cell migration was evaluated using ImageJ software (Bethesda, MD, USA; ver. 1.50i).

Collagen Determination
Collagen content in cell and tissue lysates were determined with the Sircol™ Collagen Assay Kit (Accurate Chemical and Scientific, New York, NY, USA) as previously described [81]. In brief, collagen from samples of the renal cortex were collected and concentrated using acid pepsin digestion (0.1 mg/mL in 0.5 M acetic acid). Acid extracts were then incubated for 30 min with 0.1% Sirius red dye in picric acid at room temperature. The precipitated collagen-dye complex was pelleted by centrifuging at 12,000× g for 10 min. The supernatant was removed, and the pellet was washed with an acid/salt buffer and centrifuged again at 12,000× g for 10 min. The pellet was retained and solubilized with an alkali buffer and absorbance was measured at 560 nm.

Determination of Protein Expression by Western Blot Analysis
Cell lysates were obtained as described in [79,80]. In brief, after treatments cells were pelleted and lysed in RIPA buffer containing protease and phosphatase inhibitors for 30 min on ice, and centrifuged at 12,000× g at 4 • C for 15 min. Protein content was determined using the Bradford method; 25 µg proteins were separated using SDS-PAGE gradient gels (40-20% running gel; BioRad Cat #4561096) and electroblotted onto polyvinylidene fluoride (PVDF) membranes (Immobilon-P, Millipore-sigma; Merck Millipore Burlington, MA, USA; Cat #IPVF00010). The membranes were blocked with 2% BSA in Tris-buffered saline-Tween (TBST) buffer (0.1% Tween-20) for one hour and incubated overnight at 4 • C with the primary antibodies (Table 1). Membranes were then washed and incubated with appropriate secondary antibodies for 1 h. HRP-conjugated β-actin (1:5000) was used as a determinant of equal protein loading. After a series of washes in TBS-Tween buffer, protein bands were visualized by chemiluminescence with an ECL Prime luminescence kit (Amersham, UK) using an Alpha Innotech Fluorochem imager (ProteinSimple, San Leandro, CA, USA). Densitometry measurements were done using ImageJ software (Bethesda, MD, USA; ver. 1.50i).

Determination of Protein Expression by ELISA
LLC-PK1 cells treated for 48 h with TGF-β1 or CsA in the presence or absence of chrysin were lysed with RIPA buffer containing protease inhibitors. Lysates were centrifuged at 12,000× g for 15 min at 4 • C and supernatants were used. Phosphorylation of Akt and GSK was assayed using multispecies Instant One™ ELISA Kits for Akt (phospho/total) and GSK-3β (phospho/total) (Invitrogen, Carlsbad, CA, USA).

Statistics
Statistical analyses were conducted using the GraphPad Prism version 5 for Mac OS X (GraphPad Software, La Jolla, CA, USA; Apple Inc., Cupertino, CA, USA). All experiments were repeated at least four times. Data are expressed as mean ± SEM unless specified otherwise. Differences between the means were analyzed by one-way ANOVA followed by Tukey's post hoc test. The criterion for statistical significance was set at p < 0.05.