A previous communication explains the synthesis and structure of tHGA (Lee et al., 2017). Briefly, a well-stirred mixture of phloracetophenone (1.000 g, 6 mmol), geranyl bromide (0.876 g, 4.80 mmol), and anhydrous potassium carbonate (0.415 g, 3.00 mmol) in dry acetone (3.5 ml) was refluxed for 6 h. The reaction mixture was filtered and evaporated under reduced pressure to give an oily orange residue that was purified by flash column chromatography on Si gel (petroleum ether-ethyl acetate, 10:1) to afford tHGA as a light-yellow powder; m.p. 128–130°C. Purity was more than 99%.

The human bronchial epithelial cell line, BEAS-2B was purchased from American Type Culture Collection (ATCC, USA). The human eosinophilic leukemia cell line, EoL-1 was purchased from RIKEN BioResource Center (RIKEN, Japan). BEAS-2B cells were grown at 37°C, 5% CO₂, in DMEM supplemented with 6 mM L-glutamine, 10% FBS and 100 U/ml streptomycin and penicillin. BEAS-2B cells were subcultured at 80% confluency to avoid squamous epithelial differentiation. EoL-1 cells were grown at the same atmospheric conditions but in RPMI-1640 medium supplemented with 2 mM L-glutamine, 10% FBS and 100 U/ml streptomycin and penicillin. Cells with passage number below seven were used in all experiments.

The eosinophil-induced EMT model was adopted from Yasukawa et al. (2013). Prior to each experiment, BEAS-2B cells were cultured to 60–70% confluency in 6-well plates and serum-starved with 1% FBS for 24 h. EoL-1 cells were subjected to differentiation and maturation in media supplemented with 0.5 mM sodium n-butyrate (BA) at 5 × 10⁵ cells/ml for 5 days. Non-cytotoxic doses of tHGA were determined prior to further experiments by the MTT viability assay following 24 h incubation of cells with varying doses of tHGA. For the co-culture experiments, serum-starved BEAS-2B cells were pretreated with serially-diluted tHGA (30 to 7.5 μM) in 0.1% DMSO for 1 h. Following removal of tHGA treatment, cells were washed with sterile PBS. A total of 2 × 10⁶ cells/well of BA-differentiated EoL-1 in RPMI-1640 media were added to BEAS-2B cultures for 48 h. Subsequently, EoL-1 cells were removed from BEAS-2B cells by gentle pipetting and washed thrice with PBS prior to assay. Co-culture experimental groups were as follow:

For TGF-β-induced EMT experiments, BEAS-2B were induced by adding 5 ng/ml recombinant TGF-β1 (Merck Millipore, USA) for 48 h. Experimental groups for TGF-β-induced experiments were as follow:

In tHGA target identification experiments, serum-starved BEAS-2B cells were pretreated with 30 μM tHGA or respective internal control inhibitors (Table 1) in 0.1% DMSO for 1 h. Treatments were removed and cells were washed with PBS prior to co-culture of BA-differentiated EoL-1 with BEAS-2B for 1 h. EoL-1 cells were removed from BEAS-2B cells by gentle pipetting and washed thrice with PBS prior to assay. Experimental groups for tHGA target identification experiments were as follow:

BEAS-2B cells were photographed at 400x magnification. Images of three arbitrary fields were taken for each treatment group. Quantitative methods of morphological changes in EMT were adopted from Ren et al. (2015). Briefly, the centroid of a cell is determined. The radius ratio was obtained by dividing the maximum radius with the minimum radius of a cell measured from the centroid to the cell membrane. All measurements were taken with a Leica Microsystems microscope imaging software (Leica Microsystems, Germany). All whole cells in each field were measured.

Whole cell extracts of BEAS-2B cells were obtained with Chemicon Total Protein Extraction Kit (Merck Millipore, USA) whereas protein samples of cytoplasmic and nuclear fractions were extracted with NucBuster Protein Extraction Kit (Merck Millipore, USA) according to the manufacturer's instructions. Protein samples of 20 μg were electrophoresed on 8–12% SDS-polyacrylamide gels and transferred to 0.2 μm polyvinylidene fluoride membranes using a wet transfer system at 0.35 mA for 1.5 h (Bio-Rad Laboratories, USA). Membranes were blocked with 5% bovine serum albumin in Tris-buffered saline-tween 20 (TBS-T) for 1 h prior to overnight incubation at 4°C with respective antibodies (Table 2) in 5% BSA TBS-T. Following washing with TBS-T, membranes were hybridized with horse radish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody (1:5000; Santa Cruz Biotechnology, USA) or HRP-conjugated goat anti-mouse secondary antibody (1:5000; Santa Cruz Biotechnology, USA) for 1 h at room temperature. Membranes were then incubated with chemiluminescence reagent (Thermo Scientific, USA) for 1 min and imaged in a gel documentation system (Vilber Lourmat, Germany). Membranes were stripped and reprobed as required. Band intensities were quantified by ImageJ Image Processing Software (NIH, USA) and normalized by comparison to β-actin or Lamin A/C.

Total RNA was isolated and purified from BEAS-2B cells following co-culture with EoL-1 for 48 h using Qiagen RNeasy Plus Mini Kit (Qiagen, USA) according to the manufacturer's protocol. RNA concentration and purity were determined by using the Implen Nanophotometer P300 (Implen, USA) while RNA integrity was examined by formaldehyde agarose gel electrophoresis. RNA samples of 100 ng were used with Qiagen One-Step reverse transcriptase-polymerase chain reaction (RT-PCR) kit (Qiagen, USA) according to the protocol recommended by the manufacturer. The master mix was performed in an Eppendorf thermal cycler (Eppendorf, Germany) for reverse transcription at 50°C for 30 min, initial PCR activation at 95°C for 15 min, denaturation at 94°C for 30 s, annealing at 57°C for E-cadherin, collagen I, fibronectin and β-actin for 30 s, 59°C for vimentin and β-actin for 30 s, and 65°C for TGF-β and β-actin for 30 s, elongation at 72°C for 1 min and final extension at 72°C for 10 min. The sequences of the primers (IDT, USA) are listed in Table 3. PCR was carried out for 35 cycles. The reaction products from PCR were examined by 1.8% agarose gel electrophoresis containing 0.01% of ethidium bromide. PCR products in each gel electrophoresis were ran in parallel to a Low Molecular Weight DNA Ladder (NEB, USA). Band intensities were quantified by ImageJ Image Processing Software (NIH, USA) and normalized by comparison to the RT-PCR product of β-actin mRNA.

Spent media from co-culture experiments was collected and centrifuged to remove EoL-1 cells and debris. The concentration of secreted TGF-β1 was quantified using a commercial enzyme-linked immunosorbent assay kit (R&D Systems, USA) according to manufacturer's protocol and subjected to correction with media native TGF-β1 levels.

The binding activity of activator protein (AP)-1 in nuclear fractions was assayed using an electrophoretic mobility shift assay. AP-1 consensus double stranded oligonucleotide (5′-CGC TTG ATG ACT CAG CCG GAA-3′) (Santa Cruz Biotechnology, USA) was end-labeled with Pierce Biotin 3′ End DNA Labeling Kit (Thermo Fisher Scientific, USA). Briefly, 10 μg of nuclear protein samples were incubated with 100 fmol of oligonucleotides at room temperature for 20 min and separated on a 6% polyacrylamide gel. The specificity of binding was determined by 200-fold unlabeled oligonucleotide competition. The AP-1-DNA complexes were transferred onto 0.2 μm polyvinylidene fluoride membranes and incubated with chemiluminescence reagent (Thermo Scientific, USA) for 1 min and imaged in a gel documentation system (Vilber Lourmat, Germany). Band intensities were quantified by ImageJ Image Processing Software (NIH, USA).

All experiments were performed a minimal of three times with triplicate samples. Data were analyzed using one-way analysis of variance. Significant differences between experimental groups were tested using Dunnett's post hoc test by comparing to induced-controls (Group C –eosinophils or TGF-β-induced BEAS-2B cells). All data is expressed as mean ± S.E.M. Differences are considered significant when P < 0.05.

As established previously (Yasukawa et al., 2013), the morphology of BEAS-2B cells was altered from cobble-stone shaped to spindle fibroblast-like morphology following co-culture with EoL-1, an indication of EMT (Figure 1A). Pretreatment with both 30 and 15 μM tHGA inhibited EMT as evident from the morphology and radius ratio of BEAS-2B cells (Figure 1C). Bronchial epithelial cells lose protein expression of the epithelial marker E-cadherin and start expressing the mesenchymal marker vimentin following co-culture with eosinophils. Treatment with tHGA prevented the loss of epithelial marker and mesenchymal marker upregulation from occurring (Figures 1B,D,E). Data further demonstrates that prevention of alterations in epithelial and mesenchymal marker expression were due to the effect of tHGA on gene expression (Figures 1F,G).

The expression of fibronectin and collagen I which are uniquely associated with mesenchymal phenotype are shown in Figure 2. Western blots revealed that tHGA, particularly at a concentration of 30 μM, was effective in preventing fibronectin and collagen I protein expression which were significantly increased following EoL-1 co-culture. A similar trend was noted in the expression of the respective genes.

As described by Hosoki et al. (2014) and Yasukawa et al. (2013), TGF-β1 production was enhanced following co-culture of BEAS-2B with EoL-1. We demonstrated that that although monoculture of either cell line did not exhibit high TGF-β1 secretion, co-culture of BEAS-2B and EoL-1 significantly enhanced the release of TGF-β1 as determined by ELISA (Figure 3A). Interestingly, tHGA suppressed the secretion (Figure 3A) and mRNA expression (Figures 3B,C) of TGF-β1 by BEAS-2B co-cultured with EoL-1, in a concentration-dependent manner.

Since it has been suggested that EMT of eosinophil-induced bronchial epithelial cells is TGF-β dependent (Yasukawa et al., 2013), we assessed the effect of tHGA upon BEAS-2B cells induced by recombinant human TGF-β. Exposure of BEAS-2B cells to recombinant TGF-β induced altered morphology, loss of E-cadherin expression and increased vimentin expression. However, we observed no significant suppression of TGF-β-induced EMT following treatment with tHGA (Figure 4).

It was established previously that the activation of MAPK and PI3K pathways leads to the expression of TGF-β in BEAS-2B after co-culture with EoL-1 (Yasukawa et al., 2013). Figure 5 shows the effect of tHGA upon BEAS-2B cell MAPKs following EoL-1-induction. Induction successfully activated phosphorylation of all MAPKs. tHGA treatment caused significant inhibition of JNK phosphorylation (Figure 5A) without any effect upon ERK1/2 (Figure 5B) and p38 (Figure 5C).

We examined the activation of the PI3K/AKT pathway and found that interaction between EoL-1 and BEAS-2B cells resulted in the phosphorylation of PI3K and AKT at both phosphorylation sites (Thr308 and Ser473) (Figures 6A–C,E). Activated AKT can phosphorylate GSK-3β at Ser9 which then inactivates GSK-3β (McCubrey et al., 2014). Since GSK-3β is associated with increased stability and availability of c-Jun for activation via phosphorylation (Wei et al., 2005), we also examined whether EoL-1/BEAS-2B co-culture leads to GSK-3β phosphorylation and whether this was blocked following tHGA treatment (Figures 6D,E). It was apparent that tHGA inhibited the phosphorylation of all these molecules.

AP-1 structurally consists of a homodimer of c-Jun or heterodimer of c-Jun and Fos or ATF protein families (Hess, 2004), it is a major regulator of TGF-β1 transcription (Sullivan et al., 2009). Following eosinophil induction c-Jun was phosphorylated and translocated into the nucleus of BEAS-2B cells. Treatment with tHGA inhibited c-Jun phosphorylation and translocation (Figures 7A,B). Since tHGA inhibited c-Jun phosphorylation we confirmed blockade of downstream AP-1 DNA binding through an electrophoretic mobility shift assay. As shown in Figures 7C,D, AP-1 DNA binding in BEAS-2B cells increased following EoL-1 co-culture whereas tHGA treatment caused a significant decrease in AP-1 DNA binding.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.