16-Hydroxycleroda-3,13-Dien-15,16-Olide Induces Apoptosis in Human Bladder Cancer Cells through Cell Cycle Arrest, Mitochondria ROS Overproduction, and Inactivation of EGFR-Related Signalling Pathways

A clerodane diterpene compound 16-hydroxycleroda-3,13-dien-15,16-olide (CD) is considered a therapeutic agent with pharmacological activities. The present study investigated the mechanisms of CD-induced apoptosis in T24 human bladder cancer cells. CD inhibited cell proliferation in a concentration and time-dependent manner. CD-induced overproduction of reactive oxygen species and reduced mitochondrial membrane potential, associated with reduced expression of Bcl-2 and increased levels of cytosolic cytochrome c, cleaved PARP-1 and caspase-3. In addition, CD treatment led to cell cycle arrest at the G0/G1 phase and inhibited expression of cyclin D1 and cyclin-dependent kinases 2 and 4 and led to increased levels of p21, p27Kip1 and p53. All of these events were accompanied with a reduction of pEGFR, pMEK1/2, pERK1/2, pAkt, pmTOR, pP70S6K1, HIF-1α, c-Myc and VEGF. RNAseq-based analysis revealed that CD-induced cell death was characterised by an increased expression of stress and apoptotic-related genes as well as inhibition of the cell cycle-related genes. In summary, CD induces apoptosis in T24 bladder cancer cells through targeting multiple intracellular signaling pathways as a result of oxidative stress and cell cycle arrest.


Introduction
16-hydroxycleroda-3, 13-dien-16, 15-olide (CD) is one of the major active compounds of clerodane diterpenes and has been found to exhibit numerous pharmacological functions, such as anti-lipogenic [1], anti-leishmanial [2], anti-fungal [3], anti-inflammation [4], and anti-cancer activities [5]. Polyalthia longifolia is a tall evergreen tree native to Pakistan, India, and Sri Lanka that To investigate whether CD-induced cell death is associated with apoptosis, we used AO/EB double staining and the annexin V-PI assay for visualisation and quantification of apoptosis. CDinduced cell shrinkage and apoptosis, and necrosis was rare after 24 h of treatment ( Figure 2A). Compared with the control group, CD-treated T24 cells had high levels of early and late apoptotic cells, but lower levels of necrotic cells (<2%) after 24 h of treatment ( Figure 2B). To investigate whether CD-induced cell death is associated with apoptosis, we used AO/EB double staining and the annexin V-PI assay for visualisation and quantification of apoptosis. CD-induced cell shrinkage and apoptosis, and necrosis was rare after 24 h of treatment ( Figure 2A). Compared with the control group, CD-treated T24 cells had high levels of early and late apoptotic cells, but lower levels of necrotic cells (<2%) after 24 h of treatment ( Figure 2B).

CD Suppresses MMP and Triggers ROS Production
To examine whether CD-induced cell death involves mitochondria-dependent apoptotic mechanisms, the fluorescent dyes Rh123 and JC-1 were used to measure the MMP. As shown in Figure 3A, vehicle-treated BC cells exhibit strong green fluorescence in the mitochondria, whereas treatment with CD eliminated Rh123 staining starting at concentrations of 10 µM ( Figure 3A). As shown in Figure 3B, there are reduced JC-1 aggregates and monomers in CD-treated cells, especially at 20-40 µM ( Figure 3B). We also analysed the MMP using flow cytometry and observed decreased fluorescence intensity after treatment with CD ( Figure 3C,D). In summary, compared with the vehicle-treated control, both Rh123 and JC-1 staining in T24 cells were reduced.

CD Suppresses MMP and Triggers ROS Production
To examine whether CD-induced cell death involves mitochondria-dependent apoptotic mechanisms, the fluorescent dyes Rh123 and JC-1 were used to measure the MMP. As shown in Figure 3A, vehicle-treated BC cells exhibit strong green fluorescence in the mitochondria, whereas treatment with CD eliminated Rh123 staining starting at concentrations of 10 µM ( Figure 3A). As shown in Figure 3B, there are reduced JC-1 aggregates and monomers in CD-treated cells, especially at 20-40 µM ( Figure 3B). We also analysed the MMP using flow cytometry and observed decreased fluorescence intensity after treatment with CD ( Figure 3C,D). In summary, compared with the vehicle-treated control, both Rh123 and JC-1 staining in T24 cells were reduced.  was used as a positive control for confirming the successful inhibition of MMP. Bar graphs show values normalised as percentages relative to the control. Data are expressed as mean ± SD of three (Rh123) and four (JC-1) independent experiments. * p < 0.05,** p < 0.01, *** p < 0.001 as compared with the control group (0 µM).
The loss of MMP leads to an energy imbalance and triggers ROS production. We used the fluorescent dye DCFH-DA as an indicator of peroxide and superoxide accumulation. After treatment with CD, we observed a concentration-dependent increase in ROS production ( Figure 4A,B). Quantitative analysis showed that treatment with 30 and 40 µM of CD significantly elevated ROS production to 1.5 times basal levels ( Figure 4B). Pre-treatment with the antioxidant reagent N-acetyl cysteine (NAC, 0.25-1 mM) reversed CD-induced cytotoxicity in a concentration-dependent manner ( Figure 4C). CD-induced mitochondria ROS production can be observed immediately 8 h after treatment ( Figure 4D) and the effect can be blocked by NAC. The representative photos were taken from one of the three independent experiments. Flow cytometry analysis of MMP by (C) Rh123 (D) JC-1 staining. Carbonyl cyanide 3-chlorophenylhydrazone (CCCP) was used as a positive control for confirming the successful inhibition of MMP. Bar graphs show values normalised as percentages relative to the control. Data are expressed as mean ± SD of three (Rh123) and four (JC-1) independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001 as compared with the control group (0 µM).
The loss of MMP leads to an energy imbalance and triggers ROS production. We used the fluorescent dye DCFH-DA as an indicator of peroxide and superoxide accumulation. After treatment with CD, we observed a concentration-dependent increase in ROS production ( Figure 4A,B). Quantitative analysis showed that treatment with 30 and 40 µM of CD significantly elevated ROS production to 1.5 times basal levels ( Figure 4B). Pre-treatment with the antioxidant reagent N-acetyl cysteine (NAC, 0.25-1 mM) reversed CD-induced cytotoxicity in a concentration-dependent manner ( Figure 4C). CD-induced mitochondria ROS production can be observed immediately 8 h after treatment ( Figure 4D) and the effect can be blocked by NAC.

CD Triggers the Expression of Pro-Apoptotic Proteins and Inhibits Anti-Apoptotic Proteins
CD treatment led to increased cytochrome c release from mitochondria to the cytosol ( Figure  5A-C). Therefore, investigated downstream signalling molecules in the caspase-related pathway by western and found that treatment with CD enhanced expression of cleaved caspase-3 (17 and 19 kDa subunits), cleaved PARP-1 and pH2A.X ( Figure 5D-H). CD treatment also inhibited Bcl-2 antiapoptotic protein expression. These results suggest that CD induces mitochondrial-dependent apoptosis in BC cells through activation of the caspase-3 dependent pathway. Scale bar = 50 µm. Bar graphs show values normalised as percentages relative to the control. Data are expressed as mean ± SD of four independent experiments. In (B), ** p < 0.01 compared with control group (0 µM). In (C), # p < 0.05 compared with control group; *** p < 0.001 compared with CD group.

CD Triggers the Expression of Pro-Apoptotic Proteins and Inhibits Anti-Apoptotic Proteins
CD treatment led to increased cytochrome c release from mitochondria to the cytosol ( Figure 5A-C). Therefore, investigated downstream signalling molecules in the caspase-related pathway by western and found that treatment with CD enhanced expression of cleaved caspase-3 (17 and 19 kDa subunits), cleaved PARP-1 and pH2A.X ( Figure 5D-H). CD treatment also inhibited Bcl-2 anti-apoptotic protein expression. These results suggest that CD induces mitochondrial-dependent apoptosis in BC cells through activation of the caspase-3 dependent pathway.

Effects of CD on Cell Cycle Progression, Cyclins and Cyclin-Dependent Kinases (CDKs)
To determine whether CD inhibits proliferation, we assessed cell cycle markers in CD-treated T24 cells. Compared with the control (0 µM CD, 39.8% of cells in G0/G1), CD treatment led to a cell cycle arrest of T24 cells (50.1%, 59.95%, 70.9%, and 73.9% of cells in arrest at 10, 20, 30, and 40 µM concentrations of CD, respectively, Figure 6A) in the G0/G1 phase after 24 h of treatment. This was associated with a decrease of the number of cells in the S and G2/M phases of the cell cycle ( Figure 6A). Accordingly, the G1 phase regulatory proteins, including cyclin D1, CDK2, CDK4, p21, p27Kip1, and p53 were analysed using western blot. CD treatment reduced protein levels of CDK2, CDK4, and cyclin D1, but increased levels of CDK inhibitors, p21 and p27Kip1, in a concentration-dependent manner after 24 h of treatment ( Figure 6B-G). The protein p53 was also elevated via CD treatment ( Figure 6H).

Effects of CD on Cell Cycle Progression, Cyclins and Cyclin-Dependent Kinases (CDKs)
To determine whether CD inhibits proliferation, we assessed cell cycle markers in CD-treated T24 cells. Compared with the control (0 µM CD, 39.8% of cells in G0/G1), CD treatment led to a cell cycle arrest of T24 cells (50.1%, 59.95%, 70.9%, and 73.9% of cells in arrest at 10, 20, 30, and 40 µM concentrations of CD, respectively, Figure 6A) in the G0/G1 phase after 24 h of treatment. This was associated with a decrease of the number of cells in the S and G2/M phases of the cell cycle ( Figure  6A). Accordingly, the G1 phase regulatory proteins, including cyclin D1, CDK2, CDK4, p21, p27Kip1, and p53 were analysed using western blot. CD treatment reduced protein levels of CDK2, CDK4, and cyclin D1, but increased levels of CDK inhibitors, p21 and p27Kip1, in a concentration-dependent manner after 24 h of treatment ( Figure 6B-G). The protein p53 was also elevated via CD treatment ( Figure 6H).

CD Modulates the Epidermal Growth Factor Receptor-Mediated Signalling Pathway
To determine whether CD induces apoptosis by modulation of the EGFR pathway, we measured the expression of pEGFR (Tyr1173), pAkt1 (Ser473), pmTOR (Ser2481), pMEK (Ser 217/221), pERK1/2 (Thr202/Tyr204), and pP70S6K1 (Thr 389) as well as the corresponding total protein levels of each protein. This data suggests that CD inhibits the expression of proteins involved in multiple oncogenic pathways, including the EGFR, Akt, mTOR-P70S6K1, and MEK-ERK pathways. We assessed levels of downstream targets of EGFR, c-Myc, and HIF-1α and found that after 24 h of treatment, CD significantly suppresses c-Myc and HIF-1α expression in a concentration-dependent manner ( Figure  8A-D). Treatment with CD also reduced the expression of vascular endothelial growth factor (VEGF). Figure 8E displays the predicted pathway through which CD induces apoptosis, cell cycle arrest, and inhibits proliferation. All of the proteins in this network are listed as having strong interactions in the string database (string-db.org). This data suggests that CD inhibits the expression of proteins involved in multiple oncogenic pathways, including the EGFR, Akt, mTOR-P70S6K1, and MEK-ERK pathways. We assessed levels of downstream targets of EGFR, c-Myc, and HIF-1α and found that after 24 h of treatment, CD significantly suppresses c-Myc and HIF-1α expression in a concentration-dependent manner ( Figure 8A-D). Treatment with CD also reduced the expression of vascular endothelial growth factor (VEGF). Figure 8E displays the predicted pathway through which CD induces apoptosis, cell cycle arrest, and inhibits proliferation. All of the proteins in this network are listed as having strong interactions in the string database (string-db.org).

Expression Profiling of CD-Triggered Cell Death
As certain morphological signs of apoptosis were observed after treatment of CD on cultured T24 cells (Figures 2-5), we investigated how the cells responded to CD at an earlier time (12 h) through whole-genome expression profiling. The total mapped reads were on average 45,771,919 and 43,699,180 with 95.32% and 95.99% mapping rates for CD treated cells and the control samples, respectively. The threshold to the absolute confect score was set at 0.77, which resulted in a list of 485 significantly expressed genes, including 308 up-regulated and 177 down-regulated ones (Table S1 (Supplementary Material)). Within those up-regulated ones, 77 genes involved in apoptosis and 23 in autophagy were found ( Figure 9A). To further determine the detailed responses after CD treatment, functional enrichment analysis was performed on g-Profiler using terms of biological process in the Gene Ontology database (http://geneontology.org/) [16] As shown in Figure 9B, upregulated genes were involved in the biological processes in response to cell stress, protein unfolding,

Expression Profiling of CD-Triggered Cell Death
As certain morphological signs of apoptosis were observed after treatment of CD on cultured T24 cells (Figures 2-5), we investigated how the cells responded to CD at an earlier time (12 h) through whole-genome expression profiling. The total mapped reads were on average 45,771,919 and 43,699,180 with 95.32% and 95.99% mapping rates for CD treated cells and the control samples, respectively. The threshold to the absolute confect score was set at 0.77, which resulted in a list of 485 significantly expressed genes, including 308 up-regulated and 177 down-regulated ones (Table S1 (Supplementary Material)). Within those up-regulated ones, 77 genes involved in apoptosis and 23 in autophagy were found ( Figure 9A). To further determine the detailed responses after CD treatment, functional enrichment analysis was performed on g-Profiler using terms of biological process in the Gene Ontology database (http://geneontology.org/) [16] As shown in Figure 9B, up-regulated genes were involved in the biological processes in response to cell stress, protein unfolding, and organic substances stimulus. In addition, genes involved in cell death and apoptosis were also up-regulated ( Figure 9B). On the other hand, down-regulated genes were mainly playing roles in regulating the mitotic cell cycle, such as cell division, nuclear division, chromatid segregation, and so on ( Figure 9C). Interestingly, when looking into the signaling pathway of the cell cycle, we noticed an overall suppression of cyclins (CCNE2, CCNA2, CCNB1) and other factors in multiple stages except for the inhibitory CDKN1A (protein p21), which exhibited more than 4-fold up-regulation after the CD treatment ( Figure 9D). These features found in the expression profile showed that the T24 cell was responding to stress and under suppression of the cell cycle upon the CD treatment.
Molecules 2020, 25, x FOR PEER REVIEW 11 of 17 and organic substances stimulus. In addition, genes involved in cell death and apoptosis were also up-regulated ( Figure 9B). On the other hand, down-regulated genes were mainly playing roles in regulating the mitotic cell cycle, such as cell division, nuclear division, chromatid segregation, and so on ( Figure 9C). Interestingly, when looking into the signaling pathway of the cell cycle, we noticed an overall suppression of cyclins (CCNE2, CCNA2, CCNB1) and other factors in multiple stages except for the inhibitory CDKN1A (protein p21), which exhibited more than 4-fold up-regulation after the CD treatment ( Figure 9D). These features found in the expression profile showed that the T24 cell was responding to stress and under suppression of the cell cycle upon the CD treatment.

Discussion
BC is the tenth most common cancer worldwide, with approximately 550,000 new cases annually [9,17]. However, little progress has been made in the development of new treatments in recent years. Here, we demonstrate that a clerodane diterpene compound isolated from P. longifolia (termed CD) inhibits cell proliferation, stimulates cell cycle arrest, and induces mitochondria-dependent apoptosis. This is associated with inactivation of the EGFR, Akt, mTOR and MEK1/2-ERK1/2 signalling pathways, as well as their downstream effectors c-Myc, HIF-1α and VEGF. Genes with altered expression are color-coded for the log2 fold changes (red: up-regulated; blue: down-regulated; +p phosphorylation; −p dephosphorylation). The KEGG pathway map was acquired from Kyoto Encyclopedia of Genes and Genomes (https://www.genome.jp/kegg/) and the plot was performed by Cytoscape.

Discussion
BC is the tenth most common cancer worldwide, with approximately 550,000 new cases annually [9,17]. However, little progress has been made in the development of new treatments in recent years. Here, we demonstrate that a clerodane diterpene compound isolated from P. longifolia (termed CD) inhibits cell proliferation, stimulates cell cycle arrest, and induces mitochondria-dependent apoptosis. This is associated with inactivation of the EGFR, Akt, mTOR and MEK1/2-ERK1/2 signalling pathways, as well as their downstream effectors c-Myc, HIF-1α and VEGF.
Apoptotic events can be triggered by intrinsic and extrinsic signals [18]. Recent work has focused on inducing mitochondrial dysfunction as a therapy for cancer in recent years [19]. Reduced MMP is an early step in cell death and can be caused by ROS overproduction [19,20]. Our results suggest that CD triggers a loss of MMP and ROS overproduction as well as subsequent release of cytochrome c into the cytosol and activation of caspase-3, PARP-1, and pH2A.X. Thus, our results support the hypothesis that CD induces programmed cell death, at least in part, through mitochondrial-dependent intrinsic apoptosis. On the other hand, the membrane phospholipid containing a high level of polyunsaturated fatty acid is susceptible to ROS attack, which is called lipid peroxidation. ROS-induced lipid peroxidation propagates cytotoxicity and is associated with apoptosis, autophagy, and ferroptosis [21]. RNAseq data indicated that after exposure to CD for 12 h, the MAPK pathway and genes related to ferroptosis were also up-regulated ( Figure S1 (Supplementary Material)). Recently, ferroptosis has appeared to be a new approach for manipulating cancer cell death [22]. Research has been shown that activation of the MAPK pathway may sensitize cancer cells to ferroptosis [22]. Moreover, the ferroptotic cell death is driven by iron-dependent lipid peroxidation which accumulates ROS and causes cell death. It raises the higher possibility that CD might induce ferroptosis prior to apoptosis in the T24 BC cell; however, more experiments are needed to clarify. Additionally, recent studies demonstrated that excessive induction of autophagy causes autophagic cell death, overcoming drug resistance in leukaemia and bladder cancer [23,24]. Now we are focusing on the detailed autophagic pathway to determine the connection between ROS, apoptosis, and autophagy, induced by CD treatment in T24 cells.
The effects of CD have been examined on several other cancer cell lines, including oral cancer [5], leukaemia [7], and renal cell carcinoma [8]. The current study is the first to observe the inactivation of the EGFR pathway. In BC, approximately over 50% of tumour specimens presented higher EGFR expression and were associated with recurrent, high-grade, and high-stage patients, as well as having a worse prognosis [25]. It is worth noting that even in the absence of EGFR, the downstream signaling pathways (MAPK and Akt) still activate [14]. Our findings show that CD not only blocked EGFR activation but attenuated two main signaling pathways, as well as preventing HIF-1α, c-Myc, and VEGF expression. In vivo studies have shown that the inhibition of EGFR [26], VEGF [27], and mTOR [28] suppresses bladder tumour growth. These results imply that CD might serve as a good candidate for bladder cancer treatment. Further studies of the specific molecular mechanisms that are involved in CD-induced anti-cancer activity in BC are required.
In conclusion, the present study demonstrates that CD inhibits cell proliferation and induces mitochondrial-dependent apoptosis via an EGFR-mediated signalling network.

Plant Authentication and Extraction
The leaves of P. longifolia were collected locally from Kaohsiung, Taiwan in the summer of 2013 and were verified by Prof. Yi-Chen Chia. The extraction of 16-hydroxycleroda-3,13-dien-15,16-olide was performed as described previously, and the corresponding NMR data is shown by Liu et al. [8].

Morphology Observation and Cell Viability Assay
The morphological changes were observed using a phase-contrast microscope (Leica, Wetzlar, Germany) at 100× magnification.

Acridine Orange and Ethidium Bromide (AO/EB) Double Staining
A dual fluorescent staining solution (3 µL) containing 100 µg/mL AO (Molecular probes TM , Thermo Fisher Scientific Inc., Waltha, MA, USA, #A1301) and 100 µg/mL EB (Sigma, #E1510) was added to treated cell suspensions (97 µL) and added to a slide and coverslipped. Cell morphology was examined using a fluorescent microscope with cooled-CCD (NIKON, Tokyo, Japan) at 200× and 400× magnification. AO permeates cells, resulting in cell nuclei fluorescing green, while EB is only taken up by cells with disruptions to cell membrane integrity [29]. Cells that die from necrosis have a structurally normal orange nucleus.

Fluorescence Staining for MMP and Mitochondrial ROS
T24 cells were treated with different concentrations of CD for 24 h. After treatment, cells were incubated with Rh123 (Invitrogen TM , 10 µg/mL) and JC-1 (Molecular probes TM , 2.5 µM) for 30 min at 37 • C, respectively. For the detection of mitochondrial superoxide, MitoSOX TM red indicator (2.5 µM, Molecular probes TM , #M36008) was performed on the T24 cells for 10 min at 37 • C. Staining was visualised using a cooled-CCD (NIKON, Japan) at 200× and 400× magnifications.

Cell Cycle Analysis
Cells (3 × 10 5 cells per well) were serum-starved for 24 h prior to incubation with different concentrations of CD. After 24 h of treatment, cells were collected and washed with ice-cold PBS and fixed overnight with 70% ethanol at −20 • C. Cells were then centrifuged at 2000× g for 10 min, washed with ice-cold PBS and incubated with the staining buffer (0.2 mg/mL RNase A, 20 µg/mL PI, and 0.1% Triton X-100) for 30 min at 37 • C in the dark. Stained cells were measured using a flow cytometer (Beckman Coulter) and the results were analysed using the Expo32 ADC analysis software (Beckman Coulter).

RNA Extraction and Expression Profiling
T24 cells were treated with or without CD at the concentration of 40 µM for 12 h and then were purified with an RNA kit (Geneaid Biotech Ltd., Taipei, Taiwan) following the instructions. The RNA samples (about 1 µg) were sent to GENEWIZ (Suzhou, China) for RNA sequence analysis. Two experimental replicates for each condition were performed, and the purified RNA samples were sequenced with NovaSeq TM platform (Illumina ® , San Diego, CA, USA) for the depth of 50 million paired-end reads. The resulting sequence reads were filtered to remove those of short fragments (<75 bp) or with low sequencing quality (Q < 20). The qualified sequence reads were aligned and quantified using the Hisat2 (v2.0.1) program [30] with its implemented bowtie2. The dispersion and difference of expression levels between the CD-treated cells and the untreated were estimated via Deseq2 package [31] using the raw read count data, where a minimum of 10 counts in total across all samples was set as the criteria for the estimation. The resulting log2 fold changes were further evaluated for the confidence bounds using Topconfects [32] The false discovery rate was set to 0.05 and the step size was set to 0.01 for scoring and ranking of the confidence bounds. For enrichment analysis, a confect score of 0.77 was implemented as a threshold to select the genes on the top of the ranked list. The threshold was determined so that the resulting list contained approximately (yet not more than) 500 genes with significantly altered expression. Up-regulated and down-regulated genes were separated and submitted onto the web based g-Profiler for functional enrichment analysis [16].

Statistical Analysis
Quantitative data are expressed as means ± SD from at least three independent experiments using GraphPad Prism 7.0 software (GraphPad Software, San Diego, CA, USA). One-way analysis of variance with Dunnett's post-hoc was applied to compare the treated group to the control group. p-values of <0.05 were considered statistically significant.