Pomegranate Extract (POMx) Induces Mitochondrial Dysfunction and Apoptosis of Oral Cancer Cells

The anticancer effect of pomegranate polyphenolic extract POMx in oral cancer cells has rarely been explored, especially where its impact on mitochondrial functioning is concerned. Here, we attempt to evaluate the proliferation modulating function and mechanism of POMx against human oral cancer (Ca9-22, HSC-3, and OC-2) cells. POMx induced ATP depletion, subG1 accumulation, and annexin V/Western blotting-detected apoptosis in these three oral cancer cell lines but showed no toxicity to normal oral cell lines (HGF-1). POMx triggered mitochondrial membrane potential (MitoMP) disruption and mitochondrial superoxide (MitoSOX) generation associated with the differential downregulation of several antioxidant gene mRNA/protein expressions in oral cancer cells. POMx downregulated mitochondrial mass, mitochondrial DNA copy number, and mitochondrial biogenesis gene mRNA/protein expression in oral cancer cells. Moreover, POMx induced both PCR-based mitochondrial DNA damage and γH2AX-detected nuclear DNA damage in oral cancer cells. In conclusion, POMx provides antiproliferation and apoptosis of oral cancer cells through mechanisms of mitochondrial impairment.

The qRT-PCR primer information for antioxidant-and mitochondrial biogenesisrelated genes, as well as GAPDH, are provided in the top and bottom of Table 1, respectively. In reference to housekeeping gene GAPDH, the relative mRNA expression (log 2 ) of these genes was calculated according to the 2 −∆∆Ct method [40]. In brief, ∆Ct is calculated as (Ct value of a target gene-Ct value of GAPDH gene), where the target genes are antioxidant-and mitochondrial biogenesis-related genes. When the Ct value of a target gene is undetectable (>50 cycles; qRT-PCR is performed for 50 cycles), it was assigned 50 cycles for further calculation. The ∆∆Ct is the difference in ∆Ct between the drug treatment and untreated control, which is ∆∆Ct = ∆Ct (drug treatment) − ∆Ct (control). Table 1. Primer sequences and amplicon lengths for antioxidant-and mitochondrial biogenesis-related genes.

Mitochondrial Mass
For mitochondrial mass measurement, cells were stained by 300 nM MitoTracker TM Green FM (Thermo Fisher Scientific) at the requirement (37 • C, 30 min) and analyzed by Accuri C6 flow cytometer (Becton-Dickinson) using FL1 channel as described [43].

Quantitative PCR (qPCR): mtDNA Copy Number
Total genomic DNA from cells incubated with POMx for 24 and 72 h was prepared according to the OMEGA Bio-Tek user manual of the E.Z.N.A. ® Tissue DNA kit (Norcross, GA, USA) [44]. Using the nuclear DNA (nDNA) gene GAPDH as a reference, the relative copy numbers of mtDNA such as NADH-ubiquinone oxidoreductase chain 1 (ND1) and ND5 genes [45] were analyzed using the 2 −∆∆Ct method [40] after qPCR reaction in a touch-down program [34]. The PCR information for the mtDNA copy number is listed in Table 2. SLR-qPCR was applied to assess mtDNA damages [46]. Using SLR-qPCR, the copy numbers of two DNA fragments with different lengths, i.e., small (ND1/ND5) and long (ND1-L/ND5-L) fragments, were measured for calculating mtDNA damage (lesions per 10 kb DNA between ND1 and ND5) by the formula: (1-2 −∆ (long Ct-short Ct) ) × 10,000 (bp)/length of the long fragment (bp) [46]. The primer and PCR amplicon information for mitochondrial DNA damage is provided in Table 2.

DNA Damage: γH2AX
The level of double-strand break marker for DNA damage (γH2AX) was analyzed [47]. p-Histone H2A.X (Ser 139) at 500X dilution was chosen as the primary antibody purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) to detect γH2AX at 4 • C for 1 h. Subsequently, a secondary antibody conjugated by Alexa 488 was used in its flow cytometry application (BD Accuri C6; FL1 channel).

Statistics
One-way ANOVA processed all the statistics after Tukey's HSD post hoc tests using JMP ® 12 software to compare different groups [49]. Treatments without overlapping low cases are regarded as significant differences.

DNA Damage: γH2AX
The level of double-strand break marker for DNA damage (γH2AX) was analyzed [47]. p-Histone H2A.X (Ser 139) at 500X dilution was chosen as the primary antibody purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) to detect γH2AX at 4 °C for 1 h. Subsequently, a secondary antibody conjugated by Alexa 488 was used in its flow cytometry application (BD Accuri C6; FL1 channel).

Statistics
One-way ANOVA processed all the statistics after Tukey's HSD post hoc tests using JMP ® 12 software to compare different groups [49]. Treatments without overlapping low cases are regarded as significant differences.

HPLC profile of POMx and Three Main Bioactive Components
The contents for punicalin, punicalagin, and ellagic acid of POMx-capsules were analyzed by HPLC using authentic reference compounds. The linear equations of three main compounds were y = 10 7 x − 69,990 (R 2 = 0.9997), y = 5 × 10 6 x − 47,955 (R 2 = 0.9998), and y = 7 × 10 6 x − 21,969 (R 2 = 0.9999), respectively. The results show that POMx contains punicalagin 26.582 mg/g, ellagic acid 47.857 mg/g, and punicalin 8.375 mg/g ( Figure 1). Figure 1. HPLC profile of POMx and the contents of its three main bioactive components. The HPLC profile for punicalagin, ellagic acid, and punicalin was provided as well as their contents within POMx (mg/g). STD means standard.

Antiproliferation of Oral Cancer Cells Following POMx Incubation
Cell viability detected by ATP assay in oral cancer cells after POMx (0, 50, 75, 100, and 125 µg/mL) treatment for 24 h is dose-responsively decreased (Figure 2A). IC 50  Cell viability detected by ATP assay in oral cancer cells after POMx (0, 50, 75, 100, and 125 μg/mL) treatment for 24 h is dose-responsively decreased (Figure 2A). IC50 value at 24 h ATP assay for POMx in oral cancer cells (Ca9-22, HSC-3, and OC-2) are 80.53, 100.34, and 108.12 μg/mL, respectively. Moreover, longer exposure to POMx for 72 h decreases more viability to oral cancer cells than that of the 24 h treatment. In contrast, normal oral cells (HGF-1) show only a mild decrease after 72 h exposure to POMx. Similarly, cell viability detected by trypan blue assay in oral cancer and normal oral cells after POMx (0 and 100 μg/mL) treatment for 0, 12, 24, and 72 h are time-dependently decreased ( Figure 2B). In addition, it was noted that cell viabilities for oral cancer cells (Ca9-22, HSC-3, and OC-2) are lower than that of normal oral cells. Figure 2C shows that 24 and 72 h POMx incubations of oral cancer cells induce abnormal cell morphology while normal oral cells (HGF-1) retain normal morphology. Accordingly, POMx has a selective killing effect on oral cancer cells but less harmful to normal oral cells. Similarly, cell viability detected by trypan blue assay in oral cancer and normal oral cells after POMx (0 and 100 µg/mL) treatment for 0, 12, 24, and 72 h are time-dependently decreased ( Figure 2B). In addition, it was noted that cell viabilities for oral cancer cells (Ca9-22, HSC-3, and OC-2) are lower than that of normal oral cells. Figure 2C shows that 24 and 72 h POMx incubations of oral cancer cells induce abnormal cell morphology while normal oral cells (HGF-1) retain normal morphology. Accordingly, POMx has a selective killing effect on oral cancer cells but less harmful to normal oral cells.

Cell Cycle Change of Oral Cancer Cells Following POMx Incubation
After POMx incubations (0, 50, and 100 µg/mL) for 24 and 72 h, the patterns for cell cycles in three oral cancer cell lines are shown ( Figure 3A). For 24 h POMx incubation, HSC-3 and OC-2 cells show slightly sub-G1 accumulations but not for Ca9-22 ( Figure 3B). For 100 µg/mL POMx incubation, all these three cell lines show a decrease in the G1 phase. HSC-3 and Ca9-22 cells show an increase to G2/M, while OC-2 cells show an increase to S phase and G2/M decrease.
For 72 h POMx incubation, HSC-3 cells show dramatic sub-G1 and S phase accumulations but show a decrease in G1 and G2/M phases ( Figure 3B). Ca9-22 and OC-2 cells show moderate subG1 and G2/M accumulation but show decreased G1 phase compared with the control.
Accordingly, POMx differentially disturbs cell cycle distribution of oral cancer cells between 24 and 72 h, and POMx at 72 h induces more subG1 accumulation (apoptosis-like) than at 24 h.
After POMx incubations (0, 50, and 100 μg/mL) for 24 and 72 h, the patterns for cell cycles in three oral cancer cell lines are shown ( Figure 3A). For 24 h POMx incubation, HSC-3 and OC-2 cells show slightly sub-G1 accumulations but not for Ca9-22 ( Figure 3B). For 100 μg/mL POMx incubation, all these three cell lines show a decrease in the G1 phase. HSC-3 and Ca9-22 cells show an increase to G2/M, while OC-2 cells show an increase to S phase and G2/M decrease.
In addition, POMx induces relatively more apoptosis in oral cancer cells than in normal oral cells. Moreover, apoptosis proteins such as cleaved PARP and BAX are increased, and the anti-apoptosis proteins such as Bcl-2 and Bcl-xL are decreased after 72 h POMx incubation ( Figure 4C).
Moreover, the AO-detected autophagy of three oral cancer cell lines is decreased by POMx during 12, 24, and 72 h incubations compared with the control, suggesting that POMx may inter-regulate apoptosis and autophagy. Accordingly, 72 h POMx incubation induces more apoptosis than 24 h for oral cancer cells. Moreover, POMx induces more apoptosis in oral cancer cells than in normal oral cells, especially for 100 µg/mL at 72 h.

MitoMP of Oral Cancer Cells Following POMx Incubation
After POMx incubation (0, 50, and 100 μg/mL) for 24 h, the patterns for MitoMP in oral cancer and normal (HGF-1) cell lines are shown ( Figure 5A). The MitoMP (−) (%) of these three oral cancer cells dose-responsively increase after POMx incubation ( Figure 5B). Moreover, POMx induces more MitoMP (−) (%) in three oral cancer cells than normal oral cells.  After time course treatments of POMx, the dynamics of flow cytometry patterns for MitoMP in these oral cancer cells are shown ( Figure 5C). The MitoMP (−) (%) of these three oral cancer cells is increased over time (12,24, and 72 h) after POMx incubation compared with the control ( Figure 5D). Moreover, POMx induces more MitoMP (−) (%) in three oral cancer cells than in normal oral cells throughout the time course. Accordingly, POMx causes higher MitoMP destruction in oral cancer cells than in normal oral cells.
After time course treatments of POMx, the flow cytometry patterns for MitoSOX in oral cancer and normal oral cells are shown ( Figure 6C). The MitoSOX (+) (%) of these three oral cancer cells was increased over time (12,24, and 72 h) after POMx incubation compared with the control, while it was unchanged in normal cells at 12 and 24 h and decreased at 72 h ( Figure 6D). Accordingly, POMx induced higher MitoSOX generation in oral cancer cells than normal oral cells.
causes higher MitoMP destruction in oral cancer cells than in normal oral cells.

MitoSOX Generation of Oral Cancer Cells Following POMx Incubation
After POMx incubations (0, 50, and 100 μg/mL) for 24 h, the patterns for MitoSOX in oral cancer (Ca9-22, HSC-3, and OC-2) and normal oral (HGF-1) cells are shown ( Figure  6A). The MitoSOX (+) (%) of these three oral cancer cells were dose-responsively increased after POMx incubation while remaining unchanged in normal cells ( Figure 6B). After time course treatments of POMx, the flow cytometry patterns for MitoSOX in oral cancer and normal oral cells are shown ( Figure 6C). The MitoSOX (+) (%) of these three oral cancer cells was increased over time (12,24, and 72 h) after POMx incubation compared with the control, while it was unchanged in normal cells at 12 and 24 h and decreased at 72 h ( Figure 6D). Accordingly, POMx induced higher MitoSOX generation in oral cancer cells than normal oral cells.

Mitochondrial Mass of Oral Cancer Cells Following POMx Incubation
After 24 h POMx incubation (0, 50, and 100 µg/mL), the patterns for Mitotracker in three oral cancer cell lines are shown ( Figure 8A). The Mitotracker (+) (%) of these three oral cancer cells were decreased after POMx incubation compared with the control ( Figure 8B).
After time course treatments of POMx, the flow cytometry patterns for Mitotracker in these oral cancer cells are shown ( Figure 8C). The Mitotracker (+) (%) of these three oral cancer cells are decreased at 12 and 24 h POMx incubation compared with the control, although it was slightly increased at 72 h ( Figure 8D).
were consistently expressed at 72 h POMx incubation.

Mitochondrial Mass of Oral Cancer Cells Following POMx Incubation
After 24 h POMx incubation (0, 50, and 100 μg/mL), the patterns for Mitotracker in three oral cancer cell lines are shown ( Figure 8A). The Mitotracker (+) (%) of these three oral cancer cells were decreased after POMx incubation compared with the control ( Figure  8B). After time course treatments of POMx, the flow cytometry patterns for Mitotracker in these oral cancer cells are shown ( Figure 8C). The Mitotracker (+) (%) of these three oral cancer cells are decreased at 12 and 24 h POMx incubation compared with the control, although it was slightly increased at 72 h ( Figure 8D).
After time course treatments of POMx, the mitochondrial resident protein (TIMM22 and TOMM20) expressions were detected in oral cancer cells ( Figure 8E). For 24 h POMx incubations, the TIMM22 and TOMM20 are almost unchanged in oral cancer cells. However, although TOMM20 remains unchanged, TIMM22 is upregulated at 72 h POMx incubations, consistent with Mitotracker detection. Accordingly, Mitotracker detections and After time course treatments of POMx, the mitochondrial resident protein (TIMM22 and TOMM20) expressions were detected in oral cancer cells ( Figure 8E). For 24 h POMx incubations, the TIMM22 and TOMM20 are almost unchanged in oral cancer cells. However, although TOMM20 remains unchanged, TIMM22 is upregulated at 72 h POMx incubations, consistent with Mitotracker detection. Accordingly, Mitotracker detections and protein expressions are differentially regulated at 24 h POMx incubation but in a consistently regulated manner at 72 h.

Mitochondrial DNA Copy Number, Lesion and Biogenesis of Oral Cancer Cells Following POMx Incubation
In addition to MitoMP, MitoSOX, and mitochondrial mass as described above (Figures 5-8), other mitochondrial functions such as mitochondrial DNA copy number, lesion, and biogenesis were further examined in POMx-incubated oral cancer cells (Figure 9). After POMx incubations (0, 50, and 100 µg/mL) for 24 and 72 h, the relative mtDNA copy numbers of three oral cancer cell lines were dose-responsively decreased ( Figure 9A). mtDNA damages between ND1 and ND5 genes were higher in oral cancer cells following 24 and 72 h POMx incubation than those of the control ( Figure 9B). Moreover, the mRNA expressions of all tested mitochondrial biogenesis genes (TFB2M, TFAM, POLRMT, and TUFM) were downregulated by 24 h POMx compared with the control (Figure 9C, left). Among these biogenesis genes, TUFM was dramatically downregulated by 72 h POMx ( Figure 9C, right). The protein expressions of these mitochondrial biogenesis genes were consistently downregulated at 24 and 72 h POMx ( Figure 9D). Therefore, POMx downregulates gene expressions for mitochondrial biogenesis in oral cancer cells.

Mitochondrial DNA Copy Number, Lesion and Biogenesis of Oral Cancer Cells Following POMx Incubation
In addition to MitoMP, MitoSOX, and mitochondrial mass as described above (Figures 5-8), other mitochondrial functions such as mitochondrial DNA copy number, lesion, and biogenesis were further examined in POMx-incubated oral cancer cells (Figure 9). After POMx incubations (0, 50, and 100 μg/mL) for 24 and 72 h, the relative mtDNA copy numbers of three oral cancer cell lines were dose-responsively decreased ( Figure 9A).

γH2AX-Detected DNA Damage of Oral Cancer Cells Following POMx Incubation
After POMx incubations (0, 50, and 100 µg/mL) for 0, 24, and 72 h, the patterns for γH2AX in three oral cancer cell lines were shown ( Figure 10A). The γH2AX (+) (%) of these three oral cancer cells was slightly increased at 24 h POMx incubation and dramatically increased at 72 h POMx incubation compared with the control ( Figure 10B).
( Figure 9C, left). Among these biogenesis genes, TUFM was dramatically downregulated by 72 h POMx ( Figure 9C, right). The protein expressions of these mitochondrial biogenesis genes were consistently downregulated at 24 and 72 h POMx ( Figure 9D). Therefore, POMx downregulates gene expressions for mitochondrial biogenesis in oral cancer cells.

Discussion
We found that POMx showed antiproliferation, apoptosis, oxidative stress, mitochondrial impairment, and DNA damage to several kinds of oral cancer cells. The detailed mechanisms for the POMx-induced antiproliferation are discussed in the following. Accordingly, POMx triggers γH2AX-detected DNA damage in oral cancer cells.

Discussion
We found that POMx showed antiproliferation, apoptosis, oxidative stress, mitochondrial impairment, and DNA damage to several kinds of oral cancer cells. The detailed mechanisms for the POMx-induced antiproliferation are discussed in the following.

POMx Has a Selective Antiproliferation Function towards Cancer Cells with Safety to Normal Cells
POMx provides antioxidant-rich natural products [51] and shows anticancer effects on several cancer cells [11,16,52]. This is partly explained by antioxidants having dual functions to reduce or induce oxidative stress at physiological or high concentrations [14]. The present study shows that the IC 50 values at the 24 h ATP assay for POMx incubated three oral cancer cell lines (Ca9-22, HSC-3, and OC-2) were 80.53, 100.34, and 108.12 µg/mL, respectively ( Figure 2). Similarly, IC 50 values at 72 h MTS assay for POMx for prostate (C4-2, PC3, and ARCaPM) [11] were 42, 78, and 161 µg/mL, respectively. Moreover, trypan blue assay in addition to ATP assay confirms viability results of POMx in oral cancer and normal cells.
The safety of POMx is well documented. For example, normal human prostatic epithelial PrEC cells showed no cytotoxicity (95% viability) to POMx [11]. At the 72 h MTT assay, pomegranate fruit extract (PFE) (50-150 µg/mL) showed antiproliferation against lung cancer cells with 53% viability but no cytotoxic effects on normal bronchial epithelial cells with 90% viability [15]. The pomegranate juice and oil showed antiproliferation and apoptosis in prostate cancer cells but no cytotoxicity in normal prostate epithelial cells [53]. Similarly, the normal oral cells (HGF-1) show higher viability in both ATP assay and trypan blue assay (Figure 2A,B) than the three oral cancer cell lines of this study. Punicalagin and ellagic acid, two main components of POMx, induce apoptosis of colon cancer cells without affecting normal colon cells [54]. Therefore, POMx and other pomegranate-derived natural products provided selective killing against several cancer cells and did not show side effects on normal oral cells.

POMx Inhibits Antioxidant Signaling to Generate Oxidative Stress
When the pro-oxidant level is higher than the antioxidant level, cellular oxidative stress is generated. In addition, mitochondrial impairment may change antioxidant gene expressions [55]. For example, advanced glycation end products were reported to inhibit the cellular antioxidant system and trigger oxidative stress [50].
Similar to the present study, the mRNA expressions for several antioxidant genes (NFE2L2, GCLC, TXN, CAT, SOD1, HMOX1, and NQO1) were downregulated at 24 h POMx for three oral cancer cell lines ( Figure 7A). Moreover, protein expressions for these antioxidant genes are downregulated at 72 h POMx treatment for oral cancer cells (HSC-3 and OC-2) but slightly upregulated for Ca9-22 cells ( Figure 7B). These results suggest that mRNA and protein expressions for antioxidant signaling may be differentially regulated between different oral cancer cell lines. Under these differential regulation modes, oxidative stress such as MitoMP depletion and MitoSOX generation were upregulated at 12, 24, and 72 h POMx in three oral cancer cells (Figures 5 and 6). Therefore, antioxidant pathways play a vital function in POMx induced oxidative stress in the present study.

POMx Induces Mitochondrial Impairment in Oral Cancer Cells
In addition to MitoMP and MitoSOX, mitochondrial mass, mtDNA copy number, mtDNA lesion, and mitochondrial biogenesis were also changed after POMx incubation in the present study. Similarly, Resveratrol may mitigate neurotoxicity following Rotenone treatment through promoting mitochondrial mass and DNA copy number [56]. Thus, there is a complex interaction between these mitochondrial functions.
Modulating mitochondrial function is associated with apoptosis. In view of mitochondrial mass change, several treatments may subsequently induce apoptosis. For example, TNFα decreases mitochondrial mass and induces apoptosis in human dermal microvascular endothelial cells (HMEC-1) [57]. In the present study, POMx shows similar results for oral cancer cells. The Mitotracker-detected mitochondrial mass is downregulated at 12 and 24 h POMx treatment but slightly upregulated at 72 h ( Figure 8C,D). Similarly, Western blotting shows that mitochondrial resident protein TIM22 is upregulated at 72 h POMx. Accordingly, the mitochondrial mass is dynamically changed over time after POMx treatment of oral cancer cells. The role of POMx-induced mitochondrial mass change warrants a detailed investigation in the future. mtDNA copy number change may regulate apoptosis. Increasing mtDNA copy number may inhibit apoptosis. In contrast, reducing mtDNA copy number was shown to induce ROS generation and apoptosis in tumor cells [58]. Similarly, 24 and 72 h POMx treatment increased oxidative stress and decreased mtDNA copy number ( Figure 9A) in oral cancer cells, leading to apoptosis.
Change of mitochondrial biogenesis change regulates apoptosis. Biogenesis may increase mitochondrial mass and DNA copy number [56] and is associated with apoptosis [63,64]. Similarly, 24 and 72 h POMx treatment inhibits mRNA and protein expressions for mitochondrial biogenesis of gene expression (TFB2M, TFAM, POLRMT, and TUFM) in oral cancer cells ( Figure 9C,D). This finding supports the notion that a decrease in mitochondrial biogenesis reduces the mitochondrial mass ( Figure 8). Moreover, mitochondrial fission factor (MFF) overexpression in breast cancer cells decreases both mitochondrial mass and activity [65]. Since POMx downregulates mitochondrial biogenesis ( Figure 9C,D) and mass (Figure 8), it is possible that POMx treatment causes mitochondrial fission and leads to apoptosis of oral cancer cells. It warrants a detailed investigation of the role of mitochondrial fission in POMx treatment for oral cancer cells in the future.

POMx Induces Apoptosis but Inhibits Autophagy in Oral Cancer Cells
POMx and pomegranate leaf extract (PLE) respectively induce apoptosis in human prostate [10] and lung [21] cancer cells. However, no caspase experiments were performed before the present study. Ethanol extracts of pomegranate fruit (PEE) induced apoptosis by cleaving Cas-3 and raising Bax/Bcl-2 ratio in urinary bladder cancer T24 cells [22]. Consistently, 72 h POMx induced apoptosis for oral cancer cells by the results of annexin V expression ( Figure 4A) and Western blotting ( Figure 4C).
The autophagy pathway is activated to guarantee the elimination of damaged mitochondria to maintain cell survival. In the case where autophagy is reduced, this may lead to cell death without the elimination of damaged mitochondria. This rationale is partly supported by our finding that the increase of apoptosis is accompanied by a decrease in AO-detected autophagy ranging from 12 to 72 h POMx treatment ( Figure 4E). These results warrant a detailed investigation of the impact of mitophagy or autophagy upon POMx treatment of oral cancer cells.

Conclusions
In the present study, the antiproliferation of POMx was evaluated using several types of oral cancer cells, and its detailed mechanisms related to mitochondrial function were explored. POMx treatment shows antiproliferation and apoptosis associated with downregulating antioxidant gene expression and triggering mitochondrial impairment, causing ATP depletion, MitoMP disruption, and MitoSOX generation as well as decreases in mitochondrial mass, mtDNA copy number, and mitochondrial biogenesis. Moreover, both nuclear and mitochondrial DNA damages were induced by POMx incubation in oral cancer cells. In conclusion, POMx provides antiproliferation and apoptosis effects on oral cancer cells through impaired mitochondrial functioning. Acknowledgments: The authors thank our colleague Dr. Hans-Uwe Dahms (https://biology.kmu. edu.tw/index.php/en-GB/faculty-members/92-hans-uwe-dahms) for editing the manuscript.

Conflicts of Interest:
The authors declare that there are no conflict of interest among them.