Antitumor Effect of Periplocin in TRAIL-Resistant Human Hepatocellular Carcinoma Cells through Downregulation of IAPs

Cortex periplocae is the dried root bark of Periploca sepium Bge., a traditional Chinese herb medicine. It contains high amounts of cardiac glycosides. Several cardiac glycosides have been reported to inhibit tumor growth or induce tumor cell apoptosis. We extracted and purified cortex periplocae and identified periplocin as the active ingredient that inhibited the growth of TNF-related apoptosis-inducing ligand-(TRAIL-) resistant hepatocellular carcinoma cells. The antitumor activity of periplocin was further increased by TRAIL cotreatment. Periplocin sensitized TRAIL-resistant HCC through the following two mechanisms. First, periplocin induced the expression of DR4 and FADD. Second, the cotreatment of TRAIL and periplocin suppressed several inhibitors of apoptosis (IAPs). Both mechanisms resulted in the activation of caspase 3, 8, and 9 and led to cell apoptosis. In addition, intraperitoneal injection (IP) of periplocin repressed the growth of hepatocellular carcinoma (HCC) in xenograft tumor model in mice. In summary, periplocin sensitized TRAIL-resistant HCC cells to TRAIL treatment and resulted in tumor cell apoptosis and the repression of tumor growth in vivo.


Introduction
Liver cancer is the �h leading malignancy in men and the ninth in women worldwide [1]. e incidence of liver cancer is highly correlated to chronic local in�ammation and cirrhosis. erefore, factors that stimulate in�ammation in liver, including alcohol uses, infection of hepatitis B and C viruses, and fatty liver diseases, are strongly correlated to the pathological progression of liver cancer.
Current treatments for liver cancer are limited. Hepatectomy can be used in early stage liver cancer patients with functional liver. Liver transplantation can help patients with damaged livers, but matching suitable donors is not easy. Although these surgical operations work well in early stage diseases, they are not helpful for patients with cancer cells spread out of the liver. Chemotherapy and internal radiation therapy are also options for liver cancer treatment. However, they may do damage to other tissues and organs as well. Targeted therapy is a more speci�c treatment for cancer. Sorafenib, a multireceptor kinase inhibitor with antiangiogenic activity, is the standard treatment for advanced hepatocellular carcinoma (HCC) that cannot be removed with surgery. Although it extends the median overall survival in patients with advanced HCC for nearly 3 months, sorafenib does not increase the median time to symptomatic progression in patients with advanced HCC [2]. erefore, novel treatments for HCC are strongly in need.
With the advancement of techniques in extraction, isolation, and recognition of compounds from plants, scientists started to search for antitumor components from herb medicine [3][4][5][6]. Cortex periplocae is the dried root bark of Periploca sepium Bge. It contains several cardiac glycosides and can be used in the treatment of various heart conditions. Recent studies also suggest that periplocin, a cardiac glycoside extracted from cortex periplocae, can inhibit cell growth in colon cancer cells and lung cancer cells [7,8].
TNF-related apoptosis-inducing ligand (TRAIL) is a member of the tumor necrosis factor superfamily. It is also known as CD253 and APO-2L. TRAIL binds to the death 2 Evidence-Based Complementary and Alternative Medicine receptors DR4 and DR5 and induces cell apoptosis [9][10][11]. erefore, TRAIL is a potential candidate for cancer treatment [12]. In addition, phases 1 and 2 clinical trials for speci�c monoclonal antibodies against DR4 and DR5 have provided promising results [13].
Although TRAIL is a promising chemotherapeutic target for cancers, resistance to TRAIL-induced apoptosis has been reported in several different cancers, including colorectal cancer, breast cancer, liver cancer, and pancreatic cancer [14][15][16][17]. Several different mechanisms are proposed for TRAIL resistance [18]. Ways to overcome TRAIL resistance are still under investigation [19,20]. We sought to investigate the effect of periplocin in sensitizing TRAIL-resistant HCC cell lines in this study.

Western Blot.
Total cellular lysates were prepared by using RIPA lysis buffer. Proteins in cell lysates (50 g) were separated on 4-12% SDS-polyacrylamide minigels and electrotransferred to a PVDF membrane by iBlot Dry Blotting System (Invitrogen, Carlsbad, CA, USA). Antibodies used in this study were as follows: anti-Caspase 3, 8, and 9 antibodies

Viability Assay (MTT).
Cells were seeded at 10 4 cells in 100 uL medium per well in a 96-well plate and incubated (37 ∘ C, 5% CO 2 ) overnight. Drugs of interest were added to each well and incubated (37 ∘ C, 5% CO 2 ) for 2 days. MTT solution (5 mg/mL, Sigma Chemical Co., St. Louis, MO, USA) was added to each well at a �nal concentration of 0.5 mg/mL and incubated (37 ∘ C, 5% CO 2 ) for 1-2 hours. Aerward, 100 uL of 10% SDS (Fluka, St. Louis, MO, USA) was added to each well and incubated at room temperature overnight.
2.5. Determination of ROS Production. ROS production was monitored by �ow cytometry using DCFH-DA. is dye is a stable compound that readily diffuses into cells and is hydrolyzed by intracellular esterase to yield DCFH, which is trapped within cells. us, the �uorescence intensity is detected to quantify the amount of peroxide produced by the cells. To investigate the effect of periplocin and the combination of periplocin and TRAIL on generating intracellular ROS in HA22T/VGH, cells were pretreated with N-acetyl-cysteine (NAC) (30 mM) for 30 min and followed by periplocin (0.3, 0.03 M) alone or together with TRAIL (100 ng/mL). DCHF-DA (100 uM) were added to periplocin-treated cells with or without H 2 O 2 (200 uM) for 1-2 hr. Green �uorescence was excited using an argon laser by �ow cytometric analysis [21].

2.�. �uanti�cation of A�o�tosis by Anne�in V�P�.
Aer treated with periplocin alone or together with TRAIL for 24 hours, HA22T/VGH cells were washed and resuspended in the staining buffer and examined with the Vybrant Apoptosis Assay Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. e cell suspension was incubated with 2.5 L of Annexin V and 1 L of propidium iodide at room temperature for 15 min. e stained cells were analyzed by �uorescence activated cell sorter (FACS) analyses with a FACSCalibur �ow cytometer, and data were analyzed using CellQuest soware (BD Biosciences, San Jose, CA, USA).

2.�. �uanti�cation of Apoptosis by Sub-G1 Pea�.
Trypsinized cells were washed with ice-cold PBS and �xed in 70% ethanol at −20 ∘ C for at least 1 h. Aer �xation, cells were washed twice, incubated in 0.5 mL of 0.5% Triton X-100/PBS at 37 ∘ C for 30 min with 1 mg/mL of RNase A, and stained with 0.5 mL of 50 mg/mL propidium iodide for 10 min. e �uorescence emitted from the propidium-DNA complex was quantitated by FACSCalibur �ow cytometer (BD Biosciences, San Jose, CA, USA).

FACS Analysis.
Cells were incubated with dye-labeled monoclonal antibodies (mAb) against target molecules for 30 min on ice. Stained cells were then washed twice and resuspended in cold buffer and analyzed with a FACScan �ow cytometry (BD Biosciences, San Jose, CA, USA). More than 1 × 10 5 cells were analyzed for each sample, and the results were processed by using WinMDI 2.8 soware (Scripps Research Institute, La Jolla, CA, USA).

In Vivo Efficacy Study.
All experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC number: ITRI-IACUC-2012-010M, Industrial Technology Research Institute of Taiwan, HsinChu, Taiwan. SCID (CB17/Icr-Prkdcscid/CrlBltw) mice were purchased from BioLASCO Ltd. (Ilan, Taiwan). Huh-7 cells (3 × 10 6 cells per mice) in 100 L mix (equal volumes of PBS and Matrigel) were implanted subcutaneously (sc.) into the right �ank of female SCID mice (6-8 weeks old). Tumor sizes were measured with calipers, and tumor volumes (mm 3 ) were calculated with the following formula: 2 /2 (where is the longest diameter and is the shortest diameter).

Histology and Immunohistochemistry.
At the end of the study, mice were sacri�ced, and tumor samples were collected, �xed in formalin, and embedded in para�n as tissue sections. Tissue sections were stained with hematoxylin and eosin (H and E) for general tissue morphology evaluation. e antihuman Ki67 antibody (1 : 500 dilution, IS-626, Dako, Glostrup Denmark) and antihuman cyclin-D1 antibody (1 : 500 dilution, IS-626, Dako, Glostrup, Denmark) were used in immunohistochemistry staining. Staining procedure was completed by using Autostainer Link 48 system (Dako, Glostrup Denmark). Five �elds of every tumor sample were randomly selected, and the percentage of the Ki67-positive and cyclin-D1-positive cells was calculated to evaluate the proliferation of tumor samples.

Periplocin as the Active Ingredient in Cortex Periplocae in Inhibiting the Growth of Hepatocellular Carcinoma Cells.
Cortex periplocae (CP) is a traditional medicine capable of inhibiting cancer cell growth. To identify the active ingredients in CP that are responsible for its activity in inhibiting the growth of hepatocellular carcinoma (HCC) cells, we isolated and puri�ed CP. Aer several rounds of puri�cation, we found a group of compounds named CP-1 to 6 as major components in the fraction that can actively inhibit tumor cell growth (Figure 1(a)).
To further identify the active ingredients in the fraction, the pure compounds in the active fraction was examined one by one, and periplocin (CP-1) was identi�ed as the most potent compound in inhibiting tumor cell growth with IC 50 at 0.027 M (Figure 1(b)). e cotreatment of TRAIL and periplocin or periplogenin (CP-5) strongly enhances the growth inhibiting activity of periplocin and periplogenin (Figure 1(b)). Interestingly, periplocin is less toxic to normal cells. e cell viability of PBMC was more than 80% when treated with 300 g/mL periplocin (data not shown).

e Combination Treatments of Periplocin and TRAIL-Induced Apoptosis in TRAIL-Resistant HCC Cells.
Although TRAIL is a promising anticancer drug, more and more TRAIL-resistant cancers were reported. We sought to determine if periplocin can sensitize TRAIL-resistant HCC cells to TRAIL treatment. As shown in Figure 2(a), TRAIL or periplocin alone had little effect on the viability of HCC cells, but the combination of these two drugs showed cytotoxicity to TRAIL-resistant HCC cells.
To study if the combination treatments of periplocin and TRAIL sensitize TRAIL-resistant HCC cells and induce HCC apoptosis, HA22T/VGH cells were stained with Annexin V and PI to characterize cells in early and late stages of apoptotic processes accordingly. As shown in Figure 2(b), periplocin treatment increased the ratio of Annexin V and PI positive HCC cells. Cotreatment of TRAIL and periplocin further increased the ratio of Annexin V and PI positive cells. erefore, TRAIL and periplocin synergistically induced cell apoptosis in HCC cells.
Moreover, the accumulation of cell debris aer apoptosis was demonstrated by sub-G1 population in cell cycle analysis. Consistent with our previous results, periplocin dose dependently increased sub-G1 population in HCC cells, while the addition of TRAIL further increased the sub-G1 population in HCC cells (Figure 2(c)).  and data not shown) TRAIL induces cell apoptosis through interaction with death receptors DR4 and DR5 signaling. It was reported that compounds which upregulate DR4 and DR5 in HCC cells could sensitize TRAIL-resistant HCC cells to TRAIL treatment [22]. To investigate if periplocin sensitizes TRAIL-resistant cells through the same mechanism, the expression levels of DR4 and DR5 in HA22T/VGH cells with or without periplocin treatments were detected. Periplocin increased DR4 expression and further induced FADD expression in HA22T/VGH cells 8 hours aer treatment. However, periplocin did not induce DR5 expression in HA22T/VGH cells (Figure 3(a)). e results were veri�ed by FACS analysis (data not shown).

Periplocin and TRAIL Cotreatment Induces Apoptosis in
TRAIL binds DR4 and activates FADD. Activated FADD induces the cleavage of several proapoptotic proteins and activates them. We examined the activation of several apoptosis-related proteins, including the cleavage of BID, caspase 8, caspase 3, and PARP in HA22T/VGH. As expected, periplocin or TRAIL treatment alone had little effects on   the cleavage of BID, caspase 8, caspase 3, and PARP. e combination of periplocin and TRAIL treatments strongly increase the cleavage of all these apoptosis-related proteins (Figure 3(b)). Similar results were observed in Huh7, another HCC cell line (Figure 3(c)).
To further con�rm the importance of caspase activation in periplocin-regulated cell apoptosis with or without TRAIL treatment, caspase inhibitors were added to HA22T/VGH cells prior to periplocin and/or TRAIL treatments. Inhibitors against caspase 3, caspase 8, and caspase 9 partially rescued cell survival repressed by periplocin and/or TRAIL treatments, and pan inhibitor against all three caspases completely blocked cell death induced by periplocin and/or TRAIL treatments (Figure 3(d)).

Periplocin and TRAIL Cotreatment Induces Apoptosis in HA22T/VGH Cells by Inhibiting IAPs.
Another reported mechanism for sensitizing TRAIL-resistant cell lines to TRAIL treatment is through regulating proteins involved in apoptotic pathways [23]. Since the cleavage of caspases can be induced either through intrinsic (mitochondrial mediated) or extrinsic (death receptor mediated) apoptotic pathways, the role of mitochondrial-mediated caspase cleavage was �rst investigated in HA22T/VGH with periplocin and TRAIL treatments. e expression levels of Bcl-2 protein family and several mitochondria dependent apoptotic regulators, including Bax, Bad, Mcl-1, and apaf-1, were detected before and aer periplocin and/or TRAIL treatments. However, the treatments of periplocin and/or TRAIL did not affect the expression of apoptotic regulators (Figure 4(a) and supplemental Figure 2A). Nevertheless, the combination treatment of periplocin and TRAIL activated caspase 9 in two different HCC cell lines (Figures 3(c) and 4(a) and supplemental Figure 2 ). Since the combination treatment of periplocin and TRAIL activated several caspases, we examined the effect of periplocin and TRAIL on members of inhibitors of apoptosis (IAP) family. e expression of several IAP family members, including cIAP-1, XIAP, and survivin, was repressed by the combination treatment of periplocin and TRAIL in HA22T/VGH cells (Figure 4(b) and supplemental Figure 2 ).

Periplocin Represses Tumor Formation In Vivo.
In addition to in vitro mechanistic studies, we also veri�ed the potency of TRAIL and periplocin on repressing tumor growth in vivo. Since TRAIL was expressed in NK cells in mice, we treated tumor-bearing mice with only periplocin in this in vivo study [24,25]. To test the antitumor activity of periplocin in vivo, HCC cells were subcutaneously injected into SCID mice, and periplocin was intraperitoneally (IP) injected daily two weeks aer the initial injection of tumor cells. Since the mice were well-tolerant to 5 mg/kg periplocin aer 2 weeks of treatment, the dose of periplocin were raised to 20 mg/kg daily throughout the study. Periplocin was able to inhibit HCC growth in xenogra model. Tumor growth inhibition (TGI, %) of periplocin treatment was 51 ± 11% aer 24 days treatment, (Figure 5(a)). When treated with periplocin, the mice body weight was slightly decreased at the dose of 20 mg/kg when compared to the vehicle group. However, the body weight kept at around 90 percent of control group, and no further body weight loss was observed ( Figure 5(b)). ese data showed that periplocin strongly inhibited tumor growth of Huh-7 tumors without obvious side effects.
In Ki67 immunochemistry analysis, periplocin could inhibit the Ki67 expression in tumor samples ( Figure 5(c)). Aer selecting �ve random �elds and calculating the Ki67 positive cells of every tumor sample, the percentage of Ki67-positive cells in vehicle group was 57.3 ± 0.67%. e percentage of Ki67-positive cells in periplocin-treated group was 22.78 ± 10.09%. e result showed that periplocin could signi�cantly inhibit the Ki67 expression in Huh-7 tumors, which suggested that periplocin inhibited tumor growth in vivo.
To further verify the periplocin-inhibited tumor growth in vivo, we also examined the expression of cyclin-D1 in Huh-7 tumors. In cyclin-D1 immunochemistry analysis, the percentage of cyclin-D1-positive cells in vehicle group was 76.87 ± 2.93%. e percentage of cyclin-D1-positive cells in periplocin-treated group was 58.85 ± 5.05%. e result showed that periplocin could inhibit the cyclin-D1 expression in tumor samples ( Figure 5(c)). To verify the role of periplocin in cyclin-D1 expression, Huh-7 cells were seeded on Millipore Millicell EZ slide (2 × 10e4/well) for 24 h and subsequently treated with different concentrations of periplocin for 24 h. As expected, periplocin dosedependently repressed cyclin-D1 expression in Huh-7 cells ( Figure 5(d)).

Discussion
Periplocin is a cardiac glycoside structurally similar to digoxin. e pharmacological function of periplocin is also similar to digoxin and has been used to treat heart diseases. Nevertheless, digoxin was shown to block cancer growth through inhibiting HIF-1 signaling pathway in cancer cells [26]. Indeed, Digitalis was used as treatment for breast cancer patients and reduced cancer recurrence rate [27,28]. Compounds structurally similar to digoxin also possess antitumor activity [29]. Several studies suggest that cardiac glycoside induces cell apoptosis. Digoxin induces apoptosis through activating Cdk5 [30]. In addition, Digitalis was reported to induce mitochondria-dependent apoptotic pathways in guinea-pig cardiomyocytes [31].
TRAIL induces cell apoptosis via DR4 and DR5, activates Fas-associated death domain (FADD) and caspase 8, and signals through both mitochondria-dependent and -independent pathways [32]. ere are three antagonistic decoy receptors DcR1, DcR2, and osteoprotegerin that interact with TRAIL but cannot transmit apoptotic signal. erefore, cancer cells could gain TRAIL-resistance by overexpressing DcR1, DcR2, or osteoprotegerin [33][34][35]. In addition, defects in adaptor protein FADD or caspase 8 can also lead to TRAIL resistance [36,37]. In this study, we demonstrated that periplocin induced DR4 and FADD expression in TRAIL-resistant HCC cells. e combination treatment of TRAIL and periplocin activated caspase 8, the key caspase for both mitochondria-dependent andindependent apoptotic signaling pathways, in TRAIL resistant HCC cells.
e combination treatment of periplocin and TRAIL induced HCC cell apoptosis through activating IAP. IAPs are members of a protein family that regulate apoptosis. Currently there are 8 known members in this family, and X-linked inhibitor of apoptosis protein (XIAP) is the best characterized member in the family. XIAP blocks apoptosis by binding and inactivating caspase 3, 7, and 9 [38]. Other IAP, including c-IAPl, c-IAP2, and survivin, have been shown to bind to caspase 3 and 7 [39]. Indeed, we observed the repression of XIAP, c-IAP1, and survivin, and the deactivation of caspase 3 and 9 by TRAIL and periplocin treatments in this study. e expression of TRAIL by NK cells was demonstrated both in vitro and in vivo [24,25,40,41]. We took advantage of that and designed an experiment that combined endogenous TRAIL and exogenous periplocin treatment. e injected periplocin was able to repress HCC tumor growth in vivo.
While purifying cortex periplocae, we identi�ed six compounds with potential antitumor activity ( Figure 1). Among these compounds, periplocin (CP-1) and periplogenin (CP-5) are the two compounds with cytotoxicity against cancer cells, and periplocin is more potent than periplogenin. Structurally, periplocin is different from periplogenin by only one disaccharide residue. It was reported that saccharide residues are involved in the recognition of plant root surfaces by zoospores [42]. erefore, it is not surprising that saccharide residues can affect protein interactions. However, detailed mechanisms of how periplocin interacts with cell surface molecules and how does the disaccharide residue regulate cell apoptosis require further investigation.
In traditional herb medicine, patients are usually treated with multiple raw materials. Sometimes it is hard to purify one single active ingredient since multiple components are required to achieve therapeutic goals. In this study, we puri-�ed an active ingredient, periplocin, with cytotoxicity against HCC cells from cortex periplocae. Interestingly, maximum cytotoxicity against TRAIL-resistant HCC was achieved by combining the treatment of periplocin and TRAIL. e idea of the combination treatment is consistent with the concept of traditional medicine. Provided the mechanistic studies, the combination treatments of active ingredients from herb medicine and chemical synthesized compounds or protein drugs could be potential treatment options for drug-resistant cancers.

Conclusion
In this study, we demonstrated that periplocin could sensitize TRAIL-resistant HCC to TRAIL treatment, and the combination treatment of TRAIL and periplocin can induce apoptosis in TRAIL-resistant HCC. Furthermore, we showed that periplocin sensitized TRAIL-resistant HCC cell lines to TRAIL through the following two mechanisms. First, periplocin induced the expression of DR4 and FADD to activate proapoptotic signaling pathways. Second, the cotreatment of TRAIL and periplocin suppressed several IAP, which also led to the activation of proapoptotic signaling pathways. Our working model is shown in Figure 6. Further studies are required to apply periplocin clinically.

Con�ic� of �n�eres�s
e authors claim no con�ict of interests.