Purified vitexin compound 1, a new neolignan isolated compound, promotes PUMA‐dependent apoptosis in colorectal cancer

Abstract Purified vitexin compound 1 (VB1, a neolignan isolated and extracted from the seed of Chinese herb Vitex negundo) is an effective antitumor agent and exhibits promising clinical activity against various cancers including colorectal cancer. However, it remains unknown about the precise underlying mechanism associated with the antitumor effect of VB1 and how it triggers apoptosis in cancer cells. Here, we demonstrated that VB1 promoted apoptosis via p53‐dependent induction of p53 upregulated modulator of apoptosis (PUMA) and further to induce Bax (Bcl‐2‐associated X protein) activation and mitochondrial dysfunction in colon cancer HCT‐116 and LoVo cells. Deficiency in p53, PUMA, or Bax abrogated VB1‐induced apoptosis and promoted cell survival in HCT‐116 cells. Furthermore, the combination of VB1 with chemotherapeutic drugs 5‐fluorouracil (5‐FU) or NVP‐BZE235 resulted in a synergistic antitumor effect via PUMA induction in HCT‐116 cells. VB1 significantly suppressed the cell proliferation of wild‐type (WT) HCT‐116 and LoVo cells in vitro and tumor growth in vivo. The results indicate that p53/PUMA/Bax axis plays a critical role in VB1‐induced apoptosis and VB1 may have valuable clinical applications in cancer therapy as a novel anticancer agent used alone or in combination with other chemotherapeutic drugs.


| INTRODUCTION
Colorectal cancer is the third most frequent malignancy and the third cause of cancer-related deaths. 1 It affected approximately 95 270 people and resulted in 49 190 deaths in the United States in 2016. 2 Although the treatment of colorectal cancer has made substantial improvement in the past decades, patients with metastatic colon cancer have only about 3 years of median overall survival, 3,4 because most patients eventually develop acquired drug resistance. [5][6][7][8] Drug resistance and doselimiting toxicity limit the efficacy of clinical response to chemotherapy. 9,10 Therefore, there is an urgent need to develop novel chemical entities that have better anticancer activity, less toxicity, or complement current chemotherapeutic drugs to improve their anticancer efficacy against colorectal cancer.
In this study, we assessed the effects of VB1 on cell proliferation inhibition and apoptosis induction in human colorectal cancer in vitro and in vivo. We also studied the expression, function, and molecular mechanism of PUMA during VB1induced apoptosis in both wild-type (WT) and p53, PUMA, and Bax knockout (KO) cells. Furthermore, we investigated the combinational therapy of VB1 plus 5-fluorouracil (5-FU) or VB1 plus NVP-BZE235 in colon cancer cells, and demonstrated the indispensability of PUMA/Bax axis in these processes. Finally, the antitumor activity of VB1 was analyzed in the xenograft model in vivo, and showed a positive therapeutic effect. For the first time, our results revealed that VB1 suppressed colorectal cancer growth via p53/PUMA/Bax signaling pathway, which makes VB1 as a potential antitumor drug in clinical application in the future.

| Transfection studies
The plasmids of expressing PUMA and GFP-Bax were kindly provided by Dr. Jian Yu. 14 HCT-116 cells were cultured in a 6-well plate at a density of 10 6 /well. Cells were transfected with pGFP-Bax or pPUMA following Polyjet ™ transfection manufacturer's instructions. The medium was replaced with fresh culture medium after 8 hours. One day post-transfection, cells were treated with medium (control) or 10 μmol/L VB1 for 24 hours followed by related experiments.

| Xenograft mouse model
Five-week-old SPF Grade female nu/nu mice (Vital River Lab Animal Technology Co. Ltd., Beijing, China, Certificate No. SYXK2013-0001) were housed in sterile microisolator cages (five per cage) with free access to water and food ad libitum. All animal experiments were carried out followed the protocols approved by Central South University Animal Use and Care Committee (Changsha, Hunan, China). 1 × 10 6 cells were injected s.c. into both flanks of mice. Mice were administered by i.p. injection of VB1 40 mg/kg every other day for 2 weeks when tumors were measurable, whereas the same volumes of normal saline (NS) were used as vehicle control. Mice were euthanized when tumors reached ~1.0 cm 3 (1000 mg) in size. Tissues of tumors were collected and examined. The protein was extracted using a Total

| Analysis of cell viability and apoptosis
Cells were cultured in 96-well microplate at a density of 5 × 10 3 cells/well for 24 hours. Cell viability was assessed with Cell Counting Kit-8 (CCK-8) (7Sea Biotech, Shanghai, China) at indicated time post-treatment following the manufacturer's instructions. The absorbance value at 450 nm (OD450) was read with a 96-well plate reader (DG5032, Hua Dong, Nanjing, China), to determine the cell viability.
For colony formation assay, cells were cultured in 6-well plate at a density of 5 × 10 4 cells/well for 24 hours. The cells were then treated with indicated concentrations of drugs and medium (control) for 24 hours. Medium was changed every 2 days. Colonies were visualized with crystal violet staining at Day 14.
For analysis of apoptosis by nuclear staining, cells were cultured in a 3.5-cm dish, rinsed with phosphate-buffered saline (PBS) twice and then 500 μL DMEM containing 5 μg Hoechst 33342 was added into the plates and incubated for 15 minutes in an incubator. Apoptosis was assessed through microscopic visualization of condensed chromatin and micronucleation. Apoptosis indices were calculated as the percentage of apoptotic cells among one hundred cells in a randomly selected portion. The positive rate of apoptotic cells was calculated by GD-10.0 image analysis system.

| Flow cytometry
HCT-116 and LoVo cells were suspended in 1 × 10 6 cells/ mL, and 5 μL of Annexin V and propidium iodide staining solution were added to 300 μL of the cell suspension. After the cells were incubated at room temperature for 15 min in the dark, stained cells were assayed and quantified using a FACSort Flow Cytometer (Beckman Coulter, Brea, CA, USA). Cell debris was excluded from the analysis by an appropriate forward light scatter threshold setting. Compensation was used wherever necessary.

| Co-immunoprecipitation
HCT-116 cells were cultured in 10-cm dish at a density of 8 × 10 6 cells/dish for 24 hours. Cells were then treated with 10 μmol/L VB1 for 24 hours, and the same volume of medium was used as control. Cells were harvested and lysed with lysis buffer (25 mmol/L HEPES, 125 mmol/L K-acetate, 2.5 mmol/L-acetate, 2 mmol/L DTT, 0.4% Tx-100, 2X Phosphatase Inhibitor, Protease Inhibitor, Na Orthovanadate 400 μmol/L, pH = 7.2). To detect the interaction between PUMA and Bax, anti-PUMA antibodies (~4 μL) were firstly added to 400 μL cell lysates and mixed on a rocker at 4°C for 4 hours. The immunocomplexes were captured by the addition of protein G/A-agarose (Roche Applied Sciences) mixed at 1:10 ratio, followed by an additional 1 hour incubation. The beads were washed three times by PBS and then collected by centrifugation at 300 g for 5 minutes. After the final wash, the beads were mixed with 60 μL of 2× Laemmli sample buffer, heated at 100°C for 5 minutes, and analyzed by Western blotting.

| GFP-Bax translocation assay
HCT-116 cells were cultured in a 3.5-cm dish at a density of 10 6 /dish for 24 hours and then transfected with plasmid GFP-Bax (conc. 902.4 ng/μL) following Polyjet ™ transfection protocol (please see Transfection method section). One day post-transfection (please see the 2.2 Transfection Studies), cells were treated with medium (control) and 10 μmol/L VB1 for 24 hours. After mitochondria were stained with MitoTracker ® Red CMXRos (Shanghai Yesen), the distribution pattern of GFP-Bax with or without VB1 treatment was imaged under a Olympus FV1000 confocal microscope. Bax redistribution was examined by matching fluorescence emitted by GFP-Bax (green) and MitoTracker (red). The cells, which exhibited intense punctate staining of GFP overlapping with the distribution of MitoTraker, were considered for mitochondrially localized Bax.

| Statistical analysis
Statistical analysis was carried out using GraphPad Prism V software (GraphPad Software Inc., La Jolla, CA, USA). P value was calculated via either Student's t test or one-way analysis of variance (ANOVA). Statistical significance was defined as P-value <0.05 (marked as *). Figure 1A) significantly inhibited the proliferation of HCT-116 and LoVo cells in a concentration-and timedependent manner ( Figure 1B), and the results were further confirmed by colony formation assay ( Figure 1C). Next, the effect of VB1 on cell apoptosis was evaluated by Western blotting. The cleaved caspase 3, a critical apoptotic marker, was markedly increased after the cells were treated with 2.5, 5, 10, or 20 μM of VB1, respectively ( Figure 1D). Chromatin condensation was obviously shown in HCT-116 and LoVo cells at 24 hours after VB1 treatment ( Figure 1E). FACS analysis also showed that cell apoptosis increased remarkably in VB1-treated group compared with control group in HCT-116 cells (P < 0.05) ( Figure 1F,G).

PUMA expression
We first examined the effect of VB1 on the major members of Bcl-2 family at the protein level by Western Blotting. VB1 concentration-dependently increased the expression of PUMA and slightly increased the expression of Bim. However, VB1 did not affect the expressions of other Bcl-2 family members such as Bcl-xl, Mcl-1, Bcl-2, and Bax ( Figure 2A). PUMA induction was dose-and timedependent, and VB1 could induce it at the concentration as low as 2.5 μmol/L and as early as 3 hours (Figure 2B,C). The results were also confirmed in LoVo cells ( Figure 2D,E). We then further examined the transcription factors that transactivate PUMA, such as FoxO3a, p73, p65, and p53 to investigate their role in VB1-induced PUMA expression. The results showed VB1 did not change the expression of p-FoxO3a, FoxO3a, p65 S276, p65 S536, p65 and p73 ( Figure 2F), and no translocation of p65 or FoxO3a occurred by immunofluorescence analysis after VB1 treatment ( Figure S2). The data indicate that they play little or no role in PUMA induction. Both HCT-116 and LoVo cells contain WT p53, we found that VB1 treatment resulted in p53 upregulation in these cells ( Figure 2B-E). Furthermore, the effect of VB1 on PUMA induction was almost completely blocked in p53-KO cells, compared to that of the parental cells ( Figure 2G,H). These results imply that PUMA induction by VB1 is mediated by p53-dependent mechanism.

| P53/PUMA pathway played a vital role in VB1-induced cell apoptosis
To examine the activation of p53/PUMA pathway in VB1-induced apoptosis, parallel studies have been performed in PUMA-KO and p53-KO HCT-116 cells. The results showed p53 expression but not cleaved caspase 3 was upregulated by VB1 in PUMA-KO HCT-116 cells ( Figure 3A), indicating that PUMA contributed to VB1induced apoptosis but did not affect expression of p53, and VB1 increased expressions of PUMA and cleaved caspase 3 in the WT but not in the p53-KO HCT-116 cells ( Figures 3A and S3a), implying that p53 is indispensable for VB1-induced PUMA upregulation and apoptosis. The chromatin condensation of p53-KO and PUMA-KO HCT-116 cells was significantly lower than that of WT HCT-116 cells detected by Hoechst 33342 staining after VB1 treatment. (Figure 3B,C). Furthermore, apoptotic resistance to VB1 was also confirmed by Annexin V/propidium iodide staining in p53-KO and PUMA-KO cells ( Figure  S3b,c). The cell viability and colony formation were also significantly increased in p53-KO and PUMA-KO HCT-116 cells after VB1 treatment ( Figure 3D-F). Therefore, these results suggest that p53/PUMA pathway is necessary for VB1-induced apoptosis in colon cancer HCT-116 cells. We and others reported before that PUMA-induced cell apoptosis was associated with activation of Bax directly and indirectly. 29,30 Here, we further investigated the relationship between PUMA and Bax in VB1 treated HCT-116 cells. As shown in Figure 4A, there was a direct interaction between PUMA and Bax and both of them were significantly elevated by VB1. To clarify the role of Bax activation in VB1-induced apoptosis, the expressions of p53 and PUMA as well as cleaved caspase 3 were investigated in the WT and Bax-KO HCT-116 cells following VB1 treatment. As shown in Figure 4B, the expressions of p53 and PUMA increased in both WT and Bax-KO cells by VB1 ( Figure 4B). However, cleaved caspase 3 was increased in WT HCT-116 cells only whereas remained unchanged in Bax-KO HCT-116 cells after VB1 treatment ( Figure 4B). Furthermore, chromatin condensation was significantly inhibited in Bax-KO HCT-116 cells ( Figure 4C). Consistent results were also obtained by cell viability ( Figure 4D) and colony formation assays ( Figure 4E). Finally, we observed VB1 induced the mitochondrial translocation of GFP-Bax in HCT-116 cells through confocal microscopy ( Figure 4F) to further confirm Bax activation by VB1. Therefore, the data indicate that Bax is an essential initiator for cell apoptosis induced by VB1.

BZE235 to induce PUMA-dependent apoptosis
After we demonstrated the pro-apoptotic role of PUMA and tumor growth inhibition by VB1, we further studied the effect of VB1 in combination with other chemotherapeutic drugs on apoptosis. We selected a DNA-damaging drug 5-FU and a novel dual PI3K/mTOR inhibitor NVP-BZE235 for induction of PUMA expression and apoptosis in HCT-116 cells (Figure S4a-d). The results showed that VB1 combined with 5-FU or NVP-BZE235 led to higher expression of PUMA and cleaved caspase 3, compared to mono treatment ( Figures 5A,B and S4e). In addition, combination therapy induced higher levels of apoptosis in WT HCT-116 cells, but not in PUMA-KO cells ( Figure 5C,D). Moreover, the apoptotic effect of combination was completely restored by addition of exogenous PUMA ( Figure 5E,F), implying PUMA is essential for the effect of combination treatment-induced apoptosis, which is consistent with our previous study. 45 5-FU greatly increased phospho-Akt but NVP-BZE235 reduced Akt phosphorylation ( Figure  S4f). However, phospho-Akt was greatly reduced with the treatment of VB1 combined with 5-FU or NVP-BZE235 ( Figure 5G,H). These data demonstrate that PUMA mediates the effect of chemosensitization of VB1 and robust induction of PUMA can enhance apoptosis, thereby improve the therapeutic efficacy.

| VB1 inhibited colon cancer growth in vivo
Since VB1 could inhibit cell growth and promote PUMAmediated apoptosis in vitro, we further evaluated its in vivo inhibitory effect on tumor growth in vivo of HCT-116 bearing nude mice. As shown in Figure 6A, VB1 significantly suppressed tumor growth compared to the control group after i.p. injection of 40 mg/kg VB1 every other day for 2 weeks. The tumor sizes in VB1-treated mice were roughly 1/5 of that in vehicle-treated mice on day 13 ( Figure 6B). The weights and volumes of tumors in VB1-treated group were 75% less than that of tumors in the control group ( Figure 6C,D). PUMA and p53 were significantly induced following VB1 treatment in HCT-116 tumors ( Figure 6E). These data show significantly antitumor activity of VB1 in vivo against HCT-116 tumor xenografts and are consistent with the results achieved in vitro.

| DISCUSSION
Lignans are a group of complex polyphenolic antioxidants in plants and have various pharmacological effects including antioxidant, antimicrobial, and anti-inflammatory. 46 An accumulation of evidence has demonstrated that lignans arise as one of the most promising dietary agents for cancer prevention and therapy. 40,47,48 Previous studies have demonstrated that VB1 acts as a novel antitumor agent for tumor chemoprevention and chemotherapy by targeting cell cycle arrest, antiangiogenesis and apoptosis induction in various cancers both in vitro and in vivo. 41,43,49,50 Although it has been reported that VB1 induces apoptosis through inhibition of multiple protein kinases and signal transduction pathways, 42,43 the underlying mechanisms of how VB1 induced apoptosis in colon cancer cells are still not fully clarified. Here, we report the anticancer efficacy of VB1 and associated mechanism in colorectal cancer cells. We demonstrated the anticancer efficacy of VB1 against HCT-116 and LoVo cells with greater anticancer effect than that of 5-FU (P < 0.05), a commonly used chemotherapeutic agent in clinic. We are the first to show a role of the p53/PUMA/Bax pathway in VB1-induced apoptosis in colon cancer WT HCT-116 cells. The expression of p53-dependent PUMA directly affects the activation of Bax and the mitochondrial apoptotic pathway.
PUMA plays an important role in mediating apoptosis. In this study, we found that VB1 significantly induced apoptosis and suppressed cell growth in WT HCT-116 and LoVo cells but not in PUMA-KO HCT-116 cells, indicating VB1 induces apoptosis via PUMA-dependent pathway. Further studies prompted that PUMA promoted VB1 induced Bax activation and the effect of VB1 on apoptosis was significantly attenuated in Bax-KO HCT-116 cells. Bim induction was also slightly increased by VB1 in HCT-116 cells by Western blotting analysis. However, the expressions of Mcl-1, Bcl-xl, and Bcl-2 remained unchanged in HCT-116 cells. As Bax is almost undetectable in LoVo cells, we speculated that whether Bak (another Bcl-2 family member like Bax) will play a critical role in mitochondrial outer membrane permeabilization (MOMP) following PUMA induction. As a result, we found that VB1 treatment led to dramatic Bak oligomerization, which is consistent with our hypothesis that after upregulated by p53, PUMA may activate Bak and cause its conformational change in LoVo cells, subsequently, Bak forms oligomer channels in the mitochondrial membrane for cytochrome c release, caspase activation and apoptosis. Further study may be needed to clarify whether other Bcl-2 family proteins and the inhibitor of apoptosis family of proteins (IAP) such as survivin are involved in cell apoptosis induced by VB1.
Previous studies have shown that NF-κB pathway 22 and p73 were involved in PUMA activation. 15,16 However, the phosphorylation level of p65 and the expression of p73 remained unchanged after VB1 treatment. Furthermore, it is well established that FoxO3a can bind to p53-binding sites in the PUMA promoter, [17][18][19] our data indicated that FoxO3a was not activated by VB1 in HCT-116 cells. In addition, the expressions of PUMA and cleaved caspase 3 as well as apoptosis induction by VB1 were significantly reduced in p53-KO HCT-116 cells compared to that of WT HCT-116 cells, suggesting that PUMA induction is through p53-dependent pathway. However, we also found that VB1 could inhibit cell growth and induce cell apoptosis in mutate p53 colon cancer cells (data not shown). Therefore, other cell-killing mechanisms of VB1 need further investigation.
Acquired drug resistance is the main reason for poor clinical efficacy of chemotherapy drugs, and the combination of different anticancer mechanisms can effectively reduce the occurrence of drug resistance. Thus, it is urgent to identify and develop effective combination regimens in cancer therapy to prevent and overcome drug resistance. Here, we used the combination of VB1 with chemotherapeutic drugs 5-FU or NVP-BZE235. Consistent with our previous studies, 5-FU induced apoptosis via a p53-dependent pathway, whereas NVP-BZE235 induced apoptosis via a p53-independent pathway. 45 Our results showed that VB1 increased the expression of p53-dependent PUMA in HCT-116 cells. The result may explain the synergistic effect of 5-FU and VB1 on apoptotic induction due to upregulation of PUMA through p53dependent pathway. More importantly, VB1 could further suppress Akt activation when combined with 5-FU to completely inhibit Akt activation induced by 5-FU (Figures 5 and  S4). NVP-BZE235 has been reported to upregulate PUMA in HCT-116 cells in a p53-independent pathway. 45 The combination of VB1 and NVP-BZE235 greatly induced the expression of PUMA via both p53-dependent and independent pathways. However, whether Akt inhibition contributes to PUMA induction remains unclear and needs to be further investigated.
Taken together, our studies demonstrated the anticancer effects of VB1 against HCT-116 tumor xenografts in vivo and investigated its mechanisms in colon cancer HCT-116 and LoVo cells. We also confirmed the indispensable role of p53/PUMA/Bax axis in apoptotic induction and therapeutic efficacy of VB1. Loss of p53, PUMA, or Bax may lead to cancer cells resistance to VB1 in cancer therapy. Therefore, we provide a novel insight for the anticancer mechanism of VB1 and scientific rationale for further development of VB1 as a novel anticancer drug used alone or plus on other chemotherapeutic drugs in the therapy of colorectal cancer clinically.