|Home | About | Journals | Submit | Contact Us | Français|
Limited success has been achieved in extending the survival of patients with metastatic and hormone-refractory prostate cancer (HRPC). There is a strong need for novel agents in the treatment and prevention of HRPC. We have shown that flavokawain B (FKB), a kava chalcone, is about 4 to 12 fold more effective in reducing the cell viabilities of androgen receptor (AR)-negative, HRPC cell lines DU145 and PC-3 than AR–positive, hormone-sensitive prostate cancer cell lines LAPC4 and LNCaP, with minimal effect on normal prostatic epithelial and stromal cells. FKB induces apoptosis with an associated increased expression of proapoptotic proteins: death receptor-5, Bim and Puma, and a decreased expression of inhibitors of apoptosis protein: XIAP and survivin. Among them, Bim expression was significantly induced by FKB as early as four hours of the treatment. Knockdown of Bim expression by short-hairpin RNAs attenuates the inhibitory effect on anchorage-dependent and-independent growth and caspase cleavages induced by FKB. These findings suggesting the effect of FKB, at least in part, requires Bim expression. In addition, FKB synergize with TRAIL for markedly enhanced induction of apoptosis. Furthermore, FKB treatment of mice bearing DU145 xenograft tumors results in tumor growth inhibition and increases Bim expression in tumor tissues. Together, these results suggest robust mechanisms for FKB induction of apoptosis preferentially for HRPC and the potential usefulness of FKB for prevention and treatment of HRPC in an adjuvant setting.
A significant portion of prostate cancer (PCa) patients are curable either by surgery or by radiotherapy when detected early (1, 2), and many patients require no active therapy but are managed by expectant management. For some patients progression and metastasis require androgen-deprivation therapies (ADT), including orchiectomy or administration of leutinizing hormone-releasing hormone agonists/antagonists (1, 2). The majority of these PCa is initially androgen-sensitive and responds to ADT (1, 2). However, ADT rarely cures patients, but will eventually result in a resistant phenotype to androgen blockade with more aggressive disease (1, 2), which represents a major clinical challenge (1, 2). Recently, docetaxel-based chemotherapy has shown minor survival benefit and has become the remaining treatment option for hormone-refractory PCa (HRPC) (3). As effective treatment for HRPC is still not available, new agents that are particularly designed for the prevention and treatment of HRPC are highly desired.
Multiple mechanisms have been reported to account for progression of PCa to a hormone-refractory stage (4–12). In the majority of these hormone refractory cases, AR signaling is aberrantly activated (4). The reported mechanisms for the aberrant AR activation include AR amplification, mutation, splicing and/or co-regulator interactions, as well as through peptide growth factors and ligands for G-protein-coupled receptors (4–6). But there are also reports about extensive loss of AR expression through its promoter hypermethylation in about 20–30% of HRPC. In addition, HRPC has been associated with neuroendocrine differentiation, with some neuroendocrine differentiated PCa cells being AR-negative (9). Furthermore, PCa initiating cells may consist of a very small subpopulation of AR-negative, stem/progenitor cells. These PCa initiating cells are resistant to current anti-hormonal therapy, radiotherapy, and/or chemotherapies and therefore contribute to the recurrence of PCa (10). Taken together, we can argue that both AR-negative and -positive PCa cells should be targeted in order to develop more effective treatment and prevention approaches for HRPC.
Many phytochemicals have shown promising anticancer results with little or no toxicity to normal cells (13). In addition, most of these phytochemicals are constituents of the human diet or are taken as dietary supplements (13–17). Therefore, some of these phytochemicals may have the potential for supplementation of traditional main therapies for treatment and prevention of HRPC. Kava (Piper methysticum) is an ancient crop of the western Pacific. The root extract of kava has been part of the Pacific Islanders’ culture for thousands of years, serving as a beverage, medicine and in soci-religious functions similar to wine in Western cultures (18). Consumption of traditional aqueous kava preparation correlates with low and uncustomary sex ratios (more cancer in women and men) of cancer incidences in three kava-drinking countries: Fiji, Vanuatu, and Western Samoa (19). We recently demonstrated that flavokawains from kava extracts are strong apoptosis inducers against bladder cancer cells and that flavokawain A partially requires Bax activation for its apoptotic effect (20). However, the mechanism for Bax activation induced by flavokawains and the potential role of flavokawains as anti-cancer agents in PCa is not known. In addition, there is very little reported data about agents that can particularly target HRPC cell lines.
In this study, we demonstrate for the first time that flavokawain B (FKB) is significantly more effective in inhibiting the growth of AR-negative, HRPC cell lines (DU145 and PC-3) than AR-positive, hormone-sensitive cell lines (LNCaP and LAPC4), with a minimal effect on the growth of primary prostatic epithelial and stromal cells derived from normal prostates. In addition, we show that FKB up-regulated expression of up-stream Bax activators, including Bim, Puma and DR5; and down-regulated expression of survivin and XIAP for induction of apoptosis. Furthermore, FKB demonstrated in vivo anti-tumor efficacy and induced Bim expression in tumor tissues.
The LNCaP, LAPC4, DU145, and PC-3 cell lines were obtained from ATCC and cultured in RPMI 1640 medium with 10%FBS. Normal prostate epithelial (PrECs) and stromal cells (PrSCs) were obtained from Clonetics Inc. and maintained in PrEBM and SCBM medium, respectively (Cambrex). Pure flavokawain A and B (99%) were from LKT Laboratories, Inc. Flavokawain C was purchased from INDOFINE Chemical Company, Inc.. Flavokawain A, B and C were dissolved in DMSO, aliquoted, and stored at −80°C. Antibodies for DR5, DR3, DR4, XIAP and survivin were from Cell Signaling Technology, Inc. Antibodies against BAX, Bcl2, Bclx/l and beta-actin were from Santa Cruz Biotechnology, Inc. Bim antibody was purchased from Calbiochem. Inc. Anti-Bax 6A7 antibody, which recognizes only the open form of Bax, was from Sigma. Thymidine, 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT), cycloheximide and propidium iodide were from Sigma. RNAazol B was purchased from Tel-Test. The Reverse Transcription System kit was from Promega.
The MTT assays were performed as previously described (20). Dose–response curves for growth inhibition were generated as a percentage of vehicle-treated control. The combination index (CI) was calculated to determine whether the flavokawains interact synergistically, additively, or antagonistically (21). The nature of the interaction between flavokawain A, B and C was analyzed using the median effect/combination index analysis described by Chou and Talaly (21). Mixtures of flavokawain A and B, flavokawain A and C, and flavokawain B and C were made so that two compounds were equipotent. Three sets of experiments for each compound combination were carried out. The CI was calculated as follows: CIAB = ICA,B/ICA + ICB,A/ICB+ICA,B*ICB,A/ICA*ICB, or CIAC = ICA,C/ICA+ICC,A/ICC+ ICA,C*ICC,A/ICA*ICc, or CIBC = ICC,B/ICc + ICB,C/ICB+ICC,B* ICB,C/ICC*ICB. ICA, ICB and ICc are the concentrations of flavokawain A, B, and C, respectively, needed to produce a given level of growth inhibition when used alone, whereas ICA,B, ICB,A, ICA,C, ICC,A, ICC,B, and ICB,C are the concentration needed to produce the same effect when used in combination. The concentration needed to produce a given level of growth inhibition (% effect) was determined by non-linear least square regression (GraphPad PRISM). The CI values were computed automatically to indicate the degree of synergy or antagonism.
Cells were treated with 0.1% DMSO, 8.8 μM FKB, 100 ng/ml TRAIL, or 100ng/mol TRAIL plus 8.8 μM FKB for 24 hours. After stated treatments, cells were stained with FITC-conjugated Annexin V and propidium iodide in PBS according to the manufacturer’s protocol (PharMingen). All analyses of cells were done using appropriate scatter gates to exclude cellular debris and aggregated cells. Ten thousand events were collected for each sample stained with Annexin V.
The Cell Death Detection ELISA kit (Roche) was used to detect apoptosis. Cells seeded in six-well plates were treated with 0.1% DMSO or FKB (4.4, 8.8 and 16.7 μmol/L) for 24 hours. The cells were lysed and centrifuged. Then, the supernatant was transferred into anti-histone-coated microtiter plate and incubated with anti-DNA peroxidase antibody. After removal of the unbound antibodies, the nucleosomes were quantified by the peroxidase reaction using 2,2′-azino-di(3-ethylbenzthiazolin-sulfonat) as substrate. A microtiter plate reader at 492 nm read the color intensity.
Apoptosis was confirmed by using the Caspase-Glo® 3/7, Caspase-Glo® 8 and Caspase-Glo® 9 Assay (Promega) according to the manufacturer’s instructions. Cells were plated in a 24-well plate and treated with 0.1% DMSO or FKB (4.4, 8.8 and 16.7 μmol/L) for 24 hours. Then 100 μL of Caspase-Glo® 3/7, Caspase-Glo® 8 or Caspase-Glo® 9 reagent was added on to each well and the luminescence of each sample was measured in a luminometer (GloMax ®-MultiDetection System).
Total RNA was isolated from PCa cells using the RNAazol B method as described (22). Real-time quantitative PCR amplification reactions for DR5 mRNA levels were carried out using MyiQ system (Bio-Rad) as described previously (22) and the sequences of primer sets are available upon request. Data were analyzed by using the comparative Ct method, where Ct is the cycle number at which fluorescence first exceeds the threshold. The Ct values from each sample were obtained by subtracting the values for β-actin Ct from the DR5 Ct value. The variation of β-actin Ct values is <0.5 among different samples. One difference of Ct value represents a 2-fold difference in the level of mRNA. Specificity of resulting PCR products was confirmed by melting curves.
Clarified protein lysates (20–80 μg) were denatured and resolved by 8–16% SDS-PAGE. Proteins were transferred to nitrocellulose membranes, and probed with antibodies and visualized by an enhanced chemiluminescence detection system.
Total protein (200μg) was precleared with protein G plus-agarose and then precipitated with 2 μg anti-Bim or IgG antibody overnight at 4°C. Agarose beads were washed four times with lysis buffer and resuspended in SDS-PAGE 2x sample buffer. Proteins were eluted by boiling the beads and subjected to immunoblotting of active Bax protein.
Two different Bim targeting sequences were designed using the RNAi Designer program (Invitrogen), and then subcloned into a pENTR/U6 vector according to the Gateway protocol. After that, these RNAi cassettes were transferred into a pBLOCK-iT 3-DEST vector using Clonase™ to catalyze a recombination reaction (Invitrogen). Negative controls used shRNA targeting sequences complementary to the LacZ gene without any match to human genome sequences. These shRNAs can effectively knock down transfected LacZ genes. All constructs were transfected using Lipofectamine 2000 (Invitrogen) and selected with G418. The pooled stable transfectants were examined.
FKB was formulated in 10% grain alcohol (is pure alcohol made from fermented grains and used as a solvent in commercial kava root extracts) in 0.9% saline and given by gavage. NCR-nu/nu (nude) mice were obtained from Taconic. DU145 cells were concentrated to 1× 106 per 100 μL PBS and injected s.c. into the right flank of each mouse. After seven days tumors were approximately the size of 120 mm3, the mice bearing DU145 tumors were randomly divided and pair matched into treatment and control groups of 10 mice each. Daily dosing was begun with vehicle or 50 mg/kg FKB. Because there were no in vivo data regarding FKB before this study, the dose of FKB (50 mg/kg/d) was used according to a pharmacologically effective and non-toxic dose of flavokawain A (a close analogue of FKB in kava extracts) (20). Body weight, diet, and water consumption were recorded thrice weekly throughout the study. Once xenografts started growing, their sizes were measured every three day. The tumor volume was calculated by the formula: 0.5236 L1(L2)2, where L1 is the long axis and L2 is the short axis of the tumor. At the end of the experiment, tumors were excised and weighed, blood was collected and all were stored at −80°C until additional analysis.
Comparisons of caspase activities, OD values for cell death and cell viabilities between treatment and control were conducted using Student’s t test. For tumor growth experiments, repeated-measures ANOVA was used to examine the differences in tumor sizes among treatments, time points, and treatment-time interactions. Additional post-tests were done to examine the differences in tumor sizes between control and FKB treatment at each time point by using conservative Bonferroni method. All statistical tests were two sided.
Flavokawain A, B and C constitute about 0.46%, 0.015% and 0.012% of kava extracts, respectively (23). To determine which flavokawain(s) or their combination(s) could be the most potent agent against HRPC, we first examined the effect of individual flavokawains and the 1:1/31:1/38 mixture of flavokawain A:B:C (The ratio of flavokawains in the mixture is similar to that naturally occurred in kava extracts) on the growth of an AR-negative DU145 cells. Figure 1A shows that FKB is the most potent growth-inhibitory agent among the individual flavokawains and the flavokawain mixture. Figure 1B shows that combination indexes for different flavokawain combinations on cell growth inhibition are more than 1.0, suggesting there is no synergistic effect.
We subsequently examined the effect of FKB on the growth of different PCa cell lines and prostatic cells derived from a normal prostate by primary culturing. Figure 1C shows that FKB at a concentration of 17.6 μM inhibits the growth of AR-negative, hormone refractory PCa cell lines (DU145 and PC3) by about 90%, and partially reduces the growth of AR-positive, hormone sensitive PCa cell lines (LNCaP and LAPC4) by about 32–50% (Students’ t test, Ps<0.01). At the same concentration, FKB has minimal effect on the growth of prostatic epithelial and fibroblastic cells from a normal prostate (less than 6%, Ps>0.05). The IC50 of FKB treatment for 48 hours on different PCa cell lines are estimated to be 32 (LAPC4), 48.3 (LNCaP), 6.2 (PC-3) and 3.9 μM (DU145). AR-negative cell lines (PC-3 and DU145) are approximately 4 to 12 fold more sensitive to FKB’s effect than AR-positive cell lines (LAPC4 and LNCaP). This selective killing effect of FKB on AR-negative, hormone-refractory PCa cell lines suggest a potentially novel strategy for prevention and treatment of HRPC.
Figure 1D & E shows that treatment of AR negative DU145 and PC-3 cells grown in 10% FBS with FKB results in a highly significant to complete inhibition of their growth in a time-dependent manner. An inhibitory effect of FKB was evident at 1 day, but a more profound effect was observed during 3–5 days of treatment. The 1.1- and 2.2-μM concentrations of FKB shows 34 and 80% inhibition in growth of DU145 cells and 12 and 64% inhibition in growth of PC-3 cells, respectively. Cells treated with 8.8 and 17.6 μM of FKB shows complete growth inhibition, respectively (Figure 1D &E). At these concentrations of FKB, cells stopped growing as early as 3 days, with a small reduction in initial cell density (Figure 1D & E).
To determine whether the growth inhibitory effect of FKB is through the induction of apoptosis, we examined the apoptotic morphology of control- and FKB-treated cells under light- and fluorescence microscopes. Figure 2A&B shows typical apoptotic morphologies in FKB-treated cells treated but not with control- treatment. The apoptotic morphologies include cell shrinkage and rounding up, cell membrane blebbing, as well as nuclear fragmentation and condensation. In addition, FKB resulted in a significant increase in both early (Annexin V staining only, right-lower panels) and late apoptosis (Annexin V and PI staining, right-upper panels) populations when compared to control treatment (Figure 2C, 41.7 ± 4% and 23.6 ± 2% apoptotic cells in FKB treated DU145 and PC-3 cells, respectively, vs. 4.5 ± 0.3% and 1.07 ± 0.2% in control treated DU145 and PC-3 cells; Students’ t test, P<0.01). Cell death ELISA experiments further confirmed that FKB causes cell death in a dose-dependent manner (data not shown). In addition, FKB causes PARP cleavage in both DU145 and PC3 cells (data not shown). Together, these results provide a firm conclusion that FKB induces apoptosis in DU145 and PC3 cells.
Apoptosis can be induced by the extrinsic pathway associated with death receptor stimulation on the cell surface (25), and by the intrinsic pathway characterized by the involvement of mitochondrial dysfunction (26). While caspase 3 is an effecter caspase, initiator caspase 8 and 9 are activated by death receptors and mitochondrial releasing factors, respectively (24). Figure 2D shows that FKB increases caspase 8, 9 and 3 by about 30 to 143% compared to vehicle control treatment, indicating that both death receptor- and mitochondrial- mediated apoptotic pathways are activated (24).
Figure 3A&B shows that FKB specifically increases the protein and mRNA expression of DR5, without affecting the expression of DR4 and DR3 (data not shown). Additionally, PC-3 cells treated with 8.8 μM FKB and 100 ng/ml TRAIL reduces cell viabilities by 46 % and 22 %, respectively, while combination of both agents leads to a marked decrease of cell viability by 76% (Figure 3C, Students’ t test, Ps<0.01). Similarly, Annexin V staining shows that either 8.8μM FKB or 100 ng/ml TRAIL treatment alone results in about 22% or 23% of PC-3 cells undergoing apoptosis, whereas combination of both agents increase the percentage of apoptotic cells to 49% (Figure 3D, Students’ t test, P<0.01). Together, these results suggest that FKB potentiates the apoptotic effect of TRAIL ligand via induction of DR5.
The BH3-only proteins, including Bad, Bim/Bod, Bid, Bmf, Noxa, and Puma, are immediate upstream triggers for Bax activation (26). We investigated which BH3-only proteins are responsible for FKB-induced activation of Bax. FKB treatment significantly increases the protein expression of Bim and Puma (Figure 4, panel a and b) without affecting the expression of other BH3-only proteins, including Bad, Bid and Noxa, in DU145 and PC3 cells (data not shown). Consistently, the increase in Bim and Puma by FKB treatment is associated with an increase in active Bax (Figure 4, panel c). In addition, FKB decreases the expression of inhibitors of apoptotic proteins: XIAP and survivin (Figure 4, panel d and e). Furthermore, FKB treatment causes an increased formation of a Bax/Bim immunocomplex in DU145 cells. Together, these results indicate that FKB activates the mitochondrial-mediated apoptotic pathway by changing the balance between proapoptotic and antiapoptotic proteins.
Figure 4C shows that FKB treatment of DU145 cells exhibits a time-dependent effect on the expression of apoptotic and inhibitors of apoptotic proteins. Bim and Puma expression were induced as early as 4 hours of FKB treatment. At 8 hours of FKB treatment, a decrease in XIAP and survivin were observed. The induction of DR5 and active Bax were only shown at 16 hours of FKB treatment. These results suggest that Bim is an early and up-stream event for FKB activation of apoptosis in DU145 cells.
We next examined whether Bim is, at least in part, required for the growth inhibitory and apoptotic effect of FKB. Figure 5A shows that Bim protein expression was inhibited up to 75–95% by stable transfection of Bim ShRNA plasmids compared with a control LacZ ShRNA transfection. The control-transfected DU145 and PC3 cells did not show any difference in cell growth compared with their parental cells (data not shown). Figure 5B shows that cells with stable suppression of Bim expression are more resistant to the growth inhibitory effects of FKB than those transfected with LacZ ShRNA (Student’s t test, Ps < 0.05). Knockdown of Bim expression was also shown to attenuate the inhibitory effects of FKB on colony formation in soft agar (Student’s t test, Ps < 0.05; Figure 5C and Supplementary Figure 1). In addition, the degree to which FKB attenuated cell viability was associated with the level of Bim expression (Figure 5B&C). Furthermore, in control-transfected PC3 cells, 8.8 μM FKB treatment results in about 180 to 240% increase of cell death and caspase 3/7/9 activities compared to 0.1% DMSO treatment, respectively (Student’s t test, Ps < 0.05; Figure 5D), whereas in Bim ShRNA-transfected PC3 cell lines, FKB at the same concentration does not cause significant cell death nor caspase 3/7/9 activations (Student’s t test, Ps >0.05; Figure 5D). This result indicates the FKB-induced cell death and caspase 3/7/9 activation in PC-3 cells is blocked by Bim suppression. Together, these results provide strong evidence that Bim is a critical target for the apoptotic and growth inhibitory effects of FKB.
Finally, Figure 6A demonstrates the effects of oral administration of 50 mg/kg FKB daily for 24 days on mice bearing established DU145 tumors resulted in a significant decrease in the growth rate of tumors compared with control group (P < 0.05, ANOVA). The wet tumor weights in control- and FKB–treated group recorded at the end of the treatment are 465 ± 49 and 199 ± 25 mg, respectively (Figure 6B; mean ± SE; P = 0.045, Student’s t test). FKB treatment reduced tumor growth by 67%. The body weight gain and diet and water consumption of the FKB–treated mice were similar to the control group of mice (Supplementary figure S2). In addition, the mice did not show any gross abnormalities on necropsy at the end of the treatment. Immunohistochemical analysis shows that there is a strong increase of Bim expression in tumor tissues in the FKB group compared to the control group. Consistently, western blotting analysis of tumor lysates revealed higher levels of Bim protein in the FKB-treated tumor tissues than those treated by control. Together, FKB demonstrated in vivo antitumor activity, which was accompanied by increased expression of Bim.
Because of lower cell proliferation rates in clinical PCa compared to most other cancers (27), induction of apoptosis could be a more effective mechanism for elimination of PCa cells (28–31). We demonstrated that FKB has robust mechanisms in induction of apoptosis via targeting both death receptors and the balance of pro- and anti-apototic proteins. FKB increases DR5 expression, leading to activation of the death receptor mediated apoptotic pathway. The TRAIL, a DR5 ligand, is considered an effective anticancer agent, as it selectively induces apoptosis in a variety of tumor cells, yet is relatively nontoxic to normal cells (25). We further demonstrated that FKB exerts a synergistic apoptotic effect when combined with TRAIL. The mechanism for FKB mediated DR5 expression remains unclear. In addition to up-regulation of DR5 mRNA expression, FKB was found to increase the mRNA expression of Puma and p21/WAF1 in PC3 cells (data not shown). DR5, Puma and P21/WAF1 are common p53 target genes (32). Because the PC3 cell line harbors a p53 gene deletion, it is unlikely that FKB activates the p53 for its effect on DR5 and Puma expression. The apoptotic effect of FKB is also not dependent on PTEN as there is no significant difference for the apoptotic effect observed in PTEN-wild type DU145 vs -mutant PC3 cells. In addition, FKB did not change the PTEN expression in DU145 cells (data not shown). Instead, we recently reported that flavokawain A regulated the expression of DR5 and Puma mRNA via increased expression of p73, a p53 family member, in bladder cancer cell lines (33). Further experiments are therefore in progress to determine whether FKB can stabilize p73 protein in p53 mutant PCa cells to explain its effect on gene transcription of DR5 and Puma. Altogether, these data suggest that combination of TRAIL and FKB may represent a novel therapeutic strategy against HRPC.
Bim directly initiates the BAX-mediated mitochondrial apoptosis via binding to stabilized α-helix of BCL-2 domains of BAX (34). We have shown that Bim forms an immunocomplex with active Bax protein, which presumably allows permeabilization of mitochondria with subsequent release of cytochrome C and others for activation of the caspase 9/3 cascade. In addition, knockdown of Bim was shown to markedly attenuate anchorage-dependent and -independent cell growth inhibitory effects and caspase apoptotic effects of FKB.
Bim does play a role in tumorigenesis. In mice, the loss of Bim can accelerate c-myc oncogene-driven lymphomagenesis (35, 36). In patients with mantle cell lymphoma and renal cell carcinoma, the loss or reduced expression of Bim has also been found (37, 38). Studies suggest Bim to be an important convergent point for both Akt and ERK regulating apoptotic signaling (39–42). Combination of an mTOR inhibitor rapamycin and an ERK inhibitor PD0325901 resulted in substantially enhanced antitumor effects particularly for hormone-refractory prostate tumors in the mouse model and led to increased expression of Bim in tumor tissues (43). We also demonstrated that FKB significantly induces Bim expression in tumor tissues while exhibiting antitumor effects. Together, these results suggest a role for Bim expression in antitumor mechanisms, and the potential usefulness of Bim protein as a surrogate biomarker for monitoring the apoptotic and anti-tumor effects of FKB in future studies.
In summary, we have demonstrated the strong inhibitory effect of FKB on the growth of PCa cell lines with more potency to AR negative, HRPC cell lines. FKB induces apoptosis via robust mechanisms; including (i) increases the expression of DR5 leading to activation of the death receptor pathway and (ii) upregulates Bim and Puma expression and down-regulates XIAP and survivin expression resulting in activation of the BAX-initiated mitochondria pathway. Additionally shown, FKB reduces tumor growth in vivo and induces the expression of Bim in tumor tissues. It appears that the growth inhibitory effect of FKB, at least in part, requires Bim expression. Based on these findings, we propose that FKB, either its derivatives or in combination with TRAIL may be an effective treatment modality for those PCa patients with hormone-refractory disease.
This work was supported by AICR grant 41493, NIH grants CA129793 and CA122558 (to X. Zi).