|Home | About | Journals | Submit | Contact Us | Français|
Despite state of the art cancer diagnostics and therapies offered in clinic, prostate cancer (PCa) remains the second leading cause of cancer-related deaths. Hence, more robust therapeutic/preventive regimes are required to combat this lethal disease. In the current study, we have tested the efficacy of Andrographolide (AG), a bioactive diterpenoid isolated from Andrographis paniculata, against PCa. This natural agent selectively affects PCa cell viability in a dose and time-dependent manner, without affecting primary prostate epithelial cells. Furthermore, AG showed differential effect on cell cycle phases in LNCaP, C4-2b and PC3 cells compared to retinoblastoma protein (RB−/−) and CDKN2A lacking DU-145 cells. G2/M transition was blocked in LNCaP, C4-2b and PC3 after AG treatment whereas DU-145 cells failed to transit G1/S phase. This difference was primarily due to differential activation of cell cycle regulators in these cell lines. Levels of cyclin A2 after AG treatment increased in all PCa cells line. Cyclin B1 levels increased in LNCaP and PC3, decreased in C4-2b and showed no difference in DU-145 cells after AG treatment. AG decreased cyclin E2 levels only in PC3 and DU-145 cells. It also altered Rb, H3, Wee1 and CDC2 phosphorylation in PCa cells. Intriguingly, AG reduced cell viability and the ability of PCa cells to migrate via modulating CXCL11 and CXCR3 and CXCR7 expression. The significant impact of AG on cellular and molecular processes involved in PCa progression suggests its potential use as a therapeutic and/or preventive agent for PCa.
Elucidating the mechanisms and developing therapies against prostate cancer (PCa) has been a long-standing interest for many researchers. Currently offered chemotherapeutic regime for hormone refractory PCa are often associated with hepatotoxicity and renal failure. Natural compounds are gaining popularity in the ‘war on cancer’ over conventional chemotherapies due to dearth of effective therapies without side effects. Compounds obtained from microbial and plant sources can be less toxic and more beneficial than conventional anti-cancer agents. As a result, many of these agents are in clinical trials; though certain limitations are associated with them.1,2 A major limitation of natural compounds is their poor bioavailability. However, Andrographolide (AG), a diterpenoid extracted from Andrographis paniculata, has an advantage over other natural agents due to its better absorption and hepatoprotective effects. Independent pharmacokinetic studies showed that 90% of orally administered AG is absorbed in the blood and 40% in tissues and cells.3,4
AG functions mainly by manipulating cell cycle and cytokine signaling. Studies have shown that it inhibits NF-kB and reduces pro-inflammatory cytokines IL-2, IL-6, TNF-α and IFN-γ in PCa.5,6 Role of these cytokines on host and cancer cells during disease progression has been well proven. Cytokines mediated signals can cross talk with chemokines and multiple chemokine-receptors are shown to be associated with tumor progression and metastasis.7-17 Additionally, cell cycle regulation is governed by calcium flux, which in turn is altered by chemokine signaling. In this manuscript we have presented evidence that AG exerts its anticancer effects by modulating cell cycle as well as chemokine receptors (CXCR3 and CXCR7) and their common ligand CXCL11 in PCa cells.
Effect of AG was tested on PCa cells (LNCaP, C4-2b, PC3 and DU-145 cell lines) differing in androgen dependence and metastatic potential. Normal prostatic epithelial cells (PrEC) were used as control. Cell viability of PCa cell lines with or without AG treatment using MTT assay displayed differential susceptibility, IC50 ~20 µM (PC3 and DU-145 cells) and ~50 µM (LNCaP and C4-2b cells), to AG compared to untreated cells; whereas the effect of AG on PrEC cells was minimal (Fig. 1).
Cell cycle progression is facilitated and governed by an extensive cascade of cyclins and cyclin-dependent kinases and phosphatases. AG treatment showed no significant change in cyclin E2 levels in LNCaP or C4-2b cells. However, a decrease in cyclin E2 levels was observed in PC3 and DU-145 cells, 24 hr following AG treatment (Fig. 2). Cyclin A2 was higher in LNCaP, C4-2b and PC3 cells after 6 hr AG treatment, which was sustained till 24 hr in C4-2b and PC3. However, in DU145 cells cyclin A2 levels decreased after 6 hr AG treatment and marginal increase was observed after 24 hr.
On the other hand, Cyclin B1 increased in LNCaP and PC3 after addition of AG but C4-2b cells showed a reduction in cyclin B1 levels after 24 hr and no change was observed in DU-145.
Reduction in phosphorylation of (Ser807/811 and Ser795) Rb (retinoblastoma protein) in LNCaP, C4-2b and PC3 was noted after AG treatment. A reduction in CDC2 phosphorylation (Tyr15) was observed in all the 4 PCa cell lines 24 hr after AG treatment, with C4-2b cells showing reduction in CDC2 phosphorylation as early as 6 hr. Phosphorylation of H3 (Ser10, Ser28, and Thr11) was drastically reduced in C4-2b cells as early as 6 hr but increased in PC3 cells after 24 hr AG treatment. However, AG had no effect on the phosphorylation status of H3 in LNCaP and only negligible phosphorylation of H3 was detected in DU-145 cells. Phosphorylated (S642) Wee1 was reduced in LNCaP cells after 24 hr of AG treatment, while reduction in Wee1 phosphorylation in C4-2b cells was observed after 6 hr. Interestingly, PC3 cells showed an increase in the phosphorylation of Wee1 at 6 hr, but this increase was diminished by 24 hr of AG treatment. No such change was observed in DU-145 cells (Fig. 2).
The level of cyclin dependent kinase inhibitor-p21 was determined in all PCa cell lines. Following AG treatment, a reduction in p21 was observed in LNCaP, C4-2b and DU-145 cells and this reduction was remarkable in DU-145 compared to LNCaP and C4-2b, but no change was observed in PC3 cells that express very low levels of p21.
To further understand the significance of these changes in cell cycle regulators, we studied the impact of AG on PCa cell cycle. Distribution of PCa cells in cell cycle phases was affected after 24 hr AG treatment (Fig. 3). Specifically, LNCaP (G2%: 9.19 to13.9%), C4-2b (G2%: 11.4 to12.7%) and PC3 (G2%: 10.4 to 12.9%) cells were stalled in G2/M phase, while DU-145 cells showed G1 phase arrest (G1%: 40 to 57.4%). Consequently, G1/G2 ratio varied significantly after AG treatment (Table 1).
Multiple chemokines are involved in dissemination and homing of many cancer including PCa.10-17 We found that CXCR3 expression in LNCaP cells was decreased after 6 hr of AG treatment, but this suppressive effect was short lived and the receptor expression recovered after 24 hr. In case of PC3 and DU-145 cells, reduced expression of CXCR3 was maintained from 6 to 24 hr after AG treatment. CXCR7 also showed similar expression patterns following AG treatment (Fig. 4) as CXCR3 and both share a common ligand-CXCL11, although the reduction of CXCR7 expression was not as prominent as CXCR3 in LNCaP cells. The reduction of CXCR3 and CXCR7 expression after AG treatment corresponded with an increase in expression of CXCL11 in PC3 and DU-145 cells, but CXCL11 was below detection in LNCaP cells suggesting a different regulatory mechanism of chemokine receptor(s) expression in hormone-responsive LNCaP (Fig. 4).
Cell viability was determined after blocking CXCR3/CXCR7 with monoclonal antibody to understand the significance of AG mediated reduction in these chemokine receptors (Fig. 5). Growth of the PCa cells was significantly stunted after receptor blockade but stimulation of these receptors with CXCL11 did not show a significant effect on growth of PCa cell line. Migratory potential of PCa cells when treated with suboptimal dose of AG (20 uM - LNCaP and C4-2b and 7.5uM - PC3 and DU-145, respectively) was drastically reduced. There was no significant change in AG induced inhibition of PCa cell migration when CXCL11 was used as chemo-attractant. Blocking CXCR3 prior to AG treatment increased the migration of LNCaP cells toward CXCL11 (Fig. 6). As expected, the migration of control cells significantly decreased after CXCR3 blocking.
Failure to exit the mitotic phase or a derailed cell division cycle can culminate into the development of neoplasia; therefore, most anti-cancer drugs target cell cycle. Unfortunately, these cancer drugs are limited due to their off-target effects and corresponding toxicities. To address this issue, we have used a natural compound, AG (Andrographolide) that has anti-inflammatory and anti-tumor activity in this study.5,6 Unlike other natural compounds, this bioactive constituent is readily absorbed in vivo3,4 and have protective effects on liver and kidney.18,19 We assessed the efficacy of AG on different PCa cell lines. LNCaP and C4-2b cells are less sensitive to this compound, as compared to PC3 and DU-145 cells. Thus, AG has potential to target PCa cells, which are aggressive in nature. Our results suggest that AG affects PCa cell growth by multiple mechanisms and more importantly, it could be effectively used in advanced PCa, which is less responsive to chemotherapies.
AG affects PCa cells survival by arresting their cell cycle.5 We show that it imposes this effect by modulating specific cyclin and cyclin dependent kinase activities. This compound reduces the expression of cyclin D1 and CDK4 and hence the formation of active cyclin D1/CDK4 complex in human colorectal carcinoma Lovo cells.20 AG also induces p27 while decreasing CDK4 expression in rheumatoid arthritis fibroblast-like synoviocytes.21 Our data show effect of AG on Cyclin E2, a G1 phase cyclin that is expressed in a tissue specific manner and at very low levels in non-neoplastic cells. Tumor-derived cells express significantly high levels of cyclin E2.22 Our data show a decrease in cyclin E2 expression by PC3 and DU-145 cells, after AG treatment. Cyclin E2 was not significantly modulated in LNCaP and C4-2b cells. Similar effect of AG was found in lung tumor of VEGF – A165 transgenic mice.23 Cyclin E2 associates with and activates CDK2 to promote phosphorylation of Rb, to drive cells from G1 to S phase. A reduction in cyclin E2 would presumably reduce Rb phosphorylation. We show that LNCaP, C4-2b and PC3 cells treated with AG have reduced Rb phosphorylation (Ser807/811 and Ser795). Hypo-phosphorylated Rb following AG treatment would support sequesteration of E2F, which in its free state enables progression to S phase by upregulating cyclin E and cyclin A expression. AG could also affect expression and activity of cyclin E and CDK2 via GSK3-β and c-Myc.24,25
Cyclin A regulates cell cycle at multiple stages by interacting with CDC2 and CDK2. The cyclin A/CDK2 complex is required for DNA replication and subsequent S phase. The availability of cyclin A/ CDC2 is a rate-limiting step for mitosis. It is apparent that AG treatment increased cyclin A2 in PCa cells, although at different time points. Free-E2F induced expression and degradation during prometaphase are 2 factors regulating cyclin A2 levels.26 Based on Rb phosphorylation status, it is the degradation of cyclin A2 that is affected by AG. Degradation of cyclin A2 during pro-metaphase is required for mitotic exit. Hence, increased cyclin A2 observed in PCa cells after AG treatment implicates that these cells will undergo mitotic arrest.
Cyclin A complexes with CDK to phosphorylate CDC25 and consequently activate cyclin B-CDC2 complex. Cyclin B1 levels increase after AG treatment in LNCaP and PC3 cells and decrease in C4-2b cells. DU-145 cells show no change in their cyclin B1 levels after AG treatment. Studies have shown overexpression of cyclin B1 in various cancers, including prostate27 and it is a poor prognostic marker.28 Nevertheless, high levels of cyclin B1 during early stages of cancer boost immune system by promoting T cells and antibody production.29 Moreover, nuclear accumulation of cyclin B1 is required for mitosis. AG induced cyclin B1 in PC3 and LNCaP cells could be restricted to the cytoplasmic fraction depending on its phosphorylation status.30 Besides, our data also show that AG reduces expression/phosphorylation of CDC2, which is sufficient to stop mitotic division leading to G2/M arrest.31 Similar findings have been reported by others using human glioblastoma cells.32 Since anaphase-promoting complex degrades cyclin B1 before cell cycle proceeds, increased levels of cyclin B1 observed in LNCaP and PC3 cells, may be due to a halt at ‘pre’ anaphase stage of mitosis. Thus AG affects cell cycle in LNCaP and C4-2b cells mainly via cyclin A2 and B1 whereas in DU-145 cells cyclin E related regulation is more important. In PC3 cells, AG modulates all 3 cyclins.
Additionally, it is known that cyclin A2 and B1 expression positively correlates with H3 phosphorylation that characterizes chromatin condensation during initiation of mitosis.33,34 AG treatment increased H3 (Ser10, Ser28, and Thr11) phosphorylation in LNCaP and PC3 cells, but completely abolished it in C4-2b cells, which corresponded well with cyclin B1 levels in these cells. Wee1 (kinase) activity also regulates cell cycle by phosphorylating H3 and CDC2. Phosphorylation of Wee1 at S642 deactivates it. AG treatment affected Wee1 phosphorylation status in all the 4 cell lines to varied extent. Neither H3 nor CDC2 phosphorylation status correlated with Wee1 phosphorylation status. This could be the consequence of AG induced increase in another phosphatase like PP2A.35 Our unpublished data shows AG affects Chk2 that could inhibit CDC25 family of protein, which in turn dephosphorylate (Wee1 phosphorylated) CDC2.36-38 Activity of these phosphatases and kinases synchronize the firing of replication origin during S phase, which consequently controls the length of S phase.39
Shortening of S phase could also be due to reduced p21 as low p21 levels are required for release from S phase arrest and cells with high p21 move slowly through S phase. Expression levels of p21 are low in PC3 cells and AG did not affect p21. Levels of p21 are reduced in all other PCa cells after AG treatment, remarkably so in mutated p53 bearing DU-145 cells. Decrease in p21 levels could be a consequence of the effect of AG on Hsp90, which is important for p21 protein stability.40 AG is reported to interfere with the binding of hsp90 with other proteins and promote Hsp90 degradation.41,42 On the other hand, this effect of AG on p21 levels could be due to its enhanced degradation by other factors.43 In either case, it is clear that despite low levels of p21 AG is capable of arresting PCa cell cycle by affecting the cyclin/CDK complexes.
All cells show G2 phase arrest after AG treatment, except DU-145 cells, which lacks CDKN2A locus and hence have no p16 (p16INK4a) or p14arf. These proteins lead to G1 phase arrest by influencing CDK4-cyclin D and CDK6-cyclin D complexes. These cells also lack Rb1. Absence of all these proteins could lead to the G1 phase arrest in these cells instead of G2 as seen in other PCa cells.
We elucidated the effect of AG on chemokine receptors expressed by PCa cells. CXCR3 was significantly reduced after AG treatment; CXCR7 was also affected. Aksoy et al. 44 demonstrated that surface expression of CXC chemokine receptor CXCR3 is confined to S + G2/M in human airway epithelial cells; however, we could not detect any such cell cycle phase specific expression of CXCR3 (data not shown). CXCL-9, −10 and −11 are 3 non-redundant ligands that bind to CXCR3. Of these, CXCL11 has higher affinity toward CXCR3 and also binds to CXCR7. Interestingly, AG induces CXCL11 expression in PC3 and DU-145 cells. The reduction in CXCR3 expression could be due to AG induced expression of CXCL11 in AR-independent PCa cell lines as observed in other study.44 AG is known to modulate, IFN-α, -β and –γ, cytokines which induce CXCL11.45 Hence, increased CXCL11 expression in AG-treated cells could be due to its effect on IFNs. In addition to this, imitating AG effect by blocking of CXCR3, CXCR7 or addition of CXCL11 in PCa cells reduces proliferation. Neutralizing CXCL11 in AG-treated cells significantly increased the proliferation of PC3 cells (unpublished data). This implicates that AG affects proliferation in other PCa cells via factors other than CXCL11. Migration of all 4 PCa cells was drastically hindered after treatment with sub-lethal doses of AG. As expected, blocking CXCR3 significantly decreased migration of PCa cells toward CXCL11. But, this effect was not seen in AG treated cells as AG treatment itself reduces CXCR3 expression. When CXCR3 was blocked in LNCaP cells prior to AG treatment migration under CXCL11 gradient was higher than only AG treated LNCaP cells. This result implies that in LNCaP cells, where CXCR7 was not significantly affected by AG, CXCR7 compensates for blocked CXCR3. All this together indicates that CXCL11 induced migration is governed by CXCR3 (with an exception of LNCaP) in PCa cells whereas CXCR7 is important for viability. These results show that AG hinders key mechanisms involved in cancer progression by interfering with CXCR3/7-CXCL11 axes.
In conclusion, ability of AG to impede cancer growth by impacting key cell cycle regulators and chemokines, which are otherwise difficult to target due to their crucial physiological role, rationalizes its application as a potent therapeutic/preventive agent.
Andrographolide was purchased from LKT laboratories. PC3 and LNCaP cell lines were obtained from the ATCC (Manassas, VA, USA). Normal prostate epithelial cells (PrEC) were obtained from Clonetics-Biowhittaker (Walkersville, MD, USA) and cultured in PrEMB medium (Clonetics-Biowhittaker). PC3 cells were cultured at 37°C with 5% CO2 in F12 K medium with 2 mM L-glutamine and with 10% fetal bovine serum (FBS) (Sigma, St Louis, MO, USA). LNCaP, C4-2b and DU-145 cells were cultured in RPMI-1640 with 10% FBS. All the experiments were carried out in presence of 1% FBS containing respective medium unless specified.
For cell cycle analysis, PCa cells (1 × 106) treated with and without AG were washed thrice with fluorescence-activated cell-sorting (FACS) buffer (PBS supplemented with 2% fetal bovine serum). Cells were then stained with propidium iodide (Cell signaling) as per manufacturer's instruction. Cells were analyzed by flow cytometry using a Guava easy cyte flow cytometer (Millipore) and flow Jo 10.0.6 software (Treestar Inc.).
PCa cells were seeded at a density of 2 × 104/100ul/well in 96 well plates. Next day, keeping the volume constant, cells were treated with different concentration (0–100μM) of AG. Cell proliferation was assessed 24, 48 and 72 hr using MTT assay.46 Briefly, MTT (20 μL of 5 mg/ml in PBS) was added in each well and plate was incubated at 37°C for 3 hr. Resultant formazan crystals were dissolved in 100 μL of DMSO and absorbance was measured at 570 nm. To validate the data the experiment was repeated 3 times.
Migration studies were performed using BD Biocoat migration chambers (Becton-Dickson Labware). Inserts were activated with serum free DMEM for 2 hr at 37°C with 5% CO2. Next, 0.5 × 105 cells were added to the top chamber of inserts and 100ng/ml CXCL11 (Peprotech, NJ) was added as chemo-attractant in the bottom chamber. To determine if the migration of PCa cells is mediated specifically by CXCR3-CXCL11 interaction, cells pre-incubated with 1.0 μg/ml anti-CXCR3 antibody (R&D Systems) were added to the top chamber in one well of Matrigel or control inserts and allowed to migrate under chemotactic gradient of CXCL11 for overnight at 37°C and 5% CO2. After incubation, non-migrating cells on the upper surface of the membrane were removed with a cotton swab. Cells at the bottom surface of the insert were fixed with 100% methanol for 2 min, stained for 2 min with 1% Toluidine blue O in 1% Borax (Fisher Scientific), and rinsed twice with de-ionized water. Migrated cells were counted by microscopy at 40X magnification. All experiments were repeated 3 times to validate the results.
Results of viability and migration assays were analyzed by one-way ANOVA. Values were considered significantly different at a level of 0.05.
No potential conflicts of interest were disclosed.
Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number R21CA169716 and in part by U01CA179701, SC1CA180212 and Morehouse School of medicine FACS core. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.