Malignant tumors remain a significant health threat, with death often occurring as a result of metastasis. Cell adhesion is a crucial step in the metastatic cascade of tumor cells, and interruption of this step is considered to be a logical strategy for prevention and treatment of tumor metastasis. Celastrol [3-hydroxy-24-nor-2-oxo-1(10),3,5,7-friedelatetraen-29-oic acid], a quinone methide triterpene from the medicinal plant Tripterygium wilfordii, possesses antitumor activities, whereas the underlying mechanism(s) remains elusive. Here, we found that celastrol inhibited cell-extracellular matrix (ECM) adhesion of human lung cancer 95-D and mouse melanoma B16F10 cells. This inhibition was achieved through suppressing β1 integrin ligand affinity and focal adhesion formation, accompanied by the reduced phosphorylation of focal adhesion kinase (FAK). In understanding the underlying mechanisms, we found that celastrol activated p38 mitogen-activated protein kinase (MAPK) by phosphorylation before the decrement of phosphorylated FAK and that this action was independent of the presence of fibronectin. Using 4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)1H-imidazole (SB203580), a specific inhibitor of p38 MAPK, the effects of celastrol on β1 integrin function, cell-ECM adhesion, and phosphorylation of FAK were partially attenuated. In addition, focal adhesion-dependent cell migration and invasion were both inhibited by treatment with celastrol. Finally, the antimetastatic activity of celastrol was examined in vivo using the B16F10-green fluorescent protein-injected C57BL/6 mouse model, as indicated by decreased pulmonary metastases in celastrol-administrated mice. Taken together, these data demonstrate for the first time that celastrol exerts potent antimetastatic activity both in vitro and in vivo, and they provide new evidence for the critical roles of p38 MAPK in the regulation of integrin function and cell adhesion.
The maintenance, promotion, and disruption of cell adhesion are particularly important in cancer progression and metastasis. Adhesion occurs not only in malignant cell detachment from the primary carcinoma but also in tumor cell attachment to distant tissue, and multiple cell adhesion molecules are involved in these events (Pascho et al., 2009). In the former case, adhesion between cell and cell is down-regulated, and classic cadherin molecules such as E-cadherin are linked to the actin cytoskeleton through linker molecules, including α- or β-catenin (Lorch et al., 2007). In the latter case, cell surface molecules mediate the cell-extracellular matrix (ECM) attachment. Among these molecules, integrin clusters, the most important molecules, are responsible for forming a stable membrane platform with a high avidity for the ECM on the outside of the cell; and on the inside of the cell, they have multiple binding sites for adaptors and enzymes that become compartmentalized into plasma membrane-associated complexes (Streuli and Akhtar, 2009). Accordingly, owing to its initial roles in cancer metastasis, interruption of cell-ECM adhesion has been considered as a potential strategy for cancer prevention and treatment. Unfortunately, in spite of the cell-ECM adhesion and their components frequently being described for various tumors, less attention has been focused on potential manipulations on cell-ECM adhesion or the related signaling transductions, which may contribute to the control of tumor progression and metastasis.
Integrins, composed of α and β transmembrane subunits (Taddei et al., 2003), are the main receptors mediating the anchorage of cells to the components of the ECM. The attachment of cells to certain ECM proteins, such as fibronectin (FN), collagen, and laminin, leads to the clustering of integrins, followed by recruiting cytoplasmic proteins, such as focal adhesion kinase (FAK) and paxillin. The interaction of β subunit of integrins and FAK causes kinase autophosphorylation at Tyr397 of FAK (van Nimwegen and van de Water, 2007) and then fully phosphorylates and activates FAK to contribute to the assembly of focal adhesion complex (Huang et al., 2006), which is critical for coordinated movement and is the major factor influencing cell signaling and cancer metastasis (Brakebusch and Fässler, 2003). Therefore, inhibition of the integrin signaling pathway is of great interest in the development of drugs for the effective treatment of cancer.
Celastrol [3-hydroxy-24-nor-2-oxo-1(10),3,5,7-friedelatetraen-29-oic acid; Fig. 1A], also known as tripterine, was identified decades ago in traditional Chinese medicine from the plant Tripterygium wilfordii and is used for the treatment of cancer and inflammatory diseases (Sethi et al., 2007). Preclinical research demonstrates that celastrol exhibits significant in vitro and in vivo activity against malignant tumor cells and xenografts (Yang et al., 2006), and the mechanism of action involved has been studied extensively. Zhang et al. (2008) showed that celastrol is a novel Hsp90 inhibitor that disrupts the Hsp90-Cdc37 interaction in the superchaperone complex to exhibit antitumor activity in vitro and in vivo. Sethi et al. (2007) showed that celastrol potentiates TNFα-induced apoptosis and suppresses invasion of tumor cells by inhibiting NF-κB-regulated gene products and transforming growth factor-β-activated kinase 1-mediated NF-κB activation. Moreover, the ability of celastrol to inhibit all-trans-retinoic acid-caused adhesion between leukemia cells and endothelial cells by modulating the expression level of E-selectin and vascular adhesion molecular-1 has also been reported previously (Xu et al., 2007).
These findings suggested that celastrol might possess antimetastasis activity, especially by modulating cell adhesion. However, to date, the effects of celastrol on cell-ECM adhesion and integrin have not been addressed, despite the central roles of integrin and focal adhesion complex formation in tumor metastasis.
This study was designed to examine the potential effects of celastrol on the integrin signaling pathway during cell adhesion and to detect the antimigration and anti-invasion effects of celastrol in vitro and its antimetastasis activities in vivo. In addition, given the emerging correlation among intracellular mitogen-activated protein kinase (MAPK), AKT, and adhesion (Hess et al., 2005; Zeng et al., 2006), particularly the role of each in the regulation of integrin function (Bauer et al., 2006), the participation of the MAPK and AKT pathway in integrin signaling and adhesion in celastrol-treated cancer cells was further investigated.
Materials and Methods
Celastrol (Fig. 1A) was synthesized (>99% purity) by Professor Wei Lu (East China Normal University, Shanghai, China) and dissolved in dimethyl sulfoxide (DMSO) as a stock solution at 50 mM. The stock solution was kept frozen in aliquots at −20°C and thawed immediately before each experiment. DMSO, thiazolyl blue tetrazolium bromide (MTT), and poly-l-lysine (PLL) were purchased from Sigma-Aldrich (St. Louis, MO). Fibronectin and SB203580 were obtained from Merck Biosciences (Darmstadt, Germany).
Cell Lines and Cell Culture.
Mice melanoma cell line B16F10 and human lung cancer cell line 95-D were obtained from the Chinese Academy of Sciences, Shanghai Institute of Cell Biology (Shanghai, China) and maintained in RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen), l-glutamine (2 mM), penicillin (100 IU/ml), streptomycin (100 μg/ml), and HEPES (10 mM), pH 7.4, in a humidified atmosphere of 95% air plus 5% CO2 at 37°C.
Cell Proliferation Assay.
Cell viability was assessed by MTT assay. In brief, cells were seeded into 96-well plates and cultured on fibronectin, and then they were exposed to serial concentrations of celastrol for different times (hours). Next, cells were incubated with MTT (5 mg/ml; 20 μl/well) for 4 h, and the formazan granules generated by live cells were dissolved in DMSO and shaken for 10 min. The absorbance was measured at 570 nm using a Multiskan spectrum microplate spectrophotometer (Thermo Fisher Scientific, Vantaa, Finland). The inhibition rate on cell proliferation was calculated for each well as follows: (A570control cells − A570treated cells)/A570control cells × 100%.
Cell Adhesion Assay.
The cell adhesion assay was done as described previously (Zhou et al., 2008). In brief, fibronectin or poly-l-lysine was diluted in sterile water and applied to 96-well plates overnight at 4°C. Nonspecific binding sites were blocked with 1% bovine serum albumin for 30 min. Cells were serum-starved for 45 min; detached with 2 mM EDTA in PBS; plated in triplicate onto wells in serum-free medium with vehicle, celastrol, or blocking antibody against β1-integrin (10 μg/ml, 6S6; Millipore, Billerica, MA); and allowed to adhere for 1 or 2 h according to the cell type at 37°C. Nonadherent cells were washed away, and adherent cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. After extensive washing, dye was extracted with 10% acetic acid and quantified by measuring absorbance at 595 nm on a multiwell spectrophotometer (Thermo Fisher Scientific). The rate of adhesion was calculated using the following formula: ratio of adhesion = (A595celastrol − A595blank)/(A595control − A595blank) × 100%.
Wound Healing Assay.
B16F10 and 95-D cells were seeded in 24-well plates and cultured until 70 to 80% confluent. Using a pipette tip, we made a straight scratch to represent an artificial wound. After treatment with different concentrations of celastrol, the migration of cells across this artificial wound was assessed.
A B16F10 and 95-D cell invasion assay was performed in a Transwell Boyden chamber (Corning Life Sciences, Teterboro, NJ) using a polycarbonate filter with a pore size of 8 μm (coated with 30 μg/well Matrigel; BD Biosciences, San Jose, CA) (Zhang et al., 2005). A cell suspension (3 × 105 cells/ml) was added to each well. After 8-h attachment, the top chamber was replaced with serum-free medium treated with different concentrations of celastrol or with control [0.05% DMSO (v/v)]. The bottom chamber contained 0.6 ml of RPMI 1640 medium supplemented with 5% fetal bovine serum. After 18-h incubation at 37°C, all of the nonmigrant cells were removed from the top chamber; the migrated cells were then fixed with 90% ethanol and then stained with 0.1% crystal violet in 2% methanol. The stained cells were subsequently photographed and quantified under a microscope (model DFC300 FX; Leica, Wetzlar, Germany) and then extracted with 10% acetic acid. The absorbance values were determined at 595 nm, and the rate of migration was calculated using the following formula: ratio of migration = (A595celastrol − A595blank)/(A595control − A595blank) × 100%.
95-D cells were plated onto glass culture slides coated with fibronectin or poly-l-lysine in serum-free medium, allowed to adhere for 2 h, and then incubated with various concentrations of celastrol with or without pretreatment with SB203580 (10 μM) for 30 min. Cells were fixed with 4% paraformaldehyde and permeabilized with PBS containing 0.1% Triton X-100. After blocking with 5% bovine serum albumin for 30 min, cells were incubated with primary active β1-integrin-recognizing antibodies 12G10 (Abcam plc, Cambridge, UK) or total β1-integrin-recognizing antibody K-20 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at a dilution of 1:100 for 1 h; washed three times with PBS; and then incubated with the Alexa Fluor 488-conjugated secondary antibodies (Invitrogen) in the dark. The nuclei were visualized by staining with 4′6-diamidino-2-phenylindole (DAPI). Pictures were obtained using a microscope (model DFC300 FX; Leica).
Expression of Cell Surface Integrins.
95-D cells were seeded onto six-well plates in complete medium and cultured overnight. The growth medium was then replaced with serum-free RPMI 1640 medium, and the cells were incubated in the presence of 4 μM celastrol for 2 h. Cells were detached, washed with 5% bovine serum albumin three times, and incubated on ice for 1 h with primary active β1-integrin-recognizing antibodies 12G10 (Abcam plc). Cells were then washed and incubated with secondary fluorescent anti-mouse antibody conjugated with Alexa Fluor 488 (Invitrogen) for 60 min at 4°C. The cells were then washed three times and resuspended in ice-cold PBS. The fluorescence intensity was measured with a flow cytometer.
95-D cells were seeded onto six-well plates in complete medium and cultured overnight. Then, the cells were detached using 5 mM EDTA and resuspended in PBS. Cells were incubated with MnCl2 with or without celastrol for 2 h at 4°C. Mn2+ was used to activate integrins, and then 50 μl of the treated cells was mixed with 50 μl of the primary antibody 12G10 (20 μg/ml; Abcam plc) (specific for activated integrins) and incubated for 60 min on ice. After three washing steps with PBS with 1% fetal calf serum, the cells were stained with the Alexa Fluor 488-conjugated secondary antibodies in the dark for 30 min. After three washing steps, the cells were resuspended in a volume of 300 μl of PBS. The fluorescence intensity was measured with a flow cytometer.
Preparation of Total Cell Extracts and Western Blot Analyses.
Cells (5 × 105/well) were treated with celastrol for serial times and concentrations. Proteins were extracted with lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 0.1% SDS, 0.5% deoxycholic acid, 0.02% sodium azide, 1% Nonidet P-40, 2.0 μg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride). The lysates were centrifuged at 10,000g for 30 min at 4°C. Equivalent amounts of proteins were analyzed by SDS-polyacrylamide gel electrophoresis. Proteins on gels were then electrotransferred to polyvinylidene difluoride membranes (Millipore) and probed with primary antibodies to integrin (K-20), p38, AKT, extracellular signal-regulated kinase (ERK), FAK, β-actin (Santa Cruz Biotechnology, Inc.), p-p38 (Thr180/Tyr182), p-AKT (Ser473), p-ERK (Thr202/Tyr204) (Cell Signaling Technology, Danvers, MA), p-FAK (Tyr397) (Invitrogen), and horseradish peroxidase-labeled secondary antibodies (Southern Biotech, Birmingham, UK). Proteins were visualized using enhanced chemiluminescence detection reagents (GE Healthcare, Buckinghamshire, UK).
Intravenous Injection of B16F10-GFP Cells in Mice and Administration of Celastrol.
For green fluorescent protein (GFP) gene transduction, B16F10GFP cells were achieved from stable transfection with pZsGreen-1-GFP (Clontech, Mountain View, CA) by Lipofectamine 2000 (Invitrogen) and G418 (Roche Diagnostics, Mannheim, Germany) selection (800 μg/ml) and maintenance (400 μg/ml) (Yang et al., 2010). Mice were then injected intravenously with a single dose of 2 × 105 B16F10-GFP cells in a total volume of 0.2 ml in RPMI 1640 medium (Yang et al., 2010). The next day, the mice were randomized to control and treated groups (three mice in each group), and they received vehicle [1% DMSO/7% Cremophor/ethol (3:1) and 92% PBS i.p. administration] or celastrol (2.0 or 4.0 mg/kg i.p. administration) every 2 days. Mice with tumor implants were killed 3 weeks later. The lungs were removed and photographed under a fluorescent macroimaging system. Numbers of pulmonary nodules were counted under a fluorescence stereomicroscope.
Data are presented as mean ± S.D., and significance was assessed with a Student's t test. Differences were considered significant at p < 0.05.
Celastrol Inhibits Fibronectin-Mediated Cell Adhesion and Spreading.
Previous studies demonstrated that celastrol modulated the expression of metastasis-related proteins, including MMP-2 and MMP-9 (Sethi et al., 2007). Accordingly, we hypothesized that celastrol possessed antimetastasis abilities and would impose effects on metastasis-related events, including cell-ECM adhesion, cell migration, and invasion. First, using mouse melanoma carcinoma B16F10 cells and human metastasis lung cancer 95-D cells as models, the activity of celastrol on cell adhesion to FN, the most common component of the ECM and typical integrin ligand, was examined from three independent experiments. Figure 1, B and C, shows that celastrol inhibited cell adhesion to fibronectin in a concentration-dependent manner. When 4 μM celastrol was applied on 95-D cells for 1 h, the ratio of cell adhesion declined to 41% from 100% (4 μM versus control; p < 0.001); similarly, celastrol (8 μM; 2 h) down-regulated B16F10 cell adhesion to fibronectin from 100 to 23% (8 μM versus control; p < 0.001). In contrast, the cell adhesion to PLL, an event that did not engage with integrin, was not affected by celastrol treatment (Fig. 1, B and C). We next asked whether the inhibitory effect of celastrol on fibronectin-mediated cell adhesion was due to cytotoxicity. An MTT assay was used to evaluate cell viability in parallel with the adhesion assay, and the results from three independent experiments (Fig. 1, B and C) showed that both B16F10 and 95-D cells treated with equivalent concentrations of celastrol as that used in the adhesion assay (0, 1, 2, 4, and 8 μM exposure for 2 h in B16 cells; and 0, 0.5, 1, 2, and 4 μM exposure for 1 h in 95-D cells) maintained a high level of viability, thereby excluding the impact of cytotoxicity on the reduced cell adhesion to fibronectin.
Because the formation of focal adhesion is a crucial step in cell adhesion and spreading (Carragher and Frame, 2004), we further investigated the effect of celastrol on fibronectin-dependent formation of focal adhesion. 95-D cells were plated onto fibronectin-coated or PLL-coated slides in serum-free RPMI 1640 medium. Figure 1D shows that fibronectin mediated cell spreading with focal adhesions localized at the periphery of cells; in contrast, cells cultured on PLL-coated slides displayed a rounded morphology and failed to assemble focal adhesions. Upon treatment with celastrol, the fibronectin-dependent formation of focal adhesion was dramatically reduced, which was accompanied by the rounding up of cells and the retraction of cells from the substratum. These data collectively implicated the inhibitory effects of celastrol on fibronectin-mediated cell-ECM adhesion and the focal adhesion formation.
Celastrol Reduces the Activity of β1 Integrin.
The observation that celastrol suppressed fibronectin-dependent cell-ECM adhesion inspired us to explore the possible cellular target(s) of this compound and thus provide insight into the molecular mechanism underlying the antitumor activity possessed by celastrol. Numerous studies report that integrins are the major receptors for ECM, making up 24 combinations of 18α and 8β subunits (Huveneers et al., 2007), and that the protein expression levels, affinity for ligands, cell surface distribution, and downstream signaling pathway are associated with the ability of integrins to evoke cell-ECM attachment. Among these molecules, β1 integrin is the prototypic fibronectin receptor (Zeng et al., 2006); therefore, we focused on this molecule to explore the mechanisms of celastrol-interrupted cell adhesion.
The expression level of β1 integrin in celastrol-treated cells was first evaluated using Western blot, and the results showed that the protein expression level of β1 integrin subunit was not affected by celastrol treatment (Fig. 2, C and D). Next, we investigated whether the modulation of function rather than protein expression level of integrin was responsible for the suppressed cell adhesion caused by celastrol. To determine the effect of celastrol on activated integrins, we performed immunofluorescence assays using the antibody 12G10 that specifically recognizes activated β1 integrin (Kim et al., 2008). As shown in Fig. 2A, celastrol exposure (4 μM) significantly down-regulated integrin activity of 95-D cells compared with the control group, suggesting that celastrol reduced the activated β1 integrin in a concentration-dependent manner.
To clarify the effect of celastrol on integrin-mediated cell adhesion, we used a cell adhesion assay and introduced MnCl2 to further validate our hypothesis. Preincubation of the cells with 1 mM Mn2+ would uniformly activate integrins by inducing an active conformation (Humphries et al., 2005). As expected, three independent experiments showed that Mn2+ pretreatment restored cell adhesion to fibronectin, which was strongly inhibited by celastrol (Fig. 2B). When exposed to 4 μM celastrol, the adhesion ratio of 95-D cells in celastrol and celastrol plus Mn2+ groups was 41 and 85% (p < 0.001), respectively, whereas adhesion of B16F10 cells in these two groups was 24 and 85% (p < 0.001), respectively. In contrast, pretreatment with EDTA, which complexes divalent cations and thus hinders integrins from establishing their active conformation, imposed no effect on celastrol-treated cells, as indicated by the low level of adhesion in all the groups of 95-D cells (Fig. 2B). These data further demonstrated that upon celastrol exposure, the inactivation of β1 integrin mediated the suppressed adhesion to fibronectin and that the inhibited adhesion could be rescued by Mn2+ pretreatment.
The aforementioned data prompted us to examine whether β1 integrin-associated signaling pathways were affected by celastrol treatment. It has been reported that when integrins bind to ECM proteins, they cluster in the plane of the cell membrane, associate with signaling complexes, and relay the signals to downstream factors, including FAK and paxillin. As shown in Fig. 2, C and D, upon exposure to celastrol, the phosphorylation of FAK on the Tyr397 site, which was an early consequence of the integrin-ligand interaction, was suppressed in a concentration- and time-dependent manner, with the reduction of p-FAK (Tyr397) being initiated after 30 min in both 95-D and B16F10 cells.
Taken together, these results demonstrated that celastrol down-regulated the function of β1 integrin rather than the protein expression level, reduced the downstream phosphorylation of FAK, and resulted in the subsequent loss of cell adhesion to fibronectin. The loss of β1 integrin function and cell adhesion to fibronectin could be abated by Mn2+ pretreatment.
Celastrol Activates p38 Mitogen-Activated Protein Kinase.
Our results showed that celastrol potentially inhibited integrin-mediated cell adhesion; decreased β1-integrin ligand affinity; and blocked the phosphorylation of FAK (Tyr397), a crucial component of integrin signaling complexes. Therefore, the major challenge to us was to explore the mechanism leading to the prohibition of β1 integrin function in celastrol-exposed cells.
It is well established that the activation of ERK, p38 MAPK, and AKT play important roles in integrin-mediated bidirectional signaling pathways. For this reason, we first examined whether celastrol can modulate the expression level and phosphorylation status of these kinases. Figure 3 shows that celastrol triggered the phosphorylation of p38 MAPK in a concentration- and time-dependent manner. In contrast, it imposed no effect on phosphorylated ERK and AKT, implying that the activation, i.e., the phosphorylation of p38 rather than that of ERK and AKT, was regulated upon celastrol treatment. It is interesting to note that the phosphorylation of p38 was initiated after 10 min of exposure to celastrol, which is earlier than that observed for celastrol-induced dephosphorylation of FAK (Tyr397) (Fig. 2, C and D). The observed sequential phosphorylation status alteration of these two proteins indicated that, upon the treatment with celastrol, p-p38 was up-regulated before the reduction of p-FAK (Tyr397) in 95-D and B16F10 cells, raising the possibility that p38 activation resulted in the subsequent inhibition of integrin signaling transduction.
Activated p38 MAPK Partially Contributes to the Inhibition of Integrin Activity and Cell Adhesion Caused by Celastrol.
The aforementioned data showed that the phosphorylation of p38 increased before the down-regulation of p-FAK (Tyr397); thus, in this context, we investigated the role of the p38 signaling pathway in the regulation of integrin activity and cell adhesion to fibronectin in celastrol-treated cells.
Our results demonstrated that the phosphorylation levels of ERK, p38, and AKT in the absence of fibronectin were similar to those in the presence of fibronectin (Fig. 4), suggesting that the phosphorylation status of these kinases was independent of fibronectin; in others word, p38 probably located upstream of β1 integrin in the corresponding signaling pathways. To test this hypothesis, SB203580, a specific inhibitor of p38, was introduced to examine the role of p38 in celastrol-inhibited cell adhesion. Pretreatment with SB203580 partially reversed the suppression of fibronectin-mediated cell adhesion by celastrol (0.5, 1, 2, and 4 μM). The ratio of cell adhesion was rescued in 95-D cells, as indicated by 92, 86, 71, and 58%, respectively, compared with 87, 68, 65, and 41% without pretreatment of SB203580 (Fig. 5A). Statistical analyses of the data from three independent experiments indicated that in 4 μM celastrol-treated groups, SB203580 significantly enhanced the adhesion ratio (p < 0.05). Similar observations were achieved in B16F10 cells: the ratio of cell adhesion was ascended to 95, 90, 81, and 58% after treatment of celastrol in combination with SB203580; whereas without SB203580, the adhesion ratio was 86, 77, 41, and 24% (4 μM, without SB203580 versus with SB203580; p < 0.05; 8 μM, without SB203580 versus with SB203580, p < 0.01) (Fig. 5A). In addition, 12G10 antibody and fluorescence-activated cell sorting analysis were used to confirm whether the attenuation of prohibited cell adhesion to FN was attributed to the rescue of integrin activity caused by p38 inhibition. Preparation of the cells for the fluorescence-activated cell sorting experiment resulted in inactive integrins. Therefore, the activity of integrin was triggered by pretreatment with 1 mM Mn2+, which uniformly activates integrins by inducing an active conformation (Humphries et al., 2005). Figure 5B shows that SB203580 partially restored the loss of β1 integrin activity elicited by celastrol, without affecting the protein expression level, and that it attenuated the decrease of FAK phosphorylation (Tyr397) (Fig. 5, C and D) in celastrol-exposed 95-D and B16 cells.
To further confirm that p38 MAPK is located upstream of β1 integrin, an antibody blocking the activity of β1 integrin was introduced. Under our experimental conditions, the p-p38 level was not induced by the inactivation of β1 integrin by using 6S6, and the combination of 6S6 and celastrol caused no more elevation of p-p38 than that in celastrol-treated cells, thus excluding the regulatory effects of β1 integrin on p38 MAPK. Taken together, these findings provided evidence that activated p38 MAPK was integral to the regulation of integrin signaling transduction and that the suppression of β1 integrin activity was due, in part, to the activation of p38 MAPK.
Celastrol Inhibits Cell Migration and Cell Invasion.
As demonstrated in Fig. 2, C and D, celastrol treatment caused a dose- and time-dependent decrease of p-FAK (Tyr397). Because the phosphorylation of Tyr397 on FAK has been reported to be linked to integrin clustering and is regarded as a critical factor in the initiation of cell migration and invasion (Schlaepfer et al., 2004), we next examined whether celastrol had the ability to inhibit cell migration and cell invasion. To address this question, a wound healing assay was performed. After treatment with serial concentrations of celastrol, the migration of cells across an artificial wound was assessed. In the untreated controls, cell spreading from the edges of the wound could be shown at 0 h, and near complete closure of the wound edges occurred at 24 h. With the increasing concentration of celastrol, the migration inhibition capacity was enhanced; in particular, there was virtually no migration in the 4 μM celastrol-treated group at 24 h (Fig. 6A). Simultaneously, the cell viability of the cells treated with indicated concentrations of celastrol for 24 h was measured using an MTT assay to exclude the impact of cytotoxicity on cell migration. The results showed that little cytotoxicity was triggered under our experimental conditions (Fig. 6A).
A Transwell assay was used to investigate the effect of celastrol on tumor cell invasion. We seeded the tumor cells into the top wells of a Matrigel invasion chamber in the absence of serum. The bottom compartment contained 0.6 ml of RPMI 1640 medium supplemented with 5% serum. As shown in Fig. 6B, celastrol treatment at serial concentrations (0.5, 2, and 4 μM) for 18 h dramatically suppressed the invasion activity of 95-D cells by approximately 70, 28, and 11%, respectively. The invasion capability of B16F10F10 cells was inhibited after exposure to different concentrations (1, 2, and 4 μM) of celastrol by approximately 78, 54, and 21%, respectively.
Celastrol Inhibits Cancer Metastasis In Vivo.
The above-mentioned findings collectively demonstrated that celastrol exhibited significant antiadhesion, antimigration, and anti-invasion activities on 95-D and B16F10 cells. We were thus encouraged to evaluate the potential inhibitory effects of celastrol on cancer metastasis in vivo. First, we isolated stable, highly expressed GFP transductants of B16F10-GFP cells (Fig. 6C) as reported by our previous study (Yang et al., 2010). Then, we examined the effect of celastrol on the development of pulmonary metastases by using the B16F10-GFP-c57 mouse model (Yang et al., 2010). Several mice that developed lung metastases were analyzed by measuring nodule counts. The results demonstrated that the pulmonary nodule counts reduced in a dose-dependent manner. In 2 and 4 mg/kg celastrol-administrated mice, the average pulmonary nodule counts were 30.3 and 18.3 for each mouse, respectively, which were significantly lower than values from the vehicle-treated group (37.3) (Fig. 6D), demonstrating the in vivo antimetastasis capability possessed by celastrol.
Cancer is the most common disease worldwide. The prognosis and the overall survival are mainly determined by the progression of cancer metastasis rather than by the primary carcinoma, even in patients with an isolated tumor (Hervy et al., 2006). As an end-stage malignant disease, cancer metastasis relapse is often associated with resistance to therapy after systemic treatments (Nguyen et al., 2009). Thus, interfering with metastasis has been regarded as a promising strategy to improve the current cancer treatment and therapy. However, exploration and development of novel antimetastatic agents remain major challenges.
Celastrol, extracted from the Celastraceae family of plants, is of particular interest recently because of its superior antitumor capabilities, mentioned by different reports, against a variety of tumors, including prostate cancer, breast cancer, pancreatic cancer, and leukemia, through modulating proteasome activity (Yang et al., 2006), heat shock response (Hieronymus et al., 2006; Zhang et al., 2008), and NF-κB signaling pathways (Sethi et al., 2007). Some of these reports also mentioned the ability of celastrol to inhibit tumor metastasis; nevertheless, to date, no study has described its antimetastatic activities comprehensively or characterized the underlying mechanisms profoundly. In this study, for the first time, we systematically evaluated the effects of celastrol represented in metastatic-related events, including cell-ECM adhesion, migration, invasion, and the development of pulmonary metastases in vivo, by using corresponding assays and models. The anti-invasion activity achieved in this study was in accordance with the reported inhibitory effects of celastrol on TNF-induced MMP-9 and vascular endothelial growth factor (Sethi et al., 2007). Together with previous reports, this natural compound was identified to be a potent candidate as a therapeutic agent to prevent cancer metastasis, and the molecular mechanisms remained to be fully established.
β1 integrin belongs to a class of cell surface receptors that play diverse roles in mediating multiple aspects of malignant cell behavior, particularly in cell adhesion, invasion, and migration (Zutter et al., 1993; Gui et al., 1996; Shaw, 1999; Berry et al., 2004; White et al., 2004). Aberrant expression and improper localization of β1 integrin at the cell surface have been observed in mammalian cancer cells and are indicated to be closely related to their metastatic behavior (Tawil et al., 1996). Moreover, inhibition of β1 integrin is demonstrated to abrogate the formation of metastasis in gastric and breast cancer models (Elliott et al., 1994; Fujita et al., 1995; Kawamura et al., 2001). From a clinical perspective, antibodies targeting β1 integrin have been tested in preclinical models for cancer treatment purposes, and some even entered clinical trials. However, few small-molecule antagonists have been developed to inhibit β1 integrin.
In this study, celastrol, as a natural small-molecule compound, was demonstrated to target β1 integrin, and we found that the effects of celastrol imposed on β1 integrin were responsible for its antiadhesion capability, which may provide an understanding of its antimetastatic mechanisms. Our results showed that although the protein expression levels of β1 integrin were left intact, the active conformation β1 integrin recognized by the specific antibody 12G10 was significantly decreased when the cells were treated with celastrol. In addition, using a cell adhesion assay, celastrol disrupted the interaction between β1 integrin and fibronectin, which was restored by Mn2+. Moreover, phosphorylation of FAK, a kinase located downstream of β1 integrin, was blocked by celastrol treatment. These data collectively revealed that via suppressed β1 integrin function, celastrol subsequently resulted in the loss of cell adhesion to fibronectin and the reduction of activated downstream signaling proteins such as FAK, thereby exerting its antimetastatic activity.
In another study by our group (H. Zhu, X.-W. Liu, W.-J. Ding, D.-Q. Xu, Y.-C. Zhao, W. Lu, Q.-J. He, and Bo Yang, submitted for publication), celastrol was shown to activate p38 MAPK via phosphorylation at Thr180 and Tyr182 within 2 h, indicating that the MAPK pathway responded rapidly upon celastrol treatment. These findings provided a clue for us to explore the potential interactions between integrin and p38 MAPK, an essential signaling molecule. Although the molecular mechanisms underlying integrin-fibronectin interactions are unclear, p38 MAPK has been reported to play crucial but yet elusive roles in regulating the integrin function. A recent study reported that CD26, an antigen with a key role in T-cell biology, depleted by small interfering RNA transfection in T-anaplastic large cell lymphoma karpas 299, made the cells lose the ability to adhere to fibronectin and collagen I. Specifically, this CD26-induced cell adhesion was illustrated to be regulated by p38 MAPK-dependent activation of β1 integrin (Sato et al., 2005). In contrast, Da Silva et al., (2003) claimed that p38 was involved in preventing β1 integrin function, which suggested that p38 inhibited β1 integrin rather than activated it. In our study, celastrol activated p38 MAPK in a time- and concentration-dependent manner. And SB203580, a specific inhibitor of p38 MAPK, partially reversed the inhibition of β1 integrin affinity for 12G10 and the reduction of p-FAK (Tyr397) caused by celastrol treatment. These data suggested that activated p38 was responsible, in part, for celastrol-induced suppression of integrin ligand affinity and downstream phosphorylation of FAK. These findings not only enrich the study of celastrol-induced inhibition of integrin function but also offer the first-hand evidence that p38 MAPK is involved in integrin-mediated cell adhesion.
Given the central role that integrin and its downstream molecules, such as FAK, play in cancer metastasis (Hehlgans et al., 2007), we further addressed whether celastrol exhibited inhibitory activities on cell migration and invasion. Cell migration is a complex, highly regulated process requiring the continuous formation and disassembly of adhesion, and it has been implicated in the invasion of tumor cells (Lorch et al., 2007). FAK, a critical downstream substrate of integrin, has been regarded as an important contributor to cell migration and invasion (Schlaepfer et al., 2004). The work presented here identifies the decrease in tumor cell migration and invasion in vitro induced by celastrol. The inhibition effects probably resulted from the shift in the phosphorylation and activation status of FAK. A previous study by Sethi et al. (2007) suggested that suppression of TNF-induced invasion of tumor cells could be due to down-regulation of MMP-9 as a result of celastrol treatment. However, Tsutsumi et al. (2008) concluded that cell surface Hsp90 plays an important role in modulating cancer cell migration, which is independent of the function of the intracellular Hsp90 pool, and that small-molecule inhibitors of surface Hsp90 may provide a new approach to targeting the metastatic phenotype. Because celastrol has been identified as an Hsp90 inhibitor (Zhang et al., 2008), the possibility of Hsp90-mediated antimigration and anti-invasion abilities of celastrol also could not be excluded. A more precise relationship between the integrin pathway and celastrol-induced inhibition of cancer cell migration and invasion should be explored.
In summary, our study found that cell-ECM adhesion significantly diminished as a result of celastrol treatment, accompanied by the inactivation of β1 integrin, reduction of p-FAK (Tyr397), and activation of p38 MAPK (Fig. 7). The loss of β1 integrin activity, decrease in p-FAK, and suppression of adhesion to fibronectin were partially rescued by inhibiting p38 MAPK by using a specific inhibitor, denoting the involvement of p38 MAPK in the activity manipulation of β1 integrin induced by celastrol. Moreover, the achieved distinct inhibitory effects on cell migration and invasion in vitro and the antimetastatic capacity of celastrol in vivo upon celastrol treatment further merit the development of this natural compound as a therapeutic anticancer agent and provide clues for the chemical modification on celastrol to explore more efficient drug candidates. Our present work not only lends new insight into the anticancer activities and mechanisms of celastrol but also gives clues that p38 MAPK acts as a regulator in integrin-mediated cancer metastasis, although the detailed mechanisms remain to be investigated.
This work was supported by the National Natural Science Foundation of China [Grants 30801406, 30873096], Zhejiang Natural Science Foundation [Grant R2080326], and Zhejiang Provincial Programme for the Cultivation of High-Level Innovative Health Talents.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- extracellular matrix
- focal adhesion kinase
- heat shock protein
- tumor necrosis factor
- nuclear factor-κB
- mitogen-activated protein kinase
- dimethyl sulfoxide
- thiazolyl blue tetrazolium bromide
- phosphate-buffered saline
- extracellular signal-regulated kinase
- green fluorescent protein
- matrix metalloproteinase.
- Received January 8, 2010.
- Accepted May 13, 2010.
- Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics