The targeting of αvβ3 is a promising therapeutic strategy for suppressing tumor metastasis. However, it is unclear whether the therapeutic efficacy could be influenced by metastasis-promoting factor(s) in vivo. Here we report that Toll-like receptor 4 (TLR4) ligand released from damaged tumor cells or bacteria had a negative effect on the therapeutic effect of a recombinant CBD-HepII polypeptide of fibronectin (CH50) that suppresses tumor metastasis by targeting αvβ3. The TLR4 ligand could antagonize the inhibitory effect of CH50 on tumor cell adhesion and invasion by promoting the expression and activity of αvβ3 in tumor cells. The TLR4 ligand also reduced the antimetastasis effect of CH50 by promoting tumor cell survival in circulation. Moreover, TLR4 ligands released by tumor cells in circulation could increase the survival and proliferation capacity of tumor cells after extravasation, resulting in the formation of more metastatic nodules. The effect of TLR4 signaling was mainly mediated by nuclear factor-κB (NF-κB). Inhibiting NF-κB could abrogate the negative effect of TLR4 ligand, and augment the inhibitory effect of CH50 on tumor metastasis. Consistently, the combination of NF-κB inhibitor and CH50 significantly inhibited metastasis of tumor cells in vivo and prolonged the survival of mice. The findings in this study suggest that the combination of NF-κB inhibitor and αvβ3 antagonist would be a novel therapeutic option for the prevention of tumor metastasis. (Cancer Sci 2012; 103: 1319–1326)
Metastases, rather than primary tumors, are responsible for most cancer deaths. For better efficacy of tumor therapy, there is an urgent need for the development of therapeutic strategies to suppress tumor metastasis. The process of tumor metastasis consists of a series of steps, all of which must be successfully completed to give rise to a metastatic tumor. Therefore, inhibiting the key step in the metastatic process could effectively inhibit tumor metastasis. The development of antimetastasis drugs has been specifically aimed at the important molecules associated with tumor cell migration and invasion. Some targets for the antimetastasis therapy have been found in recent years.[2-5] Among them, integrin αvβ3 has unique advantages as a therapeutic target, as it could mediate tumor cell invasion, arrest in circulation, and extravasation into distant organs.[6, 7] Moreover, targeting αvβ3 not only inhibits tumor cell metastasis but also suppresses tumor angiogenesis.[8-10] Antagonists of αvβ3 have been developed in the past decade.[11-15] The preparation and design of small peptides, antibodies, and antagonists targeting αvβ3 are still in progress.[16-18]
Although targeting αvβ3 is effective in suppressing tumor metastasis in animal models,[16, 19] the successful treatment of tumor metastasis by targeting αvβ3 is seldom reported in clinical trials. The existence of metastasis-promoting factor(s) in vivo might be an important reason for the unsatisfactory efficacy. Therefore, for better efficacy of antimetastasis therapy by targeting αvβ3, it is important to inhibit the effect of the factor(s) that could antagonize the effect of therapeutic agents targeting αvβ3.
Toll-like receptor 4 (TLR4) is a pattern recognition receptor expressed on immune cells and tumor cells.[20-22] It has been found that silencing TLR4 in tumor cells could reduce the metastatic potential of tumor cells. A well known TLR4 ligand, LPS, has been found to promote tumor metastasis after surgery. Given that the TLR4 ligand could exist in vivo due to surgery, damage of tumor cells, or the existence of bacteria in tumor,[22, 24-26] the TLR4 ligand may have a negative effect on antimetastasis therapy based on the targeting of αvβ3.
In this study, we investigated the impact of TLR4 ligand on antimetastasis therapy based on targeting αvβ3. Our results showed that TLR4 signaling augmented the adhesion and invasion capacity of tumor cells by promoting the expression and activity of αvβ3, and also increased the apoptosis resistance of tumor cells, thus to attenuate the therapeutic effect of a recombinant CBD-HepII polypeptide of fibronectin (CH50) that inhibits tumor metastasis by targeting αvβ3.[27, 28] The activation of nuclear factor-κB (NF-κB) was crucial for the effect of TLR4 signaling. Inhibiting NF-κB could abolish the negative effect of TLR4 ligand on CH50 treatment, suggesting that NF-κB inhibitor might be very important in improving the therapeutic efficacy of tumor metastasis by targeting αvβ3.
Materials and Methods
Animals and cell lines
Female C57BL/6 mice, 6–8 weeks old, were purchased from the Center of Medical Experimental Animals of Hubei Province (Wuhan, China) for studies approved by the Animal Care and Use Committee of Tongji Medical College (Huazhong University of Science and Technology, Wuhan, China). The C57BL/6 background melanoma B16F1 cell line was purchased from the China Center for Type Culture Collection (Wuhan, China) and cultured according to their guidelines.
Reagents and plasmids
Lipopolysaccharide, resveratrol, and fibrinogen were purchased from Sigma-Aldrich (St. Louis, MO, USA). 6-amino-4-(4-phenoxyphenylethylamino) quinazoline (QNZ), Bay 11-7082, SB203580, and wortmannin were purchased from Merck4Biosciences (Calbiochem, Darmstadt, Germany). Bortezomib was purchased from Millennium Pharmaceuticals (Cambridge, MA, USA). Recombinant polypeptide CH50 was prepared as described previously. Plasmid pCH510, a eukaryotic expression vector of CH50, was constructed and preserved in our laboratory.
Preparation of molecules from damaged tumor cells
Tumor cells were washed and suspended in PBS to a final concentration of 5 × 107/mL. After four rounds of freeze–thaw cycles followed by vortexing for 30 s, the cells were removed by centrifugation. The supernatant contained a mixture of molecules from damaged tumor cells (DTC-Ms). The concentration of DTC-Ms was defined by the concentration of protein, which was determined using Coomassie Bradford reagent (Thermo Fisher Scientific, Rockford, IL, USA).
Matrigel invasion assay
B16 cells (2.5×105) were placed in the upper compartments of modified Boyden chambers (Transwell; Corning, Corning, NY, USA) in which the Transwell filter inserts were coated with Matrigel. After a 24-h incubation, B16 cells that had migrated to the lower surface of the membranes were fixed, stained, and counted under a microscope from randomly chosen fields. The average number of cells per field was calculated.
For downregulation of TLR4 or NF-κB p65 subunit, B16 cells were transduced with TLR4 shRNA or NF-κB p65 shRNA lentiviral particles, or control shRNA lentiviral particles (Santa Cruz Biotechnology, Santa Cruz, CA, USA) according to the manufacturer's protocol. After selection with puromycin, the cells were used for further experiments.
Western blot assay
Western blot assay was done as described previously. Antibodies were purchased from Santa Cruz Biotechnology and Cell Signaling Technology (Danvers, MA, USA).
Luciferase reporter assay
The NF-κB activity in cells was detected with a NF-κB luciferase reporter vector as described previously. Briefly, 1 × 106 B16 cells were transiently transfected with plasmid pSV40κB-luc with Lipofectamine Plus reagents (Invitrogen, Carlsbad, CA, USA). The cells were cotransfected with a β-galactosidase reporter plasmid driven by CMV promoter to normalize experiments for transfection efficiency. The cells were cultured in the presence of QNZ (20 nM), and harvested 24 h after transfection. When indicated, LPS and DTC-Ms were added to the culture for 10 h before harvest. The cells were extracted with reporter lysis buffer (Promega, Madison, WI, USA) for the luciferase assay according to the manufacturer's instructions.
Analysis of gene expression by real-time RT-PCR
Total RNA was extracted from cells with TRIzol reagent (Invitrogen). The relative quantity of mRNA was determined by real-time RT-PCR assay as described previously. The mRNA of GAPDH was used as internal control. The primer sequences were: αv, sense 5′-CTTCTCGGTGGTCCTGGTA-3′, antisense 5′-ATTGCTTGTGCAGTCCGTG-3′; β3, sense 5′-AGAAACAGAGCGTGTCCCGTAA-3′, antisense 5′-TGGGTCTTGGCATCCGTGGT-3′; and GAPDH, sense 5′-GGCAAATTCAACGGCACAGT-3′, antisense 5′-AGATGGTGATGGGCTTCCC-3′.
The B16 cells were added (1 × 105/well) to 96-well plates precoated with fibrinogen. After a 2-h incubation at 37°C and the removal of non-adherent cells, the adherent cells were detected by LDH assay. The results were expressed as A570 values.
Analysis for actin polymerization
The B16 cells were incubated in a Matrigel-coated plate for 5 h. The cells were fixed in 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, then stained with rhodamine–phalloidin (Invitrogen) according to the manufacturer's protocol to visualize polymerized actin.
The B16 cells were cultured (1 × 106/well) for the indicated time in 6-well plates precoated with poly-2-hydroxyethyl methacrylate (poly-HEMA) (10 mg/mL; Sigma). The cells were then stained with phycoerythrin–annexin V (BD Biosciences, San Diego, CA, USA) and analyzed by flow cytometry. For the assay of B16 cell apoptosis in circulation, B16 cells were labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE and injected into the tail veins of mice (2 × 106 per mouse). Five and 12 h later, the whole blood was collected from mice. After lysis of red blood cells, the left cells were stained with phycoerythrin–annexin V and analyzed by flow cytometry. The percentage of apoptotic B16 cells (annexin V+ CFSE+ cells/total CFSE+ cells) was calculated.
Assay of tumor cell arrest in lung
Carboxyfluorescein succinimidyl ester-labeled B16 cells (5 × 105) were injected into the tail veins of mice. Lungs were harvested from mice 5 and 24 h after tumor cell injection. Frozen sections were prepared and analyzed by fluorescence microscopy. Fluorescent spots were counted from randomly chosen fields in the sections of each mouse.
In vivo gene transfection
Plasmid was prepared and analyzed as described previously. The in vivo transfection with plasmid pCH510 or control plasmid pcDNA3.1 was carried out by i.v. injection. Naked plasmid DNA (200 μg in 2 mL saline) was injected into the tail veins of mice within 10–15 s.
Animal experiments and treatment protocols
B16 cells (5 × 105) were injected into mouse tail veins on day 0 (d0). The mice of treatment groups received an i.v. injection of pCH510 (200 μg per injection) or i.p. injection of bortezomib (0.8 mg/kg) or Bay 11-7082 (20 mg/kg), once every 2 days from −d1 to d11. The mice of control groups received the injection of an equal volume of saline or equal amount of pcDNA3.1 plasmid. To evaluate the inhibitory effect of CH50 and bortezomib in the process of tumor metastasis, mice were inoculated by i.v. injection of 5 × 105 B16 cells on d0, followed by i.v. injection of 2 × 105 B16 cells on d3 and d6. The mice of the treatment group received an i.v. injection of pCH510 on −d2, d1, and d4, and/or i.p. injection of bortezomib on −d1, d2, and d5. Saline and pcDNA3.1 plasmid were used as control. The mice were killed at the indicated time. The tumor nodules on the surface of lung were counted. In a parallel survival rate follow-up experiment, the number of living mice was recorded.
Soft agar assay
B16 cells were cultured (1 × 106/well) for 24 h in 6-well plates precoated with poly-HEMA. The cells were harvested and resuspended in 0.33% agar medium containing 10% FBS. One milliliter of the cell suspension containing 104 cells was plated over the bottom 0.6% agar layer in 6-well plates in triplicate. Then 14 days later, the colonies of B16 cells were counted under a microscope.
Results are expressed as the mean value ± SD and interpreted by one-way anova. Differences were considered to be statistically significant when P < 0.05.
Nuclear factor-κB plays crucial role in promoting effect of TLR4 signaling on invasive migration of melanoma cells
To investigate whether TLR4 ligand may influence the inhibitory effect of CH50 on tumor metastasis, we first analyzed whether TLR4 ligand could promote the invasive migration of B16 melanoma cells. B16 cells were treated with LPS, a well known TLR4 ligand released from bacteria, and the DTC-Ms that contain multiple endogenous TLR4 ligands. The treatment of B16 cells with either LPS or DTC-Ms could significantly promote the invasive migration of the cells (Fig. 1a). The effect of DTC-Ms was mediated by TLR4 signaling, as downregulating TLR4 expression in B16 cells abrogated the effect of both LPS and DTC-Ms (Fig. 1b,c). To ascertain the crucial signaling pathway for the effect of TLR4 signaling, we treated B16 cells with LPS in the presence of QNZ (NF-κB inhibitor), resveratrol (TRIF inhibitor), wortmannin (PI3K inhibitor), and SB203580 (p38MAPK inhibitor). Of these, only QNZ completely abrogated the effect of LPS (Fig. 1d). At the concentration used in the experiments, QNZ was not toxic to B16 cells (data not shown), but inhibited the NF-κB activity in B16 cells and abrogated the effect of LPS and DTC-Ms (Fig. 1e). The crucial role of NF-κB was further proved by downregulating the expression of the NF-κB p65 subunit in B16 cells (Fig. 1f,g). Importantly, LPS and DTC-Ms antagonized the inhibitory effect of CH50 on the invasion capacity of B16 cells in the absence, but not presence, of QNZ (Fig. 1h). These results suggested that NF-κB was crucial for both basal invasive migration of B16 cells and the promoting effect of TLR4 signaling on invasive migration of B16 cells.
Toll-like receptor 4 ligand augments the expression and activity of αvβ3 by activating NF-κB
CH50 inhibits tumor metastasis by interfering with the expression and activity of αvβ3.[27, 28] We next wondered whether TLR4 ligand may antagonize the effect of CH50 by influencing the expression and activity of αvβ3. For this purpose, we treated B16 cells with LPS or DTC-Ms in the absence or presence of CH50. The expressions of αv and β3 were increased by LPS and DTC-Ms (Fig. 2a). CH50 suppressed the expressions of αv and β3, whereas its effect was antagonized by LPS and DTC-Ms (Fig. 2b). Inhibiting NF-κB with QNZ completely abrogated the effect of LPS and DTC-Ms. Consistent with the effect on αv and β3 expressions, LPS and DTC-Ms increased the adhesion of B16 cells to fibrinogen, a ligand for integrin αvβ3 but not other integrins, and antagonized the inhibitory effect of CH50 (Fig. 2c), suggesting that the activity of αvβ3 was increased by LPS and DTC-Ms. Moreover, LPS and DTC-Ms also increased the polymerization of actin in B16 cells in response to ECM molecules, and reduced the inhibitory effect of CH50 (Fig. 2d). QNZ also abrogated the effect of LPS and DTC-Ms on the adhesion and actin polymerization of B16 cells. Therefore, TLR4 ligand could antagonize the inhibitory effect of CH50 on the expression and activity of αvβ3 integrin by activating NF-κB.
Toll-like receptor 4 ligand enhances survival of tumor cells in circulation by activating NF-κB
In addition to the expression and activity of αvβ3, prolonging the survival of disseminating tumor cells in circulation directly increases metastasis. We then examined whether TLR4 ligand could promote tumor cell survival in circulation. The results showed that LPS and DTC-Ms downregulated the expression of Bax, but increased the expression of Bcl-xL in B16 cells (Fig. 3a). Although the expression of Bcl-2 was not significantly influenced, the ratio of Bcl-2/Bax was increased, which is in favor of tumor cell survival. Accordingly, the pretreatment with LPS and DTC-Ms reduced the apoptosis of tumor cells cultured under anchorage-independent conditions (Fig. 3b). In line with this, pretreatment of B16 cells with LPS and DTC-Ms reduced the apoptosis of B16 cells in blood (Fig. 3c). The effect of LPS and DTC-Ms was abrogated by QNZ, suggesting that TLR4 ligand promoted tumor cell survival in circulation by activating NF-κB.
Nuclear factor-κB inhibitor augments the inhibitory effect of CH50 polypeptide on extravasation of tumor cells
Tumor cell arrest within the vasculature and the following invasive migration are the prerequisites for extravasation of tumor cells from the blood. In accordance with the increased survival of tumor cells in blood and the increased expression and activity of αvβ3, LPS and DTC-Ms promoted the arrest and following invasion of B16 cells (Fig. 4a,b), evaluated by the number of fluorescent spots in lung tissues 5 and 24 h after B16 cell injection. Consistently, LPS and DTC-Ms increased the number of metastatic nodules in lung (Fig. 4c). In vivo expression of CH50 suppressed B16 cell arrest and following invasion, and reduced the number of metastatic nodules in lung, whereas pretreatment of B16 cells with LPS and DTC-Ms could antagonize the inhibitory effect of CH50 (Fig. 4e). Inhibiting NF-κB with QNZ completely abrogated the effect of LPS and DTC-Ms (Fig. 4d,e), suggesting that inhibiting NF-κB could augment the inhibitory effect of CH50 on extravasation of tumor cells.
Nuclear factor-κB inhibitor suppresses the effect of TLR4 ligands in circulation
We next investigated whether the TLR4 ligand released from damaged tumor cells in blood might also influence the inhibitory effect of CH50 on tumor metastasis. To do this, we injected mitomycin C (MMC)-treated B16 cells into the tail veins of mice. High-mobility group box 1 protein (HMGB1) and heat shock protein 70 (Hsp70) were detected in blood after i.v. injection of MMC-treated B16 cells (Fig. 5a), indicating the release of TLR4 ligand from tumor cells. We then injected CFSE-labeled B16 cells the tail veins of mice. Interestingly, pre-injection of MMC-treated B16 cells did not influence the retention of B16 cells in lung or the inhibitory effect of CH50 on tumor cell retention (Fig. 5b), but resulted in the increase of metastatic nodules in lung (Fig. 5c). We then cultured B16 cells for 24 h under anchorage-independent conditions in the absence or presence of LPS or DTC-Ms. Neither LPS nor DTC-Ms influenced the apoptosis of tumor cells within 24 h (data not shown). However, the cells cultured in the presence of LPS and DTC-Ms formed many more colonies after further culture in soft agar (Fig. 5d), indicating that the survival and proliferation capacity of tumor cells was increased by stimulation with TLR4 ligand under anchorage-independent conditions. Inhibiting NF-κB with QNZ abrogated the effect of LPS and DTC-Ms. Consistently, the negative effect of pre-injection of MMC-treated B16 cells on CH50 treatment was abrogated by bortezomib, a clinically used drug that inhibits the activation of NF-κB, and Bay 11-7082, which also inhibits NF-κB activation (Fig. 5e, Table S1).
Nuclear factor-κB inhibitor and CH50 cooperatively inhibit metastasis of tumor cells in vivo
We next investigated whether targeting both NF-κB and αvβ3 than targeting NF-κB or αvβ3 alone in the process of tumor metastasis might be more effective in suppressing tumor metastasis. For this purpose, we used a model of repetitive injection of B16 cells into the tail vein, and stopped the treatment 1 day before last injection of tumor cells. The repetitive injection of B16 cells significantly increased the number of tumor nodules in the lung of mice. In vivo expression of CH50 or injection of bortezomib alone could inhibit the metastasis of tumor cells, reducing the number of metastatic nodules in the lung (Fig. 6a). However, the combination of CH50 with bortezomib was much more effective, evaluated by a further reduced number of metastatic nodules in lung (Fig. 6a) and prolonged survival of mice (Fig. 6b). Therefore, targeting both αvβ3 and NF-κB was much more effective than targeting NF-κB or αvβ3 alone in preventing tumor metastasis.
Our present findings strongly suggest that inhibiting NF-κB in tumor cells could augment the inhibitory effect of the drugs targeting αvβ3 on tumor metastasis. Targeting αvβ3 is a promising strategy to suppress tumor metastasis, as αvβ3 mediates tumor cell invasion, arrest in circulation, and extravasation into distant organs.[6, 7] Our previous study showed that recombinant polypeptide CH50 could suppress tumor metastasis and tumor angiogenesis by blocking αvβ3 and reducing αvβ3 expression.[27, 28] In this study, we found that inhibiting TLR4 signaling in tumor cells could further augment the inhibitory effect of CH50 on tumor metastasis. Toll-like receptor 4 signaling could directly promote the expression and activity of αvβ3 in tumor cells, thus augmenting αvβ3-mediated adhesion and invasion of tumor cells, and antagonizing the inhibitory effect of CH50 on the invasive migration of tumor cells. Therefore, abrogating the effect of TLR4 signaling could augment the inhibitory effect of CH50 on tumor metastasis. In recent years, many αvβ3-targeting drugs, including small molecule antagonists and antibodies, have been developed.[11-18] These drugs are usually designed for blocking αvβ3 by simply binding to αvβ3. Abrogating the effect of TLR4 signaling may also increase the therapeutic efficacy of these drugs.
The damage of cells could cause the release of intracellular molecules from cells. Some of these molecules have been identified as TLR4 ligands, including HMGB1, Hsp60, Hsp70, Hsp22, gp96, S100A8, and S100A9.[25, 34, 35] Therefore, the molecules released from damaged tumor cells are the important source of endogenous TLR4 ligand. The TLR4 ligand in the milieu of primary tumor may promote tumor metastasis by promoting intravasation and extravasation of tumor cells. In this study we found that the TLR4 ligand in circulation could also promote tumor metastasis by increasing the survival and proliferation capacity of tumor cells after extravasation. After intravasation, large numbers of tumor cells in circulation could be damaged by chemotherapeutics or shear force in blood. Therefore, TLR4 ligand could be released into blood by these damaged cells. Although the practical concentration of single type TLR4 ligand might be very low in peripheral blood, multiple TLR4 ligands released from tumor cells might increase the actual concentration of TLR4 ligands in blood. Moreover, it has been found that tumor-released S100A8 could bind specifically to endothelial cells, and stimulate endothelial cells to express and release SAA3, another TLR4 ligand, thus increase the concentration of TLR4 ligand in local microvessels of target organs of metastasis. Moreover, TLR4 ligand may also be released to blood by bacteria. The stimulation of tumor cells by TLR4 ligand in circulation might not significantly influence the extravasation of tumor cells, but could promote the formation of tumor nodules by promoting the survival and proliferation capacity of tumor cells. To this end, although the number of tumor cells migrating into the tissue could be reduced by the drugs targeting αvβ3, the increased survival and proliferation capacity of tumor cells could result in the formation of more metastatic nodules based on the same number of tumor cells migrating into the tissue. The survival and proliferation of tumor cells after extravasation are crucial for the success of tumor metastasis. Therefore, inhibiting the effect of TLR4 ligand in circulation may augment the therapeutic effect of antimetastasis drugs that only suppress the invasive migration of tumor cells.
Triggering of TLR4 could activate several signaling pathways in cells, including the MyD88/NF-κB, TRIF, PI3K, and p38MAPK pathways.[38-41] Our data showed that the activation of NF-κB is indispensable for the metastasis-promoting effect of TLR4 ligand. The activation of NF-κB is necessary for tumor cell invasion, apoptosis resistance, and proliferation after extravasation. Inhibiting NF-κB in tumor cells could completely abolish the effect of TLR4 ligand on metastasis. Given that TLR4 ligand could activate dendritic cells, and play important roles in the initiation of antitumor immunity,[42, 43] completely blocking TLR4 ligand in vivo might not be an appropriate strategy in antitumor therapy. Alternatively, targeting NF-κB might be feasible for reducing the promoting effect of TLR4 ligand on tumor metastasis. Some NF-κB inhibitors are currently used in combination with chemotherapy and radiotherapy for cancer treatment.[44, 45] Although NF-κB plays a key role in many physiological processes, the side-effects of NF-κB inhibitor are not as extensive as imagined, probably due to different sensitivities between tumor cells and normal cells, or different delivery efficiencies of the drugs to tumor cells and normal cells. Importantly, our data showed that bortezomib, used in a safe therapeutic dosage, could significantly increase the therapeutic effect of CH50 which inhibits tumor metastasis by targeting αvβ3 integrin.
Given that the release of TLR4 ligands might be an inevitable event in the process of tumor therapy due to the damage of tumor cells by chemotherapy and radiation therapy, or tissue trauma and inflammation caused by surgery, or postoperative infection,[26, 46-48] the negative effect of TLR4 signaling on antimetastasis therapy might be a general obstacle for tumor therapy. Therefore, targeting NF-κB may be crucial for the treatment of cancer with antimetastasis drugs. Treatment with NF-κB inhibitor in vivo may not achieve the same effect as treatment in vitro, due to limited delivery efficiency. Nevertheless, the administration of NF-κB inhibitor could significantly increase the therapeutic effect of drugs targeting αvβ3 integrin, which plays crucial roles in the metastasis of many types of cancer cells. The combination of αvβ3-targeting drugs and NF-κB inhibitor could be a novel strategy for the comprehensive treatment of these types of cancer.
This work was supported by the National Natural Science Foundation of China (Grant Nos. 30830095, 30771974, 30772589), and the National Development Program (973) for Key Basic Research of China (Grant No. 2009CB521806).