Background
Malignant tumor cells produce various growth factors that induce angiogenesis to supply nutrition for their own growth. Thus molecules that inhibit angiogenesis are good candidates for anti-tumor agents [
1]. Indeed, some studies in which angiogenesis was targeted have provided encouraging results. Recently, however, it was reported that monotherapy with the monoclonal antibody bevacizmab, which targets vascular endothelial growth factor (VEGF), or an endogenous anti-angiogenic agent such as endostatin produced only moderate suppression of tumor growth compared to a combined therapy that included a cytotoxic agent [
2,
3]. These observations suggest that a molecule with both cytotoxic and anti-angiogenic activities may have a stronger anti-cancer effect. However, such a molecule has not been identified.
Chondromodulin-1 (ChM1) is a 25 kDa glycoprotein that is expressed mainly in cartilage. ChM1 shows anti-angiogenic activity and has been suggested to inhibit endothelial cells from invading cartilage [
4‐
6]. Recently, we reported that the ectopic administration of ChM1 dramatically suppresses tumorigenesis
in vivo [
7], which suggests that ChM1 acts directly against tumor cells. ChM1 can have either a positive or negative effect on cell proliferation: It promotes the proliferation of chondrocytes and osteoblasts [
4,
6,
8], but suppresses growth of endothelial cells and T-cells [
5,
9].
ChM1 also promotes anchorage-independent growth of chondrocytes [
6]. Anchorage-independent growth is a characteristic of non-adherent cells, including oncocytes [
10], chondrocytes [
11‐
14], and hemocytes [
15,
16]. On the other hand, transforming growth factor β (TGFβ) also modulates cell growth both positively and negatively. TGFβ promotes anchorage-independent growth of chondrocytes [
6,
11,
12], but suppresses or promotes anchorage-independent growth of tumor cells depending on the type and state of the cells [
14,
17‐
19]. Thus, ChM1 may also suppress tumor cell growth.
Anchorage-dependent signaling involves extracellular matrix-integrin complexes and their downstream molecules such as Erk, Akt, and GSK3β, which are shared with the signaling pathway activated by cytokine receptor stimulation [
10,
20‐
22]. Abnormality in this signaling pathway, of tumor suppressor proteins, or a combination of both, constitutively activates oncocytes, thereby inducing anchorage-independent tumor growth. The cytokine signaling pathway involving the Signal Transducers and Activators of Transcription (STAT) protein [
3,
23‐
27], a latent transcriptional factor activated by the Janus Kinase (JAK) family of tyrosine kinase, is also modified in various types of tumor cell. Therefore, the STAT signaling pathway may also be involved in the putative action of ChM1.
The aim of the present study was to determine whether ChM1 has a direct action on tumor cells. Here, we report that ChM1 directly suppresses tumor cell anchorage-independent growth by inhibiting the anchorage-independent STAT signaling pathway.
Methods
Reagents and antibodies
Anti-ChM1 polyclonal antibody, kindly provided by Dr. Hiraki (Department of Molecular Interaction and Tissue Engineering, Institute for Frontier Sciences, Kyoto University, Kyoto, Japan), was used for western blot analysis [
28]. Other primary antibodies were purchased from Cell Signaling Technology Inc (Boston, Massachusetts,) to detect cell-cycle related proteins, Erk (p44/p42), Akt, GSK3β, and their phosphorylated forms. The constructs of pSTAT RE-TK hRluc (STAT-luc), pISRE RE-TK hRluc (ISRE-luc), and pGAS RE-TK hRluc (GAS-luc), provided by Dr. Yokoyama K., and obtained from RIKEN BioResource Center, Tsukuba, Japan, were used for the luciferase reporter assays.
Cell culture
HEK 293, HepG2 (human hepatocellular carcinoma), HeLa (human uterine cervical adenocarcinoma), and PC-3 (human prostate adenocarcinoma) cells were obtained from the American Type Culture Collection (Rockville, MD). The human osteosarcoma cell line, NOS-1, which is osteoid inducible in xenografted tumors in nude mice, was established previously from a 16 year-old male Japanese patient [
29,
30]. HepG2 and HeLa cells were cultured in DMEM (Gibco/Invitrogen, Grand Island, NY), PC-3 cells in Ham's F12K (Sigma-Aldrich, Inc., St. Louis, MO), and NOS-1 cells in RPMI (Gibco) supplemented with 10% fetal bovine serum (FBS) at 37°C under 5% CO
2 in air. Human umbilical vein endothelial cells (HUVECs) and normal human dermal fibroblasts (NHDFs) were obtained commercially (Cambrex, Walkersville, MD). HUVECs were grown in EGM2 medium (EBM complete medium with supplements; Sanko-Junyaku, Tokyo, Japan), and NHDFs in FGM2 medium (FBM medium with supplements; Sanko-Junyaku) at 37°C under 5% CO
2 in air. Cells were used at passages 2 through 4 after acquisition.
DNA synthesis assay (BrdU incorporation)
HUVECs and NHDFs were harvested with trypsin/EDTA and suspended in EGM2 and FGM2 as appropriate. The cells were seeded at 3 × 104 cells/ml into a 96-well multi titer plate (100 μl/well) and cultured for 24 hours. The cells were then starved in 0.5% FBS containing Opti-MEM for 12 hours and stimulated with 10 ng/ml FGF-2 (PeproTech EC, London, UK) in either the presence or absence of 25 μg/ml rhChM1 for another 24 hours. Cells were labeled with BrdU during the last 3 hours of this incubation. HepG2 cells were harvested with trypsin/EDTA and suspended at a density of 5 × 103 cells/ml in 10% FBS containing DMEM. HeLa cells were harvested similarly and suspended at a density of 6 × 104 cells/ml. Cells were then seeded into a 96-well multi titer plate (100 μl per well), and cultured for an additional 36 hours. The medium was replaced with one containing either 10 μg/ml or 25 μg/ml rhChM1, BrdU was added, and the cells were cultured for 6, 12 or 24 hours. BrdU incorporation by the cells was measured at least in triplicate at each time point using a cell proliferation ELISA BrdU colorimetric kit according to the manufacturer's instructions (Roche, Basel, Switzerland). Absorbances at 450 nm, referenced at 655 nm, were measured using a Model 680 Microplate Reader (Bio-Rad, Hercules, CA)
Adenovirus preparation
The human ChM1 cDNA expression vector (pcDNA3-ChM1) was provided by Dr. Hiraki (Department of Molecular Interaction and Tissue Engineering, Institute for Frontier Sciences, Kyoto University, Kyoto, Japan). This cDNA was inserted into a cassette cosmid carrying an adenovirus type-5 genome lacking the E1A, E1B and E3 regions, and in which the Swa I cloning site is flanked by the CAG promoter at the 5' end and by a rabbit globin poly (A) sequence at the 3' end (Takara Bio Inc., Shiga, Japan). In 293 cells, recombination between the homologous regions of the linearized transfer cosmid vector and the adenovirus genome resulted in formation of the complete adenoviral recombinant that contains the ChM1 cDNA. Before use in experiments, the adenovirus was purified by sequential centrifugation in double CsCl step gradients as previously described [
31]. Titers of viral stocks were determined by a plaque assay of 293 cells. Viral suspensions were stored at -80°C. The virus was thawed on ice prior to use.
Adenovirus treatment in vivo
Six- to eight-week old BALB/c athymic nude mice (Clea Japan Inc., Tokyo, Japan) were used. Animal experiments were performed in accordance with the institutional guidelines of the university committee on the use and care of animals. Mice were inoculated with 5 × 106 HepG2 cells in the flank and tumors were allowed to grow to a volume of 150 mm3. Animals were divided into three treatment groups: Ad-ChM1 injection (n = 6); Ad-LacZ injection (n = 6); and injection of control vehicle (Phosphate-Buffered Saline, PBS) (n = 6). Adenovirus vectors (1 × 108 plaque forming units/100 μl) were injected directly into the foci center on days 0, 2 and 4 of treatment. Tumor length and width were measured with calipers over a period of five weeks. Tumor volume was calculated as (length × width2)/2.
Counting the number of total cells and viable cells in vitro
Approximately 0.5~2.5 × 104 cells were plated onto 35 mm culture plates and cultured for 24 hrs. Cells were then infected with Ad-ChM1 or Ad-LacZ as a control, at an appropriate multiplicity of infection (MOI) and were further cultured. The MOI for each cell line was selected to produce the optimum effect of ChM1 without cytotoxicity by Ad-LacZ. The total number of cells was counted using a hemocytometer at 24 hrs, 36 hrs, 48 hrs, and 72 hrs after infection with adenoviruses. Viable cells were identified using the trypan blue exclusion method and were counted at each sampling interval. These experiments were carried out at least in triplicate.
Anchorage-independent growth assay
HepG2 and HeLa cells were cultured on 35 mm culture plates and infected with adenoviruses as described above. Six hours (6 hrs) after adenovirus infection, colony formation assays were performed. The cells were detached and suspended in a culture medium containing 0.68% malting agar (Sigma-Aldrich). The cell suspension was then plated on culture medium containing 0.4% agarose that had been allowed to harden beforehand. The cells were cultured in a volume of 300 μl for 21 days with changes to fresh medium every 3 to 4 days. The numbers and sizes of the colonies were measured under a phase contrast microscope on days 7, 14, and 21 of culture. The experiment was carried out in triplicate.
Western blotting
To detect ChM1 and cell-cycle related proteins, the culture medium was collected and subjected to trichloroacetic acid (TCA) precipitation. The pellet produced by TCA precipitation was resolved in RIPA buffer containing a protease inhibitor cocktail (Sigma-Aldrich) and PMSF (500:1:1). For whole cell extracts, cells were scraped, lysed with RIPA buffer, and the lysate diluted with an equal volume of buffer containing 2-mercaptoethanol. Xenografted tumor specimens were harvested 48 hrs after adenoviral infection, followed by homogenization in lysis buffer (8 M Urea, 0.2% sodium dodecyl sulfate, 0.8% Triton X-100, 3% 2-mercaptoethanol). Insoluble materials were removed by centrifugation. The supernatants were boiled for 3 min and stored at -20°C. For SDS-PAGE, proteins (10~80 μg/lane) were run on 8~15% polyacrylamide slab gels and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% skimmed milk in TBS with 0.1% Tween 20 (TBS-T) for 1 hr at room temperature, and incubated overnight at 4°C with primary antibodies diluted at 1:1000 to 1:4000. After washing at least twice in TBS-T, the membranes were incubated with a horseradish peroxidase-conjugated secondary antibody (diluted 1:25000 with TBS-T) at room temperature for 1 hour. The membrane was washed twice with TBS-T and immunoreactivity was visualized using the Immobilone western blotting detection system (Millipore, Billerica, MA).
Films of cell-cycle related proteins were developed and scanned, and bands were analyzed as a ratio of target protein/α-tubulin control using the Scion Image for Windows program, version 4.02.
Luciferase reporter assay
Three reporter constructs were obtained from RIKEN BRC (Tsukuba, Japan). The nucleotide sequences of the response elements were as follows: 5'-gatccagttcccgtcaatcg-3' for STAT, 5'-gatccagaaacaaaaacaag-3' for ISRE, and 5'-gatccttccgggaattctgggaag-3' for GAS. These constructs express Renilla luciferase. We prepared a reference construct by digesting the HSV-TK promoter, between the BamH1 site and Hind III sites, from the pRL-TK vector (Promega Corporation, Madison, WI, USA) that expresses Renilla luciferase, and cloning this fragment into the pGL4.18 [luc2p/Neo] vector (Promega) that expresses Firefly luciferase.
Cells were infected with virus and cultured for 12 hours. They were then washed twice with culture medium and then transfected with various luciferase expression vectors by the lipofection method using Fugene-6 (Roche Diagnostics). Twenty-four hours after transfection, the cells were harvested and a Dual-Luciferase TM reporter assay system (Promega) was used for sequential measurement of Firefly and Renilla luciferase activities using the specific substrates beetle luciferin and coelenterazine, respectively. Quantification of luciferase activities and calculation of relative ratios were carried out using a luminometer (TD-20/20, Turner Designs, Sunnyvale, CA). In these experiments, at least three independent transfections were performed.
Statistical analysis
Student's t-test was performed for group comparisons of BrdU incorporation on HUVECs and NHDFs. Fisher's protected least significant difference procedure was performed after repeated-measures analysis of variance (ANOVA) for group comparisons of BrdU incorporation on HepG2 and HeLa cells, and for the comparison of luciferase reporter assays on HepG2, HeLa, and HUVECs.
Discussion
Previously, we reported that rhChM1 inhibits growth of chondrosarcomas
in vivo [
7], but our understanding at that time was that the mechanism of the inhibitory effect was solely due to the anti-angiogenic activity of ChM1. In this study, we demonstrated that ChM1 has
in vivo and
in vitro anti-tumor activity against the hepatocyte tumor cells, HepG2, and that the effect is due not only to its anti-angiogenic activity but also to direct inhibition of tumor cell growth. Moreover, our results showed that the Jak/STAT signaling pathway is one of the targets of ChM1 action.
Monotherapy with the anti-VEGF antibody, bevacizmab, or an endogenous anti-angiogenic agent such as endostatin caused only a moderate suppression of tumor growth compared with a combined therapy with a cytotoxic agent [
2,
3]. These results indicate that a molecule with both anti-angiogenic and direct cytotoxic activity should be superior for the treatment of patients with malignant tumors. In this regard, our finding that ChM1 has the ability not only to inhibit angiogenesis, but also to inhibit tumor growth is of interest. ChM1 is the first example of an endogenous molecule with both anti-angiogenic and cytotoxic activities and our results suggest that this molecule warrants further
in vivo study in the future.
In addition to its anti-angiogenic activity, ChM1 is also known to have chondrocyte modulating activity [
4,
6], bone remodeling activity [
34], and T-cell suppressing activity [
9]. In particular, ChM1 also promotes the anchorage-independent growth of chondrocytes [
6]. Anchorage-independent growth is a characteristic of non-adherent cells, including oncocytes [
10], chondrocytes [
11‐
14], and hemocytes [
15,
16]. As is shown in Figure
2, the growth of HeLa cells cultured on plates was not affected by ChM1, whereas the growth of HepG2, PC-3 and NOS-1 cells was significantly suppressed. In contrast, the growth of HeLa cells cultured in soft agarose gel was suppressed by ChM1 in a similar fashion to HepG2 cells, although the effect on HeLa cells was slightly less (Figure
3B). These data indicate that ChM1 inhibits the anchorage-independent growth of tumor cells.
Moreover, our observations also provide some suggestion as to why the results of plate culture produces conflicted with those obtained from soft agarose gel culture. The luciferase reporter assay, carried out on cells cultured on plates, demonstrated that ChM1 suppressed the promoter activity of STAT-luc and GAS-luc in HeLa cells to a similar extent as in HepG2 cells and HUVECs. This appears to be inconsistent with the fact that ChM1 inhibited the growth of HepG2, but not HeLa cells cultured on plates. When the basal promoter activities of STAT-luc and GAS-luc were examined, however, HepG2 cells were found to have the highest levels, followed by HUVECs. In contrast, the basal levels of HeLa cells were much lower than that of the other cells(data not shown). Thus, the basal promoter activities of STAT-luc and GAS-luc may be negligible in HeLa cells. Taken together with the observation that the growth of HeLa cells on plates was not affected by ChM1, these data suggest that ChM1 inhibits the anchorage-independent growth of cells, and, therefore, its effect on cells cultured in soft agarose gel may be achieved by inhibition of the Jak/STAT pathway. When cells are cultured on plates, however, the effect of ChM1 on cell growth varies depending upon the degree to which the cells rely on the Jak/STAT pathway for growth. Thus, the growth of HeLa cells cultured on plates was unaffected by ChM1, since anchorage-dependent growth plus the anchorage-independent non-Jak/STAT pathway may contribute to growth. This explanation is consistent with our observation that phosphorylation of Akt, Erk and GSK3β, signaling molecules downstream of integrin-mediated signal transduction [
10,
20‐
22] and the anchorage-independent non-Jak/STAT pathway, was not affected by ChM1.
However, it is unclear how ChM1 activates intracellular signaling pathways and whether there are specific receptors for ChM1. We have shown that ChM1 suppresses the promoter activity of STAT-luc and GAS-luc, but not of ISRE-luc. ChM1 may act through one or more of the following mechanisms: 1) by recruiting protein tyrosine phosphatase family members such as SHP which inactivate Jak; 2) by recruiting SOCS and/or PIAS to degrade STAT dimers; or 3) by directly or indirectly inhibiting cofactors that form complexes with STAT dimers [
24,
35]. Obviously, further study is required to examine these mechanisms.
The cytotoxic action of ChM1 may be due to growth arrest, apoptosis or a combination of both. Our results strongly indicate that ChM1 mainly causes growth arrest. First, ChM1 inhibited DNA synthesis (Figure
1) and suppressed cell proliferation during culture on plates (Figure
2C), as well as in soft agar (Figure
3). Second, ChM1 down-regulated proteins such as cyclinD1, cyclinD3, and cdk6 that promote cell division, and up-regulated cdk inhibitors such as p21
cip1 (Figure
2D). Third, cells treated with ChM1 were mostly viable and the number of apoptotic cells was negligible throughout the culture period (data not shown). Taken together, these data suggest that the cytotoxic effect of ChM1 is mainly due to cell cycle arrest, and that apoptosis does not play an important role, if any. To some extent, our data contradict a recent observation that ChM1 induces apoptosis of vascular endothelial cells [
36]. The reasons for this inconsistency are not clear at present, but may be due to the use of different cell types and/or experimental conditions in the two studies. Possibly, the effect of ChM1 varies between cell types depending on differences in cell cycle regulation and the balance of signaling pathways that can be directly or indirectly affected by the protein (see discussion above). Our study suggests that ChM1 suppresses the growth of tumor cells by directly arresting the cell cycle and that apoptosis does not play a major role.
Conclusion
We have demonstrated that ChM1 produces an anti-tumor effect not only by inhibiting angiogenesis but also by inducing growth arrest of tumor cells, and by directly suppressing the proliferation of tumor cells in an anchorage-independent manner. However, ChM1 did not alter the phosphorylation of the downstream molecules at which the signaling pathways through receptors for growth factors and cytokines converge with the anchorage-dependent pathway. The mechanism of the induced growth arrest appears to involve the anchorage-independent Jak/STAT pathway.
ChM1 is the first example of an endogenous molecule that possesses two different anti-tumor actions. Our results clearly indicate that this molecule warrants further study in vivo.
Acknowledgements
Dr. Hiraki Y., Department of Molecular Interaction and Tissue Engineering, for providing the polyclonal antibody and for valuable suggestion for an experimental technique.
pSTAT RE-TK hRluc, pISRE RE-TK hRluc, and pGAS RE-TK hRluc deposited by Dr. Yokoyama K. were provided by the DNA Bank, RIKEN BioResource Center, Tsukuba, Japan, with the support of the National Bio-Resources Project of Ministry of Education, culture, Sports, Science and Technology, Japan(MEXT).
Dr. Matsuda J. PhD, Venture Business Laboratory, Niigata University, Niigata, Japan, provided valuable opinions and techniques about producing and the arrangement of the figure.
This study was supported by a Grant for Promotion of Niigata University Research Project (to HK), by a grant from the Japan Orthopaedics and Traumatology Foundation, Inc. (to NE and AO) and by a Grant-in Aid from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) (No.16390531) (to HK).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
HM and HK conceived the project and contributed to its design, HM mainly performed the experiments and wrote/reviewed/edited the manuscript, HK performed adenovirus preparation and animal experiments, and co-wrote part of the manuscript. TY and OI supervised the project and contributed to its design, analysis and interpretation of the data. MMA supported cell culture and western blotting analysis, interpreted the data, for which they are due special thanks. TH, HK, HY supervised and supported the project particularly aspects involving ChM1. NK contributed to amplification of the adenovirus and suggested the experiment. AO contributed to the finance of the project. NE contributed to the finance of the project, conceived and supervised the project. HK contributed to organization of the project, analyzed and interpreted the data and co-wrote the manuscript. All authors read and approved the final manuscript.