Cell lines

Osteosarcoma (Saos-2, U2OS, and SJSA), rhabdomyosarcoma (RH30, RH3 and RD2), leiomyosarcoma (SK-LMS-1), human foreskin fibroblast (HFF), and human skeletal muscle myoblast (HSMM) cell lines were purchased from American Type Culture Collection (ATCC). CW9019, a rhabdomyosarcoma cell line, was a gift from Dr. Fred Barr (Department of Pathology, University of Pennsylvania). All cell lines were maintained in 1× DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicilin/streptomycin/amphotericin B (Fisher Scientific International) at 37°C, aired with 5% CO₂. HSMM cells were grown in SkBM-2 basal medium supplemented with SkGM-2 singleQuots kit according to the manufacture's protocol (Cambrex Bio Science, Walkersville, MD, USA).

Cancer tissue microarray immunohistochemistry

To examine whether Stat3 is activated in rhabdomyosarcoma, osteosarcomas, and other soft-tissue sarcomas, we stained osteosarcoma (n = 113), rhabdomyosarcoma (n = 64) and other soft-tissue sarcoma (n = 87) tissue samples on tissue microarray slides from different providers (Biopathology Center of Columbus Children Research Institute, Biomax, and Cybrdi) using immunohistochemistry with a p-Stat3(Y705)-specific monoclonal antibody (Cell Singling Tech., Danvers, MA). The immunohistochemistry method was described previously [5, 15]. Nuclear p-Stat3 expression levels were scored as 0, 1, 2, and 3 according to the immunohistochemical staining intensity. The nuclear staining intensity was scored on the following scale: 0, no staining; 1, weak staining; 2, moderate staining; and 3, intense staining. Since all the normal tissues stained were scored as 0 and occasionally 1, samples stained with scores 0 or 1 were considered as negative, whereas sarcoma samples with scores 2 and 3 were graded as positive. The intensity of immunostaining was evaluated only when more than 50% nuclei showed p-Stat3 expression. Scoring of immunostaining intensity was completed by two to three independent observers (F. H., G. C., and J. L.). Discrepant scores between the two or three observers were rescored to arrive at a single final score. Light microscopic images were documented using a LEICA DM-4000B fluorescent microscope (Leica Microsystems, Bannockburn, IL) with an attached Diagnostic RT-KE 2 MP digital camera (Diagnostic Instruments, Sterling Heights, MI).

Western blot

Cells were collected at 4°C in cold harvest buffer supplemented with proteinase inhibitor cocktails and spun down at 3000 × g for 5 min. Cell pellets were lysed in RIPA lysis buffer as described previously [15]. Protein concentrations were quantitated using BCA protein assay kit from Pierce, Inc. (Rockford, IL) according to the manufacture's protocol. Fifty or 100 μg of cellular proteins were resolved on 10% PAGE gels in electrophoresis buffer and transferred to Hybond™-p membrane (Amersham Biosciences, Piscataway, NJ) using transfer buffer with a constant 100 V. The membranes were then blocked using 5% nonfat dry milk in TBST (Tris-HCl, pH7.5, Tween, 0.1%) for 30 min at room temperature (RT) and were incubated with primary antibody over night at 4°C or for 1 hour at RT using concentrations recommended by the manufacture. The membranes were washed three times in 1× TBST for 5 min each time. Proteins of interest were visualized using an ECF™ western blotting kit (Amersham Biosciences, Piscataway, NJ) according to the manufacture's protocol. Incubation of secondary antibody and anti-fluorescein were carried out both in presence of 1× TBST with 2% nonfat dry milk. The fluorescent signals were scanned and documented using a Storm 860 scanner (Molecular Dynamics, Sunnyvale, CA). Antibodies were purchased separately and used for Western blots of FLAG (Sigma, St. Louis, MO, USA), GAPDH (Chemicon International, Temecula, CA), Stat3, and p-Stat3 (Y705) (Cell Signaling Tech., Danvers, MA).

Transduction of dominant negative Stat3 Y705F in cancer cells

The construction and infection of recombinant Adenovirus/CMV-dnStat3 Y705F (rAd/dnStat3) is described previously [14, 16]. DnStat3 was generated from Stat3 by changing the tyrosine at position 705 into phenylalanine. Its protein product cannot be activated through tyrosine phosphorylation that is crucial for dimerization. The clone is tagged with a FLAG marker. About 2 × 10⁵ U2OS, SaoS2, SJSA, RD2 and RH30 cells were transduced with rAd/dnStat3 or a negative control viral vector, rAd/CMV-eGFP (rAd/eGFP) (Applied Viromics, Fremont, CA) with multiplicities of infection (moi) of 400, 100, and 10 based on TCID50 using 293T cells. For cell growth experiments, cells in 5 random fields of view (100× magnification) were enumerated on days 2, 4, and 6 post-infection of rAd/eGFP and rAd/dnStat3. Cell growth rates were presented as percentages of untransduced controls. Each data point was averaged from triplicate experiments.

Treatment of STA-21 and cell viability assay

Approximately 5000 RD2 and RH30 cells were grown in 100 μl 10% FBS-supplemented DMEM medium in 96-well flat-bottomed plates overnight. Treated cells were exposed to STA-21 (30 μM) that was dissolved in dimethyl sulfoxide (DMSO) before being added to the medium. Cell viability was analyzed by the MTT [3-(4, 5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2H-tetrazolium bromide] (Sigma, St. Louis, MO, USA) assay in three replicates. At the endpoint, cells were treated with MTT (1 mg/ml) for 3–4 hours. Colorimetric quantification was determined by an EL808 Ultra Micro-plate Reader (Bio-Tek Instruments, Inc) after the addition of formazan dissolved in 25% N, N-dimethylformamide and 10% SDS under light-proof conditions overnight.

Caspases 3, 8, and 9 immuno-fluorescent staining and acridine orange staining

1 × 10⁵ cells (U2OS or RD2) were seeded on sterile coverslips in a 6-well plate overnight. The cells were transduced by either rAd/eGFP or rAd/dnStat3 for 3 or 4 days and then fixed using methanol/acetone (v:v = 1:1). Three washes followed the fixation using 1× PBS buffer. During the third wash, the coverslips were transferred to a new 6-well plate. For immuno-fluorescent staining, the cells were blocked in 1× PBS with 10% normal horse serum for 1 hour and incubated with primary rabbit antibodies that recognize cleaved-caspase-3 (Asp175), cleaved-caspase-8 (Asp374), or cleaved-caspase-9 (Asp330) (Cell Singling Tech., Danvers, MA) with 1:100, 1:50, and 1:100 dilutions, respectively. Excess antibodies were removed using 3 washes of 1× PBS with constant agitation, 10 minutes for each wash. Secondary goat anti-rabbit IgG(H+L) Alexa Fluo^R 594 antibody (Invitrogen, Carlsbad, CA) (1:1000 dilution) was incubated with 1% bovine serum albumin (BSA) in 1× PBS for 1 hour at RT. Unbound antibody was washed off three times using 1× PBS. Nuclei were counter-stained using 4'-6-Diamidino-2-phenylindole (DAPI) (100 ng/ml) in distilled H₂O for 5 min and then rinsed three times with 1× PBS, 10 min for each wash. The cleaved caspase positive cells were scored from three independent fields of view (100× magnification) and presented in averaged percentages of total cells (DAPI staining) with standard deviations from triplicate experiments.

Acridine orange staining was previously described[17]. The cells were incubated with 1 mg/ml acridine orange (Sigma, St. Louis, MO) for 15 min before 3 washes of 1× PBS. The fluorescence and phase-contrast microscopic photographs were documented using LEICA DM-IRB inverted fluorescent microscope (Leica Microsystems, Bannockburn, IL) with an attached Diagnostic RT-SE6 monochrome digital camera (Diagnostic Instruments, Inc, Sterling Heights, MI).

Elevated Stat3 phosphorylation in rhabdomyosarcoma, osteosarcoma and other soft-tissue sarcoma tissues and cell lines

Our results indicated that Stat3 phosphorylation levels were elevated in osteosarcoma, rhabdomyosarcoma and other soft-tissue sarcoma tissues and cell lines (Figure 1 &2). The sarcoma tissues used here were classified according to clinicopathological data in Table 1 and 2. Osteosarcoma tissue microarray slides had a total of 113 specimens. Soft-tissue sarcoma microarray slides had a total of 151 specimens (Table 1). Rhabdomyosarcoma is the most common soft tissue sarcoma of childhood. Based on histological criteria, it can be classified into two major subtypes, alveolar rhabdomyosarcoma (ARMS) and embryonal rhabdomyosarcoma (ERMS). Rhabdomyosarcoma tissue microarray slides had a total of 64 specimens in which 32 of them were ARMS and another 32 specimens were ERMS. The patient ages of these cases were between 0 and 19 (Table 2). Normal tissues did not stain for p-Stat3 (Figure 1A: a &1c) and sarcoma tissues stained positively in nuclei, cytoplasm, or both (Figure 1A: b, d and insets in b, d). The percentages of p-Stat3 positive samples were 19% (21/113) of osteosarcoma, 27% (17/64) of rhabdomyosarcoma, and 15% (22/151) of other soft-tissue sarcoma samples.

We also investigated the status of p-Stat3 in sarcoma cell lines. Analysis of Stat3 phosphorylation in these cells lines was carried out using Western blots with GAPDH as an internal protein loading control (Figure 1B &1C). Western blots with a p-Stat3 specific antibody revealed that Stat3 was phosphorylated in several rhabdomyosarcoma, osteosarcoma, and leiomyosarcoma cell lines. These included RD2, RH30, CW9019, SMS-CTR, Saos-2, SKLMS-1, U2OS, SJSA, as well as IFN-γ-treated HeLa cells serving as a positive control. P-Stat3 levels in RH3 and a negative control cell line, HFF, were relatively low or undetectable. However, these two p-Stat3 negative cell lines contained similar levels of total Stat3 with the other p-Stat3 positive cell lines. Elevated Stat3 phosphorylation crucial for Stat3 activation was observed in most of the sarcoma cell lines we screened.

rAd-mediated transduction of dnStat3 in rhabdomyosarcoma and osteosarcoma cell lines

Since elevated levels of Stat3 phosphorylation was found in sarcoma tissues and cell lines, we subsequently investigated the role activated Stat3 may play in cell growth or survival of sarcoma cell lines. We introduced dnStat3 into rhabdomyosarcoma and osteosarcoma cell lines using an adenoviral vector delivery system. RD2 and SJSA cells were infected with rAd/dnStat3 (moi = 10, 100, and 400). FLAG-tagged dnStat3 expression levels in sarcoma cells were detected in Western blots probed with an anti-FLAG antibody. Two days post-infection, dnStat3 was expressed in SJSA and RD2 cells in a dose-dependent manner, but not in untransduced cells and cells transduced with rAd/eGFP (moi = 400) (Figure 3A &3B). The transduction efficiency of rAd vector on these cells was determined by infection of rAd/eGFP. Greater than 90% of cancer cells showed green fluorescence by day 4 post-infection with rAd/eGFP (moi = 400) (Figure 3C and Saos-2 data not shown).

Targeting Stat3 signaling pathway using dnStat3 and STA-21 suppressed cell growth and viability in rhabdomyosarcoma and osteosarcoma cells

Osteosarcoma and rhabdomyosarcoma cell growth and viability were significantly suppressed in the presence of dnStat3 or STA-21. Sarcoma cells were transduced with either rAd/eGFP or rAd/dnStat3 (moi = 400). Growth of cells with/without transduction was normalized to untransduced controls at days 2, 4 and 6 post-transduction (Figure 4A). The growth rates of untransduced cells were set at 100%. There were minor adverse effects by rAd/eGFP on cell growth as observed in osteosarcoma and rhabdomyosarcoma cells when moi of 400 was used. However, all sarcoma cells transduced with rAd/dnStat3 showed dramatic reduction in cell growth being less than 20 or 40% of untransduced controls (day 4 or 6 post-infection) (Figure 4A). To investigate whether the dnStat3 cell growth inhibition effects were specific to sarcoma cells, a normal human skeletal muscle myoblast cell line (HSMM) was transduced with the same dose of rAd/dnStat3 and rAd/eGFP (moi = 400). Interestingly, at day 6 post infection, there were no observable changes in cell growth and morphology of HSMM (data not shown).

STA-21 is a novel small inhibitor that prevents Stat3 from dimerization, translocation into the nucleus, and its signaling pathway [13]. The cell viability of RH30 and RD2 were also greatly reduced after 4 and 5 days' exposure to STA-21 (Figure 4B). Approximately 12% and 48% untreated cell viability were detected in RH30 and RD2 cells, respectively, exposed to 30 μM STA-21 for 4 days. After 5-day treatment, STA-21 inhibitory effects were even more significant that viabilities of drug-treated RH30 and RD2 cells were only 1.8% and 11% of the untreated, respectively.

Apoptosis induced by dnStat3 and STA-21 through caspase cleavage pathways

Rhabdomyosarcoma and osteosarcoma cell lines transduced with rAd/dnStat3 not only suffered cell growth and viability inhibition but also total cell number reduction suggesting that cell death was induced by the expression of dnStat3 (Figure 4C and data not shown). In addition, massive accumulation of vacuoles occurred in the cytoplasm of these cells (Data not shown). These prompted us to explore if necrosis or apoptosis contribute to the cell death. In order to investigate the mechanism underlying the rhabdomyosarcoma and osteosarcoma cell death, RD2 and U2OS cells were fixed at day 4 post transduction of rAd/eGFP or rAd/dnStat3. The fixed cells were then subjected to acridine orange staining and anti-cleaved-caspase immuno-fluorescent staining for necrosis and apoptosis evaluations, respectively. For acridine orange staining, there was no difference between control cells and transduced cells with rAd/dnStat3 (data not shown) indicating that necrosis is not involved. For apoptosis evaluation, there was no observable difference in cleaved caspase immuno-staining among negative controls (untransduced or transduced with rAd/eGFP) in U2OS and RD2 cells lines (Figure 5A &5B). However, cell death caused by the transduction of dnStat3 appeared to be apoptosis as cleaved caspases 3, 8, and 9 were observed in increased portions of dnStat3-expressing RD2 and U2OS sarcoma cells (Figure 5A, B &5C; cleaved caspase 9 staining for U2OS and cleaved caspase 8 staining for RD2, data not shown). In RD2 cells, 4-day dnStat3 expression induced 23%, 44%, and 52% of cells to undergo apoptosis through cleaved caspases 3, 8 and 9, respectively. Similarly in U2OS cells, transduction of dnStat3 caused caspases 3, 8, and 9 cleavages in 27%, 42%, and 63% of cells, respectively (Figure 5C). The apoptotic effect through the blocking of Stat3 signaling in sarcoma cells was further confirmed using STA-21. STA-21-treated rhabdomyosarcoma cells, RH30 and RD2, showed increased portions (58.2% and 35.4%) of cells undergoing apoptosis through the caspase 3 cleavage pathway (Figure 5D &5E). In contrast, normal HSMM cells were not affected by the STA-21 treatment for the same duration (Figure 5D &5E). Apparently, cytotoxicity to normal cells by targeting the Stat3 signaling pathway would be very minimal.