Combination Effects of Salvianolic Acid B with Low-Dose Celecoxib on Inhibition of Head and Neck Squamous Cell Carcinoma Growth In vitro and In vivo

Head and neck squamous cell carcinoma (HNSCC) represents >90% of all head and neck cancers (1). Although tobacco exposure is the primary etiologic environmental agent associated with HNSCC, development of this tumor is associated with pathways that are also active in the inflammatory response (2). Chronic inflammatory microenvironment gradually leads to cellular proliferation and multiple genetic alterations, inducing early malignant changes from normal epithelial cell to cancer cell (3). Thus, it remains an appealing strategy to develop effective, nontoxic, and affordable novel pharmacologic agents for preventing development of both HNSCC and second primary HNSCC (4–6).

Celecoxib is a nonsteroidal anti-inflammatory drug (NSAID) that helps to control inflammation by inhibiting cyclooxygenase-2 (COX-2) activity (7). Celecoxib was hailed as a promising cancer prevention drug and has been tested in several clinical studies (8–11). However, it was found to be associated with a dose-dependent cardiovascular morbidity, which limited its long-term use as a cancer prevention agent (7, 12–14).

Salvianolic acid B (Sal-B) is a leading bioactive component of Salvia miltiorrhiza Bge (Danshen or Tanshen) and used as a quality control ingredient and active marker for S. miltiorrhiza Bge products by the National Pharmacopoeia Council of China. S. miltiorrhiza Bge is a well-known Chinese herbal medicine that has been used for thousands of years to treat cardiovascular diseases, chronic hepatitis, liver fibrosis, and tumoral diseases without undesirable side effects (5, 15–19). Based on the theory of traditional Chinese medicine, S. miltiorrhiza Bge has its effect by promoting blood circulation to repair damaged tissues. Additionally, S. miltiorrhiza Bge and Sal-B have been shown to inhibit inflammation (5, 19, 20), coagulation (21), oxidation (22), and tumor growth (5, 23–25). Our previous study found that Sal-B widely suppresses proliferation of human cancer cell lines, including HNSCC, prostate, breast, and liver cancers (5). These anticancer activities are through not only selectively inhibiting COX-2 mRNA expression but also effectively inducing apoptosis in cancer cells (5). However, there was a report indicating that Sal-B plays as an antioxidant to modulate hydrogen peroxide–induced apoptosis through the phosphoinositide 3-kinase–Akt–Raf–mitogen-activated protein/extracellular signal-regulated kinase kinase–extracellular signal-regulated kinase pathway (26). As a promising natural anticancer agent, more studies are needed to understand the anticancer mechanism and clinical application.

In recent years, combination treatment has received increased attention (27–29) because it has the potential to enhance the therapeutic effect and reduce toxicity by lowering the dose required for each agent (30). In this study, we tested the hypothesis that combination of Sal-B and low-dose celecoxib may result in a more pronounced anticancer effect in HNSCC cell growth when compared with either agent alone using in vitro and in vivo studies.

Sal-B was isolated from S. miltiorrhiza Bge as previously described (5). Celecoxib was extracted from commercially available 200 mg Celebrex capsules with ethyl acetate aqueous solution (5).

HNSCC cell lines (JHU-06, JHU-011, JHU-013, and JHU-022) were obtained from The Johns Hopkins University. The HNSCC cells were grown in RPMI 1640. The cultures were supplemented with 10% fetal bovine serum and an antibiotic-antimycotic mixture (100 IU/mL penicillin and 100 μg/mL streptomycin; Cellgro). All cells were grown in 5% CO₂ at 37°C and subcultured at an initial density of 1 × 10⁵/mL every 3 to 4 days. All experiments were done with cells in the logarithmic phase of growth.

Cells (10,000 per well density) were seeded in flat-bottomed 96-well cell culture plates in a proper medium with 2% fetal bovine serum. The cells were then treated with 100 μmol/L Sal-B and 10 μmol/L celecoxib individually and with 50 μmol/L Sal-B plus 5 μmol/L celecoxib in combination for 24 hours. Twenty microliters of MTT (5 mg/mL) solution (Sigma) were added to each well at 37°C for 4 hours. After removing the media, 200 μL of DMSO were added and left for 30 minutes at room temperature to dissolve the formazan crystals in shaker. The absorbance at 560 nm was measured in a microplate reader (Bio-Rad).

To analyze cell cycle, JHU-022 cells were cultured without serum for 24 hours and then treated with Sal-B (100 μmol/L), celecoxib (10 μmol/L), or a half-dose of combination for 24 or 48 hours. The treated cells were collected, washed, and fixed in chilled 80% ethanol. The fixed cells were washed twice in PBS and then incubated with a solution containing 100 μg/mL RNase at 37°C water bath for 30 to 45 minutes. Twenty-five to 50 μL of propidium iodide (final concentration, 50 μg/mL) were added to the cells, which were incubated at 37°C water bath for 15 minutes. The cell cycle was analyzed by flow cytometry. Ten thousand cells per sample were analyzed for both assays. To analyze levels of COX-2 protein, the treated HNSCC cells were collected, washed, and fixed in chilled 80% ethanol. The fixed cells were washed twice in PBS and then incubated with a solution containing COX-2–FITC (1:200; Chemical Co.) in the dark for 30 minutes at room temperature. The COX-2 level was determined by FACStar flow cytometry (Becton Dickinson & Co.).

Four-week-old male athymic nude mice (nu/nu) were obtained from Harlan Sprague Dawley, Inc. They were provided Harlan Teklad #2018 Global 18% Protein Rodent Diet and water ad libitum. Mice were housed in temperature-controlled rooms (74 ± 2°F) with a 12-hour alternating light-dark cycle. JHU-013 cells (2 × 10⁶/100 μL/mouse) were injected s.c. into the lower backs of the mice using a 25-gauge needle in left and right sides. All animals were then randomly assigned to four groups (five mice per group) with topical treatment for 24 days following JHU-013 cell inoculation: Sal-B–treated (80 mg/kg/d), celecoxib-treated (5 mg/kg/d), combination-treated (40 mg/kg/d Sal-B and 2.5 mg/kg/d celecoxib), and untreated control group by i.p. administration every day. Body weights were measured once a week. Autopsy in these animals revealed tumor masses, and tumor size and weight were measured. Guidelines for the humane treatment of animals were followed as approved by the Howard University Animal Care and Use Committee.

Paraffin-embedded JHU-013 cancer cell tissue sections were used for the histologic determination of tumor structure. Tissue sections from each group of rats were stained with H&E (American HistoLabs, Inc.).

Paraffin-embedded nude mouse xenograft tissue was done using LSAB2 System-HRP kit following the standard manufacturer's protocol (DakoCytomation). Paraffin-embedded tumor specimens were deparaffinized in xylene and rehydrated in graded alcohol. Endogenous peroxidase was blocked using hydrogen peroxide for 20 minutes. The slides were incubated with primary antibody of proliferating cell nuclear antigen (PCNA; 1:100; EMD) overnight at 4°C and then incubated with second antibody for 2 hours at room temperature. Binding streptavidin–horseradish peroxidase (HRP) was 20 minutes at room temperature. Staining time of 3,3′-diaminobenzidine solution depended on the sample condition. The stained slides were visualized on a microscope. Images were captured with an attached camera linked to a computer. For quantitative analysis of immunohistochemistry and H&E staining, 10 fields were randomly selected for each slide from three different mice. We manually counted the number of positive cell and total number of cells. The percentage of positive cells of the total number of cells was calculated for each image.

RNA was extracted from the treated cells based on the RNeasy Mini kit protocol (Qiagen). First-strand cDNA was synthesized by PCR using a SuperScript III system with oligo(dT) primers (Invitrogen). The expression levels of COX-2 were measured by real-time reverse transcription-PCR using Taqman Universal PCR Master Mix (Applied Biosystems) with COX-2 primer (6-FAM–labeled probe, Hs00153133_m1). A β-actin (4333762T) primer was used as an endogenous control. The real-time PCR was done in triplicate for each gene. AmpliTaq Gold Enzyme was first activated at 95°C for 10 minutes followed by 40 cycles, including denaturation at 95°C for 15 seconds and annealing at 60°C for 1 minute. The data were analyzed using MxPro-Mx3000p software and analyzed using 2^−ΔΔCT method (31).

Prostaglandin E₂ (PGE₂) levels were evaluated as an indication of level of COX-2 activity. JHU-013 cells were treated with Sal-B (100 μmol/L), celecoxib (10 μmol/L), and combination (50 μmol/L Sal-B plus 5 μmol/L celecoxib) in RPMI 1640 with 2% fetal bovine serum. The culture medium was collected, and the level of PGE₂ was determined by a Correlate-EIA Prostaglandin E₂ Enzyme Immunoassay kit (Assay Designs) at absorbances (Bio-Rad) of 405 and 570 nm. The PGE₂ level was calculated as a unit of pg/mL as per instructions provided by the manufacturer of the assay kit.

Apoptotic cell death in deparaffinized tissue sections (American HistoLabs) was determined by terminal deoxynucleotidyl transferase (TdT)–mediated dUTP nick end labeling (TUNEL) technique using TdT-FragEL DNA Fragmentation Detection kit (EMD). This method is based on the specific binding of TdT to 3′-OH end of DNA and ensuring synthesis of polydeoxynucleotide polymer. Briefly, deparaffinized sections were digested using proteinase K, and endogenous peroxidase activity was blocked using 3% hydrogen peroxide in 10 mmol/L Tris (pH 8.0). The sections were then placed in equilibration buffer and incubated with TdT enzyme in a humid chamber at 37°C for 1.5 hours. The reaction was terminated with stop buffer and incubated with HRP for 30 minutes. The apoptotic nuclei were stained by 3,3′-diaminobenzidine and observed under microscope. Ten fields were randomly selected for each slide. Each field was then photographed at a magnification of ×20. We manually counted the numbers of positively stained nuclei for TUNEL. The percentage of positive cells of the total number of cells was calculated for each image.

Whole-cell lysates were prepared from treated JHU-013 and JHU-022 cells incubated with radioimmunoprecipitation assay lysis buffer kit (Santa Cruz Biotechnology), and protein concentrations were quantified using a Bio-Rad protein assay. Whole-cell proteins (30 μg) were separated by 8% SDS-PAGE gel, transferred to a polyvinylidene difluoride membrane (Amersham Corp.), and then probed sequentially with antibodies against the following proteins: COX-2, epidermal growth factor receptor (EGFR), p53, Bcl-2, and β-actin (Sigma). Washed blots were then incubated with HRP-conjugated anti-rabbit, anti-mouse, or anti-goat antibody (Santa Cruz Biotechnology) for 1 hour at room temperature. Blots were developed by a peroxidase reaction using the enhanced chemiluminescence detection system (Bio-Rad).

Results were presented relative to untreated controls. Values represent the mean ± SD of a minimum of three replicate tests. Data were analyzed by the Duncan test following the ANOVA procedure when multiple comparisons were made. Differences were considered significant when P < 0.05.

Four HNSCC cell lines (JHU-06, JHU-011, JHU-013, and JHU-022) were used for testing the effectiveness of the combined therapeutic regimen (Fig. 1A). The cells were first treated with Sal-B (100 μmol/L) alone and celecoxib (10 μmol/L) alone or a half-dose of Sal-B (50 μmol/L) and celecoxib (5 μmol/L) combination for 24 hours. After treatment, cell viability was measured using the MTT assay. Overall, both the individual agent and the combination treatments elicited decreased cell viability and proliferation in all four HNSCC cell lines. There was a significant reduction of cell viability in JHU-013 and JHU-022 cell lines with the combination treatment relative to the individual agent treatments (Fig. 1A).

Cell cycle activity was analyzed after JHU-022 cells were treated with Sal-B (100 μmol/L) alone, celecoxib (10 μmol/L) alone, or Sal-B (50 μmol/L) with celecoxib (5 μmol/L) in combination for 24 and 48 hours, respectively (Fig. 1B; Table 1). Again, all treatments resulted in increased G₀-G₁ phase and decreased G₂-M and S phases of the cell population. The G₀-G₁ phase was markedly higher in the combination treatment relative to either Sal-B or celecoxib treatment alone in 24 hours (68% combination versus ∼59% Sal-B, 63% celecoxib) or the 48-hour treatment (73% combination versus ∼63% Sal-B, 69% celecoxib). Interestingly, the combination treatment inhibited S phase significantly as well (Table 1).

The combination effect was further investigated in JHU-013 solid tumor xenografts in immunodeficient mice. The daily treatments were initiated a day after JHU-013 cells were transplanted into the nude mice. The doses for individual treatments were 40 mg/kg/d Sal-B and 2.5 mg/kg/d celecoxib. The single-agent treatments consisted of 80 mg/kg/d Sal-B or 5 mg/kg/d celecoxib administered via i.p. injection every day for 24 days. During this period, each mouse was manually examined for tumor volume at least four times. The combination treatment was most effective in suppressing tumor growth in athymic nude mice (Fig. 2A). Sal-B alone also resulted in significant inhibition of tumor growth. Animals treated with celecoxib alone generally did not show a reduction in tumor growth compared with the untreated group. It should be noted that one mouse in the celecoxib alone group died on experimental day 22. The average tumor weight in the combination group (0.100 ± 0.005 g) was significantly lower than the untreated control groups (0.208 ± 0.007 g), the Sal-B group (0.162 ± 0.010 g), and the celecoxib group (0.196 ± 0.213 g; P < 0.01; Fig. 2B).

Histologic sections were obtained from the JHU-013 xenografts and subsequently used to evaluate the single-agent and combination treatments on cell proliferation. A large necrotic area in the center of the tumor mass was present in all tumors from the control and single treatment groups, whereas the limited necrosis was found in combo treatment group that tumors were growing more slowly and did not outgrow their blood supply. The loose necrotic structure was visualized by H&E staining (Fig. 3A). PCNA, a proliferation biomarker, was analyzed by immunohistochemical staining (Fig. 3B). Statistics showed expression levels of PCNA that were significantly reduced in treated groups compared with untreated control (P < 0.05). However, there was no significant difference between the single and combination treatments (Fig. 3C).

The combination effect on the inhibition of COX-2 expression in HNSCC cell lines was also investigated in vitro and in vivo. COX-2 mRNA and protein levels were significantly reduced by either the single or the combination treatment in JHU-013 and JHU-022. Western blot and quantitative real-time PCR analyses showed that the combined treatment was significantly more effective at inhibiting COX-2 expression than either single-agent treatment (Fig. 4A). There was an ∼13-fold reduction of COX-2 mRNA as a result of combination treatment compared with untreated JHU-013 cells, whereas the levels of COX-2 mRNA were reduced only approximately 1- or 2-fold in the single treatment groups. Flow cytometry analysis illustrated that COX-2 protein levels were markedly reduced in the treated JHU-013 cells and in the cells obtained by disaggregation of JHU-013 solid tumor xenografts (Fig. 4B). The COX-2 protein levels in the isolated xenograft cells were significantly reduced in the treated groups. In contrast, the lowest COX-2 protein level was observed in the Sal-B–treated group. There was a broad peak of cell distribution in the combination group with a majority of low levels of COX-2 protein. There was 32% reduction in PGE₂ due to treatment with 100 μmol/L Sal-B alone and 40% reduction with 10 μmol/L celecoxib treatment compared with untreated JHU-013 cells (Fig. 4C). In addition, the PGE₂ level was significantly reduced by ∼63% when JHU-013 cells were exposed to a combination of Sal-B (50 μmol/L) and celecoxib (5 μmol/L). The average level of PGE₂ was 292.7 pg/mL in combination treatment, 541.2 pg/mL in Sal-B treatment, and 471.4 pg/mL in celecoxib treatment, respectively.

Cell apoptosis was determined by Western blot and TUNEL assay in the cultured HNSCC cells and JHU-013 xenografts. First, we measured the expression of the apoptosis-related proteins p53 and Bcl-2 after JHU-022 cells were exposed to Sal-B, celecoxib, and the combination (Fig. 5A). Our results showed that either the Sal-B single treatment or the combination resulted in an increased p53 protein level and decreased Bcl-2 level in JHU-022 cells. These results were not observed for celecoxib treatment alone. A significant change in the levels of proapoptotic p53 protein and antiapoptotic Bcl-2 protein was observed in the combination-treated cells. In addition, there was a significant reduction of EGFR protein level in the combination-treated cells. EGFR is one of the fundamental elements associated with COX-2 pathway contributing to the tumor growth and progression in many solid tumors. The apoptotic events were further detected by TUNEL assay in JHU-013 solid tumor xenografts (Fig. 5B). The low-dose combination treatment of Sal-B and celecoxib led to a higher number of apoptotic cells than the untreated or single treatment groups (Fig. 5C).

HNSCC ranks among the six most common cancers in the world and is a significant cause of cancer morbidity and mortality (32). According to the American Cancer Society, ∼35,000 new cases are diagnosed and 7,600 deaths occur annually in the United States. Despite advances in treatment, the overall 5-year survival rate has briefly improved in recent years, but the incidence of oral cavity and pharynx cancers continues to increase. In addition, African-American males have suffered the highest incidence of HNSCC and the lowest survival compared with any racial/ethnic group (33). It urgently needs an effective, nontoxic, and affordable novel pharmacologic agent or drug to prevent development of HNSCC.

The idea that chronic inflammation shares pathways active in malignancy and can work in concert with genotoxic agents and participate in carcinogenesis is noted in many solid tumor systems. Therefore, blocking inflammation is thought to be a logical approach to cancer prevention therapy (34). COX-2, which has aberrant expression in many inflammation-associated tumors, is believed to be related to the development of cancer. Additionally, its principal metabolite PGE₂ seems to have pleiotropic effects on carcinogenesis, such as inducing tumor cell growth, inhibiting apoptosis, promoting tumor angiogenesis, and decreasing immunosurveillance (35). Therefore, COX-2 could be an ideal target for development of chemopreventive agents for cancer prevention; meanwhile, alterations to the COX-2/PGE₂ pathway are widely accepted, which have key roles in promoting tumor development.

Celecoxib is a selective COX-2 inhibitor, and its effectiveness for cancer prevention and therapy has been proven by a plethora of data, including colorectal cancer, prostate cancer, breast cancer, and HNSCC (36, 37). Although several tests of COX-2 inhibitors for cancer prevention were halted because of unacceptable cardiotoxicity, the debate on celecoxib reuse as a chemopreventive agent is appearing to intensify. Celecoxib has been shown to cause adverse cardiovascular effects when taken at high doses over a long duration (38). A report from a large-scale clinical trial indicated that the increased cardiovascular risk in celecoxib was only noted at daily doses of ≥400 mg (12, 37). In other words, the cardiovascular risk can be limited by reducing the dose of celecoxib, below 400 mg daily dose. In our study, with translation celecoxib dosage from human to mouse, the celecoxib dose in combination or single treatment is much lower than the suggested safe dose limit.

The combination of conventional cytotoxic drugs with other chemopreventive agents is an evolving trend in clinical practices for cancer treatment and prevention (39–41). Many studies reported that combination treatment not only could lead to increased anticancer efficacy but also reduced drug side effects. The strategy of this study is to apply a combination approach, Sal-B with low-dose celecoxib, to suppress HNSCC cell growth. There are three reasons why we were interested in the combination treatment: First, we found that Sal-B is capable of inhibiting tumor growth via blocking COX-2 expression and promoting apoptotic pathways (5); second, the combination with low dose of celecoxib may enhance the anti–COX-2 capacity in both COX-2 activity and expression; and third, the cardiovascular side effects of celecoxib may be also limited by reducing the dose of celecoxib and cardiovascular protection of Sal-B.

The outcomes showed that the combination treatment of Sal-B with low-dose celecoxib greatly enhanced the inhibition of HNSCC cell proliferation and cell viability compared with Sal-B or celecoxib alone in either cultured cells or tumor xenografts (Figs. 1 and 2). However, the response to the combination treatment varies in different HNSCC cell lines (Fig. 1A). It may depend on COX-2 expression level in the cell lines. For example, JHU-011, JHU-013, or JHU-022 cell lines have higher levels of COX-2 than JHU-06 (5). These three cell lines are more sensitive to the combination treatment than JHU-06. In addition, the combination treatment resulted in increased cell populations in G₁ phase and decreased populations in S phase in cell cycle experiment (Fig. 1B; Table 1). Cell proliferation was inhibited via DNA synthesis blockage. The results using tumor xenografts provided further support for the observations from cultured cells. The average tumor weight and volume in the combination group was significantly lower than that observed for the untreated and single-agent treatment groups. However, a large variation of tumor size was observed in the celecoxib alone–treated group (Fig. 2B). Although this phenomenon also occurred in our previous study, it remains to be determined if this death was due to celecoxib-related toxicity. Furthermore, a mouse in this group suffered dramatic body weight loss and died on day 22. These results indicate that the observed safety of celecoxib as used in this study is worthy of further investigation. S. miltiorrhiza Bge is regarded as a “panacea” for cardiovascular protection during its use over thousands of years in China due to its ability to promote blood circulation and remove blood stasis. Therefore, Sal-B, the major functional component of S. miltiorrhiza Bge, may be able to mediate the cardiovascular risk of celecoxib. The mechanism of celecoxib increasing cardiovascular risk is not clear, but there is some evidence that celecoxib might lower the production of prostacyclin without affecting the synthesis of thromboxane A₂, contributing to a prothrombotic state (42). Sal-B exerts cardiovascular prevention via making balance with prostacyclin and thromboxane A₂ (43). However, more studies are needed to provide further insight into this possibility.

Our previous study has shown that Sal-B effectively inhibits the levels of COX-2–related mRNA and protein in HNSCC (5). In this study, the enhanced effects of Sal-B and celecoxib on inhibition of COX-2 expression were clearly shown in in vitro tests, such as the levels of COX-2 mRNA and protein (Fig. 4A). In in vivo tests, the lowest COX-2 protein level was observed in the Sal-B–treated group analyzed by flow cytometry; however, a broad peak of low level of COX-2 protein was found in the combination-treated group (Fig. 4B). It is recognized that different tumor growth phases and the tumor microenvironment may cause a large range of COX-2 expression. Furthermore, COX-2 catalyzes the synthesis of prostaglandins, including PGE₂, from arachidonic acid. Overexpression of COX-2 in tumor cells is seen in conjunction with increased PGE₂ levels in HNSCC (44). PGE₂, as a principal metabolite of COX-2 pathway, affects multiple aspects of cell physiology involving cell survival, proliferation, migration, and invasion (45). The expression level of PGE₂ is significantly inhibited by the combination treatment (Fig. 4C). It has been reported that the COX-2/PGE₂ pathway likely plays a role along with other signaling pathways in regulating cancer development (34). Our studies show that the combination treatment exhibited enhanced inhibition of COX-2/PGE₂, induction of apoptosis, and regulation of EGFR expression (Fig. 5A).

In conclusion, Sal-B in combination with low-dose celecoxib can enhance the anticancer efficacy, particularly in chronic inflammatory-associated HNSCC, compared with high-dose Sal-B or celecoxib alone. The in vitro and in vivo lines of evidence show that the enhanced effects of the combination treatment on cell proliferation and cell death of HNSCC cells are through multiple pathways, such as sequential inhibition of the COX-2/PGE₂/EGFR pathway and induction of apoptosis. Additionally, the cardiotoxicity of celecoxib may be limited by lowering its dose. This combination strategy can be considered a new approach in targeting inflammatory-associated HNSCC prevention and treatment.

No potential conflicts of interest were disclosed.

We thank Dan Zhang and Yongheng Dong for their excellent technical assistance and Liang Shan for valuable discussion.

Grant Support: NIH/National Cancer Institute grant CA118770, RCMI 2 G12 RR003048, and Tianjue Gu Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.