Docetaxel suppresses invasiveness of head and neck cancer cells in vitro

Abstract

The combination of docetaxel, cisplatin, and fluorouracil significantly enhances the survival of head and neck cancer patients compared to cisplatin and fluorouracil. We hypothesized that docetaxel may affect invasiveness of the head and neck cancer cells in addition to its tumor‐killing effect. Two different head and neck cancer cell lines (HEp‐2 and Ca9‐22) were treated with docetaxel at IC₁₀ and IC₅₀ concentrations. Cell migration and invasive growth was evaluated by wound healing assay and three‐dimensional (3D) culture of multicellular tumor spheroids, respectively. Expression levels of possible downstream effectors for cell migration/invasiveness were measured by immunoblotting in conditions with or without docetaxel. Docetaxel, but not cisplatin, suppressed filopodia formation compared with no treatment (control) condition. Consistent with this, docetaxel suppressed two‐dimensional (2D) cell migration and 3D cell invasion compared with control or cisplatin. Only docetaxel treated cells exhibited thick tubulin bundle and had lower activity of Cdc42, a member of the Rho family of small GTPases. In conclusion, Docetaxel treatment suppressed migration and invasiveness of head and neck cancer cells in vitro, which is likely to be mediated by regulating Cdc42 activity. (Cancer Sci 2010; 00: 000–000)

Squamous cell carcinoma (SCC) of the head and neck accounts for 5% of cancers in adults worldwide.^{( 1 )} The disease is potentially curable at an early stage, but most patients present with locally advanced disease. Only 30–50% of patients with locally advanced disease live for 3 years after surgery and radiation therapy. Locoregional recurrences or distant metastases develop in 40–60% of them.^{( 2} , ³ , ⁴ , ^{5 )} Moreover, radical resection of the tumor, which results in loss of the larynx, severely compromises quality of life in such patients. In order to increase the survival rate and preserve voice and swallowing function over surgery and radiation alone, various supplementary chemotherapy regimens have been developed.^{( 6} , ⁷ , ^{8 )} Recently it was reported that the combination of docetaxel, cisplatin, and fluorouracil (TPF) followed by chemoradiotherapy significantly enhances survival of head and neck cancer patients compared to cisplatin and fluorouracil (PF) followed by chemoradiotherapy.^{( 9 )} Estimates of overall survival at 3 years were 62% in the TPF and 48% in the PF groups. It was also reported that incidence of distant metastasis is lower in the TPF (5%vs 9%), although it does not reach statistical significance. We therefore hypothesized that docetaxel may affect migration and invasiveness of the head and neck cancer cells. The mechanism of antitumor effect of the taxanes, such as docetaxel is thought to be the stabilization of microtubules during mitosis, which leads to blockade of cell division and eventually cell death. However, the effect of docetaxel on cell invasiveness has not been evaluated in head and neck cancers. We demonstrate here that docetaxel inhibited migration of head and neck cancer cells in vitro, suggesting that microtubule plays an important role in tumor cell motility.

Material and Methods

Cell culture.  Two human head and neck SCC cell lines were used. HEp‐2 cells derived from larynx was obtained from the American Type Culture Collection (Manassas, VA, USA). Ca9‐22 cells derived from gingiva was obtained from the Japanese Collection of Research Bioresources (Osaka, Japan). HEp‐2 and Ca9‐22 cells were cultured in Eagle’s Minimum essential medium with 10% fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin. The cells were maintained in a humidified incubator with 5% CO₂ at 37°C.

Proliferation assay.  The anti‐proliferative effects of cisplatin (Wako, Osaka, Japan) and docetaxel (Toronto Research Chemicals, Toronto, ON, Canada) were analyzed in water soluble tetrazolium assay (a modified MTT assay). Cells were plated in 96‐well plates at 1000 cells per well. After 48 h of incubation, the cells were exposed to a drug for 1 h. After washing with PBS, the cells were grown in culture medium for an additional 96‐h period. The number of viable cells was determined by addition of WST‐8 (Dojindo laboratories, Kumamoto, Japan) as per the manufacturer’s instructions. The IC₁₀ and IC₅₀ values, being the drug concentrations at 10% and 50% cell growth inhibition compared with control cell growth, were used in all further experiments.

Migration assay.  HEp‐2 and Ca9‐22 cells were seeded in 90‐mm dishes at 5 × 10⁵ cells per dish. At 48 h of incubation, cells were treated with culture medium without drug (control), cisplatin, or docetaxel for 1 h. After washing, cells were maintained in culture medium for an additional 72‐h period, and a scratch wound was applied with a sterile 200‐μL pipette tip. Wound width was measured in four areas at time 0 (immediately after wounding) and after 12 or 24 h of incubation. At the end of the incubation, the cells were fixed with 4% paraformaldehyde for 10 min and stained for α‐tubulin (1:20; Calbiochem, La Jolla, CA, USA) and phalloidin (200 units/mL; Invitrogen, Carlsbad, CA, USA) (see Immunocytochemistry).

Three‐dimensional (3D)‐gel culture assay.  HEp‐2 cells were seeded in a 96‐well low attachment plate (Smilon celltite spheroid U; Sumitomo Bakelite, Tokyo, Japan) at 2 × 10³ cells per well. Multicellular tumor spheroids were observable around 48 h after plating. The spheroids were placed into the extracellular matrix (Matrigel; BD Biosciences, San Jose, CA, USA) in Transwell filters (Corning, Corning, NY, USA). After 48 h of incubation in the Transwell, the spheroids were treated with cisplatin, docetaxel, or culture medium (control) for 1 h at IC₁₀ and IC₅₀. After washing, cells were grown in culture medium for an additional 96‐h period.

Western blot analysis.  HEp‐2 cells and Ca9‐22 cells were seeded at 1 × 10⁶ cells in a 90‐mm dish, allowed 2 days for attachment, and treated with docetaxel and cisplatin at IC₁₀ for 1 h. At 0, 24, and 48 h of culture (only immediately after treatment in the evaluation of Rho GTPases), cells were washed twice with ice‐cold PBS and were harvested by scraping in cell lysis buffer (50 mM Tris‐HCl, 150 mM NaCl, 0.5 mM EDTA, 1% Triton X‐100) with protease inhibitors (Roche Applied Science, Mannheim, Germany). After homogenization and centrifugation for 30 min at 13 000 g , the supernatants (30 μg of protein) were subjected to SDS–PAGE, followed by transferring to Hybond‐P PVDF membrane (GE Healthcare, Little Chalfont, UK). The membrane was incubated with either anti‐phospho Ezrin/Radixin/Moesin (ERM), cofilin, LIM domain kinase 1 (LIMK1) (Cell Signaling Technology, Danvers, MA, USA), Cdc42 (1:1000; Cytoskeleton), Rac1 (1:250; Cytoskeleton, Denver, CO, USA), or RhoA (1:500; Cytoskeleton). After washing, it was incubated with horseradish peroxidase‐conjugated antirabbit IgG or antimouse IgG (Jackson Immuno Research Laboratories, West Grove, PA, USA). The images were obtained with Fujifilm LAS‐4000 mini (Fujifilm, Tokyo, Japan).

Immunocytochemistry.  HEp‐2 cells were seeded on glass coverslips in a 12‐well plate at 3 × 10⁴ cells per well. After 48 h of culture, the cells were treated with docetaxel or cisplatin at IC₁₀ for 1 h. After washing, the cells were fixed with cold methanol for 10 min. The fixed cells were made permeable with 0.2% Triton X‐100 (Sigma‐Aldrich, St. Louis, MO, USA) in PBS and incubated with a monoclonal antibody against Cdc42 (1:200; Cytoskeleton), Rac1 (1:20; Cytoskeleton), or RhoA(1:100; Cytoskeleton) in 10% goat serum. After washing, the coverslips were incubated with FITC‐labeled antimouse IgG antibodies (1:200; Invitrogen) and DAPI (1:500; Roche). After washing, coverslips were mounted in Dakocytomation Fluorescent Mounting Medium (Dako, Carpinteria, CA, USA) and were observed with a laser confocal microscope FV1000‐D (Olympus, Tokyo, Japan).

Rho GTPases activity assay.  A G‐LISA Cdc42, Rac, and Rho‐A activation assay biochemistry kit (Cytoskeleton) was used for measuring Cdc42, Rac, and Rho‐A activity. HEp‐2 cells and Ca9‐22 cells were seeded at 5 × 10⁵ cells per dish in 90‐mm dish, allowed 2 days for attachment and treated with docetaxel or cisplatin at IC₁₀ for 1 h. Immediately after washing twice with ice‐cold PBS, the cells were harvested by scraping in lysis buffer and cells were processed rapidly on ice. The lysates were clarified by centrifugation at 7300–17 800g at 4°C for 2 min. Protein concentration was determined according to the manufacturer’s protocol, and the cell extracts were adjusted to 0.25 mg/mL for Cdc42 assay and 1.5 mg/mL for Rac/RhoA assay.

Statistical analysis.  Welch’s t‐test was used to compare all experimental results. A P‐value less than 0.05 was considered significant.

Results

Determination of IC₁₀ and IC₅₀ in HEp‐2 and Ca9‐22 cells.  The antiproliferative effects of cisplatin and docetaxel in HEp‐2 and Ca9‐22 cells were determined by a modified MTT assay (WST assay). Both drugs inhibited cell growth in a dose‐dependent manner (Fig. 1). HEp‐2 cells were more sensitive to cisplatin than Ca9‐22 cells, while Ca9‐22 cells were more sensitive to docetaxel. The IC₁₀ and IC₅₀ concentrations were used in the following experiments (Table 1).

Figure 1.

 Growth curve after 1‐h drug exposure followed by a 96‐h incubation in cisplatin or docetaxel treatment. Control cell growth = 1.0. Each data point represents mean ± SE.

Table 1.

 IC₁₀ or IC₅₀ values in two cell lines

Cell line	Anticancer agent	IC10	IC50
HEp‐2	Cisplatin	2 μM	20 μM
	Docetaxel	5 nM	23 nM
Ca9‐22	Cisplatin	10 μM	40 μM
	Docetaxel	6 nM	11 nM

IC₁₀ or IC₅₀ values after 1‐h drug exposure followed by a 96‐h incubation in two head and neck cell lines.

Docetaxel inhibits the migration of head and neck cancer cells.  Migration was assessed in a wound healing assay. In HEp‐2 cells, wound closure relative to no treatment condition was 100.8 ± 3.5% and 71.4 ± 7.8% (mean ± SE) after cisplatin and docetaxel treatment at IC₁₀, respectively (P < 0.001). In Ca9‐22 cells, wound closure was 102.3 ± 3.1% and 58.7 ± 5.6% after cisplatin and docetaxel treatment at IC₁₀, respectively (P < 0.001) (Fig. 2). Docetaxel significantly inhibited the migration as compared to cisplatin treatment or no treatment in both cell lines.

Figure 2.

 Migration assay at IC₁₀ in two head and neck cancer cell lines. Migration rate compared to control cell migration rate. Each data point represents mean ± SE. **P < 0.001.

Docetaxel inhibits the invasiveness of multicellular tumor spheroids.  Because 3D multicellular tumor spheroids provide a microenvironment that mimics tumors in vivo,^{( 10 )} spheroid culture has been used to evaluate tumor cell invasiveness and/or its responsiveness to anticancer drugs.^{( 11} , ¹² , ¹³ , ¹⁴ , ^{15 )} At IC₁₀ determined in monolayer culture, both cisplatin or docetaxel did not affect filopodia formation. However, at IC₅₀ also determined in monolayer culture, docetaxel, but not cisplatin, significantly decreased filopodia formation in HEp‐2 cells in spheroid culture (Fig. 3). It has been shown that tumor cells in spheroid culture require higher concentrations of chemotherapeutic agents to inhibit their growth than those in monolayer culture. Therefore, it is not surprising that a higher concentration of docetaxel was required to suppress cell invasion in spheroid than in monolayer culture. In fact, the maximum plasma concentration of docetaxel in cancer patients treated by this drug is reported to be 1000‐fold higher than the concentration used in our experiment.^{( 16 )} Thus, we do not think that the docetaxel concentration in our experiment was high enough to non‐specifically affect various cell functions. Because Ca9‐22 cells formed cysts instead of spheroids under the culture condition, we could not evaluate the effect of docetaxel or cisplatin on the invasiveness of this cell line. Taken together with the result of wound healing assay, docetaxel, but not cisplatin, treatment inhibited head and neck cancer cell motility and invasiveness.

Figure 3.

 Three‐dimensional‐Gel culture of HEp‐2 cell spheroids at IC₅₀ concentration. HEp‐2 cell spheroids were treated with cisplatin or docetaxel at IC₅₀ followed by 96‐h incubation. Bars, 250 μm.

Docetaxel suppresses filopodia formation.  By phalloidin staining, the filamentous actin structure of HEp‐2 cells and Ca9‐22 cells was visualized (Fig. 4). Filopodia formation was suppressed when cells were treated with docetaxel. But no gross abnormality was found in the actin filament structure. Ezrin/radixin/moesin (ERM) proteins link the cortical cytoskeleton to the plasma membrane. In their active conformation (i.e. phosphorylated ERM), the N‐terminal ERM domain binds to the cytoplasmic tails of transmembrane proteins, and the C‐terminal ERM association domain binds to actin filaments.^{( 17 )} Using a p‐ERM antibody, the levels of active ERM proteins were evaluated by western blotting after no treatment, cisplatin, or docetaxel treatment at IC₁₀. There was no significant difference in the level of p‐ERM among cisplatin, docetaxel, and no treatment over a time course up to 48 h (data not shown). We also investigated the cofilin pathway as a regulator of the actin cytoskeleton. Cofilin is a small protein that is able to bind both G‐actin and F‐actin, and regulated by LIM kinase 1 and its related kinases. The levels of cofilin and LIMK1 were evaluated over a time course of treatment at IC₁₀. The levels of these proteins were also not significantly different among cisplatin, docetaxel, and no treatment (data not shown).

Figure 4.

 Staining of α‐tubulin (anti‐α‐tubulin), F‐actin (phalloidin), and nucleus (DAPI) in HEp‐2 cells and Ca9‐22 cells treated by cisplatin, docetaxel, or no treatment at IC₅₀. Arrows, thick tubulin bundle; arrow heads, filopodia. Bars, 50 μm.

Tubulin bundle is formed by docetaxel treatment.  Docetaxel acts as a microtubule stabilizer so that tubulin monomers cannot break apart from microtubules for remodeling.^{( 18 )} Cytotoxic effect is believed to be due to inhibition of centrosome‐associated microtubule turnover that eventually affect cell division.^{( 19 )} But docetaxel may exert similar effect on cytosolic, non‐centrosome associated microtubules. We therefore examined the structure of microtubules. Tubulin bundle formation was noted in docetaxel treatment (Fig. 4), as has been previously demonstrated with docetaxel or paclitaxel,^{( 20} , ^{21 )} but not in cisplatin treatment. This deformed microtubular structure may be sufficient to inhibit cell movement. However, the fact that docetaxel treatment inhibited filopodia formation suggests that there might be a connection between microtubular deformity and the actin remodeling process.

Docetaxel inhibits Cdc42 activity and its subcellular local‐ization.  Rho GTPases are molecular switches that regulate many essential cellular processes, including actin dynamics, gene transcription, cell‐cycle progression, cell adhesion, tumor progression, and invasiveness.^{( 22} , ²³ , ^{24 )} Of Rho‐GTPases, Cdc42 has been previously implicated in connecting microtubular input to actin filament organization.^{( 25 )} Cdc42 also promotes leading‐edge extension through activation of Rac. Rac, and Rho are implicated in the formation of lamellipodia and stress fiber, respectively.^{( 26 )} Thus, we examined the activity of Cdc42, Rac, and RhoA in the cells that underwent cisplatin or docetaxel treatment or no treatment. At IC₁₀ concentration, docetaxel significantly decreased Cdc42 and Rac activity in HEp‐2 cells, but not RhoA activity (Fig. 5a). The total amount of Cdc42, Rac, and RhoA was not significantly different among these three conditions (Fig. 5b).

Figure 5.

 (a) Colorimetric assay of Cdc42, Rac, and RhoA activity in HEp‐2 cells. The levels of activated Cdc42, Rac, and RhoA in HEp‐2 cells were evaluated immediately after 1 h of indicated treatment. Each data point represents mean ± SE. *P < 0.05, **P < 0.01. (b) Western blot analysis of rho family GTPases. (c) Immunostaining of Cdc42, Rac1, and RhoA in HEp‐2 cells treated by cisplatin, docetaxel, or no treatment at IC₁₀. Nucleus was stained by DAPI. Arrows, Cdc42 localized at plasma membrane of HEp‐2 cells. Bars, 25 μm.

It is reported that Cdc42 is activated in a thin band at cell edges extending filopodia.^{( 27 )} Consistent with the results of activity assay, Cdc42 localizing at the plasma membrane was decreased after docetaxel treatment at IC₁₀ (Fig. 5c). Localization of Rac1 and RhoA had no apparent changes after treatment compared to control.

Discussion

We have shown that IC₁₀ and/or IC₅₀ concentration of docetaxel, but not cisplatin, inhibited migration of head and neck cancer cell lines both in 2D and 3D cultures. These results suggest that docetaxel inhibits local invasion of head and neck cancer cells. Although the effect of docetaxel on cell migration and invasiveness of ovary cancer cells,^{( 28 )} umbilical vein endothelial cells,^{( 29 )} and glioma cells^{( 30 )} has been described, its effect on head and neck cancer cells has not been evaluated.

Moreover, we also revealed that docetaxel treatment decreased Cdc42 activities as compared to cisplatin treatment or control in HEp‐2 cells. Actin polymerization and filament elongation at the front, coupled to actin filament contraction at the rear, are thought to provide the major driving forces for migration in animal cells. Persistent and efficient directed migration requires additional cellular changes involving polarization of the microtubule cytoskeleton and secretory pathway.^{( 31} , ^{32 )} Among Rho GTPases, Cdc42 was reported to control the polarity of actin and microtubule through distinct signal transduction pathways.^{( 25 )} On the other hand, microtubule or actin may direct Cdc42 activation to specific peripheral locations.^{( 33} , ³⁴ , ³⁵ , ^{36 )} Rac1/Cdc42 capture microtubule plus ends through IQ motif containing GTPase activating protein 1 (IQGAP1) and CAP‐GLY domain containing linker protein 1 (CLIP‐170 also known as CLIP‐1), and form a tripartite complex.^{( 37 )} This complex links microtubules and special cortical regions, which is essential for cell polarization or inducing multiple leading edges.^{( 37 )} Moreover, both upstream guanine‐nucleotide‐exchange‐factor activators of Rac1/Cdc42, including guanine nucleotide exchange factor‐H1 (GEF‐H1),^{( 38 )} Vav,^{( 39 )} and Lfc,^{( 40 )} and the downstream Rac1/Cdc42 effectors MLK2 and JNK,^{( 41 )} either bind to tubulin or localize to microtubules in cells.^{( 42 )} These mechanisms would explain how growing microtubule plus ends could provide active Rac1/Cdc42 to the leading cell edge. We propose that the abnormal tubulin bundle induced by docetaxel lead to suppression of Cdc42 activity by losing the ability to direct Cdc42 activation to intended locations of plasma membrane. Decreased Cdc42 localization at the plasma membrane after docetaxel treatment (Fig. 5) is consistent with this possibility. This decreased Cdc42 activity was likely to affect actin filament formation at the leading edge, resulting in decreased filopodia (Fig. 4).

In conclusion, these results suggest that docetaxel treatment has the benefit of reducing local invasion in addition to its tumor‐killing effect, giving experimental support for clinical data that TPF is better than PF for treating head and neck cancer.