Deguelin Potentiates Apoptotic Activity of an EGFR Tyrosine Kinase Inhibitor (AG1478) in PIK3CA-Mutated Head and Neck Squamous Cell Carcinoma

Head and neck squamous cell carcinoma (HNSCC) is known to be intrinsically resistant to inhibitors for epidermal growth factor receptor (EGFR). Until now, clinical outcomes for HNSCC using EGFR inhibitors as single agents have yielded disappointing results. Here, we aimed to study whether combinatorial treatment using AG1478 (EGFR tyrosine kinase inhibitor) and deguelin, which is a rotenoid isolated from the African plant Mundulea sericea, could enhance the anti-tumor effects of AG1478 in HNSCC. For Ca9-22 cells with EGFR, KRAS, and PIK3CA wild types, AG1478 alone suppressed both phosphorylated levels of ERK and AKT and induced apoptosis. On the contrary, for HSC-4 cells with EGFR and KRAS wild types, and a PIK3CA mutant, AG1478 alone did not suppress the phosphorylated level of AKT nor induce apoptosis, while it suppressed ERK phosphorylation. Forced expression of constitutively active PIK3CA (G1633A mutation) significantly reduced the apoptotic effect of AG1478 on the PIK3CA wild-type Ca9-22 cells. When HSC-4 cells with the PIK3CA G1633A mutation were treated with a combination of AG1478 and deguelin, combination effects on apoptosis induction were observed through the inhibition of the AKT pathway. These results suggest that the combination of EGFR tyrosine kinase inhibitor with deguelin is a potential therapeutic approach to treat PIK3CA-mutated HNSCC.


Introduction
Epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase that activates intracellular signaling pathways, including the phosphatidylinositol 3-kinase (PI3K)/AKT pathway (survival signal) and the ERK pathway (proliferation signal), and is abundantly expressed in the majority of epithelial malignancies, including head and neck squamous cell carcinoma (HNSCC) [1]. Because the elevated expression of EGFR in HNSCC correlates with poor prognosis, EGFR signaling has been thought to be

AG1478 Suppressed the Phosphorylation of EGFR in a Dose-Dependent Manner in HSC4 Cells
In this study, we used two representative HNSCC cell lines with different PIK3A gene statuses (Table 1). At first, we tested the dose-dependency of AG1478 in the down-regulation of EGFR phosphorylation. Figure 1 clearly shows that AG1478 inhibited the phosphorylated level to 56.8% at 0.1 µM, and completely at 10 µM or more.  [22] and Kozaki et al. (2006) [14]. ** WT: wild type. phosphorylation. Figure 1 clearly shows that AG1478 inhibited the phosphorylated level to 56.8% at 0.1 μM, and completely at 10 μM or more.  [22] and Kozaki et al. (2006) [14]. ** WT: wild type. Because the activation of EGFR leads to the activation of intracellular signaling pathways, including the PI3K/AKT pathway (survival signal), the inhibition of EGFR activation by AG1478 may induce apoptosis, dependent on the PIK3CA gene status. Therefore, we examined the effects of AG1478 on cancer cells with different PIK3CA gene statuses. Expectedly, HSC-4 cells with a mutant PIK3CA gene did not undergo apoptosis by AG1478 treatment, which was enough of a concentration for the inhibition of EGFR phosphorylation ( Figure 2A). However, for Ca9-22 cells with the WT PIK3CA gene, AG1478 induced cleavage of PARP (c-PARP) as a consequence of apoptosis, apparently at 5 μM, and clearly at 10 μM for 72 h ( Figure 2B). This difference was also shown by a cell viability assay. Reduction of cell viability by AG1478 was shown at a much higher rate in Ca9-22 cells than in HSC-4 cells ( Figure 2C). These data suggested that the apoptotic effect of AG1478 was deeply associated with the PIK3CA gene status.  Because the activation of EGFR leads to the activation of intracellular signaling pathways, including the PI3K/AKT pathway (survival signal), the inhibition of EGFR activation by AG1478 may induce apoptosis, dependent on the PIK3CA gene status. Therefore, we examined the effects of AG1478 on cancer cells with different PIK3CA gene statuses. Expectedly, HSC-4 cells with a mutant PIK3CA gene did not undergo apoptosis by AG1478 treatment, which was enough of a concentration for the inhibition of EGFR phosphorylation ( Figure 2A). However, for Ca9-22 cells with the WT PIK3CA gene, AG1478 induced cleavage of PARP (c-PARP) as a consequence of apoptosis, apparently at 5 µM, and clearly at 10 µM for 72 h ( Figure 2B). This difference was also shown by a cell viability assay. Reduction of cell viability by AG1478 was shown at a much higher rate in Ca9-22 cells than in HSC-4 cells ( Figure 2C). These data suggested that the apoptotic effect of AG1478 was deeply associated with the PIK3CA gene status.

Deguelin Suppressed Phosphorylation Level of AKT in Dose-and Time-Dependent Manner in HSC-4 Cells
We next examined the inhibitory effect of deguelin on phosphorylation levels of AKT in HSC-4 cells. Figure 6 shows that the inhibitory effect of deguelin on AKT phosphorylation in HSC-4 cells was dose-and time-dependent in HSC-4 cells ( Figure 6). Inhibition of AKT phosphorylation was obvious with 1 µM of deguelin treatment for 12 h, and this inhibition was prolonged to at least 24 h ( Figure 6A). The inhibitory effect of deguelin for 12 h on AKT phosphorylation was significantly dose-dependent ( Figure 6B). The best-fitting line on the semi-log plot of Figure 6B indicated a strong exponential relation between the deguelin concentration and AKT phosphorylation.

Deguelin Suppressed Phosphorylation Level of AKT in Dose-and Time-Dependent Manner in HSC-4 Cells
We next examined the inhibitory effect of deguelin on phosphorylation levels of AKT in HSC-4 cells. Figure 6 shows that the inhibitory effect of deguelin on AKT phosphorylation in HSC-4 cells was dose-and time-dependent in HSC-4 cells ( Figure 6). Inhibition of AKT phosphorylation was obvious with 1 μM of deguelin treatment for 12 h, and this inhibition was prolonged to at least 24 h ( Figure 6A). The inhibitory effect of deguelin for 12 h on AKT phosphorylation was significantly dose-dependent ( Figure 6B). The best-fitting line on the semi-log plot of Figure 6B indicated a strong exponential relation between the deguelin concentration and AKT phosphorylation.

The Combination of AG1478 and Deguelin in HSC-4 Cells
Finally, we tested the effects of the combination of AG1478 and deguelin in HSC-4 cells.

Combination of AG1478 and Deguelin Was Stronger Than AG1478 Alone with Regard to the Effect on the Active Form of AKT, and the Combination Was Stronger than Deguelin Alone with Regard to the Effect on the Active Form of ERK in HSC-4 Cells
To test whether deguelin has activity for the sensitization of AG1478, we treated cells with AG1478 in the presence of deguelin to see the active forms of AKT and ERK. The 2 h combination treatment of deguelin and AG1478 significantly reduced phosphorylated levels of ERK, but not AKT, compared to the control or deguelin alone ( Figure 7A). Longer treatment (12 h) with the combination significantly reduced AKT phosphorylation relative to the control or AG1478 alone ( Figure 7B).

The Combination of AG1478 and Deguelin in HSC-4 Cells
Finally, we tested the effects of the combination of AG1478 and deguelin in HSC-4 cells.

Combination of AG1478 and Deguelin Was Stronger Than AG1478 Alone with Regard to the Effect on the Active Form of AKT, and the Combination Was Stronger than Deguelin Alone with Regard to the Effect on the Active Form of ERK in HSC-4 Cells
To test whether deguelin has activity for the sensitization of AG1478, we treated cells with AG1478 in the presence of deguelin to see the active forms of AKT and ERK. The 2 h combination treatment of deguelin and AG1478 significantly reduced phosphorylated levels of ERK, but not AKT, compared to the control or deguelin alone ( Figure 7A). Longer treatment (12 h) with the combination significantly reduced AKT phosphorylation relative to the control or AG1478 alone ( Figure 7B).

The Combination of Deguelin and AG1478 Significantly Induced Apoptosis Relative to the Control or Alone in HSC-4 Cells
Finally, we determined the combination effect of deguelin and AG1478 on the apoptosis of HSC-4 cells. As shown in Figure 8A, deguelin (1 μM) alone significantly increased c-PARP levels to 2.1-fold higher than the control, while AG1478 (1 μM) alone slightly increased c-PARP levels to 1.1-fold higher than the control. When the cells were treated with both drugs, the c-PARP level increased up to 5.0-fold compared to the control, 4.4-fold compared to AG1478 alone, and 2.4-fold compared to deguelin alone, respectively. Furthermore, as shown in Figure 8B, AG1478 at 1 μM alone did not significantly reduce the proliferation of HSC-4 cells (see Figure 2C); however, the combination of deguelin at 0.1 μM and AG1478 at 1 μM significantly reduced the viable cell number of HSC-4 cells relative to the control or alone. Even a low dose of deguelin (0.1 μM) can significantly enhance the anti-tumor efficacy of AG1478.

The Combination of Deguelin and AG1478 Significantly Induced Apoptosis Relative to the Control or Alone in HSC-4 Cells
Finally, we determined the combination effect of deguelin and AG1478 on the apoptosis of HSC-4 cells. As shown in Figure 8A, deguelin (1 µM) alone significantly increased c-PARP levels to 2.1-fold higher than the control, while AG1478 (1 µM) alone slightly increased c-PARP levels to 1.1-fold higher than the control. When the cells were treated with both drugs, the c-PARP level increased up to 5.0-fold compared to the control, 4.4-fold compared to AG1478 alone, and 2.4-fold compared to deguelin alone, respectively. Furthermore, as shown in Figure 8B, AG1478 at 1 µM alone did not significantly reduce the proliferation of HSC-4 cells (see Figure 2C); however, the combination of deguelin at 0.1 µM and AG1478 at 1 µM significantly reduced the viable cell number of HSC-4 cells relative to the control or alone. Even a low dose of deguelin (0.1 µM) can significantly enhance the anti-tumor efficacy of AG1478.

The Combination of Deguelin and AG1478 Significantly Induced Apoptosis Relative to the Control or Alone in HSC-4 Cells
Finally, we determined the combination effect of deguelin and AG1478 on the apoptosis of HSC-4 cells. As shown in Figure 8A, deguelin (1 μM) alone significantly increased c-PARP levels to 2.1-fold higher than the control, while AG1478 (1 μM) alone slightly increased c-PARP levels to 1.1-fold higher than the control. When the cells were treated with both drugs, the c-PARP level increased up to 5.0-fold compared to the control, 4.4-fold compared to AG1478 alone, and 2.4-fold compared to deguelin alone, respectively. Furthermore, as shown in Figure 8B, AG1478 at 1 μM alone did not significantly reduce the proliferation of HSC-4 cells (see Figure 2C); however, the combination of deguelin at 0.1 μM and AG1478 at 1 μM significantly reduced the viable cell number of HSC-4 cells relative to the control or alone. Even a low dose of deguelin (0.1 μM) can significantly enhance the anti-tumor efficacy of AG1478.

Discussion
Overexpression of EGFR is observed in a majority of studied HNSCC cases [1]. Thus, the use of EGFR inhibitors, such as cetuximab, gefitinib, and erlotinib, has been expected to be an applicable strategy for HNSCC therapy. However, their monotherapy yielded low response rates in clinical trials [8,9]. A promising solution to improve the clinical response rate may be through the combination of EGFR inhibitors with other treatment modalities. Here, we demonstrated that deguelin has a combination effect with AG1478 in PIK3CA mutant HNSCC cells. Previously, deguelin alone was reported to induce apoptosis via AKT signaling inhibition, autophagy, and CDK4/Survivin degradation at high doses (50 or 100 µM) [20] and to suppress invasive ability at a low dose (0.1 µM) in HNSCC cells [23]. Although deguelin was reported to sensitize HNSCC cells to 5-fluorouracil (FU) [20], to the best of our knowledge, this is the first report that deguelin has a combination effect with an EGFR TKI for induction of apoptosis in HNSCC cells.
Previously, EGFR inhibitor-based therapy has been developed in combination with other inhibitory drugs targeting PI3K signaling. Rebucci et al. [24] demonstrated that the combination of cetuximab (EGFR antibody) with a PI3K inhibitor could be a good therapeutic option in PIK3CA-mutated HNSCC. Furthermore, D Amato et al. [25] reported that the dual PI3K/mTOR inhibitor, PKI-587, enhances sensitivity to cetuximab in Detroit562 cells (HNSCC) having the PIK3CA mutant. On the other, in some cases, it has been reported that the combination of erlotinib (EGFR TKI) and everolimus (mTOR inhibitor) did not show significant benefits in unselected patients with platinum-resistant metastatic HNSCC [26], and that six out of 12 patients administered erlotinib and temsirolimus (mTOR inhibitor) withdrew within six weeks due to toxicity or death due to poor tolerance, prompting early closure of the trial [27]. Thus, combination therapy with an EGFR inhibitor and PI3K pathway inhibitor still seeks to provide better benefits for not only clinical efficacy, but also low side effects. Here, we investigated a new EGFR-based combination therapy using the natural compound deguelin, showing that the PIK3CA mutation is associated with sensitivity to EGFR TKIs, and that deguelin has a combination effect with an EGFR TKI for the induction of apoptosis in HNSCC cells. Our data suggest that EGFR inhibitor-based therapy combined with deguelin showed a better benefit for HNSCC patients with the PIK3CA mutation, which are often resistant to EGFR TKI monotherapy. Therefore, the mutation of PIK3CA should be determined as a biomarker to administer deguelin for better clinical outcomes in cancer patients.
In this study, we indicated that deguelin at 1 µM reduced not only p-AKT expression, but also p-ERK expression. Previously, we reported that deguelin at 1 or 10 µM induced apoptosis [21] in HSC-4 cells, which were not sensitive to AG1478. Indeed, in this study, the viable cell number was almost the same between the combination of deguelin at 1 µM and AG1478 at 1 µM, and deguelin at 1 µM alone; however, this combination significantly reduced the viable cell number relative to AG1478 alone at 1 µM (data not shown). Taken together, this fact suggested that a high-dose deguelin monotherapy could be used for PIK3CA-mutated HNSCC patients, or the combination of low-dose deguelin and EGFR TKI could be used for PIK3CA-mutated HNSCC patients.
In HNSCC, EGFR and KRAS mutations seem to be rare [12,28]; however, mutation of PIK3CA occurs in 9%-11% of patients [13,14] and has been shown to activate the AKT signaling pathway [14]. In this study, we aimed to elucidate the importance of the EGFR signaling pathway in HNSCC using two kinds of cell lines with EGFR and KRAS wild-type genes with different gene statuses of PIK3CA [14,22].
Although KRAS mutation has been established as a potential biomarker for predicting the efficacy of erlotinib in lung cancer [11], little is known about a predictive marker for EGFR TKIs in HNSCC. Here, we clearly demonstrated that the PIK3CA mutation is associated with sensitivity to EGFR TKIs. Signal transduction through the PI3K/AKT pathway is mediated by the conversion of phosphatidylinositol 4,5-bisphosphate (PIP2) to phosphatidylinositol 3,4,5-triphosphate (PIP3) by PI3K. This reaction is antagonized by the phosphatase and tensin homolog deleted from chromosome ten (PTEN). In prostate cancer, PTEN expression is associated with EGFR inhibitor sensitivity [29]. On the contrary, PTEN expression in HNSCC was not associated with sensitivity to the EGFR inhibitor [15]. Therefore, further study was needed to determine the role of the PI3K signaling status in the efficacy of EGFR inhibitors in HNSCC.
horse radish peroxidase-conjugated avidin and electrochemiluminescence reagent (Amersham, Buckinghamshire, UK). The signals detected by anti-GAPDH antibody, used as the loading control.

Vector Construction and Transfection
To express mutant PI3K (E545K, constitutively active mutation) (PIK3CA G1633A) as a GFP-tagged protein, the PIK3CA mutant gene in pBABEpuro-HA-PIKCA-E545K1 was placed into pAcGFP1-Hyg-C1 vector using the In-Fusion™ cloning system according to manufacturer's protocols. Briefly, the PIK3CA E545K insert without the HA region was amplified by PCR with the primer set: 5 -AAG CTT CGA ATT CTG ATG CCT CCA CGA CCA-3 (forward) and 5 -CGG TAC CGT CGA CTG GTT CAA TGC ATG CT-3 (reverse) and cloned into pAcGFP1-Hyg-C1, which has been linearized by PCR with the primer set: 5 -CAG TCG ACG GTA CCG CGG GC-3 (forward) and 5 -CAG AAT TCG AAG CTT GAG CTC GAG ATC TGA-3 (reverse) (underlined sequences were indicated for In-Fusion™ enzyme recognition tag sequences consisting of 15 mer). The mutant PIK3CA gene inserted in the pAcGFP1-Hyg-C1 vector was termed as pPIK3CA-GFP. The sequences were confirmed by DNA sequencing. The vector was transfected into cells with Xfect Transfection Reagent. The experimental protocols for the introduction of the vectors into Escherichia. coli (DH5α or JM109) or human cancer cells were approved by the ethical committee of Ohu University (Koriyama, Japan).

Protein Assay
The protein content in the lysates were measured using a DCTM Protein Assay Kit with bovine serum albumin as the standard.

Statistical Analysis
Statistical significance was calculated using Student's t-test. p-values less than 0.05 were considered significant.

Conclusions
In conclusion, at first we demonstrated that the PIK3CA mutation is associated with sensitivity to EGFR TKIs. Second, we demonstrated the combination effect of AG1478 and deguelin in PIK3CA-mutated HNSCC. Therefore, our data in this study suggested that combination therapies using EGFR TKIs and deguelin are new therapeutic approaches to treat PIK3CA-mutated HNSCC.