In vitro effects of gamma-secretase inhibition in HPV-positive and HPV-negative head and neck squamous cell carcinoma

All experiments were conducted in three HNSCC cell lines, two HPV-negative (FaDu and Cal27, hypopharyngeal and tongue squamous cell carcinoma, respectively) and one HPV-positive cell line (SCC154, tongue squamous cell carcinoma) [19], which were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were split every 5–8 days, depending on the seeding density. All cell lines were used only up to the 30th passage due to the risk of accumulating additional mutations. The cultivation was performed using Dulbecco’s modified eagle’s medium (DMEM), 1% Penicillin/Streptomycin (P/S), and 10% fetal calf serum (FBS), which were obtained from Gibco (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). The incubation environment for the cells was 37 °C and 5% CO2 (Hera Cell 240, Heraeus Holding GmbH) cultivated in DMEM. Cells were split by removing the medium, washing the cells with Dulbecco’s Phosphate Buffered Saline (DPBS) (Gibco, Thermo Fischer Scientific, Waltham, Massachusetts, USA) and 0.05% trypsin – 0.53 mM EDTA solution (Sigma-Aldrich, St. Louis, Missouri, USA) to detach the cells. The inhibitor PF was acquired from Selleckchem (Houston, Texas, USA) [20]. In all experiments, three biological replicates were conducted.

In order to determine the cytototoxic effects of PF and calculate its half-maximal inhibitory concentration (IC50) value of PF, we performed a dose–response assay in 96-well plates (Sarstedt, Nurnbrecht, Germany) for each cell line. In particular, 5000 cells/well were seeded in 100 μl DMEM. After 24 h (h) of incubation, cells were exposed to gradually increasing concentrations of the inhibitor. The inhibitor was reconstituted in dimethyl sulfoxide (DMSO) (Sigma-Aldrich, St. Louis, Missouri, USA) and diluted in cell culture medium (2.5–40 μM for FaDu and Cal27, and 1.25–20 μM for SCC154 with five replicates per dose). Vehicle control was conducted using 0.1% DMSO. PF treatment was followed by irradiation with 2, 4, and 8 Gray (Gy) with the YXLON device (YXLON, International GmbH, Hamburg, Germany). After 72 h, the medium containing the inhibitor/vehicle control was removed and replaced with 100 μl of 56 μM resazurin (Sigma-Aldrich, St. Louis, Missouri, USA) solution. FaDu and Cal27 were then incubated for 90 min (min), while SCC154 were incubated for 180 min. Measurements were performed using the TECAN Spark reader (TECAN Spark, Tecan Group Ltd, Maennedorf, Switzerland).

In order to assess the anti-migratory potential of the PF, we used 24-well plates (Greiner Bio-One, Frickenhausen, Germany) with Ibidi inserts (Greiner Bio-One, Frickenhausen, Germany) to create the initial gap. For this, 70,000 Cal27 and FaDu cells were seeded in Ibidi chambers (140,000/well) and 250,000 SCC154 cells/ibidi chamber in starvation medium (DMEM with 1% P/S and 1% FBS). Ibidi inserts were removed after reaching 100% cell confluency. Upon removal of the Ibidi inserts, wells were washed with DPBS and introduced with starving medium containing two different inhibitor concentrations (17 and 23 μM for Cal27 and FaDu, 10 and 20 μM for SCC154), as well as the vehicle control (0.1% DMSO). Visualizations using the TECAN Spark plate reader were utilized after 0, 24, and 48 h. Calculations and analysis of the obtained images were performed with ImageJ.

Furthemore, to determine a single-cell’s ability to form colonies after treatment, we performed a clonogenic assay. In particular, 1000 SCC154 cells/well and 350 cells/well for Cal27 and FaDu were seeded in 12-well plates (Sarstedt, Nurnbrecht, Germany) containing 1 mL of medium. Treatment was performed 24 h after seeding (23 μM PF for Cal27 and FaDu and 20 μM for SCC154). After 72 h of treatment, cells were incubated for 10 (Cal27 and SCC154) and 14 days (FaDu). Subsequently, wells were washed with DPBS to remove cell debris, and the visualization was performed with the TECAN Spark plate reader.

Caspase 3/7 Glo Assay (Promega, Fitchburg, Wisconsin, USA) was conducted to assess the pro-apoptotic effects of the inhibitor. For Cal27 and FaDu, 7.500 cells/well and 15,000 SCC154 cells/well in 100 μL DMEM were seeded into 96-well plates (Sarstedt, Nurnbrecht, Germany). After 24 h, the medium was cells were treated. After 72 h, the medium was collected and mixed in a 1:1 ratio with Caspase 3/7 Glo reagent in three conditions: blank (media + reagent, no cells), tested sample (media containing inhibitor, treated cells + reagent), and negative control (media containing vehicle control + reagent and treated cells). We prepared the reagent according to the manufacturer’s instructions (Promega). Cells were incubated for 30 min and the luminescence was measured with the TECAN Spark plate reader.

In order to compare differences in cell viability between different inhibitor concentrations, we conducted a two-way ANOVA analyses. On the other hand, differences in the migration, the colony formation, and the apoptosis assays were analyzed with a one-way ANOVA. For both, multiple comparison testings were performed on all possible pairwise means. We used GraphPad Prism (GraphPad Software, San Diego, California, USA) for graphical representation. All figures show mean values ± standard deviation (SD). Synergistic interactions between the inhibitor and radiation therapy were evaluated using the online tool Synergy Finder using a zero-interaction potency (ZIP) model.

In order to assess the cytotoxicity of the PF in HNSCC, we performed the dose–response assay. For these, 5000 cells/well were seeded into 96-well plates with in 100 μl DMEM for each cell line. After 24 h of incubation, we treated cells with increasing concentrations of the inhibitor. We observed significant anti-proliferation effects of PF in all three cell lines (Fig. 1). In particular, the IC50 concentration for SCC154 was 20 μM. Meanwhile, it was 23 μM for both, the Cal27 and the FaDu cell line, indicating a slightly higher sensitivity of the HPV-positive cell line. Regarding synergy, cell viability data obtained from the dose–response assay was used to calculate synergy scores. Synergy reflects an excess response due to drug and radiation interactions. Importantly, scores higher than 10 are interpreted as synergistic interactions, scores between 10 to − 10 as additive effects, and scores under − 10 suggest antagonistic effects. Here, we observed mostly additive and synergistic effects of radiation and PF treatment in all three cell lines. Furthermore, the strongest synergy levels shown for SCC154 (Fig. 2).

We utilized a wound healing assay to assess the effects of PF on the migratory potential of three HNSCC cell lines and graphically depicted these results in Fig. 3, showing mean gap closure ± SD. As aforementioned, the corresponding IC50 values of PF were used for treatment. In particular, we used 24-well plates and the gap was created with Ibidi insets. In total, strong anti-migratory effects were observed in all three cell lines, particularly in FaDu and Cal27. Specifically for Cal27, treatment with 23 μM PF significantly reduced migration ability after 24 h and 48 h. Similarly, strong anti-migratory potential was shown in FaDu cells after treatment with 23 μM of PF and gap closure was significantly reduced at the 24 and 48 h time points (p < 0.0001 and p = 0.0255, respectively). A slightly less potent anti-migratory potential of PF was observed in SCC154. Still, the gap closure significantly decreased 24 h and 48 h after treatment with 20 μM of PF (p = 0.0118 and p = 0.0287, respectively). Furthermore, similar effects were observed in all cell lines after treatment with lower inhibitor concentrations.

Next, in order to assess the colony forming abilities, we performed the clonogenic assay with 1000 SCC154 cells/well and 350 cells/well for Cal27 and FaDu seeded into 12-well plates. Cal27 and SCC154 cells were treated with three PF concentrations and a DMSO vehicle control. In total, strong anti-clonogenic properties of PF were shown in both cell lines. These results are shown in Fig. 4. The surviving fraction reflects the percentage of surviving colonies normalized to the vehicle control value. PF treatment of the FaDu cell line did not deliver reproducible results, and therefore, we excluded these from the analysis. One-way ANOVA multiple comparison test showed significant differences for increasing inhibitor concentrations in both cell lines.

Finally, we assessed the pro-apoptotic potential of PF with the Caspase 3/7 Glo assay. For this, 7500 Cal27 and FaDu cells/well and 15,000 SCC154 cells/well were seeded into 96-well plates in 100 μL DMEM. All three cell lines were treated the corresponding IC50 PF concentration. Notably, the treatment induced apoptosis in all three cell lines. In particular, a significant increase was observed in FaDu and SCC154 cells. These results are depicted in Fig. 5.