T2R5 agonist phendione decreases cell viability and induces apoptosis in head and neck squamous cell carcinoma

Abstract

Bitter taste receptors (T2Rs), a family of G-protein coupled receptors, are emerging as potential therapeutic targets in head and neck squamous cell carcinoma (HNSCC). Phendione, a known T2R5 agonist, has not been previously investigated in HNSCC. Here, we show that phendione activates endogenously expressed T2R5 in HNSCC cells and ex vivo tumor samples, inducing sustained calcium responses, reducing cell viability, and promoting apoptosis through a T2R5-dependent mechanism. Analysis of The Cancer Genome Atlas data revealed that high T2R5 expression in HNSCC tumors correlates with improved long-term disease-specific survival, suggesting a potential tumor-suppressive role for T2R5. These findings highlight T2R5 as a promising therapeutic target in HNSCC and support further investigation of phendione or other T2R5 agonists as potential anti-cancer agents.

Introduction

Head and neck squamous cell carcinomas (HNSCCs) are a group of malignancies arising from the epithelial tissues of the oral cavity, pharynx, larynx, and sinonasal cavity (1). HNSCC ranks as the seventh most common cancer globally, with a five-year average survival of 60–70%. Major risk factors for HNSCC include tobacco and alcohol use, as well as human papilloma virus (HPV) (1,2). Despite public health efforts, such as HPV vaccination and smoking cessation initiatives (3,4), the incidence of HNSCC is projected to rise by approximately 30% annually by 2030 (2,5). Current treatments, including surgery, chemotherapy, and radiation therapy, are often associated with significant long-term impacts on patients’ quality of life (6), underscoring the need for novel, less invasive therapeutic strategies.

Bitter taste receptors (taste family 2 receptors or T2Rs) are a family of G-protein coupled receptors (GPCRs) with 26 functional isoforms (7,8). While T2Rs are traditionally associated with detecting bitter compounds on the tongue, they are also expressed in extra-oral tissues (9–11) where they may regulate diverse physiological and pathological processes. Recently, we and others have shown that T2Rs regulate cell proliferation in malignancies such as HNSCC (10,12–15). T2Rs frequently induce calcium (Ca²⁺) release from endoplasmic reticulum (ER) stores via inositol-1,4,5-triphosphate (IP3) (16,17). In HNSCC, T2R isoform 14 and 4 (T2R14 and T2R4) activation induce apoptosis through mitochondrial Ca²⁺ overload, reactive oxygen species (ROS) production, and caspase-3/7 activation (13,14). Notably, preclinical evidence suggests that T2R14 agonist, lidocaine, may activate apoptosis in cancer cells and/or inhibit metastasis (18–20), prompting clinical trials to evaluate its efficacy in solid tumors including breast cancer (21) and HNSCC (22).

Unlike T2R14, which is activated by over 100 known bitter compounds (23,24), other T2R isoforms demonstrate greater ligand selectivity. T2R5, encoded by the TAS2R5 gene has only 12 identified agonists, including heterocyclic metal chelator 1,10-phenanthroline (25–28). Among T2R5 agonists, the derivative 1,10-phenanthroline-5,6-dione (phendione) exhibits relatively high potency for T2R5 in human airway smooth muscle cells (29). Phendione-based therapies have also been studied as effective antibacterial, antifungal, and antitumor agents (30–34). In contrast to cisplatin, a commonly used chemotherapeutic agent that is first line for HNSCC, phendione does not cause double stranded DNA-breaks (35–37). Instead, phendione exerts its potential antitumor effects through a distinct mechanism of action that has yet to be fully characterized.

Some T2Rs, including T2R5, are upregulated in malignant tissues compared to normal tissue (13), but their functional roles in cancer remain underexplored (13). It is unknown, for example, if all T2Rs in HNSCC cells can activate apoptosis. Given that T2R14 and T2R4 activation causes cell death in HNSCC and phendione is a relatively specific and potent T2R5-agonist, we hypothesized that phendione would induce cell death and decrease cell viability in HNSCC via T2R5. We investigated the role of T2R5 in HNSCC to evaluate the potential of phendione or other T2R5 agonists as chemotherapeutic agents. Using in vitro and ex vivo models, we demonstrate that phendione induces intracellular Ca²⁺ responses, reduces cell viability, and promotes apoptosis through T2R5-dependent mechanisms. Furthermore, we explore the association between T2R5 expression and patient survival, offering new insights into the therapeutic potential of T2Rs in HNSCC.

Materials and Methods

Cell Culture

Human cell lines SCC47 (lateral tongue, HPV+) (UM-SCC-47; Millipore SCC071, RRID:CVCL_7759), FaDu (hypopharynx, HPV-) (ATCC HTB-43, RRID:CVCL_1218), RPMI 2650 (nasal septum, HPV-) (ATCC CCL-30, RRID:CVCL_1664), SCC4 (tongue, HPV-) (ATCC CRL-1624, RRID:CVCL_1684), and SCC90 (base of tongue, HPV+) (ATCC CRL-3239, RRID:CVCL_1899) were cultured submerged in high glucose DMEM (Corning; Glendale, AZ, USA) supplemented with 10% FBS (Genesee Scientific; El Cajon, CA, USA), 1% penicillin/streptomycin mix (P/S) (Gibco; Gaithersburg, MD, USA), and 1% nonessential amino acids (NEAA)(Gibco). Primary gingival keratinocytes (PCS-200–014) were grown submerged with Dermal Cell Basal Medium (PCS-200–030) with Keratinocyte Growth Kit (PCS-200–040) from ATCC.

Tissue was acquired from HNSCC patients undergoing oncologic surgeries after obtaining written informed consent. The study was conducted in accordance with the Declaration of Helsinki with approval from the Institutional Review Board of the University of Pennsylvania (IRB #417200). Patient ages ranged from 51.1 to 77.0 years; all included participants were male. Patients were only included if they were undergoing curative-intent surgery for primary HNSCC and had adequate additional tissue available for research. Patients were excluded if they had pathologies other than HNSCC. No patient or sample attrition occurred during the study. Sex of patients was recorded but was not used as a variable in analysis as all patients were male. All specimens processed were included in the final analysis. Ex vivo human tumor samples were sectioned into 300 μm thick tumor slice cultures using a vibratome (Precisionary Compresstome^® Vibratome VF-510–0Z) with the following settings: oscillations 6, speed 3.

Immunofluorescence

Cells were grown on chambered glass coverslips (CellVis) and fixed with 4% paraformaldehyde in PBS and subsequently blocked and permeabilized in buffer containing 1% BSA, 0.1% Triton X-100 in PBS. Primary polyclonal T2R5 antibodies (Bioss) were applied at 1:250 dilution and incubated overnight at 4°C. Secondary antibody Alexa-Fluor-647 (Life Technologies, RRID: AB_2536183), were incubated for 2 hours at a 1:500 dilution. Cleaved caspase-3 antibody (Abcam, RRID: AB_725947) was used at 1:100 dilution with secondary antibody donkey α-rabbit Alexa-Fluor-546 (Life Technologies, RRID: AB_2534103) at 1:1000 for 1 hour. DAPI was used as a nuclear counterstain using Fluoromount Mounting medium (Thermo Fisher Scientific). Images were captured using the Olympus Live Cell imaging System with 40x objectives. For ex vivo tumor slices, phalloidin conjugated to Alexa fluor-488 (Life Technologies) was used as an actin counterstain. Images were processed through Fiji version of ImageJ software (RRID:SCR_002285) (38).

Reverse Transcription Quantitative PCR (RT-qPCR)

Cells were lysed in TRIzol (Thermo Fisher Scientific), and RNA was isolated and purified using the Direct-zol RNA kit (Zymo Research). Reverse transcription was performed with High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific), and TAS2R5 expression was quantified using TaqMan qPCR assays. Ubiquitin C (UBC) was selected as the housekeeping gene for normalization, due to its stable expression in cancer cells (39).

siRNA Transient Knockdown TAS2R5

SCC47 cells were transfected with siTAS2R5 (hs.Ri.TAS2R5.13) or siScramble control (IDT DsiRNA TriFecta kit) with RNAiMAX lipofectamine reagent (Thermo Fisher Scientific) and Opti-MEM (Gibco). Suspended transfections were performed on day 1, with additional transfection on day 3. Knockdown efficiency was validated by western blot, showing sufficient T2R5 protein knockdown compared to control.

TAS2R5 Overexpression

FaDu cells were transfected with a pcDNA3.1(+) vector containing full-length human TAS2R5 (Invitrogen) using Lipofectamine 3000 and P3000 reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. Cells were collected 24–48 hours post-transfection for functional assays. Overexpression efficiency was assessed by RT-qPCR, which demonstrated marked upregulation of TAS2R5 mRNA compared to empty vector controls.

Live Cell Ca²⁺ Imaging

Cells were loaded with 5 μM of Fluo-4-AM (Thermo Fisher Scientific) for 1 hour at room temperature in the dark. Imaging was performed in Hank’s Balanced Salt Solution (HBSS) buffered with 20 mM HEPES (pH 7.4) containing 1.8 mM Ca²⁺. Cells were imaged on 8-well glass bottom chamber slides using Olympus IX-83 microscope (20x 0.75 NA PlanApo objective) with FITC filters (Sutter Lambda LS), Orca Flash 4.0 sCMOS camera (Hamamatsu, Tokyo, Japan), Metamorph (Molecular Devices), and XCite 120 LED Boost (Excelitas Technologies). 1,10-phenanthroline-5,6-dione (phendione) was obtained from Fisher Scientific (Catalog No. 456080010). Ligands (phendione, 1,10-phenanthroline, procyanidin trimer C2) were dissolved in HBSS and added in real-time during imaging. Final DMSO concentration did not exceed 0.1%. Images were analyzed through Fiji ImageJ software (38).

For zero-extracellular Ca²⁺ (“0-Ca²⁺”) conditions, cells were loaded with Fluo-4-AM in HBSS, agonists tested were dissolved in Ca²⁺-free HBSS with 10 mM EGTA (pH = 7.4). At the time of ligand injection, cells were bathed in 60 μL of Ca²⁺ containing HBSS and agonist was added in a volume of 300 μL and a final concentration of ~8.3 mM EGTA. We previously calculated that this is equal to final free Ca²⁺ concentration of 2 nM using MaxChelator (Chris Patton, Stanford University, RRID:SCR_000459) (14).

For ER Ca²⁺ imaging, pCMV-ER-LAR-GECO1 (Addgene_61244) (40) was transfected into cells using Lipofectamine 3000 (Thermo Fisher Scientific) 24 hours prior to imaging. Gα_q inhibitors YM254890 (FUJIFILM Wako Chemicals) and FR900359 (Sapphire North America) were incubated at 1 μM and 10 μM, respectively, for 1 hour at room temperature in the dark. Pertussis toxin (Tocris) (50 ng/mL) was incubated overnight.

Cell Viability: Crystal Violet and ATP Detection Assays

Cells were treated with phendione in high-glucose DMEM +10% FBS, +1% P/S, +1% NEAA for 24 hours at 37° C in a humidifier incubator at 5% CO₂. Remaining adherent cells were stained with 0.1% crystal violet, washed, dried, and dissolved in 30% acetic acid. Absorbance at 590 nm was measured using a Tecan plate reader (Tecan Spark 10M; Mannedorf, Switzerland) and normalized to untreated controls.

For ex vivo tumor slices, viability changes at 24 hours were assessed using CellTiter 96 Aqueous One solution MTS Cell Viability dye (Promega). Adjacent tumor slices were incubated in phenol-free, high glucose DMEM +10% FBS, +1%P/S, +1% NEAA with or without experimental treatments, MTS cell viability dye was added at 22 hours, to both tumor slices and blanks, and incubated for an additional 2 hours for total duration of 24 hours. Adjacent tumor slices have previously been shown to be metabolically similar, serving as technical replicates (41). After incubation period, 50 μL of media from tumor slices and blanks were transferred to a new well and absorbance at 490 nm was measured using Tecan plate reader. Results were normalized to untreated controls for each patient.

An ATP detection assay (Cayman Chemical) was used to validate cell viability after phendione exposure. SCC47 cells were treated with phendione or control media for 24 hours. Cells were then placed on ice and washed with ice-cold sterile PBS. Cells were homogenized with appropriate buffers and stored in -20 °C overnight. Non-diluted samples and standards were processed per the manufacturer’s instructions and read on a white plate using Varioskan LUX luminescence plate reader. ATP concentrations for all samples were calculated based on the log-transformed readings of the linear regression of standards, outlined by Cayman Chemical kit instructions.

Apoptosis Assays

Cells were seeded into 24-well black glass-bottom plates (CellVis) to achieve ~70% confluency at the time of the experiment. Treatments were prepared in phenol-free, high glucose DMEM +10% FBS, + 1% P/S, and + 1% NEAA. CellEvent Caspase 3/7 indicator dye (Thermo Fisher Scientific) was added per manufacturer’s instructions. Images were acquired every 25 minutes for 24 hours using a Tecan Plate reader with 5% CO₂ at 37° C. Representative images were taken on microscope described above.

The Cancer Genome Atlas (TCGA) Analyses

TAS2R5 expression data were obtained from the TCGA, TARGET, GTEx combined cohort via UCSC Xena (https://xenabrowser.net/ ). DESeq2-standardized counts (RRID:SCR_015687) were used to compare TAS2R5 expression levels across datasets. For comparison of TAS2R5 expression between tumor and normal samples, log2-transformed DESeq2 standardized counts were utilized. Kaplan-Meier survival analyses were performed using Xena to evaluate the correlation of TAS2R5 expression with overall and disease-specific survival (OS and DSS). For ROC-derived cut-offs for OS and DSS curves, a binormal ROC curve was fitted using JLABROC4 (https://englab.rad.jhmi.edu/jrocfit/JROCFITi.html ). Due to the AUC being <0.5, the decision rule was inverted (lower values indicate positivity). For OS, a threshold of 3.2080 was selected because it maximized Youden’s J (TPF-FPF), with sensitivity 0.3411 and specificity 0.3992 at that point. For DSS, a threshold of 2.9140 was selected, as it maximized Youden’s J, with sensitivity 0.4415 and specificity 0.5004 at that point (42). An online and publicly available ROC Curve calculator was utilized which is available at: https://englab.rad.jhmi.edu/jrocfit/JROCFITi.html. Graphs were generated in GraphPad Prism (RRID:SCR_002798, San Diego, CA, USA).

Statistical Analyses

Statistical analyses were conducted using GraphPad Prism (San Diego, CA, USA). T-tests were used for two-group comparisons, while one-way ANOVA was applied for comparisons across multiple groups. Sample sizes were chosen based on prior publications demonstrating sufficient power to detect treatment effects in similar models; no formal power analysis was performed. Statistical tests used are indicated in the figure legends, with significance indicated as follows: P < 0.05 (*), P < 0.01 (**), P < 0.001 (***), P < 0.0001 (****). Error bars represent mean ± SEM unless otherwise indicated.

Data Availability

All data generated in this study are shown in the manuscript. The raw data points used to generate graphs and figures are available upon request from the corresponding author.

Results

Endogenous Expression of T2R5 in HNSCC Cell Lines and Patient Derived Tumor Slices

The bitter taste receptor T2R5 has been implicated in various physiological and pathological processes (29,43,44), but its expression in HNSCC remains underexplored. To determine whether TAS2R5 is endogenously expressed in HNSCC, we assessed its expression across multiple cell lines representing both HPV+ and HPV− HNSCC subtypes. We found that TAS2R5 is variably expressed in both HPV+ (SCC47) and HPV- (RPMI 2650, SCC4, FaDu) HNSCC cell lines, as detected by RT-qPCR (Figure 1A). Comparisons to non-malignant primary gingival keratinocytes showed lower expression of TAS2R5 compared with SCC47, RPMI 2650, and SCC4 cells but no significant difference compared with FaDu cells.

Figure 1: TAS2R5 is expressed in HNSCC cell lines and patient-derived tumors.

A Relative mRNA expression of TAS2R5 was measured in primary keratinocytes and compared to HNSCC cell lines FaDu, SCC4, SCC47, SCC90, and RPMI2650 by qPCR normalized to ubiquitin C (UBC) housekeeping gene, (mean ± SEM, n > 3 separate passages). B Mean area of intensity for IF of SCC47 and RPMI2650 n= 4 wells, C T2R5 protein expression measured via immunofluorescence in SCC47 (DAPI = nucleus). D Ex vivo patient-derived tumor slice from oral tongue HPV- HNSCC showing diffuse T2R5 staining (DAPI = nucleus; phalloidin = (green, actin). Each antibody was compared to secondary only control at the same microscope settings. Significance by 1-way ANOVA with Dunnett’s multiple comparison posttest. *p < 0.05, **p < 0.01, ***p < 0.001; ns, no statistical significance.

Immunofluorescence using a T2R5 antibody revealed diffuse staining in both SCC47 (Figure 1C) and RPMI 2650 (Supp. Figure 1) cells. Immunofluorescence images were quantified based on mean area intensity in SCC47 and RPMI2650, which were consistent with RT-qPCR results (Figure 1B). These findings suggest that T2R5 is endogenously expressed in HNSCC cell lines. Immunofluorescence of patient derived tumor samples (Figure 1D) also showed T2R5 staining, suggesting that T2R5 is present within HNSCC tumors.

Stimulation with Phendione Induces a Sustained Intracellular Ca²⁺ Response in HNSCC

Previous studies have identified 1,10-phenanthroline as the only compound among ~100 bitter ligands capable of activating T2R5 in heterologously expressed HEK-293T cells (26). In human airway smooth muscle cells, its derivative, 1,10-phenanthroline-5,6-dione (phendione), induces Ca²⁺ responses at much lower concentrations, suggesting that phendione is a more potent T2R5 agonist (29). To test if phendione stimulation elicits a Ca²⁺ response in HNSCC, live cell imaging was performed on cells loaded with the Ca²⁺ indicator dye Fluo-4.

In SCC47 cells, phendione stimulation at concentrations as low as 1.4 μM induced a significant intracellular Ca²⁺ response that was sustained for ≥30 minutes (Figure 2A-B). Similar responses were observed in RPMI 2650, FaDu, and SCC4 cell lines with 1.4 and 142 μM phendione (Figure 2C-E), while SCC90 cells demonstrated a significant response only at 142 μM (Figure 2F). In primary gingival keratinocytes, phendione also elicited significant Ca²⁺ responses (Figure 2G-H), but the response was oscillatory rather than sustained over time (Figure 2A). Notably, in SCC47, the sustained Ca²⁺ responses persisted for ≥5 hours (Figure 2I-J) and resulted in morphological changes like blebbing seen under increased magnification (Supp. Figure 2). These findings highlight differences in T2R5-activated Ca²⁺ response kinetics between malignant and non-malignant cells.

Figure 2: Phendione stimulation results in sustained, intracellular Ca²⁺ responses.

HNSCC cell lines were loaded with Ca²⁺ binding dye Fluo-4-AM and stimulated with T2R agonist phendione. A-B SCC47 Ca²⁺ responses over time (A, minutes) and peak (B, mean ± SEM) after phendione stimulation. C-F Peak Ca²⁺ response after phendione stimulation for RPMI2650 (C), FaDu (D), SCC4 (E), and SCC90 (F). G-H Primary keratinocyte Ca²⁺ responses over time (C, minutes) and peak (D, mean ± SEM) after phendione stimulation. Significance by 1-way ANOVA with Dunnett’s multiple comparison posttest comparing HBSS to each agonist concentration. I-J Extended evaluation of SCC47 Ca²⁺ responses to phendione demonstrate a steady increase that peaks around 4 hours. *p < 0.05, **p < 0.01, ***p < 0.001, ****p<0.0001; ns, no statistical significance.

To assess the source of Ca²⁺ mobilization, phendione stimulation was repeated under zero extracellular Ca²⁺ (0-Ca²⁺) conditions, achieved by chelating extracellular Ca²⁺ with EGTA. Under these conditions, phendione induced an immediate but transient Ca²⁺ peak that was not sustained, unlike the response observed under normal extracellular Ca²⁺ (1.8 mM Ca²⁺; Figure 3A-B). This indicates that the initial Ca²⁺ response primarily arises from intracellular Ca²⁺ stores, likely the ER.

Figure 3: Phendione intracellular Ca²⁺ responses are T2R5 dependent.

A-B To assess the source of Ca²⁺ mobilization, phendione stimulation was repeated under 0-extracellular Ca²⁺ conditions (0-Ca²⁺), achieved by chelating extracellular Ca²⁺ with EGTA. Phendione induced an immediate, but transient Ca²⁺ peak that was not sustained, suggesting that the initial Ca²⁺ response primarily arises from intracellular Ca²⁺ stores. C-D The role of ER Ca²⁺ stores was evaluated by transfecting SCC47 cells with the fluorescent ER Ca²⁺ biosensor pCMV ER LAR GECO1. Phendione stimulation caused a steady and significant depletion of ER Ca²⁺compared to HBSS control (C). Endpoint Ca²⁺ levels in the ER after 20 minutes were significantly lower in the phendione condition versus HBSS (D). E SCC47 transient knockdown of TAS2R5 was generated using siRNA (siTAS2R5) and compared to nontargeting human siRNA (Scramble). Peak Ca²⁺ responses (mean ± SEM) were significantly reduced for siTAS2R5 compared to siScramble suggesting T2R5-dependent signaling. F-G SCC47 cells were loaded with Ca²⁺ binding dye, Fluo-4, and treated with Gαq inhibitors YM254890 (F) and FR900359 (G). Gαq inhibitors significantly blunted the peak Ca²⁺ responses to phendione compared to untreated controls (mean ± SEM), suggesting Gα involvement. H-I Cyclic AMP (cAMP) production was assessed using fluorescent biosensor pcDNA3.1-pFlamindo2. In contrast to most bitter agonist signaling to Gα_i/o which leads to decreased cAMP, stimulation of SCC47 cells with phendione resulted in an increase in cAMP levels compared to the HBSS control. Isoproterenol (positive control) induced a robust increase in cAMP production. Significance by unpaired t-test. Significance by 1-way ANOVA with Dunnett’s multiple comparison posttest. *p < 0.05, **p < 0.01, ***p < 0.001; ns, no statistical significance.

The role of ER Ca²⁺ stores was further evaluated by transfecting SCC47 cells with the fluorescent ER Ca²⁺ biosensor ER-LAR-GECO1 (40). Phendione stimulation caused a steady and significant decrease of ER Ca²⁺ compared to HBSS control, as shown by a gradual decrease in fluorescence (Figure 3C). Endpoint Ca²⁺ levels in the ER after 20 minutes were significantly lower in the phendione condition versus HBSS control (Figure 3D). Together, these data suggest that phendione mobilizes Ca²⁺ from ER stores, a typical characteristic of T2R activation (7,45,46). However, the sustained Ca²⁺ response observed in SCC47 cells was absent under 0-Ca²⁺ conditions, indicating a requirement for extracellular Ca²⁺ influx during prolonged stimulation.

To evaluate Ca²⁺ responses to other previously reported T2R5-selective agonists, SCC47 cells were stimulated with 1,10-phenanthroline and procyanidin trimer C2 (procyanidin) (26,47–49). Significant dose-dependent Ca²⁺ responses were observed for both procyanidin and 1,10-phenanthroline (Supp. Figure 3). These data provide further evidence that T2R5 agonism can lead to significant, sustained Ca²⁺ responses in HNSCC cell lines.

Phendione Acts on Endogenously Expressed T2R5 Likely Coupled to Gα_q in HNSCC

To determine whether the phendione-induced Ca²⁺ response is mediated by T2R5, transient knockdown of TAS2R5 was performed using siRNA (siTAS2R5). This resulted in a significant reduction in Ca²⁺ response compared to cells transfected with a nontargeting siRNA control (siScramble), indicating that the Ca²⁺ response is primarily T2R5-dependent (Figure 3E). Successful T2R5 protein knockdown was confirmed with Western blot (Supp. Figure 4).

To further support these results, heterologous overexpression of TAS2R5 cloned in pcDNA3.1 was performed in FaDu cells. Cells with overexpressed TAS2R5 showed an increased Ca²⁺ response to phendione compared to vehicle control (Supp. Figure 5A). Successful upregulation of TAS2R5 was confirmed via RT-qPCR, which showed near 300-fold upregulation (Supp. Figure 5C). These knockdown and overexpression results align with published data demonstrating the selectivity of 1,10-phenanthroline (parent compound to phendione) for T2R5 over other T2Rs (26).

To explore the role of G-protein coupling in the phendione-induced Ca²⁺ response, live cell imaging with Fluo-4 was conducted in the presence of specific G-protein inhibitors. Two Gα_q inhibitors, YM254890 and FR900359 (50,51), significantly blunted the Ca²⁺ response to phendione compared to untreated controls (Figure 3F-G). In contrast, treatment with pertussis toxin (PTX), a Gα_i/o inhibitor that we showed reduces T2R14-mediated responses (14), did not significantly decrease the Ca²⁺ response (Supp. Figure 6A). Additionally, the T2R14 antagonist LF1 (52) did not attenuate the Ca²⁺ response to phendione (Supp. Figure 6B), further supporting that phendione targets T2R5 and not other bitter taste receptors such as T2R14.

To investigate secondary messenger signaling, cyclic AMP (cAMP) production was assessed using fluorescent biosensor pFlamindo2 (53), which enables real-time visualization of cAMP dynamics. T2Rs coupled to Gα_i/o can inhibit adenylate cyclase, leading to a decreased in cAMP (54,55). However, stimulation of SCC47 cells with phendione resulted in a steady and significant increase in cAMP levels compared to the HBSS control (Figure 3H-I). Isoproterenol, a β-adrenergic agonist, was used as a positive control to induce cAMP production (56,57). These results support that phendione does not activate Gα_i/o but rather signals through a distinct GPCR pathway, likely involving Gα_q. Ca²⁺ responses with stimulation by another T2R5 agonist, procyanidin, was also reduced by Gα_q inhibitor YM254890, further suggesting that T2R5 couples to Gα_q (Supp. Figure 3C).

Phendione Exposure Reduces Cell Viability in a Dose-Dependent Manner

Phendione was previously identified as a potential chemotherapeutic with cytotoxic effects (58). To evaluate its impact on HNSCC cells via T2R5, crystal violet assays were performed to measure cell viability following 24-hour phendione exposure. A dose-dependent reduction in cell viability was observed across HNSCC cell lines (Figure 4). Concentrations as low as 140 nM caused a significant decrease in cell viability in SCC47 and RPMI 2650 (Figure 4A-B), while SCC4 and FaDu viability was significantly reduced at 1.42 μM (Figure 4C-D). SCC90 was the least responsive, requiring a dose of 14.2 μM to decrease viability (Figure 4E).

Figure 4: T2R5 stimulation leads to HNSCC apoptosis and reduced cell viability.

A-G Cell viability was measured following 24-hour phendione exposure using crystal violet and measured at 590 nm absorbance. Dose dependent reductions in cell viability were observed in cancer cell lines (A-E) and primary keratinocyte (F; mean ± SEM, one-way ANOVA Dunnett’s multiple comparison test). G The percent viability after 142 μM phendione exposure was significantly lower for cancer cell lines compared to primary keratinocytes (mean ± SEM, one-way ANOVA Dunnett’s multiple comparison test). H ATP detection was performed following 24 hours of phendione exposure, indicating a dose dependent reduction in cell viability (mean ± SEM, one-way ANOVA Dunnett’s multiple comparison test). I SCC47 siTAS2R5 transient knockdown exhibited less robust decreases in cell viability after phendione treatment compared to control siScramble (mean ± SEM, unpaired t-test with Welch’s correction). J Viability of HNSCC tumor slices from n = 5 patients indicates significant reduction after 24 hours of treatment with 142 μM phendione. No significant reduction is observed at 1.4 μM phendione. Significance by 1-way ANOVA with Dunnett’s multiple comparison posttest. K-L To evaluate if phendione induces apoptosis, SCC47 cells were exposed to control (media only) and 142 μM phendione in media for 5 hours and caspase 3 cleavage was evaluated via immunofluorescence (K) and quantified by mean area intensity (L). M-N To further evaluate if apoptosis, SCC47 cells were stained with CellEvent, a dye that fluoresces upon activation of caspase 3/7 cleavage. Phendione increased fluorescence compared to controls over 24 hours (I, J; mean ± SEM; unpaired t-test). O Representative images at 24 hours with Hoechst dye (counter nuclear stain) and CellEvent dye in control and phendione condition. *p < 0.05, **p < 0.01, ***p < 0.001; ns, no statistical significance.

In primary gingival keratinocytes, phendione also caused a dose-dependent reduction in cell viability (Figure 4F). However, keratinocytes showed greater resistance to phendione at 142 μM phendione compared to HNSCC cell lines (Figure 4G), suggesting that, although phendione is not entirely cytoselective for cancer cells, normal cells may exhibit greater resistance to its cytotoxic effects. Calculated IC₅₀ values for phendione in HNSCC cell lines range from ~0.8 – 4.8 μM (Supp. Figure 7), consistent with previously reported EC₅₀ values in other cell lines (29,59). Cell viability effects were further validated with an ATP detection assay which showed a dose-dependent reduction in ATP concentrations in cell lysates after 24-hour exposure to phendione (Figure 4H).

To investigate whether the reduction in cell viability is mediated through T2R5, SCC47 cells with transient T2R5 knockdown (siTAS2R5) were evaluated. Cells transfected with siTAS2R5 demonstrated improved viability at 142 μM phendione compared to cells transfected with non-targeting siScramble controls (Figure 4I). To provide additional support for T2R5 specificity, FaDu cells were used for heterologous overexpression with TAS2R5 cloned in pCDNA3.1 vector to upregulate T2R5. Compared to vehicle, over-expressed T2R5 cells had a significantly increased Ca²⁺ response to phendione (Supp. Figure 5A) and were more sensitive to phendione’s cytotoxicity, as seen at the sublethal dose of 14.2 μM (Supp. Figure 5B). Validation of overexpression was confirmed via RT-qPCR, with a large ~300 fold increase (Supp. Figure 5C). These data indicate that phendione acts via T2R5 to decrease HNSCC cell viability.

To further assess phendione’s toxicity, ex vivo tumor slice cultures derived from HNSCC patients were exposed to 1.4 and 142 μM phendione for 24 hours. Cell viability was measured using MTS dye, with results normalized to controls. Phendione significantly reduced cell viability at 142 μM (Figure 4J). The demographic information for the HNSCC tumor samples can be found in Supp. Table 1. This ex vivo data corroborates findings from HNSCC cell lines, further supporting anti-tumor cytotoxic effects of phendione and potential translational applications to patients.

To evaluate if phendione induces apoptosis, SCC47 cells were exposed to phendione (142 μM) or control for 5 hours and immunofluorescence was performed for cleaved caspase-3. Compared to control, phendione treated cells had higher intensity of staining for cleaved caspase-3 (Figure 4K-L). SCC47 cells were also stained with CellEvent, a dye that fluoresces upon activation of caspase 3/7 cleavage. Phendione (142 μM) significantly increased fluorescence compared to controls, indicative of caspase 3/7 activation (Figure 4M-O). Fluorescence began increasing at ~6 hours post-treatment and steadily rose over the 24-hour period. These results suggest that phendione induces apoptosis in HNSCC cells through caspase 3/7 activation, consistent with prior reports of similar mechanisms involving other T2Rs in HNSCC (13,14).

TAS2R5 Expression in HNSCC Tumors is Associated with Long-Term Patient Survival

To evaluate the clinical implications of TAS2R5 expression in HNSCC, TCGA data were analyzed to compare patients with low versus high TAS2R5 expression levels based on bulk RNA-sequencing of tumor samples. Kaplan-Meier survival analyses demonstrated improved 5-year overall survival and disease-specific survival for cases with high TAS2R5 expression compared to low TAS2R5 expression (Figure 5A-B) which was also observed at 10 years (Supp. Figure 8). Further analysis using ROC curve-derived cut offs for high and low TAS2R5 expression highlighted that 5-year DSS, but not OS, was significantly improved in patients with elevated TAS2R5 in their tumors (Supp. Figure 9), supporting the potential of TAS2R5 as a biomarker for enhanced survival in HNSCC. These findings suggest a potential endogenous tumor-suppressive role of TAS2R5 in regulating tumor proliferation in HNSCC patients.

Figure 5: Increased TAS2R5 expression is associated with improved survival.

A-B Using bulk RNA-sequencing data from The Cancer Genome Atlas (TCGA), HNSCC patients were divided into groups based on high and low TAS2R5 expression. Kaplan–Meier survival analysis demonstrated improved 5-year overall survival (OS) (A) and disease-specific survival (DSS) (B) for cases with high TAS2R5 expression compared to low TAS2R5 expression (p = 0.014 and 0.015 by log-rank test, respectively). C Comparison of log2 transformed expression levels indicated a slight increase in TAS2R5 expression in tumor compared to adjacent normal tissue in HNSCC patients. *p < 0.05. D Comparison of log2 transformed expression levels of TAS2R5 indicate no significant difference between HPV+ and HPV– HNSCC tumors (n= 113, Welch’s t-test p = 0.61).

Our prior study highlighted differential expression of TAS2Rs between tumor and normal tissues, with TAS2R5 showing higher expression in tumor samples (13). To evaluate this further, TAS2R5 expression levels in TCGA were compared between tumor and adjacent normal tissues in HNSCC patients. Results suggest a slight but significant increase in TAS2R5 expression in tumor tissue (Figure 5C). Previously, we demonstrated that TAS2R14 expression was increased in HPV+ HNSCC tumors (14). To investigate if TAS2R5 expression is similarly influenced by HPV status, TCGA data was stratified into HPV+ and HPV− HNSCC tumors. No significant differences in TAS2R5 expression were observed between the two groups (Figure 5D). Unlike TAS2R14, TAS2R5 expression is not specifically upregulated in HPV+ tumors.

Discussion

Bitter taste receptors (T2Rs), a subclass of GPCRs, have increasingly been recognized for their crucial extra-oral roles in both normal (10,27) and malignant tissues (12,60). Recent studies have highlighted their therapeutic potential in HNSCC (13–15,61), with ongoing research that includes a phase I clinical trial led by our group to evaluate T2R14-agonist lidocaine as a neoadjuvant treatment for HNSCC (22). While T2R14 has been explored for its cytotoxic effects in HNSCC (13,14), the role of T2R5 – a highly selective receptor for specific bitter ligands (27) – remained unexplored until now. T2R5 has been implicated in extra-oral functions across multiple tissues, including the brain, uterus, skin, and lung (29,44,62,63), but its role in cancer had not been defined.

We demonstrate that phendione, a high-affinity T2R5 agonist, exerts cytotoxic effects in HNSCC via T2R5-dependent mechanisms. Phendione, a derivative of 1,10-phenanthroline (29), was previously noted to exhibit anticancer activity in epithelial cells (58). Unlike traditional metal chelators like cisplatin, phendione induces anti-cancer effects without causing DNA breaks (35), suggesting alternative cytotoxic pathways. Our data suggest that phendione activates T2R5, leading to sustained Ca²⁺ signaling, reduced cell viability, and apoptosis. These data may indicate T2R5 as a potential viable target for HNSCC therapies.

We showed two other known T2R5 agonists (procyanidin trimer C1 and 1,10-phenanthroline) induce a sustained Ca²⁺ response similar to phendione. Other T2R5 agonists, many of which are derivatives or phendione, have been studied for their effects of mitigating receptor desensitization and downregulation in human airway epithelial cells (64). This opens an opportunity for T2R5-agonist engineering for HNSCC targeting in the future.

T2R5 Expression in HNSCC

We identified endogenous expression of T2R5 in HNSCC, with varying levels across different cell lines. Primary keratinocytes exhibited the lowest T2R5 expression, potentially explaining their relative resistance to phendione-induced cytotoxicity compared to malignant cells. Ex vivo tumor slices also demonstrated T2R5 expression, supporting the clinical relevance of T2R5 as a therapeutic target. The observed differential expression between malignant and normal tissues suggests that T2R5 activation could selectively target cancer cells while minimizing off-target effects in normal tissues.

T2R5 induces Ca²⁺ Mobilization in HNSCC cells

Phendione stimulation induced a sustained intracellular Ca²⁺ response in HNSCC cells. This contrasts with the transient responses observed in primary keratinocytes and other T2R subtypes in HNSCC cells, like T2R14, where Ca²⁺ peaks more transiently (14,54). Sustained Ca²⁺ signaling has been implicated in mitochondrial dysfunction, ROS production, and caspase activation, all of which contribute to apoptosis (65–68). This could be particularly relevant for tumors with high T2R5 expression that are stimulated with a T2R5-agonist to induce or sensitize malignant cells to undergo apoptosis. Additionally, the responses to other T2R5 agonists, like 1,10-phenanthroline and procyanidin trimer C-2, further supports the role of T2R5 in the Ca²⁺ mobilization observed with phendione. Procyanidin trimer C-2, a potent polyphenol in fruits and berries (47), has been reported to exert preventative and anti-cancer effects in HNSCC (69–71). While some reports have evaluated the role of dietary intake of plant polyphenols in HNSCC (72), future studies could evaluate the effects of polyphenols that specifically stimulate T2Rs.

Interestingly, our data suggest that phendione/T2R5 activates Ca²⁺ signaling through Gα_q-mediated pathways. We have showed that denatonium, a highly bitter agonist that activates many T2Rs, is also inhibited by Gα_q-inhibitors (13). While some T2Rs are blocked by Gα_i/o inhibitors like pertussis toxin (14), T2R5-agonists are not. This G-protein coupling difference emphasizes the complexity of T2R signaling and opens new avenues for targeted therapies that can leverage specific signaling pathways. GPCRs as a whole make up the largest portion of currently approved therapeutics (73), and many current FDA-approved therapeutics are predicted to activate bitter taste receptors (23,52,74). It is thus important to characterize tissue specific T2R signaling for drug development, repurposing, and off-target effects.

GPCR signaling can converge on multiple oncogenic pathways, including ERK1/2, MAPK, and PI3K/AKT cascades (75–77). Activation of ERK1/2 in HNSCC has been linked to either proliferative or anti-proliferative effects depending on the strength and duration of signaling (78,79). Our data suggest that T2R5 couples to Gαq in SCC47 cells, raising the possibility that downstream PKC- and Ca²⁺-dependent MAPK activation could be engaged following phendione stimulation (80). MAPK signaling is well established in promoting angiogenesis and metastasis through VEGF expression and ERK/JNK activation in HNSCC (76,81). While not examined here, future studies will directly test whether T2R5-mediated Ca²⁺ mobilization modulates ERK/MAPK or PI3K/AKT activity, as this will be critical for understanding the broader oncogenic context of T2R5 signaling and for evaluating T2R5 agonists as therapeutic agents.

Cytotoxic Effects of T2R5 activation by Phendione

The dose-dependent reduction in cell viability observed with phendione treatment highlights its therapeutic potential in HNSCC. While primary keratinocytes also exhibited reduced viability, their relative resistance compared to malignant cells points to a role for differential T2R5 expression in mediating cytotoxic effects. This suggests that phendione or similar T2R5 ligands could be leveraged as selective chemotherapeutic agents, particularly in tumors with elevated T2R5 expression.

The induction of apoptosis further reinforces the potential specificity of T2R5-mediated cytotoxicity. Apoptosis is a critical mechanism for eliminating cancer cells while minimizing damage to surrounding tissues (82–84) and targeting T2R5 may provide an opportunity to exploit this pathway in HNSCC. Importantly, this approach may complement existing therapies, as apoptosis induction is often impaired in resistant tumors (85,86). Future work is needed to explore the potential of phendione or other T2R5 agonists to synergize with chemotherapeutic agents such as cisplatin, which relies on DNA damage pathways, or BAY-876, which targets GLUT1 and has demonstrated efficacy alone and with other T2R agonists in HNSCC (15,87).

Clinical Implications of TAS2R5 Expression

The positive association between high TAS2R5 expression and improved survival outcomes in HNSCC patients positions TAS2R5 as a promising prognostic biomarker. Its potential utility in stratifying patients based on tumor biology could aid in personalizing treatment regimens. For instance, the expression levels of TAS2R5 may serve as a biomarker for determining which patients require more intensive standard treatment regimens or those who can be safely de-escalated to minimize treatment side-effects. Alternatively, patients with high TAS2R5 expression might benefit from therapies targeting this specific receptor.

Interestingly, TAS2R5 expression did not differ between HPV+ and HPV− tumors, suggesting that its therapeutic relevance may extend across HNSCC subtypes. This is particularly significant for HPV− HNSCC, which is associated with poorer prognoses (88). The involvement of T2R5 in immune cell populations, such as macrophages (89–91) raises additional questions about its role in the tumor microenvironment. Understanding whether T2R5 contributes to tumor immune cell infiltration or function could provide further insights into its dual role as a therapeutic target and a biomarker (92).

Limitations and Future Directions

Although this study establishes the role of T2R5 in mediating Ca²⁺ signaling, cytotoxicity, and apoptosis in HNSCC, the lack of in vivo models represents a key limitation. Mouse models are particularly challenging for preclinical testing because of limited T2R homology, especially the absence of a direct ortholog for human T2R5 (28). To address translational relevance, we used patient-derived ex vivo tumor slice cultures, which preserve the native tumor microenvironment and have been shown to reflect clinical biology (93,94). While informative, tumor slice-based assays (e.g., MTS metabolic activity) cannot easily distinguish between effects on malignant versus stromal cells. Future work should therefore prioritize patient-derived xenografts (PDX) or engineered xenograft models expressing human TAS2R5 to bridge this translational gap.

Additionally, the variable expression of TAS2R5 across individuals underscores the importance of incorporating biomarker-driven approaches into prognostic or therapeutic development. The attenuated Ca²⁺ responses observed in SCC90 cells despite high TAS2R5 mRNA expression further highlight that receptor expression alone may not predict functional activity, reinforcing the need for future studies to define biomarkers of T2R5 signaling competence in HNSCC. The effects of phendione should also be further explored in translational contexts, including potential off-target effects such as fibrosis, a common complication of current HNSCC treatments (radiation, chemotherapy, surgery) (95),(96). Finally, exploring the combination of T2R5-targeted ligands with other treatments, such as immune checkpoint inhibitors, may expand the therapeutic landscape for HNSCC.

In conclusion, this study identifies T2R5 as a novel therapeutic target in HNSCC. Phendione is a potent T2R5 activator capable of inducing sustained Ca²⁺ responses, reducing cell viability, and promoting apoptosis. The association between high TAS2R5 expression and improved survival demonstrates potential for TAS2R5 as a prognostic marker. These findings pave the way for further exploration of T2R5 in both basic and translational cancer research, with potential to improve outcomes for HNSCC patients through precision medicine approaches.