Open Access
28 January 2025 Optical imaging provides flow-cytometry–like single-cell level analysis of HIF-1α-mediated metabolic changes in radioresistant head and neck squamous carcinoma cells
Jing Yan, Carlos Frederico Lima Goncalves, Pranto Soumik Saha, Cristina M. Furdui, Caigang Zhu
Author Affiliations +
Abstract

Significance

Radioresistance remains a significant problem for head and neck squamous cell carcinoma (HNSCC) patients. To mitigate this, the cellular and molecular pathways used by radioresistant HNSCC that drive recurrence must be studied.

Aim

We aim to demonstrate optical imaging strategies to provide flow cytometry–like single-cell level analysis of hypoxia-inducible factor 1-alpha (HIF-1α)-mediated metabolic changes in the radioresistant and radiosensitive HNSCC cells but in a more efficient, cost-effective, and non-destructive manner. Through both optical imaging and flow cytometry studies, we will reveal the role of radiation-induced HIF-1α overexpression and the following metabolic changes in the radioresistance development for HNSCC.

Approach

We optimized the use of two metabolic probes: 2-[N-(7-nitrobenz-2-oxa-1, 3-diazol-4-yl) amino]-2-deoxy-D-glucose (2-NBDG) (to report glucose uptake) and Tetramethylrhodamine ethyl ester (TMRE) (to report mitochondrial membrane potential) with both a standard fluorescence microscope and a flow cytometry device, to report the changes in metabolism between radioresistant (rSCC-61) and radiosensitive (SCC-61) HNSCC cell lines under radiation stresses with or without HIF-1α inhibition.

Results

We found that the matched HNSCC cell lines had different baseline metabolic phenotypes, and their metabolism responded differently to radiation stress along with significantly enhanced HIF-1α expressions in the rSCC-61 cells. HIF-1α inhibition during the radiation treatment modulates the metabolic changes and radio-sensitizes the rSCC-61 cells. Through these studies, we demonstrated that a standard fluorescence microscope along with proper image processing methods can provide flow cytometry–like single-cell level analysis of HIF-1α-mediated metabolic changes in the radioresistant and radiosensitive HNSCC cells.

Conclusions

Our reported optical imaging strategies may enable one to study the role of metabolism reprogramming in cancer therapeutic resistance development at the single-cell level in a more efficient, cost-effective, and non-destructive manner. Our understanding of radiation resistance mechanisms using our imaging methods will offer opportunities to design targeted radiotherapy for improved treatment outcomes for HNSCC patients.

Statement of Discovery

We demonstrate that optical imaging strategies can provide flow-cytometry–like single cell level analysis of HIF-1α mediated metabolic changes in the radioresistant and radiosensitive HNSCC cells, but in a more efficient, cost-effective, and non-destructive manner. We found that the matched HNSCC cell lines had different baseline metabolic phenotypes, and their metabolism responded differently to radiation stress along with significantly enhanced HIF-1α expressions in the radioresistant HNSCC cells. HIF-1α inhibition during the radiation treatment modulates the metabolic changes and radiosensitizers the radioresistant HNSCC cells.

1.

Introduction

Head and neck squamous cell carcinomas (HNSCCs) represent the sixth most common malignancy worldwide,1 and the overall 5-year survival rate is 50%.2 Depending on the stage and resectability, treatment options available for HNSCC patients include various combinations of surgery, radiotherapy (RT), and chemotherapy. RT alone or combined with chemotherapy has been used as a primary curative treatment prescribed for locally advanced HNSCC patients either as definitive or as adjuvant post-surgical therapy.3 Statistical data reported that more than 75% of the locally advanced HNSCC patients in the US receive RT as part of their care; however, over 50% of these RT-treated patients are prone to develop recurrence post-treatment,3 which leads to most deaths for the HNSCC patients.4 Given that HNSCC patients have a high risk for recurrence post-RT, it is crucial to understand the cellular and molecular pathways that support the selective or acquired survival of these cells in vivo, with an eventual goal of designing therapies to prevent tumor recurrence and improve the survival of HNSCC patients.

Increasing evidence shows that metabolic reprogramming may be responsible for radioresistance development.5 The tumor metabolic rewiring not only provides an unparalleled advantage to tumor cells to survive, grow, and metastasize under a hypoxic and nutrient-poor environment but also endows these cells with plasticity to adapt and escape immunosuppression and therapeutic treatment.6,7 Several groups reported that increased glucose metabolism is associated with radioresistance in breast cancers8 and lung cancers.9 Some other studies demonstrated that deregulated mitochondrial metabolism is also closely related to the radioresistance development in human cancers.10 Recent in vitro studies showed that radioresistant HNSCC cells have enhanced glycolysis and decreased oxidative phosphorylation (OXPHOS) compared with radiosensitive HNSCC cells.1113 Hypoxia-inducible factor-1α (HIF-1α), a master regulator of cellular oxygen sensing and adaptation to hypoxia, plays an essential role in tumor cell survival, growth, and spread.14 HIF-1α can be stabilized by hypoxia and also by reactive oxygen species (ROS)1517 such as those produced from RT.14 Previous studies reported that RT induced HIF-1α expression in radioresistant HNSCC cell lines.18 HIF-1α overexpression is known to enhance glycolysis and the Warburg effect, promote cancer stem cell-like characteristics, and boost treatment resistance19 and has also been shown to increase tumor angiogenesis to facilitate tumor cell proliferation, spread, and treatment resistance.20 Taking together, the HIF-1α-mediated changes in metabolism may promote tumor radioresistance and recurrence, targeting the HIF-1α, and the corresponding regulated metabolic changes will offer new opportunities to improve RT efficacy.21

In the studies described above, Seahorse assay or metabolomic analysis was used to quantify OXPHOS and glycolysis of tumor cells in vitro. Seahorse Assay22 and metabolomics23 provide valuable measurements of the metabolic phenotypes of tumor cells; however, they are limited to in vitro or ex vivo studies and cannot be used for repeated time course measurements due to their destructive nature. Although imaging tools such as PET24 or MRI25 complement the Seahorse Assay and metabolomics for live metabolic scanning, they are not suitable for cellular-level metabolic imaging due to low resolution or low sensitivity.26 Flow cytometry has also been extensively used to report cell metabolism using fluorescent probes at the single-cell level27 although it is a destructive technique with using expensive equipment. To overcome these limitations, optical imaging has emerged as a new strategy for the non-destructive analysis of multiple metabolic pathways in live cells. Optical measures of two endogenous tissue fluorophores, reduced nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FAD)28 have been explored to report the reduction-oxidation (redox) state in the electron transport chain of cancer cells.9,2931 To quantify tumor glycolysis and mitochondrial function directly and explicitly, we have explored several fluorescent probes for metabolic measurement. The 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]-2-deoxy-D-glucose (2-NBDG) has been used in tumors to report glucose uptake,32,33 similar to the clinically available FDG-PET. Tetramethylrhodamine ethyl ester (TMRE) has been utilized to quantify mitochondrial membrane potential (MMP) to study OXPHOS.34,35 Along with novel imaging processing methods, a standard fluorescence microscope may have the potential to provide flow cytometry–like single-cell level analysis of cell metabolism but in a more efficient, cost-effective, and non-destructive manner.

In this study, we demonstrate the use of two probes: 2-NBDG and TMRE with both a standard fluorescence microscope (ZOE™, Bio-Rad) and a flow cytometry device (BD Symphony A3), to report the changes in metabolism between radioresistant (rSCC-61) and radiosensitive (SCC-61) HNSCC cells under radiation stresses with or without HIF-1α inhibition.36 Our studies found that rSCC-61 cells have increased baseline glucose uptake and decreased baseline MMP compared with SCC-61 cells. Radiation treatment further enhanced glucose uptake for rSCC-61 cells but not for SCC-61 cells. Moreover, radiation treatment increased MMP for SCC-61 cells but not for rSCC-61 cells. We also observed that radiation induced overexpression of HIF-1α in rSCC-61 cells but not in SCC-61 cells. HIF-1α inhibition in the two cell lines during the radiation treatment modulates their metabolic changes and radio-sensitizes the rSCC-61 cells, which suggested that the radiation induced HIF-1α expression and the following metabolic changes may contribute to the radioresistance development in HNSCC. Through these studies, we also demonstrated that a standard fluorescence microscope along with proper imaging processing software (CellProfiler) can provide flow cytometry–like single-cell level analysis of HIF-1α-mediated metabolic changes in the radioresistant HNSCC cells, but in a more efficient, cost-effective, and non-destructive manner. This study reports the functional flexibility of our optical approach to report the key metabolic changes of radioresistant and radiosensitive HNSCC under therapeutic stress, thereby revealing the role of metabolism reprogramming in the development of resistance to cancer therapeutics. Using this optical metabolic imaging approach, we aim to create a non-destructive platform that can ultimately be translated to in vivo imaging using orthotropic HNSCC tumor models37 to study the radioresistance development mechanisms at single-cell level resolution.

2.

Materials and Methods

2.1.

Cell Culture and Radiation Treatment with HIF-1α Inhibition

A matched model of radioresistance for HNSCC including SCC-61 and rSCC-61 cell lines was used in this study.12 The SCC-61 cell line that was previously derived from a HNSCC tumor located at the base of the tongue is radiation sensitive, whereas the rSCC-61 cell line generated from SCC-61 cells is radioresistant.11 The generation of the SCC-61/ rSCC-61 matched model was described earlier in detail.11 Both SCC-61 and rSCC-61 cells used in this study were cultured in the DMEM/F12 medium (Gibco, Waltham, Massachusetts, United States) supplemented with 10% fetal bovine serum (FBS) (Gibco) and 1X penicillin streptomycin (Gibco) at 37°C and 5% CO2. rSCC-61 cells were maintained with further weekly radiation of 2 Gy. Cell medium was replaced every 2 days with fresh medium. Where applicable, an X-RAD 225XL orthovoltage preclinical irradiator was used for radiation treatment. In general, 75% of confluent cells were exposed to a total of 4 Gy radiation for all radiation treatments. In the HIF-1α inhibition treatment, 50-μM YC-1 in DMSO (Abcam) or equal volume of DMSO as control was added to the medium before the radiation treatment based on former publications.31,38 After radiation and/or HIF-1α inhibition treatments, all cells were then returned to the CO2 incubator for 24 h prior to any metabolic imaging, survival test, or western blotting experiments. A single dose of 4 Gy was used in our study as the former dose–response study showed that the SCC-61 and rSCC-61 cells would have significantly different survival rates until a single treatment dose reached up to 4 Gy.11 A former time-response study showed that HIF-1α protein expression induced by a 3 Gy radiation treatment will be significantly enhanced and stabilized in radioresistant HNSCC cells at 8 and 24 h post-radiation treatment.18 On the other hand, another time-response study showed that the radiation-induced metabolic changes in cells would occur around 24 h post-radiation treatment.39 Taken together, we selected the time point of 24 h post-radiation treatment as the time point for our metabolic imaging, survival test, or western blotting experiments.

2.2.

Survival Test and Western Blotting Experiment

To evaluate the cell survival rate after any treatment, a crystal violet colorimetric test40 was performed to test the cell viability. In brief, cells from each experimental group were first fixed with 4% paraformaldehyde for 15 min at room temperature, and then, the fixed cells were washed two times with PBS (1X, Gibco) and exposed to 0.5% crystal violet solution. After 15 min of exposure, excess crystal violet was removed, and cells were washed three times with distilled water. The content of the wells was dissolved in DMSO and read at 565 nm using the Synergy HTX Multi-Mode Microscope Reader (BioTek, Winooski, Vermont, United States). Protein extraction and western blot analysis were performed to characterize HIF-1α expression. The cells were lysed in cold RIPA buffer (Sigma-Aldrich, St. Louis, Missouri, United States) supplemented with a 1X protease inhibitor kit (Roche, Basel, Switzerland) and 1X phosphatase inhibitor (Roche). After extraction, the protein concentration of each sample was determined using the PierceTM BCA Protein Assay kit (Thermofisher Scientific, Waltham, Massachusetts, United States). All protein samples were heated in boiling water after adding 1X Laemmli Sample Buffer (Bio-Rad, Hercules, California, United States) and 5% 2-Mercaptoethanol (Bio-Rad). 30-μg protein was resolved on an 8% to 15% (depending on the molecular weight of the protein of interest) SDS-polyacrylamide gel (PAGE) (Bio-Rad) and transferred to nitrocellulose (NC) membrane (Bio-Rad). Membranes were incubated in a blocking solution (5% non-fat dried milk dissolved in Tris-buffered saline). HIF-1α (1:1000 dilution, 36169, Cell Signaling Technology, Danvers, Massachusetts, United States) and beta-actin (1:2000 dilution, MA1-140, Invitrogen, Waltham, Massachusetts, United States) were diluted in blocking solutions. The membranes were washed and incubated with appropriate secondary antibodies and assessed by SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Thermofisher Scientific™) on Tanon 5200 Chemiluminescent Imaging System. All bands were normalized to the 45-kDa beta-actin band as a loading control.

2.3.

Optical Metabolic Imaging and Flow Cytometry Experiments

Optical metabolic imaging and flow cytometry of glucose uptake and MMP on the HNSCC cells were conducted using our previously validated fluorescence probes including 2-NBDG and TMRE. The fluorescent probes including 2-NBDG and TMRE were chosen for our imaging due to their translatability to in vivo experiments as published by us before.33,35 The 2-NBDG (Biogems, Westlake Village, California, United States) is an optical analog glucose that enters the cell via glucose transporters,4143 which measures glucose uptake analogous to widely accepted PET imaging.44,45 TMRE (Biotium, Fremont, California, United States) is a cationic dye that accumulates in the mitochondrial inner membrane as a function of MMP,46 which has been utilized extensively to study mitochondrial metabolic capability.33,35 During labeling, 2-NBDG and TMRE were diluted to final concentrations of 200  μmol/L and 50  nmol/L, respectively, in glucose-free DMEM (Gibco) with 10% dialyzed FBS (Gibco) and 1X Penicillin Streptomycin (Gibco). Concentrations were chosen based on standard manufacturer protocols to not affect the metabolism of cells and minimize each probe’s toxicity.42 Before any optical imaging or flow cytometry, the regular media was removed, and the cell was washed with PBS and incubated at 37, 5% CO2 for 30 min with either 200  μmol/L 2-NBDG or 50  nmol/L TMRE dissolved in glucose-free cell media. Cells were only stained with one of two probes to prevent metabolite competition and minimize optical and biological cross-talk between metabolic probes.42 All optical metabolic images were collected using a fluorescence microscope (ZOE™ Fluorescent Cell Imager, Bio-Rad). The green channel that has an excitation peak of 488 nm (17-nm bandwidth) and emission peak of 517 nm (23-nm bandwidth) was used for 2-NBDG uptake imaging, and the red channel with an excitation peak of 556 nm (20-nm bandwidth) and emission peak of 615 nm (61-nm bandwidth) was used for TMRE uptake imaging. The gain level, contrast level, and LED intensity of the microscope were optimized and then fixed for either 2-NBDG or TMRE imaging at all different time points to ensure the images with sufficient optical signals could be obtained and compared. The optical metabolic microscope has a 20× objective, a field of view of 0.70  mm2, with a resolution of 1  μm (2.7  pixels). The cells were cultured in a 12-well plate, and the cell density was 0.2 to 0.5×106  cells per well. These factors can significantly influence the performance of cell segmentation algorithms. Higher-quality imaging of cells with appropriate cell density can improve segmentation accuracy; improper conditions may lead to difficulties in distinguishing individual cells. In flow cytometry experiments, the cells were detached by 0.25% Trypsin (Gibco) after probe labeling and then pelleted at 1500 rpm for 5 min by a centrifuge (Fisher Scientific, Hampton, New Hampshire, United States). To maximize the cell viability and signal intensity, the cell pellets were directly resuspended with cold PBS and then analyzed using a flow cytometer (BD Symphony A3, BD Bioscience, Franklin Lakes, New Jersey, United States) with optimized configurations. Specifically, 2-NBDG was excited by a blue laser (488 nm) and detected with a filter set of 505-nm long pass dichroic filter and 515  nm±20  nm band pass emission filter, whereas TMRE was excited by a yellow-green laser (561 nm) and detected with a filter set of 570-nm long pass dichroic filter and 586  nm±15  nm bandpass emission filter. DAPI was excited by an ultraviolet laser (355 nm) and detected with a filter set of 410-nm long pass dichroic filter and 450  nm±50  nm bandpass emission filter to exclude dead cells. Flow cytometry data were collected and processed using BD FACSDIVATM and FlowJo software provided by the Flow Cytometry and Immune Monitoring Core Facility at the University of Kentucky.

2.4.

Optical Metabolic Imaging Data Analysis

To provide flow cytometry–like single-cell data analysis for optical metabolic imaging, all 2-NBDG and TMRE images at different experimental groups were processed, as illustrated in Fig. 1. To be brief, we utilized publicly available open-source software CellProfiler (version 4.2.6) to analyze fluorescent microscopy images measured on in vitro HNSCC cells. To identify individual cells from fluorescence images, we initially utilized the automated ‘primary object identification’ toolbox of the cell profiler, followed by manual object identification to improve cell identification accuracy. In the primary object identification toolbox, the diameter range of the identified cells was chosen to be between 50 and 100-pixel units (18 to 38  μm) using global minimum cross-entropy thresholding as it provided the highest accuracy. Cells outside of this range were manually identified using the manual object identification toolbox in CellProfiler. The intensity distribution histograms were generated by calculating the mean intensity values of individual identified cells from each image, considering the whole cell region without isolating the nucleus. The mean intensity values of each identified cell from all the images were then saved in an Excel sheet. By default, CellProfiler outputs the individual cell intensity values within the range of 0 to 1 for the identified cells (single-cell level analysis), we have then converted each cell intensity distribution value to 8-bit by multiplying the intensity values with 255. To speed up the processes, random 1000×750-pixel windows were cropped from each image for individual cell intensity analysis. For a comparison among the different experimental groups, 10 to 20 random images from each experimental group were selected and processed using the aforementioned method. Once all images were processed by the Cell Profiler, MATLAB (MathWorks, R2023a, Natick, Massachusetts, United States) software was used to provide flow cytometry–like data analysis. Specifically, the kernel smoothing function (ksdensity) and kernel distribution fit were used to generate the frequency plot and then to provide the histogram analysis for the fluorescence intensity. The mean intensities for each experimental group were also created and compared using Student’s t-test (between two groups) or analysis of variance (ANOVA) test (among three groups). pvalues<0.05 were considered statistically significant. MATLAB Statistics Toolbox was used for all statistical tests.

Fig. 1

Optical microscope along with proper image processing software can characterize metabolism per cell.

BIOS_2_1_012702_f001.png

To quantitatively compare the optical imaging data with flow cytometry results, the histograms generated from optical images and flow cytometry data were evaluated and compared with several metrics. Specifically, the relative changes of histogram peak locations, histogram FWHM (full width at half maximum), mean intensities, and median intensities among different experimental groups were calculated and compared. The p-values were also summarized and compared between the optical imaging data and flow cytometry results.

3.

Results

3.1.

Optical Metabolic Imaging and Flow Cytometry Capture Increased Glucose Uptake and Decreased MMP in rSCC-61 Cells Versus SCC-61 Cells

To compare the baseline metabolic phenotypes of rSCC-61 cells and SCC-61 cells, both optical metabolic imaging and flow cytometry were conducted on these two HNSCC cell lines. Figure 2 shows the glucose uptake (2-NBDG uptake) and MMP (TMRE uptake) of radioresistant (rSCC-61) and radiosensitive (SCC-61) HNSCC cells characterized by both optical metabolic imaging and flow cytometry. Figure 2(a) shows the survival rates of the two HNSCC cell lines under 4 Gy of radiation treatment. The survival rate data confirmed that rSCC-61 cells are more radioresistant compared with SCC-61 cells. The representative fluorescence imaging shows that rSCC-61 cells had increased glucose uptake but decreased MMP compared with SCC-61 cells [Fig. 2(b)]. Figure 2(c)1 shows the statistical analysis using the optical imaging data collected on different batches of experiments and illustrates that the glucose uptake level in rSCC-61 cells was statistically higher compared with SCC-61 cells (p<0.0001). By contrast, Fig. 2(d)1 shows that the MMP of rSCC-61 cells was lower than that in SCC-61 cells but not statistically significant. Figures 2(c)2 and 2(d)2 show the corresponding flow cytometry data measured on the same batch of HNSCC cells. The flow cytometry results also show that rSCC-61 cells had significantly higher glucose uptake [Fig. 2(c)2] but slightly lower MMP [Fig. 2(d)2] compared with SCC-61 cells. Both optical imaging and flow cytometry results are consistent with the previous analysis of energy metabolism in these two HNSCC cell lines using Seahorse Assay.12 Figure 2(e) shows the histogram characteristics changes for the histograms generated from optical images and flow cytometry data. Generally, the histogram characteristics changes between the optical data and flow cytometry data showed that the metabolic changes between the two HNSCC cell lines are consistent although it appears that the histograms from flow cytometry had larger changes among different experimental groups compared with that from optical imaging, which suggested that flow cytometry has a higher sensitivity compared with optical microscopy in our study.

Fig. 2

rSCC-61 cells had increased glucose uptake and decreased MMP compared with SCC-61 cells. (a) Survival fractions of rSCC-61 and SCC-61 cells under 4 Gy of radiation treatment. (b) Representative fluorescent imaging shows that the rSCC-61 cells had increased 2-NBDG uptake and decreased TMRE uptake compared with SCC-61 cells. Statistical analysis of HNSCC baseline glucose uptake (c)1 and MMP (d)1 based on metabolic images. Statistical analysis of HNSCC baseline glucose uptake (c)2 and MMP (d)2 based on flow cytometry data. (e) Histogram characteristic changes between the histograms generated from optical images and flow cytometry data. Peak refers to the x-axis value (intensity) indicated by the dashed lines (probability density peak or count peak) in panels (c)1, (c)2, (d)1, and (d)2. Median refers to the median intensity of all cell populations, whereas mean refers to the mean intensity of all cell populations. The sample size for optical imaging was 10 to 20 (images) per group. The sample size for flow cytometry was three to six samples per group. The sample size for the survival test was 12 per group. Student’s t-test was used for statistical analysis.

BIOS_2_1_012702_f002.png

3.2.

rSCC-61 Cells Had Different Metabolic Changes Under Radiation Stress Compared with SCC-61 Cells, Along with Enhanced HIF-1α Expression in rSCC-61 Cells

To investigate the metabolic responses under radiation stresses, both SCC61 and rSCC-61 cells were irradiated with a 4 Gy radiation and then characterized by both optical metabolic imaging and flow cytometry. Figure 3 shows glucose uptake and MMP of SCC-61 cells and rSCC-61 cells with or without radiation treatment. The glucose uptake and MMP were quantified based on both optical fluorescence images and flow cytometry data. The histograms in Figs. 3(a)1 and 3(b)1 illustrate that radiation stress enhanced glucose uptake in rSCC-61 cells but not in SCC-61 cells. However, the average intensities in the bar figures showed that the radiation treatment upregulated the glucose uptake in both rSCC-61 and SCC-61 cells. Similarly, the histograms in Figs. 3(a)2 and 3(b)2 illustrate that the MMP was increased post-radiation treatment for SCC-61 cells but not for rSCC-61 cells, whereas the average TMRE uptake intensities showed that the radiation enhanced MMP for both cell lines. Figures 3(a)3 and 3(b)3 show the histogram characteristics changes for the histograms generated from optical images and flow cytometry data. Overall, the histogram characteristics changes between the optical data and flow cytometry data showed that the metabolic changes between the control group (0 Gy) and RT treatment group (4 Gy) are consistent for both SCC-61 cells [Fig. 3(a)3] and rSCC-61 cells [Fig. 3(b)3], whereas the histograms from flow cytometry had larger changes among different experimental groups compared with that from optical imaging, further suggesting that flow cytometry has a higher sensitivity compared with optical microscopy in our study. Figure 3 shows that the statistical analysis of the cell metabolism changes using metabolic images is always consistent with that using the flow cytometry data. To explore if the HIF-1α protein was involved in the metabolic changes under the radiation stress, western blotting experiments were also conducted on the same batch of HNSCC cell lines with or without radiation treatment. The western blotting results in Figs. 3(a)4 and 3(b)4 show that the HIF-1α expression was significantly enhanced in rSCC-61 cells (p<0.05) under radiation stress but not in SCC-61 cells, which suggests that HIF-1α may be associated with metabolic changes in the acquisition of radioresistance in HNSCC cells.

Fig. 3

rSCC-61 cells had different metabolic changes under radiation stress compared with SCC-61 cells, along with enhanced HIF-1α expression in rSCC-61 cells but not in SCC-61 cells. The 2-NBDG uptake and TMRE uptake changes were acquired from both optical metabolic images and the corresponding flow cytometry experiments. Statistical analysis of the 2-NBDG uptake changes post radiation treatment for SCC-61 cells (a)1 and rSCC-61 cells (b)1. Statistical analysis of the TMRE uptake changes post radiation treatment for SCC-61 cells (a)2 and rSCC-61 cells (b)2. (a)3 and (b)3 Histogram characteristic change between the histograms generated from optical images and flow cytometry data. Peak refers to the x-axis value (intensity) indicated by the dashed lines (probability density peak or count peak) in panels (a)1, (a)2, (b)1, and (b)2. Median refers to the median intensity of all cell populations, whereas mean refers to the mean intensity of all cell populations. Representative western blotting images and the statistical analysis of HIF-1α expression in SCC-61 cells (a)4 and rSCC-61 cells (b)4 under 4 Gy radiation treatment. The sample size for imaging was 10 to 20 (images) per group. The sample size for flow cytometry was 3 to 4 samples per group. The sample size for western blotting was four repeats (six data points in total) per group. Student’s t-test was used for statistical analysis.

BIOS_2_1_012702_f003.png

Fig. 4

Radiation induced overexpression of HIF-1α in rSCC-61 cells but not in SCC-61 cells, and the HIF-1α inhibition radio-sensitizes the rSCC-61 cells. Representative western blotting images and the statistical analysis of HIF-1α expression in SCC-61 cells (a) and rSCC-61 cells (b) under different treatments. The radiation treatment induced and enhanced HIF-1α in both rSCC-61 and SCC-61 cells but was only statistically significant for rSCC-61 cells. HIF-1α inhibition mitigates the HIF-1α inhibition expressions in both rSCC-61 and SCC-61 cells but was only statistically significant for rSCC-61 cells. The corresponding survival rates of SCC-61 cells (c) and rSCC-61 cells (d) under different treatments. All results were normalized to their corresponding control group. The sample size for all western blotting experiments was five repeats (seven data points in total) per group. The sample size for the survival test was 12 per group. Student’s t-test was used for statistical analysis.

BIOS_2_1_012702_f004.png

3.3.

Radiation Induces Significantly Increased HIF-1α Expression in rSCC-61 Cells and the HIF-1α Inhibition Radio-Sensitizes the rSCC-61 Cells

To further reveal the role of radiation-induced HIF-1α expression in the HNSCC radioresistance development, we further conducted western blotting experiments on rSCC-61 and SCC-61 cells with various treatments, including no treatment (control), 4 Gy radiation-only treatment, HIF-1α inhibition only, and 4 Gy radiation treatment plus HIF-1α inhibition. Figures 4(a) and 4(b) show the representative western blotting results of HIF-1α expressions in SCC-61 and rSCC-61 cells under the four different treatments, whereas Figs. 4(c) and 4(d) show their corresponding survival rates. Figure 4(a) shows that radiation treatment slightly upregulates the HIF-1α expression in SCC-61 cells, whereas the HIF-1α inhibition mitigates the HIF-1α expression. Figure 4(c) shows that the survival rates of SCC-61 did not change much when the HIF-1α inhibition was introduced in SCC-61 cells compared with those that only received radiation treatment. Figure 4(b) shows that radiation stress significantly upregulates the HIF-1α expression in rSCC-61 cells (p<0.005), whereas the HIF-1α inhibition significantly mitigates the HIF-1α expression (p<0.005). Figure 4(d) shows that the survival rates of rSCC-61 decrease significantly after the HIF-1α inhibition in the radiation treated rSCC-61 cells compared with the rSCC-61 cells that only received radiation treatment (p<0.005). Together, Fig. 4 suggests that HIF-1α may be a potential target to sensitize radiation response in HNSCC.

3.4.

HIF-1α Inhibition Modulates the Metabolic Changes Induced by Radiation Stress in Both rSCC-61 Cells and SCC-61 Cells

To further investigate if the radiation-induced HIF-1α overexpression in the radioresistant rSCC-61 cells was associated with metabolic changes post-radiation treatment, HIF-1α inhibition (YC-1) experiments during the radiation treatment on the two HNSCC cell lines were conducted. Figure 5 shows the metabolic changes of the two cell lines receiving different treatments including control and radiation with or without HIF-1α inhibition. Figure 5(a)1 presents typical 2-NBDG uptake images that reflect changes in the glucose uptake among three different experimental groups, whereas Fig. 5(b)1 shows the typical TMRE uptake images that reflect the MMP changes among the three experimental groups. Figures 5(a)2 and 5(a)3 report the corresponding statistical analysis on the 2-NBDG uptake among groups based on the optical metabolic images and flow cytometry data. Both optical imaging data and flow cytometry results show consistent trends for glucose uptake changes in the two cell lines under various treatments. Specifically, the radiation treatment enhanced glucose uptake in both SCC-61 and rSCC-61 cells compared with their control group, whereas the HIF-1α inhibition reversed the radiation-induced changes in SCC-61 cells but not in rSCC-61 cells. Rather, the HIF-1α inhibition further enhanced glucose uptake in rSCC-61 cells compared with the radiation-treated group. Figures 5(b)2 and 5(b)3 show the statistical analysis on the MMP among different groups based on both optical imaging data and flow cytometry data. Optical imaging data show that radiation treatment enhanced MMP in SCC-61 cells compared with their control group, whereas the HIF-1α inhibition along with radiation treatment further enhanced the MMP in SCC-61 cells compared with the radiation treatment group, as shown in Fig. 5(b)2. By contrast, the optical imaging data shows that radiation treatment did not upregulate MMP in rSCC-61 cells compared with their control group, whereas the HIF-1α inhibition plus radiation stress significantly enhanced the MMP in rSCC-61 cells compared with the radiation treatment group [Fig. 5(b)2]. The flow cytometry results of Fig. 5(b)3 show that the TMRE uptake changes in the two cell lines from different experimental groups are generally consistent with the optical imaging data. Figures 5(a)4 and 5(b)4 show the histogram characteristics changes for the histograms generated from optical images and flow cytometry data. Overall, the histogram characteristics changes between the optical data and flow cytometry data showed that the metabolic changes among the control group (0 Gy), RT treatment group (4 Gy), and RT treatment plus HIF-1α inhibition group (4 Gy+ YC-1) are generally consistent for both HNSCC cell lines.

Fig. 5

HIF-1α inhibition using YC-1 modulates the metabolic changes induced by radiation stress for rSCC-61 and SCC-61 cells. Representative 2-NBDG uptake imaging (a)1 and TMRE uptake imaging (b)1 at baseline, 4 Gy radiation treatment, and the combination of 4 Gy radiation treatment with HIF-1α inhibition for SCC-61 and rSCC-61 cells. Statistical analysis of 2-NBDG uptake changes among the control group, radiation treatment group, and radiation plus HIF-1α inhibition group using optical imaging data (a)2 and flow cytometry data (a)3. Statistical analysis of TMRE uptake changes among the control group, radiation treatment group, and radiation plus HIF-1α inhibition group using optical imaging data (b)2 and flow cytometry data (b)3. (a)4 and (b)4 Histogram characteristics changes between the histograms generated from optical images and flow cytometry experiments from optical images and flow cytometry experiments. Peak refers to the x-axis value (intensity) indicated by the dashed lines (probability density peak or count peak) in panels (a)2, (a)3, (b)2, and (b)3. Median refers to the median intensity of all cell populations, whereas mean refers to the mean intensity of all cell populations. The sample size for optical imaging was 10 to 20 images per group. The sample size for flow cytometry was three to six samples per group. * represents p<0.05, ** represents p<0.01, and *** represents p<0.001. The ANOVA test was used for statistical analysis. Both the metabolic imaging and flow cytometry experiments were repeated independently at least two times, which all yielded consistent trends for the metabolic changes under various treatments.

BIOS_2_1_012702_f005.png

4.

Discussion

Interest in tumor metabolism continues to grow in the field of cancer research as metabolic reprogramming of tumors has been recognized as one of the major cancer hallmarks.47 Despite a variety of metabolic tools that report on different endpoints to piece together a narrative on metabolic reprogramming of tumor cells that escape therapies, there still exists a significant unmet need for a translational technique with the ability to directly link in vitro and in vivo results for cancer research. For example, the most commonly used metabolic tool for in vitro cell studies is the Seahorse Assay, which perturbs cell metabolism with special chemicals, following which the oxygen consumption and extracellular acidification rates are measured.4857 Metabolomics can simultaneously screen a large number of metabolites and map metabolic networks from in vitro cells and ex vivo tissues to even human subjects but requires sample extraction.23,58 Because of their destructive nature, it is impractical to conduct multiple metabolic measurements on the same population of cancer cells or in vivo using either Seahorse Assay or metabolomics. FDG-PET24 is commonly used for in vivo measurement of glucose uptake (FDG-PET), and magnetic resonance spectral imaging25,26 has been explored to report mitochondrial metabolism and glycolysis by monitoring P31 or C13 nuclei in metabolites in vivo, but they have either low resolution or low sensitivity. Optical metabolic imaging could potentially fill this critical gap due to its capability to directly measure the metabolic parameters from in vitro cells, ex vivo tissues, and in vivo tumor models at subcellular level resolution in a non-destructive manner.

Autofluorescence of reduced nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FAD) have been explored as a non-destructive approach to report the reduction-oxidation (redox) state of cells29,30,59,60 by looking at the ratio of the two (FAD/NADH) and then provide an indirect measure of the balance between glycolysis and OXPHOS. FAD- and NADH-based label-free autofluorescence techniques have been explored extensively by others for cancer applications.29,30,59,60 However, autofluorescence imaging typically requires expensive two-photon microscopy to avoid photon damage on cells,61 and autofluorescence signal from cells is relatively weak. To address some of the limitations associated with the autofluorescence technique, we have exploited metabolic probes-based approaches to measure cell glucose uptake and mitochondrial function directly and explicitly. Specifically, we utilized fluorescence probes such as 2-NBDG and TMRE to measure glucose uptake and mitochondrial function of cells with the use of commonly available single-photon microscopes at a lower cost. These labeling-based techniques are relatively easy to implement and are compatible with live-cell imaging, providing distinct functional information. However, they involve the use of external probes that require a labeling procedure, and their signals are not endogenous, potentially limiting their ability to reflect the natural metabolic state.

To demonstrate the capability of optical techniques, here, we utilized our optical imaging methods along with the flow cytometry technique to capture the metabolic changes of radioresistant and radiosensitive HNSCC cells in vitro. For the first time, we demonstrated that a standard fluorescence microscope along with proper imaging processing software (CellProfiler) can provide flow cytometry–like single-cell level analysis of HIF-1α-mediated metabolic changes in the radioresistant HNSCC but in a more efficient, cost-effective, and non-destructive manner. It should be noted that the optical imaging method is theoretically non-destructive under conditions where photon damage is negligible, minimizing any potential perturbations to cell metabolism that could arise from destructive techniques. Moreover, due to its non-destructive nature, the optical technology is suitable for repeatable and longitudinal studies on the same population of in vitro cells or patient-derived organoids.62 To demonstrate the proof-of-concept, we utilized CellProfiler for single-cell analysis from fluorescence images in this study. CellProfiler has been widely used in the research community as open-source software with a user-friendly interface and a minimal requirement for programming knowledge. CellProfiler offers flexibility to choose from a variety of cell segmentation algorithms, such as threshold algorithms, morphology algorithms, and watershed algorithms, or a combination of these, depending on user needs. It also allows for the customization of parameters to improve segmentation accuracy. However, it should be noted that other open-source software tools, such as Ilastik and Icy, and deep learning-based platforms such as CellPose, DeepCell, and StarDist, can also be employed for single-cell analysis. The performance of these tools may vary depending on the specific dataset, and each has its advantages for different types of analysis.

The classical notion of ineffective radiotherapy is primarily attributed to hypoxia.63,64 However, increasing evidence shows that metabolic reprogramming may also be responsible for the development of radioresistance in cancers.5 HIF-1α plays an essential role in tumor cell survival, growth, and metastasis.14 On the one hand, HIF-1α regulates the expression of genes involved in OXPHOS and energy production and promotes the Warburg effect,14 enhanced glycolysis, and lactate secretion in the presence of oxygen in many proliferating tumor cells. On the other hand, HIF-1α also enhances angiogenesis to facilitate tumor growth.20 HIF-1α can be stabilized by hypoxia and also by ROS1517 produced from RT.14 Previous studies reported that the RT induced significantly increased HIF-1α in many types of human cancer cells, including breast cancer,65 lung cancer,9 and HNSCC.18 Former in vitro studies of the matched model of radioresistance for HNSCC used in this study (rSCC-61 and SCC-61) showed that radioresistant HNSCC cells have enhanced glycolysis and decreased OXPHOS compared with their parental radiosensitive cells.11,12 Based on the above findings, it is reasonable to hypothesize that the radiation induced HIF-1α overexpression, and the following metabolic changes might be responsible for the development of radioresistance in HNSCC.

In this study, we utilized optical metabolic imaging and a matched model of radiation resistance for HNSCC (SCC-61 and rSCC-61) to understand the role of radiation-induced HIF-1α and the following metabolic changes between the radiosensitive and radioresistant HNSCC cells. Our in vitro cell studies found that the radioresistant HNSCC cells have increased baseline glucose uptake and decreased MMP compared with the radiosensitive cells. The optical imaging data are consistent with the flow cytometry data and the published Seahorse Assay results.11,12 We also found that radiosensitive HNSCC cells (SCC-61) have increased glucose uptake and increased MMP under radiation stress. It is widely accepted that mitochondria produce more ROS at high MMP;66 thus, the increased MMP in SCC-61 cells suggested a significantly increased ROS induced by RT as reported previously,11 which can lead to increased oxidative stress. By contrast, radiation treatment of the radioresistant HNSCC cells (rSCC-61) further enhanced glucose uptake while unchanged MMP along with significantly increased HIF-1α expressions. The unchanged MMP in rSCC-61 cells suggested that radiation treatment did not induce excess ROS production in the rSCC-61 cells, in agreement with the published data.11 RT-enhanced glucose uptake in rSCC-61 cells can compensate for the loss of energy production due to unchanged MMP to sustain cell growth under RT. Earlier studies reported that radiation induces HIF-1α overexpression to promote glycolysis in radioresistant lung cancer cells that may help them escape from the RT.9,31 These RT-induced metabolic events in rSCC-61 cells may well be caused by the observed HIF-1α overexpression as HIF-1α inhibition modulated the metabolic changes induced by radiation. Together, these data suggest that the radiation-induced HIF-1α overexpression and associated metabolic changes may contribute to the radiation resistance in rSCC-61 cells. The HIF-1α inhibition significantly promoted the MMP in rSCC-61 cells along with the survival tests shown in Fig. 4(d), which confirmed that the HIF-1α inhibitor (YC-1) could potentially radio-sensitize the HNSCC cells, a finding that supports future therapeutic investigations.31 It was interesting to observe that the HIF-1α inhibition further enhanced glucose uptake in rSCC-61 cells in our study. Though we are not fully clear on the mechanism behind this phenomenon yet, these observations may be supported by compensatory or adaptive metabolic mechanisms that reported in other human cancer cells.67,68 For example, a former study reported that inhibiting HIF-1α might activate alternative glycolytic drivers such as c-Myc or AMPK, which can enhance the expression of glucose transporters or enzymes to promote glucose uptake even in the absence of HIF-1α activity.67 Another study showed that radioresistant cells can rely on HIF-1-independent adaptations for survival, such as activating pathways involving PI3K-Akt, mTOR, or others, which would compensate for the loss of HIF-1α activity and promote glucose metabolism.68 Nevertheless, we will continue to explore the potential mechanism behind this interesting phenomenon in our future study.

The work described here reported an effective imaging strategy for the direct study of two key metabolic endpoints (glucose uptake and MMP) related to tumor metabolism and has the potential to be expanded to image other endpoints such as fatty acid uptake69 in our future work. This study demonstrated that a standard fluorescence microscope along with proper image processing software can provide flow cytometry–like analysis of cell metabolism at the single-cell level. It should be noted that the inherent differences between these two techniques pose some challenges for direct quantitative comparisons. However, we still provided some comparisons of the results obtained by these two technologies using multiple matrices, including standard deviation, p-value, histogram peak location changes, mean and median intensity changes, and histogram FWHM changes between experimental groups. These quantitative matrices reveal some variations in the readouts between the two techniques. Specifically, we noticed that the standard deviations for metabolic parameters from fluorescence images were slightly larger than those from flow cytometry, and the relative changes of histogram peak locations, mean and median intensities, and FWHM from fluorescence images were smaller than those from flow cytometry data. However, the p-values among various experimental groups generated from fluorescence images are generally comparable with those from flow cytometry data. These discrepancies in the readouts can be caused by several factors: (1) In flow cytometry experiments, cells were suspended for individual cell intensity readout, whereas microscopy images were taken while retaining the cells in their native conditions; (2) Flow cytometry results were generated using significantly larger cell populations (20,000) compared with microscopy imaging (1000); and (3) The microscope results were highly dependent on the image quality, microscope configuration, and the choice of cell segmentation software; (4) The flow cytometry is more sensitive compared with the standard fluorescence microscope used in our study. Nevertheless, our results showed that microscopy provided similar trends between experimental groups compared with flow cytometry, but in a more efficient, cost-effective, and non-destructive manner.

The selection of proper techniques for measuring individual cellular metabolism depends on the specific needs of the experiment, as each method has its own advantages and limitations. Flow cytometry measures individual cell intensity in suspension, passing cells sequentially through a laser beam, making it ideal for high-throughput population-based analysis. Optical microscopies capture cellular fluorescence intensities within their native microenvironment, offering the unique advantage of spatial and contextual information that flow cytometry cannot provide. A detailed comparison of these two techniques for cellular metabolism analysis is shown in Table 1 to better understand the trade-offs and make informed decisions.

Table 1

Comparison of flow cytometry and microscopy for cellular metabolism analysis.

MetricsFlow cytometryMicroscopy
Sample typesCellsCells, ex vivo tissue
Special preparationDetachment, suspensionNone
SensitivityVery highHigh (nM for TMRE, μM for 2-NBDG)
Signal-to-background ratio10 for 2-NBDG, 1000 for TMRE20 for 2-NBDG, 80 for TMRE
ThroughputMillion cells per testThousands of cells per test
QuantificationStatisticalParametric, can be statistical
MorphologicalNoYes
ComplexityModerateLow
Cost in instrumentationHalf million USDLess 10k USD

It should be noted that tumor cell metabolism is highly dependent on its micro-environment. Although in vitro cultures remain critically important in cancer research, they cannot fully replicate the tumor microenvironment, which limits their applications to translation studies. The next step to understand the role of metabolic reprogramming in radioresistance development for HNSCC should be to image in vivo the landscape of 2-NBDG and TMRE changes in tumor tissues with relevant vasculature and stromal environment using an orthotropic HNSCC tumor model.37

5.

Conclusion

We demonstrated optical imaging strategies to characterize the metabolic changes of radioresistant and radiosensitive HNSCC cells at a single-cell level under therapeutic stresses. We found that radioresistant and radiosensitive HNSCC cell lines had significantly different metabolic changes in response to radiation stress. The radiation stress also promoted HIF-1α expression in the radioresistant cells. We also found that HIF-1α inhibition during the radiation treatment modulated the metabolic changes induced by radiation stress and radio-sensitizes the radioresistant HNSCC cells, which suggested that the radiation induced HIF-1α overexpression, and the associated metabolic changes may be responsible for the development of the radiation resistance phenotype. Our study demonstrated that optical imaging techniques could be a powerful tool for studying the role of metabolic reprogramming in the development of resistance to cancer therapeutics at the single-cell level in a more efficient, cost-effective, and non-destructive manner.

Disclosures

The authors have no relevant financial interests in this paper and no potential conflicts of interest to disclose.

Code and Data Availability

A detailed description of the entire CellProfiler pipeline for microscopy imaging processing is provided in the Supplementary Material. Data and code to generate figures are available: https://github.com/cgzhu123/HNSCC-metabolic-imaging.

Acknowledgments

This work was supported by generous funding from the National Institute of Dental and Craniofacial Research (NIDCR) and the National Institute of General Medical Sciences (NIGMS), branches of the U.S. National Institutes of Health (Grant No. R01DE031998). The authors thank the University of Kentucky Flow Cytometry and Immune Monitoring Shared Resource Facility (Grant No. P30CA177558), for the use of its facilities and services. The funders had no role in study design, data collection or analysis, decision to publish, or preparation of the paper.

References

1. 

J. Ferlay et al., “Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008,” Int. J. Cancer, 127 (12), 2893 –2917 https://doi.org/10.1002/ijc.25516 IJCNAW 1097-0215 (2010). Google Scholar

3. 

S. Marur and A. A. Forastiere, “Head and neck cancer: changing epidemiology, diagnosis, and treatment,” Mayo Clin. Proc., 83 (4), 489 –501 https://doi.org/10.4065/83.4.489 MACPAJ 0025-6196 (2008). Google Scholar

4. 

K. D. Miller et al., “Cancer treatment and survivorship statistics, 2022,” CA Cancer J. Clin., 72 (5), 409 –436 https://doi.org/10.3322/caac.21731 CAMCAM 0007-9235 (2022). Google Scholar

5. 

L. Tang et al., “Role of metabolism in cancer cell radioresistance and radiosensitization methods,” J. Exp. Clin. Cancer Res., 37 (1), 87 https://doi.org/10.1186/s13046-018-0758-7 (2018). Google Scholar

6. 

C. Jose, N. Bellance and R. Rossignol, “Choosing between glycolysis and oxidative phosphorylation: a tumor’s dilemma?,” BBA-Bioenergetics, 1807 (6), 552 –561 https://doi.org/10.1016/j.bbabio.2010.10.012 (2011). Google Scholar

7. 

T. Epstein et al., “Separation of metabolic supply and demand: aerobic glycolysis as a normal physiological response to fluctuating energetic demands in the membrane,” Cancer Metab., 2 7 https://doi.org/10.1186/2049-3002-2-7 (2014). Google Scholar

8. 

F. Zhao et al., “Inhibition of Glut1 by WZB117 sensitizes radioresistant breast cancer cells to irradiation,” Cancer Chemother. Pharm., 77 (5), 963 –972 https://doi.org/10.1007/s00280-016-3007-9 CCPHDZ 0344-5704 (2016). Google Scholar

9. 

K. Alhallak et al., “Optical imaging of radiation-induced metabolic changes in radiation-sensitive and resistant cancer cells,” J. Biomed. Opt., 22 (6), 060502 https://doi.org/10.1117/1.JBO.22.6.060502 JBOPFO 1083-3668 (2017). Google Scholar

10. 

W. W. Kam and R. B. Banati, “Effects of ionizing radiation on mitochondria,” Free Radic. Biol. Med., 65 607 –619 https://doi.org/10.1016/j.freeradbiomed.2013.07.024 FRBMEH 0891-5849 (2013). Google Scholar

11. 

N. Bansal et al., “Broad phenotypic changes associated with gain of radiation resistance in head and neck squamous cell cancer,” Antioxid. Redox. Signaling, 21 (2), 221 –236 https://doi.org/10.1089/ars.2013.5690 (2014). Google Scholar

12. 

J. Mims et al., “Energy metabolism in a matched model of radiation resistance for head and neck squamous cell cancer,” Radiat. Res., 183 (3), 291 –304 https://doi.org/10.1667/RR13828.1 RAREAE 0033-7587 (2015). Google Scholar

13. 

J. C. Fleming et al., “HPV, tumour metabolism and novel target identification in head and neck squamous cell carcinoma,” Br. J. Cancer, 120 (3), 356 –367 https://doi.org/10.1038/s41416-018-0364-7 BJCAAI 0007-0920 (2019). Google Scholar

14. 

C. Wigerup, S. Pahlman and D. Bexell, “Therapeutic targeting of hypoxia and hypoxia-inducible factors in cancer,” Pharmacol. Ther., 164 152 –169 https://doi.org/10.1016/j.pharmthera.2016.04.009 (2016). Google Scholar

15. 

J. K. Brunelle et al., “Oxygen sensing requires mitochondrial ROS but not oxidative phosphorylation,” Cell Metab., 1 (6), 409 –414 https://doi.org/10.1016/j.cmet.2005.05.002 1550-4131 (2005). Google Scholar

16. 

R. D. Guzy et al., “Mitochondrial complex III is required for hypoxia-induced ROS production and cellular oxygen sensing,” Cell Metab., 1 (6), 401 –408 https://doi.org/10.1016/j.cmet.2005.05.001 1550-4131 (2005). Google Scholar

17. 

K. D. Mansfield et al., “Mitochondrial dysfunction resulting from loss of cytochrome c impairs cellular oxygen sensing and hypoxic HIF-alpha activation,” Cell Metab., 1 (6), 393 –399 https://doi.org/10.1016/j.cmet.2005.05.003 1550-4131 (2005). Google Scholar

18. 

H. Lu et al., “The anti-EGFR antibody cetuximab sensitizes human head and neck squamous cell carcinoma cells to radiation in part through inhibiting radiation-induced upregulation of HIF-1alpha,” Cancer Lett., 322 (1), 78 –85 https://doi.org/10.1016/j.canlet.2012.02.012 CALEDQ 0304-3835 (2012). Google Scholar

19. 

A. J. Majmundar, W. J. Wong and M. C. Simon, “Hypoxia-inducible factors and the response to hypoxic stress,” Mol. Cell, 40 (2), 294 –309 https://doi.org/10.1016/j.molcel.2010.09.022 MOCEFL 1097-2765 (2010). Google Scholar

20. 

T. Hashimoto and F. Shibasaki, “Hypoxia-inducible factor as an angiogenic master switch,” Front. Pediatr., 3 33 https://doi.org/10.3389/fped.2015.00033 (2015). Google Scholar

21. 

T. W. Meijer et al., “Targeting hypoxia, HIF-1, and tumor glucose metabolism to improve radiotherapy efficacy,” Clin. Cancer Res., 18 (20), 5585 –5594 https://doi.org/10.1158/1078-0432.CCR-12-0858 (2012). Google Scholar

22. 

D. A. Ferrick, A. Neilson and C. Beeson, “Advances in measuring cellular bioenergetics using extracellular flux,” Drug Discov. Today, 13 (5–6), 268 –274 https://doi.org/10.1016/j.drudis.2007.12.008 DDTOFS 1359-6446 (2008). Google Scholar

23. 

E. J. Want et al., “Global metabolic profiling of animal and human tissues via UPLC-MS,” Nat. Protoc., 8 (1), 17 –32 https://doi.org/10.1038/nprot.2012.135 1754-2189 (2013). Google Scholar

24. 

G. Kelloff et al., “Progress and promise of FDG-PET imaging for cancer patient management and oncologic drug development,” Clin. Cancer Res., 11 (8), 2785 –2808 https://doi.org/10.1158/1078-0432.CCR-04-2626 (2005). Google Scholar

25. 

K. Glunde and Z. M. Bhujwalla, “Metabolic tumor imaging using magnetic resonance spectroscopy,” Semin. Oncol., 38 (1), 26 –41 https://doi.org/10.1053/j.seminoncol.2010.11.001 (2011). Google Scholar

26. 

R. J. Gillies and D. L. Morse, “In vivo magnetic resonance spectroscopy in cancer,” Annu. Rev. Biomed. Eng., 7 287 –326 https://doi.org/10.1146/annurev.bioeng.7.060804.100411 ARBEF7 1523-9829 (2005). Google Scholar

27. 

R. J. Arguello et al., “SCENITH: a flow cytometry-based method to functionally profile energy metabolism with single-cell resolution,” Cell Metab., 32 (6), 1063 –1075 e1067 1550-4131 (2020). Google Scholar

28. 

C. Zhu et al., “Early prediction of skin viability using visible diffuse reflectance spectroscopy and autofluorescence spectroscopy,” Plast. Reconstr. Surg., 134 (2), 240e –247e https://doi.org/10.1097/PRS.0000000000000399 (2014). Google Scholar

29. 

J. Hou et al., “Correlating two-photon excited fluorescence imaging of breast cancer cellular redox state with seahorse flux analysis of normalized cellular oxygen consumption,” J. Biomed. Opt., 21 (6), 060503 https://doi.org/10.1117/1.JBO.21.6.060503 JBOPFO 1083-3668 (2016). Google Scholar

30. 

A. T. Shah et al., “Optical metabolic imaging of treatment response in human head and neck squamous cell carcinoma,” Plos One, 9 (3), POLNCL 1932-6203 (2014). Google Scholar

31. 

D. E. Lee et al., “A radiosensitizing inhibitor of HIF-1 alters the optical redox state of human lung cancer cells in vitro,” Sci. Rep., 8 8815 https://doi.org/10.1038/s41598-018-27262-y SRCEC3 2045-2322 (2018). Google Scholar

32. 

K. Yamada et al., “A real-time method of imaging glucose uptake in single, living mammalian cells,” Nat. Protoc., 2 (3), 753 –762 https://doi.org/10.1038/nprot.2007.76 1754-2189 (2007). Google Scholar

33. 

C. Zhu et al., “Simultaneous in vivo optical quantification of key metabolic and vascular endpoints reveals tumor metabolic diversity in murine breast tumor models,” J. Biophotonics, 12 e201800372 https://doi.org/10.1002/jbio.201800372 (2018). Google Scholar

34. 

S. W. Perry et al., “Mitochondrial membrane potential probes and the proton gradient: a practical usage guide,” Biotechniques, 50 (2), 98 https://doi.org/10.2144/000113610 BTNQDO 0736-6205 (2011). Google Scholar

35. 

C. G. Zhu et al., “Near-simultaneous intravital microscopy of glucose uptake and mitochondrial membrane potential, key endpoints that reflect major metabolic axes in cancer,” Sci. Rep., 7 13772 https://doi.org/10.1038/s41598-017-14226-x SRCEC3 2045-2322 (2017). Google Scholar

36. 

J. Yan et al., “Optical imaging of HIF-1α mediated metabolic changes in the radio-resistant head and neck squamous carcinoma cells,” in Tech. Digest Ser., Optica Biophotonics Congr.: Biomed. Opt. 2024 (Transl., Microsc., OCT, OTS, BRAIN), JM4A.52 (2024). Google Scholar

37. 

P. S. Saha and C. Zhu, “Diffuse reflectance spectroscopy for optical characterizations of orthotopic head and neck cancer models in vivo,” Biomed. Opt. Express, 15 (7), 4176 –4189 https://doi.org/10.1364/BOE.528608 BOEICL 2156-7085 (2024). Google Scholar

38. 

S. H. Li et al., “A novel mode of action of YC-1 in HIF inhibition: stimulation of FIH-dependent p300 dissociation from HIF-1α,” Mol. Cancer Therapeut., 7 (12), 3729 –3738 https://doi.org/10.1158/1535-7163.MCT-08-0074 (2008). Google Scholar

39. 

A. Krysztofiak et al., “Metabolism of cancer cells commonly responds to irradiation by a transient early mitochondrial shutdown,” iScience, 24 (11), 103366 https://doi.org/10.1016/j.isci.2021.103366 (2021). Google Scholar

40. 

M. Feoktistova, P. Geserick and M. Leverkus, “Crystal violet assay for determining viability of cultured cells,” Cold Spring Harb. Protoc., 2016 (4), pdb.prot087379 https://doi.org/10.1101/pdb.prot087379 (2016). Google Scholar

41. 

C. Zou, Y. Wang and Z. Shen, “2-NBDG as a fluorescent indicator for direct glucose uptake measurement,” J. Biochem. Bioph. Methods, 64 (3), 207 –215 https://doi.org/10.1016/j.jbbm.2005.08.001 JBBMDG 0165-022X (2005). Google Scholar

42. 

M. C. Madonna et al., “Optical imaging of glucose uptake and mitochondrial membrane potential to characterize Her2 breast tumor metabolic phenotypes,” Mol. Cancer Res., 17 (7), 1545 –1555 https://doi.org/10.1158/1541-7786.MCR-18-0618 (2019). Google Scholar

43. 

M. Hassanein et al., “Development of high-throughput quantitative assays for glucose uptake in cancer cell lines,” Mol. Imaging Biol., 13 (5), 840 –852 https://doi.org/10.1007/s11307-010-0399-5 (2011). Google Scholar

44. 

P. E. Valk et al., “Cost-effectiveness of PET imaging in clinical oncology,” Nucl. Med. Biol., 23 (6), 737 –743 https://doi.org/10.1016/0969-8051(96)00080-7 (1996). Google Scholar

45. 

M. D. Farwell, D. A. Pryma and D. A. Mankoff, “PET/CT imaging in cancer: current applications and future directions,” Cancer, 120 (22), 3433 –3445 https://doi.org/10.1002/cncr.28860 CANCAR 0008-543X (2014). Google Scholar

46. 

L. C. Crowley, M. E. Christensen and N. J. Waterhouse, “Measuring mitochondrial transmembrane potential by TMRE staining,” Cold Spring Harb. Protoc., 2016 (12), pdb.prot087361 https://doi.org/10.1101/pdb.prot087361 (2016). Google Scholar

47. 

P. S. Ward and C. B. Thompson, “Metabolic reprogramming: a cancer hallmark even Warburg did not anticipate,” Cancer Cell, 21 (3), 297 –308 https://doi.org/10.1016/j.ccr.2012.02.014 (2012). Google Scholar

48. 

S. P. M. Crouch et al., “The use of ATP bioluminescence as a measure of cell-proliferation and cytotoxicity,” J. Immunol. Methods, 160 (1), 81 –88 https://doi.org/10.1016/0022-1759(93)90011-U JIMMBG 0022-1759 (1993). Google Scholar

49. 

R. C. Scaduto and L. W. Grotyohann, “Measurement of mitochondrial membrane potential using fluorescent rhodamine derivatives,” Biophys. J., 76 (1), 469 –477 https://doi.org/10.1016/S0006-3495(99)77214-0 BIOJAU 0006-3495 (1999). Google Scholar

50. 

M. Wu et al., “Multiparameter metabolic analysis reveals a close link between attenuated mitochondrial bioenergetic function and enhanced glycolysis dependency in human tumor cells,” Am. J. Physiol.-Cell Physiol., 292 (1), C125 –C136 https://doi.org/10.1152/ajpcell.00247.2006 (2007). Google Scholar

51. 

A. V. Kuznetsov et al., “Analysis of mitochondrial function in situ in permeabilized muscle fibers, tissues and cells,” Nat. Protoc., 3 (6), 965 –976 https://doi.org/10.1038/nprot.2008.61 1754-2189 (2008). Google Scholar

52. 

V. Chen et al., “Bezielle selectively targets mitochondria of cancer cells to inhibit glycolysis and OXPHOS,” Plos One, 7 (2), e30300 https://doi.org/10.1371/journal.pone.0030300 POLNCL 1932-6203 (2012). Google Scholar

53. 

S. Akakura et al., “Cancer cells recovering from damage exhibit mitochondrial restructuring and increased aerobic glycolysis,” Biochem. Biophys. Res. Commun., 448 (4), 461 –466 https://doi.org/10.1016/j.bbrc.2014.04.138 BBRCA9 0006-291X (2014). Google Scholar

54. 

N. Lynam-Lennon et al., “Altered mitochondrial function and energy metabolism is associated with a radioresistant phenotype in oesophageal adenocarcinoma,” Plos One, 9 (6), e100738 https://doi.org/10.1371/journal.pone.0100738 POLNCL 1932-6203 (2014). Google Scholar

55. 

J. S. Xie et al., “Beyond Warburg effect—dual metabolic nature of cancer cells,” Sci. Rep., 4 4927 https://doi.org/10.1038/srep04927 SRCEC3 2045-2322 (2014). Google Scholar

56. 

Y. Y. Fan et al., “A bioassay to measure energy metabolism in mouse colonic crypts, organoids, and sorted stem cells,” Am. J. Physiol.-Gastrointest. Liver Physiol., 309 (1), G1 –G9 https://doi.org/10.1152/ajpgi.00052.2015 (2015). Google Scholar

57. 

S. Y. Wang et al., “2-Deoxy-D-glucose can complement doxorubicin and sorafenib to suppress the growth of papillary thyroid carcinoma cells,” Plos One, 10 (7), e0130959 https://doi.org/10.1371/journal.pone.0130959 POLNCL 1932-6203 (2015). Google Scholar

58. 

K. Sellers et al., “Pyruvate carboxylase is critical for non–small-cell lung cancer proliferation,” J. Clin. Invest., 125 (2), 687 –698 https://doi.org/10.1172/JCI72873 (2015). Google Scholar

59. 

A. J. Walsh et al., “Optical metabolic imaging identifies glycolytic levels, subtypes, and early-treatment response in breast cancer,” Cancer Res., 73 (20), 6164 –6174 https://doi.org/10.1158/0008-5472.CAN-13-0527 CNREA8 0008-5472 (2013). Google Scholar

60. 

A. J. Walsh et al., “Quantitative optical imaging of primary tumor organoid metabolism predicts drug response in breast cancer,” Cancer Res., 74 (18), 5184 –5194 https://doi.org/10.1158/0008-5472.CAN-14-0663 CNREA8 0008-5472 (2014). Google Scholar

61. 

M. Skala and N. Ramanujam, “Multiphoton redox ratio imaging for metabolic monitoring in vivo,” Methods Mol. Biol., 594 155 –162 (2010). Google Scholar

62. 

J. Yan et al., “Portable optical spectroscopic assay for non-destructive measurement of key metabolic parameters on in vitro cancer cells and organotypic fresh tumor slices,” Biomed. Opt. Express, 14 (8), 4065 –4079 https://doi.org/10.1364/BOE.497127 BOEICL 2156-7085 (2023). Google Scholar

63. 

I. J. Hoogsteen et al., “The hypoxic tumour microenvironment, patient selection and hypoxia-modifying treatments,” Clin. Oncol., 19 (6), 385 –396 https://doi.org/10.1016/j.clon.2007.03.001 (2007). Google Scholar

64. 

S. Rockwell et al., “Hypoxia and radiation therapy: past history, ongoing research, and future promise,” Curr. Mol. Med., 9 (4), 442 –458 https://doi.org/10.2174/156652409788167087 (2009). Google Scholar

65. 

D. M. Aebersold et al., “Expression of hypoxia-inducible factor-1alpha: a novel predictive and prognostic parameter in the radiotherapy of oropharyngeal cancer,” Cancer Res., 61 (7), 2911 –2916 CNREA8 0008-5472 (2001). Google Scholar

66. 

J. F. Turrens, “Mitochondrial formation of reactive oxygen species,” J. Physiol., 552 (2), 335 –344 https://doi.org/10.1113/jphysiol.2003.049478 JPHYA7 0022-3751 (2003). Google Scholar

67. 

N. T. Moldogazieva, I. M. Mokhosoev and A. A. Terentiev, “Metabolic heterogeneity of cancer cells: an interplay between HIF-1, GLUTs, and AMPK,” Cancers, 12 (4), 862 https://doi.org/10.3390/cancers12040862 (2020). Google Scholar

68. 

S. H. Lee, M. Golinska and J. R. Griffiths, “HIF-1-independent mechanisms regulating metabolic adaptation in hypoxic cancer cells,” Cells, 10 (9), 2371 https://doi.org/10.3390/cells10092371 (2021). Google Scholar

69. 

M. C. Madonna et al., “In vivo optical metabolic imaging of long-chain fatty acid uptake in orthotopic models of triple-negative breast cancer,” Cancers, 13 (1), 148 https://doi.org/10.3390/cancers13010148 (2021). Google Scholar

Biography

Jing Yan received her BS degree in biotechnology from Southern Medical University, Guang Dong, China, in 2016. She is currently pursuing a PhD in biomedical engineering at the University of Kentucky, Lexington, Kentucky, United States. Her research interest includes the development of non-destructive imaging and spectroscopic techniques for cancer applications.

Carlos Frederico Lima Goncalves obtained his PhD in molecular biology from the Federal University of Rio de Janeiro in 2016. He joined Dr. Zhu lab in 2021 as a postdoc fellow to study the HNSCC radiation biology using in vitro cancer models. His research interests include cancer biology and cancer metabolism pathways.

Pranto Soumik Saha received his BS degree in biomedical engineering from Khulna University of Engineering & Technology, Khulna, Bangladesh, in 2021. He is currently pursuing a PhD in biomedical engineering at the University of Kentucky, Lexington, Kentucky, United States. His research interest includes the development of optical spectroscopy and imaging techniques for biomedical applications including cancer research.

Cristina M. Furdui received her PhD from the University of Nebraska, Lincoln, in 2002. She is currently a professor of Molecular Medicine at Wake Forest University. Her research interests include radiation tolerance, head and neck cancer, redox biology of inflammation and cancer, biospecimen preservation, and biomarker clinical assay development.

Caigang Zhu received his PhD from Nanyang Technological University in Singapore. He was a postdoc at Duke University and currently is an associate professor in the Department of Biomedical Engineering at the University of Kentucky. His research interests include optical spectroscopy and imaging technology development and their applications for cancer research.

CC BY: © The Authors. Published by SPIE under a Creative Commons Attribution 4.0 International License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

Funding Statement

Jing Yan, Carlos Frederico Lima Goncalves, Pranto Soumik Saha, Cristina M. Furdui, and Caigang Zhu "Optical imaging provides flow-cytometry–like single-cell level analysis of HIF-1α-mediated metabolic changes in radioresistant head and neck squamous carcinoma cells," Biophotonics Discovery 2(1), 012702 (28 January 2025). https://doi.org/10.1117/1.BIOS.2.1.012702
Received: 21 August 2024; Accepted: 3 January 2025; Published: 28 January 2025
Advertisement
Advertisement
KEYWORDS
Flow cytometry

Radiotherapy

Optical imaging

Glucose

Histograms

Mode conditioning cables

Microscopy

Back to Top