If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Pharmaceutical Sciences1, University of Pittsburgh, Pittsburgh, PA 15261, USA.University of Pittsburgh Medical Center Hillman Cancer Center2, Pittsburgh, PA 15232, USA.
In solid tumors like head and neck cancer (HNC), chronic and acute hypoxia have serious adverse clinical consequences including poorer overall patient prognosis, enhanced metastasis, increased genomic instability, and resistance to radiation-, chemo-, or immuno-therapies. However, cells in the two-dimensional monolayer cultures typically used for cancer drug discovery experience 20%-21% O2 levels (normoxic) which are 4-fold higher than O2 levels in normal tissues and ≥10-fold higher than in the hypoxic regions of solid tumors. The oxygen electrodes, exogenous bio-reductive markers, and increased expression of endogenous hypoxia-regulated proteins like HIF-1α generally used to mark hypoxic regions in solid tumors are impractical in large sample numbers and longitudinal studies. We used a novel homogeneous live-cell permeant HypoxiTRAK™ (HPTK) molecular probe compatible with high content imaging detection, analysis, and throughput to identify and quantify hypoxia levels in live HNC multicellular tumor spheroid (MCTS) cultures over time. Accumulation of fluorescence HPTK metabolite in live normoxic HNC MCTS cultures correlated with hypoxia detection by both pimonidazole and HIF-1α staining. In HNC MCTSs, hypoxic cytotoxicity ratios for the hypoxia activated prodrugs (HAP) evofosfamide and tirapazamine were much smaller than have been reported for uniformly hypoxic 2D monolayers in gas chambers, and many viable cells remained after HAP exposure. Cells in solid tumors and MCTSs experience three distinct O2 microenvironments dictated by their distances from blood vessels or MCTS surfaces, respectively; oxic, hypoxic, or intermediate levels of hypoxia. These studies support the application of more physiologically relevant in vitro 3D models that recapitulate the heterogeneous microenvironments of solid tumors for preclinical cancer drug discovery.
Head and neck cancers (HNC) are estimated to be the 6th most common cancer worldwide accounting for 890,000 new cases and 450,000 deaths per annum, and it is projected that 54,010 new patients will develop HNC and 10,850 will die of these cancers in 2021 in the USA. [
] Seven drugs are approved for HNC therapy, and while 5-year survival rates have improved modestly from 55% to 66% in the last 30 years, cure rates remain around 50%, and patients with advanced, recurrent, or metastatic HNC have a median survival of only 6-12 months. [
] Surgical resection and chemo-radiotherapy are the front-line therapies for localized or locoregionally confined HNC, and high dose cisplatin is the standard chemotherapy regimen for advanced metastatic HNC. [
] Cetuximab targets and blocks the epidermal growth factor receptor (EGFR) which is overexpressed in 80-90% of head and neck squamous cell carcinoma (HNSCC) tumors and is associated with poor patient progression-free and overall survival. Unfortunately, the clinical benefits of cetuximab have been relatively modest for most HNSCC patients. [
Improving Response Rates to EGFR-Targeted Therapies for Head and Neck Squamous Cell Carcinoma: Candidate Predictive Biomarkers and Combination Treatment with Src Inhibitors.
] The immune checkpoint antibody pembrolizumab (Keytruda®) which blocks the PD-1 receptor is well tolerated and shows clinically relevant antitumor activity in patients with recurrent or metastatic HNSCC, but only 16% of patients responded to treatment. [
Antitumor activity of pembrolizumab in biomarker-unselected patients with recurrent and/or metastatic head and neck squamous cell carcinoma: results from the phase Ib KEYNOTE-012 expansion cohort.
Safety and clinical activity of pembrolizumab for treatment of recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-012): an open-label, multicentre, phase 1b trial.
] The low response rates and limited efficacy of existing drugs emphasizes the need to discover new more effective HNC therapies. However, the discovery and development of new anti-cancer drugs is a difficult, lengthy, costly, and high-risk endeavor. On average, the complete process of drug discovery research and development, clinical trials, and regulatory approval for new therapeutic entities requires 10 to 14 years to complete, costs $2.6 billion or more, and succeeds only 12% of the time. Despite substantial investments in cancer research, drug discovery and development, the overall probability of success (POS) for new oncology drugs is ≤5%, ≥4-fold lower than therapeutic areas where the POS can be >20%: ophthalmology, infectious, and cardiovascular diseases. [
] Some of the major barriers to the clinical realization of cancer drug efficacy include patient-to-patient heterogeneity at the genetic, histologic, cell differentiation state, or clonal population composition of tumors, innate and/or acquired drug resistance, and attributes of the tumor microenvironment that limit drug distribution and effectiveness. [
] Despite these myriad challenges, precision medicine approaches employing a variety of omics technologies (genomics, proteomics, transcriptomics, and metabolomics) to identify biomarkers and stratify patient populations have improved success rates in some cancer clinical trials. [
Another approach to improve clinical success rates is to select and prioritize better drug candidates by implementing more physiologically relevant in vitro and in vivo models for cancer lead discovery and drug development. [
3-Dimensional culture systems for anti-cancer compound profiling and high-throughput screening reveal increases in EGFR inhibitor-mediated cytotoxicity compared to monolayer culture systems.
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
] For solid tumors, chronic and acute hypoxia have serious adverse clinical consequences including poorer overall patient prognosis, enhanced metastasis, increased genomic instability, and resistance to radiation-, chemo-, or immuno-therapies. [
] A lack of oxygen is rare in normal tissues but is characteristic of most solid tumors that are typically supplied by abnormal, randomly formed, or poorly organized blood vessels which access outer regions of tumor masses leaving inner cores poorly vascularized. [
] At O2 levels ≤2%, physiological hypoxia, tissues attempt to restore or maintain their preferred O2 level by upregulating hypoxia-activated genes and/or by increasing blood flow by vasodilation. [
] Normal cells adapt to hypoxia by reprogramming specific pathways to conserve energy and switch from oxidative phosphorylation to anaerobic glycolysis as their primary energy source. [
] At hypoxic O2 levels, cellular sensors stabilize hypoxia-inducible factors (HIF-1α, HIF-2α, or HIF-3α) which interact with ubiquitous HIF-1β to bind to hypoxia DNA response elements that modulate the transcription of >100 genes regulating critical cellular functions like angiogenesis, metabolism, proliferation, apoptosis, immunity, migration, and metastasis. [
] Peripheral tissues have median O2 levels ranging from 3.4% up to 6.8%, with 5% O2 providing an accurate approximation of normal tissue oxygenation (physoxia). [
] For >30 years, high throughput screening (HTS) growth inhibition (GI) assays conducted in tumor cell line panels have driven the discovery of novel cytotoxic compounds and/or the confirmation of the cytotoxicity of molecularly targeted agents. [
Implementation of the NCI-60 human tumor cell line panel to screen 2260 cancer drug combinations to generate >3 million data points used to populate a large matrix of anti-neoplastic agent combinations (ALMANAC) database.
] However, two-dimensional (2D) monolayer cultures maintained and assayed in normal incubators experience atmospheric O2 levels (normoxic) 20%-21%, 4-fold higher than O2 levels in normal tissues and ≥10-fold higher than in the hypoxic regions of solid tumors. [
] Three-dimensional (3D) in vitro culture models that recapitulate solid tumor architectures and microenvironments like hypoxia are therefore being deployed to identify and prioritize better cancer drug candidates. [
3-Dimensional culture systems for anti-cancer compound profiling and high-throughput screening reveal increases in EGFR inhibitor-mediated cytotoxicity compared to monolayer culture systems.
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
Multicellular tumor spheroid (MCTSs) cultures are 3D models that resemble avascular tumor nodules, micro-metastases, or the intervascular regions of large solid tumors with respect to morphology, cell-cell and cell-extracellular matrix (ECM) contacts. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] MCTSs exhibit volume growth kinetics, develop gradients of nutrient distribution and O2 concentration which give rise to diverse microenvironments with differential proliferation and drug distribution zones. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] Consistent with cells in solid tumors, the oxygenation status of distinct cells in MCTS cultures is determined by the size (diameter, volume, and surface area) of the spheroid, the distance that cells are from the surface, and the rates of O2 diffusion and consumption. [
] Hypoxia is clinically important in HNC, where it is a strong predictor of poorer patient prognosis, radiation therapy failure, immune evasion, and tumor recurrence. [
Maximizing the value of cancer drug screening in multicellular tumor spheroid cultures: a case study in five head and neck squamous cell carcinoma cell lines.
] Most of the studies that have provided insights into cellular responses to hypoxia have been performed in 2D cultures equilibrated in regulated gas chambers at hypoxic O2 levels. [
] However, cells in 2D monolayers cultured in gas chambers at uniform specific O2 levels do not experience the gradients, heterogeneity, or oscillations in O2 levels that cells in solid tumors encounter. [
] O2 levels can be measured directly by inserting polarographic needle oxygen electrodes into tissues or tumors, or indirectly by the bio-reduction of nitroimidazole compounds such as pimonidazole (PIMO) in hypoxic cells to form adducts with macromolecules detected by immuno-histochemistry or immuno-fluorescence. [
] However, O2 electrodes are not practical for measuring O2 levels in large numbers of samples or over time, and PIMO staining provides an indirect static assessment of O2 levels at a throughput and capacity limited by the requirement for extensive sample processing. We have previously characterized MCTS cultures generated in hydrogel microwell arrays and 384-well U-bottomed ultra-low attachment microtiter plates (ULA-plates) and compared their anticancer drug responses to 2D monolayer cultures of the same HNSCC cell lines.[
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] This manuscript describes the characterization of a novel cell-permeant molecular probe, HypoxiTRAK™ (HPTK), to identify and label endogenous hypoxic regions in live HNSCC MCTSs cultured under normoxic conditions. HPTK was converted to a fluorescent metabolite that accumulated in the cells of HNSCC MCTSs in a concentration and time dependent manner which was readily detected and quantified by high content imaging and analysis. Accumulation of the HPTK fluorescent metabolite in live MCTSs correlated with hypoxia detection by PIMO and HIF-1α staining in fixed MCTSs. Having established that normoxic HNSCC MCTS cultures contain endogenous hypoxic regions, we investigated whether hypoxia-activated prodrugs (HAPs) might be more efficacious in MCTS versus normoxic 2D monolayer cultures.
Materials and methods
Reagents
Formaldehyde (37%) was purchased from Sigma-Aldrich (St. Louis, MO). Cobalt (II) chloride hexahydrate and 99.9% high-performance liquid chromatography grade dimethyl sulfoxide (DMSO) were obtained from Alfa Aesar (Ward Hill, MA). Dulbecco's Mg2+- and Ca2+-free phosphate-buffered saline (PBS) was purchased from Gibco (Grand Island, NY). Dulbecco's modified Eagle's medium (DMEM) and Dulbecco's modified Eagle's medium / Ham's F12 50/50 (DMEM/F12) was purchased from Corning (Manassas, VA). Fetal bovine serum (FBS), L-glutamine, penicillin, and streptomycin (P/S) were purchased from Thermo Fisher Scientific. CellTiter-Blue® (CTB) was purchased from Promega Corporation (Madison, WI). Hoechst 33342, Calcein AM (CAM), Ethidium Homodimer (EHD) and Image-iT™ Green Hypoxia Reagent (ITGHR) were purchased from Life Technologies (Thermo Fisher Scientific, Waltham, MA). An HP3-100 Omni Kit containing pimonidazole HCl (PIMO) and antibodies that bind to pimonidazole adducts in hypoxic tissues was purchased from Hypoxyprobe (Burlington, MA). The HypoxiTRAK™ (HT10500, HPTK) molecular probe was kindly provided by Roy Edward of Biostatus (Shepshed, Loughborough, United Kingdom). HIF-1a (D1S7W) XP® Rabbit monoclonal antibody (36169S) was a gift from Cell Signaling Technology (Danvers, MA). Doxorubicin and Cisplatin were obtained from the National Cancer Institute (NCI) Developmental Therapeutics Program (DTP). [
Implementation of the NCI-60 human tumor cell line panel to screen 2260 cancer drug combinations to generate >3 million data points used to populate a large matrix of anti-neoplastic agent combinations (ALMANAC) database.
Confirmation of selected synergistic cancer drug combinations identified in an HTS campaign and exploration of drug efflux transporter contributions to the mode of synergy.
] Tirapazamine (TPZ) was purchased from Apexbio Technology LLC (Houston, TX). Evofosfamide (TH-302, EFF) and Deferoxamine mesylate were purchased from MedChemExpress LLC (Monmouth Junction, NJ).
Cells and tissue culture
The four human HNSCC cell lines were provided by Dr. Jennifer Grandis of the HNC SPORE at the University of Pittsburgh Medical Center Hillman Cancer Center and were maintained in a humidified incubator at 37°C, 5% CO2, and 95% humidity: Cal33, FaDu, UM-22B, and OSC-19. All cell lines were cultured in DMEM supplemented with 10% FBS, 1% L-glutamine, and 1% P/S. The culture medium for the FaDu and OSC-19 cell lines was also supplemented with 1% non-essential amino acids. HNSCC cell lines were passaged or used to generate MCTS cultures after isolated cell suspensions were prepared by dissociating cells in tissue culture flasks with trypsin and centrifugation at 270 x g for 5 min at room temperature, and resuspension in growth media. Viable trypan blue excluding cells in the suspension were counted using a hemocytometer.
Generation of HNSCC multicellular tumor spheroids in ultra-low attachment microtiter plates
MCTSs were produced by seeding HNSCC cell lines into 384-well U-bottomed ultra-low attachment (ULA-plates) microtiter plates (Corning, Tewksbury, MA) as described previously. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] Briefly, 384-well ULA-plates were rehydrated by adding 50 µL of serum free medium (SFM) to wells and incubating for 15 minutes in a humidified incubator at 37°C before the media was removed and 45μL of a single-cell suspension of HNSCC cell lines, seeded at 2,500 or 5,000 cells/well in growth medium, were transferred into wells using a Matrix automated multichannel pipette (Thermo Fisher Scientific, Waltham, MA), ULA-plates were centrifuged at 17 x g for 1 minute, and then placed in an incubator at 37°C, 5% CO2 and 95% humidity for the indicated time periods. For HNSCC MCTS cultures maintained in ULA-plates beyond 3 days, spent media was exchanged for fresh medium every 3 days using a Janus MDT Mini (PerkinElmer, Waltham, MA) automated liquid handler platform equipped with a 384-well transfer head. Each medium exchange cycle consisted of 2 × 20 µL aspiration and discard steps followed by 2 × 20 µL fresh media dispense steps. Three media exchange cycles achieved ∼ 85% media exchange and a uniform volume of 45uL per well.
High content imaging analysis of HNSCC multicellular tumor spheroid morphology, viability, growth and hypoxia probe staining
Images of HNSCC MCTSs were acquired on an ImageXpress Micro (IXM) automated wide field high content imaging platform and analyzed with algorithms from the integrated MetaXpress Imaging and Analysis software (Molecular Devices LLC, Sunnyvale, CA), as described previously. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] The IXM optical drive uses a 300 W Xenon lamp broad spectrum white light source and a 1.4-megapixel 2/3" chip Cooled CCD Camera and optical train for fluorescence and transmitted light phase contrast imaging. The IXM has Zero Pixel Shift (ZPS) filter sets: DAPI, FITC/ALEXA 488, CY3/TRITC, CY5, and Texas Red. Single images of HNSCC MCTSs were sequentially acquired using a 4X Plan Apo 0.20 NA objective in both the transmitted light (TL) and fluorescent image acquisition modes: DAPI, FITC, Texas Red or Cy5 channels. We used the IXM automated image-based focus algorithm to acquire both a coarse focus (large µm steps) set of images of Hoechst-stained objects in the DAPI channel for the first MCTS to be imaged, followed by a fine (small µm steps) set of images to select the best focus image. In subsequent wells and channels, only a fine focus set of images were acquired to select the best focus Z-plane. MCTS morphology and growth were assessed daily by acquiring 4X TL images on the IXM, and we used the line-scan tool of the MetaXpress software to measure HNSCC MCTS diameters and the pixel intensity profiles of the various fluorescent probes.
Live/dead staining of HNSCC multicellular tumor spheroids
We used the Calcein AM (CAM) live cell and Ethidium Homodimer (EHD) dead cell reagents to label viable and/or dead cells within HNSCC MCTS cultures, as described previously. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] HNSCC MCTS cultures were incubated with a cocktail of the Hoechst (8μg/mL), CAM (2.5μM), and EHD (5μM) reagents for 1h, and single images of HNSCC MCTSs were sequentially acquired on the IXM using a 4X objective in both the TL and fluorescent image acquisition modes: DAPI, FITC and Texas Red channels. We used the multiwavelength cell scoring (MWCS) image analysis module to analyze HNSCC MCTS fluorescent images as described previously. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] To create a whole MCTS mask we set the approximate minimum width of the Hoechst stained nuclei of the MCTS to be 150 µm with an approximate maximum width to be 550 µm and applied a threshold intensity above local background of 70. After applying user defined background average intensity thresholds, typically 50-70 in both the FITC and Texas Red channels, the MWCS module image segmentation created total MCTS masks in all three fluorescent channels and quantified the mean average (MAFI) and integrated fluorescence intensity (MIFI) of the CAM signal in the FITC channel and the EHD signal in the Texas Red channel. MAFI and MIFI values represent the average and total pixel fluorescent intensities in channels 1, 2 or 3 within MCTS masks of positively stained MCTSs above pre-set background thresholds.
Hypoxia staining of live and processed 2D monolayer and multicellular tumor spheroid HNSCC cultures
Live pre-formed 3-day HNSCC MCTS cultures or HNSCC cells seeded directly into 384-well ULA-plates were incubated with 0, 10, 30, 50 or 100 nM HypoxiTRAK™ (HT10500, HPTK) for 1-4 days and single images of HNSCC MCTSs were sequentially acquired on the IXM using a 4X objective in the TL and Cy5 fluorescent image acquisition modes. Pseudo-color fluorescence intensity data visualizations of Cy5 channel images were also used to illustrate HPTK fluorescent metabolite accumulation and distribution in HNSCC MCTS cultures. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
] The relative fluorescent intensities of the pixels in the image were represented as distinct colors with the “hotter” and “brighter” colors (low to high, yellow, red, white) representing higher intensity signals and cooler colors (low to high, purple, cyan, green) representing lower intensity signals. Alternatively, live HNSCC MCTS cultures were incubated with 2.5 µM Image-iT™ Green Hypoxia Reagent (ITGHR) for 6h or 1-4 days and single images of HNSCC MCTSs were sequentially acquired on the IXM using a 4X objective in the TL and FITC fluorescent image acquisition modes. Live 2D HNSCC monolayer cultures were stained similarly with HPTK or ITGHR.
To stain HNSCC MCTS with pimonidazole (PIMO), 1- and 3-day MCTS cultures were incubated for 3 h with 100 µM PIMO from an Omni Kit purchased from Hypoxyprobe (Burlington, MA). MCTSs were then fixed and stained for 1 h with 3.7% formaldehyde containing 8µg/mL Hoechst. The fixative was aspirated and fixed MCTSs were washed with 6 × 40 µL PBS exchanges using a Janus MDT Mini automated liquid handler platform equipped with a 384-well transfer head. MCTSs were then permeabilized with 0.5% triton-X 100 in PBS for 1 h. After 4 × 40µL washes with PBS on a Janus MDT Mini platform, MCTSs were blocked for 16 h at 4°C with 0.1% Tween-20 in PBS. Blocked MCTSs were then incubated with 1:100 dilution of affinity purified anti-pimonidazole rabbit primary antibody (Omni Kit) and incubated at 4°C for 16 h. After 4 × 40µL washes with PBS on a Janus MDT Mini platform, the MCTSs were incubated with a 1:1000 dilution of AlexaFluor-488 conjugated goat anti-rabbit antibody for 1 h at room temperature. Following a final 4 × 40 µL wash cycle on a Janus MDT Mini platform, single images of HNSCC MCTSs were sequentially acquired on the IXM using a 4X objective in the TL and FITC fluorescent image acquisition modes. 2D monolayer HNSCC cultures were stained with PIMO using the same protocol.
To stain HNSCC MCTSs for HIF-1α, 1- and 3-day MCTSs were fixed and stained for 1 h with 3.7% formaldehyde containing 8µg/mL Hoechst. The fixative was aspirated and fixed MCTSs were washed with 6 × 40 µL PBS exchanges on a Janus MDT Mini platform equipped with a 384-well transfer head. Fixed MCTSs were then permeabilized with 0.5% triton-X 100/PBS for 1 h. After 4 × 40µL washes with PBS on a Janus MDT Mini platform, MCTSs were blocked for 16 h at 4°C with 0.1% Tween-20 in PBS. Blocked MCTSs were then incubated with a 1:200 dilution of HIF-1a rabbit monoclonal primary antibody (mAb) at 4°C for 16 h. After 4 × 40µL washes with PBS on a Janus MDT Mini platform, the MCTSs were incubated with a 1:1000 dilution of AlexaFluor-488 conjugated goat anti-rabbit antibody for 1 h at room temperature. Following a final 4 × 40 µL wash cycle on a Janus MDT Mini platform, single images of HNSCC MCTSs were sequentially acquired on the IXM using a 4X objective in the TL and FITC fluorescent image acquisition modes. To chemically induce hypoxia in control wells, MCTSs were treated with 500 µM cobalt chloride or 100 µM deferoxamine for 24h before processing as described above. 2D monolayer HNSCC cultures were stained for HIF-1α using the same protocol.
To quantify HPTK, ITGHR, PIMO and HIF-1α staining we used the MWCS image analysis module to create a whole MCTS mask from the Hoechst-stained nuclei as described above and applied this mask to quantify the MAFI and MIFI values from the images acquired in the appropriate fluorescent channel; HPTK signal (Cy5), ITGHR signal (FITC), PIMO (FITC) and HIF-1α (FITC). For 2D monolayer cultures, we used the MWCS image analysis module to create individual nuclear masks of the Hoechst-stained nuclei and applied this mask together with background thresholding to quantify the MAFI and MIFI values of the cells from images acquired in the other fluorescent channels.
Analysis of HNSCC 2D monolayer and multicellular tumor spheroid culture viability and growth inhibition using the cell titer blue® reagent
The homogeneous CellTiter-Blue® (CTB) cell viability reagent was used to measure cell viability and/or growth inhibition of 2D monolayer and MCTS HNSCC cultures as described previously. [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] For 2D monolayer cultures, HNSCC cell lines were harvested, counted and seeded into 384-well assay plates at 750 cells per well in 45 µL of tissue culture media and cultured overnight at 37°C, 5% CO2, and 95% humidity. On the next day, 5 µL of test compounds at the indicated concentrations or plate controls diluted in SFM were transferred to assay plates on the Janus MDT Mini liquid handler equipped with a 384-well transfer head and assay plates were then cultured for an additional 72h in an incubator at 37°C, 5% CO2, and 95% humidity. Maximum control wells (Max controls, n=32) received 5 µL of DMSO (0.25% DMSO final) and minimum control wells (Min controls, n=32) received 200μM doxorubicin plus 0.25% DMSO. After the 72h compound exposure period 10 μL of the CTB cell viability detection reagent was dispensed into the wells of HNSCC assay plates, and incubated for 4h at 37°C, 5% CO2 and 95% humidity before capturing the relative fluorescent unit (RFUs) signals (Ex. 560 nm/ EM. 590 nm) on a SpectraMax M5e (Molecular Devices, LLC, Sunnyvale, CA) micro-titer plate reader platform. For MCTS growth inhibition assays HNSCC cell lines were seeded at 2,500 or 5,000 cells per well into 384-well ULA-plates in 45μL of growth medium incubated at 37°C, 5% CO2 and 95% humidity for 3 days. [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] After 3 days spent media was exchanged for fresh media as described above and test compounds at the indicted concentrations were transferred to the wells using a Janus MDT Mini platform equipped with a 384-well transfer head and the plates were returned to the incubator for an additional 72h. Maximum and minimum control wells received 0.25% DMSO and 200μM doxorubicin plus 0.25% DMSO respectively. After the 72h compound exposure period 10 μL of the CTB cell viability detection reagent was dispensed into the wells of HNSCC MCTS assay plates, and incubated for 4h at 37°C, 5% CO2 and 95% humidity before capturing RFU signals on a SpectraMax M5e micro-titer plate reader platform.
Data processing, analysis and curve fitting
For HNSCC 2D monolayer and MCTS growth inhibition (GI50) assays, the mean maximum DMSO control wells (Max controls n=32) and 200μM doxorubicin mean minimum plate control wells (Min controls, n=32) were used to normalize the data from compound treated wells and to represent uninhibited growth and 100% cytotoxicity, respectively. The GI50 data were fit to a non-linear sigmoidal log (inhibitor) vs. normalized response variable slope model using the equation: Y=100/(1+10^((LogIC50-X) *Hillslope))), where y was the percent growth inhibition and x was the corresponding log10 of the compound concentration. The GI50 is the concentration of compound that gives a 50% response, halfway between 0% and 100%. The Hillslope describes the steepness of the curve. All curve fitting, linear regression analysis, and graphs were created using the GraphPad Prism 6 software.
Results
HypoxiTRAK™ staining of live HNSCC multicellular tumor spheroid cultures
To investigate whether HNSCC MCTS cultures generated in 384-well ULA-plates and cultured under normoxic conditions would endogenously develop hypoxic regions capable of reducing the HPTK probe and accumulating its fluorescent metabolite, we conducted a series of HPTK concentration and time course experiments (Fig. 1). The first experiments were conducted in HNSCC MCTSs that had been allowed to form for 3 days prior to the addition of the HPTK probe because we speculated that established MCTSs were more likely to have developed hypoxic microenvironments. Fig. 1A shows images of FaDu MCTSs that were formed after seeding 5,000 cells per well in 384-well ULA-plates and culturing for 3-days before the media was exchanged and fresh media containing 30 nM HPTK was added to the MCTSs which were then returned to the incubator for up to 4 more days. Transmitted light, DAPI and Cy5 channel images of the established FaDu MCTSs were acquired after 4, 24, 48, 72 and 96 hours of continuous exposure to the HPTK probe. Transmitted light images of preformed FaDu MCTSs exposed to 30 nM HPTK for up to 4 days in culture were identical to those acquired in wells not exposed to the probe at all time points (Fig. 1A). The morphology, size (500-600 µm) and changes in FaDu MCTS diameters over time in culture ± HPTK were consistent with our previous observations with these MCTSs,[
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] indicating that the probe was not acutely cytotoxic. Both the area and intensity of the fluorescent staining observed in Cy5 channel images of FaDu MCTSs incubated with 30 nM HPTK increased with longer times in culture, whereas images of MCTSs not exposed to the probe exhibited only background staining (Fig 1A). Color composite images of the fluorescent HPTK metabolite (red) and Hoechst DNA (blue) stains indicated that the metabolite accumulated predominantly in the inner core regions of FaDu MCTSs (Fig. 1A). Fig. 1B shows a similar set of images acquired for up to 4 days after 5,000 FaDu cells were seeded directly into 384-well ULA-plates to form MCTSs in the presence of 50 nM HPTK. Based on the 3-day preformed MCTS data presented above we selected 50 nM HPTK as a suitable probe concentration to include during MCTS formation. The HPTK probe did not interfere with the ability of FaDu HNSCC cells to form MCTSs under these conditions, nor did it alter the morphology, size, and increase in MCTS diameters over time in culture (Fig. 1B). FaDu cells continuously exposed to HPTK throughout the MCTS formation and culture period exhibited fluorescent staining in Cy5 channel images that increased in area and intensity with longer times in culture (Fig. 1B), whereas Cy5 images of MCTSs not incubated with the probe only displayed background fluorescent staining. Color composite images of the HPTK fluorescent metabolite and Hoechst stained MCTSs again indicated that the metabolite accumulated preferentially in FaDu MCTS inner cores.
Fig. 1Hypoxia Staining of Live Head and Neck Squamous Cell Carcinoma Multicellular Tumor Spheroids (MCTSs) by the HypoxiTRAK™ (HPTK) Probe. (A) HPTK staining in FaDu MCTSs pre-formed in ultra-low attachment plates. FaDu MCTSs that were formed after seeding 5,000 cells per well in 384-well ULA-plates and culturing for 3-days before the media was exchanged and fresh media containing 30 nM HPTK was added to the MCTSs which were then returned to the incubator for up to 4 more days. Transmitted light, DAPI and Cy5 channel images of established FaDu MCTSs were acquired after 4, 24, 48, 72 and 96 hours of continuous exposure to the HPTK probe or media alone are presented. Color composite images of the DAPI and Cy5 channel images of the FaDu MCTSs continuously exposed to 30 nM HPTK for the indicated time periods are also presented. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. (B) HPTK staining during FaDu MCTS formation in ultra-low attachment plates. Transmitted light, DAPI and Cy5 channel images were acquired daily for up to 4 days after 5,000 FaDu cells were seeded directly into 384-well ULA-plates and allowed to form MCTSs in the presence of 50 nM HPTK. For comparison, only images from 96-hour control MCTSs that did not receive the probe are presented. Color composite images of the DAPI and Cy5 channel images of the FaDu MCTSs continuously exposed to 50 nM HPTK for the indicated time periods are also presented. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. (C) Mean Average (MAFI) and (D) Mean Integrated (MIFI) Fluorescent Intensity values for fluorescent metabolite accumulation in FaDu MCTSs versus HPTK concentration. (E) MAFI and (F) MIFI values for fluorescent metabolite accumulation in MCTSs versus time of HPTK exposure. The multiwavelength cell scoring (MWCS) image analysis module was used to create a whole MCTS mask from the Hoechst-stained nuclei and this mask was used to quantify the MAFI and MIFI values of the HPTK fluorescent metabolite in Cy5 channel images of the live MCTS cultures. For pre-formed MCTSs (PFM), data from a single replicate from a single experiment are presented. In the MCTS formation (MF) experiment, the data represent the mean ± sd of triplicate determinations. HPTK exposure times (, black) 24 hours, (, red) 48 hours, (, blue) 72 hours, and (, green) 96 hours. HPTK concentrations , black) 0 nM, (, red) 10 nM, (, blue) 30 nM, (, purple) 50 nM, and (, green) 100 nM. The MAFI and MIFI values of HPTK fluorescent metabolite staining in MCTSs increased linearly (r2 ≥ 0.9) with respect to both HPTK concentration and exposure time.
Fig. 1Hypoxia Staining of Live Head and Neck Squamous Cell Carcinoma Multicellular Tumor Spheroids (MCTSs) by the HypoxiTRAK™ (HPTK) Probe. (A) HPTK staining in FaDu MCTSs pre-formed in ultra-low attachment plates. FaDu MCTSs that were formed after seeding 5,000 cells per well in 384-well ULA-plates and culturing for 3-days before the media was exchanged and fresh media containing 30 nM HPTK was added to the MCTSs which were then returned to the incubator for up to 4 more days. Transmitted light, DAPI and Cy5 channel images of established FaDu MCTSs were acquired after 4, 24, 48, 72 and 96 hours of continuous exposure to the HPTK probe or media alone are presented. Color composite images of the DAPI and Cy5 channel images of the FaDu MCTSs continuously exposed to 30 nM HPTK for the indicated time periods are also presented. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. (B) HPTK staining during FaDu MCTS formation in ultra-low attachment plates. Transmitted light, DAPI and Cy5 channel images were acquired daily for up to 4 days after 5,000 FaDu cells were seeded directly into 384-well ULA-plates and allowed to form MCTSs in the presence of 50 nM HPTK. For comparison, only images from 96-hour control MCTSs that did not receive the probe are presented. Color composite images of the DAPI and Cy5 channel images of the FaDu MCTSs continuously exposed to 50 nM HPTK for the indicated time periods are also presented. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. (C) Mean Average (MAFI) and (D) Mean Integrated (MIFI) Fluorescent Intensity values for fluorescent metabolite accumulation in FaDu MCTSs versus HPTK concentration. (E) MAFI and (F) MIFI values for fluorescent metabolite accumulation in MCTSs versus time of HPTK exposure. The multiwavelength cell scoring (MWCS) image analysis module was used to create a whole MCTS mask from the Hoechst-stained nuclei and this mask was used to quantify the MAFI and MIFI values of the HPTK fluorescent metabolite in Cy5 channel images of the live MCTS cultures. For pre-formed MCTSs (PFM), data from a single replicate from a single experiment are presented. In the MCTS formation (MF) experiment, the data represent the mean ± sd of triplicate determinations. HPTK exposure times (, black) 24 hours, (, red) 48 hours, (, blue) 72 hours, and (, green) 96 hours. HPTK concentrations , black) 0 nM, (, red) 10 nM, (, blue) 30 nM, (, purple) 50 nM, and (, green) 100 nM. The MAFI and MIFI values of HPTK fluorescent metabolite staining in MCTSs increased linearly (r2 ≥ 0.9) with respect to both HPTK concentration and exposure time.
Fig. 1Hypoxia Staining of Live Head and Neck Squamous Cell Carcinoma Multicellular Tumor Spheroids (MCTSs) by the HypoxiTRAK™ (HPTK) Probe. (A) HPTK staining in FaDu MCTSs pre-formed in ultra-low attachment plates. FaDu MCTSs that were formed after seeding 5,000 cells per well in 384-well ULA-plates and culturing for 3-days before the media was exchanged and fresh media containing 30 nM HPTK was added to the MCTSs which were then returned to the incubator for up to 4 more days. Transmitted light, DAPI and Cy5 channel images of established FaDu MCTSs were acquired after 4, 24, 48, 72 and 96 hours of continuous exposure to the HPTK probe or media alone are presented. Color composite images of the DAPI and Cy5 channel images of the FaDu MCTSs continuously exposed to 30 nM HPTK for the indicated time periods are also presented. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. (B) HPTK staining during FaDu MCTS formation in ultra-low attachment plates. Transmitted light, DAPI and Cy5 channel images were acquired daily for up to 4 days after 5,000 FaDu cells were seeded directly into 384-well ULA-plates and allowed to form MCTSs in the presence of 50 nM HPTK. For comparison, only images from 96-hour control MCTSs that did not receive the probe are presented. Color composite images of the DAPI and Cy5 channel images of the FaDu MCTSs continuously exposed to 50 nM HPTK for the indicated time periods are also presented. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. (C) Mean Average (MAFI) and (D) Mean Integrated (MIFI) Fluorescent Intensity values for fluorescent metabolite accumulation in FaDu MCTSs versus HPTK concentration. (E) MAFI and (F) MIFI values for fluorescent metabolite accumulation in MCTSs versus time of HPTK exposure. The multiwavelength cell scoring (MWCS) image analysis module was used to create a whole MCTS mask from the Hoechst-stained nuclei and this mask was used to quantify the MAFI and MIFI values of the HPTK fluorescent metabolite in Cy5 channel images of the live MCTS cultures. For pre-formed MCTSs (PFM), data from a single replicate from a single experiment are presented. In the MCTS formation (MF) experiment, the data represent the mean ± sd of triplicate determinations. HPTK exposure times (, black) 24 hours, (, red) 48 hours, (, blue) 72 hours, and (, green) 96 hours. HPTK concentrations , black) 0 nM, (, red) 10 nM, (, blue) 30 nM, (, purple) 50 nM, and (, green) 100 nM. The MAFI and MIFI values of HPTK fluorescent metabolite staining in MCTSs increased linearly (r2 ≥ 0.9) with respect to both HPTK concentration and exposure time.
We used the multi-wavelength cell scoring (MWCS) image analysis algorithm to quantify the mean average (MAFI) and integrated (MIFI) fluorescent intensities of the HPTK fluorescent metabolite staining in Cy5 channel images of 3-day preformed (PFM) FaDu MCTSs continuously exposed to 0, 10, 30, or 100 nM HPTK throughout 4-days of normoxic culture conditions (Fig. 1C-1F). The MAFI and MIFI values of HPTK fluorescent metabolite staining in preformed FaDu MCTSs increased linearly (r2 ≥ 0.9) with respect to both HPTK concentration (Fig. 1C & 1D) and longer exposure times (Fig. 1E & 1F). The MAFI and MIFI values of HPTK fluorescent metabolite staining in FaDu cells continuously exposed to 50 nM HPTK throughout the MCTS formation (FM) and culture period also increased linearly (r2 ≥ 0.9) with respect to longer HPTK exposure times (Fig. 1E & 1F). HPTK concentration and exposure time course experiments conducted in MCTSs generated from Cal33, OSC19, or UM-22B HNSCC cell lines produced similar image and quantitative data sets (data not shown). Collectively these data demonstrated that HNSCC MCTS cultures established and cultured under normoxic conditions in 384-well ULA-plates converted the HPTK probe and accumulated its fluorescent metabolite in the inner cores of MCTSs in a concentration and time dependent manner. At concentrations up to 100 nM and continuous exposure up to 4 days, neither the HPTK probe nor its metabolite interfered with the ability of HNSCC cells to form MCTSs, nor did they alter the morphologies, sizes, or growth of the MCTSs.
HypoxiTRAK™ Staining of Multicellular Tumor Spheroids formed by different HNSCC Cell Lines
We selected 4 HNSCC cell lines that generate MCTSs with distinct growth phenotypes and morphologies to investigate whether these attributes might influence either the amount of HPTK fluorescent metabolite formed or its distribution. [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] The linear growth rates of the MCTSs formed by FaDu, UM-22B, and Cal33 HNSCC cell lines have been characterized as rapid, moderate, and slow respectively, while OSC19 MCTS have a progressive slow death phenotype. [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] The 4 HNSCC cell lines selected form relatively large MCTSs (≥500-600 µm diameters) when seeded at 5000 cells per well but with different morphologies; FaDu and Cal33 cell lines formed condensed MCTSs with a smooth and even periphery, the OSC19 cell line produced rounded MCTSs with uneven perimeters, and the UM-22B cell line formed large loose cell aggregates with irregular outer margins. [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] Fig. 2 shows images and quantitative data sets acquired 3 days after 5,000 FaDu, Cal33, OSC-19, or UM-22B HNSCC cells were seeded into 384-well ULA-plates to form MCTSs in the presence of 50 nM HPTK. The transmitted light images of the MCTSs indicated that the continuous presence of the HPTK probe did not interfere with the ability of the different HNSCC cell lines to form MCTSs in 384-well ULA-plates, nor did it alter either their morphologies or sizes (Fig 2A). MCTSs formed by the FaDu, Cal33 and OSC19 cell lines in the presence of HPTK exhibited fluorescent metabolite staining in Cy5 channel images that was substantially higher than the background levels observed in MCTSs not exposed to the probe (Fig. 2A). In contrast, the loose cell aggregates formed by the UM-22B cell line exhibited much weaker HPTK fluorescent metabolite staining in Cy5 channel images barely above no probe background levels (Fig. 2A). In pseudo-color fluorescence intensity data visualizations of the Cy5 channel images the “hotter” and “brighter” colors (low to high, yellow, red, white) represent higher intensity signals and cooler colors (low to high, purple, cyan, green) represent lower intensity signals. Both the Cy5 channel pseudo-color visualizations and the color composite images of the HPTK fluorescent metabolite and Hoechst stained MCTSs indicated that the metabolite accumulated preferentially in the inner cores of FaDu and Cal33 MCTSs, appeared evenly distributed throughout OSC19 MCTSs, and UM-22B cell aggregates exhibited much lower staining levels (Fig 2A).
Fig. 2HypoxiTRAK™ (HPTK) Hypoxia Staining in Live MCTS Cultures of different Head and Neck Squamous Cell Carcinoma (HNSCC) Cell Lines. (A) HPTK staining of HNSCC MCTSs formed in ultra-low attachment plates Transmitted light, DAPI and Cy5 channel images were acquired 3 days after 5,000 cells of the Cal33, FaDu, OSC-19 or UM-22B HNSCC cell lines were seeded directly into 384-well ULA-plates and allowed to form MCTSs in the presence of 50 nM HPTK. For comparison, similar grey scale images from 3-day control MCTSs that did not receive the probe are presented. Color composite images of the DAPI and Cy5 channel images of the MCTSs continuously exposed to 50 nM HPTK for 72 hours are also presented. Pseudo-color pixel intensity visualizations of the Cy5 channel images are also presented where the relative fluorescent intensities of the pixels are represented as distinct colors with the “hotter” and “brighter” colors (low to high, yellow, red, white) representing higher intensity signals and cooler colors (low to high, purple, cyan, green) representing lower intensity signals. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. Line-scan fluorescent intensity plots of DAPI and Cy5 Images from 3-day (B) Cal33 (C) FaDu, (D) OSC-19 and (E) UM-22B HNSCC MCTSs ± the HPTK Probe. Line-scan fluorescence intensity plots were created using the line scanning tool of the MetaXpress image analysis software to draw a line across the DAPI and Cy5 images and plot the derived fluorescent intensity values versus distance in µm across the image to provide an intensity profile of the accumulation and distribution of Hoechst and the HPTK fluorescent metabolite in MCTSs; (, blue) DAPI channel no probe control, (, cyan) DAPI channel HPTK treated, (, dark red) Cy5 channel no probe control, (, red) Cy5 channel HPTK treated. Representative data from one independent experiment are presented. (F) MAFI and (G) MIFI values for fluorescent metabolite accumulation in 3-Day HNSCC MCTSs ± the HPTK Probe. The multiwavelength cell scoring (MWCS) image analysis module was used to create a whole MCTS mask from the Hoechst-stained nuclei of 3-day MCTS cultures of Cal33, FaDu, OSC-19 or UM-22B HNSCC cell lines and this mask was used to quantify the MAFI and MIFI values of the HPTK fluorescent metabolite in Cy5 channel images of the live MCTS cultures; ( black) no probe control, ( red) HPTK treated. The data represent the mean ± sd of triplicate determinations from a single experiment.
Fig. 2HypoxiTRAK™ (HPTK) Hypoxia Staining in Live MCTS Cultures of different Head and Neck Squamous Cell Carcinoma (HNSCC) Cell Lines. (A) HPTK staining of HNSCC MCTSs formed in ultra-low attachment plates Transmitted light, DAPI and Cy5 channel images were acquired 3 days after 5,000 cells of the Cal33, FaDu, OSC-19 or UM-22B HNSCC cell lines were seeded directly into 384-well ULA-plates and allowed to form MCTSs in the presence of 50 nM HPTK. For comparison, similar grey scale images from 3-day control MCTSs that did not receive the probe are presented. Color composite images of the DAPI and Cy5 channel images of the MCTSs continuously exposed to 50 nM HPTK for 72 hours are also presented. Pseudo-color pixel intensity visualizations of the Cy5 channel images are also presented where the relative fluorescent intensities of the pixels are represented as distinct colors with the “hotter” and “brighter” colors (low to high, yellow, red, white) representing higher intensity signals and cooler colors (low to high, purple, cyan, green) representing lower intensity signals. A scale bar indicating 300 µm is shown in the upper left image and representative images from one experiment are presented. Line-scan fluorescent intensity plots of DAPI and Cy5 Images from 3-day (B) Cal33 (C) FaDu, (D) OSC-19 and (E) UM-22B HNSCC MCTSs ± the HPTK Probe. Line-scan fluorescence intensity plots were created using the line scanning tool of the MetaXpress image analysis software to draw a line across the DAPI and Cy5 images and plot the derived fluorescent intensity values versus distance in µm across the image to provide an intensity profile of the accumulation and distribution of Hoechst and the HPTK fluorescent metabolite in MCTSs; (, blue) DAPI channel no probe control, (, cyan) DAPI channel HPTK treated, (, dark red) Cy5 channel no probe control, (, red) Cy5 channel HPTK treated. Representative data from one independent experiment are presented. (F) MAFI and (G) MIFI values for fluorescent metabolite accumulation in 3-Day HNSCC MCTSs ± the HPTK Probe. The multiwavelength cell scoring (MWCS) image analysis module was used to create a whole MCTS mask from the Hoechst-stained nuclei of 3-day MCTS cultures of Cal33, FaDu, OSC-19 or UM-22B HNSCC cell lines and this mask was used to quantify the MAFI and MIFI values of the HPTK fluorescent metabolite in Cy5 channel images of the live MCTS cultures; ( black) no probe control, ( red) HPTK treated. The data represent the mean ± sd of triplicate determinations from a single experiment.
The line scanning tool of the MetaXpress image analysis software was used to draw lines through MCTS DAPI and Cy5 images and export the fluorescent intensity values versus distance in µm to generate fluorescence intensity profile graphs. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] The line-scan fluorescence intensity plots of Hoechst and HPTK fluorescent metabolite accumulation in 3-day HNSCC MCTS cultures formed and cultured ± 50 nM HPTK are presented in Fig. 2B (Cal33), 2C (FaDu), 2D (OSC19), and 2E (UM-22B). The fluorescence intensity plots of the Hoechst images (DAPI channel) of all 4 HNSCC MCTSs were essentially superimposable irrespective of the presence or absence of the HPTK probe. In contrast, the fluorescence intensity plots of the HPTK fluorescent metabolite images (Cy5 channel) indicated that only FaDu, Cal33 and OSC19 MCTSs exposed to HPTK had substantial metabolite staining above no probe background levels (Fig. 2B, 2C, and 2D). The loose irregular cell aggregates formed by UM-22B cells exposed to HPTK exhibited very low fluorescence intensity profiles barely higher than the background levels in aggregates not exposed to the probe (Fig. 2E). The fluorescence intensity plots of Hoechst and HPTK fluorescent metabolite distribution in 3-day HNSCC MCTSs, confirmed that the metabolite preferentially accumulated in the inner cores of FaDu and Cal33 MCTSs, but was more evenly distributed throughout OSC19 MCTSs (Fig. 2B, 2C, and 2D). The average and integrated fluorescent intensity values of HPTK fluorescent metabolite staining in 3-day HNSCC MCTS cultures ± 50 nM HPTK are presented in Fig. 2F and 2G, respectively. All four HNSCC MCTS cultures formed and cultured in the presence of HPTK exhibited significantly higher fluorescent metabolite MAFI values (p-value < 0.05) than MCTSs not exposed to the probe, but only FaDu, Cal33 and OSC19 MCTSs produced significantly higher MIFI values (p-value < 0.05) than no probe controls. All three HNSCC cell lines that established large compact MCTSs with either smooth or slightly irregular perimeters after 3 days of normoxic culture conditions in 384-well ULA-plates converted the HPTK probe into a fluorescent metabolite that accumulated in the MCTSs. In contrast, the large irregular loose cell aggregates formed by the UM-22B cell line under these conditions exhibited much weaker HPTK fluorescent metabolite staining, barely above background levels. There was no obvious correlation between HNSCC MCTS growth phenotypes and the ability to produce and accumulate the fluorescent HPTK metabolite. The fluorescent intensity of HPTK metabolite staining in MCTSs was substantially reduced after fixation (data not shown).
Comparison of HypoxiTRAK™ staining to different hypoxia staining methods in HNSCC multicellular tumor spheroid and 2D monolayer cultures
We wanted to determine whether the HPTK fluorescence metabolite accumulation detected in live HNSCC MCTS cultures was consistent with the detection of hypoxia by more established methods such as pimonidazole (PIMO) or HIF-1α staining. We also evaluated the Image-iT™ Green Hypoxia Reagent (ITGHR) which is reportedly another live cell hypoxia detection reagent. To compare with MCTS cultures, we used 2D monolayer cultures of the same HNSCC cell lines incubated under the same conditions, exposed to the same hypoxia reagents, and processed using the same protocols. We evaluated the four different hypoxia detection methods in MCTS and 2D monolayer cultures prepared from all four HNSCC cell lines, but for the sake of brevity only the data for the FaDu cell line are presented. Fig. 3 shows the images and quantitative data sets obtained from 3-day FaDu MCTS and 2D monolayer cultures that were continuously exposed to either 50 nM HPTK, 2.5 µM ITGHR, or were incubated with 100 µM PIMO for 3 hours prior to processing. To detect the presence of PIMO adducts or HIF-1α expression, MCTS and 2D monolayer cultures were fixed, permeabilized and processed with the appropriate primary and secondary antibodies for immunofluorescence detection. To validate the HIF-1α primary antibody and immunofluorescence detection method, we showed that 24-hour exposure of FaDu 2D monolayer cultures to commonly used chemical hypoxia-mimetic agents, either 500 µM cobalt chloride (CoCl2) or 100 µM of the iron chelator deferoxamine (DFO), significantly increased the nuclear accumulation of HIF-1α in these cells (Suppl. Fig 1). All four hypoxia detection reagents and methods produced substantial fluorescent staining in the images of 3-day FaDu MCTSs that was not present in the corresponding images of 2D monolayer cultures; Cy5 images for HPTK (Fig 3A), or FITC images for ITGHR (Fig 3B), PIMO adducts (Fig 3C), and HIF-1α (Fig 3D).
Fig. 3Hypoxia Staining in Live FaDu MCTS and 2D Monolayer Head and Neck Squamous Cell Carcinoma (HNSCC) Cultures. (A) HypoxiTRAK™ (HPTK), (B) Image-iT™ Green Hypoxia Reagent (ITGHR), (C) Pimonidazole (PIMO) and (D) HIF-1α. Live FaDu MCTSs and 2D monolayer cultures were continuously exposed to 50 nM HPTK or 2.5 µM ITGHR for 3 days before images were acquired on the IXM HCS platform. To stain with PIMO, cultures were exposed to 100 µM PIMO for the last 3 hours of the 3-day culture period before fixation and processing for immunofluorescence detection. To stain for HIF-1α 3-day cultures were fixed and processed for immunofluorescence detection. For HPTK, DAPI and Cy5 channel images were acquired. For ITGHR, PIMO and HIF-1α, DAPI and FITC channel images were acquired. Grey scale individual channel and color composite images of the DAPI and probe channels are presented for MCTS and 2D monolayer cultures ± the different hypoxia probes, HPTK or ITGHR, or no primary antibody controls for PIMO and HIF-1α. Scale bars indicating 300 µm are shown in the upper left images of MCTS and 2D monolayer cultures respectively. MAFI values (E) HPTK, (F) ITGHR, (G) PIMO, (H) HIF-1α and MIFI values (I) HPTK, (J) ITGHR, (K) PIMO, (L) HIF-1α for FaDu MCTS and 2D Monolayer HNSCC Cultures. The multiwavelength cell scoring (MWCS) image analysis module was used to quantify the MAFI and MIFI values of the hypoxia probes in the Cy5 or FITC channel images of FaDu MCTSs and 2D monolayers after 24 or 72 hours of culture. ( black) no probe or no primary antibody control, ( red) HPTK treated, ( green) ITGHR treated, ( purple) PIMO, and ( orange) HIF-1α. The data represent the mean ± sd of 6-8 replicate determinations from a single experiment.
Fig. 3Hypoxia Staining in Live FaDu MCTS and 2D Monolayer Head and Neck Squamous Cell Carcinoma (HNSCC) Cultures. (A) HypoxiTRAK™ (HPTK), (B) Image-iT™ Green Hypoxia Reagent (ITGHR), (C) Pimonidazole (PIMO) and (D) HIF-1α. Live FaDu MCTSs and 2D monolayer cultures were continuously exposed to 50 nM HPTK or 2.5 µM ITGHR for 3 days before images were acquired on the IXM HCS platform. To stain with PIMO, cultures were exposed to 100 µM PIMO for the last 3 hours of the 3-day culture period before fixation and processing for immunofluorescence detection. To stain for HIF-1α 3-day cultures were fixed and processed for immunofluorescence detection. For HPTK, DAPI and Cy5 channel images were acquired. For ITGHR, PIMO and HIF-1α, DAPI and FITC channel images were acquired. Grey scale individual channel and color composite images of the DAPI and probe channels are presented for MCTS and 2D monolayer cultures ± the different hypoxia probes, HPTK or ITGHR, or no primary antibody controls for PIMO and HIF-1α. Scale bars indicating 300 µm are shown in the upper left images of MCTS and 2D monolayer cultures respectively. MAFI values (E) HPTK, (F) ITGHR, (G) PIMO, (H) HIF-1α and MIFI values (I) HPTK, (J) ITGHR, (K) PIMO, (L) HIF-1α for FaDu MCTS and 2D Monolayer HNSCC Cultures. The multiwavelength cell scoring (MWCS) image analysis module was used to quantify the MAFI and MIFI values of the hypoxia probes in the Cy5 or FITC channel images of FaDu MCTSs and 2D monolayers after 24 or 72 hours of culture. ( black) no probe or no primary antibody control, ( red) HPTK treated, ( green) ITGHR treated, ( purple) PIMO, and ( orange) HIF-1α. The data represent the mean ± sd of 6-8 replicate determinations from a single experiment.
Fig. 3Hypoxia Staining in Live FaDu MCTS and 2D Monolayer Head and Neck Squamous Cell Carcinoma (HNSCC) Cultures. (A) HypoxiTRAK™ (HPTK), (B) Image-iT™ Green Hypoxia Reagent (ITGHR), (C) Pimonidazole (PIMO) and (D) HIF-1α. Live FaDu MCTSs and 2D monolayer cultures were continuously exposed to 50 nM HPTK or 2.5 µM ITGHR for 3 days before images were acquired on the IXM HCS platform. To stain with PIMO, cultures were exposed to 100 µM PIMO for the last 3 hours of the 3-day culture period before fixation and processing for immunofluorescence detection. To stain for HIF-1α 3-day cultures were fixed and processed for immunofluorescence detection. For HPTK, DAPI and Cy5 channel images were acquired. For ITGHR, PIMO and HIF-1α, DAPI and FITC channel images were acquired. Grey scale individual channel and color composite images of the DAPI and probe channels are presented for MCTS and 2D monolayer cultures ± the different hypoxia probes, HPTK or ITGHR, or no primary antibody controls for PIMO and HIF-1α. Scale bars indicating 300 µm are shown in the upper left images of MCTS and 2D monolayer cultures respectively. MAFI values (E) HPTK, (F) ITGHR, (G) PIMO, (H) HIF-1α and MIFI values (I) HPTK, (J) ITGHR, (K) PIMO, (L) HIF-1α for FaDu MCTS and 2D Monolayer HNSCC Cultures. The multiwavelength cell scoring (MWCS) image analysis module was used to quantify the MAFI and MIFI values of the hypoxia probes in the Cy5 or FITC channel images of FaDu MCTSs and 2D monolayers after 24 or 72 hours of culture. ( black) no probe or no primary antibody control, ( red) HPTK treated, ( green) ITGHR treated, ( purple) PIMO, and ( orange) HIF-1α. The data represent the mean ± sd of 6-8 replicate determinations from a single experiment.
The average and integrated fluorescent intensity values of the 4 fluorescent hypoxia stains quantified using the MWCS image analysis algorithm in 1- and 3-day MCTSs and 2D monolayer FaDu cultures are presented in Fig. 3E to 3L. Although 3 of the 4 hypoxia detection reagents did not appear to negatively impact MCTS cultures at the concentrations and exposures used, continuous exposure to 2.5 µM ITGHR for 72 hours resulted in smaller MCTSs and was cytotoxic to FaDu 2D monolayer cultures (Fig. 3A, 3B, 3C, and 3D, and Suppl. Fig. 2). In MCTS cultures, all four hypoxia stains exhibited significantly higher MAFI values (p-value < 0.05) than no probe or no primary antibody controls, and HPTK and ITGHR MAFI values were substantially higher in 3-day MCTS cultures than in 1-day cultures (Fig. 3E, 3F, 3G and 3H). For PIMO and HIF-1α staining the increase in MAFI values between 1-day and 3-day MCTS cultures were less dramatic (Fig. 3G and 3H). In 2D monolayer cultures however, MAFI values for the four hypoxia stains were substantially lower than in MCTS cultures, with only the ITGHR and PIMO probes producing MAFI values above background controls, and there was no difference in values between 1-day and 3-day cultures (Fig. 3E, 3F, 3G, and 3H). Compared to background controls, all four hypoxia stains also exhibited significantly higher MIFI values (p-value < 0.05) in MCTS cultures and MIFI values were substantially higher in 3-day MCTS cultures than in 1-day cultures (Fig. 3I, 3J, 3K, and 3L). In contrast, the MIFI values for the four hypoxia stains in 2D monolayer cultures were so much lower than in MCTS cultures that the data did not register on the same Y axis scale (Fig. 3I, 3J, 3K, and 3L). The accumulation and detection of the HPTK fluorescence metabolite in live HNSCC MCTSs cultured under normoxic conditions correlated with the development of hypoxia detected by both PIMO and HIF-1α staining, and by the ITGHR live cell hypoxia detection reagent. Although it was expected that none of the hypoxia reagents would produce positive staining in 2D HNSCC monolayers cultured under normoxic conditions, both the ITGHR and PIMO reagents consistently produced higher staining than no probe or no primary antibody backgrounds, albeit at much lower levels than in MCTS cultures (Fig. 3 and Suppl. Fig. 3).
Cancer drug induced growth inhibition in 2D monolayer and multicellular tumor spheroid HNSCC cultures
To explore whether HNSCC MCTS cultures might be more sensitive than 2D monolayer cultures to the growth inhibitory and cytotoxic effects of the hypoxia-activated prodrugs tirapazamine (TPZ) and evofosfamide (EFF), we used the cell titer blue (CTB) metabolic activity reagent to monitor cell viability, growth and drug-induced growth inhibition as described previously. [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] We included two cytotoxic anti-cancer drugs for comparison, doxorubicin and the HNC standard of care drug cisplatin. We selected the FaDu and Cal33 HNSCC cells lines because MCTS cultures prepared from these cell lines display growth phenotypes, [
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] and exhibited strong hypoxia staining under these conditions (Figs 1-3). Fig. 4 shows representative GI50 curves produced in 1-day 2D monolayer and 3-day preformed MCTS cultures of the FaDu and Cal33 HNSCC cell lines exposed to the 4 drugs at the indicated concentrations for an additional 72 hours of incubation. In 2D monolayer cultures, doxorubicin, cisplatin, and EFF exhibited complete concentration response curves (two asymptotes) with good quality curve fits (r2 values >0.95), efficacies >80%, and calculable GI50s (Fig. 4 & Table 1). However, even though the curve fits were good and yielded calculable GI50s, TPZ produced incomplete curves (only one asymptote) in 2D monolayer cultures with efficacies in the 70%-80% range (Fig 4 & Table 1). Doxorubicin, and cisplatin also exhibited complete concentration response curves with good quality curve fits, efficacies >80%, and calculable GI50s in MCTS cultures (Fig. 4 & Table 1). However, EFF and TPZ produced incomplete curves in MCTS cultures with lesser quality curve fits (r2 ∼ 0.85 to 0.9), efficacies in the 30%-90% range, and calculable GI50s except for TPZ in FaDu MCTSs (Fig 4 & Table 1). Consistent with our previous studies,[
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
] doxorubicin and cisplatin GI50s were substantially lower in 2D HNSCC monolayers than in MCTS cultures, on average ≥ 60-fold and ≥ 9.9-fold respectively (Fig 4 & Table 1). 2D HNSCC monolayer cultures were also more sensitive than MCTS cultures to EFF (2.6-fold) and TPZ (1.5-fold) (Fig 4 & Table 1).
Fig. 42D Monolayer and Multicellular Tumor Spheroid Cell Titer Blue® Growth Inhibition Curves for Doxorubicin, Cisplatin, Evofosfamide and Tirapazamine in FaDu (A-D) and Cal33 (E-F) Head and Neck Squamous Cell Carcinoma Cell Lines. For 2D-monolayer cultures, the FaDu and Cal33 HNSCC cell lines were seeded into 384-well assay plates and cultured for 24h before they were exposed to the indicated concentrations of doxorubicin, cisplatin, evofosfamide or tirapazamine for 72h prior to the addition of CTB and measurement of the RFU signals. For MCTS cultures, the FaDu and Cal33 HNSCC cell lines were seeded in 384-well ULA-plates and after 3 days in culture the MCTSs were exposed to the indicated concentrations of the drugs for 72h prior to the addition of CTB and measurement of the RFU signals. The mean maximum (0.5% DMSO) and minimum (200 µM doxorubicin + 0.5% DMSO) plate control CTB RFUs were used to normalize the RFU data from the compound treated wells as % inhibition of growth and the GI50 data were fit to a non-linear sigmoidal log inhibitor concentration versus the normalized response variable slope model using the GraphPad Prism 6 software. The normalized mean ± SD (n=3) growth inhibition data from duplicate wells for each compound concentration are presented. The data and curve fits for 2D monolayer (, black circles) and MCTS (, red circles) cultures are indicated in black and red, respectively. Representative experimental data from one of three independent experiments are shown.
Table 1Head and Neck Squamous Cell Carcinoma 2D Monolayer and Multicellular Tumor Spheroid Culture Growth Inhibition 50 (GI50) Determinations.
HNSCC Cell Line
FaDu
Cal33
Culture Conditions
2D Monolayers
3D MCTSs
2D Monolayers
3D MCTSs
Compounds
Mean GI50
sd
n
Mean GI50
Sd
n
Mean GI50
sd
n
Mean GI50
sd
n
Doxorubicin
0.15
0.04
3
6.1
1.74
3
0.13
0.02
3
10.7
9.56
3
Cisplatin
16.5
2.46
3
73.2
17.85
3
11.4
2.05
3
176
9.89
3
Evofosfamide
24.3
4.05
3
65.6
15.05
3
52.6
17.2
3
135
18.9
3
Tirapazamine
117
2.55
2
>200µM
-
2
110
25.6
2
162
24.1
2
GI50 = growth inhibition 50 concentration.
Mean GI50 = mean GI50 from 2-3 independent experiments determined from 10-point serial dilution series conducted in triplicate wells per drug concentration.
sd = standard deviation of the mean GI50.
n = number of independent experiments.
> = greater than the maximum concentration tested.
It has been shown that including morphology and dead cell parameters in a multi-parameter drug impact score with cell viability readouts improves the stratification of drug responses and maximizes the value of these more physiologically relevant 3D tumor cultures. [
] Transmitted light, Hoechst, Calcein AM (CAM, live cell), and Ethidium homodimer (EHD, dead cell) images of FaDu MCTS cultures exposed to the indicated concentrations of EFF or TPZ for 72h are presented in Fig. 5A and 5B, respectively. The corresponding CTB RFU's and MAFI values for the CAM and EHD fluorescent stains of EFF and TPZ treated FaDu MCTS cultures normalized to the signals in DMSO control wells are presented in Fig. 5C and 5D, respectively. Transmitted light, CAM and EHD images of FaDu MCTSs exposed to 7.4 and 2.5 µM EFF for 72 hours, concentrations below where the CTB viability reagent registered inhibition, showed that MCTS morphologies were altered (size, shape, and density), live cell staining was reduced, and dead cell staining was increased (Fig. 5A). Furthermore, the normalized EFF concentration responses for the increase in dead cell EHD staining and loss of live cell CAM staining were substantially left-shifted relative to the CTB inhibition readout (Fig. 5C). Similar sets of images and quantitative data from FaDu MCTSs exposed to TPZ for 72 hours, showed that MCTS morphologies were altered, live cell staining was reduced, and dead cell staining was increased at concentrations lower than where CTB signals were substantially impacted (Fig. 5B and 5D). These data indicated that EFF and TPZ both had a more pronounced impact on HNSCC MCTS cultures at lower concentrations than the CTB viability data indicated.
Fig. 5Impact of Evofosfamide and Tirapazamine Exposure on FaDu Head and Neck Squamous Cell Carcinoma Multicellular Tumor Spheroid Cultures. Transmitted light, Hoechst, Calcein AM or Ethidium Homodimer images of FaDu HNSCC MCTS cultures treated with (A) Evofosfamide or (B) Tirapazamine. Grey scale transmitted light, Hoechst, Calcein AM (live) and Ethidium homodimer (dead) images together with live(green)/dead(red) color composite images of FaDu MCTS cultures exposed to DMSO or the indicated concentrations of Evofosfamide or Tirapazamine for 72 h are presented. All scale bars represent 300 µm. MCTS Cell Titer Blue® Growth Inhibition Curves and Normalized MIFI Calcein AM (live) or Ethidium Homodimer (dead) signals of FaDu MCTSs exposed to (C) Evofosfamide or (D) Tirapazamine. The CTB growth inhibition data and curve fits for FaDu MCTS cultures exposed to the indicated concentrations of evofosfamide or tirapazamine are indicated in blue (blue circle, ). The normalized mean ± SD (n=2) growth inhibition data from duplicate wells for each compound concentration are presented. The MIFI Calcein AM (green, ) and Ethidium homodimer (red, ) signals of FaDu MCTSs exposed to evofosfamide or tirapazamine were normalized and expressed as % of DMSO controls. The normalized MIFI data from singlicate wells for each drug concentration are presented. Representative data from one of three independent experiments are shown.
Fig. 5Impact of Evofosfamide and Tirapazamine Exposure on FaDu Head and Neck Squamous Cell Carcinoma Multicellular Tumor Spheroid Cultures. Transmitted light, Hoechst, Calcein AM or Ethidium Homodimer images of FaDu HNSCC MCTS cultures treated with (A) Evofosfamide or (B) Tirapazamine. Grey scale transmitted light, Hoechst, Calcein AM (live) and Ethidium homodimer (dead) images together with live(green)/dead(red) color composite images of FaDu MCTS cultures exposed to DMSO or the indicated concentrations of Evofosfamide or Tirapazamine for 72 h are presented. All scale bars represent 300 µm. MCTS Cell Titer Blue® Growth Inhibition Curves and Normalized MIFI Calcein AM (live) or Ethidium Homodimer (dead) signals of FaDu MCTSs exposed to (C) Evofosfamide or (D) Tirapazamine. The CTB growth inhibition data and curve fits for FaDu MCTS cultures exposed to the indicated concentrations of evofosfamide or tirapazamine are indicated in blue (blue circle, ). The normalized mean ± SD (n=2) growth inhibition data from duplicate wells for each compound concentration are presented. The MIFI Calcein AM (green, ) and Ethidium homodimer (red, ) signals of FaDu MCTSs exposed to evofosfamide or tirapazamine were normalized and expressed as % of DMSO controls. The normalized MIFI data from singlicate wells for each drug concentration are presented. Representative data from one of three independent experiments are shown.
Although clinical trial studies utilizing O2 electrodes have provided valuable information on oxygen tension (pO2) levels in normal tissues and tumors, the procedure is invasive and impractical for either longitudinal measurements or large sample numbers. [
] Hypoxic cells bio-reduce exogenously added nitroimidazole compounds such as PIMO or EF5 to adducts that irreversibly bind to proteins and DNA which can be detected by immuno-histochemistry or immuno-fluorescence. [
] Elevated PIMO staining occurs at O2 levels <9.2% but increases exponentially at O2 levels ≤1.3% which is below the 2% O2 threshold of physiological hypoxia. [
] PIMO staining provides an indirect readout of pO2 in tissues or tumor samples exposed to the reagent in vivo or in vitro but requires extensive sample processing before it can be qualitatively or quantitatively assessed. [
Maximizing the value of cancer drug screening in multicellular tumor spheroid cultures: a case study in five head and neck squamous cell carcinoma cell lines.
] Harvested tissue or tumor samples are fixed, sectioned, treated to reduce non-specific binding, exposed to specific primary antibodies to PIMO-adduct complexes, washed, incubated with secondary antibodies that are conjugated to enzymes or fluorophores, and then washed again. Immuno-detection of elevated expression and colocalization to PIMO positive regions of endogenous hypoxia responsive proteins such as HIF-1α, lactate dehydrogenase-5, glucose transporter-1, or monocarboxylate transporter-4 have been used to mark hypoxia in HNC tumors. [
Maximizing the value of cancer drug screening in multicellular tumor spheroid cultures: a case study in five head and neck squamous cell carcinoma cell lines.
] Increased expression and colocalization of multiple hypoxia protein markers were considered more definitive than individual protein analyses due to concerns that some proteins might not be expressed in all tumor types or that expression may be modulated by other stress pathways and might not be specific for hypoxia. [
Maximizing the value of cancer drug screening in multicellular tumor spheroid cultures: a case study in five head and neck squamous cell carcinoma cell lines.
] Higher throughput non-invasive homogeneous hypoxia probes with sufficient spatial resolution to dynamically measure hypoxia levels in tumor regions over time would be valuable tools to study the role(s) and contributions of hypoxic microenvironments to cancer progression and drug resistance.
Non-invasive imaging modalities to assess hypoxia based on oxygen quenching of phosphorescence or medical imaging technologies like positron emission tomography (PET) and magnetic resonance (MR) are presently in development, primarily for in vivo applications. [
] The HypoxiTRAK™ (HPTK) molecular probe described here shares spectral characteristics with the cell permeant anthraquinone DRAQ5™ nucleic acid fluorescent dye. [
Characteristics of a novel deep red/infrared fluorescent cell-permeant DNA probe, DRAQ5, in intact human cells analyzed by flow cytometry, confocal and multiphoton microscopy.
] HPTK is reduced by oxygen-binding hemoproteins to a far-red fluorescent metabolite that accumulates in cells at biologically relevant levels of hypoxia (<3 % O2; pO2 23 mm Hg), but not in cells under normoxic conditions. [
] Images of A549 monolayers cultured in a gas chamber at 1% O2 levels with 100 nM HPTK for 4 days exhibited strong fluorescent metabolite staining readily quantifiable by flow cytometry. [
] Confocal images of HepG2 hepatic and OE21 esophageal cancer cell lines cultured for 4 hours in the presence of 100 nM HPTK under hypoxic (<0.1% O2) conditions showed an increase in red fluorescence compared to normoxic (21% O2) cultures. [
] We show here that HNSCC MCTSs generated and/or incubated under normoxic conditions in 384-well ULA-plates converted the non-fluorescent HPTK probe to a fluorescent metabolite which accumulated in MCTSs (Fig. 1, Fig. 2, Fig. 3). Both the area and intensity of the fluorescent metabolite staining of MCTS cultures continuously exposed to HPTK increased linearly with longer incubation times and higher probe concentrations (Fig. 1). The production and accumulation of the HPTK fluorescence metabolite by live normoxic HNSCC MCTS cultures strongly correlated with hypoxia detection by both PIMO and HIF-1α staining in fixed MCTSs (Fig. 3). By comparison, all three hypoxia detection methods produced negligible staining in normoxic 2D monolayer cultures. Color composite images of HPTK fluorescent metabolite and Hoechst stained MCTSs, together with pseudo-color visualizations and fluorescence line scan intensity plots indicated that the fluorescent metabolite accumulated preferentially in the inner cores of FaDu and Cal33 MCTSs but appeared distributed throughout OSC19 MCTSs (Fig 2). Continuous exposure to ≤100 nM HPTK for up to 4 days did not interfere with the formation of MCTSs by HNSCC cells in 384-well ULA-plates, nor did it alter the morphology, size, or growth of established MCTSs, indicating that neither HPTK nor its fluorescent metabolite were acutely cytotoxic. Only HNSCC cell lines that formed large densely packed MCTSs generated and accumulated the fluorescent HPTK metabolite (Fig. 2). The UM-22B cell line which forms large irregular loose cell aggregates in ULA-plates only showed weak staining barely over background, and there was no apparent correlation between HNSCC MCTS growth phenotype and HPTK fluorescent metabolite staining.
Novel azide-based click chemistry has been pursued for use in bioreductive drugs and/or candidate hypoxia markers, and non-fluorescent azide-containing molecules were described that could be reduced to fluorescent amides in cells lines cultured at O2 levels <1%, but not under normoxic conditions. [
] In hypoxic conditions, azide containing molecules bound to CYP450 iron centers to form nitrenoid species that were reduced by NADPH to fluorescent amide species. [
] The non-fluorescent azide-containing molecule CH-02 was reduced to a fluorescent amide in HepG2 hepatic and OE21 esophageal cancer cells lines cultured at O2 levels <1%, but not in normoxic conditions. In 10- to 14-day 500-600 µm diameter normoxic HCT116 colorectal carcinoma MCTSs that were exposed to 20 µM of CH-02 and 40 µM PIMO for 8 hours, confocal images showed colocalized fluorescence amide and PIMO staining in fixed MCTS sections. [
We also evaluated the Image-iT™ Green Hypoxia Reagent (ITGHR), reported to be a live-cell permeable irreversible end-point assay probe that fluoresces in environments with O2 concentrations <5%, but where intensity increases at lower O2 levels. [
] A precursor of the ITGHR probe exhibited increased fluorescence in tumor cell lines cultured in low (2.5% or 5%) O2 levels for 24 hours, and in tumor tissues transplanted into nude mice. [
] HNSCC MCTSs cultured under normoxic conditions that were continuously exposed to the ITGHR reagent exhibited substantial fluorescent staining that increased with longer times in culture, consistent with HPTK, PIMO and HIF-1α hypoxia staining (Fig. 3 and Suppl. Fig. 3). However, 2D HNSCC monolayers cultured under normoxic conditions also exhibited substantial ITGHR staining, albeit at much lower levels than in MCTS cultures (Fig. 3 and Suppl. Fig. 3). Continuous exposure of HNSCC cultures to 2.5 µM ITGHR for 72 hours in ULA-plates resulted in smaller MCTSs than controls and was cytotoxic in 2D monolayer cultures (Fig. 3B and Suppl. Fig. 2). In MCTS cultures pulse labeled with ITGHR for 6 hours to reduce toxicity, staining declined with longer culture periods (data not shown), which was inconsistent with the three other hypoxia markers. Although ITGHR exhibited fluorescent staining in HNSCC MCTS cultures, concerns over cytotoxicity, staining in normoxic 2D monolayer cultures, and the decline in MCTS staining observed at longer culture periods with the pulse labeling protocol lessened overall confidence in the probe.
Oxygen electrodes, exogenous bio-reductive markers, and increased expression of endogenous hypoxia-regulated proteins have revealed hypoxic regions in HNC tumors that are associated with poorer patient prognosis and therapeutic outcomes. [
Maximizing the value of cancer drug screening in multicellular tumor spheroid cultures: a case study in five head and neck squamous cell carcinoma cell lines.
] When tumor oxygen concentrations decline to mild hypoxic levels (2.6% O2, 20 mmHg) HNC cells become more radioresistant, and at severe hypoxia levels (0.5% O2, 3mmHg) radiation therapy sensitivity drops precipitously. [
] Hypoxic HNC tumor cells that have reduced cell cycle progression and proliferation rates show enhanced resistance to chemotherapy agents directed at tumors with high growth factions. [
] Cells in hypoxic tumor regions that are distant from functioning blood vessels are exposed to sub-efficacious drug concentrations due to drug diffusion limitations and/or sequestration by intervening cells. [
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
] HIF-1α is a transcriptional activator of PD-L1 in myeloid and tumor cells, and hypoxia promotes the recruitment of regulatory T cells (Tregs), tumor associated macrophages (TAMs), production of PGE2, IL-6 and IL-10, and the immunosuppressive activities of myeloid cells in the tumor microenvironment. [
] In HNC a hypoxic tumor microenvironment contributes to the resistance to both radiation and chemotherapies while enhancing the ability of tumor cells to evade the immune system. [
] Hypoxia exerts a strong selection pressure promoting genome instability and genetic heterogeneity leading to the survival and proliferation of cells with a greater propensity for angiogenesis, escape from apoptosis, and metastasis. [
] The adverse clinical consequences of acute and/or chronic hypoxia in HNC prompted the evaluation of hypoxia-activated prodrugs (HAPs) to selectively target hypoxic tumors populated by the most aggressive and therapy resistant cells. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] In hypoxic conditions, HAPs undergo an initial one electron reduction by endogenous oxidoreductases to their active cytotoxic products which cause irreversible DNA damage ultimately leading to cell death. HAP bio-reductive activation is rapidly reversed in the presence of O2 resulting in hypoxic selectivity. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] We selected two HAPs with promising preclinical in vitro tumor cell line cytotoxicity and in vivo tumor xenograft data that have advanced to clinical trials for solid cancers including HNC. Evofosfamide (EFF, TH-402) (1-Methyl-2-nitro-1H-imidazol-5-yl)-methyl-N,N′-bis(2-bromoethyl)-phosphorodiamidate is cleaved in hypoxic cells to release an azole derivative and a bromo-iso-phosphoramide (Br-IPM) radical that alkylates DNA inducing intra- and inter-strand crosslinks and activating apoptosis. [
] In normoxic conditions, the Br-IPM radical anion rapidly reacts with O2 to regenerate the inactive EFF parent. Tirapazamine (TPZ, SR-4233), 3-amino-1,2,4-benzotriazine-1,4 dioxide is activated at very low O2 levels by intracellular reductases to a reactive radical species that induces DNA single- and double-strand breaks, chromosomal aberrations, and cell death. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] In normoxic conditions, the TPZ radical is back oxidized to the parent molecule in a futile cycle thereby conferring hypoxic selectivity.
The preclinical in vitro tumor cell line cytotoxicity data used to support the clinical application of HAPs in hypoxic tumors is almost exclusively generated by comparing hypoxic 2D monolayer cultures in regulated gas chambers to normoxic cultures, where hypoxic cytotoxicity ratios (normoxic/hypoxic) are characteristically large. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] EFF has exhibited hypoxia selective cytotoxicity in many tumor cell lines including 21 HNSCC cell lines where the median air/N2 hypoxic IC50 ratio was 360-fold. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] The HPTK, PIMO and HIF-1α data presented here for normoxic HNSCC MCTS cultures (Fig. 1, Fig. 2, Fig. 3) is consistent with previous PIMO and endogenous hypoxia responsive protein staining that have shown that MCTSs contain hypoxic regions. [
] When we compared the in vitro cytotoxicity of EFF and TPZ in HNSCC tumor cell lines cultured as normoxic 2D monolayers to MCTSs with endogenous hypoxic regions, MCTSs were 2.5-fold less sensitive than the 2D monolayers to EFF and 1.5-fold less sensitive to TPZ (Fig. 4 and Table 1). However, relying solely on cell viability readouts can under-estimate drug effects in MCTS cultures, [
] and the inclusion of morphology and dead cell staining showed that both HAPs had impacted HNSCC MCTSs more profoundly than the CTB GI50 data indicated (Fig. 5). Nevertheless, the adjusted hypoxic cytotoxicity ratios in HNSCC MCTSs were much smaller for EFF (10- to 20-fold) and TPZ (<10-fold) than have been reported for hypoxic 2D monolayers in gas chambers. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] In addition, residual CTB signals and CAM staining showed that numerous viable cells remained in HNSCC MCTS cultures after exposure to EFF or TPZ (Fig. 5).
In solid tumors, hypoxia can be either chronic and/or acute, distinguished by how long cells are hypoxic but also by the root cause. [
] Chronic or diffusion limited hypoxia occurs when proliferative pressure pushes cells away from perivascular areas and O2 becomes limited by diffusion and/or consumption by intervening cells. [
] Acute hypoxia occurs when sudden reductions in blood flow produces a temporary lack of O2 and nutrients due to leaky or collapsed blood vessels or increased intra-tumoral interstitial pressure. [
] Restoration of blood flow and reperfusion of acute hypoxia can trigger reactive oxygen release and increased levels of DNA damage, chromosomal breaks, and genomic instability. [
] Since O2 levels progressively decrease as distances from blood vessels increase, at least three distinct O2 microenvironments exist within solid tumors; cells close to blood vessels will be oxic, cells closest to necrotic regions and furthest away from blood vessels will be fully hypoxic, and intervening cells will experience intermediate levels of hypoxia. [
] Cells in solid tumors subject to acute hypoxia can also experience substantial oscillations and variability in O2 levels. Cells in MCTS cultures also experience heterogeneous O2 levels depending upon the relative distances of cells from the MCTS surface and the rates of O2 diffusion and consumption by intervening cells. [
] In addition to fully hypoxic cells capable of bioreducing HAPs to their cytotoxic forms, solid tumors and MCTS cultures contain cells with intermediate O2 levels that may be less efficient at activating the prodrug, and oxic cells where the active species is rapidly back oxidized to the inactive parent molecule. [
] Thus, cells in solid tumors and MCTS cultures located in regions with different O2 levels will be exposed to different concentrations of the activated prodrug. In contrast, cells in 2D monolayers cultured in regulated gas chambers at uniform hypoxic O2 levels will be equally capable of converting HAPs to their cytotoxic forms and will be exposed to the same concentrations of the activated drug. [
] Although the in vitro hypoxic selectivity of the EFF and TPZ in HNSCC MCTSs with endogenous hypoxic regions were much less compelling than for hypoxic 2D monolayer cultures, the MCTS data would seem to be more consistent with the clinical outcomes of these agents. Despite producing promising results in pre-clinical studies and early-Phase clinical trials, EFF and TZP both failed to demonstrate clinical benefits in Phase-III studies in combination with existing chemo- or radiation-therapies for HNC, or for other cancers. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] It was proposed that future HAP clinical trials would benefit from the development of dependable markers of tumor hypoxia in vivo that could be used to identify and select/stratify patients with tumors bearing a significant proportion of hypoxic cells for treatment. [
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
] Similarly, it would seem prudent to incorporate preclinical testing of HAPs in more physiologically relevant in vitro 3D models that more faithfully recapitulate the heterogeneous O2 microenvironments of solid tumors to better assess and/or predict their potential clinical applications and outcomes.
In conclusion, the data presented herein demonstrate that the HypoxiTRAK™ molecular probe is a homogeneous live-cell permeant reagent which provides a dynamic readout of longitudinal hypoxia levels in live MCTS cultures that is compatible with HCS detection, analysis, and throughput. Accumulation of the HPTK fluorescence metabolite in live normoxic HNSCC MCTS cultures correlated with the detection of hypoxic regions by both PIMO and HIF-1α staining. The hypoxic cytotoxicity ratios for the hypoxia activated prodrugs evofosfamide and tirapazamine in HNSCC MCTSs were much less impressive than have been reported for hypoxic 2D monolayers in gas chambers, and many viable cells remained in MCTS cultures after HAP exposure. These studies support the application of more physiologically relevant in vitro 3D models that recapitulate the heterogeneous microenvironments of solid tumors for preclinical cancer drug discovery.
Declaration of Competing Interest
The authors declare no conflict of interest with respect to the research, authorship, and publication of this article.
Acknowledgments
The studies were supported in part by a Development Research Project award (Johnston, PI) from the Head and Neck Cancer Spore P50 (Ferris and Grandis, CA097190) of the University of Pittsburgh Medical Center Hillman Cancer Center. Thanks to Roy Edward from Biostatus for providing the HypoxiTRAK™ molecular probe for these studies.
Improving Response Rates to EGFR-Targeted Therapies for Head and Neck Squamous Cell Carcinoma: Candidate Predictive Biomarkers and Combination Treatment with Src Inhibitors.
Antitumor activity of pembrolizumab in biomarker-unselected patients with recurrent and/or metastatic head and neck squamous cell carcinoma: results from the phase Ib KEYNOTE-012 expansion cohort.
Safety and clinical activity of pembrolizumab for treatment of recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-012): an open-label, multicentre, phase 1b trial.
3-Dimensional culture systems for anti-cancer compound profiling and high-throughput screening reveal increases in EGFR inhibitor-mediated cytotoxicity compared to monolayer culture systems.
High-content screening comparison of cancer drug accumulation and distribution in two-dimensional and three-dimensional culture models of head and neck cancer.
Implementation of the NCI-60 human tumor cell line panel to screen 2260 cancer drug combinations to generate >3 million data points used to populate a large matrix of anti-neoplastic agent combinations (ALMANAC) database.
High content screening characterization of head and neck squamous cell carcinoma multicellular tumor spheroid cultures generated in 384-well ultra-low attachment plates to screen for better cancer drug leads.
Maximizing the value of cancer drug screening in multicellular tumor spheroid cultures: a case study in five head and neck squamous cell carcinoma cell lines.
Confirmation of selected synergistic cancer drug combinations identified in an HTS campaign and exploration of drug efflux transporter contributions to the mode of synergy.
Characteristics of a novel deep red/infrared fluorescent cell-permeant DNA probe, DRAQ5, in intact human cells analyzed by flow cytometry, confocal and multiphoton microscopy.
Tirapazamine, cisplatin, and radiation versus cisplatin and radiation for advanced squamous cell carcinoma of the head and neck (TROG 02.02, HeadSTART): a phase III trial of the Trans-Tasman Radiation Oncology Group.
, black) 24 hours, (00008-3&id=gr1a.jpg){kind=link}
00008-3&id=gr1b.jpg){kind=link}
00008-3&id=gr1c.jpg){kind=link}
black) no probe control, (00008-3&id=gr2a.jpg){kind=link}
00008-3&id=gr2b.jpg){kind=link}
00008-3&id=gr3a.jpg){kind=link}
00008-3&id=gr3b.jpg){kind=link}
00008-3&id=gr3c.jpg){kind=link}
, black circles) and MCTS (00008-3&id=gr4.jpg){kind=link}
00008-3&id=gr5a.jpg){kind=link}
00008-3&id=gr5b.jpg){kind=link}
