Single-Spheroid Imaging Tracks Secretion Suppression Before Later Cell Death
A combined optical assay followed VEGF-A capture, growth and viability around individual human colorectal tumor spheroids. Brefeldin A reduced secretion during an early viable interval, while later cell death restricted the treated-group analysis to five hours.
Measuring a Secreted Signal Alongside the Cells Producing It
A study led by the École Polytechnique Fédérale de Lausanne combined two imaging approaches to ask how a tumor spheroid’s secretory response relates to its growth and viability. One channel detected vascular endothelial growth factor A, or VEGF-A, captured near the spheroid. Fluorescence channels followed its cellular structure and membrane damage. The arrangement allowed repeated measurements of individual spheroids rather than pooling all of their secreted material into one endpoint.
The main drug-response experiments used spheroids made from the established human HCT-116 colorectal cancer cell line. A separate multicolor demonstration used CAL-27 tongue cancer cells. These were cell-line spheroids, with the principal statistical comparison centered on HCT-116, rather than a patient-derived organoid response cohort.
The sensing surface contained gold nanohole arrays coated with antibodies against VEGF-A. Binding near the surface changed transmitted light, producing an optical secretion map. The same instrument acquired fluorescence images from above the spheroid. Open microwells kept individual spheroids in position, and nearby empty wells supplied background and drift references.
An Early Inhibitor Response Had to Be Distinguished From Toxicity
The investigators imaged the spheroids for 12 hours at ten-minute intervals. Nuclear fluorescence defined projected size, and a membrane-impermeant SYTOX dye indicated dying cells. A U-Net neural network, which segments image regions, identified an exclusion zone around each spheroid so that movement, detached cells and debris would not be counted as secreted-protein signal.
Brefeldin A served as a secretion-pathway inhibitor. It reduced VEGF-A signal and spheroid growth, but the treated spheroids later developed viability loss and debris. The investigators therefore restricted their principal treated-group analysis to the first five hours, when viability remained stable. Untreated spheroids could be followed over the full 12-hour period.
Population-level secretion comparisons included 32 untreated and 23 treated spheroids. The reported mean maximum optical secretion signal was approximately 1.02 × 10⁸ arbitrary units in untreated spheroids and 1.76 × 10⁷ in treated spheroids. The corresponding mean secretion rates were approximately 1.00 × 10⁷ and 1.92 × 10⁶ arbitrary units per hour. These are sensor-signal measures of captured molecules, rather than absolute VEGF-A mass or the complete secretion into the surrounding volume.
The different observation windows matter when interpreting those maxima. The combined viability channel showed when an inhibitor-response comparison would become confounded by dying cells. Separately, the authors report an ELISA check consistent with suppressed VEGF-A secretion after treatment.
Structural Measurements Added Context Without Explaining All Secretion Variation
Projected-area analysis found positive average growth in untreated spheroids and a small negative average slope in treated spheroids. The growth-rate comparison used 32 untreated spheroids but only 13 treated ones: ten treated trajectories were excluded because their linear fits had a coefficient of determination below 0.6. That reduced denominator is separate from the 23-spheroid secretion analysis.
The investigators also collected three-dimensional fluorescence stacks at the beginning and end of the experiment. These supported cell-count, volume and dead-cell analyses in the untreated group. The treated group’s 12-hour stacks were excluded from this comparison because viability was already compromised after five hours.
Among untreated spheroids, mean live-cell counts increased from approximately 163 to 236, while the mean dead-cell fraction rose from 6.09% to 9.78%. Size, cell count and volume were strongly related, but their correlations with the captured secretion signal were only modest. Morphology therefore supplied useful context without serving as a complete surrogate for secretory activity.
The study’s optical design also defines what can be measured. Antibodies capture molecules at a two-dimensional surface, whereas spheroids secrete into three dimensions. Signal depends on diffusion, capture and the available surface as well as cellular production. Multiple simultaneous biomarkers and more complex tissue models are proposed extensions; the demonstrated drug comparison used one secretion inhibitor and one principal tumor cell line.