Resolve a drug-associated secretion change in the context of growth and emerging cytotoxicity.
Model & Design
Human HCT-116 colorectal cancer cell-line spheroids for primary response analysis; CAL-27 cells for a separate multichannel demonstration.
Experimental Format
Anti-VEGF-A gold-nanohole sensor with open microwells, fluorescence and brightfield imaging; 10-minute time resolution over 12 hours, with treated analysis restricted to a five-hour viable interval.
Model Components
Component
Spheroids and inhibitor comparator
Description
Untreated HCT-116 spheroids and brefeldin-A-treated spheroids; nuclear/cytoplasmic dyes and SYTOX death indicator.
Component
Spatial optical analysis
Description
Reference wells correct drift; U-Net segmentation defines exclusion zones to avoid movement/debris signal; z-stacks characterize untreated 3D structure at zero and 12 hours.
Measurements & Endpoints
VEGF-A-associated optical signal and spatial capture area
Signal-derived secretion rate
Projected size and growth slope
Live/dead cell counts and 3D structure
Analysis
Use optical intensity changes as relative captured-protein signals, not absolute total secretion. Analyze drug response within the period of maintained viability. U-Net is an image-segmentation neural network; the pipeline trained on 132 annotated images and tested on 17 additional images.
Published Evidence
Study Design
Untreated versus secretion-inhibited spheroids with concurrent viability monitoring; separate ELISA cross-check reported.
Study Sample
Secretion: n = 32 untreated and 23 treated spheroids. Growth rate: n = 32 and 13 after ten low-fit treated trajectories were excluded. Untreated 3D comparison: n = 32.
Comparison
Untreated spheroids and nearby empty reference wells; treated viability defines the usable interval.
Finding
Mean maximum sensor signals were 1.02 × 10⁸ versus 1.76 × 10⁷ a.u.; mean rates were 1.00 × 10⁷ versus 1.92 × 10⁶ a.u./h. Treated viability deteriorated after five hours, excluding 12-hour treated 3D outcomes; untreated mean live-cell counts rose from 163 to 236.
Interpretation & Limits
Untreated and treated maximum-signal windows differ.
Two-dimensional antibody capture samples only part of three-dimensional secretion.
A sensor signal is not absolute VEGF-A mass.
Main drug evidence uses one established tumor cell line and one inhibitor.
Growth-fit exclusion changes the treated denominator.
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.