Escribe una PREreview

The capacity of engineered nanoparticles (NPs) to traverse biological barriers represents a fundamental challenge in nanomedicine, targeted drug delivery, and nanotoxicology. Existing characterisation methods – Transwell® permeability assays, inductively coupled plasma atomic emission spectroscopy (ICP-AES), fluorescence microscopy, and flow cytometry – each suffer from inherent limitations including low throughput, marked inter-laboratory variability, or the requirement for particle labelling that alters physicochemical properties and compromises translational relevance. The present work describes and validates a fully integrated analytical platform that couples a standard cross-flow microfluidic chamber incorporating sequential porous-membrane cell barriers with label-free brightfield light microscopy and a multi-stage artificial intelligence / machine learning (AI/ML) pipeline. Intracellular nanoparticle accumulation, principally within lysosomes, produces characteristic organelle darkening that can be detected without additional labelling, segmented with high fidelity by a pre-trained ResAt-UNet convolutional neural network (IoU = 0.85, precision = 93.2 %, recall = 86.9 %), and converted into quantitative transport-efficiency (TE) and barrier-integrity metrics. PLGA-coated 15 nm SPIONs at 100 µg/mL achieved the highest TE values across the experimental matrix – 10.8 ± 1.5 % in HUVEC and 3.4 ± 0.6 % in hCMEC/D3 barriers under field-free conditions, rising to 13.2 ± 1.6 % and 4.1 ± 0.7 % respectively under 1 T magnetic guidance (a realistic +22 % relative gain). Concentration response was non-monotonic, generally plateauing between 100 and 250 µg/mL and declining at 500 µg/mL as receptor-mediated uptake saturated and sub-lethal toxicity began to compromise barrier tightness, although some variability in the plateau region was observed across formulations. The 100 and 150 nm carriers under 1 T magnetic field crashed barrier integrity through apical aggregation and magneto-mechanical stress, and were assigned N/A under the quality-control rule. Beyond a single endpoint, the platform delivers four interrelated NTE descriptors – transport fraction, transport rate, mean intracellular residence time, and transport heterogeneity index – whose biological plausibility is established by their coherent physicochemical scaling. Positioned as a New Approach Methodology (NAM), the platform offers a pathway for reducing animal experimentation in early-stage nanocarrier development.