Magnetic lanthanide sensor with self-ratiometric time-resolved luminescence for accurate detection of epithelial cancerous exosomes.

Fluorescence-based LB (liquid biopsy) offers a rapid means of detecting cancer non-invasively. However, the widespread issue of sample loss during purification steps will diminish the accuracy of detection results. Therefore, in this study, we introduce a magnetic lanthanide sensor (MLS) designed for sensitive detection of the characteristic protein, epithelial cell adhesion molecule (EpCAM), on epithelial tumor exosomes. By leveraging the inherent multi-peak emission and time-resolved properties of the sole-component lanthanide element, combined with the self-ratiometric strategy, MLS can overcome limitations imposed by manual operation and/or sample complexity, thereby providing more stable and reliable output results. Specifically, terbium-doped NaYF nanoparticles (NaYF:Tb) and deformable aptamers terminated with BHQ1 were sequentially introduced onto superparamagnetic silica-decorated FeO nanoparticles. Prior to target binding, emission from NaYF:Tb at 543 nm was partially quenched due to the fluorescence resonance energy transfer (FRET) from NaYF:Tb to BHQ1. Upon target binding, changes in the secondary structure of aptamers led to the fluorescence intensity increasing since the deconfinement of distance-dependent FRET effect. The characteristic emission of NaYF:Tb at 543 nm was then utilized as the detection signal (), while the less changed emission at 583 nm served as the reference signal (), further reporting the self-ratiometric values of and (/) to illustrate the epithelial cancerous features of exosomes while ignoring possible sample loss. Consequently, over a wide range of exosome concentrations (2.28 × 10-2.28 × 10 particles per mL), the / ratio exhibited a linear increase with exosome concentration [(/) = 0.166 lg () + 3.0269, = 0.9915], achieving a theoretical detection limit as low as 24 particles per mL. Additionally, MLS effectively distinguished epithelial cancer samples from healthy samples, showcasing significant potential for clinical diagnosis.