Exosomes contain a wealth of proteomic information, presenting
promising biomarkers for the noninvasive early diagnosis of diseases,
especially cancer. However, it remains a great challenge to accurately
and reliably distinguish exosomes secreted from different types of
cell lines. Fluorescence immunoassay is frequently used for exosome
detection. Nonspecific adsorption in immunoassays is unavoidable and
affects the reliability of assay results. Despite the fact that various
methods have been proposed to reduce nonspecific adsorption, a more
effective method that can eliminate the influence of nonspecific adsorption
is still lacking. Here, we report a more convenient way (named SR-TFC)
to remove the artifacts caused by nonspecific adsorption, which combines
tricolor fluorescence labeling of target exosomes, tricolor super-resolution
imaging, and pixel counting. The pixel counting method (named CFPP)
is realized by MATLAB and can eliminate nonspecific binding sites
at the single-pixel level, which has never been achieved before and
could improve the reliability of detection to the maximum extent.
Furthermore, as a proof-of-concept, profiling of exosomal membrane
proteins and identification of breast cancer subpopulations are demonstrated.
To enable multiplex breast cancer phenotypic analysis, three kinds
of specific proteins are labeled to obtain the 3D phenotypic information
on various exosomes. Breast cancer subtypes can be accurately identified
according to the super-resolution images of some clinically relevant
exosomal proteins. Worth mentioning is that, by selecting other biomarkers,
classification of other cancers could also be realized using SR-TFC.
Hence, the present work holds great potential in clinical cancer diagnosis
and precision medicine.