Selective isolation and noninvasive analysis of circulating cancer stem cells through Raman imaging

Circulating cancer stem cells (CCSCs), a rare circulating tumor cell (CTC) type, recently arose as a useful resource for monitoring and characterizing both cancers and their metastatic derivatives. However, due to the scarcity of CCSCs among hematologic cells in the blood and the complexity of the phenotype confirmation process, CCSC research can be extremely challenging. Hence, we report a nanoparticle-mediated Raman imaging method for CCSC characterization which profiles CCSCs based on their surface marker expression phenotypes. We have developed an integrated combinatorial Raman-Active Nanoprobe (RAN) system combined with a microfluidic chip to successfully process complete blood samples. CCSCs and CTCs were detected (90% efficiency) and classified in accordance with their respective surface marker expression via completely distinct Raman signals of RANs. Selectively isolated CCSCs (93% accuracy) were employed for both in vitro and in vivo tumor phenotyping to identify the tumorigenicity of the CCSCs. We utilized our new method to predict metastasis by screening blood samples from xenograft models, showing that upon CCSC detection, all subjects exhibited liver metastasis. Having highly efficient detection and noninvasive isolation capabilities, we have demonstrated that our RAN-based Raman imaging method will be valuable for predicting cancer metastasis and relapse via CCSC detection. Moreover, the exclusion of peak overlapping in CCSC analysis with our Raman imaging method will allow to expand the RAN families for various cancer types, therefore, increasing therapeutic efficacy by providing detailed molecular features of tumor subtypes. [Display omitted] •RANs are designed for detection, isolation, and analysis of CCSCs and CTCs.•CCSCs and CTCs are distinguished without signal overlapping by Raman imaging.•Noninvasive isolation of detected cells is conducted by using restriction enzyme.•The secondary tumor's subtype can be predicted with differentiated CCSC's profile.