Monitoring and Targeting Metastasis Through Circulating Tumor Cells: From Molecular Profiling to Natural Killer Cell-Based Therapeutics

Abstract

Circulating tumor cells (CTCs) are the seeds of cancer metastasis and can be accessed through a simple blood draw. Recent developments in nanomaterials and microfluidic devices have enabled the invention of highly sensitive microfluidic platforms such as the graphene oxide (GO) chip. Such novel microfluidic devices have drastically advanced the isolation and characterization of CTCs, revealing that these rare cells may reflect disease status in cancer patients. However, further investigations in the context of clinical treatment studies are still needed to fully elucidate the value of CTCs as biomarkers. Additionally, targeting these rare CTCs might be a feasible therapeutic strategy to control cancer metastasis. In this thesis work, I first optimized the GO chip workflow to isolate CTCs from 16 metastatic bladder cancer patients. Interestingly, I found that patients with CTCs > 3/mL had worse overall survival. Additionally, through amplicon-based targeted sequencing and a novel, customized bioinformatic analysis pipeline to remove variances from contaminating white blood cell (WBC) signals, I found that several tumor-related genes, including KRT5, KRT10, MMP-2, and AKR1C2, were significantly upregulated in patients with metastatic disease. This similar workflow was then applied to longitudinal studies in stage III non-small-cell lung cancer (NSCLC), in which I developed analytical methods to correlate CTC counts and RNA expressions with clinical outcomes. CTCs were isolated and enumerated from 26 NSCLC patients and 6 time points across chemoradiation and immunotherapy treatment. The PD-L1 expression and RNA profile of these CTCs were characterized using on-chip immunofluorescent staining and microarray, respectively. I found that some CTC metrics, including a larger decrease in CTC counts after therapy started, and having ≤ 50% PD-L1+ CTCs at single time points, significantly predicted longer, progression-free survival. Compared to pretreatment, significant upregulations of genes related to cancer metastasis, invasion, proliferation, and resistance were found in CTC samples during chemoradiation. To explore potential therapies against these residual metastatic seeds, I hypothesized the critical role of NK cells in eliminating CTCs based on existing literature and investigated the CTC–NK interaction using novel single-cell resolution characterization methods. Through quantified NK cytotoxicity, 2 in-house CTC lines were both highly sensitive to NK-mediated killing compared to other lung-cancer-cell lines. By calculating an NK and epithelial to mesenchymal transition (EMT) score from bulk and single-cell RNA sequencing of CTC samples, I found a strong correlation between the NK-sensitive phenotype and EMT. In other words, CTCs, especially those with mesenchymal phenotypes, are highly sensitive to NK-mediated killing and have NK-sensitive signatures at the RNA and protein levels. Lastly, I adapted the GO platform to isolate patient-specific NK cells and generate a novel therapeutic agent, NK-derived exosomes. Compared to NK cells, exosomes can potentially infiltrate biological barriers and are less influenced by the immunosuppressive tumor microenvironment. In a small cohort of NSCLC patients, I isolated NK cells and CTCs and produced NK exosomes through a 12-hour, on-chip incubation. The NK cell and NK exosome concentrations showed correlations with bloodborne CTC numbers. And preliminary results then demonstrated that the NK exosomes harvested from the NK-GO chip had cytotoxic effects on in-house derived CTCs. The findings in this thesis strongly demonstrate the value of CTCs as biomarkers and therapeutic targets in clinical cancer management. Further, the methods developed in this work could be further applied to larger studies to advance the clinical utility of CTC-based liquid biopsies.