Immediate sample fixation increases Circulating Tumour Cell (CTC) capture and preserves phenotype in head and neck squamous cell carcinoma: towards a standardised approach to microfluidic CTC biomarker discovery

Introduction: Research demonstrates strong evidence that circulating tumour cells (CTCs) can provide diagnostic and/or prognostic biomarkers in head and neck squamous cell carcinoma (HNSCC) and a potential tool for therapeutic stratification. However, the question still remains as to the optimum method of CTC enrichment and how this can be translated into clinical practice. We aimed to evaluate the Parsortix microfluidic device for CTC enrichment and characterisation in HNSCC, seeking to optimise a sample collection and processing protocol that preserves CTC integrity and phenotype.

Method: Spiking experiments of the FaDu and SCC040 HNSCC cell lines were used to determine the Parsortix capture rate of rare “CTC-like” cells. Capture rates of cancer cells spiked into EDTA blood collections tubes (BCTs) were compared to the Transfix fixative BCT and Cytodelics whole blood freezing protocol. The Lexogen Quantseq library preparation was used to profile gene expression of unfixed cells before and after microfluidic enrichment and enriched cell line spiked Transfix blood samples. An antibody panel was optimised to enable immunofluorescence microscopy CTC detection in HNSCC patient Transfix blood samples, using epithelial (EpCAM) and mesenchymal (N-cadherin) CTC markers.

Results: Across a spiked cell concentration range of 9–129 cells/mL, Parsortix demonstrated a mean cell capture rate of 53.5% for unfixed cells, with no significant relationship between spiked cell concentration and capture rate. Samples preserved in Transfix BCTs demonstrated significantly increased capture rates at 0 h (time to processing) compared to EDTA BCTs (65.3% vs. 51.0%). Capture rates in Transfix BCTs were maintained at 24 h and 72 h timepoints, but dropped significantly in EDTA BCTs. Gene expression profiling revealed that microfluidic enrichment of unfixed cell lines caused downregulation of RNA processing/binding gene pathways and upregulation of genes involved in cell injury, apoptosis and oxidative stress. RNA was successfully extracted and sequenced from Transfix preserved cells enriched using Parsortix, demonstrating epithelial specific transcripts from spiked cells. In a proof-of-concept cohort of four patients with advanced HNSCC, CTCs were successfully identified and visualised with epithelial and epithelial-mesenchymal phenotypes.

Conclusion: We have optimised a protocol for detection of CTCs in HNSCC with the Parsortix microfluidic device, using Transfix BCTs. We report a significant benefit, both in terms of cell capture rates and preserving cell phenotype, for using a fixative BCT- particularly if samples are stored before processing. In the design of large cohort multi-site clinical trials, such data are of paramount importance.