Hydro-Seq enables contamination-free high-throughput single-cell RNA-sequencing for circulating tumor cells

Abstract

Molecular analysis of circulating tumor cells (CTCs) at single-cell resolution offers great promise for cancer diagnostics and therapeutics from simple liquid biopsy. Recent development of massively parallel single-cell RNA-sequencing (scRNA-seq) provides a powerful method to resolve the cellular heterogeneity from gene expression and pathway regulation analysis. However, the scarcity of CTCs and the massive contamination of blood cells limit the utility of currently available technologies. Here, we present Hydro-Seq, a scalable hydrodynamic scRNA-seq barcoding technique, for high-throughput CTC analysis. High cell-capture efficiency and contamination removal capability of Hydro-Seq enables successful scRNA-seq of 666 CTCs from 21 breast cancer patient samples at high throughput. We identify breast cancer drug targets for hormone and targeted therapies and tracked individual cells that express markers of cancer stem cells (CSCs) as well as of epithelial/mesenchymal cell state transitions. Transcriptome analysis of these cells provides insights into monitoring target therapeutics and processes underlying tumor metastasis.

Fig. 1

Hydro-Seq, a high capture efficiency scRNA-seq platform for contaminated rare samples. a–e Design of Hydro-Seq technology. Integrated microfluidic circuit design with valve controls for high-efficiency cell capture and contamination removal. a For circulating tumor cells (CTC) application, Hydro-Seq chip is designed with 16 identical branch channels in parallel. Each branch channel consists of 50 chambers for bead-cell pairing, totaling 800 chambers per chip. b A closer view highlights the arrangement of parallel chambers. To minimize area consumption, each chamber shares the valve with its neighboring chamber. The entrance valve has a peak height of 45 µm and an area of 200 × 200 µm to enable bead and cell loading. The cell and bead valves are realized in a height of 15 µm and an area of 100 × 100 µm for high-density chamber arrangement. c, d At the upstream and downstream of branch channels, the valve-controlled wash channels enable channel washing during sample loading and delivery of lysis buffer to the entrance channel for lysis. e Enlarged view of a microfluidic bead-cell pairing chamber. f–j Hydro-Seq operation flow. f Bead capture valve is closed during sample loading. The smaller red blood cells flow through the capture sites, while a larger cell (larger than 12 µm) blocks the channel for cell capture. g After cell loading, the bead capture valve is then opened to wash the contaminants away. h After removing contaminants in the chamber, beads are loaded for pairing. i Lysis buffer is introduced into the chamber. After closing all the valves, the beads are moved to the cell capture site for mRNA capture. j By opening all the valves and introducing a back flow, the beads can be retrieved to the inlet for downstream preparation

Fig. 2

Species-mixing experiment and CTC loading in Hydro-Seq. a–d Species-mixing experiment. a Chambers with beads paired to a mouse cell (3T3 with green fluorescence) and a human cell (HEK with red fluorescence). Fluorescent imaging was applied to examine the pairing condition and verify the capture. b Histograms of the percent cross-species contamination of mixed mouse and human cells. Cells with over 90% transcripts mapped to human were labelled red and cells with over 90% transcripts mapped to mouse were labeled green. c Species-mixing analysis highlights the single-cell resolution RNA-sequencing with zero cellular cross contamination. Each red dot represents a mouse cell and each blue dot represents a human cell. d tSNE plot of single-cell expression analysis for HEK cells from two independent Hydro-Seq experiments. (Experiment 1 labeled with red color and experiment 2 labeled with cyan color) Each dot represents a cell. Cells from two experiments are evenly dispersed and do not show any clustering, indicating good reproducibility of Hydro-Seq. e–h CTC loading in Hydro-Seq. e After CTC enrichment, single-cell RNA-sequencing (scRNA-seq) of the CTCs is still challenged by the presence of many background blood cells. (Scale bar: 25 µm) f Erythrocytes flowing through the chamber during sample loading. (Scale bar: 25 µm) g During the washing cycle, the bead valve is opened to remove other blood cells through the bead capture flow channel, achieving contamination-free single CTC isolation for bead pairing. h Pairing a bead to a single CTC for scRNA-seq. Source data are provided as a Source Data file

Fig. 3

Immunostaining and single-cell RNA-sequencing of CTCs. a With CD45 and Hoechst staining, the CD45 positive nucleated cells were identified as leukocytes. (Scale bar: 15 µm) b tSNE plot of all CTCs from patient samples processed by Hydro-Seq. (666 CTCs from 21 patient samples with each color representing individual patient sample.) c The reproducibility test of Hydro-Seq processing. For the two tubes of blood drawn from the same patient and processed on the same day, comparable CTC counts were achieved from the two experiments (Exp. 1: 13 CTCs with red color; Exp. 2: 12 CTCs with cyan color). The expression profiles of housekeeping, cell proliferation, epithelial, mesenchymal, and other genes are consistent, showing good reproducibility. Source data are provided as a Source Data file

Fig. 4

Gene expression and clustering of breast CTCs. a–d Breast cancer therapy related genes: a human epidermal growth factor receptor 2 (HER2/Erbb2), b estrogen receptor (ESR1), c Androgen (AR), and d progesterone receptor (PGR). e–h MET-related genes: e E-cadherin (CDH1), f Epithelial Cell Adhesion Molecule (EPCAM), g Keratin-8 (KRT8), and h Keratin-18 (KRT18). i, j Cell cycle genes: i c-jun (JUN), and j cyclin D1 (CCND1). k–n EMT-related genes: k Phosphoinositide-dependent kinase-1 (PDK1), l Plasminogen activator inhibitor-1 (SERPINE1), m EMT transcription factor (ZEB2), and n transforming growth factor β (TGFB1). o The clustering and separation of HER2+ MET-like and HER2- EMT-like CTCs. p–u Stemness related genes: p CD44, q CD24, r vimentin (VIM), s pan-ALDH isoforms (ALDH), t ALDH1a3, and u CD90 (THY1). (Each dot represents one CTC. Green color represents the lowest expression, and red color represents the highest expression. The expression is logarithmically normalized. 666 CTCs from 21 patient samples were plotted based on tSNE clustering method.) Source data are provided as a Source Data file

Fig. 5

CTC intra-patient heterogeneity of EMT-like and MET-like states. a 78 CTCs from the same patient sample were plotted based on tSNE clustering method, demonstrating a clear separation between MET-like and EMT-like CTC populations. b Heatmap shows the significant gene expression differences between two populations in EMT, MET, and CSC related genes. Source data are provided as a Source Data file