Identifying EGFR-Expressed Cells and Detecting EGFR Multi-Mutations at Single-Cell Level by Microfluidic Chip

Abstract

EGFR mutations companion diagnostics have been proved to be crucial for the efficacy of tyrosine kinase inhibitor targeted cancer therapies. To uncover multiple mutations occurred in minority of EGFR-mutated cells, which may be covered by the noises from majority of un-mutated cells, is currently becoming an urgent clinical requirement. Here we present the validation of a microfluidic-chip-based method for detecting EGFR multi-mutations at single-cell level. By trapping and immunofluorescently imaging single cells in specifically designed silicon microwells, the EGFR-expressed cells were easily identified. By in situ lysing single cells, the cell lysates of EGFR-expressed cells were retrieved without cross-contamination. Benefited from excluding the noise from cells without EGFR expression, the simple and cost-effective Sanger's sequencing, but not the expensive deep sequencing of the whole cell population, was used to discover multi-mutations. We verified the new method with precisely discovering three most important EGFR drug-related mutations from a sample in which EGFR-mutated cells only account for a small percentage of whole cell population. The microfluidic chip is capable of discovering not only the existence of specific EGFR multi-mutations, but also other valuable single-cell-level information: on which specific cells the mutations occurred, or whether different mutations coexist on the same cells. This microfluidic chip constitutes a promising method to promote simple and cost-effective Sanger's sequencing to be a routine test before performing targeted cancer therapy.

Keywords: EGFR mutation; Microfluidic chip; Single-cell analysis; Tyrosine kinase inhibitor.

Fig. 1

The schematic view of microfluidic chip and its operation. a The cell mixture is pumped into a microfluidic chip which consists of three layers: the microfluidic channel, the cell trapping array and the cell lysate collecting chamber. The flow velocity was 3 μL min⁻¹. b Cells are trapped in microwells. The square microwell is designed to fit only one cell. c The chip was fluorescently examined to identify cells with specific protein expression. d Blue balls represent negative cells which exhibit only blue color of DAPI, while green balls represent positive cells which exhibit both green color of FITC and blue color of DAPI. e The cells in microwells are lysed by inputting cell lysis solution through the microfluidic channel. The cell lysate is directed to cell lysate collecting chambers. f The cell lysates are separately collected from cell lysate collecting chambers for the following DNA amplification and sequencing. To clearly demonstrate the chip structure, the schematic figures do not follow the exact well number and dimension

Fig. 2

The fabrication of microfluidic chip. a The fabrication process of the microfluidic chip which consists of three layers. They are indicated by a number 1, 2, and 3, respectively. b SEM images of empty silicon microwells (left column) and microwells occupied by single cells (right column). Scale bar: 20 µm

Fig. 3

The optimization of cell capture efficiency. The cell capture efficiency was defined as the ratio between captured cells and all cells pumped into the microfluidic chip. a The relationship between cell capture efficiency and side length of square microwells. Other operation parameters are: 3.2 × 10⁵ Cells mL⁻¹ and 3 min trapping time. b The relationship between cell capture efficiency and cell trapping time. Other operation parameters are: 25 µm in side lengths, and 3.2 × 10⁵ Cells mL⁻¹. c The relationship between cell capture efficiency and cell density. Other operation parameters are: 25 µm side lengths and 3 min trapping time. In a, b and c, data are expressed as the mean ± SD from 3 independent assays. d Trapped single cells were stained by DAPI, which are blue spots. Scale bar: 100 µm

Fig. 4

Fluorescently identifying, amplifying and sequencing MCF-7 cells. a The fluorescent images of cells in microwells. Upper row is an area contains only MCF-7 cells which exhibit both DAPI (blue) and FITC (green) staining; Middle row contains only HEK-293T cells which exhibit only DAPI (blue) staining; Lower row contains both HEK-293T cells and MCF-7 cells. Scale bar: 50 µm. b DNA amplification products of STR domain were verified by fluorescence image of agarose gel. Left two columns are the empty run and PBS reaction, both as negative control; The third and fourth left columns are the products from pure HEK-293T and MCF-7 cells, without the chip processing, both as positive control. Three right columns are products from chip areas contain no cells, only HEK-293T cells and only MCF-7 cells, sequentially. c The sequencing results for STR domain from the cells lysates from areas, respectively, contain only HEK-293T and MCF-7 cells. (Color figure online)

Fig. 5

Detecting EGFR multi-mutations. a The fluorescent images of cells in microwells. Upper row is an area contains only EGFR-expressed cells which exhibit both DAPI (blue) and FITC (green) staining; middle row contains only HEK-293T cells which exhibit only DAPI (blue) staining; lower row contains both HEK-293T cells and EGFR-expressed cells. Scale bar: 100 µm. b The Sanger’s sequencing provides accurate mutation information: In exon 19, NCI-H1650 cells show a deletion mutation (Del E746-A750); in exon 20, NCI-H1975 cells show a point mutation (T790M); in exon 21, NCI-H1975 cells show a point mutation (L858R). c For NCI-H1975 cells, mutations in both exon 20 and 21 are detected; for NCI-H1650 cells, mutation in exon 19 is also detected. In addition, no false-positive result on either HEK-293T or A549 cells is detected. (Color figure online)