Label-free capture of breast cancer cells spiked in buffy coats using carbon nanotube antibody micro-arrays

Abstract

We demonstrate the rapid and label-free capture of breast cancer cells spiked in buffy coats using nanotube-antibody micro-arrays. Single wall carbon nanotube arrays were manufactured using photo-lithography, metal deposition, and etching techniques. Anti-epithelial cell adhesion molecule (EpCAM) antibodies were functionalized to the surface of the nanotube devices using 1-pyrene-butanoic acid succinimidyl ester functionalization method. Following functionalization, plain buffy coat and MCF7 cell spiked buffy coats were adsorbed on to the nanotube device and electrical signatures were recorded for differences in interaction between samples. A statistical classifier for the 'liquid biopsy' was developed to create a predictive model based on dynamic time warping to classify device electrical signals that corresponded to plain (control) or spiked buffy coats (case). In training test, the device electrical signals originating from buffy versus spiked buffy samples were classified with ∼100% sensitivity, ∼91% specificity and ∼96% accuracy. In the blinded test, the signals were classified with ∼91% sensitivity, ∼82% specificity and ∼86% accuracy. A heatmap was generated to visually capture the relationship between electrical signatures and the sample condition. Confocal microscopic analysis of devices that were classified as spiked buffy coats based on their electrical signatures confirmed the presence of cancer cells, their attachment to the device and overexpression of EpCAM receptors. The cell numbers were counted to be ∼1-17 cells per 5 μl per device suggesting single cell sensitivity in spiked buffy coats that is scalable to higher volumes using the micro-arrays.

Figure 1

Schematic of sensing mechanism. Cooperative binding of anti-EPCAM antibodies to their corresponding receptors in cells on top of nanotube biosensors creates ‘spikes’ in electrical conductance.

Figure 2

Device characterization. (a) Optical image of the wafer with 60 devices manufactured using photolithography, metal deposition and etching; (b) scanning electron micrograph of carbon nanotubes; (c) Raman spectrum of semiconducting carbon nanotubes showing a large G band, small D band and 2D band. The I_G/I_D ~30 was observed in these nanotubes; (d) Test set up showing six devices on a chip and compares it to a penny. The source, drain and reference electrodes are observed.

Figure 3

Electrical characteristics of nanotube devices. (a) SEM image of the actual device fabricated from 4, 2, and 1 µg of carbon nanotube film and their corresponding high magnification; (b) film resistance before and after annealing at 250 °C; (c) histogram suggesting high degree of control in device resistance over 58 devices.

Figure 4

Understanding semiconducting nanotube network. (a) Real time response of device to concentrations of NH⁴⁺, 300 nM to 1.3 mM, and Hg²⁺, 30 pM to 13 µM, ions; (b) voltage sweep before and after exposure to NH⁴⁺ and Hg²⁺ ions; (c) concentration sensitivity of voltage sweep for NH⁴⁺ and Hg²⁺ ions; (d) normalized signal conductance versus concentration of Hg²⁺ ions suggesting Langmuir-adsorption isotherm. The inset shows the percentage sensitivity versus concentration for Hg²⁺. Similar results were seen for NH⁴⁺ ions suggesting same sensing mechanism.

Figure 5

Functionalization of antibodies. (a) Schematic of the PASE functionalization protocol; (b) SEM image of positive control suggesting high degree of antibody functionalization; (c) SEM image of negative control suggesting no functionalization; (d) comparison of different functionalization protocols including PASE–antibody, streptavidin–biotin and amine-polymer-NP conjugation.

Figure 6

Testing in cell cultures. (a) Normalized device conductance of anti-EpCAM functionalized device with SKBR3 cancer cells; (b) normalized device conductance of anti-EpCAM functionalized device with MCF7 cancer cells; (c) normalized device conductance of anti-EpCAM functionalized device with MCF10 A normal cells; (d) representative graph of normalized device conductance of anti-IgG functionalized device with SKBR3, MCF7 and MCF10A cells.

Figure 7

Merge of sensor arrays with statistical classifiers in training set: Normalized conductance versus time data from the arrays grouped into two categories. (a) Plain buffy coat; (b) buffy coat spiked with MCF7 cells.

Figure 8

Merge of sensor arrays with statistical classifiers in blinded set: Normalized conductance versus time data from the arrays grouped into two categories. (a) Plain buffy coat; (b) buffy coat spiked with MCF7 cells.

Figure 9

Heat map. Summary of the relationship between electrical signatures and the cellular-proteomic features namely overexpression of EpCAM in spiked buffy coats versus buffy coats. The statistical classifier naturally partitions the buffy versus spiked buffy coats suggests specific interactions are quite unique in their electrical signatures compared to non-specific interactions and establishes a relationship between electrical conductance data with proteomic features.

Figure 10

Cell capture using nanotube devices based on optical microscopy. (a) Optical microscopy of spiked cells in buffy coats using nanotube devices; (b) optical microscopy of plain buffy coats.

Figure 11

Cell capture using nanotube devices using confocal microscope. Representative confocal images from six devices imaged shown) of captured cells in spiked buffy coats ranging from 1 to 17 cells per 5 µl sample.