2.2.1. Single Cell Immunoassay
Aside from carbohydrates and nucleic acids, proteins are one of the most common macromolecules in cells. Among these molecules, proteins are the most diverse molecules, playing a variety of biological roles: communication of information within and among cells, protection of cells against infection, catalysts for chemical reactions, and as structural components, to name but a few [
42]. Therefore, there is great interest in quantifying, identifying and isolating proteins, in order to understand the plethora of unknown mechanisms in which they are involved. Conventionally, the methods of Lowry and Bradford were employed to quantify total protein content [
43,
44]. However, these methods do not permit the identification of specific proteins involved in the processes of living cells. Subsequently, antibodies were utilized to identify specific proteins [
45], and Southern blot, Northern blot and Western blot analyses were developed to detect DNA, RNA, and proteins, respectively. The Western Blot was then adapted to detect single-cell proteins differentiated by molecular weight; this enabled the interrogation of more than 1000 cells in less than 4 h and multiplexed measurements of up to 11 proteins [
46]. Single-cell Western blot analysis, however, relies on separating the single-cell protein lysate using a polyacrylamide gel coating on a glass microscope slide, which can be destructive. Flow cytometry, microfluidics technologies, and surface methods such as ELISPOT were also studied as single-cell functional proteomics tools and have been extensively reviewed elsewhere [
47]. The use of functionalized nanopipettes as a platform for label-free identification of biomolecules such as proteins has been strongly recommended. Also, protein-based recognition elements, such as antibodies and enzymes, can be functionalized in the sensing zone and further used for sensing of various molecules [
48] (a summary of sensing applications is given in
Section 2 below). Functionalized nanopipettes can then be inserted into the single cell and used to monitor proteins in that cell. An antibody-labeled nanopipette shows excellent potential for the longitudinal interrogation of single cells. Implementation of this technology is on the cutting edge of advances in developing methods to combat human diseases. In addition to the proteomic approach, incorporating aspiration and sequencing of molecules from the nanopipette biopsy could identify significant disease-resistant variant genes. Therefore, the nanopipette can serve as a platform for integrated analyses in genomics, transcriptomics, proteomics and metabolomics.
2.2.3. Single Cell Aspiration
Extraction of molecules from a single cell by means of the nanobiopsy platform relies on electrowetting within a nanopipette. In brief, when an organic solution fills a nanopipette and the device is inserted into an aqueous solution, a liquid-liquid interface is created at the tip. Once voltage is applied between these two solutions, a force is produced at the interface that causes the solution to enter into or leave from the nanopipette [
2,
52]. Under this condition, when a negative potential is applied, the solution moves toward the lumen of the nanopipette, and when a positive potential is applied, the solution moves to the outside of the nanopipette. In these interrogations, the amount of aspirated material from the cell compartment was estimated to be around 50 fL, or approximately 1% of the volume of a cell [
11]. As mentioned above, we integrated the nanopipette platform into a scanning ion conductance microscope (SICM) system that automatically positions the nanopipette above the cell of interest [
27]. While in aqueous solution, the nanopipette is biased with a positive voltage to prevent the solution from flowing towards the lumen of the nanopipette. This voltage generates an ion current between the liquid-liquid interface, which can be used as the input of a feedback loop integrated with custom-built software. The software controls the movement of the nanopipette, continuing to approach the cell until a drop in the ionic current is detected, indicating the tip is at close proximity to the surface of the cell [
27]. When a reduction of the electric current is detected, the software stops movement in the direction of the cell and lowers the nanopipette at high speed (100 μm/s). This movement inserts the nanopipette into the cell membrane. The voltage applied to the nanopipette is then switched to 500 mV for 5 s, causing aspiration of cell cytoplasm into the nanopipette. Subsequently, a switch to 100 mV stops the influx, but does not induce the efflux of the aspirated content [
11]. Nanopipettes fabricated from multiple-barreled capillaries allow the simultaneous injection of dye as molecules of biological interest are aspirated from the cell. Because of the small size of the device (approx. 50 nm), injury to cells from the nanopipette is minimal. Sequential delivery of multiple dyes has demonstrated the ability of the nanopipette platform to interrogate the single cell numerous times without fatally damaging the cell.
Figure 2 shows the injection of multiple dyes into a single cell. Seger and collaborators demonstrated the ability of cells to survive for 27 h after the exposure [
27]. These injections suggest the potential application of the nanopipette platform in multiple interrogations of the single cell, without lethal damage, which can be critical for the development of single-cell drug resistance studies. Another study showed the use of nanopipettes to detect genes that were not previously described in the body of a neuron [
53] by finding the compartmentalization of mRNA molecules in different parts of neurons. For the mRNA molecule of interest to be interrogated, it must first be sequenced.
2.2.4. Nanogenomics
The nanopipette can also be employed to aspirate cell contents from the same single cell multiple times during its lifetime to study molecular dynamics. This platform was previously validated to isolate molecules such as RNA for cDNA synthesis and qPCR, and our group became one of only a few to have performed next-generation sequencing (NGS) from the species extracted with nanopipettes. Nashimoto’s research group has shown device automation in the ZYX axis for isolation of mRNA molecules [
54]. Guillaume-Gentil has demonstrated the identification of metabolites and enzymes using atomic force microscopy and also validated mRNA aspiration using qPCR [
55]. However, analytical techniques such as NMR and MS spectrometry for the detection of single-cell molecules are still limited. In 2007, Luo and Li reported on the identification of 12C/13C-dansyl labeled metabolites by means of MALDI-MS in a minimum of 100 cells [
56]. The group was able to detect subpopulations of heterogeneous tissue, but technical limitations of the method did not allow single-cell resolution. Guillaume-Gentil also reported the utilization of atomic force for aspiration and detection of mRNA molecules [
57]. Cao and collaborators demonstrated longitudinal interrogation of single cells, sampling GFP and RFP transcripts from cells [
58]. These techniques, on the one hand, relied on the observation of aspiration by fluorescence or qPCR amplification. Genes of interest, on the other hand, are not always tagged with fluorescent protein to identify protein localization. Also, not all RNA molecules involved in genetic mechanisms are expressed as proteins. However, it is not rare that all the genes of a cell must be interrogated. To successfully identify the highest possible number of genes involved in drug resistance, interrogation of cells can only be accomplished using next-generation DNA sequencing platforms.
To show that nanopipettes did not affect in the function of cells upon piercing the cell membrane, human BJ fibroblasts were treated with Ca
2+ agent Fluo4 AM, and fluorescent microscopy was used to show the localization of Ca
2+ ions before, during and after nanopipette biopsy [
11]. Optic microscopy images showed that the procedure was minimally invasive, generating only a small change of Ca
2+ during nanobiopsy. The cell recovered a few seconds after the process, reaching Ca
2+ concentrations that matched pre-aspiration levels. By contrast, Actis et al. demonstrated that micropipette aspiration caused dramatic changes in the concentration of Ca
2+ ions in the cell [
11]. The low interference of nanopipettes results from the minimal interaction of the nanopipette with the surface membrane of the cell, in contrast with the highest surface of communication and damage demonstrated by micropipettes.
It is important to note that nanopipette aspiration is based on a voltage-controlled influx of material and not adsorption of molecules to the walls of nanopipette. PCR amplification of DNA templates was not observed if negative voltage was not applied to the nanopipette during single-cell interrogation and when aspiration was performed in the bulk solution. This is the critical element that differentiates the nanobiopsy technology from AFM-based platforms. Both Wickramasinghe’s and Osada’s groups used AFM probes to extract RNA from cells in culture, either based on physisorption or hybridization of complementary RNA immobilized onto the probe [
59,
60]. We can foresee that the use of nanopipettes to aspirate limited copies of mitochondrial DNA from a living cell might provide the basis for less invasive and more accurate monitoring of disease progression. The potential of nanobiopsy is also such that the foundation can be established for the development of new classes of drugs to attenuate diseases as diverse as Parkinson’s and Alzheimer’s Disease. The nanopipette can be used as a platform for cancer research and clinical management, elucidating the role of heterogeneity in primary tumor tissues and systemically identifying critical parameters in disease progression and potential metastatic states [
61,
62]. By combining the nanopipette platform with downstream sequencing implementation, gene expression inside single cells can be longitudinally investigated, and the effect of drug mechanisms on mutation-selection can be better examined.
The nanopipette platform also allows subcellular interrogation. By using different dyes in the cellular nucleus or by staining the cytoplasm, enabling the isolation of the nucleus, it is possible to target the two compartments differentially. The following pictures in
Figure 3 show cells stained with mitochondria dye (mitotracker orange). The chromosomal region can be distinguished from the cytoplasmic by observing the white granulocytes that correspond to the interaction of mitochondrial proteins with the dye. The nucleus is depicted as circular black orifices without mitochondria. The nanopipette was inserted into the dark orifice, corresponding to the cellular nucleus.
To control the sequencing process of nanopipette aspiration downstream, we implemented the addition of External RNA Controls Consortium (ERCC) spike-in controls with samples collected from the cells. ERCC controls are a set of RNA standards for use in microarray, qPCR and sequencing applications [
63]. These molecules are artificial poly-adenylated RNA, used in library preparation protocols before cDNA synthesis. We detected increased variability as the number of detected reads decreased. However, the nanopipette biopsy was able to identify reads mapping to the human genome [
53]. Therefore, coupling the nanopipette platform with the sequencing of mRNA molecules showed the ability of the nanopipette platform to successfully identify low-abundant molecules in the context of gene expression, a capability essential for single-cell interrogation. ERCC spike-ins were used to show the ability of nanopipettes to isolate cellular RNA molecules for sequencing. Reads were used that mapped to at least one spot in the human genome, as described by Actis et al. [
11] and Toth et al. [
53].
Figure 4 shows the number of reads (counts) mapped to the ERCC reference as a function of ERCC concentration.
After separation of the ERCC counts from reads proceeding from cellular content, reads mapped to the human reference genome were plotted as Principal Component Analysis (PCA) results of cellular expression, showing the clustering pattern of the nuclear aspirations of single cells. The PCA of gene expression in the nuclear nanobiopsy samples, using both non-processed and pre-processed gene counts, are shown in
Figure 5. The sequencing reads were aligned against the human reference genome using the STAR aligner, and the HTSeq package was used to count the number of mapped reads. Using the limit of detection (10 reads per detected transcript), reads were input to DESeq2 (HL = HeLa transcriptome library; MBL = MDA-MB-231 transcriptome library; NL = iCell neuron library; MCL = MCF-7 transcriptome library).
Figure 5A shows the PCA of gene expression in the nanobiopsy samples. Libraries MBL1, MBL9, MBL12, and MBL14 were considered outliers and removed from downstream analysis. The
Figure 5B graphs are plotted from PCA runs with reads log-transformation, with the aim of mitigating the variation effect of highly expressed genes or any biases possibly introduced during the cell nanobiopsy procedure, library preparation or sequencing run. In
Figure 5C we focused on the areas from
Figure 5A that have more clustering structure for better visualization. Furthermore,
Figure 5D was plotted to give a closer look to the clustered areas from
Figure 5C.
Figure 5C,D illustrate the expression profile of the four cell types are different from one another.
It was not clear to what extent the MDA-MB-231 cells and MCF-7 cells were distinguishable using PCA in
Figure 5. Therefore, we plotted the MDA-MB-231 cells and MCF-7 separately in (
Figure 6A–C).
Figure 6B represents focused areas of
Figure 6A,
Figure 6C represents focused areas of
Figure 6B, for more clustering structure.
Figure 6E represents focused areas of
Figure 6D for more clustering structure. These results suggest that, although the number of detected reads is small per sequenced libraries, nanopipette technology detects the similarities of same-cell type.
Figure 6D–F support the conclusion of MDA-MB-231 vs. MCF-7 comparison.
Figure 6E represents focused areas of
Figure 6D,
Figure 6F represents focused areas of
Figure 6E, with more clustering structure.
To determine the identity of more abundant genes in the MDA-MB-231 and MCF-7 cells, we extracted RefSeq IDs with more than 200 reads in at least one of the 39 sequenced libraries from the dataset, and checked the presence of the genes in both MDA-MB-231 and MCF-7 cells.
Table 1 represents the ability of nanobiopsy to resolve the identity of a cell type by detecting highly abundant transcripts associated with ubiquitous biological processes, as an example, genes associated with glucose metabolism (UGP2, ENO1), ribosomal protein synthesis (RPLP0, EEF1A1, NPM1), protein folding (HSP90AA1), protein degradation (POMP), DNA binding (H3F3B), and drug resistance by cancer cells (AXL) [
64,
65,
66,
67,
68,
69]. More specifically, genes ENO1, H3F3B and HSP90AA1 are important cancer drivers in human cells.