Multidisciplinary single-cell optical cytometry
Xuantao Su
DOI: https://doi.org/10.1002/cyto.a.24513
IF: 4.714
2021-01-01
Cytometry Part A
Abstract:Cytometry is interdisciplinary in that it measures the physical or chemical signals of cells or particles flowing in a fluidic stream.1 The first generation of commercial flow cytometry performs impedance measurements for particle counting. With the invention of laser technology, light scattering from single cells or particles in flow can be measured, making the optical flow cytometry attractive. Compared with label-free light scattering measurements, the first fluorescence flow cytometer was developed in 1969. However, it is not until the groundbreaking development of engineered green fluorescence probe2 and the many other fluorescence labels that optical flow cytometry was widely used in the field of biomedicine. Recently, we have witnessed the burgeoning multidisciplinary single-cell cytometry of mechanical cytometry,3 photoacoustic cytometry,4 and mass cytometry,5 with more on the novel developments of optical cytometric technologies.6, 7 In the June issue of Cytometry Part A in 2021, “Quantitative Single-Cell Optical Technologies”,8 we embraced the studies of single cells with various optical technologies such as flow cytometry and super-resolution microscopy. For example, Wei et al. have further developed the in-vivo flow cytometry by adopting 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) amino]-2-deoxy-d-glucose (2-NBDG) to label tumor cells, and found the CTCs of mice in-vivo have a high level of 2-NBDG.9 Su and colleagues have shown that the deep learning technology is potentially important for intelligent diagnosis of cervical cancer by developing a light scattering pattern-specific convolution net cytometer (LSPS-net Cytometer).10 This issue will focus on the interdisciplinary fields of nanotechnology, deep learning and microfluidics, which echoes the intrinsic multidisciplinary property of cytometry. Nanotechnology, also called nanoscience, studies matter with size ranging from 1 to 100 nm by multidisciplinary approaches. The amalgamation of nanotechnology and cytometry heavily resides in the labeling of the cells or particles, in order to help for the cytometric measurements with a higher sensitivity or resolution. Due to the quantum effects, different sizes of semiconductor nanoparticles (quantum dots) emit variant colors, which can be used for biological labelling. Compared with fluorescence labelling, the quantum dots are not limited to blinking and provide multi-color labels with higher emitting intensity; however, they may suffer from the fabrication complexity, decoration problems, and high toxicity.11 Gold nanoparticles have been widely used as versatile bio-labels, thanks to their strong light scattering and surface plasmon resonance. The nanoparticles have been shown in cytometry for many interesting works such as three-dimensional intracellular visualization.12 In flow cytometry, the cells are flowing in a single cell profile by hydrodynamic focusing, where the diameter of the sample fluidic stream can be narrowed to the size of a cell. Thus, conventional flow cytometry explores the microfluidics. Moreover, it is known to all that microfluidics controls the fluidics at micro-scale or nano-scale. In flow cytometry, the samples measured are usually in the scale of microliters or nanoliters, which requires the technology of microfluidics. The microfluidics also couple well with optics and electronics. In this case, the microfluidic technology may help to advance the core technology of flow cytometry and make flow cytometry compact, portable, and high-efficiency.13, 14 Deep learning, which is inspired by the neuron network of the human brain for studying, is an important step forward in machine learning toward the artificial intelligence, especially in biomedicine. The deep learning technology incorporates well with flow cytometry for two main reasons. Firstly, big data from a large number of cells in a due time can be obtained by high-throughput flow cytometry, for example, several thousand cell images per second by an imaging flow cytometer. This requires robust and automatic algorithms for big data analysis. Secondly, the development of novel methods to extract more information of cells such as label-free flow cytometry relies on data mining that may be better achieved by deep learning.10, 15, 16 The integration of nanotechnology in cytometry is presented in this issue with nanoparticles and fluorescence resonance energy transfer. Plasmonic nanoparticle (PNP) may provide high sensitivity for disease diagnosis. Evans et al. have reviewed the recent development of PNP-based technologies, and described their PNP-based “digital” cytometry methodology for neurological disorder characterization (Evans et al., pp. 1067–1078). Toward the development of efficient biomarkers, biofunctionalized nanospheres are combined with imaging flow cytometry to measure immune cell signaling in subcellular regions of interest (Thaunat et al., pp. 1079–1090). Fluorescence resonance energy transfer (FRET), the energy transfer between two adjacent molecules (less than 10 nm), has been shown with great applications in biomedicine.2 Davis et al. have reported the measurements of FRET in living cells by using both conventional flow cytometry and the spectra cytometry17. With quantitative FRET imaging, the inhibitory priority of Bcl-xL to Bad, tBid and Bax has been evaluated by using live-cell imaging assay (Chen et al., pp. 1091–1101). However, concerns have been raised when performing FRET measurements with flow cytometry (Lambert et al., pp. 1102–1106). The incorporation of microfluidics with optics is demonstrated by developing a microfluidic cytometer for white blood cell counting (Li et al., pp. 1107–1113). For their microfluidic cytometer, polydimethylsiloxane is used for chip fabrication and optical fibers and microlens are integrated with the microfluidics, where the small angle forward scattering, side scattering and fluorescence signals of single cells controlled by on-chip hydrodynamic focusing are measured. Combining impedance and optics in flow cytometry, Yue et al. have reported that specific membrane capacitance and cytoplasmic conductivity as intrinsic bioelectrical markers can be measured to classify tumor subtypes by their constriction channel (with a width and a height that were smaller than the cell diameter) based impedance cytometry (Yue et al., pp. 1114–1122). Deep learning is adopted to cytometry for effective analysis of the big cytometric imaging data and for label-free cell classification in the current issue. Liu et al. have combined flow cytometry with an artificial neural network (MCellNet) to classify the species that have similar morphologies with high accuracy (Liu et al., pp. 1123–1133). In order to measure cell viability, the 2D light scattering method is combined with the deep learning algorithm for label-free classification of live and dead colonic adenocarcinoma cells (Yang et al., pp. 1134–1142). Key Research and Development Program of Shandong Province (Major Science and Technology Innovation Project), Grant number: 2019JZZY011016; National Natural Science Foundation of China (NSFC), Grant number: 91859114. The peer review history for this article is available at https://publons.com/publon/10.1002/cyto.a.24513.