The new method for three dimensional imaging of cells without fluorescence staining. Credit: Alain Herzog
Fluorescent dyes, or stains, are frequently used to label intracellular components such as nuclei when analyzing and screening cells using methods like confocal microscopy and flow cytometry. However, fluorescent dyes can be costly, staining methods are often time-consuming and the labeling process can sometimes cause damage to the cells being investigated. Researchers from Ecole Polytechnique Federale de Lausanne (EPFL), the Consiglio Nazionale delle Richerche (CNR), the University Federico II and CEINGE-Biotecnologie Avanzate have now developed a new stain-free 3D imaging method that not only allows intracellular components to be observed, but allows the nuclei of single cells to be identified while in motion during flow cytometry, opening up new possibilities for diagnostic and drug screening methods.
The technique relies on the use of quantitative phase imaging, a holographic imaging approach that reveals the phase delay incurred as the light beam of a microscope passes through matter. When phase data is collected from multiple different angles and processed computationally, a 3D map of the refractive index of each voxel in the 3D image can be generated through a process called learning tomography. However, previous applications of quantitative phase imaging for bioimaging lacked the specificity to reveal individual intracellular components. The researchers sought to overcome this limitation by using a new artificial intelligence method called computational segmentation based on statistical inference (CSSI), which uses a deep convolutional neural network to better group voxels with similar refractive indices, and assemble these clusters into shapes that can be further classified, according to co-corresponding author Demetri Psaltis. This way, intracellular components with different refractive indices, such as different types of nuclei, can be resolved.
The team showed how the method could be used in high-throughput single-cell screening applications by using quantitative phase imaging with tomographic phase microscopy and flow cytometry. By using a flow velocity gradient to rotate the cells in fluidic channels 50-100 μm across, the cells could be observed using a stationary beam and detector to detect the phase delay, estimate the orientation of the cell and generate the 3D refractive index maps via the learning tomography algorithm, explained co-corresponding author Pietro Ferraro. The method achieved a transverse resolution of half a micron to one micron, allowing the researchers to observe structures as small as protein aggregates as well as assess the size of the nucleus and outlines of single cells, said Psaltis. The method was validated by comparing results to those obtained using confocal fluorescence microscopy. This research was published in Nature Photonics.
This stain-free technique for observing intracellular components in flow cytometry could be used in applications such as diagnostic testing and drug development. The method could allow for the detection of circulating cancer cells through liquid biopsies, both to identify cancer types and as an early diagnostic tool for cancer metastasis. It could also be used to screen drug candidates, such as those designed to break down cross-linked proteins that are associated with diseases like Parkinson’s; by running treated cells repeatedly through the imaging setup, the breakdown of cross-linked proteins could be observed in real time.