Single-cell Imaging Research Articles

Artificial intelligence-assisted diagnosis of hematologic diseases based on bone marrow smears using deep neural networks

Background and objectivesThe morphological examination of bone marrow (BM) cells is essential in both diagnosing and treating various hematologic diseases. However, it is still done manually with a heavy workload. An artificial intelligence-assisted diagnosis support system of BM cells is highly required to reduce the workloads of examiners and improve the reproducibility of the results. MethodsIn this paper, we proposed an artificial intelligence-assisted diagnosis support system of morphological examination based on bone marrow smears including cells detection, classification and prediction of leukemia types. For cell detection, we trained the novel YOLOX-s model to locate cells precisely and obtain single cell images. For cell classification, we regarded it as a fine- grained classification task and proposed a novel architecture called MLFL-Net utilizing multi-level features. Furthermore, we predicted the leukemia types on a dataset including 40 normal people (BM transplantation donors) and 40 patients of different kinds of acute leukemia according to the World Health Organization (WHO) standard. ResultsWe constructed a large-scale data set of 11,788 fully-annotated micrographs from 728 smears and 131,300 expert-annotated single cell images. With the data set, the detection model achieved 0.9797 AUC and 4.33% box placement error. For cell classification, the total accuracy of our proposed MLFL-Net reached 89.53% which outperformed all the other related models in identifying cell categories. In the meantime, we took acute leukemia as an example to explore the leukemia types prediction procedure of hematological disease. It generated the same diagnostic prediction as the experts gave for 92.5 percent of the cohort. ConclusionThis Artificial Intelligence-assisted system can be implemented to aid in clinical decision making and accelerate diagnosis. The method will contribute to promote the intelligence and modernization of BM cytomorphology, which has vital significance of the development of the medical career.

Integrated intracellular organization and its variations in human iPS cells

Understanding how a subset of expressed genes dictates cellular phenotype is a considerable challenge owing to the large numbers of molecules involved, their combinatorics and the plethora of cellular behaviours that they determine1,2. Here we reduced this complexity by focusing on cellular organization—a key readout and driver of cell behaviour3,4—at the level of major cellular structures that represent distinct organelles and functional machines, and generated the WTC-11 hiPSC Single-Cell Image Dataset v1, which contains more than 200,000 live cells in 3D, spanning 25 key cellular structures. The scale and quality of this dataset permitted the creation of a generalizable analysis framework to convert raw image data of cells and their structures into dimensionally reduced, quantitative measurements that can be interpreted by humans, and to facilitate data exploration. This framework embraces the vast cell-to-cell variability that is observed within a normal population, facilitates the integration of cell-by-cell structural data and allows quantitative analyses of distinct, separable aspects of organization within and across different cell populations. We found that the integrated intracellular organization of interphase cells was robust to the wide range of variation in cell shape in the population; that the average locations of some structures became polarized in cells at the edges of colonies while maintaining the ‘wiring’ of their interactions with other structures; and that, by contrast, changes in the location of structures during early mitotic reorganization were accompanied by changes in their wiring.

Cathodic Electrochemiluminesence Microscopy for Imaging of Single Carbon Nanotube and Nucleolin at Single Tumor Cell.

Cathodic electrochemiluminesence (ECL) microscopy based on luminol analog L012 was originally established to implement the imaging of a single nanotube and nucleolin on a single tumor cell. This microscopy utilizes multiwalled carbon nanotubes (MWCNTs) as advanced coreactant accelerators to efficiently convert dissolved oxygen (O2) and H2O2 into reactive oxygen species (ROS) due to excellent electrocatalytic properties. The produced ROS could oxide L012 into an excited state of L012 leading to a bright cathodic ECL illumination, thereby promoting ECL imaging of MWCNTs at a low triggering potential. After being modified with AS1411 aptamers, MWCNTs@AS1411 probes were incubated with tumor cells for specific ECL imaging of nucleolin on the plasma membrane, which permits cathodic ECL microscopy for label-free bioassays without ECL tags. The L012-based cathodic ECL microscopy with a moderate operating potential and label-free characteristics provides a universal approach in single nanomaterial and single-cell imaging and analyses.

Antibiotic perseverance increases the risk of resistance development

The rise of antibiotic-resistant bacterial infections poses a global threat. Antibiotic resistance development is generally studied in batch cultures which conceals the heterogeneity in cellular responses. Using single-cell imaging, we studied the growth response of Escherichia coli to sub-inhibitory and inhibitory concentrations of nine antibiotics. We found that the heterogeneity in growth increases more than what is expected from growth rate reduction for three out of the nine antibiotics tested. For two antibiotics (rifampicin and nitrofurantoin), we found that sub-populations were able to maintain growth at lethal antibiotic concentrations for up to 10 generations. This perseverance of growth increased the population size and led to an up to 40-fold increase in the frequency of antibiotic resistance mutations in gram-negative and gram-positive species. We conclude that antibiotic perseverance is a common phenomenon that has the potential to impact antibiotic resistance development across pathogenic bacteria.

Optical photothermal infrared spectroscopy and discrete wavenumber imaging for high content screening of single cells.

Major drawbacks of direct mid-infrared spectroscopic imaging of single cells in an aqueous buffer are strong water absorption, low resolution typically above 10 μm, and Mie scattering effects. This study demonstrates how an indirect detection principle can overcome these drawbacks using the optical photothermal infrared (O-PTIR) technique for high-resolution discrete wavenumber imaging and fingerprint spectroscopy of cultivated cells as a model system in a simple liquid sample chamber. The O-PTIR spectra of six leukemia- and cancer-derived cell lines showed main IR bands near 1648, 1547, 1447, 1400, 1220, and 1088 cm-1. Five spectra of approximately 260 single cells per cell type were averaged, the O-PTIR data set was divided into leukemia-derived cells (THP-1, HL 60, Jurkat, and Raji) and cancer cells (HeLa and HepaRG), and partial least squares linear discriminant analysis (PLS-LDA) was applied in the spectral range 800-1800 cm-1 to train three classification models. A leukemia versus cancer cell model showed an accuracy of 90.0%, the HeLa versus HepaRG cell model had an accuracy of 95.4%, and the model for the distinction of leukemia cells had an accuracy of 75.4%. IR bands in linear discriminants (LDs) of the models were correlated with second derivative spectra that resolved more than 25 subbands. The IR and second derivative spectra of proteins, DNA, RNA and lipids were collected as references to confirm band assignments. O-PTIR images of single cells at a 200 nm step size were acquired at 1086, 1548, and 1746 cm-1 to visualize the nucleic acid, protein, and lipid distribution, respectively. Variations in subcellular features and in the lipid-to-protein and nucleic acid-to-protein ratios were identified that were consistent with biomolecular information in LDs. In conclusion, O-PTIR can provide high-quality spectra and images with submicron resolution of single cells in aqueous buffers that offer prospects in high-content screening applications.

Enhanced electrochemiluminescence imaging of single cell membrane proteins based on Co3O4 nanozyme catalysis.

The development of enhanced strategies with excellent biocompatibility is critical for electrochemiluminescence (ECL) imaging of single cells. Here, we report an ECL imaging technique for a single cell membrane protein based on a Co3O4 nanozyme catalytic enhancement strategy. Due to the remarkable catalytic performance of Co3O4 nanozymes, H2O2 can be efficiently decomposed into reactive oxygen radicals, and the reaction with L012 was enhanced, resulting in stronger ECL emission. The anti-carcinoembryonic antigen (CEA) was coupled with nanozyme particles to construct a probe that specifically recognized the overexpressed CEA on the MCF-7 cell membrane. According to the locally enhanced visualized luminescence, the rapid ECL imaging of a single cell membrane protein was eventually realized. Accordingly, Co3O4 nanozymes with highly efficient activity will provide new insights into ECL imaging analysis of more biological small molecules and proteins.

Broadband CARS high-throughput single-cell imaging

Broadband Coherent anti-Stokes Raman Scattering (BCARS) enables the whole vibrational spectrum of cytologically prepared samples to be obtained using a hyperspectral raster scan approach. This technique has the potential to enable high-throughput automated detection of cell abnormalities. Images are distorted by the non-resonant background which requires a treatment for proper analysis. Using statistical denoising and phase retrieval returns Raman spectra similar to that of a spontaneous Raman measurement. Here, we present our work using this method for single-cell imaging of PEO1 ovarian adenocarcinoma cells prepared with the ThinPrep processor which enables label-free Raman cytology.

Tumor microenvironment-responsive gold nanodendrites for nanoprobe-based single-cell Raman imaging and tumor-targeted chemo-photothermal therapy

Nanodendrite particles (NDs) with densely branched structures and biomimetic architectures have exhibited great promise in tumor therapy owing to their prolonged in vivo circulation time and exceptional photothermal efficiency. Nevertheless, traditional NDs are deficient in terms of specific surface modification and targeting tumors, which restrict their potential for broader clinical applications. Here, we developed coronavirus-like gold NDs through a seed-mediated approach and using silk fibroin (SF) as a capping agent. Our results demonstrate that these NDs have a favorable drug-loading capacity (∼65.25%) and light-triggered release characteristics of doxorubicin hydrochloride (DOX). Additionally, NDs functionalized with specific probes exhibited exceptional surface-enhanced Raman scattering (SERS) characteristics, enabling high-sensitivity Raman imaging of unstained single cells. Moreover, these NDs allowed for real-time monitoring of endocytic NDs for over 24 ​h. Furthermore, ND@DOX conjugated with tumor-targeting peptides exhibited mild hyperthermia, minimal cytotoxicity, and effective targeting towards cancer cells in vitro, as well as responsiveness to the tumor microenvironment (TME) in vivo. These unique properties led to the highest level of synergistic tumor-killing efficiency when stimulated by a near-infrared (NIR) laser at 808 ​nm. Therefore, our virus-like ND functionalized with SF presents a novel type of nanocarrier that exhibits significant potential for synergistic applications in precision medicine.

In Vivo Voltammetric Imaging of Metal Nanoparticle-Catalyzed Single-Cell Electron Transfer by Fermi Level-Responsive Graphene.

Metal nanomaterials can facilitate microbial extracellular electron transfer (EET) in the electrochemically active biofilm. However, the role of nanomaterials/bacteria interaction in this process is still unclear. Here, we reported the single-cell voltammetric imaging of Shewanella oneidensis MR-1 at the single-cell level to elucidate the metal-enhanced EET mechanism in vivo by the Fermi level-responsive graphene electrode. Quantified oxidation currents of ~20 fA were observed from single native cells and gold nanoparticle (AuNP)-coated cells in linear sweep voltammetry analysis. On the contrary, the oxidation potential was reduced by up to 100 mV after AuNP modification. It revealed the mechanism of AuNP-catalyzed direct EET decreasing the oxidation barrier between the outer membrane cytochromes and the electrode. Our method offered a promising strategy to understand the nanomaterials/bacteria interaction and guide the rational construction of EET-related microbial fuel cells.

Site-selective heat boosting electrochemiluminescence for single cell imaging.

In operando visualization of local electrochemical reactions provides mechanical insights into the dynamic transport of interfacial charge and reactant/product. Electrochemiluminescence is a crossover technique that quantitatively determines Faraday current and mass transport in a straightforward manner. However, the sensitivity is hindered by the low collision efficiency of radicals and side reactions at high voltage. Here, we report a site-selective heat boosting electrochemiluminescence microscopy. By generating a micron-scale heat point in situ at the electrode-solution interface, we achieved an enhancement of luminescence intensity up to 63 times, along with an advance of 0.2 V in applied voltage. Experimental results and finite element simulation demonstrate that the fundamental reasons are accelerated reaction rate and thermal convection via a photothermal effect. The concentrated electrochemiluminescence not only boosts the contrast of single cells by 20.54 times but also enables the site-selective cell-by-cell analysis of the heterogeneous membrane protein abundance. This electrochemical visualization method has great potential in the highly sensitive and selective analysis of local electron transfer events.

Reversibly Tuning Electrochemiluminescence with Stimulated Emission Route for Single-Cell Imaging

Electrochemiluminescence (ECL) has established itself as an excellent transduction technique in biosensing and light-emitting device, while conventional ECL mechanism depending on spontaneous emission of luminophores lacks reversibility and tunable emission characters, limiting the universality of ECL technique in the fields of fundamental research and clinical applications. Here, we report the first observation of stimulated emission route in ECL and thus establish a reversible tuning ECL microscopy for single-cell imaging. This microscopy uses a focused red-shifted beam to transfer spontaneous ECL into stimulated ECL, which enables selective and reversible tuning of ECL emission from homogeneous solution, single particles, and single cells. After excluding other possible competitive routes, the stimulated ECL emission route is confirmed by a dual-objective system in which the suppressed spontaneous ECL is accompanied by the enhanced stimulated ECL. By incorporating a commercial donut-shaped beam, the sharpness of single-cell matrix adhesion is improved 2 to 3 times compared with the counterpart in confocal ECL mode. The successful establishment of this stimulated emission ECL will greatly advance the development of light-emitting device and super-resolution ECL microscopy.

Multiplexed single-cell imaging to characterize melanoma brain metastases under therapeutic pressure

Multiplexed single-cell imaging to characterize melanoma brain metastases under therapeutic pressure

Fission yeast obeys a linear size law under nutrient titration.

Steady-state cell size and geometry depend on growth conditions. Here, we use an experimental setup based on continuous culture and single-cell imaging to study how cell volume, length, width and surface-to-volume ratio vary across a range of growth conditions including nitrogen and carbon titration, the choice of nitrogen source, and translation inhibition. Overall, we find cell geometry is not fully determined by growth rate and depends on the specific mode of growth rate modulation. However, under nitrogen and carbon titrations, we observe that the cell volume and the growth rate follow the same linear scaling.

CellVisioner: A Generalizable Cell Virtual Staining Toolbox based on Few-Shot Transfer Learning for Mechanobiological Analysis.

Visualizing cellular structures especially the cytoskeleton and the nucleus is crucial for understanding mechanobiology, but traditional fluorescence staining has inherent limitations such as phototoxicity and photobleaching. Virtual staining techniques provide an alternative approach to addressing these issues but often require substantial amount of user training data. In this study, we develop a generalizable cell virtual staining toolbox (termed CellVisioner) based on few-shot transfer learning that requires substantially reduced user training data. CellVisioner can virtually stain F-actin and nuclei for various types of cells and extract single-cell parameters relevant to mechanobiology research. Taking the label-free single-cell images as input, CellVisioner can predict cell mechanobiological status (e.g., Yes-associated protein nuclear/cytoplasmic ratio) and perform long-term monitoring for living cells. We envision that CellVisioner would be a powerful tool to facilitate on-site mechanobiological research.

Dispersive phase microscopy incorporated with droplet-based microfluidics for biofactory-on-a-chip.

Biomolecular imaging of intracellular structures of a single cell and subsequent screening of the cells are of high demand in metabolic engineering to develop strains with the desired phenotype. However, the capability of current methods is limited to population-scale identification of cell phenotyping. To address this challenge, we propose to utilize dispersive phase microscopy incorporated with a droplet-based microfluidic system that combines droplet volume-on-demand generation, biomolecular imaging, and droplet-on-demand sorting, to achieve high-throughput screening of cells with an identified phenotype. Particularly, cells are encapsulated in homogeneous environments with microfluidic droplet formation, and the biomolecule-induced dispersive phase can be investigated to indicate the biomass of a specific metabolite in a single cell. The retrieved biomass information consequently guides the on-chip droplet sorting unit to screen cells with the desired phenotype. To demonstrate the proof of concept, we showcase the method by promoting the evolution of the Haematococcus lacustris strain toward a high production of natural antioxidant astaxanthin. The validation of the proposed system with on-chip single-cell imaging and droplet manipulation functionalities reveals the high-throughput single-cell phenotyping and selection potential that applies to many other biofactory scenarios, such as biofuel production, critical quality attribute control in cell therapy, etc.

Real-time fluorescence imaging flow cytometry enabled by motion deblurring and deep learning algorithms.

Fluorescence imaging flow cytometry (IFC) has been demonstrated as a crucial biomedical technique for analyzing specific cell subpopulations from heterogeneous cellular populations. However, the high-speed flow of fluorescent cells leads to motion blur in cell images, making it challenging to identify cell types from the raw images. In this study, we present a real-time single-cell imaging and classification system based on a fluorescence microscope and deep learning algorithm, which is able to directly identify cell types from motion-blur images. To obtain annotated datasets of blurred images for deep learning model training, we developed a motion deblurring algorithm for the reconstruction of blur-free images. To demonstrate the ability of this system, deblurred images of HeLa cells with various fluorescent labels and HeLa cells at different cell cycle stages were acquired. The trained ResNet achieved a high accuracy of 96.6% for single-cell classification of HeLa cells in three different mitotic stages, with a short processing time of only 2 ms. This technology provides a simple way to realize single-cell fluorescence IFC and real-time cell classification, offering significant potential in various biological and medical applications.

Single-cell resolved ploidy and chromosomal aberrations in nonalcoholic steatohepatitis-(NASH) induced hepatocellular carcinoma and its precursor lesions

Nonalcoholic steatohepatitis (NASH)-induced hepatocellular carcinoma (HCC) and its precursor, nonalcoholic fatty liver disease (NAFLD) are an unmet health issue due to widespread obesity. We assessed copy number changes of genes associated with hepatocarcinogenesis and oxidative pathways at a single-cell level. Eleven patients with NASH-HCC and 11 patients with NAFLD were included. Eight probes were analyzed using multiplex interphase fluorescence in situ hybridization (miFISH), single-cell imaging and phylogenetic tree modelling: Telomerase reverse transcriptase (TERT), C-Myc (MYC), hepatocyte growth factor receptor tyrosine kinase (MET), tumor protein 53 (TP53), cyclin D1 (CCND1), human epidermal growth factor receptor 2 (HER2), the fragile histidine triad gene (FHIT) and FRA16D oxidoreductase (WWOX). Each NASH-HCC tumor had up to 14 distinct clonal signal patterns indicating multiclonality, which correlated with high tumor grade. Changes frequently observed were TP53 losses, 45%; MYC gains, 36%; WWOX losses, 36%; and HER2 gains, 18%. Whole-genome duplications were frequent (82%) with aberrant tetraploid cells evolving from diploid ancestors. Non-tumorous NAFLD/NASH biopsies did not harbor clonal copy number changes. Fine mapping of NASH-HCC using single-cell multiplex FISH shows that branched tumor evolution involves genome duplication and that multiclonality increases with tumor grade. The loss of oxidoreductase WWOX and HER2 gains could be potentially associated with NASH-induced hepatocellular carcinoma.

An isolated single-particle-based SECM tip interface for single-cell NO sensing

As a key factor in cellular metabolic processes, nitric oxide (NO) is a challenging target for in situ real-time monitoring due to its transient property and short diffusion distance. Scanning electrochemical microscopy (SECM) has unique advantages in single-cell analysis, which can obtain the electrochemical current by scanning the cell surface with a tip microelectrode. In particular, it can further improved the electrochemical response by enhancing the interface properties of its tip. Here, an interface design strategy based on platinum single nanoparticle (Pt NP) was developed, and fluorinated self-assembled monolayers (SAMs) were used to further improve its performance. This modified tip was used as an SECM probe for NO concentration monitoring and morphological imaging of single MCF-7 cells. It has the high sensitivity (164.7 μA/μM·cm2) and good selectivity for NO detection, which benefits from the efficient catalytic properties of Pt NPs and high mass transport and hydrophobic antifouling properties of the interface. Notably, it shows a superior performance in detecting the fluctuation of NO released by a single MCF-7-cell under the stimulation of cadmium (Cd), which demonstrates a promising method for using a single-particle-based tip in SECM applications.

An improved Vision Transformer model for the recognition of blood cells

Leukemia is a common, multiple and dangerous blood disease, whose early diagnosis and treatment are very important. At present, the diagnosis of leukemia heavily relies on morphological examination of blood cell images by pathologists, which is tedious and time-consuming. Meanwhile, the diagnostic results are highly subjective, which may lead to misdiagnosis and missed diagnosis. To address the gap above, we proposed an improved Vision Transformer model for blood cell recognition. First, a faster R-CNN network was used to locate and extract individual blood cell slices from original images. Then, we split the single-cell image into multiple image patches and put them into the encoder layer for feature extraction. Based on the self-attention mechanism of the Transformer, we proposed a sparse attention module which could focus on the discriminative parts of blood cell images and improve the fine-grained feature representation ability of the model. Finally, a contrastive loss function was adopted to further increase the inter-class difference and intra-class consistency of the extracted features. Experimental results showed that the proposed module outperformed the other approaches and significantly improved the accuracy to 91.96% on the Munich single-cell morphological dataset of leukocytes, which is expected to provide a reference for physicians' clinical diagnosis.

Label-free single-cell live imaging reveals fast metabolic switch in T lymphocytes

Label-free single-cell live imaging reveals fast metabolic switch in T lymphocytes