Advanced microfluidic droplet manipulation based on piezoelectric actuation

As droplet-based microfluidic devices evolve, the demand for simple-to-fabricate droplet manipulation modules increases. Of these modules, droplet sorting has drawn much attention due to its ability not only to enrich, but also to selectively isolate droplet subpopulations of interest. In this paper, we present an innovative piezoelectric-driven droplet sorter that is simple to fabricate, reproducible and robust, which provides extensive control over spatio-temporal droplet pattern. This degree of control is demonstrated by sorting droplets of alternating volumes and by grouping defined number of droplets into traveling clusters. The ability to automatically sort droplets is demonstrated by computerized detection and sorting of droplets based on their color. The sorter performance was investigated and found to work on a wide range of sorting parameters. The sorter is created by a single step fabrication process and does not rely on complex electronics or optics. These advantages simplify the adoption of droplet-based microfluidic technology by the scientific community and provide an ideal platform for single cell assays.     

Au nanoparticles enhanced fluorescence detection of DNA hybridization in picoliter microfluidic droplets

This work reports a facile microfluidic device for Au-nanoparticle enhanced fluorescence detection of tiny amount of nucleotides within droplets in a high-throughput way. Droplets containing single strand DNA probe and relevant complementary strands DNA(cDNA) are generated in flow-focusing manner and the hybridization between them is realized in droplets flowing along a long serpentine channel. In order to find the optimal experimental condition, finite element method simulation is used to predict the interface evolution between the two phase liquids. Based on the fluorescence emited by intercalator reacted with the generated double-strand DNA(dsDNA), the target cDNA with a concentration of 1nM can be detected in droplets. And when we adopt Au nanoparticles to immobilize DNA probe which can amplify the fluorescence intensity, 10pM completary DNA could be detected. Due to the advantages in high-throughput and compartmentalization of this droplet platform, the detection procedure can be finished in 3 h. Our method shows good potential application in facile, sensitive, low cost and fast DNA detection for applications in personal health care.     

Droplet-based microsystem for multi-step bioreactions

A droplet-based microfluidic platform was used to perform on-chip droplet generation, merging and mixing for applications in multi-step reactions and assays. Submicroliter-sized droplets can be produced separately from three identical droplet-generation channels and merged together in a single chamber. Three different mixing strategies were used for mixing the merged droplet. For pure diffusion, the reagents were mixed in approximately 10 min. Using flow around the stationary droplet to induce circulatory flow within the droplet, the mixing time was decreased to approximately one minute. The shortest mixing time (10 s) was obtained with bidirectional droplet motion between the chamber and channel, and optimization could result in a total time of less than 1 s. We also tested this on-chip droplet generation and manipulation platform using a two-step thermal cycled bioreaction: nested TaqMan® PCR. With the same concentration of template DNA, the two-step reaction in a well-mixed merged droplet shows a cycle threshold of ～6 cycles earlier than that in the diffusively mixed droplet, and ～40 cycles earlier than the droplet-based regular (single-step) TaqMan® PCR.     

Generation of monodisperse alginate microbeads and <Emphasis Type="Italic">in situ</Emphasis> encapsulation of cell in microfluidic device

A microfluidic method for the in situ production of monodispersed alginate hydrogels using chaotic mixing is described. Aqueous droplets comprising of alginate and calcium as a cross-linking agent were formed as an immiscible continuous phase, and then the alginate and calcium in the droplet came into contact and were rapidly mixed. Gelation of the hydrogel was achieved in situ by the chaotic mixing of the droplets in the microfluidic device. Important operating parameters included: the capillary number (Ca) and the flow rate of the continuous phase, which mainly influenced the formation of three distinctive flow regimes, such as fluctuation, stable droplets, and laminar flow. Under the stable formation of droplets regime, monodispersed alginate microbeads having a narrow size distribution (below 3% of CV) were produced in the microfluidic device and the size of the microbeads, ranging from 60 to 95 μm, could be easily modulated by varying the flow rate, viscosity, and interfacial tension. In addition, this approach can be applied to the encapsulation of yeast cells in alginate hydrogels with a high monodispersity. This simple microfluidic technique for the production of monodispersed hydrogels and encapsulation of biomolecules shows strong potential for use in biosensors, cell sensors, drug delivery systems, and cell transplantation applications.     

Using circulating tumor cells to evaluate the efficacy of irreversible electroporation for unresectasble pancreatic cancer

We used circulating tumor cells (CTCs) as biomarkers to evaluate the efficacy of pre-irreversible electroporation (IRE) and post-IRE for unresectable pancreatic cancer (PC). Real-time qPCR was used to detect potential biomarker genes in CTCs, and magnetic-activated cell sorting (MACS) and fluorescence-activated cell sorting (FACS) were performed on 43 patients with PC who underwent IRE. Some patients experienced adverse reactions within 30 days of the operation, including arrhythmia (6.9%), intraoperative transient change of blood pressure (25.5%), cough (11.6%), nausea and vomiting (23.3%), ascites (25.6%), fever (9.3%), and pain of puncture point (60.5%). The number of CTCs decreased significantly with postoperative time (P < 0.01). Delta cycle threshold values for the CTC-related genes CEA, Ep-CAM, and CK19 increased significantly after IRE. Furthermore, the expression of CEA, Ep-CAM, and CK19 decreased significantly with time after IRE (P < 0.01). Detecting CTCs by RT-qPCR and FACS combined with MACS has significant diagnostic and prognostic value for evaluating the efficacy of IRE in patients with unresectable PC.     

Intravenous lipid infusion results in myocardial lipid droplet accumulation combined with reduced myocardial performance in heparinized rabbits

The question addressed in this study was whether a relation between myocardial lipid droplet accumulation and depressed myocardial function existed following intralipid infusion for 45 min in open chest, anaesthetized rabbits. One group of rabbits (n = 8) received intralipid infusion whereas a control group (n = 8) received sodium chloride. Local myocardial performance was obtained by sonomicrometry and the fractional volume of myocardial lipid droplets was measured by morphometric methods. The fractional volume of lipid droplets was 0.667 ± 0.1160% in the intralipid group compared with 0.318 ± 0.080% in the control group (P < 0.03). Cardiac output and stroke volume fell 26% (P < 0.0001) and 34% (P < 0.000), respectively, as a result of intralipid infusion. However, myocardial blood flow obtained by radiolabelled microspheres remained unchanged. Local myocardial function was reduced for both segments after intralipid infusion; maximal systolic shortening was reduced from 15.63 ± 1.45 to 12.07 ± 1.55% (P < 0.002) in the circumferential segment and from 9.46 ± 1.17 to 7.40 ± 0.53% (P < 0.05) in the longitudinal segment. The end-diastolic length of the circumferential segment was reduced by 3% (P < 0.05) after intralipid infusion. The reduced end-diastolic length of circumferential segments together with unchanged left ventricular end-diastolic pressure might indicate reduced left ventricular end-diastolic compliance. We conclude that acute intralipid infusion in rabbits results in myocardial lipid droplet accumulation and depressed local myocardial function.     

Bubble Jet agent release cartridge for chemical single cell stimulation

We present a new method for the distinct specific chemical stimulation of single cells and small cell clusters within their natural environment. By single-drop release of chemical agents with droplets in size of typical cell diameters (d <30 μm) on-demand micro gradients can be generated for the specific manipulation of single cells. A single channel and a double channel agent release cartridge with integrated fluidic structures and integrated agent reservoirs are shown, tested, and compared in this publication. The single channel setup features a fluidic structure fabricated by anisotropic etching of silicon. To allow for simultaneous release of different agents even though maintaining the same device size, the second type comprises a double channel fluidic structure, fabricated by photolithographic patterning of TMMF. Dispensed droplet volumes are V?=?15 pl and V?=?10 pl for the silicon and the TMMF based setups, respectively. Utilizing the agent release cartridges, the application in biological assays was demonstrated by hormone-stimulated premature bud formation in Physcomitrella patens and the individual staining of one single L 929 cell within a confluent grown cell culture.     

Automated quality control and cell identification of droplet-based single-cell data using dropkick

A major challenge for droplet-based single-cell sequencing technologies is distinguishing true cells from uninformative barcodes in data sets with disparate library sizes confounded by high technical noise (i.e., batch-specific ambient RNA). We present dropkick, a fully automated software tool for quality control and filtering of single-cell RNA sequencing (scRNA-seq) data with a focus on excluding ambient barcodes and recovering real cells bordering the quality threshold. By automatically determining data set–specific training labels based on predictive global heuristics, dropkick learns a gene-based representation of real cells and ambient noise, calculating a cell probability score for each barcode. Using simulated and real-world scRNA-seq data, we benchmarked dropkick against conventional thresholding approaches and EmptyDrops, a popular computational method, showing greater recovery of rare cell types and exclusion of empty droplets and noisy, uninformative barcodes. We show for both low- and high-background data sets that dropkick''s weakly supervised model reliably learns which genes are enriched in ambient barcodes and draws a multidimensional boundary that is more robust to data set–specific variation than existing filtering approaches. dropkick provides a fast, automated tool for reproducible cell identification from scRNA-seq data that is critical to downstream analysis and compatible with popular single-cell Python packages.

Single-cell RNA sequencing (scRNA-seq) allows for untargeted profiling of genome-scale expression in thousands of individual cells, providing insights into tissue heterogeneity and population dynamics. Droplet-based platforms that involve microfluidic encapsulation of cells in water–oil emulsions (Klein et al. 2015; Macosko et al. 2015; Zheng et al. 2017) have grown widely popular for their robustness and throughput. The use of barcoded poly-thymidine capture oligonucleotides provides information for assigning eventual sequencing reads to each droplet downstream from bulk library preparation. Because of the low cellular density required to avoid doublets (i.e., two or more cells captured in the same droplet), the vast majority of droplets are empty, ideally containing only tissue dissociation buffer and a barcoded RNA-capture bead with no cellular RNA. However, during tissue dissociation, cells may die and lyse, shedding ambient mRNA into the supernatant that is then captured as background in droplets containing cells and so-called “empty droplet” reactions. Ultimately, a droplet-based scRNA-seq data set contains up to hundreds of thousands of barcodes that correspond to these “empty droplets,” which include sequenced material from ambient RNA alone.To prepare these data for downstream analysis, empty droplets and other uninformative barcodes with little to no molecular information must be removed. Often, computational biologists will define manual thresholds on global heuristics such as total counts of unique molecular identifiers (UMIs) or the total number of genes detected in each barcode in order to isolate high-quality cells. Although these hard cutoffs may generally yield expected cell populations and remove the bulk of populational noise in low-background samples, they are highly arbitrary, batch specific, and generally biased against cell types with low RNA content or genetic diversity (Lun et al. 2019). Furthermore, lenient thresholds often yield filtered data sets with populations of dead and dying cells or empty droplets with high ambient RNA content, especially in encapsulations with high background resulting from tissue-specific cell viability and dissociation protocols. These cell clusters may be gated out manually by the experienced single-cell biologist, but they will distort dimension-reduced embeddings and alter statistical testing for differential gene expression if left unchecked.Here we introduce dropkick, a fully automated machine learning software tool for data-driven filtering of droplet-based scRNA-seq data. dropkick provides a quality-control (QC) module for initial evaluation of global distributions that define barcode populations (real cells vs. empty droplets) and quantifies the batch-specific ambient gene profile. The dropkick filtering module establishes initial thresholds on predictive global heuristics using an automated gradient-descent method and then trains a gene-based logistic regression model to assign confidence scores to all barcodes in the data set. dropkick model coefficients are sparse and biologically informative, identifying a minimal number of gene features associated with empty droplets and low-quality cells in a weakly supervised fashion. The following study aims to show how dropkick outperforms basic threshold-based filtering and a similar data-driven model (Lun et al. 2019) in recovery of expected cell types and exclusion of empty droplets, with robustness and reproducibility across encapsulation platforms, samples, and varying degrees of noise from ambient RNA.     

Ganglion cell density and oil droplet distribution in the retina of brown-eared bulbul (Hysipetes amaurotis)

This study was intended to determine the number and density of both retinal ganglion cells and the oil droplets of cone photoreceptor cells in brown-eared bulbul (Hysipetes amaurotis). For this study birds were killed with proper dose of anesthetic (pentobarbital, 30 mg/kg), and the eyes were removed from the orbital cavity to isolate the retina. For the ganglion cell study retinal whole-mount specimens were prepared and stained with 0.1% cresyl violet. The different types of oil droplets were counted from color microphotographs of freshly prepared retinal samples. The mean total number of ganglion cells was estimated at approximately 2.5×10⁶; with an average density of 16 523 cells/mm². Two high-density areas, namely the central area (CA) and the dorso-temporal area (DTA), are located in the central and dorso-temporal retinas, respectively, in bulbuls (24 032 cells/mm² in the CA; 23 113 cells/mm² in the DTA). Small ganglion cells persisted in the highest density areas, whereas the largest soma sizes were found in the lowest density areas of the retina. Four types of different colored oil droplets — red, orange, green and clear — were identified with an average density of 29 062/mm². Among the different colors, the green oil droplets had a significantly higher population (13 083/mm²) than the others across the retina. The central retina had a significantly higher number of all types of oil droplets, at a density of 60 552/mm². The density and size of the different colored oil droplets were inversely related across the regions of the retina. Taken together, it is concluded that the CA of the retina is an excellent quality area for visual perception due to peak density of ganglion cells and oil droplets. Moreover, each specific oil droplet makes a distinct contribution to visual perception, thereby ensuring that the bird has a retina that best matches its natural environment and feeding behavior.     

Influence of process conditions during impulsed electrostatic droplet formation on size distribution of hydrogel beads

In the medical applications of microencapsulation of living cells there are strict requirements concerning the high size uniformity and the optimal diameter, the latter dependent on the kind of therapeutic application, of manufactured gel beads. The possibility of manufacturing small size gel bead samples (diameter 300 microm and below) with a low size dispersion (less than 10%), using an impulsed voltage droplet generator, was examined in this work. The main topic was the investigation of the influence of values of electric parameters (voltage U, impulse time tau and impulse frequency f) on the quality of obtained droplets. It was concluded that, owing to the implementation of the impulse mode and regulation of tau and f values, it is possible to work in a controlled manner in the jet flow regime (U> critical voltage UC). It is also possible to obtain uniform bead samples with the average diameter, deff, significantly lower than the nozzle inner diameter dI (bead diameters 0.12-0.25 mm by dI equal to 0.3 mm, size dispersion 5-7%). Alterations of the physical parameters of the process (polymer solution physico-chemical properties, flow rate, distance between nozzle and gellifying bath) enable one to manufacture uniform gel beads in the wide range of diameters using a single nozzle.     

Facilitated expansion of human embryonic stem cells by single-cell enzymatic dissociation

Traditionally, human embryonic stem cells (hESCs) are propagated by mechanical dissection or enzymatic dissociation into clusters of cells. To facilitate up-scaling and the use of hESC in various experimental manipulations, such as fluorescence-activated cell sorting, electroporation, and clonal selection, it is important to develop new, stable culture systems based on single-cell enzymatic propagation. Here, we show that hESCs, which were derived and passaged by mechanical dissection, can be rapidly adjusted to propagation by enzymatic dissociation to single cells. As an indication of the stability of this culture system, we demonstrate that hESCs can be maintained in an undifferentiated, pluripotent, and genetically normal state for up to 40 enzymatic passages. We also demonstrate that a recombinant trypsin preparation increases clonal survival compared with porcine trypsin. Finally, we show that human foreskin fibroblast feeders are superior to the commonly used mouse embryonic fibroblast feeders in terms of their ability to prevent spontaneous differentiation after single-cell passaging. Importantly, the culture system is widely applicable and should therefore be of general use to facilitate reliable large-scale cultivation of hESCs, as well as their use in various experimental manipulations. Disclosure of potential conflicts of interest is found at the end of this article.     

一种能与Fas结合的融合型九肽生物学活性分析

目的鉴定从噬菌体随机肽库筛选所获得的融合型多肽3A8的生物学活性。方法制备融合型多肽3A8，经纯化，采用ELISA检测融合型多肽3A8对Fas．Fc的特异性结合；细胞ELISA检测3A8能否与Jurkat细胞表面天然Fas受体结合；^3H-TdR掺入法检测3A8对Jurkat细胞增殖的影响；流式细胞仪检测细胞凋亡。结果ELISA结果显示融合型多肽能与Fas．Fc特异性结合，呈浓度剂量依赖关系。细胞ELISA实验证实3A8能与Jurkat细胞表面Fas受体结合。^3H—TdR掺入法检测发现3A8可抑制Jurkat细胞增殖，抑制率为50％。经3A8处理细胞经PI染色流式细胞仪分析，发现25．41％的Jurkat细胞发生凋亡，明显高于未经处理的对照组。结论从噬菌体随机肽库获得的融合型多肽3A8能与Jurkat细胞表面受体Fas结合，诱导细胞凋亡。     

Hyaline droplet formation in the renal epithelium of patients with haemoglobinuria

Renal hyaline droplet formation is described in two patients with haemoglobinuria. The droplets were shown to contain haemoglobin by a histochemical method. The findings are correlated with the results of recent experimental studies which have demonstrated that renal hyaline droplets are phagosomes containing absorbed protein and not products of cellular degeneration as was previously thought.     

Fat droplet formation in rat lactating mammary gland and mammary carcinomas viewed by freeze-fracture.

The freeze-fracture morphology of the endoplasmic reticulum and Golgi apparatus membranes was analyzed in lactating rat mammary glands and in mammary carcinomas induced by 7,12-dimethylbenzanthracene or N-nitrosomethylurea. Membranes in close proximity with fat droplets present a spectrum of transformations, from the normal, wavy, particle-rich appearance of cytoplasmic membranes to rigid, particle-free bilayers. The following sequence of events is proposed for the biogenesis of fat droplets: (1) formation of particle-free areas in the endoplasmic reticulum and Golgi membranes; (2) apposition of the bilayered lipid membranes to the growing fat droplet; and (3) progressive conversion of the membrane lipids (or amphypathic lipid precursors) into triglycerides at the periphery of the fat droplet. Our results suggest that membranes are not only involved in the synthesis and secretion processes, but that some of their components are incorporated in the fat droplets and contribute to the secretory product itself.     

Impedance feedback control of microfluidic valves for reliable post processing combinatorial droplet injection

Droplet microfluidics has found use in many biological assay applications as a means of high-throughput sample processing. One of the challenges of the technology, however, is the ability to control and merge droplets on-demand as they flow through the microdevices. It is in the interest of developing lab-on-chip devices to be able to combinatorically program additive mixing steps for more complex multistep and multiplex assays. Existing technologies to merge droplets are either passive in nature or require highly predictable droplet movement for feedforward control, making them vulnerable to errors during high throughput operation. In this paper, we describe and demonstrate a microfluidic valve-based device for the purpose of combinatorial droplet injection at any stage in a multistep assay. Microfluidic valves are used to robustly control fluid flow, droplet generation, and droplet mixing in the device on-demand, while on-chip impedance measurements taken in real time are used as feedback to accurately time the droplet injections. The presented system is contrasted to attempts without feedback, and is shown to be 100% reliable over long durations. Additionally, content detection and discretionary injections are explored and successfully executed.     

Automated single-cell sorting system based on optical trapping

We provide a basis for automated single-cell sorting based on optical trapping and manipulation using human peripheral blood as a model system. A counterpropagating dual-beam optical-trapping configuration is shown theoretically and experimentally to be preferred due to a greater ability to manipulate cells in three dimensions. Theoretical analysis performed by simulating the propagation of rays through the region containing an erythrocyte (red blood cell) divided into numerous elements confirms experimental results showing that a trapped erythrocyte orients with its longest axis in the direction of propagation of the beam. The single-cell sorting system includes an image-processing system using thresholding, background subtraction, and edge-enhancement algorithms, which allows for the identification of single cells. Erythrocytes have been identified and manipulated into designated volumes using the automated dual-beam trap. Potential applications of automated single-cell sorting, including the incorporation of molecular biology techniques, are discussed.     

Induction of lipid droplet accumulation in cardiac muscle cells of guinea pigs and mice: An analysis of the effects of reserpine and fasting

Reserpine has been demonstrated in previous studies to induce the accumulation of lipid droplets in ventricular cardiac muscle cells of bats and this effect has been attributed to mediation by the sympathetic nervous system. The present study was done to evaluate the species specificity of this action of the drug. Reserpine caused lipid droplet accumulation in cardiocytes of guinea pigs and mice. However, in contrast to the action of the drug in bats, this action of reserpine in guinea pigs and mice could not be antagonized by chemical sympathectomy with 6-hydroxydopamine or by treatment of animals with phenoxybenzamine, atropine or hexamethonium. Like reserpine, fasting for as little as 24 hours induced lipid droplet accumulation in cardiocytes of guinea pigs and mice. This effect also could not be prevented by treatment with 6-hydroxydopamine indicating that the sympathetic nervous system was not involved in its mediation. Force-feeding guinea pigs given reserpine prevented the accumulation of lipid droplets normally induced by this drug. It is concluded that the lipid droplets accumulation in the hearts of guinea pigs that follows administration of reserpine is not due to a direct cardiotoxicity of the drug but is secondary to the failure of drug-treated animals to eat. Since the autonomic nervous system appears to mediate reserpine's cardiotoxicity in bats, the species difference revealed by these experiments probably results from an underlying physiological difference in the nervous systems of these animals.     

The visual cells of the chicken as revealed by phase contrast microscopy

Phase microscopic investigations of Kolmer-fixed, depigmented sections of the adult chicken retina have provided photomicrographic evidence of the existence of three different photoreceptors: single rods, single cones, and double cones. The rod extends the entire thickness of the visual cell layer and is characterized by a uniformly thick outer segment and a hyperboloid-containing inner segment which is devoid of an oil droplet. The single cone is the shortest element; it contains a red oil droplet. The double cone consists of two unequal members, a tall, slender chief cone and a broad accessory cone. The chief component contains a large yellow oil droplet, whereas the accessory cone houses a small, oval, yellowish-green droplet and a characteristically large, oval paraboloid. The rod hyperboloid and the accessory cone paraboloid contain glycogen. No colorless droplets have been observed. Owing to the close association between oil droplet color and cone type, three colored layers of oil droplets are formed within the thickness of the retina: a proximal row of red droplets (the short, single cones), an intermediate layer of yellowish-green droplets (the accessory cones), and a distal row of yellow droplets (the tall chief cones).     

Lipid droplet accumulation in cardiac muscle cells of the bat: Potential auto-toxicity of the cardiac sympathetic innervation

The previously described ability of reserpine and parachlorophenylalanine to induce the accumulation of lipid droplets in ventricular cardiac muscle cells of the bat was investigated. Lipid droplet accumulation was assessed qualitatively by light microscopy and quantitatively by morphometric analysis of electron micrographs. An hypothesis that the action of the drugs was an indirect one, mediated by the cardiac adrenergic innervation, was framed and tested. Lipid droplet accumulation occurred during a time of intense sympathetic activity, that of arousal from hibernation. The ability of the two drugs to produce the effect was antagonized by prior sympathectomy with 6–dopamine. The effect was mimicked by administration of exogenous norepinephrine together with inhibitors of its catabolic enzymes, monoamine oxidase and catechol-omethyl transferase. These observations are all consistent with the initial hypothesis and raise the possibility that endogenous norepinephrine in the cardiac sympathetic innervation might be, at least potentially, auto-toxic.     

Lipid droplet accumulation in cardiac muscle cells of the bat: potential auto-toxicity of the cardiac sympathetic innervation.

The previously described ability of reserpine and parachlorophenylalanine to induce the accumulation of lipid droplets in ventricular cardiac muscle cells of the bat was investigated. Lipid droplet accumulation was assessed qualitatively by light microscopy and quantitatively by morphometric analysis of electron micrographs. An hypothesis that the action of the drugs was an indirect one, mediated by the cardiac adrenergic innervation, was framed and tested. Lipid droplet accumulation occurred during a time of intense sympathetic activity, that of arousal from hibernation. The ability of the two drugs to produce the effect was antagonized by prior sympathetectomy with 6-hydroxy-dopamine. The effect was mimicked by administration of exogenous norepinephrine together with inhibitors of its catabolic enzymes, monoamine oxidase and catechol-o-methyl transferase. These observations are all consistent with the initial hypothesis and raise the possibility that endogenous norepinephrine in the cardiac sympathetic innervation might be, at least potentially, auto-toxic.