Parylene C is a processable and biocompatible polymer, which has been widely used in the microelectromechanical system (MEMS). Recently, the Parylene MEMS techniques have been implemented in various biomedical microdevices to improve the device performance or to enable unique function. In this talk, we will show our recent achievements in Parylene MEMS, including, high throughput liquid biopsy based on Parylene C micropore array, Parylene C caulked PDMS (pcPDMS) technique, fluorescent pipette enabled by enhanced Parylene C autofluorescence.  

Precise manipulation of particles and biological cells is an essential process in various biomedical research fields, industrial and clinical applications. Among various force fields applied for microfluidic cell manipulation, acoustic waves have superior propagating properties in solids and fluids, which can readily enable non-contact cell manipulation in long operating distances. In addition, acoustic fields are advantageous to high power laser beams for non-contact optical tweezing in terms of biocompatibility, throughput and setup simplicity. Exploiting acoustic waves for fluid and cell manipulation in microfluidics has led to a newly emerging research area, acoustofluidics. In this presentation, I will talk about particle and cell manipulation in microfluidics using high frequency surface acoustic waves (SAW). In particular, I will discuss a unique design of a focused IDT (FIDT) structure, which is able to generate a highly localized SAW field on the order of 20 µm wide. This highly focused acoustic beam has an effective manipulation area size that is comparable to individual micron-sized particles. Here, I demonstrate the use of this highly localized SAW field for single cell level sorting using sub-millisecond pulses and selective capture of cells.

Single cell analysis is a powerful tool to unveil cell heterogeneity, which may play a important role in series events in cancer progression, such as tumor initiation, metastasis and clinical chemotherapy resistance. To profile subtle biological variations at individual cell level, methods that allows single cell imaging and corresponding gene expression are still urgently required. Here, we adapted previously developed microfluidic techniques for massive single cell imaging and implemented this strategy to conduct single cell apoptosis analysis and followed by picking up individual cells from the micro-well, performing single cell RNA sequencing, to gain single cell imaging and RNA sequencing data correspondingly. We utilized drug resistant breast cancer cell line and wild type to demonstrate the capability of this approach, which can realize thousands of single cell apoptosis analysis and single cell RNA-seq at the very interested individual cells. From images to corresponded single cell gene expression data, the multiple drug resistant genes has been validated as the main contributors to the phenotype of doxorubicin resistance.

Monocytes represent a highly heterogeneous leukocyte population with the ability to differentiate into macrophages, a major cell type involved in the pathogenesis of atherosclerotic plaque in cardiovascular diseases [1,2]. Label-free analysis of their native cellular phenotypes and functions not only reduces assay cost and time, but is also essential for understanding disease progression and development of novel therapeutic strategies. Impedance cytometry is an emerging technology for high throughput cell phenotyping based on intrinsic electrical properties without the use of antibodies [3,4]. While parallel electrodes offer higher detection sensitivity as compared to co-planar electrodes [3,4], it is limited by laborious microfabrication. Herein, we present the development of an efficient microfluidics impedance cytometer using coplanar electrodes for rapid monocyte identification and phenotyping based on differentiation status.

The microfluidics impedance cytometer consists of a two-inlet, two-outlet polydimethylsiloxane (PDMS) microchannel (30 μm (width) × 20 μm (height)) bonded on patterned coplanar electrodes (20 μm in width with 20 μm separation gap). Sample and sheath fluid are injected into the device to hydrodynamically focus the cells at the channel center prior electrical detection. As the cell moves through the detection region, it disrupts the electric field generated by coplanar electrodes, thereby causing a change in electrical impedance. Using our setup (Fig. 1), the change in electrical impedance was quantified based on current change, and various cellular information were extracted at different frequencies of the excitation signal. We measured two impedance parameters namely the 1) opacity (ratio of impedance magnitude at 0.3MHz (|Z_LF|) to impedance magnitude at 1.7MHz (|Z_HF|)) which reflects cell membrane capacitance, and 2) |Z_LF| which characterize cell size.

For identification of different blood cell types,  human monocytes and lymphocytes purified using Dean Flow Fractionation (DFF) [5], and diluted red blood cells (RBCs) samples were separately injected into the microdevice. As shown in figure 2, different cell types were clearly differentiated based on |Z_LF| due to distinct cell size differences (monocytes: 10 – 12 µm, lymphocytes: 7 – 8 µm, and red blood cells: 6 – 8 µm). Characterization of primary monocytes and leukemic THP-1 monocytic cell line also showed distinct differences in size and opacity (Fig. 3). Lastly, we compared the impedance profile of THP-1 and differentiated macrophages (PMA stimulated), and found that macrophages were more heterogeneous in size with significantly lower opacity than THP-1, demonstrating the feasibility of real-time assessment of monocyte differentiation using impedance-based sensing (Fig. 4).

In conclusion, the developed microdevice enables continuous label-free monocyte phenotyping using coplanar electrodes configuration with sufficient sensitivity. With the design simplicity and low cost fabrication, we envision our method will greatly facilitate immunology research and point-of-care monocyte profiling in patients with cardiovascular diseases.

In recent years, we have developed several versatile analysis platforms for the manipulation and sensing of biomolecules, particularly for low-copy number molecule detection. In the first scenario, sub-30 nm insulating nanoconstriction, serving as molecular dam operating under the balance of negative dielectrophoresis (DEP), electrophoresis, and electroosmosis, enables protein enrichment of 10⁵-fold in 20 seconds [1], which can then be coupled with graphene-modified electrodes for sensitive electrochemical detection of peptides, cancer biomarkers, and cortisol [2-4] (Fig. 1). In the second scenario, an array of electrode nanogaps with sub-10 nm gap size function as templates for AC DEP-based molecular trapping, plasmonic hot spots for surface-enhanced Raman spectroscopy as well as electronic measurements, and fluorescence imaging (Fig. 2), demonstrated with R-phycoerythrin [5] and Alzheimer’s disease candidate biomarkers A-beta 40 and 42 peptides. In the third scenario, we implemented nanoslit as a cost-effective nanofluidic-based immunosensor for low-noise real-time kinetic measurement of fluorescently labeled protein binding (Fig. 3), with a limit of detection down to 1 pM, regardless of the analyte size [6]. Further, a 10 nm deep sub-nanoliter fluidic nanochannels is developed on germanium crystal for attenuated total reflection infrared (ATR-IR) spectroscopy for ultralow volume (~650 pL) molecular characterization [7] (Fig. 4). Our platforms open up simple ways for low-concentration or low-volume sample analysis.

Mechanical force is pervasive in regulating cell physiology and morphogenesis. For example, cell contractility is known positively correlated with the density of actin stress fibers and the size of focal adhesions, and directly associated with cell viability and differentiation. Recently, a variety of cell mechanics was reported with left-right (LR) asymmetry, e.g. rightward-biased cell migration [1]. Such biased mechanics suggests a cell-level mechanism of how LR asymmetry in tissue architectural is formed [2]. Interestingly, because actomyosin cytoskeleton was found important for both expression of LR asymmetry and generation of cellular forces, it predicts a type of cellular force that mediates the LR-biased mechanics and eventually coordinates the formation of LR asymmetry in tissue architecture. To address this, we report a nanowire magnetoscope that reveals a rotating force – torque – exerted by cells [3]. Ferromagnetic nanowires were deposited into cell culture and spontaneously internalized by cells. With a uniform and horizontal magnetic field around the cell culture, nanowires inside the cell were first aligned with the magnetic field and subsequently rotated in clockwise (CW) or counterclockwise (CCW) direction due to the cellular torque (Fig. 1). Importantly, this torque was found with LR bias depending on cell types. While NIH 3T3 fibroblasts and human vascular endothelial cells (hVECs) exhibited CCW torques, C2C12 mouse myoblast cells showed a CW-biased torque (Fig. 2). Moreover, using the quantitatively measured torque and the analysis of subcellular actin distribution, we found that an actin ring composed of transverse arc and radial fibers is the key factor determining the LR bias of cellular torque (Fig. 3 and Fig. 4). Together, our finding of LR biased cellular torque measured by the nanowire magnetoscope offers a new approach for characterizing cell’s rotational force and a fundamental framework explaining single cell’s LR asymmetry. Furthermore, we will discuss how LR asymmetry was regulated by microenvironment cues, i.e. stiffness [4], and how it was utilized in controlling cell orientation and formation of tissue-like architecture, paving the way for rebuilding artificial tissue constructs with inherent LR asymmetry in the future.

Our laboratory has been interested in developing label-free methods to probe the innate information in immune cells, as well as see how they respond to stimulation. Such information can be highly useful either to classify whether disease is present, or to understand the nature and features of a disease. A complicating factor in trying to understand the features of cells, as well as features of populations of cells is the issue that each cell generally has slightly different morphology as well as chemical makeup.

We developed the first imaging system to combine digital holographic microscopy (quantitative phase) with Raman analysis [1]. The two modes give complementary information on the morphological and chemical features of a cell. Phase measurements are also more dominated by proteins in the cell, whereas Raman measurements show stronger contributions from lipids in the cell. From these results, we were also able to determine that for use in high-throughput analysis, we could simplify the measurement approach. The use of quantitative phase to rapidly image the cells, while Raman is used not in an imaging mode, but in a point-wise manner to collect the “spectral signature” of each cell, allows quantitative morphological features to be extracted, with high signal to noise Raman information. In fact, the smart scanning of the cell, where the Raman excitation spot is scanned rapidly through the cell and then descanned before the detector, can be ideal. This approach allows all the Raman shifted signal to be optically binned at the detector, minimizing readout noise and maximizing the spectral information gained. The variance in measured data was spread throughtout a higher number of principal components when using this optical binning approach that we termed “hybrid scanning”. These results have applications for other scientists working towards high thoughput methods that are challenging due to signal-to-noise considerations.

We will present these results in this presentation, and give an overview of how our methods might be applied in a microfluidic environment, along with some of the challenges that result. Ideally, these methods could be used for fully label-free analysis of the degree of activation of macrophages, as well as the identification and statistics of different lymphocyte types in biofluids. To this end, we recently showed the degree to which different lymphocyte cell lines can be classified, using only single point Raman analysis, with relatively high throughput.

REFERENCES:

[1] N. Pavillon, A. J. Hobro and N. I. Smith, "Cell Optical Density and Molecular Composition Revealed by Simultaneous Multimodal Label-Free Imaging", Biophys. J. 105(5), pp. 1123-1132 (2013).

[2] N. Pavillon and N. I. Smith, "Maximizing throughput in label-free microspectroscopy with hybrid Raman imaging", J. Biomed. Opt. 20(1), pp. 016007-1-016007-10 (2015).

[3] A. J. Hobro, Y. Kumagai, S. Akira and N. I. Smith, "Raman spectroscopy as a tool for label-free lymphocyte cell line discrimination", Analyst 141, pp. 3756-3764 (2016).

Corneal infections are vision threatening diseases caused by microorganisms, usually by bacteria or fungi. Corneal infections often lead to impairment in sight or even blindness without suitable and prompt treatment. Rapid and accurate diagnosis is critical to begin appropriate treatment; however, in about 50% of patients it is impossible to make a diagnosis of the pathogen. The conventional slit lamp examination cannot achieve make definite diagnosis. An ophthalmologist usually needs to take a tiny tissue and stain it to identify the category. In addition, traditional microbiology culturing is required to culture the organism for diagnosis. However, these methods are time consuming, for example fungi often require more than 2 weeks and often produce no results. We report the use of spontaneous Raman spectroscopy in diagnosing corneal infection caused by three prevalent types of fungi, including Candida, Aspergillus, and Fusarium. Raman spectroscopy was conducted first on fungal spores from different species and the spectra showed significant differences. The spectra of rabbit corneas infected by different fungi were then obtained with a larger laser focus. Using multi-variable analysis, infected corneas can be distinguished from healthy corneas rapidly, non-invasively and with high accuracy.

How to intoto extract the spatiotemporal information of massive cells when imaging whole organisms is a key challenge to modern light microscopy. Unlike popular epifluorescence microscopy methods, light-sheet fluorescent microscopy (LSFM) recently emerges as a technique of choice that uniquely limits excitation to the vicinity of the focal plane, providing high axial resolution and fast acquisition rate while minimizing the background and photo-damage. Here we present a voxel super-resolving reconstruction strategy in conjunction with an oblique light-sheet scanning approach to achieve three-dimensionally (3-D) enhanced resolution as well as reduced aliasing better suited to volumetric imaging of multi-cellular organisms. This reported method, termed oblique scanning light-sheet microscopy, uses fast GPU-based computation to address the general challenge of high-throughput, high-resolution microscopy that is originally coupled to the physical limitation of the system optics. As demonstrated by imaging of brain and neuron structures,  our method currently offers improved spatial resolution by ~3 folds compared to conventional LSFM images, high speeds up to nearly 2 volumes per minute, and the ability to circumvent the tradeoff between a over 100 mm³ large volume of macro specimen and ~2 μm isotropic resolution of cells.

We present an image-based cell-cycle analysis at single-cell precision measured by a multimodal time-stretch imaging flow cytometer with a throughput >10,000 cells/sec. Flow cytometry is a potent tool for cellular phenotyping, which hold the keys to understand cellular functions [1]. However, it lacks the ability to detect and quantify biophysical phenotypes of cells, an effective intrinsic marker to probe a multitude of cellular processes. A notable example is cell growth, regarded as “one of the last big unsolved problems in cell biology” [2]. Although quantitative-phase microscopy (QPM) enables cell growth studies by quantifying the cell size and dry mass in a label-free manner, it has largely been restricted to adherent cell analysis with a low imaging throughput of ~100’s cells [3,4]. This limitation hampers high-throughput single-cell analysis which is now an unmet need in detection and analysis of rare metastatic cancer cells in a large population (thousands to even millions of cells) [3-6].

Leveraging its ultrafast frame rate, biophysical single-cell imaging based on time-stretch technology has shown its potential in scaling the imaging throughput by at least 2 orders of magnitude higher than the current techniques [7-9]. Here we present a further advancement by developing a multimodal time-stretch imaging flow cytometer for high-throughput image-based cell cycle analysis of cancer cells (metastatic breast cancer cell line, MDA-MB-231). The system features both QPM and fluorescence detection of individual suspended cells, flowing in a polydimethylsiloxane-based microfluidic channel at a high throughput of 10,000 cells/sec. Its configuration is similar to that reported in ref. [8-10] except an additional module for fluorescence excitation and detection. Fig. 1 shows some representative bright-field and quantitative-phase cell images captured by the system.

Not only can the system performs biophysical phenotyping inferred from the QPM, but also quantify the DNA content of single-cells with DNA-specific fluorescence labels. The combined information can be utilized for cell-growth monitoring and characterization of cell-cycle phases. The fluorescence signal is first used to identify the cell-cycle phase of individual cells and revealed the distribution of the cell-cycle phases in the whole population with G1/S/G2M phase, as 57.3%, 18.4% and 24.3% respectively (Inset of Fig. 2). Cell growth is then quantified by the cell dry mass which represents the protein content of each cell [11]. Throughout the cell-cycle, a progressive increase in dry mass from G1 (286±4 pg), via S (357±8 pg) to G2/M (438±11 pg) is observed (p <0.01) (Fig. 2). Furthermore, our analysis combining dry mass and fluorescence signal reveals that faster cell growth occurs in G1 and G2/M phases, in comparison to that during the S phase. It is consistent to the dominating action of DNA content duplication in this phase [12]. In summary, this integrated time-stretch imaging flow cytometer (QPM plus fluorescence detection) presents a powerful tool for large-scale single-cell analysis based on both molecular signatures (e.g. DNA content) and biophysical markers (e.g. dry mass) – opening a new paradigm in single-cell analysis of basic biology and new mechanistic insights into disease processes, not limited to cancer cell growth.