Hongshen Ma

Macrophages are highly plastic immune cells which morph into different phenotypes depending on their surrounding environment. The two main paradigms for macrophage phenotyping are phenotyping by function and by origin. Phenotyping based on function separates macrophages into three polarizations: naïve (M0), pro-inflammatory (M1), and anti-inflammatory (M2). Phenotyping based on origin distinguishes macrophages from their place of origin. This can be from monocytes created in the bone marrow, or local proliferation from tissue resident macrophage populations. These phenotypes have different functions and can be associated with different metabolic states, which can be detected via immunofluorescence assays or autofluorescence. Detecting these phenotypes can be a powerful tool in the diagnosis of different diseases, as many diseases impact macrophage populations in the affected tissue. Currently macrophage phenotyping is difficult and requires use of unreliable surface markers or expensive sequencing techniques. In this thesis we use fluorescent microscopy-based images for the training and validation of different image-based algorithms for effective classification. We first imaged THP-1 derived M0/M1/M2 macrophages for brightfield and fluorescently stained mitochondrial images. We then applied this dataset to multiple state-of-the-art deep learning models, achieving 5-fold validation accuracy of 80.50% using PNASNet-5. We also explore phenotyping murine tissue resident macrophages, bone-marrow monocyte-derived macrophages, and monocytes. This was conducted by imaging for autofluorescence and applying to an original convolutional neural network (MacNet) as well as using CellProfiler feature engineering and classical classifiers. We achieve 97.72% validation accuracy using MacNet and 98.58% validation accuracy on CellProfiler features in random forest.

View record

A fundamental challenge to the scalability of microfluidic chemotaxis assays is the requirementfor time-lapse imaging to continuously track migrating cells. Current assays are not suitable fordrug testing and drug screening applications that require the ability to perform hundreds ofexperiments in parallel. End-point chemotaxis assays wherein cells are aligned at fixed startingpoint before migration have been proposed as an alternative to continuous tracking. Previousmethods developed to align cells use fluid flow to pattern cells in traps or constrictions, requiringexternal instrumentation and subjecting the cells to high shear stress. By patterning cells throughcentrifugation in our microfluidic device, alignment can be achieved without precise flow controlwhile applying minimal shear stress to cells. This technique is insensitive to cell geometry and iscapable of handling rare cells in the liquid sample. Additionally, as the chemical gradient in theassay is generated through passive diffusion, the stand-alone device can be placed in theincubator and subsequently imaged to obtain the migration characteristics in each of the 12devices on our substrate. The device was used to observe the response of human neutrophils togradients of fMLP. Our study reveals the potential to leverage cell alignment throughcentrifugation to develop highly scalable, end-point microfluidic chemotaxis assays.

View record

Chemotaxis is the migration of cells in response to a chemical stimulus. This phenomenon is a part of many physiological and pathological processes, such as the neutrophilic response to bacterial invasion, as well as tumour invasion and metastasis. Since the 1960s, assays have been developed to study cell chemotaxis. Early assays mostly measured the number of cells migrating across a membrane, but do not allow tracking of individual cells. Recently, microfluidic assays have enabled single-cell tracking, but they are only able to maintain a stable chemical gradient for a limited time, or require continuous perfusion to maintain a chemical gradient.Here, we developed a microfluidic chemotaxis assay which is capable of maintaining a stable chemical gradient for an extended period of time without the need for a fluid flow system. This capability is achieved by forming a linear chemoattractant gradient in a hydrogel prepolymer in a microchannel, then polymerising the hydrogel by exposure to UV light, thereby fixing the gradient in place. Cells are dispensed on top of the polymerised hydrogel and the cell response to the chemoattractant gradient is observed. Compared to many existing chemotaxis assays, this device requires significantly less time and user expertise to operate. Two versions of the hydrogel-stabilized chemotaxis assay have been developed. Version 1 is manufactured using polydimethylsiloxane (PDMS), while Version 2 is manufactured using 3D printing of translucent epoxy resin. For both versions, the manufacturing methods and operation protocol have been optimised to achieve a device reliability of 80%. Version 2 dramatically reduced the number of failure modes and simplified device operation. A simulation study was conducted to better understand the diffusion process that forms the chemical gradient, and verify the gradient profile applied to the cell sample.

View record

Monoclonal antibodies have become a dominant biopharmaceutical in recent years, with sales expected to exceed $125 billion by 2020. Antibody therapies have been used to treat a wide range of diseases including cancer, multiple sclerosis, and rheumatoid arthritis, with higher specificity and lower toxicity than other chemotherapeutics. The potential promise of antibody therapies has necessitated a significant need to develop improved production technologies in order to shorten the timelines for development, testing, and clinical trials.Modern methods of monoclonal antibody production involve transfecting an antibody gene expression cassette into a host cell line for production, where the cassette is randomly integrated into the genome. This random integration results in a heterogeneity between transduced cells, resulting in significant variability in antibody production rate within the cell population. Additional screening and selection processes are therefore needed to optimize the productivity of the antibody-producing cell line. While several strategies have been developed to select high-producing cell lines, each existing strategy suffers from problems such as long timelines, indirect selectivity, complex procedures, and proprietary processes.We developed a technology named Selective Laser Gelation (SLG) capable of selectively arresting the growth of individual target cells. This capability is enabled by localized gelation using an infrared laser to utilize the unique inverse solution-gel transition of methylcellulose solutions. Phase-transition hysteresis enables the retention of localized gels after the laser is removed. Methylcellulose solution limits the diffusion of secreted antibody from individual cells and small colonies, and when combined with a fluorescently-conjugated secondary antibody, the produced antibody can be visualized and quantified. This capability is then used to selectively preserve high antibody-producing cells while arresting the growth of low-producing cells. In this thesis, we first modeled the thermodynamics of laser heating on a methylcellulose solution. We then developed an experimental apparatus and software to test the SLG procedure, which we used to show that the SLG process can selectively inhibit the growth of selected cells. Finally, we use the SLG process to increase overall antibody productivity within a shorter timeline than current methods by selecting high-producing antibody-secreting cells.

View record

Circulating tumour cells (CTCs) are cancer cells shed from a primary tumour site into the bloodstream, where they have the potential to invade other tissues in the body, and thus become the seed of metastases. CTCs have great potential to monitor disease progression and guide cancer treatment, but a key technical challenge for their isolation and characterization is their extreme rarity in blood. CTCs are commonly enriched using immunoaffinity, which while being highly selective, may fail to capture cells that have weak antigen expression. The biophysical properties of CTCs offer a compelling alternative to immunoenrichment. CTCs are much larger in size than erythrocytes, but are similar to leukocytes. Owing to their epithelial origin however, CTCs are likely to be more rigid than leukocytes which allows for deformability based methods to separate these cells. Previously, our group has demonstrated the continuous flow microfluidic ratchet device for deformability based separation of CTCs. Here, an improved version of the device has been developed to be compatible with pre-enrichment methods, allowing for a dramatic increase in throughput. While similar in principle to the previous version, this work specifically improves the design of the sample infusion area to increase the points of contact between the sorting matrix and sample inlet, in order to prevent the accumulation of cell debris. Using this new design, epoxy resin devices and supporting instrumentation were developed to provide a pathway towards scale-up production and automation. These combined improvements allow biology laboratory technicians to enrich CTCs without significant training. The improved device is capable of capturing > 80% CTCs from whole blood at a throughput of 1 mL/hr, which when combined with a red blood cell lysis pre-enrichment step, increases to 8 mL/hr. Finally, devices were used to enrich CTCs from patients with metastatic castration-resistant prostate cancer. CTCs were found in 3 out of 11 patients, with an average count of 78. The enriched cells were further processed to perform single cell genomic sequencing where CTCs were found to contain driver mutations including those commonly associated with prostate cancer.

View record

Circulating tumor cells (CTCs) are exceedingly rare cancer cells shed from tumors into the bloodstream, where they have the potential to invade other tissues to seed metastases. CTCs are difficult to isolate but their critical role in tumor metastasis, as well as their proven prognostic value has attracted tremendous interest in recent years. While many methods have been developed to isolate CTCs, a major bottleneck to their clinical application has been the precise identification and characterization of these cells, owing to their tremendous phenotypic heterogeneity. To address these formidable challenges, a number of microscopy techniques have been applied to gather large amounts of information about captured cells. However, these studies are currently limited by two major concerns: First, due to the phenotypic plasticity of tumor cells, there may be significant variability in the properties of CTCs as observed using microscopy. Second, if the CTCs are subjected to multi-parameter analysis, the high-content data may be too expansive to analyze with a reasonable amount of time and effort. In this thesis, I developed an efficient and customizable spectral image cytometry platform to collect multi-spectral data from immunofluorescence micrographs of cell samples enriched for CTCs in order to quickly and easily analyze this information to facilitate CTC identification and characterization. This work includes the development of software tools to convert microscopy data for processing, to segment the images into single cell images, to rank potential CTCs, and to provide a user interface for rapid augmented review. The performance of this software platform has been evaluated by analyzing multi-spectral fluorescence imaging data previously collected by our group from ten patients with castrate resistant prostate cancer, and then comparing the result to unassisted manual reviews performed by blinded reviewers. The final CTC identification counts closely matched manual analysis with a slight increase in verified CTC counts, which is likely a result of the comprehensive nature of the automated screening process. The average computation time is 4.5 minutes per sample, which is faster than the time required to acquire the imaging data, and thus allows operators to quickly review results between acquisitions.

View record

Red blood cell (RBC) deformability plays an important role in the pathogenesis of Plasmodium falciparum malaria, and therefore could potentially enable simple, rapid, and reagent-free biophysical assays. A key challenge, however, is that pathological cells often only represent a small fraction of the sample, which requires testing a large number of individual cells to enable their detection. Additionally, it is often desirable to perform multiple assays simultaneously, which require technologies capable of parallelized analysis. Traditional technologies for analyzing RBC deformability are limited by their experimental difficulty, extensive instrumentation requirements, as well as their lack of throughput and parallelizability. Here, a new microfluidic mechanism called trans-dispersion is developed to address these issues, enabling a high-throughput and parallelized analysis of RBC deformability. The trans-dispersion mechanism transports single RBCs through a series of constrictions in a microfluidic channel, where their transit speed is a function of their deformability. This process is analogous to gel-electrophoresis, where the migration speed of molecules depends on their length. To ensure a sensitive and consistent measurement, the geometry of the constriction is sized such that the transiting cell forms a temporary seal with each constriction while supporting microchannels ensure consistent forces are applied to each deformation channel. After undergoing repeated deformations, the final position of each RBC, indicating its deformability, is determined using simple bright-field microscopy and automated image processing, and thereby resulting in a repeatable, high-throughput and parallelized process. The performance of this mechanism was evaluated by detecting changes in RBC deformability resulting from chemical degradation, malaria parasitism and exposure to anti-malarial drugs. This device can distinguish variation in RBC deformability following chemical degradation using small concentrations (0.0005%) of glutaraldehyde (GTA). P. falciparum-infected RBCs (iRBCs) show distinct deformability curves compared to the uninfected controls. The linear correlation between the parasitemia and the percentage of non-transiting cells could potentially be used to infer the parasitemia of clinical specimen. Furthermore, this device was able to simultaneously assess the efficacy of several antimalarial compounds; showing that rigidification of P. falciparum-iRBCs can potentially be used to evaluate antimalarial drug efficacy, as well as serve as a functional screen for new antimalarials.

View record

Immunoenrichment of conventional circulating tumor cells (CTCs) may fail to capture cells with poor antigen expression. Micropore filtration is a compelling label-free alternative to separate CTCs based on their biophysical characteristics rather than biochemical characteristics. However, this strategy is prone to clogging of the filter microstructure, which dramatically reduces selectivity after processing large numbers of cells. Our group previously reported the resettable cell trap (RCT) mechanism to perform micropore filtration in a way that is resistant to clogging. We improved the selectivity of this label-free mechanism by filtering the samples multiple times on chip and dramatically improving the throughput by parallelization. The resettable cell trap device is a microfluidic mechanical constriction with adjustable apertures that can capture CTCs based on their distinct size and deformability. It can also be periodically cleared to release the trapped cells to prevent clogging. Three identical cell traps are aligned in series which improves selectivity by removing leukocytes that non-specifically adhere to the surface of microchannels.We validated this mechanism by doping UM-UC13 bladder cancer cells into diluted whole blood at a density of 1 UC13 to 1000 leukocytes. The first filtration step achieved 183-fold enrichment and 93.8% yield. The second and third traps together provided an additional enrichment of ~5 without significant change in yield. Furthermore, additional filtration steps provide even greater enrichment. In patients with metastatic castration-resistant prostate cancer (mCRPC, n=24) and localized prostate cancer (LPC, n=18), CTCs were successfully identified using the resettable cell trap device followed by single-cell spectral analysis. We additionally compared the RCT device to the CellSearch® System, the only FDA approved commercial CTC enumeration platform. The microfluidic RCT device identified 83.3% (20/24) patients with >=5 CTCs per 7.5 ml of blood with a mean of 329 counts. Within the same patient group, the CellSearch only measured >=5 CTCs in 37.5% (9/24) patients with a mean of 23 CTCs per 7.5 ml of blood. The RCT device identified significant more CTCs and positively identified more mCRPC patients than the CellSearch system.

View record

Circulating tumor cells have been implicated as potential seeds of cancer metastasis and have strong prognostic and diagnostic value in cancer therapy. The primary challenge in CTC characterization is their extreme rarity in circulation relative to leukocytes. Conventional strategies employ CTC immunoenrichment that is highly selective but may fail to enrich for CTCs with poor antigen expression. However, CTCs exhibit unique morphological characteristics that distinguish them from leukocytes and deformability-based sorting mechanisms represent a compelling label-free CTC enrichment strategy. Our group previously reported the microfluidic ratchet mechanism capable of highly selective deformability based cell separation without clogging. Here, we developed a continuous version of this process that obviates the need for microvalves and operate with dramatically increased throughput. Implementation of the microfluidic ratchet consists of a matrix of funnel constrictions with microchannels for flow control. The openings of the funnel constrictions are gradually reduced from the bottom row to the top row. Cells enter at the bottom-left of the funnel matrix and are driven by a rightward flow simultaneously as a vertical oscillatory flow. Each cell traverses through the funnel matrix in a step-wise diagonal path until reaching a limiting funnel size. CTCs are the least deformable cells and reach their limiting funnel size relatively quickly. Leukocytes are more deformable and travel to a smaller funnel region. Finally, erythrocytes are extremely deformable and exit through the top row. We evaluated the selectivity of this mechanism using UM-UC13 bladder cancer cells doped into whole blood from healthy donors. UM-UC13 cells were enriched by ~10⁴ relative to leukocytes, with ~90% capture efficiency, and thus demonstrate significantly greater selectivity than separation based solely on size. We used the microfluidic ratchet device to enumerate CTCs from 58 samples with 52 patients with castrate resistant prostate cancer, in parallel with CellSearch, and 6 healthy control samples. The CTC capture rate is significantly higher for our device, which detected ≥5 CTCs in 67.3% of patients with an average count of 256, while the CellSearch system detected ≥5 CTCs in 40.4% of patients with an average count of 74.

View record

Red blood cell (RBC) deformability plays an important role in the pathology of various diseases, including malaria, hemoglobinopathies, and micronutrient deficiencies. Specifically, in malaria, the analysis of RBC deformability presents new approaches for detecting infections and for rapidly evaluating the response to drugs by patients. A key challenge, however, is that the infected RBCs represent only a small subpopulation of clinical blood specimens. Therefore effective detection of infection and analysis require methods that can measure a large number of individual RBCs. Traditional technologies for measuring RBC deformability either cannot evaluate single cells to identify diseased subpopulations or do not have sufficient measurement throughput to detect rare subpopulations. In additions, they require delicate experiments, expensive equipment, and skilled technicians. To address these issues, we developed a new microfluidic mechanism, known as the Multiplexed Fluidic Plunger (MFP), to measure RBC deformability using many microscale-tapered constrictions in parallel. The deformability of each RBC is determined by the threshold pressure required to squeeze the cell through a constriction. Our mechanism overcomes a key challenge where the pressure applied to each cell is dependent on the presence or absence of other cells and thereby produces in an inconsistent measurement result. We devised a mechanism to avoid this error and showed that a consistent measurement is obtained independent of constrictions occupancy. Furthermore, the sensitivity of the MFP device is comparable or superior to existing techniques since it can distinguish control and 0.0005% glutaraldehyde-treated RBCs (p
View record

The ability to separate cells based on biomechanical properties such as size and deformability is emerging as a potential alternative to biochemical methods for cell separation, particularly in cases where biochemical markers are unknown or expressed at low levels. The separation of circulating tumor cells (CTCs) is an example problem where this type of technology is important because the cell surface markers currently used to capture these cells are known to be unreliable. The performance of existing biomechanical cell separation techniques is currently hindered by clogging, which reduces specificity of the separation process. We previously demonstrated a microfluidic ratchet mechanism that overcomes the reversible nature of low Reynolds number flow. In this thesis, we leverage this mechanism to prevent clogging while preserving high selectivity by periodically clearing the filter microstructure to create an automated microfluidic platform that demonstrates the size and deformability-based separation of cultured human bladder UC13 cancer cells from white blood cells (WBCs). This platform has two components: the first is a size-based hydrodynamic concentrator, which performs an initial sample preparation step to reduce the sample volume while removing a fraction of the contaminant WBCs. The second is an automated cell separation device where cells are transported through a 2D array of ratcheting funnel constrictions and sorted using an oscillatory flow. We evaluate the ability of this platform for separating rare cancer cells doped into WBCs at low concentration to assess the potential of this technology for biomechanical separation of CTCs. Specifically, using a sample where cancer cells are doped into WBCs at a ratio of 1:1000, the combined system achieved a cancer cell yield of 96.0±0.1%; the outlet had a purity of 75±3%; and the population of cancer cells in the mixture was enriched by a factor 3000 (+643, -278).

View record

Techniques for the separation of cells from heterogeneous samples that do not rely onbiological labels are important in applications where specific labels are unknown orunavailable. However, limitations of existing label-free separation techniques haveprevented their widespread adoption. Those techniques that separate based on cell sizetypically offer high throughput but lack specificity. Those that separate based on acombination of cell size and deformability have superior selectivity, but are slow and proneto clogging.This work reports a microfluidic device that employs novel resettable cell traps to separatecells based on size and deformability. The resettable cell trap is a microchannel withcontrollable cross-section, featuring recesses to temporarily store captured cells. Larger andless deformable cells flowing through a cell trap with constricted cross-section will beselectively captured due to size restriction, and can be released back into the flow forcollection by enlarging the channel cross-section. Smaller and more deformable cells willsimply pass through the constricted channel. The ability to enlarge the trap and purge it ofcaptured cells enables long term operation without clogging. The cell separation devicepresented is able to separate UM-UC13 cancer cells from human leukocytes with highenrichment (~100x), retention (~90%) and throughput (450,000 cells/hour). Serial separationusing this mechanism provides extremely high enrichment (~2500x) without sacrificingretention. The mechanism is also shown to resolve size differences of 1 µm betweenpolystyrene microspheres. The resettable cell trap is an improvement upon existingtechnology, providing greater enrichment than possible through size-based techniques whileimproving throughput and eliminating problems caused by clogging that are typical offiltration based techniques.

View record

The deformability of single cells can be used as a biomarker to evaluate the status of many diseases including cancer, malaria, and arthritis. Traditional techniques for measuring single cell deformability, such as micropipette aspiration, optical tweezers, and atomic force microscopy, involve delicate experiments performed by highly-skilled technicians using specialized equipment. This thesis presents a new mechanism for measuring the deformability of single cell using the pressure required to deform single cells through a micro-scale constriction. This technique is in principle similar to the micropipette aspiration, but involves considerably simpler operation, is less prone to errors, and requires less specialized equipment and technical skill. The ability of this mechanism to measure single cell deformability is initially verified by testing neutrophils, which demonstrated similar results and measurement precision as micropipette aspiration. Subsequently, this device was used to study the deformability of human red blood cells and specifically, the decrease in deformability of red blood cells parasitized by Plasmodium falciparum, the most common species of the parasite that causes malaria. Finally, this device was used to measure the directional asymmetry associated with the deformation of single cells along the direction of the funnel taper and against the direction of the funnel taper. This asymmetry was used to create a microfluidic ratchet to enable unidirectional transport of cells from a fluctuating fluid flow.

View record

The separation of biological cells using non-chemical methods is important to many areas of medicine and biology. Filtration through microstructured constrictions is one such method where cells can be separated by a combination of size and deformability. This technique, however, is limited by unpredictable variations of the filter hydrodynamic resistance as cells accumulate in the microstructure. Applying a reverse flow to unclog the filter will undo the separation and reduce filter selectivity because of the reversibility of low-Reynolds number flow. This work introduces a microfluidic structural ratchet mechanism to separate cells using oscillatory flow through a 2-dimensional array of funnel-shaped structures. Devices are fabricated using multi-layer soft lithography of polydimethylsiloxane (PDMS) and flow is controlled using pressure sources and on-chip membrane valves. An iterative procedure of design and testing is used to produce a final device which is characterized by the sorting and separation of L1210 mouse lymphoma cells (MLCs), peripheral blood mononuclear cells (PBMCs) from healthy donors, as well as polystyrene microparticles. The ability of this mechanism to sort and separate cells/particles based on size and deformability is investigated and confirmed. Additionally, the spatial distribution of cells after sorting is demonstrated to be repeatable and the separation process is shown to be irreversible. This mechanism can be applied generally to separate cells that differ by size and/or deformability.

View record