Multi-modal microfluidics for cancer cells bio-sensing
The main focus of this research will be developing cells-on-chip biosensors to study multi-modal biophysical attributes for cancer cells as well as the anticancer drug treatment studies. My experiences on biosensors are using microelectrodes and constriction-based microfluidic channels to detect the bioelectrical parameters of cells, and using microfluidic channels to detect the biomechanical parameters. Microfluidics and machine learning technologies are able to be integrated to study both the cancer cell lines and patient primary tumor cells. Generally, transformed cells are “softer” and hence more deformable than their healthier counterparts that may facilitate motility. Aggressive cancer cells acquire the ability to deform and squeeze through tissue matrix and access the circulatory system and subsequently adhere to and extravagate blood or lymph vessels to establish secondary tumors suggesting that the metastatic potential of cancer cells is related to their mechanical properties. Moreover, ion channel expression and plasma lipid composition are often altered in breast tumor cells, which in turn modulate the cell membrane capacitance and resistance. We demonstrated a cyclic constriction microfluidic device for detecting breast cancer cells using bioelectrical properties of single cells. These studies suggest that cell biophysical properties can be used as potential “label-free” biomarkers to provide more detailed cancer diagnoses. Started from 2017, we developed a platform for cell analysis by surface enhanced Raman scattering (SERS). We found that the Raman Effect can detect the progress of metastatic breast cancer cell mitosis procedure at single cell level. In addition, nano-antennas for Raman scattering can collect bimolecular information of cells at different cell cycles.
High-throughput microfluidic chip on label-free miRNA detection for glioma diagnosis
We developed our own selective lysing technology for mixed samples with both free-floating miRNAs and exosomal miRNAs. Accurate diagnosis of glioblastoma (GBM) and metastatic brain tumor (MBT) is a critical step in personalized medicine in cancer research and cancer treatment. Current diagnosis of GBM is invasive and challenging in identifying the grade of glioma. Moreover, differentiating between glioblastoma and metastatic brain tumor is still inaccurate in clinical practices. The cerebrospinal fluid (CSF) and the micro RNA (miRNA) within CSF have become a useful source of diagnosis of GBM and MBT. The microRNA in CSF can be used as markers for glioma. Direct assay of CSF for miRNA analysis with minimal influence from the cells are challenging by current techniques. Even though CSF is useful in evaluating the treatment progress for brain tumor, due to the technique limitations, CSF is not broadly used as treatment monitoring source. Therefore, a technique with high resolution miRNA detection in CSF can facilitate the diagnosis of brain tumor, and possibly benefit the monitoring of the treatment progress, such as evaluation of the drug response by miRNA variations. The long term goal of this research is to develop a microfluidic device system that is capable of pre-processing the CSF samples and detect the miRNAs accurately and rapidly using the advantage of microelectromechanical system (MEMS) technique and the multi-physical attributes biosensors. The objective of this research is to develop an integrated microdevice system to detect the low concentration of miRNAs in CSF samples. The rationale is that the biophysical attributes of each kind of miRNA binding to the specific single-stranded DNA (ssDNA) will cause a shift in bioelectrical properties of the targeted sensing region, which can be monitored by impedance spectroscopy. Together with kernel-based machine learning algorithms, the sensing data can be considered as high-dimensional non-parametrical parameters, which can be accurately identified as a unique signature for GBM or MBT. The main hypothesis is that this device can identify the miRNAs within CSF, and correlates the miRNAs as markers for GBM and MBT, so that a clinical relevance can be established. (1) Microdevice development for CSF pre-processing and miRNA sensing. The MEMS technique can integrate multi-physical biosensors into a microdevice for detecting the low concentration miRNAs. First, the pre-processing device will have an integrated multi-constriction microchannel arrays for cell trapping with the assist of dielectrophoresis trap. Second, the cyclic deformability microfluidic channels with capacitive electrodes can inspect the bioelectrical data of the CSF and selectively perform the exosomes lysis to harvest miRNAs with the residual cells intact. This procedure can minimize the interference from the cells in miRNA detection. Finally, the arrays of micro-cavities with ion exchange membrane with ssDNA will capture miRNAs from the CSF. Additional microelectrodes will be used for sensing each micro-chamber. The miRNAs in CSF can bind to the ssDNA, and further alter the bioimpedimetric measurement results with the increasing of miRNA concentrations. Our published results show that the miRNA can be accurately detected with a concentration of 1 pM. (2) Multiple kinds of miRNA sensing. The microdevice can be further integrated with multiple parallel sensing channels for multiple kinds of miRNAs. We hypothesis that this device can be integrated in larger scales to attract different kinds of miRNAs with equal probability. We can achieve this goal by designing centrosymmetric microfluidic channels and micro-cavity patterns. Two symmetric pairs of electrodes can be used in differential measurement on a cavity with captured miRNAs for a higher sensitivity than direct impedance measurement. (3) miRNA sensing in CSF with heterogeneous cell mixtures. This microfluidic system will be tested with CSF with lymphocytes and miRNAs. The clinical relevance between the GBM levels or MBT and microfluidic miRNA detection results will be tested. The results can be used to further optimize the machine learning algorithms. The future of this research can be expanded into brain tumor treatment evaluation. Taking CSF samples are easier than invasive biopsy testing. Furthermore, with the assist of improved machine learning and artificial intelligence, this research can be developed into a diagnostic tool for personalized medicine. The expected outcome of this proposal will be a microdevice with multi-physics biosensors for miRNA detection in CSF samples for the diagnosis of GBM level and MBT. The future tasks will be focused on bringing this device and technology from laboratory for personalized medicine for clinical diagnosis and possibly the treatment evaluation after validating the clinical relevance between our technology and clinical diagnose results. In addition, this technique is capable of analyzing the miRNA trace from SARS-CoV-2 to study the possible diseases caused by COVID-19.
3D microfluidics for cancer cell migration and anti-cancer drug responses studies
Most microfluidic devices for cell analysis are 2D microchannels with plenty of features. Improving the analysis resolution and throughput and developing novel applications become priorities in cells-on-chip researches. The PDMS microfluidic fabrication technologies limited the ability of generating 3D microchannels by soft-lithography and bonding. 3D microchannels always accompanied by complex fabrication process flow with multiple alignments and bonding, which significantly decreased the yields of successful devices. Scientists explored other chemical methods for bonding PDMS to multiple materials by specific coatings and treatments. However, the stacked 2D microfluidic channels by connecting via through-holes are not ideal 3D microfluidic channels. Specifically, for the cancer cells metastasis studies or circulating tumor cells studies, a real 3D microfluidic matrix with meander microfluidic channels with round cross-sections can better mimicking the human blood circulatory system. Together with a programmable pressure pump to simulate the heart pumping procedure, this state-of-art 3D microfluidic matrix illustrated in Figure 5 can provide a better in vitro environment for cancer cells metastatic progression or CTC migration. This microfluidic matrix molded from 3D wax-printed vascular structures can achieve the requirement of simulating blood circulatory system. This 3D microfluidic matrix may significantly improve the similarity of cells in vivo.
The 3D printed sacrificial wax struts can be used for PDMS replica molding to simulate the blood vessel matrix in human body. This 3D microfluidic matrix is highly mimicking the human blood vessels in the following aspects:
(1) The cross-sections of the microfluidic channels are round rather than rectangular by SU-8 molding; (2) The flow pumping in main channel is pulsed flow rate rather than constant pressure or constant flow rate in most lab-on-chip approaches; (3) The 3D microfluidic matrix has more flexible design methods to generate capillary vessels rather than stacking multiple layers of 2D microfluidic channels;
In addition, the 3D microenvironment can simulate the migration of circulating tumor cells. The walls can be molded with biomaterials to detect secondary tumors initiated by CTCs. Proper imaging technology can estimate the size or infiltration status of tumor cells. Other than the above mentioned research focuses, my experiences of using biomaterials for membrane in microchip, and my 3D printing experience using self-prepared biomaterials can provide me more possible bio-related research based on lab-on-a-chip and additive manufacturing. There are two additional future research projects that I can immediately start.
Previous Projects
Postdoctoral Research (Virginia Tech)
My work focused on studying the bioelectrical and biomechanical properties of breast cancer cells and circulating tumor cells (CTCs) using microfluidic techniques. The objective of this work is to develop a high-throughput microfluidic chip to expose cells to sequential deformations and to identify differential biomechanical and bioelectrical properties of tumor and normal breast cells. This work results in “mechanical modulatory signatures” that quantify degrees of resistance of breast cells in response to applied sequential forces. We have completed tests of human breast tumor cell lines and through our collaboration with a clinical team, and initiated testing of primary breast cells derived from patient biopsy samples. The application of the work addresses the important challenges of distinguishing normal breast cells from tumor cells with metastatic potential, and to use these devices to identify drug resistant tumor cells which could help physicians choose appropriate drug treatments on an individual patient basis. Aggressive cancer cells, which possess a less organized cytoskeleton, often have acquired the ability to deform and squeeze through complex tissue matrices and gain access to the circulatory and lymphatic systems, where they can spread to distant sites and establish secondary tumors. This suggests that the metastatic potential of cancer cells is related to their mechanical properties. These studies suggest that cell biophysical properties can be used as potential “label-free” biomarkers to provide more detailed cancer diagnoses. We developed a platform for cell analysis by surface enhanced Raman scattering (SERS). We found that the Raman Effect can detect the progress of metastatic breast cancer cell mitosis procedure at single cell level. In addition, nano-antennas for Raman scattering can collect bimolecular information of cells at different cell cycles. The biophysical attributes are able to distinguish cancerous and normal cells, or even different subtypes of cancer cells. Furthermore, the anti-cancer drug treatment studies on the cancer cells can also be achieved on these kinds of multi-modal microfluidic techniques.
Postdoctoral Research (University of Notre Dame)
In this project, we aim to explore the potential of coupled oscillator networks made of living cardiac muscle cells, or bio-oscillators, as collective computing components for solving computationally hard problems such as optimization, learning and inference tasks. Cardiac muscle cells are electrically active components that can initiate and relay electrical signals without loss. More interestingly, they spontaneously beat (i.e. oscillate) at a stable pace, and when coupled with each other, they synchronize to a locked, steady frequency. We started with a pair of coupled bio-oscillators and characterize their synchronization behavior. Then we investigated a lithography based microfabrication method to scale up to complex networks with many nodes and edges. In order to demonstrate the communication between two clusters of cardiomyocytes (CMs), we patterned the CMs into a pair of nodes of 100~1000 CM cells with a bridge connection. CMs, the major functional cell of the myocardium through its ability to fire action potentials, constitute only one-third of the population, while the rest were made of non-excitable cells such as the cardiac fibroblasts (CFs). Conduction across the myocardium and dynamic function of the CMs were influenced by the various factors including CM-CF crosstalk. We designed and fabricated novel microdevices to fulfill the function of detection of CM action potentials with adjustable micropatterns with the scale of 100~1000 cells. This project is aiming to build a bio-nano hybrid computational component with data processing capabilities and microelectrode array read outputs for programmable and configurable computing.
Ph.D Research
My Ph.D research topic is “Cell-free Artificial Photosynthesis System”. I participated The objective of this research is to create a cell-free artificial platform for harvesting light energy and transforming the energy to organic compounds. In order to achieve this objective, we take the approach of mimicking the photosynthetic processes of a di-cotyledons plant leaf and integrating them into a compact system using microfabrication technology. We envision integrating the “artificial leaves” to create a compact energy harvesting system with high efficiency. (1) Light reaction are realized in a microfluidic platform that consists of two fluid chambers separated by a planar membrane with embedded proteins that convert light energy into ATP; (2) Dark reaction are realized in another microfluidic platform that consists of two chambers (one for liquid containing ATP and the other for air flow) separated by a membrane that can transport CO2 from the air to the liquid chamber; (3) Glucose synthesis and storage unit were developed by mimicking sponge mesophyll found in a leaf; (4) Interconnection and integration of the above-mentioned three key components will be investigated to generate a complete cell-free artificial platform for effective energy harvesting. This research brings together expertise in advanced manufacturing (bio- and microfabrication, additive manufacturing), biochemistry and biomaterials, and system control and integration.