Label-free single cell electro-mechano-phenotyping using microfluidics impedance cytometry

Abstract

Cell biophysical properties are related to cellular functions and behaviours in diseases, and assessment of these intrinsic properties can be exploited as novel biomarkers for biomedical diagnostics. While label-free cell analysis requires minimal sample preparation and is more cost effective, there are limited biophysical tools that can characterize both mechanical and electrical properties of single cells at high throughput (> 1000 cells/min). In this thesis, we report a novel microfluidic impedance deformability cytometer for multi-frequency (0.3 ~ 12 MHz) electro-mechano-phenotyping of single cells (cell size, deformability, membrane opacity, nucleus opacity) based on hydrodynamic cell stretching. By applying Uniform Manifold Approximation and Projection (UMAP) as a dimension reduction technique for single cell impedance signatures, we demonstrated its utilities in 2 key biomedical applications including 1) rapid induced pluripotent stem cells (iPSCs) profiling in cell therapy, as well as 2) leukocyte characterization in type 2 diabetes mellitus (T2DM) for cardiovascular risk stratification. In regenerative medicine, the tumorigenicity risk from residual iPSCs in differentiated cellular products remains a critical challenge and advocates an unmet need for rapid spinal cord progenitor cells (SCPCs) safety profiling. We developed a first-in-kind supervised UMAP model for continuous impedance-based monitoring and quantification of low abundance (~ 1%) iPSCs during differentiation to SCPCs. Additionally, we identified cell membrane opacity (Day 1) as a novel early intrinsic biomarker to predict SCPC differentiation efficiency at Day 10. We envision this label-free and optics-free platform technology can be further developed as a versatile, cost-effective process analytical tool (PAT) to monitor or assess stem cell quality and safety in regenerative medicine. With increasing aging population, T2DM subjects often suffer from chronic low-grade inflammation and are at higher risks of developing cardiovascular diseases. The paradigm in VII diabetes health management is shifting from glucose-centric monitoring to cardiovascular risk assessment. Complete blood count and C-reactive proteins are established inflammatory biomarkers but mechanobiology of diabetic immune cells remains poorly understood. As a proof of concept for impedance-based immunoprofiling in T2DM, we developed a novel clinically friendly label-free microfluidic workflow that uses a spiral inertial cell sorter (Dean Flow Fractionation) to achieve efficient size-based fractionation of leukocyte subtypes (neutrophils, lymphocytes, monocytes) for leukocytes impedance profiling from whole blood within 2 hours. The developed platform was validated in vitro and in vivo using diabetic mouse models (db/db), and in a clinical cohort of healthy individuals, pre-diabetes, T2DM without cardiovascular disease (CVD) complications (T2DM), and T2DM patients with history of cardiovascular complications (DM-CVD). Distinct differences in biophysical properties of leukocytes were observed with increasing diabetes severity, and pro-inflammatory phenotypes of DM-CVD neutrophils were also confirmed by bulk RNA sequencing. Principle component analysis (PCA) of impedance-based biophysical parameters also led to distinct clustering of T2DM patients with CVD, thus suggesting its potential as a novel point-of-care blood assay for CVD risk stratification and personalized medicine to improve diabetes care. Taken together, the developed impedance cytometry coupled with UMAP data analysis offers a rapid sample-to-answer solution for single cell biophysical characterization. With its multi-frequency single-cell profiling of up to 9 biophysical attributes, including both cell mechanical and electrical attributes, this method demonstrated a first-in-kind workflow that effectively identifies tumorigenic potential in SCPCs and stratifies cardiovascular risk in T2DM patients.