Study on dielectrophoretic behavior of a cell with sampling-based analysis for antimicrobial susceptibility test

Abstract

Antimicrobial susceptibility tests are used to evaluate antibiotics effectiveness for treating bacterial infection. These infections may contain drug resistant cells which can render certain antibiotics powerless. For effective patient treatment, a rapid, simple, and reliable susceptiblity test is critically needed. These susceptibility tests can be catagorized into either phenotypic or genotypic methods. Genotypic methods search for specific sequences known to cause the resistances. This method can be very rapid with high sensitivity and specificity, but requires skilled personnel and expensive equipment. Also, for many antibiotics, the genetic bases of resistance are highly complex or poorly understood. Phenotypic tests are based on detection of bacterial growth in the presence of antibiotics. Although very reliable and relatively cheap, standard methods require trained personnel and can be relatively slow, taking weeks for certain strains to culture for results. A rapid and simple method is needed to monitor the effective treatement of bacterial infections. Dielectrophoresis (DEP) has been previsouly used to separate live and dead cells due to having different electric properties, and can be an approach to shorten and simplify the phenotypic method. DEP behavior of a cell is associated with the electric field-induced polarization of a cell. The cell polarization is dependent on cell characteristics such as size, shape and electrical properties, as well as enviromental parameters including electric field frequency, electrode geometry, and medium properties. Therefore, to differentiate cells according to viability, one must understand the complex relationship between cell polarization and cell property change. However, current DEP studies are insufficient to provide a comprehensive evaluation of the multiple cell characteristic parameters. In this study, a novel sampling approach is presented to understand the DEP behavior of drug-treated cells. An algorithm using a sampling method is proposed to approximate diverse cell properties. With the algorithm, the cell properties are efficiently estimated from DEP experimental data. The estimation algorithm is evaluated in comparison to conventional approaches. The adavantages of the algorithm are shown by estimating various combinations of cell properties, which cannot be accomplished with conventional approaches. A sensitivity analysis using a sampling method is conducted to investigate the relation between the electric field-induced polarization of a cell and the variation of cell properties. For the experimental parameters of applied frequency and medium conductivity, the sensitive properties determining cell polarization are specified. Through the analysis, 4 transition condistions reversing cell polarization are found. By exploring the transition condition, the changes of each cell property are estimated. Based on the estimation algorithm and the sensitivity analysis, dielectrophoretic behavior of Mycobacterium bovis bacillus Calmette-Guérin (BCG) cells is studied in response to heat-killing and drug treatment of rifampin (RIF) and isoniazid (INH). As a surrogate marker of Mycobacterium tuberculosis, the BCG cell is chosen for a drug-susceptibility test. The experimental parameters are designed on the basis of the sensitivity analysis. The medium conductivity (σm) and the frequency (f) for a crossover frequency (fxo1) test are decided to detect the change of σm-fxo1 in conjunction with the drug mechanism. Statistical modeling is conducted to estimate the distributions of viable- and nonviable cells from the discrete measurement of fxo1. Finally, the parameters of the electrophysiology of BCG cells, Cenvelope and σcyto, are extracted through the sampling algorithm. The estimated change of the electrophysiological parameters due to heat- and drug treatment is supported by the SEM images of BCG cells. The proposed experimental method and the corresponding algorithm can be beneficial for a drug-susceptibility test for tuberculosis. By further developing the DEP characterization approach, the electrokinetic separation of viable and nonviable cells is studied. The cell behavior observed in an experiment is analyzed through a numerical simulation. In consideration of frequency-dependent DEP and AC electroosmosis (ACEO), the capturing and trapping behavior of cells are successfully analyzed. The optimal condition for separating viable- and nonviable BCG cells is found by understanding the electrokinetics of cells.