Here we present the enrichment of a cell population from a large sample volume using 3D carbon-electrode dielectrophoresis. There is a critical need for technologies capable of processing large sample volumes in order to extract few targeted particles. An immediate application is the isolation of the pathogenic population causing sepsis in patients. In clinical diagnosis, it currently takes at least 24 hours  detect the pathogens, as the patient sample needs to go through a number of steps, including culture. Here, we take initial steps towards isolating a diluted targeted particle population from a relatively large sample volume in a relatively short duration. We process samples featuring an increasingly diluted (10⁵-10²) yeast cell suspension. We also explore the limits on the flow rate that can be used towards minimizing the assay time.

The fabrication of the 3D carbon-electrode DEP (carbonDEP) device has been detailed by our group before¹. The experimental sample used here featured yeast cells suspended in a sugar solution of low conductivity. Dilutions of the growing culture with clean buffer was used to obtain targeted concentrations (10², 10³, 10⁴ and 10⁵ cells/ml).

Varying volumes of the yeast sample were flowed through the carbonDEP device at 10 µl/min using a syringe pump. In case of sample with cell concentration 10² cells/ml, a total volume of 4 ml was flowed through the carbonDEP device, as we were targeting to capture minimum of 200 cells. The carbon electrodes were stimulated with a sinusoidal signal of magnitude 20 Vpp and 100 kHz frequency to implement a positiveDEP trapping force of all yeast cells. After processing the cell suspension, the inflow was changed to buffer media to wash the trapped cells using 100 μl of buffer. At that point, the electrodes were turned off to release the trapped cells and elute them for their retrieval at the end of the device. A total of 34 fractions (20 µl each) were retrieved from the chip and were divided into 1) sample, where yeast cells were retained from the sample solution (fractions 1-25); 2) Washes, where trapped cells were washed with  buffer (fractions 26-30); and 3) Eluates, where trapped cells were eluted upon turning off the electric field (fractions 31-34 ). A hemocytometer, with a limit of detection of 10⁴ cells/ml, was used to obtain cell concentration. In order to quantify the impact of flow rate on the trapping efficiency, experiments featuring a cell concentration of 10² cells/ml were performed with flow rate of 20 μl/min and 30 μl/min.

The results from the experiments featuring different cell concentrations are shown in Figure 1. The capacity of this particular chip design is 10⁴ cells/ml judging from the saturation point achieved in all cases. The cases we are most interested on are when the concentration is below 10³ cells/ml. In this case, an enrichment up to 5638.7% can be observed. In other words, we demonstrate the capability of isolating as few of 750 cells originally present in 4 ml of media in just 20 μl of clean buffer. This result is promising as this technique can be employed in enhancing the sample pre-treatment and cell enrichment steps in the clinical process of pathogen detection. Time of assay is of crucial importance in clinical diagnosis. Figure 2 shows results when attempting to enrich a sample with 10² cells/ml original concentration at different flow rates. At 10 μl/min enrichment is 5638.7% but the assay takes up to 7 hours. As expected, the degree of enrichment is inversely proportional to the flow rate in the channel. However, an increase of flow rate, from 10 to 30 μl/min, drastically reduces the time assay from 7 hours to just 2 hours, without a significant impact on the amount of enrichment (2915%) which is still 350 cells out of 1000 cells originally present in 4 ml. Thus a 3 time increase on the flow rate only leads to a decrease of 51% in the enrichment.

Ongoing work is on using these experimental results to validate a computational model of our device to facilitate its optimization. Of special interest is to understand and quantify the interaction between the DEP and drag forces on a bacterial population as the chip design and different experimental parameters vary. The ultimate goal is to be able to process up to 5 ml of sample in less than two hours.

References:

1.  R. Martinez-Duarte, P. Renaud, and M. J. Madou, Electrophoresis, 32, 2385–92 (2011) http://www.ncbi.nlm.nih.gov/pubmed/21792991.