# Also in the Article

Selection of buffer conductivity and working frequency of the DEP-D unit

Procedure

During typical practice, cells are attracted towards a stronger electric field by positive DEP (pDEP), or they are repelled towards a weaker electric field by negative DEP (nDEP). Theoretically, there is a crossover frequency (fcross) at which DEP force experienced by the cells switches from pDEP to nDEP regimes. At fcross, DEP force is equal to zero since Re(fCM) of cells is zero28. LCBs are used to see this switching24,29. High conductivity buffers (HCBs) are defined as buffers having conductivity higher than that of cell cytoplasm24. When HCBs, such as phosphate buffered saline (PBS, σ = 1500 mS/m for 1X concentration30), are used, only nDEP occurs since cells experience only negative Re(fCM) values in HCBs, therefore pDEP is not an option for these buffers. We used relatively HCBs at a conductivity close to the estimated cytoplasmic conductivity of wild type cells (e.g., σ = 200 mS/m) to enable switching between nDEP and pDEP becomes less sensitive to small variations in cell cytoplasmic conductivity. MDR cells having higher cytoplasmic conductivity than that of wild type cancer cells as well as the conductivity of HCB were trapped in DEP cages by pDEP in this LOC system. On the other hand, wild type cells having cytoplasmic conductivity not higher than HCB were not trapped. By this approach, a threshold was created for selection of MDR cells by the conductivity of an HCB. Therefore, the range for working frequencies of DEP-D unit to select MDR cells was enlarged.

To better explain relatively HCB utilization approach, a case study was carried out via MATLAB simulations for Re(fCM) of K562/wt and K562/imaR cells. The size and electrical properties of cells, previously obtained by Labeed et al. and Demircan et al.17,25 and presented in Supplementary materials (Sect. 4, Table S1), were used in a single shell cell modeling31. LCB and HCB had conductivities of 2.5 mS/m and 200 mS/m, respectively. In LCB, fcross of K562/wt cells, having 210 mS/m cytoplasmic conductivity, was 43.7 MHz (Fig. 3a). When the cytoplasmic conductivity of a cell was changed to 215 mS/m, Re(fCM) value became 0.008 at the same frequency. In case 43.7 MHz was chosen as working frequency for obtaining zero Re(fCM) for K562/wt cells, a pDEP force could be exerted on a cell instead of zero DEP force for even a 2.4% change in cytoplasmic conductivity. If the buffer velocity, i.e. flow velocity, $vm,$ is low enough, trapping of wild type cells becomes inevitable at this frequency. To prevent this, higher frequencies can be chosen to stay on the safe side. However, Re(fCM) of K562/imaR cells also decreased with increasing frequency (Fig. 3a, blue line). If the buffer velocity is high enough, the trapping of K562/imaR cells becomes impossible, which is not wanted. The frequency range in the case of LCB was determined to be between 43.7 MHz and 59.8MHZ at which K562/imaR cells had Re(fCM) values of 0.25 and 0.10, respectively, to trap K562/imaR cells by pDEP. If the same comparison was achieved in a relatively HCB, having conductivity close to the cytoplasmic conductivity of K562/wt cells (σ = 200 mS/m), the frequency interval was defined between 1.4 MHz and 45.6 MHz. Additionally, the increase in Re(fCM) of wild type cells was two-fold (from 0.008 to 0.016) and the rate of change in Re(fCM) values of both cell types was low with reduced slope in Re(fCM) plot (Fig. 3b).

Single shell cell modeling to obtain Re(fCM) characteristics of cells having cytoplasmic conductivities of 210, 215, referring K562/wt cells, and 374 mS/m, referring K562/imaR cells, in LCB (conductivity, σ = 2.5 mS/m) (a) and relatively HCB (σ = 200 mS/m) (b). (c) Selection of the working frequency (8.6 MHz) for K562/wt and imaR cells in relatively HCB (σ = 200 mS/m). (d) Selection of the working frequency (6.2 MHz) for CCRF-CEM/wt and doxR cells in buffers (at different conductivity values: 110, 125, and 160 mS/m).

The level of MDR (i.e., high-level laboratory or clinically relevant models) may change the electrophysiological properties of cancer cells32. We have previously reported that high-level laboratory MDR model, K562/imaR cells, have 1.8 times higher ion concentration than that of K562/wt cells on average25. According to the literature, K562/wt cells have cytoplasmic conductivity in the range of 210–240 mS/m17. By using this reference value in the direct relation between cytoplasmic conductivity and ion concentration33, the estimated cytoplasmic conductivity values of K562/imaR, CCRF-CEM/wt, and CCRF-CEM/doxR cells were listed in Table Table2.2. The working frequency was selected as 8.6 MHz for K562 cells since at this frequency, K562/imaR cells have maximum Re(fCM) while K562/wt cells have the lowest Re(fCM) in a buffer at a conductivity of 200 mS/m (Fig. 3c).

Radii25 and estimated cytoplasmic conductivity ranges of K562/wt, K562/imaR, CCRF-CEM/wt, and CCRF-CEM/doxR cells used in the determination of working frequency of the DEP-D unit in HCBs.

In our previous study25, no statistically significant difference was reported between the cytoplasmic ion concentration of CCRF-CEM/wt and CCRF-CEM/doxR cells as clinically relevant MDR model. Using our LOC system with appropriate HCB, we hypothesized that if proper buffer conductivity and frequency are determined, statistically different trapping behavior between CCRF-CEM/wt and CCRF-CEM/doxR cells can be obtained. Therefore, the dielectrophoretic responses of CCRF-CEM/wt and CCRF-CEM/doxR cells were examined in 3 different buffers with 110, 125, and 160 mS/m conductivities. The conductivity values were identified based on the minimum expected cytoplasmic conductivity of CCRF-CEM/wt and CCRF-CEM/doxR cells (185 mS/m and 229 mS/m, respectively, Table Table2).2). The frequency was selected as 6.2 MHz, at which CCRF-CEM/doxR cells have maximum Re(fCM) value, which ensures the highest pDEP force on resistant cells to trap them on electrodes (Fig. 3d).

Note: The content above has been extracted from a research article, so it may not display correctly.

Q&A