3D-electrode integrated microsieve structure as a rapid and cost-effective single neuron detector

The first technique to achieve in vitro cell culture was developed by Ross Harrison in the first decade of the twentieth century. Beginning with this invention, animal and human cell cultures have a crucial role in the life sciences, including cancer research, stem cell studies, and pharmacology, to name just a few.¹ Providing elaborated understanding of intra- and extracellular mechanisms and cellular interactions in their specific tissue formations are key to innovative health care applications. In the past three decades, achievement of cell culturing on microelectromechanical systems (MEMS)-based devices, such as early developments of microelectrode arrays (MEAs), micro total analysis systems, and lab-on-a-chip, offers a great number of interactive analysis techniques to unravel tissue physiology. This type of integrated devices enables far more efficient analytical methods than conventional cell culture systems.² These platforms included originally only the development of two-dimensional (2D) cell integration but recently opened up to the advancement of three-dimensional (3D) microenvironments for the purpose of in vivo-like cell culture requirements, today called organ-on-chip. These microsystems have the general advantages of high precision, small amount of sample and reagent usage, cost- and time-effectiveness, and combination of different materials having divergent physical and chemical properties to make more convenient milieu for cells like those found inside the body.^2,3

Microsieve-based electrode arrays, inspired by one of the most common types of MEMS-based platforms, which is the planar MEAs, used to interrogate electrogenic cells, found in the heart or the nervous system, have a great appeal. Microsieve-based electrode arrays are advantageous over standard microtiter plates or a simple planar spotted protein or cell-binding microarrays because of their direct means of digitalizing the readout of signals at the single cell level. While all these devices combine considerably high parallelization in cell assay technology, only electrical readouts are completely label-free.^4,5 In the case of microsieves, hydrodynamic flow can be used to trap a single cell inside a 3D micropore.⁶ When a cell is trapped into the pore, hydrodynamic resistance increases, and this prevents the trapping of other cells on the same location of the sieve. This provides high spatial control over cell trapping.⁷ Previously, in our group, the fabrication procedure of silicon microsieves was developed by using advanced silicon micromachining techniques.⁷ Next, we developed a passive pumping method for the parallel trapping of single neurons onto a 3D microsieve, eliminating the active pumping devices, reducing footprint, and decreasing cell damage.⁸ Last but not least, we further developed a simple replica molding protocol with a subsequent laser ablation step to translate the concept in a more cost-effective polymer platform for cell culture experiments.⁹ Figure 1 presents this microsieve array made of NOA81 [Fig. 1(a)] in our group by focusing on the structure of a microsieve in the top view [Fig. 1(b)]. Whereas in the silicon-based platform, electrodes, coupled to the contact lines of an MEA, can be integrated with relative ease into the sidewalls of the 3D micropores of the device to make an electrophysiological measurement of neurons possible, this process is not straightforward when working with polymers as a substrate material. Therefore, we demonstrated that also dynamic optical readouts can benefit from the microsieve concept by utilizing Ca²⁺ imaging in SH-SY5Y cell cultures.⁹ 

In this study, we investigated if this simple polymeric microsieve structure could be utilized as a disposable add-on for a reusable 3D electrode array by carrying out a passively aligned click-on hybrid assembly step (applied for a provisional patent in the US with the number of 63/056809 on July 27, 2020). Since the use of integrated silicon micromachining and thin-film technology for the fabrication of electrically functionalized microsieves was a major challenge and is still very expensive at the current scale of device production also, the fundamental biomedical research in this field of neuroscience is severely hampered. While still working toward the integration of electrodes for neuroelectrophysiological recordings in our polymer-based microsieve that uses a transparent optical adhesive, NOA81 (Norland Products, Inc.), as a material,⁹ we present here an alternative route for its utilization in an impedance spectrometry-assisted neuropharmacological screening assay. Using microsieve arrays as a 3D cell culture substrate, unfortunately, not all the pores are always invaded by cells, although the capture ratio was as high as ∼80%.⁹ Therefore, it would be an advantage to know the exact distribution of captured cells throughout the entire pore layout of the sieve in a real-time manner, which will enhance statistical relevance of these biological experiments. However, optical monitoring of each of the several 100 microsieve pores is a labor-intensive way to say go/no-go for further analysis in long-term cultures as required in the culture of neurocircuits derived from stem cells. Hence, in this study, the concept and feasibility of a rapid and label-free platform to monitor the cell placement distribution in a real-time manner by containing 3D electrodes is introduced. In this platform, the polymer-based microsieve part will be pluggable and cost-effective, which means that it can complement a relatively expensive 3D electrode apparatus used for the impedance measurement in a high-throughput and robust cell culture microenvironment. The specifically designed impedance sensors will then be reusable for the large number of microsieve cultures that are generally needed in a biological study without any rigorous cleaning steps by means of the elimination of cell-electrode touching interfaces.

The proposed device has two pluggable parts: (i) Disposable polymer-based microsieves and (ii) reusable 3D electrodes [Fig. 2(a)]. Cells are trapped in 3D microsieve pores by passive pumping, and biological assays can be achieved. In our original design, the 3D microsieves have a square top opening of 20 μm edge size and a round bottom opening of about 3 μm, respectively, connected by a funnel-like channel transforming from an inverted pyramidal shape into a conically tapered hole.⁹ For reducing complexity in our model, we simplified this structure in the polymer microsieves by a prismatic shape with a top and bottom opening of 20 and 3 μm, respectively. In addition, the 100 μm thick polymer microsieve contains holes from the back side of 20 × 20 μm² square openings to plug-in the 3D electrodes during assembly [Fig. 2(b)].

Biological cells are dielectric particles. Therefore, they change the impedance of a solution in which they are immersed. At low frequencies, only the cell diameter determines the amount of variation of an impedance signal. By using this phenomenon, Coulter invented the Coulter Counter, offering a label-free and real-time cell counting principle.¹⁰ This method was further developed, and it was shown that at high frequencies, cell electrical properties affect the impedance response.¹¹ This provides a method to characterize a cell by its electrical properties when it flows through a set of electrodes, like in optical flow cytometry. This technique was invented and named as impedance flow cytometry (IFC) by Gawad et al.¹¹ The IFC has been used for the analysis of different types of cells, including cancer cells, stem cells, and bacteria, in lots of research studies, such as drug effect analysis, diagnosis, and cell viability and differentiation.^12,13

Our platform will work based on this powerful technique to track cell movement and detect cell position on microsieve arrays by using the simple configuration presented in Fig. 3(a) and eliminating labor-intensive microscopy effort. When a cell moves into or out of a microsieve pore, there will be an impedance change [Fig. 3(b)]. The trend of the impedance change will differ based on the movement direction. Therefore, if a cell goes inside microsieve and then move outside of it, the impedance change will have a characteristic fingerprint [Fig. 3(b)]. This will be determined by means of real-time measurements. As a result of these measurements, go/no-go mapping of a microsieve by a microsieve is determined by this smart and rapid platform [Fig. 3(c)].

Biological cells deflect electric field lines between two electrodes at which a cell is interleaved. The quantification of this deflection can be carried out by impedance variation. In order to detect cells in between electrodes via impedimetric measurements, the variation amount in impedance, caused by cell existence, should be measurable. In this study, finite element modeling (FEM) of the proposed system was accomplished to see if it is feasible to achieve this type of electrical analysis. In order to make quantitative interpretation, electric field and impedance of a thin-film sidewall and 3D electrodes were compared.

FEM enables solving complex equations for complex structures numerically. In our structure, the electric field is not uniform between electrodes due to the presence of insulator layers. Therefore, the numeric solution is the most accurate way to evaluate the electric field in our proposed system. Electric Currents’ interface of the AC/DC module of COMSOL Multiphysics 5.5 in the frequency domain with current conservation, electrical insulation, terminal in the current type, and ground boundary conditions was used to model the electrical field strength in a pore. This interface of COMSOL solves current conservation equation based on Ohm's law and neglects inductive effects. Domain equations are as follows:¹⁴ 

where ɛ₀, ɛ_r, and σ are the electrical permittivity of the free space, relative permittivity, and conductivity, respectively. J is √−1, and ω is the angular frequency of the applied signal. V, E, and D denote the electric potential, electric field, and displacement, respectively.¹⁴ ∇ is the gradient operator.

Electrodes were defined as perfect conductors at the boundary of the microsieves' pores and have infinite thickness to decrease memory need and calculation time. Inner and outer boundaries of microsieves' pores were modeled as thin-film sidewall and 3D electrodes, respectively. Solution domain electrical properties were chosen like phosphate-buffered saline (PBS), which has 1.4 S/m conductivity with the permittivity constant of 80.¹⁵ The microsieve material was modeled as polydimethylsiloxane, chosen from the material library, having a permittivity constant of 2.75 and a conductivity of 10⁻¹⁶ S/m. By applying 1A current and analyzing the system response at 10 kHz in the frequency domain, 2D and 3D solutions of sidewall and 3D electrode integrated microsieves were carried out. The system was simulated at such a relatively low frequency since it is enough to determine cell existence. Here, cells were modeled as an air-filled sphere having a 5 μm diameter. Cell coordinates [given in Figs. 4(b) and 4(e)] were kept constant for x and z directions, while they were swept in the y direction from 30 μm to 120 μm with a 2 μm step size at both 2D and 3D simulations. By observing convergence plots, a free triangular meshing style was chosen with 1.26 μm, 2.52 nm, and 1.1 as the maximum element size, minimum element size, and element growth rate, respectively, in 2D simulations. In the 3D simulation, free tetrahedral meshing was used with 4.41 μm, 189 nm, and 1.35 as the maximum element size, minimum element size, and element growth rate, respectively. The software solved the system for the spatial distribution of electric potential. Electric field and impedance quantification were obtained during postprocessing by using the Results interface.¹⁴ 

2D and 3D simulations of thin-film sidewall and 3D electrode integrated microsieves were carried out. Table I presents mesh properties, number of degrees of freedom values, and solution time specifically to the simulation. Impedance results show that in terms of the impedance variation trend, there was no difference between 2D and 3D simulations (Figs. 4 and 5, respectively). However, the values of impedances [Figs. 4(c) and 4(f)] did not model the impedance magnitude of a real structure. Therefore, comparisons between thin-film sidewall and 3D electrode structures were carried out via 3D simulations even solution times were significantly longer.

Figure 5 presents the electric field distribution through the center line without the cell [Figs. 5(a) and 5(c)] and impedance variation [Figs. 5(b) and 5(d)] for thin-film sidewall and 3D electrodes, respectively, while a cell, having a 5 μm radius, moves from inside to outside of a microsieve pore. Electric field had more strength (∼3.4 times) and increasing trend in 3D electrode design, while it decreased in the sidewall type since the distance between electrodes increases in sidewall electrodes, but the insulator thickness decreases even if the distance between electrodes is constant in 3D ones.

In order to determine the amount of the impedance change due to existence of a cell as an indicator for cell detection efficiency, the following formula was used:¹⁶ 

where |z| denotes impedance magnitude in this equation.

The impedance magnitude variation of thin-film sidewall electrodes [Fig. 5(b)] was calculated as 3.5%, having 2.7 kΩ baseline impedance, while the one calculated for 3D electrodes was 3.4% and baseline impedance (58.9 kΩ) was still in the measurable range [Fig. 5(d)]. These results proved that without compromising detection efficiency, proposed 3D electrodes can be integrated with microsieves in this pluggable, rapid, real-time, and label-free platform. In the integration, the usage of lithographically aligned electrodes as a master mold for backside openings ensures minimizing a potential air gap between the electrode wall and the polymer enclosing the 3D pore in the microsieve.

This study presents a feasibility study of a platform, integrating 3D electrodes and microsieves in a pluggable manner, based on FEM simulations. Simulation results indicate that the proposed system will work having the same detection efficiency with the previously studied thin-film sidewall electrodes, although it has a thick insulator layer between electrodes and cells. This system will give real-time results without using any labeling methods and become cost-effective due to reusability of electrodes. As an outlook, we will focus on the development of the fabrication method for this pluggable platform. Next to fabrication, simulation results will be verified via biological cell analysis.

This work has received funding from the European Union’s Horizon 2020 research and innovation programme H2020-FETPROACT-2018-01 under Grant Agreement No. 824070 and the Eurostar’s MINDMAP project under Grant Agreement No. E113501.