Colloids and Surfaces B: Biointerfaces 52 (2006) 14–21 Interfacing SH-SY5Y human neuroblastoma cells with SU-8 microstructures Ze-Zhi Wu a,b , Yiping Zhao a , William S. Kisaalita b, a Department of Physics and Astronomy, Faculty of Engineering, The University of Georgia, Athens, GA 30602, USA b Cellular Bioengineering Laboratory, Faculty of Engineering, The University of Georgia, Athens, GA 30602, USA Received 24 February 2006; received in revised form 30 April 2006; accepted 1 June 2006 Available online 6 June 2006 Abstract Microwell structures were fabricated using SU-8 photoresist for engineering a quasi-three-dimensional (quasi-3D) microenvironment for cultured neuronal cells. SH-SY5Y human neuroblastoma cells were successfully integrated into microwells of a nominal diameter of 100 m, with or without 10-m wide microchannels connecting neighboring microwells, in an aspect ratio (ratio of structure depth over width) of approximately 1. With the help of polyethylene glycol stamping and laminin coating, a neuronal-like network was achieved by integrating populations of SH-SY5Y cells with a microwell network pattern. Resting membrane potential establishment was evaluated with confocal microscopy and the potentiometric fluorescent dye tetramethylrhodamine methyl ester. It was found that the intra/extracellular fluorescent intensity ratio (R) was 2.4 ± 1.4 [n (number of cells measured) = 112] for SH-SY5Y cells on flat SU-8 substrates on day 5 into differentiation, which was not significantly different from the ratio on day 13 into differentiation, 2.0 ± 1.8 (n = 104) (P > 0.05). For cells in the microwell network structures, R was 4.8 ± 4.7 (n = 51) and 3.9 ± 3.2 (n = 62) on days 5 and 13 into differentiation, respectively (P > 0.5). Cells within the network structures had higher R ratios than on flat substrates, for either day 5 or 13 into differentiation (P < 0.01). These results demonstrated that the well network structures, or topographically patterned substrates, were more suitable formats for promoting SH-SY5Y cell resting membrane potential establishment than flat substrates, suggesting the potential to control cellular function through substrate topography engineering. © 2006 Published by Elsevier B.V. Keywords: Photolithography; Quasi-three-dimensional; Resting membrane potential; Potentiometric; Confocal microscopy 1. Introduction In the fields of tissue and cellular engineering, one of the most rapidly advancing sectors is that of cell-based microde- vices [1,2]. For these devices to serve their purposes, it has been expected that the response of the integrated cells should mirror physiologically or pathophysiologically what happens in vivo. In the in vivo condition, cells of a tissue compart- ment normally live in an extracellular matrix (ECM) meshwork with three-dimensional (3D) and high aspect ratio topographical textures. Three-dimensionality provides cells with characteris- tic topographical cues in the cellular microenvironments and thus enables cells to differentiate into specific phenotype and Corresponding author. Tel.: +1 706 542 0835; fax: +1 706 542 8806. E-mail addresses: zezhiwu@hotmail.com (Z.-Z. Wu), zhaoy@physast.uga.edu (Y. Zhao), williamk@engr.uga.edu (W.S. Kisaalita). maintain specific functions that are usually impossible under two-dimensional (2D) culture conditions [3,4]. To this end, a simple 2D flat substrate may not be regarded as accurately rep- resenting the in vivo situations and consequently may not be an ideal format for cell-based microdevices. Microfabrication techniques offer unique approaches for pat- terning cells and engineering cellular microenvironment. In the field of neuron-based microdevices, microwell structures have been fabricated for patterning neuronal networks or positioning neurons on top of embedded electrodes, with glass [5], silicon [6,7], polyester photoresist [8], polymer elastomer [9,10] and agarose gel [11]. However, in most of these studies, the scale of the networks and the number of addressable cells are small due to the limited number of probing electrodes. Therefore, the systems are not well suited for high-throughput screening with microscopic systems. Additionally, most of these structures were characterized by low aspect ratios (less than one) or depths (of only 1–2 cell diameters) [5–8]. In a few cases where a large 0927-7765/$ – see front matter © 2006 Published by Elsevier B.V. doi:10.1016/j.colsurfb.2006.06.001