Proceedings of The Physiological Society

Physiology 2012 (Edinburgh) (2012) Proc Physiol Soc 27, PC237

Poster Communications

Patterning human dopaminergic neurons on photolithographically engineered silicon dioxide wafers functionalized with pre-adhered HEK293 cells

M. A. Hughes1, A. Bunting2, A. F. Murray2, M. J. Shipston1

1. Centre for Integrative Physiology, University of Edinburgh, Edinburgh, United Kingdom. 2. School of Engineering, Institute for Integrated Micro and Nano Systems, University of Edinburgh, Edinburgh, United Kingdom.


  • A: Schematic illustrating photolithographically generated parylene pattern (black, 100μm diameter node with 300μm 'cross hair'. B: Following application of HEK 293 cell suspension. C: Four days following secondary co-culture with pre-differentiated LUHMES cell suspension.

The ability to interface neurons with silicon semiconductors remains a significant challenge but promises to allow bidirectional control of neural circuits and computers. A fundamental pre-requisite is the ability to define the spatial organization of patterned cells on a silicon chip. Parylene-C, a polymer used commercially to coat printed circuit boards, has previously been used to pattern primary murine hippocampal cells (1). Photolithographically defined arrays of parylene-C (on a silicon dioxide background) are activated by immersion in fetal calf serum; after which cells show preferential adhesion to parylene and repulsion from bare SiO2. However, the underlying patterning mechanism is as yet uncharacterized. Moreover, these cultures are restricted by rapid glial overgrowth which overwhelms patterned neurons. We sought to pattern a purified source of neurons, questioning whether functional neurons can remain viable in this isolated ‘on-chip' context or whether they require a supporting cell substrate. The LUHMES (Lund Human Mesencephalic) cell line was chosen as a source of homogenous post-mitotic neurons. Two indices were derived to assess cell patterning. Parylene Adhesion Index (PAI) was calculated by dividing the surface area of cell material on parylene by total surface area of parylene within a given Region Of Interest (ROI). Each ROI consisted of an iteration of the parylene geometric pattern (figure A) surrounded by a square area of SiO2. A SiO2 Repulsion Index (SRI) was calculated by dividing surface area of cell material on SiO2 by total area of SiO2 in a given ROI, and subtracting the result from 1. ‘Perfect' cell patterning on parylene results in a PAI of 1 (complete cell coverage of all parylene) and SRI of 1 (complete absence of cell material from SiO2). Data reported are means ±S.D. Both undifferentiated and pre-differentiated LUHMES failed to adhere, nor show any morphological signs of differentiation, when cultured directly on-chip (undifferentiated LUHMES: PAI 0.02±0.05, SRI 1.0±0.01; differentiated LUHMES: PAI 0.0±0.0, SRI 1±0.001). We therefore sought a different cell type with which to pre-pattern parylene areas. HEK 293 cells pattern with high fidelity (PAI 0.46±0.2, SRI 0.98±0.01, figure B). Subsequent application of pre-differentiated LUHMES cells resulted in their adhesion to pre-established HEK 293 cell clusters (figure C). Moreover, this co-culture environment promoted morphological differentiation with neurites extending between islands of adherent cell somata. HEK 293 cells appear to fulfill a role analogous to glia, dictating cell location and generating a pro-neuronal niche. By refining our protocol, we aim to improve the resolution of this patterning platform and to gain control of the direction of neurite outgrowth.

Where applicable, experiments conform with Society ethical requirements