Neuron adhesion on a silicon chip probed by an array of field-effect transistors.

TLDR: It is shown that the electrical properties of cell adhesion can be probed by taking advantage of the neuron-silicon junction, and the resulting extracellular voltage profile is evaluated from the modulation of the source-drain currents.

Abstract: Adhesion of neurons to surfaces in biological tissue (eg, to glia cells, in synapses) or in cell culture (on glass or silicon) may seriously change the features of the electrical signals Transmembrane ion currents have to flow along the narrow cleft between membrane and surface They give rise to a drop of voltage that may affect, in turn, the ion channels [1] In the present paper we show that the electrical properties of cell adhesion can be probed by taking advantage of the neuron-silicon junction [2,3] We place the neuron on an oxidized surface of silicon with an integrated array of field-effect transistors with open metal-free gate oxide [cf Fig 1(a)] An ac voltage is applied to the neuron This voltage couples into the cleft through the conductance and the capacitance of the membrane The resulting extracellular voltage profile is evaluated from the modulation of the source-drain currents Neuron-silicon assembly— A view of the silicon chip is shown in Fig 2(a) The 16 metal-free transistors are arranged in two rows Their distance is 52 mm The area of the transistor channels is 18 mm 3 18 mm Zones of recessed oxide separate the transistors (They give rise to a modulation of the surface profile by about 250 nm) Each drain has its own lead All transistors share a common source Care was taken that the geometric aspects and the electrical properties (bias voltages, sensitivity, and threshold voltage) of the transistors were similar to the larger transistors, used in previous experiments [2,3] The chip s10 mm 3 10 mmd was stuck on a plate with copper leads After wedge bonding a Plexiglas chamber was attached We cleaned the exposed surface of the chip by hot basic hydrogen peroxide The gate region was coated with poly-L-lysine The chamber was filled with culture medium (Leibowitz-15) Retzius cells from the segmental ganglia of the leech hirudo medicinalis were isolated [4] and treated with collagenase and dispase before the experiment Then a neuron was placed on the array as illustrated in Fig 2(b) We kept the bath on ground potential using an AgyAgCl electrode [cf Fig 3(b)] A bias voltage of VES › 230 V was applied between electrolyte and bulk silicon The source was kept at bulk potential The drain-source voltage was VDS › 220 V, which caused a source-drain current of about ID › 50 mA A change DVES › 10 mV of the voltage between electrolyte and source induced a current modulation of DID › 2045 mA as checked before the attachment of a neuron The neuron was contacted with a AgyAgCl electrode by fusing the membrane with a patch pipette filled with electrolyte [3,5] The intracellular potential was held at ‐50 mV Cell impedance—We superposed ac stimuli of 15 mV amplitude to the holding potential No action potentials were elicited under these conditions We measured the complex amplitudes V P of the voltage and I P of the current at the head of the pipette [Fig 3(a)] using two lock-in amplifiers A Nyquist plot of the impedance Z › V P yI P of an attached neuron is shown in Fig 3(b) in a frequency range from 1 to 15 000 Hz The double