Calcium imaging

Cells cultured in microfluidic devices as described above were washed with ACSF (130 mM NaCl, 10 mM Na-HEPES, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 30 mM Glucose, pH 7.4; supplemented with 1 mM probenecid for imaging Fura-2). For ratiometric calcium imaging, DRG or DH neurons were loaded with 2.5 µM Fura-2 dissolved in ACSF for 1 hour in the incubator with occasional rocking and washed before imaging.

An inverted fluorescent microscope (Nikon Eclipse TE200) with 20X Plan Fluor, 0.5 NA objective was used for all live imaging experiments. For imaging GCaMP, a FITC filter was used and cells were excited at 488 nm wavelength using a Random Access Monochromator (PTI) with Xenon Short Arc 75 W lamp, while for imaging Fura-2 at 340 and 380 nm wavelengths, in both cases fluorescence images were acquired by Hamamatsu Orca FLASH 4.0 at 2 frames per second using EasyRatioPro (PTI) software. To deliver electrical stimuli, a Neurolog NL800A stimulus isolators driven by Neurolog signal generators (Digitimer) configured to deliver biphasic stimuli was connected to coiled Pt wires submerged into each wells on the side of the corridor in the Periphery compartment. It was empirically determined (data not shown) that 5 Hz biphasic pulses (2 ms for each phase) are optimal for exciting the majority of the DRG axons and a train of 10 pulses at 2x threshold amplitude (i.e. 40–80 μA) elicit a robust response in the DH neurons (concentrations of DMSO up to 0.2% were tested as a control over an imaging period of up to 3 hours). The train of pulses was delivered three times for each drug combination with at least 2.5 minutes in between the trains and at least 10 minutes after each drug application or wash sequence.

Image analysis was done in FIJI. To generate an image of responder cells, an average image of 10 frames after the stimulus was divided (or subtracted) by an average of 10 frames before the stimulus to generate ΔR/R₀ (or ΔF) images (macro used for the analysis are deposited at https://github.com/RaoufLab/NV-ImageJ-Scripts), where ΔR/R₀ is the change in the fluorescence intensity ratio, F₃₄₀/F₃₈₀ for Fura-2 experiments (or ΔF/F for GCaMP)_. For ease of visualization “Green Fire Blue” LUT is used throughout the paper with the calibration bar next to each image. For DH neurons, cells that responded to electrical stimulation of the periphery compartment in the absence of any blockers were selected for further analysis. The mean change in ΔR/R₀ for each ROI was then calculated for each cell in the presence of the blockers. If the ΔR/R₀ was reduced by more than 50%, the cell was considered as responding to the blockers. Cells in which the baseline had significantly changed during the experiment were discarded. For DRG axons, a threshold mask was applied to the ratioed images of axons during electrical stimulus with no blockers, and the same threshold mask was then applied to the images from the same field responding to electrical stimulus in the presence of the blockers. The area of responding axons (where all pixels with intensity great than 50 a.u.) was compared and quantified to assess proportion of axons that were inhibited by the blockers.