Transient Magnetothermal Neuronal Silencing Using the Chloride Channel Anoctamin 1 (TMEM16A)

Abstract

Determining the role and necessity of specific neurons in a network calls for precisely timed, reversible removal of these neurons from the circuit via remotely triggered transient silencing. Previously, we have shown that alternating magnetic field mediated heating of magnetic nanoparticles, bound to neurons, expressing temperature-sensitive cation channels TRPV1 remotely activates these neurons, evoking behavioral responses in mice. Here, we demonstrate how to apply magnetic nanoparticle heating to silence target neurons. Rat hippocampal neuronal cultures were transfected to express the temperature gated chloride channel, anoctamin 1 (TMEM16A). Spontaneous firing was suppressed within seconds of alternating magnetic field application to anoctamin 1 (TMEM16A) channel expressing, magnetic nanoparticle decorated neurons. Five seconds of magnetic field application leads to 12 s of silencing, with a latency of 2 s and an average suppression ratio of more than 80%. Immediately following the silencing period spontaneous activity resumed. The method provides a promising avenue for tether free, remote, transient neuronal silencing in vivo for both scientific and therapeutic applications.

Keywords: anoctamin 1; chloride channel; magnetogenetic; magnetothermal; remote silencing.

FIGURE 1

Magneto-thermal silencing scheme. (A,B) Illustrate the principle of magnetogenetic silencing. Magnetic nanoparticles (brown) were encapsulated in polymer (PMMA). Fluorescent dye (Alexa Fluor 647) molecules were bound to Neutravidin molecules (green; 5:1 molar ratio). The Neutravidin molecules were covalently attached to the nanoparticle polymer layer. These particles were then attached to the neuronal cell membranes via biotin-avidin bonding formed with the membrane attached biotinylated IgM antibodies. When exposed to alternating magnetic fields (AMF), the particles heat, opening the temperature gated membrane Ano1 channels (B). This causes an influx of Chloride ions, leading to membrane hyperpolarization.

FIGURE 2

Methodology for spike estimation and latency calculation. (A) (Top) Bleach corrected, normalized GCaMP6f data. (Middle) Predicted action potential events from GCaMP6f trace (black sticks) and (bottom) the overlay of regenerated trace obtained from convolving predicted action potential events with single AP calcium peak (black, broken) with the normalized GCaMP6f trace. (B) The average (n = 3 peaks) signal from three smallest peaks recorded from the same neuron in a single recording. (C) Histogram of the residual of the GCaMP6f trace and the regenerated trace in (A), fitted with a Gaussian curve. The sigma of the fit was 1.42 ± 0.05. (D) Extension of (A), showing a horizontally magnified view of GCaMP6f peak fitting. Convolution of the two estimated APs (black bars) with the average single peak profile (B) gives the reconstructed GCaMP6f peak (black broken). (E) Representative numerical integration (top, black) of the GCaMP6f plot (bottom, green). Suppression in firing is indicated by a reduction in slope of the integration plot. Dotted lines show linear fits of three distinct sections of the trace [red: before suppression (left of 1), green: during suppression (between 1 and 2), blue: after resumption, following suppression (right of 2)]. Slopes of these lines give average rate of Ca²⁺ influx, during the indicated periods. The points of intersection of these lines give the times corresponding to the beginning (1) and ending (2) of suppression.

FIGURE 3

Ano1/TMEM16A expression does not alter GCaMP6f peaks. (A) Ano1/TMEM16A expression in rat hippocampal neurons, visualized with the mCherry tag (red) and GCaMP6f fluorescence (green) overlay (right). (B) Average single gCAMP6f peaks recorded in Ano1/TMEM16A^+/- neurons (top and bottom, respectively) at 37°C, respectively (n = 4). (C) GCaMP6f peak rise times for Ano1/TMEM16A^+/- neurons were found to be 0.25 ± 0.03 s and 0.28 ± 0.05 s, respectively. Peak decay half-life times for Ano1/TMEM16A^+/- neurons were 0.27 ± 0.01 s and 0.29 ± 0.06 s, respectively. No significant change in peak characteristics was found (n = 4, all cases). Color scheme for Ano1/TMEM16AC^+/- data follows the scheme used in (B).

FIGURE 4

Thermal silencing of Ano1/TMEM16A⁺ hippocampal neurons. (A) Comparison of representative GCaMP6f traces between Ano1/TMEM16A^+/- neurons at 36, 37, 38, and 39°C (left and right, respectively). (B,C) Action potential firing rate (per 5 s) comparison in Ano1/TMEM16A^+/- neurons at various bath temperatures. Significant suppression in firing rate was observed at 38°C (p = 0.0163) and 39°C (p = 0.0032) in Ano1/TMEM16A⁺ neurons, compared to 37°C (T-test). In Ano1/TMEM16A^- neurons similar comparison yielded non-significant changes (p = 0.7067 and p = 0.1682 at 38 and 39°C, respectively). (D) Firing rates of Ano1/TMEM16A⁺ and Ano1/TMEM16A^- neurons were not significantly different at 36°C (p = 0.6628) or 37 °C (p = 0.4526), while firing rates at 38 and 39°C varied significantly between Ano1⁺ and Ano1^- neurons (p = 0.0227 and 0.0051, respectively). ^∗p < 0.05, ^∗∗p < 0.005.

FIGURE 5

Nanoparticle heating on the neuronal membrane. (A) Fluorescence micrograph showing magnetic nanoparticles coated with Alexa fluor-647 labeled NeutrAvidin attached to the neuronal membrane in a rat hippocampal culture via biotinylated anti-A2B5 antibody. (B) Plot shows average temperature rise in magnetic nanoparticle decorated membrane under 5 s magnetic field (n = 3). Bottom axis shows time relative to AMF start, and top axis shows 5 s time bins when significant temperature points are achieved (e.g., 2–7 s = 1–3°C, 7–12 s = 3–1°C, 17–22 s <1°C above the starting temperature).

FIGURE 6

Magneto-thermal silencing of hippocampal neurons. (A) Representative traces of GCaMP6f fluorescence intensity averaged over pixels containing the soma of Ano1/TMEM16A⁺, MNP⁺ neurons, exposed to AMF (28.87 ± 1.03 kA/m, 412.5 kHz; gray, solid). Black sticks under each plot indicate respective calculated action potential (AP) events. (B) Binned (mean ± SEM, n = 9) AP firing rates (binned over 2 s, Hz) of Ano1/TMEM16A⁺, MNP⁺ neurons. Starting temperature was 37°C and AMF (28.87 ± 1.03 kA/m, 412.5 kHz; gray, solid) was applied for 5 s. (C,D) Binned (mean ± SEM, n = 9) AP firing rates (2 s bins, Hz) of Ano1/TMEM16A^-, MNP⁺ (n = 15, C) and Ano1/TMEM16A⁺, MNP^- (n = 5, D) neurons are shown. Starting temperature was 37°C and AMF (28.87 ± 1.03 kA/m, 412.5 kHz; gray, solid) was applied for 5 s in both cases. (E,F) Shows the comparison between AP firing rates recorded in 5 s time intervals corresponding to various temperature ranges achieved during magnetothermal membrane heating (and subsequent cooling) in Ano1/TMEM16A⁺, MNP⁺ neurons (see Figure 5B). Connected plots in E compares AP firing rates at different time bins after AMF application (starting at 0 s) with AP rates before AMF, in Ano1/TMEM16A⁺, MNP⁺ neurons. Overlaid blue markers show mean ± SEM (n = 20). AP rates before AMF application (37°C) was 3.05 ± 0.68 AP per 5 s; AP rates during 2–7 s (38–40°C) was 0.65 ± 0.25 AP per 5 s; AP rates during 7–12 s (38–39°C) was 0.90 ± 0.35 AP per 5 s; and AP rates during 17–22 s (37–38°C) was 2.30 ± 0.51 AP per 5 s. Significance level of each comparison is indicated alongside. The firing rate suppression was significant in 2–7 s and 7–12 s intervals (p = 0.0032 and 0.0094, respectively), while insignificant rate change was observed in the 17–22 s bin (p = 0.3858), unpaired T-test. Box and whisker plot in (F) summarizes the results in (E). Boxes span from 2nd to 3rd quartile (box top, 75% and box bottom, 25%), while whiskers indicate 10th and 90th percentiles. The black lines dividing the boxes indicate the median, while mean ± SEM values are overlaid in blue. (G,H) Shows the comparison between AP firing rates recorded in 5 s time intervals corresponding to various temperature ranges achieved during Magnetothermal membrane heating (and subsequent cooling) in Ano1/TMEM16A^-, MNP⁺ neurons. Connected plots in (G) compares AP firing rates at different time bins after AMF application (starting at 0 s) with AP rates before AMF, in Ano1/TMEM16A^-, MNP⁺ neurons. Overlaid blue markers show mean ± SEM (n = 15). AP rates before AMF application (37°C) was 2.20 ± 0.50 AP per 5 s; AP rates during 2–7 s (38–40°C) was 2.13 ± 0.48 AP per 5 s; AP rates during 7–12 s (38–39°C) was 2.53 ± 0.49 AP per 5 s; and AP rates during 17–22 s (37–38°C) was 3.53 ± 0.61 AP per 5 s. Significance level of each comparison is indicated alongside. The firing rate change was significant in all cases: 2–7 s, p = 0.9245; 7–12 s, p = 0.6364; and 17–22 s, p = 0.1044, unpaired T-test. Box and whisker plot in (H) summarizes the results in (G). Boxes span from 2nd to 3rd quartile (box top, 75% and box bottom, 25%), while whiskers indicate 10th and 90th percentiles. The black lines dividing the boxes indicate the median, while mean ± SEM values are overlaid in blue. ^∗p < 0.05, ^∗∗p < 0.005.

FIGURE 7

Firing suppression ratio and latency. (A) GCaMP6f signal integration (trapezoidal) in Ano1/TMEM16A⁺, MNP⁺ neurons (mean ± SEM, n = 12). AMF is indicated by the gray bar. (B) Box plot showing ratios of slopes obtained from integration plots in Ano1/TMEM16A⁺, MNP⁺ neurons. Suppression ratio is given as change in slope during suppression (silenced) and after the resumption of firing (resumed), compared to the initial slope. Ratio obtained during the silenced period was –82.04 ± 8.84% and following resumption was –8.98 ± 28.98% (mean ± SEM). The p-value between the silenced and resumed suppression ratio was 0.028, (n = 12, t-test). (C) Box plot showing the latency of silencing and the period of silencing, following 5 s of AMF in Ano1/TMEM16A⁺, MNP⁺ neurons. Latency was 1.882 ± 0.477 s following the start of AMF application, while the period was 12.05 ± 2.477 s (mean ± SEM). (D) GCaMP signal integration (trapezoidal) in Ano1/TMEM16A^-, MNP⁺ neurons (mean ± SEM, n = 16). AMF is indicated by the gray bar. (E) Box plot showing ratios of slopes obtained from integration plots in Ano1/TMEM16A^-, MNP⁺ neurons. The ratios obtained in similar time periods as indicated in (C) yield 6.02 ± 7.74% and 12.88 ± 44.20% during the silenced and resumed periods, respectively. No significant difference between these rates was found (p = 0.615, n = 16, t-test). All AMFs were 28.87 ± 1.03 kA/m r.m.s at 412.5 kHz. Boxes span the second and the third quartile, and the whiskers point at 10th and 90th percentile. Black bars in boxes indicate medians. ^∗p < 0.05.