Drug-driven AMPA receptor redistribution mimicked by selective dopamine neuron stimulation

Abstract

Background: Addictive drugs have in common that they cause surges in dopamine (DA) concentration in the mesolimbic reward system and elicit synaptic plasticity in DA neurons of the ventral tegmental area (VTA). Cocaine for example drives insertion of GluA2-lacking AMPA receptors (AMPARs) at glutamatergic synapes in DA neurons. However it remains elusive which molecular target of cocaine drives such AMPAR redistribution and whether other addictive drugs (morphine and nicotine) cause similar changes through their effects on the mesolimbic DA system.

Methodology/principal findings: We used in vitro electrophysiological techniques in wild-type and transgenic mice to observe the modulation of excitatory inputs onto DA neurons by addictive drugs. To observe AMPAR redistribution, post-embedding immunohistochemistry for GluA2 AMPAR subunit was combined with electron microscopy. We also used a double-floxed AAV virus expressing channelrhodopsin together with a DAT Cre mouse line to selectively express ChR2 in VTA DA neurons. We find that in mice where the effect of cocaine on the dopamine transporter (DAT) is specifically blocked, AMPAR redistribution was absent following administration of the drug. Furthermore, addictive drugs known to increase dopamine levels cause a similar AMPAR redistribution. Finally, activating DA VTA neurons optogenetically is sufficient to drive insertion of GluA2-lacking AMPARs, mimicking the changes observed after a single injection of morphine, nicotine or cocaine.

Conclusions/significance: We propose the mesolimbic dopamine system as a point of convergence at which addictive drugs can alter neural circuits. We also show that direct activation of DA neurons is sufficient to drive AMPAR redistribution, which may be a mechanism associated with early steps of non-substance related addictions.

Figure 1. Cocaine drives the insertion of GluA2-lacking AMPARs via its effect on the DAT.

(A) Single unit extracellular in vivo recordings (above) and corresponding firing rate plots (below) of VTA neurons during a single i.p. injection of 15 mg/kg cocaine in either WT (left) or DAT_KI (right) mice. Black bar denotes injection time, (a) and (b) denote points from which example traces were taken. (B) The resulting inhibition of neuron firing rate observed in WT mice (38±3.3%) was not present in DAT_KI mice (94.9±1.4%). n = 4–5, t₍₇₎ = 16.5, p<0.0001. (C) Representative AMPAR excitatory postsynaptic currents recorded at −60, 0 and +30 mV (normalized to +40 mV AMPAR component) and RIs (D) of WT and DAT_KI mice 24 h post cocaine injection. Linearity corresponds to and RI of 1. Mean RI = 1.95±0.17 in WT, and 1.12±0.08 in DAT_KI; F_(2–22) = 9.8, p<0.001, n = 5–9). All data are expressed as mean ± sem.

Figure 2. Addictive drugs cause rectification via AMPAR redistribution.

(A) Representative traces of AMPAR excitatory postsynaptic currents recorded at −70, 0 and +40 mV. Examples are shown from recordings 24 h post injection. (B) Individual and averaged normalized rectification indices (RIs) (mean ± s.e.m) of saline and each drug treatment. RIs of morphine (2.12±0.27), nicotine (2.06±0.23) and cocaine (1.72±0.14) groups were significantly different from the saline (1.12±0.08) control group (F_(3,35) = 4.93, p<0.01, ANOVA. n = 7–15). (C) Representative electron micrographs of VTA sections from saline- or drug-treated animals. Large profiles (arrows) represent tyrosine hydroxylase (TH) immunoreactivity in dendrites (Den) forming asymmetrical synapses with boutons (b), and small profiles (arrowheads) represent GluA2 immunoreactivity. (D) Number of small profiles plotted against the distance from the postsynaptic density. (E) Same as in (C) but staining against PSD 95. (F) Same quantification as in (D) but for PSD 95.

Figure 3. Light pulses are sufficient to mimic burst firing of dopamine (DA) neurons in vivo.

(A) Representative single unit recording (above) and peristimulus time histogram (below, 5 ms bins) of a VTA DA neuron during a single light pulse (black markers; 4 ms, one sweep every 2 s) (n = 7). (B) The same VTA DA neuron as in (A) responding to 5 light pulses at 20 Hz. (C) Light pulses (black markers) are sufficient to drive action potentials, which do not differ in waveform characteristics from spontaneously occurring action potentials (above). Average percentage of action potentials generated by consecutive light pulses (below). Note the decrease in fidelity of action potential firing with increasing numbers of light pulses. (D) A GABAergic VTA neuron, which was recorded in close proximity to light-responsive DA neurons, exhibiting no response to five 4 ms light pulses (n = 7).

Figure 4. Expression of ChR2 causes light-activated currents in DA VTA neurons.

(A) Representative whole cell voltage clamp recordings of a DA VTA neuron. Following identification of cell type by the presence of an I _h current (left) responses to light pulses (black lines) at 4 ms (middle) or 100 ms (right) in the presence of TTX were tested (n = 10). (B) Same as in (A) but a representative non-DA VTA neuron (note lack of I _h (left); n = 6). (C) Digital micrograph showing YFP labeling of neurons within the VTA, together with putative GABAergic unlabeled neurons (asterisks).

Figure 5. In vivo stimulation of dopamine neurons is sufficient to drive AMPAR redistribution.

(A) Protocol of light stimulation in vivo. (B) Whole-cell voltage-clamp recordings made ex vivo 24 h post in vivo stimulation protocol. Representative traces of AMPAR excitatory postsynaptic currents recorded at −60, 0 and +30 mV (below). (C) Individual and averaged normalized rectification indeces (RIs) (mean ± s.e.m) following light stimulation. RI of DAT Cre− = 1.12±0.14 (n = 5). RI of DAT Cre+ = 2.45±0.32 (n = 5). p<0.01, Mann-Whitney U Test. (D) Protocol of intra-VTA infusion and light stimulation in vivo, (E) Same as in (B) following intra-VTA infusion and light stimulation, (F) Same as in (C). RI of DAT-Cre+ saline injected = 1.79±0.23 (n = 4). RI of DAT Cre+ SCH23390 injected = 0.99±0.09 (n = 5). p<0.05. Error bars represent s.e.m. Error bars are smaller than the symbol for some data points.