2.1.

Animals and Experiments

A transgenic DAT-Cre mouse line was crossed with a floxed ChR2(H134R)-EYFP mouse line to produce offspring (both from Jackson Laboratories). These offsprings were genotyped (Transnetyx) and those that tested positive for the presence of Cre were used in experiments ($N=8$ animals). Brain slices were cut from adult mice 2 to 3 months in age, both male and female. For the slicing procedure, the mice were anesthetized with a solution of ketamine ($10\mathrm{mg}/\mathrm{ml}$) and xylazine ($1\mathrm{mg}/\mathrm{ml}$) and decapitated. The brain was quickly removed and submerged in an ice-cold oxygenated artificial cerebro-spinal fluid (ACSF) cutting solution containing (in mM): 125 NaCl, 2.7 KCl, 25 ${\mathrm{NaHCO}}_3$, 1.22 ${\mathrm{NaH}}_2{\mathrm{PO}}_4$, 10 dextrose, 2 CaCl, 2 ${\mathrm{MgSO}}_4·7{\mathrm{H}}_2\mathrm{O}$, 1 ascorbic acid. Coronal slices, $300\mu \mathrm{m}$ in thickness, were cut with a vibrating blade microtome (Leica VT1000S, Leica Biosystems) in the same ice-cold ACSF cutting solution. The slices were incubated for 30 min at 32°C and then maintained at room temperature (20°C) in oxygenated ACSF cutting solution.¹⁹ Animal care and experiments were performed in accordance with the National Research Council Guide for the Care and Use of Laboratory Animals (Brown University Institutional Animal Care and Use Committee).

2.2.

Histology and Immunochemistry

For histological analysis, the transgenic mouse was terminally anesthetized with Beuthanasia-D. Perfusion was performed transcardially with saline followed by 4% paraformaldehyde in phosphate-buffered saline (PBS) solution. The brain was extracted and sequentially fixed in 4% paraformaldehyde in PBS solution (1 day) and then 30% sucrose solution (until brain sunk in solution). The brain was then frozen with dry ice and cut into $40\text{-}\mu \mathrm{m}$ coronal slices with a microtome. For tyrosine hydroxylase (TH) immunostaining, slices were sequentially washed in phosphate buffer (PB; twice, 5 min each), PBS (three times, 5 min each), and then left in block solution for 1 h. The block solution was prepared with 0.1% Tween (Sigma–Aldrich, Missouri), 0.25% Triton-X (Sigma–Aldrich, Missouri), and 10% normal donkey serum (EMD Millipore, MA) in PBS. Slices were then immersed in primary antibody (1:1000 AB152 Anti-Tyrosine Hydroxylase in block solution, EMD Millipore, Massachusetts), covered with tin foil, and rotated for 2.5 days in a cold room (4°C). Slices were thoroughly washed five times in PBS (5 min each) and again blocked in block solution for 1 h. The slices were then transferred to secondary antibody solution (1:500, Alexa Fluor® 594 donkey anti-rabbit IgG in block solution; Life technology, CA) for 2 h. Afterward, they were again washed in PBS (three times, 5 min each) and PB (twice, 5 min each), and mounted on microscope slides. Specific slices containing VTA ($-3.2\mathrm{mm}$ anterior–posterior) and NAc ($+1.0\mathrm{mm}$ AP) were selected and examined under a Zeiss LSM 510 Meta confocal laser scanning microscope to verify the colocalization of ChR2-EYFP with TH.

2.3.

Optogenetic Stimulation of Brain Slices

Slices containing the NAc ($+1.0\mathrm{mm}$ AP) were placed in a recording chamber under a microscope (Eclipse E600FN, Nikon, Melville, New York) and superfused with room temperature oxygenated ACSF recording solution containing (in mM): 125 NaCl, 2.7 KCl, 25 ${\mathrm{NaHCO}}_3$, 1.22 ${\mathrm{NaH}}_2{\mathrm{PO}}_4$, 10 dextrose, 1 CaCl, 1 ${\mathrm{MgSO}}_4·7{\mathrm{H}}_2\mathrm{O}$. The expression of ChR2-EYFP was observed under the microscope with an excitation wavelength of 490 nm by a Polychrome 5000 lamp from Till Photonics (Bavaria, Germany). After confirming opsin expression in the slices, excitation of the dopaminergic cells was performed with a blue laser (473 nm, CrystaLaser, Reno, Nevada), the output of which was coupled to the microscope and projected through the objective onto the slice. Home-made carbon fiber electrodes (CFE) were inserted under the microscope into the NAc area close to the anterior commissure of the coronal forebrain slices ($+1.0\mathrm{mm}$ AP) with the aid of a micromanipulator. Under constant light from the laser, we centered the light around the CFE to ensure that the dopaminergic projections in the NAc are receiving the light excitation when laser pulses are sent through the objective. The light power was measured by a power meter (PM100D, Thorlabs, Newton, New Jersey).

2.4.

Fast-Scan Cyclic Voltammetry

Carbon fibers (Goodfellow, Coraopolis, Pennsylvania) were vacuum aspirated into glass pipettes (Sutter Instrument, Novato, California) and subjected to heat assisted pulling (Micropipette puller P97, Sutter Instrument). Exposed carbon fibers were trimmed to around $50\mu \mathrm{m}$. These homemade carbon-fiber microelectrodes were connected to a Dagan CHEM-CLAMP voltammetry amplifier through a 1 MOhm head stage ($N=0.01$). We used two National Instrument PCI cards: NI PCI 6024e was used for voltammetry data acquisition and command voltage waveform output; NI PCI 6321 was used for synchronization and triggering of optical stimuli. The data acquisition software used was Demon Voltammetry.²⁰ The head stage, electrode, and recording media were placed inside a Faraday cage to prevent extraneous electric noise. The command voltage of FSCV was scanned from $-0.4$ to 1.2 V and back to $-0.4\mathrm{V}$ at $400\mathrm{V}/\mathrm{s}$ with a sampling frequency of 10 Hz. Cyclic voltammograms in the 1 s before stimulation onset were averaged and used as the baseline for background subtraction. We examined background subtracted cyclic voltammograms [Fig. 2(c)] in each recording to identify the typical oxidation peak at 0.6 V and reduction peak at $-0.3\mathrm{V}$, respectively, to confirm that the recorded chemical species is DA, with minimal interference from other chemicals.²¹ Electrodes were calibrated by $1\text{-}\mu \mathrm{M}$ standard DA solution.^22,23 The concentration traces of recorded DA were generated using the calibration factors. In the optogenetic experiments, we paused for at least 5 min between each FSCV recording session to allow DA recovery and ensure consistency among multiple DA release events. The FSCV setup was turned on 10 min before applying stimulation to ensure a stable baseline in each session. As a control, we performed the FSCV recording experiment with the electrode placed in the NAc in wild type animals without ChR2, and did not observe any induced DA signals.

2.5.

Data Analysis

DA concentration traces recorded by FSCV were generated in Demon Voltammetry.²⁰ The FSCV color plots follow the standard two-dimensional (2-D) representation with time along the $X$ axis, command voltage along the $Y$ axis and pseudo color along the $Z$ axis showing the oxidation current measured from electrodes (background subtracted). These 2-D color plots were also generated in Demon Voltammetry. Data across multiple animals were grouped together for investigation of different light stimulation parameters; the amplitudes of DA release were normalized to allow comparison across animals with respect to one highest value we recorded in each animal. The exact peak DA levels depend on factors such as local density of dopaminergic neuron terminals; and therefore, vary with the electrodes’ different position in the NAc tissue, among different slices, and across animals. The average peak DA level observed was estimated by electrode calibration to be 524 nM ($\mathrm{SEM}=53\mathrm{nM}$; $N=8$). For curve fitting and modeling, data was exported and analyzed in MATLAB®. The model we adapted for kinetic analyses is the Michaelis–Menten model, which is the classical enzyme kinetics model commonly used to characterize DA uptake.²⁴ To fit the model, an objective function calculating the sum of squared error between recorded data and simulated data was created in MATLAB®. Simulated data was obtained by MATLAB® differential equation solver, given the Michaelis–Menten equation in Eq. (1).

3.1.

Direct Light Stimulation Targeting Dopaminergic Cell Terminals Elicits DA Release in Transgenic Mouse Brain Slices

We used optogenetics to evoke transient, relatively high concentrations²⁵ of extracellular DA release in the NAc area of eight adult transgenic mice. Only mice identified as positive (+) for Cre [i.e., expressing ChR2(H134R)-EYFP] by genotyping were used. Brain slices were checked under a fluorescent microscope to confirm opsin expression before each recording experiment. Opsin expression was found in the VTA and SN in the midbrain, and in the NAc and CPu in the forebrain area [Fig. 1(a)], which is in agreement with previous studies on dopaminergic cell distribution in the mouse brain.²⁶ Expression in other areas was not observed in the forebrain and midbrain slices. Histology experiments were done with immunostaining of TH to verify the colocalization of ChR2-EYFP with TH in both the VTA and NAc. Dopaminergic cell bodies stained with TH and showing yellow fluorescent protein (YFP) fluorescence were found in the VTA [Fig. 1(b)]. In the NAc, which contains the axons and terminals of DA neuron, YFP fluorescent neuronal projections overlapped well with the TH-labeled structures [Fig. 1(c)]. Such traces were not seen in the anterior commissure [bottom right corner, Fig. 1(c)], which does not contain dopaminergic projections. These histological results confirmed that our optogenetic stimulation of ChR2 was indeed affecting dopaminergic projections, as expected in these transgenic animals conditionally express ChR2 by DAT-Cre.

Fig. 1

ChR2-EYFP expression in transgenic mouse brain slices. (a) Fluorescent microscope images (excitation wavelength for YFP: 490 nm) show existence of YFP fluorescence in caudate putamen (left) and nucleus accumbens (NAc) (middle; AC: anterior commissure) in forebrain coronal slices ($+1.0\mathrm{mm}$ AP; scale bars are 0.2 mm). Expression is also seen in midbrain coronal slices of SN/VTA (right; $-3.1\mathrm{mm}$ AP; scale bar is 0.5 mm). (b) Histological analysis with tyrosine hydroxylase (TH) immunostaining in the VTA ($-3.2\mathrm{mm}$ AP) where DA cell bodies can be seen. TH is shown in red (left); YFP fluorescence is in green (middle); merged image is on the right. Scale bars indicate $20\mu \mathrm{m}$. (c) Same histological images in the NAc ($+1.0\mathrm{mm}$ AP) show dopaminergic neuron projections (but not cell bodies). Scale bars indicate $20\mu \mathrm{m}$.

To measure phasic terminal DA release events modulated by optogenetics, FSCV electrodes were inserted into brain slices at the NAc. Light was delivered through the objective forming a $d=2.4\mathrm{mm}$ illuminated spot around the electrode tip. As analog input controlled laser pulses were delivered onto the brain slices, simultaneous FSCV recording tracked the extracellular concentration of DA. As shown in [Figs. 2(a) and 2(b)], an abrupt increase of DA concentration followed by gradual decay over a few seconds was observed following a 25-ms light pulse. The background subtracted cyclic voltammograms displayed clear oxidation and reduction peaks at $+0.6$ and $-0.3\mathrm{V}$ [Fig. 2(c)], which are known to correspond to DA. The optogenetic induction of immediate DA release was reproducible in all $N=8$ adult transgenic animal subjects with different peak concentrations. During multiple stimulation events in one brain slice when the electrode was maintained at the same position, the time profiles of DA release showed little variation as seen in Fig. 2(b) (error bars are $\pm \mathrm{SEM}$, peak concentration at $705\pm 11\mathrm{nM}$). DA signals were not seen in control experiments with the CFE placed out of the brain slice [Fig. 2(d)]. The optical stimulation power required to induce DA release was found to be relatively low ($<1\mathrm{mW}/{\mathrm{mm}}^2$), which minimizes potential light artifacts at our typical stimulation parameters. The absence of any artifacts during induced DA release combined with the cell-type specificity in optogenetics enables us to investigate clean DA dynamics in the NAc, as well as to obtain accurate estimates of any kinetic parameters.

Fig. 2

Optogenetically induced dopamine (DA) release in NAc and the modulation effect of light power. (a) Recorded standard fast-scan cyclic voltammetry (FSCV) two-dimensional (2-D) color plot under a 25-ms square pulse light stimulation at $t=6.2\mathrm{s}$. $Z$ axis shows background-subtracted voltammetry current in pseudocolor. (b) Mean DA concentration trace under the same stimulation parameter with small vertical bars represent $\pm \mathrm{SEM}$ ($n=7$ recordings in the same slice). (c) Background subtracted cyclic voltammogram at optical DA release peak; oxidation current is seen at 0.6 V and reduction current around $-0.3\mathrm{V}$. (d) One control experiment where the electrode is not inserted in the brain slices but under the same light stimulation parameters shows no DA signals and minimal light artifact. (e) Optically induced DA peak level as a function of light power, across three animals. ($n=3\times 6=18$ recordings in total; recordings are separated by 5 to 10 min to allow DA recovery. Vertical bars represent ±SEM. Pulse width is fixed at 100 ms.) (f) Stacked DA concentration traces under the six different light power in one of the animals.

3.2.

Optogenetic Modulation of Scalable DA Release Events

It is known that the spontaneous release of DA during animals’ natural behaviors has more variability, and is often lower in concentration²⁷ than the DA release induced by an electrical pulse. This indicates that only a portion of the stored presynaptic DA pool is typically released into the extracellular space during natural behavior. To mimic the DA release events in naturally behaving animals, it is essential to achieve scalable DA release with certain precision. We investigated the effects of altering the optical stimulation power and light pulse width. Across three transgenic animals, a series of stimulations with different light powers were delivered with simultaneous FSCV, ranging from 0.09 to $1\mathrm{mW}/{\mathrm{mm}}^2$. To minimize the effects brought by light pulse width, we used square pulses fixed at 100-ms long. As demonstrated in Figs. 2(e) and 2(f), the amplitude of the induced DA concentration increased as the light power was increased. A good linear fit was not obtained between the DA levels and light power density in Fig. 2(e), indicating a nonlinear and complex correlation.

We also observed an increased DA response as the light pulse width was increased while the light power was held constant ($1\mathrm{mW}/{\mathrm{mm}}^2$) across four transgenic animals. In Fig. 3(a), short light pulses of 3-ms induced relatively small DA release. However, a sharp increase of DA level was seen as the light pulse width increased. This increased DA release reached a plateau around 15 ms. This indicates that when the light stimulation is sufficiently long ($>15\mathrm{ms}$), the amount of released DA reaches a saturation point for the effect of pulse width. Previous studies in electrically elicited DA release^24,28 have been using the number of stimulation pulses as the primary approach to modulate DA release concentrations. We found that increasing the pulse width in optogenetics can result in a similar effect as multipulse stimulations as shown in Fig. 3(b). Across three transgenic animals, four pulses of 4 ms (25 Hz) stimulation increased the amount of DA release to around 200% of that from a single 4-ms pulse; whereas a longer 25-ms pulse also elicited higher DA release. The release levels between four pulses of 4 ms (25 Hz, takes a total time of 124 ms) and a single 25 ms pulse is similar (one-way ANOVA and multiple comparison test, $\alpha =0.05$). Therefore, optical stimulations appear to be more versatile and effective since longer pulse width or stronger light power can lead to increased DA release, in addition to increasing the number of pulses. However, applying large quantity of stimulation pulses over an extended period has the potential to result in even higher DA release;^12,24 whereas increasing the pulse width beyond 15 ms does not seem to further increase DA release concentrations.

Fig. 3

The effects of light pulse width and waveform shape on the concentration of DA release. (a) Optically induced DA peak level as a function of light pulse width across four animals ($n=4\times 6=24$ recordings in total; recordings are separated by 5 to 10 min to allow DA recovery. Vertical bars represent $\pm \mathrm{SEM}$. Light power is fixed at $1\mathrm{mW}/{\mathrm{mm}}^2$). (b) The induced DA release levels from light stimulations of a 4-ms pulse, four pulses of 4 ms (25 Hz), and a 25-ms pulse; $n=7$, 10, 22 recordings respectively, across three animals. DA level evoked by the 4-ms pulse is different than the other two cases (one-way ANOVA and multiple comparison test, $\alpha =0.05$; concentration normalized with respect to DA release of 25 ms pulses. (c) Standard FSCV 2-D color plots under a 1-s square pulse (top) versus a 200 ms ramp $+1\text{-}\mathrm{s}$ square pulse stimulation (bottom). $Z$ axis shows background-subtracted voltammetry current in pseudo color; data is from single recordings. (d) DA peak level as a function of added ramp time to the 1-s square pulse. (In $n=4\times 6=24$ total recordings across four different animals. Recordings are separated by 5 to 10 min to allow DA recovery.) $Y$-axis is normalized with respect to the DA release induced by nonramp square pulses in each animal. Among adjacent ramp durations, DA peak level is significantly different between 50 and 75 ms ramp waveforms (one-way ANOVA and multiple comparison test, $\alpha =0.05$).

3.3.

Temporal Shape of Light Stimulus Waveform Affects the Amplitude of Transient DA Response

In agreement with previous findings,^10,11 our results indicate that higher extracellular DA concentrations can be induced by increasing either the intensity, the duration, or the number of light pulses, all of which correspond to increased light energy delivery into the targeted brain area. However, we found that controlling the energy delivery is not the only way to modulate DA release; the light pulse waveform also plays an important role. We demonstrate here that the optogenetically evoked transient DA signal is significantly dependent on the stimulus waveform at the onset of the light stimulus. We performed a series of FSCV recordings where the temporal shape of the rising edge of the light pulse was varied. The abrupt rising edge of a square pulse was replaced by a short linear ramp of varying slope at the onset of 1-s light stimulations. These temporal waveforms with initial ramps ranging from 12.5 to 300 ms were applied as analog inputs to the blue laser driver. All stimulations shared the same 1-s square pulse shape after the initial ramp. A comparison is shown in Fig. 3(c) between a stand-alone 1-s square pulse and a 1-s square pulse with a 200 ms initial ramp. Lower DA response is observed in the latter case despite the fact that more total light energy was delivered to the targeted brain area. More systematically, we investigated this effect in four transgenic mouse subjects by gradually increasing the ramp duration [Fig. 3(d)]. Results show that peak DA concentration varies dramatically with ramp duration. For example, the released DA concentrations are significantly different between waveforms with 50 and 75 ms added-ramp (one-way ANOVA and multiple comparison test, $\alpha =0.05$). In fact, any of the DA release from 12.5, 25, and 50 ms ramp waveforms is larger than that from 75, 100, and 300 ms ramp waveforms (one-way ANOVA and multiple comparison test, $\alpha =0.05$). Longer ramps preceding the square pulse invariably resulted in lower peak DA concentrations, even though more total light energy is delivered. Our findings thus demonstrate the induced transient DA efflux is critically dependent on the shape of temporal waveforms of optogenetic stimulation at the moment of light onset. On the other hand, difference in induced extracellular DA concentration is not necessarily a result from change in the amplitude and length of light stimulations.

3.4.

Recovery of Presynaptic DA Revealed by Optogenetics and Kinetic Analyses

As shown, upon the onset of optical stimulus, stored presynaptic DA is released, causing a transient local high concentration of extracellular DA in the submicromolar range. The extracellular DA then returns to the cytosol through DAT uptake to be prepared for a future release event. The DA must also be sequestered into synaptic vesicles to be recovered into a readily releasable state, in addition to simply being transported back into the cytosol.²⁹ The releasable presynaptic DA pool is recovered dynamically during this process. Optogenetics is an ideal tool to unfold the dynamics behind the recovery of releasable presynaptic DA because conventional electrical stimulation also excites nondopaminergic neurons in the brain which may interfere with extracellular DA release, whereas optogenetics allows the selective activation of dopaminergic cells. For this optogenetic study, we used stimulation protocols with series of two consecutive light pulses applied at varying intervals. The voltammetry signature of DA is validated by the FSCV colorplot [Fig. 4(a)]. As shown in Fig. 4(b), the initial light pulse in each recording always elicits about the same peak DA concentration. As the interpulse interval is increased from 2 to 40 s, the second pulse elicits DA concentrations of gradually increasing amplitude. At short interpulse intervals of 2, 3, and 4 s, the readily releasable DA is less than 20% of the original amount, indicating depletion of releasable presynaptic DA. For recording with 40-s interpulse interval [Fig. 4(c)], the second pulse elicits DA level to 70% of that from the initial pulse, suggesting most of the releasable DA pool is restored. Complete recovery can be achieved on a scale of minutes, since 5-min intervals between the trials were sufficiently long so that the initial DA peak for each trial remained nearly constant in magnitude ($543\pm 27\mathrm{nM}$). The recovery of DA in the two transgenic mouse subjects studied in this manner showed consistent results [Fig. 4(d)]. When compared with the extracellular clearance of DA recorded by FSCV [Fig. 4(d), red line], the recovery of readily releasable presynaptic DA takes place on a much longer time scale.

Fig. 4

Recovery of releasable presynaptic DA pool and extracellular DA kinetic analyses. (a) FSCV color plot in one recording in which two light pulses of optogenetic stimulation were set 6 s apart. (b) Series of recorded DA concentration traces under two light pulses with interpulse intervals of 2, 3, 4, 6, 10, 15, 20 s. Stimulations were kept at $1\text{-}\mathrm{mW}/{\mathrm{mm}}^2$ power and 1 s in length; data is from a set of seven recordings in one of the animals from d. Recordings are separated by 5 to 10 minutes to allow DA recovery. (c) Same DA concentration trace at 40 s interpulse interval. (d) Recovered releasable presynaptic DA on second pulses as a function of interpulse interval in two adult transgenic mice (blue and black). Red line indicates the clearance of extracellular dopamine molecules (i.e., DA concentration recorded by FSCV). (e) In the uptake kinetic analysis, simulated (red) versus recorded (black) FSCV DA concentration signal. FWHH, $T_{20}$, $T_{80}$ is shown.

We further performed kinetic analyses on the uptake and recovery process of optogenetically induced DA concentration data. Michaelis–Menten kinetics (Methods) is used to model the DA uptake. The estimated best fit parameters were found to be the following: $V_{\text{max}}$ is $0.969\mu \mathrm{M}/\mathrm{s}$ taking $K_m$ of $0.21\mu \mathrm{M}$³⁰ [$R^2=0.918$, Fig. 4(e)]; the first-order rate constant $k$ ($k=V_{\text{max}}/K_m$) is $4.6{\mathrm{s}}^{-1}$. This estimation is similar to the previously reported electrically evoked DA uptake rate constant $k$ of 4.0 to $7.8{\mathrm{s}}^{-1}$ in the mouse NAc area³¹ and also lower than the reported rate constant in the CPu area.¹⁵ In addition, several other parameters describing the uptake kinetics such as full-width at half height (FWHH), $T_{20}$ (the time of 20% substrate clearance), and $T_{80}$ (the time of 80% clearance)²⁰ were also extracted from the data. Optogenetically induced DA release peaks by single 1-s square pulses have a mean FWHH of $1.48\pm 0.02\mathrm{s}$, $T_{20}$ of $0.68\pm 0.03\mathrm{s}$, and $T_{80}$ of $2.52\pm 0.10\mathrm{s}$. We also applied Michaelis–Menten kinetics analysis to the releasable DA pool recovery data. In this case, a $K_m$ value of $1\mu \mathrm{M}$ is used^32,33 to fit the model, yielding $V_{\text{max}}$ values of 0.0585 and $0.0592\mu \mathrm{M}/\mathrm{s}$, respectively, for the two transgenic animals [Fig. 4(d), $R^2=0.922$, 0.881]. The recovery of releasable presynaptic DA seems to be adequately fitted with the Michaelis–Menten model. The $V_{\text{max}}$ value obtained for recovery of releasable DA was much lower than the value obtained for DA uptake, which indicates that uptake into cytosol alone is not sufficient to prepare DA molecules for another immediate release. Therefore, this validates that other processes are involved in the recovery of releasable presynaptic DA.