[ad_1]
Mice
C57BL/6 J mice were purchased from Charles River. DAT-IRES-cre (B6.SJL-Slc6a3tm1.1(cre)Bkmn/J), GAD-IRES-cre (Gad2tm2(cre)Z) and µOR fl/fl (B6;129-Oprm1tm1.1Cgrf/KffJ or Oprm1fl/fl) mice were from the Jackson Laboratory, µOR cre/cre (B6N-Oprm1tmT2A-eGFP/cre(ICS)/Kf or Oprm1–cre) mice were provided by B. L. Kieffer and SST-IRES-cre (Ssttm2.1(cre)Zjh/J) mice were provided by A. Holtmaat. On arrival, the mice were given a period of 7 days for habituation. Both male and female mice, aged from 8–12 weeks, were used and group housed in a temperature-controlled (21 ± 2 °C) and humidity-controlled environment (50 ± 5%), under a 12 h light/dark cycle, and provided with food and water ad libitum. After surgical procedures, mice were single housed and recovered for at least 7 days before any experimental procedure. Weights and sexes were distributed homogeneously among the groups if possible. All behavioural procedures were performed during the light cycle. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Geneva and by the animal welfare committee of the Canton of Geneva, in accordance with Swiss law.
Virus injection and implantation
Mice (age 8–12 weeks) were anaesthetized with a mixture of isoflurane (induction 3%, maintenance 1.5%, Attane) and O2 (compact anaesthesia station from Minerve) during surgery and then secured in a stereotaxic frame (Stoeling). Before craniotomy, body temperature was maintained at 37 °C with a temperature controller system, and Lacryvisc (Alcon, Switzerland) was applied to prevent eyes from dehydration. For VTA recording of the different neuronal subtypes (anterior posterior (AP): −3.28; medio–lateral (ML): −0.9; dorso-ventral (DV): −4.3; with a 10° angle) or recording of CeA µOR-expressing neurons (AP: −0.9; ML: −2.8; DV: −3.9) mice were injected with an AAV-DJ-EF1α-FLEX-GCaMP6m (respectively 400 and 150 nl) produced at Stanford University vector core. For the recording of dopamine release, an AAV5-CAG-dLight1.2 (400 nl, from Addgene) was unilaterally injected in NAc (AP: +1.5; ML: −0.7; DV: −4.3). For knockdown experiments, µOR fl/fl mice were injected with an AAV8-hSyn-cre-tdTomato or the control virus AAV5-hSyn-mCherry (150 to 400 nl) in VTA (AP: −3.28; ML: −0.9; DV: −4.3, with an angle of 10°), NAc (AP: +1,5; ML: −0,7; DV: −4.3), BLA (AP: −1.2; ML: −3.2; DV: −4.2), PVT (AP: −0.9; ML: −0.4; DV: −3, with an angle of 10°) and CeA (AP: −0.9; ML: −2.8; DV: −3.9). For ISH (RNAscope) in wild-type or µOR fl/fl mice, 250 nl of an AAV8-hSyn-cre was injected in VTA (AP: −3.28; ML: −0.9; DV: −4.3, with an angle of 10°) or CeA (AP: −0.9; ML: −2.8; DV: −3.9). For immunohistochemistry experiments, Oprm1–cre or Sst-cre mice were injected with an AAV5-hSyn-Dio-mCherry in CeA (AP: −0.9; ML: −2.8; DV: −3.9). Finally, for oGABAsi experiments, AAV5-EF1α-eArch3.0-Dio-EYFP was injected in VTA (AP: −3.28; ML: −0.9; DV: −4.3, with an angle of 10°) and for CeA optogenetic manipulation of negative reinforcement an AAV5-EF1α-ChR2(H134R)-Dio-EYFP or the control virus was injected in CeA (AP: −0.9; ML: −2.8; DV: −3.9).
During the same surgical procedure, for in vivo recording of Ca2+ and dopamine release, an optic fibre (0.4 mm diameter, MFC_400/430_0.48_4mm_ZF2.5(G)FLT, Doric Lenses) was implanted and same for optogenetic experiment (oGABAsi and negative reinforcement) (0.2 mm diameter, FOC-W-1.25-200-0.37-5.0, Inper). Three screws were fixed into the skull to secure the optical implant, then the optic fibre was lowered 200 µm above the injection site and secure using dental cement. After surgery, mice were allowed to recover for 7 days and were habituated to handling.
Behavioural apparatus
The behavioural experiment on precipitation of withdrawal (knockdown, cFOS) as well as fibre photometry recording of calcium (Ca2+) GCaMP6m took place in a custom build chamber situated in a sound-attenuated chamber (Med Associates). The experiment chambers consist of a white Plexiglas square chamber (20 × 20 × 25 cm) surmounted by a video camera (Cineplex from Plexon) recording at a rate of 40 frames per second. On top of the chamber, a white transparent piece of Plexiglas with a hole at the centre was inserted to prevent mice from escaping. For fibre photometry recording of dopamine release evoked by fentanyl 0.3 mg kg−1 and apomorphine 10 mg kg−1, the experiment took place in a transparent custom-built open field (30 × 30 × 20 cm) surmounted by a FLIR camera (Blackfly S) recording at 30 Hz. oGABAsi and negative reinforcement experiments took place in an operant chamber (ENV-307A-CT, Med Associates) situated in sound-attenuating cubicle (Med Associates) consisting of a metal/Plexiglas square chamber (15.9 × 14 × 12.7 cm) with a grid floor in which two retractable levers were present on both sides of one wall surmounted by two cues light. The apparatus was controlled and data captured using a PC running MED-PC IV (Med Associates). For CPP or RTPA experiment, a three-compartment chamber (Med Associates) was used. The apparatus consists of two chambers separated by a corridor with equal surface, but distinct walls drawings and floor texture. On top of the context, a FLIR camera recording at 30 Hz (for CPP) or a camera connected to Cineplex (Plexon for RTPA) was used. Finally, for the locomotor response to different intraperitoneal injections, the experiment took place in a transparent custom-built open field (30 × 30 × 20 cm) surmounted by a camera connected to the Cineplex system to track the centre of gravity.
Behavioural paradigm
Dependency and withdrawal precipitation
Mice were first habituated to the intraperitoneal injection of saline at least for 3 consecutive days. Then increasing dose of fentanyl 0.06, 0.12, 0.18, 0.24 and 0.3 mg kg−1 (both injections at 10 ml kg−1) were injected intraperitoneally in their home cage to create dependency. On the challenge day, mice were injected with a dose of fentanyl at 0.3 mg kg−1 and put back in their home cage for 10 min. Then the behaviour was assessed in the video-tracking apparatus for 20 min (pre-period, reward). 30 min after the intraperitoneal injection of fentanyl, naloxone was injected intraperitoneally at a dose of 5 mg kg−1 (injection at dose of 10 ml kg−1) and the mice put back in the apparatus directly to evaluate precipitation withdrawal symptoms for 20 min (post period, withdrawal). Precipitation of withdrawal was manually scored by quantifying rearings, jumps, body licking, wet-dog shakes and defecations. Furthermore, immobility (2 s of immobility) and distance travelled (in metres) were extracted from the video track.
Optogenetic experiment
For optogenetic experiments, the implanted optic fibres were connected via patch cords (oGABASI, MFO-F-W1.25-200-0.37-100, negative reinforcement, BFO-1×2-F-W1.25-200-0.37-30, Inper) to a rotary joint (FRJ_1 × 2_FC-2FC; Doric Lenses), suspended above the operant chamber. A second patch cord was connected from the rotary joint to a blue or orange DPSS laser (SDL-473–100 mW, SDL-593–100 mW, respectively; Shanghai Dream Lasers) positioned outside of the context. Laser power was typically 15–20 mW measured at the end of each patch cord. A mechanical shutter was used to control laser output (SR474 driver with SR476 shutter head; Stanford Research Systems, aligned using a connectorized mechanical shutter adapter; Doric Lenses).
oGABAsi experiment (n = 9 mice) was designed on a fixed ratio 1 schedule (FR1) consisting of 1 h session daily during the conditioning phase and then two sessions for the occlusion experiment (20 min for pre-session and 1 h for post-session). Each ALP was associated with a cue light of 2 s, and, 5 s later, a continuous laser inhibition of GABA neurons lasting 7.5 s the first 7 days and 5 s the consecutive sessions, to reduce the time of optogenetic inhibition. From the ALP to the end of the optogenetic stimulation, every press on the ALP was recorded but did not initiate a protocol of stimulation (time-out period). The occlusion experiment was realized over 15 days and started by a 20-min pre-session. Then mice were injected intraperitoneally with saline (during baseline and recovery days) or fentanyl at different doses (0.06, 0.12, 0.15, 0.21, 0.3 and 0.18 mg kg−1) before the start of the session that lasted 60 min.
The negative reinforcement experiment (n = 8 mice for the ChR2 group and n = 7 mice for the EYP group, all female) was designed on a 1 h FR1 schedule for 6 days followed by 1 h FR3 schedule for 12 days. The mouse could stop continuous optogenetic stimulation at 20 Hz (5 ms pulse every 50 ms for 1 s every 2 s) by pressing on an ALP. Each ALP was associated with a cue light that lasted 2 s and a pause of the optogenetic stimulation for 20 s. From the ALP to the end of pause of optogenetic stimulation every press on the ALP was recorded but did not initiate a protocol of stimulation pause (time-out period). The occlusion experiment was realized over 12 days consisting of 3 days of baseline followed by 9 days where an injection of fentanyl at different doses was realized every other day (0.12, 0.06, 0.09, 0.3, 0.015 mg kg−1). During the baseline or the recovery day, mice were injected intraperitoneally with saline.
CPP and RTPA
For the CPP experiment (n = 10 for the VTA knockdown group and n = 11 for CTL group), mice were habituated to saline intraperitoneal injection at least 3 days before the beginning of the behaviour. On day 1 (pre-test), mice were placed in the corridor and allowed to explore both sides of the context for 20 min. Then 6 days of 20 min conditioning were realized by intraperitoneal injection of saline or fentanyl at 0.3 mg kg−1 in a randomly assigned side of the context. On the last day, the place preference was assessed by allowing the mouse to freely explore both sides of the context (post-test). Mice were video-tracked, and the time spent in each compartment was calculated offline using a markerless pose estimation method (DLC) and a custom-made Matlab script. The centre of gravity was used to assess the time spent in each of the three compartments (corridor, saline, or fentanyl context). CPP was calculated by computing the time spent in the fentanyl compartment divided by the time spent in both compartments per session.
To achieve real-time place aversion (RTPA), a camera linked to a Cineplex system (Plexon) was used to continuously video-track the mouse within the given context. When the centre of gravity was detected on one side of the context, an uninterrupted digital signal was transmitted to an Arduino device. This digital signal was then conveyed to an Arduino device linked to a blue laser to produce the stimulation pattern utilized in the negative reinforcement task (20 Hz; 5 ms pulse every 50 ms for 1 s every 2 s). After 4–5 weeks of viral expression, mice (n = 9 for ChR2 group and n = 15 for EYFP group) were habituated for 3 days of experimenter manipulations and to the connection of the cable. On day 1 (pre-test), mice were free to explore for 20 min both sides of the context and we assessed their place preference. On days 2, 3 and 4 mice were free to explore both sides of the context for 30 min. During this phase, when the centre of gravity of the mouse entered the preferred side, a stimulation was sent until the mouse left this side of the context. On day 5 (post-test), mice were free to explore for 20 min both sides of the context where we assessed again their place preference. RTPA was calculated by computing the time spent in the stimulated compartment divided by the time spent in both compartments per session.
Locomotor response to drug injection
For the fentanyl dose–response on locomotion, mice were first habituated to saline injection for 3 days. Then we randomly daily injected fentanyl (0.06, 0.12, 0.15, 0.18, 0.21 and 0.3 mg kg−1) over 6 days and assessed the locomotor response during 1 h. For the locomotor response in VTA µOR-knockdown mice versus controls, we injected saline intraperitoneal for 3 consecutive days, followed by fentanyl (at 0.2 mg kg−1) the next 2 days. A control group was used were we injected saline intraperitoneally over 5 days.
Fibre photometry recordings
After 4–5 weeks of viral expression, mice were first habituated to handling, to the connection cable and intraperitoneal injection of saline for 3 days before testing. On the testing day, mice were connected to the fibre photometry cable and placed in the apparatus for 3 min of habituation before the start of recording. For the study of dopamine release evoked in the VTA knockdown mice vs control, 5 min of baseline fluorescence were made before the intraperitoneal injection, and then the change of fluorescence was monitored during 40 min. Mice were injected intraperitoneally for 3 consecutive days respectively with saline, fentanyl (0.3 mg kg−1), apomorphine (10 mg kg−1). For the recording of the neuronal activity (CeA and VTA) during opioid dependency and withdrawal, mice were recorded during 5 min of baseline and then 20 min after naloxone alone (5 mg kg−1), fentanyl (0.3 mg kg−1). To reduce the entangling of the cable on the challenge day, 20 min after the fentanyl intraperitoneal injection, the photoreceiver was stopped and the cable disentangled and switched on 5 min before the intraperitoneal saline or naloxone (5 mg kg−1) injection. Finally, the neuronal activity was recorded for 20 min.
Fibre photometry was performed as before, and data were collected with TDT Synapse v.84 (Tucker Davis). During recordings, excitation (470 nm, M470F3, Thorlabs) and control LED light (405 nm, M405FP1, Thorlabs) were passed through excitation filters and focused onto a patch cord. The fibre patch cord was connected to the chronically implanted fibre, and emission light (500–550 nm) was collected through the same fibre and passed onto a photoreceiver (Newport 2151, Doric Lenses). After pre-amplification by the photoreceiver (2 × 1010 V/A) the signal was digitized, demodulated and stored using a signal processor (RZ5P, Tucker Davis Technologies).
The data were analysed using MATLABR2020 (MathWorks). First, the signal during baseline acquisition originating from the 405 nm excitation source was linearly regressed to the signal originating from the 470 nm excitation source, and scaled to the 470 nm originating signal. ΔF/F was then computed as (470 nm signal – fitted 405 nm signal)/fitted 405 nm signal. Finally, the ΔF/F was binned into 10-s time bins to plot an average graph, additionally to no binning for the study of transient activity evoked by the intraperitoneal injection. Transients were detected using the Matlab function findpeaks, where peaks were defined as a prominence greater than 2 standard deviations of the ΔF/F during baseline recording. For the calculation of the area under the curve (AUC), we used the Matlab function trapz Finally, for the normalization of the AUC to apomorphine we computed the ratio of AUC evoked by apomorphine injection to the one evoked by fentanyl injection.
Histological analysis
Ninety minutes after the precipitation of withdrawal, mice were injected with a lethal dose of pentobarbital (150 mg kg−1) and perfused transcardially with PBS and 4% paraformaldehyde solution. Brains were post-fixed overnight at 4 °C. Coronal sections (60 μm) of the region of interest were cut with a vibratome. Immunostaining started by blocking slices in PBS 10% BSA and 0.3% Triton X-100 followed by 48 h incubation in PBS 3% BSA and 0.3%Triton X-100 with primary antibody: rabbit polyclonal anti-cFOS (1:5,000, from SySy, 226003). After three 15 min washes in PBS at room temperature, slices were incubated with 1:500 Alexa-conjugated secondary antibodies against rabbit (Alexa-Fluor 488, Life Technologies, A1108). Then slices were washed three times in PBS. Slices were mounted and covered on microscope slides using DAPI mounting medium vectashield. Images were obtained in a confocal laser-scanning microscopy Leica SP8 confocal microscope using additional 350-nm laser with a 40×/0.7 NA oil immersion. Analysis was performed in at least three sections per mouse per structure of interest. Semi-manual quantification of cFOS was made by an experimenter who was blind to the experimental conditions. For the visualization of dLight expression, after slicing at 60 μm, slices were incubated with a primary antibody (1:500, rabbit polyclonal anti-GFP, Invitrogen, A11122) overnight at 4 °C and the secondary antibody (1:500, Alexa goat anti-rabbit, Life Technologies, A1108) for 2 h at room temperature.
In situ hybridization
Staining for Oprm1, Slc6a3, Slc32a1, Sst and Prkcd mRNAs was performed by smFISH. Brains from 7 C57BL/6 J 12-week-old mice were rapidly extracted and snap-frozen on dry ice and stored at −80 °C until use. VTA and CeA coronal sections (14 μm) were collected directly onto Superfrost Plus slides (Fisher Scientific). RNAscope Fluorescent Multiplex labelling kit (ACDBio 323110) was used to perform the smFISH assay according to manufacturer’s recommendations. Probes used for staining are Oprm1 (ACDBio 315841), Slc6a3 (ACDBio 315441-C3), Slc32a1 (ACDBio 319191-C2), Sst (ACDBio 404631-C2) and Prkcd (ACDBio 441791-C3). After incubation with fluorescently labelled probes, slides were counterstained with DAPI and mounted with ProLong Diamond Antifade mounting medium (Thermo Fisher Scientific P36961). Fluorescence images of labelled cells were captured using sequential laser-scanning confocal microscopy (Leica SP8) and co-localization was quantified manually. For the validation of the VTA or CeA µOR knockdown we automatically counted the number of puncta per slide compared to control condition using ImageJ software.
Video data analysis
The videos, which have a resolution of 640×480 and a frame rate of 40 fps, were analysed with DeepLabCut18. From a subset of videos, we extracted 25 frames per video using the kmeans algorithm to ensure diversity and labelled them manually. The labelling comprised 15 points of interest (four corners of the box, nose, both ears, both shoulders, body centre, both hips, and base, middle and end of the tail). The labelled images were divided into a training set (90%) and a test set (10%) and a model was trained using ResNet-50 and 800,000 iterations. The resulting error was 2.24 pixels on the training set, and 6.01 pixels on the testing set. The model was used to extract the xy coordinates of the previously mentioned points of interest throughout the videos. These coordinates were corrected in the following way: the points with low confidence (<0.05) and the outliers in speed or position were replaced by a value obtained by a cubic interpolation. The whole set of coordinates was also smoothed with a moving average filter of width 5.
The body parts coordinates were used to define 14 relative variables, namely the body extension (distance between the middle of the shoulders and the middle of the hips), the distance between the shoulders, the distance between the hips, the distance between the middle of the tail and the body centre, the tail extension (distance between the base and the end of the tail), the head extension (distance between the nose and the middle of the ears), the angle between the body and the tail, the angle in the middle of the tail, the angle between the head and the body, the rotation of the body with respect to a vertical line, the distance between the centre of the mouse body and the centre of the box, the body torsion (ratio of distance between shoulder and hip on the left vs on the right), the speed and rearing.
We defined a 15th variable describing the likelihood of a jump occurring on each frame. For this purpose, we used the fact that the tracking confidence (values between 0 and 1) would drop during jumps because the mouse would leave the frame for a few milliseconds. Knowing that the tracking confidence was close to perfect while the mouse was in frame, the probability of a jump happening can be roughly approximated by P(jump) ≈ 1 − (tracking confidence). Pairing this observation with a condition on a big enough speed preceding the loss of tracking allows a refinement of the detection of jumps, as we avoid classifying bad tracking as a jump. More precisely, a sequence of consecutive frames was considered as a jump if the confidence of tracking went below the fixed threshold of 50% and the speed around the loss of tracking went above the fixed threshold of 10 cm s−1. These thresholds were defined for the automatic jump detection to closely match the jumps observed during careful examination of a few videos.
For each mouse, we thereby obtain 15 time series (one per variable). The goal is to compare them and see if there are differences between the control group and each one of the experimental groups. Since a direct comparison between time series is not possible, we use hctsa19 to perform feature extraction: it evaluates more than 7,000 operations on each time series. A given time series is hence characterized by a vector with more than 7,000 entries containing the evaluated operations. For each of the 15 variables, we assess the similarity between a certain experimental group of mice (knockdown of µORs in different brain regions) and the control group (non-knockdown) by training a linear SVM classifier with 5 repeats of 5 folds cross validation on the characterizing vectors. We compute the mean balanced accuracy: mean balanced accuracy = (sensitivity + specificity)/2. The significance of the results is obtained by comparing our original accuracy with 1,000 repeats of a classification on shuffled data.
Statistical analysis and reproducibility
Data were analysed with Microsoft excel 16.16.05 and GraphPad prism 10.0.2. Sample size were estimated with G*power (HHU). For each experiment, a minimum of two replications were conducted by experimenters. Statistical analysis was performed in GraphPad Prism 9. For all tests, the significance threshold was placed at α = 0.05. Gaussian distribution was evaluated using the D’Agostino and Pearson normality test. Multiple comparisons were first subject to mixed-factor ANOVA or Kruskal–Wallis test (defining both between- and/or within-group factors), respectively, for normally distributed and non-normally distributed data. Where significant main effects or interactions between factors were found (P < 0.05), further comparisons were made for normally distributed data by a two-tailed Student’s t-test with Bonferonni corrections applied when appropriate or a Dunn test for non-normally distributed data (that is, the level of significance equalled 0.05 divided by the number of comparisons). Mann–Whitney or Wilcoxon tests were used for non-Gaussian distributions when appropriate. For normally distributed data, single comparisons of between- or within-group measures were made by two-tailed unpaired or paired Student’s t-test, respectively.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
[ad_2]
Source link
Leave a Reply