Abstract
Cerebellum has been implicated in drug addiction; however, its underlying cellular populations and neuronal circuitry remain largely unknown. In the current study, we identified a neural pathway from tyrosine hydroxylase (TH)–positive Purkinje cells (PCTH+) in cerebellar lobule VI to calcium/calmodulin-dependent protein kinase II (CaMKII)–positive glutamatergic neurons in the medial cerebellar nucleus (MedCaMKII), forming the lobule VI PCTH+–MedCaMKII pathway in male mice. In naive male mice, inhibition of PCTH+ neurons activated Med neurons. During conditioned place preference (CPP) training, exposure to methamphetamine (METH) inhibited lobule VI PCTH+ neurons while excited MedCaMKII neurons in mice. Silencing MedCaMKII using a tetanus toxin light chain (tettox) suppressed the acquisition of METH CPP in mice but resulted in motor coordination deficits in naive mice. In contrast, activating lobule VI PCTH+ terminals within Med inhibited the activity of Med neurons and subsequently blocked the acquisition of METH CPP in mice without affecting motor coordination, locomotor activity, and sucrose reinforcements in naive mice. Our findings identified a novel lobule VI PCTH+–MedCaMKII pathway within the cerebellum and explored its role in mediating the acquisition of METH-preferred behaviors.
- addiction
- cerebellum
- conditioned place preference
- medial cerebellar nucleus
- methamphetamine
- Purkinje cells
Significance Statement
Our findings identified a novel lobule VI PCTH+–MedCaMKII pathway in the cerebellum and revealed its role in the acquisition of methamphetamine (METH)-preferred behavior. Silencing MedCaMKII efficiently suppressed the acquisition of METH CPP but produced serious deficits in motor coordination. In addition, inhibiting the Med innervated by lobule VI PCTH+ neurons during the training of METH CPP blocked the acquisition of METH CPP without influencing motor coordination, locomotor activity, and sucrose reinforcements.
Introduction
Methamphetamine (METH) is a commonly abused addictive psychostimulant. The reinforcing effects of METH, which are experienced as rewarding, have long been implicated in driving METH consumption or even addiction. Accumulating evidence from recent decades highlights the pivotal role of the cerebellum in mediating reward processing (Medina, 2019; Kostadinov and Hausser, 2022). Cerebellar granule cells encoded reward expectations (Wagner et al., 2017), while optogenetic activation of cerebellum projections has been shown to induce reward responses in mice (Carta et al., 2019). The cerebellar climbing fibers sent predictive reward signals to the cerebellar Purkinje cells (PCs; Heffley and Hull, 2019; Kostadinov et al., 2019). METH use disorders significantly altered gene expressions, neuronal activity, and structural architecture within the cerebellum both in humans (Jiang et al., 2021) and animals (Ferrucci et al., 2007; Hamamura et al., 2010; Thanos et al., 2016; Eskandarian Boroujeni et al., 2020). However, it remains largely unknown whether the cerebellum encodes the reinforcing process of psychostimulants, such as METH.
The deep cerebellar nuclei (DCN) serve as a prominent source of cerebellar output (Kang et al., 2021), sending direct or indirect projections to reward-related brain regions, including the ventral tegmental area (VTA) (Beier et al., 2015; Carta et al., 2019), striatum (Hoshi et al., 2005), and PFC (Middleton and Strick, 2001), implying its essential role in processing rewards. In rodents, the DCN is divided into three subdivisions from medial to lateral: the medial cerebellar nucleus (Med, fastigial in humans), interposed cerebellar nucleus (Int), and lateral cerebellar nuclei (Lat, dentate in humans). METH exposure overactivated microglia in the DCN of rats (Thanos et al., 2016). Specifically, both cocaine administration (Vazquez-Sanroman et al., 2015) and random cocaine-odor pairing training (Carbo-Gas et al., 2014b, 2017) activated the Med neurons in mice. These findings suggest the potential involvement of the DCN in psychostimulant addiction. However, the precise neuronal circuits of the DCN underlying addiction remain unclear.
Anatomically, the innervation of the DCN from the cerebellar cortex is exclusively from cerebellar PCs (Armstrong and Schild, 1978; Teune et al., 1998; Hirono et al., 2021). It is generally believed that PCs send GABAergic innervations to the DCN neurons, resulting in an inhibitory neuronal pathway (Hirano, 2018). In rodents, METH exposure reduced the number of PCs (Eskandarian Boroujeni et al., 2020), while cocaine exposure decreased the spontaneous discharges (Jimenez-Rivera et al., 2000) and dendritic spine density of PCs (Vazquez-Sanroman et al., 2015), indicating a functional decline of cerebellar PCs in response to psychostimulants. Tyrosine hydroxylase (TH)–positive PCs (PCTH+) have been identified both in rodents (Hess and Wilson, 1991; Takada et al., 1993; Lee et al., 2006; Choi et al., 2012; Locke et al., 2020) and humans (Fujii et al., 1994). TH mapping revealed a wide distribution of PCTH+ with specific enrichment in posterior and lateral regions associated with cognitive function, from lobule VI to lobule X (Locke et al., 2020). METH exposure dose-dependently increased the number of cerebellar PCTH+ (Ferrucci et al., 2006, 2007). An imaging study revealed an aberrant functional connectivity between cerebellar lobule VI and precuneus in METH users, which was found to be correlated with addiction severity (Li et al., 2020). Further, cerebellar lobule VI has been reported to constitute an integrative interface between motor and cognitive/emotional circuits (Belkhiria et al., 2017), and cannabis users exhibited increased thickness of lobule VI compared with healthy controls (Wang et al., 2021), implying the involvement of lobule VI in addiction. The distribution of PCTH+ projections, particularly from lobule VI to the Med, and its role in the process of drug addiction remain unknown.
In this study, we developed METH conditioned place preference (CPP) model in male mice to explore the role of PCTH+–Med pathway in the acquisition of METH CPP. In parallel, the rota-rod test, open-field test (OFT), and sucrose self-administration (SA) were used to assess motor coordination, locomotor activity, and natural rewarding in naive mice.
Materials and Methods
Animals
Male C57BL/6 mice (∼25 g, 8–10 weeks of age) were maintained on a 12 h reverse light/dark cycle with food and water available ad libitum. All mice were handled for 3 consecutive days before each experiment. All procedures were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee (IACUC) at Nanjing University of Chinese Medicine.
CPP
METH CPP procedures were performed in the TopScan3D CPP apparatus (CleverSys), which is constructed of two distinct chambers (different visual and tactile cues) separated by a removable guillotine door. The CPP procedure consisted of three phases: the preconditioning test (pretest), conditioning (CPP training), and postconditioning test (CPP test; He et al., 2022). For the pretest, mice were allowed a 15 min free exploration o the two chambers as the baseline preference, and the baseline preference was calculated for each mouse by averaging data from a 3 d pretest. Mice were equally divided into two subgroups according to their baseline preference. Conditioning in the METH group was confined to the preferred chamber (S-paired) for 45 min paired with a saline (0.2 ml, i.p.) injection and nonpreferred chamber (METH-paired) for 45 min paired with a METH (1 mg/kg, i.p.) injection on alternate days (METH on odd days and saline on even days, 8 d). The saline group received saline (0.2 ml, i.p.) injection on each CPP training day. On Day 9, cohort 1 mice were subjected to the CPP test, allowing them to freely access the two chambers without any injection for 15 min, forming S’ and M’ groups; cohort 2 mice continued to the drug-paired CPP training, forming S and M groups.
The CPP score was calculated by subtracting the time spent in the S-paired chamber from the M-paired chamber. The ΔCPP score was the test CPP score minus the pretest CPP score. Primitive density map is the raw spatial distribution map calculated through spatial filtering on animal's visiting frequency at neighbor pixels, which weight as functions of radius to the current pixel.
Rota-rod task
Motor coordination and balance were measured by the accelerating rota-rod test (4–40 rpm in 5 min; Chen et al., 2018), in which mice were placed on the 3-cm-diameter rota-rod cylinder (Ugo Basil) during the 5 min test. Mice were initially trained to maintain themselves in a neutral position on the rod, and the latency to fall off the rota-rod was calculated.
OFT
General locomotor activity was measured in an open-field arena. Mice were placed in the chamber and allowed to freely explore for 5 min. The total distance traveled was recorded and analyzed by an automated detection system (TopScan).
Sucrose reinforcement
Sucrose reinforcement experiments were conducted in the Skinner box (Harvard Apparatus). During all daily 0.5 h training sessions, each active lever press resulted in the delivery of 0.1 ml of 10% sucrose under an FR1 reinforcement schedule with a 5 s timeout. The number of active and inactive lever presses and the volume of sucrose were recorded. An inactive lever press had no programmed consequences. During the training phase, mice were water restricted but allowed to drink freely for 1 h after the daily training.
Stereotaxic surgery and viral regulatory strategies
All viruses were generated and packaged by BrainVTA. Mice were head-fixed with a stereotactic frame (RWD Life Science) under isoflurane anesthesia (2% induction, 0.5% maintenance, RWD Life Science, R510-22-10). In addition, 100 nl of viruses or fluorogold (FG, Fluorochrome, 4716905) was injected using a glass micropipette attached to an infusion pump (Drummond) over 5 min at a rate of 20 nl/min. We waited for 10 min both before and after each injection. The coordinates used here were as follows: Med (AP, −6.40 mm; ML, ±0.75 mm; DV, −3.55 mm) and lobule VI (AP, −6.80 mm; ML, 0 mm; DV, −1.20 mm).
For anterograde tracing, 100 nl of a 1:1 volume mixture of rAAV2/9-mTH-Cre (BrainVTA, PT0781) and rAAV2/9-L7-DIO-mCherry (BrainVTA, PT-5402) was injected in the lobule VI. The first vector delivered Cre recombinase under the control of a mouse tyrosine hydroxylase (mTH) promoter fragment, whereas the second vector delivered a Cre recombinase–dependent mCherry tag. For retrograde tracing, 100 nl FG (4%) was injected unilaterally in the Med.
For the designer receptors exclusively activated by designer drug (DREADD) experiments, a combination of viruses rAAV2/9-mTH-Cre (BrainVTA, PT0781) and rAAV2/9-L7-DIO-hM4D(Gi)/hM3D(Gq)-mCherry (BrainVTA, PT-5404/PT-5403) was injected into the lobule VI. For the CNO off-target verification experiments, rAAV2/9-mTH-mCherry (BrainVTA, PT-2412) was injected into the lobule VI. Subsequently, pedestal guide cannulas (27 gauge, RWD Life Science) were implanted 1 mm above the Med. Clozapine N-oxide (CNO, 1 mM, 500 nl, MedChemExpress, HY-17366) or vehicle was locally infused into the Med. During CPP training, CNO (1 mM, 500 nl, MedChemExpress, HY-17366) or vehicle was locally infused into the Med 5 min before every METH injection day (Day 1, Day 3, Day 5, and Day 7) to regulate lobule VI PCTH+–Med pathway.
For the tettox experiments, 100 nl of a 1:1 volume mixture of rAAV2/9-CaMKII-Cre (BrainVTA, PT-0220) and rAAV2/9-EF1α-DIO-tettoxlc-mCherry (BrainVTA, PT-2139; or rAAV2/9-EF1α-DIO-EGFP, BrainVTA, PT-0795) was injected bilaterally in the Med.
After the behavioral tests, an accurate assessment of virus expression was conducted, and the off-target injected mice were excluded.
Fiber photometry
A combination of viruses rAAV2/9-mTH-Cre (BrainVTA, PT0781), rAAV2/9-L7-DIO-hM3D(Gq)-mCherry (BrainVTA, PT-5403), and rAAV2/9-Ef1α-DIO-axon-GCaMP6m (BrainVTA, PT-1225) was injected into the lobule VI. Subsequently, cannulas and optical fibers (200 μm outer diameter, Thinkerbiotech) were implanted 1 mm above the Med. After 3 weeks, the GCaMP6m signals from axonal terminals of lobule VI PCTH+ within the Med were recorded at homecage. The GCaMP6m signals were recorded and analyzed by ThinkerTech TrippleColor MultiChannel Fiber Photometry Acquisition Software and ThinkerTech TrippleColor MultiChannel Fiber Photometry Analysis Package (Thinkerbiotech). The baseline fluorescence signal was recorded for 10 min prior to CNO or vehicle treatment, followed by real-time recording of the fluorescence signal for 30 min. Mergence of the raw heatmap data from a single mouse was used as a statistical point and normalized by area under the curve (AUC) normalization. The AUC is the integral underrecording duration related to the corresponding baseline at every trial.
Immunofluorescence
The mice were perfused with 0.9% saline followed by 4% paraformaldehyde (PFA) in PBS buffer. The brains were removed and postfixed in 4% PFA at 4°C overnight and then transferred to 30% (w/v) sucrose. Frozen sections (30 μm) were cut on a cryostat (Leica). Brain sections were rinsed in PBS, permeabilized with PBS with 0.3% Triton X-100 (PBS-Tx) for 30 min, and then blocked with PBS-Tx containing 5% normal donkey serum for 1.5 h. The sections were incubated with the primary antibodies overnight at 4°C, followed by the corresponding fluorophore-conjugated secondary antibodies for 1.5 h at room temperature. The following primary antibodies were used as follows: mouse anti-NeuN (1:1,000, 94403S, Cell Signaling Technology, RRID: AB_2904530), rabbit polyclonal anti-NeuN (1:500, 2407S, Cell Signaling Technology, RRID: AB_2651140), mouse anti-CaMKII (1:150, sc-13141, Santa Cruz Biotechnology, RRID: AB_626789), rabbit polyclonal anti-TH (1:250, BM4568, Boster Bio), rabbit polyclonal anti-c-fos (1:500, 226003, Synaptic Systems, RRID: AB_2231974). The secondary antibodies were used as follows: Alexa Fluor 555–labeled donkey anti-mouse IgG (1:500, A322773, Invitrogen, RRID: AB_2762848), Alexa Fluor 555–labeled donkey anti-rabbit (1:500, A32794, Invitrogen, RRID: AB_2762834), Alexa Fluor 488–labeled donkey anti-mouse (1:500, A21202, Invitrogen, RRID: AB_141607), and Alexa Fluor 488–labeled donkey anti-rabbit (1:500, A32790, Invitrogen, RRID: AB_2762833). The images of the entire brain region were captured by THUNDER Imaging Systems TCS SP8 (Leica) using a 10× objective (100× total magnification, Figs. 1E,J,P, 2B, 3G, left, 4C,D, left, 5D) or 20× objective (200× total magnification, Figs. 1B,C left, F,I,K, left, 1Q, 2C, 3E, 3G, left, 3I,N,P, 4C right, 4D right, 5D,G). Confocal z-stack images were obtained using a TCS SP8 confocal microscope (Leica) with a 63× oil-immersion objective (630× total magnification, Δz = 2 µm, 15 z-stack; Figs. 1C right, 1K left, 5F).
Referring to the Paxinos and Franklin atlas of the mouse brain, the slices containing Med or lobule VI were chosen based on anatomical landmarks representing the bregma. The whole brain slice was scanned by a microscope, and the cells expressing c-fos, NeuN, CaMKII, TH, tettox-mCherry, and hM4Di/hM3Dq-mCherry were counted within the entire Med and lobule VI in each slice.
The data were quantified manually using Las.X 3.7.4.23463 software. The statistical graphs displayed the sample size, with each sample representing one mouse. Quantification data for each mouse was obtained by averaging measurements from three slices. The final qualification data was derived from the average of all samples within the group. For c-fos quantification, the percentage of c-fos–positive cells among markers (such as CaMKII, TH, and NeuN) was assessed in lobule VI or Med. When evaluating the transfection efficiency and specificity of viruses in tettox and DREADDs experiments, the percentage of mCherry- or EGFP-positive cells (transfected cells) among marker-positive cells was determined.
Patch clamp
Slices preparation was performed as previously described (Ge et al., 2021). Mice were deeply anesthetized with isoflurane (RWD Life Science, R510-22-10) and perfused with the ice-cold cutting solution. Sagittal slices containing the lobule VI were cut at 200 μm thickness using a vibratome in a 4°C cutting solution. The slices were transferred to 37°C cutting solution for 9 min and then transferred to holding solution to allow for recovery at room temperature for 1 h before recordings. During electrophysiological recordings, the brain slice was continuously perfused with oxygenated artificial CSF (aCSF) maintained at 30°C by an in-line solution heater (TC-324C, Warner Instruments). Loose-patch electrodes were filled with aCSF, and access resistance was maintained at 20–50 MΩ throughout the experiment. Recordings were performed under current-clamp mode with 0 holding current. All signals were filtered at 4 kHz, amplified at 5× using a MultiClamp 700B amplifier (Molecular Devices), and digitized at 10 kHz with a Digidata 1440A analog-to-digital converter (Molecular Devices). All data were analyzed with the Clampfit 10.6 software (Molecular Devices).
Statistical analysis
Statistical analysis was carried out using GraphPad Prism 9.0 software. The data are presented as the mean ± SD. Statistical significance was set as *p < 0.05, **p < 0.01, and ***p < 0.001. The two-tailed parametric unpaired t tests passed the Anderson–Darling test, D’Agostino–Pearson omnibus test, Shapiro–Wilk test, and Kolmogorov–Smirnov test. The one-way repeated-measure (RM) ANOVAs (parametric) followed by Bonferroni's post hoc test and passed the Brown–Forsythe and Bartlett's tests. The two-way RM ANOVAs were corrected with Geisser–Greenhouse and followed by Bonferroni's post hoc test.
Results
Med glutamatergic neurons receive lobule VI PCTH+ innervation
To identify the pathway of lobule VI PCTH+ to the Med, a virus tracing experiment was performed in mice, and calcium/calmodulin-dependent protein kinase II (CaMKII) was used as a marker of glutamatergic neurons (Kahn et al., 2019). Firstly, an anterograde tracer with mTH promotor expressed Cre (Chan et al., 2017), and the L7 promotor expressed mCherry (Kayakabe et al., 2013) was unilaterally injected into cerebellar lobule VI to specifically transfect PCTH+ cells (Fig. 1A). As shown in Figure 1B, mChery were highly overlapped with TH staining in lobule VI. The mCherry-labeled axon terminals from lobule VI PCTH+were abundantly distributed in the Med, surrounding the CaMKII-positive glutamatergic neurons (MedCaMKII, Fig. 1C). Next, a retrograde tracer (FG) was unilaterally injected into the Med (Fig. 1D,E). The results showed that FG-positive retrogradely marked PCs were highly expressed in lobule VI (Fig. 1F), with 67.45% of Med-projecting PCs (FG+ cells) being TH+ cells and 70.52% of TH+ cells coexpressed with FG (Fig. 1G). These results indicate that there exist direct progections from lobule VI PCTH+ to Med in the cerebellum, forming the lobule VI PCTH+–MedCaMKII pathway.
Lobule VI PCTH+ sends direct projections to and negatively innervates the Med glutamatergic neurons. A, Schematics of anterograde tracer virus injection. B, Representative images of mCherry fluorescence and immunostaining with TH and DAPI in the lobule VI. Scale bars, 1 mm (left) and 100 μm (right). C, Representative images of the anterograde virus-mCherry fluorescence and immunostaining with CaMKII and DAPI in the Med. Scale bars, 300 μm (left) and 50 μm (right). D, Schematics of retrograde tracer FG injection. E, Representative image of FG fluorescence and immunostaining with TH in the Med. Scale bar, 300 μm. F, Representative images of the retrograde FG fluorescence and immunostaining with TH in the lobule VI. Scale bars, 1 mm (left) and 100 μm (right). G, Proportions of the FG and TH double-positive cells in the lobule VI. H, Schematics of virus injection, cannula, and fiber implantation. I, Representative image and quantification of fluorescent virus tag mCherry with TH immunostaining in the lobule VI. Scale bar, 100 μm. J, Representative image of fluorescent virus tag GCaMP6m and mCherry, along with cannula and fiber implantation. Scale bar, 1 mm. K, Representative images of fluorescent virus tag GCaMP6m, mCherry, and DAPI immunostaining in the Med. Scale bar, 200 μm (left) and 20 μm (right). L, Heatmap and sample traces of the GCaMp6m fluorescence (axonal terminals of lobule VI PCTH+) in the Med. M, The AUC of the normalized GCaMP signal. Two-way RM ANOVA. Ftreatment × time (3,6) = 10.77, p = 0.0079. *p = 0.00120, ***p = 0.0002, vehicle group versus that of the CNO group. n = 2 per group. N, Schematics of virus injection and cannula implantation. O, Schematics of the cannula placements in the Med. P, Representative images of fluorescent virus tag mCherry in the lobule VI and cannula implantation in the Med. Scale bars, 100 μm (left) and 1 mm (right). Q, Representative images of fluorescent virus tag mCherry and immunostaining with c-fos and NeuN in the Med. Scale bars, 300 μm (left) and 100 μm (right). R, Percentage of c-fos and NeuN double-positive neurons in the Med. Two-tailed unpaired t test, t(22) = 3.604, **p = 0.0016. n = 12 per group. The arrows indicate representative colabeled cells.
In order to validate the physiological validation of the viral vectors, we explored DREADD strategies in combination with fiber photometry in male naive mice. As shown in Figure 1H, a combination of viruses rAAV2/9-mTH-Cre, rAAV2/9-L7-DIO-hM3D(Gq)-mCherry, and rAAV2/9-Ef1α-DIO-axon-GCaMP6m was injected into the lobule VI, and cannulas and fibers were placed in the Med. As depicted in Figure 1I–K, viruses were restrictively expressed in the lobule VI and highly overlapped with TH staining, GCaMP6m, and mCherry-tagged axon terminals from lobule VI PCTH+ closely surrounded Med neurons. As shown in Figure 1L,M, compared with vehicle treatment, cannula CNO treatment activated more axonal terminals of lobule VI PCTH+ within the Med, evidenced by increased calcium signals in the Med from 10 min, and sustained up to 30 min.
To observe the regulatory role of PCTH+ on the Med neurons, a mixture of virus (AAV-mTH-Cre and AAV-L7-DIO-hM4D(Gi)-mCherry) was unilaterally injected into lobule VI, and the cannulas were placed bilaterally in the Med (Fig. 1N–P). Vehicle or CNO (1 mM, 500 nl) was locally infused through the cannula to inhibit the axonal terminals of lobule VI PCTH+ in the Med. As shown in Figure 1Q,R, CNO treatment (inhibiting the PCTH+ terminals within the Med) increased the percentage of c-fos–positive neurons in Med neurons, indicating that lobule VI PCTH+ neurons negatively innervated Med activity in the cerebellum.
To assess the potential off-target effects of CNO, we conducted a control experiment in naive mice without DREADDs. Virus rAAV2/9-mTH-mCherry, which lacks Gi or Gq elements, was injected into the lobule VI, and the cannulas were placed bilaterally in the Med (Fig. 2A). Immunostaining revealed a high overlap between mCherry-positive cells and TH staining in the lobule VI (Fig. 2B). Three weeks after virus expression, either vehicle or CNO was locally infused through the cannula. There was no difference in the proportions of c-fos and NeuN double-positive cells in the Med between the vehicle and CNO groups, indicating that CNO treatment did not exert a substantial impact on c-fos expression in Med neurons in the absence of DREADDs (Fig. 2C,D). These findings indicate that CNO does not possess any off-target potential in the current study.
CNO treatment did not exert a substantial impact on c-fos expression in Med neurons in the absence of DREADDs. A, Schematics of virus injection and cannula implantation. B, Representative image of mCherry fluorescence and immunostaining with TH in the lobule VI. Scale bar, 100 μm. C, Representative images of fluorescent virus tag mCherry and immunostaining with c-fos and NeuN in the Med. Scale bars, 300 μm (left) and 100 μm (right). D, Percentage of c-fos and NeuN double-positive cells in the Med. Two-tailed unpaired t test, t(10) = 0.1530, p = 0.8814. n = 6 per group. The arrows indicate representative colabeled cells.
METH inhibits lobule VI PCTH+ and excites MedCaMKII during CPP training
A standard METH CPP procedure was performed in male mice to assess the acquisition of METH CPP. All mice were subjected to METH injection on odd days (Days 1, 3, 5, and 7) while subjected to saline injection on even days (Days 2, 4, 6, and 8, Fig. 3A). After CPP training (Days 1–8), the mice randomly divided into cohort 1 and cohort 2. On Day 9, cohort 1 mice were subjected to CPP test, forming S’ and M’ groups, whereas cohort 2 mice were continued to CPP training, forming S and M groups (Fig. 3A). The acquisition of METH CPP was evaluated by the CPP score, which calculated by subtracting the duration spent in the S-paired chamber from the METH-paired chamber (Zhou et al., 2019).
METH inhibits lobule VI PCTH+ and excites cerebellar MedCaMKII during CPP training. A, Scheme of METH CPP model. B, CPP scores during the pretest and CPP test in cohort 1 mice. Two-way RM ANOVA. Ftreatment × session (1,18) = 9.689, p = 0.0060. ###p < 0.0001, CPP test versus pretest of the M’ group. ***p = 0.0009, CPP test of the M’ group versus that of the S’ group. n = 10 per group. C, ΔCPP scores (CPP test minus pretest) of the S’ and M groups. Two-tailed unpaired t test, t(18) = 3.11, ***p = 0.0060. n = 10 per group. D, Position density maps of a single mouse during the pretest and CPP test in cohort 1 mice. Primitive density map is the raw spatial distribution map calculated through spatial filtering on an animal's visiting frequency at neighbor pixels, which weigh as functions of radius to the current pixel. E, Immunohistochemistry for c-fos/TH/DAPI in the lobule VI following METH CPP test in cohort 1 mice. Scale bars, 1 mm (left) and 100 μm (right). F, Percentage of c-fos and TH double-positive neurons in the lobule VI PCTH+ following METH CPP test in cohort 1 mice. Two-tailed unpaired t test, t(10) = 0.0789, p = 0.9386. n = 6 per group. G, Immunohistochemistry for c-fos/NeuN/DAPI in the Med following METH CPP test in cohort 1 mice. Scale bars, 300 μm (left) and 100 μm (right). H, Percentage of c-fos and NeuN double-positive neurons in the Med following METH CPP test in cohort 1 mice. Two-tailed unpaired t test, t(8) = 0.4502, p = 0.6645. n = 5 per group. I, Immunohistochemistry for c-fos/TH/DAPI in the lobule VI following METH-induced CPP acquisition in cohort 2 mice. Scale bars, 1 mm (left) and 100 μm (right). J, Percentage of c-fos and TH double-positive cells in the lobule VI PCTH+ following METH-induced CPP acquisition in cohort 2 mice. Two-tailed unpaired t test, t(12) = 2.396, *p = 0.0337. S group, n = 8; M group, n = 6. K, Representative images of electrophysiological recordings on the lobule VI PCs following METH-induced CPP acquisition in cohort 2 mice. Scale bars, 1 mm (middle) and 50 μm (right). L, Frequency of spontaneous firing in the lobule VI PCs following METH-induced CPP acquisition in cohort 2 mice. Two-tailed unpaired t test, t(33) = 2.37, *p = 0.0241. S group, n = 21 cells; M group, n = 14 cells. M, Sample traces of spontaneous firing in the lobule VI PCs following METH-induced CPP acquisition in cohort 2 mice. N, Immunohistochemistry for c-fos/NeuN/DAPI in the Med following METH-induced CPP acquisition in cohort 2 mice. Scale bars, 300 μm (left) and 100 μm (right). O, Percentage of c-fos and NeuN double-positive neurons in the Med following METH-induced CPP acquisition in cohort 2 mice. Two-tailed unpaired t test, t(10) = 3.459, **p = 0.0061. n = 6 per group. P, Immunohistochemistry for c-fos/CaMKII/DAPI in the Med following METH-induced CPP acquisition in cohort 2 mice. Scale bars, 300 μm (left) and 100 μm (right). Q, Percentage of c-fos and CaMKII double-positive cells in the MedCaMKII following METH-induced CPP acquisition in cohort 2 mice. Two-tailed unpaired t test, t(10) = 5.639, ***p = 0.0002. n = 6 per group. The arrows indicate representative colabeled cells.
For cohort 1 mice, M’ mice had higher test scores than their pretest scores, as well as much higher ΔCPP scores than those of the S’ mice (Fig. 3B–D), indicating the acquisition of METH CPP in M’ mice. In addition, 75 min after the CPP test, the neuronal activity of Med and lobule VI PCTH+ was measured by c-fos staining. There was no difference in the proportions of c-fos and TH double-positive cells in the lobule VI PCTH+ between S’ and M’ groups (Fig. 3E,F), as well as the proportions of c-fos and NeuN double-positive cells in the Med (Fig. 3G,H). These results indicate that METH CPP test does not affect the activity of lobule VI PCTH+ and Med.
For cohort 2 mice, the activity of lobule VI PCTH+ and Med was assessed 45 min after the last drug-paired CPP training on Day 9. Compared to the S mice, M mice exhibited a reduction in the proportions of c-fos and TH double-positive cells in lobule VI PCTH+ (Fig. 3I,J). Furthermore, brain slices from M mice exhibited a decreased frequency of spontaneous firing in lobule VI PCs compared with S mice (Fig. 3K–M). Additionally, M mice displayed an increased percentage of c-fos and NeuN double-positive cells in the Med (Fig. 3N,O), as well as an increased percentage of c-fos and CaMKII double-positive cells (Fig. 3P,Q). These findings demonstrate that METH CPP acquisition inhibits lobule VI PCTH+ while exciting MedCaMKII in the cerebellum.
Silencing MedCaMKII suppresses the acquisition of METH CPP in METH-exposed mice but produces motor coordination deficits in naive mice
To determine the role of Med in the acquisition of METH CPP, a combination of viruses rAAV2/9-CaMKII-Cre and rAAV2/9-EF1α-DIO-tettoxlc-mCherry (tetanus toxin light chain, tettox; Zhou et al., 2018) was injected bilaterally into the Med to silence MedCaMKII before CPP procedure (Fig. 4A). The CaMKII promoter expressed Cre in the glutamatergic neurons and then restricted Cre-dependent tettox expression to MedCaMKII neurons. Viruses were expressed restrictedly in the Med and highly overlapped with CaMKII staining (Fig. 4B–D). As shown in Figure 4E–G, METH exposure increased the CPP score in control virus-injected but failed to affect that in tettox-injected mice when compared with the corresponding pretest, as well as induced a decrease in ΔCPP score in tettox group than the control group. There was no difference in the distance traveled on the pretest day between the tettox and control groups (Fig. 4H). When compared with control virus-injected METH mice during the METH CPP training, tettox-injected mice traveled less distance on METH injection days (odd days), whereas traveled a similar distance on saline injection days (even days; Fig. 4I–J). These results indicate that silencing MedCaMKII by tettox attenuates METH-sensitized behaviors, and has no influence on locomotor behaviors in mice.
Silencing MedCaMKII suppresses CPP acquisition in METH mice but produces motor coordination deficits in naive mice. A, Experimental scheme of METH-induced CPP acquisition. B, Schematics of virus injection and expression spread. C, Representative images and quantification of virus-EGFP fluorescence in the Med and immunostaining with CaMKII and DAPI of the control group. Scale bars, 1 mm (middle) and 300 μm (right). D, Representative images and quantification of virus-mCherry fluorescence in the Med and immunostaining with CaMKII and DAPI of the Tettox group. Scale bars, 1 mm (middle) and 300 μm (right). E, CPP scores during the pretest and CPP test. Two-way RM ANOVA. Ftreatment × session (1,19) = 5.530, p = 0.0296. ###p = 0.0003, CPP test versus pretest of the control group. ***p < 0.0001, CPP test of tettox group versus that of the control group. Control group, n = 10; tettox group, n = 11. F, ΔCPP scores (CPP test minus pretest) of the two groups. Two-tailed unpaired t test, t(19) = 2.352, *p = 0.0296. Control group, n = 10; Tettox group, n = 11. G, Position density maps of a single mouse during the pretest and CPP test. H, Distance traveled during pretest. Two-tailed unpaired t test, t(19) = 1.198, N.S., p = 0.2458. Control group, n = 10; Tettox group, n = 11. I. Distance traveled during the CPP acquisition training. Two-way RM ANOVA. Ftreatment × session (7,133) = 17.0, p < 0.0001. ***p < 0.001 versus control group. Control group, n = 10; Tettox group, n = 11. J, Position density maps of a single mouse during Days 7 and 8. K, Experimental scheme of behavior tests. L, Latency to fall during the rota-rod test. Two-way RM ANOVA. Ftreatment × session (4,64) = 5.701, p = 0.0005. *p < 0.05, **p < 0.01, ***p < 0.001, Tettox group versus control group. Day 1, p = 0.0013; Day 2, p = 0.0133; Day 3, p < 0.0001; Day 4, p = 0.0034; Day 5, p = 0.0002. Control group, n = 11; Tettox group, n = 7. M, Distance traveled during the OFT. Two-tailed unpaired t test, t(16) = 1.860, N.S., p = 0.0814. Control group, n = 11; Tettox group, n = 7. N, Position density maps of a single mouse during the OFT. O, Number of active lever presses during the sucrose reinforcement. Two-way RM ANOVA. Ftreatment × session (9,144) = 1.220, p = 0.2870. *p < 0.05 versus the control group. Day 10, p = 0.0139; Day 11, p = 0.0475. Control group, n = 11; Tettox group, n = 7. P, Number of inactive lever presses during the sucrose reinforcement. Two-way RM ANOVA. Ftreatment × session (9,144) = 4.061, p = 0.0001. Control group, n = 11; Tettox group, n = 7. Q, The volume of sucrose consumption during the sucrose reinforcement. Two-way RM ANOVA. Ftreatment × session (9,144) = 1.827, p = 0.0682. **p = 0.0035 versus the control group. Control group, n = 11; Tettox group, n = 7. The arrows indicate representative colabeled cells.
To assess the role of MedCaMKII in motor coordination, general locomotor, and natural reward acquisition, naive mice were subjected to rota-rod test, OFT, and sucrose SA, respectively (Fig. 4K). As shown in Figure 4L, tettox-injected mice showed shorter latency to fall than the controls during the rota-rod test. However, there was no difference in the distance traveled in the OFT apparatus between the tettox and control groups (Fig. 4M,N). In parallel, the two groups exhibited similar levels of active lever pressing (Fig. 4O), inactive lever pressing (Fig. 4P), and sucrose consumption during the sucrose SA test (Fig. 4Q). These results indicate that silencing MedCaMKII has no effect on locomotor activity and natural reward acquisition but produce serious motor coordination deficits in naive mice.
Activating lobule VI PCTH+ terminals within the Med blocks the acquisition of METH CPP in METH-exposed mice and does not affect motor coordination in naive mice
To examine the role of lobule VI PCTH+–Med pathway in METH CPP acquisition, a mixture of virus (rAAV2/9-mTH-Cre and rAAV2/9-L7-DIO-hM3D(Gq)-mCherry) was injected into the lobule VI. The mTH promotor expressed Cre in the catecholaminergic neurons, and the L7 promotor expressed DREADDs in the PCs, finally restricting Cre-dependent DREADD expression in the lobule VI PCTH+ and their axon terminals (Fig. 5A,B). In parallel, the cannulas were placed bilaterally in the Med (Fig. 5B–D).
Activating lobule VI PCTH+–Med pathway during CPP training blocks CPP acquisition in METH mice and has no influence on motor coordination in naive mice. A, Experimental scheme of METH-induced CPP acquisition. B, Schematics of virus injection and cannula implantation. C, Schematics of the cannula placements. D, Representative images of fluorescent virus tag mCherry and cannula implantation. Scale bar, 1 mm. E, Representative image and quantification of fluorescent virus tag mCherry with TH immunostaining in the lobule VI. Scale bar, 100 μm. F, Representative image and quantification of fluorescent virus tag mCherry with CaMKII immunostaining in the Med. Scale bar, 50 μm. G, Representative images of fluorescent virus tag mCherry and immunostaining with c-fos and NeuN in the Med. Scale bars, 300 μm (left) and 100 μm (right). H, Percentage of c-fos and NeuN double-positive neurons in the Med. Two-tailed unpaired t test, t(22) = 6.126, ***p < 0.0001. n = 12 per group. I, CPP scores during the pretest and CPP test. Two-way RM ANOVA. Ftreatment × session (1,26) = 25.41, p < 0.0001. ###p < 0.0001, CPP test versus pretest of the vehicle group. ***p < 0.0001, CPP test of the CNO group versus the vehicle group. Vehicle group, n = 16; CNO group, n = 12. J, ΔCPP scores (CPP test minus pretest) of the two groups. Two-tailed unpaired t test, t(26) = 5.041, p < 0.0001. Vehicle group, n = 16; CNO group, n = 12. K, Position density maps of a single mouse during the pretest and CPP test. L, Distance traveled during the CPP acquisition training. Two-way RM ANOVA. Ftreatment × session (7,182) = 10.88, p < 0.0001. **p < 0.01 versus saline group. Saline group, n = 16; CNO group, n = 12. M, Average distance traveled during METH and saline injected days. Two-way RM ANOVA. Ftreatment × session (1,26) = 15.47, p = 0.0006. &&&p < 0.001 versus saline injected day of the saline group. ***p < 0.001 versus the saline group on the METH injected day. Saline group, n = 16; CNO group, n = 12. N, Position density maps of a single mouse during Days 7 and 8. O, Experimental scheme of behavior tests. P, Latency to fall during the rota-rod test. Two-way RM ANOVA. Ftreatment × session (4,104) = 2.156, p = 0.0792. Vehicle group, n = 16; CNO group, n = 12. Q, Distance traveled during the OFT. Two-tailed unpaired t test, t(26) = 0.59, p = 0.5575. Vehicle group, n = 16; CNO group, n = 12. R, Position density maps of a single mouse during the OFT. S, Number of active lever presses during the sucrose reinforcement. Two-way RM ANOVA. Ftreatment × session (9,234) = 2.276, p = 0.0184. Vehicle group, n = 16; CNO group, n = 12. T, Number of inactive lever presses during the sucrose reinforcement. Two-way RM ANOVA. Ftreatment × session (9,234) = 3.180, p = 0.0012. Vehicle group, n = 16; CNO group, n = 12. U, The volume of sucrose consumption during the sucrose reinforcement. Two-way RM ANOVA. Ftreatment × session (9,234) = 2.566, p = 0.0078. Vehicle group, n = 16; CNO group, n = 12. The arrows indicate representative colabeled cells.
During the 8 days CPP training period, CNO or vehicle was locally infused through the cannula to activate the axonal terminals of lobule VI PCTH+ in the Med 15 min before each METH injection day (Days 1, 3, 5, and 7). As shown in Figure 5E,F, viruses were restrictedly expressed in the lobule VI and highly overlapped with TH staining, and mCherry-tagged axon terminals from lobule VI PCTH+ closely surrounded MedCaMKII. Compared with vehicle-treated mice, CNO-treated mice showed decreased c-fos expression in the Med, suggesting that activating the axon terminals from lobule VI PCTH+ inhibited Med neuronal activity (Fig. 5G,H). In parallel, METH injection increased the CPP test score in vehicle-treated mice but failed to affect that in CNO-treated mice when compared with the corresponding pretest (Fig. 5I,K). The ΔCPP score in the CNO group was much lower than that in the vehicle group (Fig. 5J). CNO-treated mice traveled less distance on METH injection days (Days 5 and 7) when compared with vehicle-treated mice (Fig. 5L). CNO-treated mice showed no difference in distance traveled in the CPP apparatus between saline and METH injected days (Fig. 5M,N). These results suggest that activating lobule VI PCTH+–Med pathway blocks the acquisition of METH CPP.
To assess the role of lobule VI PCTH+–Med pathway in motor coordination, general locomotor, and natural reward acquisition, naive mice were subjected to rota-rod test, OFT, and sucrose SA, respectively (Fig. 5O). As shown in Figure 5P–R, there was no difference in the latency to fall in the rod test, as well as the distance traveled in the OFT apparatus between CNO-treated and vehicle-treated mice. In parallel, there was no difference in active lever presses (Fig. 5S), inactive lever presses (Fig. 5T), and sucrose water consumption (Fig. 5U) between the two groups during the sucrose SA test. These results suggest that activating lobule VI PCTH+–Med pathway does not alter motor coordination, locomotor activity, or natural reward acquisition in naive mice.
Discussion
In the present study, we identified a direct pathway from lobule VI PCTH+ to MedCaMKII, forming lobule VI PCTH+–MedCaMKII pathway in the cerebellum. The lobule VI PCTH+ has a negative innervation on the Med in naive mice. METH exposure suppressed lobule VI PCTH+ but activated MedCaMKII. Silencing MedCaMKII disrupted the acquisition of METH CPP, along with serious deficits in motor coordination. Selectively activating the lobule VI PCTH+ terminals within the Med blocked the acquisition of METH CPP without affecting motor coordination, gross locomotor, or natural reward (Fig. 6).
Schematic diagram of the present study. The lobule VI PCTH+–MedCaMKII pathway is identified in the cerebellum. During METH CPP training, lobule VI PCTH+ was suppressed while MedCaMKII was activated in male mice. Silencing MedCaMKII during CPP training alleviated the acquisition of METH CPP but resulted in serious deficits in motor coordination. In contrast, inhibiting the subregion of Med that is innervated by lobule VI PCTH+ blocked the acquisition of CPP in METH mice without influencing the motor coordination, locomotives, and natural reward behaviors in naive mice.
Chronic METH administration caused cerebellar neurodegeneration involving a decrease in the population of PCs (Boroujeni et al., 2020; Eskandarian Boroujeni et al., 2020; Ding et al., 2022). In addition, METH administration increased TH expression in the vermis PCs and the number of PCTH+ (Ferrucci et al., 2007). A single METH injection increased c-fos expression in the PCs, whereas repeated METH administration resulted in the hypofunction of PCs (Hamamura et al., 1999). Electrophysiological studies showed that repeated METH administration decreased the rebound action potential and spontaneous firing frequency in the PCs (Ramshini et al., 2021). In the present study, our results provided the initial evidence indicating the involvement of lobule VI PCTH+ in the acquisition of METH CPP, putting new insight into the role of lobule VI PCs in METH administration.
Generally, the drug CPP model recruits the brain circuits implicated in the process of reward and conditioned memory. Growing evidence demonstrated that the cerebellum appeared to be a key node for drug-related cues (Jasinska et al., 2014; Moulton et al., 2014; Li et al., 2015; Gil-Miravet et al., 2021). In human cocaine users, cocaine-related stimuli activated the cerebellum (Grant et al., 1996; Anderson et al., 2006). In rodents, cocaine-paired olfactory cue increased the activity of the granule cells in the cerebellar vermis (Carbo-Gas et al., 2014a,b), and cocaine-odor pairing procedure triggered the Med (Carbo-Gas et al., 2014b). In nonhuman primates, PCs encoded multiple reward-related signals during reinforcement learning (Sendhilnathan et al., 2021). One recent work revealed a direct monosynaptic pathway from the cerebellar DCN to the VTA, which was implicated in reward-related behavior (Carta et al., 2019). As to natural rewards such as food, the feeding procedure excited DCN neurons, whereas activating DCN neurons suppressed food intake (Low et al., 2021). Here, we found that the lobule VI PCTH+–Med pathway was responsive to METH during CPP training. Importantly, either silencing MedCaMKII or activating lobule VI PCTH+–Med pathway blocked the acquisition of METH CPP, indicating an involvement of lobule VI PCTH+–Med pathway in mediating METH-preferred behaviors. In contrast, neither silencing MedCaMKII nor activating lobule VI PCTH+–Med pathway affected sucrose SA in mice, suggesting a specific involvement of lobule VI PCTH+–Med pathway in the METH-preferred behaviors rather than sucrose SA.
At the end of the last century, TH had been identified to be expressed in cerebellar PCs of both humans and mice (Austin et al., 1992; Takada et al., 1993; Fujii et al., 1994; Sakai et al., 1995). However, the functional role remains largely unexplored. It is well known that TH is involved in the synthesis of catecholamine neurotransmitters such as dopamine, norepinephrine, and epinephrine. A query of the Allen Mouse Brain ISH database revealed that cerebellar PCs lack the necessary enzymes required for dopamine synthesis, release, and reuptake, including dopa decarboxylase, vesicle monoamine transporter type 2 (VMAT2), and dopamine transporter (DAT). Consistently, Kim et al. (2009) found extremely weak expression of VMAT2 in the cerebellar lobule VI. Furthermore, no phosphor-TH immunoreactive PCs were observed in the vermis of the cerebellum (Lee et al., 2006). Similar to the cerebellum, striatum TH-positive neurons neither express VMAT2 and DATs, nor release dopamine, which are regarded as inhibitory interneurons (Xenias et al., 2015). Therefore, we thought that the lobule VI PCTH+ may be TH-positive inhibitory neurons instead of dopamine-synthesizing neurons.
As to the role of TH and PCTH+ in the cerebellum, Carlson et al. (2021) found that conditional knock-out of TH in fibers innervating the Lat resulted in cognitive deficits, but not motor learning deficits. Locke et al. (2020) reported that knocking out TH in cerebellar PCs impaired behavioral flexibility, social cognitive memory, and associative fear learning without influencing gross motor, sensory, instrumental learning, and sensorimotor gating functions. In the current study, activating lobule VI PCTH+ terminals within the Med attenuated the neuronal activities of Med, implying that lobule VI PCTH+ might negatively regulate Med activity through inhibitory neurotransmitters such as the GABAergic synaptic system. Further studies are required to explore the synaptic elements of lobule VI PCTH+–Med pathway. At least, we found that lobule VI PCTH+–Med was involved in the acquisition of METH CPP with no significant influence on locomotor, motor coordination, and natural reward.
METH-induced neurotoxicity and behavioral deficits are in a sex-specific manner in both humans (Cheng et al., 2023) and mice (Fan et al., 2022). It has been believed that hormonal fluctuations during the estrous cycle are one of the contributors to sex-dependent differences in addiction (Hudson and Stamp, 2011; Broderick and Malave, 2014). In this study, the utilization of exclusively male mice models to evaluate the involvement of the cerebellum in METH CPP acquisition represents a constraint. Subsequent investigations should be conducted to investigate the potential impact of sex on the regulatory effects of the lobule VI PCTH+–Med pathway on METH CPP acquisition and to ascertain the mechanisms underlying such effects. Additionally, we did not measure the phenotypes of TH− neurons in lobule VI and their projections to Med in the acquisition of METH CPP, which need to be further explored in future studies. The primary objective of this study is to determine the accurate neural projection to the Med that influences METH preference behavior, thereby providing a potential target with minimal side effects for future intervention in addictive-like behaviors.
Collectively, our findings identified a novel cerebellar lobule VI PCTH+–MedCaMKII pathway and revealed its potential role in the acquisition of METH CPP. Both MedCaMKII and lobule VI PCTH+–Med pathway were involved in mediating METH-preferred behaviors. Importantly, selectively suppressing the subpopulations of Med neurons innervated by lobule VI PCTH+ proved more effective in alleviating METH-preferred behaviors compared with inhibiting the entire Med population, as indicated by no deficits in motor coordination in naive mice.
Footnotes
This work is supported by the National Natural Science Foundation of China (82271531 and 82071495), the Open Project of Chinese Materia Medica First-Class Discipline of Nanjing University of Chinese Medicine (2020YLXK004), and Traditional Chinese Medicine Technology Development Project of Jiangsu Province, China (ZD202302 and YB201907).
↵*F.G., Z.W., and W.Y. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Xiaowei Guan at guanxw918{at}njucm.edu.cn or Jiasong Chang at 331880557{at}qq.com.
