The 3-NP was solubilized in phosphate buffer (PB) and pH-adjusted to 7.4. Intraperitoneal (i.p.) injection of 3-NP crosses the blood–brain barrier and mimics the neuropathology of HD (Hamilton and Gould, 1987; Borlongan et al., 1997, Rodríguez et al., 2010). The treatment protocol involved daily i.p. injections of 3-NP at a dosage of 15 mg/kg for a period of 5 days. Control mice, in contrast, received only PB (i.p.) throughout the treatment duration. All experiments were conducted 2 days following the administration of the last treatment injection.

Mice were anesthetized using halothane in a drop-in induction chamber and euthanized by decapitation. The brain was swiftly extracted and immersed in ice-cold (4°C) low-calcium saline, comprising 130 mM NaCl, 3 mM KCl, 5 mM MgCl₂, 26 mM NaHCO₃, 1.25 mM NaH₂PO₄, 1 mM CaCl₂, and 10 mM glucose, to protect from excitotoxicity-induced damage during the cutting process. Sagittal brain slices (300 μm) were obtained using a vibratome (PELCO, 1000 Plus Sectioning System, Ted Pella Inc., Redding, CA, United States). These sections were maintained in the same low-calcium saline solution and continuously oxygenated (95% O₂, 5% CO₂) for 1 h before electrophysiological experiments.

Following the incubation period, the slices were transferred to a recording chamber continually perfused with recording saline (130 mM NaCl, 3 mM KCl, 2 mM MgCl_2., 26 mM NaHCO₃, 1.25 mM NaH₂PO₄, 2 mM CaCl₂, and 10 mM glucose) and consistently bubbled with 95% O₂/5% CO₂. Borosilicate electrodes (3–5 MΩ) filled with an internal solution containing 72 mM KH₂PO₄, 36 mM KCl, 2 mM MgCl₂•6H₂O, 10 mM HEPES sodium salt, 1.1 mM EGTA, 0.2 mM Na₂ATP, 0.2 mM Na₃GTP, and 5 mM QX-314 were used for whole-cell voltage clamp recordings. QX-314 was present in the internal solution to inhibit voltage-gated sodium channels and avoid the influence of neuronal firing.

Medium spiny neurons (MSN) were visualized using infrared differential interference contrast (DIC) illumination on an upright microscope (BX51WI, Olympus) equipped with water immersion objective magnification (Olympus X Lum Plan Fl 20x/0.95 W, Japan) coupled to a CCD camera (charge-coupled-device, Hitachi) and identified by their membrane properties (input resistance, membrane capacitance, and time constant, see Table 1). Bicuculline (10 μM) was present in all experiments to study corticostriatal excitatory postsynaptic currents (EPSCs). In experiments aimed at assessing the role of Cdk5 in plasticity, the inhibitor roscovitine (20 μM) was applied via bath application before stimulation. For experiments targeting D1 receptors, the antagonist SCH23390 (1 μM) was applied before roscovitine. Similarly, experiments focusing on PKA involved the bath application of the inhibitor H89 (5 μM) prior to roscovitine application. Electrical stimulation was conducted in the corpus callosum using a tungsten bipolar electrode coupled to a constant voltage stimulator isolator (DS2A-MK. II, Digitimer Ltd., Hertfordshire, United Kingdom). To elicit EPSC responses, a stimulus ranging from 10 to 20 V, with a duration of 50 to 100 μs and a frequency of 0.1 Hz, was applied using an isolated stimulator. The stimulus intensity was calibrated to 50% of the maximum amplitude of the EPSC response. In all cases, a paired-pulse stimulation consisting of two stimuli (S₁ and S₂) of the same duration (100 μs) and intensity (previously calibrated), separated by an interval of 30 to 50 ms at 0.1 Hz, was delivered through the bipolar electrode (See Supplementary Figure S1).

Recorded EPSCs from MSNs were amplified and digitized using an Axopatch 200B amplifier coupled to a Digidata 1322A digitizer (Axon, Molecular Devices Corp.CA, United States). Throughout the experiment, input and access resistance were continuously monitored by evoking a transmembrane current with a voltage command (10 mV) in a whole-cell voltage clamp. Time constants were determined after whole-cell compensation, and capacitance was measured using the Axopatch dial. The access resistance (RN) was determined by fitting the I-V relation with a polynomial function. Throughout each experiment, monitoring was facilitated through the membrane test function of pClamp software at −70 mV, ~100 Hz acquisition rate, and 1 Hz filtering. Acceptance criteria included an access resistance below 30 MΩ, consistent across all experiments. Voltage holding was maintained at −70 mV.

For synaptic plasticity induction, high-frequency stimulation (HFS) consisting of three trains of 100 Hz, lasting 3 s each, with intervals of 20 s between each train, was administered. The amplitudes of EPSCs before and after the trains were analyzed in the last 10 min before the train and after the train during the last 10 min of the time course.

Mice were deeply anesthetized with halothane prior to transcardial perfusion as described above. The perfusion process involved initially flushing the circulatory system with a saline solution (NaCl 0.15 M), followed by perfusion with phosphate-buffered solution (PB 0.1 M and pH 7.4) containing 4% paraformaldehyde (PFA) in 0.15 M PB at pH 7.4. This procedure was carried out at 4°C to ensure optimal preservation of tissue integrity.

Following the removal from the skull, the brains were postfixed in the same fixative at room temperature (RT) for 2 h. Subsequently, they were transferred to a PB-sucrose solution (30%) to facilitate cryopreservation. The brains were then frozen in cold 2-methylbutane and cut into 30-μm-thick coronal sections using a cryostat. These sections were systematically collected into 24-well dishes filled with a blocking solution, preparing them for subsequent immunohistochemistry analyses (3% bovine serum albumin, 0.1% Triton X-100, and 0.025% sodium azide in PB).

The brain sections underwent an overnight incubation in a blocking serum at 4°C. Following this, the tissue slices were treated with a primary antibody targeting Cdk5 (1:500; Santa Cruz Biotechnology, Inc.), appropriately diluted in the blocking solution, and incubated for 24 h at 4°C. Subsequently, three 15-min washes were conducted using the blocking solution to ensure proper antibody removal. The tissue slices were then subjected to a 90-min incubation at room temperature with the corresponding biotinylated secondary antibody (Vector Laboratories, Burlingame, CA, United States) diluted (1:500) in the blocking solution.

Following three washes with PB, the slices underwent incubation using an avidin–peroxidase kit (VECTASTAIN® ABC-HRP, Peroxidase Staining Kit, Vector). Peroxidase activity was subsequently revealed and intensified using 3,3′-diaminobenzidine as a chromogen and nickel chloride as outlined in the supplier’s protocol. The staining process was halted after approximately 5 min by thoroughly washing the slices with PB. The treated slices were then carefully mounted on gelatin-coated slides, left to air-dry overnight, and cover-slipped with Permount (Fisher Chemical™). Specific immunostaining was detected as a diffuse brown precipitate.

To assess the quantity of Cdk5-positive neurons, brain serial coronal sections from six mice (three mice per experimental group) containing the striatum were digitized (4X, 10X, and 40X). These digitized images were captured using a light microscopy setup adapted to an image analysis system (Universal Imaging, Motic). Subsequently, the images were utilized to document the expression of Cdk5 in the striatum and subjected to analysis through ImageJ software provided by the NIH. Then, an unbiased counting frame of 2 mm² per field (10X) was generated, and the frames were obtained from the dorsal striatum (one frame per slice and nine slices per mouse brain), enabling the counting of stained cells. The brightness, contrast, and threshold settings remained unchanged, ensuring consistency, while the particle analyzer of ImageJ software was employed to accurately detect the cells. Following the comprehensive counting of all cells within the sampling frame, the total number of cells was estimated. Statistical analysis of cell counts was performed using Student’s t-test, and significance was established at p < 0.05.

To evaluate the phosphorylation status of DARPP-32 at Thr-75, DARPP-32 at Thr-34, and total DARPP-32, striatal slices from animals treated with either 3-NP (n = 5 mice) or vehicle control (n = 5 mice) were incubated in the physiological saline solution used for electrophysiological experiments, within an 8-well Petri dish that was continuously oxygenated (95% O₂ and 5% CO₂). This incubation included the addition of bicuculline (10 μM), roscovitine (20 μM), or a combination of roscovitine (20 μM) and nifedipine (5 μM) for 20 min. Following incubation, the slices were homogenized in a lysis buffer comprising 26 mM Tris–HCl, 1% Triton X-100, 1.3 M glycerol, 130 mM NaCl, and a protein phosphatase inhibitor cocktail (Complete™ mini-tab, Roche). To evaluate Cdk5, protein-level striatal tissue was processed as described above, but without the reagents tested in electrophysiological experiments.

Following collection, the samples were subjected to centrifugation for 5 min at 4,500 rpm, and the resultant supernatant was collected and stored at −70°C for subsequent analysis. Protein quantification was conducted using the Bradford method.

In the subsequent blotting procedure, 30 μg of protein was loaded onto a 10% polyacrylamide electrophoresis gel. The proteins were then transferred to PVDF transfer membranes, blocked with 10% skinny non-fat dry milk dissolved in TBS (pH 7.6) for 1 h and incubated with primary antibodies against phospho-DARPP-32^Trh-34 (GeneTex, Cat. No. GTX55025, 1:2,000), phospho-DARPP-32^Trh-75 (Cell Signaling Technology, Cat. No. 2301, 1:1,000), DARPP-32 (GeneTex, Cat. No. GTX82715, 1:1,000), Cdk5 (sc-249, Santa Cruz, 1:200–1:1,000), and β-actin (GeneTex, Cat. No. GTX629630, 1:60000–1:120000) as a load control for 12 h at 4°C. Secondary antibodies (Anti-rabbit IgG, HRP, Cell Signaling Technology, Cat. No. 7074, 1:2000–1:10000) were then applied for 2 h at room temperature. All antibodies were diluted in TBS. The protein level of Cdk5 was detected using chemiluminescence and scanned with an AC-Digit Western Blot scanner. The intensity of immune-reactive bands was analyzed using Image Studio Digits 4.0 software, while DARPP-32 was identified utilizing the chemiluminescence method with Immobilon™ Western Chemiluminescent HRP Substrate from Millipore (Billerica, MA, United States). Subsequently, images were captured using the Gel Documentation System from BioSens SC 645, a product of Shanghai Bio-Tech Co., Ltd. The relative protein level was determined by comparing the optical density of Cdk5 to the expression of β-actin on the same membrane. This ratio was then normalized to the control expression. Additionally, for DARPP-32, the ratio of phospho-DARPP-32/total DARPP-32 level was assessed and similarly normalized to control expression.

All reagents were purchased from Sigma-Aldrich Merck (Darmstadt, GER) unless otherwise stated.

A one-way ANOVA with a post-hoc test or Student’s t-test (paired or unpaired, depending on the experiment) was used for data comparisons for WB analyses and between pre-HFS and post-HFS in recorded neurons. Non-normal distribution data were analyzed using a non-parametric test. Statistically significant differences were considered at p < 0.05.

First, we analyzed whether 3-NP produced changes in Cdk5 staining and protein-level expression. The results showed that Cdk5 immunolocalization significantly increased the number of Cdk5-positive striatal cells from 3-NP-treated mice compared to that in control mice (Figures 1A,B). However, the protein level of Cdk5 in the striatum exhibited a non-significant increase compared to the control striatal tissue (Figure 1C).

Synaptic plasticity impairments have been reported in different models of HD and striatal degeneration. To test whether 3-NP administration in vivo-induced synaptic plasticity alterations in our striatal degeneration model, patch-clamp recordings from corticostriatal connections were carried out. In brain slices from seven control mice, HFS produced a significant LTD in the EPSC amplitude of all 7 medium spiny neurons (MSNs) after HFS training (Figure 2A, a1,a2). The analysis of the paired-pulse ratio (PPR) before and after HFS in control tissue, was significantly different in the PPR ratio suggestive of presynaptic mechanisms underlying LTD generation (Figure 2A, a3).

The analysis of plasticity induced in the neurons of seven 3-NP-treated mice exhibited two types of plasticity; 7 of 10 cells exhibited LTD after HFS (Figure 2B, b1). The PPR after HFS was significantly different indicating presynaptic mechanisms (Figure 2B, b3). Furthermore, the LTD observed in slices from 3-NP-treated mice was significantly smaller than that observed in control mice. Three neurons out of 10 exhibited LTP after HFS (Figure 2C), with a PPR that did not exhibit statistical differences after HFS, suggesting that postsynaptic mechanisms are involved in this LTP generation (Figure 2C, c3).

In our previous study, we established a correlation between Cdk5 activity and synaptic plasticity in the striatum. To determine whether Cdk5 plays a role in corticostriatal synaptic plasticity in slices from mice treated with 3-NP, we selectively inhibited its activity using the Cdk5 inhibitor roscovitine (20 μM). Under the influence of roscovitine, corticostriatal synapses of neurons from 3-NP-treated mice exhibited LTD and LTP (Figure 3A, a1,a2). Seven out of 11 cells from 3-NP-treated mice exhibited LTD in the presence of roscovitine after HFS (Figure 3A, a1,a2). This LTD is discrete in comparison with that observed in neurons from control mice or 3-NP-treated mice without roscovitine. The PPR analysis demonstrated that LTD induced under these conditions was not caused by presynaptic mechanisms (Figure 3A, a3). Few cells from the 3-NP-treated mice generated LTP (4/11) in the presence of roscovitine after HFS. The PPR from these experiments was statistically different after HFS (Figure 3B, b3).

In our earlier findings, we documented that Cdk5 inhibition facilitates the induction of LTP through the D1 receptor signaling pathway in striatal slices from control mice (Miranda-Barrientos et al., 2014). To investigate whether activation of D1 receptors was implicated in the changes of EPSC amplitudes observed in the presence of roscovitine in neurons from 3-NP-treated animals, we conducted recordings on these neurons in the presence of a D1 receptor antagonist SCH23390 (1 μM) and roscovitine (Figure 4A).

In one set of cells (3 of 9), the D1 inhibitor SCH23390 impeded the generation of any type of plasticity in the 3-NP slices in the presence of roscovitine (Figure 4A, a1,a2). However, the D1 antagonist did not block LTP induction in six of nine cells from 3-NP-treated animals (Figure 4B, b1,b3). The PPR was not significantly different in any of the two sets of cells, where no plasticity was developed (Figure 4A, a3) or where LTP was not prevented by SCH23390 suggestive of postsynaptic changes (Figure 4B, b3).

In particular, the administration of the PKA inhibitor H89 (5 μM) to the recording bath blocked the induction of any type of plasticity in the presence of roscovitine in eight of eight MSN cells recorded from the 3-NP-treated animal tissue (Figure 4C, c1,c2). No changes in the PPR were observed. The block of synaptic plasticity in the presence of H89 + roscovitine in 3-NP-treated neurons suggests that PKA signaling is involved in the induction of LTP generated when the Cdk5 is inhibited by roscovitine (Figure 4C).

Cdk5 phosphorylates DARPP-32 phosphoprotein on its residue Thr-75 (Bibb et al., 1999). Given that the number of positive neurons for Cdk5 increased in tissue slices from 3-NP-treated mice, we wanted to determine whether the changes observed in plasticity were due to changes in the phosphorylation state of DARPP-32. Immunoblot assays were performed to measure total-DARPP-32, as well as the Cdk5 phosphorylation site DARPP-32^Thr75 and the PKA phosphorylation site DARPP-32^Thr34 (Hemmings et al., 1984; Snyder et al., 1992) in slices from control and 3-NP-treated mice. These experiments were conducted in the presence of bicuculline, roscovitine, and roscovitine + nifedipine to block other sources of calcium that could activate Cdk5 due to membrane depolarization. Data were expressed as p-DARPP-32/total DARPP-32 and normalized to control protein level.

The phosphorylation state of DARPP-32 ^Thr34 in the presence of roscovitine was not significantly affected (Figures 5A,B); however, the level of p-DARPP-32^Thr75 was significantly reduced in the presence of roscovitine and in the presence of roscovitine + nifedipine in both control and 3-NP-treated mice slices (Figures 5A,C).

In the analysis of the relative expression of p-DARPP-32^Thr34 and p-DARPP-32^Thr75 under basal conditions, we observed a non-significant decrease in p-DARPP-32^Thr34 and there was a non-significant increase in p-DARPP-32^Thr75 in tissue from the 3-NP-treated group compared to the control group (Figure 5D).

In vivo mice treatment with 3-NP induced an increase in Cdk5-positive staining in cells within the striatum. This immunostaining increase was specifically observed in the soma of striatal cells, compared to control mice. Cdk5 is known to localize at various cellular compartments, including the axon, membrane-associated regions with activator proteins, and the cytoplasm, particularly during heightened activity. This variability in localization could explain the increased staining observed in the soma of neurons from 3-NP-treated mice, without changing significantly protein levels. Previously, it has been reported that neuronal activity modifies the activity of Cdk5 and its cellular localization (Tomizawa et al., 2000). Previous experiments involving the incubation of slices with 3-NP (100 μM) for 3 h have also shown an increase in Cdk5 (Napolitano et al., 2004).

The overexpression and/or overactivation of Cdk5 has been linked to neurodegeneration in HD (Crespo-Biel et al., 2007; Paoletti et al., 2008). The increased Cdk5 level results in the hyperphosphorylation of multiple substrates, regulating neurotransmission, cell trafficking, and cell survival (Cheung and Ip, 2012; Sheng et al., 2015; Liu et al., 2016). However, a Cdk5 increase can also have a homeostatic function on neuronal activity (Sheng et al., 2015; Pao and Tsai, 2021).

In cells from the control tissue, only LTD was observed after HFS, while neurons from 3-NP treatment modified striatal plasticity. The cells from 3-NP-treated mice exhibited both LTD and LTP, with the magnitude of LTD being smaller than that observed in the slices from the control mice. The 3-NP treatment also favored the development of LTP in a few cells. Earlier reports have shown that both in vitro and in vivo administration of 3-NP in rats reduces LTD and induces LTP in field recordings, which supports our findings (Calabresi et al., 2001; Dalbem et al., 2005; Picconi et al., 2006).

The presence of roscovitine resulted in a discrete LTD or LTP in corticostriatal synapses. These effects on striatal plasticity suggest that Cdk5 plays a role in modulating synaptic plasticity favoring LTD. The increase in Cdk5-positive cell staining may indicate a contribution of Cdk5, in part, to the changes in plasticity observed after 3-NP treatment. This is supported by the effects of the Cdk5 inhibitor roscovitine administration, which reduces the strength of LTD in most of the cells and facilitates an LTP with greater amplitude than that observed in 3-NP tissue slices alone in some cells.

The D1 receptor signaling pathway was implicated in the plasticity induced in only a subset of the recorded cells from 3-NP-treated mice. The addition of the D1 antagonist SCH23390, along with roscovitine, inhibited LTD expression in these cells but not in others that continued exhibiting LTP despite the presence of SCH23390 (Figure 4). Nonetheless, the inhibition of PKA with H89 in the presence of roscovitine blocked the induction of plasticity in all cells, indicating the involvement of PKA signaling in the induction of striatal plasticity generated in neurons from 3-NP treatment. H89 blocked plasticity in all cells because, in the striatum, PKA can be activated by the stimulation of D1 or A2a receptors in the direct and indirect pathways, respectively (Nishi et al., 2011; Ma et al., 2022). While the present study did not specifically evaluate A2a receptors, cells lacking D1/PKA signaling activation may engage in PKA signaling through the activation of A2a receptors.

In the striatum, the expression of LTD and LTP at corticostriatal synapses is known to be finely regulated by dopamine levels. Low levels of dopamine tend to facilitate LTD induction, while elevated dopamine levels favor LTP (Lovinger, 2010). Although both forms of synaptic plasticity can be induced (Lovinger, 2010) and are involved in the control of learned motor programs (Di Filippo et al., 2007), certain forms of LTP are linked to excitotoxic processes or pathological LTP (Centonze et al., 2001).

Cdk5 inhibitors have been shown to enhance dopaminergic transmission (Chergui et al., 2004), because the activations of Cdk5 reduces PKA phosphorylation of DARPP-32 in Thr-34, which is associated with striatal LTP generation (Miranda-Barrientos et al., 2014; Hernandez et al., 2016). Phosphorylation of DARPP-32 in its Thr75 residue by Cdk5 inhibits PKA signaling (Bibb et al., 1999), suggesting that under normal conditions, Cdk5 tends to promote the expression of LTD by preventing PKA activation. Our previous study demonstrated that Cdk5 inhibition facilitates LTP development (Miranda-Barrientos et al., 2014). Consequently, an increase in Cdk5 following 3-NP treatment may trigger a homeostatic mechanism aimed at preventing pathological LTP induction or changes in neuronal activity. It has been noted that Cdk5 plays a crucial role in synaptic scaling, enabling neurons to adapt to changes in activity levels (Seeburg et al., 2008). This potential regulatory response is crucial, especially considering the associated excitotoxicity stemming from altered striatal dopamine levels and a depolarized membrane potential (Zeron et al., 2002). Our 3-NP model imitates changes in the early stages of striatal degeneration; therefore, it is possible that Cdk5 activation prevents DARPP-32^Thr34 phosphorylation by PKA. To test this idea, we evaluated the phosphorylation state of DARPP-32.

The phosphorylation state of DARPP-32^Thr75 exhibited a notable reduction in the presence of both roscovitine and in the presence of roscovitine + nifedipine, with no concurrent alteration in DARPP-32^Thr34. The depolarization of MSN cell membranes by 3-NP may induce Ca²⁺ influx, potentially activating Cdk5. To discern the role of Ca²⁺ entry through L-type channels, nifedipine was employed to eliminate Ca²⁺ influx and its potential contribution to DARPP-32^Thr75 phosphorylation (Loya-López et al., 2020), but no further inhibition of phosphorylation was observed. Nonetheless, reduced phosphorylation of DARPP-32^Thr75 with Cdk5 inhibition may leave the PKA/DARPP-32^Thr34 pathway activated, thereby mitigating LTD and favoring LTP.

Concomitant with the lack of a significant increase in the protein level of Cdk5, protein expression analysis revealed a non-significant increase in the basal phosphorylation of DARPP-32^Thr75, a site phosphorylated by Cdk5 (Bibb et al., 1999). Additionally, the PKA phosphorylation site, DARPP-32^Thr34, exhibited a non-significant decrease.

In a previous study, acute administration of a higher dose of 3-NP led to DARPP-32^Thr-75 upregulation and DARPP-32^Thr34 downregulation (Napolitano et al., 2004). In our case, the administration of 3-NP in low doses in vivo for 5 days, mirroring the initial stages of neurodegeneration, did not significantly alter the phosphorylation status of DARPP-32^Thr34 and DARPP-32^Thr75. Nevertheless, the observed tendency to increase DARPP-32^Thr75 phosphorylation and decrease DARPP-32^Thr34 phosphorylation suggests a compensatory mechanism to prevent corticostriatal LTP (Miranda-Barrientos et al., 2014). This mechanism may aim to mitigate the risk of excitotoxicity arising from NMDA receptor activation by mitochondrial dysfunction (Lu et al., 2015) and confers a neuroprotective role for Cdk5 in early neurodegeneration (Pao and Tsai, 2021).