Experiments were performed on tissue obtained from 36 male Wistar rats (weighing 30–60 g) aged 4–6 weeks at the start of the protocol, obtained from the Kerman University of Medical Sciences, and housed under controlled photoperiod (lights on: 07:00–19:00 h), at 22 ± 1 °C with food and water available. All protocols and procedures using animals were approved by the Kerman University of Medical Sciences Ethic Committee (IR.KMU.REC.1399.254). Male rats (n = 6 cells/group; one cell from each rat) were divided randomly into six groups: control (Co), cannabinoid agonist (WIN; 7.5 nmol) [19], cannabinoid antagonist (AM; 20 nmol) [19], 3 acetyl pyridine (3-AP; 1 mM) [20], 3-AP + WIN, and 3-AP + AM.

Rats were euthanized by cervical dislocation, and the posterior skull was removed allowing extraction of the cerebellum in the coronal plane. The brain was sliced into sections of a 250 μm thickness using a Vibroslicer (Campden Instrument, NVSLM1, Sarasota, FL) in artificial cerebrospinal fluid (aCSF) containing (in mM) 124 NaCl, 25 NaHCO3, 10 D-glucose, 4.4 KCl, 2 MgCl2, 1.25 NaH2PO4, and 2 CaCl2, which was bubbled with 95% O2–5% CO2 (pH 7.4 ± 0.05 and osmolarity was adjusted to 300 ± 10 mOsm). The slices were immediately incubated in aCSF at 35 ± 2 °C for 30 min before being recorded within 6 h at room temperature (22 °C).

A Multiclamp 700B amplifier from Axon Instruments was used to conduct whole-cell patch clamp recordings from Purkinje cells in current clamp and voltage clamp modes. Purkinje cells were identified using the established criteria listed below. Signals were digitized using a 1440 A/D converter (Axon Instruments). Electrophysiological data were sampled at 10 kHz and filtered at 20 kHz [21]. Patch pipettes had a resistance of 5–9 MΩ when filled with an internal solution comprising (in mM) 140 potassium gluconate, 5 KCl, 10 HEPES, 2 MgCl2, 0.2 EGTA, 2 Na2ATP, and 0.4 Na2GTP. The internal solution's pH and osmolarity were adjusted to 7.3 (by KOH) and 290 mOsm, respectively. The perfusion rate of the recorded slides was 1.9–2 ml/min. Following the formation of a giga seal on the cell membrane, a brief suction was used to rupture the membrane, allowing the cell to be clamped to a holding voltage of -60 mV. The test seal function was continuously used throughout the recording to make sure the seal was stable. The number of action potentials generated after negative current injection when the cell rebounded to the resting membrane potential, and the spike latency of the first action potential were measured. In spontaneously firing neurons, action potential parameters including action potential (AP) half-width, AP frequency, AP amplitude, AHP amplitude, voltage threshold (the membrane potential from the base and start of the action potential), interspike interval (ISI), and coefficient of variation (CV) (regularity of action potential and dispersion of a probability/frequency) were measured [22].

The sag voltage in response to hyperpolarizing current pulses (amplitude of -100 pA to -500 pA) was calculated. The peak voltage deviation was divided by the amplitude of the steady-state voltage deviation using the following formula:

Sag voltage = Vpeak − Vsteady state.

Furthermore, the sag was calculated as [(peak response—steady state response) / peak response] 100. To examine the effect of a CB1R agonist/antagonist on Ih in Purkinje cells, I-V activation curves were obtained in voltage clamp mode using 520 ms of hyperpolarizing steps (50 to 140 mV in increments of 10 mV).

Drugs were applied in aCSF from a separate reservoir (also gassed with carbogen) to the recording chamber. 3-AP (Sigma, USA) was dissolved in deionized water as a × 100 stock solution containing 0.4% ascorbic acid. Purkinje cells were exposed to aCSF (control) or 1 mM 3-AP (for at least 20 min). WIN (7.5 nmol) and AM (20 nmol) were added to the bath [23].

Data are presented as mean ± SEM using Graph Pad Prism 9 (Graph Pad Software, USA). Normally distributed data were compared by a two-way ANOVA, and the Tukey post hoc analysis was used for multiple comparisons between multiple groups. Differences were considered statistically significant when P < 0.05.

In CB1R agonist/antagonist and 3-AP-exposed slices, a total of 30 Purkinje cells were recorded. The passive and active properties of 6 cells were examined for each group. Our results indicated that cells in the 3-AP (P < 0.001), 3-AP + WIN (P < 0.01), and 3-AP + AM (P < 0.01) treated groups exhibited a significantly lower resting membrane potential (RMP) (F (2, 24) = 1.446, P = 0.2554) compared to cells in the Co, WIN, and AM groups, respectively (Fig. 1A). We measured R_in as an index of how the cells would passively respond to current inputs. There were no statistical differences in R_in (F (2, 23) = 0.9682, P = 0.3947) between any of the Purkinje cell treatment groups, indicating that exposure to 3-AP whether in the presence or absence of cannabinoid agonists or antagonists, had no effect on R_in (Fig. 1B).Fig. 1

Changes in electrophysiological properties of Purkinje cells in vermis slices treated with CB1R agonist/antagonist and 3-AP. A 3-AP alone and in combination with CB1R agonist/antagonist resulted in a lower RMP compared to control, WIN, and AM respectively. B No significant differences were observed in Rin. n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM. **(P < 0.01) and ***(P < 0.001) represent a significant difference with the counterpart group

Purkinje cells in 3-AP (P < 0.001), 3-AP + WIN (P < 0.05) and 3-AP + AM (P < 0.001) exposed slices demonstrated a higher spontaneous firing frequency (F (2, 24) = 12.41, P = 0.0002) when compared to that in the Co, WIN, and AM groups, respectively; whereas, Purkinje cells in the 3-AP + AM (P < 0.01) group demonstrated a significantly lower firing frequency when compared to that in 3-AP-treated cells (Fig. 2A). Besides firing rate, another important parameter of the Purkinje cell’s function is the interspike interval (ISI). We examined this parameter by measuring the spike intervals in Purkinje cells that were spiking spontaneously. This analysis indicated that cells exposed to 3-AP (P < 0.001) demonstrated a lower ISI (F (2, 24) = 4.817, P = 0.0174) compared to the Co and AM (P < 0.05) treated groups (Fig. 2B). We also found that Purkinje cells in the AM and 3-AP + AM (P < 0.05) groups had significantly broader AP half-widths (P < 0.05) compared to those seen in the 3-AP group (Fig. 2C). The amplitude of the action potential was significantly larger in Purkinje cells in the 3-AP group compared to cells in the Co group (P < 0.05). However, our data showed a lower amplitude of the action potential in cells from the WIN (P < 0.01), AM (P < 0.001), and 3-AP + AM (P < 0.001) groups compared to the amplitude in Purkinje cells of the 3-AP group (Fig. 2D). Purkinje cells in 3-AP (P < 0.001) exposed cell groups showed a lower CV (F (2, 24) = 12.67, P = 0.0002) compared to cells in the Co. In addition, 3-AP + AM (P < 0.001) cells showed lower CV compared to that seen in cells in the AM group (Fig. 2E). We examined the voltage threshold (F (2, 24) = 1.112, P = 0.3454) and kinetics of action potentials to determine if 3-AP affects these parameters. Purkinje cells from 3-AP-exposed (P < 0.01) cells demonstrated a significantly lower voltage threshold compared to cells in the Co group (Fig. 2F). The prominent AHP seen in Purkinje cells following an action potential reflects conductance activated during the spike. Amplitudes of the AHP (F (2, 24) = 16.49, P < 0.0001) in the Purkinje cells of the 3-AP (P < 0.01) and 3-AP + WIN groups (P < 0.001) were larger in amplitude compared to amplitudes seen in Co and WIN cells, respectively. Purkinje cells in the AM (P < 0.05) and 3-AP + AM (P < 0.05) groups displayed a significantly lower AHP compared to that seen in the 3-AP group (Fig. 2G). Furthermore, traces of spontaneous activity can be seen in Fig. 3, where two action potentials from each group have been selected and then superimposed. Compared to the Co and 3-AP + AM groups, the AP amplitude, AHP, and AP frequency were significantly higher, and the half width, and threshold were lower in the 3-AP groups (Fig. 3).Fig. 2

The effect of CB1R agonist/antagonist on the spontaneous firing properties of Purkinje cells in control and 3-AP treated slices. A Electrophysiological properties shown are action potential frequency, B interspike interval, C action potential half-width, D action potential amplitude, E coefficient variant, F action potential threshold, and G AHP amplitude. n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM. *(P < 0.05), **(P < 0.01), and ***(P < 0.001) represent a significant difference with the counterpart group. #(P < 0.05), ## (P < 0.01), and ###(P < 0.001) represent a significant difference versus 3-AP

To determine if 3-AP affects the excitability of cerebellar Purkinje cells, we first measured the rheobase. Rheobase is operationally defined as the minimal amount of electrical current required to trigger an action potential when current is injected into a cell. In current clamp mode, Purkinje cells exposed to 3-AP (P < 0.05) and 3-AP + WIN (P < 0.05) showed a significantly lower rheobase (F (2, 24) = 2.056, P = 0.1499), indicating a heightened excitability compared to cells in the Co and WIN groups, respectively (Fig. 4A). However, while the 3-AP and 3-AP + WIN groups required less current to stimulate an action potential, there was not a significant difference noted in the threshold voltage noted in the RAMP protocol, which indicates the current at which an action potential was triggered (F (2, 24) = 1.615, P = 0.2198) (Fig. 4B).Fig. 4

3-AP causes a reduction in rheobase in cerebellar Purkinje cells in current clamp recordings. A Rheobase (current clamp), B RAMP threshold. n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM. *(P < 0.05) represent a significant difference with the counterpart group

We next examined the response of Purkinje cells to the effects of CB1R agonist/antagonist on spontaneous excitatory synaptic activity following exposure to 3-AP. Purkinje cells in the 3-AP (P < 0.05) and 3-AP + WIN (P < 0.05) groups showed smaller intervals of sEPSCs (F (2, 24) = 1.926, P = 0.1676) when compared to the sEPSCs intervals seen in the Co and WIN groups, respectively (Fig. 5A). Consistent with a heightened increase in excitability, the amplitude of sEPSCs (F (2, 24) = 1.717, P = 0.2009) in the Purkinje cells of the 3-AP group (P < 0.01) was increased. In addition, the combination of 3-AP and WIN (P < 0.01) amplified the action of 3-AP in enhancing sEPSP amplitudes. In Purkinje cells in the WIN (P < 0.001) and AM (P < 0.01) groups, there was a significantly lower amplitude of sEPSCs when compared to amplitudes seen in 3-AP cells. Administration of AM failed to significantly alter the effects of 3-AP on sEPSP amplitudes (Fig. 5B).Fig. 5

EPSPCs of Purkinje cells in slices exposed to CB1R agonist/antagonist in control and 3-AP treated slices. A EPSPCs interspike interval, B EPSCs amplitude is shown. n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM. *(P < 0.05) and **(P < 0.01) represent significant differences with the same group with Vehicle (DMSO). #(P < 0.05) represents a significant difference versus 3-AP

Rebound is defined as membrane depolarization occurring at the offset of a hyperpolarizing stimulus and is one of several intrinsic properties that may promote rhythmic electrical activity. Purkinje cells from 3-AP (P < 0.001) and 3-AP + WIN (P < 0.001) exposed brain slices displayed a significantly higher number of rebound action potentials (F (2, 24) = 8.932, P = 0.0013) appearing following presence of a hyperpolarizing conditioning pulse (Fig. 6A) and a reduction in the first spike latency (F (2, 24) = 10.71, P = 0.0005), (Fig. 7A) at -100 pA (P < 0.01) compared to averages for these same parameters in Purkinje cells from Co and WIN groups, respectively. In Purkinje cells in the 3-AP treatment group, which were treated with AM (P < 0.05), conversely, a lower number of rebound action potentials was noted (Fig. 6A, B) as well as an enhanced first spike latency (Fig. 7A, B) at both -100 pA (P < 0.01) and -500 pA (P < 0.05) compared to 3-AP Purkinje cells.Fig. 6

The effect of CB1R agonist/antagonist on the rebound of action potential at A -100pA and B -500 pA. n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM. ***(P < 0.001) represents the significant difference with the same group with Vehicle (DMSO). #(P < 0.05) represents the significant difference versus 3-AP

The rebound of action potentials was significantly higher in Purkinje cells from the 3-AP and 3-AP + WIN groups when compared to that seen in the Co and WIN groups at both -100 pA (P < 0.001) and -200 pA (P < 0.05). Furthermore, there was a significantly lower rebound in Purkinje cells in the 3-AP + AM group at -200 pA and -500 pA compared to 3-AP cells. Therefore, exposure to AM was capable of reducing the heightened excitability of Purkinje cells associated with 3-AP treatment.

Exposing Purkinje cells to 3-AP led to an increase in rebound spikes, which is a sign of heightened excitability, and treating them with AM decreased the rebound spikes. Thus, our data shows that evoked and spontaneous activities can be reversed by AM.

Furthermore, the prolonged, time-dependent changes evident in the spike rate/pulse of action potential after the increase in the current step from 0.1 nA to 0.5 nA indicate that an adaptation occurs in the action potential in all groups. Particularly, we did note a significant difference in the adaptation seen in the 3-AP groups (Fig. 8).Fig. 8

Population data representative of the effects on spike rate/pulse (0.1 nA–0.5 nA). n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM. *(P < 0.05) and ***(P < 0.001) represent the significant differences with the same group with Vehicle (DMSO). ###(P < 0.001) represents the significant difference versus 3-AP

The sag voltage was studied to determine the effects of CB1R modulation on Ih currents (hyperpolarization-activated currents) in Co and 3-AP treated slices. Sag is considered an indicator of the kinetics and amplitude of Ih relative to the membrane time constant. After the application of hyperpolarizing current pulses (-100 pA (F (2, 24) = 0.3777, P = 0.6894) and -500 pA (F (2, 24) = 1.931, P = 0.1668)) the %sag voltage showed no significant difference in any group (Fig. 9). This finding suggests that the effects of cannabinoids on ataxia might not involve actions at Ih.Fig. 9

The percentage of the sag potential was determined at A -100pA and B -500pA. n = 6 cells/group; one cell from each rat. The data are expressed as mean ± SEM