Impact of NMDA Receptor Overexpression on Cerebellar Purkinje Cell Activity and Motor Learning

Abstract

In many brain regions involved in learning NMDA receptors (NMDARs) act as coincidence detectors of pre- and postsynaptic activity, mediating Hebbian plasticity. Intriguingly, the parallel fiber (PF) to Purkinje cell (PC) input in the cerebellar cortex, which is critical for procedural learning, shows virtually no postsynaptic NMDARs. Why is this? Here, we address this question by generating and testing independent transgenic lines that overexpress NMDAR containing the type 2B subunit (NR2B) specifically in PCs. PCs of the mice that show larger NMDA-mediated currents than controls at their PF input suffer from a blockage of long-term potentiation (LTP) at their PF-PC synapses, while long-term depression (LTD) and baseline transmission are unaffected. Moreover, introducing NMDA-mediated currents affects cerebellar learning in that phase-reversal of the vestibulo-ocular reflex (VOR) is impaired. Our results suggest that under physiological circumstances PC spines lack NMDARs postsynaptically at their PF input so as to allow LTP to contribute to motor learning.

Keywords: NMDA; Purkinkje cell; cerebellum; compensatory eye movements; motor learning; synaptic plasticity.

Figure 1.

Generation of NR2B transgenic mice. A, Schematic representation of the expression of endogenous NMDARs at CF-PC and PF-synapses over the wild-type mouse lifetime. All experiments performed in the following figures are performed at 6 ± 1 weeks of age (for details, see Materials and Methods). Middle, Scheme depicting the expression of NMDAR (dots) in a wild-type mouse at the synapses formed on the PC dendritic tree by CFs. Synapses formed by PFs (the axons of granule cells, GC) do not have NMDARs. Right, Same schematic representation of synapses onto PCs in the L7-NR2B + Tg transgenic mouse. Note that NMDARs are present also at the PF-PC synapse. Inset, Details of the vector used to generate the two independent lines used in the study. B, Western blotting gels containing homogenates of adult cerebella, forebrains, and hippocampi of two transgenic mice (Tg1+ and Tg2+) and their control littermates (Tg1- and Tg2-). The left blot was processed with an anti-NR2B antibody, which visualizes a band at 190 kDa, the right one with an anti-NR1 antibody (120 kDa). Actin was used as loading control; note that the ratio of NR labeling to actin labeling should be taken into account for assessing the amounts of protein present. C, Confocal immunofluorescent images of NR2B-immunoreactivityin dorsal hippocampus (Hip) and cerebellar cortex (Cb) in control (left), Tg1 (middle), and Tg2 (right) mice. Note in control mice, the low level of NR2B labeling (blue) in the cerebellum as compared to hippocampal CA1 and dentate gyrus (DG). In addition, note moderate increased labeling in cerebellar molecular layer (ml) of transgenic mice. D, Low- and high-magnification images of calbindin immunoreactivity in sagittal cerebellar sections illustrating the normal appearance of cerebellar gross morphology and PCs of adult transgenic (Tg1 and Tg2) mice. E, top, High magnification of individual Golgi-stained PCs (black) of control (left) as well as Tg1 (middle) and Tg2 (right) NR2B mice, counterstained with thionin (blue). Bottom, Sholl analysis of the dendritic arborization of PCs (left) and length of their primary dendrites (right) for control (black, n = 30, N = 4), Tg1 (orange, n = 30, N = 3), and Tg2 (red, n = 30, N = 4) NR2B mice. Empty circles indicate individual data points, full circles indicate mean ± SEM.

Figure 2.

Functional NMDARs are present at six weeks of age and do not compromise PCs’ basic electrophysiological properties. A, left, Schematic representation of the recording configuration. Middle, Example traces of CF currents recorded in the presence of the AMPA antagonist NBQX (green) and subsequently of blockers of both AMPA and NMDARs (D-AP5, black) in PCs of both transgenic lines and control littermates. Bold lines are average values; shading indicates individual cell variability. Right, Quantification of NMDA current at the CF-PC synapse in control (Ctrl, black, n = 5) and transgenic animals (Tg1, orange, n = 7; Tg2, red, n = 6). B, Similar to A, with additional example traces of baseline PF-evoked EPSCs before the addition of glutamatergic receptors blockers (blue), and normalized PF-PC NMDA current quantification (Ctrl, n = 9; Tg1, n = 9; Tg2, n = 7). Note that NMDA-mediated currents are only significantly different from controls in Tg2. C, Average firing frequency elicited by somatic current injections from -65 mV in PCs of transgenic (Tg1, orange, n = 5; Tg2, red, n = 12) and control (Ctrl, black, n = 15) mice. The inset illustrates the recording configuration. D, Average amplitude of the EPSCs at the PF-PC synapse to stimuli of increasing intensity for transgenic (Tg1, orange, n = 5; Tg2, red, n = 8) and control (Ctrl, black, n = 18) mice. The inset illustrates the recording configuration. Empty circles represent individual data points, full circles are mean ± SEM; *p < 0.05, **p < 0.01, and absolute p values are indicated in the main text.

Figure 3.

PF-PC LTP is selectively affected in transgenic mice. A, LTP was induced by PF stimulation at 1 Hz for 5 min in six-week-old transgenic (Tg2, red, n = 8) and control (Ctrl, black, n = 7) mice. The normalized paired-pulse ratio (50-ms interstimulus interval) of the recordings of the same cells is plotted below. B, LTD was induced as described in A, but with concomitant CF activation (Ctrl, n = 8; Tg2, n = 8). C, Similar to A, but with the NMDAR blocker D-AP5 present in the extracellular solution (Ctrl, n = 7; Tg2, n = 8). D, Induction of IP by 5 min of PF stimulation at 1 Hz did not result in significant differences between transgenic and control mice (both N = 6). The normalized membrane potentials of the same cells are represented below. The scheme at the left of each panel depicts the respective recording configuration, while the middle example traces of Ctrl (black) and Tg2 (red) EPSCs (A–C) or action potentials (D) recorded before (t = 0 min) and after (t = 20 min) the tetanic stimuli show the plastic changes. Values are mean ± SEM; *p < 0.05; **p < 0.01; absolute p values are indicated in the main text.

Figure 4.

Motor performance is normal, but motor learning is impaired in transgenic mice. A, Distance traveled and average speed in the open field for transgenic (Tg2, red, n = 8) and control (Ctrl, black, n = 7) mice. B–D, Baseline compensatory eye movements (examples traces for 0.4 Hz, middle left) quantified by gain (middle right) and phase (right) for Tg2 mice (red, n = 9) and control (Ctrl, black, n = 10) mice: (B) OKR; (C) VOR (in the dark); and (D) VVOR (in the light), schematized on the left of each respective panel. E, left, Representation of gain-decrease training paradigm (day 1: 5 × 10 min sinusoidal, in-phase drum and table rotation at 0.6 Hz, both with an amplitude of 5°; day 2: VOR gain measurement at 0.6 Hz). Middle, Example traces of before (time point, t = 0, indicated by a) and after (t = 50 min, b) adaptation. Right, Normalized gain for VOR recorded with 10-min intervals during 50-min training session for six-week-old Tg2 mice (red, n = 8) and control (Ctrl, black, n = 11) mice on day 1 and a single measurement at day 2. F, Similar to E; following the gain-decrease protocol, for four consecutive days the six-week-old transgenic and control mice were subjected to the phase reversal protocol (5 × 10 min sinusoidal in-phase drum and table rotation at 0.6 Hz, but with drum amplitudes of 7.5° on day 2 and 10° on days 3–5, while the table amplitude was 5°). VOR responses (middle: example traces, a same as E, c: t = 50 min on day 4) are depicted as gain of eye movement multiplied by the cosine of its phase, gain*cos(phase). Negative values here indicate a phase larger than 90° and the (theoretical) goal of the training is a value of -1. Empty circles represent individual data points, full circles are mean ± SEM; p values are indicated in the main text; asterisks indicate significant difference.