Purkinje cell microzones mediate distinct kinematics of a single movement

Abstract

The classification of neuronal subpopulations has significantly advanced, yet its relevance for behavior remains unclear. The highly organized flocculus of the cerebellum, known to fine-tune multi-axial eye movements, is an ideal substrate for the study of potential functions of neuronal subpopulations. Here, we demonstrate that its recently identified subpopulations of 9+ and 9- Purkinje cells exhibit an intermediate Aldolase C expression and electrophysiological profile, providing evidence for a graded continuum of intrinsic properties among PC subpopulations. By identifying and utilizing two Cre-lines that genetically target these floccular domains, we show with high spatial specificity that these subpopulations of Purkinje cells participate in separate micromodules with topographically organized connections. Finally, optogenetic excitation of the respective subpopulations results in movements around the same axis in space, yet with distinct kinematic profiles. These results indicate that Purkinje cell subpopulations integrate in discrete circuits and mediate particular parameters of single movements.

Fig. 1. KCTD12 identifies physiologically distinct subpopulations of 9+ and 9- PCs in the flocculus.

A Sagittal sections of the flocculus (white dashed line) immunolabeled for KCTD12 (purple) and AldoC (red). White arrows indicate the KCTD12 boundary identified with immunostaining. Scale bar = 200 µm. B Schematic representation of differential protein expression among floccular Purkinje cell (PC) subpopulations. C Top: Immunolabelling of KCTD12 and AldoC in lobule IX (coronal) and the flocculus (sagittal) labeling the cerebellar modules. Bottom: Relative signal intensity of KCTD12 (purple curve) and AldoC (red curve) expression in the PC layer for the top region. Scale bar = 50 µm. D Example of biotin-filled PCs in different regions of the flocculus. Scale bar left panels = 200 µm, right panels = 100 µm. E Example cell-attached recordings from 9+ (top) and 9− (bottom) PCs. Scale bar = 100 ms. F Summary data of firing rate and CV2 for PCs in Lobule III, Lobule X, 9- flocculus and 9+ flocculus. n = 14, 35, 20, and 14 cells, respectively. G Example traces from current injections steps. Black trace indicates 0 pA and gray trace +100 pA current injections around holding current. Scale bar = 100 ms. H Summary data of current injection steps for PCs in Lobule III, Lobule X, 9- flocculus and 9+ flocculus. n = 20, 51, 32, and 36 cells, respectively. Left graph displays the average number of spikes in response to a 500 ms current pulse ranging from +100 to +900 pA. Right graph displays the average time to the first spike at each of the current injection steps. Fl flocculus, C caudal, R rostral, D dorsal, V ventral. ***: p < 0.001, **: p < 0.01; *: p < 0.05 based on one-way ANOVA with multiple comparisons (see Supplementary Table 1). All data are presented as mean values ± SEM.

Fig. 2. Novel mouse models CaMKIIα^Cre/T29 (transgenic) and Kcng4^Cre (knock-in/knock-out) to discriminate PC subpopulations.

A–C and E,F Serial transverse sections of the cerebellum of CaMKIIα^Cre/T29;Ai14 (A) and Kcng4^Cre;Ai14 (E) mouse immunolabeled for Hsp25 (B), KCTD12 (C), and AldoC (F). White brackets indicate the Hsp25 domains (B) and the corresponding TdTomato+ domains in the CaMKIIα^Cre/T29;Ai14 (A), as well as KCTD12+ domains (C). Yellow brackets indicate the TdTomato+ domains from the CaMKIIα^Cre/T29;Ai14, which do not correspond to any Hsp25+ domains, with a zoom on the flocculus (A₅ and B₅) to highlight that in this particular region, CaMKIIα^Cre/T29 expression extends beyond Hsp25 and matches KCTD12 expression (expression pattern confirmed in n = 4 CaMKIIα^Cre/T29;Ai14 mice). Blue and red brackets and dotted lines show the complementation of TdTomato +, from the Kcng4^Cre;Ai14 and the AldoC+ populations (E₁ and F₁). Scale bar = 1 mm (applies to all images). D Schematic representation of Cre expression in the CaMKIIα^Cre/T29 relative to Hsp25 and KCTD12 expression profiles. G Schematic representation of Cre expression in the Kcng4^Cre relative to AldoC expression profile.

Fig. 3. Anterograde Cre-dependent tracing from floccular regions 9+ and 9- reveals discrete long-range projections.

A Transverse sections of the flocculus of CaMKIIα^Cre/T29;Ai14 (yellow) and Kcng4^Cre;Ai14 (blue) mice with KCTD12 (purple) immunolabeling. Scale bar = 200 µm. B and C TdTomato (Cre-dependent) labelled fibers in the vestibular structure following injection of AAV1-CAG-Flex-TdTomato in the flocculus of CaMKIIα^Cre/T29 (yellow) or Kcng4^Cre (blue) animals. Scale bars = 200 µm. D Schematic of floccular domains based on PC molecular subpopulations (top) and functional zones described by Schonewille et al. (bottom). PrH prepositus hypoglossi, SuVe superior vestibular nucleus, LVe lateral vestibular nucleus, d/vMVePC dorsal/ventral medial vestibular parvicellular nucleus, MVeMC medial vestibular magnocellular nucleus, PFl paraflocculus, Fl flocculus, SpVe spinal vestibular nucleus, Lat lateral cerebellar nucleus, IntP nucleus interpositus posterior, IntA nucleus interpositus anterior, DPGi dorsal paragigantocellular nucleus of the reticular formation, Sol solitary nucleus, 4V fourth ventricle, Y group Y.

Fig. 4. Monosynaptic rabies tracing from neural subpopulations in the vestibular complex and PrH.

A Schematic of the procedure for monosynaptic tracing with genetically modified rabies virus. AAV helper (AAV1-EF1a-Flex-GTB or BA-96-AAV2/1-pAAV-Syn-Flex-nGToG-WPRE3) and rabies viruses (EnVA-G-deleted-Rabies-mCherry or BRABV-001-pSADB19dG-mCherry) were injected in the vestibular complex and/or PrH at different time points in either GAD67^Cre, GlyT2^Cre, VGluT2^Cre or SST^Cre mice. Retrograde trans-synaptically-labelled PCs are then identified in the flocculus. B Example of primary infected neurons in the PrH and MVePC (left, scale bar = 1 mm; middle, scale bar = 200 µm) following injections in a SST^Cre mouse, with trans-synaptically labelled 9 + PC in the flocculus (right, Scale bar = 200 µm). C–E Localization of the primary infected cells in GAD67^Cre/GlyT2^Cre (C), VGluT2^Cre (D) or SST^Cre (E) mice. Each color represents a different injection, and associated marks in the flocculus represent single PCs identified following retrograde trans-synaptic labelling. Distribution of PC subpopulations in the flocculus are illustrated with yellow (9+) and blue (9−). PrH prepositus hypoglossi, SuVe superior vestibular nucleus, LVe lateral vestibular nucleus, MVePC medial vestibular parvicellular nucleus, MVeMC medial vestibular magnocellular nucleus, PFl paraflocculus, Fl flocculus, SpVe spinal vestibular nucleus, DPGi dorsal paragigantocellular nucleus of the reticular formation, Gi gigantocellular nucleus of the reticular formation, 4V fourth ventricle. Schematics adapted from Paxinos & Franklin, 2001.

Fig. 5. Selective optogenetic stimulation of the 9+ and 9− PC subpopulations drives distinct kinematics.

A Schematic image of the recording settings. The left flocculus is optogenetically stimulated (blue shading indicates stimulation time) while simultaneously recording the ipsilateral eye position in X and Y coordinates. The green arrows indicate the direction of the eye movement during optogenetic stimulation of all PCs in Pcp2^Cre;Ai27 animals. The horizontal eye movement (naso-temporal, in degrees) is shown as a positive deflection. The vertical eye movement (downward) is shown as a negative deflection. B Means of horizontal (top) and vertical (bottom) eye movement traces for Pcp2^Cre;Ai27 (green), Kcng4^Cre;Ai27 (blue), and CaMKIIα^Cre/T29;Ai27 (yellow) mice. Optogenetic stimulation durations are 200, 500 or 1000 ms, indicated by blue shading. C Velocity profiles of mean magnitude eye movement response during 1000-ms optogenetic stimulation. The velocity profile shows a peak after LED onset and LED offset. The heatmap shows the statistical divergence in velocity at each time point using the t-score from the unpaired two-tailed Student’s t test. Velocity is statistically significant when the t-score surpassed t > 2.043 for Pcp2^Cre;Ai27 / Kcng4^Cre;Ai27 (df = 30), and t > 2.064 for Pcp2^Cre;Ai27 / CaMKIIα^Cre/T29;Ai27 (two-tailed t-value), indicated by the asterisks. Dotted lines indicate movement phases based on peak velocity. D Representation of the four phases (1) drive, (2) hold, (3) release, and (4) recovery obtained from the velocity traces, displaying the absolute magnitude of the movement for optogenetic stimulation of 1000 ms. Blue shading indicates the time period of optogenetic stimulation. E Maximum amplitude of the eye movement response reached at the end of optogenetic stimulation for all optogenetic stimulation durations. F Initial peak velocity of the magnitude eye movement response occurring shortly after LED onset. G Absolute mean velocity in the four phases of the eye movement. H Distance traveled by the eye during the four phases of the induced eye movement. ***: p < 0.001, **: p < 0.01; *: p < 0.05 05 based on two-way ANOVA with multiple comparisons (see Supplementary Table 1). n = 14 Kcng4^Cre;Ai27 mice, 15 Pcp2^Cre;Ai27 mice and 8 CaMKIIα^Cre/T29;Ai27 mice. All data are presented as mean values ± SEM except (G), which is presented as box plots where the median is shown as the line, boxes extend from 25th to 75th percentile and whiskers extend to min and max values.

Fig. 6. Selective optogenetic stimulation of the 9+ and 9− PC subpopulations differentially influences sensory input-driven eye movements.

A–C Schematic of the VOR, OKR and VVOR experimental design (top). During sensory stimulation in Pcp2^Cre;Ai27 (green), Kcng4^Cre;Ai27 (blue), and CaMKIIα^Cre/T29;Ai27 (yellow) mice, optogenetic stimulation was either absent or provided during the portion in which sensory stimulation evoked a naso-temporal (light red, light purple) or temporo-nasal (dark red, dark purple) eye movement. Optogenetic activation of (sub)populations of PCs during vestibular stimulation (VOR), visual stimulation (OKR), or both (VVOR) drove differential responses between genotypes. Temporal (T) and the Nasal (N) side for the direction of the eye are indicated on the right. D–F Subtraction of eye movements during sensory stimuli in the absence of optogenetic stimulation from those with optogenetic stimulation to compare the impact of stimulation in Pcp2^Cre;Ai27 (green) and Kcng4^Cre;Ai27 (blue) mice. ***: p < 0.001, **: p < 0.01; *: p < 0.05 based on two-way ANOVA with multiple comparisons (see Supplementary Table 1). n = 10 Kcng4^Cre;Ai27 mice, 15 Pcp2^Cre;Ai27 mice and 8 CaMKIIα^Cre/T29;Ai27 mice. All data are presented as mean values ± SEM.