Cover ArticleResearch Articles, Cellular/Molecular

Dendritic Inhibition by Shh Signaling-Dependent Stellate Cell Pool Is Critical for Motor Learning

Wen Li, Lei Chen, Jonathan T. Fleming, Emily Brignola, Kirill Zavalin, Andre Lagrange, Tonia Rex, Shane A. Heiney, Gregory J. Wojaczynski, Javier F. Medina and Chin Chiang
Journal of Neuroscience 29 June 2022, 42 (26) 5130-5143; DOI: https://doi.org/10.1523/JNEUROSCI.2073-21.2022

Abstract

Cerebellar inhibitory interneurons are important regulators of neural circuit activity for diverse motor and nonmotor functions. The molecular layer interneurons (MLIs), consisting of basket cells (BCs) and stellate cells (SCs), provide dendritic and somatic inhibitory synapses onto Purkinje cells, respectively. They are sequentially generated in an inside-out pattern from Pax2+ immature interneurons, which migrate from the prospective white matter to the ML of the cortex. However, little is known about how MLI subtype identities and pool sizes are determined, nor are their contributions to motor learning well understood. Here, we show that GABAergic progenitors fated to generate both BCs and SCs respond to the Sonic hedgehog (Shh) signal. Conditional abrogation of Shh signaling of either sex inhibited proliferation of GABAergic progenitors and reduced the number of Pax2+ cells, whereas persistent Shh pathway activation increased their numbers. These changes, however, did not affect early born BC numbers but selectively altered the SC pool size. Moreover, genetic depletion of GABAergic progenitors when BCs are actively generated also resulted in a specific reduction of SCs, suggesting that the specification of MLI subtypes is independent of Shh signaling and their birth order and likely occurs after Pax2+ cells settle into their laminar positions in an inside-out sequence. Mutant mice with reduced SC numbers displayed decreased dendritic inhibitory synapses and neurotransmission onto Purkinje cells, resulting in an impaired acquisition of eyeblink conditioning. These findings also reveal an essential role of Shh signaling-dependent SCs in regulating inhibitory dendritic synapses and motor learning.

SIGNIFICANCE STATEMENT The cerebellar circuit that enables fine motor learning involves MLIs of BCs and SCs, which provide dendritic and somatic inhibitory synapses onto Purkinje cells. Little is known about how their identities and numbers are determined, nor are their specific contributions to motor learning well understood. We show that MLI subtypes are specified independent of Shh signaling and their birth orders but appear to occur in their terminal laminar positions according to the inside-out sequence. This finding challenges the current view that MLI subtypes are specified sequentially at the progenitor level. We also demonstrate that dendritic inhibition by Shh signaling-dependent SC pool is necessary for motor learning.

Introduction

The cerebellum plays an essential role in fine motor learning characterized by an adaptive process that involves recurring error-evoked learning to maintain optimum motor performance (Raymond and Medina, 2018). The circuit that enables this mode of learning is composed of different classes of inhibitory and excitatory neurons that are generated during embryonic and postnatal development from spatially distinct progenitors (Palay and Chan-Palay, 1974; Zhang and Goldman, 1996; Hoshino et al., 2005; Machold and Fishell, 2005; Wang et al., 2005). However, it remains unclear how their identities and numbers are specified to generate a functional circuit.

Purkinje cells (PCs) are the sole projection neurons in the cerebellar cortex, with a dendritic plane in the molecular layer (ML) where they integrate excitatory and inhibitory input from different sources (Beckinghausen and Sillitoe, 2018). The principal excitatory neurons are granule cells with their axons extending into the molecular layer as parallel fibers that provide excitatory input to PCs. Additionally, PCs receive excitatory input from two classes of extracerebellar afferent projections; climbing fibers of the inferior olive nuclei terminate in the molecular layer and contact PCs, and mossy fibers (from the brainstem and elsewhere) synapse with granule cells and thus influence many PCs at once. The balance to the excitatory neurotransmission is provided largely by GABAergic inhibitory inputs from molecular layer interneurons (MLIs) that consist of basket cells (BCs) and stellate cells (SCs; Sotelo, 2015). BCs are located in the inner one-third of the ML, and their axons form perineuronal nests, or baskets, as well as specialized structures known as pinceau around the PC soma/axon initial segment (Somogyi and Hámori, 1976). In contrast, SCs occupy the rest of the ML and make direct contact with distal PC dendrites. Despite the apparent morphologic and positional differences between BCs and SCs, the precise location and the boundary between these two cell types have not been well defined until the recent identification of Ret, a receptor tyrosine kinase, whose expression is restricted to BCs (Sergaki et al., 2017).

Cerebellar inhibitory neurons are generated from spatially distinct germinal zones. In mice, the early born neurons, including PCs, Golgi (granule layer interneuron), and Lugaro cells, are generated from the ventricular zone (VZ) of the fourth ventricle during early to midembryonic stages (Hoshino et al., 2005; Sudarov et al., 2011). The late-born MLIs are generated sequentially but in an overlapping manner from the prospective white matter (PWM) during the late embryonic to postnatal development, with BCs emerging first and followed by SCs (Zhang and Goldman, 1996). Studies have identified stem cell-like astroglia in the PWM as the source of transient amplifying GABAergic progenitors marked by Ptf1a (Fleming et al., 2013), a bHLH transcription factor whose function is required for the specification of all GABAergic lineages, including the Pax2+ immature interneuron (Hoshino et al., 2005; Pascual et al., 2007). Pax2+ cells migrate from PWM to ML before terminally differentiating into BCs and SCs (Weisheit et al., 2006). However, it remains unclear how MLI subtype identities are determined.

The Sonic hedgehog (Shh) signaling pathway has been shown to play critical roles during cerebellar development. It is required for the rapid expansion of granule cells by promoting the proliferation of granule cell precursors (Dahmane and Altaba, 1999; Wallace, 1999; Wechsler-Reya and Scott, 1999). Additionally, Shh signaling is transiently required to maintain the proliferative capacity of multipotent radial glial cells in the VZ and astroglia in the PWM (Huang et al., 2009; Fleming et al., 2013). Unlike VZ, Ptf1a+ progenitors are proliferative and exhibit Shh pathway activity in the PWM (Fleming et al., 2013). However, the significance of this activity in the regulation of MLI pool size and cerebellar function has not been determined. Although studies have shown that MLIs play an important role in cerebellar-dependent motor learning (ten Brinke et al., 2015; Sergaki et al., 2017), it remains to be determined to what extent each MLI subtype contributes to motor learning. In this study, we investigated the role of Shh signaling and its contribution to MLI subtype allocations and cerebellar-dependent motor learning.

Materials and Methods

Mice

All procedures followed animal care and biosafety protocols approved by the Vanderbilt University Division of Animal Care in accordance with National Institutes of Health guidelines. Histologic analyses were conducted at postnatal day (P)1 to P30. Behavior experiments used 3- to 5-month-old adult mice. The following mouse lines with either sex were used in this study: (1) Gli1nlaz, nlacZ reporter was inserted into the endogenous Gli1 locus, creating a null allele and activating nlacZ expression in a pattern indistinguishable from that of Gli1 (Bai et al., 2002); (2) Ptf1aCre/+, Cre was inserted into the endogenous Ptf1a locus, creating a null allele and activating Cre expression in a pattern indistinguishable from that of Ptf1a (Kawaguchi et al., 2002); (3) Ptf1aCreER/+, CreER was inserted into the endogenous Ptf1a locus, creating a null allele and activating CreER expression in a pattern indistinguishable from that of Ptf1a (Kopinke et al., 2012); (4) SmoF/F, the loxP sequence was inserted on either side of exon 1 at the Smo locus, permitting a Cre recombinase-dependent Smo loss of function mice (Long et al., 2001); (5) R26SmoM2, a lox-stop-lox eYFP SmoM2 cassette was inserted into ROSA26 locus, permitting the expression of eYFPSmoM2 fusion protein in a Cre-dependent manner (Jeong et al., 2004); (6) Ret GFP/+, exon 1 of the Ret locus was replaced by eGFP, creating a null allele and activating eGFP expression in a pattern similar to that of Ret (Jain et al., 2006); (7) R26iDTR, a loxp cassette containing diphtheria toxin receptor (DTR) was inserted into ROSA26 locus, permitting the expression of DTR in a Cre-dependent manner (Buch et al., 2005); and (8) R26RAi9, a CAG lox-stop-lox tdtomato cassette was inserted into ROSA26 locus, permitting the expression of tdtomato under the control of CAG promoter in a Cre-dependent manner (Madisen et al., 2010).

Mice treatments

For proliferation analysis, P3 and P6 pups were injected intraperitoneally with 50 mg/kg of EdU (5-ethyl-2-deoxyuridine; Sigma-Aldrich). Brains were harvested 2 h after injection and processed for frozen tissue sections. For fate mapping studies, P0, P3, and P5 pups were injected intraperitoneally with 50 mg/kg of EdU (P5 pups only) or 50 mg/kg of tamoxifen (Sigma-Aldrich) dissolved in corn oil. Brains were harvested at P21 and processed for frozen tissue sections as described below. P0 Pups were injected intraperitoneally with 3 ng of diphtheria toxin daily for two or three consecutive days for cell ablation experiments. Brains were harvested at P6 or P21 for frozen tissue sections.

Tissue processing for histologic analysis

Brains were retrieved from P0 to P8 after cervical dislocation and from deeply anesthetized adults after transcardial perfusion with 4% paraformaldehyde (PFA). Animals were anesthetized with 100 mg/kg ketamine and 5 mg/kg of Xylazine. Brains were then fixed with 4% PFA for 48 h at 4°C and processed for paraffin embedding. Sections of 5 mm were then collected on glass slides (Fisher Scientific), and paraffin was removed using xylene in a three-wash series (1 × 10 min, 2 × 5 min). Sections were rehydrated with a descending EtOH series (100% 2×, 95, 75, 50%) 3 min each followed by two 1 min washes in H2O. Sections were stained with hematoxylin and eosin solution. Golgi stains on adult sections were performed using FD Rapid GolgiStain Kit (FD NeuroTechnologies) according to instructions from the manufacturer.

Immunohistochemistry

Immunohistochemistry analyses were performed on frozen tissue sections. Fixed brain tissues described above were rinsed with PBS and immersed in 30% sucrose solution before embedding in tissue freezing medium (optimal cutting temperature). Frozen tissues were sectioned on a Leica CM1950 cryostat at 15 mm. Endogenous peroxidases were blocked with 3% H2O2 in MeOH (5 ml 30% H2O2 in 45 ml 100% MeOH) for 10 min followed with a 1× PBS wash for 3 × 5 min. Sections were incubated with PBS blocking solution containing 10% normal goat or donkey serum and 0.1% Triton X-100 at room temperature for 1 h. A moisture chamber was prepared, and slides were placed into it. One hundred microliters of primary antibody (prepared in PBS blocking solution) per slide was added, and each was covered with a coverslip and allowed to sit overnight at 4°C. The following day, coverslips were removed, and slides were washed in 1× PBS, 3 × 5 min each. One hundred microliters of secondary antibody per slide was added, and each was covered with a coverslip and allowed to sit for 1 h at room temperature. Slides were washed in 1× PBS, 3 × 5 min each and were mounted using FluoroSave (Millipore) before imaging. For EdU detection, sections were incubated in PBS with 0.1 m CuSO4, 0.1 m tris-hydroxypropyltriazolylmethylamine (THPTA), 2 mm Azide 488, 1 m L-Ascorbic at room temperature for 30 min. Sections were washed 3 × 5 min in PBS and then incubated with 5 µg/ml DAPI at room temperature for 10 min. All fluorescent images were acquired on a Leica DMi8 microscope or Zeiss LSM 700 confocal microscope, processed using Leica LAS X or Zeiss ZEN software and analyzed using ImageJ (National Institutes of Health).

Rotarod test

Motor coordination and learning were tested using a commercially available (Harvard Apparatus) accelerating rotarod. The cylinder was 3 cm in diameter and was covered with textured rubber. Mice were confined to a section of the cylinder, 5.7 cm wide, by gray plastic dividers. The height to fall was 16 cm. Mice were placed on the accelerating rotarod whose speed gradually increased from 4 to 40 rpm over the course of a 300 s trial. The time taken for the mouse to fall from the rotating rod was recorded. Mice that fell in <15 s were given a second trial. Occasionally, mice clung to the rod, and the whole animal rotated along with it. This behavior was classified as a “rotation,” and the time at which this occurred for the first time on each trial was also recorded for each mouse. Thus, the rotarod score was defined as latency to fall or to the first rotation, whichever occurred first. Three sessions were conducted on consecutive days, with three trials per session.

Gait analysis

Mice were placed on a translucent treadmill at a standstill. The treadmill was then turned on at a speed of 20 cm/s, and still images of their paw movements and placements were taken with a high-speed video camera underneath the belt. Only 10–20 s of footage was recorded to obtain ∼40 indices of gait from the analysis software. The following parameters were measured: stance time, time elapsed while foot is in contact with the treadmill, in its stance phase; swing time, time elapsed while foot is in the air, in its swing phase; stride time and time elapsed between two successive initiations of stances; brake time and time elapsed from the first contact with the treadmill to peak of stance; deceleration; and propel time and time elapsed from peak stance to full swing.

Eyeblink conditioning in head-fixed mice

The apparatus and experimental procedure for eyeblink conditioning were previously described (Heiney et al., 2014). Briefly, mice were anesthetized with isoflurane (1.5–2% by volume) and positioned on the stereotaxic apparatus prepared for surgery. The skull was exposed, and two small holes were drilled on either side of the midline near bregma for inserting screws. A thin aluminum head plate was then placed over the bregma, and the screws were fitted into the central hole in the head plate, which was affixed to the skull by Metabond cement.

Mice were habituated to head restraint for 1 h in three habituation sessions on top of a foam cylinder with the head fixed before the start of the conditioning sessions. Mice were exposed to either a 240 ms blue LED light positioned 20 cm in front of the mouse or a 240 ms tone of white noise delivered via a speaker (4 Ω magnetic speaker, FF1, Tucker-Davis Technologies) as a conditioned stimulus (CS). The volume of the white noise was set to just below the threshold that causes transient startle movement of the eyelid. The unconditioned stimulus (US) was a periocular air puff (30–40 psi) of 40 ms duration and delivered via a 23 gauge needle placed 5 mm from the cornea of the mouse. The pressure of the periocular air puff was set for each mouse to induce a full reflexive blink as the unconditional response. The CS/US interstimulus interval was 200 ms. Mice received 100 trials (80 CS/US paired trials and 20 CS-only trials) per day for 14 d. Eyelid movement was detected under infrared illumination using a high-speed (200 or 350 frames/s) monochrome video camera (Allied Vision). Eyelid positions, ranging from 1 (fully closed) to 0 (fully opened), were determined as the fraction of eyelid closure (FEC) using custom MATLAB software and the Image Acquisition Toolbox. The FEC represents the proportion of the distance between the two eyelids. The response was considered to be a conditioned response (CR) if the FEC exceeded 0.2 after the CS onset but before the US.

Electrophysiology recording

Cerebella from P21 to P30 were retrieved and embedded in agarose while affixed to the specimen holder using adhesive glue. Sagittal slices 300 mm thick of cerebellar vermis were prepared using a vibrating microtome (VT1000S, Leica) in oxygenated prechilled medium containing 125 mm NaCl, 2.5 mm KCl, 4 mm MgCl2, 1.25 mm KH2PO4, 26 mm NaHCO3, 1 m CaCl2, and 25 mm D-glucose, pH 7.3–7.4. The slices were then transferred to a recording chamber and superfused with artificial CSF and gassed with a mixture of 95% O2/5% CO2 >1 h at ice-cold temperature before recordings. Whole-cell recordings from Purkinje cells in cerebellar lobules 4–7 (voltage clamped at −60 mV) were performed at room temperature with borosilicate glass pipettes (2–4 MΩ) pulled with a vertical micropipette puller (PC-10, Narishige). Recording electrodes were filled with internal solution containing the following (in mm): 140 CsCl, 4 NaCl, 0.5 CaCl2, 10 HEPES, 5 EGTA, 2 Mg-ATP, and 0.4 GTP, pH 7.3. For sIPSCs recording, 10 mm 6-Cyano-7-nitroquinoxaline-2,3-IPC (CNQX) and 50–100 mm amino-5 phosphonopentanoic acid (APV) were added to the external solution to block glutamate receptor-mediated sEPSCs. Membrane currents were recorded using a MultiClamp 700B amplifier (Molecular Devices) connected to a DigiData 1440 (Molecular Devices) using pClamp 10.2 software (Molecular Devices). Series resistance (8–14 MΩ) was monitored throughout the experiments, and experimental data were discarded if the value changed by >20%. All signals were filtered at 2 kHz and sampled at 5–10 kHz. sIPSCs were analyzed with a threshold of 10 pA. Postsynaptic currents were analyzed using Clampfit 11 software (Molecular Devices). Amplitudes and interevent intervals were measured as a mean of the values obtained from 5 min recording sessions and analyzed using Prism 8 software (GraphPad) for statistical analysis and graphic presentations.

ERG and visual evoked potential recordings

Mice were dark adapted overnight, dilated with 1% tropicamide for 10 min, and anesthetized with 20/8/0.8 mg/kg ketamine/xylazine/urethane. Anesthetized mice were placed on the warmed surface of the Celeris ERG system (Diagnosys), and corneal electrodes were placed on eyes lubricated with Genteal eye drops. Subdermal platinum needle electrodes were placed in the snout and tail as reference and ground electrodes, respectively. Additional platinum electrodes were placed on the right and left sides of the back of the head, near the visual cortex. The ERG and visual evoked potential (VEP) were recorded sequentially during the same session. For the ERG recording, mice were exposed to 15 flashes of 1 Hz, 1 cd.s/m2 white light. For the VEP recording, mice were exposed to 50 flashes of 1 Hz, 0.5 cd.s/m2 white light. At the end of the session, mice recovered on a warm pad and then were returned to their cage.

Quantifications and statistical analysis

ImageJ (National Institutes of Health) software was used to measure the area (µm2) for regions of interest (ROIs) and for the acquisition of cell counting. For each stage, four to six midsagittal sections (∼10 µm thick) of each cerebellum were used for quantitative analysis. For quantification of cells in the white matter, inner granule layer, and molecular layer, the entire laminate was used as ROIs. Cells were counted using ImageJ plug-in Cell Counter or Nucleus Counter according to the subcellular distribution of target antigen. For quantification of synapses, four to five 5000 µm2 rectangular ROIs in the molecular layer of lobules IV and V per sample were used. Positively stained puncta were counted by ImageJ plug-in Puncta Analyzer. For quantification of PC spine density, Differential interference contrast (DIC) images were taken on a Leica TCS SP5 microscope of sagittal cerebellar sections from Ptf1aCre; SmoF/- mice and control littermates. Images were taken at a picture size of 1024 × 1024 pixels, and a 63× objective was used for dendritic spine counting analysis. Z-stacks of individual cells were captured, and the length of dendritic spines was measured by tracing a segmented line from the base of the spine to its apex. The number of spines per visible segment was quantified manually in ImageJ and provided an index of spine density (spines per µm). Manual counting was done blindly, and only spines that had strong, clear Golgi stain were counted. All quantitative data were analyzed using Prism 8 software (GraphPad) for statistical analysis and graphic presentations. Unpaired Student's t test was used to compare the statistical difference between control and conditional mutant animals (Table 1).

Results

Shh signaling is transiently activated in Ptf1a+ progenitors in the PWM

To better define Shh signaling in GABAergic progenitors, we used β-galactosidase (β-gal) expression from GlinlacZ mice (Bai et al., 2002) to assess Shh pathway responsiveness among Ptf1a+ cells at different postnatal stages. We found that Shh signaling is transiently activated in a subset of Ptf1a+ cells from P1 to P8 (Fig. 1A), whereas contemporaneous populations of Ptf1a+ cells in the brain/brainstem were negative for the signal. At P1, the β-Gal+ and Ptf1a+ cells account for ∼19% of Ptf1a+ progenitors in the PWM, and this number increased to ∼26% at P3 (Fig. 1B,C). However, at P6, the percentage of double-positive cells declined to ∼14%, but no statistically significant changes were observed for Ptf1a+ progenitors (Fig. 1B,C). By P8, a very small population of Ptf1a+ progenitors were positive for β-gal expression (Fig. 1C). Thus, the Shh pathway is transiently activated in a subset of Ptf1a+ progenitors.

Figure 1.
Figure 1.

Shh signaling is transiently activated in Ptf1a+ progenitors and is required for MLI expansion. A, Midsagittal sections of Gli1nlacZ mice showing β-gal and Ptf1a expression in PWM at P1, P3, P6, and P8. Arrowheads indicate cells coexpress β-gal and Ptf1a. Scale bars: 50 µm. B, C, Quantification of total Ptf1a + cells (B) and the percentage of β-gal +/Ptf1a+ double-positive cells (C) in the PWM. N= 4 mice per group. D, Brains from Smo F/+ and Ptf1aCre/+; SmoF/- animals at P21 showing external views and H&E stained sections. E, F, Quantification of the overall midsagittal area of the cerebellum and the molecular layer in P21 Smo F/+and Ptf1aCre/+; Smo F/- mice; N = 3–5 mice per group. G, Midsagittal cerebellar sections from P21 Smo F/+and Ptf1aCre/+; Smo F/- mice stained with antibodies against Parv. Scale bars: 50 µm. H–J, Quantitative analysis of GABA+ DCN (H), Parv+ interneurons (I), Calbindin+ Purkinje cells (I), and Neurogranin+ Golgi cells (J) in P21 Smo F/+and Ptf1aCre/+; Smo F/- cerebella; N = 3–5 mice per group. K, The length of dendritic spines of Purkinje cells in P21 Smo F/+and Ptf1aCre/+; Smo F/- cerebella; N = 3 mice per group. All graphs displayed are mean ± SEM; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. n.s., Not significant (Extended Data Figs. 1-1, 1-2).

Figure 1-1

There is no significant difference in the number of MLIs in different control cerebella. A, Midsagittal sections from P30 animals with various genotypes stained with antibody against Parv. B, Quantitative analysis of Parv+ interneurons shown in A; N = 3 mice per group. Scale bars: 30 µm. Download Figure 1-1, TIF file.

Figure 1-2

Representative images of different cerebellar cell types in SmoF/+ and Ptf1aCre; SmoF/- mice. A, B, Sagittal sections of P30 cerebellar stained with GABA (A) and Calbindin and Neurogranin (B). Scale bars: 50 µm. Download Figure 1-2, TIF file.

To ascertain the significance of Shh signaling in Ptf1a+ progenitors, we blocked Shh reception by genetically ablating the transducer Smo using the GABAergic lineage-specific Ptf1acre line (Kawaguchi et al., 2002). Our initial analysis of the mature cerebellar cortex of Ptf1aCre; SmoF/- mutant mice at P30 detected no structural or layering alterations; thickness of the molecular layer, granule layer, overall cerebellar area, and foliation pattern was comparable to the that of the control (Fig. 1D,F). However, there is a clear reduction of hematoxylin/eosin-stained cells in the molecular layer. Indeed, immunohistochemical detection of the calcium-binding protein parvalbumin (Parv), which preferentially marks MLIs and PCs, showed a statistically significant decrease in the total number of MLIs in Ptf1aCre; SmoF/- mutant cerebella (Fig. 1D,G), whereas no difference was observed in Ptf1aCre, SmoF/-, SmoF/+, Smo+/−, Ptf1aCre/+;Smo+/− and Ptf1aCre;SmoF/+ cerebella (Extended Data Fig. 1-1). The most reduction was most prominent in the outer half of the ML, suggesting that a germinal program was biased toward the stellate interneuron subtype (see below). The number of other early born inhibitory neurons, including deep cerebellar nuclei (DCN) and PCs and Golgi cells expressing GABA, Calbindin, and Neurogranin, respectively, was not significantly changed in Ptf1aCre; SmoF/- mutants (Fig. 1H,J; Extended Data Fig. 1-2). Additionally, no significant changes in Purkinje cell spine density were observed (Fig. 1K).

Shh signaling regulates precursor Pool expansion

To understand how Shh signaling affects the MLI pool size, we evaluated the proliferation and survival of Ptf1a progenitors in the PWM. Following a short-term (1 h) labeling with the thymidine analog EdU, we observed a significant reduction of proliferating Ptf1a progenitors in Ptf1aCre; SmoF/- mutants at both P3 (∼20%) and P6 (∼34%) when compared with controls in the vermis (Fig. 2A–C). We observed no statistical difference of Ptf1a+ cells among Ptf1aCre, SmoF/-, SmoF/+, Ptf1aCre/+;Smo+/− and Ptf1aCre;SmoF/+ cerebella (Extended Data Fig. 2-1). Accordingly, the total number of Ptf1a+ progenitors in the mutant was reduced by ∼30% and ∼68% at P3 and P6, respectively (Fig. 2B,C). The proliferative effect of Shh signaling on Ptf1a+ progenitors was not restricted to the vermal region, as a similar reduction (∼39%) was also observed in parasagittal sections at P6 (Extended Data Fig. 2-2). Viability was measured with cleaved-Caspase-3, but no appreciable apoptotic cells were found in Ptf1a lineage cells at either P3 or P6 (Extended Data Fig. 2-3).

Figure 2.
Figure 2.

Shh signaling promotes Ptf1a+ progenitor proliferation and the expansion of Pax2+ immature interneurons. A–C, Ptf1a and EdU staining in PWM of Smo F/+and Ptf1aCre/+; Smo F/- cerebella at P3 and P6. Arrowheads indicate cells positive for both EdU and Ptf1a. Quantification of total Ptf1a+ cells in the PWM at P3 and P6 (B). The percentage of EdU +/Ptf1a+ double-positive cells relative to the total number of Ptf1a+ cells in PWM at P3 and P6 (C) is also quantified; N = 3 mice per group. Scale bars: 50 µm. D, Pax2 and NeuN staining in sagittal sections of Smo F/+and Ptf1aCre/+; Smo F/- cerebella. Right, A higher magnification view of boxed regions. Scale bars: 100 µm. E–H, Quantitative analysis of Pax2+ cells in different regions of the cerebellum shown in D; N = 3 mice per group. I, Ptf1a and EdU staining in PWM of SmoM2 and Ptf1aCre/+; SmoM2 mice at P6. Arrowheads indicate cells positive for both EdU and Ptf1a. Scale bars: 50 µm. J, K, Quantitative analysis of total Ptf1a + cells (J) and the percentage of EdU +/Ptf1a+ double-positive cells relative to the total number of Ptf1a+ cells (K) in PWM at P6. L, Pax2 staining in the ML of SmoM2 and Ptf1aCre/+; SmoM2 mice at P6. Scale bars: 100 µm. M, Quantitative analysis of Pax2+ cells shown in L; n = 3 mice per group. N, Ptf1a and Ccnd2 staining in PWM of SmoM2 and Ptf1aCre/+; Smo F/- mice at P6. Arrowheads indicate cells coexpressing Ccnd2 and Ptf1a. Scale bars: 50 µm. O, Quantitative analysis of the percentage of Ccnd2 and Ptf1a double-positive cells relative to the total number of Ptf1a+ cells in PWM at P6; N = 3 mice per group. All graphs displayed are mean ± SEM; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. n.s., Not significant (Extended Data Figs. 2-1, 2-2, 2-3).

Figure 2-1

There is no significant difference in number of Ptf1a+ progenitors in different control cerebella. A, Midsagittal sections from P6 animals with various genotypes stained with antibody against Ptf1a. B, Quantitative analysis of Ptf1a+ interneuron progenitors shown in A; N = 3 mice per group. Scale bars: 50 µm. Download Figure 2-1, TIF file.

Figure 2-2

Shh signaling promotes Ptf1a+ progenitor proliferation. A–C, Cerebellar parasagittal sections from P6 Smo F/+ and Ptf1aCre/+; SmoF/- mice stained with EdU and the antibody against Ptf1a. Quantification of total Ptf1a + cells in the PWM at P6 (B). The percentage of EdU +/Ptf1a+ double-positive cells relative to the total number of Ptf1a+ cells in PWM at P6 (C) is also quantified. Scale bar, 50 µm. Download Figure 2-2, TIF file.

Figure 2-3

Loss of Shh signaling in Ptf1a+ progenitors does not activate programmed cell death. Cerebellar sections from P6 Smo F/+ and Ptf1aCre/+; SmoF/- mice stained with antibodies against cleaved-caspase-3, NeuN, and DAPI. Scale bar, 50 µm. Download Figure 2-3, TIF file.

Genetic fate-mapping studies have shown that Ptf1a+ cells emerge upstream of and contribute considerably to neonatal Pax2+ immature interneuron pools (Fleming et al., 2013). The peak of Pax2 production in PWM occurs at P5, where they subsequently migrate to ML through the inner granule layer (IGL) before terminally differentiating into BCs and SCs (Weisheit et al., 2006; Wefers et al., 2018). We measured the abundance of Pax2+ cells at P6 and observed a significant reduction of their numbers in PWM, IGL, and ML (Fig. 2D–H).

The role of Shh signaling in promoting Ptf1a progenitor proliferation was further evaluated using a gain-of-function approach. We generated Ptf1acre; SmoM2 mice by crossing the Ptf1aCre driver strain to SmoM2 conditional mutants that harbor a constitutively activated form of Smo (Mao et al., 2006). Consistent with the loss of function study, we observed a significant increase in the number of Ptf1a+ progenitors (∼27%) as well as their proliferative capacity (∼15%) in Ptf1acre; SmoM2 cerebella when compared with controls at P6 (Fig. 2I–K). Accordingly, the number of Pax2+ cells is also significantly increased (Fig. 2L,M). Together, we provide strong evidence for a direct and essential role of Shh signaling in driving the proliferation of Ptf1a+ progenitors in PWM and subsequent expansion of Pax2+ immature interneurons.

Previous studies have shown that Cyclin d2 (Ccnd2) is expressed in the PWM and required for the proliferation of MLI progenitors (Huard et al., 1999), but the signal that activates its expression is unknown. We therefore examined Ccnd2 expression and found that the number of Ccnd2+ GABAergic progenitors is significantly reduced in the Ptf1aCre; SmoF/- PWM by immunohistochemistry (Fig. 2O,P). These data suggest that Shh signaling regulates Ccnd2 expression, which contributes to proliferation of Ptf1a progenitors.

Specification of MLI subtypes occurs independent of their birth orders

To precisely determine whether specific MLI subtypes are affected in Ptf1aCre; SmoF/- mutants, we acquired the RetGFP mouse line in which EGFP reporter is expressed from the endogenous Ret promoter (Jain et al., 2006). Consistent with previous studies (Sergaki et al., 2017), RetGFP expression in BCs is first detectable at P7 and continues throughout adulthood (Extended data Fig. 3-1). Outside the ML, RetGFP is expressed in the deep cerebellar nuclei (DCN) and associated fibers in PWM (Extended data Fig. 3-1), but not in Pax2+ interneuron progenitors as reported previously (Sergaki et al., 2017). The Ret+ DCN neurons are NeuN positive but GABA negative, indicating that they are not GABAergic neurons (Extended Data Fig. 3-1). Therefore, RetGFP expression is confined to mature BCs and DCN neurons. In addition to RetGFP expression, we incorporated Parv as a marker to highlight all MLIs. In this way, BCs and SCs can be unambiguously identified as RetGFP+ Parv+ and RetGFP− Parv+ cells, respectively. Using these markers, we were able to determine that SC numbers were reduced by ∼21% in Ptf1aCre; SmoF/- mutant cerebella (Fig. 3A,B) whereas no appreciable changes in BC numbers were observed when compared with those of controls (Fig. 3A,C). The reduction in SC numbers is not because of the inability of Pax2+ cells to differentiate into SCs as Pax2 expression is almost undetectable in the ML at P14 (Extended Data Fig. 3-2).

Figure 3.
Figure 3.

Shh signaling selectively regulates stellate cell pool size. A, Cerebellar sections from P21 Smo F/+; RetGFP/+ and Ptf1aCre/+; SmoF/-; RetGFP/+ mice stained with antibodies against GFP, Parvalbumin, and Calbindin. B–C, Quantitative analysis of Parv+ GFP+ (B) and Parv+ GFP− (C) cells shown in A; N = 5 mice per group. D, Cerebellar sections from P21 Ptf1aCre/+; R26RAi9; RetGFP/+ mice showing the extent of BC labeling (dtTomato+ GFP+) following TM induction at P0, P3, or P5. E–F, Quantitative analysis of the percentage of labeled BCs (dtTomato+ GFP+) and SCs (dtTomato+ GFP−) relative to total labeled (dtTomato+) cells. G, Quantitative analysis of the percentage of labeled BCs (dtTomato+ GFP+) relative to total GFP+ BCs. H, Cerebellar sections from P21 RetGFP/+ and Ptf1aCre/+; SmoM2; RetGFP/+ mice stained with antibodies against GFP and Parvalbumin. I–K, Quantitative analysis of Parv+ (I), Parv+ GFP+ (J), and Parv+ GFP− (K) cells shown in H; N = 3 mice per group. L, Cerebellar sections from P6 RetGFP/+ and Ptf1aCre/+; iDTR mice showing Pax2 distribution (green) in external granule layer (EGL), IGL, and WM following DT administration at P0–P2. M–P, Quantitative analysis of total Pax2+ cells (M) and Pax2+ cells in WM (N), IGL (O), and ML (P); N = 3 mice per group. Q, Cerebellar sections from P21 RetGFP/+ and Ptf1aCre/+; iDTR mice stained with antibodies against GFP and Parvalbumin. R–T, Quantitative analysis of Parv+ GFP+ (R), Parv+ GFP− (S), and Parv+ (T) cells shown in A; N = 3 mice per group. All graphs displayed are mean ± SEM; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. n.s., Not significant (Extended Data Figs. 3-1, 3-2, 3-3).

Figure 3-1

RetGFP is selectively expressed in BCs and DCN. A, Expression of RetGFP in the molecular layer is first detectable at P7. B, RetGFP expression in the PWM is only associated with noncellular fibers and does not colocalize with Pax2. C, D, RetGFP is also expressed in the NeuN+ DCN (C) and negative for GABA (D). Download Figure 3-1, TIF file.

Figure 3-2

Pax2 expression is no longer present in the ML of P14 cerebella. Midsaggital cerebellar sections of SmoF/F and Ptf1aCre; SmoF/- mice stained with antibodies against Ki67 and Pax2. Scale bar, 50 µm. Download Figure 3-2, TIF file.

Figure 3-3

DT administration efficiently depleted Ptf1a+ progenitors with no effect on early born Golgi cell numbers and no delay in BC to SC production. A, Cerebellar sections from P2 iDTR and Ptf1aCre/+; iDTR mice stained with antibodies against Ptf1a. B, Quantitative analysis of Ptf1a+ cells shown in A; N = 3 mice per group. C, Cerebellar sections from P21 iDTR and Ptf1aCre/+; iDTR mice stained antibodies against Neurogranin. D, Quantitative analysis of Neurogranin+ cells shown in A; N = 3 mice per group. D, Cerebellar sections from P21 iDTR and Ptf1aCre/+; iDTR mice stained with EdU and antibodies against Parv and Ret. All graphs displayed are mean ± SEM; *p < 0.05. n.s., Not significant. Scale bars: 50 µm. Download Figure 3-3, TIF file.

Classical birth dating and genetically inducible fate mapping (GIFM) studies have suggested that MLI subtype identities are correlated with their birth dates and laminar positions within the cerebellar cortex (Leto et al., 2006; Sudarov et al., 2011). To determine whether the subtype-specific effect of Shh signaling on MLIs is associated with temporal differences in Shh responsiveness of Ptf1a+ progenitors fated to generate both BCs and SCs, we generated Ptf1acreER; Ai9; RetGFP line to perform a GIFM study focusing on critical stages where Shh signaling is activated in Ptf1a progenitors. Tamoxifen (TM) administration at P0 marked both BCs (dtTomato+ RetGFP+) and SCs (dtTomato+ RetGFP−) in the molecular layer at P21, representing ∼21% and ∼79% of all marked cells, respectively (Fig, 3D,F). Moreover, marked BCs account for ∼56% of all RetGFP+ cells (Fig. 3G), indicating that the peak of BC production occurs between P0 and P1. When TM was administered at P3, marked BCs decreased to ∼30% (Fig. 3G), with a concomitant increase in marked SCs (Fig. 3F). At P5, SCs are the only labeled population present in the ML (Fig. 3F,G). Therefore, BCs and SCs are both generated at the time when Shh signaling is active in Ptf1a+ progenitors, consistent with previous Gli1CreER GIFM studies in which BC and SC subtypes are both generated from Shh-responsive progenitors (Fleming et al., 2013). Therefore, the differential effect of Shh signaling on MLI subtypes is unlikely because of temporal or regional difference in Shh responsiveness. Indeed, analysis of Ptf1acre; SmoM2 mutants with constitutive Shh pathway activation in Ptf1a+ progenitors showed that only RetGFP− Parv+ (SC) numbers are elevated whereas the RetGFP+ Parv+ (BC) pool remained unchanged when compared with the control (Fig. 3H–K).

The observation that altered numbers of Pax2+ immature interneurons in the gain and loss of Shh signaling mutants had no effect on BC pool size suggests that commitment to the BC fate from Pax2+ cells does not occur until they settle into the inner ML. Accordingly, the BC pool remains unchanged as long as there are sufficient Pax2+ cells to fill the inner ML. To further test this model, we generated Ptf1acre; RetGFP; ROSA26DTR mice designed to temporarily ablate Ptf1a+ progenitors and their descendent Pax2+ cells. The ROSA26DTR line expresses the Cre-inducible DTR that triggers cell ablation on exposure to Diptheria toxin (DT). Administration of DT at P0–P2, a time when Ptf1a+ progenitors are fated to generate most BCs, resulted in a drastic reduction of Pax2+ immature interneurons in all layers of the cerebellum at P6 (Fig. 3L–P), whereas DT treatment of Ptf1aCre/+ control had no effects. Despite ∼50% reduction of total Pax2+ cells (Fig. 3M), we observed no effect on BC numbers, whereas SCs are reduced by ∼49% at P21 (Fig. 3Q–T). The lack of perturbations in BC numbers is not because of selective depletion of Ptf1a+ progenitors at the later time period when SCs are generated as evidenced by >27% reduction of Ptf1a+ progenitors at P2 when DT was administered at P0 and P1 (Extended Data Fig. 3-3A,B). As expected, this reduction had no effect on early born Golgi cells (Extended Data Fig. 3-3C,D). We also ruled out the possibility of a delay in BCs to SCs production in DT-treated mice, which would have prolonged BCs production and -preferentially affected SC numbers. For this, we performed a classical birth-dating study on DT-treated pups (P0–P2) by administrating EdU at P5 when BCs are no longer generated (Fig. 3D–G). Analysis of EdU labeled cells at P21 revealed that only SCs (Ret− Parv+) but no BCs (Ret+ Parv+) were labeled after DT treatment (Extended Data Fig. 3-3E), indicating that the progenitor cell ablation did not elicit a delay in BCs to SCs production. Collectively, the results suggest that MLI subtypes are specified independent of their birth order but likely occur at their laminar positions, providing a mechanism by which Shh signaling is selectively promoting the expansion of SCs.

Impaired GABAergic synapses and inhibitory control over PCs

In the mature cerebellum, SCs and BCs relay inhibitory input to PCs via direct synapses on dendritic tree or somata/proximal initial segment, respectively, balancing excitatory input received from granule neurons and afferent fibers (Sotelo, 2015). To understand how the loss of Shh-dependent SCs might affect the integrity of cerebellar neural circuitry, we evaluated the distribution of specific molecular markers for inhibitory and excitatory synapses in the ML. We used a presynaptic marker, vesicular GABAergic transporters (VGAT), in combination with a postsynaptic marker Gephyrin to highlight inhibitory synapses. In the ML, the number of inhibitory synapses (puncta with colocalization) at PC dendrites was reduced by 31% in Ptf1aCre; SmoF/- mutants compared with controls (Fig. 4A–C). We also used another inhibitory presynaptic marker, glutamic acid decarboxylase 67 (Gad67), which is responsible for up to 90% of GABA synthesis in the brain (Asada et al., 1996; Kash et al., 1997), and found that its puncta are also significantly reduced (Extended Data Fig. 4-1). Consistent with unperturbed BC development, the number of inhibitory synapses at PC soma and the pinceau marked by the potassium voltage-gated channel 1.2 (Kv1.2) remains similar between Ptf1aCre; SmoF/- mutants and controls (Fig. 4D,E). However, VGluT1 (Fig. 4F,H), which labels excitatory synaptic terminals at granule cell parallel fibers, showed no significant changes, nor did VGluT2 (Fig. 4G,I), which marks excitatory synaptic terminals at climbing fibers. Altogether these findings show substantial impairment to the inhibitory component of the cerebellar system in Ptf1aCre; SmoF/- mutants, whereas the excitatory component appears to be unperturbed.

Figure 4.
Figure 4.

Loss of Shh-dependent stellate cells reduces ML GABAergic synapses. A–A2, Cerebellar sections from P21 Smo F/+ and Ptf1aCre/+; SmoF/- mice stained with antibodies against vGAT and Gephyrin. A1 and A2 are enlarged views of the boxed regions in A. PCL, Purkinje cell layer. Scale bars: 25 µm. B–C, Quantitative analysis of vGAT and Gephyrin double-positive synapses in the ML (B) and PCL (C); N = 4 mice per group. D, Cerebellar sections from P21 Smo F/+ and Ptf1aCre/+; SmoF/- mice stained with antibodies against Kv1.2 and calbindin. Scale bars: 50 µm. E, Quantitative analysis of Kv1.2 in PCL shown in D; N = 4 mice per group. F–G, Cerebellar sections from P21 Smo F/+ and Ptf1aCre/+; SmoF/- mice stained with antibodies against vGluT1 (F), vGluT2 (G), and calbindin. Scale bars: 25 µm. H, I, Quantitative analysis of vGluT1 and vGluT2 positive puncta in the ML; N = 3 mice per group. All graphs displayed are mean ± SEM *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. n.s., Not significant (Extended Data Fig. 4-1).

Figure 4-1

Loss of Shh-dependent stellate cells reduces ML GABAergic presynaptic puncta. A, Cerebellar sections from P21 Smo F/+ and Ptf1aCre/+; SmoF/- mice stained with antibodies against GAD67. Scale bars: 25 µm. B–C, Quantitative analysis of GAD67 in the ML (B) and PCL (C); N = 3 mice per group. All graphs displayed are mean ± SEM; *p < 0.05. n.s., Not significant. Download Figure 4-1, TIF file.

The significantly reduced number of dendritic inhibitory synapses in Ptf1aCre; SmoF/- mice suggests impaired neurotransmission. To directly test this, we performed whole-cell patch-clamp recordings on cerebellar slices to compare spontaneous IPSCs (sIPSCs) of PCs as a measure of inhibitory synaptic strength in wild-type (n = 10) and mutant (n = 8) mice. This analysis revealed that adult Ptf1aCre; SmoF/- mice had ∼35% decrease in the frequency of sIPSCs when compared with the control (Fig. 5A–C). Similarly, the amplitude of sIPSCs is also significant reduced in Ptf1aCre; SmoF/- mice (Fig. 5B). These findings indicate that PCs lose dendritic synapses and show attenuated inhibitory input.

Figure 5.
Figure 5.

Decreased amplitude and interevent interval of sIPSCs in Ptf1aCre/+; SmoF/- mutants. A, Representative traces of spontaneous IPSC recorded from Smo F/+ and Ptf1aCre/+; Smo F/- cerebellar Purkinje cells. B–C, Statistical analysis of sIPSCs amplitude (B) interevent interval (C). Each dot represents the mean of values obtained from 5 min recording sessions of individual PC; N = 15 cells from 11 Smo F/+ mice and 8 Ptf1aCre/+; Smo F/- mice. Values represent mean ± SEM; * p < 0.05.

Diminished capacity for associative motor learning

To determine whether the reduction of dendritic inhibitory neurotransmission on PCs might impair cerebellum-dependent functions, a series of neurobehavioral assays were performed. The cerebellum is classically known to coordinate motor function, and the most obvious manifestation of cerebellar impairment is abnormal gait. Ptf1aCre; SmoF/- mice did not demonstrate gross impairments in mobility, such as ataxia, nor were they observed to show clinical signs of tremor. When subjected to TreadScan analysis, adult Ptf1aCre; SmoF/- (n = 12) and control mice (n = 11) showed comparable gait, which is divided into the following three basic components: (1) stance (break plus propulsion), (2) swing, and (3) stride (stance plus swing), and is measured in both length and time (Fig. 6A–E). To test motor learning, these animals were also subjected to an accelerating rotarod assay. On the first day of testing sessions, mutant and control animals showed nearly identical latency to fall, indicating that no obvious differences in motor coordination, balance, or strength were detectable (Fig. 6F). However, during the subsequent days, it appeared that Ptf1aCre; SmoF/- mutants had an apparent deficiency in the ability to learn to perform this task, with statistically significant shorter latencies to fall than controls (Fig. 6F). Overall, loss of a SC subset did not appear to adversely impact gross motor function in Ptf1aCre; SmoF/- mutants and only imparted a subtle impairment to primitive motor learning.

Figure 6.
Figure 6.

Shh-dependent SC pool is required for cerebellum-dependent motor learning but not basic motor function. A–E, Quantitative gait analysis of adult Smo F/+ and Ptf1aCre/+; SmoF/- mice showing stance time (A), brake time (B), propel time (C), swing time (D), and stride time (E); N = 12 mice per group. F, Accelerating rotarod test to measure latency to fall on 5 consecutive days in Smo F/+ and Ptf1aCre/+; SmoF/- mice; N = 30 mice per group. G, Schematic of eyeblink conditioning system. The trial structure is consists of a 240 ms blue light (CS) or tone (CS) that precedes and coterminates with a 40 ms air puff (US). H–J, Group averages of trial-by-trial eyelid closure over the entire 14 d of conditioning with LED (H) or tone (J) in Smo F/+ and Ptf1aCre/+; SmoF/- mice. I–K, The percentage of conditioned responses LED (I) or tone (K); N = 5 mice per group. All graphs displayed are mean ± SEM; *p < 0.05, **p < 0.01. n.s., Not significant (Extended Data Figs. 6-1, 6-2, 6-3).

Figure 6-1

Ptf1aCre/+; SmoF/- mice showed the lower amplitude of eyelid closure. Average eyelid position traces on day 1, 4, 8, and 14 for each genotype. Error clouds are SD based on per mouse averages. Download Figure 6-1, TIF file.

Figure 6-2

Normal retinal function in the absence of Shh signaling in Ptf1a+ progenitors. A–F, Statistical analysis of electroretinogram of Smo F/+ and Ptf1aCre/+; SmoF/- mice showing amax (A), bmax (B), OP1 (C), OP2 (D), OP3 (E), and VEP N1 (F). Values are mean ± SEM; N = 8 mice for Smo F/+ and N= 7 mice for Ptf1aCre/+; SmoF/-. n.s., not significant. Download Figure 6-2, TIF file.

Figure 6-3

Retinoic acid signaling is not activated in BCs. Cerebellar sections from P21 RARE-hsp68LacZ; RetGFP/+ mice stained with antibodies against β-gal and GFP. Download Figure 6-3, TIF file.

View this table:
Table 1.

Key resources table

Proper cerebellar function is also necessary for delay eyeblink conditioning, a form of associative motor learning required for multisensory integration that is often impaired in autistic individuals (Sears et al., 1994; Medina et al., 2000). The paradigm involves pairing a neutral CS (e.g., LED light pulse or tone) with an eyeblink-evoking US (e.g., an air puff). After repeated CS-US pairing, an association is progressively established so that the eyelid CR, in the form of learned blink, occurs before the onset of the US. We, therefore, determined whether cerebellum-dependent eyeblink conditioning is compromised in Ptf1aCre; SmoF/- mice with reduced SC numbers. In this paradigm, a blue LED light pulse was used as the CS for a period of 240 ms, followed immediately by a gentle air puff for 40 ms (Fig. 6G). Eyelid movement was measured using a high-speed infrared camera interfaced with MATLAB (Heiney et al., 2014), and the eyelid position was assigned to a value between 0 (fully open) and 1 (fully closed). After 14 d of consecutive sessions of conditioning (1400 trials in total), Ptf1aCre; SmoF/- mice had a consistently lower amplitude of eyelid closure just before CS exposure when compared with control mice (Fig. 6H; Extended Data Fig. 6-1). Although both the mutant and control mice showed a gradual increase in the percentage of CRs, mutant mice had significantly lower CRs throughout the conditioning sessions (Fig. 6I).

Although the Shh pathway has not been reported to be active in the retina of Ptf1a-expressing cells and their lineage, we were concerned that the CR deficit in Ptf1aCre; SmoF/- mutants is because of compromised retina function. We, therefore, performed ERGs and VEPs to assess retinal and visual pathway activity in response to light. We observed no significant differences in maximum responses of rod photoreceptors (amax), inner retinal neurons (bmax) and oscillatory potentials (OP), or the visual cortex (VEP N1) between Ptf1aCre; SmoF/- and control mice (Extended Data Fig. 6-2), indicative of normal retinal and visual pathway function. To confirm that the mutants have a deficit in associative motor learning, we also used a tone as CS in delayed eyeblink conditioning. Similar to LED light, the mutant mice showed a diminished response to the tone as presented by the lower amplitude of eyelid closure and percentage of CRs (Fig. 6J,K). Collectively, these results indicate that the Shh signaling-dependent dendritic synapses formation plays a critical role in associative motor learning.

Discussion

MLIs constitute the majority of inhibitory interneurons in the cerebellum, but the mechanisms that regulate their subtype identities and pool sizes remain not well understood. We reveal that Shh signaling is activated in a subset of interneuron progenitors that give rise to both BCs and SCs, but surprisingly it is selectively required for the expansion of the SC pool. Through lineage tracing, genetic gain and loss of function, and cell ablation experiments, we show that molecular layer interneuron subtypes are specified independent of Shh signaling and their birth orders but appear to occur in their terminal laminar positions according to inside-out sequence. Our studies at the synaptic and behavioral levels show that dendritic GABAergic inhibition controlled by Shh signaling-dependent SC pools is critical for motor learning.

Previous studies have shown that MLIs originate from multipotent astroglia in the PWM (Silbereis et al., 2009; Fleming et al., 2013; Parmigiani et al., 2015). These astroglia act as resident stem-like cells and proliferate in response to the Shh signal from PCs to generate astrocyte precursors as well as Ptf1a progenitors (Fleming et al., 2013). Our finding that Shh signaling is directly required for the proliferation of Ptf1a+ progenitors and subsequent expansion of MLIs offers an additional mechanism by which Shh regulates MLI pool size.

It is estimated that ∼90% of all inhibitory neurons in the cerebellum are represented by late-born GABAergic interneurons (Weisheit et al., 2006). These neurons come from Pax2-expressing immature interneurons that are mostly generated during the first week of postnatal life (Weisheit et al., 2006). As Pax2+ cells descend from Ptf1a+ progenitors (Fleming et al., 2013) and emerge at the final progenitor cell division, Shh-induced proliferation of Ptf1a+ progenitors, therefore, serves as the main driver for their rapid expansion. This mode of regulation is in contrast to other regions of the nervous system where Shh pathway activity is generally excluded from postmitotic Ptf1a-expressing cells. In the spinal cord, Ptf1a expression is restricted to the postmitotic progenitors in the dorsal neural tube, whereas the Shh pathway is activated in the ventral domain (Glasgow et al., 2005; Bai et al., 2004). Moreover, ectopic Ptf1a expression is correlated with the downregulation of the Shh signaling pathway as observed in Danforth's short tail mice (Orchard et al., 2019). Similarly, in the developing retina, Shh signaling is activated in early retinal progenitors, whereas Ptf1a expression is restricted to postmitotic horizontal and amacrine cells (Fujitani et al., 2006). Indeed, the lack of Shh pathway activity in postmitotic retinal lineage cells is consistent with the absence of visual impairment in Ptf1acre/+; SmoF/- mice (Extended Data Fig. 6-3). Therefore, the Shh signaling pathway appears to be co-opted for the rapid expansion of Ptf1a+ progenitors to ensure an adequate supply of MLIs. Mechanistically, Ccnd2 is an attractive target that mediates the proliferative effect of Shh pathway activation on MLI progenitors. Unlike Ccnd1, Ccnd2 is strongly expressed in MLI progenitors, and loss of Ccnd2 impairs their production (Huard et al., 1999). Reduced expression of Ccnd2 in Smo-deficient Ptf1a progenitors is consistent with the Shh pathway regulating Ccnd2 expression.

It has been reported that cerebellar MLIs undergo limited programmed cell death during the first 2 weeks of postnatal development (Yamanaka et al., 2004). More recent studies showed that GDNF signaling is required for the survival of MLIs. Genetic deletion of GDNF receptor GFRa1 or Ret in GABAergic progenitors resulted in ∼25% reduction of MLIs (Sergaki et al., 2017). Interestingly, the loss of NeuroD2, a bHLH transcription factor required for MLI differentiation, also promotes MLI survival (Pieper et al., 2019). Thus, MLI pool size is regulated at multiple levels by distinct mechanisms, from the proliferation of multipotent astroglia and Ptf1a progenitors to the survival of MLIs.

Fate mapping and heterochronic transplantation studies have suggested that MLI identities and laminar placement link to birth date within the cerebellar cortex (Altman and Bayer, 1997; Cameron et al., 2009; Leto et al., 2009; Sudarov et al., 2011). Accordingly, a reduction in Ptf1a+ progenitors and Pax2+ immature interneurons as observed in Ptf1acre/+; SmoF/- mutants would have a profound effect on the production of both BCs and SCs. It is, therefore, unexpected to discover that Ptf1acre/+; SmoF/- mutants did not affect BC production. This is not because of a lack of Shh responsiveness of Ptf1a+ progenitors fated to generate BCs as our Ptf1aCreER GIFM study showed that BCs are largely generated from P0 to P3, when the majority of Shh responsive Ptf1a+ progenitors are present. Importantly, Gli1CreER GIFM studies have shown that both MLI subtypes are generated from Shh-responsive progenitors (Fleming et al., 2013). The fact that the increased Pax2+ cell number observed in Ptf1acre/+;SmoM2 mutants did not affect BC pool size further suggests that the commitment to BC and SC fates is not directly related to their birth order but appears to occur at laminar positions in an inside-out sequence. Accordingly, Pax2+ cells that initially populate the inner ML acquire BC identity, whereas those remaining in the outer ML assume SC fate. This model is further supported by our genetic ablation experiment in which the reduction of Ptf1a+ progenitors at the peak of BC production did not affect BC numbers despite a drastically diminished number of Pax2+ cells. We have ruled out the possibilities that the lack of BC perturbation is because of a delay in the progenitor cell depletion or timing of BC to SC production. Thus, changes in Pax2+ pool sizes as observed in Smo gain- and loss-of-function mutants will have the greatest impact on SC numbers. Our model is consistent with the plastic nature of cerebellar immature interneurons observed from heterochronic transplantation studies (Leto et al., 2009) in which P7 interneurons grafted to P1 cerebella exclusively adopt basket cell fate. Our model also explains why the preferential reduction of SCs in the ML is observed in Ccnd2 and Ascl1 mutants despite an early loss of interneuron progenitors (Huard et al., 1999; Sudarov et al., 2011).

Recent studies have shown that immature GCs play an instructive role in differentiation of SCs at the ML (Cadilhac et al., 2021). At present, Ret is the only marker that is selectively expressed in the BCs. Thus, understanding how its expression is activated may provide insight into how laminar positional information affects the identity of BCs. Ret has been studied extensively in the context of kidney and enteric nervous system development (Lake and Heuckeroth, 2013; Costantini, 2016). Its expression in the ureteric bud and enteric neural crest-derived precursor cells is activated by retinoic acid (RA) signaling (Mendelsohn et al., 1999; Simkin et al., 2013). However, analysis of RA signaling using RA responsive reporter mice indicated that it is only activated in a subset of PCs, not in the BCs (Extended Data Fig. 6-3). Another possibility is that identity of BCs is specified through an activity-dependent mechanism as proposed for the cortical interneuron specification (Wamsley and Fishell, 2017). There is evidence that the patterning of BC pinceau is shaped by PC activity (Zhou et al., 2020). In this context, Ret expression may be activated when the pinceau is established between its axon collaterals and the PC axon initial segment. Future studies are required to determine how positional information influences BC identity.

Previous studies have shown that the reduction of MLIs or the loss of their postsynaptic receptors γ2 GABA-A resulted in impaired motor learning (Wulff et al., 2009; ten Brinke et al., 2015; Sergaki et al., 2017). However, it is unclear to what extent each MLI subtype contributes to motor learning. There are critical anatomic and functional differences in how SCs and BCs provide feedforward inhibition of PCs in response to parallel fiber (PF) and climbing fiber (CF) activation. SCs primarily transmit chemical inhibition through axodendritic synapses, whereas BCs deliver both chemical and electrical field inhibition through axosomatic and axoaxonic synapses, respectively (Palay and Chan-Palay, 1974; Korn and Axelrad, 1980; Blot and Barbour, 2014). Tamoxifen-mediated temporal deletions of the GABAergic transporter have suggested that BCs and SCs may have a distinct function in regulating PC simple spike firing pattern and rate (Brown et al., 2019). Our analysis of Shh-signaling mutants provides evidence that SCs contribute significantly to motor learning. We found that the 21% reduction of SC numbers had a profound consequence on the cerebellar circuitry, resulting in a nearly 31% reduction in dendritic inhibitory synapses and significant impairment of inhibitory neurotransmission. This impairment in dendritic synapses and GABAergic input onto PCs is sufficient to cause motor learning deficit. It remains to be determined whether axosomatic and axoaxonic synapses contribute to motor learning.

Footnotes

  • This work was supported by the Vanderbilt–Ingram Cancer Center Support Grant P30 CA068485 and the National Institute of Neurological Disorders and Stroke Grant NS 097898 to C.C. We thank Jingqiong Kang, PhD, and Chun-Qing Zhang, MD, for initial help with electrophysiology experiments and Christopher Wright, DPhil, for the gift of antibody against Ptf1a. We also thank Xi Huang, PhD for insightful feedback and helpful comments.

  • The authors declare no competing financial interests.

  • Correspondence should be addressed to Chin Chiang at chin.chiang{at}vanderbilt.edu

References

  1. ↵
    2. Altman J,
    3. Bayer SA
    (1997) Development of the cerebellar system in relation to its evolution, structure, and function. Boca Raton, FL: CRC.
  2. ↵
    2. Asada H,
    3. Kawamura Y,
    4. Maruyama K,
    5. Kume H,
    6. Ding R,
    7. Ji FY,
    8. Kanbara N,
    9. Kuzume H,
    10. Sanbo M,
    11. Yagi T,
    12. Obata K
    (1996) Mice lacking the 65 kDa isoform of glutamic acid decarboxylase (GAD65) maintain normal levels of GAD67 and GABA in their brains but are susceptible to seizures. Biochem Biophys Res Commun 229:891–895.
    OpenUrlCrossRefPubMed
  3. ↵
    2. Bai CB,
    3. Auerbach W,
    4. Lee JS,
    5. Stephen D,
    6. Joyner AL
    (2002) Gli2, but not Gli1, is required for initial Shh signaling and ectopic activation of the Shh pathway. Development 129:4753–4761.
    OpenUrlCrossRefPubMed
  4. ↵
    2. Bai CB,
    3. Stephen D,
    4. Joyner AL
    (2004) All mouse ventral spinal cord patterning by hedgehog is Gli dependent and involves an activator function of Gli3. Dev Cell 6:103–115.
    OpenUrlCrossRefPubMed
  5. ↵
    2. Beckinghausen J,
    3. Sillitoe RV
    (2018) Insights into cerebellar development and connectivity. Neurosci Lett 688:2–13.
    OpenUrl
  6. ↵
    2. Blot A,
    3. Barbour B
    (2014) Ultra-rapid axon-axon ephaptic inhibition of cerebellar Purkinje cells by the pinceau. Nat Neurosci 17:289–295.
    OpenUrlCrossRefPubMed
  7. ↵
    2. Brown AM,
    3. Arancillo M,
    4. Lin T,
    5. Catt DR,
    6. Zhou J,
    7. Lackey EP,
    8. Stay TL,
    9. Zuo ZW,
    10. White JJ,
    11. Sillitoe RV
    (2019) Molecular layer interneurons shape the spike activity of cerebellar Purkinje cells. Sci Rep 9:1742.
    OpenUrlCrossRefPubMed
  8. ↵
    2. Buch T,
    3. Heppner FL,
    4. Tertilt C,
    5. Heinen TJ,
    6. Kremer M,
    7. Wunderlich FT,
    8. Jung S,
    9. Waisman A
    (2005) A Cre-inducible diphtheria toxin receptor mediates cell lineage ablation after toxin administration. Nat Methods 2:419–426. doi:10.1038/nmeth762 pmid:15908920
    OpenUrlCrossRefPubMed
  9. ↵
    2. Cadilhac C,
    3. Bachy I,
    4. Forget A,
    5. Hodson DJ,
    6. Jahannault-Talignani C,
    7. Furley AJ,
    8. Ayrault O,
    9. Mollard P,
    10. Sotelo C,
    11. Ango F
    (2021) Excitatory granule neuron precursors orchestrate laminar localization and differentiation of cerebellar inhibitory interneuron subtypes. Cell Rep 34:108904. doi:10.1016/j.celrep.2021.108904 pmid:33789110
    OpenUrlCrossRefPubMed
  10. ↵
    2. Cameron DB,
    3. Kasai K,
    4. Jiang Y,
    5. Hu T,
    6. Saeki Y,
    7. Komuro H
    (2009) Four distinct phases of basket/stellate cell migration after entering their final destination (the molecular layer) in the developing cerebellum. Dev Biol 332:309–324. doi:10.1016/j.ydbio.2009.05.575 pmid:19500566
    OpenUrlCrossRefPubMed
  11. ↵
    2. Costantini F
    (2016) Kidney development, disease, repair and regeneration. Sect Dev :41–56.
  12. ↵
    2. Dahmane N,
    3. Altaba A R-i
    (1999) Sonic hedgehog regulates the growth and patterning of the cerebellum. Development 126:3089–3100. doi:10.1242/dev.126.14.3089 pmid:10375501
    OpenUrlAbstract
  13. ↵
    2. Fleming JT,
    3. He W,
    4. Hao C,
    5. Ketova T,
    6. Pan FC,
    7. Wright CCV,
    8. Litingtung Y,
    9. Chiang C
    (2013) The Purkinje neuron acts as a central regulator of spatially and functionally distinct cerebellar precursors. Dev Cell 27:278–292. doi:10.1016/j.devcel.2013.10.008 pmid:24229643
    OpenUrlCrossRefPubMed
  14. ↵
    2. Fujitani Y,
    3. Fujitani S,
    4. Luo H,
    5. Qiu F,
    6. Burlison J,
    7. Long Q,
    8. Kawaguchi Y,
    9. Edlund H,
    10. MacDonald RJ,
    11. Furukawa T,
    12. Fujikado T,
    13. Magnuson MA,
    14. Xiang M,
    15. Wright CV
    (2006) Ptf1a determines horizontal and amacrine cell fates during mouse retinal development. Development 133:4439–4450. doi:10.1242/dev.02598 pmid:17075007
    OpenUrlAbstract/FREE Full Text
  15. ↵
    2. Glasgow SM,
    3. Henke RM,
    4. Macdonald RJ,
    5. Wright CV,
    6. Johnson JE
    (2005) Ptf1a determines GABAergic over glutamatergic neuronal cell fate in the spinal cord dorsal horn. Development 132:5461–5469. doi:10.1242/dev.02167 pmid:16291784
    OpenUrlAbstract/FREE Full Text
  16. ↵
    2. Heiney SA,
    3. Wohl MP,
    4. Chettih SN,
    5. Ruffolo LI,
    6. Medina JF
    (2014) Cerebellar-dependent expression of motor learning during eyeblink conditioning in head-fixed mice. J Neurosci 34:14845–14853. doi:10.1523/JNEUROSCI.2820-14.2014 pmid:25378152
    OpenUrlAbstract/FREE Full Text
  17. ↵
    2. Hoshino M,
    3. Nakamura S,
    4. Mori K,
    5. Kawauchi T,
    6. Terao M,
    7. Nishimura YV,
    8. Fukuda A,
    9. Fuse T,
    10. Matsuo N,
    11. Sone M,
    12. Watanabe M,
    13. Bito H,
    14. Terashima T,
    15. Wright CV,
    16. Kawaguchi Y,
    17. Nakao K,
    18. Nabeshima Y
    (2005) Ptf1a, a bHLH transcriptional gene, defines GABAergic neuronal fates in cerebellum. Neuron 47:201–213. doi:10.1016/j.neuron.2005.06.007 pmid:16039563
    OpenUrlCrossRefPubMed
  18. ↵
    2. Huang X,
    3. Ketova T,
    4. Fleming JT,
    5. Wang H,
    6. Dey SK,
    7. Litingtung Y,
    8. Chiang C
    (2009) Sonic hedgehog signaling regulates a novel epithelial progenitor domain of the hindbrain choroid plexus. Development 136:2535–2543. doi:10.1242/dev.033795 pmid:19570847
    OpenUrlAbstract/FREE Full Text
  19. ↵
    2. Huard JM,
    3. Forster CC,
    4. Carter ML,
    5. Sicinski P,
    6. Ross ME
    (1999) Cerebellar histogenesis is disturbed in mice lacking cyclin D2. Development 126:1927–1935. doi:10.1242/dev.126.9.1927 pmid:10101126
    OpenUrlAbstract/FREE Full Text
  20. ↵
    2. Jain S,
    3. Golden JP,
    4. Wozniak D,
    5. Pehek E,
    6. Johnson EM,
    7. Milbrandt J
    (2006) RET is dispensable for maintenance of midbrain dopaminergic neurons in adult mice. J Neurosci 26:11230–11238. doi:10.1523/JNEUROSCI.1876-06.2006 pmid:17065462
    OpenUrlAbstract/FREE Full Text
  21. ↵
    2. Jeong J,
    3. Mao J,
    4. Tenzen T,
    5. Kottmann AH,
    6. McMahon AP
    (2004) Hedgehog signaling in the neural crest cells regulates the patterning and growth of facial primordia. Genes Dev 18:937–951. doi:10.1101/gad.1190304 pmid:15107405
    OpenUrlAbstract/FREE Full Text
  22. ↵
    2. Kash SF,
    3. Johnson RS,
    4. Tecott LH,
    5. Noebels JL,
    6. Mayfield RD,
    7. Hanahan D,
    8. Baekkeskov S
    (1997) Epilepsy in mice deficient in the 65-kDa isoform of glutamic acid decarboxylase. Proc Natl Acad Sci USA 94:14060–14065. doi:10.1073/pnas.94.25.14060
    OpenUrlAbstract/FREE Full Text
  23. ↵
    2. Kawaguchi Y,
    3. Cooper B,
    4. Gannon M,
    5. Ray M,
    6. MacDonald RJ,
    7. Wright CVE
    (2002) The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors. Nat Genet 32:128–134. doi:10.1038/ng959 pmid:12185368
    OpenUrlCrossRefPubMed
  24. ↵
    2. Kopinke D,
    3. Brailsford M,
    4. Pan FC,
    5. Magnuson MA,
    6. Wright CVE,
    7. Murtaugh LC
    (2012) Ongoing notch signaling maintains phenotypic fidelity in the adult exocrine pancreas. Dev Biol 362:57–64. doi:10.1016/j.ydbio.2011.11.010 pmid:22146645
    OpenUrlCrossRefPubMed
  25. ↵
    2. Korn H,
    3. Axelrad H
    (1980) Electrical inhibition of Purkinje cells in the cerebellum of the rat. Proc Natl Acad Sci U S A 77:6244–6247. doi:10.1073/pnas.77.10.6244 pmid:6255484
    OpenUrlAbstract/FREE Full Text
  26. ↵
    2. Lake JI,
    3. Heuckeroth RO
    (2013) Enteric nervous system development: migration, differentiation, and disease. Am J Gastrointest Liver Physiol 305:G1–G24.
    OpenUrlCrossRefPubMed
  27. ↵
    2. Leto K,
    3. Carletti B,
    4. Williams IM,
    5. Magrassi L,
    6. Rossi F
    (2006) Different types of cerebellar GABAergic interneurons originate from a common pool of multipotent progenitor cells. J Neurosci 26:11682–11694. doi:10.1523/JNEUROSCI.3656-06.2006 pmid:17093090
    OpenUrlAbstract/FREE Full Text
  28. ↵
    2. Leto K,
    3. Bartolini A,
    4. Yanagawa Y,
    5. Obata K,
    6. Magrassi L,
    7. Schilling K,
    8. Rossi F
    (2009) Laminar fate and phenotype specification of cerebellar GABAergic interneurons. J Neurosci 29:7079–7091. doi:10.1523/JNEUROSCI.0957-09.2009 pmid:19474334
    OpenUrlAbstract/FREE Full Text
  29. ↵
    2. Long F,
    3. Zhang XM,
    4. Karp S,
    5. Yang Y,
    6. McMahon AP
    (2001) Genetic manipulation of hedgehog signaling in the endochondral skeleton reveals a direct role in the regulation of chondrocyte proliferation. Development 128:5099–5108. doi:10.1242/dev.128.24.5099 pmid:11748145
    OpenUrlAbstract/FREE Full Text
  30. ↵
    2. Machold R,
    3. Fishell G
    (2005) Math1 is expressed in temporally discrete pools of cerebellar rhombic-lip neural progenitors. Neuron 48:17–24. doi:10.1016/j.neuron.2005.08.028 pmid:16202705
    OpenUrlCrossRefPubMed
  31. ↵
    2. Madisen L,
    3. Zwingman TA,
    4. Sunkin SM,
    5. Oh SW,
    6. Zariwala HA,
    7. Gu H,
    8. Ng LL,
    9. Palmiter RD,
    10. Hawrylycz MJ,
    11. Jones AR,
    12. Lein ES,
    13. Zeng H
    (2010) A robust and high-throughput Cre reporting and characterization system for the whole mouse brain. Nat Neurosci 13:133–140. doi:10.1038/nn.2467 pmid:20023653
    OpenUrlCrossRefPubMed
  32. ↵
    2. Mao J,
    3. Ligon KL,
    4. Rakhlin EY,
    5. Thayer SP,
    6. Bronson RT,
    7. Rowitch D,
    8. McMahon AP
    (2006) A novel somatic mouse model to survey tumorigenic potential applied to the Hedgehog pathway. Cancer Res 66:10171–10178. doi:10.1158/0008-5472.CAN-06-0657 pmid:17047082
    OpenUrlAbstract/FREE Full Text
  33. ↵
    2. Medina JF,
    3. Nores WL,
    4. Ohyama T,
    5. Mauk MD
    (2000) Mechanisms of cerebellar learning suggested by eyelid conditioning. Curr Opin Neurobiol 10:717–724. doi:10.1016/s0959-4388(00)00154-9 pmid:11240280
    OpenUrlCrossRefPubMed
  34. ↵
    2. Mendelsohn C,
    3. Batourina E,
    4. Fung S,
    5. Gilbert T,
    6. Dodd J
    (1999) Stromal cells mediate retinoid-dependent functions essential for renal development. Development 126:1139–1148. doi:10.1242/dev.126.6.1139 pmid:10021334
    OpenUrlAbstract
  35. ↵
    2. Orchard P,
    3. White JS,
    4. Thomas PE,
    5. Mychalowych A,
    6. Kiseleva A,
    7. Hensley J,
    8. Allen B,
    9. Parker SCJ,
    10. Keegan CE
    (2019) Genome-wide chromatin accessibility and transcriptome profiling show minimal epigenome changes and coordinated transcriptional dysregulation of hedgehog signaling in Danforth's short tail mice. Hum Mol Genet 28:736–750. doi:10.1093/hmg/ddy378 pmid:30380057
    OpenUrlCrossRefPubMed
  36. ↵
    2. Palay SL,
    3. Chan-Palay V
    (1974) Cerebellar cortex, cytology and organization. Berlin: Springer.
  37. ↵
    2. Parmigiani E,
    3. Leto K,
    4. Rolando C,
    5. Figueres-Oñate M,
    6. López-Mascaraque L,
    7. Buffo A,
    8. Rossi F
    (2015) Heterogeneity and bipotency of astroglial-like cerebellar progenitors along the interneuron and glial lineages. J Neurosci 35:7388–7402. doi:10.1523/JNEUROSCI.5255-14.2015 pmid:25972168
    OpenUrlAbstract/FREE Full Text
  38. ↵
    2. Pascual M,
    3. Abasolo I,
    4. Meur AM-L,
    5. Martinez A,
    6. Rio JAD,
    7. Wright CV,
    8. Real FX,
    9. Soriano E
    (2007) Cerebellar GABAergic progenitors adopt an external granule cell-like phenotype in the absence of Ptf1a transcription factor expression. Proc Natl Acad Sci U S A 104:5193–5198 doi:10.1073/pnas.0605699104 pmid:17360405
    OpenUrlAbstract/FREE Full Text
  39. ↵
    2. Pieper A,
    3. Rudolph S,
    4. Wieser GL,
    5. Götze T,
    6. Mießner H,
    7. Yonemasu T,
    8. Yan K,
    9. Tzvetanova I,
    10. Castillo BD,
    11. Bode U,
    12. Bormuth I,
    13. Wadiche JI,
    14. Schwab MH,
    15. Goebbels S
    (2019) NeuroD2 controls inhibitory circuit formation in the molecular layer of the cerebellum. Sci Rep 9:1448. doi:10.1038/s41598-018-37850-7 pmid:30723302
    OpenUrlCrossRefPubMed
  40. ↵
    2. Raymond JL,
    3. Medina JF
    (2018) Computational principles of supervised learning in the cerebellum. Annu Rev Neurosci 41:233–253. doi:10.1146/annurev-neuro-080317-061948 pmid:29986160
    OpenUrlCrossRefPubMed
  41. ↵
    2. Sears LL,
    3. Finn PR,
    4. Steinmetz JE
    (1994) Abnormal classical eye-blink conditioning in autism. J Autism Dev Disord 24:737–751. doi:10.1007/BF02172283 pmid:7844097
    OpenUrlCrossRefPubMed
  42. ↵
    2. Sergaki MC,
    3. López-Ramos JC,
    4. Stagkourakis S,
    5. Gruart A,
    6. Broberger C,
    7. Delgado-García JM,
    8. Ibáñez CF
    (2017) Compromised survival of cerebellar molecular layer interneurons lacking GDNF receptors GFRα1 or RET impairs normal cerebellar motor learning. Cell Rep 19:1977–1986. doi:10.1016/j.celrep.2017.05.030 pmid:28591570
    OpenUrlCrossRefPubMed
  43. ↵
    2. Silbereis J,
    3. Cheng E,
    4. Ganat YM,
    5. Ment LR,
    6. Vaccarino FM
    (2009) Precursors with glial fibrillary acidic protein promoter activity transiently generate GABA interneurons in the postnatal cerebellum. Stem Cells 27:1152–1163. doi:10.1002/stem.18 pmid:19418461
    OpenUrlCrossRefPubMed
  44. ↵
    2. Simkin JE,
    3. Zhang D,
    4. Rollo BN,
    5. Newgreen DF
    (2013) Retinoic acid upregulates ret and induces chain migration and population expansion in vagal neural crest cells to colonise the embryonic gut. Plos One 8:e64077. doi:10.1371/journal.pone.0064077 pmid:23717535
    OpenUrlCrossRefPubMed
  45. ↵
    2. Somogyi P,
    3. Hámori J
    (1976) A quantitative electron microscopic study of the purkinje cell axon initial segment. Neuroscience 1:361–365. doi:10.1016/0306-4522(76)90127-5 pmid:1004711
    OpenUrlCrossRefPubMed
  46. ↵
    2. Sotelo C
    (2015) Molecular layer interneurons of the cerebellum: developmental and morphological aspects. Cerebellum 14:534–556. doi:10.1007/s12311-015-0648-x pmid:25599913
    OpenUrlCrossRefPubMed
  47. ↵
    2. Sudarov A,
    3. Turnbull RK,
    4. Kim EJ,
    5. Lebel-Potter M,
    6. Guillemot F,
    7. Joyner AL
    (2011) Ascl1 genetics reveals insights into cerebellum local circuit assembly. J Neurosci 31:11055–11069. doi:10.1523/JNEUROSCI.0479-11.2011 pmid:21795554
    OpenUrlAbstract/FREE Full Text
  48. ↵
    2. ten Brinke MM,
    3. Boele H-J,
    4. Spanke JK,
    5. Potters J-W,
    6. Kornysheva K,
    7. Wulff P,
    8. IJpelaar ACHG,
    9. Koekkoek SKE,
    10. Zeeuw CID
    (2015) Evolving models of pavlovian conditioning: cerebellar cortical dynamics in awake behaving mice. Cell Rep 13:1977–1988. doi:10.1016/j.celrep.2015.10.057 pmid:26655909
    OpenUrlCrossRefPubMed
  49. ↵
    2. Wallace VA
    (1999) Purkinje-cell-derived Sonic hedgehog regulates granule neuron precursor cell proliferation in the developing mouse cerebellum. Curr Biol 9:445–448. doi:10.1016/s0960-9822(99)80195-x pmid:10226030
    OpenUrlCrossRefPubMed
  50. ↵
    2. Wamsley B,
    3. Fishell G
    (2017) Genetic and activity-dependent mechanisms underlying interneuron diversity. Nat Rev Neurosci 18:299–309. doi:10.1038/nrn.2017.30 pmid:28381833
    OpenUrlCrossRefPubMed
  51. ↵
    2. Wang VY,
    3. Rose MF,
    4. Zoghbi HY
    (2005) Math1 expression redefines the rhombic lip derivatives and reveals novel lineages within the brainstem and cerebellum. Neuron 48:31–43. doi:10.1016/j.neuron.2005.08.024 pmid:16202707
    OpenUrlCrossRefPubMed
  52. ↵
    2. Wechsler-Reya RJ,
    3. Scott MP
    (1999) Control of neuronal precursor proliferation in the cerebellum by Sonic Hedgehog. Neuron 22:103–114. doi:10.1016/s0896-6273(00)80682-0 pmid:10027293
    OpenUrlCrossRefPubMed
  53. ↵
    2. Wefers AK,
    3. Haberlandt C,
    4. Surchev L,
    5. Steinhäuser C,
    6. Jabs R,
    7. Schilling K
    (2018) Migration of interneuron precursors in the nascent cerebellar cortex. Cerebellum 17:62–71. doi:10.1007/s12311-017-0900-7 pmid:29149443
    OpenUrlCrossRefPubMed
  54. ↵
    2. Weisheit G,
    3. Gliem M,
    4. Endl E,
    5. Pfeffer PL,
    6. Busslinger M,
    7. Schilling K
    (2006) Postnatal development of the murine cerebellar cortex: formation and early dispersal of basket, stellate and Golgi neurons. Eur J Neurosci 24:466–478. doi:10.1111/j.1460-9568.2006.04915.x pmid:16903854
    OpenUrlCrossRefPubMed
  55. ↵
    2. Wulff P,
    3. Schonewille M,
    4. Renzi M,
    5. Viltono L,
    6. Sassoè-Pognetto M,
    7. Badura A,
    8. Gao Z,
    9. Hoebeek FE,
    10. van Dorp S,
    11. Wisden W,
    12. Farrant M,
    13. De Zeeuw CI
    (2009) Synaptic inhibition of Purkinje cells mediates consolidation of vestibulo-cerebellar motor learning. Nat Neurosci 12:1042–1049. doi:10.1038/nn.2348 pmid:19578381
    OpenUrlCrossRefPubMed
  56. ↵
    2. Yamanaka H,
    3. Yanagawa Y,
    4. Obata K
    (2004) Development of stellate and basket cells and their apoptosis in mouse cerebellar cortex. Neurosci Res 50:13–22. doi:10.1016/j.neures.2004.06.008 pmid:15288494
    OpenUrlCrossRefPubMed
  57. ↵
    2. Zhang L,
    3. Goldman JE
    (1996) Generation of cerebellar interneurons from dividing progenitors in white matter. Neuron 16:47–54. doi:10.1016/s0896-6273(00)80022-7 pmid:8562089
    OpenUrlCrossRefPubMed
  58. ↵
    2. Zhou J,
    3. Brown AM,
    4. Lackey EP,
    5. Arancillo M,
    6. Lin T,
    7. Sillitoe RV
    (2020) Purkinje cell neurotransmission patterns cerebellar basket cells into zonal modules defined by distinct pinceau sizes. Elife 9:e55569. doi:10.7554/eLife.55569
    OpenUrlCrossRef
Back to top

In this issue

Email
Print
View Full Page PDF
Citation Tools
Respond to this article
Dendritic Inhibition by Shh Signaling-Dependent Stellate Cell Pool Is Critical for Motor Learning
Wen Li, Lei Chen, Jonathan T. Fleming, Emily Brignola, Kirill Zavalin, Andre Lagrange, Tonia Rex, Shane A. Heiney, Gregory J. Wojaczynski, Javier F. Medina, Chin Chiang
Journal of Neuroscience 29 June 2022, 42 (26) 5130-5143; DOI: 10.1523/JNEUROSCI.2073-21.2022
Reddit logo Twitter logo Facebook logo Mendeley logo

Jump to section

Keywords

Responses to this article

Respond to this article

Jump to comment:

No eLetters have been published for this article.