Article

Early molecular layer interneuron hyperactivity

triggers Purkinje neuron degeneration in SCA1
Highlights Authors
d Increased molecular layer interneuron (MLIN) activity drives Federica Pilotto,
SCA1 pathophysiology Christopher Douthwaite, Rim Diab, ...,
Benoı̂t Zuber, Sabine Liebscher,
d Inhibition of MLIN ameliorates network and behavioral SCA1 Smita Saxena
phenotypes

d Mimicking MLIN hyperexcitability in healthy mice generates

Correspondence
SCA1-like pathophysiology [email protected]
(S.L.),
d MLIN hyperexcitability is conserved in human SCA1 patient- [email protected] (S.S.)
derived GABAergic neurons
In brief
Pilotto et al. identify early
hyperexcitability of molecular layer
interneurons (MLINs) in SCA1 driving
circuit dysfunction and motor deficits.
Inhibition of MLINs in SCA1 mice induced
lasting improvements of circuit pathology
and behavior. MLIN proteomics revealed
a signature accounting for
hyperexcitability, conserved in patient-
derived GABAergic neurons.

Pilotto et al., 2023, Neuron 111, 2523–2543

August 16, 2023 ª 2023 The Author(s). Published by Elsevier Inc.
https://doi.org/10.1016/j.neuron.2023.05.016 ll
 ll
OPEN ACCESS

Article
Early molecular layer interneuron hyperactivity
triggers Purkinje neuron degeneration in SCA1
Federica Pilotto,1,2,10 Christopher Douthwaite,3,10 Rim Diab,1,2 XiaoQian Ye,3 Zahraa Al qassab,1,2 Christoph Tietje,3
Meriem Mounassir,3 Adolfo Odriozola,4 Aishwarya Thapa,1,2 Ronald A.M. Buijsen,7 Sophie Lagache,8
Anne-Christine Uldry,8 Manfred Heller,8 Stefan Mu € ller,9 Willeke M.C. van Roon-Mom,7 Benoı̂t Zuber,4
Sabine Liebscher,3,5,6,* and Smita Saxena1,2,11,*
1Department of Neurology, Inselspital University Hospital, Bern, Switzerland
2Department for BioMedical Research, University of Bern, Bern, Switzerland
3Institute of Clinical Neuroimmunology, Klinikum der Universita €t Mu
€nchen, Ludwig-Maximilians University Munich, Martinsried, Germany
4Institute of Anatomy, University of Bern, Bern, Switzerland
5Munich Cluster for Systems Neurology (SyNergy), Munich, Germany
6University Hospital Cologne, Deptartment of Neurology, Cologne, Germany
7Department of Human Genetics, Leiden University Medical Center, Leiden, the Netherlands
8Proteomics and Mass Spectrometry Core Facility, Department for BioMedical Research, University of Bern, Bern, Switzerland
9Flow Cytometry and Cell sorting, Department for BioMedical Research, University of Bern, Bern, Switzerland
10These authors contributed equally
11Lead contact

*Correspondence: [email protected] (S.L.), [email protected] (S.S.)

https://doi.org/10.1016/j.neuron.2023.05.016

SUMMARY

Toxic proteinaceous deposits and alterations in excitability and activity levels characterize vulnerable
neuronal populations in neurodegenerative diseases. Using in vivo two-photon imaging in behaving spino-
cerebellar ataxia type 1 (Sca1) mice, wherein Purkinje neurons (PNs) degenerate, we identify an inhibitory cir-
cuit element (molecular layer interneurons [MLINs]) that becomes prematurely hyperexcitable, compro-
mising sensorimotor signals in the cerebellum at early stages. Mutant MLINs express abnormally elevated
parvalbumin, harbor high excitatory-to-inhibitory synaptic density, and display more numerous synaptic
connections on PNs, indicating an excitation/inhibition imbalance. Chemogenetic inhibition of hyperexcit-
able MLINs normalizes parvalbumin expression and restores calcium signaling in Sca1 PNs. Chronic inhibi-
tion of mutant MLINs delayed PN degeneration, reduced pathology, and ameliorated motor deficits in Sca1
mice. Conserved proteomic signature of Sca1 MLINs, shared with human SCA1 interneurons, involved the
higher expression of FRRS1L, implicated in AMPA receptor trafficking. We thus propose that circuit-level def-
icits upstream of PNs are one of the main disease triggers in SCA1.

INTRODUCTION protein Ataxin-1, leading to the premature degeneration of cere-

bellar PNs.8 One critical open question is why cerebellar PNs
The mechanisms underlying the major neurodegenerative dis- selectively degenerate in SCA1. In most NDs, multiple cell types
eases (NDs) are still poorly understood. An enigmatic but uniform within neuronal circuits throughout the CNS are affected, and
finding in patients and murine models of NDs is the early alter- crucial synaptic and network dysfunction might occur before
ation in excitability of vulnerable neuronal populations and the appearance of overt pathology.9–11 Moreover, it has been
changes in the corresponding neuronal circuitry.1,2 For instance, increasingly acknowledged that non-cell autonomous impair-
in spinal muscular atrophy, motoneurons (MNs) become hyper- ments, such as early circuit-mediated alterations, might critically
excitable,3 whereas in amyotrophic lateral sclerosis, spinal govern or even trigger the pathology in disease-vulnerable neu-
MNs become hypoexcitable, losing their ability to fire repeti- rons.12 Previous studies by us and others highlighted the impor-
tively.4,5 Similarly, in spinocerebellar ataxia type-1 (SCA1) and tance of early circuit-related changes within the cerebellum that
type-2 (SCA2), Purkinje neurons (PNs) show a reduction in the affect PN activity in SCA1.6,13,14 Mainly, excitatory synaptic in-
firing rate at presymptomatic stages.6,7 The precise cause of puts onto PNs from climbing fibers (CFs), but not from parallel fi-
these alterations and whether they are involved in the degenera- bers (PFs), are altered presymptomatically, leading to reduced
tive process remains unknown. CF-PN synaptic strength and subsequent alterations in PN excit-
SCA1 is a devastating, incurable ND due to the expansion of a ability. This suggests the existence of early, selectively vulnerable
polyglutamine (PolyQ) repeat within the ubiquitously expressed circuit components, upstream of the affected PNs.6,14–16 Most

Neuron 111, 2523–2543, August 16, 2023 ª 2023 The Author(s). Published by Elsevier Inc. 2523
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 ll
OPEN ACCESS Article

(legend on next page)

2524 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

previous studies were, however, conducted in slice cultures, al- photon calcium imaging in behaving mice, running on a spherical
lowing only limited insight into how cerebellar circuit elements treadmill (Figures 1A and S1). By expressing the genetically en-
and entire networks are affected in a spatiotemporal manner in coded calcium indicator GCaMP7s, we simultaneously moni-
animals by these described molecular and synaptic alterations. tored activity of the three main inhibitory neuronal subtypes in
We hypothesized that PolyQ-repeat expansions have a more the cerebellar cortex, namely, MLINs, PNs, and morphologically
widespread impact on cerebellar circuits, thereby altering excit- putative Golgi cells (Gol), while mice were maneuvering in a virtual
atory/inhibitory (E/I) balance, compromising sensorimotor func- environment (VR), with the locomotion speed being represented
tions, and ultimately triggering PN degeneration. Notably, PN activ- by the visual flow of the VR (feedback [fb]) or running in darkness
ity is tightly shaped by GABAergic inhibitory molecular layer (Figures 1A and 1B). These recordings revealed differential im-
interneurons (MLINs).17 Although MLINs do not degenerate in pairments of baseline fluorescence (indicative of alterations in
SCA1,18 they might still contribute to PN degeneration. First resting cytoplasmic calcium concentrations), neuronal activity
in vitro evidence for an involvement of MLINs in SCA1 pathology levels, and response properties of the different neuronal types
was provided by Edamakanti and colleagues, who identified the (Figures 1C and 1D). As the typical instantaneous firing rate of
hyperproliferation of mutant cerebellar stem cells, preferentially these cell types exceeds the temporal resolution of the calcium in-
differentiating into GABAergic interneurons, thus enhancing dicator, but the fluorescence signal still scales with overall
GABAergic synaptic inputs onto PNs.13 However, the longitudinal neuronal firing,19,20 we assessed neuronal activity by computing
molecular, cellular, and circuit-level consequences of such synap- the area under curve of the DF/F trace. We found that MLINs in
tic alterations remain unclear. Sca1 mice had not only a significantly lower baseline fluores-
Here, combining in vivo two-photon calcium imaging in behaving cence (Figure 1E) but were also more active during quiet wakeful-
knockin mice, harboring an expansion of 154 CAG repeats in the ness (QW) during fb and darkness (Figure 1F). Conversely, we did
endogenous Ataxin-1 mouse locus (termed Sca1 throughout) not observe any changes in baseline fluorescence and a decrease
together with chemogenetics, histological, ultrastructural, and in spontaneous neuronal activity levels (i.e., neuronal activity
MLIN-specific proteomic analyses, we demonstrate that neurons occurring during stationary epochs, measured as area under
within the cerebellar network are differentially affected by the PolyQ the curve [AUC]/min) during QW in Gol (Figures 1G and 1H) or
expansion. Early impairments mainly involve alterations in the MLIN PNs (Figures 1I and 1J). Moreover, we found a strong increase
population, rendering them hyperexcitable, thereby affecting their in the response to locomotion (Figures 1K–1M) selectively in
response properties and compromising coding space in mice. MLINs in Sca1 (Figure 1N) and more MLINs being running respon-
Additionally, MLIN-selective proteomics from Sca1 mice revealed sive (Figure 1O). In contrast to MLIN hyperresponsiveness, Gol in
a disease-linked molecular signature, accounting for the observed Sca1 were driven less by locomotion (Figure 1P) and displayed no
changes in mutant MLIN functionality. Notably, this signature was change in the fraction of running responsive neurons (Figure 1Q).
conserved within human SCA1 patient-derived GABAergic neu- At this early disease stage, no alterations in the response to loco-
rons. Our results thus indicate that at the circuit level, molecular motion in PNs was detected (Figures 1R and 1S). However, the
changes within mutant MLINs govern early cerebellar network number of Gol and PNs captured in this study are considerably
dysfunction, promoting PNs degeneration in SCA1. lower than those of MLINs; thus, we may potentially underesti-
mate alterations of these neuronal populations.
RESULTS Recording the same neurons under anesthesia further corrob-
orated these findings (Figures S2A–S2D). We found a strong
Cerebellar neuronal activity levels are differentially reduction in the baseline fluorescence of MLINs in Sca1 mice
altered in early symptomatic Sca1 mice (Figure S2E), whereas baseline fluorescence was neither altered
To assess circuit-level deficits during the early symptomatic in Gol (Figure S2F) nor PNs (Figure S2G). We again observed an
phase in postnatal (P) day 60 Sca1 mice, we performed two- increase in spontaneous neuronal activity of MLINs (Figure S2H)

Figure 1. Simultaneous monitoring of diverse cerebellar neuronal subtypes in behaving mice

(A) Experimental design (fb, feedback; dark, darkness; iso, isoflurane anesthesia).
(B) Scheme, depicting sites of adeno-associated virus (AAV)-injection and imaging and the cerebellar cortex layers.
(C) Average projection of a quiet wakefulness (QW, left) episode and locomotion (loco, right) of the same field of view (FOV) in WT and Sca1 mice (dashed lines-
cerebellar layers).
(D) Calcium traces of molecular layer interneurons (MLINs, magenta), Gol (Gol, yellow), and Purkinje neurons (PNs, gray) during fb and darkness in WT (upper) and
Sca1 mice (lower). Gray areas-locomotion epochs.
(E–J) Baseline fluorescence and spontaneous neuronal activity (area under the curve (AUC)/min) during QW. (E) MLIN baseline fluorescence, (F) MLIN neuronal
activity (WT 1,386, Sca1 714), (G) Gol baseline fluorescence and (H) neuronal activity (WT 114, Sca1 48), and (I) PN baseline fluorescence and (J) spontaneous
neuronal activity (WT 112, Sca1 132).
(K) Heatmap depicting average neuronal activity in response to running onset for all MLINs in WT (left, blue frame) and Sca1 mice (right, red frame).
(L and M) (L) Same as in (K) for Gol and (M) for PN.
(N–S) (N) Average population activity as function of instantaneous running velocity in MLINs (superimposed by sigmoidal fit), (O) fraction running-responsive
MLINs per FOV (WT: 16 FOV [8 mice], Sca1: 9 FOV [6 mice]), (P) as in (N) for Gol, (Q) as in (O) for Gol (WT: 16 FOV [8 mice], Sca1: 6 FOV [4 mice]), (R) as in (N) for PN,
(S) as in (O) for PN (WT: 14 FOV [8 mice], Sca1: 9 FOV [6 mice]). Data in (N), (P), and (R) mean ± SEM. Data presented as box plots throughout the manuscript
display the median and interquartile range and whiskers extend to the minimum and maximum, unless differently stated. Scale bars, (B) 1 mm, (C) 100 mm,
*p < 0.05, **p < 0.01, ***p < 0.001. (ML, molecular layer; PNL, Purkinje neuron layer; GCL, granule cell layer). Statistical test description and F/t/R values are in
STAR Methods. See also Figures S1–S4.

Neuron 111, 2523–2543, August 16, 2023 2525

 ll
OPEN ACCESS Article

but a decrease in Gol activity (Figure S2I) and no change in spon- tion activity in WT and Sca1 (Figure 2C), dominated by
taneous PN activity in P60 Sca1 mice (Figure S2J). Despite lack- locomotion and whisking. However, in Sca1 mice, the variance
ing the temporal resolution to resolve the high instantaneous of PC1 explained by locomotion, air puff, and whisking was
pacemaker firing activities of 40–60 Hz,21 larger fluctuations in reduced compared with WT (Figure 2D). Cell-type-specific
the PN fluorescence traces occur, likely caused by pauses in regression analyses of PC1 revealed a significant reduction of
simple spike firing.22 Those fluctuations, however, did not differ the behavioral parameters for MLINs only, whereas for Gol it
between wild-type (WT) and Sca1 mice. was mainly the contribution of the air-puff-representation that
Because GCaMP only allows for an approximation of calcium was affected (Figure 2D). In PNs, no significant difference in
levels, we performed experiments using the ratiometric calcium the encoding of behavioral parameters was found (Figure 2D).
indicator Twitch2B,23 (Figures S3A–S3C), enabling more accu- We also assessed the information content captured in PC2,
rate measurements of actual calcium concentrations indepen- which still contributed 16% to the overall variance of the pop-
dent of expression levels. This analysis confirmed the reduced ulation activity (WT: 15.95% ± 4%, Sca1: 8.44% ± 2.6% of the
resting calcium concentration of MLINs in anesthetized SCA1 variance). PC2 was also driven by locomotion as a single param-
mice (Figure S3D), and also unraveled reduced levels in Gol, eter but mainly captured all parameters combined, thus, the dif-
but not in PNs at this early symptomatic stage (Figures S3E ference between active and quiescent phases (Figure S5). In
and S3F). We assessed neuronal activity levels and again found contrast to PC1, we could not detect any differences in the vari-
increased activity in MLINs in Sca1 (Figure S3G); however, we ance explained by the individual behavioral parameters between
found unchanged Gol activity (Figure S3H) and a slight, yet sig- WT and Sca1 mice (Figure S5).
nificant, increase in PNs (Figure S3I). As in PNs, we may mainly Neuronal activity patterns were multi-dimensional, consisting
detect fluctuations in the calcium signal due to pauses in firing of at least 10 components. We thus asked how the geometry
and complex spikes. Consequently, reduced pacemaker activity of these neuronal representations might be affected in Sca1
(as described for PNs in Sca124), potentially even appears as mice. We computed the manifolds by projecting the multi-
greater fluctuations in the fluorescent signal. dimensional neuronal population activity to low-dimensional
To probe for probable impairments of sensorimotor integration, space, consisting of the first three components.27–29 In WT
we compared neuronal activity levels during fb with those in dark- mice, two distinct structures emerged, separating QW from
ness, (Figure S4A) and observed a strong linear correlation. How- locomotion (Figure 2E), corroborating previous reports.27
ever, we found Sca1 MLINs displayed a leftward shift in the ratio of Notably, the PC space representing QW was reduced in Sca1
fb/dark activity, indicating alterations in sensorimotor integration mice (Figure 2F), and the two subspaces were less segregated,
(Figure S4A). Conversely, we found Gol to be less active in Sca1 as illustrated by the decreased Euclidean distance (Figure 2F).
mice during fb and darkness (Figure S4B), whereas Sca1 PNs Together these data highlight a compromised encoding capacity
were slightly more active during fb (Figure S4C). We also probed of the cerebellar circuit in early symptomatic Sca1 mice, an effect
sensorimotor integration by assessing neuronal responses to an primarily driven by MLIN dysfunction.
air puff, driving sensory, arousal, and startle-response signals25 To assess whether these deviations in neuronal function
(Figures S4D–S4F). We observed a significant increase in neuronal persist throughout the disease course, we performed similar ex-
activity in mutant MLINs (Figure S4G), a decreased response in Gol periments in P200 Sca1 mice (Figures S6A and S6B). Under anes-
(Figure S4H), and an increase in PNs (Figure S4I). Together, these thesia, we observed a small but significant decrease in baseline
data show that neuronal dysfunction in Sca1 mice is based on a fluorescence in MLINs (Figure S6C), no change in Gol (Fig-
combination of altered spontaneous activity and differentially ure S6D), and a pronounced increase in baseline fluorescence
impaired responsiveness to sensorimotor signals. Importantly, in PNs (Figure S6E), indicating a severe increase in cytoplasmic
the cell type showing the strongest change in Sca1 mice are calcium levels in PNs at late disease stages. Neuronal activity
MLINs, displaying increased activity levels, arguing for MLINs be- of MLINs under anesthesia was still significantly increased (Fig-
ing hyperexcitable already in early symptomatic Sca1 mice. ure S6F), which was also true for Gol (Figure S6G), whereas in
PNs, we found reduced calcium fluctuations (Figure S6H). These
Neuronal coding space is strongly reduced in Sca1 mice findings were in line with the increased MLIN activity seen in
To gain insight into cerebellar network dysfunction in early symp- awake mice during QW (Figure S6I). However, in awake mice,
tomatic Sca1 mice, we investigated neural network representa- we still observed a decrease in Gol activity (Figure S6J), which
tions of behavioral states and their distinctiveness. Several could reflect the higher inhibitory tone during wakefulness, i.e.,
neuronal populations in the cerebellar cortex encode various abolished under anesthesia, thereby potentially unmasking
sensorimotor signals in a multidimensional manner.25–27 We changes in intrinsic excitability. In PNs, we again observed
chose a principal-component analysis (PCA)-based approach decreased calcium fluctuations (Figure S6K). We quantified the
to investigate the dimensionality of the population dynamics in response to locomotion and found a strong hyperresponsiveness
behaving mice. We assembled an input matrix, consisting of of MLINs to running (Figure S6L), a hyporesponsiveness in Gol
0-centered, Z scored DF/F traces of all cells identified in a given (Figure S6M), and a non-significant trend toward compromised
field of view (FOV) and aligned it with the recordings of running locomotion responses in PNs (Figure S6N). Altogether, these
speed, whisking, pupil width, and applied air puffs (Figure 2A). data reveal that the observed hyperexcitability of MLINs remains
Analyses of the variance explained of the population activity re- conserved throughout the disease course in Sca1 mice.
vealed a multidimensional nature (Figure 2B). The first compo- Given previous reports indicating hyperexcitability in PN
nent (PC1) alone explained 50% of the variance of the popula- dendrites as a critical contributor to PN degeneration,30 we

2526 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

Figure 2. Compromised coding space of behavioral parameters in cerebellar cortex of Sca1 mice
(A) Exemplary input matrix consisting of 0-centered and Z scored DF/F traces used for dimensionality reduction derived from all cells in one FOV of a feedback
session of a WT mouse aligned with behavioral parameters (lower traces, magenta, MLIN; gray, PN; yellow, Gol).
(B) PC1, PC2, and PC3 time-series derived from input matrix shown in (A).
(C) Variance explained of the neuronal activity as a function of number of principal components in WT (blue, thin lines, individual FOV; thick line, mean) and Sca1
mice (red). PC1 explains 50% of the variance in both WT and Sca1 (WT 51.9% ± 8.6%, 16 FOV, 8 mice; Sca1 54.63% ± 11.6%, 14 FOV, 7 mice).
(D) Variance of PC1 (derived from all cells in a FOV or individual cell types), explained by the respective behavioral parameters in WT and Sca1.
(E) Representative manifold of QW and active states (as function of locomotion speed, color-coded) for WT and Sca1 mice.
(F) Euclidean distance in PC space within and across brain states (WT: 12 FOV, 8 mice, Sca1: 10 FOV, 5 mice). Data in (D) and (E) are median + interquartile range.
*p < 0.05, **p < 0.01, ***p < 0.001. See also Figures S5–S7.

assessed PN dendritic calcium signals, likely reflecting complex matured; therefore, we evaluated the expression of the cal-
spiking31 (Figures S7A and S7B). At P60, when no change in PN cium-binding protein parvalbumin (PV), a well-established
somata was observed yet, we already found evidence for MLIN marker, at P30, when Sca1 mice lack motor incoordination
increased dendritic excitability, seen in a higher fraction of active (Figure S8A). Additionally, PV-expression levels are an indicator
dendrites (i.e., dendrites displaying at least one clear calcium of neuronal activity, coinciding with plasticity-related learning
transient, Figure S7C), and an increase in dendritic calcium sig- processes.32,33 Although the majority of P30, WT MLINs ex-
nals in Sca1 mice (Figures S7D and S7E). These changes per- pressed medium levels of PV, 32% of MLINs presented high
sisted until late disease stages (P200), with more active den- levels of PV in Sca1. The fraction of mutant MLINs expressing
drites (Figure S7F) and higher activity levels (Figures S7G high levels of PV increased as a function of disease progression
and S7H). with 49% at P90 and 60% at P200 (Figure 3A). To exclude that
the observed increase in mutant MLIN numbers, displaying
Mutant MLINs exhibit early connectivity and functional high levels of PV was due to MLIN degeneration, we quantified
changes in Sca1 mice the number of MLINs. Notably, until the end-stage, mutant
We sought to gain insight into the mechanisms governing early MLIN numbers remained unchanged, suggesting that the early,
MLIN hyperexcitability in SCA1. By P30 cerebellar cortex is yet gradual, increase in PV levels, reflect changes in MLIN

Neuron 111, 2523–2543, August 16, 2023 2527

 ll
OPEN ACCESS Article

(legend on next page)

2528 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

activity (Figure 3B). Furthermore, we examined PV levels in inter- lighted several neuronal populations, with transduction being the
neurons located in the primary motor cortex, a region involved in highest in PNs, followed by granule cell layer (GCL) neurons and,
movement execution. No change in PV-expression levels or lastly, MLINs, suggesting a developmentally regulated expres-
numbers was observed in Sca1 (Figures S8B and S8C), suggest- sion of the Cre-driver34 (Figure S9B). To validate the specificity
ing that alterations in PV-expression levels were restricted to the of AAV2/8-mediated transduction, AAV2/8-hSyn-DIO-mCherry
cerebellum. or AAV2/8-hSyn-DIO-hM4D(Gi)-mCherry (inhibitory designer re-
Next, we assessed the distribution of synaptic inputs onto ceptors exclusively activated by designer drugs [DREADD]) were
MLINs and found that Sca1 MLINs displayed increased numbers injected into adult WT::PV cerebellum. Analysis of the trans-
of excitatory synapses, marked by the vesicular glutamate trans- duced area revealed widespread mCherry expression, albeit
porter 1 (VGluT1) (Figure 3C). Concomitantly, mutant MLINs restricted to MLIN somata and their perisomatic synapses onto
received fewer vesicular GABA transporter (VGAT)-positive ho- PNs extending to the axonal initial segment, forming the pinceau
motypic inhibitory inputs from adjacent interneurons (Figure 3C). (Figure S9C). Next, we injected AAV2/2-hSyn-DIO-hM4D(Gi)-
Both alterations in E/I inputs onto MLINs persisted until late dis- mCherry and AAV2/8-hSyn-DIO-hM4D(Gi)-mCherry into
ease stage. Because altered E/I connectivity on MLINs might in- WT::Pcp2 neonates, wherein Cre-dependent expression is
fluence PN responses, we examined the connectivity between solely in PNs. AAV2/2 led to extensive mCherry expression in
MLINs and PNs at P30 and found an increased number of PNs throughout the cerebellum (82.4%); however, AAV2/8
VGAT-positive inhibitory synapses surrounding PN somata in transduction was markedly lower in PNs (42.4%) and exclu-
Sca1 (Figure 3D). However, an increased number of inhibitory sively within lobules IX and X (Figure S9D). In parallel, injections
synapses on mutant PNs might reflect a delay in circuit refine- into the adult WT::Pcp2 cerebellum (lobules VI and VII) revealed
ment and maturation, with no consequences on PN function. very high percentages of AAV2/2-mediated mCherry expression
Consequently, we evaluated the MLIN-PN connectivity via serial within adult PNs (71.25%), but AAV2/8-mediated mCherry
block-face scanning electron microscopy (SBF-SEM) in symp- expression was present only within 19.75% of PNs, suggesting
tomatic mice. The 3D reconstruction of individual MLINs and that the floxed version of AAV2/8 serotype preferentially trans-
their synaptic contacts onto PN somata revealed a significantly duces adult MLINs (Figure S9D, lower graph). Lastly, quantifica-
elevated number of axosomatic synapses in Sca1. This observa- tion of the mCherry expression in 17 animals from our experi-
tion supports the notion that enhanced inhibitory connectivity mental cohort confirmed that AAV2/8 serotype transduces
between MLINs and PNs is related to higher MLIN activity levels, adult MLINs with a very high selectivity (98% of all mCherry-
and this altered connectivity is maintained throughout disease positive cells), (Figure S9E).
(Figure 3E). Overall, these results demonstrate that in Sca1 To reduce the inhibitory tone onto PNs, we injected AAV2/
mice cerebellar dysfunction is based on the disinhibition of the 8-hSyn-DIO-hM4D(Gi)-mCherry, bilaterally into the posterior
cerebellar output, likely instigated by hyperexcitable MLINs, cerebellum of P70, Sca1::PV, and WT::PV animals (Figure 4A). Ex-
and that those deficits arise presymptomatically. pressed receptors were activated using the ligand clozapine-N-
oxide (CNO) for three consecutive days, thereby suppressing syn-
MLIN-mediated inhibition modulates SCA1 pathology aptic transmission from MLINs onto PNs. CNO treatment, but not
To probe for a causal role of MLIN hyperactivity in influencing saline application, led to reduced PV-expression within MLINs, the
motor pathology, we performed chemogenetic experiments. reduction being stronger in Sca1::PV MLINs, suggesting that
To selectively target MLINs, Sca1 mice were crossed with homo- higher PV-expression in MLINs is dynamic and reversible (Fig-
zygous PV-Cre mice to generate Sca1::PV or WT::PV animals ure 4B). Notably, diminished MLIN activity counteracted Calbindin
(Figure 4A). To infect MLINs and not PNs, we injected the three reduction in Sca1 PNs (Figure S10A), a disease-specific marker for
common AAV serotypes: AAV2/1-hSyn-DIO-mCherry, AAV2/2- PN dysfunction.35 Moreover, restoration of phosphorylated
hSyn-DIO-mCherry, and AAV2/8-hSyn-DIO-mCherry into the CaMKII (P-CAMKII) expression within mutant PNs was observed,
adult WT::PV cerebellum. mCherry immunolabeling revealed indicative of ameliorated intracellular calcium signaling and PN
that, although all serotypes primarily transduced MLINs, AAV2/ activation in response to reduced PN inhibition (Figure S10B).
8 infected a very high proportion of MLINs (Figure S9A). To char- We next assessed whether the improvement in disease-asso-
acterize AAV2/8 serotype selectivity for MLINs, we performed in- ciated PN-specific hallmarks also affected symptomatic ataxia.
tracerebroventricular (i.c.v.) injections in WT::PV neonates and To exclude side effects of CNO, we administered CNO (3 mg/
analyzed the cerebellum at P35. mCherry immunolabeling high- kg) intraperitoneally (i.p.) to naive (WT) mice, 45 min before the

Figure 3. Altered inhibitory connectivity in the adult Sca1 cerebellar cortex

(A) Representative confocal images (R.I.) of MLINs stained for parvalbumin (PV) from P30, WT, and Sca1 cerebellum and quantitative analysis (Q.A.) of the
percentage of neurons expressing low, medium, and high levels of PV at P30 (presymptomatic), P90 (symptomatic), and P200 (end-stage). Chi-squared test.
(B) No difference in MLIN numbers between WT and Sca1 at different ages.
(C) R.I. of MLIN stained for PV, VGAT (inhibitory), or VGluT1 (excitatory) synaptic markers.
(D) R.I. of PNs stained for Calbindin and VGAT illustrating increased density of inhibitory synapses onto Sca1 PNs soma.
(E) Representative 3D-SBF-SEM reconstructions of WT and Sca1 PNs (blue), MLINs (green), inhibitory axosomatic synapses (red), excitatory synapses
(magenta). Right: 2D-SBF-SEM image: inhibitory axosomatic synapse (magenta) and PN soma (cyan), showing elevated inhibitory axosomatic inputs onto mutant
PNs. No. of animals: 3-6/genotype/age. Graphs (B–D) depict mean values/animal, (E) depicts mean value per PN. Scale bars, (A) 20, (C) 2, (D) 10, (E) 2 and 0.5
(mm). *p < 0.05, **p < 0.01, ***p < 0.001. See also Figure S8.

Neuron 111, 2523–2543, August 16, 2023 2529

 ll
OPEN ACCESS Article

(legend on next page)

2530 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

rotarod test, which did not cause any adverse effects (Fig- mice (Figure 4M), as did the calcium signal fluctuations in PN
ure S11A).36 Subsequently, we tested the inhibitory DREADD- somata (Figure 4N), potentially reflecting a more stable pace-
expressing mice for rotarod performance before and after the maker firing activity, which would cause less fluctuations in the
surgical procedure. At P85, mice were tested for three consecu- fluorescence trace. PN dendrites were not affected upon CNO
tive days on the rotarod after CNO/saline injection. Surprisingly, in awake mice (Figure 4O). We examined whether the compro-
suppressing synaptic transmission between MLINs and PNs did mised coding capacity would normalize upon chemogenetic
not have any effect on WT::PV animals. However, a substantial dampening of MLIN excitability. Computing the manifolds of all
improvement in motor performance was observed in Sca1::PV neurons in a given FOV, we found two subspaces (Figure 4P),
animals after CNO injections (Figure 4C). Sca1::PV mice dis- which were less discrete in Sca1::PV mice. Intriguingly, after
played improved rotarod performance, even after 24 h of CNO CNO application, the subspaces became more discrete, and
washout, suggesting some form of lasting plasticity. To exclude the Euclidean distance between both states became insignifi-
that the amelioration in motor incoordination was due to non- cant between WT::PV and Sca1::PV mice (Figure 4Q). The inves-
specific viral transduction of brain stem and spinal cord neurons, tigation of neuronal function upon MLIN silencing thus demon-
we performed immunostaining for mCherry, whose expression strates a normalization not only of activity levels in Sca1 but
remained localized within the cerebellum, with no unspecific also of the networks’ coding capacity.
spreading to other CNS regions (Figures S11B and S11C).
Additionally, we probed the impact of chemogenetic inhibition Enhancing MLIN activity is sufficient to induce SCA1-
of MLINs on neuronal function. We co-injected AAV2/8-hSyn- like pathology in WT mice
DIO-hM4D(Gi)-mCherry and AAV2/1-syn-jGCaMP7s into P75, We hypothesized that augmented inhibition could impair PN
Sca1::PV, and WT::PV mice and performed in vivo calcium function, thereby inducing motor incoordination; thus, we
imaging in P110 aged mice (Figure 4D). Although CNO applica- mimicked the altered mutant MLIN activity in WT animals. We
tion alone in control mice had no effect on neuronal activity injected excitatory DREADDs (AAV2/8-hSyn-DIO-hM3D(Gq)-
(Figures S11D–S11I), we found that 50% of MLINs in mCherry) in the WT::PV cerebellum (Figures 5A and 5B).
DREADD-expressing mice displayed reduced neuronal activity Elevated PV-expression within WT::PV MLINs, similar to those
under light anesthesia. About 20% of MLINs did not overtly in Sca1::PV MLINs, after CNO application was observed (Fig-
change their activity levels, and the remaining MLINs increased ure 5C). Importantly, we also found that the rotarod performance
their activity (Figure 4E). The overall distribution of absolute ac- of WT::PV mice that received CNO injections drastically drop-
tivity changes upon CNO shows a reduction of activity levels in ped, almost reaching levels observed in Sca1::PV mice (Fig-
a substantial proportion of MLINs (Figures 4F–4H). We observed ure 5D). Centered on the acute MLIN overactivation data and
reduced activity in Gol exclusively in WT::PV mice (Figure 4I), the fact that NDs show a progressive increase in symptoms,
whereas calcium signals in PNs somata were not affected by we mimicked this aspect by chronically activating MLINs either
the CNO-triggered silencing of MLINs (Figure 4J). However, a for 30 or 60 days in WT::PV mice (Figure 5E). Because SCA1 pa-
significant decline in calcium signals in PN dendrites in WT::PV thology is characterized by initial dysfunction in PN spines, de-
and Sca1::PV mice was observed (Figure 4K). These findings tected by reduced Homer-3 expression,14 we examined
were corroborated by recordings in awake mice, in which CNO Homer-3 levels. Notably, 30 days of CNO application led to a sig-
administration revealed a strong reduction in the spontaneous nificant reduction in Homer-3-positive puncta, which was main-
activity of MLINs during QW (Figure 4L). In awake mice, Gol ac- tained after 2 months of CNO treatment, suggesting a threshold-
tivity decreased upon CNO in WT::PV and as a trend in Sca1::PV ing effect (Figure 5F). Moreover, P-CAMKII levels in PN soma

Figure 4. Acute MLIN activity reduction influences cerebellar network responses and alleviates motor deficits
(A) Experimental design showing acute MLIN inhibition in WT::PV and SCA1::PV mice via AAV2/8-DREADD(Gi). Right: R.I. of mCherry expression within WT::PV
MLINs (yellow arrowheads), note the absence of mCherry staining in PNs (blue arrowheads).
(B) R.I. of comparable PV expression in WT::PV and Sca1::PV-expressing DREADD(Gi) MLINs after saline/CNO injection, n = 3 mice/group.
(C) Rotarod measurement reveals improved motor performance of Sca1::PV-DREADD(Gi)+CNO mice vs. saline group. WT::PV+saline n = 9; WT::PV+CNO n = 8;
Sca1::PV+saline n = 8; Sca1::PV+CNO n = 10 mice.
(D) In vivo imaging experiment design. Same neurons were recorded (fb, iso) at baseline (BL, w/o CNO, P110) and upon CNO application.
(E) Fraction of neurons with increased, decreased, or no change in activity in Sca1::PV MLIN+CNO.
(F) Absolute activity changes in MLINs upon CNO.
(G) Maximum intensity projection of a Sca1::PV mouse before and after CNO.
(H–K) (H) MLIN activity under anesthesia before and after CNO in WT::PV (left, 1,088 MLINs) and in Sca1::PV (right, 548 MLINs), (I) same as in (H) for Gol in WT::PV
(78 Gol) and Sca1::PV (38 Gol), (J) and in PN in WT::PV (102 PN) and Sca1::PV (83 PN) and (K) calcium signals in PN dendrites in WT::PV (83 PN dendrites) and
Sca1::PV (88 PN dendrites), data in (H)–(K) from WT::PV: 12 FOV, and Sca1::PV: 8 FOV, 5 mice/genotype).
(L–O) (L) QW spontaneous activity of MLINs before and after CNO in WT::PV (1,316 MLINs) and Sca1::PV (813 MLINs), (M) of Gol in WT::PV (82 Gol) and Sca1::PV
(56 Gol), (N) of PN in WT::PV (49 PN) and Sca1::PV (76 PN) and (O) calcium signals in PN dendrites in WT::PV (12 PN dendrites) and Sca1::PV (19 PN dendrites),
data from WT::PV: 17 FOV, and Sca1::PV: 17 FOV, 7 mice/genotype).
(P) Example manifold of all neurons in a FOV from a Sca1::PV mouse, depicting less discrete segregation of QW and active states (running speed color-coded),
which is improved after CNO application (right).
(Q) Euclidean distance in PC-space between the center of mass of QW and locomotion (before CNO: WT::PV: 10 FOV, Sca1::PV: 12 FOV, after CNO: WT::PV: 9
FOV, Sca1::PV 11 FOV, (5–6 mice/genotype). Graph (C) represents mean ± SEM, data in (H–O) and (Q) represent median and interquartile range. Scale bars, (A)
100, (B) 20, (G) 50 (mm). *p < 0.05, **p < 0.01, ***p<0.001. See also Figures S9–S11.

Neuron 111, 2523–2543, August 16, 2023 2531

 ll
OPEN ACCESS Article

(legend on next page)

2532 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

were significantly reduced to levels observed in Sca1 PNs (Fig- roprotective. Inhibitory DREADD AAV2/2-hM4D(Gi)-mCherry was
ure 5G). Subsequently, we examined whether the cellular expressed in WT::Pcp2 and Sca1::Pcp2 mice (Figure S12D). Acute
changes translated into motor dysfunction (Figure 5H). Rotarod CNO application reduced P-CAMKII expression in both WT and
measurements during chronic CNO application revealed a Sca1 PNs, confirming reduced intracellular calcium-signaling
consistent decline in the latency to fall in WT::PV mice (Figure 5I). and PN activity (Figure S12E). Chronic CNO treatment from P60-
Interestingly, motor incoordination was maintained at least until P70, surprisingly, anticipated PN degeneration selectively in
20 days after halting CNO treatment, suggesting that chronic Sca1 mice (Figures S12F and S12G). This differential outcome
MLIN overactivation, even in WT::PV mice, not only induces a on disease pathology dependent on the neuronal population tar-
SCA1-like phenotype but also leaves a long-lasting trace (Fig- geted, highlights the contribution of MLIN hyperresponsiveness
ure 5I). Lastly, we analyzed the same cohort of mice 130 days af- to SCA1 pathology.
ter halting CNO treatment and found that Homer-3-expression Lastly, measurement of rotarod performance, during and after
remained reduced (Figure 5J), whereas P-CAMKII levels were CNO application, revealed a striking amelioration of motor coor-
somewhat normalized within PN soma of WT::PV mice (Fig- dination deficits in Sca1::PV animals, expressing inhibitory
ure 5K). These results show that MLIN-mediated, elevated inhi- DREADD in MLINs (Figure 6F). Notably, despite the termination
bition of PNs, non-cell autonomously, promotes cellular pathol- of the CNO application at P90, Sca1::PV symptomatic animals
ogy within PNs and triggers manifestations of ataxia-like displayed a significantly improved motor performance
symptoms in WT::PV mice, lacking disease-causing mutations. compared with the untreated Sca1::PV group for another
2 months, whereas no significant effect on WT::PV motor perfor-
Chronic disinhibition of PNs is neuroprotective in mance was observed (Figure 6F). Chronic CNO application de-
Sca1 mice layed the onset of the hindlimb clasping phenotype, a measure-
Next, we hypothesized that 30 days of chronic reduction in MLIN ment that directly correlates with disease severity, with higher
activity in Sca1::PV mice should provide long-term amelioration scores indicative of symptom worsening (Figure 6G).
in cerebellar dysfunctions. To this end, we selectively expressed
inhibitory DREADDs in WT::PV and Sca1::PV MLINs at P40, fol- Proteomic analyses of adult MLIN identifies widespread
lowed by the chronic administration of CNO in drinking water synaptic alterations
(40 mg/mL) for 30 days. Analysis performed after stopping the Because functional MLIN alterations in SCA1 are strong contribu-
CNO treatment, at an age corresponding to the late disease stage tors to the degenerative process, we investigated the molecular
(Figure 6A), revealed sustained reduction of PV-expression, spe- signature, accounting for the functional changes in mutant
cifically within mutant MLINs (Figure 6B), and normalized MLINs. We undertook a FACS approach to selectively isolate adult
P-CAMKII levels, with no changes in total CAMKII levels (Figure 6C, PV-positive MLIN for proteomic analysis (Figure 7A).38 MLINs were
S12A, and S12B). Moreover, a beneficial effect on PN synaptic labeled by injecting AAV8- hSyn-DIO-mCherry into P30-35 WT::PV
structures was observed via an increase in Homer-3 immunoreac- and Sca1::PV animals. Two weeks post surgery, cerebellar tissue
tive puncta at P180, suggesting that the amelioration in pathology was dissociated, and viable single neurons were sorted for
due to reduced MLIN activity is maintained long term, despite mCherry and Hoechst (Figure S13A). We confirmed that the sorted
halting CNO-mediated modulation (Figure 6D). This overall protec- neurons were viable PV-positive interneurons by culturing them for
tive effect was also reflected by sustained PN numbers (Figure 6E), 24 h and staining for MLIN marker PV (Figure S13B). Sorted MLINs
indicating that reduced inhibitory drive onto Sca1 PNs ameliorates were lyophilized and processed for mass spectrometry, revealing
cerebellar degeneration. Next, we assessed whether MLIN modu- that the sorted neurons present typical MLIN markers (Figure 7B).
lation can alter the activation of astrocytes and microglia, which is Proteomic analyses highlighted a large number of downregulated
observed early in SCA1 pathogenesis.37 Gliosis levels were signif- proteins in Sca1 MLINs (Figure 7C; Table S1). Panther (protein anal-
icantly lower in cerebellar lobules where MLIN activity was reduced ysis through evolutionary relationships) protein class analyses of
in comparison with neighboring non-transduced lobules, indi- downregulated proteins revealed that 34.2% of those belong to
cating a general beneficial effect of reducing MLIN activity in the metabolite interconversion enzyme class, 17.8% to RNA meta-
Sca1 mice (Figure S12C). Because PN dendrites exhibited hyper- bolism protein, 11% to transporter protein class, and 9.6% to
activity, we examined whether reducing PN activity would be neu- membrane trafficking proteins (Figure 7D). As previously shown,

Figure 5. Mimicking mutant MLIN hyperresponsiveness in WT::PV mice

(A) Experimental design for acute cerebellar MLIN activation in WT::PV mice.
(B) R.I. of mCherry expression in WT::PV, lobule VIII.
(C) Mimicking mutant MLIN hyperresponsiveness in WT::PV MLINs by CNO stimulation of excitatory DREADD (Gq), increases PV levels.
(D) WT::PV animals display impaired motor performance on rotarod vs. saline group after CNO injection WT::PV+saline n = 7, WT::PV+CNO n = 10 mice).
(E) Experimental design for chronic MLIN activation in WT::PV animals.
(F) R.I. and Q.A. of reduced Homer-3 expression in WT::PV-mCherry and WT::PV-DREADD(Gq) mice after 30 or 60 days of CNO treatment.
(G) Q.A. of reduced P-CAMKII expression in WT::PV-DREADD(Gq) and WT::PV-mCherry PNs, no change in total CAMKII.
(H) Experimental design for chronic WT::PV MLIN activation, analyzed at P200 (bold arrow: longitudinal rotarod measurement throughout).
(I) Chronic CNO administration impairs motor performance in WT::PV-DREADD(Gq) mice, n = 8 animals/group.
(J) R.I. and Q.A showing reduced Homer-3 expression in WT::PV-DREADD(Gq) vs. WT::PV-mCherry.
(K) Q.A. of P-CAMKII and CAMKII expression after chronic activation of MLINs in WT::PV mice. n = 3–5 mice/analysis. Graphs (D and I) represent mean ± SEM, (F,
G, J, and K) depict mean values/animal. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars, (B) 200, (C) 10, (F) 10, (J) 10 (mm).

Neuron 111, 2523–2543, August 16, 2023 2533

 ll
OPEN ACCESS Article

(legend on next page)

2534 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

proteins involved in synaptic function are prominently altered in and cortical interneurons.40 A fraction of PV-positive iGNs ex-
Sca1 PNs.14 Similarly, MLINs presented an altered synaptic profile, pressed high levels of SORCS3 (62%), suggesting that at least
with proteins involved in synaptic transmission being shared and a fraction of iGNs also possess MLIN-like characteristics
reduced in expression in both MLINs and PNs, indicating that (Figure S14G).
mutant Ataxin-1 selectively affects the expression of synaptic pro- Subsequently, immunostaining revealed higher immunoreac-
teins of dedicated cerebellar circuit elements (Figures 7D tivity for FRRS1l and GluR2 within human SCA1 iGNs compared
and S13C). with Controls (CNTRLS) (Figure 8A). The higher expression of
Panther protein class analysis of the upregulated proteins re- FRRS1L in SCA1 iGNs directly correlated with elevated PV im-
vealed that 17.2% of those proteins belong to the metabolite munolabeling, 57% of iGNs displayed elevated PV levels (Fig-
interconversion enzyme group, 11.1% to protein modifying ure 8B). These results corroborate findings in Sca1 mice in Fig-
enzyme and 9.1% to transporter protein class (Figures 7E and ure 3A, in which 49% of P90, MLINs exhibited high PV
7F). Interestingly, we identified the increased expression of the expression. Because in vivo mutant MLINs are excessively hy-
a-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate (AMPA) perexcitable and hyperresponsive to sensorimotor stimuli, we
receptor-associated protein, ferric chelate reductase 1-like wondered whether this hyperexcitability of MLINs is a conserved
(Frrs1l), involved in AMPA receptor biogenesis and trafficking.39 feature, detectable in SCA1 iGNs. Live-cell calcium imaging was
Immunoblotting total cerebellar lysates displayed no difference performed on human iGNs that were transduced with the green
in Frrs1l expression in Sca1 (Figure S13D). However, immuno- calcium indicator AAV2/1-hSyn-GCaMP6s, 5-days post differ-
staining MLINs against Frrs1l revealed the elevated expression entiation, and calcium imaging was conducted at day in vitro
of Frrs1l in Sca1 MLINs. As further evidence, higher Frrs1l levels (DIV)12 (Figure 8C). Both CNTRL and SCA1 iGNs fell into two
positively correlated with increased PV levels in Sca1 MLINs differentially responding populations. Among the neurons that
within the posterior cerebellum (Figure 7G). Notably, Frss1l levels increased their cytoplasmic calcium levels after 50 mM KCl stim-
exhibited an anterior-posterior gradient, akin to that observed in ulations, two populations of MLINs were observed, labeled as
the degeneration of PNs. A higher fraction of posterior MLINs, ex- high and low-responding iGNs (Figure 8D). Notably, a substantial
pressing elevated levels of Frrs1l at both symptomatic (P100) and proportion, 78% of SCA1 iGNs, fell into the high-responding
late symptomatic (P200) stages, was observed (Figure 7H). category (median DF < 1.5 at 70 s), displaying prominently higher
Consequently, we assessed AMPA receptor-subunit, GluR2 calcium signals compared with the CNTRL iGNs, where only
expression in MLINs, which revealed a dramatic increase in 58% of the iGNs fell into the high-responding category. More-
GluR2 immunoreactivity on the surface of Sca1 MLINs (Figure 7I), over, the amplitude in high-responding MLINs at 70 s was higher
likely accounting for the increase in MLIN excitability. in SCA1 iGNs compared with CNTRL iGNs (Figure 8D, middle
panel). Mutant iGNs exhibited a significantly longer rise time in
Conserved hyperresponsiveness of human SCA1 the high-responding population compared with healthy iGNs
patient iPSC-derived GABAergic neurons (Figure 8D, right panel).
To validate our findings in the context of human pathology, we The high expression levels of FRRS1L in SCA1 iGNs indicated
generated induced GABAergic neurons (iGNs) from induced a probable link between FRRS1L levels and MLIN hyperrespon-
pluripotent stem cells (iPSCs)38 derived from three SCA1 patients siveness. To clarify this association, lentivirus-mediated overex-
and their unaffected siblings (Figure S14A). We evaluated the purity pression of FRRS1L (LV::FRRS1L-myc) or knockdown of FRR
of our cultures and performed an in-depth characterization of the S1L (LV::sh-FRRS1L-GFP) was performed (Figures 8E and
generated interneuron populations (Figure S14B). Two-week-old S15A). Calcium imaging revealed that FRRS1L-overexpressing
iGNs were 91% positive for the neuronal marker MAP2 and nega- CNTRL iGNs presented higher calcium signals, closely
tive for astrocyte marker GFAP (Figures S14B and S14C). Of the mimicking those observed in SCA1 iGNs (Figure 8F). By
91% MAP2-expressing neurons, 91% expressed the neurotrans- contrast, the overexpression of FRRS1L in SCA1 iGNs did not
mitter GABA (Figure S14D). Of the 91% GABA-positive neurons, further elevate calcium traces, suggesting a ceiling effect
93% were PV-positive, but lacked Calbindin and expressed (Figure S15B). FRRS1L knockdown was performed by LV car-
Kv1.2, indicating that neurons were excitable and active rying the small hairpin (sh) RNA-tagged to GFP; therefore, red
(Figures S14E and S14F). Because our protocol generated pan- calcium indicator AAV2/1-hSyn-RCaMP was used, and we first
inhibitory interneurons, we immunostained for SORCS3, an confirmed that both calcium indicators elicited similar responses
MLIN-specific marker, which is absent in cerebellar Gol, PNs, (Figure S15C). The knockdown of FRRS1L dramatically reduced

Figure 6. Chronic suppression of MLIN activity delays SCA1 pathology

(A) Experimental design for chronic MLIN inhibition.
(B) R.I. of PV expression within WT::PV and Sca1::PV MLINs. Chronic inhibition of Sca1::PV MLINs via DREADD(Gi), normalizes PV expression to WT levels.
(C) R.I. and Q.A. of restored P-CAMKII expression in PNs.
(D) 3D-isosurface reconstructions showing elevated Homer-3 and Calbindin expression within the molecular layer in Sca1::PV-DREADD(Gi) mice.
(E) R.I. of Calbindin labeled PNs from four conditions and Q.A. of sustained PN numbers in Sca1::PV-DREADD(Gi) vs. Sca1::PV-mCherry group.
(F) CNO chronic administration delays appearance of motor symptoms in Sca1::PV-DREADD(Gi) mice, WT::PV-mCherry n = 13; WT::PV-DREADD(Gi) n = 11;
Sca1::PV-mCherry n = 13; Sca1::PV-DREADD(Gi) n = 10 mice.
(G) Clasping phenotype across disease progression in Sca1::PV-mCherry and Sca1::PV-DREADD(Gi), showing reduced clasping severity in DREADD(Gi) cohort.
Graphs (C–E) depict mean values/animal, (F) represents mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars, (B) 20, (C and E) 20, (D) 4 (mm). See also
Figure S12.

Neuron 111, 2523–2543, August 16, 2023 2535

 ll
OPEN ACCESS Article

(legend on next page)

2536 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

calcium responses in both SCA1 and CNTRL iGNS (Figures 8G PNs. Notably, MLINs display irregular fast spiking, which is
and S15D). Due to the very high transduction efficiency of LVs, mainly regulated intrinsically.17 Thus, the observed increase in
FRRS1L overexpression completely abolished the low-respond- spontaneous activity of MLINs during stationary/quiescent
ing population of iGNs, and FRRS1L knockdown completely epochs could relate to increased intrinsic hyperexcitability and/
abrogated the high-responding fraction of iGNs (Figure S15E), or defective reciprocal inhibition due to reduced synaptic con-
thus highlighting the role of FRRS1L in governing MLIN hyperres- nections among MLINs. Although we found MLINs to be hyper-
ponsiveness in SCA1. excitable, we observed a reduction in Gol sensorimotor re-
sponses. Gol are inhibitory GABAergic neurons, located within
DISCUSSION the GCL, providing crucial regulation and gain control of granule
cell activity via lateral inhibition in a feedforward inhibitory
This study provides for the first-time direct evidence for a causal loop.25,45,46 Importantly, Gol receive inhibitory input from the
impact of a hitherto poorly recognized circuit element, namely MLINs, whereas both neuronal types are activated by PFs.45,47
MLINs, in the degenerative process of Sca1 PNs. Our data It is possible that the hyporesponsiveness of mutant Gol is a
demonstrate that hyperexcitability of MLINs occurs very early in consequence of MLIN hyperactivity, which could potentially
the disease process and persists until the end-stage. Together affect PF input, resulting in a vicious stimulation cycle.
with an enhanced synaptic connectivity between MLINs and Our data demonstrate that mutant MLINs exhibit aberrant spon-
PNs, it compromises cerebellar sensorimotor processing, fuels taneous activity and aberrant responses to sensorimotor input,
PN synaptic dysfunction, and causes motor deficits. Our data, us- persisting from early until late disease stages. Together with higher
ing chemogenetic modulation of MLINs, demonstrate that circuit synaptic connections onto PNs, we propose that PNs undergo a
dysfunction is sufficient to trigger SCA1-disease-like alterations dual hit by MLINs, thus strongly disrupting correlated PN activity
and that PNs remain the most vulnerable neurons within the cere- and compromising their pacemaker firing properties.24 Addition-
bellum. We further identify the molecular signature of mutant ally, we demonstrated dendritic hyperexcitability of PNs, present
MLINs, revealing that enhanced Frrs1L expression influences throughout the disease course, suggesting that increase in PN
MLIN activity responses. This work provides crucial evidence for dendritic excitability might represent a homeostatic mechanism
circuit dysfunction as an important driver of SCA1 pathophysi- in response to excessive inhibition, resulting in a maladaptive pro-
ology, and neurons, which primarily don’t degenerate, play a deci- cess. Nonetheless, we did not observe major alterations in somatic
sive role in shaping circuit output and triggering the pathology. calcium signals in PNs and found reduced spontaneous activity,
The cerebellum is a multimodal sensorimotor integration hub, along with a strong increase in cytosolic calcium levels in PNs
facilitating the fine-tuning of movements, balance, stance, and somata at late disease stages. However, the very high instanta-
cognition, all of which are compromised in SCAs.41 Our in vivo neous firing rates of PNs (20–70 Hz)48 and the rather slow kinetics
calcium imaging experiments unexpectedly unraveled a selec- of calcium indicators, primarily facilitate the detection of simple
tive sensorimotor hyperresponsiveness of MLINs, which provide spike pauses.22 Hence, our approach falls short in detecting
the main inhibitory input onto PNs. MLINs receive excitatory more subtle changes in PN activity.
sensorimotor signals through PFs and also from CFs via gluta- Moreover, several studies highlight deficits in PNs, associ-
mate spillover.17 Moreover, MLINs are reciprocally connected, ated with genes regulating calcium homeostasis in different
enabling tightly regulated feedforward and fb inhibition. Recent SCA mouse models, such as impaired transactivation of reti-
studies describe a cardinal role for MLINs in valence coding, noic acid-related orphan receptor (ROR) alpha by mutant
associative learning,36 as well as licking-related activity, modified Ataxin-1.49 Impaired calcium homeostasis is a convergent
by, e.g., motivational state.42,43 Chemogenetic-mediated MLIN pathway in SCAs, involving voltage-gated calcium channels50
inhibition reduces lick rate and precision, emphasizing the role such as SCA6, 51 SCA752, and mutations in the a-subunit of
of MLIN-mediated PN inhibition for movement control.42 Notably, Cav2.1 channel, causing episodic forms of ataxia, such as
optogenetic stimulation of MLINs can trigger motor behavior44 EA2.53,54 Furthermore, BK channels, involved in establishing
through PN inhibition, and concerted MLIN activity scales with PN pacemaker activity, are downregulated in SCA1 mice, and
locomotion speed, mediated by the silencing of PN.31 enhancing their expression restores PN pacemaker firing.24
We show that the increase in spontaneous neuronal activity of Intriguingly, a recent publication suggests a negative fb loop
MLINs in Sca1 reduces the coding space and renders brain state by which PN, in turn, inhibit MLINs;55 thus, it is conceivable
representations less distinct, thus sending ambiguous signals to that compromised PN pacemaker firing could affect MLIN

Figure 7. Frrs1l likely determines MLIN hyperresponsiveness

(A) Experimental design for isolating mCherry labeled MLINs from adult WT::PV and Sca1::PV cerebellum (n = 9 mice/genotype).
(B) MLIN-selective markers identified via mass spectrometry (MS) from one experiment (3 cerebella/genotype).
(C) Downregulated (iLFQ Log2FC 1.4) proteins in Sca1::PV MLINs.
(D) PANTHER overrepresentation test: downregulated protein classes in Sca1::PV MLINs.
(E) Upregulated (iLFQ Log2FC 1.4) proteins in Sca1::PV MLINs.
(F) PANTHER overrepresentation test: upregulated protein classes in Sca1::PV MLINs.
(G) MS-based iLFQ values of Frrs1l. R.I. of increased Frrs1l and PV expression in MLINs from WT::PV and Sca1::PV, and Pearson’s correlation.
(H) Q.A. of anterior (lobules II–V) and posterior (lobules VI–X) cerebellum for Frrs1l expression in MLINs.
(I) R.I. of GluR2 and Q.A. displaying increased GluR2 expression in Sca1::PV MLINs vs. WT::PV. n = 3–5 mice for analysis. Graphs (B and G) represent mean ±
SEM, (I) depicts mean values/animal, *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars, (G) 20, (I) 5 (mm). See also Figure S13 and Table S1.

Neuron 111, 2523–2543, August 16, 2023 2537

 ll
OPEN ACCESS Article

(legend on next page)

2538 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

activity and similar to Gol drive vicious loops of MLIN causing increased inhibitory synaptic input on PNs, particularly
hyperactivation. at pinceau synapses, thus strongly regulating action potential
Our study reveals that altered GABAergic inputs can causally generation.62 Despite a striking similarity to SCA1, with regard
contribute to circuit-level dysfunctions. Notably, reducing to excessive MLIN-mediated inhibition of PNs, EA1 lacks neuro-
GABA transmission also restored PN pacemaker activity in an degeneration, likely due to unaltered basket cell firing.63 Notably,
ethanol-exposure model.56 The selective lowering of the inhibi- altered excitability of PV-interneurons has been reported in other
tory tone by chemogenetically silencing Sca1 MLINs in our study, neurological diseases.64–68 In Alzheimer’s disease, interneuron-
lead to an immediate improvement in motor performance. This deficits not only underlie cognitive decline but also epileptic dis-
rapid improvement suggests that the Sca1 phenotype is not pri- charges.68,69 In Dravet syndrome, PV-interneurons exhibit
marily due to PN degeneration, and importantly highlights the impaired excitability, thus participating in disease pathophysi-
therapeutic potential of this cell-type-specific intervention. ology.70 In ALS and Fragile-X syndrome, PV-interneurons are
Intriguingly, acute chemogenetic-mediated MLIN silencing hypoexcitable, thereby causing cortical hyperexcitability.66,71
significantly reduces the hyperexcitability of PN dendrites, indi- SCA1, however, represents a special case, as MLINs turn hyper-
cating that this hitherto believed PN cell-autonomous alteration excitable. But why are MLINs hyperexcitable? We show that ho-
might, at least in part, constitute a maladaptive feature in motypic synaptic connections are altered, likely resulting in
response to excess MLIN-mediated inhibition. defective reciprocal inhibition among the MLIN population.
However, MLIN hyperexcitability is not the sole driver of PN Moreover, in Sca1 mice, we found strongly increased levels of
degeneration, especially as ample evidences point to PN cell- the calcium buffer PV and lowered resting cytoplasmic calcium
autonomous processes.57,58 Nevertheless, recent studies concentrations, which recently have been linked to increased
highlight alterations in non-cell autonomous processes, syner- neuronal excitability.72 We propose that altered calcium homeo-
gistically enhancing or even initiating dysfunction within PNs. stasis observed in Sca1 MLINs contribute toward higher
Excitatory PF-PN and CF-PN synaptic dysfunctions are MLIN activity. Patient-derived iGNs experiments also argue for
observed in rodent models of SCAs, indicating a conserved intrinsic alterations, and our proteomic analyses identified alter-
role for circuit elements in either triggering or governing cere- ations in synaptic molecules, which might contribute to altered
bellar neurodegeneration. Specifically, these alterations could MLIN synaptic function and neurotransmission.
affect intrinsic PN excitability, as well as alter synaptic re- FRRS1L is a recently identified, predominantly brain-localized
sponses.14,59–61 Ataxin-1 controls the expression of Cav3.1 protein, with the highest expression in the cortex, cerebellum,
and TRPC3 channels, involved in the regulation of intrinsic hippocampus, and basal-ganglia.73 Patients harboring recessive
PN excitability, and mutant Ataxin-1 renders PNs hyperexcit- loss-of-function mutations in FRRS1L gene, exhibit severe intel-
able in a cell-autonomous fashion.30 Therefore, homeostatic lectual disability, movement disorders, hypotonia, and epi-
mechanisms in response to excess MLIN-mediated inhibition lepsy.39,73,74 Notably, some patients manifest cortical and cere-
might lead to further reductions in the expression of these re- bellar neurodegeneration. FRRS1L within the ER regulates
ceptors and anticipate PN degeneration. AMPA receptor assembly via FRRS1l/CPT1c complexes,
Our chemogenetically mediated MLIN stimulation in healthy enabling multimer-formation of GluAs and their trafficking to
mice highlights the pivotal role for this circuit element in driving the plasma membrane.75 In Sca1 mice and in patient iGNs, a se-
motor deficits typical of SCA1. Importantly, it demonstrates the lective increase in FRRS1L levels indicate that FRRS1L levels
vulnerability of PNs to extended periods of excessive inhibition, can influence synaptic plasticity and shape activity patterns.
which causes long-term synaptic deficits on PNs, and, on a Importantly, FRRSIL overexpression in CNTRL iGNs alters their
longer time scale, might cause overt PN degeneration. An inter- intrinsic properties, rendering them hyperresponsive to depola-
esting pathophysiological overlap exists with Episodic Ataxia rizing stimulation. However, future studies are needed to deci-
type 1 (EA1), characterized by stress-induced attacks of spastic pher how Ataxin-1 contributes to altered FRRS1L expression.
contractions and motor dysfunction. EA1 is a channelopathy, One likely pathway for future investigation is that of Sirtuin-1,

Figure 8. iGNs activity in Control and SCA1 correlates with PV and FRRS1L expression
(A) R.I. of FRRS1L, GluR2 and PV expression in 3 human SCA1 patients’ iPSC-derived iGNs vs. 3 CNTRL lines and Q.A. of elevated. GluR2 expression.
(B) Q.A. of PV expression binned into expression classes: high, medium, and low (chi-squared test). Pearson’s correlation between PV expression and FRRS1L in
SCA1 vs. CNTRL. Q.A. of FRRS1L expression level.
(C) Experimental timeline for iGNs calcium imaging and Q.A. of differentially responding populations after 50 mM KCl stimulation (low-responding cut off: 1–1.5 at
70 s; high-responding < 1.5 at 70 s) and corresponding heatmaps. Q.A. of calcium transients showing increase in DF at 70 s in SCA1 high-responding neurons
vs. CNTRL.
(D) Significant increase in DF amplitude at 70 s in SCA1 high-responding neurons. Time to max DF was significantly higher in high-responding SCA1 populations.
n = 27–33 neurons.
(E) Timeline for calcium imaging of iGNs transduced at DIV4 with LV:FRRS1L-myc (transduction efficiency, mean: 83.06) or LV:shFRRS1L-GFP (transduction
efficiency mean: 89.60) or LV-GFP. Subsequently at DIV7 iGNs were infected with AAV2/1-GCaMP6 or AAV2/1-RCaMP. R.I. of FRRS1L overexpression or
knockdown in iGNs.
(F) Q.A. of calcium transients in CNTRL lines overexpressing FRRS1L, showing increase in DF after KCl stimulation, n = 12–15 neurons/condition.
(G) Q.A. of calcium transients in SCA1 lines after FRRS1L knockdown, displaying decline in DF after KCl stimulation. n = 12–14 neurons/condition. Box and
whisker plot in (A and B) and graphs in (C and F) show mean ± SEM, (E) percentage of mean/coverslip. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars, (A, B, and D)
15 mm. See also Figures S14 and S15.

Neuron 111, 2523–2543, August 16, 2023 2539

 ll
OPEN ACCESS Article

which normalizes calcium homeostasis, thus promoting neuro- ACKNOWLEDGMENTS

protection in SCA7.76
We are grateful to all the SCA1 patients and their families united in the Dutch
Lastly, our findings present new evidence for early cerebellar
SCA1 Families Fund for donating fibroblasts (Leiden). We thank Shenyi Jiang
circuit-related dysfunction in SCA1 and provide important (LMU) for confocal microscopy and Angelina Oestmann (UBERN) for taking
insight into enhanced inhibitory drive onto PNs triggering mo- care of the animals. The study was supported by the European Research
lecular pathology, which induces PN degeneration. Our work Council under the European Union’s Horizon 2020 research and innovation
highlights brain state-dependent alterations within the cere- program (#725825), SAND: Marie Sk1odowska-Curie actions, Innovative
bellar circuit, and, to our knowledge, for the first time, it impli- Training Network, and E-rare grant (CALSER) to S.S, and German Research
Foundation under Germany’s Excellence Strategy within the framework of
cates early deficits in sensorimotor processing as a potential
the Munich Cluster for Systems Neurology -EXC 2145 SyNergy- ID
diagnostic hallmark of SCA1.
390857198, and the Emmy Noether Program to S. Liebscher.

STAR+METHODS AUTHOR CONTRIBUTIONS

Conceptualization and writing, F.P., S.S., and S. Liebscher; investigation and

Detailed methods are provided in the online version of this paper analyses, F.P., C.D., R.D., X.Y., Z.A.Q., A.O., C.T., M.M., and A.T.; FACS, mass
and include the following: spectrometry, and analyses, S.M., S. Lagache, A.C.U., and M.H.; reagents,
R.A.M.B., W.M.C.v.R.-M., and B.Z.; supervision, S.S. and S. Liebscher.
d KEY RESOURCES TABLE
d RESOURCE AVAILABILITY DECLARATION OF INTERESTS
B Lead contact
The authors declare no competing interests.
B Materials availability
B Data and code availability
INCLUSION AND DIVERSITY
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Mice strains One or more of the authors of this paper self-identifies as an underrepresented
d METHOD DETAILS ethnic minority in their field of research or within their geographical location.
One or more of the authors of this paper self-identifies as a gender minority
B Pharmacological treatments
in their field of research. One or more of the authors of this paper self-identifies
B Accelerating Rotarod
as a member of the LGBTQIA+ community.
B Hindlimb clasping assay
B DREADDs and genetically encoded calcium indicators Received: September 14, 2022
for in vivo imaging Revised: March 17, 2023
B Intracerebroventricular (i.c.v.) injection Accepted: May 17, 2023
B Surgical procedures in adult animals at UBERN Published: June 14, 2023

B Cranial window implantation and virus injection at LMU

REFERENCES
B Two-photon imaging in awake and anesthetized mice
B Ratiometric calcium imaging 1. Leroy, F., and Zytnicki, D. (2015). Is hyperexcitability really guilty in amyo-
B Chronic CNO treatment at LMU-UBERN trophic lateral sclerosis? Neural Regen. Res. 10, 1413–1415. https://doi.
B Calcium imaging with acute CNO administration org/10.4103/1673-5374.165308.
B Immunofluorescence of rodent tissue at UBERN 2. Roselli, F., and Caroni, P. (2015). From intrinsic firing properties to selec-
B Confocal microscopy tive neuronal vulnerability in neurodegenerative diseases. Neuron 85,
B SBF-SEM 901–910. https://doi.org/10.1016/j.neuron.2014.12.063.
B iPSC differentiation into GABAergic neurons (iGNs) 3. Mentis, G.Z., Blivis, D., Liu, W., Drobac, E., Crowder, M.E., Kong, L.,
Alvarez, F.J., Sumner, C.J., and O’Donovan, M.J. (2011). Early functional
B Immunofluorescence staining of iGNs
impairment of sensory-motor connectivity in a mouse model of spinal
B FACS of MLIN
muscular atrophy. Neuron 69, 453–467. https://doi.org/10.1016/j.
B Mass spectrometry of sorted MLIN neuron.2010.12.032.
B Calcium imaging of iGNs and lentiviral-mediated
4. Leroy, F., Lamotte d’Incamps, B., Imhoff-Manuel, R.D., and Zytnicki, D.
FRRS1l transduction (2014). Early intrinsic hyperexcitability does not contribute to motoneuron
d QUANTIFICATION AND STATISTICAL ANALYSIS degeneration in amyotrophic lateral sclerosis. eLife 3. https://doi.org/10.
B Microscopy and image analysis 7554/eLife.04046.
B Data analysis of mass spectrometry 5. Saxena, S., Roselli, F., Singh, K., Leptien, K., Julien, J.P., Gros-Louis, F.,
B In vivo two-photon calcium image processing and data and Caroni, P. (2013). Neuroprotection through excitability and mTOR
analysis required in ALS motoneurons to delay disease and extend survival.
Neuron 80, 80–96. https://doi.org/10.1016/j.neuron.2013.07.027.
B Population activity dimensionality and manifold
6. Barnes, J.A., Ebner, B.A., Duvick, L.A., Gao, W., Chen, G., Orr, H.T., and
analysis
Ebner, T.J. (2011). Abnormalities in the climbing fiber-Purkinje cell circuitry
B Statistical analysis
contribute to neuronal dysfunction in ATXN1[82Q] mice. J. Neurosci. 31,
12778–12789. https://doi.org/10.1523/JNEUROSCI.2579-11.2011.
SUPPLEMENTAL INFORMATION 7. Hansen, S.T., Meera, P., Otis, T.S., and Pulst, S.M. (2013). Changes in
Purkinje cell firing and gene expression precede behavioral pathology in
Supplemental information can be found online at https://doi.org/10.1016/j. a mouse model of SCA2. Hum. Mol. Genet. 22, 271–283. https://doi.org/
neuron.2023.05.016. 10.1093/hmg/dds427.

2540 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

8. Paulson, H.L., Shakkottai, V.G., Clark, H.B., and Orr, H.T. (2017). 24. Dell’Orco, J.M., Wasserman, A.H., Chopra, R., Ingram, M.A.C., Hu, Y.S.,
Polyglutamine spinocerebellar ataxias — from genes to potential treatments. Singh, V., Wulff, H., Opal, P., Orr, H.T., and Shakkottai, V.G. (2015).
Nat. Rev. Neurosci. 18, 613–626. https://doi.org/10.1038/nrn.2017.92. Neuronal atrophy early in degenerative ataxia is a compensatory mecha-
9. Braz, B.Y., Wennagel, D., Ratié, L., de Souza, D.A.R., Deloulme, J.C., nism to regulate membrane excitability. J. Neurosci. 35, 11292–11307.
Barbier, E.L., Buisson, A., Lanté, F., and Humbert, S. (2022). Treating early https://doi.org/10.1523/JNEUROSCI.1357-15.2015.
postnatal circuit defect delays Huntington’s disease onset and pathology in 25. Gurnani, H., and Silver, R.A. (2021). Multidimensional population activity in
mice. Science 377, eabq5011. https://doi.org/10.1126/science.abq5011. an electrically coupled inhibitory circuit in the cerebellar cortex. Neuron
109. 1739.e8–1753.e8. https://doi.org/10.1016/j.neuron.2021.03.027.
10. Pradhan, J., and Bellingham, M.C. (2021). Neurophysiological mechanisms
underlying cortical hyper-excitability in amyotrophic lateral sclerosis: a re- 26. Cayco-Gajic, N.A., and Silver, R.A. (2019). Re-evaluating circuit mecha-
view. Brain Sci. 11, 549. https://doi.org/10.3390/brainsci11050549. nisms underlying pattern separation. Neuron 101, 584–602. https://doi.
org/10.1016/j.neuron.2019.01.044.
11. Ramı́rez-Jarquı́n, U.N., Lazo-Gómez, R., Tovar-y-Romo, L.B., and Tapia,
R. (2014). Spinal inhibitory circuits and their role in motor neuron degener- 27. Lanore, F., Cayco-Gajic, N.A., Gurnani, H., Coyle, D., and Silver, R.A.
ation. Neuropharmacology 82, 101–107. https://doi.org/10.1016/j.neuro- (2021). Cerebellar granule cell axons support high-dimensional represen-
pharm.2013.10.003. tations. Nat. Neurosci. 24, 1142–1150. https://doi.org/10.1038/s41593-
021-00873-x.
12. Gunes, Z.I., Kan, V.W.Y., Ye, X., and Liebscher, S. (2020). Exciting
complexity: the role of motor circuit elements in ALS pathophysiology. 28. Gobbo, F., Mitchell-Heggs, R., and Tse, D. (2022). Changes in brain activ-
Front. Neurosci. 14, 573. https://doi.org/10.3389/fnins.2020.00573. ity and connectivity as memories age. Cogn. Neurosci. 13, 141–143.
https://doi.org/10.1080/17588928.2022.2076076.
13. Edamakanti, C.R., Do, J., Didonna, A., Martina, M., and Opal, P. (2018).
Mutant ataxin1 disrupts cerebellar development in spinocerebellar ataxia 29. Vyas, S., Golub, M.D., Sussillo, D., and Shenoy, K.V. (2020). Computation
type 1. J. Clin. Invest. 128, 2252–2265. https://doi.org/10.1172/JCI96765. through neural population dynamics. Annu. Rev. Neurosci. 43, 249–275.
https://doi.org/10.1146/annurev-neuro-092619-094115.
14. Ruegsegger, C., Stucki, D.M., Steiner, S., Angliker, N., Radecke, J., Keller,
€egg, M.A., and Saxena, S. (2016). Impaired mTORC1- 30. Chopra, R., Bushart, D.D., Cooper, J.P., Yellajoshyula, D., Morrison, L.M.,
E., Zuber, B., Ru
Huang, H., Handler, H.P., Man, L.J., Dansithong, W., Scoles, D.R., et al.
dependent expression of Homer-3 influences SCA1 pathophysiology.
(2020). Altered Capicua expression drives regional Purkinje neuron vulner-
Neuron 89, 129–146. https://doi.org/10.1016/j.neuron.2015.11.033.
ability through ion channel gene dysregulation in spinocerebellar ataxia
15. Duvick, L., Barnes, J., Ebner, B., Agrawal, S., Andresen, M., Lim, J., type 1. Hum. Mol. Genet. 29, 3249–3265. https://doi.org/10.1093/hmg/
Giesler, G.J., Zoghbi, H.Y., and Orr, H.T. (2010). SCA1-like disease in ddaa212.
mice expressing wild-type ataxin-1 with a serine to aspartic acid replace-
31. Wagner, M.J., Savall, J., Hernandez, O., Mel, G., Inan, H., Rumyantsev, O.,
ment at residue 776. Neuron 67, 929–935. https://doi.org/10.1016/j.
Lecoq, J., Kim, T.H., Li, J.Z., Ramakrishnan, C., et al. (2021). A neural cir-
neuron.2010.08.022.
cuit state change underlying skilled movements. Cell 184. 3731.e21–3747.
16. Ebner, B.A., Ingram, M.A., Barnes, J.A., Duvick, L.A., Frisch, J.L., Clark, e21. https://doi.org/10.1016/j.cell.2021.06.001.
H.B., Zoghbi, H.Y., Ebner, T.J., and Orr, H.T. (2013). Purkinje cell
32. Donato, F., Rompani, S.B., and Caroni, P. (2013). Parvalbumin-expressing
ataxin-1 modulates climbing fiber synaptic input in developing and adult
basket-cell network plasticity induced by experience regulates adult
mouse cerebellum. J. Neurosci. 33, 5806–5820. https://doi.org/10.1523/
learning. Nature 504, 272–276. https://doi.org/10.1038/nature12866.
JNEUROSCI.6311-11.2013.
33. Donato, F., Chowdhury, A., Lahr, M., and Caroni, P. (2015). Early- and late-
17. Kim, J., and Augustine, G.J. (2021). Molecular layer interneurons: key ele-
born parvalbumin basket cell subpopulations exhibiting distinct regulation
ments of cerebellar network computation and behavior. Neuroscience
and roles in learning. Neuron 85, 770–786. https://doi.org/10.1016/j.
462, 22–35. https://doi.org/10.1016/j.neuroscience.2020.10.008.
neuron.2015.01.011.
18. Vig, P.J., Fratkin, J.D., Desaiah, D., Currier, R.D., and Subramony, S.H.. 34. Wang, W.X., Qiao, J., and Lefebvre, J.L. (2022). PV-IRES-Cre mouse line
(1996). Decreased parvalbumin immunoreactivity in surviving Purkinje targets excitatory granule neurons in the cerebellum. Mol. Brain 15, 85.
cells of patients with spinocerebellar ataxia-1. Neurology 47, 249–253. https://doi.org/10.1186/s13041-022-00972-1.
https://doi.org/10.1212/wnl.47.1.249.
35. Vig, P.J.S., Subramony, S.H., Burright, E.N., Fratkin, J.D., McDaniel, D.O.,
19. Chen, T.W., Wardill, T.J., Sun, Y., Pulver, S.R., Renninger, S.L., Baohan, Desaiah, D., and Qin, Z. (1998). Reduced immunoreactivity to calcium-
A., Schreiter, E.R., Kerr, R.A., Orger, M.B., Jayaraman, V., et al. (2013). binding proteins in Purkinje cells precedes onset of ataxia in spinocerebel-
Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature lar ataxia-1 transgenic mice. Neurology 50, 106–113. https://doi.org/10.
499, 295–300. https://doi.org/10.1038/nature12354. 1212/WNL.50.1.106.
20. Dana, H., Sun, Y., Mohar, B., Hulse, B.K., Kerlin, A.M., Hasseman, J.P., 36. Ma, M., Futia, G.L., de Souza, F.M.S., Ozbay, B.N., Llano, I., Gibson, E.A.,
Tsegaye, G., Tsang, A., Wong, A., Patel, R., et al. (2019). High-perfor- and Restrepo, D. (2020). Molecular layer interneurons in the cerebellum
mance calcium sensors for imaging activity in neuronal populations and encode for valence in associative learning. Nat. Commun. 11, 4217.
microcompartments. Nat. Methods 16, 649–657. https://doi.org/10. https://doi.org/10.1038/s41467-020-18034-2.
1038/s41592-019-0435-6.
37. Cvetanovic, M., Ingram, M., Orr, H., and Opal, P. (2015). Early activation of
21. Raman, I.M., and Bean, B.P. (1999). Ionic currents underlying sponta- microglia and astrocytes in mouse models of spinocerebellar ataxia type
neous action potentials in isolated cerebellar Purkinje neurons. 1. Neuroscience 289, 289–299. https://doi.org/10.1016/j.neuroscience.
J. Neurosci. 19, 1663–1674. https://doi.org/10.1523/JNEUROSCI.19-05- 2015.01.003.
01663.1999.
38. Buijsen, R.A.M., Gardiner, S.L., Bouma, M.J., van der Graaf, L.M.,
22. Ramirez, J.E., and Stell, B.M. (2016). Calcium imaging reveals coordinated Boogaard, M.W., Pepers, B.A., Eussen, B., de Klein, A., Freund, C., and
simple spike pauses in populations of cerebellar Purkinje cells. Cell Rep. van Roon-Mom, W.M.C. (2018). Generation of 3 spinocerebellar ataxia
17, 3125–3132. https://doi.org/10.1016/j.celrep.2016.11.075. type 1 (SCA1) patient-derived induced pluripotent stem cell lines
€us, I., Mues, M., Russo, L., Dana,
23. Thestrup, T., Litzlbauer, J., Bartholoma LUMCi002-A, B, and C and 2 unaffected sibling control induced pluripo-
H., Kovalchuk, Y., Liang, Y., Kalamakis, G., Laukat, Y., et al. (2014). tent stem cell lines LUMCi003-A and B. Stem Cell Res. 29, 125–128.
Optimized ratiometric calcium sensors for functional in vivo imaging of https://doi.org/10.1016/j.scr.2018.03.018.
neurons and T lymphocytes. Nat. Methods 11, 175–182. https://doi.org/ 39. Brechet, A., Buchert, R., Schwenk, J., Boudkkazi, S., Zolles, G., Siquier-
10.1038/nmeth.2773. Pernet, K., Schaber, I., Bildl, W., Saadi, A., Bole-Feysot, C., et al. (2017).

Neuron 111, 2523–2543, August 16, 2023 2541

 ll
OPEN ACCESS Article
AMPA-receptor specific biogenesis complexes control synaptic transmis- 55. Sendhilnathan, N., Goldberg, M.E., and Ipata, A.E. (2022). Mixed selec-
sion and intellectual ability. Nat. Commun. 8, 15910. https://doi.org/10. tivity in the cerebellar Purkinje-cell response during visuomotor associa-
1038/ncomms15910. tion learning. J. Neurosci. 42, 3847–3855. https://doi.org/10.1523/
40. Kozareva, V., Martin, C., Osorno, T., Rudolph, S., Guo, C., Vanderburg, C., JNEUROSCI.1771-21.2022.
Nadaf, N., Regev, A., Regehr, W.G., and Macosko, E. (2021). A transcrip- 56. Dong, G.H., Xu, Y.H., Liu, L.Y., Lu, D., Chu, C.P., Cui, S.B., and Qiu, D.L.
tomic atlas of mouse cerebellar cortex comprehensively defines cell types. (2022). Chronic ethanol exposure during adolescence impairs simple
Nature 598, 214–219. https://doi.org/10.1038/s41586-021-03220-z. spike activity of cerebellar Purkinje cells in vivo in mice. Neurosci. Lett.
771, 136396. https://doi.org/10.1016/j.neulet.2021.136396.
41. Cendelin, J. (2014). From mice to men: lessons from mutant ataxic mice.
Cerebellum. Ataxias 1, 4. https://doi.org/10.1186/2053-8871-1-4. 57. Pérez Ortiz, J.M., and Orr, H.T. (2018). Spinocerebellar ataxia type 1: mo-
lecular mechanisms of neurodegeneration and preclinical studies,
42. Gaffield, M.A., and Christie, J.M. (2017). Movement rate is encoded and
pp. 135–145. https://doi.org/10.1007/978-3-319-71779-1_6.
influenced by widespread, coherent activity of cerebellar molecular layer
interneurons. J. Neurosci. 37, 4751–4765. https://doi.org/10.1523/ 58. Ju, H., Kokubu, H., and Lim, J. (2014). Beyond the glutamine expansion:
JNEUROSCI.0534-17.2017. influence of posttranslational modifications of ataxin-1 in the pathogenesis
of spinocerebellar ataxia Type 1. Mol. Neurobiol. 50, 866–874. https://doi.
43. Astorga, G., Li, D., Therreau, L., Kassa, M., Marty, A., and Llano, I. (2017).
org/10.1007/s12035-014-8703-z.
Concerted interneuron activity in the cerebellar molecular layer during
rhythmic oromotor behaviors. J. Neurosci. 37, 11455–11468. https://doi. 59. Kuo, S.H., Lin, C.Y., Wang, J., Sims, P.A., Pan, M.K., Liou, J.-Y., Lee, D.,
org/10.1523/JNEUROSCI.1091-17.2017. Tate, W.J., Kelly, G.C., Louis, E.D., et al. (2017). Climbing fiber-Purkinje
cell synaptic pathology in tremor and cerebellar degenerative diseases.
44. Heiney, S.A., Wohl, M.P., Chettih, S.N., Ruffolo, L.I., and Medina, J.F.
Acta. Neuropathol. 133, 121–138. https://doi.org/10.1007/s00401-016-
(2014). Cerebellar-dependent expression of motor learning during eye-
1626-1.
blink conditioning in head-fixed mice. J. Neurosci. 34, 14845–14853.
https://doi.org/10.1523/JNEUROSCI.2820-14.2014. 60. Shuvaev, A.N., Belozor, O.S., Mozhei, O.I., Shuvaev, A.N., Fritsler, Y.V.,
Khilazheva, E.D., Mosyagina, A.I., Hirai, H., Teschemacher, A.G., and
45. D’Angelo, E., and Casali, S. (2012). Seeking a unified framework for cere-
Kasparov, S. (2022). Indirect negative effect of mutant ataxin-1 on short-
bellar function and dysfunction: from circuit operations to cognition. Front.
and long-term synaptic plasticity in mouse models of spinocerebellar
Neural Circuits 6, 116. https://doi.org/10.3389/fncir.2012.00116.
ataxia type 1. Cells 11, 2247. https://doi.org/10.3390/cells11142247.
€usser, M.
46. Duguid, I., Branco, T., Chadderton, P., Arlt, C., Powell, K., and Ha
61. Hoxha, E., Balbo, I., Miniaci, M.C., and Tempia, F. (2018). Purkinje cell
(2015). Control of cerebellar granule cell output by sensory-evoked Golgi
signaling deficits in animal models of ataxia. Front. Synaptic Neurosci.
cell inhibition. Proc. Natl. Acad. Sci. USA 112, 13099–13104. https://doi.
10, 6. https://doi.org/10.3389/fnsyn.2018.00006.
org/10.1073/pnas.1510249112.
62. D’Adamo, M.C., Hasan, S., Guglielmi, L., Servettini, I., Cenciarini, M.,
47. D’Angelo, E., and De Zeeuw, C.I. (2009). Timing and plasticity in the cere- Catacuzzeno, L., and Franciolini, F. (2015). New insights into the patho-
bellum: focus on the granular layer. Trends Neurosci. 32, 30–40. https:// genesis and therapeutics of episodic ataxia type 1. Front. Cell.
doi.org/10.1016/j.tins.2008.09.007. Neurosci. 9, 317. https://doi.org/10.3389/fncel.2015.00317.
48. Arancillo, M., White, J.J., Lin, T., Stay, T.L., and Sillitoe, R.V. (2015). In vivo 63. Herson, P.S., Virk, M., Rustay, N.R., Bond, C.T., Crabbe, J.C., Adelman,
analysis of Purkinje cell firing properties during postnatal mouse develop- J.P., and Maylie, J. (2003). A mouse model of episodic ataxia type-1.
ment. J. Neurophysiol. 113, 578–591. https://doi.org/10.1152/jn. Nat. Neurosci. 6, 378–383. https://doi.org/10.1038/nn1025.
00586.2014.
64. Ferguson, B.R., and Gao, W.J. (2018). Pv interneurons: critical regulators
49. Serra, H.G., Duvick, L., Zu, T., Carlson, K., Stevens, S., Jorgensen, N., of E/I balance for prefrontal cortex-dependent behavior and psychiatric
Lysholm, A., Burright, E., Zoghbi, H.Y., Clark, H.B., et al. (2006). disorders. Front. Neural Circuits 12, 37. https://doi.org/10.3389/fncir.
RORa-mediated Purkinje cell development determines disease severity 2018.00037.
in adult SCA1 mice. Cell 127, 697–708. https://doi.org/10.1016/j.cell.
65. Hijazi, S., Heistek, T.S., Scheltens, P., Neumann, U., Shimshek, D.R.,
2006.09.036.
Mansvelder, H.D., Smit, A.B., and van Kesteren, R.E. (2020). Early restora-
50. Morino, H., Matsuda, Y., Muguruma, K., Miyamoto, R., Ohsawa, R., tion of parvalbumin interneuron activity prevents memory loss and
Ohtake, T., Otobe, R., Watanabe, M., Maruyama, H., Hashimoto, K., network hyperexcitability in a mouse model of Alzheimer’s disease. Mol.
et al. (2015). A mutation in the low voltage-gated calcium channel Psychiatry 25, 3380–3398. https://doi.org/10.1038/s41380-019-0483-4.
CACNA1G alters the physiological properties of the channel, causing spi-
66. Khademullah, C.S., Aqrabawi, A.J., Place, K.M., Dargaei, Z., Liang, X.,
nocerebellar ataxia. Mol. Brain 8, 89. https://doi.org/10.1186/s13041-015-
Pressey, J.C., Bedard, S., Yang, J.W., Garand, D., Keramidis, I., et al.
0180-4.
(2020). Cortical interneuron-mediated inhibition delays the onset of amyo-
51. Riess, O., Schöls, L., Bottger, H., Nolte, D., Vieira-Saecker, A.M., trophic lateral sclerosis. Brain 143, 800–810. https://doi.org/10.1093/
Schimming, C., Kreuz, F., Macek, M., Krebsová, A., Macek, M.S., et al. brain/awaa034.
(1997). SCA6 is caused by moderate CAG expansion in the alpha1A-
67. Morello, N., Schina, R., Pilotto, F., Phillips, M., Melani, R., Plicato, O.,
voltage-dependent calcium channel gene. Hum. Mol. Genet. 6, 1289–
Pizzorusso, T., Pozzo-Miller, L., and Giustetto, M. (2018). Loss of Mecp2
1293. https://doi.org/10.1093/hmg/6.8.1289.
causes atypical synaptic and molecular plasticity of parvalbumin-express-
52. Niewiadomska-Cimicka, A., Doussau, F., Perot, J.B., Roux, M.J., Keime, ing interneurons reflecting Rett syndrome–like sensorimotor defects.
C., Hache, A., Piguet, F., Novati, A., Weber, C., Yalcin, B., et al. (2021). eNeuro 5, 0086–18.2018. eNeuro. https://doi.org/10.1523/ENEURO.
SCA7 mouse cerebellar pathology reveals preferential downregulation of 0086-18.2018.
key Purkinje cell-identity genes and shared disease signature with SCA1
68. Palop, J.J., Jones, B., Kekonius, L., Chin, J., Yu, G.Q., Raber, J., Masliah,
and SCA2. J. Neurosci. 41, 4910–4936. https://doi.org/10.1523/
E., and Mucke, L. (2003). Neuronal depletion of calcium-dependent pro-
JNEUROSCI.1882-20.2021.
teins in the dentate gyrus is tightly linked to Alzheimer’s disease-related
53. Rajakulendran, S., Schorge, S., Kullmann, D.M., and Hanna, M.G. (2010). cognitive deficits. Proc. Natl. Acad. Sci. USA 100, 9572–9577. https://
Dysfunction of the Ca(V)2.1 calcium channel in cerebellar ataxias. F1000 doi.org/10.1073/pnas.1133381100.
Biol. Rep. 2. 69. Verret, L., Mann, E.O., Hang, G.B., Barth, A.M.I., Cobos, I., Ho, K.,
54. Pietrobon, D. (2010). CAv2.1 channelopathies. Pflugers Arch. 460, Devidze, N., Masliah, E., Kreitzer, A.C., Mody, I., et al. (2012). Inhibitory
375–393. https://doi.org/10.1007/s00424-010-0802-8. interneuron deficit links altered network activity and cognitive dysfunction

2542 Neuron 111, 2523–2543, August 16, 2023

 ll
Article OPEN ACCESS

in Alzheimer model. Cell 149, 708–721. https://doi.org/10.1016/j.cell. (2016). Elucidation of the behavioral program and neuronal network encoded
2012.02.046. by dorsal raphe serotonergic neurons. Neuropsychopharmacology 41,
70. Tai, C., Abe, Y., Westenbroek, R.E., Scheuer, T., and Catterall, W.A. 1404–1415. https://doi.org/10.1038/npp.2015.293.
(2014). Impaired excitability of somatostatin- and parvalbumin-expressing 83. Dirren, E., Towne, C.L., Setola, V., Redmond, D.E., Schneider, B.L., and
cortical interneurons in a mouse model of Dravet syndrome. Proc. Natl. Aebischer, P. (2014). Intracerebroventricular injection of adeno-associ-
Acad. Sci. USA 111, E3139–E3148. https://doi.org/10.1073/pnas. ated virus 6 and 9 vectors for cell type–specific transgene expression in
1411131111. the spinal cord. Hum. Gene Ther. 25, 109–120. https://doi.org/10.1089/
71. Liu, X., Kumar, V., Tsai, N.P., and Auerbach, B.D. (2021). Hyperexcitability hum.2013.021.
and homeostasis in fragile X syndrome. Front. Mol. Neurosci. 14, 805929. 84. Walton, J. (1979). Lead aspartate, an en bloc contrast stain particularly
https://doi.org/10.3389/fnmol.2021.805929. useful for ultrastructural enzymology. J. Histochem. Cytochem. 27,
72. Paeger, L., Pippow, A., Hess, S., Paehler, M., Klein, A.C., Husch, A., 1337–1342. https://doi.org/10.1177/27.10.512319.
€ning, J.C., and Kloppenburg, P. (2017). Energy imbalance
Pouzat, C., Bru
85. Rubio, F.J., Li, X., Liu, Q.R., Cimbro, R., and Hope, B.T. (2016).
alters Ca2+ handling and excitability of POMC neurons. eLife 6. https://
Fluorescence activated cell sorting (FACS) and gene expression analysis
doi.org/10.7554/eLife.25641.
of Fos-expressing neurons from fresh and frozen rat brain tissue. J. Vis.
73. Madeo, M., Stewart, M., Sun, Y., Sahir, N., Wiethoff, S., Chandrasekar, I., Exp. 54358. https://doi.org/10.3791/54358.
Yarrow, A., Rosenfeld, J.A., Yang, Y., Cordeiro, D., et al. (2016). Loss-of-
86. Braga-Lagache, S., Buchs, N., Iacovache, M.I., Zuber, B., Jackson, C.B.,
function mutations in FRRS1L lead to an epileptic-dyskinetic encephalop-
and Heller, M. (2016). Robust label-free, quantitative profiling of circulating
athy. Am. J. Hum. Genet. 98, 1249–1255. https://doi.org/10.1016/j.ajhg.
plasma microparticle (MP) associated proteins. Mol. Cell. Proteomics 15,
2016.04.008.
3640–3652. https://doi.org/10.1074/mcp.M116.060491.
74. Shaheen, R., Al Tala, S., Ewida, N., Abouelhoda, M., and Alkuraya, F.S.
(2016). Epileptic encephalopathy with continuous spike-and-wave during 87. Cox, J., and Mann, M. (2008). MaxQuant enables high peptide identifica-
sleep maps to a homozygous truncating mutation in AMPA receptor tion rates, individualized p.p.b.-range mass accuracies and proteome-
component FRRS1L. Clin. Genet. 90, 282–283. https://doi.org/10.1111/ wide protein quantification. Nat. Biotechnol. 26, 1367–1372. https://doi.
cge.12796. org/10.1038/nbt.1511.
75. Schwenk, J., Boudkkazi, S., Kocylowski, M.K., Brechet, A., Zolles, G., 88. UniProt Consortium (2019). UniProt: a worldwide hub of protein knowl-
Bus, T., Costa, K., Kollewe, A., Jordan, J., Bank, J., et al. (2019). An ER as- edge. Nucleic Acids Res. 47, D506–D515. https://doi.org/10.1093/nar/
sembly line of AMPA-receptors controls excitatory neurotransmission and gky1049.
its plasticity. Neuron 104. 680.e9–692.e9. https://doi.org/10.1016/j. 89. Silva, J.C., Gorenstein, M.V., Li, G.Z., Vissers, J.P.C., and Geromanos,
neuron.2019.08.033. S.J. (2006). Absolute quantification of proteins by LCMSE: a virtue of par-
76. Stoyas, C.A., Bushart, D.D., Switonski, P.M., Ward, J.M., Alaghatta, A., allel MS acquisition. Mol. Cell. Proteomics 5, 144–156. https://doi.org/10.
Tang, M.B., Niu, C., Wadhwa, M., Huang, H., Savchenko, A., et al. 1074/mcp.M500230-MCP200.
(2020). Nicotinamide pathway-dependent Sirt1 activation restores cal-
90. Huber, W., von Heydebreck, A., Su €ltmann, H., Poustka, A., and Vingron, M.
cium homeostasis to achieve neuroprotection in spinocerebellar ataxia
(2002). Variance stabilization applied to microarray data calibration and to
type 7. Neuron 105, 630–644.e9. https://doi.org/10.1016/j.neuron.2019.
the quantification of differential expression. Bioinformatics 18 (Suppl 1 ),
11.019.
S96–S104. https://doi.org/10.1093/bioinformatics/18.suppl_1.S96.
77. Krashes, M.J., Koda, S., Ye, C., Rogan, S.C., Adams, A.C., Cusher, D.S.,
91. Kammers, K., Cole, R.N., Tiengwe, C., and Ruczinski, I. (2015). Detecting
Maratos-Flier, E., Roth, B.L., and Lowell, B.B. (2011). Rapid, reversible
significant changes in protein abundance. EuPA Open Proteom. 7, 11–19.
activation of AgRP neurons drives feeding behavior in mice. J Clin Invest
https://doi.org/10.1016/j.euprot.2015.02.002.
121, 1424–1428. https://doi.org/10.1172/JCI46229.
78. Dana, H., Mohar, B., Sun, Y., Narayan, S., Gordus, A., Hasseman, J.P., 92. Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery
Tsegaye, G., Holt, G.T., Hu, A., Walpita, D., et al. (2016). Sensitive red pro- rate: a practical and powerful approach to multiple testing. J. R. Stat.
tein calcium indicators for imaging neural activity. Elife 5, e12727. https:// Soc. B Methodol. 57, 289–300. https://doi.org/10.1111/j.2517-6161.
doi.org/10.7554/eLife.12727. 1995.tb02031.x.
79. Kremer, J.R., Mastronarde, D.N., and McIntosh, J.R. (1996). Computer €bener,
93. Liebscher, S., Keller, G.B., Goltstein, P.M., Bonhoeffer, T., and Hu
visualization of three-dimensional image data using IMOD. J Struct Biol M. (2016). Selective persistence of sensorimotor mismatch signals in vi-
116, 71–76. https://doi.org/10.1006/jsbi.1996.0013. sual cortex of behaving Alzheimer’s disease mice. Curr. Biol. 26,
80. Keller, G.B., Bonhoeffer, T., and Hu €bener, M. (2012). Sensorimotor 956–964. https://doi.org/10.1016/j.cub.2016.01.070.
mismatch signals in primary visual cortex of the behaving mouse. 94. Scekic-Zahirovic, J., Sanjuan-Ruiz, I., Kan, V., Megat, S., De Rossi, P.,
Neuron 74, 809–815. https://doi.org/10.1016/j.neuron.2012.03.040. Dieterlé, S., Cassel, R., Jamet, M., Kessler, P., Wiesner, D., et al. (2021).
81. Watase, K., Weeber, E.J., Xu, B., Antalffy, B., Yuva-Paylor, L., Hashimoto, Cytoplasmic FUS triggers early behavioral alterations linked to cortical
K., Kano, M., Atkinson, R., Sun, Y., Armstrong, D.L., et al. (2002). A long neuronal hyperactivity and inhibitory synaptic defects. Nat. Commun.
CAG repeat in the mouse Sca1 locus replicates SCA1 features and reveals 12, 3028. https://doi.org/10.1038/s41467-021-23187-9.
the impact of protein solubility on selective neurodegeneration. Neuron 34, 95. Kato, S., Kaplan, H.S., Schrödel, T., Skora, S., Lindsay, T.H., Yemini, E.,
905–919. https://doi.org/10.1016/S0896-6273(02)00733-X. Lockery, S., and Zimmer, M. (2015). Global brain dynamics embed the mo-
82. Urban, D.J., Zhu, H., Marcinkiewcz, C.A., Michaelides, M., Oshibuchi, H., tor command sequence of Caenorhabditis elegans. Cell 163, 656–669.
Rhea, D., Aryal, D.K., Farrell, M.S., Lowery-Gionta, E., Olsen, R.H.J., et al. https://doi.org/10.1016/j.cell.2015.09.034.

Neuron 111, 2523–2543, August 16, 2023 2543

 ll
OPEN ACCESS Article

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
Mouse anti parvalbumin Swant Cat#235; RRID: AB_10000343
Goat anti parvalbumin Swant Cat#PVG214; RRID: AB_10000345
Mouse anti calbindin Abcam Cat#ab82812; RRID: AB_1658451
Rabbit anti calbindin Abcam Cat#ab108404; RRID: AB_10861236
Rabbit anti-pCaMKII-a Santa cruz Cat#sc-12886-R; RRID: AB_2067915
Rabbit anti-CaMKII-a Lifespan Bio Cat#1178-50; RRID: AB_968870
Goat anti mCherry Origene Cat#AB0081; RRID: AB_2333094
Rabbit polyclonal anti mCherry Abcam Cat#ab183628; RRID: AB_2650480
Rabbit anti VGluT1 SYSY Cat#135303; RRID: AB_887875
Rabbit anti VGAT SYSY Cat#131003; RRID: AB_887869
Rabbit anti HOMER-3 Origene Cat#TA308627
Rabbit anti HOMER-3 SYSY Cat#160303; RRID: AB_10804288
Mouse anti GFP ABCAM Cat#AB1218;
RRID: AB_298911
Mouse anti myc Cell signaling Cat# 2276;
RRID: AB_331783
Mouse anti Frrs1l Santa cruz Cat# sc-398692
Mouse anti GluR2 Millipore Cat#MAB397; RRID: AB_11212990
Chicken anti MAP2 Sigma/Millipore Cat#AB15452; RRID: AB_805385
rabbit anti GABA Sigma Cat#A2052; RRID: AB_477652
Rabbit anti KV1.2 Thermo Fisher Scientific Cat# PA5-77578; RRID: AB_2736054
Rabbit anti-GFAP Sigma Cat#G4546; RRID: AB_1840895
Rabbit anti-SORCS3 Novus Cat##NBP1-30615;
RRID: AB_2192266
Donkey anti-Goat IgG (H+L) Cross- Invitrogen, Thermo Catalog # A-11058
Adsorbed Secondary Antibody, Alexa Fisher Scientific
Fluor 594
Neurotrace 435/455 blue fluorescent Invitrogen, Thermo Catalog number: N21479
Nissl stain Fisher Scientific
Bacterial and virus strains
AAV2/1-hSyn-GCaMP7S-WPRE Dana et al.20 Addgene: 104487-AAV1
AAV2/1-hSyn-DIO-mCherry Unpublished, a gift from Addgene: 50459-AAV1
Bryan Roth to Addgene
AAV2/2-hSyn-DIO-mCherry Unpublished, a gift from Addgene: 50459-AAV2
Bryan Roth to Addgene
AAV2/8-hSyn-DIO-mCherry Unpublished, a gift from Addgene: 50459-AAV8
Bryan Roth to Addgene
AAV2/8-hSyn-DIO-hM3D(Gq)-mCherry Krashes et al.77 Addgene: 44361-AAV8
AAV2/8-hSyn-DIO-hM4D(Gi)-mCherry Krashes et al.77 Addgene: 44362-AAV8
AAV2/2-hSyn-DIO-hM4D(Gi)-mCherry Unpublished, a gift from Addgene: 50475-AAV2
Bryan Roth to Addgene
AAV2/1-hSyn-GCaMP6s-WPRE-SV40 Chen et al.19 Addgene: 100843-AAV1
AAV2/1-Syn-NES-jRCaMP1a-WPRE-SV40 Dana et al.78 Addgene: 100848-AAV1
AAV2/1-hSyn-Twitch2B-WPRE-SV40 Thestrup et al.23 Addgene: 100040- AAV1
(Continued on next page)

e1 Neuron 111, 2523–2543.e1–e10, August 16, 2023

 ll
Article OPEN ACCESS

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
C9orf4 (FRRS1L) - Human shRNA lentiviral Origene: CAT# TL314264V
particles
C9ORF4 (FRRS1L) - Human Myc-DDK- Origene: CAT#: RC221300L1V
tagged lentiviral particles
Chemicals, peptides, and recombinant proteins
SB 431542 Stem cell technologies Cat#72234
LDN193189 Stem cell technologies Cat#72147
CHIR99021 Stem cell technologies Cat#72052
L-Ascorbic Acid Sigma Aldrich Cat#A4403
SAG Stem cell technologies Cat#73412
BDNF Stem cell technologies Cat#78005
DAPT Stem cell technologies Cat#72082
GDNF Stem cell technologies Cat#78058
IGF-1 Stem cell technologies Cat#78078
CNTF Stem cell technologies Cat#78010
Clozapine N-Oxide Tocris Cat#4936
Experimental models: Cell lines
Human induced pluripotent Buijsen et al.38 LUMCi002
stem cell (SCA1)
Human induced pluripotent Buijsen et al.38 LUMCi003
stem cell (Healthy)
Human induced pluripotent Leiden University Medical Center LUMCi034
stem cell (SCA1)
Human induced pluripotent Leiden University Medical Center LUMCi035
stem cell (Healthy)
Human induced pluripotent Leiden University Medical Center LUMCi022
stem cell (SCA1)
Human induced pluripotent Leiden University Medical Center LUMCi023
stem cell (Healthy)
Experimental models: Organisms/strains
B6.129S-Atxn1tm1Hzo/J Jackson laboratories RRID: IMSR_JAX:005601
B6;129P2-Pvalbtm1(cre)Arbr/J Jackson laboratories RRID: IMSR_JAX:008069
B6.129-Tg(Pcp2-cre)2Mpin/J Jackson laboratories RRID: IMSR_JAX:004146
Software and algorithms
Fiji https://doi.org/10.1038/nmeth.2019 RRID:SCR_002285
Imaris Oxford instruments RRID:SCR_007370
iMOD Kremer et al.79 RRID:SCR_003297
Graphpad prism http://www.graphpad.com/ RRID:SCR_002798
MATLAB R2018a version The MathWorks, Inc RRID:SCR_001622
https://de.mathworks.com/
Custom built software on Labview 2018 Keller et al.80 https://www.ni.com/de-de/shop/
SP1, 2020 SP1, National Instruments National Instruments software/products/labview.html
SciScan 1.4 - Imaging acquisition software Scientfica UK N/A
Other
IR camera Imaging source DMK 22BUC03
Ti:Sapphire laser with a DeepSee Spectra Physics MaiTai HP N/A
pre-chirp unit
Two-photon microscope, equipped Hyperscope, Scientifica N/A
with an 8 kHz resonant scanner
316 water-immersion objective (0.8 NA) Nikon N/A
(Continued on next page)

Neuron 111, 2523–2543.e1–e10, August 16, 2023 e2

 ll
OPEN ACCESS Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
PMT FEMTO DHPCA-100
Emission filter G 525/50 nm,
C/Y 460/80nm (CFP) and
560/80 nm (YFP)

RESOURCE AVAILABILITY

Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Smita
Saxena ([email protected]).

Materials availability
This study did not generate new unique reagents.

Data and code availability

d All data reported in this paper will be shared by the lead contact upon request.
d This paper does not report original code.
d Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice strains
The Atxn1154Q/2Q knock-in mice (B6.129S-Atxn1tm1Hzo/J),81 called Sca1 throughout the manuscript, PV-IRES- cre mice (B6;129P2-
Pvalbtm1(cre)Arbr/J) and Pcp2-cre (B6.129-Tg(Pcp2-cre)2Mpin/J) were obtained from Jackson Laboratory. To obtain Sca1::PV or
Sca1::Pcp2 animals, Atxn1154Q/2Q males were crossed with PV-cre or Pcp2-cre homozygous females. Animals of both genders
were equally used in the study. Mice were kept at a 12/12-hour light/dark cycle with ad libitum access to food and water and
were group housed. We regularly performed long-range PCR on Sca1 mice, confirming the length of PolyQ repeats in both colonies
as being around 154 repeats. Animal care, housing, ethical usage, and procedures were approved by the government of upper Ba-
varia and in accordance with the Swiss Veterinary Law guidelines.

METHOD DETAILS

Pharmacological treatments
Clozapine-N-oxide (Tocris, 4936) was administered IP at a dosage of 3 mg/Kg, 45 minutes prior to behavioral test (for acute MLIN
modulation) or in drinking water 40 mg/ml82 for chronic modulation of MLIN expressing the inhibitory G protein hM4D(Gi) or hM3D(Gq).
All CNO treatments were independently performed at UBERN and LMU.

Accelerating Rotarod
Rotarod test was performed between 1 pm and 6 pm. Experimental mice were habituated to the experimenter and trained on the
rotating rod at fixed speed of 5 rpm for three days prior to the first rotarod recording. The protocol consisted of 4 trials per day, in
which the rotating rod accelerated from 5 to 40 rpm within 300s, followed by 300s rest in between trials. To avoid experimenter
bias, rotarod for chronic DREADD(Gi) experiments were performed independently both at the LMU and UBERN. Experimenters
were blinded for treatment condition throughout the study and data was merged and presented together.

Hindlimb clasping assay

The tails of the mice were grabbed at the base and mice were lifted for 20 seconds. If both hind limbs were consistently splayed out-
ward, away from the abdomen it was assigned a score of ‘‘0’’. If only one of the hindlimbs was retracted near the abdomen for the
majority of the time, the score assigned was ‘‘1’’. If both hindlimbs are retracted close to the abdomen, the animal receives a score of
‘‘2’’. Finally, if both hindlimbs were tightly clinched to the abdomen the score assigned was ‘‘3’’. Clasping phenotype for chronic
DREADD(Gi) experiments was measured independently at the LMU and UBERN, data was merged and presented together.

DREADDs and genetically encoded calcium indicators for in vivo imaging

Cre dependent DREADDs and genetically encoded calcium indicators were purchased from Addgene: AAV2/1-hSyn-DIO-mCherry
(50459) viral titer 731012 vg/mL, AAV2/2-hSyn-DIO-mCherry (50459) viral titer 431012 vg/mL, AAV2/8-hSyn-DIO-mCherry

e3 Neuron 111, 2523–2543.e1–e10, August 16, 2023

 ll
Article OPEN ACCESS

(50459) viral titer 131013 vg/mL, AAV2/8-hSyn-DIO-hM4D(Gi)–mCherry (44362) viral titer 131013 vg/mL, AAV2/8-hSyn-DIO-
hM3D(Gq)–mCherry (44361) viral titer 431012 vg/mL and AAV2/8-hSyn-DIO-hM4D(Gi)–mCherry (50475) 731012 vg/mL. For calcium
imaging: AAV2/1-hSyn-GCaMP6s-WPRE-SV40 (100843) viral titer 131013 vg/mL; AAV2/1-hSyn-jGCaMP7s-WPRE (104487) viral
titer 131013 vg/mL. AAV2/1-hSyn1-Twitch2B-WPRE.SV40 (100040) viral titer 131013 vg/mL.

Intracerebroventricular (i.c.v.) injection

Injection of AAV particles in neonates (P0-P2) was performed by unilateral injection of 1.0 ml of AAV2/8-hSyn-DIO-hM4D(Gi)–
mCherry. The viral particles were injected in the left lateral ventricle, as described previously.83 A 0.1% fast green solution was added
to the vector suspension, in order to visualize the spread of the virus.

Surgical procedures in adult animals at UBERN

To induce analgesia, animals were injected IP with 0.1 mg/kg Buprenorphine 20 minutes prior to the surgical procedure. Anesthesia
was induced using an induction chamber with 5% isoflurane, animals were then placed onto the stereotaxic frame and anesthesia
was maintained at 1.5-2% isoflurane throughout the surgical procedure. Body temperature was maintained using a heating pad, and
a lubricating eye cream was applied to avoid dehydration of the corneas. The head of the animal was fixed with ear bars. Under a
dissecting microscope (Olympus, Tokyo, Japan) a medial skin incision was performed to expose the skull, the incision was extended
in order to expose bregma and lambda to align the skull. In order to expose the occipital part of the cerebellum, muscles were
cut over the medial line and muscles fibers were gently opened to access the injection site. A drill was used to make two holes
in the occipital bone. For DREADDs-mediated circuit modulation experiments, the following coordinates were used to inject
AAV2/8-hSyn-DIO-hM4D(Gi)–mCherry, AAV2/8-hSyn-DIO-hM3D(Gq)–mCherry and AAV2/8- hSyn-DIO-mCherry: AP -8.25mm, DV
-2.5mm, ML +/- 1mm, angle 57  , depth 0.3-0.4 mm from the dura. For the FACS sorting of MLIN the following coordinates were
used: AP - 8.25mm, DV -2.5mm, ML 0 and /+/-1 mm, angle 57  , depth 0.3-0.4 mm from the dura; AP -7.25mm, DV -1 mm, ML
0 and /+/-1.5 mm, angle 0  , depth 0.3-0.4 mm from the dura. A Hamilton syringe equipped with a glass pipette was used to infuse
300nl (or 500nl for FACS sorting) of virus per injection site at an infusion rate of 50nl/minute. After the completion of the injection, the
glass pipette was left in place for few minutes before withdrawal. Following the glass pipette withdrawal, the muscles and the skin
were sutured. A subcutaneous injection of warm saline was applied, and animals were placed in a pre-warmed cage and monitored
until fully awake and recovered from the anesthesia.

Cranial window implantation and virus injection at LMU

Mice of both sexes were implanted with a cranial window at P42 ± 6.70 (mean ± SD, P60 cohort) for imaging early symptomatic mice
and at P174 ± 2.91 for late symptomatic mice (P200 cohort). They received a stereotaxic injection of AAV2/1-hSyn-jGCaMP7s-
WPRE, diluted 1:10 in saline, Addgene viral prep # 104487-AAV1; RRID: Addgene_104487-AAV1 into lobule V (5Cb) & VI (6Cb) of
the cerebellar vermis. Mice were given Meloxicam (10 mg/kg) and Metamizol (200 mg/kg) 30 minutes prior to surgery, orally and
were then anesthetized with Fentanyl (0.05 mg/kg), Midazolam (5.0 mg/kg), and Metedomidin (0.5 mg/kg) given intraperitoneally.
A circular craniotomy of 3 mm diameter was centered at 1.5 mm lateral from the midline and -6.5 mm posterior to bregma overlying
lobule V & VI and injections were made at two sites at -6.5mm anterior/posterior (AP) and 0.5 mm lateral and -6.8 mm AP and 1mm
lateral from the midline (Figure 1B), with 200nl AAV injected at a depth of both 500mm and 150mm from the cortical surface at each site,
using a flow rate of 50nl/min. A 3 mm round glass coverslip (Warner Instruments) was placed over the craniotomy and sealed flush to
the surrounding skull using UV-curable dental acrylic (Venus Diamond Flow, Heraeus Kulzer GmbH). To allow head fixation during
two-photon imaging, a custom metal head bar was then attached onto the skull using dental acrylic (Paladur, Heraeus Kulzer
GmbH). At the end of the procedure the anesthesia was antagonized through application of Atipamezol (2.5 mg/kg), Flumazenil
(0.5 mg/kg) und Naloxon (1.2 mg/kg, all i.p.) and Meloxicam (10 mg/kg) was then administered orally at 12, 24, 36-, 48-, 60- and
72-hours post-op.

Two-photon imaging in awake and anesthetized mice

Early symptomatic mice were imaged at P63 ± 3.85 (mean ± SD), and late symptomatic mice at P204 ± 3.60. To habituate the mice,
each was given a minimum of three training sessions of 20 minutes each across the 7 days prior to first imaging time point, allowing
them to become comfortable running head restricted on the air-supported treadmill and engaging with the virtual reality environment.
In vivo two-photon imaging was performed using a two-photon microscope (Hyperscope, Scientifica, equipped with an 8 kHz reso-
nant scanner) at a frame rate of 30 Hz and a resolution of 512 3 512 pixels. Light source was a Ti:Sapphire laser with a DeepSee pre-
chirp unit (Spectra Physics MaiTai eHP). GCaMP7s was excited at 910 nm, with a laser power around 30–70 mW and emitted photons
detected by a GaAsP PMT with a bandpass filter in front (525/50 nm). Using a 16x water-immersion objective (Nikon), stacks consist-
ing of 15,000 frames (equivalent to 8 min) were acquired covering a field of view (FOV) of 453 x 453mm to simultaneously image mo-
lecular layer interneurons, Purkinje neurons and Golgi cells. Two-three FOVs at this resolution were imaged per mouse, each under
three distinct conditions, (1) in a virtual reality environment depicting a linear track with patterned walls consisting of lines and dots of
contrasting colour, providing visual feedback for locomotion (fb), (2) in darkness (dark), and (3) under light isoflurane anesthesia (iso).
Mice were anesthetized with isoflurane initially at a volume of 2-2.5 Vol % in pure O2 at a flow rate of 0.5 l/min. This was gradually
reduced to 1.0-1.5 Vol % over a minimum of 30 minutes induction period in order to establish a respiratory rate of between 110-130

Neuron 111, 2523–2543.e1–e10, August 16, 2023 e4

 ll
OPEN ACCESS Article

breaths per minute, which was then maintained throughout the recording. A physiological monitoring system (Harvard Apparatus)
was used to ensure body temperature was stable at 37 degrees. Several behavioral parameters were recorded synchronized with
the imaging data acquisition. As such the speed of the mouse on the ball was tracked by an optical mouse sensor (Logitech
G500s Laser Gaming Mouse) placed in front of the styrofoam ball. Pupil position and width as well as whisking events were detected
in video recordings acquired with an infrared camera (The Imaging Source, DMK 22BUC03, USB 2.0 monochrome industrial camera),
positioned to the front right of the mouse to visualize the right eye and whiskers on this side. Information of pupil position and pupil
width were derived online using custom-build software (National Instruments). Pupil width was computed post hoc using a custom-
written script (Math Works). Whisking was tracked by superimposing a detection window over mouse’s snout area. Brief puffs of
pressurized air were applied through a thin tube positioned at the right side of the mouse’s body.

Ratiometric calcium imaging

A circular craniotomy was performed at 3 mm diameter, centered at 1.5 mm lateral from the midline and -6.5 mm posterior to bregma
on mice aged P42 (± 7.27). Overall surgical procedure is as described above. Stereotaxic injections of AAV2/1-hSyn-Twitch-2B-
WPRE (containing the two fluorescent proteins mCerulean3 and cpVenusCD as a FRET pair) at two sites of -6.5mm anterior/posterior
(AP) and 0.5 mm lateral and -6.8 AP and 1mm lateral from the midline, with 200nl AAV injected at a depth of both 500mm and 150mm
from the cortical surface at each site (50nl/min, diluted 1:3 in saline). The following recovery time and habituation for imaging exper-
iments was consistent as initially described above. Twitch-2B was excited at 860 nm, with a laser power around 30-70 mW, emitted
photons were detected via two PMTs (passing through bandpass filters: 460/80nm (CFP) and 560/80 nm (YFP)) using identical
voltage settings (800 for the blue channel, 830 for the yellow channel) for all animals. Stacks consisted of 15,000 frames and were
acquired with a FOV of 453 x 453mm to simultaneously image molecular layer interneurons, Purkinje neurons and Golgi cells. Imaging
conditions used were darkness and subsequent isoflurane anesthesia, as described above. Imaging experiments were conducted at
ages P73 (± 8.20).
Imaging data was processed in a similar fashion as the GCaMP7s recordings, including motion correction and semimanual seg-
mentation to identify regions of interest. The calcium trace here corresponds to the ratio between the yellow channel and the cyan
channel (Y/C ratio).

Chronic CNO treatment at LMU-UBERN

To investigate the effects of chronic MLIN inhibition/activation, we either injected inhibitory or excitatory DREADD or control mCherry
AAV, diluted 1:6 in saline into the cerebellar vermis at the age of P40-45. For induction of anesthesia and application of analgesics, the
surgery protocol was as described for imaging experiments. Two small holes (approximately 300mm diameter) were drilled on either
side of the midline to allow injection via micropipette at the following coordinates AP -8.25mm, DV 2.5, ML +/-1mm. Injections were
made at a 57 angle to a 0.3-0.4 depth from touching the dura surface (rate 50nl/min). Clozapine-n-oxide (CNO, 40mg/ml) was admin-
istered through the drinking water with 1% sucrose, to mask the CNO taste. Mice were given 1% sucrose water 3 days prior to start of
treatment to get used to the taste, every 3 days throughout the treatment CNO solution was prepared fresh, mice were weighed and
assessed for any physical impairments. Rotarod (as described above/5-40 RPM, 300 seconds) was performed prior to treatment
start, with 4 trials (300 second intervals) over 3 days, then 4 trials on 2 consecutive days as depicted in main figure schemes. Handling
of mice (minimum twice per week) and habituation around rotarod (minimum once per week) was performed regularly throughout.

Calcium imaging with acute CNO administration

All surgical procedures were performed as described above in WT::PV mice and Sca1::PV mice at P73 (± 4.46). A circular craniotomy
of 3 mm diameter, centered at 1.5 mm lateral from the midline and -6.5 mm posterior to bregma was drilled. Injections of AAV2/1-
hSyn-GCaMP7s-WPRE were then made at two sites of -6.5mm anterior/posterior (AP) and 0.5 mm lateral and -6.8 mm AP and 1mm
lateral from the midline, with 200nl AAV injected at a depth of both 500mm and 150mm from the cortical surface at each site (50nl/min,
diluted 1:10 in saline). Injections of AAV2/8-hSyn-DIO-hM4Gi-WPRE were also made, bilaterally at -6.8 AP, +/-1mm from the midline,
150mm from the cortical surface at each site. And again, bilaterally at AP -8.25mm, DV 2.5, ML +/-1mm. The latter two injections at AP
-8.25mm were made at a 57 angle to a 0.3-0.4 mm depth from touching the dura surface (50nl/min, diluted 1:3 in saline).
Mice were then imaged at P110 (± 5.26) (habituation protocol as described above) and then the same neurons were again imaged
at P114 (± 4.47) 45 minutes following an intraperitoneal injection of clozapine-N-oxide (CNO, 3mk/kg). Mice were imaged as
described above, using the same imaging equipment, however we only assessed neuronal activity during feedback as well as under
isoflurane anesthesia (isoflurane 1.0-1.5 Vol % at a flow rate of 0.5 l/min to establish respiratory rates between 110-130 breaths per
minute, induction process and physiological monitoring as described in initial two photon imaging experiments).
For control acute CNO experiments without injection of DREADDs, Sca1 mice at age P80 underwent the exact same surgical pro-
cedures with viral injections AAV2/1-hSyn-GCaMP7s-WPRE and AAV2/8-hSyn-mCherry. The mice were then imaged at P120 and
P121 for baseline and CNO experiments respectively following procedures and conditions as the DREADDs mice.

Immunofluorescence of rodent tissue at UBERN

Mice were transcardially perfused with 4% paraformaldehyde (PFA) in 1X PBS at UBERN, while in LMU animals were transcar-
dially perfused with 1x PBS 0.05% Heparin, followed by 4% PFA; the cerebellum was isolated and kept overnight at 4 C in

e5 Neuron 111, 2523–2543.e1–e10, August 16, 2023

 ll
Article OPEN ACCESS

the same fixative solution, followed by 30% sucrose in PBS for cryoprotection until samples were used. After embedding in Tissue
Tek O.C.T compound (Bio system, 4583), cerebellar sections (50 mm) were sagitally sliced while brain was sliced coronally (50 mm)
using a cryostat. Antibodies used for immunofluorescence were: mouse anti-Parvalbumin (1:1000, Swant, 235), goat anti-Parval-
bumin (1:1000, Swant, PVG213), mouse anti-Calbindin (1:1000, Abcam, ab82812), rabbit anti-Calbindin (1:1000, Abcam,
ab108404), rabbit anti-pCaMKII-a, (1:200, Santa cruz, sc-12886), rabbit anti-CaMKII-a, 1:500 (LSB1178, Lifespan Bio), goat
anti-mCherry (1:1000, Origene, AB0081-200), rabbit anti-mCherry (1:1000, Abcam, ab183628), rabbit anti-VGluT1 (1:1000,
SYSY, 135303), rabbit anti-VGAT (1:1000, SYSY, 131003), rabbit anti-Homer-3 (1:500, Origine, TA308627), rabbit anti-Homer-3
(1:1000, SYSY, 160303) mouse anti-Frrs1l (1:100, Santa cruz, sc-398692), mouse anti-GluR2 (1:500, Millipore, MAB397). Sections
were kept for 2h in PBS solution containing 0.05% Triton X-100 and 10% normal donkey serum (NDS, Jackson immunoresearch,
017-000-121), after the antibodies were applied in PBS, 3% NDS, 0.05% Triton X-100, and incubated overnight at 4 C. Sections
were then briefly washed with PBS and incubated for 120 min at room temperature (RT), with appropriate combinations of sec-
ondary antibodies from Invitrogen.

Confocal microscopy
Confocal images were acquired using an Olympus Fluoview 1000-BX61 (Olympus, Tokyo) microscope, fitted with a 20X, 40X air
objective or 60X immersion oil objective. For the analysis of intensities, data were acquired using identical confocal settings, with
signals at the brightest cells being non-saturated.

SBF-SEM
Anesthetized mice were transcardially perfused first with 0.1 M PBS (pH=7.4) and then fixed with fixation solution (2.5% glutaralde-
hyde (GA) +2% PFA in 0.1 M Na-cacodylate, pH=7.4). Cerebellum was isolated and post-fixed in fixation solution. Bloc staining,
dehydration and embedding were performed as follows: post-fixation in 0.15 M cacodylate buffer, 1.5% potassium ferrocyanide
and 2% osmium tetroxide, followed by sample incubation with 0.64 M pyrogallol for 15min at room temperature and for 5min in a
water bath at 50  C, and subsequently rinsed with water. Samples were incubated in 2% OsO4 for 22min at room temperature
and 8min in a water bath at 50 C. After water rinses, the samples were incubated overnight in a solution of 0.15 M gadolinium acetate
(LFG Distribution, Lyon, France) and 0.15 M samarium acetate (LFG Distribution) pH 7.0. Followed by water rinses, samples were
incubated in 1% Walton’s lead aspartate,84 particularly useful for ultra-structural enzymology12 at 60 C for 30min. After staining,
the samples were dehydrated in a graded ethanol series (20%, 50%, 70%, 90%, 100%, 100%) at 4 C, each step last for 5min.
The blocks were infiltrated with Durcupan resin mixed with ethanol at ratios of 1:3 (v/v), 1:1, and 3:1, each step lasted for 2h, to finally
be infiltrated with pure Durcupan overnight. The samples were transferred to fresh Durcupan and the resin was polymerized for
3 days at 60 C. Care was taken to have osmicated material directly exposed at the block surface in contact with the glue in order
to reduce specimen charging under the electron beam. Pyramids with a surface of approximately 500 3 500 mm2 were trimmed
with a razor blade. The blocs were imaged at the SBF-SEM using a Quanta FEG 250 (FEI, Eindhoven, The Netherlands), equipped
with a 3View2XP in-situ ultramicrotome (Gatan). Images were acquired in low or high vacuum according to block quality. Analysis was
performed using Fiji, were axosomatic synapses were counted within different PNs soma. For 3D representative images iMOD soft-
ware was used and PN, MLIN and synapses were manually ROI to reconstruct the final model (around 100 and 280 consecutive im-
ages were used for WT and Sca1 models).

iPSC differentiation into GABAergic neurons (iGNs)

The Control and SCA1 iPSCs were obtained from Dr. Ronald A.M. Buijsen, and Dr. Willeke M.C. van Roon-Mom. Leiden University
Medical Center (LUMC), Leiden, The Netherlands. iPSCs were cultured in GeltrexTM (ThermoFischer) coated plates in mTeSRTM1
(StemCell technologies) media. Briefly, human iPSCs were dissociated to single cells using Accutase (StemCell technologies) and
seeded at 3X106 onto 10 cm plate with N2B27 differentiation medium (Advanced (DMEM/F12), Neurobasal (1:1) medium, 1%
Pen/strep (Gibco), 1% GlutaMAX (Gibco), 0.1mM 2- mercaptoethanol (Gibco), 1X B27 supplement (Gibco), 1X N2 supplement
(Gibco), supplemented with 10ng/ml basic fibroblast growth factor ((StemCell technologies), 20 mM SB431542 (StemCell technolo-
gies), 0.1 mM LDN193189 (StemCell technologies), 3 mM CHIR99021 (StemCell technologies), 10 mM L-Ascorbic Acid (L-AA; Sigma)
and 1X Revitacell supplement (Gibco)) to initiate the formation of embryoid bodies (EBs). On day 2, media patterning of EBs was
induced by the addition of media supplemented with 500 nM Smoothened Agonist (SAG; StemCell technologies). EBs were pelleted
and fed with fresh media on every alternate day until day 14. 10 ng/ml Brain derived neurotrophic factor (BDNF; StemCell technol-
ogies) was added from day 7 while 10 mM DAPT (StemCell technologies) was added from day 9. EBs were dissociated using trypsin
on day 16 and triturated with ice cold cell trituration and wash medium (1X PBS (Gibco), 0.45% Glucose, 0.1% Bovine Serum Albumin
(BSA; Sigma), 2mM MgCl2, 0.8mM EDTA (Invitrogen), 2.5% Fetal Bovine Serum (FBS; Sigma), 1X N2 supplement, 1X B27 supple-
ment and DNAse). Triturated EBs were then plated on poly-ornithine/laminin (Sigma) coated plates in iGNs feeding medium (Neuro-
basal medium (Gibco), 1X glutaMAX, 1X Non-essential amino acid (NEAA, Gibco), 0.1mM 2- mercapthoethanol, 1X N2 supplement,
1X Pen/strep, 1X B27 supplement, 10ng/ml glial cell derived neurotrophic factor (GDNF; StemCell technologies), BDNF 10ng/ml,
10ng/ml insulin- like growth factor (IGF-1; StemCell technologies), 10ng/ml Ciliary neurotrophic factor (CNTF; StemCell technologies)
and 10 mM AA and kept at incubator at 37 C and 5% CO2 for further maturation up to two weeks.

Neuron 111, 2523–2543.e1–e10, August 16, 2023 e6

 ll
OPEN ACCESS Article

Immunofluorescence staining of iGNs

iGNs plated on coverslips were fixed using 4% PFA for 15min and blocked for 1h with 3% bovine serum albumin (BSA) and 0.1 %
TritonX-100 in phosphate buffered saline PBS. After blocking, neurons were incubated with the following primary antibodies: mouse
anti- Parvalbumin (1:1000, Swant, 235), goat anti-Parvalbumin (1:1000, Swant, PVG213), chicken anti-MAP2 (1:1000, Sigma,
AB15452), rabbit anti-GABA (1:500, Sigma, A2052), mouse anti- GFAP (1:500, Sigma, G4546), mouse anti-Frrs1l (1:100, Santa
cruz, sc-398692), mouse anti- GluR2 (1:500, Millipore, MAB397), mouse anti-myc (1:1000, Cell Signaling, 2276), mouse anti-GFP
(1:1000, ABCAM, AB1218), rabbit anti-SORCS3 (1:500, Novus biologicals, NBP1-30615) in blocking buffer overnight at 4 C. After
washing three times with PBS, cells were incubated in blocking buffer with Alexa Fluor fluorescently labeled secondary antibodies
and DAPI for 1h at room temperature. Cells were then washed with PBS and mounted on glass slides. Images were acquired with a
confocal microscope Olympus FluoViewTM FV1000 (Olympus) fitted with a 20X or 40x air objective and 60x immersion oil objective.

FACS of MLIN
Protocol for isolation of MLIN from adult animals was performed as in Rubio et al.,85 with some optimization. Animals were deeply
anesthetized with isoflurane and decapitated, the cerebellum was extracted and placed on an ice-cold petri dish. A drop of Hibernate
A media (Gibco), supplemented with 1X B27 (Gibco) was used to cover the fresh cerebellum and a razor blade was used to mince the
tissue. The tissue was then centrifuged at 110x g for two minutes at 4 . The supernatant was discarded and 1 ml of Accutase was
added to resuspend the pellet. Tubes were incubated at 4 C for 30 minutes with hand-over-end mixing followed by centrifugation at
960x g for 2 minutes at 4 C. Supernatant was discarded and pellet resuspended in 0.6 ml of Hibernate A + B27 media. Single cells
were obtained by mechanical trituration using a glass pipette. Samples were triturated 10 times and let settle in ice to collect the su-
pernatant and the pellet was resuspended again in 0.6ml of Hibernate A + B27 media. The procedure was repeated until  3 ml of
supernatant was collected. The collected supernatant was then filtered through 100 and 40 mm cell strainer and 2 additional ml of
Hibernate A + B27 media were added to the filtered supernatant. Hoechst was added to the single cell suspension and double pos-
itive neurons for mCherry and Hoechst were sorted using MoFlo ASTRIOS EQ (Beckman Coulter) in 300 ml of sterile DPBS. Samples
were centrifuged at 400x g for 15 minutes and  250 ml of DPBS was discarded and sorted cells in approximately 50 mL DPBS were
frozen, then lyophilized, and stored at -80C until further use.

Mass spectrometry of sorted MLIN

The lyophilizate was re-suspended in 10 mL 8M Urea / 100mM Tris-HCl pH8, containing proteases inhibitor cocktail (Complete EDTA
free, Roche, Rotkreuz), reduced, alkylated and digested with additions of 100 ng, each, of LysC for 2 hours at 37 C followed by
Trypsin (both sequencing grade from Promega) at room temperature overnight as described by Braga-Lagache et al.86 The digests
were analyzed by nano-liquid chromatography on a Dionex, Ultimate 3000, (ThermoFischer Scientific, Reinach, Switzerland) coupled
to a timsTOF Pro (Bruker Daltonics, Bremen, Germany), through a CaptiveSpray source (Bruker, Bremen, Germany) with an end-plate
offset of 500 V, a drying temperature of 200  C, and with the capillary voltage fixed at 1.6 kV. A volume of 2 mL from the protein digest
were loaded onto a pre-column (C18 PepMap 100, 5 mm, 100A, 300 mm i.d. x 5mm length, ThermoFisher) at a flow rate of 10 mL/min
with 0.05% TFA in water/acetonitrile 98:2. After loading, peptides were eluted in back flush mode onto a homemade C18 CSH Waters
column (1.7 mm, 130 Å, 75 mm 3 20 cm) by applying a 90-minute gradient of 5% acetonitrile to 40% in water / 0.1% formic acid, at a
flow rate of 250 nl/min. The timsTOF Pro was operated in the Parallel Acquisition Serial Fragmentation (PASEF) mode. The mass
range was set between 100 and 1700 m/z, with 10 PASEF scans between 0.6 and 1.6 V s/cm2. The accumulation time was set to
2 ms, and the ramp time was set to 100 ms. Fragmentation was triggered at 20,000 arbitrary units (au), and peptides (up to charge
5) were fragmented using collision induced dissociation with a spread between 20 and 59 eV.

Calcium imaging of iGNs and lentiviral-mediated FRRS1l transduction

iGNs were transduced with constitutive AAV2/1-Syn-GCaMP6s-WPRE-SV40 at DIV5, and imaging was performed on DIV12 using an
Olympus Fluoview 1000-BX61 (Olympus, Tokyo) microscope fitted with a 40X water immersion objective. For FRRS1l overexpression,
iGNs were transduced at DIV4 with lentiviruses LV::FRRS1L-myc (Origene, RC221300L1V) and at DIV7 with AAV2/1-Syn-GCaMP6s-
WPRE-SV40. For the knockdown of FRRS1L, iGNs were transduced at DIV4 with LV::shFRRS1L-GFP (Origene, TL314264V) and subse-
quently transduced with AAV2/1-Syn-NES-jRCaMP1a-WPRE-SV40. Calcium imaging was performed at DIV12-14. Cells were incubated
in the recording solution containing 115mM NaCl, 5.6mM KCl, 1mM NaH2PO4, 2 mM CaCl2, 1mM MgCl2, 11mM glucose and 25mM
NaHCO3 for 10 minutes before imaging. Images were acquired every 10 seconds, the first 40 seconds were considered as baseline, after-
wards 50mM KCl was added to the imaging solution to depolarize the cells. All experiments were performed from three-five independent
culture differentiations/line. Images were analysed using Fiji, neuronal somata were assessed as ROIs and median intensity over time was
measured. To calculate DF, the median intensity values are divided by the average of the first 40 seconds of recording (F0) per single ROI.

QUANTIFICATION AND STATISTICAL ANALYSIS

Microscopy and image analysis

For parvalbumin, P-CaMKII, CaMKII, Calbindin, Frrs1l, signal intensity values were calculated over a maximal intensity projection of
several consecutive Z-stack spaced 0.5 mm. Cells were manually ROI and average intensity was measured with Fiji. For parvalbumin

e7 Neuron 111, 2523–2543.e1–e10, August 16, 2023

 ll
Article OPEN ACCESS

and Frrs1l, neurons with values of intensity below 50 arbitrary units (a.u.) were considered as low expressing, between 50 and 100 a.u.
as medium expressing and above 100 a.u. as high expressing. Intensity values were plotted using GraphPad Prism 9. For GluR2
expression analysis, 5 random spots within the cellular membrane of PV positive MLIN were measured using Fiji, the average of
the 5 points were plotted using GraphPad Prism 9. The spread of mCherry positive MLIN for chronic inhibition of cerebellar MLIN
was counted in 8 mCherry and 9 DREADD(Gi) injected animals between UBERN and LMU, multiple areas within every image (total
of 142 images, approximately 8-9 sections/animal) were used to analyze the number of positive MLIN, PN and GCL neurons.
For the analysis of VGAT and VGluT1 synapses 4 Z-stacks spaced 0.5 mm were used. Both neurons soma and synapses were re-
constructed in 3D using Imaris software (version 7.6.3), only synapses in contact with the cell membrane were counted for density
analysis. For Homer-3 synapses analysis 3 Z-stacks spaced 0.5 mm were used. Puncta were reconstructed in 3D using Imaris soft-
ware within a ROI of 35.4 mm2 and the number of synapses was plotted using GraphPad Prism 9.
Analysis of fluorescence intensity for iGNs SORCS3 signal intensity values were calculated over a maximal intensity projection of
consecutive Z-stack spaced 0.5 mm, neurons with values of intensity up to 70 a.u. were considered as non-expressing SORCS3, and
neurons with values above 70 a.u. were considered as expressing SORCS3. Background fluorescence intensity was between 20-50
a.u. For parvalbumin intensity analysis neurons were considered low expressing with values below 70 a.u., medium with values be-
tween 70 and 100 a.u. and high expressing with values over 100 a.u.

Data analysis of mass spectrometry

All samples were processed with MaxQuant87 (version 2.0.1.0), with first and main peptide search tolerance set to 20, respectively 10
ppm, and MS/MS match tolerance to 40 ppm. MaxQuant’s TIMS-DDA default instrument settings were kept. Enzyme specificity was
set to strict trypsin, and a maximum of three missed cleavages were allowed. Carbamidomethylation on cysteine was set as a fixed
modification, methionine oxidation and protein N-terminal acetylation as variable modifications. Sorted cell samples were searched
against the manually reviewed uniprot.88 Mus musculus database (release July 2020), subsequent mouse samples with the full uni-
prot database (release June 2021). Common contaminants were added in each case. Protein intensities are reported as MaxQuant’s
Label Free Quantification (LFQ) values, as well as iTop389 values (sum of the intensities of the three most intense peptides); for the
latter, variance stabilization90 was used for the peptide normalization, and missing peptide intensities, if at least 2 evidences exist in a
group, were imputed by drawing values from a Gaussian distribution of width 0.3 centered at the sample distribution mean minus 2.8x
the sample standard deviation. Imputation at protein level for both iTop3 and LFQ was performed if there were at least two measured
intensities in at least one group of replicates; missing values in this case were drawn from a Gaussian distribution of width 0.3
centered at the sample distribution mean minus 2.5x the sample standard deviation. Differential expression tests were performed
using empirical Bayes (moderated t-test) implemented in the R limma package.91 The Benjamini and Hochberg92 method was further
applied to correct for multiple testing. The criterion for statistically significant differential expression is that the maximum adjusted
p-value for large fold changes is 0.05, and that this maximum decreases asymptotically to 0 as the log2 fold change of 1 is ap-
proached (with a curve parameter of one time the overall standard deviation). Proteins consistently significantly differentially ex-
pressed through 20 protein imputation cycles were subsequently flagged. Sorting and consequent mass spectrometry experiment
was replicated in triplicate (3 WT and 3 Sca1 animals per replicate). Proteins found in at least 2 replicates experiments were used for
PANTHER analysis.
See also Table S1 related to Figure 7. List of identified proteins in adult rodent MLIN.

In vivo two-photon calcium image processing and data analysis

Image analyses were performed in Matlab (R2021a, Math Works) using custom-written routines.4 In brief, full frame images were cor-
rected for potential x and y brain displacement, and regions of interests (ROIs) were semi-automatically selected based on the
maximum and mean projection of all frames. Fluorescence intensity of all pixels within a selected ROI were averaged in each frame,
and the resulting time series (traces) low pass filtered at 10 Hz and smoothed over 3 frames. Potential neuropil signal contamination
was addressed with the following equation (FROI_comp, neuropil-compensated fluorescence of the ROI; FROI, initial fluorescence
signal of the ROI; Fneuropil, signal from the neuropil) (modified after Chen et al.,19 Liebscher et al.,93 and Scekic-Zahirovic et al.94).
FROI comp = FROI 0:8 3 Fneuropil
In case the median of Fneuropil was higher than the median of FROI, we opted for a neuropil compensation factor of 0.7. To estimate
the noise and to compute the baseline of each trace (F0), we subtracted the 8th percentile in a sliding window of 30 frames. The stan-
dard deviation of this trace divided by its median and after subtracting 1, yields the noise of each trace. F0 corresponds to the median
of all values lower than the 70% percentile of the noise band. All further analyses are performed on (FROI_comp/F0)-1 traces, reflecting
the DF/F. The activity of dedicated neurons is given as area under the curve (AUC) of the DF/F per time unit. To minimize contributions
of noise fluctuations to this metric, we only considered values exceeding 2x the noise. 8 Running responsiveness was assessed by
randomly circularly shifting the DF/F trace with respect to the speed vector 1000 times and computing the ratio of the median DF/F
during locomotion and during quiet wakefulness and comparing it to the actual activity ratio of a given ROI. Neurons for which the
actual ‘locomotion/quiet wakefulness’ activity ratio exceeded the 95th percentile of the shuffled data were considered running
responsive. Spontaneous activity reflects the AUC during quiet wakefulness. Only experiments with a minimum of 4000 frames, cor-
responding to 2.7 min, were considered in this analysis. Air puff responses were assessed in experiments with at least 2 air puffs,

Neuron 111, 2523–2543.e1–e10, August 16, 2023 e8

 ll
OPEN ACCESS Article

which were not followed by a running response in a window of 3 seconds following the puff application to avoid crosstalk between
modalities. Due to these varying sets of criteria, the effective number of neurons differs across analyses.

Population activity dimensionality and manifold analysis

To reduce the dimensionality, we applied principal component analysis (PCA) to the 0-centered and z-scored DF/F traces of all neu-
rons or of selected cell-types in each FOV in the feedback imaging sessions.27,95 To assess the correlation between the population
response pattern and the conjunct recorded behavior, a linear regression was performed for the principal components and the
different behavioral parameters, i.e., locomotion, pupil width, whisking and air puff and all measured behavior combined, to separate
active from quiescent states. To further explore the structure of the behaviorally associated population response patterns, we con-
ducted a PCA-based manifold analysis by projecting high dimensional population response pattern trajectories into 3 dimensional
PCA space. The manifold subspace clearly segregated into two distinct clusters, corresponding to quiet wakefulness and active
state (dominated by locomotion). The center of mass for the respective subspaces was computed as the mean of the coordinates
within each subspace. The Euclidean distance within a given subspace corresponds to the average length of each dot to the center of
mass. The Euclidean distance between two subspaces reflects the distance between the center of mass of each subspace. To
compare the Euclidean distances between WT and Sca1, only experiments with at least 30% of the time spent being stationary
and a minimum of 20% of the time spent running were included. Of note, as the number of ROIs in a given field of view (FOV) varied
significantly within the acute chemogenetic stimulation data set, we bootstrapped by subsampling 70 pseudorandomly selected
ROIs in each FOV for each of the 100 iterations and computed the mean values of the derived center of mass.

Statistical analysis
Analysis was done using GraphPad Prism and Matlab (R2021a, Mathworks). Statistical significances throughout the paper were eval-
uated by two-tailed, unpaired Student’s t test, Two-way ANOVA or Kolmogorov-Smirnov (KS) tests to assess differences in distribu-
tions. Post ANOVA Sidak or Tukey test was used to evaluate statistical significance throughout the paper as indicated in the respec-
tive figure legend. Values are expressed as mean ± standard error of the mean (SEM). A P-value <0.05 was considered as statistically
significant (*p <0.05, **p<0.01, ***p<0.001).

Figure Test F / t / R / etc. values

1E–1J Kolmogorov-Smirnov test
1N 2-way repeated measures ANOVA genotype F(1,2092) = 71.89
1O Wilcoxon rank sum test
1P 2-way repeated measures ANOVA genotype F(1,163) = 12.18
1Q Wilcoxon rank sum test
1R 2-way repeated measures ANOVA genotype F(1,250) = 0.17
1S Wilcoxon rank sum test
2C 2-way ANOVA genotype F(1,28) = 2.07
2D–2F Wilcoxon rank sum test
3A Chi-square
3C–3E Unpaired t-test
4B Chi-square
4C 2-way ANOVA time F(11,192)= 9.761, treatment F(1,192)=
(Sidak multiple comparison test) 97.34, interaction F(11,192)= 12.35
4H–4O Kolmogorov-Smirnov test
4Q Wilcoxon rank sum test after bootstrapping
5C Chi-square
5D Two-way ANOVA timepoints F(11,180)= 10.55, treatment
(Sidak multiple comparison test) F(1,180)= 37.88, interaction F(11,180)= 12.09,
5F and 5G Unpaired t-test
5I Two-way ANOVA time points F(4,70)= 0.9894, treatment
(Sidak multiple comparison test) F(1,70)= 38.49, interaction F(4,70)= 1.095
5J Unpaired t-test
6B Chi-square
6C–6E Unpaired t-test
(Continued on next page)

e9 Neuron 111, 2523–2543.e1–e10, August 16, 2023

 ll
Article OPEN ACCESS

Continued
Figure Test F / t / R / etc. values
6F Two-way ANOVA timepoints F(4,215)=1.030, treatment
(Tukey multiple comparison test) F(3,215)= 28,59, interaction F(12,215)= 1.039
7G Unpaired t-test
& Pearson correlation
7H Chi-square
7I Unpaired t-test
8A Unpaired t-test
8B Chi-square &
Pearson correlation
8C Multiple t-test
8D Unpaired t-test
8F and 8G Multiple t-test
Table indicating precise statistical test employed for main figures.

Neuron 111, 2523–2543.e1–e10, August 16, 2023 e10