bio tech

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/340518273

Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of

Neurological Disease in Mice

Article  in  Cell · April 2020

DOI: 10.1016/j.cell.2020.03.024

CITATIONS

39 READS

1,522

22 authors, including:

Some of the authors of this publication are also working on these related projects:

Focal Cortical Dysplasia View project

Optical and electrical characterization of GaInP/GaAs double-junction tandem solar cells View project

Jinlin Su

Institute of Neuroscience

9 PUBLICATIONS   42 CITATIONS   

SEE PROFILE

Xinde Hu

Northwest A & F University

56 PUBLICATIONS   718 CITATIONS   

SEE PROFILE

Changyang Zhou

Institute of Neuroscience

33 PUBLICATIONS   421 CITATIONS   

SEE PROFILE

Aileen Sun

Chinese Academy of Sciences

41 PUBLICATIONS   823 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Jinlin Su on 24 April 2020.

The user has requested enhancement of the downloaded file.

Article

Glia-to-Neuron Conversion by CRISPR-CasRx

Alleviates Symptoms of Neurological Disease in Mice

Graphical Abstract

Highlights

d Knockdown of Ptbp1 converts Müller glia into retinal

ganglion cells in mature retinas

d Central projections of converted retinal ganglion cells

restore visual responses

d Induction of neurons with dopaminergic features in PD

model mice

d Induced neurons alleviated motor dysfunctions in PD mice

Zhou et al., 2020, Cell 181, 1–14 April 30, 2020 ª 2020 Elsevier Inc. https://doi.org/10.1016/j.cell.2020.03.024

Authors

Haibo Zhou, Jinlin Su, Xinde Hu, ...,

Haishan Yao, Linyu Shi, Hui Yang

Correspondence [email protected] (H.Z.), [email protected] (H.Y.)

In Brief

In vivo CasRx-mediated downregulation

of a single RNA-binding protein, Ptbp1,

locally converts glia to neurons and

shows promise for treating disorders due

to neuronal loss in mice.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

Article

Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice Haibo Zhou,1,4,* Jinlin Su,1,4 Xinde Hu,1,2,4 Changyang Zhou,1,2,4 He Li,1,2,4 Zhaorong Chen,1,2,4 Qingquan Xiao,1,2

Bo Wang,1,2 Wenyan Wu,1,2 Yidi Sun,2,3 Yingsi Zhou,1 Cheng Tang,1,2 Fei Liu,1 Linhan Wang,1,2 Canbin Feng,1

Mingzhe Liu,1 Sanlan Li,1 Yifeng Zhang,1 Huatai Xu,1 Haishan Yao,1 Linyu Shi,1 and Hui Yang1,5,* 1Institute of Neuroscience, State Key Laboratory of Neuroscience, Key Laboratory of Primate Neurobiology, CAS Center for Excellence in Brain Science and Intelligence Technology, Shanghai Research Center for Brain Science and Brain-Inspired Intelligence, Shanghai Institutes

for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China 2College of Life Sciences, University of Chinese Academy of Sciences, Beijing 100049, China 3Bio-Med Big Data Center, Key Laboratory of Computational Biology, CAS-MPG Partner Institute for Computational Biology, Shanghai Institute of Nutrition and Health, Shanghai Institutes for Biological Sciences, University of Chinese Academy of Sciences, Chinese

Academy of Sciences, Shanghai 200031, China 4These authors contributed equally 5Lead Contact *Correspondence: [email protected] (H.Z.), [email protected] (H.Y.)

https://doi.org/10.1016/j.cell.2020.03.024

SUMMARY

Conversion of glial cells into functional neurons rep- resents a potential therapeutic approach for replen- ishing neuronal loss associated with neurodegener- ative diseases and brain injury. Previous attempts in this area using expression of transcription factors were hindered by the low conversion efficiency and failure of generating desired neuronal types in vivo. Here, we report that downregulation of a sin- gle RNA-binding protein, polypyrimidine tract-bind- ing protein 1 (Ptbp1), using in vivo viral delivery of a recently developed RNA-targeting CRISPR system CasRx, resulted in the conversion of Müller glia into retinal ganglion cells (RGCs) with a high effi- ciency, leading to the alleviation of disease symp- toms associated with RGC loss. Furthermore, this approach also induced neurons with dopaminergic features in the striatum and alleviated motor defects in a Parkinson’s disease mouse model. Thus, glia- to-neuron conversion by CasRx-mediated Ptbp1 knockdown represents a promising in vivo genetic approach for treating a variety of disorders due to neuronal loss.

INTRODUCTION

Neurodegenerative diseases are devastating diseases associ-

ated with the progressive loss of neurons in various parts of

the nervous system. For example, degeneration of retinal gan-

glion cells (RGCs), the sole output neurons of the retina, repre-

sents the leading cause of retinal diseases with permanent

blindness (Quigley and Broman, 2006). Trans-differentiation of

Müller glia (MG) into RGCs has been proposed to be a potential

therapy for restoring visual function. However, MG lose the

neurogenic capacity at ~2 postnatal weeks in mice and MG-to-

RGC conversion has not been achieved in mature mammalian

retinas so far (Elsaeidi et al., 2018; Laha et al., 2017; Roska

and Sahel, 2018; Ueki et al., 2015).

Recent in vitro studies showed that downregulating

the expression of a single gene, polypyrimidine tract-binding

protein 1 (Ptbp1), which encodes a RNA binding protein,

was sufficient to convert cultured mouse fibroblasts and

N2a cells into functional neurons (Xue et al., 2013). However,

in vivo neuronal conversion by downregulating Ptbp1 has not

been explored. A recently characterized RNA-guided and

RNA-targeting CRISPR protein family known as Cas13 offers

an efficient approach for manipulation of RNA transcripts

(Abudayyeh et al., 2016; Cox et al., 2017; East-Seletsky

et al., 2016; Gootenberg et al., 2017; Knott and Doudna,

2018; Konermann et al., 2018). Among Cas13 proteins, an or-

tholog of CRISPR-Cas13d (CasRx) has the smallest size and

exhibits high targeting specificity and efficiency, making

it ideal for in vivo therapeutic application (Konermann

et al., 2018).

In the present study, we showed that MG could be converted

into RGCs by injecting AAVs expressing CasRx and two guide

RNAs (gRNAs) targeting Ptbp1 mRNA in both intact and

damaged mature retinas. Notably, converted RGCs established

central projections to dorsal lateral geniculate nucleus (dLGN)

and superior colliculus (SC) and partially restored visual func-

tions in a mouse model with drug-induced retinal injury. Using

a similar approach, we demonstrated that knockdown of

Ptbp1 in the striatum locally induced neurons expressing dopa-

minergic markers with a high efficiency and alleviated motor dys-

functions caused by midbrain dopamine neuronal loss in a

mouse model of PD. Thus, CasRx-mediated Ptbp1 knockdown

could achieve efficient glia-to-neuron conversion that replen-

ishes the desired neuronal types, paving the way for future ther-

apeutic applications.

Cell 181, 1–14, April 30, 2020 ª 2020 Elsevier Inc. 1

Figure 1. Specific Knockdown of Ptbp1

mRNA In Vitro Using CasRx

(A) Schematic illustration of CasRx-mediated

knockdown of Ptbp1 mRNA.

(B) Schematic indicates the target sites of six

gRNAs in the Ptbp1 gene.

(C and D) Knockdown efficiency of different

combinations of gRNAs. The gRNA 5 and 6

showed the most potent knockdown efficiency in

both N2a cells (C) and astrocytes (D). Number

above each bar indicates the number of repeats

per group. All values are presented as

mean ± SEM.

(E and F) Expression levels in log2 (fragments per

kilo base per million mapped reads [FPKM] + 1)

values of all detected genes in RNA sequencing

(RNA-seq) libraries of CasRx-Ptbp1 (y axis)

compared to CasRx control (x axis). N2a cells (E),

n = 4 independent replicates for both groups; as-

trocytes (F), n = 2 independent replicates for both

groups.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

RESULTS

Specific Knockdown of Ptbp1 In Vitro Using CasRx To examine the efficiency of CasRx-mediated knockdown of

Ptbp1, we first screen six gRNAs for their efficiency in CasRx

editing of Ptbp1 in both N2a cells and cultured astrocytes (Fig-

ures 1A and 1B). We found that co-transfection of a vector con-

taining CasRx gene with two gRNAs 5 and 6 that target Ptbp1

exon IV and VII, resulted in 87% ± 0.4% and 76% ± 4% (SEM,

n = 5 repeats) reduction of Ptbp1 mRNA in N2a cells and

cultured astrocytes, respectively (Figures 1C and 1D). Tran-

scriptome analysis showed that Ptbp1 was specifically downre-

gulated while the transcriptional level of typical neuronal genes

remained unchanged, 2 days after transfection (Figures 1E, 1F,

and S1A).

2 Cell 181, 1–14, April 30, 2020

Ptbp1 Knockdown Converts MG to RGCs in Mature Retinas Given the finding that Ptbp1 knockdown

could convert cultured mouse fibroblasts

and N2a cells into functional neurons (Xue

et al., 2013), we next examined whether

Ptbp1 knockdown in the mature retina

could result in conversion of MG into

RGCs in vivo. The possibility of using

CasRx-induced knockdown of Ptbp1 in

glia-to-neuron conversion was examined

in the mature retina. To specifically and

permanently label the retinal MG, we in-

jected AAV-GFAP-GFP-Cre into the eyes

of Ai9 mice (Rosa-CAG-LSL-tdTomato-

WPRE) to induce tdTomato expression

specifically in MGs (Figures S1B and

S1C). We also constructed AAV-GFAP-

CasRx-Ptbp1 (with gRNAs 5+6 targeting

Ptbp1) driven by the GFAP promotor to

knockdown Ptbp1 specifically in MG

(Yao et al., 2016, 2018) and AAV-GFAP-

CasRx that does not contain Ptbp1 gRNAs as a control (Figures

2A and S1D). At 1 month after subretinal co-injection of AAV-

GFAP-CasRx-Ptbp1 and AAV-GFAP-Cre-GFP into the eyes of

5-week Ai9 mice retinas, we found that many tdTomato+ cells

co-immunostained with RGC markers Brn3a or Rbpms in retinal

ganglion cell layer (GCL) (tdTomato+Brn3a+, 18 ± 2 cells per 1 mm

3 10 mm; tdTomato+Rbpms+, 18 ± 2 cells per 1 mm 3 10 mm), but

no such cell in the retina injected with control AAV vectors (Fig-

ure 2B–2E), suggesting conversion of MG to RGCs in the mature

retina. Notably, converted RGC cells frequently showed low

expression of GFAP-driven GFP (Figure S1E), consistent with

the loss of glial identity after MG-to-RGC conversion. Interest-

ingly, we observed a fraction of tdTomato+ cells in the GCL ex-

pressing Foxp2+, Brn3c+, or Parvalbumin+ (Figures S1F–S1H),

markers of F-RGCs, RGCs subtype 3, and PV-RGCs,

Figure 2. Ptbp1 Knockdown Converts MG to RGCs in Intact Mature Retinas

(A) Schematic illustration of MG-to-RGC conversion. Vector I (AAV-GFAP-GFP-Cre) encodes Cre recombinase and GFP driven by the MG-specific promoter

GFAP and Vector II (AAV-GFAP-CasRx-Ptbp1) encodes CasRx and gRNAs. To induce RGCs, retinas (Ai9 mice, 5 weeks old) were either injected with AAV-GFAP-

CasRx-Ptbp1 or control vector AAV-GFAP-CasRx together with AAV-GFAP-GFP-Cre. Occurrence of conversion was examined around one month post-in-

jection. ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer.

(B) Representative images showing colocalization of tdTomato and Brn3a in the GCL (dash line). White arrowheads indicate the endfeet of MG, which did not

colocalize with Brn3a, and yellow arrowheads indicate the colocalization of tdTomato and Brn3a in the retinas injected with AAV-GFAP-GFP-Cre and AAV-GFAP-

CasRx-Ptbp1. Brn3a is a specific marker for RGCs. Scale bar, 20 mm.

(C) Number of tdTomato+ or tdTomato+Brn3a+ cells in the GCL at 1 month after AAV injection. AAV-GFAP-GFP-Cre plus AAV-GFAP-CasRx, n = 6 retinas; AAV-

GFAP-GFP-Cre plus AAV-GFAP-CasRx-Ptbp1, n = 7 retinas. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.

(D) Representative images showing colocalization of tdTomato with the other RGC-specific marker Rbpms in the GCL. Yellow arrowheads indicate the co-

localization of tdTomato and Rbpms. Scale bar, 20 mm.

(E) Number of tdTomato+ or tdTomato+Rbpms+ cells in the GCL at 1 month after AAV injection. AAV-GFAP-GFP-Cre plus AAV-GFAP-CasRx, n = 6 retinas; AAV-

GFAP-GFP-Cre plus AAV-GFAP-CasRx-Ptbp1, n = 8 retinas. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

respectively (Rheaume et al., 2018; Rousso et al., 2016), suggest-

ing that MG were converted into different subtypes of RGCs.

Successful induction of RGCs was also confirmed by another

strategy at 2–3 weeks after co-injecting AAV-GFAP-mCherry

and AAV-EFS-CasRx-Ptbp1 (CasRx driven by the ubiquitous

promoter EFS) into the retinas of C57BL/6 mice (Figures S2A–

S2E). Together, our results showed that RGCs could be efficiently

converted from MG via Ptbp1 knockdown in the mature retina.

Previous studies showed that overexpression of Ascl1 or a cock-

tail of transcription factors could convert MG into different types

of retinal neurons in the mature retina (Jorstad et al., 2017;

Yao et al., 2018). We also found that, besides RGCs,

MG could also be converted into amacrine cells by CasRx-medi-

ated knockdown of Ptbp1 (Figures S3A–S3C).

Cell 181, 1–14, April 30, 2020 3

Figure 3. MG-to-RGC Conversion in a Mouse Model of NMDA-Induced Retinal Injury

(A) Outline of experimental design. Retinal injury was induced in Ai9 mice 4–8 weeks of age by intravitreal NMDA injection (200 mM, 1.5 mL). Two to three weeks

after NMDA injection, AAVs were introduced by subretinal injection. Immunostaining and behavioral tests were performed 1 month after AAV injection.

(B) NMDA injection essentially depleted RGCs in the GCL. Scale bar, 50 mm.

(C) Colocalization of tdTomato and Brn3a. Yellow arrowheads indicate the co-localization of Brn3a and tdTomato in the GCL. n = 6 retinas, Scale bar, 20 mm.

(D) Number of Brn3a+ or tdTomato+ or tdTomato+Brn3a+ cells in the GCL. Uninjured retina, n = 6 retinas; GFAP-CasRx plus GFAP-GFP-Cre, n = 6 retinas; GFAP-

CasRx-Ptbp1 plus GFAP-GFP-Cre, n = 7 retinas. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.

(E) Colocalization of tdTomato and Rbpms. Yellow arrowheads indicate co-localization of Rbpms and tdTomato in the GCL injected with GFAP-CasRx-Ptbp1

plus GFAP-GFP-Cre. Scale bar, 20 mm.

(F) Number of Rbpms+ or tdTomato+ or tdTomato+Rbpms+ cells in the GCL. Uninjured retina, n = 6 retinas; GFAP-CasRx plus GFAP-GFP-Cre, n = 7 retinas;

GFAP-CasRx-Ptbp1 plus GFAP-GFP-Cre, n = 7 retinas. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.

(G) Image shows a representative RGC-like tdTomato+ cell recorded under two-photon microscope. Scale bar, 20 mm.

(H) A representative tdTomato+ ON cell spikes to LED light in response window relative to baseline windows. In total, 8 cells were recorded and 6 of them showed

response to LED light. Among these cells, 5 were ON cells and 1 was an OFF cell.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

MG-to-RGC Conversion in a NMDA-Induced Retinal Injury Mouse Model To explore whether MG-derived RGCs could replenish RGCs in

retina injury mouse model, we intravitreally injected 4- to

8-week-old Ai9 mice with N-methyl-D-aspartate (NMDA,

200 mM), which causes a near complete loss of RGCs and the

reduction of the thickness of inner plexiform layer (IPL) (Niwa

et al., 2016). Two to three weeks after NMDA injection, the

eyes were either injected with AAV-GFAP-CasRx-Ptbp1 plus

AAV-GFAP-GFP-Cre or control AAVs (Figures 3A and 3B). One

month after AAV injection, we found that the number of Brn3a+

or Rbpms+ cells (Brn3a+, 21 ± 4 cells per 1 mm 3 10 mm, as

4 Cell 181, 1–14, April 30, 2020

compared to 4 ± 1 cells per 1 mm 3 10 mm in untreated injured

retina and 117 ± 8 cells per 1 mm 3 10 mm in uninjured retina;

Rbpms+, 34 ± 3 cells per 1 mm 3 10 mm, as compared to 6 ±

1 cells per 1 mm 3 10 mm in untreated injured retina and 143 ±

5 cells per 1 mm 3 10 mm in uninjured retina) in the GCL was

significantly elevated in retinas injected with AAV-GFAP-

CasRx-Ptbp1, and the majority of these cells were tdTomato+

(Figures 3C–3F). Moreover, more than half the tdTomato+ cells

in GCL expressed Brn3a and Rbpms (Figures 3C–3F). To deter-

mine whether MG-derived RGCs integrate into the retinal circuits

and have the capacity of receiving visual information, we per-

formed cell-attached recording from MG-derived RGCs under

Figure 4. Central Projections of Converted RGCs and Partial Restoration of Visual Responses

(A) Schematic illustration of the visual pathway. RGCs send their axons via the optic nerve that relay visual signals outside of the retina to dLGN and SC in

the brain.

(B) Retinal flat-mount preparations. Yellow arrowhead indicates MG-derived tdTomato+ RGC axons. Images were tiled. Scale bar, 100 mm. Experiments were

independently repeated 3 times per group with similar results.

(C) Representative images showing tdTomato+ axons of induced RGCs in the optic nerve. Images were tiled. Scale bar, 200 mm. Experiments were independently

repeated 5 times per group with similar results.

(legend continued on next page)

Cell 181, 1–14, April 30, 2020 5

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

two-photon microscope to monitor light stimulus-evoked re-

sponses (Figure 3G). We found that 6 out of 8 cells examined

showed action potentials in response to light stimulation (Fig-

ure 3H). Among these cells, five were ON cells and one was an

OFF cell (Figure 3H). These results suggested that functional

RGCs could be converted from MG via Ptbp1 knockdown in

the injured retina.

Central Projections of Converted RGCs Restored Visual Responses In the visual system, visual information is relayed by RGC projec-

tions to the dorsal lateral geniculate nucleus (dLGN) and superior

colliculus (SC) in the brain (Figure 4A). In the CasRx-treated

NMDA-injured retina, we observed a large amount of tdTomato+

axons in the treated retina and the optic nerve, but no such axons

in control AAV-treated group (Figures 4B and 4C). Remarkably,

we found tdTomato+ axons in the dLGN and SC, which were

much more abundant in the contralateral than the ipsilateral

side of the brain (Figures 4D and 4E), consistent with expectation

that newly formed axon projections of the converted RGCs

correctly send their projections to their central target areas (As-

sali et al., 2017; Rebsam et al., 2009).

The function of central projections of MG-derived RGCs was

further examined by monitoring visual responses evoked by the

light stimulus applied to the NMDA-injured retina. Visually

evoked potentials (VEPs) were recorded in the primary visual

cortex (V1) of anaesthetized mice 1 month after AAVs injection

(Figure 4F). Striking VEPs were evoked by stimulus to the

contralateral retina in NMDA-injured mice treated with AAV-

GFPA-CasRx-Ptbp1 and AAV-GFAP-mCherry, similar to that

found in wild-type uninjured mice, whereas only weak re-

sponses were observed in the control NMDA-injured mice in-

jected with control AAV vectors (Figures 4G and 4H). This sup-

ports the notion that central projection to dLGN had restored at

least partially visual information relay to V1, presumably by

making synaptic connections with existing functional dLGN

neurons in the brain. Finally, we explored whether CasRx-medi-

ated conversion of MG to RGCs could restore vision-depen-

dent behavior that was lost by NMDA-induced retinal injury

(Figure 4I). We found that bilateral intravitreal NMDA injection

in mice resulted in a reduced duration in the dark compartment

in a light/dark preference test, consistent with a loss of vision

due to retina injury. By contrast, CasRx-mediated MG-to-

RGC conversion in both eyes of bilaterally retina-injured mice

resulted in a marked increase in the duration in dark compart-

ment, to a level close to that found for control uninjured mice

(D and E) Representative images showing that strong signals were observed in th

Experiments were independently repeated 4 times per group with similar results.

was tiled for (D) and (E). Scale bar, 500 mm (left), 50 mm (right).

(F) Schematic illustration of VEP recording (C57BL/6 strain).

(G) VEPs to light flashes in the primary visual cortex. Responses across mice from

of retinas per group is indicated.

(H) Response amplitude for wild-type (WT, C57BL/6 strain, n = 8 retinas), NMDA, a

GFAP-CasRx (n = 11 retinas), and NMDA, AAV-GFAP-mCherry, and AAV-GFAP

presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.

(I) Design of the dark/light preference test. Note that both eyes were treated with

(J) Percentage of time spent in dark chambers. WT (C57BL/6 strain), n = 13 mice; N

GFAP-CasRx, n = 12 mice; NMDA, AAV-GFAP-mCherry and AAV-GFAP-CasRx-

*p < 0.05, **p < 0.01, ***p < 0.001.

6 Cell 181, 1–14, April 30, 2020

(Figure 4J), consistent with the restoration of vision-dependent

behaviors.

Next, we determined the time course of appearance of MG-to-

RGC conversion and central projections in the intact retinas

without NMDA-induced injury, by performing immunostaining

at five different time points (1 week, 1.5 weeks, 2 weeks, 3 weeks,

and 1 month) after AAV injection. We found that tdToma-

to+Rbpms+ and tdTomato+Brn3a+ cells in the retina were first

seen at 1.5 week after the AAV injection, and the number of these

cells progressively increased over time (Figures S4A and S4B).

We noted that there was an intermediate stage showing that

induced RGCs migrated from INL to GCL at 1.5 week after the

AAV injection (Figure S4C). The time course of MG-to-RGC con-

version was also demonstrated in the retina with NMDA-induced

injury (Figure S4D). For the central projection, we found progres-

sive increase in the tdTomato+ axons in the visual pathway:

labeled axons were not found in the optic nerve at 1 week (Fig-

ure 5A) and began to appear in contralateral dLGN at 1.5 week

(Figure 5B) after the AAV injection. Labeled projections were first

observed in the contralateral, but not ipsilateral, SC by 2 weeks

(Figure 5C) and clearly observed in both contralateral and ipsilat-

eral dLGN and SC by 3 weeks, with further increased on both

sides at 1 month (Figures 5B–5D). Similar findings were also

observed in the injured retinas injected with NMDA (Figures

S5A–S5C).

CasRx-Induced Neuron Conversion in Mouse Striatum To generalize the potential usefulness of CasRx-induced glia-to-

neuron conversion in other systems, we further examined

whether knockdown of Ptbp1 in the striatum could locally

convert other types of cells into dopamine neurons, an approach

that may be useful for replenishing dopamine in the striatum due

to degeneration of dopaminergic neurons in midbrain substantia

nigra associated with Parkinson’s disease (PD). We first injected

wild-type mice with AAV-GFAP-CasRx-Ptbp1 (with gRNAs 5+6

for Ptbp1) into the striatum to downregulate Ptbp1, together

with AAV-GFAP-mCherry that fluorescently labeled astrocytes

(Figures S6A and S6B). As a control, we used AAV-GFAP-CasRx

that does not contain Ptbp1 gRNA. We found that both mCherry

and CasRx were largely expressed in astrocytes and showed a

high co-infection efficiency in the striatum, with 99% ± 1%

mCherry+ cells expressed CasRx (82% ± 2% GFAP+ cells ex-

pressed mCherry, and 95% ± 1% mCherry+ cells expressed

GFAP). The absolute number of CasRx-infected cell was 40 ±

8 Flag+ cells per 200 mm 3 200 mm 3 10 mm (Figures S6C and

S6D). The expression of Ptbp1 was downregulated in astrocytes

e contralateral dLGN (D) and SC (E), target regions of RGC axons in the brain.

Note that ipsilateral dLGN and SC showed weak tdTomato+ signals. Left panel

the same group are shown and each line represents a single retina. The number

nd AAV-GFAP-mCherry (n = 12 retinas), NMDA, AAV-GFAP-mCherry, and AAV-

-CasRx-Ptbp1 (n = 8 retinas). Each point represents a single mouse. Data are

NMDA and injected with the same AAVs 2 weeks later.

MDA and GFAP-mCherry, n = 14 mice; NMDA, AAV-GFAP-mCherry and AAV-

Ptbp1, n = 12 mice. All values are presented as mean ± SEM; unpaired t test;

Figure 5. The Time Course of Induced RGCs Sent Their Projections into Optic Nerve and Central Brain Regions

(A) Representative images showing progressive increase of tdTomato+ axons in the optic nerve without NMDA-induced injury at five different time points (1 week,

1.5 weeks, 2 weeks, 3 weeks and 1 month). The tdTomato+ axons were first seen at 1.5 week after AAV injection (yellow arrowhead). Images were tiled. Scale bar,

50 mm. Experiments were independently repeated 3 times with similar results.

(B) Progressive increase in the density of tdTomato+ axons (yellow arrowheads) in dLGN. Note that tdTomato+ axons were first observed in the contralateral dLGN

at 1.5 week after AAV injection and upper panels were tiled. Scale bar, 500 mm (upper), 50 mm (lower). Experiments were independently repeated 3 times with

similar results.

(C) Progression of tdTomato+ axons (yellow arrowheads) in SC. Note that tdTomato+ axons were first observed in the contralateral SC at 2 week after AAV

injection and upper panels were tiled. Scale bar, 500 mm (upper), 50 mm (lower). Experiments were independently repeated 3 times with similar results.

(D) Schematic showing the progressive projections of induced RGC axons over time.

Cell 181, 1–14, April 30, 2020 7

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

Figure 6. Induction of Neurons with Dopaminergic Features in PD Model Mice (A and B) Outline of the experiment (A). 6-OHDA was unilaterally injected into the medial forebrain bundle. After 3 weeks, AAV-GFAP-CasRx-Ptbp1 plus AAV-

GFAP-mCherry, AAV-GFAP-CasRx plus AAV-GFAP-mCherry, or saline was injected into the ipsilateral (relative to the side of 6-OHDA infusion) striatum of mice

infused with 6-OHDA (B). Immunostaining were performed around 1 month or 3 months after AAV injection.

(C) Confocal images of converted mCherry+TH+ (yellow arrowheads) and mCherry+DAT+ (yellow arrowheads) cells 1 month or 3 months after AAV injection. The

neurites of a converted mCherry+TH+DAT+ cell were indicated by the orange arrowhead. Note that TH and DAT are dopamine neuron-specific markers. Scale bar,

50 mm or 10 mm for inset.

(D) Percentage of mCherry+TH+ cells in mCherry+ cells. AAV-GFAP-CasRx, 1 month: n = 5 mice; AAV-GFAP-CasRx-Ptbp1, 1 month: n = 5 mice; AAV-GFAP-

CasRx-Ptbp1, 3 months: n = 3 mice.

(E) Percentage of mCherry+TH+ cells in TH+ cells, n = 5 mice per group. Scale bar, 50 mm.

(F) Percentage of mCherry+DAT+ cells in mCherry+ cells. AAV-GFAP-CasRx, 1 month: n = 5 mice; AAV-GFAP-CasRx-Ptbp1, 1 month: n = 5 mice; AAV-GFAP-

CasRx-Ptbp1, 3 months: n = 3 mice.

(G) Percentage of mCherry+DAT+TH+ cells in mCherry+TH+ cells. AAV-GFAP-CasRx, 1 month: n = 5 mice; AAV-GFAP-CasRx-Ptbp1, 1 month: n = 5 mice; AAV-

GFAP-CasRx-Ptbp1, 3 months: n = 3 mice.

(H) Representative images showing that ALDH1A1 and GIRK2 expression colocalized with mCherry+TH+ (yellow arrowheads), and calbindin expression (orange

arrowheads) did not colocalized with mCherry+DAT+. ALDH1A1 and GIRK2 are SNc A9 area dopamine neuron-specific markers, and calbindin is a VTA dopamine

neuron-specific maker. n = 3 mice per group. Scale bar, 50 mm.

(I) The percentage of mCherry+ALDH1A1+, mCherry+GIRK2+, and mCherry+calbindin+ cells in mCherry+cells, respectively. n = 3 mice per group.

(J) Percentage of mCherry+TH+ALDH1A1+ and mCherry+TH+GIRK2+ in mCherry+TH+ cells, respectively; and mCherry+DAT+calbindin+ cells in mCherry+DAT+

cells. n = 3 mice per group.

(legend continued on next page)

8 Cell 181, 1–14, April 30, 2020

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

1 week after co-injecting AAV-GFAP-CasRx-Ptbp1 and AAV-

GFAP-mCherry into the striatum (Figures S6E and S6F), and a

high percentage (48% ± 10%, SEM, n = 6 mice) of mCherry+ cells

expressed mature neuron markers NeuN at 1 month after AAV in-

jection but not in the control striatum injected with AAV-GFAP-

mCherry and AAV-GFAP-CasRx (0.97% ± 0.45%, SEM, n = 6

mice) (Figures S6G and S6H). The neuronal type of these con-

verted cells was further examined by immunostaining of cell-

type-specific markers. We found that around 50% of converted

neurons expressed glutaminase (Figures S6I and S6J), a marker

of excitatory glutamatergic neurons (Kaneko et al., 1995), very

few converted neurons expressed an interneuron subtype

marker somatostatin (SST), and no cell expressed another inter-

neuron subtype marker parvalbumin (PV) (Figures S6K and S6L).

Co-staining of dopamine neuron marker tyrosine hydroxylase

(TH) with NeuN showed that a fraction of (7.5% ± 3%, SEM)

mCherry+ cells expressed TH, but contained a low level of

NeuN (Figures S6M–S6O), similar to that found previously in ro-

dent midbrain TH+ dopamine neurons (Cannon and Greenamyre,

2009). Furthermore, we found that the expression of AAV-GFAP-

mCherry in converted neurons persisted for at least 1 month after

infection (Figure S6G), consistent with previous reports (Liu et al.,

2015; Zhou et al., 2018).

Induction of Neurons with Dopaminergic Features in PD Model Mice To explore the potential in vivo effect of CasRx-mediated

neuronal reprogramming in the striatum, we used a mouse

model of PD generated by unilateral infusion of 6-hydroxydop-

amine (6-OHDA) into the right medial forebrain bundle (Boix

et al., 2015; Heuer et al., 2012; Iancu et al., 2005; Thiele

et al., 2012). This infusion induces the loss of dopamine neu-

rons in the ipsilateral ventral midbrain and degeneration of

dopaminergic projection in the ipsilateral striatum (Figures

S7A–S7C). Three weeks after 6-OHDA infusion, AAV-GFAP-

CasRx-Ptbp1 (or AAV-GFAP-CasRx as a control) together

with AAV-GFAP-mCherry were injected into the ipsilateral

striatum. Analysis of striatal cell types was performed at

different time points after AAV injection (Figures 6A and 6B).

Interestingly, we found that 6-OHDA-lesioned mice injected

with AAV-GFAP-CasRx-Ptbp1 and AAV-GFAP-mCherry

showed a high percentage of cells expressing both TH and

mCherry at 1 month after injection, and the percentage

increased at 3 months (19% ± 0.4%, SEM, n = 5 mice, at

1 month; 32% ± 7%, SEM, n = 3 mice, at 3 months) (Figures

(K) Whole-cell recording were performed in acute striatum slices (n = 22 cells). I

action potentials (20 out of 22 cells) and voltage sag rectification upon somatic c

(L) Image shows the spontaneous synaptic currents of an induced neuron.

(M) Voltage-gated currents at hyperpolarization steps.

(N) Representative images showing mCherry+TH+ cells expressed VMAT2 (yellow

cells, n = 3 mice. Scale bar, 20 mm.

(O) Uptake of FFN206 in mCherry+ cells from the virus-injected striatum region. Im

acute brain slices 1 month after AAV injection. n > = 10 slices per group. Scale b

(P) Release of FFN206 in mCherry+ cells (yellow arrowheads) from the virus-inje

bar, 10 mm.

(Q) The statistical analysis of KCl-induced release of FFN206 in slices injected w

cells. Note that the signals of soma were quantified. The 25 analyzed cellswere o

All values are presented as mean ± SEM.; unpaired t test; *p < 0.05, **p < 0.01,

6C, 6D, S7D, and S7E). Such cells were rarely observed in

mice at 1 and 2 weeks after injection or injected with control

AAVs (Figures 6C, 6D, and S7D–S7F). In addition, around

80% of TH+ cells in the virus-injected region were mCherry+

(Figures 6E and S7G), suggesting that they were mainly

derived from astrocytes. We note that the percentage of

mCherry+TH+ cells in mCherry+ cells in wild-type (non-PD

without 6-OHDA lesion) mice was lower than that of

6-OHDA lesioned mice (Figure S7H), suggesting that endoge-

nous repair mechanisms may promote the induction of TH+

neurons after injury. We also examined whether induced neu-

rons expressed the mature dopamine neuron marker dopa-

mine transporter Slc6a3 (DAT), which is present in midbrain

dopamine neurons but absent in the lesion-induced tran-

siently TH-expressing striatal neurons (Darmopil et al.,

2008). We found a high percentage (10% ± 3%, SEM, n = 5

mice, at 1 month; 31% ± 7%, SEM, n = 3 mice, at 3 months)

of mCherry+DAT+ cells in the AAV-GFAP-CasRx-Ptbp1-in-

jected striatum but not in mice injected with the control AAV

(Figures 6C, 6F, S7D, and S7I). Further co-immunostaining

of TH and DAT revealed that the majority of mCherry+TH+ cells

expressed DAT (Figures 6C, 6G, and S7D), indicating that

most induced neurons expressed the mature dopaminergic

marker. In addition, mCherry+ cells expressed two other

midbrain dopamine neuron markers, DOPA-decarboxylase

(DDC) and forkhead box protein A2 (FOXA2) (Figures S7J–

S7M), further characterizing the dopaminergic features of con-

verted neurons.

Previous studies (Darmopil et al., 2008; Ünal et al., 2015) re-

ported the presence of TH+ interneurons in the mouse striatum

after 6-OHDA lesion. Here, we have evaluated the appearance

of induced TH+ interneurons, shown by PV+, SST+, and calreti-

nin+ (CR+) cells (Rivetti di Val Cervo et al., 2017; Ünal et al.,

2015) with mCherry expression at 3 months after AAV injection,

and found that none of these interneuron markers colocalized

with mCherry+TH+ cells, suggesting that converted TH+ neurons

were not transiently induced TH+ interneurons (Figure S7N). To

explore the subtype identity of induced neurons, we co-immuno-

stained TH with two SNc A9 area-specific dopamine neuron

markers, ALDH1A1 and GIRK2, respectively (Grealish et al.,

2014; Kriks et al., 2011; Vogt Weisenhorn et al., 2016), and

DAT with a ventral tegmental area (VTA)-specific dopamine

neuron marker calbindin (Vogt Weisenhorn et al., 2016). Our re-

sults showed that almost all induced neurons expressed

ALDH1A1 and GIRK2 but not calbindin (Figures 6H-6J),

mage shows a neuron-like mCherry+ cell has the ability to generate repetitive

urrent injection (green, 4 out of 10 cells) in current-clamp mode.

arrowheads) and percentage of mCherry+VMAT2+TH+ cells in mCherry+TH+

ages showing uptake of FFN206 (blue), a fluorescent dopamine derivative, from

ar, 30 mm (upper), 20 mm (lower).

cted striatum region. Images showing KCl-induced release of FFN206. Scale

ith AAV-GFAP-mCherry plus AAV-GFAP-CasRx-Ptbp1, n = 25 mCherry+blue+

btained from 5 slices (n = 3 mice).

***p < 0.001.

Cell 181, 1–14, April 30, 2020 9

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

suggesting that induced neurons shared many characteristics

with SNc dopamine neurons.

Whole-cell recording was performed on striatal slices of in-

jected mice. We found that the majority of neuron-like mCherry+

cells (20 out of 22 cells) were capable of generating repetitive ac-

tion potentials in response to depolarizing current injection in the

current-clamp mode (Figure 6K). We also observed spontaneous

postsynaptic currents in the voltage-clamp mode (Vc = �70 mV), indicating that converted neurons received functional synaptic

inputs (Figures 6L and 6M). Moreover, in 4 out of 10 neurons

examined, we observed delayed voltage rectification (Figure 6K)

induced by hyperpolarization-activated currents (Ih), a signature

of midbrain dopamine neurons (Engel, 2016). To determine

whether induced neurons could release dopamine, we per-

formed immunostaining and found that the majority of mCherry+

TH+ cells expressed vesicular monoamine transporter 2 (VMAT2)

(Figure 6N), an essential protein that regulates the packaging,

storage, and release of dopamine (Liu and Edwards, 1997). We

further found that many cells in the virus-injected striatum region

showed uptake of a fluorescent dopamine derivative (FFN206),

an VMAT2 substrate that is able to detect active VMAT2 in intact

cells (Hu et al., 2013; Kiyoi et al., 2018; Rivetti di Val Cervo et al.,

2017), and partial reduction of the fluorescence upon high KCl

treatment, suggesting the capability of dopamine release func-

tion of the converted cells (Figures 6O–6Q). Based on the

expression of VMAT2 in the soma, and the reduction of

FFN206 in the soma after KCl treatment, we speculate that

release of dopamine from soma is the most likely mechanism,

although release from neurites could not be excluded. Taken

together, our results showed that CasRx-mediated Ptbp1

knockdown could efficiently induce neurons showing features

of dopaminergic neurons in the striatum of PD model mice.

Induced Neurons Alleviated Motor Dysfunctions in PD Mice We further examined whether induced neurons in the striatum

could alleviate the symptoms in the 6-OHDA-induced PD mouse

model (Figure 7A). The motor functions were evaluated for drug-

induced and drug-free activities. For drug-induced activities, we

first examined apomorphine-induced contralateral rotation

behavior, which is widely used for demonstrating unilateral

dopamine neuronal loss (Brooks and Dunnett, 2009). We found

that apomorphine-induced net rotation (counted as contralat-

eral-ipsilateral rotation number) was significantly diminished in

Ptbp1-knockdown mice (injected with AAV-GFAP-CasRx-

Ptbp1 and AAV-GFAP-mCherry), as compared to control mice

(injected with AAV-GFAP-CasRx and AAV-GFAP-mCherry or

with saline), to the level comparable to that found in non-lesioned

wide-type mice (Figures 7B, S7O, and S7P). Another ipsilateral

preferred rotation behavior induced by systemic amphetamine

administration (Brooks and Dunnett, 2009), which increases

intercellular dopamine concentration by inhibiting dopamine

re-uptake of DAT in the striatum (Freyberg et al., 2016; Miller,

2011), also showed marked reduction of net rotation (counted

as ipsilateral-contralateral rotation number) and ipsilateral rota-

tion ratio (counted as ipsilateral/total rotation number) in mice in-

jected with AAV-GFAP-CasRx-Ptbp1, as compared to control

mice (Figures 7C, 7D, and S7Q). These results suggest that

10 Cell 181, 1–14, April 30, 2020

induced neurons in the striatum could release sufficient dopa-

mine to reduce motor dysfunction revealed by the drug-induced

rotation behavior in the PD model mice. In addition, we examined

two drug-free motor dysfunctions, the forelimb-use asymmetry

and motor coordination, using cylinder and rotarod tests,

respectively (Brooks and Dunnett, 2009). We found that mice in-

jected with AAV-GFAP-CasRx-Ptbp1 also showed significantly

lower percentages of ipsilateral touches of the cylinder and

longer duration on the rotarod, as compared to control mice (Fig-

ures 7E and 7F). Together, these results suggested that Ptbp1

knockdown-mediated neuron conversion in the striatum allevi-

ated the motor dysfunctions in the PD mouse model.

DISCUSSION

In this work, we showed that glial cells could be efficiently con-

verted to neurons in a region-dependent manner by CasRx-

mediated downregulation of an RNA-binding protein Ptbp1. In

a NMDA-induced retina injury model, this Ptbp1 knockdown

led to MG-to-RGC conversion in the injured retina that at least

partially restored visual responses in the central visual pathway

and a vision-dependent behavior. In a 6-OHDA-induced PD

mouse model, it led to appearance of neurons with dopami-

nergic features in the striatum, and the alleviation of motor

dysfunction associated with dopamine neuron loss in the sub-

stantia nigra. These results suggest that downregulating expres-

sion of a single RNA binding protein is sufficient to transform glial

cells into different types of neurons in brain regions undergoing

neuronal loss, providing a new approach for therapeutic

application.

The REST complex is responsible for silencing a large number

of neuron-specific transcription factors (Yeo et al., 2005), and its

activity could be suppressed by miR-124, which is strongly in-

hibited by the binding protein PTBP1 in non-neuronal cells

(Xue et al., 2013; Yeo et al., 2005). Thus, downregulation of

Ptbp1 releases miR-124 activity, which suppresses REST com-

plex, leading to expression of neuron-specific transcription fac-

tors (Xue et al., 2013). Interestingly, our finding that Ptbp1 knock-

down in glial cells led to conversion into different neuronal types

in the retina versus the striatum indicates that the activation of

neuronal type-specific transcription factors resulting from

REST suppression depends on either the glial cell type (astro-

cytes versus MG) or the environmental cues (in striatum versus

retina), or both (Grande et al., 2013; Vignoles et al., 2019). Similar

situation was found for Ascl1-induced conversion of astrocyte-

to-GABAergic neuron and MG-to-bipolar cell in the midbrain

and retina, respectively (Jorstad et al., 2017; Liu et al., 2015).

To bring in vivo direct cell type conversion closer to clinical appli-

cation in various types of brain dysfunctions, it is important to

identify the instructive signals underlying glia-to-neuron conver-

sion in the future. For clinical application, cell-transplantation

based approaches could replenish the required cell types in

various brain regions, although major hurdles remain to be over-

come, such as the low viability and integration of transplanted

cells in functional circuits (Chen et al., 2016; Grealish et al.,

2014; Kim et al., 2011; Kriks et al., 2011; Rivetti di Val Cervo

et al., 2017; Torper et al., 2013; Yoo et al., 2017). The alternative

approach of glia-to-neuron conversion in vivo has previously

Figure 7. Induced Neurons Alleviated Motor Dysfunctions in PD Mice

(A) Outline of the experiment.

(B) Net rotations (contralateral-ipsilateral) induced by apomorphine injection.

(C) Net rotations (ipsilateral-contralateral) induced by amphetamine injection.

(D) The percentage of ipsilateral rotations relative to the total number of rotations (ipsilateral/total) after systemic injecting amphetamine.

(E) The percentage of spontaneous ipsilateral touches, relative to the total number of touches.

(F) Rotarod test. Results are expressed as time (second) that mice remained on an accelerating rotarod before falling. Number above the bar indicates the number

of mice per group.

All values are presented as mean ± SEM. One-way ANOVA followed by Tukey’s test, for statistics, see also Table S1. *p < 0.05, **p < 0.01, ***p < 0.001.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

been examined by expressing one or more specific transcription

factors (Jopling et al., 2011; Mertens et al., 2016). The present

approach of CasRx-mediated Ptbp1 downregulation aims to

achieve the similar goal of conversion by knocking down a key

endogenous factor that gates the activity of REST complex.

The glia-derived neurons appeared to be stable for as long as

the observation was made (up to 3 months). Notably, we found

that MG-derived RGCs could establish central projection in the

brain and relayed visual information to the dLGN neurons.

Such demonstration of circuit integration is critical for the cell-

replacement therapy of brain dysfunction.

The short hairpin RNA (shRNA) technologies, which enable

cleavage or inhibition of desired transcripts, have significant

off-target effects (Birmingham et al., 2006; Jackson et al.,

2003). Cas13-mediated knockdown has an efficiency compara-

ble to or higher than RNAi knockdown, but has substantially

reduced off-target effects (Abudayyeh et al., 2017; Cox et al.,

2017; Konermann et al., 2018), making it potentially more useful

for therapeutic applications. In addition, CasRx has the smallest

size among Cas13 family of proteins, allowing AAV packaging

when paired with a CRISPR array (encoding multiple guide

RNAs). For the gene-editing field, the use of RNA-targeting

CRISPR system CasRx could avoid the risk associated with per-

manent DNA alteration caused by the standard CRISPR-Cas9

editing (Kosicki et al., 2018; Shin et al., 2017), thus potentially

less risky for clinical application. While previous reports claimed

that Cas13 could induce collateral cleavage of RNA in human

cells by targeting exogenous overexpressed genes (Wang

et al., 2019), it is of interest to know whether CasRx also induces

collateral damage during therapeutic applications, with only

short-term CasRx-mediated targeting of endogenous genes.

Our work also demonstrated the efficiency of RNA editing by

Cell 181, 1–14, April 30, 2020 11

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

CasRx in vivo, which may become useful not only for cell

replacement but also for treating other diseases that require

downregulation of specific gene products.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d RESOURCE AVAILABILITY

B Lead Contact

B Materials Availability

B Data and Code Availability

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

B Mice

B Cell line

B Primary cell culture (astrocyte)

d METHOD DETAILS

B Transfection, qPCR and RNA-seq

B 6-OHDA PD mouse model

B Stereotactic injection and immunofluorescence stain-

ing (brain)

B Intravitreal and subretinal injection, and immunofluo-

rescence staining (retina)

B Electrophysiology (brain slice)

B FFN206 uptake and release test

B Electrophysiology (retina)

B Visual-evoked potentials

B Apomorphine-induced rotation test

B Amphetamine-induced rotation test

B Cylinder test

B Rotarod test

B Dark light preference test

d QUANTIFICATION AND STATISTICAL ANALYSIS

SUPPLEMENTAL INFORMATION

Supplemental Information can be found online at https://doi.org/10.1016/j.

cell.2020.03.024.

ACKNOWLEDGMENTS

We thank Dr. Mu-ming Poo for helpful discussions, insightful comments, and

manuscript preparation; the Optical Imaging facility, Y. Wang, Y. Zhang and Q.

Hu; FACS facility, H. Wu and L. Quan in ION; and M Zhang in IPS. We thank

Drs. J. Zhou, S. Yu, L. De, S Wang, W Ying, Q. Wang, X Shen, Y. Lu, and Y.

Zhang. for technical assistance. This work was supported by the R&D Program

of China (2018YFC2000100 and 2017YFC1001302); CAS Strategic Priority

Research Program (XDB32060000); National Natural Science Foundation of

China (31871502 and 31522037); the Shanghai Municipal Science and Tech-

nology Major Project (2018SHZDZX05); and the Shanghai City Committee of

Science and Technology Project (18411953700 and 18JC1410100).

AUTHOR CONTRIBUTIONS

H.Z. and H. Yang jointly conceived the project, designed experiments, and su-

pervised the whole project. H.Z. designed vectors and performed experi-

ments. J.S. designed and performed experiments for PD part. X.H. designed

and performed experiments for RGC part and conducted immunostaining

for PD part. C.Z. designed vectors and verified in vitro. H.L. constructed plas-

12 Cell 181, 1–14, April 30, 2020

mids and performed experiments in vitro. Z.C. and H. Yao performed VEP.

B.W. and Y. Zhang performed recordings in retinas. W.W. and H.X. performed

recordings in brain slices. Q.X. performed immunostaining, behavior test and

RNA-seq. C.T. assisted with immunostaining. M.L. assisted with dopamine

release experiments. L.W. and C.F. assisted with immunostaining. Y.S. and

Y. Zhou analyzed RNA-seq data. S.L. isolated astrocytes. F.L. and L.S. pre-

pared AAVs. H.Z. and H. Yang wrote the paper.

DECLARATION OF INTERESTS

The authors declare no competing interests.

Received: August 12, 2019

Revised: December 18, 2019

Accepted: March 10, 2020

Published: April 8, 2020

REFERENCES

Abudayyeh, O.O., Gootenberg, J.S., Konermann, S., Joung, J., Slaymaker,

I.M., Cox, D.B.T., Shmakov, S., Makarova, K.S., Semenova, E., Minakhin, L.,

et al. (2016). C2c2 is a single-component programmable RNA-guided RNA-

targeting CRISPR effector. Science 353, aaf5573.

Abudayyeh, O.O., Gootenberg, J.S., Essletzbichler, P., Han, S., Joung, J., Be-

lanto, J.J., Verdine, V., Cox, D.B.T., Kellner, M.J., Regev, A., et al. (2017). RNA

targeting with CRISPR-Cas13. Nature 550, 280–284.

Assali, A., Le Magueresse, C., Bennis, M., Nicol, X., Gaspar, P., and Rebsam,

A. (2017). RIM1/2 in retinal ganglion cells are required for the refinement of ipsi-

lateral axons and eye-specific segregation. Sci. Rep. 7, 3236.

Birmingham, A., Anderson, E.M., Reynolds, A., Ilsley-Tyree, D., Leake, D., Fe-

dorov, Y., Baskerville, S., Maksimova, E., Robinson, K., Karpilow, J., et al.

(2006). 30 UTR seed matches, but not overall identity, are associated with RNAi off-targets. Nat. Methods 3, 199–204.

Boix, J., Padel, T., and Paul, G. (2015). A partial lesion model of Parkinson’s

disease in mice–characterization of a 6-OHDA-induced medial forebrain

bundle lesion. Behav. Brain Res. 284, 196–206.

Brooks, S.P., and Dunnett, S.B. (2009). Tests to assess motor phenotype in

mice: a user’s guide. Nat. Rev. Neurosci. 10, 519–529.

Cannon, J.R., and Greenamyre, J.T. (2009). NeuN is not a reliable marker of

dopamine neurons in rat substantia nigra. Neurosci. Lett. 464, 14–17.

Chan, K.Y., Jang, M.J., Yoo, B.B., Greenbaum, A., Ravi, N., Wu, W.L., Sán-

chez-Guardado, L., Lois, C., Mazmanian, S.K., Deverman, B.E., and Gradi-

naru, V. (2017). Engineered AAVs for efficient noninvasive gene delivery to

the central and peripheral nervous systems. Nat. Neurosci. 20, 1172–1179.

Chen, Y., Xiong, M., Dong, Y., Haberman, A., Cao, J., Liu, H., Zhou, W., and

Zhang, S.C. (2016). Chemical Control of Grafted Human PSC-Derived Neurons

in a Mouse Model of Parkinson’s Disease. Cell Stem Cell 18, 817–826.

Chiu, K., Chang, R.C., and So, K.F. (2007). Intravitreous injection for establish-

ing ocular diseases model. J. Vis. Exp. (8), 313.

Cox, D.B.T., Gootenberg, J.S., Abudayyeh, O.O., Franklin, B., Kellner, M.J.,

Joung, J., and Zhang, F. (2017). RNA editing with CRISPR-Cas13. Science

358, 1019–1027.

Darmopil, S., Muñetón-Gómez, V.C., de Ceballos, M.L., Bernson, M., and

Moratalla, R. (2008). Tyrosine hydroxylase cells appearing in the mouse stria-

tum after dopamine denervation are likely to be projection neurones regulated

by L-DOPA. Eur. J. Neurosci. 27, 580–592.

East-Seletsky, A., O’Connell, M.R., Knight, S.C., Burstein, D., Cate, J.H.D.,

Tjian, R., and Doudna, J.A. (2016). Two distinct RNase activities of CRISPR-

C2c2 enable guide-RNA processing and RNA detection. Nature 538, 270–273.

Elsaeidi, F., Macpherson, P., Mills, E.A., Jui, J., Flannery, J.G., and Goldman,

D. (2018). Notch Suppression Collaborates with Ascl1 and Lin28 to Unleash a

Regenerative Response in Fish Retina, But Not in Mice. J. Neurosci. 38,

2246–2261.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

Engel, D. (2016). Subcellular Patch-clamp Recordings from the Somatoden-

dritic Domain of Nigral Dopamine Neurons. J. Vis. Exp. (117) https://doi.org/

10.3791/54601.

Freeman, J.A., and Nicholson, C. (1975). Experimental optimization of current

source-density technique for anuran cerebellum. J. Neurophysiol. 38,

369–382.

Freyberg, Z., Sonders, M.S., Aguilar, J.I., Hiranita, T., Karam, C.S., Flores, J.,

Pizzo, A.B., Zhang, Y., Farino, Z.J., Chen, A., et al. (2016). Mechanisms of

amphetamine action illuminated through optical monitoring of dopamine syn-

aptic vesicles in Drosophila brain. Nat. Commun. 7, 10652.

Gootenberg, J.S., Abudayyeh, O.O., Lee, J.W., Essletzbichler, P., Dy, A.J.,

Joung, J., Verdine, V., Donghia, N., Daringer, N.M., Freije, C.A., et al. (2017).

Nucleic acid detection with CRISPR-Cas13a/C2c2. Science 356, 438–442.

Grande, A., Sumiyoshi, K., López-Juárez, A., Howard, J., Sakthivel, B., Aro-

now, B., Campbell, K., and Nakafuku, M. (2013). Environmental impact on

direct neuronal reprogramming in vivo in the adult brain. Nat. Commun.

4, 2373.

Grealish, S., Diguet, E., Kirkeby, A., Mattsson, B., Heuer, A., Bramoulle, Y., Van

Camp, N., Perrier, A.L., Hantraye, P., Björklund, A., and Parmar, M. (2014). Hu-

man ESC-derived dopamine neurons show similar preclinical efficacy and po-

tency to fetal neurons when grafted in a rat model of Parkinson’s disease. Cell

Stem Cell 15, 653–665.

Heuer, A., Smith, G.A., Lelos, M.J., Lane, E.L., and Dunnett, S.B. (2012). Uni-

lateral nigrostriatal 6-hydroxydopamine lesions in mice I: motor impairments

identify extent of dopamine depletion at three different lesion sites. Behav.

Brain Res. 228, 30–43.

Hu, G., Henke, A., Karpowicz, R.J., Jr., Sonders, M.S., Farrimond, F., Ed-

wards, R., Sulzer, D., and Sames, D. (2013). New fluorescent substrate en-

ables quantitative and high-throughput examination of vesicular monoamine

transporter 2 (VMAT2). ACS Chem. Biol. 8, 1947–1954.

Iancu, R., Mohapel, P., Brundin, P., and Paul, G. (2005). Behavioral character-

ization of a unilateral 6-OHDA-lesion model of Parkinson’s disease in mice. Be-

hav. Brain Res. 162, 1–10.

Jackson, A.L., Bartz, S.R., Schelter, J., Kobayashi, S.V., Burchard, J., Mao, M.,

Li, B., Cavet, G., and Linsley, P.S. (2003). Expression profiling reveals off-

target gene regulation by RNAi. Nat. Biotechnol. 21, 635–637.

Jopling, C., Boue, S., and Izpisua Belmonte, J.C. (2011). Dedifferentiation,

transdifferentiation and reprogramming: three routes to regeneration. Nat.

Rev. Mol. Cell Biol. 12, 79–89.

Jorstad, N.L., Wilken, M.S., Grimes, W.N., Wohl, S.G., VandenBosch, L.S.,

Yoshimatsu, T., Wong, R.O., Rieke, F., and Reh, T.A. (2017). Stimulation of

functional neuronal regeneration from Müller glia in adult mice. Nature 548,

103–107.

Kaneko, T., Kang, Y., and Mizuno, N. (1995). Glutaminase-positive and gluta-

minase-negative pyramidal cells in layer VI of the primary motor and somato-

sensory cortices: a combined analysis by intracellular staining and immunocy-

tochemistry in the rat. J. Neurosci. 15, 8362–8377.

Kim, J., Su, S.C., Wang, H., Cheng, A.W., Cassady, J.P., Lodato, M.A.,

Lengner, C.J., Chung, C.Y., Dawlaty, M.M., Tsai, L.H., and Jaenisch, R.

(2011). Functional integration of dopaminergic neurons directly converted

from mouse fibroblasts. Cell Stem Cell 9, 413–419.

Kiyoi, T., Liu, S., Sahid, M.N.A., Shudou, M., Maeyama, K., and Mogi, M.

(2018). High-throughput screening system for dynamic monitoring of exocy-

totic vesicle trafficking in mast cells. PLoS ONE 13, e0198785.

Knott, G.J., and Doudna, J.A. (2018). CRISPR-Cas guides the future of genetic

engineering. Science 361, 866–869.

Konermann, S., Lotfy, P., Brideau, N.J., Oki, J., Shokhirev, M.N., and Hsu, P.D.

(2018). Transcriptome Engineering with RNA-Targeting Type VI-D CRISPR Ef-

fectors. Cell 173, 665–676.

Kosicki, M., Tomberg, K., and Bradley, A. (2018). Erratum: Repair of double-

strand breaks induced by CRISPR-Cas9 leads to large deletions and complex

rearrangements. Nat. Biotechnol. 36, 899.

Kriks, S., Shim, J.W., Piao, J., Ganat, Y.M., Wakeman, D.R., Xie, Z., Carrillo-

Reid, L., Auyeung, G., Antonacci, C., Buch, A., et al. (2011). Dopamine neurons

derived from human ES cells efficiently engraft in animal models of Parkinson’s

disease. Nature 480, 547–551.

Laha, B., Stafford, B.K., and Huberman, A.D. (2017). Regenerating optic path-

ways from the eye to the brain. Science 356, 1031–1034.

Liu, Y., and Edwards, R.H. (1997). The role of vesicular transport proteins in

synaptic transmission and neural degeneration. Annu. Rev. Neurosci. 20,

125–156.

Liu, Y., Miao, Q., Yuan, J., Han, S., Zhang, P., Li, S., Rao, Z., Zhao, W., Ye, Q.,

Geng, J., et al. (2015). Ascl1 Converts Dorsal Midbrain Astrocytes into Func-

tional Neurons In Vivo. J. Neurosci. 35, 9336–9355.

McCarthy, K.D., and de Vellis, J. (1980). Preparation of separate astroglial and

oligodendroglial cell cultures from rat cerebral tissue. J. Cell Biol. 85, 890–902.

Mertens, J., Marchetto, M.C., Bardy, C., and Gage, F.H. (2016). Evaluating cell

reprogramming, differentiation and conversion technologies in neuroscience.

Nat. Rev. Neurosci. 17, 424–437.

Miller, G.M. (2011). The emerging role of trace amine-associated receptor 1 in

the functional regulation of monoamine transporters and dopaminergic activ-

ity. J. Neurochem. 116, 164–176.

Mitzdorf, U. (1985). Current source-density method and application in cat ce-

rebral cortex: investigation of evoked potentials and EEG phenomena. Physiol.

Rev. 65, 37–100.

Niwa, M., Aoki, H., Hirata, A., Tomita, H., Green, P.G., and Hara, A. (2016).

Retinal Cell Degeneration in Animal Models. Int. J. Mol. Sci. 17, E110.

Qi, Y., Dai, X., Zhang, H., He, Y., Zhang, Y., Han, J., Zhu, P., Zhang, Y., Zheng,

Q., Li, X., et al. (2015). Trans-Corneal Subretinal Injection in Mice and Its Effect

on the Function and Morphology of the Retina. PLoS ONE 10, e0136523.

Quigley, H.A., and Broman, A.T. (2006). The number of people with glaucoma

worldwide in 2010 and 2020. Br. J. Ophthalmol. 90, 262–267.

Rebsam, A., Petros, T.J., and Mason, C.A. (2009). Switching retinogeniculate

axon laterality leads to normal targeting but abnormal eye-specific segregation

that is activity dependent. J. Neurosci. 29, 14855–14863.

Rheaume, B.A., Jereen, A., Bolisetty, M., Sajid, M.S., Yang, Y., Renna, K., Sun,

L., Robson, P., and Trakhtenberg, E.F. (2018). Single cell transcriptome

profiling of retinal ganglion cells identifies cellular subtypes. Nat. Commun.

9, 2759.

Rivetti di Val Cervo, P., Romanov, R.A., Spigolon, G., Masini, D., Martı́n-Mon-

tañez, E., Toledo, E.M., La Manno, G., Feyder, M., Pifl, C., Ng, Y.H., et al.

(2017). Induction of functional dopamine neurons from human astrocytes

in vitro and mouse astrocytes in a Parkinson’s disease model. Nat. Biotechnol.

35, 444–452.

Roska, B., and Sahel, J.A. (2018). Restoring vision. Nature 557, 359–367.

Rousso, D.L., Qiao, M., Kagan, R.D., Yamagata, M., Palmiter, R.D., and Sanes,

J.R. (2016). Two Pairs of ON and OFF Retinal Ganglion Cells Are Defined by

Intersectional Patterns of Transcription Factor Expression. Cell Rep. 15,

1930–1944.

Shin, H.Y., Wang, C., Lee, H.K., Yoo, K.H., Zeng, X., Kuhns, T., Yang, C.M.,

Mohr, T., Liu, C., and Hennighausen, L. (2017). CRISPR/Cas9 targeting events

cause complex deletions and insertions at 17 sites in the mouse genome. Nat.

Commun. 8, 15464.

Su, J., Huang, P., Qin, M., Lu, Q., Sang, X., Cai, Y., Wang, Y., Liu, F., Wu, R.,

Wang, X., et al. (2018). Reduction of HIP2 expression causes motor function

impairment and increased vulnerability to dopaminergic degeneration in Par-

kinson’s disease models. Cell Death Dis. 9, 1020.

Thiele, S.L., Warre, R., and Nash, J.E. (2012). Development of a unilaterally-

lesioned 6-OHDA mouse model of Parkinson’s disease. J. Vis. Exp. (60), 3234.

Torper, O., Pfisterer, U., Wolf, D.A., Pereira, M., Lau, S., Jakobsson, J., Björ-

klund, A., Grealish, S., and Parmar, M. (2013). Generation of induced neurons

via direct conversion in vivo. Proc. Natl. Acad. Sci. USA 110, 7038–7043.

Ueki, Y., Wilken, M.S., Cox, K.E., Chipman, L., Jorstad, N., Sternhagen, K.,

Simic, M., Ullom, K., Nakafuku, M., and Reh, T.A. (2015). Transgenic

Cell 181, 1–14, April 30, 2020 13

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

expression of the proneural transcription factor Ascl1 in Müller glia stimulates

retinal regeneration in young mice. Proc. Natl. Acad. Sci. USA 112,

13717–13722.

Ünal, B., Shah, F., Kothari, J., and Tepper, J.M. (2015). Anatomical and elec-

trophysiological changes in striatal TH interneurons after loss of the nigrostria-

tal dopaminergic pathway. Brain Struct. Funct. 220, 331–349.

Vignoles, R., Lentini, C., d’Orange, M., and Heinrich, C. (2019). Direct Lineage

Reprogramming for Brain Repair: Breakthroughs and Challenges. Trends Mol.

Med. 25, 897–914.

Vogt Weisenhorn, D.M., Giesert, F., and Wurst, W. (2016). Diversity matters -

heterogeneity of dopaminergic neurons in the ventral mesencephalon and its

relation to Parkinson’s Disease. J. Neurochem. 139 (Suppl 1 ), 8–26.

Wang, Q., Liu, X., Zhou, J., Yang, C., Wang, G., Tan, Y., Wu, Y., Zhang, S., Yi,

K., and Kang, C. (2019). The CRISPR-Cas13a Gene-Editing System Induces

Collateral Cleavage of RNA in Glioma Cells. Adv. Sci. (Weinh.) 6, 1901299.

Xu, H.T., Han, Z., Gao, P., He, S., Li, Z., Shi, W., Kodish, O., Shao, W., Brown,

K.N., Huang, K., and Shi, S.H. (2014). Distinct lineage-dependent structural

and functional organization of the hippocampus. Cell 157, 1552–1564.

Xue, Y., Ouyang, K., Huang, J., Zhou, Y., Ouyang, H., Li, H., Wang, G., Wu, Q.,

Wei, C., Bi, Y., et al. (2013). Direct conversion of fibroblasts to neurons by re-

programming PTB-regulated microRNA circuits. Cell 152, 82–96.

Yao, K., Qiu, S., Tian, L., Snider, W.D., Flannery, J.G., Schaffer, D.V., and

Chen, B. (2016). Wnt Regulates Proliferation and Neurogenic Potential of

14 Cell 181, 1–14, April 30, 2020

Müller Glial Cells via a Lin28/let-7 miRNA-Dependent Pathway in Adult

Mammalian Retinas. Cell Rep. 17, 165–178.

Yao, K., Qiu, S., Wang, Y.V., Park, S.J.H., Mohns, E.J., Mehta, B., Liu, X.,

Chang, B., Zenisek, D., Crair, M.C., et al. (2018). Restoration of vision after

de novo genesis of rod photoreceptors in mammalian retinas. Nature 560,

484–488.

Yeo, M., Lee, S.K., Lee, B., Ruiz, E.C., Pfaff, S.L., and Gill, G.N. (2005). Small

CTD phosphatases function in silencing neuronal gene expression. Science

307, 596–600.

Yoo, J., Lee, E., Kim, H.Y., Youn, D.H., Jung, J., Kim, H., Chang, Y., Lee, W.,

Shin, J., Baek, S., et al. (2017). Electromagnetized gold nanoparticles mediate

direct lineage reprogramming into induced dopamine neurons in vivo for Par-

kinson’s disease therapy. Nat. Nanotechnol. 12, 1006–1014.

Zhou, H., Lin, Z., Voges, K., Ju, C., Gao, Z., Bosman, L.W., Ruigrok, T.J., Hoe-

beek, F.E., De Zeeuw, C.I., and Schonewille, M. (2014). Cerebellar modules

operate at different frequencies. eLife 3, e02536.

Zhou, H., Liu, J., Zhou, C., Gao, N., Rao, Z., Li, H., Hu, X., Li, C., Yao, X., Shen,

X., et al. (2018). In vivo simultaneous transcriptional activation of multiple

genes in the brain using CRISPR-dCas9-activator transgenic mice. Nat. Neu-

rosci. 21, 440–446.

Zhu, Y., Qiao, W., Liu, K., Zhong, H., and Yao, H. (2015). Control of response

reliability by parvalbumin-expressing interneurons in visual cortex. Nat. Com-

mun. 6, 6802.

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies

Rabbit anti-PTBP1 Thermo Fisher Scientific Cat# PA581297; RRID: AB_2788516

Rabbit-anti-NeuN Millipore Cat# ABN78; RRID: AB_10807945

Rabbit-anti-NeuN Cell Signaling Technology Cat# 24307S; RRID: AB_2651140

Guinea Pig anti-NeuN Millipore Cat# ABN90; RRID: AB_11205592

Mouse anti-Flag Sigma-Aldrich Cat# F3165; RRID: AB_259529

Rabbit anti-TH Millipore Cat# AB152; RRID: AB_390204

Rat anti-DAT Millipore Cat# MAB369; RRID: AB_2190413

Rabbit anti-DDC Abcam Cat# AB3905; RRID: AB_304145

Rabbit anti-FOXA2 Abcam Cat# AB108422; RRID: AB_11157157

Rabbit anti-GFAP Agilent Cat# Z0334; RRID: AB_10013382

Mouse anti-ALDH1A1 Proteintech Cat# 60171-1-Ig; RRID: AB_10693634

Goat anti-GIRK2 Abcam Cat# AB65096; RRID: AB_1139732

Rabbit anti-Calbindin Proteintech Cat# 14479-1-AP; RRID: AB_2228318

Goat anti-VMAT2 Everest Biotech Cat# EB06558; RRID: AB_2187855

Rabbit anti-glutaminase Proteintech Cat# 12855-1-AP; RRID: AB_2110381

Mouse anti-parvalbumin Sigma-Aldrich Cat# P3088; RRID: AB_477329

Guinea Pig anti-somatostatin Synaptic Systems Cat# 366004; RRID: AB_2620126

Mouse anti-calretinin Proteintech Cat# 66496-1-lg

Mouse anti-Brn3a Millipore Cat# MAB1585; RRID: AB_94166

Rabbit anti-RBPMS Proteintech Cat# 15187-1-AP; RRID: AB_2238431

Rabbit anti-Sox9 Millipore Cat# AB5535; RRID: AB_2239761

Rabbit anti-Pax6 BioLegend Cat# 901301; RRID: AB_2565003

Rabbit anti-Prox1 Millipore Cat# AB5475; RRID: AB_177485

Rabbit anti-Pou4f3(Brn3c) Proteintech Cat# 21509-1-AP

Rabbit anti-Foxp2 Abcam Cat# AB16046; RRID: AB_2107107

Alexa Fluor� 488 AffiniPure Donkey Anti-Rabbit IgG (H+L) Jackson ImmunoResearch Labs Cat# 711-545-152; RRID: AB_2313584

Alexa Fluor� 488 AffiniPure Donkey Anti-Mouse IgG (H+L) Jackson ImmunoResearch Labs Cat# 715-545-150; RRID: AB_2340846

Cy5-AffiniPure Donkey Anti-Guinea Pig IgG (H+L) Jackson ImmunoResearch Labs Cat# 706-175-148; RRID: AB_2340462

Cy5 AffiniPure Donkey Anti-Rabbit IgG (H+L) Jackson ImmunoResearch Labs Cat# 711-175-152; RRID: AB_2340607

Cy5-AffiniPure Donkey Anti-Rat IgG (H+L) Jackson ImmunoResearch Labs Cat# 712-175-153; RRID: AB_2340672

Cy5 AffiniPure Donkey Anti-Mouse IgG (H+L) Jackson ImmunoResearch Labs Cat# 715-175-150; RRID: AB_2340819

Bacterial and Virus Strains

AAV-GFAP-mCherry This study N/A

AAV-GFAP-CasRx This study N/A

AAV-GFAP-CasRx-Ptbp1 This study N/A

AAV-GFAP-Cre-GFP Obio technology (shanghai) corp N/A

Biological Samples

Mouse brain tissue This study N/A

Mouse eye tissue Mouse strain is listed in the

"Experimental Models:

Organisms/Strains"

N/A

Chemicals, Peptides, and Recombinant Proteins

Desipramine Sigma-Aldrich Cat# D3900

6-OHDA Sigma-Aldrich Cat# H116

(Continued on next page)

Cell 181, 1–14.e1–e6, April 30, 2020 e1

Continued

REAGENT or RESOURCE SOURCE IDENTIFIER

Apomorphine Sigma-Aldrich Cat# A4393

D-amphetamine Sigma-Aldrich Cat# A008

FFN206 Abcam Cat# AB144554

NMDA Sigma-Aldrich Cat# M3262

Deposited Data

RNaseq raw data This study PRJNA610168

Experimental Models: Cell Lines

Mouse: N2a Cell bank of SIBCB N/A

Experimental Models: Organisms/Strains

Mouse: B6.Cg-Gt(ROSA)26Sortm9(CAG-tdTomato)Hze The Jackson Laboratory JAX#007909 RRID:IMSR_JAX:007909

Oligonucleotides

gRNA 1: 50-tttgtaccgactgctatgtctgggacgat-30 This study N/A

gRNA 2: 50-ggctggctgtctccagagggcaggtcaggt-30 This study N/A

gRNA 3: 50- gtatagtagttaaccatagtgttggcagcc-30 This study N/A

gRNA 4: 50-gctgtcggtcttgagctctttgtggttgga-30 This study N/A

gRNA 5: 50-tgtagatgggctgtccacgaagcactggcg-30 This study N/A

gRNA 6: 50-gcttggagaagtcgatgcgcagcgtgcagc-30 This study N/A

Ptbp1 qPCR primers Forward: 50-AGAGGAGGCTGCCAA CACTA-30

This study N/A

Ptbp1 qPCR primers Reverse: 50-GTCCAGGGTCACTGG GTAGA-30

This study N/A

CasRx qPCR primers Forward: 5’-CCCTGGTGTCCGG

CTCTAA-30 This study N/A

CasRx qPCR primers Reverse: 50-GGACTCGCCGAAGTA CCTCT-30

This study N/A

Software and Algorithms

EthoVision XT Noldus Information Technology https://www.noldus.com/ethovision-xt

ImageJ National Institutes of Health https://imagej.net/Welcome

Prism 6 GraphPad https://www.graphpad.com/

scientific-software/prism/

Adobe Illustrator Adobe https://www.adobe.com/products/

illustrator.html

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

RESOURCE AVAILABILITY

Lead Contact Further information and requests for resources should be directed to and will be fulfilled by the Lead Contact, Hui Yang (huiyang@ion.

ac.cn).

Materials Availability All unique resources generated in this study are available upon reasonable request to lead contact with a completed Materials Trans-

fer Agreement.

Data and Code Availability RNA-sequencing data were deposited in Sequence Read Archive (SRA): PRJNA610168. No unique code was generated in this study.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice C57BL/6 mice were obtained from Shanghai SLAC Laboratory Animal Co.,Ltd, and Ai9 mice (JAX#007909) were obtained from

Jackson Laboratory. Only C57BL/6 mice were used in PD-related experiments, and both C57BL/6 and Ai9 mice were used in

e2 Cell 181, 1–14.e1–e6, April 30, 2020

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

retina-related experiments. All mice used in this study were housed in a 12 hours light/dark cycle room with water and food ad libitum.

Only male mice were used in behavior tests and the ages of mice were indicated in Figures, Figure legends and Method. All animal

experiments were performed and approved by the Animal Care and Use Committee of the Institute of Neuroscience, Chinese Acad-

emy of Sciences, Shanghai, China.

Cell line The N2a cell line was obtained from Cell bank of Shanghai Institute of Biochemistry and Cell Biology (SIBCB), Chinese Academy of

Sciences (CAS), and cultured in DMEM with 10% FBS and 1% penicillin/streptomycin in a 37�C incubator under 5% CO2.

Primary cell culture (astrocyte) Isolation and culture of astrocytes were performed as previously described (McCarthy and de Vellis, 1980). In brief, the P0 mouse was

euthanized by decapitation and the mouse brain was isolated after peeling off the skin and skull. The cortex was dissected from the

intact brain and divided into two hemispheres after removing hippocampus, meninges and blood vessels on ice. Then the tissues

were cut in pieces and digested with papain in 37�C incubator for 10 min. Digested tissues were triturated into single cell by passing through pipette tips gently, and then the dissociated cells were centrifuged (1000rpm, 5min) and the pellet was re-suspended in

DMEM/F12 medium with 10% FBS (GIBCO), 1% penicillin/streptomycin, 10ng/ml hFGF and 10ng/ml hEGF. These mixed cells

were grown in culture flasks and the medium was half-replaced every 3 days. After 7 days, the flasks were shaken in 37�C for 12 hours and the medium were discarded. Purified astrocytes were digested from flasks and seeded in 6-well plates or 10-cm dishes

for the following experiments.

METHOD DETAILS

Transfection, qPCR and RNA-seq Transient transfection was conducted with 4 mg vectors expressing Ptbp1-CasRx-GFP according to the standard procedure using

Lipofectamine 3000 (Thermo Fisher Scientific). Control plasmids do not express gRNA. Around two days after transient transfection,

GFP-positive cells were collected by fluorescence activated cell sorting (FACS) and lysed for qPCR analysis. RNA was first extracted

using Trizol (Ambion) and then converted to cDNA using a reverse transcription kit (HiScript Q RT SuperMix for qPCR, Vazyme,

Biotech). The amplification was tracked by AceQ qPCR SYBR Green Master Mix (Vazyme, Biotech). For astrocytes RNA-seq, astro-

cytes were cultured in 10-cm dishes and infected with lentivirus. Around 60000 positive cells were isolated by FACS, RNA was ex-

tracted and reverse transcribed to cDNA, which was used for RNA-seq. An average of the two repeats is presented. For N2a cells,

cells were seeded in 6-well plates and transfected with 7 mg vectors expressing gRNA-CasRx-GFP according to the standard pro-

cedure using Lipofectamine 3000 (Thermo Fisher Scientific). Two days after transfection, around 50000 GFP-positive cells per sam-

ple were collected by fluorescence activated cell sorting (FACS) and lysed for qPCR analysis. Retinas were also isolated to determine

the expression of AAVs. For N2a cell RNA-seq, N2a cells were cultured in 15-cm dish and transient transfection was conducted with

70 mg plasmids. ~500000 GFP-positive (GFP top 20%) N2a cells were collected by FACS, and RNA was extracted and then con-

verted to cDNA, which was used for transcriptome-wide RNA-seq. RNA-seq data was analyzed as previously described (Zhou

et al., 2018) and presented as the mean of all repeats. The mRNA sequencing (high-throughput) was performed using Illumina

Genome Analyzer and the adapters were removed using Trimmomatic (v0.36) during sequencing. The hisat2 (v2.0.0) was used to

map qualified reads to the mouse reference genome (mm10) with default parameters. Then, the expression levels of all mapped

genes were estimated by stringtie (v2.0) and the gene expression abundances were indicated by FPKM (fragments per kilobase

of transcript per million fragments mapped). All data could be accessed with the SRA number PRJNA610168.

6-OHDA PD mouse model The procedure was based on previous study (Su et al., 2018). In brief, adult C57BL/6 mice (aged ~10 weeks) received i.p. injection of

25 mg/kg of Desipramine hydrochloride half-hour before anesthesia. After anesthesia, mice were injected with 3 mg 6-OHDA or

saline into right medial forebrain bundle according to the following coordinates: anteroposterior (A/P) = �1.2 mm, mediolateral (M/L) = �1.1 mm, dorsoventral (D/V) = �5 mm. All mice were delivered 1 mL of 4% glucose-saline solution subcutaneously 1 hour after surgery. Mice were typically allowed to recover for 3 weeks feeding with soaked food pellets.

Stereotactic injection and immunofluorescence staining (brain) AAV-PhP.eb (Chan et al., 2017) was used in this study. Stereotactic injections were performed as previously described (Zhou et al.,

2014). The mice were placed in a stereotactic frame. Next, the skin over the skull was shaven and opened using a razor. A craniotomy

with coordinates (AP +0.8 mm, ML ± 1.6 mm) was made over the boundary of frontal and parietal bones, allowing the placement of a

injection micropipette (~20 mm outside diameter at the tip). The injection micropipette was advanced into the striatum using a micro-

manipulator and the contents in the pipette were released using a hydraulic manipulator. The viral solution containing either AAV-

GFAP-CasRx-Ptbp1+AAV-GFAP-mCherry or AAV-GFAP-CasRx+AAV-GFAP-mCherry was injected slowly (~6 minutes/ml). Mice

were injected in the striatum (AP +0.8 mm, ML ± 1.6 mm and DV �2.8 mm) with high-titer AAVs (> 1 3 1013 vg/ml, 1.5 mL for non- lesioned mice aged 8-10 weeks and 3 mL for 6-OHDA-lesioned mice aged around 13 weeks). The volume ratio between GFAP-

Cell 181, 1–14.e1–e6, April 30, 2020 e3

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

mCherry and GFAP-CasRx or GFAP-CasRx-Ptbp1 were 1:15. Immunofluorescence staining was performed at different time points

after AAV injection. The brains were perfused and fixed with 4% paraformaldehyde (PFA) overnight, and kept in 30% sucrose for at

least 12 hours. Brains were sectioned after embedding and freezing, and slices with the thickness of 30 mm were used for immuno-

fluorescence staining. Brain sections were rinsed thoroughly with 0.1 M phosphate buffer (PB). Primary antibodies: Rabbit anti-

PTBP1 (1:1000, PA581297, Invitrogen), Rabbit-anti-NeuN (1:500, ABN78, Millipore), Rabbit-anti-NeuN (1:500, 24307S, Cell Signaling

Technology), Guinea Pig anti-NeuN antibody (1:500, ABN90, Millipore), Mouse anti-Flag (1:2000, F3165, Sigma), Rabbit anti-TH anti-

body (1:500, AB152, Millipore), Rat anti-DAT (1:100, MAB369, Millipore), Rabbit anti-DDC antibody (1:400, ab3905, Abcam), Rabbit

anti-FOXA2 antibody (1:50, ab108422, Abcam) Rabbit anti-GFAP antibody (1:500, Z0334, DAKO), Mouse anti-ALDH1A1 antibody

(1:50, 60171-1-Ig, Proteintech), Goat anti-GIRK2 antibody (1:100, ab65096, Abcam), Rabbit anti-Calbindin antibody (1:1000,

14479-1-AP, Proteintech), Goat anti-VMAT2 antibody (1:50, EB06558, Everest biotech), Rabbit anti-glutaminase antibody

(1:50,12855-1-AP, Proteintech), Mouse anti-parvalbumin(PV) antibody (1:2000, P3088, Sigma), Guinea Pig anti-somatostatin-

28(SST) antibody (1:500, 366004, Synaptic Systems), Mouse anti-calretinin(CR) antibody (1:100, 66496-1-Ig, Proteintech). Second-

ary antibodies: Alexa Fluor� 488 AffiniPure Donkey Anti-Rabbit IgG (H+L) (1:500, 711-545-152, Jackson ImmunoResearch) Alexa Fluor� 488 AffiniPure Donkey Anti-Mouse IgG (H+L) (1:500, 715-545-150, Jackson ImmunoResearch) Cy5-AffiniPure Donkey Anti-Guinea Pig IgG (H+L) (1:500, 706-175-148, Jackson ImmunoResearch) Cy5 AffiniPure Donkey Anti-Rabbit IgG (H+L) (1:500,

711-175-152, Jackson ImmunoResearch) Cy5-AffiniPure Donkey Anti-Rat IgG (H+L) (1:500, 712-175-153, Jackson ImmunoRe-

search) were used in this study. After antibody incubation, slices were washed and covered with mountant (Life Technology). Images

were visualized under an Olympus FV3000 microscope.

Intravitreal and subretinal injection, and immunofluorescence staining (retina) NMDA and AAVs (AAV-PHP.eB) were introduced via intravitreal and subretinal injection, respectively, as previously described (Chiu

et al., 2007; Qi et al., 2015). For intravitreal injection, the pipettes were prepared using a puller and connected with a 1 mL syringe.

Then mice were anaesthetized with 0.35 mL 2% Tribromoethanol and one drop of 0.5% alcaine was droped on the eye before intra-

vitreous injection. Around 1.5 mL NMDA solution (200mM) was injected into the vitreous body using the pipette. After injection, oflox-

acin eye ointment was applied on the eye to prevent infection. For subretinal injection, mice were aneathetized and the pupil size was

dilated with Tropicamide Phenylephrine Eye drops. Then one drop of sodium hyaluronate was droped on the cornea to enable better

visualization. A penetration was made in the cornea under a Olympus microscope (Olympus, Tokyo, Japan) using a 30G needle. Next,

a Hamilton syringe (32G needle) was inserted into the eye via corneal perforation. To inject AAVs (high titer: > 1 3 1013 vg/ml) into the

subretinal space, an inner retina area with low density of blood vessels was penetrated with the needle and ~1 mL content was in-

jected into the subretinal space with slow speed (taking up to 20 s). After injection, the injection needle was removed slowly and

a drop of ofloxacin eye ointment was administrated. To determine the conversion in intact retinas, in total 1 mL AAV-GFAP-GFP-

Cre (0.2 ml) and AAV-GFAP-CasRx-Ptbp1 (0.8 ml), or AAV-GFAP-Cre-GFP (0.2 ml) and AAV-GFAP-CasRx (0.8 ml) were delivered to

the retina via subretinal injection (Ai9 and C57BL/6 mice, 5 weeks old). To determine the reprogramming in damaged retinas,

NMDA was dissolved in the PBS with a final concentration of 200 mM and 1.5 mL NMDA solution was delivered to the eye of Ai9

mice aged 4-8 weeks or C57BL/6 mice aged 5-6 weeks (For VEP and dark/light preference test) via intravitreal injection. Two-three

weeks after NMDA injection, AAV-GFAP-GFP-Cre together with AAV-GFAP-CasRx-Ptbp1 or AAV-GFAP-CasRx were co-delivered

to the retina via subretinal injection. To evaluate the functional rescue of injured retinas (VEP and light/dark transition test), retinal

injury was induced in mice aged 5-6 weeks (C57BL/6) by NMDA injection and AAV-GFAP-mCherry (0.2 ml) mixtured with AAV-

GFAP-CasRx-Ptbp1 (0.8 ml) or AAV- GFAP-CasRx (0.8 ml) were delivered 2-3 weeks after NMDA injection. The eyes, optic nerves

and brain were extracted at different time points after AAV injection and fixed with 4% paraformaldehyde (PFA) for 2 (for eyes and

optic nerves) or 24 (for brains) hours, then maintain in a 30% sucrose solution for 2 (for eyes) or 24 (for brains) hours. After embedding

and freezing, eyes and brains were sectioned with the thickness of 30 mm. After incubation, slices were washed and covered with

mountant (Life Technology). Images were visualized under an Olympus FV3000 microscope.

Electrophysiology (brain slice) Electrophysiological recordings were performed as previously described (Xu et al., 2014). In brief, mice were anesthetized and

perfused transcardially with ice-code NMDG artificial cerebrospinal fluid (aCSF) [NMDG aCSF (mM): NMDG 93, KCl 2.5, NaH2PO4 1.25, NaHCO3 30, HEPES 20, glucose 25, thiourea 2, sodium ascorbate 5, sodium pyruvate 3, N-Acetyl-L-cysteine (NAC) 12, CaCl2 0.5, MgSO4 10] gassed with 95% O2 and 5% CO2 at room temperature. The brains were extracted after perfusion and placed into the

ice-cold NMDG aCSF solution for 30 s. The brains were trimmed and sectioned at 300 mm thickness with the speed of 0.04–

0.05 mm/s. Brain slices were moved into a chamber filled with gassed NMDG aCSF and kept at 32-34�C for % 12 min. The slices were then transferred into a new chamber containing gassed recording aCSF [recording aCSF (mM): NaCl 126, KCl 3, NaH2PO4

1.2, NaHCO3 26, glucose 10, CaCl2 2.4, MgCl 1.3] at room temperature for one hour. Neuon-like (with clearly visualized large

soma) mCherry-positive cells were patched under the microscope Olympus BX51WI, and data was acquired using Clampex 10.

FFN206 uptake and release test FFN206 take-up and release experiments were performed as previously described (Rivetti di Val Cervo et al., 2017). Mice were anes-

thetized and perfused with artificial cerebrospinal fluid (aCSF). The brains were extracted and placed into the icy-cold aCSF solution

e4 Cell 181, 1–14.e1–e6, April 30, 2020

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

for 30 s. The brains were sectioned at 300 mm thickness. For imaging, the slices were incubated 30 minutes in the dark in ACSF

containing a final concentration of 10 mM FFN 206 (ab144554, Abcam), then FFN 206 was washed out for 30 minutes with aCSF

(osmotically balanced). Slices were imaged on an Olympus FV3000 confocal microscope, and stimulation of FFN206 release was

induced by addition KCl (wash-in) with a final concentration of 56 mM.

Electrophysiology (retina) Mice were dark-adapted overnight before euthanasia. Dissection of the retina was performed in oxygenated (95% O2 / 5% CO2) arti-

ficial cerebrospinal fluid (ACSF) containing 126 mM NaCl, 2.5 mM KCl, 1.25 mM NaH2PO4, 2 mM MgCl2, 2 mM CaCl2, 26mM NaHCO3 and 10 mM glucose under an infrared microscope at room temperature. A piece of the retina was placed with RGCs facing up in a

recording chamber on the stage of an upright microscope. The tdTomato-positive cells in ganglion cell layer were identified using

two-photon (l = 1030 nm) microscopy and targeted for loose cell-attached recording under infrared light. Pipettes (4-7 MU) for

recording were filled with ASCF and 0.25 mM Alexa488 hydradize. Recordings were acquired with a Multiclamp 700A amplifier

and pClamp 10 software suite (Molecular Devices). Signals were low-pass filtered at 1 kHz and digitized at 10 kHz. The full field light

stimulus was delivered by a white LED. After recording current pulses were injected to fill the cell for visualizing the morphology.

Visual-evoked potentials The mouse was injected with a mixture of fentanyl (0.05 mg/kg), midazolam (5 mg/kg), and meditomidine (0.5 mg/kg) intraperitone-

ally. The mouse was head-fixed in a stereotaxic apparatus, and its body temperature was maintained at 37�C by means of a heating blanket. A craniotomy (~1 mm diameter) was made above both sides of the primary visual cortex (V1) (AP �3.6 to - 3.9 mm, ML 2.2 mm), and the dura was removed. Visual stimuli were presented on a 17’’ LCD monitor (Dell P170S, max luminance of

69 cd/m2) placed 8 cm away from the eye contralateral to the recording site. The eye ipsilateral to the recording site was blocked

from visual stimulation. We displayed 100 repeats of flash stimuli (full field, 100% contrast) for 2 s with an interval of 2 s. Recordings

were made with multi-site silicon probes (A1 3 16-5mm-50-177, NeuroNexus Technologies) in V1 (AP �3.6 to - 3.9 mm, ML 2.2 mm), with the tip of the electrode reaching cortical depth at around 900 mm for each recording. Both the reference and ground wires were

placed in a small craniotomy at least 3 mm away from the recording site. The neural responses were amplified and filtered using a

Cerebus 32-channel system (Blackrock microsystems). Local field potential (LFP) signals were sampled at 2 kHz or 10 kHz with a

wide-band front-end filter (0.3 – 500 Hz). The LFP responses to the full-screen flash stimuli were used for current source density

(CSD) analysis to identify the location of cortical layer 4 (Zhu et al., 2015). To generate CSD profile, we computed the second spatial

derivative of LFP by the following equation (Freeman and Nicholson, 1975; Mitzdorf, 1985):

v 2 f

vz2 z

fðz � nDzÞ + fðz + nDzÞ � 2fðzÞ ðnDzÞ2

where f is LFP, z is the coordinate of the recording sites, Dz is th

e distance between adjacent recording sites, and nDz is the differ-

entiation grid (n = 2). Layer 4 (the granular layer) was determined as those recording sites at the initial current sink. For each mouse,

we used a layer 4 channel showing the largest mean amplitude to analyze visually evoked response.

Apomorphine-induced rotation test All tested mice were placed in the testing room for habituation before each test. After 30 minutes habituation, mice were i.p. injected

with 0.5 mg/kg apomorphine (A4393, Sigma-Aldrich) dissolved in saline. After 10 minutes, each mouse was placed in an opaque cyl-

inder (30 cm diameter) for free moving with a camera recording above for 20 min. After each trial, the cylinders were cleaned with 70%

ethanol to eliminate the olfactory cues. Raw data were obtained by counting the rotation number in the video. A rotation was defined

as a full-body turn with one hind-paw as center and without switching head direction. Both ipsi- (clockwise) and contra-lateral (coun-

terclockwise) rotation was counted. Data were quantified as the net contralateral rotation number (n = contralateral rotation – ipsi-

lateral rotation) during 20 minutes.

Amphetamine-induced rotation test All mice performed amphetamine test 1 week after apomorphine test to avoid potential result interference between these two tests.

After 30 minutes habituation, mice received i.p. injection of 2.5mg/kg D-amphetamine (A008, Sigma-Aldrich, the solvant was evap-

orated) dissolved in saline 10 minutes before starting test and then recorded with same apparatus used in apomorphine test for 40 mi-

nutes. A rotation in this test was defined as a complete 360� circling without changing direction. Videos were analyzed by EthoVision XT (version 11.5.1012) and both ipsi- (clockwise) and contra-lateral (counterclockwise) rotation was counted. Data were quantified as

the net ipsilateral rotation number (n = ipsilateral rotation – contralateral rotation) and ratio of ipsilateral rotation to total rotations (n =

ipsilateral rotation/ (ipsilateral + contralateral rotation)) during 40 minutes.

Cylinder test After habituation, each mouse was gently placed in a glass beaker (1000 mL volume) and allowed to move freely with a camera

recording in front of it for 10 min. After each trial, the beakers were cleaned with 70% ethanol to eliminate the olfactory cues. Raw

Cell 181, 1–14.e1–e6, April 30, 2020 e5

Please cite this article in press as: Zhou et al., Glia-to-Neuron Conversion by CRISPR-CasRx Alleviates Symptoms of Neurological Disease in Mice, Cell (2020), https://doi.org/10.1016/j.cell.2020.03.024

data were obtained by counting the paw touches on the wall in the video. Both ipsi- and contra-lateral paw touches on the wall were

separately counted and data were quantified as the ratio of ipsilateral touches to total touches (n = ipsilateral touches/ (ipsilateral +

contralateral touches)).

Rotarod test All mice were trained for 2 days before testing in the third day and both training and testing were started after habituation. On day 1,

mice were trained with a fixed speed of 4rpm in a period of 300 s on rotarod apparatus (Ugo Basile) for 4 times. On day 2 and 3, mice

were trained and tested with a accelerate speed from 4 to 40 rpm for 4 times. The rods were cleaned with 70% ethanol after every trial

to avoid potential odor cues. The time a mouse spent on rod before falling off was recorded as latency and the mean value of 3 longest

latencies was used for analysis.

Dark light preference test Both eyes were injured and injected with either AAV-GFAP-CasRx plus AAV-GFAP-mCherry or control AAVs (both eyes). The appa-

ratus used for the light/dark transition test consists of a box divided into a small (one third) dark section and a large (two thirds) illu-

minated section (550 lumens) with a door. Mice were allowed to move freely between the two compartments for 10 min. The time

spent in each chamber was recorded by the camera and analyzed using Ethovision XT. After each trial, the compartments were

cleaned with 70% ethanol to eliminate the olfactory cues.

QUANTIFICATION AND STATISTICAL ANALYSIS

The percentage of mCherry+TH+ cells in mCherry+ cells (WT mice without 6-OHDA) was quantified one month after AAV injection. To

avoid the possibility that the signal of converted TH+ neurons may be masked by intense TH signals of dopamine neurites in the intact

striatum, we carefully evaluate the TH signal of mCherry+ cells via layer by layer scanning (every layer is 1 mm). The co-transduction

efficiency of two AAVs (n = 5 mice) was quantified in the striatum (without 6-OHDA lesion) one week after injecting AAV-GFAP-

mCherry and AAV-GFAP-CasRx-Ptbp1. And the absolute number of AAV-GFAP-CasRx-Ptbp1 (n = 3 mice) was quantified (without

6-OHDA lesion) one month after injecting AAV-GFAP-mCherry and AAV-GFAP-CasRx-Ptbp1. The percentage of mCherry+GFAP+

cells in GFAP+ cells was quantified in the striatum (without 6-OHDA lesion) one month after injecting AAV-GFAP-mCherry and

AAV-GFAP-CasRx (n = 3 mice). The absolute number of TH+ and DAT+ cells were quantified in the striatum (with 6-OHDA lesion)

one and three months after injecting AAV-GFAP-mCherry and AAV-GFAP-CasRx-Ptbp1. Number of mice used in each experiment

were indicated in figures and figure legends. All values are presented as mean ± SEM. Unpaired two-tailed Student’s t test was used

to determine the statistical significance (p < 0.05), except Figures 6Q and S7O (paired two-tailed Student’s t test), and Figure 7 (one-

way ANOVA followed by Tukey’s test) and Figures S7P and S7Q (two-way ANOVA followed by bonferroni’s test). Randomization was

used in all experiments and no statistical methods were used to pre-determine sample sizes. Data distribution was assumed to be

normal but this was not formally tested.

e6 Cell 181, 1–14.e1–e6, April 30, 2020

Supplemental Figures

Figure S1. Characterization of AAV Specificity and Expression, Turning Off of GFP Over Time and Subtypes of Converted RGCs, Related to

Figure 1

(A) Expression levels in log2 (FPKM + 1) values of all detected genes in RNA-seq libraries of CasRx-Ptbp1 (y axis) compared to CasRx control (x axis). N2a cells,

n = 4 independent replicates; astrocytes, n = 2 independent replicates. Typical neuronal genes are depicted as black dots. (B) Specificity of AAV-GFAP-GFP-Cre.

AAV-GFAP-GFP-Cre drove GFP expression and unlocked tdTomato expression specifically in the MG of Ai9 mice. Sox9 is a MG-specific marker. Scale bar:

50 mm. (C) Confirmation of AAV expression. Percentage of GFP+ cells expressing tdTomato, and percentage of tdTomato+ cells expressing Sox9. All values are

presented as mean ± SEM. (D) qPCR analysis of infected retinas confirmed the expression AAV-GFAP-CasRx and AAV-GFAP-CasRx-Ptbp1. Note that

confirmation of AAV expression by immunostaining Flag-tag (fused with CasRx) was successful in the brain (data not shown) but not in the retinas. Number above

the bar indicates the number of repeats per group. All values are presented as mean ± SEM. (E) MG-derived RGCs eventually turned off expression of GFP over

time. White arrowheads indicate that a tdTomato+Rbpms+ cell expressed low level of GFP and yellow arrowheads indicate that a MG-derived RGC turned off

expression of GFP, at one month post-injection (Ai9 mice). Scale bar. 20 mm. Experiments were independently repeated 6 times per group with similar results. (F-

H) Staining of Foxp2, Brn3c and Parvalbumin. Note that Foxp2, Brn3c and Parvalbumin are RGC subtype markers for F-RGCs, subtype 3 RGCs and PV-RGCs,

respectively. Yellow arrowheads indicate that tdTomato+ cells colocalize with different markers and white arrowheads indicate that tdTomato+ cells do not

colocalize with different markers. Scale bar. 20 mm. Experiments were independently repeated 3 times per group with similar results.

Figure S2. Knockdown of Ptbp1 Converted MG into RGCs in the Intact Retinas of C57BL/6 Mice, Related to Figure 2

(A) Schematic illustration of induction of RGCs from MG. Vector 1 (GFAP-mCherry) encodes mCherry driven by the MG-specific promoter GFAP and Vector 2

(AAV-EFS-CasRx-Ptbp1) encodes gRNAs and CasRx under a ubiquitous promoter. To induce RGCs, retinas were either injected with AAV-GFAP-mCherry plus

(legend continued on next page)

AAV-EFS-CasRx-Ptbp1, or AAV-GFAP-mCherry alone as a control. Occurrence of conversion was examined 2-3 weeks after injection. (B) Representative

images showing colocalization of mCherry and MG marker Sox9, suggesting that AAV-GFAP-mCherry specifically expressed in MG. Scale bar, 50 mm. (C, D)

Representative images showing mCherry+Brn3a+ and mCherry+Rbpms+ cells in the GCL. n = 3 retinas per group. Scale bar, 50 mm. (E) A representative RGC-like

mCherry+ cell showed (4 out of 4 cells, all cells were ON-cell) responses to LED light.

Figure S3. Knockdown of Ptbp1 Converted MG into Amacrine Cells but No Other Types of MG-Derived Neurons, Related to Figure 2

(A) Observation of tdTomato+Pax6+ cells in the intact retinas of Ai9 mice injected with AAV-GFAP-CasRx-Ptbp1. Green arrowheads indicate that tdTomato+ cells

do not colocalize with Pax6, and yellow arrowheads indicate the colocalization of Pax6 and tdTomato. Note that Pax6 is marker for amacrine cells. Scale bar,

20 mm. (B) tdTomato+Prox1+ cells were not observed. Arrowheads indicate that tdTomato cells do not colocalize with Prox1, a marker for bipolar cells. Scale bar,

20 mm. (C) tdTomato+ cells were not observed in the layer (ONL) of photoreceptor cells. White arrowheads indicate tdTomato+ RGC like cells in the GCL, yellow

arrowhead indicates tdTomato+ amacrine like cells in the INL, and green arrowheads indicate tdTomato+ projections of MG. Scale bar, 20 mm. Experiments were

independently repeated at least three times per group with similar results.

Figure S4. Progression of MG-to-RGC Conversion, Related to Figure 3

(A) Representative images showing progressive MG-to-RGC conversion in the intact retinas (without NMDA injection) at five different time points. Arrowheads

indicate induced RGCs. Scale bar, 20 mm. Experiments were independently repeated > = 3 times with similar results. Note that representative images from

‘‘1 month’’ are also shown in Figures 2B and 2D. (B) The absolute number of tdTomato+Brn3a+ and tdTomato+Rbpms+ cells in the GCL. Note that the values from

‘‘1 month’’ are also shown in Figures 2C and 2E. All values are presented as mean ± SEM. (C) Representative images showing the intermediate stage of induced

cells (white arrowheads) at 1.5 week after AAV injection. n > = 3 mice per group. Note that Intermediate cells typically lost their projections in ONL. Scale bar,

10 mm. Experiments were independently repeated > = 3 times with similar results. (D) Representative images showing progressive MG-to-RGC conversion in the

NMDA-injured retinas at four different time points. Arrowheads indicate induced RGCs. Scale bar, 20 mm. Note that representative images from ‘‘1 month’’ are

also shown in Figures 3C and 3E. Experiments were independently repeated > = 2 times with similar results.

Figure S5. The Time Course of Induced RGCs in the Injured Retinas Sent Their Projections into Optic Nerve and Central Brain Regions,

Related to Figure 5

(A) Representative images showing progressive increase of tdTomato+ axons in the optic nerve with NMDA-induced injury at three different time points (1 week,

2 weeks, 3 weeks). Images were tiled. Scale bar, 50 mm. Experiments were independently repeated 2 times with similar results. (B) Progressive increase in the

density of tdTomato+ axons (yellow arrowheads) in dLGN. Images were tiled. Scale bar, 500 mm (upper), 50 mm (lower). Experiments were independently repeated

2 times with similar results. (C) Progression of tdTomato+ axons (yellow arrowheads) in SC. Images were tiled. Scale bar, 500 mm (upper), 50 mm (lower). Ex-

periments were independently repeated 2 times with similar results.

Figure S6. AAVs Infected Astrocytes with a High Efficiency in the Striatum and Neuronal Subtypes of Induced Neurons, Related to Figure 6

(A, B) Schematic illustration of the injection strategy. Vector I (AAV-GFAP-mCherry) encodes mCherry driven by the astrocyte-specific promoter GFAP. Vector II

(AAV-GFAP-CasRx-Ptbp1) carries CasRx under GFAP promoter and gRNAs targeting Ptbp1. Striatum was either injected with AAV-GFAP-CasRx-Ptbp1 or

control vector AAV-GFAP-CasRx together with AAV-GFAP-mCherry. Occurrence of conversion is evaluated around one-month post-injection. ST, striatum. (C)

Colocalization of mCherry and GFAP, showing that mCherry driven by the astrocyte-specific promoter labels astrocyte with a high efficiency. Histogram shows

percentage of mCherry+ cells expressing GFAP, n = 3 mice. Scale bar, 20 mm. (D) Colocalization of Flag (fused with CasRx) and GFAP, one week after AAV

injection. Scale bar, 20 mm. (E) Representative images showing Ptbp1 expression (detected by Ptbp1 antibody) was downregulated in the striatum one week after

AAV injection. Expression of CasRx (fused with Flag) is detected by antibody against Flag tag. n = 4 mice for each group. Scale bar, 10 mm. (F) Quantification of

Ptbp1 fluorescence intensity using ImageJ. The a.u. represents arbitrary unit. (G) Representative images of the striatum. NeuN is a specific marker for mature

neurons. White arrowheads indicate that NeuN expression does not colocalize with mCherry, and yellow arrowheads indicate the co-localization of NeuN and

mCherry. Scale bar: 50 mm. (H) Percentage of mCherry+NeuN+ cells in mCherry+ cells (n = 6 mice per group; t = - 4.7, p < 0.001). (I) Images show that a mCherry+

NeuN+Glutaminase+ cell (pink arrowheads) is adjacent to a mCherry+NeuN+Glutaminase- cell (yellow arrowheads). Scale bar, 10 mm. (J) Percentage of

mCherry+Glutaminase+NeuN+ cells in mCherry+NeuN+ cells. n = 5 mice. (K) Representative images showing that mCherry+NeuN+ cells rarely colocalize with

Somatostatin. Yellow arrowheads indicate mCherry+NeuN+Somatostatin- cells and Pink arrowheads indicate a mCherry+NeuN+Somatostatin+ cell. The control

AAV showed the absence of mCherry+SST+ cells, indicating that SST+ cells could not be infected by AAV-GFAP-mCherry. Experiments were repeated 5 times

independently with similar results. Scale bar, 20 mm. (L) Representative image showing that mCherry+NeuN+ cells do not colocalize with Palvabumin. Experiments

were repeated 4 times independently with similar results. Scale bar, 20 mm. (M) Observation of mCherry+TH+ cells (white arrowheads) in the intact striatum. Scale

bar, 20 mm. (N) The percentage of mCherry+TH+ cells in mCherry+ cells, n = 5 mice per group. (O) Typically, mCherry+TH+ cells showing low (upper, white ar-

rowheads) or undetectable level of NeuN expression (lower, yellow arrowheads). Scale bar, 5 mm. Experiments were repeated 5 times independently with similar

results.

All values are presented as mean ± SEM; unpaired t test; *p < 0.05, **p < 0.01, ***p < 0.001.

Figure S7. Knockdown of Ptbp1-Induced Neurons Expressing Dopaminergic Markers in a 6-OHDA-Induced Mouse Model of PD, Related to

Figure 6

(A) Outline of the experiment. (B) Staining shows depletion of TH+ neurons (green) in the ipsilateral substantia nigra (relative to the side of 6-OHDA infusion).

Images were tiled. Scale bar, 100 mm. Experiments were independently repeated 12 times with similar results. (C) DAT staining shows depletion of dopamine

neuron fibers (green) in the ipsilateral striatum. Images were tiled. Scale bar, 500 mm. Experiments were independently repeated > 10 times with similar results.

(D, E) Representative confocal images showing mCherry+TH+ and mCherry+DAT+ cells at different time points after AAV injection, and Quantification of

mCherry+TH+ and mCherry+DAT+ cells. n > = 3 mice per group. Scale bar, 50 mm. Note that the data ‘‘GFAP-CasRx, 1 month,’’ ‘‘GFAP-CasRx-Ptbp1, 1 month’’

and ‘‘GFAP-CasRx-Ptbp1, 3 months’’ are also shown in Figures 6C, 6D, and 6F. Scale bar, 50 mm. (F) The absolute number of TH+ cells in PD model mice. (G)

Confocal images of mCherry+ TH+ cells and percentage of mCherry+TH+ cells in TH+ cells, n = 5 mice per group. Scale bar, 50 mm. (H) The percentage of

mCherry+TH+ cells in mCherry+ cells. Note that the data are also shown in Figures 6D and S6N. (I) The absolute number of DAT+ cells in PD model mice. (J, K)

Confocal images of mCherry+ DDC+ cells and percentage of mCherry+DDC+ cells in mCherry+ cells, n = 5 mice per group. DDC is a dopamine neuron marker.

Scale bar, 50 mm. (L, M) Representative images showing colocalization of FOXA2 and mCherry (yellow arrowheads), and percentage of mCherry+FOXA2+ cells in

(legend continued on next page)

mCherry+ cells. n = 5 mice per group. FOXA2 is a dopamine neuron-marker. Scale bar, 30 mm. (N) Representative images showing mCherry+TH+ cells did not

colocalize with representative striatal interneuron markers PV, SST and CR. Yellow arrowheads indicate mCherry+TH+ cells, and blue arrowheads indicate PV+,

SST+ or CR+ cells. Note that SST colocalized with mCherry but not TH, consistent with the results in Figure S1H showing that knockdown of Ptbp1 could sparsely

induce SST+ neurons in the intact striatum without 6-OHDA lesion. Experiments were independently repeated 3 times with similar results. Scale bar, 20 mm. (O)

Comparison of the number of net rotations before and one month after AAV injection. Number above the dots indicates the number of mice per group. Paired

t test. For statistics, see also Table S2. The data for ‘‘1 month’’ are also shown in Figure 7. (P) Net rotations (rotations/min) induced by apomorphine injection.

Behavior was assessed at 1 month and 3 months for each mouse, n = 3 mice per group. Two-way ANOVA followed by bonferroni’s test. The data

for ‘‘1 month’’ are also shown in Figure 7. (Q) Net rotations (rotations/min) induced by Amphetamine injection, n = 3 mice per group. Two-way ANOVA followed by

bonferroni’s test. The data for ‘‘1 month’’ are also shown in Figure 7.

All values are presented as mean ± SEM; unpaired t test; *p < 0.05, **p < 0.01, ***p < 0.001.

View publication statsView publication stats