Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood...

6
Transdifferentiation of human adult peripheral blood T cells into neurons Koji Tanabe a,b,1 , Cheen Euong Ang a,b,c,1 , Soham Chanda a,b,d , Victor Hipolito Olmos a,b , Daniel Haag a,b , Douglas F. Levinson e , Thomas C. Südhof d,2 , and Marius Wernig a,b,2 a Department of Pathology, Stanford University, Stanford, CA 94305; b Institute for Stem Cell Biology and Regenerative Medicine, Stanford University, Stanford, CA 94305; c Department of Bioengineering, Stanford University, Stanford, CA 94305; d Department of Molecular and Cellular Physiology and Howard Hughes Medical Institute, Stanford University, Stanford, CA 94305; and e Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA 94305 Contributed by Thomas C. Südhof, May 3, 2018 (sent for review November 21, 2017; reviewed by Thomas Graf and Hideyuki Okano) Human cell models for disease based on induced pluripotent stem (iPS) cells have proven to be powerful new assets for investigating disease mechanisms. New insights have been obtained studying single mutations using isogenic controls generated by gene targeting. Modeling complex, multigenetic traits using patient- derived iPS cells is much more challenging due to line-to-line variability and technical limitations of scaling to dozens or more patients. Induced neuronal (iN) cells reprogrammed directly from dermal fibroblasts or urinary epithelia could be obtained from many donors, but such donor cells are heterogeneous, show interindividual variability, and must be extensively expanded, which can introduce random mutations. Moreover, derivation of dermal fibroblasts requires invasive biopsies. Here we show that human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted into fully functional iN cells, demonstrating that terminally differen- tiated human cells can be efficiently transdifferentiated into a distantly related lineage. T cell-derived iN cells, generated by non- integrating gene delivery, showed stereotypical neuronal morphol- ogies and expressed multiple pan-neuronal markers, fired action potentials, and were able to form functional synapses. These cells were stable in the absence of exogenous reprogramming factors. Small molecule addition and optimized culture systems have yielded conversion efficiencies of up to 6.2%, resulting in the generation of >50,000 iN cells from 1 mL of peripheral blood in a single step without the need for initial expansion. Thus, our method allows the generation of sufficient neurons for experi- mental interrogation from a defined, homogeneous, and readily accessible donor cell population. induced neuronal cells | direct conversion | transdifferentiation | disease modeling | iN cells A dvances in cell reprogramming and genome editing tools have provided new ways to interrogate human gene function in various human cellular contexts, such as neurons. In particu- lar, genetic engineering of embryonic or induced pluripotent stem (iPS) cells has proven powerful for dissecting the specific consequences of disease-associated mutations in controlled ge- netic backgrounds (1, 2). However, these methods cannot be expected to provide fully adequate cellular models of diseases for which highly polygenic mechanisms underlie risk. For example, large-scale genome-wide association study data suggest that 3050% of the genetic risk for each of the neuropsychiatric disorders that have been studied to date can be explained by the joint effects of thousands of common genetic variants with small in- dividual effects, such that individual patients are likely to be carrying a unique combination of many contributory variants (3). One way to study such complex genetic backgrounds in human neurons is by reprogramming patient cells to iPS cells (4). However, iPS cells have significant line-to-line variability in terms of differentiation capacity, presumably due to variations in their epigenetic and pluripotent state (57). Moreover, iPS cells are often karyotypically unstable when grown in feeder-free conditions, and their growth and formation is labor-intensive and difficult to scale from a large number of individuals. Another way to obtain neurons is by deriving induced neuro- nal (iN) cells from fibroblasts in a single conversion step, which in principle would greatly facilitate their derivation from many patients (8). However, unlike neonatal human fibroblasts, adult human fibroblasts have proven difficult to reprogram into syn- aptically competent iN cells (914). Moreover, fibroblasts are heterogeneous and ill-defined and must be expanded in vitro from invasive and painful skin biopsies to obtain sufficient numbers, increasing the risk of acquiring random genetic muta- tions during an extended culture period. Here we report that functional synapse-forming human iN cells can be induced from freshly isolated and stored adult peripheral T cells using non- integrating episomal vectors. Previous studies have shown the conversion of blood and urinary cells into various neural pro- genitor cells that only inefficiently gave rise to functional neu- rons (1521). The described conversions were accomplished with transient expression of iPS cell reprogramming factors, an Significance Recent advances in genomics have revealed that many poly- genetic diseases are caused by complex combinations of many common variants with individually small effects. Thus, building informative disease models requires the interrogation of many patient-derived genetic backgrounds in a disease-relevant cell type. Current approaches to obtaining human neurons are not easy to scale to many patients. Here we describe a facile, one- step conversion of human adult peripheral blood T cells directly into functional neurons using episomal vectors without the need for previous in vitro expansion. This approach is more amenable than induced pluripotent stem cell-based approaches for application to larger cohorts of individuals and will enable the development of functional assays to study complex human brain diseases. Author contributions: K.T., C.E.A., T.C.S., and M.W. designed research; K.T., C.E.A., S.C., and V.H.O. performed research; D.H. and D.F.L. contributed new reagents/analytic tools; K.T., C.E.A., S.C., T.C.S., and M.W. analyzed data; and K.T., C.E.A., T.C.S., and M.W. wrote the paper. Reviewers: T.G., Center for Genomic Regulation; and H.O., Keio University School of Medicine. The authors declare no conflict of interest. Published under the PNAS license. Data deposition: The sequences reported in this paper have been deposited in the Gene Expression Omnibus (GEO) database, https://www.ncbi.nlm.nih.gov/geo (accession no. GSE113804). 1 K.T. and C.E.A. contributed equally to this work. 2 To whom correspondence may be addressed. Email: [email protected] or wernig@ stanford.edu. This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1720273115/-/DCSupplemental. Published online June 4, 2018. 64706475 | PNAS | June 19, 2018 | vol. 115 | no. 25 www.pnas.org/cgi/doi/10.1073/pnas.1720273115 Downloaded by guest on October 8, 2020

Transcript of Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood...

Page 1: Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted

Transdifferentiation of human adult peripheral bloodT cells into neuronsKoji Tanabea,b,1, Cheen Euong Anga,b,c,1, Soham Chandaa,b,d, Victor Hipolito Olmosa,b, Daniel Haaga,b,Douglas F. Levinsone, Thomas C. Südhofd,2, and Marius Werniga,b,2

aDepartment of Pathology, Stanford University, Stanford, CA 94305; bInstitute for Stem Cell Biology and Regenerative Medicine, Stanford University,Stanford, CA 94305; cDepartment of Bioengineering, Stanford University, Stanford, CA 94305; dDepartment of Molecular and Cellular Physiology andHoward Hughes Medical Institute, Stanford University, Stanford, CA 94305; and eDepartment of Psychiatry and Behavioral Sciences, Stanford University,Stanford, CA 94305

Contributed by Thomas C. Südhof, May 3, 2018 (sent for review November 21, 2017; reviewed by Thomas Graf and Hideyuki Okano)

Human cell models for disease based on induced pluripotent stem(iPS) cells have proven to be powerful new assets for investigatingdisease mechanisms. New insights have been obtained studyingsingle mutations using isogenic controls generated by genetargeting. Modeling complex, multigenetic traits using patient-derived iPS cells is much more challenging due to line-to-linevariability and technical limitations of scaling to dozens or morepatients. Induced neuronal (iN) cells reprogrammed directly fromdermal fibroblasts or urinary epithelia could be obtained frommany donors, but such donor cells are heterogeneous, showinterindividual variability, and must be extensively expanded,which can introduce random mutations. Moreover, derivation ofdermal fibroblasts requires invasive biopsies. Here we show thathuman adult peripheral blood mononuclear cells, as well asdefined purified T lymphocytes, can be directly converted intofully functional iN cells, demonstrating that terminally differen-tiated human cells can be efficiently transdifferentiated into adistantly related lineage. T cell-derived iN cells, generated by non-integrating gene delivery, showed stereotypical neuronal morphol-ogies and expressed multiple pan-neuronal markers, fired actionpotentials, and were able to form functional synapses. These cellswere stable in the absence of exogenous reprogramming factors.Small molecule addition and optimized culture systems haveyielded conversion efficiencies of up to 6.2%, resulting in thegeneration of >50,000 iN cells from 1 mL of peripheral blood in asingle step without the need for initial expansion. Thus, ourmethod allows the generation of sufficient neurons for experi-mental interrogation from a defined, homogeneous, and readilyaccessible donor cell population.

induced neuronal cells | direct conversion | transdifferentiation |disease modeling | iN cells

Advances in cell reprogramming and genome editing toolshave provided new ways to interrogate human gene function

in various human cellular contexts, such as neurons. In particu-lar, genetic engineering of embryonic or induced pluripotentstem (iPS) cells has proven powerful for dissecting the specificconsequences of disease-associated mutations in controlled ge-netic backgrounds (1, 2). However, these methods cannot beexpected to provide fully adequate cellular models of diseases forwhich highly polygenic mechanisms underlie risk. For example,large-scale genome-wide association study data suggest that 30–50% of the genetic risk for each of the neuropsychiatric disordersthat have been studied to date can be explained by the jointeffects of thousands of common genetic variants with small in-dividual effects, such that individual patients are likely to becarrying a unique combination of many contributory variants (3).One way to study such complex genetic backgrounds in human

neurons is by reprogramming patient cells to iPS cells (4).However, iPS cells have significant line-to-line variability interms of differentiation capacity, presumably due to variations intheir epigenetic and pluripotent state (5–7). Moreover, iPS cellsare often karyotypically unstable when grown in feeder-free

conditions, and their growth and formation is labor-intensiveand difficult to scale from a large number of individuals.Another way to obtain neurons is by deriving induced neuro-

nal (iN) cells from fibroblasts in a single conversion step, whichin principle would greatly facilitate their derivation from manypatients (8). However, unlike neonatal human fibroblasts, adulthuman fibroblasts have proven difficult to reprogram into syn-aptically competent iN cells (9–14). Moreover, fibroblasts areheterogeneous and ill-defined and must be expanded in vitrofrom invasive and painful skin biopsies to obtain sufficientnumbers, increasing the risk of acquiring random genetic muta-tions during an extended culture period. Here we report thatfunctional synapse-forming human iN cells can be induced fromfreshly isolated and stored adult peripheral T cells using non-integrating episomal vectors. Previous studies have shown theconversion of blood and urinary cells into various neural pro-genitor cells that only inefficiently gave rise to functional neu-rons (15–21). The described conversions were accomplished withtransient expression of iPS cell reprogramming factors, an

Significance

Recent advances in genomics have revealed that many poly-genetic diseases are caused by complex combinations of manycommon variants with individually small effects. Thus, buildinginformative disease models requires the interrogation of manypatient-derived genetic backgrounds in a disease-relevant celltype. Current approaches to obtaining human neurons are noteasy to scale to many patients. Here we describe a facile, one-step conversion of human adult peripheral blood T cells directlyinto functional neurons using episomal vectors without theneed for previous in vitro expansion. This approach is moreamenable than induced pluripotent stem cell-based approachesfor application to larger cohorts of individuals and will enablethe development of functional assays to study complex humanbrain diseases.

Author contributions: K.T., C.E.A., T.C.S., and M.W. designed research; K.T., C.E.A., S.C.,and V.H.O. performed research; D.H. and D.F.L. contributed new reagents/analytic tools;K.T., C.E.A., S.C., T.C.S., and M.W. analyzed data; and K.T., C.E.A., T.C.S., and M.W. wrotethe paper.

Reviewers: T.G., Center for Genomic Regulation; and H.O., Keio University Schoolof Medicine.

The authors declare no conflict of interest.

Published under the PNAS license.

Data deposition: The sequences reported in this paper have been deposited in the GeneExpression Omnibus (GEO) database, https://www.ncbi.nlm.nih.gov/geo (accession no.GSE113804).1K.T. and C.E.A. contributed equally to this work.2To whom correspondence may be addressed. Email: [email protected] or [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1720273115/-/DCSupplemental.

Published online June 4, 2018.

6470–6475 | PNAS | June 19, 2018 | vol. 115 | no. 25 www.pnas.org/cgi/doi/10.1073/pnas.1720273115

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

Page 2: Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted

approach recently shown to induce a pluripotent intermediatestate (22).

ResultsDirect Induction of iN Cells from Peripheral Blood Mononuclear Cells.To investigate whether blood cells can be transdifferentiated toiN cells, we collected fresh blood from an adult healthy indi-vidual. We then isolated peripheral blood mononuclear cells(PBMCs) using gradient centrifugation and electroporated thesecells with episomal vectors encoding the four transcription fac-tors Brn2, Ascl1, Myt1l, and Ngn2, collectively termed theBAMN pool, which was previously found to generate iN cellsfrom human fibroblasts (9), and enhanced green fluorescentprotein (EGFP) into 3 million PBMCs. Transfected cells werethen cultured in IL-2 and CD3/CD28 antibodies containingmedia supporting T cell growth (Fig. 1A). On day 3, we placedthe nonadherent, transfected blood cells on different substrates,including primary mouse glia, mouse fibroblast SNL cell line,human primary fibroblasts, Matrigel, Polyornithine, or laminin-coated dishes. BAMN-transfected cells attached well on primary

mouse glia cells and human fibroblasts, but on no other sub-strate. Cells transfected with EGFP alone did not attach to anysubstrate (SI Appendix, Fig. S1 B and C). On day 5, we changedN3 medium

Electroporation

(BAMN+GFP)

media change

media change

media change

D0 D5 D7D3 D14 D21media change

Seeding on Glia

(IL2, activator)

blood cell medium+IL2+Activ. +IL2

Day21 Day42

Day1 Day14

0 20 40 60 80

Indu

ctio

n ef

ficie

ncy

(%)

Age

0

0.4

0.8

0

2.5

5

-IL2

-Act

iv.

N3 R

elat

ive

Num

ber o

f iN

cel

ls

+IL2

+A

ctiv

.

+IL2

-Act

iv.

Day 21

Tran

sdiff

eren

tiatio

n e

ffici

ency

(%)

0

0.5

1.5

1

2.0

0

0.5

1.5

1

Rel

ativ

e ef

ficie

ncy

ofel

ectro

pora

tion

Day 21

*

Rel

ativ

e nu

mbe

r of

iN c

ells

Day 21

*

Fres

h

4 Co

-80

Co

0

0.5

1.5

1

A

B

D

C

E F G

Fres

h

4 Co

-80

Co

Fres

h

4 Co

-80

Co

Fig. 1. Generation of neuronal cells from peripheral blood cells. (A) Experi-mental outline of iN cell induction from PBMCs. (B) Morphological changesduring iN cell induction from PBMCs. (Scale bars: 50 μm.) (C) The relativenumber of iN cells from T cells with or without T cell activator (anti-CD3/CD28),with or without IL-2, or a change to N3 media on day 3. n = 3 individuals. (D)Efficiency of iN cell induction of transduced cells from 35 individual donorswithout inhibitors at day 21. n = 1 for each donor. The number of iN cells onday 21was divided by the number of total EGFP+ cells counted on day 1. (E andF) Relative iN cell induction (E) and efficiency of electroporation (F) fromPBMCs of three individual donors that were kept at −80 °C or at 4 °C for 2 drelative to the fresh sample. *P < 0.05, paired t test. (G) Transdifferentiationefficiency of PBMCs from three individual donors kept at −80 °C or at 4 °C for2 d relative to the fresh sample. Error bars represent SD.

3sm

0

5

15

Con

tFo SB

Do

* *

*

*

Cont-TUJ1 Cont-MAP2

3sm-TUJ1 3sm-MAP2

(n = 5/8)(n = 5/9)

(n = 5/6)(n = 3/7) 20mV10mS

3sm

10

Fold

-incr

ease

indu

ctio

n ef

ficie

ncy

25ms0.5nA 1ms

0.5nA

Na+ and K+ channel currents (N3+3sm condition, Day 42)1.5

10.5

0-0.5

-80 -40 0 40

Na+

K+

12

8

4

0

2

1

0

-50

-40

-30

-202142

** 70

60

50

-30

-20

-10

** *

ns

Intrinsic membrane properties(N3 + 3sm condition)

Action potential properties(N3 + 3sm condition)

Con

t

Day 21 Day 42

Days Days Days Days Days

Voltage (mV)

Cur

rent

(nA

)

Vre

st (m

V)

Rm (G

Ohm

)

Cm (p

F)

AP

hei

ght (

mV

)

AP

thre

shol

d (m

V)

2142 21 42 21 42 21 42

A

B

D

E F

C

Fig. 2. Small molecule treatment improves iN cell conversion efficiency andmaturation. (A) Immunofluorescence analysis of iN cells with and withoutsmall molecules (3sm, three small molecules: forskolin, dorsomorphin, andSB431542; Cont, DMSO). (Scale bars: 50 μm.) (B) Fold change of improved iNcell formation on day 21 following various small molecule treatments as in-dicated. Do, dorsomorphin; Fo, forskolin; SB, SB431542. Data shown are av-erage fold changes of three independent experiments using PBMCs fromthree different donors. The fold change over the control condition was plottedbecause the absolute reprogramming efficiency was variable among the three do-nors, but the fold change was consistent. *P < 0.05, paired t test. The error barsindicate SDs. Similar results were obtained with PBMCs from another set of threedifferent donors. (C) Example traces of action potential firing recorded from PBMC-derived iN cells with or without 3sm at days 21 and 42 (n represents the number ofcells patched that shows action potentials over the total number of cells patched).The experiment was performed with cells from three different donors, yieldingsimilar results. (D) Sample traces (Left) and average values (mean ± SEM; Right)demonstrating the presence of voltage-gated Na+ and K+ channels in blood iN cellscocultured with glia with 3sm for 42 d. (Inset) Expanded view of the dotted boxedarea. (E) Intrinsic properties of membrane potential (Vrest), capacitance (Cm), andinput resistance (Rm) approach more mature values over time. (F) Maturation ofblood iN cells on extended coculture with glia in 3sm from day 21 to day 42 asdetermined by increased action potential (AP) height and threshold. Bar graphsrepresent mean ± SEM. *P < 0.05; **P < 0.01. ns, not significant.

Tanabe et al. PNAS | June 19, 2018 | vol. 115 | no. 25 | 6471

NEU

ROSC

IENCE

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

Page 3: Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted

the hematopoietic medium to the neuronal medium N3. Re-markably, after seeding of the human blood cells on murine glialcells, we noticed that glial cells deteriorated quickly. We reasonedthat the nontransfected, activated human T cells presumably be-gan to attack the mouse glia. Withdrawal of IL-2 and the T cellactivators after day 3 not only mitigated this problem, but alsoimproved reprogramming by 2.7- to 3.6-fold. Switching to neuro-nal media on day 3 also rescued glial viability but improvedreprogramming only slightly (Fig. 1C and SI Appendix, Fig. S1E).To test the general applicability of our protocol, we obtained

blood from a total of 35 healthy adult donors of various ages andethnicities and both sexes. Surprisingly, the electroporationefficiencies varied significantly (SI Appendix, Fig. S1A), butnonetheless we were able to generate morphologically complexiN cells from all tested blood samples (Fig. 1D). The reprog-ramming efficiency varied as well, but the electroporation ratedid not correlate with the reprogramming efficiency (SI Appen-dix, Fig. S1D). Unlike iPS cell reprogramming, reprogrammingto iN cells was consistently lower in aged donors (Fig. 1D).Because most biorepositories freeze patient samples, we next

tried to induce iN cells from short- and long-term stored PBMCsat cold temperature. We isolated PBMCs from a fresh whole-blood sample, reprogrammed a fraction of the cells, and frozethe remaining cells at −80 °C. Another fraction of the whole-blood sample was maintained at 4 °C for 2 d before subsequentPBMC isolation. We then reprogrammed the PBMCs isolatedfrom the stored blood sample and the frozen PBMCs. Therewere no significant differences in the reprogramming efficiency andelectroporation efficiency between fresh and frozen PBMCs, but theiN cell yield was substantially lower from PBMCs stored at 4 °C dueto decreased transfection efficiency (Fig. 1 E–G). Thus, storage offreshly isolated cells at −80 °C did not affect reprogramming.

Combined BMP and TGF-β Pathway Inhibition and PKA ActivationImproved Reprogramming Efficiency. We next sought to furtherincrease the induction efficiency of iN cells using small mole-cules. Blockade of BMP and TGF-β pathways have been shownto promote neural induction during normal development, duringES cell differentiation, and from fibroblasts (23–25). Moreover,cAMP has been reported to facilitate neuronal survival andmaturation (26). Therefore, we treated reprogramming bloodcells with compounds regulating these three pathways from day5 to day 21 (Fig. 2A). Indeed, the number of iN cells was sig-nificantly increased by the adenylyl cyclase activator forskolin(1.9-fold), the BMP pathway blocker dorsomorphin (3.7-fold),and the TGF-β pathway inhibitor SB431542 (4.6-fold) whenanalyzed on day 21 (Fig. 2B). The combination of the three in-hibitors (forskolin, dorsomorphin, and SB431542; 3sm) showedan additive effect on the number of iN cells relative to the cellsseeded (8.7-fold) (Fig. 2 A and B). These combined improve-ments yielded a reprogramming efficiency of up to 6.2%, whichtranslates into 54,265 iN cells from 1 mL of blood (Fig. 2B).

Blood iN Cells Exhibit Passive and Active Neuronal MembraneProperties. We then tested the functional properties of bloodiN cells that were generated with and without addition of the threeinhibitors. Patch-clamp recordings of blood iN cells showed theirability to fire action potentials on step-current injection at 21 and42 d after infection and expressed functional voltage-gated Na+and K+ channels (Fig. 2 C and D). As expected, 3sm treatmentand longer culture periods (day 42) yielded cells with parametersof more mature action potentials (height, threshold, faster de-polarization and repolarization kinetics) and more mature in-trinsic properties, such as increased capacitance and decreasedinput resistance, compared with control-treated cells and day21 cells, respectively (Fig. 2 C, E, and F).

ROCK Inhibition Substantially Improves Formation of NeuronalMorphologies but Does Not Improve Functional Properties. In an at-tempt to increase the reprogramming efficiencies even further, wescreened six additional compounds—IWP2, DAPT, retinoic

E

B

G

C

A

Day 21

(n = 3/6)(n = 2/6)

(n = 5/6)(n = 3/7)

20mV10mS

F

D

Cont 3sm 3sm+Rock i

Rel

ativ

e le

ngth

neu

rites

*

*

0

DM

SO

IWP

2

SU

RA

Roc

k i

DA

PT

SP

Con

t3sm

5

10

EGFP

MAP2

3sm + Rock i

Day 21

0

10

20

TUJ1 (+)MAP2 (+)30

Rel

ativ

e nu

mbe

r of

iN c

ells

DM

SO

IWP

2

SU

RA

Roc

k i

DA

PT

SP

Con

t

3sm

*

*

*

0

10

30

3smRock iIWP2

RA

----

++++

+---

++-+

+++-

++--

Rel

ativ

e N

umbe

r o

f iN

cel

ls 20

*

**

Control

3sm+Rock inhibition

H

0 100 200 300Total no of neurons

*

*

0 3 2114

DMSO

3sm+Rocki3sm+Rocki

3sm

6 12Relative Number of iN cells

DMSO3sm+Rock i

0 5 02114

w/ BDNF and GDNF w/o BDNF and GDNF

DMSO

D14

D21 3sm+Rock i3sm+Rock i

3sm

Day 42

Day 42Day 21

Fig. 3. ROCK inhibition improves morphological maturation, but notfunctional maturation. (A) Relative number of TUJ1-positive (yellow bars) orMAP2-positive (blue bars) cells with 3sm (forskolin, dorsomorphin, andSB431542) and additional small molecules from PBMCs from three individualdonors on day 21, normalized to the no treatment control condition (Cont).n = 3 individuals. *P < 0.05 relative to DMSO (only with 3sm). (B) Relativeaverage length of neurites with small molecules from PBMCs from threeindividual donors. The length was normalized to the no treatment control(Cont). n = 3 individuals. *P < 0.05 relative to the DMSO. (C) Example pic-tures of iN cells with indicated inhibitors. Green, EGFP fluorescence (Scalebars: 50 μm.). (D) iN cells generated with 3sm plus ROCK inhibitor expressneuronal markers, including MAP2. (Scale bars: 50 μm.) (E) The effectof indicated small molecule combinations on reprogramming efficiency.*P <0.05, paired t test. n = 3 individuals. (F) The effect of the duration of 3smplus ROCK inhibitor with (yellow bars) or without (blue bars) GDNF andBDNF on reprogramming efficiency. n = 3 individuals. (G) Representativetraces of action potential responses of PBMC-derived iN cells under controlcondition (N3 only; Top) and in the presence of 3sm plus ROCK inhibitorwhen cocultured with glia for 21 d (Left) or 42 d (Right). (H) Graph showingthe total number of neurons in four conditions: DMSO control (black line), 3smtreatment for 21 d (blue line), 3sm plus ROCK inhibitor for 21 d (red line), and3sm plus ROCK for 14 d and 3sm for 7 d. n = 3 individuals. *P < 0.05.

6472 | www.pnas.org/cgi/doi/10.1073/pnas.1720273115 Tanabe et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

Page 4: Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted

acid, SU5402, Y26732, and SP600125—targeting pathways pre-viously implicated in neural differentiation in combination withforskolin, dorsomorphin, and SB431542. Of the single com-pounds tested, we found that ROCK inhibition increased boththe relative number of iN cells and the morphological com-plexity of reprogramming cells compared with the DMSOcontrol (Fig. 3 A–D). Other combinations of small mole-cules did not have additional effects (Fig. 3E). Moreover, theaddition of neurotrophic factors did not substantially increasethe formation of iN cells, but long-term drug treatmentyielded more iN cells than transient small molecule treatment(Fig. 3F).Importantly, however, when we tested their functional prop-

erties, cells generated in the presence of ROCK inhibition werenot able to generate mature action potentials, unlike cells grownwithout the ROCK inhibitor (Fig. 3G). The lack of mature actionpotentials suggests that ROCK inhibition simply affects cyto-skeletal rearrangements but perturbs functional neuronal matu-ration. To assess whether transient ROCK inhibition is sufficientto increase conversion efficiency and still allow for functionalmaturation, we removed the ROCK inhibitor and 2 wk after 3smplus ROCK inhibitor treatment and found no beneficial effect(Fig. 3H).

Molecular Characterization of Blood-Derived iN Cells. To assess thetranscriptional changes induced by reprogramming, we per-formed RNA-sequencing (RNA-seq) on the donor PBMCs aswell as on EGFP+ and PSA-NCAM+ blood iN cells purified bymagnetic cell sorting. A total of 6,941 genes were differentiallyexpressed between the PBMCs and blood iN cells (Fig. 4A). Theup-regulated genes were enriched for Gene Ontology terms suchas nervous system development and synaptic transmission, whilethe down-regulated genes were enriched for cellular defense

responses (Fig. 4A). Pan-neuronal markers (TAU, TUBB3,MAP2, and NCAM) were up-regulated while blood surfacemarkers (CD8, CD45, and CD3) were silenced (Fig. 4 B and C).As expected, proproliferative genes were down-regulated andnegative cell cycle regulators were induced (Fig. 4 D and E).These results suggest that blood iN cells have silenced hemato-poietic transcriptional programs and have adopted a pure neuronalidentity.To characterize the regional and neurotransmitter identity of

PBMC-derived iN cells, we performed immunofluorescenceand considered the results in conjunction with the RNA-seqdata. The two methods yielded very consistent results. Wefound that blood iN cells expressed the vesicular glutamatetransporter (VGLUT) but not the vesicular GABA transporter,GAD65, or tyrosine hydroxylase, suggesting that blood iN cellsare excitatory neurons similar to fibroblast iN cells and Ngn2-ES iN cells (8, 9, 27) (Fig. 4 F, G, and I). Based on the RNA-seqresults, we found expression of forebrain markers and corticallayers II–V but not layers I, IV, and VI (Fig. 4 H and J). Im-munofluorescence analysis demonstrated that in fact almost allblood iN cells were positive for SATB2 and CTIP2 but negativefor REELIN, as suggested by the RNA results (Fig. 4 F and G).SATB2/CTIP2 double-positive cells are found in layer V of themouse cortex (28).

The Blood-to-Neuron Conversion Does Not Involve a ProliferativeNeural Progenitor State. Several groups have reported the con-version of blood cells into proliferative neural cells using subsetsof iPS cell reprogramming factors (20, 21). In contrast, ourreprogramming factors are not proproliferative and our ap-proach yields postmitotic neurons directly. Nevertheless, it maybe possible that even our reprogramming process involves aproliferative intermediate progenitor. To address this question,

Top significant GO termsU

preg

ulat

ed

Dow

nreg

ulat

edPBMC Blood-iN

Neural markers

0

50

MIKI67

MCM2CDK1

100

RP

KM

PBMCiN

0

50

P57IN

K4AIN

K4BIN

K4CIN

K4D

100

RP

KM

PBMCiN

0

500

250

PBMCiN

TAUTUJ1

MAP2

NCAM

RP

KM

0

500

250

PBMCiN

CD8CD45CD3

Blood markers

RP

KM

VGLUT

CTIP2

SATB2

EGFP

EGFP

EGFP

Immunofluorescence

vGLU

T

SATB2CTIP

2

REELIN THvG

ATMAP2

TUJ1

GAD650

100

50

% in

Neu

rona

l cel

ls

PRPH

150

0

4

2

RNA sequencing

0

40

20

RELN

SATB2CUX1

RORB

FEZF2CTIP

2TBR1

SOX5

RP

KM

VGLUT2

GAD67 THCHAT

DBH

RP

KM

RNA sequencing

Synaptic transmissionChromatin organizationMAPK cascadeNervous system developmentCellular component morphogenesisIntracellular protein transportSensory perception of smell

DNA recombinationTranslationRegulation of cell cycleCell proliferationCellular defense responseMitosis Defense response to bacterium

05

10152025303540

RP

KM

FOXG1OTX1

SIX3

EN1EN2

HOXB4

HOXC4

RNA sequencing

A

D

F G

B C

I J

E

H

3.002.001.000.00-1.00-2.00-3.00

PBMCiN

PBMCiN

Fig. 4. Blood iN cells express genes characteristic ofexcitatory, postmitotic neurons. (A) Heatmap show-ing 6,941 genes differentially expressed between PBMCand blood iN cells. Two biological replicates per pop-ulation, greater than twofold change and P < 0.05.Shown are the seven most significant (P < 0.05, Bon-ferroni-corrected) Gene Ontology terms among up- anddown-regulated genes using PANTHER. (B) Induction ofpan-neuronal markers. (C) Suppression of blood cellgenes. (D) Down-regulation of cell cycle activators. (E)Induction of antiproliferative cyclin-dependent kinaseinhibitors. (F) Immunofluorescence of blood iN cellsshowing expression of the excitatory marker vGLUTand subtype markers SATB2 and CTIP2 at 21 d afterinfection and cultured with 3sm on glia (Scale bars:50 μm.). (G) Quantification of day 21 blood iN cellsgrown on glia in 3sm conditions by immunofluores-cence for indicated markers. (H) Expression of region-specific markers by the PBMC-derived iN cells by RNA-seq. (I) Validation of the neurotransmitter-specificmarkers by RNA-seq. (J) Validation of the cortical sub-type-specific markers by RNA-seq.

Tanabe et al. PNAS | June 19, 2018 | vol. 115 | no. 25 | 6473

NEU

ROSC

IENCE

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

Page 5: Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted

we decided to stain the cells at different time points betweendays 3 and 21 during reprogramming with the neural progenitormarker Sox1 and the proliferation marker Ki67 (SI Appendix,Fig. S2). To our surprise, Ki67-positive cells decreased only afterthe first week of reprogramming. Nonetheless, all Ki67-positivecells were Sox1-negative, and the number of proliferative cellsdeclined rapidly after day 7. Thus, no proper neural progenitorcells are formed as transient intermediates. The initial persis-tence of proliferative cells is likely due to the cytokines that weused to activate lymphocytes.

Blood iN Cells Are Stable Without Persistent Transgene Expression.To examine whether the neuronal identity is dependent oncontinued transgene expression, we first examined whetherthe transgenes were perhaps already silenced in our original

transfection protocol (Fig. 1A). We had used a bicistronic de-livery of Ngn2 and Ascl1 linked by the T2A self-cleaving peptide.After cleavage, the T2A peptide sequence is predicted to befused in frame with Ascl1 and thus can serve as a tag of theexogenous Ascl1. Immunostaining with T2A antibodies showedthat as many as 40% of blood iN cells were T2A-Ascl1 negative,and FACS analysis demonstrated that approximately the samefraction of cells had silenced the EGFP transgene, which was co-electroporated together with the reprogramming factors (SIAppendix, Fig. S3 A and B). Because we used the mouse cDNAsfor all four reprogramming factors to reprogram human cells, wecould accurately distinguish the exogenous (mouse) factors fromthe endogenous (human) ASCL1, NGN2, BRN2, and MYT1Lgenes in our RNA-seq dataset. This analysis clearly demon-strates that the exogenous factors were effectively silenced inthe EGFP−/PSA-NCAM+ iN cell population (SI Appendix,Fig. S3C). Thus, blood iN cells can adopt a stable neuronalidentity without the need for continued expression of exogenousreprogramming factors.

Synaptically Competent iN Cells Can Be Derived from CD3+ T Cells.PBMCs consist of fairly heterogenous hematopoietic cell pop-ulations. We wondered whether iN cells could be establishedfrom a more defined cell type. After characterizing freshlytransfected cells, we found that our electroporation conditionsgreatly favor CD3+ T cells, suggesting that the vast majority ofPBMC iN cells are in fact T cell-derived (SI Appendix, Fig. S4 Aand B). To more specifically test whether T cells can be con-verted, we introduced the BAMN factors into four defined cellpopulations purified based on CD3 and CD4 expression. Indeed,morphologically complex iN cells were induced from CD3+/CD4−and CD3+/CD4+ cells, but not from CD3− cells (Fig. 5A). T cell iNcells showed passive and active neuronal membrane propertieswhen cocultured with glia for 18 d and recorded on day 21 (Fig. 5B and C and SI Appendix, Fig. S4E). Finally, we confirmed theT cell origin of PSA-NCAM FACS-purified iN cells, by dem-onstrating genomic VDJ rearrangements at TCR-β locus (SIAppendix, Fig. S4 C and D).The defining property of a neuron is its ability to make syn-

aptic connections. Therefore, we asked whether BAMN-inducedT cell iN cells would be able to form functional synapses. Localadministration of GABA and AMPA on the soma and proximaldendrites of day 21 T cell iN cells recorded in voltage-clampmode yielded prominent GABAA receptor-mediated inhibitorypostsynaptic currents (PSCs) and AMPA receptor-mediated ex-citatory PSCs, respectively (SI Appendix, Fig. S4 G and H),demonstrating the presence of functional neurotransmitter re-ceptors. To address whether the T cell iN cells have the capacityto form synapses and functionally integrate into existing neuro-nal networks, we plated EGFP-labeled iN cells at 14 d post-induction onto unlabeled human iPS cell-derived neurons. At21 d after coculture, we performed voltage-clamp recordingsfrom EGFP-positive cells and observed synaptic AMPA receptor-mediated spontaneous network activities, indicating their success-ful integration into the synaptic network (Fig. 5D). In addition,spontaneous PSCs could also be observed in these cells, as con-firmed by application of the specific AMPA and GABAA receptorantagonists CNQX and picrotoxin, respectively (Fig. 5D). Finally,evoked PSCs could also be elicited by activating surrounding axonsvia extracellular field stimulation of the vicinity (Fig. 5E), unam-biguously demonstrating that blood iN cells can receive functionalsynaptic inputs from other neurons.

DiscussionOur demonstration that adult human peripheral T cells can bedirectly converted to neurons has both conceptual and practicalimplications. During normal development, the only cells with thepotential to change lineage identity are uncommitted stem andprogenitor cells. Most reprogramming studies use heterogeneousfibroblasts as donor cells, raising the question as to whether thetransdifferentiation capability is limited to undifferentiated progenitor

Rel

ativ

e re

prog

ram

min

g ef

ficie

ncy

0

1.5

CD

3 /C

D4

+-

CD

3 /C

D4

++

CD

3 /C

D4

-+

CD

3 /C

D4

--

1.0

20ms20pA

2 s20 pA

50 ms20 pA

Spontaneous network activity

20 pA

0.5 s20 pA

Control

CNQX+Picrotoxin

Spontaneous PSCs

Rec

Bright-field

Merge

A

(n = 5/8)

(n = 6/6)

CD3 /CD4+ -

CD3 /CD4+ +

20mV20mS

Day 21B C

D E Evoked PSCsControl

CNQX+Picrotoxin

Fig. 5. Generation of synaptically competent iN cells from T cells. (A) Rel-ative reprogramming efficiency of iN cells from the four indicated PBMCpopulations. The plotted efficiency was normalized by the electroporationefficiency and the efficiency of the CD3+/CD4− cell population was set to 1(n = 3 individuals). (B) Recording configuration of EGFP-labeled blood iN cellscocultured with unlabeled iPS cell-derived neurons. The recording electrode(Rec) was placed onto an EGFP-positive blood iN cell (white arrowhead)surrounded by nonfluorescent human iPS cell-derived neurons (black ar-rowheads). (Scale bar: 50 μm.) (C) Example traces of action potential firingrecorded from iN cells derived from CD3+/CD4− (red) or CD3+/CD4+ (blue)T cells. N represents the number of cells patched that show action potentialsover the total number of cells patched. (D) Representative traces of AMPAreceptor-mediated spontaneous network activity (Top, black) recorded froma T cell-derived iN cell, indicating successful integration into the humansynaptic network. The trace in red (Bottom) represents an expanded view ofthe boxed area. Spontaneous PSCs recorded from a T cell-derived iN cell(Top, black) and subsequently blocked by CNQX and picrotoxin (Bottom,black). The trace in red (Middle) represents and expanded view of the boxedarea. This pattern was observed in 2 of the 27 cells patched. (E) Evoked PSCs(five trials) in response to extracellular field stimulation recorded from ablood iN cell (Top), which was subsequently blocked by CNQX and picrotoxinapplication (Bottom).

6474 | www.pnas.org/cgi/doi/10.1073/pnas.1720273115 Tanabe et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

Page 6: Transdifferentiation of human adult peripheral blood T ... · human adult peripheral blood mononuclear cells, as well as defined purified T lymphocytes, can be directly converted

cells (29). Our results unequivocally show that terminally dif-ferentiated cells can be transdifferentiated into another, distantlyrelated somatic lineage.The derivation of neurons from adult peripheral blood cells

also has important practical implications. Unlike fibroblasts,whose derivation requires an invasive and painful skin punchbiopsy, lymphocytes can be obtained in large numbers from asimple venipuncture, a procedure performed in almost everyhospitalized patient, often on a daily basis. Moreover, bloodsamples are stored in biorepositories in much larger numbersthan skin fibroblasts. Of relevance for blood iN cell applicationsusing such repositories, we observed that iN cells can be obtainedfrom fresh and frozen blood cells with similar efficiency.Therefore, our blood iN cell conversion described here enablesthe generation of human neurons from virtually any individual,unlike the use of fibroblasts as donor cells, which have provendifficult to obtain from certain populations, such as children andmentally ill persons. In addition, the greater accessibility allowsfor scalability of donor individuals, which will be instrumental inassessing how common, low-risk–conferring genetic variantscontribute to cellular function in complex genetic diseases. An-other advantage over fibroblasts as donor cells is that fibroblastsneed to be expanded in vitro to obtain sufficient numbers, whichmay lead to accumulation of deleterious mutations.From a mechanistic standpoint, it was unexpected to find that—

unlike iN cell transdifferentiation from fibroblasts—the earlycoculture of glia was critical for transdifferentiation of bloodcells. The effect of glia seems to be fundamentally different herethan in fibroblast reprogramming, where glial coculture does notsubstantially impact conversion efficiency rather than synapticmaturation (8, 9). In contrast, the role of glial factors affectedthe generation of iN cells in general when transdifferentiatedfrom blood cells. Since transfected PBMCs also attached onto a layerof fibroblasts but did not reprogram, we assume that the monolayerof glial cells provides secreted or cell contact-dependent factorsto the blood cells that are essential for transdifferentiation inaddition to enabling their attachment.While this paper provides a clear proof of concept that human

adult peripheral T cells can be converted to iN cells with all keybiochemical and functional properties of neurons, we note that—similar to human fibroblast iN cells—these cells exhibit less maturesynaptic properties compared with primary mouse or iPS cell-

derived iN cells (20, 21, 30, 31), and using the current protocol,our blood iN cells are exclusively excitatory. While many cellbiological processes, such as transcription, polarization, migra-tion, and subcellular transport, can already be studied in thesecells, future efforts will need to focus on improving synapticmaturation and deriving additional neuronal subtypes.

Materials and MethodsPBMCs were isolated from fresh blood donations obtained through theStanford Blood Bank from individuals of various ethnic backgrounds (Cau-casian, Japanese, Indian, South American, and African), various ages (16–78 y), and both sexes using density gradient centrifugation with Ficoll-PaquePLUS (GE Healthcare) according to the manufacturer’s instructions. PBMCswere frozen by a stem cell banker (ZENOAQ). Then 3 μg of vectors (PcxLE-Ngn2-2A-Ascl1, Brn2, and Myt1L: 0.666 μg; pCXLE-GFP: 0.5 μg; pCXWB-Ebna1: 0.5 μg) or nonreplicative vectors (pCXWB-Brn2, Ascl1, v5Myt1l,flagNgn2 and GFP: 0.5 μg) were electroporated into 3 × 106 isolated PBMCswith the Nucleofector 2b Device (Lonza) with the Amaxa Human T-CellNucleofector Kit, program V-024 (Lonza).

Transduced cells were cultured for 3 d in six-well plates in X-VIVO10medium (Lonza) supplemented with 30 U/mL IL-2 (PeproTech) and 3.4 μL/mLDynabeads Human T-Activator CD3/CD28 (Life Technologies). At 3 d afterelectroporation, 0.1–1 × 106 transduced cells were seeded on primary gliaculture in a well of a 12-well plate. Glia (1.5 × 105) were seeded in a well of a12-well plate coated with Matrigel (Corning). At 2 d after seeding, the mediumwas replaced with DMEM/F12 (Invitrogen) including N2 supplement (Gibco),B27 supplement (Gibco), and insulin (5 μg/mL; Sigma Aldrich). The mediumwas changed every 7 d. The RNA-seq files are available in the NationalCenter for Biotechnology Information’s Gene Expression Omnibus database(accession no. GSE113804). Virus generation, electrophysiology, RNA-sequencing,TCR recombination and generation of human embryonic stem cell-derivedneurons are described in detail in SI Appendix.

ACKNOWLEDGMENTS. We thank Dr. Keisuke Okita, Dr. Kazutoshi Takahashi,and members of the M.W. laboratory for important suggestions. This projectwas supported by National Institutes of Health Grants R01 MH092931 andU19 MH104172, the New York Stem Cell Foundation (NYSCF)–RobertsonPrize, and the Stanford Schizophrenia Genetics Research Fund establishedby an anonymous donor. C.E.A. was supported by California Institute ofRegenerative Medicine Training Grant and the Siebel Foundation. M.W. isa NYSCF–Robertson Stem Cell Investigator, a Howard Hughes Medical Insti-tute Faculty Scholar, and a Tashia and John Morgridge Faculty Scholar. T.C.S.is a Howard Hughes Medical Institute Investigator.

1. Turner TN, et al. (2017) Genomic patterns of de novo mutation in simplex autism. Cell171:710–722.e12.

2. Ang CE, Wernig M (2014) Induced neuronal reprogramming. J Comp Neurol 522:2877–2886.

3. Treutlein B, et al. (2016) Dissecting direct reprogramming from fibroblast to neuronusing single-cell RNA-seq. Nature 534:391–395.

4. Takahashi K, et al. (2007) Induction of pluripotent stem cells from adult human fi-broblasts by defined factors. Cell 131:861–872.

5. Nichols J, Smith A (2009) Naive and primed pluripotent states. Cell Stem Cell 4:487–492.

6. Theunissen TW, et al. (2014) Systematic identification of culture conditions for in-duction and maintenance of naive human pluripotency. Cell Stem Cell 15:471–487.

7. Shen Y, et al. (2012) A map of the cis-regulatory sequences in the mouse genome.Nature 488:116–120.

8. Vierbuchen T, et al. (2010) Direct conversion of fibroblasts to functional neurons bydefined factors. Nature 463:1035–1041.

9. Pang ZP, et al. (2011) Induction of human neuronal cells by defined transcriptionfactors. Nature 476:220–223.

10. Yoo AS, et al. (2011) MicroRNA-mediated conversion of human fibroblasts to neu-rons. Nature 476:228–231.

11. Ambasudhan R, et al. (2011) Direct reprogramming of adult human fibroblasts tofunctional neurons under defined conditions. Cell Stem Cell 9:113–118.

12. Pfisterer U, et al. (2011) Direct conversion of human fibroblasts to dopaminergicneurons. Proc Natl Acad Sci USA 108:10343–10348.

13. Son EY, et al. (2011) Conversion of mouse and human fibroblasts into functionalspinal motor neurons. Cell Stem Cell 9:205–218.

14. Caiazzo M, et al. (2011) Direct generation of functional dopaminergic neurons frommouse and human fibroblasts. Nature 476:224–227.

15. Lee JH, et al. (2015) Single transcription factor conversion of human blood fate toNPCs with CNS and PNS developmental capacity. Cell Rep 11:1367–1376.

16. Giorgetti A, et al. (2012) Cord blood-derived neuronal cells by ectopic expression ofSox2 and c-Myc. Proc Natl Acad Sci USA 109:12556–12561.

17. Castano J, et al. (2014) Fast and efficient neural conversion of human hematopoieticcells. Stem Cell Reports 3:1118–1131.

18. Tang X, et al. (2016) Conversion of adult human peripheral blood mononuclear cellsinto induced neural stem cell by using episomal vectors. Stem Cell Res 16:236–242.

19. Wang L, et al. (2013) Generation of integration-free neural progenitor cells from cellsin human urine. Nat Methods 10:84–89.

20. Matsui T, et al. (2012) Neural stem cells directly differentiated from partially re-programmed fibroblasts rapidly acquire gliogenic competency. Stem Cells 30:1109–1119.

21. Kim J, et al. (2011) Direct reprogramming of mouse fibroblasts to neural progenitors.Proc Natl Acad Sci USA 108:7838–7843.

22. Maza I, et al. (2015) Transient acquisition of pluripotency during somatic cell trans-differentiation with iPSC reprogramming factors. Nat Biotechnol 33:769–774.

23. Hemmati-Brivanlou A, Melton DA (1994) Inhibition of activin receptor signalingpromotes neuralization in Xenopus. Cell 77:273–281.

24. Chambers SM, et al. (2009) Highly efficient neural conversion of human ES and iPScells by dual inhibition of SMAD signaling. Nat Biotechnol 27:275–280.

25. Ladewig J, et al. (2012) Small molecules enable highly efficient neuronal conversionof human fibroblasts. Nat Methods 9:575–578.

26. Lepski G, Jannes CE, Nikkhah G, Bischofberger J (2013) cAMP promotes the differ-entiation of neural progenitor cells in vitro via modulation of voltage-gated calciumchannels. Front Cell Neurosci 7:155.

27. Chanda S, et al. (2014) Generation of induced neuronal cells by the single re-programming factor ASCL1. Stem Cell Reports 3:282–296.

28. Alcamo EA, et al. (2008) Satb2 regulates callosal projection neuron identity in thedeveloping cerebral cortex. Neuron 57:364–377.

29. Tanabe K, Haag D, Wernig M (2015) Direct somatic lineage conversion. Philos Trans RSoc Lond B Biol Sci 370:20140368.

30. Maximov A, Pang ZP, Tervo DG, Südhof TC (2007) Monitoring synaptic transmission inprimary neuronal cultures using local extracellular stimulation. J Neurosci Methods161:75–87.

31. Zhang Y, et al. (2013) Rapid single-step induction of functional neurons from humanpluripotent stem cells. Neuron 78:785–798.

Tanabe et al. PNAS | June 19, 2018 | vol. 115 | no. 25 | 6475

NEU

ROSC

IENCE

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020