Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into Neurons, Skeletal Myocytes, and Oligodendrocytes

Abstract

The isolation or in vitro derivation of many human cell types remains challenging and inefficient. Direct conversion of human pluripotent stem cells (hPSCs) by forced expression of transcription factors provides a potential alternative. However, deficient inducible gene expression in hPSCs has compromised efficiencies of forward programming approaches. We have systematically optimized inducible gene expression in hPSCs using a dual genomic safe harbor gene-targeting strategy. This approach provides a powerful platform for the generation of human cell types by forward programming. We report robust and deterministic reprogramming of hPSCs into neurons and functional skeletal myocytes. Finally, we present a forward programming strategy for rapid and highly efficient generation of human oligodendrocytes.

Full text

Stem Cell Reports
Report
Inducible and Deterministic Forward Programming of Human Pluripotent
Stem Cells into Neurons, Skeletal Myocytes, and Oligodendrocytes
Matthias Pawlowski,
1,2,6,7,
*Daniel Ortmann,
1,3,6
Alessandro Bertero,
1,3,6,8
Joana M. Tavares,
2
Roger A. Pedersen,
1,5
Ludovic Vallier,
1,3,4
and Mark R.N. Kotter
1,2,
*
1
Anne McLaren Laboratory, Wellcome Trust-MRC Stem Cell Institute, University of Cambridge, Cambridge CB2 0SZ, UK
2
Department of Clinical Neuroscience, University of Cambridge, Cambridge CB2 0QQ, UK
3
Department of Surgery, University of Cambridge, Cambridge CB2 0QQ, UK
4
Wellcome Trust Sanger Institute, Hinxton, Cambridge CB10 1SA, UK
5
Department of Paediatrics, University of Cambridge, Cambridge, CB2 0QQ, UK
6
Co-first author
7
Present address: Department of Neurology, University of Mu
¨nster, 48149 Mu
¨nster, Germany
8
Present address: Department of Pathology, University of Washington, Seattle, WA 98109, USA
*Correspondence: pawlowsk@uni-muenster.de (M.P.), mrk25@cam.ac.uk (M.R.N.K.)
http://dx.doi.org/10.1016/j.stemcr.2017.02.016
SUMMARY
The isolation or in vitro derivation of many human cell types remains challenging and inefficient. Direct conversion of human plurip-
otent stem cells (hPSCs) by forced expression of transcription factors provides a potential alternative. However, deficient inducible gene
expression in hPSCs has compromised efficiencies of forward programming approaches. We have systematically optimized inducible
gene expression in hPSCs using a dual genomic safe harbor gene-targeting strategy. This approach provides a powerful platform for
the generation of human cell types by forward programming. We report robust and deterministic reprogramming of hPSCs into neurons
and functional skeletal myocytes. Finally, we present a forward programming strategy for rapid and highly efficient generation of human
oligodendrocytes.
INTRODUCTION
Despite major efforts to develop robust protocols for scalable
generation of human cell types from easily accessible and
renewable sources,the differentiation of human pluripotent
stem cells (hPSCs) into specific cell types often remains
cumbersome, lengthy, and difficult to reproduce. Moreover,
the recapitulation of developmental stages in vitro yields
fetal cells that often do not reach full maturation (Cohen
and Melton, 2011). More recently, forced expression of line-
age-specific master regulators resulting in direct reprogram-
ming of somatic cell types has provided an efficient alterna-
tive to directed differentiation (Huang et al., 2014; Ieda et al.,
2010; Zhou et al., 2008). In particular, the direct conversion
of hPSCs, termed forward programming (Moreau et al.,
2016), combines the advantages of hPSC differentiation
and direct cellular reprogramming, enabling scalable and
rapid generation of human cell types (Zhang et al., 2013).
Currently available forward programming protocols are
largely based on lentiviral transduction of hPSCs, which
results in variegated expression or complete silencing of
transgenes (Darabi et al., 2012; Smith et al., 2008). Addi-
tional purification steps are usually necessary for enriching
the desired cell type. Lentiviral approaches randomly insert
transgenes into the genome bearing the risk of unwanted
interference with the endogenous transcriptional program.
Therefore, refinements to the current forward program-
ming approaches are desirable.
As the result of a systematic effort to optimize gene
expression in hPSCs, we arrived at a robust hPSC forward
programming platform by targeting all components of
the Tet-ON system required for inducible expression
of transcription factors into genomic safe harbor sites
(GSHs) (Sadelain et al., 2012). The Tet-ON system consists
of two components: a constitutively expressed transcrip-
tional activator protein responsive to doxycycline (dox)
(reverse tetracycline transactivator [rtTA]), and an inducible
promoter regulated by rtTA (Tet-responsive element) that
drives expression of the transgene (Baron and Bujard,
2000). Previous GSH-targeting strategies of the Tet-ON
system relied on introducing both elements into the
AAVS1 GSH of hPSCs, either separately (Hockemeyer
et al., 2009), or together (using an all-in-one Tet-ON vector)
(Ordova
´s et al., 2015; Qian et al., 2014). Compared with
these designs, we reasoned that targeting each of the two el-
ements of the Tet-ON system into a different GSH would
have several advantages: inducible overexpression based
on dual GSH targeting would not be affected by promoter
interference between the two transgenes (Baron and Bujard,
2000), while homozygous GSH targeting would maximize
the number of safely targeted transgene copies. Moreover,
the larger cargo capacity in each of the transgenes would
permit increased flexibility for transgene design, thus allow-
ing the insertion of large reprogramming cassettes.
Here we show that an optimized approach based on dual
GSH targeting of Tet-ON-controlled transgenes results in
Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017 jª2017 1
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

C
A
B
ED
F
G
IH
JK
- allele a
- allele b
ROSA26
e1 e2
e1
EGFP
e2
e1 e2
e1
rtTA
CAG
e2
HET
HET
e1
EGFP
e2
e1
EGFP
e2
e1 e2
e1
rtTA
CAG
e2
HET
HOM
e1 e2
e1
EGFP
e2
e1
rtTA
CAG
e2
e1
rtTA
CAG
e2
HOM
HET
e1
EGFP
e2
e1
EGFP
e2
e1
rtTA
CAG
e2
e1
rtTA
CAG
e2
HOM
HOM
- allele a
- allele b
AAVS1
(= HET-HET) (= HET-HOM) (= HOM-HET) (= HOM-HOM)
hPSCs
rtTA
hPSCs
Nucleofection Neo selection
+ clonal isolation
iEGFP
hPSCs
Puro selection
+ clonal isolation
Induced
iEGFP
hPSCs
+dox
Cas9n
guide A
Cas9n
guide B
Lipofection
ZFN
left
ZFN
right
AAVS1
rtTA
100101102103104
0
20
40
80
100
Cells (% of Max)
EGFP (FITC:log)
Sample MFI EGFP+ (%)
i-EGFP DOX 3030
1
99.8
0
10
-1
10
0
10
1
10
2
10
3
0
25
50
75
100
Doxycycline [ng/mL]
MFI (% of Max)
Dose esponse
Pluripotency
-4 -2 0 2 4 8
0
25
50
75
100
Day
MFI (% of Max)
Induction Kinetics
+ dox dox withdrawal
-Tubulin
hESC
rtTA (M2) rtTA
(HET)
rtTA
(HOM)
rtTA
(long exp.)
rtTA
(short exp.)
11/31 1/3
HOM-HOMHOM-HETHET-HOMHET-HET
WT
CAG-EGFP
CT
DOX
CT
DOX
CT
DOX
CT
DOX
0
100
1000
2000
3000
4000
MFI
EGFP expression levelrtTA expression
****
**
Pluripotency
DAPI
NKX2.5
NANOG
DAPI
DAPI
DAPI
EOMES
EGFP
EGFP
EGFP
EGFP
Merge
Merge
Merge
Merge
Definitive Endoderm
NeuroectodermLateral Plate Mesoderm
200μm
200μm
DOX
200μm
200μm
200μm
200μm
DOX
DOX
DOX
200μm
200μm
(legend on next page)
2Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

homogeneous, controllable, and extremely high expres-
sion of inducible transgenes in hPSCs. Application of the
optimized overexpression platform enabled us to develop
rapid and deterministic forward programming protocols
for mature human cell types.
RESULTS
Development of an Optimized Inducible Transgene
Overexpression Method by Dual GSH Targeting
To optimize inducible transgene overexpression from
GSHs, we generated human embryonic stem cells (hESCs)
with inducible EGFP (i-EGFP) expression. Initially, we
tested four different designs (Figure S1A). These comprised
two all-in-one targeting constructs in which both rtTA and
i-EGFP expression cassettes (third-generation Tet-ON sys-
tem) were inserted into the same allele of the AAVS1 GSH
locus. In these constructs, the rtTA expression was under
the control of either an EF1aor CAG promoter. The other
two transgene designs were based on spatial separation of
the activator and responder into two distinct GSHs (Fig-
ure S1A). For this purpose, we sequentially targeted the
rtTA cassette into the human ROSA26 GSH (Bertero et al.,
2016) and an i-EGFP transgene into the AAVS1 GSH (Fig-
ures 1A and S1A–S1E) (Hockemeyer et al., 2009). Robust
and homogeneous inducible transgene expression was
achieved only when the dual GSH approach and a CAG
promoter for rtTA expression was used (Figure S1A). Impor-
tantly, the dual GSH-targeting approach was highly effi-
cient (Table S1), and did not affect hESC self-renewal or
differentiation (Figures 1H–1K).
Using the CAG promoter-based dual GSH-targeting
approach, we selected clonal lines that carried either one
or two copies of each of the transgenes (Figure 1B), and
observed that homozygous targeting of both elements
allowed maximal inducible overexpression (Figures 1C,
1D, S1E, and S1F). Under these conditions, EGFP expres-
sion was induced homogeneously in all cells, consistent
across multiple clones, and more than 50-fold higher
compared with EGFP expression via the strong constitutive
CAG promoter (Figures 1D, 1E, S1E, and S1F). Maximal
EGFP levels were reached approximately 4 days after induc-
tion, and expression was quickly reversed upon dox with-
drawal (Figure 1F). Moreover, EGFP expression could be
titrated by adjusting the dose of dox (Figure 1G). I-EGFP
expression was highly efficient in hESCs (Figures 1E and
1H), and during germ layer differentiation (Figures 1I–1K
and S1G). There was no detectable background expression
of EGFP in the absence of dox (Figures 1D, 1E, and S1G).
Taken together, these results established that homozy-
gous dual GSH targeting of the Tet-ON system is a powerful
strategy for homogeneous and controllable expression of
inducible transgenes in hPSCs and their derivatives. We
will refer to this platform as ‘‘OPTi-OX’’ (optimized induc-
ible overexpression).
Human Induced Neurons
To test the applicability of the OPTi-OX platform for for-
ward programming of hPSCs into mature cell types, we first
chose to generate excitatory cortical neurons, as previous
studies demonstrated that these can be readily derived
by lentiviral overexpression of pro-neuronal transcription
factors in hPSCs (Zhang et al., 2013).
To this end, we generated NGN2 OPTi-OX hPSCs (Fig-
ure 2A; Table S1), and treated them with dox in chemically
defined neuronal culture medium (Zhang et al., 2013).
Induction of NGN2 expression (Figure S2A) resulted in
Figure 1. Development of an Optimized Inducible Gene Overexpression System
(A) Workflow for targeting the hROSA26 and AAVS1 loci with the Tet-ON system in hPSCs for inducible EGFP expression (i-EGFP). Cas9n,
D10A nickase mutant Cas9 endonuclease; ZFN, zinc-finger nucleases; rtTA, reverse tetracycline transactivator; TRE, Tet-responsive
element.
(B) Schematic of the four outcomes following generation of dual GSH-targeted inducible EGFP hESCs: clonal lines were categorized based
on the number of successfully targeted alleles of the hROSA26 and AAVS1 loci.
(C) Detection of the rtTA protein by western blot in heterozygous (HET) and homozygous (HOM) hROSA26-CAG-rtTA hESCs. Homozygous
targeting results in increased rtTA protein expression. hESCs with random integration of a second-generation rtTA (M2-rtTA) and wild-type
hESCs are shown as positive and negative reference. a-Tubulin, loading control.
(D) Median fluorescent intensity (MFI) of EGFP expression in the various dual GSH-targeted i-EGFP hESCs described in (B). Cells were
analyzed by flow cytometry in non-induced conditions (CTR) or following 5 days of dox. AAVS1-CAG-EGFP and wild-type (WT) hESCs were
included for comparison. Statistical analysis of dox-treated groups demonstrated that EGFP levels were highest in double-homozygous
clones (each data point, n = 1–5, represents a clonal line; mean ±SEM; one-way ANOVA with post hoc Dunnett’s test; **p < 0.01,
****p < 0.0001).
(E) Flow cytometry of EGFP OPTi-OX hESCs after 5 days of dox treatment. Non-induced cells were included as negative control.
(F and G) EGFP induction and rescue kinetics (F) and dox dose-response (G) in EGFP OPTi-OX hESCs detected by flow cytometry (n = 2
biological replicates; mean ±SEM; all values normalized to the maximum fluorescence intensity after 5 days of dox).
(H–K) Immunocytochemistry (ICC) for lineage-specific markers in undifferentiated EGFP OPTi-OX hESCs and following differentiation into
the germ layers.
Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017 3
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

downregulation of pluripotency factors and initiation of
the neuronal transcriptional program (Figure 2B). Dox-
treated cells extended neuronal processes as early as
3 days post induction. After 1 week, all cells displayed a
neuronal morphology and expressed the pan-neuronal
markers bIII-tubulin and MAP2 (Figures 2C–2E and Movie
S1). At this stage, induced neurons (i-Neurons) showed
strong expression of forebrain markers BRN2 and FOXG1,
and of glutamatergic neuronal genes GRIA4, VGLUT1,
and VGLUT2 (Figures 2B and 2F), indicative of an
excitatory cortical neuronal identity of the forward-pro-
grammed cells, consistent with previous reports (Zhang
et al., 2013). Short pulses of dox treatment for 4 days or
longer sufficed for complete conversion, and converted
cells did not rely on continuous transgene expression (Fig-
ures S2B and S2C). Importantly, we did not observe any
reduction in the efficiency of generating i-Neurons over
extended culture periods of the inducible hESCs (>25 pas-
sages, Figure 2E). Finally, we confirmed the applicability
of the NGN2 OPTi-OX system in hiPSCs (Figure 2G).
Collectively, these results demonstrated that OPTi-OX en-
ables robust and rapid forward programming of hPSCs
into cortical neurons.
Human Induced Skeletal Myocytes
To further explore the potential of OPTi-OX for forward
programming of hPSCs, we focused on generating human
skeletal myocytes. Existing protocols for the directed differ-
entiation of skeletal myocytes from hPSCs are difficult,
time consuming, and result in low and variable yields
(Chal et al., 2015). On the other hand, myogenic transdif-
ferentiation has been achieved by overexpressing the
transcription factor MYOD1 in somatic cell types, but the
ability of hPSCs to undergo MYOD1-induced forward
programming is a matter of debate (Abujarour et al.,
2014; Albini et al., 2013; Tanaka et al., 2013).
We therefore generated MYOD1 OPTi-OX hPSCs (Figures
3A and 3B and Table S1). However, induction of MYOD1
expression following dox treatment resulted in cell death
within 3–5 days, regardless of the culture medium used
B
DF
G
i-Neurons
NGN2
OPTi-OX hPSCs
Dox
i-Neurons
Day 1 Day 3 Day 7
100μ001m μm 100μm
100
102
104
106
108
Expression
relative to hPSCs
047 14
Day
Neuronal Genes
NGN2
BRN2
FOXG1
MAP2
SYP
VGLUT2
GRIA4
Day 14
DAPI/TUBB3/VGLUT1
DAPI/TUBB3
50μm
OPTi-OX iPSCs
Day 14
Control
i-Neuron
i-Neuron
(+P25)
0
50
100
Cells (%)
TUBB3
+
cells
Day 4
DAPI/TUBB3
DAPI/MAP2
Day 7 Day 14
0.0
0.5
1.0
Expression
relative to hPSCs
407
14
Day
Pluripotency Genes
NANOG
OCT4
200μm200μm200μm
200μm200μm200μm200μm
CE
A
Figure 2. Forward Programming of hPSCs into Neurons
(A) Experimental approach for conversion of NGN2 OPTi-OX hPSCs into i-Neurons.
(B) Time course of i-Neuron generation from hESCs by qPCR demonstrating the expression pattern of pluripotency factors (OCT4
and NANOG), pan-neuronal (MAP2 and SYP), forebrain (BRN2,FOXG1), and glutamatergic neuronal marker genes (VGlut2,GRIA4)(n=3
biological replicates; mean ±SEM; relative to PBGD and normalized to pluripotency).
(C) Phase contrast images illustrating the morphological changes during i-Neuron generation (a corresponding time-lapse is shown in
Movie S1).
(D) ICC for the pan-neuronal marker proteins bIII-tubulin (TUBB3) and microtubule-associated protein 2 (MAP2) during the generation of
i-Neurons.
(E) Quantification of bIII-tubulin-positive neuronal cells by ICC after 1 week of induction. Undifferentiated cells were used as negative
control, and numbers are reported for i-Neuron generation in newly isolated NGN2 OPTi-OX hESCs and after 25 passages.
(F and G) ICC for neuronal markers in i-Neurons 14 days after induction.
4Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

i-Myocytes
Dox
Retinoic Acid
MYOD1
OPTi-OX hPSCs
02468
100
102
104
Day
Expression
relative to hPSCs
Myocyte Genes
MYDO1
MYOD1 (endo)
MYOG
RYR1
DMD
DES
DAPI MYOD1OCT4/NANOG
DAPI MYOG/MHC
DAPI
DAPI MYOG/TNNT
MYOG/MHC
50μm
DOX - Day 8DOX - Day 8
200μm
10
0
10
1
10
2
10
3
10
4
Myosin Heavy Chain (PE:log)
0
20
40
60
80
100
Cells (% of Max)
Control
i-Myo
i-Myo
(+P50)
0
50
100
Cells (%)
MHC
+
cells
hPSCs
i-Myocytes
Sample
A
C
B
D
E
F
GH
I
J
0.0
0.5
1.0
1.5
2.0 MYOG
Doxycycline [ng/μL]
10
0
10
1
10
2
10
3
10
0
10
1
10
2
10
3
10
0
10
1
10
2
10
3
0
5
10
15
20
25 MYOD1 (endo)
Doxycycline [ng/μL]
DAPI
MYOD1 OPTi-OX (DOX day 5)
MYOG/MHC
Doxycycline
0 ng/μL
200μm
24 ng/μL
200μm
64 ng/μL
200μm
125 ng/μL
200μm
1000 ng/μL
200μm
0
10
20
30
40
50
Doxycycline [ng/μL]
Exression
relative to PBGD
MYOD1 (total)
MYOD1 OPTi-OX
CTRDOX Day 2
MYOD1 - DOX Day 5
CTRplus RA
200μm
200μm
200μm
200μm
(legend on next page)
Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017 5
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

(data not shown). These findings demonstrated that
MYOD1 overexpression alone was not sufficient to drive
myogenesis in hPSCs, in agreement with the postulated ex-
istence of epigenetic barriers preventing forced myogenesis
(Albini et al., 2013).
Cellular reprogramming strategies can be enhanced by
combining transcription factor overexpression with extra-
cellular signaling cues (Bar-Nur et al., 2014). We conducted
a systematic screen for pro-myogenic factors by modu-
lating key signaling cascades that are implicated in primi-
tive streak formation, somitogenesis, and myogenesis
(Figure S3A). We found that the addition of all-trans reti-
noic acid (RA) in conjunction with MYOD1 overexpression
was sufficient for rapid and deterministic conversion of
hPSCs into myogenin and myosin heavy chain double-
positive myocytes after 5 days of induction (Figures 3C
and S3A). The effect of RA was concentration dependent
(data not shown), and mediated at least in part through
the receptor isoforms RARaand RARb(Figures S3B
and S3C).
Following minor optimization of the culture condi-
tions (see the Supplemental Experimental Procedures), we
arrived at a protocol resulting in nearly pure induced skel-
etal myocytes (i-Myocytes). Reprogrammed cells developed
typical spindle-like, elongated morphology, underwent
extensive cell fusion, and exhibited strong and homoge-
neous myogenic marker expression on mRNA and protein
levels (Figures 3D–3H, S3D, and S3E; Movie S2). Further-
more, the addition of nanomolar concentrations of acetyl-
choline (ACh) or the selective ACh-receptor agonist carba-
chol resulted in muscle fiber contraction, demonstrating
the functionality of the i-Myocytes (Movie S3). Similar
results were obtained with i-Myocytes generated from
MYOD1 OPTi-OX hiPSCs (Figure S3F). Importantly, induc-
tion efficiency did not decrease over extended culture pe-
riods (>50 passages, Figure 3H), thus demonstrating the
robustness and reproducibility of this method. Finally, we
noted that the levels of the MYOD1 transgene following in-
duction positively correlated with conversion efficiency
(Figures 3I and 3J), which highlights the importance of a
robust gene-delivery method. In conclusion, these data
demonstrated that the OPTi-OX platform enables robust
and rapid forward programming of hPSCs into skeletal
myocytes.
Human Induced Oligodendrocytes
Encouraged by our results deriving neurons and myocytes,
we sought to utilize the same overexpression system to
develop a forward programming protocol for oligodendro-
cytes. Oligodendrocytes are of critical importance for CNS
function and their loss or dysfunction plays a key role in
many neurological diseases. Unlike neurons (Zhang et al.,
2013), protocols for efficient generation of human oligo-
dendrocytes from renewable sources remain an unmet
need: currently available hPSC differentiation protocols
are extremely long (up to 200 days) and yield heteroge-
neous cell populations (Douvaras et al., 2014; Stacpoole
et al., 2013; Wang et al., 2013).
We generated OPTi-OX hPSCs bearing inducible SOX10
either alone or in combination with OLIG2 in the form
of a polycistronic expression cassette (Figure 4A). Although
cells induced with SOX10 alone robustly expressed
the oligodendrocyte precursor (OPC) marker O4 after
10 days of induction, these cells failed to differentiate into
myelin-expressing cells and died (Figure 4A). In contrast,
the OLIG2-SOX10 overexpressing cells progressed from
an O4-positive progenitor stage into a mature CNP/MBP-
positive phenotype after 20 days of induction (Figure 4A).
Moreover, gene expression analysis confirmed that
OLIG2-SOX10 OPTi-OX hPSCs induced in oligodendrocyte
Figure 3. Forward Programming of hPSCs into Skeletal Myocytes
(A) Experimental approach for rapid single-step conversion of MYOD1 OPTi-OX hPSCs into skeletal myocytes (i-Myocytes) following
treatment with dox and RA.
(B) Representative ICC for MYOD1 before (CTR) and after induction with dox. This demonstrates homogeneous induction of transgene
expression, paralleled by downregulation of the pluripotency factors NANOG and OCT4.
(C) Effect of RA on myocyte forward programming compared with otherwise identical control (CTR) induction conditions (see Figure S2B for
the entire signaling molecule screen).
(D) qPCR of the temporal expression pattern of pluripotency factors (top panel) and myocyte marker genes during i-Myocyte generation
(n = 3 biological replicates, mean ±SEM; relative to PBGD and normalized to pluripotency).
(E and F) ICC for skeletal myocyte markers in i-Myocytes.
(G and H) Quantification of MHC-positive cells by flow cytometry 10 days after induction. Undifferentiated cells were used as negative
control, and figures are reported for i-Myocyte generation in newly isolated MYOD1 OPTi-OX hESCs, or in the same cells following
50 passages (+P50) (n = 3 biological replicates; mean ±SEM).
(I) qPCR for total MYOD1, endogenous MYOD1, and MYOG 2 days post induction with different dox concentrations (n = 3 biological
replicates; mean ±SEM).
(J) ICC for myogenin and myosin heavy chain following 5 days of induction with different dox concentrations. Non-converted, proliferative
cell clusters appeared when the dox concentration was lowered to 0.125 mg/mL. Further reduction of dox resulted in an increase in non-
myocyte cell populations.
6Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

A
G
IJKLM
H
B
CD
F
E
DAPI/O4 DAPI/MBP/CNP
Day 1 Day 10 Day 20
ControlOLIG2 + SOX10 SOX10
OPTi-OX
DAPI SOX10/OLIG2
200μm
200μm
200μm
200μm
200μm
200μm
200μm
200μm
200μmi-Oligodendrocyte
Precursors (i-OPCs) i-Oligodendrocytes (i-OLs)
Dox
PDGFaa
+FGF2
Dox
OLIG2-SOX10
OPTi-OX hPSCs
Faa
GF2
x
-Oligodendrocytes
i-Oligodendrocyte
Precursors (i-OPCs)
oxo
PDGF
+
D
a
Dox
P1
P2
P3
0
20
40
60
80
100
Passage no
BrdU+ cells
BrdU+/DAPI [%]
+
-
DOX
DOX
DAPI/O4
DAPI/BrdU
DAPI/A2B5/PDGFRA
100μm
200μm
Myelin Genes
0 5 811141720 30
100
101
102
103
104
105
106
Day
Expr ession
relative to hPSCs
CNP
PLP
MBP
MAG
MOG
0
50
100
Cells (%)
CNP
PLP
i-OL
i-OL
(+P50)
DAPI/PLP
50μm
DAPI/MBP/PLP
50μm
DAPI/MBP
50μm
DAPI/CNP
50μm
DAPI/MBP/PLP
200μm
200μm
PLURI
i-OPC d8
i-OPC d11
0.00
0.05
0.10
0.15
Expression
relative to PBGD
OLIG2
(endo)
****
***
PLURI
i-OPC d8
i-OPC d11
0
5
10
15
SOX10
(endo)
****
**
PLURI
i-OPC d8
i-OPC d11
0.00
0.02
0.04
0.06
0.08
0.10
NKX2.2
*
PLURI
i-OPC d8
i-OPC d11
0.00
0.02
0.04
0.06
OLIG1
*
**
PLURI
i-OPC d8
i-OPC d11
0.0
0.2
0.4
0.6
PDGFRA
***
**
PLURI
i-OPC d8
i-OPC d11
i-Neuron (ctr)
0.0
0.1
0.2
0.3
CSPG4
*
ns
Figure 4. Forward Programming of hPSCs into Oligodendrocytes
(A) ICC for inducible transgenes after 1 day of induction (left column), the OPC marker O4 after 10 days (middle), and the oligodendrocyte
markers CNP and MBP after 20 days (right).
(B) Experimental approach for rapid conversion of OLIG2-SOX10 OPTi-OX hPSCs into the oligodendrocyte lineage cells (i-OPCs and i-OLs).
(C and D) Characterization of i-OPCs by ICC for OPC surface markers (A2B5, O4, PDGFRA).
(legend continued on next page)
Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017 7
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

medium (Douvaras et al., 2014) supplemented with the
mitogens PDGFaa and FGF2 first passed through an OPC-
like stage, during which they remained proliferative and
co-expressed typical OPC markers (Figures 4B–4F). Remark-
ably, following withdrawal of mitogens, i-OPCs differenti-
ated into mature oligodendrocytes, expressing the typical
myelin-associated proteins (Figures 4G–4M).
DISCUSSION
OPTi-OX is the result of a systematic effort to optimize gene
expression in hPSCs. It relies on a dual GSH-targeting strat-
egy for the Tet-ON system, overcoming the limitations of
viral-mediated transgene delivery forward programming
protocols (Abujarour et al., 2014; Darabi et al., 2012; Zhang
et al., 2013) and allows stronger and more controlled trans-
gene overexpression compared with previous targeting ap-
proaches (Gonza
´lez et al., 2014; Hockemeyer et al., 2009;
Ordova
´s et al., 2015). Table S2 compares the gene-delivery
methods that have been used for transcription factor
expression in different hPSC forward programming ap-
proaches. Moreover, site-specific insertion of the two com-
ponents of the inducible gene expression system mini-
mizes genomic off-target effects and together with the
chemically defined medium compositions enhances the
reproducibility of the protocols.
The functionality of our platform is exemplified through
the production of several cell types. First, we show that
NGN2 and MYOD1 OPTi-OX hPSCs can be used as an inex-
haustible source for highly scalable, rapid, single-step, vi-
rus-free, chemically defined, fully reproducible, and deter-
ministic generation of i-Neurons and i-Myocytes.
Finally, we successfully applied the OPTi-OX platform to
develop a forward programming protocol for generating
human oligodendrocytes. Recent studies demonstrated
that forced expression of transcription factors allows direct
conversion of rodent fibroblasts (Najm et al., 2013; Yang
et al., 2013), and primary human fetal neural stem cells
(Wang et al., 2014) into OPCs, but the reprogramming of
renewable human cell sources into oligodendrocytes has
not been reported. While i-OPCs undergo the expected
morphological changes, and express mature markers in
monocultures in vitro, further characterization of the cells
using co-culture models and transplantation into myelin-
deficient mutants is needed.
Human oligodendrocytes are of considerable interest for
several applications. The efficiency and speed of the pre-
sented forward programming system will enable high-
throughput drug screens and toxicology testing, in vitro
modeling of hereditary leukodystrophies, and the devel-
opment of cell-transplantation strategies (Goldman et al.,
2012).
Transcription factor combinations for direct cellular
reprogramming into many cell types of clinical interest
are now available, including cardiomyocytes (Ieda et al.,
2010), pancreatic bcells (Zhou et al., 2008), and hepato-
cytes (Huang et al., 2014). We anticipate that the OPTi-
OX platform will be applicable for the generation of
many other cell types. Overall, the presented method can
provide the basis for inexhaustible, high-throughput, ho-
mogeneous, and large-scale manufacturing of many hu-
man cell types.
EXPERIMENTAL PROCEDURES
Gene Targeting
Targeting of the hROSA26 and AAVS1 locus was performed as
described recently (Bertero et al., 2016). Targeting of the hROSA26
locus was done by nucleofection. Neomycin-resistant colonies
were picked and screened by genotyping. Correctly hROSA26-
rtTA-targeted clones were subsequently targeted with the inducible
transgene cassette in the AAVS1 locus by lipofection. Resulting
puromycin-resistant colonies were picked and re-analyzed by
genotyping.
Inducible Transgene Expression and Forward
Programming
Inducible overexpression was performed with dual GSH-targeted
OPTi-OX hPSCs. Expression of inducible transgenes was prompted
(E) Characterization of i-OPCs by qPCR compared with hPSCs (PLURI). As transcription of CSPG4 (NG2) was also detected in hPSCs, we
included i-Neurons as negative control (n = 3 biological replicates; mean ±SEM; all values relative to PBGD; one-way ANOVA with post hoc
Dunnett’s test; *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, p > 0.05).
(F) Immunostaining for BrdU (left panel) and quantification of BrdU-positive cells following three serial passages of i-OPCs every 4 days
and concomitant BrdU-pulses (n = 3 biological replicates; mean ±SEM; P, passage).
(G) qPCR of the temporal expression pattern of genes encoding for the myelin-associated proteins (CNP,MAG,MBP,MOG, and PLP) during
i-OL generation. OLIG2-SOX10 OPTi-OX hPSCs were induced in oligodendrocyte medium supplemented with PDGFaa and FGF2. After 1 week
of induction, mitogens were withdrawn to enable terminal differentiation (n = 3 biological replicates; mean ±SEM; all values relative to
PBGD and normalized to pluripotency).
(H) Quantification of CNP and PLP expressing i-OLs derived from OLIG2-SOX10 OPTi-OX hPSCs after 20 days of induction by ICC. Undif-
ferentiated cells were used as negative control, and figures are reported for newly isolated OLIG2-SOX10 OPTi-OX hPSCs and after 50
passages (+P50).
(I–M) ICC providing an overview (I) and high-magnifications (J–M) of mature pre-myelinating oligodendrocytes.
8Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

by adding dox to the culture medium. For forward programming
into neurons, skeletal myocytes, and oligodendrocytes, standard
medium conditions for the derivation of the respective cell types
were used. Gene and protein expression analysis was performed
as described recently (Bertero et al., 2016). Please refer to the Sup-
plemental Experimental Procedures for details on culture condi-
tions and analysis techniques.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental
Procedures, three figures, two tables, and three movies and can
be found with this article online at http://dx.doi.org/10.1016/j.
stemcr.2017.02.016.
AUTHOR CONTRIBUTIONS
M.P. conceived the study, designed and performed experiments,
analyzed data, and wrote the first draft of the manuscript. D.O. de-
signed and performed experiments and analyzed the data. A.B. de-
signed and performed experiments, analyzed data, and wrote the
manuscript. J.M.T. performed additional experiments. R.A.P. pro-
vided expert advice. L.V. supervised and supported the study.
M.R.N.K. conceived, supervised, supported the study, and finalized
the manuscript.
ACKNOWLEDGMENTS
We thank Kosuke Yusa for providing the AAVS1 ZFN plasmids.
Research in the senior author’s laboratory is supported by a core
support grant from the Wellcome Trust and MRC to the Wellcome
Trust-Medical Research Council Cambridge Stem Cell Institute.
Further support was provided by a research fellowship from the
German Research Foundation (DFG PA2369/1-1to M.P.), a British
Heart Foundation PhD Studentship (FS/11/77/39327 to A.B.), a
Clinician Scientist Award from the National Institute for Health
Research UK (CS-2015-15-023 to M.R.N.K.), and the Qatar Founda-
tion (to M.R.N.K.). The WellcomeTrust – Medical Research Council
Cambridge Stem Cell Institute is supported by core funding from
the Wellcome Trust and MRC. The views expressed in this publica-
tion are those of the authors and not necessarily those of the NHS,
the National Institute for Health Research or the Department of
Health. Patent protection has been sought for the dual/multiple
safe harbour site approach and individual reprogramming proto-
cols detailed in the present manuscript.
Received: June 13, 2016
Revised: February 17, 2017
Accepted: February 17, 2017
Published: March 23, 2017
REFERENCES
Abujarour, R., Bennett, M., Valamehr, B., Lee, T.T., Robinson, M.,
Robbins, D., Le, T., Lai, K., and Flynn, P. (2014). Myogenic differen-
tiation of muscular dystrophy-specific induced pluripotent stem
cells for use in drug discovery. Stem Cells Transl. Med. 3, 149–160.
Albini, S., Coutinho, P., Malecova, B., Giordani, L., Savchenko, A.,
Forcales, S.V., and Puri, P.L. (2013). Epigenetic reprogramming of
human embryonic stem cells into skeletal muscle cells and gener-
ation of contractile myospheres. Cell Rep. 3, 661–670.
Bar-Nur, O., Brumbaugh, J., Verheul, C., Apostolou, E., Pruteanu-
Malinici, I., Walsh, R.M., Ramaswamy, S., and Hochedlinger, K.
(2014). Small molecules facilitate rapid and synchronous iPSC gen-
eration. Nat. Methods 11, 1170–1176.
Baron, U., and Bujard, H. (2000). Tet repressor-based system for
regulated gene expression in eukaryotic cells: principles and ad-
vances. Methods Enzymol. 327, 401–421.
Bertero, A., Pawlowski, M., Ortmann, D., Snijders, K., Yiangou, L.,
Cardoso de Brito, M., Brown, S., Bernard, W.G., Cooper, J.D.,
Giacomelli, E., et al. (2016). Optimized inducible shRNA and
CRISPR/Cas9 platforms for in vitro studies of human development
using hPSCs. Development 143, 4405–4418.
Chal, J., Oginuma, M., Al Tanoury, Z., Gobert, B., Sumara, O., Hick,
A., Bousson, F., Zidouni, Y., Mursch, C., Moncuquet, P., et al.
(2015). Differentiation of pluripotent stem cells to muscle fiber
to model Duchenne muscular dystrophy. Nat. Biotechnol. 33,
962–969.
Cohen, D.E., and Melton, D.A. (2011). Turning straw into gold:
directing cell fate for regenerative medicine. Nat. Rev. Genet. 12,
243–252.
Darabi, R., Arpke, R.W., Irion, S., Dimos, J.T., Grskovic, M., Kyba,
M., and Perlingeiro, R.C.R. (2012). Human ES- and iPS-derived
myogenic progenitors restore DYSTROPHIN and improve contrac-
tility upon transplantation in dystrophic mice. Cell Stem Cell 10,
610–619.
Douvaras, P., Wang, J., Zimmer, M., Hanchuk, S., O’Bara, M.A., Sa-
diq, S., Sim, F.J., Goldman, J., and Fossati, V. (2014). Efficient gen-
eration of myelinating oligodendrocytes from primary progressive
multiple sclerosis patients by induced pluripotent stem cells. Stem
Cell Rep. 3, 250–259.
Goldman, S.A., Nedergaard, M., and Windrem, M.S. (2012). Glial
progenitor cell-based treatment and modeling of neurological dis-
ease. Science 338, 491–495.
Gonza
´lez, F., Zhu, Z., Shi, Z.D., Lelli, K., Verma, N., Li, Q.V., and
Huangfu, D. (2014). An iCRISPR platform for rapid, multiplexable,
and inducible genome editing in human pluripotent stem cells.
Cell Stem Cell 15, 215–226.
Hockemeyer, D., Soldner, F., Beard, C., Gao, Q., Mitalipova, M.,
DeKelver, R.C., Katibah, G.E., Amora, R., Boydston, E.A., Zeitler,
B., et al. (2009). Efficient targeting of expressed and silent genes
in human ESCs and iPSCs using zinc-finger nucleases. Nat. Bio-
technol. 27, 851–857.
Huang, P., Zhang, L., Gao, Y., He, Z., Yao, D., Wu, Z., Cen, J., Chen,
X., Liu, C., and Hu, Y. (2014). Direct reprogramming of human
fibroblasts to functional and expandable hepatocytes. Cell Stem
Cell 14, 370–384.
Ieda, M., Fu, J.-D., Delgado-Olguin, P., Vedantham, V., Hayashi, Y.,
Bruneau, B.G., and Srivastava, D. (2010). Direct reprogramming of
fibroblasts into functional cardiomyocytes by defined factors. Cell
142, 375–386.
Moreau, T., Evans, A.L., Vasquez, L., Tijssen, M.R., Yan, Y., Trotter,
M.W., Howard, D., Colzani, M., Arumugam, M., Wu, W.H., et al.
(2016). Large-scale production of megakaryocytes from human
Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017 9
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016

pluripotent stem cells by chemically defined forward program-
ming. Nat. Commun. 7, 11208.
Najm, F.J., Lager, A.M., Zaremba, A., Wyatt, K., Caprariello, A.V.,
Factor, D.C., Karl, R.T., Maeda, T., Miller, R.H., and Tesar, P.J.
(2013). Transcription factor-mediated reprogramming of fibro-
blasts to expandable, myelinogenic oligodendrocyte progenitor
cells. Nat. Biotechnol. 31, 426–433.
Ordova
´s, L., Boon, R., Pistoni, M., Chen, Y., Wolfs, E., Guo, W.,
Sambathkumar, R., Bobis-Wozowicz, S., Helsen, N., Vanhove, J.,
et al. (2015). Efficient recombinase-mediated cassette exchange
in hPSCs to study the hepatocyte lineage reveals AAVS1 locus-
mediated transgene inhibition. Stem Cell Rep. 5, 918–931.
Qian, K., Huang, C.-L., Chen, H., Blackbourn, L.W., Chen, Y., Cao,
J., Yao, L., Sauvey, C., Du, Z., and Zhang, S.-C. (2014). A simple and
efficient system for regulating gene expression in human pluripo-
tent stem cells and derivatives. Stem Cells 32, 1230–1238.
Sadelain, M., Papapetrou, E.P., and Bushman, F.D. (2012). Safe har-
bours for the integration of new DNA in the human genome. Nat.
Rev. Cancer 12, 51–58.
Smith, J.R., Maguire, S., Davis, L.A., Alexander, M., Yang, F., Chan-
dran, S., Ffrench-Constant, C., and Pedersen, R.A. (2008). Robust,
persistent transgene expression in human embryonic stem
cells is achieved with AAVS1-targeted integration. Stem Cells 26,
496–504.
Stacpoole, S.R.L., Spitzer, S., Bilican, B., Compston, A., Karadottir,
R., Chandran, S., and Franklin, R.J.M. (2013). High yields of oligo-
dendrocyte lineage cells from human embryonic stem cells at
physiological oxygen tensions for evaluation of translational
biology. Stem Cell Rep. 1, 437–450.
Tanaka, A., Woltjen, K., Miyake, K., Hotta, A., Ikeya, M., Yama-
moto, T., Nishino, T., Shoji, E., Sehara-Fujisawa, A., Manabe, Y.,
et al. (2013). Efficient and reproducible myogenic differentiation
from human iPS cells: prospects for modeling Miyoshi myopathy
in vitro. PLoS One 8, e61540.
Wang, S., Bates, J., Li, X., Schanz, S., Chandler-Militello, D., Levine,
C., Maherali, N., Studer, L., Hochedlinger, K., Windrem, M., et al.
(2013). Human iPSC-derived oligodendrocyte progenitor cells
can myelinate and rescue a mouse model of congenital hypomye-
lination. Cell Stem Cell 12, 252–264.
Wang, J., Pol, S.U., Haberman, A.K., Wang, C., O’Bara, M.A., and
Sim, F.J. (2014). Transcription factor induction of human oligoden-
drocyte progenitor fate and differentiation. Proc. Natl. Acad. Sci.
USA 111, E2885–E2894.
Yang, N., Zuchero, J.B., Ahlenius, H., Marro, S., Ng, Y.H., Vierbu-
chen, T., Hawkins, J.S., Geissler, R., Barres, B.A., and Wernig, M.
(2013). Generation of oligodendroglial cells by direct lineage con-
version. Nat. Biotechnol. 31, 434–439.
Zhang, Y., Pak, C., Han, Y., Ahlenius, H., Zhang, Z., Chanda, S.,
Marro, S., Patzke, C., Acuna, C., Covy, J., et al. (2013). Rapid sin-
gle-step induction of functional neurons from human pluripotent
stem cells. Neuron 78, 785–798.
Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J., and Melton, D.A.
(2008). In vivo reprogramming of adult pancreatic exocrine cells
to beta-cells. Nature 455, 627–632.
10 Stem Cell Reports jVol. 8 j1–10 jApril 11, 2017
Please cite this article in press as: Pawlowski et al., Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into
Neurons, Skeletal Myocytes, and Oligodendrocytes, Stem Cell Reports (2017), http://dx.doi.org/10.1016/j.stemcr.2017.02.016