CRISPR Gene Editing in Stem Cells

63 128

transactivators targeting an enhancer of Oct4 to generate mouse
induced pluripotent stem cells. Notably, that study required
co-delivery of vectors directly encoding ectopic C-MYC, KLF4,
and SOX2 to achieve pluripotency (Gao et al., 2013). More
recently, we have demonstrated the direct conversion of primary
mouse embryonic fibroblasts (PMEFs) to skeletal myocytes
using a dCas9-based transactivator targeting the endogenous
Myod1 gene (Chakraborty et al., 2014). Several groups have
also applied CRISPR/Cas9-based transcriptional regulation to
direct the differentiation of human induced pluripotent and em-
bryonic stem cells (Balboa et al., 2015; Chavez et al., 2015;
Wei et al., 2016).
The above examples involve the targeted activation of a single
transcription factor to guide reprogramming or differentiation,
but many approaches require concurrent expression of multiple
factors to efficiently establish a mature phenotype (Takahashi
and Yamanaka, 2006; Vierbuchen et al., 2010). There have
been no examples demonstrating multiplex endogenous gene
activation to induce cellular reprogramming, and the versatility
of that approach for direct conversion to other cell phenotypes
is not known. Moreover, only the report of TALE transcription
factors targeting Oct4 evaluated changes to epigenetic marks
at the target loci (Gao et al., 2013), and this group later reported
that dCas9-based transcriptional activators were inefficient
at endogenous gene activation and reprogramming (Gao et al.,
2014). In this study, we tested the hypothesis that targeted
epigenetic reprogramming of the regulatory elements controlling
expression of lineage-specific transcription factors is sufficient
for direct conversion between cell types by applying dCas9-
based transactivators to the activation of endogenous genes
that directly convert PMEFs to induced neuronal cells (iNs).
RESULTS
Multiplex Endogenous Gene Activation of Neurogenic
Factors in PMEFs
Overexpression of transgenes encoding the transcription factors
Brn2, Ascl1, and Myt1l (BAM factors) has been shown to directly
convert cultured PMEFs to functional induced neuronal cells
(Vierbuchen et al., 2010). We hypothesized that the targeted
activation of the endogenous genes encoding these same
factors in their native chromatin context via a dCas9-based
transactivator could more rapidly and deterministically remodel
the chromatin at the target loci and provide an alternate method
to achieve the reprogramming of PMEFs to iNs (Figure 1A). To
achieve targeted gene activation, we used a transactivator with
both N-terminal and C-terminal VP64 transactivation domains
(VP64 dCas9VP64) (Chakraborty et al., 2014) that generated a
10-fold improvement in activation of ASCL1 in HEK293T cells
at 3 days post-transfection compared to the first-generation
dCas9 transcription factor with a single C-terminal VP64 domain
(Maeder et al., 2013b; Perez-Pinera et al., 2013) (Figure 1B). We
used VP64dCas9VP64 for the remainder of this study.
We used lentiviral delivery to constitutively express
VP64 dCas9VP64 in PMEFs. Initially, we delivered the gRNAs
through transient transfection of plasmid DNA in order to assess
stable reprogramming of cell phenotype following transient
activity of transactivators. The induction of Brn2 and Ascl1
gene expression by VP64dCas9VP64 was attained by delivering
four gRNAs targeted to the putative promoter region directly
upstream of the transcription start site (TSS). The decision to
deliver four gRNAs for each gene was based on the reported
synergistic effects of multiple gRNAs on gene activation (Maeder
et al., 2013b; Mali et al., 2013a; Perez-Pinera et al., 2013). The
optimal gRNAs were selected from a pool of eight gRNAs
through elimination screening (Figure S1A). The gRNAs targeting
regions proximal to the TSS of the Myt1l locus did not induce
detectable levels of activation, but targeting an intronic region
directly upstream of the first coding exon of Myt1l was sufficient
to activate expression (Figure S1B).
Co-transfection of 12 gRNA expression plasmids (CR-BAM),
targeting each of the three endogenous BAM factors with
4 gRNAs, into PMEFs stably expressing VP64dCas9VP64 was
sufficient to induce transcriptional upregulation of all three
endogenous genes when compared to the transfection of a
plasmid encoding firefly luciferase (pLuc; Figure 1C). We also de-
tected Brn2 and Ascl1 protein expression by western blot (Fig-
ure S1C), although we could not detect Myt1l protein using
commercially available antibodies. In addition to gRNA transfec-
tions, we transfected three plasmids encoding the BAM factor
transgenes under the control of the EF1a /HTLV promoter
(pBAM) into the same cells and observed a modest increase in
the mRNA levels of the corresponding endogenous genes
(Figure 1C).
To attain successful reprogramming, it is generally considered
necessary to express the exogenous factors at high levels (Vier-
buchen and Wernig, 2011). Therefore, we compared the total
mRNA and protein levels of Brn2 , Ascl1 , and Myt1l produced
3 days after CR-BAM and pBAM plasmid transfections (Figures
1D–1F). Despite the higher levels of transcriptional activation
from the endogenous loci by CR-BAM (Figure 1C), pBAM trans-
fection generated significantly more total mRNA encoding each
BAM factor than induction by CR-BAM, as determined by qRT-
PCR (Figure 1D). Quantitation of single-cell protein levels from
immunofluorescence staining also revealed significantly higher
single-cell levels of Brn2 and Ascl1 in cells transfected with
pBAM compared to those transfected with CR-BAM (Figures
1E and 1F).
Induction of Neuronal Cells from PMEFs via
VP64dCas9VP64-Mediated Gene Activation
Treated PMEFs were assayed for neuronal phenotypes as
detailed schematically in Figure 2A. We observed an increase
in mRNA of the early pan-neuronal marker bIII tubulin (Tuj1)
3 days after transfection with either pBAM or CR-BAM when
compared to a pLuc control (Figure 2B). We cultured the cells
for 2 weeks in neurogenic medium and analyzed expression
of pan-neuronal markers by immunofluorescence staining.
We identified cells with neuronal morphologies that expressed
Tuj1 in populations transfected with CR-BAM (Figure 2C). A
subset of Tuj1+ cells also expressed the more mature pan-
neuronal marker Map2 (Figure 2C). The generation of Tuj1+
Map2+ cells with neuronal morphologies following treatment
with VP64 dCas9VP64 and gRNAs was contingent on the addition
of a small-molecule cocktail to the medium that has been used
previously for neural differentiation of embryonic stem cells
and has been shown to improve the efficiency of the direct con-
version of human fibroblasts to neurons when used in parallel
Cell Stem Cell 19, 406–414, September 1, 2016 407