pone.0041794 1..12 - journal.pone.0041794.pdf

Elevated IKK

a

Accelerates the Differentiation of Human

Neuronal Progenitor Cells and Induces MeCP2-

Dependent BDNF Expression

Ali Khoshnan

, Paul H. Patterson

Biology Division 216-76, California Institute of Technology, Pasadena, California, United States of America

Abstract

The I

B kinase

(IKK

) is implicated in the differentiation of epithelial and immune cells. We examined whether IKK

also

plays a role in the differentiation and maturation of embryonic human neuronal progenitor cells (NPCs). We find that

expression of an extra copy of IKK

(IKK

) blocks self-renewal and accelerates the differentiation of NPCs. This coincides

with reduced expression of the Repressor Element Silencing Transcription Factor/Neuron-Restrictive Silencing Factor (REST/

NRSF), which is a prominent inhibitor of neurogenesis, and subsequent induction of the pro-differentiation non-coding RNA,

miR-124a. However, the effects of IKK

on REST/NRSF and miR-124a expression are likely to be indirect. IKK

neurons

display extensive neurite outgrowth and accumulate protein markers of neuronal maturation such as SCG10/stathmin-2,

postsynaptic density 95 (PSD95), syntaxin, and methyl-CpG binding protein 2 (MeCP2). Interestingly, IKK

associates with

MeCP2 in the nuclei of human neurons and can phosphorylate MeCP2

in vitro

. Using chromatin immunoprecipitation

assays, we find that IKK

is recruited to the exon-IV brain-derived neurotrophic factor (BDNF) promoter, which is a well-

characterized target of MeCP2 activity. Moreover, IKK

induces the transcription of BDNF and knockdown expression of

MeCP2 interferes with this event. These studies highlight a role for IKK

in accelerating the differentiation of human NPCs

and identify IKK

as a potential regulator of MeCP2 function and BDNF expression.

Citation:

Khoshnan A, Patterson PH (2012) Elevated IKK

Accelerates the Differentiation of Human Neuronal Progenitor Cells and Induces MeCP2-Dependent

BDNF Expression. PLoS ONE 7(7): e41794. doi:10.1371/journal.pone.0041794

Editor:

Nicoletta Landsberger, University of Insubria, Italy

Received

March 22, 2012;

Accepted

June 25, 2012;

Published

July 27, 2012

ß

2012 Khoshnan, Patterson. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits

unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding:

This work was supported by a grant (

#

2433) from The International Rett Syndrome Foundation. www.rettsyndrome.org. The funders had no role in

study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing Interests:

The authors have declared that no competing interests exist.

* E-mail: khoshnan@caltech.edu

Introduction

IKK

is a component of the IKK complex (

), which is an

important regulator of NF-

B pathway and plays a major role in

cell proliferation/differentiation and inflammation. IKK

also

regulates the production of active p52 NF-

B factor, which is

essential for the development of the immune system [1].

Moreover, IKK

has anti-inflammatory properties and can inhibit

the IKK

/NF-

B activity and lower the expression of inflamma-

tory cytokines [2,3]. Recent studies have identified several NF-

B-

independent functions for IKK

[4]. IKK

localizes to the nucleus

and phosphorylates proteins such as CREB binding protein (CBP),

the silencing mediator of retinoid and thyroid hormone receptor

(SMRT), forkhead box A2 (FOXA2), and

-catenin. All of these

proteins are expressed in the brain and are implicated in various

aspects of neurodevelopment [4–9]. In animal models, IKK

mediated phosphorylation of histone-3 (H3) and CBP contributes

to memory reconsolidation in the hippocampus [10]. Moreover,

IKK

phosphorylates the estrogen receptor and promotes

estrogen-regulated gene expression [11]. Estrogen is a neuroster-

oid that modulates dendritic growth and synaptogenesis in the

central nervous system [12]. Thus, IKK

may play a role in

neurodevelopment.

IKK

is constitutively active in human neuronal progenitor cells

(NPCs). Moreover, the level and activity of IKK

decreases in

neurons exposed to DNA damaging agents whereas elevation of

IKK

is neuroprotective and augments neuronal resiliency to

stress [13]. Here, we report that expression of an extra copy of

IKK

accelerates the differentiation and maturation of human

embryonic NPCs. Our data also identify IKK

as a modifier of

MeCP2, which is a prominent regulator of neuronal gene

expression [14]. Thus, manipulating the levels and activity of

IKK

may be a useful strategy to enhance neuronal differentiation

and regulate MeCP2 activity.

Results

Elevation of IKK

affects the proliferation and

differentiation of human neuronal progenitor cells

IKK

regulates the differentiation of several cell types including

epithelial and immune cells including monocytes, B cells, and

regulatory T cells [15–19]. Interestingly, the level of IKK

protein

is increased several fold during monocyte-to-macrophage differ-

entiation [18]. The focus of this study was to determine whether

elevation of IKK

alters the proliferation and/or the differenti-

ation of an embryonic human mesecephalic NPC line

(MESC2.10). Unlimited proliferation of MESC2.10 cells is

regulated by a tetracycline-regulated (tet-off) v-myc and the

addition of mitogenic factor, basic fibroblast growth factor-2

(bFGF-2). Upon shutting down the expression of v-myc by

doxycycline and removal of FGF-2, MESC2.10 NPCs can

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

differentiate into neurons expressing dopaminergic markers [20].

Expressing an extra copy of IKK

in MESC2.10 cells (IKK

)

(Fig. S1) has no visible effect on proliferation when v-myc is

expressed (data not shown). Employing a neurosphere assay,

which is used to study the self-renewal of neuronal stem cells

(NSCs) [21], we find that MESC2.10 cells proliferate for several

generations in the absence of v-myc. While IKK

cells also form

neurospheres, they are smaller in size and the numbers are

significantly reduced (Fig. 1A top panels and B). To extend these

findings, primary neurospheres were dissociated into single cell

suspensions and cultured in a second round in the presence of

FGF-2 and doxycycline. Although control NPCs form secondary

neurospheres, this property is completely lost in IKK

NPCs

(Fig. 1A. bottom panels). Thus, elevated IKK

interferes with the

self-renewal of MESC2.10 NPCs.

To examine whether the reduced proliferation of IKK

progenitors is due to precocious differentiation, we cultivated cells

on a laminin substrate in proliferating medium (

bFGF-2) with the

addition of doxycycline to repress v-myc expression, which blocks

neurosphere formation of the IKK

but not of the control NPCs

(Fig. 1A). Staining cells for the neuronal differentiation marker

tubulin III (Tuj-1), we do not find any Tuj-1-positive cells in either

control or IKK

NPCs when cells express v-myc (Fig. 1C top

panels). However, the majority of IKK

NPCs express Tuj-1 by

the 2

day after the addition of doxycycline. This is in contrast to

control NPCs, which continue to proliferate under these condi-

tions and

,

5% of the cells stain positively for Tuj-1 by the 2

and

,

45% by the 4

days (Fig. 1C, D). By the 4

day, IKK

NPCs

develop extensive neurite outgrowth (Fig. 1C lower right panel),

which is a hallmark of neuronal differentiation

in vitro

[22]. We also

plated day 6 dissociated neurospheres (Fig. 1A) on laminin and

examined for Tuj-1 after 24 h of further cultivation in the

presence of bFGF-2 and doxycycline. More than 95% of the

IKK

NPCs express Tuj-1 and develop prominent neurite

outgrowth. Under these conditions,

,

50% of the control cells also

stain positively for Tuj-1 but have no detectable neurite outgrowth

(Fig. 1E, F). Control and IKK

NPCs express Nestin, which is a

marker of proliferating NPCs (Fig. S2A). However, growth

conditions that promote the differentiation of IKK

NPCs

(Fig. 1C, E), do not significantly affect the level of Nestin. Nestin

accumulates in the neurites of dissociated day 6 IKK

neuro-

spheres whereas it is predominantly in the cytoplasm of control

cells (Fig. S2B). It is possible that the turnover and/or reduction of

Nestin expression requires a longer cultivation of IKK

NPCs.

To gain more insights in the role of elevated IKK

on NPCs

differentiation, control and IKK

cells were cultured on laminin-

coated dishes and induced to differentiate under conditions that

promote the generation of dopaminergic neurons [20]. The

majority of cells in differentiating control and IKK

are positive

for the neuronal markers Tuj-1 and MAP2 by the 4

day.

However,

,

50% of the control cells are weakly stained for the

expression of Tuj-1 and MAP2 (Fig. 2A, B). Using Western blot

analysis, we find that the level of Tuj-1 protein is

,

2.2 fold higher

in differentiating IKK

NPCs by the 2

and 4

days compared

to controls (Fig. 2C). Moreover, IKK

cells display elaborate

neurite outgrowth, which is minimal in control MESC2.10 NPCs

(Fig. 2A, B). The ability of IKK

to enhance neurite outgrowth

was further examined in a scratch lesion assay, which involves

removing cells manually and following growth into the open space

over time [23]. Differentiating IKK

NPCs generate extensive

neurite outgrowth two days after the lesion is formed whereas

outgrowth is much less in control NPCs (Fig. 2D, arrows).

Conditioned medium from differentiating IKK

NPCs has no

visible effect on the differentiation of the control cells (data not

shown), indicating that the affects of IKK

are likely cell

autonomous. However, we cannot rule out the possibility of a

low level of growth factors or labile molecules secreted by IKK

cells that may affect neurite outgrowth. Transient transfection of

embryonic rat cortical progenitor cells with IKK

also promotes

extensive neurite outgrowth, indicating that the pro-differentiating

properties of elevated IKK

are not limited to MESC2.10 human

NPCs (Supplementary Fig. 3).

While IKK

NPCs rapidly cease proliferation upon the

induction of differentiation, control cells undergo further divisions

as monitored microscopically (data not shown). To examine this

phenomenon in detail, we performed BrdU labeling, which is a

marker of DNA synthesis and cell proliferation [24]. We find that

,

50% of control NPCs incorporate BrdU at 4 days post-

differentiation. However, BrdU incorporation is minimal in

differentiating IKK

progenitors and

,

90% are post-mitotic

(Fig. 3A). Moreover,

,

40% of the 4

day differentiating control

NPCs express Ki-67, another marker of cell proliferation [25],

whereas less than 1% of IKK

cells stain positively for Ki-67 at

this time point (Fig. 3B). The majority of BrdU-positive cells stain

weakly for Tuj-1, indicating that are not fully committed to

differentiation (Figs. 3A, C). However, BrdU incorporation is

reduced dramatically upon further incubation and the majority of

control NPCs become Tuj-1 positive after 8

day in culture

(Fig. 3C). These findings are consistent with those in Figs, 1 and 2,

where elevated IKK

blocks the self-renewal of NPCs and

promotes the differentiation of MESC2.10 NPCs.

IKK

affects the REST/NRSF and miR-124 regulatory loop

We examined whether the levels or the cellular distribution of

endogenous IKK

is altered during the differentiation of control

NPCs. While the levels of IKK

do not change significantly, its

accumulation in the nuclear fraction increases in the 4

and 8

day cultures (Fig. 4A). It is relevant that levels of nuclear IKK

increase by the 2

day in differentiating IKK

NPCs (Fig. 4B,

middle panel). Thus, the rate of nuclear accumulation of IKK

may contribute to the onset of neuronal differentiation.

is a chromatin modifying kinase and is known to influence

gene expression by various means [4,10,11]. Since nuclear

accumulation of IKK

coincides with neuronal differentiation,

we hypothesized that IKK

may directly affect the expression of

key regulators of neurogenesis. One prominent modulator of

neuronal differentiation is REST/NRSF [26]. REST binds to a

consensus cis-element in the promoter of several hundred neuron-

specific genes and prevents their expression. The inhibitory

functions of REST are essential for the self-renewal of embryonic

as well as adult NSCs [27–29]. REST levels are dramatically

reduced during neuronal differentiation allowing the expression of

neurogenic proteins and non-coding RNAs [27,29]. Using reverse

transcription and real-time PCR (qRT-PCR), we find that, while

the levels of REST mRNA gradually decrease during the

differentiation of control NPCs, it is still detectable by 4 days. In

contrast, REST expression is rapidly reduced in differentiating

IKK

NPCs and is not detectable by the 2

day post-

differentiation (Fig. 4C). Western blot analysis of nuclear lysates

is consistent with the mRNA results in differentiating control and

IKK

NPCs (Fig. 4B, top panel). Thus, compared to control

NPCs, the levels of REST mRNA and protein drop more rapidly

in differentiating IKK

NPCs. REST promoters contain several

NF-

B binding sites [30]. Since IKK

regulates NF-

B [4], we

hypothesized that it may influence the binding of NF-

B to REST

promoter and affect REST expression directly. However, in gene

reporter assays with the REST promoter fused to luciferase [30],

elevated IKK

does not reduce REST promoter activity (data not

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

shown). Therefore, the effect of IKK

on reduced expression of

REST in differentiating IKK

NPCs appears to be indirect.

In NPCs REST represses miR-124 expression, which is pro-

neurogenic and has several REST binding sites in its promoter

[27]. MiR-124 is abundant in neurons and is a major determinant

of neuronal differentiation [27,31]. We hypothesized that IKK

may enhance the expression of miR-124, which may be the

underlying cause of REST reduction in differentiating IKK

NPCs. MiR-124 has several isoforms and miR-124a is well

characterized in the context of neuronal differentiation [27,31].

We find that miR-124a expression is induced in both control and

IKK

NPCs during differentiation. However, compared to

Figure 1. Effects of IKK

a

on the proliferation of MESC2.10 NPCs.

(

) Elevated IKK

impairs neurosphere formation of MESC2.10 NPCs. The

neurosphere assay was carried out as described in Methods. Representative micrographs of primary (1

u

) (top panels) and secondary (2

u

) neurospheres

(bottom panels) formed by control (C) and IKK

NPCs are shown. Assays were done in triplicate. (

) Quantification of neurospheres reveals a

significant deficit in the IKK

NPCs compared to controls. Six culture wells were counted in each condition and averaged. P value was obtained

using student’s t-test. (

) Elevated IKK

induces the differentiation of MESC2.10 cells when v-myc expression is repressed. Control and IKK

NPCs

were cultivated on laminin in the absence (time 0) or the presence of doxycycline for 2 or 4 days. Cells were stained for Tuj-1 expression.

Representative micrographs obtained with a confocal microscope are shown. The DNA stain TOTO-3 was used to identify nuclei. (

) The % of Tuj-1

positive cells in day 2 and day 4 cultures is shown. (

) IKK

neurospheres undergo spontaneous neuronal differentiation. Day 6 dissociated

neurospheres were plated on laminin and stained for Tuj-1 after 24 h. Representative micrographs obtained with a confocal microscope are shown.

(

) The % of Tuj-1 positive in day 2 and day 4 cells is shown. For D and F, total and Tuj-1 positive cells were counted in 6 different confocal images

and the % positive was calculated. P values were obtained using student’s t-test.

doi:10.1371/journal.pone.0041794.g001

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

control, primary as well as the mature miR-124a transcripts are

several fold higher in the IKK

cells (Fig. 4D, E, respectively).

miR-124a levels are inversely related to those of REST in control

and IKK

NPCs (Fig. 4B, C). Using a gene reporter assay with

the miR124a promoter fused to luciferase [27], elevated IKK

does not induce miR-124a promoter activity (data not shown).

Taken together, these findings indicate that the levels of REST

and miR-124a, which are critical determinants of neuronal

differentiation, are significantly altered in IKK

NPCs compared

to controls. However, our studies do not support a direct link

between elevated IKK

and the expression of REST and miR-

124.

We further examined differentiating NPCs for expression of

other neuron-enriched miRNAs. Interestingly, miR-7, which

promotes neurite outgrowth and is co-expressed with miR-124

in other cell models [32], is selectively induced in differentiating

IKK

NPCs (Fig. S4). The induction of miR-7 may contribute to

the extensive neurite outgrowth observed in differentiating IKK

NPCs (Figs. 2A, B; Fig. S3). Expression of other miRNAs such as

miR-132 and -133a, and -155 is not significantly altered in either

differentiating control or IKK

NPCs (data not shown).

Elevated IKK

promotes neuronal maturation

The positive effects of IKK

on neuronal differentiation raised

the question of whether it also influences neuronal maturation.

One hallmark of maturing neurons is the accumulation of MeCP2,

which regulates many aspects of neurodevelopment, and loss of

MeCP2 function is implicated in the brain disorder, Rett

Figure 2. Elevated IKK

a

promotes the differentiation of MESC2.10 cells.

(

A, B

) IKK

promotes neurite outgrowth in differentiating NPCs.

Control (C) and IKK

NPCs were differentiated on coverslips for 4 days, fixed and stained with neuronal differentiation markers Tuj-1 (A) and MAP-2

(B). Representative micrographs obtained with a confocal microscope are shown. The DNA stain TOTO-3 was used to identify nuclei. (

) Tuj-1 levels

are elevated in differentiating IKK

NPCs. Representative western blot results are shown for cytoplasmic lysates staining for Tuj-1 levels at different

time points during the differentiation of control and IKK

NPCs [diff.(days)]. IKK

was used as a loading control. Fold-change was obtained by

dividing the intensity of Tuj-1 to the corresponding IKK

, obtained by a Fluorchem 8900 (Alpha Innotech, San Leandro, CA). (

) The scratch assay

shows extensive neurite outgrowth in differentiating IKK

NPCs. Cultures on the 2

day of differentiation were wounded by a micropipette tip and

further incubated for additional two days. Cells were fixed and stained as above. Arrows point to the areas of neurite extension.

doi:10.1371/journal.pone.0041794.g002

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

syndrome [33]. We find that MeCP2 is expressed at a low level in

control NPCs, however it is more abundant (

,

6 fold higher) in

IKK

NPCs differentiated for 8 days (Figs. 5A and 6B, top

panels). MeCP2 transcription is not altered during the differen-

tiation of control and IKK

NPCs, indicating that other post-

transcriptional regulatory pathways affect MeCP2 levels (data not

shown). Expression of SCG10/Stathmin-2, a neuron-specific

microtubule destabilizing protein that promotes neurite outgrowth

and neuronal migration [34], and pre-and post-synaptic markers

such as syntaxin1 and PSD95 [35,36] is also induced in IKK

cells (Fig. 5A, panels 2–4, lanes 5–8).

Since MeCP2 is important in neurodevelopment [33], we asked

whether MeCP2 induction contributes to the differentiation of

IKK

NPCs. Towards this end, we reduced the expression of

MeCP2 in IKK

cells using a lentivirus encoding an shRNA

targeting MeCP2 (Fig. 5A, top panel, lanes 9–12). This cell line is

labeled as MeCP2 knockdown (MeCP2KD). The levels of MeCP2

in the MeCP2KD line are comparable to those detected in

differentiating control NPCs (Fig. 5A, top panel; compare lanes 1–

4 with 9–12). Knockdown of MeCP2 does not affect neuronal

differentiation since the levels of REST and miR-124a are similar

to those expressed in IKK

cells (data not shown). Moreover,

reduction of MeCP2 has no visible effect on IKK

-induced

neurite outgrowth, as observed by Tuj-1 staining (Fig. 5B).

However, IKK

-induced accumulation of PSD95 in 8

day

cultures is significantly reduced (2.4-fold) when MeCP2 expression

is silenced (Fig. 5A, panel 4, compare lanes 8 and 12). This is

consistent with previous findings in animal models where the

absence of functional MeCP2 negatively affects PSD95 levels [37].

IKK

induces MeCP2-dependent BDNF expression

MeCP2 binds to methylated CpG dinucleotides, which are

abundant in the promoters of many genes [38]. MeCP2 is also

implicated in the expression of many neuronal genes, including

BDNF, whose expression is influenced by the level as well as post-

translational modifications of MeCP2 [39,40]. We asked if

elevated MeCP2 in IKK

neurons could also affect BDNF

levels. BDNF expression can be initiated from 9 different

promoters [41] and the exon-IV promoter of human BDNF (rat

promoter III) is a well-known target of MeCP2 [39,40]. Using

Figure 3. IKK

a

inhibits the proliferation of early differentiating NPCs.

(

) Elevated IKK

accelerates NPCs commitment to differentiation.

BrdU was added on the 4

day of differentiation. 24 h later cells were fixed and stained with a rat anti-BrdU antibody (green) and the neuron-specific

marker Tuj-1. Pictures were taken with a confocal microscope. (

) As a further test of proliferation, 4

day cultures were stained with the Ki-67

antibody, which identifies proliferating cells. (

) A time-course (days) of BrdU incorporation reveals the difference in rate of decline in proliferation

between the control and the IKK

NPCs. Each time point represents 24 h of BrdU incorporation. Samples were processed as in A.

doi:10.1371/journal.pone.0041794.g003

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

qRT-PCR, we find that differentiating control NPCs do not

express high levels of BDNF (Fig. 5C). However, BDNF

transcription from exon-IV is significantly induced in IKK

differentiating NPCs and is further elevated by KCl-mediated

depolarization of 8

day IKK

neurons (Figs. 5C, D).

Knockdown of MeCP2 levels reduces both basal as well as KCl-

induced BDNF expression by

,

50% (Fig. 5C, D). Thus, IKK

promotes BDNF transcription, which is in part MeCP2-depen-

dent.

IKK

is recruited to several different promoters including NF-

B and estrogen-regulated genes [4,11]. Since BDNF levels are

elevated in IKK

neurons, we asked whether IKK

associates

with regulatory regions of the exon-IV BDNF promoter. Using

chromatin immunoprecipitation (ChIP), we find that IKK

Figure 4. IKK

a

regulates REST and miR-124a expression.

(

) IKK

accumulates in the nuclei of differentiating MESC2.10 NPCs. Representative

Western blot results for levels of endogenous IKK

in the cytoplasm (C) and nuclear (N) fractions of differentiating NPCs (top panel) are shown. IKK

was detected with a mouse anti-IKK

antibody. Nuclear LaminB1 and cytoplasmic tubulin were used as loading controls (middle and bottom panels,

respectively). (

) REST protein levels also decline faster in differentiating IKK

NPCs compared to differentiating controls. Representative western

blot results are shown from nuclear lysates for REST (top panel), IKK

(middle panel) and laminB1 (bottom panel). REST was detected with a mouse

anti-REST antibody and Anti-Flag antibody was used to detect IKK

. LaminB1 was used a as loading control. (

) After initiating differentiation, REST

mRNA levels decline faster in IKK

NPCs than in control cells. Taqman probes were used to quantify the mRNA levels at the days shown. The data are

shown relative to the level in proliferating control NPCs. GAPDH mRNA was used for normalization. Triplicate samples were averaged for each point,

and the SEMs indicated. N.D., not detected. (

D, E

) The accumulation of primary (pri-miRNA) and mature miRNA-124a are shown in D and E,

respectively. Taqman probes were used for the qPCR. Pri-miRNA was normalized to GAPDH mRNA and mature miRNA was normalized to the small

RNA, RNU6. The data are shown relative to the levels in proliferating control NPCs. P values were obtained using student’s t-test.

doi:10.1371/journal.pone.0041794.g004

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

enriched at the BDNF promoter (Fig. 6A). Moreover, CREB and

MeCP2, which bind to this element [42,43], are also abundant

(Fig. S5A, B, left panels). As expected, MeCP2 binding to the

exon-IV BDNF promoter is reduced in MeCP2KD neurons (Fig.

S5B, right panel), which coincides with the reduction of BDNF

expression (Fig. 5C, D) and suggests that the concentration of

MeCP2 may be important for the regulation of the exon-IV

BDNF promoter. Interestingly, the association of IKK

and

CREB with the BDNF promoter is not altered by knockdown of

MeCP2 levels, indicating that they may bind independently of

MeCP2 (Figs. 6A and S5A, right panels). However, we cannot rule

out the possibility that residual MeCP2, which is bound to the

exon-IV promoter in MeCP2KD neurons, may be sufficient to

recruit IKK

and CREB (Fig. S5B, right panel). Overall, these

findings support a role for IKK

in the regulation of MeCP2-

dependent BDNF expression.

Phosphorylation of MeCP2 at Ser421 has previously been

implicated in the induction of BDNF expression [39]. Using an

antibody recognizing phospho-Ser 421, we find that phosphory-

lated MeCP2 accumulates in 8

day differentiated IKK

but not

control neurons (Fig. 6B, middle panel). This time course coincides

with the elevated levels of BDNF in IKK

neurons (Fig. 5C, D).

The fact that IKK

is a kinase raised the question of whether

IKK

associates with and phosphorylates MeCP2. IKK

co-

localizes with MeCP2 in the nuclei of IKK

neurons (Fig. 6C).

Moreover, complexes containing both IKK

and MeCP2 can be

immunoprecipitated from the nuclear fraction of 8

day post-

differentiation IKK

neurons (Fig. 6B, D). Therefore, we

performed

in vitro

kinase assays using recombinant IKK

and

MeCP2 proteins. We find that IKK

phosphorylates MeCP2

(Fig. 6E). However, mass spectrometric analysis identifies phos-

phorylated Ser residues other than Ser421 (A. Khoshnan, et al.,

unpublished data). Previous studies have identified CAMK-II and

CAMK-IV as potential kinases phosphorylating Ser421 of MeCP2

[39,44]. Thus, phosphorylation of Ser421 in IKK

neurons may

be an indirect effect of IKK

. The characterization of IKK

Figure 5. IKK

a

promotes expression of markers of mature neurons and BDNF.

(

) Elevated IKK

enhances neuronal maturation. Cell

lysates from various time points (days after inducing differentiation) were examined by western blotting for the levels MeCP2, SCG10, syntaxin, and

PSD-95. A lentivirus encoding an shRNA targeting MeCP2 was used to knockdown the expression of MeCP2 in IKK

NPCs (labeled as MeCP2KD)

(lanes 9–12). IKK

was used as a loading control. A non-specific band (N.S., below the authentic band) is recognized by the anti-PSD95 antibody. (

)

Knockdown of MeCP2 expression does not influence IKK

-induced neuronal differentiation. NPCs were differentiated and examined as described in

Fig. 2A. Representative confocal micrographs of 4

day differentiating cultures are shown. (

) IKK

promotes MeCP2-dependent BDNF expression. A

time course (days) for BDNF expression during the differentiation of NPCs is shown. Taqman probes were used to quantify mRNA generated from the

exon-IV of the BDNF promoter. GAPDH was used for normalization. The data are shown relative to the level in proliferating control (day 0) NPCs. (

)

IKK

also promotes MeCP2-dependent BDNF expression following depolarization. Data are shown for cultures after 8

days of differentiation and

depolarization with 50 mM KCl for 6 hr immediately preceding harvest. The data are shown relative to the level in non-depolarized control cells.

Assays were done in triplicate and P values were obtained using student’s t-test.

doi:10.1371/journal.pone.0041794.g005

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

mediated phosphorylation of MeCP2 at Ser421 and other residues

and their effects on the activity of MeCP2 is a topic of current

work in our laboratory.

Discussion

We have identified novel functions for IKK

in enhancing the

differentiation of human NPCs. Elevated IKK

indirectly lowers

the level of REST/NRSF repressor, which is a global inhibitor of

neurogenesis [26–29]. The ability of IKK

to enhance neuronal

differentiation is further exemplified by the induction of neuron-

enriched miRNAs such as miR-124a and -7, and proteins

including MeCP2, PSD95, and BDNF, which are involved in

neurite outgrowth, neuronal maturation, and synaptic plasticity.

Thus, increasing the level and/or the activity of IKK

may be a

useful strategy to promote neuronal differentiation

in vitro

and

potentially

in vivo

. Our results also highlight a direct link between

IKK

and MeCP2, which could be instrumental in regulating

MeCP2-dependent gene expression and neurodevelopment.

Elevation of IKK

inhibits self-renewal and accelerates the

differentiation of MESC2.10 NPCs, and reduction of REST

expression may play a role. As a repressor of neuronal genes,

REST promotes the proliferation of NSCs as well as neuroblas-

toma cell lines, whereas reduction of REST induces neuronal

differentiation [26–29,45]. We propose that the effect of IKK

REST expression is indirect, since elevated IKK

does not lower

the REST promoter activity. However, REST promoters have

several NF-

B binding sites [30] and the regulation of NF-

activity by IKK

may influence REST levels under certain

physiological conditions. We have not been successful in estab-

lishing a link between IKK

/NF-

B and REST transcription,

however.

A critical step in the initiation of NSC differentiation is the

induction of miR-124, which is repressed by REST [27]. miR-124

is enriched in the brain and is recognized as the ‘‘micromanager of

neurogenesis’’

in vivo

[46,47]. Indeed, miR-124 promotes the direct

conversion of human fibroblasts into functional neurons, where it

instructs chromatin remodeling and promotes brain-specific

alternative splicing of mRNAs essential for neuronal differentia-

tion [48–50]. Thus, the reduced levels of REST and reciprocal

elevation of miR-124 in IKK

cells will likely cause global

changes in gene expression that inhibit proliferation and engage

the differentiation programming (Fig. 4). In addition, miR-124

plays an important role in synaptic plasticity and memory

formation in post-mitotic neurons in Aplysia [51].

In vivo

studies

indicate that IKK

is involved in hippocampal-dependent

memory reconsolidation [10]. It will be interesting to examine

Figure 6. IKK

a

associates with MeCP2 and is recruited to the exon-IV BDNF promoter.

(

) Flag-tagged IKK

is recruited to the exon-IV

BDNF promoter. ChIP assays were used to immunoprecipitate IKK

/DNA complexes (using anti-Flag antibody) from differentiated IKK

and

MeCP2KD neurons (day 8). The left panel is ChIP from IKK

and the right panel is from MeCP2KD neurons. Non-reactive IgGs were used controls.

DNA was amplified by PCR. Products were visualized by agarose gel-electrophoresis and ethidium bromide staining. (

) Western blots were used to

assay nuclear lysates for phosphorylation of MeCP2 at Ser421 in 8

day differentiated IKK

neurons (middle panel). The top panel shows the total

levels of MeCP2 during differentiation (0–8 days). LaminB1 was used as a loading control. (

) IKK

and MeCP2 co-localize in the nuclei of IKK

neurons. IKK

NPCs were differentiated for 6 days and stained with MeCP2 antibody (green) and an anti-Flag antibody detecting IKK

(red).

Representative micrographs obtained with a confocal microscope are shown. (

) MeCP2 co-immunoprecipitates with IKK

. Nuclear lysates from 8

day differentiated IKK

neurons were immunoprecipitated with anti-Flag beads (for IKK

) and examined for the presence of MeCP2 with an anti-

MeCP2 antibody. A non-immune mouse antibody (C-Ab) was used as a negative control for immunoprecipitation. A strong band for anti-IgG staining

is also seen. (

) IKK

phosphorylates MeCP2. Active recombinant IKK

was tested for the ability to phosphorylate MeCP2. The kinase assay was

performed as described in Methods with recombinant MeCP2 or GST as substrates. Products were visualized by SDS-PAGE followed by

autoradiography.

doi:10.1371/journal.pone.0041794.g006

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794

whether elevated expression of IKK

induces miR-124 and

enhances memory formation and learning, possibly by affecting

neurogenesis in the adult hippocampus.

IKK

accumulates in the nuclei of differentiating NPCs

(Figs. 4A, B, and 6C), and nuclear transfer of IKK

is implicated

in the phosphorylation of histone-3 (Ser10), which leads to

enhanced expression of various genes [4,10]. Our transcriptome

analysis (mRNA–seq) of differentiating control and IKK

NPCs

reveals significant changes in the expression of several hundred

mRNAs in IKK

cells; some of these encode proteins involved in

neurodevelopment and the splicing of neuron-specific mRNAs (A.

Khoshnan et al., unpublished data). Characterization of some of

these genes may shed further light on the mechanism of how

IKK

accelerates neuronal differentiation and regulates complex

epigenetic changes such as neurite outgrowth. It is intriguing that

miR-7, which is implicated in neuronal homeostasis and neurite

outgrowth [32], is selectively induced in differentiating IKK

NPCs. miR-7 also protects dopaminergic neurons against oxida-

tive stress, where it reduces the expression of

-synuclein and leads

to enhanced survival [52]. We have previously shown that IKK

protects MESC2.10 neurons against oxidative stress-induced

neuronal death and preserves the integrity of neuron-enriched

huntingtin protein, which has neuroprotective properties [13].

Thus, in addition to promoting neurite outgrowth, IKK

-induced

miR-7 may also contribute to the resiliency of neurons under

adverse environmental conditions.

The ability of IKK

to regulate MeCP2 levels and activity is

another novel aspect of this study. These interactions were

characterized in the context of BDNF expression, which is induced

by elevated IKK

and suppressed when MeCP2 levels are

knocked down (Fig. 5). BDNF plays a critical role in neuronal

differentiation and survival, miRNA processing, and synaptic

plasticity [53,54]. The MeCP2-dependent induction of BDNF

may therefore be important in these processes, which has

implications for neurological and psychiatric disorders. While

earlier studies supported an inhibitory role for MeCP2, recent

findings are consistent with a positive effect of MeCP2 on BDNF

expression [39,40,42,43]. Moreover, in animal models where

MeCP2 is inactive or deleted, BDNF levels are significantly

reduced [55,56]. Our data are also consistent with a positive effect

of elevated MeCP2 on BDNF and highlight the involvement of

IKK

Recent studies propose that MeCP2 may function both as a

repressor and activator of the same target genes, depending on its

association with other proteins. For example, MeCP2-dependent

recruitment of HDAC2 or CREB to the glial-derived neurotrophic

factor promoter can inhibit or promote gene expression,

respectively [57]. We find that IKK

associates with MeCP2

and both are recruited to the BDNF exon-IV promoter, which

may be crucial for the induction of BDNF. Thus, similar to CREB,

binding of IKK

to MeCP2 may enhance MeCP2-dependent

gene expression. Moreover, maximal BDNF expression in IKK

neurons coincides with elevated levels of MeCP2 (Fig. 5). We posit

that changes in the homeostasis of MeCP2 may dictate whether it

acts as repressor or activator of gene expression. At steady state,

MeCP2 may simply function as a chromatin organizer and control

the noise in global gene expression [38]. On the other hand, when

MeCP2 levels are elevated, it may facilitate selective gene

expression by associating by other regulatory proteins such as

IKK

and CREB. It is relevant that elevation of MeCP2 in

transgenic mice induces the expression of

,

2200 genes including

CREB [14]. Moreover, the levels of MeCP2 and its phosphory-

lation at Ser421 are increased by exogenous factors such as

amphetamine, cocaine, and the anti-depressant fluoxetine [58,59].

These findings support the dynamic nature of MeCP2 expression

in neurons and how fluctuations in its levels and/or its

phosphorylation may dictate various functions. Exogenous stimuli

including growth factors and cytokines also regulate IKK

activity

[1–4]. The elevation of MeCP2 in IKK

neurons and the

phosphorylation of MeCP2 by IKK

raise the possibility that

environmental activation of IKK

may affect MeCP2 homeostasis

and activity. Further characterization of IKK

-MeCP2 interac-

tions may shed light on the complex nature of MeCP2 activities in

neurons.

Materials and Methods

Antibodies and reagents

Anti-Tuj-1 antibody was obtained from Covance (Berkeley,

CA). Anti-Nestin antibody was purchased from R&D systems

(Minneapolis, MN). Anti-syntaxin and anti-Flag antibodies were

obtained from Sigma (St Luis, MO). Anti-NRSF/REST and anti-

PSD95 antibodies were provided by David Anderson and Mary

Kennedy at the California Institute of Technology, respectively.

Anti-IKK

was purchased from BD Biosciences (San Diego, CA).

Anti-MeCP2 antibody and recombinant active IKK

were

purchased from Millipore (Temecula, CA). Anti-phospho MeCP2

(Ser421) antibody was provided by Dr. Michael Greenberg at

Harvard medical school. Anti- IKK

and anti-MAP2 antibodies

were obtained from Cell Signaling Technology (Danvers, MA).

Anti-SCG10 antibody was produced in house. Rat anti-BrdU and

ki67 rabbit antibodies were purchased from abcam (Cambridge,

MA). Anti-laminB1 antibody, DMEM/F12, bFGF-2, N-2 and B-

27 media supplements were obtained from Invitrogen (Carlsbad,

CA). Anti-CREB antibody was purchased from Santa Cruz

Biotechnology (Santa Cruz, CA). Cell fractionation and ECL

detection kits including HRP-conjugated secondary antibodies

were from PIERCE Biotechnology (Rockford, IL). Recombinant

MeCP2 protein was purchased from Panomic (Santa Clara, CA).

Generation of MESC2.10 human neurons

The generation of MESC2.10 human NPCs has previously

been reported [20]. Briefly, NPCs were obtained from an 8 week-

old human embryo and transduced with a retrovirus encoding a

tetracycline-regulated (tet-off) v-myc to promote proliferation [20].

Propagation is in serum-free medium containing bFGF-2 [13,20].

In our studies, NPCs were propagated in dishes coated with poly-

lysine and laminin in DMEM/F12 in the presence N2 and B-27

neuronal supplements and 20 ng/ml bFGF-2 (Invitrogen). When

indicated, doxycycline (2

g/ml) was added to proliferating

medium to block the expression of v-myc. To differentiate

MESC2.10 NPCs, proliferation medium was replaced DMEM/

F12 containing N2 supplement, 2

g/ml of doxycycline and

M cAMP.

Neurosphere assays

Single-cell suspensions of MESC2.10 and IKK

NPCs (5,000

cells/ml) were cultured in DMEM/F12, containing bFGF (40 ng/

ml) and B27 supplements in the presence of 2

g/ml of

doxycycline for 6 days. For secondary neurosphere assays, primary

neurospheres were dissociated using trypL (Invitrogen) and

cultured as above. Fresh medium containing FGF-2 and

doxycycline was added every two days. Cultures were examined

using a phase contrast microscope. The number of spheres in 6

wells were counted and averaged for each condition. To stain

neurospheres for markers of differentiation, 6 days old cultures

(after the first round) were dissociated and plated on laminin

substrates for 24 h as above.

IKK

and Neuronal Differentiation

PLoS ONE | www.plosone.org

July 2012 | Volume 7 | Issue 7 | e41794