PMA activates Stat3 in the Jak/Stat pathway and induces SOCS5 in rat brain astrocytes.

Abstract

Suppressors of cytokine signaling (SOCS) family members are negative feedback regulators of the Jak/Stat pathway, which is an essential inflammatory signaling pathway. We investigated expression of eight members of the SOCS family in rat astrocytes, using two inflammatory stimulants, PMA and IFN-gamma. Only a few SOCS genes were induced by both stimulants, and we detected an increase in SOCS5 protein with PMA. PMA activated the Jnk, Erk, p38, and Jak/Stat signal pathways. In addition, it increased the level of activated-Stat3 resulting from tyrosine phosphorylation. A gel-shift assay showed that a protein in nuclear extracts from PMA-treated cells was able to bind to Stat binding elements. These results suggest that activated Stat3 binds to SOCS promoters and leads to their transcriptional induction.

Full text

Page 1

Mol. Cells, Vol. 23, No. 1, pp. 94-99






PMA Activates Stat3 in the Jak/Stat Pathway and Induces SOCS5
in Rat Brain Astrocytes

Mi-Na Hwang, Kwang Soo Kim1, Yo-Woo Choi, Ilo Jou1, and Sungpil Yoon*
Research Institute, National Cancer Center, Goyang 411-764, Korea;
1 Department of Pharmacology, Ajou University School of Medicine, Suwon 442-721, Korea.

(Received August 24, 2006; Accepted December 7, 2006)

Suppressors of cytokine signaling (SOCS) family mem-
bers are negative feedback regulators of the Jak/Stat
pathway, which is an essential inflammatory signaling
pathway. We investigated expression of eight members
of the SOCS family in rat astrocytes, using two inflam-
matory stimulants, PMA and IFN-γ. Only a few SOCS
genes were induced by both stimulants, and we detected
an increase in SOCS5 protein with PMA. PMA acti-
vated the Jnk, Erk, p38, and Jak/Stat signal pathways.
In addition, it increased the level of activated-Stat3 re-
sulting from tyrosine phosphorylation. A gel-shift assay
showed that a protein in nuclear extracts from PMA-
treated cells was able to bind to Stat binding elements.
These results suggest that activated Stat3 binds to
SOCS promoters and leads to their transcriptional in-
duction.

Keywords: Astrocyte; IFN-γ; Jak/Stat; PMA; SOCS; Stat3.


Introduction

Glial cells are the most abundant cells in the brain, and
participate in the immune response to brain injury. Astro-
cytes, a component of the glial cells, are activated and
restore the integrity of brain tissue in response to patho-
logical stimuli or extracellular signals; this process results
in a variety of inflammatory effects. Since long-term in-
flammation can cause damage to the brain and neurode-
generation, activated glial cells can both benefit and harm
brain activity (Hwang et al., 2005; Ridet et al., 1997;
Streit et al., 1999). Therefore, it is important to under-
stand which genes are up- or down-regulated and how
long they are expressed in activated glial cells, as well as

* To whom correspondence should be addressed.
Tel: 82-31-920-2461; Fax: 82-31-920-2002
E-mail: yoons@ncc.re.kr
to identify the transcription factors and/or signals that
control their expression.
Suppressors of cytokine signaling (SOCS) family mem-
bers are negative feedback regulators of the Jak/Stat path-
way, which is an essential inflammatory signaling path-
way (Igaz et al., 2001; Kishimoto et al., 1994). Phos-
phorylation of Jak leads to the phosphorylation of “signal
transducers and activators of transcription” (Stats). The
phosphorylated Stats are released from the receptor com-
plexes, form dimers, and translocate to the nucleus where
they bind to promoters in target genes and regulate tran-
scription including that of SOCS genes (Igaz et al., 2001;
Kishimoto et al., 1994; Wen et al., 1995). The SOCS fam-
ily consists of at least eight members, namely cytokine-
induced SH2 protein (CIS) and SOCS1-7. SOCS proteins
contain a central SH2 domain, a conserved C-terminus
referred to as the SOCS box, and unique N-termini (Hil-
ton et al., 1998). Once SOCS proteins are made they in-
hibit Jak/Stat signaling by repressing Jak activity or the
binding of Stats to the cytoplasmic domains of receptors
(Endo et al., 1997). Thus, SOCS members play an active
positive role in anti-inflammation of the brain.
SOCS family proteins are induced via Jak/Stat pathways
by cytokines, growth factors and several immunomodula-
tors (Starr et al., 1999). Interferon-γ (IFN-γ), a prominent
pro-inflammatory cytokine, induces expression of SOCS
family proteins as well as expression of inflammatory
mediators (Frohman et al., 1989; Vanguri, 1995). The
signal transduction pathways of IFN-γ in the brain involve
the expression of SOCS via Jak/Stat pathways (Darnell et
al., 1994). Phorbol 12-myristate 13-acetate (PMA) acti-
vates protein kinase C (PKC) and causes inflammatory

Abbreviations: EMSA, electrophoretic mobility shift assay;
IFN-γ, interferon-γ; PKC, protein kinase C; PMA, phorbol 12-
myristate 13-acetate; RT; reverse transcription; SOCS, suppres-
sors of cytokine signaling; Stat, signal transducer and activator
of transcription.
Molecules
and
Cells
©KSMCB 2007
Communication

Page 2

Mi-Na Hwang et al. 95

activation in the brain (Arai et al., 2003; Vitkovic et al.,
2005), and there is relationship between PKC activation
and the Jak/Stat signaling pathways (Aeder et al., 2004;
Cohen et al., 2005).
In this study we investigated the expression of eight
members of the SOCS family in rat astrocytes in response
to two inflammatory stimulants, PMA and IFN-γ. We also
studied which signaling pathway was activated or respon-
sible for the SOCS expression. We detected the induction
of a number of the SOCS genes and an increase in SOCS5
protein after PMA treatment. We found that the Jnk, Erk,
p38, Jak signal pathways were activated by PMA and that
PMA treatment led to the phosphorylation of tyrosine resi-
dues only on Stat3, not on other Stats. A gel-shift assay
showed that a protein in nuclear extracts from PMA-treated
cells was able to bind to Stat binding elements. These re-
sults suggest that Stat3 activation may be sufficient for
binding to SOCS promoters and for inducing SOCS tran-
scription. Therefore, we suggest that PMA-induced SOCS
expression is activated by Stat3.


Materials and Methods

Reagents Phorbol 12-myristate 13-acetate (PMA) was pur-
chased from Sigma (USA). Interferon-γ (IFN-γ) was from
Peprotech (Rocky Mountain, NJ). Antibodies against the phos-
phorylated forms of Stat1 (tyrosine-), Stat3 (tyrosine- and ser-
ine-), Stat5, and Stat6 were from Cell Signaling Technology
(USA). Antibodies against the phosphorylated forms of Jnk1,
Jak1, and c-Src were from Biosource (USA). Antibodies against
phosphorylated-p38, p38, phosphorylated-Erk1/2, Erk1/2, Stat5,
Jak1, and Jnk1 were from Santa Cruz Biotechnology (USA).
Antibodies against SOCS5 and phospho-serine-Stat1 were from
Abcam (UK). Anti-SOCS1 antibody was from Zymed (USA),
and antibodies against Stat1, Stat3, Stat6, and GAPDH were
from Stressgen (Canada).

Cell culture Primary astrocytes were cultured from 1- to 3-day-
old Sprague-Dawley rats, as described previously (Zhou et al.,
1998). Briefly, the cortices were triturated into single cells in
MEM containing 10% FBS and plated into 75 cm2 T-flasks (0.5
gemisphere/flask) for 2−3 weeks. Microglia were detached from
the flasks by mild shaking and the astrocytes remaining in the
flask were harvested with 0.1% trypsin, plated in dishes and
cultured in MEM supplemented with 10% FBS.

RT-PCR Total RNA was extracted using RNAzol B (TELTEST,
USA). First-stranded cDNA was synthesized from 2 μg of total
RNA using a reverse transcription kit (Promega, USA). To am-
plify double-stranded cDNA, PCR was performed in a 25 μl
reaction mixture containing 5 μl of the reverse transcribed cDNA.
The sequences of the PCR primers used are described in Table 1.
The GAPDH gene was amplified to normalize the results. After
the PCR reactions, 10 μl samples were electrophoresed on 1.5%
Table 1. Primer sequences for RT-PCR.
Name Sequences
INOS Yang et al., 2004

F-TCACTGGGACAGCACAGAAT
R-TGTGTCTGCAGATGTGCTGA
CIS

F-CGACTCTAGCTTCCGACTGG
R-TCAGAGCTGGAACGGGTACT
SOCS1 Choi et al., 2005

F-ACACTCACTTCCGCACCTTC
R-AGCAGCTCGAAAAGGCAGTC
SOCS2

F-AAGACGTCAGCTGGACCGAC
R-TCTTGTTGGTAAAGGCAGTCCC
SOCS3 Choi et al., 2005

F-ACCAGCGCCACTTCTTCACG
R-GTGGAGCATCATACTGATCC
SOCS4

F-TGTTCGTCCATTGAGTTGGA
R-GAAGCACTGTTGGCAGTTAT
SOCS5

F-GCCTTACAGCTGGGACTGAG
R-GGCTTTGACTGCTTGCTGTA
SOCS6

F-TCTTCTCCCATGGAGGTGTC
R-TACCACCCCTGTTTTGCAAG
SOCS7

F-GAGCCTCAGCTTTCGATCAC
R-TTGGAAATGAGCTGCGCTTC
GAPDH Choi et al., 2005

F-TCCCTCAAGATTGTCAGCAA
R-AGATCCACAACGGATACATT


agarose gels.

Western blot analysis For Western blot analysis, total proteins
were extracted by the TCA method (Reid and Schatz, 1982).
Briefly, cells in 60 mm dishes were washed three times with 5
ml PBS buffer. To each plate was added 500 μl of 20% TCA
and the cells were dislodged by scraping and transferred to Ep-
pendorf tubes. Proteins were pelleted by centrifugation for 5
min at 3000 rpm and resuspended in Tris-HCl (pH 8.0) buffer.
Protein concentrations were estimated by the dye-binding
method. The proteins were dissolved in SDS-PAGE loading
buffer (Invitrogen), and subjected to Western blot analysis.

Electrophoretic mobility shift assay (EMSA) The crude nu-
clear proteins in the cell extracts were assayed by EMSA as
described previously (Min et al., 2003). Briefly, 20 μl mixtures
containing 2 μg of nuclear extract in a reaction buffer consisting
of 8.5 mM EDTA, 8.5 mM EGTA, 8% glycerol, 0.1 mM ZnSO4,
50 μg/ml poly (dI-dC), 1 mM DTT, 0.3 mg/ml bovine serum
albumin (BSA), 6 mM MgCl2, and γ-32P-radio-labeled oligonu-
cleotide probe (3 × 104 cpm), with or without a 20-fold excess
of unlabeled probe were incubated for 30 min on ice. DNA-
protein complexes were separated on 6% polyacrylamide gels in
Tris/glycine buffer, and the dried gels were exposed to X-ray
film. The following double-stranded oligonucleotides were used
in these studies; proximal Stat binding element, 5′-CAGTTCC-
AGGAATCGGGGGGC-3′, 21bp (Auernhammer et al., 1999;
Ehlting et al., 2005); GAS/ISRE, 5′-AAGTACTTTCAGTTT-
CATATTACTCTA-3′, 27 bp (Santa Cruz Biotechnology, sc-

Page 3

96 Induction of SOCS Members in Astrocytes
















Fig. 1. PMA and IFN-γ induce members of the SOCS family in
a short period of time. Primary astrocytes were serum-starved
for 24 h and stimulated with 1 μM PMA or IFN-γ (10 U/ml) for
the indicated times. Total RNA was isolated and RT-PCR was
performed to measure the mRNA levels of iNOS, SOCS mem-
bers, and GAPDH. The primers for PCR are shown in Table 1.


2537). The 5′-end labeled probes were prepared with 40 μCi (γ-
32P) ATP, using T4 polynucleotide kinase (Promega, USA) and
purified on Sephadex G-25 Quick Spin Columns (Roche Mo-
lecular Biochemicals).


Results

Inflammatory stimulation by PMA and IFN-γ induces
certain members of the SOCS family PMA and IFN-γ
are inflammatory stimulants (Arai et al., 2003; Darnell et
al., 1994; Vitkovic et al., 2005). We observed that both
stimulants induced iNOS expression, a representative
product of inflammation in immune cells. Expression of
iNOS peaked at three hours and continued for eight hours
in response to both treatments (Fig. 1). To examine the
induction of all the members of the SOCS family we first
performed a computer search for cDNA sequences of rat
SOCS genes in PubMed but could not identify sequences
for SOCS4, SOCS5, SOC6. Using mouse SOCS cDNA
sequences we identified the corresponding rat genomic
loci and were able to design RT-primers for SOCS4,
SOCS5, and SOCS6. The primer sequences for SOCS1
and SOCS3 have been previously reported (Choi et al.,
2005), and the primers for CIS, SOCS2, and SOCS7 were
newly designed from their cDNA sequences in PubMed
(Table 1). The primer pairs generated the expected sizes
of PCR products, and we found that IFN-γ induced CIS,
SOCS1, and SOCS3 within a short time of stimulation
(Fig. 1). However, we did not detect any induction of:
SOCS2, SOCS4, SOCS6, or SOCS7. PMA induced CIS
and SOCS3 (but not SOCS1) in a manner similar to IFN-γ








Fig. 2. PMA increases SOCS5 protein. Primary astrocytes were
serum-starved for 24 h and stimulated with 1 μM PMA or IFN-γ
(10 U/ml) for the indicated times. Cells extracts were fraction-
ated by SDS-PAGE and Western-blot analysis was performed
using antibodies against SOCS1, SOCS5, and GAPDH.


stimulation. The induction by PMA was maintained for
three hours, whereas expression of CIS, SOCS1, and
SOCS3 peaked at one hour with IFN-γ. We also detected
induction of the SOCS5 gene by PMA.
We investigated protein levels using antibodies against
CIS, SOCS1, SOCS3, and SOCS5 (Fig. 1). We were
only able to detect expression of SOCS1 and SOCS5 pro-
teins (Fig. 2) and we observed no increase in SOCS1 pro-
tein in response to either treatment (Fig. 2) suggesting
that SOCS1 may be regulated by translational control.
Expression of SOCS5 (64 kDa) was detected three hours
after PMA treatment, the same time that its mRNA in-
creased (Fig. 2). The 64 kDa band was visible between
two non-specific bands (see arrow in Fig. 2A).

PMA activates the Jnk, Erk, p38 and Jak signal path-
ways, but not Src phosphorylation We investigated the
effect of PMA on the intensity and duration of phosphory-
lation of Jak1, Jnk1, Erk, and p38 since these kinases
have been shown to be associated with SOCS expression
and to mediate inflammation in astrocytes. We found that
PMA activated Jak1, Jnk1, Erk1/2 and p38 with maxi-
mum levels at one hour (Figs. 3A and 3B). We did not
detect any phosphorylation of Jnk1 or p38 in cells treated
with IFN-γ (Figs. 3A and 3B), and Erk1/2 was only
slightly activated. Previously, it was reported that PMA
induced serine/threonine phosphorylation of c-Src in
glioblastoma cells (Amos et al., 2005), but we did not
detect any phosphorylation of c-Src by PMA or IFN-γ
(Fig. 3A).

PMA phosphorylate only tyrosine residues of Stat3
among the Stat proteins Since Jak1 was activated by
PMA and IFN-γ (Fig 3B), we examined whether they in
turn phosphorylated any Stat proteins. IFN-γ signaling led
to tyrosine-phosphorylation of Stat1, Stat5, and Stat6
whereas we found that PMA only induced tyrosine-
phosphorylation of Stat3. This increased by one hour and
then declined rapidly (Fig. 3C). We did not detect any
serine phosphorylation of Stat1 or Stat3 (Fig. 3C).

Page 4

Mi-Na Hwang et al. 97

A C





B






Fig. 3. A−B. PMA activates Jnk1, Erk, Jak1, and p38. Primary
astrocytes were serum-starved for 24 h and stimulated with 1
μM PMA or IFN-γ (10 U/ml) for the indicated times. Cells ex-
tracts were fractionated by SDS-PAGE and Western-blot analy-
sis was performed using antibodies against phosphorylated-Jnk1,
c-Src, -Jak1, -Erk1/2, and -p38, and un-phosphorylated Jak1,
Jnk1, Erk1/2, and p38. Anti-GAPDH antibody was used as a
loading control. C. PMA phosphorylates only the tyrosine resi-
due of Stat3. Western-blot analysis was performed with the
samples indicated in Figs. 3A and 3B using antibodies against
tyrosine phosphorylated (pY) Stat1, Stat3, Stat5, Stat6, and Jak1,
serine phosphorylated (pS) Stat1, and Stat3, and GAPDH. The
membranes were stripped out and re-probed with antibodies
against Stat1, Stat3, Stat5, Stat6, and Jak1.


PMA and IFN-γ increase binding to proximal Stat and
GAS/ISRE elements Phosphorylated Stat dimers bind
functional GAS/ISRE elements and proximal Stat binding
elements in several inflammation-related genes. Therefore,
we investigated the effects of PMA and IFN-γ on binding
to these elements. Rat brain primary astrocytes were
treated with PMA or IFN-γ for one or three hours. Nuclear
extracts were then prepared and assayed for proximal Stat
or GAS/ISRE binding activity by EMSAs. We found that
extracts from PMA- or IFN-γ-stimulated cells shifted the
Stat bands (Figs. 4A and 4B). Binding to the GAS element
by extracts of either PMA or IFN-γ stimulated cells peaked
at three hours, whereas binding to the proximal Stat bind-
ing element reached its highest level by one hour (Figs. 4A
and 4B). Addition of anti-Stat3 antibody reduced the
shifted band formed by PMA-treated extracts (Fig. 4C).
This suggests that the anti-Stat3 antibody removed Stat3,
leaving the binding activity for the GAS/ISRE elements.
We conclude that the shifted band formed by the PMA-
treated extracts includes Stat3.


Discussion

We have investigated which SOCS genes are regulated in
A





B C







Fig. 4. Nuclear extracts of PMA or IFN-γ stimulated cells con-
tain proteins binding to the proximal Stat or GAS/ISRE ele-
ments. A. Primary astrocytes were stimulated with 1 μM PMA
or IFN-γ (10 U/ml) for one hour and three hours (marked above
the gel). The control lane was not treated with any stimulants.
After preparation of nuclear extracts, the oligonucleotide bind-
ing activity of proximal Stat elements was measured by EMSA.
B. The nuclear extracts used in (A) were tested for binding to
the GAS/ISRE element. As a control a twenty-fold excess of
non-radiolabeled probe was mixed with the radio-labeled probes
in PMA (marked as 1 h and 20X) and IFN-γ (marked as 1 h and
20X) lanes. C. Primary astrocytes were stimulated with 1 μM
PMA for three hours (marker as PMA). The control lanes were
not treated with any stimulants. Nuclear extracts (PMA or Con-
trol) were tested for binding to the GAS/ISRE element. The
EMSA was performed in the presence (marked as Anti-Stat3
Ab) or absence of 1 μg of the anti-Stat3 antibody.


the transition from resting to activated glial cells. We
investigated the expression of eight members of the SOCS
family in primary astrocyte cells using two inflammatory
stimulants, PMA and IFN-γ. Both stimulants induced
iNOS expression indicating that the astrocytes in the
presence of PMA or IFN-γ were in a state of inflamma-
tion for at least eight hours. IFN-γ induced stronger iNOS
expression, and may cause greater inflammation (Fig. 1).
Previously, it was shown that IFN-γ activates the Jak/Stat
pathway and up-regulates CIS, SOCS1, and SOCS3 (Choi
et al., 2005; Yang et al., 2004). We also observed IFN-γ
induction of these compounds for a short time after stimu-
lation (Fig. 1). However, we did not detect any induction
of SOCS2, SOCS4, SOCS6, or SOCS7. PMA induced
CIS and SOCS3 (but not SOCS1) in a manner similar to
IFN-γ. The induction by PMA was maintained for three
hours, whereas expression of CIS, SOCS1, and SOCS3
peaked at one hour with IFN-γ. These findings suggest
that the time of induction of SOCS genes is dependent on
the inflammatory stimulant. We also found induction of
the SOCS5 gene by PMA. Expression of SOCS5 with
IFN-γ stimulation was also evident at one hour, but the

Page 5

98 Induction of SOCS Members in Astrocytes












Fig. 5. Current results and models are summarized in this sche-
matic diagram. The two inflammatory stimulants, PMA and
IFN-γ, use different mechanisms to regulate the expression of
SOCS family genes in rat astrocytes. PMA activates the Jnk, Erk,
p38, and Jak signal pathways, whereas IFN-γ activates the Jak
and Erk pathways only. PMA increases activated Stat3 by tyro-
sine phosphorylation, whereas IFN-γ activates Stat1, Stat5, and
Stat6. Both stimulants activate CIS and SOCS3. PMA also in-
duces SOCS5 (but not SOCS1).


amount was minimal. It is possible that expression of
SOCS5 occurs very early after IFN-γ stimulation. How-
ever, we concluded that SOCS5 was not induced by IFN-
γ, when compared the induction times of CIS, SOCS1 and
SOCS3.
In our Western blot analysis we did not detect any pro-
teins in the size range of CIS or SOCS3. However, SOCS
proteins have been found to be too unstable for detection
by Western blot analysis. We were only able to detect ex-
pression of SOCS1 and SOCS5 (Fig. 2). Since there was
no induction of SOCS1 mRNA by PMA, detection of the
corresponding protein was not to be expected (Fig. 2).
Previously it was reported that SOCS1 translation is regu-
lated by a 5′-untranslated region (Chen et al., 2002), and
this might explain why we did not observe any increased
protein. The expression of SOCS5 (64 kDa) was detected
at three hours after PMA treatment and disappeared by
eight hours, which matched the degradation time for the
mRNA (Fig. 1). As mentioned above, the SOCS1 gene
could be regulated by translational control. Therefore,
induction of SOCS5 by PMA may be regulated at the
transcriptional level; this suggests that SOCS5 was in-
duced by PMA and translated without any regulation. It
has been shown previously that induction of CIS, SOCS1,
SOCS2 and SOCS3 occurs during inflammation of the
brain. This study is the first demonstration of the induc-
tion of SOCS5 protein in astrocytes in an inflammatory
state.
We also studied the role of the Jnk, Erk, p38 signal
pathways in SOCS expression. The Jnk, Erk, p38, Jak
signal pathways were activated by PMA whereas Erk1/2
activation was only slightly increased by IFN-γ. In addi-
tion, we did not detect any phosphorylation of Jnk1 or
p38 in cells treated with IFN-γ. These results suggest that
PMA and IFN-γ activate different signal pathways, al-
though induction of the SOCS genes was similar (Fig. 1).
Jak/Stat signaling pathways are the major pathways
that positively regulate the expression of SOCS genes. It
has been shown that phosphorylation of Stats is dependent
on the activation of Jaks, and activated Stats are responsi-
ble for induction of the SOCS genes. We investigated
which Stat members were responsible for induction of the
SOCS genes and found that PMA only activated tyrosine-
phosphorylation of Stat3. This suggests that Stat3 activa-
tion by PMA might be the major Jak/Stat pathway in-
volved. When we compare the transcriptional induction
time (three hours) of SOCS genes by PMA (Fig. 1) and of
one hour for Stat3 (Fig. 3C), the results suggest that the
Stat3-induced SOCS genes degrade the positive transcrip-
tional activators (Stat3) by feedback inhibition. Stat1 and
Stat3 have redundant functions in gene regulation. Hence
activation of Stat3 by PMA may be analogous to the
SOCS gene induction of activated Stat1 by IFN-γ. Previ-
ously, it was showed that phosphorylation of Stat3 in pri-
mary microglia occurred 15 minutes after IFN-γ stimula-
tion (Jeon et al., 2005). However, in astrocytes, we did
not detect any phosphorylation of Stat3 by IFN-γ at one
hour (Fig. 3C). It appears that the absence of Stat3 activa-
tion by IFN-γ, in contrast to the activation of Stat1, Stat5
and Stat6, is related to the cell type-specificity of astro-
cytes. Our results suggest that activation of Stat3 by PMA
may be sufficient for SOCS expression in the Jak/Stat
pathway. Although Jak1, a kinase for Stats, was activated
by both PMA and IFN-γ, there appear to be differences in
the activation patterns of the different Stat members (Fig.
3C). As shown in Figs. 3A−3B, there was also a large
difference in the levels of activation of the Jnk, p38, and
Erk signal pathways by PMA and IFN-γ. This may result in
activation of different Stats. The GAS/ISRE element is
responsible for binding of phosphorylated Stat dimers in
the promoter region of several inflammation-related genes.
It has been shown that the proximal Stat binding element in
the SOCS3 promoter is responsible for Stat activity (Chen
et al., 2002; Jeon et al., 2005). Therefore, we studied the
effects of PMA and IFN-γ on binding to the proximal Stat
and GAS/ISRE elements by EMSA. We found that PMA
and IFN-γ stimulation shifted a band containing the Stat
binding site (Figs. 4A and 4B). The shifted band was re-
duced upon addition of the anti-Stat3 antibody, suggesting
that activated Stat3 was bound to the GAS/ISRE elements
and was present in the shifted band. Considering that
PMA only activated Stat3 (Fig. 3C), we may conclude
that PMA-activated Stat3 is involved in the major Jak/Stat
pathway and is sufficient for activating the SOCS genes.
These results seem to be related to the cell type specificity
of astrocytes. Thus, we propose that Stat3 plays a major
role in SOCS induction by PMA in astrocytes. In addition,
the activation of Stat3 by PMA may have a similar role to