JNK-mediated activation of ATF2 contributes to dopaminergic neurodegeneration in the MPTP mouse model of Parkinson’s disease Qiaoying Huang, Xiaoxiao Du, Xin He, Qing Yu, Kunhua Hu, Wolfgang Breitwieser, Qingyu Shen, Shanshan Ma, Mingtao Li PII: DOI: Reference:
S0014-4886(15)30111-4 doi: 10.1016/j.expneurol.2015.10.010 YEXNR 12143
To appear in:
Experimental Neurology
Received date: Revised date: Accepted date:
18 April 2015 2 October 2015 24 October 2015
Please cite this article as: Huang, Qiaoying, Du, Xiaoxiao, He, Xin, Yu, Qing, Hu, Kunhua, Breitwieser, Wolfgang, Shen, Qingyu, Ma, Shanshan, Li, Mingtao, JNKmediated activation of ATF2 contributes to dopaminergic neurodegeneration in the MPTP mouse model of Parkinson’s disease, Experimental Neurology (2015), doi: 10.1016/j.expneurol.2015.10.010
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT JNK-mediated
activation
of
ATF2
contributes
to
dopaminergic
neurodegeneration in the MPTP mouse model of Parkinson’s disease.
IP
T
Qiaoying Huanga, 1, Xiaoxiao Dua, 1, Xin Hea, 1, Qing Yua, Kunhua Hua, Wolfgang
a
SC R
Breitwieserc, Qingyu Shenb, d, Shanshan Maa, *, Mingtao Li a, b, *
Department of Pharmacology and bGuangdong Province Key Laboratory of Brain
NU
Function and Disease, Zhongshan School of Medicine, Sun Yat-sen University, 74
c
MA
Zhongshan 2nd Road, Guangzhou 510080, China
Cell Regulation Department, CRUK Manchester Institute, Wilmslow Road
Department of Neurology, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University,
CE P
d
TE
D
Manchester, M20 4BX, United Kingdom
Number 107, Yan Jiang Xi Road, Guangzhou 510120, China
AC
*Corresponding authors.
E-mail addresses: [email protected] (M. Li), [email protected] (S. Ma)
1
These authors contributed equally to this work.
Abstract The c-Jun N-terminal kinase (JNK)/c-Jun pathway is a known critical regulator of dopaminergic neuronal death in Parkinson’s disease (PD) and is considered a potential target for neuroprotective therapy. However, whether JNK is activated 1
ACCEPTED MANUSCRIPT within dopaminergic neurons remains controversial, and whether JNK acts through downstream effectors other than c-Jun to promote dopaminergic neuronal death
T
remains unclear. In this study, we confirm that JNK but not p38 is activated in
IP
dopaminergic neurons after 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-
SC R
intoxication. Furthermore, within the dopaminergic neurons of the substantia nigra in MPTP-treated mice, JNK2/3 phosphorylates threonine 69 (Thr69) of Activating transcription factor-2 (ATF2), a transcription factor of the ATF/CREB family,
NU
whereas the phosphorylation of Thr71 is constitutive and remains unchanged. The
MA
increased phosphorylation of ATF2 on Thr69 by JNK in the MPTP mouse model suggests a functional relationship between the transcriptional activation of ATF2 and
D
dopaminergic neuron death. By using dopaminergic neuron-specific conditional
TE
ATF2 mutant mice, we found that either partial or complete deletion of the ATF2 DNA-binding domain in dopaminergic neurons markedly alleviates the MPTP-
CE P
induced dopaminergic neurodegeneration, indicating that the activation of ATF2 plays a detrimental role in neuropathogenesis in PD. Taken together, our findings
AC
demonstrate that JNK-mediated ATF2 activation contributes to dopaminergic neuronal death in an MPTP model of PD.
Keywords Parkinson’s disease, dopaminergic neurodegeneration, JNK, ATF2
Introduction Parkinson’s disease (PD) is a common neurodegenerative disease; the cardinal symptoms of PD are due to the progressive loss of dopaminergic neurons of the 2
ACCEPTED MANUSCRIPT substantia nigra pars compacta (SNpc) and the consequent loss of the projecting nerve fibers in the striatum (Lees et al., 2009).
T
c-Jun NH2-terminal kinase (JNK) is implicated in multiple paradigms of neuronal
IP
death including PD pathogenesis (Philpott and Facci, 2008) and is also considered as
SC R
a potential therapeutic target for neurodegenerative disease (Wang et al., 2012). JNK activation has been demonstrated in post-mortem brain tissue of the SNpc from PD patients (Ferrer et al., 2001) as well as in parkinsonian genetic models (Akundi et al.,
NU
2011; Cha et al., 2005; Chen et al., 2012) and multiple parkinsonian animal models
MA
induced by neurotoxins such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (Nishi, 1997; Saporito et al., 2000), 6-OHDA (Hu et al., 2011; Pan et al., 2007) and
D
paraquat (Peng et al., 2004). JNK inhibition and JNK deficiency in mice conferred
TE
neuroprotection in animal models of this disease (Hunot et al., 2004; Saporito et al., 1999). Although there is compelling evidence that activation of JNK promotes
CE P
dopaminergic neuronal death, the in situ evidence showing JNK activation in a cellspecific manner is limited. Furthermore, one research group reported that p38 was
AC
activated in dopaminergic neurons of the SNpc in an MPTP mouse model, whereas JNK was predominantly activated in microglia (Karunakaran et al., 2008). Therefore, to inform future strategies for precisely targeting JNK to protect dopaminergic neurons, it is important to clarify whether and in which cell type JNK is activated in PD. The mechanism of JNK-mediated dopaminergic neuron death is also not fully elucidated. c-Jun is the direct downstream effector of JNK that mediates dopaminergic neurodegeneration, as established in a variety of PD animal models (Brecht et al., 2005; Crocker et al., 2001; Hayley et al., 2004). c-Jun is a transcription factor belonging to the AP-1 family that acts by forming homodimers or heterodimers 3
ACCEPTED MANUSCRIPT with other AP-1 or ATF proteins (Shaulian and Karin, 2002). We previously demonstrated that activating transcription factor-2 (ATF2) is activated by JNK in
T
response to proapoptotic stimuli, dimerizes with c-Jun, and binds to CRE/ATF
IP
consensus sequences to mediate neuron death (Ma et al., 2007; Yuan et al., 2009).
SC R
ATF2 is expressed ubiquitously and most abundantly in the central nervous system (Kara et al., 1990; Maekawa et al., 1989; Takeda et al., 1991). Decreased ATF2 expression is found in the SNpc of postmortem tissue from PD patients (Pearson et al.,
NU
2005). However, whether ATF2 is involved in JNK-dependent dopaminergic neuronal
MA
death in parkinsonian models remains unclear.
In the present study, we first confirm that JNK and not p38 is activated in
D
dopaminergic neurons in an MPTP mouse model of PD and further demonstrate that
TE
ATF2 is a novel JNK substrate that promotes MPTP-induced dopaminergic neuronal
CE P
death.
Materials and methods
AC
Chemicals and reagents
MPTP·HCl and paraformaldehyde (PFA) were purchased from Sigma-Aldrich (Shanghai, China). The following primary antibodies were used in this study: anti-pJNK (Promega, Cat. #V7932), p-JNK T183/Y185 (G-7, Santa Cruz, Cat. #sc-6254), p-p38 T180/Y182 (Cell Signaling, Cat. #4511), p38 (Cell Signaling, Cat. #8690), pATF2 T69/71 (Cell Signaling, Cat. #9225), p-ATF2 T69/71 (Millipore, Cat. #05-891), p-ATF2 T71 (Cell Signaling, Cat. #9221), ATF2 (Abcam, Cat. #ab32160), ATF2 DNA-binding domain (ATF2-DBD; reported previously) (Ackermann et al., 2011), pc-Jun S63 (Cell Signaling, Cat. #9261), TH (Millipore, Cat. #AB1542), TH (Millipore, Cat. #AB152), CD11b (Millipore, Cat. #MAB1387Z), and Iba-1 (Wako, #019-19741). 4
ACCEPTED MANUSCRIPT Normal donkey serum, bovine serum albumin (BSA, IgG-free, protease-free) and donkey anti-mouse IgG (H+L) Fab fragments were purchased from Jackson
T
ImmunoResearch. The ABC kit and DAB kit are from Vector Laboratories.
IP
Animals and treatments
SC R
All animal experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Sun Yat-sen University. The mice were housed in rooms with controlled 12 h light/dark cycles, temperature, and
NU
humidity, and food and water were provided ad libitum. Male C57BL/6J mice were
MA
obtained from the Experimental Animal Center of Sun Yat-sen University. The floxed Atf2 mice backcrossed to C57BL/6 for at least 10 generations were previously
D
reported (Ackermann et al., 2011). The dopamine transporter (DAT)-Cre mutant mice
TE
backcrossed to C57BL/6 for eleven generations and the JNK1-, JNK2- and JNK3deficient mice backcrossed to C57BL/6 for six generations were obtained from The
CE P
Jackson Laboratory (stock numbers 006660, 004319, 004321, 004322). All animal treatments were performed as previously described (Wang et al., 2004;
AC
Wang et al., 2007). In brief, 8-12-week-old male mice weighing 22-28 g were treated with MPTP [30 mg/kg of free base, intraperitoneally (i.p.)] at 24-hour intervals for 5 consecutive days. The control animals received a corresponding volume of saline alone. The mice were sacrificed at the indicated time points. For immunofluorescent or immunohistochemical analysis, the mice were anesthetized with chloral hydrate (400 mg/kg, i.p.) and then perfused with ice-cold phosphate-buffered saline (PBS; 0.01 M, pH 7.4) for 3 min, followed by 4% paraformaldehyde (PFA) for 8 min, at a flow rate of 10 ml/min. The brains were removed and postfixed overnight in 4% PFA at 4°C and then cryoprotected in 20% and 30% sucrose. For western blots, the mice were anesthetized, perfused with ice-cold saline containing 1 mM Na3VO4 and 20 5
ACCEPTED MANUSCRIPT mM NaF, and then sacrificed by cervical dislocation. The brains were washed with ice-cold PBS and the ventral midbrains were rapidly dissected on ice.
T
Immunofluorescence
IP
Coronal sections (20 μm) throughout the entire midbrain were cut using a cryostat
SC R
(Leica CM1950). Free-floating tissue sections from each of the treatment groups and time points were immunostained in parallel using the same solutions to minimize variation between experiments. The sections were pre-incubated in blocking solution
NU
containing 5% normal donkey serum and 0.3% Triton X-100 in 50 mM Tris-buffered
MA
saline (TBS, pH 7.4) at room temperature for 1 h. For endogenous IgG blocking, the sections were then incubated in Fab fragments (250 μg/ml) for 1.5 h at room
D
temperature and washed three times. Primary antibodies were diluted in 1% BSA and
TE
0.3% Triton X-100 in TBS (pH 7.4) and were applied overnight at 4°C. After three washes in TBS with 0.05% Tween 20, the sections were incubated with Alexa 488-,
CE P
Alexa 555- or Alexa 647-conjugated secondary antibodies diluted in the same buffer as the primary antibodies for 1 h at room temperature. All sections were
AC
counterstained with Hoechst 33258. Following three additional washes, the sections were coverslipped with mounting media and visualized with an inverted confocal laser scanning microscope (LSM 780, Carl Zeiss, Germany). Sections with the primary antibody omitted were processed and photographed under the same conditions and used as negative controls. Immunofluorescence quantification was performed using ImageJ software. The SNpc was defined as Regions of Interest (ROI) using Polygon selection tool and the integrated optical density (IOD) of the color channal of target protein was measured. Backgrounds of images from different sections were set the same, and IOD was measured with the same threshold. The IOD in saline-treated control mice is defined as one. 6
ACCEPTED MANUSCRIPT Immunohistochemistry Coronal sections (40 μm) encompassing the entire midbrain and striatum were
T
serially collected. Sections containing the striatum and the SNpc were processed for
IP
TH staining. The sections were incubated with rabbit anti-TH antibody (1:10 000) and
SC R
then visualized with the Vectastain Elite ABC kit (Vector Labs, #PK-6101) and the DAB peroxidase substrate kit (Vector Labs, #SK-4100) following the manufacturer’s protocol. Adjacent SNpc sections were used for Nissl-staining to evaluate the survival
NU
of nigral neurons.
MA
Cell counts
Nigral dopaminergic cell were stereologically assessed using the optical
D
fractionator with the aid of Stereo Investigator (MicroBrightField Inc; Williston, VT,
TE
USA). Nigral dopaminergic groups were outlined on the basis of TH immunolabeling, with reference to a standard mouse brain atlas (Paxinos and Franklin, 2001) and the
CE P
substantia nigra cell group described in (Fu et al., 2012). TH-positive neurons with clearly identified nuclei were counted in both hemispheres of the SNpc of every
AC
fourth section throughout the entire extent of the midbrain (AP -2.7 to AP -3.8 mm). The numbers of TH- and Nissl-possitive cells were counted by investigators blinded to the treatments and genotypes as described (Baquet et al., 2009). Striatal Densitometry The density of striatal DA terminals was measured as the IOD of the striatal THimmunoreactive signal using ImageJ software. Four sections were randomly selected from those containing the striatum at the approximate level of Bregma-0.22 to 1.10 mm according to the mouse brain atlas (Paxinos and Franklin, 2001), and the IOD in the dorsal striatum of each section was measured on each side. In each section, the
7
ACCEPTED MANUSCRIPT optical densities were corrected by subtracting out the background staining in the corpus callosum.
T
Striatal 1-methyl-4-phenylpyridinium ion (MPP+) levels
IP
The measurement of MPP+ levels in the various genotypes was conducted
SC R
according to a previously published protocol (Jackson-Lewis and Przedborski, 2007). Mice were sacrificed at 90 min after one i.p. injection of 30 mg/kg MPTP, and the striata of both sides were dissected and processed for high-performance liquid
NU
chromatography (HPLC) with ultraviolet (UV) detection (wavelength 295 nm).
MA
Western blot
The ventral midbrain tissue was weighed, quickly homogenized in Laemmli buffer
D
(50 μl/mg tissue) containing Tris·Cl (62.5 mM, pH 6.8), SDS (2%, w/v), Butterfield’s
TE
phosphate buffer (BPB; 0.005%, w/v), glycerol (10%, v/v) and DTT (8 mg/ml), and then boiled at 95-100°C for 5 min. The protein lysates (30 μg) were then resolved on
CE P
10% SDS-PAGE gels, electrotransferred to PVDF (Roche) membranes, and probed with primary antibodies at 4°C overnight, followed by the appropriate horseradish
AC
peroxidase-conjugated secondary antibodies. All blots were visualized using ECL chemiluminescence (Amersham Biosciences). Quantification was performed using ImageJ software.
Statistical analysis The statistical analyses were performed using GraphPad Prism version 6.01 for Windows (GraphPad Software, La Jolla, California, USA; www.graphpad.com). All data are presented as the mean ± standard error of the mean (SEM). Unpaired twotailed t-tests were used for comparisons between the saline- and MPTP-treated groups. When there were two variables (genotype and treatment), two-way ANOVA followed by Tukey’s multiple comparisons test was used. 8
ACCEPTED MANUSCRIPT
Results
clarify
whether
and
where
JNK
is
activated,
IP
To
T
JNK is activated predominantly in dopaminergic neurons in MPTP-treated mice. we
performed
SC R
immunofluorescence with a rabbit anti-JNK antibody. As shown in Fig. 1A, faint staining of phospho-JNK was apparent in dopaminergic neurons of the SNpc in the
NU
saline-treated group. An increase in phospho-JNK was visible as early as 6 hours after administering the first dose of MPTP administration. Consecutive treatments with
MA
MPTP led to the lasting JNK activation, which peaked after five doses of MPTP injection (Fig. 1A, B). At high magnification, the phospho-JNK signal was located
D
almost exclusively in the nuclei of dopaminergic neurons (Fig. 1A inserts). The
TE
specificity of this rabbit anti-phospho-JNK antibody has been confirmed
CE P
(Supplemental Fig. 1A-C). This observation was further supported by western blot analysis (Fig. 1C).
To explain the significant differences between our and others’ findings
AC
(Karunakaran et al., 2008), we performed double immunostaining using the same mouse anti-phospho-JNK antibody (G-7, Santa Cruz). We did find increased signals in microglial cells, predominantly at the microglial surface, but this pattern of signals did not diminish upon omitting the primary antibody in the negative controls (Supplemental Fig. 2). We assumed that these increased signals were due to the antimouse IgG secondary antibody recognizing endogenous IgG on the surface of microglia, which increased in concert with microglial activation. Therefore, we used anti-mouse IgG Fab fragments to block endogenous mouse IgG. As shown in Supplemental Fig. 2, the increased signal on the surface of microglia was completely 9
ACCEPTED MANUSCRIPT blocked. These results confirm that JNK is activated mainly in the nuclei of dopaminergic neurons after MPTP treatment, suggesting that the nuclear pathway is
T
involved in JNK-dependent dopaminergic neurodegeneration.
IP
ATF2 Thr69 is phosphorylated within dopaminergic neurons after MPTP
SC R
intoxication.
The transcription factor c-Jun is a classic substrate of JNK, showing expected
NU
activation upon MPTP treatment (Fig. 2A). Our previous work has demonstrated that the dimerization of activated c-Jun and ATF2 is necessary for JNK-dependent
MA
neuronal death (Yuan et al., 2009). Therefore, we wondered whether ATF2 is also activated in parkinsonian models. To explore this possibility, we first observed the
D
phosphorylation levels of ATF2 Thr69/71 (human homolog corresponding to
TE
Thr51/53 of murine ATF2), which is essential for its transcriptional activity (Gupta et
CE P
al., 1995; Livingstone et al., 1995; van Dam et al., 1995), and found an increase in the phosphorylation of ATF2 Thr69/71 (Fig. 2B). A time-course analysis showed a timedependent increase of phospho-ATF2 Thr69/71, which peaked after the last dose of
AC
MPTP and was localized exclusively to the nuclei of TH-positive neurons in the ventral midbrain of the MPTP-treated mice (Fig. 2B, D). The increase of phosphoATF2 was temporally and spatially consistent with that of phospho-JNK (Fig. 1A, B). To confirm this result, we adopted another acknowledged phospho-ATF2 antibody that has been shown to specifically recognize Thr71-monophosphorylated ATF2 (corresponding to murine homolog Thr53). The specificity of both phospho-ATF2 antibodies has been demonstrated by previous studies (Ackermann et al., 2011; Morton et al., 2004; Ouwens et al., 2002). To our surprise, we found that phosphoATF2 Thr71 was constitutively expressed in dopaminergic neurons of control mice 10
ACCEPTED MANUSCRIPT and remained unchanged after MPTP-treatment (Fig. 2C, E). This observation was further supported by western blot analysis (Fig. 2F). Given that the phospho-ATF2
T
Thr69/71 antibody recognizes Thr69/71-dual-phosphorylated but not mono-
IP
phosphorylated ATF2 (Morton et al., 2004; Ouwens et al., 2002), we therefore
SC R
concluded that phospho-ATF2 Thr71 is constitutively expressed and that phosphoATF2 Thr69 is markedly increased in dopaminergic neurons after MPTP intoxication. Therefore, the upstream kinases that could phosphorylate the Thr69 of ATF2 may be
NU
highly valuable for the ATF2 activation in the MPTP PD model.
MA
JNK2/3 mediates the phosphorylation of ATF2 Thr69 during MPTP-induced neurodegeneration.
D
ATF2 Thr69 can be phosphorylated not only by JNK but also by p38 (Raingeaud et
TE
al., 1995). Although we found that ATF2 activation is spatially and temporally
CE P
accompanied by that of JNK (Fig. 1A,B; Fig. 2B, D), other evidence has shown that p38 was also activated in dopaminergic neurons by MPTP treatment (Karunakaran et al., 2008). To investigate whether p38 is activated in addition to JNK, we measured
AC
phospho-p38 levels in parallel with phospho-JNK in situ, using an anti-phospho-p38 antibody whose specificity has been confirmed (Supplemental Fig. 1D, E). Basal expression of phospho-p38 was detected in the nuclei of dopaminergic neurons in the SNpc (Fig. 3A, B). However, no increase was observed after consecutive doses of MPTP treatment in dopaminergic neurons (Fig. 3A, B). Results from western blot analysis were in consistant with those from immunofluorescence (Fig. 3C). These results clarify that JNK not p38 is activated in dopaminergic neurons of the SNpc after MPTP exposure. Thus, the increase in phospho-ATF2 Thr69 is very likely due to JNK activation. 11
ACCEPTED MANUSCRIPT To further confirm whether JNK phosphorylates ATF2 Thr69 in dopaminergic neurons after MPTP intoxication, we introduced JNK2/3 knockout mice because it
T
has been shown that JNK2 and JNK3 mediate the MPTP-associated stress response in
IP
dopaminergic neurons, whereas JNK1 does not (Hunot et al., 2004). In JNK2/3-
SC R
knockout mice, MPTP-induced c-Jun phosphorylation was completely blocked compared with that in wild-type mice (Fig. 4A). This finding suggests the complete loss of nuclear JNK activity in these compound mutants, whereas JNK1-deficiency
NU
has no effect on c-Jun activation (data not shown). Moreover, the increased signal
MA
detected by the ATF2 Thr69/71-dual-phosphorylation antibody was almost completely blocked in JNK2/3-null mice upon MPTP treatment (Fig. 4B, D), whereas
D
the expression level of Thr71-mono-phosphorylated ATF2 remained steady (Fig. 4C,
TE
E). Together, these results indicate that JNK2/3 mediates ATF2 activation in the MPTP model through the phosphorylation of Thr69, which is primed by the
JNK.
CE P
constitutively abundant phosphorylation of Thr71 by an undefined kinase other than
AC
The reduction in functional ATF2 alleviates MPTP-induced dopaminergic neurodegeneration. ATF2 activation in PD-related models has not been previously reported, and we were eager to evaluate whether it promotes dopaminergic neurodegeneration. For this purpose, we crossed DatCre/+ (Backman et al., 2006) mice with mice expressing a floxed allele of Atf2 (Atf2f/f) (Ackermann et al., 2011) to obtain dopaminergic neuronspecific ATF2-deficient mice. DatCre/+-induced recombination has been shown to be efficient and dopaminergic neuron-specific (Backman et al., 2006). The crosses led to the deletion of the DNA binding domain of ATF2 and thus caused the loss of 12
ACCEPTED MANUSCRIPT transcriptional function of ATF2. Immunofluorescence using an ATF2-DBD antibody (raised against the ATF2 DNA-binding domain) (Ackermann et al., 2011) showed
T
that DatCre/+ mediated the complete deletion of ATF2 in Atf2f/f;DatCre/+ (designated
IP
Atf2 homo) mice and a partial reduction of Atf2 in Atf2f/+;DatCre/+ (designated Atf2
SC R
het) mice (Fig. 5A).
We then used these ATF2 mutant mice and their control littermates
NU
(Atf2+/+;DatCre/+ and Atf2f/f;Dat+/+) to investigate the role of ATF2 in MPTP-induced dopaminergic neuronal death. A statistical analysis of the cell counts in the SNpc
MA
revealed that the basal TH-positive neuron numbers did not differ between the Atf2 het mice and control mice, but the numbers of residual TH-positive neurons after
D
MPTP treatment in Atf2 het mice were much greater than those in control mice
TE
(compared with the saline-treated group, MPTP induced a 48.2% loss of TH-positive cells in control mice and a 20.9% loss in Atf2 het mice), suggesting a protective effect
CE P
of the ATF2 knockout. In Atf2 homo mice, although a notable reduction (35.1% compared with control mice) in the basal numbers of TH-positive neurons was
AC
observed (which may be due to abnormal development of dopaminergic neurons; see discussion), the relative extent of the MPTP-induced loss of dopaminergic neurons was also alleviated (18.1% of TH-positive cells were lost in Atf2 homo mice; Fig. 5B), supporting the view that ATF2 activation promotes dopaminergic neurodegeneration. Nissl staining revealed similar trends and statistical results (39.4%, 18.7% and 10.8% losses in control, Atf2 het and Atf2 homo mice, respectively, Fig. 5B, lower right panel), suggesting the loss of actual TH-positive neurons instead of a loss of TH expression. As expected, we observed a significant loss of striatal TH-staining density in MPTP-treated control mice when normalized to the saline condition, and this was partially reversed in MPTP-treated Atf2 mutant mice (49.9%, 24.6% and 26.0% 13
ACCEPTED MANUSCRIPT reductions in control, Atf2 het and Atf2 homo mice, respectively; Fig. 5C). Similar changes of the nigral dopaminergic dendrites were also shown (Supplemental Fig. 3).
T
To determine whether the reduced loss of dopaminergic neurons or the decrease in
IP
axonal and dendritic density was due to the reduced production of MPP+ from MPTP,
SC R
we measured striatal MPP+. Indeed, we found similar levels of MPP+ in control, Atf2 het and Atf2 homo mice (Table 1). Taken together, these findings demonstrate that
NU
functional ATF2 is involved in MPTP-induced dopaminergic neuronal death.
MA
Discussion
In the present study, we used a subacute MPTP-induced PD model to investigate the
D
mechanism of dopaminergic neuronal degeneration in PD. Our principal findings
TE
were as follows: (i) we confirm that JNK, not p38, is predominantly activated in dopaminergic neurons of the SNpc; (ii) we demonstrate for the first time that ATF2
CE P
Thr69 phosphorylation is increased selectively in dopaminergic neurons by MPTP toxicity, whereas ATF2 Thr71 phosphorylation is constitutive and remains unchanged;
AC
(iii) we demonstrate that JNK2 and JNK3 mediate ATF2 Thr69 phosphorylation but have no effect on ATF2 Thr71 phosphorylation; and (iv) we demonstrate that ATF2 activation is involved in MPTP-induced dopaminergic neurodegeneration. Numerous studies have addressed the importance of JNK in mediating dopaminergic neurodegeneration in both sporadic and familial PD (Wang et al., 2012). However, direct evidence showing that phosphorylated JNK is increased in dopaminergic neurons of the SNpc is limited. Most of the previous studies used immunoblots to analyze levels of phospho-JNK in the SNpc or ventral midbrain tissue lysates in the MPTP-induced PD model (Boyd et al., 2007; Castro-Caldas et al., 2012; Chen et al., 2005; Karunakaran et al., 2007; Saporito et al., 2000; Sung et al., 2012; 14
ACCEPTED MANUSCRIPT Xia et al., 2001) and therefore could not distinguish between signals from neurons and those from other cell types. Until recently, when Karunakaran et al. reported that JNK
T
phosphorylation was predominant in microglia using a mouse antibody (Karunakaran
IP
et al., 2008), no in situ evidence in the MPTP-induced mouse model was available.
SC R
Here, we found that staining mouse tissue with a mouse antibody without blocking endogenous IgG could lead to false-positive results. We chose a rabbit anti-phosphoJNK antibody to avoid interference by endogenous murine IgG, showing that
NU
phosphorylated JNK is predominantly increased in dopaminergic neurons and is
MA
localized mainly in the nucleus after MPTP treatment. By using this rabbit antibody, we also observed activated JNK in a small number of microglial cells (data not
D
shown), which suggests a possible role for JNK in microglial activation and pro-
TE
inflammatory actions. Further studies are required to investigate the roles of JNK in different cell types during dopaminergic neurodegeneration.
CE P
We show for the first time that dual-phospho-ATF2 Thr69/71 significantly increases in dopaminergic neurons soon after MPTP injection, whereas mono-
AC
phospho-ATF2 Thr71 is constitutively expressed and remains unchanged, suggesting that Thr69 is a critical regulatory site for ATF2 activation in the MPTP model. By using JNK2/3 knockout mice, we demonstrated that JNK2/3 mediates the phosphorylation of Thr69 but not of Thr71. A two-step mechanism for ATF2 phosphorylation, in which Thr71 is phosphorylated first, thereby priming ATF2 for subsequent efficient Thr69 phosphorylation, has previously been reported (Ouwens et al., 2002). Whether Thr71-mono-phosphorylated ATF2 occurs constitutively in vivo and has functions other than acting as a priming site remains unknown. In the current study, we found that neurons in specific locations, including the SNpc, hippocampus and cortex, express high levels of Thr71-mono-phosphorylated ATF2 but negligible 15
ACCEPTED MANUSCRIPT basal expression of dual-phospho-ATF2 Thr69/71 (data not shown). As it has been pointed out that phosphorylation of Thr69 or Thr71 and transcriptional activation of
T
ATF2 is crucial for viability and development, reviewed in (Lau and Ronai, 2012), we
IP
believe that the physiological functions of Thr71-mono-phosphorylated ATF2 and its
SC R
upstream kinases in vivo may be worth investigating further.
In our study, both homozygous and heterozygous mutants of ATF2 alleviated the MPTP-induced dopaminergic neurodegeneration, revealing that ATF2 is one of the
NU
molecular targets of JNK that participate in dopaminergic cell death. However, Atf2
MA
homo mice, in which functional ATF2 was completely abolished, showed reduced numbers of dopaminergic cells in the SNpc in three-month-old mice. This observation
D
is also consistent with those of previous loss-of-function studies, in which ATF2-
TE
deficient mice demonstrate significant neurological abnormalities (Ackermann et al., 2011; Reimold et al., 1996), including a deficit of dopaminergic fibers from the SNpc
CE P
at E18.5 (Kojima et al., 2008). These results indicate that ATF2 plays a role in neuronal development, and this effect likely depends on the function of Thr71-mono-
AC
phosphorylated ATF2, as we speculate in the above paragraph. Studies of postmortem brain tissue from PD patients demonstrate that ATF2 expression is reduced in some of the remaining pigmented cells of the SNpc in comparison to those in normal tissues (Pearson et al., 2005). ATF2 downregulation can also be detected in other neurological diseases (Pearson et al., 2005) and in response to neuronal injury (Herdegen et al., 1997; Martin-Villalba et al., 1998; Robinson,
1996).
ATF2
activation
by
N-terminal
phosphorylation
and
heterodimerization promotes its ubiquitination-dependent degradation (Fuchs and Ronai, 1999). In the subacute MPTP PD model, we only detected ATF2 activation rather than decreased total protein, possibly because neuronal toxin-induced 16
ACCEPTED MANUSCRIPT dopaminergic neurodegeneration is an acute process in which the dopaminergic neurons die before ATF2 degradation can be detected. However, whether the long-
T
lasting activation of ATF2 leads to the increased degradation of this protein remains
IP
to be verified in tissue from PD patients.
SC R
Our data establish for the first time that ATF2 is a mediator of JNK-dependent cell death in dopaminergic neurons. We also confirm that JNK is activated in dopaminergic neurons, whereas p38 is not. Elucidation of the precise regulation of
NU
these MAP kinases and their downstream effectors in dopaminergic neurons and other
MA
cell types involved in neurodegeneration, such as microglia, will help us to establish
D
more effective targets and improve neuroprotective therapy for PD.
TE
Acknowledgments
This work was supported by the National Basic Research Program of China 973
CE P
Program (2011CB504105), the National Natural Science Foundation of China (U1201224, 81471289, 31371071, 81471290), and the Science and Technology
AC
Project of Guangzhou city (2014Y2-00500). There are no conflicts of interest to declare.
Figure Legends Fig. 1 JNK is activated in dopaminergic neurons in MPTP-treated mice. Mice were treated with 1, 3 and 5 doses of MPTP and sacrificed 6 h post injection (M1×6 h, M3×6 h, M5×6 h). (A-B) The detection of p-JNK T183/Y185 (green) in dopaminergic
neurons
(TH-positive,
red)
of
the
ventral
midbrain
by
immunofluorescence (A) and quantitative data are shown (B). The insert is a higher magnification of the area indicated by the small rectangle. Data are expressed as the 17
ACCEPTED MANUSCRIPT mean±SEM (B, n=5 per group). ** p
S0014-4886(15)30111-4 doi: 10.1016/j.expneurol.2015.10.010 YEXNR 12143
To appear in:
Experimental Neurology
Received date: Revised date: Accepted date:
18 April 2015 2 October 2015 24 October 2015
Please cite this article as: Huang, Qiaoying, Du, Xiaoxiao, He, Xin, Yu, Qing, Hu, Kunhua, Breitwieser, Wolfgang, Shen, Qingyu, Ma, Shanshan, Li, Mingtao, JNKmediated activation of ATF2 contributes to dopaminergic neurodegeneration in the MPTP mouse model of Parkinson’s disease, Experimental Neurology (2015), doi: 10.1016/j.expneurol.2015.10.010
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT JNK-mediated
activation
of
ATF2
contributes
to
dopaminergic
neurodegeneration in the MPTP mouse model of Parkinson’s disease.
IP
T
Qiaoying Huanga, 1, Xiaoxiao Dua, 1, Xin Hea, 1, Qing Yua, Kunhua Hua, Wolfgang
a
SC R
Breitwieserc, Qingyu Shenb, d, Shanshan Maa, *, Mingtao Li a, b, *
Department of Pharmacology and bGuangdong Province Key Laboratory of Brain
NU
Function and Disease, Zhongshan School of Medicine, Sun Yat-sen University, 74
c
MA
Zhongshan 2nd Road, Guangzhou 510080, China
Cell Regulation Department, CRUK Manchester Institute, Wilmslow Road
Department of Neurology, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University,
CE P
d
TE
D
Manchester, M20 4BX, United Kingdom
Number 107, Yan Jiang Xi Road, Guangzhou 510120, China
AC
*Corresponding authors.
E-mail addresses: [email protected] (M. Li), [email protected] (S. Ma)
1
These authors contributed equally to this work.
Abstract The c-Jun N-terminal kinase (JNK)/c-Jun pathway is a known critical regulator of dopaminergic neuronal death in Parkinson’s disease (PD) and is considered a potential target for neuroprotective therapy. However, whether JNK is activated 1
ACCEPTED MANUSCRIPT within dopaminergic neurons remains controversial, and whether JNK acts through downstream effectors other than c-Jun to promote dopaminergic neuronal death
T
remains unclear. In this study, we confirm that JNK but not p38 is activated in
IP
dopaminergic neurons after 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-
SC R
intoxication. Furthermore, within the dopaminergic neurons of the substantia nigra in MPTP-treated mice, JNK2/3 phosphorylates threonine 69 (Thr69) of Activating transcription factor-2 (ATF2), a transcription factor of the ATF/CREB family,
NU
whereas the phosphorylation of Thr71 is constitutive and remains unchanged. The
MA
increased phosphorylation of ATF2 on Thr69 by JNK in the MPTP mouse model suggests a functional relationship between the transcriptional activation of ATF2 and
D
dopaminergic neuron death. By using dopaminergic neuron-specific conditional
TE
ATF2 mutant mice, we found that either partial or complete deletion of the ATF2 DNA-binding domain in dopaminergic neurons markedly alleviates the MPTP-
CE P
induced dopaminergic neurodegeneration, indicating that the activation of ATF2 plays a detrimental role in neuropathogenesis in PD. Taken together, our findings
AC
demonstrate that JNK-mediated ATF2 activation contributes to dopaminergic neuronal death in an MPTP model of PD.
Keywords Parkinson’s disease, dopaminergic neurodegeneration, JNK, ATF2
Introduction Parkinson’s disease (PD) is a common neurodegenerative disease; the cardinal symptoms of PD are due to the progressive loss of dopaminergic neurons of the 2
ACCEPTED MANUSCRIPT substantia nigra pars compacta (SNpc) and the consequent loss of the projecting nerve fibers in the striatum (Lees et al., 2009).
T
c-Jun NH2-terminal kinase (JNK) is implicated in multiple paradigms of neuronal
IP
death including PD pathogenesis (Philpott and Facci, 2008) and is also considered as
SC R
a potential therapeutic target for neurodegenerative disease (Wang et al., 2012). JNK activation has been demonstrated in post-mortem brain tissue of the SNpc from PD patients (Ferrer et al., 2001) as well as in parkinsonian genetic models (Akundi et al.,
NU
2011; Cha et al., 2005; Chen et al., 2012) and multiple parkinsonian animal models
MA
induced by neurotoxins such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (Nishi, 1997; Saporito et al., 2000), 6-OHDA (Hu et al., 2011; Pan et al., 2007) and
D
paraquat (Peng et al., 2004). JNK inhibition and JNK deficiency in mice conferred
TE
neuroprotection in animal models of this disease (Hunot et al., 2004; Saporito et al., 1999). Although there is compelling evidence that activation of JNK promotes
CE P
dopaminergic neuronal death, the in situ evidence showing JNK activation in a cellspecific manner is limited. Furthermore, one research group reported that p38 was
AC
activated in dopaminergic neurons of the SNpc in an MPTP mouse model, whereas JNK was predominantly activated in microglia (Karunakaran et al., 2008). Therefore, to inform future strategies for precisely targeting JNK to protect dopaminergic neurons, it is important to clarify whether and in which cell type JNK is activated in PD. The mechanism of JNK-mediated dopaminergic neuron death is also not fully elucidated. c-Jun is the direct downstream effector of JNK that mediates dopaminergic neurodegeneration, as established in a variety of PD animal models (Brecht et al., 2005; Crocker et al., 2001; Hayley et al., 2004). c-Jun is a transcription factor belonging to the AP-1 family that acts by forming homodimers or heterodimers 3
ACCEPTED MANUSCRIPT with other AP-1 or ATF proteins (Shaulian and Karin, 2002). We previously demonstrated that activating transcription factor-2 (ATF2) is activated by JNK in
T
response to proapoptotic stimuli, dimerizes with c-Jun, and binds to CRE/ATF
IP
consensus sequences to mediate neuron death (Ma et al., 2007; Yuan et al., 2009).
SC R
ATF2 is expressed ubiquitously and most abundantly in the central nervous system (Kara et al., 1990; Maekawa et al., 1989; Takeda et al., 1991). Decreased ATF2 expression is found in the SNpc of postmortem tissue from PD patients (Pearson et al.,
NU
2005). However, whether ATF2 is involved in JNK-dependent dopaminergic neuronal
MA
death in parkinsonian models remains unclear.
In the present study, we first confirm that JNK and not p38 is activated in
D
dopaminergic neurons in an MPTP mouse model of PD and further demonstrate that
TE
ATF2 is a novel JNK substrate that promotes MPTP-induced dopaminergic neuronal
CE P
death.
Materials and methods
AC
Chemicals and reagents
MPTP·HCl and paraformaldehyde (PFA) were purchased from Sigma-Aldrich (Shanghai, China). The following primary antibodies were used in this study: anti-pJNK (Promega, Cat. #V7932), p-JNK T183/Y185 (G-7, Santa Cruz, Cat. #sc-6254), p-p38 T180/Y182 (Cell Signaling, Cat. #4511), p38 (Cell Signaling, Cat. #8690), pATF2 T69/71 (Cell Signaling, Cat. #9225), p-ATF2 T69/71 (Millipore, Cat. #05-891), p-ATF2 T71 (Cell Signaling, Cat. #9221), ATF2 (Abcam, Cat. #ab32160), ATF2 DNA-binding domain (ATF2-DBD; reported previously) (Ackermann et al., 2011), pc-Jun S63 (Cell Signaling, Cat. #9261), TH (Millipore, Cat. #AB1542), TH (Millipore, Cat. #AB152), CD11b (Millipore, Cat. #MAB1387Z), and Iba-1 (Wako, #019-19741). 4
ACCEPTED MANUSCRIPT Normal donkey serum, bovine serum albumin (BSA, IgG-free, protease-free) and donkey anti-mouse IgG (H+L) Fab fragments were purchased from Jackson
T
ImmunoResearch. The ABC kit and DAB kit are from Vector Laboratories.
IP
Animals and treatments
SC R
All animal experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Sun Yat-sen University. The mice were housed in rooms with controlled 12 h light/dark cycles, temperature, and
NU
humidity, and food and water were provided ad libitum. Male C57BL/6J mice were
MA
obtained from the Experimental Animal Center of Sun Yat-sen University. The floxed Atf2 mice backcrossed to C57BL/6 for at least 10 generations were previously
D
reported (Ackermann et al., 2011). The dopamine transporter (DAT)-Cre mutant mice
TE
backcrossed to C57BL/6 for eleven generations and the JNK1-, JNK2- and JNK3deficient mice backcrossed to C57BL/6 for six generations were obtained from The
CE P
Jackson Laboratory (stock numbers 006660, 004319, 004321, 004322). All animal treatments were performed as previously described (Wang et al., 2004;
AC
Wang et al., 2007). In brief, 8-12-week-old male mice weighing 22-28 g were treated with MPTP [30 mg/kg of free base, intraperitoneally (i.p.)] at 24-hour intervals for 5 consecutive days. The control animals received a corresponding volume of saline alone. The mice were sacrificed at the indicated time points. For immunofluorescent or immunohistochemical analysis, the mice were anesthetized with chloral hydrate (400 mg/kg, i.p.) and then perfused with ice-cold phosphate-buffered saline (PBS; 0.01 M, pH 7.4) for 3 min, followed by 4% paraformaldehyde (PFA) for 8 min, at a flow rate of 10 ml/min. The brains were removed and postfixed overnight in 4% PFA at 4°C and then cryoprotected in 20% and 30% sucrose. For western blots, the mice were anesthetized, perfused with ice-cold saline containing 1 mM Na3VO4 and 20 5
ACCEPTED MANUSCRIPT mM NaF, and then sacrificed by cervical dislocation. The brains were washed with ice-cold PBS and the ventral midbrains were rapidly dissected on ice.
T
Immunofluorescence
IP
Coronal sections (20 μm) throughout the entire midbrain were cut using a cryostat
SC R
(Leica CM1950). Free-floating tissue sections from each of the treatment groups and time points were immunostained in parallel using the same solutions to minimize variation between experiments. The sections were pre-incubated in blocking solution
NU
containing 5% normal donkey serum and 0.3% Triton X-100 in 50 mM Tris-buffered
MA
saline (TBS, pH 7.4) at room temperature for 1 h. For endogenous IgG blocking, the sections were then incubated in Fab fragments (250 μg/ml) for 1.5 h at room
D
temperature and washed three times. Primary antibodies were diluted in 1% BSA and
TE
0.3% Triton X-100 in TBS (pH 7.4) and were applied overnight at 4°C. After three washes in TBS with 0.05% Tween 20, the sections were incubated with Alexa 488-,
CE P
Alexa 555- or Alexa 647-conjugated secondary antibodies diluted in the same buffer as the primary antibodies for 1 h at room temperature. All sections were
AC
counterstained with Hoechst 33258. Following three additional washes, the sections were coverslipped with mounting media and visualized with an inverted confocal laser scanning microscope (LSM 780, Carl Zeiss, Germany). Sections with the primary antibody omitted were processed and photographed under the same conditions and used as negative controls. Immunofluorescence quantification was performed using ImageJ software. The SNpc was defined as Regions of Interest (ROI) using Polygon selection tool and the integrated optical density (IOD) of the color channal of target protein was measured. Backgrounds of images from different sections were set the same, and IOD was measured with the same threshold. The IOD in saline-treated control mice is defined as one. 6
ACCEPTED MANUSCRIPT Immunohistochemistry Coronal sections (40 μm) encompassing the entire midbrain and striatum were
T
serially collected. Sections containing the striatum and the SNpc were processed for
IP
TH staining. The sections were incubated with rabbit anti-TH antibody (1:10 000) and
SC R
then visualized with the Vectastain Elite ABC kit (Vector Labs, #PK-6101) and the DAB peroxidase substrate kit (Vector Labs, #SK-4100) following the manufacturer’s protocol. Adjacent SNpc sections were used for Nissl-staining to evaluate the survival
NU
of nigral neurons.
MA
Cell counts
Nigral dopaminergic cell were stereologically assessed using the optical
D
fractionator with the aid of Stereo Investigator (MicroBrightField Inc; Williston, VT,
TE
USA). Nigral dopaminergic groups were outlined on the basis of TH immunolabeling, with reference to a standard mouse brain atlas (Paxinos and Franklin, 2001) and the
CE P
substantia nigra cell group described in (Fu et al., 2012). TH-positive neurons with clearly identified nuclei were counted in both hemispheres of the SNpc of every
AC
fourth section throughout the entire extent of the midbrain (AP -2.7 to AP -3.8 mm). The numbers of TH- and Nissl-possitive cells were counted by investigators blinded to the treatments and genotypes as described (Baquet et al., 2009). Striatal Densitometry The density of striatal DA terminals was measured as the IOD of the striatal THimmunoreactive signal using ImageJ software. Four sections were randomly selected from those containing the striatum at the approximate level of Bregma-0.22 to 1.10 mm according to the mouse brain atlas (Paxinos and Franklin, 2001), and the IOD in the dorsal striatum of each section was measured on each side. In each section, the
7
ACCEPTED MANUSCRIPT optical densities were corrected by subtracting out the background staining in the corpus callosum.
T
Striatal 1-methyl-4-phenylpyridinium ion (MPP+) levels
IP
The measurement of MPP+ levels in the various genotypes was conducted
SC R
according to a previously published protocol (Jackson-Lewis and Przedborski, 2007). Mice were sacrificed at 90 min after one i.p. injection of 30 mg/kg MPTP, and the striata of both sides were dissected and processed for high-performance liquid
NU
chromatography (HPLC) with ultraviolet (UV) detection (wavelength 295 nm).
MA
Western blot
The ventral midbrain tissue was weighed, quickly homogenized in Laemmli buffer
D
(50 μl/mg tissue) containing Tris·Cl (62.5 mM, pH 6.8), SDS (2%, w/v), Butterfield’s
TE
phosphate buffer (BPB; 0.005%, w/v), glycerol (10%, v/v) and DTT (8 mg/ml), and then boiled at 95-100°C for 5 min. The protein lysates (30 μg) were then resolved on
CE P
10% SDS-PAGE gels, electrotransferred to PVDF (Roche) membranes, and probed with primary antibodies at 4°C overnight, followed by the appropriate horseradish
AC
peroxidase-conjugated secondary antibodies. All blots were visualized using ECL chemiluminescence (Amersham Biosciences). Quantification was performed using ImageJ software.
Statistical analysis The statistical analyses were performed using GraphPad Prism version 6.01 for Windows (GraphPad Software, La Jolla, California, USA; www.graphpad.com). All data are presented as the mean ± standard error of the mean (SEM). Unpaired twotailed t-tests were used for comparisons between the saline- and MPTP-treated groups. When there were two variables (genotype and treatment), two-way ANOVA followed by Tukey’s multiple comparisons test was used. 8
ACCEPTED MANUSCRIPT
Results
clarify
whether
and
where
JNK
is
activated,
IP
To
T
JNK is activated predominantly in dopaminergic neurons in MPTP-treated mice. we
performed
SC R
immunofluorescence with a rabbit anti-JNK antibody. As shown in Fig. 1A, faint staining of phospho-JNK was apparent in dopaminergic neurons of the SNpc in the
NU
saline-treated group. An increase in phospho-JNK was visible as early as 6 hours after administering the first dose of MPTP administration. Consecutive treatments with
MA
MPTP led to the lasting JNK activation, which peaked after five doses of MPTP injection (Fig. 1A, B). At high magnification, the phospho-JNK signal was located
D
almost exclusively in the nuclei of dopaminergic neurons (Fig. 1A inserts). The
TE
specificity of this rabbit anti-phospho-JNK antibody has been confirmed
CE P
(Supplemental Fig. 1A-C). This observation was further supported by western blot analysis (Fig. 1C).
To explain the significant differences between our and others’ findings
AC
(Karunakaran et al., 2008), we performed double immunostaining using the same mouse anti-phospho-JNK antibody (G-7, Santa Cruz). We did find increased signals in microglial cells, predominantly at the microglial surface, but this pattern of signals did not diminish upon omitting the primary antibody in the negative controls (Supplemental Fig. 2). We assumed that these increased signals were due to the antimouse IgG secondary antibody recognizing endogenous IgG on the surface of microglia, which increased in concert with microglial activation. Therefore, we used anti-mouse IgG Fab fragments to block endogenous mouse IgG. As shown in Supplemental Fig. 2, the increased signal on the surface of microglia was completely 9
ACCEPTED MANUSCRIPT blocked. These results confirm that JNK is activated mainly in the nuclei of dopaminergic neurons after MPTP treatment, suggesting that the nuclear pathway is
T
involved in JNK-dependent dopaminergic neurodegeneration.
IP
ATF2 Thr69 is phosphorylated within dopaminergic neurons after MPTP
SC R
intoxication.
The transcription factor c-Jun is a classic substrate of JNK, showing expected
NU
activation upon MPTP treatment (Fig. 2A). Our previous work has demonstrated that the dimerization of activated c-Jun and ATF2 is necessary for JNK-dependent
MA
neuronal death (Yuan et al., 2009). Therefore, we wondered whether ATF2 is also activated in parkinsonian models. To explore this possibility, we first observed the
D
phosphorylation levels of ATF2 Thr69/71 (human homolog corresponding to
TE
Thr51/53 of murine ATF2), which is essential for its transcriptional activity (Gupta et
CE P
al., 1995; Livingstone et al., 1995; van Dam et al., 1995), and found an increase in the phosphorylation of ATF2 Thr69/71 (Fig. 2B). A time-course analysis showed a timedependent increase of phospho-ATF2 Thr69/71, which peaked after the last dose of
AC
MPTP and was localized exclusively to the nuclei of TH-positive neurons in the ventral midbrain of the MPTP-treated mice (Fig. 2B, D). The increase of phosphoATF2 was temporally and spatially consistent with that of phospho-JNK (Fig. 1A, B). To confirm this result, we adopted another acknowledged phospho-ATF2 antibody that has been shown to specifically recognize Thr71-monophosphorylated ATF2 (corresponding to murine homolog Thr53). The specificity of both phospho-ATF2 antibodies has been demonstrated by previous studies (Ackermann et al., 2011; Morton et al., 2004; Ouwens et al., 2002). To our surprise, we found that phosphoATF2 Thr71 was constitutively expressed in dopaminergic neurons of control mice 10
ACCEPTED MANUSCRIPT and remained unchanged after MPTP-treatment (Fig. 2C, E). This observation was further supported by western blot analysis (Fig. 2F). Given that the phospho-ATF2
T
Thr69/71 antibody recognizes Thr69/71-dual-phosphorylated but not mono-
IP
phosphorylated ATF2 (Morton et al., 2004; Ouwens et al., 2002), we therefore
SC R
concluded that phospho-ATF2 Thr71 is constitutively expressed and that phosphoATF2 Thr69 is markedly increased in dopaminergic neurons after MPTP intoxication. Therefore, the upstream kinases that could phosphorylate the Thr69 of ATF2 may be
NU
highly valuable for the ATF2 activation in the MPTP PD model.
MA
JNK2/3 mediates the phosphorylation of ATF2 Thr69 during MPTP-induced neurodegeneration.
D
ATF2 Thr69 can be phosphorylated not only by JNK but also by p38 (Raingeaud et
TE
al., 1995). Although we found that ATF2 activation is spatially and temporally
CE P
accompanied by that of JNK (Fig. 1A,B; Fig. 2B, D), other evidence has shown that p38 was also activated in dopaminergic neurons by MPTP treatment (Karunakaran et al., 2008). To investigate whether p38 is activated in addition to JNK, we measured
AC
phospho-p38 levels in parallel with phospho-JNK in situ, using an anti-phospho-p38 antibody whose specificity has been confirmed (Supplemental Fig. 1D, E). Basal expression of phospho-p38 was detected in the nuclei of dopaminergic neurons in the SNpc (Fig. 3A, B). However, no increase was observed after consecutive doses of MPTP treatment in dopaminergic neurons (Fig. 3A, B). Results from western blot analysis were in consistant with those from immunofluorescence (Fig. 3C). These results clarify that JNK not p38 is activated in dopaminergic neurons of the SNpc after MPTP exposure. Thus, the increase in phospho-ATF2 Thr69 is very likely due to JNK activation. 11
ACCEPTED MANUSCRIPT To further confirm whether JNK phosphorylates ATF2 Thr69 in dopaminergic neurons after MPTP intoxication, we introduced JNK2/3 knockout mice because it
T
has been shown that JNK2 and JNK3 mediate the MPTP-associated stress response in
IP
dopaminergic neurons, whereas JNK1 does not (Hunot et al., 2004). In JNK2/3-
SC R
knockout mice, MPTP-induced c-Jun phosphorylation was completely blocked compared with that in wild-type mice (Fig. 4A). This finding suggests the complete loss of nuclear JNK activity in these compound mutants, whereas JNK1-deficiency
NU
has no effect on c-Jun activation (data not shown). Moreover, the increased signal
MA
detected by the ATF2 Thr69/71-dual-phosphorylation antibody was almost completely blocked in JNK2/3-null mice upon MPTP treatment (Fig. 4B, D), whereas
D
the expression level of Thr71-mono-phosphorylated ATF2 remained steady (Fig. 4C,
TE
E). Together, these results indicate that JNK2/3 mediates ATF2 activation in the MPTP model through the phosphorylation of Thr69, which is primed by the
JNK.
CE P
constitutively abundant phosphorylation of Thr71 by an undefined kinase other than
AC
The reduction in functional ATF2 alleviates MPTP-induced dopaminergic neurodegeneration. ATF2 activation in PD-related models has not been previously reported, and we were eager to evaluate whether it promotes dopaminergic neurodegeneration. For this purpose, we crossed DatCre/+ (Backman et al., 2006) mice with mice expressing a floxed allele of Atf2 (Atf2f/f) (Ackermann et al., 2011) to obtain dopaminergic neuronspecific ATF2-deficient mice. DatCre/+-induced recombination has been shown to be efficient and dopaminergic neuron-specific (Backman et al., 2006). The crosses led to the deletion of the DNA binding domain of ATF2 and thus caused the loss of 12
ACCEPTED MANUSCRIPT transcriptional function of ATF2. Immunofluorescence using an ATF2-DBD antibody (raised against the ATF2 DNA-binding domain) (Ackermann et al., 2011) showed
T
that DatCre/+ mediated the complete deletion of ATF2 in Atf2f/f;DatCre/+ (designated
IP
Atf2 homo) mice and a partial reduction of Atf2 in Atf2f/+;DatCre/+ (designated Atf2
SC R
het) mice (Fig. 5A).
We then used these ATF2 mutant mice and their control littermates
NU
(Atf2+/+;DatCre/+ and Atf2f/f;Dat+/+) to investigate the role of ATF2 in MPTP-induced dopaminergic neuronal death. A statistical analysis of the cell counts in the SNpc
MA
revealed that the basal TH-positive neuron numbers did not differ between the Atf2 het mice and control mice, but the numbers of residual TH-positive neurons after
D
MPTP treatment in Atf2 het mice were much greater than those in control mice
TE
(compared with the saline-treated group, MPTP induced a 48.2% loss of TH-positive cells in control mice and a 20.9% loss in Atf2 het mice), suggesting a protective effect
CE P
of the ATF2 knockout. In Atf2 homo mice, although a notable reduction (35.1% compared with control mice) in the basal numbers of TH-positive neurons was
AC
observed (which may be due to abnormal development of dopaminergic neurons; see discussion), the relative extent of the MPTP-induced loss of dopaminergic neurons was also alleviated (18.1% of TH-positive cells were lost in Atf2 homo mice; Fig. 5B), supporting the view that ATF2 activation promotes dopaminergic neurodegeneration. Nissl staining revealed similar trends and statistical results (39.4%, 18.7% and 10.8% losses in control, Atf2 het and Atf2 homo mice, respectively, Fig. 5B, lower right panel), suggesting the loss of actual TH-positive neurons instead of a loss of TH expression. As expected, we observed a significant loss of striatal TH-staining density in MPTP-treated control mice when normalized to the saline condition, and this was partially reversed in MPTP-treated Atf2 mutant mice (49.9%, 24.6% and 26.0% 13
ACCEPTED MANUSCRIPT reductions in control, Atf2 het and Atf2 homo mice, respectively; Fig. 5C). Similar changes of the nigral dopaminergic dendrites were also shown (Supplemental Fig. 3).
T
To determine whether the reduced loss of dopaminergic neurons or the decrease in
IP
axonal and dendritic density was due to the reduced production of MPP+ from MPTP,
SC R
we measured striatal MPP+. Indeed, we found similar levels of MPP+ in control, Atf2 het and Atf2 homo mice (Table 1). Taken together, these findings demonstrate that
NU
functional ATF2 is involved in MPTP-induced dopaminergic neuronal death.
MA
Discussion
In the present study, we used a subacute MPTP-induced PD model to investigate the
D
mechanism of dopaminergic neuronal degeneration in PD. Our principal findings
TE
were as follows: (i) we confirm that JNK, not p38, is predominantly activated in dopaminergic neurons of the SNpc; (ii) we demonstrate for the first time that ATF2
CE P
Thr69 phosphorylation is increased selectively in dopaminergic neurons by MPTP toxicity, whereas ATF2 Thr71 phosphorylation is constitutive and remains unchanged;
AC
(iii) we demonstrate that JNK2 and JNK3 mediate ATF2 Thr69 phosphorylation but have no effect on ATF2 Thr71 phosphorylation; and (iv) we demonstrate that ATF2 activation is involved in MPTP-induced dopaminergic neurodegeneration. Numerous studies have addressed the importance of JNK in mediating dopaminergic neurodegeneration in both sporadic and familial PD (Wang et al., 2012). However, direct evidence showing that phosphorylated JNK is increased in dopaminergic neurons of the SNpc is limited. Most of the previous studies used immunoblots to analyze levels of phospho-JNK in the SNpc or ventral midbrain tissue lysates in the MPTP-induced PD model (Boyd et al., 2007; Castro-Caldas et al., 2012; Chen et al., 2005; Karunakaran et al., 2007; Saporito et al., 2000; Sung et al., 2012; 14
ACCEPTED MANUSCRIPT Xia et al., 2001) and therefore could not distinguish between signals from neurons and those from other cell types. Until recently, when Karunakaran et al. reported that JNK
T
phosphorylation was predominant in microglia using a mouse antibody (Karunakaran
IP
et al., 2008), no in situ evidence in the MPTP-induced mouse model was available.
SC R
Here, we found that staining mouse tissue with a mouse antibody without blocking endogenous IgG could lead to false-positive results. We chose a rabbit anti-phosphoJNK antibody to avoid interference by endogenous murine IgG, showing that
NU
phosphorylated JNK is predominantly increased in dopaminergic neurons and is
MA
localized mainly in the nucleus after MPTP treatment. By using this rabbit antibody, we also observed activated JNK in a small number of microglial cells (data not
D
shown), which suggests a possible role for JNK in microglial activation and pro-
TE
inflammatory actions. Further studies are required to investigate the roles of JNK in different cell types during dopaminergic neurodegeneration.
CE P
We show for the first time that dual-phospho-ATF2 Thr69/71 significantly increases in dopaminergic neurons soon after MPTP injection, whereas mono-
AC
phospho-ATF2 Thr71 is constitutively expressed and remains unchanged, suggesting that Thr69 is a critical regulatory site for ATF2 activation in the MPTP model. By using JNK2/3 knockout mice, we demonstrated that JNK2/3 mediates the phosphorylation of Thr69 but not of Thr71. A two-step mechanism for ATF2 phosphorylation, in which Thr71 is phosphorylated first, thereby priming ATF2 for subsequent efficient Thr69 phosphorylation, has previously been reported (Ouwens et al., 2002). Whether Thr71-mono-phosphorylated ATF2 occurs constitutively in vivo and has functions other than acting as a priming site remains unknown. In the current study, we found that neurons in specific locations, including the SNpc, hippocampus and cortex, express high levels of Thr71-mono-phosphorylated ATF2 but negligible 15
ACCEPTED MANUSCRIPT basal expression of dual-phospho-ATF2 Thr69/71 (data not shown). As it has been pointed out that phosphorylation of Thr69 or Thr71 and transcriptional activation of
T
ATF2 is crucial for viability and development, reviewed in (Lau and Ronai, 2012), we
IP
believe that the physiological functions of Thr71-mono-phosphorylated ATF2 and its
SC R
upstream kinases in vivo may be worth investigating further.
In our study, both homozygous and heterozygous mutants of ATF2 alleviated the MPTP-induced dopaminergic neurodegeneration, revealing that ATF2 is one of the
NU
molecular targets of JNK that participate in dopaminergic cell death. However, Atf2
MA
homo mice, in which functional ATF2 was completely abolished, showed reduced numbers of dopaminergic cells in the SNpc in three-month-old mice. This observation
D
is also consistent with those of previous loss-of-function studies, in which ATF2-
TE
deficient mice demonstrate significant neurological abnormalities (Ackermann et al., 2011; Reimold et al., 1996), including a deficit of dopaminergic fibers from the SNpc
CE P
at E18.5 (Kojima et al., 2008). These results indicate that ATF2 plays a role in neuronal development, and this effect likely depends on the function of Thr71-mono-
AC
phosphorylated ATF2, as we speculate in the above paragraph. Studies of postmortem brain tissue from PD patients demonstrate that ATF2 expression is reduced in some of the remaining pigmented cells of the SNpc in comparison to those in normal tissues (Pearson et al., 2005). ATF2 downregulation can also be detected in other neurological diseases (Pearson et al., 2005) and in response to neuronal injury (Herdegen et al., 1997; Martin-Villalba et al., 1998; Robinson,
1996).
ATF2
activation
by
N-terminal
phosphorylation
and
heterodimerization promotes its ubiquitination-dependent degradation (Fuchs and Ronai, 1999). In the subacute MPTP PD model, we only detected ATF2 activation rather than decreased total protein, possibly because neuronal toxin-induced 16
ACCEPTED MANUSCRIPT dopaminergic neurodegeneration is an acute process in which the dopaminergic neurons die before ATF2 degradation can be detected. However, whether the long-
T
lasting activation of ATF2 leads to the increased degradation of this protein remains
IP
to be verified in tissue from PD patients.
SC R
Our data establish for the first time that ATF2 is a mediator of JNK-dependent cell death in dopaminergic neurons. We also confirm that JNK is activated in dopaminergic neurons, whereas p38 is not. Elucidation of the precise regulation of
NU
these MAP kinases and their downstream effectors in dopaminergic neurons and other
MA
cell types involved in neurodegeneration, such as microglia, will help us to establish
D
more effective targets and improve neuroprotective therapy for PD.
TE
Acknowledgments
This work was supported by the National Basic Research Program of China 973
CE P
Program (2011CB504105), the National Natural Science Foundation of China (U1201224, 81471289, 31371071, 81471290), and the Science and Technology
AC
Project of Guangzhou city (2014Y2-00500). There are no conflicts of interest to declare.
Figure Legends Fig. 1 JNK is activated in dopaminergic neurons in MPTP-treated mice. Mice were treated with 1, 3 and 5 doses of MPTP and sacrificed 6 h post injection (M1×6 h, M3×6 h, M5×6 h). (A-B) The detection of p-JNK T183/Y185 (green) in dopaminergic
neurons
(TH-positive,
red)
of
the
ventral
midbrain
by
immunofluorescence (A) and quantitative data are shown (B). The insert is a higher magnification of the area indicated by the small rectangle. Data are expressed as the 17
ACCEPTED MANUSCRIPT mean±SEM (B, n=5 per group). ** p
