Address correspondence and reprint requests to Dr Esther L. Sabban, Department of Biochemistry and Molecular Biology, Basic Sciences Building, New York Medical College, Valhalla, NY 10595, USA. E-mail: Sabban@nymc.edu
The locus coeruleus (LC) is a critical stress-responsive location that mediates many of the responses to stress. We used immunoblotting and immunohistochemistry to investigate changes in induction and phosphorylation of several transcription factors and kinases in the LC that may mediate the stress-triggered induction of tyrosine hydroxylase (TH) transcription. Rats were exposed to single or repeated immobilization stress (IMO) for brief (5 min), intermediate (30 min) or sustained (2 h) duration. Single IMO elicited rapid induction of c-Fos and phosphorylation of cyclic AMP response element-binding protein (CREB) without changing the expression of early growth response (Egr)1, Fos-related antigen (Fra)-2 or phosphorylated activating transcription factor-2. Repeated IMO triggered increased phosphorylation and levels of CREB along with transient induction of c-Fos and increased Fra-2 expression. Several mitogen-activated protein kinases were activated by repeated IMO, shown by increased phosphorylation of p38, c-Jun N-terminal kinase (JNK)1/2/3 and extracellular signal-regulated kinase (ERK1/2). ERK1 was the major isoform expressed, and ERK2 the predominant isoform phosphorylated. Repeated IMO elicited hyperphosphorylation of ERK1/2 selectively in TH immunoreactive neurons, with substantial nuclear localization. These distinct alterations in transcriptional pathways following repeated compared with single stress may be involved in mediating long-lasting neuronal remodeling and are implicated in the mechanisms by which acute beneficial responses to stress are converted into prolonged adaptive or maladaptive responses.
The locus coeruleus (LC) plays a central role in the response to stress and is rapidly activated by a variety of stressors, both intrinsic and extrinsic. Such activation is adaptive to survival from life-threatening situations and serves as a general alarm function. Ascending norepinephrine (NE) axons originating within the LC richly innervate numerous regions implicated in stress responses, including the extended amygdala, hippocampus and prefrontal cortex, areas considered critical in mediating the alertness, focus and many other cognitive and physiological changes necessary in dealing with stress (Foote and Aston-Jones 1995). Acute stress releases NE in the terminal fields of the LC projections leading to a depletion of NE and an increase in NE metabolites (Tanaka et al. 1982; Glavin et al. 1983; Smith et al. 1991). During repeated stress, brain NE levels are replenished by increased biosynthesis (Weiss et al. 1975; Kvetnansky et al. 1977).
The stress-triggered increase in TH protein in the LC coincides with increased transcription. This was shown by nuclear run-on assays of transcription (Osterhout et al. 2002) as well as TH promoter-driven reporter activity in transgenic mice following single and repeated immobilization (Serova et al. 1999), and by in situ hybridization with intron-specific probes following footshock (Chang et al. 2000). These three independent methods clearly demonstrate the importance of transcriptional activation for induction of TH in the LC.
The proximal TH promoter contains several sites that are involved in the regulation of TH transcription (Fung et al. 1992). The cyclic AMP (cAMP)/calcium response element (CRE/CaRE) is critical for basal expression of the TH gene as well as for induction by several factors including cAMP and calcium (Kilbourne et al. 1992; Kim et al. 1994; Nagamoto-Combs et al. 1997; Lewis-Tuffin et al. 2004). Chronic stress was shown to up-regulate several components of the cAMP pathway in the LC, including protein kinase A and adenylyl cyclase (Melia et al. 1994a) which, through phosphorylation of CRE-binding protein (CREB), can activate TH transcription. The activating protein (AP)-1 site within the TH gene proximal promoter is essential for the response of the gene to a number of stimuli. Stimulation of the TH gene is often associated with induction of the AP-1 transcription factor c-Fos (Icard-Liepkalns et al. 1992; Sun and Tank 2003). Numerous studies have shown that stress induces expression of c-Fos and activation of AP-1 transcriptional activity (Pacak et al. 1992; Melia et al. 1994b; Chen and Herbert 1995). Other Fos family members, such as Fos-related antigen (Fra)-2, are involved in the response to repeated stress in the adrenal medulla and may also play a role in the LC (Nankova et al. 2000). The TH promoter also contains an early growth response (Egr)1 site overlapping with an specificity protein (Sp)1 site. Although not previously examined in the LC, Egr1 is induced in the adrenal medulla by stress and can modulate TH gene transcription by interacting with AP-1 factors (Papanikolaou and Sabban 1999).
The mitogen-activated protein kinases (MAPKs) regulate several of these factors, including CREB and AP-1 factors. Our earlier work indicated that c-Jun N-terminal kinase (JNK) is activated in the LC with repeated immobilization stress (IMO); however, a recent study indicated that a single episode of restraint stress did not result in phosphorylation of JNK or p38, although it led to a marked activation of extracellular signal-regulated kinase (ERK) (Shimizu et al. 2004).
The present study examines, in parallel, the changes in several signaling pathways and transcription factors in the LC that can be involved in the regulation of TH gene expression following different periods of single and repeated IMO. The findings demonstrate differences in activation or expression of transcription factors and kinases, depending on the repetition of the stress, and implicate a variety of MAPK pathways in mediating potentially long-lasting neuronal remodeling and plasticity in response to prolonged repeated stress.
Materials and methods
All animal procedures were approved by the Institutional Animal Care and Use Committee in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication no. 86–23). Male, Sprague–Dawley rats (250–275 g) obtained from Taconic Farms (Germantown, NY, USA) were housed three to four per cage at 24°C in a humidity-controlled room, on a 12-h light–dark cycle (lights on at 06.00 hours and off at 18.00 hours) in a barrier area to minimize background stress. Food and water were available ad libitum.
IMO was accomplished by taping the forelimbs and hind limbs with surgical tape to metal mounts attached to a board as described previously (Kvetnansky and Mikulaj 1970; Nankova et al. 1994). In the acute stress condition (1 × IMO), animals were immobilized for one session of 5 min, 30 min or 2 h and killed immediately afterwards or 24 h later (1 × + 24 h). In the repeated stress conditions, animals were immobilized for 2 h daily on consecutive days before the final day (day 2 for 2 ×ΙΜΟ, day 6 for 6 ×ΙΜΟ), on which the rats were immobilized for a session lasting 5 min, 30 min or 2 h. In some experiments, rats were immobilized for 2 h on each of 5 consecutive days and killed 24 h later without additional exposure (5 × + 24 h). Control groups were not exposed to stress.
Following the last period of IMO, rats were killed by decapitation. For immunoblots the brain was dissected using a tissue slicer with a digital micrometer. The LC was punched from frontal sections, 9.2–10.4 mm from Bregma, and frozen in liquid nitrogen. For immunohistochemistry brain hemisections were submersed in ice-cold fixative (4% paraformaldehyde with 0.05 m sodium orthovanadate). Brains were cryoprotected in 15% phosphate-buffered sucrose for 24 h at 4°C and then in 30% phosphate-buffered sucrose for 24 h at 4°C, after which brains were frozen on dry ice.
Frozen LC tissue samples from individual rats were sonicated in 200 µL of hot (90–95°C) 1% sodium dodecyl sulfate. Aliquots of the homogenate were used to determine protein concentration (BCA Protein Assay; Pierce, Pierce Rockland, IL, USA). Equal amounts of protein from each sample were separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (Owl Separation System; Portsmouth, NH, USA) and transferred to nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA, USA). Each gel contained samples from three control animals.
The blots were probed with antibodies corresponding to total or phospho-specific proteins of interest: anti-rabbit TH (1 : 5000; generous gift from Dr John Haycock, Louisiana State University Health Sciences Center, New Orleans, LA, USA); anti-rabbit Fra-2 and anti-rabbit c-Fos (1 : 1000; Santa Cruz Biotech, Santa Cruz, CA, USA); anti-rabbit CREB and anti-rabbit phospho-CREB (1 : 1000; Upstate, Lake Placid, NY, USA); anti-rabbit ERK1/2, anti-rabbit phospho-ERK1/2, anti-rabbit p38, anti-rabbit phospho-p38, anti-rabbit phospho-ATF-2, anti-rabbit JNK and anti-rabbit phospho-JNK (1 : 1000; Cell Signaling Technologies, Beverly, MA, USA), anti-mouse β-actin (1 : 30 000; Sigma, St Louis, MO, USA) and anti-mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1 : 5000; Imgenex, San Diego, CA, USA). The membrane was blocked by incubation in Tris-buffered saline (TBS) solution containing 3% dry milk and 0.1% Tween 20. After incubation for 24 h with the primary antibodies at 4°C according to the manufacturer's instructions, the filter was washed for 30 min three times in TBS/0.1% Tween 20 before the addition of the specific horseradish peroxidase-conjugated secondary antibody (1 : 2000–1 : 10 000). The membrane was incubated for 1 h and washed for 10 min three times in TBS/0.1% Tween 20.
Detection was accomplished by enhanced chemiluminescence. Protein loading was assessed by re-probing the blots with β-actin or GAPDH antibody. Because the ratio between both β-actin and GAPDH is unaltered by stress, we used β-actin as our loading control for all experiments. Before re-probing, blots were stripped with Restore Western Blot Stripping Buffer (Pierce), and the membranes were washed extensively with TBS/0.1% Tween 20 (TBSTx) and blocked for 1 h in 2% bovine serum albumin or 3% non-fat dry milk before immunodetection.
The optical density and total area of the immunoreactive bands was measured utilizing Image Pro Plus analysis software (version 4.1; Media Cybernetics, Silver Spring, MD, USA). The linear range for protein and phosphoprotein immunoreactivity in the immunoblots was determined for all antibodies by using increasing amounts of protein samples. The immunoreactivity was found to increase in a linear manner with amounts of total protein ranging from 5 to 60 µg (data not shown). Subsequent experiments were performed using 10–30 µg total protein. To control for differences in exposure, blots were standardized to the mean of the control optical density values on each respective blot.
Free-floating serial sections (35 µm), cut on a cryostat, were washed for 30 min in TBS (pH 7.4). The position of the LC was controlled between groups by selecting sections in which the cell body region had the broadest width and were directly adjacent to the fourth ventricle; this occurred systematically at the same coordinate with respect to Bregma, − 9.68 mm, according to the atlas of Paxinos and Watson (1986). The sections were permeabilized with TBSTx and preincubated for 60 min with 3% bovine serum albumin in TBSTx, then incubated overnight at 4°C with primary antibody in TBSTx containing 3% bovine serum albumin. The primary antibodies used in the present studies were anti-mouse TH (1 : 1000; Immunostar, Hudson, WI, USA), anti-rabbit phospho-ERK1/2 (1 : 100; Cell Signaling Technologies), anti-rabbit phospho-CREB (1 : 200; Upstate), anti-rabbit Egr1, anti-mouse c-Fos and anti-mouse phospho-ERK (1 : 100; Santa Cruz Biotech). After TBSTx rinses, the sections were incubated overnight with fluorescently labeled secondary antibodies, anti-mouse Cy2 or Cy3 and anti-rabbit Cy5 (1 : 200; Jackson ImmunoResearch, West Grove, PA, USA) or anti-mouse Alexa 488 and anti-rabbit Alexa 594 (1 : 200; Molecular Probes, Eugene, OR, USA). Negative control experiments included omission of the primary antibody and substitution of an equivalent dilution of normal serum. Following rinses, the sections were mounted on gelatin-coated slides, dehydrated, and cover slipped with DPX mountant (VWR International, West Chester, PA, USA).
It is important to note that the modified immunohistochemical procedure used here allowed the localization of phosphorylated proteins within the brain tissue containing the LC. Immediate fixation of the brain tissue in 4% paraformaldehyde with a phosphatase inhibitor, rather than by the traditional method of perfusion, provided distinct advantages important to the study of phosphoprotein signaling. The post-translational modifications of the proteins of interest were preserved and this technique enabled examination of changes at shorter time points following the stress than is possible with perfusion. In addition, this procedure did not require anesthesia, which might interfere with the experimental findings.
The control and experimental sets contained between six and 12 animals per group. Most experiments were performed twice. For western blots, each data point was from an individual animal and quantification of the bands was performed by calibrated scanning densitometry (Model GS-800; Bio-Rad Laboratories) under conditions in which there was a linear response. The approximate molecular masses were estimated using calibrated prestained standards (Bio-Rad Laboratories). Data are presented as representative western blots and graphs depict the mean + SEM values obtained from densitometric analysis. Where statistical comparisons were made between two data sets, a paired two-tailed Student's t-test was used. In addition to comparisons of experimental groups and controls, intra-group and inter-group comparisons were made. One-way anova was used to examine intra-group significance, with a Tukey post-test examination. Two-way anova was used to examine inter-group significance. Results were considered significant for p≤ 0.05. Statistical analyses were performed using Prism 4 (Graph Pad Software, Inc., San Diego, CA, USA).
Repeated IMO leads to up-regulation of TH protein
The induction of TH protein following IMO was investigated by immunohistochemistry and immunoblot analysis. Immunostaining of TH in sections containing the LC are shown in Fig. 1(a). TH staining was detected in cell bodies and processes. Immunofluorescence detection revealed a marked increase in TH in the LC of animals following repeated stress for 2 (2 × IMO) or 6 (6 × IMO) consecutive days compared with controls.
Changes in TH at the different times of immobilization were examined by quantifying immunoblots (Fig. 1b). There was little change in TH protein levels within the 2 h of a single IMO stress, whereas following repeated IMO for 2 or 6 days resulted in a substantial increase in TH protein within the LC (4–6-fold vs. controls). The increased levels of TH were sustained at all times of the repeated immobilization. Interestingly, a significant up-regulation in the TH was observed 24 h after a single IMO for 2 h (the 1 × + 24 h time point). Thus, the induction of TH at the 2 × 5′ time point resulted from increased TH expression occurring subsequent to the first immobilization, as evidenced by the significant increase measured at 24 h after a single immobilization.
IMO-induced transcription factor activation
As transcriptional activation is likely involved in the IMO-induced increase in TH protein in the LC, we examined the time course of activation or induction of several transcription factors known to regulate TH transcription. Previous studies revealed at least three major regulatory elements on the TH promoter – AP-1, Egr1/Sp1 and CRE/CaRE motifs – that can interact with transcription factors modulated by a number of signaling pathways (Fig. 2), which may be involved in mediating the stress response in the LC.
The AP-1 factor c-Fos was found to be significantly induced within the LC with both acute and repeated IMO (Fig. 3). The levels of c-Fos protein were rapidly induced; even 5 min of a single immobilization led to a 50% induction. At the other 1 × IMO time points measured, c-Fos levels were increased over 2-fold. Exposure to repeated immobilizations further enhanced levels of c-Fos such that the levels were significantly increased compared with those following single IMO. Significantly increased levels of c-Fos, compared with control, were observed with IMO for 2 and 6 consecutive days. On the second day of repeated IMO, the levels of c-Fos were high after 5 min, and significantly higher after 30 min or 2 h of the stress. With 6 × IMO, levels of c-Fos were significantly greater than those in control and 1 × IMO animals, but less than those in the 2 × ΙΜΟ group.
In order to determine whether or not the increase observed at 5 min of stress on the sixth day represented a rapid response to that stress, or reflected a sustained increase from the previous day, we examined the levels in a group of animals that were not exposed to stress on the sixth day (Fig. 4). A day after 2 h of IMO for 5 consecutive days (5 × + 24 h), the c-Fos level had declined to near control values. However, when a parallel group of animals was exposed to 5 min of stress (6 × 5′), there was a rapid induction of c-Fos. As the animals in these two groups were treated identically except for a 5-min exposure to stress on the sixth day, the results indicate that the increase in c-Fos following repeated stress was not merely a sustained augmentation, but rather a more rapid/immediate response to the latest stressful encounter.
Although c-Fos was augmented by stress of all durations examined, levels of Fra-2 were increased only with repeated stress (Fig. 5). Immunoblot analysis revealed that Fra-2 levels were near basal values until the rats were stressed for 2 h on the second day. Six consecutive days of IMO produced a substantial increase in Fra-2 protein to more than three times control levels with all durations of IMO.
We examined the effect of IMO on Egr1 protein levels in the LC. Egr1, which is induced in the adrenal medulla in response to single and repeated IMO, has been proposed to interact with AP-1 factors in the regulation of TH transcription (Papanikolaou and Sabban 1999, 2000; Nakashima et al. 2003). We examined the levels of Egr1 in the LC following single and repeated IMO. Immunoblots showed that Egr1 levels were very low in the LC of control and immobilized animals (data not shown). Immunohistochemistry revealed that Egr1 protein increased with repeated stress but not within the LC. As seen in Fig. 6, neurons in the mesencephalic trigeminal nucleus, adjacent to the LC, were labeled with an anti-Egr1 antibody and these cells did not overlap with the TH-positive neurons.
In addition to AP-1 factors and Egr1, IMO was found to trigger phosphorylation of the transcription factor CREB in the adrenal medulla (Sabban et al. 2004). CREB has emerged as an important point of extracellular signaling convergence, especially in the nervous system (reviewed by Gonzalez and Montminy 1989; Shaywitz and Greenberg 1999; Lonze and Ginty 2002; Quinn 2002). Phosphorylation of CREB at Ser133 is necessary for CREB activation. To examine the activation of CREB in the LC, we used an antiserum that recognized the phosphorylation at Ser133. The time course of change in the amount of phosphorylated CREB protein was determined by immunoblot analysis (Fig. 7). A single 30-min session of IMO produced a greater than 2-fold increase in the phosphorylation of CREB within the LC. Increased phosphorylation of CREB (∼2 fold) was observed following a single 2-h IMO. Upon exposure of rats to a second period of IMO, the level of phosphorylated CREB was further increased at 5 min and remained high throughout the immobilization, but was significantly lower than that seen at 5 min. In the 6 × IMO group, increased levels of phospho-CREB were observed with immobilization for 5 and 30 min, but not with IMO for 2 h.
An increase in the levels of phospho-CREB may result from increased phosphorylation of existing CREB or an IMO-induced increase in CREB protein levels. CREB protein was therefore measured in LC homogenates using an antibody that recognizes the total amount of protein regardless of post-translational modifications (Fig. 8). Within 30 min of single IMO stress, a 2-fold increase in CREB was measured. With increasing duration and increasing repeated exposure to stress, levels of CREB increased to more than 3-fold control levels.
The stress-related induction of phospho-CREB was localized primarily to TH-immunoreactive neurons as shown by co-immunofluorescence (Fig. 9). In unstressed rats, neurons within the LC reacted with antibodies to TH (Fig. 9a) and phospho-CREB (Fig. 9b). When the images were merged (Fig. 9c), phosphorylated CREB was found in TH-positive neurons within the LC as well as in adjacent areas. Levels of both TH (Fig. 9d) and phospho-CREB (Fig. 9e) immunofluorescence were also examined in rats exposed to repeated stress for 6 days. In contrast to the control group, the phospho-CREB appeared to be in noradrenergic neurons (determined by TH immunoreactivity) (Fig. 9f). Interestingly, repeated stress resulted in an up-regulation of phospho-CREB and c-Fos within many of the same neurons within the LC (Fig. 10).
Another CREB family member, ATF-2, was recently shown to induce TH transcription (Suzuki et al. 2002). This factor is a ubiquitously expressed member of the CREB family, which forms homodimers or heterodimers with other transcription factors of the ATF family, such as CREB, or the AP-1 family, such as c-Fos (Hai and Curran 1991; De Cesare et al. 1995; De Cesare and Sassone-Corsi 2000). In contrast to findings for CREB, no change was observed in the levels of phosphorylated ATF-2 protein following a single or repeated immobilization (Fig. 11).
Stress-triggered activation of kinases
The above experiments revealed that AP-1 factors and CREB, but not Egr1 and ATF-2, might mediate the response of the LC to immobilization. We therefore examined the effects of stress on upstream kinases that may be involved in mediating these changes. The phosphorylation of CREB and induction of c-Fos can be mediated by several upstream kinases including members of the MAPK family: ERK1/2, JNK1/2/3 and p38 (Deak et al. 1998; Bonni et al. 1999). A single IMO led to a modest change in phosphorylation of JNK, significant for JNK1 at 30 min and JNK2/3 at 5 min (Fig. 12). However, repeated IMO triggered a robust increase in phosphorylation of all isoforms of JNK. Phosphorylation of JNK2/3 was increased about 3-fold by 5 min or more of immobilization in the 2 × IMO group. The response of JNK1 was somewhat slower.
The phosphorylation state of MAPK p38 was not significantly altered by a single IMO, but did specifically respond to the repeated stress (Fig. 13). Five minutes of repeated stress (2 × or 6 × ΙΜΟ) produced more than a 3-fold increase in p38 phosphorylation. Surprisingly, levels of phospho-p38 showed a bimodal pattern; they declined to towards baseline with IMO for 30 min and increased after exposure for 2 h to the repeated stress.
The phosphorylation of CREB can be mediated by ERK1/2 activation through receptor tyrosine kinase receptors (see Fig. 2). Blockade of ERK1/2 phosphorylation with the mitogen-activated protein kinase kinase (MEK)1/2 inhibitor SL327 has been demonstrated to significantly reduce CREB phosphorylation (Davis et al. 2000). It was therefore important to determine whether single and/or repeated IMO led to ERK activation.
ERK1 was found to be the predominant isoform of ERK in the LC (Fig. 14a); however, ERK2 was the major phosphorylated form. Apart from a slight, but significant, activation of phospho-ERK2 at 5 min, a single episode of IMO did not noticeably alter ERK signaling in the LC (Fig. 14). In contrast, repeated stress produced significant changes in the levels of ERK1 and ERK2 phosphorylation. In the 2 × IMO group, following 30 min of stress, levels of both phosphorylated ERK isoforms were induced by 50–100%. Within 5 min of stress on the sixth consecutive day, phospho-ERK1 and phospho-ERK2 levels rose to over 250% of control. Aftet 30 min of IMO on the sixth day, the phosphorylation of ERK1 and ERK2 was over 4-fold that measured in unstressed animals. Levels remained significantly high throughout the sixth exposure to stress.
The increase in phospho-ERK1/2 following IMO was specifically localized to noradrenergic neurons of the LC using TH co-immunofluorescence. The phospho-ERK1/2-immunoreactive neurons within the LC of repeatedly stressed rats (2 × and 6 × ΙΜΟ) were increased in terms of the number of labeled neurons detected and the intensity of labeling compared with that in control tissue (Fig. 15). Activated ERK was observed in a subset of LC neurons and particularly localized to TH-immunoreactive cell bodies.
In the inactive state unphosphorylated ERK is anchored to MEK in the cytoplasm (Fukuda et al. 1997). Upon phosphorylation, ERK dimerizes and translocates to the nucleus (Fukuda et al. 1997; Khokhlatchev et al. 1998; Adachi et al. 1999). To characterize the subcellular regulation of ERK following repeated stress, phospho-ERK expression was examined using confocal microscopy (Fig. 16). After 2 h of IMO on the sixth day, phospho-ERK was found in both nuclear and non-nuclear regions of the noradrenergic neurons within the LC, with substantial levels in the nucleus. Interestingly, repeated stress resulted in an up-regulation of phospho-CREB and phospho-ERK within many of the same neurons (Fig. 17).
The present study showed that acute and repeated IMOs differentially regulate the activation of several transcription factors and kinases in the rat LC. Multiple MAPK pathways were activated, including ERKs, JNKs and p38. This was especially evident with repeated IMO. Among the transcription factors that can regulate TH, single exposure to IMO elicited induction of c-Fos and activation of CREB, without changing expression of Egr1, Fra-2 or phosphorylation of ATF-2. Repeated immobilizations continued to trigger activation of CREB, which was accompanied by increased CREB levels. Transient induction of c-Fos was also observed with repeated stress exposure. Furthermore, repeated IMO led to a significant induction of Fra-2.
The IMO-induced expression of TH was found to coincide with up-regulation and activation of several factors implicated in the regulation of TH gene transcription. TH protein levels were increased with repeated exposure to IMO stress, and sustained throughout successive IMO episodes. These data are consistent with previous reports from our laboratory and others showing the induction of TH gene expression in the LC with repeated exposure to stress (Melia et al. 1994a; Kvetnansky and Sabban 1998; Rusnak et al. 1998). Even a single immobilization triggered increased TH protein; this was not observed during the 2 h of stress but was evident the next day, probably reflecting the time for protein synthesis.
A single exposure to IMO triggered rapid induction of c-Fos. As an immediate early gene c-fos is used extensively as a marker for neural activation in response to stress stimuli in the brain. The induction of c-Fos expression in a number of brain regions has been shown to depend on the intensity and duration of exposure to an acute stress. Up-regulation of c-Fos in catecholaminergic neurons of the ventrolateral medulla (VLM) and the nucleus of the solitary tract (NTS) was shown following 15 min of restraint stress (Dayas et al. 2004). In the LC, c-Fos was increased after 30 min of mild (restraint) or severe (immobilization) stress and remained higher during the entire 4-h stress, whereas it had already declined in other brain areas (Chowdhury et al. 2000). Our results revealed that even 5 min of exposure to IMO elicited a significant increase in c-Fos.
In addition to modulating c-Fos, a single immobilization led to activation of CREB. CREB directly binds to the CRE site within the TH promoter. Many studies have demonstrated the crucial role of the CRE in basal as well as stimulated TH gene expression. For example, the TH gene can be regulated by the cAMP signaling pathway and raised calcium via the CRE (Kilbourne et al. 1992; Hiremagalur et al. 1993; Kim et al. 1994; Nagamoto-Combs et al. 1997; Lewis-Tuffin et al. 2004). CREB can be activated by different kinases including ERK1/2 and protein kinase C (Johannessen et al. 2004). Although we did not examine protein kinase C activation in this study, it was noted that phospho-ERK2 was increased after 5 min of a single IMO. Taken together these data suggest that a single session of IMO can result in the activation of the ERK pathway, which may lead to phosphorylation of CREB.
Exposure of rats to twice-repeated IMO activated the same pathways as activated by single IMO, with an additional increase in another Fos family member, Fra-2, and phosphorylation of the upstream kinases ERK1/2, JNK2/3 and p38. The level of Fra-2 was significantly increased in the LC following 2 h of immobilization on the second as well as the sixth day. It has also been shown by our laboratory to be induced in the adrenal medulla with repeated stress and is able to induce TH promoter activity (Nankova et al. 2000). Interestingly, the levels of Fra-2 mRNA are reportedly increased in post-mortem tissue from the LC of individuals with major depressive syndrome (Xiang et al. 2004), suggesting that increased Fra-2 gene expression may be linked to the sustained up-regulation of TH gene expression in both stress and depression.
Although c-Fos has been well established as a marker for neuronal activation by acute stress (reviewed by Palkovits 2002), the present study revealed that repeated exposure to IMO stress over 6 days also results in significant enhancement of c-Fos. These data are in good agreement with other studies that have demonstrated an up-regulation of c-Fos following repeated immobilization or restraint stress for 10–14 consecutive days (Chen and Herbert 1995; Stamp and Herbert 1999; Medeiros et al. 2003). Importantly, we showed for the first time that c-Fos was newly induced in response to the most recent exposure to the stress, as evidenced by the fact that levels of c-Fos declined to control values 24 h after the removal of stress (5 × + 24 h) and were re-induced following subsequent exposure to another IMO.
In addition to its effect on several transcription factors, repeated IMO activated several upstream kinases. The present study revealed that all of the major MAPKs were activated in the LC with repeated IMO. The phosphorylation of ERK1/2, JNK1/2/3 and p38 was significantly increased with IMO over 2 days. Phospho-ERK1/2 has been shown to activate c-Fos and Fra-2, which both interact with the AP-1 site of the TH promoter (Murphy et al. 2004). At the same time ERK1/2 and p38 are known to phosphorylate mitogen-stimulated kinase 1/2, which then phosphorylates CREB, and might lead to activation of c-Fos. These multiple pathways leading to the activation of CREB may result in the observed increase in c-Fos protein and ultimately TH gene transcription.
ERK1 was found to be the major ERK isoform expressed in the LC. In this regard, ERK1 and ERK2 have been shown to be differentially expressed and regulated in various brain regions (Ortiz et al. 1995). Although ERK1 was the major isoform in the LC, experiments with phospho-specific antisera indicated that ERK2 is the major phosphorylated form of ERK in this brain region. Phosphorylation of both ERK1 and ERK2 was increased markedly by repeated stress. Immunohistochemistry indicated that phospho-ERK was almost exclusively localized in the TH-positive cells of the LC following repeated stress, with substantial localization of in the nucleus in rats exposed to six daily repetitions of immobilization. This is consistent with translocation of ERKs from the cytoplasm to the nucleus so that activated ERKs can directly phosphorylate and stimulate specific transcription factors (Brunet et al. 1999). Similar results indicating an up-regulation of phosphorylated ERKs within the LC following restraint stress were recently reported (Shimizu et al. 2004). ERKs have also been shown to be phosphorylated in the hippocampus and amygdala in response to immobilization and swim stress (Sananbenesi et al. 2003; Shen et al. 2004).
Exposure of rats to repeated IMO increased not only the amount of phospho-CREB but also total CREB levels. These findings are similar to those observed using another physical stress paradigm, footshock prod (Bruijnzeel et al. 2001). The increase in total CREB levels suggests a regulatory change in the signaling apparatus in response to chronic stress, which may also be related to retention of stress-related memories. CREB plays a critical role in plasticity processes underlying learning and memory, and the levels of CREB protein are considered to be a limiting factor in memory foundation (Bartsch et al. 1998; Brodie et al. 2004).
This is the first study to show stress-induced activation of specific JNK isoforms and p38 within the LC. Repeated IMO induced activation of JNKs (1 and 2/3) and p38 to more than 300% of control levels within the LC. Previously, we found that JNK enzymatic activity was also increased following repeated immobilization (Nankova et al. 1998). Other physical stressors (restraint and forced swim) have also been shown increase the levels of phospho-JNK in the hippocampus, hypothalamus and amygdala of mice (Liu et al. 2004). Both the JNK and p38 MAPK pathways have been shown to be activated by inflammatory cytokines, brain injury and ischemic insult. It is important to note that over-activation of these stress-activated kinases is known to cause neuronal degeneration and impairment of the function of the CNS and may be involved in stress-related disorders (Oo et al. 1999; Schroeter et al. 2003).
Of particular interest is the finding that all JNK isoforms examined in this study were significantly phosphorylated with repeated IMO. The physiological actions of JNK isoforms differ in the brain. JNK3 is considered a degenerative signal transducer and efficient activator of apoptosis, based on the neuroprotective events following the inhibition of JNK3 in knockout mice following excitotoxicity (Waetzig et al. 2002), 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-triggered neurotoxicity (Hunot et al. 2004), ischemia (Kuan et al. 2003) and by JNK peptide inhibitors (Borsello et al. 2003). Conversely, the embryonic lethality of JNK1 and JNK2 double knockout mice indicates that these JNKs have fundamental physiological roles during development (Kuan et al. 1999).
The finding that many signaling pathways are activated in the LC, especially with repeated IMO, indicates that there is probably a broad change in transcriptional activation in response to stress. Changes in examined transcription factors, although considered here in the context of regulation of TH gene expression, are not restricted to this gene. Some of the other stress-related targets within the LC include the neuropeptides galanin and neuropeptide Y, the biosynthetic enzyme dopamine β-hydroxylase and receptors for growth factors such as TrkB, the high-affinity receptor tyrosine kinase for brain-derived neurotrophic factor.
Several of the signaling pathways activated in the LC are similar to those observed in the adrenal medulla to be affected by IMO. For example, induction of c-Fos and phosphorylation of CREB were significant even with a single relatively brief exposure to IMO (Sabban et al. 1995; Nankova and Sabban 1999; Sabban and Kvetnansky 2001). In addition, marked induction of Fra-2 was found with repeated IMO (Nankova et al. 2000). However, the intracellular response to stress in the LC is not identical to that of the catecholamine-synthesizing cells of the adrenal medulla. Of particular note, Egr1 is markedly induced in the adrenal medulla (Papanikolaou and Sabban 1999) and is thought to interact with AP-1 factors to regulate TH transcription (Papanikolaou and Sabban 2000; Nakashima et al. 2003). However, Egr1 was not induced by IMO in the TH-positive cells of the LC.
Induced and/or activated transcription factors and kinases are likely to be involved in mechanisms involved in plasticity of the noradrenergic system following exposure to stress. Stress-induced increase in TH mRNA and TH protein in the LC after single or twice repeated stress may reflect an adaptive regulatory response, serving potentially to increase the synaptic capacity of the noradrenergic system in anticipation of subsequent stress and to induce ability to invoke appropriate behavioral adaptation (Zigmond et al. 1995). Our results suggest that in this period c-Fos, CREB, and partially ERK1/2 and JNK, may be especially important in the transcriptional activation of TH gene expression. Activity of the LC is related to increased arousal and focused attention on salient events in threatening or demanding situations (Foote and Aston-Jones 1995; Robbins and Everitt 1995). The behavioral alterations may also be considered as adaptive adjustments in anticipation of changing demands. Activation of the brain noradrenergic system during stress probably plays an important integrative function in coping and adaptation by facilitating evoked synaptic transmission in many brain regions mediating specific behavioral and physiological processes comprising the stress response (Valentino et al. 1993; Aston-Jones et al. 1999; Ziegler et al. 1999). With prolonged stress, the response may reflect a relatively greater impact of stressful stimuli on noradrenergic neurons of the LC, resulting from long-term adaptation to stress or failure to initiate appropriate behavioral strategies, and may contribute to the susceptibility to stress-related pathology. This may be mediated by the induction of c-Fos, Fra-2, CREB, and a pronounced increase in activation of JNK, ERK and p38. The dependence of expression and/or activation of these transcription factors and kinases on the frequency of stress illustrates the involvement of multiple regulatory mechanisms that may underlie changes in the plasticity of neurons in the LC.
We gratefully acknowledge the support contributed by National Institutes of Health grant NS28869 and Office of Naval Research grant N0014-02-1-0315. We wish to thank Anne Sollas and Wael Yunis for their technical assistance, Dr Victoria Arango and Suham A Kassir (New York Psychiatric Institute and Columbia University, New York, NY, USA) for excellent guidance with the immunocytochemistry and Drs Ken Lerea and Chris Leonard for use of their fluorescence microscopes.