Stress-induced changes in epinephrine expression in the adrenal medulla in vivo

Authors

  • T. C. Tai,

    1. Department of Psychiatry, Harvard Medical School and the Laboratory of Molecular and Developmental Neurobiology, McLean Hospital, Belmont, Massachusetts, USA
    2. Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California, USA
    Search for more papers by this author
    • 1

      The present address of T.C. Tai is the Division of Medical Sciences, Northern Ontario School of Medicine, Laurentian University Campus, Sudbury, ON, Canada.

  • Robert Claycomb,

    1. Department of Psychiatry, Harvard Medical School and the Laboratory of Molecular and Developmental Neurobiology, McLean Hospital, Belmont, Massachusetts, USA
    Search for more papers by this author
  • Brenda J. Siddall,

    1. Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California, USA
    Search for more papers by this author
  • Rose Ann Bell,

    1. Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California, USA
    Search for more papers by this author
  • Richard Kvetnansky,

    1. Institute of Experimental Endocrinology, Slovak Academy of Sciences, Bratislava, Slovak Republic
    Search for more papers by this author
  • Dona L. Wong

    1. Department of Psychiatry, Harvard Medical School and the Laboratory of Molecular and Developmental Neurobiology, McLean Hospital, Belmont, Massachusetts, USA
    2. Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, California, USA
    Search for more papers by this author

Address correspondence and reprint requests to Dona Lee Wong PhD, Department of Psychiatry, Harvard Medical School, Laboratory of Molecular and Developmental Neurobiology, McLean Hospital, 115 Mill Street, MRC Room 116, Belmont, MA 02478, USA.
E-mail: dona_wong@hms.harvard.edu

Abstract

Immobilization (IMMO) stress was used to examine how stress alters the stress hormone epinephrine (EPI) in the adrenal medulla in vivo. In rats subjected to IMMO for 30 or 120 min, adrenal corticosterone increased to the same extent. In contrast, EPI changed very little, suggesting that EPI synthesis replenishes adrenal pools and sustains circulating levels for the heightened alertness and physiological responses of the ‘flight or fight’ response. In part, stress activates EPI via the phenylethanolamine N-methyltransferase (PNMT) gene as single or repeated IMMO elevated PNMT mRNA. The rise in PNMT mRNA was preceded by induction of the PNMT gene activator, Egr-1, with increases in Egr-1 mRNA, protein, and protein–DNA binding complex apparent. IMMO also evoked changes in Sp1 mRNA, protein, and Sp1–DNA complex formation, although for chronic IMMO changes were not entirely coincident. In contrast, glucocorticoid receptor and AP-2 mRNA, protein, and protein–DNA complex were unaltered. Finally, IMMO stress elevated PNMT protein. However, with seven daily IMMOs for 120 min and delayed killing, protein stimulation did not attain the highly elevated levels expected based on mRNA changes. The latter may perhaps suggest initiation of adrenergic desensitization to prolonged and repeated IMMO stress and/or dissociation of transcriptional and post-transcriptional regulatory mechanisms.

Abbreviations used
AdoMet

S-adenosylmethionine

CORT

corticosterone

DA

dopamine

DOPAC

dihydroxyphenylacetic acid

ECL

enhanced chemiluminescent

EPI

epinephrine

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GMSA

gel mobility shift assay

GR

glucocorticoid receptor

IMMO

immobilization

NE

norepinephrine

PNMT

phenylethanolamine N-methyltransferase

RT-PCR

reverse transcription-polymerase chain reaction

Stress, whether environmental, physiological, or psychological, induces a sequelae of responses, beginning with cortisol and epinephrine (EPI) release into the circulation. These changes provide the heightened awareness to cope with the stress and constitute the ‘fight or flight’ response (Cannon and De La Paz 1911). Subsequently, homeostasis must be restored to ensure survival, and activation of hormonal and neural inhibitory pathways returns the stress hormones, cortisol and EPI, to basal values by suppressing their further synthesis and release.

Epinephrine is synthesized by N-methylation of norepinephrine (NE), a reaction catalyzed by phenylethanolamine N-methyltransferase (PNMT; EC 2.1.1.28), using as co-substrate and methyl donor, S-adenosylmethionine (AdoMet). As a regulator of EPI production, PNMT serves as a marker for adrenergic function. Hormones and neural stimuli have been shown to regulate the enzyme, particularly in response to stress (Viskupic et al. 1994; Wong et al. 1996, 2002b, 2004; Lee et al. 1999; Sabban and Kvetnansky 2001), with a potential target being the PNMT gene. Immobilization (IMMO) stress, 2-deoxyglucose and insulin rapidly increase PNMT mRNA levels in the adrenal medulla of rodents. Integrity of the splanchnic nerve is not required for stress-induced changes in adrenal PNMT mRNA, suggesting that neural pathways function independent of the hypothalamic–pituitary–adrenal axis to contribute to stress-elicited PNMT transcriptional activity (Baruchin et al. 1993; Viskupic et al. 1994; Wong et al. 1996, 2002b, 2004; Sabban and Kvetnansky 2001).

The PNMT gene can be activated by several transcription factors, including Egr-1 (Ebert et al. 1994; Morita et al. 1995, 1996; Tai et al. 2001), Sp1 (Ebert and Wong 1995; Her et al. 1999, 2003), the glucocorticoid receptor (GR) (Ross et al. 1990; Tai et al. 2002); AP-2 (Ebert et al. 1998; Wong et al. 1998b), Myc-associated Zinc finger transcription factor (MAZ) (Her et al. 1999, 2003), and glial cell missing transcription factor (GCM) (Tai and Wong 2003). These transcriptional regulators independently stimulate the PNMT gene, and in addition, Egr-1, AP-2, and the GR interact cooperatively to increase PNMT mRNA beyond their separate contributions.

PNMT can be controlled post-translationally as well via glucocorticoid regulation of PNMT enzyme stability (Wong et al. 1985). Corticosteroids sustain sufficient levels of the AdoMet metabolic enzymes, S-adenosylmethionine synthetase and S-adenosylhomocysteine hydrolase to ensure AdoMet availability for PNMT catalysis, while AdoMet binding to PNMT prevents its degradation.

The following studies were undertaken to identify how stress controls EPI and thereby adrenergic function in vivo. We show that repeated IMMO continues to evoke a stress response equivalent to a single IMMO with adrenal corticosterone (CORT) similarly elevated. In the adrenal medulla, sufficient dopamine (DA) precursor is available for EPI synthesis. IMMO induces a rise in the mRNA and protein of the PNMT transcriptional activators Egr-1 and Sp1 prior to or coincident with induction of PNMT mRNA and likely also enhances Egr-1 and Sp1 phosphorylation to permit their greater interaction with respective DNA-binding elements in the PNMT promoter. Finally, changes in PNMT protein, for the most part, concur with those of PNMT mRNA and appear to alter EPI biosynthesis to ensure sufficient neurohormone to sustain adrenal medullary pools and elevated circulating levels evoked by stress and required for the physiological responses required to counter the stress.

Materials and methods

IMMO stress

Male Sprague–Dawley rats (175–200 g, n = 8–10/group), obtained from Charles River Laboratories (Wilmington, MA, USA), were acclimated to a (12 : 12) day : light cycle (lights on 7:00 a.m.) and provided food and water ad libitum. All animals, control and experimental, were handled daily in order to eliminate confounding stress responses due to handling. After 1 week, animals were subjected to a single IMMO stress for 30 or 120 min and killed immediately or up to 24 h later (Kvetnansky et al. 1971). Alternatively, rats were repeatedly IMMO-stressed for 30 or 120 min daily for 6 or 7 days and killed as described. Adrenal glands were removed, quick frozen in dry ice-cooled isopentane (Sigma-Aldrich Corp., St Louis, MO, USA) and segregated into left and right adrenals for mRNA or protein determination, respectively. Tissue was stored at −70°C until use. In some cases, adrenal medulla was dissected from cortex prior to experimentation. Animal procedures were consistent with recommendations in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and fully approved by the McLean Hospital Animal Care and Use Committee.

Adrenal CORT

The right adrenal gland was sonicated in 1 mL of 50 mmol/L Tris–HCl, pH 7.4 for 15 s and cell debris was removed by centrifugation at 17 000 g for 2 min. CORT levels were determined in triplicate for each sample by radioimmunoassay as described by manufacturer (ICN Biomedical Inc., Costa Mesa, CA, USA) using 100 μL of adrenal extract diluted 1 : 200, 1 : 800, and 1 : 1600.

Quantitation of adrenal DA, dihydroxyphenylacetic acid, NE, and EPI

Right adrenal glands were sonicated in 1 mL of 0.1 mol/L perchloric acid containing 0.1 mmol/L EDTA and protein was pelleted by centrifugation at 17 000 g for 2 min. Supernatant was removed and after resuspending the pellet in 1.0 N NaOH, protein was determined as described by Bradford (1976). Perchlorate supernatants were passaged through a 45-μm filter and analyzed for catecholamine content by reverse phase HPLC using electrochemical detection (Lindley et al. 1999). Briefly, an aliquot of supernatant was directly injected onto a C18 reverse phase analytical column (5 μm, 250 × 4.6 mm; Biophase ODS, BAS, West Lafayette, IN, USA) protected by a pre-column cartridge (5 μm, 30 × 4.6 mm; BAS) and catecholamines was electrochemically detected with an analytical electrode set at +0.72 V (BAS). After normalization to adrenal protein, DA, dihydroxyphenylacetic acid (DOPAC), NE, and EPI were expressed relative to values in adrenal medulla from unstressed control rats set to 100%.

PNMT and transcription factor mRNA

Total RNA was isolated from adrenal glands or adrenal medulla using TRI REAGENTTM (Sigma-Aldrich Corp.) as per manufacturer and PNMT mRNA and transcription factor mRNA was quantified by ribonuclease protection assay (Wong et al. 1992a) or radioactive reverse transcription-polymerase chain reaction (RT-PCR) as previously described (Her et al. 2003). For RT-PCR, total RNA from adrenal medulla was treated with DNaseI (1 U/2 μg of total RNA; Ambion Inc., Austin, TX, USA) for 30 min at 37°C, and then 1 μg of total RNA was reverse transcribed with StrataScriptTM (Stratagene, La Jolla, CA, USA). PCR was performed on 100 ng of reverse transcription product in a 20 μL volume containing 20 mmol/L Tris–HCl (pH 8.4), 50 mmol/L KCl, 1.5 mmol/L MgCl2, 200 nmol/L dNTPs, 0.2 μmol/L sense and anti-sense primers, 0.1 μCi [α-32P]dATP, and 2 U of Taq DNA polymerase (Promega Corp., Madison, WI, USA). 5′ and 3′ primers included: PNMT, 5′CAGCTTCTTGGAGGTCAACCTG3′ and 5′TTATTAGGTGCCACTTCGGGTG3′; Egr-1, 5′CTCAACAGGGCAAGCATACG3′ and 5′CTGACATCGCTCTGAATAACG3′; Sp1, 5′TGAAGGCCAAGTTGAGCTCCAT3′ and 5′TTTACCAGCAGCGGATCATCAG3′; AP-2, 5′CCTGTCGCTCCTCAGCTCCA3′ and 5′AGGGAACTCGGTTTCACACAC3′; GR, 5′GCTTACATCTGGTCTCATTCC3′ and 5′GACCTCTTGAAGGATTTGGAG3′; and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), 5′ATGCTGGTGCTGAGTATGTCG3′ and 5′CATGTCAGATCCACAACGGATAC3′. Reactions were incubated for 1 min at 94°C, 1 min at 61°C, and 1 min at 72°C in a PTC-200 DNA Engine Thermocycler (MJ Research Inc., Waltham, MA, USA), repeating these cycles 35 times for PNMT and transcription factors and 18 times for GAPDH. To ensure that conditions for RT and PCR amplification were executed within the linear range of assay, RT linearity curves for production of PNMT, transcription factor, and GAPDH cDNA were generated using varying amounts of total RNA isolated from adrenal medulla and varying times at 37°C. For PCR, linearity curves for generation of PNMT, transcription factor and GAPDH amplicons from their respective cDNAs were also executed, varying denaturation, annealing and amplification temperatures and number of amplification cycles. PNMT or transcription factor amplicons were combined with GAPDH amplicon and resolved on 5% polyacrylamide gels for autoradiography using Kodak X-Omat LS film (Fisher Scientific, Springfield, NJ, USA). For RT-PCR, semi-quantitative comparisons were performed. Autoradiographic signal intensities for PNMT, transcription factor and GAPDH amplicons were determined by computerized densitometry using NIH Image software (National Institute of Mental Health, National Institutes of Health, Bethesda, MD, USA), version 1.63f on a MacIntosh G5 computer (Apple, Cupertino, CA, USA) and a Hewlett Packard Scanjet 8200 (Hewlett-Packard Company, Palo Alto, CA, USA). The PNMT and transcription factor signals were normalized respective to GAPDH signals and then relative to control or one another as appropriate.

Cytosolic and nuclear protein extract

Cytosolic and nuclear proteins were isolated from adrenal medulla as previously described (Tai et al. 2001). Medulla was minced and tissue homogenized in 400 μL of 10 mmol/L HEPES–KOH (pH 7.9) containing 1.5 mmol/L MgCl2, 10 mmol/L KCl, 0.5 mmol/L dithiothreitol, 0.2 mmol/L phenylmethylsulfonyl fluoride, and other protease inhibitors (Complete, Mini-EDTA-free protease inhibitor cocktail tablets; Roche Diagnostics, Indianapolis, IN, USA) using a 1 mL Dounce homogenizer (Bellco Glass Inc., Vineland, NJ, USA), followed by centrifugation at 1000 g for 5 min. Supernatant was retained as the cytosolic extract. The pellet containing nuclei was resuspended in 100 μL of 20 mmol/L HEPES–KOH (pH 7.9) containing 25% glycerol, 420 mmol/L NaCl, 1.5 mmol/L MgCl2, 0.2 mmol/L EDTA, 0.5 mmol/L dithiothreitol, 0.2 mmol/L phenylmethylsulfonyl fluoride, and other protease inhibitors, incubated on ice for 20 min, and debris was removed from the nuclear extract by centrifugation (17 000 g for 2 min) at 4°C. Protein was determined in cytosolic and nuclear extracts (Bradford 1976) and extracts were stored at −70°C until use.

Western analysis

Nuclear or cytosolic proteins (10 μg of extract), along with protein markers (GE Healthcare Biosciences, Piscataway, NJ, USA), were separated on 10% or 12% denaturing polyacrylamide gels in 50 mmol/L Tris–40 mmol/L glycine, pH 8.3 containing 0.1% sodium dodecyl sulfate at 100 mA for 1.5 h (Tai and Wong 2003). Transcription factor or PNMT protein was analyzed by chemiluminescence (GE Healthcare Biosciences Corp.) using specific antibodies [Egr-1, Sp1, GR, AP-2, and PNMT antibodies; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA; (Wong et al. 1987)] after transblotting to nitrocellulose in 25 mmol/L Tris–40 mmol/L glycine, pH 8.3 containing 10% MeOH at 100 V for 1 h. Blots were Ponceau stained prior to luminescent analysis to ensure equivalent loading of protein. Fluorographic signals were analyzed semi-quantitatively as described above for RT-PCR.

Gel mobility shift assay

Gel mobility shift assays (GMSAs) were performed using the following double-stranded oligonucleotides, 5′CCTCCCCGCCCCCGCGCGTCC3′, −165 bp Egr-1 binding site in the PNMT promoter (−160 to −180 bp) (Ebert et al. 1994); 5′TAGAGGGGCGGGGCTCTAGAC3′, Sp1 binding site (Christy and Nathans 1989); 5′GTCTGGGCGGGGGGGAGGGGA3′, GR-binding sequence with palindromic glucocorticoid response element (Scheidereit and Beato 1984); and 5′GATCGAACTGACCGCCCGCGGCCCGT3′, AP-2 consensus binding site (SG-2513; Santa Cruz Biotechnology Inc.), 5′ end-labeled with [γ-32P]dATP using T4 polynucleotide kinase as previously described (Ebert et al. 1994; Tai et al. 2001). Binding complexes were separated on 5% polyacrylamide gels and autoradiography was performed for complex visualization. Protein–DNA binding complexes were compared by computerized densitometry as described for RT-PCR and western analysis.

Statistical analysis

Data are presented as the mean ± SEM. Experiments were repeated at least three times, with 8–10 replicates per group. Statistical significance between experimental and control groups was determined using InStat3 (GraphPad, San Diego, CA, USA). Data were analyzed by one-way anova followed by post hoc comparisons using the Student–Newman–Keuls multiple comparisons test to compare all values against each other or Dunnett’s comparison test to compare treatment groups against controls. Results were considered statistically significant with values of p ≤ 0.05.

Results

Stress-induced changes in adrenal CORT

While many studies report circulating corticosterone levels rather than tissue concentrations, the PNMT gene is glucocorticoid-regulated (Tai et al. 2002), and the adrenal medulla is the first tissue to be exposed to glucocorticoids newly synthesized in the adrenal cortex. IMMO stressed-induced changes in adrenal CORT were therefore determined by radioimmunoassay (ICN Biomedical Inc.) to ascertain the availability of this corticosteroid for PNMT gene activation in the medulla and ensure that desensitization to IMMO stress was not occurring with repeated stress. As shown in Fig. 1, CORT was rapidly elevated by IMMO stress, with a 10.5-fold induction above control apparent immediately following 30 min of IMMO (p ≤ 0.001 relative to control). If IMMO was extended to 2 h, CORT levels were not significantly different from its elevation with 30 min of stress (p ≥ 0.05) although relative to control, the rise was only 8.5-fold (p ≤ 0.001). However, if killing was delayed for 3 h, CORT levels, which were 4.9-fold control values (p ≤ 0.05) were significantly lower than those observed for 30 min of IMMO with immediate killing (p ≤ 0.001) or 120 min of IMMO with immediate killing (p ≤ 0.01).

Figure 1.

 Immobilization (IMMO) stress-induced changes in adrenal corticosterone (CORT). Male Sprague–Dawley rats (175–200 g) were subjected to IMMO stress for 30 min and killed immediately or for 120 min and killed immediately or 3 h later. Alternatively, they were IMMO stressed daily for 30 or 120 min for 6 days and killed 24 h later or 30 or 120 min daily for 7 days and killed immediately or 3 h after the final IMMO. Right adrenal glands were isolated, homogenized in 1 mL of 50 mmol/L Tris–HCl, pH 7.4, and CORT was measured by radioimmunoassay (ICN Biomedicals Inc.) in triplicate using 100 μL of sample diluted 1 : 200, 1 : 800, and 1 : 1600. Data are expressed as the mean ± SEM (n = 8–10 per group). *, significantly different from control (p ≤ 0.05); **, significantly different from control (p ≤ 0.01); ***, significantly different from control (p ≤ 0.001); +, significantly different from intra-group comparator (p ≤ 0.05); ++, significantly different from intra-group comparator (p ≤ 0.01); and +++, significantly different from intra-group comparator (p ≤ 0.001).

When animals were subjected to daily repeated IMMOs using the same conditions, a similar pattern emerged. Specifically, seven daily IMMOs for 30 min with immediate killing, seven daily IMMOs for 120 min with immediate killing, six daily IMMOs for 120 min with 24 h delayed killing, and seven daily IMMOs for 120 min with 3 h delayed killing elevated CORT 12.1-, 10.1-, 4.8-, and 2.9-fold above control (p ≤ 0.001, 0.001, 0.01, and NS, respectively). CORT levels after seven daily 30 or 120 min IMMOs and immediate killing were not significantly different, while CORT levels after seven daily IMMOs for 120 min with 3 h delayed killing or six daily IMMOs for 120 min with 24 h delayed killing were significantly different from values under the former two conditions (p ≤ 0.001 for both). Intergroup comparisons of CORT showed no significant differences between values for single and corresponding daily repeated IMMOS for 7 days. Finally, there were no significant differences between CORT values for one and seven daily IMMOs for 120 min with 3 h delayed killing and six daily IMMOs for 120 min with 24 h delayed killing.

Together, these results show that IMMO stress, whether stress occurs once or repeatedly, markedly elevates adrenal CORT, followed by an apparent decline to basal values. These changes in CORT are consistent with previous reports for stress-induced changes in circulating CORT and ensure sufficient availability of corticosteroids for PNMT gene induction in the adrenal medulla. In addition, with repetitive IMMO daily for 7 days, desensitization was not apparent based on equivalent induction of adrenal CORT with repeated IMMO.

Stress-induced changes in DA, DOPAC, NE, and EPI

To determine whether catecholamine synthesis was sufficient to restore adrenal medullary pools as well as sustain the elevated circulating levels of EPI released from the adrenal medulla in response to stress, DA, DOPAC, NE, and EPI levels were determined in adrenal glands isolated from rats exposed to IMMO stress using reverse phase HPLC (Lindley et al. 1999). As shown in Fig. 2a, significant changes in DA and DOPAC were evident with either single or repeated IMMO for 30 or 120 min, followed by immediate killing. With a single 30 min IMMO, DA, and DOPAC were elevated 1.4- and 1.3-fold, respectively (p ≤ 0.001 relative to control), while seven repeated IMMOs daily for 30 min elevated DA and DOPAC 1.6- and 1.5-fold, respectively (p ≤ 0.001 compared with control). Extending the length of IMMO to 120 min resulted in a greater increase in DA and DOPAC for a single IMMO (1.9- and 1.8-fold, respectively; p ≤ 0.001 relative to control) but not for repeated IMMO (1.6- and 1.3-fold, respectively). However, in the case of the latter, the relative rise in DA was significantly greater than the rise in DOPAC (p ≤ 0.001).

Figure 2.

 Immobilization (IMMO) stress-induced changes in adrenal dopamine (DA), dihydroxyphenylacetic acid (DOPAC), norepinephrine (NE) and epinephrine (EPI). Male Sprague–Dawley rats were IMMO stressed, and perchlorate extracts of adrenal medullae were prepared as described in ‘Materials and methods.’ After passage through 45 μm filters, catecholamine content was measured by reverse phase HPLC (Lindley et al. 1999), normalized to tissue protein and then expressed relative to control values set to 100%. Data are expressed as the mean ± SEM (n = 8–10 per group) relative to unstressed control values. *, significantly different from control (p ≤ 0.05); **, significantly different from control (p ≤ 0.01); ***, significantly different from control (p ≤ 0.001); ++, significantly different from matched single IMMO (p ≤ 0.01); +++, significantly different from matched single IMMO (p ≤ 0.001); and a, significantly different from corresponding changes in DA (p ≤ 0.001).

In contrast, no significant change in adrenal NE content was detected by comparison with basal levels (Fig. 2b). EPI content remained unaltered as well with the exception that a slight, but significant, reduction in EPI content was apparent immediately after a single 30 min IMMO or a single 120 min IMMO with killing 3 h later (p ≤ 0.05 compared with control).

It appears, then, that in response to IMMO stress, increased DA biosynthesis provides sufficient precursor to both replenish medullary pools of EPI and sustain the elevated circulating EPI levels elicited by and critical to the stress response.

Stress-induced changes in PNMT mRNA expression

As a first step in determining whether IMMO stress alters PNMT gene expression, total RNA was isolated from adrenal medullae collected from rats subjected to IMMO as described above and PNMT mRNA determined by ribonuclease protection assay (Wong et al. 1992a) using GAPDH mRNA, which is not altered by IMMO stress, as a normalization standard (Fig. 3). Levels of PNMT mRNA rapidly rose in response to IMMO, with a significant increase apparent with 120 min of IMMO, followed by immediate killing (3.7-fold, p ≤ 0.001 compared with control). Delaying killing for 3 h after 120 min of IMMO stress resulted in highest induction for both acute and repeated IMMO (6.9-fold control, p ≤ 0.001 compared with control). Whereas a single 30 min IMMO, paired with immediate killing, did not elevate adrenal PNMT mRNA, seven daily IMMOs for 30 min increased PNMT mRNA 3.0-fold (p ≤ 0.01 compared with control, p ≤ 0.05 with respect to a single 30 min IMMO). Extending the time of repeated daily stress to 120 min further incremented PNMT mRNA levels (4.4-fold compared with control, p ≤ 0.001), with highest elevation again apparent with killing 3 h post-stress (5.5-fold compared with control, p ≤ 0.001) (Fig. 3). In addition, there was no significant difference in the magnitude of rise in PNMT mRNA for single or multiple IMMO for 120 min with immediate killing or delayed killing. Analysis of PNMT mRNA by radioactive RT-PCR yielded similar results for single and seven daily IMMOs (Fig. 5), with increases in duration of IMMO and time to killing resulting in higher PNMT mRNA expression. However, for six daily IMMOs (120 min of IMMO with killing 24 h later), PCR analysis showed robust induction of adrenal medullary PNMT mRNA by comparison with ribonuclease protection assay results. Similar induction was also observed by PCR for 30 min of IMMO and killing 24 h later. These differences, in part, may relate to animal to animal variation and their individual time courses for stress responsiveness. We did not examine PNMT mRNA changes in the same samples using both ribonuclease protection assays and RT-PCR in these studies. In general, however, when we have performed so in the past, similar results have been obtained.

Figure 3.

 Stress-induced changes in adrenal phenylethanolamine N-methyltransferase (PNMT) mRNA. Male Sprague–Dawley rats were immobilization (IMMO) stressed as described in Fig. 1 and total RNA was isolated from the left adrenal using TRI REAGENTTM (Sigma-Aldrich Corp.). PNMT mRNA was determined by ribonuclease protection assay using 1 μg of total RNA as previously described (Wong et al. 1992a). Data are expressed as the mean ± SEM (n = 8–10 per group). **, significantly different from control (p ≤ 0.01) and ***, significantly different from control (p ≤ 0.001).

Figure 5.

 Stress-induced changes in adrenal Egr-1, Sp1, glucocorticoid receptor (GR), and AP-2 mRNA. Male Sprague–Dawley rats were immobilization stressed as described in Fig. 1, total RNA was isolated from single adrenal medallae, and Egr-1, Sp1, GR, AP-2, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (control) mRNA was determined by radioactive RT-PCR (Her et al. 2003) as described in ‘Materials and methods.’ Changes in phenylethanolamine N-methyltransferase (PNMT) mRNA in the same samples, determined by RT-PCR, are also depicted for reference. Representative autoradiogram (n = 6).

Thus, IMMO stress causes a marked and robust elevation in PNMT mRNA. The intensity of the stress response was not diminished by 7 days of repeated IMMO stress, consistent with the absence of adaptation to the stressor.

Stress-induced changes in PNMT gene transcriptional regulators

If the IMMO stress-induced increase in PNMT mRNA is associated with stimulation of PNMT gene expression, it should be preceded by a rise in PNMT gene transcriptional activators. The immediate early gene transcription factor Egr-1 has been shown to be an important contributor to PNMT gene activation in response to a variety of stimuli (Morita et al. 1995, 1996; Tai et al. 2001; Tai and Wong 2002, 2003; Wong and Tai 2002; Wong et al. 2002a,b). Changes in Egr-1 were therefore examined in total RNA isolated from adrenal medullae of IMMO-stressed rats using ribonuclease protection assays. As shown in Fig. 4, stimulation of Egr-1 mRNA was apparent immediately following 30 min of IMMO with a 25.9-fold maximum elevation relative to control (p ≤ 0.001). A significant but lesser increase in Egr-1 occurred with 120 min of IMMO, followed by immediate killing (5.5-fold relative to control, p ≤ 0.001). Repeated daily IMMO stress for 30 min, followed by immediate killing, also significantly induced Egr-1 mRNA (19.3-fold relative to control, p ≤ 0.001). As with PNMT mRNA, the results were corroborated by radioactive RT-PCR (Fig. 5). When RT-PCR was used to investigate IMMO stressed-induced changes in Sp1 mRNA, induction was apparent for all acute paradigms, with levels highest for 30 min of IMMO and immediate killing, and 120 min of IMMO with delayed killing (∼2.0-fold) and for six daily IMMOs for 30 or 120 min with killing 24 h later. In contrast, no significant differences were observed in Sp1 mRNA with seven daily IMMOs, irrespective of duration or time to killing or in GR or AP-2 mRNA with any of the IMMO stress paradigms (Fig. 5).

Figure 4.

 Stress-induced changes in adrenal Egr-1 mRNA. Egr-1 mRNA was determined by ribonuclease protection assay (Wong et al. 1992a) using 1 μg of total RNA from the samples described in Fig. 3. Data are expressed as the mean ± SEM (n = 8–10 per group). *, significantly different from control (p ≤ 0.05) and ***, significantly different from control (p ≤ 0.001).

To determine whether stressed-associated Egr-1, Sp1, GR, and AP-2 protein expression corresponded to respective mRNA expression, enhanced chemiluminescent (ECL) western analysis was performed on nuclear extracts isolated from individual adrenal medullae of IMMO stressed rats (Tai et al. 2001). In keeping with its role as an immediately early gene transcription factor, Egr-1 protein was markedly and rapidly induced, with highest elevation apparent immediately following 120 min of IMMO. Egr-1 protein was still elevated 3 h after termination of 120 min of IMMO but 33% less abundant (Fig. 6). With 30 or 120 min of IMMO stress daily for 6 days, Egr-1 protein was also elevated 24 h after the final stress but less than observed for 120 min of IMMO with immediate killing. If animals were exposed to one additional day of IMMO, Egr-1 protein levels were clearly lower. Abundance was highest with 30 min of daily IMMO stress for 7 days, followed by immediate killing (33% levels seen with single 120 min IMMO) and declined with increasing duration of stress or time to killing. Specifically, Egr-1 protein after 7 days of 120 min of IMMO stress was 15% the levels seen with a single 120 min IMMO and not significantly different from basal if killing occurred 3 h after the final 120 min IMMO. As apparent from the fluorogram, two protein bands were observed for Sp1, representing the 95 and 105 kDa forms of Sp1 previously described (Ebert and Wong 1995; Her et al. 1999). As occurred for Egr-1, both Sp1 proteins were also elevated by a single IMMO for 30 min (immediate killing) or 120 min (immediate or delayed killing) and six daily IMMOs for 30 min (killing 24 h later) with an ∼2.0-fold increase by comparison with control levels. However, no significant induction of either was observed with IMMO for 120 min daily for 6 days with killing 24 h later or for 7 days irrespective of stress duration or time to killing. Finally, no changes in GR or AP-2 protein were apparent for any of the stress treatments, consistent with the absence of change in their respective mRNAs.

Figure 6.

  Stress-induced changes in phenylethanolamine N-methyltransferase gene transcriptional activators and respective protein–DNA binding complex formation. Male Sprague–Dawley rats were immobilization stressed as described in Fig. 1 and nuclear extracts prepared from single adrenal medullae as described in ‘Materials and methods.’ Enhanced chemiluminescent western analysis was performed using 10 μg of nuclear protein and transcription factor-specific antibodies [Egr-1, Sp1, glucocorticoid receptor (GR), and AP-2 antibodies; Santa Cruz Biotechnology Inc.], and is depicted as the upper of the paired bands. Gel mobility shift assays were also executed by combining 5 μg of nuclear extracts and radiolabeled, double-stranded oligonucleotides encoding the Egr-1, Sp1, GR, and AP-2 consensus binding sequences as described in ‘Materials and methods’ (Tai and Wong 2003). Representative fluorograms (enhanced chemiluminescent western analysis) and autoradiograms (gel mobility shift assay) (n = 6).

Egr-1 and Sp1 are transcription factors that require phosphorylation for binding and transcriptional activation (Williams and Tjian 1991; Cao et al. 1992, 1993). Hence, a change in their phosphorylation state is an important determinant of their functionality as genetic regulators of PNMT. As antibodies specific for the phosphorylated forms of these proteins were unavailable, GMSAs were executed for each of the above transcription factors using the nuclear extracts above and [32P]-labeled oligonucleotides encoding the consensus binding sequences for each transcription factor as described in ‘Materials and methods’ to assess their ability to form protein–DNA complexes. In general, the abundance of Egr-1 and Sp1 protein–DNA complexes mirrored the pattern of change for their proteins as determined by ECL western blotting (Fig. 6) with one exception. While the 95 and 105 kDa forms of Sp1 were unaltered by six daily IMMOs for 120 min with killing 24 h later, their potential for protein–DNA binding complex formation increased ∼2.0-fold. Thus, the extent of IMMO stress and duration post-stress markedly affects the amount of Sp1–DNA binding complexes generated, indicating that stress may alter the phosphorylation state of Sp1 as well and thereby titer its ability to act as a stress-induced transcriptional activator of the PNMT gene. Finally, no differences in GR or AP-2 protein–DNA complex formation were observed irrespective of the stress treatment.

Thus, Egr-1 and Sp1 appear to be two transcriptional mediators of the stress responsiveness of the PNMT gene. For both, changes in protein and extent of protein phosphorylation are essential to their role as stress-induced transcriptional activators.

Changes in PNMT protein

Results described thus far support the hypothesis that enhanced biosynthesis of EPI via transcriptional activation of the PNMT gene may be an important regulatory component of the stress responsiveness of this neurohormone. However, it must also be demonstrated that induction of PNMT mRNA leads to appropriate changes in PNMT enzyme. To investigate the effects of IMMO stress on PNMT protein, ECL western analysis was performed on cytosolic protein extracts isolated from individual adrenal medullae of the IMMO stressed rats. In general, as shown in the representative fluorogram in Fig. 7, the pattern of stressed-induced change in PNMT protein reflected changes observed for PNMT mRNA. PNMT protein was significantly elevated by a single IMMO for 30 or 120 min, with a more marked rise apparent 3 h after 120 min of IMMO. Highest elevation of PNMT occurred with repeated stress, six times daily for 30 or 120 min with killing 24 h after the final IMMO. PNMT was similarly elevated if IMMO was repeated one additional day for 30 or 120 min with immediate killing. However, while PNMT protein was significantly elevated in response to 30 min of IMMO with immediate killing, mRNA was not significantly altered from control values. In addition, PNMT protein was closer to basal levels of expression following 120 min of IMMO with 3 h delayed killing even though PNMT mRNA was the highest for all of the seven daily IMMO paradigms. The former may indicate that IMMO stress activates post-transcriptional mechanisms, which stabilize PNMT protein (Wong et al. 1992b) and the latter perhaps the onset of adaptation to IMMO stress, despite continued elevation of CORT.

Figure 7.

 Stress-induced changes in adrenal phenylethanolamine N-methyltransferase (PNMT) protein. Cytosolic extracts were prepared from individual right adrenal glands as described in ‘Materials and methods’ and enhanced chemiluminescent western analysis performed using anti-PNMT antibody (Wong et al. 1987). Representative fluorogram (n = 6).

Thus, changes in PNMT protein induced by IMMO stress are also consistent with EPI biosynthesis and transcriptional control of the PNMT gene being one regulatory mechanism mediating adrenergic responsiveness to stress.

Discussion

The neurohormone EPI is an important component of the stress response, activating physiological processes required to meet the challenge of a stress. It has also emerged as a critical constituent in stress-related disorders such as cardiac disease and depression (Lorita et al. 2002; Stone et al. 2003; Gold et al. 2006). While the primary source of this neurohormone is the adrenal medulla, limited clusters of adrenergic neurons in the brainstem (Cooper et al. 1991) exert powerful influences on cardiac function (Singewald and Philippu 1996), and PNMT and EPI synthesis can also occur in heart tissue itself (Ebert et al. 1996; Ebert and Thompson 2001; Krizanova et al. 2001; Ziegler et al. 2002). In addition, EPI, through activation of central α1b receptors, seems to modulate serotonergic, dopaminergic, and noradrenergic neurotransmitter systems identified as targets for drugs effective in treatment of stress-related mood disorder (Stone et al. 2003).

One way in which EPI is regulated is via its biosynthesis. Previous studies suggested that the latter occurs in part through control of PNMT gene transcription (Viskupic et al. 1994; Wong et al. 1996, 2002a,b, 2004; Sabban and Kvetnansky 2001). Findings reported here provide further evidence that transcriptional regulation may be important in orchestrating adrenergic responses to stress. We show that IMMO stress rapidly and markedly increases PNMT mRNA expression. The latter is preceded by activation of at least two PNMT gene transcriptional regulators, Egr-1 and Sp1, and their enhanced binding to their consensus elements in the PNMT promoter. Downstream of PNMT gene activation, corresponding changes in PNMT protein are induced. As adrenal medullary EPI remains unchanged and elevated circulating EPI is sustained in response to stress (Dronjak et al. 2004), sufficient biosynthesis of EPI appears to occur to replenish medullary pools and support stress-elicited, elevated levels in the blood stream.

While EPI levels in the adrenal medulla are sustained during stress, CORT is rapidly and markedly elevated but elevation is transient. In particular, a maximum 10.5-fold rise in CORT above control values was induced by 30 min of IMMO stress (p ≤ 0.001). If IMMO was extended to 120 min or sampling delayed for 3 h after 120 min of stress, the magnitude of rise was not significantly different, but a downward trend towards basal levels was apparent with extended IMMO duration and time post-stress. An identical pattern was observed with repeated IMMO stress for 7 days, with no significant differences in CORT induction between single and multiple IMMO stress paradigms. The rise in adrenal CORT clearly suffices to sustain the high circulating levels essential for coping with the stress and provides glucocorticoids for GR activation and thereby gene induction. Thirty minutes of stress evokes maximum glucocorticoid stimulation as CORT levels seem lower when IMMO is increased to 2 h. The latter further suggests that adrenal CORT changes, rather than representing absolute changes, may reflect stress-induced temporal changes in this hormone, which should exhibit a rapid, marked rise, followed by a decline. As repeated IMMO stress elicits a similar pattern of CORT induction, adrenergic responses to IMMO stress apparently do not desensitize to 30 or 120 min of daily IMMO at least through 7 days.

Immobilization stress also induced a similar, rapid, and marked induction in PNMT mRNA expression. Highest stimulation was observed 3 h after 120 min of IMMO, and the pattern of PNMT mRNA induction was identical for single and repeated IMMO. These results are consistent with the findings of Viskupic et al. (1994). While the investigators only reported statistical comparisons relative to unstressed animals and not inter-treatment group comparisons, results depicted for changes in PNMT mRNA suggested that duration of IMMO stress was less important than time from initiation of stress to killing. Specifically, there appeared to be no significant difference between 5 and 120 min of IMMO as long as killing occurred 2 h after the stress was begun. Our findings concur for both single and repeated IMMO except that we observed a greater peak in PNMT mRNA and markedly so for a single IMMO (∼7.0 vs. 3.0-fold, respectively). There does appear to be some variation in PNMT mRNA changes in responses to six daily IMMOs for 30 or 120 min between samples assessed by ribonuclease protection assay versus RT-PCR. These differences are likely real as the total RNA samples were from different animals. In the past when both assay methodologies were used to determine PNMT mRNA in the same samples, results have been similar, whereas cross-experiment differences have been observed. As described above for CORT changes, PNMT mRNA induction may more appropriately represent temporal PNMT gene responses to stimulus activation. We have previously shown that a variety of stimuli maximally induce PNMT gene expression 6–8 h post-stimulus (Morita et al. 1995, 1996; Morita and Wong 1996). We have also observed that splanchnic denervation does not affect the rise in PNMT mRNA elicited by IMMO stress (data not shown), consistent with the conclusion by Viskupic et al. (1994) that innervation of the adrenal medulla is not required for stress-mediated PNMT mRNA stimulation. Betito et al. (1994) have reported that mild, acute stress elicits similar changes in adrenal PNMT expression as well (Betito et al. 1994).

To demonstrate that changes in steady state levels of PNMT mRNA may be due to IMMO stress activation of PNMT gene transcription, we investigated the effects of IMMO stress on the PNMT transcriptional activators Egr-1, Sp1, the GR, and AP-2. In the case of the immediate early gene transcription factor Egr-1, a very important regulator of the PNMT gene (Ebert et al. 1994; Morita et al. 1995, 1996; Morita and Wong 1996; Wong et al. 1998a,b, 2002a,b; Tai et al. 2001; Tai and Wong 2002, 2003; Wong and Tai 2002), mRNA and protein levels rapidly and transiently rose. As demonstrated by other investigators (Cao et al. 1992), phosphorylation of Egr-1 likely contributes to its interaction with consensus DNA target sequences, increasing potential for transcriptional activation. A similar response was observed for Sp1 with induction of mRNA, protein, and protein–DNA complex formation being markedly elevated for all acute IMMO paradigms as well as for 30 min of daily IMMO for 6 days. However, for six daily IMMOs for 120 min with killing 24 h later, a marked increase in Sp1 mRNA and protein–DNA complex occurred, while protein expression remained unaltered. In addition, seven daily IMMOs, irrespective of duration of IMMO or time to killing, had no effect on Sp1 mRNA, protein, or protein–DNA complex. Sp1 is a ubiquitous and abundant transcription factor, which like Egr-1, requires phosphorylation for functionality (95 and 105 kDa proteins). It is unclear to what extent available antibodies interact with both phosphorylated and de-phosphorylated forms of this protein and if so, with the same avidity. Thus, changes in Sp1 protein and protein–DNA binding complex formation may arise from induction of protein, phosphorylation of existing protein and/or a combination of the two. We are currently determining whether we can distinguish between these possibilities by comparing western analysis and GMSA results from phosphatase-treated nuclear extracts to those from untreated nuclear extracts. In contrast, GR and AP-2 mRNA, protein, and protein–DNA binding complex formation were not altered by IMMO stress. However, the absence of induction does not preclude their participation in stress-induced activation of PNMT gene transcription as we have previously shown that Egr-1, AP-2, and the GR, when bound to their respective consensus elements in the rat PNMT promoter, can interact cooperatively to synergistically activate PNMT promoter-driven transcription (Ebert et al. 1998; Wong et al. 1998b; Tai et al. 2002).

Finally, changes in PNMT mRNA, for the most part, were also consistently reflected in PNMT protein expression for acute and chronic IMMO with two exceptions. In the case of a single IMMO for 30 min followed by immediate killing, there was no significant elevation in adrenal medullary PNMT mRNA and yet, PNMT protein was clearly elevated above control values. We have previously reported that chronic exposure to corticosteroids can stabilize PNMT protein by ensuring sufficient levels of the co-substrate and methyl donor, S-adenosyl-methionine (Wong et al. 1985). Glucocorticoids apparently sustain levels of the AdoMet metabolic enzymes and sufficient AdoMet so that its binding to PNMT protects the enzyme from proteolytic degradation. While such cellular events require long-term changes, it is possible that short-term responses associated with 30 min of stress may also elicit changes, which prevent PNMT degradation so that the net effect is higher steady state levels of PNMT protein. The reverse situation is observed for daily IMMO for 120 min for 7 days, followed by delayed killing. PNMT protein is higher than basal values but less than anticipated given that mRNA levels were the highest for the three 7-day stress paradigms. We have previously reported similar hormonally induced discrepancies between PNMT mRNA and protein (Wong et al. 1992b). While the marked elevation of PNMT mRNA still supports the potential for increased EPI biosynthesis, the reduction in protein expression relative to mRNA may be indicative of the initiation of desensitization to IMMO stress and/or a similar circumstance of dual and separate transcriptional and post-transcriptional control of PNMT expression.

In summary, our findings demonstrate that an important mechanism for sustaining adrenergic function during stress is orchestrated through PNMT gene regulation. IMMO stress increases PNMT mRNA and protein rapidly and markedly preceded by activation of the PNMT transcriptional regulators, Egr-1 and Sp1. These changes sustain EPI synthesis, thereby maintaining adrenal medullary pools and providing sufficient release of this stress hormone to support the physiological responses necessary to invoke adaptive changes to counter the adverse effects of stress.

Acknowledgements

This work was supported by a grant from the National Institute of Diabetes, Digestive Disease and Kidney Disorders, DK51025, the Spunk Fund Inc., the Sobel and Keller Fund, and McLean Hospital (DLW) and Slovak grant, APVV Grant No. 0148-06 (RK).

Ancillary