Background
Catechol-
O-methyltransferase (COMT) metabolizes catecholamines and thereby acts as a key modulator of dopaminergic and adrenergic/noradrenergic neurotransmission [
1,
2]. Converging lines of evidence have revealed an important role of COMT in the etiology and pathogenesis of a wide variety of central nervous system (CNS) disorders [
2‐
4]. Recently, COMT has also been implicated in the regulation of pain perception [
5,
6]. Myofacial pain patients exhibit lower COMT activity relative to controls [
7], and COMT inhibition increases pain sensitivity in rodents by promoting catecholamine stimulation of β
2- and β
3-adrenergic receptors [
8].
The COMT protein exists in two major forms: a shorter soluble form (S-COMT) and a longer membrane-bound form (MB-COMT). They are encoded from one gene by two mRNA transcripts (1.3 and 1.5 kb in human, 1.6 and 1.9 kb in rats) regulated by the proximal
P1 and distal
P2 promoters, respectively [
9‐
11]. Only the longer transcript was found in the brain [
12] with the predominant protein being MB-COMT. However, S-COMT protein is also expressed in the brain from the longer MB-COMT mRNA isoform
via a leaky scanning mechanism [
13]. Though their sequences are largely homologous, MB-COMT has approximately a 10-fold greater affinity for dopamine and noradrenaline relative to S-COMT [
14]. Seven novel COMT mRNA variants have also been detected in brain, however, they likely to exist at much lower levels than the primary transcript [
15]. Although recent reports describe a neuronal expression of COMT [
16], it is primarily considered a glial enzyme [
17‐
19].
A significant role for glia in mediating pain has been implicated by studies of patients with persistent pain conditions and animal models of pain [
20‐
22]. Proinflammatory cytokines are produced and released by activated microglia and astrocytes in the CNS as well as by immune cells at the site of injury or inflammation. [
23‐
26]. TNFα is widely considered to be the prototypic proinflammatory cytokine due to its principal role in initiating the cascade of cytokines and growth factors involved in the inflammatory response [
20]. Tissue levels of TNFα have been correlated with pain report in a number of painful diseases [
27‐
29]. TNFα activates NF-κB, which is the pivotal regulator of cellular inflammatory responses [
30‐
32]. Specifically, the NF-κB pathway plays one of the major roles in injury or inflammation-evoked activation of astrocytes [
23,
33,
34]. Within the nervous system, NF-κB is most frequently composed of two DNA-binding subunits (p65/Rel A and p50) that form a complex with the inhibitory subunit IκB which normally retains NF-κB within the cytoplasm of unstimulated cells [
35]. Signal-induced phosphorylation, ubiquitination, and degradation of IκB triggers NF-κB nuclear translocation and DNA binding. Phosphorylation of IκB is mediated by the IκB kinase (IKK) complex, which consists of two catalytic subunits, IKKα and IKKβ, and the regulatory subunit IKKγ [
36]. Gene knock-out studies have established an essential role for IKKβ in TNFα-induced activation of NF-κB [
37].
A growing number of reports reveal a crucial role of NF-κB in nociception. NF-κB activity is increased in animal models of neuropathic and inflammatory pain [
38‐
42]. A specific IKK inhibitor reverses heightened pain sensitivity to noxious (hyperalgesia) and normally innocuous stimuli (allodynia) [
43]. Increased neuropathic and inflammatory pain is suppressed by pretreatment with an NF-κB inhibitor [
39,
44]. Interestingly, selective inactivation of NF-κB in glial cells or astrocytes leads to decreased pain and better functional recovery [
45‐
47].
Despite increasing evidence for an important role of NF-κB in pain regulation, very few studies have addressed the mechanisms whereby this pathway exerts its effects on nociception [
38,
39,
41,
43,
48]. We hypothesized that NF-κB regulates expression of COMT, an enzyme known to contribute to enhanced pain states. Thus, the present study explored the relationship between the NF-κB pathway and COMT expression in order to gain an understanding of the cellular mechanisms underlying inflammatory pain.
Discussion
In the present report, we provide the first demonstration that COMT gene expression is downregulated by TNFα in primary rat astrocytes at both protein and mRNA levels. As the P2-COMT promoter controls expression of MB-COMT, the main COMT transcript in brain, this promoter was cloned from human genomic DNA and transfected in H4 astroglioma cells. The activity of the cloned promoter was substantially suppressed by TNFα in a time-dependent manner.
A number of putative regulatory elements have been described in
P1- and
P2-COMT promoters, including estrogen response (ER) elements [
50] that likely mediate estradiol-dependent downregulation of COMT expression in cell culture [
55].
P2-COMT also contains abundant methylation sites associated with cancer development [
56], schizophrenia, and bipolar disorder [
57]. We identified a novel putative regulatory site – a κB consensus sequence that is a potential target for TNFα-dependent NF-κB activation.
Next, we demonstrated that TNFα-dependent COMT downregulation was indeed mediated by the NF-κB pathway. Transient expression of p65, the essential component of NF-κB complexes, or IKKβ, the major positive regulator of NF-κB activition, significantly decreased P2-COMT reporter expression. In addition, H4 IκBα-SR cells lost the ability to regulate P2-COMT promoter expression in response to TNFα treatment. The TNFα-mediated suppression of endogenous COMT expression was also abrogated in H4 IκBα-SR cells. Moreover, we confirmed that TNFα activated NF-κB in H4 astroglioma cells through the canonical IκB degradation pathway to trigger p65 nuclear translocation and DNA binding.
Our data strongly suggest that the putative κB binding site 5'-GGGGACGCCC-3' at position -109 of the P2-COMT promoter region is a functional site for NF-κB-mediated regulation of COMT expression as deletion of the P2-COMT region containing this site abrogated TNFα-dependent inhibition of P2-COMT activity in H4 astroglioma cells. Furthermore, competition experiments performed with the wild type or mutant site-specific oligonucleotides showed that TNFα indeed induced recruitment of p65 to this κB consensus binding site of the promoter.
Although NF-κB-mediated activation of transcription is well known, the mechanisms of NF-κB-mediated repression are poorly established. Probably, the best studied example of transcriptional repression by NF-κB complex is described for Dorsal transcription factor, a
Drosophila Rel family member that can either activate or repress gene expression through the recruitment of coactivators such as CBP or corepressors such as Groucho [
58]. Furthermore, a number of examples have been reported in mammals. NF-κB can repress transcription by competing with steroid receptors for a common promoter
cis-DNA element [
59]
via N-myc recruitment to the glutamate transporter gene promoter [
53] and through inhibiting histone H4 acetylation at the cytochrome P-450 1A1 promoter [
60]. Thus, further experiments should be conducted to address the specific mechanism underlying NF-κB-dependent inhibition of
COMT gene expression. Interestingly, consequent deletions of 5'fragments of the
P2-COMT promoter led to a significant increase in overall basal promoter activity. This result would suggest the presence of a number of putative negatively regulating elements along the
P2-COMT promoter other than ER- and κB-response elements. Although, to date, no studies have systematically searched for regulators of COMT expression, this finding clearly warrants further research.
Our results demonstrating that COMT expression is downregulated in astrocytes under inflammatory conditions are in line with those of other studies showing a positive correlation between astrocyte activation and exaggerated pain responses [
24,
25,
61]. Intrathecal injection of gp120 (human immunodeficiency virus-1 envelope glycoprotein) induces mechanical allodynia
via the release of proinflammatory cytokines and NF-κB activation in spinal cord astrocytes, but not in microglial cells or neurons [
62]. Selective inactivation of astroglial NF-κB in transgenic mice expressing a dominant negative form of the inhibitor IκBα leads to a dramatic improvement in functional recovery after contusive spinal cord injury (SCI) [
46] and decreases formalin-induced pain [
47]. Additionally, several recent studies report cell type-specific NF-κB activation by cytokines. For example, in rat brain cultures IL-1 induces NF-κB activation in astrocytes, but not in neurons [
63,
64]. Taken together, these studies unequivocally link NF-κB activation in astrocytes to pain states.
Although activation of the NF-κB pathway has been deemed critical for the development of pain [
40‐
42], there are few reports studying NF-kB-dependent pro-nociceptive signaling. Historically, these studies have focused on NF-kB-dependent up-regulation of pro-inflammatory cytokines [
25,
65], cyclooxygenase-2 (COX-2) [
39,
43], inducible and neuronal nitric oxide synthases (iNOS and nNOS) [
38,
41], c-src [
48], and c-fos [
38]. However, recent studies from our group demonstrated that genetic variants of
COMT coding for low enzymatic activity are associated with heightened experimental pain sensitivity and the onset of a myofacial pain condition in humans [
5]. Additionally, pharmacologic inhibition of COMT in a rat model of inflammation resulted in elevated pain sensitivity [
8]. Together, these data suggest that an NF-κB-mediated decrease in COMT expression is likely to contribute to heightened pain sensitivity under inflammatory conditions. A series of
in vivo experiments further addressing this hypothesis are currently being conducted in our laboratory.
Methods
Cell culture and reagents
Primary astrocytes were isolated and cultured as described earlier [
68]. Human H4 astroglioma cells were obtained from ATCC (HTB-148) and cultured in DMEM (Sigma), 10% fetal bovine serum (FBS; HyClone) and 1× penicillin-streptomycin (Invitrogen). H4 cells stably expressing IκBα-SR were a generous gift from Dr. Baldwin (UNC) and generated as described previously [
53]. All oligonucleotides were obtained from MWG-Biotech AG. The pCMV-SPORT-M, pCMV-SPORT-M-p65 and pCMV-SPORT-M-IKKβ expression vectors were a generous gift provided by Dr. Romanov (Attagene) and 3x-κB/luc construct was a gift from Dr. Baldwin (UNC).
Quantitative real-time RT-PCR
Total RNA was isolated using the Trizol reagent (Invitrogen), treated with RNase free-DNase I (Promega) and reverse transcribed with random primers by Superscript III (Invitrogen). The cDNA was amplified with SYBR Green PCR master mix (Applied Biosystems) using forward and reverse PCR primers (5'-CCAGAGGAGACCCCAGACC-3' and 5'-ACAGCTGCCAACAGCAGAG-3', respectively, for human MB-COMT; 5'-GGAAATCGTGCGTGACATC-3' and 5'-CATGGATGCCAAGGATTC-3', for human β-actin; 5'-CCAGAGGAGACCCCAGACC and 5'-ACAGCTGCCAACAGCAGAG-3', for rat MB-COMT; and 5'-TGCGGGTCATAAGCTTGC-3' and 5'-CGATCCGAGGGCCTCACTA-3' for rat 18S rRNA) in S2 Real Time PCR machine (Eppendorf). PCR reactions were performed in triplicate. Three independent experiments were performed, and the result of a representative experiment is shown. MB-COMT mRNA levels were normalized to β-actin RNA or 18S rRNA as an endogenous control.
Western blot analysis
10–50 μg of protein lysates from whole cells, nuclear and cytoplasmic extracts, normalized for protein content using a BCA Protein Assay Kit (Pierce), were run on precast Novex Tris-Glycine gels (Invitrogen), blotted onto nitrocellose (Whatman), and blocked in TBST with 5% nonfat dry milk. The following antibodies were used: COMT (Chemicon, AB5873), β-actin (I-19) (Santa Cruz, SC-1616), IκBα (C-21) (Santa Cruz, SC-371), p65 (Cell Signaling, #3034), and p50 (Santa Cruz, SC-7178). Chemiluminescence was detected in ImageQuant-ECL Imaging System (GE Healthcare) and images were analyzed using ImageQuant TL software (GE Healthcare). Blots from three independent experiments were densitometrically analysed and the values normalized to the β-actin control, with untreated group set to 100%.
Cloning of human P2-COMT distal promoter
Primers
U1 5'-
CCTACGCGT GCTCCTCTGGCGGAAAGGAA-3' and
D1 5'-
CGAAGATCT ACCTCTCCCGCGACGGCCCG-3', with added Mlu I and Bgl II restriction sites, respectively, were used to amplify
P2-COMT from 50 ng of human genomic DNA with GeneAmp PCR kit (Applied Biosystems). The 1.5 kb PCR product was digested by Mlu I and Bgl II restrictases (NEB), gel-purified, ligated into pGL3 Luciferase Reporter Vector (Promega) using Rapid DNA Ligation Kit (Roche) and transformed into competent
E. coli DH5α cells (Invitrogen). Recombinant plasmids were isolated using EndoFree Plasmid Kit (Qiagen) and sequenced at UNC sequencing facility. Putative regulatory elements were determined with TFSearch database
http://www.cbrc.jp/research/db/TFSEARCH.html.
Construction of serial 5'-end deletions of human P2-COMT/Luc clone
Serial deletions were generated by PCR amplification of corresponding fragments from
P2-COMT/Luc clone using forward primers, containing Mlu I restriction site, and reverse primers, containing Bgl II site,
U2 5'-
CCTACGCGT GCGGACACCCTCACGAGGACA-3' and
D1, respectively, for
Del. 1, and
U3 5'-
CCTACGCGT CCACCGGAAGCGCCCTCCTA-3' and
D1 for
Del. 2. The amplified fragments were digested by Mlu I and Bgl II, purified from the agarose gel and cloned into pGL3 reporter vector. Deletions were confirmed by sequencing. The nucleotide numeration was based on Tenhunen et al. [
11].
Transient transfection, luciferase and β-galactosidase assays
Cells were seeded into 12-well plates (5 × 104cells/well) and transfected with 500 ng of total DNA using FuGene 6 reagent (Roche). Normally, up to 400 ng of P2-COMT luciferase reporter and 30 ng of control plasmid for transfection efficiency (pSV-β-galactosidase vector, Promega) were used for transfection. The amount of DNA was kept constant by addition of pCMV-Sport-M vector with no insert. Cells were treated by TNFα (R&D Systems) and harvested 48 h after transfection. Luciferase activity was determined using Luciferase Assay System (Promega) and normalized for transfection efficiency by measuring the β-galactosidase activity using a β-Galactosidase Enzyme Assay System (Promega). Transfections were performed in triplicate, and a representative experiment is shown.
ELISA for activated NF-κB
NF-κB activation was measured using TransAM NF-κB p65 Chemi Kit (Active Motif). Cell lysates were tested for their ability to bind to a plate-immobilized oligonucleotide containing a κB consensus binding site (5'-GGGACTTTCC-3'). Competition experiments were performed with the wild-type (ACCGCG GGG ACGCCCG GGG ACGCCC CGACC) and mutant (5'-ACCGCG CTC ACGCCCG CTC ACGCCC CGACC) oligonucleotides specific to P2-COMT κB binding site and κB wild-type and mutated consensus oligonucleotides provided by the manufacturer. The wild-type but not mutated oligonucleotides were expected to compete with NF-κB for binding. Chemiluminescence was measured in 1420 Multilabel Counter Victor3 (PerkinElmer). Nuclear extracts were prepared using Nuclear Extract Kit (Active Motif).
Statistical Analysis
Protein, mRNA, and promoter activity data were analyzed by paired t-test and analysis of variance (ANOVA) with post-hoc tests. P < 0.05 was considered to be statistically significant.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
IET and LBD conceived of the study, and IET, AGN, SW, MC, and LQ performed experiments. IET, AGN, and LBD participated in writing of the manuscript. All authors read and approved the final manuscript.