nucleotide–binding oligomerization domain family proteins
mammalian target of rapamycin
pathogen-associated molecular patterns
pattern recognition receptors
The induction of cyclooxygenase-2 (COX-2) protein expression was assessed in human polymorphonuclear leukocytes (PMN) stimulated via receptors of the innate immune system. Peptidoglycan (PGN) and mannan, and at a lower extent the bacterial lipoprotein mimic palmitoyl-3-cysteine-serine-lysine-4, induced COX-2 protein expression. In contrast, lipoteichoic acid and muramyldipeptide were irrelevant stimuli. The mRNA encoding COX-2 was present in resting PMN at an extent quite similar to that detected in stimulated PMN, whereas the expression of COX-2 protein was undetectable. Treatment with the phosphatidylinositol 3-kinase inhibitor (PI3K) wortmaninn, the mammalian target of rapamycin (mTOR) inhibitor rapamycin, and the translation inhibitor cycloheximide blocked the induction of COX-2 protein in response to mannan and PGN, whereas the transcriptional inhibitor actinomycin D did not show a significant effect. These results disclose a capability of pathogen-associated molecular patterns to induce the oxidative metabolism of arachidonic acid more robust than that of PMN archetypal chemoattractants, since mannan and PGN make it coincidental the release of arachidonic acid with a rapid induction of COX-2 protein regulated by a signaling cascade involving PI3K, mTOR, and the translation machinery. This mechanism of COX-2 protein induction expression in PMN is substantially different from that operative in mononuclear phagocytes, which is highly dependent on transcriptional regulation.
PMN are essential elements of the innate immune system due to their wide predominance in peripheral blood and their rapid mobilization into tissues in response to microbial infection. PMN can be activated by a variety of stimuli, which includes cytokines, IgG class antibodies, complement factors, extracellular matrix components, adhesion molecules, and microorganism constituents. This group of stimuli is of chief importance, since early protection against pathogens relies on the recognition by phagocytes of unique pattern molecules, termed pathogen-associated molecular patterns (PAMP), which are recognized through pattern-recognition receptors (PRR) by the host innate immune system. The TLR family (for review see 1, 2), the nucleotide-binding oligomerization domain (NOD) family proteins (for review see 3–5), the lectin C-type receptors such as the β-glucan receptor dectin-1 6, and the mannose receptor (MR) family (for review see 7, 8) include a wide array of PRR able to interact with many structural signatures expressed in microorganisms.
The study of the response of PMN to PAMP has focused on the induction of cytokines, the activation of oxidant production, the release of granular components, and the modulation of apoptosis 9–11. However, few data are available regarding the activation of the arachidonic acid (AA) cascade, yet this seems of interest in view of the important role of eicosanoids in connecting innate and adaptive immunity 12–15 and the attention paid to the pharmacological modulation of both the eicosanoid receptors and the enzymes involved in their biosynthesis in many clinical settings 16, 17. Moreover, current understanding of PMN biology has been modified by recent findings indicating that the life span of PMN can be prolonged by proinflammatory agonists 10, as well as by the depiction of mechanisms of translational control of the expression of specific proteins, which endow the PMN with the potential for rapid protein synthesis from constitutive mRNA without requiring new transcript generation 18–20. The possibility that this mechanism could be operative in PAMP-dependent responses and might influence AA metabolism through the expression of COX-2, the inducible isoform of COX, is a challenging hypothesis. It could provide a rationale for pharmacological intervention with the macrolide antibiotic rapamycin, an inhibitor of the mammalian target of rapamycin (mTOR) 21, or alternatively, an explanation for some of the untoward effects of rapamycin.
In a previous study, we have observed that peptidoglycan (PGN) and mannan are potent stimuli for AA release and PGE2 production in human PMN 22. Since PGE2 can be produced both by COX-1, the constitutive isoform of COX, and by COX-2, the inducible isoform, we have addressed the effect of a set of PAMP signatures on the expression of COX-2 in human PMN. We have found the presence of preformed mRNA encoding for COX-2 in resting PMN and an early and lasting induction of COX-2 protein in response to both mannan and PGN. Since the expression of COX-2 protein was almost completely blocked by the inhibitor of phosphatidylinositol 3-kinase wortmannin, by the mTOR inhibitor rapamycin, and by the translation inhibitor cycloheximide, we propose the existence of a mechanism of COX-2 induction expression in human PMN mediated by the mTOR route.
PAMP signatures are strong inducers of COX-2 protein in human PMN
Since PMN have a limited life span and gene expression induction assays might entail long incubation periods, the appearance of apoptosis was assessed by looking at both phosphatidylserine surface display and propidium iodide staining. Less than 10% of total cells showed positive staining with annexin V-FITC at the end of the isolation procedure, whereas this figure steadily increased along the incubation period to reach ∼ 96% positive cells at 24 h. In contrast, propidium iodide staining after 24 h of incubation only showed 18% of positive cells (Fig. 1A), thus indicating that PMN isolated by conventional procedures are viable and can be used in experiments entailing long incubation periods, since no massive late apoptosis was observed within 24 h.
Human PMN were found to contain high amounts of COX-2 mRNA, as judged from RT-PCR reactions carried out in samples obtained both in resting cells and after stimulation with agonists acting through a varied array of PRR. Of note, differences in the extent of COX-2 mRNA expression could hardly be detected in RT-PCR reactions (Fig. 1B). These results were first observed with primers selected from exons 5 and 7, but similar results were obtained with primers selected from exons 1 and 2 (see below). Real-time RT-PCR of COX-2 mRNA showed a CT value of 18.33 ± 0.73 in control cells, 16.09 ± 0.53 (mean ± SD, n = 5) in PGN-treated PMN, and 17.1 ± 0.33 in mannan-treated PMN, i.e. a set of values, which after normalization for the expression of β-actin mRNA only represents a 2.46-fold increase for PGN- and a 1.8-fold induction for mannan-treated PMN. These findings are different from those observed in macrophages, where mannan and PGN increased COX-2 mRNA 363- and 163-fold, respectively (Table 1), thus indicating either robust transcriptional activation or mRNA stabilization.
|Mannan 10 mg/mL||1.8||302||363|
|PGN 10 μg/mL||2.46||179||163|
Since recent studies have suggested that PMN respond to chemical mediators by initiating the translation of preformed mRNA 19–21, these findings were considered as an indication that a similar molecular mechanism might be implicated in the regulation of COX-2 protein. Mannan and PGN produced a dose-dependent induction of the expression of COX-2 protein (Fig. 1C). Of note, this was also observed with platelet-activating factor (PAF) (Fig. 1D), which is the chemical mediator first associated to this mechanism of signal-dependent translation of constitutive messenger RNA in PMN 19. However, the level of COX-2 protein expression induced by PAF was somewhat lower, thereby suggesting a more robust COX-2 protein induction in response to PRR activation than to G-protein-coupled receptors, the class of receptors to which PAF receptors belong 23. COX-1 protein showed the same level of expression in the absence and presence of several stimuli, but well below the level detected in platelets (Fig. 1E), which are the archetypal source of COX-1. N-palmitoyl-S-[2,3-bis(palmitoyloxy)-2R,S)-propyl]-cysteinyl-(S)-seryl-(S)-lysyl-(S)-lysyl-(S)-lysyl-(S)-lysine trihydrochloride (Pam3CSK4), which is an agonist of TLR2/TLR1 heterodimers, showed a less robust effect and lipoteichoic acid, an agonist of TLR2/TLR6 heterodimers, failed to elicit COX-2 protein induction (Fig. 1C, rightmost lanes). Muramyldipeptide (MDP), which is the archetypal ligand for NOD2 also failed to induce COX-2 expression (Fig. 2A). Since interaction between NOD2 and specific TLR pathways has been reported as a mechanism of cooperation in the innate immune response, which might lead to the synergistic activation of host cells 24–26, the effect of the combined addition of both Staphylococcus aureus PGN and MDP was assessed. As shown in Fig. 2A, this combination of agonists did not modify the effect elicited by PGN alone. By contrast, the endogenous ligand of the MR mucin-3 induced COX-2 expression and enhanced the effect of a suboptimal concentration of mannan (Fig. 2B, right panels).
PGN elicited detectable COX-2 protein induction as soon as 30 min after addition of the stimulus and this remained almost unchanged from 1 to 18 h. A similar trend was observed for C3bi-coated zymosan and mannan, although a decreasing tendency was observed around 18 h in response to these agonists (Fig. 2B). To define more widely the proinflammatory context elicited by PRR in which COX-2 induction occurs, the expression of other inflammatory proteins was assessed with human Ab array membranes. The expression of IL-1β, IL-8/CXCL8, and MIP-1β/CCL4 increased above their basal expression level in response to mannan (Fig. 2C), thus agreeing with the induction of these cytokines in human PMN stimulated with other agonists 19. Taken together, the most likely interpretation of these results is that the MR might be involved in the response to mannan, and that PGN contains a structural signature not acting on NOD2 nor mimicked by lipoteichoic acid and Pam3CSK4, which could act via the TLR route in combination with an additional catch-up receptor(s) and/or by an yet ill-defined TLR2-independent route.
Mechanisms involved in COX-2 induction in human PMN
Since PMN are terminally differentiated cells which contain regulators of transcriptional control and show signal-dependent activation of mRNA translation 20, we put forward the hypothesis that COX-2 mRNA could be one of those mRNA controlled in the same manner, because abundant expression of COX-2 mRNA was detected in resting human PMN. Contrary to this view, one could argue that the calculation of the predicted secondary structure energy of the 5′- untranslated region (UTR) of COX-2 mRNA is - 36.94 kcal/mol, as judged from the application of RNAfold software 27, 28 to sequences available in data banks (Fig. 3A). This value is lower than that usually associated with transcriptional regulation (-50 kcal/mol, ); however, it fits well with those reported for many transcripts detected using cDNA library arrays which are regulated at the transcriptional level in human monocytes adherent to P-selectin 21. The presence of the 5′-UTR in COX-2 transcripts in human PMN was confirmed by real-time PCR with a set of primers spanning the first 20 nucleotides of exon 1 and exon 2 of COX-2, which gave similar results to PCR reactions using the primers selected from exons 5 and 7 (Fig. 3B). Moreover, the presence of four tracts of 5–8 consecutive pyrimidine bases (underlined in the sequence shown in Fig. 3A) is an additional feature strongly suggesting the possibility of translational control by mTOR 19. As shown in Fig. 4A, preincubation of PMN with 100 nM rapamycin inhibited the induction of COX-2 elicited by complement-coated zymosan, PGN, and mannan, thereby suggesting that the mTOR route could be implicated in the translational regulation of COX-2 protein induction. Since mTOR is integrated in a signaling cascade, the proximal component of which is PI3K, the effect of the PI3K inhibitor wortmannin was addressed. As shown in Fig. 4B, a significant inhibition of COX-2 induction was produced by wortmannin as well as by the translation inhibitor cycloheximide. As shown in Fig. 4C, PGN induced a time-dependent threonine phosphorylation of eIF4E-binding protein (4E-BP1). This provides further evidence of the involvement of the mTOR route, since the phosphorylation of this translation inhibitor by mTOR disrupts its binding to eIF4E and activates cap-dependent translation 29.
Additional mechanisms of COX-2 mRNA regulation were addressed using actinomycin D to inhibit transcription. As shown in Fig. 5A, actinomycin D did not influence the induction of COX-2 protein elicited by mannan and PGN, whereas it fully inhibited the response to LPS. Since COX-2 mRNA stability in some cell types is regulated at the 3′-UTR, PMN were incubated in the presence and absence of 1 μg/mL actinomycin D for 30 min before the addition of PGN to address the half-life of COX-2 mRNA. In the absence of actinomycin D, 53 ± 7% of the starting COX-2 mRNA was detected in control cells versus 70 ± 9% (mean ± SD, n = 3) in PGN-treated cells, 2 h after addition of the stimulus. Actinomycin D treatment induced a further drop of remaining mRNA in vehicle-treated cells, whereas this additional drop was hardly observed in PGN-treated PMN (Fig. 5B). Of note, the decay of β-actin and GADPH mRNA was also significant in actinomycin D-treated PMN (not shown). Further assessment of transcriptional regulation of COX-2 expression was carried out by looking at the effect of 2-hydroxy-4-trifluoromethylbenzoic acid, an inhibitor of both NF-κB and NF-AT 30, 31, which is an useful tool to address in a single-step transcriptional regulation since both transcription factors have been involved in COX-2 regulation in different cell types 32, 33. As shown in Fig. 5C, hydroxy-4-trifluoromethylbenzoic acid lacked a significant effect on COX-2 protein expression in response to all of the stimuli tested, what might suggest that the transcriptional regulation of COX-2 in human PMN is not the main mechanism influencing COX-2 protein expression.
To ascertain whether the above described mechanisms could also be operative in mononuclear phagocytes, experiments were conducted in both human monocytes and monocyte-derived macrophages. Induction of COX-2 protein expression was produced by PGN in monocytes; however, the time course was somewhat different from the findings observed in PMN, since COX-2 expression steadily increased up to 4–8 h (Fig. 6A). Unlike the results obtained in PMN, rapamycin did not influence COX-2 protein expression in monocytes nor in macrophages (Fig. 6A and B), whereas actinomycin D significantly blocked COX-2 protein induction expression in response to zymosan, mannan, PGN, and laminarin (Fig. 6B, lower panel). Taken collectively, these results strongly suggest that different mechanisms can be involved in COX-2 regulation in PMN and mononuclear phagocytes. Noteworthy, IL-8 protein expression was only marginally affected by rapamycin treatment of PMN (Fig. 6C), therefore indicating that rapamycin does not exert a general inhibitory effect on protein induction expression in PMN.
The present results extend the functional relevance of receptors for a variety of PAMP in human PMN regarding their effect on AA metabolism and, in combination with the results obtained in another study 22, unveil a close association between AA release and the induction of COX-2. This indicates that the supply of unesterified AA generated under these conditions is coincidental with the enhanced expression of COX-2 protein. Furthermore, it allows an integrated metabolic route for the production of prostaglandins rather than the usual pathway triggered by archetypal chemoattractants, e.g. formylated peptides and PAF, where the activation of their cognate G-protein coupled receptors is largely linked to the 5-lipoxygenase pathway due to the constitutive status of this enzyme and its post-translational regulation by Ca2+34 and phosphorylation 35. Noteworthy, the stimuli used in this study elicited the production of both LTB4 and PGE2 22, thus fitting well with the existence of two programs for AA metabolism in PMN: (i) the constitutive metabolic route conveying AA into the 5-lipoxygenase pathway, and (ii) the inducible route linked to COX-2 induction expression. The response to PAMP, although showing some features of the first program, fits best with the second program. This fact is of functional relevance, since most studies on AA release have been carried out with xenobiotics as calcium ionophores and phorbol esters, or combining agonists of G-protein-coupled receptors with substances producing a prolongation of Ca2+ transients, for instance cytochalasin B 36 and thapsigargin 37. On this basis, PGN- and mannose-containing molecular patterns emerge as pathophysiologically relevant inducers of AA metabolism in human PMN, at a similar extent as the most functional stimuli so far tested, i.e. opsonized zymosan 38. In contrast, our data do not support a role for lipoteichoic acid and MDP. In the light of these findings, COX-2 induction expression upon microbial invasion seems to be a hallmark of host PMN response to microorganisms containing mannose, e.g. Candida albicans, Leishmania donovani, and Pneumocistis carinni, β-glucan, e.g. Candida albicans, and PGN, a component mainly found in Gram-positive bacteria, where it accounts for around 90% w/w of the cell wall. This fact helps explain why PGN can reach very high concentrations in tissues infected by Gram-positive bacteria 39 and be a major etiological factor in the pathogenesis of pyogenic inflammation or in bacterial-infection-induced inflammatory diseases such as psoriasis 40. LPS and adenosine have been found to induce COX-2 protein expression in human PMN previously 41–43, yet these are agonists distinct from those used in the present experiments, because LPS is a typical activator of the TLR4/NF-κB route and adenosine acts through the adenosine receptor A2A.
Regulation of COX-2 protein expression is a tightly regulated process involving both transcriptional and post-transcriptional events, which may show cell type-dependent mechanisms. Regarding post-transcriptional mechanisms, regulation by both 5′- and 3′-UTR seems possible. The structural hallmarks of 5′-UTR implicated in post-transcriptional control are a 5′-terminal oligopyrimidine tract and a stable secondary structure defined by calculated stabilities of –50 kcal/mol or greater 18, 44. Whereas the secondary structure energy of the 5′-UTR of COX-2 is more consistent with transcriptional control, the presence of polypyrimidine tracts and/or as yet unidentified binding proteins may make COX-2 mRNA responsive to translational control via 5′-UTR in PMN. In addition, the detection of COX-2 mRNA in resting PMN, but not the protein, the effect of rapamycin at the concentrations tested 45–48, and the phosphorylation of 4E-BP1, which allows eIF4E to recognize the 7-methyl cap of mRNA and the initiation of translation, support post-transcriptional regulation.
Regulation of COX-2 protein by the 3′-UTR has been described at least in monocytes, macrophages, and neoplastic cells since COX-2 mRNA contains 23 copies of the AUUUA RNA instability element, which makes this mRNA short-lived and suitable for post-transcriptional regulation by mechanisms dependent on at least the p38-MAP kinase 49 and the PI3K routes 50. Interestingly, a recent report using a preparation of platelets and monocytes has shown that efficient induction of COX-2 protein requires combination of signals involving activation of κB-driven transcriptional activation and stabilization of COX-2 mRNA by silencing the AU-rich mRNA element 51. Whereas cells with a lasting life span might need combination of mechanisms for a timely regulation of COX-2 protein expression, the existence of rapid regulatory mechanisms in short-lived cells such as PMN seems most appropriate.
Using actinomycin D to block transcription we have found a stability of COX-2 mRNA in PMN above that reported for human monocytes 49 and alveolar macrophages 50. Moreover, studies in these cells required preincubation with LPS for periods of 4 to 24 h to allow induction of COX-2 mRNA, whereas in the present study the preincubation period could be suppressed due to the presence of preformed COX-2 mRNA. Since both the p38-MAP kinase and the PI3K routes may be triggered by PGN, it cannot be ruled out that mRNA stability could be an additional mechanism involved in the regulation of the expression of COX-2 protein in PMN, because COX-2 mRNA shows a lower decay in PGN-treated PMN than in vehicle-treated cells. However, our findings cannot be uniquely explained on this basis, since COX-2 mRNA is detected in resting PMN at a level not found in monocytes and macrophages, and the induction of COX-2 protein in PMN shows a more rapid time course than that observed in monocytes under identical conditions 41. Unlike actinomycin D, all of the drugs acting on the cascade involving PI3K, mTOR, and the translation machinery showed an almost complete inhibition of COX-2 protein expression, thus suggesting that COX-2 is another element of the array of constitutive messenger RNA translated by proinflammatory stimuli in human PMN 19.
The MR seems to be the most likely PRR for mannose-containing molecular patterns 22, but regarding PGN, three different types of receptors could be taken into account. In a previous study, we have confirmed the expression of TLR2, TLR1, TLR6, NOD2, and PGN-recognizing protein (PGRP)-S 22, all of which have some binding capacity for PGN or its elementary building blocks. As regards the TLR2 route, PGN is an archetypal ligand 39, even though this view has been challenged recently after a report has shown that the ability of PGN to activate TLR2 might be lost after removal of lipoproteins and lipoteichoic acids 52, thereby pointing to the intracellular receptors NOD1 and NOD2 as the actual PGN receptors in view of their capacity to interact with peptidic moieties of diaminopimelic acid-type PGN, the type detected in Gram-negative bacteria, and MDP, respectively. However, the lack of COX-2 protein induction by MDP is an argument against a chief involvement of NOD2, although a more complex mechanism involving cooperation of receptors could be operative.
The possible involvement of PGRP-S is a challenging hypothesis within a rapidly moving field. Recent studies have disclosed the role of PGRP-S in infection defense on both intracellular 53 and extracellular bacteria, since PGPR-S is present in neutrophil extracellular traps and can bind to both Gram-positive and Gram-negative PGN 54. On this basis, it seems likely that notions already established in insect paradigms, for instance, the PGN catch-up receptor function of Drosophila PGPR-SA, could be extended to the human model 55. Irrespective of the PRR involved, the present data disclose a central role for mannan- and PGN-based signatures in the induction of inflammatory responses in human PMN by showing a strong AA-releasing activity coupled to COX-2 induction and to the production of other inflammatory proteins. Translation of preformed mRNA by an mTOR-dependent route seems to be the most likely mechanism explaining the induction of some of those proteins.
Materials and methods
Mucin-3, zymosan particles, soluble mannan from Saccharomyces cerevisiae, muramyldipeptide (N-acetylmuramyl-L-alanyl-D-isoglutamine), Staphylococcus aureus PGN, and lipoteichoic acid were from Sigma Chemical (St. Louis, MO). The synthetic palmitoylated mimic of bacterial lipopeptides N-palmitoyl-S-[2,3-bis(palmitoyloxy)-2R,S)-propyl]-cysteinyl-(S)-seryl-(S)-lysyl-(S)-lysyl-(S)-lysyl-(S)-lysine trihydrochloride (Pam3CSK4) was from Invivogen. Rabbit Ab reactive to 4E-BP1 phosphorylated at Thr-70 was from Cell Signaling Technology (#9455).
Cell purification and culture
Human PMN were isolated from buffy coats of healthy volunteer donors by centrifugation into Ficoll cushions and sedimentation in Dextran T500. Monocytes were isolated by centrifugation onto Ficoll cushions and adherence to plastic dishes. Differentiation of monocytes into macrophages was carried out by culture of adhered monocytes for 2 weeks in Primaria six-well dishes (BD Biosciences), in the presence of 5% human serum. Since the experimental design could entail long incubation periods with regard to the life span of PMN, the occurrence of constitutive apoptosis was addressed by detecting the surface display of phosphatidylserine with the annexin V-FITC apoptosis detection kit of BD PharMingen to detect early apoptosis, and with propidium iodide vital staining to assess late apoptosis. The samples were analyzed using FACScan and CellQuest program (Becton Dickinson). The protocol of study has been approved by the Bioethical Committee of the Spanish Research Council and Centro de Hemoterapia y Hemodonación de Castilla y León.
Immunoblots of COX-2
The amount of protein in each cell lysate sample was assayed using the Bradford reagent and 50 μg of protein of each sample were loaded on each lane of a 10% SDS/PAGE gel. Proteins were transferred to nitrocellulose membranes using a semi-dry transfer system. The membranes were blocked with dry milk and used for immunoblot with a goat polyclonal antiserum (SC-1745) from Santa Cruz Biotechnology. This was followed by incubation with donkey anti-goat IgG-horseradish peroxidase-conjugated Ab. Detection was performed using the Amersham (Little Chalfont, UK) ECL system.
Assay of proinflammatory protein release by PMN
Culture supernatants were incubated with RayBio® human Inflammation Antibody Array III and Human Cytokine Antibody Array III from RayBiotech, and developed according to the manufacturer's instruction. For this purpose, membranes were incubated with medium obtained from the incubation of PMN with mannan and PGN for 18 h, followed by washing, incubation with HRP-conjugated streptavidin, and detection reaction with the ECL system. Blot densities were assayed using Quantity One® software from Bio-Rad laboratories (Hercules, CA), and normalized with the positive controls incorporated into the membranes. Only variations upon control values showing larger than twofold increase on basal densities were considered positive.
RT-PCR assays for COX-2
Total cellular RNA was extracted by the TRIzol method (Life Technologies, Grand Island, NY). cDNA first strand was synthesized from total RNA by reverse transcription reaction. The reaction mixture contained 0.2 mg/mL total RNA, 2.5 µL H2O, 20 U of RNasin ribonuclease inhibitor, 4 µL buffer 5X, 2 µL DTT 0.1 M, 4 µL dNTP 2.5 mM, 1 µL hexanucleotide 0.1 mM, and 200 U of Moloney-murine leukemia virus reverse transcriptase. The reaction was carried out at 37°C for 60 min in a volume of 20 µL. The cDNA was amplified by PCR in a reaction mixture containing 2 µg of cDNA template, 10 µL H2O, 2.5 µL buffer 10X, 0.75 µL MgCl2 50 mM, 1.0 µL dNTP 2.5 mM, 1.25 µL of each forward and reverse primers and 0.25 µL of Taq DNA polymerase 5 U/μL. The amplification profile for detection included 1 cycle of initial denaturation at 94°C for 5 min, 30 cycles of denaturation at 94°C for 30 s, primer annealing at 60°C for 30 s, and extension at 72°C for 30 s; 1 cycle of final extension at 72°C for 7 min. The expression of β-actin was used as a control for the assay of constitutively expressed genes. PCR products were identified by automatic sequencing of the DNA eluted from the agarose gel by excision of the band under UV light followed by purification using a QIAquick PCR purification kit (Qiagen, Valencia, CA). The oligonucleotide primers used in the PCR reactions were selected from different exons and are shown in Table 2.
Real-time RT-PCR of COX-2
Genomic DNA was removed from purified RNA by treatment with DNase (Turbo-DNA freeTM, Ambion) and used for RT reactions. The resulting cDNA was amplified in a PTC-200 apparatus equipped with a Chromo4 detector (Bio-Rad) using SYBR Green I mix containing HotStart polymerase (ABgene). The primers for this reaction were selected in exon 1 and exon 2 with the purpose of detecting the 5′-UTR of COX-2 transcripts in exon 1. Cycling conditions were as follows: 94°C for 5 min, followed by 40 cycles of 94°C for 30 s, 60°C for 30 s, and extension at 72°C for 30 s. β-Actin and GAPDH were used as housekeeping genes to assess the relative abundance of COX-2 mRNA, using the comparative CT (circle threshold) method for relative expression. This method allows quantifying the relative expression for a given cDNA using the formula 2–ΔCT, where ΔCT = ΔC - ΔC. Thus, 1 arbitrary unit (AU) corresponds to the expression of the housekeeping gene and the results of relative expression changes between treatments can be expressed by the ratio 2–ΔΔCT56. This quantitative approach was initially applied to the study of COX-2 mRNA stability with actinomycin D; however, whereas the expression of mRNA of both β-actin and GADPH did not show significant changes upon stimulation with PGN and mannan, these mRNA were significantly reduced upon actinomycin D treatment. For this reason, results were finally expressed as percentage of the initial COX-2 mRNA according to the formula: (amount of mRNA at a time/amount of mRNA at time 0) × 100.
This work was supported by grants from Plan Nacional de Salud y Farmacia (grant SAF2004–01232), Junta de Castilla y León (Grant CSI05C05), Red Brucella, Red Respira, and Red Temática de Investigación Cardiovascular from Instituto de Salud Carlos III. N.F. is under contract within the Ramón y Cajal Program (Ministerio de Educación y Ciencia of Spain and Fondo Social Europeo), A.G.V. was the recipient of a grant from Instituto de Salud Carlos III, I.V. was the recipient of a grant from Banco de Santander-Central-Hispano.