Molecular Basis for Defining the Pineal Gland and Pinealocytes as Targets for Tumor Necrosis Factor

The pineal gland, the gland that translates darkness into an endocrine signal by releasing melatonin at night, is now considered a key player in the mounting of an innate immune response. Tumor necrosis factor (TNF), the first pro-inflammatory cytokine to be released by an inflammatory response, suppresses the translation of the key enzyme of melatonin synthesis (arylalkylamine-N-acetyltransferase, Aanat). Here, we show that TNF receptors of the subtype 1 (TNF-R1) are expressed by astrocytes, microglia, and pinealocytes. We also show that the TNF signaling reduces the level of inhibitory nuclear factor kappa B protein subtype A (NFKBIA), leading to the nuclear translocation of two NFKB dimers, p50/p50, and p50/RelA. The lack of a transactivating domain in the p50/p50 dimer suggests that this dimer is responsible for the repression of Aanat transcription. Meanwhile, p50/RelA promotes the expression of inducible nitric oxide synthase (iNOS) and the production of nitric oxide, which inhibits adrenergically induced melatonin production. Together, these data provide a mechanistic basis for considering pinealocytes a target of TNF and reinforce the idea that the suppression of pineal melatonin is one of the mechanisms involved in mounting an innate immune response.


INTRODUCTION
The pineal gland is considered the endocrine component of the circadian timing system because it transduces light/dark cycle information into the nocturnal melatonin surge. Melatonin, a highly conserved molecule, acts as an antioxidant in primitive taxa, while in mammals, in addition to its protective effect, it exerts chronobiotic functions (Hardeland and Fuhrberg, 1996;Tan et al., 2010). During the past two decades, several groups have shown the regulatory role of melatonin in the defense response (Lissoni et al., 1994;Nelson et al., 1997;Maestroni, 1998;Guerrero and Reiter, 2002). In addition, the synthesis of melatonin by extra-pineal tissues that are related to defense responses, such as activated polymorpho-and mono-nuclear cells in the blood (Carrillo-Vico et al., 2004), in the peritoneum (Martins et al., 2004), and in the colostrum (Pontes et al., 2006), has been clearly shown. Melatonin's effects include inhibition of nuclear factor kappa B (NFKB), a key transcription factor mediating the mounting of the inflammatory response (Gilad et al., 1998;Beni et al., 2003;Huang et al., 2008;Li et al., 2009;Tamura et al., 2009). Therefore, melatonin is now considered to play an important role in the modulation of inflammatory responses.
More recently, the mounting of inflammatory responses were shown to involve the suppression of the nocturnal melatonin surge , reinforcing a putative bidirectional communication between the pineal gland, and the immune system (Skwarlo-Sonta et al., 2003;Markus et al., 2007). However, it remains unknown whether the pineal gland is able to respond to inflammatory mediators or whether it contains the receptors and the downstream mechanism(s) that mediate the pro-inflammatory agent-induced suppression of the nocturnal melatonin surge.
In the rat pineal gland, the nuclear translocation of NFKB presents a daily rhythm . At the light/dark transition, there is an abrupt reduction of nuclear NFKB content, which remains low until the next light phase, at which point the nuclear level of NFKB rises continuously until the next light/dark transition. This cycle is regulated on a circadian basis by the internal clock; in animals kept in constant darkness, the reduction in nuclear NFKB content occurs at the subjective day/night transition.
The inducible regulation of gene transcription is a central element in the defense of multicellular organisms against environmental, mechanical, chemical, and microbiological stresses. During resting conditions, NFKB inhibitory protein (NFKBI) binds to NFKB dimers, impairing their nuclear translocation (O'Neill and Kaltschmidt, 1997). Activation of membrane receptors leads to NFKBI phosphorylation and ubiquitination, and its subsequent proteasomal degradation. Free NFKB dimers translocate to the nucleus, bind to the kappa B element of gene promoters and induce or repress the transcription of target genes. The five different subunits of NFKB express REL homology domains, which are responsible for binding to the DNA kappa B element. These subunits may or may not contain a transactivating domain (TAD), which is essential for inducing promoter activation. The subunits p50 and p52 have no TAD, whereas RelA, RelB, and c-Rel express the TAD (Ghosh and Hayden, 2008;O'Dea and Hoffmann, 2009). Lipopolysaccharide (LPS), a pathogen-associated molecular pattern, suppresses the synthesis of melatonin and induces TNF production in the rat pineal gland by activating toll-like receptor 4 (TLR4), which triggers the nuclear translocation of NFKB (da Silveira Cruz- Machado et al., 2010). Moreover, TNF transiently inhibits the noradrenaline-induced transcription of the arylalkylamine-N -acetyltransferase gene (Aanat ), which is the enzyme that converts serotonin to N -acetylserotonin, the immediate precursor of melatonin (Fernandes et al., 2006). In this study, we aimed to identify the downstream pathway that mediates TNFinduced melatonin suppression in the rat pinealocyte. The present paper shows the cellular distribution of TNF receptors, the nuclear translocation of NFKB dimers induced by TNF in pinealocytes and the functional expression of this pathway.

ANIMALS
Prepubertal female Wistar rats were kept under a 12:12 h light/dark cycle (lights on at 07h30, named zeitgeber time zero -ZT0) with water and food provided ad libitum. The animals were killed by decapitation between ZT9 and ZT10. All experiments were carried out in compliance with the ethical standards of our institution (Ethics Committee of the Institute of Bioscience of the University of São Paulo; license 081/2008) and the recommendations of the National Council on Experimental Animal Control (CONCEA).

ORGAN CULTURE
The pineal gland was cultured according to Ferreira et al. (1994). Briefly, the glands were cultivated for 48 h in a 24-well plate containing BGJb medium (2 mM glutamine, 100 U/mL penicillin, 10 μM streptomycin, 37˚C, 95% O 2 , 5% CO 2 ). Each well (200 μL) contained one gland, and the medium was replaced every 24 h. The length of culture permitted the complete denervation of the gland (Parfitt et al., 1976).

CELL VIABILITY TEST
Cell viability was assessed using an MTT assay. Pinealocytes seeded at a density of 0.5 × 10 5 cells/well in 96-well plates (150 μL/well) were either treated or not treated with TNF (80 μM) and MTT (0.5 mg/mL) for 4 h at 37˚C. NFKB-DNA binding and iNOS activity were inhibited by incubating the cells with PDTC (25 μM) or 1400 W (1 μM), respectively. The formazan dye used was dissolved in dimethyl sulfoxide (DMSO 100%). The plates were read in a microplate spectrophotometer (SpectraMAX 250, Molecular Devices, CA, USA) at 540 nm. Cell viability was defined based on the percentage of optic density (OD) in each group as compared to the control group.

IMMUNOHISTOCHEMISTRY
For tissue preparation, the animals were deeply anesthetized by intramuscular injection of ketamine (160 mg/kg) plus xylazine (40 mg/kg) and perfused transcardially with 150 mL saline solution followed by 1000 mL of cold 4% paraformaldehyde fixative solution, pH 9.5. Each pineal gland was removed from the skull and cryoprotected in the same fixative solution plus 20% sucrose at 4˚C for 3 days, followed by cryoprotection in PBS plus 30% sucrose at 4˚C. The pineal glands were embedded in Tissue-Tek freezing medium and stored at −80˚C for no longer than 2 weeks before sectioning with a cryostat (30 μm). The immunohistochemistry assay was performed with free-floating sections incubated in the blocking solution (1% BSA, 0.3% Triton X-100 in PBS) for 1 h at room temperature. The sections were then incubated with the primary antibodies under constant agitation for 48 h at 4˚C, rinsed with PBS (30 min) and incubated with the secondary antibodies for 90 min. After rinsing, the pineal sections were mounted on gelatin-coated slides. The negative staining controls omitted the primary antibodies. Under these conditions, staining was completely abolished.

IMMUNOCYTOCHEMISTRY
Immunocytochemistry was performed as described by da Silveira Cruz- Machado et al. (2010). Briefly, pinealocytes were cultivated on chamber slides (8 wells, 0.5 × 10 5 cells, 18 h), fixed in 4% cold paraformaldehyde, permeabilized with 0.5% saponin and incubated with blocking buffer (1% BSA, 0.5% saponin, and 0.3 M glycine in PBS) to avoid non-specific staining. The primary antibodies were diluted in PBS plus 1% BSA and incubated for 18 h at 4˚C, rinsed with PBS and then incubated with fluorescent secondary antibodies for 1 h at room temperature. Next, the cell nuclei were stained with DAPI (300 μM, 5 min) at room temperature. Controls were prepared by omitting the primary antibodies. Under this condition, the staining was completely abolished.

Supershift assay
Nuclear factor kappa B subunits were identified from a pool of four samples that were either stimulated or not stimulated with TNF (30-90 ng/mL, 5 min) and incubated with 2 μg/mL of rabbit polyclonal affinity-purified antibodies for RelA, p50, p52, c-Rel, RelB, and Bcl3 from Santa Cruz Biotechnology (sc-109x, sc-114x, sc-298x, sc-70x, sc-226x, and sc-185x, respectively) for 45 min at room temperature before the addition of a 32 P-NFKB probe. EMSA was carried out as described previously.

TNF-INDUCED NITRIC OXIDE PRODUCTION IN ISOLATED PINEALOCYTES
Nitric oxide production was detected in isolated pinealocytes using confocal laser-scanning microscopy and the intracellular indicator DAF-FM, which forms a fluorescent product after reacting with nitrite ions produced by the spontaneous oxidation of nitric oxide (Kojima et al., 1998).
Nitric oxide was detected in pinealocytes using the protocol described by Tamura et al. (2009). Pinealocytes were grown on glass coverslips (200 μl DMEM, 37˚C, 5% CO 2 ) for 17 h, and then incubated for 1 h in a saline solution: NaCl 150 mM, KCl 5 mM, CaCl 2 2 mM, NaHCO 3 15 mM, glucose 11 mM, and Larginine 0.1 mM at pH 7.4. Next, TNF (80 ng/mL), the antagonist of iNOS (1400 W, 1 μM) or NFKB (PDTC, 25 μM) was added. TNF was incubated for 2 h, PDTC and 1400 W were pre-incubated for 30 and 50 min, respectively, before adding TNF. During the final 50 min after adding TNF, the cells were loaded in the dark with DAF-FM DA (5 μM). The cells were then washed to remove excess probe and mounted on the stage of an inverted microscope equipped with a 40× oil-immersion objective.
Nitric oxide fluorescence was measured using an argon laser with excitation and emission wavelengths of 488 nm and 515-530, respectively, using a confocal laser-scanning microscope (ZEISS LSM 510). The fluorescence was quantified in three different fields per well, counting 7-10 cells/field. The cell perimeter was defined as the region of interest (ROI), and the increase in nitric oxide production was estimated as the percentage by which the nitric oxide donor, SNP (1 mM), increased intracellular nitric oxide.

DATA ANALYSIS
Data are presented as the mean ± SEM. Gel shift assays were quantified using ImageJ. Statistical analyses were performed using ANOVA followed by the Newman-Keuls test. Values of p < 0.05 were considered statistically significant.

TNF-R1 DISTRIBUTION IN THE DIFFERENT CELL TYPES IN PINEAL PARENCHYMA
The rat pineal gland parenchyma is composed of different cell types, most of which are pinealocytes (approximately 90%) and glial cells. First, we found that immune-like TNF-R1 was diffusely expressed in pineal gland sections, as shown in two different glands (Figures 1A,D). A diffuse image using the same antibody was observed in renal cell carcinoma cryosections (Harrison et al., 2007) and melanoma cell (Gray-Schopfer et al., 2007).
To identify the cell types, the astrocytes and microglia were identified with selective antibodies (GFAP for astrocytes and ED-1 for microglia). Astrocytes and microglia have specific localizations in the pineal gland. The immunoreactivity for GFAP was confined to the proximal region, specifically near the pineal stalk (Figure 1B), whereas the immunostaining for ED-1 revealed a diffuse pattern (Figure 1E). When the images of the two cell markers were merged with TNF-R1 immunostaining, co-localization with astrocytes ( Figure 1C), and microglia ( Figure 1F) was observed.
Analysis of the images at higher amplification shows that almost all astrocytes present immunoreactivity to TNF-R1; however, in the case of microglia, only some of the cells marked with ED-1 were immune-stained with antibodies against TNF-R1 (Figure 2).

TNF-R1 IS EXPRESSED IN PINEALOCYTES
To confirm the presence of TNF-R1 in pinealocytes, these cells were isolated and immunostained for TNF-R1 (Figure 3). We observed a dynamic variation in the immunoreactivity to TNF-R1, depending on the length of time that the pinealocytes were incubated with TNF (80 ng/mL; Figure 3). The 30-min incubation period did not change the fluorescence intensity. However, incubation for 60 or 180 min resulted in 60% reduction in immunofluorescence intensity. The MTT viability test assured that the concentrations of TNF used did not lead to cell death. The MTT assay showed that cell death was observed only at much higher concentrations of TNF and that the other pharmacological tools used, such as PDTC and 1400 W, also did not result in cell death. Therefore, we may conclude that the membrane detection of immunoreactivity for TNF-R1 is altered by long-lasting incubation with TNF. Thus, the downstream reactions triggered by activation of TNF-R1 were determined by incubating pinealocytes or pineal glands for less than 30 min.

TNF REDUCES NFKBIA
One of the downstream events that transduce TNF-R1 signals is the nuclear translocation of NFKB dimers, which requires degradation of the inhibitory protein that sequesters these dimers in the cytoplasm. We observed a significant reduction of the  represents nuclei stained by DAPI. The yellow merged images show positive co-localization of TNF-R1 in astrocytes (C) and microglia (F). The filled arrow (F) indicates that microglia (red) are negative for TNF-R1, and the arrowhead shows positive co-localization between microglia and TNF-R1 (yellow). The circle shows a pinealocyte expressing TNF-R1 (green). Scale bar = 20 μm fluorescence intensity of the immunostaining for NFKBIA in isolated pinealocytes incubated for 10 min with TNF (80 ng/mL, Figure 4). Thus, TNF triggers the NFKB signaling throughout the degradation of NFKBIA.

TNF INDUCES NUCLEAR TRANSLOCATION OF NFKB IN RAT PINEAL GLAND
The effect of TNF on NFKB nuclear translocation was determined using EMSA in the cultured pineal glands stimulated with TNF (30 ng/mL) for 5-60 min. The NFKB nucleotide probe revealed two complexes (C1 and C2), and TNF stimulation transiently increased the nuclear translocation of C1 (Figure 5). Significant increases in the translocation of the C1 subunit were observed at 5-min intervals.
The translocation of NFKB subunits was evaluated by incubating cultured pineal glands with TNF (30 ng/mL) for 5 min. The supershift assay for DNA-protein complexes was conducted with specific antibodies for p50, RelA, p52, c-Rel, RelB, and Bcl-3. Only the C1 complex was supershifted; therefore, we could not identify C2 (Figure 6).
Nuclear extracts from non-stimulated glands were supershifted with p50 and RelA antibodies, but not with the other antibodies tested. p50 antibodies supershifted all complexes with C1, whereas RelA antibodies promoted a partial shift (Figure 6A). Figure 6B. It is interesting to note that we observed two supershifted bands for p50, but only one for RelA. On the other hand, p50 supershifted all of the C1 bands, but RelA antibodies only partially supershifted the C1 bands. Therefore, we conclude that in cultured pineal gland both p50/p50 and p50/RelA dimers are found in the nuclear extract.

Autoradiogram is shown in
The effect of TNF was evaluated by incubating glands for 5 min in various concentrations of TNF (10, 30, 60, and 90 ng/mL). TNF induced the nuclear translocation of both dimers (p50/p50 and p50/RelA; Figure 7). Lower concentrations increased the nuclear translocation of p50/RelA, whereas higher concentrations were required to translocate both dimers.

EXPRESSION OF iNOS AND PRODUCTION OF NITRIC OXIDE
The expression of iNOS was evaluated using immunofluorescence and pharmacological methods in isolated pinealocytes. Isolated pinealocytes were incubated for 120 min with TNF (30-90 ng/L). The cells were incubated sufficiently long to allow iNOS and nitric oxide to accumulate in amounts detectable by our techniques.
The expression of iNOS was TNF dose-dependent (Figures 8A,B) and was blocked by inhibition of NFKB with PDTC (25 μM, Figure 8C). This de novo synthesized enzyme was responsible for the TNF-induced nitric oxide production, as it was blocked by a selective iNOS antagonist, 1400 W (1 μM, Figure 8D).

www.frontiersin.org FIGURE 3 | The TNF-R1 expression in dispersed pinealocytes. (A)
Representative immunocytochemistry images of cultured pinealocytes showing the constitutive expression of TNF-R1 (stained in red, left-hand side). Incubation with TNF (80 ng/mL) for 60 min reduces the detection of TNF-R1 (right-hand side). Blue staining represents fluorescent nuclei stained using DAPI. (B) Quantification of the fluorescence of pinealocytes stimulated (gray bars) or not stimulated (white bars) with TNF (80 ng/mL) for 30-180 min. Data are expressed as the mean ± SEM; n = 4 independent cultures. In each well, we counted 21 cells in three different, randomly chosen fields. *p < 0.05 vs. no TNF. Scale bar = 20 μm.

DISCUSSION
The pineal gland, which is classically considered a neuro-humoral transducer of photic environmental information, is now considered an integral player in the immune response Skwarlo-Sonta et al., 2003;Markus et al., 2007). The indolamine melatonin has long been recognized as an immune-modulatory agent (Guerrero and Reiter, 2002;Hotchkiss and Nelson, 2002). However, only recently has the pineal gland per se been studied with regard to its role in the mounting of the inflammatory response . In the context of the immunepineal axis, the suppression of pineal melatonin synthesis favors the mounting of an inflammatory response. The presence of a receptor repertoire necessary for mediating the response to LPS and TNF (da Silveira Cruz- Machado et al., 2010) reinforces the idea that this gland participates in the regulation of the innate immune response. Here we studied the role of the pinealocytes and the mechanism triggered by the pro-inflammatory cytokine TNF as a first approach toward understanding the cellular mechanisms by which the pineal gland contributes to the innate immune response.

FIGURE 4 | NFKBIA expression in dispersed pinealocytes. (A)
Representative immunocytochemistry images demonstrating expression of NFKBIA (stained in red) in cultured pinealocytes in the absence (left-hand panel) or in the presence of TNF (80 ng/mL, for 10 min; right-hand panel). Blue staining represents the fluorescent nuclei issued with DAPI. (B) Quantification of the fluorescence of pinealocytes stimulated (gray bars) or not stimulated (white bar) with TNF (80 ng/mL) for 1-60 min. Data are shown as the mean ± SEM; n = 3 independent cultures. *p < 0.05 vs. no TNF (white bar). Scale bar = 10 μm.
Tumor necrosis factor is a major pro-inflammatory mediator (Wajant et al., 2003) that signals through TNF-R1 and TNF-R2. TNF-R1 is expressed in many cells and tissues, whereas TNF-R2 is mostly found in cells of the immune system (Hehlgans and Männel, 2002). Soluble TNF has high affinity for TNF-R1 whereas membrane-bound TNF interacts with TNF-R2 (Grell et al., 1998;McCoy and Tansey, 2008). In the present study, we tested the effects of exogenous TNF; therefore, we focused our attention on TNF-R1. First, we determined the cellular localization of TNF-R1 in the pineal gland. Then, after confirming its presence in the pinealocytes, we explored the signal transduction pathway that translates pinealocyte responses to TNF.
The three most important cells in the rodent pineal gland are astrocytes, microglia, and pinealocytes. Astrocytes, located in the stalk along the entrance site of the arteries and veins, are immunestained by GFAP (Luo et al., 1984;Zang et al., 1985;Schröder and Malhotra, 1987;Berger and Hediger, 2000;Jiang-Shieh et al., 2003), whereas several subtypes of cells that are labeled by specific antibodies and diffused alongside pinealocytes in the pineal parenchyma are collectively named microglia (Jiang-Shieh et al., 2003). Astrocytes and microglia were identified in the present work based on the immunoreactivity to GFAP and ED-1, respectively, whereas pinealocytes were isolated and cultured to confirm the Frontiers in Endocrinology | Cellular Endocrinology expression of TNF-R1 in these cells. All three cell types expressed TNF-R1 under resting conditions. Therefore, TNF signaling could integrate their responses. In this context, it is interesting to mention that the pineal gland could respond to circulating cytokines as well as TNF produced by the gland itself, as observed in cultured pineal glands stimulated with LPS (da Silveira Cruz- Machado et al., 2010).
The time-dependent reduction in the immunoreactivity of TNF-R1 when the isolated pinealocytes were incubated for 30-180 min suggests a reduction in the number of receptors available in the membrane. Although we did not pursue this subject, two mechanisms involved in reducing the number of TNF receptors available in the membrane are known (for reviews, see Higuchi and Aggarwal, 1994;Hehlgans and Männel, 2002). The first mechanism is related to receptor shedding, leading to a protein that binds circulating TNF. This mechanism is preferential for the reduction of TNF-R2. The second mechanism involves internalization, which is mandatory for TNF-induced apoptosis. Because activation of NFKB is independent of internalization, the time interval chosen for analyzing NFKB activation was shorter than that involved in receptor internalization.
The presence of TNF-R1 in pinealocytes strongly indicates that TNF directly affects melatonin production. We have observed previously that incubation of cultured rat pineal glands with TNF inhibits the noradrenaline-induced transcription of Aanat and the synthesis of N -acetylserotonin, the immediate precursor of melatonin (Fernandes et al., 2006). These in vitro observations were corroborated by clinical data that show a suppression of the nocturnal melatonin surge in patients with high levels of circulating TNF, such as in the presence of sepsis (Mundigler et al., 2002), myocardial stroke (Domínguez-Rodríguez et al., 2002), and mastitis (Pontes et al., 2006). In addition, an analysis of the levels shifted bands corresponding to each column in the representative gel. The first and second peaks correspond to complexes C2 and C1, respectively. Anti-RelA and anti-p50 supershifted the C1 complex in a partial and a total manner, respectively. Each point was obtained from a pool of 3-4 glands.
www.frontiersin.org  of TNF and melatonin in mothers whose infants were delivered by cesarean section, a condition that induces an acute inflammatory response, has shown that the daily melatonin rhythm was recovered only in mothers with no circulating TNF (Pontes et al., 2007). Therefore, the hypothesis that activation of TNF-R1 triggers a transduction pathway that inhibits noradrenaline-induced melatonin production is supported by experimental and clinical findings.
Here, we show for the first time that activation of TNF-R1 in pinealocytes reduces the expression of NFKBIA, allowing the migration of NFKB to the nuclei, as observed in immunecompetent cells and neurons (Lawrence, 2009). In rat pineal glands, TNF induced the translocation of two NFKB dimers: p50/p50 and p50/RelA. The major difference between these two dimers is that TAD is present only in p50/RelA (Hayden and Ghosh, 2004). The binding of the dimer p50/p50 to the putative kappa B element in the promoter of Aanat  could explain the inhibition of its transcription, leading to the inhibition of the melatonin biosynthetic pathway (Fernandes et al., 2006). The dimer p50/RelA induces the transcription of a set of genes involved in the inflammatory response. Among several other proteins, iNOS is induced by p50/RelA in immunecompetent cells and in many other cell types, such as neurons (Bethea et al., 1998;Arias-Salvatierra et al., 2011), endothelial cells (Tamura et al., 2009), and muscle cells (Katsuyama and Hirata, 2001). We evaluated the expression of iNOS and the production of nitric oxide by isolated pinealocytes stimulated with TNF. Taking into account the pharmacological inhibition of iNOS expression and nitric oxide production by PDTC and 1400 W, respectively, we concluded that the effect of TNF is mediated by NFKB. FIGURE 9 | Schematic representation of TNF's dual effect on rat pinealocytes. TNF stimulates TNF-R1, leading to internalization of the receptor and activation of the NFKB pathway. This activation is mediated by the degradation of NFKBIA and depends on its phosphorylation (Neumann and Naumann, 2007). Next, we observed the nuclear translocation of the p50/p50 and p50/RelA dimers. p50/p50 NFKB acts to inhibit gene transcription, whereas p50/RelA activates gene transcription. We have shown previously that TNF inhibits the transcription of Aanat, which encodes the key protein for melatonin synthesis (Fernandes et al., 2006). In fact, TNF inhibits melatonin synthesis. Here, we also show that the nuclear translocation of p50/RelA induces the transcription of iNOS, a hallmark enzyme of inflammatory responses. We suggest that the mechanism of action of TNF, as determined in the present paper, is the molecular basis for understanding the inhibition of melatonin synthesis that occurs at the beginning of an innate immune response.
Three different isoforms of nitric oxide synthase are expressed in the pineal gland; the endothelial and neuronal isoforms are constitutive and signal through cyclic GMP (Lin et al., 1994;Maronde et al., 1995;López-Figueroa and Møller, 1996a;Spessert et al., 1998;Jacobs et al., 1999), whereas the induced isoform is not present in basal conditions (Jacobs et al., 1999) and signals thorough the nuclear translocation of NFKB (this work; Kaur et al., 2007). The constitutive NOS are localized in blood vessels and near the stalk of the rat pineal gland (López-Figueroa and Møller, 1996b), whereas iNOS, as mentioned above, is found in the pinealocytes. In addition, the production of nitric oxide by activation of the constitutive NOS or stimulation of the gland with 8-Br-cyclic GMP (Maronde et al., 1995;Spessert et al., 1998) has no effect on the production of melatonin induced by adrenoceptor stimulation. On the other hand, production of nitric oxide by iNOS (this work) or the use of nitric oxide donors, such as sodium nitroprussiate and 3-morpholino-sydnonimine-1 (SIN-1), which releases a high concentration of nitric oxide, inhibits adrenergically induced melatonin production. Therefore, this could be a second mechanism involved in suppression of the nocturnal melatonin surge by TNF.
Together, these data clearly indicate that the activation of TNF-R1 in pineal glands leads to the degradation of NFKBIA and the nuclear translocation of the dimers p50/p50 and p50/RelA (Figure 9). Although these dimers act through different mechanisms, we propose that both mechanisms could suppress adrenergically induced melatonin synthesis. The p50/p50 dimer directly inhibits Aanat gene transcription, whereas the p50/RelA dimer inhibits melatonin synthesis by inducing the production of NO in the pinealocytes.
In conclusion, this study provides mechanistic evidence for considering the pineal gland a key participant in the innate immune response. Understanding how the synthesis of melatonin is suppressed to permit the proper mounting of an inflammatory response is the initial step in evaluating why the daily rhythm of melatonin is not restored in some, but not all, chronic diseases. Considering the pinealocytes a target for cytokines improves our understanding of the role played by the pineal gland in organism defense.

AUTHOR CONTRIBUTION
Claudia Emanuele Carvalho-Sousa: Acquisition and analysis/interpretation of data, writing the manuscript; Sanseray da Silveira Cruz-Machado: acquisition of data, writing the paper; Eduardo Koji Tamura: acquisition of data; Pedro A. C. M. Fernandes: acquisition of data, writing the manuscript; Luciana Pinato: acquisition of data; Sandra M. Muxel: acquisition of data; Erika Cecon: acquisition of data; Regina P. Markus: Concept/design, data interpretation, writing the manuscript. All authors approved submission of the manuscript to FCE.

ACKNOWLEDGMENTS
This work was supported by grants from the National Council for Research and Development (CNPq) and the Foundation of Research of the Estado de São Paulo (FAPESP; thematic grant 07/07871-6). We thank Débora Aparecida Moura for technical assistance. Claudia Emanuele Carvalho-Sousa, Sanseray da Silveira Cruz-Machado and Erika Cecon are PhD fellows and Eduardo Koji Tamura is a post-doctoral fellow from FAPESP. Regina P. Markus is a senior fellow of CNPq.