Dietary melatonin selectively reverses age-related changes in cortical cytokine mRNA levels, and their responses to an inflammatory stimulus

The basal levels of expression of mRNA of cytokines, interleukin-6 (IL-6) and tumor necrosis factor (TNF-(cid:1) ), in the cerebral cortex of 5 and 26 month-old male B6C3F1 mice have been compared. In addition, the responsivity of animals of differing age to an inﬂammatory stimulus (lipopolysaccharide, LPS) has been studied. Basal levels of both of these cytokine mRNAs were elevated in aged animals relative to the younger group. However LPS administration led to a robust increase in cytokine mRNA levels in the younger animals but in aged mice, there was either an unchanged (IL-6) or a depressed (TNF-(cid:1) ) response. Administration of dietary melatonin (200 ppm) to aged mice for 6 weeks prior to sacriﬁce, resulted in reduction of basal levels of cytokine mRNA to values found in the younger animals. Furthermore, following administration of LPS to melatonin fed animals, cerebral cytokine mRNA levels were signiﬁcantly elevated rather than being unchanged or depressed. Taken together these ﬁndings reﬂect a trend in the cortices of melatonin-treated aged mice, to more closely approximate the status of younger mice. For comparative purposes, parallel studies were carried out using an immunologically active organ (spleen) and a non-neural organ with a low rate of cell turnover (heart muscle). In both these tissues, basal levels of cytokine mRNAs of animals of either age were very low, and there was a marked positive response to LPS. Dietary melatonin had no effect on the responses of TNF-(cid:1) mRNA to LPS but attenuated the reaction of splenic IL-6 mRNA, thus bringing the response closer to that of the younger mice. © 2002 Elsevier Science Inc. All rights reserved.


Introduction
There is evidence that the immune system may be compromised during aging [12].In the central nervous system, part of this impairment may be expressed as an inappropriate and prolonged response to inflammatory stimuli [7].Microglial cells also become progressively active with normal aging [39,45].Lipopolysaccharide-induced chronic inflammation can cause extensive astrogliosis in the temporal lobe regions of the rat brain [51].Such activation of the astroglial cells is associated with changes, which mirror those seen in the AD brain [21,51].Age-related declines of function have also been associated with higher levels of oxidative damage [10,23,44].Various antioxidant treatments have been proposed in an attempt to retard deleterious changes associated with senescence [16,18,20,40].
However, many antioxidants such as ␣-tocopherol and coenzyme Q [27] have limited access to the brain [50].
In view of its combination of lipophilic and hydrophilic properties, melatonin can readily cross the blood-brain barrier and has been proposed as a potentially neuroprotective agent [36].Melatonin has been reported to have antioxidant activity [25,26].It can act directly by scavenging reactive species such as hydroxyl [47], H 2 O 2 [46], NO [48], and peroxynitrite [6,56]; or it can act indirectly by inducing antioxidant enzymes [25,38].Extended deprivation of melatonin can lead to elevated levels of oxidative damage [37].This agent can reduce the extent of injury to the CNS caused by a wide range of neurotoxicants including iron-induced necrosis [25], mercury [30], kainate [28], cyanide [52] and paraquat [29].Melatonin levels decrease with age [33].Since melatonin can prolong survival time of mice [2,31] albeit associated with increased spontaneous tumor incidence according to one report [2] -this is likely to be relevant to the aging process [36], even if the exact mechanism is incompletely understood [54].We have found dietary melatonin can prevent age-related changes in both cerebral mitochondrial function [42] and in levels of production of nitric oxide and reactive oxygen species (ROS) [9].This agent has also been reported to reverse age-related behavioral changes [4].However, melatonin may not be a direct antioxidant [26], but may act by enhancing gene expression of antioxidant enzymes [3].
A relation between oxidative stress and inflammatory events within nervous tissues has often been described [8,18].In the current study, the effect of an inflammatory stimulus (LPS) on cortical levels of mRNAs for cytokines IL-6 and TNF-␣, was studied in mice of two ages.Melatonin was shown to wholly or partially reverse age-related changes in both resting levels and inflammation-induced fluxes of these parameters.

Animal treatment
Male B6C3F1 mice, a hybrid between C57BL/6 and C3H from Harlan Labs (Indianapolis, IN), aged 5 months (young group) and 26 months (old group), were housed six per cage and were maintained on a 12 h light/dark cycle in a temperature controlled (20 Ϯ 1°C) room.Food and water were provided ad lib. to 3 animals per group.All food was in the form of pellets manufactured by compressing thoroughly mixed, finely powdered ingredients (Dyets Inc., Bethlehem, PA).Food was stored at 4°C in sealed, plastic bags as received from the manufacturer.The minimal basal diet (Dyets #101101) consisted of 50% sucrose and 26% casein (w/w) as well as a minimal salt and vitamin mix.For One group of mice was fed basal diet supplemented with 200-ppm (w/w) melatonin (Dyets #101475; melatonin from Sigma, St. Louis, Mo) for 6 weeks; based on measured food consumption, this provided an average melatonin dosage of 29 mg/kg/day for the young animals and 23 mg/kg/day for the old.Control groups were fed unsupplemented basal diet.

RNA extraction
Mice were killed by cervical dislocation; tissues were excised quickly, immediately frozen on dry ice (brains) or in liquid nitrogen (other organs) and stored at -70°C.Total RNA was extracted using the TRI REAGENT Kit (Molecular Research Center, Inc., Cincinnati, OH), following the manufacturer's protocol.RNA concentrations were determined by absorption at 260-nm wavelength.Purity was monitored by measuring the ratio of absorbance at 260 nm to that at 280 nm.

Preparation of cDNA probes
A rat IL-6 cDNA-containing plasmid (gift from Drs. Wolfgang Northemann and Georg Fey, Scripps Research Foundation, La Jolla, CA) was transformed into JM101 cells and then digested with PstI and BamHI restriction enzymes to obtain the insert for hybridization.
A murine TNF-␣ cDNA-containing plasmid (gift from George N. Davetelis) was transformed into JM101 cells.The TNF cDNA was inserted in the pUC9 plasmid at the Pst I and Bam HI restriction sites.For the purpose of hybridization, however, Pst I and Eco RI were used as restriction enzymes for the purification of the insert from the vector because Eco RI cleaves off a portion of the 3Ј-untranslated region of the insert that contains a TTATTTATT consensus sequence that is also common to IL-1 and other cytokines [7] and might result in unwanted cross-hybridization.

Northern blot analysis
Aliquots of total RNA (10 g each, as determined from absorbance at 260 nm wavelength and verified by gel ethidium bromide fluorescence intensity) were denatured with formaldehyde and formamide, electrophoresed on 1.2% agarose gel containing 6% formaldehyde and transferred onto nylon Zeta-probe blotting membranes (Bio-Rad Laboratories, Hercules, CA).This standardization was selected in preference to use of housekeeping genes, actin-␥ or glyceraldehyde phosphate dehydrogenase (GAPDH).This was because these genes were responsive to varying age and to melatonin supplementation (data not shown).Membranebound RNA was then hybridized with one of the cDNA probes labeled with [ 32 P]dCTP using the RTS Radprime System (Life Technologies, Rockville, MD) to yield a specific activity of approximately 10 9 cpm/g.The membranes were autoradiographed for periods varying from 8 h to 7 days at -70°C on x-ray film (X-OMAT AR, Kodak, and Rochester, NY).A densitometer (Eagle Eye image-processor combined with DNA Scan signal analysis software, Stratagene, San Diego, CA) was used to quantify the signals as area-integrated optical density.

Statistical analyses
Differences between groups were assessed by one-way Analysis of Variance followed by Gabriel's Test.The acceptance level of significance was P Ͻ 0.05 using a twotailed distribution.

Results
IL-6 and TNF-␣ mRNA expression levels were measured in brain and also, for comparative purposes, parallel studies were carried out using an immunologically active tissue (spleen) and a non-neural tissue with a similarly low rate of cell turnover (heart muscle).

Basal levels
Unstimulated levels of both of these cytokine mRNAs were elevated in the brains of aged animals relative to the younger group (Figs. 1, 2).This was not the case for heart (Fig. 3) and spleen (Fig. 4).

Lipopolysaccharide (LPS) treatment
LPS administration led to a robust increase in cerebral cytokine mRNA levels in the younger animals, but not in aged mice, where LPS resulted in either an unchanged (IL-6, Fig. 1) or a depressed cytokine mRNA level (TNF-␣, Fig. 2).In both heart and spleen of animals of either age, levels of both cytokine mRNAs were greatly elevated following LPS injection (Figs. 3, 4).

Dietary melatonin
Administration of dietary melatonin (200 ppm) to aged mice for 6 weeks prior to sacrifice, resulted in restoration of basal levels of cerebral cytokine mRNA to values found in the younger animals.Furthermore, in melatonin-treated older mice, cerebral cytokine mRNA levels were significantly elevated by administration of LPS rather than being unchanged or depressed (Figs. 1, 2).In heart and spleen, dietary melatonin had no effect on the responses of TNF-␣ mRNA to LPS (Figs. 3, 4).The reaction of splenic IL-6 mRNA to LPS, was attenuated in melatonin-treated aged mice, thus bringing the response closer to that of the younger mice (Fig. 4).

Discussion
In all tissues studied, basal levels of cytokine mRNAs of young animals were very low, but in the brain, were ele-Fig.1. Levels of IL-6 mRNA in cerebral cortex of 5 and 26 month old mice.Values were densitometrically quantitated and are expressed as units of integrated optical density Ϯ standard error (n ϭ 3).Some animals received a prior injection of lipopolysaccharide (LPS) while two groups of older animals were fed a melatonin-containing diet.* ϭ differs significantly from corresponding value for 5 month-old mice.† ϭ differs significantly from value for corresponding group not treated with LPS.Fig. 2. Levels of TNF-␣ mRNA in cerebral cortex of 5 and 26 month old mice.Values were densitometrically quantitated and are expressed as units of integrated optical density Ϯ standard error (n ϭ 3).Some animals received a prior injection of lipopolysaccharide (LPS) while two groups of older animals were fed a melatonin-containing diet.* ϭ differs significantly from corresponding value for 5 month-old mice.† ϭ differs significantly from value for corresponding group not treated with LPS.Fig. 3. Levels of IL-6 mRNA and TNF-␣ in heart of 5 and 26 month old mice.Values were densitometrically quantitated and are expressed as units of integrated optical density Ϯ standard error (n ϭ 3).Some animals received a prior injection of lipopolysaccharide (LPS) while two groups of older animals were fed a melatonin-containing diet.* ϭ differs significantly from corresponding value for 5 month-old mice.† ϭ differs significantly from value for corresponding group not treated with LPS.Fig. 4. Levels of IL-6 mRNA and TNF-␣ mRNA in spleen of 5 and 26 month old mice.Values were densitometrically quantitated and are expressed as units of integrated optical density Ϯ standard error (n ϭ 3).Some animals received a prior injection of lipopolysaccharide (LPS) while two groups of older animals were fed a melatonin-containing diet.* ϭ differs significantly from corresponding value for 5 month-old mice.† ϭ differs significantly from value for corresponding group not treated with LPS.vated with age.In line with the current study, there are several other reports of a modest elevation of non-stimulated IL-6 levels within the aging brain [19,53].There is previously published evidence that systemically administered LPS is able to rapidly and directly target the CNS [21,15,22,49].
In tissues from young animals there was a marked positive response of TNF-␣ expression to LPS, and this was maintained with age in heart and spleen.Our young-animal results parallel reported changes in serum TNF-␣ protein levels, which were increased five-fold by LPS administration [5].However, we found this reactivity to be absent in the cortex of aged mice.Our finding of a weaker response to LPS in the cortex of older animals is in accord with the reported ability of LPS to increase the number of activated microglia and impair behavior in young, but not old rats [17].
In aged mice, a large response of IL-6 mRNA expression to LPS treatment occurred in heart and spleen tissue.This heightened reaction has also been described in several other tissues including lung and kidney [43].In contrast, cerebral IL-6 mRNA levels of aged mice following LPS exposure, were greatly attenuated relative to the young.This suggests that, while hyper-reactivity of immune responses may characterize many aged organ systems, this is not the case for the central nervous system.It appears that in the brain, aging may be associated with high basal levels of cytokine mRNAs, rather than increased susceptibility to induction.This high background and attenuated response implies a reduced signal to noise ratio, which may have adverse physiological consequences.
Melatonin had no pronounced effect on the basal or induced levels of mRNAs for cytokines of the non-neural tissues studied.However, in the cortices of melatonintreated aged mice, levels of both cytokine mRNAs were depressed to values paralleling those present in younger mice.
There is evidence that melatonin can enhance immune processes in cells derived from aged animals [24,55].Chronic melatonin administration led to an increased IL-1 and TNF-␣ production and enhanced T-cell-promoted antigen presentation by splenic macrophages [32].This sensitization of circulating components of the immune system by melatonin was not paralleled in spleen tissue but was reflected within the CNS following melatonin treatment by enhanced positive reactivity to an inflammatory stimulus (Figs I and 2).
The effect of melatonin upon the levels of cerebral cytokine peptides has not yet been investigated but is an obvious and necessary sequel to the current work.Although ROS play a significant role in inflammation and melatonin can clearly scavenge ROS [34], the regional specificity of the findings described here suggests that antioxidant effects did not directly mediate the effects observed.
It is likely but not proven, that restoration of inflammatory responses within the brains of aged animals, to those found in younger animals, is a beneficial event.This is supported by the report that normal age-related increases in circulating TNF-␣ and IL-6 may be exacerbated with neurological disease [13].
An optimal situation may be low intrinsic levels of cytokines accompanied by a rapid ability to respond to a stimulus.However, a response that is excessively prolonged, may be a component of several neurodegenerative disorders associated with aging, notably Alzheimer's disease [7].Therefore, the temporal nature of the effect of melatonin upon induction of inflammatory cytokines is being investigated in isolated cell lines of neural origin.