TNFα siRNA Reduces Brain TNF and EEG Delta Wave Activity in Rats

Abstract

Tumor necrosis factor alpha (TNFα) is a pleiotropic cytokine with several CNS physiological and pathophysiological actions including sleep, memory, thermal and appetite regulation. Short interfering RNAs (siRNA) targeting TNFα were incubated with cortical cell cultures and microinjected into the primary somatosensory cortex (SSctx) of rats. The TNFα siRNA treatment specifically reduced TNFα mRNA by 45% in vitro without affecting interleukin-6 or gluR1-4 mRNA levels. In vivo the TNFα siRNAα reduced TNFα mRNA, interleukin-6 mRNA and gluR1 mRNA levels compared to treatment with a scrambled control siRNA. After in vivo microinjection, the density of TNFα-immunoreactive cells in layer V of the SSctx was also reduced. Electroencephalogram (EEG) delta wave power was decreased on days 2 and 3 on the side of the brain that received the TNFα siRNA microinjection relative to the side receiving the control siRNA. These findings support the hypothesis that TNFα siRNA attenuates TNFα mRNA and TNFα protein in the rat cortex and that those reductions reduce cortical EEG delta power. Results also are consistent with the notion that TNFα is involved in CNS physiology including sleep regulation.

INTRODUCTION

Tumor necrosis factor alpha (TNFα) is involved in several brain-regulated physiological processes including sleep, body temperature and appetite (Obal and Krueger 2003). TNFα brain mechanisms are relatively understudied in part due to the difficulty of manipulation of endogenous TNFα levels. TNFα and TNFα mRNA brain levels vary with sleep propensity (Bredow et al., 1997; Floyd and Krueger 1997). Microinjection of TNF into the hypothalamus (Kubota et al., 2001) or locus coeruleus (De Sarro et al., 1997), two areas involved in sleep regulation, promotes non-rapid eye movement sleep (NREMS). TNFα expression in cortical neurons, as measured by immunohistochemistry, is enhanced after increased neural activity (Fix et al., 2006). Further, injection of TNFα onto the surface of the cortex enhances the amplitudes of electroencephalogram (EEG) delta (0.5-4 Hz) waves during NREMS (Yoshida et al., 2004). Such data implicate cortical TNFα in sleep regulation although there is a clear need for new tools to study endogenous TNFα.

RNA-mediated interference (RNAi) is a sequence-specific RNA silencing mechanism found in plants and animals (Sioud et al., 2004). The RNAi pathway is triggered in mammalian cells by the presence of double stranded RNA or small interfering RNA molecules (siRNA) (Fire et al., 1998; Elbashir et al., 2001). An RNA-induced silencing complex (RISC) unwinds the siRNA duplex. The specific degradation is guided by the antisense strand and proceeds with the sense strand siRNA binding to the complementary mRNA site for cleavage by Argonaut-2 (Flynn et al., 2004). The cleavage of the sense strand siRNA and target mRNA results in self-amplifying new siRNA intermediates that continue the mRNA target degradation. siRNAs are useful to probe gene function and structure both in vitro and in vivo and are potential new therapeutic agents (Lewis et al., 2002). siRNA effects are reversible; in vivo they begin about 1 day after administration and last 2-3 days (Dillon et al., 2004; Ryther et al., 2005). Compared with the antisense technology the effects of siRNA appear to be more selective and persistent (Hough et al., 2003; Miyagishi et al., 2003). The experimental and therapeutic use of siRNAs is currently expanding rapidly. In this report we use a TNFα siRNA with the goal of determining if TNF siRNA reduces TNF mRNA and protein expressions in brain and whether those actions manifest in one of TNF’s biological activities, the sleep electroencephalogram (EEG).

We show herein that a TNFα-siRNA reduces in vivo cortical TNFα mRNA and TNF immunoreactivity and in vitro TNFα mRNA in cortical cells. We also show that the TNFα siRNA reduces EEG delta wave power during NREMS. Results are consistent with the hypothesis that TNFα is constitutively expressed in brain and has a role in the sleep EEG.

RESULTS

In vitro RT-PCR analyses

An initial screening of TNFα-siRNAs using cultured cells showed that TNFα-siRNA targeting nucleotides 228-247 and 476-494 reduced TNFα mRNA. Therefore these two siRNAs were pooled for all other in vivo and in vitro experiments. One way ANOVA indicated that TNFα siRNAs induced a significant 45% suppression of TNFα mRNA [F_{(2, 34)} = 16.9, p < 0.00001] compared to either scrambled siRNA-treated or saline controls (Fig. 1). In contrast, the in vitro levels of interleukin 6, GluR1, 2, 3 and 4 mRNAs were not significantly affected by either the scrambled siRNA or TNFα siRNA treatments (Fig.1).

Figure 1.

RT-PCR for TNFα mRNA, IL6 mRNA and GluR1, 2, 3, 4 mRNA levels in the rat fetal cortex primary cell cultures (right, in vitro) after transfection with saline or 20 pmol scrambled siRNA (Con) or 20 pmol TNFα siRNA, see Methods Section for details. Means ± SE were determined from four independent transfections each measured in triplicate. Left; in vivo: one side of the brain received scrambled siRNA (Con) and the other side received TNFα siRNA. Results are expressed as relative change compared with cyclophillin; 7 rats were used. Asterisks indicate statistical significance using the Students’ paired t-test (p < 0.05)

In vivo RT-PCR analyses

Compared to values obtained after scrambled siRNA injections, the TNFα siRNA induced a significant 37% decrease (t = -4.24; p < 0.0001) in TNFα mRNA in the somatosensory cortex (SSctx) (Fig. 1). The TNFα siRNA, unlike the in vitro results, also significantly reduced other transcripts such as IL6 mRNA (56% less) and GluR1 mRNA levels (12% reduction) (Fig. 1) compared to those found in the opposite SSctx (t = -5.59; p < 0.001, t = -2.36; p < 0.05, respectively). The in vivo levels of GluR 2, 3 and 4 mRNA did not significantly change after the TNFα siRNA injection (Fig. 1).

TNFα immunoreactivity after TNFα siRNA treatment

TNFα siRNA treatment of the SSctx reduced the number of darkly stained TNFα-immunoreactive cells in layer V within about 1.0 mm of the injection site (Fig. 2). The TNFα-IR cells illustrated in Figure 2 were about 0.7 mm from the TNFα siRNA injected site (Fig. 2A,C) and they appeared less dense within the cytoplasm than on the scrambled siRNA injected side (Fig. 2B,D). Quantitative evaluation of these TNFα-immunoreactive cells indicated a significant 33% decrease in the density of these cells (density of TNFα siRNA -injected cells = 0.140 ± 0.001; density of scrambled siRNA -injected cells = 0.212 ± 0.019; t = -7.6261; p < 0.002).

Figure 2.

Microinjection of TNFα siRNA into layer V of the somatosensory cortex reduced the immunoreactivity in TNFα-IR pyramidal neurons in layer V (bar below indicates affected region) (A) relative to the opposite side that was injected with a scrambled siRNA control (B). The darkly stained TNFα-IR cells are evident throughout layer V on the side receiving the scrambled siRNA control injection but are reduced in number within 1 mm of the injection of TNFα siRNA (A above the bar relative to B above the bar). The photograph A illustrates the region between 0.7 and 1.3 mm from the injection site. Bar in A and B = 0.3 mm. C and D: Higher power magnification of layer V neurons within the 1 mm region affected by the siRNA injection illustrates the differences in the TNF-IR staining between the siTNF treated (C) and the scrambled siRNA treated side (D) of the brain sections. Bar = 0.025 mm.

Effect of TNF siRNA on sleep and EEG slow wave activity (SWA)

The localized microinjection of the TNFα siRNA into the SSctx did not alter duration of either NREMS or rapid eye movement sleep (REMS) across the 5 day recording period (data not shown). The effect of TNFα siRNA on regional EEG SWA during NREMS is illustrated in Fig 3. The first 23 h after injection (day 1) the rats exhibited their normal diurnal rhythms of EEG SWA during NREMS and there were no significant differences between the two sides of the brain (treatment F_(1,6) = 0.60, p = 0.47; time, F_(5,30) = 1.29, p = 0.29) [treatment × time interaction, F_(5,30) = 1.12, p = 0.37]). In contrast on day 2, there was a significant decrease in EEG SWA during NREMS during the light period on the side that received the TNFα siRNA in comparison with the side that received scrambled siRNA treatment (F_(1,6) = 6.65, p = 0.04; time, F_(5,30) = 0.49, p = 0.78; treatment × time interaction, F_(5,30) = 0.82, p = 0.99). On day 2 on the side that received the TNFα siRNA, the EEG SWA during NREMS was significantly lower (TK test) in the first 4 h after light onset but was not significantly different at later times. Similar results were obtained on day 3 (treatment F_(1,6) = 6.52, p = 0.04; time, F_(5,30) = 1.19, p = 0.38; treatment × time interaction, F_(5,30) = 3.04, p = 0.02). On day 3 the significant decrease in EEG SWA during NREMS was sustained during the first 8 h of the light period (TK test). The EEG SWA returned to control levels by 8 h after light onset on day 3. On day 4, there was also a significant decrease in EEG SWA during NREMS on the side that received the TNFα siRNA compared to the side that received scrambled siRNA (treatment F_(1,6) = 2.47, p = 0.17; time, F_(5,30) = 0.98, p = 0.30; treatment × time interaction, F_(5,30) = 3.08, p = 0.02). On Day 4, as on Day 2, the significant difference occurred during the first 4 h after light onset (TK test). The TNFα siRNA did not significantly affect EEG SWA during NREMS during the dark period. Using time blocks corresponding to the 12 h light and 11 h dark periods there were no significant differences during the dark periods (Fig. 3, right). In contrast, EEG SWA was significantly reduced by TNFα siRNA on days 2 and 3 (t = -2.52; p < 0.05; and t = -2.42; p < 0.05, respectively) during the light period. By day 4 there were no significant differences in EEG SWA during NREMS during the light period (Fig. 3 right). There were no differences in EEG SWA during wakefulness or REMS at any time after TNFα siRNA treatment. Finally, EEG power in higher frequencies (>4 Hz) was not altered by the TNFα siRNA (data not shown).

Figure 3.

Time course of EEG SWA during NREMS expressed as a percentage of baseline EEG SWA determined the day before Day 1. A significant decrease in EEG SWA was observed on Days 2-4. Left graphs: On Day 1, the TNFα siRNA was injected at time 0 on one side of the cortex (open circles) and a scrambled siRNA was injected onto the other side (closed circles). Right graphs: EEG SWA during NREMS calculated for the 12 h light period and the 11 h dark period. The significant decreases only occur in the light phase on Days 2 and 3 after microinjection of the TNFα siRNA (grey bar) relative to the scrambled siRNA (black bar) only.

DISCUSSION

The major findings reported herein are that the TNFα siRNA, when injected directly into the cortex, locally reduced endogenous TNFα mRNA levels and reversibly reduced EEG SWA during NREMS unilaterally. These findings suggest that TNFα is involved in spontaneous EEG SWA and that EEG delta activity is generated in part within the cortex. Similar conclusions were reached previously after the demonstration that injection of TNFα onto the SSctx enhanced EEG SWA regionally whereas injection of the TNF soluble receptor reduced EEG SWA that had been enhanced by sleep loss (Yoshida et al., 2004). EEG SWA is posited to be indicative of sleep intensity (Borbely and Tobler 1989) because it is enhanced during the deep sleep following sleep deprivation (Pappenheimer et al., 1975). Microinjection of TNFα onto the surface of the SSctx also increases the amplitude of surface evoked potentials induced by whisker stimulation (Churchill et al., 2005a) in a fashion similar to the increase in surface-evoked potentials during quiet sleep (Rector et al 2005), suggesting that increases in extracellular TNF can alter the physiological responses to neural stimulation. Such enhanced potentials are characteristic of a cortical column sleep-like state (Rector et al., 2005).

TNFα plays a role in physiological sleep (Obal and Krueger, 2003). Systemic or central injection of TNFα enhances NREMS and EEG delta wave power during NREMS (Obal and Krueger, 2003; Krueger et al., 1998). In contrast, inhibition of TNFα reduces spontaneous sleep and attenuates sleep rebound after sleep loss (Takahashi et al., 1996). In humans, circulating levels of TNFα correlate with sleepiness in health and disease (Dreisbach et al., 1998; Imagawa et al., 2004; Moss et al., 1999; Darko et al., 1995; Obal and Krueger 2003). The soluble TNF receptor reduces fatigue in rheumatoid arthritis patients (Franklin et al., 1999) and in sleep apnea patients reduces sleepiness (Vgontzas et al., 2004). Collectively, such data implicate TNFα in sleep regulation and other brain processes. Current data are consistent with this conclusion because the TNF siRNA inhibited cortical TNFα expression and EEG SWA.

After TNFα siRNA treatment, the unilateral reductions in EEG SWA during NREMS occurred during daylight hours. The reason for this time-of-day effect remains unknown. However, a similar time-of-day-dependent reduction in duration of spontaneous NREMS occurs in mice lacking the TNF 55 kD soluble receptor (Fang et al., 1997). While EEG SWA activity during NREMS is in part regulated independently from duration of NREMS, the two parameters often covary with each other (Fang et al., 1999). For instance, after sleep loss both are enhanced (Pappenheimer et al., 1975). Further, there is now a relatively large literature indicating that localized sleep intensity (EEG SWA) is dependent upon neural activity within the region during prior wakefulness and that activity enhances neuronal expression of TNFα in the activated areas. Thus, for example, stimulation of the right hand enhances EEG delta wave power in the left somatosensory cortex during the subsequent first NREMS period (Kattler et al., 1994). Stimulation of a facial whisker enhances the number of TNFα- and Fos-immunoreactive cells in the activated somatosensory cortex barrel but not in the adjacent unstimulated barrel columns (Fix et al., 2006). In rats spontaneous levels of cortical TNF mRNA and protein are highest at the beginning of the light period (Floyd and Krueger 1997; Bredow et al., 1997). This is likely the end result of greater activity levels during the night time hours. Further, spontaneous EEG SWA also is highest at the beginning of daylight hours when rat NREMS duration is greatest, likely a manifestation of higher TNFα levels. Thus, the current finding that the TNFα siRNA reduced EEG SWA during NREMS during daylight hours is consistent with prior data. Current data also strongly suggest that TNFα plays a role in the local regulation of the sleep EEG and sleep intensity.

There are several methodological issues that limit interpretation of the in vivo EEG SWA results. Thus, because the TNFα siRNA was injected into brain parenchyma the effective dose would decrease the further the distance from the injection site. There could be differential effects of reducing endogenous TNF by different magnitudes. High and low doses of exogenous TNFα have differential effects on neuroprotection (see below). Further, the time courses of TNFα siRNA-induced effects could also be dependent on dose because TNFα induces its own production (Churchill et al., 2006). In this study we did not localize the topographical distribution of the TNFα siRNA-induced changes in EEG SWA beyond comparing the two cerebral hemispheres. Studies are underway to determine if the actions of the TNFα siRNA on EEG delta power are localized only to the small region of TNFα knockdown.

There are several possible mechanisms by which TNFα siRNA could decrease EEG SWA. First, TNFα enhances AMPA receptor expression and activity (Beattie et al., 2002; De et al., 2003; Furukawa and Mattson 1998; reviewed Pickering et al., 2005) and TNFα siRNA injections into the cortex in vivo reduce the AMPA receptor subunit, GluR1, mRNA (Fig. 2). Thus, the TNFα siRNA may decrease AMPA receptor expression and these receptors are posited to be involved in the generation of EEG synchronization (Bazhenov et al., 2002). Second, the reduction of TNFα immunoreactivity was mainly in large pyramidal neurons in layer V. These neurons show a burst firing mediated long-term depression and endocytosis of AMPA receptors, which may provide a mechanism by which synaptic weights may be reconfigured during NREMS (Czarnecki et al., 2006). Cortical pyramidal neurons also provide an excitatory cortical output and influence the cortical-subcortical circuitry important in generating EEG synchronization (Steriade and Amzica, 1998). Finally, although topical application of TNFα enhances cerebral blood flow (Shibata et al., 1996), a reduction in cerebral blood flow due to reduced levels of TNFα induced by TNFα siRNA cannot explain the decreases in EEG slow wave activity, because cerebral blood flow is negatively correlated with EEG SWA (Kjaer et al., 2002; Dang-Vu et al., 2005).

The lack of change in mRNA levels for interleukin-6 and the AMPA receptor subunit mRNAs after TNFα siRNA treatment in vitro suggests that the TNFα siRNA was somewhat specific in silencing TNFα mRNA levels. However, this specificity is likely an artifact of cell culture because TNFα activates nuclear factor kappa B and other transcription factors which in turn could secondarily enhance expression of multiple genes. For instance, TNFα enhances IL6 expression (Ghezzi et al., 2000), and in our hands, the TNFα siRNA reduced IL6 mRNA. Further, the dynamics of the effects of TNFα siRNA on expression of the mRNAs examined or others is unknown; it seems likely that after the primary effect on TNFα mRNA there would be multiple secondary or tertiary effects, each with their own time course. Finally, the specific effects of the TNFα siRNA likely would depend on TNFα levels. Thus, for example, high levels of TNFα are associated with apoptosis and neuronal death while lower levels of TNFα are neuroprotective (De et al., 2005; Pickering et al., 2005).

The motivation for pursuing this line of research is to understand and develop a mechanism to transiently silence TNFα mRNA in vivo and in vitro and show that by doing so a sleep phenotype, EEG SWA, is altered. The current data show the effectiveness of using the TNFα siRNA. Previously, another sleep phenotype, duration of REMS was altered by an siRNA. Thus, Chen et al., (2006) showed using an orexin siRNA a reduction in orexin mRNA and an enhancement in REMS selectively in the dark period after microinjection into the lateral hypothalamus. Those data coupled with the current results, clearly indicate that selected sleep phenotypes can be targeted via use of siRNAs and thereby offer potential new experimental and therapeutic tools. Regardless, current data support the hypothesis that TNFα siRNA will silence TNFα mRNA and thereby reduce the levels of TNFα immunoreactivity in the rat cortex and that TNFα plays an important role in EEG synchronization in localized regions of cortex.

MATERIALS AND METHODS

Somatosensory cortical (SSctx) cell culture and transfection: To validate that the TNFα siRNA affects TNFα mRNA cultured rat cortical cells were used. Primary cultures of rat fetal cortical cells were prepared via a non-enzymatic dissociation as described previously for the hypothalamus with slight modifications (De et al., 2005). For the cortical cultures, 2-4 × 10⁵ cells/well were plated in 6 well plates coated with 100 μg/ml of poly-l-ornithine in 0.15 M borate buffer (pH 8.4). The cells were grown in Dulbelcco’s modification of Eagle’s medium (DMEM) with 10% fetal bovine serum for two days in a humidified atmosphere of CO₂: air (5:95) at 37°C. Then the medium was replaced every other day after plating the cells with serum-free DMEM supplemented with 1 μM human transferrin, 5 μg/ml insulin, 100 μM putrescine, and 60 nM sodium selenite (sf-HDMEM). At 13 days after plating, the medium was replaced with 0.8 ml antibiotic free basic DMEM medium for 3 h before transfection. A real-time reverse transcriptase polymerase chain reaction (RT-PCR) assay was used to screen different TNFα siRNAs. Treatment of the cells was performed in 3-4 different culture preparations from 1-2 mother rats. The treatment conditions were 1) without siRNA control (C), 2) with a Silencer Negative Control #1 called scrambled siRNA (Ambion, Austin,TX) or 3) a pool of two different TNFα siRNA sense: 5′-GACCCUCACACUCAGAUCAtt-3′, antisense: 5′-UGAUCUGAGUGUGAGGGUCtg-3′ (228-247) and sense: 5′-GCCGAUUUGCCACUUCAUAtt-3′, antisense: 5′-UAUGAAGUGGCAAAUCGGCtg-3 (476-494) (Ambion). Both of these TNFα siRNAs were used in subsequent experiments. The cells were transfected using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Briefly, 2 μl of 20 pmol siRNA was mixed with 100 μl of Opti-MEM (Invitrogen). In a second tube, 2 μl of Lipofectamine was mixed with 100 μl of Opti-MEM and incubated at room temperature for 5 min. These two solutions were gently mixed then incubated for 20 min. This lipid-siRNA formulation was taken to a final volume of 1 ml with DMEM. After a 4 h incubation, the cells were replated into the growth medium. Twenty four hours post-transfection, TNFα, interleukin-6 (IL6), cyclophillin and glutamate receptors (gluR) 1-4 mRNA levels were measured by real-time RT- PCR.

Animals and surgery

Male Sprague–Dawley rats weighing (300–350g) were obtained from Taconic Farms, Inc. (Germantown, NY). Rats (n = 7) were housed individually with a 12:12 h light-dark cycle (lights on at 0900 h; about 300 lux) with an ambient temperature of 23 ± 1°C. Food and water were available ad libitum. Surgical implantations of the guide cannulae and EEG electrodes were performed using ketamine and xylazine anesthesia, 87 and 13 mg/kg, respectively, as previously described (Yasuda et al., 2005b). The stainless steel jewelry screw EEG electrodes (Plastics One Inc, Roanoke, VA) were placed bilaterally just under the dura above the somatosensory cortex (Bregma coordinates: 2.5 mm AP, 5.5 mm ML) and bilateral microinjection guide cannulae were placed in layers V-VI as previously described (Yasuda et al., 2005b). Another EEG electrode used as the common reference was implanted 10 mm posterior to bregma on the midline over the cerebellum. Electromyogram (EMG) electrodes were implanted in the dorsal neck muscles. Lead wires were fitted into a miniature plug and attached to the skull with dental cement (Duz-All; Coralite Dental Products, Skokie, IL). Post-surgery recovery was monitored for ten days with rats in their home cages placed inside environmental chambers (Hotpack 352600, Philadelphia, PA). Rats had relatively unrestricted movements during the recovery period.

Recording and analyses of EEG delta activity

A flexible tether (Plastics One) connected the EEG and EMG electrodes to an electronic swivel as described previously (Yasuda et al., 2005a) and rats were given three days to acclimate to this experimental condition. During each of the acclimation days, the rats were handled for 10 min during the 40 min period just prior to light onset. Recordings began at light onset (0900 h). The vigilance states NREMS, rapid eye movement sleep (REMS) and wake were determined off line in 10 sec epochs using criteria published previously (Zhang et al., 1999). On the first day, the EEG baseline was recorded and on the following day microinjections were made on each side of the brain. One side received TNFα siRNA (0.025 nmol of each siRNAs in 2 μl total) while the opposite side received 2 μl of scrambled siRNA (0.050 nmol). Each injection took about 10 min. Lateralization was counterbalanced such that four rats were injected with TNFα siRNA on the right side and three rats on the left side. Each rat was removed from the tether and its environmental chamber and exposed to lights during the microinjection. Experiments were performed in series using 3-4 rats to minimize the time between treatment and light onset (0820 and 0900 h). After injection, EEG and EMG were recorded for 5 days (23 h/day with 1 h/day for cage cleaning and data backup). The average power of EEG slow wave activity (SWA) (1/2 – 4 Hz) during NREMS throughout the 23 h control-recording period was normalized to 100% for each animal and EEG SWA data are expressed as percent of control determined on the baseline day as previously described (Yasuda et al., 2005a).

In vivo mRNA expression

Seven days after completion of the EEG recordings described above, the same rats (n = 7) were again microinjected at the same site and with the same doses of TNFα siRNA or scrambled siRNA as were previously used in the EEG studies. EEG SWA responses to the initial TNFα siRNA were over by day 4 after the initial injection (Fig. 3, right side). This second injection took place one week later. Rats were sacrificed 24 h after the injection and the SSctx was dissected and immediately frozen in liquid nitrogen. The dissection of the SSctx was 2 mm anterior to posterior and 2 mm medial to lateral and 1 mm dorsal to ventral. Whole tissues were stored at -80°C until preparation for RT-PCR analyses.

RT-PCR

Total RNA was extracted, cDNA synthesized and RT-PCR performed as previously described (Taishi et al 2004). Briefly, DNAse I-treated total RNA (2 μg) was reverse transcribed with oligo (dt) 18 primer using superscript III for first-strand cDNA synthesis (Invitrogen) according to the manufacturer’s instructions. Primers were designed using Primer3 (www.genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi) and if possible, an intron was spanned. A standard curve and efficiency analysis of the primers were performed and the primers efficiencies were above 90%. The sequences of primers used are listed in Table 1. The real time PCR reaction was performed in triplicate using the PLATINUM Quantitative PCR SuperMix-UDG (Invitrogen) with 5 μl of diluted cDNA (5 ng, total RNA) in a final reaction volume of 25 μl. The amplification was followed by 40 cycles of denaturation at 94°C for 15 s, annealing at 58°C for 15 s and extension at 72°C for 15 s. Finally, a melting curve was generated by stepwise increases in temperature (0.5°C increase every 10 s) for 80 cycles starting at 55°C. The threshold cycle (C_t) was determined using SYBR Green fluorescence and gene expression was evaluated by means of a comparative C_t method (User Bulletin #2 ABI PRISM 7700 sequence detection system, PE Applied Biosystems). The fold-change between the mRNA expression levels of the experimental and the control samples was calculated as previously described (Churchill et al., 2006). The dissociation curves of each primer pair used showed a single peak and the samples tested after the PCR reactions had a single expected DNA band in an agarose gel analyses.

Table 1.

A list of gene, primer and sequence used to detect mRNA levels of rat

Gene	Gene bank	Primer	Sequence 5’-3’
IL6	NM_012589	Sense	ggagtgctaaggaccaagacca
		Antisense	aggtttgccgagtagacctca
TNFα	X66539	Sense	gacaaggctgccccgactatgtgctc
		Antisense	tgatggcggagaggaggctgactttc
GluR1	NM_031608	Sense	acttcctcgtacacagccaacc
		Antisense	agacaccatcctctccacagtca
GluR2	M85035	Sense	Gtcctggcatgggaatgaat
		Antisense	atcggatgcctctcaccact
GluR3	M85036	Sense	caatgtggcaggcgtgttctat
		Antisense	tgtgagtttcatgcgtttgg
GluR4	NM_017263	Sense	tcagtgaggcaggcgtctta
		Antisense	cgagtccttgggtccacattc
Cyclophillin A	NM_008907	Sense	aaatgctggaccaaacacaaa
		Antisense	ctcatgccttctttcaccttc

Immunohistochemistry and quantitative analyses

Another group of five rats were used for immunohistochemical analyses of TNFα. These rats also received bilateral microinjection cannulae positioned just above the SSctx as described above for the EEG/mRNA experiments. TNFα siRNA and scrambled siRNA were injected into opposite sides of the cerebral cortex just prior to 0900 as described above. Forty eight hours after the microinjections, the rats were sleep deprived for 6 h in order to enhance TNFα expression (Bredow et al., 1997). Twenty four hours later, the rats were anesthetized and cardiac-perfused with isotonic saline, followed by 4% paraformaldehyde. The brains were removed and post-fixed for 6 h, and then sunk in 20% sucrose overnight. The brains were frozen in crushed dry ice and stored at -80°C until sectioning. Brains were sectioned coronally with the sliding microtome at 30 μm thickness and tissue sections were treated for immunohistochemistry using a goat polyclonal anti-rat TNFα (0.5 μg/ml, R&D Systems, Inc Minneapolis, MN) and a biotinylated secondary antibody as previously described (Churchill et al., 2005b). Avidin-biotin-peroxidase reagent (ABC kit at 1:200) and diaminobenzidine tetrahydrochloride (DAB) were from Vector Labs (Burlingame, CA). Negative controls were performed by omitting the primary antibody or pre-incubating the primary antibody (250 ng) with 750 ng recombinant TNFα (R&D Systems) for 24 h prior to the 3-day incubation with the sections. No specific IR-labeling was observed in the sections incubated without primary antibodies and a significant reduction in the labeling was observed when the primary antibody was preabsorbed with the recombinant TNFα. Western analyses of rat brain extracts indicated that the anti-rat TNFα antibody reacted with the 17 kD and 26 kD forms of the TNFα (data not shown).

Photomicrographs of the SSctx adjacent to the injection sites from 2 different sections were prepared using a Leica DMLB microscope with a Spot digital camera. Quantitation of the uncalibrated optical density for each layer V neuron within each figure was determined using Image J (NIH). The counts of 40 neurons/per section were averaged on each side of the brain and two sections in the region of the injection sites were averaged for each rat. The data for each of the rats (n =5) was statistically evaluated using a paired Students t-test. The comparison was evaluated between TNFα siRNA and scrambled siRNA-injected sides of the same rat brains.

Statistical analysis

Data are presented as means ± standard error. Two-way repeated measures analysis of variance (ANOVA) was used to compare 4 h block values of NREMS, REMS, and EEG SWA among the various groups. For the EEG SWA, treatment effect (between scrambled siRNA and TNFα siRNA injection on two sides of each rat) and time effect were the two factors of the two-way ANOVA. When significant differences were detected using the Tukey-Kramer multiple-comparison test (TK test), individual differences were evaluated between TNFα siRNA and scrambled siRNA injected. The time blocks corresponding to the 12 h light and 11 h dark periods for EEG SWA were compared by means of the paired t-test between TNFα siRNA- and scrambled siRNA–injected sides. Since variation in sleep parameters with time of day are well known, only the group or treatment effects are reported herein. The changes in levels of mRNA were also analyzed using of the paired Students’ t-test for in vivo data or a one-way ANOVA for in vitro data. An α-level of P < 0.05 was considered to be significant in all tests.

Abbreviations

TNFα

tumor necrosis factor alpha

SSctx

somatosensory cortex

EEG

electroencephalogram

siRNA

small interfering RNA

NREMS

non rapid eye movement sleep