|Home | About | Journals | Submit | Contact Us | Français|
Many mammalian cell types show daily rhythms in gene expression driven by a circadian pacemaker. For example, cultured astrocytes display circadian rhythms in Period1 and Period2 expression. It is not known, however, how or which intercellular factors synchronize and sustain rhythmicity in astrocytes. Because astrocytes are highly sensitive to vasoactive intestinal polypeptide (VIP), a neuropeptide released by neurons and important for the coordination of daily cycling, we hypothesized that VIP entrains circadian rhythms in astrocytes. We used astrocyte cultures derived from knock-in mice containing a bioluminescent reporter of PERIOD2 (PER2) protein, to assess the effects of VIP on the rhythmic properties of astrocytes. VIP induced a dose-dependent increase in the peak-to-trough amplitude of the ensemble rhythms of PER2 expression with maximal effects near 100nM VIP and threshold values between 0.1 and 1 nM. VIP also induced dose- and phase-dependent shifts in PER2 rhythms and daily VIP administration entrained bioluminescence rhythms of astrocytes to a predicted phase angle. This is the first demonstration that a neuropeptide can entrain glial cells to a phase predicted by a phase response curve. We conclude that VIP potently entrains astrocytes in vitro and is a candidate for coordinating daily rhythms among glia in the brain.
In mammals the circadian system is composed of a hierarchical arrangement of oscillators including a master pacemaker located in the suprachiasmatic nuclei (SCN). Other oscillators have been reported in the central nervous system such as the olfactory bulb (Granados-Fuentes et al., 2004a; Granados-Fuentes et al., 2004b; Abraham et al., 2005), the retina and other areas of the hypothalamus (Tosini and Menaker, 1996; Abe et al., 2002) as well as in peripheral organs such as liver and lungs (Schibler et al., 2003; Schibler, 2003; Fuller et al., 2008; Lamia et al., 2008). These cells can display intrinsic circadian rhythms as seen for example in cultured fibroblasts (Balsalobre et al., 1998; Balsalobre, 2002), hepatocytes (Lavery et al., 1999), leukocytes (Arjona and Sarkar, 2005; Du et al., 2005), cells in the retina (Tosini and Menaker, 1996; Gerhold and Wise, 2006; Ruan et al., 2006; Ruan et al., 2008), muscle cells (Durgan et al., 2005; Chalmers et al., 2008) and astrocytes (Prolo et al., 2005). In the absence of synchronizing signals, each of these cells run at their own pace leading to desynchrony among cells and damping of the population rhythms (Nagoshi et al., 2004; Welsh et al., 2004; Carr and Whitmore, 2005; Matsumoto et al., 2007). In contrast, SCN neurons synchronize to each other (Aton and Herzog, 2005) and likely entrain peripheral rhythms either directly or indirectly by synaptic or hormonal signals (Silver et al., 1996).
The mechanism for intracellular generation of circadian rhythms is based in part on a transcription-translation feedback loop in which CLOCK and BMAL1 proteins induce the expression of the Period (Per1, Per2 and Per3) and Cryptochrome genes (Cry1 and Cry2) whose protein products interact and inhibit CLOCK/BMAL1 transcriptional activity. Bioluminescent reporters of the transcriptional or translation activity of these clock genes provide a method for monitoring circadian cycling in real time.
Limited data have implicated glia in the functioning of the circadian system both in vivo, (Bennett and Schwartz, 1994; Castel et al., 1997; Moriya et al., 2000) and in vitro (van den Pol et al., 1992; Prosser et al., 1994; Shinohara et al., 1998; Shinohara et al., 2000). In addition, SCN astrocytes have been shown to display circadian rhythms in vivo (Lavialle and Serviere, 1993; Lavialle and Serviere, 1995; Leone et al., 2006; Becquet et al., 2008). Finally, cultured cortical astrocytes display circadian rhythms that can be sustained by co-culturing them with SCN explants, suggesting that neuronal signals synchronize daily rhythms among astrocytes (Prolo et al., 2005). The factors that play a role in neuron-to-astrocyte circadian communication are not known.
VIP is a neuropeptide released by neurons found in the OB, retina, SCN and sparsely distributed throughout the neocortex (Sims et al., 1980; Stopa et al., 1984; Gall et al., 1986; Okamoto et al., 1992). Because VIP is required for coordination of circadian rhythms both in the SCN and in behavior (Aton et al., 2005) and because it is also required for the astrocyte-mediated modulation of the circadian rhythms in the GnRH system (Gerhold and Wise, 2006), we hypothesized that VIP could be a an important factor in neuronal regulation of circadian cycling in astrocytes.
We tested the hypothesis that VIP entrains circadian rhythms in astrocytes. We found that VIP shifts circadian rhythms in PER2 expression in a dose- and phase-dependent manner and that daily VIP administration sustains and entrains circadian cycling in cultured astrocytes. We conclude that VIP is a strong candidate for coordinating daily rhythms in glia in those areas of the brain where it is released.
Homozygous Period2::Luciferase (PER2::LUC) knock-in mice (postnatal days 2 to 6, congenic on C57BL/6J background) expressing PER2 protein fused with firefly (P. pyralis) luciferase (Yoo et al., 2004) were maintained in a 12/12 h light/dark cycle in the Danforth Animal Facility at Washington University. All of the procedures were approved by the Animal Care and Use Committee and conformed to National Institutes of Health guidelines.
Pure astrocyte primary cultures were obtained following methods previously described (Noble and Mayer-Proschel, 1991; Abe et al., 2002; Prolo et al., 2005). Briefly, neonatal PER2::LUC pups were decapitated; the heads were sprayed with 70% ethanol to minimize contamination. Brains were dissected, placed in ice cold Hank's balanced salt solution (HBSS) and transferred to a laminar flow hood where 2mm by 2mm pieces of cortex were excised from the parietal area after careful removal of the meninges. Cortical pieces were then minced and transferred to a solution of 0.05% trypsin, 0.25nM EDTA in phosphate buffer for 15 min at 37°C. After 15 min the supernatant was collected and passed through a 75 μm nylon mesh, washed with HBSS and collected in a 50 ml conical tube containing 3 ml of fetal bovine serum (FBS). This process was repeated once and the collected suspension was then centrifuged at 0.4 gn for 10 min. The pellet was resuspended in growth medium (Dulbecco's Modified Eagle Medium (DMEM) supplemented with 4mM L-glutamine, 10% FBS, 100 U/ml penicillin and 100μg/ml streptomycin). The cells were then transferred to a culture flask containing growth medium and fed twice per week until confluent (after about 14 days). Astrocytes where then passaged as needed up to 4 times and transferred to 35mm dishes (BD Primaria™, BD Biosciences, CA) for the experiments. When medium was exchanged, the flasks were hit against the bench to remove other cell types that do not strongly adhere to the flask surface. We tested purity of cultures by immunolabeling for the astrocytes-specific marker, glial fibrillary acidic protein (GFAP) using published methods (Prolo et al., 2005) and determined that 95% or more of the cells were GFAP positive.
PER2::LUC activity was recorded as described previously (Abe et al., 2002) with minor modifications. Briefly, four days after the cells were plated in 35mm dishes (1×105 cells per dish at 100 cells/mm2), growth medium was exchanged with 2ml of recording medium (DMEM, 10mM HEPES, 2% B27, 10% FBS, 4mM L-glutamine, 4% NaHCO3 and 0.1 mM D-luciferin). The dishes were sealed with high vacuum silicone grease (VWR, Batavia, IL, USA) and placed under photomultiplier tubes (model HC135−11MOD; Hamamatsu, Shizouka, Japan) in a culture incubator at 34°C. In each experiment 8 to 10 cultures were simultaneously recorded. Dishes were sealed at all times except during treatments when they were opened for less than 60 seconds to add 10 μl of recording medium, VIP or gastrin-releasing peptide (GRP) dissolved in recording medium. Bioluminescence counts were collected and summed every 6 minutes using custom made acquisition software.
DMEM, GRP, HBSS, HEPES and penicillin/streptomycin solution were purchased from Sigma (St. Louis, MO, USA). B27 and FBS were purchased from Invitrogen, (Carlsbad, CA, USA). VIP was obtained from Bachem (King of Prussia, PA, USA) and D-Luciferin was purchased from Xenogen corp. (Cranbury, NJ, USA)
Bioluminescence data were processed by “cross-over analysis” as previously described (Abe et al., 2002). Briefly, a 24-h running average was subtracted from the raw data to reduce trends in the baseline that occur over subsequent days. A 3-h running average and a 24-h running average were calculated from the detrended data set. The rising phase crossing of these two smoothed lines provided a phase marker for each cycle. To compare data from different cultures, we subtracted the minimum bioluminescence between 24- and 48-h of recording from the data and then divided by the maximum obtaining an amplitude of 1 for the second cycle of all cultures. We measured the trough-to-peak amplitude of the second cycle for each culture and treatment.
To assign phases to the time of treatment and measure the phase shifts, the peak of the last cycle prior to the treatments was defined as CT 12. We determined peak (acrophase) times from a sine function (period 20 to 28 h) fitted to each day's activity using Clocklab software (version 2.61, Actimetrics, IL, USA). The phase of the rhythm after drug application was determined by fitting the time of the peaks from four consecutive days starting from the 2nd cycle after the treatment. Phase shifts were measured as the difference in time between pre- and post-treatment phases with negative values representing phase delays and positive values representing phase advances. We plotted the phase shifts centered on CT 0 to allow a linear fit to the resulting PRC which we used to predict the phase angle of entrainment.
We tested the hypothesis that a neurotransmitter coordinates daily rhythms in astrocytes. We treated cultured cortical astrocytes with candidate neurotransmitters after their circadian cycling had damped to a low amplitude. Application of 200 nM VIP increased the peak-to-trough amplitude of the ensemble rhythms (one way ANOVA, p<0.005) compared to vehicle and to 200 nM gastrin releasing peptide (GRP), a neuropeptide also expressed in the SCN (Fig. 1). We also found a significant effect of 200 nM VIP on the time of peak bioluminescence of the restored rhythms. The second bioluminescence peak occurred 27.5±3.3h after the treatment with vehicle while VIP-treated cultures peaked at 32.5±0.7 h after treatment (mean ± SD, p<0.05 two sample t-Test assuming different variances). These results suggested that VIP can reset circadian timing in astrocyte cultures.
We tested the response of cultured glia to a range of VIP concentrations. We measured changes in peak-to-trough amplitude of the ensemble bioluminescence rhythms as a function of VIP dose (Fig 2). Threshold VIP doses for detectable increases in PER2 rhythm amplitude ranged between 0.1 and 1 nM with similar results in a total of four independent experiments. The maximal response, reached near 100 nM was nearly five times the amplitude of untreated cultures. We conclude that VIP specifically increased the amplitude of the ensemble rhythms and phase shifted circadian oscillations in astrocyte cultures compared to vehicle and GRP.
To further characterize VIP effects on the phase of circadian rhythms in astrocytes, we measured steady state shifts in the time of the daily peak of PER2::LUC expression as a function of VIP dose. In three independent experiments cultures were treated either with vehicle (n = 6 cultures) or increasing doses of VIP ranging from 0.1 nM to 1 μM (n = 6 cultures per dose) at CT 10 during their third cycle of PER2::LUC bioluminescence after a full medium exchange. We found a dose-dependent effect of VIP on the time of peak bioluminescence (one way ANOVA, p<0.05, Fig. 3A and B). Fitting the data with a sigmoidal dose-response curve revealed a threshold near 0.1 nM, an EC50 of 4.5 ±1.8 nM and saturating responses at 100−250 nM VIP. Thus, the dynamic ranges for changes in amplitude and phase were similar (0.1 to 100 nM VIP).
We next examined the dependence of VIP-induced changes on the circadian time of administration. We generated a phase response curve (PRC) and found that 200 nM VIP treatments between CT 8 and 13 delayed while treatments between CT13 and 20 advanced subsequent PER2::LUC rhythms (Fig. 3C). The large shifts were consistent with a strong resetting (type 0) PRC and differed from the negligible and phase-independent effects of vehicle (−0.3 ± 0.6 h, mean ± SD, n = 11).
We used the PRC to predict if and how the astrocyte circadian oscillator entrains to daily VIP stimulation. We assumed the oscillator will entrain if and when a daily VIP stimulus falls at a time to compensate for the difference between the oscillator's intrinsic period and the 24-h period of a daily stimulus. For example, untreated astrocytes expressed a period of 23.6 ± 0.4 h (mean ± SD, n = 12). We therefore predicted they would entrain when they delay by 0.4 h per day (Fig. 4). The PRC in Figure 3 predicts that will happen when VIP is given near CT 4. Relative to the peak of PER2 bioluminescence (CT 12) the phase angle of VIP entrainment would be 8 hours.
To test if VIP can entrain astrocytes to a predicted phase angle, we treated cultures daily for 6 days with either VIP (200 nM), vehicle or no treatment starting two days after a total medium exchange (Fig. 4). Vehicle-treated cultures did not maintain a constant phase angle relative to the time of treatment, showed a period shorter than 24 h (p<0.05, one sample t-Test) and no significant difference in their period during and after the daily treatment (23.6 ± 0.3 h vs 23.7 ± 0.2 h, p> 0.05). In contrast, after the first treatment, VIP-treated cultures showed a small advance and then, over days 2−6, peaked 7.8 ± 0.7 h after VIP administration. These cultures adopted a near 24-h period (23.9 ± 0.2 h, mean ± SD) which was significantly longer than their period on the days after the last VIP administration (23.3 ± 0.4 h, p< 0.05, two sample paired Student's t-test). Thus, VIP rapidly entrained astrocytes to a stable phase angle consistent with that predicted by the PRC.
Our data provide the first evidence of circadian effects of a neuropeptide on mammalian glia. We found that VIP synchronizes and sustains rhythmicity in astrocytes in vitro. A pulse of VIP augments the amplitude of the ensemble rhythm of PER2 expression and phase shifts these rhythms in a dose- and time-dependent manner. Daily VIP entrains circadian timing in astrocytes. We found that VIP-induced changes in the amplitude and phase of astrocytes cultures have similar dose dependency with dynamic changes in the nanomolar range. These concentrations are consistent with physiological levels in the brain and activity on astrocytes and neurons (Olah et al., 1994; Jozwiak-Bebenista et al., 2007; Nowak et al., 2007). This was in contrast to significantly smaller effects of GRP, another neuropeptide known to phase shift the SCN (Piggins et al., 1995; McArthur et al., 2000; Antle et al., 2002; Karatsoreos et al., 2006) or vehicle. These findings indicate that VIP is a potent entrainment factor for cultured astrocytes and suggest that it could play a role as a neuron-to-glia coupling signal in vivo.
The results here implicate VIP in coordinating circadian rhythms in cortical astrocytes. Although VIP neurons have been noted in the SCN, OB, retina and scattered throughout the neocortex (Sims et al., 1980; Stopa et al., 1984; Gall et al., 1986; Okamoto et al., 1992), little is known about circulating levels of VIP in the brain. Furthermore, other than the circadian modulation of GFAP in SCN astrocytes (Lavialle and Serviere, 1993), there has been no investigation into circadian rhythms of astrocytes in the brain. For example, cortical slices have been reported to be arrhythmic when Per1::luciferase expression is measured using photomultiplier tubes (Abe et al., 2001). It is not yet known if the astrocytes in these explants were circadian, but undetected, or arrhythmic. The evidence that cortical glia can express entrainable circadian rhythms in vitro (this report and Prolo et al., 2005) should motivate futurein vivo investigations. Future studies should focus on a potential role for VIP from the SCN or other circadian neurons (e.g. the OB or even, perhaps, the neocortex) to entrain cortical glia either through synaptic or humoral communication.
VIP binding and VIP-specific receptors have been described previously in cortical astrocyte cultures (Martin et al., 1992; Ashur-Fabian et al., 1997; Pilzer and Gozes, 2006; Goursaud et al., 2008). In vivo, one study noted high expression of VIP receptors in rat cortex which, however, failed to colocalize with GFAP immunoreactivity (Joo et al., 2004). Further research is needed to address in vivo expression and function of receptors for VIP and other intercellular signals including, for example, neuropeptides with some affinity for VIP receptors, like pituitary adenylate cyclase activating polypeptide or peptide histidine isoleucine.
The large effects of VIP on the timing and amplitude of glial circadian rhythms could reflect that VIP strongly resynchronizes otherwise asynchronous astrocytes or induces clock gene expression in otherwise arrhythmic astrocytes or both. Studies on fibroblast cultures with other stimuli support the first alternative (Izumo et al., 2003; Gachon et al., 2004; Nagoshi et al., 2004; Welsh et al., 2004) although the combination of resynchronization and amplification of rhythmicity has also been suggested (Izumo et al., 2006; Pulivarthy et al., 2007). It is interesting to consider whether VIP affects all astrocytes similarly. For example, light-induced changes in the amplitude of circadian transgenic fibroblasts have been related to differences in the phase-responses of cells within the same culture (Pulivarthy et al., 2007; Ukai et al., 2007). Here, we report VIP-induced increases in PER2 expression of astrocyte cultures that were no longer expressing measurable circadian oscillations. In non-damped cultures (as in Fig. 3), we found that the phase estimation was reliable, but measurement of amplitude changes was not. Single-cell imaging could reveal if VIP affects synchrony among astrocytes, amplitude in individual cells or both.
Finally, astrocytes may be directly involved in the known role of VIP-mediated synchrony among SCN neurons and in circadian behavior (Harmar et al., 2002; Colwell et al., 2004; Aton et al., 2005; Dolatshad et al., 2005; Brown et al., 2005). For example, astrocytes surround VIP-containing neurons in the SCN and release a large variety of intercellular signaling molecules in response to VIP (Brenneman et al., 2003; Furman et al., 2004; Masmoudi-Kouki et al., 2006; Masmoudi-Kouki et al., 2007). That their response to VIP depends on time of day emphasizes the role of astroglia as circadian oscillators within a network of oscillators.
Thanks to Tatsiana Simon for her technical assistance and to Dr. Joseph Takahashi for the generous gift of founder PER2::LUC mice. This work was supported by NSF grant IOB6425445, NIMH grant 63104 and Children's Discovery Institute grant MCII2008-130.