Nirinjini Naidoo, Center for Sleep & Circadian Neurobiology, University of Pennsylvania, 2100 Translational Research Laboratories, 125 S. 31st Street, Philadelphia, PA 19104, USA. Tel.: +215 746-4811; fax: +215 746-4814; e-mail: email@example.com
Fragmentation of wakefulness and sleep are expected outcomes of advanced aging. We hypothesize that wake neurons develop endoplasmic reticulum dyshomeostasis with aging, in parallel with impaired wakefulness. In this series of experiments, we sought to more fully characterize age-related changes in wakefulness and then, in relevant wake neuronal populations, explore functionality and endoplasmic reticulum homeostasis. We report that old mice show greater sleep/wake transitions in the active period with markedly shortened wake periods, shortened latencies to sleep, and less wake time in the subjective day in response to a novel social encounter. Consistent with sleep/wake instability and reduced social encounter wakefulness, orexinergic and noradrenergic wake neurons in aged mice show reduced c-fos response to wakefulness and endoplasmic reticulum dyshomeostasis with increased nuclear translocation of CHOP and GADD34. We have identified an age-related unfolded protein response injury to and dysfunction of wake neurons. It is anticipated that these changes contribute to sleep/wake fragmentation and cognitive impairment in aging.
Age-related declines in sleep depth and sleep quality are expected outcomes in healthy aging, where it is generally assumed that daytime sleepiness is a consequence of poor nighttime sleep quality (Buysse et al., 1992; Huang et al., 2002; Bliwise et al., 2005). However, sleep homeostasis is less robust in older individuals (Niggemyer et al., 2004). Thus, an alternative explanation for impaired sustained wakefulness in the daytime in older individuals is that aging imparts a decline in wake neuron function. Fragmentation of daytime wakefulness correlates with cognitive declines and reductions in perceived quality of life (Carskadon et al., 1982; Dement et al., 1982; Foley et al., 2007). Elucidating mechanisms by which aging contributes to impaired wakefulness is, therefore, of great clinical significance.
Injury to wake-active neurons (WAN) has been observed in neurodegenerative processes and in aging. Specifically, degeneration of the locus coeruleus (LC) noradrenergic neurons is observed in Alzheimer’s disease (AD) and also in Parkinson’s disease (PD) (Ohayon & Partinen, 2002; Lee et al., 2004, 2007; Abbott et al., 2005; Gjerstad et al., 2006; Hawley et al., 2009). A substantial loss of orexinergic neurons, with reduced orexin levels in both cerebral spinal fluid and in the prefrontal cortex has been identified in Parkinson’s disease (Fronczek et al., 2007; Thannickal et al., 2007). Orexinergic neuron injury and loss has also been identified in Huntington’s disease, where huntingtin inclusion bodies have been observed in orexinergic but not adjacent neurons (Aziz et al., 2008). In aging, axonopathy and inclusions have been demonstrated in both noradrenergic and orexinergic wake neurons (Ishida et al., 2000; Zhang et al., 2002b, 2009). While diverse neuronal groups are susceptible to age-related changes and neurodegeneration, there are distinct advantages in studying mechanisms underlying age-related injury to wake neurons: electrographically quantifiable sleep and wake activity, discrete populations of neurons delineated by specific neurotransmitter synthesizing enzymes, and the ability to index function of select wake neurons with c-fos response to wakefulness.
In this series of studies, we characterized wake in young and old mice finding wake instability and impaired wake responses to novelty in older mice. Both orexinergic and noradrenergic wake neurons in young mice evidence an unfolded protein response (UPR), as supported by the presence of phosphorylated PERK. In older mice, this response progresses with translocation of CHOP and increased GADD34. Herein, we identify age-dependent endoplasmic reticulum dyshomeostasis in wake-active neurons. This process is expected to contribute to the vulnerability of these wake neurons in aging and in age-related neurodegenerative processes.
Materials and methods
Young (2 months) and old (12 and 24 months) male C57BL/6J (B6) mice were studied. Mice were maintained in a 12-h light/dark cycle (with lights on at 7 a.m.) and housed in conditions of 23.5 ± 1 °C and 35–45% humidity. Food and water were provided ad libitum. The methods and study protocols were approved in full by the Institutional Animal Care and Use Committee of the University of Pennsylvania and conformed with the revised NIH Office of Laboratory Animal Welfare Policy.
Behavioral state and sleep latency recording and analysis
Surgical implantation of electrodes and electrophysiological recordings followed previously described methods (Veasey et al., 2000, 2004a). Following third postoperative recovery, mice were connected to counter-weighted recording cables in individual cages. Ten days after surgery, sleep recordings were initiated. Ability to move freely within the entire cage and stand on hind limbs to explore the top of the cage was confirmed in all mice studied. Baseline sleep was recorded for 5 days to ensure stable sleep/wake activities across days. Day 5 of stable baseline was used to analyze sleep/wake times. On recording day 6, a baseline murine multiple sleep latency test (MMSLT) was performed (nap opportunities every 30 mins between 2 p.m. and 4 p.m.) to measure wakefulness after sleep as previously described (Veasey et al., 2004b). Briefly, mice were awakened at 1:50 and kept awake until 2 p.m. when they were allowed a 20-min nap opportunity. This was repeated at 2:30, 3, and 3:30 p.m. This MMSLT has been validated to detect increased sleepiness (Veasey et al., 2004b). For each of the four nap opportunities, the sleep latency was scored as the onset to the first 2-min sleep period. These four latencies were averaged for each animal. The following day, enforced wakefulness by gentle handling (soft brush strokes or introduction of murine nest building materials into the cage) was performed for 6 h, followed by a second murine multiple sleep latency test (MSLT) (circadian time held constant across MSLT’s). Primary behavioral state measures were total wake time/24 h, average wake bout lengths in the light and dark period, and wake bout length frequency distribution and sleep latency in the MSLT, measured with two-way anova for n = 10/group.
Novel social encounter wake assay
The locus coeruleus influences the wake response to novel environment (Leger et al., 2009). To assess whether mice maintain wakefulness during a novel social encounter in the subjective day, a second group of mice was implanted, and 10 days following electrode implantation, recordings were obtained. After establishing a 48-h stable baseline, a nonlittermate 8-week-old B6 male nonbreeder mouse was introduced into the cage for the 12 h of the dark period while recordings continued. Primary endpoints were total wake time for the 12 h of the social encounter.
Immunohistochemistry was used to characterize the waking c-fos response in locus coeruleus and orexinergic neurons and to characterize the ER stress response in these same groups of neurons. Mice were perfused, and brains were cryopreserved using previously published techniques with 40-μm sections collected in 1:6 series (Zhu et al., 2007). Orexinergic neurons in the lateral hypothalamus were labeled with monoclonal mouse anti-Orexin-A antibody (Orex-A; MAB763, R&D Systems, Minneapolis, MN, USA) and noradrenergic locus coeruleus with anti-tyrosine hydroxylase (TH) antibody (#22941; Immunostar, Hudson, WI, USA).
To measure the c-fos response to wakefulness, mice were randomized to either undisturbed sleep across a time of heightened sleep (1–4 p.m.) or enforced wakefulness by gentle handling across the same time span. This gentle handling was accomplished by providing the mice with bedding materials and occasionally stroking the mice with a soft paintbrush. For the enforced wakefulness group, wake was electrographically confirmed for the entire 3-h period. At the end of the 3-h conditions, mice were deeply anesthetized and perfused; brains were postfixed, cryopreserved, and sectioned. Neurons were double-labeled with the above neuronal markers for orex-A or TH and c-fos protein (Ab-5; EMD Calbiochem, San Diego, CA, USA). Orex-A and TH were labeled with Alexa fluor 488 (green), and the c-fos secondary antibody was tagged with Alexa fluor 594 (red) (Molecular Probes, Carlsbad, CA, USA). All immunopositive orex-A or TH-positive neurons with visible nuclei in six rostral–caudal sections across the nucleus/mouse were analyzed. Cells were scored as c-fos positive if c-fos labeling in the nucleus was more intense than background (Image J, NIH). Three scorers, blinded to condition, counted numbers and c-fos labeling, and scores were averaged (90–98% agreement). For each mouse, greater than 150 neurons/region were examined. The mean percentage of c-fos-labeled neurons was compared with two-way anova (n = 5 mice, age, and behavioral state condition). A Bonferroni correction was performed for the three wake groups analyzed. C-fos positivity was confirmed examining the same sections with confocal microscopy using 1-μm cuts to confirm nuclear localization (see Fig. 3).
To characterize ER stress in wake-active neurons, adjacent sections of locus coeruleus and lateral hypothalamus were examined in young and old mice for baseline sleep conditions. Having previously identified activation of the PERK sensor pathway in cortical neurons across prolonged wakefulness, this pathway was assessed using rabbit polyclonal p-PERK antibody (sc-32577, Santa Cruz, Santa Cruz, CA, USA). BiP, the major ER chaperone, was measured in the same populations as a second index of a compensated ER stress response using a rabbit polyclonal (SC-13968, Santa Cruz). To examine ER dyshomeostasis, CHOP and GADD 34 were examined, using mouse monoclonal (SC-575, Santa Cruz) and rabbit polyclonal (SC825, Santa Cruz) antibodies, respectively. Specificity of the primary antibodies has been substantiated in our previous studies (Zhu et al., 2007; Naidoo et al., 2008).
Aging induces wakefulness instability
Hourly wake percentiles and total wake time for lights-on and lights-off periods are shown in Fig. 1 and summarized for 24 h, lights on and lights off in Table 1. Total wake time for 24 h did not differ with aging, 721 ± 20 min in the old and 746 ± 34 min in the young mice, t = 0.5, NS. Hourly percentiles of wake did not vary with age (Fig 1A). Despite no change in total wake time, there were significant age effects on both the number of wake bouts and the duration of wake bouts (Fig. 1 and Table 1). Old mice had 25% more wake bouts in the lights-on period (t = 2.8, P < 0.05) and approximately 100% more wake bouts in the lights-off period (t = 5.1, P < 0.001). The duration of wake bouts during the lights-off period was significantly shorter in older mice, t = 4.4, P < 0.001, as summarized in Table 1. The diurnal ratio for sleep time was reduced in old mice: sleep in the lights-on relative to lights-off periods was 1.6 ± 0.1 in old mice and 2.5 ± 0.3 in young mice, t = 2.7, P < 0.05. In summary, age-dependent reductions occurred in wake stability, particularly in the active (lights-off) period. Examples of individual sleep/wake patterns and age-dependent wake instability across the beginning of the dark period are provided in the lower panels of Fig. 1B.
Table 1. Comparison of sleep wake data in young and aged mice over 24 h and during the lights off and lights on periods. Average ± standard deviation shown
Minutes total sleep/24 h
693 ± 31
718 ± 36
Minutes wake/24 h
746 ± 34
721 ± 20
Sleep bout number
158 ± 13
238 ± 15
Wake bout number
158 ± 13
238 ± 14
Avg sleep bout duration (min)
2.3 ± 0.2
1.5 ± 0.1
Avg wake bout duration (min)
3.7 ± 0.7
1.6 ± 0.3
Minutes total sleep
481 ± 11
445 ± 18
238 ± 11
274 ± 18
Sleep bout number
210 ± 13
267 ± 12
Wake bout number
210 ± 13
268 ± 12
Avg sleep bout duration (min)
2.4 ± 0.2
1.7 ± 0.2
Avg wake bout duration (min)
1.2 ± 0.1
1.0 ± 0.1
Minutes total sleep
211 ± 28
273 ± 15
508 ± 28
446 ± 15
Sleep bout number
105 ± 15
208 ± 12
Wake bout number
106 ± 15
209 ± 12
Avg sleep bout duration (min)
2.1 ± 0.2
1.3 ± 0.1
Avg wake bout duration (min)
6.2 ± 1.2
2.2 ± 0.2
Sleep latency response to enforced wakefulness is intact in aged mice
Baseline sleep latency assessed in the latter half of the lights-on period showed a reduced sleep latency in older mice (11 ± 1 min in old mice and 14 ± 1 min in young mice, t = 3.5, P < 0.01 (Fig. 2). In contrast, following 6-h enforced wakefulness, both age groups responded similarly with reductions in sleep latency (to 6.8 ± 2 in old mice, P < 0.05; and to 6.6 ± 1 in young mice, P < 0.001). There were no age-related differences in the sleep latency after sleep deprivation, t = 0.1, NS). In summary, older mice show slightly lower sleep latencies at baseline, but showed intact homeostatic sleep response to enforced wakefulness.
Wake response to novel social encounter is impaired in aged mice
Noradrenergic brainstem neurons are responsive to novel environmental stimuli (Leger et al., 2009), and the locus coeruleus is believed to contribute significantly to the wakefulness response to novel encounters (Delini-Stula et al., 1984; Harro et al., 2000). In an effort to characterize noradrenergic wake neuron function, we examined the wake response to a 12-h dark period novel social encounter. Results are summarized in Fig. 2B. Young mice responded to the novel social encounter with a marked increase in wake time (from 67% ± 6 at baseline to 95 ± 7% with new mouse, t = 7, P < 0.001). In contrast, old mice showed no difference (68% ± 5 vs. 73% ± 6, t = 2, not significant (NS)). Social interactions were observed for the first hour of the encounter and did not vary with age. No aggressive behavior was observed for either group. In summary, young mice showed a marked increase in wakefulness in the active (lights-off period) to a novel social encounter, while older mice did not demonstrate a lasting increased wakefulness in response to a novel social encounter.
C-fos response to wakefulness is diminished in orexinergic and catecholaminergic neurons with aging
In an effort to examine functional responses to wakefulness within the orexinergic and noradrenergic LC neurons, nuclear c-fos immunoreactivity was measured in each cell type across resting and enforced wakefulness. Young vs. old mice which were allowed undisturbed rest showed no differences in the percentage of c-fos immunoreactive orexinergic neurons (Fig 3A,B). Both groups slept 50–60% of the 4 h prior to perfusion. Young mice kept awake for 3 h showed a robust increase in c-fos immunoreactivity in orexinergic neurons (t = 4, P < 0.01), while old mice kept awake did not show an increase in orexinergic c-fos (=1, NS), despite comparable wake time (> 95%) for the 3 h prior to perfusion. As observed with orexinergic neurons, there were no differences in the percentage of c-fos immunoreactive noradrenergic neurons in the locus coeruleus in the resting young and old mice, and in response to 3 h enforced wakefulness, young mice showed a large increase in LC c-fos immunoreactivity (t = 5, P < 0.001). In contrast, no difference was observed between c-fos immunoreactivity in old mice across wake and sleep conditions (t = 2, NS). Comparing wake responses for both nuclei in young vs. old, significantly greater numbers of orexinergic (t = 4, P < 0.01) and noradrenergic (t = 3, P < 0.01) neurons were c-fos positive in younger mice (summarized in Fig. 3).
ER stress is evident in wake neurons in young mice and increases with aging
To ascertain whether ER stress was evident in the less functional aged wake-active neuronal groups, we first examined phosphorylated-protein kinase RNA-like endoplasmic reticulum kinase (p-PERK) immunoreactivity. PERK is a serine/threonine transmembrane kinase that is responsible for repressing protein synthesis. PERK is activated via autophosphorylation upon dissociation from BiP, when BiP is recruited to chaperone misfolded proteins. Activated PERK phosphorylates eukaryotic initiation factor 2α (eIF2α) leading to a stalled 43s ternary complex, inhibition of protein translation, and hence reduced protein load and ER stress. When ER stress persists PERK activation also leads to pro-apoptotic signaling and activation of ATF4 (Activating transcription factor 4), CHOP, and GADD34 (uncompensated ER stress). Young mice had p-PERK present in about 40% of neurons. Older mice had p-PERK immunoreactivity in approximately 80% of orexinergic and noradrenergic neurons, P < 0.05 (Fig. 4). To more fully characterize the ER stress response in young and old mice, we examined BiP immunoreactivity, a compensated ER stress response, and CHOP and GADD34, markers of ER dyshomeostasis when p-PERK is elevated. BiP, measured in the orexinergic neurons, evidenced no change in older mice (present in both groups in 90–95% of neurons, without a difference in intensity). There was little to no CHOP staining in the neurons of the young mice, while aged animals displayed considerable CHOP immunolabeling in orexinergic neurons and in nonorexinergic lateral hypothalamic neurons. Consistent with activation of CHOP, CHOP was localized in the neuronal nuclei (Fig. 5). To provide further support for CHOP activation, we stained additional sections for GADD34 + orexin-A. GADD34, a protein phosphatase 1 (PP1)-interacting protein, transcriptionally activated by CHOP, causes PP1 to dephosphorylate eIF2α, relieving the translational block imposed by eIF2α phosphorylation (Brush et al., 2003; Szegezdi et al., 2006). Removal of the translational block in still stressed cells aggravates ER stress by increasing the protein folding load. Higher levels of GADD34 were found in the neurons of the aged mice compared with that in the young mice (70% ± 4 vs. 14% ± 4, P < 0.01).
We have identified specific wake impairments in aged mice that are consistent with orexinergic and noradrenergic neuronal injury. Although we cannot exclude involvement of other wake neurons, both populations studied evidence three striking findings: compensated ER stress in young mice at baseline, impaired c-fos response to wakefulness, and ER dyshomeostasis in older mice. Collectively, these findings support a high UPR burden in select populations of wake-active neurons that progresses with age, imparting substantial injury to, and dysfunction of these neurons. This progression of ER dyshomeostasis in wake-active neurons with aging is expected to contribute to orexinergic and noradrenergic locus coeruleus injury in neurodegenerative disorders and to cognitive decline in healthy aging.
Contrasting baseline sleep/wake activity in old and young mice, the most striking difference in the present studies was the instability of sleep and wake states, as manifested by short bout lengths and greater numbers of bouts, particularly in the dark period. Specifically, there was a doubling of the transitions into and out of wakefulness in the aged animal’s active period. Behavioral state instability is observed in mice with deficient orexin signaling (Chemelli et al., 1999; Sawai et al., 2010), and there is some evidence from previous reports that there are age-related changes in orexin neurons in both numbers of neurons and projection densities, where the total number of orexinergic neurons (labeled with either orexin-A or orexin-B) is reduced at 24 mos relative to 2 mos in the rat (Sawai et al., 2010). In the macaque monkey, however, orexinergic neuronal counts across the entire lateral hypothalamus reveal no differences with age although orexinergic axons to the locus coeruleus are diminished with age (Downs et al., 2007). Orexinergic terminal densities in the basal forebrain, locus coeruleus, dorsal raphe, and laterodorsal tegmentum and pedunculopontine tegmentum have also been shown to decline significantly in the aged cat (Zhang et al., 2002a, 2005). Consistent with reduced axons, tissue levels of orexin-A and orexin-B are significantly reduced with age (Porkka-Heiskanen et al., 2004). Findings in the present study extend the age effect on orexinergic neurons to include two functional characteristics: heightened sleep/wake instability and impaired c-fos responses to wakefulness in orexinergic neurons. Most importantly, the present work identifies a vulnerability of orexinergic neurons to ER stress in the young that progresses with age. This age-related increase in the UPR is found, however, not only in wake neurons, but in surrounding neurons and glia, suggesting that the age-related ER stress effects are not limited to wake neurons. Nonetheless, the observation that the c-fos response to wakefulness is impaired in these mice strongly supports the concept that wake neurons are impaired in aging. In further support, we observed reduced wakefulness in aged mice during a novel social encounter, where young mice responded with marked increases in wakefulness.
Previous studies have documented age-related injury to noradrenergic locus coeruleus neurons including neurofibrillary tangles and neuronal loss in several species including human and non-human primates (Vijayashankar & Brody, 1979; Brody, 1980; Sladek et al., 1982; Sturrock & Rao, 1985). In the aged mouse, loss of locus coeruleus neurons correlates with cognitive function. In the present study, we demonstrate an age-dependent locus coeruleus wake impairment: the wake response to a novel social encounter, in parallel with the impaired c-fos response and ER dyshomeostasis in the locus coeruleus wake neurons. Thus, aging imparts ER dyshomeostasis in at least two key wake-active neuronal populations, and in each area, the UPR changes occur in parallel with functional consequences consistent with the cell group injured. Whether other wake neuronal populations (histaminergic, dopaminergic, serotonergic, and cholinergic groups) are also injured and rendered dysfunctional will require further study.
The age-dependent lowered resistance to ER stressors identified in the present study could be explained by several mechanisms through which age lowers resistance to ER stressors. Recent studies examining the effect of age on the ER stress response indicate that there is a diminished protective response with reduced BiP the major ER chaperone, impaired calcium handling, and at the same time robust pro-apoptotic signaling (Li & Holbrook, 2004; Paz Gavilan et al., 2006; Hussain & Ramaiah, 2007; Naidoo et al., 2008). Reductions in BiP with aging have been observed in cortical neurons (Naidoo et al., 2008) and in hippocampal neurons (Paz Gavilan et al., 2006). One consequence of low levels of active BiP is a decreased capacity of the ER to handle protein load, both nascent and misfolded proteins. In the present study, we did not observe a reduction in BiP in orexinergic neurons in old mice. In addition, no age-related differences were observed in native gel westerns to assess BiP monomeric (active) and multimeric (inactive forms). However, in aged mice, we observed increased p-PERK, supporting insufficient or inadequate BiP for protein handling. In light of the unchanged BiP levels, despite increased p-PERK, it is possible that the monomeric BiP in aged mice has a posttranslational modification, rendering the chaperone less effective. Alternatively, age-related increases in misfolded proteins are not met with increased synthesis of BiP. The elevated CHOP and GADD34 in orexinergic neurons in aged mice support ER dyshomeostasis. In support, depletion of BiP in Purkinje cells leads to the induction of CHOP and GADD34, feedback suppression of eIF2α phosphorylation, and apoptotic cell death (Wang et al., 2009). Aged animals subjected to ER stress exhibit increased CHOP activation; A role for CHOP in mediating apoptosis in response to ER stress is well established (Wang et al., 1996; Zinszner et al., 1998). It has been reported that elevated CHOP sensitizes cells to oxidative insults (Ikeyama et al., 2003) and that ectopic expression of CHOP in rat fibroblasts leads to increased levels of ROS (McCullough et al., 2001). Targets of CHOP include Bcl-2, GADD34, ERO1a, and TRB3 (Szegezdi et al., 2003). Bcl-2 confers protection against lethal ER stress and apoptosis (Szegezdi et al., 2006) as well as oxidative stress (Lee, 2001). CHOP is known to repress Bcl-2. GADD34, which we have found to increase with age in the present study and with prolonged waking in the cortex in a previous study (Naidoo et al., 2008), promotes resumption of protein synthesis by removing the translational block. We observed a similar progression in motoneurons in response to hypoxia/reoxygenation in that motoneurons with basal ER stress progressed to dyshomeostasis with CHOP activation (Zhu et al., 2008).
In summary, orexinergic and noradrenergic wake neurons show increased ER stress upon aging and evidence ER dyshomeostasis with increased CHOP and GADD34 protein. Wake-active neurons process large amounts of neuropeptides, and because of metabolic demands, these peptides may be exposed to increased oxidative stress, particularly with aging. Alternatively, it is possible that orexinergic and noradrenergic neurons have impaired calcium buffering capacity upon aging. Having identified progressive ER stress in orexinergic and noradrenergic wake neurons with functional impairment, we provide a focus for exploring mechanisms of age-related injury of orexinergic and noradrenergic wake neurons, neurons essential for optimal cognitive performance.
This work was supported in part by NIA AG032500 and HL079588.