University of Groningen

Transcription

1 University of Groningen Habituation to a test apparatus during associative learning is sufficient to enhance muscarinic acetylcholine receptor-immunoreactivity in rat suprachiasmatic nucleus van der Zee, Eelke; Biemans, BAM; Gerkema, Menno P.; Daan, S Published in: Journal of Neuroscience Research DOI: /jnr IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2004 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Van der Zee, E. A., Biemans, B. A. M., Gerkema, M. P., & Daan, S. (2004). Habituation to a test apparatus during associative learning is sufficient to enhance muscarinic acetylcholine receptor-immunoreactivity in rat suprachiasmatic nucleus. Journal of Neuroscience Research, 78(4), DOI: /jnr Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Journal of Neuroscience Research 78: (2004) Habituation to a Test Apparatus During Associative Learning Is Sufficient To Enhance Muscarinic Acetylcholine Receptor-Immunoreactivity in Rat Suprachiasmatic Nucleus Eddy A. Van der Zee, 1,2 * Barbara A.M. Biemans, 1 Menno P. Gerkema, 1 and Serge Daan 1 1 Department of Animal Behaviour, University of Groningen, Haren, The Netherlands 2 Department of Molecular Neurobiology, University of Groningen, Haren, The Netherlands The suprachiasmatic nucleus (SCN) is engaged in modulation of memory retention after (fear) conditioning, but it is unknown which pathways and neurotransmitter system(s) play a role in this action. Here we examine immunocytochemically whether muscarinic acetylcholine receptors (machrs), mediating cholinergic signal transduction in the SCN, are involved. For this purpose, machr immunoreactivity (machr-ir) was studied in the SCN after various stages of passive shock avoidance (PSA) and active shock avoidance (ASA) training and, for ASA, at various posttraining time points. machr-ir was significantly enhanced in SCN neurons as a result of the training procedure, and the number of machr-positive glial cells in the SCN increased significantly. The increase in machr-ir as a result of PSA and ASA training was not due to fear conditioning or the number of correct avoidances (in case of ASA training) but rather to behavioral arousal as a consequence of (brief) exposure to a novel environment (the test apparatus). This finding was confirmed by a cage-change experiment in which the rats were allowed to stay in a novel cage for 15 min or 24 hr. Only the brief exposure to the fresh cage triggered alterations for SCN machrs 24 hr later. These results shed new light on a possible function of the cholinergic system in the SCN mediated by machrs in relation to modulation of memory processes and demonstrate that behavioral arousal during (the habituation stage of) a learning task is sufficient to alter the machr system in the SCN Wiley-Liss, Inc. Key words: active shock avoidance; cholinergic signal transduction; glial cells; novelty; passive shock avoidance; cage change The suprachiasmatic nucleus (SCN) of the anterior hypothalamus is the master circadian clock (biological clock) in mammals (Rusak and Zucker, 1979), controlling and synchronizing many biochemical, physiological, and behavioral processes. Studies in the 1970s suggested that it plays a role in learning and memory processes; these studies used passive shock avoidance (PSA) and active shock avoidance (ASA) as associative learning tasks (Holloway and Wansley, 1973a, b; Wansley and Holloway, 1976; Stephan and Kovacevic, 1978). The pathways and neurotransmitter systems through which the SCN is engaged in memory processes are not clear, but the cholinergic system is a possible candidate. The cholinergic basal forebrain system, acting via muscarinic acetylcholine receptors (machrs), is involved in various aspects of PSA and ASA learning (Lo Conte et al., 1982; Dekker et al., 1991; Van der Zee and Luiten, 1999). The major function of the cholinergic system is to evaluate sensory stimuli for their importance (Sarter and Bruno, 1994; Blokland, 1996). Cholinergic neurons are activated by behaviorally salient and arousing stimuli and play a key role in behavioral arousal and attention toward stimuli crucial for (associative) learning (Hasselmo, 1995; Acquas et al., 1996; Wenk, 1997). Bina and coworkers (1993) demonstrated that cholinergic neurons of the medial septum, nucleus basalis, and diagonal band project to the SCN. These cholinergic projections consist of fibers containing large numbers of varicosities and make axosomatic and axodendritic synaptic contact with SCN neurons (Kiss and Halásy, 1996). Acetylcholine (ACh) levels in the SCN do not show an endogenous circadian fluctuation but these levels can be enhanced fivefold within 30 min by arousing an animal Contract grant sponsor: NWO (Netherlands Organization for Scientific Research); Contract grant number: GpD ; Contract grant sponsor: Vernieuwingsimpuls; Contract grant number: *Correspondence to: Eddy A. Van der Zee, Department of Molecular Neurobiology, University of Groningen Biological Center, Kerklaan 30, 9751 NN Haren, The Netherlands. e.a.van.der.zee@biol.rug.nl Received 20 July 2003; Revised 21 July 2004; Accepted 13 August 2004 Published online 5 October 2004 in Wiley InterScience (www. interscience.wiley.com). DOI: /jnr Wiley-Liss, Inc.

3 Novelty and Cholinergic Signaling in SCN 509 (Murakami et al., 1984). Cholinergic agonists can phase shift in a phase-dependent manner the locomotor rhythms in rats and hamsters (Earnest and Turek, 1985; Meijer et al., 1988; Wee and Turek, 1989; Bina and Rusak, 1996) and neuronal firing activity in rat SCN slices (Liu and Gillette, 1996). These actions are predominantly mediated by machrs. Furthermore, cholinergic depletion of the SCN by nucleus basalis lesioning reduces the synthesis and expression of SCN neuropeptides, indicating a role in the regulation of metabolic activity of SCN neurons (Madeira et al., 2004), and alters the responses of the SCN to phase shifting effects of light (Erhardt et al., 2004). Localization of machrs in the SCN has been demonstrated by ligand binding (Kobayashi et al., 1978; Rotter et al., 1979; Bina et al., 1998). In contrast to the case for many other neuroactive substances, the expression of machrs in the SCN (like ACh levels) does not reveal daily fluctuations in rats and hamsters (Van der Zee et al., 1991; Bina et al., 1998). The distribution of machrs in the rat SCN has been described in detail by using the monoclonal antibody M35, which binds to all five machr subtypes (Van der Zee et al., 1991; Carsi- Gabrenas et al., 1997). Using subtype-selective antibodies, we reported that SCN neurons express at least m1, but not m2, machrs, although both subtypes are present on axon terminals within the SCN (Van der Zee et al., 1999). The aim of the current study was to examine 1) whether the SCN is cholinergically engaged by a learning task via its muscarinic receptors, 2) which aspects of the learning task are critical, and 3) whether the neurochemical changes correlate with learning and memory performance. Behaviorally induced alterations in the immunocytochemical expression of machrs in learning and memory-related brain regions (e.g., hippocampus, neocortex, and amygdala) have been described and examined in detail in rats following PSA and ASA learning, but the SCN was not studied (for review see Van der Zee and Luiten, 1999). Here we analyzed whether associative learning changes machr expression in the SCN. For this purpose, we examined machr-ir in the SCN after various stages of PSA and ASA training (experiment 1) and for ASA at 10 different posttraining time points (experiment 2) in order to determine the time course of the changes observed in experiment 1. Experiment 3, in which rats were placed in a novel cage, served to determine the impact of a brief (15 min) or prolonged (24 hr) exposure to a novel environment on machr-ir in the SCN. MATERIALS AND METHODS Animals and Assignment to Experimental Groups Male Wistar rats (body weight 300 g; 121 in total) were obtained from the breeding colony at our institute (Department of Animal Physiology, University of Groningen). Animals were group housed (5 7 individuals per cage). Food and water were available ad libitum in a temperature-controlled environment of 21 C 1 C on a 12:12 hr light/dark cycle, with lights on from 8:30 to 20:30 hr. PSA and cage changes were performed between 10:30 and 12:30 [Zeitgeber Time (ZT) 2 and 4, and ASA was performed between 10:30 and 16:30 hr; ZT 2 and 8]. The experiments were approved by the Animal Experimental Committee of the University of Groningen and the animals have been acquired and cared for in accordance with the guidelines published in the NIH Guide for the Care and Use of Laboratory Animals. In experiment 1, 48 rats were used. For PSA, rats were habituated to the test apparatus (habituated group H PSA ;n 6); habituated and shocked twice by a mild foot shock (trained group T PSA ;n 10); or habituated, shocked twice by a mild foot shock, followed by a retention trial (tested group TE PSA ; n 8). Experimentally naïve rats served as controls (naïve group N PSA ;n 6). For ASA, rats were habituated to the shuttle box (habituated group H ASA ;n 6) or habituated and trained by receiving 60 trials in the shuttle box (trained group T ASA ;n 6). Experimentally naïve rats served as controls (naïve group N ASA ;n 6). All experimental rats were killed 24 hr after their last exposure to the test apparatus. In experiment 2, 55 rats were used. Fifty rats were trained for ASA and killed 2, 8, 16, 20, 22, 24, 32, 40, or 48 hr or 13 days (13 24 hr) after training [trained groups T ASA2, -8, -16, -20, -22, -24, -40, -48 (n 5); T ASA32 (n 6); T ASA13days (n 4)]. Five rats served as naïve cage controls. In experiment 3, 18 rats were used. Twelve rats were placed in a novel cage in the beginning of the light period either for 15 min (followed by a return to the home cage for 23.5 hr; n 6) or for 24 hr (n 6). All rats were killed 24 hr following the start of the experiment, and 6 rats served as control rats that did not receive a cage change. For cage changing, the animals were provided with a novel cage, containing fresh sawdust, food, and water. All 18 rats were kept in their home cage for 14 days prior to the experiment. Behavioral Procedures for PSA and ASA A step-through-type passive avoidance apparatus was used to investigate one-trial associative learning. Briefly, the apparatus consisted of a dark compartment ( cm) connected via a small opening to an elevated, well-lit platform. Access from the platform to the dark compartment was regulated by a sliding door in the opening. The floor of the dark compartment was made of stainless steel bars, through which a scrambled foot shock could be delivered. Two habituation trials were given on 2 consecutive days, 24 hr apart. The first habituation trial (day 1) started with a 5-min adaptation to the dark compartment. Immediately afterward, the rat was placed on the illuminated platform and allowed to enter the dark compartment. The sliding door was then closed, and the rat stayed for another 3 min in the dark. During the second habituation trial (day 2), the rat was placed directly on the illuminated platform. After entering the dark compartment the sliding door was closed. The animal was allowed to stay in the dark for 3 min once more. Twenty-four hours later (day 3), the training trial was performed. The preshock entry latency was measured, and the rat received two scrambled electric foot shocks (0.6 ma a.c. for 3 sec) 30 sec apart, shortly after entering the dark compartment with the sliding door closed. The rat was removed from the dark compartment 30 sec after the last foot shock. The H PSA group was

4 510 Van der Zee et al. also placed in the dark compartment for 30 sec but did not receive a foot shock. Twenty-four hours (day 4) later, H PSA and T PSA groups were killed as described below, whereas the TE PSA group was placed on the platform to measure the postshock entry latency to the dark compartment with a cutoff latency of 5 min. Twenty-four hours later (day 5), these animals were killed. As such, H PSA and T PSA animals were killed 72 hr and TE PSA animals 96 hr after their first exposure to the test apparatus, but all experimental subjects were killed 24 hr after their last exposure. Active shock avoidance conditioning, a multitrial associative learning task, was performed in a two-way shuttle box ( cm). All experimental animals were habituated to the shuttle box for 5 min. The sound of a buzzer served as the conditioned stimulus (CS), presented for 5 sec prior to the onset of the unconditioned stimulus (US) of a scrambled electric foot shock (0.3 ma a.c.) delivered through the grid floor of the shuttle box for 3 sec. When the rats crossed the barrier within 5 sec after the onset of the CS presentation, the CS was terminated immediately, no foot shock was applied, and the response was recorded as an avoidance. If the rat did not make an avoidance, it either received the foot shock for 3 sec or escaped from it during the 3-sec delivery by crossing to the other side of the shuttle box. Eight seconds after the onset of the CS, both the CS and the US were terminated when a rat failed to make the response within that time. The animals received 60 trials in one daily session, with an intertrial interval varying between 10 and 30 sec (20 sec on average). Immunocytochemical Procedure Rats were deeply anesthetized with 6% sodium pentobarbital and transcardially perfused with 100 ml heparinized saline, followed by 300 ml fixative composed of 2.5% paraformaldehyde, 0.05% glutaraldehyde, and 0.2% picric acid in 0.1 M phosphate buffer (PB; ph 7.4). The brains were removed from the skull and cryoprotected by overnight storage in 30% sucrose in 0.1 M PB at 4 C. Brains were cut coronally on a cryostat into 25- m sections. (The brain of one animal from the T ASA32 group was discarded because of a failed perfusion.) Thereafter, immunostaining was carried out on the free-floating sections for machrs by using the monoclonal mouse anti-machr IgM M35 raised against purified bovine machr protein as described in detail elsewhere (Van der Zee et al., 1989). M35 does not discriminate between the five machr subtypes (Carsi-Gabrenas et al., 1997; Van der Zee and Luiten, 1999). In short, tissue sections were preincubated for 15 min in 0.1% H 2 O 2 in phosphate-buffered saline (PBS) and immersed in 5% normal rabbit serum (NRS) in PBS for 30 min to reduce aspecific binding in the following incubation step. Next, the sections were incubated with the first antibody (M35), diluted 1:200 in 1% NRS overnight at 4 C under gentle movement of the incubation medium. After the primary incubation, sections were rinsed in PBS and again preincubated with 5% NRS for 30 min before the secondary incubation step in biotinylated rabbit IgG anti-mouse- IgM [mu-chain directed, F(ab ) fraction; Zymed, South San Francisco, CA], diluted 1:200 in PBS for 2 hr at room temperature. Thereafter, the sections were thoroughly rinsed in PBS and incubated in streptavidin-horseradish peroxidase (HRP; Zymed) diluted 1:200 in PBS for 2 hr at room temperature. Finally, after rinsing in PBS and 0.05 M Tris buffer, the sections were processed by the diaminobenzidine (DAB)-H 2 O 2 reaction (30 mg DAB and 0.003% H 2 O 2 /100 ml 0.05 M Tris buffer, ph 7.4), guided by a visual check. Control experiments were performed by the omission of the primary antibody, yielding immunonegative results. Cresyl violet background staining was performed on M35-processed tissue sections of experimentally naïve animals (N PSA and N ASA,T PSA and T ASA of experiment 1) to estimate the number of principal SCN cells. For fluorescent double labeling, sections were labeled for glial fibrillary acidic protein (GFAP; 1:100; Amersham, Arlington Heights, IL) with incubation steps as described above, and with phycoerrythrin-conjugated goat anti-mouse IgG (1:50; Tago, Burlingame, CA) as secondary antibody. After completion of GFAP staining, the sections were immunoprocessed for machrs as described above, with fluorescein isothiocyanate (FITC)-conjugated streptavidin (1:50; Zymed) instead of HRP conjugate. Data Analysis and Quantification The optical density (OD) of M35-ir was determined in the SCN of all animals (experiments 1 and 2). Five levels along the rostrocaudal axis of the SCN were distinguished, with levels 1 and 5 corresponding to bregma levels 0.92 and 1.50 mm, respectively, according to the brain atlas of Paxinos and Watson (1986). Rostrocaudal levels 2, 3, and 4 were then equally distanced from each other between levels 1 and 5. Brain sections containing the SCN were then assigned to these levels. SCN M35-ir OD was determined in brain sections corresponding to the five levels of all animals from experiment 1 (one brain section per level per animal; the obtained values of the left and right SCN were averaged). Based on the outcome of this rostrocaudal analysis, we decided to perform all further OD measures on brain sections corresponding to level 3 (the middle portion of the SCN; two brain sections per animal were analyzed and the values of the left and right SCN were averaged). In addition to the SCN, the OD of M35-ir was also determined of the medial preoptic area (mpoa), the retrochiasmatic area (RChA), and the lateral hypothalamus (LH) of animals from experiment 1. The mpoa was analyzed at bregma level 0.80 mm, and the RChA and LH at bregma level 1.80 mm. The OD was expressed in arbitrary units corresponding to gray levels with a Quantimet 600 image-analysis system. The value of background staining was measured in the corpus callosum, which was relatively devoid of M35-ir. The relative OD was calculated by the formula [(OD area OD background )/ OD background ], thus correcting for variability in background staining between sections. The number of machr-positive SCN neurons and machr-positive astrocytes was determined in both the left and the right SCN of one brain section (rostrocaudal level 3, representing the mid-scn region), and the counts were summed per individual [all experimentally naïve (N ASA and N PSA ;n 12); T PSA (n 10); T ASA (n 6) rats from experiment 1 were analyzed]. Neurons and astrocytes were counted only when the nucleus was present in the section (the longest diameter of the SCN neurons ranged from 7.5 m to 23.0 m). The sample area was delineated by an ocular grid containing a square of m. Total number of SCN neurons in this sample

5 Novelty and Cholinergic Signaling in SCN 511 Fig. 1. Photomicrographs of muscarinic acetylcholine receptor immunoreactivity in the SCN of behaviorally naïve rats (N PSA ; a,d) 24 hr after passive shock avoidance retention test (TE PSA ; b,e) and 24 hr after active shock avoidance training (T ASA ; c,f); d, e, and f are enlargements of the SCN area shown in a, b, and c, respectively. 3V, third ventricle; OC, optic chiasm. Scale bars 48 m in a, b, and c, 12 m ind,e,andf. area was based on cresyl violet background staining as mentioned above. Statistical Evaluation A Wilcoxon signed rank test (WSR), a one-way ANOVA for effect of treatment followed by post hoc pair wise comparison (Dunn s test), or a two-way ANOVA followed by post hoc multiple comparison (Bonferroni) was used for statistical analysis when appropriate. P 0.05 was used as an index of statistical significance in all cases. All tests were applied two-tailed, and all data are presented as mean SEM. RESULTS machr-ir in Experimentally Naïve Rats (Experiment 1) Loosely scattered machr-positive cells with associated dendritic profiles were found within the SCN (Fig. 1a,d) as well as in the mpoa and, to a lesser degree, in the RChA and LH, as described earlier (Van der Zee et al., 1989, 1991). The machr-positive cells belonged to a subpopulation of relatively large, often somewhat elongated, monopolar or bipolar neurons. No apparent differences in distribution or morphological appearance of these cells were found in the dorsoventral axis of the SCN. machr immunostaining, as determined by OD, differed little along the rostrocaudal axis in experimentally naïve rats (see Fig. 3). No differences in any of these measures were obtained between the N PSA and the N ASA groups. Some immunolabeled neurons were present in the optic chiasm. Few machr-ir glial cells were found in the SCN, but those observed often were associated with the large blood vessels or found at the border of the optic chiasm.

6 512 Van der Zee et al. Fig. 2. Performance in behavioral tasks. A: Passive shock avoidance test. Preshock latencies (black bars) to enter the dark compartment are displayed for the three experimental groups (H PSA : habituated rats; T PSA : trained rats; TE PSA : rats receiving training and test). Postshock latency (gray bar) is displayed for the TE PSA group receiving the test trial 24 hr later. The dots at 300 sec represent the two rats that did not enter the dark compartment before the cutoff criterion. **P B: Active shock avoidance test. Number of correct avoidances is plotted per block of 10 trials. The performance of all rats of experiment 2 together (n 50) is displayed as well as the performance after splitting the group into responders and nonresponders (rats making three or fewer correct avoidances in the last block of 10 trials). Behavior and Alterations in machr-ir in Experimentally Tested Rats (Experiment 1) PSA: behavior. All rats displayed exploratory behavior during the habituation period to the test apparatus. PSA preshock entry latency did not differ among the various groups (Fig. 2A; H PSA sec; T PSA sec; TE PSA : sec; one-way ANOVA: P 0.05). The memory retention test of the eight TE PSA animals revealed that two rats did not enter the dark compartment before the 5 min cutoff time, whereas the remaining six animals stayed on the platform for sec on average (Fig. 2A). Postshock entry latency was significantly increased compared with preshock entry latency (P 0.01; WSR test), indicating that the animals associated the dark compartment with the foot shock delivery. PSA: immunostaining. Enhanced machr-ir was found in SCN cell soma and their dendritic profiles in H PSA,T PSA, and TE PSA rats (a TE PSA rat is shown in Fig. 1b,e). H PSA,T PSA, and TE PSA groups revealed significantly enhanced machr-ir compared with the behaviorally naïve group (Fig. 3A; one-way ANOVA effect of experimental treatment: P 0.01). Although the OD seems to increase further in the trained and tested groups (T PSA and TE PSA ) compared with the habituated group (H PSA ), post hoc testing revealed that all groups differed significantly from the naïve group (H PSA P 0.01; T PSA and TE PSA P 0.05) but that no significant differences were present among the H PSA,T PSA, and TE PSA groups. Densely stained cells were homogeneously distributed throughout the dorsoventral axis of the SCN. The number of machr-ir neurons in TE PSA rats increased significantly compared with that in naïve rats (Table I). The staining intensity for machrs was more pronounced at the anterior than at the posterior part of the SCN (twoway ANOVA main effect of SCN level: P 0.05; and of experimental treatment: P 0.001), as revealed by OD measures (analyzed for the TE PSA group; Fig. 4A), but the rostrocaudal gradient was present only in the TE PSA group (interaction effect of SCN level treatment: P 0.05). Post hoc testing revealed that machr staining in TE PSA rats differed significantly from that in naïve rats at all levels. No correlation was found between the postshock entry latency as a measure of memory retention (ranging from 83 to 300 sec) and the OD values for machr-ir in the SCN (Spearman rank correlation: r 0.13; P 0.68). ASA: behavior. All animals exposed to the twoway shuttle box displayed exploratory behavior. During ASA training, a gradual increase in number of correct avoidances occurred over time that varied considerably among individuals. On average, rats reached correct avoidances (69.2%) within 60 trials (data not shown). ASA: immunostaining. An increase in machrir was found in the SCN in all H ASA and T ASA rats, and this was comparable in distribution and appearance to that found for PSA (a T ASA rat is shown in Fig. 1c,f). No differences in ir were obtained with respect to time of training, which varied from ZT2 to ZT8. The number of machr-ir neurons in T ASA rats increased significantly compared with that in naïve rats (Table I). The impact of 5 min of habituation to the shuttle box was examined in the H ASA group and of an additional ASA training session in de T ASA group. Figure 3B shows that the exposure to the test apparatus, and further training, induced a significant increase (one-way ANOVA: P 0.01) in machr-

7 Novelty and Cholinergic Signaling in SCN 513 ir compared with N ASA animals. Post hoc testing, however, revealed that H ASA rats did not differ from T ASA rats (one-way ANOVA: P 0.05). A significant increase in immunostaining in T ASA rats as measured by OD was seen throughout the rostrocaudal axis of the SCN (Fig. 4B; two-way ANOVA effect of experimental treatment: P 0.001; no effect of SCN level) compared with N ASA rats. The distribution along the rostrocaudal axis was similar in both groups (no interaction effect of SCN level experimental group). This increase in machr-ir in T ASA rats was less pronounced and showed less variation along the rostrocaudal axis than the increase seen in PSA-tested animals, although it did not differ significantly from that of TE PSA for all SCN levels. No correlation was found between the number of correct avoidances and the OD values in the SCN (Spearman rank correlation: r 0.41; P 0.50). Alterations in machr-ir in the Medial Preoptic and Retrochiasmatic Area and Other Brain Regions (Experiment 1) In addition to changes in machr-ir in the SCN, other brain regions showed changes as well. Within the hypothalamus in the near vicinity of the SCN, changes were observed in the mpoa, RChA, and LH of TE PSA and T ASA animals compared with N PSA animals. OD measures revealed a significant increase for both learning tasks in the mpoa and RChA (one-way ANOVA effect of experimental procedure: P 0.01 for both regions), but not in the LH (Fig. 3C). The increase in the RChA was less pronounced and more variable among individuals than that of the mpoa (RChA: TE PSA and T ASA vs. N PSA : P 0.01 and P 0.05, respectively; mpoa: TE PSA and T ASA vs. N PSA : P 0.001). Though beyond the scope of the current study, distinct patterns of the changes in machr-ir were present in the forebrain after PSA and ASA performance. Enhanced machr-ir was seen in the SCN in both paradigms, but characteristic and localized changes were also found elsewhere after each task. In PSA-tested rats, changes were present in neocortex (in a columnar fashion; Van der Zee et al., 1994; Van der Zee and Luiten, 1999), piriform cortex, and thalamic regions (e.g., the ventrolateral thalamic nucleus and ventral posteromedial thalamic nucleus), whereas no changes were found in the paraventricular nucleus or (dorsal) hippocampus or various other brain regions. In ASA-tested rats, changes were seen in neocortex (but to a lesser extent than after PSA), amygdala (decrease in the central nucleus Fig. 3. Relative optical density of machr-ir at rostrocaudal level 3 of the SCN (A,B). A: PSA. Habituated, trained, and retention-tested rats are compared with the values for naïve rats. B: ASA. Habituated and trained rats are compared with the values for naïve rats. C: Relative optical density of machr-ir in mpoa, RChA, and LH of naïve and PSA- or ASA-trained rats. Significant increases compared with experimentally naïve rats are indicated by asterisks: *P 0.05, **P 0.01, ***P

8 514 Van der Zee et al. TABLE I. Number of machr-ir Neurons in the Midregion of the SCN (Rostrocaudal Level 3) and the Estimated Percentage of Total SCN Neurons That These Cells Represent machr-positive neurons Percentage of total Number neurons Naïve rats (N ASA N PSA ;n 12) PSA-trained rats (T PSA ;n 10) *** 51 ASA-trained rats (T ASA ;n 6) ** 39 Number of machr-ir cells was significantly enhanced after PSA and ASA training (Experiment 1) compared with naïve rats. **P ***P and increase in corticomedial nucleus; Roozendaal et al., 1997; Van der Zee and Luiten, 1999), paraventricular nucleus, dentate gyrus of the dorsal hippocampus, and thalamic regions (e.g., the ventrolateral thalamic nucleus and ventral posteromedial thalamic nucleus), whereas no changes were found in various other brain regions. The mpoa and RChA in habituated rats (H PSA and H ASA ) revealed changes in machr-ir similar to those seen in animals that were used for associative learning (TE PSA and T ASA ). In contrast, habituated rats revealed no apparent changes in the other brain regions described above, which revealed associative-learning-characteristic alterations in machr-ir (Van der Zee and Luiten, 1999). Time Course of Enhanced machr-ir in the SCN of ASA-Trained Rats (Experiment 2) Behavior. The number of avoidances varied considerably among animals, ranging from 0% correct avoidances (two animals; one of the T ASA20 and one of T ASA32 group) to 90% correct avoidances. A distinction was made between slow and fast learners (respectively referred to as nonresponders and responders) on the basis of the number of avoidances in the last block of 10 trials (Fig. 2B). With all groups of experiment 2 taken together, rats reached correct avoidances (56.8%) on average within 60 trials (Fig. 2B). Immunostaining. Brain sections were studied at rostrocaudal level 3 for machr-ir at 2, 8, 16, 20, 22, 32, 40, and 48 hr after training as well as 13 days (13 24 hr) posttraining (Fig. 5A). OD values of SCN machr-ir in the N ASA and T ASA24 rats in this experiment (experiment 2) were comparable to those of naïve and ASA-trained rats of experiment 1, replicating the previous findings. A significant effect of time point was found (one-way ANOVA: P 0.001). Post hoc tests showed that T ASA2,T ASA8, and T ASA16 groups did not differ from N ASA animals and that T ASA20 was significantly higher (P 0.05) than at all previous time points. Furthermore, T ASA24,T ASA32, and T ASA48 animals differed significantly from all time points before 24 hr (P 0.01 in all cases), but not from each other. Although T ASA13days had an intermediate level of relative OD, this time point did not differ from either Fig. 4. Relative optical density of machr-ir of five equidistant rostrocaudal levels through the SCN of PSA-tested (A) and ASA-trained (B) rats compared with experimentally naïve rats. In case of PSA, a significant increase in machr-ir was found at levels 1 4, representing the most caudal part of the SCN. In case of ASA, a significant increase in machr-ir was found at all levels. Significant increases compared with experimentally naïve rats, as revealed with post hoc testing, are indicated by asterisks: *P 0.05, **P 0.01, ***P naïveort ASA24 animals (P 0.1). It should be noted, however, that, in the four animals of the 13-day group, two had a high relative OD (4.96 and 6.46) resembling T ASA24,T ASA32,T ASA40, and T ASA48, whereas the other two were low (0.92 and 0.95), resembling naïve animals. In addition to OD measures, the numbers of machr-ir astrocytes were counted at rostrocaudal level 3 (Fig. 5B). Immunoreactive astrocytes were clearly discernible from neurons because of higher staining intensity and characteristic morphology (this study; see also Van der Zee

9 Novelty and Cholinergic Signaling in SCN 515 Fig. 6. Fluorescent double labeling of machrs and GFAP in the SCN. Expression of machrs in astrocytes is demonstrated by colocalization of GFAP and machrs (large arrows). Small arrows point to machr-positive neurons. Fig. 5. A: Relative optical density of machr-ir at rostrocaudal level 3 of the SCN of naïve rats and ASA-trained rats killed at different posttraining time points. B: Number of machr-positive astrocytes in the SCN at rostrocaudal level 3 of the same rats as shown in A. Significant increases compared with naïve animals are indicated by asterisks: *P 0.05, **P et al., 1991, 1993). Expression of machrs in SCN astrocytes is demonstrated in Figure 6 by way of fluorescent double labeling. Enhanced numbers of immunopositive astroglial cells (one-way ANOVA main effect of time point: P 0.001) were encountered in all T ASA groups compared with N ASA animals (P 0.05 in all cases), with the exception of T ASA13days. In the latter group, the two animals with high machr-ir OD also revealed high numbers of machr-positive astrocytes (202 and 184), whereas the two individuals resembling naïve animals for OD also had low numbers of astrocytes (43 and 39). The astroglial changes were primarily limited to the SCN and were not obviously present in the mpoa and RChA (data not shown). The variation in number of correct avoidances enabled us to analyze the potential relationship between the OD and the rate of acquisition (expressed as number of correct avoidances). Because alterations in machr-ir were fully expressed at 24 hr posttraining and later time points, whereas, at 20 hr posttraining and earlier time points, no changes had occurred yet, correlations were studied in two clusters of animals. The first cluster consisted of T ASA2, T ASA8, T ASA16, and T ASA20 and the second cluster of T ASA24,T ASA32,T ASA40, and T ASA48.In both cases, no significant correlation was found (cluster 1: r ; P 0.258; cluster 2: r ; P 0.447), indicating that neither the initial nor the enhanced level of machr-ir relates to acquisition. machr-ir and a Novel Cage (Experiment 3) Cage changes were performed during a brief period (15 min) for comparison with the habituation periods in a test apparatus during a learning task or for 24 hr for comparison with classic cage change experiments and nonphotic effects on circadian rhythms (Mrosovsky, 1988). The OD and the number of astrocytes for machr-ir were determined in the mid-scn region of

10 516 Van der Zee et al. TABLE II. Optical Density of machr-ir and Number of machr-positive Astrocytes in the Midregion of the SCN of Rats That Received a Novel Cage for 15 Minutes or 24 Hours and Their Home-Cage Controls Number of machr positive machr OD astrocytes Home-cage controls (n 6) Minutes novel cage (n 6) ** ** 24 Hours novel cage (n 6) Optical density of machr-ir and number of machr-positive cells in the SCN are significantly changed as a consequence of placement in a novel cage for a brief period. **P all animals (Table II). A cage change significantly enhanced both machr-od and the number of machrpositive astrocytes in the 15-min group (P 0.01) but not in the 24-hr group. DISCUSSION The results of the present study demonstrate that the cholinergic system in SCN neurons and astroglial cells is affected as a result of the training procedure for passive and active shock avoidance learning, two forms of associative learning. The essential findings of the present study are as follows. 1) Enhanced machr-ir takes place in SCN neurons and glial cells after associative learning. 2) The increase in machr-ir following PSA or ASA training is not due to fear conditioning per se (i.e., the association of the foot shock with the place of delivery in PSA, or the coupling of a tone and a foot shock in ASA) but rather to exposure to a novel environment (the test apparatus). 3) A brief but not a prolonged exposure to a novel environment (a cage change) does affect SCN machr-ir in a way similar to a brief exposure to a test apparatus. 4) The increase in machr density is not related to performance measures of learning and memory, insofar as no correlations were found in any of the experiments. 5) The increase in SCN machr-ir is differentiated in time for glial cells and neurons; glial cell changes are seen 2 hr after the last training trial and neuronal changes not before the 22 hr posttraining time point. Anatomical Aspects of Enhanced machr-ir Immunocytochemical examination of the distribution of the m1 subtype of machrs has shown that nearly all SCN cells express machrs in experimentally naïve rats (Van der Zee et al., 1999). In contrast, only a small subset of relatively large SCN cells is strongly M35 positive, representing approximately 8% of the SCN cell population. M35 binds to internalized machrs, with no selectivity to machr subtype (Raposo et al., 1987; Carsi- Gabrenas et al., 1997; Van der Zee and Luiten, 1999). This suggests that most SCN cells have few internalized machrs (experimentally naïve home-cage control situation) and that this number increases significantly after (a brief) exposure to a novel environment. After training for PSA or ASA, the proportion of M35-positive cells increased to about 40 50%. This finding implies that the changes in the SCN are massive but presumably not yet maximal; approximately 50 60% of the SCN cells seem to be unaffected by the training procedure. The induction of machr changes does not seem to depend on time of day of training or perfusion within the light period. In the ASA experiments, the time of day of training and perfusion varied between ZT2 and ZT8. Although the physiology of the SCN might be changed within this period, no differences in machr-ir were observed in relation to ZT. This is in line with previous observations that circadian variation in machr immunostaining when using the M35 antibody was not present in the SCN (Van der Zee et al., 1991; unpublished observations). Similarly, Bina and coworkers (1998) showed the absence of daily fluctuations in machrs in Syrian hamster SCN. It remains to be tested, however, whether training in the dark period induces similar changes in the SCN. A strong increase in the number of machr-positive glial cells is found after ASA training (and PSA training alike; data not shown). This change was induced within 2 hr after training, suggesting that it is a direct consequence of the behavioral arousal of the animals in the test apparatus. As with the neuronal changes seen after hr, high numbers of immunostained glial cells remain present up to (at least) posttraining time 48 hr. The decline of this enhanced immunostaining seems to occur concurrently in glial cells and neurons, as seen in two of the four animals studied 13 days after training. Glial cells are supposed to be part of a synchronizing mechanism in the SCN, possibly through intercellular Ca 2 waves (Van den Pol et al., 1992; Welsh and Reppert, 1996), and/or may regulate SCN glutamate content (Lavaille and Serviere, 1995). Our current knowledge of neuron glia interactions in SCN function is limited (Morin et al., 1989; Prosser et al., 1994; Tamada et al., 1998), but the current results potentially point to a role of astroglia in regulating SCN function in relation to novelty and arousal. It is tempting to speculate that these affected glial cells induce the neuronal changes seen almost one circadian cycle later, because these neuronal alterations are initiated intrinsically in the SCN, without further external stimuli. Evidently, the relation between the astroglial and neuronal changes in this context awaits further investigation. Aspects of Fear Conditioning Causing Enhanced machr-ir in the SCN The results of the different aspects of PSA training show that the changes of machrs in the SCN are due not to the conditioning aspect but rather to the initial novelty aspects of the learning task. Habituation of the animals to the PSA test apparatus enhanced machr-ir to the same extent as seen in animals receiving the additional foot shock or this foot shock in combination with a retention trial 24 hr later. Similarly, the results of ASA

11 Novelty and Cholinergic Signaling in SCN 517 show that habituation to the shuttle box triggers the changes. Transferring a rat from its home cage to a fresh cage for 15 min (but not 24 hr) also induced machr alterations. Apparently, brief exposures to a novel environment are sufficient to set in motion the machr changes, whereas prolonged exposure to a novel environment prohibits or counteracts this change. Interestingly, a short-lasting, significant increase in c-fos mrna expression in the SCN is induced within 30 min following the brief exposure to a novel environment (Emmert and Herman, 1999), showing SCN sensitivity for novelty through the expression of immediate early genes. There was no correlation between learning and memory performance and machr-ir in the SCN. One might argue that, although the SCN may be permissive for (time-dependent) memory retention (Stephan and Kovacevic, 1978), this nucleus most likely does not play a key role in acquisition processes or consolidation of memory for conditioning. The SCN may modulate memory retention by either facilitating memory retention at 24-hr intervals following training or inhibiting memory retention at other time points (Stephan and Kovacevic, 1978), possibly by an output signal generated in the machrpositive cells (Biemans, 2003). In contrast to the SCN, however, changes in machr-ir in the amygdala after ASA and PSA training are associated with number of foot shocks, and the temporal dynamics of these alterations (neuronal changes occur within 2 hr in the hippocampus and within 8 hr in the amygdala) differ from those seen in the SCN (Van der Zee et al., 1997; Roozendaal et al., 1997; Van der Zee and Luiten, 1999). This indicates that machrs in brain regions other than the SCN are more closely involved in conditioning aspects and learning and memory processes. Fear conditioning experiments in mice revealed a circadian modulation of context-dependent but not tonecued fear conditioning (Valentinuzzi et al., 2001), also suggesting a link between SCN functioning and environmental cues. Circadian modulation of the expression of fear memory persists for several day, even under constant light conditions, demonstrating the endogenous nature of the rhythms (Chaudhury and Colwell, 2002). These studies clearly indicate modulation of fear memory by the circadian system in mice, although the expression of the underlying rhythm differs from that in rats (Holloway and Wansley, 1973a, b; Wansley and Holloway; 1976). M35 Binding Characteristics and Possible Functional Consequences of Enhanced machr-ir in the SCN Enhanced machr-ir detected by M35 binding has previously been explained by internalization of the machrs and the subsequent exposure of the epitope for the monoclonal antibody M35 (for review see Van der Zee and Luiten, 1999, and references therein). Hence, the monoclonal antibody M35 detects neurons affected by the experimental treatment of the animal, in this study by briefly exposing a rat to either the PSA or the ASA test apparatus. The responsible mechanisms and pathways in the SCN are unknown, but machrs are internalized when phosphorylated. Stimulation of machrs by ACh results in such phosphorylation through the activation of different kinases, pointing to cholinergic signal transduction as the prime candidate. In this context, it is relevant to note that, in old rats with a deteriorated cholinergic system, no machr changes could be induced in the SCN by ASA training (Biemans et al., 2003). However, it should be noted that other neurotransmitters (e.g., glutamate) acting via the same types of kinases can also lead to changes in the phosphorylation state of machrs (Van der Zee and Luiten, 1999). Previously, we described the functional impact of enhanced machr-ir in the hippocampus, neocortex, and amygdala (Van der Zee and Luiten, 1999). Internalized machrs make cholinoceptive cells less sensitive to ACh and, hence, less sensitive for behavioral arousal accompanied by enhanced levels of ACh. Enhanced machr-ir detected by M35 binding also reflects enhanced intrinsic information processing among the affected neurons within a brain region (Hasselmo and Bower, 1993; Hasselmo, 1995; Van der Zee and Luiten, 1999). In case of the SCN, it is of interest to consider the affected SCN cells as a neuronal substrate for time stamping. The time of event initiates, through cholinergic signaling, a circadian output rhythm in the SCN neurons phase locked to the time of event instead of phase locked to the light-dark cycle, employing, for example, one or more clock genes and a peptide as output. The internalized machrs block further cholinergic input and, therefore, protect these cells from additional input by which the initiated rhythm could be disrupted. Indeed, different phase relationships between SCN neurons have been recorded (Herzog et al., 1998; Schaap et al., 2003; Quintero et al., 2003). However, further experiments are required to begin to test such a hypothesis. It is unlikely that the machr changes are related to (nonphotic) phase shifts. PSA training performed under free-running conditions did cause SCN machr changes, but no phase shifts in locomotor activity were observed (Biemans et al., unpublished observations). Cage changes (those of 24 hr) can induce nonphotic phase shifts (Mrosovsky, 1988) but do not lead to SCN machr changes. Functionally, machr changes in the SCN may indicate that familiarization with a novel environment in a short and restricted moment in time is placed in a temporal, circadian context. This mechanism may be important for time place association and incorporation of daily habits in the behavioral repertoire of an animal (Daan, 1981). One may speculate that machr-positive neurons of the SCN link incoming environmental information to time of day. Individual cholinergic neurons of the basal forebrain system innervate both memory-related brain regions (e.g., hippocampus and neocortex) and the SCN (Bina et al., 1993). Interestingly, SCN-lesioned rats are behaviorally aroused for much longer in a novel environment than intact rats (Buijs et al., 1997), suggesting that

12 518 Van der Zee et al. SCN processing of environmental novelty has adaptive values. The present results indicate a new functional aspect of cholinergic signal transduction in the SCN. Additional research, however, is needed to shed more light on the relevance and mechanisms of this finding. REFERENCES Acquas E, Wilson C, Fibiger HC Conditioned and unconditioned stimuli increase frontal cortical and hippocampal acetylcholine release: effects of novelty, habituation, and fear. J Neurosci 16: Biemans BAM A time to remember. Consequences of ageing on the circadian memory modulation in rodents. PhD Thesis, University of Groningen. Biemans BAM, Van der Zee EA, Daan S Age-dependent effects of conditioning on cholinergic and vasopressin systems in the rat suprachiasmatic nucleus. Biol Chem 384: Bina KG, Rusak B Muscarinic receptors mediate carbachol-induced phase shifts of circadian activity rhythms in Syrian hamsters. Brain Res 743: Bina KG, Rusak B, Semba K Localization of cholinergic neurons in the forebrain and brainstem that project to the suprachiasmatic nucleus of the hypothalamus in rat. J Comp Neurol 335: Bina KG, Rusak B, Wilkinson M Daily variation of muscarinic receptors in visual cortex but not suprachiasmatic nucleus of Syrian hamsters. Brain Res 797: Blokland A Acetylcholine: a neurotransmitter for learning and memory? Brain Res Rev 21: Buijs RM, Wortel J, Van Heerikhuize JJ, Kalsbeek A Novel environment induced inhibition of corticosteron secretion: physiological evidence for a suprachiasmatic nucleus mediated neuronal hypothalamoadrenal cortex pathway. Brain Res 758: Carsi-Gabrenas JM, Van der Zee EA, Luiten PGM, Potter LT Non-selectivity of the monoclonal antibody M35 for subtypes of muscarinic acetylcholine receptors. Brain Res Bull 44: Chaudhury D, Colwell CS Circadian modulation of learning and memory in fear-conditioned mice. Behav Brain Res 133: Daan S Adaptive daily strategies in behavior. In: Aschoff J, editor. Handbook of behavioural neurobiology, vol 4: biological rhythms. New York: Plenum. Dekker AJAM, Connor DJ, Thal LJ The role of cholinergic projections from the nucleus basalis in memory. Neurosci Biobehav Rev 15: Earnest DJ, Turek FW Neurochemical basis for the photic control of circadian rhythms and seasonal reproductive cycles: role for acetylcholine. Proc Natl Acad Sci U S A 82: Emmert MH, Herman JP Differential forebrain c-fos mrna induction by ether inhalation and novelty: evidence for distinctive stress pathways. Brain Res 845: Erhardt C, Galani R, Jeltsch H, Cassel J-C, Klosen P, Menet JS, Pévet P, Challet E Modulation of photic resetting in rats by lesions of projections to the suprachiasmatic nuclei expressing p75 neurotrophin receptor. Eur J Neurosci 19: Hasselmo ME Neuromodulation and cortical function: Modeling the physiological basis of behavior. Behav Brain Res 67:1 27. Hasselmo ME, Bower JM Acetylcholine and memory. Trends Neurosci 16: Herzog ED, Takahashi JS, Block GD Clock controls circadian period in isolated suprachiasmatic nucleus neurons. Nat Neurosci 1: Holloway FA, Wansley R. 1973a. Multiphasic retention deficits at periodic intervals after passive-avoidance learning. Science 180: Holloway FA, Wansley R. 1973b. Multiple retention deficits at periodic intervals after active and passive avoidance learning. Behav Biol 9:1 14. Kiss J, Halásy B Synaptic contacts between cholinergic afferents and suprachiasmatic neurones of the rat. Neuroreport 7: Kobayashi RM, Palkovits M, Hruska RE, Rothschild R, Yamamura HI Regional distribution of muscarinic cholinergic receptors in rat brain. Brain Res 154: Lavialle M, Serviere J Developmental study in the circadian clock of the golden hamster: a putative role of astrocytes. Brain Res Dev Brain Res 86: Liu C, Gillette MU Cholinergic regulation of the suprachiasmatic nucleus circadian rhythm via a muscarinic mechanism at night. J Neurosci 16: Lo Conte G, Bartolini L, Casamenti F, Marconcini-Pepeu I, Pepeu G Lesions of cholinergic forebrain nuclei: changes in avoidance behavior and scopolamine actions. Pharmacol Biochem Behav 17: Madeira MD, Pereira PA, Silva SM, Cadete-Leite A, Paula-Barbosa MM Basal forebrain neurons modulate the synthesis and expression of neuropeptides in the rat suprachiasmatic nucleus. Neuroscience 125: Meijer JH, Van der Zee E, Dietz M The effects of intraventricular carbachol injections on the free-running activity rhythm of the hamster. J Biol Rhythms 3: Morin LP, Johnson RF, Moore RY Two brain nuclei controlling circadian rhythms are identified by GFAP immunoreactivity in hamsters and rats. Neurosci Lett 99: Mrosovsky N Phase response curves for social entrainment. J Comp Physiol A 162: Murakami N, Takahashi K, Kawashima K Effect of light on the acetylcholine concentrations of the suprachiasmatic nucleus in the rat. Brain Res 311: Paxinos G, Watson C The rat brain in stereotaxic coordinates, 2nd ed. Centrecourt, Australia: Academic Press. Prosser RA, Edgar DM, Heller HC, Miller JD A possible glial role in the mammalian circadian clock. Brain Res 643: Quintero JE, Kuhlman SJ, McMahon DG The biological clock nucleus: a multiphasic oscillator network regulated by light. J Neurosci 23: Raposo G, Dunia I, Marillo S, André C, Guillet JG, Strosberg AD, Benedetti EL, Hoebeke J Redistribution of muscarinic acetylcholine receptors on human fibroblasts induced by regulatory ligands. Biol Cell 60: Roozendaal B, Van der Zee EA, Hensbroek RA, Maat H, Luiten PGM, Koolhaas JM, Bohus B Muscarinic acetylcholine receptor immunoreactivity in the amygdala II. Fear-induced plasticity. Neuroscience 76: Rotter A, Birdsall NJM, Burgen ASV, Field PM, Hulme EC, Raisman G Muscarinic receptors in the central nervous system of the rat. I. Technique for autoradiographic localization of the binding of the [ 3 H]propylbenzilylcholine mustard and its distribution in the forebrain. Brain Res Rev 1: Rusak B, Zucker I Neural regulation of circadian rhythms. Physiol Rev 59: Sarter MF, Bruno JP Cognitive functions of cortical ACh: lessons from studies on trans-synaptic modulation of activated efflux. Trends Neurosci 17: Schaap J, Albus H, vanderleest HT, Eilers PHC, Détári L, Meijer JH Heterogeneity of rhythmic suprachiasmatic nucleus neurons: implications for circadian waveform and photoperiodic encoding. Proc Natl Acad Sci U S A 100: Stephan FK, Kovacevic NS Multiple retention deficit in passive avoidance in rats is eliminated by suprachiasmatic lesions. Behav Biol 22: