Exposure to forced swim stress does not alter central production of IL-1

Brain Research 972 (2003) 53–63 www.elsevier.com / locate / brainres

Research report

Exposure to forced swim stress does not alter central production of IL-1 Terrence Deak*, Cherie Bellamy, Leah G. D’Agostino Behavioral Neuroscience Program, Department of Psychology, State University of New York at Binghamton, Vestal Parkway East, Binghamton, NY 13902 -6000, USA Accepted 21 February 2003

Abstract In recent years, there has been increasing recognition that pro-inflammatory cytokines play a role in behavioral and physiological alterations produced by exposure to psychological stressors. Indeed, increases in central IL-1 production have been observed following stressors such as inescapable tailshock and social isolation, while no changes in IL-1 have been observed following other stressors (e.g., exposure to a predator). The goal of the following work was to establish whether exposure to the forced swim test (FST), a commonly used animal model of behavioral despair / depression, leads to an increase in central or peripheral production of IL-1. Briefly, adult male Sprague–Dawley rats (n58 per group) were forced to swim for 15–30 min (25 8C) and killed at various intervals (ranging from immediately to 24 h) following stressor termination. Brains (hippocampus, hypothalamus, posterior cortex) and multiple peripheral tissues (pituitary, adrenals, spleen, plasma) were then dissected and frozen for subsequent measurement of IL-1 using a commercially available enzyme-linked immunosorbent assay. No observable increases in IL-1 were found in rats that were forced to swim acutely, or in rats that were re-exposed to the forced swim stressor 24 h later. These data suggest that exposure to forced swim does not lead to an increase in central production of IL-1, suggesting that the central IL-1 system is unlikely to play a role in mediating behavioral consequences of this stressor. However, these data do not exclude the possibility that other pro-inflammatory cytokines (such as IL-6 and TNF-alpha) might be produced in response to forced swim exposure.  2003 Elsevier Science B.V. All rights reserved. Theme: Neural basis of behavior Topic: Stress Keywords: Interleukin-1; Brain; Forced swim test; Depression; Stress; Rat; Corticosterone

1. Introduction When an organism encounters a pathogen, immune cells are quickly recruited to the site of infection. These immune cells secrete pro-inflammatory cytokines, such as tumor necrosis factor-a (TNF), interleukin-1b (IL-1), and interleukin-6 (IL-6) into the tissue surrounding the infection [5,52]. These pro-inflammatory cytokines perform several key functions for effective host defense: (1) they mobilize and activate other immune cells very rapidly so that the infection can be contained and cleared, (2) they

*Corresponding author. Tel.: 11-607-777-5918; fax: 11-607-7774890. E-mail address: [email protected] (T. Deak).

communicate to the brain that an infection has occurred, and (3) they elicit systemic reactions to suppress pathogen proliferation and enhance immune function. Of particular interest here is that many of the systemic reactions to infection are directly mediated by the central nervous system (CNS) [18,29,45,46]. For instance, changes in behavior that are normally associated with sickness (i.e., sickness behaviors) such as adipsia and aphagia [41], reduced social and sexual interaction [3], and decreased exploration [4] are all coordinated by pro-inflammatory cytokines in the brain. Likewise, the induction of peripheral physiological adjustments (i.e., acute phase responses) such as changes in liver protein synthesis, leukocytosis, changes in plasma ion levels, and fever are also coordinated by brain cytokines (see Ref. [28] for a review). Thus, cytokines play a critical regulatory role in the generation

0006-8993 / 03 / $ – see front matter  2003 Elsevier Science B.V. All rights reserved. doi:10.1016 / S0006-8993(03)02485-5

54

T. Deak et al. / Brain Research 972 (2003) 53–63

and maintenance of sickness responses to infection [29,45,46]. Cytokines not only mediate sickness responses to infection, but they are also involved in behavioral and physiological responses to psychological stressors. Of these cytokines, IL-1 has emerged as one of the most important for mediating stress-related changes in immune function and behavior. For instance, increases in IL-1 have been reported following diverse stressors such as inescapable tailshock [37,38], social isolation [44], immobilization [35], and some forms of restraint [33]. Interestingly, stressinduced increases in IL-1 have been observed in both peripheral tissues (such as blood, pituitary, and spleen) as well as in the CNS (hypothalamus, hippocampus, and cortex) [38]. However, the regional specificity of these effects is highly dependent upon the nature of the stressor employed, especially with regards to IL-1 increases within the CNS [37,44]. Nevertheless, from these studies it is clear that psychological stressors can induce pro-inflammatory cytokine production both centrally and peripherally. Perhaps a more important issue is whether cytokine production in response to stress is functionally related to the behavioral and physiological outcomes of stressor exposure. Indeed, there is strong evidence that this is the case. For instance, blockade of central cytokines by central administration of alpha-melanocyte stimulating hormone (a-MSH; a functional blocker of IL-1 [7]) blocks a wide range of behavioral and physiological adaptations observed following inescapable tailshock, including adipsia, aphagia, changes in acute phase proteins, and fever [34]. Along these same lines, central administration of IL-1 receptor antagonist blocks the enhancement of fear conditioning and the interference with escape responding that is normally observed following inescapable tailshock [31]. Furthermore, IL-1 receptor antagonist also blocks impairments in contextual fear conditioning that are produced by social isolation [44], and reduces the ACTH and hypothalamic monoamine response to immobilization stress [47]. Finally, central injection of IL-1 produces many of the same behavioral, neurochemical, and physiological consequences as stressor exposure [12,28]. In light of these findings, the potential role of cytokines such as IL-1 in mediating the normal behavioral and physiological responses to commonly used animal models of stress is of utmost importance. Given recent evidence that pro-inflammatory cytokines mediate some behavioral and physiological responses of stressor exposure (described above), we sought to evaluate whether exposure to the forced swim test would alter IL-1 expression both in the central nervous system as well as in several stress-responsive peripheral tissues. If forced swim exposure were to induce central IL-1 production, then it would provide a conceptual basis for examining a functional role for IL-1 in mediating behavioral consequences of forced swim exposure. Furthermore, a thorough spatial and temporal examination of potential changes in IL-1

production—using comparable techniques as have been used for the assessment of IL-1 alterations produced by other stressors—will provide an important comparison that will aid in the determination of the stressor characteristics which might be more generally required for the observation of stress-induced IL-1 responses in the CNS. Additionally, the present investigation is especially relevant given the widespread use of the forced swim test as an animal model of behavioral despair / depression [2,8,9,13,42], and recent reports of heightened pro-inflammatory cytokine activity in depressive populations (see Ref. [48] for a review). Thus, we conducted a series of experiments in order to assess whether forced swim exposure would affect IL-1 levels in the central nervous system as well as in several stress-responsive peripheral tissues which are known to produce pro-inflammatory cytokines under various conditions [22,49].

2. Materials and methods

2.1. Subjects Adult male Sprague–Dawley rats (350–450 g) were born and bred at the animal facility in the Department of Psychology using breeders originally obtained from Harlan. Colony conditions were maintained at 2261 8C with a 14:10 light–dark cycle (lights on 06:00–20:00). Animals were group housed in standard Plexiglas bins (2–3 rats per cage) and had ad libidum access to food and water.

2.2. Forced swim stressor The forced swim stressor was administered in a similar manner as described elsewhere [9,42]. Briefly, rats were removed from their home cages and transported to a separate treatment room. They were immediately placed in a cylinder (45 cm high, 20 cm diameter) of water filled 30 cm high that was meticulously maintained at 2561 8C. Rats were forced to swim for 15–30 min, after which they were immediately killed by rapid decapitation (unanesthetized) for tissue harvest or dried briefly with a towel and returned to their home cages (depending on experimental protocol, see below). As is standard in the literature, all rats were exposed to the forced swim stressor in a cylinder that had been freshly cleaned and disinfected prior to the session.

2.3. Systemic injections of LPS Lipopolysaccharide (from Escherichia coli serotype 0111:B4) was purchased from Sigma (St. Louis, MO, USA). LPS was initially diluted in sterile, pyrogen-free saline (0.9%) and aliquots were stored at 220 8C until needed. On the day of experimentation, a frozen aliquot was thawed and diluted to the required concentration(s),

T. Deak et al. / Brain Research 972 (2003) 53–63

55

also in pyrogen-free physiological saline. All injections were administered intraperitoneally (i.p.).

by a shocker (BRS / LVE, Model SGS-004, Beltsville, MD, USA) driven by a personal computer interface.

2.4. Tissue collection and measurement of IL-1

2.6.1. Experiment 1: LPS validation Prior to the examination of whether the forced swim stressor affects IL-1 levels in central or peripheral tissue compartments, we conducted a pilot study to validate our tissue extraction and IL-1 measurement procedure. This was necessary at least in part because we intended to extract and measure IL-1 in sites that we have not previously examined (spleen and adrenal glands), thus expanding the repertoire of tissue compartments which might be involved in stress-induced IL-1 production. For this validation, rats (n55–6 per group) were injected with 0, 10, or 100 mg / kg LPS as described above, and killed by rapid decapitation 2 h later. This time point was chosen because previous research has clearly demonstrated that IL-1 levels are high both centrally and peripherally 2 h after LPS injection [16,17]. Tissue was harvested and dissected as described above and stored at 270 8C until the time of assay.

Brain regions and peripheral tissues were quickly dissected on a cold plate and frozen immediately on dry ice. All tissue was stored at 270 8C until the time of assay. On the day of the assay, each tissue was thawed in 0.25–1.0 ml of ice cold buffer (pH 7.2, 4 8C) containing 50 mM Tris, 1 mM EDTA, 6 mM MgCl 2 100 mM amino-n-caproic acid, 5 mM benzamidine-HCl, 0.2 mM phenylmethylsulfonyl fluoride. After adding buffer to the tissue, samples were sonicated for 10 s using an ultrasonic dismembrator (Fisher, Model 100) at a setting of 5. Sonicated samples were then centrifuged for 10 min at 20,800 rcf (4 8C). Afterwards, supernatants were collected and IL-1 was measured using a commercially available rat IL-1b enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, Minneapolis, MN, USA). Total protein content was measured in each sample using the method of Bradford [6]. All data are expressed as pg of IL-1 / 100 mg total protein, except plasma samples which are expressed as pg IL-1 / ml plasma. Each tissue compartment was analyzed using a separate ELISA procedure, so relative differences across tissue compartments should be made with the utmost caution. This procedure has been previously validated as an effective method for the detection of IL-1 in both central and peripheral tissue compartments [19,37,38,44]. The inter- and intra-assay coefficients of variation were both less than 9% for all IL-1 ELISAs.

2.5. Radioimmunoassay for corticosterone ( CORT) Total plasma CORT levels were measured by radioimmunoassay using rabbit antiserum (antibody B3-163; Endocrine Sciences, Tarzana, CA, USA) as in our previous work [10,11]. This antiserum has very low cross reactivity with other glucocorticoids and their metabolites. The assay sensitivity was 0.5 mg / dl (assay volume520 ml plasma). Both the intra-assay and inter-assay coefficients of variation were less than 7.5%.

2.6. Footshock paradigm Rats were exposed to 80 inescapable footshocks (2.0 mA, 90 s variable ITI, 5 s each) over the course of approximately 2 h. The footshock chambers were standard operant chambers measuring 23.5 cm320.5 cm318 cm (L3W3H; Ralph Gerbrands, Model C, Arlington, MA, USA) that were adapted to deliver shock through the grid floor (18 bars distributed 1 cm on center with a diameter of 1.0 mm each). All current was delivered to the grid floor

2.6.2. Experiment 2: Short time course Since many sickness responses are mediated by production of pro-inflammatory cytokines, and exposure to psychological stressors can also induce pro-inflammatory cytokine production, we sought to determine whether exposure to forced swim stress would also increase levels of IL-1 in both the CNS and periphery. Thus, rats (n58 per group) were forced to swim as previously described for 0 (non-stressed controls), 15, or 30 min. Immediately after the swim session or control treatment, rats were killed by rapid decapitation and tissue was quickly dissected as in previous work [10,38].

2.6.3. Experiment 3: Long time course The findings of experiment 2 suggest that exposure to forced swim does not affect central or peripheral IL-1 levels. However, the data thus far do not negate the possibility that a delayed increase in IL-1 (perhaps several h after stressor termination) might be observed [42]. Furthermore, previous studies reporting stress-induced increases in central IL-1 were much longer in duration (1–3 h [37,44]), which might suggest that stress-induced increases in IL-1 require several h from stressor initiation to develop. Thus, in order to test whether forced swim exposure leads to a delayed increase in IL-1 production, rats (n58 per group) were forced to swim for 15 min and were killed either 2 or 4 h following stressor termination. A separate group of non-stressed controls were used for comparison. For this experiment, the time of the swim session was varied so that all rats would be killed at the same time of day. This was necessary because previous

56

T. Deak et al. / Brain Research 972 (2003) 53–63

studies indicate that central IL-1 levels follow a circadian rhythm [37,50].

2.6.4. Experiment 4: Re-exposure to the forced swim stressor The results of experiments 2 and 3 indicate that a single exposure to the forced swim stressor does not increase IL-1 levels. Nevertheless, it was still possible that reexposure to the forced swim stressor (as is common practice in the literature) would increase central production of IL-1. To test this, a 232 design was employed where control and swim treatment were varied on day 1 versus day 2. This meant that rats (n58 per group) were randomly assigned to one of four treatment groups: swim on day 1 and another swim session on day 2 (swim re-exposure), swim on day 1 with no swim on day 2 (24 h after swim), no swim on day 1 with swim on day 2 (immediately after swim), or no swim on either day (nonstressed controls). This design expands the previously established time courses from experiments 2 and 3, while adding the critical test of whether re-exposure to the swim environment increases IL-1 production. All swim exposures (irrespective of whether they were administered on day 1 or day 2) lasted for 15 min. Rats that had to be returned to their home cages were quickly dried beforehand and killed 24 h later. All other rats were killed immediately after their final swim session or control treatment. 2.6.5. Experiment 5: Footshock and cues associated with footshock Given the clear lack of an increase in IL-1 levels that was observed in the previous three experiments, and other reports of stressors that fail to affect central IL-1 levels (e.g., Ref. [40]), we sought to replicate previous reports of stress-induced IL-1 production [38] using a similar regimen of footshock exposure. To do this, rats were exposed to footshock as described above (referred to as the ‘footshock’ group) using parameters very similar to those previously shown to increase central production of IL-1 following tailshock [37,38]. In addition, a separate group of rats were exposed to the cues associated with the footshock experience of a conspecific by being placed in an identical operant chamber situated 6 cm away for the duration of the shock session. Thus, rats in the ‘cues only’ group were exposed to the stress of being transported to another room and the novelty of the operant chamber, as well as the sight, sound, and smell of another rat being shocked (always it’s cage mate). A third group of nonstressed controls remained in their home cages until the time of sacrifice. This triadic design has been referred to as a ‘stress communication’ paradigm, and has recently been employed to dissociate the psychological and physical components of stressor exposure [14,21,32]. All rats were killed immediately following stressor or control treatment

and tissue was dissected and processed as previously described.

3. Results

3.1. Experiment 1: LPS validation The goal of experiment 1 was to verify a priori that our method of tissue harvest, protein extraction, and IL-1 measurement was sensitive enough to detect treatment differences following LPS injection, an effect which has been well documented [17,37,38,44]. Data from experiment 1 was analyzed using a single factor analysis of variance (ANOVA) design (see Fig. 1). Injection of LPS produced a dose dependent increase in IL-1 levels of the pituitary gland [F(2,14)512.64, P,0.001], spleen [F(2,14)510.26, P,0.01], adrenals [F(2,14)59.601, P, 0.01], and plasma [F(2,14)54.53, P,0.05]. Post hoc analyses using Fisher’s PLSD confirmed significant increases above saline injected controls in both the low dose (10 mg / kg) and high dose (100 mg / kg) LPS groups in these peripheral compartments. Although a similar trend was observed in the central tissue compartments, the increase in IL-1 was only statistically reliable following i.p. administration of 100 mg / kg LPS in the hypothalamus [F(2,14)54.375, P,0.05] and posterior cortex [F(2,14)5 3.92, P,0.05]. The tendency towards an increase in IL-1 levels of the hippocampus did not achieve overall statistical significance [F(2,14)51.831, P.0.05]. Statistically reliable increases in plasma levels of corticosterone were observed following injection of either dose of LPS [F(2,14)514.515, P.0.001].

3.2. Experiment 2: Short time course This experiment was designed to assess whether exposure to an acute bout of forced swim stress would produce an immediate increase in either central or peripheral IL-1 levels. Data were analyzed using a single factor ANOVA, and post hoc tests were performed using Fishers PLSD where an overall significant ANOVA was achieved (see Fig. 2). No observable changes in either central [hypothalamus: F(2,21)50.569, P.0.05; posterior cortex: F(2,21)50.263, P.0.05; hippocampus: F(2,21)50.408, P.0.05] or peripheral [pituitary gland: F(2,21)50.093, P.0.05; spleen: F(2,21)50.744, P.0.05; adrenals: F(2,21)51.229, P.0.05; plasma: F(2,21)51.489, P. 0.05] IL-1 levels were found in rats that were forced to swim for 15 or 30 min. Exposure to the forced swim stressor did, however, significantly increase plasma levels of corticosterone [F(2,21)596.571, P,0.0001], confirming that the swim session was in fact stressful in nature. These data clearly demonstrate that exposure to the forced swim stressor has no immediate effect on central or peripheral production of IL-1.

T. Deak et al. / Brain Research 972 (2003) 53–63

57

Fig. 1. Dose dependent increases in IL-1 levels were observed in all peripheral tissue compartments following either 10 or 100 mg / kg LPS (i.p.). Similar effects were observed in the hypothalamus and posterior cortex, although the effect was only statistically reliable in the high dose group. An asterisk (*) denotes statistically significant differences relative to saline injected controls.

Fig. 2. Rats (n58 per group) were forced to swim for 0, 15 or 30 min and killed immediately thereafter for tissue harvest. No observable increases in either central or peripheral IL-1 levels were observed in any of the tissue compartments examined. An asterisk (*) denotes statistically significant differences relative to non-stressed controls.

3.3. Experiment 3: Long time course

in central [hypothalamus: F(2,21)51.765, P.0.05; posterior cortex: F(2,21)51.564, P.0.05; hippocampus: F(2,21)50.681, P.0.05] or peripheral [pituitary gland: F(2,21)50.339, P.0.05; spleen: F(2,21)50.028, P.0.05; adrenals: F(2,21)52.206, P.0.05; plasma: F(2,21)5 0.370, P.0.05] IL-1 levels were observed 2 or 4 h following stressor termination. By 2 h following the swim

This experiment sought to determine whether acute exposure to 15 min of forced swim would produce a delayed increase in IL-1 levels. Once again, data were analyzed using a single factor ANOVA (see Fig. 3). Similar to the previous experiment, no observable changes

58

T. Deak et al. / Brain Research 972 (2003) 53–63

3.4. Experiment 4: Re-exposure to the forced swim stressor This experiment was designed to assess whether repeated exposure to the forced swim stressor for 15 min on each of 2 days would affect IL-1 levels both centrally and peripherally. The data from experiment 4 were analyzed using a 232 ANOVA design where treatment on day 1 (non-stressed controls versus 15 min swim exposure) was crossed with treatment on day 2 (also non-stressed controls versus 15 min swim exposure). Consistent with the findings of the previous two studies, virtually no effects of forced swim exposure were observed in IL-1 levels (see Fig. 4). Specifically, there were no reliable main effects or interactions in any of the following tissue compartments: hypothalamus: F(1,28)50.397, P.0.05; hippocampus: F(1,28)50.609, P.0.05; pituitary gland: F(1,28)50.02, P.0.05; spleen: F(1,28)50.293, P.0.05; adrenals: F(1,28)50.073, P.0.05; plasma: F(1,28)50.583, P. 0.05. In contrast, a small but reliable reduction in IL-1 levels was observed in the posterior cortex immediately and 24 h after a single exposure to the forced swim stressor [F(1,28)512.552, P,0.01]. As expected, plasma corticosterone levels were significantly elevated immediately after the first swim session as well as after the second exposure to forced swim [F(1,28)5803.072, P,0.0001]. It is perhaps important to note that the magnitude of the corticosterone response to forced swim exposure did not differ significantly based on whether it was the primary exposure to forced swim, or the second exposure to forced swim.

3.5. Experiment 5: Footshock and cues associated with footshock

Fig. 3. Rats (n58 per group) were forced to swim for 15 min, quickly dried with a towel, and returned to their home cages. Rats were then killed for tissue harvest at 2 or 4 h following termination of the forced swim stressor. A separate group of non-stressed controls (‘none’) were used for comparison. No significant changes in IL-1 were observed at any time following the forced swim stressor.

session, plasma corticosterone levels had completely returned to baseline levels [F(2,21)50.530, P.0.05]. These data provide further evidence that IL-1 levels are relatively unaffected by the forced swim stressor.

This experiment was designed to assess whether footshock or the stress cues associated with footshock would increase central IL-1 levels. All data were analyzed using a single factor ANOVA, and post hocs were performed using Fishers PLSD where an overall significant ANOVA was achieved (see Fig. 5). Exposure to footshock led to significantly elevated IL-1 levels in the hypothalamus [F(2,21)56.305, P,0.01] and spleen [F(2,21)59.65, P, 0.01], as well as a significant reduction in IL-1 levels of the posterior cortex [F(2,21)55.046, P,0.05]. IL-1 levels in all other compartments were unaffected by stressor exposure (hippocampus [F(2,21)51.062, P.0.05]; pituitary [F(2,21)50.069, P.0.05]; adrenals [F(2,21)50.144, P.0.05]; plasma [F(2,21)50.234, P.0.05]. As expected, plasma levels of corticosterone were significantly elevated immediately after footshock and exposure to the cues associated with footshock [F(2,21)559.322, P,0.0001]. It is interesting to note that the only reliable effects observed in the ‘cues only’ group was an increase in corticosterone; IL-1 levels were completely unaffected by this treatment.

T. Deak et al. / Brain Research 972 (2003) 53–63

Fig. 4. Rats (n58 per group) were assigned to one of four treatment groups: non-stressed controls (‘no swim’), 15 min acute swim that were killed immediately after stressor termination (‘imm. post’), 15 min swim and killed 24 h after the swim session (‘24 hrs’), or re-exposure to the swim session for an additional 15 min on day 2 (‘swim twice’). A small but reliable suppression of IL-1 levels was observed in the posterior cortex of rats immediately and 24 h following stressor termination, and corticosterone levels were significantly elevated in rats immediately after an acute swim session as well as after re-exposure to the swim stressor 24 h later. An asterisk (*) denotes statistically significant differences relative to non-stressed controls.

59

Fig. 5. Rats (n58 per group) were exposed to 80 inescapable footshocks (‘footshock’), were placed in close proximity to it’s cage mate who was in fact receiving footshock (‘cues only’), or remained in their home cages (‘control’). Immediately after stressor or control treatment, all rats were killed and tissue collected for later processing. IL-1 levels were significantly increased in the hypothalamus and spleen of the footshock group, while being suppressed following the same treatment in the posterior cortex. Corticosterone levels were also significantly increased following footshock as well as in the ‘cues only’ group. An asterisk (*) denotes statistically significant differences relative to non-stressed controls.

60

T. Deak et al. / Brain Research 972 (2003) 53–63

4. Discussion The forced swim test is perhaps one of the most commonly used animal models of behavioral despair, and has been used extensively as a pre-clinical diagnostic tool for the assessment of novel anti-depressants [2,8,9,13,42]. Despite its widespread use in this manner, very little is known about the neural mechanisms underlying the normal progression to immobility that is observed during forced swim exposure. As such, the present set of experiments sought to examine whether forced swim exposure would lead to elevated levels of IL-1 in a regionally selective manner. To our knowledge, the data provided herein are the first to examine whether exposure to the forced swim test affects central and / or peripheral production of IL-1. The data from experiment 1 clearly indicate that our method of tissue harvest, sample processing, and IL-1 detection is sensitive and reliable. This demonstration is obviously a critical prerequisite for any statements to be made about effects that might have been observed in subsequent studies. Robust increases in tissue content of IL-1 were observed in nearly all tissues examined, with the hippocampus being the only exception. Although the peripheral tissue compartments appeared to be much more sensitive to the relative dose of LPS that was administered, trends for such dose responsive effects were clearly evident within the central nervous system as well. It is perhaps important to note that these trends towards increased IL-1 levels in the central nervous system at the low dose of LPS might have achieved statistical significance had more subjects been utilized per group. However, given the preliminary nature of this study and the fact that we were simply looking for validation of our assay sensitivity, it seemed judicious to use as few animals as reasonably possible. Nevertheless, the results of experiment 1 clearly support the sensitivity of our IL-1 detection method. The goal of experiment 2 was to determine whether acute exposure to the forced swim stressor (of varying duration) would affect IL-1 levels in the central nervous system and periphery. In this experiment, no effects whatsoever were observed in any of the tissue compartments examined when tissue was collected immediately after stressor termination. This time point was chosen because the largest increases in central IL-1 production observed following inescapable tailshock were observed immediately following stressor termination, which then declined in a reasonably linear fashion thereafter [38]. Despite the lack of an effect on IL-1 levels immediately following acute exposure to the forced swim stressor, it was still possible that forced swim could produce a delayed and / or protracted induction of IL-1 that would be observed much later. The theory here would be that forced swim exposure sets into motion a sequence of physiological events which would eventually lead to increased production of central IL-1, which might then produce behavioral alterations in the post-stressor period. Thus, we

examined whether central IL-1 levels would be elevated at 2 or 4 h following stressor termination. Clearly, no reliable increases in IL-1 production were observed in any of the compartments examined. In fact, the only reasonable trends in the data were towards a decrease in tissue content of IL-1 in the hypothalamus and adrenal glands, although these were not statistically reliable either. It is quite possible that an increase in central production of IL-1 might have been observed following forced swim exposure had the duration of the stressor been 2–3 h (as with other stressors for which increases in central IL-1 have been reported). However, it is unclear at present whether a previously untrained rat could swim for that long, and our primary interest was whether the standard forced swim stressor (typically no more than 15–30 min) employed in the literature was capable of inducing central IL-1 production. The final forced swim experiment sought to determine whether re-exposure to the forced swim environment 24 h later would affect central IL-1 levels. In order to account for the possibility that the initial forced swim exposure might alter IL-1 levels 24 h later, we employed a 232 design where treatment on day 1 (control versus 15 min swim) was crossed with treatment on day 2 (also control versus 15 min swim). Clearly the most interesting and critical issue here is whether the re-exposure to the forced swim stressor 24 h later (relative to initial swim exposure) would lead to increased production of IL-1. However, the data obtained from experiment 4 do not support such a conclusion. Indeed, the only reliable interaction stemming from IL-1 data was a suppression of IL-1 levels in the posterior cortex immediately and 24 h after swim exposure. A similar trend was also observed in the hypothalamus, but this was not statistically reliable. Thus, it appears as though repeated exposure to the forced swim stressor does not lead to a sensitization in the neural substrates which govern stress-induced increases in IL-1. In order to more generally address the nature of the stressor which may be necessary for the observation of stress-induced increases in IL-1, we exposed rats to 80 footshocks over the course of approximately 2 h. As has previously been shown with inescapable tailshock, exposure to the footshock stressor led to reliable increases in hypothalamic and splenic IL-1. In previous work, these two compartments have proven to be the most sensitive regions for the observation of stress induced increases in IL-1 ([37,38] and T. Deak, unpublished observations). Moreover, the spatial distribution of IL-1 induction in the footshock animals more closely resembles the spatial distribution of IL-1 induction that has been observed following tailshock exposure (increase in hypothalamus and spleen with the two electroshock paradigms [38]) than social isolation (increase only in hippocampus [44]), as one might expect given the nature of the stressors employed. In contrast, no changes in IL-1 levels were observed in rats that were exposed to the stress cues

T. Deak et al. / Brain Research 972 (2003) 53–63

(visual, olfactory, and auditory) in combination with transport / novelty of the situation (‘cues only’ group). Although no systematic behavioral observations were recorded from the ‘cues only’ group, rats in this group showed prominent signs of fear (freezing, startle, defecation, etc.) that were evident throughout the 2 h session, but that dissipated markedly during the latter half. Parenthetically, this might also explain the lower plasma corticosterone levels observed in the cues only group (relative to the footshock group), since the magnitude of the corticosterone response in animals of the ‘cues only’ group is dependent upon the magnitude of behavioral reactivity in the animal to which footshock is actually applied [20]. Nevertheless, these data suggest that the effects of stressors on central IL-1 might depend at least in part upon the physical nature of the stimulus, and represent an important replication of stress-induced increases in IL-1. With that said, it is important to note that this study examined only a single time point, and that there is good evidence to suggest that exposure to stress cues alone as in the present study (‘cues only’ group) might produce a confluence of physiological and behavioral alterations in a much more delayed manner (perhaps 2–3 days later [24]). Taken together, these data suggest that IL-1 production is not a part of the normal repertoire of physiological changes induced by forced swim. Nevertheless, the possibility remains that IL-1 may be involved in mediating behavioral consequences of forced swim exposure, but at levels which were not detectable with our ELISA method. However, the observation of increased IL-1 levels using our ELISA procedure after such a low dose of LPS (10 mg / kg), and following the footshock stressor in experiment 5 would clearly argue against a ‘lack of sensitivity’ explanation. One could also argue that an increase in IL-1 might have been observed in specific CNS sites following forced swim that were not examined herein. However, given such a clear lack of effect of forced swim, we consider that to be an unlikely (albeit still possible) outcome. Another alternative is that although IL-1 levels were relatively unaffected by forced swim exposure, other critical components of the IL-1 signaling pathway may have been affected such as the IL-1 receptor or the IL-1 receptor accessory protein [51]. Indeed, future studies will likely be necessary to examine the possible role for these other factors. The lack of a role for central IL-1 does not preclude a possible role for other cytokines or pro-inflammatory mediators in the generation and maintenance of behavioral responses to forced swim. Indeed, it should be noted that exposure to a novel environment [26], restraint, footshock, or cues that had been associated with footshock [53] all produce increased plasma levels of IL-6. In light of these findings, it would clearly be interesting to examine inflammatory mediators such as IL-6 under the precise conditions and at similar time points following stressor exposure as in the experiments reported herein. However,

61

the measurement of other inflammatory mediators was not possible in these particular tissue samples due to the limited quantity of sample retrieved, difficulties associated with isolation and detection in CNS tissues, and the differential sample processing that may be required for assays of these other mediators. In light of the present data, one might question whether the forced swim stressor—as employed in the present work—was actually stressful in nature. This, in fact, was our precise motive for the measurement of plasma corticosterone levels. As expected, exposure to forced swim produced a very reliable increase in plasma corticosterone that was observed immediately after the forced swim session (experiments 2 and 4), and had completely resolved by 2 h after stressor termination (experiment 3). Importantly, the relative magnitude of the increase was comparable to those previously reported for this stressor [1,25]. It is interesting to note that forced swim exposure did not produce an increase in basal levels of corticosterone 24 h later as has been reported with some other stressors [15,23,39]. Furthermore, the corticosterone response to the first swim session did not differ significantly from that of the second swim session, indicating that repeated exposure to the forced swim stressor does not lead to habituation or sensitization within the hypothalamic pituitary adrenal axis, at least under the present conditions. In comparing forced swim with other stressors that induce central production of IL-1, it is also necessary to take into account the relative role of glucocorticoids as a modulator of stress induced cytokine production. Although initial reports of central IL-1 increases following inescapable tailshock were observed in adrenalectomized rats [37], similar findings have since been reported in adrenal intact rats following multiple stressors, including tailshock [38], social isolation [44], and restraint [33]. Although the magnitude and duration of central IL-1 increases following stressor exposure appears to be much larger in the absence of glucocorticoids, it is clear that central IL-1 responses to stress have in fact been observed with normally functioning glucocorticoid responses to stress. Indeed, a similar effect of glucocorticoids has been observed in noveltyinduced IL-6 responses in plasma [36]. With that said, it is quite possible that forced swim might induce central IL-1 production in adrenalectomized rats. However, this issue is beyond the scope of the present investigation as our goal was to determine whether forced swim would induce central IL-1 production in rats with properly functioning glucocorticoid responses to stress. Finally, it is necessary to discuss the current data within the context of other stressor paradigms during which a role for IL-1 has been evaluated. Both inescapable tailshock and forced swim exposure have been argued as animal models of learned helplessness and / or behavioral despair [30,43], yet clearly divergent effects on central IL-1 production have been obtained using the two stressor models. One might interpret this disparity of findings as

62

T. Deak et al. / Brain Research 972 (2003) 53–63

evidence that the behavioral consequences of forced swim exposure are qualitatively different from those observed following the tailshock paradigm, and that the sequelae of forced swim and tailshock are likely mediated by altogether different neural mechanisms. Along these same lines, it is interesting to note that comparable increases in central IL-1 levels were observed in rats exposed to either inescapable or escapable tailshock, suggesting that the controllability of the stressor is not the critical determinant for the observation of stress-induced increases in IL-1 [27]. There are in fact other precedents in the literature of stressors that do not affect central IL-1 levels, including exposure to a predator [40] and at lest one form of restraint [27]. The present data broaden our current understanding of pro-inflammatory cytokine involvement in mediating the consequences of stressor exposure by (a) excluding forced swim from the class of stressors which increase central IL-1 production, (b) validating the notion that certain stressors (especially those which employ electroshock) can in fact elicit central IL-1 responses, and (c) suggesting that the behavioral consequences of forced swim exposure are not mediated by IL-1. Clearly, the relative role of proinflammatory cytokine involvement in mediating the neurobehavioral consequences of common stressor paradigms must continue to be the subject of thorough and systematic scrutiny in future studies.

Acknowledgements The author wishes to thank T.W. Yep and Kelly A. Bordner for their excellent technical contributions to the following work. This research was supported by NIH grant MH65959.

References [1] E. Abel, Physiological correlates of the forced swim test in rats, Physiol. Behav. 54 (1993) 309–317. [2] A. Armario, M. Gil, J. Marti, O. Pol, J. Balasch, Influence of various acute stressors on he activity of adult male rats in a holeboard and in the forced swim test, Pharmacol. Biochem. Behav. 39 (1991) 373– 377. [3] R. Avitsur, E. Cohen, R. Yirmiya, Effects of interleukin-1 on sexual attractivity in a model of sickness behavior, Physiol. Behav. 63 (1997) 25–30. [4] R. Avitsur, Y. Pollak, R. Yirmiya, Administration of interleukin-1 into the hypothalamic paraventricular nucleus induces febrile and behavioral effects, Neuroimmunomodulation 4 (1997) 258–265. [5] H. Baumann, J. Gauldie, The acute phase response, Immunol. Today 15 (1994) 74–80. [6] M.M. Bradford, A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein– dye binding, Anal. Biochem. 72 (1976) 248–254. [7] A. Catania, J. Lipton, a-Melanocyte-stimulating hormone peptides in host responses. From basic research to human research, Ann. N. Y. Acad. Sci. 680 (1993) 412–423.

[8] T. Connor, P. Kelliher, Y. Shen, A. Harkin, J. Kelly, B. Leonard, Effect of subchronic antidepressant treatments on behavioral, neurochemical, and endocrine changes in the forced-swim test, Pharmacol. Biochem. Behav. 65 (2000) 591–597. [9] J. Cryan, I. Lucki, Antidepressant-like behavioral effects mediated by 5-hydroxytryptamine (2c) receptors, J. Pharmacol. Exp. Ther. 295 (2000) 1120–1126. [10] T. Deak, K.T. Nguyen, C.S. Cotter, M. Fleshner, L.R. Watkins, S.F. Maier, R.L. Spencer, Long term changes in mineralocorticoid and glucocorticoid receptor occupancy following exposure to an acute stressor, Brain Res. 847 (1999) 211–220. [11] T. Deak, K.T. Nguyen, A.L. Ehrlich, L.R. Watkins, S.F. Maier, J. Licinio, M.-L. Wong, G.P. Chrousos, E. Webster, P.W. Gold, The impact of the nonpeptide corticotropin-releasing hormone antagonist antalarmin on behavioral and endocrine responses to stress, Endocrinology 140 (1999) 79–86. [12] A.J. Dunn, Interactions between the nervous system and the immune system: Implications for psychopharmacology, in: F.E. Bloom, D.J. Kupfer (Eds.), Psychopharmacology: The Fourth Generation of Progress, Raven Press, New York, 1995, pp. 719–733. [13] H. Einat, F. Clenet, A. Shaldubina, R. Belmaker, M. Bourin, The antidepressant activity of inositol in the forced swim test involves 5-HT2 receptors, Behav. Brain Res. 118 (2001) 77–83. [14] Y. Endo, K. Shiraki, Behavior and body temperature in rats following chronic foot shock or psychological stress exposure, Physiol. Behav. 71 (2000) 263–268. [15] M. Fleshner, T. Deak, R.L. Spencer, M.L. Laudenslager, L.R. Watkins, S.F. Maier, A long-term increase in basal levels of corticosterone and a decrease in corticosteroid-binding globulin following acute stressor exposure, Endocrinology 136 (1995) 5336– 5342. [16] M. Hansen, K. Nguyen, M. Fleshner, L. Goehler, R. Gaykema, S. Maier, L. Watkins, Effects of vagotomy on serum endotoxin, cytokines, and corticosterone after intraperitoneal lipopolysaccharide, Am. J. Physiol. 278 (2000) R331–R336. [17] M. Hansen, K. Nguyen, L. Goehler, R. Gaykema, M. Fleshner, S. Maier, L. Watkins, Effects of vagotomy on lipopolysaccharideinduced brain interleukin-1beta protein in rats, Auton. Neurosci. 85 (2000) 119–126. [18] B.L. Hart, Biological basis of the behavior of sick animals, Neurosci. Biobehav. Rev. 12 (1988). [19] T. Hosoi, Y. Okuma, Y. Nomura, Electrical stimulation of affarent vagus nerve induces IL-1b expression in the brain and activates HPA axis, Am. J. Physiol. 279 (2000) R141–R147. [20] M. Ishikawa, C. Hara, S. Ohdo, N. Ogawa, Plasma corticosterone response of rats with sociopsychological stress in the communication box, Physiol. Behav. 52 (1992) 475–480. [21] M. Ishikawa, S. Ohdo, H. Watanabe, C. Hara, N. Ogawa, Alteration in circadian rhythm of plasma corticosterone in rats following sociopsychological stress induced by communication box, Physiol. Behav. 57 (1995) 41–47. [22] A.M. Judd, R.M. Mac Leod, Adrenocorticotropin increases interleukin-6 release from rat adrenal zona glomerulosacells, Endocrinology 130 (1992) 1245–1254. [23] G. Kant, J. Leu, S. Anderson, E. Mougey, Effects of chronic stress on plasma corticosterone, ACTH and prolactin, Physiol. Behav. 40 (1987) 775–779. [24] M. Katsura, Y. Mohri, K. Shuto, A. Tsujimura, M. Ukai, S. Ohkuma, Psychological stress, but not physical stress, causes increase in diazepam binding inhibitor (DBI) mRNA expression in mouse brains, Mol. Brain Res. 104 (2002) 103–109. [25] P. Kelliher, T. Connor, A. Harkin, C. Sanchez, J. Kelly, B. Leonard, Varying responses to the rat forced swim test under diurnal and nocturnal conditions, Physiol. Behav. 69 (2000) 531–539. [26] L.G. LeMay, A.J. Vander, M.J. Kluger, The effects of psychological stress on plasma interleukin-6 in rats, Physiol. Behav. 47 (1990) 957–961.

T. Deak et al. / Brain Research 972 (2003) 53–63 [27] S. Maier, K. Nguyen, T. Deak, E. Milligan, L. Watkins, Stress, learned helplessness, and brain interleukin-1 beta, Adv. Exp. Med. Biol. 461 (1999) 235–249. [28] S. Maier, L. Watkins, Cytokines for psychologists:implications of bidirectional immune-to-brain communication for understanding behavior, mood, and cognition, Psychol. Rev. 105 (1998) 83–107. [29] S. Maier, L. Watkins, D. Nance, Multiple routes of action of interleukin-1 on the nervous system, in: R. Ader, D. Felten, N. Cohen (Eds.), Psychoneuroimmunology, Academic Press, New York, 2000, pp. 563–583. [30] S.F. Maier, M.E.P. Seligman, Learned helplessness: theory and evidence, J. Exp. Psychol. Gen. 105 (1976) 3–46. [31] S.F. Maier, L.R. Watkins, Intracerebroventricular interleukin-1 receptor antagonist blocks the enhancement of fear conditioning and interference with escape produced by inescapable shock, Brain Res. 695 (1995) 279–282. [32] S. Makino, T. Shibasaki, N. Yamauchi, T. Nishioka, T. Mimoto, I. Wakabayashi, P.W. Gold, K. Hashimoto, Psychological stress increased corticotropin-releasing hormone mRNA and content in the central nucleus of the amygdala but not in the hypothalamic paraventricular nucleus in the rat, Brain Res. 850 (1999) 136–143. [33] M. Mekaouche, L. Givalois, G. Barbanel, P. Siaud, D. Maurel, F. Malavel, A. Bristow, J. Boissin, I. Assenmacher, G. Izart, Chronic restraint enhances interleukin-1-beta release in the basal state and after an endotoxin challenge, independently of adrenocorticotropin and corticosterone release, Neuroimmunomodulation 1 (1994) 292– 299. [34] E. Milligan, K. Nguyen, T. Deak, J. Hinde, M. Fleshner, L. Watkins, S. Maier, The long term acute phase-like responses that follow acute stressor exposure are blocked by alpha-melanocyte stimulating hormone, Brain Res. 810 (1998) 48–58. [35] M. Minami, Y. Kuraishi, T. Yagaguchi, S. Nakai, Y. Hirai, M. Satoh, Immobilization stress induces interleukin-1b mRNA in the rat hypothalamus, Neurosci. Lett. 123 (1991) 254–256. [36] L.E. Morrow, J.L. McClellan, C.A. Conn, M.J. Kluger, Glucorticoids alter fever and IL-6 responses to psychological stress and to lipopolysaccharide, Am. J. Physiol. 264 (1993) R1010–R1016. [37] K. Nguyen, T. Deak, S. Owens, T. Kohno, M. Fleshner, L. Watkins, S. Maier, Exposure to acute stress induces brain interleukin-1b protein in the rat, J. Neurosci. 18 (1998) 2239–2246. [38] K. Nguyen, T. Deak, M. Will, M. Hansen, B. Hunsaker, M. Fleshner, L. Watkins, S. Maier, Timecourse and corticosterone sensitivity of the brain, pituitary, and serum interleukin-1b protein response to acute stress, Brain Res. 859 (2000) 193–201. [39] J.E. Ottenweller, B.H. Natelson, D.L. Pitman, S.D. Drastal, Adrenocortical and behavioural responses to repeated stressors: toward an animal model of chronic stress-related mental illness, Biol. Psychiatry 26 (1989) 829–841. [40] C. Plata-Salaman, S. Ilyin, N. Turrin, D. Gayle, M. Flynn, T.

[41]

[42]

[43] [44]

[45]

[46]

[47]

[48] [49]

[50]

[51]

[52]

[53]

63

Bedard, Z. Merali, H. Anisman, Neither acute nor chronic exposure to a naturalistic (predator) stressor influences the interleukin-1beta system, tumor necrosis factor-alpha, transforming growth factorbeta1, and neuropeptide mRNAs in specific brain regions, Brain Res. Bull. 51 (2000) 187–193. C.R. Plata-Salaman, S.M. French-Mullen, Intracerebroventricular administration of a specific IL-1 receptor antagonist blocks food and water intake suppression induced by lipopolysaccharide, Physiol. Behav. 51 (1992) 1277–1279. R. Porsolt, G. Anton, N. Blavet, M. Jalfre, Behavioural despair in rats: a new model sensitive to antidepressant treatments, Eur. J. Pharmacol. 47 (1978) 379–391. R. Porsolt, M. LePichon, M. Jalfre, Depression: a new animal model sensitive to antidepressant treatments, Nature 266 (1977) 730–732. C. Pugh, K. Nguyen, J. Gonyea, M. Fleshner, L. Watkins, S. Maier, J. Rudy, Role of interleukin-1 beta in impairment of contextual fear conditioning caused by social isolation, Behav. Brain Res. 106 (1999) 109–118. N. Rothwell, S. Hopkins, Cytokines and the nervous system II: actions and mechanisms of action, Trends Neurosci. 18 (1995) 130–136. B. Schobitz, E. DeKloet, F. Holsboer, Gene expression and function of interleukin 1, interleukin 6 and tumor necrosis factor in the brain, Prog. Neurobiol. 44 (1994) 397–432. F. Shintani, T. Nakaki, S. Kanba, K. Sato, G. Yagi, M. Shiozawa, S. Aiso, R. Kato, M. Asai, Involvement of interleukin-1 in immobilization stress-induced increase in plasma adrenocorticotropic hormone and in release of hypothalamic monoamines in the rat, J. Neurosci. 15 (1995) 1961–1970. A. Sluzewska, Indicators of immune activation in depressed patients, Adv. Exp. Med. Biol. 461 (1999) 59–74. B.L. Spangelo, A.M. Judd, R.M. Mac Leod, D.W. Goodman, P.C. Isakson, Production of interleukin-6 by anterior pituitary cells in vitro, Endocrinology 126 (1990) 582–586. P. Taishi, S. Bredow, N. Guha-Thakurta, F. Obal, J. Krueger, Diurnal variations of interleukin-1b mRNA and b-actin mRNA in rat brain, J. Neuroimmunol. 75 (1997) 69–74. L. Watkins, M. Hansen, K. Nguyen, J. Lee, S. Maier, Dynamic regulation of the proinflammatory cytokine interleukin-1b: molecular biology for non-molecular biologists, Life Sci. 65 (1999) 449– 481. J.T. Whicher, C.I. Westacott, The acute phase response, in: J.T. Whicher, S.W. Evans (Eds.), Biochemistry of Inflammation, Vol. 18, Kluwer, Boston, MA, 1992. D. Zhuo, A.W. Kusnecov, M.R. Shurin, M. DePaoli, B.S. Rabin, Exposure to physical and psychological stressors elevates plasma interleukin 6: relationship to the activation of the hypothalamic– pituitary–adrenal axis, Endocrinology 133 (1993) 2523–2530.