INTRODUCTION

Veterans returning from the first Gulf War registered a number of nonspecific somatic (eg headaches, rashes, joint pain, fatigue, gastrointestinal distress) and cognitive (lack of concentration, confusion, short-term memory loss) complaints referred to as Gulf War Illness (GWI). GWI is shrouded in controversy stemming from the nonspecific nature of complaints, the lack of consistent biological markers, reporting biases, recollection biases, and issues regarding deployment.

The search for causes has focused attention on factors that could directly induce illness or persistent symptoms. One such factor is pyridostigmine bromide (PB), a carbamate inhibitor of acetylcholinesterase used extensively as a pretreatment under the threat of nerve gas exposure. At the recommended dose, PB is considered relatively safe, because PB does not readily penetrate the blood–brain barrier (Lallement et al, 1998; Beck et al, 2001, 2003). The safety of PB treatment is less certain in the presence of other chemical exposures (Abou et al, 1996; Hoy et al, 2000), stress (Friedman et al, 1996; Grauer et al, 2000; Servatius et al, 2000; Sinton et al, 2000; Kant et al, 2001; Song et al, 2002; Beck et al, 2003), and in individuals with increased sensitivity to PB (Servatius et al, 1998; Beck et al, 2001). Safety notwithstanding, a role for PB in GWI would still be in doubt in that the signs and symptoms of GWI are not merely reducible to cholinergic toxicity.

Is a direct biological path necessary? Ferguson and Cassaday suggested that the symptoms of illness arose secondary to Pavlovian conditioning. They noted the resemblance of the symptoms of GWI to the acute-phase responses of proinflammatory cytokines, in particular interleukin (IL)-1β (Ferguson and Cassaday, 1999). In their formulation, GWI is presumed to be the result of reactivation of IL-1β through situational reminders of the precipitating event (Ferguson and Cassaday, 2001). If true, proinflammatory activation should be evident, or sensitized in veterans with GWI. Contrary to this prediction, elevated IL-1β in sick veterans is lacking (Soetekouw et al, 1999).

Despite the failure of the specific prediction, a conditioning model may still apply to the involvement of both PB and IL-1β. Both IL-1β and PB constitute salient interoceptive stressors. Treatment with of IL-1β results in fever, fatigue, malaise, and anhedonia (Anisman and Merali, 1999). PB produces a number of unpleasant effects: nausea, stomach cramping, fatigue, excessive lacrimation, diarrhea, and urinary urgency (Sharabi et al, 1991). Thus, the potential exists for the side effects of PB (Romano and King, 1987) and IL-1β to serve as unconditional stimuli (USs) for Pavlovian conditioning.

Here we focused our attention on the associability of PB and IL-1β as interoceptive stressors with contextual stimuli. Contextual fear or anxiety arising from exposure to stress is typically observed in rodent models as freezing (Fanselow, 1990) or as an exaggerated acoustic startle response (ASR) (Davis, 1990). The advantage of the latter is that an exaggerated ASR is a positive sign of anxiety that is distinct from malaise. To expose contextual learning, a discrimination paradigm was employed. Four otherwise identical chambers were distinguished by either visual or olfactory cues. Exposure to stressors occurred in one of two contexts. Conditioning should be apparent as increased ASRs in the context concordant with stressor exposure. Moreover, such learning should persist, that is, subsequent exposures to contextual cues should elicit potentiated ASRs. The associative strength of contextual features will depend on the nature of the US and the form and saliency of available conditional stimuli (CSs) (Garcia et al, 1966, 1968; Garcia and Koelling, 1966). Therefore, contexts were distinguished by visual or olfactory cues.

METHODS

Subjects

Adult male Sprague–Dawley rats (300–450 g) were obtained from Charles River (Willington, MA); rats were allowed 2 weeks to acclimate to laboratory conditions. Rats were individually housed in specially designed chambers (eight rats per chamber) that attenuate sound, and maintain airflow and temperature. Rats were fed ad lib Purina Rodent Chow with free access to water. Rats were maintained on a 12 : 12 light:dark cycle with light onset at 0800. All procedures conformed to the Institutional Animal Care and Use Committee of the Department of Veterans Affairs, New Jersey Health Care System.

Materials and Apparatus

The startle apparatus (holders (EO5-15), platforms (E45-11), amplifier, and interface) were obtained from Coulbourn Instruments (Allentown, PA). The software package (Labview) (used for the generation of acoustic stimuli, stimulus delivery, and signal acquisition) and the A/D card (6024E) were obtained from National Instruments (Austin, TX).

Four custom experimental boxes were used for stressor exposure and ASR testing. The boxes are constructed of 2 mm thick PVC, which sandwiches a 3.5 cm foam core. The boxes have smooth inner surfaces with inner dimensions of 26 × 29 × 43 cm. A house light (3 W) is affixed to the top panel as well as a baffled ventilation fan. The front panel is windowed for observation.

Stressors

Interoceptive stress was manipulated by delivering either peripheral cholinesterase inhibitors or proinflammatory cytokines through intraperitoneal (i.p.) injections. PB and neostigmine bromide (NB) (Sigma Chemical, St Louis, MO) were freshly prepared the morning prior to injection and dissolved in physiological saline. Preliminary work established levels of PB and NB that inhibited plasma butyrylcholinesterase activity by 30%. At the levels of PB and NB administered, rats exhibit mild signs of cholinergic overstimulation: barely noticeable muscle fasciculation, increased lacrimation, and loose stool. IL-1β (specific activity >5 × 108 U/mg) was obtained from PeproTech (Rocky Hills, NJ). Preliminary work showed that IL-1β (1.0 and 3.0 μg/kg) induced a mild but significant hyperthermia that lasted for 2–4 h in rats. An exteroceptive stressor was devised that would be continuously arousing for 1 h without unduly harming the rat. A clip, part of the system for delivery of tailshock (Ottenweller et al, 1992) but without the electrodes, was attached to the rat's tail. Observation of the rat indicated that the clip remained uncomfortable throughout the exposure session, with intermittent gnawing and circling behavior. Neither the clip nor the gnawing by the rat damaged a single rat's tail.

ASR

Rats were placed in holders that sit on force transducers as previously described (Servatius et al, 1998). The open cover of the holders is formed by a tubular alignment of rods, spaced 0.3–2 cm apart, allowing the rat to observe their surrounds. Each session consisted of 60, 102-dB white noise bursts (100-ms duration, immediate rise/fall). The interstimulus interval ranged from 25 to 35 s. Response magnitude was calculated on a trial-by-trial basis (Beck et al, 2002). For each stimulus presentation, a response threshold for whole-body response was computed as the average rectified activity 200 ms prior to stimulus onset plus six times the standard deviation of that rectified activity. Response amplitudes, the maximum rectified activity within 125 ms after stimulus onset, were only recorded when poststimulus activity exceeded the response threshold. For trials in which activity did not reach this criterion, ‘not available’ was recorded.

Procedures

Rats were stratified on body weight and randomly assigned to groups. Stressor exposure occurred on DAY 1. For interoceptive stressors, rats received a single i.p. injection. For clip exposure, a clip was attached to the base of the tail. In all cases, rats were placed in the holders used for ASR testing. The experimental chambers were distinguished by either visual or olfactory cues. To manipulate visual cues, masking tape was affixed to the walls in irregular patterns in two of the boxes; the other two boxes remained unadorned. Thus, wall adornment was the basis for distinguishing chambers in experiments manipulating visually distinguished contexts. To manipulate olfactory cues, segments of commercially available car air fresheners (strawberry or peppermint; Car-Freshener Corporation, Watertown, NY) were affixed to the ceiling of the chambers. Segments of the car air fresheners were of equal area; new segments were used each day of exposure. Order of exposure for this initial period was counterbalanced across conditions. The duration of exposure was 1 h. After this exposure session, rats were returned to their home cages.

The following day (DAY 2), the first ASR test commenced. The rats were placed in the chambers for 5 min prior to the first ASR trial. The SAME designation refers to rats whose ASR testing occurred in a context concordant with initial exposure. The DIFF designation refers to rats whose ASR testing occurred in a chamber not experienced during initial exposure. A second ASR test occurred 2 weeks after stressor exposure (DAY 15). For all rats, the DAY 15 test occurred in a chamber concordant with the DAY 2 test.

Data Analysis

A startle session was reduced to 12 blocks of five responses each. The first block represents initial reactivity; the magnitude of these responses is typically higher with greater variability than the subsequent blocks. In that we were particularly interested in initial reactivity, the initial block of ASRs were separately analyzed. For these data, between-subject (Treatment × Context) analyses of variance (ANOVAs) were performed. For the subsequent blocks, split-plot ANOVAs with repeated measures were performed. With the ASR protocol used herein, typically ASRs decrement substantially from the first block to subsequent blocks indicating habituation. For some, ASRs increment during the subsequent blocks, suggesting the development of sensitization or dishabitutation. Dunnett's and Dunn's tests (tDs) were used to compare SAME and DIFF responses within treatments.

RESULTS

PB with Visual Cuing

We administered to rats PB at 0.1 or 1.0 mg/kg or saline vehicle through intraperitoneal (i.p.) injection. Preliminary data showed that these doses correspond to 15 and 35% inhibition of plasma butyrylcholinesterase activity, respectively. These doses produce inhibition in the range recommended for PB as a pretreatment under the threat of nerve gas.

On DAY 2, PB-treated rats displayed exaggerated ASRs in the DIFF context both initially and over the rest of the test session (see Figure 1). For initial responses, a significant main effect of Context, F(1,30)=7.5, was qualified by the Treatment × Context interaction, F(2,30)=4.1, all p's <0.05. Overall responses also differed. The main effect of Context, F(1,300)=5.6, and the Treatment × Context × Block interaction, F(20,300)=1.7, was significant, all p's <0.05. Contrary to expectation, startle reactivity was enhanced in a context that differed from the context in which rats received PB exposure.

Figure 1
figure 1

Comparison of ASRs on DAY 2 and DAY 15 in rats given PB (0.1 or 1.0 mg/kg i.p.) or vehicle (VEH) injections in contexts distinguished by visual cues. On DAY 2, both initial ASRs and overall ASRs were exaggerated in DIFF compared to SAME in PB-treated rats, but not VEH rats (n=6 per group). No differences among groups were detected on DAY 15.

Full size image

Enhanced reactivity was transitory. For DAY 15, initial reactivity did not differ. Although analyses of the overall responses indicated a significant Treatment × Context × Block interaction, F(20,300)=2.3, p<0.05, the significant contrasts were not germane. Thus, the enhanced ASRs of PB-treated rats on DAY 2 dissipated by DAY 15.

PB Compared to NB with Visual Cuing

NB is similar to PB, but with a shorter half-life and greater potency. A direct comparison of PB and NB indicates whether effects attributable to PB are unique to its pharmacokinetic properties or generalize to other cholinomimetics. In addition, the comparison allows for a demonstration of the reproducibility of the phenomena with PB. Inasmuch as the greatest ASRs were observed to 0.1 mg/kg, this dose of PB was compared to an equivalent dose of NB (0.016 mg/kg).

On DAY 2, initial reactivity was exaggerated in both PB- and NB-treated rats in DIFF compared to SAME contexts (see Figure 2). The main effect of Context, F(1,48)=9.9, was qualified by the Treatment × Context interaction, F(2,48)=3.6, all p's <0.05. For overall responding, only the main effect of Block was significant, F(10,480)=2.7, p<0.01. Specific comparison indicated that PB-treated rats exhibited exaggerated ASRs in DIFF compared to SAME (tD=2.25, p<0.05). Treatment with AChE inhibitors led to exaggerated ASRs in contexts that differed from those of treatment.

Figure 2
figure 2

Comparison of ASRs on DAY 2 and DAY 15 in rats given PB (0.1 mg/kg), NB (0.16 mg/kg), or VEH injections in contexts distinguished by visual cues. On DAY 2, initial ASRs were exaggerated in DIFF in NB-treated rats, while overall responses were exaggerated in PB-treated rats in DIFF (n=9 per group). No differences among groups were detected on DAY 15.

Full size image

Again, enhanced reactivity was transitory. No significant effects were noted in either initial or overall response on DAY 15. Although enhanced reactivity was apparent in rats given cholinomimetics and tested in DIFF, the overall magnitude of these effects was reduced in comparison to those observed in Experiment 1.

IL-1β with Visual Cuing

Exaggerated ASRs observed after PB and NB treatment in the presence of visual cues are presumed to be attributable to the unwell feelings produced by cholinergic overstimulation. IL-1β is known as an adjunct to sickness behavior, inducing fever, malaise, and anhedonia (Anisman and Merali, 1999). Thus, IL-1β is an interoceptive stressor (Maier and Watkins, 1998). The visual cuing experiment was preformed as above, with the exception that rats were treated with IL-1β (1.0 or 3.0 μg/kg i.p.) or a VEH injection. These doses produce mild hyperthermia, and the higher dose leads to facilitated acquisition of the classically conditioned eyeblink response within 2 h of injection (Servatius and Beck, 2003).

On DAY 2, initial responses did not differ among groups (see Figure 3). Overall responses only differed as a function of Block, F(10,300)=2.4, p<0.01. Thus, treatment with IL-1β did not affect reactivity in either context. As for DAY 15, no differences were detected in either initial or overall responses, all p's >0.05. Thus, treatment with IL-1β did not induce enhanced reactivity, in SAME or DIFF, regardless of treatment context and day of testing.

Figure 3
figure 3

Comparison of ASRs on DAY 2 and DAY 15 in rats given IL-1β (1.0 or 3.0 μg/kg i.p.) or VEH injections in contexts distinguished by visual cues. No differences were detected among groups either DAY 2 or DAY 15 (n=6 per group).

Full size image

Exteroceptive Stressor and Visual Cuing

Contextual learning in the face of exteroceptive stressors is commonly studied. For example, exposure to footshock engenders robust behavioral changes within the exposure context indicative of fear (Iwata and Ledoux, 1988; Fanselow and Tighe, 1988) including the potentiation of the ASR (Davis, 1989; Richardson and Elsayed, 1998). Manipulation of exteroceptive stress with regard to contextual fear typically involves punctate stressors. To manipulate exteroceptive stress in a manner similar to interoceptive stressors, we sought a stressor that would continually engage the rat for an hour, but would not unduly harm or injure the rat. Therefore, a polyurethane clip was attached to the base of the rat's tail. The clip appeared to be annoying to the rat (inducing circling, gnawing, and biting sporadically throughout the 1 h period).

No changes in reactivity were apparent across contextual cues or as a function of stressor exposure on DAY 2, all p's >0.05 (see Figure 4). Similarly, no differences in ASR were noted on DAY 15, all p's >0.05.

Figure 4
figure 4

Comparison of ASRs on DAY 2 and DAY 15 of rats with a clip affixed to their tail (C) or no clip (NC) in contexts distinguished by visual cues. No differences were detected among groups either DAY 2 or DAY 15 (n=6 per group).

Full size image

PB with Olfactory Cues

Enhanced reactivity in PB-treated rats in the presence of novel visual cues was unexpected. Rats may not readily attribute interoceptive stimuli associated with cholinergic overstimulation to the visual features of their surroundings. The classic work of Garcia et al (1966) suggests a natural linkage between sickness and olfactory and gustatory sensations. Thus, fear or anxiety attributable to the surrounding in which one became sick (Garcia and Koelling, 1966; Holder and Garcia, 1987) may be more apparent when contexts are differentiated by odors (Herzog and Otto, 1997; Otto et al, 2000). Therefore, one would expect that ASR potentiation should be observed in SAME contexts with olfactory cues.

To simplify the experimental design, a single dosage of PB was chosen. In the first two experiments manipulating PB in visually distinguished contexts, PB was administered at 1.0 and 0.1 mg/kg. There was concern that the effects of 0.1 mg/kg may be too subtle, and there was a robust effect in Experiment 1, which was less apparent in Experiment 2. Therefore, we treated the rats with PB at 0.5 mg/kg, approximately midway between the two effective doses of 0.1 and 1.0 mg/kg. Two novel olfactory cues (strawberry and peppermint car air fresheners) were used to distinguish the chambers.

On DAY 2, initial reactivity did not differ. However, exaggerated ASRs were evident in PB-treated rats during the remainder of the test session (see Figure 5). The main effect of Treatment was significant, F(1,20)=4.3, p<0.05. Thus, PB-treated rats exhibited enhanced reactivity in both SAME and DIFF on DAY 2.

Figure 5
figure 5

Comparison of ASRs on DAY 2 and DAY 15 in rats given PB (0.5 mg/kg i.p.) or VEH injections in contexts distinguished by odors. On DAY 2, initial responses did not differ, but the overall ASRs of rats given PB were exaggerated regardless of context (n=6 per group). The line below the significance symbol indicates a main effect comparison between PB-treated and VEH-treated rats. On DAY 15, both initial and overall ASRs of rats given PB and tested in SAME were exaggerated compared to those tested in DIFF.

Full size image

On DAY 15, discrimination between SAME and DIFF was evident. For initial responses, the main effects of Treatment, F(1,20)=10.5, and Context, F(1,20)=4.9, were qualified by the Treatment × Context interaction, F(1,20)=7.3, all p's <0.01. Overall responses yielded a similar pattern, with PB-treated rats exhibiting exaggerated ASRs in SAME. The main effects of Treatment, F(1,20)=5.7, and Block, F(10, 200)=2.3, were significant, all p's <0.05. Specific comparisons demonstrated that the ASRs of PB-treated rats in SAME were greater than those in DIFF, tD=3.1, p<0.01. PB-treated rats discriminated between SAME and DIFF olfactory contexts 15 days after exposure.

IL-1β with Olfactory Cues

Contextual learning was apparent when odors distinguished contexts in which the rats experienced cholinergic overstimulation. To determine generalizability of contextual learning with interoceptive USs, rats were treated with IL-1β (3.0 μg/kg i.p.) or vehicle. To simplify the design, only the 3.0 μg/kg dose of IL-1β, one that facilitates acquisition of the classically conditioned eyeblink response in rats (Servatius and Beck, 2003), was administered. Again, the chambers were differentiated by peppermint or strawberry scents.

On DAY 2, initial reactivity did not differ (see Figure 6). However, exaggerated ASRs in rats treated with IL-1β developed over time. The main effects of Context, F(1,20)=13.2, and Block, F(10,200)=2.3, were significant. Specific comparison indicated that IL-1β-treated rats in SAME had larger ASRs than those in DIFF, tD=2.9, p<0.05.

Figure 6
figure 6

Comparison of ASRs on DAY 2 and DAY 15 in rats given IL-1β (3.0 μg/kg i.p.) or VEH injections in contexts distinguished by odors. On DAY 2, initial ASRs did not differ (n=6 per group). Overall ASRs were exaggerated in SAME relative to DIFF, an effect predominantly derived from rats given IL-1β. On DAY 15, initial ASRs did not differ; however, overall ASRs were exaggerated in SAME relative to DIFF contexts in IL-1β-treated rats.

Full size image

Exaggerated reactivity in SAME was still apparent in IL-1β-treated rats on DAY 15. Although the analysis of initial responses did not indicate significance, specific comparison demonstrated that IL-1β-treated rats tested in SAME displayed greater ASRs than those tested in DIFF, tD=2.0, p<0.05. A similar pattern was observed in overall ASRs. The Treatment × Context interaction, F(1,20)=4.6, was significant, p<0.05. Again, the ASRs of IL-1β rats tested in SAME were greater than those tested in DIFF, tD=3.3, p<0.01. Similar to PB treatment, rats treated with IL-1β displayed exaggerated ASRs 2 weeks after acute exposure in contexts congruent with treatment.

DISCUSSION

Interoceptive stimuli can be powerful USs for associations with contextual features. Such Pavlovian relations are clearly evident in development of anticipatory nausea in cancer patients or animal models of illness such as treatment with LiCl in rats. In both these cases, the interoceptive stressor is nausea. Do associations form with mild interoceptive stimuli, those that produce sensations less than frank illness? We examined whether visual or olfactory features would form associations with mild interoceptive stimuli accompanying peripheral cholinesterase inhibitors and proinflammatory cytokines.

Visual Cues

Contrary to our initial expectations, rats given peripherally acting cholinomimetics exhibited exaggerated ASRs in novel surroundings. Exaggerated ASRs were evident early in testing and dissipated over the initial test session. Increased reactivity in the presence of changed visual features assumes that rats attended to the visual features during initial exposure. However, those features were not directly associated with physiological disturbances secondary to cholinergic overstimulation. This finding was replicated in rats given PB and reproduced in rats given NB, albeit with a smaller degree of difference in reactivity compared to the first experiment. Changes in reactivity, within either SAME or DIFF environments, were not apparent 2 weeks after initial exposure. We do not know if the apparent increased reactivity to novel visual features persists after the initial test inasmuch as the contexts of the retest were congruent with the initial test session. Thus, exposure to cholinomimetics that do not cross the blood–brain barrier leads to transiently increased reactivity in a novel environment distinguished by visual stimuli.

Exposure to an exteroceptive stressor did not lead to exaggerated ASRs either in SAME or DIFF when environments were distinguished by visual cues. Observation of the rat indicated that focus of the rat's attention was on the tailclip itself with bouts of activity to remove the tailclip. Thus, the US was explicit, causing specific discomfort to the rat. Under these conditions, the place in which stress occurred appears to be irrelevant. Contextual learning during periods of footshock or tailshock exposure is apparent later as freezing (Iwata et al, 1986; Fanselow and Tighe, 1988), exaggerated ASRs (Davis, 1989), and facilitated classical conditioning (Shors and Servatius, 1997). A rat is anxious or fearful in the surroundings concordant with shock exposure. The presence of punctate conditional signals attenuates contextual fear (Baldi et al, 2004). Thus, knowledge that allows for prediction or control attenuates contextual fear (Jackson and Minor, 1988). In the case of the tailclip, the source of irritation is continuous, known, and observable. Contextual features may be ignored as a potential source of the mild irritation because the source of the irritation is so predictable in this instance.

Potentiated ASRs were not observed in visual contexts after concomitant IL-1β treatment. The doses of IL-1β we used herein produce mild hyperthermia, on the order of 0.3–0.4°C, that lasts for several hours (unpublished observations). Our behavioral data suggest that these levels of IL-1β serve as interoceptive stressors leading to changes in ASRs (Beck and Servatius, 2003) and classical conditioning (Servatius and Beck, 2003), as observed after exposure to inescapable shocks (Servatius and Shors, 1994; Servatius et al, 2001). These data suggest that as interoceptive USs, PB and IL-1β can be dissociated with respect to associative learning of contexts. However, a direct test of this possibility would require evidence that the aversive sensations attributable to cholinergic overstimulation and proinflammatory activation are equivalent.

Manipulation along the interoceptive/exteroceptive dimension was not wholly orthogonal or independent. A stressor is considered exteroceptive or interoceptive based on the initial source of stimulation. Exteroceptive stimuli involve the five basic senses of touch, smell, vision, taste, and hearing. Interoceptive stimuli arise from within the body (eg gastric distention, inflammation, fever). Exteroceptive stressors—such as tail pinch, foot shock, and tailshock—will induce a cascade of physiological adjustments secondary to exteroceptive stressors. Indeed the tailclip would be expected to induce activation of the sympathetic nervous system (Antelman and Szechtman, 1975), hypothalamic–pituitary–adrenal axis (HPAA) (Kirby et al, 1997), and possibly a proinflammatory response. On the other hand, PB and IL-1β were administered through i.p. injection, thus having an exteroceptive component. While we did not include noninjected controls, the levels of ASR in vehicles were similar to those obtained in other experiments performed at a similar time using noninjected rats. Cessation of afferent traffic during stressor exposure, whether exteroceptive or interoceptive, would be necessary to address directly this point.

Olfactory Cues

Interoceptive sensations attributable to PB and IL-1β treatment were readily associated with odor. Treatment with both PB and IL-1β led to potentiated ASRs when the contexts for testing were concordant with that of exposure. In contrast to the exaggerated responses in visually distinguished novel contexts, exaggerated responses in olfactory contexts developed over the initial test session. These data suggest that the reaction to the odor previously paired with interoceptive stressor required several minutes to develop.

Exaggerated ASRs in SAME were apparent 2 weeks after stressor exposure. Here, both initial ASRs and session-wide ASRs were exaggerated. The increased reactivity specifically in the presence of odors paired with interoceptive stressors is consistent with the induction of fear or anxiety, reinforced through a recapitulation of interoceptive stimuli.

Odors discretely paired with exteroceptive stressors (footshock) (Otto et al, 1997; Richardson et al, 1999; Paschall and Davis, 2002), but not after pairing with illness from LiCl (Richardson and McNally, 2003), potentiate ASRs. The lack of contextual learning after treatment with LiCl paired with odors could be accounted for by a number of procedural and design differences between the Richardson and McNally studies and the present work. For one, our protocol involved a single pairing of the interoceptive stressor and context; Richardson and McNally used several pairings over a 2-week period prior to the initial startle test. For another, the discrimination evaluated differed. Our protocol evaluated treated or vehicle rats in one of two odor-distinguished contexts. Richardson and McNally contrasted the ASRs of rats given LiCl in the experimental chamber or home cage, and then tested in the experimental chamber. Also, the procedures varied in the ASR test. Our ASR protocol delivers twice the number of acoustic bursts as the Richardson and McNally protocol. In that the exaggerated ASRs of the initial test were evident later in the test session, this difference in ASR test duration (number) may have played a role. In addition, the discrimination was much more apparent during the second session conducted 2 weeks after the initial; Richardson and McNally evaluated the ASR in a single session conducted the day after the last pairing.

In addition to the learning apparent in contexts differentiated by odors, rats treated with PB exhibited a transient increased reactivity within the novel odor. This increased reactivity was not apparent after IL-1β treatment. Together with the increased reactivity in visually differentiated contexts, it seems that treatment with cholinomimetics alters reactivity to novel features in the environment. Future research needs to examine the stability, generalizability, and robustness of these transient changes in reactivity.

Implications

Dissociation between visual- and odor-differentiated contexts in producing potentiated ASRs is in keeping with the classic work of Garcia and Koelling (1966). Whereas it is likely that the environmental cues used to distinguish contexts vary in saliency (wall adornment and odors), cue saliency would not provide a good explanation for why rats reacted to novelty in visually distinguished contexts and reacted in odor-distinguished contexts congruent with treatment. Here, we held the challenge, ASR, constant across experiments to facilitate comparisons among treatments. Different challenges, such as chemicals, infectious, or biological agents, may reveal distinct patterns of enhanced reactivity relative to contextual features.

A second dissociation was evident between cholinomimetics and IL-1β. Cholinomimetics affect reactivity in both visually and olfactory-distinguished contexts. However, IL-1β only affected reactivity in olfactory-distinguished contexts. While further research would be necessary to validate this impression, it is intriguing that learning related to exteroceptive features could depend on the nature of the interoceptive stressor.

The dissociation between cholinomimetics and IL-1β may reflect their differences in the peripheral organ and tissues affected or the central processing of the respective interoceptive changes. Proinflammatory cytokines comprise acute-phase responses for infections, leading to a cascade of immune (eg IL-1 receptor antagonist, IL-6, and IL-10) and neurohormonal responses (sympathetic and HPAA activation). Systemic administration of IL-1β increases norepinephrine, serotonin, and dopamine turnover in the same regions as many stressors: paraventricular nucleus of the hypothalamus, locus coeruleus, central amygdala, and prefrontal cortex (Brebner et al, 2000). Brain stem and amygdala nuclei mediate HPAA responses (Xu et al, 1999). On the other hand, PB will affect cholinergic transmission at autonomic ganglia as well as postganglionic parasympathetic terminals, and the neuromuscular junction. Beyond this broad characterization, specific mapping of brain activation following PB treatment has not been performed, likely due to the peripheral nature of PB. Therapeutic levels of PB treatment will induce a mild increase in plasma corticosterone in rats (Servatius et al, 2000); however, PB is known to induce growth hormone and inhibit adrenocorticotrophin (ACTH) levels in humans by direct actions on the pituitary (Llorente et al, 1996). The elevated corticosterone levels in rats may reflect the emotional components attributable to PB treatment overwhelming the direct suppressive actions on ACTH or species-specific actions. Nonetheless, both IL-1β and PB treatment (and the tailclip for that matter) will induce an HPAA response; however, the behavioral patterns differ between treatments arguing against corticosterone influencing learning in this paradigm. A level of processing sensitive to contextual features, such as the amygdala, hippocampus, entorhinal cortex, and periaquaductal grey, likely mediates the learning and enhanced reactivity in this paradigm.

Unexplained illnesses (chronic fatigue syndrome, fibromyalgia, GWI, multiple chemical sensitivity) have in common the appearance of nonspecific symptoms without an underlying medical explanation. In the case of GWI, there existed the possibility that a single common feature could be identified, given that the lives of troops are more structured than civilians. However, none has been identified. One relatively unexplored possibility is that learning accounts for the generation and maintenance of nonspecific symptoms. Learning processes (eg second order or higher learning and occasion setting) may fill the gap between underlying medical disturbances and the appearance of symptoms. The present data showing first-order conditioning with PB and IL-1β as USs provide initial support for such a hypothesis.

In summary, the aversive properties of even mild interoceptive stressors (those that do not produce frank illness) supported contextual learning, the nature of which depended on the features of the context. Contextual learning, especially in the presence of odors associated with interoceptive stressors, persisted for weeks after a single pairing.