Peripheral lipopolysaccharide administration impairs two-way active avoidance conditioning in C57BL/6J mice

Abstract

Peripheral administration of lipopolysaccharide (LPS) or interleukin-1 (IL-1) may lead to alterations of CNS function and behavioral changes designated “sickness behavior.” Further, some experiments show evidence of LPS- and cytokine-mediated alterations in learning and memory. The current series of experiments examined the effects of a single or repeated intraperitoneal LPS injections, at a number of doses and time points before or after test sessions, on behavior in a two-way active avoidance conditioning paradigm. Subjects were able to avoid the mild shock stimulus, escape it, or fail to respond to it. Subjects treated with LPS at many, but not all, of the time points sampled showed impaired learning, by exhibiting significantly fewer avoidance responses than controls. Furthermore, an LPS-induced increase in non-cued inter-trial interval crossings was observed during the later days of testing, suggesting that a greater percentage of their avoidance responses was not conditioned and their behavior was less efficient. Taken together, the results suggest that LPS-treated animals showed a diminished association between conditioned stimulus (CS) and unconditioned stimulus (US). These results support the theory that peripheral immune stimuli may induce deleterious effects on learning, and extend the work to a negatively reinforced operant procedure.

Introduction

Numerous functional links between the nervous system and the immune system have been elucidated by scientists in diverse disciplines [1]. One key mediator of immune system effects on the nervous system is the proinflammatory cytokine, interleukin-1β (IL-1β) [2], [3], [4], [5]. IL-1 is expressed following exposure to a pathogen or lipopolysaccharide (LPS; an immunostimulatory component of Gram-negative bacterial cell wall, often used as an experimental immune stimulus). Although IL-1 cannot completely account for all the behavioral effects of LPS [6], [7], [8], some of these effects can be blocked through administration of an IL-1 receptor antagonist [5], [9], suggesting a key role for this cytokine in mediating many behavioral effects of LPS.

LPS and IL-1 have been shown to induce a variety of neurobiological effects, including alterations in monoamine turnover in various brain regions [10], [11], [12], [13], activation of the hypothalamic–pituitary–adrenal (HPA) axis by triggering corticotropin-releasing factor (CRF) release from the hypothalamus [14], [15], [16], and alteration of long-term potentiation (a neurobiological correlate of learning) in CA1, CA3, and the dentate gyrus of the hippocampal formation [17], [18], [19], [20].

Peripheral or central administration of LPS, IL-1β, or IL-6 induces a complex array of behaviors including anorexia, anhedonia, decreased locomotion and exploratory behavior, increased anxiety, somnolence, and general behavioral depression [2], [21], [22]. These “sickness behaviors” are thought to reflect an adaptive state of altered motivation, rather than the passive response of an illness-debilitated animal [5], [23], [24]. Further, exposure to LPS and some cytokines may lead to alterations in central processes involved in learning and memory. For example, a number of studies have reported that administration of live bacteria, LPS, or IL-1β alters performance in the Morris water maze (MWM; a hippocampus-dependent spatial learning paradigm), and that LPS-, bacteria-, or cytokine-treated animals generally have shown increased latencies and swim distances in finding the hidden platform [25], [26], [27], [28], [29], [30]. However, other investigators found that the primary deficit in MWM performance for LPS-treated male and female mice was a significant reduction in swim speed, with no effect or only a modest effect upon learning per se [31] (Sparkman et al., submitted for publication). Further, another study that examined behavior in the MWM showed that intracerebroventricular (i.c.v.) infusion of IL-1β had no effect on memory, and infusion of interleukin-1 receptor antagonist (IL-1Ra) actually caused memory impairment, potentially indicating a facilitating role for IL-1β in normal learning and memory processes [32]. Consistent with these findings and others (e.g., [33], [34]), Gibertini [35] has argued that low doses of IL-1 tend to facilitate learning, but that high doses impair it.

To further explore this phenomenon, other investigators have utilized different testing paradigms to examine LPS and IL-1 effects upon learning. For example, learning/memory impairments following LPS and/or IL-1 exposure were demonstrated in an autoshaping procedure [36], a Y-maze procedure [37], and were shown particularly nicely in a contextual fear conditioning procedure [5], [9], [38], [39]. However, again, the behavioral effects of LPS and interleukin-1 are not completely straightforward. For example, one group that examined behavior in a delayed-type matching-to-sample procedure and Y-maze spatial memory noted performance effects due to motivational factors, but no clear effects on learning per se [40]. Moreover, Yirmiya et al. [32] found that i.c.v. infusion of IL-1β caused an improvement in passive avoidance conditioning, while IL-1Ra caused a memory impairment. Therefore, whether or not functionally pleiotropic cytokines are harmful or beneficial to neural and behavioral function likely depends upon a number of important factors, such as dose, gender, age, species, timing of administration, and physiological context, along with aspects of the behavioral testing parameters and other aspects of individual laboratories [9], [32], [38], [40], [41], [42].

Despite encouraging inroads made thus far, clearly more work is needed to explore and characterize LPS- and cytokine-induced effects on learning and memory. The current study was designed to investigate the effects of peripheral LPS exposure on behavior in a well-established test of learning that has not previously been utilized to explore learning effects of LPS administration, two-way active avoidance conditioning. This task utilizes a negatively reinforced operant response, and is often thought to be primarily striatum-dependent. However, lesions of the hippocampus or fimbria fornix, or administration of anticholinergic agents strongly alter performance in this task, in part via behavioral disinhibition and increased general activity [43], [44], [45], leading to decreased response efficiency. Further, since this testing paradigm combines many testing trials per day across multiple testing days with sufficient motivation to perform the task (even in animals experiencing sickness behavior), without the considerable thermoregulatory and motor demands of a water maze, it may be better able to tease apart labile alterations in behavior that are due to non-learning performance decrements from actual learning decrements. Moreover, since this task shares some interesting similarities to the autoshaping and fear conditioning paradigms (e.g., involvement of both Pavlovian and operant components as in the autoshaping procedure, and a specific Pavlovian fear conditioning component that works against the animals learning the operant two-way active avoidance task), these conceptual similarities may facilitate comparisons with these earlier data sets that may buttress these findings or provide alternate vantage points. Lastly, it has been shown that certain autoimmune-disordered mice, which display unusual cytokine expression profiles [46], [47], [48], exhibit performance decrements in this task. These decrements have been suggested to be in part due to factors such as altered attention, motivation, or emotionality [49], [50].

Our major hypothesis was that even a single peripheral administration of LPS during early phases of the testing (e.g., prior to the first day of testing, or the memory consolidation associated with that day's learning) would produce distinguishable performance and learning deficits in subjects tested in this task, while LPS given at time points following the early days of training would produce only non-learning performance decrements or possibly no effects at all. Further, we hypothesized that animals given repeated injections of LPS would show additive effects of the LPS exposure only until day 3 of testing, when tolerance to the LPS would become apparent.