In an attempt to tease out just how cytokines affect learning and memory, Maier and his colleagues studied rats given so-called fear conditioning. Rats were taken out of their normal cage, placed in a new one, and then a tone was sounded, followed by a mild electric shock sent through the floor. After a couple of days back in their habitual cage, the animals were returned to the "new" cage, but this time, played only the tone. Normally, the rats can remember that both the "new" cage - the "context" - and the tone are associated with adverse stimuli, and respond by "freezing" with fear in anticipation of a shock, says Maier.

However, animals injected with lipopolysaccharide (LPS), a consituent of bacterial cell walls known to increase cytokine signaling in the blood and in the brain, did not react when they were put inside the "new" cage, but did freeze when they heard the tone. The hippocampus is the region of the brain involved in the recognition of place and the consolidation of contextual memories, says Maier, suggesting that the LPS somehow triggers an effect in the hippocampus.

LPS induces a cytokine called interleukin-1- (IL-1), which can also be made in the brain, so the researchers tried several ways to increase IL-1 in the hippocampus. In every case, whether IL-1 was induced by injecting a compound from yeast-cell walls (prompting an immune response), by isolating the rats from their cage-mates after the initial treatment, or by injecting IL-1 directly into the hippocampus, inducing IL-1 prevented animals from remembering the environment and context of their initial shock, but did not prevent them from learning to associate the tone and the electric shock. By blocking IL-1 receptors in the brain, however, Maier's team could prevent any of these techniques from interfering with contextual memory.

Further studies showed that raising IL-1 levels in the hippocampus interfered with the formation of memories if given within an hour, within three hours, and within 24 hours of the initial shock, but had no effect if given 48 hours after the experience. In other words, IL-1 interfered with three important stages in memory formation, characterized by glutamate release, protein synthesis, and neural growth and rearrangement, respectively.

Only one compound, brain-derived neurotrophic factor (BDNF) is known to have clearly defined roles in each of those stages of memory consolidation, said Maier, so he looked to see if IL-1 might affect the expression of growth factors. Social isolation - removing animals from their companions, thus causing stress - reduced the amount of BDNF in the hippocampus. Furthermore, "this effect is blocked by an intrahippocampal IL-1-receptor antagonist," said Maier, which "suggests to us that cytokines within the hippocampus do interfere with memory, but don't do so directly."

Several other cytokines are found in the brain, and it is unclear how their function mimics or opposes IL-1, he said. For that reason, and because cytokine antagonists are short-lived, Maier suspects that a cytokine blocker would not be terribly effective as a way of improving memory or treating disease that impairs memory. However, he and his colleagues are now beginning to test whether gene therapy can stimulate release of an opposing cytokine that shuts down IL-1 action.

Depression, sleep, stress, and cancer chemotherapy all increase cytokine levels in the brain, Maier notes, and all of these situation are associated with memory problems. "So much can be explained through cytokines," agreed William A. Banks of St Louis University School of Medicine, Missouri, who reported at the meeting that cytokines injected into the blood, not into the brain, also impaired learning.

"The theme of cytokines over and over is that they work as a network. It looks like a mess from the outside, but I think when we understand [the cytokine system] better, we can integrate a variety of different things that just seem like puzzles now," said Banks.