Evidence for infiltration of peripheral cells into the brain has been equivocal. Doubts lingered over methodology when studies used irradiation to ablate peripheral immune cells and then replaced them with fluorescently labeled cells to test if they could reach the brain (see ARF related news story). Irradiation can damage the blood-brain barrier and allow cells across. Even in the absence of irradiation, results were at odds. Mouse models of Alzheimer's disease lacking the chemokine receptors that drive migration of peripheral monocytes clear plaque more slowly than usual (see ARF related news story), suggesting that peripheral cells enter the brain. Yet other data suggests that peripheral immune cells are distinct from those in the CNS in AD models (see Mildner et al., 2011).

Notwithstanding the ambiguities, evidence has grown that peripheral cells can sometimes infiltrate the CNS, though they typically avoid a healthy brain. In experimental models of multiple sclerosis, for example, monocytes enter the brain and contribute to degeneration of myelin (see Ajami et al., 2011). At the Venusberg meeting, Richard Ransohoff from the Cleveland Clinic, Ohio, showed how CCR2-positive cells strip myelin from neurons in the CNS (see ARF related news story). Only peripheral monocytes express CCR2, a chemokine receptor that helps guide these cells to sites of damage.

What prompts peripheral cells to enter the brain? Nicholas Varvel, who works at the Hertie Institute, University of Tubingen, Germany, posited that neurodegeneration can be that trigger. Varvel chose a model of epilepsy to study infiltration of circulating monocytes. In his poster, he described how intraperitoneal injection of kainic acid into FVB mice, a strain particularly sensitive to this neurotoxin, induces robust epileptic activity. Kainate also causes rampant neurodegeneration in the CA3 region of the hippocampus within three days. By crossing FVB mice with a CCR2 reporter strain developed in Ransohoff's lab, Varvel tested if this neurodegeneration recruited circulating monocytes.

The CCR2 mice make red fluorescent protein (RFP) driven by the chemokine receptor promoter. In the crosses, Varvel saw recruitment of RFP-positive cells to the CA3 within two to three days of inducing epilepsy. "The finding suggests that not only do the cells migrate into the brain, but that they are also functional when they get there, because they are clustering around sites of neuronal damage," said Varvel. In fact, he was able to show this in a second way. By backcrossing to the FVB strain, the neurodegenerative profile changed to envelop mostly cortical neurons. The CCR2 cells now traveled to the cortex, again targeting sites of damage. "This supports the idea that these monocytes are functional in the brain," said Varvel.

Could peripheral cells similarly enter the brain in Alzheimer's disease? People who have AD are susceptible to epilepsy-like seizures, and many transgenic AD mouse models have seizure activity as well. "That might be difficult to measure," suggested Varvel. He noted that his is a model of acute toxicity, whereas AD is chronic.

If circulating monocytes are entering the brain in response to neuronal death, then to what signals are they reacting? Varvel found that glia in the damaged zones made CCL2, the chemokine that binds the CCR2 receptor. But ultimately, neurons must be sending out signals, probably fractalkine, suggested Varvel. Fractalkine and its cognate CX3CR1 receptor on glia are a primary neuron-glial signaling pathway in the brain.

In peripheral immune cells, other pathways are at work. Mari Shinohara, from Duke University, Durham, North Carolina, reported that inflammasome signaling drives infiltration of circulating immune cells into the brain in experimental autoimmune encephalomyelitis (EAE), a model of multiple sclerosis. Inflammasomes are large complexes that promote caspase 1 cleavage and maturation of interleukin 1β (IL-1β) and IL-18, and together the two cytokines boost expression of chemokines and chemokine receptors that guide monocyte migration. That may be how the inflammasome helps direct peripheral cells into the CNS, said Shinohara.

Previously, researchers in her lab had found that interferon- β (IFN-β), a first-line treatment for MS, suppresses a specific inflammasome that goes by the mouthful "nod-like receptor (NLR) family, pyrin domain-containing 3 type inflammasome," or NLRP3 for short (see Inoue et al., 2012). When researchers in Shinohara’s group ablated NLRP3 in mice, T cells produced fewer chemokines and chemokine receptors than usual, and injections of myelin oligodendrocyte glycoprotein (MOG) failed to induce EAE. Compared to wild-type mice, NLRP3 knockouts had fewer IL-17-producing Th17 cells, which are thought to be particularly inflammatory and promote a variety of autoimmune diseases, including MS. Her group found that NLRP3 helped antigen-presenting cells activate Th17 cells.

Do fewer Th17 cells explain why these mice resist EAE? Not exactly. In Bonn, Shinohara showed that when she injected additional Th17 cells from MOG-immunized NLRP3 knockouts into the blood to replenish cell counts to normal, she still failed to induce EAE. Taking the same number of Th17 cells from MOG-immunized wild-type mice and injecting them into the periphery, however, did restore the immune response that leads to central encephalomyelitis. Those experiments indicated that quality, rather than the quantity, of the cells was crucial for EAE. The key was cell migration. "The Th17 cells activated by NLRP3-negative antigen-presenting cells are pathogenic but unable to migrate into the CNS," said Shinohara. When she injected Th17 cells from immunized NLRP3 knockout mice directly into the CNS, then the animals did develop EAE.

The work not only suggests that the NLRP3 inflammasome may have ramifications for MS treatment strategies. Only half of MS patients respond to IFN-β, suggesting that there may be alternate pathways driving cell migration into the CNS. In fact, Shinohara found that more aggressive immunization with the myelin oligodendrocyte glycoprotein antigen induced EAE even in the NLRP3 knockout. "It seems that there is also an inflammasome-independent pathway that induces EAE and possibly MS," said Shinohara. "If we could understand that other pathway, we may be able to target it for treatment."

Researchers at the meeting felt this was a good example of how inflammation and inflammatory signals in the periphery can lead to consequences in the central nervous system. "We need to develop a better understanding of how systemic inflammation affects the brain," said meeting organizer Michael Heneka, University of Bonn. He recently reported that among patients at an intensive care unit, cognition declined over the next six to 24 months in those who had developed sepsis. Magnetic resonance imaging also showed atrophy of the hippocampus (see Semmler et al., 2013).—Tom Fagan.