8 November 2008. Constant baiting can enrage even the most placid. Is that true also of microglia in the Alzheimer brain? Two papers in this week’s Journal of Neuroscience add to growing evidence that accumulating amyloid-β inflames these otherwise quiescent cells, turning them from helpers to hinderers. Researchers led by Michele Mazzanti at the University of Milan, Italy, show how Aβ peptides induce chloride channel activity in isolated microglia and increase production of reactive oxygen species (ROS), which have been implicated in neurodegeneration. Whether the same sequence of events occurs in vivo is not yet clear, but Mazzanti told ARF that he is studying the effect of knocking down the same chloride channel in an AD mouse model. In the second paper, researchers led by Javier Vitorica at the University of Seville, Spain, report that as Aβ oligomers accumulate in mice, microglia switch from an “alternative,” protective phenotype to a more “classic” pro-inflammatory one. “I think that is very interesting,” said Joseph El Khoury, Harvard Medical School, in an interview with ARF. “This biphasic microglial role is becoming more of an emerging paradigm.”
In fact, El Khoury and colleagues recently reported a similar finding—that as AD mouse models age, microglia become less apt to phagocytose Aβ and more likely to produce pro-inflammatory cytokines (see ARF related news story). “This group has shown that not only do microglia fail to clear Aβ, they also are stimulated by Aβ to become toxic,” said El Khoury.
Vitorica, first author Sebastien Jimenez, and colleagues characterized microglial markers in PS1/APP double transgenic mice (PS1M146L and APP751SL) at six and 18 months (see Blanchard et al., 2003). The younger animals showed little sign of the classic pro-inflammatory phenotype. TNFα, FAS ligand, nitric oxide synthase, cyclooxygenase 2, and most other classic markers were expressed at normal levels as in microglia from wild-type mice. In contrast, YM-1, a marker of the proposed alternative differentiation pathway, was elevated in microglia from six-month-old transgenic animals; YM-1-positive cells were also seen infiltrating and surrounding Aβ plaques. By 18 months, however, the picture had changed. All inflammatory markers tested were significantly higher in microglia from PS1/APP animals, with TNFα expression almost 25-fold higher than in microglia from wild-type animals. Cellular morphology also changed to a more activated form, with obvious cell hypertrophy and shorter, thicker cell processes. Interestingly, YM-1 expression remained, even at 18 months.
The findings suggest that as the animals age, pro-inflammatory microglia emerge while alternatively differentiated microglia also remain in the picture. In support of this, Jimenez and colleagues, using immunohistochemistry, found that most microglia surrounding plaques in 18-month-old transgenic mice are negative for TNFα, while in plaque-free areas they had a pro-inflammatory cytokine profile. What this means is not clear, but infiltration of peripheral microglia into the brain has been seen in mouse models and may help explain the different phenotypes observed (see ARF related news story).
What could trigger the phenotypic switch? For one, the authors considered the role of infiltrating T cells. Though they did find an increase in T cell markers in the brain of 18-month-old transgenic mice, their profile (higher IL-10, TGFβ1, IL-4) indicates anti-inflammatory rather than pro-inflammatory signaling. Next, they considered Aβ itself, particularly soluble Aβ (see ARF related news story). Though six-month-old transgenic animals have abundant Aβ deposits, the researchers found only very low levels of the soluble peptide in these animals. By 18 months, however, when the pro-inflammatory switch has been flicked, soluble Aβ had jumped 15-fold, as had soluble Aβ aggregates as measured with the monoclonal antibody Nu-1 (see Lambert et al., 2007). Western blots also indicated the emergence of oligomers at 18 months, particularly 6-mers and 12-mers.
To test if Aβ oligomers could indeed affect microglial phenotype, Jimenez and colleagues incubated mixed astro-microglial cultures with monomeric and oligomeric preparations of Aβ. They found a dose-dependent increase in TNFα expression in response to oligomeric, but not monomeric samples. Soluble extracts from 18-month-old (but not six-month-old) PS/APP mice had the same TNFα-inducing effect, which could be eliminated by first depleting extracts of Aβ using antibodies (6E10 or A-11, which recognizes oligomers). Together, the data suggest that soluble oligomers of Aβ can induce a change in microglial gene expression.
Do these findings relate to the physiology or treatment of Alzheimer disease? The authors suggest that the alternate microglial phenotype might be protective, while the activated, pro-inflammatory microglia are cytotoxic. This idea has been around for some time. “If you found a target that could stop the neurotoxic effect but not the phagocytic effect, that would be terrific,” said El Khoury. An alternative approach, recently reported by a different group, may be to block TGFβ signaling pathways in peripheral cells. That appears to enhance recruitment of amyloid-eating immune cells into the brain (see ARF related news story). “That would be a nice way to do it, if you could enhance recruitment of cells that clear Aβ, while at the same time blocking the toxic effect,” said El Koury.
In the second paper, Mazzanti and colleagues show how Aβ can induce microglia to produce ROS. They previously reported that Aβ increases expression of the chloride intracellular channel 1 (CLIC1) in microglia (see Novarino et al., 2004). “CLIC1 is a protein that normally stays in the cytoplasm, but we found that upon stimulation with Aβ it moves to membrane and increases the permeability of chloride ions,” said Mazzanti. That in turn leads to activation of the plasma membrane NADPH oxidase, which generates ROS. “The good thing is, if we downregulate the channel current we are able to decrease the production of ROS,” said Mazzanti.
First author Rosemary Milton, University College London, and colleagues discovered the CLIC1/ROS connection when they treated cells from the BV2 immortalized microglia line with Aβ25-35 peptides. Mazzanti told ARF that the scientists used these peptides because they were more stable and reliable in their hands, but had since obtained the same results with Aβ1-42. Treating the cells with Aβ25-35 led to an increased electrical current across the cell membrane. A CLIC1 inhibitor blocked half of it, and a generic chloride channel blocker eliminated another 25 percent. Using a green fluorescent protein-tagged CLIC1 construct, the researchers were able to show that Aβ induced a redistribution of the chloride channel from the cytoplasm to the cell membrane: CLIC1 appears to be one of those rare membrane proteins that is synthesized as a soluble protein and then inserted into membranes post-translationally (see Tulk et al., 2002). After fixing the cells, CLIC1 antibodies confirmed the presence of CLIC1 in the cell membrane and a FLAG epitope in the CLIC1 N-terminal showed that it spanned the membrane.
The authors wondered if Aβ, CLIC1 translocation, and ROS production are related. The reasons are that Aβ has been shown to induce the production of ROS by NADPH oxidase in microglia (see McDonald et al., 1997), and that this oxidase, which pumps electrons across the cell membrane to the extracellular space to generate superoxide, requires a compensatory transfer of ions to prevent the cell from becoming depolarized. Milton and colleagues found that Aβ25-35 and Aβ1-42 (but not Aβ35-25) increased ROS production in BV2 cells. The CLIC1 blocker IAA94, a CLIC1 antibody, or silencing the CLIC1 gene using siRNA all counteracted this effect.
El Khoury commented that this discovery of a new protein involved in Aβ toxicity is interesting, but questioned how it will pan out in vivo. Mazzanti agreed. “This is just an in vitro study. We have already started to work with a mouse model in which we hope to show that blocking CLIC1 is in some way neuroprotective,” he said.—Tom Fagan.
Jimenez S, Baglietto-Vargas D, Caballero C, Moreno-Gonzalez I, Torres M, Sanchez-Varo R, Ruano D, Vizuete M, Gutierrez A, Vitorica J. Inflammatory response in the hippocampus of PS1m146L/APP751SL mouse model of Alzheimer’s disease: age-dependent switch in the microglial phenotype from alternative to classic. J. Neurosci. 2008 November 5; 28:11650-11661. Abstract
Milton RH, Abeti R, Averaimo S, DeBiasi S, Vitellaro L, Jiang L, Curmi PMG, Breit SN, Duchen MR, Mazzanti M. CLIC1 function is required for beta-amyloid-induced generation of reactive oxygen species by microglia. J. Neurosci. 2008 November 5; 28:11488-11499. Abstract