Blocking the signaling of TGF-β in peripheral macrophages allowed those immune-system scavengers to enter the brain and clear amyloid deposits, mitigating some symptoms in an AD mouse model, according to an advance online report in Nature Medicine. The study, by senior author Richard Flavell of Yale University School of Medicine in New Haven, Connecticut, suggests that macrophages can enter the brain in significant numbers and directly attack amyloid plaques. The finding holds out the promise of peripheral therapy, which would circumvent the troublesome requirement of broaching the blood-brain barrier. But some commentators said more work needs to be done to prove definitively that the cells responsible for the plaque clearance did indeed come from the bloodstream and that they could be deployed safely in the brain.
Flavell and his team were expecting quite different results when they devised the protocol for this study. First author Terrence Town of Cedars-Sinai Medical Center in Los Angeles, California, said he thought the interruption of TGF-β signaling would provoke an inflammatory response. That effect had already been reported in microglia, immune cells that reside in the brain (Brionne et al., 2003). Town used the CD11c-DNR mouse, in which the TGF-β signaling pathway is blocked only in macrophages, dendritic cells, and natural-killer cells. The mouse was made in Flavell’s laboratory, which specializes in studies of T cell activation and differentiation and autoimmune disease. By crossing the TGF-β-blocked mice with APP-transgenic mice, Town hoped to explore whether and how TGF-β signaling in those peripheral cells contributes to AD. “We thought we would get mice that were just full of inflammation, and the disease would get worse,” Town said. “We were utterly surprised when we saw the opposite.”
What Town and his colleagues report instead is no evidence of an inflammatory response and a seven- to eightfold increase in the percentage of peripheral macrophages infiltrating the brains of the double-transgenic mice, as tagged by the CD45, CD11b, and CD11c markers and other measures. They also saw a reduction in plaque and soluble amyloid-β of up to 90 percent in both cerebrovascular and parenchymal tissue. “The amyloid load was strikingly reduced, not only histologically, but also biochemically,” said Town.
“These data fit very well with the novel concept that systemic innate immune cells have the capacity to fight against toxic proteins, but they do not do it in an efficient manner,” wrote Serge Rivest of Laval University in Quebec City in an e-mail. Rivest had previously suggested that macrophages could be instrumental in attacking plaque in the brain. “Improving both the infiltration and immune properties of these cells will hopefully soon be a very effective new therapy to cure AD.” (See extended comment below.)
Yet in behavioral tests, the intervention had only a modest effect on the spatial learning and memory deficits in mice that most closely resemble AD symptoms in people. There was no appreciable difference in performance in the Morris water maze, for example, and only a modest improvement in spontaneous Y-maze alternation, a measure of spatial working memory. Town and his colleagues did see a complete mitigation of the hyperactivity associated with the Tg2576 phenotype, which is thought to be brought on by cortical and/or hippocampal injury in these Alzheimer’s mice. But that hyperactivity has no established human correlate in AD.
Joseph El Khoury, for one, is not disturbed by the lackluster behavioral results. “The AD models in mice aren’t perfect,” said El Khoury, of Massachusetts General Hospital in Boston. “We should focus on the significant amelioration of pathology, rather than the behavioral effects.”
More questions have come regarding the authors’ assertion that peripheral macrophages are responsible for the plaque clearance. “We are not convinced that CD11c is a marker that can distinguish microglia from invading peripheral cells,” noted Mathias Jucker of the University of Tuebingen in an e-mail. Tony Wyss-Coray of Stanford University School of Medicine in Palo Alto, California, whose work Town credits as his inspiration, said he likewise lost faith in that marker’s ability to identify peripheral cells after reading Karen Bulloch’s recent study about CD11c cells (Bulloch et al., 2008). Furthermore, Wyss-Coray’s work shows that TGF-β has a neuroprotective effect mediated by the microglia (Wyss-Coray et al., 2001; Tesseur et al., 2006). “That doesn’t diminish what they found at all,” said Wyss-Coray, because TGF-β, like other cytokines, can have different effects on different cell types. “And whether these cells came recently from the blood or were already in the brain, I don’t care. They have a very striking effect on the pathology.”
But the questions of the provenance and brain-homing capacity of these cells do have implications for treatment in humans, where researchers have struggled to find AD therapies that can penetrate the blood-brain barrier. Town said that three different lines of evidence implicate macrophages rather than microglia in this study. In addition to the panel of immune markers, he ran tests showing that the microglia did not have the TGF-β signaling blockade, while the peripheral macrophages did. And in brain-tissue sections, the CD45+ cells appeared smooth and round, quite distinct from the process-ridden, irregularly shaped microglia, said Town. He and his colleagues saw these round cells in the brains’ blood vessels, escaping the blood vessels, surrounding amyloid deposits, and engulfing them. “These peripheral macrophages were entering the brain and devouring the plaque,” Town said. “We caught them in the act.”
“It's pretty convincing that the cells are coming from the blood,” El Khoury agreed. He suggested parabiosis experiments of the kind performed previously (see Ajame et al., 2007; Mildner et al., 2007) to demonstrate conclusively whether the relevant factors were peripheral. To El Khoury, the big advance here is demonstrating that the immune system can be revved up to take apart amyloid deposits without simultaneously promoting potentially damaging inflammation. In the paper’s supplementary material, fluorescence-activated cell sorting characterization indicates that the vast majority of macrophages that entered the brain are of the anti-inflammatory subtype Ly-6C–. “The question has been, how can you increase the phagocytosis and clearance of plaque without causing all this damage?” said El Khoury. “That’s what they did in this really impressive paper.”
Pritam Das of the Mayo Graduate School of Medicine in Jacksonville, Florida, is also persuaded that the cells are peripheral macrophages. In an e-mail he called the findings “fascinating” but cautioned that the effects on neurodegeneration need to be explored (see extended comment below).
“Certainly, one of the concerns of having long-term peripheral infiltration of macrophages (regardless of their phenotype), is unwanted reactions on other cell types/tissues in the brain,” Das wrote.
For his part, Flavell agreed that any clinical application of this technique would need to be approached with caution, but he also noted that antibodies to TGF-β have already been demonstrated to be safe in Phase 1 clinical trials for fibrotic disease.—Karen Wright
Karen Wright is a freelance writer in New Hampshire, U.S.