The two best-known signs of Alzheimer’s, in the brains of its victims, are the plaques of amyloid beta protein and tangles of tau protein. But the disease also features chronic inflammation. Cells known as microglia—neural cousins of pathogen-eating macrophages of the bloodstream—swarm around amyloid plaques and dying, tangle-ridden neurons. They seem helpful, gobbling up amyloid beta as well as disease-damaged cells. But does their immunological enthusiasm also cause harm to healthy cells—could it accelerate the disease or even help to initiate it? Scientists have debated these questions for more than two decades, without any firm resolution. Now a burst of new research suggests that inflammation does, indeed, play a major role in Alzheimer’s—and that targeting specific elements of that inflammation could be useful in treating or preventing the disease.
Friend or foe?
Wherever it occurs in the body, chronic inflammation is a double-edged sword. The initial inflammatory response is meant to defend tissues against molecular foes such as viruses, cancerous cells, and harmful amyloid protein aggregates. But the longer it lasts, the more this inflammation stresses and kills healthy, “innocent bystander” cells. Over time—as in rheumatoid arthritis, for example—the inflammation can become self-sustaining.
Since the late 1980s, various studies have found hints that the chronic inflammation found in Alzheimer’s hastens the disease process, and may even be a disease trigger. A history of serious head injury, which typically causes brain inflammation, is known to be a risk factor for Alzheimer’s. Systemic infection—another cause of inflammation—also appears to accelerate the disease. Several epidemiological studies have found that older people who use anti-inflammatory drugs regularly appear to have significantly lower incidences of Alzheimer’s.
The value of those epidemiological studies came into question several years ago, when more rigorous placebo-controlled clinical trials of anti-inflammatory drugs—ibuprofen, naproxen, and celecoxib, for example—failed to show signs of helping people who already have Alzheimer’s dementia or early cognitive impairment. In some cases these drugs apparently accelerated the course of the disease. Yet in a little-publicized study, published in late 2011, naproxen seemed to have a marked effect in preventing the disease: It reduced the incidence of Alzheimer’s among elderly people who started out cognitively normal and took the drug for more than two years.
“We really need to understand the [Alzheimer’s] inflammation reaction better—what is good and what is bad,” says Irene Knuesel, a researcher at the University of Zurich. Her laboratory and others have been discovering important clues about which is which.
Stop the bad, boost the good
Frank Heppner, Burkhard Becher, and their laboratories in Germany and Switzerland reported in November 2012 on experiments with two closely related inflammation-promoting proteins, IL-12 and IL-23. The proteins are among those pumped out by microglia when the cells become immunologically active, and the researchers found signs that the proteins exist at elevated levels in the cerebrospinal fluid of people with Alzheimer’s. Blocking the two proteins in young “Alzheimer’s mice”—in this case a two-mutant-gene model that is normally quite resistant to amyloid-reducing therapies—prevented most of the usual buildup of amyloid plaques. Blocking these inflammatory proteins in older Alzheimer’s mice, whose brains were already plaque-ridden, reduced soluble, more toxic forms of amyloid beta, and reversed the mice’s cognitive deficits.
A therapy, ustekinumab, that simultaneously blocks IL-12 and IL-23 (by blocking a molecular subunit found in each) is already FDA-approved for treating psoriasis. Heppner, whose laboratory is at Charité teaching hospital in Berlin, would like to understand better how the IL-12/23-blocking treatment really works against the disease, and improve upon the strategy if possible. He suspects that microglial-produced IL-12 and IL-23 may stem the activity of astrocytes, helper cells that also have a role in neuroinflammation and the clearance of amyloid beta. In any case, he says, given the availability of an approved and apparently well-tolerated IL-12/23-blocking drug, “I think it’s not unfeasible to even go directly into patients [for clinical trials] right now.”
Researchers have recently described three other anti-inflammatory approaches that also seem to work well in the same Alzheimer’s mouse model—by reducing brain inflammation, amyloid beta deposition, and cognitive impairments.
One of these approaches, funded in part by the Dana Foundation, involves the blocking of a protein called NLRP3. This molecular “switch” within microglia detects abnormal, disease-associated molecular structures, including those found on amyloid beta aggregates. When it finds them, it helps to shift microglia to a state in which they release pro-inflammatory compounds as well as cell-harming hydrogen peroxide and other reactive oxygen molecules. In this study, the researchers found signs that NLRP3 is abnormally activated in the brains of people with Alzheimer’s and with the related disorder called Mild Cognitive Impairment. In Alzheimer’s mice that were genetically engineered to lack NLRP3, microglia were shifted back towards a more normal, non-inflammatory, “housekeeping” state, in which they consume much more amyloid beta and secrete neuron-nourishing proteins.
“I think that NLRP3 activation is a key step in innate immune activation in the brain, and we’re now screening for inhibitors of that pathway by using libraries of known drugs and new peptides,” says Michael T. Heneka of the University of Bonn, who was a senior author of the NLRP3 study and the two others.
In the second study, Heneka and his colleagues targeted a microglial protein called MRP14, which also helps to shift microglia to an inflammatory state. In the third, they used compounds to activate both the PPAR-γ and RXR receptors on microglia, which respectively reduce inflammatory signaling and increase the microglial uptake of amyloid beta. There are already FDA-approved drugs that activate PPAR-γ and RXR; the diabetes drug pioglitazone activates PPAR- γ, while the skin cancer drug bexarotene activates RXR. (Bexarotene is also being investigated as a booster of the amyloid-beta clearing protein apo-E.) In principle, the fact that both drugs already have been proved reasonably safe in humans eliminates one hurdle to their testing in Alzheimer's cases.
The new inflammation hypothesis
How each of these inflammation pathways relates to the Alzheimer’s disease process is not yet clear. But in the MRP14 study, Heneka and his colleagues found that the protein doesn’t just push microglia into an activated, inflammatory state. It also somehow increases the presence and activity of a neuronal enzyme, BACE1, that helps to produce amyloid beta, and thus increases amyloid beta production. Heneka suspects that inflammation helps to start the disease process by boosting the production of amyloid beta—and then helps to sustain the process by reducing the ability of microglia to remove amyloid beta. A key point is that amyloid beta accumulation seems to be both a cause and an effect of chronic inflammation.
“Usually inflammation is there to limit the invasion of foreign bacteria or viruses or parasites, and once these are removed, then the inflammation is resolved by anti-inflammatory mechanisms,” Heneka says. “But in the [Alzheimer’s] brain, there is the constant detection of the ongoing amyloid beta deposition, which does not allow the inflammation to resolve.”
Inflammation as an actor
The idea that inflammation doesn’t just add to Alzheimer’s, but also helps to get it started, has been supported by the recent development, in Knuesel’s laboratory, of mice that seem to mimic the process. Unlike other Alzheimer’s-model mice, Knuesel’s have not been genetically engineered to produce excess amyloid beta or mutant tau. They are normal lab mice that have been injected twice—once in the womb and later in adulthood—with virus-mimicking molecules that cause chronic neuroinflammation.
Analyses of the brain changes in these mice, and comparisons to brain tissue from people with Alzheimer’s, suggest to her that inflammation may help trigger Alzheimer’s by exacerbating a common age-related problem with neurons. As they get older and their functions become less efficient, neurons lose their ability to transport and properly dispose of proteins. [See recent Dana briefing paper.] This decline is apt to show up first in neurons’ output stalks, or axons, which, being long and thin, are particularly vulnerable to a disruption of their internal transport systems. If enough proteinaceous waste builds up in an axon, it will swell—a feature often seen in Alzheimer’s brains—and may try to bubble off a waste-filled granule, which will then be consumed by nearby microglia.
Inflammation worsens this problem in several ways: First, it increases the production of amyloid-beta in inflamed regions. Second, it stresses neurons and hastens the age-related decline of their protein-transport and disposal systems. Third, it pushes microglia into a reactive, inflammatory state and thus reduces their ability to clear up axons’ expelled waste.
Eventually this proteinaceous waste builds up outside damaged axons. Much of it will consist of amyloid beta and its aggregates. Some of these aggregates will clump together in plaques, and all of them will further inflame microglia, leading into a vicious cycle of amyloid beta deposition, inflammation, and neuronal damage. Ultimately the damaged axons will degenerate, their downstream synaptic terminals will die, the rest of their neuronal bodies will die, and the neuronal networks to which they belonged will stop functioning properly. In some cases, especially familial, early-onset Alzheimer’s, excess production or aggregation of amyloid beta might be the main initiator of the process – but for ordinary late-onset Alzheimer’s, aging and inflammation might be the most common triggers.
Conversely, says Knuesel, a healthy neural immune system, with properly working microglia, may help to ensure a gradual, healthy brain aging. “We have to accept that as we age a certain synaptic loss will occur, and some neurons will die,” she says. “The big question is: How can we prevent the complete collapse?”