A New Class of Culprit in Neurodegeneration


by Jim Schnabel

July 3, 2012

Perhaps the biggest surprise to emerge from recent studies of the human genome is that hardly any of our DNA—less than 1.5 percent—contains codes for the making of proteins. A far larger portion, as much as one-quarter of the human genome, is transcribed instead into strips of RNA that function without ever being translated into proteins. Scientists are just starting to discover what these “non-coding” RNAs (ncRNAs) do within cells, but already it seems very likely that disruptions of their functions contribute to major diseases, including neurodegenerative diseases.

“I think that we’ve seen just the tip of the iceberg so far, with respect to the role of these non-coding RNAs in neurodegenerative diseases,” says Maral Mouradian, a neurologist at the Robert Wood Johnson Medical School in new Brunswick, New Jersey.

“We’re starting to look at these things and realize that they’re performing all kinds of functions that hadn’t been noticed before,” says Rory Johnson, a genomics expert at the Centre for Genomic Regulation in Barcelona, Spain.

Non-coding RNAs in Alzheimer’s

Many non-coding RNAs are known as microRNAs (miRs), because of their short length of about 22 nucleotides. They work in what is called the RNA interference pathway, mostly serving as tiny dimmer switches, dialing down the influence of certain genes. They accomplish this feat by binding to their target genes’ messenger-RNA transcripts and triggering the cleavage of the transcripts by enzymes or otherwise preventing the gene transcripts from being translated into proteins.

As a postdoctoral student at the Catholic University of Leuven in Belgium several years ago, Sébastien Hébert examined brain tissue from a small group of recently-deceased elderly people with common, late-onset, “sporadic” Alzheimer’s. He and his colleagues reported that, compared with the brains of age-matched people without Alzheimer’s, the Alzheimer’s brains had lower levels of various miRs—including three that, Hébert showed, could regulate (dial down) the production of the enzyme BACE1. This enzyme is necessary for the release of Alzheimer’s-associated amyloid beta protein from its larger precursor protein. When BACE1 is overproduced, amyloid beta is overproduced too—and thus becomes more likely to aggregate and cause disease. Hébert found that the Alzheimer’s brains with abnormally high BACE1 levels had abnormally low levels of these BACE1-regulating miRs—and this relationship between BACE1 and the miRs showed up in neuronal cell cultures too.

“It was known before we published the study that BACE1 is increased in at least some sporadic Alzheimer’s patients, but the reason was unknown,” says Hébert, now a researcher at the University of Laval in Quebec. “So this result of ours helps to fill in the explanatory gap a bit.”

Two of the three miRs belong to the miR-29 family, whose functions also include the dialing down of gene transcripts that trigger neuronal apoptosis—a programmed cellular suicide, often triggered by exposure to toxins, inflammation, or other stresses. “If you had less of this miR in the brain, you would be more sensitive to all toxic stimuli, and thus could be more susceptible to Alzheimer’s; that’s one of the hypotheses that I’m working on,” Hébert says.

Non-coding RNAs in Parkinson’s

Several studies so far have pointed to miRs as possible contributors to Parkinson’s disease. Mouradian’s laboratory has been one of the most active in this area. “There’s a lot of complexity that needs to be better understood, and we’ve only been studying this for the past four or five years,” she says. “But one microRNA seems to be consistently important in downregulating the expression of alpha synuclein”—the protein whose amyloid aggregates have long been linked to Parkinson’s.

That microRNA is miR-7, which Mouradian and her colleagues recently found to be a major regulator for alpha synuclein production in neurons. When Mouradian’s team added miR-7 to cell cultures, alpha synuclein levels dropped and the cells showed fewer signs of the oxygen free-radical damage (“oxidative stress”) that alpha synuclein is known to induce. By contrast, when the researchers used an inhibitor of miR-7, alpha synuclein levels rose.

So far, no study has shown that miR-7 levels are diminished in Parkinson’s brains. However, miR-7 slows alpha-synuclein production by attaching to a major regulatory region of the alpha synuclein messenger RNA known as the 3’ untranslated region (3’-UTR). Genome-wide association studies already have linked Parkinson’s to genetic variations in the corresponding region of the alpha synuclein gene. These variations may inhibit miR-7, and perhaps other miRs, from binding and performing their dimmer-switch functions—allowing alpha synuclein levels to rise perilously even when the relevant miR levels seem normal. Even moderate increases in alpha synuclein levels due to an extra copy of the alpha synuclein gene lead to Parkinson’s—in an accelerated, early-onset form.

Mouradian emphasizes that years of work still need to be done to tie miR-7 or other ncRNAs to Parkinson’s, although the evidence so far—and the newly revealed importance of miRs generally—strongly suggests that there are critical links. “I suspect that the amount of alpha synuclein in certain vulnerable neuron types has to be regulated tightly during our lifetimes, and the miR system is a very important part of that regulation, so that variations in the miRs themselves or in their UTR binding sites could have important impacts on disease risk, at the very least,” she says.

Non-coding RNAs in Huntington’s disease

Huntington’s disease differs from Alzheimer’s and Parkinson’s in many ways, yet its paths of neurodegeneration also appear to involve ncRNAs.

Huntington’s is purely an inherited disorder, whose gene mutations result in an abnormal, elongated form of the huntingtin protein. Mutant huntingtin can form amyloid aggregates that are similar to those seen in Alzheimer’s and Parkinson’s. Just how huntingtin aggregates are toxic to brain cells isn’t yet clear. But the mutation appears, among other things, to set off a molecular cascade resulting in the shut-down of vital neuronal genes—and there is evidence that stretches of the genome encoding important ncRNAs are shut down too.

One of the jobs of normal huntingtin is to bind to a protein called REST and thus keep REST out of the nucleus of neurons. REST is a transcription factor that, when it does get into a cell nucleus, represses the activity of important neuronal genes; it is supposed to be active only in non-neuronal cells, where it prevents them from behaving like neurons. Mutant and/or aggregated huntingtin apparently does a poor job at keeping REST out of the nucleus in neurons, with the result, says Rory Johnson, that “REST gets into the nucleus and switches off key neuronal genes—and that probably causes a lot of the neurodegeneration.”

The crucial genes that REST shuts off inappropriately include the gene for BDNF, a growth factor that is known to help brain cells stay healthy. But as Johnson and his colleagues have been finding over the past few years, the production of regulatory microRNAs such as miR-9 and miR-132 is also reduced in Huntington’s. Moreover, in studies published in 2010 and 2012, he and his colleagues sifted through genomic data from prior studies of Huntington’s brain tissue, and found evidence that in Huntington’s neurons, potentially even more important ncRNAs, known as “long” non-coding RNAs, are suppressed—and some of these contain specific genomic binding sites for REST, which suggests that REST directly switches them off.

The functions of long ncRNAs are far from being completely understood. But Johnson and his colleagues believe that studying ncRNA-related pathways in Huntington’s is likely to reveal much about how the disease kills neurons and why certain neuronal types are so vulnerable. In principle, resupplying a mix of ncRNAs that are suppressed in Huntington’s—via an artificial, cell-infecting virus, as in gene therapies—could even treat the disease. Certainly researchers in other areas of medicine are now racing to develop ncRNA-based therapies.

Johnson also suspects that this long-ignored layer of cell biology could help researchers understand the differences between the brains of humans and the brains of other animals. Such differences have deep scientific importance, and also practical importance: they tend to make animal models of human brain diseases far less useful than researchers would like. Long non-coding RNAs, for example, are often expressed in the human brain in high amounts and yet are not found at all in the brains of mice. “A popular idea at the moment is that the evolution of these long non-coding RNAs has to some extent been involved with, or has even been necessary for, the evolution of human cognition,” says Johnson.