Exercise drives up expression of a brain chemical that helps learning and memory, and people with Alzheimer’s and other neurodegenerative diseases have less of it. Researchers reported in the October 10 Cell Metabolism that exercise-induced surges in this molecule, brain-derived neurotrophic factor (BDNF), are activated by the same molecular pathway that stimulates fat metabolism in muscle. Furthermore, the scientists, led by Bruce Spiegelman and Michael Greenberg of Harvard Medical School, showed they could increase BDNF in the brains of mice by manipulating the metabolic pathway in the liver. Though it remains unclear if the same can be done in people, the findings suggest a way to boost BDNF in the brain without injecting it with drugs, cells, or viruses. Broadly speaking, the study outlines how physical activity helps the brain. Other scientists found the data intriguing. “This paper is exciting and seems to be a first step toward understanding why exercise is advantageous not only for physical but also mental health,” wrote Volkmar Lessmann, Otto-von-Guericke University, Magdeburg, Germany, in an email to Alzforum (see his full comment below).
Rather than hunting directly for molecules that upregulate BDNF, Spiegelman and colleagues came at the mechanism from the exercise angle. Over the last few years, the researchers figured out how physical exertion helps white fat cells burn energy. Compared with sedentary animals, mice that exercised on a running wheel expressed more of a transcriptional co-activator in their leg muscles. Called PPARγ co-activator 1α (PGC-1α), this factor upregulates FNDC5, a membrane protein that is cleaved and secreted by white fat cells into the blood as the hormone irisin. The peptide activates proteins that rev up metabolism in muscle (ARF news story on Boström et al., 2012). The scientists were curious whether exercise invokes a similar pathway in the brain. If it does, “maybe we can manipulate it to prolong the life of neurons and preserve their connections to prevent dementia,” first author Christiane Wrann told Alzforum.
To test that idea, she and colleagues analyzed brain tissue from sedentary mice and from animals given access to a running wheel. After 30 days of scampering about five kilometers per day, the latter group of mice were noticeably quicker and leaner, and expressed more FNDC5 mRNA in their hippocampus, Wrann said. Consistent with their 2012 study in muscle, the researchers showed that hippocampal FNDC5 correlated well with PGC-1α mRNA levels in the brains of mice from birth through 30 days of age. In cell culture experiments, inducing PGC-1α expression with adenoviruses or knocking it down with short-hairpin RNAs produced corresponding increases and decreases in FNDC5 transcript levels, suggesting that PGC-1α activates FNDC5 in primary cortical neurons.
Having connected exercise with PGC-1α upregulation and FNDC5 activation in the hippocampus, and given the wealth of evidence for exercise boosting BDNF in the brain (Zuccato and Cattaneo, 2009; Murray et al., 1994; ARF conference story), Wrann and colleagues wondered if FNDC5 turns on BDNF. Forcing FNDC5 expression in mouse cortical neurons increased irisin secretion and, sure enough, generated more BDNF mRNA and protein. Conversely, silencing FNDC5 sent BDNF levels plummeting. The findings suggest that FNDC5 works upstream of BDNF to activate its expression.
The most important experiment was the one showing that peripheral application of FNDC5 induced the same brain changes as exercise, Wrann said. Using adenoviruses to express FNDC5 in the liver, the researchers looked for BDNF changes in the hippocampus a week later. Not only were BDNF transcript levels increased in the treated mice compared to controls, but hippocampal expression of other neuroprotective factors (e.g. cFos, Arc, Zif268) rose as well. The researchers did not test whether irisin reached the brain, though they did find that overexpressing FNDC5 in the liver drove the hormone up in the blood. “It could be that irisin induces another molecule that then crosses the blood-brain barrier and induces BDNF changes,” Wrann said. “This needs to be explored in more detail.” In unpublished experiments, Wrann has detected FNDC5 mRNA in primary human neurons and neuronal cell lines.
“It makes sense that a molecule involved in energy generation and fat metabolism would go in and signal to the brain," said Nicole Berchtold of the University of California, Irvine, who worked there with Carl Cotman to show that running increases BDNF expression and neurogenesis in the hippocampus (see ARF news story). Trying to influence brain BDNF through peripheral manipulations might prove effective “because it mimics the way BDNF is normally regulated or supplied in the brain,” Berchtold noted. However, she cautioned that the strategy may not work if the system for regulating BDNF malfunctions, as it may in aging or AD.
In mouse and primate models of AD and aging, infusions of BDNF have been shown to curb synapse loss and restore cognitive deficits (see ARF news story; ARF news story). However, these studies supplied BDNF directly into the brain through injections or stem cells, and such gene delivery methods have proven problematic in human trials of other neuronal growth factors (see ARF news story).
Though it is not practical to measure brain BDNF levels in living people—for instance, before and after exercise—functional magnetic resonance imaging (fMRI) studies show possible correlations between brain function and serum BDNF (see ARF news story on Erickson et al., 2010; Voss et al., 2012). A recent study that measured circulating BDNF in adults who used an indoor rower suggests that plasma levels of the trophin increase two- to threefold during exercise, and that the brain contributes 70–80 percent of circulating BDNF (Rasmussen et al., 2009). Berchtold said her team has unpublished postmortem microarray data showing increased expression of BDNF and other plasticity genes in the brains of active seniors relative to those who rarely exercised.
For their part, Wrann and colleagues are trying to develop a stable form of irisin that can be injected intravenously into mice. If such molecules can improve memory and synaptic plasticity, they may consider testing them in disease models, she said.—Esther Landhuis