20 December 2011. Conventional wisdom says too much stress is a dangerous thing, weakening the heart, immune system, and perhaps the brain. But how bad is it, really? Does stress increase the risk of developing Alzheimer’s disease, or worsen dementia? If so, what kind of stress, how long, how severe? While clear answers are not in, emerging data do link stress and AD more tightly together. For example, epidemiological studies finger stress as a risk factor for cognitive decline. Stress hormones can impair memory and damage the hippocampus over time, and people with AD tend toward higher levels of circulating cortisol, the main human stress hormone. Recent animal data show that psychological distress can worsen the pathology of AD.

“Overall, the evidence is accumulating that chronic adverse stress is bad for the brain during aging,” said Mark Mattson at the National Institute on Aging in Baltimore, Maryland. However, it is uncertain if stress somehow causes decline, and if it does, what the mechanisms might be. “I think the best evidence is for stress and cortisol as exacerbative, rather than causative factors, and it is unclear what the magnitude of the effect is. It seems to me to be important work to follow up on,” Robert Sapolsky at Stanford University, Palo Alto, California, wrote to ARF. One possibility is that stress acts as a “second hit,” pushing vulnerable brains toward dementia faster, suggested Osborne Almeida at the Max Planck Institute of Psychiatry, Munich, Germany. One thing the experts agree on: The relationship between stress and dementia is complex.

As people around the world plunge into the busyness—some say stress—of the pre-holiday countdown, this three-part series offers an overview of the topic. After all, stress is a life factor adults can control, at least in some measure. Part 1 of this series covers human data that connect stress and dementia, and describes a new longitudinal study that may help clarify that link. Part 2 summarizes animal studies, including new papers on tau, that demonstrate stress accelerates neuropathology. Part 3 delves into recent work uncovering specific mechanisms and implicating particular receptors in good and bad effects of stress, as well as possible therapeutic approaches under consideration.

First, the basics: The body’s response to stress is controlled by the hypothalamic-pituitary-adrenal (HPA) axis. In response to external stressors, cells in the hypothalamus release corticotrophin-releasing factor (CRF). This hormone then acts on pituitary cells to stimulate the secretion of adrenocorticotropic hormone (ACTH), which circulates through the blood to the adrenal glands on the kidneys. The adrenal glands then dump glucocorticoids, the primary mediators of stress, into the bloodstream. In humans, the main glucocorticoid is cortisol; in rats, it is corticosterone. Glucocorticoids feed back on the hypothalamus and pituitary to dial down production of CRF and ACTH, and also mediate many of the physiological effects of stress, such as raising blood glucose and depressing the immune system.

In healthy people, cortisol levels rise and fall throughout the day, peaking in the morning and again in the afternoon. They also vary from one day to the next, depending on activities. Glucocorticoids play crucial physiological roles, for example, in maintaining arousal and in memory encoding and consolidation (see, e.g., Roozendaal, 2000; Wilhelm et al., 2011). The steroids bind to two different brain receptors, the glucocorticoid receptor (GR) and the mineralocorticoid receptor (MR), which activate distinct signaling pathways. Being transcription factors, the receptors allow glucocorticoids to turn genes on or off. But glucocorticoids are not the only mediators of stress effects in the brain. For example, CRF itself and its receptors are widely expressed in brain tissue (see Baram and Hatalski, 1998), and may also play a role. The takeoff point for researchers’ interest in stress and AD is the observation that the short-term stress response is healthy and adaptive, but the effects of long-term, chronic stress may be a different story.

Does Stress Make Us Vulnerable to Dementia?
For people, a body of evidence hints that excess stress could contribute to the development of dementia. Numerous epidemiological studies, many from Robert Wilson at Rush University, Chicago, Illinois, indicate that people who are more prone to psychological distress face a higher risk for cognitive decline and AD (see, e.g., ARF related news story on Wilson et al., 2007; Simard et al., 2009; Johansson et al., 2010; and Wilson et al., 2011). Research has also turned up some genetic links between stress and AD. For example, a rare single nucleotide polymorphism that increases expression of 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1), an enzyme that regenerates glucocorticoids and thus amplifies their effects, confers a sixfold increased risk for sporadic AD (see de Quervain et al., 2004 and AlzGene entry). Au contraire, a genetic variant of the glucocorticoid receptor in about 7 percent of the population seems to protect its carriers from the harmful effects of the steroids, and greatly lowers their risk of dementia (see ARF related news story and van Rossum et al., 2008). Cortisol levels are higher in people who carry the ApoE4 AD risk allele than in non-carriers (see Peskind et al., 2001).

Some clues exist as to how stress might harm the brain. Small studies have shown that prolonged cortisol elevation correlates with reduced brain volume in the hippocampus and prefrontal cortex, two areas affected early in AD (see Lupien et al., 1998; Pruessner et al., 2010). Intriguingly, some animal studies appear to tie into this human observation by showing that chronic stress represses hippocampal neurogenesis and makes cells more vulnerable to damage and death, suggesting a possible mechanism behind brain shrinkage (see, e.g., Stein-Behrens et al., 1994; Mirescu and Gould, 2006; and Joëls et al., 2007). Could one culprit behind lessened neurogenesis in stressed brains be a lack of brain-derived neurotrophic factor (BDNF)? Some evidence hints at this. “It’s interesting that many of the neurons that are vulnerable in AD, such as hippocampal CA1 neurons, express high levels of glucocorticoid receptors,” Mattson noted. “Cortisol acting through GR has been shown to reduce expression of BDNF [in hippocampal neurons]. That’s generally thought to be a bad thing, because BDNF has important roles in learning and memory.”

Stress exacerbates other disorders that drive up risk for AD, such as depression (see Ownby et al., 2006; Aznar and Knudsen, 2011), diabetes (see AlzRisk summary; Pavlatou et al., 2008), and metabolic syndrome (see Tamashiro et al., 2011). Diabetic rats and mice maintain higher levels of blood glucocorticoids after a stressful situation than normal animals do, and reducing these levels improves learning and normalizes neurogenesis and synaptic plasticity (see ARF related news story and ARF related news story).

Metabolic syndrome, which is characterized by insulin resistance and abdominal fat, correlates with a higher risk of cognitive problems and AD (see, e.g., Craft, 2007; Whitmer et al., 2007; and García-Lara et al., 2010). A recent study by Philip Landfield’s group at the University of Kentucky, Lexington, showed that monkeys with metabolic syndrome are more sensitive to glucocorticoids than controls are (see Blalock et al., 2010). Monkeys with some degree of metabolic syndrome also dialed down insulin pathway genes and had poorer mitochondrial function and increased neuroinflammation, all of which have been linked to AD.

“We really need to understand the relationship of glucocorticoids to insulin in the brain, and to metabolic syndrome. That’s because in many systems throughout the body, glucocorticoids and insulin have antagonistic roles in energy metabolism. It’s possible that some of the decline in insulin function [in AD] is related to glucocorticoid activity,” Landfield told ARF. Intranasal insulin is currently under investigation as an AD treatment (see ARF related news story).

AD and Cortisol: The Chicken and the Egg
Whether or not stress leads to AD, people with the disease typically have higher levels of cortisol in their blood than healthy people do (see, e.g., Davis et al., 1986; Masugi et al., 1989; Hartmann et al., 1997; Swanwick et al., 1998; Umegaki et al., 2000). Some studies have correlated high plasma cortisol levels and other HPA abnormalities in AD with faster disease progression and worse memory problems (see Weiner et al., 1997; Csernansky et al., 2006; Elgh et al., 2006; Peavy et al., 2011). It is not known how common elevated blood cortisol is in AD. To answer this, Kim Green at the University of California, Irvine, plans to screen about 300 patient samples collected at the UC Irvine Alzheimer’s Disease Research Center. He will correlate cortisol levels with neuropathology data. AD patients with elevated cortisol might be good candidates for trials of anti-glucocorticoid-based therapies, Green noted.

A tough question is, Which comes first, high plasma cortisol or AD? In other words, does a malfunctioning stress system lead to AD, or does the neuropathology of dementia disrupt normal stress physiology? While both could be true, some evidence supports the latter idea. In Green’s studies on triple transgenic AD mice (3xTgAD), pathology preceded abnormalities in the HPA axis (see Green et al., 2006). Green and colleagues propose that glucocorticoid feedback to the hippocampus, which dampens the stress system, may be lost as the hippocampus degenerates in AD. They speculated that increasing glucocorticoid levels may accelerate pathology and further damage the hippocampus, leading to a vicious cycle of disease progression. This hypothesis leaves open the possibility that midlife stress may precipitate AD, Green noted. He believes both processes may occur, perhaps by different mechanisms. One way to address whether high blood cortisol levels increase risk would be to look at people with Cushing’s disease, whose hyperactive pituitary glands elevate their cortisol throughout life, and see if they are more prone to cognitive problems in late life, suggested Jenna Carroll at the University of Pennsylvania, Philadelphia.

Searching for Answers In Longitudinal Data
To clarify the question of midlife stress and dementia risk, Scott Moffat at Wayne State University is making use of data collected by the Baltimore Longitudinal Study of Aging, run by the NIA. He has analyzed the cortisol content in some 6,000 urine samples obtained from about 2,000 people over the course of 20 years, with a few samples going back as far as 40 years. To minimize the effect of daily cortisol fluctuations, each sample was collected over 24 hours. Moffat will correlate urine cortisol with cognitive test results over time, and imaging data in a subset of the group.

Susan Resnick at NIA works on the cognitive and imaging parts of the Baltimore study. She told ARF they have MRI measures of brain volume and positron emission tomography (PET) measures of brain function for about 150 people, as well as amyloid imaging in a smaller group. This study will be one of the largest to look at the relationship among longitudinal cortisol levels, brain imaging, and cognition, Resnick noted. She expects to publish the data next year, and hopes they will shed some light on what effects, if any, midlife cortisol levels have on late-life cognitive outcomes. Resnick points out, however, that cortisol is but one important factor to consider. “The next steps are to understand the effect that cortisol has on other physiological parameters,” she said, for example, on lipids, inflammatory markers, and glucose utilization. “You’re not going to understand cortisol in isolation.” For a look at what the animal data show, see Part 2 of this series.—Madolyn Bowman Rogers.

This is Part 1 of a three-part series. See also Part 2 and Part 3. Download a PDF of the entire series.