# Why Stress Is Bad for Your Brain

> An annotated edition of this paper on Fermat's Library, where scientists and readers explain its ideas, equations and history in the margins. Annotated edition: https://fermatslibrary.com/s/why-stress-is-bad-for-your-brain

- **Authors:** Robert M. Sapolsky
- **Published:** 1996
- **Topic:** Biology
- **Original:** https://s3-us-west-1.amazonaws.com/fermatslibrary/uploads%2F1623006035772-zqinjre2ok-7248013db8f03d094d26a07ee379d189%2Fstress_brain_sapolsky1996.pdf
- **Pages:** 2
- **Annotations:** 9 comments in 9 threads
- **Featured:** Fermat's Library Journal Club
- **Annotated edition:** https://fermatslibrary.com/s/why-stress-is-bad-for-your-brain

## Annotations

### Annotation 1 (page 1)

Beside the passage: “Why Stress Is Bad for Your Brain”

**Annotation** by Micael Oliveira (2021-06-06):

> #### TL;DR
>
> > ***"But if you get chronically, psychosocially stressed, you're going to compromise your health. So, essentially, we've evolved to be smart enough to make ourselves sick."***
>
> Stress is defined as existing or anticipated threat of well-being. Our bodies' stress response evolved to help us get out of short-term physical emergencies - like if a lion is chasing you, you run. Such reactions compromise long-term physical health in favor of immediate self-preservation.
>
> Unfortunately, when confronted with non-life-threatening stressors, such as troubleshooting the internet connection worrying about money etc, modern humans turn on the same stress response. Modern lifestyles represent challenges that render individuals susceptible to physical and mental disorders.
>
> An individual's vulnerability and resilience may be determined by both genetic and epigenetic factors which influence our ability to cope with stress - ex: gender, early life experiences.
>
> The most easily measureable and critical physiological response to stress involves the release of glucocorticoids (GCs). GCs can exert profound modulatory effects on a variety of brain functions from early development through to late life.
>
> The hypothalamus, pituitary, and adrenal glands play an essential role in the adaptive response to psychogenic (ex: fear) and physical stressors (ex: cellular lesion or pathogen invasion).
>
> In this paper Sapolsky points out to the effects that prolonged exposure to GCs could have on the hypothalamus - ***"If you turn your stress response on for too long you get sick." ***
>
>
> > ***"Furthermore, if you're chronically stressed, all sorts of aspects of brain function are impaired, including, at an extreme, making it harder for some neurons to survive neurological insults," [...] "Also, neurons in the parts of the brain relating to learning, memory and judgment don't function as well under stress. That particular piece is what my lab has spent the last 20 years on."***

### Annotation 2 (page 1)

Beside the passage: “ndrome 1”

**Annotation** by Micael Oliveira (2021-06-06):

> Robert M. Sapolsky is an American neuroendocrinology researcher and the recipient of a MacArthur Foundation genius grant. He is currently a professor of biology, and professor of neurology and neurological sciences and, by courtesy, neurosurgery, at Stanford University. He is also the author of several works of nonfiction such as "Why Zebras Don't Get Ulcers", "Behave", "A Primate's Memoir".
>
> !["sapolsky"](https://news-media.stanford.edu/wp-content/uploads/2017/05/08115014/sapolsky_qa.jpg)
>
> Robert Sapolsky Stanford page: [Robert Sapolsky Webpage](https://profiles.stanford.edu/robert-sapolsky)

### Annotation 3 (page 1)

Beside the passage: “Sustained stress can have numerous ~athologic effects. Among the molecules that mediate such effects are the adrenal steroid hormones, including the human glucocorticoid (GC) hydrocortisone. Along with epinephrine (adrenaline) and norepinephrine, GCs are essential for surviving acute physical stress (evading a …”

**Annotation** by Micael Oliveira (2021-06-08):

> Here is a great lecture by Robert Sapolsky recorded on September 22, 2016:
> [!["sapolsky"](https://i.imgur.com/B1OiH7S.png)](https://youtu.be/D9H9qTdserM?t=95)

### Annotation 4 (page 1)

Beside the passage: “ndrome 1”

**Annotation** by Grégoire Jadi (2021-06-15):

> A long period of exposure to stress is long:
> - 1000 days = 2.7 years
> - 2000 days = 5.5 years
> - 3000 days = 8.2 years
> - 4000 days = 10.9 years
>
> (Depression timeline)

### Annotation 5 (page 1)

Beside the passage: “Sustained stress can have numerous ~athologic effects. Among the molecules that mediate such effects are the adrenal steroid hormones, including the human glucocorticoid (GC) hydrocortisone. Along with epinephrine (adrenaline) and norepinephrine, GCs are essential for surviving acute physical stress (evading a …”

**Annotation** by Micael Oliveira (2021-06-08):

> ### Glucocorticoids
>
> Glucocorticoids are a type of corticosteroid hormone that is very effective at reducing inflammation and suppressing the immune system. Inflammation is the way our immune system responds to harmful substances and trauma and is part of our healing process. However, if the usual control mechanisms that turn the process of inflammation off aren’t functioning properly and it continues unabated, our tissues can become damaged. Continued inflammation is associated with many chronic conditions.
>
> Hormones such as GCs and adrenaline are released by primates to respond to stressful situations. These hormones which instantaneously increase the heart rate and energy level. The stress response is incredibly ancient evolutionarily.
>
> In the short term stress hormones are brilliantly adapted to help you survive an unexpected threat. According to the author of this paper: ***"You mobilize energy in your thigh muscles, you increase your blood pressure and you turn off everything that's not essential to surviving, such as digestion, growth and reproduction. You think more clearly, and certain aspects of learning and memory are enhanced. All of that is spectacularly adapted if you're dealing with an acute physical stressor—a real one."***
>
> > ***"Fish, birds and reptiles secrete the same stress hormones we do, yet their metabolism doesn't get messed up the way it does in people and other primates."***

### Annotation 6 (page 1)

Beside the passage: “It is not clear whether the atrophy is associated with trauma (combat or abuse) or with succumbing to PTSD (which occurs in 5 to 20% of such traumatized individuals). In the Gurvits study, control groups consisted of healthy volunteers (matched for age, education, and other characteristics) and matched veterans with a …”

**Annotation** by Marcos Pereira (2021-06-17):

> Alternatively, there is some evidence suggesting the atrophy is not a consequence of the trauma, but the reverse:
>
> "Some studies shows correlation of reduced hippocampus volume and posttraumatic stress disorder (PTSD) [...] This finding was not replicated in chronic PTSD patients traumatized at an air show plane crash in 1988 (Ramstein, Germany).[141] It is also the case that non-combat twin brothers of Vietnam veterans with PTSD also had smaller hippocampi than other controls, raising questions about the nature of the correlation.[142] A 2016 study strengthened theory that a smaller hippocampus increases the risk for post-traumatic stress disorder, and a larger hippocampus increases the likelihood of efficacious treatment.[143]"

### Annotation 7 (page 1)

Beside the passage: “A first example, recently published by Sheline and colleagues at Washington University School of Medicine, concerns major depression (4). Approximately half of depressive patients studied secrete abnormally high amounts of GCs. Although investigators had searched with magnetic resonance imaging (MRI) for hippocampal …”

**Annotation** by Micael Oliveira (2021-06-08):

> > ***They have found significant reductions in the volume of both hippocampi (12% in the right and 15% in the left) when comparing individuals with a history of depression to age-, educa- tion-, gender-, and height-matched controls. No change in overall brain volume was ob- sewed.***

### Annotation 8 (page 2)

Beside the passage: “relevant to question whether the high-dose GC reeimes used to control manv autoimmune akd inflammatory diseases have neuropathological consequences. (Both therapeutic and experimental administration of GCs to humans results in memory impairment.) In addition, in the rodent the extent of lifetime GC exposure can …”

**Annotation** by Micael Oliveira (2021-06-08):

> More literature on this topic:
>
> - Here is a Review Article published in 2016: [Chronic Stress and Glucocorticoids](https://downloads.hindawi.com/journals/np/2016/6391686.pdf)
> - Here is an 2007 interview with Robert Sapolsky [Robert Sapolsky discusses physiological effects of stress](https://news.stanford.edu/news/2007/march7/sapolskysr-030707.html)

### Annotation 9 (page 2)

Beside the passage: “How persistent are the changes? Although the Cushingoid atrophy reverses with correction of the endocrine abnormality (6), in the PTSD and depression studies, the atrophy occurred months to years after the trauma or the last depressive episode, and at a time when patients did not hypersecrete GCs. Thus, these …”

**Annotation** by Jonas Hirshland (2021-08-19):

> Do methods exist to reduce GC secretion? Would they hinder ability to handle stress?

## Full text

Extracted from the PDF. [Annotation N] marks where each annotation above sits.

#### Page 1

Why [Annotation 1] Stress Is Bad for Your Brain

Robert M. Sapolsky

Sustained stress can have numerous ~athologic effects. Among the molecules that [Annotation 3] mediate such effects are the adrenal steroid hormones, including the human glucocorticoid (GC) hydrocortisone. Along with epinephrine (adrenaline) and norepinephrine, [Annotation 5] GCs are essential for surviving acute physical stress (evading a predator, for example) but they may cause adverse effects when secretion is sustained, such as when waiting to hear about a grant renewal (1 ). Excessive exDosure to GCs has adverse effects in the rodent brain, particularly in the hippocampus, a structure vital to learning and memory and possessing high concentrations of receDtors for GCs (2). A few davs of stress . . or GC overexposure "endangers" hippocampal neurons, compromising their ability to survive seizures or ischemia; as the likely underpinning of this, the steroids worsen the poor regulation of glutamate and calcium that occurs during such neurologic insults. Over the course of weeks, excess GC reversibly causes atrophy of hippocampal dendrites, whereas GC overexvosure for months can cause vermanent loss of hippocampal neurons. Although a few studies suggest that similar effects occur in the brains of primates (3), there has been virtually no evidence for GC-induced damage in the human. Some new, exciting studies present the first such evidence.

A first example, recently published by Sheline and colleagues at Washington University School of Medicine, concerns major depression (4). Approximately half of depressive patients studied secrete abnormally high amounts of GCs. Although investigators had searched with magnetic resonance imaging (MRI) for hippocampal atrophy in depressives, these studies could not distineuish the " hippocampus from neighboring structures or used geriatric de~ressives with brain-wide atropgy from an ;nay of diseases. The authors of the new studv reDort MRIs with far , . more resolution than in previous studies and have excluded individuals with neurologic, metabolic, or endocrine diseases. They have found significant reductions in the volume of both hippocampi (12% in the right and 15% in the left) when comparing individuals with a history of depression to age-, educa- [Annotation 7] tion-, gender-, and height-matched controls. No change in overall brain volume was obsewed. The individuals studied had been

The author is in the Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA.

de~ression-free for months or decades and. at thi time of the study, had normal GC concentrations. The investieators ruled out sev- - era1 confounding variables: alcohol or substance abuse, electroconvulsive therapy, and current use of antidepressants. Remarkably, there was a significant correlation between the duration of the depression and the extent of atrophy (see figure, top panel).

A similar relation is seen in patients with Cushing's syndrome: GCs are overproduced as a result of a hypothalamic, pituitary, adrenal, or pulmonary tumor, and there is bilateral hippocampal atrophy (5). Unfortunately, for control values the authors of this study had to relv on com~arisons with ~ublished data from MRI scans. However, as an impressive internal control, among the Cushingoid individuals, the extent of GC hypersecretion correlated with the extent of hippocampal atrophy (which also correlated with the extent of impairment in hippocampal-dependent cognition) (see figure, middle panel). No atrophy occurred in the caudate nucleus, a brain region with few GC receptors (6). Additional evidence of the relation between GCs and hippocampal function has emerged from studies of individuals with posttraumatic stress disorder (PTSD). In Vietnam combat veterans with PTSD, Bremner and colleagues found a significant 8% atrophy of the right hippocampus (and near significant atrophy of the left) (7). In a study in Biological Psychiany (in press), Guwits, Pitman, and colleagues also examined Vietnam veterans with PTSD and found significant 22 and 26% reductions in volumes of the right and left hippocampi, respectively (8). Finally, in another study, also in press in Biological Psychiatry, Bremner et al. found a 12% atrophy of the left hippocampus in adults with PTSD associated with childhood abuse (with near significant atrophy in the right hippocampus) (9). The studies controlled for age, gender, education, and alcohol abuse-and the Bremner studie-ruled out depression as a confounding variable as well. There is some uncertainty as to the anatomical specificity of the effect. In the studies by Bremner, the results were only presented as absolute hippocampal volume, and there were nearly as large (but nonsignificant) reductions in volumes of the amygdala, caudate nucleus, and temporal lobe. However, the study by Guwits et al. showed hippocampal atrophy after correction for whole-brain volume, with no atrophy in the amygdala.

ndrome [Annotation 2] [Annotation 4] 1

brain atrophy? Fielati& between hippocampal volume and (top) duration of depression among individuals with a history of major depression [from (41, (mlddle) extent of cortisol hypersecretion among Cushingoid patients [adapted from (5)], and (bottom) duration of combat exposure among veterans with or without a history of posttraumatic stress disorder [from (a]. Cortisol is another term for the human GC hydrocortisone.

It is not clear whether the atrophy is associated with trauma (combat or abuse) or with succumbing to PTSD (which occurs in 5 to 20% of such traumatized individuals). In the [Annotation 6] Gurvits study, control groups consisted of healthy volunteers (matched for age, education, and other characteristics) and matched veterans with a history of combat exposure but no PTSD. In the combat veterans, both with and without PTSD, longer durations of combat were associated with smaller hippocampi (see figure, bottom panel). However, because the PTSD patients sustained longer combat exposure than did the controls who had experienced combat but did not have PTSD, it was impossible to dissociate combat from PTSD as a predictor of atrophy. In contrast, in the Bremner combat study (in which there was no non-PTSD combat control group), combat duration did not predict extent of atrophy. Finally, in the childhood abuse study (in which there were no non-PTSD childhood abuse controls), it was not pos-

SCIENCE VOL. 273 9 AUGUST 1996

#### Page 2

sible to dissociate the PTSD from the trauma.

Each of these studies has some weaknesses, but they are countered by complementary strengths in the other studies.

Are GCs the damaging agents? Depression is accompanied by numerous physiological abnormalities, and it has not been demonstrated that the hippocampal atrophy occurs only among depressives who overproduce GCs. Moreover, among individuals with PTSD, there is no information as to the extent of the GC stress response during the trauma (or what additional physiological changes occur then). Thus, in these cases, it is not clear whether GCs mediate the atrophy. However, as noted, the defining abnormality in Cushing syndrome is GC excess, making it a likely culprit in causing atrophy.

How persistent are the changes? Although the Cushingoid atrophy reverses with correction of the endocrine abnormality (6), in the PTSD and depression studies, the atrophy [Annotation 9] occurred months to years after the trauma or the last depressive episode, and at a time when patients did not hypersecrete GCs. Thus, these long-standing changes could conceivably represent irreversible neuron loss.

The PTSD and depression studies present a problem of causality. Given the cognitive role of the hippocampus, a smaller hippocampus might be more likely to lead to being assigned frontline combat duty rather than a skilled task at headquarters. Furthermore, eiven the evidence of de~ression as a disorder - of "learned helplessness," a smaller hippocampus might predispose toward depression (that is, less cognitive capacity to detect efficacious coping responses and thus greater vulnerability to learned helplessness). Finally, PTSD individuals, before joining the military, had high rates of learning disorders and delayed developmental landmarks that could reflect cerebral atrophy (10). Thus, a small hippocampus could be a cause, rather than a consequence, of the trauma or stressor in these studies. However. there is no ~lausible way in which a small hippocampus' predisposes one toward the pituitary or adrenal abnormalities of the Cushingoid patients, or toward being a victim of childhood abuse.

Should this literature ultimatelv show that sustained stress or GC excess can damage the human hippocampus, the implications are considerable. It would then become

STATs Find That Hanging Together

Can Be Stimulating

Stewart Leung, Xiaoxia Li, George R. Stark

Transcription factors~ctivate the synthesis of messenger RNAs from DNA, therebv changing ;he function of cells. A few year; ago, a new family of transcription factorsthe STATs (signal transducers and activators of transcription)-was described that mediates the action of a large and vastlv im- - portant class of signaling molecules, the cytokines and growth factors. Each cvtokine or growth fac;or activates a distinct skt of genes to produce very distinct effects on the cell, yet there are only a limited number of STATs to mediate these signals. How do these few STATs generate a specific response for each cytokine or growth factor? Part of the answer to this puzzle is provided in a report by Xu et al. in this week's issue of Science (1 ) . The STATs exist as latent transcription factors in the cytoplasm. After binding of the growth factor or cytokine to its receptor, the STAT is activated by tyrosine phospho-

The authors are n the Department of Molecular Boogy Research lnsttute, Cleveland Clinic Foundatlon, Cleveland, OH 44195, USA. E-mail. starkg@ cesmtp.ccf org

rylation (24); it then migrates to the nucleus, binds to s~ecific DNA elements. and activates the transcription ofnearby genes. The six STAT family members form homoor heterodimers in which the phosphotyrosine of one partner binds to the SH2 (SRC homology 2) domain of the other (5). These dimers bind to palindromic GAS sequences that have similar affinities for different STATs.

The new work by Xu et al. (1) describes how each cytokine elicits a specific transcriptional response when each must use a limited number of factors and when the target DNA - elements distinguish relatively poorly among these factors. In investigating a region of the human interferon-y (IFN-y) gene that contains clusters of GAS elements, these authors found that homodimers of STATs 1,4, 5, and 6 all bind, but with different footprints. Their observations suggest that STAT dimers may cooperate in binding to clustered GAS elements and that the details of this cooperation may help to determine the cytokine specificity of the response.

The STAT proteins share blocks of ho-

relevant to question whether the high-dose GC reeimes used to control manv autoimmune akd inflammatory diseases have neuropathological consequences. (Both therapeutic and experimental administration of GCs to humans results in memory impairment.) In addition, in the rodent the extent of lifetime GC exposure can influence the likelihood of "successful" hippocampal and cognitive aging (1 1 ); similar issues must be examined concerning our own dramatic differences in cognitive aging. [Annotation 8]

References

1 A Munck eta/. Endocr Rev 5, 25 (1984) 2 B. McEwen, Prog. Brain Res 93, 365 (1992), R Saoolskv. Semin Neurosci 6 323 11994)

Sapolsky et a/, ibid. 10, 2897 (1990); A Magar~nos eta/, ibid 16, 3534 (1996) 4 Y. Sheline eta/., Proc. Natl Acad. Sci U.S.A 93, 3908 (1 996). 5 M Starkman el a/, Biol Psychiatry 32, 756 (1 992). 6. M. Starkman, personal communlcaton 7 J, Bremner eta/, Am J Psychiatry 152, 973 (1995) 8. T. Gurvits eta/, Bin/ Psychiatry, in press 9 J Bremner eta/., ibid in press. 10. T Gurv~ts eta/, J Neuropsychiatry Ciin Neuroscl 5. 183 119931. 11. M J ~eaney el a/., Science 239, 766 (1988); A. ssa eta/., J Neuroscl 10, 3247 (1990).

mology, arrayed over their entire 800-amino acid length, and it is likely that similar domains have similar functions: (i) The SH2 domain near residue 600 is highly conserved, as is a tyrosine near residue 700, which becomes phosphorylated upon activation. In addition to binding the phosphotyrosine of another STAT, the SH2 domain also mediates the binding of STATs to specific phosphotyrosine residues of activated cytokine receptors (6-8). (ii) The COOH-termini of STATs mediate transcriptional activation, and phosphorylation of a serine residue in this region of STATs la, 3, 4, and 5 enhances this activity (9). In contrast, the acidic COOH-terminal region of STAT2 can activate transcription without phosphorylation (1 0). (iii) STATs contain a DNA binding domain near residues 400 to 500 (1 1). (iv) STAT2-STAT1 heterodimers bind to an additional protein, p48, to form the major transcription factor generated in response to IFN-a. The region comprising residues 150 to 250 of STAT1 interacts with p48 (1 2). Other STAT dimers may also interact with p48 (or similar proteins) to form more complex oligomeric transcription factors. Xu et al. (1 ) have found a new function for the NH2-terminal domains of STATs 1 and 4: Mediating cooperative binding of these STATs to tandem GAS sites. Deletion of 90 amino acids from the NH2-terminus of STAT4 did not affect its binding to a single GAS site but abolished the cooperative binding of two STAT4 dimers to a double site. Furthermore, a peptide representing the

SCIENCE VOL. 273 9 AUGUST 1996

---

Source: [Fermat's Library](https://fermatslibrary.com/) — https://fermatslibrary.com/s/why-stress-is-bad-for-your-brain. More annotated papers: https://fermatslibrary.com/llms.txt
