Inflammation and cortisol response in coronary artery disease

Abstract

Atherosclerosis is characterized by chronic inflammation involving autoimmune components. The degree of inflammatory activity, as detectable both within the atherosclerotic plaque and in the circulation, is associated with plaque destabilization and atherothrombotic complications. Endogenous glucocorticoids are modulators of innate and acquired immune responses, and as such play a key role in the reciprocal interaction between neuroendocrine and immune systems. Abnormalities in hypothalamic-pituitary-adrenal axis (HPA) function have been described in several chronic inflammatory disorders, and evidence has emerged lately that HPA dysfunction may be implicated also in the pathogenesis of coronary artery disease. This review is an outline of knowledge gained so far by previous studies of glucocorticoids in coronary atherosclerosis and myocardial infarction. The results consistently point towards a dysregulated cortisol secretion that may involve a failure to contain inflammatory activity. A dysfunctional HPA axis and its possible implications for coronary artery disease progress, including the hypothetical link between stress and inflammation, are discussed.

• Coronary artery disease
• cortisol
• glucocorticoids
• HPA axis
• inflammation
• stress

Atherosclerosis—an inflammatory disease

Inflammation has a major influence on the development of atherosclerosis. In the atherosclerotic lesion, an excessive production of proinflammatory T helper (Th) 1 cytokines, like interferon (IFN)-γ and tumour necrosis factor (TNF)-α, is considered to drive the development towards plaque rupture and acute thrombotic complications 1, 2. Activated T cells, mainly of the Th1 type, and macrophages are frequently found in so-called high-risk lesions 3, 4. The Th1 polarization is associated with an overexpression of matrix-degrading enzymes, such as cathepsins and matrix metalloproteinase (MMP)-9 in the shoulder region of the plaque, thereby contributing to the weakening of the fibrous cap 5–7. Th1 cytokines also stimulate the secretion of prothrombotic and procoagulant factors 8. On the other hand, there is increasing evidence from experimental studies that anti-inflammatory cytokines, like interleukin (IL)-10, favour development towards plaque stability and a less prothrombotic and procoagulant state 8, 9. The enhanced proinflammatory activity is not only detectable within the arterial wall. A number of cross-sectional reports have shown an increased systemic inflammatory activity in patients with stable as well as unstable coronary artery disease (CAD), though most pronounced in unstable conditions. Such data include increased numbers of leukocytes, in particular activated T helper cells, and increased levels of serological markers, like C-reactive protein (CRP), proinflammatory cytokines, and MMP-9 10–15. On the other hand, the low serum levels of IL-10 in patients with unstable CAD compared to patients with stable disease have been suggested to reflect a less efficient atheroprotection 16. The clinical importance of a proinflammatory state is further supported by a large number of epidemiological studies showing that elevations of inflammatory markers, like CRP and IL-6, are predictors of cardiovascular events 17–19. Conversely, high levels of IL-10 have been associated with an improved outcome in CAD patients 20. Altogether, data indicate that an imbalance between proinflammatory and anti-inflammatory systems has an important role in the clinical outcome of patients with atherosclerotic disease. A repeated or continuous exposure to potential self-antigens, like oxidized lipoproteins, may explain why inflammatory activity in atherosclerosis is not resolved. However, even if inflammation in atherosclerosis is initiated and driven by an antigen-specific response, it is likely that other mechanisms affecting the anti-inflammatory homeostasis are involved. During the last two decades, our understanding of the interactions between the hypothalamic-pituitary-adrenal (HPA) axis and immune-mediated inflammation has largely expanded. As will be reviewed herein, neuroendocrine dysregulation may contribute to both increased susceptibility to inflammatory/autoimmune disease and sustainment of the inflammatory process.

Key messages

• Hypothalamic-pituitary-adrenal (HPA) axis dysfunction, mainly involving a blunted HPA response, is associated with inflammatory/autoimmune disease activity.

• A flatter diurnal cortisol rhythm and a blunted HPA axis response is found in patients with preclinical as well as clinically established coronary artery disease (CAD).

• The dysregulated cortisol secretion in CAD patients is associated with a systemic inflammatory activity, including an increased inflammatory response to stress. HPA axis dysfunction may, thereby, have implications for CAD progress.

Neuroendocrine regulation of inflammation

One major mechanism by which the central nervous system regulates the immune system is through the HPA axis and its end-product, glucocorticoids 21–23. Three cytokines—TNF-α, IL-1 and IL-6—account for most of the HPA axis-stimulating activity either directly through the systemic circulation or indirectly by, for example, activating neural mechanisms. The release of cytokines is not only triggered by immune stimuli but also by non-immune stress, presumably stimulated by catecholamines acting through β2-adrenergic receptors. The first step in HPA axis activation is the release of corticotropin-releasing hormone (CRH) from intrahypothalamic neurons. CRH travel from the hypothalamus via the hypophyseal–portal blood vessels to the anterior pituitary gland where it acts via specific receptors to trigger the release of the adrenocorticotrophic hormone (corticotrophin, ACTH) from specific ACTH-producing cells into the systemic circulation. ACTH in turn acts on the adrenal cortex to initiate the synthesis of cortisol, which is released immediately into the systemic circulation by diffusion. The magnitude of the HPA response to incoming stimuli is tempered by the glucocorticoids which act at the levels of the pituitary gland and hypothalamus to suppress the synthesis and release of ACTH and CRH, thus forming a negative feedback loop (Figure 1). Cortisol is the predominant glucocorticoid in man. It also constitutes the active form while cortisone is its inactive precursor. The glucocorticoids exert wide-spread actions in the body, which are essential for the maintenance of homeostasis and enable the organism to prepare for, respond to, and cope with physical and emotional stress 24. They are important regulators of immune and inflammatory processes and are required for numerous processes associated with host defence. These properties underlie many of the stress-protective actions of the steroids as they quench the pathophysiological responses to tissue injury and inflammation and, thereby, prevent them from proceeding to a point where they threaten the survival of the host.

Abbreviations

Table

ACTH	adrenocorticotrophic hormone
CAD	coronary artery disease
CRH	corticotropin-releasing hormone
CRP	C-reactive protein
GR	glucocorticoid receptor
HPA	hypothalamic-pituitary-adrenal axis
IFN-γ	interferon-γ
IL	interleukin
MAPK	mitogen-activated protein kinase
MMP	matrix metalloproteinase
Th	T helper
TNF-α,	tumour necrosis factor-α

Figure 1.  Schematic diagram illustrating the bidirectional communication between the neuroendocrine and immune systems. Tumour necrosis factor (TNF)-α, interleukin (IL)-1, and IL-6 stimulate the secretion of corticotropin-releasing hormone (CRH) from hypothalamic neurons; at high concentrations or over a prolonged period they may also stimulate the secretion of adrenocorticotrophic hormone (ACTH) from the pituitary gland and cortisol from adrenal cortex. The hypothalamic-pituitary-adrenal axis activity is modulated by a negative feedback system through which the release of CRH and ACTH is suppressed by cortisol itself. Cortisol inhibits the production of all three inflammatory cytokines and also inhibits most of their effects on target tissues. The solid lines indicate stimulation, the dotted lines indicate inhibition.

Initially, glucocorticoids were thought to have mainly immunosuppressive effects. In 1948 it was shown for the first time that a synthesized version of cortisone was capable of reversing the inflammation of rheumatoid arthritis 25. However, it is important to recognize that glucocorticoids in pharmacological doses exert different effects than they do under physiological conditions 26. Pharmacological doses (higher concentrations than physiological) are anti-inflammatory or immunosuppressive at virtually every level of immune and inflammatory responses, whereas physiological levels of glucocorticoids should rather be considered immunomodulatory. In vitro, their role at doses slightly higher than physiological is mainly exerted through genomic effects, among which are the induction of anti-inflammatory proteins annexin-1 (also called lipocortin 1) and mitogen-activated protein kinase (MAPK) phosphatase 1, and the repression of transcription of cyclo-oxygenase 2 27. Glucocorticoids also block the transcriptional activity of nuclear factor κB, which is a major factor involved in the regulation of cytokines and other immune responses 28. As a consequence of the inhibition of nuclear factor κB, the expression of cytokines like IL-1, IL-6, IFN-γ, and TNF-α is downregulated. The net effect of glucocorticoids is a shift of cytokine production from a primarily proinflammatory to anti-inflammatory pattern, roughly corresponding to Th1 and Th2, respectively. The shift of Th1 to Th2 is considered to be due mainly to the downregulation of Th1 cytokines, thus allowing a dominant expression of Th2 cytokines 29, 30.

The transcriptional actions of glucocorticoids are mediated by supposed diffusion of the steroid hormone across the cell membrane and its binding to intracellular glucocorticoid receptors (GRs) 26, 31. Two human isoforms of the GR have been identified, termed GR-α and GR-β, which originate from the same gene by alternative splicing of the GR primary transcript 27, 32. GR-β is the predominant isoform of the receptor and the one that shows steroid-binding activity. The possible physiological role of GR-β is currently a matter for debate but it regulates GR-α activity in vivo and has been shown to act as a dominant negative inhibitor of GR-α activity.

Methodological considerations in the study of HPA axis activity

Cortisol shows a robust diurnal pattern in healthy adults with the strongest secretory activity of the adrenal cortex during the early morning hours. Peak cortisol levels are observed shortly after awakening, with steadily decreasing values thereafter, except for sizable, short-term increases in response to stimuli like lunch meal, exercise, or threat-provoking stressors. The nadir of cortisol secretion is reached around 02.00 or 03.00 with only minimal levels of the steroid detectable. The diurnal rhythm in the rate of cortisol secretion may thus explain the diurnal changes in disease-associated immune responses. For example, the delayed hypersensitivity reaction, which is particularly responsive to glucocorticoids, is most pronounced in the evening when cortisol secretion is low and least pronounced in the morning when secretion is high 33. Others have shown a circadian rhythm in the disease activity in rheumatoid arthritis with maximal activity between 02.00 and 04.00 34.

The measurement of free cortisol in 24-hour urine samples serves as an integrated measure of free cortisol concentrations during the entire day. To determine the circadian patterns of cortisol secretion or random cortisol concentrations, measurements of free cortisol can be performed in blood or saliva. However, measurements over 2–6 days are considered necessary to achieve reliable trait measures, since state factors may bias data from a single day 35. Free cortisol in saliva has been shown to have the same diurnal rhythm as blood cortisol. Furthermore, it has been shown that the transfer of cortisol from blood to saliva is rapid with a reflection in saliva of a cortisol increase in blood within 60 seconds and a state of equilibrium within 5 minutes 36, 37. Due to several advantages over blood cortisol analyses, such as stress-free sampling and laboratory independence, the determination of salivary cortisol has become the method of choice in basic research and clinical environments. It was first introduced to psychobiological stress research almost three decades ago, and since then a number of studies have investigated the association between psychological factors and cortisol in saliva 38. The activity of the HPA axis is often evaluated by measuring the levels of blood or salivary cortisol in response to different stimulation procedures such as CRH test, physical exertion or psychological stress. However, since the responses may show a large heterogeneity, standardized laboratory stress protocols that combine different stressors are preferable 39.

Cortisol in coronary artery disease

A growing body of evidence suggests an association between HPA axis activity and coronary atherosclerosis. A population-based prospective study showed that an elevated cortisol:testosterone ratio, based on one early morning blood sample, increased the risk of CAD mortality and incidence 40. However, data from earlier studies in CAD patients are both sparse and inconsistent. Positive correlations between morning plasma cortisol and the degree of CAD, as assessed by angiography, have been demonstrated 41–43, while others have failed to find associations between morning cortisol and number of diseased coronary vessels 44 or between 24-hour urinary cortisol and ischaemia, as assessed by stress echocardiography 45. On the other hand, we recently determined the total cortisol output in 24-hour urine collections and found significantly increased levels in patients with stable conditions of CAD when compared to age- and gender-matched healthy subjects (Figure 2) 46. One limitation is that neither single measurements of morning cortisol, nor 24-hour urine collections, provide any information on cortisol reactivity or daily cortisol profile. A dysregulation of the HPA axis can take the form of a smaller decline in cortisol throughout the day, i.e. a flatter diurnal slope. Although the determinants or consequences of having a flat cortisol rhythm are not clear, it is generally considered a result of long-term HPA overstimulation. A flat rhythm has also been proposed to constitute a marker of disease progression. Abnormal circadian rhythms have been observed in patients with cancer and, as an example, patients with metastatic breast cancer whose diurnal cortisol rhythms are flattened have earlier mortality 47. In a population-based study of middle-aged men, Rosmond and Björntorp showed that clusters of established risk factors for cardiovascular disease were tightly associated with a pathological HPA axis, characterized by low diurnal cortisol variability and a poor lunch-induced cortisol response 48. A number of studies have confirmed the link between subtle alterations in cortisol secretion and separate cardiovascular risk factors, including smoking, abdominal obesity, and hypertension 49–53. However, the most consistent findings have been reported in smokers and involve an increased cortisol release throughout the day as well as an attenuated HPA axis response to stress 51, 52. In a population-based study using a cross-sectional design, six salivary cortisol samples were collected from middle-aged adults on a single day, from awakening to bedtime. Results showed that the flatter the cortisol slopes throughout the day, the greater the likelihood of any coronary calcification 54. In addition, Rosmond and co-workers showed that the 5-year incidence of cardiovascular-related events was significantly higher in men with an abnormal cortisol slope along with low testosterone levels, compared to men with a normal hormone pattern 55. In order to evaluate the diurnal variation in cortisol secretion in patients with established CAD, we measured cortisol in saliva twice daily for 3 consecutive working days, the first sample taken 30 min after awakening and the second in the evening before going to bed. The morning cortisol levels did not differ between patients and controls, while the evening cortisol levels were significantly higher in patients, thus clearly indicating a flatter cortisol profile over the day in CAD patients (Figure 3) 46.

Figure 2.  The total daily cortisol production, as assessed by free cortisol in 24-hour urine collection, in apparently healthy individuals versus patients with coronary artery disease (CAD) 46, P<0.05. Box plots summarize the median, interquartile range, and minimum and maximum values.

Figure 3.  The salivary cortisol levels 30 minutes after awakening (morning) and at bedtime (evening) in apparently healthy individuals versus patients with coronary artery disease (CAD) 46, P<0.05. Values are given as mean (SD).

The association between HPA dysfunction and inflammation in CAD

Cytokines, like IL-1 and IL-6, can increase the glucocorticoid secretion by providing the synthesis and release of CRH and ACTH in hypothalamus and pituitary gland but also by enhancing the secretion of cortisol in adrenal cortex. Cortisol, in its turn, inhibits the production of TNF-α, IL-1, and IL-6. The interaction between the HPA axis and the immune system may result in reciprocally protective adaptations. For instance, immune suppression in Cushing's syndrome is mild, suggesting the development of tolerance to glucocorticoids. In animals with chronic inflammatory disease the hypercortisolism is mild rather than severe 56. However, disturbances at any level of the HPA axis or glucocorticoid action may lead to an imbalance of the system and an enhanced susceptibility to infection and inflammatory/autoimmune diseases. The association between a blunted HPA axis and susceptibility to autoimmune/inflammatory disease has been clearly shown in many animal models, e.g. when comparing two highly inbred rat strains, Fischer rats and Lewis rats 57–59. The Lewis rats are highly susceptible to a wide variety of autoimmune/inflammatory diseases, while Fischer rats are resistant to these diseases. The Lewis rats exhibit a blunted HPA axis response, compared to Fischer rats with an excessive HPA response compared to outbred rats. In Lewis rats treated with low-dose dexamethasone or transplanted intracerebroventricularly with fetal hypothalamic tissue from Fischer rats, the autoimmune disease was markedly attenuated 60. The abnormalities in Lewis rats may also have parallels in humans. In patients with CAD, we found that serum levels of IL-6 and CRP were strongly correlated to evening salivary cortisol but not to morning cortisol or 24-hour urinary cortisol 46. In addition, the circulating levels of MMP-9 were significantly correlated to evening cortisol levels and remained so after adjustment for CRP and IL-6 (unpublished observations, Nijm J, Jonasson L, 2007). These findings suggest that the low-grade systemic inflammation in CAD patients is associated with a flatter diurnal cortisol slope. In a previous clinical study, serum cortisol and IL-6 were measured in blood samples taken between 09.00 and 12.00 in patients with both stable and unstable conditions of CAD. In two-thirds of patients, the cortisol levels were found to be ‘inappropriately’ normal with concomitant high IL-6 levels thus giving rise to the interesting speculation that endogenous cortisol production was insufficient to limit inflammation in these patients 61. However, the correlations between cortisol and inflammatory markers do not prove a cause-and-effect relationship, and there are inherent difficulties in evaluating the appropriateness of a given level of cortisol for a particular level of on-going inflammation in individuals. In addition, the possible effects of medication on cortisol levels should be considered. However, data from previous studies investigating the effects of beta-blockers and statins have been consistent, giving no evidence for alterations in basal cortisol levels or ACTH-induced cortisol response 62–66.

The measurement of HPA axis response to stress is considered to be an important piece of information to include for a more precise study of HPA function. In patients with rheumatoid arthritis, the basal morning cortisol levels did not differ compared with healthy subjects but, still, the patients with rheumatoid arthritis showed a lower cortisol response after insulin-induced hypoglycaemia 67. Similar results have been demonstrated in patients with atopic dermatitis and systemic lupus erythematosus. Although the basal morning cortisol levels in these patient groups did not differ from healthy subjects, their ACTH and cortisol responses to acute psychological stress or insulin-induced hypoglycaemia were significantly lower 68, 69. When CAD patients were exposed to acute physical stress (a maximal bicycle exercise test), we found that their cortisol response was significantly attenuated compared to age- and gender-matched control subjects, thus indicating a hyporeactive HPA axis in the patients. Similarly, the cortisol response to a standardized laboratory stress test (‘anger recall’ followed by an arithmetic test) was markedly attenuated in CAD patients compared to controls (Figure 4A) 46. The differences between patients and controls remained significant even after adjustment for possible confounding factors such as smoking and treatment with beta-blocker or statin. The study of stress response in CAD patients is the first to demonstrate a blunted cortisol response in patients with atherosclerotic disease, but findings are in line with previous results from the population-based LiVicordia study (Linkoping-Vilnius Coronary Risk Assessment Study) 70. In the 1990s, Lithuanian middle-aged men had a 4-fold risk for CAD mortality compared with Swedish men. The aim of the LiVicordia study was, therefore, to compare both traditional and new possible risk factors for CAD in 50-year-old men from each of the cities Vilnius, Lithuania and Linköping, Sweden. Small differences were found in traditional risk factors between Swedish and Lithuanian men but, interestingly, Lithuanian men exhibited a significantly lower cortisol response to a standardized laboratory stress test. In addition, the LiVicordia study showed a higher prevalence of subclinical atherosclerosis in Lithuanian men, as assessed by ultrasound measurements of intima-media thickness 71.

Figure 4.  A: Salivary cortisol response, given as percentage increase, to physical stress (grey boxes) and psychological stress (white boxes) in apparently healthy individuals versus patients with coronary artery disease (CAD), P<0.01 for both tests. B: Inflammatory response, given as percentage increase in C-reactive protein (CRP), to physical stress (grey boxes) and psychological stress (white boxes) in apparently healthy individuals versus patients with CAD, P<0.01 for both tests 46.

In the CAD patients, serum levels of CRP and IL-6 were determined before and 24 hours after the stress tests 46. None of the inflammatory markers differed between patients and controls at base-line. Neither were there any differences in IL-6 levels between patients and controls after 24 hours, probably reflecting that IL-6 levels had returned to base-line levels. The hepatic synthesis of CRP is largely under the regulation of IL-6. In humans, CRP is increased following IL-6 infusion, reaching a peak level 21 hours after the cessation of IL-6 infusion 72. In the CAD patients, acute physical stress induced a significant increase in CRP in the patients compared to the controls after 24 hours. Similarly, the psychological stress test induced a significant increase in CRP in the patients but not in the controls (Figure 4B). It also deserves to be emphasized that all CAD patients participating in the stress tests were on long-term treatment with statin, a drug with well known CRP-lowering effects 73. Altogether, the findings support the hypothesis that a hypofunctional HPA axis in CAD patients fails to suppress stress-induced inflammatory activity. In an earlier study, patients with rheumatoid arthritis showed a failure to increase cortisol secretion following surgery, despite high levels of IL-1β and IL-6, compared to subjects with chronic osteomyelitis, the latter a joint disease not associated with systemic inflammation 74. The authors thus proposed that the activity of HPA axis in patients with rheumatoid arthritis was insufficient to inhibit on-going inflammation. In agreement with this, it was recently shown that an acute psychological stress test induced an increase in CRP in patients with rheumatoid arthritis, but not in patients with chronic osteomyelitis 75. Interestingly, these authors speculated that the inflammatory reaction to stress could underlie the increased risk for myocardial infarction in patients with rheumatoid arthritis.

There is a strong association between external triggers and onset of myocardial infarction and sudden cardiac death beyond that expected by chance alone. Several studies have shown that both physical exertion and acute emotional stress such as a burst of anger are potent triggers of myocardial infarction 76, 77. In one multicentre study, possible triggers were identified by 49% of the population; the most common were emotional upset (18%) and moderate physical activity (14%) 78. Lately, a case-control study with 11,119 patients with a first myocardial infarction and 13,648 age- and sex-matched controls from 52 countries (the INTERHEART study) investigated the relation of psycho-social factors to the risk of myocardial infarction. People with myocardial infarction reported a significantly higher prevalence of stress factors including stress at work and at home, financial stress, and major life events in the past year 79. The mechanism by which stress increases the risk of myocardial infarction is not clarified, but the interplay between stress and inflammation could play an important role. Thus, one intriguing possibility is that individuals with a blunted cortisol response will be more susceptible to atherosclerosis progression and plaque instability due to an ‘abnormally’ high inflammatory response to stressful stimuli.

A suppressed release of cortisol in response to stressful stimuli may be one mechanism linking HPA dysfunction to inflammatory disease, and glucocorticoid resistance in the target tissue may be another 80, 81. Glucocorticoid resistance is characterized by increased cortisol secretion without clinical evidence of hypercortisolism. The circadian rhythm of cortisol is intact albeit at higher concentrations, and there is resistance of the HPA axis to dexamethasone treatment. Up to 30% of patients with autoimmune or chronic inflammatory diseases have a form of glucocorticoid resistance, yet the mechanisms are not fully established. In experimental studies, the treatment with TNF-α or IL-1 leads to the development of glucocorticoid resistance by inducing a relative overexpression of GR-β 82. GR gene variants may also be related to changes in cortisol sensitivity. The haplotype 3, characterized by the GR-9β polymorphism, leads to an increased GR-β expression by generating a more stable GR-β messenger RNA transcript 83. This polymorphism has been found to be associated with rheumatoid arthritis in a small observational study 84. Moreover, in a large population-based prospective cohort, haplotype 3 was recently shown to be associated with a more active proinflammatory state, as assessed by elevations in CRP and IL-6, and greater carotid intima-media thickness 85. Interestingly, in this study the carriers of haplotype 3 also showed a more than 2-fold risk of myocardial infarction during a 9-year follow-up.

To conclude, coronary atherosclerosis is associated with hypercortisolism, involving a flatter diurnal rhythm. In addition, patients with clinically established CAD show a blunted cortisol response to stress, a feature previously described in autoimmune diseases, like rheumatoid arthritis. The relationship between a flat diurnal cortisol pattern and an attenuated cortisol response to stress is not clarified but both phenomena have been associated with reduced activity of the HPA axis. In the CAD patients, the low diurnal cortisol decline correlates with systemic inflammatory activity and, upon stressful stimuli, their attenuated cortisol response is accompanied by a significant increase in CRP. It remains unclear if the HPA dysfunction is mainly a secondary phenomenon due to a long-lasting inflammatory process or if it is more actively involved in the pathogenesis of CAD. Yet, the role of the HPA axis in atherogenesis has to be demonstrated in animal models. However, chronic or repeated immune insults on the HPA axis function may eventually lead to a poorer regulation of the inflammatory response, possibly earlier and more pronounced in genetically predisposed individuals. The imbalance between anti-inflammatory and proinflammatory mechanisms may thus lead to an increased susceptibility to immune insults and a maintenance of a proinflammatory state. It is a well known fact that a proinflammatory state is closely related to atherosclerosis and risk of cardiovascular events. Although the clinical consequences of a hyporesponsive HPA axis are still unknown, it is intriguing to speculate that it may favour the development towards plaque instability, thereby providing a link between stress and inflammation in CAD.

Acknowledgements

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.