Does HPA dysfunction predispose to RA?

Adrenal glucocorticoids, secreted in response to pituitary adrenocorticotropic hormone (ACTH) stimulation, are considered among the key factors involved in regulation of immune responses. Thus dysfunction of the HPA axis has been suspected to be involved in the onset or perpetuation of chronic inflammation in RA. It has been suggested that inherited or acquired down-regulation of the HPA axis essentially would create a predisposing environment for autoimmunity development.

Specific gene variants have been associated with several changes in HPA axis reactivity resulting in suboptimal cortisol levels during challenges. In general, genes involved in HPA function were only rarely studied in association with RA. Corticotrophin-releasing hormone (CRH) promoter polymorphisms were among the first neuroendocrine genes thought to be associated with RA. These early studies were performed using the candidate gene approach, and small cohorts had very low statistical power. The association of the CRH promoter polymorphism was not confirmed in subsequent large-scale genome-wide association studies in RA. Similarly, a suspected glucocorticoid receptor polymorphism failed to be confirmed in RA. In addition to genetic predisposition, some early life events appear to program HPA function in adulthood. Thus the process could contribute to development of various diseases including RA as seen in animal models of arthritis. Yet, specific data are lacking addressing HPA programming and RA predisposition.

In the context of chronic inflammation, up-regulated HPA function with higher production of cortisol would be anticipated in RA and other inflammatory diseases. Inappropriately low cortisol unable to dampen ongoing inflammation in RA has been conceptualized as a relative adrenal hypofunction. Inflammatory cytokines per se were found to have specific effect on adrenal steroids synthesis. Therefore, some of the observed subtle HPA variations could be attributed to ongoing inflammation.

Clinical evidence for HPA dysfunction in RA

In general, clinical studies in RA demonstrate normal HPA function, which has been considered inappropriately normal for the given level of inflammation. Although subtle differences in endocrine parameters were detected in the clinical studies in RA, their significance for immune system modulation remains unclear.

Interpretations of the inappropriately normal HPA function in RA range from an innate deficiency in the NEI loop effector component, which would be independent of ongoing inflammation, to a direct modulation of endocrine function by inflammatory cytokines. In addition to HPA axis control mechanisms, synthetic glucocorticoids are used extensively in patients with RA. Recent data suggest efficacy of these drugs in alleviating symptoms of inflammation, and in retarding erosive damage. It becomes clear that the net effect of low-dose glucocorticoids in the treatment of RA favors the beneficial aspects of these drugs over the negative aspects. These clinical findings further reinforce importance of physiologic regulation of the HPA axis in controlling disease activity and progression.

During the past 20 years, a great effort has been made in searching for evidence of improper HPA axis function in RA as demonstrated in animal models of arthritis. In early case-controlled human studies, there were no conclusive differences in urinary corticosteroid metabolites or in corticosteroid secretion in response to ACTH stimulation between RA patients and healthy controls. Neither did circadian secretion of cortisol and ACTH show any differences. Elevated cortisol levels were reported in premenopausal female patients with RA previously not treated with glucocorticoids. On the other hand, another study showed normal serum and normal 24-hour cortisol and elevated ACTH concentrations indicating defective adrenal gland function in untreated RA patients. In a group of 15 patients with clinical symptoms of less than 1 year duration, elevated C-reactive protein and erythrocyte sedimentation rate, and normal cortisol, ACTH, dehydroepiandrosterone (DHEA), and dehydroepiandrosterone sulfate (DHEAS) was observed compared with age- and sex-matched controls. The authors interpreted their findings that in RA patients, the HPA axis is functionally defective already in early stages of the disease, as evidenced by the inappropriately low cortisol levels compared with the ongoing inflammation. In RA patients without prior prednisolone treatment, serum levels of DHEA and cortisol were similar to controls, and serum levels of DHEAS were significantly lower as compared with controls.

The evaluation of HPA axis response to various stimuli yielded controversial results. Chikanza and colleagues showed lower diurnal cortisol levels in RA patients and lower cortisol response to surgical stress as compared with control patients with osteomyelitis and osteoarthritis. Normal results of the CRH stimulation test in these patients indicated normal pituitary and adrenal function. Based on these results, the authors suggested impaired hypothalamic function. The authors also proposed that impaired HPA axis together with observed elevated prolactin levels before and after surgery in RA patients might lead to proinflammatory hormonal status with possible involvement in RA pathogenesis. A similar study, however, did not detect differences in ACTH, cortisol, and prolactin levels before and after surgery in RA and osteoarthritis patients.

The insulin-induced hypoglycemia resulted in a minor tendency to lower interval-specific cortisol response in RA patients not treated with glucocorticoids. In a study by Eijsbouts and colleagues, basal plasma and salivary and urinary cortisol levels were not different between patients with RA and healthy controls. During the insulin-induced hypoglycemia, adrenocorticotropic hormone ACTH levels were similar, but cortisol levels were consistently lower in RA patients than in healthy controls. In the authors’ controlled investigation of glucocorticoid-naïve premenopausal RA females using the insulin-induced hypoglycemia, basal levels and hypoglycemia-stimulated responses of several adrenal steroids were studied. When compared with age- and body mass index (BMI)-matched healthy females, RA patients had lower basal DHEAS levels and, unexpectedly, a tendency to higher stimulated cortisol response. ACTH response to the hypoglycemia was comparable between RA patients and controls. An evaluation of basal levels of ACTH and cortisol, and subsequent response to CRH stimulation in newly diagnosed RA patients did not detect significant differences compared with healthy controls. Decreased response of DHEA and DHEAS to low-dose ACTH and ovine CRH stimulation in untreated RA females in the follicular phase of the cycle supports the concept that adrenal rather than pituitary function is impaired. Using a bicycle ergometer task, a cold pressor task, and a computerized Stroop color-word interference test as stimuli, RA patients tended to have a less-pronounced ACTH response and had a significantly smaller cortisol response than healthy controls in reaction to the stressors.

Clinical evidence for HPA dysfunction in RA

In general, clinical studies in RA demonstrate normal HPA function, which has been considered inappropriately normal for the given level of inflammation. Although subtle differences in endocrine parameters were detected in the clinical studies in RA, their significance for immune system modulation remains unclear.

Interpretations of the inappropriately normal HPA function in RA range from an innate deficiency in the NEI loop effector component, which would be independent of ongoing inflammation, to a direct modulation of endocrine function by inflammatory cytokines. In addition to HPA axis control mechanisms, synthetic glucocorticoids are used extensively in patients with RA. Recent data suggest efficacy of these drugs in alleviating symptoms of inflammation, and in retarding erosive damage. It becomes clear that the net effect of low-dose glucocorticoids in the treatment of RA favors the beneficial aspects of these drugs over the negative aspects. These clinical findings further reinforce importance of physiologic regulation of the HPA axis in controlling disease activity and progression.

During the past 20 years, a great effort has been made in searching for evidence of improper HPA axis function in RA as demonstrated in animal models of arthritis. In early case-controlled human studies, there were no conclusive differences in urinary corticosteroid metabolites or in corticosteroid secretion in response to ACTH stimulation between RA patients and healthy controls. Neither did circadian secretion of cortisol and ACTH show any differences. Elevated cortisol levels were reported in premenopausal female patients with RA previously not treated with glucocorticoids. On the other hand, another study showed normal serum and normal 24-hour cortisol and elevated ACTH concentrations indicating defective adrenal gland function in untreated RA patients. In a group of 15 patients with clinical symptoms of less than 1 year duration, elevated C-reactive protein and erythrocyte sedimentation rate, and normal cortisol, ACTH, dehydroepiandrosterone (DHEA), and dehydroepiandrosterone sulfate (DHEAS) was observed compared with age- and sex-matched controls. The authors interpreted their findings that in RA patients, the HPA axis is functionally defective already in early stages of the disease, as evidenced by the inappropriately low cortisol levels compared with the ongoing inflammation. In RA patients without prior prednisolone treatment, serum levels of DHEA and cortisol were similar to controls, and serum levels of DHEAS were significantly lower as compared with controls.

The evaluation of HPA axis response to various stimuli yielded controversial results. Chikanza and colleagues showed lower diurnal cortisol levels in RA patients and lower cortisol response to surgical stress as compared with control patients with osteomyelitis and osteoarthritis. Normal results of the CRH stimulation test in these patients indicated normal pituitary and adrenal function. Based on these results, the authors suggested impaired hypothalamic function. The authors also proposed that impaired HPA axis together with observed elevated prolactin levels before and after surgery in RA patients might lead to proinflammatory hormonal status with possible involvement in RA pathogenesis. A similar study, however, did not detect differences in ACTH, cortisol, and prolactin levels before and after surgery in RA and osteoarthritis patients.

The insulin-induced hypoglycemia resulted in a minor tendency to lower interval-specific cortisol response in RA patients not treated with glucocorticoids. In a study by Eijsbouts and colleagues, basal plasma and salivary and urinary cortisol levels were not different between patients with RA and healthy controls. During the insulin-induced hypoglycemia, adrenocorticotropic hormone ACTH levels were similar, but cortisol levels were consistently lower in RA patients than in healthy controls. In the authors’ controlled investigation of glucocorticoid-naïve premenopausal RA females using the insulin-induced hypoglycemia, basal levels and hypoglycemia-stimulated responses of several adrenal steroids were studied. When compared with age- and body mass index (BMI)-matched healthy females, RA patients had lower basal DHEAS levels and, unexpectedly, a tendency to higher stimulated cortisol response. ACTH response to the hypoglycemia was comparable between RA patients and controls. An evaluation of basal levels of ACTH and cortisol, and subsequent response to CRH stimulation in newly diagnosed RA patients did not detect significant differences compared with healthy controls. Decreased response of DHEA and DHEAS to low-dose ACTH and ovine CRH stimulation in untreated RA females in the follicular phase of the cycle supports the concept that adrenal rather than pituitary function is impaired. Using a bicycle ergometer task, a cold pressor task, and a computerized Stroop color-word interference test as stimuli, RA patients tended to have a less-pronounced ACTH response and had a significantly smaller cortisol response than healthy controls in reaction to the stressors.