The endocrine system and it's effect on Inflammation

[Mike Bartolatz]

IMMUNE SYSTEM, EFFECTS ON THE ENDOCRINE SYSTEM
Chapter 29 - Jason A. Berner, M.D., Dimitris A. Papanicolaou, M.D.
March 11, 2003 Index
Contributors
Search



INTRODUCTION
Over the last two decades significant progress has been made on the study of the effects of the immune system on the endocrine system. It has become evident that most of the effects of the immune system on the endocrine system are mediated by cytokines. Initially, cytokines were thought to be products of the immune system alone that had immune and hematological functions only. However, it has become apparent that cytokines are part of a neuroendocrine and immune system network, in which both the nervous and endocrine systems participate. Furthermore, cytokines are being secreted by non-immune cells as well, such as adipocytes.


The inflammatory cytokines interleukin-6 (IL-6) and--to a lesser extent--tumor necrosis factor-alpha (TNF-a) and IL-1 have been shown to have endocrine, autocrine, and paracrine roles. IL-6, in particular, has drawn much attention in the endocrine field for it is the most "endocrine" of all cytokines. This chapter will focus on the interactions between cytokines, especially IL-6, on different endocrine systems.

IMMUNE SYSTEM AND THYROID DISEASE

Euthyroid Sick Syndrome

The term "Euthyroid Sick Syndrome" (ESS) has been used for more than thirty years to describe a pattern of thyroid hormone alterations during non-thyroidal illness. The degree of thyroid function abnormalities correlates with disease severity. Conditions associated with ESS include starvation, sepsis, surgery, trauma, and chronic degenerative diseases, to name a few. The characteristic laboratory abnormalities with ESS include low triidothyronine (T3) and/or free T3 (fT3), elevated reverse T3 (rT3), normal thyroid stimulating hormone (TSH), and normal serum thyroxine (T4) or free T4 (fT4) concentration. The more severe the illness is the more extensive the hormonal changes are.

Both low T3 and low T4 states have been reported. The typical laboratory findings that appear in the low T3 state include low peripheral T3 and/or fT3 concentration and elevated rT3 concentration along with normal TSH levels. These thyroid hormone changes are the result of inhibition of hepatic type-1 5¢ deiodinase (D1) that facilitates conversion of T4 to T3 and of rT3 to diiodothyronine (1). Thus the cause of the decreased T3 concentration in ESS is decreased T3 production, whereas the cause of the increased rT3 concentration is the result of attenuated degradation. Factors that can precipitate a low T3 include systemic illness, decreased food intake and several catabolic conditions, including diabetes mellitus. Prolonged and severe illness is marked by a decrease in circulating total T4 along with low T3 and high rT3. Very low T4 levels carry a poor prognosis and have been associated with an increased likelihood of death.

The cytokines that play a key role in the hormonal responses observed during ESS are IL-1, TNF-a, interferon-gamma (IFN-g) and, especially, IL-6. These cytokines have been shown to cause inhibition of iodide uptake in vitro (Figure 1) (2). TNF-a, IL-1beta, and--to a lesser degree-- IFN-g inhibited both constitutive and TSH-stimulated Na/I-symporter gene expression in rat thyroid FRTL-5 cells. Na/I-symporter gene expression correlated well with TSH-stimulated iodide uptake in FRTL-5 cells, suggesting that the above cytokines modulate iodide uptake in these cells by modulating the Na/I symporter mRNA expression. In addition to effects on iodide uptake, cytokines have been shown to decrease thyrocyte growth (3), organification of iodide (4, 5), synthesis of thyroglobulin (6, 7), and thyroid hormone release in vitro (8).


Figure 1. Inhibition of Na+/I- gene expression resulting in decreased iodide uptake by thyrocytes.

Besides the effects mentioned on inhibition of thyroid gland function, cytokines have an effect on hepatic D1 activity as well (Figure 2). As previously mentioned, the role of D1 is to peripherally convert T4 to T3 and rT3 to diiodothyronine. In ESS, there is a decrease in D1 activity causing decreased T3 and increased rT3 concentrations. However, the exact mechanism of decreased D1 activity in ESS is unclear. In vitro studies, involving the effects of cytokines IL-1b, IL-6, and TNF-a on D1 levels in rat thyroid FRTL-5 and liver cells, produced controversial results. The D1 activity in rat thyroid FRTL-5 was inhibited by these cytokines (9), whereas liver D1 activity was surprisingly increased (10).


Figure 2. IL-6 inhibition of 5´Deiodinase-I resulting in decrease in T3 and increase in rT3 concentration.

While the in vitro studies of cytokines on D1 activity are controversial, the in vivo studies revealed evidence that cytokines inhibit D1 activity directly or indirectly. To delineate the effects of IL-6 on D1 activity, studies were done using IL-6 knockout mice (i.e. lacking IL-6) (11). Listeria monocytogenes infection or turpentine injection was used to induce the ESS state. The decrease in serum T3 concentration was attenuated in IL-6 knockout mice compared to wild-type animals. This was associated with only a modest decrease in hepatic D1 activity (compared to wild-type animals), implying that IL-6 played a significant role in the pathogenesis of ESS in that model. In another study, immunoneutralization of IL-6 was found to attenuate lipopolysaccharide (LPS)-induced decrease of hepatic D1 activity, implying that IL-6 mediated the effect of LPS on that liver enzyme (12). In the same study immunoneutralization of IL-1, TNF-a, or IFN-gamma had no effect on D1 activity.

In exploring the effects of cytokines on the hypothalamic-pituitary unit, in vitro studies demonstrated that IL-1b and TNF-a inhibited TSH release from the pituitary (Figure 3) (13). It appears that IL-1b inhibits TSH release from the pituitary indirectly through stimulation of K+-mediated release of somatostatin from the hypothalamus. IL-6 demonstrated no such effect on TSH secretion or somatostatin release, implying that this cytokine had no direct effect on the hypothalamic-pituitary unit of the thyroid axis (14).


Figure 3. Cytokine-mediated decrease in TSH secretion.

In animal studies, administration of IL-1b to rats resulted in a decrease in TSH release with decreases in both basal and TRH-stimulated TSH levels. Both pro-TRH and b-TSH mRNA levels were decreased as well (15). TNF-a administration to rats had a similar effect (16, 17). After IL-6 was administered to rats, TSH decreased. However, there was no change in hypothalamic pro-TRH mRNA levels, or in stored b-TSH in the pituitary (18). These data, coupled with the fact that IL-6 had no effect on TSH release in vitro, suggest that the observed decrease of circulating TSH in vivo following IL-6 administration was the result of an indirect- probably through stimulation of the HPA axis- rather than a direct action of this cytokine on the TRH-TSH unit.

Clinical trials further support the notion that IL-6 plays a central role in the pathogenesis of ESS. One prospective study that involved septuagenarian patients admitted for emergency surgery demonstrated that half of those patients had low serum T3 concentrations preoperatively. Patients with ESS had increased mortality and higher circulating IL-6 and C-reactive protein levels along with decreased albumin concentrations compared to those who were euthyroid. Plasma IL-6 concentration correlated strongly with serum rT3 concentration (19). Another study involving surgical patients revealed that postoperative plasma IL-6 concentration correlated negatively with both serum fT4 and serum fT3 concentration (20). Furthermore, Boelen et al. showed that serum IL-6 concentration correlated negatively with both fT4 and fT3 concentration and positively with serum rT3 concentration in patients admitted to the general medicine ward (21).

The central role that cytokines play in the pathophysiology of ESS has been further demonstrated in studies involving cytokine administration to humans. TNF-a administration to healthy volunteers resulted in a decrease serum T3 concentration and an increase in serum rT3 concentration (22). Unlike IL-6, serum TNF-a did not correlate with any of the typical thyroid parameters (low T3, increased rT3, decreased TSH levels) seen in ESS in correlative studies (19, 23). Therefore, the changes to the thyroid hormone profile that occurred during TNF-a administration might be indirect (i.e. through TNF-a increasing circulating IL-6 levels) rather than direct. Administration of IFN-a to normal volunteers resulted in a decrease in serum T3, and increase in serum rT3 and a decrease in serum TSH concentrations, probably through increase in IL-6 concentration (24). IL-6 administration to healthy volunteers resulted in an increase in serum rT3 and a decrease in serum T3 concentrations, most likely due to a depressive effect of IL-6 on hepatic D1 activity (25). Serum TSH concentration was decreased following IL-6 administration. The nadir of serum TSH concentration coincided with the peak of cortisol concentration following IL-6 administration, suggesting that IL-6 might have mediated its effect on TSH indirectly through stimulation of the HPA axis.

The Role of IL-6 in Amiodarone-Induced Thyroid Disease

Amiodarone, an iodine-rich cardiac drug, may induce thyrotoxicosis or hypothyroidism, not only in patients with underlying thyroid disease but also in people with apparently normal thyroid function. Amiodarone can induce thyrotoxicosis through two different mechanisms. One mechanism involves a Jod-basedow-like effect of increased iodide uptake and thyroid hormone synthesis. The other mechanism involves thyroid gland destruction resembling subacute thyroiditis. A cross-sectional study in patients with amiodarone-induced thyrotoxicosis (AIT) revealed that serum IL-6 concentration was elevated in patients with AIT without a goiter or circulating thyroidal autoantibodies (AIT-) compared to patients that had AIT in the presence of a goiter or thyroidal autoantibodies (AIT+) (26). AIT- patients had a very low (<3%) 24-hour thyroidal radioiodine uptake suggesting that a subacute thyroiditis-like mechanism was responsible for the thyrotoxicosis in these patients. To determine if plasma IL-6 concentration was elevated in other destructive processes besides AIT, serum IL-6 concentration was measured in patients after undergoing fine needle aspiration of the thyroid, percutaneous ethanol injections into thyroid nodules, or radioactive iodine treatment. Serum IL-6 concentration increased significantly following any of the above procedures, suggesting that IL-6 could be used as a marker of thyroid destructive processes, regardless of etiology (27).

Postpartum Thyroid Disease

Postpartum thyroiditis is a syndrome of transient or permanent thyroid dysfunction occurring in the first year after delivery due to an autoimmune inflammation of the thyroid (28). The disease is initially characterized by hyperthyroidism, which eventually progresses to hypothyroidism. The pathogenesis of the disease is autoimmune attack of the thyroid gland typically associated with thyroidal autoantibodies. Anti-thyroid peroxidase antibody (anti-TPO) and anti-thyroglobulin antibody (anti-Tg) are the major antibodies associated with postpartum thyroiditis. The titer of anti-TPO tends to be higher than that of anti-Tg when both antibodies are present (29, 30). Of these two antibodies mentioned, the anti-TPO is the one that best correlates with the development of postpartum thyroiditis. Thyroid dysfunction in women with antibodies to TPO was a function not only of the titer of anti-TPO antibodies but also of their ability to activate complement (31). The ability to activate complement is thought to be central to the association of anti-TPO antibodies with postpartum thyroiditis. Reports collectively demonstrate that 30-60% of anti-TPO-positive women in pregnancy subsequently develop postpartum thyroiditis. Although anti-TPO antibodies have a strong association with postpartum thyroiditis, relationship of actual causality between anti-TPO antibodies and postpartum thyroiditis has not been established (28).

One plausible explanation for postpartum thyroiditis is that during pregnancy there is a shift from T-helper 1 (TH1 or Type 1) to T-helper 2 (TH2 or Type 2) cytokine production, followed by a "rebound" shift back to Type 1 after delivery (Figure 4). Type 1 cytokines (e.g. IL-12, IFN-g) are pro-inflammatory and have been implicated in the pathogenesis of several autoimmune diseases, whereas Type 2 cytokines (e.g. IL-4, IL-10, IL-13) are anti-inflammatory. It has been suggested that Type 2 cytokine production is teleologically important during pregnancy for the fetus not be immunologically rejected (32, 33). Elenkov et al. recently showed that ex vivo IL-12 production after LPS stimulation is decreased during the third trimester of pregnancy and becomes increased during the postpartum period (34). This study further supports the notion that the rebound shift back to Type 1 cytokine production after delivery may be so extensive that autoimmune thyroid failure is triggered.


Figure 4. Th1/Th2 balance before, during and after pregnancy

Progesterone has been shown to effect pregnancy-associated immunomodulation via alteration of the Type 1/Type 2 cytokine balance (35). Progesterone's immunological effects are partially exerted through progesterone-induced blocking factor (PIBF) (a protein fraction of 34 kD) (36, 37). This factor is produced by progesterone-influenced lymphocytes. PIBF has immunomodulating properties that include being able to regulate perforin expression by NK cells (37-39). In addition, PIBF affects Type 1/Type 2 balance via an increased production of IL-3, IL-4, and IL-10 and decreased production of IL-12 from lymphocytes and macrophages in vitro. Treatment of mice with anti-PIBF antibody resulted in a shift toward Type 1 cytokine production along with increased rates of pregnancy resorption (40). The effects of estradiol on the Type 1/Type 2 balance are less clear, as both pro- and anti-inflammatory effects on CD4+ T cells by this hormone have been reported (41, 42).

Type 1 versus Type 2 cytokine balance is influenced during pregnancy by cortisol as well. Pregnancy is characterized by suppression of the hypothalamic CRH neurons, due to the occurrence of hypercortisolism in the third trimester. At that time, the pituitary-adrenal unit is driven by placental CRH, which, unlike hypothalamic CRH, is positively regulated by glucocorticoids (43). It has been postulated that the teleological significance of the hypercortisolism of pregnancy is the suppression of the mother's immune system so that the fetus is not rejected as a foreign body. After delivery, circulating cortisol levels decrease abruptly. This causes women to be relatively hypocortisolemic in the early postpartum state, until HPA axis activity returns to normal (44). The hypocortisolemic state is associated with an increase in Type 1 relative to Type 2 cytokine production. The subsequent increase in Type 1 cytokine production after delivery may be responsible for the exacerbation of certain autoimmune diseases (45).

Hashimoto's Thyroiditis and Graves' Disease

The predominance of TH1 versus TH2 CD4+ lymphocytes appears to play a significant role in the pathogenesis of autoimmune diseases as mentioned previously. The antibodies involved in autoimmune thyroid disease are anti-TPO and anti-Tg in Hashimoto's thyroiditis and anti-TSH receptor antibody in Graves' disease. Fisfalen et al. demonstrated that 65% of T cell clones derived from Hashimoto's thyroiditis (HT) patients were found to be reactive to TPO and 59% of T cell clones derived form Graves' disease patients were found to be reactive to TSH receptor (TSH-R) (46). However, antigen reactivity was not disease-specific: 16% of clones from HT patients were reactive to TSH-R antigens and 41% of clones derived from GD patients were reactive to TPO antigens. TH1 cells were predominant in the T cell clones derived from HT patients, whereas TH1 and TH0 cells were equally present in the clones derived from Graves' disease patients. Interestingly, anti-TPO and anti-Tg-specific clones established from HT thyroid tissue produced high levels of IFN-g, a Type 1 pro-inflammatory cytokine (46). TH1 cells may also affect autoimmune thyroid disease through induction of thyrocyte apoptosis- which appears to be a major mechanism of thyroid tissue damage- indirectly through IL-1b production by activated macrophages (47). In addition, thyrocytes themselves can produce inflammatory cytokines such as TNF-a, TGF-b, IL-1, IL-6, and IL-8, which also can cause thyrocyte destruction.

THE IMMUNE SYSTEM AND DIABETES MELLITUS

Type I Diabetes Mellitus

Insulin dependent diabetes mellitus (IDDM) or type I diabetes is an autoimmune disease, in which there is progressive and selective destruction of the insulin-producing beta cells in the islets of Langerhans of the pancreas. Similar to diseases such as Hashimoto's thyroiditis and Graves' disease, T cells appear to play a major role in this process. Although the exact mechanism by which T cells are involved in the pathogenesis of IDDM has not been determined, three different mechanisms have been proposed. One mechanism involves molecular mimicry-activated T-cell proliferation. This mechanism is based on the assumption that epitopes of proteins expressed by infectious agents can be shared by unrelated molecules encoded by host genes. Therefore, the immune system mistakes these unrelated "self" molecules for epitopes of these infectious particles and incorrectly attacks cells expressing such epitopes on their cell surface. One particular antigen thought to be involved in this process in IDDM is GAD65. GAD65 has been found to share a six amino acid-long segment (PEVKEK) with the coxsackievirus B4 P2-C protein. However, T-cell clones from diabetics have failed to show reactivity to both GAD65 and coxsackievirus B4 P2-C in vitro (48). A second mechanism thought to trigger molecular mimicry-activated T-cell proliferation is "bystander" T-cell proliferation. This mechanism involves the stimulation of non-antigen-specific T cells by various cytokines during infection simply because they are in the area. The cytokines thought to be involved in this nonspecific stimulation are interferon alpha (IFN-a) and interferon beta (IFN-b) (49). A third theory involving a superantigen-mediated T-cell proliferation mechanism proposes that autoreactive T-cells can be inappropriately primed to react against self-structures through an encounter with a superantigen. Superantigens are proteins, or segments of proteins able to simultaneously bind the major histocompatibility complex (MHC) class II molecule expressed at the surface of an antigen-presenting cell and the T-cell receptor (TCR) molecule on the helper T cell. According to this theory, viral infections could influence development of IDDM through a superantigen effect by recruiting and activating T cells with autoreactive potential. With regard to T helper 1 vs. T helper 2 cell involvement in the pathogenesis of IDDM, early reports attributed a powerful diabetogenic effect to cytokines derived from TH1 cells, such as IFN-g (50). Type 2 cytokines, on the other hand, appear to be protective against the development of IDDM. Specifically, a recent study showed that the Type 2 cytokine IL-4 protected NOD mice from developing insulitis or diabetes. The potential role of a Type 2 cytokine deficiency in the pathogenesis of IDDM in humans was further demonstrated in another study involving diabetic siblings. In that study, T-cell clones from normal twins secreted both IFN-g and IL-4, whereas T-cell clones from diabetic twins secreted only IFN-g (51).

Type II Diabetes Mellitus

Type II or non-insulin-dependent diabetes mellitus (NIDDM) is one of the most common metabolic disorders in the United States. It is associated with hypercholesterolemia, atherosclerosis, hypertension, kidney disease, and coronary artery disease. Obesity is strongly related to Type II diabetes and insulin resistance as well. Insulin resistance- the hallmark of NIDDM- is the impaired ability of insulin to effect glucose uptake by cells. The etiology of insulin resistance is not fully understood but several studies have shown that systemic inflammation may be playing a significant role in the pathogenesis of this disorder (52). Serum C-reactive protein (C-RP) concentration has been shown to be elevated in elderly patients who subsequently developed impaired glucose tolerance or diabetes (53). Furthermore, the Women's Health Study demonstrated that women with elevated circulating levels of IL-6 and/or C-RP were at greater risk for the development of type-2 diabetes compared to women with low C-RP and/or Il-6 levels (54). Even when BMI and fasting insulin levels were accounted for, there was still a significant correlation between these markers of inflammation and the development of Type II DM (55). The adipose tissue appears to be a major source of circulating IL-6 in humans. Plasma IL-6 concentration correlates well with body mass index (56, 57). Plasma IL-6 has been shown to be elevated in obese people with insulin resistance (57, 58). IL-6 secretion by cultured human adipocytes correlates with markers of adipogenesis and is hormonally regulated (glucocorticoids suppress, whereas catecholamines and insulin stimulate IL-6 secretion) (59). In another study, IL-6 administration to 15 normal male volunteers induced dose-dependent increases in fasting blood glucose thought to be secondary to glucagon and other counterregulatory hormone releases (60).

TNF-a has received considerable attention with regard to insulin resistance as well and may be a key mediator of its pathogenesis. Adipose tissue-expressed TNF-a has been shown to be elevated in obese people with insulin resistance (58, 61). TNF-a appears to be an important mediator of insulin resistance in obese animals through its overexpression in adipose tissue (62, 63). When given chronically, TNF-a has been shown to cause insulin resistance in normal rats. Immunoneutralization of TNF-a by administration of a TNF receptor antibody resulted in a two- to threefold enhancement of peripheral glucose uptake in response to insulin in fa/fa rats (62). At the molecular level, chronic exposure of adipocytes to low doses of TNF-a led to a dramatic decrease in the insulin-stimulated autophosphorylation of the insulin receptor (IR) and the phosphorylation of insulin receptor substrate 1 (IRS-1) (64). Furthermore, IRS-1-mediated inhibition of insulin receptor tyrosine kinase activity in the presence of TNF-a has been demonstrated. Treatment of cultured murine adipocytes with TNF-a has been shown to induce serine phosphorylation of IRS-1 and convert IRS-1 into an inhibitor of the IR tyrosine kinase activity in vitro (65). The decrease in the catalytic efficiency of the IR correlated strongly with inhibition of insulin-stimulated glucose uptake. Hotamisligil et al. concluded that TNF-a plays an inhibitory role on the insulin-stimulated tyrosine kinase phosphorylation cascade. The question of whether TNF-a induces insulin resistance directly or indirectly through inhibitors of tyrosine kinase or counterregulatory hormones on muscle, fat, and liver in vivo needs further investigation. TNF-a has also been shown to down-regulate glucose transporter GLUT4 mRNA levels in adipocyte and myocyte cultures as well (62, 66, 67). However, gross quantitative regulation of glucose transporters in muscle, the major site for glucose disposal, has not been studied in patients with NIDDM.

TNF-a may also be playing a role in the hyperlipidemia observed in Type II DM. TNF-a has been shown to have profound effects on whole body lipid metabolism (68-70). Circulating triglycerides and very low density lipoproteins in rats and humans are increased after administration of TNF-a (69, 70). In addition to TNF-a, IL-1 and IFN-g also stimulate fatty acid synthesis. IL-6 influences fat metabolism as well. Studies have suggested that certain polymorphisms in the promoter region of the IL-6 gene can affect lipid levels through changes in IL-6 gene transcription and ultimately IL-6 production (71). IL-6 has been proposed to cause an increase in circulating lipid levels probably through a decrease in peripheral lipoprotein lipase activity (72). However, the effects of IL-6 on lipid metabolism are much more complex. Administration of recombinant human IL-6 to healthy volunteers was followed by a reduction in circulating triglyceride levels in normal volunteers. In the same study, total cholesterol as well as apoprotein B levels were decreased after administration of IL-6. IL-6 levels correlated negatively with cholesterol levels after major surgery (73, 74), in patients undergoing hemodialysis (75), or after myocardial infarction (MI) (76). IL-6 may cause the acute decrease in cholesterol after major surgery or myocardial infarction perhaps by decreasing hepatic lipoprotein production (77). Thus chronic exposure to high IL-6 levels may lead to hyperlipidemia, whereas acute exposure may have the opposite effects on serum lipid levels.

CALCIUM METABOLISM AND THE IMMUNE SYSTEM

Osteoporosis

Osteoporosis is a disease of bone metabolism typically occurring in postmenopausal women, in which bone resorption is favored over bone formation. The two cells involved in dynamic bone metabolism are osteoblasts and osteoclasts. Osteoblasts function to lay down new bone and osteoclasts function to reabsorb old bone. Cytokines have been implicated in the pathogenesis of osteoporosis. Osteoblasts produce the cytokines IL-6 and IL-11 (78). In addition, many cytokines such as IL-1, IL-3, IL-6, IL-11, and TNF along with colony stimulating factors are implicated in osteoclast development (79-82).

IL-6 plays a major role in osteoclast development and function. IL-6 is produced by both stromal cells and osteoblastic cells in response to stimulation by systemic hormones such as parathyroid hormone (PTH), PTH-related peptide (PTH-RP), thyroid hormones and 1,25-dihydroxyvitamin D3. Growth factors involved in resorption, such as transforming growth factor-beta (TGF-b), and other cytokines- like IL-1 and TNF- increase IL-6 production as well (83). IL-6 has been shown to stimulate osteoclast formation and bone resorption in fetal mouse bone in vitro (84, 85) and, along with IL-1, it stimulates bone resorption in vivo (86). Evidence has been shown to suggest that IL-6 influences not only immature osteoclasts but mature osteoclasts as well. Human osteoclastoma cells have receptors for IL-6, which stimulates the resorptive activity of these cells (87). Furthermore, IL-6 has been shown to play a role in the abnormal bone resorption observed with multiple myeloma (88), Paget's disease (89), rheumatoid arthritis (90), and Gorham-Stout disease (91). Effects of increased osteoclast-induced bone resorption are not solely reserved to IL-6, but IL-6 family cytokines- such as leukemia-inhibitory factor (LIF) - as well. It appears that LIF acts on osteoclasts indirectly via stimulating IL-6 release by osteoblasts, resulting in an increase in bone resorption (92).

In postmenopausal osteoporosis, there is a dramatic increase in the number of osteoclasts with the decline of circulating estrogens. IL-6 is again thought to play a significant role in the pathogenesis of this process. Estrogens have been shown to inhibit the production of IL-6 from cultured bone marrow stromal and osteoblastic cell lines (93). The inhibitory effect of estrogen on IL-6 production is mediated through inhibition of IL-6 gene transcription via an estrogen-receptor-mediated effect on the transcriptional activity of the proximal 225-bp sequence of the promoter (94, 95). Studies on IL-6 knockout mice (i.e. mice lacking IL-6) have demonstrated that the presence of IL-6 is essential for the development of postmenopausal osteoporosis. Unlike wild-type mice, IL-6 knockout mice did not show cellular changes in the marrow and trabecular bone sections and were protected from the loss of trabecular bone after the loss of sex steroids (96, 97).

TNF-a has also been shown to influence bone resorption involved in postmenopausal osteoporosis. One study revealed that augmented production of TNF-a by T-cells following ovariectomy led to increased macrophage stimulating factor- (M-CSF) and RANK ligand-induced osteoclastogenesis (98). Furthermore, ovariectomy failed to induce bone loss or stimulate osteoclastogenesis in T-cell deficient nude (without a thymus) mice. The effects of TNF-a on bone resorption were then further investigated in TNF knockout mice. Ovariectomy was found to induce rapid bone loss in wild-type mice but failed to do so in TNF-a-deficient (TNF-/-) mice. In addition, adoptive transfer of wild type (wt) T cells to T cell-deficient nude mice restored the bone loss induced by ovariectomy in these mice, whereas reconstitution with T cells from TNF-/- mice failed to do so (99). The authors concluded that bone marrow T cells (through production of TNF-a) may be playing a major role as mediators of the bone-wasting effects of estrogen deficiency in vivo. The data on IL-6 and TNF-a knockout mice are intriguing. It appears that lack of either TNF-a or IL-6 is protective against the development of osteoporosis due to sex-steroid deficiency. It is plausible that TNF-a stimulates IL-6, which in turn mediates osteoclast bone resorption. TNF-a stimulates production of IL-6 by several cell types, including bone marrow stromal cells and osteoblasts (100). IL-6- on the other hand- does not stimulate TNF-a secretion (101). Therefore TNF-a may be exerting some of its actions through IL-6, which may explain why both TNF-a and IL-6 knockout mice show decreased osteoclast-induced bone resorption.

Estrogen has been shown to decrease TNF-a gene expression by blocking Jun NH(2)-terminal kinase (JNK) activity (102). By blocking JNK, there is decreased phosphorylation of c-Jun and JunD- which are nuclear transcription factors that along with AP-1 help form the initiation complex at the TNF promoter site. This results in the subsequent inhibition of TNF-a gene transcription. It appears that an intact p55 TNF receptor-I is necessary for estrogen deficiency-induced osteoporosis. In a recent study, ovariectomy failed to decrease the bone mineral density of P55 knockout mice (99).

Besides having a direct role on osteoclastogenesis, TNF-a has been shown to inhibit osteoblast differentiation as well. The transcription factor RUNX2 is a critical regulator of osteoblast differentiation. A recent study showed that incubation of fetal calvarial precursor cells with TNF-a led to a 50-90% reduction in the expression of all isoforms of RUNX2 mRNA (103).

Besides influencing osteoporosis due to estrogen deficiency by itself, TNF-a may also influence other cytokines to mediate osteoclast-induced bone resorption. IL-1 is one cytokine that may be influenced by TNF-a. IL-1 plays a role in estrogen deficiency-mediated osteoporosis by stimulating osteoclast activity through direct targeting of mature osteoclasts rather than by stimulating osteoclastogenesis like TNF-a does. Studies have shown that TNF-a neutralization blocks IL-1 production whereas IL-1 neutralization does not block TNF-a. Thus it is likely that increased bone marrow levels of IL-1 observed in ovariectomized mice are a result of increased TNF-a production. TNF-a and IL-1 have also been shown to induce production of IL-7 (104). IL-7 has been shown to influence osteoclastogenesis through T cells by RANK-ligand-dependent and independent mechanisms.

It appears that IL-6 and IL-8 play a major role in thyrotoxicosis-induced osteoporosis as well. Besides the effects of IL-6 on bone resorption already mentioned, IL-8 appears to play a role in bone resorption as well with IL-8 receptors being expressed by osteoclasts (105). A rise in IL-6 and IL-8, but not in IL-1b, IL-11, or TNF-a, in patients with thyrotoxicosis due to Graves' disease (Gd) or toxic multinodular goiter (TMNG) has been observed (106). In the same study, patients with thyroid carcinoma (CA) on TSH suppressive therapy had significantly raised circulating levels of IL-6 and IL-8 compared to controls. In both groups of patients plasma levels of IL-6 and IL-8 correlated with serum T3 and free T4 concentrations but not with the bone turnover markers, namely urinary deoxypyridinoline (Udpd) or bone-specific alkaline phosphatase (b-aLP). Both IL-6 and IL-8 have also been shown to be released by human bone marrow stromal cell cultures containing osteoblast progenitor cells in response to T3 (105). Therefore, the increased plasma concentrations of IL-6 and IL-8 seen in thyrotoxicosis are most likely caused by T3 stimulation of bone osteoblasts despite the inability of bone markers to correlate with acute changes in thyroid hormone status.

Bone resorption in primary hyperparathyroidism appears to be mediated by cytokines as well. Cytokines such as IL-6, interleukin-1b, and TNF-a had been examined with respect to their relationship to biochemical markers of bone turnover in patients with primary hyperparathyroidism, hypoparathyroidism, normal controls, and in patients who had undergone parathyroid adenectomy. IL-6, IL-6 soluble receptor (IL6-sR), and TNF-a were all found to be elevated in patients with primary hyperparathyroidism. After parathyroid adenectomy, the circulating levels of each of these cytokines fell into the normal range. The hypoparathyroid patients had lower levels of IL-6, IL-6sR, and TNF-a compared to normal controls (107). The cytokine that especially correlated with serum levels of PTH and with biochemical markers of bone resorption- such as serum deoxypyridinoline, type I collagen carboxyterminal telopeptide, urinary pyridinoline, and urinary deoxypyridinoline- was circulating IL-6. Multiple regression analysis further confirmed that IL-6 and not TNF-a was independently predictive of bone resorption in these patients. (107).

Besides having a direct action on bone resorption, IL-6 may also have an indirect action on bone resorption in hyperparathyroidism by its influence on IL-11. Both IL-6 and IL-11 belong to a family of cytokines that shares the use of gp130 receptor for intracellular signaling (108, 109). In vitro studies have further established that IL-11 plays a prominent role in bone homeostasis. Like IL-6, IL-11 is also produced by osteoblasts in response to osteotropic factors such as PTH and 1,25(OH)2-vitD. In addition, IL-11 has been found to inhibit osteoclast apoptosis and stimulate osteoclast formation (110-112). Although PTH stimulates the production of both cytokines by human osteoblast-like cells, IL-6 has been found to be elevated in hyperparathyroidism, while IL-11 levels were significantly reduced. Also, after parathyroidectomy, circulating levels of IL-6 dropped while those of IL-11 increased. When PTH was infused into rodents, there was a significant decline in mean circulating levels of IL-11, whereas IL-6 levels increased. Furthermore, pretreatment of cells with neutralizing serum to IL-6 enhanced PTH-induced IL-11 production, compared with the effect of pretreatment with non-immune sera. These data indicate that IL-6 negatively regulates IL-11 production in vivo and in vitro. Analysis of steady-state mRNA levels in SaOS-2 cells indicated that this effect is posttranscriptional. Since both IL-6 and IL-11 stimulate osteoclast formation, down-regulation of IL-11 by IL-6 may help modulate the resorptive response to PTH (113).

EFFECTS OF THE IMMUNE SYSTEM ON THE STRESS SYSTEM

The HPA Axis

A two-way communication exists between the hypothalamic pituitary-adrenal axis and the immune system. The HPA axis is activated in states of inflammation or infection (Figure 5). This activation is mediated by the inflammatory cytokines TNF-a, IL-1, and IL-6, which are secreted in tandem in response to various infectious and noninfectious stimuli. The inflammatory cytokines are produced by a variety of cells, including monocytes, macrophages, astrocytes, endothelial cells, and fibroblasts. During inflammatory states, these cytokines activate the HPA axis through stimulation of the CRH neurons of the parvocellular nucleus in the hypothalamus (114, 115). The stimulation of the hypothalamic-pituitary unit leads to secretion of cortisol by the adrenal glands, resulting in termination of the inflammatory cascade. Although all three inflammatory cytokines have the capacity to activate the HPA axis, it appears that IL-6 is the critical component of this cascade. Studies in rats have demonstrated that immunoneutralization of IL-6 abolishes the effects of the other two cytokines on the HPA axis (116). TNF-a and IL-1 stimulate the production of IL-6 and IL-6 in turn stimulates the HPA axis. While acute stimulation with IL-6 stimulates the HPA axis through activation of the hypothalamic CRH neurons, chronic exposure to IL-6 can stimulate directly the corticotroph cells of the pituitary and the adrenal cells as well.


Figure 5. Inflammation, trauma, or acute illness activate the hypothalamic-pituitary-adrenal axis. Both IL-6 and cortisol produce the ESS.

Glucocorticoids appear to inhibit IL-6 secretion at the transcriptional level through interaction of the ligand-activated glucocorticoid receptor with nuclear factor-Kappa B (NF-kB). This demonstrates that glucocorticoids and IL-6 participate in a feedback loop, in which IL-6 stimulates glucocorticoid release and glucocorticoids subsequently feedback and inhibit IL-6 release. This explains the inverse relationship with diurnal variation between circulating IL-6 and glucocorticoid levels (117). Furthermore, acute hypocortisolism has been shown to result in a four- to five-fold elevation of circulating IL-6 and TNF-a levels. In a study involving patients with Cushing disease studied before and after transsphenoidal adenomectomy, cytokines were measured during the hypercortisolemic, hypocortisolemic, and eucortisolemic (while patients were on glucocorticoid replacement) states (118). While the patients were hypocortisolemic, plasma IL-6 concentration rose, while they experienced symptoms of glucocorticoid deficiency, which are part of the "steroid withdrawal syndrome". This syndrome consists of pyrexia, headache, anorexia, nausea, fatigue, malaise, arthralgias, myalgias, and somnolence of variable degree. Interestingly, IL-6 levels did not rise in patients who did not become hypocortisolemic after surgery (and did not develop symptoms consistent with the withdrawal syndrome), indicating that hypocortisolism was necessary for the rise in IL-6. Glucocorticoid replacement was followed by a dramatic decrease of IL-6 levels, which was concomitant with relief of the observed symptoms. Another study involving administration of IL-6 to humans reproduced the symptomatology associated with the steroid withdrawal syndrome, rendering further support to the notion that IL-6 plays a central role in the development of symptoms of adrenal insufficiency (119).

Similar to IL-6, LIF can stimulate the hypothalamic pituitary axis. LIF is a multifunctional cytokine of the IL-6 cytokine family, sharing the common gp130 receptor subunit together with IL-6, interleukin-11, oncostatin-M, ciliary neurotrophic factor and cardiotrophin-1. Both LIF and its receptor have been found to be expressed in the pituitary gland during development (120). Furthermore, LIF-binding sites (LIFR) have been found on one third of ACTH-positive cells and approximately 20% of GH-positive cells of the pituitary. In several tissues, LIF, LIFR, and gp130 mRNA expression is stimulated by various inflammatory stimuli, whereas LIF gene expression is negatively regulated by glucocorticoids. LIF stimulates ACTH secretion in vitro and in vivo. Murine corticotroph AtT-20 cells exhibit a 2-4 fold increase in ACTH secretion during incubation with 1 nM LIF for 24 hours. Incubation of AtT-20 cells with corticotropin-releasing hormone (CRH) results in a 3 to 7 fold rise in ACTH secretion. Combined with LIF, CRH causes a further 2 to 3 fold increase of ACTH secretion in comparison to treatment with CRH alone.

The Autonomic Nervous System

Catecholamines have been found to stimulate IL-6 secretion through a beta-adrenergic mechanism in rats (121). IL-6 has been shown to play a role in the stress response to exercise in humans as well. In one study involving high-intensity treadmill exercise tests on 15 male volunteers, the subjects were given placebo, hydrocortisone or dexamethasone before starting their exercise. Plasma epinephrine and norepinephrine concentrations peaked at 15 minutes after the start of exercise, whereas plasma IL-6 concentrations peaked twice at 15 and at 45 minutes after the onset of the test run. There was no difference in either the epinephrine or norepinephrine peaks among the three treatments, but the net area under the curve for IL-6 was smaller after pretreatment with hydrocortisone or dexamethasone than after pretreatment with placebo. A positive correlation was observed between peak plasma epinephrine and norepinephrine levels and IL-6 plasma levels at 15 minutes (122). This demonstrates that IL-6 secretion is likely stimulated during exercise by catecholamines, whereas exogenous glucocorticoids attenuate this effect without affecting the catecholamine levels.

Go back to ADRENAL PHYSIOLOGY AND DISEASES