| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Excellence Center for Research, Transfer and High Education De Novo Therapies (M.R., S.R., M.S., P.R.), University of Florence, 50121 Florence, Italy; and Unit of Internal Medicine and Endocrinology (M.R., L.C.), Istituto Superiore per la Prevenzione e Sicurezza del Lavoro Laboratory for Endocrine Disruptors, Fondazione Salvatore Maugeri, Istituto di Ricovero e Cura a Carattere Scientifico, Chair of Endocrinology, University of Pavia, 27100 Pavia, Italy
Correspondence: Address all correspondence and requests for reprints to: Mario Rotondi, M.D., Ph.D., Unit of Internal Medicine and Endocrinology, Istituto Superiore per la Prevenzione e Sicurezza del Lavoro Laboratory for Endocrine Disruptors, Fondazione Salvatore Maugeri, Istituto di Ricovero e Cura a Carattere Scientifico, Via S. Maugeri 4, 27100 Pavia, Italy. E-mail: mrotondi{at}fsm.it
 |
Abstract |
|---|
(IFN-) inducible chemokines (CXCL9, CXCL10, and CXCL11) and their receptor, CXCR3, play an important role in the initial stage of autoimmune disorders involving endocrine glands. The fact that, after IFN- stimulation, endocrine epithelial cells secrete CXCL10, which in turn recruits type 1 T helper lymphocytes expressing CXCR3 and secreting IFN-, thus perpetuating autoimmune inflammation, strongly supports the concept that chemokines play an important role in endocrine autoimmunity. This article reviews the recent literature including basic science, animal models, and clinical studies, regarding the role of these chemokines in autoimmune endocrine diseases. The potential clinical applications of assaying the serum levels of CXCL10 and the value of such measurements are reviewed. Clinical studies addressing the issue of a role for serum CXCL10 measurement in Graves’ disease, Graves’ ophthalmopathy, chronic autoimmune thyroiditis, type 1 diabetes mellitus, and Addison’s disease have been considered. The principal aim was to propose that chemokines, and in particular CXCL10, should no longer be considered as belonging exclusively to basic science, but rather should be used for providing new insights in the clinical management of patients with endocrine autoimmune diseases.
-inducible CXC chemokines and their receptor CXCR3 agonists in vitro inhibit CXCL10 production induced by proinflammatory cytokines |
I. Introduction |
|---|
inducible chemokines, which interact exclusively with the chemokine receptor CXCR3. All chemokines possess the ability to attract and recruit distinct types of cells in different organs or tissues. To exert such a function, many chemokines are released into the bloodstream, where they can be detected and also quantitated. Chemokines exhibit their peculiar function of attraction and recruitment of different cell types during physiological processes of maturation and trafficking of immune cells throughout different lymphoid organs, but they also play an important role in inducing, maintaining, and amplifying the inflammatory reactions. Therefore, the ability of chemokines to attract and recruit different immune cells in inflamed tissues is important for the protection against infectious agents. However, chemokines can also have a dangerous effect for the body by maintaining and amplifying chronic inflammatory reactions, when the invading agent cannot be rapidly removed or neutralized, as well as by sustaining chronic immune responses against self-antigens which are responsible for autoimmune diseases. For this reason, the assessment of chemokines in inflamed tissues may help with understanding the pathophysiological mechanisms involved in these disorders. More importantly, chemokines are produced in the inflamed tissue by both infiltrating and resident cells, with a strict relationship with the phases of inflammation. The enhanced production of chemokines in the inflamed tissue(s) and the relative blood flow of the inflamed district are both responsible for the increased concentrations of the same chemokines in serum and other biological fluids. Therefore, it is reasonable to think that at least in some diseases, the detection and quantitation of chemokines in biological fluids may provide a useful tool for monitoring the phase and the severity of the disease.
This review will be focused on the possible role of the so-called CXCR3-binding chemokines in autoimmune endocrine disorders. The main reason for this choice stems from the fact that recently CXCR3-binding chemokines were extensively investigated and were found to exhibit strong variations both in inflamed tissues and in the serum during the different phases of autoimmune endocrine diseases. This is probably due to the fact that the production of all CXCR3-binding chemokines by resident cells is stimulated by IFN-. IFN- also induces the local recruitment of inflammatory cells, which express the CXCR3 receptor and are, in turn, able to produce IFN-. This sequence of events results in an increased production of the same group of chemokines, thus establishing an important loop for the maintenance and amplification of inflammatory reactions. Therefore, CXCR3-binding chemokines probably play a pathogenic role in autoimmune endocrine disorders by influencing the development and/or by amplifying the inflammatory process responsible of these diseases. Moreover, due to their increase in biological fluids and to the variations of their levels according to the different phase of the disease, the measurement of CXCR3 chemokines in the serum may represent a useful tool for monitoring the activity of the inflammatory process.
|
II. The Chemokines |
|---|
-granules. Subsequently, from 1984 through 1989, cDNAs for structurally related proteins, including IFN--induced protein 10 (IP-10) (2), JE (3), IFN--induced monokine (Mig) (4), regulated on activation, normal T cell expressed and secreted (RANTES) (5), I-309 (6), KC (7), and macrophage inflammatory protein-1 (MIP-1) (8), were cloned by investigators looking for cell differentiation- and activation-associated genes. Thus, the existence of a gene family was established before identifying their functions (9, 10, 11). The discovery of the neutrophil-targeted chemokine IL-8 represented a landmark in immunology because it was the first leukocyte subtype-selective chemoattractant to be found (12, 13). The discovery of IL-8 also promoted the search for functions of other chemokines on leukocyte chemotaxis as well as the discovery of new family members. The interest in the field grew with the subsequent reports of macrophage chemotactic protein CCL2, CCL5, and CCL11, the first important chemokines active on monocytes, T cells, and eosinophils, respectively (14, 15, 16, 17). As the number of family members expanded, various short-lived collective terms were used, including "the platelet factor (PF)-4 family" (9), "the small inducible cytokine family" (10), and "the intercrines" (11). Finally, in 1992 at the Third International Symposium on Chemotactic Cytokines in Baden, Germany, the term "chemokine," a short neologism for "chemotactic cytokines," was coined and accepted as standard (18). The nomenclature for chemokines is based on the configuration of a conserved amino-proximal cysteine-containing motif. Based on this system, there are currently four branches of the chemokine family: CXC, CC, CX3C, and C (where X is any amino acid) (Table 1
) (19, 20). The transmission of chemokine-encoded messages is mediated by specific cell-surface G protein-coupled receptors with seven transmembrane domains. At present, the human chemokine receptor system consists of 20 different receptors (Table 1). In 2000, a new nomenclature system for chemokines and chemokine receptors was approved by the Nomenclature Committee of the International Union of Pharmacology (NC-IUPHAR) (Table 1) (21).
|
B. The CXC chemokine family
CXC chemokines have four conserved cysteines and are distinguished by the presence of one amino acid between the first and second cysteine. The CXC chemokine subfamily includes 14 different members whose encoding genes are clustered on human chromosome 4, with few exceptions (22). Most members of the CXC chemokine family exhibit chemotactic properties toward neutrophils and lymphocytes, and are unique in that they constitute a family showing positive or negative activity on the control of angiogenesis (23). CXC chemokines can be further divided into two groups (ELR+ and non-ELR) according to the presence or absence of the tripeptide motif glutamic acid-leucine-arginine (ELR) N-terminal to the first cysteine residue. Interestingly, as shown by site-directed mutagenesis, the presence or the absence of an ELR motif in the chemokines-amino acid sequence seems to correlate with their angiogenic or angiostatic activity, respectively (24). Thus, ELR+ CXC chemokines have been linked to angiogenesis (25, 26), whereas the ELR- CXC chemokines, including CXCL4, CXCL9, CXCL10, and CXCL11, antagonize angiogenesis (23). Furthermore, ELR-CXC chemokines, such as CXCL13, CXCL9, CXCL10, and CXCL11, are powerful chemoattractants for lymphocytes. Recently, a novel CXC chemokine receptor with angiogenic potential was identified and named as CXCR7 (27).
Another classification scheme was based on the function of chemokines and their expression pattern. According to these criteria, two groups of chemokines were identified. The first group includes the so-called inflammatory/inducible chemokines, which are regulated by proinflammatory stimuli, such as lipopolysaccharide and primary cytokines (i.e., IL-1 and TNF-), and orchestrate innate and adaptive immune responses. Inflammatory chemokines control the recruitment of effector leukocytes in infection and inflammation sites, in tissue injuries, and in tumors.
The second group includes the homeostatic/constitutive chemokines, which are important in lymphocyte and dendritic cell trafficking, and in immune surveillance. Homeostatic chemokines navigate leukocytes during hematopoiesis in the bone marrow and thymus; during initiation of adaptive immune responses in the spleen, lymph nodes, and Peyer’s patches; and during immune surveillance in healthy peripheral tissues (19).
The finding that several chemokines cannot be assigned unambiguously to either one of the two functional categories led to the characterization of a third group of chemokines, which were referred to as "dual-function" chemokines (19, 20). Dual-function chemokines participate in immune defense functions (i.e., are up-regulated under inflammatory conditions) and also target noneffector leukocytes, including precursor and resting mature leukocytes, at sites of leukocyte development and immune surveillance. Many dual-function chemokines are highly selective for lymphocytes and have a role in T cell development in the thymus, as well as in T cell recruitment to inflammatory sites.
Genes encoding for inflammatory CXC chemokines are typically found in a major cluster on human chromosome 4, whereas genes for homeostatic chemokines are located alone or in small clusters on different chromosomes (23).
Studies on the expression of chemokines in different species showed that none of the mammalian CXC chemokines, except CXCL12 and CXCL14, possesses orthologs in any other vertebrate class, including birds. This finding suggests that the fine regulation of inflammatory responses is a recent acquisition in the evolution. Indeed, some orthologs of human CXC chemokines are not represented even in mice (28).
The role of CXC chemokines in several types of inflammatory and autoimmune disorders has been largely investigated and was recently reviewed (20). Converging evidence suggests that a subgroup of CXC chemokines, sharing binding to the same receptor, CXCR3, play a role in the pathogenesis of endocrine autoimmune diseases. Thus, this review will focus mostly on the role of the chemokine receptor CXCR3 and its binding chemokines in endocrine autoimmune diseases.
The main messages of this section are:
C. The IFN--inducible CXC chemokines and their receptor CXCR3
Three CXC chemokines were found to share the property to be induced by IFN-. They were initially called "IFN--inducible protein 10" (IP-10) (2), Mig (4), and IFN- inducible T cell chemoattractant (I-TAC) (29). In the new nomenclature, the three chemokines were named as CXCL10, CXCL9, and CXCL11, respectively (21).
All three chemokines were found to bind a unique receptor named CXCR3, which was discovered in 1995 on a genomic clone isolated by PCR-based homology hybridization. The gene was named GPR9, was originally incorrectly mapped to human chromosome 8p11.2–12 (30), and was later correctly mapped to chromosome Xq13 (31). The rank order of binding affinity is CXCL11 > CXCL10 > CXCL9. Initially, CXCR3 was found to be expressed on a subset of circulating T cells, B cells, and natural killer cells, and among T cells, mainly on type 1 T helper (Th1) cells (32, 33). In subsequent studies, it was found that CXCR3 was expressed not only by immune cells, but also by resident cells (34) such as human mesangial cells (35), human liver stellate cells, vascular pericytes (36), and human microvascular endothelial cells (37).
More recently, a distinct, receptor, deriving from an alternative splicing of the CXCR3 gene was identified and named as CXCR3-B (38). CXCR3-B not only binds CXCL10, CXCL9, and CXCL11, but also acts as functional receptor for the orphan CXC-chemokine CXCL4, which exclusively interacts with CXCR3-B. The interaction of chemokines with CXCR3 mediates their chemotactic and immune effects, whereas the binding to the splicing variant CXCR3-B accounts for their angiostatic effect (38). To add to the complexity of CXCR3 biology, another variant of human CXCR3 has been identified, which is generated by posttranscriptional exon skipping. This receptor was named CXCR3-alt and binds CXCL11, but its biological role is still unknown (39). The main aim of this review is to discuss the role of IFN- inducible chemokines (CXCL9, CXCL10, and CXCL11) and their classic CXCR3 receptor; therefore, the biological effects resulting from the interaction between the alternative variant of CXCR3 and their ligands will be limited to their angiostatic effects.
The main messages of this section are:
. |
III. Main Biological Actions of CXCR3-Binding Chemokines |
|---|
T cells were originally divided into two main subsets which are named as CD4+ T helper (Th) and of CD8+ cytotoxic T (Tc) cells. Subsequently, two different types of CD4+ Th cells, known as type 1 Th (Th1) and type 2 Th (Th2), were recognized (Fig. 1). Th1 cells produce cytokines, such as IL-2, IFN-, and lymphotoxin-, which result in the activation of macrophages, in the production of complement-fixing and -opsonizing antibodies, and also in cytotoxicity (41). By contrast, Th2 cells have been thought to play a regulatory rather than protective role, inasmuch as cytokines produced by these cells (i.e., IL-4 and IL-13) have an inhibitory effect on the production of Th1 cytokines, as well as on several functions of activated macrophages (41). It should be noted, however, that Th1 and Th2 cells do not represent clearly distinct lineages of Th cells, as CD4+ and CD8+ T cells, but extremely polarized forms of a much more heterogenous Th cell response. Moreover, their phenotype of cytokine production in humans is not always so clearly polarized as in mice.
|
The fact that Th1 cells produce IFN-, which induces the production by different cell types of CXCL9, CXCL10, and CXCL11, and that these chemokines in turn can attract and recruit Th1 cells, suggests the existence of a loop between IFN--producing Th1 cells and resident cells producing CXCR3-binding chemokines (42). Th2 cells express different chemokine receptors, such as CCR4 and CCR8, thus being recruited in target tissues by CCL17, CCL22 (both ligands for CCR4), and CCL1 (ligand of CCR8). Based on these findings, it can be proposed that chemokines interacting with T cells via CXCR3 may induce a recruitment of Th1 cells into the inflamed tissues. On the other hand, chemokines interacting with different chemokine receptors on T cells may recruit Th2 cells, which are responsible for allergic inflammation.
Further studies support the concept that the role of CXCR3-binding chemokines in the regulation of the immune response goes far beyond their powerful chemotactic activity on activated lymphocytes. A large body of experimental evidence emphasizes the role of CXCL10 in the initiation and amplification of host alloresponses (43). CXCL10-deficient mice have impaired T cell responses, impaired contact hypersensitivity, and limited inflammatory cell infiltrates. They are also unable to control viral infections (44). CXCL10 not only mediates leukocyte recruitment, but also drives T cell proliferation to allogenic and antigenic stimulation and IFN- secretion in response to antigenic challenge (42). Accordingly, CXCL10 up-regulates the production of Th1 cytokines and down-regulates the production of Th2 cytokines (45). The final result is a strong up-regulation of inflammatory reactions characterized by the production of IFN-, thus exerting important protective activity against infections sustained by intracellular bacteria and some viruses, which is provided by Th1 cells (Fig. 2). This also results in a down-regulation of allergic inflammation that is provided by Th2 responses.
|
.
B. Angiogenesis
CXCR3-binding chemokines are powerful inhibitors of angiogenesis (23). The major receptor mediating the angiostatic effect of CXC chemokines is CXCR3-B. CXCR3-B is expressed in human microvascular endothelial cells (37, 46) during the late S-phase of the cell cycle on through mitosis, representing the first example of a chemokine receptor expression linked to a particular phase of the cell cycle (37). In vivo, the expression of CXCR3-B in small vessels (37, 47, 48), is higher in inflamed and neoplastic tissues compared with normal tissues (37).
Angiostatic CXC chemokines were shown to inhibit angiogenesis in several experimental models (49, 50, 51) and to participate in the control of angiogenesis during physiological repair of tissue injury (52). CXCL9 and CXCL10 are specifically expressed during the late phase of wound healing repair, to help prevent unlimited vessel growth without blocking other repair processes involved in wound healing (23).
CXCR3-binding chemokines are also involved in the pathogenesis of proliferative diabetic retinopathy (53). The levels of CXCL10 were found to be significantly higher in vitreous samples from patients with inactive, compared with active, proliferative diabetic retinopathy. This suggests that decreased levels of this angiostatic chemokine might favor retinal angiogenesis during diabetic retinopathy (53).
Overall, the angiostatic effect of CXCR3-binding chemokines is strictly dependent upon the expression of CXCR3-B, the alternatively spliced form of the classic CXCR3 receptor (38). Yet, the expression of CXCR3-B has not been evaluated in either normal or pathological endocrine glands.
The main messages of this section are:
|
IV. CXCR3-Binding Chemokines in Healthy Subjects and in Nonendocrine Immune-Mediated Pathological Conditions |
|---|
is observed. Commercial ELISA kits for the measurements of CXCL9 and CXCL11 are also available.
A. CXCR3-binding chemokines in healthy subjects
The circulating concentrations of CXCR3-binding chemokines in humans have been extensively studied only for CXCL10, both in health and disease, whereas data regarding the serum levels of the other two chemokines (CXCL9 and CXCL11) are still limited. Studies performed in large series of healthy adult subjects found mean serum levels of CXCL10 ranging from 70 to 90 pg/ml (54, 55). Variations between healthy subjects may be estimated by SD values of approximately 50 pg/ml (54, 55). These figures are comparable to those reported in the normal subjects used as controls in clinical studies investigating CXCL10 in patients with endocrine and nonendocrine diseases (which will be quoted when discussing the specific disease). Nevertheless, it should be noted that in the latter studies healthy subjects were rarely screened for circulating autoantibodies, and therefore it is not always possible to exclude the presence of subclinical autoimmune disorders which, at least in some cases, might have affected the serum concentrations of CXCL10. It is now known that in euthyroid chronic autoimmune thyroiditis (CAT), the most frequent subclinical autoimmune condition in humans, especially in middle-aged women, the serum levels of CXCL10 are significantly increased compared with healthy controls proven to be negative for thyroglobulin (Tg) and thyroid peroxidase (TPO) antibodies (Ab). Euthyroid CAT might represent the most frequently undetected condition biasing the results of CXCL10 in apparently healthy subjects. Many other, less easily detectable abnormalities (e.g., other subclinical autoimmune diseases) may lead to similar problems. The question arises as to how healthy subjects should be selected when comparing their serum CXCL10 levels to those observed in a specific pathological condition. To complicate the issue further, it should be noted that there are currently no data regarding fluctuations of CXCL10 in the serum of individual healthy subjects. As far as our current knowledge permits, we will try to define some physiological variables, which were found to influence the serum levels of CXCL10.
1. The role of gender.
Clinical studies, in which healthy subjects were screened for excluding subclinical thyroid abnormalities by means of thyroid ultrasound (US) and tests for circulating Tg Ab and TPO Ab, found no gender-related differences in circulating concentrations of CXCL10. In the absence of such a screening for subclinical autoimmune conditions, higher serum levels of CXCL10 might be expected in females, due to the well-known greater prevalence of autoimmunity in women than in men. In clinical studies, the potential bias resulting from gender-related differences may be reduced, at least in part, by performing a strict sex-matching between the subjects to be evaluated.
2. The role of age.
An age-related dysregulation of the immune system has been extensively reported by studies performed in humans and experimental animal models of aging (56, 57, 58, 59). The influence of aging on circulating concentrations of CXCL10 was evaluated in two clinical studies (54, 55). Healthy subjects aged from 10 to 80 yr, proven to be negative for circulating thyroid antibodies and with no evidence of other autoimmune diseases, were studied. In a multiple linear regression model including age, body mass index (BMI), systolic and diastolic blood pressure, glycemia, total high-density and low-density lipoprotein cholesterol, triglycerides, TSH, Tg Ab, and TPO Ab, only age was significantly related to serum levels of CXCL10, a positive correlation being found between the two variables.
3. The role of body weight and BMI.
Although the serum levels of CXCL10 have not been specifically investigated in obesity, studies evaluating healthy subjects reported no change in serum concentrations of CXCL10 in relation to BMI (54). The issue remains open because patients with morbid obesity were not studied.
4. Practical points.
The above-described physiological changes in the serum levels of CXCL10 must be taken into account when this chemokine is measured in different pathological conditions. In diseases such as hepatitis C virus (HCV) hepatitis, primary biliary cirrhosis, or end stage renal diseases, the serum levels of CXCL10 are extremely high; thus, the comparison with healthy subjects might not be biased by gender or age. On the contrary, when studying pathological conditions in which the serum levels of CXCL10 are significantly, but only slightly higher than in healthy controls, such as endocrine autoimmune diseases, the above factors must be considered to reduce potential sources of error. As a consequence, matching patients and controls for gender and age appears critical to avoiding misleading results.
B. CXCR3-binding chemokines in some immune-mediated pathological conditions
In this section, we will briefly review the major findings obtained in nonendocrine diseases, which are either crucial for a better comprehension of the role of chemokines in human pathology or exemplify the applications of their assays in the clinical practice. Consistent with the aim of this review, description of data from basic studies will be limited to essential information, whereas findings obtained in clinical studies will be more extensively described. The latter data support the view that measuring CXCL10 in the serum is useful in several nonendocrine diseases, both autoimmune and nonautoimmune, such as chronic HCV hepatitis. In different clinical settings, the serum levels of CXCL10 proved to be useful as an index predicting the course and severity of the disease, as a marker of its activity, as a predictor of treatment outcome, and as a parameter for choosing the best therapeutic option. The fact that autoimmunity is responsible for most endocrine disease indicates that the results obtained by assaying CXCL10 in autoimmune nonendocrine disorders might be transferred to endocrinopathies. This remains to be done in that, as we will see in the subsequent sections, few clinical studies have been performed so far in endocrine patients.
1. CXCR3-binding chemokines in HCV-induced chronic hepatitis.
The first published clinical study in which the serum levels of CXCL10 were evaluated in human disease is the investigation by Narumi et al. (60), reporting significantly increased circulating concentrations of CXCL10 and CCL2 in patients with HCV compared with healthy subjects. The serum concentrations of both chemokines were found to be significantly higher in patients with chronic active hepatitis C compared with those with chronic persistent hepatitis C. Subsequent studies performed in patients with different types of liver diseases demonstrated that the serum levels of CXCL10 were significantly higher in patients with autoimmune hepatitis, primary biliary cirrhosis, and both hepatitis B virus and HCV chronic hepatitis than in healthy controls (61). Circulating CXCL10 concentrations were found to be significantly correlated with the serum levels of aspartate and alanine aminotransferases. This finding suggested a relationship between the serum levels of CXCL10 and the necroinflammatory activity of hepatitis. Several clinical studies evaluated the changes of serum CXCL10 in patients with HCV hepatitis undergoing IFN- therapy (60, 62). In their first report, Narumi et al. (60) had demonstrated that the serum levels of CXCL10 significantly decreased after IFN- treatment, but only in cured patients, as assessed by the normalization of serum aminotransferases and by the disappearance of HCV from serum for 6–12 months after stopping therapy (60). In nonresponders to IFN-, the basal serum levels of CXCL10 were significantly higher than in responders to therapy and remained high throughout the treatment (60). A subsequent study reported that the serum levels of CXCL10 and CCL4, but not those of CXCL9, decreased significantly in HCV patients showing a virological response to IFN- treatment (63). A recent study simultaneously evaluated the three CXCR3-binding chemokines (CXCL9, CXCL10, and CXCL11) in plasma samples collected at 1 wk before treatment (baseline), 29 d after starting therapy, and 6 months after completion of a course of pegylated IFN, with or without ribavirin (64). The principal interest for this study stems from the fact that it is the only published experience in which the three IFN- inducible chemokines were assessed simultaneously, thus allowing a comparison of the relative importance of CXCL9, CXCL10, and CXCL11. At baseline, the serum concentrations of CXCL9, CXCL10, and CXCL11 were higher in patients with HCV hepatitis compared with healthy controls, the greatest increase being found for CXCL10. After successful antiviral treatment, the serum levels of CXCL10 and CXCL9, but not those of CXCL11, significantly decreased (64). After completion of IFN- treatment, sustained responders had circulating levels of CXCL10 similar to healthy subjects, whereas in nonresponders to therapy, the serum levels of CXCL10 remained elevated. Pretreatment levels of CXCL9 did not differ in responders compared with nonresponders and declined during therapy in both groups. No significant association was found between pretreatment levels of CXCL11 or between its changes in serum and the outcome of treatment.
Taken together, these results demonstrate that the circulating levels of each of the three IFN--inducible chemokines are differently regulated during IFN- therapy. Of the three chemokines that bind CXCR3, CXCL10 is the one most closely associated with the outcome of treatment with IFN- for HCV-related hepatitis, and its pretreatment levels may predict the likelihood of a favorable response (64).
The main messages of this section are:
in HCV hepatitis results in a long-lasting normalization of circulating CXCL10, which is mainly due to a decreased lymphocytic infiltration of the liver. treatment.
2. CXCR3-binding chemokines in allograft rejection.
Growing evidence suggests that CXCL10 is critical in promoting and amplifying host alloresponses responsible for acute allograft rejection (65, 66, 67, 68, 69, 70). In CXCL10- or CXCR3-gene-deficient mice, cardiac transplants are not acutely rejected and undergo permanent engraftment (43, 65). Accordingly, neutralization of CXCL10 with mAbs prolongs the allograft survival in both cardiac and small bowel models of allograft rejection (65, 66). Furthermore, the intragraft expression of CXCL10 has been reported in association with the rejection of renal (67), lung (68), and cardiac (69, 70) allografts. Thus, the importance of CXCL10–CXCR3 interactions in the pathogenesis of graft failure appears to be clearly demonstrated in multiorgan models.
Recent evidence indicates that CXCR3 and CXCL10 are also highly expressed in conjunction with the development of chronic rejection, also named as chronic allograft vasculopathy (71). Indeed, in addition to its potent effects on immune responses (23, 31, 32, 40, 72, 73, 74, 75), CXCL10 also alters vascular endothelial and smooth muscle cell functions (23, 35, 36, 37, 40, 76, 77), thus promoting the development of chronic allograft nephropathy. Pretransplant serum levels of CXCL10 were measured in kidney graft recipients to verify its value in predicting the recipient’s risk for graft rejection and transplant failure (78, 79). Patients with normally functioning grafts showed significantly lower pretransplant serum levels of CXCL10 compared with patients who experienced graft failure, and lifetime analysis showed significantly lower 5-yr survival rates of the grafts with increasing pretransplant serum levels of CXCL10. Furthermore, frequency of acute rejection episodes in the first month after transplant significantly increased in relation to increasing pretransplant serum levels of CXCL10. In particular, patients with serum CXCL10 levels greater than 150 pg/ml showed a nearly 2-fold greater frequency of rejection. Rejection episodes were not only more frequent, but also more severe in patients showing high pretransplant serum levels of CXCL10 (78). Recently, patients developing chronic allograft vasculopathy were also shown to have significantly higher pretransplant serum concentrations of CXCL10 than patients with normally functioning grafts (78, 79). Multivariate analyses indicated that high serum levels of CXCL10 were a significant risk factor for acute graft rejection and graft failure (78). Taken together, these results indicate that high pretransplant serum levels of CXCL10 may predict the risk for the development of acute rejection and chronic allograft vasculopathy. Accordingly, the urinary levels of CXCL9 and CXCL10 are a sensitive and specific predictor for acute rejection and also mirror the response to antirejection therapy (80, 81). High urinary levels of CXCL10 in the first days after transplant also predict acute rejection, as well as short and long-term graft function (82). Thus, the measurement of CXCR3-binding chemokines in serum or urine may be useful to select those patients requiring more aggressive immunosuppressive regimens.
The main messages of this section are:
3. CXCR3-binding chemokines in multiple sclerosis (MS).
Several chemokine receptors, and among them CXCR3, were shown to be highly expressed in brain samples obtained at autopsy from patients with MS (83, 84, 85), suggesting that CXCR3 might be responsible for the recruitment of autoaggressive T cells. In line with this interpretation, CXCL9 and CXCL10 were found to be significantly elevated in the cerebrospinal fluid (CSF) of MS patients, being positively correlated with the CSF cell counts. The relevance of elevated levels of CXCL10 and CXCL9 in the CSF of MS patients was further supported by the uniform detection of CXCR3+ lymphocytes in the perivascular inflammatory cuffs of brain lesions (83, 84, 85). The accumulation of these cells was directly related to the demyelinating process. The demonstration that the chemotactic activity toward CD4+ T cells specific for a myelin basic protein peptide is mediated by CXCL10 (86) and the notion that IFN- is a potent inducer both of CXCL10 and of clinical relapses of MS provided evidence for a pathogenetic role of CXCL10 in this disease.
Th1- and Th2-oriented chemokines were sequentially measured in the serum and the CSF of patients with MS. CXCL10 and CCL2 were chosen as prototype chemokines for a Th-1 and a Th-2 phenotype, respectively. The measurement of CXCL10 and CCL2 in the serum and CSF of MS patients showed that these chemokines had a different behavior in relation to the activity of the disease. CXCL10 was higher in the serum and the CSF of patients with acute MS and lower in those with a stable phase of the disease. An opposite pattern characterized the CCL2 secretion profile, with high levels being found in the serum and the CSF during the active phases of MS and with a decline in the stable phase of the disease. These findings indicate an involvement of both chemokines, with reciprocal changes according to the clinical phase of MS (85, 87, 88). Because CXCL10 is mainly related to Th1 responses, the increase of CXCL10 in the serum and the CSF of patients during the acute phases of MS fits with the notion that IFN- mediates the immune changes leading to an exacerbation of the disease.
The main messages of this section are:
|
V. CXCR3-Binding Chemokines in Endocrine Autoimmune Diseases |
|---|
, and lymphotoxin-, that result in the activation of macrophages, production of complement-fixing and -opsonizing antibodies, and also cytotoxicity. By contrast, another polarized subset of Th cells, known as Th2, has been thought to play a protective role, inasmuch as cytokines produced by these cells (i.e., IL-4 and IL-13), play an inhibitory effect on the production of Th1 cytokines, as well as on several functions of activated macrophages. Th cells able to produce both Th1 and Th2 cytokines have been named type 0 Th (Th0). Both Th1 and Th2 cells collaborate with B cells for the production of antibodies. However, Th1-induced antibodies differ from those detectable during Th2 responses because of the different subclasses. In mice, Th1 lymphocytes induce B cells to produce mainly IgG2a, whereas Th2 cells induce the production of IgG1 and IgE. In humans, the situation is less clearly dichotomic, but it is known that Th2 responses are characterized by IgE and IgG4, whereas Th1 responses promote the production of IgG1 and IgG3 subclasses. IgG1, which represent the major subclass of human IgG in the serum, are complement-fixing and -opsonizing antibodies, and therefore they contribute, together with activated macrophages, to the phagocyte-dependent protection against infectious agents. Usually, IgG1 also represent the major subclass among autoantibodies, and this is the reason why high levels of autoantibodies are commonly observed in patients with diseases characterized by strong Th1 response.
As mentioned in Section III.A, Th1 cells mainly express CXCR3 as a chemokine receptor and can be recruited into target tissues by CXCL9, CXCL10, and CXCL11. Th2 cells express different chemokine receptors, such as CCR4 and CCR8, thus being recruited in target tissues by CCL17, CCL22 (both ligands for CCR4), and CCL1 (ligand of CCR8). The demonstration of IFN--producing T cells and of CXCR3-binding chemokines in target tissues of organ-specific autoimmune disorders, including those affecting the endocrine glands, has suggested the existence of an important pathogenetic loop. The concept is based on the role of these chemokines in recruiting Th1 cells and in maintaining and amplifying chemokine production by Th1 cells through IFN- production.
It is worth noting that an impressive series of data obtained both in experimental animal models and in human diseases have shown that when Th1 responses, because of their severity and/or chronicity, become dangerous for the body, they can be shifted to a less polarized profile (Th0) or even to responses characterized by the prevalent production of Th2 cytokines. Likewise, established Th2 responses can be shifted to a less polarized profile or even to a prevalent Th1 profile. This phenomenon is known as immune deviation (90).
In the last few years, a novel subset of Th cells has been discovered and named Th17 or ThIl-17 (Fig. 1
) (91). These cells appear to be distinct from Th1 and Th2 cells because of peculiar mechanisms of development and possible functions (92, 93). Although Th1 cells mainly develop in response to IL-12 produced by dendritic cells and Th2 cells develop due to the early presence of IL-4, Th17 cells develop in response to the production of IL-23, IL-6, and TGFß1 by dendritic cells. Th17 cells have been recently suggested to play a pathogenic role in autoimmune diseases on the basis of data obtained in animal models, such as experimental autoimmune encephalomyelitis (which is considered as the equivalent of MS) and collagen-induced arthritis (a model of rheumatoid arthritis). Their role in human endocrine autoimmune diseases remains to be established (94). Thus, in our discussion, we will only take into account the body of experimental evidence suggesting that in these disorders the effector responses are apparently mediated by Th1 cells.
B. Autoimmune thyroid diseases
1. Background.
The thyroid is a major target for autoimmunity. Human autoimmune thyroid disorders (AITD) are characterized by reactivity to self-thyroid antigens, which may be expressed as destructive inflammatory or antireceptor autoimmunity (95) and encompass the clinical spectrum of Graves’ disease and CAT (96, 97, 98). Graves’ disease shares many immunological features with CAT, both diseases being characterized by lymphocytic infiltration of the gland, which can result in tissue destruction (99, 100). One of the histopathological hallmarks of thyroid glands affected by AITD is leukocytic infiltration, mainly by mononuclear cells, including T and B lymphocytes and macrophages (95, 101). In AITD, the lymphocytic infiltrate is also an important site of thyroid autoantibody synthesis (95, 102). Lymphocytes mediate important inflammatory effects, such as the release of cytokines (95). The cellular makeup of the infiltrate varies with the type of AITD, the stage of the disease, and the therapy used, but it is also patient-dependent. This cellular infiltrate sometimes organizes itself into germinal centers that share many of the features of lymph node germinal centers (101, 103, 104). Intrathyroidal lymphocytes play a central role in the pathogenesis of AITD, but the mechanisms by which different lymphocytic subsets are recruited and arrested in the thyroid tissue are only partially understood. To the best of our knowledge, the recruitment of lymphocytes in AITD is a multistep process involving adherence and migration across the endothelium, trafficking through the interstitium, and finally moving toward the thyroid follicular cells (105, 106). Leukocyte extravasation involves the combined action of adhesion molecules, such as selectins and integrins, and chemotactic factors, mainly chemokines (107). In AITD, infiltrating lymphocytes and endothelial cells bear an enhanced expression of various adhesion molecules, pointing to lymphocyte function-associated antigen-1/intercellular adhesion molecule-1, very late antigen-4/vascular cell adhesion molecule-1, and selectin/selectin ligands adhesion pathways as predominant in lymphocyte migration to the thyroid (108). Studies evaluating cytokines in AITD have demonstrated the production of IL-1, IL-2, IL-6, IL-10, IFN-, and TNF- by infiltrating T cells and macrophages (109, 110, 111, 112, 113, 114, 115). However, the specific role of these molecules in the pathogenesis of AITD is still debated (115). In addition, the thyroid follicular cells themselves produce many cytokines (116, 117, 118, 119, 120).
2. Chemokines in AITD.
In 1992, Weetman et al. (121) first described the production of chemokines by cultured thyroid follicular cells. They demonstrated that thyrocytes stimulated by IFN-, TNF-, or IL-1 produce IL-8, a CXC chemokine (121). A subsequent study showed that human thyrocytes in primary culture, upon stimulation with IL-1, TNF-, or IFN-, produce CCL2 (122). Although the highly organized lymphomononuclear cell infiltration present in AITD suggested an involvement of chemokines in their pathogenesis, several years passed before endocrinologists pointed their attention toward these new molecules. It was not until 2000 that the expression of chemokines in AITD was studied in detail and evidence was provided as to their pathogenetic role, at least in the initial phases of these disorders.
The interest in IFN- inducible chemokines (CXCL9, CXCL10, and CXCL11), and their receptor (CXCR3), originated from an investigation aimed at evaluating the antiangiogenetic effects of these molecules. To this purpose, the expression by human endothelial cells of CXCR3 and its ligands was studied in normal tissues and in specimens from diseased organs (37). CXCR3 was detected in a small number of vascular wall cells from normal tissue specimens including thymus, liver, kidney, and gut. Thyroid specimens were obtained from normal tissue and from Graves’ glands. By both immunohistochemistry and in situ hybridization, a higher signal for the protein and the mRNA of CXCR3 was detected in endothelial cells from Graves’ glands, but not from normal thyroids (37).
In 2001, Garcià-Lòpez et al. (123) first demonstrated the production of CXCR3-binding chemokines by human thyrocytes in primary cultures after stimulation with IFN-. In the same culture system, CCL2 and CCL5 were secreted in response to TNF-. In basal conditions, CXCL10 and CXCL9 were not detected in the surnatants from thyroid follicular cells, but their secretion was induced by IFN- and synergistically increased by TNF- addition. As compared with autologous peripheral blood lymphocytes (PBL), intrathyroidal lymphocytes from AITD patients showed a higher expression of CXCR3 and of the receptors for CCL5 and CCL2, CCR2, and CCR5, respectively. T lymphoblasts expressing CXCR3 showed an increased migration to supernatants of IFN- stimulated thyroid follicular cells, which was abolished by neutralizing antibodies directed to CXCL9 and CXCL10, as well as to their receptor, CXCR3. Taken together, these data suggested a role for thyroid follicular cells, through the production of CXCL10, CXCL9, and CCL5, in the recruitment of specific subsets of activated lymphocytes (123).
By using immunohistochemistry, a statistically significant increase of CXCL10 and CXCL9 was found in thyroid tissue specimens obtained from Hashimoto’s glands, compared with normal thyroid tissue (123). By contrast, in patients with Graves’ disease, the intrathyroidal chemokine expression pattern was highly variable, with only a few subjects expressing high levels of CXCL10 and CXCL9, as assessed by immunohistochemistry (123).
A clear-cut demonstration that CXCL9 and CXCL10 were hyperexpressed in Graves’ glands was obtained using combined in situ hybridization and immunohistochemistry (124). The expression of the mRNAs for CXCL10 and CXCL9 in normal thyroids, as well as in thyroids from patients with autoimmune and nonautoimmune hyperthyroidism (Graves’ disease and toxic adenoma), was assessed by in situ hybridization (Fig. 3). The quantitative evaluation of CXCL10 and CXCL9 mRNAs, performed by a computerized video image analysis system, provided evidence that the expression of the mRNAs for the two chemokines was significantly higher in thyroid glands from Graves’ patients compared with normal thyroids or toxic adenoma glands. The wide variability in the expression of chemokines reported by Garcià-Lòpez et al. (123) in Graves’ disease was confirmed and was related to the duration of the disease. A statistically significant increase of CXCL10 expression was found in the thyroid of patients with recent onset (<2 yr) compared with patients with long-standing (>2 yr) disease, in whom the expression of CXCL10 did not differ from that observed in normal thyroid specimens (124).
|
, CXCL10, and CXCL9 in the same samples. RT-PCR revealed that the expression of these molecules was highly heterogeneous, being the mRNA levels for CXCL10 and CXCL9 strictly related to those of the IFN-. The latter were also higher in patients with recent-onset (<2 yr) Graves’ disease. Multiple double-label immunohistochemistry was used to identify the cellular source of chemokines and showed that CXCL10 and CXCL9 were highly expressed by both thyroid follicular cells and infiltrating mononuclear cells (Fig. 3). The CXCR3 receptor was found only in inflammatory and endothelial cells (Fig. 4) (124).
|
In summary, these early studies demonstrated a role for CXCR3-binding chemokines and their receptor in AITD by evaluating chemokine expression at the mRNA and at the protein level, both in the thyroid and in primary cultures of thyrocytes (126, 127). The subsequent steps for defining the role played by chemokines in AITD were provided by clinical studies that evaluated the serum levels of CXCL10 in large series of patients with Graves’ disease or CAT (124, 128, 129, 130, 131, 132).
3. CXCR3-binding chemokines in Graves’ disease.
The first observation of increased serum levels of CXCL10 in patients with Graves’ disease was reported in 2002 (124). Serum samples were collected from 50 unselected Graves’ patients with different duration of their disease, as well as from 25 healthy controls. All Graves’ patients had been treated with methimazole (MMI) at variable doses and were euthyroid at the time of serum analysis. Corticosteroid treatment was an exclusion criterion. Mean CXCL10 serum levels were significantly higher in patients with Graves’ disease compared with healthy subjects, even if there was a large overlap of CXCL10 results between the two groups. The serum concentrations of CXCL10 were inversely correlated with the duration of Graves’ disease, the highest levels being found in patients with recent-onset disease. By contrast, no correlation was observed between the serum levels of CXCL10 and other clinical or biochemical parameters such as sex, age, and titers of circulating Tg Ab or TPO Ab. Interestingly, the reduction of CXCL10 serum levels in long-standing (>2 yr) Graves’ disease was associated with a slight increase in the serum concentrations of CCL22, a chemokine associated with Th2 immune responses (133, 134).
The analysis of the CCL22/CXCL10 ratio demonstrated that a longer duration of Graves’ disease was associated with an increase of the CCL22/CXCL10 ratio in the serum of Graves’ patients (124). Thus, in the late phase of Graves’ disease, an increase in the CCL22/CXCL10 ratio, mainly due to a CXCL10 decline, is observed both in the thyroid gland and in the serum. This phenomenon parallels the reduction of intrathyroidal IFN- mRNA expression (124).
a. Changes in serum levels of CXCL10 in relation to thyroid function and treatment in Graves’ disease.
Following the observation that the serum levels of CXCL10 are increased in Graves’ disease, several clinical trials were designed with the aim of systematically evaluating the serum chemokine status in Graves’ patients in relation to their thyroid function and treatment (130, 131, 132). The final goal was to relate the findings of circulating CXCL10 to the clinical phenotype and to evaluate possible relations between the serum levels of CXCL10 and the two major therapeutic strategies used in Graves’ disease: medical treatment and thyroid removal. Although high serum levels of CXCL10 are not a specific feature of Graves’ disease, having been reported in several endocrine and nonendocrine autoimmune or even nonautoimmune human diseases, the results provided by the following clinical studies support the view that measuring CXCL10 serum levels in Graves’ patients may be useful.
The first study retrospectively evaluated 103 patients with Graves’ disease but with no clinical signs or symptoms of inflammatory ophthalmopathy (132). Graves’ patients were recruited irrespective of their thyroid function or drug treatment. Thirty of them were hyperthyroid and untreated. Fifty-five patients were on MMI treatment for 1–28 months, and the remaining patients were euthyroid, being in remission after a previous course of MMI. Healthy subjects, patients with euthyroid CAT, patients with nontoxic nodular goiter, and hyperthyroid patients with toxic nodular goiter served as controls.
The mean serum levels of CXCL10 were significantly higher in Graves’ patients than in healthy subjects or patients with nontoxic multinodular goiter, but they did not differ from those found in patients with euthyroid CAT. Among Graves’ patients, the serum levels of CXCL10 were significantly higher in those older than 50 yr, in patients with a hypoechoic pattern of the thyroid at US, and in those with an increased thyroid blood flow. Thyroid volume was unrelated to circulating CXCL10. No significant correlation was observed between the levels of CXCL10 and the titers of Tg Ab, TPO Ab, or TSH-receptor (TR) Ab in serum. However, high serum levels of CXCL10 were mainly observed in Graves’ patients who were strongly positive for TR Ab.
Hyperthyroid patients with Graves’ disease had significantly higher serum CXCL10 levels than those who were euthyroid or hypothyroid. Graves’ patients with untreated hyperthyroidism had significantly higher serum CXCL10 levels than those who were hyperthyroid or euthyroid while taking MMI (166 ± 125, 124 ± 41, and 94 ± 35 pg/ml, respectively). The serum levels of CXCL10 did not significantly differ in hyperthyroid Graves’ patients who were untreated compared with those who relapsed after a previous course of MMI (176 ± 125 and 155 ± 97 pg/ml, respectively). Euthyroid patients on MMI or in remission after medical treatment showed similar serum levels of CXCL10.
This retrospective study confirmed that the serum levels of CXCL10 are increased in patients with Graves’ disease, being strongly associated with the hyperthyroid phase of the disease, and do decrease when euthyroidism is restored by MMI treatment (132). In agreement with these findings, high levels of CXCL10 cosegregated with high TR Ab titers. Furthermore, high serum levels of CXCL10 were found to be strongly associated with a marker of disease activity, such as the increased thyroid blood flow. In this regard, the question might be raised of how the huge blood flow of Graves’ glands would fit with the high expression of an angiostatic chemokine, such as CXCL10. The development of new vessels during an inflammatory process results from a balance between angiogenic and angiostatic factors (23). In Graves’ glands, new vessels develop due to extremely high local concentrations of vascular endothelial growth factor produced by thyroid follicular cells in response to thyroid-stimulating antibodies (135). In this setting, the angiostatic effect of CXCL10, which requires binding to the splicing variant B of CXCR3 (see Section III.B), would be easily overcome by the preponderant role of vascular endothelial growth factor, an extremely powerful angiogenetic factor (136).
Patients with Graves’ disease in remission after a previous course of MMI therapy showed serum levels of CXCL10 similar to healthy controls or to euthyroid patients with nontoxic multinodular goiter. The reduction of circulating CXCL10 in patients rendered euthyroid by MMI treatment could be ascribed to the well-known immunomodulatory effect of antithyroid drugs (137). MMI, besides its ability to decrease thyroid hormone production (138), has been shown to interfere with some immunological abnormalities typical of Graves’ hyperthyroidism. The immunosuppressive effect of MMI is highlighted by the reduction of circulating thyroid antibodies, which occurs during medical treatment with this drug, and by the consistent percentage (nearly 30%) of Graves’ patients entering prolonged remission after a course of medical therapy (138, 139). These immunological effects of MMI might be mediated, at least in part, by an action on chemokine production, resulting in a decreased lymphocytic infiltration of the gland. Indeed, a milder lymphocytic infiltration was reported in Graves’ glands after medical treatment (140). Patients with newly diagnosed or relapsing hyperthyroidism had comparable serum concentrations of CXCL10. The increase in serum concentrations of CXCL10 during relapses of hyperthyroidism would be in line with a novel activation of the Th1-mediated immune response and might be taken as an index of an impending relapse of hyperthyroidism after MMI treatment. The increase of circulating CXCL10 in the active phases of Graves’ disease is in agreement with findings in MS showing that serum levels of CXCL10 are higher at disease onset
Top
Abstract
I. Introduction
