Analysis of the cumulative changes in Graves’ disease thyroid glands points to IFN signature, plasmacytoid DCs and alternatively activated macrophages as chronicity determining factors

Abstract

Graves’ disease (GD) is a chronic autoimmune process in the thyroid gland and involves IFN and IFN driven immune activation. Assuming the thyroid gland is the main site stimulating the autoimmune response, we investigated the role of IFNs and other factors in the chronic evolution of GD by comparing the transcriptomic profiles of thyroid glands from short clinical course (SC), long clinical course (LC) cases, and control glands (C). Over 200 differentially expressed genes of the immune system were identified. Results were extensively analyzed bioinformatically and validated by qPCR in 31 glands. The analysis indicated that GD involved a progressive accumulation of changes with clearly distinct profiles in the SC and LC glands. Humoral response, antigen presentation and chemokines & cytokines were overall the most represented gene ontology categories in LC cases. Ingenuity Pathway Analysis pointed to a few inflammatory pathways in SC cases whereas LC cases involved numerous complex pathways, such us “communication between innate and adaptive immune cells” and “autoimmune thyroid signaling”. A broad IFN signature consisted of the over-expression of 74 and 84 type I and type II IFN responsive genes respectively (overall 96 out of 211, 45%), but many of these genes can also be directly activated through cytoplasmic viral receptors. For the first time, plasmocytoid dendritic cells were identified in GD thyroid, but surprisingly, the main producers of IFN-alpha were cells with a myeloid cell phenotype. In addition, cells with the phenotype of alternatively activated macrophages were detected in abundance in GD thyroids, confirming data from the transcriptomic analysis. Collectively, these results confirmed the role of IFNs, suggested other natural immunity triggers, identified new cell types in the local disease process, and expanded our knowledge of the processes that may determine the chronicity of GD.

Introduction

Autoimmune Thyroid Diseases (AITD) include Hashimoto thyroiditis (the paradigm of organ specific autoimmunity) and Graves’ disease (GD), have a prolonged chronic course, and very rarely cure spontaneously [1], [2]. The transient course of autoimmunity in experimentally induced animal models suggests that thyroid tissue damage is not sufficient by itself to maintain the autoimmune responses [3]. One major question in AITD -and in organ specific autoimmune diseases in general—is the perpetuating factors that maintain the autoimmune response despite regulatory mechanisms that should lead to its resolution [4]. Analogously, the response to chronic infections is maintained by the persistence of the pathogen which is the source of both non-self antigen and of Pathogen Associate Molecular Patterns (PAMPs) that activate antigen presenting cells. As no pathogens have been convincingly isolated from the target tissue of organ specific autoimmune diseases, endogenous Damage Associated Molecular Patterns (DAMPs) rather than PAMPs are the candidates to explain the persistent immune response.

Numerous observations indicate that the site that controls, at least in part, the autoimmune response in GD is the thyroid gland. For example, the immune response in GD declines after thyroidectomy [5], and GD thyroid tissue engrafted into immunodeficient mice produces anti- thyroid antibodies [6]. The thyroid tissue itself undergoes many changes in GD, such as thyroid follicular cells (TFCs) express HLA class I and class II [7], [8], TAPs and LMPs [9], HLA-DMA [10], ICAM-1 and NCAM [11], [12], [13], [14], CD40 [15], Fas and Fas-L molecules [16] and also produce CXCL12 [17] and IL-16 [18]. In addition, GD entails the frequent extensive infiltration of the thyroid gland by lymphocytes, macrophages and dendritic cells making up germinal centers containing Thyroid Peroxidase (TPO) and Thyroglobulin (TG) specific B-lymphocytes [19]. Abundant naive T cells with features of recent thymic emigrants are present in GD thyroid tissue up to approximately 30 months, but not in later specimens [20]. Intrathyroidal lymphoid follicles, some with germinal centers, develop in the glands after one year of clinical course ([19] and unpublished observation). Collectively, these observations support the argument that the thyroid is an important site in the pathogenesis of GD, although they do not exactly fit the “aberrant expression of HLA-DR” hypothesis of AITD [21], [22]. These characteristics of GD imply that the factors determining chronicity can be identified by comparing the molecular profile of the thyroid glands at the start of GD and at a chronic stage of the autoimmune process.

The well validated clinical observation that the therapeutic administration of type 1 interferons can trigger AITD [23], [24] suggests that interferons play a crucial pathogenic role in thyroid autoimmunity as in other types of autoimmune diseases [25], [26]. Many of the aforementioned changes in TFCs from GD glands can be attributed to IFNs, as shown in vitro [27], [28]. Recently, typical interferon-inducible genes have been reported elevated in PBLs from recent onset GD patients [28]. However, the presence of an IFN signature and of IFN-dependent pathways activation has not been investigated in the thyroid glands of these patients, a pathogenically important and accessible site.

Inescapably, current knowledge of the immunopathology and pathogenesis of GD is biased by the dominant paradigms on the organization of the immune response and tolerance. Genome wide transcriptomic profiling can provide a non-biased perspective of the processes that are affecting autoimmune tissue [29]. In this study, we have analyzed GD glands by combined genome wide transcriptomic profiling with qPCR, bioinformatics, and immunohistological analysis to address three questions: 1) Is there an IFN effect/signature in GD glands including the activation of the antigen processing and presentation pathways?; 2) Is the continuous generation of DAMPs the cause of the chronic nature of GD, and can they be identified?; 3) Are additional elements of the immune response so far unknown playing a relevant role in the pathogenesis of GD?