Maria Magdalena Zorro Manrique

Celiac disease (CeD) is an autoimmune disease triggered by a grain protein called gluten that is frequently present in the Western diet. The disease primarily occurs in the small intestine, where it causes a severe inflammatory response. A lifelong gluten-free diet has emerged as the only available therapy for CeD. Unfortunately, particularly for young patients, this therapy is difficult to adhere to due to social ‘stigma’ and because the diet may lead to nutritional deficiencies. Another problem associated with ingesting gluten-free food items is that these may actually contain ‘hidden’ sources of gluten. To design effective treatments and preventative measures, a deeper understanding of the mechanisms underlying the disease is required.

As CeD is an intestinal autoimmune disease, disease etiology comprises deregulation in several cell types that leads to inflammation and degradation of the small intestinal lining. These cell types include cells from the adaptive immune system such as CD4+ T cells, which produce interferon gamma (IFNg), a key cytokine that recruits and activates immune cells; intraepithelial cytotoxic lymphocytes (IE-CTLs), the effector cells responsible for killing epithelial cells; B cells, which secrete autoantibodies and present gluten peptides to the CD4+ T cells; and cells from the innate immune system, such as macrophages and dendritic cells, which contribute to the activation and migration of innate and adaptive immune cells, primarily through the cytokines they produce. Lastly, intestinal epithelial cells may also play a role, directly or indirectly, by allowing the passage of ions and nutrients into the body and by expressing molecules/cytokines that influence the differentiation and function of immune cells.

After several decades of study, researchers have found that the genetic variants associated with CeD risk through genome-wide association studies (GWAS) change expression of certain genes called candidate genes. Candidate genes may alter the function of immune cells and intestinal epithelial cells that reside in gut, thus contributing to the risk of developing CeD. The goal of this thesis research was to gain more insights into the function of some of these candidate genes and the biological processes in which they are involved. To do so, we have integrated in silico methods with experiments performed in the lab using cells and environmental conditions thought to be relevant for CeD.

Chapter 1 presents an overview of CeD and the major cell types, cytokines, immune molecules and genetic and environmental factors that are thought to contribute to the disease. It also discusses the current limitations on accurately pinpointing candidate genes and some approaches to overcome this issue. Finally, an outline of the chapters is provided.

Although most CeD candidate genes exhibit immune functions, it is known that non-immune genes such as LPP may contribute to the disease. Previous in silico studies have suggested that this poorly characterized gene may be involved in intestinal barrier function, which is known to be disturbed in CeD patients. In Chapter 2 we studied the role of LPP in intestinal epithelial cells. We generated intestinal epithelial cell lines with reduced expression of LPP (called LPP knockdown). These cells presented alteration of essential functions such as proliferation, permeability of the cellular barrier and the capacity to form organized structures when compared to cells with normal levels of LPP (called controls). Surprisingly, the LPP knockdown cells presented an enhanced inflammatory response to IFNg, a cytokine known to fuel the inflammation in CeD. Thus, these results suggest that LPP may contribute to disease by affecting the non-immune and immune functions performed by epithelial cells.

In Chapter 3, we identified 118 CeD candidate genes by performing expression quantitative trait loci (eQTL) analysis and subsequently applying different statistical approaches to link these eQTL genes to genetic risk for CeD. Subsequent analysis suggested that these genes affect general and immune-specific processes in the different cell types present in the gut. Interestingly, one of these candidate genes, TRAFD1, was found to be a central regulator of 40 genes implicated in the IFNg signaling and the Major Histocompatibility Class I (MHC-I) antigen processing/presentation pathways. Of note, these biological pathways are known to participate in the abnormal activation of immune cells, specifically the IE-CTLs that are responsible for killing healthy epithelial cells, resulting in intestinal tissue damage in CeD. Although further experimental validation for many of the candidates may be needed, the results in this chapter provide a starting point that can lead towards a better understanding of the role of genetic variants associated with CeD in the immune response in CeD pathogenesis.

CeD arises when the immune system breaks its tolerance to the otherwise innocuous gluten peptides, thus leading to the subsequent inflammatory response of the adaptive and innate immune system. One of the factors that may contribute to this break in tolerance is deregulation of cytokine production. In Chapter 4, we studied how cytokines that are known to be upregulated in autoimmune disease-affected tissues (IFNb, Il-15 and IL-21) alter the gene expression of IE-CTLs, the cells responsible for intestinal destruction in CeD. We found that although each cytokine induced specific patterns of gene expression in IE-CTLs, all the stimuli promoted the expression of genes associated with IFN and IFNg signaling, both of which are processes essential for immune cell activation and perpetuation of inflammation. Due to the fact that the cytokines we used for the experiments are also present in the environment of autoimmune disease–affected tissues, this result indicates that IFN signaling may be a common mechanism that could lead to aberrant activation of IE-CTLs when they are exposed to abnormal inflammatory conditions. This assumption was further confirmed by performing an enrichment analysis of genes differentially expressed in IE-CTLs upon cytokine stimulation in genomic loci associated with several autoimmune diseases. Here we observed that the genes affected by cytokine stimulation in IE-CTLs are enriched in genes found in loci associated with ten common autoimmune diseases. Interestingly, for most of the autoimmune diseases tested, the disease-associated genes were found to be enriched in IFN signaling genes. This indicates the relevance of IE-CTLs and IFN signaling not only in CeD, but also in other autoimmune conditions, which points to possible therapeutic approaches to ameliorate the abnormal inflammation.

With the advent of next generation sequencing technologies, several types of non-coding RNA molecules, which are encoded in our genetic background, have been uncovered. Among them, long non-coding RNAs (lncRNA) and microRNAs have garnered a lot of attention due to their role in regulating biological processes such as gene expression and immune cell differentiation and activation (reviewed in Chapter 5). Furthermore, several reports have indicated that alterations in the expression of these RNA molecules may be associated with multiple cancers and autoimmune diseases. Despite the biological relevance of these transcripts, their precise functions remain poorly understood.

To better understand the role of lncRNAs in the innate immune system, we characterized one lncRNA, RP11-291B21.2, in IE-CTLs (Chapter 6). We studied the expression pattern of RP11-291B21.2 in different immune cell populations derived from blood and found that this lncRNA is uniquely expressed in CTLs. We also found that the level of RP11-291B21.2 was higher when the CTLs are less activated, suggesting that this lncRNA could participate in T cell activation. This observation was further confirmed by combining co-expression network analysis (an in silico method to predict gene function) with experiments in cells with a reduced expression of RP11-291B21.2 (knockdown). Here we observed impaired production of some cytokines relevant for CeD (IFNg and TNF) and reduced expression of several genes associated with cell activation and inflammation in the RP11-291B21.2 knockdown when compared with the controls. Although we do not fully understand how RP11-291B21.2 influences the activation of IE-CTLs, our results constitute an interesting basis for further study of this lncRNA in the biology of CTLs and in autoimmune-mediated diseases where abnormal CTL activation is associated with tissue destruction.

Finally, in Chapter 7, we summarize and integrate the major findings of this thesis. Furthermore, we discuss some of the gaps regarding the study of candidate genes and biological pathways associated with CeD and how novel technologies could be used to fill these gaps and confirm hypotheses derived from in silico analysis or in vitro models to move us further towards the design of novel therapeutic interventions for CeD.