Molecular investigations into the patho-aetiology of giant cell arteritis

Abstract

Giant cell arteritis (GCA) is the most common form of vasculitis in people over 50 years of age. It is characterized by inflammation of the medium and large arteries of the head and neck. When the ophthalmic artery is involved, it can lead to sudden irreversible visual loss. GCA is one of the few true ophthalmic emergencies and has potential life threatening complications such as stroke, aortic dissection and rupture. Diagnosing GCA can be challenging in the acute setting as symptoms and signs are variable often leading to a delay in diagnosis. The gold standard to diagnose GCA is a temporal artery biopsy (TAB), an invasive surgical procedure which confirms the diagnosis histologically. In the absence of specific biochemical markers to identify GCA in the acute setting, cases with a high index of suspicion are treated with high dose corticosteroids. The pathogenesis and genetic architecture of GCA remains largely unknown.

The overriding hypothesis of this work is that both genetic and environmental factors contribute to the development of GCA. The work described in this thesis uses multiple techniques including population modelling, analysis of patient cohorts, and modern molecular techniques to provide novel contributions that help define the disease burden, environmental and genetic risk factors as well as the immunological processes occurring in active disease.

A systematic review performed at the start of this thesis, predicted that the future worldwide burden of GCA will increase in view of ageing population. By 2050 more than 3 million people will have been diagnosed with GCA and about 500,000 will be rendered visually impaired, highlighting the need to improve our understanding of this disease.

Research described in this thesis includes the recruitment of a large patient cohort to investigate the genetic architecture of GCA through a genome-wide association study (GWAS). For complex diseases such as GCA, the number of potential candidate genes is overwhelmingly large. In the last twelve years, GWASs have revolutionised the gene-mapping field. Recruitment of the large cohort of patients with GCA required for GWAS is challenging. It is a rare disease, patients are elderly and many live in residential care, are deceased, or have major medical co-morbidities. Such factors make it difficult to enrol patients in studies where a blood sample is required for DNA extraction. To circumvent this issue, an alternative source of DNA from patients with GCA was investigated; DNA extracted from archived formalin fixed paraffin embedded (FFPE) TAB tissue.

An international collaborative group, the Global GCA Genomics Consortium was established and 2084 archived FFPE TAB specimens from 30 pathology centres across four countries were recruited. In addition, 394 blood samples were obtained from patients with biopsy proven GCA over a 3-year recruitment phase. Epidemiological data from this cohort was used to investigate whether there was seasonal variation in the incidence of GCA which could indicate an environmental aetiology. No seasonal preponderance was demonstrated. In addition, an observational study based on the administrative burden involved in setting up a large collaborative nationwide non-interventional study highlighted the need to streamline ethics nationally to prevent duplication of work.

For the GWAS, DNA samples extracted from FFPE blocks were genotypes using Illumina SNP microarray chips. After stringent assessment of sample suitability (meeting inclusion criteria), DNA and genotyping quality control, a total of 1,212 DNA samples, provided a genotyping success rate >95%. These 1,212 GCA cases were compared to 9,657 historic controls from the UK Household Longitudinal Study for the discovery phase of the analysis. Association analyses for 8,075,195 SNPs, identified two genetic loci at the genome-wide significance threshold of p<5x10-8, including the HLA region, where the top hit was rs9272417 (p-val = 8x10-51) which is associated to the gene expression of HLA DQA1 in the tibia artery and the aorta. The novel locus outside the MHC region was rs2351254 (OR = 0.52; p-val = 1.54x10-8) on chromosome 15. This SNP is associated with the expression of MFGE8 in the tibial artery and aorta. The association of this locus replicated in two separate cohorts, 177 cases from an independent Australian cohort and 352 cases from the United Kingdom (OR = 0.71; p-val = 0.056 and OR = 0.39; p-val = 0.003).

This thesis also describes work to investigate the immunopathogensis of GCA. GCA is presumed to be an autoimmune disease in which T lymphocytes play an important role. High resolution RNA-sequencing (RNA-seq) permits the study of gene expression, giving an insight into the cellular processes driving inflammation during the acute phase of this disease. This technique was used to investigate RNA expression in CD4+ and CD8+ T-lymphocytes from patients with active and quiescent GCA over the course of 12 months. Serial blood samples were collected from 16 patients with GCA and 16 age-matched controls. Studying the circulating expression profiles in these T cells of both patients and controls was used to identify molecular associations with GCA disease status. These could potentially identify a new biomarker of active disease, increase the sensitivity and specificity of haematological monitoring of active disease, and forgo the need of an invasive procedure to diagnose this disease.

The results of this RNA-sequencing study show cell type-specific transcript expression profiles, novel gene-phenotype associations, and uncover important biological pathways in GCA. Through monitoring expression profiles of patients over a 12-month period, polynomial modelling analyses identified 179 and 4 statistically significant transcripts with altered expression profiles (FDR < 0.05) between cases and controls in CD4+ and CD8+ populations, respectively. Genes included LMBR1L, UAP1L1, KCNMB4, PTCD2 and THRAP3. Gene correlations were found between both symptoms and biochemical markers used in the acute setting for predicting long-term prognosis. 15 genes were shared across 3 phenotypes in CD4 and 16 across CD8 cells. In CD8, IL32 was common to 5 phenotypes: a history of Polymyalgia Rheumatica, visual disturbance and raised neutrophils at time of presentation, bilateral blindness and death within 12 months. In the acute phase, such gene-phenotype relationships identified could act as disease biomarkers, provide insight to potential disease severity and guide in initiating appropriate patient management.

In summary, this work has identified novel gene associations with GCA through a GWAS and has led to the discovery of interesting transcripts profiles through an RNA-seq approach. The results derived from this research have provided the basis for future functional work to investigate the role of these genes in GCA pathogenesis. This research could assist future studies to explore avenues for disease screening, prevention, monitoring and treatment.