Stromal Cell Diversity Drives Pathogenic Mechanisms in Rheumatoid Arthritis and Pre-Disease Onset

Abstract

Rheumatoid Arthritis (RA) is a chronic, progressive, inflammatory autoimmune disease primarily characterised by joint damage and disability and is frequently associated with comorbidities including diabetes and cardiovascular disease (CVD). In the last two decades, targeted biotherapeutics have proven successful in the treatment of RA, leading to significantly improved outcomes. Recent research has also focused on 'individuals at-risk’ (IAR) of developing RA, an approach that may provide important insights into the evolution of RA. Despite these advances in therapies and early intervention, a significant proportion of patients experience suboptimal responses, no response, or suffer adverse events that limit therapeutic impact. Currently, it is difficult to predict who will respond to treatment or who will develop severe, erosive disease. This difficulty arises from the complex microenvironment of the joint, where interactions between immune and stromal cells drive pro-inflammatory responses. Furthermore, recent studies have suggested that based on autoantibody status and differential cellular predominance in the synovium, that several different pathotypes exist under the umbrella of RA. One such pathotype is associated with a predominance of synovial fibroblasts (FLS), key stromal cells involved in immune regulation, metabolic re-wiring, and invasive mechanisms within the joint. Indeed, studies suggest that this pathotype may be refractory to current treatments. However, little is known about their phenotype or transcriptional profile in health and disease, particularly in relation to autoantibody status. Therefore, the aim of this thesis is to functionally characterise distinct FLS populations in RA patients (autoantibody positive vs autoantibody negative), the findings of which may have implications for treatment strategies. In Chapter 2, the role of the mammalian target of Rapamycin (mTOR) in mediating RA FLS activation and function was examined. Transcriptomic analysis of healthy donor and RA synovial tissue revealed dysregulated expression of several key components of the mTOR pathway in RA. Moreover, the expression of phospho-ribosomal protein S6 (pS6), a major downstream target of mTOR, was increased in RA FLS compared to HC FLS. In the presence of TNFα, RA FLS displayed heightened phosphorylation of pS6 and were responsive to mTOR inhibition via Rapamycin. Rapamycin effectively altered RA FLS cellular bioenergetics by inhibiting glycolysis and the expression of rate-limiting glycolytic enzymes. Furthermore, mTOR signaling was shown to mediate RA FLS migratory and invasive mechanisms, with these functions abrogated in the presence of Rapamycin. Finally, transcriptomic analysis identified a significant upregulation in several genes involved in the Hippo-Yes-associated protein (YAP) pathway in RA synovial tissue, with predicted convergence on the mTOR pathway identified. Crosstalk between the mTOR and YAP pathways was shown to mediate RA FLS invasive mechanisms. Together these data identify mTOR as a key driver of RA FLS pathogenic mechanisms, effects that were, in part, mediated via crosstalk with the Hippo-YAP pathway. In Chapter 3, the phenotypic characterization of the recently identified circulatory ‘pre- inflammatory mesenchymal’ (PRIME) cells was examined in RA and IAR as potential systemic cellular biomarkers of disease. Additionally, site-specific functional differences in PRIME cells were also investigated. The frequency of PRIME cells (CD45-CD31-PDPN+) was increased in RA->RA+ compared to HC. An increase in the frequency of RA+ PRIME cells positive for immunoregulatory markers CD200+ and HLA-DR+ was observed compared to HC. Enhanced adhesive capacity of PRIME cells was observed in RA, reflected by the increased expression of ICAM-1, CD44, VCAM-1 and CD61. RA PRIME cells also displayed increased expression of established FLS subset/sub-lining markers THY-1, CD34, and FAPα. Similar to RA, in IAR, increases in HLA-DR+ and ICAM-1+ PRIME cells were observed, as were the THY-1+, FAPα+ and CD34+ subsets. Site-specific analysis demonstrated that PDPN+ cells were significantly more abundant in synovial tissue compared to circulation. These PDPN+ cells demonstrated greater frequency of CD61, ICAM-1, CD200R1, CD200, pmTOR and pS6 in peripheral blood compared to synovial tissue, while their expression of CD44, HLA-DR, and pAKT was enriched in the synovial tissue. In summary, PRIME cells could function as a potential cellular biomarker in IAR and as indicators of disease severity in RA. Moreover, further investigation comparing the characteristics of circulating PRIME cells to PDPN+ cells in the joint may provide insight into which specific FLS subsets arise from these circulating mesenchymal precursors as opposed to a tissue-resident origin. Finally, additional research is required to better understand the origins of PRIME cells to elucidate the early stages of joint inflammation and joint spreading. In Chapter 4, FLS diversity was further examined focusing on the site of inflammation— the synovium. PDPN+ FLS were enriched in RA+<RA-<HC. Analysis of FLS-specific subset markers demonstrated that CD55+ (lining layer marker) FLS were significantly increased in RA- compared to RA+ and HC, while THY1+ (sub-lining) and FAPα+ (activation) FLS were increased in RA+ compared to HC. An increased frequency of HLA-DR+ and ICAM-1+ FLS was observed for RA+ patients compared to HC, whereas CD61+ and pAKT+ FLS were associated with RA- patients. RA+ demonstrated increased frequency of the activated SL subset (CD55-THY1+FAPα+), while RA- were associated with the activated LL subset (CD55+THY1-FAPα+). CD55+THY1-FAPα+ subsets demonstrated increased expression of Ki67+ and pAKT+, while CD55+THY1+FAPα+ subsets demonstrated increased expression of VCAM-1+, CD61+, CD44+, and YAP+ cells. The CD55- THY1+FAPα+ sub-lining subset was associated with increased expression of HLA-DR, CD200, ICAM- 1, CD34, and pS6. Finally, CD55-THY1-FAPα+ demonstrated increased frequency of CD200R1+, ICAM-1+, Cadherin 11+, PDGFRα+, and pmTOR+ cells compared to the other subsets. Enrichment of these subsets and specific activation markers were also observed in IAR. In parallel, scRNA-seq identified 14 FLS clusters between HC and RA synovial tissue. Cluster 3 (POSTN+COL3A1+), a sub- lining cluster was enriched in RA compared to HC and was characterised by genes involved in collagen and ECM reorganization, as well as key signaling pathways NFkB, TGF-β, and NOTCH. Receptor-ligand interaction analysis for this cluster identified interactions with macrophage- and T cell-derived ligands including IL-1b, VEGF and TGF-β. Functional assays demonstrated IL-1β- induced expression of the NOTCH ligands DLL4 and JAG-1, with addition of the NOTCH inhibitor DAPT abrogating NOTCH-induced pro-inflammatory responses. Identification of differential FLS subsets and their associated function in the HC vs IAR vs RA synovium will better facilitate our understanding of their contribution to RA disease pathogenesis. Furthermore, stratifying RA patients by seropositivity will give better insight into the pathogenic differences between these two disease cohorts, with the potential for better tailored treatments. RA is associated with several comorbidities including the development of CVD. Therefore, in Chapter 5, circulatory endothelial dysfunction and serum biomarkers associated with inflammation and cardiovascular/metabolic processes were analysed. Circulating endothelial cells (CECs) were found to be elevated in both RA and IAR compared to HC. Most CECs identified originated from the microvasculature with a majority positive for CD36 expression. RA CECs demonstrated impaired capacity for wound healing with decreased expression of PDGFRα and YAP compared to HC. CECs have a pathogenically activated phenotype, evident by increased expression of FAPα and THY-1. RA CECs demonstrated trending increases in immunoregulatory and adhesive markers. Analysis of serum biomarkers demonstrated increases in inflammatory, metabolic, and tissue injury soluble mediators in RA, some of which were also increased in IAR. Analysis of circulatory CECs and serum biomarkers could act as potential biomarkers in IAR and RA patients for disease onset and progression but may also have implications for those at greater risk of developing a cardiovascular comorbidity, further contributing to disease pathogenesis. In conclusion, the data outlined in this thesis provide a significant in-depth analysis of FLS subsets and a better understanding of the unique phenotypic function of FLS that may differ in RA pathotypes. Moreover, the data indicate a role for metabolic reprogramming of FLS in inducing the resolution of inflammation. Finally, cellular and soluble markers in the circulation may be useful predictive markers of RA disease onset and progression, in addition to associated CVD. Taken together, the data in this thesis significantly enhance our understanding of RA disease pathogenesis, with implications for biomarker discovery and the development of new therapeutic approaches for difficult-to-treat patients with RA.