부산대학교 도서관

Antibodies in patients with PNS or antibody-mediated autoimmune disease are polyclonal and bind different protein epitopes. Some are linear and flexible while others are rigid or made up of loops that only come together in three-dimensional space in the native protein structure. Methods where the antigen is denatured, such as western blot, where the detergent is used to extract the target from cells, or reagents, such as formaldehyde, are used to preserve test substrates are inferior to live cell-based assays (CBA). Regardless of whether the epitope is linear and flexible, rigid, or conformational, the native protein provides targets for both antibody types. Modifying the target structure in any way reduces the test accuracy, hence live CBAs are the gold standard for antibody detection in patients with autoimmune antibody-associated diseases. These tests can also be applied to soluble molecules such as leucine-rich, glioma-inactivated 1 (LGI1) by linking them to the cell membrane. In this thesis, live assays for the detection of autoantibodies to soluble cytokines, integral membrane proteins, and onconeural antigens were developed. The association of cytokine antibodies and thymoma is concluded, and the ability to express cytosolic and nuclear proteins on the surface of a cell for antibody detection is demonstrated for the first time for onconeural antibodies and a multiplex flow assay for the detection of antibodies to aquaporin-4 (AQP4) or myelin oligodendrocyte glycoprotein (MOG) in the same sample has been developed. The cytokine assay includes an extracellular myc-tag to evaluate the surface expression of each cytokine. Assays for autoantibodies to IL-6, IL-12, IL-17A, IL-17F, IL-22, IL-23, IFNα2, IFNα8, and IFN⍵1 were developed and autoantibodies to IFNα2 and IL-12 were found most frequently in patients with a thymoma. In a cohort of 126 autoimmune myasthenia gravis (aMG) patients, the assay sensitivities were 80% and 64% respectively. In the same cohort, 93.3% of aMG patients with both autoantibodies had been diagnosed with thymoma. It was hoped that these cytokine antibodies could predict relapse in patients after thymectomy, but the endpoint titres of autoantibodies to IFNα2 and IL-12 from most TAMG patients were above the cut-offs years after patients' thymectomy. Regarding the detection of cytokine autoantibodies, there was excellent correlations between the newly developed live CBAs and luciferase immunoprecipitation system (LIPS), with the benefit that the multi-subunit cytokines, e.g., IL-12 and IL-23, could be assayed by our live CBAs as IL-12B homodimer, or IL- 12 and IL-23 heterodimers. Multiple live CBAs for onconeural antibodies were developed including HuD, NOVA1, CRMP5, CDR2L, Amphiphysin, and ZIC4, all of which were expressed above the surface of HEK293T cells. They were incorporated into a vector that included a CASPR2 signal sequence for surface expression and a CASPR2 transmembrane and cytosolic domain to retain the proteins on near the cell surface and an eGFP tag to locate transfected cells in this transient transfection system. The same strategy did not work for Ma1, Ma2, CDR2, SOX1, and KLHL11 but they did over-express as cytosolic antigens. The main reason for the surface expression of these targets is to eliminate the background of cytosol and nuclear staining seen in fixed CBAs and immunohistochemistry (IHC). Our live CBAs identified all antibody-positive samples sent from two centres of excellence in Spain and the USA (cohorts with 69 serum and 21 CSF samples in total). In addition, our live CBAs didn't pick up 8/12 samples, which had been identified seropositive by line blots only (HuD Ab+, n=1, NOVA1 Ab+, n=1; CDR2 Ab+, n=5; Amphiphysin Ab+, n=1), from patients without high-risk PNS and tumour. Therefore, our eGFP-tagged live CBAs are diagnostically more specific than line blots for the prediction of tumours and are multiplexable compared to tissue-based fixed IHC. In order to quantitate and multiplex these assays a quadruplex CBA and flow cytometry assay were developed. The quadruplex CBA depended on specific localisation of transfection reagent while the flow cytometry system required the addition of colour to distinguish cells transfected with different antibody targets. CellTrace Violet dye proved useful for colour-coding cell populations. The quadruplex assay (HuD, NOVA1, Amphiphysin and CDRL2) by either CBA or flow cytometry was able to reproduce data from individual assays. Lastly, we applied a DsRed2 reporter to live flow cytometry CBA (F-CBA) for detecting autoantibodies to AQP4 M23 and MOG α1 isoforms. Our live AQP4 F-CBA was very sensitive with a high signal-to-noise ratio, but our live MOG F-CBA had a higher background, as expected from the literature. This duplex test requires further optimisation before being implemented in a routine diagnostic laboratory. To summarize, this thesis narrates the developments and screening results of our live and permeabilized CBAs for detecting autoantibodies to soluble cytokines, onconeural antigens and transmembrane proteins. The CBAs we developed could also benefit clinicians to predict the corresponding tumours for early and better management.