UCSF

Encephalitis and associated conditions (meningitis, myelitis, etc.) are a collection of diseases characterized by acute or chronic inflammation of the brain and central nervous system (CNS). Diagnosing and treating encephalitis can be challenging given the litany of potential infectious, ischemic, metabolic, autoimmune, neoplastic, paraneoplastic, parameningeal and toxic causes. Diagnostic approaches in infectious and autoimmune etiologies – which account for an overwhelming majority of cases – are generally candidate-based panels of nucleic acid (PCR), serological, or cell-based assays with limited throughput and often suffer from low sensitivity and specificity. Frequently, patients receive no definite diagnosis, are treated empirically at great cost to the individual and greater health system, or receive treatments contraindicated to their specific condition. Novel, unbiased, and high-throughput diagnostic methods are needed to address these challenges.

Here I present my efforts to develop rapid, cost effective, unbiased, and high-throughput methods leveraging next generation sequencing (NGS) technologies to illuminate the underlying causes of infectious and autoimmune-mediated CNS inflammation. I demonstrate the design and implementation of a phage display and immunoprecipitation (PhIP-Seq) assay to map the antigenic determinants of autoantigens in two paraneoplastic neurological disorders, the anti-Hu and anti-Yo syndromes. The assay, leveraging a rationally designed library of phage clones expressing >700,000 peptides encompassing the entire human proteome, also serves as a discovery platform to identify novel disease-associated antigens and antibody binding signatures that may eventually serve as valuable diagnostic and prognostic biomarkers in cancer detection and immune repertoire profiling.

To address challenges in identifying infectious etiologies, I present my efforts in applying metagenomic and metatranscriptomic NGS (mNGS) to patient cerebrospinal fluid (CSF) samples. The low nucleic acid content of CSF and the sterile, privileged nature of the CNS present unique opportunities and challenges for metagenomic assays. The complex, noisy, and polymicrobial datasets generated by mNGS require careful analysis to determine which, if any, of the identified microbes represent a true pathogen versus environmental contamination. Failure to make this distinction can result in spurious disease associations with organisms later determined to be laboratory contaminants. I demonstrate the necessity and efficacy of a straightforward statistical framework coupled with careful study design for identifying microbial pathogens in a series of challenging cases of subacute or chronic infectious meningitis, as well as for analyzing publicly-available data from recent mNGS diagnostic and brain microbiota studies.