1721. A Transcriptional Signature of Acute Aspergillus Infection Offers High Diagnostic Accuracy Despite the Presence of Immunosuppression

BACKGROUND: Invasive aspergillosis (IA) is a major cause of critical illness in immunocompromised (IC) patients. However, current fungal testing methods have significant limitations and there is a clear need for new diagnostic options. Disease-specific gene expression patterns in circulating host cells show promise as novel diagnostics; however, it is unknown whether such a “signature” exists for IA. Additionally, there is a need for better understanding of the effect of iatrogenic immunosuppression (present in most cases of IA) on such host response-driven biomarkers. METHODS: Male BALB/c mice were separated into an Aspergillus fumigatus inhalational exposure group and a placebo group. These two groups were each subdivided into three additional sets based on immunocompromised status (no immunosuppression, cyclophosphamide, and corticosteroids) for a total of six experimental groups. Mice were sacrificed 4 days post-infection. Whole blood was assayed for transcriptomic responses via microarray. Bayesian techniques were utilized to develop classifiers of IA and leave one out cross-validation was used to estimate predictive probabilities. RESULTS: Aspergillus infection triggers a powerful response in non-IC hosts, with 2996 genes differentially expressed between IA and controls. We generated a 146-gene expression classifier able to discriminate between non-IC mice with IA and uninfected non-IC mice with 100% accuracy. However, the presence of immunosuppressive drugs exhibited a strong confounding effect on the transcriptomic classifier that was derived in the absence of immunosuppression. After controlling for the genomic effects of immunosuppressive drugs, we were able to generate a 187-gene classifier with a sensitivity of 100% and specificity of 97% across all IC states. CONCLUSION: The host transcriptomic response to IA is robust and highly conserved. Pharmacologic perturbation of the host immune response unsurprisingly has powerful effects on gene expression-based classifier performance and must be taken into account when developing novel diagnostics. When appropriately designed, host-derived peripheral blood transcriptomic responses to IA demonstrate the ability to accurately diagnose Aspergillus infection, even in the presence of immunosuppression. DISCLOSURES: All authors: No reported disclosures.