Discovery and Detection of miRNA Biomarkers for Viral Infection

Despite its rare occurrence, viral infection in pharmaceutical production cultures still poses a great threat to both pharmaceutical companies economically and to the public health. However, current methodologies used to determine viral contamination have many drawbacks in addition to being both time and labor intensive. In this thesis, we make a new connection between pharmaceutical CHOK1 production cultures and the ability of the miRNA transcriptome to indicate cellular stress – particularly in the case of viral infection. Monitoring of the host cell miRNA transcriptome is a real-time, direct measurement of host cell response to the viral infection and may be utilized in a new technique for fail-fast adventitious agent detection. 

Here, we quantified virulence of 5 different viruses (Reo3, MMV, PI2, EMC, and VSV) in CHO-K1 cells. CHO-K1 cells did not show cytopathic effects after being inoculated with PI2, however, the other 4 viruses were used to identify both general and virus-specific biomarkers for infection. Overall, 10 miRNAs (cgr-miR-93-3p, cgr-miR-32-5p, cgr-miR-33, cgr-miR-17-5p, cgr-miR-24-5p, cgr-miR-18a-5p, cgrmiR-450b-5p, cgr-miR-542-3p, cgr-miR-21-3p, and cgr-miR-29b-3p) have shown promise as biomarkers for general viral infection in CHO-K1 cultures compared to healthy, uninfected cells. In addition to passing through the double filtering method of fold change and p-value, these miRNAs showed excellent discrimination ability between infected and uninfected cultures with a receiver operating curve (ROC) analysis. 

The concept of host cell miRNA biomarkers was also leveraged to lay the groundwork for a more efficient diagnostic method for ocular surface infection. Using herpes simplex virus type 1 (HSV-1) infection of human corneal epithelial cells (HCLEs), miRNA biomarkers were identified to describe this technology. Overall, the top miRNA biomarker for an early host response to HSV-1 infection in HCLE cells was found to be hsa-miR-378f. Additionally, both hsa-miR-615-3p and hsa-miR-495-3p showed a linear increase in their fold change over time and were good classifiers without a proclivity towards false positives or false negatives. 

Finally, we also explored the use of DNA and LNA probes as a more inexpensive option compared to the γPNAA probes previously used with the MTE method. MTE is a quicker and more cost-effective option to replace qPCR and Next Generation Sequencing in quantifying miRNA expression. As a fail-fast method, it could be easily used at numerous points during an upstream pharmaceutical manufacturing process without a large time or capital investment. It was observed that secondary structure of short nucleic acids played a large role in their detection strategy using MTE. Unstructured short nucleic acids could be detected using DNA probes; however, structured targets needed the stronger probe chemistry of LNA to overcome their self-complementarity. The ideal substitution percentage for LNA probes was found to be in the 30% range to optimize binding and prevent aggregation.